Kraus

and
Carver

ELECTROMAGNETICS
SECOND EDITION
Kraus and Carver
Ht/'W

SECOND
EDITION

MtGRAW-HIU
KOGAKUSHA

'

''

'

'

'

SYMBOLS, PREFIXES,

AND ABBREVIATIONS
See Appendixes A-l (Units) and A-2 (constants and conversions) for more detailed information

A
A

ampere

K,

angstrom = 10" 10

K,k

astronomical unit

atto=10"' 8

l,L

length (scalar),

vector potential,
area,

AR

axial ratio

AU
a

Wb m"

susceptance,

U m"

nulls

capacitance,

cc

capacitance/unit length, F m"

a constant, c = velocity of light
cubic centimeter

degree Celcius

D,

electric flux density,

Cm"

distance,

deg

degree, angle

dB

decibel

dBi

decibels over isotropic

di

element
element
element
element
element

d\

ds

ds

dv
E,

emf

.-:

F,

10 log(/' 2 //' 1 )

m
m
m

of length (scalar),
of length (vector),
of surface (scalar),
of surface (vector),
of volume (scalar),

electric field intensity,

electromotive force,
electric charge, C
farad

<u

emf (electromotive

force),

velocity,

s"

energy, J

log

common

Wb

weber

energy density,

X
X

reactance,

x
x

unit vector in

M,

m,

polarized

logarithm (base 10)

mega = 10 6
-1
magnetization,

Am

polarization state

M(e,

t)

m
m

magnetic (dipole) moment,

meter
3
milli = 10"

min

minute

N
N,n

newton
number

Np

neper

nano

unit vector

A m2

Y
Y
y

(integer)

10" 9

z
z
z

normal to a surface

Cm"
Cm"

polarization of dielectric,

moment,

polarization state

P
P

power,

z<

P(y, o)

power per solid angle,

rad
power pattern, dimensionless

-P.

zL
7

"2

A
10* gauss

Jm' 3

reactance/unit length,

Q m"'

coordinate direction
admittance, CI
admittance/unit length, Q
unit vector in y direction
coordinate direction

m"

impedance,
impedance/unit length, Q m"
intrinsic impedance, conductor, O
per square
CI
intrinsic impedance, dielectric,
per square
load impedance, Cl
transverse impedance, rectangular
waveguide, CI
transverse
impedance, cylindrical
waveguide, H
intrinsic impedance, space, O per
square
characteristic impedance, transmis-

revolution

coordinate direction, also red shift

(alpha) angle (deg or rad)
(beta) angle, deg or rad; also phase

gram

radius,

henry
- '
magnetic field,
half-power beam width
hertz = 1 cycle per second

unit vector in r direction

rad
rad 2

radian
square radian = steradian
"2
Poynting vector,

Am

current,

joule

current density,
kelvin

Am

Urn -1

S,

m;

Wm

distance,

second (of time)

steradian = square radian

m;

= sr

resistivity,

S,s
sr

sion line,

also coordinate direction

S
-2

also surface area,

= rad 2

(xi)

m"'

'

'

(pi)

(rho) electric charge density,

also

m"

(nu)

Ci

resistance/unit length, fi

v
l

right elliptical polarization

Zo

Wb

linkage,

flux

right circular polarization

10 9

lambda)

(capital

Mo

direction

RCP
REP

(lambda) wavelength,
(mu) permeability, H m"
permeability of vacuum, H
2
(mu) mobility,
s" V"
"
(mu) micro = 10 6

unit vector in z direction

Zo

turn

m"

(kappa)

watt

H~

N m"

(theta) unit vector in 6 direction

reluctance,

Zti,

m"

of refraction
(theta) angle, deg or rad

<M

permeance,

F m~

vacuum, F

(eta) index

w
w

left elliptically

electric dipole

Hz

giga

permittivity of

resistance,

N m~

conductance, U
conductance/unit length,

J,/

voltage (also emf),

(epsilon) tensor permittivity,

natural logarithm (base e)

15

volt

deg or rad

(epsilon) permittivity,

R
R

10"

G
G

V
V

= 2-njX

angle, deg or rad

(delta) angle,

charge,

flux unit

Hz
I /,

magnetostatic potential,

Nm

e,<?

femto

frequency,

torque,

constant

(gamma)

fu

HPBW

time, s

tera=

In

10'

pico= 10"' 2

H,

circularly polarized

tesla

radiation pressure,

force,

force per volume,

left

V m"

m
length (vector), m

LCP
LEP

P,
2

H m"

= Wb m"

T
T
T

&
&

coulomb

C,c

fe

Wb m~

BWFN beam width, first

1!

T=

susceptance/unit length,

C
c
c

magnetic flux density,

B
B
/

k
kg

'

unit vector

Am -1

L
L

B,

sheet-current density,

a constant
3
kilo = 10
kilogram
inductance,
inductance/unit length,

A, a

3.14

mass

density,

kg

Cm"

3
;

" 3

reflection coefficient, dimensionless

P>

surface charge density,

PL

linear charge density,

Cm"
-1

Cm

(sigma) conductivity, O m"

(tau) tilt angle, polarization ellipse,
deg or rad
transmission coefficient, deg or rad
(phi) angle, deg or rad
(phi) unit vector in 4> direction
(chi) susceptibility, dimensionless
(psi) angle, deg or rad
1

4>

x
P

n
n
u

magnetic flux, Wb
omega) ohm
2
(capital omega) solid angle, sr or deg
(upsidedown capital omega) mho
(capital

(U=

1/0

= S,

Siemens)

(omega) angular frequency

rad

s "

2nf),

McGRAW-HILL ELECTRICAL AND ELECTRONIC

ENGINEERING SERIES

Koenig, Tokad, and Kesavan Analysis of Discrete Physical Systems

Kraus: Antennas
Kraus and Carver Electromagnetics
Kuh and Pederson: Principles of Circuit Synthesis
Kuo: Linear Networks and Systems
Ledley: Digital Computer and Control Engineering
LePage: Complex Variables and the Laplace Transform for Engineering
LePage and Seely: General Network Analysis
Levi and Panzer: Electromechanical Power Conversion
Ley, Lutz, and Rehberg: Linear Circuit Analysis
Linvill and Gibbons: Transistors and Active Circuits
:

Frederick Emmons Terman, Consulting Editor

W. W. Harman, J. G. Truxal, and R. A. Rohrer, Associate Consulting Editors

Angelakos and Everhart: Microwave Communications

Angelo: Electronic Circuits
Angelo: Electronics: BJTs, FETs, and Microcircuits
Aseltine: Transform Method in Linear System Analysis
Atwater: Introduction to Microwave Theory

Littauer: Pulse Electronics

Belove, Schachter, and Schilling: Digital and Analog Systems, Circuits, and Devices:

An

Introduction
Bennett: Introduction to Signal Transmission

Beranek: Acoustics
Bracewell: The Fourier Transform and

Its

Meishl: Principles of Electromechanical-energy Conversion

Application

Brenner and Javid: Analysis of Electric Circuits

Carlson: Communication Systems: An Introduction
Communication
Chen: The Analysis of Linear Systems
Chen: Linear Network Design and Synthesis
Chirlian: Analysis and Design of Electronic Circuits
Chirlian: Basic Network Theory

to Signals

and Noise

Chirlian: Electronic Circuits: Physical Principles, Analysis, and Design

Chirlian and Zemanian: Electronics
Clement and Johnson: Electrical Engineering Science
Cunningham: Introduction to Nonlinear Analysis
D'Azzo and Houpis: Feedback Control System Analysis and Synthesis
Elgerd: Control Systems Theory
Elgerd: Electric Energy Systems Theory: An Introduction
Eveleigh: Adaptive Control and Optimization Techniques
Eveleigh: Introduction to Control Systems Design
Feinstein: Foundations of Information Theory
Fitzgerald, Higginbotham, and Grabel: Basic Electrical Engineering
Fitzgerald, Kingsley, and Kusko: Electric Machinery
Electrical Measurement Analysis
Friedland, Wing, and Ash: Principles of Linear Networks
Gehmlich and Hammond: Electromechanical Systems
Ghausi: Principles and Design of Linear Active Circuits
Ghose: Microwave Circuit Theory and Analysis
Glasford: Fundamentals of Television Engineering
Greiner: Semiconductor Devices and Applications
Hammond and Gehmlich: Electrical Engineering

Frank:

Hancock: An Introduction to the Principles of Communication Theory

Harman: Fundamentals of Electronic Motion

Harman Principles of the Statistical Theory of Communication

Harman and Lytle: Electrical and Mechanical Networks
:

Hayashi: Nonlinear Oscillations in Physical Systems

Hayt: Engineering Electromagnetics
Hayt and Kemmerly: Engineering Circuit Analysis
Hill: Electronics in Engineering
Javid and Brown Field Analysis and Electromagnetics
Johnson: Transmission Lines and Networks
:

X ^V

Lynch and Truxal: Introductory System Analysis

Lynch and Truxal: Principles of Electronic Instrumentation
Lynch and Truxal: Signals and Systems in Electrical Engineering
McCluskey: Introduction to the Theory of Switching Circuits
Manning: Electrical Circuits

in Electrical

Millman and Halkias: Electronic Devices and Circuits

Millman and Halkias: Integrated Electronics: Analog and Digital
Millman and Taub: Pulse, Digital, and Switching Waveforms

Circuits

and Systems

Minorsky: Theory of Nonlinear Control Systems

Mishkin and Braun: Adaptive Control Systems
Moore: Travelling-wave Engineering
Murdoch: Network Theory
Oberman: Disciplines in Combinational and Sequential Circuit Design
Pettit and McWhorter: Electronic Amplifier Circuits
Pettit and McWhorter: Electronic Switching, Timing, and Pulse Circuits
Reza: An Introduction to Information Theory
Reza and Seely: Modern Network Analysis
Ruston and Bordogna: Electric Networks: Functions, Filters, Analysis
Ryder: Engineering Electronics
Schilling and Belove: Electronic Circuits: Discrete and Integrated
Schwartz: Information Transmission, Modulation, and Noise
Schwartz and Friedland: Linear Systems
Seeley: Electromechanical Energy Conversion
Seifert and Steeg: Control Systems Engineering
Shooman: Probabilistic Reliability: An Engineering Approach
Siskind: Direct-current Machinery
Skilling: Electric Transmission Lines
Stevenson: Elements of Power System Analysis
Stewart: Fundamentals of Signal Theory
Strauss: Wave Generation and Shaping
Su: Active Network Synthesis
Taub and Schilling: Principles of Communication Systems
Terman: Electronic and Radio Engineering
Terman and Pettit: Electronic Measurements
Thaler and Pastel: Analysis and Design of Nonlinear Feedback Control Systems
Tou Digital and Sampled-data Control Systems
Tou Modern Control Theory
Truxal: Automatic Feedback Control System Synthesis
Tuttle: Electric Networks: Analysis and Synthesis
Valdes: The Physical Theory of Transistors
Weeks: Antenna Engineering
Weinberg: Network Analysis and Synthesis
:

McGRAW-HILL
KOGAKUSHA, LTD.
Tokyo
Auckland
Dusseldorf

Johannesburg

London
Mexico

New

Delhi

Panama
Sao paulo
Singapore
Whistling swan family with father, right, and mother, center, both equipped with
migration transmitter and wrap-around dipole antenna. Juvenile swan or cygnet
is at left.
The units radiate pulses of 1-mW peak power at 222
and have a
100-km operating range with four months of service life. The units weredeveloped
to study flight pattern, flight altitude, air temperature, heart rate, and body-core
temperature during migrations. These magnificent 10-kg (22-lb) birds make
14,000-km round trips each year between Chesapeake Bay and Alaska. Typically
they fly in formation for distances of 1000 km at altitudes of 3 km in nonstop
flights lasting 1 hours (average speed 1 00 km/h).
The tracking aircraft following
such a flight must land twice to refuel. The transmitter, including batteries,
weighs only 90 g (3 oz) and is secured to the upper middle back by four teflon
straps which go under the wings and around the body.
flexible wire dipole
antenna runs part way along two of the straps. The system does not interfere
with the swan's activities in flight or on the water, and is almost invisible beneath
the feathers. In the photograph the transmitter cases have been retouched for
clarity and the location of the dipoles is indicated by dashed lines.
The swan at

MHz

Sydney
INTERNATIONAL STUDENT EDITION

the right also has a large plastic neck band for visual identification. The asymmetrical mounting of the dipole antenna results in vertically polarized radiation
when the swan is on the water, and in horizontally polarized radiation when in

change from

monitoring
(Photograph by John W. Hamblen,
Applied Physics Laboratory, Johns Hopkins University. See W. A. GoodandJ. W.
Hamblen, APL Technical Digest, vol. 10, no. 2, pp. 2-10, Nov.-Dec. 1970.)
flight.

station that the

vertical to horizontal polarization alerts the

swan has become airborne.

JOHN

D.

KRAUS

Taine G. McDougal Professor of Electrical

Engineering and Astronomy
Director, Radio Observatory
The Ohio State University

KEITH

R.

CARVER

Associate Professor of Electrical Engineering

Senior Engineer, Physical Science Laboratory

New Mexico

State University

ELECTROMAGNETICS
SECOND EDITION

'1

McGRAW-HILL
KOGAKUSHA, LTD.
Tokyo
Auckland
Diisseldorf

Johannesburg

London
Mexico

New

Delhi

Panama
Sao paulo
Singapore

Sydney
INTERNATIONAL STUDENT EDITION

JOHN

D.

KRAUS

Taine G. McDougal Professor of Electrical

Engineering and Astronomy
Director, Radio Observatory
The Ohio State University

KEITH

R.

CARVER

Associate Professor of Electrical Engineering

Senior Engineer, Physical Science Laboratory

New Mexico

State University

ELECTROMAGNETICS
SECOND EDITION

OTHER McGRAW-HILL BOOK COMPANY BOOKS BY JOHN D. KRAUS

1950
Antennas
Radio Astronomy

1966

Library of Congress Cataloging in Publication Data

Kraus, John Daniel, 1910Electromagnetics.

(McGraw-Hill

electrical

and

electronic engineering

series)

Bibliography: p.
Electromagnetic theory. I. Carver, Keith R.,
1.
Title.
II.
1940joint author.
72-13520
530.1 '41
QC661.K72 1973

ISBN 0-07-035396-4 /

ELECTROMAGNETICS
INTERNATIONAL STUDENT EDITION
for
rights by McGraw-Hill Kogakusha, Ltd,,
manufacture and export. This book cannot be re-exported
from the country to which it is consigned by McGraw-Hill.
Exclusive

1973 by McGraw-Hill, Inc. All rights reserved.

Copyright
Copyright 1953 by McGraw-Hill, Inc. All rights reserved.
No part of this publication may be reproduced, stored in a
retrieval System, or transmitted, in any form or by any
means, electronic, mechanical, photocopying, recording,
or otherwise, without the prior written permission of
the publisher.

KOSA1DO PRINTING

CO.. LTD.

TOKYO JAPAN

CONTENTS

Preface

..

xvii
1

Introduction

/-/

Electromagnetics

1-2

Dimensions and Units

Fundamental and Secondary Units
How to Read the Symbols and Notation
Equation Numbering

1-3

1-4
1-5
1-6

Denned

1-7

Dimensional Analysis
Thumbnail Electromagnetics

1-8

SPEMP Chart

2
2-1

2-2
2-3

The

Static Electric Field: Part 1

12

Introduction

The Force between Point Charges and Coulomb's Law

and Staticness

Idealness

2-4

Electric Field Intensity

2-5

Positiveness, Right-handedness,

and Outwardness

yy

CONTENTS

V1U

The

2-6

Electric Field of Several Points

Charges and the Principle of

Superposition of Fields

The
The

2-7
2-8

2-9

Electric Scalar Potential

Electric Scalar Potential as

Relation of Electric

Field

a Line Integral of the

Electric Field

Lines and Equipotential Contours;

Orthogonality
2-10 Field of

Two

Equal Point Charges of Opposite Signs and of Same

Sign
2-11 Charge Density and Continuous Distributions of Charge
2-12 Electric Potential of Charge Distributions and the Principle of

Superposition of Potential
2-13

The

Electric Field as the Gradient of the Electric Potential

2-14 Gradient in Rectangular Coordinates

2-15 The Electric Dipole and Electric-dipole

Moment

2-16 Electric Flux

2-77 Electric Flux over a Closed Surface; Gauss'

Law

2-18 Single Shell of Charge

2-19 Conductors and Induced Charges
2-20 Conducting Shell
2-21 Capacitors and Capacitance

2-22 Boundary Relations at a Conducting Surface

Problems
3

The

Static Electric Field: Part

3-1

Introduction

3-2

Homogeneity, Linearity, and Isotropy

3-3

Dielectrics

3-4

The

and Permittivity

Electric Field in a Dielectric

3-5

Polarization

3-6

Artificial Dielectrics

3-7
3-8

Boundary Relations
Table of Boundary Relations

3-9

Parallel-plate Capacitor

3-10 Dielectric Strength

3-11 Energy in a Capacitor

3-12 Energy Density in a Static Electric Field

3-13 Fields of Simple Charge Configurations
3-14 Field Distributions

3-15 Field of a Finite Line of Charge

3-16 Field of an Infinite Line of Charge

56

CONTENTS

IX

3-17 Infinite Cylinder of Charge

3-18 Infinite Coaxial Transmission Line
3-19

Two

Infinite Lines

of Charge

3-20 Infinite Two-wire Transmission Line

3-21 Infinite Single-wire Transmission Line

Mapping of

3-22 Graphical

Static Electric Fields; Field Cells

3-23 Divergence of the Flux Density

3-24 Maxwell's Divergence Equation

3-25 Divergence Theorem

3-26 Divergence of

D and P in a Capacitor

3-27 The Laplacian Operator; Poisson's and Laplace's Equations

N

Problems

The Steady

Electric Current

4-1

Introduction

4-2

Conductors and Insulators

4-3

The

4-4

4-7

Ohm's Law
Power Relations and Joule's Law
The Electric Circuit
Resistivity and Conductivity

4-8

Table of Conductivities

4-9

Current Density and Ohm's

4-5
4-6

109

Electric Current

Resistance and

4-10 Kirchhoff's Voltage

and

Law
Law and the

at

a Point

Difference between Potential

Emf

4-11 Tubes of Current

4-12 Kirchhoff's Current

Law

4-13 Divergence of J and Continuity Relations for Current

4-14 Current and Field at a Conductor-Insulator Boundary

4-15 Current and Field at a Conductor-Conductor Boundary

4-16 Current Mapping and the Resistance of Simple Geometries;

Conductor

Cells

4-17 Laplace's Equation for Conducting Media

Problems
5

The

Magnetic Field of Steady Electric Currents

Static

on a Magnet
Magnet on a Current-carrying Wire

5-1

Introduction; Effect of a Current

5-2

Effect of a

5-3

The Magnetic
Savart

Law

Field of a Current-carrying Element;

140

The

Biot-

i
.1

5-5

The Magnetic Field of an Infinite Linear Conductor

The Force between Two Parallel Linear Conductors; Definition

5-6

The Magnetic

5-7

Magnetic Flux

5-8

Magnetic Flux over a Closed Surface

5-9

Magnetic Field Relations in Vector Notation

5-4

of the

Ampere
Field of a Current-carrying
\j/

Loop

m and Magnetic Flux Density

5-10 Torque on a Loop; Magnetic

'

Moment

The Solenoid
5-12 Inductors and Inductance

5-11

5-13 Inductance of Simple Geometries

5-14 Ampere's
5-15 Ampere's

Law and H
Law Applied

to a Conducting

Medium and Maxwell's

Equation
5-16 Magnetostatic Potential

U and Mmf F

5-17 Field Cells and Permeability

5-18 Energy in an Inductor
5-19 Energy Density in a Static Magnetic Field

5-20 Magnetic Poles

5-21 Curl

5-22 Maxwell's First Curl Equation

5-23

Summary

of Operations Involving

Comparison of Divergence and Curl

5-25 The Vector Potential
5-26 A Comparison of Static Electric and Magnetic Fields
5-24

Problems

The

Static

Magnetic Field of Ferromagnetic Materials

6-1

Introduction

6-2

Bar Magnets and Magnetic Poles

6-3

Magnetic Materials

6-4

Relative Permeability

6-5

Magnetic Dipoles and Magnetization

6-6

Uniformly Magnetized Rod and Equivalent Air-filled Solenoid

The Magnetic Vectors B, H, and
Boundary Relations
Table of Boundary Relations for Magnetic Fields

6-7
6-8

6-9

6-10 Ferromagnetism
6-11 Magnetization Curves

198

CONTENTS

Xl

6-12 Hysteresis
6-13 Energy in a Magnet
6-14 Permanent Magnets
6-15 Table of Permanent Magnetic Materials

6-16 Demagnetization

6-17 The Magnetic Circuit; Reluctance and Permeance

6-18 Magnetic Field Mapping; Magnetic Field Cells

6-19 Comparison of Field

Maps

in Electric, Magnetic,

and Current

Cases
6-20 Gapless Circuit
6-21 Magnetic Circuit with Air

6-22 Magnetic

Gap

Gap Force
Gap

6-23 Permanent Magnet with

6-24 Comparison of Electric and Magnetic Relations Involving

Polarization and Magnetization

Problems

Laplace's and Poisson's Equations and Boundary-value Problems

7-1

Introduction

7-2

Solution of Laplace's Equation in Rectangular Coordinates

7-3

Example

7-4

Uniqueness

7-5

Point-by-point, or Iterative,

7-6

The

Parallel-plate Capacitor

7-7

Example
Example

7-8

Digital-computer Solution

7-9

Analog-computer Solution

7-70

Example

Method

2:

The

3:

Square Trough with Different Potential on Each Side

4:

Infinite

Square Trough

Conducting Sheet between

Two Conducting

7-11 Solution of Laplace's Equation in Cylindrical

Coordinates
7-12

Example

5:

262

Planes

and Spherical
i

Coaxial Line

7-13 Poisson's Equation

Example 6: Parallel-plate Capacitor with Space Charge

The Theory of Images
7-16 Example 7: Charged Cylindrical Conductor over an Infinite Metallic
Ground Plane
7-17 Example 8: Current-carrying Conductor over Infinite Metallic
Ground Plane
7-14
7-15

Problems

CONTENTS

Xii

Time-changing Electric and Magnetic Fields

305

8-1

Introduction

8-2

Faraday's

8-3

Maxwell's Equation from Faraday's Law: Integral

8-4

Moving Conductor

8-5

General Case of Induction

8-6

Examples of Induction

8-7

Stokes'

8-8

Maxwell's Equation from Faraday's Law: Differential

8-9

Mutual Inductance and Self-inductance

8-10

The Transformer

Law
in

Form

a Magnetic Field

Theorem

Form

8-11 Alternating-current Behavior of Ferromagnetic Materials

8-12 Eddy Currents

8-13 Displacement Current
8-14 Maxwell's Equation from Ampere's

Law: Complete Expression

8-15 Dielectric Hysteresis

8-16 Boundary Relations

8-17 General Field Relations

8-18 Comparison of Electric and Magnetic Field Relations

Problems

The Relation between Field and Circuit Theory;

Maxwell's Equations

9-1

Introduction

9-2

Applications of Circuit and Field Theory

9-3

The

9-4

Maxwell's Equations as Generalizations of Circuit Equations

9-5

Maxwell's Equations in Free Space

Series Circuit;

348

Comparison of Field and Circuit Theory

9-6

Maxwell's Equations for Harmonically Varying Fields

9-7

Tables of Maxwell's Equations

Problems

10

Plane Waves in Dielectric and Conducting Media

10-1

Introduction

10-2

Plane Waves and the

10-3

Solutions of the

10-4

Table

Wave Equation
Wave Equation
of Solutions of the Wave Equation

10-5

Phase Velocity

10-6

Index of Refraction

10-7

Group

Velocity

363

CONTENTS

10-8
10-9

10-10

Xlii

Impedance of Dielectric Media

Impedance of Transmission-line Cell

Two

Plane Waves Traveling in Opposite Directions; Standing

Waves
10-11 Energy Relations in a Traveling

Wave

10-12 The Poynting Vector

10-13 Energy Relations in a Standing

Wave

10-14 Conductors and Dielectrics

10-15

Wave Equation

for Conducting

Media

10-16 Depth of Penetration

10-17 Relaxation Time

10-18 Impedance of Conducting Media

10-19 The Poynting Vector in Conducting Media

10-20 Circuit Application of the Poynting Vector

10-21 General Development of the

Wave Equation

Problems
11
//-/

Wave

Polarization

426

Introduction

1 1-2

Linear, Elliptical, and Circular Polarization

11-3

Poynting Vector for Elliptically or Circularly Polarized Waves

11-4

The

11-5

Partial Polarization

11-6

Cross Field

Polarization Ellipse and the Poincare Sphere

and the Stokes Parameters

Problems

12

Wave

Reflection, Refraction,

12-1

Introduction

12-2

Plane Wave,

and

Diffraction

12-3

Normal Incidence
The Terminated Wave

12-4

Linearly Polarized Plane Wave, Oblique Incidence

12-5

Elliptically Polarized

12-6

Plane Wave, Oblique Incidence

Huygens' Principle, and Physical Optics

12-7

Geometrical-optics Concepts

12-8

Scattering from a Conducting Strip

12-9

Geometrical Theory of Diffraction

445

Problems
13
13-1

Transmission Lines, Waveguides, and Resonators

Introduction

482

MV

CONTENTS

13-2

Coaxial, TWo-wire, and Infinite-plane Transmission Lines

13-3
13-4

The Infinite Uniform Transmission Line

Comparison of Circuit and Field Quantities

73-5

Characteristic-impedance Determinations

13-6

The Terminated Uniform Transmission Line

13-7

Transmission-line Charts

13-8

A/4 Transformer

13-9

Transformer Bandwidth

13-10

Wave

Reflections

on a 2/4 Transformer

13-11 Further Time-domain Studies of X/4 Transformer

13-12 Power Flow on a Transmission Line

13-13 Circuits, Lines, and Guides

13-14

TE Mode Wave

Comparison

in the Infinite-parallel-plane

Transmission Line

The Hollow Rectangular Waveguide

13-16 The Hollow Circular Cylindrical Waveguide
13-17 Hollow Waveguides of Other Cross Section

13-15

13-18 Attenuation at Frequencies Less than Cutoff

13-19 Attenuation at Frequencies Greater than Cutoff

13-20 Waveguide Devices

13-21 Waveguide

Iris

Theory

13-22 Intrinsic, Characteristic,

and Wave Impedances

Traveling to a Plane Boundary

13-23

Waves

13-24

Open Waveguides

13-25 Cavity Resonators

Problems

14
14-1

Antennas and Radiation

Introduction

14-2

Retarded Potentials

14-3

The Short Dipole Antenna

14-4

Radiation Resistance of a Short Dipole

14-5

Directivity, Gain,

14-6

Array Theory

and

Effective Aperture

14-6.1

Two

14-6.2

Pattern Multiplication

14-6.3

Binomial Array

14-6.4

Array with n Sources of Equal Amplitude and Spacing

Array with n Sources of Equal Amplitude and Spacing:

14-6.5

Isotropic Point Sources

Broadside Case

601

CONTENTS

14-6.6

'

XV

Array with n Sources of Equal Amplitude and Spacing:

End-fire Case

.14-6.7

14-6.8

Graphical Representation of Phasor Addition of Fields

Simple Two-element Interferometer

14-7

Continuous Aperture Distribution

14-8

Fourier Transform Relations between the Far-field Pattern and

the Aperture Distribution

14-9

Linear Antennas

14-10 Fields of A/2 Dipole Antenna

14-11 Traveling-wave Antennas

14-12 The Small

Loop Antenna

14-13 The Small Helical Antenna

14-14

14-15

The Helical-beam Antenna

Beam Width and Directivity of Arrays

14-16 Scanning Arrays

14-17 Frequency-independent Antennas
14-18 Reciprocity
14-19 Self and Mutual Impedance and Arrays of Dipoles
14-20 Reflector and Lens Antennas
14-20.1

Infinite-flat-sheet Reflector

14-20.2

Finite-flat-sheet Reflector

14-20.3

Thin Reflectors and Directors

14-20.4

Corner Reflectors

14-20.5

Parabolic Reflectors and Lens Antennas

14-20.6

Some Comments on Corner

Reflectors vs. Parabolic

Reflectors

Complementary Antennas
Horn Antennas

14-21 Slot and

14-22

14-23 Aperture Concept

14-24 Coherence
14-25 Friis Formula and Radar Equation

14-26 Radio Telescopes, Noise Power, and Antenna Temperature

14-27 Rotating Bar Magnet as a Very-low-frequency Radiator (Pulsar)
14-28 Antennas for Polarization Measurements

Problems
15

Particles

<

and Plasmas

15-1

Introduction

15-2

Charged

Particle in a Static Electric Field

716

XVI

CONTENTS

15-3

Charged

15-4

The Cathode-ray Tube

The Cyclotron
The Mass Spectrograph
Table of Charge and Mass for Common Particles
Plane Waves in an Ionized Medium in the Presence of a Magnetic

15-5
15-6

15-7
15-8

Particle in

a Static Magnetic Field

Field

15-9

Faraday Rotation

15-10 Magnetohydrodynamic

Waves

Problems
16
16-1

Moving Systems and Space-Time

749

Introduction

16-2

Simultaneity

16-3

Space-Time Concept

16-4

Coordinate Transformation

16-5

Invariance in Space-Time

16-6

Four-dimensional Field Formulation

16-7

Fields of a

16-8

Maxwell's Stress Tensor

Moving System

Problems

Appendix

Units, Constants,

and a Few Mathematical

785

Relations
A-l

Units

A-2

Constants and Conversions

A-3

Trigonometric Relations

A-4

Hyperbolic Relations

A-5

Logarithmic Relations

A-6
A-7

Series

Approximation Formulas for Small Quantities

A-8

Solution of Quadratic Equation

A-9 Vector Identities
A-IO Recurrence Relations for Bessel Functions
A-ll Coordinate Diagrams

Appendix

Bibliography

801

Appendix

Answers to Starred Problems

803

Index

813

PREFACE

The wide and continued use of the first edition of Electromagnetics, including a foreign
translation, has given us the incentive to prepare a second edition.

Many

users have

told us of features they wish to have retained in a

new edition. But new advances

and changes in emphasis and teaching methods have made the addition of new material
and revision of the old desirable. The aim of the book, as before, is to present the
basic elements of electromagnetic theory for an introductory fields course. The topics
have been selected from a wide variety of subjects, but these have been chosen to illustrate important concepts without becoming encyclopedic.
As prerequisites the stu-

dent

is

assumed to have a knowledge of introductory physics and mathematics through

and integral calculus. Also a course in vector analysis is essential either

differential

beforehand or concurrently.

The introduction (Chap. 1) includes sections on units and dimensions, dimenand two teaching aids: Thumbnail Electromagnetics and SPEMP chart.
Modernized metric (SI) units and nomenclature are used throughout the book. The
five chapters (2 to 6) on static electric and magnetic fields and steady currents form the
sional analysis

foundation of field theory.

time-changing

well's equations.

Three chapters (7 to

9) discuss

boundary- value problems,

and the relation of field and circuit theory, culminating in MaxProblem solutions by a variety of methods including analytical,

fields,

PREFACE

XViU

and computer (both digital and analog) are discussed. Chapters 10 to 12

wave reflection,
cover waves in dielectric and conducting media, wave polarization and
and
(wavefront)
optics
Treatments using both physical
refraction, and diffraction.
explain
and
14
geometrical optics (ray-path) methods are included. Chapters 13
graphical,

transmission

lines,

waveguides, resonators, antennas, and radiation.

New sections

in

reflectometry, aperture
these chapters include such topics as bandwidth, time-domain
radio telescopes.
concepts, array theory, radar equation, antenna temperature and
(or relaspace-time
and
Chapters 15 and 16 treat particles, plasmas, moving systems,

Topics in these chapters include particles in static electric and magnetic fields,
wave propagation in a magnetized plasma, Faraday rotation, magnetohydrodynamic
(MHD) waves, the space-time concept, fields of a moving system, and radiation preselectric or magnetic field which
It is shown that there is no such thing as a pure
sure.

tivity).

on outer space
Appendix
and systems in relative motion, these ideas are of fundamental importance.
A contains tables of important constants, equations, and formulas. There is also a
Appendix B is a brief biblivery complete tabulation of units and their equivalents.

retains

identity for all observers.

its

ography, and Appendix

listing

With

the increasing emphasis

C gives answers to the starred problems.

There

is

a complete

relations
of symbols inside the front cover and some frequently used vector

back cover.
Almost half of the illustrations are new. The problem sets, which are completely
Many extend or
revised and expanded, are an important feature of the new edition.
are many which
included
problems
of
variety
wide
the
Among
text.
supplement the

inside the

pertain to

modern

real-world engineering situations,

e.g.,

problems involving a study

and/or the design of an operational device. Some of these

of
solutions (few real-world problems have exact answers).
indefinite
or
have multiple
There are also many
class discussion or term papers.
for
used
can
be
Such problems
engineering feasibility

problems of the more conventional type, yielding

gradations of

difficulty.

A considerable

definite numerical

number of problems

answers with

all

are arranged for solu-

by computers. Answers are given to one-half of the problems. There is much

intends to solve
to be gained by becoming familiar with the problems, whether one
them or not. Many topics are included which are not discussed elsewhere in the
tion

book and some very interesting and thought-provoking ideas are presented. We
recommend "problem reading for pleasure."
The worked examples, numbering about 100, are valuable instructional aids.
These examples are of great assistance in understanding the theory and how to
apply

it

to practical situations.

The book
theory.

There

is
is

designed to provide

flexibility for

course needs in introductory

adequate material for a one-year course. The

For example, a short course can cover the

adapted for shorter courses.

material

is

field

also well
first six,

omission of
seven, eight, nine or ten chapters either completely or with the

all

or

PREFACE

XIXi

part of the following sections: 3-14 through 3-22; 4-14; 4-15; 4-16; 5-25, 5-26; 6-12

through 6-24; 7-10 through 7-17; 10-10; and 10-13 through

for a short course

is

to omit all or

most of these

10-21.

Another option

sections but to include parts of Chaps.

13 to 16; for example, Sees. 13-1 through 13-6 (transmission lines); 14-1, 14-2, 14-5,
14-6,

and

and 14-7 (antenna aperture and array theory); 15-2 and 15-3 (charged particles);
Antennas are remarkable
Still other options are possible.

16-2 (simultaneity).

and space, and some familiarity with them

The antenna sections (14-1, 14-2, 14-5, 14-6,
and 14-7), indicated above, use nothing more advanced than phasor addition.
Although great care has been exercised, some errors in the text or figures are
inevitable.
Anyone finding them will do us a great service by writing us so that they

in that they can interface between a circuit

is

desirable even in a short fields course.

can be corrected in subsequent printings.

We

many teachers and a generation of students who have used

and have generously supplied us with comments and suggestions.
Professor Charles F. Fell, University of Nevada, made many recommendations and is
responsible for the definitions of idealness, staticness, etc., used in Sees, 2-3 and 2-5.
the

first

are grateful to

edition

Professor Clarence

W.

Schultz, University of Connecticut, offered

numerous sugges-

and supplied the method used with Fig. 6-6 for developing the relation between
B, H, and M. We are much indebted to Professors Chen-To Tai, University of Michi-'
gan, and John N. Cooper, Navy Post Graduate School and Professors Louis L. Bailin,
tions

John D. Cowan,

Jr;

Daniel B. Hodge, Hsien Ching Ko, Curt A. Levis, Edward

M.

Kennaugh, Robert G. Kouyoumjian, William H. Peake, Leon Peters, Jr., Jack H.

Richmond, and Thomas A. Seliga of The Ohio State University. The suggestions by
several anonymous reviewers chosen by the publisher were also most useful.
Finally,

one of us

for her assistance

the years of

its

(J.

D. K.), expresses

on the manuscript and

his sincere appreciation to his wife, Alice,

for her patience

and encouragement during

preparation.

JOHN
KEITH

KRAUS
CARVER

D.
R.

INTRODUCTION

ELECTROMAGNETICS DEFINED

1-1

Electromagnetics embraces electricity, magnetism, electric fields, magnetic

fields, and
electromagnetic waves. In circuit theory the emphasis is on the voltage
at a pair of
terminals and the current through a conductor, but in electromagnetics
the emphasis

on the space between conductors and the electric and magnetic fields in this
The field approach of electromagnetics is essential for an understanding of

is

guides, antennas, waves in space,

and

particle-field interactions.

also provides a basic insight into the operation

circuit

and

space.

wave-

Electromagnetics

characteristics of such basic

elements as capacitors, inductors, and resistors.

The title of this book could

have been " Electric and Magnetic Fields," "Fields and Waves," ox "
Electromagnetic
Theory and Applications." All this and more is meant by the word Electromagnetics,
which was used as the title for the first edition of this book (1953) and is used

again as

the

title

for this edition.

'

v^

[CHAP.

ELECTROMAGNETICS

DIMENSIONS AND UNITS

1-2

When you can measure what you are speaking about

Lord Kelvin is reported to have said
about it; but when you cannot measure it,
something
know
numbers
you
in
and express it
when you cannot express it in numbers your knowledge is of a meagre and unsatisfactory
kind it may be the beginning of knowledge but you have scarcely progressed in your thoughts
:

to the stage of science whatever the matter

may

be.

might be added that before we can measure something, we must define its
dimensions and provide some standard, or reference unit, in terms of which the quantity

To

this

it

can be expressed numerically.

dimension defines some physical characteristic. For example, length, mass,
The dimensions of length, mass, time,
time, velocity, and force are dimensions.
are considered as the fundamental
intensity
luminous
and
electric current, temperature,

dimensions since other dimensions can be defined in terms of these six. This choice
represent the dimenLet the letters L, M, T, I, F, and
is arbitrary but convenient.

sions of length, mass, time, electric current, temperature,

and luminous

intensity.

Other dimensions are then secondary dimensions. For example, area is a secondary
dimension which can be expressed in terms of the fundamental dimension of length
2

squared (L

of force are

).

As other examples, the fundamental dimensions of velocity are LfT and

ML/T 2

A unit is a standard or reference by which a dimension can be expressed numera unit in terms of which the dimension of length can be
expressed, and the kilogram is a unit in terms of which the dimension of mass can be
For example, the length (dimension) of a steel rod might be 2 meters and
expressed.
ically.

its

Thus, the meter

is

mass (dimension) 5 kilograms.

1-3

The

FUNDAMENTAL AND SECONDARY UNITS

units for the fundamental dimensions are called the fundamental or base units.

In this book the International System of Units, abbreviated SI, is used.f In this
system the meter, kilogram, second, ampere, kelvin, and candela are the base units for
the six fundamental dimensions of length, mass, time, electric current, temperature,
and luminous intensity. The definitions for these fundamental units are

Length equal to

Meter (m)

,650, 763.73

wavelengths in vacuum corresponding

to the 2p 10 -5ds transition of krypton 86.

The International System of Units is the modernized version of the metric system.
The abbreviation SI is from the French name Systeme Internationale d'Unites.
For the complete
Pub. 330, 1971.

official

description of the system see U.S. Natl. Bur. Stand. Spec.

INTRODUCTION

SEC. 1-3]

Kilogram (kg)

Equal to mass of international prototype kilogram, a platinum-

iridiium

mass preserved

artifact

among

Second (s)

at Sevres, France.

This standard kilogram

is

the only

the SI base units.

Equal to duration of 9, 1 92,631 ,770 periods of radiation correspond-

ing to the transition between two hyperfine levels of the ground state of cesium

The second was formerly defined

133.

The

earth's rotation rate

transition
differ

much more

is

by about

is

as 1/86,400 part of a

mean

solar day.

gradually slowing down, but the atomic (cesium-133)

constant and

is

now

The two standards

the standard.

second per year.

Ampere (A) Electric current which if flowing in two infinitely long parallel
vacuum separated by 1 meter produces a force of 200 nanonewtons

wires in

per meter of length (200

nN m"

*=

2 x 10" 7

N m _1

).

Kelvin (K)

Temperature equal to 1/273.16 of the triple point of water (or

triple point of water equals 273.16 kelvins).f
Candela (cd) Luminous intensity equal to that of 1/600,000 square meter of
a perfect radiator at the temperature of freezing platinum.

The

units for other dimensions are called secondary, or derived, units

based on these fundamental units (see Table

The material

in this

book

electric current (dimensional, symbols L,

units for these dimensions are the basis of

called the meter-kilogram-second-ampere (mksa) system,

The complete
of which
3
or 10~
.

now

T, /).

a subsystem of the

SI.

and submultiples of the SI units be stated in steps of 10 3

Thus, the kilometer (1 km = 10 3 m) and the millimeter (1 mm = 10" 3 m)

is

that multiples

(= 10

~2

m)

not.

is

proper SI designation for the width of motion-picture film

list

M,

what was formerly

SI involves not only units but also other recommendations, one

are preferred units of length, but the centimeter

For a

and are

A-l, in Appendix A).

deals almost exclusively} with the four fundamental

dimensions length, mass, time, and

The four fundamental

2, Sec.

35

is

of the preferred units see Appendix A, Sec. A-l, Table

For example, the

mm,

not 3.5 cm.

1.

In this book rationalized SI units are used. The rationalized system has the
advantage that the factor 4n does not appear in Maxwell's equations although it does
appear in certain other relations. A complete table of units in this system is given
as Table 2, Sec. A-l, in

Appendix A.

The

table

lists

dimensions or quantities alpha-

under each of the following headings: Fundamental, Mechanical, Electrical,

and Magnetic. For each quantity the mathematical symbol (as used in equations),
description, SI unit and abbreviation, equivalent units, and fundamental dimensions
betically

are listed.

Note that the symbol for degree is not used with

kelvins.

ture of water

However, the degree sign

is

373 kelvins (373 K), not 373K.

Thus, the boiling temperais

retained

with degrees Celsius.

t

A couple of equations involve temperature.

None

involve luminous intensity.

[CHAP.

ELECTROMAGNETICS

It is

suggested that as each

to the table and, in particular,

new

quantity and unit

is

discussed the student refer

familiar with the fundamental dimensions for

become

the quantity.

1-4

HOW TO READ THE SYMBOLS AND NOTATION

book

In this

quantities or dimensions,

or resistance R, are always in

boldface as vectors and
(scalar).

or f .t

Ex

is

scalars, like

Quantities which

may

as scalars, e.g., electric field

charge Q, mass

M,

be vectors or scalars are

(vector) or

E(=

|E|)

Unit vectors are always boldface with a hat (circumflex) over the letter like
A dot over the symbol indicates explicitly that the quantity is harmonically

time-varying (a phasor).

but

italics.

italics

which are

E is a space vector
= %EX Ex = E^"").

Thus,

a scalar-phasor (E

and time phasor (vector-phasor),

roman type, i.e., not italic; for example, H for henry, s for second, or
The abbreviation for a unit is capitalized if the unit is derived from
a proper name; otherwise it is lowercase (small letter). Thus, we have C for coulomb
but m for meter. Note that when the unit is written out, it is always lowercase even
though derived from a proper name. Prefixes for units are also roman, like n in nC
See Table 1, Sec. A-l, in Appendix A
for megawatt.
in
for nanocoulomb or
Units are in

A for ampere, t

for complete

EXAMPLE
means

lists

of

MW

prefixes.

D = i200 pC m ~ 2

that the electric flux density

D is a vector in the positive x direction with a magnitude

of 200 picocoulombs per square meter

10

" 10

coulombs per square meter).

K=10V

EXAMPLE

means

V equals
V (roman) for volts.

that the vpltage

voltage and

(=

10 volts.

Distinguish carefully between

(italics)

for

few places in the book where tensors appear they are indicated by a boldface
with a bar over it like F. In longhand notation a vector may be indicated by
a bar over the letter and a tensor by two bars.

t In the
letter

longhand notation no distinction is usually made between quantities (italics)

However, it can be done by placing a bar under the letter to
indicate italics or writing the letter with a distinct slant.

% In

and units (roman).

INTRODUCTION

SEC. 1-6]

EXAMPLE
means
and

S = 4Wm-2 Hz-'

that the flux density

also be written
less

5=4 W/m

S (a scalar) equals 4 watts per square meter per hertz. This can
is more direct
/Hz or 4 W/(m 2 Hz), but the form W m -* Hz
1

ambiguous.

Note

that for conciseness, prefixes are used

where appropriate instead of

Thus, the velocity of light would be given as c

exponents.

_1

However,
megameters per second) and not 3 x 10
s
-1
8
exponential form (3 x 10 m s ) would be used.
.

300

Mm

in solving a

-1
s

(300

problem the

The modernized metric (SI) units and the conventions used herein combine to
and unambiguous notation, and if the reader is attentive to the
details, he will find it has both elegance and beauty.
give a concise, exact,

1-5

EQUATION NUMBERING

Important equations and those referred to in the text are numbered consecutively
beginning with each section.
section

its

refers to

number

Chap.

is

Ayhen reference

made

to an equation in a different

preceded by the chapter and section number.

14, Sec. 15,

Eq.

(3).

would read simply

15 of Chap. 14

is

(3).

reference to this

Note

Thus, (14-15-3)

same equation within Sec.

and section numbers are

that chapter

printed at the top of each page.

DIMENSIONAL ANALYSIS

1-6
It is

a necessary condition for correctness that every equation be balanced dimension-

ally.

For example, consider the hypothetical formula

M ,
T~ DA
where

M = mass
L=

length

D = density (mass per unit volume)

A =

area

The dimensional symbols

for the

left

side are

dimensional symbols for the right side are

ML
U

_
-

M
~L

M/L, the same

as those used.

The

ELECTROMAGNETICS

[CHAP.

Therefore, both sides of this equation have the dimensions of mass per length, and

the equation
correct;

is

balanced dimensionally.

i.e., it is

not a

sufficient

condition for correctness, and

This

is

not a guarantee that the equation

condition for correctness.

it is

It is,

is

however, a necessary

frequently helpful to analyze equations in this

way

to determine whether or not they are dimensionally balanced.

Such dimensional analysis

of a quantity

are.

is

also useful for determining

For example, to find the dimensions of

what the dimensions

force,

we make

use of

Newton's second law that

Force

= mass

x acceleration
*

Since acceleration has the dimensions of length per time squared, the dimensions of
force are

Mass x

length

Time 2
or in dimensional symbols

Force

1-7

ML

= -=T2

THUMBNAIL ELECTROMAGNETICS

Here are the basic equations of electromagnetic theory as covered

in the

book.

Ultimately the student should be able to recognize each equation, be able to picture
the physical circumstances surrounding

its

development,

know

the dimensions (and

and know how to apply it. For example, (31) should

be recognized as Ampere's law, which equates the magnetic field
integrated around
units) in

which

it is

expressed,

a closed contour or loop to the current through the loop (or integral of the current
density over a surface
Fig. 1-1.

bounded by the

loop).

diagram should be drawn, as

in

Units should be stated and the application to a simple problem outlined.

The student is urged to refer frequently to these equations while studying the
book and as he progresses to check each one he feels he understands. These equations
will also

be useful as

a-

brief course outline or reference while preparing for tests or

examinations.

(,)

Q,Q,
"-'3

(2)

O,

-i,-'^>

<3)

""-<"<>'"
(4)

'h

INTRODUCTION

SEC. 1-7 J

(5)

E=-VK

(9)

E=-VF-^
8t

(6)

V=-(E-dl
(10)

(7)

V=-fi4ne r

B = VxA

Ho rrr J x

<">

--SJTJT
4tt JJJ

(15)

(16)

(20)

Ql
K =7r-e/^cos0

<k

<>

D = eE

(17)

D-ds = jjj pdv

< 21 )

rfB

(8)

B = nH

(18)

V-D = p

(22)

\/,

V2 K =

(28)

C=y

(25)

(29)

V 2 K=-A4

= em /

(26)

(30)

|j>

J ds

Closed contour
1-1

Amp6re's law

(31).

<//

rf/

sin

Anr 1
<rE

=jJB'ds
<jEB'<fe

VB =

V-J =

Surface (area
bounded by
contour

FIGURE

**

</i>

(27)

(31)

Illustrating

^-j

^
dB =

(23)

(24)

^O

(19)

I=$H-dl = fjj-d*

A)

fn-d\

ELECTROMAGNETICS

(32)

[CHAP.

-d-*~k

(33

*-?-

>

V*

(34)

H=J
V-A =

(35)
(36)

(40)

En =Et2
Dnl -Dn2 =p,
B,,, = 2? 2
(38)
(37)
(39)
D=e E+P
(41) B =
(H + M)
(42) B = VxA

= H, 2

//rl

/i

A x

(43)

(44)

-=C

--

(45)

(46)

L=

_=7

L=

(47)

^
(48)

= ieE 2

(49).

w.

(53)

<*A
<U=--

(50)

= iCV 2

wm = lnH 2

(51)

,_ V-^K.il-/
v

ff

^m = L/

(52)

= ^H-dl.Jj(j+^-A

(59)

VxH = J+

rr 5B

VxE=-^

(57)

(58)

* = -^-

ExH

(64)

+ Et
E -,

VSWR =

Sav = \ Re (E x H*)

W=VI

(65)

(66)

= i Re

JP,,

K/*

(67)

4=

U = j(yxB)-dl- jjy-ds

$ = i(ExH*)

(63)

2nR

(60)

(61)

S=

s //B.A--//^.*

(54)

(56)

H=K

H==K

(55)
v
'

(62)

=E

sin (art

V J=

(68)

fix)

^
of

\=^

(69)

\ZC

(72)

K=

cos

(70)

MM
-2

(73)

j,-^

VxVxE-y

(71)

FT = },4 e S(l

rf

<

'> +

< 2 >

(76)

E=

cos A/A/)
(74)

C75)

rf-WW+'a

Q= <^>-<^>

(77)

1/

<EjE 2 cos

<5>

(78)

K=S

<sin 2c>

SEC **7]

INTRODUCTION

(79)

^_ZFK =

(83)

(86)

= %=

Z=

(80)

"

Z=

(84)

= /()
N
+(t^) ->V

w*-|-^
(91)

Q=

(97)

(1

p=

(92)

T^dPJdi

If

s -'

+ e" + e> + e* +

(82)

!!
=

<->
(93)

*><-*)

-l

-lt

e = f(y)exp(^co S 0)dy
L = ^jr,dl

(101)

w=

(104)

fcT

Zs =

(102)

^o J

= $JW,*)<i

Sin(

Tmin

(110)

F=

c, p

(103)

(105)

S =

^=

=-^-

= t^j

(108)

(111)

F=wa = w

1)

= d = -|*.
2
(

/I,

)av

F=

(112)

= .M=piM

?E

= e(vxB)

W = eV

(113)

(114)

(118)

(120)

,/

(115)0,*

+ i(ixB-y
J-G
at
p

V^J-.
ax.

(121)

-r^ =
dx

(116)

(119)

^I =
(122)

= Jl-(-^V

(117)v

4^
J

HPBW = -L

(100)

(109)

dv

evB

= ^4-"
^
P,
r**r

(106)

Ae

/(

"ff

42,,

.4 C

(107)

(99)

sin(i/>/2)

(98)

/oc

~je"" w sme
x

rfs

(85)

60jt/

.-MSi

2tc

(87)

= .

(81)

_ (/ c )2

^-=
dx m

(123)

=
dx m

-^

10

[CHAP.

ELECTROMAONETICS

(124)

\E

+ (v x

B)|

(125)

D-J =

(126)

= y Bx -

B'

(127)

x EJ

F=D-TmB

SPEMP CHART
Solution Paths for Electromagnetic Problems

Physical situation

EM

module

Understanding
and reevaluation

Problem statement
and solution strategy

TL

Boundary values
or constraints

Maxwell's equations

Continuity equation

Coulomb's law

Laplace's and

Biot-Savart law
Other laws

Poisson 's equations

Moving systems

Maxwell's stress

Invariance

tensor

Wave equation

Scalar and

Poynting vector

vector potential

Lorentz transformations
Einstein four-

dimensional formulation

L.
Programmable

Analytical

method

digital

Hybrid

computer

Analog computer
(such as Teledeltos

Graphical

method

paper, antenna pattern

range, or other device)

Solution

Display
'
1

Equation or
numerical result
1

Table

L_

Graph

Maps

-1

Appicationto physical solution or otherwise

Three- dimensional model

INTRODUCTION

SEC. 1-8 J

-11

SPEMP CHART

1-8

In the preceding section the fundamental equations of electromagnetic theory are

In applying these equations to a problem a path

presented.

achieve a solution and display the results.

As

Problems.

which

SPEMP

The

usually available.

followed in order tc
is

chart shows Solution Paths for Electromagnetic

the student progresses through the book, he should refer to this chart,

will clarify the steps

magnetic (EM) theory

is

and procedures.

Note that the part related to electro(EM module) while the part

contained within the dashed line

outside (analogous to the "

main frame ")

an appropriate module

substituted.

if

is

choice of methods and displays

is

is

general and can be used for any discipline

PROBLEMS
*

l-l

Give

the symbols

M,

(a) the

L, T,

dimensional description,

and

/,

and

(c)

dl

where

-r
at

(b)

the dimensional formulas in terms of

the SI units for each of the following:

F dl

length and

where

= time

F= force

d[

dx
1-2

Give the information requested

jjjpdv

in

Prob.

1-1

Q2
*

jE-rfl

4ne

47tc

'

for each of the following:

Answers to starred problems are given

in

Appendix C.

bii

THE STATIC ELECTRIC

2-1

FIELD:

PART

INTRODUCTION

In this chapter the basic relations for static electric fields in free space are discussed.

They include Coulomb's law and the force per charge, Gauss' law and the electric
flux, and the superposition of fields.
The equivalence of the electric field to the
gradient of the electric potential and of the electric potential to the line integral of the
electric field are explained.
The orthogonality of field lines and equipotential contours
'is pointed out.
Shells of charge and capacitance are also considered.
Also mentioned are the simplifying concepts of staticness and idealness and the conventions
of positiveness, right-handedness, and outwardness.

2-2

THE FORCE BETWEEN POINT CHARGES

AND COULOMB'S LAW

group of charged

Even a

particles,

i.e.,

atoms or

single electron has a finite size.

By " finite "

is

meant " not

electrons, occupies a finitef volume.

However,
infinitesimal."

it is

often convenient to regard a

sec. 2-2J

THE STATIC ELECTRIC FIELD! PART

small, concentrated region of charged particles as a point charge.

-13

This assumption

no appreciable error provided the size of the volume occupied by the charged
particles is small compared with the other distances involved.
The basic experiment of electrostatics was first reported by Coulomb in
leads to

1785,f

using small charged bodies which

may

be regarded as point charges. The results of

this experiment are given by Coulomb's law, which states that the force F between

two point charges

and

Q2

proportional to the product of the charges and in-

is

versely proportional to the square of the distance r between them,

F=k-Q1Q2
where k

is

line

(N)

a constant of proportionality.

distance this law

As suggested

connecting the charges.

(1)

Because of the inverse-square

said to be an inverse-square law.

is

effect

of

The force is in the direction of the

in Fig. 2- la, the force

is

two charges are of the same sign, but, as suggested in

inward (attractive force) if the two charges are of opposite sign.

force) if the
is

i.e.,

outward (repulsive
Fig. 2-16, the force

F
(a)

FIGURE

2-1

(a)

Two

(fc)

opposite sign.

point charges of

same

sign

and

F
"*

*_

J"+

(*)

Q,

In the International System the constant of proportionality

is

given by

k =

4ne
where e is the permittivity J of the medium in which the charges are situated. By
dimensional analysis of (1) we find that e has the dimensions of capacitance per length,

or in dimensional symbols T*I 2 /MI?. The SI unit for permittivity is the farad per
meter (F

m _1

The

).

permittivity of
e

8.85

x l(T

vacuum
,2

Fm

is

-1

=8.85pFm"

si^xlO-'Fffl-^J-nFm3o7t

t Charles A. de

36rc

Coulomb, History Royal Acad.

Sci. (France), 1785, pp.

569 and 579.

t Also called the dielectric constant.

For a discussion of capacitance

capacitor.

see Sec. 2-24.

See also Sec. 3-22 on the field-cell

14

[CHAP. 2

ELECTROMAGNETICS

The permittivity of air is substantially the same as for vacuum. In this chapter it is
assumed that the medium is air or vacuum, and so to be more explicit the permittivity
e could be written e throughout. However, the same equations apply to a medium
of infinite extent provided the medium is isotropic, homogeneous, and linear. The
general situation of material media for which e

Force

is

a vector;

vector equation

and

i.e., it

#e

discussed further in Chap.

is

has both magnitude and direction.

also substituting the value of k,

Rewriting

3.

(1) as

we have

(2)

4ner

where

F=

Nf

force,

unit vector (see Fig. 2-1) pointing in direction of line joining the charges

F = tF)
=
charge
1, C
Qj
=
charge
C
2,
Q2
e = permittivity of medium, F m~
r = distance between point charges, m
(thus,

This

is

the complete vector expression for Coulomb's law stated in rationalized SI

To demonstrate

units.

the application of this law

let

us consider the following

problems.

EXAMPLE

A negative point charge of /*C is situated in air at the origin of a rectangular

A second negative point charge of 100 pC is situated on the positive x
1

coordinate system.

axis at the distance of 500

solution

mm from the origin.

What

is

the force on the second charge?

By Coulomb's law the force

F = i-

(-10- 6)(-10-4)

(4irX0.5 2)(10- 9 /367r)

= + &3.6 N
That

is,

there

EXAMPLE
in air.

is

What is

a force of 3.6

N (0.8 lb) in the positive x direction on the second charge.

Two point charges of 1 C each and

of the same sign are placed

the magnitude of the repulsive force?

One newton

equals the force required to accelerate

kg

mi"'.

mm apart

SBC. 2-4)

THE STATIC ELECTRIC FIELD: PART

solution

From Coulomb's

law the force

1

This force

is sufficient

to

lift

gravity equals the

mass

AND

weight of 9.2 x 10 11 metric tons

weight of 10 12 U.S. tons

(kg) times 9.8

IDEALNESS

2-3

10' 5

millions of

Empire

State buildings, or roughly all the buildings

[Note that the force

m s -2

45

is

F= 4we
- * =9x
a
x*A
10 6

in the United States simultaneously.

N at the earth's surface due to

.]

STATICNESS

The concept of a point charge employed in Sec. 2-2 is an idealization. Jn many

situations discussed in science and technology idealness is invoked in this manner in
order to simplify the problem.
In Chaps. 2 and 3

it is

important to note that

The implication here

discussed.

to one another.

This

is

is

that all charges

also something of

developing the basic concepts of

an

it is

the static electric field which is

and

objects are at rest with respect

idealization but

field theory.

one which

is

useful in

In later chapters the situation under

conditions of motion will be treated.

ELECTRIC FIELD INTENSITY

2-4

Consider a positive

electric point

coordinate system.

If

of

it is

charge Q t placed rigidly at the origin of a polar

another positive point charge Q 2 is brought into the vicinity
acted upon by a force. This force is directed radially outward and becomes

greater as
forces

Q2

may act.

approaches Q^. It may be said that Q t has a field around it where

The nature of this field is suggested by the vector diagram of Fig. 2-2,

the length of the vector being proportional to the force at the point (at center of vector).

Dividing (2-2-2) by
charge,

Q2

puts the equation in the dimensional form of force per

i.e.,

F
Q2

force

charge

which has the dimensional symbols

ML _ ML
T 2 Q~T*1

16

'

[CHAP. 2

ELECTROMAGNETICS

FIGURE

2-2

1,

/
\

+
0,

\
X

Q x with vectors indicating

magnitude and direction of associated
Point charge

electric field.

If

Q2

is

a positive

test

charge, the resulting force per unit charge

electric field intensity^ E.

Q2 =

where

The

Q,

Q2

Aner 2

According to

which

(1)

positive test charge.

SI unit of electric field intensity

is

field intensity

(1),

is

meter (V

the point charge

proportional to

the newton per

is

electric potential (Sec. 2-7),

electric field intensity is the volt per

defined as the

Thus,

appear after the discussion of

sity

is

and

Q
is

is

m"

coulomb (N C" ')

As

will

an equivalent unit for the

).

surrounded by an

electric field

inversely proportional to r

a vector having the same direction as the force

2
.

of inten-

The

electric

but differing

in

dimensions and in numerical magnitude.

It may
It is not implied by (1) that the positive test charge has a value of 1 C.
have any convenient value since the ratio of the force (newtons) to the test charge
(coulombs) is independent of the size of the charge provided the test charge does not
disturb the field being measured. As noted in Example 2 above, 1 C represents a
much larger charge than is ordinarily encountered in static problems. Thus, if we

attempted to use a

we

test

seek to measure.

test

charge of
Therefore,

C,
it is

we would tend to disturb the charges whose field

necessary to use small test charges; in fact, the

charge should be small enough to ensure that

charge configuration whose

field is to

be measured.

Also called the electric field strength.

it

does not appreciably disturb the

THE STATIC ELECTRIC FIELD: PART

SEC. 2-5]

If the test charge is

made

small enough,

so that the ultimate value of the electric

size,

AF on a positive test charge AQ

force

as the charge approaches zero;

it

may be

17

regarded as of infinitesimal

field intensity at

divided by the charge

a point becomes the

(A0 with the limit taken

i.e.,

E=

lim
ac->o

AF

Ay

(2)

Actually the smallest available test charge

is an electron.
Since this is a finite charge,
cannot be measured with unlimited accuracy. Although this is of
importance in atomic problems, it need not concern us in the large-scale, or macroscopic, problems treated in this book.
In practice, E would be measured with a small
but finite test charge, and if this charge is small enough, E would differ inappreciably
it

follows that

from that measured with an

infinitesimal or vanishingly small test charge as implied

in (2).

EXAMPLE A negative point charge

What

coordinate system.
3

is

10

nC

is

situated in air at the origin of a rectangular

the electric field intensity at a point on the positive

axis

m from the origin ?

By

solution

(1)

the field intensity

is

given by
[

_.X

That

is,

10 -"

- io^cJ
"

(47rx3 2)(10-/367r)

"^

jj c

NC-'forVm- )
the electric field intensity is 10 N C -' (or 10 V m -') and is in the negative x direction.
-il0

POSITIVENESS, RIGHT-HANDEDNESS,

2-5

AND OUTWARDNESS
It will

be convenient at the outset to adopt several conventions.

It is

convenient to define the electric

charge.

Thus, the electric

field at

field,

as in Sec. 2-4, in terms of positive

any point

is

in the direction

of the force on a

positive test charge placed at that point.

It is convenient to standardize on right-handed coordinate systems.

For
example, the rectangular coordinate system in Fig. 2-3a is right-handed since

turning the positive x axis into the positive y axis and proceeding the direction of
a right-handed screw, we move in the positive z direction. This set of coordinate

axes

may

accordingly be termed a positive

system in Fig. 2-36

is

set.

The rectangular coordinate

x axis into the positive

left-handed since turning the positive

[CHAP. 2

ELECTROMAGNETICS

18

A
(6)

(a)

FIGURE
(a)

axis,

we must proceed

like

2-3

Right-handed and

(b) left-handed coordinate systems.

a left-handed screw to

move

in the positive z

direction.

around
a consequence of the positiveness convention that the electric field
we
that
is
consequence
further
A
2-2).
Fig.
a positive charge is outward (see
the
surface
as
closed
on
a
point
any
at
normal
take the positive direction of the
It is

note that the

outward normal. Thus, applying Gauss' law (Sec. 2-17), we shall
density
over a
flux
electric
outward
integral of the normal component of the
enclosed.
closed surface yields the positive charge
inherent in our definitions and, as discussed above,
outwardness.
also associated with the definitions of right-handedness and

The concept of positiveness

is

2-6

is

THE ELECTRIC FIELD OF SEVERAL POINT CHARGES

AND THE PRINCIPLE OF SUPERPOSITION OF FIELDS

a linear function of the value of the charge,

point charge are linearly superposable by
it follows that the fields of more than one
fact
may be stated as the principle of superthis
vector addition. As a generalization,
Since the electric field of a point charge

is

position applied to electric fields as follows:

The

total or resultant field at

a point

is

the vector

sum of the

individual

component

fields at the point.

Thus, referring to Fig. 2-4, the

P is E
is

field intensity

the vector

of the point charge

at the point

The total field at P due to both point charges

and of the point charge Q 2 is E 2

sum of E t and E 2 or E, as indicated
.

in the figure.

sec. 2-6]

THE STATIC ELECTRIC FIELD: PART

FIGURE

19

positive point charge of

nC

is

+Qi

situated in air at the origin (x

and a negative point charge of 2 nC is situated on the y axis 1

y = 1), as shown in Fig. 2-5. Find the total electric field intensity
2 m from the origin (x = 2, y = 0).

N-10- 9

FIGURE

2-4

Vector addition of fields due to two

equal point charges of the same sign
to give resultant, or total field E.

EXAMPLE A

= 0, y = 0),
(x = 0,

m from the origin

at the point

P on the x axis

2-5

Vector addition of

fields

due to two unequal point charges of opposite sign to

give resultant, or total field E.

solution

from

The

vector value of the electric field E, due to the charge at

(2-4-1),

E,=i-

10-

(477x2 2 )(10- 9 /36tt)

= i2.25NCThe magnitude of the

field

E 2 due
2

to the charge at (0, 1)

is

-2xlQ- 9

~(47TX2.24 2 )(10-736ir)

= -3.59 NC"

(0,

0)

is,

20

ELECTROMAGNETICS

The

vector value of

[CHAP. 2

E2

is

given by

E 2 = 3.59 cos

= -x3.58

+ y3.59 sin a

^4 + y3.59^

= -x3.2 + yl.6NC-'
= unit vector in x direction
y = unit vector in y direction.

where &

The

total vector field

E can

be obtained by graphical vector addition of E, and

ically as follows

or

in

E = x(2.25-3.2)+.yl.6

both rectangular and polar forms

E = -x0,95 + yl.6 =

2-7

E 2 or analyt-

1.86/120.7

NC"

THE ELECTRIC SCALAR POTENTIAL

Consider two points, jq and x 2

direction.

situated in a uniform electric field

moved

Let a positive test charge at x 2 be

The

as in Fig. 2-6.

field exerts

the charge against the force.

in the negative

parallel to the

x direction to x t

a force on the charge so that it requires work to move

The amount of work per unit charge is equal to the

force per unit charge (or field intensity E) times the distance through which the charge
is

moved.

Thus,

E(x 2

*i)

= work

(joules/coulomb, J

per unit charge

C"

(1)

The dimensions are

Force

work

x length
charge

Charge

ML

L _

ML

T 'Q~ T 3 I

or
In SI units the relation

is

Newtons

Coulomb

x meters

joules

coulomb

The dimensions of work per charge are those of potential. In our example
work or energy per unit charge required to transport the test charge
from x 2 to x is called the difference in electric potential^ of the points x 2 and x
The point x x has the higher potential since it requires work to reach it from point x 2
Thus, moving from x 2 to x l (opposite to E), we experience a rise in potential. The unit
(Fig. 2-6), the

quantity.
For
t Potential, in general, is a measure of energy per some kind of unit
above
example, the difference in gravitational potential at sea level and 100
sea level is equal to the work required to raise a 1-kg mass from sea level to a height
Potential is a scalar quantity, i.e.,
against the earth's gravitational field.
of 100

it

has magnitude but no direction.

SEC. 2-7]

THE STATIC ELECTRIC FIELD: PART

21

FIGURE

of electric potential
is

2-6

Linear path

Path

as.

in, uniform electric field.

F is the volt (V) and is equal

expressible either in joules per

coulomb or

Newtons

- x meters
Coulomb

a;

>

UC

to

"
'

Hence,

electric potential

in volts.

joules

coulomb
-

volts

Dividing by meters, we obtain

Newtons =

volts

Coulomb

meter

Thus, the electric

field intensity

= electric field intensity

expressible in either newtons per

is

coulomb or

in

volts per meter.

EXAMPLE
If the distance

Let the uniform electric

x2

solution

x,

is

From

100

field in Fig. 2-6

mm, what

is

(1) the electric potential is given

K=10x
That

is,

the potential of x,

is 1

0.1

(Fig. 2-7).

of 10

m _1

by

= V
1

higher than the potential of x 2

Consider next the case of a nonuniform

positive point charge

have an intensity

the potential difference of the two points?

The

field

such as exists in the vicinity of the

electric field

is

radial

and

is

inversely

proportional to the square of the distance r from the charge Q. The energy per
coulomb required to move & positive test charge from r2 to rx along a radial path equals
the potential difference, or rise, K21 between the points. This is given by
21

=
J*

'dV= - f'Edr

(V)

/
E,

"/

/
+Q

V
*i

FIGURE

rz

2-7

Linear path in nonuniform electric

Path

field.

[CHAP. 2

22

ELECTROMAGNETICS

The

negative sign takes into account the fact that the motion from r 2 to r l

'

to the direction of the

Substituting the value of

field.

=
where

Q
Q_/}__
47r\r,

is

opposite

(2-4-1) yields

(2)

,N
J_\
r2 /

V = potential at point r
V2 = potential at point r 2
i

The

potential difference in (2)

the positive test charge from r 2 to

r,

/1

r dr _=
_JL f'dr
2
4ne J, 2 r

E from

to r 2

the field does

r,

is

positive since

against the

work must be expended

However,

field.

the charge and there

work on

is

fall

if

to

the motion

move
from

is

in potential (negative

potential difference).
If the point r 2 (Fig. 2-7) is

Thus,

at zero potential.

(2)

removed

to infinity,

we can

arbitrarily define

it

to be

becomes
(V)
4tt/

This potential
It is inversely

the

called the absolute potential of the point r x

proportional to the distance from

work per coulomb required

point

2-8

is

(3)

to the point

due to the charge Q.

and is, by definition,

r,

to bring a positive test charge

from

infinity to the

a-j.

THE ELECTRIC SCALAR POTENTIAL AS A

LINE INTEGRAL OF THE ELECTRIC FIELD

In Sec. 2-7 the test charge

the path followed

is

is

moved via the shortest path between two points.

immaterial since the potential difference

the difference in potential of the two end points of the path.

is

Thus, referring to Fig.

2-7, the potential at the point r t with respect to the potential at r 2 is said to

be single-

can have only one value regardless of the path taken from r 2 to r y
When the path of the test charge is not parallel to E but at an angle 0, as in
Fig. 2-8, the potential difference V21 between the points x 2 and x x is equal to the
path length x 2 *i multiplied by the component of E parallel to it. Thus, V2l =
valued;

Actually,

determined solely by

(x 2

i.e., it

-Xi)E
If

cosd.

the test charge

no work

is

is

moved

performed and hence

important property of

perpendicular to the direction of the

this

path

is

said to be

fields that equipotential

and

field (0

an equipotential

field lines are

line.

orthogonal.

90),

It is

an

SEC. 2-8

THE STATIC ELECTRIC FIELD: PART

FIGURE

2-8

Linear path in uniform electric

angle 0.

field at

yy

'

*'

Let us consider next the case where the path of the test charge
the potential difference between the end points of the path
the infinitesimal element of path length dl and the
integrated over the length of the path

from a to

in Fig. 2-9, the infinitesimal potential rise

dl

is

is

curved.

x'

Then

given by the product of

component of

For the path

b.

dV

is

23

in the

parallel to

uniform

it,

field

between the ends of the path element

given by

dV =
where 6

is

-E cos Gdl= -E-dl

the angle between the path element and the

(1)

field.

A potential rise (positive

dV) requires that the component of the motion parallel to E be

opposed to the field. Hence the negative sign in (1). By integrating (I) between the
points a and b we obtain the potential rise Vab between the points a and b. Thus,
potential difference

'

dV =

- Va = -

cos

dl

= -

dl

(2)

we note that contributions to the work occur

component parallel to E. Hence, the work in going from
work increments along the zigzag-path elements parallel to

Referring to the zigzag path in Fig. 2-9,

only

when

the path has a

sum of

a to b

is

E (0 =

180) taken in the limit as the increments

the

the

tion

approach zero with no work contribufrom the zigzag-path elements perpendicular to E (0 = 90).
The integral involving dl in (2) is called a line integral. Hence, the potential

rise

between a and b equals the

of

line integral

along the curved path from a to

b.

As

written in (2), the result can be expressed either in scalar form (with cos 0) or in
vector form as a scalar or dot product,

EXAMPLE
(a

uniform

In Fig. 2-9

field).

solution

Let x,

From

Consider

+Q

now

m.

be everywhere

Find

Vab

in the

+x direction

and equal to 10

r"

m -'

Ecos0dl = -\

Edx = Ex = + \0V
t

the situation where the path of the test charge

nonuniform.

as in Fig. 2-10.

(2)

K06 = -

electric field is

let

The

Let the nonuniform

field intensity

field

is

curved and the

be produced by a point charge

due to a point charge

is

given by (2-4-1).

24

[CHAP. 2

ELECTROMAGNETICS

FIGURE

2-9

Curved path

Substituting this in (2) and also putting dr

cos 9

dl,

in

where dr

a uniform electric

is

an

field

infinitesimal

element of radial distance, gives

Q_r'*_Q_(l_l\

v""
Putting b

/-j

along a radial

EXAMPLE
= 400 mm

and a

r2

Anchor 2
makes

4ne\b

(V)

this result identical

(3)

a)

'

with (2-7-1), where the path

is

line.

Let the positive charge Q, Fig. 2-10, be equal to 223 pC. Also let
= 100 mm. The medium is air. Find the absolute potential V. at a,

and b

the absolute potential

at b,

and the potential

rise

Vat

SOLUTION

v.=

v--

4lTo

vab = Vb ~
zero (0

to

move a

test

to

V.

15

charge along an equipotential contour or surface

= 90). The maximum amount

moving normal

^ 20

b*

4tt

The work

^ 5V

of work

an equipotential surface.

per unit distance

is

is

performed by

This coincides with the direction of the

field.

The work

to transport a test charge

around any closed path in a static field is

Thus, the upper and lower limits

zero since the path starts and ends at the same point.

of the integrals in

(2)

become the same, and the

result

is

zero.

property of the

THE STATIC ELECTRIC FIELD: PART

sec. 2-9]

FIGURE

25

2-10

Curved path

in

a nonuniform

electric

field.

static electric field

zero.

It

is,

then, that the line integral of this field around

a closed path

follows that the potential difference between any two points

is

is

independent

of the path.

2-9

RELATION OF ELECTRIC FIELD LINES AND

EQUIPOTENTIAL CONTOURS; ORTHOGONALITY

field line indicates the direction

into the

field.

If the test

charge

is

of the force on a positive

released,

it

moves

test

In a uniform field, the field lines are parallel, as in Fig. 2-11.

gives

no information about the

charge introduced

in the direction

intensity of the field.

It indicates

of the

only the direction.

However, by measuring the work per coulomb required to move a positive

+- H

field line.

A single field line

test

charge

r-

10 V
_J

10

FIGURE

20

30

40

50 mm

2-11

Field lines (solid) and equipotential lines (dashed) of a uniform electric

field.

]f

26

ELECTROMAGNETICS

along a

[CHAP. 2

field line the potential differences

larger the potential difference

the

In a uniform

field.

The

along the line can be determined.

between two points a unit distance apart, the stronger

field the potential difference

per unit length

is

constant, so

that the equipotential lines (which are orthogonal to the field lines) are equally spaced.
-1
In the example of Fig. 2-11, the electric field intensity is 0.2
so that the
,

V mm

equipotential contours at 1-V intervals are parallel lines spaced 5

the lines

is

arbitrarily taken as

mm apart.

One of

having a zero potential so that the potentials shown are

relative to this line.

Consider

now

the case of a nonuniform field such as exists in the vicinity of the

positive point charge

it

moves

radially

Q in

Fig. 2-12.

away from Q,

If

positive test charge

released in this field,

In Fig. 2-12 this

varies inversely as the square of the distance.

fact that the field lines

become more widely separated

The absolute

is

potential

is

so that the field lines are radial.

The
is

field intensity

suggested by the

as the distance from

Q increases.

from Q. If Q 10 pC,
are then as shown by the concentric

inversely proportional to the distance

the equipotential contours for 20, 10,

5,

and

circles in Fig. 2-12.

>h

It is to

be noted that a potential

rise is

always in the opposite direction to E.

20

FIGURE

30 mm

2-12

Field lines (solid) and equipotential lines (dashed) of a nonuniform electric

field.

THE STATI C ELECTRIC FIELD: PART

SEC. 2-10]

27

FIELD OF TWO EQUAL POINT CHARGES OF

OPPOSITE SIGNS AND OF SAME SIGN

2-10

P due to two point charges + Q and - Q is equal to the

vector sum of the fields at P due to each of the charges alone. This is illustrated in
The potential V at P is equal to the algebraic sum of the potentials at P
Fig. 2-13.
The

electric field at

due

to.

a point

each charge alone.

A map

of the

+Q

Fig. 2-13 for point charges

contours are given in volts for

constitutes

an

and equipotential contours (dashed) is shown in

and - Q separated by 127 mm. The equipotential

field lines (solid)

electric dipole

Q=

The charge configuration

140 pC.

with a charge separation of 127

in Fig. 2-13

mm.

two equal point charges of the same sign

A map of the field lines (solid) and equipotential contours
(positive) as in Fig. 2-14.
(dashed) is shown for a charge separation of 127 mm. The equipotential contours
In contrast to the above

are given in volts for

Q=

let

us consider

140 pC.

The only

difference

between the charge configura-

and that in Fig. 2-13 is that the lower charge is positive.

Near each charge the effect of the other charge is small, and the equipotentials
For intermediate distances the
are circles like those around an isolated point charge.
Of particular interest is the figureequipotentials have the shapes shown in Fig. 2-14.
eight-shaped equipotential (V = 39.5 V) that crosses itself at the point P, where
E = 0. This point is called a singular point. At such a point, field and equipotential

tion of Fig. 2-14

lines are not perpendicular.

Note that in three dimensions the (dashed) equipotential lines are surfaces
generated by rotating Figs. 2-13 and 2-14 about the (vertical) axis through the charges.

2-11

CHARGE DENSITY AND CONTINUOUS

DISTRIBUTIONS OF CHARGE

The electric charge density p

volume.

is

equal to the total charge

Q in a volume

v divided

by the

Thus,

(1)

Electric charge density has the dimensions of charge per unit volume, or in

dimensional symbols
-3
meter (Cm ).

g/L3

The

SI unit of charge density

is

the coulomb per cubic

28

ELECTROMAGNETICS

[CHAP. 2

FIGURE

2-13

and potential variation around an electric dipole consisting of a

and negative charge of 140 pC separated by 127 mm. The solid lines are
lines, and the dashed lines are equipotential contours, with their potential

Electric field

positive
field

level indicated in volts.

By assuming

that electric charge can be continuously distributed throughout

a region we can also define the value of the charge density p at a point P as the charge
AQ in a small volume element Av divided by the volume, with the limit of this ratio
taken as the volume shrinks to zero around the point P.

In symbols,

THE STATIC ELECTRIC FIELD PART

SEC. 2-11]

FIGURE
pC

29

2-14

Electric field

140

and

potential variation

separated by 127

mm.

The

around two equal positive charges of

and the dashed lines

solid lines are field Jines,

are equipotential contoursi with their potential level indicated in volts.

lim
Au-.0

(2)

Af

This gives the value of p at a point and hence defines p as a point function.
It will be convenient to use this definition of p, but it is to be noted that

based on the assumption that the

electric

charge

is

electric

charge

is

not continuously distributed but

continuously distributed.
is

it is

Actually

associated with discrete particles

30

ELECTROMAGNETICS

[CHAP. 2

(electrons or atoms) separated

by

finite

atomic distances.

Nevertheless, the assump-

tion of a continuous charge distribution (an idealization) leads to

provided the region contains

many atoms

compared with atomic dimensions.

large

no appreciable error

or electrons and the distances involved are

The assumption of continuous charge

distribution can be applied to the large-scale, or macroscopic, problems treated in this

book but would not be

applicable to problems

on atomic

structure,

where the non-

continuous nature of the charge distribution must be taken into account.

The charge

is sometimes called a volume charge

from surface charge density and linear charge density. The

density p, discussed above,

density to distinguish

it

surface charge density p s gives the charge per unit area (coulombs per square meter,

m -2

at a point

density

on a continuous surface distribution of charge.

p L gives the charge per unit length (coulombs per meter,

on a continuous

(2),

linear charge

at a point

Both p s and p L are point functions which

with a surface or line element substituted for the volume

line distribution

can be defined as in

The

C m" 1 )

of charge.

element.

ELECTRIC POTENTIAL OF CHARGE

DISTRIBUTIONS AND THE PRINCIPLE
OF SUPERPOSITION OF POTENTIAL

2-12

Since the electric scalar potential due to a single point charge

value of

its

charge,

it

linearly superposable

may

is

a linear function of the

more than one point charge are

addition.
As a generalization, this fact

follows that the potentials of

by

scalar (algebraic)

be stated as the Principle of Superposition applied to electric potential! as follows

The

total electric potential at

a point

is

the algebraic

sum of the

individual

com-

ponent potentials at the point.

Thus,

if

only the three point charges

Q u Q2

and

Q3

are present in Fig. 2-15,

the total electric potential (work per unit charge) at the point

Vp
where

rt

r2
r3

= distance
= distance
= distance

from
from
from

JL/ei +

fc

47T\ri

r2

P is

given by

e.)
(1)
r3 )

Q to P
Q 2 to P
Q z to P
x

This can also be expressed with a summation sign.

V, =

Thus,

7Z-lT

t Although electric scalar potential

for brevity.

is

implied, the

(2)

word

scalar will usually be omitted

THE STATIC ELECTRIC HELD: PART

sec. 2-12]

dl

31

Linear charge
distribution

Ir

Point charges

\r.

*v*
2

Z^7 P
r
r
is

Surface charge

//
/
J
/

distribution

Volume
charge
distribution

dvS

FIGURE

2-15

Electric potential at
line, surface,

If the

charge

is

is

the algebraic

sum of

not concentrated at a point but

Fig. 2-15, the potential at

the potentials due to the point,

and volume distributions of charge.

P due

is

distributed along a line as in

to this linear charge distribution

is

--\^dl
J
4ne

where p L
dl

The

When the

is

carried out over the entire line of charge.

charge

is

distributed over a surface, as in Fig. 2-15, the potential at

caused by this surface charge distribution

ds

= surface charge density,

= element of surface, mJ

integration

is

is

v-km**
The

(3)

= linear charge density, Cm -1

= element of length of line, m

integration

where p s

(V)

(V)

Cm -2

carried out over the entire surface of charge.

(4)

32

ELECTROMAGNETICS

[CHAP. 2

For a volume charge

distribution, as in Fig. 2-15,

(5)

where p
do

The

=*

(volume) charge density,

element of volume,

integration
If

is

Cm -3

taken throughout the volume containing charge.

the point charges, the line charge distribution, the surface charge distribution,

and the volume charge distribution of

total electric potential at the point

position principle the algebraic

Fig. 2-15 are all present simultaneously, the

due to

sum of

v = vp

all

of these distributions

the individual

component

is

by the super-

potentials.

+vL +v +v

(6)

"-=(?Ht'' + JJt* + /JJ'*)

We may

Thus,

(7)

also write (7) as

V=YQG
G = X/Aner.
G is the electrostatic

+ jp L Gdl+ jjPsGds + jjjpG

dv

(8)

where
.

Green's function and

is

equal to the potential for a 1-C

point charge.

EXAMPLE As shown in Fig. 2-16 a square that is m on a side in air ha? a point charge
Q, = + pC at the upper right corner, a point charge Q 2 = 10 pC at the lower right corner,
and a line distribution of charge of uniform density p L = + 10 pC m "' along the left edge.
1

Find the potential at the point

P at

the center of the square.

lm

^^
FIGURE

2-16

Line and point charges for example

illustrating
superposition of electric
potential.

THE STATIC ELECTRIC FIELD: PART

SEC. 2-13]

solution

The

potential at

The

potential at

P caused

by the

P due to the

point charges

line distribution

of charge

10dy

V0.5 2 +j> 2

4tT " J
.
" y= 0.:

The

total potential at

P is

33

is

is

= + 0.158 V

then

V= V + Kt = + 43mV
p

The principle of superposition stated for the special cases of potential in this
section and for fields in Sec. 2-6 can be applied, in general, to any quantity which is
linearly related to its cause.
The electric fields or potentials at a point are linear functions of the charge producing them and hence are superposable (by vector addition
for fields and scalar addition for potential).

THE ELECTRIC FIELD AS THE

GRADIENT OF THE ELECTRIC POTENTIAL

2-13

The

potential rise between

two points along an

gradient of the potential in the

a slope

is

same way

a measure of the gradient of the slope.

potential at a point

is

electric field line is

a measure of the

that the elevation rise between

More

defined as the potential rise

AV

two points on

specifically the gradient

of the

across an element of length A/

along a field line divided by A/, with the limit of this ratio taken as A/ approaches
Thus,

zero.

Gradient of

More
line with

V=

completely, the gradient of

magnitude as given

in (1).

is

dV
=

AV
lim
Ai-o Al

(1)

dl

a vector whose direction

is

Since a potential rise occurs

against the electric field (see Sec. 2-8), the direction of the gradient

is

along a

field

when moving

opposite to that
*

of the field.

Thus,

"""

Gradient of
In

more symbolic notation we can

where grad

V=

grad

V = \V = -E

(2)

V.
As will be discussed in the next section,
can also be expressed with the operator del, or nabla, V, as VK

stands for the gradient of

-E

write

Gradient of

the gradient of

V=

34

ELECTROMAGNETICS

[CHAP. 2

GRADIENT IN RECTANGULAR COORDINATES

2-14

In this section a relation for gradient will be developed in rectangular coordinates.

To do

this,

consider the electric potential distribution of Fig. 2-17.

coulomb to bring a
This

104 V.

of both
is

is

its

variation

is

The

The work per

of coordinates)

is

indicated by the equipotential contours.

The field

potential elsewhere

parallel,

no variation with respect to z (normal to page).

and equally spaced.

At P the

by the vector E, perpendicular to the equipotential

now

(at origin

a function

at P.

Thus, the contours are straight,

Consider

is

the absolute potential

x and y, and

uniform.

positive test charge to the point

There

is

electric field is as indicated

line.

the change in potential along an infinitesimal element of path

length in the x direction (y

= constant). Then

= E cos a = E
dx

= angle between E and x axis

Ex = component of E in x direction

where a

Likewise, for an infinitesimal element of path length in the

dV
dy

direction,

= E cos P E

= angle between E and y axis

Ey = component of E in y direction

where

fi

FIGURE

2-17

Potential distribution with electric field

at

a point P.

THE STATIC ELECTRIC HELD: PART

sec. 2-14]

By

the principle of superposition the total field

sum of the component fields

at the point.

Hence,

dV

dV\

dx

dy)

at the point

E = kEx + fEy = -|*-^ + y = -grad

Thus, the gradient in

this rectangular

derivatives of the potential

EXAMPLE
and by

Suppose that

m _1

in the

added

Therefore,

E has

to the positive

the vector

V= -VK

two-dimensional case

in Fig. 2-17 the potential decreases

direction.

a magnitude of 2.24
axis (a

is

35

is

equal to the x and y

vectorially.

Find the

by 2

V m _1

in the

direction

electric field E.

grad V = x2 yl
E = -grad V= x2 + yl = 2.24/26.6 V

and

V m -1 and

is

directed at an angle of 26.6 with respect

= 26.6).

The two-dimensional case discussed above can readily be extended to three

Thus, as shown in Fig. 2-18, there are field components at the origin

dimensions.

in the three coordinate directions as follows

8V
xx = x\Ecosa = -&dx

y,

= ycos/?=-y

dV
dy~

tEz = IE cosy = -t

FIGURE

8V
dz

2-18

Components of

electric field in rectan-

gular coordinates.

'x

36

[CHAP. 2

ELECTROMAGNETICS

By the

principle of superposition the total field

component

fields,

dV

= -(*"
dx

where the relation

origin

is

the vector

sum of the

It is

dV\

+S^.+^)
Bz
By

in the parentheses is the

nates for the gradient of V.

the product of

E at the

or

0)

complete expression in rectangular coordi-

often convenient to consider that this expression

and the operator

V=

del (V).

*A + yA +
By

ex

is

Thus, in rectangular coordinates

The quantity V is a vector operator. It is meaningless

product of V and V yields the gradient of V. That is,

dV

dV

BV

ox

oy

dz

(2)

dz

uJtil applied.

Taking the

VF = *^-+y^- + 2 = gradF=-E

(3)

Expressions for the gradient in other coordinate systems are given inside the back
cover.

2-15

THE ELECTRIC DIPOLE AND

ELECTRIC-DIPOLE MOMENT

The combination of two equal point charges of opposite

distance

is

called

an

electric dipole,

and the product Ql

sign separated
is

by a small

called the electric-dipole

If the two charges were superposed, the resultant field would be zero.
However, when the two charges are separated by even a small distance /, there is a

moment.

finite resultant field (see Fig. 2-13).

By regarding

the separation between the charges as a vector

the negative to the positive charge as in Fig. 2-19, the dipole

as a vector Ql with the magnitude Ql

I,

pointing from

moment can be expressed

and the direction of 1.

Referring to Fig. 2-19, the potential of the positive charge at a point

v.-

4ner 1

The

potential of the negative charge at

is

1-4
4ner
2

P is

THE STATIC ELECTRIC FIELD: PART

sec. 2-15]

37

Axis of
dipole

FIGURE

2-19

Electric dipole.

The

total potential

V at P is then
4ne Vi

If the point
lines r u r,

and

is

at

a large distance compared with the separation

r 2 are essentially parallel,

we have very

cos

where

=
=

and

r2

distance from center of dipole to point

/,

so that the radial

nearly that

= r + - cos

angle between axis of dipole and r

Hence, the potential

at a distance r

from an

electric dipole is

Ql cos e

V=

4ner

(V)

assumed that r is much greater than

neglected compared with those in r 2
where

r2 /

it is

(1)

/ (r

> /)

so that terms in

1
I

can be

It is instructive

moment, the
employed.

to consider (1) as the product of four factors involving the dipole

angle, the distance,

and a constant

characteristic of the system of units

Thus,

V=

Ql

cos0
r

Dipole

Angle

Distance

moment

factor

factor

Ane

(2)

Constant

Expressions for the potential and also the electric field of dipoles (and quadrupoles
and higher-order configurations) always contain these four kinds of factors.

38

ELECTROMAGNETICS

To

[CHAP. 2

find the electric field of the dipole of Fig. 2-19

Thus,

the gradient discussed in the preceding section.

potential given

by

18V

.dV

=
=
=

convenient to

make

use of

us take the gradient of the

dO

~Ql

Ql cos

dr

let

obtainingf

(1),

E:
where

it is

2ner

sin 6

Aner

(3)

unit vector in r direction (see Fig. 2-20)

unit vector in

direction

separation of dipole charges

FIGURE

2-20

Component fields and total

distance r from an electric

According to

(3),

the electric field has

field

at a

dipole.

two components, as shown

Thus

in Fig. 2-20,

one in

the r direction (,) and one in the 9 direction (Ea ).

E=
or

,=

Ql cos e

and

2ner 3

tE,

+ 0

(4)

E =

In these equations the restriction applies that r

Ql

sinfl

-A^~

>

(Vm

>

(5)

/.

By a similar procedure to that followed above it is possible to obtain the potential

and field relations for more complex charge configurations, e.g., the quadrupole or
octopole shown in Fig. 2-21 (see Probs. 2-15 and 2-16). The potential and field variations with distance for these configurations are listed in Table 2-1, along with the

distance factors for a dipole

t See inside

and a

single charge.

back cover for gradient

We

note that the higher the order

in spherical coordinates.

THE STATIC ELECTRIC FIELD: PART

sec. 2-16]

39

Octopole

Quadrupole

e
FIGURE

2-21

Charge configuration (a) for a quadrupole and (ft) for an octopole.

(o)

(b)

of the charge configuration the more rapidly the potential or field falls off with distance.

We

also note that for

any given configuration the

field falls off

more

rapidly than the

potential.

DISTANCE FACTOR OF ELECTRIC

POTENTIAL AND ELECTRIC FIELD
FOR DIFFERENT CHARGE CONFIGURATIONS

Table 2-1

Electric

Configuration

potential

Electric

field

Single charge

Dipole

Quadrupole
Octopole

2-16

ELECTRIC FLUX

Referring to Fig. 2-22, imagine that the two point charges

+ Q and Q are connected

by tubes of electric flux or chargelike flow between them.

any tube

is

\l>

where

D = average
A=
More

The

electric flux

ij/

through

a constant given by

flux density,

= DA

Cm"

cross-sectional area of tube,

generally

(C)

m -2

we have

^ = |JD-rfs = \\Dn

ds

(C)

(1)

40

ELECTROMAGNETICS

[CHAP. 2

is

^
FIGURE

2-22

two point charges

of opposite sign with flux tubes joining
them. The tubes follow the electric field
Electric field between

lines.

Each tube has a constant amount

Infinite plane

of flux.

This relation states that the electric flux through any surface equals the integral of the
flux density over the surface.
In Fig. 2-22 ij/ is constant along any tube.
If all space
is filled

with flux tubes connecting

+Q

and

- Q,

the total flux through

passing through the infinite plane separating the charges

integrating

D over the

infinite

is

all

equal to Q.

the tubes

Likewise

plane yields Q, or

*=JjD-*=e

(C)

(2)

infinite

plane

Consider next the

This

as in Fig. 2-23|.

electric field

is

around a

single isolated positive point charge

E=*

or

Aner

The second equation has

density.

It

given by

cE

4w 2

the dimensions of charge per unit area, or electric flux

follows that

n=t

4nr 2

D = cE

or

We

(3)

(4)

imagine that the field lines end on an equal negative charge distributed over the
of a sphere at infinity.

inside

THE STATIC ELECTRIC FIELD: PART

sec. 2-17]

FIGURE

41

or

2-23

Electric field of an isolated positive point

charge Q. The flux density

r is the

sity

same

D at a radius

as the surface charge den-

on the sphere

if

were uniformly

distributed over the sphere.

where

D = electric flux density,

Cm -2

= permittivity of medium, F m _1
E = electric field intensity, V m -1

According to
This

is

(4),

the flux density and field intensity are vectors with the

true for all isotropic media,

i.e.,

same direction.
media whose properties do not depend on

direction.

Since 4nr

of

equals the area of a sphere of radius

D at the radius r is identical

the charge

r, it

follows that the magnitude

with the surface charge density which would occur

if

were distributed uniformly over a sphere of radius r instead of concen-

trated at the center (see Fig. 2-23).

2-17

ELECTRIC FLUX OVER A CLOSED SURFACE;

GAUSS'

LAW

Referring to Fig. 2-24,

let

imaginary sphere of radius

surface element ds

is

a positive point charge

r.

The

infinitesimal

be situated at the center of an

amount of

electric flux

through the

given by
dt//

=D

ds

Integrating this over the sphere of radius r then gives the total flux through the sphere

$ = jj D

ds

(1)

42 ELECTROMAGNETICS

FIGURE

[CHAP. 2

2-24

Point charge

at origin

of spherical

coordinate system.

We

note that

D is everywhere normal to the sphere, so that

D

where

D=

D|

ds

ds

From

(1)

and

= D ds

ds

= scalar magnitude of the vector D

= scalar magnitude of the vector ds

(2-16-3)

*SI Anr
From

Fig. 2-24 ds

(r dff)(r d(/> sin 6)

ds

sin d

(2)

d6

d4>,

tended by the spherical element of surface area ds

and the
is

dsjr

solid angle dil sub-

dl

sin

dO

d<f>;

therefore

<]/

$- [[do.
4n J J

An

Q
= ^-[-cos0]S
4n

Thus, the total

electric flux

D=

Jo

d<j>

d0 =

Q
^x2x27t
=S
An

(3)

over the sphere (obtained by integrating the normal

flux density

2/4t'-2 by the area of the sphere (4nr 2 ).

serves to illustrate a

D is

2"

f sin e de

JQ

D over the sphere) is equal to the charge Q enclosed by

We could have obtained the result in this case more simply by multiplying

component of the
the sphere.

f
JQ

However, the above development

also be applied to cases where

more general procedure which can

not constant as a function of angle.

THE STATIC ELECTRIC FIELD: PART

SEC. 2-18]

The

result in the

above example

a statement of Gauss' law for a special

is

43

case.

A general statement 6f Gauss' lawf for electric fields is:

The surface integral of the normal component of the
any closed surface equals the charge enclosed.
Thus, in symbols

D cos 9 ds = JJ D

ds

= Q

electric flux density

over

(C)

(4)

where Q is the total or net charge enclosed. This charge may

volume integral of the charge density p, so that

also be expressed as the

JJ

jjD-

A-ffjpdv-Q

(5)

where the surface integration is carried out over a closed surface and the volume
integration throughout the region enclosed.
An alternative notation is

<j>

where s
v

=
=

= i>

ds

To

(6)

or volume, integral throughout region enclosed

is

the basic theorem of electrostatics.

of the inverse-square law (Coulomb's law).

l/r

= Q

double, or surface, integral over closed surface

triple,

Gauss' law

vary as

p do

2
,

Thus,

the total flux over a surface enclosing

illustrate

Gauss' law, several situations

It is

if

will

it

a necessary consequence

for a point charge did not

would not equal the charge.

be analyzed with

its

aid in the

following sections.

2-18

SINGLE SHELL OF CHARGE

Referring to Fig. 2-25a, suppose that a positive charge

an imaginary spherical
by integrating

shell

of radius r t

Q is

The medium

uniformly distributed over

is air.

Applying Gauss' law

D over a spherical surface (radius r dr) just inside the shell of charge,
x

we have

E-ds =

6<t

(1)

Js

since the charge enclosed

is

zero.

It

follows from symmetry that

Propounded by Karl Friedrich Gauss

in 1813.

inside the shell

44 ELECTROMAGNETICS

[CHAP. 2

'Spherical shell of

charge

+Q
(a)

(6)

FIGURE

2-25

Uniformly charged spherical shell with

graphs showing variation of electric
field intensity E, electric potential V, and
surface charge density p, as a function of
radial distance

is

zero.

(d)

r.

Applying Gauss' law to a spherical

shell (radius r y

dr) just outside the shell

of charge, we have, neglecting infinitesimals,

ej E-ds =

E4nr l 2

= Q

(2)

E=

or

7his value of field intensity

is

(3)

identical with that at a radius r,

from a point charge Q.

We can therefore conclude that the field outside the shell of charge is the same as if the
charge Q were concentrated at the center. Summarizing, the field everywhere due to
a spherical

shell

of charge

is

E=
E=
The

variation of

E
t

inside

4ne r 2

as a function of r

is

<

r,

outside

illustrated

(4)

by

>

r,

(5)

Fig. 2-256. f

Note that a point charge at the origin gives an infinite E as r->0 but a surface
charge of finite area at a radius r t gives a finite E as r-vri. This is because the
volume charge density p of a point charge is infinite, whereas the surface charge
density p* of the shell of charge is finite. In the present case ps
QUirr^.

SEC. 2-19]

THE STATIC ELECTRIC FIELD: PART

The absolute

potential at a radius r outside the shell

V= Introducing the value of

E from

"

dx

(6)

Q
the shell, where r

r 1(

4ne

inside the shell

therefore the potential

is

is

(Vmr

(7)

we have

VSince

given by

(5),

4ne

At

is

45

zero,

it

(8)

no work to move a

requires

test

charge inside and

constant, being equal to the value at the shell.

Summarizing,

the electric potential everywhere due to a spherical shell of charge of radius r

1

V=

inside

<

r,

(9)
v
>

47t ri

V=

outside

4ne

>

is

<\Q\
v
'

r,

The variation of V as a function of r is illustrated by Fig. 2-25c. The variation of

the surface charge density p s is shown by Fig. 2-25d. The surface density is zero
everywhere except at r = r u where it has the value 2/4W, 2 , as indicated by the vertical
or spike.

line,

It is

to be noted that the potential

at the shell

(/

However, the

r^.

is

continuous, both (9) and (10) being equal

electric field

is

discontinuous, jumping abruptly

from zero just inside the shell to a value Q/4ne r t 2 just outside the
from the assumption that the shell of charge has zero thickness.

2-19

shell.

This results

CONDUCTORS AND INDUCED CHARGES

A conductor can conduct, or convey, electric charge.

In static situations a conductor

may

is

all

be defined as a

medium

parts of a conductor

aluminum, and

When

in

which the

must be

silver are

at the

electric field

same

potential.

always zero.

It

follows that

Metals such as copper, brass,

examples of conductors.

a metallic conductor

is brought into an electric field, different parts of

the conductor would assume different potentials were it not for the fact that electrons
flow in the conductor until a surface charge distribution is built up that reduces the
total field in the conductor to zero.f
This surface charge distribution is said t o consi st

The

electrons in the outermost shell of the

that they migrate readily

atoms of a conductor are so loosely held

from atom to atom under the influence of an electric field.

46...

[CHAP. 2

ELECTROMAGNETICS

The

of induced charges.

field in

which the conductor

may

placed

is

be called the

applied field E, while the field produced by the surface charge distribution
called the induced field E,
field in

The sum of the applied and induced

Although the

the conductor equal to zero.

total field inside the

zero after the static situation has been reached, the total

induced charges are in motion,

To summarize, under
and

its

potential

is

i.e.,

field is

a total

conductor

is

not zero while the

while currents are flowing.

static conditions the electric field in

a constant.

may be

fields yields

Charge may

reside

on

a conductor

is

zero,

the surface of the conductor,

and, in general, the surface charge density need not be constant.

Conducting
shell

FIGURE

2-26

Conducting shell of wall thickness b

with point charge Q at center.

2-20

An

Induced charges

CONDUCTING SHELL
uncharged conducting

initially

thickness b

a) is

shown

a hole in the shell which

of inner radius a and outer radius b (wall

shell

is

plugged after the charge

in the conducting wall to be zero requires

E
or

+
E,

The induced

field

is

is

E,

an induced

be

The point charge has

inside.f
field

field

Ea For the total field E

E inside the wall such that
.

=E=

(1)

= -E

(2)

produced by a distribution of induced negative charges on the

inner shell wall and induced positive charges

Fig. 2-26.

+Q

This might be done by introducing the charge through

Let this be called the applied

radial electric field.

Let a point charge

in cross section in Fig. 2-26.

placed at the center of the shell.

a*

on the outer

shell wall, as suggested in

Let us apply Gauss' law to this situation to determine quantitatively the

magnitude of these induced charges.

t This

an

is

an idealized version of an experiment

ice pail.

first

performed by Faraday, using

THE STATIC ELECTRIC FIELD: PART

sec. 2-20]

47

Conducting
shell

FIGURE

2-27

Conducting

shell of wall thickness b

with surfaces of integration.

Suppose that an imaginary sphere designated S t with a radius a

just inside the inner wall of the shell as in Fig. 2-27.
integral of the

normal component of

Applying Gauss' law to the sphere

have, since the total field

in the

S2

g?s

of radius a

is

situated

the surface

+ Q.

That

= +Q

fife

conductor

dr

By Gauss' law

over this sphere must equal

is

(3)

dr just inside the conductor,

we

(4)

zero,

cf>

is,

ds

Si

Thus, the total charge inside the sphere

is

situated

on the inner surface of the

S2 must be zero.
shell wall.

It

follows that a charge

Since the shell was originally un-

charged, this negative charge Q, produced by a migration of electrons to the inner

surface,

of the

must leave a deficiency of electrons, or positive charge Q, on the outer surface

It is assumed that the surface charges reside in an infinitesimally thin

shell.

layer.

Applying Gauss' law to the sphere

surface of the shell,

S3

of radius b

dr just outside the outer

we then have

ds

= +Q

(5)

S3

To summarize,

+ Q at the center of the shell induces an exactly

Q on the inner surface of the shell, and this in turn results

the charge

equal but negative charge

in

an equal positive charge

The

flux tubes originating

inside of the shell.

on

were present.

distributed over the outer surface of the shell.

no total field and no flux in the shell wall. Outside the

from the charge +Q on the outer surface as though no
The variation of the component fields E (applied) and E,

There

is

shell the flux tubes continue

shell

(+Q)

+Qal the center end on the equal negative charge on the

48

[CHAP. 2

ELECTROMAGNETICS

Conducting shell

Potential

FIGURE

2-28

Conducting shell of wall thickness b a

with graphs showing variation of applied
induced field ,, total field E,

field ,

potential

K, surface

with charge

Surface
charge
density

charge density p s

and

at center (b, c, d,

e)

and with charge only on outside of shell

and /)

(/, g, h,

(induced) as a function of r
field

is

shown

If a

illustrated

is

by

Fig. 2-2Sb.

The

variation of the total

in Fig. 2-2Sc.

conducting wire

is

connected from the inner surface of the

shell to the

charge

+ Q at the center, electrons will flow and reduce the charge at the center and on the
inner surface to zero.
However, the charge + Q remains on the outer surface of
This
results
in
an applied field only external to the shell (r > b) and of the
the shell.
same value as
is

before.

There

is

no induced

identical with the applied field

This

final result

the charge

The

+Q
is

is

field

zero for r

whatsoever.

<b

as

Thus, the total

shown by

might have been achieved more simply in the

first

field

Fig. 2-28/ and g.

place by applying

to the outside of the originally uncharged conducting sphere.

variation of

of the sphere

and

V and ps as

a function of r

only on the outside of the shell

is

when the charge + Q

is

at the center

and e, while the variation when the charge

shown by Fig. 2-28A and i.

indicated in Fig. 2-28d

is

THE STATIC ELECTRIC FIELD: PART

SEC. 2-21]

CAPACITORS AND CAPACITANCE

2-21

capacitor^

an

is

By

of two conductors separated by an

electrical device consisting

insulating or dielectric

its

49

medium.

definition the capacitance of a capacitor

the ratio of the charge

is

conductors to the potential difference between them.

of a capacitor

on one of

is

Q=c

v
where

Thus, the capacitance

(1)

Q = charge on one conductor

V = potential difference of conductors

Dimensionally

(1) is

Charge

charge

energy/charge

Potential

charge

= capacitance

energy

or in dimensional symbols

The

SI unit of capacitance

is

MI} jT 1

_ I 2 T*
~ ML2

the farad (F).

Thus,

C V -1

equals

F, or

Coulombs
Farads
Volt
In other words, a capacitor that can store

has a capacitance of

C charge with a potential difference of 1 V

Since a capacitor of

F.

ordinarily used in practice, microfarads (juF)

FIGURE
difference

capacitor

and

with

voltage

plate charge Q.

area of each plate

separation is d.

is

The

A, and the plate

EXAMPLE A parallel-plate capacitor consists

by distance

d, as

shown

capacitance

t Formerly called a condenser.

much

larger than

is

T
j

'

f
i

_Jr_

-Q

of two

in cross section in Fig. 2-29.

find the capacitance.

is

+Q
~

2-29

Parallel-plate

and picofarads (pF) are commonly used.

flat

metal sheets of area

If the

medium has a

separated

permittivity

e,

50

ELECTROMAGNETICS

solution

If

[CHAP. 2

a voltage

V is applied

between the

V= Ed
where

E is the electric field

between the plates.

Q = DA
where
tance

is

the flux density between the plates.

plates,

we have

(V)

The charge on one

plate

is

given by

(C)

But

D=

eE; so

6= eEA,

and the capaci-

is

Ed

Thus the capacitance is proportional to the permittivity e, and the plate area A and is inversely
proportional to the plate-separation distance d. It was assumed that the field was uniform
between the plates and zero outside (no fringing). Actually any fringing effects will be small
if the plates are large compared to d.

2-22

BOUNDARY RELATIONS AT A
CONDUCTING SURFACE
a thin imaginary volume element be constructed at the
The medium outside the conductor may, for example, be air.
The volume element has
half in the conductor and half in air.

Referring to Fig. 2-30,

surface of a conductor.

The volume element

an area

is

let

parallel to the

normal to the surface.

conductor surface but has an infinitesimal thickness dl

According to Gauss' law, the normal component of the

FIGURE

2-30

Conductor-air boundary with cross section through small volume element half
in the conductor and half in air.

Conductor

flux

THE STATIC ELECTRIC FIELD: PART

sec. 2-22]

density

ps

is

D over the volume element must equal the total charge Q enclosed.

the surface charge density,

Now D in the conductor is zero,

n

Thus,

if

we have
<j>

51

Dds= Q = p

and so the

(1)

integral reduces to the

normal component

Hence

of the flux density in air multiplied by the area A.

DA=p A
s

Dn=P

or

(2)

This important boundary relation states that the normal component of the flux density

D at a conducting surface equals the surface charge density.

Both D and p s have the

dimensions of charge per area and are expressed in coulombs per square meter
If a thin

conducting sheet

is

introduced normal to an electric

on the sheet so that the

are induced

The value of the induced

D at the sheet.

field,

original field external to the sheet

surface charge density

s is,

by

(2),

Hence, one can interpret the flux density

(Cm

2
).

surface charges
is

undisturbed.

equal to the flux density

D at a point as equal to the

charge density p s which would appear on a thin conducting sheet introduced normal
to

at the point.

Referring, for example, to the thin conducting sheet normal to

the field in Fig. 2-31, the relation of

On
On

left side:

right side:

D and p

where n

is

left side

and normally outward on the

each side

s is

the unit vector normal to the surface.

is

as follows:

D = ~Aps
D = +&ps
right.

equal to the charge density p s

Thus

D is

normally inward on the

The magnitude of the

flux density

on

-Thin sheet

FIGURE

2-31

Thin conducting sheet placed normal to

field has an induced surface charge density

p s equal to the

field at the sheet.

densities

flux density

of the

The surface charge

on the two sides of the sheet are

equal in magnitude but opposite in sign.

Surface charge
density p.

52

[CHAP. 2

ELECTROMAGNETICS

PROBLEMS
Note

2-1

everywhere.

Find the force

F on

a positive point charge of 100

pC at a distance 100 mm from

a positive point charge of 10 nC.

A positive point charge of 10 pC is located at x = 0, y = 100 mm and another such

= 0, y = ~200 mm. Find the magnitude and direction of the force F on a
positive point charge of 10 nCatx= 300 mm, y = 0.
*
2-3 Find the electric field intensity E and the absolute potential Kat a distance 100 mm

2-2

charge at x

from a

positive point charge of 10 nC.

2-4

Two positive point charges of 1 C each are separated by 2a.

between find the

2-5

by

A positive point

At the point halfway between

2a.

At the point halfway

E (magnitude and direction) and electric potential V.

charge of 1 C and a negative point charge of 1 C are separated

electric field

find the electric field

(magnitude and direction) and

the electric potential V.

2-6
are 3

1.6

point charge of

/u.C is

at the origin,

m in the +y and y directions.

find the electric field

2-7

at the point

charge of

2-8

2-9

+3 nC

= 0,y = ~2m.

electric potential

At a point 4
(magnitude and directiop) and

V at

the point

is

at the point

Find the

positive point charges of

the

^C

direction

electric potential V.

= 0j = 2m,

electric field

= 0, y

and

m from the origin in

E

and a charge of

nC
1

is

(magnitude and direction) and the

m.

Find the unit normal to the plane 3x + 4y + 5z = 10.

A potential distribution is given by V = 6x + 3 V. What

is

the expression for

E?

What is its vector value (magnitude and direction) at the point (0,0) and at (10,0), that is,
x = 0, y = and x = 10, y = 0?
*
2-10 A potential distribution is given by V=3y ll2 \. What is the expression for E ?
What is its vector value (magnitude and direction) at the points (0,0), (4,0), and (0,4) ?
What is the expression
*
2-11 A potential distribution is given by V = 7y 2 + \2x V.
for E? What is its vector value (magnitude and direction) at points (0,0), (5,0), (0,3), and
(5,3)?

2-12
sion for

E?

A potential distribution is
What

is its

given by

V= 5l(x + y + z
3

3
)

V.

What

is

the expres-

vector value at the points (0,0,4) and (5,4,3) ?

2-13

A potential distribution is given by V= kr in

2-14

Two

sin 6.

Find E.

small-diameter 10-g dielectric balls can slide freely on a vertical plastic

Each ball carries a negative charge of 1 /xC. (a) Find the separation if thelowerball
from moving. (6) What is the dipole moment ?
2-15 (a) Show that at a large distance from a quadrupole (as in Fig. 2-2\a) the potential

thread.
is

restrained

is

given by
2
C/ sin0cos0

27T r

* Answers to starred problems are given in Appendix C.

THE STATIC ELECTRIC FIELD: PART

PROBLEMS]

where

Q=

53

individual charge

= spacing between charges

r = radial distance
= angle from x or y axis to
/

Find the

(6)

radial line (quadrupole oriented as in Fig. 2-21 a)

electric field E.

As an extension of Prob. 2-15 find the potential Fand

from an octopole as in Fig. 2-216.

2-16
distance r

electric field

E at

a large

A charge of +2Q is located at the point x = 0, y = and a charge + Q at the

= 0, y =
Find the point for which E = 0.
2-18 A charge of +2Q is located at the point x = 0, y=
and a charge Q at
the point x = Q, y = \. Find the line in the xy plane for which V = 0. Draw a graph.
2-/P A dipole in a uniform field experiences no translational force. However,
2-17

point

it

does experience a torque tending to align the dipole axis with the
dipole of moment ql'ma uniform field
the dipole axis

and the

0,

where d

is

that for a

the angle between

field.

nC is uniformly distributed throughout a spherical

E and K everywhere, (b) Draw a graph of "and V
a function of radius from the center of the sphere to a radius of 400 mm.
2-21 A sphere of radius R has a charge-density function p = kr 2
Find E and Kin2-20

positive charge of 10

volume of radius
as

E this torque is qlE sin

Show

field.

R = 100 mm.

(a)

Find

and outside the sphere.

side

2-22

Note

that four answers are required.

shell of thickness r 2

r,

has a charge-density function p

E and Kfor all values of r. Note that six

and V as a function of r to a distance 3r 2

answers are required,

= kr.

Draw

(b)

Find

(a)

a graph of

2-23

Show that the vectors A = x + 4y +

2-24

Find

= 45 and

*
y=

<f>

A B

where

A = 3i + 2y + z

3z and

and

B = 4x + 2y 4z are perpendicular.
is

a unit vector in the

r direction at

30.

E = ibc + yy + zz,

2-25

If

2-26

Point charges are situated as follows

m; and

- 30 nC

at

find the total electric flux over a sphere of radius R.

= 0, y = 4

m.

(a)

Find

10

nC

V, E,

at

x=

0,

at

and

= 0, 20 nC at x = 3,
= 3, y = 4 m. Give

1, 3.5, and 5 m?
volume of side length 2/ has a uniform charge density p. Find
the integral of the normal D over one face ( / < x < /; l<,y<^02-28 Using Gauss' law, find the total electric charge inside a cubical volume 1 m on a
side situated in the positive octant with three edges coincident with the x, y, z axes and one
corner at the origin if the electric flux density D is given by (a) D = x(x -I- 3) and (b) D =

units.

(6)

2-27

S(y

is

the total electric flux over spheres of radius

cubical

+ 4).
2-29

on a

What

Using Gauss' law, find the

total electric charge inside a cubical

2
at the origin if the electric flux density D is given by (a) D = x2x
2 2 2
3 3
=
=xxyz
ix
z
D x(x + 3) + f(y + 4) + z(z + 5); (d) D
+ yx y z + y

one corner
(c)

volume 2 m
and

side situated in the positive octant with three edges coincident with the x, y, z axes
;

(Jb)

D = xxyz;

2-30

Eight equal positive or negative point charges are located at points (x,y)'of

(100,100), (100,250), (100,400), (250,100), (250,400), (400,100), (400,250),

Using a computer program and one of the charge configurations of

and (400,400) mm.

Fig.

P2-30 (or other

54

ELECTROMAGNETICS

[CHAP. 2

FIGURE P2-30
Computer-generated potential contour maps for 12 different configurations of 8
charges of the same magnitude with signs as indicated. The heavy line is the zeropotential contour.

map of the area < x < 500, <y <

mm and label all contours with the proper potential (voltage). The actual map can be
drawn to some convenient jize smaller than 500 by 500 mm by applying a scale factor. All

combinations), obtain a computer-plotted contour

500

charges are

nC, so that the potential at any point

4n

**

is

(V)
r

the static electric field: part

problems]

Make

the contour interval

or 2 V.

intervals except in interesting regions,

55

In Fig. P2-30 contours are shown only at 5- or 10-V

where the interval is

The heavy lines in Fig. P2-30

V.

are the zero-potential contours.

Referring to the potential contour

2-31

map of Prob.

points (50,50), (50,450), (150,150), (150,350), (200,200),

2-30, determine the field

E at the

(300,200),

(200,300), (250,250),

and (450,450) mm by noting the contour orientation

(VK) and also by the vector addition of the field contributions from each of the
Compare the values of E obtained as the gradient of V from the map and by
eight charges.
vector addition of the field contributions from the individual charges. Add arrows of appropriate length and orientation on the map to indicate the value of the field E at the above
(300,300), (350,150), (350,350), (450,50),

and

interval

points.

Complete the

2-32

map of Prob.

2-30 by drawing the electric

field lines

(with arrows)

not a two-dimensional problem (V varies perpendicular to page), so that an area bounded by Kand E lines for a constant voltage increment

as in Figs. 2-13

will

and

2-14.

Note

that this

is

not form curvilinear squares over the entire

2-33

If the

map

(see Sec. 3-22).

mass of the earth were uniformly distributed

of the same radius as the earth, show that the gravitational

field

a thin spherical shell

in

would be unchanged

outside

but inside would be zero.

2-34

(a)

What electric charge would be required on the earth and the moon

to balance

same ratio as the masses? Take the

mass of the earth as 6 x 10 24 kg, of the moon as 7 x 10" kg, and their separation as 400 Mm.
_
-2
2
The gravitational constant in the analog to Coulomb's law is 6.7 x 10 " N m kg

their gravitational attraction if the charges

were

in the

(b) If

the charges were of opposite sign what

2-35

An

2-36

The

is

the dipole

moment ?

Find its capacitance.

surrounded by an ionized shell, or ionosphere. If it is assumed to
be equivalent to a conducting shell at a height of 500 km, what is the capacitance of the earthshell

isolated metal sphere has a radius R.

earth

is

combination?

Show

2-37

on the inverse-square

that Gauss' law depends

law, that

is,

Eccl/r 2

for a point charge.

Two

2-38

concentric shells have radii r t and r 2

uted over the shell of radius

radius r 2

ru

and charge

Apply Gauss' law to

+ Q2

is

find (a) the field

Charge

+Q

is

uniformly distrib-

uniformly distributed over the shell of

everywhere and

(6) the potential

everywhere.

2-39

What

2-40

A thin-walled

is

the capacitance of the earth?

metal sphere of 10

liters

Take the
capacity

earth's radius as 6.4

is filled

Mm.

with nitrogen at atmos-

If one electron is removed from each molecule of the gas (and removed to a
from
the sphere) find the potential of the sphere. The medium outside the
large distance
sphere is air. Take the density of the nitrogen gas as 3 x 10" molecules per liter.
Find their capaci2-41 Two concentric spherical metal shells have radii r, and r 2

pheric pressure.

tance.

The
2-42

machine.

shell thickness is negligible.

Describe

how electrostatic fields

See, for example, A.

D. Moore,

are employed in the operation of a xerographic

Electrostatics, Sci.

Am., 226: 47 (March 1972).

THE STATIC ELECTRIC FIELD: PART

3-1

INTRODUCTION

This chapter extends the theory introduced in Chap. 2 to include material media.'
The conditions of homogeneity, linearity, and isotropy are pointed out. The influence
of a dielectric medium on field distributions is discussed, and the concepts of electrostatic

energy and energy density are introduced.

The fields of several charge distribuIntroduction ofthe divergence concept leads to Maxwell's divergence
equation in D, which is then applied to the parallel-plate capacitor.
tions are found.

3-2

HOMOGENEITY, LINEARITY, AND ISOTROPY

A medium

is

homogeneous

if its physical characteristics

(mass density, molecular

do not vary from point to point. If the medium is not homogeneous,

may be described as inhomogeneous, nonhomogeneous, or heterogeneous.
A medium is linear with respect to an electrostatic field if the flux density D is

structure, etc.)
it

proportional to the electric field intensity E.

This

D =e

is

E.

Here the factor

permittivity e

nonlinear.

may

or permittivity,

not always be constant.

If

is

the case in free space, where

a constant.

it

is

In material media the

not, the material

is

said to be

THE STATIC ELECTRIC FIELD: PART 2

SEC. 3-3]

An

one whose properties are independent of direction.

isotropic material is

Generally, materials whose molecular structure

However,

crystalline

57

is

randomly oriented

media or certain plasmas may have

will

be isotropic.

directional characteristics.

Such materials are said to be nonisotropic or anisotropic.

book concepts are usually developed first for the case where the medium
linear, and isotropic.^
Later the ideas may be extended to cases where
one or more of these restrictions no longer hold.
In this

is

homogeneous,

DIELECTRICS

3-3

AND PERMITTIVITY

In a conductor the outer electrons of

an atom are

from atom to atom under the influence of an

easily detached

and migrate

readily

In a dielectric, on the other

electric field.

hand, the electrons are so well bound or held near their equilibrium positions that they

cannot be detached by the application of ordinary

field

produces no migration of charge in a

makes

good

dielectrics act as

insulators.

Hence, an

electric fields.

dielectric,

electric

and, in general, this property

Paraffin, glass,

and mica are examples of

dielectrics.

An

important characteristic of a dielectric

mittivity of a dielectric

is

is its

permittivityl

convenient to use the relative permittivity er of the dielectric,

permittivity to that of

er
e

Whereas
is

or

is

a dimensionless

The

often
its

(1)

m _1
m _1

),

the relative permittivity e r

ratio.

relative permittivity
is

is

the value ordinarily given in tables.

shown

in

low-frequency) fields and, except for

air is

it is

the ratio of

-e

expressed in farads per meter (F

permittivity of a few media

er for

= relative permittivity of dielectric

= permittivity of dielectric
= permittivity of vacuum = 8.85 pF
e

i.e.,

Thus

vacuum.

where

Since the per-

e.

always greater than the permittivity of vacuum,

so close to unity that for

Table

3-1,

vacuum or

The

where the values are for

air,

are approximate.

most problems we can consider

relative

static (or

Note that

air equivalent to

vacuum.
t Dielectrics of this type are sometimes designated Class A dielectrics.
However, the permittivity is not always a conJ Also called the dielectric constant.
stant, as one might be led to infer by this term (see Chap. 8).

PERMITnvnTES OF DIELECTRIC MEDIAt

Table 3-1

Relative

Medium

permittivity e.

Vacuum

It
.0006

Air (atmospheric pressure)

Styrofoam

1.03

Paraffin

2.1

Plywood

2.1

Polystyrene

2.7

Amber

3.0

Rubber

Plexiglas

3.4

Dry sandy
Nylon

soil

3.4

(solid)

3.8

Sulfur

Quartz

5
5

Bakelite

Formica
Lead glass
Mica
Marble

6
6
6

10

Flint glass

Ammonia
Glycerin

22
50

Water

81

(liquid)

(distilled)

Rutile (Ti0 2 )

89-173

Barium titanate (BaTi0 3 )

Barium strontium titanate (2BaTi0 3 lSrTi0 3 )
Barium titanate zirconate (4BaTi0 3 lBaZr0 3 )
Barium titanate stannate (9BaTi0 3 lBaSn0 3 )
:

1.200K
lO.OOOTf

13,000ft
20,000tt

For a comprehensive tabulation of permittivities, see A. R.

"Dielectric Materials and Applications," pp. 301-370,
Press, Cambridge, Mass., 1954.

By

Von
The

Hippel,

M.I.T.

definition.

Crystals, in general, are nonisotropic;

direction.

substance.

Its

i.e.,

their properties vary with

an example of such a nonisotropic crystalline

relative permittivity depends on the direction of the

Rutile

is

applied electric field with relation to the crystal axes, being 89 when the
field is perpendicular to a certain crystal axis and 173 when the field is
parallel to this axis.

crystals

e,

14.

For an aggregation of randomly oriented

rutile

All crystals, except those of the cubic system, are

nonisotropic to electric fields; i.e., their properties vary with direction.

Thus, the permittivity of many other crystalline substances may vary
with direction. However, in many cases the difference is slight. For
example, a quartz crystal has a relative permittivity of 4.7 in one direction and 5.1 at right angles. The average value is 4.9. The nearest
integer is 5, and this is the value given in the table.
H The permittivity of these titanates is highly temperature-sensitive.
The above values are for 25C. See, for example, E. Wainer, High
Titania Dielectrics, Trans. Electrochem. Soc, 89 (1946).
ft K. W. Plessner and R. West, High-permittivity Ceramics for Capacitors, in J. B. Birks and J. H. Schulman (eds.), " Progress in Dielectrics,"
vol. 2, John Wiley & Sons, Inc., New "York, 1960.

58

THE STATIC ELECTRIC FIELD: PART 2

SEC. 3-5]

In free space, the electric

field inside
if

we

field is

space

electric field in free

But,

THE ELECTRIC FIELD IN A DIELECTRIC^"

3-4

dielectric

is

defined as force per unit

or other material

charge/ This implies

However,

a measurable quantity.

medium may be

very

to

that the

measure the

difficult

59

electric

or impractical.

confine our attention to the external effects of the dielectric, such internal

measurements become unnecessary provided a theory can be formulated for the behavior of the dielectric which produces agreement with external measurements. Thus,
a distinction should be
free space)

and an

made between an

electric field as

chapter a theory for the

to the external field

3-5

electrostatic field in

dielectric is

a dielectric).f

In this

developed and then related

by means of boundary conditions.

no migration of charge when a

is

dielectric is placed in

an

electric

there does occur a slight displacement of the electrons with respect to their

nuclei so that individual

atoms behave

When

like very small, or atomic, dipoles.%

these atomic dipoles are present, the dielectric

state

a measurable quantity (as in

(as in

POLARIZATION

Although there
field,

electric field as

a theoretical quantity

is

said to be polarized or to be in a

When the field is removed and the atoms return to their normal,

of polarization.

or unpolarized, state, the dipoles disappear.

In the above simple picture an
electric dipole,

i.e.,

atom of a

dielectric material is represented

by an

a positive point charge representing the nucleus and a negative

point charge representing the electronic charge, the two charges being separated by a
small distance.

The

electrons revolve in orbits

around the nucleus and act

When

negatively charged cloud surrounding the nucleus.

the cloud surrounds the nucleus symmetrically, as in Fig. 3- la,

is

like

atom is unpolarized,
and the dipole moment

the

zero (the equivalent positive and negative point charges have zero displacement).

Under the

influence of an electric field the electron cloud

asymmetrical, as in Fig.
picture, the

3-lfc,

and the atom

is

becomes

polarized.

slightly displaced or

According to our simple

atom may then be represented by the equivalent point-charge dipole of

Fig. 3-lc (dipole

moment =

Ql).

Electricity and Magnetism at the College Level (Report of the

Coulomb's Law Committee of the American Association of Physics Teachers),

The Teaching of

Am. J. Phys., 18 (1):5 (January 1950).

The dipoles may also be of molecular size.
when a field is applied, and this may result

In a liquid the molecules are free to turn

in

relatively large permittivity.

Water is

an example.

When

polarization, in a dielectric persists in the absence of an applied electric field,

the substance

is

permanently polarized and

an example of an electret.

electric crystal is

is

called an electret.

strained piezo-

60

ELECTROMAGNETICS

[CHAP. 3

UNPOLARIZED ATOM

(a)

Negatively
charged cloud

POLARIZED ATOM

FIGURE

(b)

3-1

Unpolarized atom as in (a) becomes

polarized as in (b)

when

-Q +Q

electric field is

applied. Equivalent dipole

is

shown

EQUIVALENT DIPOLE
(Moment =Ql)

e-e

at

(c)

(c).

Consider the dielectric slab of permittivity e in Fig. 3-2a situated in vacuum.

Let a uniform
i.e.,

field

be applied normal to the slab.

induces atomic dipoles throughout the slab.

tive charges

zation

is

This polarizes the

The

of adjacent dipoles annul each other's effects.

to produce a layer of negative charge

dielectric,

In the interior the positive and neganet result of the polari-

on one surface of the slab and a layer

of positive charge on the other, as suggested in Fig. 3-2a.

The

moment

effect

of the atomic dipoles can be described by the polarization

per unit volume.

P=
where p

= QL

is

the net dipole

moment

rectangular volume of surface area

this

P or dipole

Thus,

QL

J
-

(1)

in the

A and

volume

thickness

v.

(v

For example, consider the

= AL) in Fig. 3-2a. For

volume
p

where p sp
faces.

is

QL

Q
(2)

the surface charge density of polarization charge appearing on the slab

Thus,

has the dimensions of charge per area, the same as D.

The value of P

in (2) is

an average for the volume

v.

To

define the

meaning of

P at a point, it is convenient to assume that a dielectric in an electric field has a continuous distribution of infinitesimal dipoles,

i.e.,

a continuous polarization, whereas the

dipoles actually are discrete polarized atoms.

The assumption of a continuous

61

THE STATIC ELECTRIC FIELD: PART 2

sec. 3-5]

Capacitor plates;

*mm r

f-A

Area

4
Charges
on plates

^i&v?
m&-r*
-Dielectric slab-

Induced.
"'J.

->

Charges

+ Charges

on this face

on this face

'.

charges
on faces

y+(-

of slab

^-g-^-^

(b)

(a)

FIGURE

3-2

(a) Dielectric slab in

tric slab in

distribution leads to

uniform electric field.

no appreciable error provided

volumes containing many atoms or dipoles,

now

(6) Parallel-plate capacitor

a continuously polarized

dielectric, the

that

we

restrict

our attention to

to macroscopic regions.

i.e.,

Assuming

value of P at a point can be denned as the

moment/) of a small volume Ay divided by the volume, with the

Aw shrinks to zero around the point. Thus,

net dipole

as

with dielec-

lower part.

limit taken

lim

(3)

At,-.oAt;

Consider

now

a uniform electric

a parallel-plate capacitor with plates

field in

separated by a distance d, as in the cross-sectional view of Fig. 3-26.

between the plates so that the

the upper part of the capacitor

part

is filled

electric field

is

vacuum

There

(or air) with permittivity e

with a dielectric of permittivity

space between the plates, but in Fig. 3-26 there

is

a voltage

E = V/d everywhere. The medium

e.
is

The

dielectric

in

The lower

completely

fills

the

a gap in order to show the charges on

the plates.
In the upper part of the capacitor (Fig. 3-26)

DQ = e E

we have
(4)

62

[CHAP. 3

ELECTROMAGNETICS

D = electric flux density in

where

vacuum (or air-filled)

-1
of vacuum = 8.85 pF m

= permittivity
E = V/d = electric field

intensity,

part of capacitor,

Cm -2

V m -1

In the lower, dielectric-filled part of the capacitor the electric field polarizes
the dielectric causing a surface charge density p sp to appear

on both

faces of the di-

These bound charges induce free charges of opposite sign on the

electric slab.

capacitor plates (compare upper and lower parts of Fig. 3-26).

charge surface density

on the

plates

is

increased by p sp

As a

result the free-

Therefore, in the dielectric

we have

Dd = e E + p
but from

(2)

p sp

= P, and

(5)

si

so

Dd = t E+P
Dd = electric flux

where

C m -2
pF m"

density in dielectric,

= permittivity of vacuum = 8.85

_1
E = electric field intensity, V m
-2
P = polarization (of dielectric), Cm

Equation

(6) implies the

subscript to

(6)

presence of dielectric (because of the

D is redundant.

and so the

term),

Therefore, (6) can be written

D=e E+P
Although developed for the
relation

which applies

In the dielectric

(7)

special case of a parallel-plate capacitor, (7)

is

in general.

we can

also write

D = Dd = eE

1
where
(6)

e is the permittivity

and

(8)

of the dielectric material in farads per meter.

Equating

(8) gives

eE

or

The

a (vector)

ratio

PjE

is

also

= o+

E+P
or

(9)

-=e-

(10)

sometimes written

Xo

(11)

THE STATIC ELECTRIC HELD: PART 2

SEC. 3-6]

where %

Comparing

the electric susceptibility (dimensionless).

is

Thus, the susceptibility /

EXAMPLE

Find the

for

and

(11) gives

= e,-1

(12)

vacuum, for which

er

1.

and polarization

relative permittivity e P

Assume each

Fig. 3-26.

(10)

63

plus or minus sign represents

nCm"

P
!

for the dielectric slab of

and A

SOLUTION

e,

= e =
e

By

inspection of Fig. 3-26

D =*EE =
Do
5

we can

(dimensionless)

write

P = p sp =
This result can also be obtained as follows.

P=D

From
1

m~

nC

2
i

(6)

- E

Therefore,

P = 5-3 = 2nCm" 2
the same as before.

media

In isotropic

and hence

scalar
in

is

D=e E+P

and

same

general, not in the

Thus,

scalar.

is

are in the

same

direction, so that their quotient is a

In nonisotropic media, such as crystals,

direction, so that e

a general relation, while

no longer a

is

D = tE

is

P and E

are,

Chap.

15).

scalar (see

a more concise expression,

which, however, has a simple significance only for isotropic media or certain special
cases in nonisotropic media.

3-6

ARTIFICIAL DIELECTRICS

Certain of the properties of a dielectric material

dielectrics,

radio waves. f

microscopic
scopic

may

be simulated with

artificial

which were developed as a material for lenses for focusing short-wavelength

size.

Whereas the true

size,

the

For example, the

t

W.

dielectric consists

artificial dielectric consists

E.

of atomic or molecular particles of

of discrete metal particles of macro-

artificial dielectric

may

consist of a large

number of

Kock, Metallic Delay Lens, Bell Syst. Tech. J., 27: 58-82 (January 1948). See
J. D. Kraus, "Antennas," p. 390, McGraw-Hill Book Company,

also discussion in
New York, 1950.

64

ELECTROMAGNETICS

[CHAP. 3

Metal
spheres

FIGURE
Slab of

3-3

artificial dielectric consisting

metal spheres in

metal spheres, as in Fig.

3-3,

lattice

of

arrangement.

arranged in a three-dimensional, or

which simulates the arrangement of the atoms of a true

lattice, structure

dielectric but

on a much

larger scaled

The

permittivity of

calculated.

an

This calculation

significance of polarization
electric field

many

artificial dielectric

and

The

field

a true dielectric, and each sphere

I,

of the

of

artificial dielectric consisting

induces charges on the individual spheres as sug-

may

equal to the net dipole

be represented by an equivalent dipole of charge

The

as in Fig. 3-46.

polarization

moment per

P=
where

now be

illustration

Thus, the spheres become analogous to the polarized atoms of

gested in Fig. 3-4a.

(3-5-1),

of metal spheres will

application in a practical problem. Let a uniform

its

be applied, as in Fig. 3-3, to a slab of

metal spheres.

q and length

made

approximate but provides a good

is

of the

by

artificial dielectric is,

unit volume, or

Nql

0)

N = number of spheres per unit volume, m~ 3

ql = dipole moment of individual sphere, C m

From

(3-5-10) the permittivity

is

Introducing the value of P from

given by

(1) into (2),

(2)

we have
Nql
(3)

According to

(3),

the permittivity of the

artificial dielectric

can be determined

if

number of spheres per unit volume and the dipole moment of one sphere per
unit applied field are known. Proceeding now to find the dipole moment of one sphere
the

t If the spheres are

hollow (or

metal disks or strips are used), the artificial dielectric

corresponding slab of true dielectric.
a principal advantage of the artificial-dielectric material.

slab

may be much

This

is

if

lighter in weight than the

65

THE STATIC ELECTRIC FIELD: PART 2

SEC. 3-6]

()

(b)

FIGURE

3-4

(a) Individual
trie

it is

and

sphere of

assumed that each sphere

spheres, but this

is

E * -,9

artificial dielec-

(6) equivalent dipole.

is

in a

uniform

0-1

Axis of_
dipole

+q

This neglects the interaction of

field.

negligible provided the sphere radius

small

is

compared with the

spacing between spheres.

The

potential

at a point in a

V = -

f
Jo

uniform

given by

field is

= -Er cos

cos 6 dr

(4)

= radial distance from origin (taken at center of dipole)

= angle between field direction or axis of dipole and radial line (see Fig.
According to (4), V = at all points in a plane through the origin and normal
where

Equation

(4) gives the potential at

the potential

Vd

of a dipole in

air

a point in a uniform

is,

from

Assuming

field.

3-46)
to E.

that r

>

/,

(2-15-1),

ql cos 6
(5)

The

total potential

is,

by superposition, the sum of

V=V

+ Vd = -Er cos

The metal sphere has only induced charges on its

and negative charge), so that its potential is zero.

0= -EacosO +

and

or

(5)

'

a I cos

A
Ane

j
i

(6)

surface (equal

Thus, for

=-

47te

Solving for ql/E,

(4)

amounts of

a, (6)

positive

reduces to

(7)

flr

we have

| = 47t

(8)

66

ELECTROMAGNETICS

Introducing (8) into

[CHAP. 3

(3)

we

obtain e
er

where

4ne

Na 3

+ 4nNa

or
(9)

= permittivity of artificial dielectric

= relative permittivity of artificial dielectric
number of spheres per unit volume
= radius of sphere

e
er

N
a

Both the unit volume and the radius are expressed

Since the volume v of the sphere
er

where

is

4na 3 /3,
1

in the

same unit of length.

can also be written

(9)

+ 3vN

(10)

N = number of spheres per unit volume

volume

of sphere (in same units as unit volume)

Thus, the permittivity of the

artificial dielectric

depends on both the number of spheres

per unit volume and the size of the spheres.

3-7

BOUNDARY RELATIONS

In a single

medium

the electric field

is

continuous.

That

is,

the

field, if

changes only by an infinitesimal amount in an infinitesimal distance.

boundary between two

different

magnitude and direction.

It is

media the

not constant,

However,

at the

may change abruptly both in

in many problems to know the

electric field

of great importance

These boundary relations are discussed in

relations of the fields at such boundaries.

this section.
It is

convenient to analyze the boundary problem in two parts, considering

first

the relation between the fields tangent to the boundary and second the fields normal
to the boundary.

Taking up

media of
It is

first

the relation of the fields tangent to the boundary,

permittivities e x

and

e2

assumed that both media are perfect

of the two media are zero.

let

two

dielectric

be separated by a plane boundary as in Fig. 3-5.

insulators,

i.e.,

the conductivitiesf

ffi

and

<

Ax

"
Medium

i
'

Ay

FIGURE

E ",

e,

cr,

3-5

The tangential

electric field is

E, 2

continuous

across a boundary.

<7 2

Consider a rectangular path, half in each medium, of

For discussion of conductivity

Medium 2
e,

see Sec. 4-7.

67

THE STATIC ELECTRIC FIELD PART 2

sec. 3-7]

Pi

Medium

Ay

FIGURE

3-6

The normal component of the

flux dencontinuous across a charge-free

boundary.

sity is

length

Ax parallel to

average electric
average

the boundary

field intensity

field intensity

and of length Ay normal to the boundary.

tangent to the boundary in

tangent to the boundary in

medium

medium 2 be E t2

1
.

be

Let the

E tl

and the

The work per

unit charge required to transport a positive test charge around this closed path
line integral

of E around the path ($E

work along

zero the

though a

E around

By making

rfl).

the segments of the path normal to the boundary

finite electric field

may

exist

normal to the boundary.

the rectangle in the direction of the arrows

is

The

*(!

is

zero even

According to

(2) the tangential

electric field is continuous across

If

under

medium 2

is

(1)

(2)

components of the

electric field are the

same on

In other words, the tangential

such a boundary.

a conductor (o 2

static conditions

and hence

(2)

the field E, 2 in

0),

medium

2 must be zero

reduces to

(1 =0
According to

(3),

Fig. 3-6.

(3)

the tangential electric field at a dielectric-conductor boundary

Turning our attention next to the

dielectric

of

then

<2

both sides of a boundary between two dielectrics.

the

line integral

E n Ax E t2 Ax =
or

is

Ay approach

the path length

is

zero.f

normal to the boundary, consider two

separated by the xy plane as shown in

fields

media of permittivities e and e 2

It is assumed that both media are perfect insulators
t

(a,

a2

0).

Suppose

t This assumes that no currents are flowing. If currents are present, E in the conductor is not zero, unless the conductivity is infinite, and (2) applies rather than (3).

In Chap. 8 the relations of (2) and (3) are extended to include time-changing fields,
and it is shown that the relation ,, = Etl of (2) holds with static or changing fields
for the boundary between any two media of permittivities, permeabilities, and conFurthermore, for changing fields the relation
ductivities i, pi, <*i and e 2 /i 2 o 2
,

Ei=

of (3) is restricted to the case where the conductivity of medium 2 is infinite

oo).
This follows from the fact that a time-changing electric field in a con(<7 2
ductor is zero only if the conductivity is infinite.
t

68

ELECTROMAGNETICS

[CHAP. 3

that an imaginary box

Az.

Let

Dnl

medium, of area Ax Ay and height

constructed, half in each

is

be the average flux density normal to the top of the box in medium

and D 2 the average flux density normal to the bottom of the box in medium 2.
Dnl is an outward normal (positive), while Dn2 is an inward normal (negative). By
Gauss' law the electric flux or surface integral of the normal component of D over a
closed surface equals the charge enclosed.
By making the height of the box Az
approach zero the contribution of the sides to the surface integral is zero. The total
flux over the

box

is

then due entirely to flux over the top and bottom surfaces.

average surface charge density on the boundary

is

ps

If the

we have on applying Gauss'

law

Dnl Ax Ay -

D 2 Ax Ay

A,i

According to

(4) the

boundary between two

This

is

p,

Ax Ay

Dm = Ps

(4)

normal component of the flux density changes at a charged

dielectrics

by an amount equal

to the surface

charge density.

usually zero at a dielectric-dielectric boundary unless charge has been placed

there by mechanical means, as by rubbing.

If the

boundary

is

free

from charge, p s
A,,

According to

(5), the

and

(4)

reduces to

A,2

(5)

normal component of the flux density

is

continuous across the

charge-free boundary between two dielectrics.

If

medium 2

According to
boundary
It is

is

is

a conductor, D 2

the

(6),

and

A.i

(4)

reduces to
(6)

Ps

normal component of the flux density at a dielectric-conductor

equal to the surface charge density on the conductor.^

important to note that p s in these relations refers to actual electric charge

finite distances from equal quantities of opposite charge and not to

separated by

p sp due to polarization. The polarization surface charge is produced

by atomic dipoles having equal and opposite charges separated by what is assumed

surface charge

to be an infinitesimal distance.

It is

not permissible to separate the positive and

negative charges of such a dipole by a surface of integration, and hence the volume

must always contain an integral (whole) number of dipoles and, therefore, zero net
charge. Only when the positive and negative charges are separated by a macroscopic distance (as

on

trie

opposite surfaces of a conducting sheet) can

we

separate

8 it is pointed out that the relation D nl

D 2
D 2
P* of (4) and D,
hold with static or time-changing fields for any two media of permittivities,
permeabilities, and conductivities t u p t , a t and 2, ^2. "2-

t In

of

Chap.
(5)

THE STATIC ELECTRIC FIELD: PART 2

sec. 3-7]

69

them by a surface of integration. This emphasizes a fundamental difference between

the polarization, or so-called "bound" charge, on a dielectric surface and the true charge
on a conductor surface. In a similar way the boundary relation for polarization is

If

medium 2

is

free space,
/*.!

Equations

(6)

and

(8) are written

more

D-ft=p
The minus

(8)

-Ps

generally as

and

-P-n = p sp

sign in the polarization relation results

from the

polarization charge at a dielectric surface the polarization

the surface normal

EXAMPLE

is

(9)

is

fact that with positive

directed inward while

directed outward (see Sec. 2-5).

Boundary between two

dielectrics.

Let two isotropic dielectric media

and 2 be separated by a charge-free plane boundary as in Fig. 3-7. Let the permittivities
be <fi and e 2 and let the conductivities a, = a 2 = 0. Referring to Fig. 3-7, the problem is
to find the relation between the angles a, and a z of a static field line or flux tube which
,

traverses the boundary.

solution

For example, given

ii, to find

<x 2

Let

= magnitude of D in medium
D = magnitude of D in medium 2
Ei = magnitude of E in medium
E2 = magnitude of E in medium 2
Z)i

Normal

Medium 2

FIGURE

3-7

Boundary between two dielectric media

showing change in direction of field line.

e,

cr.

70

ELECTROMAGNETICS

[CHAP. 3

In an isotropic medium,

and

According to the boundary

have the same direction.

relations,

D al ^D

and

Etl =E,

and

Dn2 = D

(10)

Referring to Fig. 3-7,

D,

cos xi

>,

Etl = Ei

while
Substituting (11)

and

and

sin a,

(12) into (10)

and dividing the

Di cos oti
Ei sin a,
But Di

i, and

D2 =

e2

,__!_

tan a 2
,i

e2

= relative permittivity of medium

P2 ''= relative permittivity

= permittivity

of

= 4).

EXAMPLE
iri

Fig. 3-7

is

Then when a

a2

(12)

resulting equations yields

D 2 cos

a2

a2

sin

e,i

(13)

m
(14)
e r2

e r 2 60

is

30, the angle

air (,i

a2

in

Boundary between a conductor and a

Since

= E2 sin

of vacuum

a conductor.

solution

(11)

medium 2

Suppose, for example, that medium

sulfur (,2

<x 2

so that (13) becomes

tan

where

E, 2

cos

Find

1),

while

dielectric.

is

medium

is

a slab of

66.6.

Suppose that medium 2

ai.

medium 2

According to the boundary

medium 2

is

a conductor,

D = E2 =
2

under

static conditions.

relations,

Di=p

or

,,1

= Ps
1

and

,1

Therefore

ai

Etl =
-1
= tan ~' -=
=
tan
Ei

It

follows that a static electric field line or flux tube at a dielectric-conductor boundary

is

always perpendicular to the conductor surface (when no currents are present).

This fact

is

of fundamental importance in

field

mapping

(see Sec. 3-22).

3-8

71

THE STATIC ELECTRIC FIELD: PART 2

sec. 3-9]

TABLE OF BOUNDARY RELATIONS

Table 3-2 summarizes the boundary relations for

static fields

developed in the pre-

ceding section.

Table 3-2

BOUNDARY RELATIONS FOR STATIC ELECTRIC

FIELDSt

Field

component

Boundary

Tangential

E = En

0)

Any two media

Tangential

,i

(2)

Medium

Condition

relation

ti

is

a dielectric;

medium

is

conductor

Normal

A,i

-D.2

Norma!

D.l

D, 2

Normal

Di

= p.

= p,

(3)

Any two media

(4)

Any two media

(5)

Medium

with charge at boundary

with

no charge

at

1 is a dielectric; medium 2
conductor with surface charge

boundary
is

and (4) apply in the presence of currents and also for time-varying fields
The other relations, (2) and (5), also apply for time-changing situations provided

f Relations (1), (3),

(Chap. 8).
oa.
a1

3-9

PARALLEL-PLATE CAPACITOR

Consider the parallel-plate capacitor shown in Fig.

3-8.

From

Sec. 2-21

we have

eA

0)
I

= permittivity of medium
A = area of plates, m 2
/ = separation of plates, m

where e

between capacitor

FIGURE

3-8

Parallel-plate capacitor.

plates,

Fm"

72

ELECTROMAGNETICS

[CHAP. 3

Introducing the relative permittivity er and the value of e

we obtain

for the

capacitance of the parallel-plate capacitor

r

C = 8.85^where

er

is

the relative permittivity of the

EXAMPLE A

(pF)

medium between

parallel-plate capacitor consists of

a side and separated by 10

mm.

plate.

the plates.

two square metal

plates 500

mm

on

A slab of sulfur (e, = 4) 6 mm thick is placed on the lower

mm

plate, as indicated in Fig. 3-9.

and the upper

(2)

This leaves an air gap 4

thick between the sulfur slab
Find the capacitance of the capacitor. Neglect fringing of the field

at the edges of the capacitor.

solution

The foil

slab.

that

it

is

Imagine that a thin metal

is

of negligible thickness) the

.may now be

foil is

not connected to either plate.

series,

an

capacitance C , and a sulfur-filled capacitor of 6

suggested in Fig. 3-96.

The capacitance of the

C- =
The capacitance of the

8.85Ae,

8.85

-J-=

foil is

the capacitor

field in

regarded as two capacitors in

placed on the upper surface of the sulfur

Since the

is

normal to

E (and assuming

undisturbed.

air capacitor

The capacitor

mm plate spacing and

of 4

mm plate spacing and capacitance

air capacitor

0.5 2

0.004

is,

from

C, , as

(4),

=553pF

sulfur-filled capacitor is

8.85

0.5 2

x 4

= 475PF
1

0.006

'

The

total capacitance of two capacitors in parallel is the sum of the individual capacitances.
However, the total capacitance of two capacitors in series, as here, is the reciprocal of the
sum of the reciprocals of the individual capacitances. Thus, the total capacitance C is

given by

i__L + J_
C~ C C,
a

s4

mm

Ml
^6

FIGURE

.*

mm

L r
I

Air

~T"

(a)

"(b)

3-9

(a) Parallel-plate capacitor

capacitor.

with sulfur slab and air gap and

(6) equivalent series

THE STATIC ELECTRIC FIELD: PART 2

SEC. 3-10]

= 553 x 1,475
C= _C^
553 + 1,475
C. + C,

is

DIELECTRIC STRENGTH

3-10

The

73

field intensity

E in a dielectric cannot be

increased indefinitely.

field intensity that

dielectric

If a certain value

down.f The maximum

can sustain without breakdown is called its dielectric

exceeded, sparking occurs and the dielectric

is

said to break

strength.

In the design of capacitors

is

it

difference that can be applied before

this

important to

breakdown

know

occurs.

the

maximum

For a given

potential

plate spacing

breakdown is proportional to the dielectric strength of the medium between the

The radius of curvature of the edge of the capacitor plate is another factor,

plates.

since this curvature largely determines the

maximum

field intensity that

occurs for a

given potential difference (see Sec. 3-23).

Many
that

if

These types have the advantage

capacitors have air as the dielectric.

breakdown occurs, the capacitor

is

not permanently damaged.

For applica-

tions requiring large capacitance or small physical size or both, other dielectrics are

employed.
listed in

The

Table

dielectric strengths

3-3.

The

of a number of

dielectric strengths are for

common

a uniform

dielectric materials are

field,

and the materials

are arranged in order to increasing strength.

Table 3-3

Dielectric strength

Material

Air (atmospheric pressure)

Oil (mineral)

15

Polystyrene
Rubber (hard)

20

Bakelite

25
30
30
30

Paraffin

Quartz (fused)

Mica

As E

15

Paper (impregnated)

Glass (plate)

MVm-'

21

200

gradually increased, sparking occurs in air almost immediately when a

value of field is exceeded if the field is uniform (E everywhere parallel), but a
glowing, or corona, discharge may occur first if the field is nonuniform (diverging)
with sparkover following as E is increased further.
is

critical

74

ELECTROMAGNETICS

ENERGY IN A CAPACITOR

3-11
It

requires

To

[CHAP. 3

work to charge a

Hence energy is stored by a charged capacitor.

capacitor.

determine the magnitude of this energy, consider a capacitor of capacitance

charged to a potential difference

between the two conductors.

Then from

(2-21-1)

where q

is

the charge

infinitesimals

it is

on each conductor.
work

the infinitesimal

= CV

(1)

Potential

dW per

is

work per

charge.

infinitesimal charge dq.

In terms of

That

is,

dq
Introducing the value of

V from

(2) in (1)

we have

dW = ^dq
If the, charging process starts
is

delivered, the total

work

(3)

from a zero charge and continues

is

the integral of
1

(3),

This

is

the energy stored in the capacitor.

final

charge

or
2

w -rJ>
__ =
r

until

By

Q
57r
2 C
1

(2-21-1) this relation

(4)

can be variously

expressed as

Q
C

where

,1

(5)

W = energy, J
C = capacitance, F
V = potential difference, V
Q = charge on one conductor, C

3-12

ENERGY DENSITY

A STATIC ELECTRIC FIELD

IN

Consider the parallel-plate capacitor of capacitance C shown in Fig. 3-10.

charged to a potential difference V between the plates, the energy stored is
JV

= CV 2 =iQV

In what part of the capacitor is

The question may now be asked
in the electric field between
The
energy
is
stored
The
answer
is:
stored?
:

When

it is

(1)

the energy

the plates.


THE STATIC ELECTRIC FIELD: PART 2

sec. 3-12]

Area

J2
I

A
Thin metal
sheets

y*-&v

FIGURE

3-10

FIGURE

Energy

stored in the electric field between

Small cubical

is

75

3-11

volume of capacitance

A/.

the capacitor plates.

To demonstrate this, let us proceed as follows. Consider the small cubical volume
Av (=A/ 3 ) between the plates as indicated in Fig. 3-10. This volume is shown to a
larger scale in Fig. 3-11.

(of area A/

of metal
field will

now

2
)

foil

The length of each

side

is

and the top and bottom

A/,

are parallel to the capacitor plates (normal to the field E).

are placed coincident with the top and bottom faces of the volume, the

be undisturbed provided the sheets are sufficiently thin.

The volume Ay

constitutes a small capacitor of capacitance

= ,A/
AC=1^
A/
The

faces

If thin sheets

potential difference

AV

of the thin sheets

is

(2)

given by

AV = E A/
Now

the energy

(3)

AW stored in the volume Ay from

AW = iACAV
is,

(3-11-5),

Substituting (2) for

AC and

(3) for

AV

in (4),

AW = $eE

we have
2

Av

(5)

Dividing (5) by Av and taking the limit of the ratio

we obtain

the energy per volume, or energy density,

shrinks to zero.

(4)

A W/Av as Ar approaches zero,

w at the point around which Av

Thusf

w=

AW
lim
Ad->0

==

ie 2

(J

m"

AtT

For the more general case of a nonisotropic medium

in the same direction,

w=

iD.E

(7)

in

which

D and E may

not be

(6)

76 ELECTROMAGNETICS

[CHAP. 3

Now the total energy

W stored by the capacitor of Fig. 3-10 will be given by the

integral of the energy density

over the entire region in which the electric

field

has

a value;

W
where the integration
plicity

it is

is

\wdv = ~ \eE 2

dv

(8)

taken throughout the region between the

assumed that the

field is

fringing of the field at the edges of the capacitor.

Thus, evaluating

W = ieE Al = iDAEl = $QV

2

where
This

A = area of one capacitor plate, m

/ = spacing between capacitor plates,

no

(J)

(9)

obtained by integrating the energy density throughout the volume between

the plates of the capacitor,

/In

is

(8),

result,

3-13

For sim-

plates.

uniform between the plates and that there

is

identical with the relation given

FIELDS OF SIMPLE

many problems

it is

desirable to

For example,

by

(3-11-5).

CHARGE CONFIGURATIONS
know

the distribution of the electric field and the

breakdown value
medium, sparking, or corona, can occur. From a knowledge of the
field distribution, the charge surface density on conductors bounding the field and the
capacitance between them can also be determined.
In the following sections the field and potential distributions for a number of
simple geometric forms are discussed. The field and potential distributions around
point charges, charged spheres, line charges, and charged cylinders are considered
first.
The field and potential distributions of these configurations can be expressed
by relatively simple equations. The extension of these relations by the method of
images to situations involving large conducting sheets or ground planes is then
Finally, the field and potential distributions for some conductor conconsidered.
associated potential.

if

the field intensity exceeds the

for the dielectric

figurations which are not easily treated mathematically are

method known

as field mapping.

the use of computers

3-14
Field

Many

field

found by a simple graphical

problems are conveniently solved by

such methods are discussed in Chap.

7.

FIELD DISTRIBUTIONS
and potential distributions may be presented

potential

along a reference line

For example,
and of the electric

in various ways.

a graph of the variation of the magnitude of the electric

field

may give the desired information.

This

is

illustrated

THE STATIC ELECTRIC FIELD PART 2

sec. 3-14]

77

4irrj

V
-

FIGURE

3-12

Variation of electric field

sphere of radius r t

E and

potential

of an isolated charged conducting

by the curves for E and V in Fig. 3-1 2a for the field and potential along a radial line
extending from the center of a charged conducting sphere of radius r,. Or the field

and

potential distributions

3-126.

In this

map

may

be indicated by a contour map, or graph, as in Fig.

the radial lines indicate the direction of the electric

the circular contours are equipotential lines.

between contours

is

a constant.

field,

while

In this diagram the potential difference

78

[CHAP. 3

ELECTROMAGNETICS

dz

-Line of charge

FIGURE
Thin

line

3-13

of charge of length la.

FIELD OF A FINITE LINE OF CHARGE

3-15

now

the field produced by a thin line of electric charge. Let a positive

uniformly as an infinitesimally thin line of length 2a with
distributed
be
Q
center at the origin, as in Fig. 3-13. The linear charge density p L (charge per unit

Consider
charge

length)

is

then

Pt

(1)

coulombs per meter when Q is in coulombs and a is in meters.

P on the r axis, the infinitesimal electric field dE due to an infinitesimal length of wire dz is the same as from a point charge of magnitude p L dz.

where p L

is

in

At the point

Thus,

dE =

Anel

where

I
I

=Jr 2 + z 2
= unit

vector in direction of

(2)

sec. 3-16]

THE STATIC ELECTRIC FIELD: PART 2

Since the z axis in Fig. 3-13

is

an

axis of

symmetry, the

field

has only z and

com-

These are

ponents.

and

dEr = dE cos

= dE

(3)

dEz = dE sin

= dE

(4)

The resultant or total r component Er of the field at a point on

by integrating (3) over the entire line of charge. That is,
pr C(*

Hence the

PjJ_

the r axis

is

obtained

dz

+ z 2y
VFTz

4neJ- a Jt r 2

2 3
)

Pifl

Inerjr 1

By symmetry the

4neL a

E.=

and

(5)

a2

component Ez of the field at a point on the r axis

E at points along the r axis is radial and is given by

resultant z

total field

2neryjr

This relation gives the

field

is

zero.

Pifl

E =

(6)

as a function of r at points on the r axis for a finite line of

charge of length 2a and uniform charge density p L

3-16

79

FIELD OF AN INFINITE LINE OF CHARGE

Consider that the line of charge in Fig. 3-13 extends to infinity in both positive and
negative z directions.

and

By

dividing the numerator

a become infinite the

found to be

letting

charge

is

and denominator of (3-15-6) by a

due to an infinite line of positive

electric field intensity

(1)

2ner

The potential difference

from the infinite line of charge
a positive

test

V21
is

charge from r 2 to

between two points at radial distances

r2

and

r,

then the work per unit charge required to transport

r,.

Assume

that r 2

>

rt

This potential difference

80

is

[CHAP. 3

ELECTROMAGNETICS

given by the line integral of

r 2 if the line

of charge

is

Er from r2 to r u the potential at r

being higher than at

Thus

positive.

^ L^'-tlT
or

(2)t

21

3-17
If the

2ne

INFINITE CYLINDER OF
charge

is

rr

2;r

CHARGE

distributed uniformly along a cylinder of radius r x instead of concen-

trated along an infinitesimally thin line, the field external to the cylinder
(3-16-1) for

The

r>r

Inside the cylinder,

>

r,

Inside the cylinder the potential

3-18

and
is

given by

E = 0.

potential difference between the cylinder

given by (3-16-2), where r 2

is

and points outside the cylinder

is

p L is the charge per unit length of the cylinder.

the

same as the

potential at the surface (r

r t ).

INFINITE COAXIAL TRANSMISSION LINE

coaxial transmission line consists of

by the cross section of Fig.

much can be learned about
static conditions.

3- 14a.

two conductors arranged

This

is

common

coaxially, as

type of transmission

shown
and

line,

its properties from a consideration of its behavior under

Let a fixed potential difference be applied between the inner and

FIGURE

3-14

Coaxial transmission

=n

line.

becomes
we put r=r then V=0a.tr n. That is, the surface for which r
a reference surface with respect to which the potential values are measured.

t If

THE STATIC ELECTRIC FIELD: PART 2

SEC. 3-19]

outer conductors of an infinitely long coaxial line so that thexharge

of one

The

line is

pL

The

field is

81

Q per unit length /

confined to the space between the two conductors.

and the equipotential lines are concentric circles, as indicated

The magnitude of the field at a radius r is given by (3-16-1), where
a<,r <b, and where p L is the charge per unit length on the inner conductor. The
field lines are radial,

in Fig. 3-146.

potential difference

between the conductors

is,

from

(3-16-2),

K = ^-ln*
2ne
Since capacitance

by length
(C

m -1

/,

).

we have

is

Cjl

given by the ratio of charge to potential,

The

(Q/l)/V.

is

C=

pL

the permittivity of the

Since e

e er

where

Q/V.

Dividing

ratio Q/l equals the linear charge density

Hence, the capacitance per unit length C/l of the coaxial

i=T
where e

(1)

2ne

(Fm

wm

>

medium between conductors.

pF m -1 (2) can be expressed more

8.85

pL

line is

(2)

conveniently

as

T
where

er

3-19

=
=
=

55.6cr

h^)

relative permittivity

inside radius

24.2e r

(3)

of medium between conductors

of outer conductor

radius of inner conductor (in

TWO

_ 1s
(pFm
}

iogTw

same

INFINITE LINES OF

units as b)

CHARGE

Let two infinite parallel lines of charge be separated by a distance 2s as in Fig. 3-15.

Assume that the linear charge density of the two lines is equal but of opposite sign.
The resultant electric field E at a point P, distant rt from the negative line and' r 2
from the

positive line, is then the vector sum of the field of each line taken alone.
Let the origin of the coordinates in Fig. 3-15 be the reference for potential.

Imagine that only the positively charged

tial

difference between

P and

the origin

r+

line is present.

Then, from (3-16-2) the poten-

is

=^ln*
2ne
r

(1)

82

[CHAP. 3

ELECTROMAGNETICS

v=o

E^
JrPfx V)

\\r

^^ "n

Origin

-^>**'

>

Y+v

^^*.

Equipotential

FIGURE
Two

3-15

infinite lines

of charge separated by a distance

2s.

Similarly for the negatively charged line

s

Pi
= - In ,

V_

2ne

With both
algebraic

lines present the total potential difference

sum of (1) and

(2),

in (3)

is

a constant,

(3) is

K_

origin

is

the

(3)

r2

the equation of an equipotential line.

if (3) is

The form

transformed in the following

From (3)
,

In

= 2neV
/"!

r2

-1

and
Since 2neV/p L

K.

In

2ne

of the equipotential line will be more apparent

manner.

V between P and the

or

V = V+ +
If

(2)

ry

is

(4)

Pl

= # y ln-

a constant for any equipotential

(5)

line,

the right side of (5)

is

a constant

Thus
e

2n<V/ PI.

=K

and

^ = Kf2

(6)

SEC. 3-20]

The coordinates of the point P in Fig.

r2

=y/(s

x)

+y

2
.

THE STATIC ELECTRIC FIELD: PART 2

83

and

3-15 are (x, y), so that r t =s/(s

Substituting these values of r x

and

x)

r 2 in (6), squaring,

and

re-

arranging yields

x2
Adding j 2 (A 2
side, we have

i)

/(K

l)

K 2 + 1\ 2

+y

(*-xrn)
the equation of a circle having the

is

(x

where

x,

y
A
r

(8)

(*, y) is

its

2Ks

and

(9), it

w.

(9)
\

follows that the equipotential curve passing through the

on the x

axis at

IKs

pi

do)

a distance from the origin

2

in Fig. 3-16.

is

shown

in Fig. 3-15.

approaches zero and h

so that the equipotentials are smaller circles with their centers

nearly at the line of charge.

the plane

equipotential line of radius r with center at (h, 0)

s,

left

-fern)

increases, corresponding to larger equipotentials, r

approaches

The potential

x =

is

This
is

more

by the additional equipotential circles

zero along the y axis. That is, V at x = 0.
Thus,
is

illustrated

the reference plane for potential.

Field lines are also

potential circles

and

shown

in Fig. 3-16.

These are everywhere orthogonal to the

also are circles with their centers

on the y

axis.

INFINITE TWO-WIRE TRANSMISSION LINE

3-20

The

on the

a circle of radius

center

An

(7)

form

a:

As

+ y2 =

- h) 2 +y 2 =r 2

'

with

= coordinates of point on circle

= x coordinate of center of circle
= radius of circle

Comparing
point

to both sides of (7) to complete the square

This

2xs v2

discussion of two infinite lines of charge in the previous section

is

easily extended

to the case of an infinite line consisting of two parallel cylindrical conductors or wires.

This

is

a type of transmission line

commonly used

in practice,

and much can be learned

84 ELECTROMAGNETICS

[CHAP. 3

v=o

FIGURE
Field

an

about

its

properties

3-16

and equipotential

infinite

lines

around two

two-conductor transmission

infinite parallel lines

of charge or around

line.

from a consideration of its behavior under

static conditions.

Let

a fixed potential difference be applied between the conductors so that the charge per
unit length of each conductor

The

surface of the wire

pL
an equipotential

is

is

circle in Fig. 3-16 will coincide

surface,

and therefore an equipotential

with the wire surface.

Thus, the heavy

The

radius r and center-to-center spacing 2h can represent the two wires.

potential distributions external to the wire surfaces are the

produced by two infinitesimally thin

inside the wires

The charge

is

is,

if

field

of

and

the field were

The field
on the surface.
on the wire surface but has higher density on

lines

of charge with a spacing of

of course, zero, and the potential

not uniformly distributed

same as

circles

is

2s.

the same as

the adjacent sides of the conductors.

The

potential difference

between them

is,

from

(3-19-3)

Vc

between one of the conductors and a point midway

and

(3-19-6),

Vc =
The value of K can be expressed

in

2ne

In

(1)

terms of the radius r and half the center-to-center

spacing h by eliminating s from (3-19-10) and (3-19-11) and solving for K, obtaining

THE STATIC ELECTRIC HELD: PART 2

SEC. 3-21]

*-*+.The

V2c

potential difference

To

then

on one conductor to the

we take the

difference of potential

i.e.,

7 " VZ ~
er

is

find the capacitance per unit length, C/l, of the two-conductor line

between the conductors,

where

between the two conductors

ratio of the charge per unit length

85

log[

relative permittivity

+ V(A/r)*-l]

r)

of the

(PF

medium surrounding

m~

(4)

the conductors, dimen-

sionless

=
=

h
r

conductor radius (same units as h)

INFINITE SINGLE- WIRE TRANSMISSION LINE

3-21

half center-to-center-spacing

single-wire transmission line with

ground return

is

another form sometimes used.

Let the conductor radius be r and the height of the center of the conductor above

ground be
that the

Assume

h.

ground

The

field

is

and

theory of images.^

that the conductor has a positive charge

potential distribution of this type of line

Thus,

if

the ground

is readily found by the

removed and an identical conductor with
below ground level as the other conductor

is

p L per unit length placed as far

above, the situation is the same as for a two-conductor line (Fig. 3-16).

charge
is

conductor which replaces the ground

The
by

field

and

is

The

called the image of the upper conductor.

potential distribution for the single conductor line

is

then as illustrated

Fig. 3-17.

The

and the ground is as

The capacitance per unit length C/l

difference in potential between the single conductor

given by (3-20-1) or by one-half of (3-20-3).

is

p L per unit length and

at zero potential.

twice the value given by (3-20-4), or

f
t

=
log[

r)

The theory of images

^W-l]

is

discussed

more

(pFm_1)

fully in Sec. 7-15.

0)

[CHAP. 3

ELECTROMAGNETICS

86

x +-

--

^--1-*-

\
i

\
i

FIGURE

3-17

Conductor over a metallic ground plane.

GRAPHICAL MAPPING OF STATIC

3-22

ELECTRIC FIELDS; FIELD CELLS

Not

all

above.

conductor configurations can be treated mathematically as readily as those

Other methods are discussed in Chap. 7. However, for two-dimensional

problemst a very

effective graphical-field-mappingt

experience in graphical field mapping

is

valuable because

such important fundamental properties of

1

2
3
4

method
it is

is

applicable.

Some

based on and emphasizes

static electric fields as the following:

and potential lines intersect at right angles.

The surface of a conductor is an equipotential surface.
The field meets a conducting surface normally.
In a uniform field, the potential varies linearly with distance.
Field

By a two-dimensional problem is meant one in which the conductor configuration

can be shown by a single cross section, all cross sections parallel to it being the same.
A. D. Moore, " Fundamentals of Electrical Design," McGraw-Hill Book Company,

New

York, 1927.

Except at singular points, as at point

P in

Fig. 2-14.

THE STATIC ELECTRIC FIELD PART 2

sec. 3-22]

Conductor

87

40 V

30 V

20 V

10

V
Conductor 2

FIGURE

3-18

Cross section of two sheet conductors with partially completed

flux tube

is

and the

parallel to the field,f

field

electric flux is constant

map.

over any

cross section of a flux tube.

tube of flux originates on a positive charge and ends on an equal negative

charge.

Graphical field mapping will be introduced with the aid of an example. Consider
two charged sheet conductors 1 and 2 as shown in cross section in Fig. 3-18. The
sheets extend infinitely far to the

two-dimensional problem,

all

left

and

right

and also normal to the page.

This

is

cross sections parallel to the page being the same.

Therefore, the field and potential distribution everywhere between the sheets will be

known if it can be found for a two-dimensional cross section such as shown in Fig.

3-18.

Let the potential difference between the conductors be 40 V, with the upper conductor
positive

and the lower conductor

of b the

field is

To

at zero potential.

the

uniform, so that equipotential surfaces 10

left

of a and to the right

V apart are equally

spaced

and 40 V. Between a
and b the conductor spacing changes, and the equipotentials may be drawn tentatively
as shown by the dashed lines.
The next step in the mapping procedure is to draw field lines from conductoi 1
as indicated, the conductor surfaces being equipotentials at

to conductor 2 in the uniform field region to the

between the equipotentials.

square

is

A stack,

In this

way

left

of a with the spacing equal to that

the region

the end surface of a rectangular volume, or

is

divided into squares.

cell,

Each

of depth d into the page.

or series, of squares bounded by the same field lines represents the sidewall
t

The

sidewalls of a flux tube are field lines.

88

ELECTROMAGNETICS

[CHAP. 3

T
Flux tube with

cells in series

Remainder
Flux tube
with 9 Vi
cells in series

FIGURE

3-19

Cross section of two sheet conductors with completed

three-dimensional view of a field cell.

field

map.

Inset

shows

of a rectangular flux tube extending between the positive charge on one conductor to
the negative charge

on the

other.

drawing field lines as nearly normal

The

field

map

is

next extended to the right by

to the equipotentials as possible, with the field lines

spaced so that the areas formed are as nearly square as possible.

revisions of the tentative equipotentials between a

and b and

should be possible to remap the region to the right of a so that

lines are

everywhere orthogonal and the areas between the

After one or two

also of the field lines,

field

it

and equipotential

lines are all squares or

The completed field map is shown in Fig. 3-19.

By a curvilinear square is meant an area that tends to yield true squares as it is
subdivided into smaller and smaller areas by successive halving of the equipotential
interval and the flux per tube.
A partially subdivided curvilinear square is illustrated
curvilinear squares.

in Fig. 3-20.

A field map, such as shown in Fig. 3-19, divides the field into many squares each
of which represents a side of a field cell. These field cells have a depth d (into the
page) as suggested by the three-dimensional view of the typical field cell in Fig. 3-19.
The
cell

cell

has a length

(parallel to the field)

and a width

are the walls of a flux tube (parallel to the

field),

b.

The

sidewalls of a field

while the top and bottom coincide

89

THE STATIC ELECTRIC FIELD: PART 2

sec. 3-22]

Equipotential
lines

FIGURE

3-20

Partially subdivided curvilinear square.

As

with equipotential surfaces.

their

curvilinear cells are subdivided into smaller cells,

depth d (into the page), the same as the larger

may

The subdivided

end areas tend to become true squares.

cells

are always of

Thus, afield cell, or simply a

cell,

are applied to the equipotential surfaces of a field

cell,

cells.

be defined as a curvilinear square volume.

If thin sheets

of metal

we have a field-cell

foil

capacitor.

The capacitance

C of a

parallel-plate capacitor

is

"

(1)

e permittivity of medium
A = area of plates
/ = spacing of plates

where

Applying

this relation to

the capacitance

of the

field-cell

capacitor with a square end (b

C =
Dividing by d,

we

e is

/),

ed

obtain the capacitance per unit depth of a field

Co

where

we have

for

field cell

the permittivity of the

medium

in farads per meter.

(2)

cell

as

(3)

90

ELECTROMAGNETICS

[CHAP. 3

Conducting
sheets

FIGURE

3-21

field-cell

capacitor has a capacitance

per unit depth equal to the permittivity

e of the

For

medium.

air (e

capacitance of the capacitor

=e

the

shown

is

8.85 pF.

Thus, the significance of the value of e

depth of a

field-cell capacitor.

medium of air

(or

For example, a

is

that

the capacitance per unit

it is

field-cell

capacitor of unit depth in a

vacuum) has a capacitance of 8.85 pF.

Such a capacitor is

illustrat-

ed in Fig. 3-21.

Any

field cell

can be subdivided into smaller square-ended

cells in parallel as in series.

large or small, exactly square or curvilinear

In a
cells

field

map, such

cells

with as

Hence the capacitance! per unit depth of any

as in Fig. 3-19,

is

equal to

e.

most of the area

with four in series for each flux tube.

These

many

field cell,

is

cells all

divided into "regular"

have the same potential

them (io V). Hence these cells may be defined as cells of the same
The remaining area of the map consists of a fractional, or remainder, flux tube.
This tube is also divided into cells. These cells are of two kinds, both different from
those in the rest of the map. One kind of cell in the remainder flux tube has about
4.3 V across it and the other kind about 1 V across it.
There are nine 4.3-V cells and
difference across

kind.

four 1-V

cells.

Any

field

properties are

cell

has the same capacitance per unit depth.

common

to field cells of the

same

have the same potential differences across them.

ends of those

cells are

kind.

These

cells

Many

additional

of the same kind

In uniform fields the areas of the

the same, but in nonuniform fields the areas will not be the

same.
t

The capacitance of an
greater than 8.85

isolated capacitor such as

shown

pF because of fringing of the field.

only a portion of a more extensive

field,

and

its

in Fig. 3-21 is

However, a

somewhat

field cell represents

sides are parallel to the field (no

fringing).

understood that this capacitance is that which would be obtained if the field
made into a field-cell capacitor by placing thin sheets of metal foil coincident
with its equipotential surfaces (if no conductor is already present).

} It is

cell were

E STATIC ELECTRIC HELD:

SEC. 3-22]

Since the capacitance per unit depth of any

it

of the same kind

cell

same kind

follows that the electric flux through any cell of the

91

the same,

the same (Q/d

Thus, the 10-V cells in Fig. 3-19 have a flux of lOe C/unit depth, while the

VC/d).
4.3-

is

is

PART 2

and 1-V

cells

have 4.3e and e C/unit depth, respectively.

Now the average flux density D at the equipotential surface of a field cell is given
by
Z>

= g=p s

(Cm" 2 )

ba

where

Q = total

charge on

through
b

equipotential surface of field cell

= total

flux

ifr

cell,

width of

foil at

(4)

cell,

d = depth of cell,

m
m

Cm"

equipotential surface,
p 4 = average surface charge density on foil at
Hence, the average flux density is inversely proportional to the field-cell or flux-tube
width. Also the average surface charge density p s at a conducting surface is inversely
.

proportional to the width of the

the spacing of conductors

of a; so in the uniform

field cell

and 2 to the

field

or flux tube at the surface.

right of b in Fig. 3-19

field

For example,

4 times that to the

left

of a the surface charge density p s is 4

region to the right of b. The surface charge

region to the

times the value of p s in the uniform

is

left

even smaller than to the right of b in the region of concave conductor

density
near
b and somewhat larger than to the left of a in the region of convex
curvature
is

conductor curvature near

or

Since E = D/e,
length (E = V/l).

same kind

is

a.

proportional to the

cell

width,

W (=$QV) stored in any

cell

of the

the field intensity is also inversely

Furthermore, the energy

the same.

It

follows that the average energy density

proportional to the area of the end of the

example, the energy density in the

uniform

cell

field

(=bl

for a square-ended

region to the

left

is

inversely

cell).

For

of a in Fig. 3-19

is

16 times the energy density in the uniform field region to the right of b.
To summarize, the properties of an accurate electric field mapt ore as follows:
1

4
5

The capacitance of any field cell is the same.

The capacitance C per unit depth of any field cell is equal to the permittivity
e of the medium.
The potential difference across any field cell of the same kind is the same.
through any field cell of the same kind is the same.
The flux
over any cross section of a flux tube is the same.
The flux
The average flux density D in any cell of the same kind is inversely proportional
i//

\ji

to the width of the cell or flux tube.

t In a single

medium of uniform

permittivity.

92

ELECTROMAGNETICS

[CHAP. 3

The average charge density ps

kind

is

The average

at the conducting

boundary of any

of the same

cell

inversely proportional to the width of the cell at the boundary.

field intensity

E in any cell of the same kind is inversely proportional

to the cell width.

The energy stored in any cell of the same kind is the same.
The average energy density w in any cell of the same kind is inversely proportional
to the area of the end of the cell.
(This is the area that appears in the field map.)

9
10

In order to test the accuracy of a field map, and hence the accuracy with which
the above properties hold for a particular

map, the curvilinear squares of the map can

be further subdivided by halving the equipotential interval and halving the flux per

tube, as in Fig. 3-20.

true squares, the field

rectangles, the

If the smaller regions so

is

accurately

mapped.

produced tend to become more nearly

However,

equipotential lines should intersect orthogonally.

rule be observed at all stages of

making a

field

is

Often

digital

equivalent to graphical field

iteration is continued until

graphical field

EXAMPLE
end

it is

Also

become
and

field

important that this

better to erase

and

In graphical field mapping an

as important as a pencil.

With analog and

be 40

the regions tend to

It is especially

map.

begin again than to try to revise an inaccurate map.

eraser

if

map is inaccurate and another attempt should be made.

mapping

is

computers and methods discussed in Chap.

mapping

is

With the

carried out.

no further change occurs

in the

7,

digital

map.

a process

computer,

Experience in

a valuable prelude to computer mapping.

mm

Referring to Fig. 3-19,

let the conductor separation at ff be 10

and at gg
have a depth (into the page) of 200 mm. If the conductors
gg, and if fringing of the field is neglected, find the capacitance C of the result-

mm, and let the conductors

at

#and

ing capacitor.

solution

The medium

in the capacitor is air.

The method of solution

of capacitors formed by the individual

Each

cell

will

be to evaluate the

combination

series-parallel

cells.

has a capacitance

Co

The capacitance between

=e

rf= 8.85 x 0.2

1.77

pF

the ends of each flux tube with 4 cells in series

1.77

-j-

then

= 0.443 pF

The capacitance between the ends of the remainder

1.77

is

^5 = 0.191

flux tube with 9.25 cells in series

PF

is

93

THE STATIC ELECTRIC HELD: PART 2

SEC. 3-23]

There are fifteen 4-cell tubes and one remainder

(9.25-cell) tube.

Hence the total capacitance

C between ff and gg is the sum of the capacitances of all the flux tubes,

or

C= (15 X 0.443) + 0.191 = 6.83 pF

The above

calculation

capacitance of unity.

On

simplified if each cell

somewhat

is

T + 935
and the
of a

total actual capacitance

cell,

arbitrarily assigned a

is

is

given by

86units

the product of this result

and the actual capacitance

or

C = 3.86

(8.85

Yet another method of calculation

is

is

this basis the total capacitance in arbitrary units

is

0.2)

= 6.83 pF

to use the relation that the total capacitance

given by

C = -C

(5)

where

N = number of cells (or flux tubes) in parallel

n = number of cells in series

Co = capacitance of one cell
and where all cells are of the same
of the 10-V

cells,

x
C= 1^
4

Note that

if

we had wanted

coinciding with the equipotentials

in series

Thus

kind.

in the

above example, counting

in terms

we have
(8.85

0.2)

6.83

pF

the capacitance of a capacitor with conductors

m and n (Fig. 3-19) and of 200 mm

would be reduced to two and the capacitance doubled.

depth, the cells

In this

way

the

capacitance of any conductor configuration conforming to the equipotential surfaces

of a

field

3-23

map can

be easily calculated.

DIVERGENCE OF THE FLUX DENSITY D

In Sec. 2-17 Gauss' law

that the normal

is

applied to surfaces enclosing finite volumes, and

component of the

flux density

equals the electric charge enclosed.

enclosing infinitesimal volumes

Let Ay be a small but

we

finite

throughout the volume, the charge

By an

it is

shown

integrated over a closed surface

extension of this relation to surfaces

are led to a useful relation called divergence.

volume.

AQ

If

we assume a uniform charge

enclosed

is

density

the product of the volume charge

94 ELECTROMAGNETICS

density p

[CHAP. 3

and the volume Av.

integral of the

By Gauss' law

normal component D of the

the charge enclosed

is

also equal to the

flux density over the surface of the

volume

Thus,

Av.

D ds
s

and

= AQ

p Av

(1)

A, ds
(2)

If the charge density

is

not uniform throughout Av,

we may

Av shrinks to zero, obtaining the charge density p

The limit of (2) as Av approaches zero
collapses.
written div D or V D.
Hence

take the limit of (2) as

at the point
is

around which Av

called the divergence of

D,

j.D.ds
lim
A.,-0

Whereas the

Av

integral of the

divD = p

If the

charge

is

zero at a point,

of D

is

follows that the charge density

D is zero at that point.

now

is

zero and also that

important to note that the divergence

It is

expression.

more formal way, developing it as a differenvolume Ax Ay Az = Av is placed in an electric field with flux

discuss divergence in a

A small

density D, having

shown

D over a finite volume yields

D gives the free charge density at a point.

a scalar point function.

Let us
tial

it

13)

normal component of

the free charge enclosed, the divergence of

the divergence of

(Cm -3 )

components

in Fig. 3-22.

The

D x Dy
,

and

total flux density

D = ZD X +

in the three coordinate directions as

D is related to its components by

yZ>

+ %D Z

Z^^s^^t^
FIGURE

3-22

Construction used to develop differential

expression for divergence of D.

(4)

95

THE STATIC ELECTRIC HELD: PART 2

SEC. 3-23]

The normal outward component of D at the back face is Dx since the field is directed
inward. If the field changes between the back and front faces, the normal component
of D at the front face can, by Taylor's theorem, be represented by an

dDx Ax

D * + -8
When Ax

is

Dx Ax 2

-T is^r

Dx Ax 3
+
-8si
d

(5)

may

be neglected, so that

very small, the square and higher-order terms

at the front face

we have

infinite series,

the normal component of D

for

Dx + -^Ax

(6)

ox

In like

manner the normal component of D

side face

at the left side face

is

Dy and at the right

is

Dy + -^Ay

(7)

dy

Similarly at the bottom face

Dz

it is

and at the top face

D +
flux of

and over the front

D over the

face

(8)

-Dx AyAz

(9)

flux of

Dx+

iit

Ax AyAz

(10)

D over the entire volume,

we obtain

for the total flux

= ( -Dx + D X + ^ Ax jA.v Az

+ (-Dy

+D +^Ay\AxAz + (-D +D +^Az\AxAy

z

(11)

simplifies to

A1
Ai^

/8DX

Gauss' law

(or integral of the

we know

8Dy

dDz \

+ ^ + ^)Ax Ay Az
[^
oz
ox
dy
\

From

is

is

Adding up the outward

which

-d7

back face

Aij/

Az

'

The outward

is

(12)

that the total electric flux over the surface of the

normal component of D over the surface of the volume)

is

volume
equal to

96 ELECTROMAGNETICS

the charge enclosed.

[CHAP. 3

The charge enclosed

density p over the volume.

Dividing by Av and taking the limit as

ofD.

Av approaches

im

and

D ds = SD dD1 dDI
+
+

f'

Av

divD

dx

flD,
=

dx
center

member

of (15)

is

dy

dy

= p

--

dz

(15)

differential relation for the divergence

The divergence of
or dot product of the operator V and D. That

divD =
is

(14)

dz

SD
8D
+ * + Z

in rectangular coordinates.

This

we obtain the divergence

zero,

Thus,

At>^o

The

also equal to the integral of the charge

is

Therefore

D=

expressed

is,

V-D

(16)

readily seen by expanding (16) into the expressions for

of

can also be written as the scalar,

+ 4)(4 + ^
dy

V and

D.

,(*>.+**>,+">.)

Then

d7)

Performing the multiplication indicated in (17) gives nine dot-product terms as follows

ox

dy

dz

dz

dx

dx

dy
(18)

The dot product of a

unit vector

another vector at right angles

is

on

itself is unity,

zero.

(18) vanish, but the three involving

&

dy

dz

but the dot product of a vector with

Accordingly, six of the nine dot products in

it, y
y, and t 1 do not, and the product in

dicated by (17) becomes

_ _

dDx

dDy

dDz

dx

dy

dz

(19)

The dot product of the operator V with a vector function is the divergence of the
The quantity V-may be considered as a divergence operator. Thus the

vector.

97

THE STATIC ELECTRIC FIELD: PART 2

SEC. 3-24]

divergence operator applied to a vector function yields a scalar function

For example,

D (divergence of D) is given in rectangular coordinates by the right side of (19) and

is

scalar,

If

being equal to the charge density p.

is

known

everywhere, taking the divergence of

enables us to find the

sources (positive charge regions) and sinks (negative charge regions) responsible for the
electric flux,

and, hence, for D.

MAXWELL'S DIVERGENCE EQUATION

3-24

The .relation

(3-23-15) that

V-D=p

(1)

was developed by an application of Gauss' law

fundamental
set

to

an

infinitesimal volume.

This relation

differential relation for static electric fields.

of four differential relations

known

are developed in later chapters.

In a region free from charge, p

as Maxwell's equations.

EXAMPLE

As a simple

is filled

nonelectrical

with air under pressure.

quickly, the air rushes out.

It is

It is

the

one of a

The other

three

= 0, and

V-D=0

cylinder

is

(2)

example of divergence consider that a long hollow

If the cover over one end of the cylinder is removed

apparent that the velocity of the air will be greatest near

the open end of the cylinder, as suggested by the arrows representing the velocity vector v

Suppose that the flow of

in Fig. 3-23d.

component.

air is free

directly proportional to x, as indicated

K is

is

independent of y but

is

by
|v|

where

from turbulence, so that v has only an x

Let us assume that the velocity v in the cylinder

a constant of proportionality.

=vx = Kx
The

(3)

question

is:

What

is

the divergence of v

in the cylinder?

solution

Apply the divergence operator to

(3),

giving

V.T-g-JT
Hence, the divergence of v

is

(4)

equal to the constant K.

velocity field can be represented graphically

by

with the density of the lines indicating the magnitude of v.

lines

The

showing the direction of v

velocity field in the cylinder

98

ELECTROMAGNETICS

[CHAP. 3

>
(a)

(b)

FIGURE

3-23

The velocity v of air rushing from a tube

has divergence (a) and

(b).

When

air

flows with uniform velocity through a

tube open at both ends, as at

divergence of v

represented in this
their source,

is

way

the

(c),

(c)

zero.

is

illustrated in Fig. 3-236.

v increases as a function of x.

This situation

v lines originate,

(3) to

is

concisely expressed

x.

i.e.,

have

This indicates that

by div v = K. That is,

and this tells us that,

K throughout the cylinder,

the divergence of v has a constant value

assuming

We note that

throughout the cylinder, the number increasing with

be correct, the source of the velocity

field

provided by the expanding air

is

uniformly distributed throughout the cylinder.

on the other hand, both ends of the cylinder are open and air passes through with
velocity everywhere, v x equals a constant and the divergence of v is zero in the
cylinder.
In this case, the source of the velocity field must be somewhere external to the
cylinder, and the velocity field diagram would be as shown in Fig. 3-23c.
If more lines enter a small volumef than leave it or more leave it than enter, the field
has divergence. If the same number enter as leave the volume, the field has a zero divergence.
If,

the

same

3-25

DIVERGENCE THEOREM

From Gauss' law

(2-17-6)

we have
<j>

where

ds

<j>

p do

(1)

D is integrated over the surface s and p is integrated throughout the volume

enclosed by s.
t In the limit

an infinitesimal volume.

THE STATIC ELECTRIC FIELD: PART 2

SBC. 3-26]

From

(3-24-1) let us introduce

D for p in (1),

T>-ds

The

relation stated in (2)

is

i>

99

obtaining

\Ddv

(2)

the divergence theorem as applied to the flux density D,

or Gauss' theorem (as distinguished from Gauss' law).

This relation holds not

but also for any vector function.

In words, the divergence

only for D, as in

theorem

(2),

states that the integral

of the normal component of a vector function over a

closed surface s equals the integral of the divergence of that vector throughout the volume
v enclosed

3-26

As

by the surface

s.

DIVERGENCE OF D AND P IN A CAPACITOR

further illustrations of the significance of divergence

parallel-plate capacitor of Fig. 3-24.

plates except for the small air gaps.

the surface of the plates.

slab of paraffin

let

us consider the charged

fills

the space between the

True charge of surface density p s

Polarization charge of surface density p sp

is
is

present

on

present on

the surface of the paraffin.

According to

(3)

of Table

Dnl
where

of

3-2, the relation

-D =p
n2

at a

boundary

is

given by
(1)

Dnl = flux density in air gap

Dn2 = flux density in conducting plate =

p s = true surface charge density

Suppose that the true charge p s is distributed uniformly throughout a thin layer
of thickness Ax, as suggested in Fig. 3-24. Then the total change AD in flux density
from one

side of the layer to the other

.i

But when

Ax

is

is

given by

A.2

A>

(2)

small,

AD X

dDx

Ax

(3)

(4)

dx

Therefore, (1) becomes

dDx -l'
dx

Ax

100

ELECTROMAGNETICS

[CHAP. 3

Capacitor plates-Air

gaps-

gSiK *
rtlf."

axis

*-'*

sR?.'.
i'
$.":>

mi
g||8
J

iV

It

Charge layers*[-i!i
ri-

1-1

ll
II

-MV-D-/0

FIGURE
Cross

3-24

section

ii

through

parallel-plate

capacitor with paraffin slab showing the

variation of the flux density D, charge

"

density p, electric field E, polarization P,

and polarization charge density p along
axis

Ax

!!

-7-P = ft,

between the plates. The thickness

of the charge layers is greatly

!i

'!
i

exaggerated.

where p is the volume charge density

component, dDJdx = div D. Thus,

dD
~dH

Hence the change of D with distance

in the

charge layer,

Since

has only an

= V-D = p

(in the

(5)

charge layer) equals the divergence of

and also the volume charge density. It follows that if the charge layer is infinitesimally
thin (A;* - 0), then V D and p approach infinity.
However, it is more reasonable

to consider that the charge layer

the

D and p may be
x

large, they are

axis of the capacitor

is

is

of small but

not

infinite.

shown graphically

finite thickness,

The

variation of

in Fig. 3-24.

so that although

D and

D along

THE STATIC ELECTRIC FIELD: PART 2

SEC. 3-26]

At

the paraffin surface

constant, but both

is

and

change.

From

101

(3-5-7)

P = D- E

Now

the change in polarization

Thus, from

polarization.

(6)

equal to the surface charge density p sp due to

is

(3-7-7),

P*l-Pn2=-Psp

Pnl = polarization in paraffin

P 2 = polarization in air gap
p sp = polarization surface-charge

where

Assume

the total change

Ax

is

uniformly distributed throughout

is

Then

at the paraffin surface, as suggested in Fig. 3-24.

APa in polarization from

Pnl

But when Ax

density

that the polarization surface charge

a thin layer of thickness

(?)

one side of the layer to the other

is

given by

-Pn2 =AP =APx

(8)

small,

dP
Therefore, (7) becomes

----'.
the volume density of polarization charge in the layer at the paraffin

where p p

is

surface in

coulombs per cubic meter.

in that

it

Since

charge.

div P.

Polarization charge differs

cannot be isolated, whereas true charge can.

has only an

x component, which

from true charge p

In this sense

it is

a function only of x,

is

fictitious

dPJdx =

Thus,

VP =

dx
If the polarization distribution is

uniform,

(11)

V P=

0.

distance (in the charge layer) equals the divergence of

p p of polarization charge.
small but

finite

The assumption of a

thickness results in a value of

The divergence of

that

yields the sources of the

divergence of P yields the sources of the polarization

The

variation of E, P,

graphically in Fig. 3-24.

and

may

be large but not

field (true charge),

is

of

infinite.

while the

field.

V P along the x axis

Hence the change of P with

and also the volume density

polarization charge layer that

V P

of the capacitor

is

illustrated

102

ELECTROMAGNETICS

It

[CHAP. 3

may be shown that the potential Vp in an unbounded volume due to a polariza-

tion distribution

is

given by

Vp

V-P
=-\
4ne
1

dv

(12)

Thus, when both true charge and polarization are present and the distribution of both
are fixed, the total potential

VT

unbounded volume

in the
r

V-P

4ne

=
P=
e =
r =

where p

true

volume charge

polarization,

Cm

- V-

-
1

VT =

or

4ne

Jv
V

4n

density,

is

dv

*dv
dv

(13)

Cm -3

-2

pFm"

permittivity of

vacuum =

distance from

volume element containing charge or polarization to point

at

which

VT is

8.85

to be calculated,

The volume integration is taken over

The field intensity E is then

all

regions containing charge or polarization.

E = -VKr
Whereas
tivity c, (13) is

dielectric

(2-12-5) applies to

more

media,

i.e.,

general since

(14)

a single homogeneous
it

dielectric

medium of permit-

can be applied also to space with several different

a nonhomogeneous medium.

For the case of a

single

homogene-

ous medium, (2-12-5) and (13) are equivalent.

3-27

THE LAPLACIAN OPERATOR;

POISSON'S AND LAPLACE'S EQUATIONS

As an extension of the divergence operator we

si-an) operator.
From (3-24-1)

Since

are led to the laplacian (la-plah'-

V-D = p

(1)

D= -eVV

(2)

V VV = - -

(3)

D = eE and E = - W, we have

Introducing (2) in

(1) gives

PROBLEMS]

This

E STATIC ELECTRIC FIELD:

PART 2

103

The double operator (divergence of

squared) and is called the laplacian operator.

Poisson's (pwas'-awns) equation.

is

the gradient)

also written as

is

Thus Poisson's equation can be

(del

written

(4)
e

If

= 0,

reduces to

(4)

V 2 V=0
which

known

is

(5)

as Laplace's equation.

In rectangular coordinates
2

v
o v
d

.,

lb?
The
if

static potential distribution for

o v
cfv

"a?

cfy
o v

...

dz

any conductor configuration can be determined

boundary

a solution to Laplace's equation can be found which also satisfies the

This

conditions.

is

discussed in Chap.

7.

PROBLEMS
*

3-1

floflCm"

flux density
ized,

plane slab of dielectric

what are

3-2

(a)
flat

in slab,

and

slab of sulfur

(e) field intensity

The

3-3

left

(b)

There

is

If

is

nC

m "2

what

with a

field

uniformly polar-

is

moment of the

slab?

is air-filled.

The

flat

is

is filled

plate separation

Find E, D, and

10-mm

is

with a dielectric of

10

mm, and

both halves, e

in

plate separation

there

= 3e

is

and a 100-V

and thickness 5 mm on the lower

Find E, D, and P in theairspace and in the dielectric.

a dielectric of permittivity
(a)

air), (rf) field intensity

slab (in air) ?

= 3*
= 5) of
e

large extent.

faces perpendicular to the direction of

a long needle-shaped cavity with

10

and

density outside the slab (in

a function of the distance between the plates,

Two cavities are cut in a dielectric medium (t r

D=

the total dipole

(c) flux

outside the

a thin disk-shaped cavity with

Cavity 2
air.

(b)

V as

Find

half of a horizontal parallel-plate capacitor

The space above is air-filled,

3-5

is

and

horizontal parallel-plate capacitor has a

potential difference.
plate.

placed normal to a uniform

= 4) is placed

potential difference between the plates.

3-4

5) is

normal to a uniform field. If the polari"2

what are (a) polarization
on the slab surfaces is 0.5 C

(e,

permittivity e while the right half

100

the polarization in the slab

zation-surface-charge density p lp
in the slab, (b) flux density in the slab,

(<>

If the slab occupies a volume of 0.1

its

axis parallel to D.

the magnitude of

E at

The

(a) the center

center of cavity 2?
* Answers to starred problems are given

in

Appendix C.

Cavity

D in the dielectric.

cavities are filled with

of cavity

and

(b)

the

104

ELECTROMAGNETICS

3-6

[CHAP. 3

parallel-plate capacitor

The space between the plates

is

is filled

is

square and has a plate separation of

with a dielectric

= 25).

(a,

If

kV

mm.

potential difference

applied to the plates, find the force in newtons squeezing the plates together.

The

3-7

electric field

above a slab of

in air

dielectric (e r

with respect to a normal to the plane surface of the slab.

What

an angle of 30

5) is at

the angle between

is

E and

the normal inside the slab ?

Three plane

3-8
together.
,

=2

If

dielectric slabs

E in air is at an

slab, find the angle

showing the path of the

of equal thickness with

= 2, 3, and 4 are sandwiched

e,

angle of 30 with respect to a normal to the plane surface of the

of

line

Draw

with respect to the normal inside the slabs.

a figure

through the sandwich.

What is the magnitude of E, D, and P 400 mm from a point charge of 1 nC in an

medium (a) of air and (6) of a dielectric with = 9?
3-10 What is e, for an artificial dielectric made of a uniform cubical lattice of metal

3-9
infinite

spheres in air

if

Show

3-11

arranged

in

mm in diameter

the spheres are 30

and spaced uniformly 60

mm between

and z directions?

centers in the x, y,

maximum

that the

a uniform cubical

possible

e,

for

lattice in air is 2.57.

an

It is

of metal spheres

artificial dielectric

assumed that there

is

no

interaction

between spheres and that they are almost touching.

3-12

horizontal parallel-plate capacitor of 'area

A dielectric

potential difference V.

midway between

3-13

o[l

+( e

has a plate separation d and a

e,

and thickness

t is

placed

the plates [air space above and below the slab of thickness (dt)/2].

Find E, D, and

slab of relative permittivity

and

in the slab

air spaces.

Neglect fringing.

parallel-plate capacitor of area

y/d)],

where y

Find E, D, P, and

V as

at

one plate and v

is filled

with a dielectric of permittivity

= d at the

other plate.

Neglect fringing.

Make a graph
showing the variation of e, E, D, P, and V as a function of y. (c) Find the capacitance.
(d) Find the polarization-volume-charge density p p as a function of y and the polarizationsurface-charge density p sp at y =
and y = d. (e) Make a graph of p p vs. y.
(a)

3-14

Repeat Prob. 3-13 for the case where the plate

the slab but the other plate

to a distance

at

= 1d with

is air.

114)

is

Neglect fringing.

remains in contact with

the space between y

= rfand

N mm ~

to 1-m-square

3-6).

Find the capacitance

if

10

mm plate separation.

The

a sheet of powdered rutile (take

(a) sheet
mm thick
mm air), (b) sheet 2 mm thick, (c) sheet 4 mm thick, (d) sheet 6 mm thick,
mm thick, (/) sheet 10 mm thick (g) Draw a graph of the capacitance vs. sheet

placed between the plates under the following conditions:

(remaining 9
(e) sheet 8

mm thick (see Prob.

A parallel-plate capacitor is 200 by 200 mm with

3-16

medium
e,

moved

Design an electrostatic press for applying pressures of

plywood sheets 10

is

(b)

2d air-filled.
3-1 5

'

a function of distance y between the plates,

is

thickness.

3-17

(a) If

100

is

applied to the capacitor of Prob. 3-16, calculate the energy stored

for the six conditions listed,

for the six conditions.

(b)

Find the energy density

in the dielectric

and

in the air space

THE STATIC ELECTRIC HELD: PART 2

problems]

is

the

amount of energy which can be

What

3-19
if

the sphere

as 3

MV m

capacitor would be required to store

What volume of

3-18

This

is

is

300

the

kW

105

hr of energy?

stored in a typical automobile storage battery.

maximum voltage to which an isolated

mm in diameter and situated in air?

metal sphere can be charged

Take the

dielectric strength

of

air

_I
.

3-20

parallel-plate air capacitor

is

charged to 1,000 V, the potential source dis(a) What is the energy

connected, and the plates separated to twice their original spacing,

stored in the two cases?

3-22
distance
e

=e

(6)

Account for the difference

in energies.

Relate Prob. 3-20 to the voltage buildup in a storm cloud in a rising air current.
The permittivity of the dielectric in a parallel-plate capacitor is a function of the

3-21

y between the

plates (y

at

one plate and y

= yi

at the other plate) as given

by

Find the capacitance.

e".

3-23 A charge Q is uniformly distributed in a circular ring of radius r,. Find

E and V at the center of the ring.
Find
*
3-24 A charge Q is uniformly distributed in a square ring of side length
ring.
E and V at the center of the
Find
3-25 A charge Q is uniformly distributed in a half-circular ring of radius r
E and V at the origin.
*
3-26 A charge Q is uniformly distributed along a line of length / situated coincident
Find V at a point x,
with the x axis with one end at the origin and the other end at x =
/.

/.

on the x

axis

where x >l.
i

Repeat Prob. 3-26 for the case where the linear charge density varies as x
Map the field lines and equipotential contours of a coaxial transmission

3-27
3-28

consisting of a circular inner conductor of diameter

d symmetrically

conductor of square cross section with an inner side dimension of

symmetry only one octant (45 sector) need be mapped.
Referring to Prob. 3-28, (a) what

3-29
(6) If

100

is

applied to the

line,

what

is

is

line

located inside an outer

3d.

Note that thanks

to

the capacitance per meter length of the line?

the potential of a point

midway between

the inner

conductor and a corner?

3-30

A and

B(r =
3-31

If

100

V is applied to the plates of the capacitor of Fig. P3-30, find .Eat the points

10 mm).
If

100

V is applied to

the asymmetrical coaxial transmission line of Fig. P3-31,

2,5r

FIGURE

P3-30

Parallel-plate capacitor with metal half-cylinders attached.

106

[CHAP. 3

ELECTROMAGNETICS

FIGURE

P3-31

Cross section of asymmetrical coaxial

transmission

find

Vat the point

What

3-32

shown

in Fig.

line.

P.
is

The

potential of the outer conductor

is

zero.

the ratio of the capacitance per unit depth of capacitor

P3-32?

to capacitor

Neglect fringing.

A coaxial transmission line consists of an inner conductor of diameter d situated

3-33

symmetrically inside a triangular outer conductor of equilateral cross section with inner
side lengths all equal to 3d.

(a)

Map the field lines and equipotential

contours,

(b)

What

is

the ratio of surface charge density at the center point of a side of the outer conductor to
that at a point
line,

what

is

on the side midway from the center to a corner ? (c) If 100 V is applied to the
midway from the inner conductor to a corner? The

the potential at a point

outer conductor

3-34

is

at zero potential,

(d)

What

is

the capacitance per unit length ?

A high-voltage conductor is brought through a grounded metal panel by means

of the double concentric capacitor bushing shown in Fig. P3-34.

strength 3

MV m _1

)-

The medium is

Neglect fringing and thickness of sleeves,

(a)

air (dielectric

Find the outer sleeve

L which equalizes the voltage across each space, (b) Find L which results in the same
maximum field in each space, (c) What is the maximum working voltage for each
condition ? (d) If the inner sleeve (200 mm long) were removed, what would be the maximum
length

value of

working voltage?

(e) If

the

number of concentric

sleeves

is

increased in

number so

that

FIGURE P3-32
Cross section of two capacitors.

THE STATIC ELECTRIC FIELD: PART 2

problems]

107

-Metal panel
High- voltage

50mm

conductor

40 mm

30mm

MrUVJ m:!.
i

200 mm-

FIGURE

P3-34

Concentric capacitor bushing.

the spacing between sleeves becomes smaller, what is the ultimate working voltage ?

graphs of E

vs.

radius for the conditions in parts

(a), (b), (d),

and

(/) Draw

(e).

A charge Q is uniformly distributed in a thin circular ring of radius r\. Derive

and Kas a function of radius r (a) in the plane of the ring and (b) perpendicthe plane of the ring along its axis, (c) Draw graphs of E and V for both cases to

3-35

expressions for
ular to

a distance of 4r t
3-36

charge

is

(a)

Derive expressions for

(b)

Draw graphs
3-37

i.e.,

of

uniformly distributed over a plane square area of side length

/.

E and V as a function of distance normal to the center of the area.

E and V to

a distance of

5/.

Repeat Prob. 3-36 for the case where the charge has a cosine distribution,

the charge density

is

maximum at

the center of the area and tapers as a cosine function

to zero at the edges.

3-38

Map

the field lines and equipotential contours for the following situations:

+Q and +2Q separated by a distance d. (b) Two point charges

+ Q and 2Q separated by a distance d. (c) Four point charges of equal magnitude form-

(a)

Two

point charges

Four equal point charges arranged as in Fig. 2-2\a

array of equal positive and negative point charges
arranged in a uniform rectangular grid. Adjacent charges are of opposite sign. Use a
computer with machine or cathode-ray-tube plotter. (/) Repeat (e) for a uniform hexagonal
grid.
Note that (e) and (/) are but two of many configuration which can give interesting
ing a quadrupole as in Fig. 2-2la.

but

all

of the same sign,

(e)

An

(rf)

infinite

displays with a computer.

3-39

Find the divergence of the following vector functions

= x sin az + y cos ax + iy
(b) G = x cos ay + yx i sin az
3-40 Show that V P = (c - l)e V E.
3-41 A sphere of radius r^ has a uniform charge density p. (a) Find D
as a function of radius, (b) Make a graph of D and V D to a radius of 3ri.
(a)

and

V D

ELECTROMAGNETICS

108

[CHAP. 3

3-42

Repeat Prob. 3-41 for the case where the charge density p

3-43

Show that the total

= kr,

where k

con-

stant.

acitance of the largest

capacitance of n unequal capacitors is (a)

when they

are connected in parallel

of the smallest when they are connected in

(6) less

(c)

A = &3x + y4y 3 + z5z 2

C = x2 + ylx 2y 2 z 2 Iz ~

(e)

E = f sin< +

(g)

G = ir "' + 6 sin

series.

(b)
l

(d)

+ zz
4> sin 9 sin
sin
2

4>cos<

<f>

(/)
(h)

<j>

B = \2xy 2 z - y3xz + iAx

D = xy + yz + ixy
F = rlr 112 <j>sin <f>cos + tan
H = rz + 6 cos 9 sin + 4>r" 2

U=-xT

(6)

U=x y*z

(c)

V= -

(d)

V=-

+ y 12

(cylindrical coordinates)

V=e
W=x

(cylindrical coordinates)

<f>

(/)

(spherical coordinates)

(spherical coordinates)
V=
e
W=xy + xz
.

-yz

(h)

Problems 2-28 and 2-29 were to be solved by integrating

3-46

of the cubical volume

the divergence of

(D-rfs=

Q).

and integrating

Now

over the surface

and 2-29 by first taking

volume of the cube (J| Jv D dv =

solve Probs. 2-28

this result over the

H\pdv = Q).

A parabolic line of charge with vertex at the origin and focus at x = m extends
and y directions. If p L nCm"', find Kand E at the focus. An

3-47
3

i.

<f>

tf>

(a)

(g)

the cap-

Find the gradient and laplacian of the following scalar functions

3-45

(e)

more than

than the capacitance

Find the divergence of the following vector functions

3-44
(a)

and

m in both +y

approximate solution can be obtained by dividing the parabolic-line charge into segments

and adding

their contributions.

Repeat Prob. 3-47 for the case where the parabolic

3-48

line

of charge extends only

= 0toy + Jim.
3-49 A point charge + Q is situated at the center of a dielectric sphere of permittivity

from y

and radius R.

Kand in

surface charge density required

the dielectric and in the air outside. (6) Determine the

on the surface of the sphere to maintain the same field inside

the sphere but reduce the

outside to zero.

What is

static

Find

field

A typical 12-V automobile lead acid storage battery can store kW hr of energy.

3-50
(a)

(a)

the smallest practical volume in which

form?

(6)

What

is

W hr of energy can be stored in electro-

the ratio of this volume to that of the storage battery?

THE STEADY ELECTRIC CURRENT

4-1

INTRODUCTION

Electric charge in

charge

is

motion constitutes an

by

carried

electrons.

carried by ions, both positive

by electrons and

and

electric current.

In metallic conductors the

In liquid conductors (electrolytes) the charge

negative.

In semiconductors the charge

is

is

carried

holes, the holes behaving like positive charges (see Sec. 4-7).

In this chapter the important relations governing the behavior of steady electric
currents in conductors are discussed.t

The

fields associated

static fields.

static fields

with steady currents are constant with time and hence are

In Chaps. 2 and 3 the discussion was almost entirely concerned with

having

In this chapter the

all

associated charges stationary,

fields are also static,

i.e.,

with no currents present.

but steady direct currents

may

be present.

t Specifically, a steady direct current is meant. This should not be confused with a
steady-state current, which may imply a time-changing current that repeats itself
periodically.

110

4-2

ELECTROMAGNETICS

[CHAP. 4

CONDUCTORS AND INSULATORS

In some metals, like silver and copper, there

occupied

of the atom.

shell

from atom

This electron

is

but one electron in the outermost

is

so loosely held that

it

migrates easily

atom when an electric field is applied. Materials that permit such

electron motion are called conductors.
Silver and copper are examples of good
conductors, their resistance to electronic motion being relatively slight. Not all
good conductors have only one electron in the outermost occupied shell. Some have
two, and a few, such as aluminum, three. However, in all conductors these electrons
are loosely bound and can migrate readily from atom to atom.
Such electrons are
to

often called true charges.

may

In other substances, however, the electrons

be so firmly held near their

normal position that they cannot be liberated by the application of ordinary

These materials are called
insulator

may produce no

Although a

dielectrics or insulators.

migration of charge,

insulator, or dielectric (Sec. 3-5),

their equilibrium positions.

it

field

fields.

applied to an

can produce a polarization of the

a displacement of the electrons with respect to

i.e.,

The charges of an

insulator are often called

bound or

polarization charges in contrast to the free, or true, charges of a conducting material.

Certain other materials with properties intermediate between conductors and

Under some conditions such a substance may

sufficient field may become a

insulators are called semiconductors.

an insulator but with the application of heat or

act like
fair

4-3

conductor.

THE ELECTRIC CURRENT

When an isolated conducting object is introduced into an electric field, charges migrate
(currents flow) until a surface charge distribution has been built
total field in the

and the applied

conductor to zero.
field is

flowing.

field

infinitely

The

continuous across a boundary.

to migrate parallel to the

not isolated

long conductor, such as a metal wire, in a

field.

The field causes

component of the

electric

the electrons in the conductor

Since the electrons are negatively charged, they

in a sense opposite to the field direction.

If there are n electrons per

of the conductor and their average velocity

isoms" 1

is

field

the boundary relation (1) of Table 3-2 that the tangential

field is

that reduces the

E in the conductor is not zero if current is

E is the same inside and outside the conductor. This follows from

as in Fig. 4-1

Rather,

up

however, the conducting object

maintained, current will continue to flow in the conductor.

For example, consider an

uniform

If,

move

meter of length

then the total charge per second

However, they may be torn off by mechanical means such as rubbing.

THE STEADY ELECTRIC CURRENT

SEC. 4-4]

111

.Wire

E
1

^-Electron

-*.

"7\

the charge of each electron.

The

FIGURE
Infinite

4-1

conductor in uniform

passing a fixed point on the wire

electric current / in the wire is

V-

field.

is

nqv,

where q

is

then defined as

I=-nqv

(1)

The electric current is, by definition, taken to be in the opposite direction to the electron
motion. Hence the negative sign in (1). Electric current has the dimensions of
current, or in dimensional

symbols /

= Q/T).

The SI

unit of current

is

the ampere

Thus,

(A).

Amperes

coulombs

second

That

is,

lCs"

4-4

a current of

A is produced by charge flowing by a fixed point at the rate of

1
.

RESISTANCE

In 1826 Georg Simon

voltage

AND OHM'S LAW

Ohm

experimentally determined the relations between the

over a length of conductor and the current / through the conductor in

terms of a parameter characteristic of the wire (see Fig.

the resistance R,

is

defined as the ratio of the voltage

R=The

latter

form

is

resistance
It is

and

assumed that

such a resistance

is

said to

This parameter, called

/.

Thus,

Ohm's

law,

(1)

which

between the ends of a conductor

the current

4-1).

to the current

V = IR

or

the usual statement of

difference or voltage

is

states that the potential

equal to the product of

its

I.

is

independent of

obey Ohm's law.

/; that is, J? is

There

a constant.

Conversely,

exist circuit elements,

however,

whose resistance is not a constant. Such elements are said to be

nonlinear, and a V-\s.-I diagram is required to display their behavior.
Nevertheless,
the resistance of the nonlinear element is still defined by R = V/I, but R is not independent of / and the resistance does not obey Ohm's law.
such as

rectifiers,

112

ELECTROMAGNETICS

[CHAP. 4

Resistance has the dimensions of potential divided by current or, in dimensional

symbols,

Ml} IIT 3

The

SI unit of resistance

Volts

Ampere
That

is,

the resistance of a conductor

difference of

is

is

the

ohm

Thus

(CI).

= ohms
a current of

fl if

A flows when

JOULE'S

LAW

Referring again to Fig. 4-1, the potential difference or voltage

of the conductor

move

is

per unit time, or power P.

work
= power
=
time

Current x time

VI

or

the

power

Multiplying by the current / yields the work

d.

x current

is

across the length

Thus,

Work

This

equal to the work per unit charge (current x time) required to

a charge through the distance

the watt (W).

a potential

applied between the ends of the conductor.

POWER RELATIONS AND

4-5

is

dissipated in the length

=P

(1)

d of the conductor.

The

SI unit of power

is

Hence,
Volts x amperes

watts

(W)

or in dimensional symbols

ML2 _ Ml}
Introducing the value of

V from Ohm's law

(4-4-1) into (1) yields

P = I 2 R\

(2)

Thus, the work or energy dissipated per unit time in the conductor

product of

its

resistance

and the square of the current

is

given by the

This energy appears as

/.

heat in the conductor.

The energy

W dissipated in the conductor in a time T

W = PT = I RT

is

then

(3)

THE STEADY ELECTRIC CURRENT

sec. 4-6]

113

W = energy, J

where

P = power, W
/ = current, A
R = resistance, Q
T = time, s
This relation
If

it is

4-6

The

is

known as Joule's law.

not constant, I

is

It is

assumed that P is constant over the time

THE ELECTRIC CIRCUIT

discussion in the preceding sections concerns an infinitely long conductor along

which a
length

field

d as

is

applied (Fig. 4-1).

in Fig. 4-2a

Consider

is

now

a cylindrical conductor of

finite

between two large conducting blocks of negligible resistance

maintained at a constant potential difference

field

T.

integrated over the time interval T.

along the conductor.

This

field is

V by a battery.

given by

E=

This produces a uniform

V/d.

As long

as this field

maintained in the conductor, current flows that has a value

*!

-WAWV

(6)

FIGURE

4-2

conductor of length d
between end blocks, and (b) schematic
diagram.
(a) Cylindrical

[CHAP. 4

ELECTROMAGNETICS

114

Assuming that the resistance of the wires connected to the battery is negligible
compared with R, the potential difference V is equal to the voltage appearing across
the terminals of the battery. The arrangement of Fig. 4-2a may then be represented
by the schematic diagram of Fig. 4-2b.
This is a diagram of a closed electric
ing of a resistor of resistance

The

resistance of a conductor

conductor

is

made

but also on

different types of substances,

only of the substance.

The

of the most elementary form consistV.

depends not only on the type of material of which the

To facilitate comparisons between
its shape and size.

it is

convenient to define a quantity which

resistivity

rialjs numericallvequal to the

is

is

characteristic

Theresistiyity of

such a quantity.

the cube between two heavy blocks of negligible resistance, as in Fig. 4-3.
current /through

is

amate-

of a homogeneous unit cube of the material

This condition may be produced by clamping

resist ance

uniform current distribution.

with_.a

circuit

and a battery of voltage

AND CONDUCTIVITY

RESISTIVITY

4-7

the cube, the resistivity

S of the material

is

given by

S=

With a

V/I,

where

the potential between the blocks.

Unit cube

FIGURE

4-3

Unit cube between end blocks.

In SI units, this measurement

If

two

while

is

in

ohms

for a cube of material

on a

cubes are placed in series between the blocks, the resistance measured

if

sistance

two cubes are placed

in parallel, the resistance is $S.

of a rectangular block of length

R=

SI

or

and cross section

S~

It

a,

side.

is

25,

follows that the re-

as in Fig. 4-4,

is

(0
I

SEC. 4-7]

THE STEADY ELECTRIC CURRENT

^?^
h

FIGURE

4-4

From (1)

of the block material.

resistivity

Resistance x area

= resistance

Length
Thus,

resistivity

~~p^
;

J S" /"

Block of conducting materia).

where S is the

J 1

S'

resistivity

has the dimensions

x length

has the dimensions of resistance times length and in SI units

(J2 m).

is

ex-

pressed in ohm-meters

The

a function of the temperature.

resistivity is

varies nearly linearly with absolute temperature.

range from a reference or base temperature

In metallic conductors

it

Over a considerable temperature

S is

the resistivity

given approximately

by

S=S U +
where

ol(T

- T )]

T = temperature of material ,C
T = reference temperature (usually 20C)
S = resistivity at temperature T Sim
= temperature coefficient of resistivity, numerical
,

a.

EXAMPLE
coefficient

For copper the

resistivity

= 0.0038C -'.

Find the

The

S at

solution

resistivity

5=

1.77

X 10

at

20C

resistivity at

a temperature
"8

[1

is

units

17.7 nfl

_1

and the corresponding

30C.

T is

+ O.0O38(r- 20)]

(12

m)

At a temperature of 30C,

5=
This

is

1.77

x 10 ""[I +0.0038(10)]=

an increase of nearly 4 percent over the

At temperatures near absolute zero (T

superconducting, an effect
resistivity

first

resistivity at

Kf

nOm

20C.

or -273C) some metals become

observed by H. Kamerlingh Onnes in 1911.

drops to zero, and the magnetic

sition is very abrupt.

18.4

field is expelled,

so that

B = 6.

The

The
tran-

Niobium

(also called columbium) becomes superconducting

K, but copper and gold are not superconducting, at least
not at temperatures down to 0.05 K. Some intermetallic compounds like Nb 3 (Al-Ge)
at 9.2

K, aluminum

at 1.2

become superconducting at the relatively high temperature of about 21 K.

t

The

SI unit for absolute temperature

is

the kelvin (K) (see Sec. 1-3).

If materials

16

[CHAP. 4

ELECTROMAGNETICS

are developed which

become superconducting

at 25

K or more, a technological revoluby

tion will occur since this will permit use of superconductors cooled

relatively

inexpensive liquid hydrogen (boiling temperature 20.4 K).f

The

reciprocal of resistance
is

reciprocal

ohm

called

is

is

an upsidedown

reciprocal of resistivity

S is
a

resistivity is

nient to deal with

in

its

mhos per meter (y

capital

omega
That

conductivity a.

given in

is

(y).
is,

(la)

e.g.,

it is

where

more conve-

often

parallel circuits are

expressed in ohm-meters, the conductivity

is

Since

l/R.

1-JS~ Ra

reciprocal, the conductivity,

G=

expressed in reciprocal ohms.

convenient in certain applications,

Since resistivity

involved.

is

is,

a mho (ohm spelled backward); conductance

mhos, and the SI symbol

Although the

That

conductance G.

is

expressed in ohms, conductance

resistance

The

is

expressed

_1
).

The resistance^ of a rectangular block,

as in Fig. 4-4, of material of conductivity

o-is

R=
or the conductance

aa

(2)

$2)

bl ockis
is
of the block

= -1 =
G

where a

(ohms,

aa

conductivity of block material,

(mhos, y)

(3)

_1
ym

= cross-sectional area of block, m 2

= length

of block,

semiconductor contains electrons and holes (vacant spaces in the semi-

conductor

lattice).

like positive charges.

electrons move opposite to E, but

From J = cE the conductivity may be

The

J
where p

is

the holes

move with E

expressed.

=
PE PV

(4)

the electron (or hole) charge density, in coulombs per cubic meter.

and
the drift velocity v of the electron (or hole) per unit field
2 _1
V -1 ). The mobility of electrons and holes is different.
the mobility (m s
quantity

we may

/i is

is

The
called

Hence,

write for the conductivity of a semiconductor

t B. T. Matthias, The Search for
23-28 (August 1971).

High-temperature Superconductors, Phys. Today, 24:

THE STEADY ELECTRIC CURRENT

sec. 4-8]

= PeUe +

PhUh

117

(Um -1 )

(5)

= charge density of electrons, Cm -3

= mobility of electrons, m 2 s _1 V -1
e
-3
p h = charge density of holes, Cm
-1
2
H h = mobility of holes, m s" V

where pe
fi

4-8

The

TABLE OF CONDUCTIVITIES
conductivities

a of some

common

materials are listed in Table 4-1 for a temper-

ature of 20C and for a superconductor at temperatures below 21 K.

TABLE OF CONDUCnvmES

Table 4-1

Conductivity,

Urn-

Substance

Type

Quartz, fused
Ceresin wax

Insulator
Insulator
Insulator
Insulator
Insulator

~10" 17
~io- 17
~io- ,J
~io- 15
~io- 15
~io-' 5

Distilled water

Insulator
Insulator
Insulator
Insulator

Seawater

Conductor

Tellurium

Conductor
Conductor
Conductor
Conductor
Conductor

Sulfur

Mica
Paraffin

Rubber, hard
Glass
Bakelite

Carbon
Graphite
Cast iron

Mercury

Nichrome
Constantan
Silicon steel

German

silver

Lead

Zinc
Tungsten

Duralumin
Aluminum, hard-drawn
Gold
Copper

Conductor
Conductor
Conductor
Conductor

Silver

Conductor

Nb 3 (AI-Ge)

Superconductor

Phosphor bronze
Brass

~4

~5x
~3 x

10 2
10*

~10 s
~10 6
10 6
10 6
10 6
10 6
10 6
6
5 x 10

Conductor
Conductor
Conductor
Conductor
Conductor
Conductor
Conductor
Conductor
Conductor
Conductor

Tin

~io- 12
~io-9
~io-4

2x
2x
3x

9x

10 6
10 7
7
1.1 x 10
7
1.7 x 10
7
1.8 x 10

5.7

x
x
x
x

10 7
10 7
10 7
10 7

6.1

10 7

3
3.5
4.1
,

~00

"

118

4-9

[CHAP. 4

ELECTROMAGNETICS

CURRENT DENSITY AND OHM'S LAW AT A POINT

If the current is distributed

current density

uniformly throughout the cross section of a wire, the

uniform and

is

is

That

sectional area a of the wire.

given by the total current / divided by the crossis,

jJ-

(1)

Current density has the dimensions of current per area and in SI units
-2
amperes per square meter (Am ).
If the current

is

not uniformly distributed,

(1) gives the

is

expressed in

average current density.

However, it is often of interest to consider the current density at a point. This is

denned as the current AI through a small area As divided by A* with the limit of this
Hence, the current density at a point is given by
ratio taken as As approaches zero.
,

It is

density

assumed that the surface As

is

at the point

is

AI

r
lim
As-0 AS

(2)

normal to the current

direction.

The current

a vector point function having a magnitude equal to the current density

and the direction of the current

Consider

now

at the point.

a block of conducting material as indicated in Fig. 4-5.

Let a

and cross section a be constructed around

a point P in the interior of the block with a normal to J as indicated. Then on
applying Ohm's law to this cell we have V = IR, where V is the potential difference
small imaginary rectangular

between the ends of the

cell

of length

But

cell.

V = El and

= Ja;

so El

= JaR

and

(3)

aR
By making the cell
at the point P,

enclosing

P as

small as

we

wish, this relation can be

made to apply

and we can write

FIGURE

= ctE

(4)

4-5

Block of conducting material with small

imaginary cell enclosing the point P.

jf

mzL

THE STEADY ELECTRIC CURRENT

SEC. 4-10]

Equation
cussion

at a point and relates the current density J at a point to the

and the conductivity a of the material. In the above disassumed that the conducting material is homogeneous (same material

(4) is

total field

119

Ohm's law

at the point

it is

throughout), isotropic (resistance between opposite faces of a cube independent of

the pair chosen), and linear (resistance independent of current).

KIRCHHOFF'S VOLTAGE LAW AND THE

DIFFERENCE BETWEEN POTENTIAL AND EMF

4-10

Consider the simple

The

circuit consists

shown by

electric circuit

of

a resistor

and the

the schematic diagram in Fig. 4-6.

battery.

The current is /at all points in

we have from Ohm's

At any point in the conducting material of the circuit

the circuit.

law at a point that J/a

E, where

E is the total field

E may

In general the total field

at the point.

be due not only to

other causes such as the chemical action in a battery.

static

To

charges but also to

indicate this explicitly,

we write

E = Ec + Ee
where

Ec =

static electric field

that the field

Ee =

\V),
a

is

Ec

this is

to indicate explicitly

by other causes as by a battery; the subscript e

to indicate explicitly that

is

is

due to charges

electric field generated

is

Whereas

is

(1)

due to charges the subscript c

it is

an emf-producing

field

derivable as the gradient of a scalar potential due to charges (E c

not the case for

Ee

Substituting (1) in (4-9-4),

writing J = I/c,

the cross-sectional area of the conductor, and noting the value of

a from

where

(4-7-2),

we have
-=
where Rjl

is

Iy = Ec + Ee

the resistance per unit length in

point in the circuit.

ohms per

(2)

meter.

This applies at any

Integrating around the complete circuit gives

JEC

dl

+ j>Ee

FIGURE

dl

= lj-dl

4-6

Series circuit of battery

resistance.

(3)

and external
<{j

120

The

ELECTROMAGNETICS

first

term

[CHAP. 4

zero;

is

the line integral of a

i.e.,

closed circuit (see Sec. 2-8, last paragraph).

the line integral of E e around the circuit
total electromotive force, or emf-f

is

field due to charges is zero around a

However, the second term involving

not zero but

1J r of the

present example, by chemical action in the battery.

flow since an electric

The right-hand

field

Ec

due to charges

is

is

equal to a voltage called the

The

circuit.

If it

in the

not able to maintain a steady current.

IR drop around

side of (3) equals the total

Ee is produced,

field

were absent, no current would

the circuit.

Hence

(3)

becomes

^t = IR T
where

R T is the total resistance of the

battery

(4)

circuit (equals

if

the internal resistance of the

zero).

is

In general, for a closed circuit containing

many

resistors

and sources of emf,

X o = /Zj?
c

This

is

Kirchhoff's voltage law.

around a closed

it

sum of the emfs

ohmic or IR drops around the

states that the algebraic

sum of

circuit equals the algebraic

the

Kirchhoff's voltage law applies not only to an isolated electric circuit as

circuit.%

any single mesh (closed path) of a network.

in Fig. 4-6 but to

To

In words

(5)

emf from the scalar potential V, the symbol "V (script V) is used
and IT are expressed in volts, so that either may be referred to as

distinguish

Both

for emf.

a voltage
It is

if

one does not wish to make a distinction between potential and emf.

to be noted that the scalar potential

static field

Ec

while the

emf

"Vt

is

equal to the line integral of the

equals the line integral of

Ee

Thus, between any

two points a and b,

Vab = Vb and
In ( 6 )
is

<Vah

Kb

is

= -

=\\

"

Ec

d\

(6)

-dl

(7)

independent of the path of integration between a and

b,

but HJ^, in

(7),

not.
t Also called electromotance.
X In time-varying situations,

where the

circuit

dimensions are small compared with

The algebraic sum of the instantaneous

emfs around a closed circuit equals the algebraic sum of the instantaneous ohmic drops
the wavelength, Kirchhoff's law

around the

is

modified

circuit.

An open-circuited battery (no current flowing) has a terminal potential difference

V equal to its emf IT. The potential V is as given by (6). As explained in the
examples that follow, Ec and Ee have opposite directions in the battery. Therefore,
in order that

Ki,

lT (Ib for

an open-circuited

battery, (7) has

no negative

sign.

THE STEADY ELECTRIC CURRENT

SEC. 4-10]

Ec

For closed paths,

around the circuit.

EXAMPLE

d\

and

Ee

d\

V,. , where

Let the circuit of Fig. 4-6 be redrawn as in Fig. 4-7a.

internal resistance
in the battery is

R it

and

it

will

be convenient in

this

uniform between the terminals (c and d).

to be at zero potential.

The

resistor has

Ec = 0. The field

The

the total

field

Ee

The point b (or c) is taken arbitrarily

a uniform resistance

and the wires connecting

(ct

= oo).

Hence, in

E has a value only in the battery, being zero elsewhere.

the variation of the potential

emf

battery has an

example to assume that the

the resistor and the battery are assumed to have infinite conductivity
the wire,

lJ r is

121

Find

V around the circuit.

(a)

FIGURE

4-7

(b)

of battery and external

resistance with (6) graph showing vari(a) Series circuit

ation of potential around circuit.

the

solution By Kirchhoff's
sum of the IR drops. Thus

voltage law the

Ohm's law

the emfs around the circuit equals

= IRo + IRi

\3

In the resistor

sum of

R^+Rr

Ee = 0, but Ec has a value (as discussed in connection with Fig.

(4-9-4) to the resistor (between

a and

b),

we have

E,
(TO

where a
/o

= conductivity of resistor material (assumed uniform)

= distance from a to b

Integrating

from a to b

yields

f
J.

or

'o

Ec

.rfI

r*0
= /92
la

f dl
J.

4-2).

Applying

122

ELECTROMAGNETICS

[CHAP. 4

^Trough

^Trough

FIGURE

upper

level

lower level

4-8

Hydraulic analog for electric circuit of Fig. 4-7.

where

Vmt is the potential

difference between

to c with perfectly conducting wires,

a and

appearing across the terminals of the battery.

Since point a

b.

Vcd ~ Vab

where

Vcl

is

is

connected to

d and b

the potential difference

we have

Therefore,

= Vct + IR,
Vci = U - IR

17

or

Thus, the potential difference appearing between the terminals of the battery

emf 17 of the

battery minus the drop

V around

variation of the potential

It is

instructive to

the above example (Fig. 4-7)

horizontal trough at what

is

equal to the

to the internal resistance of the battery.

the circuit

compare the

analogous hydraulic systems.

potential in Fig. 4-7.

IR due

is

The

as indicated in Fig. 4-76.

electric circuits of the

above examples with the

Thus, a hydraulic system analogous to the circuit of

is

may

shown

Between b and c there

in Fig. 4-8.

be called a lower

Between c and d there

is

level,

pump which

other liquid against the gravitational field in the same

raises positive charge against the static electric field

an open

raises the water or

manner as the

Ec

is

corresponding to the ground

battery in Fig. 4-7

Thus, the water in the upper

trough has a higher potential energy ihan the water in the lower trough in the same
way as the charge in the wire between d and a in the electric circuit of Fig. 4-7 is at a
higher potential than the charge in the wire from b to
in a horizontal frictionless trough at

an upper

conducting wire between these points in Fig. 4-7.

c.

level

From d to a the water moves

corresponding to the perfectly

From a to b the water falls through

a pipe to the lower level and in so doing gives up the energy

pumped

to the upper levef.

The pipe

it

acquired in being

offers resistance to the flow of water,

energy given up by the water appears as heat.

and the

This energy is analogous to that appear-

ing as heat in the resistor of Fig. 4-7 owing to charge falling in potential from a to

In this analogy the

b.

pump does work, raising the water against the gravitational field in

THE STEADY ELECTRIC CURRENT

SEC. 4-11]

the same

manner

as the chemical action in the battery does

(against the electrostatic field

Ec

work per

123

unit charge

and internal resistance R,) equal to the emf of the

battery.

In a single-cell battery with two electrodes the

field

Ec

is

largely confined to a

thin layer at the surface between the electrode and the electrolyte and

between the two electrodes.

electrolyte

preceding example

is

could be approached

connected in

series

Thus, the potential variation

not representative of an actual two-electrode

if

each battery consisted of a large number of

between c and d

is

zero in the

assumed

cells

of small

it

emf

in Fig. 4-7.

TUBES OF CURRENT

4-11

In Chap. 2

we

Fig. 4-9.

Let

Let us

discussed tubes of flux.

namely, that of tubes of current.

all

A tapered

now

consider an analogous concept,

section of a long conductor

the space in the conductor be

filled

is

shown

also parallel to the electric field.

current tube, the total current I

in

Each tube

with current tubes.

everywhere parallel to the current direction and hence, from the relation J
is

in the

although

cell

is

ctE,

Since no current passes through the wall of a

through any cross section of a tube

is

a constant.

Thus

/=/! J

= current density, Am -2
a = cross section of tube (over

ds

= constant

(1)

where J
If

is

which J

is

integrated),

constant over the cross section and normal to

it,

m2

then I

= Ja

or, referring

to

the current tube of rectangular cross section in Fig. 4-9,

I

= Jbd

where b and d are the tube cross-sectional dimensions

(see Fig. 4-9).

Current tube

Conductor

FIGURE

4-9

Tapered section of a long conductor showing current tube.

124

ELECTROMAGNETICS

If all

[CHAP. 4

of the conductor

is

divided

up into current tubes, each with the same current

is I = I n, where n is the number of

the total current / through the conductor

current tubes.

Surfaces normal to the direction of the current (or field) are equipotential

The

surfaces.

by a distance

resistance
is

uniform

V between two equipotential surfaces separated

by Ohm's law equal to the current I in a current tube times the

potential difference

/ is

of a section of tube of length

uniform), the resistance

(field

R=
where

Thus,

/.

R is,

from

aa

V=

I R.

If the current density

(4-7-2), given

by

(Q)

(2)

= length of tube section, m

conductivity of conducting medium, Urn -1
a = cross-sectional area of tube, m 2
/

4-12

KIRCHHOFF'S CURRENT LAW

Whereas

flux tubes in a static electric field begin

and end on

electric

charge and hence

are discontinuous, the tubes of a steady current form closed circuits on themselves

and hence are

To

continuous.

describe this continuous nature of steady currents,

That is, it has no sources or sinks (ending

and end on electric charges in a static electric
field.
As a consequence as much current must flow into a volume as leaves it. Thus,
in general, the integral of the normal component of the current density J over a closed
surface s must equal zero, or
it is

said that the current

places) as

do the

is

flux tubes,

solenoidal.

which

start

|j-rfs

This relation

is

volume may be

for steady currents

entirely inside a

(1)

and applies to any volume.

conducting medium, or

it

may

For example, the

be only partially

filled

The conductors may form a network inside the volume, or they

may meet at a point. As an illustration of this latter case, the surface S in Fig. 4-10
encloses a volume that contains five conductors meeting at a junction point P. Taking
the current flowing away from the junction as positive and the current flowing toward
with conductors.

the junction as negative,

we have from

(1) that

h ~h ~h +h-h =

THE STEADY ELECTRIC CURRENT

sec. 4-13]

FIGURE

125

4-10

Junction point of several conductors.

In other words, the algebraic

sum of

the currents at

a junction

is

zero.

This

is

Kirchhoff's current law, which can be expressed in general by the relation

I/ = o

4-13

(2)

DIVERGENCE OF J AND CONTINUITY

RELATIONS FOR CURRENT

Consider the small volume element Av shown in Fig. 4-11 located inside a conducting

medium.

The

In general,

it

has three rectangular components that vary with position.

ment here

is

formally the same as for

FIGURE

current density J

is

a vector having the direction of the current flow.

D in

Sec. 3-23

V- J

and

The develop-

yields

(1)

4-11

Construction used to develop differential

expression for divergence of J.

./

126

[CHAP. 4

ELECTROMAGNETICS

AJ

FIGURE

Q4-12

Construction for the continuity relation

between current and charge.

It states that steady currents have no sources or sinks.

divergence is everywhere zero is said to be solenoidal.
whose
Any
(4-12-1)
is
an integral relation involving J for a finite region.
Equation
Equation (1) is a differential relation involving J at a point. Both equations are
expressions of the continuous nature of J. Both are statements of Kirchhoff's

This

is

a point relation.

vector function

current law.

Let us digress briefly to consider the situation

assumed above.

Then

if

the current

and the

(4-12-1) does not necessarily hold,

is

not steady as

difference

between

the total current flowing out of and into a volume must equal the rate of change of
electric

charge inside the volume.

(rate of decrease of charge).

is

a net flow of current out o/the volume

must equal the negative

(positive current flow)

p Av, where p

Specifically,

Now

the total charge in the volume

the average charge density.

J ds =

4>

rate of change of charge with time

Av of

Fig. 4-1

is

Therefore

dp Av

Js

(2)

dt

Ay approaches

imit as

dp
(3)

This

is

the general continuity relation^ between current density

density p at a point.

Consider

For steady currents dp/dt

now the

situation

shown

terminates inside a small volume Av.

J over the volume

and

(1).

where a wire carrying a current /

Applying (2) to this situation, the integral of
in Fig. 4-12,

yields the net current entering or leaving the volume.

that / is entering the volume,

Assuming

we have

J-ds= -I
Since p

J and the charge

(3) reduces to

Av equals the totaP charge

Q inside the

(4)

volume,

dp

dQ

dt

dt

(5)

t Also called the equation of conservation of charge.

THE STEADY ELECTRIC CURRENT

SEC. 4-14]

and

Substituting (4)

127

(5) in (2) yields

'=?
This

is

CURRENT AND FIELD AT A

CONDUCTOR-INSULATOR BOUNDARY

4-14

The

the continuity relation between the current and charge in a wire.

between the current density J and the

relation

ductor

is

oE.

Consider

now

electric field intensity

in a con-

the situation at a conductor-insulator boundary as

Assuming that the conductivity of the insulator is zero, J = in the

At the boundary, current in the conductor must flow tangentially to the
boundary surface. Thus, on the conductor side of the boundary we have
in Fig. 4-13.

insulator.

(1)

= component of electric field tangential to boundary = |E|

= component of current density tangential to boundary = J
a = conductivity of conducting medium

where E,
J,

Insulator

FIGURE

4-13

Insulator-conductor boundary.

By

the continuity of the tangential electric field at a boundary, the tangential field on

the insulator side of the boundary

When
body as

is

also E,

current flows, a conductor of finite conductivity

it is

in the static case with

no currents present.

is

not an equipotential

For example, the potential

varies along a current-carrying wire with uniform current density as suggested in

Fig. 4-14.

The arrows

indicate the field

and current

directions, while the transverse

^Wire

FIGURE

4-14

Section of long wire.

(('<
"""

^ rnninrttontiale
Equipotentials

128

ELECTROMAGNETICS

[CHAP. 4

lines are equipotentials.

separated by a distance

IR drop,

to the

of a length
is

that

Since
/

V=

is,

E is

uniform, the potential difference

along the wire

of the wire.

IR, where lis the current in the wire and

The

field is

of two points

This potential difference

is El.

is

also equal

R is the resistance

the same both inside and outside the wire and

entirely tangential (and parallel to the axis of the wire).

superimposed on

If

this situation there is a static electric

charge distribution at

the boundary surface due to the proximity of other conductors at a different potential,

a component of the electric

is

E normal to the conductor-insulator boundary also

field

present on the insulator side of the boundary.

then the vector

The

sum of the normal component E and

In the conductor, E

0,

and the

shown

through the outer conductor.

this field is relatively

a relatively strong

field is

En

is large,

weak, as suggested by the short arrows

In the insulating space between the inner and outer conductors there

field

due to the voltage applied

static electric field.

It is relatively

It is

entirely

strong, as suggested

at the

two components E and E, added

surface

is

and equal to E

high, E,
.

end of the cable.

normal to the surfaces and

by the long arrows for E

the surface of the inner conductor (Fig. 4-15) the total field

the metal in the cable

For

through a part of

in Fig. 4-15

The field in the conductor is entirely tangential (and

Since the conductivity of
and is indicated by E,

parallel to the axis of the cable)

exist

is

Current flows to the right in the inner conductor and returns

a long coaxial cable.

for

the boundary.

field is entirely tangential to

instance, consider the longitudinal cross section

the conductor

total field in the insulator

the tangential component E,

vectorially as in Fig. 4-16.

may

be so small that

However, the

size

E is

is

is

may
This

indicated by

At a point

then the

at

sum of the

If the conductivity of

substantially

normal to the

of E, has been exaggerated in Fig. 4-16

'Equipotentials

FIGURE

4-15

Longitudinal cross section of coaxial transmission

line.

Equipotentials are shown

by the dashed lines. The arrows indicate the direction of the normal and tangential field components, E and E and the current density J.

THE STEADY ELECTRIC CURRENT

sec. 4-15]

129

Insulator

~Et '-.

FIGURE

Conductor-.

4-16

E at insulator-conductor boundary (point P in Fig. 4-15) resolved into

normal and tangential components. The tangential component E, is due to the
current in the conductor (Jt /a) while the normal component E is due to the
surface charge induced by the voltage between the inner and outer conductors.
Total

in order to

field

show the

slant of the total field

lines across the entire insulating

more

clearly.

The shape of the

total field

space between the inner and outer conductors

suggested in Fig. 4-17, with the slant of the

field at the

is

conductors greatly exaggerated.

Equipotential surfaces are indicated by the dashed lines.

Equi- -^
potentials

Outer
conductor

s- Inner

_ conductor

Field lines

FIGURE

4-17

Longitudinal cross section of coaxial

transmission line showing equipotentials

(dashed) and total

4-15

field lines (solid).

CURRENT AND FIELD AT A

CONDUCTOR-CONDUCTOR BOUNDARY

"Consider the conductor-conductor boundary shown in Fig. 4-18 between two media

of constants a u

e,

flowing from one

and a 2

e2

medium to the

In general, the direction of the current changes in

other.

130

ELECTROMAGNETICS

[CHAP. 4

Current or
field line

Normal

FIGURE

4-18

Boundary between two different conducting media showing change in direction

of current or

field line.

For steady currents we have the boundary

relation
(1)

where Jal
J 2

From

= component
= component

of current density normal to boundary in

of current density normal to boundary in

relation (1) of Table 3-2

we

also

medium
medium

have

E n = E t2
where

It

E tl = component
E t2 = component

of
of

field

tangent to boundary in

field

tangent to boundary in

(2)

medium 1
medium 2

follows that

(3)
o-j

where Jn
J, 2

Dividing

= component
= component
(3)

by

a2

of current density tangent to boundary in

of current density tangent to boundary in

medium
medium

(1) gives
Jti

Jti
(4)

<T l

J 1

<T 2

J2

tan a
or

tan a 2

where a t and a 2 are as shown

Chap. 12).

refraction (see

in Fig. 4-18.

(5)

a2
This relation

is

similar to Snell's

law of

THE STEADY ELECTRIC CURRENT

sec. 4-16]

131

CURRENT MAPPING AND THE RESISTANCE OF

SIMPLE GEOMETRIES; CONDUCTOR CELLS

4-16

If the current density

is

uniform throughout a conductor,

its

resistance

is

easily

For example, consider the homocalculated from its dimensions and conductivity.
It has /' = 1 m,
shown
in Fig. 4-19.
a
geneous rectangular bar of conductivity
against high=
bar
are
clamped
=
the
faces
of
400 mm, and d 400 mm. If the end
b'
conductivity blocks, as in Fig. 4-3, the field and current density throughout the bar
will

be uniform.

The

resistance

R of the bar is
/'

odb'

0.1 6ct

R=
where a

is

The

()

mhos

the conductivity of the bar in

(1)

per meter.

resistance of the bar can also be calculated

by dividing the

into square areas each representing the end surface of a conductor

the cells are equipotentials.

The

to the current direction.

The top and bottom

resistance

*o =
where

S is the

d equals the

resistivity

resistivity

side of the bar

The

sides

of

surfaces of the cell are parallel

of such a

cell is

given by

1__S

(la)

ad

"odb

of the bar material.

S of the

cell.

Hence the product of R and the depth

bar material, or

d=S

(2)

Current
tube

Cell

FIGURE

4-19

Conductor divided into current tubes.

Vertical lines are equipotentials.

132

ELECTROMAGNETICS

[CHAP. 4

Taking the reciprocal of

(2) yields

Go

That

is,

the conductance per unit depth of a conductor

of the medium.

Then

Referring to Fig. 4-19,

the total conductance of the bar

Number

of

cells (or

Number
From

of a

cell

total

conductance of the bar

is

current tubes) in parallel

of

total resistance

is

da

is

cell is a,

= 0.4<x

(13)

(4)

then

=ok = 625S

The method of calculating the

(13)

in arriving at (1) since

nonuniform.

conductor

cells will

&

(fi)

resistance or conductance of the bar

it

is

more

by means of

general than the

can be applied not only to uniform current

more

distributions (as here) but also to the

is

so that the actual

given by

evaluating the series-parallel combination of conductor cells

bution

the reciprocal, or

method used

13.

10

cells in series

G = -rV? = tt>0-4<t = 0. 1 6<7

The

equal to the conductivity

is

of bar material

G =
The

cell is

the conductance of each cell equal

conductance per unit depth of a conductor

(3) the

conductance

let

general situation where the current distri-

In a nonuniform distribution the sides of

be curvilinear squares.

many

or

all

Their area and arrangement

determined by graphical current-mapping techniques that are

procedures discussed in Sec. 3-22.

of the

may

be

like the field-mapping

Graphical current-mapping techniques can be applied to any two-dimensional

i.e., to a conductor whose shape can be described by a single cross section

problem,
with
is

all

other cross sections parallel to this one being identical to

actually electric field

field

mapping

have the same direction

in

it.
Current mapping
a conducting medium since the current and the

in isotropic

The following fundamental

media (J

<tE).

properties are useful in current mapping:

and equipotentials

Current

Current flows tangeritially to an insulating boundary.

The

lines

total current

intersect at right angles.

through any cross section of a continuous current tube

is

constant.

In a uniform current distribution the potential varies linearly with distance.

Current tubes are continuous.

THE STEADY ELECTRIC CURRENT

SEC. 4-17]

With these properties

mind one

in

divides a conductor cross section into current

tubes and then by equipotentials into conductor

same
mapping in an

cells

with sides that are squares or

method described
medium.

curvilinear squares, using the

trial-and-error

connection with

insulating

field

133

in Sec. 3-22 in

In calculating the conductance of a conductqr with a nonuniform current

map

distribution a current

is first

The conductance

made, as discussed above.

is

then given by

G = -G

(6)

where

= number of cells (or current tubes) in parallel

= number of cells in series = number of cells per
G = conductance of each cell as given by (4)
TV

The accuracy of the conductance (or its reciprocal,

on the accuracy with which the curvilinear squares

EXAMPLE A

the resistance) depends primarily

are

mapped.

homogeneous rectangular bar of conductivity a has the dimensions shown

is identical with the one of Fig. 4-19 except that two cuts have been

This bar

in Fig. 4- 20a.

made

across the full width of the bar, as indicated.

Find the resistance of the bar when

ends are clamped between high-conductivity blocks as in Fig.

solution

map made

tube

longitudinal cross section of the bar

with the result shown in Fig. 4-206. t

is

drawn

to scale

of one conductor

and there are 4 tubes

cell is

Thus, comparing this result with

4-17

From

(5) for

There are 13

(6) the resistance

(4)

or (la) the

cells in series in

of the bar

a tube,

is

the uniform bar, the slots in the bar produce an

resistance.

LAPLACE'S EQUATION FOR CONDUCTING MEDIA

For steady currents

get ctV (VF) =

we

its

1/0. 4er (fi).

Hence, from

in parallel.

increase of 30 percent in

R =

and a current

portion of one quadrant has been

further subdivided to test the accuracy of the curvilinear squares.

resistance

its

4-3.

V J=

0,

but J

<tE; so

a\ E =

0.

Recalling that

E = - VK,

or

V2 V =
t Although the entire cross section of the bar has been
that a map of only one quadrant is required.

(1)

mapped, the symmetry

is

such

134

ELECTROMAGNETICS

[CHAP. 4

400 mm

^Curvilinear

if

vz^

Equipotentials

Conducting bar with notch and

(b) current

FIGURE
(a)

c ell

(*)

4-20

map.

ratio of cells in series to cells in parallel multiplied

This

Laplace's equation.

is

fields,

and since

it

It

was derived previously

also applies here,

it

Resistance of bar equals

by the resistance of each cell.

in Sec. 3-27 for static electric

follows that problems involving distributions

of steady currents in conducting media can be handled in the same

involving static field distributions in insulating media.

an unknown current
found that also

and

distribution,

satisfies

if

way as problems
we have a conductor with

a solution to Laplace's equation can be

the boundary conditions,

rent distribution in the conductor.

If

If this

is

we can obtain the potential and

we can nevertheless

not possible,

curfind

the approximate potential and current distribution in two-dimensional problems by

mapping as discussed in Sec. 4-16. From a knowledge of the current

we can determine the resistance, the maximum current density, and other

graphical current
distribution

items of practical importance for a given conductor configuration.

In conducting media, current tubes and the conductivity a are analogous to the
flux tubes

and permittivity

e in insulating

= oE

media.

(Am

Thus

in conducting

media we have

-2
)

(2)

the steady electric current

problems]

while in insulating media

we have

(Cm" 2 )

D = eE
It is

medium

also to be noted that in a conducting

^=<r
rfis

the conductance per unit depth of

(Urn'

(3)

medium, or

a conductor cell equals the conductivity of the

where

(4)

the depth of the cell (see Fig. 4-19), while in an insulating

tance per unit depth of a dielectric field

d =
For a

is

there

is

(Fm"

(5)

medium of

For a

permittivity e there are

static electric field in

current of density J

analogous dielectric situation, and vice versa.

1

and 2

in Fig. 3-19 is a

per unit depth between ff and gg

is

For example,

also a solution for the

if

the

conductor of conductivity

medium between

a, the

conductance

given by

= 1^ a =
%
4
d

is

no

a conducting

Since both fields obey

oE.

Laplace's equation, a solution in the conductor situation

conductors

medium the capaci-

equals the permittivity e of the medium, or

D = eE.

a flux density

medium of conductivity a

cell

a dielectric

static electric field in

currents, but there

135

3.86<x

further discussion of Laplace's equation

is

CUm"

given in Chap.

7.

PROBLEMS
*

4-1

between

in length has 50 mV applied

80
in cross section by 2
Find the following quantities and give units in each case: (a) resistance

ends.

field

E, (f) power loss P in the

4-2

Repeat Prob.
If

of the bar,

(c)

current

/,

(d) current density /, (e) electric

power loss per unit volume, (h) energy loss

" 3
Take T= 20C and p = 2GC m

bar, (g)

velocity v of electrons.

4-3

mm

A copper bar 20 by

of the bar, (b) conductance

(i) drift

its

4-1 for

= x3_yz A m "

W per hour,

2
,

T= 100C.

find the current / through a square 2

m on a

side with

one

and other corners at (0,2,0), (0,0,2), and (0,2,2).

20-mm-thick L-shaped copper block has the dimensions

corner at the origin

shown in Fig. P4-4.

4-4 A
At 20C find the resistance R and conductance G (a) between faces 1 and 2, (6) between sides
(d) If 1 mV is applied
3 and 4, and (c) between front and back surfaces (parallel to page),
at
the points Pi and P 2
field
E
electric
and
density
J
current
the
between faces 1 and 2, find
The point Pi is 10 percent of the distance along the diagonal from the inside corner, and P z

136

ELECTROMAGNETICS

[CHAP. 4

110mm

40mm
70mm

110mm

70mm

FIGURE

P4-4

40mm

L-shaped block.

10 percent of the distance from the outside corner,

is

density J at

4-5

If the

L-shaped block of

P4-4

Fig.

is

is

80

mm long,

What

(a)

is

and the

is

the ratio of the current

form a

straight bar

20 by 40

mm in cross

the resistance between the ends of this bar ?

the ratio of the resistance between the faces

4-4]

What

cut along the diagonal at the bend and one

half turned over, the two halves can be joined to

section by

(e)

P2 ?

to that at

and 2 of the L-shaped bar

resistance of the straight bar [part (a) of this problem] ?

(c)

(b)

What

[part (a) of Prob.

Why

is this

ratio

not unity?

4-6

Two aluminum

bars 20 by 20

by being overlapped 20 mm.

combination

map

(field

At 20C

required) ?

20

mm in cross section by

as

two bars of (a) placed end

mm

(a)

(b)

in cross section

what

What

is

is

by 100

mm long

are joined

the resistance between the ends of this

the resistance of a continuous bar 20 by

mm long [same length as overlapped bars in (a)] ? (c) What is

the resistance of a continuous bar 20 by 20 mm in cross section by 200 mm long [same length
than in

80

to end] ?

(rf)

Why is resistance in (a) greater than in (b) but less

(c) ?

4-7

and three

batteries are connected in series as shown in Fig.

emf *Ui = 1.5 V and the electrolyte or internal resistance
Ri = 1 fi, for the second battery the emf HJj = 2 V and the internal resistance R 2 = 0, and
for the third battery the emf 1J 3 = 3 V and the internal resistance R 3 = 1 Q.
The first two

P4-7.

(a)

For the

batteries

resistance

battery the

first

have single

'

while the third has three cells in series, each cell of 1 V emf and
Assume that half the total emf of a cell occurs at each electrode, and

cells,

i CI internal resistance.

Answers to starred problems are given

in

Appendix C.

THE STEADY ELECTRIC CURRENT

problems]

-wvwvw-

%=2

n>,=i.5

<U,=3

ll HhhN

137

ll

ft,

FIGURE

P4-7

Series circuit.

connections between cells have negligible resistance. Draw a graph such

showing the variation of potential with position between points a and c when
and also when R = 0. Take V= at the point b. (6) Referring to the circuit

assume that

all

as in Fig. 4-7,

Ro

= 4.5 H

of Fig. P4-7,

let

the emfs be as indicated, and

let

Ri

4-8
is

Four wires meet

4-9

it

common

The current

junction point.

in wires

each and flowing away from the junction, while the current in wire 3

toward the junction.

at

.5 Q, R 2 = 2 Q, and /? 3 = 3 il.
Draw
R = 6.5 O and also when R = 0.

a graph of the variation of potential with position when

which

is

What

are the current magnitude

and direction

A nichrome bar 40 by 40 mm in cross section by

is

the resistance between the ends of the bar ?

solid bar (without hole) ?

(c)

What

is

and 2

A flowing

in the fourth

wire?

mm long has a hole through

The hole 30 mm in diameter.

100

symmetrically situated at the center of a long side.

At 20C (a) what

is

is

(6)

What is the resistance of a

the length of a solid bar having the same resistance

as in (a) ?

4-10

bar and strip are connected as shown in Fig. P4-10d.

conductivity, while the strip conductivity

clamped against a

large, infinitely

is

assumed to be

its

resistance to be the

4-11
strip

Why

is

same as when connected

is

made

finite

end of the bar

is

the bar would need to be lengthened

to the strip.

the resistance of the bar of Fig. P4-10 larger

than when contact

The bar has

If the

conducting block as in Fig. P4-106 instead of connected

to the strip as in Fig. P4-10a, determine by what length

for

infinite.

when

it is

connected to the

with the block ?

mm
Bar

x:-!

mm-'

'

Strip

(a)

Block position
equivalent to strip

(6)

FIGURE
Bar and

P4-10

strip.

[CHAP. 4

ELECTROMAGNETICS

138

Deduce the

4-12

relation

by applying the divergence theorem to

Show

4-13
relation /

*
10

rfs

=
by Eq.

(4-3-1) leads to the continuity

Q = positive charge.

where

Demonstrate that the source of the emf energizing the coaxial line of Fig. 4-17
end by showing that if the source were at the right end, the field lines would be

at the left

bowed

that the definition of current given

= dQ/dt,

4-14
is

in the opposite direction.

4-15

Show

4-16

(a)

that at a conductor-conductor

A wire 2 mm in

boundary oi/a 2

diameter has a resistance of

1 ii

= E

A is flowing in the wire. What is the electric field E in the wire?

surface charge of 8

pC

m~

is

applied to the wire, what

direction) just outside the surface of the wire?

The

4-17

current direction at the flat

angle of 45 with respect to the surface in

is

(b) If

the electric field

The medium

= J,ilJ, z
A current

2 /Ei

per 100 m.

a uniform

of

static

E (magnitude and

outside the wire

is air.

boundary surface between two media makes an

medium

1.

What

is

the angle between the current

and the surface in medium 2? The constants for the media are a = 1 MO m
= e in medium 1 and a = 1 ku m _1 and e = 3 in medium 2.
~2
what is the surface charge
4-18 If the cujy ent density J in medium 1 is 20 A m
-'

direction

and

boundary

density at the

Two

4-19

centers.

is

Prob. 4-17?

long parallel zinc-plated iron pipelines have a spacing of 4

pipes are half buried in the ground as indicated in Fig. P4-19.

The

of the pipes

in

500

mm.

The conductivity of

the ground (sandy

soil) is

100 fj.13

between

The diameter

m _1

With-

out drawing a field map, find the resistance between the two pipes per meter of length. Hint:
Note the analogy between this situation and the static electric field between two parallel
cylindrical conductors.

4m
-Pipe lines

FIGURE

J
-~-----_^7^

500rnm

P4-19

Pipelines half buried in the ground.

If the
washer of thickness / has an inner radius n and outer radius r 2
determine the resistance (a) between the inner and outer edges, (6) between
the flat surfaces, and (c) around the washer (same as resistance between the edges of an
infinitesimally thick saw cut through one radius of the washer).

4-20

conductivity

is

flat

a,

For an applied emf "V determine the current density J as a function of radius
and (c) of Prob. 4-20.
4-22 A flat washer of thickness t has an inner radius ri and a square outer edge of
length / = 3n. If the conductivity is a, determine the resistance (a) between inner and
4-21

for cases (a)

side

flat surfaces, and (c) around the washer (same as resistance between the edges of an infinitesimally thick saw cut through one radius of the washer).

outer edges, (b) between the

In general, a surface charge

4-23

(conductivities

<ji

Show

flowing.

and

and

<r 2

is

present on the boundary between two conductors

permittivities ei

and

that the surface charge density p s

\cri

where

is

4-24

C m~ 3

(a)

At 30C,

intrinsic

'

_l

Find the current

(a) if the wire is

in

-i

V"

velocities

If the electron

and of the
a wire of

what

is

(6) If

and hole charge

an

electric field

which current

m~ *.

is

densities for the

two

is

both

applied

holes.

mm

cross section with an applied field of 0.1

(c) If

wires.

(6) if the

the electron mobility for copper

the charge density for the copper wire?

and charge

densities are
_1

of 10

of germanium with properties as in Prob. 4-24 and

copper with conductivity of 57 MJ5

s

respectively) across

semiconductor germanium has an electron mobility of 0.4

s" 1 V

find the velocity of the electrons

m~

given by

<JzJ

find the conductivity of the material.

4-25

e2

is

the normal component of current density.

and hole mobility of 0.2

4

139

the steady electric current

problems]

(d)

Compare

is

wire

is

0.003

V
of

the electron drift

THE STATIC MAGNETIC FIELD OF STEADY

ELECTRIC CURRENTS

INTRODUCTION; EFFECT OF A CURRENT ON A MAGNET

5-1

static electric

electric current,

charge has an electric

field,

as discussed in Chaps. 2

on the other hand, possesses a magnetic

carrying a current / has a magnetic field surrounding

When
field

this field

is

field.

it,

For

and

3.

An

instance, a wire

as suggested in Fig. 5-la.

explored with a pivoted magnet or compass needle, the magnetic

produces an aligning force or torque on the needle such that the needle always

normal to a radial line originating at the center of the wire. This

parallel to the magnetic field.
If one moves in the direction of the
found that the magnetic field forms closed circular loops around the wire.

orients itself

orientation
needle,

it is

is

The

direction of the magnetic field is taken to be the direction indicated as

" north " (N) by the compass needle, as in Fig. 5-\b. The relation of the magnetic field
direction to the current direction can be easily

remembered by means of the right-hand

With the thumb pointing in the direction of the current, as in Fig. 5-2, the
fingers of the right hand encircling the wire point in the direction of the magnetic field
rule.

or lines of magnetic flux.

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-2]

FIGURE
(a)

5-1

Magnetic

a current.

around wire carrying

Cross section perpen-

field

(6)

Compass
needles

The current is flow-

dicular to the wire.

(b)

(a)

ing out of the page.

5-2

141

EFFECT OF A MAGNET ON A
CURRENT-CARRYING WIRE

Consider the uniform magnetic

field between the poles of a permanent magnet as in

compass needle may be used to explore this field as suggested. If a
conductor with current / is placed in the field, it will be acted on by a force

Fig. 5-3. f

straight

F=BIL

or

F = LIB

(N)

(1)

where L = length of conductor in magnetic field, m

/ = current in conductor, A
B = factor involving the magnetic field

Field or
flux line

FIGURE

5-2

Right-hand rule relating direction of

field or flux lines (fingers) to direction of
current / (thumb).

Due to fringing effects the field will not be uniform at the edges. However, if the
pole cross section is large compared to the gap spacing, the effects will be small.

142

ELECTROMAGNETICS

[CHAP. 5

/Compass needle

%b

OSZ
p-i

Pole faces

FIGURE

5-3

Linear conductor of length L and current

uniform magnetic field between pole

There is a force on
the conductor normal to the page
I'm.

pieces of a magnet.

jPermanent magnet!

Vj

(inward).

In Fig. 5-3, the force

B=
Thus,

is

in a direction perpendicular to the

F
IL

force

current

(NA-'nT

moment

B can be described as a force per current moment.

B,

From

or T)

(1)

(2)

However, as will be discussed

customary to call it the magnetic flux density.

-2
square meter (Wb m ), or tesla (T) (1 Wb m~ 2 = 1 T).
later,

page (inward).

it is

The

unit

is

the weber per

In Fig. 5-3, 1, B, and F are mutually perpendicular. If I is not perpendicular to

we find that F is a function of $ (see Fig. 5-4). In general, for any elemental current

element

dF=BIdl sin
This equation and

F = BIL are

(3)

</>

the basic motor equations.

B
>-

FIGURE

CT

5-4

The force on a current element is normal

to the plane containing the element

B.

and

^Current
element

dF(

inward)

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-3J

5-3

143

THE MAGNETIC FIELD OF A CURRENT-CARRYING

ELEMENT; THE BIOT-SAVART LAW

Let the aligning torque on an arbitrarily small perfectly mounted magnetic needle be
used to measure the field B produced by an incremental current-carrying element of
length A/, as in Fig. 5-5.

incremental

is

From

a function of

/,

these measurements

A/,

r,

and

AB =
where

A: is

9, as piven

we

find (for r

>

A/) that the

by

IAI sin 6

(1)

a constant of proportionality given by

(2)

4n

By dimensional analysis of (1) we find

/i is the permeability of the medium.
n has the dimensions of flux per current divided by length. It will be shown in

where
that

Sec. 5-12 that inductance has the

dimensions of flux per current.

meability has the dimensions of inductance divided by length.

Therefore per-

The SI

unit for

Magnetic needle
to page)

FIGURE

5-5

B produced by short current-carrying element A/ as

a function of radius r, angle 0, current /, and length A/. B is determined by
measuring the aligning torque on an arbitrarily small magnetic needle (normal
to page) at P. The wires supplying current to the element A / are always arranged
so that they are radial with respect to P, and thus they do not affect B at P.

Arrangement for measuring

144

[CHAP. 5

ELECTROMAGNETICS

FIGURE

5-6

Construction for calculating flux density

B at a point P due to a current / in a long
conductor.

permeability

is

The

the henry per meter.t

= 4tc

^
Introducing (2) in

more fundamental

(1)

x 1(T 7

and writing

permeability of

vacuum

is

H m" = 400ti nH m _1
1

infinitesimals instead

of incrementals, we obtain the

relation

dB =

H Idl sin

(3)

4tc

The

dB is normal to the page (inward at P).

we wish to know B at a point P, as in Fig.

direction of

In case

5-6,

due to a current / in a
we assume that

long, straight or curved conductor contained in the plane of the page,

the conductor
in series.

from

all

is

The

made up of elements

total flux density

these elements

and

is

or segments of infinitesimal length dl connected

B at the point P is

then the

expressed by the integral of

al rsin

sum of the

(3).

contributions

Thus

(4)

t Recall that permittivity e has the dimensions of capacitance per length and
pressed in farads per meter. Note that
l/4*r

Ho/4it

where c

is

_ j2

/*o o

the velocity of light or radio waves equal to 300

Mm s -1

is

ex-

sec. 5-4]

where

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

B = flux density at P, N A~ m" or T

_1
H = permeability of medium, H m
/ = current in conductor, A
dl = length of current element, m
r = distance from- element to P, m
9 = angle measured clockwise from positive
'

'

direction of radius vector r extending

The

integration

is

145

<!>
direction of current along dl to

from

dl to

carried out over the length of the conductor.

Equations

(3)

and

(4)

are statements of the Biot-Savart law.

5-4

THE MAGNETIC FIELD OF AN

INFINITE LINEAR CONDUCTOR
B

/*

'<

The magnetic

field

infinite length

with a current / can be obtained from (5-3-4).

right

dl sin

or flux density

The geometry

is

of the wire

into the page.

=r

is

dO and

shown

R = r sin

This

thin linear conductor of

as indicated,

Since

case becomes

A11R

and

it,

that

is,

over the entire

Integrating gives

B = ~[-cos
4nR

0]l

= ^-2
4nR

=0

FIGURE

at the

according to the right-hand rule.

^f -de = ^-fsinede

between the angles

length of an infinite wire.

from a

With the current /

is

6, (5-3-4) in this

47cv
is

a radius

in Fig. 5-7.

B=
where the integration

at

5-7

Construction for finding flux density

near a long straight wire.

inward

146 ELECTROMAGNETICS

[CHAP. 5

B=

or

THE FORCE BETWEEN TWO PARALLEL LINEAR

CONDUCTORS; DEFINITION OF THE AMPERE

Consider a length
as in Fig. 5-8.
/' in

(1)

B = flux density, N A -1 m _1 or T
fi = permeability of medium, H m~
/ = current in conductor, A
R = radial distance, m

where

5-5

JlJ_

2nR

of two long parallel linear conductors in air spaced a distance

Assume

that conductor

the opposite direction.

conductor
to the

were in

2.

It

The

follows that there

carries a current / and conductor 2 a current

due to conductor 1 are into the page at

a force to the right on conductor 2 and a force

flux lines
is

on conductor 1. That is, the conductors are repelled. If the currents

the same direction, the forces would be reversed and the conductors would be

left

attracted.

The magnitude

F of the force on a

length

F=I'b[

FIGURE

dl

of conductor 2

= I'Bl

5-8

Force between two long parallel conductors in air.

is

0)

7'

"/'

THE STATIC ELECTRIC FIELD OF STEADY MAGNETIC CURRENTS

sec. 5-6]

= current in conductor 2
=
B flux density at conductor 2 produced
Introducing the value of B from (5-4-1) gives
where

/'

by current / in conductor

(2)

2nR
where

F = force on length

of conductor

2,

/= current in conductor 1, A
/' = current in conductor 2, A
R = separation of conductors, m
H = permeability of vacuum or air =
Since (2)

is

symmetrical in / and

same magnitude

as the force

Dividing (2) by

F on

/',

nHm"

on a length

x 10

of conductor

is

of the

on

either conductor as

ir
(3)

2nR

and introducing the value for n

= R = m and F = 2

2.

yields the force per unit length

F=2x
If /

4007t

conductor

= /,

the force

f _n

If/'

147

1(T 7

N, then / =

(N)

R
1

A.

This measurement

(4)

is

used to define

the ampere in the SI system (see Sec. 1-3).

5-6

THE MAGNETIC FIELD OF A

CURRENT-CARRYING LOOP

Let the loop be in the xy plane with

its

center at the origin, as in Fig. 5-9, so that the z

The loop has a radius R and current /. At the

the contribution dB produced by an element of length dl of

axis coincides with the loop axis.

point

the loop

on the loop
is,

from

axis

(5-3-3),

dB =
where 6
the loop

is

is

the angle between dl

in

fil

dl sin

4nr 2

(1)

and the radius vector of length r. It is assumed that

\i.
The direction of dB is normal to

a medium of uniform permeability

the radius vector of length

r,

that

is,

at

an angle ^ with respect to the loop or z

axis.

148

ELECTROMAGNETICS

[CHAP. 5

Loop

axis

Loop with
current /

FIGURE

5-9

Construction for rinding flux density

on axis of current loop.

The component dBz

in the direction of the z axis

dB = dB cos

is

given by

= dB -

(2)

From

Fig. 5-9

we

note that

these values into (1)

and

90, dl

= R d<j>,

substituting this value for

and

dB

= y/R 2 +

in (2),

2
.

Introducing

we have

UlR
4n(R 2 + z 2 )

dBz
The

total flux density

Bz

in the z direction

The element

loop.

Integrating this

(3)

then the integral of (3) around the entire

dl also produces a

filR

B = B,=

47t(R

At

is

d(l>

component of dB normal to the axis of the

component for all elements around the loop yields zero because
Hence, Bz equals the total flux density B at the point P as given by

loop.

of symmetry.

W2

the center of the loop, z

0,

2
r

2 312
)

2*

nIR 2
__
** " 2(R 2 + z 2 3 ' 2
)

and

B^

(5)

2R

where

(4)

B = flux density at center of loop, N A -1 m _1

_1
H = permeability of medium, H m
/ = current in loop, A
R = radius of loop, m

or

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

SEC. 5-7]

Normal
area

f-y

to

149

'

'

r+r

/ i\*k.F<

j,

S-7 Magnetic
or
field

flux lines

FIGURE

5-10

Flux

and area A.

lines

MAGNETIC FLUX

5-7

The magnetic

field

quantity

i//

AND MAGNETIC FLUX DENSITY

introduced above as a force per current

also be regarded as a magnetic flux density,

we

describe

moment can

and this term is in more common use. Thus,

as the magnetic flux per unit area or

B = ^f
A
In (1)

assumed that the magnetic

is

it

generally

we have

(1)

field lines are

perpendicular to A.

More

(see Fig. 5-10)

\}i

= BA

cos a

(2)

m = magnetic flux through area A

B = magnitude of magnetic flux density B
a = angle between normal to area A and direction

where

\j/

of

Dimensionally
,

force

current

moment

x area

or in dimensional symbols

MLL2
T
The SI

B is

unit

of magnetic

flux

i//

the weber per square meter

If

is

is

ML2
IT 2

IL

the weber (Wb).

(Wb m~ 2 )

The

unit of magnetic flux density

or tesla (T) (dimensional symbols

not uniform over an area, the simple product

surface integral, so that, in general,

il,

(2)

M/IT 2 ).

must be replaced by a

we have

=jJBcosctds

(3)

150

ELECTROMAGNETICS

Where ds

[CHAP. 5

element of surface area

B = magnitude of B
a

= angle

Equation

between normal to ds and direction of B

can also be written as a scalar or dot product.

Thus,

(3)

*.-JjB-*
where

m = magnetic flux, Wb
B = magnetic flux density, Wb m -2 or T or N A -1 m -1
ds = vector with direction normal to surface element ds and

\j/

MAGNETIC FLUX OVER A CLOSED SURFACE

5-8

flux tubes

of a

static electric field originate

and end on

other hand, tubes of magnetic flux are continuous;

This

sinks.

To

magnitude equal

m2

to area of ds,

The

(4)

is

a fundamental difference between

i.e.,

electric charges.

static electric

describe this continuous nature of magnetic flux tubes

density

is

solenoidal.

Since

volume as leave it.

the result must be zero, or
enter a

it is

may be regarded

and magnetic
is

the

fields.

said that the flux

flux tubes

must

carried out over a closed surface,

B-ds =

<j>

This relation

is

it

many magnetic

continuous, as

Hence, when (5-7-4)

On

they have no sources or

(Dt

as Gauss' law applied to magnetic fields [compare with

(2-17-6) for electric fields].

It

follows that the divergence of

B equals

zero.

That

is,

V-B =
Both

(1)

and

relation for a finite

(2) are expressions

volume and

Maxwell's equations
given by

(2)

of the continuous nature of B,

(2) the relation at

(in differential form).

a point.

Equation

The same equation

(1).

The symbol

f, indicates

an integral over a closed surface.

(1)

being the

(2) is

in integral

one of
form is

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

SEC. 5-9]

5-9

MAGNETIC FIELD RELATIONS IN VECTOR NOTATION

A linear current-carrying conductor placed

force

151

Fona.

length

of conductor that

F = IB sin
where

<()

is

in

a uniform magnetic

given,

dl

from

(5-2-3),

= IBL sin

field experiences

by

(1)

(j>

F = force, N
/ = current in conductor, A
B = flux density of field, T
L = length of conductor, m
= angle between I and B
<j>

Equation

(1) is

involved.

The

a scalar equation and relates only the magnitudes of the quantities

force F is perpendicular to both I and B. For example, let the

conductor be normal to a uniform magnetic field of flux density B as in Fig. 5-1 la.
lines, as
If the current in the conductor is flowing out of the page, it produces flux
indicated, so that the flux density

The

resulting force

is

is

increased below the wire and weakened above.

therefore upward, as suggested in Fig. 5-116.

(a)

(b)

FIGURE

5-11

Force F on current-carrying conductor

in uniform magnetic field.

152

ELECTROMAGNETICS

[CHAP. 5

*
^Conductor of
length

(a)

FIGURE

(b)

5-12

Relation between current direction

I, field

and force F.

direction B,

we have F in the
when I is in the positive x direction and B in the positive y direction.
of I is not perpendicular to the direction of B but is as shown in Fig.

Relating the directions to the coordinate axes as in Fig. 5- 12a,

positive z direction
If the direction

5-126, the force

where

<p

direction of

force

is

is still

in the positive z direction with a

magnitude given by

(1),

equals the angle measured from the positive direction of I to the positive

(counterclockwise in Fig. 5-126).

in the positive z direction if sin

A more concise method

is

</>

With

<j>

measured

is

by means of the

way, the

in this

positive.

of expressing the relation

vector, or

cross product.

The cross product of two vectors

is

defined as a third vector whose magnitude

is

equal to the product of the vector magnitudes and the sine of the angle between them.

The direction of the third vector

is

perpendicular to the plane of the two vectors and

such a sense that the three vectors form a right-handed

in

set.

Using the vector product, we can state the relation as

F=
For an elemental length of conductor

this

d =
where d

= vector

= vector
dl = length
B = vector

(I

indicating magnitude

ductor,
I

(I

x B)L

(2)

becomesf

x B)

and

dl

(4)

direction of force

on element of con-

indicating magnitude

of conductor,

and

direction of current in conductor,

and

direction of the flux density,

indicating magnitude

t For a volume distribution of current

rfF

where

rfF is the force

we have

= (JXB)d

on the volume element do

at which the current density

(3)
is

J.

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-9]

FIGURE

153

5-13

Relation for finding

B at

a point

P due

to a current /in a conductor of anyshape.

Equation

and the

(4)

combines

in

one expression the relations between both the magnitudes

directions of the quantities involved, whereas (1) related only the magnitudes.

Equation (5-3-4) gives the magnitude of B at a point as produced by a current /

conductor contained in a single plane.

in a straight or curved

A more general relation

applying to a conducting wire of any shape, as in Fig. 5-13, can be stated with the aid

of the vector product:

-J
where

Ixf
dl

B = flux density at P, T
_1
H = permeability of medium, H m
I = current in conductor (vector pointing
element

dl),

(5)

in positive direction of current at

unit vector pointing from element dl to point P,

= distance from dl to P, m
dl = infinitesimal element of length of conductor, m
t

dimensionless

The

integration in (5)
If the current

is

is

carried out over the length of conductor under consideration.

distributed throughout a volume, the flux density

is

given by

Jxf

-sill
where J
(5),

and

is

dv

the current density in a volume element dv at a distance

(6) are basic magnetic field relations.

(6)

r.

Equations

(4),

'

154

ELECTROMAGNETICS

[CHAP. 5

fefr
Axis of
rotation

(a)

(6)

FIGURE

5-14

Rectangular loop in

field

of uniform flux density B.

TORQUE ON A LOOP; MAGNETIC MOMENT

5-10

When

a current loop

tend to rotate

it.

is

placed parallel to a magnetic

The

field,

forces act

on the loop that

tangential force times the radial distance at which

it

acts

moment, on the loop. Torque (or mechanical

moment) has the dimensions of force times distance and is expressed in newton-meters
(N m).
Consider the rectangular loop shown in Fig. 5-14a with sides of length / and d
situated in a magnetic field of uniform flux density B.
The loop has a steady current /.
According to (5-9-4), the force on any element of the loop is
is

called the torque, or mechanical

d = {lxV)dl
If the plane

of the loop

is

an angle

at

/?

with respect to B, as indicated in the cross-

sectional view of Fig. 5-146, the tangential force

F,=
The

total torque

|F| cos

on the loop

is

is

dl

= IBl cos P

(2)

then

T=
But Id equals the area

= IB cos P

J?

(1)

2F,-

= IBld cos

(3)

of the loop; so

T =IAB cos p
According to

(4),

the torque

is

proportional to the current in the loop, to

to the flux density of the field in which the loop

The product IA

(4)

is

its

in (4) has the dimensions of current times area

magnetic moment of the loop.

Its

area,

and

situated.

dimensional symbols are II?, and

it is

and

is

the

expressed in

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

SEC. 5-11]

m 2 Let
T=mB cos

amperes times square meters (A

letter

Thenf m = IA,

m.

and

).

moment by

us designate magnetic

fi,

155

the

or

T= mB siny
where y

is

(5)

the angle between the normal to the plane of the loop and the direction of B
If the

(see Fig. 5-14*).

If the

loop has

magnetic moment

plane of the loop and with

its

is

N turns, the magnetic moment m = NIA.

regarded as a vector

m with direction n normal to the

positive sense determined

in direction of current, thumb in direction of

ft),

by the right-hand rule

(fingers

the torque relation of (5) can be

Thus

expressed in a more general form using the vector product.

T=mx B
T = torque on loop, N m
m = nw = &IA = magnetic moment

where

B = flux

T is considered

the loop as given by

m into

turning

of loop,

density of field in which loop

The torque

is

A m2

stituated,

to be a vector coinciding with the axis of rotation of

The

x B.

(6)

direction of the torque

on

the loop

obtained by

THE SOLENOID

5-11

A helical coil, or solenoid, is often used to produce a magnetic field.

the flux density for such a

length

and radius

with the radius

Fig. 5-15*.

N turns of thin wire carrying a current

(Fig. 5- 15a).

of the

coil.

The spacing between

with a linear current density

is

/.

is

The

coil

has a

small compared
is

shown

in

sufficiently small, or if the wire is replaced

K = NI/l (Am -1

).

negligible spacing
in the coil

This

is

conducting sheet, as in Fig. 5-1 5d, with total current

find the flux density

turns

cross section through the solenoid

If the spacing between turns

by a thin conducting strip of width l/N and with

in Fig. 5- 15c, one may consider that the current

To

Let us calculate

coil.

Let the coil consist of

is

B.

between turns as

produces a current sheet

equivalent to a single turn of

NI(=KJ).

B at the center of the solenoid,

let

a section of the coil of

length dx, as in Fig. 5- 15c, be regarded as a single-turn loop with a current equal to

Kdx = -rdx
t Although the loop in Fig. 5-14 has a rectangular area, the relation
regardless of the shape of the loop area.

(1)

m = lA

applies

156

ELECTROMAGNETICS

[CHAP. 5

weeN turns

(a)

oooooooooo

OOOOOOOQOOO

Hh

(b)

Current sheet of total

current

=Kl=NI

*dx

(c)

FIGURE

5-15.

(a) Solenoidal coil, (A) cross section, (c) coil with strips

and

(d) equivalent current

sheet.

Top

Single
turn

V^Single

FIGURE

.-4-*

5-16

Solenoid in

field

turn

of uniform flux density

Solenoid axis

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

SEC. 5-11]

From (5-6-4) the flux density dB at the

dx at a distance x from the center

dB
The total flux

density

157

center of the solenoid due to this loop of length

is

-iWTxW2dx

(2)

B at the center of this coil is then equal to this expression integrat-

ed over the length of the

That

coil.

is,

+'/3

HMR*

dx

Performing the integration gives

BJgL*
^fAtf
If the length of the solenoid

is

much

UNI

(4)

greater than

radius

its

(/

>

R), (4) reduces to

NI

B = *r- = nK

= Kv

and

(5)

B = flux density, T
_1
fi = permeability of medium, H m
N number of turns on solenoid, dimensionless
/ = current through solenoid, A
= length of solenoid, m
K = sheet current density, Am -1

where

Equations

(4)

and

(5) give the flux

changing the limits of integration in

of the

coil

(3) to

>R

and / we obtain the

^L=

iJr 1 +

is

By

one end

(6)

this reduces to

B^
which

flux density at

(on the axis) as

B
For

density at the center of the solenoid.

= iK

(7)

one-half the value at the center of the coil as given by (5).

now calculate the maximum torque tending to rotate a solenoid placed in a

Let us

The torque is maximum when the solenoid

field of uniform flux density.
normal to the direction of B, as in Fig. 5-16. The axis of rotation is at the

magnetic
axis

is

ELECTROMAGNETICS

158

[CHAP. 5

FIGURE

5-17

Solenoid and magnetic flux

Assuming

center of the solenoid.

on a

tangential force F,

that the solenoid

single straight

F,

The

segment of

is

of square cross section, the

turn

is

given by

= IBd cos

(8)

net torque due to 2 turns, one at a distance x above the center of the solenoid and

another at an equal distance below,

is

T= 41Brd cos

then

/?.

But cos P

djlr; so

T=2Id 2 B = 2IAB
where

A = d2

is

the center of the solenoid.

equal to Eq. (9) times Nj2, where

torque

Tm

This torque

the area of the solenoid.

of the turns from

is

That

is

(9)
is

independent of the distance

Hence, the total torque on the solenoid

the number of turns.

This

is

the

maximum

is,

Tm = NIAB = m'B
where

5-12

An

lines.

rri

= NIA

is

the magnetic

moment of the

(10)

solenoid.

INDUCTORS AND INDUCTANCE

inductorf

is

a device for storing energy in a magnetic

field.

It

may be

regarded

as the magnetic counterpart of a capacitor, which stores energy in an electric

As examples, loops, coils, and solenoids are inductors.

The magnetic lines produced by a current in a solenoidal
as suggested in Fig. 5-17.
figure links the current

An

Each

N times.

line that passes

coil

field.

form closed loops,

through the entire solenoid as

If all the lines link all the turns, the total

in the

magnetic

sometimes called an inductance. However, it is usual practice to

an inductor. This makes for uniform usage when we
speak, for example, of an inductor of 1 M inductance, a capacitor of 1 pF capacitance,
or a resistor of 1 fl resistance.
inductor

is

refer to a coil or solenoid as

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-13]

flux linkage

A (capital lambda) of the coil is equal to the total magnetic flux

the coil times the

\ji

159

m through

number of turns, or
Flux linkage

By definition the

inductance

= A = N\li m

(Wb

turns)

L is the ratio of the total

magnetic flux linkage to the

current / through the inductor, or

L=
This definition
as

air.

As

W A

satisfactory for a

is

discussed in Chap.

6,

(1)

1
medium with a

constant permeability, such

however, the permeability of ferrous media is not

is defined as the ratio of the infinitesimal change

constant, and in this case the inductance

in flux linkage to the infinitesimal

change in current producing

it,

or

dA
(2)

Inductance has the dimensions of magnetic flux (linkage) divided by current.

unit of inductance

is

the henry (H).

Henrys

The dimensional symbols

5-13

The

Thus,

webers

ampere

for inductance are

ML

jI

T2

INDUCTANCE OF SIMPLE GEOMETRIES

The inductance of many inductors can be readily calculated from

their geometry.

examples, expressions for the inductance of a long solenoid, a toroid, a coaxial line,

As
and a two- wire

line will

In Sec. 5-1
less

1 it

was shown

than at the center.

However,

be derived in

this leakage

is

This

is

this section.

that the flux density

B at the end of a

long solenoid

is

caused by flux leakage near the ends of the solenoid.

mostly confined to a short distance at the ends of the solenoid,

very long, one may, to a good approximation, take B constant

so that if the solenoid is

over the entire interior of the solenoid and equal to
The total flux linkage of a long solenoid is then

HN
A = N^ m = NBA =

its

value at the center (5-11-5).

IA
(1)

160

ELECTROMAGNETICS

[CHAP. 5

Thus, the inductance of a long solenoid (see Fig. 5-17)

_A_nN

is

A
(2)

where

L=

inductance of solenoid,

A = flux
/

linkage,

Wb

turns

current through solenoid,

_1
H = permeability of medium, H m
N = number of turns on solenoid, dimensionless

A = cross-sectional area of solenoid, m 2

/ = length of solenoid, m

EXAMPLE
length of 500

Calculate the inductance of a solenoid of 2,000 turns

mm on a cylindrical paper tube 40 mm in diameter.

solution

From

L=

(2) the
4tt

inductance of the solenoid

x 10" 7 x 4 x 10 6 x

ff

x 4 x 10

If a long solenoid
is

magnetic

obtained.
lines

is

12.6

flux linkage

mH

itself,

a toroidal

the toroid has a uniform winding of

many

or

coil,

turns, the

of flux are almost entirely confined to the interior of the winding,

being substantially zero outside.

calculate

air (/*=/x

bent into a circle and closed on

When

is

is

0.5

toroid,

wound uniformly over a

The medium

If the ratio R/r (see Fig. 5-18)

is

large,

as though the toroid were straightened out into a solenoid.

one may
Thus, the

is

N\p
Ym
m

2nR

,. = HNI
= NBA
N^ nr 2i =

\iN 2 Inr 2

= nN2R
2nR

(3)

N turns

FIGURE
Toroid.

5-18

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-13]

The inductance of the

toroid

is

then

~1~
where

161

fiN

(4)

2R

L inductance of toroid, H
H = permeability (uniform and constant) of medium
JV = number of turns of toroid, dimensionless
r = radius of coil (see Fig. 5-18), m
R = radius of toroid, m

inside coil,

H m _1

Consider next a coaxial transmission line constructed of conducting cylinders of

The

radius a and b, as in Fig. 5-19.

current on the outer conductor
radius r

is

the

same

is

current on the inner conductor

The

of the same magnitude.

as at this radius

from a long

is /.

The return

flux density

B at

straight conductor with the

any

same

current, or

B (at r) =

(5)

2nr

The

total flux linkage for a length

d of line

is

then

d times

the integral of (5) from the

inner to the outer conductor, or

r
=
A = d\Bdr
J
,

dul

dr
dul
b

= ^-ln\
In
a
2n
r

Ja

Hence, the inductance of a length d of the coaxial

L=-=
I

In 2n
a

fid

line is

(#)

(7)

or the inductance per unit length (L\d) for the coaxial line

= fln
^
d 2n a

(6)

is

given by

(Hm-)

(8)

Longitudinal section

Inner

Cross

conductor

section

MMlH!lgMWMIllMttt>JlHllllli.. l.l-.^MHi:iWl -WJHJ

r-

16

FIGURE

5-19

Coaxial transmission

line.

162

ELECTROMAGNETICS

[CHAP. 5

Cross section

Longitudinal section

j^iftSsfr&fc ijtss&izmiis;

mm

-#-2<j

Sspv
1

Conductorss

\
#/&^r^!& "^/"^^ffp^ftffTi*

-B

.1..

i, ^l'.

SS*?@K i2fe

1,
1*

H
FIGUF

inward

5-20

Two-wi re transmission

line.

= permeability (uniform and constant) of medium inside coaxial line, H m -1

= inside radius of outer conductor
a = outside radius of inner conductor (in same units as b)

where n
b

It is

assumed that the currents are confined to the

the case

when

radii

a and

b.

This

is effectively

the walls of the conductors are thin.

Evaluating (8) for an

air-filled line (ji

da

-,

0.2 In

=n

we have

),

(^Hm"

0.46 log a

(9)

Let us consider finally a two-wire transmission line as illustrated in Fig. 5-20.

The conductor radius

is

a,

one of the conductors the

and the spacing between centers is D. At any radius r from

The
flux density B due to that conductor is given by (5).

due to both conductors for a length d of line

from a to D, or

total flux linkage

integral

of

(5)

A=

2d

Bdr =

D ar
dr
Hid Cr"
r

where

line is

is

given by

dim-)

= permeability (uniform and constant) of medium, H m _1

D = spacing between centers of conductors
a = radius of conductors (in same units as D)
\i

(11)

or the inductance per unit length of line (L/d)

-ln

(10)

^ln*

L=

d times twice the

Hence, the inductance of a length d of the two-conductor

then

uia
D
u
= nld
In -

Ja

is

(12)

THE STATIC ELECTRIC FIELD OF STEADY MAGNETIC CURRENTS

sec. 5-14]

It is

assumed that the current

when

is

confined to a radius

a.

This

is

163

effectively the case

the walls of the conductors are thin.

Evaluating (12) for a

medium of air

(ji

=n

da

-=

0.4 In

)>

we nave

(^Hm -1 )

- = 0.92 log -

(13)

AMPERE'S LAW AND H

5-14

According to

(5-4-1), the flux density

(Fig. 5-21)

given by

is

B at a distance R from
B=

a long straight conductor

Hi

(1)

2nR

= permeability of medium, H m~
=
/
current in wire, A

where n
If

is

now

integrated

around a path of radius

I
J

d\

= - idl =
2nR

j>B-dl

or

enclosing the wire once,

- 2nR = nl

(2)

ftI

(3)

2nR

Relation (3) holds not only in the example considered but also in
the integration

is

Equation

over a singly closed path.

of the medium by introducing the vector

FIGURE

we have

(3)

all

may be made

cases where

independent

H defined as follows:
B
(4)

5-21

Relation of flux density

B to

current

/.

164

ELECTROMAGNETICS

According to

(4),

(CHAP. 5

H and B are vectors having the same direction.

This

is

true for all

isotropic media.

The

quantity

is

called the magnetic field

H, the vector H, or simply H.

It

has

the dimensions of

The dimensional symbols

for

Flux density

current

Permeability

length

H are IjL.

In SI units

is

expressed in

Webers/meter 2

amperes

Webers/ampere-meter

meter

Introducing (4) into (3) yields

H
where

d\

(5)

H = magnetic field, Am -1
= infinitesimal element
/ = current enclosed, A

d\

This relation

known

is

of path length,

as Ampere's law.

around a single closed path

is

In words

it

states that the line integral

ofH

equal to the current enclosed.

In the case of a single wire the integration always yields the current / in the wire
regardless of the path of integration provided only that the wire

by the path.

As

illustrations, integration

Fig. 5-22 yields /, while integration

these paths

do not enclose the

is

completely enclosed

around the two paths

around the paths

at (c)

and

at (a)

and

(b) in

(d) yields zero since

wire.

kj)
(b)

*H
FIGURE

5-22

Line integral of

Wire

around closed paths

equals current in wire

when paths en-

5=^f

and (b) but is zero when

the paths do not enclose the wire (c) and

close the wire (a)

(d).

(d)

THE STATIC MAGNETIC FIELD OF STEAD Y ELECTRIC CURRENTS

sec. 5-14]

EXAMPLE
1

Am

-1
.

The magnitude of

Find the current

solution

According to

a radius of

(5),

the current in the wire

d\

linear conductor

is

given by

is

= H X 2nR = 2tt A

solid cylindrical conductor of radius

Derive expressions for

from a long

in the wire.

/= |h

EXAMPLE

at

1 65

has a uniform current density.

H both inside and outside the conductor.

Plot the variation of

H as a

function of radial distance from the center of the wire.

solution

See Fig. 5-23a.

Outside the wire

(r

> R)

I
H = 2nr

Inside the wire the value of /fat a radius r

r.

Thus, inside the wire (r

is

(6)

determined solely by the current inside the radius

< R)
H-

(7)

2-nr

where

/'

= I(r/R) 2

is

the current inside radius

HAt the surface of the wire

is

= R,

and

(8)

Therefore, inside the wire

r.

(8)

2irR 2

equals

(6).

A graph of the variation of H with r

presented in Fig. 5-236.

./Wire
Current
'Out of page

FIGURE

5-23

and outside current-carrying

wire (Example 2).
inside

166

ELECTROMAGNETICS

FIGURE

[CHAP. S

5-24

Rectangular path
current density J.

5-15

medium with

in

AMPERE'S LAW APPLIED TO A CONDUCTING

MEDIUM AND MAXWELL'S EQUATION

Ampere's law as discussed in the preceding section may be applied to the more general
situation of a path inside a conducting medium.
Thus, suppose that the origin of the
coordinates in Fig. 5-24

direction as shown.

situated inside a conducting

is

the current density in the

extent.

Let

per square meter) in the positive y

According to Ampere's law, the line integral of
around the

rectangular path enclosing the area

this case, the current

medium of large

medium be J (amperes

(Fig. 5-24) is equal to the current enclosed.

/enclosed by the path

is

In

given by the integral of the normal com-

ponent of J over the surface A, or

This expression

is

d\

jj J ds

I
(1)

a generalization of Ampere's law and constitutes one of Maxwell's

equations in integral form.

5-16

The

MAGNETOSTATIC POTENTIAL U AND MMF F

line integral

of the

static electric field

E c around a closed path is

JEe -dl=0
fThe symbol Ec

zero.t

That

is,

(1)

indicates explicitly a static electric field as produced by electric

charges, as distinguished from an emf-generating field E as, for example, in a battery.
In Chaps. 2 and 3 only E fields were considered, and so for simplicity no subscript
was used, it being understood in those chapters that E means Ec .

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

SEC. 5-16]

and can be derived from a related scalar potential

Fields of this type are called lamellar

function.

Thus,

Ec

which

is

167

due to charges,

is

derivable as the negative gradient of a

scalar potential V, or

EC
Between any two points along a path

=-VK

(2)

in the field

we have

Ec

J
Although the
closed loops),

it

static

magnetic

can be treated

outside current regions

like

d\

field is

= V - V2

(3)

not lamellar (since magnetic flux

a lamellar

field if

and do not enclose any

paths of

current.

lines

form

integration are entirely

Thus, when no current

is

enclosed,

&

d\

(4)

H can then be derived from a scalar magnetic potential function

Under this condition,

(or magnetostatic potential) U.

That

is,f

H=-Vl/
Between any two points along a path

(5)

in the field

we have

2
f
Ji

The

scalar potential

d\

U has the dimensions

Ui

- U2

(6)

of

Current

x distance

current

Distance

Hence,

U is expressed in

amperes.

Returning

now

to a further consideration of electric fields,

Sec. 4-10 that if

emfs

exist in

where

E = total field, V m
'U = total emf around

we have

learned in

a path of integration,

d\

= ctf

(7)

path,

VD =

V 2 V=0.

in charge-free regions, we obtain Laplace's equation

t Since
B
0; so if no current is enclosed,
In a magnetic field we always have
2
0.
write Laplace's equation in the magnetostatic potential as

U=

we may

168

ELECTROMAGNETICS

In a magnetic
that when current

[CHAP. 5

we may write an analogous

field

relation, based

on Ampere's

law,

enclosed by a path of integration,

is

I=F

K-dl =

(A)

(8)

where the quantity F, called the magnetomotance or magnetomotive force (mmf), is

If the path of integration in (8) encloses a number of

equal to the current enclosed.

turns of wire each with a current / in the

same

direction, (8)

JH-dl = NI = F
where

(A

may be

written

turns)

(9)

N = number of turns of wire enclosed, dimensionless

/ = current in each turn,

The product NI

is

expressed in ampere-turns, and the

mmf in

this case

has the same

units.

The above

When

relations for electric

the integration

is

and magnetic

summarized

fields are

restricted to current-free regions

in

Table

5-1.

and to paths that are

U and mmf F are the same. The requirement that the path
not link the current can be met by introducing a hypothetical barrier surface in the
not closed, the potential

magnetic

field

Table 5-1

COMPARISON OF ELECTRIC AND MAGNETIC FIELD RELATIONS

through which the path

Relation

Electrostatic fields

Closed path

<j>Ec .</l

Gradient of scalar

is

not allowed to pass.

Magnetostatic

<l>

EC =-VK

fields

H- d\ = 0; no

current enclosed (Fig. 5-25)

H = VU (Am"

(Vm-1)

For example, imagine

); in

current-free region

potential

Integral between
two points

Closed path

JE

|E

.dl=V

rfl

= IT

-V

(V)

d\
Jt
currents

(V)

d\

Ui

- U2

= /= F

(A) path avoids

;

all

(A) ; path encloses current

<J>

(Fig. 5-26a)

or

d\

<J>

current

= NI = F

(A

turns); path encloses

N times (Fig. 5-266)

THE STATIC MAGNETIC FIELD OF STEAD Y ELECTRIC CURRENTS

sec. 5-16J

69

(a)

Wire

(6)
Path of

MMres

integration

FIGURE

of loop

FIGURE

5-25

Path of integration enclosing no current (see

Table 5-1).

that a long conductor

5-26

Path of integration enclosing current /.

Cross section through 5-turn loop showing
path linking the 5 turns (see Table 5-1).

(a)
(6)

normal to the page as

in Fig. 5-27 carries

a current

/.

Let a

from the wire an infinite distance to the

left, as suggested in the figure.
Now the integral of H from points 1 to 2 yields the
current / provided 2 and 1 are separated by an infinitesimal distance. Thus
barrier surface be constructed that extends

d\

I/,

The requirement of (4)

is still satisfied

- U2 = /

(A)

since the line integral of

path 1231 that avoids crossing the barrier

is

zero.

That

(10)

H around the closed

is,

H-dl =

To

(ID

infinity_

Barrier^
surface

FIGURE 5-27
Conductor and barrier

surface.

Path of integration

170

ELECTROMAGNETICS

[CHAP. 5

FIGURE

5-28

Current-carrying wire showing magnetic equipotentials (radial)

and

Both

field lines (circles).

U and F are scalar functions

is, U is a single- valued
.

integration; that
fact that the

The potential

If a current-carrying wire

path of integration (multiple linking), the result

Hence

F is

This follows from the

path of integration never completely encloses the current and

to current-free regions.

valued since

U is independent of the path of

function of position.

its

is

is

encircled

called the

mmf F.

magnitude depends on the number of times the path

not, in general,

is

restricted

more than once by the

It is

multiple-

encircles the wire.

independent of the path of integration.

In Fig. 5-27 the barrier surface represents a magnetic equipotential plane.

point

is

If

taken arbitrarily as zero potential, then the potential of point 2 on the other

side of the barrier

is /.

Hence, we

may construct two surfaces as in

Fig. 5-28,

one with

U = and the other U = I. Other equipotential surfaces are also drawn in Fig. 5-28
for U = 1/4, U = 1/2, and U = 37/4. The equipotential surfaces are everywhere normal to

H and extend from the surface of the wire to infinity.

the interior

5-17

of the

FIELD CELLS

AND PERMEABILITY
of two

Consider a transmission

line

The

w and

strips

have a width

transmission line

They do not extend into

wire.

is

shown

flat parallel

a separation

/.

conducting strips as in Fig. 5-29a.

Each

/.

The

The field between the

strips

strip carries

in cross section in Fig. 5-296.

a current

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-17]

FIGURE

5-29

Parallel-strip transmission line (a) in perspective

Magnetic

tion, (c)

171

and (6)

field cell (or transmission-line cell)

in cross sec-

with strips of

width equal to spacing.

is

uniform, except near the open

field

If equipotentials are

sides.

region with a spacing equal to the separation

of the

line,

drawn

in the

uniform

we may regard

the line

composed of a number of field-cell transmission lines (or transmission-line

arranged in parallel. Each transmission-line cell has a square cross section,

as being
cells)

as in Fig. 5-29c.

The

current in each strip of a line

entire line.

of line

is

Thus, across one

given by

A = Bid.

cell

cell is /'

HI =

/'.

Now

(//w)/

where /

is

The inductance of this length of line

L =
or the inductance per unit length1

Bid

T'~ HI

the current in the

the total flux linkage per length

= Hd

is

then

(1)

is

7-'

(2)

172

For

ELECTROMAGNETICS

air

fi

fi

4O0n

[CHAP. 5

nH m -1

medium has an inductance per

so that a field-cell transmission line with air as the

-1
1
400n nH
, or 1.26 nH m"

unit length of

Thus, the permeability n of a medium may be interpreted as the inductance per

unit length of a transmission-line cell filled with this medium.

EXAMPLE

Using the

concept, calculate the inductance and also capacitance

field-cell

per unit length of the coaxial transmission line shown in cross section in Fig. 5-30.

The

line is air-filled.

solution

The inductance per

unit length of the coaxial line

= - = !

(Hm-

fi

= number of line cells in parallel

= permeability of air = 1.26 fM m~

given by

(3)

where L /d = inductance per unit length of transmission-line

n

is

cell

Dividing the space between the coaxial conductors into curvilinear squares, we obtain 9.15
line cells in parallel.

Thus

1.26

9.15

As a check, we note
inner so that from (5-13-9)

we

the

same

is

twice the radius of the

get

a
is

that the radius of the outer conductor

which

0.138 /uHrn-

= 0.46 log 2 = 0.138 |iiH m~

result as obtained above.

FIGURE

5-30

Coaxial transmission line divided into

9.15 field-cell lines in parallel.

Outer
conductor

Field cell

THE STATIC MAGNETIC HELD OF STEADY ELECTRIC CURRENTS

sec. 5-18]

The capacitance per

173

unit length of the coaxial line of Fig. 5-3ft is given by

C
(4)

where

C /d= capacitance

per unit length of line cell (same as capacitance per unit length
of field-cell capacitor; see Sec. 3-22)

n
to

= number of line cells in parallel

= permittivity of air = 8.85 pF m~

Thus

-7

= 9.15x8.85 = 81 pFm-

Using the relation of (3-18-3),

24.2

d
which

5-18

An

is

the

same as obtained by the

ENERGY

IN

81

cell

pFm"

method.

AN INDUCTOR

inductor stores energy, as

With the switch

log 2

may

be demonstrated with the aid of the circuit of Fig.

lamp is lighted. When the switch is opened, the

increases momentarily in brilliance because the magnetic energy stored in the
magnetic field of the inductance induces a current as the field collapses. The (mag5-31.

s closed the

lamp

netic) energy delivered to the

lamp

is

Wm = ^Pdt = ^VIdt
From

circuit

theory the voltage across the inductor

Wm = LJldI=iLI

FIGURE

(J)

is

given by

(1)

V = L dljdt;

(J)

5-31

Circuit for demonstrating energy storage

in

a magnetic

field.

so
(2)

[CHAP. 5

ELECTROMAGNETICS

174

Thus, the magnetic energy has the dimensions of inductance times current squared.
Since

L = A//,

the energy stored can be variously expressed as

^=*l/2 =*a/ 4t
where

(3)

Wm = energy stored, J
L = inductance of inductor, H
/ = current through inductor, A
A = flux linkage, Wb turns

5-19

ENERGY DENSITY

IN

The energy possessed by an inductor

is

A STATIC MAGNETIC FIELD

stored in

its

magnetic

of side length A/ and volume Av

field.

Let us find the

Consider a small unit cube

density of this energy as a function of the flux density B.

A/ 3 situated in a magnetic

field as in Fig. 5-32a.

Let thin metal sheets be placed on the top and bottom surfaces of the cube, each with

a current A/ as indicated.

Also

let all

the surrounding space be

The

cubes, as suggested by the cross section of Fig. 5-326.

filled

with such

directions of the current

flow on the sheets are indicated by the circles with a dot (current out of page) and
circles

with a cross (current into page).

Each cube can be regarded as a magnetic field-cell transmission line of length

(into page) of A/.
Each cell has an inductance AL = \i A/. The field H is related to
the current A/ by H A/ = A/. The energy stored in each cell is, from (5-18-3),

AWm = iAL AJ
It

(J)

follows that

AWm = inH

A/

= inH 2

Dividing (1) by Av and taking the limit of the ratio

we

cell

(1)

AWJAv

obtain the energy per volume, or energy density,

point around which the

At;

wm

of volume Ay shrinks to zero.

*f

= hm

Av-0 Av

= inH 2

as

Ay approaches

of the magnetic

zero,

field at

the

Thus

(2)

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-20]

175

(a)

J7

Metal
sheets

Jl_
'
Current

Jia

-*-B

a_a_
U

-*-B

directions

FIGURE
(a)

(b)

a_

5-32

Small cubical volume in a magnetic

(6) Cross section through region

field.

filled

5-20

with

many such

cubes.

MAGNETIC POLES

A bar magnet can be represented by two magnetic poles of pole strength Q m separated
by a distance

resulting in a dipole

be separated (see Chap.

6),

moment Q m

I.

Since the poles of a magnet cannot

an isolated magnetic pole

is

not physically realizable.

This contrasts with electric charges (positive and negative), which can be separated.
Nevertheless, an isolated magnetic pole

is

a convenient fiction and will be utilized in

the next section to illustrate the concept of curl.

F=
where

Qm

5-21

CURL

is

Hi,

as the

is

on the pole

B&

or

(1)

relates the line integral

of

B around

the current enclosed by the loop.

work per

closed path.

write for the force

the pole strength in ampere meters.

Equation (5-14-3)

where /

We

unit pole required to

move an

a finite closed path, or loop, to

Let us regard the line integral of

isolated magnetic pole

around the

176

Although

relations involving finite paths are useful in circuit theory,

quently desirable in
is

[CHAP. 5

ELECTROMAGNETICS

field

it

fre-

is

Curl

theory to be able to relate quantities at a point in space.

such a point relation and can be regarded as an extension of Ampere's law so that

it

applies at a point.

Consider a small plane area As in a conducting medium with a current A/

flowing through the area and normal to

The meaning of the

it.

curl of

B may

then

be expressed as follows: The magnitude of the curl ofB is equal to the ratio of the work
per unit magnetic pole (carried around the boundary of the area) to the area As as As
approaches zero.

Further, the curl of

plane of the area.

hm

AI =
The

is

a vector with a direction normal

current density

= hm

As

a s ->o

where J

Thus, the magnitude of the curl of

is

to the

given by

= nJ
As

As-o

A// As as As

(1)

-*

current through area A*

direction of the curl of

Equation

is

normal to the area As.

(1) gives the total curl

of

B if J

is

normal to the plane of the loop.

If

not normal to the plane of the loop, (1) gives only one component of the total curl
expression, which will be developed in the following paragraphs.

is

Suppose that the rectangular coordinate system shown

inside a conducting

medium of large

extent.

in Fig. 5-33

is

situated

Let the current density in the

medium

be J and the component of the current density in the x direction Jx

AI through

the small area

Ay Az

(Fig. 5-33)

is

The

total current

then

Jx AyAz = AI
This current produces a magnetic
at the

j>

axis be

not uniform,
Fig. 5-33)

field.

(2)

Let the flux density along edge

By and the flux density along edge 4 at the z axis be Bz

its

value at edges 2 and 3

may be

expressed to a

first

of the area

If the field

approximation

is

(see

by

dB

* Ay
Bz +-^Ay

oy

now

j
and

dB

y
B,y + ^Az
B
oz

(3)

work performed per unit magnetic pole carried around the

The
total work is equal to the sum of the increments of work
periphery of the area.
edges.
Each increment of work equals the force per unit pole
along each of the four
unit pole moves.
The total work will be calculated
multiplied
distance
the
(B)
by the
path
in
a counterclockwise direction, as
magnetic
moved
around
the
per unit
pole
around the path since a right5-33.
This
positive
direction
is the
shown in Fig.
Consider

the

THE STATIC MAGNETIC HELD OF STEADY ELECTRIC CURRENTS

sec. 5-21]

FIGURE

177

5-33

Construction for finding x component of

curl of B.

handed screw rotated in this direction will advance in the positive x direction.
work to move the unit pole along edge 1 is By Ay and along edge 2 is

(* +
The work along edge

'

The

Az

3 is

(-.+%")
the minus sign indicating that the motion

is

Az.

The

total

work equals

is

the

against the

sum of

field.

The work along edge

these four increments, and this

equal to n times the total current through the area as given by

(2).

is

Thus,

3B,

j>B-dl = By Ay + Bz Az + -r1 AyAz

dy

dB

By Ay-?AyAz-Bz Az = )xJx Ay Az
oz

(4)

from which

B dl=
-

Dividing by the area

Ay Az and

(^-^) AyAz = ^ Aj,Az

taking the limit of the ratio

Work done around

periphery of area

Area

(5)

178

[CHAP. 5

ELECTROMAGNETICS

as the area approaches zero,

is

the

jB-dl

,.

hm

AyA*-+o
(6) is the

In this instance

obtain by definition the curl of B.

Therefore,

curl of B, written curl, B.

x component of the

Equation

we

Ay Az

complete

BB

Z
-r1
dy

dB

~ ~ry =
dz

f*

Jx

curl*

we have

differential expression for curl

it

(6)

B if the current flows

components flowing in the

y and z directions, curl B also has components in these directions. Deriving them in an
identical manner to that above and adding them vectorially, we obtain the complete
However,

only in the x direction.

if

the current also has

expression for the curl

This

is

equal to the vector

sum of the

curl

Dividing by

/x,

C-M-)

-*(-) +

current-density components, or

B = ti(%Jx +

del (V)

is

z)

= nJ

(8)

H=J

(9)

conveniently expressed in vector notation as the cross product of the operator

and H,

that

is,

dy

dx

U = kH x + fH
which

fJy

we have
curl

Curl

<"

dz

(10)
z

yields

v(f-fW(-M-f)
:

HJX + yJ + tJz

or

V x

H can also

=J

(11)

xH = J

(12)

be conveniently expressed in determinant form as

VxH

Jy

dz

(13)

Hx H

H,

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CUMENTS

sec. 5-21]

/>'

179

Vertical

or z axis

-Trough

Vxv

.*

Vxv -o|

(d)

{9)

^
v-

- M.

FIGURE

Water trough

The

curl

5-34
for

Example

of a quantity

is

y^-

a point function.

Therefore, according to (12),

H has a nonzero value only at points where the current density J not zero. At
a point inside a wire carrying a steady current, V x H equals the current density J at
the point, but at a point outside the wire V x H = 0.
V x

is

EXAMPLE
is

shown

trough

A rectangular trough carries water in the x direction. A section of the trough

The width of the

Find the curl of the velocity v of the water for two assumed conditions, (a) The
everywhere uniform and equal to a constant, i.e.,

in Fig. 5-34a, the vertical direction coinciding with the z axis.

is b.

velocity

is

&K

(H)

180

[CHAP. 5

ELECTROMAGNETICS

= unit vector in positive x direction,

K = a constant, m s "

dimensionless

where x

top view of the trough

The

shown

is

in Fig. 5-346 with the positive

fact that the velocity v is constant

by the graph of
(b)

v, as a function of

The velocity

varies

direction

downward.

suggested by the arrows of uniform length and also

is

y in Fig. 5-34c.
at the edges of the trough to a

from zero

maximum at the center,

the quantitative variation being given by

= x^sin^

(15)

K=& constant, m s "

where

The

width of trough, m

sinusoidal variation of v

is

suggested by the arrows in the top view of the trough in Fig.

5-34e and also by the graph of v x as a function of y in Fig. 5-34/

solution

Equation (14) can be reexpressed

(a)

= xy x

where

v x is the

component of velocity

in

direction.

terms involving v x , namely, 8v x /dz and dv x /8y.

and hence
(b)

Vxv=

is

Thus

Since v x

everywhere in the trough (see Fig.

= xvx

Equation (15) can be reexpressed v

vx

Since v x

(16)

is

vx

= K.

The

curl of v has

5-34rf).

Thus

= Ksin^

(17)

not a function of z, the derivative BvJBz

= 0.

However,

v x is

a function of y so that

Kn
-=
-COS
8v

and we have for the

try

(18)

8y
curl of v

Try
Krr

cos V X v = - z,

i.

is

the unit vector in the positive z direction;

the curl of v
center

it is

is

in the negative z direction

in the positive z direction.

,,

(19)

where

two

a constant, both terms are zero

i.e.,

at the left of the center of the trough

(downward in Fig. 5-34a), while to the right of the

The variation of the curl of v across the trough is

presented graphically in Fig. 5-34#.

A physical interpretation of the curl

of v in the above example can be obtained

with the aid of the curl-meter, or paddle-wheel, devicef of Fig. 5-35. If this device
is inserted with its shaft vertical into the trough with the assumed sinusoidal variation
f H.

H.

Skilling,

Sons, Inc.,

New

" Fundamentals of Electric Waves," 2d

York, 1948.

ed., p. 24,

John Wiley

&

THE STATIC MAGNETIC HELD OF STEADY ELECTRIC CURRENTS

sec. 5-21]

FIGURE

5-35

Paddle wheel for measuring

for the velocity of the water [part (b) of

left

of the center of the trough (position

at the right

Example
1

1], it

in Fig. 5-34e)

curl.

spins clockwise

when

it is

at the

and counterclockwise when it is

of the center of the trough (position 2 in Fig. 5-34e), corresponding to

At the

negative and positive values of curl.

center of the trough (position 3 in Fig.

5-34e) the curl meter does not rotate since the forces

This corresponds to the curl of v being zero.

wheel shaft

181

The

on the paddles are balanced.

rate of rotation of the paddle-

proportional to the curl of v at the point where

is

At any point the

rotates fastest near the edges of the trough.

it is

inserted.

Thus,

rate of rotation

is

it

also

maximum with the shaft vertical (rather than inclined to the vertical), indicating that
V x v is in the z direction. It is assumed that the paddle wheel is small enough to

avoid affecting the flow appreciably and to indicate closely the conditions at a point.
If the curl

assumed

meter with shaft vertical

in part (a) of

EXAMPLE

will

1, it

is

inserted in water with uniform velocity, as

not rotate

(curl v equals zero).

Consider a current-carrying conductor of radius R as shown in cross section

The current is uniformly distributed so that the current density J is a constant.

in Fig. 5-36of.

Example

Taking the axis of the wire in the z direction,

zJ

where / is the

total current in the

t
2

(Am-)

conductor in amperes.

Find the curl of

(20)

H both inside and

outside the wire.

solution

Example 2 of
Since

is

The

variation of

Sec. 5-14.

entirely in the

The

H as a function of radius was worked out for this case in

H shown in the graph of Fig. 5-366.

variation found for

direction,

is

we have

H = 6ff,
/

where

/
2nr

outside conductor

H.2irR 2

inside conductor

(21)

182

ELECTROMAGNETICS

[CHAP. 5

Wire

(a)

(b)

iTr

(e)

FIGURE

5-36

Conducting wire (Example

Using the expression for curl

in cylindrical coordinates (see inside

V xH=

and zero outside

5-22

The

back cover), we have

outside conductor

VXH = 2
Hence, the curl of

2).

inside conductor

(23)

H has a value only where there is current, being a constant in the conductor

(see Fig. 5-36c).

MAXWELL'S FIRST CURL EQUATION

relation derived

from Ampere's law

in Sec. 5-21 that

VxH = J
is

(22)

one of Maxwell's equations.

the field

Equation

to the current density

integral form, as given

by

(1) is

at a point.

(5-15-1), relates

(Dt
a differential expression and relates

The corresponding expression

in

H around a finite closed path to the total

current passing through the area enclosed.

The
is a special form of the more general relation given in (8-14-4).
more general equation has an additional term involving the displacement current

t Equation (1)
density.

is present only for time-changing

considered here, (8-14-4) reduces to (1).

However, a displacement current

so that for steady

fields, as

fields

THE STATIC MAGNETIC HELD OF STEADY ELECTRIC CURRENTS

SEC. 5-23]

183

Thus far, we have encountered three of Maxwell's four equations applying at a

They are V D = p, V B = 0, and (1). The fourth relation, (8-8-2), is
also an equation involving curl, so that (1) may be referred to as Maxwell's first curl
equation and (8-8-2) as the second.

point.

5-23

SUMMARY OF OPERATIONS INVOLVING V

We have discussed four operations involving the operator V (del), namely, the gradient,
divergence, laplacian,

and

curl.

divergence of the gradient (V

Although the laplacian can be resolved into the

/= V

V/), this operation

is

of such importance as to

These operators with their differential equivalents in

rectangular coordinates are listed below with /representing a scalar function and F
warrant

listing

it

separately.

a vector function.

grad/= V/= x J^+ y

+ *zf
dz
dy
dx

Gradient:

(1)

Gradient operates on a scalar function to yield a vector function.

div

Divergence:

SFX
F = V F = - +

dx

5F
-r2

dy

dF

+ -r1z

(2)

dz

Divergence operates on a vector function to yield a scalar function

Laplacian:

The

div(grad/)

laplacian operates

on a

=V

(Vf)

fH
f
= \ 2f = f
- +
H j +
j^
dz
dx 2
dy 2
d

(3)

scalar function to yield another scalar function.f

Curl operates on a vector function to yield another vector function.

t In rectangular coordinates it is also possible to interpret the laplacian of a vector

function as the vector sum of the laplacians of the three scalar components of the
vector.

Thus

V 2 F = xV 2 Fx + $V2Fr + tV2F,
However, in no other coordinate system

is this

simple interpretation possible.

184

5-24

ELECTROMAGNETICS

[CHAP. 5

A COMPARISON OF DIVERGENCE AND CURL

Whereas divergence operates on a vector function to

yield a scalar function, curl

operates on a vector function to yield a vector function.

There

is

another important

Referring to the differential relation for the divergence in (5-23-2), we

note that the differentiation with respect to x is on the x component of the field, the
difference.

y is on the y component, etc. Therefore, to have

divergence the field must vary in magnitude along a line having the same direction as
differentiation with respect to

the field.f

Referring to the relation for curl in (5-23-4),

we note, on the other hand, that

on the y and z components of the field, the
on the x and z components, etc. Therefore, to have

the differentiation with respect to

differentiation with respect to

is

y is
must vary in magnitude along a line normal to the direction of the field 4
This comparison is illustrated in Fig. 5-37. The field at (a) is everywhere in the
y direction. It has no variation in the x or z directions but varies in magnitude as a
function of y. Therefore this field has divergence but no curl. The field at (b) is also
everywhere in the y direction. It has no variation in the x and y directions but does
vary in magnitude as a function of z. Therefore, this field has curl but no divergence.
curl the field

Let us

now

discuss the significance of operations involving

sider the divergence of the curl

of a vector function;

V (V x

where F

is

first

twice.

First con-

F)

(1)

any vector function given in rectangular coordinates by

F = &FX
If we

i.e.,

take the curl of F,

we

+ $Fy + IFZ

obtain another vector.

this vector, the result is identically zero.

V (? x

Next taking the divergence of

Thus
F)

(2)

In words (2) states that the divergence of the curl of a vector function is zero. As a
corollary we may say that if the divergence of a vector function is zero, then it must be
the curl of

some other

vector function.

t This is a necessary but not a sufficient condition that a vector field have divergence.
For example, the
field due to a point charge is radial and varies as 1/r 2 but has no
divergence except at the charge. If, however, the field is everywhere in the
direc-

tion, as in Fig. 5-37<j,

and

varies only as a function

of

y,

then this

field

does have

divergence.
X This is a necessary but not a sufficient condition that a vector field have curl.
For
example, the
field outside of a long wire varies in magnitude as \jr and
has a
direction normal to the radius vector; yet
has no curl in this region. If, however,
the field is everywhere in the y direction, as in Fig. 5-376, and varies only as a function

of

r,

then this

field

does have curl.

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-24]

185

,,Z

x (out)
*

^-_

Vector

field

with divergence

Vector

(a)

field

with curl

(b)

FIGURE

5-37

Examples of fields with divergence and with

For example, the divergence of

is

always zero everywhere;

i.e.,

V-B =
Therefore,

(3)

can be expressed as the curl of some other vector function.

designate this other vector function by A.

Let us

Then

B=V x A
The function

curl.

(4)

A in (4) is called the vector potential and is discussed in more detail in the

next section.

Let us consider another operation involving

gradient of a scalar function/.

That

Vx
Taking
result is

first

the gradient

found to be

of/ and then

identically zero.

twice, namely, the curl of the

is,

(5)

(\f)

the curl of the resulting vector function, the

Thus

Vx(V/)eO

(6)

As a
(6) states that the curl of the gradient of a scalar function is zero.
any vector function which is the gradient of some scalar function has no curl.
For example, we recall from (2-13-2) that the static electric field due to charges

In words

corollary,

Ec is

derivable as the gradient of a scalar potential V.

E= -\V

Thus
(7)

186

ELECTROMAGNETICS

[CHAP. 5

It follows, therefore, that

Ec

the curl of

is

zero, or

V x Ec =
If a vector field has
field

on

Ec is

The

lamellar.

no

curl,

positive charges (as sources)

other hand,

if

solenoidal.

Its flux

Finally,

a vector

it is

it is

flux tubes

field

(8)

said that the field

of such

fields are

is

lamellar.

discontinuous.

and terminate on negative charges

has no divergence such as B,

it is

Thus the electric

They originate

(as sinks).

On the

said that the field

is

tubes are continuous, having no sources or sinks.

important to note that from

sources (p) of the electric field, and from

Vx

D = p,

the divergence

of It finds the

H = J the curl of H finds the sources (J)

of the magnetic field.

THE VECTOR POTENTIAL

5-25

According to

(5-9-6), the

magnetic flux density

distribution, as in Fig. 5-38,

is

at a point

P produced

by a current

!([ J
**r
J x
'"
H

B
where

given by

B=

flux density, T
-1
=
permeability
of medium (uniform), H m
H
J = current density at volume element, A m~ 2
f = unit vector in direction of radius vector r, dimensionless
r = radius vector from volume element to point P, m
dv = volume element, m 3

Carrying out the integration over the entire volume occupied by the current-carrying
conductor gives the total flux density B at P due to the current.
In Sec. 5-24

we noted

possible to express

that since the divergence of B

as the curl of

some other

vector.

is

always zero,

Thus from

it

should be

(5-24-4)

we can

write

B=VxA
where
If

we

(2)

A is called a vector potential since it is a potential function that is also a vector.t

also

make

VA=0
t

The

potential function

relation

Ec = V V)

is

(3)

V from which the electric field Ec can be derived (by the

a scalar quantity, and hence V is a scalar potential.

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sbc. 5-25]

187

Volume element dv
inside conductor

Current-carrying

conductor

FIGURE

5-38

Construction for finding flux density

at P.
B is into the page at P.

A is completely

Taking the curl of

defined.

(2) yields

VxVxA = VxB =
By

(4)

/xJ

-9)
the vector identity for the curl of the curl of a vector (see Appendix A, Sec. A-9)

equation (4) becomes

V(V A)

Introducing the condition of

(3), this

- V 2 A = nJ

(5)

reduces to

V 2 A = -n3
or in terms of the three rectangular components of

WA
2

Equation

(7) is the vector

+ yV 2

^+W A
2

sum of three

(6)

A and

= -^i/, + y/ + 2/z )

scalar equations.

Hence,

S 2 A X = -nJx
\ 2 A y = -fiJ,
\ 2 A 2 = -/i/,
Each of these

relations has the

same form as Poisson's equation.

to the three equations of (8) are

(7)

(8)

Therefore solutions

ELECTROMAGNETICS

188

[CHAP. 5

'

dv
in
<-///t
47t JJJ r
.

'-s/JJt*
Taking the vector sum of the components for

A in

(9) gives

(10)

is

According to (10), the vector potential A at a point due to a current distribution

equal to the ratio J/r integrated over the volume occupied by the current distribu-

tion, where J is the current density at each volume element dv and r is the distance
from each volume element to the point P, where A is being evaluated (see Fig. 5-38).
If the current distribution is known, A can be found.
Knowing A at a point, the flux

density

at that point

From

(2)

is

we note

Magnetic

then obtained by taking the curl of

that

A has the

flux density

x distance

Hence, the vector potential

current

can be expressed in webers per meter or newtons per

The dimensional symbols

EXAMPLE

as in (2).

distance

ampere.

dimensions of

for

A are ML/ IT 2

Consider a short copper wire of length

/ and a cross-sectional area a situated

shown in Fig. 5-39. The current density
hypothetical situation that J is uniform through-

in air coincident with the z axis at the origin, as

is

in the positive z direction.

Assume

the

out the wire and constant with respect to time. Find the magnetic flux density B everywhere
at a large distance from the wire, using the vector potential to obtain the solution.

(10),

solution The vector potential A at any point P produced by the wire is given by
where the ratio J/r is integrated throughout the volume occupied by the wire. Since

we wish
distance.
(r

>

I).

to find

it suffices to find A at a large

be large compared with the length of the wire

only at a large distance r from the wire,

Specifically the distance r should

Then, at any

poirit

P the distance

r to different parts

of the wire can be considered

constant and (10) written as

=JIf j

<fo

<>

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-25]

FIGURE

189

5-39

Construction for finding the vector potential

A and flux density B due to a short

current-carrying wire.

Now J is everywhere in

the z direction and also

JJ/

where / = Jz a

is

* = 1 JIj/ ,-*

the current in the wire.

this result in (11),

we

is

Thus J

uniform.

<fl

-t

= tJt

and

/I / dl

(12)

Completing the integration

in (1 2)

and

substituting

obtain

A = 2 Poll

i.A z

(13)

477T

A = vector potential at distance r from wire, Wb m~

unit vector in positive z direction, dimensionless
~
Ho = permeability of air = 4007T nH m
/ = current in wire, A
/ = length of wire, m
r = distance from wire, m
Equation (13) gives the vector potential A at a large distance from the wire.
*

where

i.

in the positive z direction as indicated

to the distance r from the wire.

It is

by the unit vector

i.

not a function of angle

and
(<j>

is

It is

everywhere

inversely proportional

or 8 in Fig. 5-39).

190

ELECTROMAGNETICS

of A.

Having found the vector potential A, we obtain the flux density

In rectangular components the curl of A is given by

[CHAP. 5

by taking the curl

v-=#-t) + #-H(t-f)
A has only a z component,

Since

(14) reduces to

VXA = iy!-J^
dx
8y
Now r = Vx 2 + y 2 + z 2

Therefore

= TAn
f

oy

+ ^F^+>'
oy
2

= ^7
SA Z

and

(15)

dx

fi

Il 3

4tt

(x

dx

,/2

-^/Z

(16)

x-

(17)

4ir

Poll
)= -!-2
+y + z 2\-,i2
An r
2

1'2

Introducing these relations in (15) and noting the geometry in Fig. 5-39,

Anr 2 \

B=

Anr 2

we have

VxA^^

(19)

Anr 2

B = magnetic flux density at distance r and angle 6, T

$ = unit vector in direction (see Fig. 5-39), dimensionless
= angle between axis of wire and radius vector dimensionless
"
Ho = permeability of air = 4007r nH m
/= current in wire, A
/ = length of wire, m
r = distance from wire to point where B is being evaluated, m

where

4>

r,

According to
<t>

direction.

the z axis.

That

(19), the flux density

is,

produced by the wire

is

everywhere in the

the lines of magnetic flux form closed circles concentric with

The planes of the

circles are parallel to the

magnetic flux at a distance r from the origin

is

xy plane.

One such

indicated in Fig. 5-39.

line

of

According to

and inversely proportional to r 2

Although the result of (19) could have been written down almost directly from
(1), without using the vector potential explicitly, the above example serves to illustrate
the manner in which the vector potential can be applied. Employing the vector
potential in the above example is analogous to using a 10-ton steam hammer to crack a
walnut. However, on many problems of a more difficult nature the vector potential
(19),

is

is

also proportional to sin 6

indispensable.

THE STATIC MAGNETIC FIELD OF STEADY ELECTRIC CURRENTS

sec. 5-26]

191

A COMPARISON OF STATIC ELECTRIC

5-26

AND MAGNETIC
to

It is instructive

compare

FIELDS

electric

ences and their similarities.

and magnetic

partial

and to note both

fields

comparison

is

relations developed in the preceding chapters for static fields.

relations for nonstatic fields

Table 5-2

comparison of

given in Sec. 8-18.

is

A COMPARISON OF STATIC ELECTRIC

AND MAGNETIC FIELD EQUATIONS

Description of equation

Force

Basic

their differ-

given in Table 5-2 involving

for

relations

and solenoidal

lamellar

Magnetic

Electric fields

fields

= (IxB)<//

F=GE

rfF

V x Ec = Of

V'B =

EC =-VK

B= VXA

F=QB

fields

Derivation from scalar or vector potential

4ire J

Constitutive relations

Source of

electric

and magnetic

ATT

v r

ivT

D = eE

B = /xH

V'D = p

VXH = J

we = iE 2 = iED

w m = iixH 2

L=

fields

Energy density
Capacitance and inductance

Capacitance and inductance

per unit length of a cell

<J>E-rfl

JE

Derivation from scalar poten-

A
7

=e

d=

rf

Closed path of integration

= iBH

&H-dl=F = NI

= 1T

-dl

<p

rfl

H=-V/

Ee=-VK

no current enclosed
in current-free region

tials

Ec

is

fields

the static electric field intensity (due to charges).

(not due to charges) may also be present.

(without subscript) implies that emf-producing

192

ELECTROMAGNETICS

PROBLEMS

[CHAP. 5

(ji

(i

everywhere)

Two long parallel linear conductors carry 100 A. If the conductors are separated
mm, what is the force per meter of length on a conductor if the currents flow (a) in

5-1

by 20

opposite directions and (b) in the same direction?

A long linear conductor with current of 10 A is coincident with the z axis The
+ z direction. If B = 3 + y4 T, find the vector value of the force F

5-2

current flows in the

per meter length of conductor.

Show that (5-9-4) can also be written dF = (B X I) dl.

A square wire loop 2 m on a side is situated with edges coincident with the positive
axes and one corner at the origin. A long straight wire with 50 A flowing in the

5-3

5-4

x and y
+ x direction

is

situated in the

(straight wire crosses

wise, as viewed

5-5 (a)

from the

If

xy plane and

loop without touching

+ z direction, find

B = z6 sin (nx/2)

parallel to the

is

axis at a distance of

loop current

If the

it).

is

20

the vector value of the force on the loop.

sin (iry/2) T, find the total

magnetic flux over a square

m on a side with edges coincident with the positive x and y axes and one corner at

area 2
origin,

(b) If

5-6

B = iJc/r

T, what

is

flowing clock-

the

the magnetic flux through a circle of radius r ?

thin linear conductor of length

carrying a current /

is

coincident with the

The medium surrounding the conductor is air. One end of the conductor is at a
distance y from the origin and the other end at a distance y 2
Show that the flux density
due to the conductor at a point on the x axis at a distance Xi from the origin is

axis.

B=
Note

that

if

Vxi 2

+ yi 2 /

B = fi

( y t

=y

2 ),

and

if

Xi

>

I,

Il/Anx*.

A uniform cylindrical coil, or solenoid, of 2,000 turns is 600 mm long and 60 mm

5-7

in diameter.

the

yi

\V Xl 2 +y 2 2

the center of the conductor coincides with the origin

the expression reduces to

yz

h-qI I

4"*

coil, (b)

end of the

If the coil carries

on the

axis at

a current of 15

one end of the

mA,

coil, (c)

find the flux density (a) at the center of

on the

axis halfway

between the center and

coil.

5-8 Calculate and plot a graph of B as function of position along the axis of the solenoid of Prob. 5-7 from the center of the solenoid to a distance of 100
beyond one end.

mm

5-9

Show that the flux density at a point P on the axis of a uniform solenoid is given by

NI
where

Ct

O,

Cl 2

= solid angle subtended from point P by left end of solenoid ( = 2n if P is

at left

end of solenoid)

n 2 = solid
right

Note that

angle subtended from point

P by

right

end of solenoid

( =-

2n

if

is

at

end of solenoid)

at the center

of a long slender solenoid

fli

* Answers to starred problems are given

= n 2 <^ 4tt so that B = /a

in

Appendix C.

Nil I =

/u

K.

problems]

the static magnetic field of steady electric currents

A solenoid 300 mm long and 15 mm in diameter has a uniform winding of 2,500

a uniform field of 4 T flux density -and a current of 2 A is

5-10

If the solenoid is placed in

turns.

is (a) maximum
moment of solenoid ?

passed through the solenoid winding, what

torque on solenoid, and

193

(c)

magnetic

force

on

solenoid, (b)

maximum

A uniformly wound solenoid of 5,000 turns is 1 m long by 15 mm in diameter.

= 2 A, what is (a) the flux density B, (6) magnetic field H, and (c) magnetic

5-11

If the current /

wm

energy density

at the center of the solenoid ?

density K, (e) the total magnetic energy

side carrying

5-14
flux density

An

infinite

B at

the focus.

has an inner radius r t and outer radius

r2

and

(c)

flux density

around the washer

(circular

loop current).

at the center of a square wire loop 2

m on a

infinite

wire of parabolic shape carries a current

wire of hyperbolic shape carries a current

/.

Find the magnetic

/.

Find the magnetic

B at the foci.
Show

that the magnetic flux density at a focus of

wire with current /

where e

the equivalent sheet current

a current of 3A.

An

5-15
flux density

5-16

flat surfaces,

Find the magnetic

5-13

is

B as a function of radius r if a current /flows (a) between inner and outer

edges, (b) between the

What

A conducting washer of thickness

5-12

Find the flux density

(d)

W and (J) the inductance LI

is

an

elliptically

shaped loop of

given by

= b/a = /axial ratio

b = semiminor axis
a = semimajor axis
r = a distance to focus from center of ellipse
1

+ \/ 2)vr + (-v 2+l)

) .{ + 2vV
(2 + V2)y-(V2+lK

,-

(2

G = 2V2 + yha )

and

,-.

1
1

sin"

V2y 2 -1

where

VT^-,

5-17

(a)

What

uniform flux density

(b)

What is

5-18

the magnetic

What

is

the

is

maximum

torque on a square loop of 200 turns in a

B = 4 T? The loop
the

is

150

moment of the loop ?

maximum torque on a

mm on

field of
a side and carries a current of 8 A.

small coil of magnetic

moment 10"'

Am" 2

situated near the center of a long air-filled solenoid of 1,500 turns per meter with a current

of 2

A?
5-19

Referring to Fig. P5-19,

situated in air

is

show

that the force between the

two current elements

given by

dF =

!^lMdl1&iX(&iXi)

(1)

194

ELECTROMAGNETICS

[CHAP. 5

FIGURE

P5-19

Geometry of short current-carrying elements

for finding force between them.

= force on element 1 due to current I2 in element 2, N

= permeability of air, H m~
dh, dh = lengths of current elements 1 and 2, respectively, m
1 h = currents in elements 1 and 2, respectively, A
r = distance between elements, m
Sj = unit vector in direction of current 7i in element 1, dimensionless
& 2 = unit vector in direction of current I2 in element 2, dimensionless
? = unit vector in radial direction (from element 2 to 1), dimensionless
Show further that dF = & F dF, where dF is given by
where rfF

fj.o

1,

fiLo

dF=\dV\
where

9i

I2 dli It

= angle between 8 2 and t (see Fig.

= angle between di'and & 2 X t
fi,

and

aF
is

sin 02 sin

P5-19)

(S 2

f)

sin

where & r

dh

2 sin

the unit vector in the direction of force dF.

8t

the static magnetic field of steady electric currents

problems]

fr-^"
FIGURE

195

/Loo P 2

P5-20
for finding torque.

Geometry of loops

noted that these equations give the force on element 1 due to the presence of
That is, they are not symmetrical with respect to elements
However, with two closed circuits the force, as given by an integral of (1), is the

It is

element 2 but not vice versa.

1 and 2.
same for both

every action there

tion s

loop

is

1 is

is

an equal (and opposite) reaction,

Two loops

5-20

Thus, in the case of actual circuits Newton's third law, that to

circuits.

If the separa-

T on loop

2 due to

Calculate the inductance of a uniform, 5,000-turn solenoidal coil 500

mm long

compared with the

given by

large

of the loops, show that the torque

size

/x

where

is satisfied.

in cross section in Fig. P5-20.

shown

are arranged as

mm'

m = magnetic moment of loop 1

= magnetic moment of loop 2

m'

5-21

and of 10

mm

radius.

The medium

is air.

Calculate the inductance of an air-filled toroidal coil of 1 ,000

5-22
area with a

radius of 500

mean

Two

5-23

mm2 cross-sectional

The toroid has a uniform winding of 10,000

mm.

identical 100-turn circular coils

m in radius have

turns.

their axes coincident

and are spaced 1 m apart, forming a Helmholtz pair. Both carry 10 A in the same sense.
along the axis of the coils from the center of one coil
Calculate and plot the variation of
of H, along the axis of a single
to the center of the other. Also calculate and plot the variation

100-turn coil

m in radius,

due to a current / in the coil.

of the Helmholtz pair and with

halfway between the coils

of the pair. Also let / have such a value that

from the Helmholtz pair at this point.

may be

area so that each

Let the single coil be situated

its

axis coincident with the axis

H at the center of the single coil

is

the

same as

Assume that the coils have negligible cross-sectional

represented by a thin single-turn loop.

The field map for a coaxial transmission line with asymmetrically located inner
conductor is shown in Fig. P5-24. If there is 10 V between the conductors and the sheet5-24

K at the point Piis2Am-',

current density

P2

(6)

EatP3

line, (/)

the magnetic field

(c)

inductance per unit length of

(i)

magnetic flux density

P3

(j>)

Hal Pi,

at

Pi

(m) magnetic energy density

magnetic flux
5-25

ductivity

<(i

( j)

wm

Pi

()

m per unit length of line,

to find (a) the electric field

(e)

(k) electric

wm

and

at

P3

density

at

energy density

(o)

E at

capacitance per unit length of

charge

(q) total current

Pi

at

(A)

Pi

D
,

at

(I)

P3

w. at

Q per unit length of ,line,

/.

of Prob. 5-24 is filled with a conductor of uniform conthe inner and outer conductor per unit length of line,
between
flows

If the transmission line

a and

map

HatP3

line, (g) electric flux

BatP 3

at

use the
id)

196

ELECTROMAGNETICS

[CHAP. 5

FIGURE P5-24
Coaxial line with asymmetrically located inner conductor.

find (a) current density

JatP 2

(d) shunt conductance

conductors have

5-26

currents of 10

Draw

a field

5-27

eter of 50

(c)

if per unit length of line, and

assumed that the transmission-line

shunt resistance

line.

It is

infinite conductivity.

in opposite directions.

map

for

line joining the wires

two

Fig. 3-16 for

J at P 3

transmission line consists of two long thin parallel conductors that carry

magnetic potential.

(b)

per unit length of

The conductors are spaced a distance 2s apart.

wires.
Show both If lines and lines of equal

a plane normal to the

Indicate the value of potential for each equipotential line.

(Compare

be arbitrarily taken to have zero potential.

this

Let the

map with

parallel lines of charge spaced a distance 2s.)

A long straight tubular conductor of circular cross section with an outside diam-

mm and wall thickness of 5 mm carries a direct current

inside the wall of the tube, (b) just outside the wall of the tube,

of 50 A.

and

(c) at

Find
(a) just
a point in the tube

wall halfway between the inner and outer surfaces.

5-28 (a) Prove that V (V x F) = 0, where F = xFx + fFy + IF

(b) Prove
V x (V/) = 0, where/is a scalar function, (c) V x H = J for steady currents. Show
V-J = 0. (</) E = -VK for static fields. Show that VxE( = 0. (e) Given
B = V x A, show that V B= 0.

that
that

that

5-29

(a)

Deduce Maxwell's

curl equation

from Ampere's law and apply

x and y axes and one corner

H = t3x 2yAm-K

at the

it

to find

m on a side with edges coinciding with the positive

-1
origin for a field H = x2y Am
(b) Repeat for

the total current /through a square area 2

THE STATIC MAGNETIC HELD OF STEADY ELECTRIC CURRENTS

problems]

197

FIGURE P5-34
Conducting strip in magnetic field for
determination of Hall effect voltage.

5-30
at
(6)

(a)

Using the vector potential and

a distance of 2
Find

5-32
5-33
at the point

5-34

magnetic

curl,

deduce the magnetic

flux density

for the situation in (a) using the Biot-Savart law.

5-31

its

m normal to a short current element of length L with current of 10 mA.

Show that the field F = xy + yx has zero curl.

2
2
find V x F and the path of V x F.
If F = ibc + ylyz - zx
2
If F = x2x + $4xy z, {a) find the curl of the curl of F; (b) evaluate
2
(2,2,2) (c) repeat (a) and (6) using V(V F) - V F.
,

this result

(a)

field

When

a current / flows as shown (Fig. P5-34) in a

normal to the strip, there

is

XB=

= (IXB)/=/(lxB):

flat

conducting

a force per unit charge to the

g(v

B)

or

right.

strip

with

Thus,

F
-=VXB

numbers of positive and negative charge carriers, the points Pi and P 2

In metals the carriers are negative charges (electrons),
potential
difference.
no
develop
and Pi becomes negative with respect to P,. This is the Hall effect. Show that the emf
developed between Pi and Pi is given by 13 = IB/nqd, where n is the number of electrons
per cubic meter in the metal, q is the electron charge, and d is the thickness of the strip. (A) If
/= 50 A, B = 2 T, w = 25 mm, and d= 0.5 mm, find the Hall voltage between P t and Pi
If there are equal

for a copper strip.

Take

n= 8.4 x

10 28

m" 3

for copper,

(c)

Why

does the width

of

the strip not appear in the above equation for the Hall voltage?

5-35

A transmission line consists of two parallel 10-mm-diameter conductors separatmm. If there is a potential difference V = 100 kV between the conduc-

ed (on centers) by 500

tors,

is

there a line current / which results in a balance between the electric

forces acting

on the wires?

If so,

what

is its

value?

and magnetic

THE STATIC MAGNETIC FIELD OF

FERROMAGNETIC MATERIALS

6-1

INTRODUCTION

Magnetic fields are present around a current-carrying conductor. They also exist
around a magnetized object such as an iron bar magnet. The field of the iron bar
is not produced by current circuits of the type considered in Chap. 5, but currents
are the cause.

In the bar the currents flow in circuits of atomically small dimen-

In contrast to these microscopic

sions.

of macroscopic

An

circuits, the circuits

considered in Chap. 5 are

size.

around the nucleus of an atom forms a tiny

all atoms have
revolving electrons, we might suppose that all substances would exhibit magnetic
effects.
However, such effects are very weak in most materials. Nevertheless there
is a group of substances, including iron, nickel, and cobalt, in which magnetic effects
electron revolving in

electric current loop.

its

orbit

Since a current loop has a magnetic field and

are very strong.

orbital

These substances are called ferromagnetic materials. Both the

motion and the electron spin (or rotation of the electron around its own

axis) contribute to the

electron, or charge,

magnetic

motion

is

effect,

the spin being particularly important.

equivalent in

its effect

This

to a tiny current loop, which

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

SEC. 6-3]

FIGURE
New

(a)

N
S "
a

N
n

6-1

poles appear at each point of divi-

199

(b)

S N

S N

S N

S
S

^^^" ^^^^ ^^^^ "^^^^

sion of a bar magnet.

(C)

magnet or magnetic dipole. Although the effect of each atomic

current loop is very small, the combined effect of billions of them in an iron bar results
in a strong magnetic field around the bar.
acts as a miniature

6-2
If

BAR MAGNETS AND MAGNETIC POLES

an iron bar magnet

is

freely suspended,

This end

that one end points north.

or simply

its

will

turn in the earth's magnetic

often convenient to call a

It is

field

so

called the north-seeking pole of the magnet

The other end of the magnet has a pole of opposite

north pole.

called a south pole.

is

it

polarity

north pole a positive pole

and a

south pole a negative pole.

All magnetized bodies have both a north and a south pole.

For example, consider the long magnetized iron rod of

isolated.

rod has a north pole

half,

new

these

is

one end and a south pole at the other.

poles appear, as in Fig. 6-16, so that there are

cut in half,

The

if

atom

two magnets.

rod
If

is

This
cut in

each of

four magnets, as in Fig. 6-lc, each with a north and a

for this

is

which acts

that the ultimate source of ferromagnetism

is

an

isolated,

fact that

it

like

would

still

have a north and a south pole.

magnetic poles cannot be isolated, whereas

an important point of difference between

6-3

If the

a tiny magnet with a north and a south pole.

the cutting process could be continued to atomic dimensions and a

circuit

Therefore, even
single iron

we obtain

The reason

south pole.

atomic current

is

at

They cannot be
Fig. 6-la.

electric

electric charges can,

charges and magnetic poles.

MAGNETIC MATERIALS

All materials

group these

show some magnetic

effects are

Depending on
Sec. 6-7)

is

is

opposed to the applied

in the

With the exception of the ferromagnetic

their magnetic behavior, substances

paramagnetic, or ferromagnetic.

ization

effects.

weak.

field,

same direction as the

show only weak magnetic

effects.

can be

classified as diamagnetic,

In diamagnetic materials the magnetization (see

while in paramagnetic materials the magnet-

field.

The

materials in both groups, however,

Materials in the ferromagnetic group, on the other

ELECTROMAGNETICS

200

[CHAP. 6

hand, show very strong magnetic

as the field, the

same as

effects.

The magnetization is in the same direction

Most of this chapter deals with

for paramagnetic materials.

the magnetic fields of the ferromagnetic materials.

number of substances

behavior.

Many

are classified in Table 6-1 according to their magnetic

substances

show such weak magnetic

However, vacuum

nonmagnetic.

is

effects that

the only truly nonmagnetic

they are called

medium.

RELATIVE PERMEABILITY

6-4

In dealing with

many media,

it is

often convenient to speak of the relative permeability

H r defined as

(1)

where

= relative permeability, dimensionless

= permeability, H m -1
H = permeability of vacuum = 400 inHm'
\i T

to be noted that the relative permeability

It is

The

relative permeability of

is

vacuum or

a dimensionless ratio.

free space is unity

relative permeability

of diamagnetic substances

is

magnetic substances

it is

than

1.

ferromagnetic materials

may

be as large as

The

generally

much

slightly less

The

greater than

by

than

The

definition.

1,

while for para-

relative permeability of the

1

and

in

some

special alloys

million.

relative permeability of

tively constant

of

slightly greater

is

diamagnetic and paramagnetic substances

and independent of the applied

dielectric substances is

field,

much

independent of the applied

is

rela-

as the relative permittivity

electric field intensity.

How-

of ferromagnetic materials varies over a wide range for

also depends on the previous history of the specimen (see

ever, the relative permeability

different applied fields.

Hysteresis, Sec. 6-12).

definite quantity for

the

maximum may

considered in

more

It

However, the maximum

relative permeability is a relatively

a particular ferromagnetic material although in different materials

occur at different values of the applied

field.

This subject

is

detail in Sec. 6-11.

In Table 6-1, the relative permeabilities nr are listed for a number of substances.
in order of increasing permeability, and they are also

The substances are arranged

classified as to

maximum

group type.

The value given

relative permeability.

for the ferromagnetic materials

is

the

sec

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

6-5]

201

Table 6-1

Substance

Group type

Bismuth

Lead
Copper
Water

Diamagnetic
Diamagnetic
Diamagnetic
Diamagnetic
Diamagnetic

0.99983
0.99998
0.999983
0.999991
0.999991

Vacuum

Nonmagnetic

Air

Paramagnetic
Paramagnetic
Paramagnetic
Ferromagnetic

It
1.0000004
1.00002
1.0008

Silver

Aluminum
Palladium
2-81 Permalloy powder (2 Mo, 81 Ni)t

Cobalt
Nickel
Ferroxcube 3 (Mn-Zn-ferrite powder)
Mild steel (0.2 C)
Iron (0.2 impurity)
Silicon iron (4 Si)

78 Permalloy (78.5 Ni)

Mumetal (75Ni, 5Cu, 2Cr)
Purified iron (0.05 impurity)

Supermalloy

By

(5

Mo, 79 Ni)

Relative permeability p.

130

Ferromagnetic
Ferromagnetic
Ferromagnetic
Ferromagnetic
Ferromagnetic

250
600
1,500
2,000
5,000

Ferromagnetic
Ferromagnetic
Ferromagnetic
Ferromagnetic
Ferromagnetic

7,000
100,000
100,000
200,000
1,000,000

definition.

Remainder is iron and impurities.

X Percentage composition.
Used in transformer applications with continuous tape-wound (gapless) cores.

MAGNETIC DIPOLES AND MAGNETIZATION

6-5

A loop of area A
distance

Qm

I,

from

where

Qm

the magnetic

with current / has a magnetic

this
is

moment of IA.

The

fields at

loop are identical with those of a bar magnet of dipole

the magnetic pole strength and

moment of the bar

is

/ is

a large

moment

the pole separation, provided

equal to that of the loop (see Fig. 6-2). Thus,t

QJ = IA

(1)

Magnet
Area

FIGURE

Bar magnet of moment QI and equivalent current

t This equivalence

maximum
torque

T=

6-2

loop of

moment

Loop

IA.

may be shown from the fact that a current loop in a field B has a
T= IAB [from (5-10-4)] while for a bar magnet the maximum

torque

QIB.

It

follows that for equal torques the

moments must be

equal.

202

ELECTROMAGNETICS

[CHAP. 6

Axis of rod

Area

rr~B
tr~m

t~i

IE3

is
m is~si i Hi
fj51 issi rea [jsi

[j~~si

A-

it~si isi ii~si issi i7~si

esi
gsi

gsi

i;si s~si
ijsi ijsi

-*

FIGURE

6-3

Uniformly magnetized rod with elemental magnetic dipoles.

It was Ampere's theory that the pronounced magnetic effects of an iron bar
occur when large numbers of atomic-sized magnets associated with the iron atoms

are oriented in the

same

is

not important

containing large numbers of them.

as miniature current loops.

Thus, they

In either case,

we

if

be regarded as tiny magnets or

sufficient to describe

Consider the long iron rod shown in cross section in Fig.

same

direction, as suggested in the figure.

as one of uniform magnetization.

The

dipoles) can be conveniently described

which

defined as the magnetic-dipole

is

effect

them by

6-3.

This state of affairs

Assume

meter

(Am -1

that

and are oriented

may be described

of the atomic magnets (or magnetic

by a quantity called the magnetization

moment

M,

Thus

per unit volume, f

IA
=
M = mV = QJ
V
V
where

their

Q m l or as IA.

the atomic magnets are uniformly distributed throughout the rod

in the

precise

confine our attention to regions

may

it is

magnetic moment, which can be expressed either as

all

The

direction so that their effects are additive.

nature of the tiny magnets

(2)

Q m I is the net magnetic (dipole) moment in volume v.

Magnetization has the dimensions of both magnetic-dipole moment per volume
and of magnetic pole strength per area (TL 2 /L 3 = I/L). It is expressed in amperes per
).

The value of

in (2) is an average for the volume v.

To define
at a point,
convenient to assume that the iron rod has a continuous distribution of infinitesimal magnetic dipoles, i.e., a continuous magnetization, whereas the dipoles actually
it is

are of discrete, finite

size.

Nevertheless, the assumption of continuous magnetization

leads to

no appreciable error provided that we

taining

many magnetic

value of

At>

at

restrict our attention to volumes conThen, assuming continuous magnetization, the

dipoles.

a point can be defined as the net dipole moment

of a small volume

divided by the volume with the limit taken as At; shrinks to zero around the point.

The magnetization

is

analogous to the

electric polarization

(Sec. 3-5).

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

SEC. 6-6]

203

Thus

M=

lim
a->o

If

is

(Am" 1 )

^av

known as a function of position

moment of the rod is given by

(3)

in a nonuniformly magnetized rod, the

total magnetic

(Am 2 )

m = \ Mdv
where the integration

EXAMPLE
dipoles

If the

is

volume of the rod.

long uniformly magnetized rod of Fig. 6-3 has N' elementa magnetic

of moment Am,

solution

carried out over the

(4)

From

find the magnetization of the bar.

(2) the

magnetization

is

M = Am = N" Am
v

where

M = magnetization, Am -1
volume, number m"
N" = N'/v = elemental dipoles per
both an average value and also
case the magnetization M
3

unit

In this

is

rod since the magnetization

6-6

UNIFORMLY MAGNETIZED ROD AND

is

the value anywhere

assumed uniform.

in the

EQUIVALENT AIR-FILLED SOLENOID

Consider that a long bar magnet consists of miniature current loops as in Fig. 6-4,
one in place of each tiny bar magnet of Fig. 6-3, with the moment IA of each loop
equal to the

moment Qm I of each magnet. Assuming that there are n

we have

loops in a single

cross section of the rod (as in the end view in Fig. 6-4),

nA'

=
A=

where A'
Further,

let

=A

area of elemental loop

cross-sectional area of rod

us assume that there are

the side view in Fig. 6-4).

N such sets of loops in the length of the rod (see

Then

nN =
where n

= number

N'

of loops in a cross section of rod

N = number of such sets of loops

N'

(1)

= total number

of loops in rod

(2)

204

ELECTROMAGNETICS

[CHAP. 6

^-

Side view

Uniformly
rod

End view

^magnetized

DDDDl

aDBDDO

.COCO

V>a?
Area

Wsets

loops

of loops

FIGURE

6-4

Uniformly magnetized rod with elemental current loops.

It

follows that the magnetization

M of the rod

is

given by

m _ N'lA' NI nA' _ NI
M = ~v~
LA
~L^'~L- K
= equivalent sheet current
L = length of rod, m

where K'

density

(3)

on outside surface of rod,

Am -1

Referring to the end view of the rod in Fig. 6-4, it is to be noted that there are
equal and oppositely directed currents wherever loops are adjacent, so that the
currents have no net effect with the exception of the currents at the periphery of the

As a

rod.

result there is the equivalent

of a current sheet flowing around the rod, as

suggested in Figs. 6-4 and 6-5a.

Although the

sets

This sheet has a linear current density K' (A m" 1 ).

of current loops are shown for clarity in Fig. 6-4 with a large

spacing, the actual spacing

is

assume that the current sheet

of atomic dimensions, so that macroscopically

continuous.

This type of a current sheet

solenoid with

many

ticular Fig. 5-1 5d).

is

effectively

what we also have

NI

(Am"

= number of turns in solenoid, dimensionless

= current through each turn, A
L = length of solenoid, m

TV

The

in the case of

turns of fine wire, as in Fig. 6-56 (see also Fig. 5-15 and in parThe actual sheet-current density
for the solenoid is

K=
where

we can

is

sheet-current density

K is expressed in amperes per meter.

(4)

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

SEC. 6-7]

205

Rod

))ffl-));))ffl)))))

'f

(a)
Solenoid

FIGURE

Q3MSMB

6-5

Uniformly magnetized rod and

(a)

(b)

(b)

equivalent solenoid.

If the solenoid
6- 5a,

and

K=

if

B = HqH;

of Fig. 6-56

is

K', the solenoid

and noting

(5-11-5),

the
is

same length and diameter

we have

In air

at the center of the solenoid that

NI

= iioK=n

B=n

as the rod of Fig.

the magnetic equivalent of the rod.

(T)

(5)

At the center of the rod

B=
From

(5)

fi

K'

= fi

(T)

=
we have H = K, and from (6) we have
we note that H is perpendicular to K and

Vectorially

more

6-7

ft

is

These are scalar

perpendicular to K'.

is

relations.

The

ft

and

K'

Mx

ft

(7)

the unit vector normal to the plane containing the field vectors.

THE MAGNETIC VECTORS,

Consider an

From

K'.

general vector relations are

K=
where

(6)

air-filled

toroid of area

B, H,

and radius

AND M
R

with JV

turns, as in Fig. 6-6.

(5-13-3)

^0=^0
But from

(6-6-5) iV

1/2nR

1
(1)

2nR

K, the sheet-current density.

B = no K = n

Hence,
(2)

206

ELECTROMAGNETICS

[CHAP. 6

Cross-sectional

area

A
Aircore

FIGURE

6-6

Toroid with
turns of coarse winding
(producing B ) and N turns of fine
winding (producing Bm ).

N turns
Nm turns

same winding were placed on an iron ring of the same area and radius, the value
would increase. Imagine now that in place of an iron ring we have the same
air-filled toroid of Fig. 6-6 but with the ring's effect replaced by another toroidal
If the

of

winding of

Nm turns.

The magnetic

field

from

Bm = Ho
But

this

(3)

Nm l/2nR = K', the equivalent sheet-current density.

follows that the total

B is

Hence, from (6-6-6)

(4)

given by

B = B + Bm = n (H+M)

(5)

B = ii (H + M)

or

where

B=

is

2nR

Bm = n K'=n
It

winding

magnetic flux density,

(6)

H = magnetic field, Am -1

M = magnetization, Am -1
Although developed for the
applies in general.

special case of a toroid, (6)

Dividing by

is

a (vector) relation which

H gives
M =

^o(l+^)

(7)

Dividing by the permeability n

l*r

The

207

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

SEC. 6-7]

ratio

M/H is also

gives

or

M
=

ff

/i r

,
(8)

sometimes written

= Xm
where Xm
have

is

(9)

Comparing

the magnetic susceptibility (dimensionless).

(8)

and

(9),

=Hr~l

Xm

(10)

Thus, the magnetic susceptibility x m = for vacuum, for which fir = 1.

and
are in the same direction, so that their quotient
In isotropic media

scalar

and hence

\i is

in general, not in the

B = /* (H + M)

In nonisotropic media, such as crystals,

scalar.

same

direction,

and

\i is

a general relation, while

is

we

not a scalar (see Chap.

B = //H

is

M and H

is

are,

Hence,

15).

a more concise expression,

which, however, has a simple significance only for isotropic media or certain special
cases in nonisotropic media.

single iron crystal is nonisotropic, but

most iron specimens

consist of

an aggre-

gate of numerous crystals oriented at random, so that macroscopically such specimens

may

be treated as though they were isotropic.

In such cases

B = //H

can be applied

as a strictly macroscopic, or large-scale, relation.

Since

V B=

0,

we

have,

on taking the divergence of

V-H=-VM
If the divergence of

We

recall

a vector

field is

from the polarized

(6),

(Am -2)

not zero, the

field

dielectric case (Sec.

(11)

has a source, or place of origin.

3-26) that

V-P= pp

indicates that the polarization field originates

on

volume density pp)

In an analogous manner,

that the

at the dielectric surface.

which

the polarization charge (of apparent

(1 1)

indicates

H field originates where the magnetization field M ends and that the H field

ends where the

The

field originates.

locations where

This occurs at the ends of the rod in Fig.

H (or V

M)

is

not zero

tions of the magnetic poles of a magnetized object.

magnetized rod, as in Fig.

6-3, are at the

may be

6-3.

regarded as the loca-

Thus the poles of a uniformly

end faces of the rod.|

t In ordinary magnets with flat ends the magnetization tends to be nonuniform near the
edges. Entirely uniform magnetization is possible in spherically or elliptically shaped

magnetic objects. However, the assumption of uniform magnetization is a good

approximation for a long homogeneous rod magnet, since the magnetization is nearly
uniform over all of the rod except near the edges at the ends of the rod. In actual
magnets with flat ends the effective separation between the pole centers is slightly less
than the physical length of the magnet.

208

ELECTROMAGNETICS

[CHAP. 6

Taking the curl of

(6),

we have

V x B = n (V x H) +

V x B = Ho J +

or

Where

there

is

no magnetization,

ju

jio(V

(V x

(13) reduces to

x M)

(12)

M)

(13)

V xB=

/i

The

J, as in (5-21-8).

has the dimensions of current density (amperes per square meter) and
-2
represents the equivalent current of density J' (Am
) flowing, for example, in a very

curl of

around the

thin layer

cylindrical surface

current density for this sheet

is

of a uniformly magnetized rod.

K' =J' Ax (A

Thus

layer of current of average density J'.

VxB=

m _1

(13)

/io(J

),

where Ax

is

The

linear

the thickness of the

becomes
J')

(14)

= actual current density, as in current-carrying wire, Am -2

-2
J' = equivalent current density, as at the surface of magnetized bar, Am
The flux density B is always the result of a current or its equivalent. For
example, the magnitude of B at the center of a long slender iron rod surrounded by a

where J

long solenoid

is,

from

(6-6-5)

and

(6-6-6),

B = (K+K')

(T)

(15)

K = sheet-current density due to solenoid current, Am -1

where

K'

In

many

producing

equivalent sheet-current density due to magnetization of rod,

it,

potential A,

cases

we can

as in (15).

which

conveniently express

In general,

we can

directly in terms

also express

Am -1

of the currents

in terms of the vector

Thus

in turn is related to the currents.

B=VxA
If

(16)

both conduction currents and magnetization are present,

fi

fJ+

J'

=s/^*

A:

<">

where

=VxH

EXAMPLE 1
area A -nr 1
.

The coil

is

and

J'

=Vx

(A

m~ 2)

Referring to Fig. 6-7a, a toroidal coil has a radius

The

coil

R and

made of many turns N of fine insulated wire with a current /.

ing the variation of B,

a cross-sectional

has a very narrow gap, as shown in the gap detail in Fig. 6-8o.

M, H, and fi along the line

of radius

JR at

Draw graphs show-

the gap (centerline of

coil).

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-7]

^Air core

'Iron ring
with winding

Iron ring

FIGURE

6-7

"

Toroidal coil with gap. (ft) Permanently magnetized iron ring with gap.
cored toroidal coil with gap.
(a)

solution

Neglecting the small effect of the narrow gap,

around the inside of the

209

Since

entire toroid.

R > r,

its

magnitude

is

(c)

Iron-

substantially uniform

is,

from

(5-11-5), given

approximately by

NI
2nR

B = Ho
Where

is

No

l*

(T)

(18)

the magnitude of the linear sheet-current density, in amperes per meter.

of the magnitude

M=

''

along the centerline of the coil at the gap

ferromagnetic material

0, as indicated in Fig. 6-8c.

Since

M=

0,

is

shown

is

present, so that the magnetization

It

follows that

M=

and also

A graph

in Fig. 6-86.
is

negligible

and

V H = 0.

we have

f*o
1*0

Therefore, the magnitude of

is

K K

f*0

NI
2nR

(Am-

constant and equal to the sheet-current density

(19)

K of the

The permeability everywhere is /* (Fig. 6-8e).

It is to be noted that B is continuous and that in this case H is also continuous since
there is no ferromagnetic material present. Both B and H have the same direction everywhere
coil winding, as indicated in Fig. 6-8rf.

in this case.

EXAMPLE
of the same

2
size

now that the toroidal coil of Example 1 is replaced by an iron ring

and also with a gap of the same dimensions, as suggested in Figs. 6-76 and

Consider

210

[CHAP. 6

ELECTROMAGNETICS

Af

(C)

M=0

JK
H

(d)

(e)

FIGURE

6-8

Magnitudes of magnetic quantities along the

(see Fig. 6-7a) (Example 1).

Assume

6-9a.

magnitude to
/*,

and

that the ring has a uniform

K for the toroid in Example

V H along

solution
the right side.

around the

permanent magnetization

ring has a north pole at the

left

side of the

Neglecting the small effect of the narrow gap,

and also across the

K' on the surface of the ring.

interior of the entire ring

M and K'

Hence,

that

in

is

a toroid

equal in

Draw graphs showing the variation of B, M, H,

B = noM = HaK'
1.

gap

the centerline of the ring at the gap.

The

equivalent sheet-current density

where

coil centerline at the

gap.

both examples.

Its

gap and a south pole at

is substantially uniform
It is

due

entirely to the

From (6-6-3), K'

(T)

are, according to the stated conditions, equal to

B is the same in

= M.

Thus
(20)

K for the solenoid in Example

value at the gap

is

illustrated in Fig. 6-96.

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-7]

211

North pole

FIGURE

6-9

Variation of magnetic quantities along the centerline at the gap in a permanently

magnetized iron ring (see Fig. 6-76) (Example 2).

M = K', but outside the ring and the gap Af = Suppose the
M from zero K' the gap occurs over a short distance Ax rather than as a
The graph for M then as shown

In the ring,

change

in

in

to

0.

at

square step function.

in Fig. 6-9c.

is

M = 0; so
B =
H=
K'

Outside the ring and in the gap

(Am-

that

212

[CHAP. 6

ELECTROMAGNETICS

Inside the ring

H = -M

(Am-

(21)

1*0

or approximately

The

tive.

From

0.

is

and

In the air gap

negative.

is

According to
this equals the

p.

and negative because

The

=p.

variation of

(6-7-11), the divergence of

H equals

the negative divergence of

(A

mm"

small

is

that

M changes linearly

in

magnitude over

this thickness,

Ax

also that

we have on

(23)

v.H=4^-' = P-

or

( 24 >

Ax

where the upper sign

in front of K' applies if

M decreases and H

increases in proceeding across

The variation of V H along the

pole volume density p m has a value only in

in a positive direction (from left to right).

Hence

illustrated in Fig. 6-9/

of assumed thickness

Ax

at the sides of the gap,

the

at the sides

and for

this

both examples.

is

the

centerline

the layers

This locates the poles of the ring magnet

of the gap.

Km Example

In the gap, the directions of

in the toroid

reason the iron surfaces of the gap are called pole faces.

Since K'm this example equals

in

Ax is very

the centerline

M = dx = jAx = - p m

Thus

Assuming

at the gap.

and assuming

the cross-sectional diameter Ir of the ring,

compared with

same as

M, and

(22)

zero everywhere except at the layers of assumed thickness

This

Ax

is

across the

p.

apparent pole volume density p m in the ring on both sides of the gap.

V-H=-V-M=p m

is

small and nega-

suggested in Fig. 6-9e.

is

small

is

6-9rf.

is

(6-7-7) the permeability in the ring is large

compared with
gap

H = K' K' = The exact value of H not zerot but

H across the gap illustrated in Fig.

variation of

of Example

1,

but

1, B and H in the gap have identical values

B and H are the same. In the iron ring, B

smaller and in the opposite direction.

is

An H direction opposite to that of B is characteristic of conditions inside a permanent magnet.

EXAMPLE

Suppose now that the iron ring of the previous example has wound over it
Example 1 with the gap in the toroid coinciding with the gap in the ring,
as shown in Fig. 6-7c and in the gap detail of Fig. 6-10a. The combination constitutes
an iron-cored toroid as contrasted with the air-cored toroid of Example 1. Let the sheet
current density for the toroid winding be A' as in the first example. Further, let the induced
magnetization added to the permanent magnetization in the ring yield a total uniform magnet3

the toroidal coil of

ization (permanent

the variation of B,

and induced) that

p., and V

M, H,

The above

is

equal in magnitude to 4K.

Draw

graphs showing

H along the centerline of the ring at the gap.

analysis is approximate since

it

neglects the effect of the gap.

See Sec. 6-23.

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-7]

213

XM=J'

V-H=
-V-M= ft

FIGURE

6-10

Variation of magnetic quantities along coil centerline at the gap in an iron-cored

toroid (see Fig. 6-7c) (Example 3).

solution
effect

ring

In this case the total magnetization

of the narrow gap, the flux density

and across the gap.

It is

that

M changes

M = 4K, and

is

M = K' = 4K.

substantially uniform

Neglecting the small

around the

inside of the

given (see Fig. 6-106) by

B=
In the ring

is

fi.

in the

(K + K')
gap

M=

linearly over a short distance

= 5/* K
0, as

(T)

shown in

Ax at

Fig. 6-10c.

the pole faces.

(25)
It is

again assumed

214

ELECTROMAGNETICS

[CHAP. 6

In the gap

H= =5K

(26)

Mo
In the ring

H = -M
and so we have very nearly
depicted in Fig.

6-10rf.

that

(27)

H = 5K AK K.

In the gap fi=fj.

The

variation of

= /*o(l + 50 = /io(l+^ =

/*

H across the gap

is

In the ring (see Fig. 6-10e)

(28)

5/tio

The divergence of H or pole volume density /> is given by the negative of the divergence
This has a value of 4K/Ax over the assumed pole thickness Ax at the pole faces.
This is illustrated in Fig. 6-10/ The fact that V H = m at the pole faces is also indicated
p
of

M.

H = 0.
In this example, B and H have the same direction
In the ring, however, H
weaker than in the gap.
Elsewhere

in Fig. 6- 10a.

both

in the

gap and

in the ring.

is

In this example, the toroid has a sheet-current density of

has an equivalent sheet-current density around

0.

curved surface of K'

),

and the ring

= 4K(Am-

).

H = J (A m" as suggested in Fig. 6-10a. Elsewhere

At the curved surface of the ring Vx M = J'(Am"'). Elsewhere V x M = 0.

Inside a wire of the toroidal coil

VxH =

its

K (A m"

Vx

In the last two examples involving ferromagnetic material

the magnetization, or

M,

lines originate, or

it is

to be noted that

have their source, on a south (negative)

The H lines, on the other

on a north pole and end on a south pole. Thus,

pole and end on, or have as a sink, a north (positive) pole.

hand, originate, as in Example

2,

H has a positive value at a north pole, while V M has a positive value at a south

pole.

As a final example

let

us compare the

fields

around a solenoid and the equivalent

permanently magnetized rod.

EXAMPLE

long uniform solenoid, as in Fig. 6-1 la,

is situated in air and has NI

permanently magnetized iron rod (Fig. 6-1 le) has the same
dimensions as the solenoickand has a uniform magnetization
equal to Nil I for the solenoid.
Draw graphs showing the variation of B, M, and
along the axes of the solenoid and the
rod. Also sketch the configuration of the fields for the two cases.

ampere-turns and a length

/.

solution

K = NI/l,

Since the rod and solenoid have the

the two are magnetically equivalent.

The B

same dimensions and

fields for

M = K' =

both are the same every-

215

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-7]

P le

wmm-*

-North

/Magnetized rod

South.

^VXH=J

^-Solenoid

/yxM-j/ P "

^=^S
-TT~~-V- H= - VM
V-H=-V-M=-/>, |-^t

^^(/)
M

M=0
I.

(C)

JL
FIGURE

(flO

(d)

(h)

7^

6-11

Solenoid and equivalent magnetized rod showing

where, and the

H fields for both are the same

the solenoid

long compared with

is

its

fields

along axis (Example

Assuming

outside the solenoid and rod.

diameter, the flux density at the center

is

4).

that

nearly given

by

NI

At the ends of

(29)

the solenoid

B = i/x K
The magnitude of B

(30)

at other locations along the solenoid axis

with a suitable change in

The

limits.

can be obtained from

variation of B along the solenoid axis

is

(5-1 1-3)

shown graph-

The variation along the rod axis is the same (Fig. 6-11/).
= everywhere (Fig. 6-1 lc). In the rod the magnetization
For the solenoid case,
assumed to be uniform, as in Fig. 6-llg.
= K at the center and = iK
For the solenoid case, H = B//x everywhere, so that

ically in Fig. 6-116.

is

at the ends.

the rod,
tion

is

is

The
the

variation of

same as for

along the solenoid axis

as suggested in Fig. 6-11 h.

distance

Ax

at the

Inside the rod

the solenoid.

ends of the rod.

It is

assumed

The

that

direction of

is

shown

in Fig. 6-1 Id.

Outside

H = (B/fio) M, so that the varia-

M changes from

H in the rod

is

to

K over a short

opposite to that for B.

V x H = J, as indicated in Fig. 6-1 la. On the

V x M = J', as suggested in Fig. 6-1 \e. In the solenoid case,
V H = everywhere. In the rod case, V B = everywhere, but V H =

Inside the wires of the solenoid winding,

cylindrical surface of the rod,

V B = and
_ V M = p at
-

the

end faces of the rod.

216

ELECTROMAGNETICS

[CHAP. 6

field

<^r^>

*~= Solenoid :S

field

field

zero

Solenoid

v*

(air core)

(air core)

-(a)

field

Solenoid

=5

(air core)-

(b)

-(c)-

Rod
--N

()

FIGURE

6-12

and equivalent permanently magnetized rod. The B fields are the

and rod [see (a) and (</)]. The
field is zero everywhere
except inside the rod [see (6) and ()]. The H fields are the same outside both
solenoid and rod but are different inside [see (e) and (/)].
Fields of solenoid

same

for both solenoid

H fields for the two cases are sketched in Fig. 6-12.

H directed from the north pole to the south pole.

The B, M, and
that inside the rod

is

have, in general, different directions in the rod,

case.

Here

ju,

loses

its

to be noted

Since

M and B

simple scalar significance in this

can be obtained by vector addition, using (6-7-6). As an example,

at
6-12/ is obtained by the vector addition of B//i and
as in Fig. 6-12#.
Although the magnetization is based on the actual magnetization phenomenon,
often simpler and more convenient to ignore the mechanism of the phenomenon

the point Pin Fig.

it is

It is

and use the permeability n to describe the characteristics of the magnetic medium.
This is particularly true where \i can be treated as a scalar. In this case
n is determined experimentally from a sample of the material. However, since ft is not a

constant for ferromagnetic materials but a function of

and also the previous history
of the sample, the methods for dealing with ferromagnetic materials require special
consideration (see Sec. 6-10 and following sections).

6-8

BOUNDARY RELATIONS

In a single medium the magnetic field is continuous. That is, the field, if not constant,
changes only by an infinitesimal amount in an infinitesimal distance. However, at
the boundary between two different media, the magnetic field

both in magnitude and direction.

may change

abruptly

THE STATIC MAGNETIC HELD OF FERROMAGNETIC MATERIALS

sec. 6-8]

FIGURE

217

6-13

Construction for developing continuity

relation for

It is

normal component of B.

convenient to analyze the boundary problem in two parts, considering

separately the relation of fields normal to the

Taking up

first

boundary and tangent to the boundary.

the relation of fields normal to the boundary, consider two media

of permeabilities ^j and n 2 separated by the xy plane, as shown in Fig. 6-13. Suppose

an imaginary box is constructed, half in each medium, of area Ax Ay and height

that

Az.

Let

Bnl

be the average component of

B normal to

the top of the box in

medium

and B 2 the average component of B normal to the bottom of the box in medium 2.
Bnl is an outward normal (positive), while Bn2 is an inward normal (negative). By
1

Gauss' law for magnetic

zero.

fields (5-8-1), the total

magnetic flux over a closed surface

In other words, the integral of the outward normal components of

closed surface

is

is

over a

By making

the height Az of the box approach zero, the conbox to the surface integral becomes zero even though
components of B normal to the sides. Therefore the surface

zero.

tribution of the sides of the

there

may be

finite

integral reduces to

Bnl Ax Ay

B 2

Ax Ay =

or

0)

B.-

According to

(1),

the

normal component of the flux density

is

continuous across the

boundary between two media.

Turning

now

to the relation for magnetic fields tangent to the boundary, let

two media of permeabilities

.6-14.

fi t

and n 2 be separated by a plane boundary, as

Consider a rectangular path, half in each medium, of length

Ax

in Fig.

parallel to the

boundary and of length Ay normal to the boundary. Let the average value of
medium 1 be tl and the average value of
tangent to

tangent to the boundary in

the boundary in

medium 2 be

t2

According to

a closed path equals the current /enclosed.

(5-14-5), the integral of

By making

H around

the path length' Ay approach

218

ELECTROMAGNETICS

[CHAP. 6

-Ax1

tl

'***'
f

T* *** Medium 2

Current sheet of
-.linear density
perpendicular

'

to

FIGURE

Medium

H hui

Ay

'.'".

'':

.'.'.

'.
;

P}.

'.'

;'":'.'.':'''.''.'.'
::'.''.'. ..:.:.:
'.'.'.:''.'

''.

page

6-14

Construction for developing continuity relation for tangential component of H.

zero, the contribution of these segments of the path

becomes zero even though a

may exist normal

Hn Ax Ht2 Ax = I or

line integral

Hn -H

t2

where

The

to the boundary.

finite field

=K

(Am"

Ax

then reduces to

(2)

the linear density of any current flowing in an infinitesimally thin sheet at

is

the surface, f

According to
is

(2),

the change in the tangential component of

equal in magnitude to the sheet-current density

that

K is normal to H;
This

to the page.

is

that

is,

H across a boundary

K on the boundary.

It is

to be noted

the direction of the current sheet in Fig. 6-14

is

normal

expressed by the vector relation

ftxH=K
where A

unit vector normal to boundary, dimensionless

H = change in magnetic field

K = sheet-current density at
If the field

related to
is

zero

below the boundary

across boundary,
-1

boundary,

is

zero

0), the

K = 0,

current

K related to

H, 2

Am"" 1

Am

(Ht2 =

0), (3) indicates that

Hn will be into the page in Fig. 6-14, while

(Hn =

If

(3)

will

if

the current

the field above the boundary

be out of the page.

then

Hn - H,.

(4)

t If J is the current density in amperes per square meter in a thin sheet of thickness
Ay', then
is defined by

AT

where J-*- oo as Ay'-* 0.

= /A/

(Am"

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-8]

219

are continuous across the boundary

(4), the tangential components of
between two media provided the boundary has no current sheet.

According to

If

medium

1 is

a nonconductor, and

if

Hn = K
where

2 is

0,

(5)

the sheet-current density in amperes per meter in

ary (into the page in Fig. 6-14).

(5) is

t2

When medium

is

air

medium 2

and medium 2

at the

is

bound-

a conductor,

approximated at high frequencies because the skin effect restricts the current
its surface (see Chap. 10).

in the conductor to a very thin layer at

EXAMPLE

as in Fig. 6-15.

Consider a plane boundary between two media of permeability /x, and fi 2 ,

Assume that the media
Find the relation between the angles a t and <x 2
.

are isotropic with

solution

and

From

H in the same direction.

the boundary relations,

and

H,i

cos ai

and

B 2

H = Hi sin a.

and

H, 2

Bml

From

n.

H,

(6)

Fig. 6-15,

Bnl Bi
tl

= B2 cos a 2
= Hi sin a 2

(7)
(8)

magnitude of B in medium 1
B2 = magnitude of B in medium 2
Hi magnitude of H in medium 1
H2 = magnitude of H in medium 2

where Bi

Substituting (7)

and

(8) into (6)

and dividing

yields

tan ai
(9)

tan a 2

fr2

Normal to-*boundary

j
FIGURE

6-15

Boundary between two media of

ent permeability showing change
rection of magnetic field line.

differ-

in di-

yi
/

Medium

Boundary

Medium 2
>2

Field line

BorH

220

ELECTROMAGNETICS

[CHAP. 6

B
Air

:'.':.'

..".-

FIGURE
B

!.'.'.;'.;'.'

':

'

Iron

'..'.':

:>b

6-16

lines at air-iron

(a)

boundary.

(b)

= relative permeability of medium 1, dimensionless

= relative permeability of medium 2, dimensionless

where pn
fi T 2

Equation
at the

.'

(9) gives the relation between the angles a t and

boundary between two media, f

EXAMPLE

Referring to Fig. 6-16,

medium

let

iron with a relative permeability of 7,000.

boundary (a 2 = 0),
a 2 = 85, find ai

find a,,

(b) If

(a) If

be air (p r

in the iron

1)

is

<x

for

H lines

and

and medium 2 be soft

normal on the

incident

in the iron is nearly tangent to the surface at

an angle

solution

(a)

From

(9)

tan a!

tan

When

<x 2

(b)

<x 2

7.000

Pr2

tan a 2

(10)

= 0, a! = 0, so that the B line in air is also normal to the boundary (see Fig.

6-16a).

When 2 =85, we

the direction of B in air

tion in the iron

is

have, from (10), that tan cd =0.0016, or ai =0.1. Thus,

almost normal to the boundary (within 0.1) even though its direc-

nearly tangent to the boundary (within 5) (see Fig. 6-166). Accordingly,

for many practical purposes the direction
in air or other medium of low relative
permeability may be taken as normal to the boundary of a medium having a high relative
is

o/BorH

permeability.

This property

is

reminiscent of the one for

or

at the

boundary of a

conductor.

B and
have the same direction (/* a scalar). In the
absence of magnetization, as in air, B and
have the same direction. When
magnetization is present, as in a soft-iron electromagnet, B and
also tend to have
the same direction. However, this is not the situation in a permanent magnet.

t This relation applies only if

6-9

221

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-10]

TABLE OF BOUNDARY RELATIONS

FOR MAGNETIC FIELDS

Table 6-2 summarizes the boundary relations for magnetic fields developed in Sec.

Table 6-2

BOUNDARY RELATIONS FOR MAGNETIC

6-8.

FIELDSt

Field

component

Boundary relation

Normal

Bi

Normal

\i^iHi

Tangential

H H K
AxH = K

Tangential
Tangential

= Bi

tl

= \l,iHi
t2

H = Hti
H =K
ti

tl

Condition
(1)

Any two media

(2)

Any two media

(3)

Any two media

at

(3a)

(4)

Any two media

(5)

H,2

t These relations apply for both static

with current sheet of infinitesimal thickness

boundary
with

no current

sheet at

boundary

0; also medium 2 has a current sheet of infinitesimal

thickness at boundary;
tl and
2 are normal to each
other

and time-varying

fields (see

Chap.

8).

FERROMAGNETISM

6-10

Although magnetic

known

effects in

most substances are weak, the group of substances

as ferromagnetic materials exhibits strong magnetic effects.

of these materials

is

not a constant but

is

The permeability
field and of

a function both of the applied

the previous magnetic history of the specimen.

In view of the variable nature of the

permeability of ferromagnetic materials, special consideration of their properties

is

needed.
In ferromagnetic substances the magnetic effects are produced by the motion

of the electrons of the individual atoms.

The

net effect

is

stance such as iron these atomic magnets over a region of

to

make an atom of a

In a ferromagnetic sub-

ferromagnetic substance act like a miniature bar magnet.

many atoms

tend to orient

themselves parallel to each other, with north poles pointing one way.
called a magnetic

domain and

is

spontaneously magnetized.

The

size

This region

is

of a domain

depends on conditions but usually contains millions of atoms. In some substances

the shape appears to be like a long, slender rod with a transverse dimension of microscopic size but lengths of the order of a millimeter or so.
small, but not atomically small, bar magnet.

Thus, a domain acts

like

222

ELECTROMAGNETICS

FIGURE

[CHAP. 6

6-17

Six directions of easy magnetization in

an iron

crystal.

In an unmagnetized iron crystal the domains are parallel to the direction of

easy magnetization, but since as

many have north

other, the external field of the crystal

tions of easy magnetization.

That

is,

way

poles pointing one

there

is

a positive and negative direction

along each of the three mutually perpendicular crystal axes (Fig. 6-17).
the polarity of the domains in an unmagnetized iron crystal

the highly schematic diagram of Fig. 6- 18a.

as the

In an iron crystal there are six direc-

zero.

is

single

may

Therefore

be as suggested by

N represents a domain with a

north pole pointing out of the page and a single S a domain with a south pole pointing
out of the page.

If the crystal is

placed in a magnetic

one of the

field parallel to

domains with polarity opposing or perpendicular

to the field become unstable and a few of them may rotate so that they have the same
With further increase of the field more domains change over,
direction as the field.
directions of easy magnetization, the

each as an individual unit, until when

magnetic saturation

magnetized to a

is

maximum

extent.

tions after the applied field

netized.

all

the domains are in the

reached, as suggested by fig. 6-186.

is

If the majority of the

removed, the specimen

is

In fact,

state.

if

the temperature

is

same

direction,

crystal

domains retain

is

then

their direc-

said to be permanently

Heat and mechanical shock tend to return the

magnetized

The

mag-

crystal to the original un-

raised sufficiently high, the

themselves are demagnetized and the ferromagnetism disappears.

This

is

domains

called the

Curie point (about 770C for iron).

Magnetization which appears only in the presence of an applied field may be

spoken of as induced magnetization, as distinguished from permanent magnetization,
which is present in the absence of an applied field.

6-11

MAGNETIZATION CURVES

The permeability

p.

of a substance

is

V-

given by

B
jj

HO

V-r

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

SEC. 6-11]

223

S--N

N-

t
S

S--N

N*-S

(a)

t
S

S-*N S--N S-*N S--N

FIGURE
(a)

6-18

Domain

polarity in

Arrows

iron crystal.

an unmagnetized

S--N

S-N

S-*-N S-*-N
(b)

indicate direction

single
represents
of magnetization.
a domain with a north pole pointing out
of the page; a single S represents a domain with a south pole pointing out of

S-N

S--N S--N S--N

S --N

S--N

S-N S*-H

Condition after crystal is

saturated by a magnetic field directed to
the page.

(6)

Applied magnetic field

the right.

B=

where

magnitude of flux density,

H = magnitude of field H, A m~
=
H =

fi

permeability of

vacuum = 400 n nH

relative permeability

m _1

of substance, dimensionless

The permeability \i or the ratio B/H is not a constant for ferromagnetic materials.
Therefore, to illustrate the relation of B to H, a graph showing B (ordinate) as a
The line or curve showing B as a function of
function of
(abscissa) is used.

on such a

BH chart is called a magnetization curve.

slope of the curve, which

is

given by dB/dH, but

is

It is

to be noted that

fi is

not the

equal to the ratio B/H.

a magnetization curve for an iron sample, a thin, closed ring may

uniform primary winding is placed over the ring, forming
be cut from the sample.
an iron-cored toroid, as in Fig. 6-19. If the number of ampere-turns in the toroid is

To measure

NT, the value of

H applied to the ring

NI

H= T
where

= 2nR and R

applied to the ring

is

may

the

mean

is

(A turns

')

radius of the ring or toroid.

be called the magnetizing force.

This value of

is

Hence, in general,

224

ELECTROMAGNETICS

[CHAP. 6

Primary

Secondary

Fluxmeter

Cross sectional
area "*

Iron ring''

FIGURE
Rowland

sometimes called by

this

method of obtaining magnetization

curve.

The flux density B in the ring may be regarded as

and is measured by placing another (secondary)
6-19, and connecting it to a fluxmeter. f
For a given

name.

the result of the applied field

coil

6-19
ring

over the ring, as in Fig.

change in H, produced by changing the toroid current 7, there is a change in magnetic

flux ^ m through the ring.
Both and B are substantially uniform in the ring and neg-

\j/ m = BA, where A is the cross-sectional

area of the ring, and the resulting change in the flux density B in the ring is given by

ligible outside.

B = ^JA

Therefore the change in flux

where

\j/

is

measured by the fluxmeter.

This ring method of measuring

magnetization curves was used by Rowland in 1873.

A typical magnetization curve for a ferromagnetic material is shown by the solid

curve in Fig.

dashed

The specimen

6-20ar.

change was noted

in

lines are also

meabilities

n r of

1, 10,

magnetization curve

is

as

H was

in this case

initially

unmagnetized, and the

By way of comparison, four

given by
5

B=

0.

shown in Fig. 6-20a, corresponding to constant relative per100, and 1,000.

The relative permeability at any point on the

^ = -^=7.9xl0 |
H
HoH
where

was

increased from

ordinate of the point,

(dimensionless)

H = abscissa of the point, Am -1

A graph of the relative permeability

fi r as a function of the applied field H, corresponding to the magnetization curve in Fig. 6-20a, is presented in Fig. 6-206. The maximum
relative permeability, an* therefore the maximum permeability, is at the point
on the

magnetization curve with the largest ratio of

t

to H,

The fluxmeter operates on the emf induced

flux

through

it

changes.

This

is

designated

"max n";

when

the magnetic

in the secondary

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

SEC. 6-11]

225

2|-

10,000

5000

(Am- 1 )

FIGURE

6-20

(a) Typical

magnetization curve and (b) corresponding relation of relative permea-

bility to

applied field

H.

occurs at the point of tangency with the straight line of steepest slope that passes
through the origin and also intersects the magnetization curve (dash-dot line in Fig.

it

6-20a).

The magnetization curve

for

iir

for air or

(almost coincident with the

vacuum would be

H axis) in Fig.

given by the dashed line

6-20a.

The

difference in the

226

ELECTROMAGNETICS

[CHAP. 6

Hard magnetization

FIGURE

6-21

Regions of easy and hard magnetization

of initial-magnetization curve.

ordinate

between the magnetization curve of the ferromagnetic sample and the

ordinate at the same i/ value

on

ferromagnetic material times

fi

the

nr =

The magnetization curve shown

That
is

is,

the material

is

condition

is

M of the

in Fig. 6-20a

is

an

initial-magnetization curve.
field

is

As

applied.

B rises rapidly at first and then more slowly. At sufficiently

H the curve tends to

The magnetization curve

become

flat,

as suggested

by

This

Fig. 6-20a.

starting at the origin has a finite slope giving

initial

fields.

initial-magnetization curve

steep section

an

Therefore the relative-permeability curve in Fig. 6-206 starts with a

permeability for infinitesimal

The

equal to the magnetization

called magnetic saturation.

permeability.
finite

line is

completely demagnetized before the

increased, the value of

high values of

and

(2)

the

flat section,

of the curve (Fig. 6-21).

The

may

be divided into two sections:

the point

P of division

(1)

the

being on the upper bend

steep section corresponds to the condition of easy

netization, while the flat section corresponds to the condition

of

difficult,

mag-

or hard,

magnetization.

Ordinarily a piece of iron consists not of a single crystal but of an aggregate of

The

small crystal fragments with axes oriented at random.

of iron

may

be represented schematically as in Fig. 6-22.

situation in a small piece

Here a number of crystal

fragments are shown, each with a number of magnetic domains, represented in most
cases by

a small square.

by the heavy

lines,

The boundaries between

and domain boundaries by the

direction of the crystal axes.

crystal fragments are indicated

light lines,

In Fig. 6-22a, not only

is

which also indicate the

the piece of iron unmagnetized,

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

SEC. 6-11]

227

Unmagnetized

Partly

magnetized

Applied field

Saturated

FIGURE

6-22

Successive stages of magnetization of a

polycrystalline specimen with increasing
field

Arrows indicate direction of mag-

netization of domains.

An N represents

a domain with a north pole pointing out

of the page; an S represents a domain
with a south pole pointing out of the
page.

but also the individual crystal fragments are unmagnetized. The domains in each
three
crystal are magnetized along the directions of easy magnetization, i.e., along the
crystal axes.

However, the polarity of adjacent domains

magnetization of each crystal

is

is

opposite, so that the total

negligible.

in the direction indicated by the

With the application of a magnetic field
arrow (Fig. 6-22) some domains with polarities opposed to or perpendicular to the

228

ELECTROMAGNETICS

FIGURE

[CHAP. 6

6-23

is reduced when domains

go from (a) aligned condition to (6) oppo-

Total energy

(a)

sitely oriented condition.

applied field

(6)

become unstable and rotate quickly to another direction of easy magsame direction as the field, or more nearly so. These changes take

netization in the

on

place

the steep (easy) part of the magnetization curve.

domains have changed

over,

is

as suggested in Fig. 6-22b.

The

result,

after all

This condition corresponds

P on the magnetization curve (Fig. 6-21).

With further increase in the applied field, the direction of magnetization of the
domains not already parallel to the field is rotated gradually toward the direction of
H. This increase in magnetization is more difficult, and very high fields may be
required to reach saturation, where all domains are magnetized parallel to the field,
roughly to that at the point

as indicated in Fig. 6-22c.

This accounts for the flatness of the upper (hard) part of

the magnetization curve.

The tendency of adjacent magnetic domains to be oppositely magnetized can

be understood from energy considerations. Thus, when adjacent domains are oriented the same, as in Fig. 6-23a, the total energy

is

oppositely oriented, as in Fig. 6-236, the energy

increased.

is

decreased.

When
The

all

domains are

situation can be

by performing an experiment with two bar magnets arranged to slide easily

Let both magnets be placed side by side and oriented the same, as in Fig.
If the left magnet is held but the right magnet is released, it will move to the

illustrated

on a

rod.

6-24a.

As the right magnet moves

on the rod to the position shown in Fig. 6-24c. The opposite
and the right magnet moves back to the left until it comes to rest

right, as in Fig. 6-246, since the adjacent like poles repel.

farther away,

poles

now

it

will rotate

attract,

N N

N S

S S

S N

(b)

( i)

FIGURE

(c)

i)

6-24

Experiment with 2 bar magnets

illustrates the decrease in total energy.

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-12]

Flux density

229

(T)

FIGURE

6-25

Hysteresis loop.

against the

than at the

left

6-24*/.
The pair of magnets now has less energy
The decrease in total energy accounts for the work done
moving away, rotating, and moving back.

magnet, as in Fig.

start (Fig. 6-24a).

magnet

by the

right

6-12

HYSTERESIS

in

If the field applied to a specimen

flux density

tion curve.

is

increased to saturation and

is

then decreased, the

B decreases, but not as rapidly as it increased along the initial-magnetizaThus, when H reaches zero, there is a residual density, or remanence, B

(Fig. 6-25).

In order to reduce

This

is

B to zero,

called the coercive force.

a negative

As

is

field

Hc

must be applied

(Fig. 6-25). f

further increased in the negative direction,

the specimen becomes magnetized with the opposite polarity, the magnetization at
t

By

reversing the battery polarity (Fig. 6-19).

230

ELECTROMAGNETICS

FIGURE

[CHAP. 6

6-26

and hard mag-

Hysteresis loops for soft

netic materials.

first

being easy and then hard as saturation

approached.

is

zero again leaves a residual magnetization or flux density

zero a coercive force

+H

must be applied.

Bringing the

and

field to

to reduce

With further increase

to

in field, the

specimen again becomes saturated with the original polarity.

The phenomenon which causes

to lag behind H, so that the magnetization

curve for increasing and decreasing fields

is

loop traced out by the magnetization curve

the substance

loop

is

is

is

is

called hysteresis,

and the

called a hysteresis loop (Fig. 6-25).

If

carried to saturation at both ends of the magnetization curve, the

called the saturation, or major, hysteresis loop.

on the saturation loop

loop

not the same,

is

is

called the retentivity,^

called the coercivity.

The

residual flux density

and the coercive force

Thus, the retentivity of a substance

is

the

value which the residual flux density can attain and the coercivity the

on

this

maximum
maximum

value which the coercive force can attain.

reached on the

BH

For a given specimen no points can be

diagram outside the saturation hysteresis loop, but any point

inside can.

In

magnetized, materials the hysteresis loop is thin, as suggested in

a small area enclosed. By way of comparison, the hysteresis loop of
a hard magnetic material is also shown, the area enclosed in this case being greater.
soft, or easily

Fig. 6-26, with

Turning our attention to the permeability

Fig. 6-27a.

6-Tlb.

At

p.,

consider the hysteresis loop of

The corresponding graph of p as a function of

= 0, it is apparent that p becomes infinite. On

The term

retentivity is also

density B, to the

maximum

is

as

shown

in Fig.

the other hand,

when

sometimes used to mean the ratio of the residual flux

flux density

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-12]

FIGURE

231

6-27

(a) Hysteresis loop,

(6)

Corresponding

permeability curve.

B = 0, n = 0. Under

such conditions, the permeability u becomes meaningless.

must be confined to situations where it has significance, e.g.,

It is to be noted that the term maximum permeamagnetization
curve.
initial
the
permeability for an initial-magnetization
maximum
specifically
the
bility signifies

Therefore the use of

/i

curve and not for a hysteresis loop or other type of magnetization curve.
Another type of magnetization curve for which \i has a definite meaning

normal

magnetization curve.

This curve

loops, obtained by cycling the field

shown

in Fig. 6-28, the field

saturation hysteresis loop.

range

H over

is

field is

size.

ly

Thus, as

obtaining the

next cycled slowly several times over a

The curve passing through

normal magnetization curve (Fig. 6-28).

is

the

repeated for successively smaller ranges in H, obtaining a series

of loops of decreasing

and

successively smaller ranges.

changed slowly over the range

The

is

the locus of the tips of a series of hysteresis

obtaining after a few reversals a repeatable hysteresis loop of smaller

This process

size.

is

is

This curve

is

the tips of these loops

useful since

it is

characteristic of the particular type of ferromagnetic material.

magnetization curve

is

is

the

reproducible

The normal

actually very similar in shane to the initial-maenetization curve.

232

ELECTROMAGNETICS

[CHAP. 6

Retentivity

FIGURE

6-28

Normal magnetization curve with

6-13

ENERGY

IN

relation to hysteresis loops.

A MAGNET

A specimen of iron with residual magnetization contains

energy since work has been

performed in magnetizing it. The magnetic energy w m per unit volume of a specimen
brought to saturation from an originally unmagnetized condition is given by the integral

of the initial-magnetization curve expressed (see Table 5-2) byf

HdB

wm =

(Jm -3 )

'n

The dimensional

relation for (1)

is

I M _ M
L IT 2 ~ UT 1

t Integrating yields

wm = i/iH 2

as in Table 5-2.

(1)

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-13]

233

(a)

FIGURE

FIGURE

6-29

Energy

Energy density areas for (a) soft and

(b) hard magnetic materials.

6-30

lost

in magnetization cycle

is

proportional to area enclosed by hysteresis loop.

MILT 2

where

has the dimensions of energy density, which

Thus, the area between the curve and the

cubic meter.

energy density.
soft)

This

is

substance which has been carried to the point

larger shaded area in Fig. 6-296.

If

more work

On

bringing

by the crosshatched areas in Fig.

is

is

expressed in joules per

axis

is

a measure of the

indicated in Fig. 6-29a for an easily magnetized (magnetically

magnetically hard substance takes

as indicated

in the magnetization process.

to magnetize, as indicated

H to zero

some energy

is

by the

released,

6-29.

increased and decreased, so that the magnetization of a specimen re-

peatedly traces out a hysteresis loop, as in Fig. 6- 30a, the area enclosed by this loop
represents the energy per unit volume expended in the magnetization-demagnetization

process in one complete cycle.

In general the specimen retains some energy in stored

magnetic form at any point in the

teresis

loop and back to

However, in going once around the hyswhich the energy will again be the same, energy

cycle.

this point, at

234

[CHAP. 6

ELECTROMAGNETICS

proportional to the area of the loop

is

lost.

This energy

expended

is

the crystal fragments of the specimen and appears as heat.

If

no

in stressing

hysteresis

were

present and the initial-magnetization curve were retraced, the area of the loop would

be zero (Fig. 6-306).

The magnetization-demagnetization process could then be

accomplished with no loss of energy as heat in the specimen, assuming that eddy
currents (see Sec. 8-12) are negligible.

PERMANENT MAGNETS

6-14
In

many

applications permanent magnets play an important part.

In dealing with

permanent magnets the section of the hysteresis loop in the second quadrant of the
BH diagram is of particular interest. If the loop is a saturation or major hysteresis
loop, the section in the second quadrant
6-3 la).

This curve

is

is

called the demagnetization curve (Fig.

characteristic curve for a given magnetic material.

intercept of the curve with the

axis

is

the

maximum

The

possible residual flux density

B
Retentivity-^

Demagnetization^
curve

j"*""'^

'/////A
y/////,

Saturation
hysteresis
loop

(a)

A///A,
-H

FIGURE

6-31

Demagnetization curves.

and

is

negative.)

(B

is

positive

W/A
J\s Coer(:ivity

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-14]

235

1.4

~
i-

10

0.8

^\

**~

1.2

V Y^

f\

y\ax~y

0.6

= 0.4
0.2

/
o
8
I

H
FIGURE

(Am-

6-32

Demagnetization and

Bff product (J nT 3 )

BH product curves for Alnico 5.

axis is the maxior the retentivity, for the material, and the intercept with the
permanent-magnet
It is usually desirable that
coercive force, or the coercivity.

mum

materials have a high retentivity, but

it is

also important that the coercivity be large

so that the magnet will not be easily demagnetized.

In Fig. 6-3 lfc, three demagnetization curves are shown.

Curve

represents a

material having a high retentivity but low coercivity, while curve 2 represents a
material which

Curve

is

just the reverse;

3 represents a material

relatively high retentivity

The maximum

is

it has a low retentivity and high coercivity.

a compromise between the other two, having

coercivity.

BH product, abbreviated BHmal

for a permanent magnet.

criterion, for

and

i.e.,

which

In fact,

it is

is

also a quantity of importance

probably the best single figure of merit, or

judging the quality of a permanent magnet material. Referring to Fig.

curves 1 or 2.
m x is greater for curve 3 than for either

BH
for a substance indicates the maximum energy density (in
product
BH
The maximum
A magnet at BHmxx delivers a given
stored
in the magnet.
meter)
cubic
joules per
6-316,

it is

apparent that

a minimum of magnetic material.

Since the product BH has the dimensions of energy density,

flux with

maximum

value the

maximum

called the energy product and

its

The product BH for any point

on the demagnetization curve

is

it is

sometimes

energy product.

proportional to the

area of the shaded rectangle, as shown in Fig. 6-3 la.

Figure 6-32 shows the demagnetization curve for Alnico 5, one of the best
permanent-magnet materials, which is an alloy containing iron, cobalt, nickel, aluminum, and copper. A curve showing the BH product is also presented. The

236

ELECTROMAGNETICS

maximum

[CHAP. 6

BH product is about 36,000 Jm" 3

and occurs

at a flux density of

about

(see point P).

A discussion of the operating point of permanent magnets

6-15

is

given in Sec. 6-23.

TABLE OF PERMANENT MAGNETIC MATERIALS

Representative materials for permanent magnets are given in Table 6-3.

order of increasing

terials are listed in the

cobalt, copper, iron,

of over 2

MA m _1

and

either

maximum

BH

product.

cerium or samarium have been cast with coercivities

cr
FIGURE

6-33

Demagnetization

of bar-magnet

effect

field.

6-16

The ma-

Magnets of

,::

.7)

y'^-u

DEMAGNETIZATION

A bar of ferromagnetic material that has

magnetized spontaneously.

a bar magnetized so that a north pole

orientation of a single

Table 6-3

a residual flux density tends to become de-

The phenomenon

domain

is

is

indicated,

and

by Fig. 6-33, which shows

and a south pole at the right. The

illustrated

is

at the left

it is

evident that the external field of

PERMANENT MAGNETIC MATERIALS

Retentivity,

Materialf

Chrome steel
0.4 Mn)

(98 Fe, 0.9 Cr, 0.6 C,

Coercivity,

Am"

BHmax

Jm"

1.0

4,000

1,600

0.2

0.6

72,000
76,000

4,800
12,000

0.7

44,800

13,600

Alnico 5 (Alcomax) (51 Fe, 24 Co, 14

1.25

44,000

36,000

Ni, 8 Al, 3 Cu)

Platinum cobalt (77 Pt, 23 Co)

0.6

290,000

52,000

Oxide (57 Fe, 28 O, 15 Co)

Alnico 12 (33 Fe, 35 Co, 18Ni, 8Ti,
6A1)

Alnico 2 (55 Fe, 12 Co, 17 Ni, 10 Al,

6Cu)

Compositions

in percent.

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-16]

237

Iron keeper

w
(a)

FIGURE
U-shaped

(b)

FIGURE

6-34

magnet with and without

6-35

Partial hysteresis loop.

keeper.

the bar magnet opposes this

domain and, hence,

will

netization

is

reduced

demagnetizing

if

but

this case there is

effect

the

little

can be

magnet

is

in the

demagnetizing
still

tend to turn

The tendency

polarity, thereby partially demagnetizing the bar.

form of a U, as

field

further reduced

or reverse

it,

for this

in Fig. 6-34a, since in

along the side of the magnet.

by means of a

The process of removing the permanent magnetization of a specimen so

zero under conditions of zero

is

or deperming.

evident that

coercive force

value

B=

and

decreasing

on each

field

slightly

by increasing

is

that the

called demagnetization

the residual flux density will rise to

H=
it

H field

can be reduced to zero by the application of the

Although

//to

some

might be possible to end up at

more than the coercive force and then
it

to zero, as suggested by the dashed lines, the process requires an accurate

B and

longer but

perming by reversals.
zero

but on removing this

as suggested in Fig. 6-35.

knowledge of

It is

The

soft-iron keeper placed

across the poles, as in Fig. 6-346.

residual flux density

its

demag-

H and the hysteresis loop.

more simply applied method
In this method,

is

called demagnetization or de-

+ H is brought to a smaller maximum amplitude

reversal so that eventually the specimen

field,

as suggested

by

Fig. 6-36.

is left in a demagnetized state at

Although such a demagnetization procedure

can be completely carried out in a matter of seconds with a small magnetic specimen
such as a watch (using ac fields), many seconds or even minutes may be required for
each reversal for large magnetic objects because of the slow decay of the induced eddy
currents and the reluctance of the domains to change polarity.

currents

is

discussed further in Sec. 8-12.

The matter of eddy

[CHAP. 6

ELECTROMAGNETICS

238

FIGURE

6-36

Demagnetization by reversals.

THE MAGNETIC CIRCUIT;

RELUCTANCE AND PERMEANCE

6-17

An

electric circuit

we may

circuit,

Consider

emf in

taken in

first

an

the circuit

parallel,

is

equal to the total

=
RT =

From

(1)

total emf,

paths.

circuit.

Hence, by analogy,

Or

all

the flux tubes of

considered as a magnetic

electric circuit carrying

is

through which the current flows.

a magnetic

may be

HJt

where 1J r

circuit

and form closed

consider that a single flux tube

a magnetic

total

forms a closed path or

flux tubes are continuous

Magnetic

a current

IR drop.

= TR T

/.

By

circuit.

Kirchhoff's law the

Thus
(1)

total resistance, SI

the total resistance

is

(2)

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

SEC. 6-17]

Consider

now a magnetic circuit.

by (2), we may, by

239

Corresponding to the resistance of an

analogy, define a quantity for the magnetic

electric circuit as given

Thus

circuit called the reluctance St.

t = 11

(3)

= total reluctance of magnetic circuit

FT = total mmf of magnetic circuit, A
m = flux through magnetic circuit, Wb

where

3l T

i/f

In general, the total flux

^m

in

a magnetic

>l,

circuit is given

= jJB-ds

by

(Wb)

(4)

B = flux density, T
= element of surface, m 2

where

as

The

integration

is

carried out over the cross-sectional area of the flux tube or tubes

that constitute the circuit.

where

is

If

uniform over the entire cross section,

is

\//

= BA,

the cross-sectional area of the circuit in square meters.

Reluctance has the dimensions of current per magnetic

flux,

or in dimensional

symbols

T 2 / _ T 2 /2
The

relation

T2 T2 /ML 2 has the dimensions of the reciprocal of inductance.

unit for reluctance

The
henrys.

is

the reciprocal henry

reciprocal of reluctance 9t

Hence, from

is

(H

').

called the

permeance 9, which

is

(5)

& T is the total permeance of the circuit in henrys.

The
complete
(3)

expressed in

(3),

^ r = ^- = F
where

Thus the

total

mmf of a magnetic circuit is equal to the line integral of H around the

circuit,

and

this in turn is equal to the

ampere-turns enclosed.

Therefore,

becomes

_*-_'
^T

where

NI is

'I'm

the ampere-turns enclosed.

^m

"Am

<-.)

240

[CHAP. 6

ELECTROMAGNETICS

The above discussion concerns the total reluctance of a circuit. Let us cona portion of a magnetic circuit. In an electric circuit, the
resistance R between two points having no emfs between them is given by
sider next the reluctance of

R=j
V=
/=

where

(O)

(7)

potential difference between the points,

current in circuit,

In the analogous magnetic case, the reluctance 3% between two points in a magnetic
circuit is given

by
!*

= -

(H" 1 )

(8)

Am

where

is

(5-16-6) for

the magnetic potential difference between the points in amperes.

U and (4) for

\ji

From

we have
C2

J>
,

(9)

fjB-<fc

where

is

integrated between the

two points

and

(1

between which we wish to

2)

find the magnetic potential difference U.

When the
(9)

circuit

has a uniform cross section of area

and the

field is

uniform,

reduces to

=
= ^1
BA fiAA
where

(H" 1 )

(10)

reluctance between points 1 and 2, H -1

/ = distance between points
and 2, m
A = cross-sectional area of magnetic circuit, m 2
_1
H = permeability of medium comprising the circuit, H m

8t

The permeance

9 between the points

and 2

'-a-?

is

given by

<

Reluctances in series are additive in the same

additive.

For reluctances

way

(ID

that resistances in series are

in parallel the reciprocal of the total reluctance

is

equal to

sum of the reciprocals of the individual reluctances. For reluctances in parallel

it is usually more convenient to use permeance, the total permeance being equal to the
sum of the individual permeances.

the

j
sec. 6-17]

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

Ends

241

of block.

30 mm

(a)

50 mm

50 mm

"i^ Block 2

"

30 mm

"

,^

H,

FIGURE

100 mm

=^^^

Block 2

40 mm

6-37

Block

/t,

B,

U.

iiJUilo-

200 mm

Rectangular iron blocks.

EXAMPLE

30 mm

I*a*

200 mm

20 mm

Block 1

B-

Find the reluctance and permeance between the ends of the rectangular
B is uniform throughout the block and
normal to the ends. The permeability of the block is uniform and has a value /*, = 500/u
o
where fc is the permeability of vacuum.
1

block of iron shown in Fig. 6-37a, assuming that

The

solution

reluctance of the block

/

"
The permeance ^

0.1

500x4ttx

fiA
is

is,

10- 7 x

from

(10),

" 1
1 06X 1 SH
15x 10-*~
-

the reciprocal of 01; so

&

1.06

x 10*

= 9.4 /xH

EXAMPLE

2 Find the total reluctance and permeance between the ends of the seriesconnected rectangular iron blocks shown in Fig. 6-37*, assuming that B is uniform
throughout
the blocks and normal to the ends. The permeability of each block is uniform,
the value in
block

being

/x t

solution
block 2

= 500/i o and in block 2, /x 2 = 2,000/x o

The

reluctance

^,

of block

is

given in Example

The

1.

is

~ (i 2 A
The

0.2
~~

2,000 x 4tt

total reluctance 9t T equals the

x lO" 7 x 15 x 10"*

sum of the

= 0.53x

10 5

H-

individual reluctances; so

^r = i + ^2=(1.06 + 0.53)x

10 s

= 1.59 x

lO'H" 1

reluctance of

[CHAP. 6

ELECTROMAGNETICS

242

The

total

permeance

EXAMPLE

Find the

total reluctance

and permeance between the ends of the

parallel-

connected rectangular iron blocks shown in Fig. 6-37c, assuming that B is uniform in each
block and normal to the ends. The permeability of each block is uniform, the value in

block

being

/i,

and

in block 2 being

The permeance 9>i of block

first.

^Aj _ 500

4tt

The permeance of block 2

H
~

total

more convenient

to calculate the total

1 is

=6

2g
'

0.2

is

A2

2,000

4tt

x 1Q- 7 x 10 x 1Q- 4 _
'

0.2

= 0> + &> 2 = (6.28 +

t

total reluctance is then given

12.6)

x 10- 6

18.9

pH

by
1

1.89

6-18

it is

permeance equals the sum of the individual permeances; so

&> T

The

= 2,000/i O

x 1Q- 7 x 20 x 1Q-*

~~~l

The

/* 2

Since the blocks are in parallel,

solution

permeance

= 500^

x 10~ 5

5.3x 10*

H-

MAGNETIC FIELD MAPPING;

MAGNETIC FIELD CELLS

The examples

in the preceding section illustrate

how the reluctance

or permeance

may

be found for sections of a magnetic circuit that have a uniform cross section and
uniform field. In two-dimensional problems where the field and cross section are

nonuniform the magnetic field configuration, and consequently the reluctance or

permeance, can also be found provided the permeability may be considered constant.
Graphical field-mapping techniques such as are employed in Sees. 3-22 and 4-16 are
applicable to such situations.

The following
1

The

field

angles.

basic properties are useful in magnetic field mapping:

(H or B)

lines

and the magnetic potential (U)

lines intersect at right

Sec. 6-18]

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

243

Equipotentials

FIGURE
Magnetic

At

6-38
Walls of flux tube

field cell.

the boundary between air and iron (or other high-permeability

the field lines

on the

medium)

of the boundary are substantially perpendicular to

air side

the boundary surface.

The boundary between

air and iron (or other high-permeability medium) may

be regarded as an equipotential with respect to the air side of the boundary but

not, in general, with respect to the iron side.

In a uniform

field

the potential varies linearly with distance.

magnetic flux tube

is

cross section of the tube

Magnetic

parallel to the field,

is

and the magnetic

flux over

any

a constant.

flux tubes are continuous.

With these properties in mind a two-dimensional magnetic field can be divided

and then by equipotentials into magnetic field cells with sides

into magnetic flux tubes

that are squares or curvilinear squares, using the trial-and-error

Sec. 3-22 in connection with electric field

magnetic field

two others by the

cell is

bounded on two sides by equipotential surfaces and on

For instance, the sides of the magnetic
of a flux tube, while the top and bottom surfaces

The field is parallel

to the sides

The permeance of a magnetic

surfaces, is, from (6-17-11),

surfaces.
tential

in

sidewalls of a flux tube.

field cell in Fig. 6-38 are the walls

are equipotentials.

method described

mapping.

&o =
and the permeance per

HA

unit depth

\ild

-j-

and normal to the top and bottom

measured between the equipo-

field cell, as

lid

(H)

(1)

is

0n

=H

(HnT 1 )

(2)

244

[CHAP. 6

ELECTROMAGNETICS

''''-:-'

'-.'

Iron'

/*

-;V'.;v.

-';*vi-:

; :;>'

:'[':.:.

Remainder

'':':'''.'.::''
'.''.

.*' ''

'"*'".

ti

Air

iri
<j^r
;

"""""it

Pole
faces

flux-;:

!-V-?.tubeV':.v:!

:::

':':; '.':.*,-.:.";'V<7

.''Vequipotentials' .".-'-/Nw.

Xyt&
(a)

A^
BorH
::lines

(b)

FIGURE
Magnetic

6-39
field in air

gap (Example

1).

where n is the permeability of the cell medium in henrys per meter. Thus, the value of n
for a medium is equal to the permeance per unit depth of a magnetic field cell of that
medium. For example, a magnetic field cell in air has a permeance per unit depth of
1
1
Thus, if d in Fig. 6-38 equals 1 m and the medium
or 1.26 fiH m"
40071 nH m"
.

is air,

the permeance of the cell

cells

is

1.26 /iH.

can be subdivided into smaller square-ended cells with as many

Hence the permeance per unit depth of any field cell,
in parallel as in series.

Any

field cell

large or small, exactly square or curvilinear,

All cells with the same flux through

is

equal to

\i.

them may be defined

as magnetic field cells

It follows that the magnetic potential difference across

of the same kind.

the same kind is the same.

To

illustrate

some of

the principles of magnetic field mapping,

let

all cells

of

us consider

three examples involving three variations of a two-dimensional problem.

magnetic circuit has an air gap of nonuniform separation, as suggested

The iron has a uniform depth d into the page of 1 m. The geometry of the
in Fig. 6-39a.
gap is identical with the region between ff and gg in the capacitor of Fig. 3-19. Find the

EXAMPLE

permeance of the
solution

air gap, neglecting fringing

It

may be assumed

of the

field.

that the iron permeability

is

much

greater than

gap will be perpendicular to the air-iron boundary and this

the
boundary can be treated as a magnetic equipotential. Since the geometry of the gap is
so that the field lines in the

245

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-18]

Equivalent current

/sheet ( out of page

Air

Equipotentials

Equivalent current
sheet ( into page

FIGURE

6-40

Iron bar of nonuniform cross section with internal

same as

map

that for the capacitor in Fig. 3-19, the field

present case, noting that the field lines here are

or

in Fig. 3-19

field

may

(Example

2).

also serve in the

H lines and the equipotentials are sur-

shown in Fig. 6-396.

With the exception of the cells in the remainder flux tube, all the field cells are of the
same kind, and the permeance of the air gap is given in terms of cells of the same kind by

faces of equal magnetic potential U, as

N
(3)

where

N = number of field cells (or flux tubes) in parallel, dimensionless

n

= number of field cells

&o = permeance
The remainder

in series, dimensionless

cell,

9i cells in series, while the other flux tubes have

0.43 of the width of a full tube, and N=\5 + 0.43

flux tube has

remainder tube
total

of one

is

4/9i

permeance of the gap is then 0> r = (15.43/4)^o = 3.86^

m, the permeance of one cell is 0> a =/i. */= 1.26 x

cell is 1

permeance

4.

Hence

= 15.43.

the

The

Since the depth of each

=1.26

/M

and the

total

is

0>t

= 3.86

x 1.26

= 4.86 /*H

assumed in this example that there is no fringing of the field. For an actual gap there
would be fringing at the edges, and the actual permeance of the gap would be somewhat

It is

larger than given above.

EXAMPLE

Let the problem of the above example be modified to that shown in Fig. 6-40.

Here the gap of the

may be

first

example

is

replaced by iron and the iron poles by

air.

regarded as part of a magnetic circuit of iron extending further to the

The iron
left

and to

246

[CHAP. 6

ELECTROMAGNETICS
Sheet current

K(A m-i)
/(out

of page)

Conducting
strip-.

Equipotentials

Sheet current
_1

K(Am

(into

FIGURE

6-41

Cross section of strip transmission

The iron extends

the right as suggested by the dashed lines in Fig. 6-40.

normal to the page, with the cross section at any depth identical to that
that the iron has a uniform permeability /* which is much larger than /i
between the surfaces indicated by the dash-dot

map

page)

line

to

(Example

a depth of

in Fig. 6-40.

3).

Assume

Find the permeance

lines ff and gg.

same as

Example 1 (Fig. 6-396)

shown in Fig. 6-40. It is
assumed that /* is so much greater than /t that in the iron the H field at the air-iron boundary
The total permeance
is substantially parallel to the boundary as indicated by the map.
between #and gg is, from (3), 0> = (4/15.43)0> o = 0.259 X/ixl= 0.259 x 1.26/*, = 0.326/*,
solution

The

field

for this

problem

is

the

for

except that the field and equipotential lines are interchanged, as

/*H,

where

/x, is

EXAMPLE

the relative permeability of iron.

Let the problem of the preceding example be modified to that of a two-strip

transmission line having the same cross section as the gap of Example

of Example

As shown

2.

in Fig. 6-41, the

two conducting

strips

and the iron

circuit

extend normal to the page

with a sheet of steady current flowing outward on the upper strip and an equal current

The medium in which the strips are located is air.

Find the inductance of a 1-m length of the line.

flowing inward on the lower strip.

edge

effects.

solution

Neglecting edge

effects, t

the field

map between

the strips

is

Neglect

identical with

that for the iron circuit in Fig. 6-40.

t If the conducting strips are extended an infinite distance to the left and right, as
suggested by the dashed lines in Fig. 6-41, the field configuration is precisely as
indicated.

The

field

In Example 2 the

between the

strips is

may be

produced by the currents on the

strips.

regarded as due to an equivalent current

sheet at the surfaces of the iron bar normal to the page (Fig. 6-40).
field in

the iron

THE STATIC MAGNETIC HELD OF FERROMAGNETIC MATERIALS

SEC. 6-19]

If

along
line

its

each

map

the

cell in

regarded as a strip transmission line with sheet currents

is

upper and lower surfaces, the inductance

(normal to the page in Fig. 6-41)

is,

from

L =n
The

LT

total inductance

a length d of

m of the singlccell

by

d= 1.26 ftH

of a 1-m length of the line

for

(5-17-2), given

is

then

x 1.26= 0.326

15.43

6-19

247

iM

COMPARISON OF FIELD MAPS IN ELECTRIC,

MAGNETIC, AND CURRENT CASES

Graphical

mapping was discussed in Sec. 3-22 for electric fields, in Sec. 4-16 for
and in Sec. 6-18 for magnetic fields. The technique is similar

field

currents in conductors,
in

all

these cases.

Of particular

significance

is

the fact that a field

map

for a certain

two-dimensional geometry can be applied to numerous problems having this geometry.

An

which the

illustration
field

map

of

was provided by the three examples

this

in Sec. 6-18, in

of Fig. 3-19 for a capacitor yielded the solution for the permeance

of the volume with the

field

applied both transversely and longitudinally.

The map

also gave the inductance of a conducting-strip transmission line.

The same map

can, in addition, supply the value of the conductance of a con-

ducting bar with the current flowing transversely and with the current flowing longitudinally.

The same map can

To summarize,

also be applied to heat-

and fluid-flow problems.

sketches are given in Fig. 6-42, showing six different problems of

same geometry for which solutions are supplied by one field map. The actual
The geometry of
is shown in Fig. 6-42a, being omitted in the other sketches.
the problems is that of the capacitor of Fig. 3- 19, which was also used in the problems
the

map

of Figs. 6-39 to 6-41.

In Fig. 6-42a the
transverse.

map

represents the electric field in a capacitor with the field

In Fig. 6-42b the

map

represents the electric field in a conducting bar

with current flowing transversely, while in Fig. 6-42c the current flows longitudinally.
In Fig. 6-42d the

map

represents the magnetic field in the air gap between

two iron

an iron bar with the

field applied longitudinally.
In Fig. 6-42/ the map represents the field between
two conducting strips acting as a transmission line with current flowing normal to
pole faces, while in Fig. 6-42e

the page.

it

represents the magnetic field in

For each case the capacitance, conductance, permeance, or inductance per

unit depth (normal to the page)

Fringing of the

is

given, as appropriate for the particular problem.

field is neglected in all cases.

248

.3
V3

i
E
a

3
8

21

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-19]

It is also

of interest to compare the significance of the

cells

(square or curvilinear)

of the field maps for the different problems we have considered.

per unit depth of an

electric field cell equals the permittivity e

249

Thus, the capacitance

of the medium; the cona of the medium;

ductance per unit depth of a conductor cell equals the conductivity

permeance per

of a magnetic field cell equals the permeability ft of the

medium; and the inductance per unit length of a transmission-line cell equals the perthe

meability

of Table

ft

unit depth

of the medium.

These relationships are summarized in the

This table also has columns headed Flow

6-4.

lines,

Flow

last

tubes,

column

and Equi-

Under Flow lines are listed the quantities having the direction of flow
and under Flow tubes the quantities equal to the total flux through a tube.

potentials.
lines

Table 6-4

IMPORTANT FIELD-MAP QUANTITIES

Field

Flow

Electric

DorE

Electric flux

Current

JorE

Current I

Flow tubes

lines

<ji

Equi-

Value of

potentials

(per unit depth)

K(V)

Permittivity e

K(V)

Conductivity

(Fm-

cell

(Urn" 1 )

BorH

Magnetic

Magnetic

flux

>fi

U{A)

Permeability

/x

(Hm- 1 )
Heat

Temperature

Heat per time

Temperature

Thermal con-

Fluid flow (non-

gradient
Velocity

Mass per time

Velocity po-

Density

ductivity
tential

turbulent; incompressible)

EXAMPLE

Apply the above analogies

to find the capacitance of

an

air capacitor

by a

resistance measurement.

solution

The

uniform conductivity

capacitor plates are immersed in a large tank

a,

and the dc resistance

In general, the conductance

G of a certain geometry is given

G-od N
where

N = number of cells in parallel

= number of cells
d = depth of cells
n

in series

filled

with a liquid of

R is measured between the plates.

by

0)

250

An

ELECTROMAGNETICS

[CHAP. 6

actual capacitor with the

Fig. 6-42a

and

same geometry has the same

C=e
where

N and n are the same as in

rfn

(2)

Hence, on dividing

(1).

C = ~G =
a

where

configuration (compare

field

b); so the capacitance

(2)

by

(1),

^
oR

(3)

C = capacitance of actual capacitor, F

= permittivity of air = 8.85 pF m~
a = conductivity of liquid, Urn -1
R = l/G = measured resistance, CI

Thus when a is known

of an

air capacitor

(it

can be measured with a rectangular volume), the capacitance

can be obtained from

(3)

by a resistance measurement.

GAPLESS CIRCUIT

6-20

Consider the magnetic circuit of a closed ring of iron of uniform cross section

around the ring and that we wish to know how large the

wound uniformly
product NT of the number

of turns and the current must be to produce a flux density

B in the ring.

and mean length

The

coil

/.

Suppose that a

on the ring

coil

of insulated wire

in Fig. 6-43a forms a toroid.

is

In the toroid we have, from

(5-11-5),

= permeability (assumed uniform) of medium inside of toroid, H m -1

N = number of turns, dimensionless
/ = current, A
/ = mean length of toroid, m
R = mean radius of toroid, m
Dividing by n, we have NT = HI ampere- turns. If a certain flux density Bis desired in

where n

the ring, the corresponding

the required

EXAMPLE
/ = 600 mm.
From

H value

is

number of ampere-turns

An

taken from a

calculated

BH curve for the ring material and

from

iron ring has a cross-sectional area

NI =

A=

HI.

1,000

mm 2 and a mean length

Find the number of ampere-turns required to produce a flux density

BH curve for the iron, H = 1,000 Am-'atfi=lT.

B=1T.

251

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sbc. 6-20]

(a)

FIGURE

(b)

6-43

Closed iron ring (a) with uniform winding and (6) with concentrated winding.

solution

From NI = HI
NI =

The

may

coil

1,000

0.6

= 600 A turns

could be 100 turns with a current of 6

or 1,000 turns with 600

mA.

be uniformly distributed around the ring, as in Fig. 6-43a, or concentrated

The
in

coil

a small

sector, as in Fig. 6-43/>.

The required number of ampere-turns can

also be found

by calculating the

reluctance of the ring circuit, as illustrated by the following example.

EXAMPLE 2 Find the number of ampere-turns required

B = 1 T by first evaluating the reluctance of the ring.
solution

From

for the ring of

Example

for

(6-17-6)

3t-

We also have H-dl=Hl = NI.

It

(2)

BA

follows that

NI = 0tBA, where
(3)

fxA

Since

= 1,000

A m"

when

B=1T,
B

1,000

mH m"

(4)

252

It is

ELECTROMAGNETICS

[CHAP. 6

to be noted that the relative permeability for this case

^=
Introducing (4) in

(3)

and also the value of

4^F=
/

and

is

795

gives the reluctance of the ring as

^ = io-^ io-^ 6xlo5H

Hence, the required number of ampere-turns

M=9tBA=6x
as obtained in

Example

(5)

"'

(6)

is

10 s x

x 10- 3 =600

MAGNETIC CIRCUIT WITH AIR GAP

6-21

shown in
By the continuity of the normal

Let a narrow air gap of thickness g be cut in the iron ring of Sec. 6-20, as
Fig. 6-44a.

The gap detail

component of
neglected.

as

assumed

iron

is

presented in Fig. 6-446.

the flux density in the gap

Neglecting the fringing involves but

here.

The

field

in the

gap

is

same as

the

is

then

little

in the iron if fringing is

error where the gap

B/fi

is

narrow,

while the field Hi in the

is

from which

The number of ampere- turns required

magnetic circuit with gap, as in Fig. 6-44a,
be obtained directly.

For

once around the magnetic

That

to produce a certain flux density

is

instance, according to (5-16-9), the line integral of

circuit equals the total

mmf F,

is,

EXAMPLE
air

or ampere-turns enclosed.

H-dl = F = NI

an

in

a problem for which the solution can

(3)

Let the iron ring of Fig. 6-44 have a cross-sectional area

gap of width g

= 2 mm,

and a mean length

= 2-nR = 600 mm,

Find the number of ampere-turns required to produce a flux density

A=

1,000

mm 2

including the air gap.

B=1

T.

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-21]

if
FIGURE

253

<

6>

6-44

Iron ring with air gap.

We have

solution

NI = JH-dl=H (l-ff)+Hg g

(4)

= H field in
= H field in

where H,

H,

From

Hence

iron

gap

BH curve for the iron, H, = 1,000 Am -1

(4)

NI = H,W -g)
where p,

and from

(2)

we know H,

in

terms of

t .

becomes

= 795

is

+n

g]

(5)

the relative permeability of the iron ring at

NI =

1,000[(0.6

- 0.002) + 795

0.002]

B=

T.

Therefore,

= 2,188 A turns

introduction of the narrow air gap makes it necessary to increase the ampere-turns from
600 to 2,188 to maintain the flux density at 1 T.
The problem can also be solved by calculating the total reluctance of the magnetic

The

circuit.

Thus, from

(4)

we have

Nl= \^
fj.A

m-

g)

Hgg
+ ^A
A
fio

(6)

254

ELECTROMAGNETICS

[CHAP. 6

NI = BA(, + &,)

and
where M,

= (/ g)lixA = reluctance of iron

^ = sIh-o A = reluctance

6-22

part of circuit

of air gap

MAGNETIC GAP FORCE

Referring to Fig. 6-44, the effect of the magnetic field

close the air gap.

That

is,

force between magnetic pole pieces

developed.

is

wm = \
is

small,

we may assume a uniform

is

is

(Jm- 3 )

2 n

gap

fields find appli-

In this section an expression for the

devices.

density of energy stored in a magnetic field

Wm stored in the gap

to exert forces which tend to

Forces produced by magnetic

numerous electromechanical

The

is

the magnetic poles of opposite polarity at the sides of the

gap are attracted to each other.

cation in

If the

(7)

(1)

field in the air gap.

The

total

energy

then

Wn = wm Ag = %^-

(J)

(2)

2/*o

where

A = area of gap, m 2
g = width of gap, m

Suppose now that the iron ring

held open by a force

gap by an
is

F as

infinitesimal

in Fig. 6-44

in Fig. 6-45.

amount dg while

perfectly flexible, so that the

gap must be

same time the current through the

at the

increased to maintain the flux density

increased by the infinitesimal

is

If the force is increased so as to increase the

coil

constant, the energy stored in the gap

is

amount

B2A

dWm = dg
Equation

(3)

has the dimensions of energy.

force times distance, which in this case

the poles.

It is

is

F dg,

(3)

But energy may also be expressed as

where F is the attractive force between

equal in magnitude to the force required to hold them apart.

Thus

dg

Fdg = -

2/i

or

B2A
F =^T-

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

sec. 6-23]

FIGURE

6-45

F=

attractive force,

B = flux density, T
A = area of gap, m
_1
H = permeability of air = 4007t nH m
Dividing by the gap area A yields the pressure P.

F
P=-

That

is,

(NrrT 2 )

ill

(5)

2/^0

PERMANENT MAGNET WITH GAP

6-23

Suppose

is

that a closed iron ring

first

toroidal coil

iron

^Gap

Hron

Forces at air gap.

where

wound on

is

When

the ring.

magnetized to saturation with a uniform

the coil

equal to the retentivity (see Fig. 6-47).

is
If,

removed, the flux density

somewhere on the demagnetization curve

information

is

line integral

of

needed to locate

this point.

(Fig. 6-47) (see also Sec. 6-14).

H once around a magnetic circuit

j

d\

(l

may be

This

-g) +

is

where

d\

NI.

Since

air

which

Further

obtained as follows.

The

NI = 0,

Hg=
g

H il-g)=-H g
i

in the

however, the system has an

gap, as in Fig. 6-46, the flux density has a smaller value as given by a point
lies

255

(1)

= H field in the iron

= 2nR = total length
g = width of gap
Hg = H field in gap
t

Thus Hi and

of magnetic circuit (including gap)

are in opposite directions, as indicated in Fig. 6-46.

If leakage is

256

ELECTROMAGNETICS

[CHAP. 6

Demagnetization
curve

^Coercivity

FIGURE

FIGURE

6-46

Permanently magnetized ring with

B is

neglected,
ratio

B/H

t ,

uniform around the

This ratio of the flux density

the shearing line, as

shown

Multiplying (1) by n

circuit.

or the permeability of the iron,

6-47

Demagnetization curves for permanent magnet.

air gap.

and solving for the

we obtain

= -A*o-l-g

B to the field H

in the iron gives the slope

The

in Fig. 6-47.

of a

line called

intersection of this line with the de-

magnetization curve determines the position of the iron on the magnetization curve
(point P).
This location is a function of the ratio of the iron path length (/ - g) to the

gap length

g.

In most permanent-magnet applications, where

tively constant,

a moderate demagnetizing

line.

The

it

desired that

remain

rela-

applied to the iron, moving the

On

position of the iron to P' (Fig. 6-47).

point P" on the shearing

field is

ring

removing the field, the iron moves to the

magnet is now said to be stabilized, and when

fields less than about the difference of

between points P' and P" are applied to the
ring and then removed, the iron will always return to approximately point P". Under
these conditions the iron moves along a minor hysteresis loop, as suggested by the

dashed

COMPARISON OF ELECTRIC AND MAGNETIC RELATIONS

INVOLVING POLARIZATION AND MAGNETIZATION

6-24

It is

lines in Fig. 6-47.

interesting to

compare the magnetic

relations

where magnetization

with the corresponding electric relations where polarization

This

is

done

in

Table

6-5.

P is

is

present

present (see Chap.

3).

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

Table 6-5

257

COMPARISON OF EQUATIONS INVOLVING

POLARIZATION P AND MAGNETIZATION M

Description of equation

Dipole-moment

relations

and permeability

Relative permittivity and

meability

B=

Relation to polarization charge

density and to equivalent
current density

(H

= --i

X"

T/2y =

Scalar and vector potentials

V=-4ne

= Mr -

V X M = J'

V.P = p

Poisson's equations

+ M)

per-

Susceptibilities

/Li

M = /*o(l+^j

+E

D= e E+P

Flux density

Permittivity

Magnetic case

Electric case

_P
c

V 2 l/ = VM= -V-H

pv.p
e

dv

+VXM

4ttJ v

* Answers to starred problems are given in Appendix C.

PROBLEMS
*

6-1

magnetic

field

6-2

magnetized needle of 20
of 50 //T flux density.

magnetized needle of 30

A m

Find the

A m2

Qm

placed coincident with the

moment is situated in
maximum torque on the needle.

magnetic

magnetic moment

the origin in a magnetic field of flux density

I is

B = yBa

axis at a large distance

If

is

situated with

a uniform

center at

its

a short bar magnet of

x from the

moment

origin, find the angle

between the needle and the y axis assuming that the needle pivots freely around

its

center

point.

A uniformly magnetized bar with a volume of 0.01 m 3 has a magnetic moment of

If the flux density B = 50 mT in the bar, what is the value of H in the bar?
500 A m
6-4 A bar magnet in a uniform magnetic field is acted on only by a torque, there being

6-3

no

translational force

on

the magnet.

In a nonuniform

field,

there

is

a net translational force.

mm

long with
Find the maximum value of this force on a uniformly magnetized bar magnet 5
2
bar
long
slender
one
pole
of
a
very
from
100
situated
a magnetic moment of 2 A
magnet having a pole strength of 500 A m.

mm

258

ELECTROMAGNETICS

Two cavities are cut in a ferromagnetic medium (/x, = 200) of large extent. Cavity
flat faces perpendicular to the direction of B in the ferro-

6-5
1

is

[CHAP. 6

a thin disk-shaped cavity with

magnetic medium.

Cavity 2

cavities are filled with air.

1

and

(b) the center

Two

6-6

and

(b) the

what

is

the magnitude of

axis parallel to B.

its

Find the force per unit length of conducor

relative permeability

d carry a current
the medium is

if (a)

= 200.

/*,

ferromagnetic media are separated by a plane boundary.

Medium

and medium 2 a relative permeability ^, = 1,500.

magnetic field direction in medium 2 is at an angle of 10 with respect to the surface,
angle between the field direction and the surface in medium 1

relative permeability

needle.

Take

What

is

= 200

10

Am

moment

magnetic

is

has a
If the

find the

mm from a compass

brought 500

the maximum angular deviation of the needle which the magnet can produce?

the horizontal

6-9
5

/x r

A bar magnet with

6-8

The

H at (a) the center of cavity

of cavity 2?

medium everywhere has a

Two

6-7

a long needle-shaped cavity with

B = 2 T,

long straight parallel conductors separated by a distance

/ in opposite directions.
air

is

If

component of the

earth's field as

mT.

Design a magnetic levitation system to raise 1,000 kg

at the earth's surface

by

mm using (a) permanent magnets and (b) electromagnets.

6-10

steel pipeline

runs east-west at a constant depth below a

undisturbed earth's magnetic

field in the

region has a dip of 70 (angle of

flat

ground.

B from

and a declination of 0 (angle of B from north). Locate the pipe and its depth
dip measurements are obtained with the dip needle
m above the ground

if

The

horizontal)

the following

Meters south of reference mark

Dip

6-11

permeability

is

angle, deg

a function of x and

is

given by

H-

= H'o + H-i

6-13

flux densities

on

10

12

14

16

75

77

74

70

65

62

66

69

flat

B,

The

Po

moment

of 50

If the bar

Am

magnet

is

is

situated parallel to a

500

mm from

the wire,

sheet of linear current density

= 5 mT

separates two media with magnetic

B2 = 20 mT in opposite directions
/i,i = 50 and jx r2 = 100.
What is Kl

and

parallel to the sheet.

The

Show that at the center of a long uniformly magnetized bar of uniform cross= magnetization and / = length of bar.
A that H 2MA/wl 2 where
Show that the permeability at the center of a long permanently magnetized

sectional area

6-15

the bar magnet.

relative permeabilities are

6-14

bar.

small bar magnet with magnetic

very long straight wire with current of 50 A.

find the torque

rectangular iron bar has a length Xi and a cross-sectional area A.

Find the permeance of the

6-12

71

rod of uniform magnetization

cross-sectional area of rod.

is

given by

[1

(ir/ 2/2/4)].

where /

= length of rod and A =

THE STATIC MAGNETIC HELD OF FERROMAGNETIC MATERIALS

find

conductor

B and H as

of 25, 50, 100, 125, 150, and 200

uniform and that

is

that the flux density

D
db

= 400
to

mm. Assume that

in the iron tube,

250

tube of

-noQ^lsind
\

ana

-.

(ft)

200 A,

the current density

Sketch the variation

mm.

from a bar magnet

Q m = dipole moment
6 = angle between bar axis and
assumed that r >

where

It is

/x r

a function of radius from

Show

6-17

is

is

B and H at radii

in the

of

enclosed by a concentric iron

of 50 mm radius
mm and outer radius 150 mm. (a) If the current in the conductor

A copper conductor

6-16

inner radius 100

259

given by

is

fi o

QTm lcos0
5

radius vector of length r

I.

Show

6-18

that the flux density

where /

Q lA
= <-,
u.

B,

from a single-turn wire loop

sin 6

and

4*r 3

B =
r

\t-olA

is

given by

cos 6

-z

2irr

j
3

= current

A = area
IA = magnetic moment
6
It is

= angle between axis of loop

assumed

that r

much

is

and radius vector of length

A small bar magnet is situated in

6-19

greater than the diameter of the loop.

air at

a distance

d from

the plane surface of a

The bar axis makes an angle 6 with respect

If the dipole moment of the bar magnet is Q m l, find the
to a perpendicular to the surface.
force and torque on the bar magnet due to the magnetization induced in the medium of per-

medium of

meability

large extent

magnetized rod 10

Am -1

mm

6-21

mm in diameter and 30 mm long has a uniform magnetiz-

Find the magnetic moment of the bar.

in diameter and 75
uniformly magnetized rod 25

ation of 1,200

moment

/*.

fj..

6-20

and permeability

kA

mm long has a magnetic

Find the equivalent sheet-current density at the surface of the rod.

Demonstrate that the force between two small bar magnets varies as the inverse

of 12

6-22

fourth power of their separation and

is

independent of their relative orientation.

Assuming that the demagnetization curve of a certain ferromagnetic material

line, what is the maximum BH product if the retentivity is 1 T and the coercivity

6-23(a)
is

a straight

*
?
(b) Prove that this is the maximum value.
According to Lord Rayleigh, the bottom part of the normal magnetization curve
+ aH 2 where f*< is the initial permeability (at = 0, B = 0) and a is a
is given by B = ft,
constant. Assume that this relation applies to the initial-magnetization curve of an iron

is

20

kA m~
6-24

What

specimen.

from

H=

to

mm.

the expression for the energy density in the iron after the field

H = Hi ?

An

6-25

of 200

is

Assume

that the specimen

is initially

raised

mm 2 and a mean radius

mm wide. Find the number of

iron ring has a uniform cross-sectional area of 150

The

ring

is

continuous except for an air gap

ampere-turns required on the ring to produce a flux density B

fringing.

is

unmagnetized.

When B = 0.5T

in the iron,

= 250.

= 0.5 T in the air gap.

Neglect

260

ELECTROMAGNETICS

[CHAP. 6

0.5
t

0.15 ml

1.8

2.8

0.5

0.5

-3.8

_*A
,0.5

fr- ln

0.5

1
1.2

>-|

2m-

m-

Side view

End view

(a)

(b)

FIGURE

P6-29
Cyclotron magnet.

A ring has a cross-sectional area of 200 mm 2 and a mean path length of 350 mm.
Two-thirds of the length of the ring is composed of ferrite (ji, = 1,000) and the remaining
one-third of mild steel Q*, = 2,000). Find the total magnetic flux in the ring if
is wound
6-26

it

with a coil of 750 turns carrying

and

an

steel sections are equivalent to

6-27

dimensions
ix,

100mA.

Consider that the two butt joints between the ferrite

air

gap of 100

/urn.

Find the reluctance and permeance between the ends of an iron bar having the
(a)

shown

(a)

How much

in Fig.

4-20 assuming

/*,

= 3,000,

(b)

shown

in Fig.

P4-4 assuming

= 5,000.
6-28

greater

is

the reluctance of the block of Fig. 4-20

uniform rectangular block without the notches assuming the same

much

greater

is

the reluctance of the block of Fig. P4-4

fx.,

compared

in

compared

both cases ?

(jt>)

to

How

to a uniform rectangular

block obtained by cutting the one in Fig. P4-4 along the diagonal at the bend, turning one

and joining both halves to form a gapless straight bar?

A cyclotron magnet has the dimensions shown in Fig. P6-29. The pole pieces
are cylindrical with tapered ends. The diameter of the gap is 1 m, and the gap spacing is
150 mm. How many ampere-turns are required in each of the two windings shown to
produce a flux density of 1 T in the air gap ? Assume that the magnet is made of iron with
half over,

6-29

f*,

= 2,000 and thatthere is no fringing at the gap.

As a

Neglect leakage along the magnet structure.

further simplification take the effective length of sections as the length

the centerline (dashed line in Fig. P6-29a).

Take

measured along

the diameter of the tapered section of the

poles as the average diameter.

6-30
Fig. P6-30.

An

electromagnet consists of a U-shaped iron yoke and iron bar, as shown in

A thin copper

sheet

on

the top of the bar prevents iron-to-iron contact

between

THE STATIC MAGNETIC FIELD OF FERROMAGNETIC MATERIALS

261

Yoke

Bar

FIGURE

P6-30

U-shaped electromagnet.

bar and yoke.

area

is

0.01 5

of the bar) ?
6-31

If the

magnetic flux through the circuit

m 2 per pole, what

is

is

15

mWb and the yoke-bar contact

the weight which the yoke will support (including the weight

Neglect fringing.
(a) If the

contact area of the electromagnet of Prob. 6-30 is reduced to 0.005

by means of tapered sections on the yoke, what

is

the weight which the yoke will support?

same as before and neglect fringing. (/>) In practice what

prevents the attractive force from increasing indefinitely as the contact area is reduced ?
6-32 What is the contact pressure (a) in Prob. 6-30 and (b) in Prob. 6-31a?
6-33 An iron ring magnet of 0.02 m 2 cross-sectional area and 300 mm radius has
*
a 1-mm air gap and a winding of 1,200 turns. If the current through the coil is 6 A, what is
the force tending to close the gap? Take r = 1,000 for the iron and neglect fringing.
6-34 Show that the permeability of a permanent magnet with air gap can be expressed
=
=
by
-ftoC^,/^), where /*<> = permeability of air, 0>, = permeance of air gap, and 0>

Assume

that the total flux

is

the

/j.

ju.

permeance of empty space occupied by the magnet.

A long conducting tube of negligible wall thickness carries a longitudinal current

K = 1\2ttR, where / = total current and R = radius of tube. Find
> R). (d) What boundinside
<
R), (6) at r = R, and (c) outside the tube
the tube (r
H (a)
ary relation is satisfied at r = /??
6-35

of uniform sheet density

(/

6-36

may

For may purposes the magnetic

be regarded as

if it

field

of the earth at ionospheric heights or above

were produced by a short magnetic dipole (or small current loop)

situated at the center of the earth with

/j.

fj.

everywhere (except within the dipole).

If the

component of B at the earth's surface is 20 /iT at a point where the dip angle of the
field (angle from horizontal) is 72, (a) find the magnetic moment required for the dipole
(or loop) at the center of the earth. Take the earth's radius as 6.37 Mm. (b) Using this model,
above the earth's surface in the equatorial and polar
calculate B at a distance of 40
horizontal

Mm

directions.

LAPLACE'S

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

7-1

INTRODUCTION

In this chapter
first

we

examples space

tionf (V

V = 0)

shall consider
is

a number of electrostatic

which

satisfy

field

problems.

In the

and we seek solutions to Laplace's equathe boundary conditions. In a later example space

free of charge (p

0),

and a solution to Poisson's equation (V 2 K= p/e) is obtained.

Because of the important part that boundary conditions play, these problems are often

charge

is

present,

called boundary-value problems.

The

solution of a boundary- value problem

is

usually facilitated if it

is set

up in a
For

coordinate system in which the boundaries can be specified in a simple manner.

instance, a

problem involving a rectangular object may be most readily handled with

rectangular coordinates, a cylindrical object by cylindrical coordinates, an elliptical

by elliptical-hyperbolic coordinates, etc. The boundaries in many practical

problems cannot be simply expressed in any coordinate system, and often in such cases
resort must be made to other methods such as graphical, point-by-point, computer,
object

or experimental methods.
6-18 are

all

Thus, the

field

solutions of Laplace's equation.

t See Sec. 3-27.

maps obtained

in Sees. 3-22, 4-16,

and

LAPLACE'S

SEC. 7-2]

7-2

POISSON*S EQUATIONS

AND

AND BOUNDARY-VALUE PROBLEMS 263

SOLUTION OF LAPLACE'S EQUATION

IN RECTANGULAR COORDINATES

In rectangular coordinates Laplace's equation (V

dx 2
Let us assume that

dy

V = 0)

becomes

dz 2

as the product of three functions X, Y,

V can be expressed

and Z,

or

V = XYZ
where

(2)

X = function of x only
Y= function of y only
Z = function of z only

Substituting (2) into (1)

we

get

Dividing by

d 2Z

d2 Y

\_d^Y
1
d^X
+
+
l

X~dx
Since the
variable

a- r

Y dy

\_d?Z
i a-z.
(

Zdz

on the left-hand side is a constant (=0) and each

may write
independent, each term must equal a constant. That is, we

sum of

is

2~

,,,

and so we have

XYZ separates the variables,

1

the three terms

X dx 2
a2
d

= ^X
-A
dx

(6)

and

similarly

-j~t

o2

a Z
d7 = ^

and
a\

where

The problem now

is

Hence, the procedure we

separation of variables.

ai

>

(8)

(9)

separately.
to find a solution for each of the three variables
have been following is often referred to as the method of

264

ELECTROMAGNETICS

[CHAP. 7

A solution of (6) is

X = C^ IX + C2 e~" x
where

verified
It

C2

and

are arbitrary constants that must be evaluated

Either term in (10)

conditions.

a solution, or the sum

is

by substituting the solution

K = ((V + C2 e-

Cu C2

etc.,

aiX

are constants.

from the boundary

may be

a solution, as

is

in (6).

follows that a general solution of (1)

1

where

(10)

ai *

)(C 3 e

is

C^-'^C^

Thus, solutions

may

Qe -

"3

*)

take the form of exponential,

trigonometric, or hyperbolic functions.

7-3

EXAMPLE

1:

THE PARALLEL-PLATE CAPACITOR

Consider a parallel-plate capacitor as shown in cross section in Fig.

are infinite in extent

and are separated by a distance xt

zero potential and the right-hand plate at potential

Vt

The

7- la.

The

plates

left-hand plate

at

is

Let us apply Laplace's equa-

tion to find the potential distribution between the plates.

There

is

no variation in

potential in the y

and z directions, so

that the

problem

is

one-dimensional and Laplace's equation (V 2 K = 0) reduces to

d2 V
(1)

~dP~~

For the second derivative of V with respect to x

must be equal to a constant. Thus, we have

to be zero the

first

derivative

dV

dx"

dV = Ct

or
Integrating (3),

Cl

we

(2)

dx

(3)

write

jdV=C!Jdx
v = dx + C2
V=C^x+

or

The boundary conditions

(4)

are
(1)

V=

(2)

V=V

at*

==

atx

==

*i

LAPLACE'S

SEC. 7-3]

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

265

(a)

(b)

FIGURE

7-1

(a) Parallel-plate capacitor and (b) potential variation.

From

the

first

boundary condition

(4)

becomes

=
Hence,

C2 = 0. From the

+ C2

second boundary condition

(5)

(4)

becomes
(6)

so that

-^
C
Cl_
*t

(7)

266

ELECTROMAGNETICS

[CHAP. 7

Introducing the values for

and

C2 from (7) and (5) into (4), we see that the solution

for the potential between the capacitor plates

V=

is

(V)

(8)

x.

Thus, the variation of

V between

the plates

is

a linear function of x, as illustrated in

Fig. 7-16.

This result could have been anticipated since

field

E between

we would

expect a uniform electric

the plates with magnitude

E=
x

(Vm _1

(9)

with the potential at any point given by

V = Ex
The constant

is

(10)

Ordinarily

we

to Laplace's equation to solve this simple problem, but

its

thus equal to the electric field between the plates.

would not have resorted

application here serves to illustrate the procedures involved in applying Laplace's

equation while using a

UNIQUENESS

7-4
If

a function

it is

minimum of mathematics.

V satisfies

a unique solution;

Laplace's equation (V 2 V

i.e., it is

= 0)

and the boundary conditions,

No

the only possible solution.

other function will

satisfy these requirements.!

Let us reexamine the solution (7-3-8) for the parallel-plate capacitor with respect
to

its

uniqueness.

It satisfies

the

boundary conditions, and

it

also satisfies Laplace's

equation, as can be confirmed by substituting (7-3-8) into (7-3-1) and performing the

The

indicated differentiation.

of the form

V=

Qx + C 2

solution

is

Vx = CY x, where Ct = V \xx
x

will also satisfy Laplace's

function

equation but will require

dif-

V = C2 at x = and V = Vt + C2 at x = x
Let us try a function of the form V = Qx 2
This function can satisfy the same
boundary conditions as V = Cy x, but it does not satisfy Laplace's equation since substituting it in (7-3-1) gives a nonzero result (d 2 Vfdx 2 = 2Ct ).
It is apparent that any

ferent

boundary conditions, that

is,

graphically obtained or computer-generated field

problem

is

likewise a unique solution.

map

of a two-dimensional

LAPLACE'S

sec. 7-5]

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS 267

Infinite parallel plates

V=V +C 2
l

FIGURE
The

7-2

potential

conditions;

Cix satisfies
and the boundary

variation

Laplace's equation

C x + C2 satisfies Laplace's
t

equation but not the boundary con2

ditions (unless Ci = 0); while Cix satisfies the boundary conditions but not
Laplace's equation.

power of x other than unity will not be a solution. The various functions we have
discussed are shown graphically in Fig. 7-2. Many different functions might be
tried as solutions, but none except V = (VJxJx satisfies Laplace's equation and the
boundary conditions. // is the unique solution.
More elegant and rigorous proofs of uniqueness are available, but the above
discussion illustrates the general approach, which involves a demonstration that solutions other than the unique one do not satisfy both Laplace's equation and the boundary conditions.

7-5

POINT-BY-POINT,

OR ITERATIVE, METHOD

In this section a point-by-point method of solving Laplace's equation

The method

is

is

discussed.

approximate, but by using sufficient points and by iteration, or repeti-

tive application, the solution

can be made as accurate as desired.f

t J. B. Scarborough, " Numerical Mathematical Analysis," 6th ed., pp. 391-422, Johns
Hopkins Press, Baltimore, Md., 1966.

268

ELECTROMAGNETICS

[chap. 7

3>V,

&y

Vi

by

FIGURE

4'

7-3

V4

Construction to find potential at point P

with respect to four surrounding points.

The

basic step

is

to find a solution of Laplace's equation at a point by noting

the slope of the second derivative of the potential

the point.

dx 2
If the potential variation

one and

first

orthogonal directions around

is

is

given by

8y

dz

(1)

independent of z, the problem reduces to a two-dimensional

(1) simplifies to

The

V in

In rectangular coordinates Laplace's equation

term in

(2) is the

V_

(2)

V with respect to x, that is, the

V with respect to x.f Similarly, the second term
of change of V with respect to y. The sum of these

second partial derivative of

rate of change of the rate of change of

the rate of change of the

two terms must be zero.
is

rate

Consider a two-dimensional potential distribution around a point P, as in

Let the potential at P be equal to V and at the four surrounding points

Fig. 7-3.

be Vu

V2 V3
,

2 and P.
(

V -

^i)/Ax

is

F4

as shown. Now (K2 - V )/Ax is the slope of V between points

approximately equal to dV/dx (becomes exact as A* ->()). Also
the slope of V between points P and 1
The difference of these slopes

and

This

is

Or

since d V/8x

= E,8 2 V/8x2 is equivalent to the negative of the rate of change of

E with respect to x.

SEC. 7-6]

AND

LAPLACE'S

(per distance increment Ax)

is

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

approximately equal to d 2 V/dx 2

in the preceding paragraph, the significance

Hence, as discussed

of Laplace's equation

Kin the x direction and the difference of the

be equal and opposite in sign. Thus, we have
in slopes of

d(dV/8x)

d(dV/dy)

dx

dy

269

is

that the difference

slopes in the

y direction must

(3)

or
[(V2

- F )/As] -

[(V

VJ/Ax]

[(F,

__

- V

)/Ay]

Ax
Letting

Ax =

Ay,

[(V

- V4 )/Ay]

Ay

(4)

we have
Vi

+ V2 +

V *

and
If we

know the potential at points

the potential at the point

P is

i(V1
1,

+ F4 - 4V x

(5)

+ V2 + F3 + FJ

(6)

F,

2, 3,

and 4, then according to Laplace's equation,

as given by (6).

In other words, the potential at

is

the average of the potentials at the four neighboring points.f

EXAMPLE

7-6

To

2:

THE INFINITE SQUARE TROUGH

method consider the infinitely long square trough of

The sides and bottom of the trough
are at zero potential. A cover, separated by small gaps from the trough, is at a potential of 40 V.
Let us use the point-by-point method to obtain the potential distribution.
First, we find the potential at the center of the trough from (7-5-6) as
illustrate the point-by-point

sheet metal

shown

in cross section in Fig. 7-4a.

40

+o+o+Q
4

Next,

iqv

we find the potential at the center of the four quadrants of the trough.

For this
xy axes 45 and take the potential at the cover-trough gap as 20 V
(average of 40 V and 0). Again from (7-5-6) we have

we

rotate the

2Q

+ 40 + 0+

10

_ i75v

t In a three-dimensional problem we would use six surrounding points (four as shown

in Fig. 7-3 plus one a distance Az from
into the page and one a distance Az from
out of the page; Az
Ajc
A^) and the potential at
would be given by

=
=
Ko
i(K, + V2 +

V3 + V*+ V,+ V6)

270

[chap. 7

ELECTROMAGNETICS

^7

40V

t!!_

40V

20

20

-10-

(o)

contour
/

40V

40

40

40

J^ft
-20V
-17.5

7 5

25

21.2

\\l7.2^21.1"17.2/

17.5

10

7.5-

7 5

38

^--lO

2.8

"

~"

5-

'

3.9

2.8

(d)
(c)

FIGURE

7-4

Application of point-by-point method to determine potential at nine points inside

an infinite square trough.

for the upper

and

left

and

right quadrants, as in Fig. 7-46,

0+10 +
for the lower

left

and

=2 .5V

Next,

right quadrants.

and

we can

find

at four

more

points,

with voltages as shown in Fig. 7-4c (xy axes returned to usual orientation).
The procedure is now repeated starting at the upper left, obtaining

4O

21.2

7.5

_ 172v

4
This

new

value of

V is now

used to recalculate the potential at the adjacent point to

the right as

40+17.2+ 10+17.2 =
21

AND

LAPLACE'S

sec. 7-6]

FIGURE

POISSON's EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

7-5

(a) Infinite square trough, (b) Slope or gradient

(c)

271

of potential in x-direction.

Slope or gradient of potential in ^-direction.

way proceeding left to right and top

no longer change. This gives the most
accurate solution to Laplace's equation we can get by this method with the number of
points used. A still more accurate solution can be obtained by using a larger number
See Fig. l-4d.f

to

All points are recalculated in this

bottom over and over again

until the values

of points (subdividing the area of the trough with a finer grid). As the distance
between points (Ax or Ay) approaches zero, the solution can be made to approach an
exact one.

computers.

This method

is well adapted for very accurate calculations using digital

However, manual calculations with relatively coarse grids can often

provide solutions of

accuracy for

sufficient

many

purposes.

as above, of the potential at nine points in this trough example

Thus, a determination,
is sufficient

for drawing

the potential contours as in Fig. 7-4d.

Referring to the

first

step (determination of

V at the center of the

trough), let us

Using the center point and four neighboring points, as in

Fig. 7-5a, we get the slopes of the potential in the x and y directions as shown in Fig.
l-5b and c. (We assume Ax = Ay = 10 units.) Thus, AVjAx = 10/10 = +1, and

reexamine the problem.

t In practice

it

is

column format

convenient to tabulate the voltage obtained in each iteration in

each grid point so that a running record is kept. When the

at

voltages stabilize, the iteration can stop.

272

ELECTROMAGNETICS

[CHAP. 7

FIGURE

AV/Ay =

10/10

7-6

Infinite

square trough with different po-

tentials

on each

or 30/10

side.

Laplace's equation

3.

is

satisfied since the difference

of the slopes in the x direction equals the negative of the difference of the slopes in the

direction. f

Thus,

+ l-(-l) =

+ l-( + 3)= -2

and

We

note in this problem that

dinate,

7-7

it

if

the slope change

must be concave upward

is

concave downward in one coor-

in the other coordinate (see Fig. 7-56

and

c).

SQUARE TROUGH WITH

DIFFERENT POTENTIAL ON EACH SIDE

EXAMPLE

Consider an
Fig. 7-6.

3:

infinitely

The four

long square trough of sheet metal as shown in cross section in

sides have small

potentials of 100, 60, 20,

and

gaps between them, and they have individual

V, as indicated.

Let us use the point-by-point method

with a 9 x 9 grid of 81 points to obtain the potential distribution.

There are 32

on the four sides, where potentials are determined as part of the boundary
conditions, and 49 ( = 7 x 7) points in the interior space, whose potentials must be
determined by the point-by-point method. A manually obtained solution is shown
in Fig. 7-7 with potential contours and field lines drawn in.
points

Note that in the one-dimensional case the difference in slopes in one coordinate
must be zero; that is, Kmust vary linearly with distance as in the parallelplate-capacitor example (Sec. 7-3). In the three-dimensional case all one can say is
that the sum of the slope differences in the three orthogonal coordinate directions
must be zero.

direction

LAPLACE'S

sec. 7-8]

AND POESON'S EQUATIONS AND BOUNDARY-VALUE PROBLEMS 273

FIGURE

100

100

100

100

80

100

100

60

100

7-7

Solution of configuration of Fig. 7-6 using point-by-point method and 49 interior

points.

Potential contours (heavy)

and

field lines (light

and with arrows) are

also shown.

7-8

DIGITAL-COMPUTER SOLUTION

The method discussed in the preceding section is an iterative technique. For greater
accuracy many more points and more iterations are required. The digital computer is
ideally suited for making such computations.
An algorithm can be developed for
systematic computation of the averages involved in the relaxation
(7-5-6)].
is

map

Having developed a

suitable algorithm,

one constructs a flow

method
chart,

[see

which

of the strategy to be pursued, and from the flow chart a series of program

statements

is

written.

This

is

To demonstrate the method, a

the computer program.

flow chart and a Fortran IV programf are given for the square trough with different
potential

on each

side

(Example

3).

grid of only

(as in Fig. 7-7) in order to simplify plotting

and

to

7x7

make a

interior points

direct

is

used

comparison with

Fig. 7-7 possible.

This flow chart and program were written by Dr. Robert

University Radio Observatory.

S.

Dixon, Ohio State

274

ELECTROMAGNETICS

[CHAP. 7

SEC. 7-8]

LAPLACE'S

AND

POEfSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS 275

Fortran

C
C
C

ffi

Program

PROGRAM 10 CALCULATE ELECTRIC POTENTIAL BY THE ITERATION NETHOO.

COMMON Z*9,9)
DIMENSION 22(9.9)
THE DESIRED POINTS LIE ON A 9 X 9 SQUARE GRID IN THE FIRST QUADRANT.
I.E. - Zll.l) IS IN THE LOWER LEFT CORNER.
DATA ITERAT/O/
FILL IN THE KNOWN VALUES ON THE BOUNDARIES.
DO 1 J=2.8
ZI1.JI-60.
Z(9,JI=Z0.
DO 2 1*2 <8
ZII.ll-O.
ZII.91-100.
Z(l, 11.30.
Z<1,9>80.
Z<9,9)'60.
Z(9, 11=10.
CALCULATE THE CENTER VALUE. CASE 0. 9 X 9 SQUARE.
Z ( 5 ,5 1 AVERAG 5 ,5 ,0 ,4 >
CALCULATE THE 4 SURROUNDING POINTS. CASE 1. 5 X 5 SQUARES.

<

Z(3,7)*AVERAG<3,7,1,2>
ZI7,7).AVERAG(7,7,1,2)
ZI7,3).AVERAG(7,3,1,2)
Z13.3)-AVERAGI3.3.1,2>
CALCULATE THE 4 SURROUNDING POINTS* CASE 0* 3 X 3 SQUARES.
ZI3,5)*AVERAG(3.5,0,2>
Z(5,7)'AVERAG<5,7,0,2>
Z(7,5)AVERAG<7,5,0,2>
Z(5,3)>AVERAG<5, 3,0.2)
NOW GO TO THE
SQUARES. CASE 1.
DO 3 IROW2.8.2
00 3 IC0L=2.8.2
Z( IC0L.1R0W)AVERAG<IC0L,IR0W,1,1)
SQUARES IN THE EVEN NUMBEREO ROWS. CASE 0.
NEXT THE IN-BETWEEN
DO 4 IROW=2,8,2
DO A ICOL>3,7,2
Z( ICOL.IROW)>AVERAGIICOL.IROW.O. 1)
THEN THE IN-BETWEEN
SQUARES IN THE OOD NUMBERED RONS. CASE 0.
00 5 IROW>3,7,2
DO 5 1C0L>2,8,2
Zi ICOL.IROWI'AVERAGIICOL. IRON. 0.1)
AT THIS POINT, ALL POINTS HAVE BEEN GIVEN AN INITIAL VALUE.
SO WE START ITERATING ALONG UNTIL THE VALUES STOP CHANGING.
WE DO ONE COMPLETE ITERATION AT A TIME, THEN CHECK TO SEE IF THE LARGEST
CHANGE IS LESS THEN 0.1 VOLT. CASE 0.
DO b I>2*8
DO t J-2.S
221 I.JI'AVERAGtI.J.O.ll
MOVE THE NEW VALUES BACK TO THE ORIGINAL ARRAY, AND FIND THE LARGEST
DIFFERENCE.
01FF2-0.
DO 7 1.2.8
DO 7 J-2,8
0IFF1=Z2(I,JI-ZII,J>
IFI0IFF1.GT.DIFF2)D1FF2=DIFF1

3X3

'

4
C

C
C
C

C
8

6
C
C

C
C

2X2
2X2

Z< I.JI.22CI ,JI

TEST FOR DIFFERENCE. REITERATE AS NECESSARY.
KEEP COUNT OF THE NUMBER OF ITERATIONS.

ITERAT-1TERATH
IF1DIFF2.GT.0.DG0 TO

WRITE OUT THE Z ARRAY ANO THE NUMBER OF ITERATIONS.

WRITE! 6, 11)
NRITEI6,12)1TERAT
ITERATIONS.'./////)
FORMATI
SOLUTION OBTAINED AFTER SIS,
12
FORMAT! 1H1)
11
J=10
00 10 K-1,9
J.J-1
10
WRITEI6.9) (ZII .J). I> 1.9)
F0RMATI1X.9I3X.FS.1)///)
9
GO HOME.
C
CALL EXIT
STOP
END
C

FUNCTION TO COMPUTE THE AVERAGE OF FOUR POINTS, CENTERED AT

SUBSCRIPT (I.J). SPACED "1SPACE" POINTS AWAY.
ICASE'O. THE A POINTS ARE IN THE CARDINAL DIRECTIONS FROM THE
CENTER.
IF 1CASE.1. THE 4 POINTS ARE IN THE 45 DEGREE DIRECTIONS FROM THE
CENTER.
FUNCTION AVERAGII.J.ICASE.ISPACEI
COMMON ZI9.9I
AVERAG.0.25(Z(I-ISPACE,J-1CASESISPACEI
X
Z(I-ISPACE*ICOSE,J*ISPACE

C
C

IF

C
C

Z11ISPACE,J*1SP4CE*ICASEI
*2<I+ISPACE*ICASE,J-ISPACEII

X
X

RETURN
END

276

[CHAP. 7

ELECTROMAGNETICS

The computer-calculated voltages should be printed directly in a square 7x7

The contours can be drawn manually or by an automatic computer-controlled

grid.

plotter.

compare the times required using the manual and computer

Thus, the manually constructed graph (as in Fig. 7-7) can be produced in

It is instructive

techniques.

to

an hour or two, although

time can be considerably reduced with the aid of a vest-

this

The programmable

pocket or desk-top micro- or minicomputer.

digital

computer

do the computations required in seconds, but automatic plotting may require

minutes.
However, many hours may be required to construct the flow chart and write
the program statements, and it takes additional time before the program is ready to
run on the computer. The advantages of the programmable computer is that once a
program has been successfully executed, large numbers of similar problems (as for
will

Furthermore,

other boundary voltages in Fig. 7-7) can be solved very quickly.

map,

more

detail

7-9

ANALOG-COMPUTER SOLUTION

is

desired for all or part of the

In an analog computer the problem

is

if

this is readily obtained.

simulated in analog form.

Thus, in the solution

of an electromagnetic field problem we can use the analogy between Laplace's equation

and Kirchhoff's current

shown

sum

in Fig. 7-8a.

Hence, to determine the potential

law.

Example

the center of the trough of

we connect four

2,

resistors

V of the point P at
of resistance

According to Kirchhoff's current law at a point

of the currents at a junction

is

as

(Sec. 4-12), the

zero, or

h +h+h+h=
From Ohm's law

I1

=(V 1

V)/R, I2

V - V

= (V2 -

V)/R,

V3 - V

V2 - V

(1)

etc.

so

we have

V - V

4
= n
~hR + -hiR + -hiR + -hi
R
1

or

+ V4 -4V =

(3)

V = i(F, + V2 + V3 + VJ

and
which

+V2 +V

(2)
y '

is

the same result

hoff's current

we obtained by applying

Laplace's equation.

(4)

Hence, Kirch-

law at a point and Laplace's equation must be equivalent.

readily seen since Kirchhoff's current

This

is

law from (4-13-1) can be expressed as

V J=

(5)

LAPLACE'S

sec. 7-9]

^-Gap

AND

POISSON'S EQUATIONS

AND BOUNDARY- VALUE PROBLEMS 277

&

y3 =40V

I>

40V

Gap^M

ir

Resistance paper

-v\*

=0

4r
V. =
iooooj Probe

Digital

High-resistance
voltmeter

voltmeter

(DVM)
(c)

(b)

(a)

FIGURE

7-8

of circuit analysis to determination of potential at center of infinite

square trough, (b) Experimental measurement of potential at nine points inside
trough and (c) at any point using resistance paper and digital voltmeter (DVM).

(a) Application

= o-E

But J

and

E = - SV, and

so

we have

V-J = <tV-E = (tV 2 K =

(6)

V2 K =

(7)

or

which

is

Laplace's equation.

If the grid is

made finer by using more resistors and junction points,

distribution can be determined

doing

this using

practical

more

accurately.

a high-resistance voltmeter

and very accurate method

is

is

An

shown

the potential

experimental arrangement for

in Fig. 7-86.

An

even more

to use the equivalent of an infinite grid of points,

namely, a sheet of uniform-resistance paper such as Teledeltos paper. The outline of

the trough is drawn on the resistance paper with silver (conducting) paint, a battery

connected as in Fig. 7-8c, and the potential contours traced out with a high-resistance
voltmeter and probe as suggested.

To

facilitate the

marked, and then more points

the readout, etc.

in Fig. 7-8c

(DVM) can be employed.

and equipotential contours desired at

until 5.000

appears on the readout, the point

measurements a

For example, with 40 V applied as

5-V intervals, the probe is moved

digital voltmeter

at this value are found, next progressing to 10.000

is

on

278

ELECTROMAGNETICS

[CHAP. 7

We have discussed analytical,

and analog-computer
Each method has its advantages and disadvantages. In solving a particular problem one should consider all facets of the
various techniques and then select the one best suited for the particular problem.
graphical, digital-computer,

procedures for solving Laplace's equation.

7-10

EXAMPLE 4: CONDUCTING SHEET

BETWEEN TWO CONDUCTING PLANES

Referring to Fig. 7-9, two infinite parallel conducting plates are spaced a distance

An

infinitely

long conducting

strip is

a.

placed between the plates and normal to them,

shown in the figure.

The width of the strip is only very slightly less than the spacing between the
plates.
The strip is insulated from the plates. Let the two plates be connected
together and a constant potential V be applied between the plates and the conducting
strip.
The medium between the plates is air. Suppose that the plates are at zero
potential and that the strip is at a positive potential of 1 V ( V = IV).
The problem is to
as

find the potential distribution in the region

strip, as

M between the

plates to the right of the

indicated in the cross section of Fig. 7-9Z.|

Plates

FIGURE

7-9

conducting strip between two infinite parallel conducting plates (a) in perspective view and (>) in cross section.

Infinite

t This problem is similar to that for the square trough of Example 2 (Sec. 7-6) except
that here the trough is not closed at the bottom (sides infinitely long). Also in the

present problem the trough has been turned

on

its side.

LAPLACE'S

SEC. 7-10]

It is possible to find
2

F= 0)

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

279

the potential distribution by solving Laplace's equation

It is most convenient to handle this

problem in rectangular coordinates, the relation of the conductor boundaries to the
coordinate axes being as in Fig. 7- 10a. Expanding Laplace's equation in the two
rectangular coordinates of the problem (x and y), we have

(V

boundary conditions.

subject to the

This differential equation

potential with respect to

order and

first

degree.

x and

However,

a partial differential equation of the second

It is

y.

most general way of expressing the variation of

the

is

equation does not

this

For

particular potential distribution in the problem.

of the differential equation which

this

tell

us anything about the

we must obtain a

solution

appropriate to the boundary conditions of the

is

These boundary conditions are

problem.

v=0

(1)

K=

at

(2)

V=

(3)

V=l

at

(4)

V=

at

<x<

oo

0<jc<
x=

oo

0<y<a
X

= 00
<y<a

Proceeding now to find a solution of (1) by the method of separation of variables,

let

us assume that a solution for

V can

be expressed as

V=XY
where

(2)

X = function of x alone
Y=

function of y alone

Substituting (2) into (1)

and dividing by XY, we have

ld2X

X
In (3) the variables are separated.

two terms

is

a constant

(zero),

dx 2

Since

^
l

Y dy 2

(3)

X and Y are independent and the sum of the

each term alone must .equal a constant.

Thus,

we can

write

and
2
where k equals a constant and k

^=- k

k 2 = 0.

(5)

280

ELECTROMAGNETICS

[CHAP. 7

V-0
"

V-l

^> Plates

Strip

V=0
ati-oof*

F=0

(0,0)

(a)

(i)

-v

x=0
Intermediate

x large

(b)

FIGURE
(a)

7-10

Boundary conditions

for potential-distributionproblem of Fig. 7-9.

(b) Potenvariation between plates as obtained from solution of Laplace's
equation at
different distances (x) from the strip
tial

These equations

may

be rearranged to the form

fx -k X=0
dx
2

(6)

d2 Y

-^ + k ^Y ==

and

(7)

Thus, the second-order partial differential equation of

(1) has been reduced to two
second-order ordinary differential equations, each involving but one variable.
These
equations, (6) and (7), have the solutions

X=
and

Y=

One may
(7), in

de** +

C^

ky

C2 e~ kx

+ C4 e-

readily confirm that these are solutions

each case obtaining an identity.

(8)

jk "

(9)

by substituting them into

(6)

and

Introducing (8) and (9) into (2) yields the

general solution

V = C C 3 fe** + C 2 C 3 e- kxeJk" + C,C4 **-*> + C2 C4 e^e'**

t

(10)

which reduces to

V=

C\<* ixm

C e-" (x j
2

(11)

LAPLACE'S

SEC. 7-10]

where C[ = Ct C 3 or
chosen in x + jy.

QC4

AND

and

POISSON'S EQUATIONS

C2 = C2 C3

Because of boundary condition

second term of

(1 1)

or

C2 C^,

(V

(4)

applies in our problem, f

AND BOUNDARY- VALUE PROBLEMS

x=

at

281

depending on which sign

C[

oo),

= 0,

is

so that only the

Also, using de Moivre's theorem,

(1 1)

then becomes

V=C

e-**(cos

To satisfy boundary condition (1) (V =

That

part of (12).

ky 7

aty

0)

sin ky)

(12)

we should retain only the imaginary

is,

V=C'2 e~ kx sin ky
This
It

is

(13)

a particular solution of Laplace's equation appropriate to our problem.

indicates that the potential

To

as a sine function of y.

V falls

satisfy

off exponentially with

boundary condition

(2)

x and

also that

V=

at

y=

it

varies

a) requires

that

k=

(14)

where n

is

a positive integer

(1, 2, 3,

.).

Introducing (14) into (13) yields

K=C^-"^ sin
/a

(15)

now

All the boundary conditions are

x=

0), that

does not

is,

at

x=

for

satisfy this requirement,

more general

V=C

e-" xla sin

^+C
a

We
2

y between

and

e-

at

Obviously (15)

a.

and hence a more general solution

solution can be obtained

for different integral values of n.

(V =

satisfied except for condition (3)

values of

all

required.

is

by taking the sum of expressions

like (15)

then have

2 *x' a

sin

+ C 3 e- 3nxla sin

^+

(16)

where Cu C2 etc., are new constants.

Equation (16) can be expressed more concisely by
,

n==0

V=

Cn e- n%xla sin

n=l

t If

we were

of the

mtv

(17)

interested in the potential distribution to the left instead of to the right

strip (Fig. 7-9), the

boundary condition

is

at

x=

oo

so that C'2

= 0.

282

ELECTROMAGNETICS

The solution for

V given by (16) or (17) is still incomplete since the coefficients Cu C2

To

are not evaluated.

etc.,

(V

[CHAP. 7

at

x=

0),

condition (3)

so that (16) reduces to

Now

we impose boundary

find their values

by the Fourier
f(by)

=C

a + C

sin

2 sin

- + C
a

3 sin

^ +

(18)

sine expansion,

=a

sin

by

+ a2

sin

2by

+ a3

sin

3by

+ an

sin

nby

(19)

where
an
It

follows that for f{by)

"
f

/(*y) sin

A>> rf(*y)

for n

nn

odd

for w even

Therefore, for f (by)

1,

(19) reduces to

Comparing

= - (sin by + $ sin 3by + | sin 56y +

(20) with (18),

it

C 1 =-,C 2 =0,C 3
x

This

is

(20)

follows that

= 0,C
=,Ct
in

Introducing these constants into (16),

V = 1 e-"** sin

=^-,etc.

and

=-

5n

we have

^a +
e"3-/in

sin

^+
a

i- *--/

sin

5n

^+
a

(21)

the complete solution of Laplace's equation for the potential

the boundary conditions of the problem.

position between the plates

Although an

and to the

It gives

the potential

V appropriate to
V as a function of

right of the strip (Fig. 7-9).

number of terms is required for an exact representation of

the potential distribution as a function of x and y, an approximate solution
of practical
value can be obtained with a finite number of terms. Each term attenuates
at a different rate.
Since the higher terms fall off more rapidly with x, only a few terms are
infinite

LAPLACE'S

SEC. 7-11]

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

283

Strip

-Equipotential

FIGURE

7-11

Electric field lines (with arrows)

and equipotentials

in space

between conducting

plates.

needed to give a good approximation except where x

is

Thus, when x

rapid at large x.

is

sm(Tty/a), since the contribution of the

function

The

= 0, V =
of y at x =

At x

ed.

small.

That

harmonics higher than the

is

first

is,

convergence

nearly equal to

may be neglect-

and a. The variation of V as a

and at a large value of x is shown at (1) and (4) in Fig. 7-10i.
two intermediate values of x are presented at (2) and (3). The

distributions at

is

very large, the distribution

for

all

values of y between

shown by the solid curves, the dashed curves giving the approxiby four terms (n = 1, 3, 5, 7) of the series at (2) and by
mate
=
It is apparent that as x decreases, the effect of the
1
and
at
two terms (n
3)
(3).
important.
more
higher terms becomes
The potential distribution can also be presented by means of equipotential
actual distribution

is

distribution as obtained

contours with orthogonal

field lines, as

suggested in Fig. 7-11.

It is

apparent that

graphical, digital-computer, and analog-computer methods could have been used

instead of the above analytical

method to obtain an engineering solution for the poten-

tial distribution.

7-11

SOLUTION OF LAPLACE'S EQUATION IN

CYLINDRICAL AND SPHERICAL COORDINATES

Laplace's equation in cylindrical coordinates

cover)

(r,<f>,z)

(see Fig.

A-l and inside back

is

dr 2

\_dV_

r dr

l_cPV
r

2
8<j>

dz 2

(1)

284

ELECTROMAGNETICS

[CHAP. 7

Using the method of separation of variables

coordinates) a solution

(as illustrated in Sec. 7-2 for rectangular

is

V = V (Ciekz + C2 e~ kz)(C 3 cos n<f> + C4 sin 0)[C5 Jn(kr) + C6 N(kr)]

where n

Jn (kr)

=
=

(2)

integer, dimensionless

Bessel function of

first

kind and of order

with argument kr, dimen-

n,

sionless

N (kr) =
n

Ci,

C2

and of order n

Bessel function of second kind

function), with
-1
constant,

argument

(also called a

Neumann

kr, dimensionless

k=
m
V = constant, V
etc. = constants, dimensionless

Laplace's equation in spherical coordinates

back cover)

(r,6,<f>)

A-l and inside

(see Fig.

is

dr

+
r

sin

Olhj?

2dV
r~dr~

cot0

5F_

~ 80 =

~r1

Using the method of separation of variables, we find that a solution

V = F (d r n + C2 r-<" +1 ))(C3

cos

m<}>

+ C4

sin

w^)[C5 iJm (cos

6)

(3)

is

+ C6 Qm(cos 6)]
(4)

Where n, m =
iy(cos 0) =

integers, dimensionless

associated Legendre function (solid zonal harmonic) of the

first

kind,

dimensionless

gB

(cos 6)

associated Legendre function (solid zonal harmonic) of the second kind,

dimensionless

Cu Cz

etc.

= constants,

dimensionless

V = constant, V
The numerical

calculations involved in solutions like (2)

carried out with a digital

computer especially
functions such as J(kr) or P m (cos 0).

7-12

EXAMPLE

if

and

(4) are

conveniently

subroutines are stored for needed

60AXIAL LINE

5:

Consider the section of coaxial transmission

origin) is short-circuited,
is a,

line shown in Fig. 7-12a.

One end (at the
and the other end is open. The radius of the inner conductor

and the inside radius of the outer conductor

is b.

The conductivity of the inner

sec. 7-12]

LAPLACE'S

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS 285

Cross section

Longitudinal section

IE3

;"

ar finite

W////AV////////////////////////A W7ZtfA

z=0
(a)

Longitudinal section

FIGURE
(a) Section

7-12

of coaxial transmission

(with arrows)

conductor

is finite,

circuiting disk is

Cross section

and

(b) Field distribution

line,

showing field

lines

equipotentials.

but the conductivity of the outer conductor and of the short-

assumed to be

A constant voltage V

infinite.

t is

applied between the

and outer conductor at the open end of the line. The length zt of the line is
long compared with its radius (zj > b). The problem is to find the potential V everywhere inside the line, except near the open end.
inner

The boundary

conditions for this

problem are

(1)

V=

at

(2)

K=

at

z =
a<,r<b
r

=b
<,

(3)

V=V

at'

< zt

[CHAP. 7

ELECTROMAGNETICS

286

There

is

also the condition that at

any distance z from the origin the

its surface (r = a) is given by

electric field in-

side the inner conductor and also along

(!)

f-25.
Bz
Zi
from which we have on integration that

F = ^z

(2)

Z-

At z

= zu

(2)

reduces to the third boundary condition above.

wish to find a solution of Laplace's equation (V V = 0) that satisfies the

boundary conditions. By symmetry V is independent of <t>; so expanding Laplace's
2

We

equation in the other two cylindrical coordinates of the problem

d

dp

J
2

lev <Py
+ --T
+ 7&dz
p dp

Using the method of separation of variables,

(r

and

z),

we havet

&

let

= RZ

(4)

Vo

where

R = function only

of

Z = function only of z
V =

constant,

Introducing (4) into

V
(3)

and dividing by

ld 2 R

R
The

last

term

is

dr

RZ yields

2
J_dR + ld Z
rRdr Z dz 2 ~

a function only of z.

(5)

Thus we can write

d2Z

Z
-^ = a 2,
where a z

a constant.

derivative in (6)
a,

From

(1) it

(6)

follows that since

VJzt

must be zero and hence az must be zero.

is

a constant the second

solution of (6) for

is

Z=C z + C
t

(7)

f Although V is dependent only on p and z, this problem is not two-dimensional in

the sense that the problem of Example 5 is two-dimensional. Here the potential
distribution for a longitudinal plane through the axis differs from the distribution for
all

planes parallel to

it.

LAPLACE'S

SEC. 7-12]

The

last

term of (5)

AND

may now be

AND BOUNDARY- VALUE PROBLEMS

POISSON's EQUATIONS

set equal to zero, so that the

dR

equation reduces to

'

lP + Tr=

287

(8)

A solution is
R = C3
Introducing (9) for

R and (7) for Z in (4)

In r

+ C4

gives the solution for the potential

= (C z + C )(C
2

V =C

or

52

In r

(9)

In r

+ CA)

(10)

+ C6 z+C1 In r + C8

(11)

*o

where

C5 C6
,

etc.,

are

new

the boundary conditions.

To

constants.

When

evaluate these constants,

condition (1)

(V=

at z

= 0)

is

we

introduce

introduced, (11)

becomes

= C7 In r + C8
For

be

(12) to

satisfied for all values

(12)

of r requires that

C7 = C8 = 0.

Thus our

solu-

tion reduces to

=C
y

z\nr +

C6 z

(13)

Introducing

now boundary condition (2) (V

at r

= b), we have

= Cs z In b + C6 z
From boundary

condition (3)

(V =

-+

From

(14)

and

(15)

we

K, at

= C5 z1

=z

lna

and

(14)
r

= a) we

get

+ C 6 z1

(15)

find that

C5 = ti

and

KoZtlnfa/ZO

C =
fi

FoZjlnte/fr)

When the values for these constants are introduced into (13) the complete solution for
the problem

is

This solution

satisfies

w,

z ln(b/r)

Laplace's equation and the boundary conditions and hence

must represent the potential

distribution at all points between the inner

and outer

288

ELECTROMAGNETICS

conductors

[CHAP. 7

(a<r<,b)

except near the open end.

This potential distribution

is

portrayed in Fig. 7-126, the relative potential being indicated for the equipotential
Electric field lines (with arrows) are also

lines.

shown, being normal to the equi-

potentials.
It is interesting

to note in Fig. 7-126 that although the field lines are normal to the

perfectly conducting surfaces (outer conductor

normal to the
in the inner

finitely

conductor

and

short-circuiting disk), they are not

The current and field direction

Comparison of Fig. 7-126 should
4-15 and 4-17 both inner and outer con-

conducting inner conductor.

is

to the left

be made with Figs. 4-15 and 4-17.

ductors are assumed to have

finite

- z direction).
In Figs.

conductivity.

POISSON'S EQUATION

7-13

In the preceding sections, the volume charge density p was assumed to be zero. We
now consider Poisson's equation (see Sec. 3-27), which applies to problems involving

space charge (p

0).

Poisson's equation

is

(1)

where

V = potential, V
-3
p = volume free-charge density, Cm
1
e = permittivity of medium, F m~

Poisson's equation

is

applies at a point.

From

a second-order inhomogeneous differential equation which

the theory of differential equations, it must have a solution

of the following form:

V particular solution +

complementary solution

The particular solution must satisfy the inhomogeneous equation (1) whereas the
complementary solution (which contains two linearly independent solution terms)
must satisfy the homogeneous equation, i.e., Laplace's equation. In the first five
examples of this chapter the volume charge density was zero, and the solution of
Laplace's equation was the complementary solution.

volume charge density

is

In Example 6 of Sec. 7-14, the

is found.

not zero, and a particular solution

In rectangular coordinates, (1) becomes

d2 V
dx-

dy

dz 2

(2)

LAPLACE'S

SEC. 7-14]

If

V is independent

of y and

AND

z, (2)

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS 289

becomes

d2 V
ax*

( 3)

EXAMPLE

6: PARALLEL-PLATE
CAPACITOR WITH SPACE CHARGEt

7-14

Consider the parallel-plate capacitor shown in Fig.

applied between the plates as indicated.

7- 13a

Vt

with potential difference

If the left-hand plate is coated with suitable

material and both plates enclosed in a glass envelope and evacuated, an electron current

passes between the plates

when

called a diode or two-element

cathode,
tial

and the right-hand

the left-hand plate

vacuum

The

tube.

plate or collector

is

is

heated.

the anode.

variation between the plates subject to the applied

there

is electric

charge between the plates,

This arrangement

left-hand plate or emitter

we must

is

is

the

Let us solve for the poten-

boundary conditions.

Since

use Poisson's equation.

V = and a positive voltage V = Vt be

When the cathode is heated, electrons are emitted, so that there

Let the cathode be held at potential

applied to the anode.
is

a volume charge density p(x) between the

From

d2 V
dx 2

p
(1)

must

al eqination

In addition

plates.

(7-13-3), Poisson's equation for this case is

it is

(1)

V=

(2)

V= V

sat isfy tl

at* =
l

at

jc

= xt

assumed that emitted electrons experience a repulsive force from other

and are barely able to escape from the cathode

electrons already between the plates

(initial velocity zero).

field

E(=dVfdx)

is

The

zero at

result of this space-charge limitation is that the electric

x=

0.

This can be expressed as a third boundary con-

dition

(3)

dV
=
E= -

dx

K. R. Spangenberg, "Vacuum Tubes,"

York, 1948.

at

x=

p. 170,

McGraw-Hill Book Company,

New

290 ELECTROMAGNETICS

[CHAP. 7

-||+

v=o

V=V,

Cathode

Anode

x=0
(a)

*ib7' *"= O.Laplace

~j

p*0,Poisson

Space charge limited:

dv
=
dx Oat

x=0

(b)

FIGURE

7-13

(a) Parallel-plate diode with space charge.

As
energy

it

discussed in Chap. 15,

acquires from the

we can equate

(6) Potential

and charge

the kinetic energy of an electron to the

field, i.e.,

imv 2

=eV

where m and e are the mass and charge of the electron, respectively
It is assumed in (2) that the initial velocity of the electron is zero.

The

electrons in

distribution.

(2)

(see

Table

15-1).

motion constitute a steady current of density

J = pv

(Am -2 )

(3)

Solving for v from (2) and substituting into (3) gives

= JT

/
I

(4)

AND

LAPLACE'S

SEC. 7-14]

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

291

Substituting (4) into Poisson's equation (1) yields

(5)

!b?~7 ^2eV
we have

Multiplying by 2dVfdx and integrating,

(dV\ 2

where

is

a constant.

4J

Applying the boundary conditions

=
Hence,

where

C = 0.
x

C2

is

a constant.

C2 = 0.

+C

(1)

and

(3),

we

get
(7)

Integrating again gives

From

the boundary condition (1)

+ C2

=
Hence,

Solving for

V from

(8)

(9)

we havet

HiMr*'"
From boundary

condition (2) this becomes

Dividing (10) by (11) yields

-*er
which

is

(12)

the unique solution to Poisson's equation satisfying the three boundary

conditions.

This variation of

in Fig. 7-136.

Note

V between the plates is shown by the lower solid curve

dy/dx is zero at x = as a result of the space-

that the slope

charge limitation [boundary condition

(3)].

line gives the potential variation in the

For comparison the upper solid (straight)

absence of space charge (solution of Laplace's

t Equation (10) is the particular solution to Poisson's equation.

solution is zero for this problem.

The complementary

292

ELECTROMAGNETICS

[CHAP. 7

The

equation, see Fig. 7-1).

ratio

xjx u which

difference

between solutions

in the exponent for the

is

unity for the solution to Laplace's equation

is

tion to Poisson's equation.

and

4/3 for the solu-

Solving (11) for J gives

(13)

This three-halves power-voltage relation for the diode current

space-charge law.f

Substituting (10) into

(4),

we can

is

the Child-Langmuir

solve for the charge-density

variation, obtaining

pocx" 2/3
This variation

is

as suggested

(14)

by the dashed curve

(See Prob. 7-43 for

in Fig. 7-1 3b.

a semiconductor junction example.)

THE THEORY OF IMAGES

7-15

Consider an arbitrary distribution of charge in space, with a

surface such as the

V=

10

Confining our attention to the volume v

2-13).

follows by superposition that the potential

to and exterior to the volume.

remain unchanged

will

if

If

It

tial

V inside v is a result of charges both interior

the charge exterior to v, V inside

an equivalent surface charge distribution

V=

In

effect, the

at

is

placed on the

lower charge of the dipole

any point in a volume

If

we

bounded by a charged

start

out with some

surface, that poten-

can also be produced by an equivalent charge distribution exterior to

distribution

on the

surface

(0

10 contour surface.

follows that the converse of this argument holds.

known potential

equipotential

we now remove

surface, as suggested in Fig. 7-146.

has been spread over the

known

shown in Fig. 7- 14a (same as Fig.

bounded by the V = 10 surface, it

surface of the dipole

now

v,

the charge

The charges exterior to v are called

3-21.
The examples of the following

being removed.

image charges, as previously discussed in Sec.

sections illustrate the theory Of images.

In Fig. 7-14 the

V=

charge

is

the image of the

+Q

charge with respect to the

10 surface, or for that matter to any of the equivalent potential surfaces.

for example,
plane.

if

the

V=

surface

is

chosen,

we have

charged cylindrical conductor over a

flat

Thus,

the case of a charge over a

plane

is

flat

discussed in the next

section.
t D. C. Child, Discharge from Hot CaO, Phys. Rev., 32: 492-51 1 (May 191 1); I. Langmuir, The Effect of Space Charge and Residual Gases on Thermionic Currents in

High Vacuums. Phys.

Rev., (2) 2: 450-486

(December

1913).

LAPLACE'S

sec. 7-16]

FIGURE
(a) Field

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS 293

7-14

and

potential distribution of an electric dipole.

(6) Distribution inside

V= 10 contour remains unchanged when an equivalent surface charge is

on

the contour

contour

7-16

is

same

and

the lower charge of the dipole

as in (a), but field outside

is

is

removed.

placed

Field inside

zero.

EXAMPLE 7: CHARGED CYLINDRICAL

CONDUCTOR OVER AN INFINITE
METALLIC GROUND PLANE

Consider a long thin uniformly charged cylindrical conductor at a distance s from an

infinite metallic

figuration

is

ground plane, as in

Fig. 7-1 5a.

From

the theory of images this con-

we have two

equivalent to the upper half of Fig. 7-156, where

parallel conductors

of opposite charge separated by a distance

It is instructive

to examine

first

identical

2s.

the two-conductor system.

The behavior of

the electric field of these two conductors can be found by superimposing the electric
fields

due to

either

conductor alone.

This procedure gives the electric

vector superposition of the two components.

Sec. 2-13,

is

(An

to find the scalar electrostatic potential

alternative

V as

field

as the

method, discussed in

the scalar superposition of

294

ELECTROMAGNETICS

[CHAP. 7

(6)

FIGURE

7-15

(a) Field of

conductors.

conductor over a ground plane, (b) Field of two oppositely charged

Field above flat equipotential surface is same as in (a).

Then the electric field is found as the

The electric field lines due to either conductor alone would
point radially outward from the conductor. Thus, on the equipotential surface
equidistant from the two conductors, a simple graphical addition at any point shows
that the resultant electric field is directed downward, as suggested in Fig. 7-156.
Thus, since E is orthogonal to the surface, it must be an equipotential surface.f
the potential due to each conductor alone.
negative gradient of V.)

According to the previously developed image theory, the fields above the
unchanged if we remove the lower conductor and place a surface

surface will remain

charge distribution on the surface.

Ps

This surface charge

is

= A, = - o En

(1)

The
from
the negatively charged conductor is E_
If each line charge is now replaced by a
cylindrical conductor of radius r, positioned so that its surface (shown dashed in
Consider

electric field

first

the case where both conductors are thin lines of charge.

from the

positively charged

conductor

is

E+

and the

electric field

t Therefore, without formally solving for all equipotential surfaces,

from symmetry considerations where one such surface is.

we

already

know

LAPLACE'S

SEC 7_J7]

Fig. 7-15*) coincides with

conductors

As shown

will

AND

POISSON'S EQUATIONS

an equipotential

be undisturbed.

in Fig. 7-156, the

AND BOUNDARY-VALUE PROBLEMS 295

surface, the field outside the cylindrical

From (3-16-1) and

Fig. 3-15,

normal component of the

electric field

can be found from

superposition as

=2cos<|E + =

2cos<

But since cos

<f>

= s/yfs 2 + x2

(3)

is,

(0

+ x2

(Cm- 2 )

(5)

density
(5) indicates that the equivalent surface

surface

at

(4) into (1) gives

tial

is

(3)

^^+x

2
ne s

P^>
Equation

+ x2

becomes

En =
Thus, substituting

L
,

2ne os/s 2

strongest at the point of closest approach

of charge on the equipoten-

to the

upper conductor, that

= 0.

x
If

we now imagine

that a large, thin, perfectly conducting

flat

plane

is

slipped

undisturbed since they are

in along the equipotential surface, the field lines will be
already normal to this metallic sheet, thus satisfying the boundary conditions. Since
the field lines have not been disturbed,

it

follows from the preceding discussion that

plane.
a charge distribution according to (5) will be set up on this conducting ground
its
conductor
and
the
Thus, the field above the ground plane is identical to the field of

image without the ground plane.

7-17

EXAMPLE 8: CURRENT-CARRYING CONDUCTOR

OVER INFINITE METALLIC GROUND PLANE

of static
to this point, the discussion of images has been limited to distributions
parallel
current
a
air
carrying
/
in
conductor
charges. Consider now a filamentary

Up

a height s above, a metallic ground plane, as shown in Fig. 7-16a. By conpositive

vention, the current / flowing to the right is identified with the motion of
v,
that
is,
velocity
a
having
charges to the right, the charges

to,

and

at

PL v

(A)

(1)

296

ELECTROMAGNETICS

[CHAP. 7

Conductor

Conductor

t~

-Conductor

Ground plane

Ground plane

Air

Air

Air

:-^

Image

i.

'.

'.

..^r ::
Image s.

(a)

(b)

FIGURE

/'

and

vertical,

is

slant current-carrying

below the ground plane, equal

direction, as suggested in Fig. 7- 16a,

right.

Hence, the current I

left.

This
will

magnitude but flowing

in

conductor over a

The motion of positive charge

imaged by the motion of negative charge to the

to the motion of positive charge to the

current

(c)

the charge per unit length of conductor.

is

to the right

7-16

Image of a horizontal,
ground plane.

where p L

.:-;l : -v;;

is

equivalent

have an image

in the opposite

i.e.,

/'=-Pl=-7

(A)

(2)

Next consider a filamentary conductor carrying a current I perpendicular to a

ground plane, as shown in Fig. 7-166. Using similar reasoning, it can be

metallic

seen that this filamentary conductor has a vertical image conductor carrying current

same same

shown. The image of a conductor making an arbiground plane can be found by resolving it into vertical
and horizontal components, as shown in Fig. 7- 16c.
in the

direction, as

trary angle with a metallic

PROBLEMS
*

7-1
strip

strip transmission line is

shown

conductor supported by a medium of

ground plane.

If 10

electric flux density

a curvilinear square
squares of the

map

/x r

in cross section in Fig. P7-1.

and

e,

It

consists of a

= 100 on a large flat conductor or

V is applied between the conductors, find (a) the electric field E, (/>) the

D, and
field

(c)

map

in (d) for

the capacitance per unit length C/l (use geometry),

(neglect fringing)

comparison with

and

(c).

(e)

(d)

Draw

use a count of the series and parallel

Give

units for all results.

PROBLEMS

AND

LAPLACE'S

POISSON'S EQUATIONS

AND BOUNDARY- VALUE PROBLEMS

FIGURE
"

7-2

The medium
current
(c)

10

is

the line

If 10

is air.

mA, find (a) electric field at

magnetic flux density

B at Pi

P7-1

Strip transmission line.

Ground plane

cylindrical coaxial transmission line

filling

297

and P 2

is

shown

is

in cross section in Fig. P7-2.

applied between the conductors and the total

A and P

(b) electric flux density

(d) magnetic field

H at Pi and P

D at P
(e)

and

P2

exact capaci-

tance per unit length, C//, (/) exact inductance per unit length Ljl, {g) total electric flux per
unit length Q/l,

and (A) total magnetic

map (both V and E lines)

flux

per unit length

series and parallel squares of the map of (i) for

and inductance values of (e) and (/). Give units for

of the

An

7-3

elliptical coaxial

transmission line

Draw a curvilinear square

and 8 V, and (;') use a count

comparison with the exact capacitance
4, 6,

all results.

shown

is

The medium filling the line has p, = 50 and e r = 60.

and the total current is 5 mA, find (a) E, D, B, and
at

(/')

tfijl.

with equipotential contours at 2,

If

in cross section in Fig. P7-3.

is

applied between conductors

P2 ,Pz, and Pt (16 values),

and inductance per unit length LI I, and (c) total electric
length (Q/l and </>//). Give units for all results, (d) Draw a
points /\,

(b) capacitance per unit length C/l

and magnetic

flux per unit

curvilinear square field

(Teledeltos paper)

map.

The

field

map may

be obtained by graphical or experimental

methods and the answers to the other parts deduced from the map.

Or an exact solution may be obtained, in which case a count of the series and parallel squares
on the map should be made for comparison.
*

7-4

FIGURE

The

sides

FIGURE

P7-2

Cylindrical

shown
and bottom of the trough are 100 mm, and the cover

transmission line consists of a square trough with insulated cover, as

in cross section in Fig. P7-4.

coaxial

transmission

line.

Elliptical

* Answers to starred problems are given

in

P7-3
coaxial

Appendix C.

transmission

line.

298

ELECTROMAGNETICS

[CHAP. 7

mm gaps
t

t
Cover

FIGURE

Trough

P3

P7-4

Square-trough transmission

line.

100

mm

2-mm gaps from the trough. The sides and bottom of the trough are at zero
and the cover is at 10 V. The total current is 1 A. (a) If the medium is air, find
at points P u P 2 and P 3 along the center line of the trough (point P 2 is at the
E, D, B, and
is

separated by

potential,

center of the trough),

1-V intervals,

Show

7-5

Ko

7-6

P)

(c)

Draw

(b)

What

is

Ax

that if

a curvilinear square

with equipotential contours at

Ay

in Fig. 7-3 the potential

at the center point

is

given by

V*

Vi
-

Show

that

in Fig. 7-3 is given

if all

length values in Fig. 7-3 are unequal the potential

at

(point

by

V3

Vi

V2

(1+ *>(.+!)

(l+-')(l+^

map

the capacitance per unit length?

= AxijAxi
= A*i Ax 2
c = Ay IAy*
d = Ay I A/*
Ax = distance from point P at to
Ax = distance to 2
Ays = distance to 3 and so forth
7-7 A long rectangular metal trough

(1

V*

+ C)(l+^

(l

+c -)(l+^

where a
b

in Fig. P7-7.

The trough

is

with insulated cover

at zero potential,

the potential at any point (x,y) inside

is

and the cover

given by

4V

sin

x sinh
x
Xi
t

nodd

nn sinh

niryi

is

is

shown

in cross section

at potential V.

Show

that

LAPLACE'S

PROBLEMS

_,,,

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

299

7-8 If the dimensions of the rectangular trough of Prob. 7-7 are Xi = 160 mm,
= 240 mm, with 4-mm gaps between the cover and sides of the trough, and the lid potential

point-by-point manual
is 40 V, determine the potential distribution inside by means of (a)
method with 4x6 mesh, (A) point-by-point method using a digital computer with 20 x 30
mesh (or larger number), (c) resistance (Teledeltos) paper method, and id) exact analytical
(e) Compare the four methods, in particular at points 20, 40, 60, 80, and 100 mm
solution,
along a bisector from a lower corner, as indicated in Fig. P7-7. Discuss accuracy and

reasons for differences.

25, 30,

drawing

in

P7-9.

lines so as to

7-9

Draw

(J)

the equipotential contours for

V based on one or more of the four methods,

and 35

(g)

V=

1.25, 2.5, 5, 10, 15, 20,

Complete the map of (/) by

form curvilinear squares.

long triangular trough with insulated cover

is

shown

in cross section in Fig.

The trough is at zero potential, and the cover is at 8 V.

All sides are equal.

mine the potential distribution inside by any of the methods mentioned in Prob.
a curvilinear square

7-10
Fig. P7-10.

map

(^Deter(A)

Draw

with equipotential contours at 1-V intervals.

long half-cylindrical trough with insulated cover

The trough

7-8.

is

potential distribution inside

shown

is

in cross section in

and the cover is at 8 V. (a) Determine the

methods mentioned in Prob. 7-8. (b) Draw a

at zero potential,

by any of the

map with equipotential contours at 1-V intervals.

What sheet-current distribution on the surface of a

curvilinear square

7-11

(a)

uniform magnetic

field inside?

/*

= /x

be arranged on the sphere to produce

7-12

everywhere.

(6)

this result ?

steady electric surface current given by

a sphere of radius R.

up by the current
Find Ci and C 2

set

is

is

a constant, and

given by

sphere will provide a

How should the turns of a winding

fi

= fi

B = id toxr<R

K = 4>fco sin

everywhere.

and

exists on the surface of

The magnetic flux density

B = (C 2 /r 3 )(z - f

3 cos 0) for r

> R.

long thin linear conductor carrying a current / extends along the plane
boundary between two media, air and iron. Assuming that the permeability of the iron is
uniform, show that the magnetic field If in air at a radius r from the conductor is
7-13

H =
it is given by H, = poll
conductor
from
the
radius
r
at
a
while
in
the
iron
/*o)wr]
+
P-IIV.H[(/* + MoVr].
7-14 A long square trough is shown in cross section in Fig. P7-14. The sides and
top and bottom are separated by small gaps.

The

and the top

sides are at zero potential,

is

Determine the potential distribution inside and draw a curvilinear square map if
the bottom is (a) at 40 V, (b) 40 V, and (c) 20 V. Any of the methods mentioned in Prob.

at

40 V.

7-8

may be

Compare with Fig. 7-7.

conducting rectangular trough with electrode is shown in cross section in
If the trough is at zero potential, find the required shape of the electrode at

used.

7-15
Fig. P7-15.

below the electrode is of the form V= C

Reduce the equation for the electrode to its simplest form and cone truct
an accurate graph of the trough and electrode for the case Xi =yi.
*
7-16 (a) Find the potential distribution between two infinite metal cones as shown
The potential of the outer cone is zero and of the inner
in cutaway view in Fig. P7-16.
(b) Find the
The half-angle of the inner cone is 6, and of the outer cone is O
cone is Vt
potential K, such that the potential in the trough
sin

kx sinh Icy.

'

300

ELECTROMAGNETICS

[CHAP. 7

Ad

lOd

>

l3d

6d

d<*

2d
(26)

(25)

(24)

(27)

3d

4.

3d]
*o

(30a)

(29)

(306)

(30c)

3d

n-ld
av*M

(30d)

d
*o

5d

id^^

*d

'

:::.-::*?

FIGURE

P7-7 to P7-38
to which each sketch relates

The problem
theses.

(31)

3d

J-01d
(32)

is

indicated by the

(38)

number

in

paren-

PROBLEMS

LAPLACE'S

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

potential distribution between a single cone at potential Vi

7-17

is

flat

metal sheet at zero

at 8 V.

a special case of (a) for which the outer cone half-angle 9

is

The trough has a

cover

and a

= 90.
A long metal trough with insulated cover is shown in cross section in Fig. P7-17.

This

potential.

301

The trough and

central partition.
(a)

tioned in Prob. 7-8

partition are at zero potential,

Determine the potential distribution

may be

used,

(b)

Draw

and the

Any of the methods mensquare map with equipotential

inside.

a curvilinear

contours at 1-V intervals.

Repeat Prob. 7-17 for the long metal trough with partition and insulated cover

7-18

and

fin

shown

in cross section in Fig. P7-18.

Repeat Prob. 7-17 for the irregular trough with insulated cover shown

7-19

in cross

section in Fig. P7-19.

/ = 25 mm has left and right
media of perfect conductivity (o- = oo), as indicated in Fig. P7-20. The
dimensions are h = 20 mm, l 2 = 80 mm, / 3 = 80 mm, and U = 20 mm. If the block con_
ductivity a = 10 U m ', find (a) the resistance of the block and (b) the current density / at
the points P lt P 2 P 3 P*, and ^5 at 10-mm intervals along the bisector of the corner if

7-20

rectangular conducting block of thickness

sides butted against

100

mV is applied

is

the same, find J at the five points (Pj

as

0,

shown

at potential

7-23

is air,

7-24

of

r,

is

map

in Prob. 7-8

shown

metal sheets

edges of the sheets

at zero potential,

and the upper

and charge Q is enclosed by a dielectric shell of

> rC), and outer radius r 3 If the medium else.

in cross section in Fig. P7-24.

r.

If the potential difference

with equipotential contours at 1-V intervals.

may

The medium

be used.

shown

in Fig. P7-25.

map

(a)

Any

and

(b)

draw a

of the methods

is air.

V, enclosed in a grounded conductor of square cross

Determine the potential distribution and

with equipotential contours at 1-V intervals.

(/>)

The medium

draw a

is air.

Repeat Prob. 7-25 for the case where two of the conductors are at zero potential

(grounded) while the other two are at

7-27

is

left

transmission line consists of four cylindrical conductors, two at a potential

curvilinear square

sheet

infinite flat

The

8 V, (a) determine the potential distribution inside

+ 8 V and two at a potential of 8

7-26

between two

coaxial transmission line consists of a rectangular outer conductor and a

curvilinear square

section, as

If the total

and compare with values in Prob. 7-20.

find the potential distribution as a function of radius

between conductors

7-25

etc.)

The lower

metal sphere of radius

cylindrical inner conductor, as

mentioned

x .

relative permittivity e,, inner radius r 2 (r 2

where

P2

in cross section in Fig. P7-22.

are separated by an infinitesimal gap.

is

Determine the potential distribution

7-22

meeting at an angle

sheet

=60 mm and 1% = 240 mm.

Repeat Prob. 7-20 for the case where h

7-21

current

between the ends of the block.

+8

and

conducting wire of radius a

is

V, as suggested in Fig. P7-26.

situated in air in an originally uniform field

Eo perpendicular to the wire. The wire has an insulating coating of radius b and permittivity
e, as shown in Fig. P7-27.
Find the potential V everywhere outside the insulating coating
and the potential V everywhere inside the insulating coating.
t

ELECTROMAGNETICS

302

7-28
infinite

The

[CHAP. 7

electric field

is

uniform and equal to

The

plane conducting sheet.

cut in the sheet,

show

Ea

and

normal to one side of an

is

zero on the other side.

field is

on the

that the surface charge density

side

If

slot

of width j

is

is

0E0
r*

and on the

side

where the

field

\\+;

was

-(s/2) 2 ]"

originally zero

is

given by

eoEo

Wwhere

the distance

is

7-29

shown

in Fig. P7-29.

everywhere above the sheet was normal to

above the sheet and hemisphere.

and hemisphere,

along the

flat

7-30

(c) Plot

(6)

"

placed on a

is

flat

The medium is air, and

and equal to E
Find Kas a function of r and 6 everywhere

it

(a)

Find the surface charge density

at all points

on the

a graph of the surface charge density on the hemisphere and

sheet to a distance of 5a.

Determine the potential distribution and draw a curvilinear square

transmission line consisting of (a) cylindrical conductor of diameter

infinite

conducting sheet of

Before the hemisphere was introduced, the electric

the sheet and hemisphere are at zero potential,

sheet

slot.

conducting hemisphere of radius a

infinite extent, as
field

from the center of the

Cs/2)

metal ground plane as in Fig. P7-30a,

Fig. P7-306, (c) cylindrical

in Fig. P7-30c,

and

(Jb)

placed

map

3rf

for a

above an

insulated cylindrical conductor, as in

conductor above a metal ground plane with

(d) insulated cylindrical conductor

dielectric slab, as

above metal sheet with

dielectric

slab, as in Fig. P7-30rf.

7-31

Determine the potential distribution and draw a curvilinear square

transmission line consisting of a cylindrical inner conductor situated

two grounded
7-32

infinite

map

for a

midway between

metal sheets, as in Fig. P7-31.

Determine the potential distribution and draw a curvilinear square

map

for a

transmission line consisting of two cylindrical conductors with an infinite dielectric slab of
permittivity e

7-33

two

flat

midway between them,

as in Fig. P7-32.

long cylindrical wire of radius a

conducting sheets of

infinite extent

wire runs parallel to the corner.

(a) find the potential

is

If the sheets are

V everywhere in

situated at a distance

which

the corner,

intersect to

grounded and the wire

(b)

Draw

d from

its

center to

form a square corner.

is

at

a potential

a curvilinear square

map

The

t ,

for the

d 10a. (c) Find the surface charge density p s along the sheets for the situation
and plot a graph of p s to a distance of 5d from the corner.
7-34.
*
A thin conducting spherical shell of radius f i is cut into two hemispheres
separated by a very small air gap. If one hemisphere is at a potential Vi and the other at a
potential V2 find the potential everywhere (a) outside the sphere and (6) inside the sphere.
*
7-35 A slot of width s is cut in an infinite horizontal flat conducting sheet. Before
the slot was cut, the electric field above the sheet was everywhere vertical and equal to E
while below the sheet the field was zero. At what distance below the center of the slot is
the potential K equal to (a) 0.1Eo s, (b) 0mEs, and (c) 0.00lEo s7
case where
in (6)

PROBLEMS

LAPLACE'S

7-36

AND

POISSON'S EQUATIONS

AND BOUNDARY-VALUE PROBLEMS

transmission line consists of two thin conducting strips 100

mm.

The medium

mm

303

wide and

Use a conformal transformation to find the

exact potential and field distribution in and around the transmission line and draw a curvilinear square map.
(b) Use an experimental (Teledeltos paper) method to obtain a curvilinear
square map. For this make the strips 1 or 2 mm thick. [In part (a) zero thickness is assumed.] (c) Compare results and discuss reasons for differences.
7-37 The electric field E in the uniform field region between the horizontal plates of
separated by 20

a large parallel-plate air capacitor

is air.

is

10

(a)

kV m" '.

Find

at a point in the fringing field

halfway between the plates and at a horizontal distance from the edges of the plates of
(a) d, (b) 10tf,

and

(c) lOOd,

where d

is

the plate spacing.

Assume

that the plates have zero

thickness.

7-38
is

If

but not in contact with

the plate

is

shown

half-cylindrical metal tube of radius a,

at zero potential.

a metal plate at a potential

it,

show

is

in cross section in Fig. P7-38,

placed across one end of the tube

that the potential inside the tube at a large distance z

from

of the form

K=0-

<R

"",>z sin</

where C = constant
J = first-order Bessel function
Ru =first root of .A(=3.832)
i

Assume

that the tube

is

infinitely long.

Develop a computer program so that the potential distribution for problems

like 7-7 to 7-10, 7-14, 7-17 to 7-20, etc., will appear on a cathode-ray oscilloscope screen for
conductor configurations outlined on the screen by a light pen or "magic pencil." For
7-39

changes

in the

7-40
right-angle

conductor configuration note the changes

(a)

Prepare a curvilinear square

bend connected as

in Fig. P7-40.

map

in the potential distribution.

of a conductor of uniform thickness with

The map may be constructed by experimental

Take a = 3b. (b) From the map

(Teledeltos paper), iterative, or graphical methods.

find the resistance of the configuration in terms of the sheet resistance per square (resistance

of a square of the sample of the conductor of the same thickness),

Compare

(c)

the result

of (b) with that given by Chawla's formula [B. R. Chawla, Proc. IEEE, 60:151 (January 1972)].

<

FIGURE

P7-40
Conductor with right-angle bend.

.-.'

.'.'

.-

v'**

304

[CHAP. 7

ELECTROMAGNETICS

Use

7-41

the

point-by-point (iterative)

symmetrical coaxial transmission

inner conductor of one third the

7-42

and

V=

line consisting

method

to

= x u find

the electric field

in

size.

In a one-dimensional device the charge density p

at

map

of a square outer conductor and a square

= p a (xlxi).

If

at

V.

A semiconductor rod extending in the x direction has a constant cross section A

= 0. If the charge density on both sides of the junction is given by
)find (a) the voltage V as a function of x and (6) the capacitance of the
p = p L x{x + Xi
at x = 0.
Take = at large x and K =
junction between the points x = \00 x
=
10
x
to
jc
^,
versus
and
V
graphs
of
E,

Plot
(c)
p,
7-43

with a junction at x
2

3 '2

8
TIME-CHANGING ELECTRIC AND
MAGNETIC FIELDS

8-1

INTRODUCTION

In the preceding chapters the principles of static electric and magnetic fields have been
considered.
discussed,

In this chapter electric and magnetic fields that change with time are

and a number of new

more important of these

circuit

in a

due to a change of magnetic

conductor moving

and concepts are introduced. Some of the

emf induced in a closed
linking it; (2) a relation giving the emf induced

relations

are (1) Faraday's law, which gives the

in a

flux

magnetic

field; (3)

Maxwell's displacement current, which

represents an extension of the current concept to include charge-free space, (4) the
equivalent conductivity of a dielectric, and (5) an extension of the boundary relations

developed in earlier chapters to include time-varying situations.

8-2

FARADAY'S LAW

In Chap. 5

About

we observed

that a current-carrying conductor produces a magnetic

1831 Michael Faraday in

dendently that the reverse effect

London and Joseph Henry

is

also possible,

That

is,

field.

Albany found indea magnetic field can produce

in

[CHAP. 8

ELECTROMAGNETICS

306

Wire loop

FIGURE

8-1

Relation between decreasing flux density

B and induced current / in loop.

(decreasing)

a current in a closed circuit but with the important qualification that the magnetic
flux linking the circuit

must be changing.

Consider, for example, the closed wire loop

flux density

B is

A magnetic field

in Fig. 8-1.

normal to the plane of the loop.

If

B is directed upward and

ing in magnitude, a current I flows in the wire in the direction indicated. f

that the current

of

and

I is

induced by the magnetic

is

given by the right-hand rule (Fig. 5-2).

decreasing in magnitude.

If

B is

upward

directed

tude, the direction of the induced current

When
in the loop
8-2).

On

the applied flux density

is

in

loop

the other hand,

is

is

is

when B

always

in

is

field

relation

This

is

for the case

as before but

It is

said

between the direction

is

where

is

increasing in magni-

opposite.

decreasing in magnitude, the current induced

such a direction as to produce a

such a direction as to produce a

in the

The

field.

with

decreas-

field

which tends to increase

increasing, the current induced in the loop

opposing B

(Fig. 8-26).

Thus

(Fig.
is

in

the induced current

such a direction as to produce flux opposing the change

in

(Lenz's law).

The changing magnetic

field

around the loop

yields

field

an emf

produces an
"Xf.

nj

That

<fc

electric field E,,.

Integrating this

is,

d\

(1)

When the loop is open-circuited, as in Fig. 8-3fl, this emf appears between the terminals.

Loop-

B(

B( decreasing)

(b)

(a)

FIGURE

increasing)

FIGURE

8-2

Induced currents for decreasing and

8-3

Single-turn loop,

increasing flux density B.

f

direction is that of the conventional current, which

of motion of the electron current.

The

is

opposite to the direction

SEC. 8-2]

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

The
that

electric field

Whereas

charges.

E c = - VK,

is,

Ec

Ee should be distinguished from an electric field Ec due to

can be described as the gradient of an electric potential V,

the field

emf-producing field.

307

Ee

The

cannot.

electric field

Ee may

be regarded as an

This type of field can result from the chemical action in a battery

as discussed in Sec. 4-10.

It

also results, as

we

see here,

from a changing magnetic

field.

The

P in

an

some

V is a single-valued function of position. That is, a point

due to charges has a single potential value V with respect to

electric potential

electric field

Ec

reference point P', such as infinity or the ground.

V=-(
is

independent of the path by which

is

not a single-valued function.

Ec

Ec

is,

given by

-dl

(2)

integrated from P' to P.

is

That

IT

This value of

the

emf IT between two

However, the emf

points 1 and 2 as

given by
2

<V

Ee

-dl

(3)

'i

Ee is integrated from 1 to 2. For example, if the

and 2 of the wire loop in Fig. 8-3 are infinitesimally close together, the
emf IT between them as obtained by integrating Ee from 1 to 2 around the loop is
equal to the line integral of E e around the closed loop. But if E is integrated from 1
e
does depend on the path by which

terminals

to 2 directly across the gap, the resulting

The
magnetic

field

producing the emf

emf induced in a

the total

emf is negligible.
emf induced

quantitative relation between the

magnetic flux linking the

is

closed circuit
circuit.

in

given by Faraday's law.-f

a closed loop and the

According to

this law,

equal to the time rate of decrease of the total

Thus, in symbols,
is

= total emf, V
= total flux, Wb
t = time, s

where IT
\j/

The negative sign indicates that the emf and current directions are positive (the righthand rule relates positive directions) with respect to the direction of the field when the

Michael Faraday, "Experimental Researches in Electricity," B. Quaritch, London,

1839.
Equation (4) is also sometimes referred to as Neumann's law.

^
308

[CHAP. 8

ELECTROMAGNETICS

This flux line passes

through surface once

This flux line passes

through surface once

Surface cutaway

from part of
turn to show
how surface
is attached to
turn below

Surface

bounded
by wire

This flux line passesthrough su rface twice

FIGURE

is

field,

lines

pass through surface4times

8-4

Circuit with 4-turn coil.

surface.

These flux

The wire loop forms

Part of the surface of

the boundary of a single continuous

turn has been cut

away

to

show how

the surface

bounded by the turn below.

and hence the

flux, is decreasing

with time.f

This situation

is

indicated in

Fig. 8-2a.

The
density

total flux

through a

circuit

is

equal to the normal component of the flux

over the surface bounded by the

circuit.

That

is,

the total magnetic flux

is

given by

*-JJbt

as

(5)

write

N turns where each turn links the same amount of flux one can also
A = Aty, where A = total flux linkage and m - flux linked by each turn.

Thus,

(4)

With loops of

ip

becomes

dA

= -"-3T =
d<l>m

Ur

(4a)

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

SEC. 8-3]

The

surface over which the integration

periphery of

conductor

and

Equation

is

the surface

bounded by the

(5) applies to a closed single-

any number of turns or loops. It is important to note that any

with any number of turns forms the boundary of a single surface (see Fig.

circuit of

closed circuit
8-4)

carried out

is

the circuit, as in Fig. 8-4.

\\

Thus, integrating

ds over this surface yields the total flux.

surface in Fig. 8-4 yields

all

over the

Lines of flux passing through the surface only

the flux.

once are integrated once, but those linking

all

4 turns are integrated 4 times since they

Substituting (5) in (4) leads tof

pass through the surface 4 times.

V --5J. B '*
= induced emf, V
B = flux density, T
2
ds = surface element, m
/ = time, s
When the loop or closed circuit

where

309

(6)

'IT

is

stationary or fixed, (6) reduces to

(7)

This form of Faraday's law gives the induced

change of

for a loop or circuit that

emf due

specifically to

a time rate of

fixed with respect to the observer.

is

This

is

sometimes called the transformer induction equation.

8-3

From

MAXWELL'S EQUATION FROM

FARADAY'S LAW: INTEGRAL FORM
(8-2-1)

and

(8-2-7)

we ha vet
nj

d\

= induced emf, V
E = Ee = emf-producing electric field, V
dY element of path, m
B = flux density, T
2
ds = element of area, m
= time, s

= -

"

(i)

where IT

m _1

The symbol

To

J,

indicates a double or surface integral (J J) over a surface

symbol

Ee

s.

be used only where it is desirable to

Since here the integral
indicate explicitly that the field is an emf-producing type.
of the electric field is an emf IT, it is obvious that E is emf-producing and to be
The same situation applies in (8-4-2).
explicit could be written E c
simplify the notation, the

will

310

ELECTROMAGNETICS

This relation

appears in

is

[CHAP. 8

referred to as Maxwell's equation as derived from Faraday's law.

(1) in

its

integral form.

The

corresponding differential relation

is

It

given in

Sec. 8-8.

According to

(1),

circuit is equal to the

density
total

8-4

the line integral of the electric field around a fixed closed

normal component of the time rate of decrease of the flux

B integrated over a surface bounded by the circuit.

emf HT induced

Both are also equal to the

in the circuit.

MOVING CONDUCTOR IN A MAGNETIC FIELD

As discussed

in Sec. 15-3, the force

F on a

particle of electric charge e

velocity v in a magnetic field of flux density

F=
Suppose that the charged particle

is

moving with a

B is

xB

(N)

(1)

situated in a wire

moving with a velocity v through

Dividing (1) by e, we obtain

ey

a magnetic field of flux density B, as suggested

in Fig. 8-5.

the force per charge or electric field intensity E,f or

E=-=vxB
e

(Vrn

-1
(2)

The magnitude of E is given by E = vB sin 9, where 6 is the angle between v and B

The electric field E is of the emf-producing type and is normal to the
plane containing v and B. For example, in Fig. 8-5, v is in the positive y direction, and
(Fig. 8-6).

is

in the positive z direction.

direction or along the wire.|

wire (Fig. 8-5)

is

Hence, crossing v into

yields

The emf IT induced between two

in the positive

points

and 2 on the

then

1T

E-dl=j(yxB)-dl

(V)

(3)

= emf induced over a length / of wire, V

E = electric field, V m _1
d\ = element of length of wire, m
_1
v = velocity of wire, m s
B = flux density of magnetic field, T

where HT

(or E) for this motional case should be primed (as in Sec. 16-7) or
t Properly the
otherwise distinguished from the
for the stationary case, but for simplicity here no

distinction is

assumed that the conductivity of the wire

tangent to the wire need not be zero.

t It is

made.
is finite

so that the component of

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

sec. 8-5]

311

Z(out)
y
*''

^Wire

FIGURE
An emf

fl

B (out)

{
FIGURE

8-5

Wire

8-6

wire and direction of B.

the

is

V = El = vBl

(V)

(3)

and

(3) is the

more

directions of the wire,

is

also

(4)

(4) are motional-induction or flux-cutting laws giving the

induced in a conductor moving with respect to the observer in a magnetic

It is

is

the length of the wire in meters.

Equations

Equation

where v, B, and the wire are mutually perpendicular,

same for all parts of the wire; (3) reduces to

straight wire

uniform, and v

/ is

Relation between direction of motion of

induced in a wire moving

across a magnetic field.

where

is

For a

assumed

its

emf

field.

general form, while (4) applies to the special case where the

motion, and the magnetic field are all mutually perpendicular.

of the wire have the same value of v and that

in (4) that all parts

uniform.

The relations may be used to find the emf induced in any part of a circuit due
to its motion through a magnetic field. They also can be applied to find the total
emf induced in a closed circuit that is moved or deformed in a magnetic field that does
not change with time. For a closed circuit (3) becomes
17

where 17

8-5

is

the total

=jE

emf induced

d\

= j (v x

B) d\

(5)

in the circuit.

GENERAL CASE OF INDUCTION

Equation (8-3-1) gives the emf induced in a closed circuit due to the time rate of change
of

(transformer induction).

circuit

due to

its

motion.

Equation (8-4-5) gives the emf induced in a closed

kinds of changes are occurring simultaneously,

When both

312

i.e.,

is

ELECTROMAGNETICS

[CHAP 8

when B changes with time and the circuit is also

sum of the emfs given by (8-3-1) and

equal to the

V=

<

(v

x B) d\

>

The

in motion, the total

emf induced

(8-4-5), orf

f
J. at

ds

(1)

term of the right-hand member gives the emf induced by the motion, while
emf induced by the time change in B. The line integral
in the first term is taken around the entire circuit, while the surface
integral in the
second term is taken over the entire surface bounded by the circuit.
first

the second term gives the

Equation

(I) is a general relation and gives the correct value

of total induced emf
For the special case of motion only, dB/dt = 0, and (1) reduces to

in all cases.

U=j(yxB)-dl

For the

special case of time

To summarize,

=j

(I)

IT

(II)

V=

(III)

IT

(j>

(v

x B)

8-6

0,

and

(1)

reduces to

the induction relations are

x B) d\ - j

change of flux density only, v

(v

= -

(2)

d\

ds

ds

general case

motion only (motional induction)

B change only (transformer induction)

EXAMPLES OF INDUCTION

In this section seven examples are given in which the total

circuit (total induction) is calculated.

in all cases.

The

general relation

emf induced

(I)

in a closed

gives the correct result

In some cases (motion only or time change only)

(II)

or

(III)

alone

is

sufficient.

A general development of (1) is given in Chap. 16 [see (16-7-32)]. Equation

(1)
can also be derived from (8-2-6) by means of the Helmholtz
formula. See C-T Tai
On the Presentation of Maxwell's Theory, Proc. IEEE, 60: 936-945 (August 1972).
Note that (8-2-6) involves a total derivative and that to evaluate it one
must resolve it
into two terms, as in (1), using the Helmholtz formula.

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

sec. 8-6]

313

^LArea A
V
Sliding conductor

FIGURE

FIGURE

8-7

Fixed loop of area

(for

Example

8-8

Sliding conductor

1).

(for

for increasing loop area

Examples 2 and

3).

Consider the fixed rectangular loop of area A shown in Fig. 8-7. The
1
B is normal to the plane of the loop (outward in Fig. 8-7) and is uniform over the
of the loop. However, the magnitude of B varies harmonically with respect to time as

EXAMPLE
flux density

area

given by

B = Bo cos tot
where B a = maximum amplitude of B, T
"
u> = radian frequency ( = 2nf, where / = frequency), rad s
t = time, s
Find the

total

emf induced

solution

from

(III),

This

the total

is

This

emf appears

a pure case of

emf induced

EXAMPLE

in the

change only, there being no motion.

loop

cm
8B

ds

zero and by

Hence,

is

= AwBo sin cot

(V)

(2)

J, 8t

at the terminals of the loop (Fig. 8-7).

is

in the loop.

17

calculated by (II)

(1)

(I) is identical

Since the velocity v

with that in

= 0,

the

emf

(2).

8-8.
The width / of the loop
x is increased uniformly with time by moving the sliding conductor
The flux density B is everywhere the same (normal to the plane
at a uniform velocity v.
with respect to time. Find the total emf induced in the loop.
and
constant
is
of the loop)
is

constant, but

Consider the rectangular loop shown in Fig.

its

solution
Hence, from

length

This

is

a pure case of motion only, the flux density

Xf

The
the

(y

B) d\

= vBl

being constant.

(V)

entire emf in this case is induced in the moving conductor of length

emf by (III) is zero and by (I) is identical with (3).

EXAMPLE

(II),

(3)

/.

Since 8B/8t

= 0,

Consider again the same loop with sliding conductor discussed in the
The flux density B is normal to the plane of the loop and is

preceding example (Fig. 8-8).

314

ELECTROMAGNETICS

[CHAP. 8

Fixed

view

<\

-*,-

Conducting^

FIGURE

Top

*m

;-:-vt

8-9

loop

with

Examples 4 and

sliding

~\"

dk

(for

strip

Sliding

^^contacts
ids

strip

'"

"

''

Side
v jew

5).

The sliding conductor moves with a uniform velocity v. These

same as in the preceding example. However, in this case let the magnitude

uniform everywhere.
conditions are the

of the flux density

B vary harmonically

with time as given by

B Bo cos wt
Find the

emf induced

total

solution

This

IJm due to the motion

is

is

a case involving both motion and a time-changing B.

from (II), by

The emf

given,

1J.

The emf IT due

(4)

in the loop.

LT,

from

= vBl = vlBo cos wt

B) d\

B is,

to a time-changing

(y

(III),

ds

= mxlBa sin wt

sum of

the emfs of (5)

(5)

(6)

J, ot

According to

(I)

emf IT

the total

is

the

O=|(vXB).rfl-

= vB

COS wt

and

(6),

or

Y' ds

+ eoxBo

sin

mt

= Bo 1 Vv* + (wx) 2 sin (cot + 8)

where S

= tan ~

(7)

(fc/cux)

= instantaneous length of loop

EXAMPLE

The

sliding contacts

circuit for

stationary, but the strip

flux density

mves

B is normal to

for clarity the loop is

circuit.

strip, as

and length x t

longitudinally with a uniform velocity

The width of the strip is /,

shown with a

slightly greater

is

suggested in Fig. 8-9.

the strip and the plane of the loop.

time and is uniform everywhere.

induced in the

a rectangular loop of width

through a thin conducting

the

It is

v.

completed by

The loop is
The magnetic

constant with respect to

same as for the loop, although

width in Fig.

8-9.

Find the

total

emf

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

ggC g_gl

This

solution
is

another case of motion only.

is

Therefore from

(II)

the total

315

emf

given by

U=|(vXB)-rfl = t)B/

The

entire

emf

in this case

arrangement of Example 4

provided by the Faraday disk generator (see Prob.

is

8-11).

Consider now the same loop and strip as in the preceding example (Fig.
the magnitude of the flux density vary harmonically with time as given by

EXAMPLE
let

induced in the moving strip and appears at the terminals.

= 0, the emf by (III) is zero and by (I) is identical with (8). A variation of the

Since SBjSt

but

is

(8)

8-9),

B = Ba cos tot
Find the

emf

total

This case involves both motion and a time-changing B.

solution

emf IT due

to the motion

(III)

the

From

(II)

emf C17, due

= I (y X B)

d\

= vBl = vB

to a time-changing

cos

tot

J.,

(I),

the total

emf ctf

is

the

sum of (10) and

1

cos

tot

(11),

or

+ wxiBo I sin tot

= Bo I vV + {u> Xl y sin (tot +8)

where S

= tan"

EXAMPLE
Fig. 8-10a.

(11)

8t

= cUm + CUT, = vB

<\f

(10)

B isf

SD

-ds^wxiBolsmwt

According to

the

is

*$

From

(9)

induced in the circuit.

(12)

(v/toxt).

Consider next a rotating rectangular loop in a steady magnetic field as in

The loop rotates with a uniform angular velocity to radians per second.

This arrangement represents a simple ac generator, the induced emf appearing at terminals
connected to the slip rings. If the radius of the loop is R and its length /, find the total

emf induced.
solution

Since this

Referring to Fig. 8-10A,

is

it is

a case of motion only, the

At low
effect

poor.

emf can be obtained from

(II).

given by

IT

total

= | (v X B)

d\

= 2vBi sin

(13)

frequencies the effect of eddy currents in the strip can be neglected. The
will be even less if the strip is very thin and its conductivity

of eddy currents

ELECTROM AONETICS

316

[CHAP. 8

Axis

Slip rings

X
B
&''

FIGURE

AC

8-10

generator (for Examples 6 and 7):

(a) perspective view;

(6)

cross

section

(b)

perpendicular to axis.

Since 6

= tor,

we have
'V

The

factor 2

field,

= ZcoRIB sin cot

necessary because there are two conductors of length

is

the emfs in both aiding.

Since 2RI

IT

= A,

moving through the

the area of the loop, (14) reduces to

= <dBA sin tor

= 0, the emf calculated by (III) is zero.

Since dB/dt

(14)

(15)

Hence from

(I)

the

emf

is

as given

by

(14) or (15).

EXAMPLE

Consider finally the same rotating loop as in the preceding example with the

modification that
velocity).

varies with time as given

When t = 0, B =

solution

emf c\3 m due

the

by

B = Bo sin tot (to same as rotation angular

(Fig. 8-106).

Find the

total

emf induced.

This case involves both motion and a time-changing B.

to the

motion
C

From (III)

and 6

emf

U = 2cdRJBo sin 2 cot = wRIBo coRlBo cos 2cot

\3,

due to a time-changing

V, =

From

(II)

the

is

ItoRJBo cos 2

cot

(16)

B is
coRlBo

coRlBo cos Icot

(17)

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

sec. 8-7]

From

(I)

the total

emf "XJ

is

given by the

1J

sum of (16) and

= <V m + nj, = -2wRlB

(17),

317

or

cos Imt

(18)

The emf in this example is at twice the rotation, or magnetic field, frequency. It is to be noted
that the emf calculated from either (II) or (III) alone contains a dc component. In adding
the emfs of (II) and (III) the dc component cancels, yielding the correct total emf given by ( 1 8)

STOKES'

8-7

THEOREM

In Sec. 8-3 Maxwell's equation from Faraday's law

equation

may be

Stokes' theorem, which

is

developed in

this section

stated in integral form.

is

transformed from an integral to a

differential

and applied

This

form by means of

to Maxwell's equation

in Sec. 8-8.

Consider a square of area As in the xy plane as in Fig. 8-11.

Ex

and Ey

Let the electric

The work per coulomb required to

move a charge around the perimeter of the square is given by the line integral of E
around the perimeter. This work equals the total emf around the perimeter. That

field

have components

as shown.

OJ

= <b E

d\

(1)

Dividing by the area A* and taking the limit of this ratio as As approaches zero yields
the curl of E normal to As at the point around which

As shrinks

to zero (see Sec. 5-22).

Thus

E-dl
lim
As-0

where (V x E)

is

the

(V x E)

(2)

As

component of the

curl of

E normal to the area As.

Ey

FIGURE

8-11

Small rectangular area.

Ex

318

ELECTROMAGNETICS

[CHAP. 8

Vi

^ *?
1 o r>
"> 1
*")">

FIGURE

8-12

Ex

Illustration for Stokes' theorem.

now a

x^

as

shown

divided into infinitesimal areas as suggested.

From

Consider

surface of area

carry a charge around an infinitesimal loop divided

of

E at the point.

tions to the total

If the curl

work

of E

is

in Fig. 8-12.
(2) the

by

area

its

is

integrated over the entire area

cancel except for the

work along

Let the area be

work per coulomb

to

equal to the curl

x y lt
t

all

contribu-

the periphery of the area

*i.Vi-

The

situation here

Fig. 8- 13a,

whose

is

as suggested in Fig. 8-135.

are very close together.

their fields cancel.

analogous to that of a single current loop with current

/,

same as a mesh of current loops, each with a current /,

It is assumed that the adjacent sides of the small loops

effect is the

Since the currents in adjacent sides are oppositely directed,

The only currents whose

effects are

not canceled are those along

the periphery of the area of radius R.

annoHD

^np

(a)

FIGURE

(6)

8-13

(a) Single current

loop and

(6) equivalent

mesh of small

current loops.

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

sec. 8-8]

x^ in Fig. 8-12, the integral of the normal component

Returning now to the area

of the curl over the area
the area.

That

x^ must equal the line integral of E around the periphery of

is,

<j>E-rfl=f(VxE)-<fe
Dimensionally

Force
x distance
Since force x distance
if

the curl of

is

the periphery of the

work/charge

x area

area

work,

balanced dimensionally.

(3) is

integrated over an area

same area

That

s.

d\

it is

is

understood

taken around

j (V x E)

ds

(4)

surface
of s

of s

by (3) or

The relation expressed

In (3)

the line integral of E

s,

is,

periphery

(4)

is

called Stokes' theorem as applied to electric

In general, Stokes' theorem states that the line integral of a vector function

fields.

around a closed contour

is

equal to the integral of the normal component of the curl

of that vector function over any surface having

8-8

(3)

of the form

(3) is

Charge

that

319

the contour

C as

its

bounding edge.

MAXWELL'S EQUATION FROM FARADAY'S LAW:

DIFFERENTIAL FORM

By means of
of

Stokes' theorem (8-7-4)

for the line integral of

let

in (8-3-1).

us substitute the surface integral of the curl

That

is,

5B

JJ(VxEWS =-j]--.,*
Since ds in (1) applies to any surface element,
in (1) are equal.

is

arbitrary

and therefore the integrands

Thus

\xE=
This

it is

(1)

SB
(2)

a7

Maxwell's equation, in differential form, as derived from Faraday's law.

integral

form of the equation was given

in (8-3-1).

The

320

ELECTROMAGNETICS

[CHAP. 8

(a)

(b)

FIGURE

8-14

Toroidal coil with two windings.

8-9

MUTUAL INDUCTANCE AND SELF-INDUCTANCE

Consider that two uniform toroidal

turns

electrical

are

wound

interwound as in Fig. 8-14a. Coil 1 of

2 of JV2 turns by a fine wire. There is

coils are

indicated by a heavy wire

Nt
no

is

and

coil

coils.
The ring-shaped form on which the coils
assumed to have a constant permeability n- Coil 1 will be called the
primary winding and coil 2 the secondary. A schematic diagram of the arrangement
is

shown in

connection between the

is

Fig. 8-146.

If the primary current Iy

terminals of the secondary coil

is

not changing.

It is

is

is

constant in value, the

zero, since the flux

assumed that all

emf

"Vl 2

appearing at the

produced by the primary coil

of the magnetic field produced by Ix is confined
\j/

ml

to the region inside the toroidal windings.

Suppose now that the resistance

increases.

is

decreased at a constant rate so that 7t

This increases the magnetic flux

ip ml
Disregarding the negative sign,
Faraday's law that the magnitude of the emf
induced in coil 2 and
appearing at its terminals is
.

we have from

V,=j\r.
dt

(1)

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

SEC. 8-9]

where

321

If a long solenoid is
ml is the magnetic flux produced by the primary coil.
on itself, we obtain a toroid. Assuming that the radius r of the toroid is
large compared with the radius s of the winding (Fig. 8-14a), the flux density B may be
\l/

closed

considered constant over the interior of the winding.

from

(5-11-5),

we

see that the total flux

Obtaining the magnitude of

ml through the toroid

is

>

2
+ ml = Bns =

where

\j/

uNiLns
uN,I.A
^-~
= ^~-

(Wb)

(2)

A = area of winding cross section = ks 2 m 2

/ = mean length of the toroidal coil = 2nr, m
,

Substituting (2) in (1) gives

*,-**
According to
turns

lt

the

(3), the secondary emf IJj is proportional to the number of primary

number of secondary turns N2 the permeability ft of the medium inside
,

the winding, the cross-sectional area

the primary current Ix and

winding.

is

of the winding, and the time rate of change of

inversely proportional to the average length / of the

When we put

M=NN

(3)

(3)

ff

(4)

reduces to

^=A/
Dimensionally

(5)

(5) is

current

Emf=Mr

time
f
M = current
x time = resistance x time
erfl

or

M = ohm-seconds = henrys
Thus, M has the dimensions of inductance, and since M involves two
or

the mutual inductance of the two coils.

coils, it is called

322

ELECTROMAGNETICS

[CHAP. 8

The inductance L discussed

fore, in contrast,

self-inductance

where It
coil

is

the current in the

is

From

in previous sections involves

The emf

called the self-inductance.

a single

form to

(5),

There-

\J 1 applied to a coil of

This relation involving the self-inductance of a

coil.

which involves the mutual inductance of two

(5-13-4) the self-inductance of a toroid

,uA
L = N2
=

N2

coils.

is

(H)

= lF

(7)

N = number of turns of toroid, dimensionless

where

3t

1/nA

From

= reluctance of region enclosed by toroid winding, H -1

(4) the

mutual inductance

M=iV
where

M of two

coils (as in Fig. 8-14)

N N2

N,N 2

uA

^T=7^r = ir

(8)

N
N

= number of primary turns, dimensionless

=
number of secondary turns, dimensionless
2
01 = reluctance of magnetic circuit linking primary
H" 1
t

and secondary windings,

battery and resistance be connected across coil 2 (Fig. 8-14), and

1

is

(H)

Consider next the converse of the situation described above.

coil

coil.

<

L t is

similar in

is

be

left

open.

Then

the

emf

I?! at the terminals of coil

IT,

= N,

let

That

is, let

the

the terminals of

is

(9)

at

where

i/% 2 is

the magnetic flux produced by the secondary coil.

Wml

But

l A
= nN2 2
J

and so

(9)

becomes

^1 = ^i^2
or

u.A

dh

y -f

^ i=M ^

_.
(io)

(U)

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

SEC. 8-10]

Thus, from

(5)

and

(11),

M = dlV,

17,
(12)

dljdt

2 jdt

Therefore,

if

323

a given time rate of change of current in the primary induces a certain

voltage in the secondary, the

same time rate of change of current applied to the second-

ary will induce the same voltage in the primary.

In effect this

is

a statement of the

reciprocity theorem as applied to a special case.

If the current

It varies harmonically with time (alternating current), we have

(13)

For harmonic variation ofIt and I2

(12)

becomes
17 2

17,1

M= jcoi

(14)

17.

17,

J2

yco/j

and

(15)

where Zm

8-10

is

the mutual impedance in ohms.

THE TRANSFORMER

Suppose that the two-coil arrangement of

shown, while the primary

coil is

Fig. 8-15 has the secondary coil open, as

The emf 172

connected to an alternating emf Uj.

appearing at the secondary terminals has a magnitude

17 2 =JV.

where

\j/

(1)

dt

= number of turns of secondary coil

= magnetic flux through transformer core
/-Iron core

>

Nl\

'

FIGURE

8-15

Transformer.

Pri.

'."'.'*
*

Sec.

^2

324

primary resistance

If the
is

[CHAP. 8

ELECTROMAGNETICS

negligible, the counter

is

emf CU 1

equal in magnitude to the applied

emf

across the primary terminals

jt

O l = Ni

where

is

the

(2)

~f

number of turns of the primary

Dividing

coil.

(1)

by

(2) yields

^=^
According to

(3) the ratio

of the secondary to the primary emf

of the number of secondary turns

ratio

is

the

same

(3)

is

equal to the ratio

number of the primary turns

to the

The

for effective (rms) voltages as for instantaneous.

Because the arrangement of Fig. 8-15 (or Fig. 8-14) can transform an emf or
voltage from one value to another,

the transformer
that

\j/

is

m links all

an

ideal

is

is

called a transformer.

It is

may be approached

connected to a high-resistance

In the present discussion

assumed to be zero so
assumed that the recurrent is drawn from the
is

further

very small and that negligible

This condition

secondary.

it is

in the sense that flux leakage

primary and secondary turns.

sistance of the primary

secondary

one

in practice

where a transformer

circuit.

EXAMPLE An ideal transformer has a turns ratio of 2, that is, NijNi = 2. An ac emf
of 10 V rms is applied to the primary. Find the emf or voltage appearing at the secondary
terminals.

solution

From

(3) the

secondary voltage

<\J 2

8-11

= cU =2x
1

is

10

= 20 V(rms)

ALTERNATING-CURRENT BEHAVIOR
OF FERROMAGNETIC MATERIALS

We noted in

Chap. 6 that the permeability of iron

is

not a constant.

In spite of

the permeability of the iron in an iron-cored coil carrying alternating current

taken as a constant for certain applications, but

its

this,

may be

value, in this case, needs further

explanation.

Where n
is

is

not a coastant,

we have from

(5-12-2) that the inductance

of a

coil

given by

L=

(H)

(1)

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

SEC. 8-11]

'

325

Slope- ordinary
permeability

at Bmax.

j- Slope =5^

Slope -ordinary permeability

at

FIGURE

8-16

Hysteresis loop illustrating ordinary

If there

is

no

flux leakage,

lBi,Hi)

A = N\ji m

and

differential permeabilities.

so

L=

(2)

dl

For a toroidal type of coil, d\/i m = A dB, and dl = I dHIN, where

and / equals the mean length of the coil. Therefore (2) becomes

equals the area

N A dB
2

L
In (3)

dB/dH has

hysteresis curve.

permeability,

point P.

the dimensions of permeability.

At some point P, as

BJH

Since

-TdH

lt

which

dB/dH

is

(3)

It is

in Fig. 8-16, this

is

equal to the slope of the

different

from the ordinary

equal to the slope of the line from the origin to the

involves infinitesimals,

it is

sometimes called the

infinitesimal

or differential permeability.
If alternating current

is

applied to an iron-cored coil so that the condition of the

moves around a hysteresis loop (Fig. 8-16) once per cycle, the slope dB/dH varies
over a wide range and the instantaneous value of the inductance will, from (3), vary
over a corresponding range. Under these conditions it is often convenient to consider
the average inductance (over one cycle) as obtained from (3), using the average value

iron

ELECTROMAGNETICS

326

[CHAP. 8

5\K
^7

Bl

Bi

FIGURE
Minor

AR
t

v Minor

loop

8-17

hysteresis loop illustrating incre-

mental permeability.

of the slope dBldH.

of

This

is

equal to the ordinary permeability at the

maximum

value

Thus

attained in the cycle (see Fig. 8-16).

N A/dB\ _N 2 A
2

where n

= BmnIHmx^

The above
coil,

is

the ordinary permeability at

discussion

alternating current

through the

is

(4)

\dHj a

is

i? max

for alternating current only through the coil.

superimposed on a

the situation

is

hysteresis loop

and

dBldH
is

is

a small

The magnetic condition

as suggested in Fig. 8-17.

of the iron then follows a minor hysteresis loop as indicated.

value of the slope

If

relatively large steady, or direct, current

In this case the average

given by the line passing through the tips of the minor

called the incremental permeability

\dH) t

'

n inc

Referring to Fig. 8-17,

(5)

AH

The incremental permeability

is much less than the ordinary permeability fij^ for

a point at the center of the minor loop in Fig. 8-17.

In (8-9-7) the self-inductance

of a toroidal

L = iV-2

coil

of TV turns

is

given as

fiA

(H)

(6)

and

in (8-9-8) the

shaped core

is

mutual inductance

M of two

coils

mounted, for example, on a ring

given as

M=NN
t

fxA

(H)

2
I

(7)

TIME-CHANGING ELECTRIC AND. MAGNETIC FIELDS

SEC. 8-12]

where

N
N

327

= number of primary turns

= number of secondary turns

For iron cores and ac operation these

relations

may be used

to calculate the average

L or average M provided the appropriate value of n is used. As discussed above, this

value of

/x is

equal to the average value of

dB/dH over the

operating range of the iron

on a BH diagram.
Although the above discussion has concerned toroidal coils wi'.h uniform core
cross section A, Eqs. (6) and (7) may be applied in the more general case, where the
iron core is of nonuniform cross section and the magnetic circuit may include an
air gap, by reexpressing them as

L=^
where

81

is

ary

M==
is

coils.

(8)

the reluctance of the closed magnetic circuit through the coil,

and
where 01

(H)

^P

(H)

(9)

the reluctance of the closed magnetic circuit linking the primary and second-

The above

relations are applicable both to uniformly distributed coils

to concentrated coils (as in Fig. 8-16) provided that flux leakage

reluctance of the magnetic circuit

8-12

is

is

calculated as discussed in Chap.

negligible.

and
The

6.

EDDY CURRENTS

When large conducting specimens are subjected to transformer or motional induction,

They flow in closed paths in the specimen
and are called eddy currents. In accordance with Lenz's law, the eddy current tends
to oppose the change in field inducing it.
Eddy currents result in joule heating in the conducting specimen. The energy

currents tend to be induced in the specimen.

loss

due to eddy currents

in the ferromagnetic cores of ac devices

is

in addition to the

energy lost in the magnetization-demagnetization process (proportional to the area of

the hysteresis loop), as discussed in Sec. 6-14. In order to reduce the eddy currents
in iron-cored ac devices, the core is

commonly made of thin

sheets or laminations of

Thus the eddy currents are confined

reduced. Each sheet is continuous in the

iron insulated electrically from each other.

to individual sheets,

and the power

loss is

direction of the magnetic flux through the core, but because of

relatively large reluctance.

By

stacking a sufficient

number of

its

thinness

it

has a

sheets in parallel the

To
total reluctance of the magnetic circuit can be reduced to the desired value.
reduce eddy currents to a minimum, iron wires are sometimes used in place of sheets,
while at radio frequencies powdered-iron cores are commonly employed.

328

ELECTROMAGNETICS

[CHAP. 8

(a)

FIGURE

(6)

8-18

and (b) Conduction current through resistor and displacement current through
capacitor,
(c) Conduction and displacement currents through conducting dielectric medium.
(a)

8-13

DISPLACEMENT CURRENT

In this section a

new concept is introduced,

and a capacitor

voltage applied to a resistor

that of displacement current.

Consider a

in parallel, as in Fig. 8-1 8a.

The nature

of the current flow through the resistor is different from that through the capacitor.
Thus a constant voltage across a resistor produces a continuous flow of current of
constant value. On the other hand, there wili be a current through the capacitor only
while the voltage

is

changing.

For a voltage V across a

resistor

we have a

in parallel, as in Fig. 8- 18a,

of resistance

R and

C
= V/R

capacitor of capacitance

current through the resistor given by

i1

and a current through the capacitor given by

dt

(1)

dt

The instantaneous charge Q in the capacitor is given by Q = CV.

The current through the resistor is a conduction current, while

the current

" through " the capacitor may be called a displacement current.

Although the current
does not flow through the capacitor, the external effect is as though it did, since as
much current flows out of one plate as flows into the opposite one. This circuit
concept may be extended to three dimensions by supposing that the resistor and capacitor elements each occupies a volume as in Fig. 8-1 Sb.

Inside each element the electric field

by

its

length

d.

resistor equals the

That

is,

E=

V/d.

E equals
From

product of the electric

inside the resistor element;

it is

Fringing of the field

the voltage

(4-9-4) the current density

field

also equal to

E and the conductivity a

it

is

neglected.

V across the element divided

Jt

inside the

of the

medium

divided by the cross-sectional area

Or

J^Eo =

(2)

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

SEC. 8-13]

The dimensional form of (2)

in SI units

Amperes
-

Meter 2

The capacitance of a

the plates and

V=

Ed, into

is

mhos

volts

meter

meter

parallel-plate capacitor

the spacing between them.

is

329

is

C = eAjd,

where

is

the area of

Substituting this value for C,

and

(1) yields

eAd dE

Dividing (3) by the area

dE

= -TTt = A Tt

(3)

gives the relation that the current density

the capacitor equals the permittivity of the nonconducting

element multiplied by the time rate of change of the electric

'2

r
=J

~A

The dimensions of

inside

field.

Thus

dE

tA\
(4)

Tt

(4) in SI units are

Amperes
5-

Meter
Recalling that

J2

medium filling the capacitor

D = eE,

(4)

farads

volts/meter

x
second

meter

becomes

'-"
Jt is a conduction current density Jcond while J2 is a displacement
Also, since the current density J, the electric displacement D,
Jdis
and the electric field intensity E are actually space vectors, which all have the same
direction in isotropic media, (2) and (5) may be expressed in more general form as
In this example
current density

Jcond

and

jdisp

= *E

(6)

do
_ = __

(7)

dE

As a final step suppose that instead of having two separate elements in parallel,
one of which acts like a pure resistance and the other like a pure capacitance, we have
only one, which has both capacitance and resistance. Thus, as in Fig. 8- 18c, there is
a single element filled with a conducting dielectric so that both conduction and displacement currents are present.

Then
Jtotal

the total current density J, ota

Jcond

+ Jdisp

is

()

ELECTROMAGNETICS

330

[CHAP. 8

The concept of displacement

in

empty

due

Here the conduction current

space.

was

current, or displacement-current density,

introduced by James Clerk Maxwell to account for the production of magnetic

zero,

is

and the magnetic

fields

fields

are

entirely to displacement currents.

8-14

MAXWELL'S EQUATION FROM AMPERE'S LAW:

COMPLETE EXPRESSION

According to Ampere's law, the

the current enclosed.
this current

currents,

is

may

line integral

of

H around a closed contour

Where both conduction and displacement currents

Thus

the total current.

be extended as follows

(5-15-1),

is

equal to

are present,

which applies only to conduction

when both conduction and displacement

currents

are present:

j>

d\

(J cond

Js

flWl = /

or

line integral

of

on the

E + g).

(2) is

as

(1)

ds

(2)

around the boundary of the surface s

Each term in (2)
(2).

taken on the right side of

is

has the dimensions of current.

of

left side

over which the surface integral

ff

The

+ Jdisp)

The conduction current through the

surface s

is

given by

cE

ds

while the displacement current through the surface

s is

given by

3E
J
Equation

(2) is the

er-'ds
dt

complete expression in integral form of Maxwell's equation

derived from Ampere's law.

It is

where J without a subscript

is

also often written

dl

3
is {

I)-

d*

(3)

understood to refer only to conduction current density.

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

sec. 8-14]

By an

331

application of Stokes' theorem to (3) or by an extension of (5-21-12) to

include displacement currents, the complete expression in differentialform of Maxwell's

equation derived from Ampere's law

is

dD
VxH = J + dt

Vx

or

H = <tE + e

(4)

BE
(5)
~Bt

It

(E

=E

should be noted that when the electric

That

rature.

field varies

harmonically with time

conduction and displacement currents are in time-phase quad-

sin cot), the

is,

= aE =

crE

sin cot

(6)

8B
BE

= -T- = a>eE cos cot

and

Thus, when
current

mum

is

cot

zero.

= 0,

the displacement current

On the other hand, when

and the displacement current

maximum

one-quarter cycle

(cot

(7)

dt

dt

is

cot

zero.

is

maximum and

k/2, the

n/2) before the conduction current,

situation in a circuit having a resistor

"through" the capacitor

is

a maxi-

Since the displacement current

the displacement current leads the conduction current by 90.

the current

the conduction

conduction current

and a capacitor

This

is

it

is

is

said that

similar to the

in parallel (Fig. 8- 18a) in

which

leads the current through the resistor by 90.

This phase difference can also be readily shown by expressing the time variation

of the

field

byf

E=E

e J ".

The displacement current

BE

e = ejcoE
Maxwell's equation

(5)

is

then

e J "=jcoeE

(8)

then becomes

V x

H = oE +jcoeE = (p +jcoe)E

(9)

fin using this notation it is understood that the instantaneous value of the
by the imaginary part of
e 1 "'. Thus

field is

given

(instantaneous)

=E

Im

e J "

sin ait

which is equal to cos ait (that is, Re eJ" =

cos ait). However, in any given analysis one convention or the other should be
adopted and used consistently. Here it is understood to be the imaginary part.

One might

also use the real part of e

J ",

[CHAP. 8

ELECTROMAGNETICS

332

j in the displacement-current term and its absence in the conductioncurrent term signifies that the displacement current is advanced in phase by 90 with

The operator

respect to the conduction current.

DIELECTRIC HYSTERESIS

8-15

In dielectrics that are good insulators the dc conduction current may be negligible.
However, an appreciable ac current in phase with the applied field may be present
This

because of dielectric hysteresis.

sis in

phenomenon

is

analogous to magnetic hystere-

Materials, such as glass or plastics, which are

ferromagnetic materials.

good

may consume considerable energy in alternating

this way is sometimes applied in industrial radio-

insulators under static conditions

The heat generated

fields.

in

frequency heating processes.

As explained

in Sec. 3-5, the electron

cloud of an atom in a dielectric becomes

when an electric field

and the atom is said to be

slightly displaced or asymmetrical

(moment Ql),
removed, the atom returns

electric dipole
field is

to

its

is

applied.

polarized.

normal, or unpolarized,

This produces an

When

state.

the electric

If the electric

again applied but in the opposite direction, the dipole will be reversed. Thus,
an alternating field is applied to a dielectric atom, the dipole goes through the

field is

when

successive stages suggested in Fig. 8- 19a.

An

equivalent mechanical system

is

shown

The large sphere represents the large mass of the nucleus. The small
in Fig. 8-196.
sphere (attached by a spring and moving through a tunnel in the large sphere) represents
the small mass of the electron cloud.

The atom

constitutes

friction) coefficient d,
is

an electromechanical system with mass m, damping (or

and tension (or spring) constant

s.

The behavior of the system

described by |

m^+d^ + sl^gEoe^'
at

where

=
=

(1)

at

dipole length, or separation

dipole charge

term involves mass times acceleration, the second term damping coefficient
The
and the third term spring constant times displacement. The rightvelocity,
times
hand side is the driving force (qE = peak force) resulting from the harmonically
first

varying applied electric

field.

t P. Debye,
1929.

Equation

(1) is

" Polar Molecules," chap.

a second-order differential equation

5,

Chemical Catalog Company,

Inc.,

New York,

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

sec. 8-15]

333

Applied field

Unpolarized atom

E=

/ N*-Polarized atom

/^~\

/ +

-q

+q

Moving mass,
charge,

FIGURE

Atomic dipole

Equivalent mechanical system

(a)

(6)

atomic dipole in alternating

mechanical system.

solution

damped harmonic

electric field

oscillation with

and

(6) equivalent

a driving function.

/""

(2)

(q/m)E

where
co

co

co

-co 2

moment per
2

co
is

the

number of

unit

volume

d = 0).
when E -

(1)

0.

From

is

(Nqi/nOEoe*

p = Nql =

(for

can be determined by solving

co

(3-6-1) the polarization or dipole

But =

(3)

+ (jcodlm)

= driving radian frequency

= resonant or natural radian frequency =y/s/m

The resonant frequency

where

and

is
/

where

8-19

(a) Behavior of

(of standard form) for a

-q

co

(4)

+ (Jiodjm)

polarized atoms per unit volume.

Now

+ PjE,

e-"';so

+
co

Nq 2jm
-co 2 + (jcodlm)

(5)

334

[CHAP. 8

ELECTROMAGNETICS

But writing

=C

-je"

we have
(Nq
V"I
2
(w
(o

= [,1+
,

(Nq

<W We

/c
le

2
- co 2 )
m)(co
~"/
p"W"0

-
- co

**-- 2
+'
+ (cod/m)

2 2
)

(f.\

()

m)(codlm)

le

2 2

+ M/0

"> )

(7)

observe that e has real and imaginary parts which are both dependent on
=
Putting J = <tE and e = e'-Je" in Maxwell's equation

VxH

frequency/)-

J + jcoeE, we obtain

Vx
It is

apparent that

e"

H =j<oe'E +

(imaginary part of

is

small in a

good

(co large) losses

constitutes

dielectric for

which a

(8)

coe")E

= and therefore coe" = 0) power loss

small.

is

(8)

involved in a frequency-dependent term

(co

However, at high frequencies

The sum of a and we"

significant.

what may be called the equivalent conductivity

a'.

Thus,

coe"

(9)

can be expressed

V x
It

can become larger as me" becomes

a'

Now

e) is

At dc

with the dimensions of conductance.

(<r

H=J

(10)

total

follows that

J total

(r'E+j(oe'E

(11)

Referring to Fig. 8-20, we see that the total-current density JtoM is the sum of a
conduction-current density (a'E) and a displacement-current density coe'E in time-

From

phase quadrature.

Fig. 8-20

we have

tan 5

(12)

coe

The quantity tan 5

is

is

called the loss tangent.

the power factor (PF).

Thus

(for small

PF =

cos

Also the cosine of the angle 6

= 90 -

d)

<5)

~ tan = -^<5

(13)

coe

and 8-33 illustrate how e' and * change with frequency. Over very
wide frequency ranges (radio to ultraviolet) materials may exhibit several such resonances and associated permittivity changes due to various vibrational modes.

t Probs. 8-32

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

sec. 8-15]

FIGURE

8-20

Time-phase diagram for

dielectric

with

losses.

The power

dissipated per unit

power

volume

medium

area

'

conduction current

is

current voltage

volume

EXAMPLE

= JE = a'E 2

(Wm -3 )

Find the average power dissipated per cubic meter in a nonconducting

solution

is

10

a'

2tt

It is

= coe" and tan 8 = o-'/cue' = e'/e', or

e' ; so a' = toe' tan 8cue tan 8, or

= 0,

and

a'

x 10 7 x 4 x

dissipated per unit

if

E=

e"

= ' tan 8

kV

dielectric

m _1

rms

MHz.

Since a

tan 8 is small, * <$ e'

The power p

(14)

length

with relative permittivity of 4 and a loss tangent of 0.001

and the frequency

335

o-'

8.85

volume

= 10 s

is

x 10" 12 x 10" 3 =2.22/iO

m"

and

since

then

2.22

x 10- 6

= 2.22 W m" 3

to be noted that the current density and the field intensity have the

same

Although both may have the same space direction,

time phase. Thus, in the above discussion
have
different
may
scalar
magnitudes
their
the space directions are the same and are fixed, but the scalar magnitude /total of the
space direction in isotropic media.

total-current density leads the scalar magnitude

by the phase angle

may be

9.

of the electric

represented by vectors separated in direction by an angle

vector used to represent the time phase of a scalar quantity

A dot () is sometimes added when

is

field intensity in

a complex function of time,

i.e.,

it is

is

0.

time

and E
Such a pseudo

Thus, on a time phase diagram, such as Fig. 8-20,

J, olal

often called a phasor.

desired to indicate explicitly that the quantity

a phasor.

336

ELECTROMAGNETICS

[CHAP. 8

Thusf

E = &E = &E
=

where &

In (15) the quantity

"

(15)

unit (space) vector in direction of

= E eJ" = vector representation

E = amplitude of E

eJ

of time phase (phasor)

However,

a scalar function of the time.

e"*' is

counterclockwise one revolution per cycle.

writing

E in its equivalent polar form.

E=E
which

interpreted to

is

mean a

This

That

and

cot

is

ti/2)

in the positive

may be made more

obvious by

is,

[mt_

(16)

vector (or phasor) of magnitude

axis).

It is
is

at

an angle

direction.

later (t

At

understood that the instantaneous magnitude of the

Referring to

eJat or

phasors

(1 1), it is

E Q /cot. It
at t = 0.
As

understood that

= &E

E in this

it is

taken equal
is

smcot

equation

(17)
is

a phasor as given by

follows that Fig. 8-20 shows the position of

E and

the other

a function of time the entire diagram rotates counterclockwise

one revolution per cycle, and the instantaneous value of any of the quantities
by its projection on the y axis.

8-16

T/2 the

electric field intensity

When

to the imaginary part, the instantaneous value of the electric field intensity
e J "

etc.

given by the real (Re) part or imaginary (Im) part of E.

E = &ImE = &ImE

cot

= 0,
= T/4

Hence, when t

One-quarter of a period

direction.

the phasor has rotated 90 to the positive

phasor has rotated a total of 180 to the negative x direction,

may be

that rotates

with respect to some reference direction (usually the real or x

the phasor

it

represented on the complex plane by a vector (or phasor) of magnitude

is

given

BOUNDARY RELATIONS

The boundary
components of

relations given in Tables 3-2

static electric

and magnetic

This can be shown as follows.

Consider

and 6-2 for the tangential and normal

fields also

first

hold for time-varying

fields.

the tangential components E, of the

The quantity E is not only a space vector but also a function of time (phasor).
Thus, it may be called a vector-phasor and designated by E (= &E). In general, we
shall use the dot over a quantity only where we wish to indicate explicitly that it is a
complex function of the time, i.e., a phasor. The fact that a quantity is a space
vector is indicated by using boldface (heavy) type. In longhand notation a space
vector may be designated by a bar placed above or below the letter.

SEC. 8-16]

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

electric field (see Fig. 3-5).

path, which

is

337

Instead of using the relation that E dl =

for a closed
we should, in the time-varying

true for static fields due to charges,

case, use Maxwell's equation

from Faradays' law

o)

-*--J.f-*

is a flux density B normal to the rectangular path (half in each medium)

and
B changes with time, then f E dl is not zero if the path encloses a finite area. However, it is assumed that the dimension Ay approaches zero so that E
n and E, 2 are

If there

separated by only an infinitesimal distance.

Therefore the area of the rectangle

approaches zero, and the surface integral of dB/dt vanishes. Thus the work around
the path is given by E n Ax - E t2 Ax = 0, as before, and it follows that E = E

t2

holds for both static and time-varying situations.

Consider next the tangential components of the

field (see Fig. 6-14).
Instead
dl =
of using the relation
ds for steady
J
fields,
we
should,
in
the
time-varying
JJ
case, use Maxwell's equation from Ampere's law in its complete form,

JH-dl^j^J + ^yds
If there

is

a time-changing

(2)

normal to the rectangular path (half in each medium),

However, it is assumed that the dimension Ay

there will be a contribution due to D.

approaches zero, so that the surface integral of dD/dt vanishes.

current density

vanishes as

an

J may

also change with time.

Ay approaches

its

zero unless the conduction current

infinitesimally thin layer at the

linear density

However,

conductor surface.

In (2) the conductionsurface integral also

is

assumed to

exist in

Thus, for a sheet current of

K at the surface

Hn -H

=K

t2

(3)

as before, while in the absence of such a sheet

Hn = H

(4)

t2

as before.

Thus, the relations for the tangential

H field of Table

6-2 hold for both

and time-changing situations.

The formal approach in obtaining the continuity relations for the normal
components of D and B is the same under time-varying conditions as for static conditions, and the relations given in Tables 3-2 and 6-2 apply under both conditions.
static

fields.

Table 8-1 summarizes the boundary relations developed for electric and magnetic
These relations apply under all situations, except as noted.

338

ELECTROMAGNETICS

[CHAP. 8

Table 8-1

BOUNDARY RELATIONS FOR ELECTRIC AND MAGNETIC

FIELDS

Field

Condition

component

Boundary

Tangential

E,i

Tangential

Tangential

H =H

Tangential

fin

Ht2 =-K

(4)t

Current sheet at boundary

Tangential

=K

(5

Medium

Tangential

#,i

(6)

Medium

Normal

Di

D 2

(7)

Any two media

with charge at boundary

Normal

Di

D 2

(8)

Any two media

with no charge at boundary

Normal

Di

=p

(9)

Medium

relation

= E, 2

ti

ti

~-

= P

(1)

Any two media

(2)

Medium

(3)

Any two media

is

a perfect conductor (o 2

2 is a perfect conductor (o 2
with current sheet at surface

2 has infinite permeability

(no currents)

is

(/x 2

<)t

a>)

>)

a perfect conductor with charge

at surface

Normal

Bi

= B2

(10)

Any two media

Under static conditions it suffices for medium 2 to be a conductor (a 2 finite). However,

co (see Chap. 10).
for E, 2 to be zero under time-varying conditions requires that u 2

Note

K and

that although

the

are measured parallel to the boundary,

n X H,
Thus, in vector notation (5) is expressed by

components of

they are normal to each other.

unit vector normal to the boundary.

where n

K=

GENERAL FIELD RELATIONS

8-17

In Chap. 5

it

was shown

that the divergence of the curl of a vector function

is

zero.

Thus,

V (V x

F)

(1)

any vector function with no divergence must be the curl of some other
vector function. Thus, if V G = 0, we can write G = V x F, where F is some other
vector function. As an example, V B = 0, so that B can be expressed as the curl

As a

corollary,

of a vector potential (B

=Vx

A).

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

SEC. 8-17]

It

was also shown

in

339

Chap. 5 that the curl of the gradient of a scalar function/

Thus V x (\f) = 0. As a corollary, any vector function with no curl is

the gradient of some scalar function. Thus, if V x F = 0, we can write F = Vgr,
where g is a scalar function. As an example, the curl of the static electric field due to
electric charges is zero (V x E = 0).
It follows that a static electric field due to charges
is

zero.

may be expressed

as the gradient of a scalar function.

That

is,

E = \V,

where

is

the electric scalar potential.

According to Maxwell's equation derived from Faraday's law, we note, however,

that in time-changing situations the curl of the electric field is not zero but

is

equal to

Thus

the time rate of decrease of B.

VxE=-Since

An
(2)

V x E is

E= -

not zero, the relation

additional term

is

This

required.

is

(2)

not sufficient for time-varying

may be found

as follows.

Since

fields.

B=Vx

A,

becomes

8(\ x A)

(3)

at

V x (e +

from which

=0

(4)

Since the curl of the expression in parentheses in (4) equals zero,

E+
where/is a

scalar function.

function, a relation

varying situations.

is

^ = V/

satisfies

V,

so that from (5)

we have

static fields this

reduces to

may vary with

time,

If the

(6)

E = - W, as it should. In the general case, where the

E is given by both a scalar potential V and a vector potential

For
field

(6).

V is taken to be this scalar

the requirements for both static and time-

dA
E--VK--

A, as in

must be equal to

(5)

If the electric scalar potential

obtained that

Thus let/=

it

Thus we can write

the gradient of a scalar function.

time variation

is

harmonic,

(6)

E= -VV -jcoA

becomes
(7)

340

[CHAP. 8

ELECTROMAGNETICS

When the
and magnetic

vector potential

fields

and the

scalar potential

can be obtained under

static

are known, the electric

or time-varying situations from the

relations

5A
-W-

E=

(Vnr

(8)

dt

B = Vx A

and

dv

where

A=f

and
It is

(T)

An

assumed that the distance

3
[

-dv

(9)

(V)

(Wbm"

(10)

(ID

r in the expressions for

and

is

small

with a wavelength, so that propagation-time effects can be neglected.

the case, the propagation time must be considered and the

used for p and

8-18

J, as explained in

Chap.

compared

If this

is

not

more general retarded form

14.

COMPARISON OF ELECTRIC AND

MAGNETIC FIELD RELATIONS

In Table 5-2 a comparison

is

made of electric and magnetic

these apply in static or slowly time-varying situations.

field equations.

All of

Certain of the relations can

be extended so as apply under rapidly time-varying conditions. These relations are

It is to be noted that under static conditions the time derivatives
listed in Table 8-2.

Table

COMPARISON OF THE ELECTRIC AND MAGNETIC FIELD

RELATIONS FOR TIME-CHANGING SITUATIONS

8-2

Description of
equation

Curl equations
(point relations)

Closed path of

Electric field

Magnetic

dB
VXE=--

0D
VXH = J+-

in<tf

tegration

Derivation of
fields

scalar

from
and vector

= ^E.rfl=-j^.rfs F=jfH.rfl = [(j+f).*

SA
E=-VK--

potentialsf

Kand A

field

are as indicated in (8-17-10) and (8-17-11).

B = VXA

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

PROBLEMS

341

are zero, and these relations reduce to the corresponding special cases given in Table
5-2.
These static relations (Table 5-2) are also applicable in time-changing situations

provided the variations are slow enough to permit the time derivatives to be neglected.
In more rapidly time-varying situations, where the time derivatives cannot be neglected,
the expressions of Table 8-2 must be employed.

PROBLEMS
*

8-1

(a)

3-turn loop with 0.5

appearing at the terminals of the loop ?

is

(6) If the

emf at

mT

_1

field

100

emf
mV, what

mm on a side to

develop a

If the flux density changes 5

normal to the plane of the loop.

area situated in air has a uniform magnetic

s

the loop terminals

what
is

is

the

the rate of change of the magnetic field ?

8-2 How many turns

maximum emf of 10 mV rms

B = 60
*

the loop rotates 30 r s~

uniform magnetic

through the
the results

8-4

field

with a velocity v

field

B=

on a graph of emf
Is the direction

100

8-5

Take

vs. the

distance

'

for

Find the emf induced

in the

in Fig. P8-5.

support point

of the current / in the closed loop at the right of Fig. P8-4

when

the switch

is

(a) closed

and

ment or half amplitude.

determine the peak

The

velocity of

any point on the pendulum

at a distance r

(m s" 2 )]. Using

terminals if d = 100 mm.

emf appearing

at the

this

r .

40 mm
*.

_-

/
Loop

h"-*-^

FIGURE
Moving

P8-3

loop.

from

value for the period,

opened?

= a)d(r/R) cos <at, where d is the maximum horizontal displaceIf the length R of the pendulum is 4 m, its period T at the earth's

100 mm-

80 mm-

(b)

given by v

is

surface will be about 4 s [T(s)= 2tt V/?(m)/9.8

(x

loop and plot

20 < x < 120 mm.

wire pendulum with brush swings normal to a uniform magnetic field of 250

mT, as shown

magnetic field?

mT extends over a square area 100 mm on a

A rectangular wire loop 40 by 80 mm is moved

mm s -

clockwise (cw) or counterclockwise (ccw)

its

in the earth's

100

with zero field outside.

side, as in Fig. P8-3,

ftT.

8-3

are required for a square loop 100

if

FIGURE

FIGURE

P8-4

Pendulum.

Stationary loops.

* Answers to starred problems are given

in

Appendix C.

P8-5

342

[CHAP. 8

ELECTROMAGNETICS

Rim

Pole piece

Wire

FIGURE
Moving

FIGURE

P8-7

Dc

Moving spoke.

Find the emf induced

8-6
netic field

FIGURE

P8-6

wire.

P8-9

machine.

a straight wire moving perpendicular to a uniform mag-

in

The magnetic

with velocity v as in Fig. P8-6.

field is

confined to the radius

of the pole pieces of a magnet.

A conducting wheel of rim and

8-7

single spoke rotates perpendicular to a uniform

The magnetic field is confined to a radius R of the pole pieces

of a magnet. An external circuit makes contact with the axle and rim through brushes,
(b) If a current / is
(a) If the wheel is rotated JVrs"', find the emf induced in the circuit,
(c) If the current flows as shown, is
sent through the circuit, find the torque on the wheel,
the torque cw or ccw ?
8-8 If the one-spoked wheel of Prob. 8-7 is replaced by a solid metal disk, what
difference (if any) occurs in both generator and motor cases ?
8-9 If a rectangular loop rotating in a uniform magnetic field B is equipped with a two-

magnetic

field

B (Fig. P8-7).

segment commutator, as in Fig. P8-9, the arrangement can function as either a dc generator
or a motor,

(a) If the

loop rotates

N r s~

find the average

current / flows in the loop, find the average torque,

the rotation direction (cw or ccw).

8-10

find the

(a)

area

loop develops an emf of 10

If the

rms magnetic

8-11

field

is

dc voltage generated,

(b) If

the current flows as shown, find

Assume commutator gap

A 5-turn wire loop of 0.5 m

radio wave.

(c) If

is

negligible.

situated in air in the presence of

2-MHz

mV rms when oriented for maximum response,

H of the wave.

thin copper disk 300

to a constant, uniform magnetic field

mm

in diameter

B = 600 mT.

is

situated with

30 r

If the disk rotates

its

plane normal
',

find the

emf

One brush contacts

developed at the terminals connected to brushes as shown in Fig. P8-1 1
This arrangement is called
the periphery of the disk and the other contacts the axle or shaft.
.

a Faraday disk generator.

where

B 600 mT
8-12

(a)

= 1 u m"

(6) If the
u>

= 2w x

magnetic field varies with time, as given by B

5 rad s"

*,

find the

emf developed

rectangular nonconducting trough 3

brackish water 250

and

mm

deep with a velocity v

If conducting plates 500

=3 m

=B

sin cat,

at the terminals.

wide carries a uniform flow of

-1
.

The water has a conductivity

mm wide by 250 mm deep are placed at the sides of

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

343

Shaft

Brushes

n>'

250 mm

Loop
Long wire

*500mrrH

FIGURE

FIGURE

P8-11

the trough and a magnetic field

/ flowing in a load resistor R,

(A)

FIGURE

P8-12

B = T is impressed, as shown in Fig. P8-12, find the current

= 10 Q. Neglect resistance of plates and connecting wires.
1

Find the power delivered to the load,

delivered to a load

if

R,

is

P8-17

Wire and loop.

Water generator.

Faraday disk generator.

variable ?

(c) Is this

the

maximum power which can be

(d) Discuss the practical aspects of using a similar system

power directly from the flow of water in a large river using the earth's
magnetic field, (e) Could the system be used to measure the velocity of flow of a river?
Can
If so, explain how such variables as depth and conductivity could be accommodated. (/)
to generate electric

the arrangement of Fig. P8-12 be modified so that

8-13
at 5 r s"

1
.

maximum rms emf induced

of 1 5

it

can be used to

pump a conducting liquid ?

A short bar magnet of 6 A m 2 magnetic moment rotates around its center point
A 50-turn loop of 1 m 2 area is located 8 m from the bar magnet. Find the
in the

loop by the magnet.

8-14

A toroid of 500 turns has a mean radius of 250 mm and a radius for the winding

mm.

Find the average inductance for

(a)

an

air

core and (6) an iron core with average

relative incremental permeability of 750.

the mutual inductance of an ideal transformer if a 60-Hz current of

rms at the secondary terminals?
rms applied to the primary induces 6
8-16 (a) Find the mutual inductance per unit length for two long straight parallel

8-15

What

is

one wire is /, find the emf induced per unit

length in the other wire if it is at a separation s and approaching at a velocity v.
8-17 Find the mutual inductance between a long straight wire and a rectangular wire

wires with separation

s.

(b) If the current in

loop as in Fig. P8-17.

8-18
r,

and

Find the mutual inductance between two concentric circular wire loops of radius
when the radius of the inner loop is much smaller than the outer loop and (A)

r 2 (a)

without this restriction.

8-19

and

(2,3)

m.

has corners at (x,y) values of (0,0), (2,0), (0,3),

normal to the loop varies as a function of position, as given

fixed square 8-turn wire loop

If

a magnetic

field

344

ELECTROMAGNETICS

[CHAP. 8

^Tracks^

t
'

'

'

'

'

'

/
Plan view of wiring
between tracks

Iron slab

(a)

(b)

FIGURE P8-20
Linear motor for track-guided vehicle.
by

B = 5 sin (ttx/2) cos (ny/3), find the rms emf induced in the loop if B also varies harmoni-

cally with time at 500 Hz.

8-20 (a) Design a linear dc motor for propelling and braking a track-guided
vehicle.
The wheels are not to be used for traction or braking except in emergency. The
vehicle
contains no power source other than that for magnetic field generation,
and no power is
supplied to the vehicle from any external source.

from the

vehicle

Currents in the track region are controlled

Figure P8-20 shows a suggested scheme, (b) What

by radio command.

current and magnetic field are required to accelerate a 50

hr" 1 in 60 s?

The

track width

is 1.5

m.

Neglect

x 10 3 kg

friction.

vehicle

from

to 100

km

[For other arrangements see

E. R. Laithwaite, Linear Motion Electrical Machines, Proc. IEEE,

58:531-542 (April 1970).]
8-21
copper conductor of diameter d
10
is formed into a
radius

.8

A
= 215 mm

and connected

The medium

mA m" \

is air.

and

if

If the loop is

circular loop of

the equivalent (series) resistance of the loop-capacitor circuit

is

Q, find

Springs^
Koprmgb.

500 mm-

FIGURE

P8-21

Radio loop.

The

= 100 mm) and spaced a distance j = 5mm apart.

placed in the field of a 40-MHz radio wave with H =

capacitor plates are square (side length

mm

to a parallel-plate capacitor, as in Fig. P8-21.

FIGURE

Bar

P8-22

Floating bar.

Weight

TIME-CHANGING ELECTRIC AND MAGNETIC FIELDS

PROBLEMS

345

rms displacement current density in the capacitor and (b) the rms voltage across the
(c) Repeat (a) and (b) if the frequency of the wave is adjusted to the resonant

(a) the

capacitor,

frequency of the

8-22

Assume

circuit.

only about 0.0002 O, but

is

its

Note: The dc resistance of the loop

unchanged.

is

radiation resistance (see Chap. 14)

straight conducting bar with weight attached is

is

about 3

Ci.

suspended by wire springs

in a uniform magnetic field B as in Fig. P8-22.

The length of the bar is 500 mm. Find the
current / (magnitude and direction) required to " float " or balance the bar and weight if

B=2T
*

and the mass of the bar and weight is 5 kg.

Find the displacement-current density of a magnetic

8-23

H = Ho
y

sin(co?

fix),

H = xH

(b)

sin

2x

A parallel-plate air capacitor has a

8-24

fiy) + tH

sin(cut
1

sin

field in air

2x

cos(<o/

,000-fl resistor connected

given by (a)

j8v).

between the centers

The plates are 100 mm square and separated by 1 mm; If 10 V rms is applied
to the capacitor, find (a) rms displacement current through the capacitor, (b) rms conduction
current, (c) total current, and (d) power dissipated in resistor at each of the following frequenof the plates.

cies:

(1)

Hz,

the quantities
fringing.

(2)

kHz,

(a), (b), (c),

(3)

and

MHz, and

(d)

GHz.

(4) 1

and four rows

Make

a table with four columns for

for the four different frequencies.

Neglect

Discuss the significance of the results.

Given that the average power dissipated per unit volume of a medium is
m~ 3 where 8 is the time-phase angle between E and D, show that the
average power dissipated in a parallel-plate capacitor is VI cos 6 (W), where V =
8-25

E(dDjdt) cos 6
total

voltage across capacitor

and / are rms

*
er

and

= current

through capacitor.

Neglect fringing.

E, D, V,

values.

8-26 A parallel-plate capacitor is filled with a dielectric of 0.003 power factor and
= 10. The plates are 250 mm square, and the distance between is 10 mm. If 500 V rms at
MHz is applied to the capacitor, find the power dissipated as heat.
8-27 A coil of 60 turns and 1,000 mm 2 area has its plane normal to a magnetic field.

is measured to be 1 mC when the coil is flipped 180,

The coil and attached measuring circuit have a resistance of 100 fi.
8-28 The average flux density inside the electron orbit of a betatron is given by B =
60 mT, where w = 2irx60 rad s" l
If the radius of the electron orbit is 120 mm, find (a)

If the

charge flowing through the coil

find B.

velocity of electrons at

= ms

electron per revolution, (c)

volts) after

also

ms.

Neglect

if

at

= 0, (b) energy per unit charge imparted to the

number of revolutions

in

ms, and

(f)

electron energy (in electron

relativistic effects.

8-29 Is it true that any machine which functions satisfactorily as a generator will
work as a motor when a voltage is applied which is of the same type as that generated?
8-30 A bar magnet is attached perpendicular to an axle shaft which extends along the

centerline of a cylindrical container of conducting fluid, as in Fig. P8-30.

(a) If

/ is passed axially through the conducting fluid, will the device operate as a motor ?

a current
(b) If the

magnet is rotated, will the device operate as a generator ?

8-31 A Faraday disk generator with superconducting solenoid is shown in cross
section in Fig. P8-31.
The disk has a radius of 100 mm, and it rotates at 3,600 r min -1
shaft with

B3

T, find (a) the required solenoid sheet-current density and (b) the
(See Machine Design, Nov. 11, 1971, p. 18.)
If

emf

generated.

346

ELECTROMAGNETICS

[CHAP. 8

^Shaft

Disk electrode

\
Solenoid
o o o o o o o

Bar magnet

ShafU

oooooooo

HI

r v

o o o o o o o

Disk

o o o o o o o

Disk electrode

FIGURE
Magnet

FIGURE

P8-30

8-32

(a)

The

P8-31

Faraday disk

in conducting fluid.

permittivity of ice is

a function of both temperature and frequency.

In E. R. Pounder, "Physics of Ice," p. 129, Pergamon Press,

relative permittivity is given

by

er

= ei je"

+ ^'f
2
1 + aY

j1

where

in superconducting solenoid.

New

and the Drude-Debye

(ei

and

'

York, 1965, the complex

relations

by

e 2 )af
+ aY 2

T = temperature, C
/= frequency, Hz

and where

for pure water ice e t

applicable from

to 10

= 75,

GHz and from

= 3, a = 1.2 x 10" 4e"- ,r(s). The relations are

The effect of impurities is not included,
to 70C.

e2

but they would probably increase the magnitude of the imaginary part

complex

relative permittivity of

pure water

ice at

0C

at 0,

10,

1,

(e?)-

Calculate the

and 100 Hz;

1, 10,

and

and 100 MHz. (b) Calculate the equivalent conductivity a' as a function
of frequency (0 to 100 MHz). Take a = 0. (c) Make a log-log graph of ej e? and a'
ef and a' vs. log frequency.
as a function of frequency; also a linear graph of
8-33 Setting the constants Nq 2 /e m and dim equal to unity, calculate the variation of
Display results graphically
ei and " from (8-15-6) and (8-15-7) as a function of to for u> = 1
100 kHz; and

1,

10,

e', ,

in

both log-log and linear-log form, as in Prob. 8-32.

8-34 A
= 65 y'15.

to

homogeneous 200-cc

bake the potato?

200-g

Idaho potato has a

relative

permittivity

= 30 kV m" at 2.45 GHz (alongside potato), how long will take

Take = 0. [See N. E. Bengtsson and P. O. Risman, Dielectric

If (rms)

it

'

o-

Foods at 3 GHz, /. Microwave Power, 6(2): 107-123 (1972).]

Metal detectors operate on several principles. In one design a coil and capacitor
form the resonant circuit of an oscillator at say 100 kHz. The introduction of a metal
object into the field of the coil changes, or detunes, the oscillator frequency, and this change
Properties of

8-35

is

noted.

Another design operates on the null

principle.

Thus, two

flat exciter coils

are

TME-CHANOINO ELECTRIC AND MAGNETIC

PROBLEMS

placed side by side in the same plane.

modulated

at

~500

(balanced out).
is

Both are connected to the

Hz), but the coils are so connected that the

field,

a voltage

is

induced in the search

Design a device for detecting automobiles for

(b)

coil,

When

a metal object

Its rectified
is

is

zero

output

introduced into

producing an audible 500-Hz tone,

traffic-light

347

(~ 100 kHz

between them

A third detector, or search coil, is situated at this point.

connected to an audio amplifier and transducer.

the

oscillator

field

FIELDS

(a)

control at a highway intersection.

Design a device for locating buried pipes or coins, (c) Discuss the pros and cons of
above applications, (rf) Can these detectors

frequency-shift vs. null-principle designs for the

be used to locate

dielectric bodies

or anomalies?

THE RELATION BETWEEN FIELD AND

CIRCUIT THEORY; MAXWELL'S EQUATIONS

9-1

INTRODUCTION

In circuit theory

we deal with circuit

current / through them.

and J) and

elements, the voltage

we

In field theory

In low-frequency-circuit theory

of one quantity,
ance.

its

resistance R.

Thus the voltage

it is

Its length, area,

difference

and

cross-sectional area

is

and shape are of secondary importis, from Ohm's law,

between the ends of the rod

the field-theory point of view

a point in the rod.

From Ohm's law

E=^
where J

in

(1)

the current through the rod.

From
at

convenient to describe the rod in terms

V = IR
where /

H,

their values as a function of position.

Consider, for instance, a short rod of length

Fig. 9-1.

V across them, and the total

deal with the field vectors (E, D, B,

= conduction current density,

= conductivity, Urn -1

we

consider the value of the electric field

at a point (see Sec. 4-9)

(Vm"

Am -2

(2)

THE RELATION BETWEEN FIELD AND CIRCUIT THEORY MAXWELL'S EQUATIONS

SEC. 9-1]

349

FIGURE 9-1
Conducting rod.

Now,

we

integrating (2) over the length of the rod,

between the ends.

That

obtain the voltage difference

V = JE-dl = j--dl
For a uniform rod with uniform current density

where J A

=I=

Jl

aA

current through rod,

(4)

this

(3)

becomes

(V)

(4)

II a A = R = resistance of rod, fi
A = cross-sectional area of rod,

Thus, from

m2

we have

V = IR
Starting with field theory,
Historically, this

we have

and other

(5)

arrived at the circuit relation

known

as

were postulated and

circuit relations

Ohm's

It

law.

verified first.

Then, as a generalization, they were extended to apply to the more general

situation.

is,

field

follows, therefore, that circuit relations are simply special cases of field

may be deduced from them. Although field relations are more general,
much simpler to use circuit equations wherever they are applicable.

equations and
it is

usually

Equation

(1) is

a pure circuit relation.

relation.

Many

mixture.

Such mixed

On

the other hand, (2)

is

a pure

field

equations are not. purely one or the other but are a combination or

connection between

relations are necessary, for example, in order to provide a

field

and

circuit theory.

Two

important equations that provide

such connecting links are

and

Equation

V = JE-dl

(V)

(6)

= ju-dl

(A)

(7)

(6) relates

V (a circuit

two points to the line integral of E

(7), which is Ampere's law, relates
field quantity) around a closed path.

quantity) between

(a field quantity) between those points.

Likewise

/ (a circuit quantity) to the line integral of

H (a

350

[CHAP. 9

ELECTROMAGNETICS

FIGURE 9-2
Coaxial transmission line with Transverse ElectroMagnetic (TEM) mode.
E-lines

are radial and solid.

H-lines

are circles and dashed.

APPLICATIONS OF CIRCUIT

9-2

AND FIELD THEORY

field relations are applicable in general, circuit relations are usually

While

convenient wherever

V and

more

/ have a simple, well-defined significance.

Thus, in determining the capacitance of a capacitor of irregular shape with the

map (see Fig. 3-19) we are in effect directing our attention to the

aid of a graphical field

field

and

its

may

at

terminal circuit element of capacitance

size

However, once we have

value as a function of position in the capacitor.

determined the capacitance, we

and shape of the capacitor and the

low frequencies consider

C with
field

it

a voltage difference

configuration within

it

as simply a twoV.

The

physical

are then relegated

to positions of secondary importance.

As another

illustration let us consider the coaxial transmission line

conditions, one where

V and / are useful

shown

in

quantities

two
and one where they are not. The coaxial line has an inner conductor of radius a
and an outer conductor of inside radius b. With a steady potential difference between
cross section in Fig. 9-2 under

the conductors the electric field lines are radial, as shown.

the magnetic field lines
difference

are circles, as indicated.

between the inner and outer conductors

V=

"

dt

In (1)

= JH-dl =

a current /

by

is

flowing,

(9-1-6) the potential

is

(V)

(1)

Likewise from (9-1-7) the current / in the inner conductor

If

Now

*Hrd0

(A)

is

(2)

J"

V is independent of the path between the conductors, while in (2) /is independent
Hence, V and / have a simple, definite
r provided it is between a and b.

of the radius

significance in this case

and are useful

quantities.

THE RELATION BETWEEN FIELD AND CIRCUIT THEORY; MAXWELL'S EQUATIONS

SEC. 9-2]

351

V
v$v

>

\j<

'l<

><j>/"

"

Longitudinal section
i

'

(b)

FIGURE 9-3
Coaxial transmission line with higher-order Transverse Magnetic

The

field

(TEM) field

shown

configuration

This type of

ponent in the axial direction).

mode

in Fig. 9-2 is called

because the electric and magnetic

and

possible under steady conditions

wavelength
lengths)
possible.

is

field

axial direction.

configurations

field

Although coaxial

in Fig. 9-3 to exist.

is

known

some

high for the

show the

field configuration.

Magnetic (TM) mode because the magnetic field

For

field has a longitudinal component.

ductors as obtained by (1)

depends on the radius

may become

r at

which

is

this

field

mode

or configuration shown

is

This

field is called

axial)

a Transverse

entirely transverse, while the electric

mode, the voltage

between the con-

negligible, while the current / obtained

integrated.

simple significance and are not as useful as the

Hence

V and / no

field quantities

concept occurred here when

breakdown of the circuit

became comparable with the wavelength.
t

modes become
components in the

as higher-order

seldom used under such conditions,

lines are

sufficiently

where the

At higher frequencies (shorter wave-

Both a cross (or transverse) section and a longitudinal (or

section are needed to

(2)

the only configuratipn or field

for time-varying situations

are characterized by having

suppose that the frequency

a Transverse ElectroMagnetic

are entirely transverse (no com-

field is

of the order of 4b or greater.f

more complex
These modes

(TM) mode.

by

longer have a

The

themselves.

the transverse dimensions

is, the frequency is so high that a disturbance traveling with the velocity of
can travel only about a distance equal to the diameter (2b) in one-half period.
In free space a wave has a wavelength A in meters that is related to the frequency /in
_1
For
c//, where c= 300
s
hertz by the velocity of light c as follows: A
a further discussion of field modes in lines and guides see Chap. 13.

That
light

Mm

352

ELECTROMAGNETICS

[CHAP. 9

B
Inductor

A=LI

Resistor

T
1
i

Capacitor

tfjr;

c=f
.

Generator

cyr

FIGURE 9-4
Series circuit with resistance, inductance, capacitance
illustrate

and ac generator used to

equivalence of field and circuit theory.

THE SERIES CIRCUIT; COMPARISON

OF FIELD AND CIRCUIT THEORY

9-3

From

we may

(4-10-1)

related to emfs

and a

express the total electric field

field

Etotal

Etotal

as the

induced by charges and currents.

= Ee + E

or

sum of a

Ee

field

Thus,

Etotal E

(1)

From (4-9-4)
"total

and from

(2)

(8-17-8)

E=-W-^

(3)

dt

Consider

now

a series circuit of resistance, inductance, and capacitance connected to

a generator as shown in Fig.

confined to the resistor,

and

all

emf to

all

9-4.

It is

assumed that

inductance to the inductor,

the generator.

all

all

Let us apply the above

resistance in the circuit

is

capacitance to the capacitor,

field relations to this circuit.

THE RELATION BETWEEN FIELD AND CIRCUIT THEORY; MAXWELL'S EQUATIONS

SEC. 9-3]

First substitute (2)

and

Then

(3) in (1).

integrate all terms completely

around the

to convert field into circuit quantities

circuit in

a clockwise direction.

left side yields

two terms on the

emf 17 of the generator. Also

VK = -E, we have

the

= I/a and E =

D/e.

The
d I

where L

Af

A=
way

In this

(5)

d\

last

term

may

aa

But Ae/d

=R

first

(5)

dI I

be reexpressed asf
j,

dI

inductance of circuit

dv =

vector potential

becomes

Dd
dl
I
+ + L
aa
.

11

integrating the

dtJ

j.

QT
<U

Now

(4)

and noting that

right

a
But J

we

This gives

JEe -dl = j>--dl + j\V-dl + j^-dl

Integrating the

353

and

= C and

D=

Q/A

the charge

so

dt

(7)
v '

we have

Q = \ldt;

so (8) takes the form

V - + Ij/A + L|

(9)

Thus, beginning with the

we have

field relations (1), (2), and (3) as applied to a series circuit,

arrived at the familiar circuit relation (9) for a series circuit.

The transformation of (6) may be also made with the aid of Stokes' theorem,
that

B = V X A and

A = LI, as follows:

s* A J. <****-/. -*--i.f
where

L = inductance
/

of loop

= Current through loop

recalling

354

[CHAP. 9

ELECTROMAGNETICS

For harmonic variation of the current with respect to time

(/

=I

Jat

we obtain

U = IR + ^-+ja>LI

(10)

jcoC

<V

01

In deriving

same

(1 1)

the assumption

at all parts of the circuit.

the circuit instantaneously.

length, this

is

= IR + jl(a>L-)

was made

(11)

that at any instant the current

This implies that a disturbance

If the circuit length is small

a satisfactory assumption.

However,

if

is

is

the

propagated around

compared with the wave-

the circuit length

is

appreciable

compared with the wavelength (say at least A/8), the variation in current and phase
around the circuit may become significant. Under these circumstances the simple
It is also to be noted that
circuit concepts tend to become inadequate and inaccurate.
the above circuit treatment ignores the phenomenon of radiation, which is so important
at

high frequencies (see Chap.

14).

There are certain exceptions to the above statement that

inadequate

when

the circuit length

is

circuit

concepts become

comparable with the wavelength.

circuit concepts are successfully applied to the long transmission line.

For example,
In this case,

the distributed inductance and capacitance are represented by suitable lumped elements
(see

Chap.

13).

Although the length of the

line

can be

many

wavelengths,

nificant that even in this extension of circuit theory the treatment

lines

9-4

is

is

it

sig-

adequate only for

with transverse dimensions that are very small compared with the wavelength.

MAXWELL'S EQUATIONS AS GENERALIZATIONS

OF CIRCUIT EQUATIONS!

In the remainder of this chapter a

number of

relations developed in the preceding

chapters are brought together and considered as a group.

as Maxwell's equations, consist of four expressions

These

relations,

known

one derived from Ampere's law,

one from Faraday's law, and two derived from Gauss' law. These equations are of
profound importance and? together with boundary, continuity, and other auxiliary
relations,

form the basic tools

for the analysis of

most electromagnetic problems.

James Clerk Maxwell, "A Treatise on Electricity and Magnetism," 2

University Press, London, 1873; 3rd ed., 1904.

vols.,

Oxford

THE RELATION BETWEEN FIELD AND CIRCUIT THEORY; MAXWELL'S EQUATIONS

sec. 9-4]

In Chap. 5 Ampere's law relating the line integral of

to the current / enclosed

was given

355

around a closed path

as

JH-

d\

(1)

Replacing the current / by the surface integral of the conduction-current density J

over an area bounded by the path of integration of H,

j
In Chap. 8 this relation was

d\

we have the more general relation

J ds

(2)

made even more general by adding a displacement-current

density to the conduction-current density so that (2) becomes

f
This relation
given in

its

is

d,=
/i

(3) is

fH

j+

(3)

Maxwell's equation as derived from Ampere's law.

integral form, the line integral of

ing the surface

Hence,

called

In circuit parlance a closed path or loop

j.

a mesh

relation.

In (3)

it is

H being taken over a closed path bound-

Applying Stokes' theorem to

is

(3),

often called a mesh.

we obtain

the corre-

sponding point relation

VxH= J+ dD
8t
Equation

(4) is

(4)

a differential relation and relates the

field quantities at

a point.

It is

the differential form of Maxwell's equation as derived from Ampere's law.

In Chap. 8 Faraday's law relating the emf IT induced in a circuit to the time rate
of decrease of the total magnetic flux linking the circuit was given as

(5)

dt

Replacing the flux linkage

circuit,

we have

the

by the surface

0
Replacing 1J in

(6)

integral of

over the area bounded by the

more general equation

by the

-s/.-

line integral

of

ds

(6)

E around the

circuit,

we have

the

still

more

general relation (for stationary circuits) that

d\

= -

8B

J, dt

ds

(7)

356

ELECTROMAGNETICS

This

field relation is

[CHAP. 9

a generalization of Faraday's

circuit

called Maxwell's equation as derived from Faraday's law.

form,

from

(7)

mesh

i.e., it is

equation.

law

It is

The corresponding point

(5).

Equation

(7) is

given in (7) in integral

relation can be obtained

by an application of Stokes' theorem, yielding

dB

V x E=

(8)
~di

Equation

(8) is

differential relation

and

relates the field quantities at a point.

It

is

In Chap. 2 Gauss' law relating the surface integral of the electric flux density

the differential form of Maxwell's equation as derived from Faraday's law.

to the charge

enclosed was given as

ds

=Q

(9)

Replacing Q in (9) by the volume integral of the charge density p throughout the
volume enclosed by the surface s, we can write (9) in a more general form as

ds

(J>

This

field relation is

a generalization of Gauss' law and

field equation as derivedfrom Gauss' law.

to a finite

volume

v.

p dv

In (10)

Applying (10) to an

it

is

called

Maxwell's

electric

appears in integral form and applies

infinitesimal

corresponding differential relation that relates the

(10)

volume, we can obtain the

field quantities at

a point as given

by

V-D = p
Equation (11)

is

Maxwell's

electric field

(11)

equation as derived from Gauss' law in

differential form.

For magnetic fields the surface integral of B over a closed surface

Thus the magnetic counterpart of Gauss' electric field relation (9) is

ds

s yields zero.

(12)

THE RELATION BETWEEN FIELD AND CIRCUIT THEORY; MAXWELL'S EQUATIONS

sec. 9-4]

The corresponding

357

differential or point relation is

V-B =
Equations (12) and (13)

may be

(13)

referred to as Maxwell's magnetic field equations as

derived from Gauss' law, (12) being the integral and (13) the differential form.
The development of Maxwell's equations as generalizations of circuit relations
involves both inductive

tion"

is

and physical reasoning.

Maxwell's equations are

rigorous.

based on them have been found

in

It is

justified

not implied that the " deriva-

by the

innumerable cases to be

fact that conclusions

in excellent

agreement with

experiment, in the same way that the earlier circuit relations are justified within their
more restricted domain by the excellent agreement of conclusions based on them with

perhaps worth recalling that Maxwell's equations were not generally

accepted for many years after they were postulated (1873). His curl equations
(involving V x E and V x H) implied that time-varying electric and magnetic fields

experiment.

It is

empty space were interdependent, a changing electric field being able to generate a
magnetic field, and vice versa. The inference from this is that a time-changing
electromagnetic field would propagate energy through empty space with the velocity
of light (see Chap. 10) and, further, that light is electromagnetic in nature. Radio

in

was 5 years (1888) before Hertz demonstrated

that electromagnetic (or radio) waves were possible, as predicted by Maxwell.
There is no guarantee that Maxwell's equations are exact. However, insofar
waves were unknown

at the time,

and

it

as the precision of experimental measurements allow, they appear to be, and therefore

we may regard them

as exact.

Along with Maxwell's equations

importance
tioned

in dealing

Ohm's law

certain other fundamental relations are of

with electromagnetic problems.

Among

these

may

be men-

at a point (4-9-4)

(14)

ffE

the continuity relation (4-13-3)

8P

(J 5)

the force relations

F = <7E
(16)

dF =

(I

x B)

dl

358

ELECTROMAGNETICS

and the

[CHAP. 9

constitutive relations between

and

D and between B and H as given by

D = eE = e E + P

(17)

B = /xH = fi (H + M)

9-5

(18)

MAXWELL'S EQUATIONS IN FREE SPACE

In the preceding section, Maxwell's equations are stated in their general form.

For
J and the charge density p
In integral form the equations are

the special case of free space, where the current density

are zero, the equations reduce to a simpler form.

dl

ds

(1)

dt

dB
(2)

D-ds =

(3)

B-ds =

(4)

In differential form the equations are

(5)
~8t

Vx

9-6

If

E=

- dB

(6)

It

VD =

(7)

V-B =

(8)

MAXWELL'S EQUATIONS FOR

HARMONICALLY VARYING FIELDS

we assume

that the fields vary harmonically with time, Maxwell's equations can be

expressed in another special form.

Thus,

if

D varies

00^

with time as given by

(1)

THE RELATION BETWEEN FIELD AND CIRCUIT THEORY; MAXWELL'S EQUATIONS

SEC. 9-7]

^^jwDo^'^jtoD

then

359

(2)

at

When

the

same assumption

is

for B, Maxwell's equations in integral

made

form

reduce to

JH-d\ =

j,

d\

((T+jwe)j E-ds

-join

ds

(4)

D-ds = j pdv

ds

(3)

(5)

(6)

H = {a +7'we)E

(7)

In differential form they are

V x

V x E = -jconH

(8)

V D=p

(9)

VB =

(10)

TABLES OF MAXWELL'S EQUATIONS

9-7

Maxwell's
Maxwell's equations are summarized in Tables 9- 1 and 9-2. Table 9- 1 gives
are
equations
The
form.
differential
9-2
in
Table
equations in integral form and
case
steady
case,
harmonic-variation
case,
stated for the general case, free-space
(static fields with
fields but with steady conduction currents), and static case
(static

no

currents).

field quantities

In Table 9-1 the equivalence

and

is

also indicated between the various

emf

the electric potential V, the

V, the magnetic potential U,

conduction current
the mmf F, the total current 7total , the displacement current 7disp , the
that Maxwell's
noted
should
be
It
flux
magnetic
and
the
ty m
the electric flux $,
.

/cond

or bodies at
equations as tabulated here apply specifically to stationary systems

rest.

360

J3

II

3
<

II

CO

CO

S3

-G->

*-e-.
II

ll

9-

E
-s-

<x

II

U,

Q.

II

Q.

II

?
II

8
a

*-e4

"-e-^

-e-^

"-e4

-e^

o
s

-3-

8
*

II

ll

8
Q

x
3

CA

*-e-^

-3-

a.

II

II

s
H

II

II

*n
t>

II

II

II

II

II

II

-3-

-3-

-3-

-3-

-3.

"?

8
K

SI*

OS |<o l<

"1

'
1
1

>>

II

tfl

'

II

ll

II

II

>

<-e-

^
w

-e->

ll

II

II

B
1)

ll

II

13

KT
II
II

5?
"*3

1
w

.^.^

SI*
+

l-s

II

>

ii

*>

-J
*

"1

3,
11

II

II

II

"3

8.

<

<

t*-T

II

,^>

Q|<5
CO I"3
.

e
E

to

3
e

*--.
n

II

-e-.
II

II

t.

WW/
|-T3

X
<*,

k.

II

h.

73

,o

=//

S/

3/ s
/

CO

.sjS

c '5
E !i
3 >

1
55

00

THE RELATION BETWEEN FIELD AND CIRCUIT THEORY; MAXWELL'S EQUATIONS

MAXWELL'S EQUATIONS IN DIFFERENTIAL FORM

Table 9-2

Dimen^^^^sions

^^^
Case

361

From Faraday

Electric current

Electric potential

area

Magnetic

Electric flux

SB
VxE=--

V.D =

VxE=--

V-D =

V-B =

V X H = (ct +ytoe)E

VXE=

V.D = p

VB =

VxH = J
VXH =

VxE =
VXE =

VD = p

V-B =

VD = p

V-B =

0D
Free space

flux

volume

volume

area

dD
VXH = J + -

General

From Gauss

From Ampere

VB =

/>

Harmonic
variation

Steady
Static

ja>ij.H

PROBLEMS
9-1

based on

with Ampere's law, derive Maxwell's equation in integral form

(a) Starting
this law.

(b)

Obtain the corresponding

differential or point relation

by applying

Stokes' theorem.

9-2

on

(a) Starting with

this law.

(6)

Faraday's law, derive Maxwell's equation in integral form based

Obtain the corresponding

differential or point relation

by applying Stokes'

theorem.
9-3
integral

(a) Starting

form based on

9-4
in integral

(a) Starting

with Gauss' law for electric

this law.

(6)

fields,

with Gauss' law applied to magnetic

form based on

derive Maxwell's equation in

Obtain the corresponding

this law.

(b)

differential or point relation.

fields,

derive Maxwell's equation

Obtain the corresponding

differential or point

relation.

9-5

Why are

9-6

(a) State

Maxwell's equations not completely symmetrical ?

Maxwell's equations in their general integral form,

(b)

Derive them

for harmonically varying fields.

9-7

(a) State

Maxwell's equations in their general

differential form,

(b)

Derive them

for harmonically varying fields.

9-8

A long cylindrical conductor of radius R and a = oo carries a current I = I sin cat.

radius r (for r < R and r>R) find (a) the conduction-current density

As a function of

J(r), (b) the displacement-current density JAf),

and

(c)

* Answers to starred problems are given

the magnetic flux density B(r).

in

Appendix C.

362

[CHAP. 9

ELECTROMAGNETICS

F^

+
J^ T

dl'

fr
dl

FIGURE P9-14
Two-conductor transmission

9-9

of radius

As

a function of radius r (for r

Ja {r) and (6)

The magnetic field between two pole pieces of radius R varies as B = B [1
2
radius
function
of
as
a
density
displacement-current
the
Assume R ^ A. Find
i(r) ] sin to/.
the displacement-current density

line

and separation d has a voltage

< R) find (a)
the magnetic field H(r). Take d< R.

A parallel-plate capacitor with plates

V = V sin wt.

applied at the center as given by

'

2 conductor
transmission

line.

9-10

r (for r

< R).
A

9-/7

long solenoid of radius

carries a sheet-current density 7C

Find the displacement-current density J as a function of radius

9-12

Show

/-

(for r

=K

sin <r.

< R).

that

8A
'

J 8/

9-75

Show

that the expression for the low-frequency inductance

= (A/7)

d\

reduces for a conducting circuit to Neumann's low-frequency inductance formula

d
s.
di
4tt

9-14
separation

A
D and

two-conductor transmission line of length

conductor radius

to Fig. P9-14, apply

a.

The conductors

has a conductor (center-to-center)

are thin-walled tubes.

Neumann's low-frequency inductance formula (Prob.

the inductance of the line

9-13) to

Referring

show

that

is

(H)
a

it

Compare with

(5-13-12).

Assume

>

D, and neglect end

effects.

theory to a circuit consisting of a resistor, inductor, and capacitor

connected in parallel with an alternating-current source 7. Proceeding in an analogous
9-75

manner

Apply

field

to that used for the series circuit in Sec. 9-3,

R
where

+C

dt

V = voltage across parallel combination

R = resistance of resistor
C = capacitance of capacitor
L = inductance of inductor

Li

show

Vdt

that

10
PLANE WAVES IN DIELECTRIC AND
CONDUCTING MEDIA

INTRODUCTION

10-1

The interdependence of electric and magnetic fields is demonstrated in a striking

manner by an electromagnetic wave propagating through space. In such a wave the
time-changing magnetic
field,

which

field

in turn generates

may be

regarded as generating a time-varying electric

a magnetic

field,

and as the process

propagated through empty space at the velocity of

The

first

last

is

part of this chapter deals largely with a plane

wave

in free space

This discussion also applies to any losslesst isotropic dielectric medium.

(vacuum).
In the

repeats, energy

light.

part of the chapter propagation through a conducting

medium

(<x

= 0)

is

considered.

10-2

PLANE WAVES AND THE WAVE EQUATION

a wave propagating toward the reader (out of the page) are indicated
The directions of E and H are everywhere perpendicular. In a uniform
plane wave E and H lie in a plane and have the same value everywhere in that plane.

The field

lines for

in Fig. 10-1

t Referring to Sec. 8-15, lossless implies that e*

o=

0.

= 0,

so that

c',

and further that

364

[CHAP. 10

ELECTROMAGNETICS

t^

H
FIGURE

10-1

wave
Plane
traveling
reader (out of page).

wave of this type with both

approaching

H transverse to the direction of propagation

and

called a Transverse Electro Magnetic

Referring to Fig. 10-2, assume that a plane wave

the

axis.

magnetic

The

field

Since

is

has only a component

H only a component H

polarized in the

density

electric field

we

is

(TEM) wave.
is

in the z direction.

direction (vertically polarized) (see

traveling in the direction of

Ey

in the
It is

Chap.

direction

said that this

and the

wave

is

11).

medium, the conduction-current

Thus Maxwell's equation from Ampere's law reduces to

are dealing with a nonconducting

zero.

VxH =

(1)

5?

or in rectangular coordinates

For a plane wave traveling in the' x direction the only components of (2) that contribute
are

'8HZ

dDy

(3)

Therefore

8HZ

5
C

dx
Maxwell's equation from Faraday's law

VxE =

(4)

dt
is

5B
(5)

It

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

sec. 10-2]

365

^
Direction of

propagation

FIGURE

10-2

Field components of plane

wave with

relation to coordinate system.

or in rectangular coordinates

JdEz

JdEx

dEy\

JdEy

8EZ\

dEx\

For a plane wave traveling in the x direction the only components of (6) that contribute

_.

aref

2^
=
dx

(7)

dt

Therefore

BEy

8x
Equation

(4) relates the

= -n

8H,
(8)

dt

space derivative of

ating (4) with respect to time

H By differentiand (8) with respect to distance x, H can be eliminated

and an expression obtained for

obtain, from (4),

Ey

in terms

of

d_(dH,

and

Jt

and, from

Ey

to the time derivative of

while (8) relates the space derivative of Ey to the time derivative of

x.

Proceeding in this way,

we

Ey

dt

(9)

(8),

dx 2
t If

we had

^ dx \

specified only that the

and

that the

that

B and hence H must

wave

travels in the

wave

(10)

dt )

is

E in the y direction
follows from (6) and also from (7)

linearly polarized with

direction,

it

be in the z direction.

366

ELECTROMAGNETICS

[CHAP. 10

Dividing (10) by fi yields

d

(1 1)

and

d_

(11)

()

dx

does not matter whether we differentiate

(9), it

with respect to

of

Ey

dx 2

ft

Since in

or vice versa, the left-hand side of (9)

first
is

with respect to

x and then

equal to the right-hand side

follows that

it

Equation (12)
of the electric

d 2 E

at*

lie

and

is

(12)

dx

and time variation of the

relates the space

field intensity

3 2

called a

wave equation in

Ey

scalar

magnitude

It is, in fact,

Ey

a scalar

wave equation of the simplest form.

Differentiating (4) and (8) in the reverse order, i.e., (4) with respect to x and
(8) with respect to t, we can eliminate Ey and obtain a wave equation in
t as

dt

Both
(13) has

(12)

many

and

(13) are of the

2
fie

(13)

dx

A wave equation as given by (12) and

same form.

important physical applications and

equation, having been integrated

by him

in 1747.

is

If

sometimes called B'AlemberVs

Ey in (12) is a transverse displace-

ment, the equation can represent the motion of a disturbance on a stretched

This was D'Alembert's problem.

Ey

If

is

can describe the motion of small oscillations of

Ey

string.

a mechanical compression, the equation

air in

a narrow pipe.

In our case

represents the scalar magnitude of the electric field intensity of a plane electro-

magnetic wave progressing in the x direction, and the equation

way of describing
Let us

now

is

the most general

the motion of this field as a function of time and space.

introduce a quantity v in (12) such that

(14)

Equation (12) then becomes

8

Ey

dt

'

Ey

1S

(15)

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-3]

Dimensionally (15)

367

is

Volts

Meter-second 2

so that

_,/,

volts

meters

second
Thus,

it appears that v has the dimensions of velocity.

This velocity is a characteristic
of the medium, being dependent on the constants \i and e for the medium. For
_1
free space (vacuum) v is approximately equal to 300
s

Mm

SOLUTIONS OF THE WAVE EQUATION

10-3

The wave equation

To apply

(10-2-1 5) is a linear partial differential equation of the second order.

the equation, a solution

must be found

for

Ey

Methods of

type of equation are discussed in texts on differential equations.

to say that if

we take

here

the following trial solution

where

solving this

It will suffice

= 2rt/A
= wavelength
m = a constant (to
t = time

sin P(x

mi)

(1)

fi

we

find

be determined)

on substitution in (10-2-15) that

(1) is

a solution provided that

m = v
where v

is

the velocity.

Hence a general

Ey =

sin

P(x

(2)

solution for (10-2-15)

vt)

+ sin /?(* -

vt)

is

(3)

Either term alone is a solution, or the sum, as in (3), is a solution. This can be verified
by taking the second derivatives of the solution in terms of t and x and substituting

them

in (10-2-15).

Since v

=fk,

it

pv
Thus,

(3)

follows that

= y fl = 2ti/=

co

(4)

can also be expressed

Ey =

sin(/?x

cot)

+ sin(/?x cot)

(5)

[CHAP. 10

368 ELECTROMAGNETICS

Suppose that the

significance of (6)

values of the time

instant of time

period

first

t.

term of

can be

(3) is

T/4 or

oat

=
=

<at

That

vt)

is,

(6)

= 0.

Ey as a function of x for several

Ey = sin fix. The curve for this

Then

2n

shown

is

lit

left,

wave is unity.
The point P

is

we can

or in the negative

in Fig. 10-36.

(7)

One-half period

later,

Focusing our attention on the

curve of Fig. 10-3c.

Fig. 10-3

7 (= T4 = 2

of one of the waves, as indicated by the point P,

traveling to the

as a solution.

by evaluating

= (2nf)t

n/2 rad

n, yielding the

P moves to the left. From

this

P(x

itself

and

shown by Fig. 10-3a. Next consider the situation one-quarter

when t = T/4, where T is the time of one period. Then

M=

crest

considered by

sin

illustrated

First let us take

0vt

(c)

is

later, i.e.,

The curve for

t = T/2, and

i-i

iS'

Ey =
The

(a)

= sin (fix + cut) at three instants of time: t = 0, t=T/4,

= T/2. A constant-phase point P moves to the left as time progresses.

Curves for
t

=o

we

note that as time progresses,

thus interpret (6) as representing a wave

The maximum value of Ey

direction.

a point of constant phase and

is

for

characterized by the condition

that

x+

vt

= constant

(8)

+o=

(9)

Taking the time derivative of (8) gives

dx
dt

dx
and

(10)
~dt

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

sec. 10-3]

369

t=0

FIGURE

(a)

10-4

Curves for E, = sin (fix

- a>t) at three instants of time t = 0,t = T/4, and t = T/2.

A constant-phase point P moves to the right as time progresses.

In (10), dx/dt

is

the rate of change of distance with respect to time, or velocity,

of a constant-phase point.

is

traveling in the negative

Next consider the

Hence, v

We

called the phase velocity.

last

= 0,

the velocity of a constant-phase point and

is

negative,

is

which means that the wave

direction.

term of (3) as a solution by

Ey =
Putting in values for

is

note also that v

sin fi(x

T/4,

vt)

itself.

Then

= sin(j8x - cot)

(11)

and T/2, we obtain from (11) the curves of Fig. 10-4.

moves to the right as time progresses. Hence (1 1)

Here a constant-phase point P

wave traveling in the

represents a

positive x direction.
we set x vt equal to a constant and proceed
and (10), we find in this case that
dx
If

di=

in the

same manner as

+V

for (9)

< 12 >

Thus, the wave travels with a velocity v in the positive x direction.

To summarize, a negative sign in x vt or in fix cot is associated with a wave
to the right, while a positive sign is associated with a wave to the left. Accordingly,
when solutions with both positive and negative signs are given, as in (3), two waves are
represented, one to the

the

left

and one

to the right,

sum of both waves.

Let us now treat in somewhat more

and the complete solution

detail the

number of forms can be used which

displacement. Four such forms are
direction.

Ey = sin(/fo cot)
Ey = sm(cot fix)
Ey = cosOSx - cot)
Ey = cos(cot fix)

wave

is

equal to

traveling in the positive

are equivalent except for a phase

370

[CHAP. 10

ELECTROMAGNETICS

C0S( fa-tut)

and

-cos(wt-0x)

s \nl0x-wt)

Ev

FIGURE

10-5

Four forms of wave expression

at

= 0.

^sin(wt-flx)

These can be rewritten so that their relation to the first

recalling that sin(-w) = -sin u = sin(w + n) and cos(-w)

Ey =

sin{cot

Ey =

cos(/to

Ey =

cos(cot

sin(/5x

cot)

sinl fix

cot

fix)

sinl)Sx

cot

fix)

cot

is

more apparent.

= cos u, we

Thus,

have

n)

(14)

-J

-J

The relation of the four forms of (13) is illustrated graphically in Fig. 10-5, where they
are compared for t = 0. On substituting other values of t into the four equations, it
The only
is seen that in each case the equation represents a wave traveling to the right.
of
displacement
n/2 or n.
phase
there
is
a
difference between them is that in some cases
If the

phase displacement

is

disregarded, any of the four forms given by (13)

can be selected to represent a wave traveling in the positive x direction.

the form

we choose

Ey = sin(a>f -

(15)

fix)

This choice has the advantage that the term with the time
traveling in either the positive or the negative

Thus
If

now we

far, it

Since

maximum

amplitude as

sin(cot

t is

positive for

waves

direction.

maximum

has been assumed that the

specify the

Ey = E
explicitly.

Suppose, then,

amplitude of

Ey

is

unity.

we have
(16)

fix)

/=

1/T, (16) can be expressed in a form in which the period

For the sake of symmetry, let us also put jS = 2njk, obtaining

Ey = E Q s.m{l7C^,-2n^

appears

(17)

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

sec. 10-4]

37

These expressions, (16) and (17), represent a wave traveling in the positive x
The corresponding expressions for a wave traveling in the negative x

direction.

direction are

and

The

Ey = E

sin(a>t

Ey = E

sinl27t

fix)

- + In

(18)

(19)

jj

solutions of the

We

solutions.

wave equation given by (16) and (18) are trigonometric

can also express the solution in exponential form. Thus

E^Eoe*"****
where

(20)

understood that the instantaneous value of the field is given by either the
imaginary or the real part of the exponential function. Thus, taking the imaginary
part (Im), we have
it is

Ey = E Im
If the real part (Re) is taken,

we

*><'-**>

=E

sm(cot

fix)

(21)

=E

cos(at

fix)

(22)

obtain

= E Re e

11 *"

" x)

Taking the second derivatives of (20) with respect to

(20)

10-4

is

and

x,

it

may

be verified that

indeed a solution.

TABLE OF SOLUTIONS OF THE WAVE EQUATION

In Table 10-1 both trigonometric and exponential solutions are given for a plane wave
in a lossless medium.
Solutions are listed for waves traveling to the left (negative x
direction) and to the right (positive x direction).
Solutions given by the sum of two

such independent solutions are also included.

Table 10-1

SOLUTIONS OF WAVE EQUATION

Trigonometric form

Exponential form

Wave

to right

Ey = Ey sin(o>< . fix)

E,

= Eye' "" -*>

Wave

to left

Ey = E

E,

=E

Ey

= Ea *">'+*> + ,<*' -*>

Two

waves

Ey =

sin(a)/

+ fix)

Ea sin(a>r + fix) + E

sin(cu/ --fix)

<s

'

eJim,+ 'x >

372

[CHAP. 10

ELECTROMAGNETICS

PHASE VELOCITY

10-5

We have seen that x It

follows that cot

vt is

fix is

a constant for a point of constant phase in a traveling wave.

That

a constant.
cot

fix

is,

and x must vary together, so that

= constant

(1)

Differentiating (1) with respect to time to find the velocity of the constant-phase point,

as

done

in (10-3-9), yields

o>-/J$ =

(2)

at

dx

CO

(3)

H'l
Thus, the phase

That

co/P

where A

of a constant-phase point,

velocity, or velocity

has the dimensions of velocity

is

more apparent

if it is

is

given by

cojfi.

reexpressed

wavelength

/= frequency
The product A/has the dimensions of wavelength (distance) times frequency (reciprocal
of time).

Thus, Xf has the dimensions of distance per time, or velocity.

(10-2-14) we have that the phase velocity is

From

co_

(5)

Equation
ability

p.

medium of permeFor free space (vacuum) the velocity is a well-known

designated by c and usually called the velocity of light). Thus

(5) gives

and

the phase velocity of a wave in an unbounded

permittivity

constant (usually

e.

=300Mms _1

(6)

N/VoCo

The

SI unit for the permeability of vacuum

H
Therefore the permittivity of

= 4007t nH m
vacuum
e

is

_1

(exactly)

is

= -^=8.85pFmHoc

(7)

SEC. 10-5]

For other media the phase

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

373

velocity relative to the velocity of light, or relative phase

velocity, is

VjUoo

(dimensionless)

(8)

s/W**

where

\ir

= relative permeability of medium

= relative permittivity of medium

The phase

velocity of a plane

or less than the velocity of

may have

wave

light (p

<

in

lossless

medium

v is

is

equal to

In general, however, the phase velocity

1).

values both greater and less than the velocity of light.

hollow metal waveguide

For example,

in

always equal to or greater than c (see Chap. 13).

two waves of the same frequency

If

an unbounded

travel with the

same

velocity in opposite

directions or with different velocities in the same direction, the phase velocity of the

resultant

wave

the velocity
trical

is

it is

phase

shift

not a constant but varies as a function of position.

usually

most convenient,

In measuring

at radio frequencies, to determine the elec-

(by phase-comparison methods) between two points, one of which

is

taken as a reference.

For a wave traveling

d(f>

or phase advancement

constant-phase point to

in the positive

d<j>

move this

distance

_L

dt=
where

T is

position

is

x direction there

is

in a positive distance dx.

is

an

infinitesimal phase lag

The time required

for a

then

d<j)

(8)

(9)

Therefore the phase velocity as a function of

the time for one period.

given by

dx

~dt

For a wave

dx
to
= ~
= ~
(TI2n)d4>
d<i>jdx

traveling in the positive

direction d<j>/dx

is

negative,

and hence v

'

is

positive.

As compared with

(10), the

velocity as averaged over

by

an

phase velocity wjp, as in

integral

(5), is

number of wavelengths.

an average phase

Thus, dividing (10)

we obtain a relative phase velocity p which is a function of position as given by

co/p,

^-Wdx
where p
X

<">

_1

= 2n/k, m
= wavelength in free

Both v in (10) and p

space,

in (1 1) are useful

m
where the phase velocity is a function of position.

374

10-6

ELECTROMAGNETICS

[CHAP. 10

INDEX OF REFRACTION

In optics the index of refraction

defined as the reciprocal of the relative phase

is

r\

That is,

velocity/).

= -= Jfi,

=-=

For nonferrous media

V/C

very nearly unity so that

/i r is

r\

EXAMPLE
for paraffin

The phase

= Je

(2)

Paraffin has a relative permittivity

and also the phase

solution

(1)

rt

velocity of

a wave

e,

in

= 2.1.

Find the index of refraction

an unbounded medium of

paraffin.

The index of refraction

7)

= Via = 1.45

= ^==207 Mms-

velocity

V2.1

EXAMPLE

Distilled water has the constants

as 0,

er

= 81,

fi,

= \.

Find

i)

and

v.

SOLUTION

= V81 = 9

=-== 0.1 lie = 33.3 Mms"

V81
The index of refraction given for water in the above example
At light frequencies, say for sodium light

frequencies (/->0).

index of refraction

is

5893 A), the

This difference was at one time cited as invalidating

Maxwell's theory.

The explanation

not a constant but

is

the value at low

(A

observed to be about 1.33 instead of 9 as calculated on the basis

of the relative permittivity.

light frequencies e,

is

for the difference

is

that the permittivity e

is

At zero frequency c, 81, but at

The index of refraction and permittivity of many

a function of frequency.

1.33

1.77.

other substances also vary as a function of the frequency, as explained in Sec. 8-15.

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-7]

10-7

375

GROUP VELOCITYf

Consider a plane wave traveling in the positive x direction, as in Fig. 10-2.

total electric field

Let the

be given by

Ey = E

cos(cot

Px)

(1)

Suppose now that the wave has not one but two frequencies of equal amplitude
expressed by

m + Aco
and
It

co

follows that the

Aco
these

P values corresponding to

and
For frequency

two frequencies are

A/?

corresponding to

co

Aco

A/5

corresponding to

co

Aco

E'y

=E

cos[(co

Aco)t

(P

AP)x]

(2)

E;

=E

cos[(co

Aco)t

(P

AP)x]

(3)

and for frequency 2

Adding

gives the total field

Ey

E;

e;

(4)

or

Ey =

{cos[(tu

AoJ)t

0?

AP)x]

cos[(a>

- Aco)t - (fi - AP)x]}

Multiplying out (5) and by trigonometric transformation

Ey = 2

cos(co

- p o x) cos(Aco t -

The two cosine factors indicate the presence of

imposed on a more rapid one.

beats,

we

get

AjS x)

i.e.,

(5)

(6)

a slow variation super-

For a constant-phase point

a>

dx

p x
co

-- =
-r=
at
p

and

constant
,_.

tf=/oA

(')

Leon Brillouin, " Wave Propagation in Periodic Structures," chap. 5, McGraw-Hill

Book Company, New York, 1946; J. A. Stratton, " Electromagnetic Theory," p. 330,
McGraw-Hill Book Company, New York, 1941.

376

ELECTROMAGNETICS
[CHAP. 10

where v

is

the phase velocity.

to a constant,

Setting the

we have

argument of the second cosine factor equal

Aw t - Ap x = constant
dx

and

Aco
(8)

where u

the phase velocity of the wave envelope,

which is usually called the group
In the above development we can consider
co + Aco and co - Aco as the
two sideband frequencies due to the modulation
of a carrier frequency co by a
frequency Aco, the carrier frequency being
suppressed.
In nondispersive media the group velocity
is
is

velocity

the same as the phase velocity

an example of a lossless, nondispersive medium,
and in free space
u-v-c. However, in dispersive media the phase and group
velocities differ
A dispersive medium is one in which the phase velocity is a function
of the frequency (and hence of the free-space wavelength).
Dispersive media are of two types

Free space

is

Normally

length

that

In these media the change in phase

velocity with waveis, dvjdX > 0.
For these media u < v.

dispersive.

positive; that

is

Anomalously

dispersive.

wavelength

negative; that

is

In these media the change in phase

velocity with
is, dv/dX < 0.
For these media u > v.

The terms normal and anomalous are arbitrary,

the significance being simply
anomalous dispersion is different from the type of
dispersion described as normal
For a particular frequency (bandwidth vanishingly
small)

dp

Aco
dco
u= hm -
=
,

aco-o

But

w = 2ti/ = 2nfl/X = pv;

or

It

also be

shown

dco

d(Pv)

dp

dp

PR

(1 1)

^ V

(10)

dv

+p
p

dp

= v-X

(11)

dv

di
Equations

dv
dfi

that

velocity function.

(9)

so

may

Ap

(12)

and (12) are useful in finding the group velocity

for a given phase-

/
PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-7]

EXAMPLE A 1-MHz (300 m

lossless

wavelength) plane wave traveling in a normally dispersive,

medium has a phase velocity

as a function of wavelength

is

377

at this frequency of

Mm s~ \

300

The phase

velocity

given by

v=kV\
where k

is

Find the group

a constant.

solution

From

(12) the

velocity.

group velocity

is

.do

or

= v(l i)

Hence

k=^ = 150 Mms-'

2

To

illustrate graphically the difference

between phase and group

velocity, let us

consider a wave of the same phase-velocity characteristics as in the above example

and assume,

further, that the

amplitude, where
carrier

modulated

/ =

100

at

wave has two frequencies,

and A/= 100 kHz. This

MHz

kHz

instantaneous magnitude of

Ey

+ A/and/
is

with the carrier suppressed.

A/, of equal

equivalent to a

From

(6)

1-MHz

graphs of the

as a function of distance (plotted in meters) are pre-

t = 0, t T/4, and t = T/2.

The point
P is a point of constant phase of the wave proper and moves with the phase velocity v.

sented in Fig. 10-6 for three instants of time,

The point P' is a point of constant phase of the envelope enclosing the wave and moves
with the group velocity
has

moved a

That

is

u.

It is

distance d' which

apparent that in one-half period (T/2) the point P'

is

to say, the group velocity u

one-half the distance

is

d moved by the point P.

v.
The intelligence

one-half the phase velocity

conveyed by the modulation moves with the velocity of the envelope,

i.e.,

at the

group

velocity.t

The difference between phase and group velocities is also illustrated by a crawling
The humps on his back move forward with phase velocity, while the

caterpillar.

caterpillar as

a whole progresses with group velocity.

(steady-state) wave the group velocity

However, if the wave consists of two or more frequencies, or a
frequency group, as in a modulated wave the group velocity may be observed because

For a single-frequency constant-amplitude

is

not apparent.

the wave amplitude

that

may

is

nonuniform and the individual waves appear to form groups

be enclosed by an envelope, as in Fig. 10-6.

t In a lossless

medium

the energy is also conveyed at the group velocity.

378

ELECTROMAGNETICS

[CHAP. 10

period
later

"2

period
later

FIGURE

10-6

Constant-phase point P of the wave moves with phase velocity v, while point P'
on the envelope moves with group velocity u. In this example the group velocity
is one-half the phase velocity.

IMPEDANCE OF DIELECTRIC MEDIA

10-8

Thus

far our attention has

also the magnetic field

find
in

been focused on the

electric field

Ey

Let us

now consider

of the same plane wave (see Fig. 10-2) and, in particular,

how Hz is related in magnitude to Ey

A solution of the wave equation (10-2-13)

H. is

=H

sin(eo<

fix)

This solution represents a wave traveling in the positive

Ey representing

a wave in the positive

x direction

Ey = E

sin(c<tf

is

(1)

direction.

solution for

given by (10-3-15) as

fix)

(2)

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

sec. 10-8]

To

find

how

in (2)

Ey we

related to

is

(3)

dx

Ey

(10-2-8) that

dH,

dv

Substituting (2) for

from

recall

dt

performing the indicated differentiation, and then

in (3),

379

in-

tegrating with respect to time yields

sin((ot

- fix)

(4)

fico

Taking the ratio of

to

Ey

to

for a single traveling wave, as given

by the

ratio of (2)

(4), we obtain

~
E

or

(5)

yle

JJu

*=Ji H

(6)

For comparison we can now write

Ey = E

and

It is

apparent that

differ

Ey and

by a factor ^Jn/e or

sin(a)t

= J- E

its

sm(cot

(V

fix)

m"

(Am

fix)

are identical functions of

(7)

')

x and

(8)

t,

but their magnitudes

reciprocal.

The dimensions of (5) expressed

in SI units are

Volts/meter

volts

Amperes/meter

ampere

= ohms

Thus, y/n/e has the dimensions of an impedance, and we can write

w?
where

is

called the intrinsic impedance of the

Z = Z,

(9)

medium.

376.731

120w

For

il

free space

(vacuum)

(10)

'

380

ELECTROMAGNETICS

[CHAP. 10

(a)

Direction of propagation

(b)

FIGURE

10-7

Instantaneous values of E, and H, along x axis (a) at time / =

and (6) \
period later. In this interval the point P has advanced A/4 to the right.

If

space and

and

H are in time phase, Z

is

a pure resistance.

To emphasize

media.

all lossless dielectric

This

is

the case for free

the fact that the impedance

is

a pure resistance, one can speak of the intrinsic resistance (instead of impedance) of
free space and of lossless dielectric media.
If E and
are not in time phase, as in

conducting media, the ratio of

intrinsic

complex, so that the more general term

impedance must be used in connection with such media.

Introducing

(9) in (5),

to

is

we have

(11)

EXAMPLE
for a plane

If the

wave

solution

magnitude of

H in a plane wave

is 1

A m~

1
,

find the

magnitude of

in free space.

From

(11)

E = ZH = 376.7

= 376.7 V m"

The instantaneous values of Ey and z along the x axis for a plane wave prox direction are illustrated in Fig. 10-7. Figure 10-7a shows the
condition at the time t = 0, while Fig. 10-76 shows the conditions one-quarter period
later (t = T/4).
The maximum values of Ey and z (E and
) are shown to be
gressing in the positive

equal.

Hence,

if

the

medium

is

free space, the scale in volts per

meter along the y

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

sec. 10-9]

axis should

be 377 times the scale in amperes per meter along the z

would be equal, however,

if

10-7 both the magnitudes

and

x axis.

Since

we

the relations of

the

medium had an

directions of

are considering a plane

Ey

and

along

Ey and Hz are

wave

axis.

impedance of

intrinsic

shown

The

to the

scales

In Fig.

1 2.

for points along the

traveling in the direction of the

all lines parallel

381

axis are the

x axis,

same as those

shown.

10-9

IMPEDANCE OF TRANSMISSION-LINE CELL

Imagine a plane wave traveling out of the page (toward the reader) with space divided

up by thin conducting strips into transmission-line cells, as suggested in Fig. 10-8.

The strips are normal to E, and they extend infinitely far normal to the page. The field
H is parallel to the strips. At the surface of each strip there is a current sheet of linear

K (normal to the page) that is equal in magnitude to H. The currents on

opposite sides of one strip are opposed as shown.
plane (TEM) wave traveling

density

through space divided up into line cells as in Fig. 10-8 has the same values of E and
everywhere as a wave in empty space (no strips present as in Fig. 10-2), it being assumed
that the strips are of infinitesimal thickness.

spacing are equal (both equal to

/),

Assuming

also that the strip width

so that the cell cross section

of the strips are infinitesimally close together, the voltage

is

and

square and the edges

between two

strips

of one

cell is

V=
The

total current / flowing

El

(1)

on the bottom surface of the top

is given by

strip

of a

cell (or

top

surface of the bottom strip of a cell)

Dividing

(1)

by

= Hl

(2)

(2) yields

(3)

"*

H
m
3

ft

8
i

H
FIGURE

transmission-line cells.

*-

Conducting
strips

10-8

Plane wave traveling out of page with

space divided by conducting strips into

<

Current on
strips

382

ELECTROMAGNETICS

[CHAP. 10

Thus, for a single traveling wave the

intrinsic

impedance

E/H of a wave

characteristic impedance V/I of a single transmission-line cell.

(10-8-11)

is

equal to the

Furthermore, from

we have

Z=
For a transmission-line

cell

we have from

unit depth (normal to the page)

is

(4)

(5-17-2) that the inductance of the line per

equal to the permeability of the medium,

i.e.,

k=V
while

we have from

(5)

(3-22-3) that the capacitance of the line cell per unit depth equals

the permittivity of the medium, or

C=

(6)

d
Accordingly,

we can

write

(7)

Z = characteristic impedance of line cell, fl

= inductance per unit length of line cell, H m _1
_1
C/d = capacitance per unit length of line cell, F m

where

Ljd

To summarize,
intrinsic

and

the characteristic impedance

impedance of the medium filling the

e

If the

so the characteristic impedance of the line

A<o

10-10

of a transmission
cell.

line cell is

medium

is

equal to the

free space,

p.

fi

is

= 376.7 Q

TWO PLANE WAVES TRAVELING

(8)

IN

OPPOSITE DIRECTIONS; STANDING WAVES

Thus

far,

we have

considered only a single traveling wave, such as a wave moving in

now examine the situation which exists

the positive or the negative

x direction.

when

traveling in opposite directions, such as the negative

there are

two waves

Let us

and

SEC. 10-10]

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

Medium 2

Medium

383

wave

Incident

FIGURE

10-9

Relation

of

incident,

reflected,

and

transmitted waves.

Assume that the two waves are of the same frequency and of
The condition that the waves be of the same frequency and form
is automatically fulfilled if one wave is a reflection of the other since both then originate
from the same source.
positive

directions.

sinusoidal form.

Referring to Fig. 10-9, assume that space

divided into two media,

is

A wave originating in medium

with a plane boundary between as shown.

1
1

and
and

2,

in-

cident on the boundary

is said to be the incident wave.

The wave reflected from the
boundary back into medium 1 is called the reflected wave. If the reflection of the
incident wave at the boundary is not complete, some of the wave continues on into
medium 2 and this wave is referred to as the transmitted wave.

Ey

In the solution of the wave equation for

terms, the

first

representing a

wave

second a wave in the positive x direction (to the

tial

solution in Table 10-1, let the incident

Ey0 = E
and the

reflected

wave

where S

= time-phase lead of Eyl

of incident wave

of reflected wave

by

(10-3-5) there are

direction (to the left)

Referring

two

and the

now to the exponen-

(traveling to the left) be given

by

" +fix)

ei(a

(1)

by

E^

<ot-px + 5)

with respect to

point of reflection

E = amplitude
E = amplitude

right).

wave

(traveling to the right)

6,1

as given

in the negative

(2)

Ey0 at x = 0, that is,

= phase shift at

384 ELECTROMAGNETICS

[CHAP. 10

2E

\'"

Bo
'

//

//
/?*

/ /'-fy

V\

/2r

3ir

V
~E

^\

i/ vv
t=

^yV
XP//

-2E

FIGURE

10-10

Pure standing wave showing

The dot is used here to indicate explicitly that Ey0 and Eyl
The total electric field Ey is
t, x, and <5.f
Ey

Ey at various instants

of time.

are complex functions of

EyQ + Eyl

(3)

The instantaneous magnitude of the fields is obtained by taking either the real
(Re) or imaginary (Im) parts of (1) and (2). Thus, taking the imaginary parts, the
total instantaneous electric field is

Ey = Im Ey = E
If

<5

sin(wr

fix)

sin(<?

-px + S)

(4)

or 180 (4) can be expanded as follows:!

Ey = E

sin cot cos fix

Collecting terms,

+E

cos

cot sin fix

+E

sin cot cos fix

- El

cos

cot sin /?*

(5)

we have

Ey = (E +

)sin cot cos fix

(E

- EJcos cot sin fix

(6)

a perfect conductor, the reflected wave is equal in magnitude to the

is taken to be at the boundary between media 1 and 2, the
incident wave. If x =
If

medium 2

is

&

and yl are scalar space components of the total

t However, both
In the case being considered, E has only one component. Thus E
t

sin(o6) = sincos b cos a

sin 6.

vector E.

field

= $y

SEC. 10-10]

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

P
l.5E ~

385

"""^

Envelope^x "
-

1.0Eo

,.iV\

0.5Eo

\t
0.5

^^

^<=^T_

ffix-

*p\

1.0Eo

\^_^<^v

1.5E

FIGURE

Envelope

10-11

Standing-wave envelope for E t = -0.5Eo with associated (traveling) wave at

four instants of time: t = 0, t = T/8, t = 774, and f = 3778.

boundary relation for the tangential component of E requires that

El
at the boundary (5 = 180).
Thus (6) becomes

=-E

Ey = 2
This represents a wave which

cos

Ey

0,

so that

cot sin fix

(7)

The values of Ey at a
The instantaneous values at a particular
The peak value of the wave is the sum of the incident
stationary in space.

is

particular instant are a sine function of x.

point are a cosine function of f.

and
\E

reflected
\

is

peak values or 2E

a pure standing wave.

The space and time

curves of Fig. 10-10.

not

variations of

It is

Ey for

is

for

which

E =
x

associated with resonators.

a pure standing wave are shown by the

to be noted that a constant-phase point, such as P, does

move in the x direction but remains

Now let

wave of this type

stationary

This type of wave

at

us examine the conditions

incident wave, say one-half as large.

a fixed position as time passes.

when

Then,

the reflected

E = -0.5E
x

case at four instants of time gives the curves of Fig.

1 0- 1 1

wave

is

smaller than the

Evaluating (6) for this

The curves show the values

of E as a function of fix at times equal to 0, |, i, and f period. The peak values of

Ey range from 1 .5 at t = to 0.5o at t = i period. The peak values as a function

of x as observed over an interval of time greater than one cycle correspond to the
This envelope remains stationary, but focusing our attention

envelope as indicated.

on a constant-phase point

travels to the

also be noted that the velocity with which

constant.

i period

left.

It will

Between time

P moves

of the wave, we note that the total instantaneous wave

and | period P moves about 0.05A

about 4 times as

far,

or about 0.2A

(0.4ji).

(0. lrc),

P moves is not

while in the next

Although the average

386

[CHAP. 10

ELECTROMAGNETICS

velocity of the constant-phase point

is

the same as for a pure traveling wave,

instantaneous magnitude varies between values which are greater and

its

less.

To summarize, there are two Ey waves, one traveling in the negative x direction
and another one-half as large traveling in the positive x direction. The waves reinforce
each other at some points and subtract from each other at other points. The resultant
x direction.
The envelope of the instantaneous curves in Fig. 10-11 can be called a standingwave curve, or envelope. At any position fix the maximum value of the field at some

wave

travels in the negative

time during the cycle

equal to the ordinate value of the envelope.

is

To calculate the value of the standing-wave envelope, we may proceed as follows.

In (6) put

A = ( + )cos fix
B = ( - ,)sin fix

(8)

Expanding

sin cot

and cos

A
Equation

(6)

sin

cot

mt

(9)

in terms of exponentials,

+ B cos cot =

JA

+ B2

we can show

sm(cot

that
(10)

y)

can then be written

Ey =
Expanding

Ey =

(1 1)

JA 2 + B 2 sin(cot + fix)

by means of

y/(E

+ E^ 2

cos

(8)

and

fix

(1 1)

(9) yields

j)

sin

fix sin(a>f

fix)

(12)

The maximum value of Ey at some position fix as observed over an interval of at least
Thus for the shape of the standing-wave
one period occurs when sin(cot + fix) = 1
have
of
E
we
envelope
y
.

Ey =
Ordinarily

we

s/(E

are not so

Ey) cos

much

fix

(E

,)

sin

fix

(13)

interested in the shape of the standing-wave

envelope as given by (13) as in the ratio of the maximum to minimum values for the
envelope, which is called the Standing- Wave Ratio (SWR). The potential or voltage at

and so the SWR in this case may be

The maximum value of the
envelope corresponds to the sum of the amplitudes of the incident and reflected waves
( + i), while the minimum corresponds to the difference between the two ( E t ).
With this information we can determine the fraction of the incident Ey wave which is
reflected, forming the reflected wave, and also that which is transmitted (see Fig. 10-9).

any distance x

will

be proportional to the

field,!

referred to as the Voltage Standing- Wave Ratio

t Provided

E is integrated

(VSWR).

over equal paths

/(K=JE</1

= El).

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-10]

.As

will

be noted

knowledge

later, this

is

387

of value in determining the nature of the

conditions at the point of reflection.

VSWR we can write

Thus, for the

+ El
E Et

VSWR - max
Emi

(14)

VSWR is unity. When the reflected

VSWR is infinite. Hence for all
intermediate values of the reflected wave, the VSWR lies between
and infinity.
When
wave

the reflected wave

is

is

zero (E t

equal to the incident wave (E t

the

0),

the

),

The
coefficient.

of (2) to

wave to the incident wave is defined as the reflection

the point of reflection (x = 0) and at the time t = 0, the ratio

ratio of the reflected

Thus, at

(1) is
.

The dot

indicates that

E ie}i

Eyl

EJ8

a complex function.

Thus, in general, p expresses both

wave with respect to the incident wave and also the
phase shift 8 at the point of reflection. The magnitude of p can range between and
1 with phase angles between
and +180.

is

the magnitude of the reflected

In this discussion,

more

specifically

may

we

are dealing with the values of the electric

be called the

field.

Hence,

field reflection coefficient.

Rewriting (14) and substituting (15) gives

VSWR

1
=

+ (E IE )
- (EJEo)
t

+ |p|
1-\P\
l

(16)

Solving for \p\ gives an expression for the magnitude of the reflection coefficient in

terms of the

VSWR:
IpI=p =

VSWR -

07)

vswrTT

In Fig. 10-12, standing-wave envelopes are presented for three magnitudes of

The amplitude
p = 0, 0.5, and 1
The curves show Ey as a function of position
in terms of both fix and wavelength.
For complete reflection (p = 1) we have a pure
standing wave with a VSWR of infinity. For zero reflection (p = 0), the VSWR
is unity, and E is constant as a function of position.
For a reflection coefficient
y
of 0.5, the curve varies between 1.5 and 0.5 so that VSWR = 3. In general, the standing-wave envelope is not a sine curve, the minimum being sharper than the maximum.
the reflected

wave as given by reflection

of the incident wave

is

taken as unity.

coefficients

388

[CHAP. 10

ELECTROMAGNETICS

0.25X

FIGURE

0.75A.

10-12

Standing-wave envelopes for three magnitudes of reflection

0.5,

and

coefficient,

= 0,

1.

by the curve for p = 0.5. However, in the limiting condition of

p = 1 the curve does have the form of a rectified sine function (|sin fix\). Also, as
is approached, the curve approximates a sinusoidal variation.
the condition p =

This

is

illustrated

ENERGY RELATIONS

10-11

From

(3-12-7) the energy density

we

IN

at

a point in an

we = $eE
where

A TRAVELING WAVE

(Jm-

electric field is

(1)

_1

e = permittivity of medium, F m
_1
E = electric field intensity, V m
From (5-19-2) the energy density wm

at a point in a

(Jm" 3 )

wm = inH 2
where n

= permeability

H = magnetic

of medium,

field,

A m"

Hm

(2)

medium

-= ft
H yje
H from (3) in

field is

In a traveling wave in an unbounded, lossless

Substituting for

magnetic

(2),

(3)

we have

wm = ipH 2 =

ieE

= we

(4)

SEC.

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

HM2]

389

Thus the electric and magnetic energy densities in a plane traveling wave are equal,
and the total energy density w is the sum of the electric and magnetic energies. Thus

w=

We

+ wm = \cE 2 + iuH2

(5)

(Jm" 3)

(6)

w = eE 2 =

or

10-12

iiH

THE POYNTING VECTOR

Continuing the discussion of the preceding section, any increase in the energy per
Likewise any decrease must
unit volume must be produced by an inflow of energy.
be equal to an outflow (principle of conservation of energy). Thus, for a small
volume Av the decrease in energy as a function of time may be expressed asf

-j (hE

+ ipH 2 )Av = jS-ds

(W)

(1)

where S

is

volume Av

the energy per unit area passing per unit time through the surface of the
in watts per square meter.

Dividing

(1)

by Av and taking the

limit as

Av approaches

V-S=-|( 2 + iu/f2
The quantity S has the dimensions of power per
per square meter.
flow but also

its

It is

a vector since

direction.

It is

it

zero,

(2)

unit area

and

indicates not only the

is

expressed in watts

magnitude of the energy

usually called the Poynting vector.

VxH = ?
ot
first

E-(VxH) = E-y

We

J.

assume no

and

we have

VxE=-?

(3)

at

equation with

and

loss or generation

On

obtain

Returning to Maxwell's equations for nonconducting media,

Writing the scalar product of the

we

E and of the

second with

H(VxE)=-H-

H gives
(4)

of energy within the volume.

the Transfer of Energy in the Electromagnetic Field, Phil. Trans.,

174:343 (1883); Oliver Heaviside, "Electromagnetic Theory," vol. 1, p. 78, Ernest
Benn, Ltd., London, 1893.

H. Poynting,

390

ELECTROMAGNETICS

Subtracting the

first

[CHAP. 10

from the second, we have

H-(VxE)-E-(VxH)=-(e'? + H-?)

(5)

at

at

By means of the conversion formula

G
the two terms on the

-F

(V x F)

left side

(V x G)

(F x G)

(6)

of (5) can be expressed as one

V(ExH)=
v

-|E +H-
Bt)

D = cE

and

,. (e

B=

and

fiH,

(7)

at

For isotropic media,

=V

(7)

takes the form

= -[e .*!> + h.2>]

(8)

Recalling from calculus that

w
we

dw 1 dw 2
Tt = 2-a7

(9)

see that (8) can be written

Taking the time-derivative operator outside the brackets, we have

V (E x H) =

A (leE 2 + inH
at

Comparing

(11) with (2)

2
)

shows that

Ex H = S
where

(1 1)

E = electric field intensity, Vm"

H = magnetic field, A m~
S = instantaneous Poynting vector,

(12)

The dimensional form of (12)

in SI units

Volts

x
Meter

W m~

is

amperes

watts

meter

meter 2

In words, this important relation indicates that the rate of energy flow per unit area
in a

wave

is

directed

normal to the plane containing

E and H and

has a magnitude in

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-12]

watts per square meter equal to

It is

EH sin 6,

where 6

is

the angle between

391

and H.

often helpful to regard the Poynting vector as a surface power density.

For a plane wave traveling

in the positive

direction, as in Fig. 10-2, (12)

becomes
y x

tHz = ZSX = S

should be mentioned that in some situations

It

the integral of the normal

total

power through the

component of E x
That

surface.

S
where

is

the

above relation

Ex

on a

flow, e.g., in a static magnetic field superimposed

(13)

H does not represent energy

static electric field.

However,

H over a closed surface always gives the

is,

ds

=P

(W)

(14)

power flowing out of closed surface s. The Poynting vector S in the

is the instantaneous power density, and P in (14) is the instantaneous

power.

When E and
power.

This

is

H are changing with time,

period and dividing by one period.

as follows.

we

are often interested in the average

obtained by integrating the instantaneous Poynting vector over one

It is

also readily obtained using complex notation

In complex notation the complex Poynting vector

S=

iExH*

E = & E E = k E E e3""
& E = unit vector in E direction
H* = & H H* = k H H e- Ji "-
k H = unit vector in H direction
f = phase angle between E and H*
(15) H* is called the complex conjugate of H,

is

given byf

(15)

where

In

H = &H H

e +Ji

where
o, - i)

'

(16)

The quantity H and its complex conjugate H* have the same space direction and the
same amplitude
but they differ in sign in their phase factors. Assuming that the
space directions of E and H (or H*) are normal to each other, the complex Poynting
vector is normal to the plane containing E and H* and is

where n

is

S = itE

E and

H*.

the unit vector normal to

eK

(17)

Although no distinction in notation is made in the symbol S for the Poynting vector
between instantaneous and complex values, the meaning should be clear from the
context.

392

ELECTROMAGNETICS

Now

[CHAf. 10

the average Poynting vector

Sav

is

given by the real part of the complex

Poynting vector, or

Sav = Re S =
where
that

is

ifl

(W m~ 2)

cos

the time-phase angle between electric and magnetic

E and

one uses the rms

are the amplitudes or

Pav

fields.

peak magnitudes of the

values, the factor \ in (18)

The average power

(18)

is

It is

understood

fields.

If,

instead,

omitted.

flowing outward through a closed surface can

now

be

expressed as

P..

<|>

Re S

ds

= i j Re(E x

H*) ds

Dividing the Poynting vector by the energy density,

the dimensions of velocity.

(W)

we

(19)

obtain a quantity with

Thus
Poynting vector
velocity

(20)

Energy density
or in dimensional symbols

This

is

the energy velocity v en

the phase velocity

MT~

ML-^T-

In nondispersive media the energy velocity

In a nondispersive lossless

v.

EH

equal to

In dispersive media that are also lossless the energy velocity

is

equal to the group

In absorbing media (not lossless) where the absorption

velocity.

is

medium we have

is

not small the

group velocity tends to lose its simple significance. However, the energy velocity
still has a simple, definite meaning.
Whereas the phase velocity (and also the group velocity) may assume values
greater or less than the velocity of light, the energy velocity never exceeds the velocity
_1

oflight(300Mms

EXAMPLE
1

W m"

2
.

).

plane traveling wave in free space has an average Poynting vector of

Find the average energy density.

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-12]

solution

The average energy

w, y

393

density

= r^-r- 3.33 nJm"

velocity

The energy

density

an average

is

The instantaneous value may be any

The difference between the average

value.

value between zero and twice the average value.

and instantaneous values

is

more

discussed in

detail in the following paragraphs.

Returning now to a further consideration of a plane wave traveling in the positive

x direction

medium,

in a lossless

vector equation (13).

let

us substitute (10-8-7) and (10-8-8) in the Poynting

This yields

>-&'' sm

(Wm -2

(cot-Px)

(21)

is the instantaneous power per unit area.

At a fixed position, fix = constant,
and this power pulsates with the passage of time as a sine-squared function. The
peak power per unit area occurs when

This

sin (cot

fix)

(22)

giving

P^kSx = J-E
where

is

the intrinsic resistance of the

area, the instantaneous value given

by

=^

medium.

To

(23)

find the average power per unit

(21) is integrated over

one period and divided

by the length of one period, which yieldsf

Average SX

Comparing

(24)

and

(23)

\JW = ^

2(average

(24)

shows that

Peak Sx

The energy flow per

Sx)

(25)

power surface density, of a plane

x direction is illustrated in Fig. 10-13. Here the instantaneous values of Sx are shown at two instants of time, t =
and t = T/8. The
values of the instantaneous power per unit area are given over a distance of 1 A in the
x direction. Confining our attention to one of the curves, say for t = 0, we note that
the power surface density is a pulsating quantity, with two pulses per wavelength.

wave

unit time per unit area, or

traveling in the positive

The same

result

may

be obtained from (18) by noting that in a

lossless

medium

394

[CHAP. 10

ELECTROMAGNETICS

Wm- 2

ix

i*

FIGURE 10-13
Instantaneous Poynting vector magnitude for plane wave traveling in positive
and I period later.
direction at two instants of time: t =

Hence, at a fixed position there are two pulses per period or

curves for

and

T/8,

we note

that the pulses

move

power flow

A/8 in the interval of period, indicating a

cycle.

Comparing the

to the right a distance of

in the positive

direction.

To obtain the instantaneous energy density of the wave, the instantaneous

Poynting vector is divided by the velocity of the wave. Thus, the instantaneous
energy density

is

given by

w=-=
Since l/v

yfejx, (26)

-E
n

sin (cot

(26)

fix)

becomes

w - eE
By

v \

steps similar to those for (23)

sin (cot

fix)

(Jm" 3 )

(27)

and (24) the peak and average energy

Peakw =
Average

w = \eE

(Jm -3 )

2
2

(Jm"

densities are

(28)

(29)

Table 10-2 summarizes the relations developed above for the Poynting vector

and energy density of a single plane traveling wave.

The instantaneous energy-density distribution of the plane wave traveling
the positive x direction has the same form as the Poynting vector curves shown
Fig. 10-13.

The

per cubic meter.

ordinate, however, for energy-density curves

is

in
in

expressed in joules

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

sec. 10-12]

From

the energy point of view,

it is

395

interesting to consider the energy per pulse

Here we consider that one pulse is A/2 long. Its boundaries are defined
by the positions where the energy density is zero. Since we are dealing with a plane
2
wave of infinite extent, let us confine our attention to a volume 1
(parallel to the
of the wave.

wavefront) by one pulse length (in the x direction).

Since one pulse length

is

A/2

volume (one pulse long) is obtained by multiplying the average

energy density by A/2 expressed in meters. Thus
long, the energy in this

Energy

where A

is

EXAMPLE
wavelength

(average w) -

The average energy

2
is

500

mm,

From

wave
volume one pulse length by

density of a plane traveling

find the energy in a

is

the energy in a

is 1

m~

If the

2
.

(30)

Energy

0.25

(30)

the wavelength.f

solution

This

(J)

volume

=iJ

by one pulse length long, or the energy

in

a volume of

3
.

Table 10-2

POYNTING VECTOR AND ENERGY-DENSITY

RELATIONS FOR A PLANE TRAVELING WAVE
IN A DIELECTRIC MEDIUM
Poynting vector S,

Condition

Instantaneous

Wm"

/-

sin (o>/

- fix)

Peak

Average

Total energy density w,

Jm~ 2

iff-ifi*-

sin (cu/

= V-ffo 2

ieE

Each pulse contains w p jhf quanta, where w, = energy per pulse

= 6.63 X 10 ~ 34 J s, and /= frequency in hertz.

constant

- fix)

= i/xHo 1

in joules,

= Planck's

396

[CHAP. 10

ELECTROMAGNETICS

It is

Ey and z have a sine distribuThe Poynting vector S and the energy density w, however,

to be noted that the instantaneous values of

tion as a function of x.

have a sine-squared distribution

The energy

density

w which we have

traveling

wave

is

equally divided at

been discussing

all

the total energy density

times between the electric energy density and

This relationship

the magnetic energy density.

is

This energy density in the case of the plane

of the plane electromagnetic wave.

given by (10-11-5).

is

Equation (21) for the magnitude of the Poynting vector of a plane traveling
wave is written in terms of E. Since the power is equally divided between electric

and magnetic forms, we can write

and
where S'e
S

s;

= ^-o 2 sin>r-j8*)

S'm

sin ((ot

(31)

(32)

fix)

= electric power per unit area

= magnetic power per unit area

and S'm are identical. Since S'e = S'm for the plane
Thus
can
traveling wave, the total power
be expressed as twice the value of S'e

The phase

factors for both S'e

S=

S'e

where the Poynting vector magnitude S

Likewise, the energy density
netic

is

S'm

= 2S'e

(33)

equal to the total power per unit area.

is

equally divided between the electric

we and mag-

wm , so that
2

we = eE
wm = inH

and

we = wm for a plane traveling wave,

Since

sin (eo<

sm 2 (<ot the total

(34)

fix)

(35)

fix)

is

twice

we

Thus

w = we + wm = 2we = 2wm

(36)

10-13

ENERGY RELATIONS IN A STANDING WAVE

Next

us consider the energy and power relations for two plane waves traveling in

let

Assume that both waves are polarized with E in the y direction.

wave travels in the negative x direction and has an amplitude
other wave travels in the positive x direction and has an amplitude E v

opposite directions.

Assume

further that one

while the

In this case the instantaneous value of

E =E

sia(cot

Ey

resulting

fix)

from the two waves,

sm(cot

fix)

is

given by
(1)

PEANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-13]

We may find

a corresponding relation for

as follows.

397

Let us start with (10-2-8),

i.e.,

*&--!*
h
dx

Substituting
respect to

/,

Ey from (I) into (2),

(2)
K
'

dt

differentiating with respect to x,

and integrating with

we obtain

-I- E

sin(<of

The magnitude of the Poynting vector

fix)

Et

I/-

sia(cot

(3)

fix)

is

Sx = Er Hz
Substituting (1)

and

(3) in (4) yields

Sx =
According to

E > Et

(5),

-I- [E
V

sin (a>t

+ fix) - , 2 sin 2

ft

the net Poynting vector

is

on a plane boundary

at

M-

in the negative

Furthermore, the net Poynting vector

Poynting vectors for the two traveling waves.

incident

(4)

is

(5)

fix)]

direction provided

equal to the difference of the

Suppose that the wave to the left is

The wave to the right then

(as in Fig. 10-9).

reflected wave.
If the medium to the left of the boundary is a perfect
we have the condition at the boundary that E t = E resulting in a pure
standing wave to the right of the boundary. We note that for this condition the net
Poynting vector is zero, and hence no power is transmitted.
It is interesting to examine the condition of a pure standing wave (E 1 = )
in more detail, particularly from the standpoint of concentrations of energy.
Accordingly, let us find the values of the electric and magnetic energy densities separately.
Substituting (10-10-7) into (10-11-1), we obtain, for the electric energy density of a

becomes a

conductor,

pure standing wave,

we = 2eE
Taking

(3),

cos

cot sin

fix

expanding, collecting terms, and putting

= -2H

E lyJeln =

sin cot cos fix

(6)

we

get

(7)

Substituting this in (10-1 1-2) yields the value of the magnetic energy density of a pure

standing wave,

wm = 2fiH

sin

cot

cos fix

(8)

398

[CHAP. 10

ELECTROMAGNETICS

Zero
time

>

ta

%-

Energy

\w.

all

electric

3r

Zv
fix

2-

i
8

period

Direction
- uiretuun

of
ui

poynting vector

Energy half

later

electric, half

magnetic

-J-

period

Energy all
magnetic

,
1|-

later

FIGURE 10-14
Total electric and magnetic energy densities at three instants of time for a pure
standing wave. Conditions are shown over a distance of lA (fix = 2rr). There
is no net transmission of energy in a pure standing wave.

and

Comparing

(6)

magnetic

zero,

is

are A/4 apart.

(8),

and

we

see that the electric energy density

is

maximum when the

maximum

Furthermore, the points where they are

vice versa.

In other words, the electric and magnetic energy densities of a pure

This condition is typical of a pure

standing wave are in space and time quadrature.

The energy

resonator (see Sec. 13-25).

form to the magnetic.

EneTgy in

oscillates

back and forth from the electric

spoken of as reactive or

this condition is often

It is not transmitted but circulates from one form to the other.

Simultaneously with the change from the electric to the magnetic form of energy
there is a space motion of the energy back and forth over a distance of A/4. These

stored energy.

relations are

shown

graphically in Fig. 10-14.

at three instants of time,

0,

T/8,

and T/4.

Here the energy densities are shown

The dashed curves show the instan-

and the solid curves show the

instantaneous magnetic energy density w m as evaluated from (8).
Finally, let us find an expression for the magnitude of the Poynting vector of a
pure standing wave. To do this, we substitute (7) and (10-10-7) in (10-12-13),

taneous

electric

energy density

we as

evaluated from

(6),

obtaining

Sx = -4

cos

cot sin cot

cos fix sin

fix

(9)

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

sec. 10-14]

Putting

in

terms of E gives

Sx = 4 /- E

cos

and the peak value of the Poynting vector

Peak

From an
At
left

inspection of (10)

this instant the position

as

vector

shown by the arrow

maxima and

electric

it is

cos

cor sin a>t

fix sin fix

(10)

is

>->*

(11)

S,

Sx is a maximum at cot = rc/4 ( period).

maximum is at fix = n/4 (A/8) and is directed to the

clear that

of one

The other arrows

in Fig. 10-14.

illustrate that at r

T/8 the energy

indicate other Poynting

flowing from the regions of

is

energy density to those of magnetic energy density.

Referring to Fig. 10-1

does not

move

or localize.

we note that in a standing wave, a constant-phase point

As a result, the energy tends to bunch up,

1,

with uniform velocity.

This condition becomes extreme with a pure standing wave for which a

constant-phase point
associated with

10-14

399

is

stationary.

a nonuniform or

Hence, localized concentrations of energy are

stationary phase velocity.

CONDUCTORS AND DIELECTRICS

In the preceding sections of this chapter the discussion was confined to waves in non-

conducting or lossless media.

Assume

Let us

that a plane traveling

normal incidence, as shown in

wave

now

consider the case of conducting media.

strikes the

boundary of a conducting medium

at

A portion of the incident energy is reflected,

Fig. 10-15.

-: ';'.v.
:

Conducting mediumv'./-;/

nFy
Incident

wave

Reflected

FIGURE

iI

wave

; '-:*;

"''V

'

10-15
;

Plane wave entering conducting

medium

at

normal incidence.

-;:.

Boundary of
con ducting medium

Wave

jf

:_- :,-...

;;

transmitted into

conducting

'-.

medium

':"-'"
'

;-

\Vv-/

'"''-:

.;.

;,

400

ELECTROMAGNETICS

[CHAP. 10

while the remainder enters the conducting medium.

Let us disregard the reflected

wave and focus our attention on the transmitted wave.

According to Maxwell's

first

curl equation,

H=J+^
ot

V x
In nonconducting media J
Recalling that

J =

ctE,

we

(1)

but in conducting media J

0,

may

not be negligible.

becomes

see that (1)

H = <tE +

<3D

V x

(2)

ot

For a

linearly polarized plane

wave

traveling in the

direction with

in the

direction, the vector equation (2) reduces to the following scalar equation involving

the field

components

Ey and

--dx-This equation has two terms in

that

is,

Ey = Eoe

10

Ey

Assuming

dx
in (4) each

(3)

-o7

that

Ey is a harmonic function

= oEy +jweEy

(4)

have the dimensions of current density, which

amperes per square meter.

The term oEy

space rate of change of

current densities.

equals the

When

Thus, according to

If the conductivity is zero, the conduction-current

(1)

we

^>

(2)

a>

~a

(3)

a> <?

the

>

= a,

is

term vanishes

not equal to zero,

the conduction current,

If

0,

the

medium

is

would be modified in the case of a lossy dielectric

~ a', but for condition (3) we have cue' X cue and
and hence we can write coe< a, as indicated.

t Referring to Sec. 8-15, condition (1)

a' and condition (2) to <oe'
to we'
a'

If

is much greater than

medium behaves like a dielectric.

the displacement current

(1),

expressed in

sum of the conduction- and displacement-

and we have the condition considered in previous sections.

one may arbitrarily define three conditions as follows :f

as in condition

is

represents the conduction-current density,

while the teTmjcoeEy represents the displacement-current density.

(4) the

of time,

becomes

", (3)

dHz
The terms

aE> +

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-14]

a perfect, or

For a not equal to zero the medium

lossless, dielectric.

However,

imperfect, dielectric.

if coe

> a,

much

is

a lossy, or

behaves more like a dielectric than any-

it

thing else and may, for practical purposes, be classified as a dielectric.

hand, when the conduction current

is

401

On

the other

greater than the displacement current,

(3), the medium may be classified as a conductor.

Under conditions
midway between these two, when the conduction current is of the same order of magnitude as the displacement current, the medium may be classified as a quasi conductor.
We can be even more specific and arbitrarily classify media as belonging to one

as in condition

of three types according to the value of the ratio a I coe as follows

< ji-

Dielectrics

coe

pL < < 100

Quasi conductors

coe

Conductors:

100

<
coe

where a
e
co

The

-1

= conductivity of medium, Urn

= permittivity of medium, Fm"
= radian frequency =2nf, where /is
1

ratio crfcoe
It is

medium

is

the frequency,

Hz

dimensionless.

to be noted that frequency

acts like a dielectric or

is

an important factor

in determining

ground (Ohio) for which er = 14 (at

Assuming no change in these values as a function of frequency, the
rural

different frequencies

is

as tabulated

Hz

At

of 30

kHz

rural

10' (A

ground behaves

GHz it acts like a dielectric.

In Fig. 10-16 the ratio aI coe

common

media.

1.3

x 10 4
1.3

=10 mm)

like

At
is

Ratio a/aie

10 3
10 7

ratio a/coe at three

Frequency,

whether

For example, take the case of average

_1
low frequencies) and a = 10~ 2 U

a conductor.

4.3x10"*

a conductor, while
1

at the

microwave frequency

MHz its behavior is that of a quasi conductor.

plotted as a function of frequency for a

number of

In preparing Fig. 10-16 the constants were assumed to maintain

their low-frequency values at all frequencies.

The curves in

Fig. 10-16 should therefore

not be regarded as accurate above the microwave region since the constants of media

may vary with frequency,

particularly at frequencies of the order of

GHz and higher.

402

[CHAP. 10

ELECTROMAGNETICS

N=l

Frequency

FIGURE
Ratio

list

10

11

12

13

14

15

17

16

18

=10" Hz

10-16

<r/<o

as a function of frequency for

some common media

of the low-frequency constants for the media of Fig. 10-16

is

(log-log plot).

presented in

Table 10-3.
Referring to Fig. 10-16,

we note

quencies far above the microwave

that copper behaves like a conductor at fre-

region.

at frequencies above about 10

On the other hand, fresh

MHz. The a/coe ratios for

water acts like

seawater, rural

a dielectric
ground, and urban ground are between the extremes of copper and fresh water.

Table 10-3

TABLE OF CONSTANTS FOR SOME

COMMON MEDIA
Medium
Copper
Seawater
Rural ground (Ohio)

Urban ground
Fresh water

Relative permittivity
dimensionless

e,

Conductivity a,

Urn-

5.8

80
14

80

x 10 7

10

~2

10"*
10

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-15]

WAVE EQUATION FOR CONDUCTING MEDIA

10-15

An

electromagnetic wave

in a

is

rapidly attenuated in a conducting

good conductor the attenuation

wave penetrates the conductor only

This depth of the penetration
it,

let

403

us

From

first

medium.

In fact,

so rapid that at high radio frequencies the

is

to a very small depth.

is

To

a matter of considerable interest.

calculate

Ey for a plane wave in a conducting medium.

develop a wave equation in

a solution of this equation, an expression for the depth of penetration

is

then

obtained.

in the

From Maxwell's curl equations we have

x direction with E in the y direction

for a linearly polarized

dHt

and

wave

BEy

(2)

&T--"-*

Differentiating (1) with respect to

and

(2)

with respect to x, we obtain

S_dH _ _ SEy_ 8%

--^=--^-i-d
dt
dt
dt~8x

(3)

1
fi

Since the order of differentiation

is

traveling

d%
ox

dH z

ox

dt

immaterial, the

left

side of (3)

is

equal to the right

side of (4) so that

d%_ a& =
2-^--^

\3 2 Ey
2
H dx

This

is

the

dt

(5)

dt

wave equation in , for a plane wave in a conducting medium. It is more

wave equation (10-2-12), developed for the case of a nonconducting
Whereas (10-2-12) has two terms, (5) has three, the third one involving

general than the

medium.

the conductivity.

Assuming harmonic variation of Ey with respect to

Ey = E

et"

t,

we may

write

(6)

404 ELECTROMAGNETICS

Taking the

first

[CHAP. 10

and second

derivatives of (6) with respect to

and

substituting these

values in (5) gives

1

- -T-rV +

a>

u ox

which,

(7)

(8)

upon rearranging terms, becomes

-^
dx
2

Let

Then

eE v -j(OffE

(8)

(Jcofia

2
(o fi)Ey

=jwpo

2
a> jte

(9)

reduces to

.-'*"
This equation

is

a simplified form of

(5).

The time

harmonic variation with time being assumed, y

is

(10)

does not appear

explicitly,

called the propagation constant.

A solution of (10) for a wave traveling in the positive x direction is

Ey = E
For conductors,

<r

>

coe,

e- x

(11)

so that (9) reduces to

2

fajcona

andf

(12)

(13)

Thus, y has a real and imaginary part. Putting y = a + jp, we see that a, the real part,
is associated with attenuation and ft, the imaginary part, is associated with phase.
Substituting the value of y

E,

from

= E exp[-(l + ,-)^*]

(1 3) in (1 1) gives

=E exp(-^*)ex P (-;^*)

t Note:

-_

V2

=1/45

(M)

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-16]

In (14) the attenuation factor

is

405

given by

jcona

and the phase factor by

x
p\J
-A/-?-*
where

< 16 >

w = radian frequency =27t/, rad s"

-1
/; = permeability of medium, H m
a = conductivity of medium, Om"
x = distance, m
1

complex operator, dimensionless

Equation (14)
positive

is

a solution of the wave equation for a plane wave traveling in the

direction in a conducting

medium.
x.
The

magnitude and phase as a function of

It gives

field

the variation of

Ey

in

attenuates exponentially

both

and

is

retarded linearly in phase with increasing x.

10-16

DEPTH OF PENETRATION

Continuing the discussion of the preceding section,

let

us

now

obtain a quantitative

measure of the penetration of a wave in a conducting medium.

Referring to Fig.

wave that penetrates the conducting medium, i.e., the transmitted

wave. Let x =
at the boundary of the conducting medium, so that x increases
positively into the conducting medium.
Let (10-15-14) be written in the form
10-15, consider the

E,

where 5
surface

At a

=E

= y/2/oifia. At x = 0, Ey =
on the conducting medium.

distance

e-

x/i

Now

e- Kx/i)

This
5 in

is

(1)

the amplitude of the field at the

(1)

has the dimension of distance.

= 8 the amplitude of the field is

\Ey \=E e- i

=Eo\

(2)

406 ELECTROMAGNETICS

Thus,

Ey

decreases to 1/e (36.8 percent) of

to a distance

Hence 3

&.

As an example,
wave
1/e

[CHAP. 10

is

its initial

called the 1/e depth

wave penetrates

value, while the

ofpenetration.

consider the depth of penetration of a plane electromagnetic

incident normally

on a good conductor, such

as copper.

Since co

= 2nf,

the

depth becomes

(3)

For copper

/tr

1,

so that m

Putting these values in

(3),

1.26 jiH

we obtain

'.

The conductivity a

58

MU m

1
.

for copper

-^
where d

1/e

depth of penetration,

/= frequency,

Hz

Evaluating (4) at specific frequencies,

At 60 Hz:

At

c5

MHz:

<5

At 30 GHz:

<5

Thus, while at 60

Hz the

1/e

we

find that

= 8.5xl(r 3 m
= 6.6xKT 5 m
= 3.8xl0- 7 m

depth of penetration

is

8.5

mm, the penetration decreases

in inverse proportion to the square root of the frequency.

(30
is

GHz)

the penetration

often called skin

is

only 0.00038

mm,

At 10

or less than $ /an.

mm

This

wavelength

phenomenon

effect.

Thus, a high-frequency

field is

damped

out as

it

penetrates a conductor in a

shorter distance than a low-frequency field.f

we can speak of other depths for

an arbitrary fraction of its original value. For
which the field is 0.01 (1 percent) of its original value.

In addition to the 1/e depth of penetration,

which the

electric field decreases to

example, consider the depth at

This depth

is

obtained by multiplying the 1/e depth by 4.6 and

may

be called the

percent depth of penetration.

is analogous to the way in which a rapid temperature variation at the surface of a

thermal conductor penetrates a shorter distance than a slow temperature variation.

t This

407

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

1(M6 ]

sef

Phase velocity

is

given by the ratio co/0.

phase velocity in the conductor

is

fi

l/<5

so that the

is

vc

Since the \/e depth

In the present case,

= co5=,
\ op

(*)

small, the phase velocity in conductors

small.

is

apparent

It is

from (5) that the velocity a function of the frequency and hence of the wavelength.
Hence, conductors
In this case, dv/dX is negative, where X is the free-space wavelength.
is

media

are anomalously dispersive

is

(Sec. 10-7).

The ratio of the velocity of a wave in free space to that in the conducting medium
the index
the index of refraction for the conducting medium. At low frequencies

for conductors

To

is

very large.

find the wavelength X r in the conductor,

Xc
In

(6),

we have from

(6)

the wavelength in conductors

index for a
It is

medium of copper

in the

small for conductors,

is

small.

is

Values of the l/e depth,

Hence the wavelength

Since the l/e depth

times the l/e depth.

is 2tc

or

cod,

2nd

both Xc and 5 are in the same units of length.

conductor

(5) that /A,.

percent depth, wavelength, velocity, and refractive

are given in Table 10-4 for three frequencies.

interesting to note that the electric field

damped to

is

percent of

its initial

about \X in the metal.

penetration depth is inversely proportional to the square root of the
the
Since
thin
sheet of conducting material can act as a low-pass filter for electroa
frequency,
amplitude in

magnetic waves.

Table 10-4

PENETRATION DEPTHS, WAVELENGTH, VELOCITY,

AND REFRACTIVE INDEX FOR COPPER
Frequency
60

mm

Wavelength in free space A (m)

l/e depth,

8.5

percent depth,

Hz

3.9

Wavelength in conductor A*,

s" 1
Velocity in conductor ve ,
Index of refraction, dimensionless

5.3

"

x 10 3
x 10-*
x 10~ 2
x 10

300
6.6

3
4.1

4.1

3.2

9.5

MHz

7.3

x 10 - 5
x
x
x
x

10"*
10"*
10
10 J

30

GHz

10

mm

x 10 - 1
x \Q~*
2.4 x 10"
7.1 x 10*
3
4.2 x 10
3.8

1.7

408

ELECTROMAGNETICS

r
,

RELAXATION TIME

10-17

Thus far, the behavior of electromagnetic waves in conducting media has

been discussed from the standpoint of depth of penetration, velocity,
and so forth. It is
instructive to consider the problem from another point of
view, namely, from that of
the behavior of a charge configuration placed in the conducting
medium.

Consider a conducting
arbitrary shape

and

density.

mutual repulsion of the like

conducting medium. Now

medium of infinite

extent, in which is placed a charge of

Imagine that the charge is released and, because of the
charges of which it is composed, spreads out through the

let us determine how long it takes for such a charge

of
density p to decrease in density to l/e of its original value.

According to

(4-13-3), the continuity relation

between current density and charge

density is

(1)

dt

From

Maxwell's equation

D = p, and from D = eE
(2)

But J

<rE,

so that (2) becomes

V-J=^
From

(1)

and

(3)

it

(3)

follows that

dp

+ -p =

dt

(4)

A solution of this equation is

p
as

may

be readily

stituting in (4).

Tr

verified

= p e-^

by taking the

first

lc) '

(5)

derivative with respect to time

and sub-

Let us put

has the dimension of time.

At

= 0, p = Po

which

Whenr=T

is

the

initial

charge density

= Poe _, =Po~

(7)

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-18]

409

Tr is the time required for the charge density to decrease to 1/e of its initial value.
The quantity Tr is called the relaxation time.
In a lossless dielectric a = 0, so that Tr is infinite. Hence the charge maintains
On the other hand, for a conductor such as copper,
its original density indefinitely.
=
58 MU m" and e = 8.85 pF m~ we find that
for which a

Thus,

'

=1.5 x 10" 19

Tr

This very short interval corresponds to the length of one period for x-rays. Thus, at
radio and microwave frequencies the relaxation time is much less than the period.

The

result of the short relaxation time is that the

conductor cannot maintain a charge

configuration long enough to permit propagation of a wave more than a very short
distance into the conductor.

10
is

19

Hz, the relaxation time

frequency is sufficiently high, of the order of

about the same length as the period and propagation

When the
is

possible.

IMPEDANCE OF CONDUCTING MEDIA

10-18

The behavior of conducting media toward plane electromagnetic waves can be confrom yet another point of view, the standpoint of impedance.
A solution of the wave equation for the electric field of a linearly polarized plane
wave traveling in the x direction in a conducting medium with E in the y direction is
of the form

sidered

y = E
while for the magnetic field a solution

eJ

"- yx

(I)

is

H = H eH"--'x
the lag in time phase of H with respect to

(2)

where
to ti z

is

or the lead of

ty with respect

Taking the

medium.

ratio of (1) to (2) yields the intrinsic

impedance

Zc of the conducting

Thus

According to

(3),

the magnitude or

modulus of the intrinsic impedance is equal to

field, and the phase angle of the impedance is

the ratio of the electric to the magnetic

equal to

^t
t

The dot on
It is

2C

indicates explicitly that the

impedance

a complex function of the phase angle

is

also a

complex quantity.

410 ELECTROMAGNETICS

To

[CHAP. 10

Zc in terms of the constants of the conducting medium, we proceed

Maxwell's equation from Faraday's law for a plane wave with components

evaluate

as follows.

Ey and 6, is
dEy

&--* dti
IT

Taking the x derivative of

(1)

and the

derivative of (2)

(4)

and substituting

in (4) yields

yy =j{icoHz
The

intrinsic

impedance

2C

is

then given by

^ _ &, _

When we

recall

(6)

J<op
(6)

from (10-15-13) that for a good conductor

Eq.

(5)

(Ofia

+j)

becomes

l+ j
2 = if =

//"

^2 V a

(7)
<r&

Since
!

^-l/45-

the intrinsic impedance can also be expressed as

.-Z.lt-J?2/4F
It

follows that the magnitude of the intrinsic impedance of a conductor

(8)

is

given by

and the phase angle by

= 45

(10)

Whereas the intrinsic impedance of a perfect dielectric medium is a pure resistance (E and fi1 in time phase), we note from (8) that the intrinsic impedance of a
y
conductor is a complex quantity,
z lagging Ey in time phase by very nearly 45.

PLANE WAVES IN DIELECnUC AND CONDUCTING MEDIA

SEC. 10-18]

411

inductance in
analogous to the situation in a circuit having resistance and
medium
conducting
the
Therefore
voltage.
applied
series where the current lags the
lc
writing
explicitly
by
expressed
can
be
behaves like an inductive impedance. This

This

is

in terms of

The

its resistive,

intrinsic

or

real, part

impedance

R and its reactive, or imaginary, part

may

Multiplying and dividing in (7) by

e,

But

and nr for the

That

is,

be expressed in yet another form, as follows.

e

and n

2 = 12
Substituting e r

X.

j^

ratios in (12),

gives

SJL

(12)

we have

vW^o s the intrinsic resistance of free space,

which equals 376.7

^"W^?

fi,

(n)

so that

<14)

= relative permeability of medium, dimensionless

= relative permittivity of medium, dimensionless
_1
a = radian frequency = 2nf, rad s
-1
= permittivity of medium, F m
_1
a = conductivity of medium, U m
The magnitude or modulus of 2C is

where

fir

er

ZC =|Z C =
|

The

ratio <we/<7 or its reciprocal

tors

ox/a is

of 3

GHz

in Sec. 10-14.

mm

=ii/x

Z c = ll x

V2

(15)

For good conduc-

Take, for instance, copper. At a frequency

-9
and
as unity, we

Taking
mm), ax/o = 2.9 x 10
to be
impedance of copper at 100

(wavelength 100

or

was discussed

very small (ajax very large).

find the intrinsic

376.7

376.7 x

0.02

5.4x10" 5

= 0.02/45 CI

// r

er

412 ELECTROMAGNETICS

The magnitude of the

[CHAP. 10

is

much

less

impedance of copper is Zc = \ZC = 0.02 Q. These

a conducting medium such as copper the ratio of E to
z
y

intrinsic

results indicate that for

than for free space.

The small value of

vanishes.

If

Zc

(perfect conductor),

is infinite

Zc =

for copper suggests that the conducting

behaves like a short circuit to the electromagnetic

and

medium

field.

In the above discussion on the impedance of conducting media, we introduced

a value of y into (6) based on the assumption that <x> we. However, in the case of

dielectric

(6) gives

medium we >

a,

and

y=jP

Substituting this value of y into

=jco^/fie.

the intrinsic impedance for a dielectric

medium

equal to

y/fi/e.

Some

of the important relations for conductors developed in this and preceding

sections are summarized in Table 10-5.

Table 10-5

PENETRATION DEPTHS, VELOCITY,

WAVELENGTH, INDEX OF REFRACTION,
IMPEDANCE, AND RELAXATION TIME IN

CONDUCTING MEDIA

- depth
e

= 8 = -7== =
Virfu.0

percent depth

= 4.6 S

Phase velocity = vc

(m)
V*<*

On)

= coS =

/2oj
/

(m

a/'7'*
Wavelength

= XC = lirh

Index of refraction

Relaxation time

-q

t where

a>

= = g =

=T, =

(dimensionless)

(2)

= 2vf= radian frequency, rad s _1

/i

.26 /*H

present

_1

(s)

/= frequency, Hz
= permeability X

(m)

= Z = / /45
'-t-J?

Impedance

(<7> <ue)t

= permittivity, F m _1
= conductivity, m -1
rjf

m _1 if no ferromagnetic material is

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

sec. 10-19]

10-19

THE POYNTING VECTOR IN CONDUCTING MEDIA

Assume

that a plane

wave

is

traveling in a uniform conducting

the value of the Poynting vector

polarized and traveling in the positive

in the z direction, as in Fig. 10-17.
at

medium.

Let us find

Suppose that the wave

for the wave.

direction with

in the

is

direction

Let the boundary of the conducting

413

linearly

and

medium be

= 0, and let

4 = o^w
6 =H

and

From

'

(1)

- i)
J( '0,
(2)

(10-12-18) the scalar magnitude of the average Poynting vector

Sav = Re S = iRe .y if* = }

Say = }E

or

where

cost

E = amplitude of electric field, V m -1

H = amplitude of magnetic field, Am -1
= phase difference between electric and
,

is

Re e*

(3)

(Wm- 2)

magnetic

(4)

fields,

rad or deg

Conducting

medium

Wave
direction

Hz

FIGURE

10-17

Field components of wave in conducting

medium.

x=0

[CHAP. 10

414 ELECTROMAGNETICS

The Poynting vector

is

entirely in the

values, the factor \ in (4)

is

omitted.

are rms
x direction in this case. If E and
For a conducting medium, t, is very nearly 45.

Returning to a further consideration of the average Poynting vector, we see that the
intrinsic impedance t c is equal to /ft z so that (3) can be expressed
y

S = iReti,a?ic =

i\fl,\

ReZc = iH 2 Rec

(Wm" 2 )

(5)

a very useful relation since if the intrinsic impedance Zc of a conducting

at the surface are known, it gives the average
medium and also the magnetic field
Poynting vector (or average power per unit area) into the conducting medium.
This

is

EXAMPLE A plane 1-GHz traveling wave in air with peak electric field intensity of
_1
Find the average power absorbed
is incident normally on a large copper sheet.
1 V m
by the sheet per square meter of area.
solution

First let the intrinsic

impedance of copper be calculated

at

GHz.

From

(10-18-14)
X
i = - 376.7 /* /^
c

For copper

/*,

= e, = 1
Re

and a

= 58 MO m"

H for the incident wave.

real part of

the average

C is

8.2

at the sheet (tangent to the surface).

=2

2=

2x1
376?7

Am

"

mrt
This

is

very nearly

power per square meter into the sheet

S.,

Thus
/f<'

From (5)

Hence the

^ = (c0 s45)(376.7)y^^||lP^ =

Next we find the value of

double

*.

=H^y

8.2

x 10-'

is

= 116 nW m- 1

The power from a wave absorbed by a conducting medium may

also

be con-

medium. Since J = oE,

manner
as the electric field E.
in
the
same
varies
medium
in^he
density
J
current
the
Thus for a plane wave incident normally on a conducting medium, as in Fig. 10-17,

veniently regarded in terms of the current induced in the

the current-density variation

is

expressed by

J-jHe-'lU-M*

(Am" 2 )

(6)

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

sec. 10-19]

rJo

FIGURE

10-18

Relative magnitude of electric field

penetration 8 for a plane

The abscissa gives

The wavelength in

where J

is

415

wave

E or current density J as a function of depth of

traveling in

the penetration distance

x direction
x and

the current density at the surface of the

10-18.

infinitely far in the positive

medium.
(8).

the conductor equals 2w8.

medium

(x

of the magnitude of J (or E) as a function of the distance x

ponential curve in Fig.

into conducting

expressed in l/e depths

is

= 0). The variation

portrayed by the exAssuming that the conducting medium extends

x direction, the total

is

current per unit width (in z direction)

induced in the conducting medium by the wave is given by the integral of the magnitude
of J from the surface of the

K=

medium

(x

That

0) to infinity.

C\J\ dx = J fe-*ls dx = J

is,

(7)

K= current per unit width (sheet-current density), Am -1

where

= current density at surface, Am -2

5 = l/e depth of penetration, m

Thus the area under the exponential curve

step curve.

be obtained

It
if

is

equal to the area under the

K per unit width

is

J maintained the constant amplitude J from the

and were zero elsewhere.

tion 5.

in Fig. 10-18

follows that the total current

equal to what would

surface to a depth 5

This gives added significance to the l/e depth of penetra-

[CHAP. 10

ELECTROMAGNETICS

416

The average power absorbed per

unit area of the conducting

medium

is,

from

(5),

Sav

or since

= K and Re Zc =

= i^O

R- e

R,

(WnT 2 )

S = i**R
where

(8)

A:

(9)f

= current per unit width, as in (7), A m _1

R = (ad = resistance of square sheet of conducting medium

K= J

of thickness

5, SI

per square

The

resistance

in (9) is

Fig. 12-3,

sometimes referred to as the skin

resistance, since at

be confined to a very thin layer. Referring to

the resistance of a square sheet of the medium of thickness 5 = h as

high frequencies the current

is

may

measured between two opposite edges.

10-20
In

field

CIRCUIT APPLICATION OF THE POYNTING VECTOR

theory

we deal with

point functions such as E, H, and S.

the electric and magnetic fields at a point and

dealing with waves in space

it is

hand, in dealing with circuits

quantities such as V, I,

and P.

equal to the line integral of

it is

is

more convenient

usually

That

is,

E between

is

E and H give

the power density at a point.

In

On the other

convenient to use such point functions.

to employ integrated

the voltage between two points and

is

the points, or

-J.
The quantity J

Thus,

E-rfl

the current through a conductor, which

(1)

is

equal to the integral of

around the conductor or

= j> H

'

The quantity

is,

for example, the

power

d\

dissipated in a load

(2)

and

is

equal to the

integral of the Poynting vector over a surface enclosing the load, or

'

= j S-ds

Note the similarity between (9) and the circuit relation for the power dissipated
2
impedance with a real part (resistance) R as given by P = i/ /f (W).

(3)

in

an

sec. 10-20]

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

El

\\

417

E
Power flow

*-C

Power

( power flow

Iff s

flow_

B+

_|_^
s

Iff

S
-Infinite

imaginary plane

(a)

-Infinite perfectly

conducting sheet
(6)

FIGURE
Simple

10-19

circuit

showing Poynting vector or power flow

to resistor (load) (a) without

and

(b)

lines

from battery (source)

with intervening conducting sheet,

Consider as an illustration the simple circuit of Fig. 10-19a, consisting of a

battery connected by a conductor with current / to a load of resistance R.

emf of the

circuit is confined to the battery

of the conductor

All the

and all resistance to the load (conductivity

is infinite).

an electric field E extending between the conductors (J E dl = V,

where E is integrated from the upper to the lower conductor) and a magnetic field
dl =
around the conductors (j
At any point the
/), as suggested in Fig. 10-19.
There

exists

(in watts per square meter) is given by the cross product of E and
The direction of the Poynting vector is suggested at several points
in Fig. 10-19a.
Around the battery (power source) the Poynting vector is outward
(positive).
Around the resistor (load) the Poynting vector is inward (negative).

Poynting vector

H (S = E x H).

418

[CHAP. 10

ELECTROMAGNETICS

Over an imaginary plane through the middle of the circuit the Poynting vector is to
the right (from battery to load). The total power leaving the battery or entering the
load

is

given by the integral of the Poynting vector over a surface enclosing the battery
a surface enclosing the load (negative power), or over an infinite

(positive power), over

plane dividing the circuit as suggested.

The

power

total

into the load

given by

is

P=j

(E x H)

ds

(4)

over a surface enclosing the load, such as the cylindrical

are everywhere normal to each
surface shown in Fig. 10- 19a. The fields E and
They are also both tangent to the cylindrical surface enclosing the resistor
other.

where the integration

is

or load.

Thus, from the geometry in this case

V=JE
J = $ H

(5)

P=-VI

(6)

= voltage between ends of the

d\ = current through load, A

d\

reduces to

P=-JE-dlJH-dl
or

where

(4)

cylinder (or load),

Thus, integrating the Poynting vector over the resistorf yields the same power VI
However, from the field point of view the power entering
as given by circuit theory.
the load

is

negative

power

since

is

inward, or negative, over the surface enclosing

the load.
In Fig. 10-19aflow lines of the Poynting vector (power flow lines) are shown.
is

evident that the

power flow

is

through the empty space surrounding the

the conductors of the circuit acting as guiding elements.

view

we

usually think of the

simplification

From

the circuit point of

power as flowing through the wires but

and does not represent the actual

It

circuit,

this

is

an over-

situation.

Suppose now that a perfectly conducting infinite sheet is placed across the middle
The sheet is continuous except for two small openings
through which the circuit conductors pass. The power flow through the sheet is
zero (E in sheet is zero), and all power flow from the battery to the load is through the
of the circuit, as in Fig. 10-196.

two holes containing the circuit conductors. However, E is very large in these
openings, and S is also large. The integral of S over these openings again equals the
The flow lines of the Poynting
total power flowing from ,the battery to the load.
vector (power flow lines) for this case are as suggested in Fig. 10-19Z>.
t

We have neglected the contribution of power flowing through the flat end surfaces of
the cylinder. However, this can be
very small compared to its length.

made

negligibly small if the cylinder diameter

is

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

SEC. 10-21]

Consider

now

and b is replaced by an alternator

The average (or total power) is now given

that the battery in Fig. 10-19a

whose emf varies harmonically with time.

by

419

P =

ReS-rfs

<f

i(f

Re(ExH*)-(/s

= iRejE-dl|H*-dl

(7)

Integrating over the alternator givesf

Pav = J Re

V=

where

VI*

=iV

Re e = iV

h cos {

(8)

V ^at

= / -*-
V amplitude or peak value of V
I = amplitude or peak value of J
i = time-phase angle between current and
cos = power factor
/

Integrating (7) over the load,

we

voltage

obtain

^.V=-Wocos
which

is

equal in magnitude but opposite in sign to

Since the load impedance

Z=

V/I,

we

(9)

(8).

also have (omitting the negative sign)

that

P = i/
R

where

is

ReZ = iV*

the resistance of the load (real part of

Z)

in

(10)

ohms.

GENERAL DEVELOPMENT OF THE WAVE EQUATION

10-21

In this chapter we have dealt with plane waves (traveling in the x direction).

The wave

equation was developed for this special case, and appropriate solutions were obtained.

A more
shown

general development of the

that for a plane

wave

it

wave equation

will

now be

given,

and

it

will

be

reduces to the expressions obtained previously.

Maxwell's curl equations are

dE
8D
= .E + VxH = J + _

and

xE=-f=-f

t E, H*, V, and /* are complex quantities (phasors).

ted.

(1)

(2)

For

simplicity the dot is omit-

420

Taking the curl of

But by a vector
of

[CHAP. 10

ELECTROMAGNETICS

(3)

(2)

identity,

H from (1) gives

Vx

and introducing the value of

meaningful only in rectangular coordinates, the left-hand side

can be expressed

Vx
Equating

(4)

and

(3),

(V x E)

and noting that

(4)

we can

no

(4)

free charge

V E = 0, we

get

5E
(5)

variation of the field with time, (5) reduces to

V+

V2E -

From

- V2E

+ ^-^
^^3F "*W

V 2 E = ( - co
or

E)

in space having

V2 E =
With the assumption of harmonic

= V(V

jmiiffjE

(6)

E=

(7)

also write

VxVxE + y

E=

wave equations in E. In (5) the time is explicit, while

harmonic variation with time being assumed. These
wave equations incorporate all four of Maxwell's equations. Maxwell's two curl
equations are the starting point for the wave equation, and (5) to (8) satisfy them.
Equations

(5) to (8) are vector

in the other three

it is

implicit,

Maxwell's two divergence equations are also

satisfied.

For a plane wave traveling in the x direction with

Eq.

(5)

E in the

v direction (E

^^-M*
d

which

= yy)

reduces to

is

the

same

as (10-15-5).

Ey

is the same
wave equations.

If

(9)

If<r=0,
d

which

dEy

a E

Ey

as obtained for a lossless

does not change with time (E

Ey

medium

in (10-2-12).

These are scalar

static),

V E=
2

(11)

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

PROBLEMS

If

E is a harmonic function of time, (9) becomes

2
E
-TT~

-<o 2 neEy +ja>noE,

y?-y

or

which

421

is

the

same as

(12)

Ey =

(13)

(10-15-10).

PROBLEMS
A plane traveling wave has a peak electric field 0 =6 V m _1 If the medium
= 1 and e, = 3, find (a) velocity of wave, (b) peak Poynting vector, (c) average
is lossless with

10-1

/i,

Poynting vector,

10-2

medium

If the
(c)

(</)

impedance of medium, and

A plane traveling
is

lossless with

(e)

peak value of the magnetic

100-MHz wave has an average Poynting

=2

/x P

and

e,

= 2,

field

vector of 5

H.

W m~

2
.

find (a) velocity of wave, (b) wavelength,

electric field E, and (e) rms magnetic field H.

Does the term perfect vacuum (or free space) imply the absence of everything?
10-4 Fora 10-MHz traveling wave with E =6Vm"' find (a) average Poynting vector,
peak energy density, and (c) mass equivalent per unit volume. The medium is free space.
10-5 Show that for a traveling wave the peak energy density stored in the electric

impedance of medium, (d) rms

10-3

*
(4)

field

equals that stored in the magnetic

From Maxwell's

field.

wave equation in
for a plane
wave traveling in the positive x direction in a medium with constants ft = /x e = e and
a = 0. The electric field is in the y direction. (Jo) Obtain a solution for the case of harmonic
variation and prove that it is a solution.
*
10-7 A plane 3-GHz wave is incident normally on a large sheet of polystyrene
How thick must the sheet be to retard the wave in phase by 180 with respect
( p = 2.7).
to a wave which travels through a large hole in the sheet?
-2
*
10-8 The earth receives 2.0 gcal min -1 cm
of sunlight, (a) What is the Poynting
vector in watts per square meter? (b) What is the power output of the sun in sunlight assuming that the sun radiates isotropically? (c) What is the rms electric field E at the earth
10-6

(a)

curl equations derive the

assuming that the sunlight

at

is all

a single frequency ?

(d)

How long does it take the sunlight

= 14.3 g cal min" '.)

Take the earth-sun distance as 150 Gm. (1
1.4-GHz radio wave received from an extragalactic radio source has an

to reach the earth?

10-9

W m -2 Hz -1
m 2 over a bandwidth of MHz?

average Poynting vector per unit bandwidth (or flux density) of 10~ 26

How much power is intercepted by

an area of 1,000

normal to the wave direction,

area

is

and

radiates isotropically,

what

is

the

(b) If the radio

source

is

in

(a)

The

at a distance of 10 9 light-years

power radiated by the source

* Answers to starred problems are given

Appendix C.

in

1-MHz bandwidth?

422 ELECTTROMAGNETICS

One light-year equals

medium is free space,
ments for

[CHAP. 10

the distance light or radio waves travel in

(c)

year.

Assume

that the

How does this power compare with the earth's total power require-

electricity, heating, industry, transportation, etc. ?

3-GHz

10-10

What is

radio waves from the sun have a flux density of 10

-20

W m -2 Hz

'.

1-GHz bandwidth assuming the flux density is constant

over this bandwidth? (b) What is the rms electric field E assuming that the power in the
1-GHz bandwidth is at a single frequency? (c) What is the radio power output of the sun
for the 1-GHz bandwidth assuming that the sun radiates isotropically? Take earth-sun
{a)

the Poynting vector for a

Gm.

distance as 150

10-11

(a) Referring to

What is
(c)

Prob. 10-8, what

Take the

of sunlight at the earth?

sunlight

is

the Poynting vector per unit bandwidth

band as 400

to 800

nm

(4000 to 8000 A),

the ratio of this Poynting vector to that given in Prob. 10-10 for

(b)

3-GHz radio waves?

According to Planck's radiation law, the radiation brightness 38 from a blackbody is given by

= ^f^rr\

(Wm-'Hz-'rad-)

= Planck's constant = 6.63 x 10" 34 J s

/= frequency, Hz
"
c = velocity of light or radio waves = 300 Mm s
k = Boltzmann's constant = 1.38 x 10"" J K"
T = blackbody temperature, K
Taking T = 6 kK for the sun, would you say that the sun behaves as a blackbody (or

where h

Planck)

radiator over the above optical-radio range?

10-12

According to quantum theory, electromagnetic waves are transmitted by

3*
J s and f= frequency
hf, where h = Planck's constant = 6.63 x 10"

photons of energy
in hertz.

Find the number of photons per second which are intercepted by an area 1 m
-2
~ 2
for two cases
(a) frequency

Wm

square for a wave with average Poynting vector of 1

= 2 GHz and (b) wavelength = 600 nm.

The area is normal to the wave

direction,

(c)

advantage do radio waves have over optical waves for transmission of information

wave

is

so weak that the

10-13

The earth

all frequencies),

(a)

number of photons

receives about 1.5

What

is

is

What
if

the

small?

kW m~

of power from the sun (integrated over

the sun's total power output assuming

it

radiates isotropically?

Take the earth-sun distance as 150 Gm. (b) What is the total power received by the earth?
Take the earth's radius as 6.4 Mm. (c) If the sun's mass is 2 x 10 30 kg and its mass is
converted to radiant energy according to Einstein's relation (energy
efficiency,

how long can

10-14

the sun radiate at

electric field

E at the earth, (c) the time

is

at

2-GHz

waves.

percent

present level ?

its

spaceship at lunar distance from the earth transmits

power of 10

= mc 2 )

radiated isotropically, find (a) average Poynting vector at the earth, (b)

ship to the earth.

it

If

rms

takes for the radio waves to travel from the space-

(Take the earth-moon distance as 380

Mm.)

(d)

How many

photons

on the earth from the spaceship transmitter?

*
10-15 At 100 kHz a lossless medium has constants n, = 1 and e p = 81. These
the ratio of the impedance
constants are approximated by distilled water. Find (a) Z/Z
per unit area per second

fall

of the

medium

to that of free space, (b) A/A

the ratio of the wavelength in the

medium

to

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

that in free space, (c) v/v

and

the ratio of the

wave

velocity in the

medium

423

to that in free space,

(d) the index of refraction.

10-16 Compute and plot curves for the amplitude of the resultant electric field due
two plane waves traveling in the positive and negative x direction at three instants of
time / = 0, 778, and 7/4, where T = period. The waves have only E, components. The
1
wave in the positive x direction has an amplitude E = 1 V m" and the other wave an
amplitude ,=0.4Vm"'. Extend the plot over a distance of at least 1 A in the x direction,
to

(a)

In which direction does a constant-phase point

ratio ?

Assuming that the smaller wave

(c)

is

move?

(6)

What

is

the standing-wave

a reflection of the larger, what

the magnitude

is

of the reflection coefficient?

10-17

plane wave in free space

is

normal incidence from an infinite

The amplitude of E of the incident

reflected at

perfectly conducting sheet, producing a standing wave.

wave

What

is

is

mV m"

12

(a)

'.

How

far

from the sheet

is

maximum?
maximum value of

the Poynting vector a

the average value of the Poynting vector? (c)

What

is

the

(6)

the

Poynting vector?

10-18

A plane wave is reflected at normal incidence from a boundary surface. The

and of the reflected
field E of the incident wave is lOmVm"
Find the VSWR when (a) E, = 2 mV m" (b) fi = 5mV m~ \ and (c) E =
1

amplitude of the electric

wave

is

E,

mV m"

',

Under what conditions is there a pure standing wave?

10-19 A 50-mW 11-GHz transmitter radiates isotropically. At a distance of 5 km
*
find (a) rms electric field E, (A) rms magnetic field H, (c) average power per unit area (Poynting
2
vector), (rf) average energy density, and (c) energy contained in a volume of unit area (1 m )
perpendicular to the wave direction by 10A in the direction of propagation. The medium
8

'.

(d)

is air.

70-20
current

/.

A cylindrical resistor of length

(a)

Show

/,

that the Poynting vector

radius a,
is

and conductivity a

carries

an rms

directed normally inward with respect to

(b) Show that the integral of the Poynting vector over the surface
PR, where jR is the resistance of the resistor.

the surface of the resistor.

of

the resistor equals

10-21

disks of radius

which

parallel-plate capacitor

a and separation distance

d.

is

(a)

being charged consists of two

Show

flat

that the Poynting vector

normally inward with respect to a surface enclosing the capacitor.

2
2
of the Poynting vector over this surface equals idna e dE /dt.

(6)

Show that

is

circular

directed

the integral

The medium filling the

Under what circumstances would the integral of the

Neglect fringing, (c)

is air.
Poynting vector over the surface enclosing the capacitor be zero, even though the charge

capacitor

on the capacitor
10-22

not zero?

is

(a) Referring to part (a)

of Prob. 10-21, what circumstances are required for

the Poynting vector to be directed normally outward with respect to the surface bounding
the capacitor? (b) Referring to part (a) of Prob. 10-20, what circumstances are required
for the Poynting vector,

if

not zero, to be everywhere directed normally outward with respect

to the surface of the resistor?

10-23
is

Show

that in a standing

equal to the geometric

V'u m i.

mean of

wave the average phase velocity of the

maximum and minimum velocities;

the

resultant field
that

is,

v.,

424

ELECTROMAGNETICS

Show

10-24

[CHAP. 10

when the incident-wave amplitude E

that

Eu

reflected-wave amplitude

approximately E,

of two waves of

this type

E = E ReeJia, -*">.

E = 2E

becomes

+ j8,

/3

and

that the resultant

+ toj

a>

/Ji) may

/J

and

<oi

co

be expressed

fio x) cos(a>i? pix).

Show that u can be expressed also in the
velocity u = da>/dfZ.
~
u = v + fi (do Idp) v\ (dv/dX) = dfjd\
where v = phase velocity.

cos(<o

'

The group

10-26

following forms

Find the group velocity for a 100-MHz wave

10-27

medium

Show

of equal amplitude and of two frequencies

(and two corresponding phase constants given by

as

greater than the

E + 1 cos 2fix.

Given a plane wave for which

10-25

much

is

the standing-wave-envelope expression of (10-10-13)

which the phase velocity c

for

For a plane wave

10-28

in

= 2x

dielectric

10 7 A 2/3

m s"

in a normally dispersive, lossless

*.

medium show

be

that the energy velocity can

Z is the intrinsic impedance of the medium.

= 0.1 O m -1
= 1, e, =40. Assuming
constants

expressed as 1/eZ or Zip, where

10-29

medium has

<r

that

/*,

do not change with frequency, does the medium behave like a conductor or
a dielectric at (a) 50 KHz and (b) 10 GHz?
_1
10-30 A plane 1.59-GHz wave with rms electric field E = 3 V m
is traveling in
a medium with constants (r = 0.OlUm" 1 ,^, = , = l. Find the rms values of (a) conductionthese values

current density, (b) displacement-current density, and

is

A medium

10-31

not change with frequency, find the 1/e and

2

(A)

MHz,

(c) 3

ZjZ

(d)

What

Um

do

percent depths of penetration at (a) 60 Hz,

GHz.

At 500

10-32

(c) total-current density,

unit area in the direction of the wave propagation ?

_1
has constants a
10 2
3.
If the constants
2, <-,
, p,

power flow per

the average

MHz

medium has

constants

the ratio of the intrinsic impedance of the

medium

\x,

= 5,

medium

eP

= 7,

= Urn -1

Find

to that of free space, (b) A/A

(a)

the

wave
wave attenuation in decibels for a 5-mm thickness, and (/) reflection coefficient p for a wave in air
incident normally on a large flat surface of a half-space filled with the medium.
10-33 At 200 MHz a solid ferrite-titanate medium has complex constants ft, =
*
15 (1 -y"3) and e r = 50(1 yl). Find (a)Z/Z (*) A/A (c) v/vo (d) 1/e depth, (e) attenuation
for a 5-mm thickness, and (/) p, where the symbols and circumstances are as explained in
ratio of the wavelength in the

velocity in the

medium

to that in free space, (c) v/v

the ratio of the

to that in free space, (d) 1/e depth of penetration, (e)

Prob. 10-32.
10-34

(a)

A medium

of large extent with constants <r=10" 3

terminated by an infinitely conducting sheet situated at

the
1

medium

is

incident normally

on

Vm"'at the sheet, draw a graph

wave,

(b) reflected

10-35

20

dB and

(6)

10-36

wave, and

the sheet.

of

(c) total

= 0. A

/,

= e, =

is

plane 1.59-GHz wave in

If the electric field of the incident

E from x =

(at the sheet) to

wave

is

= 4 m for (a) incident

wave.

In Prob. 10-34 at what distance from the sheet

down 40 dB

O m -1

is

the reflected

wave

(a)

down

with respect to the incident wave at the same distance?

Antarctica and Greenland are covered with ice of great thickness.

Radioby sending radio waves

down through the ice and obtaining a reflection from the ground under the ice. It is assumed

equipped surface vehicles are used to measure the thickness of the

ice

PLANE WAVES IN DIELECTRIC AND CONDUCTING MEDIA

PROBLEMS

425

that the ice is homogeneous and of uniform thickness over a flat perfectly conducting earth.
Take the constants for the ice at 30 MHz as ft, = 1 and e P = 3 jO.Ol. (a) Find the ratio of
the power received per unit area to the power radiated at 30 MHz for a 2-km thickness of ice.
Express in decibels, (b) Compare this value with the ratio in decibels if the ice were replaced
by air. The difference is the attenuation due to the presence of the ice. In (a) and (b)
assume that the vehicle transmitter radiates isotropically and that there is no loss (mismatch)
at the air-ice interface,
(c) What is the travel time for the wave down and back through the
ice? (d) Would it be more accurate to determine the ice thickness by pulsing the transmitter
and measuring the delay time of the reflected pulse, as in a radar system, or by measuring the
wave attenuation? (e) Which scheme would be more practical?

10-37

By a technique

waves for exploring the

10-38
1

GHz,

like that in Prob. 10-36 discuss the possibilities

interior of the

of using radio

moon.

If distilled water has constants

/* r

calculate (a) the l/e depth of penetration

l,

and

er

=81 and power

(b) the 1

factor

= 0.05

at

percent depth of penetration

at this frequency.

10-39 (a) A 1-GHz traveling wave in air has an electric field =100 MV m -1
wave is incident normally on a flat copper sheet, how thick must the sheet be to reduce
the wave transmitted through the sheet to the nonhazardous level oflOOWm -2 ? Hint:
.

If the

Consider the copper sheet as a short section of lossy transmission line connected between two
lossless lines
levels see

of characteristic impedance

W. W. Mumford, Some

Proc. IRE, 49:427-447 (February 1961)].

in the sheet ?

accomplished?

(c)

Z = 376.7

CX.

[Regarding hazardous radiation

Technical Aspects of Microwave Radiation Hazards,

(b)

What power

is

dissipated per square meter

Will dissipation of this power be a problem, and

if so,

how can

it

be

11
WAVE POLARIZATION

11-1

INTRODUCTION!

In this chapter the polarization of electromagnetic waves

elliptical,

and

circular

polarization are considered, and

is

discussed.

their relation

Linear,

on the Poincare

developed for coherent, completely

polarized waves and then extended to the case of partially and completely unpolarized
or incoherent radiation. Finally Stokes parameters and coherency matrices are
sphere

is

explained.

These concepts are

first

introduced for handling waves of this kind.

11-2

LINEAR, ELLIPTICAL,

AND CIRCULAR

POLARIZATION
Consider a plane wave traveling out of the page (positive z direction), as in Fig. 1 1-1 a,
This wave is said to be linearly
field at all times in the y direction.
polarized (in the y direction). As a function of time and position the electric field of a
with the electric

more detailed discussion of wave polarization is given in J. D. Kraus, " Radio

Astronomy," McGraw-Hill Book Company, New York, 1966.

WAVE POLARIZATION

sec. 11 -2J

Linear

Elliptical

polarization

polarization

427

Circular
polarization

r. ftk*

>

(c)

FIGURE

11-1

Linear, elliptical, and circular polarization.

linearly polarized

page)

is

wave

(as in Fig. 11 -la) traveling in the positive z direction (out

of

given by

Ey = E2 sin (cot - pz)

In general the electric

field

(1)

of a wave traveling in the z direction

may have

both a

y component and an x component, as suggested in Fig. 11-16. In this more general

situation the wave is said to be elliptically polarized.
At a fixed value of z the electric
vector

E rotates as a function

the polarization

ellipse.

The

ellipse is called the axial ratio

Two
Fig.

of time, the tip of the vector describing an

ratio

(AR).

ellipse called

of the major to minor axes of the polarization

Thus, for the wave in Fig. 11-16,

AR = E1 \E

extreme cases of elliptical polarization correspond to circular polarization, as in

1 1-1

c,

and

and AR =

1,

linear polarization, as in Fig.

while for linear polarization

1 1-1

E =
t

a.

For circular polarization ,

and AR = oo.

In the most general case of elliptical polarization the polarization ellipse

have any orientation, as suggested

in Fig. 11-2.

The

= E2
may

wave may
the x direction and

elliptically polarized

be expressed in terms of two linearly polarized components, one

in

one in the y direction. Thus, if the wave is traveling in the positive z direction (out
of the page), the electric field components in the x and y directions are

Ex = j sm(wt - Pz)
Ey = E2 sin(cof - Pz +
where

(2)

d)

(3)

E = amplitude of wave linearly polarized in x direction

E2 = amplitude of wave linearly polarized in y direction
= time-phase angle by which Ey leads Ex
i

<5

Combining

(2)

and

(3) gives the

instantaneous total vector

E = kE sin(wr - Pz) + y2
l

sin(cot

field

E:

- Pz + S)

(4)

428

ELECTROMAGNETICS

[CHAP. 11

Polarization ellipse

FIGURE

11-2

Polarization ellipse at

and amplitudes

At z =

0,

Ex = E

sin cot

(or

tilt

angle t showing instantaneous components

peak values)

and

E2

(5) eliminates cot,

ES

Equation
is

(7) describes

ellipse is t.

Intro

(6)

<5

(7)

2 cos 5

E E2
X

a (polarization)

The

= ^/l (EJE^

sin

, 2

ellipse, as in Fig. 11-2.

the semimajor axis, and the line segment

of the

= sin 2
tA
2

bEx Ey + cEy2

1-

i 2

aEx

sin

cot

(5)

<5)

and on rearranging we obtain

2E-x E y cosS

il2
or

Ey

<5).

From the relation for Ex we have sin cot = EJE X and cos

where

and

Expanding Ey yields
Ey = E2 sin(cot +
E = 2 (sin cot cos + cos cot sin

and

<5

ducing these in

Ex

OB is

sin

The

5
line

the semiminor axis.

segment

The

tilt

OA

angle

axial ratio is

OA
~
AR=
OB

(1

< AR <

oo)

(8)

WAVE POLARIZATION

SEC. 11-31

429

Wave approaching
V

=90"

w*=0

FIGURE

11-3

Instantaneous orientation of electric field

vector

E at two instants of time for a left

circularly polarized

wave which

is

ap-

proaching (out of page).

(o)

(b)

E = 0, the wave is linearly polarized in the y direction. If E2 = 0, the wave is

and \ =E2 the wave is also linearly
If 5 =
linearly polarized in the x direction.
If

polarized but in a plane at an angle of 45 with respect to the

If

wave

the

= E2 and

is left

polarized.

+90, the wave

= +90

that
Fig.

as in Fig.

and when 5
and for z = and

Thus, at a fixed position

= 0)

The

+90,

<5

is

right circularly

we have from

(2)

{at = 90) E = lE^

and
,

(3)

as in

the electric field vector rotates clockwise

According to the IEEE

(viewing the wave approaching).

to left circular polarization. f

(z

= 45).
When =

axis (t

90, the wave

One-quarter cycle later

l-3a.

is circularly polarized.

circularly polarized,

For the case d

E = y2
1-36.

definition, this corresponds

opposite rotation direction

(<5

90) corre-

sponds to right circular polarization.

If the wave

is

viewed receding (from negative z axis in Fig.

appears to rotate in the opposite direction.

1 1

-3),

the electric vector

Hence, clockwise rotation of

with the

wave approaching is the same as counterclockwise rotation with the wave receding.
Thus, unless the wave direction is specified, there is a possibility of ambiguity as to
whether the wave is left- or right-handed. This can be avoided by defining the
polarization with the aid of helical-beam antennas (see Chap. 14).

handed helical-beam antenna radiates

(or receives) right circular

Thus, a right-

(IEEE) polarization. J

A right-handed helix, like a right-handed screw, is right-handed regardless of the position

11-3

from which

it is

viewed.

There

is

no

possibility here

of ambiguity.

POYNTING VECTOR FOR ELLIPTIC ALLY

OR CIRCULARLY POLARIZED WAVES

In complex notation with

all

complex quantities (phasors) indicated

dot, the complex Poynting vector

S = it x H*
t This
%

IEEE

explicitly

by a

is

(1)

definition is opposite to the classical optics definition.

A left-handed helical-beam antenna radiates (or receives) left circular (IEEE) polarization.

430 ELECTROMAGNETICS

[CHAP. 11

The average Poynting vector

is

the real part of

or

(1),

Sav = Re = i Re t, x
Referring to Fig. 11-2,

ft*

(2)

x and y components

the elliptically polarized wave have

let

with a phase difference 5 as given by

Ex = E ella'Ey = E 2 ei(a"- pz+a)

fiz)

(3)

At z

the total electric field (vector)

it

Note that

is

a complex vector (phasor- vector) which

EX and yy where

The

Si is

H-field

is
is

component associated with

the phase lag of

H field (vector) at z =
ft

= $H y -

The complex conjugate of ft

is

is

// 1 e

Ex

(6)

The

H-field

component associated

= H 2 ei ^'- fz+d - i)

for a

Hti x

Ex

/(w '- /,z - )

with respect to

IIx
total

two com-

and a complex part or phasor (indicating time phase).

and Ex is the phasor.

Ey

resolvable into

the vector

//J

where {

is

each component has a vector part

(indicating space direction)

Thus, in StEx

The

(5)

= unit vector in x direction

= unit vector in y direction

ponent complex vectors

with

then

is

ZE^ " + y 2 eKw,+i)

t = EX + $Ey =
where

(4)

wave traveling

= yH

(ra,

1 e''

(7)

in the positive z direction

- HH 2 ^

' )

is

then

mt+> -^

(8)

equal to (8) except for the sign of the exponents.

That

is,

ft*

Substituting (5)

and

= yi^e - -"

" - *'

(9) in (2) gives the

- HH2 e- J(a" +d -v

average Poynting vector at z

Sav = i Re[(* x $)EX H* -

(y

(9)

as

x ), #*]

= i*Re(Ex H; + Ey ti*x )
where 2

is

the unit vector in the z direction (direction of propagation of wave).

(10)
It

follows that

Sav = it(E Hi Re e" +

t

= **(,#!+
It is

to be noted that

Sav

is

E2 H2 Re e1 *)

tf2 ) cos {

independent of S.

(11)

SEC. 11-4]

WAVE POLARIZATION 431

In a lossless

medium

EJHi = E2 jH2 = z

>

where Z, the

(electric

intrinsic

and magnetic fields in time phase) and

impedance of the medium, is real; so

+E H
+ H )Z = itH Z
the amplitude of the total H

Sav =ii{E l H l

= \%{H

H = y/H^ + H

where

is

2)

field.

E = -J~E 2 + E2

is

We can also

write

E2

*.-*.*'+*
where

(12)

(13)

the amplitude of the total

field.

EXAMPLE An elliptically polarized wave traveling in the positive z

direction in air has

and y components

(Vm -1 )
Ex = 3 sin(wt fiz)
(Vm"
E, = 6 sin(./ - |Sz + 75)

Find the average power unit per area conveyed by the wave.
solution

which from

The average power per

(13) has a

unit area

is

equal to the average Poynting vector,

magnitude

IE 2

E 2 + E2 *
1

Z
=3 V m -1

"~2Z~~2
From

the stated conditions the amplitude Ei

Also for

11-4

air

Z = 376.7

H.

and the amplitude

i=6Vm"'.

Hence

THE POLARIZATION ELLIPSE AND THE

POINCARE SPHERE

In the Poincare spheref representation of wave polarization, the polarization state

is

described by a point on a sphere where the longitude and latitude of the point are
related to parameters of the polarization ellipse (see Fig. 11-4) as follows:

Longitude
Latitude
t

= 2t
= 2e

(I)

H. Poincare, "Theorie mathematique de la lumiere," G. Carre, Paris, 1892; G. A.

Deschamps, Geometrical Representaton of the Polarization of a Plane Electromagnetic Wave, Proc. IRE, 39: 540 (May 1951).

432

ELECTROMAGNETICS

[CHAP. 11

Polarization state

M(e,T)orP(7,>

FIGURE

11-4

Poincare

sphere showing

relation

of

angles e, t, y, and 8.

where x
e

The

= tilt

angle, 0

<t<

180

= cot-'C+AR), -45 < e <, +45

axial ratio

(AR)

is

negative for right-handed

and

positive for left-handed

(IEEE)

polarization.

on a sphere can also be expressed in

drawn
from a reference point on the
terms of the angle subtended by the great circle
equator (see Fig. 11-4) as
and
the
equator and the angle between the great circle

The

polarization state described by a point

follows:
Great-circle angle

= 2y

where y
5

Fig.

(2)

= tan"^/-^), 0 <, y < 90

= phase difference between Ey and Ex - 180 <

Equator-to-great-circle angle

< + 180

The geometric relation of t, e, and y to the polarization ellipse is illustrated

The trigonometric interrelations of t, e, y, and <5 are as follows :f
1 1-5.

cos2y

in

cos 2e cos 2t

tan2e
tan S =
sin 2t

(3)

= tan 2y cos 5

tan2r

sin2e = sin 2y sin 5

=

one can determine y and S or vice versa. It is convenient to describe the polarization state by either of the two sets of angles (e,t) or (y,5) which

Knowing

and

x,

See M. Born and E. Wolf,

relations involve spherical trigonometry.
"Principles of Optics," pp. 24-27, The Macmillan Co., New York, 1964.

t These

WAVE POLARIZATION

SEC. 11-4]

FIGURE

11-5

Polarization ellipse showing relation of angles e, y,

describe a point

on the Poincare sphere

433

x.

Let the polarization state as a

(Fig. 11-4).

function of e and t be designated by M(e,x) or simply

and

M and the polarization

functioirof y and d be designated by P(y,d) or simply P, as in Fig.

1-4.

state as

Two

special

cases are of interest.

Case

For d = or 8 = 180, Ex and

on the equator represents a

Ey

that any point

(e

= 0)

the polarization

is

linear

and

are exactly in phase or out of phase, so

state

of linear polarization.

in the

direction (t

= 0),

At

the origin

as suggested in

On the equator 90 to the right the polarization is linear with a tilt angle
1 l-6a.
=
45,
while 180 from the origin the polarization is linear and in the y direction
t
=
90).
(t
See Fig. 1 l-6a and b. One octant of the Poincare sphere is shown in
Fig.

Fig.

l-6a

Case 2

and the full sphere is shown in

For S

= 90

and E2

= Ey

Fig.

(2y

l-6b in rectangular projection.

90 and 2c

amplitudes but are in phase quadrature, which

tion.

is

90)

Ex and Ey have equal

the condition for circular polariza-

Thus, the poles represent a state of circular polarization, the upper pole

representing

left

circular polarization

tion, as suggested in Fig. 11 -6a

Cases

and 2 represent

and

limiting conditions.

the upper hemisphere describes a

left circular at

and the lower pole right circular (IEEE) polariza-

b.

left elliptically

the pole to linear at the equator.

In the general case any point on

polarized

wave ranging from pure

Likewise, any point on the lower

434 ELECTROMAGNETICS

[CHAP. 11

Left circular polarization c

= 45

Left elliptical polarization

Linear polarization

r=45,e=0
Linear polarization
-

= 22%,e=0

Linear polarization

r=0,=0
(a)

FIGURE
One

hemisphere describes a right

11-6 a
octant of Poincare sphere with polarization states.

elliptically polarized

are

shown by

ellipses

with appropriate

the Poincar6 sphere in Fig. 11-6 a

As an

tilt

and

Fof an antenna

angles t and axial ratios

to a

AR at points on

may

it

be shown that

wave of arbitrary polarization

V = k cos

is

given byf

MM.
(4)

MM = angle subtended by
of wave
M = polarization
of antenna
M = polarization

great-circle line

right

b.

application of the Poincare sphere representation

the voltage response

where

wave ranging from pure

Several elliptical states of polarization

circular at the pole to linear at the equator.

from polarization

state

M to M

state
state

k=
The

constant

polarization state of the antenna

radiated by the antenna

field strength

when

it is

is

defined as the polarization state of the

transmitting.

The

factor

of the wave and the size of the antenna.

MM

= 0,

An

wave

in (4) involves the

important result to

matched to the wave (polarization state of

a
wave same as for antenna) and the response is maximized. However, if
a = 180,
the response is zero. This can occur, for example, if the wave is linearly polarized in
the y direction while the antenna is linearly polarized in the x direction; or if the wave
note

is

that if

the antenna

is

MM

WAVE POLARIZATION

SEC. 11-5]

ee

Left circular polarization

-=45

**b

435

1=22 *

Left

&

elliptical

0.

polarization

V
T=90

-*

7^

Right/ ")

(}

polarization

t=45

r=0

T=135

Linear

= 0

r=90*

elliptical

{>

polarization

-- --

Right circular polarization

(6)

FIGURE 11-6 6
Rectangular projection of Poincare sphere showing

full

range of polarization

states.

is

polarized while the antenna

left circularly

generally

we may say

that an antenna

is

is

right circularly polarized.

More

blind to a wave of opposite (or antipodal)

polarization state.

11-5

PARTIAL POLARIZATION AND

THE STOKES PARAMETERS

The previous

E ir E2

and

sections deal with completely polarized waves,

<5

i.e.,

waves for which

are constants (or at least slowly varying functions of time).

radiation from a single-frequency (monochromatic) radio transmitter

is

of

The

this type.

However, the radiation from many celestial radio sources extends over a wide frequency range and within any bandwidth A/ consists of the superposition of a large
number of statistically independent waves of a variety of polarizations. The resultant
wave is said to be incoherent or unpolarized. For such a wave we can write

Ex = E (t) sin cot

Ey = E2 (t) sin[cot + 5(t)]

0)

t G. Sinclair,

The Transmission and Reception of

IKE,3: 151(1950).

(2)

Elliptically Polarized

Waves, Proc.

436 ELECTROMAGNETICS

where

[CHAP. 11

the time functions are independent.

all

with time are slow compared with that of the

The

variations of E^t),

E2 (t), and <5(f)

mean frequency /(co = 2nf) and

are of

A wave of this type could be generated by connecting

the order of the bandwidth A/.

one noise generator to an antenna which is linearly polarized in the y direction and
another noise generator to an adjacent antenna which is linearly polarized in the x
If the waves from both antennas have the same average power at an
direction.
observing point, the total wave at this point will be completely unpolarized.

The most
polarized;

general situation of wave polarization occurs

i.e., it

may

when a wave

is

partially

be considered to be of two parts, one completely polarized and

the other completely unpolarized.

To

deal with partial polarization

introduced in

convenient to use the Stokes parameters,

it is

852 by Sir George Stokes.f

The Stokes parameters /, Q,

U,

and

V are

defined as follows

=S=

S,

*,-^ + <e

>
(3)

(4)

(5)

V = - (E E2
t

sin <5>

S<sin

2<r>

(6)

S = total Poynting vector magnitude for the wave

Sx Poynting vector component of wave polarized in x direction
S Poynting vector component of wave polarized in y direction
E t = amplitude of electric field component of wave polarized in x direction

where

E2 = amplitude of electric field component

Z = intrinsic impedance of medium
The

angles 5,

c,

and

t are as defined in the previous section.

indicate the time average.

functions of time.

of wave polarized in y direction

It is

understood that in general

The angle brackets

<5, e, and t are

Eu E2

Thus, for example,

T
2

<, >

G. Stokes,

On

= ^ [ [i(0] 2 dt
1 Jq

(7)

the Composition Resolution of Streams of Polarized Light from

Different Sources, Trans.

Camb.

Phil.

Soc,

9(3):

399-416 (1852).

SEC.

WAVE POLARIZATION

H-5]

= Sy

For a completely unpolarized wave Sx

may

and

and

437

E2 are uncorrelated.

It

be shown that under these conditions

<1 2

cos 5}

<' 1 '2 sin 5>

and we have
I

=S

Q=0
u=o
v=o
S

where
is

the total Poynting vector of the wave.

is

The condition

a requirement for a completely unpolarized wave.

For a completely polarized wave E u E 2 6, e, and
,

constant, so that the Stokes parameters as in (3) to (6)

Thus, for example, we would have

For a
t

= e = 0)

may

Q = U= V=

be considered to be

do not require time

averages.

(Et 2 y = E*.

linearly (completely) polarized

wave with

in the

direction

(E2

= 0, and

the Stokes parameters are

I=S
Q=S
u=o
K=
For a
e

= 0)

linearly (completely) polarized

wave with

in the

direction

the Stokes parameters are

=5

C/

v=o
For a

left circularly

(completely) polarized

wave (E t

7=5

2=0
C/

V=S

= E2

= 90)

(E t

= 0,

90,

438

ELECTROMAGNETICS

By

[CHAP. 11

consideration of other polarization states

it

show

possible to

is

Stokes parameter /is always equal to the total power (density) of the wave,

that the

Q is equal

power in the linearly polarized components (in x or y directions), U is equal to

the power in the linearly polarized components at tilt angles t = 45 or 135, and V is
equal to the power in the circularly polarized components (left- or right-handed).
It is often convenient to normalize the Stokes parameters by dividing each
parameter by S, obtaining the normalized Stokes parameters s s lt s 2 and s 3 where
The normalized Stokes parameters
s = I/S = \,s 1 = QjS, s 2 = U/S, and s 3 = V/S.
11-1.
the
cases
discussed
summarized
in
Table
for
above are
If any of the parameters Q, U, or V (or s u s 2 or s 3 ) has a nonzero value, it
The degree ofpolarizaindicates the presence of a polarized component in the wave.
to the

tion

d is

defined as the ratio of completely polarized

polarized

d=-

or

d--

power

power

vV +

2
2
2
_s/Q + u + V

power, or

s2

+ *3

(8)
1

sf + s 2 2 + s 3 2 = 1 indicates a completely
polarized wave. A partially polarized wave may be regarded as the sum of a comThus, we may write for a
pletely unpolarized wave and a completely polarized wave.
The condition

2
2
2
Q +U + V =

total

0<rf<l

total

power to the

partially polarized

or

wave
ri

r*oi
Si

-di

d cos 2e cos 2t
d cos 2e sin 2t
d sin 2e

s2

Ls3 J

Table 11-1

(9)

NORMALIZED STOKES PARAMETERS FOR SEVEN

WAVE-POLARIZATION STATES
Completely polarized waves

Normalized

unpolarized

parameter

wave

So
Si

Si
S3

Circularly polarized

Linearly polarized

Completely

Stokes

= 0
1
1

= 90

= 45

-10
1

t=135

-1

Left-hand

Right-hand

-1

WAVE POLARIZATION

SEC. 11-5]

439

term on the right of (9) represents the unpolarized part and the second
term the polarized part of the wave.
Consider now a wave with Stokes parameters s ,s u s 2 s 3 incident on an antenna
with Stokes parameters a a, a 2 a 3 .f The parameters a ,a u a 2 ,a 3 are those for a

where the

first

wave radiated by the antenna when it is transmitting. It follows that the total power
available (to a receiver) from the antenna due to the incident wave is given by

P = iA e S[a

a3 ]

a2

a,

(10)

P = \A e S(a

or

aiSi

a 2 s2

A e = effective aperture of antenna (see Chap. 14),

-2
S = Poynting vector of incident wave, W m
result may also be expressed as

where
This

P = \A e S(\ +d cos

where

MM

a is

01)

m2

MM

(12)

a)

P = iA e S(l -d) + dSA e cos 2

or

(W)

a 3 s 3)

MM
(13)

the angle subtended by the great-circle line between the wave- and antenna-

on the Poincare sphere. In (13) the first term on the right side
power and the second term the polarized power. This
form emphasizes the fact that only one-half the incident unpolarized power is available
to a receiver but all the completely polarized power is available.

polarization states

represents the unpolarized

By arrangement of (13) we have

- + d cos
cwhere Fis the factor

in parentheses,

^Aa S = FA S
c

which may be

(W)

(14)

called the wave-to-antenna coupling

MM

a
factor. This factor is a function of the degree of polarization and the angle
and gives the fraction of the power A e S which is received. It is dimensionless with
and 1. Its possible range of values as a function of the degree of
values between

Ko, On the Reception of Quasi-monochromatic Partially Polarized Radio

Waves, Proc. IRE, 50: 1950 (September 1962).

t H. C.

440 ELECTROMAGNETICS

[CHAP. 11

0.2

0.4

0.3

0.5

0.6

0.7

0.8

0.9

1.0

Degree of polarization, d

FIGURE

11-7

Chart showing wave-to-antenna coupling factor F vs. the degree of polarization

d for various values of the wave-to-antenna angle MM. The region of possible
values is shaded. These values range from a perfect match (MM. = 0) to a
complete mismatch (MM.
180).
No values are possible outside this region.
.

Note

that for completely polarized

for a completely unpolarized

difference.

polarization

is

(After

waves (d=

wave (</=

1)

matching

0) the condition of

is

important but

match makes no

H. C. Ko.)

shown by the shaded area

in Fig. 11-7.

Thus, for a perfect match

{MMa = 0) F ranges from \ to as d goes from to 1, while for a complete mismatch (MMa = 180) F ranges from i to
as d goes from
to
Between these
extremes (MM =
and MMa = 180) we have conditions of partial mismatching
1

1.

MM

between antenna and wave, for example,

= 90.
a
Equation (14) may be interpreted as signifying that the response P of the system
is given by the maximum available power A S times the factor F which characterizes
e
the receiving antenna.

The

of all observing or measuring

be modified from the optimum
or ideal result by a quantity characterizing the observing or measuring system.
If we define four new parameters
relation (14)

is

typical

systems in that the response (or observed result)

*i

*2i

= i(-s + *i)
= i(s2 -js3 )

and four

similar parameters (a

available

power with

s 12
s 22

may

= i(s2 +js 3 )
= i(s - s )

(l5)

a 12 a 21 a 22 ) for the antenna, we can express the

,

2x2 matrices as

P=A

S Tr (pu

fliilpn

-Stall

U21

022.1 L*21

*22j/

(16)

WAVE POLARIZATION

SEC. 11-6]

where Tr

signifies

the trace,

P=A
More

i.e.,

sum of the diagonal elements of a square matrix or

the

S(a 11 s 11

+a 12 s21 +a 2i s 12 +a22 s22 )

(17)

=A SF

(18)

concisely

P = A e STr

where

F is

{[,,][*]}

(W)

the wave-to-antenna coupling factor, as in (14).

tions are similar to ones used in optics,

which

receiver as a detector of radio photons.

is

appropriate

The above matrix relawe regard the antenna-

if

In optics the matrices are called coherency

A further discussion of the response of an antenna to waves of arbitrary

matrices.^

polarization

11-6

441

is

given in Sec. 14-28.

CROSS FIELD

The preceding sections have dealt with the polarization of waves in which the directions
of the electric

field

vector at a given point are entirely perpendicular to the direction

of propagation.

Another situation can occur

parallel

in

which the

to the propagation direction.

exists in the case

component

there are electric field

of propagation.

components

to this direction, so that cross field

If the

medium

is

forward near the surface, so that

However,

in the near field

present (see Chap. 14).

medium

Also cross

and normal
field

may

be

with a plane wave traveling parallel

not perfectly conducting, the direction of E

field

of an antenna

in both the direction of propagation

is

present near the surface of a conducting

component

a plane

of a single plane wave in free space since such a wave has no

in the direction

to the surface.

electric vector rotates in

This condition, called cross field,% never

has a component

En

is tilted

normal to the surface and a

tangent to the surface (in the direction of propagation) (see Chap.

13).

may

be present wherever two waves of the same frequency

traveling in different directions cross, e.g., when a wave is reflected from a conducting
In general, cross

field

surface at an oblique angle (see Chap. 12).

M. Born and

E. Wolf, " Principles of Optics,"

The Macmillan Company, New York,

1964.
%

A. Alford, J. D. Kraus, and E. C. Barkofsky, chap. 9, p. 200, in Radio Research

Laboratory Staff, H. J. Reich (ed.), " Very High Frequency Techniques," McGrawHill Book Company, New York, 1947; see also J. D. Kraus, "Electromagnetics,"
1st ed., sec. 9-15, McGraw-Hill Book Company, New York, 1953.

442 ELECTROMAGNETICS

[CHAP. 11

PROBLEMS
*

11-1

wave

traveling normally out of the

of two linearly polarized components

resultant wave:

(a)

of the polarization
clockwise?

What

wave

and

is

the resultant

Ey = 2 cos(o>/ + 90).

For the

What is the angle between the major axis

x axis? (c) Does E rotate clockwise or counter-

the axial ratio?

(Z>)

and the positive

wave left- or right-handed?

ellipse

(d) Is the

11-2

is

page (toward the reader)

Ex = 3 cos <at

traveling normally out of the page

is the resultant of two circularly

For the resultant wave: (a) What
is the axial ratio ?
(b) What is the angle between the major axis of the polarization ellipse
and the positive x axis ? (c) Does E rotate clockwise or counterclockwise ? (d) Is the wave

polarized components E,

left-

= 2eJm

and E,

).

or right-handed?

A wave

11-3

traveling normally out of the page

polarized waves, one with components E*

ponents E,
(6)

= 4e -J<'+*'

What

x axis?
handed?

is

(c)

11-4
traveling

= 3e "
J

11-6

J ".

two

elliptically

For the

resultant

wave:

(a)

What

is

the axial ratio?

and the positive

the wave left- or right-

Does

E rotate

Show

that the instantaneous Poynting vector of a plane circularly polarized

is

Show

An

clockwise or counterclockwise?

(d) Is

a constant.
that the average Poynting vector of

that of a linearly polarized

= 4e~

the resultant of

the angle between the major axis of the polarization ellipse

wave

11-5

and ,

is

= 5 cos cot and Ey = 3 sin <ut and another with com-

wave

if

the

elliptically polarized

a circularly polarized wave

wave

in

is

twice

E is the same for both waves.

a medium with constants a = 0, /x = 2, e = 5

maximum electric

field

components (normal to the direction of propagation and normal to each other)

of amplitude 3 and 4 A m~\ Find the average power conveyed through an area of 5 m 2
normal to the direction of propagation.
has

//-field

11-7

flat

An

elliptically polarized

(AR = 2) 3-GHz wave

in air is incident

conducting sheet with intrinsic-impedance magnitude of 10 mfl.

instantaneous peak electric field

E = 100 mV m~ ',

If the

normally on
wave has an

find the average Poynting vector into the

sheet.

11-8

Three waves have the following characteristics:

elliptically polarized

with

d=

i,

AR=3,

= 135,

(a)

antennas produce the following types of waves when transmitting:

polarization, (2) right elliptical polarization with
zation.

unpolarized,

(b)

AR = 3, x = 45,

(1) linear horizontal

(3) right circular polari-

If all the antennas

have unit effective aperture and all waves have unit Poynting
power for the nine wave-antenna combinations a\, a2, a3, b\, bl, etc.
x 3 row-column display with top row a\ a2, a 3, second row bl bl, etc.

vector, find the received

Arrange answers in 3
11-9

left

Three

(c) left circularly polarized.

Ten waves have

the following characteristics

(a)

d=Q

(b)

(c)

d=i, AR=oo,t=45

(d)

d=i, AR=ao, t=0

a-=i,AR = -l

* Answers to starred problems are given in Appendix C.

WAVE POLARIZATION 443

PROBLEMS

(g)

rf=l,AR=+l
d=l,AR=o,T = 135

()

rf=i,AR=4,T=30

(e)

= 1,AR=co,t=90
rf=i,AR = 10,T=0
0) </ = *,AR = 3,t = 120

(/)

<*

(A)

these waves.
Find the normalized Stokes parameters and the coherency matrices for
45
received
by six antennas of
=
=
is
3, t
11-10 (a) A wave for which d = 0.4, AR
unit effective aperture with polarization as follows

(1) linear horizontal, (2) linear vertical,

left circular,
(3) linear slant (45), (4) linear slant (135), (5)

wave

has unit Poynting vector, find the six

antennas of part (a) respond equally?

all six

and

(6) right circular.

If the

a wave to which
If so what are the wave parameters? (See

power

responses.

(6) Is there

wave power

in

Sec. 14-28.)

What

11-11

11-12
for

(a)

each of five waves charac-

1,0,0,0, (6) 1, 1, 0, 0, (c) 1, 0,

- i, 0, (rf)

1, i,

1/V2, 1/V2.

i, 0, (e) 1, 0,

wave

the percentage of polarized

by the following Stokes parameters:

terized

is

(a)

which

What

are the

Q = U2 = V
2

wave
?

(6)

characteristics (t

Draw

and AR)

for

a completely polarized

the polarization ellipse.

A wave of unit Poynting vector with coherency matrix

11-13

[;:]
is

incident

on a

right circularly polarized receiving antenna of unit effective aperture.

Find

the received power.

The degree of polarization d of a wave can be resolved into a linearly polarized

2
=
2
component d, = VJl 2 + s 2 and a circularly polarized component dc = Vs 3 with d
which
2
for
polarized
wave
2
completely
for
a
What are the Stokes parameters
Vdi + dc
11-14

di

=d

and

for

which

si

= s2 ?

a graph as in Fig. 11-7 and plot points on the graph for each of the
11-15 (a)
of Prob. 11-8. (6) Add points to the graph for the six
combinations
wave-antenna
nine
11-13.
combinations of Prob. ll-10a. (c) Add another point for the combination of Prob.
11-7?
Fig.
of
(shaded)
{d) Do all points fall inside the inner triangular region

Draw

11-16

conducting
sheet

and

(a)

A plane traveling wave is incident at an angle of 45 to an infinite, perfectly

flat sheet.

The

electric field is

parallel to the direction

everywhere contained in a plane normal to the

Plot the axial ratio of the cross-

of wave propagation.

a function of distance from the sheet for a distance of

at distances of 0A, JA, |A, JA, and 1A from the sheet.

field ellipse as
field ellipses

1A.

(b)

Draw the cross-

One (y wave) is traveling

other
(x wave) is traveling
The
in the positive y
At the origin, E for the y wave
in the positive x direction and is left circularly polarized.
that E for the x wave is in the negative z
is in the positive z direction at the same instant
11-17

Two waves of equal magnitude intersect at the origin.

direction

direction.

11-18

What is
Draw

and

is

right circularly polarized.

the locus of the resulting

the polarization ellipse

E at the origin?

and the antenna response pattern of a

polarized antenna for waves with axial ratios as follows:

AR = 12, and(rf)AR=oo.

(a)

AR = 1,

(b)

linearly

AR=4,

(c)

444 ELECTROMAGNETICS

11-19

[CHAP. 11

Describe the characteristics (AR,

left-

or right-handed,

having Stokes parameters as follows:

(1)

1,0,0,1

(2)

1,0,0,-1

(3)

1,1,0,0

(4)

1,-1,0,0

(5)

1,0,1,0

(6)

1,0,-1,0

(7)

1,0.577,0.577,0.577

(8)

1,0,0,0

(9)

1,0,0,-*

(10)

1,*,0,0

(12)

i,-*,o,i

(11)

1,0,*,0

11-20

Demonstrate the

validity of Eqs. (11-4-3).

tilt

angle) of waves

12
WAVE REFLECTION, REFRACTION,
AND DIFFRACTION

12-1

INTRODUCTION

is given to wave reflection from and refraction into

and metallic surfaces. The plane wave normally incident on a boundary
between two media is first discussed, with the treatment then extended to include
plane waves obliquely incident on boundaries. The difference between physical-optics
(wavefront) and geometrical-optics (ray-path) methods is illustrated. The geometrical theory of diffraction is discussed and applied to a simple example.

In this chapter, consideration

dielectric

12-2

PLANE WAVE, NORMAL INCIDENCE

Consider a linearly polarized wave traveling in the positive x direction with E in the
in the z direction. The wave is incident normally on the boundary

y direction and H

between two media, as in Fig. 12-la. Assume that the incident traveling wave has
Part of the incident wave is, in general,
field components and il at the boundary.
The reflected
reflected while another part is transmitted into the second medium.
t

components
components , and 6,

traveling

wave has

wave has

field

field

r and

fi r at the boundary.

at the boundary.

The transmitted

446

[CHAP. 12

ELECTROMAGNETICS
V

Medium

Medium 2

Boundary
2 ,M 2

,<72

Incident

Hi

(a)

A
Transmitted
*-

Ey,
Reflected
-*

Hr

FIGURE

12-1

(&)

Plane wave incident normally on

boundary between two media and (b)
analogous transmission line.
(a)

From

the continuity of the tangential field components at a boundary

+ =
r

and

tfi

and magnetic
of the medium. Thus

The

electric

fields

i=z

(1)

H =#
r

(2)

of a plane wave are related by the intrinsic impedance

5=-z

t
l

ti t

$=z

x
l

H,

(3)

The impedance of the reflected wave (traveling in the negative x direction)

From (2) and (3)
be negative Z t and of the incident wave, positive Z x

is

taken to

= = -
^2

or

^1

^2
2
f
i_
z\
zT
*

(4)

^1
*i

(5)

WAVE

sec. 12-2]

Multiplying (1) by

Z 2 /Z x

Z,
(5)

and

(6),

we

AND DIFFRACTION

^E

+ ^E,

(6)

Zt

get

*K)-t*

(7)

2Z 2

or

where

447

gives

Z2

Adding

REFLECTION, REFRACTION,

z2 + z
f

(8)
t

called the transmission coefficient.

is

It

follows that

2^ 2
-E
~%~z +z
2
>

(9)
t

Subtracting (5) from (6) gives

(10)

Substituting E,

from

and solving for

(8) into (10)

E =
r

where p

is

From

(9)

and

h^E^p

called the reflection coefficient.^

E we

have

Ei

(11)

follows that

Z 2 -Z t
Z2 + Z

It

(12)

(12)

(12a)

The situation (Fig. 12-1) of a plane wave incident normally on a boundary

between two different media of infinite extent, with intrinsic impedances Zt and Z2 is
analogous to the situation of a guided wave on an infinite transmission line having
,

an abrupt change

in

impedance from

to

Z2

(Fig. 12-16).

The transmission and

reflection coefficients for voltage across the transmission line are identical to those given
t This

is

the

same as the

reflection coefficient discussed in Sec. 10-10.

448

ELECTROMAGNETICS

and

in (9)
tic

[CHAP. 12

(12) if the intrinsic

impedance of the

impedance Z2 of medium 2
right of the junction.

Returning

now

impedance Zj of medium

is

This

taken to be the characteris-

(Fig. 12-16)

and the

intrinsic

taken to be the characteristic impedance of the line to the

is

discussed further in Chap.

to the case of a plane

between two media of

1 is

of the junction

line to the left

3.

wave incident normally on the boundary

infinite extent as in Fig. 12- la, let

us consider several special

cases.

Case

Assume that medium 1 is air and medium 2

(8) we have the approximate relation

is

a conductor, so that Zt

> Z2

Then, from

,^
But from

(3) this

(13)

becomes

H.Z^^H^

(14)

H,*2H

(15)

from which
t

Thus, for a plane wave in air incident normally on a conducting medium, the magnetic
field

is,

good approximation, doubled

to a

follows that lir

boundary

(in

x Hi,

medium

so that there

in intensity at the

boundary.

a nearly pure standing wave to the

is

left

It

also

of the

1).

Case 2 Consider now the opposite situation, where medium 1 is a conductor and
medium 2 is air so that Z ^ Z2
Then, from (8) we have approximately

2
t

(16)

Thus, for a wave leaving a conducting medium, the electric

the boundary.

(VSWR =

It

follows that

,,
r

so that there

oo) immediately to the left of the

owing to the attenuation of waves in medium

moves away from the boundary (to the left).

Case

In case

1 it is

perfect conductor).

assumed that ZX

Then from

the transmission coefficient x

field is

transmitted into

magnetic

the

= 0.

(in

medium

Further

However,

1).

= (medium 2 a

=
1, and from (9)
p

Consider now thatZ2

Thus, the wave

2.

nearly doubled at

VSWR decreases rapidly as one

(12) the reflection coefficient

medium

field intensity exactly

>Z

boundary
1,

field is

a nearly pure standing wave

is

r =

is

completely reflected, and no

,, and

doubles at the boundary.

H =H
r

t ,

This situation

so that the
is

analogous

to a short-circuited transmission line.f

t Although p, t, t , and lt are, in general, complex and, hence, should be written with
a dot, the dot will frequently be omitted for simplicity.

WAVE

SEC. 12-3]

Case 4

In case 2

Z2

it is

REFLECTION, REFRACTION,

assumed that ZX -4,Z2

Then from (12) p

AND DIFFRACTION

449

Consider now the hypothetical situa-

= +1

and from (9) t = 2. Thus the

wave is completely reflected, but E, = +E so that the electric field intensity at the
boundary is exactly doubled. This situation is analogous to an open-circuited transtion where

is infinite.!

t ,

mission

line.

Case 5

Assume

dielectrics 0u x

and from

that both

= p2 = p

).

medium

Then

it

and medium 2 are

from (12) that

lossless

nonferromagnetic

follows

\Ai/2

C l/ 2

(9) that

T=

rr7^:
=Z

Case

(18)

6 Take now the case where Z2

Then p = and z = 1 , so that the wave
t
propagates into medium 2 without any reflection. This situation is similar
to that on a
continuous transmission line of uniform characteristic impedance.
.

THE TERMINATED WAVE

12-3

In the preceding section

we considered various reflection and transmission situations

a boundary. In all cases there was a reflected wave except where the two
media
were of the same impedance, and in this case the wave was entirely transmitted.
at

No

case was considered in which the incident wave is terminated so that

no wave will be
transmitted or reflected. This case deserves special mention and is
considered in this
section.

The intrinsic impedance of free space is 376.7

on more physical significance

for free space takes

This concept of an impedance

fi.

we

consider the properties of a

having a resistance of 376.7 fi per square. Material so treated is often
called space paper or space cloth.
It should be noted that the resistance is not per
square centimeter or per square meter but simply per square. This is
equivalent to
if

resistive sheet

saying that the resistance between the edges of any square section
of the material is
Hence the resistance between the opposite edges of the small square of

the same.|

It is to be noted that for free space the intrinsic

impedance is only 376.7 Q. To
obtain a higher impedance would require that
1 such as in ferromagnetic media.
ft.
X If the large square in Fig. 12-2 is 4 times the area of the small, it is equivalent to two
small squares in series connected in parallel with two more small
squares

>

Thus, the resistance of this large square

is

the

same

in series.
as that of the small square.

[CHAP. 12

450 ELECTROMAGNETICS

376.7

Zero -resistance
bars

(a)

(6)

FIGURE

12-2

Space cloth has a resistance of 376.7

space cloth in Fig. 12-2a

is

large square in Fig. 12-26.

376.7

Q,,

as

is

Q per square.

also the resistance between the edges of the

In this illustration

it is

assumed that the edges are clamped

with zero-resistance bars and that the impedance of the leads is negligible.
The conductivity of the material required for a sheet of the space cloth depends

on the thickness of the

is expressed by

sheet.

Thus the resistance

J_ = _

ahl

where

R of a square section as in Fig.

(O)

2-3

(1)

ah

= length of side, m
= area of edge, m 2
h thickness of sheet, m
-1
a = conductivity of sheet, Urn
/

It

follows that the required conductivity

Consider
wave.
right

is

is

Rh

376.7/j

(Urn

-1
)

now the behavior of a sheet of space cloth placed in the path

(2)

of a plane

Suppose, as shown in Fig. 12-4a, that a plane wave in free space traveling to the
incident normally on a sheet of space cloth of infinite extent.

FIGURE

12-3

Square of space cloth of thickness

h.

WAVE REFLECTION REFRACTION, AND DIFFRACTION

sec. 12-3]

J
A?,-l

FIGURE

12-4

Infinite

sheet

of space cloth

(a)

*-i

Incident

Reflected

Transmitted

wave

wave

wave

V,-l

K-t

Plane wave traveling to right incident

normally on sheet of space cloth and (A)
analogous transmission-line arrange(a)

451

V.-f

-T
Infinite

2o

()

transmission line

J-

ment.

\ we have from (12-2-9)

Taking the amplitude of the incident wave as 1 V
is a transmitted wave continuing to the right of the sheet of amplitude

that there

E,

2 x 188.3

tE,

188.3

and from

(12-2-12) that there

is

376.7

fv-

a reflected wave to the

376.7

left

^ = ^ = 188.3 + 376.7 =-3 Vm

188.3

It is

AA

of the sheet of amplitude

_1

to be noted that the impedance presented to the incident

wave at the sheet is

the resultant of the space cloth in parallel with the impedance of the space behind

This

is

one-half of 376.7, or 188.3

it.

SI.

apparent that a sheet of space cloth by itself is insufficient to terminate a

This may also be seen by considering the analogous transmission arrangement

It is

wave.

shown

in Fig. 12-46.

In order completely to absorb or terminate the incident wave without reflection

or transmission, let an infinite, perfectly conducting sheet be placed parallel to the
space cloth and A/4 behind it, as portrayed in Fig. 12-5a. Now the impedance
presented to the incident wave at the sheet of space cloth

is

ance of the sheet in parallel with an infinite impedance.

arrangement

results in the total absorption

reflection to the left of the space cloth.

376.7

fi,

is

this

of the wave by the space cloth, with no

There is, however, a standing-wave and

energy circulation between the cloth and the conducting sheet.

transmission-line arrangement

being the imped-

As a consequence,

The analogous

illustrated in Fig. 12-56.

In the case of the plane wave, the perfectly conducting sheet effectively isolates
the region of space behind

it

from the

effects

of the wave.

In a roughly analogous

452

ELECTROMAGNETICS

[CHAP. 12

Incident

wave

*(a)

^Infinite sheet
of space cloth

-Infinite perfectly

conducting sheet

Incident

wave
Shorting

bar

Infinite

(6)

transmission
line

FIGURE

12-5

Plane wave traveling to right incident normally on sheet of space cloth backed
by conducting sheet is absorbed without reflection, (b) Wave traveling to right
on transmission line is absorbed without reflection by analogous arrangement.

(a)

manner, the shorting bar on the transmission

line

reduces the wave beyond

it

to a

small value.

may

transmission line

also be terminated

by placing an impedance across

equal to the characteristic impedance of the line, as in Fig. 12-4fc,

and disconnecting the line beyond it. Although this provides a practical method of

the line which

is

terminating a transmission

wave

in space because

it is

A region

termination.

line,

there

is

no analogous counterpart

in the case

of a

not possible to "disconnect" the space to the right of the

of space

may

only be isolated or shielded, as by a perfectly

conducting sheet.

lossy mixture of

a high-^

material can be used effectively for

for free space 0,/e,

wave

enters

it

1).

(ferrite) material

wave absorption

and a high- (barium

if

the ratio

The mixture constitutes a physical

The velocity of the wave

without reflection.

fije is

titanate)

equal to that

discontinuity, but the

is

reduced and large

attenuation can occur in a short distance.

12-4

LINEARLY POLARIZED PLANE

WAVE, OBLIQUE INCIDENCE

We have previously considered a linearly polarized plane wave normally incident on an

interface

between two media.

In this section, attention

is

given to a linearly polarized

plane wave obliquely incident on a boundary between two media, as shown in Fig.

WAVE

sec. 12-4]

REFLECTION, REFRACTION,

AND DIFFRACTION

Medium

1:

Medium 2:

453

t 1 ,Mi,<^i,Zi,'7i

2 ,M2><'2<z 2il2

refraction

'Transmitted

FIGURE
Geometry

12-6
in the plane of incidence (xy plane) for linearly polarized

wave

at

oblique incidence and for perpendicular polarization.

12-6.

The

reflected

wavelf

incident

wave (from medium 1) makes an angle of 0, with the y axis, the

1) makes an angle of r with the y axis, and the transmitted

wave (medium

makes an angle of 0, with the negative y

axis.

perpendicular to the plane of incidence

Consider two
xy plane) and (2) the electric field parallel to the plane of incidence. These waves
are said to be perpendicularly-polarized and parallel-polarized, respectively. The field
vectors shown in Fig. 12-6 are for the case of perpendicular polarization. It is clear
that any arbitrary plane wave can be resolved into perpendicular and parallel comcases: (1) the electric field

(the

ponents.

Referring to Fig. 12-6, the x'y' axes are orthogonal, with the x' axis oriented

along the direction of the incident wave.

x'

The

transmitted

of refraction.

= x sin

wave

is

It is

seen that

+ y cos

(1)

also called the refracted wave.

Hence, 6t

is

called the angle

454 ELECTROMAGNETICS

and that a unit vector

[CHAP. 12

in the y' direction

y'

can be expressed as

=-&cosfl,+ ysinO,

(2)

Perpendicular Case
Consider an incident perpendicularly
direction,

(_L)

polarized wave propagating in the negative x'

i.e.,

and

Substituting (1)

(2) into (3)

E,

= l^e" rf

(3)

H,

= *'|V"'

(4)

and

(4) gives

E = tE expUP^x sin 0, + y cos 0,)]

(5)

H,

The

= ( - 4 cos

reflected fields are

From

(12-2-12)

and

y sin

0,)

- expUp^x sin 0, + j cos 0,)]

denoted by (E r ,Hr) and the transmitted

(12-2-9) the reflection

dicularly () polarized

wave

and transmission

fields

(6)

by (E,,H

coefficients for

( ).

a perpen-

are (at origin)

(7)

E,
(8)

*L

By developing
it

relations similar to (1)

Et

and

(2) for the reflected

and transmitted waves,

follows that

EP =

2p x E expUP^x sin Qr

-y cos

E,

= 2^ ^q exp[jp 2 (x sin 0, +

H,

= (-4 cos 0, +

(ft

cos 0,

y sin

)p

- expl/p^x sin
v cos

(9)

r)]

-y cos

(1 1)

0,)]

sin 0,)t

-5 sxp\jp2 (x

Z2

(10)

r)]

sin 0,

+ ^ cos

0,)]

(12)

The boundary conditions are as follows: (1) the tangential components of the
both media must be equal at y = 0, and (2) the tangential components

electric fields in

WAVE

sec. 12-4 J

of the magnetic
written using

expOjff,

From

(12-2-12a),

equal,

i.e.,

and

+p x = T x/?!

From

the

first

exp(JPi

I*

the angle

and using

Condition

455

can be

(1)

sin 0j

sin 0,)

= x x exp(jp 2 x sin 0,)

(13)

follows that the exponential arguments are

sin 6 r

/?!

= p 2 sin 0,

all

(14)

equality,

er

i.e.,

AND DIFFRACTION

(11), to yield

x sin 6,) + p

= 0.

both media must be equal at y

fields in

(5), (9),

REFLECTION, REFRACTION,

of reflection

is

(15)

equal to the angle of incidence.^

From the second

equality

(15),

sinfl,

^1 sin

(16)
0,

1i

where ri 1 and r\ 2 are the indexes of refraction of medium 1 and medium 2, respectively.
Equation (16) is known as Snell's law% and is a relation of fundamental importance in
geometrical optics. For a lossless medium the index of refraction n can be written as
equal to y/nr er

and

law becomes

Snell's

^-sin0

sin 0,

ll 2

EXAMPLE

(17)

If a wave is incident
from air onto polystyrene, calculate the angle of transmission B,
change polystyrene and air and repeat the calculation.

angle of

8,

solution

From

Polystyrene has a relative permittivity of 2.7.

= 30

From

air

|0.= /JL

e,

From

i=e

onto polystyrene,

f*i=Mo, i=2.7c

(16)

polystyrene onto

air, ei

= 2.7e

=
,

(0.5)

=0.304

2.7

17.7

/*i

= /*o

= *o

/*2

= /*o

= V2/7 (0.5) = 0.822

8, = 55.2

sin d,

t This is sometimes called Snell's law of reflection,

t This

is

sometimes called Snell's law of refraction.

f*2

at

an

Inter-

=/*<>

456

ELECTROMAGNETICS

Boundary condition

[CHAP. 12

(2)

can be written using

and

(6), (10), (12),

(14) with

y=

0, to

yield

-COS

+ p COS 0, = Tj.

0j

COS

Z,

(18)

0,

Z2

but from (12-2-12a)

and we have on

Px

(19)

*L

and solving for p

substituting (19) into (18)

Z2 cos Of Zj cos

~Z

where Zj and
It is

Z2

cos

+ Zt

cos

medium

are the impedances of

that

0,

and medium

2,

'

respectively.

seen that the previously derived reflection coefficient for normal incidence,

(12-2-12),

lossless

is

obtained as a special case of (20)

medium 2

If

is

nonmagnetic

a perfect conductor,
dielectrics, (20)

cos

Px

when

Z2 =

0,

0.

= 1.

and p x

If

both media are

becomes

0,-

-- sin 2

-Vfe/i)

cos 6 t

+ Vfe/i)

sin

(21)

Provided medium 2 is a more dense dielectric than medium

under the square root will be positive and p will be real.

If,

(e 2

>

ej, the quantity

however, the wave

is

from the more dense medium onto the less dense medium (q > e2 ), and if
2
sin 0j ^ e 2 /e u then p L becomes complex and \p x = 1.
Under these conditions, the
incident wave is totally internally reflected back into the more dense medium. f The
incident angle for which p = 1/0 is called the critical angle ic
From (21) it is
incident

seen that this happens

when

the radical

is

zero;

flte-sindefines the critical angle.

Using

Snell's law,

imaginary,

it is

For

all

seen that

i.e.,

(22)

angles greater than the critical angle, \p =1.

t >
ic , then sin 0, > l,t and cos 0, must be
\

when

i.e.,

cos 0,
t It can be
X

shown

y/l

that this

is

sin

0,

= JA

(23)

true for either perpendicular or parallel polarization.

Although the sine is greater than unity and the cosine is imaginary, there
interpretation of the resulting field in such a case.
See Example 2.

is

a simple

WAVE

SEC. 12-4]

A = y/(ej2 ) sin 2

where

now be

medium can

written,

E,

where

from

the less dense

electric field in the less

exp( ay) exp(y/? 2

A=

I-

(a s[v^e 2

medium

sin

x sin

457

dense

0,

6,)

(24)

(25)

has a magnitude x L E

decaying exponentially

away from the surface (y direction) and propagating without

tion.
Waves whose fields are of the form of (24) are called
results

AND DIFFRACTION

(11), as

= t E
2

The

a real number.

is

E in

Thus,

REFLECTION, REFRACTION,

loss in the

surface waves.

direc-

These

can be summarized by the Principle of Total Internal Reflection as follows.

is incident from the more dense onto the less dense medium at an angle

When a wave

equal to or exceeding the critical angle, the wave will be totally internally reflected

and

will also

be accompanied by a surface wave

dense medium.

in the less

2 Distilled water is a dielectric having the constants e, 81 p, = 1

If a wave
from water onto a water-air interface, calculate the critical angle. If the incident
m" ' and the incident angle is 45, calculate the magnitude of the field strength in

EXAMPLE
is

incident

Ex = V
1

the air (a) at the interface

solution

From

and

(b) A/4

away from the

interface.

(22),

= sin-' VJ^=6.38

lc

For0, =45,
sin d,

V8l" (0.707)

= 6.36

cos

= Vl - 6.36 2

=76.28

6,

so that

39.49
2"V
=
6.28 = ^

t,

at the interface

(6)

A/4

- 1 + - 1 + - 77 ~
- 1-42/- 44.64
0.707 + V Jf - 0.5

Therefore, the magnitude of the field strength

(a)

(Npm-

^o

Ao

away from the

E,

- 5

is:

.42

V m"

interface:

\E,

indicating that the surface

= 1.42 exp(- ^- ^\ = 73.2 ;*V m"

wave

is

very tightly

bound

to the water surface.

458

ELECTROMAGNETICS

Parallel

Case

Consider

now

[CHAP. 12

the case of parallel

Fig. 12-6 but with E,

by replacing H; by

(||)

The geometry

polarization.

is

the

same as

in

Er and E, parallel to the plane of incidence as would be obtained

H, by

f ,

and H, by E,

It

can be shown that the fields are

given by

Ej

= (- cos 0; +

H,

= - 1 -5 exp[/p\(x sin 6, + y cos 9,)]

E,

= ( $. cos

=2p,|

E,

0,

0j)

y sin

expljp^x

= (- cos

H, =

0,

y sin

r)pn

sin 0,

y sin

- >> cos

#<

+ y cos

equals the angle of reflection and that Snell's

+ y cos

0,)]

y cos

6r )]

(26)

(28)

(29)

r )]

exp[y/? 2 (jc sin

conditions, as before,

(27)

exp^^x sin

0,)T|| is

2t|| - exp^jC-f sm

By matching boundary

exp[jp t (x sin 9

6,+y cos

0,)]

(30)

(31)

#<)]

found that the angle of incidence

it is

law (16) holds.

It

can also be shown

that
1

The

reflection coefficient

is

lossless

nonmagnetic

Pll

It is

||

= 1

if

cos

6,

n
COS0;

"

(32)
ll

"

Z2 cos 6, Zj cos 9
cos
+ Z2 cos 9,

~ Z1

becomes
2

+vW*i) - sin Oi
2
f VW*i) - sin 9

-(e2 /i)cos0

medium 2

(33)

dielectrics

(Cj/Ci)

and to be p

P\\

found to be

P][

which for

cos 6 t

(34)

a perfect conductor.

is

of especial interest that, for parallel polarization,

it is possible to find an
and the wave is totally transmitted into medium 2. This
the Brewster angle 9 iB can be found by setting the numerator of (34)

incident angle so that

angle, called

pu

equal to zero, giving

iB

sin

^fo

+(ca /ei)

tan

^t

(35)

WAVE

sec. 12-4]

REFLECTION, REFRACTION,

AND DIFFRACTION

Perpendicular

459

Parallel

-r

'e r

/Pj.

=81110,2

JT

=81,'10,2>

90

FIGURE

90

30

60

ii

and

,.

=81

60

30

parallel polarization vs. angle

= 81), flint glass (e, =

incidence 6 t for distilled water (e,

=2
= 10

12-7

Reflection coefficient for perpendicular

6,

,.

*_

at the Brewster angle (see

Example

10),

and

paraffin (ep

of

= 2).

3).

wave comcomponents and incident at the Brewster

angle produces a reflected wave with only a perpendicular component. Thus, a
circularly polarized wave incident at the Brewster angle becomes linearly polarized on

The Brewster angle

is

also sometimes called the polarizirtg angle since a

posed of both perpendicular and

parallel

reflection.

EXAMPLE 3 A parallel-polarized wave is incident from air onto (a) distilled water
(e = 81), (b) flint glass (e = 10), and (c) paraffin (c, = 2).
Find the Brewster angle for each
r

of these

cases.

SOLUTION
(a)

e tB

= tan " Vsi = 83.7

(6)

6, B

= sin -

(c)

6 tB

= sin "

'

The magnitude and phase angle of the

Fig. 12-7 for

and

waves incident from

paraffin (er

air

tan" 1 VTb = 72.4

tan

"l

Vl = 54.7

reflection coefficiehts

onto

distilled

water

(.

p x and

p||

are

shown

81), flint glass (er

in

10),

= 2). Note that the phase angle for p x is klways n, whereas for p,< the

460 ELECTROMAGNETICS

[CHAP. 12

phase angle jumps from

The

to

at the Brewster angle.

has unit magnitude at grazing incidence (8

coefficient

quantities pu

and p x as given

in (21)

Furthermore, the reflection

= 90).

and

(34) are

sometimes called the

Fresnel (fray-nel') reflection coefficients.

ELLIPTICALLY POLARIZED PLANE WAVE,

OBLIQUE INCIDENCE

12-5

We now consider an elliptically polarized plane wave

The problem

ary.

is

obliquely incident

on a bound-

to find the magnitude and polarization of the reflected and

transmitted waves.

The

incident electric field

(E tl ) components, as shown in
orthogonal

field

composed of both

is

inset

of

<

is

the angle

by which

Thus, as in Sec.

tudes.
(^,(50,

Ey

leads

'

(1)
>

(2)

Ez and Ei and E,

1-4, the incident

component ampliby

ellipse specified

where

(3)

polarization ellipse can also be specified in terms of

angle 1

are the
n

wave has a polarization

y^tan- 1
The

As viewed from the origin, the

components are

Ez = E iX e>"
Ey = , eJ m,+d
where 5

and perpendicular

parallel (j||)

Fig. 12-8.

From

(1 1-4-3)

we can

relate

(AR, ,t,)

its

axial ratio

AR

and

tilt

to (y f ,(5 f)

2t,-

tan 2y, cos 5

sin 2 (

sin 2?; sin

cos 2fi

cos 2e, cos 2t;

(6)

tan 6,

(7)

tan

(4)

(5)

tan 2e,
sin 2%i

where
If the

e,

wave

is

= cot" * ( + AR,)

right elliptically polarized, the

polarized, the plus sign

is

minus sign

(8)
is

used;

if it is left elliptically

used.

Considering the reflected wave as viewed from point

P (see inset B), the orthog-

onal field components are

E2 = Erl e* =
Ey.

= -r

||

e*

p J a expL/M

IP,,

|,

+ 4J\

expy{wt

*,,

(9)

dt

w)]

(10)

WAVE

sec. 12-5]

REFLECTION, REFRACTION,

AND DIFFRACTION

461

Reflected

FIGURE

12-8

Geometry

in the plane of incidence (xy plane) for elliptically polarized

wave
on a plane surface. The incident electric field has parallel
{E, M ) and perpendicular (E .) components which appear as in inset A when viewed
from the origin O. The reflected and transmitted fields shown in insets B and C
are viewed looking toward the origin from points P and Q, respectively.
incident obliquely

where

<f>^

cients

P||

and are the phase angles of the parallel and perpendicular reflection coeffiand p respectively. The phase angle by which E leads Ez is thus given
r
<j>

by
S,

and

The

<(>,,)

01)

---(a-'")

(12)

<5j

(4> L

similarly [see (3)],

relations (11)

by an observer

and

(12) specify the polarization state of the reflected wave, as seen

at point P.

To

and tilt angle of the polarization

and (8) are used with the subscript i

find the axial ratio

of the reflected wave, the relations

replaced by the subscript r.
ellipse

(4), (5),

[CHAP. 12

462 ELECTROMAGNETICS

Reflected

FIGURE 12-9a

[Example

1(a)]

Right circularly polarized (RCP) wave incident on a perfect conductor. The

No wave is transmitted.
reflected wave is left circularly polarized (LCP).

can be shown that an observer at point

see a transmitted wave with polarization ellipse given by
In similar fashion,

it

a,

= *, +

where

|,

and

coefficients

EXAMPLE
air

onto

(a)

and

(a)

wave

for these

and perpendicular transmission

When medium

two cases?
1 is

air

pn= 1 /180
180

(14)

A right circularly polarized wave (RCP) is incident at an angle of 45 from

What is the polarization
(b) polystyrene (c = 2.7).

solution

will

respectively.

a perfect conductor and

state of the reflected

For a

C)

( 13 >

tfi-f||)

x are the phase angles of the parallel

-r

(see inset

-(&[-)

tan

right circularly polarized wave,

+ (180 -

180)

= 90

-c

is

is

a perfect conductor,

= 1/180
S, = 90, so that from (11), 3, = 90
y = tan" *(T x 1) = 45. Therefore the

px

= 45,

and from

Note that when

and medium 2

(12),

used in this section to denote the

lilt

angle

it is

associated with

the incident, reflected, or transmitted wave (that is, t, , -c p , or t,) but when it is
used to denote the transmission coefficient it is associated with either a parallel

or perpendicular case (that

is, T|,

or t_J.

WAVE

SEC. 12-5]

REFLECTION, REFRACTION,

AND DIFFRACTION

wave is left circularly polarized (LCP), as shown in Fig. 12-9a. Conversely,

wave had been LCP, the reflected wave would be RCP.
(6) When medium 1 is air and medium 2 is polystyrene, then from (12-4-34)

reflected

463

if

the

incident

_ - 2.7(0.707) +V2.7- 0.5 =

2.7(0.707) + V2.7 - 0.5
and from

(12-4-21)

0.707
0.707
Therefore,
8,

-V2.7- 0.5
+ V2.7 - 0.5

= 2.81.

Thus the

=0

= -90 + 180 + (180 - 180) = 90

Substituting these values into (4), (5),

and AR,

gQO

reflected

and (8) (with subscripts / replaced by r), we find t, = 0

wave is left elliptically polarized (LEP), as shown in

Fig. 12-96.

EXAMPLE

wave whose electric

an angle of 45 from

linearly polarized

A)

(see Fig. 12-8, inset

is

incident at

= 2.7).

What

field vector bisects the zy' axis

(a) Since

pN

1 /180

a perfect conductor,

(a)

the polarization state of the reflected

is
and (b) polystyrene (e r
two cases? (c) If the angle of incidence onto polystyrene
is the polarization state of the reflected wave?

solution

onto

air

and px

is

wave

for these

58.7 (Brewster angle),

what

1 /180 ,

= 0 + 180 + (180 - 180) = 180

y, = tan- (T>< 1)=45
8P

From

(4),

tan 2t p

wave

reflected
(b)

is

= 00, so that t, = 135

linearly polarized with

From Example
p

y, = tan= 0.815, so

so

that

cos

it

wave
z

is

and from

(5)

and

= 0.126 /180

and

= 19.6 and
= 70.4. From (5)

(0.126/0.354)

that t,

px

0,

= 00.

Therefore the

and y"

axes.

8,

= 0.354/180

= 180. From
(8), AR = co.

and

linearly polarized with the electric field vector

When

AR,

1,

= 6 = 58.7,
p =0
fB

P
Therefore the reflected wave

is

0.520- 1.40
0.520+1.40

= "-TO^_?_

perpendicularly polarized.

(4),

tan 2t,

= tan 39.2

Therefore the reflected

making an angle of

axis.
(c)

(8),

the electric field vector bisecting the

70.4 with the

464 ELECTROMAGNETICS

[CHAP. 12

Reflected

FIGURE

12-96 [Example 1(6)]

Right circularly polarized (RCP) wave incident on dielectric slab of polystyrene.

The reflected wave is left elliptically polarized (LEP) with major axis of the
polarization ellipse horizontal

and the transmitted wave right elliptically polarized

axis of the polarization ellipse in the (vertical) plane of incidence. (See Prob. 12-15.) If the angle of incidence is increased to 58.7
(= Brewster angle) the angle of reflection increases to 58.7 and the angle of

(REP) with major

refraction increases
reflected

12-6

wave

from 25.5 to 31.3. Also the polarization ellipse of the

a straight horizontal line (linear J_ polarization).

collapses to

HUYGENS' PRINCIPLE AND PHYSICAL OPTICS

Huygens' principlef

states that each point

on a primary wavefront can be considered

to

be a new source of a secondary spherical wave and that a secondary wavefront can be
constructed as the envelope of these secondary spherical waves, as suggested in Fig.
This fundamental principle of physical optics can be used to explain the
apparent bending of radio waves around obstacles, i.e., the diffraction of waves.
12-10.

t C. Huygens, "Traite de la lumifcre," Leyden, 1690. English translation

Thompson, London, 1912, reprinted by The University of Chicago Press.

by

S. P.

WAVE

sec. 12-6]

REFLECTION, REFRACTION,

AND DIFFRACTION 465

Point source

FIGURE

12-10

Illustrating

Huygens' principle of physi(point-to-wave

optics

cal

correspon-

prjmary
wavefront

Secondary
wavefront
envelope

dence.)

diffracted ray

is

one that follows a path that cannot be interpreted as either

reflection

or refraction.

As an example,

consider a uniform plane wave incident on a conducting half-

plane, as in Fig. 12-1 la. f

Huygens' principle,

We want to calculate the electric field at point P by using

i.e.,

E=

j dE

(1)

over

xaxis

where dE

is

the electric field at

origin O, as in Fig. 12-1 lb,

P due to

a point source at a distance x from the

i.e.,

dE = e-Jf+dx

(2)

so that

(3)

If 8

< r,

it

Ja

follows that

X2
2r
t J.

(4)

D. Kraus, "Radio Astronomy," pp. 194-198, McGraw-Hill Book Company,

New York,

1966.

466

[CHAP. 12

ELECTROMAGNETICS

Eo

Plane

wave

^Conducting
half plane

(6)

(a)

FIGURE

12-11

Diffraction of a plane

When we

let

k2

2jrX

and u

wave by a conducting

half-plane.

kx, (3) becomes

(5)

which can be rewritten as

_

E = Eo

Kr

The

integrals in (6)

WVoC

-i*u*i2

du

e -J*uV2

Jo

du \

(6)

have the form of Fresnel integrals! and can be written

E = h. e -J*Q -ft -

[C(ko) -jSQca)]}

(7)

Kr

where

C(ko)

cos

Fresnel cosine integral

(8)

du

Fresnel sine integral

(9)

mi 2

sin

Jo

du

Ka

S(ko)

7114

/*

Jo

Mathematical Functions,"

Steguh, "Handbook of
t See M. Abramowitz and I. A.
Government Printing Office,
pp. 295-329, National Bureau of Standards, U.S.
35-38, Dover Publica1964. E. Jahnke and F. Erode, "Tables of Functions," pp.
tions, New York, 1943.

WAVE

sec. 12-6]

REFLECTION, REFRACTION,

FIGURE

AND DIFFRACTION

467

12-12

Physical-optics solution to half-plane diffraction pattern.

The power

density at

P is then

EE*

<~=-2z=

Soii[i -

wheret

C(ko)]

So

+ [i-

S(ko)]

(Wm" 2 )

2
}

(10)

E X
HZr

(11)

The

half-plane diffraction curve, (10), is shown in Fig. 12-12.

Note that when
ko = oo, which corresponds to no edge present, S = S when ko = 0, 5av = 5 /4;
and when xa = +oo, which corresponds to complete obscuration of the plane wave
by the half-plane, 5av = 0. Furthermore, the power density does not abruptly go to
;

zero as the point of observation goes from the illuminated side

side (tea

>

0); rather, there are fluctuations

(ica

<

0) to the

shadow

followed by a gradual decrease in power

density.

Another

illustration

of Huygens' principle occurs in holography.

photograph only amplitude information

is

recorded.

information and phase information are recorded.

illuminated with coherent light,

it

In an ordinary

In a hologram both amplitude

Thus, when the hologram

is

generates waves (in both amplitude and phase)

which, in accord with Huygens' principle, produce a three-dimensional picture, t

t Since this

1/Vr
t

is

D. Gabor,
Leith and
1972).

cylindrical (rather than spherical) wave, the field

(see Sec. 12-7),

A New

J.

and the power density must diminish

must diminish as

as 1/r.

Microscopic Principle, Nature, 161: 777 (May 15, 1948); E. N.

Upatnieks, Progress in Holography, Phys. Today, 25(3): 28-34 (March

468 ELECTROMAGNETICS

[CHAP. 12

Ray paths

Primary
wavefront

FIGURE
The

12-13

principle

/
of

geometrical

optics

wavefront

(ray paths).

12-7

Secondary

GEOMETRICAL-OPTICS CONCEPTS!

Imagine that a family of ray paths

is

drawn from each point on a primary wavefront to

The ray paths are

corresponding points on a secondary wavefront, as in Fig. 12-13.

perpendicular to the wavefronts and are in the direction of the Poynting vector at

each point.

Geometrical-optics theory uses a point-to-point ray correspondence

between two successive positions of a wavefront, in contrast to the physical-optics

theory of Sec. 12-6, which postulates a point-to-spherical-wave correspondence.
Put simply, physical optics involves wavefronts while geometrical optics involves
ray paths.
us reconsider the example of Sec. 12-6, where a plane wave
on a straight edge, as in Fig. 12-14. According to the theory of geometrical
optics, the power density on the right side of the edge will drop abruptly to zero (solid
line) as an observer goes from the illuminated region to the shadow region, in contrast
to the physical-optics solution (dashed line), which predicts fluctuations followed by
a gradual decrease. The geometrical-optics approximation can be considered as the

To

is

illustrate this, let

incident

1
high-frequency lhr' of the physical optics approximation.

The material

in this section is based

State University.

on

lecture notes

by Prof. Leon

Peters, Jr.,

Ohio

WAVE

sec. 12-7J

REFLECTION, REFRACTION,

AND DIFFRACTION

469

Distance

from edge

-Geometrical
optics

Illuminated

***- Physical

region

optics

Plane
Relative

wave

power density

Shadow
region

^Conducting
half plane

FIGURE 12-14
Comparison of geometrical- and
wave by a half-plane.

In geometrical optics

physical-optics solutions to diffraction of plane

assumed that the power flow between a

it is

set

of rays

In Fig. 12-15, for example, four rays from a point source are
passing through small concentric spherically shaped wavefronts of areas A y

remains constant.

shown
and A 2

A2

According to geometrical optics, the power passing through areas

must remain constant (conservation of power), i.e.,

A x and

5^=52^2

where

\E2

and
where

l^l
2Z

= power density at area A x

and

E2

= power density at area A 2

are the electric fields at

A t and A 2

Taking the

S2 /S, from

dance of the medium.

(2)

2
\

2Z

(1)

ratio

(3)

respectively,
(1)

and

and Z

is

the impe-

substituting (2)

and

(3),

we have
\E2

2
\

l*i|

_A

A2

Point source

FIGURE

12-15

Tube of rays from a point

source.

(4)

470 ELECTROMAGNETICS

[CHAP. 12

Caustic lines

FIGURE

12-16

tube

Astigmatic

of rays

from non-

spherical reflector.

Also we note that the ratio of areas

is

A2
Combining

(5)

and

(4),

we

(R,

(5)

+ rf

obtain

\E2

= \EA

R,
(6)

Rt+r

Taking into account the phase-delay factor, we have

E 2 =\E t
or

Ri
\

+ r)

(8)

= |iK j7Hil

(9)

We now

consider a

necessarily spherical.

(7)

E^E.-^-e-iK'

where

more

general situation in which the wavefronts are not

is shown with a
unequal to the radius of curvature in the

In Fig. 12-16, a curved reflecting surface

radius of curvature in the xy plane which

is

The incident rays 1, 2, 3, and 4 are all parallel to the x axis. After reflecand 2 then cross (focus) at P, rays 3 and 4 focus at P', rays 1 and 3 focus
and rays 2 and 4 focus at Q. The tube of rays is then said to be astigmatic,

xz plane.

tion, rays

at Q',

-j/KRi

Rt +


REFLECTION, REFRACTION,

WAVE

sec. 12-7]

AND DIFFRACTION

471

Caustic lines

FIGURE

12-17
for astigmatic tube of rays

Geometry

with caustic

and

lines

lines.

PP' and QQ' are called caustics. An astigmatic tube of rays is shown in
two wavefronts of areas A t and A 2 respectively. It follows that

Fig. 12-17 along with

R R2
i

A2

(R,

r)(R 2

(10)
r)

Substituting (10) into (4) gives

1/2

|2l

U+rxi +

|ll

Introducing the phase-delay term,

OO
r)J

we have
(12)

where
is

is

the complex electric field at the reference wavefront

special case of (12)

For

when

R2 = R

A x .\

Note

that (8)

cylindrical curved surfaces with axes parallel to the z axis, the radius of

curvature in the xz plane becomes

infinite,

so that

Rt =

For

oo.

this case, (12)

becomes

^G^P"*
t

(13)

in (12) assumes r > 0. If > r > R 2 a constant-phase factor of

90 is added to the exponential term of (12). If K 2 > r > Jf, a phase shift of
180 is added. If r = J?, or r = Ri (caustic points), the value of E% cannot be

The phase term

found by geometrical-optics methods.

472

ELECTROMAGNETICS

[CHAP. 12

In general, a geometrical-optics expression for the field can be written as

E=E

eJ * F(r) e~ J"

(14)

E = rectangular component of electric field, Vm"'

E = amplitude of field component at reference point, r = 0, V m -1
0o = phase of field component at reference point, r = 0, rad
F(r) = spatial attenuation factor, dimensionless
= 2n/X, m -1
r = distance from reference point, m

where

That

is,

E(r)

illustrate the geometrical-optics

conducting

strip

the strip, as
If

x attenuation factor x phase-delay factor

SCATTERING FROM A CONDUCTING STRIP

12-8

To

= reference field

of width D.

shown

concept,

in Fig. 12-18, so that the electric field

is

perpendicular to the page.

D > A, the geometrical-optics expression (12-7-13) can be used to find the field at a
The point

point P.

for the reflected

field

and

is

assumed to be sufficiently far from the strip and line source

be essentially parallel. From (12-7-13) the direct

direct rays to

is

= Ei
<

where
field at
|

us consider an infinite perfectly

let

A line source is located a distance d from the center of

R = OQ

and

point Q.

at

It is

QP, as shown

Id and A 1

2
|

= IR.

E =

\E

e~
\

is

equal to the

to a reference amplitude

2
^o=|i| ^,

O and

optics,

(2)

of length

perpendicular to the

Therefore

\Et\=Jl\E
Also

and where

According to geometrical

strip.

cylindrical wavefronts originating at

A =

(1)

in Fig. 12-18a,

a fixed point such as the center of the

where for

-jfro
'

conveneint to relate the amplitude \E 1

page,

JjT7

Substituting this

E* =lEol

and

(3)

(3) into (1) gives

Jirh

~ mro+ * )

(4)

WAVE

sec. 12-8]

REFLECTION, REFRACTION,

AND DIFFRACTION

473

Reflected ray

A/ Line
<y source

/.^^

Cylindrical

wavefront

(o)

Conducting strip

Reflected rays

'.

Line source

d-*-

FIGURE

12-18

(a) Infinite

conducting strip of width

perpendicular to page with infinite line

source parallel to
Line source and

It is

at distance d.

it

image

line

(b~)

source.

also convenient to refer the phase to the center of the strip

shown

r as

where

in Fig. 12-18a,

gives

E,=
Furthermore,

made

in the

if r

is

\E
\

= r + R + dcos 0.

P1

by defining the length

Substituting this into (4)

exp( jfir) exp(jpd cos 0)

large in comparison with R, the substitution

denominator, so that

(S)

(5)

R + r mr can

SiH^ol^-expC- jfir) exp(jpd cos 0)

which holds for
for

(it

O)

all

except those in the shadow of the

angles

7i

strip.

(6)

Thus

(6) is valid

where
O

For

be

becomes

= tan

outside this range (behind strip),

-x^/2

Ed = 0.

(7)

474

ELECTROMAGNETICS

Q'.

By the method of images

The

[CHAP. 12

cylindrical wavefront passing

through

(Sees. 7- 1 5

and

also intersects the reflected ray at

7- 1 6) the reflected

coming from an image

line

the line source at O.

Thus, the reflected field

source at O' (Fig. 12-186) which

is

ray can be viewed as

180 out of phase with

is

given by

Er = -\E \I- exp( -jpr)exp( -jftd cos 0)

-0 O

which holds over the range

range

|0|

<

EG = E + E,=

alone.

<

When

>

|0|

|0|

>

When n ^

is

0|

>

ER = 0.

given by the

Thus in the
(6) and (8):

sum of

j- exp(-jPr)[expUPdcos 0) - exp(-jpdcos 6)]

\E

=
In the range n

<,

the total geometrical-optics field

(8)

2/|

>

/- e-" sin

(^cos

the reflected field

o (incident-

and

is

0)

(9)

zero, so that

reflected-ray

EG

shadow

is

given by (6)

zone), the total

field is zero.

GEOMETRICAL THEORY OF DIFFRACTION

12-9

In recent years, a very powerful extension of geometrical optics has been developed,

known

as the geometrical theory of diffraction.

Its basis lies in

action of electromagnetic energy with a material

nomenon provided

the body

is

body

is

the fact that the inter-

essentially

a localized phe-

not too small in terms of a wavelength.

Thus, the

geometrical-optics field (Sec. 12-7) represents an important contribution to the fields

up by such an interaction. However, this must be supplemented by various other

The two most prominent supplementary contributors are those of
edge diffraction and creeping-wave diffraction. The latter represents the mechanism
by which energy is diffracted to the shadow region of a curved metallic surface such as
set

contributions.

a cylinder.f
In this section, attention

is

confined to edge diffraction.

to be a very special case, it has been

Although this appears

found to be applicable to a wide variety of geom-

and C. E. Ryan, Jr., Empirical Formulas for E-plane Creeping Waves

on General Smooth Conducting Bodies, IEEE Trans. Antennas Propag., AP-18:
432-434 (May, 1970).

f L. Peters, Jr.,

WAVE

sec. 12-9J

REFLECTION, REFRACTION,

AND DIFFRACTION

475

Incident ray

(b)

(o)

FIGURE

12-19

Cone of diffracted rays from edge of a conductor (6) plane of diffracted rays
when edge is hit by normally incident ray. [After J. B. Keller, J. Opt. Soc. Am.,
;

(a)

52 (2)

1 1 6-1

30 February

962).]

because of the local nature of the diffracted fields. The Law of Edge Diffraction
incident
states that in a homogeneous medium a diffracted ray and the corresponding
etries

ray

equal angles with the edge at the point of diffraction. They lie on opposite
12-19
the plane normal to the edge at the point of diffraction, as suggested in Fig.

make

sides

of

for a conducting half-plane.f

The

diffracted field

modifying (12-7-14),

E'D of a conducting half-plane can be found by

i.e.,

E'D

= 9\E

\er- ll2 e-'

= diffraction coefficient, m 1/2

E = amplitude of incident field at r = 0, V
= phase of incident field at r = 0, rad
r = distance from the edge, m
1
$ = 2n\k, m"

where

slightly

(1)

3)
t

m -1

<j>

If Pr

cient

is
is

large

and a plane wave

is

incident normally

on the edge, the

diffraction coeffi-

given approximately by

0= (sec

e -j*i*

ijlnfi

t J. B. Keller, Geometrical
(February 1962).

Theory of

esc

+ a\
I

Diffraction, /. Opt. Soc.

(2)

Am., 52: 116-130

476

ELECTROMAGNETICS

[CHAP. 12

Incident

Edge

FIGURE

12-20

Edge-on

view

of

plane

obstacle showing incident

conducting

and

Conductor
diffracted

ray.

where 6 and a are the angles shown

electric field is parallel to the edge,

parallel to the edge.

in Fig. 12-20.

and the minus

The approximation

where more complex expressions for the

(2) is

In (2) the plus sign

sign

is

used

if

is

used

the

if

the magnetic field

is

not valid near the shadow boundaries,

diffracted field

must be used.f

In order to illustrate the procedure for calculating the diffracted

us

field, let

return to the example of a conducting strip illuminated by a line source, as discussed in

Sec. 12-8.

and

2, as

The
shown

total diffracted field

is

From

in Fig. 12-21.

the

sum of the

diffracted fields

(1) it follows, therefore, that

from edges

the total diffracted

field is

(3)

where

2 U @ 2 = diffraction coefficients associated with edges

|

It

and 2

R = V(D/2) + d
, = amplitude of electric field at edges
2

follows from the geometry of Fig. 12-21 that rt

- (D/2) sin 0.

= r + (D/2)

sin

and

r2

Furthermore, from (12-8-3),

s,l-V!'*l

(4)

"
See
t A. Sommerfeld, Optics," pp. 258-262, Academic Press Inc., New York, 1954.
also P. M. Russo, R. C. Rudduck, and L. Peters, Jr.,
Method for Computing

E-plane Patterns of Horn Antennas,

(March

1965).

IEEE

Trans. Antennas Propag., AP-13: 219-224

WAVE

sec. 12-9]

REFLECTION, REFRACTION,

Incident-ray

477

,'Reflected-ray

shadow boundary

shadow boundary

FIGURE

AND piFFRACTION

12-21

Geometry for calculation of edge-diffracted field from a conducting strip illuminated by a line source.

When we

shadow boundary

note from Figs. 12-20 and 12-21 thatf

(Xj

(X

a2
a2

and using

\ Reflected-ray

Incident-ray /

shadow boundary /

(2),

0i
t

+ 0o
O

= +

the diffraction coefficients

(5)
(6)

- lit

(7)

- In

(8)

become

-JK/4

V* + sec 6 +

CSC-

(9)

4w
and

V2
z

^
4tc

The

sec

esc

2+ 0\
o

(10)

subscripts 1 and 2 on a and 8 refer to these angles as measured with respect to

edges 1 and 2, that is, a t is the same as a and 8, is the same as
in Fig. 12-20
(edge 1 in Fig. 12-21), while a 2 and 92 are related to the corresponding geometry
for edge 2.

478

ICHAP. 12

ELECTROMAGNfenCS

1.0

0.1

0.01

0.001

0.0001

boundary
Reflected shadow boundary-* Mncident shadow
1.0

0.1

\E

N.

Diffracted field

0.01

0.001

0.0001

1.0

0.1

Total field

\E\

0.01

0.001

0.0001

iso"

go-

FIGURE

12-22

Magnitudes of geometrical field (Ea ), diffracted field (ED), and total field (E)
for line source and conducting strip of Fig. 12-21 for the case where D = 10A
and d= A/4. Dashed lines are for region in which the approximation (12-9-12)
is

If r

and

not valid.

>

D, the square-root terms in

r2

[given before

(4)]

and

(4), (9),

become J^ x
and (10) into (3), we

J7 Jr.

(3)

Substituting

r,

get

'<*/4 + /m).-j7r

sec

sec

^-2 - esc ^jexpl -jp - sin 01

hzl + esc hj^ +Jp \ sin 0)]

(11)

WAVE

PROBLEMS

which

is

REFLECTION, REFRACTION,

AND DIFFRACTION

On

accurate except for angles near the shadow boundaries.

boundary, the

field

from (12-6-10)

The

the conducting strip.

diffracted field,

the

shadow

one-half the value of the field in the absence of

is

total electric field

is

sum of the geometric

the

field

and the

i.e.,

E = EG + ED
where

479

G is given
also Fig. 14-48.

by

ED by (1 1), as

and

(12-8-9) or (12-8-6)

(12)

shown

See

in Fig. 12-22.

PROBLEMS
*

A plane traveling wave in air with rms electric field E = 100 mV m" is incident
normally on a large body of saltwater with constants a = 3 O m~', p. = 1, e, = 80. If
_1
for
the constants are independent of frequency, find the depths at which E=\ jjN m
l

12-1

frequencies of (a) 10 kHz, (6)

MHz, and

100

(c)

MHz.

plane traveling wave in air (medium

with magnetic

Hi

is

incident

normally on the plane boundary of a large conducting medium (medium 2) with

intrinsic

12-2

impedance 10

4-y'lO

mO.

(a) Calculate

1)

field

without approximation the magnitude of the magnetic

H at the boundary in medium 2. (b) Compare this result with

that H = 2//i and determine the error involved in the approximation.

field

the approximation

12-3

Calculate the conductivity required for a sheet of space cloth

square) which

12-4

is 1

(Z

= 377

CI

per

mm thick.

half-space of air

(medium

and a

1)

half-space of dielectric

(medium

2) are

A plane linearly polarized 2-GHz traveling wave in medium

incident normally on the copper sheet. The
(air) with electric field =10Vm"' rms
= e = and
10 /xm thick. The
copper sheet has constants a = 58 MO m" and
=
=
e, = 5.
Determine the rms value of
a
constants for medium 2 (dielectric) are
0,
separated by a sheet of copper.

is

/*,

/*,

(a)

1,

it is

1,

just inside the copper sheet adjacent to medium 1 (air), (b) JSjust inside the copper sheet
medium 2 (dielectric), (c) E in medium 2 at a distance of 2 m from the copper

adjacent to
sheet,

and (d)

H at same point as in

(c).

A large flat sheet of ferrous-dielectric material is backed by aluminum foil.

MHz the constants of the ferrous-dielectric medium are p, = e, = 6 j6. How

12-5

At 500
thick must the sheet be for a 500-MHz wave (in air) incident on the sheet to be reduced upon
Take o- = 0.
reflection by 30 dB if the wave is incident normally.
*
on the
12-6 A linearly polarized plane traveling wave in air is incident at an angle
The constants of the dielectric medium
flat surface of a dielectric medium of large extent.
are a = 0, fi T = 1, e, = 8. Calculate the magnitude of the reflected field E, and the transmitted field E, relative to the incident field E, as a function of 0, when E, is (a) parallel and
(

(6) perpendicular to the

plane of incidence,

ordinates vs. 6 as abscissa,

t

(d)

What

(c)

value of

Prepare graphs showing E,/Et and E,\Ei as

0, is

the Brewster angle?

* Answers to starred problems are given

in

Appendix C.

480

[CHAP. 12

ELECTROMAGNETICS

T^

-*^b
~~^Receiving antenna
1

Xi

&2

Ground

v
1

r*~

Fig. P12-10

Microwave

12-7
the

flat

are

An

(a)

unpolarized plane traveling

wave

surface of a dielectric medium of large extent.

= 0,

reflected

;n r

= 1,

c,

= 8.

in air is incident at

an angle

circuit

on

0,

The constants of the dielectric medium

Calculate the ratio of the transmitted Poynting vector S, and the

Poynting vector S, relative to the incident Poynting vector S, as a function of

Prepare graphs showing S,/St and S,/Si as ordinates

vs. 6, as abscissa,

(b) Is there

8,

any

wave becomes (linearly) polarized?

Repeat Prob. 12-7 with circumstances reversed, i.e., with the unpolarized
wave originating in the dielectric medium, (b) Is there an angle or angles for which there
angle for which the reflected or transmitted
12-8

is

(a)

no transmitted wave?

12-9 (a) A 60-MHz radio transmitter which radiates isotropically is situated 20 m

below the surface of a deep freshwater lake with constants f*, = 1, e, = 80 j4. If the
and the polarization is circular, what is the ratio of the power received
power radiated is 100
per unit area at a radius of 1 km as a function of the zenith angle from to 90 ? The radial

distance

is

measured from a point on the water surface

directly

Express the power ratio in decibels and plot this ratio as ordinate
(6) Discuss the practical aspects

vs.

above the transmitter.

zenith angle as abscissa.

of radio communication over such paths,

(c)

Compare

the situation of this problem with transmission from the earth to an extraterrestrial point

above the ionosphere

12-10

(a)

a transmitter on a

(see

typical
tall

Chap.

15).

microwave communication

circuit for

AM, FM,

or

TV

between

building and a distant receiver involves two paths of transmission,

ground reflection (length n + r 2 ),

and d = 5 km. For a frequency of 100 MHz
calculate the ratio of the power received per unit area to the transmitted power as a function
of the height hi of the receiving antenna. Plot these results in decibels as abscissa vs. h 2 as
one a

direct path (length r

as suggested in Fig. P12-10.

and one an
Let hi

indirect path with

= 300 m

ordinate for three cases with. transmitting and receiving antennas both (a) vertically polarized,
(b) horizontally polarized,

and

(c) right circularly

polarized for h 2 values from

to 100

m.

and that the receiving antennas are also

Consider that the ground is flat and
aperture.
same
effective
all
the
isotropic and
have
(b) Compare the results for the three types of polarization and show
perfectly conducting,

Assume

that the transmitting antenna

is

isotropic

WAVE

that circular polarization

is

best

AND DIFFRACTION

REFLECTION, REFRACTION,

481

from the standpoint of both the noncriticalness of the height

hi and of the absence of echo or ghost signals.

Thus, for horizontal polarization the direct

and ground-reflected waves may reinforce at a certain height h 2 (path-length difference equals
an odd number of half wavelengths), but if this path-length difference yields delays of the
order of microseconds, objectionable ghost images can appear on a

TV screen,

(c)

Extend

the comparison of (b) to consider the effect of other buildings or structures which

may

produce additional paths of transmission.

12-11

0.0001 13

Repeat Prob. 12-10 for the case where the ground has constants a

(a)

m _1

fi,

= 1,

er

= 5.

(b)

What

difference in effect does the

ground have

in the

two

cases?

When a

12-12

12-13
strip

12-21

D = 5A illuminated

12-22.

\b) Repeat for

When

12-14

the

moon

by an

is 1

m.

if

infinite line

and

total field of an infinite conducting

source at a spacing

occults (passes in front of) a distant celestial radio source,

is

obtained, as in Fig. 12-12.

distance as 380

Mm.

(b)

Repeat

source (zero subtended angle) and compare with the pattern for
as a convolution (see Kraus,

"Radio Astronomy,"

and plot
The wavelength

(a) Calculate

the source subtends an angle of 8 seconds of arc.

Take the earth-moon

See Figs.

1A.

d = 5A.

a Fresnel type of diffraction pattern

the occupation pattern

traveling in air enters

the wavelength change but not the frequency?

(a) Calculate the geometrical, diffraction,

of width

and

wave of frequency /and wavelength A

radio

why does

denser medium,

of a point

(a) for the case

Part (a)

(a).

may be done

pp. 194-198) or by superimposing the

patterns of several point sources situated within an angle of 8 seconds of arc.

In Example

12-15

of Sec. 12-5 determine the axial ratio

the polarization ellipse of the transmitted

(/>)

wave

for the case where the angle of incidence

is

AR and the

(a) for the conditions

tilt

angle t of

of the problem and

increased to the Brewster angle,

(c) If

the

W m -2 what are the power densities of the reflected

power density of the incident wave is 5

and transmitted waves for cases a and bl
12-16

right circularly polarized

50 from the normal to a

AR,

flat dielectric

wave of 10

W m~

slab of large extent.

in air
If e p

is

incident at

=5

an angle of

for the dielectric find

and power density for (a) the reflected wave and (b) the wave
(c) Repeat parts (a) and (b) if the angle of incidence
adjusted to the Brewster angle, (d) In parts (a) and (6), is the wave left-or right-handed?

the axial ratio

tilt

angle

t,

transmitted into the dielectric slab,

is

12-17

Referring to the steps leading to (12-8-5)

show

that r

=r +R+d

cos

9.

13
TRANSMISSION LINES, WAVEGUIDES,

AND RESONATORS

13-1

INTRODUCTION

a close analogy between plane waves in unbounded media

and guided waves on transmission lines. The subject oftransmission lines is considered

As discussed earlier,

there

further in this chapter,

is

and additional comparisons are made between waves

and waves on transmission

transmission line

may be

energy from one point to another.

with

maximum

in space

lines.

defined as a device for transmitting or guiding

Usually

it is

desired that the energy be conveyed

efficiency, losses in heat or radiation being as small as possible.

Transmission lines

may be

on
Thus transmission

classify transmission lines

they can transmit.

of many forms and shapes.

the basis

of the
lines

It will

field configurations,

may be

be convenient to
or modes, which

divided into two main groups:

(1) those capable of transmitting transverse electromagnetic

those able to transmit only higher-order modes. In a

TEM

(TEM) modes and (2)

mode both the electric

and the magnetic fields are entirely transverse to the direction of propagation. There
Higher-order
is no component of either E or H in the direction of transmission.
direction of
in
the
component
one
field
least
have
at
always
modes, on the other hand,
transmission.

All two-conductor lines such as coaxial or two-wire transmission lines

TRANSMISSION UNES, WAVEGUIDES,

SEC. 13-2]

are examples of
dielectric

TEM-mode

mode types.
may be classified

transmission lines

as follows:

TEM-mode type E and H entirely transverse, e.g., all two-conductor types such
:

as coaxial

types, while hollow single-conductor waveguides or

rods are examples of higher

To summarize,

AND RESONATORS 483

and two-wire

Higher mode type

sion, e.g.,

lines

E or H or both have components in the direction of transmis-

hollow single-conductor waveguides and dielectric rods

In the above discussion

In

inclusive expression.

we have used "transmission

common

line" as a general,

all-

present-day usage, however, "line" or "trans-

TEM

modes,
mission line " is usually restricted to those devices which can transmit
while " guide " or " waveguide " is employed for those devices which can transmit only
higher-order modes.

13-2

COAXIAL, TWO- WIRE,

TRANSMISSION LINES

AND

INFINITE-PLANE

The most common forms of TEM-mode transmission lines are the coaxial and twowire types. Many other forms, in fact an infinite variety of them, are also possible.
However, all may be regarded as derived from a basic or parent form. Thus let us
consider the infinite-parallel-plane transmission line as the basic two-conductor type.

A cross
TEM wave, E

This type consists of two parallel-plane conducting sheets of infinite extent.

section through such a line
is

everywhere normal and

An

is

shown

Considering only a

in Fig. 13- la.

H everywhere parallel to the sheets.

line is provided by
Here the sheets have been

approximation to the infinite-parallel-plane transmission

the parallel-strip line

shown

in perspective in Fig. 13-16.

reduced to form long parallel

trayed by Fig. 13-lc.

strips

of width

b.

cross section of this line

In the region between the strips

However, E and H also extend outside

H lines forming loops that enclose each strip.

as for the infinite-sheet line.

between the

Now

strips,

let

the

the strips of the line of Fig. 13-lc be curved

the edges, as suggested by Fig. 13-W.

two-conductor transmission

As another

line

shown

Continuing

13-1/.

por-

the region

away from each other at

we end up with the

this process,

in Fig. 13-lc.

variation, let the strips of the line

direction, as suggested in Fig.

is

E and H are oriented the same

of Fig, 13-lc be bent

Continuing

this process,

we

in the

same

arrive at the

coaxial transmission line portrayed in Fig. 13-lg.

Thus we may regard both the two-wire transmission line (Fig. 13-le) and the
coaxial line (Fig. 13- Iff) as forms that can be derived from the parallel-plane type.

484

ELECTROMAGNETICS

[CHAP. 13

/H

line

line

Infinite

conducting
sheets

n
a \l

(6)

(c)

FIGURE

13-1

Evolution of two-wire and coaxial transmission lines from infinite-parallel-plane

type by means of transitional forms.

In the

first

part of this chapter the properties of two-conductor transmission lines

are developed by an extension of ordinary circuit theory to take into account the

finite

and comparisons are made with corresponding

relations for plane waves in space. The development applies to coaxial, two-wire, or
any two-conductor type of transmission line carrying only TEM waves. Next higherorder mode transmission is considered on a two-conductor infinite parallel-plane transmission line. This leads logically to wave transmission in a closed rectangular pipe,
or waveguide. Closed hollow guides of circular and other cross section are also
discussed.
Guiding waves by open surfaces is then considered. The remainder of
velocity of propagation along the line,

the chapter deals with resonators, in particular the cavity resonator.

AND RESONATORS

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-3]

485

T=tdV
V

FIGURE

13-2

Two-wire transmission

dx

line.

THE INFINITE UNIFORM TRANSMISSION LINE

13-3

Consider the uniform two-wire transmission

line

shown

in Fig.

In earlier

13-2.

chapters expressions were developed for the capacitance and inductance per unit
length of such a

line.

If the line is

shunt conductance that

not lossless,

it

have a series resistance and a

will also

may need to be considered.

The net effect of the

tance and inductance can be expressed by the series impedance

Z = R+jcoL = R+jX
R = series resistance, flm" 1
L = series inductance, H m _1
_1
co = radian frequency 2nf, s
1
X series reactance, Qm"

where

The net

effect

(1)

is

expressed by the

Thus

length.

(U

m _1

dx of the

line,

Y=G +ja>C = G +jB

where

Thusf

of the shunt conductance and capacitance

Y per unit

shunt admittance

(finr

series resis-

Z per unit length.

(2)

_1

G = shunt conductance, U m
C = shunt capacitance, F m _1
_1
co = radian frequency =2nf, s
_1
B = shunt susceptance, 15 m

Consider

now an

infinitesimal section

varying wave be present

current through the line

the section dx
section, or

is

on the
be /

equal to the

line.

(Fig. 13-2).

The

IZ change per unit

dV = IZdx where / is

and

let

a harmonically

Let the voltage across the line be

voltage change

dV over

V and

the

the length of

length multiplied by the length of the

the line current.

Thus

dV

r = IZ

(3)

dx

The change

in current dl between the ends of the section

dx

is

equal to the shunt

current VY flowing across the line from one wire to the other, multiplied by the length

of the section, or dl =

VY dx where V is the voltage across the line.

dx=

t Note that here

length.

if,

Thus

VY

G, X, B, L, C, Z, and

(4)

Y are

distributed quantities,

i.e.,

per unit

486

ELECTROMAGNETICS

[CHAP. 13

and

Differentiating (3)

with respect to x,

(4)

we

obtain

dZ dl
dZ
I-r + Z
=I
+ ZVY

dx 1

dx

dx

dx

(5)

-V Y -V
dx 2

dx

dx

(6)

dx

On a uniform line Z and Fare independent of x (do not vary along the line), and
so dZjdx and dYjdx are zero.f

Thus, for a uniform line (5) and (6) reduce to

-ZYV=0

(J)

-ZYI =

(8)

dx'

and
dx-

Equations

(7)

and

a uniform transmission

(8) are the basic differential equations,

line.

equations of the second order with constant coefficients.

way of expressing

or wave equations, for

In mathematical terminology they are linear differential

They are the most general

the natural law relating the variation of voltage and current with

distance along a uniform transmission line.

However, they tell us nothing specifically

about the voltage or current distribution on a particular transmission line. For this
we must first obtain a solution appropriate to the imposed conditions. As a trial
solution of (7) let us substitute

V = eyx
f

(9)

For a nonuniform (tapered) transmission

be retained. From (3)

so that the

first

line the

AV

dx

terms with dZ/dx and dYjdx must

(6a)

term on the right side of (5) can be written

1

dx

dZ dV

Z dx

dx

d(ln

Z) dV

dx

dx

(6b)

Substituting (66) in (S) gives

d1 V
dx 2
In a like

manner

(6)

rf(ln

Z) dV

dx

dx

ZYV=0

(6c)

ZYI=0

(6d)

becomes

d2 l
dx1

Y) dl
dx
dx

d(ln

Equations (6c) and (6d) are basic differential equations for a nonuniform transmission line. For a uniform line they reduce to (7) and (8).

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-3]

AND RESONATORS 487

from which

fl=y*e?*
= y2 V
2

(10)

ax

Thus,

(7)

becomes
2

(y

(11)

-Zy =

(12)

and

Equation

-ZY)e yx =

known

(12),

Jz~Y, so that the general solution for

V=
where

and

C2

two unequal roots

as the auxiliary equation, has

is

(7)

Q exp(,/zr x) + C

are constants.

Thus

+ y/ZY and

(9) is

exp( -

JZY x)

a solution provided y

(13)

= +>/ YZ.

same fashion as (7), we obtain a solution for / similar in

form to (13) but having two more constants. Instead of solving for / in this manner,
let us proceed along another avenue of approach and obtain a solution for /from (13).
If (8) is solved in the

To do

this, let (13)

be differentiated with respect to

=C
from which

it

is

obtain

(14)

-^= expUzYx) - -L= exp(- VZT x)

(15)

y/Z/Y

a solution for the current.

To

evaluate the constants,

we note from

(13) that

0,

V is the

+ C2

instantaneous voltage at the point

this voltage as the

sum of two voltages which,

Let

vary harmonically with time.

quantities

we

fZY expiJZY x) - C2y/ZY exp( y/ZYx) = IZ

V=C
where

(3),

ls

JZJY
when x =

Recalling also

follows that

This

x.

Q and C

with respect to time.

(16)

unequal in amplitude and

V2 be the amplitudes of the voltages. The

and

are constants with respect to

Thus we can put

= 0on the line. We may regard

in general, are

x but may be regarded

C = V^""
t

and

C2 = V2 eJ"".

as variables

Substituting

these into (13) and (15) yields

V=V

and

e*

'

exp( N/ZT x)

+ V2 e*

'

exp( -

JZY x)

I=-L=exp(JzYx)- ^=exp(-y/ZYx)
yJZ/Y
JzfY

(17)

(18)

488

ELECTROMAGNETICS

[CHAP. 13

The quantity y/ZY =

complex, with a real part

a,

called the phase constant.

is

called the propagation constant.

called the attenuation constant

In general

/?

(20)

Vz?

(21)

and

p = Im

Introducing (19) into (17) and (18) and rearranging,

y y

giXgHtot + fx)

guSgji.ait + Px)

(19)

= Re sJZY

or

is

Thus

y=^ZY= a +jfi

and

it

and an imaginary part

JZIY

_^_

we obtain

g-aXgjXmr-Zlx)

(22)

-3Xgj(.at-px)

(23)

y/Z/Y

Equation (22)

is the solution for the voltage on the transmission line.

The
two terms. The first term, involving cot + fix, represents a wave traveling
in the negative x direction along the line.
The magnitude of this wave at x = and
x
t =
is Vu and the factor e*
indicates that this wave decreases in magnitude as it

solution has

proceeds in the negative x direction.

wave traveling in the positive x

and t = is V2 and
at x =
,

magnitude as

it

proceeds in the positive x direction.

attenuation factors,
gKat-tx)

aTC

The

The second term, involving cot

a being the attenuation constant.

pnase fac tors,

fix,

represents a

direction along the line. The magnitude of this wave

the factor e "* indicates that this wave decreases in

The factors e" and e~*x are

The factors eJ^ mt+fx) and

being the phase constant.

(I

two terms, the first term representing a current

wave traveling in the negative x direction and the second term a current wave traveling
in the positive x direction.
The total current at any point is the resultant of the two
traveling- wave components.
solution for /in (23) also has

now to a single wave traveling in the negative x direction

by the first terms of (22) and (23), we note that V and / are identical
functions of x and t. The amplitudes differ. Taking the ratio of the voltage V
Confining our attention

as represented

across the line to the current /through the line for a single traveling wave,

the impedance

which

is

we

obtain

called the characteristic, or surge, impedance of the line.

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-3

That

AND RESONATORS

489

is,

(24)

This impedance

admittance

is

Y per

a function of the

impedance Z per unit length and shunt

Z and Y as in (1) and (2), we obtain, from

series

Expanding

unit length.

(24),

lR +jcoL
(O)

\G+jcoC

When R and G are small or when the frequency is large,

(25)

so that coL

>R

and coC

G,

(25) reduces to

(O)

(26)

Z = characteristic impedance, Q
_1
L = series inductance, H m
C = shunt capacitance, Fm"
In (26) Z is entirely real, or resistive, so that in this case we may, to be explicit,

where

speak of the characteristic resistance

of the

That

line.

is,

for this case

(27)

In general,

when

and

characteristic impedance, should

real

= R ) and the term

When R and G are

cannot be neglected,

be used.

However,

if

is

complex and the term

R and G are negligible, Z

may be used.
negligible, (25) may be

is

characteristic resistance

small, but not

reexpressed approxi-

mately in the following form:

Thus

for this case

is,

=/HG\2wC

in general, complex.

(28)

2coL}\

However,

if

G_R
C~
Z

is real.

This situation

is

Heaviside's condition for a distortionless line.

(29)

490

[CHAP. 13

ELECTROMAGNETICS

Table 13-1

CHARACTERISTIC IMPEDANCE OF
TRANSMISSION LINES

Condition

Characteristic impedance,

General case

z _ /_ /Z
+jaiC

Jc[ l+J \2a>C

Small losses

Lossless case,

0, G =
R=

Za =

2wl)\

=R

above for the characteristic impedance of a uniform

transmission line are summarized in Table 13-1.
The phase velocity v of a wave traveling on the line is given by co/fi. That is,

The

relations developed

<o

CO

(30)

If the line

is

lossless

or

R < coL and G

-4 toC,

(31)

= series inductance, H m
-1
C = shunt capacitance, F m

where L

13-4

COMPARISON OF CIRCUIT AND FIELD QUANTITIES

the
It is interesting to compare some of the relations for transmission lines developed in
preceding section with corresponding relations for waves developed in earlier chapters.
For example, consider the transmission-line equations

dV

= ZI =

(R +j(oL)I

(V

m -1

(1)

dx
dl_

and

=YV=(G+jcoC)V

dx

(Am" 1 )

(2)


TRANSMISSION LINES, WAVEGUIDES,

sec. 13-4]

AND RESONATORS 491

The corresponding relations for a plane wave traveling in the x direction with
are
the y direction, as obtained from Maxwell's two curl equations,
dEy

dx

(Vm" ! )

= -jconHz

(4)

dx

In these relations harmonic variation with time

with respect to time have been performed,

Comparing (1) with (3), we note that

Z=R

in

(3)

(Anr" 2 )

^L=-(a+jaie)Ey

and

+jcoL for the

corresponds to

assumed, and the differentiations

j<n\i in

Comparing (2) with

+ja>C for the line.

line.

Y =G

is

(4),

the

we

wave case corresponds to

wave

see that a +ja>e for the

Further comparisons of the circuit relations used for lines and the field relations
employed for waves are made in Table 13-2. The first column gives the circuit
quantity and the last column the corresponding field quantity.
indicates the SI units,

which are the same for corresponding

Table 13-2

The center column

and field relations.

circuit

COMPARISON OF CIRCUIT AND FIELD RELATIONS

Circuit relation

SI units

Field relation for space

Z=R+ju>L

flm-'

ju>p

Y=G+jwC

Urn"

a +j<D

for transmission line

Characteristic impedance

Velocity

= j=

Series inductance

=L

Shunt capacitance
Voltage

Distance

Current

Distance

=C

I1
l-f.

Intrinsic

impedance

ms _1

Velocity

Hm-

Permeability

Fm-

Permittivity

Vm"

Electric field

Am-'

1
,

ju.

=
=E

Magnetic field

=H

wave

492

ELECTROMAGNETICS

[CHAP. 13

CHARACTERISTIC-IMPEDANCE DETERMINATIONS

13-5

The correspondence of L
Thus,

if

space

is

to

unit length (or depth) of a

depth) of a

The

/i

and

to c (see Table 13-2)

divided up into transmission-line

cell,

while

is

cells,

\i is

is

of particular

interest.

the series inductance per

the shunt capacitance per unit length (or

cell.

and

transmission-line-cell concept (see Sees. 5-17

10-9)

is

of considerable

value in connection with the determination of the characteristic impedance of lossless

transmission lines operating in the
sion line portrayed in Fig. 13-3.

TEM mode.

For

instance, consider the transmis-

Let the space between the conductors be divided into

by graphical field-mapping methods.

curvilinear squares

Each square represents the

end of a transmission-line cell of characteristic impedance

characteristic impedance of the line is given by

Z Z
n

where

Z = y/Ji/e.

Then, the

(")

(1)

N = number of cells in series

= number of cells in parallel

Z = characteristic impedance of one cell ( = y/fi/e, where n is permeability and e
n

is

This method

permittivity of medium filling line)

may

be applied to two-conductor transmission lines of any shape. The

any shape can, also be obtained by

characteristics of lossless high-frequency lines of

a simple dc measurement.

For example,

if

we wish

to find the inductance or capaci-

tance per unit length or the characteristic impedance (or resistance) of the line
in Fig. 13-3

by

this

method, the conductor cross section

drawn

is

shown

to scalet with con-

ducting paint (such as silver paint) on a sheet of resistance (Teledeltos) paper of

uniform resistance

per square, as suggested in Fig. 13-4.

Then by connecting

the

Field cells

:i

FIGURE

Coaxial transmission line of noncircular

cross

^'

13-3

section

divided

into

curvilinear

- r^ *

-1

Inner conductor

squares.

t Since only the shape

size.

is

'^C''.Y,,-.

important, the cross section

v
Outer conductor

may be scaled

to

any convenient

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-5]

AND RESONATORS

/ Resistance
'

FIGURE

tance

Rm

is

Since

paper

Rs O/square

13-4

Determination of characteristic impedance of transmission line by simple dc

measurement.

terminals of an

493

ohmmeter

to the inner

Ohmmeter

and outer conductors, as

indicated, a dc resis-

measured.f

Rm =

(N/n)

Rs

RJR = N/n

and

Rm

follows that

(Hm-

L = n-f

C=e

it

R,

(Fm-

(2a)

(2b)

(2c)

Thus,

if

the line

is air-filled,

Jn/e =

^/fi /e

376.7 il

and

(2c)

becomes

376.7

Z = -=- Rn,
K

(O)

(3)

Hence,

if

" space paper" (R

s

376.7 il per square)

z =
and the ohmmeter reads

EXAMPLE
line

cells

shown

is

used

(Q)

directly the characteristic

(4)

impedance of the

line.

Find the characteristic impedance (or resistance) of the

in Fig. 13-5.

The

lossless coaxial

line is air-filled.

solution Dividing the space between the conductors into curvilinear squares or
by graphical field mapping, we obtain a total of 18.3 squares in parallel and 2 in series.
t Supposing that the resistance paper between the conductors were marked off in
curvilinear squares, Rm would be given by K
(N/n)R,, where JV= number of

squares in series, n
i.e.,

resistance of

=
= number of squares in parallel, and R = resistance per square,

one square.

494

[CHAP. 13

ELECTROMAGNETICS

FIGURE
Coaxial

13-5

transmission

with

line

18.3

and 2

transmission-line cells in parallel

Outer conductor

in series.

The

characteristic

impedance of the

impedance of each
line

of Fig. 13-5

cell is

CI

the characteristic

= - Zj=:r!= 376.7= 41.2 O

18.3

drawn to scale with conducting paint on a sheet of space paper

would be measured directly on an ohmmeter connected between the inner and outer

a cross section of this line

41.2

(1),

is

it

If

Hence, from

376.7 Q.

is

conductors.

When this

resistance-measurement method

is

applied to open types of line such

as a two-wire line, the sheet of resistance material should extend out to a distance

large compared with the line cross section if accurate results are to be obtained.
Although graphical, analog-computer, and digital-computer methods can be
applied to two-conductor lines of any shape, some configurations yield to a simple
Thus, for the case under consideration, where the characteristic impedcalculation.

that

is

ance

(O)

the value of

obtaining

(5)

can be determined by a knowledge of

L from

coaxial line (as in Fig. 13-5)

is

Z = i-

In

from
given by

and

(5-13-8)

H In e

and

for the line.

(3-18-2), the characteristic

= 0.367

(fi)

Thus,

impedance of a

(6)

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-5]

AND RESONATORS

FIGURE

495

13-6

Two-wire transmission line.

If there

is

no ferromagnetic material present,

/i

Z =-^E log-

a
b

For an

=
=
=

relative permittivity

(6)

reduces to

(fi)

(7)

fe.

where cr

and

ji

of medium

filling line

outside radius of inner conductor

inside radius of outer conductor

air-filled line er

1,

and

(7)

Z =

becomes

1381og-

(SI)

(8)

EXAMPLE
its

The

air-filled

coaxial line in Fig. 13-5 has a radius ratio b/a

= 2.

Find

characteristic impedance.

From (8)

solution

Z
The value obtained

=41.4

138 log 2

previously by graphical methods agrees well with this exact value.

In a similar way, the characteristic impedance can be obtained for a two-wire

as in Fig.

3-6.

Thus,

if

D > a,

_
0.73
a

Z =i
If there is

no ferromagnetic material present, n

276
D
=
=log

where er

relative permittivity

If the

n and

(Q)

(9)

(9)

reduces to

(Q)

(10)

of medium

D = center-to-center spacing (see Fig.

a

/* log

n\l

13-6)

radius of conductor (in

medium

is air,

er

same units as D)
and (10) becomes

Z = 2761og-

(Ci)

(11)

The

line,

we have

characteristic

impedances obtained above are summarized

in

Table

13-3.

496

[CHAP. 13

ELECTROMAGNETICS

CHARACTERISTIC IMPEDANCE OF COAXIAL AND TWO-WIRE

Table 13-3

LINES
Characteristic impedance,

Type of line
with

Coaxial

(filled

Coaxial

(air-filled)

Two-wire

(in

medium of relative

permittivity

z ^iil,

medium

of relative permittivity e r)

It is

38 log -

V r

(in air)

(D >

(see Fig.

this section that the line

is

3-5)

(see Fig 13-6)

"

2 = 276 log -

a)

assumed throughout

(see Fig 13-5)

z=-^llog

(D>a)
Two-wire

(see Fig.

lossless (or

R<

ml.

3-6)

and

G < (oC) and also that the currents are confined to the conductor surfaces to which the
small
This condition is approximated at high frequencies owing to the
radii refer.
depth of penetration.

This condition

by the use of thin-walled tubes.

It is

may

also

also

at

low frequencies

that the lines are

operating in the

be approximated

assumed

TEM mode.

13-6

THE TERMINATED UNIFORM TRANSMISSION LINE

length.
In Sec. 13-3 the line was considered to be of infinite

situation

where a

line

of characteristic impedance Z

is

Let us

now

analyze the

terminated in a load impedance

The load is at x = 0, and positive distance x is measured to the

The total voltage and total current are expressed as the resultant
left along the line.
on an infinite transmission
of two traveling waves moving in opposite directions as
regarded as the
However, on the terminated line the wave to the right may be
line.
reflected wave
the
wave, with
incident wave and the wave to the left as the reflected
Z
related to the incident wave by the load impedance L
between the
At a point on the line at a distance x from the load let the voltage
traveling
to the right
incident wave
wires and the current through one wire due to the
due
current
and
voltage
be the
be designated V and / respectively. Let V, and /,
voltage
resultant
The
wave traveling to the left that is reflected from the load.

ZL

as in Fig. 13-7.

to the

AND RESONATORS 497

TRANSMISSION UNES, WAVEGUIDES,

sec. 13-6]

Component waves
*

FIGURE

V at

a point on the

Incident

Reflected

wave

wave

wave

TV

13-7

Terminated transmission

That

Total

line is

i=0

line.

equal to the

sum of the

voltages

V and Vx

at the point.

is,

V= V + Vv
yx

imt

V = V e (the factor e
Vx = V e~ yx+Ji (the factor

where

is

e*

(1)

understood to be present)
a

'

is

understood

= propagation constant = a+jfi

= phase shift at load
load (x = 0) we have V = V and Vt =

ito

be present)

At the

V^e*

=V

L^, so that at

the load

(2)

where p

the reflection coefficient for voltage (dimensionless).

is

V=V (e'* + p
i

The resultant current J at a point on the

and 7 t at the point. That is,
i

where

=I

I
/t

It

follows that

e-y*)

line is

(3)

equal to the

sum of the

currents

=i +h

(4),

yx ~ JS

= iie -r*+M-*)
= phase difference

between current and voltage

At the load

/<=*.
'o
*o
where p

is

the reflection coefficient for current (dimensionless).

/

(5)

It

follows that

e-'V + Pie'n

(6)

498

[CHAP. 13

ELECTROMAGNETICS

Now p v and p may be expressed in terms of the characteristic impedance Z

and

the load impedance

ZL .f

note that at any point on the line

Thus we

/,
7 = K =
z
o
-r T~ l b=

Vo

while at the load (x

v, _

~T~~T'&

(7)

= 0)
(8)

It

from

follows

(4) that at

the load

_K

_ V _?i _

But

V= V + V

so

+V

V -

(9)

V,
(10)

zL
VJ V

fr

we have

v
Solving for

Vo

2L

yields

-Z

V,

z,

Vo

Zl + z

'

For real load impedances ZL ranging from to

In a similar way it can be shown that

Pi= - Z

(11)

Pv

oo,

p v ranges from

= ~Pv

+Z

-1

to

in value.

(12)

any point x on the line gives the impedance Zx at the point

relation (12) in (6)
looking toward the load. Taking this ratio and introducing the

The

for

/,

we

ratio V/i at

obtain

z
Noting

(7)

and

(11),

we can

fi.e-"

(13)

Pv e

reexpress this as

ZL + Z tanhyx

..ev

*=r fi*-

(1)

+ZL t&nhyx
Z

(14)

are (in general) also complex quantities, the

be omitted. If it becomes necessary to indicate
done by the use of bars (thus the
the absolute value, or modulus, of ZL this will be
2
2
v
Jfi
).
R
\Z
equals
L
L
absolute value of ZL

Although the impedances

dot over the letter

ZL

and

will, for simplicity,

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-6]

where Zx

impedance

x looking toward

at distance

AND RESONATORS

499

load, Ci

Z = characteristic impedance of line, Q

ZL = load impedance, CI
= propagation constant =
x = distance from load, m

a +jfi,

This

is

ZL

and

oo

z*=;

tanh yx

If the line

is

short-circuited,

ZL =

to be noted that, in general, y

is

cosh
tanh yx

cos

ocx

that the product of the

tanh yx

(16)

+ jfi).

Thus

+j cosh ax sin fix

px +j sinh ax sin fix

tanh ax +y'tan

(17)

fix

(18)

+j tanh ax tan /?x

impedance of the

line

when

it is

open-circuited and

short-circuited equals the square of the characteristic impedance,

7
where

(15)

(14) reduces to

complex (= a

it is

coth yx

sinh ax cos fix

tanh yx

Note

(14) reduces to

=Z

and

Zx = Z

when

Zx at a distance x from the load.

the general expression for the impedance

If the line is open-circuited,

It is

m _1

i.e.,

7 7

(19)

Zoc = Zx for open-circuited line (ZL = oo)

Z = Zx for short-circuited line (ZL = 0)
sc

If the line

is

lossless (a

0),

the above relations reduce to the following:

In general,

ZX =Z
When the

line isjjpen-circuited

(ZL

ZL +jZ tan fix

Z +jZL tan fix
oo),

-jZ
j tan

When the

line

is

short-circuited

(ZL

(20)

cot fix

(21)

fix

= 0),

Zx =jZ

tan fix

(22)

[CHAP. 13

ELECTROMAGNETICS

500

FIGURE

13-8

Terminated transmission

line.

We note that (19) is also fulfilled on the lossless line.

an

lines,

such as coaxial and two-wire

uniform
load

transmission line

ZL (see Fig.
On

They give the input impedance Zx of a

impedance Z terminated in a

lines.

x and

of length

characteristic

These relations are summarized

13-8).

in

Table

(VSWR)

a lossless line the voltage standing-wave ratio

VSWR =
It

Furthermore, the impedance for

open- or short-circuited lossless line is a pure reactance.

The impedance relations developed above apply to all uniform two-conductor

is

13-4.

given by

(23)

follows that

VSWR =

(24)

V*-V

l-(VJVQ )

But
Vt

VSWR =

and so

i
1

Table 13-4

+
-

l/U

(26)

\Pv\

INPUT IMPEDANCE OF TERMINATED TRANSMISSION LINEf

General case (a

Load condition

Any

(25)

\Pv\

value of load

ZL

== 0)

Lossless case (a

Zt +Z tan yx

"""""Zo+Zj.tanhy*

"x

,
"

ZL +JZo tan fix

Z +jZL tan fix

Open-circuited line

(ZL ==

00)

Zx = Za coth yx

Zx = jZ

Short-circuited line

(ZL -=

0)

Zx = Z

Z=jZ

y=a+J p where a = attenuation

tanh yx

constant in nepers per meter and

>
constant in radians per meter, where A is the wavelength.

= 0)

cot fix

tan fix

= 2jt/A = phase

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-6]

FIGURE

AND RESONATORS

501

13-9

Junction of transmission lines of different characteristic impedance.

where p is the reflection coefficient for voltage. This relation is identical with that
given by (10-10-16) for the VSWR of plane waves. Solving (26) for the magnitude of
the reflection coefficient gives
Ip "

It is

often of interest to

of the incident wave.

That

is,

This

VSWR I= vswrTT
,

know the voltage Vat the load in terms of the voltage V

is

given by the transmission coefficient for voltage t

at the load

V=i

or

*.

it

may

(28)

as suggested in Fig. 13-7 or Fig. 13-8,

be the impedance presented by another

as suggested in Fig. 13-9.

= -kyo

The load impedance may be a lumped element,

or

(27)

line

of characteristic impedance

In the latter case (28) gives the voltage

Zu

K of the wave trans-

mitted beyond the junction.

may be shown

It

that the coefficient *

is

related to

ZL and Z

by

2ZL

where

ZL = load impedance presented to line of characteristic impedance Z

pv

As

(29)

reflection coefficient for voltage

ZL ranges from
It

to oo, t ranges

from

to 2.

also follows that the transmission coefficient for current t t

*'-f-z^z- 1+ >'
As

ZL ranges from

is

given by

(30)

to oo, i varies from 2 to 0.

The relations for reflection and transmission coefficients developed in this section
are summarized in Table 13-5.
t

502

ELECTROMAGNETICS

[CHAP. 13

RELATIONS FOR REFLECTION

TRANSMISSION COEFFICIENTS

Table 13-5

AND

2 Z
-~zL +z
1.

Reflection coefficient for voltage

Reflection coefficient for current

P'~z + zL -

Transmission coefficient for voltage

^-z0+ zL ' x+p

Transmission coefficient for current

Voltage standing-wave ratio

Zq

Zl

2Zt

(VSWR)

2Z

'

'-z +zL

- 1+ P'

|p|

l-|p|

|p,|

l-|pi|

VSWR-1
Magnitude of reflection

13-7

coefficient

IP.I-

IP'I-VSWR +

TRANSMISSION-LINE CHARTS

Transmission-line calculations are often tremendously facilitated by the use of trans-

In particular, the rectangular and the Smith impedance charts

mission-line charts.

are extremely useful in making calculations

rectangular impedance chart

is

on uniform

lossless transmission lines.

illustrated in Fig. 13-lO.f

The rectangular

coordinates on this chart give the normalized resistance R as abscissa and the nor-

X as ordinate for points on the transmission line, while the closed

circles indicate the VSWR on the line and the partial circles the distance in wavelengths
malized reactance

from the load. The normalized resistance R n

divided by the characteristic resistance R of the

Rn =
The normalized reactance

H is

That

X divided by R

(dimensionless)

For methods of constructing

this chart see, for

1946.

or

(2)

example, M.I.T. Radar School

Reintjes (ed.), "Principles of Radar," pp. 8-64,

Company, New York,

is,

(1)

equal to the actual reactance

Staff, J. F.

equal to the actual resistance

line.

(dimensionless)

X
X =
t

is

McGraw-Hill Book

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-7]

AND RESONATORS

FIGURE

503

13-10

Rectangular impedance chart.

Thus, the normalized impedance Z

The chart can

Z = R

is

related to the actual impedance

hj

Z by

(dimensionless)

also be used for admittances, the normalized admittance

(3)

Yn being

given by

Yn = Gn +jBn
An

example

EXAMPLE
shown

will

be given to

reflected

Z = Ro = 100

wave

illustrate

(dimensionless)

(4)

the use of the rectangular chart.

Referring to the terminated transmission line with short-circuited stub

in Fig. 13-11, the load

impedance

=YR =

at

Q.

A (VSWR =

ZL = 150 +y50

fl.

The

line

Find the shortest values of

1).

and stubs have a characteristic

and d2 for which there is no

rfi

504

ELECTROMAGNETICS

[CHAP. 13

FIGURE

13-11

Terminated transmission line with single

matching stub. Both the stub position
rfi

solution

and

its

length

is

are adjustable.

The normalized impedance of

-?

The chart

d2

150
-

Ho

now entered at the point

the load

+/50

loo

is

= 1.5+;0.5

1.5 +y'0.5, as indicated

most of the rectangular and circular coordinate

is on the VSWR circle for which VSWR = 1 .77.

lines are

by Pi

in Fig. 13-12.

omitted in this

Hence, the

For

figure.

VSWR at B is

.77.

clarity

Point Pi
Since the

_^
r^N

0.58

0.194

\f

If P

*
i

Rn

-0.58

^VSWR = 1.77

1,

-2
To -oo

FIGURE

13-12

Worked example
chart.

of single-stub tuner (Fig. 13-11) using rectangular impedance

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-7]

stub

is

connected in parallel with the

line, it will

so moving from P, A/4 (90 electrical degrees) to

AND RESONATORS

505

be advantageous to work in admittances;

P2

(see Fig. 13-10 for distance scale) converts

Z, to a normalized admittance r =0.6 ./0.2 at the load.f Next, moving a distance

di =0.194A to the point A (or from P% to P 3 on the chart), the normalized admittance is
Y. = 1.0 + jO.58. If the stub length is now adjusted so that the stub presents a normalized

admittance of -y0.58 at A, the total admittance at

the line

is

(VSWR =

matched

normalized admittance Y.

that the distance required

Thus

= y'0.58

from a short

the required stub length

1.0+y'0.58 -/0.58

move from P 3

(pure susceptance)
circuit

(Y

is

= yo)

Pt
Ps

to

indicated at

1.0+/0and

on the

chart.

and we note

to obtain this value

is

0.167A.

is

rf2

The

is

This corresponds to a

1).

=0.167A

required distance of the stub from the load as obtained above

is

rf,=0.194A

We

note that matching with the stub can be accomplished from any point on the vertical

line containing

P*

Thus,

both di and d2 would be

A Smith charti

is

we might have moved beyond P 3

the

way

to

P6

However,

In this chart the rectangular diagram

illustrated in Fig. 13-13.

has been transformed (see Prob. 13-14) so that

The chart coordinates

circular periphery.

all

larger.

all

impedance values

fall

within the

give the normalized resistance and reac-

VSWR circles are usually not included but can be constructed as needed
with a compass centered on the center point of the chart. For example, the VSWR = 2
shown in Fig. 13-13. The VSWR unity at the center of the chart and infinity
circle
line are R = 1/
Intercepts of the VSWR circle with the Xn =
at the periphery.
VSWR and R = VSWR. The distance from the load indicated around the periphThe

tance.

is

is

is

from a load for which R < 1

and X = the impedance must correspond to some value on the straight line constructed from the center of the chart to the peripheral point marked 0. 1 wavelength
The fact that distance is proportional to angular position around the
in Fig. 13-13.
For

ery of the chart.

periphery

is

EXAMPLE

instance, at

a distance of

one of the principal advantages of

Repeat Example

A/ 10

this

type of chart.

using a Smith chart.

solution The points P u P % etc., on the rectangular impedance chart of Fig. 13-12
have been transferred to the Smith chart of Fig. 13-13. Thus, for Z* = 1.5 +/0.5 we enter
Transforming to admittances, we proceed to Pi, obtaining Y =0.6 y'0.2 at the
at Pi.
,

Moving A/4 on the chart transforms an impedance

same location on the line,

to an admittance (or vice versa)

at the
t P.

H. Smith, Transmission Line Calculator, Electronics, 12: 29-31 (January 1939).

506

ELECTROMAGNETICS

FIGURE

[CHAP. 13

13-13

Smith impedance chart and worked Example 2.

load.

We then move 0.194A along the line to A

Connecting the stub brings us to P*.

position of

P s (Yn = ;'0.58)

of the real

axis).

EXAMPLE

Fig. 13-14.

The

,4

on the chart or
(0.1 67 A)

and reading off the distance between

to Y

1.0

3-1 2).

+70.58.

obtained by noting the

is

position at which the stubs connect to the line

and d2

it

and 0.25A

(at right

end

(VSWR=1).

are adjustable.

is

fixed, as

This kind of arrangement

is

shown, but

called a double-

= 50 +J100 Q. The line and stubs have a characteristic impedance

Find the shortest values of di and dz such that there is no reflected wave

The load ZL

Z = Ro = 100 Q.
at

P3

Fig.

Consider the terminated line with the two short-circuited stubs portrayed in

the stub lengths, di

stub tuner.

or to

The stub length

(Same problem as in

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-7]

FIGURE

AND RESONATORS

507

13-14

Double-stub tuner with short-circuited

stubs.

solution

The normalized value of

the load impedance

50-K/100
=

is

0.5+7 1.0

100

The

chart (Fig. 13-15)

Constructing a

is

entered at this normalized impedance, as indicated by the point Pi.

VSWR curve

VSWR

through P,, we note that the

P u we

at

(Fig. 13-14)

is

4.6.

Pi halfway around the constant

VSWR circle from /,. Thus, the normalized load admittance is 0.4 -/0.8. Now, moving
clockwise along the constant VSWR circle from P % a distance A/4 away from the load (toward
Next, constructing the diametric line through

the generator),

we

arrive

back at

mittance of the main line

must be

zero,

we may

JYt Thus

locate

at the point

(looking toward the load)

on the

0.5 +/1.0.

is

anticipate the fact that the admittance of the

line the

normalized ad-

Since the reflection at

main

line at

(without

marked Ci (Fig. 13-15). Therefore,

fall on this circle rotated back
marked
C2
(counterclockwise) A/8 to the position indicated by the circle
total
admittance to move
will
the
cause
d
length
2
The admittance added by the stub of
on
the
circle C 2 we can
end
up
order
to
In
line.
constant-conductance
from Pi along a
to P 3 results
Moving
at
P
arriving
right,
the
A
to
at
P
or
arriving
the
to
left,
3
either
move
the stub of length

rf,

connected) must

at the junction of the stub of length

fall

rf 2

on the

circle

the admittance must

in shorter stubs ; so

we will make the stub of such length as to bring the total admittance to P3

This requires a stub admittance (pure susceptance) of

Yn =-j(1.0 - 0.14) =

-7O.86

A short-circuited stub has an infinite VSWR so that the admittance at points along the stub
are

on the

(point

s ).

circle at the periphery

of the chart.

At the short

circuit the

admittance

is infinite

Therefore, in order to present a value

Ps )
at P s )

y=-y'0.86 (point
(noting that the outside scale on the chart reads 0.25

the stub length must be given

by

di = 0.388 - 0.25 = 0.138A

We arrive back at P,
line lengths,

D on the line is A/4 from the

some other point on the chart.

because the point

we would

arrive at

load.

For other

508

ELECTROMAGNETICS

[CHAP. 13

FIGURE

13-15

Worked Example

Double stub tuner of

3.

Fig.

in order to

is

make

is

the line

To summarize,

is

= y'0.73

falls at

point

Pa

There-

=0.40 -0.25 =0.1 5A

this stub brings the total

P9 ) and

A value

given by
di

Connecting

equal to 100 +y'0 n.

fore the length of the stub

(point

Smith chart.

= 1.0 +y'0.73.

Y.

the total

actual impedance at

using

VSWR curve from P 3 to P 1 we find that the line

Hence a stub admittance of Yn =-/0.73 is required
normalized admittance at A equal to 1.0 +j0, and therefore the

Next, moving along the constant

admittance at

13-14

matched

admittance (or impedance) to the center of the chart

(VSWR =

1).

the required stub lengths are

dt

=0.15A

d2 =0.138X

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-8J

We

AND RESONATORS

509

note from Figs. 13-14 and 13-15 that

at

(4.6

VSWR

at

to

P* instead of

(point

on stubs

00

we had moved

and Pi)

II

If

B (points P and P2 )
D and A (points P 3

between

2.05

to

P 3 we
,

P9 )
P6

(points

and Pa)

would have ended up with longer

stubs,

namely,

rf,=0.443A

and also a

13-8

larger

A/4

and

rf2

=0.364A

VSWR between D and A.

TRANSFORMER

There are many situations where a A/4 section of transmission line of suitable impedance may be useful. Suppose that we wish to connect a transmission line of 100 12
This can be done
characteristic impedance to a load of 400 +j0 12, as in Fig. 13-16.
with a A/4 transformer of proper characteristic impedance Z,
note that

when x =

A/4 (fix

= it/2) (line

Zx =
Here Z t
while

is

Zx

the geometric

assumed

must be equal to 100

+jQ

CI.

It

z, = 7*00

(13-6-20)

In the present example

we

(1)

ZL = 400 + y'0 11,

follows that

(100)

200

Therefore, a A/4 section of line of characteristic impedance Z,

the desired transformation, eliminating a reflected wave on

An

From

Z,=4%ZX

or

mean ofZL and Zx

lossless),

= Rt =

interesting application of the A/4-transformer principle

which can be used to eliminate plane-wave

200

12

provides

the 100-12 line.

is

to the A/4 plate,

Thus, for example, a plane

reflection.

1
z = iooa

z,

rjz*!.=
t =40o+joa
'

FIGURE

13-16

A/4 transformer.

510

ELECTROMAGNETICS

[CHAP. 13

A/4 plate

{Dielectric
Air

e,

=l

!t r

medium 1

= 4|
(a)

\*^*\
Analogous transmission

Z = 376.7fJ

Z[=

line

Z 2 =188fl

26612

-X/4 plates-

Dielectric

medium

Air

=1

Air

r=2

<

=4

, = 2.

r=l

(b)

<^4 |

K^H
Analogous transmission

= 376.7Q Z,=
266

FIGURE

line

Zi=

= 376.78

266

J)

13-17

(a) A/4 plate at interface

of air and dielectric

medium and

(b)

two A/4 plates for

transmission through a dielectric medium.

wave

in air incident normally on a half-space filled with a lossless dielectric medium

of relative permittivity e r = 4 will be partly reflected and partly transmitted.
The
reflection can be eliminated, as shown in Fig. 13-1 la, by placing a A/4
plate between

the air and the dielectric

medium provided
Zy

where

the plate has an intrinsic impedance

^JZqZ.!

Z = intrinsic impedance of air

Z2 = intrinsic impedance of dielectric medium

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-9]

AND RESONATORS

511

In the present example,

Z = h- =
V

376.7 1

Therefore the intrinsic impedance of the plate must be

Z = Jz Z2 = 266 n
t

and

must be

relative permittivity

its

-(I)-

The analogous transmission-line equivalent is also shown in

no ferromagnetic material is present, and so /i = n

that

plate

It is

3- 1 la.

It is

assumed

to be noted that the

A/4 thick as measured in terms of the wavelength in the plate.

is

For

reflectionless transmission

finite thickness,

a A/4 plate

tric-air interface.

is

of a plane wave through a dielectric

medium of

required on both sides to prevent reflection at each dielec-

The arrangement

transmission-line equivalent

13-9

Fig.

is

is

suggested in Fig. 13-176, where the analogous

also shown.

TRANSFORMER BANDWIDTH

With the A/4 transformer discussed in the preceding section (see Fig. 13-16) the
wave is eliminated at the design frequency (or wavelength). However, there

reflected

be reflection

will

400

device. Referring to Sec.

+ /On, Zj = 2000,
Suppose that

300

In other words, the transformer

at adjacent frequencies.

frequency-sensitive

and

Z =

system

this

13-8

and

Fig. 13-16

we have

is

ZL =

loon.

(UHF)

operating at ultrahigh frequency

is

with

MHz as the design frequency for which reflection is to be zero (line assumed loss-

less).

At 300 MHz, A

chart, as in Fig. 13-18,

m, and a A/4

we enter at

section

is

250

mm

long.

Using a Smith

the normalized impedance of the load,

i.e.,

normal-

impedance of the A/4 section, which is equal to

P on the chart. At 300 MHz the A/4
impedance
at the input (or left) end of the
section is 0.25A long, so that to find the
along
the VSWR = 2 circle A/4, which
transformer section we move clockwise from P
ized with respect to the characteristic

is

+j0 = (400 + j0)/200].

[

halfway around.

impedance

This

is

the point

This point, marked 300

at the input,

which

is

0.5

+j0.

MHz in Fig.

The

actual

13-18, gives the normalized

impedance

plied by the characteristic impedance of the transformer section

is

= 200

this value multifi)

or 1 00 +J0

CI.

512

ELECTROMAGNETICS

[CHAP. 13

FIGURE

13-18

Smith chart analysis of performance of A/4 transformer over a band of frequencies.

On

the Smith chart the normalized impedance with respect to the 100-fi line is then
+J0, which is at the center of the chart. Thus, the system is matched and the
VSWR = 1 on the 100-fi line.
1

Suppose we wish to find the

cies in the

VSWR

range between 200 and 400

At

this

less

than A/4 long.

frequency the wavelength

Its

is

length at 200

entering the Smith chart at P,

of this A/4 transformer at adjacent frequen-

MHz.

Consider first the situation at 200

longer,

MHz

and the transformer section

is 0.1

67A

MHz.
now

is

=0.25/(300/200)], so that again

we move only 0.1 67A along the VSWR = 2 circle to

13-18.
To refer this normalized impedance to

the

marked 200 in Fig.

the
100-fi line we multiply by 2 (as before), which moves us to the point P'
The resulting
VSWR on the 100-Q line is thus about 2.1. Calculating the VSWR at other frepoint

quencies in the range 200 to 400 MHz, as shown on the Smith chart of Fig. 1 3- 1 8, we
can draw a VSWR-vs.-frequency curve as in Fig. 13-19 (solid line). We note that the

VSWR =

(zero reflection) at 300

MHz

but that the

VSWR

rises to

more than 2

at

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-9]

2.0

AND RESONATORS

513

-\

Single X/4
\transformer

1.8

1.6

1.4

\\

\
\
\

Double */4
transformer

200

\\

225

250

i
275

i.

300

is

^_^_

325

350

375

400

MHz

VSWR

with single and double A/4 transformers over a 2 to

A/4 transformer has a much wider bandwidth.

The double

bandwidth.

1.2

13-19

Comparison of

200 and 400

Frequency,

FIGURE

\\

N\
1.0

1.2

/
/

MHz. As indicated in the figure, the bandwidth for which the VSWR <>
MHz. This yields a bandwidth with respect to the design frequency of

about 50

Us x
One method of

requirement of the geometric mean.

impedance seen looking either way

VSWR

SI

16.7%

increasing the bandwidth of the transformer

XIA transformers in series, as in Fig.

and 283

100

[=^200 (400)]

13-20a.

Thus, >/400 (100)

at the

midpoint A.

It is

The

to connect

two

satisfy the

200 Q, which should be the

Then, 141.4

= ^100 (200)]

are the impedances required for the two sections.

for this double (or two-section) A/4 transformer

in Fig. 13-19.f

is

The impedances chosen

is

shown by the dashed

The
line

evident that this combination possesses considerably greater

VSWR can be found

283] and

+y0 [=

(400 +j0)/

impedance at A (Fig. 13-20a), as above for the single X/4 section.

then repeated for the second A/4 section to get the impedance at the

circle to get the

The process
input.

with a Smith chart by entering at 1.41

moving clockwise the length of the transformer on a constant-VSWR

is

514

ELECTROMAGNETICS

[CHAP. 13

A/4-

A/4-

FIGURE

Zl

= 400+?'0fl
(a)

13-20

Double A/4 transformer dimensions and impedances, (6) constant conductor

spacing but different conductor diameter, and (c) constant conductor diameter but

(a)

different spacing.

The bandwidth

bandwidth.

for

which the

VSWR <

1.2

about

is.

50

MHz = 375 (

225), or
-Ll^ x 100
TOO"

= 50%

Over the 200- to 400-MHz range the

VSWR does not exceed about 1 .5 for the two-section type. With a two-wire transmission line, shown schematically in Fig. 13-20a, the different impedances might be

compared to

16.7 percent for the single section.

achieved in practice with a constant spacing but different conductor diameter, as in

Fig. 13-206, or with a constant

13-20c, or

by using

dielectrics

conductor diameter and different spacing, as in Fig.

of different permittivity between the conductors, or

some combination of these methods.

With a transformer consisting of three A/4 sections connected in series the bandwidth could be increased further.! As we add still more sections, the discontinuity
(impedance step) between sections becomes less, until in the limit we would approach
an infinitely long, smooth, gradually tapered transmission line with no reflection over
an

infinite

bandwidth.

To

(We are assuming, of course, that the load impedance remains

justify the

three sections

ratios are related as the coefficients

1, 3, 3, 1

for three sections;

1,

CI, i.e.,

of the binomial

4, 6, 4,

Z Z

and 3 of the
t ,
2
the logarithms of the impedance

geometric-mean requirement the impedances

would be 119, 200, and 336

series (1, 2,

for four sections, etc.).

for

two

sections;

Thus, we have the

requirement that

_ = og- = logZ-t

3 log

from which

Zt =

119,

Z% = 200,

and

400

Z-$

Z% = 336 A.

Transmission," pp. 57-61, McGraw-Hill

3 log

See

J.

C. Slater, " Microwave

Book Company, New York.

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-10).

AND RESONATORS

515

(a)

(6)

-lXormore-

FIGURE

13-21

Transformers with gradual taper for wide bandwidth.

hence

finite

However,

bandwidth.

we must

In practice

constant with frequency.)

if

deal with lines of finite length

the transition

made smoothly and

is

and

gradually

by changing the conductor diameter or the spacing, as suggested in Fig. 13-2 la and
The
Fig. 13-216, it may be possible to achieve a very wide bandwidth in practice.

impedance of the

characteristic

line

can be

made

to follow an exponential taper.

necessary, however, that the transition be at least

of operation (see Prob. 13-12).

We may conclude that

into a long gradual transition will provide

It is

long at the lowest frequency

spreading out a discontinuity

broadband matching.

WAVE REFLECTIONS ON A

13-10

A/4

TRANSFORMER

The reflectionless property of the A/4 transformer is achieved by adjusting reflections

from two points (one at each end of the transformer) so that they balance out at the
design frequency. Let us investigate this in more detail using the wave-reflection
method.
Consider again the A/4 transformer of Sec. 13-8 but with a short section of 400-Q

between the A/4 transformer and the load, as suggested

line inserted

in Fig. 13-22.

This permits us to discuss waves to the right of junction 2 without altering the problem.

We need to consider three waves: A approaching junction from the left, B approaching junction from the right, and C approaching junction 2 from the left. In a general
1

situation
right,

load).

we would

also need to consider a

but here this wave

The

reflection

is

wave

D approaching junction 2 from the

absent since the load and line are matched (no reflection at

and transmission

coefficients (p

and

t) for

each wave are as

indicated in Fig. 13-22.

Assume

that the initial incoming

lines are lossless.

Then

wave from the

at junction 1,

will

left is

equal to

be transmitted and \

and that the

reflected.

516

ELECTROMAGNETICS

[CHAP. 13

1000

200Q

IV-

400Q
-

40011

r=i

Residual

0.333V

i_

=0.037

a80

4
27

=1 ,777V

Total

=1-975

& =0.0041

--

ri^aoi

'

81

4
243
4

5^=0.0046

etc.

FIGURE
Wave

the

f-V wave

arrives at junction 2, four-thirds of

one-third reflected, giving rise to a

traveling back to the left
arrives

= 1.997

^J

729

4
"2187

When

1456
729

16

729

back at junction

-^-V wave proceeding

from junction
it

will

2, as in Fig.

13-22

reflections

it

will

on a A/4 transformer.

be transmitted and

to the load

13-22.

When

and a f-V wave

this f~V wave

be reversed in phase (having traveled A/2, or 1 80) and

two-thirds of it will be transmitted and one-third reflected (with sign change) resulting
in a

r-\ wave traveling to the left from

as indicated in Fig. 13-22.

junction

and a +-jV~V wave to the

right,

This ^--V wave will travel to junction 2 and be partly

transmitted and partly reflected, and the process will continue indefinitely.

Values of the voltages transmitted and reflected at each junction for several
reflections are

column

shown

in Fig. 13-22 (proceeding

the residual voltage of the

lists

wave

from top to bottom). The left-hand

from junction 1.

traveling to the left

Except for the i-V reflected wave (due to the incident 1-V wave) all waves traveling to
the left have negative voltage and after each round trip of the waves on the A/4 section
the residual

wave becomes

the second, and 1/2187

must wait an

infinitely

to exactly zero.
after three

round

less.

It is

1/27

after the third.

V after

It is

the

first

round

trip,

1/243

long time before the reflected wave on the 100-fi line

But as a practical matter

trips

it

V after

evident that, in principle at least,

will

is

V (= 1/2187)
MHz (A = m) the

be reduced to 0.00046

of the wave on the transformer.

At 300

we

reduced

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-11]

A/4 transformer
for one round
trips

is

5 ns.

250

trip

is

Still

and

intervals,

is

after

mm long, and for a lossless line (v = c = 300 Mm s"

16.7

x 10

-10

sufficient

time we

been achieved and the reflection

The
Waves are

(=0.5/3 x 10

lower levels of reflected wave

AND RESONATORS

is

),

517

the time

Or 1.67 ns; and for three round

will

be achieved after longer time

steady-state conditions have

may assume that

substantially zero.f

action of the A/4 transformer

is

like that

of a resonator (see Sec. 13-25).

back and forth between the junctions, the wave traveling to the
1.5
V and the one traveling to the left being equal to 0.5 V. The
right being equal to
student can confirm these values by adding up the wave fractions in Fig. 13-22. The
resulting

As

reflected

VSWR = 2 on the A/4 section.

indicated in Fig. 13-22, the

load) will

add up to 2

after

wave

to the right of junction 2 (traveling into the

sufficient

number of wave

reflections

on the

trans-

former. Thus, the A/4 transformer transforms the 1-V input wave to a 2-V wave into
the load. It may also be said that the transformer makes the 400-fi load appear as

100

Q at

13-11

its

input (junction

1).

FURTHER TIME-DOMAIN STUDIES

OF A A/4 TRANSFORMER

The behavior of the A/4 transformer

as a function of time, discussed in the preceding

can be investigated further with the aid of time-distance diagrams, as in

Assume an incident wave of frequency / which is a step function of
Fig. 13-23.
magnitude 1 V arriving at junction 1 at time t = 0. This unit step function is given by

section,

V(i)

f1

for

>

\0

for

<

Proceeding from top to bottom in Fig. 13-23,

we can

follow the progress of the trans-

mitted and reflected traveling waves at successive instants of time leading ultimately
to the steady-state condition shown at the bottom of the figure. The ordinate gives
the voltage magnitude of the traveling waves except for the lowest graph, where it
gives the magnitude of the total voltage. Note the standing wave on the transformer.

Consider next that a square pulse of duration f t and magnitude 1 V is applied.^

Proceeding from top to bottom in Fig. 13-24, we can follow the progress of this pulse
It is clear that the transformer
at successive instants of time until it has disappeared.
t

The time for two round trips is the same as for one cycle.
condition is substantially achieved within a few cycles.

The

pulse

is

given by

VW

(1

forO</<f!
forr<0and />r t

Thus, the steady-state

'

518

ELECTROMAGNETICS

[CHAP. 13

Junction

100

Junction

2008

t=-a /V

^400

4O0Q

i-^

-J(=\/4)-

t=x,/V
.,."

-6^1 /qsiS-kiJiWaaJ

=(*,+

*W

'*:' Vl|A

.'-.

-^3

1/3

--

^r*

-~t

ii

4/9"-

<=(* 3 +2/)/V

;"
i

;>

:!'"*

'

'

''"

'

jj .;

'"

'

''.'..^i6-j

:,

**0"

3S4/9,... , ,.-

t=(x 3 +3l)/V
'

.....

Bfel/27.,:^

t=(x 2 +4l)/V

.40/81,! ,,J*

;uuM.i.,,ti,iMii!i!!i5a

.:^'--::-^364^*':-.---- .-.
,

:-.:^r-:^^i/2431,!-

?43

'

'''
.

ii

ii

,,

-.-.:-.':-.

v-

-.. -:-

. ...--.
,

"40/811

matutm

WtM

.v.-.'.

-,

',

',.

V'

160 ;

4(liJi

'
.i.'..27

**

81 V

':'

.;
,

**0.,.,

,-,,

-,w

= a>
?

'

~ !"'
'

F.
!>V/
:'M
f
T72iv;;.

'

"'
''

""

"^
'
'
'

"

'!

*-i

^4jq:. ..,,:...:

f ..^-

Standing wave
.

=<o

IV VSWR = 1

FIGURE

= lV

VSWR = 1

2V

13-23

Time-domain analysis of A/4 transformer showing progress of wave along transmission line and effect of two junctions. Waves traveling to right are shaded
light and those to left dark.
Time increases by successive intervals except for
lowest graph which gives total voltage (sum of both traveling waves).

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-11]

AND RESONATORS

519

4000

FIGURE 13-24
reflected
Progress of short pulse along transmission line with resulting pulses
and transmitted at junctions, (a) Situation at successive instants.
on continuous basis. Note that an incident pulse splits into a
transmitted pulse at each junction.

(6) Situation

reflected

and

A CRT display connected to the line at the

would show a series of echo pulses of decreasing magnitude: The first

echo from junction 1, the second from junction 2, the third from 2, 1, and 2, the
each enfourth from 2, 1, 2, 1, and 2, etc. The magnitude of the pulses after
the
counter with a junction is indicated in (o) by the t, p values. For example,
and then
the pulse traveling through junction 1, reflected from junction 2
transmission
reflected from junction 1, has a magnitude rp'p'V where t is the
left

coefficient

bottom of

at
1 (left to right), p' is the reflection coefficient
the reflection coefficient at junction 1 (right to left) [see

through junction

junction 2, and p"

(a)

and

is

(b)].

Fig. 13-24
to function as a reflectionless device for a pulse as short as indicated in
There ds
velocity).
wave
the
and
v
is
transformer
of
the
length
/
is
the
where
l/v,

fails
(f

no way

for the pulse reflected

function

and reduce

its

from junction 2 to catch up with the one reflected from

Only for the steady-state condition or for a

magnitude.!

line is of
t Hence, a Smith chart analysis of a
very short pulses.

no value

if

the line

is

to be used with

520

ELECTROMAGNETICS

[CHAP. 13

Same power through each square:

hence same power through two
^shaded squares

\i

Two conductor transmission line

(Figure 13-25a)

(t 1 > l/v) can the multiple reflections reduce the wave reflected to the
from junction 1 to its desired low value. To avoid reflections with short pulses
one must construct lines without discontinuities. Conversely, one can use short pulses
to locate and measure the magnitude of discontinuities on transmission lines. To
obtain short enough pulses for such diagnostic measurements, it may be desirable
to use frequencies which are much higher than the design frequency.

very long pulse

left

13-12

The

POWER FLOW ON A TRANSMISSION LINE

electric

and magnetic

fields

around a two-conductor transmission

in Fig. 13-25a (see also Fig. 3-16).

(perpendicular to page)

is

At any point the power flow

given by the Poynting vector.

line

of Fig. 13-25a,

we can

shown

E/H = Z (=^Jno/e =
= ' 2 /Z ) is proportional to

But

376.7 1 for air or vacuum), so that the Poynting vector

the square of the electric field E.

line are

parallel to the line

Calculating the Poynting vector for the transmission

obtain the power-flow

map

of Fig. 13-256.

The contours

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-12]

AND RESONATORS

521

(Figure 13-256)

FIGURE

13-25

(a) Electric field

map

for infinite 2-conductor lossless transmission line with

rms. (b) Power-flow map; the contours give the

potential difference of 10

The same power is

in watts per square meter.
transmitted through each curvilinear square in (a). The two conductors are
between centers. The maximum
25.4
in diameter and are spaced 76.2
power density is between the conductors and close to each one as shown by
power density (Poynting vector)

mm

mm

the contours in (b) and by the fact that the smallest curvilinear squares are so
located in

indicate

25.4

meter for the case where each conductor is

and a potential differdiameter with spacing between centers of 76.2

power flow

mm in

(a).

in watts per square

mm

ence between conductors of 10

V rms.

It is

interesting to note that the

maximum

Inteconcentrated between the two conductors and

conductors
perpendicular
the
to
grating the Poynting vector over an infinite plane
yields the total power transmitted. The total power can also be obtained from the

power flow

close to each one.

is

V and the characteristic

(= V jZc = 10 2 /215).

interconductor voltage
case 465

mW

impedance

Zc

of the

line,

or in this

522

ELECTROMAGNETICS

The

[CHAP. 13

map of Fig. 13-25a has been drawn with true curvilinear squares (see
An important property of this map is that the same power is transmitted

field

Sec. 3-22).

Thus, if we know the power through any square and

number of squares, we can obtain the total power transmitted. Or conversely,
from a knowledge of the total power and the number of squares we can obtain the
power per square. In the present case there are 3 x 5% = 15J squares in each quadthrough each curvilinear square.

the

rant, or

a total of 63 squares.

infinity,

appear at

power per square

(Note that

least partially in Fig.

is

7.4

At low frequencies a concept of

practical.

all

squares, including those which extend to

3-25a.)

The total power is 465

mW, and the

mW (=465/63).

CIRCUITS, LINES,

13-13

AND

GUIDES: A COMPARISON

currents, voltages,

and lumped

circuit

elements

Thus, for the simple circuit of Fig. 13-26a, consisting of a generator

R, circuit theory involving lumped elements can be used.

At somewhat higher frequencies these ideas can be extended

is

G and

resistor

lines

tributed constants of the line are considered.

of Fig.

line

satisfactorily to

of considerable length provided that the velocity of propagation and the

dis-

Thus, the behavior of the transmission

3-266 can be handled by an extension of circuit theory involving distribu-

ted elements, as in the preceding sections.

Consider

now

another type of transmission system, as shown in Fig. 13-26c,

consisting of a hollow cylindrical or rectangular pipe or tube of metal.

ask can such a pipe convey electromagnetic energy ?

:

If our

to simple circuits or transmission lines, as in Fig. 13-26a

Suppose we

experience were restricted

and

b,

our answer would be

(a)

,<g
Sim pie circuit

><

?#

fto

Transmission

(W

line

(c)

Waveguide

FIGURE

13-26

Comparison of

(a) circuit, (6)

single-conductor waveguide.

two-conductor transmission

line,

and

(c)

hollow

TRANSMISSION UNES, WAVEGUIDES,

sec. 13-14]

AND RESONATORS

no, since there is only a single conductor and no return circuit for the current.
if

523

However,

our experience were restricted to optics, our answer would be yes, since

light will

pass through a straight metal pipe and light consists of electromagnetic waves of

extremely high frequency (10

The complete answer

16

is

Hz).

yes and no, depending on the frequency.

reasoning further, we might deduce that

cies

but

will

if the

metal pipe

will

transmit extremely high frequencies, there must be

frequency at which there

is

Carrying our

not transmit low frequen-

some intermediate

a transition from one condition to the other.

In the

on waveguides we shall find that this transition, or low-frequency

occurs when the wavelength is of the same order of magnitude as the diameter

following sections
cutoff,

of the pipe.
In explaining the transmission of electromagnetic energy through the pipe of
Fig. 13-26c

it is

found that the

two-conductor transmission
is

circuit theory

necessary to direct our attention to the

and magnetic fields

E and

realize that the energy

and

is

which worked for lumped

lines is inadequate.

circuits and
For the hollow metal pipe or tube it

empty space

H inside the tube.

From

inside the tube

and to the electric

the field-theory point of view

actually conveyed through the

empty space

we

inside the tube

that the currents or voltages are merely associated effects.

13-14

MODE WAVE IN THE INFINITE-PARALLEL-PLANE

TRANSMISSION LINE
TE

As an introduction to waveguides let us consider an infinite-parallel-plane transmission

line,

in

as in Fig. 13-27.

This

is

a two-conductor

TEM mode with E in the z direction.

line

which

However,

is

capable of guiding energy

at sufficiently high frequencies

it

K
Sh,
Direc tion of

propa gation

FIGURE

13-27

Transmission system consisting of two conducting planes parallel to the xy plane.

The planes are assumed to be infinite in extent (infinite parallel-plane transmission
line).

524

ELECTROMAGNETICS

[CHAP. 13

Max

y Max

Wavefronts
of

wave

A
Wave

Zero

field

lines

Wavefronts
of

FIGURE
Two

wave

13-28

plane

TEM

waves traveling

in free

space in different directions.

can also transmit high-order modes, and this type of transmission between the parallel
planes serves as a good starting point for our discussion of higher-order modes.

Devices which can transmit energy only in higher-order modes are usually called
waveguides.

Consider the higher-order

mode where

direction, with transmission in the

component.

Since

Ey

is

the electric field

direction;

i.e.,

is

everywhere

in the

the electric field has only an

transverse to the direction of transmission, this

mode

Ey
is

Although E is everywhere transverse,

has longitudinal, as well as transverse, components. Assuming perfectly conducting
sheets, boundary conditions require that E vanish at the sheets.
However, Ey need
y
designated a transverse electric (TE) mode.

not be zero at points removed from the sheets.

erties

of a

TE wave

It is

possible to determine the prop-

of the type under discussion by regarding

it

as being

made up of

TEM waves reflected obliquely back and forth between the sheets.
First, however, consider the situation that exists when two plane TEM waves of

two plane
the

same frequency traveling in free space

It is assumed that the waves are

Fig. 13-28.

page.

intersect at

an angle, as suggested in
E normal to the

linearly polarized with

Wavefronts, or surfaces of constant phase, are indicated for the two waves.

The

solid lines

directed out

field is a maximum with E

may be regarded as representing the crests of

(marked "max") show where the

from the page.

These

lines

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-14]

AND RESONATORS

525

Conducting sheet-

(a)

Conducting sheet-*

-2~l

^7.

^-^

Conducting

sheet^

(b)

-?-

FIGURE

Conducting sheet'

13-29

Wave

paths and (b) wavefronts between infinite parallel conducting sheets

acting as a waveguide.
(a)

the waves.

where

is

The dashed
of

lines

maximum

(marked " min ") show where the field

is

a minimum,

absolute magnitude but directed into the page.

These

i.e.,

lines

may

be regarded as representing the troughs of the waves. Wherever the crest of

one wave coincides with the trough of the other wave there is cancellation, and the
resultant E at that point is zero.
Wherever crest coincides with crest or trough with
trough there is reinforcement, and the resultant E at that point doubles. Referring to
Fig. 13-28,

it is

therefore apparent that at

all

points along the dash-dot lines the field

always zero, while along the line indicated by dash and double dots the
reinforced and will have

Since

is

its

maximum

field will

value.

zero along the dash-dot

lines,

boundary conditions

will

be

satisfied at

plane, perfectly conducting sheets placed along these lines normal to the page.

waves, however, will

is

be

now be reflected at the

The

sheets with an angle of reflection equal to

the angle of incidence, and waves incident from the outside will not penetrate to the
region between the sheets. But if two plane waves (A and B ) are launched between the
sheets from the left end, they will travel to the right via multiple reflections between the
sheets, as suggested

by the wave paths

in Fig. 13-29a (see also Sec. 12-4).

The wavewave paths) for these waves are as indicated in Fig. 1 3-296. Here
the field between the sheets is the same as in Fig. 13-28, with solid lines indicating that
E is outward (a maximum) and dashed lines that E is inward (a minimum). At the

fronts (normal to the

sheets the resultant

is

always zero.

Although the two component waves we have been considering are plane
mode waves, the resultant wave belongs to a higher-order TE mode. It

TBMis

an

526

[CHAP. 13

ELECTROMAGNETICS

Conducting sheets,

A'

Conducting sheet

FIGURE

13-30

Component waves between

infinite-parallel-plane conducting sheets acting as a

waveguide.

TE-mode wave that it will not be transmitted unless the

The critical wavelength at which transmission is no
short.

important property of the

wavelength

is

sufficiently

longer possible is called the cutoff wavelength.

which

will

now be

It is

possible by a very simple analysis,

given, to calculate the cutoff wavelength as a function of the sheet

spacing.

Referring to Fig. 13-30,

waves traveling

in the x'

let

and x"

the

TE

wave be resolved

directions.

into

two component

These directions make an angle

TEM
with

direction

x axis). The electric field is in the y

The spacing between the sheets is b. From Fig. 13-30 we
note that E'y of the x' wave and Ey of the x" wave cancel at a point such as A at the
conducting sheet and reinforce at point B midway between the sheets provided that the

the conducting sheets (and the

(normal to the page).

'

distance

C'B

wave

in

CB = BD =
where A

is

the wavelength of the

TEM

*o

(1)

unbounded space

filled

with the

into the page (negative) at the

same medium as between the sheets. Thus, if E'y is

point C and Ey is out of the page (positive) at the point D, the two waves will cancel at
A. They will also reinforce at B since by the time the field -Ey moves from C to B
More generally we can write
the field -Ey will have moved from C" to B.
'

CB

nk
(2)

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-14]

where n

is

an integer

(1, 2, 3,

. .

It

.)-t

AND RESONATORS

527

follows that
h

nl

ABsin0 = -sin0 =

(3)

2b

or

= sin

(4)

= wavelength, m
= spacing of conducting sheets, m
n = l,2, 3, ...
= angle between component wave direction

where k

According to

(4),

we

and conducting

note that for a given sheet separation

b,

that can be transmitted in a higher-order mode occurs when

is

the cutoff wavelength k oc of the higher order mode.

Kc =
Each value of n corresponds to a

sheets

the longest wavelength

90.

Thus, for

This wavelength

90,

2b
~

(5)

particular higher-order

mode.

When n =

1,

we find

that
(6)

Lr = 2b
This

is

the longest wavelength which can be transmitted between the sheets in a higher-

order mode.

That

is,

the spacing b must be at least A/2 for a higher-order

mode to be

transmitted.

When n =
n

= 2, we

1,

the

wave

is

said to be the lowest of the higher-order types.

have the next higher-order

mode and
Kc = b

Thus the spacing b must be

n

3,

A oc

= $b,

at least \X for the

When

for this case

(7)

= 2 mode

to be transmitted.

For

etc.

Introducing (5) in (4) yields sin 9

For n even, the

field

A /A oc , or

sin-

-^

halfway between the sheets

side of the centerline.

(8)

is

zero, with

maximum

fields either

528

[CHAP. 13

ELECTROMAGNETICS

\>~ ^oc.

it

if

if

e=90

&

(a)

/WWW

\>

Ao

*-

e<90

\c,

^ ^ AOC)

90
(C)

FIGURE

13-31

(a) to (c) Reflection of

waves between
show-

id) Triangle

walls of waveguide,

ing relative magnitude of phase velocity v,

group velocity u, and energy velocity ue
in the guide, to

phase velocity

(d)

of the

t>

component wave (equal to phase velocity

of wave in an unbounded medium).

Hence, at cutoff for any

mode

the component waves for this

conditions
Kc) the angle 9 = 90. Under these
sheets,
the
between
forth
and
back
mode are reflected
(A

and do not progress in the x direction. Hence there is a standing

wave between the sheets, and no energy is propagated. If the wavelength X is slightly
progresses in the x direction although
less than X oc 9 is less than 90 and the wave

as in Fig. 13-31,

making many

reflections

further reduced, 9

from the

becomes

transmission for this

sheets, as in Fig. 13-316.

As

the wavelength

less, as in Fig. 13-31c, until at very short

mode approaches the

conditions in an

is

wavelengths the

unbounded medium.

constant-phase point of the TE wave

It is apparent from Fig. 13-30 that a
moves in the x direction with a velocity v that is greater than that of the component
waves is the same as for a wave
waves. The phase velocity v of the component
space between the conducting
fills
the
as
in an unbounded medium of the same kind

TEM

sheets.

That

is,

Cms" 1 )

7/

fie

where n

medium, H m
_1
of medium, F m

permeability of
permittivity

(9)

SEC. 13-14]

From

TRANSMISSION LINES, WAVEGUIDES,

Fig. 13-30

it

AND RESONATORS

follows that

- = - = cos
or

According to

= ?=

phase velocity v of a

(11), the
is

(10)

AA

the wavelength

529

(ms

TE

')

(1 1)

wave approaches an

increased toward the cutoff value.

infinite

On the other hand,

value as

v approaches

medium as the wavelength becomes very short.

Thus, the phase velocity of a higher-order mode wave in the guide formed by the sheets
is always equal to or greater than the velocity in an unbounded medium.
The energy,

the phase velocity v in an unbounded

however,
"en

is

cos

propagated with the velocity of the zigzag component wave. Thus

Accordingly, the energy velocity v en is always equal to or less than the

6.

velocity in

an unbounded medium.f

cutoff, the

phase velocity becomes

When,

infinite

for instance, the wavelength approaches

while the energy velocity approaches zero.

This is another way of saying that the wave degenerates into a standing wave and
does not propagate energy at the cutoff wavelength or longer wavelengths. The
relative magnitudes of the various velocities are shown by the triangle in Fig. 13-31rf.
Since the wavelength

mode
bounded medium by

the higher-order

is

proportional to the phase velocity, the wavelength A of

is given in terms of the wavelength A
in an un-

in the guide

are

< 12 >

=co^

The phase velocity and group (or energy) velocity in the guide as a function of
shown in Fig. 13-32. As 9 approaches 90, the phase velocity becomes infinite

while the energy velocity goes to zero.

The velocities are expressed in terms of the

of the wave in an unbounded medium. The situation here is

phase velocity v

analogous to the action of water waves at a breakwater.

Fig. 13-33, a water

Thus, as suggested in

plume moves along the breakwater where a wave

The waveguide behaves

like a lossless dispersive

vo

where u = group velocity

= energy velocity
= phase velocity in unbounded
v = phase velocity in guide

,
v

medium

medium.

It

crest (constant-

follows that

530

[CHAP. 13

ELECTROMAGNETICS

1.5d,

0.5 u -

FIGURE

13-32

Phase and group velocity as a function of wave angle. The ordinate gives t;
and u in terms of the velocity v for a wave in an unbounded medium of the same
type as fills the waveguide.

Plume (velocity =v)

FIGURE

Breakwater

13-33

plume of water moves along a breakwater with a phase

greater than the wave velocity v

velocity v that

is

sec. 13-14]

>

TRANSMISSION LINES, WAVEGUIDES,

AND RESONATORS

531

Conducting

/ sheets

Ev

U h

<-

rr

(a)

FIGURE

(b)

<)

13-34

(a) Infinite-parallel-plane transmission line acting as a waveguide for

is in y direction with wave in x direction (out
of page). The

of two parallel conducting sheets separated by a distance

sheets introduced

normal to

phase point) strikes the breakwater.

wave

velocity v

The

The plume

TE

wave.

guide consists

Ey

The

velocity can

(c)

b.

(b) Additional

Hollow rectangular waveguide.

velocity v of the

become

plume is greater than the

becomes 90.

infinite if

infinite-parallel-plane transmission line

we have been considering is an

and not a type to be applied in practice. Actual waveguides for higherorder modes usually take the form of a single hollow conductor. The hollow rectangular guide is a common form.
The above analysis for the infinite-parallel-plane transidealization

mission line

is of practical value, however, because the properties

of TE-mode waves,
such as are discussed above, are the same in a rectangular guide of width b as between

two

infinite parallel

planes separated by a distance b. This follows from the fact that

conducting sheets are introduced normal to E between the parallel planes,
is not disturbed.
Thus, if a TE-mode wave with electric field in the direction

if infinitely

the field
is

traveling in the

normal to

Ey

direction, as indicated in Fig. 13-34a, the introduction of sheets

as in Fig. 13-34*, does not disturb the field. The conducting sheets

now form a complete

enclosure of rectangular shape. Proceeding a step further,

the sheets beyond the rectangular enclosure be removed, leaving the
hollow
rectangular waveguide shown in Fig. 13- 34c. The cutoff wavelengths for the
TE
modes as given by (5) for the infinite-parallel-plane line also apply for this rectangular
let

b.
For the type of TE modes we have thus far considered (E comy
ponent only) the dimension d (Fig. 13-34c) is not critical.
Although the above simple analysis yields information about cutoff wavelength,
phase velocity, etc., it gives little information concerning the field configuration and

guide of width

fails

to consider

for example,

more complex higher-order modes of wave transmission in which,

both y and z components. To obtain complete

E is transverse but with

532

[CHAP. 13

ELECTROMAGNETICS

information concerning the waves in a hollow waveguide, we shall solve the wave
equation subject to the boundary conditions for the guide. This is done for the

hollow rectangular guide in the next section.

13-15

THE HOLLOW RECTANGULAR WAVEGUIDEt

In Sec. 13-14 certain properties of an infinite-parallel-plane transmission line and of a

hollow rectangular guide were obtained by considering that the higher-mode wave
component waves and then applying the boundary condiconsists of two plane

TEM

component of the resultant E must vanish at the perfectly

conducting walls of the guide. This method could be extended to provide more
complete information about the waves in a hollow waveguide. However, in this
solution of the wave
section we shall use another approach, which involves the
for
the tangential comcondition
boundary
equation subject to the above-mentioned

tion that the tangential

ponent of E.

method we

In this

Maxwell's equations and develop a wave equation

This choice of coordinates is made in order that the

start with

in rectangular coordinates.

boundary conditions for the rectangular guide can be easily applied later. The
to time and a wave
restrictions are then introduced of harmonic variation with respect
made of the type
choice
is
Next a
traveling in the x direction (direction of guide).
may consider a
Thus
we
of higher-order mode of transmission to be analyzed.
magnetic
(TM) wave for
transverse
transverse electric (TE) wave for which Ex = or a
must be an
that
there
know
which H* = 0. If, for example, we select the TE type, we

Hx component, since a higher-mode wave always has a longitudinal field component

then convenient to write
and Ex being zero means that Hx must have a value.
of a scalar -wave
solution
Next a
the remaining field components in terms of Hx
rectangular
guide.
the
of
the boundary conditions
obtained that
equation in Hx
It is

fits

is

This solution

is

substituted back into the equations for the other field

components

varia(E ,Ez ,Hy and z ). In this way we end up with a set of equations giving the
solution
This
method
of
tion of each field component with respect to space and time.
many problems.
is very general and can be applied to
,

On the Passage of Electric Waves through Tubes, Phil. Mag., 43:

125-132 (February 1897); L. J. Chu and W. L. Barrow, Electromagnetic Waves in
Hollow Metal Tubes of Rectangular Cross Section, Proc. IRE, 26: 1520-1555
(December 1938). A general dyadic Green's function method for solving transmis" Dyadic Green's Funcsion line and waveguide problems is developed by C-T Tai,
Scranton, Pa.,
tions in Electromagnetic Theory," International Textbook Company,

t Rayleigh, Lord,

1971.


TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-15]

We

shall

waveguide.

develop the method

First,

in detail for

however, the procedure

Start with Maxwell's equations.

Apply
Apply

Select the type or

Hx #
5

restriction of
restriction

TE

waves

be outlined

will

in a

AND RESONATORS

533

hollow rectangular

in step

form as follows:

harmonic variation with respect to time.

of harmonic variation and attenuation with respect to

transmission (TE in this case; so

mode of wave

Ex =

x.

and

0).

Find equations for other four

Hx

components (Ey

field

Ez

and

z)

in

terms of

Develop scalar wave equation for

Solve this wave equation for

Substitute

each

field

Hx

subject to boundary conditions of waveguide.

Hx back into equations of step

5,

giving a set of equations expressing

component as a function of space and time.

plete solution

Beginning

This constitutes the com-

of the problem.

now

with step

of the procedure, we have from Maxwell's curl equa-

tions in rectangular coordinates the following set of six scalar equations

dH z

8H y

~Sy~

dz

dH x

oEx

dHz aE >-

dz

dx

dHy

dH x

dx

dy

dEz

'

dE x
dz

dEy
dx

From Maxwell's

(1)

(2)

(3)

dHx
=
-dT

(4)

(5)

(6)

dt

'

8EV

-lz-

8E*
-d7

dE*

dy

dEx

'

dx

dEx

"dy

dt

dHz
+fl
dt

divergence equations in rectangular coordinates

we have

in

space free from charge the following two scalar equations:

^ + ^ + ^_0

(7)

d
8
d
J
+ Jly + J!i = o

(8)

dx

dx

dy

dy

dz

dz

534

ELECTROMAGNETICS

[CHAP. 13

now that any field component varies harmonically

may attenuate with distance (steps 2 and 3). Thus,
waves traveling in the positive x direction, we have, for

Let us assume
distance

and

attention to

the field

also

component

Ey is expressed

with time and

confining our
instance, that

by

E,

E^at ->x

(9)

= propagation constant = a +jP

= attenuation constant
=
P phase constant

where y
a

When the restriction

of (9)

is

introduced into the equations,

dH,

~dj

dz

dHv
" ~ai ~

(ff

reduce to

(10)

+ yHz -(a + jwe)Ey =

-yHy -^-(

(11)

(12)

-^--^+j(OliHx =

(13)

ff

dEz

dEv

dy

dz

-^ + yE

dz

yEy

The above

x
=
+ J(0)E
*

(1) to (8)

+jcoe)Ez

+jco(iHy

(14)

=0
-^+jconH
dy

(15)

eight equations can be simplified

by introducing a

series

impedance

Z and

shunt admittance Y, analogous to a transmission line (see Table 13-2), where

Z=-ja>n

(fim

-1
)

(Urn' 1 )

Y=ff+jcoe

(18)
(19)

Substituting these relations in (10) to (17) yields

dH_
-^
+ yH -YE
Tz
z

(21)

TRANSMISSION LINES, WAVEGUIDES,

r
SEC. 13-15]

AND RESONATORS

535

-yHy - -^-YEz = o

ill)

(23)

(24)

K-^-ZH. =

(25)

dz

dy
d

- + ?E,-ZH,

r,

SEy

dEz

dHv

dHz =

(26)

(27)

These are the general equations for the steady-state field of a wave traveling in the
the
x direction. No restrictions have as yet been made on the mode of the wave or
condithe
shape of the guide. We are now ready to proceed with step 4 and introduce
tion for a

TE wave

that

Ex = 0.

The equations then reduce to

dHz

dHy

dy

dz

(28)

d
-^
+ yH -YEy =
oz

(29)

-g-YE. =

(30)

- yHr

d
Jk-^-ZH
dz
dy

(31)

yEz - ZHy =

(32)

-ZH

(33)

^+^=

(34)

-yEy

dz

dy

dH
dH "
-^ + i7
+
v

Proceeding to step
is

expressed in terms of

5, let

Hx

,in

(35)

us rewrite these equations so that each

To do

this,

we note from

^=-^
=5
H
H
v

(11)

field

component

(32) and (33) that

(36)

536

The

ELECTROMAGNETICS

ratio

[CHAP. 13

EJHy or EJH

Z is

a quantity which corresponds, in the case of a waveguide,

to the characteristic impedance of a transmission line.

transverse field components,

waveguide.

it

may be

Zyz of the

Thus

Z' ,yz ~
=Tr=-TT=-H~ H ~ ~ y =~
Z

Introducing (37) into (30) and solving for

manner we have, from

(37)

()

In a like

Since (36) involves only

called the transverse-wave impedance

yields

-1
dH
-^- YZyz -^
~dy~
:

(38)

(29),

*-^J- YZ
y

Now,

and

substituting (39) into (37),

(39,

dz

yI

we obtain

substituting (38) into (37) gives

-Z

dHx

vz

Equations (38) to (41) express the four transverse

Hx

This completes step

Proceeding
derivative of (38)

now

to step 6,

/d

Hx

This

shape.

is

a partial

wave equation

differential

in

terms of

Hx

by taking the y
This yields

(43)

reduces (43) to

H + __,
H +
_,
scalar

in

Hx\

yz )ffx

= y(v - YZyz)

in

substituting both in (35).

-^ + -^r + y(y- YZ

Putting k 2

components

we can obtain a wave equation

and the z derivative of (39) and

or

field

5.

Hx

This completes step

It
6.

^
,

equation of the second order and

applies to a

(44)

first

degree.

It is

TE wave in a guide of any cross-sectional

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-15]

FIGURE

AND RESONATORS

537

13-35

Coordinates
waveguide.

hollow

for

Step 7

is

rectangular

to find a solution of (44) that satisfies the

waveguide under consideration, which

boundary conditions for the

a hollow rectangular type, as shown in

is

Assuming that the

and the height is y1
walls are perfectly conducting, the tangential component of E must vanish at the
guide surface. Thus, at the sidewalls Ey must be zero, and at the top and bottom
surfaces Ez must be zero. The problem now is to find a solution of (44) subject to
these boundary conditions. The method of separation of variables can be used in
obtaining the solution. Thus, x in (44) is a function of y and z. Hence we may seek
Fig. 13-35.

The width of

the guide

is

zu

a solution of the form

Hx =
where

Y=a

function of y only, that

Z a function

is,

YZ

(45)

Y =f(y)

of z onlyf

Substituting (45) in (44) gives

dy
Dividing by

YZ

dz

(46)

to separate variables gives

ld 2 Z

+ = =-k
2+
Z dz 2
Y dy
1

(47)

The first term is a function of y alone, the second term is a function of z alone, while
k 2 is a constant. For the two terms (each involving a different independent variable)
to equal a constant requires that each term be a constant. Thus we can write

\d2 Y_ ~
Ai
lly 2" ~
t

(48)

Sometimes the notation f(y) or Y(y) is used to represent a function of y only and
However, to simplify notation, the symbols Y and
f(z) or Z(z) a function of z only.
are used in Eqs. (45) to (55), inclusive, to indicate functions only of .v or z, re-

Y and Z in these equations should not be confused with admittance and

impedance, for which these symbols are also used.
spectively.

A
538

ELECTROMAGNETICS

[CHAP. 13

\<PZ

and

- ~ A2

~HI
Zdz

where

and

A2

are constants.

(49)

follows that

It

(50)

Equations (48) and (49) each involve but one independent variable.
(48)

solution of

is

Y=c

sin

bt y

(51)

Substituting (51) in (48) yields

h=JA~i
Hence

(51)

is

a solution provided (52)

and

(53) are each

solution

may be

Z = c3
Substituting (54)

CiC 3 sin

and

*jA\y

sin

(55) into (45),

sin S/A 2 z

(53)

a solution for Y, their sum

manner a

c t c4 sin

*J

is

c 2 cos b t y

r=c sin^Mj^ +
In the same

Another solution

is fulfilled.

Y=
If (51)

(52)

c2

we obtain

c 2 c 3 cos

-Ja 2 z

the solution for

sin

(54)

Z as

c 4 cos

-jA^y

yJAiy cos

also a solution, or

coSy/A^y

written for

is

(55)

Hx as

y/A 2 z
c 2 c^ cos -JAy

y cos sJA^z

(56)

and (41) and introducing the boundary conditions that

and 2 = at y = and y = j^, we find that only the last
the boundary conditions and then only provided that

Substituting (56) in (40)

Ey =

at z

and

term of (56) can

=z

satisfy

Ia~-

and

= mn

A-,

(57)

(58)

where

m and n are integers (0,

The

to different integers.

2, 3,

the time

c 2 c4

That

solution.

= a constant.

solution for

may

assumed
then becomes

integer or

mnz

nny
cos

If (59)

this gives the variation

539

Hx now assumes the form

is

.,..

cos

(59)

multiplied by a constant factor,

the factor should not involve

is,

Accordingly, (59)

t.

They may be equal to the same

.)

solution for

Hx (y,z) = H
where

AND RESONATORS

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-15]

y or z although

it

may

it is still

involve

x and

be multiplied by the exponential factor in (9) since

x and t. The complete

for the fields with respect to

(60)

This completes step

To perform

7.

step 8, Eq. (60)

substituted into (38)

is

through (41), giving the solutions for the transverse field components as

Hy =

Z^ ^
in

cos

-^2
Ev = yz

yZ

H mn

mnz

sin

yt

y^Ho nn ^ nny
k

yx

yx

mnz

nny
cos

z,

ja
e "

(62)

zt

>>!

(61)

zi

nny
yH mn
-7- cos

sin
H = -p
2
z
z

^^<-,*

yi

yi

yt

(63)

e'

z,

^ mnz ^^
z,

which may be added Ex = 0, are the solutions we have

sought for the field components of a TE mode in a hollow rectangular guide of width z t
Equations (60) to

(64), to

and height y^

This completes step 8f.

Turning our attention now to an interpretation of the solutions for the field components, let us consider the significance of the integers m and n. It is apparent that

for

m=

and n

= we have only three field components HX ,HZ

and

Ey and, further,

components has no variation with respect to y but each has a halfFor example, Ey has a sinusoidal variation across
cycle variation with respect to z.
the guide (in the z direction), being a maximum in the center and zero at the walls,
that each of these

and has no variation

t

Note

as a function of y.

that to be explicit in (61) to (64) we could write

y (y,ZrX,t) instead of
z , etc., but for simplicity we do not.

H (y,z,x,t) instead of H

H,

540

ELECTROMAGNETICS

[CHAP. 13

If m = 2, there is a variation of two half-cycles (full-cycle variation) of each field

component as a function of z. When n = 1 there is a half-cycle variation of each field
component with respect to y. Hence we may conclude that the value of m or n
,

number of half-cycle variations of each field component with respect to z

Each combination of m and n values represents a different field
configuration or mode in the guide.
Since we are dealing here with TE modes, it
is convenient to designate them by adding the subscript mn so that, in general, any
TE mode can be designated by the notation TEm where m is the number of halfindicates the

and

y, respectively.

cycle variations in the z direction (usually taken as the larger transverse dimension of

the guide) and n

is

the

number of

half-cycle variations in the

direction (usually

taken as the smaller transverse dimension of the guide).

Case

1:

TE 10 mode

For

this

above, only three components

ponents for the

TE 10 mode

m=

mode

Ey

Hx

and

and n

0,

and we have, as mentioned

The six field com-

that are not zero.

are

Ex =

TE mode
k

zl

condition

z1

Ez =

(65)

Hx = H

cos

iu"-> x

tf,=0

The

Zi

z,

components as a function of z is portrayed in Fig. 1 3-36a.

no variation with respect to y. This mode has the longest cutoff wavelength
of any higher-order mode, and hence the lowest frequency of transmission in a hollow
rectangular waveguide must be in the TE 10 mode. In Fig. 13-37a the field configuration of the TE 10 mode is illustrated for a guide cross section and in Fig. 13-376 for a
variation of these

There

is

longitudinal section of the guide (top view).

Case 2 TE 20 mode The variation of the field components as a function of z for the
TE 20 mode (m = 2, n 0) is shown in Fig. 13-366. The field configuration for a
TE 20 mode is shown in cross section in Fig. 13-37c and in longitudinal section (top
:

view) in Fig. l3-37d.

FIGURE

AND RESONATORS

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-15]

13-36

waveguide.

(Wave

traveling

z=0

z=b

z=0

z=b

Variation of field components for TEio

and TE 20 modes in a hollow rectangular

541

m=2
=0

of

out

(a)

page.)

(b)

TE 2(

TE,
yz plane
(wave out
of page)

~V

lines solid

lines

dashed

ox

<>x

(c)

\u

II

II'

KJ

It

lines

down
it
xz plane
!

!t

1
'

<"*,

I
'

A;?

'

'

'

lines-

,'%
i

'^

rN

i i

'

'

(d)

(b)
13-37

Field configurations for

guide.

up

FIGURE

Ji liii
'

\Js_y
-T

6V.O

'
I

lit*

v y

(wave down
page)

TE,

and

TE 20 modes

in

a hollow rectangular wave-


542

ELECTROMAGNETICS

[CHAP. 13

Hx

tf

Hz

FIGURE

13-38

TE, ,
(Wave

Variation of field components for

mode

in a square waveguide.

traveling out of page.)

Case

3:

TEi i mode

For this mode m

and the field components are given

by

Ex =
E

TE mode

condition

yZ v ,H n

ny
nz ,,_
cos
sin e"*

f2,
k

zx

yZ H n

E=- ^

Hx = if

ny
nz

cos

yi

e"

yi

cos nz
e"*7tv

cos

(66)
yx

Zl

J>1

fc

sin

yi

fc

yx

yt

yi

zi

H2= ^oJl cos ^ sin ^,-.

*

Zl

.Vl

Zl

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-15]

AND RESONATORS

543

For this mode five field components have a value, only Ex being everywhere and always
The variation of the five field components with respect to r and y is shown in
zero.
Fig. 13-38.

It is

assumed that the guide has a square cross section

field configuration for the

TE n mode in a square

guide

(y^

z t ).

The

illustrated in cross section

is

(end view) in Fig. 13-39a and in longitudinal section (side view) in Fig. 13-396.
The solution we have obtained tells us what modes are possible in the hollow

However, the particular mode or modes that are actually

on the guide dimensions, the method of exciting the guide,
depend
any case

rectangular* waveguide.

present in

and the
is

The

irregularities or discontinuities in the guide.

equal to the

sum of the

Returning

have from

now to

(50), (57),

of

fields

modes

all

(43), (18),

we

a consideration of the general significance of the solution,

and

(58) that

'er)'From

resultant field in the guide

present.

and

(19),

k2

k2

(67)

given by

is

41

Assuming a lossless dielectric medium

(67) and (68) and solving for y yields

+jcoe)

-jmnia

in the guide,

(68)

we can put a =

Then equating

0.

->/' '-"*
,

(69)

Position of

end view

gTO''H|H;;rr.v_::;H||

r-/-^
z

z,

>r

'

Side view through section

II

a-a

(b)

(a)
13-39

Field configurations for

and

a
End view

FIGURE

IV

H lines are dashed.

TEi i mode

in

a square waveguide.

lines are solid,

544

ELECTROMAGNETICS

[CHAP. 13

At sufficiently low frequencies the last term in (69) is smaller than the sum of the
two terms under the square-root sign. It follows that for this condition y is real
and therefore that the wave is attenuated. Under this condition it is said that the
wave (or mode) is not propagated.
At sufficiently high frequencies the last term in (69) is larger than the sum of the
first two terms under the square-root sign.
Under this condition y is imaginary, and
therefore the wave is propagated without attenuation.
At some intermediate frequency the right-hand side of (69) is zero, and hence
first

= 0.

This frequency

is

called the cutofffrequency for the

At frequencies higher than

frequencies lower than cutoff the

To summarize
1

2
3

mode propagates
mode is attenuated.

cutoff this

mode under consideration.

without attenuation, while at

the three cases

At low frequencies, m small, y real, guide opaque, wave does not propagate
At an intermediate frequency, a> intermediate, y = 0, transition condition, cutoff
At high frequencies, co large, y imaginary, guide transparent, wave propagates
Referring to (69),

it is

2
to be noted that ~Ja> ne

is

equal to the phase constant p

dielectric material as fills the

wave traveling in an unbounded medium of the same

guide. Thus we can write
for a

= Jk 2 - p

(m" 1 )

= ^/co 2 ^ = 2n/A = phase constant in

= wavelength in unbounded medium
k = y/innly^ 2 + (mrc/Zj) 2

where P
k

Thus, at frequencies higher than cutoff

ff

>

= Jk 2 -P

unbounded medium

and

k,

(70)

=JP

(71)

= 2n\X = phase constant in guide, m _1

A = wavelength in guide, m
At sufficiently high frequencies (fi > k) we note that the phase constant ft in the guide

where

fi

approaches the phase constant P in an unbounded medium.

frequencies less than cutoff /? < k, and

whete a

is

= Jk 2 - p

the attenuation constant.

On

the other hand, at

(72)

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-15]

At

sufficiently

low frequencies

approaches a constant value

<^

(/3

k)

we

= 0.

AND RESONATORS

545

note that the attenuation constant a

k.

At the cutoff frequency,

/}

and

k,

(nn\

Thus, at cutoff

/mn\ 2
(73)

It

follows that the cutofffrequency

is

(74)

and the cutoff wavelength

is

2n

^(nnlytf

where k oc

more

is

2
(m)

Kc

(mn/z,)

^/(n/y,)

unbounded medium

the wavelength in an

concisely, the cutoff wavelength).^

at the cutoff frequency (or,

Equations (74) and (75) give the cutoff

TEmn mode in a hollow rectangular

TE 10 mode is

frequency and cutoff wavelength for any

instance, the cutoff wavelength of a

is

identical with the value

At

found in Sec. 13-14 since z l

frequencies above cutoff

(/?

>

follows that the phase velocity vp in the guide

is

b.

C77,

equal to

Vo

For

k)

W^=^V-g=)'-0f)
It

guide.

(76)

-2*i

This

(75)

(m/zj 2

(m

s-

(78)

JX-inkvllytf-Wollztf

2ir}X 0C
t Note that k
A, A , and X oc when A

If this-value

< A^

of k

is

introduced, (71) can be used to relate

546

ELECTROMAGNETICS

[CHAP. 13

2.5

teJ teJ

TE I0 /

2.0

FIGURE

13-40

1.5

Relative phase velocity v/v

or relative

Zyz/Zd as a function

transverse impedance

for TE modes in a
hollow square guide (height y t equal to
width z t ).

of the wavelength A

0.5z,

Wavelength

2.0z,

1.5z,

1.0,

t>o

or

y/l

where v
X
k BC

(79)

~ {hike?

= 1/vi"6 = phase velocity in unbounded medium (=

= wavelength in unbounded medium
= cutoff wavelength

300

Mm

-1

for air)

The ratio v/v as a function of the wavelength 1 is shown in Fig. 13-40 for several TE
modes in a hollow waveguide of square cross section (j^ = Zj).
In the above analysis there is no attenuation whatsoever at frequencies above
This results from the assumption of perfectly conducting guide walls and a
cutoff.

medium filling the guide. However, if the walls are not

medium is not lossless, or both, there is attenuation.

lossless dielectric

perfectly

conducting or the

If the guide is filled with air, the dielectric loss

is

usually negligible

compared

with losses in the guide walls, so that the attenuation at frequencies greater than cutoff
is

mainly determined by the conductivity of the guide walls.

walls are not perfectly conducting

means

The fact that the guide

component E of the

that the tangential

However, for walls made of

a good conductor, such as copper, E, will generally be so small that the above analysis
(based on E, 0) is not affected to any appreciable extent. However, as a result of

electric field is

not zero at the walls but has a

the finite wall conductivity a

wall conductivity

is

is

not zero.

high (but not

finite value.

Thus, in most practical problems where the

infinite)

the field configuration in the guide, the

y may have both a real and an imaginary part at frequencies greater than cutoff
can be shown by solving (68) for y under these conditions, with a not equal to zero.

t That

TRANSMISSION UNES, WAVEGUIDES,

SEC. 13-15]

wavelength

A,

the phase constant

/?,

the phase velocity

with high accuracy on the assumption that the walls have

The small

earlier in this section.

v, etc.,

power

being assumed that the H-field distribution

is

lost

the

can

all

may

547

be calculated

infinite conductivity, as

(but not zero) attenuation

separately, using (10-19-5) to find the

AND RESONATORS

done

then be calculated

per unit area in the guide wall,

same as with

it

perfectly conducting

walls.

Finally, let us determine the value of the transverse-wave

modes

in a rectangular

impedance

Zyz

TE

for

Thus, from (37)

hollow guide.

Zyz

J^.

(80)

At frequencies higher than cutoff y =j'P; so

where

Zd = intrinsic impedance of dielectric medium filling guide

= JJ[e = 376.7 fl for air
A = wavelength in unbounded medium
k ac

The

cutoff wavelength

ratio of

Zyz (transverse- wave impedance) to Zd (intrinsic impedance) as a function

of the wavelength A

is

shown

in Fig. 13-40 for several

guide of square cross section (y l

TE modes

in a hollow wave-

z,).

TE-mode waves have been considered. To find the field relations

(TM) mode waves we proceed precisely as in the eight-step
list given earlier in this section except that where TE appears we substitute TM and
where Ex appears we substitute Hx and vice versa. In the TM wave Hx = 0, and the
Thus

far only

for transverse magnetic

component is Ex
This analysis will not be carried through here
However, it may be mentioned that (75) for the cutoff wave(see Prob. 13-39).
length applies to both TE and TM waves, as does (78) for the phase velocity, but this is
not the case with (81) for the transverse impedance (see Prob. 13-42). The notation
for any TM mode, in general, is TM m where m and n are integers (1, 2, 3,
It is
.).
longitudinal field

to be noted that neither

lowest-frequency

TM n

nor n

may

be equal to zero for

TM

waves.

Thus, the

TM wave that will be transmitted by a rectangular waveguide

is

the

mode.

We

have seen that each

mode

of transmission in a waveguide has a particular

cutoff wavelength, velocity, and impedance.

When

the frequency

is

high enough to

548

ELECTROMAGNETICS

[CHAP. 13

(a)

V
(b)

(a)

(e)

(*>)

TE 20

(d)

mode

FIGURE

FIGURE

13-41

Rectangular waveguide with

TE I0 mode

only.

permit transmission in more than one mode, the resultant

of the individual

mode

13-42

Rectangular waveguide with TE 20 mode

induced from TE I0 mode by asymmetrically placed projection (screw).

the

field is

sum

of the fields

fields in the guide.

For example, suppose that a rectangular waveguide, as shown in cross section

is excited in the TEi
mode. The variation of Ey across the guide

Fig. 13-41a,

sinusoidal, as

shown

TE 20 mode can also

in Fig. 13-416.

be transmitted.!

appear provided that the guide

asymmetries and irregularities
the

TE 10 -mode

energy into

If only the

TE 10 mode is

perfectly regular.

is

will

Suppose now that z t exceeds

be present, and these

TE 20 -mode

energy.

Thus,

screw projects into the guide as in Fig. 13-42a, the total

asymmetrical, as suggested in Fig. 13-426.

This total

will
if

no

TE 20

will

in practice certain

tend to convert some of

an asymmetrically located

Ey field

field

is

11, so that the

excited,

However,

in

may

will

tend to become

be resolved into

TE 10

and TE 2 o components as shown in Fig. 13-42c and d. If both TE 10 and TE 20 modes

can be transmitted, the field in the guide beyond the screw location will have energy in
both modes.

In effect the screw

is

a receiving antenna that extracts energy from the

TE 10 -mode wave and reradiates it so as to excite the TE 20 mode. However, if

frequency is decreased so that only the TE 10 wave can be transmitted, the asym-

incident

the

metric

field (Fig. 13-426) will exist

But y t

< Ao/2,

only in the vicinity of the screw and farther

so that no

TE

mode (E

in z direction)

is

transmitted.

down the

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-15]

AND RESONATORS

549

m(z)

FIGURE 13-43
TE and TM modes in a

Possible

hollow rectangular waveguide as a function of

At three times the cutoff frequency for the TE, mode there are
seven modes which will pass (see text at end of Sec. 13-15) and one mode (TE30)
frequency.
at cutoff.

guide the

mode

field will

be entirely in the

transmission, a waveguide

of transmission.!

is

TE 10 mode.

To avoid

the problems of multiple-

usually operated so that only one

For instance, to ensure transmission only in the

mode

TE 10

is

capable

mode,

zl

must be less than 1A and y l less than A/2. But to allow transmission of the TE 10
mode, Zj must exceed A/2. Hence zt must be between A/2 and 1A, and a value of
0.7 A is often used since this is well below 1A and yet enough more than A/2 so that the
and transverse impedance values are not too critical a function of frequency.
recall that at cutoff (z t = A/2) the velocity and impedance approach infinite values.

velocity

We

The height >>, may be as small as desired without preventing transmission of the TE 10
mode. Too small a value of yu however, increases attenuation (because of power lost
It is
in the guide walls) and also reduces the power-handling capabilities of the guide.
=
of
a
rectangular
and
TM
modes
TE
often the practice to make y t
zJ2. Many
m
m
guide for which y t = zJ2 are shown in Fig. 13-43. The slant scale gives the frequency
t

The lowest-frequency mode

that a guide can transmit

is

called the dominant mode.

550 ELECTROMAGNETICS

[CHAP. 13

dominant mode (TEj ) Thus, if the frequency

note from Fig. 13-43 that the following modes will be trans-

relative to the cutoff frequency for the

is

3 times this value,

we

mitted: TE 10 TE TE 20 TE n
(TE 30) is at its cutoff frequency.
The relations derived in this
,

Fig. 13-35) are

summarized

in

TM U TE2
,

and

TM 21

An

mode

additional

section for a hollow rectangular waveguide (see

Table 13-6.

Table 13-6

Name

RELATIONS FOR

TE MODES IN
HOLLOW RECTANGULAR WAVEGUTDESt

of relation

Relation

t-azM'+ff)' (m

Cutoff frequency

v
"

w v
Wh) +Wll
>

(m)
).

*
'

(ml
(U1)

Vi-(A /A^
Vo

Phase velocity
"

Vl - (bAo/2*) 2 - (mA

Vi-(A
Vl where v

Transverse-wave impedance

Zj z

/A c)

/2z,)

(ms"

(/c //y_

1/V^i.e

Vi - (nX

Zt
y - (mA

/2yi

/2z,)

Vl - (A/Ae) 2
z

Vl - (Uff
where Za

(O)

V/i/ c

TM modes except for the transverse-wave im-

t All the relations also apply to

pedance relation. The velocity

and impedance

apply not only to rectangular guides but also to

conductor guides of any shape.

relations involving (A /A c) 2

TE modes

in

hollow

single-

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-16]

AND RESONATORS

551

THE HOLLOW CYLINDRICAL

WAVEGUIDEt

13-16

hollow (circular) cylindrical

Consider the problem of describing wave propagation in a
This problem is most easily handled with a cylindrical
waveguide of radius r
procedure is similar to that used. in
coordinate system, as shown in Fig. 13-44. The
waveguide. We assume time-harmonic
the preceding section for the rectangular
interior medium (a = 0) containing
variation, perfectly conducting walls, and a lossless
.

no charge (p

= 0).

Maxwell's two
Maxwell's two curl equations yield six scalar equations, and

divergence equations yield two scalar equations.

In cylindrical coordinates these are

as follows:

8
- + VE4 -ZH =

(1)

d
- yEr - J-ZHt =

(2)

r dip

h + lE.-l?g-ZH.-0
v
dr

(3)

d<t>

+ ytt*-YE
-
o<p

(4)

8
-yH - -^-YE* = 0'
r

dr

I^_I_re
r

A+^+
8

dr

dr

(5)

or

(6)

d<t>

l
r

8
J-yE
=

CO-

d<p

oq>

_1

impedance = -join, CI m
-1
=
admittance = jioe, Urn
shunt
Y
E* = E^*"- *, etc.
y = a +jf} = propagation constant

where

Z = series

steady-state
These eight relations are the general equations for the

wave

cylindrical coordinates.
traveling in the z direction as expressed in

At

field

of a

this point

Waves in Hollow Tubes of Metal,

L. Barrow, Transmission of Electromagnetic
Southworth, Some Fundamental
Proc IRE, 24: 1298-1328 (October 1936); G. C.
1937).
Experiments with Wave Guides, ibid., 25: 807-822 (July

552

ELECTROMAGNETICS

[CHAP. 13

Direction of wave propagation

FIGURE

13-44

Coordinates for hollow cylindrical waveguide.

we can

limit the

(13-15-27),

H* =

c i c3

we

either a TE wave or a TM wave.

Selecting the TE wave
same manner as for a rectangular waveguide following
the solution for the
z component to be

problem to

and proceeding

in the

find

sm n 4> Jn (kr) +

c 2 c 3 cos

n Jn (kr)

+
where k
n

= s/y 2~+ ZY =
integer

Equation
tion.

Eq

(9) is

a solution for

The appropriate
at r

becomes

^/y

solution

and that the

infinite at r

co

it(f>

N (kr) + c 2 c4 cos
n

n<t>

but any term or combination of terms

must

is

N(kr)

(9)

ne

satisfy the

is

also a solu-

boundary conditions, which are that

the guide are

fields in

and so

c^ sin

finite, f

not a suitable solution.

The Neumann function

Thus,

(9)

can be reduced

to

= i/

where

(cos

c2 c 3

The boundary condition

n<f>

+j sin n<f>)Ja (kr)

(10)

C
-^

E =

at r

=r

nXE =

can also be expressed

atr

/-

H
TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-16]

When

the sine term

z is

maximum where

0.

If

(11)

cos n<j>Jn {kr)

n/2 and 3n/2 provided n

This choice means that

<l>

553

dropped, (10) simplifies to

is

H
=

AND RESONATORS

4>

and

we had dropped

(j>

n and

is

zero where

the cosine term instead of the

would be rotated 90 in the guide. Assuming that the

would suffice (but both would be required
polarized wave in the guide). The six components aret

sine term, the conditions for

field orientation is arbitrary, either solution

for a circularly

nyH
k

Jn(kr)ej

sjn

dJJpr)

= yJhZr*
cos n(j)
2
k

Ez =

mt -y,

=z

(12 )

^^ = _

dr

TE-mode condition

= Zl^
CO
2
k

s^^^<-

cos n<p

3)

(14)

(15)

dr

H^^sinn^Ukr)^'--

(,

(16)

Ukr)^-

(17)

variation of the fields with time and distance has also been expressed explicitly
imt ~ yz
written using
Alternative expressions for E+ and
r can be
(proportional to e
).

The

the recurrence relations (see Appendix, Sec.

<

A- 10):

= k[ukr)-Jn+l (kr)]

(18)

Equations (12) to (17) are general expressions for a TE-mode wave in a cylindrical
at r = r has not been imposed.
However, the boundary condition E+ =
guide.

Applying

this condition requires that

dJK {kr)

at r

(18a)

dr

and hence that

(19)

t In (12)

and

(13)

'

H<,

H,

554 ELECTROMAGNETICS

[CHAP. 13

J^kr)
(a)
1

<J/l(tr)

dr

(b)

zp- -j.m
FIGURE

13-45

and its derivative as a function of kr. (6) Bessel

function relations involved in determining E, and E+ variation for TEj i mode.
(a) First-order Bessel function

where

k' is the rth root

situation

is

of the derivative of the Mth-order Bessel function.

illustrated in Fig. 13-45a for

The

first

The

three roots of the derivative

= k'nr = 1.84, 5.33, and 8.54. Thus, the

= 1.84, k'12 = 5.33, and k'13 = 8.54, corresponding to wave modes TEn
and TE 13 in the guide. In general, a TE mode in a cylindrical guide is designa-

of the first-order Bessel function occur at kr

roots are k'u

TE 12
ted

TEr where the subscript n indicates the order of the Bessel function and r indicates
For TE 01 and TE 02 modes, n = and the roots of (18) for this
,

the rank of the root.

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-16]

AND RESONATORS

555

Er <t sin 4>i>^

= 180'

0=0

0=180

FIGURE

13-46

(a) Variation

TE, i mode
cylindrical

case are k'nr

3.832

and

of electric

in cylindrical

field

components E, and E* with

waveguide with component

7.016.

radial distance r for

(b) Electric field vectors for

waveguide,

fields

E, and

TE, , mode in

See

These are the same as the zeros of J^kr).

Fig. 13-45a.

mode is as
Referring to (13), the variation of E+ with r (at <j> = 0) for the TE U
90)
given by
is
as
shown in Fig. 1 3-45fc. From (1 3) the variation of Er with r (at <f> =
the curve for J^kr^kr.

The amplitudes of

the two field components,

thus be represented by curves, as in Fig. 13-46a, or by

Note that E+ is a maximum at $ = 0 and 180 and zero at

and E^ can
,

arrows, as in Fig. 13-46*.

90 and 270 while

Note also that at

= 90 and 270 and zero at = and
is a maximum at
and amplitude.
direction
the center of the guide (r = 0) Er and * both have the same
0

<f>

is

tf>

are added, the total electric field for the TE n in the

direction of
as suggested by the heavy arrows in Fig. 13-46* with the
In a
13-47a).
Fig.
also
lines
(see
as indicated by the curved

When the Er and E+ components

cylindrical guide

180.

the total electric field

556

ELECTROMAGNETICS

[CHAP. 13

180'

(Dominant mode)

FIGURE

13-47

Electric field lines (solid)

field lines

and magnetic

(dashed) in hollow cylindrical

(Low-loss

waveguide for (a) TEi , (dominant) mode

and (b) TE

(low-loss)

mode.

mode)

(b)

manner the transverse magnetic field configuration for the TE n mode can be
deduced from the variation of r and H^ with r [from (15) and (16)], with the result
similar

shown in Fig. 13-47a.

The solution we have obtained for the TE n mode is but one of an infinite
number of possible wave modes in a cylindrical guide. The particular mode or modes
that are actually present in
exciting the guide,

equal to the

TE 0l mode
Thus

and

sum of the
is

any case depend on the guide dimensions, the method of

discontinuities in the guide.

fields

of

all

modes

present.

The resultant field in the guide is

The field configuration for the

illustrated in Fig. 13-476.

far only

TE modes

the field relations for

in a cylindrical guide

TM-mode

waves we

have been considered.

To

find

and express the

remaining field components in terms of the longitudinal field Ez
A wave equation in
Ez is developed and solved to satisfy the boundary condition, which requires that
Jn(kr) =
at r = r
The first three roots k(=kr ) for the case n = 1 are k n = 3.832,
k 12 = 7.016, and fc 13 = 10.173. The fields for TM modes are given in Prob. 13-40.
let

in (1) to (8)
.

The roots

(or eigenvalues) for the

TM modes written k

(unprimed) correspond

to zero values of the Bessel function /(&/), whereas the roots for the

TE modes written

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-16]

correspond to zero values of the derivative (with respect to

k' (primed)

The

function.

relations for

numerical values

As seen

mode

since

wise the

it

listed in

TE

Table

in Fig. 13-48, the

and

TM modes are displayed in

TE 01 mode

Let us

constant,

now

and

and Table 13-7 and

of the Bessel function.

TE 03

should properly be designated

for cylindrical guides has been pointed out by

designations persist

of the Bessel

Fig. 13-48

should logically be designated the

refers to the second root of the derivative

TE 02 mode

r)

557

their

13-7.

corresponding to a trivial (nonexistent) mode.

waveguide.

AND RESONATORS

The

first

root

TE 02
Like-

is k'

0,

This inconsistency in the nomenclature

C-T Tai,t but the common and

conform to

Fig. 13-48

this

common

incorrect

usage.

examine the conditions necessary for propagation inside a cylindrical

Substituting (19) into

k =

2
2
Jy + m \i

and solving for the propagation

we obtain

-*w-

2
a> /i

Table 13-7

= a+jfi

(20)

CYLINDRICAL WAVEGUIDE

MODES
Eigenvalues

Cutoff
wavelength,

k'

Knr

A oc

2.405

2.61 ra

Mode
designation

TM 01
TE

3.832

(low loss)

TM 02
TE 02
TE, ! (dominant)

7.016
1.840

TM
TE 12

TE 21

3.054

TE 22
TE 3

6.706
4.201

TM 21
,

TM 31
TE*,

TE 51

Tai, On the Nomenclature of TE

IRE, 49: 1442-1443 (September 1961).

C-T

3.832

1.64r

7.016

0.89r
2.06r

5.135

1.22r

1.18ro

0.94/v,

1.49r

6.379

0.98r o

7.588

0.83r

5.318

TM 4l
6.416

modes

in

1.14/y,

0.89ro
3.41r

5.330

TM 12

1.64i-o

5.520

1.18ro

0.98/o

a Cylindrical Waveguide, Proc.

ELECTROMAGNETICS

558

[CHAP. 13

Root3of^y^=0;
y(fcr)=max

Frequency relativeto cut off for dominant mode (TEn)

FIGURE 13-48
Possible TE and TM modes

in a

C-T Tai.)

At

frequency.

(After

there are 9

one mode

As

modes which

(TM 02 )

hollow cylindrical waveguide as a function of

3 times the cutoff frequency for the

will pass (see text

TEt mode
,

paragraph preceding Sec. 13-17) and

at cutoff.

in the rectangular guide, there are three conditions:

2
3

At low frequencies, m small, y real, guide opaque (wave does not propagate).
At an intermediate frequency, w intermediate, y = 0, transition condition (cutoff).
At high frequencies, to large, y imaginary, guide transparent (wave propagates).

Putting y

in (20),

we

find for the cutoff frequency

Jc
"2-K.yJ''fie

*:

and

(Hz)

cutoff wavelength

(21)

2nr

Kc

(m)

(22)

TE n mode

For the

k'

cutoff wavelength for the

The

of the guide.

Table

listed in

AND RESONATORS

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-16]

559

= k'tl = 1.84, so that k^ = 2?tr /1.84 = 3.41r

Thus, the
TE n mode corresponds to a wavelength 3.41 times the radius
.

cutoff wavelengths for various

modes

in a cylindrical guide are

13-7.

At frequencies above cutoff

2

p=Ja>* ie-(^)

(radm- 1 )

From

(23)

and

(22)

we

get for the wavelength in the guide (in z direction)

(m)

"

where 1
X oc

velocity in the guide (v

i^olKc)

_m_
'

where

of same type that

=fkg ) we

(24)

= wavelength in unbounded medium

= cutoff wavelength, m

For the phase

(23)

guide,

obtain

v <>

(ms"

Vi-aoMoc)

fills

(25)

l/y/fie

Equations (24) and (25) are identical to those derived earlier for the rectangular
waveguide (see Table 13-6). They also apply to waves in hollow guides of any
cross section.
It

will

be noted that the roots k' (also called eigenvalues) are not regularly

spaced, in contrast to the case for the rectangular guide.

and cutoff wavelengths for some

modes

TE

and

are also illustrated in Fig. 13-48.

frequency than any other mode (including

mode for a

cylindrical guide.

The TE

TM

modes

in

Table 13-7 gives the roots

a cylindrical guide.

The TE n mode

will

TM modes) and hence

mode

is

These

propagate at a lower
is

called the dominant

of interest because of its low attenua-

tion characteristics in practical guides having finite wall conductivity.

For

this

mode

the attenuation decreases monotonically with increasing frequency (see Sec. 13-19).

Referring to Fig. 13-48,

pass the dominant

TE41

An

additional

that if the frequency

is

3 times that required to

the guide will pass the following

modes

TE n TM 01 TE 21 TM n TE 01 TE 31
mode (TM 02 ) is at its cutoff frequency.

the dashed line in Fig. 13-48):

and

we note

mode (TE n ),

(to the left

TM 21

of

TE, 2

560

ELECTROMAGNETICS

[CHAP. 13

(a)

FIGURE

13-49

Forms of hollow single-conductor waveguides showing how square

(6), cylin-

and elliptical (d) shaped waveguides may be derived from rectangular

(a).
Types (e) and (/) may also be derived from (a).

drical (c),

type at

13-17

HOLLOW WAVEGUIDES OF OTHER CROSS SECTION

In earlier sections

we considered

These are

rectangular and cylindrical waveguides.

only two of an infinite variety of forms in which single-conductor hollow waveguides

may be made. For example, the waveguide could have an ellipticalf cross section, as
in Fig. 13-49d, or a reentrant! cross section, as in 13-49/.

All these forms and

many

tangular type (Fig. 13-49a).

others

may

be regarded as derivable from the rec-

Thus the square cross section

case of the rectangular guide.

By

(Fig. 13-496)

is

a special

bending out the walls the square guide can be

transformed to the circular shape (Fig. 13-49c). By flattening the circular guide the
On the other hand, by bending the top and
elliptical form of Fig. 13-49d is obtained.

bottom surfaces of the rectangular waveguide inward the form shown in Fig. 13-49e
A still further modification is the reentrant form in Fig. 13-49/. The
is obtained.
value of regarding these as related forms

is

that often certain properties of a guide of a

J. Chu, Electromagnetic Waves

9 (September 1938).

t L.

X S. B.

in

Hollow

Elliptic Pipes

of Metal,

J.

Appl. Phys.,

Cohn, Properties of Ridge Wave Guide, Proc. IRE, 35: 783-789 (August

1947).

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-18]

particular shape can be interpolated approximately

AND RESONATORS

561

from the known properties of

waveguides of closely related shape.

For example, the longest wavelength that the square guide (Fig. 13-496) will
This is for the TE 10 mode. This information can be used to
is equal to 2b.

transmit

predict with fair accuracy the longest wavelength that a circular guide can transmit.

Thus,

if

the cross-sectional area of the square guide

is

taken equal to the area of the

circular guide,

'
where d is the diameter of the circular guide.
or b = A oc /2, and so we have

(1)

Now koc =

2b for the square waveguide

-
X oc

or

= s/nd=\ .lid =

(2)

3.54r

(3)

as the cutoff wavelength for a circular waveguide of diameter

approximate value

is

d or radius

r.

This

only 4 percent greater than the exact value of 3.41 r (see Table

13-7).

The procedure
guide of any shape

is

for carrying out a complete analysis of the properties of a wave-

formally the same as in the preceding sections for the rectangular

and circular waveguides.

usually most convenient, however, to set up the equations

It is

in a coordinate system such that the waveguide surfaces can be specified by a fixed

value of a coordinate.

Thus, as we have seen, a rectangular guide

is

conveniently

handled with rectangular coordinates, the guide surfaces being specified by

y = yu

= 0,

and z

zt

Likewise, a circular waveguide

cylindrical coordinates, the guide surface being specified

Fig. 13-49rfand e, these shapes

can

is

by

y=

0,

readily analyzed using

r

Referring to

be analyzed using elliptical-hyperbolic coordinates.

However, if the guide surface cannot be specified in a simple manner, the application
of the boundary condition (E, = 0) may be very difficult.

13-18

ATTENUATION AT FREQUENCIES
LESS THAN CUTOFF

It has been shown that at frequencies less than cutoff, waves are not transmitted
through hollow single-conductor guides but are attenuated. Let us now calculate the

magnitude of

this attenuation.

From

(13-15-72) the attenuation constant for a

rectangular guide at frequencies less than cutoff

is

M+'-'

562

[CHAP. 13

ELECTROMAGNETICS

Noting (13-15-75), we can reexpress

this as

(Npm "' )

-^O'-'-W'-'
where A
X oc

(2)

= wavelength in unbounded medium, m

= cutoff wavelength, m

rectangular guides but to

attenuation constant a as given in (2) applies not only to
hollow single-conductor guides of any cross-sectional shape.

The

If the

>

frequency is much less than cutoff (X

A oc), Eq. (2) reduces to the approxi-

mate relation

^
where X oc

is

(Npm-

(3)

Since, in dealing with voltage,

the cutoff wavelength in meters.

Np =

8.68 dB,f

2n x 8.69

54.6

(dBm" 1 )

EXAMPLE A certain waveguide has a cutoff wavelength A

ation per meter along the guide for

solution

Since A

>

a
This

is

of 100

an applied wavelength A of

A oc Eq.

(3)

or

(4)

= 20tt Npm"

1
,

(4)

mm.

Find the attenu-

m.

can be used, yielding

or 546

a very high rate of attenuation, the applied

dBm'

field falling

to a negligible value in a very

short distance.

cutoff is illustrated in
simple attenuator operating at frequencies less than
metal tube, acting as a waveguide, has loops
longitudinal section in Fig. 13-50.
transmission lines into and
arranged at each end, as shown, to couple from coaxial
on a movable plunger so that the
out of the waveguide. One of the loops is mounted
If the applied wavelength X is much longer
distance between the loops is variable.
not in too close proximity,
than the cutoff wavelength l oc of the guide and the loops are
and X is
For instance, if koc = 100
the attenuation is as given by (3) or (4).
of
centimeter
dB
per
increases
5.45
or more), the attenuation
much greater (1
the
but
has
useful
very
is
attenuator
outward movement of the plunger. This type of

mm

means a reduction to l/ of the original

1 neper (1 Np) attenuation
times tne
Conversely, an increase of 1 Np means an increase to e(=2.7183)
Hence, for voltages 1 Np is equal to 20 log e=8.69 dB.
original value.

Note that a
value

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-19]

AND RESONATORS

563

f Moveable plunger

Coupling loops N

^Coaxial

^Circular wave guide

^Coaxial line

FIGURE

13-50

Attenuator for use at frequencies

disadvantage of a high insertion

in a coaxial line.

Since

fi

0,

loss, i.e.,

there

is

line

a large

no change

initial

in

attenuation

less

than cutoff.

when

inserted

phase with change in plunger

position.

ATTENUATION AT FREQUENCIES
GREATER THAN CUTOFF

13-19

If the

waveguide has perfectly conducting walls and the medium

lossless, there is

no attenuation

at frequencies greater

than

filling

cutoff.

the guide

Thus, a

= 0,

is

and

from (13-15-71) we have for a rectangular guide that

or

The phase constant f} as given in (2) applies not only to rectangular guides but to hollow
single-conductor guides of any cross-sectional shape.

The behavior of the phase constant /? for this case and for the case discussed in
compared on the composite graph in Fig. 13-51. Here the propagation
constant y is shown as ordinate vs. A in an unbounded medium as abscissa. The real
Sec. 13-18 is

part of v(=a)

is

plotted as the solid curve above the

(=/?) as the dashed curve below the axis.

zero,

and

ft

approaches an

infinite

value that

At very
is

axis

and the imaginary part

short wavelengths (A

equal to

/?

for an

-+0), a

is

unbounded medium.
At still longer wave-

= Xoc) f} is zero.
j8 decreases until at cutoff (A
remains
zero,
but
a
does
not.
At
sufficiently
long wavelengths (A > A oc), a
fi
approaches a value of 2n/X 0C as indicated. This diagram applies to lossless hollow
As A

increases,

lengths,

single-conductor guides with cross sections of any shape.

Actual guides are not

Fig. 13-51.

copper,

/? is

However, for

lossless, so that
air-filled

is

not zero for X

< loc

as indicated in

guides of a good conducting material, such as

substantially as indicated for X

< Xoc

while a

is

small but not necessarily

564

[CHAP. 13

ELECTROMAGNETICS

FIGURE

13-51

Composite graph showing attenuation constant a and phase constant /3 for a

of the wavelength A
lossless hollow single-conductor waveguide as a function
At wavelengths less than the cutoff wavelength
in an unbounded medium.
At wavelengths longer
(A c) there is no attenuation (a = 0) but a phase shift p.
than cutoff there is no phase shift (J3 = 0) but an attenuation a.

negligible.

To calculate a for this case, we note (see Fig.

in the guide varies with the distance

P=P

It

factor 2 in the exponent

is

power

in the direction of transmission as given by

e- 2 * x

(W)

(3)

present because power

is

guide,

proportional to

field

squared.

follows that

2
In

3-52) that the average

P = average power at reference point (x = 0), W

x = distance in direction of transmission through

where

The

(4),

-dP/dx represents

particular location, while

- dP/dx
P

the decrease in

is

(NpnT

power per

(4)

unit distance along the guide at a

the power transmitted through the guide at that loca-

tion.f

Thus, in words the attenuation constant in nepers per unit distance is expressed by

power

~
t

lost per unit distance

twice the power transmitted

this case is due to an actual power loss

It is to be noted that the attenuation in
than cutoff no joule-heating
(joule heating of guide walls), whereas at frequencies less
inability of the guide to transmit
effect is involved, the attenuation being due to the

the higher-order

mode (wave

reflected).

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-19]

AND RESONATORS

565

xf(Sav)

Guide
walls

FIGURE
Power

fl(2||

dx13-52

.Hn

lost in walls

medium

>

direction

of waveguide results in

x=0

\(Sm ).

attenuation.

If the

Wave

"(Sav)z

|-

filling

the guide

the decrease in power per unit distance

is lossless,

equal to the power lost per unit distance in the walls of the guide.

dP

jy

This

dl

where (Sav ) w

The

is

(5)

the average Poynting vector into the wall (average with respect to time).

surface integral in (5)

waveguide (Fig.

3-52).

is

taken over a strip of length dx of the interior surface of the

The

line integral in (5) is

(same path as that for strip).

taken around the inside of the guide

In general, the average Poynting vector

is

Sav = iRe E x H*

Since

and

medium

is

(see Fig.

3-52)

Hn H \Z

Hn

Re Zc

(7)

absolute value (or magnitude) of component of

H tangent to conduct-

(Sav)w

\Hn =

(6)

are normal, the magnitude of the average Poynting vector into the

conducting wall

where.

is

is

i Re

2
|

ing surface of guide walls

Re Zc =
Introducing

real part of intrinsic

impedance of conducting wall medium

= x//ku/2ct

(7) in (5) yields

-i-^Ji"i"
Now the

power

traveling through the guide (in

P = jj(S)x ds
where

in this case the surface integral

is

direction)

is

(W)

(9)

taken over the guide cross section.

It

follows
'

that

P = iRcZyz jj\H, 2

2
\

ds

(10)

566

ELECTROMAGNETICS

where

t2

[CHAP. 13

absolute value of

through guide

Re ZyI =

real part of transverse

Therefore, the attenuation constant a

plane

impedance of guide
in general, given

is,

Re Zc )\H

tl

"

H tangent to a cross-sectional

component of

(Fig. 13-52)

dl

(Npm -1 )

2ReZyz tf)H, 2

dS

by

(11)

Zc = real part of intrinsic impedance of guide walls (conductor)

Re ZyI = real part of transverse impedance of guide

where Re

Hn

guide walls (integrated around

|

t2

H tangent to

absolute value of component of

conducting surface of

interior surface of guide)

absolute value of component of

tangent to plane of cross section

through guide (integrated over cross-sectional area) (Fig. 13-52)

Equation (11) applies to any mode in any guide.

stant

must be calculated using

(1 1),

with values of

the field distribution for that mode.

material,
is

we may assume, with but

same

the

For each mode the attenuation con-

tl

and

l2

If the guide walls are of

little

corresponding to

good conducting

error, that the H-field distribution used in (1 1)

The following example

as for perfectly conducting walls.

illustrates

an

application of (11) to a simple problem.

EXAMPLE

Find the attenuation constant for a 300-MHz

TEM

parallel-plane transmission line with a spacing between planes of 100

walls are

made of copper, and

For a

solution
so

(1 1)

the

medium between

the planes

wave

in

an

mm. The

infinite-

planes or

is air.

TEM wave the transverse impedance equals the intrinsic impedance;

becomes

2ReZe

r'

\H,i\

dy
(12)

2ReZ,f

\Ht2

2
\

dzdy

= arbitrary distance along conducting wall (see Fig. 13-27)

= spacing between walls, m
Re Zc = real part of intrinsic impedance of conducting walls, H
Re Zi = real part of intrinsic impedance of dielectric medium between walls (Zd is entirely

where yi
zi

real for lossless

The
in

integral with

both walls

is

H, s

medium)
power

involves the

lost in

one wall of the

The total power lost

For a TEM wave H is

line.

twice this; hence the factor 2 in the numerator.

everywhere parallel to the walls and normal to the direction of propagation, so that both

TRANSMISSION UNES, WAVEGUIDES,

sec. 13-19]

AND RESONATORS

567

0.000001

GHz

MHz

FIGURE

THz

13-53

Attenuation a in decibels and nepers per meter as a function of frequency for

various modes in an air-filled infinite-parallel-plane waveguide of copper with

100-mm spacing between planes. Note that the TEM mode and the TM m0
modes shown cannot occur in a single hollow rectangular guide but the TE
modes can. At frequencies below cutoff for the TE 10 mode (1.5 GHz) only the
TEM mode is transmitted but many TE and TM modes can be transmitted above
and at sufficiently high frequencies the TE modes have the lowest attenuation.

H,i and H, 2 are perpendicular to the page instead of as suggested

= \H 2 = a constant. Hence, (12) reduces to

that \H,i

in Fig. 13-52.

It

follows

For copper

at 300

MHz, Re Zc

- yi = KeZ
-
= -KeZ
z Re Z
Re Z y Zi
c

(13)

= 4.55 mfJ, while for air Re Z = 376.7

= 1.2x lO-'Npma = 1.04x lO-'dBm-

CI.

Therefore

<x

or

Thus, the attenuation amounts to about

The attenuation

in decibels per

dB km"

is

shown

in Fig. 13-53.

meter (and in nepers per meter) as a function of

frequency for a number of modes in an

copper

air-filled infinite-parallel-plane

The spacing between planes is 100 mm.

guide of

We note that

568

there

ELECTROMAGNETICS

is

[CHAP. 13

TEM

no cutoff for the

decreasing frequency.
tion although the

mode and

However,

TE 20 and

and higher-order modes

at 100

attenuation decreases for this

GHz the TE 10 mode has the

higher-order

modes can

also be transmitted.

have higher attenuation than the

all

TE 10

mode with

lowest attenua-

The TE 20

mode.

EXAMPLE

2 Find the attenuation constant for the TE r mode in a circular cylindrical

waveguide of radius a at frequencies above cutoff. Determine this attenuation in decibels
per meter for the TE i and TE n modes in a 50.8-mm-diameter copper pipe.

From

solution

(13-16-16)
2

\H, l

and (13-16-17)

\H,(r

and from (13-16-15) and

\H, 2

a TE r mode

= a)\ 2 +\H^r = a)\ 2

Ho

for

2
|

J\ka) Los 2 n$

+ M^ sin 2 n<f\

(13-16-16),

= \H

2
\

+\Ht

2
\

=!M^ ^^)]V
cos

sin2

MW)

Substituting (14)

and

J- ,(t "' [/,''. "' "* <"> +

2"

and

n<j>

d<p

/.''.

"-'

* *']
(16)

2*

sin

= 0,

n^O
n=

(0
(C

n</> d<f>

[it

Jo

TE modes

(2tt

cos 2

Jo

for

that
r

Thus

(15)
}

(15) into (11) gives

* R. Z,

We note

(14)

=1

[*

1, 2, 3,

.)

2
J (.ka)TT[\ +(P n 2 /k*a 2 )1
_ k* Re Z a
2
2^ KeZ'*rildMkr)V^nl
rf\ dJJjkr)Y n rU
I2rl^l^
J
2

The

(17)t

integral in the denominator of (18) is evaluated in E. V. Bohn, " Introduction to

Electromagnetic Fields and Waves," p. 446, Addison-Wesley Publishing Company,
Inc., Reading, Mass., 1968.

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-19]

Evaluating (17) and referring to Tables 13-6 and 13-8 for

Re Z *
r

^^

ReZ

filled

= Virplo V/ = 2.63 x
The

waveguide.

Evaluating (18) for

value of

TE

and

k' r

lO

-7

V/(Q)

TE U modes

TE

increases.

This trend also occurs for any

minimum at///c
which n

(low-loss)

mode

and Re Zt

(18)

= 377 n

for an air-

13-7.

yields the attenuation curves of Fig. 13-54.

that for the

for copper

can be found from Table

569

gives

~aReZ<Vl-(fc [f)> [ffi + VZF^*]

where Re Zc

AND RESONATORS

Note

the attenuation decreases montotonically as the frequency

= 3.1. A minimum

TE 0r

The TE 1X (dominant) mode has a

TE modes for

mode.

at this///, ratio also occurs with all

^ 0.

Parameters of rectangular and cylindrical waveguides are summarized in

Table

13-8.

Table 13-8

RECTANGULAR AND CIRCULAR CYLINDRICAL

WAVEGUIDE PARAMETERS

Symbol

Name

Attenuation constant:

Equation (rectangular or cylindrical)

Below cutoff

Above

cutoff

Phase constant

vp

Phase velocity

(Np m - 1 )

A/ \\cj

ReZc S\H

tl

frequency

|8t

"

frequency

\*dl
z

2ReZtt\Ht2

f-M-
"'

CO

Bo

Vi-(A

id

Transverse-wave

"

impedance

Cutoff wavelength

y = propagation constant

=a +jfi.

Cylindrical

Z<
Vi-(A

Z<
/A c)

'*

TEm

(ms-

TE mode only
A

m _I

/A 0C )>

Rectangular

(Npm-

ds

mode

(O)
/A c)

TE r mode
.

A oc
or TM,

Vi-(A

2nr

2irr
= -p-OT-r

TE r mode

(k'),

(m)

TM, mode (k)

570

ELECTROMAGNETICS

[CHAP. 13

0.04

0.03 -

m
S

0.02

FIGURE

0.01-

13-54

Attenuation a in decibels per meter as a

function of frequency for TE i (low-

TEn (dominant) modes in a

copper waveguide of 50.8
inside diameter (see Example 2).
and

loss)

mm

circular

Frequency(GHz)

WAVEGUIDE DEVICES

13-20

Several basic waveguide devices are discussed in this section, namely, terminations,

power

and guide-to-line transitions. Reference should be made to more

books for detailed treatments of these and other waveguide devices.!
matched termination' for a rectangular or circular waveguide is shown in

dividers,

specialized

cross section in Fig. 13-55a.

card of resistance material

is

placed transversely in

the guide A/4 from a metal plate capping the end of the guide.

The

similar to that discussed in Sec. 12-3 for the terminated wave.

For zero

is

situation here
reflection

is
it

necessary only that the card have a resistance per square equal to the transverse-

wave impedance of the

guide.

The termination of Fig. 13-55a

design frequency for which the card-to-end-plate distance

frequencies.

This termination

is

a narrowband device.

is

will

be matched

at the

A/4 but not at adjacent

To provide a broadband

termination a wedge of resistance material can be used, as suggested in Fig. \3-55b.

The length of the wedge should be of the order of a wavelength or more.

In applications where waveguides feed two or more antennas a power divider may
be required to divide the power in a predetermined ratio. Figure 13-55c shows a
power divider for a rectangular waveguide (with TE 10 mode) delivering twice the
power to the lower branch as compared to the upper. The division is achieved by
inserting a thin septum which divides the guide cross-sectional area in the ratio 2:1.
Note that the septum is perpendicular to the direction of the electric field vector E
so that it does not disturb, the field configuration in the guide. The height of the
guide to the right of the septum in each branch
t

is

increased gradually over a dis-

N. Marcuvitz, " Waveguide Handbook," McGraw-Hill Book Company, New York,

1949; J. C. Slater, "Microwave Electronics," D. Van Nostrand Company, Inc.,
New York, 1950.

571

TRANSMISSION LINES, WAVEGUIDES, AND. RESONATORS

sec. 13-21]

SIDE VIEWS

END VIEWS

Narrow-band termination

FS3

/U

Resistance card'

'

<)

x/4

Broadband termination

Septum

Lower branch
Rectangular guide to coaxial

line

(d)

TE 10 mode
Circular guide to coaxial line
()

FIGURE
(a)

13-55

and (6) Waveguide terminations,

to-coaxial-line transitions.

(e)

power

divider,

The arrows with

and (</) and (e) waveguide-

indicate the direction of the

electric fields in the guides.

tance of several wavelengths back to the standard height h.

connected into nonreflecting loads or antennas, there

will

If

both branches are

be no reflection from

this

power divider and the device can be used over a broad band of frequencies.
At some point in most waveguide systems it is necessary to convert to TEMmode transmission on a coaxial line. Two waveguide-to-coaxial-line transitions are
shown in Fig. 1 3-55d and e. The one in Fig. 1 3-55d provides a transition from a TEj
mode in a rectangular guide to a coaxial line, while the one in Fig. 13-55e provides a
transition from a TM 01 mode in a circular waveguide to a coaxial line.
13-21

WAVEGUIDE

A partial opening, or iris,

IRIS

in a

THEORY

waveguide presents a discontinuity which

to placing a susceptance across a transmission line.

distance from a load or other discontinuity,

the

iris

can act as a matching device.

and

if

If the iris

is

is

analogous

located the proper

the susceptance of the

iris is

correct,

572

ELECTROMAGNETICS

[CHAP. 13

Opening

-7\z

Iris

(a)

Opening

(6)

FIGURE

13-56

Rectangular waveguide with three types

of iris openings equivalent to (a) shunt

t.

capacitance, (6) shunt inductance, and

(c)

(e)

a parallel-tuned circuit.

Consider the cross section of a rectangular guide (with

slot

openings as in Fig. 13-56.

sented by the

iris is

It

TE 10 mode)

and

iris

or

can be shownf that the normalized susceptance pre-

given by

B=j2Z 2

\\H -H' ds
t

(1)

l\E -E^ds
t

= transverse field components of dominant mode in guide (without iris)

= transverse components of the actual field at iris
El,
The quantity Z is the characteristic impedance of the guide.J The numerator is integrawhere

t ,

H,
H'
t

iris, and the denominator is integrated over the

and H, can be calculated simply from the guide
geometry, but the determination of the actual field components (E[ and H' ) is more
difficult since they are the resultant of many higher-order modes which exist in the

ted over the metallic surface of the

opening.

The

field

components

proximity of the

iris.

t Slater, ibid., p. 118.

%

characteristic impedance usually implies a two-conductor transmission

can be extended to the dominant mode in a waveguide. For example, in
a rectangular waveguide with TE l0 mode it is equal to the ratio of the maximum

The term
line

but

it

voltage across the guide to the wall current.

If the

in (1)

is

opening

small compared to the metal area of the

is

For the
(positive),

and the

3-566 with

acts like

iris

at all),

As

large.

effect is like

is

full

the denominator

reduced to zero,

horizontal slot the susceptance

a capacitor across a transmission

full vertical slot

the susceptance

an inductance across a transmission

is

For the

line.

made

properly, the total, or net, susceptance can be

a parallel-tuned

circuit across

bandpass-filtering action with

no

is

a transmission

if

of

iris

and the

slot

of Fig. 13-56c there

the dimensions are chosen

zero.

line,

iris

capacitive

For the

line.

inductive (negative),

are both capacitive and inductive susceptances, and

like

iris,

the opening

goes to zero.

of Fig. 13-56a with

iris

is

573

Conversely, a large opening gives a small B, and as the opening

infinity.

becomes complete (no

small and the susceptance

approaches

Fig.

AND RESONATORS

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-22]

In this case the effect

is

and the device provides a

reflection (complete transmission) at the resonant or

design frequency.

13-22

INTRINSIC, CHARACTERISTIC,

AND WAVE IMPEDANCES

We

have dealt with three types of impedances:

The intrinsic impedance refers to

an unbounded medium. Thus, for a

intrinsic, characteristic,

E to H for

impedance.

the ratio of

wave

lossless dielectric

in

Intrinsic

For vacuum or

impedance

impedance

For a good conductor the

/

on an

(TEM)

medium

ill

= Zd = =

/-

(2

per square)

(1)

air

Intrinsic

The

and wave

a plane

characteristic

infinite

= Zd = Z

intrinsic

=/ = 376.7

(Q per square)

impedance equals

Zc = ,/ ficoja

/ 4S

impedance refers to the ratio of the voltage

two-conductor transmission

integral of the electric field

(2)

v eo

line.

V to

The voltage V

is

the current

equal to the

E along any path between the conductors, while the current

around one of the conductors (Amimpedance of a two-conductor line can also be

expressed in terms of the series resistance R, series inductance L, the shunt conductance
G, and shunt capacitance C. Thus, for a two-conductor line

is

equal to the integral of the magnetic field

pere's law).

The

characteristic

Characteristic

impedance

=Z

jE-dl

~I~H-dl

JR+jcoL
\IG+j<oC

(fi)

(3)

574

ELECTROMAGNETICS

[CHAP. 13

For a transmission-line
impedance EjH.

cell

the characteristic impedance V/I equals the intrinsic

The wave impedance refers to the ratio of an electric field component to a magnetcomponent at the same point of the same wave. For a TEM wave the wave
impedance is the same as the intrinsic impedance, but for higher-order modes, as in a
hollow single-conductor waveguide, there can be as many wave impedances as there
are combinations of electric and magnetic field components.
In the preceding sections
we have confined our attention to the transverse-wave impedance, e.g., the ratio of
ic field

Ey to H

for a

direction of

TE mode

in a rectangular guide (both

wave propagation

x).

Ey

and

are transverse to the

Thus,

Transverse-wave impedance

= Zyz

(il)

(4)

In a cylindrical waveguide the transverse-wave impedance equals Zr<^(= E,/H^).

The transverse- wave impedance of a waveguide is a function of the intrinsic impedance
of the medium filling the guide and the guide dimensions as given in Tables 3-6 and
13-8.
As the transverse guide dimensions become very large compared to the wave1

length, the transverse-wave

impedance of the guide approaches the

intrinsic

imped-

ance of the medium.

Whereas the characteristic impedance is basically a circuit quantity (= V/I), the

impedance and the wave impedance are field or wave quantities involving the

intrinsic

ratio of electric to magnetic field

13-23

components.

WAVES TRAVELING PARALLEL

TO A PLANE BOUNDARY

In Sec. 13-19

we

considered the attenuation due to the power lost in the walls of a

phenomena associated with this power loss or

power flow are discussed in more detail.
Consider the plane boundary between two media shpwn in Fig. 1 3-57'a, assuming
that medium 1 is air and medium 2 is a perfect conductor. From the boundary condition that the tangential component of the electric field vanishes at the surface of a
perfect conductor, the electric field of a TEM wave traveling parallel to the boundary
must be exactly normal to the boundary, as portrayed in the figure. However, if
waveguide.

medium

In this section some of the

2 has a

finite

conductivity c, there will be a tangential electric field

boundary, and, as a result, the

electric field

& forward tilt, as suggested in Fig.

electric fields, the field

on both

Ex

at the

of a wave traveling along the boundary has

3-576.

sides of the

From the continuity relation for tangential

boundary

is

Ex

AND RESONATORS

TRANSMISSION LIMBS, WAVEGUIDES,

sec. 13-23]

Medium

Medium

Direction of

(air)

575

(air)

energy flow
Direction of

energy flow

N90

Hz
:.

Medium 2

(perfect conductor):
<T

Medium 2 (conductor

= 00

finite

with
conductivity cr)

(b)

(a)

FIGURE

13-57

TEM wave traveling to right (a) along surface

and

The

direction

(6)

along surface of

medium

of perfectly conducting

medium

finite conductivity.

and magnitude of the power flow per unit area are given by the

The average value of the Poynting vector

Poynting vector.

Sav =

Sy
relation of

Ex

is

(1)

(Fig. 13-576) the

power

into the conductor

is

average value per unit area isf

its

=-iReEx H*

(or H*),

(Wm"

iReExH*

At the surface of the conducting medium

in the negative y direction, and from (1)

The space

with

and Sy

is

(2)

shown

in Fig. 13-58a.

But

(3)

where

Zc

is

the intrinsic impedance of the conducting

written

S,= where

W H* Re Z
z

= Hz0 eJ ^'' i) -''x

~
H* = Hz0 e~ lJl "~ i) yx = complex conjugate
= phase lag of Hz with respect to E x

relation for the Poynting vector in (4)

At

the surface of the conducting

is

the

medium

flowing parallel to the surface (x direction)

that (2) can be

= - 1//20 2 Re Zc

(4)

'

The

medium, so

of

same

as given in (13-19-7).

(Fig. 13-576) the

power per

Sx = i Re Eyfft
t

unit area

is

(5)

The component of the average Poynting vector

simplify notation we shall write 5, for (Sav) y
.

in the

y direction

is

(S, v ),,, but to

"

576

ELECTROMAGNETICS

[CHAP. 13

Medium

(air)

Medium 2
(a)

FIGURE

(6)

(c)

13-58

Fields and Poynting vector at surface of a conducting

medium with wave traveling

parallel to surface.

The space

relation of

Ey

(or //*),

and Sx

is

illustrated

by

But

Fig. 13-586.

wr z

(6)

<

where

Zd is the

intrinsic

impedance of the

Sx =
The

total average

Poynting vector

is

relation of

Sav

to

its

zo

inward at an angle

Hzo
^-

(SL

x and y components

t.

This angle

field (see Fig.

is

It

follows that

ReZd

(7)

ReZ, is

also the

(8)

is

to be

It is

not parallel to the surface but

same as the angle of forward

If medium 2

3-576).

y ReZ,)

illustrated in Fig. 13-58c.

noted that the average power flow (per unit area)

average electric

(air).

then

Sav = kSx + yS =
The

medium

dielectric

tilt

were perfectly conducting,

of the

would

be zero.
It is

This

is

of interest to evaluate the

done

EXAMPLE
in air

Find the forward

tilt

angle t for a vertically polarized

From

(8) the

tilt

angle t

t
3

angle t for a couple of practical situations.

3-GHz wave

traveling

along a sheet of copper.

solution

At

tilt

in the following examples.

GHz, we have

is

= tan

(Re Z4
T

.,

ReZ

= tan

llf

(9)

ReZ,

Re Z = 14.4 mil.
= 376.7 Ci.) Thus

for copper that

independent of frequency.

given by

The

l7^= 000220

intrinsic

impedance of

air is

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-23]

AND RESONATORS

577

Although t is not zero in the above example, it is very small, so that E is nearly
normal to the copper surface and S nearly parallel to it. This small value of tilt is
typical at most air-conductor boundaries but accounts for the power flow into the
conducting medium.

medium,

If the conductivity

may amount

of medium 2

is

very low, or

a dielectric

if it is

few degrees.

Thus the forward tilt of a vertically

polarized radio wave propagating along poor ground is sufficient to produce an appreciable horizontal electric field component.
In the Beverage, or wave, antenna this
horizontal component is utilized to induce emfs along a horizontal wire oriented
z

to a

parallel to the direction of transmission of the wave.

Example

In contrast to

1,

which medium 2

in

considers the case of fresh water as

EXAMPLE
in air

Find the forward

tilt

medium

is

copper, the following example

2.

angle t for a vertically polarized

3-GHz wave

traveling

along the surface of a smooth freshwater lake.

solution

At

GHz

the conduction current in fresh water

with the displacement current, so that the lake

relative permittivity e,

= 80.

may

is

negligible

be regarded as a dielectric

compared

medium of

Thus
1

= tan -1 -F=

In this case the forward

tilt

of 6.4

of the direction of the electric

The angle

t discussed

=6.4

V80

sufficient to

is

be readily detected by a direct measurement

field.

above

is

an average value.

In general, the instantaneous

direction of the electric field varies as a function of time.

In the case of a

wave

in air

Ey and Ex are in phase octature (45 phase difference),

of time the total field E may be in the x direction and period

traveling along a copper sheet,

so that at one instant

later

it

will

be in the y direction.

(See Fig.

E describes a cross-field ellipse (see Sec.

wave

in air traveling along a

ellipse is

of

copper sheet

With time the locus of the

3- 59a.)

1-6), as

(in the

direction) as in

not to scale, the abscissa values being magnified 5,000 times.

for various values of

Poynting vector for

magnified 5,000 times.

around the

wt are indicated.

this case is

ellipse twice

It is

shown

tip

of

3-GHz
Example 1
The

portrayed in Fig. 13-59* for a

The

in Fig. 13-60.

The

positions

variation of the instantaneous

Here the ordinate values are

to be noted that the tip of the Poynting vector travels

per cycle.

Whereas copper has a complex intrinsic impedance, fresh water, at the frequency
considered in Example 2, has a real intrinsic impedance. It follows that the E and
x
Ey components of the total field E are in time phase so that the cross-field ellipse in
this case collapses to

a straight line (linear cross

field)

with a forward

tilt

of 6.4.

578

[CHAP. 13

ELECTROMAGNETICS

r= 0.002'
Relative
-electric
field

intensity

(a)

FIGURE

13-59

Magnitude variation with time of E, and E components of E in air at the

surface of a copper region for a 3,000-MHz TEM wave traveling parallel to the
surface,
(b) Resultant values of E (space vector) at 22.5 intervals over one
The
cycle, illustrating elliptical cross field at the surface of the copper region.
wave is traveling to the right. Abscissa values are magnified 5,000 times com(a)

pared with the ordinate values.

FIGURE

13-60

Poynting vector in air at a point on the

surface of a copper region for a 3,000MHz TEM wave traveling along the surface (to right). The Poynting vector is
shown at 22.5 intervals over one-half
cycle.
The ordinate values are magnified
5,000 times compared with the
67.5

abscissa values.

13-24

OPEN WAVEGUIDES

we have seen that a wave traveling along an air-conductor or

boundary
has
a longitudinal (Ex) component of the electric field, resulting
air-dielectric
Hence the Poynting vector is not entirely
total
electric field.
the
forward
tilt
of
in a
In the previous section

TRANSMISSION LINES, WAVEGUIDES,

SEC. 13-24]

parallel to the

medium,

AND RESONATORS

579

boundary but has a component directed from the air into the adjacent
This tends to keep the energy in the wave from

as suggested in Fig. 13-58c.

spreading out and to concentrate

surface wave.
in free space.

it

near the surface, resulting in a bound wave, or

The phase velocity of such a bound wave is always less than the velocity
Although the field of this guided wave extends to infinity, such a

large proportion of the energy

may be confined within a few wavelengths of the surface

that the surface can be regarded as an open type of waveguide.

however, that even though the forward-tilt

surfaces, the

effect is present

It

along all

should be noted,

finitely

conducting

bound wave may be of negligible importance without a launching device

of relatively large dimensions (several wavelengths across) to

initiate

the wave.

If

smooth and perfectly conducting, the tangential component of

the electric field vanishes, there is no forward tilt of the electric field, and no tendency
whatever for the wave to be bound to the surface.
In 1899 Sommerfeldf showed that a wave could be guided along a round wire of
finite conductivity.
Zenneckf pointed out that for similar reasons a wave traveling
along the earth's surface would tend to be guided by the surface.
The guiding action of a flat conducting surface can be enhanced by adding corthe surface

is

perfectly

rugations or a dielectric coating or slab.

can act alone as an

open waveguides are discussed

these

If the dielectric slab

effective nonmetallic guide.

flat

conducting corrugations, as in Fig. 13-6 la.

The

(s -4 A).

slots

vertically with electric field

sufficiently thick,

surface of infinite extent with transverse

The corrugations have many teeth per

TEM wave traveling

between the teeth can support a

component

Ex

Thus, each

slot acts like

a short-circuited

section of a parallel-plane two-conductor transmission line of length

lossless materials, the

impedance

onto the corrugated surface

is

Assuming

a pure reactance, or
(n)

(1)

Zd = intrinsic impedance of medium filling the slots, fi

er relative permittivity of the medium, dimensionless
A

= free-space

wavelength,

d = depth of the

slots,

The slots may be regarded as storing energy from the passing wave.

<

d.

Z presented to a wave traveling vertically downward

Z=jZd tan l^d

where

it

number of

in this section.

Consider a perfectly conducting

wavelength

The

is

characteristics of a

90 the surface

is

When Inje, d/X

inductively reactive.

t A. Sommerfeld, Fortpflanzung elektrodynamischer Wellen an einem zylindrischen

Leiter, Ann. Phys. u. Chem., 67: 233 (December 1899).

Zenneck, Uber die Fortpflanzung ebener elecktromagnetischer Wellen langs

und ihre Beziehung zur drahtlosen Telegraphies Ann.
Phys., (4), 23: 846-866 (Sept. 20, 1907).

t J-

einer ebenen Leiterflache

580

ELECTROMAGNETICS

[CHAP. 13

Direction of _

guided wave
Total

Slot

Teeth

/ \

TEM wave traveling

in

Conductor

+ and -

direction

(= standing wave)
(a)

I"

rection of

gu idedwave'
'

d
'

'*
'.

'.'

^m

'\

/
Conduct

Dielectric

or

(b)

FIGURE

13-61

Wave

guided by corrugated surface.

dielectric coating of thickness d.
(a)

Consider next a
tric

of thickness

(dominant)

mode
d the

as e~"

Wave

guided by conductor with

perfectly conducting surface with coating of lossless dielec-

The

d, as in Fig. 13-616.f

electric field configuration for

wave launched parallel to the surface

thick coating
y

flat

(6)

fields attenuate

is

shown.

For a

TM

sufficiently

perpendicular to the surface (in the y direction)

where

If

2ti

TV

(Np

m"

(2)

A. Schelkunoff, Anatomy of "Surface Waves," IRE Trans. Antennas Propag.,

AP-7: S133-139 (December 1959); R. F. Harrington, "Time-Harmonic Electromagnetic Fields," p. 168, McGraw-Hill Book Company, New York, 1961.

t S.

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-24]

FIGURE

AND RESONATORS

581

13-62

waveguide showing electric field configuration for dominant mode

and higher-order mode (TM,). The higher mode is equivalent to TEM

Dielectric-slab

(TM

waves which have

For

er

total internal reflection as suggested at right.

2 this gives over 50

dB

attenuation per wavelength.

Assuming a

effectively confined close to the surface.

tion in the

direction will be

now

Consider

slab of thickness d,

due

Thus, the

field is

lossless dielectric, the attenua-

entirely to radiation.

a completely nonmetallic guide consisting of a lossless dielectric

as in Fig. 13-62. The situation here is the same as in Fig. 13-616

but with the conducting surface removed from the lower side of the dielectric slab.

The

electric field configuration for

to the surface
it

shown

is

can be propagated

TM

(dominant)
This

at the left in Fig. 13-62.

at all frequencies

down

mode wave launched parallel

mode has no cutoff; i.e.,

TM

modes
As an
shown at

to zero frequency. Higher-order

(with low-frequency cutoffs) are possible at sufficiently high frequencies.

example, the electric

the right in Fig.

field

3-62.

configuration for the

These higher modes

TM,

may

internal reflection (see Sec. 12-4) of two plane

(higher-order)

mode

is

be regarded as the resultant of total

TEM waves, as suggested in Fig.

13-62.

analogous to that discussed in Sec. 13-14 for higher-order modes

For modes higher than TMj there are
in an infinite-parallel-plane transmission line.
more closed electric field loops inside the dielectric slab, but the field outside is the
same. Higher-order modes are also possible for the fiat conducting surface with

The

situation

is

dielectric coating (Fig.

modes
tric.

in these

13-616).

open guides, theTe

Above the
is

Below the cutoff frequency the mode can

even for a lossless

dielectric.

cutoff frequency for the higher-order

unattenuated propagation for a lossless dielecstill

propagate, but there

This implies losses due to radiation.

is

attenuation

582

ELECTROMAGNETICS

[CHAP. 13

E lines

Long coated
conducting wire

Wire diameter

3\'

0.02\

"Horn type

Horn type

with coating

launcher

0.0005 A thick

FIGURE

collector

13-63

Single coated-wire

open waveguide or "G-string. " (After Goubau.)

In the preceding paragraphs

we have discussed guiding by open flat structures of

z direction (perpendicular to the page). Guiding can also occur
along round wires of metal, of dielectric, or a combination of both. As an example,
let us consider the guiding action of a single round
conducting wire with dielectric
infinite extent in the

coating.

As shown by Goubau, f

type of waveguide.

However, to

this

arrangement forms a

relatively efficient

open

guided wave along the wire with good

efficiency requires a relatively large launching device, its function
being to excite a

mode,

closely related in

wavelengths.

Hence

form to the guided mode, over a diameter of perhaps

er

The

type of guide

this

A dielectric-coated
in Fig. 13-63.

initiate the

single-wire

is

several

practical only at very high frequencies.

waveguide with typical dimensions is illustrated

of a layer of enamel of relative permittivity

dielectric coat consists

3 having a thickness of only 0.0005/1.

The wire diameter is 0.02 L The conand along the wire guide is suggested

figuration of the electric field lines in the launcher

in the figure.

The mode on

the wire

is

TM type, but

it is

like a

plane

TEM wave to a

considerable distance from the wire.

Wires wound

form of long helices are

in the

type waveguides (see Sec. 14-14).

also effective single-conductor openHelix diameters as large as 0.4 2 can be used

successfully.

A linear wire made of a nonconducting dielectric material also forms an effective

open type of waveguide. J

TM,

type like in Fig.

The wave modes

in this case

3-62 for a dielectric slab.

may

be of the

If the dielectric wire

is

TM

or the

of sufficiently

small diameter (in terms of wavelengths), most of the energy is guided in

the air outside
However, if the wire or rod is sufficiently large in diameter, most of the

the wire.

energy
e,

may be conveyed

inside the rod.

2.5) diameters of the order of XX or

t G. Goubau, Surface
Phys., 21: 1119-1128
%

For rods of moderate permittivity (say

more are required to confine most of the

Waves and Their Application

(November 1950).

to Transmission Lines, /.

Appl

D. Hondros and P. Debye, Elektromagnetische Wellen an dielektrischen Drahten,

Ann. Phys., 32: 465-476 (1910); R. M. Whitmer, Fields in Non-metallic
Guides'
Proc. IRE, 36: 1105-1109 (September 1948).

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-25]

AND RESONATORS

583

feasible as
energy to the inside of the rod. Hollow dielectric wires or tubes are also
action.
guiding
effective
for
required
be
waveguides, but larger diameters may
For efficient transmission of energy by a guiding system the attenuation should

With two-conductor transmission lines this requires that the series resistance
R and the shunt conductance G be small. The conductor separation must also be
small compared to the wavelength in order that radiation losses be negligible. Under

be small.

these conditions the fields vary as \/r

and the power density

line (see Fig.

where

r is the distance perpendicular to the line

Thus, most of the power flow is close to the

Waves carried in a single hollow conducting waveguide will be
varies as 1/r

the guide walls are perfectly conducting and the material filling the
Perfectly conducting walls also prevent any radiation from the
lossless.

unattenuated
guide

3-25*).

is

if

With open guides losses due to radiation tend to become significant, and
modes which confine the power flow close to the guiding surface are desirable for
High attenuation of the fields perpendicular to guides is also
efficient transmission.
guide.

important to reduce coupling or cross talk between adjacent transmission systems.

13-25

CAVITY RESONATORS

The purpose of transmission

energy

and waveguides

energy storage device.

As such

frequencies a parallel capacitor

To make

circuit.

it is

and inductor, as

this

is

to transmit electromagnetic

on the other hand, is an

equivalent to a resonant circuit element. At low

from one point to another.

efficiently

resonant

lines

resonator,

in Fig. 13-64a, are

combination resonate

at shorter

inductance and capacitance can be reduced, as in Fig. 13-64*.

the inductance

still

further, as in Fig. 13-64c.

The

used to provide a
wavelengths the

Parallel straps reduce

limiting case

is

the completely

enclosed rectangular box or cavity resonator shown in Fig. \3-64d. In this cavity
resonator the maximum voltage is developed between points 1 and 2 at the center of
the top and bottom plates.

Simple

LC

X
X

C)
(a)

Quasi-cavity

circuits

Enclosed cavity

Id)

(c)

FIGURE

13-64
Evolution of (enclosed) cavity resonator from simple

LC circuit.

584

ELECTROMAGNETICS

[CHAP. 13

*<s>
Open
transmission

FIGURE

Coaxial

line

line

resonator

resonator

13-65

Resonator consisting of (a) 2-conductor

open transmission line and (b) enclosed

0J

coaxial line.

(a)

(6)

Resonators can also be constructed using sections of open- or short-circuited

lines, as in Fig. 13-65.
The type at (a) uses a two-conductor transmission

transmission
line,

while that at (b) uses a coaxial

(open) type

is

resonators of Fig.

of Fig. 13-64^ the

The

line.

The disadvantage of the two-conductor

that there can be a small but significant loss due to radiation.

fields

must be

in higher-order

basic principle of a cavity resonator

pure standing wave of Sec. 10-13.

entirely electric to entirely

modes.

was described

Here the energy

magnetic twice per

rectangular cavity resonator in

more

In the

TEM mode, whereas in the cavity resonator

3-65 the fields are in the

Let us

cycle.

in

oscillates

connection with the

back and forth from

now consider the case of a

and determine the resonant frequency and

Q. It is convenient to begin by recalling the situation for a TEm0 -mode wave in a
hollow rectangular waveguide. Referring to Fig. 13-66, let a TE -mode wave
m0
traveling in the -x direction be incident on a conducting plate across the guide at
x = 0, producing a pure standing wave in the guide. This standing wave is the
resultant of two traveling waves of equal amplitude traveling in the negative x direction
(incident wave) and in the positive x direction (reflected wave).
The fields of these
traveling waves are given by
detail

Ey = jJ^H

sin

kt zeKa, * x

Hx = H cos k z e "
H = Jk H sin k z e

>

x)

J(o

(2)

(1)

/(a "

x)

(3)

where k z

-x

mnlz 1
The plus sign in the exponential refers to a wave traveling in the
and the minus sign to one traveling in the +x direction. Adding the
of the two traveling waves to obtain the fields of the standing wave, we obtain
.

direction

fields

Ey

-JPZy:

2pZ

y:

sin

sin

kz z

jpx

(e

kz z sin

fix

c"'V

e1 "

(4)

(5)

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-25]

AND RESONATORS

585

Wave guide

Reflected traveling wave

(a)

FIGURE

13-66
view of rectangular waveguide closed with plate at x = and (b)
the
cross-sectional top view with additional plate at x = x t trapping wave inside
(a) Perspective

cavity.

Inserting another conducting plate across the guide at

ln/xi.

Noting that the transverse-wave impedance

E.

= -^ H

sin

kx x

Proceeding in like manner for the magnetic

Hx = -2H
Hz =
With conducting

'"

eJ

(6)

components, we get

k x x cos k z z

*+</*>]

(7)

h: HoCOS kx x sin jt, ze * + Wi

waveguide

at

x=

We

(8)

and x = x the wave is trapped

and magnetic fields are in times

note that the electric

for E), as
in exponent for
z but not
x and

in the rectangular cavity.

(rt/2

field

p = kx =
we get

requires that

plates across the

phase quadrature

sin

kz z

sin

x = xx

Zyz = (o/i/P = con/k x

is

typical of a

standing wave.

The mode of a TE wave in a rectangular cavity would in general be designated

direction, m in the z
as a TE, m mode, where / refers to variations of the fields in the x
above discussion,
=
the
in
Since we assumed n
direction, and n in the y direction.
Now k = kz =
the designation appropriate to our example would be TE, m0
.

+ tV

but v

= -P 2
k

so

(a

= 0)

= -k

and P

to

= kx

= -k

Thus,
2

(2nfy

(m

(9)

JVIXiY + (m/ztf

586

ELECTROMAGNETICS

[CHAP. 13

Path of
equivalent

TEM wave

FIGURE
(a)

13-67

The resonant wavelength of a square

cavity

is

equal to the diagonal distance.

The path of the equivalent

also shown.

(6) Electric

TEM wave is
and magnetic

configuration in square cavity resonator with TE, 10 mode, (c) Variation

of Ey and
z field components across

field

centerline of cavity.

where X

is

the resonant wavelength.

TE 110 mode

in a square-box resonator

(jtj

zt )

is

is

for a

given by

-===\A\ Xl

Thus, the resonant wavelength

in Fig. 13-67a.

For example, the resonant wavelength

(m)

(10)

equal to the diagonal of the square box, as suggested

The resonant frequency

given by

/= c/A = (3

x 10 8 )/A Hz.

The
and magnetic field configuration in the resonator is as indicated in Fig. 13-676
and c. There is no variation in the y direction (n 0). The wave inside the cavity is
equivalent to a TEM-mode wave reflected at 45 angles, as in Fig. 13-67a, assuming
that the box is infinitely long in the j> direction (perpendicular to the page in Fig. 1 3-67a).
electric

is

TRANSMISSION LINES, WAVEGUIDES,

sec. 13-25]

AND RESONATORS

587

++++

(6)

(d)

Charges

Currents

FIGURE

13-68

Electric and magnetic energy along centerline of cavity resonator.

all

(/

In principle the

the energy

= 774) all

is

the energy

y dimension

is

is

in

noncritical for the

mode

being considered (

but in practice too large a value of y could permit modes with

direction (n

At one
Fig.

bottom surfaces of the

is all

in

(t

field variations in

later

= 0),
the y

0) the energy is all in electric form, as suggested in

and negative charges on the top and

cavity, as in Fig. 13-686.

One-quarter cycle later

magnetic form (Fig. 13-68c) with

sidewalls, as in Fig.

To

0).

instant of time

13-68a, with accumulation of positive

energy

At time

and (b), while one-quarter period

the magnetic form (c) and (d).

in the electric form, (a)

find the

(f

electric currents flowing

J/4) the

down

the

3-68<aT.

of the cavity resonator

we note

that

by

definition

total energy stored

Q = 2n decrease in energy in
The energy

situation as a function of time

is

cycle

To get the total

we ( = ^e Ey 2 ) over the interior

presented in Fig. 13-69.

energy stored we can integrate the electric energy density

volume of the cavity when Ey is a maximum, and to get the decrease in energy we can
power into the walls of the cavity and multiply this by the period
=
is
I//)It
assumed
that the medium filling the cavity is air or vacuum.
Thus, we
^X
integrate the average

have

Q-

2nW
T( -dW/dt)

2n JJJ

we

dv

TtReZe jf\H,\ 2 ds

(ID

588

ELECTROMAGNETICS

[CHAP. 13

, Total
'

energy

Electric

energy W
'Magnetic energy

Wm

Periods

FIGURE

13-69

Decrease of stored energy with time in a resonator. The energy oscillates from
all electric energy at one instant to all magnetic energy period later.

where H,

is

the magnetic field component tangent to the cavity wall.

Noting

(6)

and

performing the integration for the total energy stored gives

W = W. = ie
=

2COtioHo
(
\

kz

/2(D(i
/2cofi

ie x y y 1 zA

Y C JC
J

\
1

f'sin

**

sin

k.zdx dy dz

Jo

\(lz l

= HoHo^iyiZi

[\mxj

(J)

(12)

The factor dW/dt in (1 1) is the power lost in the walls. It is convenient to calculate
this as the power lost in three pairs of faces: two perpendicular to the x direction, two
perpendicular to the y direction (top and bottom), and two perpendicular to the z
Thus,

direction.

Px =

2 x i

Py =

Re Zc f"

x l z Re Z,
1

Ch

2
z

dy dz

= 2H

y lZl Re

Zc (^-\

m-\

(14)

Pz = 2H 2 x iyi ReZc
where

Zc

is

(13)

(15)

the intrinsic impedance of the conducting material forming the walls of

the cavity.
Substituting these results in (11),

with a TE, m0

we have

for the

of a rectangular cavity

mode

e= R7z:

Wi

2
(IzJmxJ +
1

2(hilmx 1 ) y i z 1

[(Izjmxi)

+ ljx^ +

2x l y 1

(16)

sec. 13-25]

TRANSMISSION LINES, WAVEGUIDES,

AND RESONATORS

589

Waveguide^
Cavity resonator

Cavity resonator

r~u

Hole^ri

TL

Coaxial lines (o)

Magnetic field

line

(6)

Electron

/
i

beam

Electron

beam

E
"I

(c)

FIGURE

(d)

13-70

Cross sections through rectangular cavity resonators showing coupling to (a)

coaxial lines, (b) to a hollow rectangular waveguide, and (c) and (rf) to electron

beams.

For a square cavity

(Xj

Zj)

and a

TE no mode

2/l 7l/
X1.V1Z1
Q = Re Z
+
20^
XtZj + x^i)
c

2a*o

rt

Re Zc
or

fi

mm

(17)

of cavity

volume of cavity
(18)

5 interior surface area of cavity

= 2 Re Zc /con = 1/e depth of penetration.

= z, = 100 mm and y = 50 mm, we have that
and Q = 17,500 (dimensionless).

where 5
If *!

volume of cavity
interior surface area

x lO" 2 /^//
the resonant wavelength k 141
For copper 3

6.6

Methods for coupling are illustrated in Fig. 13-70 for a square cavity with TE
U0
mode. Couplings to coaxial lines are shown in Fig. 1 3-70a. These involve an electric
probe at the center of the cavity (E = max) and a current loop at the wall (H = max).
Coupling to a hollow rectangular waveguide can be via a hole in the cavity wall, as
in Fig. 13-706.

Coupling to an electron beam can be accomplished with holes in the

590

ELECTROMAGNETICS

[CHAP. 13

Here the electrons move

To reduce the transit time for the

top and bottom surfaces of the cavity, as in Fig. 13-70c.

parallel to the electric field

may

electrons the cavity

where

it is

maximum.

be modified as in Fig. 13-70d.

PROBLEMS
*

13-1

A transmission line of 50

ance of 100 +/100 H.

the

VSWR

(c)

Find the
13-2

on the

Check

line.

reflection

characteristic

Q.

(a) Calculate the

and the one

this calculation

and transmission

impedance

is

terminated in an imped-

impedance at a point |A from the load,

in (a) using

(b)

Calculate

a Smith chart,

coefficients for voltage.

Show that a transmission line having no attenuation must also have a nonreactive

characteristic impedance.

13-3

L=

fM m~

uniform transmission

and

',

C=l nFm"

1
.

R=

has constants

line

At 3.18 kHz

10

mQ m _1 G =
,

find (a) the characteristic

fxO

m _1

impedance of

wave propagation on the line; (c) the percentage to which

wave decreases in 10 km:
13-4 Find the high-frequency characteristic impedance Z of an air-filled coaxial
transmission line having a radius ratio b\a = 3, where b is the inside radius of the outer con-

the line; (b) the phase velocity of

the voltage of a traveling

ductor and a

13-5

the outside radius of the inner conductor.

is

Find the characteristic impedance

(a)

(6)

for a two-conductor transmission line

mm

and conductor diameter of 6 mm.

What is the upper frequency limit at which this line should be used if losses due to radiation

in air with center-to-center conductor spacing of 50

are to be minimized?

13-6

strip transmission line consists

of a thin conducting strip 5

mm wide mounted

on a 1-mm-thick dielectric substrate of the same width placed against a large flat conducting
ground plane. If e, 2.7 for the dielectric material, (a) what is the characteristic impedance
of the line? (b) What is the ratio of the wave velocity to that of light?
13-7

A coaxial transmission line consists of a cylindrical inner conductor of diameter

d and a symmetrically
3d.
(e r

situated outer conductor of square cross section with side dimension

Find the characteristic impedance

= 1.13).

(c)

What

is

the line

if

the ratio of the

wave

is

(a) air-filled

and

(b) filled

with polyfoam

velocity to that of light for the case in (6)?

Z = R a = 150 CI is connected
impedance Zi to a load of 250 +y"100 l.
Find d and Zi which match the load to the 150-O line.
13-9 A double-stub tuner, shown in Fig. PI 3-9, is constructed of 200-Q line (Z =
*
R = 200 fl). Find the shortest lengths rfi and d2 which match the load ZL = 400 +/200 O
*

13-8

transmission line of characteristic impedance

through a section of length

d and

characteristic

to the 200-O line.

13-10

150-O transmission

Find di,d%,Ri, and

R2

line

is

connected to two loads as shown in Fig. P13-10.

such that the loads receive equal powers but with voltages of opposite

phase.

* Answers to starred problems are given

in

Appendix C.

PROBLEMS

TRANSMISSION LINES, WAVEGUIDES,

AND RESONATORS

591

FIGURE PI 3-9

FIGURE P13-10

Double-stub tuner.

Line with two loads.

13-1 1 The arrangement for a three-port lossless circulator

is suggested in Fig. PI 3-1 1
All three segments of the circulator have equal lengths
/ and characteristic resistance R
Port 1 is connected to a transmission line of characteristic resistance
R , and ports 2 and 3
are terminated in
If a wave of unit voltage is incident at port 1, find the
value of the
.

reflected voltage r at port

r

when

(a) I

= 0?

= A/2 and (6) / = A/4,

(c) Is there

a length

for which

13-12 Show that the input impedance to a lossless exponential

line of length y connected as in Fig. PI 3-1 2 to a load L is given by

where

Zo (0)

is

the characteristic impedance of the line at the load.

Rn

Z(y)-

FIGUREP13-11

FIGURE PI 3-1

Three-port circulator.

Exponential

line.

592

ELECTROMAGNETICS

[CHAP. 13

= 50 Q

Coaxial line(fl

150
-A/W-*-

Ro

Load

3000
150

vw*-2X

FIGURE PI 3-1
Coaxial double-stub tuner.

FIGURE

PI 3-1
Line with series resistances.

Referring to Fig. PI 3-1 3, find (a) the power in the load and (b) the

13-13

and minimum voltages on the 2A section for two cases:

and

(2)

(both sections)

The

13-14

reflection coefficient for voltage

(both sections)

maximum

= 300

CI,

on a

line

of characteristic impedance

is

= Z-Zo _ (Z/Zo)-l _ Z-1

Z + Zo (Z/Z )+l Z. +
1

Z = load impedance, CI
= normalized impedance = R +jX

Z
It

150 Q.

where

(1)

dimensionless

follows that

Pv

are obtained as functions of the real and imaginary parts of p\ a map

plane yields the Smith impedance chart (see Sec. 13-7).
and ratio of the wave velocity to the
13-15 Find the characteristic impedance
(a)
Prob. 7-1, (6) Prob. 7-2, (c) Prob. 7-3,
line
of
velocity of light v/v for the transmission

Show

that

of R and

if

R and

in the

JT

/5

take e
{d) Prob. 7-24, (e) Prob. 7-25, (/) Prob. 7-26, (g) Prob. 7-30 (four cases),
3e , and 0') Prob. 7-33.
(h) Prob. 7-31, (/) Prob. 7-32, take e

= 3<r

Find the characteristic resistances for a binomial three-section A/4 transformer for matching a 500-O load to a 50-Q line. (6) Plot the impedance presented to the
50-C1 line (normalized to 50 Ci) on a Smith chart over the frequency range 0.5/o <f<, 1.5/o
where / is the center frequency, (c) Plot the VSWR on the 50-O line as
of
13-16

in steps

(a)

0.1/

ordinate vs. the above frequency range as abscissa, (rf) For comparison also plot the
is used.
over the same frequency range for the case where only a single A/4 section

13-17
transmission
find the

two

(VSWR =

(a)
line.

sets

double-stub tuner (Fig. P13-17)

If the

load impedance

(b)

Find the

constructed of
CI,

air-filled 50-ii

= 200 mm, and h =

coaxial

50

mm,

d2 which will match the load to the line at 300 MHz

VSWR at A at 20-MHz intervals over the bandwidth 200

of stub lengths, di and

at A),

is

ZL = 40 + y"80

VSWR

PROBLEMS

to

400

TRANSMISSION LINES, WAVEGUIDES,

MHz for both sets of stub lengths.

Draw a graph

respect to frequency.

(linear scale) as abscissa for

Assume

of the

that the load

VSWR

AND RESONATORS

impedance

is

593

constant with

as ordinate (log scale) vs. frequency

VSWR
VSWR curves of part (6).
Which combination of stub lengths results in the wider bandwidth for VSWR < 2 ? One of
the curves has a VSWR = oo point close to the VSWR = 1 point at 300 MHz.
Might this
values at 11 frequencies,

both

(c)

of stub lengths.

sets

Each curve should have

Discuss the significance of the

have a practical application?

13-18

+;450 n.
the

main

a distance di from the load. The stub length

to match the load to the line.

line at

values of d%

A two-conductor 300-H transmission line is connected to a load of 450

A short-circuited single-stub tuner of the same type of line is connected across
is

d2

Find the shortest

and d2

A plane 1-GHz wave is incident normally on the plane surface of a half-space

= and e = 3.5. Find (a) the thickness in millimeters and

13-19

of material having constants

p.,

(b) the relative permittivity required for

a matching plate which

will eliminate reflection

of

the incident wave.

A quartz window (e = 5)

13-20

a vacuum chamber into the

at least 3

air.

mm

is used to pass a linearly polarized laser beam from

For adequate mechanical strength the window must be

thick.
Design an arrangement that will allow transmission of the laser beam
through the window without reflection. A A/4 coating can be applied to the air
side of the
window but not to the vacuum side. The wavelength is 700 nm. Hint
Utilize the Brew:

ster angle.

Prove that the power through each curvilinear square in Fig. 1 3-25 is the same.
Overhead long-distance high-power transmission lines radiate electromagnetic

13-21

13-22
interference,

generate poisonous ozone, and otherwise disrupt the ecology.

Design a

transmission line for handling terawatt

power levels over megameter distances which avoids

Assume that a 25-K superconductor is available. Hint: Consider a

the above effects.

buried thermally insulated liquid-hydrogen-cooled dc line operating at levels

of a few volts.
Alternatively consider a liquid-hydrogen-cooled cylindrical waveguide operating
at 100

GHz

in

a higher order mode or modes (see Sec.

13-23

4-7).

In Prob. 13-22 superconductors are considered for low-loss power transmission.

is that superconductive cables can provide faster response
(wider band-

Another advantage

width) for transmission of information.

with nanosecond response.

Design a superconductive two-conductor cable

H. J. Jensen, Nuclear Test Instrumentation

See, for example,

with Miniature Superconductive Cables, IEEE Spectrum, 5:91-99 (September

1968).
13-24 Show that when the series resistance R and the shunt conductance G of
mission line are small but not negligible, the attenuation constant

,R lc
2a/ L
'

and the phase constant as

An

13-25

a>

a trans-

written as

G JL
2/J C

VLC.

hollow rectangular waveguide has cross-sectional dimensions

80 mm. (a) Find the cutoff frequencies for the following modes:

air-filled

=40 mm and zi =
TEM, TE 10 TE 20 TE

yi

/9

may be

i,

TE02, TEn,

TE 2l and TE l2
,

(b)

Find the ratio of the wave

594

[CHAP. 13

ELECTROMAGNETICS

velocity in the guide to the velocity in free space for each of the

the waveguide

order

modes

at

yi

1 .5

An

13-26

= 40 mm and

fc

if

/=

1 .5

fc

(c) If

hollow rectangular waveguide has cross-sectional dimensions

air-filled

= 50 mm.

zi

TE

transmit of the

modes

of copper find the attenuation in decibels per meter for each of the higher

is

and

(a)

frequencies below 8

At

TM type?

(6)

Find the

GHz what

ratio of the

wave

modes

will this

guide

velocity in the guide to

modes if /= 1.7/c
The wavelength region around 10 mm is commonly referred to as K band,
30 mm as X band, and 100 mm as S band. Design a rectangular waveguide for transmitting
TEio mode K-band waves with minimum attentuation and with minimum change in velocity
over the wavelength range 9 to 1 1 mm. The guide should be incapable of passing any modes
other than the TEio (such as TE i) in this wavelength range and any waves at all at wave-

the velocity in free space for each of the

13-27

lengths greater than

An

13-28

Your result should be in the form of a drawing with dimensions.

mm.

air-filled cylindrical

waveguide has a diameter d

TM

= 50 mm.

(a)

Find the

TMu, TMu, TEoi, TE 02

modes: TMoi,
02
(6) Find the ratio of the wave velocity in the guide to the velocity in free
TE,i, and TE12
(c) If the waveguide is of copper find the attenuaspace for each of the modes if /= 1 .3 fc
cutoff frequencies for the following

tion in decibels per meter for each of the

An

13-29

below 5

GHz

ratio of the

if/=1.2/c

modes

at 1.3

waveguide has a diameter d = 75

guide transmit of the TE and

air-filled cylindrical

what modes

wave

will this

mm.

TM

At

type?

frequencies

(b)

Find the

velocity in the guide to the velocity in free space for each of the

modes

13-30

of thickness

A perfectly conducting flat sheet of large extent has a dielectric coating (e = 3)

d = 5 mm. Find the cutoff frequency for the TMo (dominant) mode and its
r

attenuation per unit distance.

13-31 Find the reflection angle 6 of the TEio rectangular-waveguide mode in Prob.
13-25 at a frequency (a) 1.1 times the cutoff frequency and (b) 2.5 times the cutoff frequency.
communications service using mobile units wants communication to be
13-32 (a)

maintained even when their radio-equipped automobiles and trucks are in a vehicular tunnel.
which can
If the smallest tunnel diameter encountered is 5 m, what is the lowest frequency

be employed?

(b)

No

frequencies ?

13-33

What

could be installed in the tunnel to permit communication at lower

frequency converters are permitted.

A metal-clad building has window openings

1 .5

by

0.5

m.

What

is

the longest

wavelength which can readily penetrate into the interior of the building?
plane 3-GHz wave in air is traveling parallel to the boundary of a conducting
13-34
*

The constants for the conducting medium are

parallel to the boundary.
and /x r = e, = 1. If the traveling-wave rms electric field E = 75 mV m~ \
the average power per unit area lost in the conducting medium.
13-35 Design a waveguide for handling gigawatt power levels with minimum attenu-

medium with
a

= 10 7 U m~

find

ation over a

20-km distance

modes which

are of

13-36

much

Identify the type

wave components

in Fig.

GHz. Consider using a cylindrical guide in a mode or

number than the dominant mode.
of device and explain the function for as many of the micro-

at

higher order

P13-36 as possible.

PROBLEMS

TRANSMISSION LINES, WAVEGUIDES,

AND RESONATORS

595

-f

WM

1
3
2

596

[CHAP. 13

ELECTROMAGNETICS

13-37

of the power

where d

is

Show

that for a dielectric-coated conductor, as in Fig.

Pt

transmitted in the dielectric

power transmitted

to the

13-616, the ratio

in the air

P.

is

given

the thickness of the dielectric coating.

modes for radio

Discuss the possibilities of using dielectric-slab surface-wave
that the moon
Note
more).
km
or
distances
(1,000
over
long
moon
communication on the
Fernald, Post-occultation
and
D.
L.
Salisbury
W.
W.
example,
for
See,
ionosphere.
has no
(Nov. 12, 1971);
Reception of Lunar Ship Endeavour Radio Transmission, Nature, 234:95
Earth of HF
the
on
Propagation
and
also A. F. Wickersham, Jr., Generation, Detection
13-38

Radio Surface Waves, Nature, 230:125-130 (Apr. 5, (1971).

wave in a hollow rectangular
13-39 Show that the field components for a
by
given
are
13-35)
(see
Fig.
conductor waveguide

and

VHF

TM

Ex = E

rvny

sin

rrmz

Zx

mnz
rnry
sin
= yE rm cos
zi

E,

k 2yx

my

Zx

^ nw

gJm _ yx

zi

yx

TM mode condition

Hx =

- y r ,Ec mn
k2

z.

yx

zi

yY^Eo

njr

mry

k2

yx

yx

that the field

mnz gJM _

nny

ny

,,_ ,JC
e"'

yx

= -yEom,
sjn
2

Show

,_
1X

e Ja"

sin

)>1

13-40

mnz gJlat _ yx
zx

components for a

TM wave in a hollow cylindrical wave-

guide (see Fig. 13-44) are given by

dJ^kr~)

yEo

Et

^ al _^

= _ ^S. Sin 4>J (kr)e>"-

n

k r

=E
H.=0
E,

single-

cos

n<

try"-"

TM mode condition

= nyEoYr* ^ ^
k 2r

)<;]m

PROBLEMS

TRANSMISSION LINES, WAVEGUIDES,

Show

13-41

equal to

TE wave

impedance of a

in

597

a rectangular guide

is

Z = Z/A/Ao), where Z* is the intrinsic impedance of the medium, A the wavelength

yI

and A

in the guide,
fills

that the transverse

AND RESONATORS

the wavelength in an

unbounded medium of the same material

that

the guide.

Show

13-42

that the transverse impedance of a

TM

wave

in

a rectangular guide

is

equal to

*J*-'-ffi-*]>-
Show

13-43
in

an

that the attenuation constant for a

infinite-plane transmission line or guide

= 2ReZ
5"
~J~a

where Re

(A,/2rf)

/
rfReZ, Vl

TEio wave

at frequencies

above cutoff

is

-a

(Np

/2d) 2

m-

')

Z = real part of intrinsic impedance of wall medium (conductor), O

= real part of intrinsic impedance of medium filling guide (dielectric),
d = wall spacing, m
A = wavelength in unbounded medium, m
c

Re Zt

Show that the attenuation constant

13-44
in

an

for a

TE wave at frequencies above cutoff

infinite-parallel-plane transmission line of spacing

ReZ

is

is

- (A/A 0C) 2

the cutoff wavelength.

13-45

medium.

2(A /A OC ) 2

rfReZ, Vl
where A oc

CI

A TEM wave is traveling in air parallel to the plane boundary of a conducting

that if K = p v, where K is the sheet-current density in amperes per meter,

Show

charge density in coulombs per square meter, and v the velocity of the wave
in meters per second, it follows that
= H, where is the magnitude of the field of the
p, is the surface

wave.
13-46

Show that the average power transmitted by a perfectly conducting

mode is given by

rectangular

guide operating in the TEio

r av
13-47

Show

that the average

guide operating in the

TEi mode
i

is

4tt 2

power transmitted by a
given by

perfectly conducting cylindrical

wtf\Ho\*r,*
82
13-48

Write the

with y and z for a

13-49
r

and

<j>

for a

field

Write the

field

TEi 2 wave

equations and sketch the variation of each

a square conducting guide (y x zi).

equations and sketch the variation of each

TEi 2 wave

field

component

in

in a cylindrical conducting guide of radius r

field
.

component with

598

ELECTROMAGNETICS

13-50

Write the

[CHAP. 13

field

equations and sketch the variation of each

with y for a TEi wave in a dielectric slab-on-conductor guide, where y

The

conductor.

dielectric-slab thickness

find the attenuation constant

are

d=A

In an infinite-parallel-plane

13-51

made of copper.
13-52 What is

for a

and

/2,

air-filled

field

component

distance from the

is

relative permittivity

is e,

transmission line of 15

mm

its

TEM wave and a TEio

wave

at 12

GHz.

= 5.
spacing

The planes

the attenuation constant a in decibels per meter for a hollow single-

conductor waveguide at an applied frequency 0.85 times the lowest cutoff frequency for the
guide?
In an infinite-parallel-plane transmission line show that at a wavelength A

13-53
less

than cutoff, the attenuation constant for a

TMio wave

is

2a

Vl-(A/2i) 2

= attenuation constant for TEM wave

parallel-plane spacing
*
13-54 A 500-MHz TEM traveling wave in a strip transmission line has an electric
The line consists of a copper strip 10 mm wide separated mm
field E = 500 mV m ~ rms.
= 2.7). (a) Find
from a flat copper ground plane by a 1-mm-thick ribbon of polystyrene
where a
b

(<r r

the average power in the direction of transmission,

guide walls per meter of length,

(d)

What

What

the characteristic impedance?

is

13-55

V between

(c)

is

(e)

What

In a hollow rectangular guide with TE

the top

(ft)

Find the average power into the

the attenuation constant in decibels per meter?

is

the

wave

velocity?

mode show that the ratio of the voltage

and bottom of the guide (at the middle) to the longitudinal current / on
is an impedance given by

the upper or lower inside surface

= height of guide
= width of guide
Z = transverse impedance

where y t

zi

13-56

Show that the

attenuation constant for a

in a hollow rectangular waveguide of height yi

Re Z.[(Aq/A 0C) 2 +
z, Re Z, Vl - (A

a
13-57

Show

-Vo

-Vo

A oc

at frequencies

above cutoff

zi is

z l jly x \
/A) 2

that the group velocity u in a hollow rectangular single-conductor wave-

guide (equal to the velocity of energy transport)

where A

TE wave

and width

= wavelength in unbounded medium

= cutoff wavelength

is

given by

TRANSMISSION LINES, WAVEGUIDES,

PROBLEMS

Both A and X cc should be distinguished from

follows that uv

= Do 2

where v

A, the

AND RESONATORS

Note

wavelength in the guide.

the phase velocity in the guide and v

is

is

599

that

it

the phase velocity

an unbounded medium.

in

Compare

13-58

the

values attainable with a parallel inductor-capacitor circuit,

a short-circuited section of two-conductor

a cavity resonator, a quartz crystal, an
transition of

ground

An

13-59

= zi = 80 mm,

square, xi

13-60

diameter

and
its

is

both ends by

at

silver-plated inside,

dominant mode
(a)

at

molecule, and a cesium

atom

(a)

line,

(hyperfine

133).

resonator operates in the

= 35 mm.

The

TEno

cavity

is

mode. The cavity

made of copper and

(a) resonant frequency, (b) resonant wavelength,

cavity resonator

d closed

13-61

cesium

with height yi

Find

gold-plated inside.

state of

air-filled cavity

a short-circuited section of coaxial

line,

ammonia

and

is
is

Q.

(c)

constructed of a short section of cylindrical tubing of

is

flat plates

What

is

separated by d/2.

the required value of

band (10 mm)?

(b)

What

is

the

The cavity is made of brass

d if the cavity is to operate in

Q1

Radio-frequency dielectric heaters are widely used in

molding of plastics. The heating

Another application of this heating

many

industrial

processes, particularly for injection

results

electric hysteresis (see Sec. 8-15).

effect is for

from the

di-

cooking

using microwave ovens.

power source.

The oven consists of a cavity resonator connected to a microwave

Design such a microwave oven operating at 2,450 MHz with 750
radio-

frequency power and an oven volume of 200 by 400 by 400 mm. (a) What maximum rate
of temperature rise will be possible for 3 kg of dielectric material with complex permittivity
r

= 4 +yi

heat

if its specific

is

J g-

C~ l ?

(6) If the

metal screen, give the screen specifications, material, wire

radiation loss through the screen to less than 10

important that the screen wires be bonded at

all

oven has a window covered by a

size,

mW cm -2

and hole size required to keep

(See Prob 10-39.) Also is it

contact points?

(c)

Describe the standing-

oven and what means (such as fan or paddles) are used to " stir "
the standing-wave pattern, {d) How will the edges of the oven door be sealed to prevent

wave mode or modes

in the

radio- frequency leakage?

13-62
cations,

1-MV
line,

and

Select a suitable type of transmission line or

line lengths as follows

buried power

line,

(a)

length 10 km, (c) 60-Hz low-voltage flexible power

50 m, (d) 500-Hz to 5-kHz bandwidth audio

50

km

(e)

100-kHz two-conductor 600-O

distance with 100 such circuits placed in

2-MHz buried 50-fl

25-MHz 50-Q line from
(/)

waveguide for frequencies, appli-

dc low-voltage measurement, length 5 m,

line

line

circuit,

(b)

ordinary telephone application over

compact bundle with minimum cross

from 100-kW transmitter

from 50-kW transmitter

dc

(~10-kW)

to antenna, distance

to antenna, distance 500

m, (g)

talk,

km,
5- to

shortwave antenna to receiver, distance 500 m, (h) 54- to 890-MHz

from receiving antenna to TV set through low electromagnetic noise, distance 50 m,
(*') same as (A) but through high noise, distance 200 m, 0') 1 .42-GHz line or guide with (3-MHz
bandwidth between radio-telescope antenna and receiver, distance 10 m, lowest possible loss,
(k) 3-GHz (S band) line or guide with 10-MHz bandwidth from 100-kW radar transmitter
to antenna, distance 10 m, (/) 30-GHz (K band) line or guide with 10-GHz bandwidth between
300-ti line

and receiver, distance 10 m, (m) 100-GHz line or guide with 10-GHz

bandwidth for data transmission, between amplifiers, distance 5 km. Some of the types to
radio-telescope antenna

choose from are as follows:

two-conductor parallel or twisted

pair,

two conductors in

shield,

600

ELECTROMAGNETICS

[CHAP. 13

coaxial line (dielectric-filled), coaxial line (air-filled), strip line, waveguide (rectangular or
circular) in

dominant or higher-order modes.

material in

all cases.

13-63

and the

line

reflected

V
V= V +

from the load has voltage

total voltage

and

selection, including attenuation, cost, etc.

Let the load be at x =

wave has voltage V, and the wave
Use a computer to calculate and plot the magnitude
V, at 100-mm intervals to a distance of 5 m from the

A uniform transmission line is

extend to positive values of x.

and phase of the

Specify conductor or guide dimensions

Also explain reasons for

connected to a load.

An

incident

load (50 points) for

= V e"eHm+ x
V = -$V e-'*e "-">
V = 10 V
/= 300 Mhz (A = m)
j = 0.3Npm-'
'>

V,

J(

V as ordinate vs. x as abscissa for one graph and phase

180 through 0 to +180) as ordinate vs. x as abscissa for second graph.
Use same scale for x in both graphs and display both graphs together with magnitude graph

Plot results with magnitude of

angle (from

below the phase graph.

13-64
(6)

(a)

Make

a power-flow contour

map like

that in Fig.

3-256 using a computer.

Repeat using a conductor spacing twice that in Fig. 13-256 but with the same conductor

diameter and the same total power.

13-65

wave to

Demonstrate the equivalence of (12-4-21) and (12-4-34) for a normally incident

(13-6-11) for a transmission line.

13-66

Hold your hand between your eye and a lamp or

finger close to but not quite touching

when you look at the finger-thumb

parallel boundaries.

your thumb.

gap.

bright area with your fore-

Explain the phenomenon you observe

Hints: Consider higher order

Consider diffraction.

modes between two

14
ANTENNAS AND RADIATION

14-1

INTRODUCTION

In discussing transmission lines and waveguides in Chap. 13 our attention was focused

on guiding energy along the system.

Little consideration

the loss of energy from the system into free space.

guides are usually

made

was given to

While transmission

radiation,
lines

i.e.,

or wave-

so as to minimize radiation, antennas are designed to radiate

(or receive) energy as effectively as possible.

shown in Fig. 14-la, connected to a radioAlong the uniform part of the line, energy is
guided as a plane TEM-mode wave with little loss. The spacing between wires is
assumed to be a small fraction of a wavelength. At the right the transmission line
As the separation approaches the order of a wavelength or more,
is opened out.
the wave tends to be radiated so that the opened-out line acts like an antenna which
launches a free-space wave. To be more explicit, the region of transition between a
guided wave and a free-space wave may be defined as the antenna. We have described
two-wire transmission line

is

frequency generator (or transmitter).

the antenna as a transmitting device.

around, and the antenna

wave and a guided wave.

is

As a

receiving device the definition

is

turned

defined as the region of transition between a free-space

602

ELECTROMAGNETICS

[CHAP. 14

lines

Transmission

line

Generator
transmitter
\

Guided (TEM)wave
One dimensional wave
Free space wave

Transition region

radiating

or antenna

in

three dimensions

Transmission

line

(6)

antenna

Guided (TEM) wave

Space

\
Transmission

(e)

Dipole

line

(d)

Horn antenna
Guided

(TE-|

mode) wave

Ground
' e'

plane

Waveguide

I
FIGURE

14-1

The antenna
wave or vice

as a region of transition between a guided

wave and a

free-space

versa.

maximum separation of 1A
maximum power beamed to

If the wires separate gradually, as in Fig. 14- la, to a

or more, the arrangement will be an

the right, as suggested.

antenna of Fig. 14-16.

we

Modifying

efficient radiator

this

with

antenna to one with straight wires yields the

Folding the wires perpendicular to the transmission

arrive at the simple dipole antenna

of Fig.

14-lc.

The

transition

line,

from a guided

wave to a space wave is more abrupt with the dipole antenna (Fig. 14-lc) than for the
antenna of Fig. 14- la, and there is more reactive or stored energy in the vicinity of the
dipole antenna.

ANTENNAS AND RADIATION

SEC. 14-1]

603

Main lobe axis

Main lobe or main beam

beam
(HPBW)

Half power

width

Beam width between

first nulls

(BWFN)

Minor lobes

(o)

Main lobe

OdB

-3d3

-lOdB

(6)

FIGURE
(a)

14-2

Antenna power pattern

in polar coordinates

and

linear scale; (b) antenna

pattern in rectangular coordinates and decibel (logarithmic scale).

Both patterns

are the same.

An
also

we

an

opened-out section of hollow rectangular waveguide or horn (Fig.

effective radiator or antenna.

arrive at the

more abrupt

opens into the center of a

flat

Bending the horn

transition

shown

sides

in Fig. 14-le,

back into a

14-1*/) is

flat

plane,

where the waveguide

conducting plane.f

The response of an antenna as a function of direction is given by the antenna

pattern.
By reciprocity (see Sec. 14-18) this pattern is the same for both transmitting
and receiving conditions.
t

For a more complete discussion of antennas and their basic properties see, for examD. Kraus, "Antennas," McGraw-Hill Book Company, New York, 1950.

ple, J.

604

ELECTROMAGNETICS

[CHAP. 14

The pattern commonly consists of a number of lobes, as suggested in Fig. 14-2a.

The lobe with the largest maximum is called the main lobe, while the smaller lobes are
referred to as the minor lobes or side and back lobes.
If the pattern

no change

in pattern

is

measured

from the antenna for there to be

Measurements
which are a function of both angle and

sufficiently far

with distance, the pattern

at lesser distances yield near-field patterns,

The pattern may be expressed

distance.

in

is

the far-field pattern.

terms of the

field intensity (field pattern)

or in terms of the Poynting vector or radiation intensity (power patterns).

14-2a

To show

a power pattern in polar coordinates.

is

Figure

the minor-lobe structure in

more

detail the pattern can be plotted on a logarithmic or decibel scale (decibels

below main-lobe maximum). Figure 14-26 is an example of a pattern on a decibel

scale in rectangular coordinates.

The pattern

in Fig. 14-26

is

the

same as the one

in

Fig. 14-2a.

single pattern, as in Fig. 14-2,

would be

sufficient to

variation of radiation with angle provided the pattern

mean,

in the case of Fig.

revolution of the one

However,

completely specify the

shown around the pattern

in practice

pendicular to the main-lobe axis

If the pattern

axis.

map

is

required to

is

show

is

linearly polarized in

patterns

may

the pattern in

H plane (containing H).

As an example,

might be linearly polarized

in

E plane

in the

This assumes that the antenna

If this

is

not the case, more

the dominant radiation from an antenna

one principal plane, but the radiation from some minor

lobes might be linearly polarized in the principal plane at right angles.

antenna might be

its

two patterns perpendicular to each other and permay suffice. These mutually perpendicular patterns

one of the principal planes.

be required.

a figure of

is

not symmetrical,

through the main-lobe axis are called the principal-plane patterns, one
(containing E) and one in the

This would

symmetrical.

14-2a, that the three-dimensional pattern

a three-dimensional diagram or a contour

entirety.

is

Or

the

elliptically polarized.

In this chapter the most important parameters of an antenna are considered.

These include the gain (or

directivity), the terminal

radiation resistance), the pattern,

and antenna temperature.

beam

width,

beam

impedance

(real part equals the

area, effective aperture, bandwidth,

variety of antenna types are also discussed.

include the dipole, loop, helix, reflector antennas, arrays, and the interferometer.
short dipole

14-2

is

important theoretically and

is

considered

first (in

These

The

Sec. 14-3).

RETARDED POTENTIALS

In dealing with antennas or radiating systems the propagation time

great importance.

Thus,

if

an alternating current

Fig. 14-3, the effect of the current

is

not

felt

is

is

a matter of

flowing in the short element in

instantaneously at the point

but only

ANTENNAS AND RADIATION

sec. 14-2]

FIGURE

605

14-3

Short current-carrying element.

after

an

interval equal to the time required for the disturbance to propagate over the

distance

r.

Accordingly, instead of writing the current / as

I

I sin

cot

which implies instantaneous propagation of the

(1)

effect

of the current, we can introduce

the time of propagation (or retardation timet), a s done by Lorentz,

[/]

where

is

[I]

Equation

(2) is a

The time

(r/c).

It is

r,

is

(2)

The brackets

[]

may

be added, as here, to

retarded.

caused by a current

difference r/c

where

is

H)

co

and

at a

earlier

time

statement of the fact that the disturbance at a time

distance r from the element

t

/ sin

called the retarded current.

indicate explicitly that the current

the distance

and write

is

is

we

that occurred at

an

the interval required for the disturbance to travel

_1

the velocity of light (300

to be noted that

[/]

Mm

).

dealt with retarded quantities in

Chap. 10

in

connection

with wave propagation, although the term retarded was not used.

For example,

Chap. 10 a solution of the wave equation

similar in

/?(

given that involves sin(co/

form to the trigonometric function

sin col

where

is

cojc

InjX)

is

in

which

is

in (2) since!

sin (cot

fir)

(3)

the phase constant.

t Called retardation time because the phase of the wave at P

the phase of the current in the element by an angle airlc.
t

fix),

The expression
The

fix)

Chap. 10

is

retarded with respect to

refers to a plane wave traveling in the x

or sin(tof
fir) refers to a spherical wave
An important point of difference between a plane
traveling in the radial direction.
and a spherical wave is that a plane wave suffers no attenuation (in a lossless medium)
but a spherical wave does because it expands over a larger and larger region as it

direction.

propagates.

sin (utt

in

relation sin u>[t

(r/c)]

606

ELECTROMAGNETICS

[CHAP. 14

In complex form (2) ist

[/]

In the

more general

=l

e Jml

'- Wc

situation

e''

""-

where the current

(A)

(4)

distributed

is

we may

write for

the retarded-current density

[J]

=J

eJ

' c

'- (r/c)]

=J

Kat -

(A

m~ 2 )

(5)

Introducing this value of current density in (5-25-10) for the vector potential,

we

obtain a retarded vector potential that

is

applicable in time- varying situations where

the distances involved are significant in terms of the wavelength.

vector potential

M^f^^f ^-""
4n

Jv

4nJ v

Likewise the scalar potential

V can

Jv

= retarded scalar potential, V

= Po e o[ -(r/c)1 = retarded charge
'"

(Wbm-)

,,

(6)

'

dv

(7)

Cm -3

density,

THE SHORT DIPOLE ANTENNA

14-3

A short linear conductor is often called a short dipole.

short dipole
is

the retarded

be put in the retarded form

4ne

[p]

is,

M = jMwhere [V]

That

is

is

always of finite length even though

vanishingly short,

Any

it is

linear antenna

an

is

In the following discussion, a

may

be very short.

If the dipole

infinitesimal dipole.

may

be regarded as being composed of a large number of

short dipoles connected in series.

dipole

it

Thus, a knowledge of the properties of the short

useful in determining the properties of longer dipoles or conductors of

complex shape such as are commonly used

Let us consider a short dipole like that shown in Fig. \4-4a.

very short compared with the wavelength

provide capacitive loading.

The

more

in practice.

(/ -4 A).

The

length

/ is

Plates at the ends of the dipole

short length and the presence of these plates result

in a uniform current / along the entire length

energized by a balanced transmission

line, as

of the dipole.

shown.

It is

The

dipole

may

be

assumed that the transmis-

is understood that the instantaneous value of current is given either by the imaginary (Im) or the real (Re) part of the exponential expression in (4).

t It

ANTENNAS AND, RADIATION

sec. 14-3]

End

607

plate

_r"

FIGURE
ductor

Transmission

14-4

(a) Short-dipole

+Q

line

antenna fed by 2-con-

transmission

line

and

(b)

-Q

its

(a)

equivalent.

(b)

sion line does not radiate, and its presence will therefore be disregarded.
Radiation
from the end plates is also considered to be negligible. The diameter d of the dipole
is small compared with its length (d <^ /).
Thus, for purposes of analysis we may

consider that the short dipole appears as in Fig. 14-4*.

a thin conductor of length

Here

it

consists simply of

with a uniform current / and point charges q at the ends.

According to (4-13-6), the current and charge are related by
/

dq
'

(1)

dt

Let us

now proceed

dipole of length

At any point

as in Fig. 14-5.

Ee

E$, and

air or

to find the fields everywhere

as shown.

P the

It is

14-5

Relation of dipole antenna to coordinates.

its

dipole.

Let the

center at the origin,

electric field has, in general, three components,

assumed that the medium surrounding the dipole is

vacuum.

FIGURE

around a short

be placed coincident with the z axis and with

608

ELECTROMAGNETICS

From

[CHAP. 14

(8-17-11) the electric field intensity

E=

at

any point

(Vm-

expressed

is

by

(2)

dt

where

V=

electric scalar potential at

A = vector potential
From

point P,

at point P,

(5-25-2) the magnetic field

Wb m _1

H at any point P

H= Vx

is

(Am -1 )

(3)

Ho

= permeability of air (400n nH m _1 )

A = vector potential at point P, Wb m _1
If the scalar potential V and the vector potential A at the
electric and magnetic fields E and H at P can be determined
where n

Since

we

point

are

known, the

by means of (2) and

(3).

are interested in the fields not only at points near the dipole but also at

and

distances that are comparable to

u = Vx
1

(AnT

LAj

use the

Thus we have

-MA]

-\[V]
dt

and

we must

larger than the wavelength,

retarded potentials given in (14-2-6) and (14-2-7).

(Vm'

(4)

(5)

Ho

^ = T~ J-

wheref

[A]

^f

J ^

47t J

The

dv

(5a)

(Wbm -1 )

dv
r

(Sb)

and magnetic

fields due to any configuration of currents and charges

where
the retarded scalar potential t^] is a quantity that
(5),
depends only on the charges (stationary) and the retarded vector potential [A] is a
quantity that depends only on the currents. Equation (5) indicates that the magnetic

electric

are given by (4) and

field

depends only on the currents, while

depends on both the currents and the charges.

determining the radiation

However,

field (at large distances

need be considered.

tion) only the currents

(4) indicates that

it

will

the electric field

be shown later that in

from a current and charge

distribu-

Since the retarded potentials will be used

exclusively in the following development, the brackets will be omitted for the sake

of simplicity,

it

being understood that the potentials are retarded.

More
tion

concisely

e- J > r/4nr.

A = no

Jo

e-*

"

dv,

where

G = free

space

Green's

func-

ANTENNAS AND RADIATION

sec. 14-3]

FIGURE

14-6

Geometry

for short dipole.

609

-HW-d

We shall now proceed to find the electric and magnetic fields everywhere from a
short dipole by

first

determining the retarded vector and scalar potentials and then

substituting these values in (4)

and

(5)

and performing the indicated operations.

Referring to Fig. 14-5 or 14-6, the current

it

is

entirely in the z direction.

follows that the retarded vector potential has only a z component.

A.=

lAI

^
An

where 7

1/2

e K-M

J -i/2

Its

Hence,

value

is

dz

(6)

amplitude (peak value in time) of current (same at

all

points along

dipole),

nH m _1
=
dz
element of length of conductor, m
co = radian frequency (=2nf, where /= frequency,
t = time, s
s = distance from dz to point P (see Fig. 14-6), m
_1
(=27t/l)
P = phase constant, rad m

H = permeability of free space =400 n

Hz)

from the dipole is large compared with its length (r > /), and if the
compared with the length (A > /), we can put s = r and neglect the

If the distance

wavelength

is

large

magnitude and phase differences of the contributions from the

wire.f Thus (6) becomes

different parts

of the

(7)

4w
t If r is large compared with
in the denominator in (6)

but A

is

not large compared with

and neglect the

/,

we may put

differences in magnitude.

=r

However,

we should retain s in the exponential expression since the difference in

phase of the contributions may be significant.

in such cases

610

ELECTROMAGNETICS

FIGURE

[CHAP. 14

Dipole

14-7

Relations for short dipole

when r>/.

The electric charge is confined

we

to the retarded scalar potential,

to the ends of the dipole ; so turning our attention

find its value to be

(V)

4ne

where q
jt
s2

From

(8)

= amplitude (peak value in time) of charge at ends

= distance from upper end of dipole to P, m
= distance from lower end of dipole to P, m

of dipole,

(1)

q=
=q
1= I

where q
It

s2

st

\Idt

"~ l s) =
retarded charge,
"~ Ps) =

e Ku

'

J(a

retarded current,

follows that q

/j(o

so that

V=

(8)

/o

(9)

can be reexpressed
/

e H-i

(10)

Ane

j(o
fa) \

Si

When r > the lines of length s and s2 from the ends of the dipole to the point
P may be considered parallel, as shown in Fig. 14-7, so that ij = r (1/2) cos and
s 2 = r + (//2) cos 0.
Substituting these into (10) and clearing fractions yields
/,

V=

I eiK"-**

4ne

jco

[r

+ (f/2) cos

0]exp(; j 01 cos 0)

[r

- (//2) cos 0]exp(-j j fil cos 6)

(ID


ANTENNAS AND RADIATION

SEC. 14-3]

where the term

with r

since r

(I

>

2
cos 0)/4 in the denominator has been neglected in comparison

By de Moivre's theorem,

/.

/e/(--Wr/
y=

"l
Since

it is

BlcosQ

COS

becomes
I

Bl cos

(1 1)

Bl cos 9\/
+;smi
r +-cos0

-^

cos!

jS/cosfl

dipole (A

611

0\/

smiL

-J )( r -2

assumed that the wavelength

is

much

\1

(12)

jj

greater than the length of the

!> /),

=
Bl cos 9
-

cos

nl cos 9

cos

2-

and

sin

(13)

A.

picosO

-^

as

Bl cos6
'^

(14)

Introducing (13) and (14) into (12) reduces the expression for the scalar potential
to
/

/e'

'"-' r)

cos8

/l

4ne c

y?)

(V)

= amplitude (peak value in time) of current, A

= length of dipole, m
to = radian frequency ( = 2nf, where / = frequency, Hz)
B = phase constant, rad m" ( = 2n/X)
= time, s
9 = angle between dipole and radius vector of length r to

(15)t

where /

point P, dimension-

less

= permittivity of free space =8.85 pF m _1

-1
c = velocity of light =300 Mm s
j = complex operator = ^J \
r = distance from center of dipole to point P,

Equation (15) gives the retarded scalar potential and (7) the retarded vector
potential at a distance r and at an angle 6 from a short dipole.
The only restrictions
are that r $> I and A > /. Before substituting these values in (4) and (5), let us express

in polar coordinates.

Thus, in polar coordinates (see Fig.

E = tEr + &Ee +

Now in

14-5),

fyEj,

(16)

A = tA + 6A e + fat

(17)

polar coordinates
r

Note

that l/e c

ft

= V^

/e

= 376.7 CI.

612

ELECTROMAGNETICS

[CHAP. 14

FIGURE

14-8

Ar

Resolution of vector potential into

and

In our case

Ae

components.

A^ =

has only a z component, so that

and from

0,

Fig. 14-8

= A z cos 6
A e = A z sin 6
A,

In polar coordinates

we

(17ft)

also have for the gradient of the scalar potential

*8V

\dV

dr
It

(17fl)

dV

39

(18)
r sin

d<j>

follows from (4) and the above relations that the components of

Er =

-ja>A r

EB8

jcoA
J
e

dV
or

IdV
r

E* = -jcoA*
introducing the value of

find that

E+

(since

V is

E.=

and

(7)

ic /(

"-" r)

le

Ho"- fr) sin

dO

V from

(20)

dV
6

(21)
d<j>

(15) into these equations,

so that dV/dcj)

cos 8 (

0)

and

we

also that

\cr 2

2ne

8V

and

(19)

r sin

d<p

<j>,

or

are

dV

sin

dV

A z from

86

independent of

Er =

= jcoA,

r sin

Now

-jcoA z cos

(22)

jcor /

/jw

(23)

\c r

47t

Turning our attention

now

polar coordinates the curl of

f

to the magnetic field,

AJ _ d(A

SO

d4>

r a(sin
L

J cor J

we can

calculate

it

by

In

(5).

is

*
r sin

cr

T3A,

rsin0

[#

8(r sin

dr

t) i

0)^1
\

j>

\ 8{rA e )

r [

dr

_ 8AJ
dd\

'

ANTENNAS AND RADIATION

sec. 14-3]

Since

A+ =

0, the first

note that A, and

we

and fourth terms of

Ae

component.

only a

operations,

and

Introducing

{Ma), and {lib)

third terms of
and
so that the second

(24) are zero.

are independent of

Thus, since only the

(24) are also zero.

<j>,

two terms

last

A r and A e

substituting this result into (5),

IHI

tf

le

From

613

iim, ' llr)

(7),

V xA

in (24) contribute,

has

into (24), performing the indicated

we

find that

and that

(jco

sin

(t

An

(25)

?)

Thus the electric and magnetic fields from the dipole have only three components
Er E, and H^,. The components Et Hr and He are everywhere zero.
When r is very large, the terms involving 1/r 2 and 1/r 3 in (22), (23), and (25) can
,

be neglected in comparison with terms in 1/r. Thus, in the far field Er is

and we have effectively only two field components, Ee and H$ given by

negligible,

le^

ja>I

n
"*
of

Ee

le>

(26)
r

iM - fr) sin

to

h'$

as given

(ll

Anr

by (26) and

*L = -L =

is

.l
J

4ncr

H4
This

SQZ./H^.,^

c r

4rte

ja>I

sin

l?o

^m

(27),

sin

we obtain

376.7 fi

(27)

fo

(28)

the intrinsic impedance of free space.

Ee and H$ are in time phase in the far field. Thus, E and

of the spherical wave from the dipole are related in the same manner as
Both are also proportional to sin 0. That is, both are a
in a plane traveling wave.
maximimum when = 90 and a minimum when 9 = (in the direction of the dipole
This variation of Ee (or H^) with angle can be portrayed by a field pattern
axis).
It is

to be noted that

in the far field

p of the radius vector being proportional to the value of the

from the dipole. The pattern in Fig. 14-9a is onehalf of a three-dimensional pattern and illustrates that the fields are a function of 9
The pattern in Fig. 14-96 is two-dimensional and represents
but are independent of

as in Fig. 14-9, the length

far field

(Eg or

H^,) in that direction

<f>.

a cross section through the three-dimensional pattern. The three-dimensional farfield pattern of the short dipole is doughnut-shaped, while the two-dimensional pattern

has the shape of a figure of eight.

614

ELECTROMAGNETICS

[CHAP. 14

FIGURE

14-9

(b)

Three-dimensional and (b) twodimensional field pattern of far field

(a)

(Ee or H$) from a short

From

(22), (23),

two components,
magnetic

field

H^

and
and

(25)

Ee

we note

Dipole

dipole.

that for a small value of r the electric field has

both of which are

Thus, in the near field,

E and

in time-phase

quadrature with the

H are related as in a standing wave.

Ee

and

other so that the total electric

field

vector rotates in a plane parallel to the direction of

At intermediate

distances,

can approach time-phase quadrature with each

propagation and containing the dipole, exhibiting the phenomenon of cross

field (see

Sec. 11-6).

In the far field the energy flow

outward.

This energy

equator (6

90).

is

is real.

In the near

That

is,

the energy flow

As a function of angle

radiated.
field

the energy flow

is

it

is

is

always radially

maximum

largely reactive.

energy flows out and back twice per cycle without being radiated.

There

is

at the

That

is,

also angu-

energy flow (in the 9 direction). This energy picture is suggested by Fig. 14-10,
where the arrows represent the direction of energy flow at successive instants.f
Let us now consider the situation at very low frequencies. This will be referred

lar

Noting that 1 =j<oq

to as the quasi-stationary case.

we can

express the field

com-

ponents as

le

'- Pr)
(M '"
Km
cos

2ne

The instantaneous

direction

and time

the instantaneous Poynting vector

/jco

1\

Vr 5

r )

(29)

rate of energy flow per unit area is given

(=E X

H).

by

ANTENNAS AND RADIATION

SEC. 14-3]

615

Polar axis

9=0

Near

field

Far
field

Equator
Reactive i
Keacnve
energy flow/

FIGURE

le

">smO

in near

(or dc) casef the field

-** flow

far regions of short dipole.

J>

c r

(30)

cr

_.^_^lg^
/e

the frequency approaches zero

Also e

energy

+ +^)
r]

sine /jm

\ cr

47i

can be neglected.

and

(J <"-'"- )

As

fl=90
__

..

iim, - l

4ne

Ht

'
.
p
Radiated

14-10

Energy flow

^^
^~^-

\>

Short
dipole

*-

K<a, ~ fr)

(to

- 0), the terms with a> in the

approaches unity.

numerator

Thus, for the quasi-stationary

components become

Er = q-x

lcosO

2rt

(32)

sin 6

E* = q-7Z1T
3
1

4ne

(33)

7 /sin

** i^rThe

electric field

components, (32) and

(34)

same as (2-15-4) and

component H$ in (34) is

(33), are the

(2-15-5) for a static electric dipole, while the magnetic field

Since these fields vary as 1/r 2 or

equivalent to (5-3-3) for a current element.

are effectively confined to the vicinity of the dipole and radiation

frequencies in the far

H^) vary as
fields

field,

These

1/r.

is

negligible.

1/r

3
,

they

At high
(Ee and

however, we note from (26) and (27) that the fields

radiated and hence are often called the radiation

fields are

of the dipole.
t

For
r^>

this case the

/

and hence

wavelength

in this case

is

very large (A-> oo) so that

A> r.

A> > >

/.

We also have

616

ELECTROMAGNETICS

[CHAP. 14

The expressions
marized in Table

for the fields

from a short

dipole, developed above, are

> / and
E^ = Hr =

In the table the restriction applies that r

14-1.

>

/.

sum-

The

components not listed are everywhere zero, that is,

= 0.
g
If we had been interested only in the far field, the development following
(7)
could have been much simplified. The scalar potential V does not contribute to the
far field, so that both E and H can be determined from A alone.
Thus, from (4),
three field

Ea

arid

H+

H* = H
l

of the far

field

= Ee/Z =

can be obtained very simply from

-(jcolZ )A e

obtained directly from

(5)

and

where

Z =J

Ee from this.

376.7

fi.

-jcoA e and

Or H+ may be

Thus
1

"*=|H| =

je

Ea = E =

IVxAI

(35)

Mo
and, neglecting terms in

1/r

2
,

=|E| =ZU, = -|VxA|

(36)

fl

Mo

Ee field

Referring to (26) for the far

the expression into

its

six basic factors.

of a short dipole,

60ti

jeJOt-M

~~

7
A

it is

instructive to separate

Thus,

sin

(37)

fl

Magni- Current Length Distance

Phase

Pattern

tude

where

60;r is

a constant (magnitude) factor, I

is the dipole current, l\A is the dipole

"~
the distance factor, je Kv
is the phase
the pattern factor giving the variation of the field with angle.
In

length in terms of wavelengths,

factor,

and

sin 6

is

1/r

any antenna

general, the expression for the field of

Table 14-1

is

will involve these six factors.

FIELDS OF A SHORT DIPOLE

Component

General expression

/ofe

Jto '~'" )

27re

E.

/o

fe

Jto, -'" )

47re

COS0

/ 1

\cr 2

sin0 Ijw
\c 2 r

//-') sing Ija,

ff*
4tt

Far

lcr

Quasi-stationary

field

jo

your 3 /

cr 2

/cur 3 /

1\

+ rV

'

cos 8

27rer 3

j60TrI e ]( "-i" , sin8

-<"'>sin0
/7o e J

2r

1 sin

47rr

sin 6

4m- 2

ANTENNAS AND RADIATION

sec. 14-4]

The

antennas or large antennas of other shape

such short dipoles.

Longer

Table 14-1 are those for a short dipole.

field relations in

Hence the

fields

integrating the field contributions

may

be regarded as

617

linear

being made up of many

of these larger antennas can be obtained by

from

all

the small dipoles making

up the antenna

(see Sec. 14-9).

RADIATION RESISTANCE OF A SHORT DIPOLE

14-4

By taking the

surface integral of the average Poynting vector over any surface enclosing

an antenna the

total

power radiated by the antenna

7>=J*S av
where

P = power

radiated,

is

obtained.

Thus

(W)

-<fc

0)

W
W

m~ 2
Sav = average Poynting vector,
The simplest surface we might choose is a

sphere with the antenna at the center.

Since the far-field equations for an antenna are simpler than the near-field relations,
it

will

be to our advantage to

dimensions of the antenna.

make

In this

the radius of the sphere large

compared with the

way

lies in

the surface of the sphere

the far

field,

and only the far-field components need be considered.

Assuming no losses, the power radiated by the antenna is equal to the average
power delivered to the antenna terminals. This is equal to iI 2 R, where I is the
amplitude (peak value in time) of the current at the terminals and j? is the radiation
Thus P = iI 2 R, and the radiation resistance is
resistance appearing at the terminals.

K=
where

is

the radiated

Let us

now

power

P=
field

(2)

(11)

in watts.

carry through the calculation, as outlined above, in order to find the

radiation resistance of a short dipole.

In the far

IP

only

Ee

and

The power

radiated

JSav -c/s = -jRe(ExH*)

H^

is

ds

(3)

are not zero, so that (3) reduces to

P=

-J

Re

Ee H*i-

(4)

618

ELECTROMAGNETICS

where

is

[CHAP. 14

the unit vector in the radial direction.

P=
where

Ee

H$ Z;

so

and H% are complex,

(5) becomes

P=Since

Re

Z = -J

ii

Re

J"

and dy

je

H%

where the angles

and

<j>

are as

But

d6

in the far

ds

ds; so

H^

Now Ee =

Re Z ds

(6)

d</>,-\

f'fWr*

Jo

(5)

being the complex conjugate of

sin

RcEe H*ds

sin

OdOd*

(7)

'o

shown

From

field.

-f

H+Hl Z ds = J H,
1

P=
\
I J?*
V e

amplitude) of the

Thus the power flow

(normal to surface of sphere of integration).

field is entirely radial

in Fig. 14-5

(14-3-27) this
col

and

H^\

is

the absolute value (or

is

sin

l"*'~ 4r
Substituting this into

(7),

(8)

we have

(9)

Upon

integration (9)

becomes

p=

few?

This

is

the

(10)

Yht

power radiated by the short dipole.

power P from (10) into (2)

Substituting the

yields for the radiation resistance

of

the short dipole

Since

Jn

le

376.7

\20n Q, (11) reduces to

Jt

t Since

Vfi

/t

=
ai

20(/?/)

\20ttX

80k 2 (j\

E9IH^, we may

(SI)

also write

(12)


ANTENNAS AND RADIATION

SEC. 14-4]

EXAMPLE

Find the radiation resistance of a dipole antenna

solution

From

619

A/ 10 long.

(12)

R = 80tt 2 ( tV) 2 = 7.9 n

If there

is

any heat

loss in the

antenna due to

conductivity of the dipole or

finite

to losses in the associated dielectric structure, an equivalent loss resistance i? loss will

appear such that the terminal resistance

is

given by

X = -Rioss + K
where R ]oss =
Rr =

()

(13)

loss resistance, fl

radiation resistance, fl

Suppose, for instance, that

terminal resistance

is

loss

then 8.9

fl

fl

The

for the A/10 dipole of the above example.

(= 1.0 +

The antenna

7.9).

Power radiated
=
k = a
Power input

efficiency

is

R.

Rr <+ R

3a )
los ,

- = 89 percent.

8.9

A
i? )oss

longer dipole with larger radiation resistance would be

more

efficient

provided

remained small.

The

radiation resistance of antennas other than the short dipole can be calculated

as above provided the far field

is

known

as a function of angle.

(6) the radiation resistance at the terminals of an antenna

is

Thus, from

(2)

and

given by

(14)

where

H\

amplitude of far

If we integrate the

an antenna, we

H field, A m

'

complex Poynting vector

both a

= ^E x H*) over a surface enclosing

power radiated
and an imaginary part equal to the reactive power. Whereas the real part, or radiated
power, is the same for any surface enclosing the antenna, the imaginary, or reactive,
power obtained depends on the size and shape of the surface enclosing the antenna.
shall obtain, in general,

real part equal to the

For a large surface lying only

in the far field the reactive

surface lying in the near field

may be of considerable

very thin linear antenna,

it

it

the square of the terminal current yields

zero, but for a

is

turns out that if the surface of integration

to coincide with the surface of the antenna, the complex

the radiation resistance.

power

magnitude.

In the case of a
is

collapsed so as

power so obtained divided by

the terminal impedance R +jX, where R is

620

ELECTROMAGNETICS

[CHAP. 14

DIRECTIVITY, GAIN,

14-5

A useful

way of describing

the pattern of an antenna

of the main lobe at a particular

power

(BWFN)

imum

or

in terms of the angular

is

beam

is

The beam

usually the one given.

widths 10 or 20

dB below the pattern max-

are also sometimes used.

The antenna power pattern as a function of angle can be expressed

component of the average Poynting vector multiplied by the square of
r at

width

Referring to Fig. 14- 2a, the angle at the half-

level.

beam width (HPBW)t

level or, half-power

width between first nulls

AND EFFECTIVE APERTURE

which

(11-3-12)

it is

and

as the radial

the distance

Thus, regardless of the wave polarization we have from

measured.

(11-3-13) that

P(^) = S

i^|^r

i// (0,0)Zr

(Wsr"

(1)

= radial component of Poynting vector (time-average value), W m~ 2

_1
E{9,4>) = total transverse electric field as a function of angle, V m
-1
H(6,(j)) = total transverse magnetic field as a function of angle, A m
r = distance from antenna to point of measurement, m
Z = intrinsic impedance of medium, 1 per square

where Sr

It is

assumed that the medium

is

The quantity

independent of distance.

(W sr -1

or

W rad"

2
).

In the far

lossless.

entirely radial (fields entirely transverse),

and

P(6,4>)

Dividing P{0,4>) by

its

and

the Poynting vector

field

H vary as

1/r,

so that P(0,0)

is

is

often called the radiation intensity

is

maximum

value LP(0,$) max ],

we obtain

the normalized antenna power pattern,

P* > =

Im^

normalized antenna power pattern

coinciding with the direction

antenna pattern

is

is

0.

(dimensionless)

shown

pattern solid angle or

Thus, integrating

beam

in Fig. 14-11 with pattern

highly significant

to integrate the normalized

obtain a solid angle.

(2)

way of

maximum

describing an

power pattern with respect to angle and

Pn {Q,<j>)

over

47t sr

(rad

2
)

we

get the total

solid angle

.^=2rc ,0 = rc

nA =

f
J

4=

P(0, 4>)sin

0d6d<t>=

J 8=0

t Also called the 3-dB beamwidth.

f l>(0,4>) dO.
JJ

(sr)

(3)

ANTENNAS AND RADIATION

sec. 14-5]

=0

sin

621

S d

dQ = sinfd0d<t>

<t>

FIGURE

14-11

direction (zenith).

explicit, the

term beam solid

Antenna power pattern with maximum aligned with the

Although the term

angle

is

total pattern solid angle

customarily used for

l A

would be

The beam solid angle ClA is the angle through which all the power from a transmitwould stream if the power (per unit solid angle) were constant over this

ting antenna

maximum value. It turns out that for typical patterns this

approximately equal to the half-power beam width (HPBW), as suggested in

angle and equal to the

angle

is

Fig. 14-12.

Beam solid
angle

FIGURE

14-12

Antenna power pattern and

to the

beam

solid angle.

its

relation

QA

Half power

beam width
(HPBW)

622

ELECTROMAGNETICS

[CHAP. 14

main lobe

Integrating over the

"m =

gives the main-lobe solid angle\

dSl
if PJP,4>)

(sr)

(4)

main
lobe

It

Qm

follows that the minor-lobe solid angle

(total pattern) solid angle

is

and the main-beam

0 =

given by the difference of the

aA - ilM

all

(5)

= 0), then ilA = QM

Pn (9,<l>) = for all 6 and

directions)

For an isotropic
and aA = 47t.t
D, which may be defined
.

Another important antenna parameter is the

maximum

beam

is,

(sr)

no minor lobes (Qm

If the antenna has

antenna (radiation same in

as the ratio of the

That

solid angle.

directivity

<t>,

radiation intensity (antenna transmitting) to the average

radiation intensity, or

D=

maximum

radiation intensity

The average radiation intensity is given by

The total power radiated is equal
4rcsr.
grated over 4n sr. Hence,

f(e,<a,ax

D=

W/4n

,<

= P(9,<f)) mi x
Pav

average radiation intensity

the total

power radiated

We

(6)

W divided

by

to the radiation intensity P(9,<f>) inte-

4rc-P(0,4>) max

dtl
JJ P(d,<p)
4*
A.-7T

4rc

4jc

JJ
4*

(dimensionless)

[W4>)im

4>)m.x]

note that the denominator of the last

dn

JJ pjji,4>)
4*

member of

(7) is

(7)

da

the

beam

solid angle as

given by (3) ; so

D=
Thus, the

directivity

4jt

of an antenna

divided by the antenna

from the pattern.

The

beam

(dimensionless)

is

equal to the solid angle of a sphere (4n

solid angle

directivity

is

(8)

aA

In this relation, directivity

a unique dimensionless quantity.

is

sr)

derived

It indicates

no clearly defined minimum exists the extent of the main lobe may
be somewhat indefinite, and an arbitrary level such as 20 dB can be used to delineate it.

t In patterns for which

The beam

efficiency

e may be defined as

M = Cl M /n A

SEC. 14-5]

how

ANTENNAS AND RADIATION

well the antenna concentrates

power into a

623

limited solid angle; the smaller the

angle the larger the directivity.

EXAMPLE

Calculate the directivity of an isotropic antenna (uniform radiation intensity

in all directions).

For an

solution

isotropic antenna P,(0,4)

Thus, the directivity of an isotropic antenna

can attain.

EXAMPLE

and noting

The

Thus,

(14-3-27).

(3)

unity.

and

This

C1 A

is

= 4n;

so

the smallest value the directivity

Calculate the directivity of a short dipole.

solution

by

is

=1

and

short dipole has only an

its

normalized power pattern

component of the magnetic

JT/o"

477

sin3(?

=_3

(u)

Hence, the directivity of a short dipole

much

given

(8),

n = ^-_
D

1.5 times as

field as

is

as

if

the

is f.
That is, the maximum radiation
power were radiated uniformly in all directions.

intensity is

Directivity is based entirely on the shape of the far(or radiated-) field pattern.
The antenna efficiency is not involved. However, the power gain, or simply
the gain, of
an antenna does involve the efficiency. It is defined as follows:

Gain

=G=

maximum radiation intensity

maximum radiation intensity from a

refer-

'

'

ence antenna with the same power input

Any

convenient type of antenna

antenna

is

lossless isotropic

gain (designated

) is

may be taken

as the reference.

type (radiation intensity uniform in

If the reference
all

directions), the

given by

0=

< 13 >

= maximum radiation intensity from antenna under consideration

P = radiation intensity from a lossless (100 percent efficient) isotropic antenna

where P'm

with same power input

624

ELECTROMAGNETICS

[CHAP. 14

Direction of

D=

(e

propagation of
plane wave

Receiving

antenna

FIGURE

(a)

14-13

Terminated receiving antenna immersed in field of plane traveling wave

and (b) equivalent circuit.
(a)

Since P'm

is

(b)

related to the radiation intensity

the antenna efficiency factor k, same as

G =

Pm

of a 100 percent

efficient

antenna by

in (14-4-13a),

= kD

(14)

Thus, the gain of an antenna over a lossless isotropic type equals the directivity if the
antenna is 100 percent efficient (k = 1) but is less than the directivity if any losses are
present in the antenna (k

The

gain

<

1).

(over an isotropic source)

isotropic) to distinguish

dBi (decibels over

antenna, as discussed in Sec. 14-20.

antenna

when expressed
it

The gain

can be written

in decibels

from the gain over a reference A/2

(in dBi) equals the directivity

if

the

is lossless.

The
infinity (1

directivity

<D<

oo).

is

On

never

less

than unity.

Its

value must

the other hand, the gain (G or

G ) may

lie

between

lie

between

and

and

infinity.

transmitting antenna radiates energy.

hand, collects energy.

In this connection

it

is

receiving antenna,

on the other

often useful to consider that the

receiving antenna possesses an aperture or equivalent area

over which

it

extracts

energy from a passing radio wave.

Thus, suppose that a receiving antenna is immersed in the field of a plane travelThe antenna is terminated in a load of impedLet the aperture A of the antenna be defined as the ratio of the
ance ZL = R L + jXL
The
received power to the power density (or Poynting vector) of the incident wave.
ing wave, as suggested in Fig. 14-13a.
.

received

power

is

equal to I
1

RL

R,

where /

is

the terminal current.

received

Therefore

power
(15)

power density of incident wave

SFc 14-S1'

ANTENNAS AND RADIATION

A = aperture, m2
/ = rms terminal current, A
S = Poynting vector (or power density)
R L = load resistance, fi

625

where

Replacing the antenna by

emf
shown in

equivalent
circuit

its

and impedance

Fig. 14-136.

(16)

is

equivalent, or Thevenin, generator having an

z^ (=

^ +jx^ w

fflay

dfaw

^^

(A>

Z^ZA

C6)

substituted into (15),

A=
where

U = rms emf induced by passing wave, V

ZL = load impedance, 2
ZA = antenna impedance
When

W m~

The terminal current /is

I=
where

of incident wave,

RA = antenna

follows that

it

RL
SKRa + Rl) + (Xa +Xl) 2

resistance,

( 17>

XA = antenna reactance, Q
RL =

load resistance,

XL = load reactance, CI
The emf 17 will be greatest when the antenna is
oriented for maximum response
Under these conditions the maximum power will be
transferred to the load when
Xl= ~ xa and RL = Rr It is assumed that the antenna is

mum aperture,

known

{RA

effective aperture

EXAMPLE

soluton
the short dipole

Ae

Find the

17 2 R,L
" "

4SR

these conditions

Ae

IT 2

4SRr

as given

effective aperture of

we need

maximum when

(m 2 )

to

know

IT, S,

is parallel

Hence

= El

(18)

all antennas.

a short dipole antenna.

the dipole

17

we

by

has a unique, simply defined value for

Referring to (18),
is

= Rr). Under

as the effective aperture

A.=
The

RA is
obtain the maxi-

lossless, so that

entirely radiation resistance

(V)

and*,.

The emf induced

to the incident electric field

in

(19)

[CHAP. 14

ELECTROMAGNETICS

626

FIGURE

0~o

14-14

a.\

Radiation from aperture A, with uni-

form

field

The Poynting

Em

vector

is

s = |!

(Wm- 2 )

Zo

where Z

is

vacuum) (=

the intrinsic impedance of medium (air or

the radiation resistance

(20)

V^

/o).

From (14-4-1 1)

is

R= Vl^o^r
0

(n)

(21)

Off

and R,

Substituting these values for O, S,

into (18) gives the effective aperture of a short

dipole:

A e = i-X 2 =0.U9\ 2
Thus, regardless of

and deliver

it

to

its

how

small the dipole

is, it

can

terminal impedance or load.

(22)

power over an aperture of 0.1 19A

assumed here that the dipole is lossless.

collect

It is

conductor,
However, in practice, losses are present due to the finite conductivity of the dipole
so that the actual effective aperture

is less

(R A

Consider next that an antenna with

1 A

If the

as suggested in Fig. 14-14.

radiated

effective aperture

field

A e has the beam solid angle

Ea is constant over the aperture, the power

is

P=
where

= Rr + Ri,s)-

Z is

^A.

(W)

the intrinsic impedance of the medium.

Then the power radiated

is

(23)

Let the

field at

a radius

be

r.

given by

P=

r ClA

Equating (23) and (24) and substituting Er

(W)

(24)

= E.AjrX [see (14-7-5)] yields the important

relation
I

= wavelength, m
2
A e = effective aperture, m
ilA = beam solid angle, sr

where A

= A c lA

(m 2 )

(25)

SEC. 14-6]

ANTENNAS AND RADIATION

According to

(25), the

product of the effective aperture and the

equal to the wavelength squared.

at

If A e

is

beam

known, we can determine

627

solid angle is

QA (or vice versa)

a given wavelength.

From

(25)

and

(8)

it

follows that

(26)

Here the

is based on the aperture.

Aperture
Three expressions have now been given for the

directivity

Sec. 14-23.

_P(0,4>) max
"

Pav

is

discussed further in

directivity

D.

They are

from pattern

(27)

from pattern

(28)

from aperturef

(29)

S1A

4*

14-6

ARRAY THEORY

Much

of antenna theory (and almost all of array theory) involves

little
the proper addition of the field contributions from all
parts of an antenna.
deal with the field (not power) since we must
include

more than
We must

both magnitude and phase.

14-6.1

Two

Isotropic Point Sources

Consider two isotropic point sources separated by a distance

d, as in Fig. 14-15a.
A
is an idealization representing here
an isotropic radiator occupying zero
volume. By reciprocity (see Sec. 14-18) the pattern of
arrays of such sources (transmitting case) will be identical with the pattern when
the array is used as a receiving
antenna. Let the two point sources be identical
(in amplitude)
point source

and in the same

also that both sources have the same polarization;
i.e., let both be
linearly polarized with E perpendicular to the
page. When the reference point for
phase is taken halfway between the sources, the far field in
the direction
is given
phase.

Assume

by

E^E^* +Eie
12

- J * 12
(1)

Here A. is independent of ohmic losses. Thus, the gain

same as given by (29) for a lossless antenna but is less if

of the antenna

losses are present.

is

the

628

[CHAP. 14

ELECTROMAGNETICS

0=0

Field pattern

FIGURE

14-15

Two isotropic point sources separated

by a distance d and (b) field pattern
(a)

when sources are of equal amplitude and

(b)

in phase with A/2 separation.

where

E = far electric field at distance r due to source 1

E2 = far electric field at distance r due to source 2
= pd cos 9 = (2n d\K) cos 9
ty
x

sources as
is the phase-angle difference between the fields of the two
E2
When
14-15a).
(see
Fig.
angle
9
the
line
at
t =
radius-vector
the
measured along

The quantity ^

we have

= 2j
For a spacing of A/2 the

field

iW +

pattern

is

e j<H2

as

shown

2j cos

"A

(2)

in Fig. 14-156.

been taken at
If the reference for phase for the two sources in Fig. 14-15 had
would
pattern
field
far
resultant
sources),
the
the
source 1 (instead of midway between
be

E=E

+E

(3)

ANTENNAS AND RADIATION

SEC. 14-6]

and for

629

E E2
t

E = 2E

cos

3 * 12

^e

= 2E

cos

| L ijf/2

(4)

The field (amplitude)

pattern is the same as before, but the phase pattern is not. This
because the reference was taken at the phase center (midpoint of array) in developing
(2) but at one end of the array in developing (4).
is

14-6.2

Pattern Multiplication

We

assumed above that each point source was isotropic (completely nondirectional).
have directional patterns which are identical, the resultant pattern is given by (14-6.1-2) [or (14-6.1-4)], where E is now also a function of
y
angle [Ex = E(0)]. The pattern E(&) may be called the primary pattern and cos
0/2 the
If the individual point sources

secondary pattern or array factor.

Multiplication,

This is an example of the Principle of Pattern

which may be stated more generally as follows
The total field pattern
:

of an array of nonisotropic but similar sources

the product

of the individual source

pattern and the pattern of an array of isotropic point sources each located at the phase
center of the individual source and having the same relative amplitude and phase,
while
the total phase pattern is the sum of the phase patterns of the individual
sources and the
is

array of isotropic point sources.^

14-6.3

From

Binomial Array
(14-6.1-2) the relative far-field pattern

point sources spaced k\2 apart

is

of two identical m-phase isotropic

given by

'= cos(-cos0l

(1)

For convenience we have set Et = \ in (14-6.1-2) so that the pattern is normalized.

This pattern, as shown in Fig. 14-156, has no minor lobes. If a second identical
array of two sources is placed A/2 from the first, the arrangement shown in Fig. 14- 16a
is

obtained.

The two sources at the center should be superimposed but are shown
By the principle of pattern multiplication, the resultant pattern

separated for clarity.

is

given by
'

t It is

= cos 2 |-cos0l

assumed that the pattern of the individual source

array as

when

it is

isolated.

(2)

is

the

same when

it is

in the

630

[CHAP. 14

ELECTROMAGNETICS

FIGURE
(a)

14-16

Binomial array with source ampli-

tudes

1:2:1:

and

(6)

binomial array

The

with source amplitudes 1:3:3:1:.

spacing between sources is A/2.

as

shown

If this three-source array with amplitudes 1:2:1

in Fig. 14- 16a.

with an identical one at a spacing A/2,

the arrangement of Fig. 14-166

is

is

arrayed

obtained with

the pattern

= COS3 ^COS0)
as

shown in
and

1:3:3:1,

Fig. 14-166.
it

also has

(3)

This array has effectively four sources with amplitudes

no minor

lobes.

Continuing

this process,

obtain a pattern with arbitrarily high directivity and no minor lobes

Table 14-2

it is

if

possible to

the amplitudes

PASCAL'S TRIANGLE

1
1

13
14
1

15
1

6
10

10
15

20

15

ANTENNAS AND RADIATION

sec. 14-6]

631

To distant
point

rfcosf

$=0

FIGURE

14-17

Array of it isotropic point sources of equal amplitude and spacing.

of the sources
coefficients

in the array

correspond to the coefficients of a binomial

by

are conveniently displayed

internal integer

is

the

sum of the adjacent ones above.

.E

series, f

These

Each
The pattern of the array is then

Pascal's triangle

(Table

14-2).

= cos' - /|cos0]
,

(4)

where n

is the total number of sources.

Although the above array has no minor lobes, its directivity is less than that of
an array of the same size with equal-amplitude sources. In practice most arrays are

designed as a compromise between these extreme cases (binomial and uniform).

Array with n Sources of Equal Amplitude and Spacing

14-6.4

The binomial array

a nonuniform array.

is

For a uniform array, as in

with n isotropic sources of equal amplitude and spacing, the far

E=E

+ei+ + eJ2 * +

E=E

or

where

(l

\j/

= fid cos 6 +

d=
S

---

Fig. 14-17,

field is

e""- 1 *)

*-

(1)

(2)

spacing between sources

= progressive phase
t
t

J. S.

difference

between sources

Stone, U.S. Pats. 1,643,323

This is a polynomial of degree n

phasor f. Thus (1) becomes

and

1,715,433.
,

as can be seen

+ r+

f1

+f

by putting e1 * equal

to a unit

632

ELECTROMAGNETICS

[CHAP. 14

Multiplying (1) by e J * yields

Ee>*
Subtracting (3) from

(1),

1)^/2,

EoieJ*

e12 *

+ e^* +

e J'"*)

(3)

we have

is

is

If the center of the array

phase angle, (

ein*

-F **WM

n*

(4)

chosen as the reference for phase, instead of source

eliminated.

1,

the

If the sources are nonisotropic but similar,

represents the primary or individual source pattern, while

sin(nt/f/2)/sin(t/>/2) is

the

array factor.

For
pattern

isotropic sources

and the center of the array as reference

for phase the

is

sin(wiA/2)
(5)

sinr/2)

As

\j/

-* 0, (5) reduces to

E = nE
This

is

the

maximum

value of the

In the direction of the

Dividing

(5)

by

maximum

(6) yields

Referring to
7

i.

(4),

It is

from a

single source.

= 8

is satisfied.

(7)

sin(i/r/2)

= 2kn

mji

(but

provided

or

fid cos

field

or f}d cos 8

the null directions of the pattern occur for e Jn *

This requires that

= null angle
= 1, 2, 3,

^=

1 sin(ni/r/2)

nE

where 6
k

n times the

the condition

the normalized field pattern

"

e*

field.

(6)

cos

mn, where

(8)

'[(*-')=]
m=

1, 2, 3,

(9)

.)

ANTENNAS AND RADIATION

SEC. 14-6]

Array with n Sources of Equal Amplitude and Spacing:

Broadside Case

14-6.5

maximum

In the direction of the pattern

the condition fid cos 6

For a broadside array (maximum at 6

When 6 is replaced by its complementary angle
angles are given by

fied.

(8

ro
If the array

is

large, so that

The

L = length of array, m
L = (n \)d nd if n is

first nulls (y 01 )

(BWFN)

must be

satis-

in phase

y (see Fig. 14-17) the null

(1)

small),

is

_ k K k
"^dTx

(2)

Lfx

large

occur when k

Hence, the beam width between

1.

first

nulls

is

*-d-_

The more commonly used parameter

the half-power

is

about one-half (more nearly 0.44) the

Thus,

hpru/
HPBW

BWFN

BWFN
=
2

The

field
is

114.6

BWFN-2, 0I

A/2 apart

must be

= sin-^-j

nd P kk (and
y

= 5

90) the sources

0).

where

633

(3)

beam width (HPBW), which

is

of a long uniform broadside array.

A
rad

573
(4)

L/k

Llk

pattern of a broadside array of 20 sources of equal amplitude spaced

shown

in Fig. 14-18a.

three-dimensional pattern

is

The

BWFN

is

11.5,

and

the

HPBW is 5.1.

The

a (disk-shaped) figure of revolution obtained by rotating

the pattern of Fig. 14-18a around the array axis.

14-6.6

Array with n Sources of Equal Amplitude and Spacing:

End-fire Case

For an

end-fire array

(maximum

at 6

0) the condition fid cos

the progressive phase difference 5 between sources be fid.

COS

or

= -TTT
naj/.

h^'^J^fk)

= 5

From

requires that

(14-6.4-8)

(1)

&

634

ELECTROMAGNETICS

[CHAP. 14

=90

*' Main lobe

HPBWi5.1

Array axis

JL

-=9.5X-

20 sources

Broadside array

n=20
d=\l2

(a)

HPBW = 34
/Arrayaxis

Main lobe

,40sources
-9 = 0

L=9.75X
End-fire array

jj=40

d=X/4
(6)

FIGURE

14-18

Broadside array of 20 point sources of equal amplitude and A/2 spacing with
field pattern; (b) (ordinary) end-fire array of 40 point sources of equal amplitude
and A/4 spacing with field pattern. Note that both arrays are of approximately
(a)

the

same length

(9.5 vs. 9.75A).

For a long array (nd > kX)

2k

0o= W/A
where L

= (n

\)d

nrfif

is

large.

2k
(3)

\L/X

The first nulls (6 0l ) occur when k

Hence,

(4)

The

field

is

'

635

is shown in
The three-dimensional

pattern of an end-fire array of 40 sources spaced A/4 apart

The

Fig. 14-186.

pattern

ANTENNAS AND RADIATION

'

SEC. 14-6]

BWFN

is

52,

and the

HPBWf

is

34.

a (cigar-shaped) figure of revolution obtained by rotating the pattern of Fig.

14-186 around the array axis.

Although the two arrays

in

Fig. 14-18 are approximately the

(broadside array 9.5A, end-fire array 9.75A)

smaller for the broadside array (5 as
in the figure.

The pattern of

width in the plane of the

On

the page.

we note

compared

but the

much

is

to 34) but only in the plane

the broadside array

figure,

same length

beam width

that the

shown

disk-shaped with a small

is

beam width is

beam

360 in a plane perpendicular to

the other hand, the pattern of the end-fire array

is

ellipsoidal,

or cigar-

shaped, with the same pattern in both principal planes (in plane of page and perpendicular to page through array axis), and, as will be
1

and

2), the directivity of the end-fire array

From

(14-5-28) the directivity of

is

shown below (compare Examples

greater.

an antenna

is

given by

An

exact evaluation of the beam solid angle ClA requires an integration as

However, a very simple approximation^, can be made using the HPBWs

in (14-5-3).
in the

two

principal planes as follows:

4jt

"HP 0HP

where 0 P
< H p

HP
cr

= HPBW in
= HPBW in
= HPBW in
= HPBW in

In connection with

(6)

^HP $HP

6 plane, rad
<j>

plane, rad

plane, deg
<t>

plane, deg

note that

(6)

4?t sr

EXAMPLE

41,253

=An rad 2 = 4?r(57.3 2 ) deg 2 =

41,253 deg

For the9.5A broadside array discussed above, with 08,

(7)

= 5.1 and $ P == 360,

calculate the approximate directivity.

tThis
array

HPBW is
HPBW x

BWFN,
t

calculated exactly from (14-6.4-7).

For a long uniform end-fire
BWFN, while for a long uniform broadside array
XJ
as given in (14-6.5-4).

HPBW

Minor lobes are ignored, and 0 r Hr

Hence (6) may be in error by
<fi

angle.

is

only approximately equal to the

1 dB or so.

beam solid

636

ELECTROMAGNETICS

[CHAP. 14

From

solution

(6)

* 57FliF= 22 5(orl3 5dB)

-

EXAMPLE

For the 9.75A

(8)

end-fire array discussed above, with 0g P

= </>& = 34,

calculate

the approximate directivity.

From

solution

(6)

**35r^.-3"(i<*>
This directivity

14-6.7

is

dB more than

(9)

for the broadside array of approximately the

same

length.

Graphical Representation of Phasor Addition of Fields

At the beginning of Sec. 14-6 it was stated that much of array theory is basically little
more than the proper addition of the field contributions from all parts of an antenna.
This can be emphasized by analyzing several simple antenna arrays using only graphical

methods.
Consider the array of two isotropic in-phase sources,

spaced a distance d

A/2 (fid

180), as

shown

and

in Fig. 14-19a.

2,

of equal amplitude

We assume that both

sources have the same polarization, e.g., linear polarization with the electric field

The field pattern of the

vector perpendicular to the page.

the addition of these fields

let

the phasors

E2

have the condition given that

=E

fields

This

is

and E2 represent the

x
radiated by the sources. We

shown by arrows

done with phasors.

magnitude and phase of the two sources or of the
Thus,

array can be determined by

with proper account taken of their relative phase.

/_0 at the sources.

The arrows

(phasors)

rotate counterclockwise (ccw) on the complex plane once per cycle (360 per cycle).

In the direction 8
so that the total
(A/2) sin 45

field

90 (broadside),

E = 2E

them of 127 (= 180

E2

and

are received without phase change,

In the direction 6

0.35A closer to the distant observer than source

as reference for phase,

the total field

in this direction.

E = 0.

and

1.

= 45, source 2 is
When E is taken
1

E2

sin 45),

add as suggested, with an internal angle between

giving a total E = 0.9 Ex
In the direction 9 = 0,
.

This can be seen graphically from Fig. 14-196.

At time

0,

E2 are as shown at the sources. Let E be taken as reference for phase. Then
one-quarter cycle later (t = 7/4) the radiated field E has traveled A/4 (with no phase
change), but during this interval E2 has advanced 90. By the time t = T/2, the radiated field E has traveled A/2 and reached source 2. In the meantime the phase of E2
Hence, the
at source 2 has advanced a total of 180 and is in phase opposition to E
x

and

ANTENNAS AND RADIATION

sec. 14-6]

637

1=45'

= 90

EM^E=0.9E,

\
E=2E,

9=0

E=0

E2

E-y

= 45
E=1.8E,

/ E2
Field pattern

3- fcI

FIGURE
and

=0

K
\E*

E=2E,

'

= 270

14-19

Graphical addition of fields for array of two isotropic point sources of

equal amplitude and same phase and (c) for end-fire array of two isotropic point
(a)

(fc)

sources of equal amplitude

total field

radiated in the direction 6

and phase quadrature.

= 0

is zero, as suggested.
Adding phasors
we obtain the field pattern shown in Fig. 14-19a.t
method, we find that the fields radiated from an array of two isotropic

for other directions,

Using this

sources of equal amplitude spaced A/4 apart with source 2 lagging source
are as suggested in Fig. 14- 19c.
t

In the direction 6

= 0, E2

and

add

by 90

in the

same

The magnitude of the arrows (phasors) is considered constant as a function of distance

(even though fields do attenuate with distance) because we are interested in the
difference of the field components at a distant point. Thus, when observing at a
distance of thousands of wavelengths there is a negligible difference in magnitude

from sources separated A/2, as above, and this

phase difference cannot be neglected.

effect

can be ignored.

However, the

638

ELECTROMAGNETICS

[CHAP. 14

Field pattern-*-

48.2

~|h
FIGURE

-9 = 0

14-20

Field pattern of array of six isotropic

in-phase point sources of equal amplitude with A/2 spacing (see Table 14-3 for

graphical additional of fields in different

directions).

phase as shown (we take source

E2

and

total field
is

add

in

E = 2E

to the right (0

In the direction 9

as reference for phase).

E2

phase opposition (we take

= 0)

and

is

180,

as reference for phase). Thus, the

zero to the

left.

The complete pattern

a cardioid, as indicated.

EXAMPLE Calculate the pattern of a broadside array of six isotropic point sources
d = A/2 and draw the phasor (addition) diagram for the pattern maxima and nulls.
The
when

solution
(roots) occur

pattern

is

given by (14-6.4-5) with S

= 0.

From

(14-6.4-8) the nulls

COS0

cos 6

or

with

(1)

= -

(2)

The minor-lobe maxima occur approximately when

^ 2*+iy
(

where k

= 1, 2, 3,

"Phasor addition"

The values of k,

6,

in the table illustrates

and

the various conditions (see also Sec. 14-6.7).

close

on themselves.

The

field pattern

ifr

how

(3)

are related as in Table 14-3,

and the column

the unit phasors from each source add for

Note

of the array

that in the null directions the phasors

is

shown

in Fig. 14-20.

ANTENNAS AND RADIATION

sec. 14-6]

639

In this example and also in other cases considered in this section the total
field is

n times the

radiation.

This

field
is

from a

possible to have a total

tion

Table 14-3

Condition

is

single source

the ordinary situation.

E < nE

= nEJ in the direction of the maximum

(E

However,

in the direction

sometimes used to obtain increased

it

should be noted that

of maximum radiation,

it is

This condi-

directivity.

PARAMETERS FOR LINEAR ARRAY OF SIX ISOTROPIC

IN-PHASE POINT SOURCES OF EQUAL AMPLITUDE AND
SPACING rf= A/2. FIELD PATTERN IS IN FIG. 14-20
k

Phasor addition

=6
Maximum

(9=9bj\

90

Unitphasors

\7*=- 60

Null

70.5

j (60)

> = 60

E=
(0

= 70.5)

J -60
(9

= -70.5)

-90

Maximum

60

^(.90)

lf-

= 1.41

of minor

(9

90

= 60)

(9

= -60)

lobe

Null

48.2

f77-(120)

^j~
(9=48.2)

V*"
(9

= -48.2)

t^=150

Maximum
of minor
lobe

= -120

= 1.04

33.7

|7r(150)

'zr.150

(9=33.7)
(0=-33.7)

Null

0,180

7r(180)

=0

(9

^ = 180

\
fc-

-*

= 0)

(9

= 180)

^ = -180

640

ELECTROMAGNETICS

[CHAP. 14

Pattern

envelope

Interferometer
field pattern

Short dipole
field pattern

, Short

Shortdipole^v

'*

(element 1)

dipole

(element 2)

Transmission

FIGURE

line

14-21

Simple two-element interferometer using short dipoles spaced 10A showing

patterns of the individual (short-dipole) elements and of the interferometer.

14-6.8

The

Simple Two-element Interferometer

interferometer

observations.

It

is

a specialized type of antenna used widely for radio-astronomy

provides a good example of pattern multiplication and array theory.

Consider two short-dipole antennas with spacing s connected in phase, as in

From

Fig. 14-21.

(14-6.1-2) the field pattern

E = 2E

where ^

By

fis

tern

cos

given by

(1)

cos

pattern multiplication

(14-3-37)],

is

and

is

the field pattern of the short dipole [=sin0, see

cos(i^/2) is the array factor.

Hence, the complete interferometer pat-

is

E = 2 sin

cos/y cos

(2)
J

When

is

(7tj/A)sin y 01

replaced by

its

complementary angle

y,

the

first

null

occurs

when

u\2 or
?oi

_,

sin

(3)

2sjX

ANTENNAS AND RADIATION

sec. 14-6]

641

Interferometer
field pattern

Element

Element 2

Steps

in pattern multiplication

A
Pattern of

Interferometer

individual

pattern

element
Pattern of 2
isotropic sources

spaced 20X(array factor)

(b)

FIGURE

14-22

Two-element interferometer using 40-point-source end-fire arrays spaced 20A

showing patterns of the individual (end-fire array) elements and of the inter(a)

ferometer.

(6) Steps in pattern multiplication to obtain interferometer pattern.

Individual element pattern (at

left)

multiplied by array pattern (center) yields

interferometer pattern (at right).

If

we assume

that the spacing s

is

many

wavelengths, the

BWFN = 2y 01 = -L rad =
s/A

BWFN

is

then

deg

(4)

s/A

This

BWFN is one-half the value for a uniform array of n sources as given by (14-6.5-3).

The

interferometer field pattern

is

illustrated

in

Fig.

14-21

for s

1(U.

The

642

ELECTROMAGNETICS

BWFN

[CHAP. 14

of this interferometer pattern

the multilobed pattern

is

most applications more

An

is,

from

equal to

(4),

The envelope of
(= sin 0). In

5.7.

the same shape as the short-dipole pattern

would be

directional individual element patterns

interferometer array of this type

shown

is

desirable.

Here two 40-source

in Fig. 14-22a.

one described in Sec. 14-6.6) are arrayed with a spacing of

Each 40-source array has a field pattern as shown (see also Fig. 14-18*). When

end-fire arrays (like the

2(U.

connected as a two-element broadside array (or interferometer), as in Fig. 14-22a, the

result is

a multilobed pattern with

BWFN = 2.9

The

(=57.3/20).

steps in the

pattern multiplication are illustrated in Fig. 14-226.

In radio-astronomy applications an interferometer is used for receiving radiation

frdm celestial radio sources as they drift across the sky through the antenna pattern.
The detection and location of the source are facilitated by the small beam widths of
the lobes of the patterns.!

CONTINUOUS APERTURE DISTRIBUTION

14-7

In the preceding pages

finite

we have

dealt with arrays of point sources,

number of sources separated by

tinuous array of point sources,

i.e.,

separated by infinitesimal distances.

array of point sources

tion).

In this

is

finite distances.

an array of an

By Huygens'

infinite

i.e.,

arrays of a

us consider a con-

let

number of point sources

principle (Sec. 12-6) a continuous

equivalent to a continuous field (or current-sheet distribu-

way we can extend array

field distributions

Now

analysis to include the radiation patterns of

across apertures of electromagnetic horns or parabolic dish antennas.

Consider the aperture distribution of length a in the y direction and width x t

in the

x direction

of the

field

The magnitude
The field is in
a distance from an

(perpendicular to page), as suggested in Fig. 14-23.

over the aperture

is

a function of y but

the same phase over the entire aperture.

elemental aperture area dx dy

is,

dE = -jm[dA x] =

from

is

independent of x.

The infinitesimal field

at

jcoA and (14-3-7),$

--^y e-Jfr^, dx dy

(v

m" )

(1)

= x component of retarded vector potential [see (14-2-6)], Wb m -1

_1
E(y) = electric field distribution as function of y across aperture, V m
Z = intrinsic impedance of medium, fl

where [A x ]

The

total field in the

6 direction at a distance

from, the center of the aperture

much simplified. For a more detailed discussion see, for example,

D. Kraus, "Radio Astronomy," chap. 6, McGraw-Hill Book Company, New
York, 1966.

t This discussion is
J.

Note

that E(y)/Z

= EJy)IZ H, = Jx Zi, where z,

current sheet in the aperture.

Thus, in (14-3-7)

is

let /

the thickness of the equivalent

= J, Zi dy and ldx.

ANTENNAS AND RADIATION

sec. 14-7]

643

Continuous
aperture
distribution

FIGURE

14-23

Aperture of width a and amplitude distribution 0).

is

obtained by integrating (1) over the aperture (extent a by

y cos

i(a"-'lr )

E(0)

-j(onx i e

The magnitude of E(0)

is

= 90) we

doubled.

is

= r^

= Ea ]

(3)

(3) reduces to

expOfiy cos &) dy

Ea ax

(4)

Ea A
(5)

the aperture area in square meters.

is

0)dy

have

radiation at

This

(2)

+/2

P"

E(0)

A =ax

E(y)cxp(jpy cos 6) dy

E(y)cxp(jlly cos

ZT A J-a/2

(maximum

+ "2

distribution \E(y)

E(0)

where

then r

J -a/2

^7lS
Y

axis (9

Hr^a,

thent

m
For a uniform aperture

4xr

On

x t ).

and

= 90

but none at

If the aperture

is

unidirectional

270), the value of E(0) should

the value which was used in deriving (14-5-25).

be

Integrating (4)

gives

E(0)

Ea A

sin[(j8a/2)cos 0]

2r X

(Pa/2)cos0

This gives the far-field pattern as a function of

of field intensity

(6)

for a continuous, uniform distribution

Ea over a rectangular aperture of area A and length a (in the y direc-

tion).
t

Note that o>iil4nZ=

1/2A.

644

ELECTROMAGNETICS

[CHAP. 14

Referring to the array of n isotropic sources of equal amplitude and spacing

we

discussed earlier in Sec. 14-6.4,

recall that the field pattern

E{0)

was given

in (14-6.4-5) as

sin m///!

=E

(7)

sin y\2

= fid cos 6 + S
E = field from individual source
For the broadside case, = 0, and (7) becomes

where

xji

<5

wm = m? sin[(M2)cos 01

.(0)

sin[(M2)cos

Let

a'

= length

Then

of array.

if a' is large, a'

attention to angles near 90 (main

E(9)

beam

= nE

(8)

6]

direction),

(n - \)d nd.
Restricting our
we can rewrite (8) as

sm[(pa'/2)cos 0]
(9)

(jSa'/2)cos

Comparing this expression with (6), we see that for a large aperture and at angles near
main beam, the field pattern of a continuous aperture is identical with the pattern
of an n source broadside array of same length (a = a'). Also the fields have the same

the

absolute value provided

nE = EaA/2r

l.f

FOURIER TRANSFORM RELATIONS BETWEEN

THE FAR-FIELD PATTERN AND THE
APERTURE DISTRIBUTION*

14-8

A one-dimensional aperture distribution E(y) and its far-field pattern E(6) are reciproFor a

cal Fourier transforms.

finite

aperture one of these transforms

may

be written

-.+0/2

E{B)

E(y) expijfiy cos 0) dy

(1)

J -a/7.

This

is

identical with (14-7-3) except for a constant factor.

field patterns

E(9) of several

Dividing by

field distributions

n,

we

Ea
J

-
~

E2/-

A"

2r A

the field at a distance r from a source of effective aperture

This is equivalent to (14-7-5).
in the aperture.

where

far-

get

F
io

Examples of the

E(y) are presented in Fig. 14-24.

is

A and

field

H. G. Booker and P. C. Clemmow, The Concept of an Angular Spectrum of Plane

Waves and Its Relation to That of Polar Diagram and Aperture Distribution,
Proc. IEE London, (3) 97:11-17 (January 1950); C. H. Walter, "Traveling Wave
Antennas," McGraw-Hill Book Company, New York, 1965.
Equation (14-7-3)

is

an absolute

relation,

whereas

(1) is relative.

ANTENNAS AND RADIATION

sec. 14-8]

645

Field pattern

Aperture distribution

EM)

E(y)

Uniform
77,

'A

Triangular

srTlMfrTTT^

Cosine

Cosine squared

AT"
Inverse taper

Edge

FIGURE

14-24

Seven different antenna aperture distributions with associated

far-field patterns.

646

ELECTROMAGNETICS

[CHAP. 14

Taking the uniform aperture distribution and

we

ence,

its

find that the tapered distributions (triangular

pattern (Fig. 14-24a) as refer-

and cosine) have

larger

beam

widths but smaller minor lobes, while the most gradually tapered distributions
(cosine squared and gaussian) have still larger beam widths but no minor lobes.
On
the other hand, an inverse taper, as in Fig.

minor lobes than for the uniform

larger
its

14-24/", yields

beam width but

a smaller

Carrying the inverse taper to

distribution.

extreme limit results in the edge distribution of Fig. 14-24^, which

that of a two-element interferometer.

The

is

equivalent to

beam width one-

pattern in this case has a

half that of the uniform distribution but side lobes equal in amplitude to the

main

(central) lobe.

A useful property of (1) is that an aperture distribution can be taken as the sum
of two or more component distributions with the resulting pattern equal to the sum of
the transforms of these distributions.

LINEAR ANTENNAS

14-9

In Sec. 14-3 the dipole antenna was analyzed for the case where

its

length was small

compared to the wavelength (say 2/10 or less). Let us now consider

antennas of any length L. It is assumed that the current distribution
This

is

a satisfactory approximation provided the conductor diameter

A/200 or

linear dipole
is
is

sinusoidal.

small, say

less.

Referring to Fig. 14-25,

we

shall calculate the far-field pattern for

thin linear center-fed dipole antenna of length L.

It will

a symmetrical

be done by regarding the

antenna as being made up of a series of elemental short dipoles of length dy and current
/ and integrating their contribution at a large distance. The form of the current distribution

on the antenna

is

found from experimental measurements to be given

approximately by
/

where
is

/ is the value

used when y

fields

= /

sin

of the current at the current

< and

(L/2)

-yis

(A)

[T(? J')]

used

of a short dipole, we find the far

(1)

maximum point, and where (L/2) + y

when y >

0.

(electric) field

Referring to Table 14-1 for the

of an antenna of length

L to

be

,L/2

Ee = ksmd\

I ezpiJPy cos

8)

dy

(2)

'-L/2

where k

= {jWk exp[j(cot - Pr

yields

Ee

)]}/r

= J60[I

Substituting (1) for / in (2)

IcosKpL cos
\

6)/2]

ihTe

- cos (/?L/2)

and integrating

(3)

ANTENNAS AND RADIATION

sec. 14-9]

647

To distant
fioint

Sinusoidal
current distribution

FIGURE

14-25

Geometry

for linear center-fed dipole antenna of length L with sinusoidal current

distribution.

where
field

[I

pattern

I exp{jco[t
is

(r /c)]} is

The shape of the

For a A/2 dipole antenna

the retarded current.

given by the factor in braces in

(3).

farthis

factor reduces to

EM

cos[(?t/2) cos 9]

(4)

with the shape shown in Fig. 14-26a.

sin

This pattern

is

only slightly more directional

than the sin 9 pattern of an infinitesimal or short dipole (shown dashed).

HPBW of the A/2 antenna

is

78,

compared with 90 for a short

The

dipole.

antenna
field pattern

-|-

A antenna
field pattern

Current distribution

I)

Current direction

Short-dipole field

pattern (for comparison)

(6)

(o)

FIGURE

14-26

(a) Far-field pattern of X/2 center-fed dipole

l.SA center-fed dipole antenna.

antenna and

(b) far-field pattern

of

648

ELECTROMAGNETICS

[CHAP 14

For a 1.5A dipole the pattern factor

cosftw cos 9)

W =
shown

as

in Fig.

phase of the far

(5)

sin

'

With the midpoint of the antenna

14-26*.

a constant r

field (at

is

shifts 180 at

each

lobes being indicated by the plus and minus signs.

dimensional pattern

is

as phase center, the

null, the relative

In Fig. 14-26a

a figure of revolution of the one

phase of the

and b the

three-

shown with the dipole

as

axis (y axis in Fig. 14-25).

EXAMPLE

Find the

solution

directivity of

a A/2 linear dipole.

For the A/2 dipole we have from

D = TTF7nri7\ =
JJ P(,8,4>) dO.

7i

[(tt-/2) cos
cos 22r/

^-
2

2tt

is

1.64/1.5

1.09

(=0.4 dB)

To find

the radiation resistance

is

m
6]
-

sin 6

sine (Si)

(=2.15 dB)

1.64

(6)

do

and cosine

an

infinitesimal or short

(Ci) integrals.

of the A/2 dipole the average Poynting vector

integrated over a large sphere, yielding the

equated to I 2 R, where /

(14-5-7) that

greater than the directivity of

Note The integration of (6) involves

dipole.

is

and

sin

Ja

This

(4)

power

the rms terminal current.

radiated. This

Performing

power

then

is

this calculation for

the A/2 dipole yields

R=
where Cin

is

30 Cin In

the modified cosine integral.

30 x 2.44
This

is

73

Q (Z =

(7)

the resistance presented by a A/2

dipole to the transmission-line terminals (Fig. 14-26).

also contains a positive reactance of 42.5

fi

The antenna impedance

resonant

(X =

results in

a slight reduction of the radiation resistance (to approximately 70

0) the length can be shortened a few percent.

radiation resistance for the

FIELDS OF

14-10

To

1.5/1

A/2

The

fi.

of the fields radiated by an antenna, the electric field

of a A/2 antenna are shown in Fig. 14-27 at four successive instants of time.

FIGURE

14-27

(rf) /

= |r where T= period.

time advances.
f

about 100

this also
1).

DIPOLE ANTENNA

Electric field configuration for a A/2

at

is

However,

illustrate the configuration

lines

and

linear dipole

To make the antenna

73 +y'42.5 H).

r/4

is

antenna at four instants of time

= 0,

(b)

7/8, (c)

P, P",

= 7/4,

and P" as

The strength of the electric field is proportional to the density of the lines. The field
shown on the book cover. (From "A Resonant Spheroid Radiator" produced at

also

the Ohio State University for the National

Prof.

(a)

Note outward movement of constant-phase points

Edward M. Kennaugh; diagrams

Committee for Electrical Engineering Films; Project Initiator,

courtesy of Prof. John D. Cowan, Jr.)

649

650

ELECTROMAGNETICS

Although the

[CHAP. 14

were computer-generated (and plotted) for a prolate spheroid

fields

with A/2 interlocal distance, they are substantially the same as for a thin A/2 center-fed
dipolef except in the immediate proximity of the antenna.
the antenna current

is

zero,

One-eighth of a period later

and the charge

(t

At time

J/4 (Fig. 14-27c) the current has reached

is

zero.

At

At

(Fig. 14-27a)
is

maximum.

T/S) the current has started to flow, as suggested in

charge at the ends

Fig. 14-276.

at the ends of the dipole

= fT (Fig.

its

peak value, and the

\A-21d) the current continues to flow in

same direction but with reduced magnitude. At t = T/2 the current is again zero,
and the charge is a maximum at the ends. The field configuration at t = T/2 is
identical with that shown in Fig. 14-27a except that the arrows and signs are reversed.
At t = |- T the current has started to flow in the reverse direction (up), and the field is
identical to that shown in Fig. 14-276 except that all arrows are reversed.
Note that
the

the lines in Fig. 14-27 are the actual field

Compare with

angle).

Referring to
P,

moves

radially

all

parts of Fig. 14-27,

c).

field

pattern

(field intensity vs.

in

point, such as

\ period (traveling approximately at

However, careful measurements show that

c close to the antenna, like

parallel plates

not the

we note that a constant-phase

outward approximately A/2

the velocity of light,

than

lines,

Fig. 14-26a with the dipole turned vertical.

moves

faster

a higher-order-mode wave confined between the

of a waveguide, but farther out the velocity becomes exactly equal to

field lines are not shown.

They
and perpendicular to the plane of
the figure.
From the standpoint of power flow the situation is similar to that shown
by the arrows in Fig. 14-10 for the short dipole with maximum power radiated outward
This is real power flow. There is also reactive
in the equatorial plane (6 = 90).
(nonradiating) power flow, oscillating back and forth close to the antenna, as suggested

c.%

Figure 14-27

is

incomplete in that the magnetic

consist of circular loops concentric with the antenna

in Fig. 14-10.

14-11

TRAVELING- WAVE ANTENNAS

In Sec. 14-9 on the linear antenna a sinusoidal current distribution was assumed.

This type of current distribution

is

closely approximated if the antenna

ductor diameter <A/200) and provided the antenna

t

is

is

thin (con-

unterminated.

For the thin A/2 dipole the E component would be as obtained from an equation
but one which holds at small distances (see General expression in
Table 14-1). Similarly calculating an E, component makes it possible to determine
the total field at any point.
like (14-9-2)

This

may be

inferred

(P, P', P") are

from the

more than

fact that the distances

between similar phase points

A/2 apart near the antenna but are exactly A/2 apart farther

out.

For a more detailed discussion of traveling-wave antennas,

op.

cit.

see, for

example, Walter,

ANTENNAS AND RADIATION

SEC. 14-11]

651

Standing wave
(sinusoidal current distribution)

Outgoing wave

Linear antenna

Transmitter

(o)
circuit

graveling wave

Outgoing wave

(current amplitude)

(6)

Transmitter

No reflected wave

3-

-Termination

(c)

(d)

Infinitesimal
Field pattern of infinitesimal dipole

dipole

FIGURE

14-28

Long linear antennas (a) without and

(b)

and

(c)

with termination; (d) equivalent

end-fire array of isotropic point sources.

Consider a thin linear antenna several wavelengths long situated a distance h

above a perfectly conducting ground plane, as in Fig. 14-28a, with a transmitter con-

The antenna-ground-plane combination acts like an openThe transmitter sends a traveling wave down the line
(to the right), which is reflected at the open end, causing an equal-amplitude wave to
travel back (to the left).
As a result a standing wave is set up on the line with a
sinusoidal current distribution. The current is zero at the open end and is a maximum
nected at the

left

end.

circuited transmission line.

A/4

from the open end, as suggested in Fig. 14-28a (see also

assumed in Sec. 14-9 on the linear dipole antenna.

Sec. 13-6).

This

is

the

situation

Let us

now

connect a resistance

antenna-ground-plane transmission
ground, as in Fig.

14-28Z>.

line

equal to the characteristic resistance of the

between the open end of the antenna and the

This converts the antenna to a terminated type.

There

is

652

ELECTROMAGNETICS

[CHAP 14

an outgoing traveling wave

amplitude

and there

(to the right) but (ideally) no reflected wave.

The current
constant over the length of the antenna provided losses are negligible
a progressive uniform phase change along the antenna. An arrangement

is
is

for establishing a traveling

suggested in Fig. 14-28c.

wave on a long linear antenna remote from the ground is

The midpoint of a A/2 antenna is used as a ground equiva-

lent.

The pattern of a

we

linear antenna with traveling

regard the antenna as being

made up of an

wave can be

readily determined if

number of seriesconnected infinitesimal dipoles operating as an end-fire array, as in Fig. 14-28<f.

Thus, from Sees. 14-6.6 and 14-6.4 the far electric field pattern of such an array is given
array of a large

by

-h
bFtm-F
e (V)
where f

= fid cos

fid

pd(cos

W
sin

2)
(1)
v
'

0-1).

sm(\j//2)

The first factor (E

represents the pattern

of the infinitesimal elements of the array, which have the same pattern as a short
dipole (E = sin 0). The second factor is the normalized pattern of an array of n
isotropic sources.

[Ee (6)]

is

By

the principle of pattern multiplication the resultant pattern

the product of these two factors.

from a large n and small

d, (1)

For large

[=d(n

1)

nd],

resulting

can be expressed approximately as

(/sL/2)(l

For example, the pattern of a 9.752

'

cos 0)

linear antenna with traveling

wave

is

shown

in

Fig. 14-29a as the product of the pattern of

an infinitesimal dipole and the pattern of

an end-fire array of 40 isotropic sources with A/4 spacing and uniform progressive
phase shift 5 = -fid= -iz/2. This is the same array as in Sec. 14-6.6 (see also Fig.

The pattern of the array of 40

14-186).

(end-fire direction).

= 0

However, the

isotropic sources has a

infinitesimal dipoles

maximum

at

making up the array have a

= 90. The resulting (product) pattern is of

shown in Fig. 14-29a (pattern at right). The first
minor lobe is only 4.7 dB down. The large minor lobes are characteristic of long
linear antennas.
By contrast, the first minor lobe of the isotropic-source array is
null at

and a maximum

at

conical shape with cross section as

13.5

dB down (see Fig. 14-29a, center pattern).

The maximum radiation for the 9.751 linear antenna is at

two such

linear antennas in the

form of a

Fig. 14-29Z>, a unidirectional antenna

balanced two-conductor transmission

antenna

is

is

1 5. By arranging
with about 30 included angle, as in

obtained which can be conveniently fed by a

line.

The

resulting pattern of this terminated

suggested in Fig. 14-296 (pattern at right).

653

sSag
oo
K
-2

13

"^

CA

>

o
it.

cct

s &<
arr

obtai

form

re

_ ^3 ^v
c C
a

CO

rt

c
h

i
ca

i*

*5

"OSS
u

1*1
S
(t

* s
t
of

pattern

inear

on

(5

^vi

CL

~;

re

-S'3
o c
.S-3

ov -a
of

"5

Co

short

tern

J
1

'

ON ts
<N CO
"*

fo
lication

(6)

sources.

r-field

06

^3
E3EK
-a

654

[CHAP. 14

ELECTROMAGNETICS

The vertical patterns of these antennas when operated parallel to a ground plane,
by multiplying the patterns of the isolated antenna
by that of a pair of isotropic sources, one at the center of the antenna and the other (as

as in Fig. 14-286, can be obtained

its

image) an equal distance below the ground plane (see Sec. 14-20.1).

THE SMALL LOOP ANTENNA

14-12

Consider a small circular loop antenna of diameter d, as in Fig. 14-30a. Its dimenassumed small compared with the wavelength (d < A), so that the current

sions are

of constant amplitude and phase around the loop. This small circular loop may be
2
2
approximated by a square loop of equal area [s = n(d/2) ], as suggested in Fig.
14-306.
In this way the loop can be analyzed as four short dipoles whose properties

is

we

Let the square loop be oriented with respect to the

considered in Sec. 14-3.

coordinates as in Fig. 14-306 and

The

c.

field in

the yz plane

from the two short dipoles 2 and 4. Short dipoles

since their fields are exactly equal and in opposite phase

is

produced by the

and

do not conyz

radiation

tribute

at all angles in the

Since the short dipoles (2 and 4) are nondirectional in the yz plane, the field
is the same as for two isotropic point sources in opposite phase.

plane.

pattern in this plane

Thus, the

electric field pattern is given

E(6)

E2 = EA

where

is

by

= E2 e- ll

-Ei e *

i 12

(1)

the electric field from the individual short dipole.

Assuming

s <^X,it follows that

E(0)= -JEzPs
Introducing the value of the far

field

E2

E(6)

sin

(2)

of a short dipole from Table 14-1, we get

^Ps

sin

(3)

rX

where

is

But

the length of the short dipole.

in

our case

=s

and

=A

(=area of

loop); so

(0) =

-jsmO
T

(Vm

(4)

A.

the value of the electric field from a small loop antenna of area A (square or
The
It is the same in any plane perpendicular to the plane of the loop.
circularf).

This

is

approximation (square for circular loop) becomes exact as the loop dimensions become
very small.
t

Or any other

shape.

sec. 14-13]

ANTENNAS AND RADIATION

655

Dipole3

Dipole2

FIGURE

14-30

(a) Circular

14-13

and

(6) square

loop antennas;

(c)

geometry for calculating far

field.

THE SMALL HELICAL ANTENNA

A wire wound in a helical coil is shown in Fig.

small compared with the wavelength (d

14-31.

Assuming that the diameter is

the radiated field can be readily determined by regarding the helix as a series of short linear
dipoles and small loops connected in series, as in Fig. 14-316. The field from one
turn of the helix is then the sum
of the fields from a short dipole and a small loop with equal
in-phase currents. From
Table 14-1 the field of the short dipole is

._.

X),

.607t/ /

Ee {&) =y

sin 9

(1)

Sin 9 patterns

X to page
Loop
(e)

FIGURE

14-31

(a) Small-helix

antenna with (b) equivalent arrangement of small loops

dipoles and (c) associated far-field patterns.

and short

656

[CHAP. 14

ELECTROMAGNETICS

while from (14-12-4) the field of a small loop oriented as in Fig. 14-31c

is

^) = I^sin0
The

Ee (6)

electric field

while the electric field

of the short dipole

patterns are independent of 0; that

about the

We

vertical axis.

in general, the field

lies in

E^B) of the small loop

perpendicular to the page.

Both

contains

(1)

&j

while (2) does not.

Thus,

have two orthogonal components which are

will

Dividing

time-phase quadrature.

the plane of the page (Fig. 14-31c),

is

they are figures of revolution of the ones shown

is,

note also that

of the helix

(2)

by

(1)

(2),

we obtain

in

for the axial ratio

(3)

If the condition 11

magnitude

= 2nA

the radiation

is fulfilled,

but

of the field varies as sin

where the

(except

on

ever, if

the axial length of the helix

axis,

This

field is zero).|

is

is

directions

all

for a small helix of one turn.

small, the field pattern

is

Then the

circularly polarized.

circularly polarized in

is

How-

and polarization of a
This can be seen from

be substantially the same as for one turn.

the principle of pattern multiplication where the multiturn helix is represented by a
row of point sources each with the pattern of one turn. If the row of point sources is
short compared with the wavelength, there will be but little difference between its

number of turns

pattern

14-14

will

and the pattern of one source.

THE HELICAL-BEAM ANTENNA*

In the preceding section

maximum

we

discussed the radiation properties of a small helix.

radiation of a small helix

the helix diameter

entirely different

is

larger, such that the circumference

mode of radiation

beam antenna with maximum

circularly polarized

length of the helix;

on

is

broadside (perpendicular to helix

axis,

i.e.,

occurs.

and

In this case the helix acts as an end-fire

the directivity

Helical

If

approximately U, an

radiation in the axial direction.

Antenna

The

radiation

is

of the helix increases linearly with the

doubling the length doubles the

f H. A. Wheeler,

is

The

axis).

directivity.

To

distinguish

for Circular Polarization, Proc. IRE, 35: 1484-

1488 (December 1947).

t

D. Kraus, Helical Beam Antenna, Electronics, 20: 109-111 (April 1947); also
D. Kraus, Helical Beam Antennas for Wide Band Applications, Proc. IRE, 36:
1236-1242 (October 1948). For extensive treatment see J. D. Kraus, "Antennas,"
chap. 7, McGraw-Hill Book Company, New York, 1950.

J.

J.

sec. 14-14]

ANTENNAS AND RADIATION

657

Ground plane
Charge

|-tof x

Traveling v

"

V,

Current

End-fire array (n isotropic sources)

Ordinary end-fire array pattern (for comparison)

(<*)<

CO

Helical

Single-turn

End-tire array

pattern (cos $)

(increased directivity)

beam

antenna pattern

(20turns,a= 12)

FIGURE

14-32

and (6) Dimensions for helical-beam antenna, (c) End view of helix
with
instantaneous direction of rotating electric field (E).
(</) Far-field pattern of
helical-beam antenna obtained by pattern multiplication of
single-turn pattern
with end-fire array of isotropic sources with increased
directivity.
The pattern
for an ordinary end-fire array of the same length is
shown for comparison.
(a)

this end-fire

mode of helix

operation from the broadside

mode of the

small helix,

we

refer to the end-fire helix as a helical-beam antenna.

The helical-beam antenna is conveniently fed with a coaxial transmission

and square or circular ground, as suggested in Fig. 14-32a. The
important

line

dimen-

sions of the helix are

its

diameter D, circumference C, spacing between turns S, turn

length L, and pitch angle

circumference (C

= nD)

a.

These are related as

in Fig.

14-326.

When

the helix

approximately U, the instantaneous charge and current

distribution on a single turn will be as suggested in the end
view of Fig. 14-32c, with
the resultant
tric field

E field

is

as indicated.

As

the

wave

travels

outward on the helix, the elecon axis. The helical-beam

vector rotates, producing circular polarization

658

ELECTROMAGNETICS

[CHAP. 14

The

antenna acts as a traveling-wave antenna without requiring a termination.

current distribution

of a

essentially that

is

single outward-traveling

wave of uniform

amplitude; the wave reflected back toward the feed from the open end of the helix
is

very small.

The pattern of a helical-beam antenna can be

readily calculated if

it is

regarded

By the

as an end-fire array of isotropic sources, each source representing one turn.

principle of pattern multiplication the actual helix pattern

turn pattern has

much

little effect,

cos

This

6.

then the product of this

The single-turn pattern may be approxima-

array pattern and the pattern of one turn.

ted as a cosine function, E(9)

is

is

an oversimplification, but the

single-

especially for long helices, because the array pattern

Thus, we

sharper than the single-turn pattern.

cvm
E(9)

may

is

so

write for the field pattern

sin(ntfr/2)

cos

(1)

sin(i/'/2)

where

/>

= pS cos

- f

(2)

We

note that

from the

(1) differs

linear traveling-wave antenna pattern as given in

(14-11-2) in that the single-turn helix pattern (cos 0) has a

the same direction as the array

dipole pattern has a null
in the conical pattern

maximum)

on the
is

If all fields

added

in

is,

phase on

Ee (0)

axis,

maximum),

resulting

less directivity.

circularly polarized

pattern for both components, that

the axis (in

while for the linear antenna the infinitesimal

axis (in the direction of the array

of Fig. 14-29a and

Since the radiation

maximum on

on

axis

and symmetrical,

(1) is

the

and E^O).

we would have

the ordinary end-fire condi-

which requires ^ =0, 2n, An, etc. In our case ip = -In. Comparing a
with the actual measured pattern shows that
pattern calculated using this value of
tion,

\//

the actual pattern

is

much

sharper or

more

directional (see Fig. 14-32rf).

To

obtain a

calculated pattern which agrees with the actual one, approximately n/n additional

phase change per turn

is

required, so that the total phase change per turn isf

xjj=-2n--

(3)

The

fulfillment

of

this

requirement implies that the wave on the helix travels with

a velocity which automatically

satisfies (3).

Thus, we have from

_ 27t _* = /?5
n

_^

(2)

and

(3)

(4)

Hansen and Woodyard have shown that the additional ir/ phase change results
increased directivity; W. W. Hansen and I. R. Woodyard, A New Principle
Directional Antenna Design, Proc. IRE, 26: 333-345 (March 1938).

in
in

'

ANTENNAS AND RADIATION

SEC. 14-14]

= v/c = relative phase velocity pf wave on helix,

velocity of wave on helix, m s"
~
c = velocity of light or radio waves = 300 Mm s
L - length of turn, m
S = spacing between helix turns, m

where p

659

dimensionless

Solving (4) for p,

we

get

P=

(S/A)

(2

In a typical case where the helix dimensions are

(5)

l)/2

C = A,

oc

12,

and n

20,

we obtain

a phase velocity p = 0.82. This calculated value agrees with the actual measured
value.
It is a remarkable property of the helical-beam antenna that the relative phase

wave traveling out on the helix adjusts automatically, so that increased

obtained over a considerable range of pitch angles (5 < a < 20) and

velocity p of the
directivity

over a 2 to

is
1

bandwidth

higher by increasing the

(ik<C < \X) provided

n > 3. The directivity can be made

The practical limit here is that the dielectric
should not support other modes to any appreciable extent.

number of turns.

structure holding the helix

The

patterns shown in Fig. 14-32*/ are those of a 20-turn helix with a = 12 and an
ordinary end-fire array of the same length with 20 sources each with a cos 6 pattern.

The approximate

directivity

of a helical-beam antenna

is

given byf

(6)

For a 20-turn

helix with typical parameters

C = A and a =

12 the directivity

D = 63

(or 18 dB).

With a pitch angle a = 12, S = 0.213A; so there are nearly 5 turns per waveand the 20-turn helix is 4.3A long (=nS = 20 x 0.2m). An ordinary end-fire

length,

array of this length has

than one-fourth of this directivity. Because of this fourimprovement of a helical-beam antenna over an ordinary
antenna of the same length, the helical-beam antenna can be called a superless

fold directivity, or gain,

end-fire

gain antenna.

The high

gain, large bandwidth, simplicity,

helical-beam antenna have

tions.

made

it

and

circular polarization{ of the

indispensable for space-communication applica-

For example, helical-beam antennas are used on the moon for transmitting
t Kraus, "Antennas," p. 197.
t

With linearly polarized antennas, the plane of polarization of the wave from the
transmitting antenna should be aligned with the receiving antenna
in

[MM=0

(11-4-4)].

(MMa =

When
180).

the planes are perpendicular, the received power goes to zero

to rotation of a space vehicle or to Faraday rotation in the

Due

ionosphere (Chap. 1 5), the plane of polarization may rotate and cause signal fading
This fading is overcome with circularly polarized antennas.

and dropouts.

660

ELECTROMAGNETICS

FIGURE

[CHAP. 14

14-33a
Helical-beam antenna with 18 dBi gain placed on Fra Mauro highlands of the
moon by Apollo 14 astronauts for continuous transmission of telemetry data
back to earth from the ALSEP experiment. The helix conductor is supported
(National Aeronautics and Space Administration.)
by a thin plastic tube.

ANTENNAS AND RADIATION

sec. 14-14]

661

***Si
MlllIllS8E

WMm

:;
\.

>

.....

*
.._*

~s

.\

.....

1/

FIGURE

14-33*

Broadside array of 96 helical-beam antennas at Ohio State University Radio

Observatory.

662

ELECTROMAGNETICS

[CHAP. 14

telemetry data back to earth from the

Package) equipment.
the

moon by

shown
later

ALSEP

(Apollo Lunar Surface Experiment

A photograph of one of these helical-beam antennas placed on

the Apollo 14 astronauts, Alan Shepard,

in Fig. 14-33a.

Apollo groups.

tions, as in radio

Jr.,

and Edgar Mitchell, is

set up on the moon by

Other helical-beam antennas have been

The

many other applica300-MHz Ohio State University

helical-beam antenna also has found

astronomy.

The

original 200- to

radio telescope (in 1952) consisted of a broadside array of 96 helical-beam antennas,

as

shown

400-MHz

in Fig. 14-33&, with each helix of 11 turns

and

m long. The 300- to

number of helical-beam

1.5

radio telescope of the University of Texas has a

Each array

antenna arrays in an interferometer arrangement.

consists of 100 helices

each of 40 turns.

14-15

BEAM WIDTH AND DIRECTIVITY OF ARRAYS

In the preceding sections

we have

discussed the directional properties of a

types of uniform arrays, both broadside and end-fire.

tions for the

BWFN,

the

HPBW, and

directivity for

number of

Table 14-4 summarizes

rela-

a uniform linear broadside

array and an ordinary end-fire array of isotropic sources and for a helical-beam antenna

(end

fire

listed for

with increased directivity).

In addition the

beam widths and

directivity are

a unidirectional square-aperture broadside array with uniform distribution.

Table 14-4

ARRAY BEAM WIDTHS AND

Type of array

fire

Square-aperture
broadside array
(unidirectional)

L = array

51

i/A

/A

Djrectivity

162

108

VZ/A

VZ/A

3.51/A

52

78
(C/A) 2

VZ/A

(C/A) 2

115

51

L/X

L/A

VZ/A

(D"f
13 (Z./A) 2

length for end-fire arrays or side length for broadside array

C = circumference of helix
A

115

(ordinary)

Helical-beam antenna
(end fire with increased directivity)

HPBW

2.2L/X

Broadside

End

BWFN

DIRECTIVITIESf

= wavelength

ANTENNAS AND RADIATION

SEC. 14-16]

663

SCANNING ARRAYS

14-16

In Sees. 14-6.5 and 14-6.6

we considered broadside and end-fire arrays.

side array all sources are in

phase

= 0),

(S

progressive phase difference between sources

sources.

Actually the

In the broad-

while in the (ordinary) end-fire array the

Pd,

is

beam maximum can be

where dis the spacing between

any arbitrary direction max if

tilted in

the phase difference 5 satisfies the relation

<5=-/Wcos0max
For a broadside array
3

= fid. To

0.707 pd. By
beam

max

direct the

90 and

maximum

<5

(1)

For an

0.

end-fire array B max

radiation at 45 (0 max

= 0

and

45) requires that d

controlling the phase difference between elements of

angle can be adjusted or scanned without turning the antenna.

an array the

Such arrays are

sometimes called scanning arrays or antennas.

An interesting example of the scanning antenna is the slow-wave backward anglefire grid-structure

array

shown

in Fig. 14-34, f in

The

accomplished by shifting the frequency.

which the beam scanning action

shown

array

in Fig. 14-34o

has

is

7x4

Each mesh has dimensions / by s. The arrows indicate the approximate

The array may be placed a distance h above a
ground plane and fed from one end, as suggested in Fig. 14-34a and b. Typical parammeshes.

instantaneous current directions.

The approximate is-plane pattern can

/ = 2.75s, h = 0.3s, and $ 0.35/1.
be calculated assuming each mesh to be represented by a point source with a short

eters are

dipole (sin 6) pattern.

Thus, from (14-6.4-5) the .E-plane pattern

Ee(0)

.\j'

sin

is

(2)

sm(iA/2)

where

i/r

Ps cos 8

(3)

Assuming a traveling wave moving along the wires with velocity c, the phase
between successive elements such as 1 and 2 (Fig. 14-34a, b, and c) is

difference

S=~Pj-Ps
For the

direction of maximum radiation

In our case

\j/

= 2n,

max

it is

(4)

required that

\ji

0,

2n, 4n,

etc.

so (3) becomes

2* /

T v cos

~2~f

(5)

D. Kraus, A Backward Angle-fire Array, IEEE Trans. Antennas Propag., AP-12:

48-50 (January 1964); J. D. Kraus, U.S. Patent 3,290,688 Dec. 6, 1966; J. D. Kraus
and K. R. Carver, Wave Velocities on the Grid-structure Backward Angle-fire
Antenna, ibid., 509-510 (July 1964).

t J.

664

ELECTROMAGNETICS

[CHAP. 14

H plane
Feed point
(a)

Radiating

Field pattern

elements

E plane

(6)

"Coaxial feed line

^^ Maximum

field

s=0.35X
()

12

or typically

(/

FIGURE

14-34

Backward

angle-fire grid array; (a) plan

and (W and

(c) side

views.

2.755)

cos

max

= 2.37 -

(6)

S/A.

As shown in Fig. 14-35, the beam direction 0^ ax (complementary angle to max ) varies
from about 15 to 83 for changes in s from 0.3 to 0.4A. Thus, a wavelength change
of 14 percent (=(0.05/0.35) x 100) swings the beam through nearly 70.
The above

analysis

is

oversimplified, but values calculated as above are in

agreement with measurements.

sion line, the extra path length //2 has the effect of slowing
(left

good

When the antenna structure is regarded as a transmis-

to right in Fig. 14-34) in the ratio s/[s

(//2)].

down the phase velocity

Typically

(/

= 2.75s) the effective

ANTENNAS AND RADIATION

sec. 14-17]

0.30

0.32

0.34

0.36

0.38

0.40

665

0.42

s/X

FIGURE

14-35

Angle of maximum radiation 6' max from backward angle-fire grid antenna as a
function of frequency (expressed in terms of the wavelengths of the mesh spacing
s/\).
A wave velocity equal to c is assumed on the structure.

phase velocity
structure.

only about 0.4c, so that the antenna can be considered as a slow-wave

Since the phase change between successive elements (such as 1 and 2) is less
is

than 360 {typically around 300

(360/A)[(//2)

s]},

radiation occurs at a backward

angle (toward feed).

14-17

FREQUENCY-INDEPENDENT ANTENNAS

frequency-independent antenna may be defined as one for which the impedance and
pattern (and hence directivity) remain constant as a function of the frequency.

Consider the arrangement in Fig. 14-36, which consists of an adjustable A/2

made with two drum-type pocket rulers. If L is adjusted to approximately A/2

dipole

Sliding contact

FIGURE

14-36

frequency-independent antenna

made

of adjustable roll-up pocket rulers set

so that L equals A/2 at the operating
frequency.

Drum-type
pocket rulers'

Transmission

line

666

ELECTROMAGNETICS

[CHAP. 14

at the frequency of operation, the

impedance and pattern remain the same.

w should also be adjusted, but this

speaking, the element thickness or width

small and for

many purposes may be

negligible.

This simple antenna

requirement that the antenna should expand or contract

in

order to be frequency-independent, or

if

Strictly
effect is

illustrates the

proportion to the wavelength

in

the antenna structure

is

not mechanically

adjustable as above, the size of the active or radiating region should be proportional to

the wavelength.

Rumseyf proposed that an antenna which is defined only in terms of angles

would be frequency-independent. % This requirement can be met with a variety of
physical structures such as a pair of concentric equiangular spirals.

The

log-periodic antenna

is

another example of a frequency-independent

antenna, and one employing dipoles^f

is

shown

in Fig. 14-37.

increase along the antenna so that the included angle a

and spacings of adjacent elements are scaled so

is

The

dipole lengths

a constant, and the lengths

that

(1)

where &

is

At a wavelength near the middle of the operating

range, radia-

from the central region of the antenna, as suggested

in Fig. 14-37.

a constant.

tion occurs primarily

The elements

in this active region are

by the arrows) below the elements.

element 8 arrives at element

7,

about A/2 long, and the phasing

By

the phase of 7 has advanced 90

When

is

(as indicated

the time the field radiated to the

and

left

its field

from

adds

in

from elements 8 and 7 arrive at

element 6, its phase has advanced so that the fields from all three elements add in
phase, producing a large resultant field radiating to the left, as suggested at the bottom
of Fig. 14-37. Examining the field radiated to the right from element 6, we find that
it combines with the fields from elements 7 and 8 to produce only a small resultant
phase to that from element

field radiating to

direction

ments

9,

the right.

8.

Thus,

these fields

maximum radiation

(they present a large impedance to the line).

and

1 1

is

to the left or in the

backward

from the wave entering the antenna on the transmission, or feed, line. Ele10, and 11 are in the neighborhood of 1A long and carry only small currents

mean

that the antenna

t V.

is

The

small currents in elements

effectively truncated at the right

9,

10,

of the active region.

H. Rumsey, "Frequency Independent Antennas," Academic Press

Inc.,

New York,

1966.
%

Such antennas possess a wide instantaneous bandwidth, whereas the antenna of

Fig. 14-36 has a narrow bandwidth but is tunable.

So called because the logarithm of (1) yields a constant logarithmic increment for
each step or period in the structure.

H D. E.

Isbell,

IRE

Trans. Antennas Propag.,

AP-8: 260-267 (1960).

ANTENNAS AND RADIATION

sec. 14-17]

Inactive region (I

667

> X/2)

Inactive

(transmission line)
region (2< X/2)

Radiated
fields
*-

-Resultant

FIGURE

(active region)

= ll,a

Resultant

~0

= 30,fc = 1.25, D = 6, BW:3 to 1

14-37

Log-periodic frequency-independent antenna of 8 dBi gain with 1 1 dipoles showing

active (central) region and inactive regions (left and right ends).

Any

small fields from elements

backward

directions.

9, 10, and 1 1 also tend to cancel in both forward and

However, some radiation may occur broadside since the

currents are approximately in phase. The elements at the left (1, 2, 3, etc.) are less
than A/2 long and present a large capacitative reactance to the line. Hence, currents
in these elements are small and radiation is small.
Thus, under the condition assumed

above that the wavelength

portion of the antenna
the wavelength
is

is

is

is

near the center of the operating range, only the middle

active, while the left

increased, the active region

decreased, the active region

moves to the

quency only a fraction of the antenna

and right ends are relatively inactive. If

moves to the right, and if the wavelength

left.

Thus,

we

see that at

any given

fre-

used (where the dipoles are about A/2 long).

At the short-wavelength limit of the bandwidth only 15 percent of the length may
be used, while at the long wavelength limit a larger fraction is used but still less than

50 percent.

k=

is

For the log-periodic antenna of

1.25, the directivity is

Fig. 14-37 the included angle a =

about 6 (=8 dB), and the bandwidth about 3 to 1.

30,

668

ELECTROMAGNETICS

RECIPROCITY

14-18

By

[CHAP. 14

reciprocity

it

follows that the pattern, directivity, aperture, and terminal impedance

of an antenna are the same transmitting or

distribution

on an antenna

To show

is

In general, however, the current

receiving.

not the same for transmission as for reception.

that reciprocity applies to antennas consider the

(Fig. 14-38a) with

media everywhere that are

two antennas
and isotropic.

and 2

Let a
zero-impedance transmitter of frequency /be connected to the terminals of antenna 1
(Fig. 14-38cr), producing a current I and inducing a voltage V
y
2l at the (open) terminals
of antenna 2. Now let the transmitter be transferred to the terminals of antenna 2
(Fig. 14-386),

linear, passive,

producing a current I2 and inducing a voltage

V12

at the

open terminals

of antenna

1.

section, the

antenna arrangement of Fig. 14-38a and b can be replaced by the net-

work of

Since any four-terminal network can be reduced to an equivalent

Fig. 14-38c.

Since reciprocity

is

easily

demonstrated for

this

network,

it

follows that

(i)

and that reciprocity applies to antennas.

In words, the reciprocity theorem for antennas states that if a current Iy at the terminals of antenna 1 induces a voltage V21 at the
{open) terminals of antenna 2 and a current I2 at the terminals of antenna 2 induces a
voltage

V12

at the (open) terminals of antenna

1,

then if I2

/,,

V12 = K21

Antenna 2

Antenna 2

Antenna

In Fig.

Antenna

Zn

^22
1

Terminals of J

Zm

Terminalsof

antenna 2

(c)

FIGURE
(a)

and

14-38

(b) Reciprocity

between two antennas,

(c)

Equivalent

circuit.

ANTENNAS AND RADIATION

sec. 14-18]

669

Receiver

FIGURE

14-39

measurement on observation
illustrating reciprocity for antenna

Pattern
circle

patterns.

14-38c Z, i

and

is

the self-impedance of antenna

Zm is the

it

If all

media involved are

Z22 is
1

'

(2)

linear, passive,

and

isotropic, reciprocity holds

and

follows that the transmitting and receiving patterns of an antenna are identical.

Referring to Fig. 14-39,

observing

having

its field

at

field at the

its

let

the pattern of antenna

with a receiving antenna

center.f

observation

The voltage

(transmitting) be measured

at points

at the terminals of

If the

circle.

around an observation
antenna

measurement procedure

changing transmitter and receiver so that antenna

it

the self-impedance of antenna 2,

and 2 as given by

Vu

Zm ^12
Pattern

mutual impedance of antennas

B is a measure of the
is

reversed by inter-

B transmits and antenna A receives,

follows from reciprocity that the pattern of antenna A, observed by

B, will be identical to that obtained

Directivity
antenna.

and aperture

when antenna A

Directivity

In general, the directivity

is

D=

was discussed

is

PJ0,<j>) is the

in Sec. 14-5 for a transmitting

An
(3)

normalized power pattern.

The same result

its

moving antenna

transmitting.

equal to

\\pj9,<i>)da

where

by

circle

center point.

is

obtained by keeping antenna

B fixed and rotating antenna A about

670

ELECTROMAGNETICS

Thus,

[CHAP. 14

As shown

depends only on the shape of the power pattern.

in the

preceding subsection, an antenna has the same pattern for both transmission and

Hence, the directivity determined with an antenna's transmitting pattern

reception.
is

identical with the directivity

determined with the antenna's receiving pattern.

Accordingly, the term directivity can be applied to both transmitting and receiving
antennas, the directivity of an antenna being the same for both situations.

According to (14-5-29), the

directivity

Ae

effective aperture

of an antenna

Hence, the term

of the antenna times a constant.

applied to both transmitting and receiving antennas,

effective aperture of

a transmitting antenna

the

is

equal to the

may be

being understood that the

it

same

is

effective aperture

as

its effective

aperture

when

receiving.

Impedance When an antenna is transmitting, it may be excited at only one point.

However, when used for reception, the antenna is excited over its entire extent by the
received wave.f As a consequence the current distribution on the antenna is, in
general, not the

same for transmission and

behaves as the same circuit regardless of the

of an antenna

is

the

same

reception.

for transmission

transfer

when

the antenna

is

is

Zr

so that the impedance

This means that

and reception.

terminal impedance of a transmitting antenna

maximum power
Z%

However, the antenna always

mode of excitation,

if

the

the load impedance required for

receiving

is

equal to

its

complex con-

jugate

That T must be the same for transmission and reception can also be seen by
considering a circuit or network of many meshes. Although the currents in the
network are dependent on the location or locations of the emfs, the

circuit

impedances

are independent of the distribution of the emfs.

14-19

AND MUTUAL IMPEDANCE

AND ARRAYS OF DIPOLES
SELF

The impedance presented by an antenna

two-terminal network.
replaced by

its

This

is

to a transmission line can be represented by a

illustrated in Fig.

14-40, in

which the antenna

is

equivalent impedance connected to the transmission-line terminals.

This impedance to which the transmission line connects

driving-point impedance

ZT

If the

antenna

is

isolated,

i.e.

is

called the terminal or

remote from other objects,

and is lossless, its terminal impedance is the same as the self-impedance of the antenna
(Z r = Z n ). This impedance has a real part equal to the radiation resistance R IX and
t Furthermore, the manner in which the receiving antenna
direction of the incident wave.

is

excited depends

on the

ANTENNAS AND RADIATION

SEC. 14-19]

671

..-Antenna
--Terminals

Transmission
line

FIGURE

14-40

Transmission line with antenna and with

equivalent

an imaginary part

lumped (terminal) impedance

called the self-reactance

Xn

^Equivalent

|^f

If there are

impedance

nearby objects, such as

other active antennas, the terminal impedance will be modified to include contributions
involving the mutual impedance between the antennas and currents flowing in them.

In Sec. 14-9 the radiation resistance of a thin A/2 dipole was calculated and found
to be equal to 73 Q.

The method used involved a

calculation of the (real)

radiated by integrating the Poynting vector over a large sphere (in far field)

reactance term resulted from this calculation.

To

power

No

include the antenna reactance

requires integration (in effect) of the Poynting vector over a closed surface collapsed

down

so that

radiating)

it

just

power

fits

over the antenna conductors.

circulating

tion, and, in general, this

In this

way

the reactive (non-

back and forth near the antenna enters into the integrawill result in a reactance term.
Such calculations are

tedious and will be omitted here.f

a thin linear A/2 center-fed dipole

As
is

stated following (14-9-7) the self-impedance of

given by

Z = R n +jXu =

73 +J42.5

(1)

The terminal impedance of a thin linear center-fed dipole antenna as a function

of the frequency is shown by the spiral solid line in the impedance diagram of Fig.
14-41.J The length-to-diameter ratio (l/d ratio) for this antenna (solid curve) is
With change in frequency the impedance varies over a wide range of values.
2,000.
At frequencies where //A 0.5 the resistance is small, and the reactance changes
rapidly from large negative to large positive values with increase in frequency (increase in 1/1).
At resonance //A = 0.48 and Zn = 70 +j0 il. At higher frequencies
where //A x 1 the resistance is several thousand ohms, and the reactance may also be of
comparable magnitude. Another (second) resonance occurs when //A = 0.93, and for
this condition Z n = 3,300 +j0 2.
At still higher frequencies the impedance values
again become small, and another (third) resonance occurs when //A = 1.48. For this
t

For a detailed discussion

see, for

example, Kraus, "Antennas," chaps. 9 and 10.

Admittance Diagrams for Antennas and the Relation between Antenna

Theories, Harv. Univ., Cruft 1Mb. Tech. Rep. 46, June, 1, 1948.

t Erik Hallen,

672 ELECTROMAGNETICS

[CHAP. 14

-50

FIGURE

14-41

Self-impedance diagram (R,X) as a function of frequency (expressed in //A ratios

indicated along curves) for linear antennas with length-diameter {l/d) ratios of
60 (dashed) and 2,000 (solid). The diagram at the left is an enlargement of the
region of

first

and

third resonances near the origin of the

main diagram.

{After

Hallen.)

Zn =

+j0

With further increase in frequency the antenna impedAt the odd-numbered resonances the resistance is small, and at the even-numbered it is high. If the antenna thickness is increased
(l/d ratio smaller), the impedance variation with frequency is smaller.
This is illustrated by the dashed spiral curve in Fig. 14-41, which is for a center-fed dipole with l/d
ratio of 60.
The resistance values of this dipole at its first and third resonances are
not much different from those for the thinner dipole but the resistance values at the
second and fourth resonances are greatly reduced. As the antenna thickness is increased further, the impedance variation becomes still less.

resonance

100

fi.

ance goes through further resonances.

simple explanation for this tendency follows from a consideration of the

V-type antenna shown in Fig. 14-42.

If a

two-conductor transmission

line is gradually

ANTENNAS AND RADIATION

sec. 14-19]

FIGURE

673

14-42

Two-conductor transmission line opened

By keeping ratio
constant the arrangement presents
a constant impedance
over a wide
out into antenna.

Djd

frequency range.

ratio of conductor spacing D to diameter d kept constant, the

impedance of the line will also be constant. Hence, there will be no
wave on the line provided none occurs at the end, and this will be essentially

opened out and the

characteristic
reflected

the case

if

is

sufficiently large

(D >

end of the structure.

A) at the

In the opened-out

region of the thick conductors the transmission line acts as an antenna, and most

of the power

impedance

is

will

this value will

radiated in a beam to the right. Accordingly, the antenna terminal

be equal to the characteristic resistance of the transmission line, and
remain substantially constant at all frequencies for which the above

conditions hold.

Thus, for an antenna like that in Fig. 14-42, the impedance spiral

reduces to a single point representing the characteristic impedance

the opened-out transmission

Let us

now

is

Ez = current

and

antenna
IX ,I2

given by

-z -z

z
,

resistance) of

consider the mutual impedance of the two antennas of Fig. 14-43.

The mutual impedance

where Iz

(all

line.

= terminal

field at

~ Kzi

"Y LEdz

M*2

(2)

distance z along antenna 2 induced

by current

in

currents of antennas

and

2, respectively

J
FIGURE
Two

14-43

parallel linear A/2 center-fed dipole

antennas with A/2 spacing.

IF

axis

674

ELECTROMAGNETICS

[CHAP. 14

20 -10

FIGURE

10

20

30

40

60

50

70

800

14-44

Mutual-impedance diagram (R m X) as a function of element spacing s/X for

two parallel A/2 center-fed dipoles with length-diameter (l/d) ratios of 73 (dashed)
and 1 1 ,000 (solid). (After Tai and Carter.)
,

The

integration

omitted here.

is

over the length

of antenna

However, for the case of two

2.

The evaluation of

(2) will

parallel thin linear X\1 center-fed

be

antennas

spaced A/2, as in Fig. 14-43,

Zm = Z12 =Z 21 = -13-;29fi
The mutual impedance of two parallel
is shown by the spiral curves of Fig.

spacing

antenna elements with an l/d ratio of 11,000.

ratio of 73.

As

(3)

linear A/2 antennas as a function

14-44.f

The

The dashed

of

solid spiral curve is for

spiral

curve

is

for an l/d

the spacing s between the antennas increases, the mutual impedance

the self-resistance of an antenna

is

(i? m =
m = 0). It is to be noted that whereas
always positive, the mutual resistance may be either

rapidly spirals in around the origin

positive or negative.

The mutual impedance Zm of two antennas is usually a complicated function of

and orientation of the antennas. However, in the case of two
short-dipole antennas the relation is relatively simple and illustrates clearly the effect
the size, separation,

C-T

Tai, Coupled Antennas, Proc. IRE, 36: 487-500 (April 1948); P. S. Carter,
and Applications to Antenna Problems, ibid.,

Circuit Relations in Radiating Systems

20: 1004-1041 (June 1932).

ANTENNAS AND RADIATION

sec. 14-19]

675

Dipole

FIGURE
Two
tion

14-45

short dipoles of length

j,

/,

separa-

and orientation angles 6 and

9' for

mutual-impedance equation.

of these parameters.
dipoles

is

Referring to Fig. 14-45, the mutual impedance of the two

given by
607t/

(sin

sin 0')(sin

ps

sX

Note that the

first

factor

is

+j cos /fa)

(4)

a magnitude factor, the second an orientation factor, and

the third Aperiodic, or complex, factor with real and imaginary parts.

when

the separation distance s

imaginary.

and

Both dipoles are

now

in

and n is odd,
the same plane and

Zm is real;

but

if

assumed that

Note
n

is

also that

even,

Zm is

X and

I <$s
of uniform magnitude and constant phase on each dipole.
an array of two identical parallel thin linear center-fed dipole

also that the current

Consider

= nX/4

it is

/ -4

is

antenna elements to be driven with equal in-phase currents and fed by a two-conductor
300-2 transmission line. This can be accomplished by the arrangement of Fig. 14-46.
The voltage at the terminals of element 1 is

V I ZU + I2 Z12
1

FIGURE

(5)

14-46

Broadside array of two thin A/2 centerfed dipoles with A/2 spacing showing
field patterns in

both principal planes.

^V-_-^E
E plane
Side

n <>

Driving point
lor array

Front

676

ELECTROMAGNETICS

[CHAP. 14

= terminal current in element 1

= terminal current in element 2
Z u = self impedance of element 1
Z12 = mutual impedance between elements

where Ix

I2

Likewise,

we have

for the terminal voltage of element 2

V2 =I2 Z22 + I Z21

(6)
Z22 is the self-impedance of element 2. We require h= I2 and we also know
that Z u = Z22 and Z 12 = Z21
Therefore,
V =h(Zll +Z12 )=V2
(7)
and the terminal impedance Z of element and Z2 of element 2 are equal and given by
l

where

Zi

For the case where

A/2

and

= Zn + Z

M
= A/2

12

= Z2

(8)

Zn = 73 +J43 CI
Z12 = -13-y29n

(9)

(10)

Therefore,

Z2 = Z n + Z12 = 73-13
If each element

shortened

is

+y(43

29)

60 +./14

the self-reactance (43

slightly,

CI)

and terminal

57

+y0

(12)

In order that the array will present a resistive load of 300

CI at the

in Fig. 14-46) the transmission-line sections

terminals can be constructed so that their length a

impedance

Z = ^57

x 600

reduced more

In so doing the

resistance will be reduced, so that

Z =
array (point

(1 1)

may be

than the mutual reactance so as to yield a total reactance of zero.

self-resistance

185

CI.

driving point of the

between P and the element

1/4 and their characteristic

Thus, each A/4 section transforms the element

terminal resistance of 57 CI to 600 CI and the two lines in parallel yield a driving-point

impedance

Z = 300 +j0 CI.

The above array

consisting of

equal in-phase currents has

terns are as

shown

two

maximum

in Fig. 14-46

E9 {&) = 2fc/

H plane

Ee

(14-9-4), while the

antenna elements fed with

The

principal-plane pat-

and are given by:

Eplane:

The shape of the

parallel thin A/2

radiation broadside.

cosr(7t/2)
u ;

(<t>)

is-plane pattern

=
is

cos 01J
-

(13)

n cos &

2klt cos

(14)

the same as for a A/2 linear antenna, as given by

shape of the //-plane pattern

is

the same as for two equal isotropic

in-phase point sources spaced A/2 apart, as given by (14-6.1-2) with </= A/2.

ANTENNAS AND RADIATION

SEC. 14-19]

The

of the above array can be calculated from (14-5-7) by integrating

directivity

the power pattern over

now

677

4?t sr.

An alternative method of calculating the directivity will

be described which employs the

and mutual

self-

resistances of the antenna ele-

ments.

Assuming no heat

losses, the

power input to the array

P = 2/

2
1

(* 11

+ R i2

(Fig. 14-46)

is

given by
(15)

Hence, the individual element currents are

(16)

''-''-AteTTJQ
same power

If the

supplied to a single isolated A/2 antenna,

is

its

current

is

In the

(i7)

H plane the pattern of the single A/2 antenna

EHVI

The maximum

(<l>)

kI

is

nondirectional and

J~

radiation from the array occurs in the direction

Now R n =

73

and

R 12 = - 13

Array gain
This gain

Q.
/
/

given byf

=k y

(18)

<f>

90 (Fig. 14-46).

Hence, the gain of the array over the single A/2 antenna as reference
ratio of (14) to (18) with 4> = 90, or

Array gain

is

2kl
- t
kl

is

given by the

2R ly

/
'

(19)

Ru + R l2

"V

Therefore,

2 x 73

1.56

(=1.9 dB)

13

with respect to a A/2 antenna (with same power input) as reference. The
of a linear A/2 (dipole) antenna has been calculated earlier as 1.64 (=gain
of A/2 antenna over isotropic source). Therefore, the directivity of the array of Fig.
is

directivity

14-46

is

D=

1.56

x 1.64

= 2.56

(=4.1 dB)J

same directivity that would be obtained by integrating the power pattern of

An sr, yet we have carried out no such integration, at least explicitly.
However, such an integration or its equivalent is implicit in the self- and mutual

This

is

the

the array over

resistances.

The

When

subscript

HW on E in (18) stands for half wavelength antenna.

expressing the gain over an isotropic source (or directivity) in decibels,

be given in decibels over isotropic (dBi), or in this case 4.1 dBi.

it

can

678

ELECTROMAGNETICS

[CHAP. 14

conducting sheet

Infinite

Field pattern

Finite conducting

sheet

Thin linear element

H plane

Image

Linear antenna
Linear antenna

Linear antenna

(driven element)

(driven element)

(driven element)
i

(c)

(6)

(a)

Incoming wave

Passive corner
Active corner reflector

reflector

'

Linear antenna

Driven element

(driven element)

Reflected

wave

(e)

(d)

FIGURE

14-47

Linear antenna (a) with infinite reflecting sheet and (6) with sheet of finite extent.
In (c) the reflecting sheet has degenerated to a thin linear element, (d) Corner
reflector

element).

14-20

with (active) driven element and (e) passive corner reflector (no driven
(/) Driven element and parabolic reflector.

REFLECTOR AND LENS ANTENNAS

In this section a variety of reflector antennas are considered.

is

also discussed.

It is

The

lens-type antenna

convenient to treat both reflector and lens types together

since the design of both involves the principle of equality of path length.

Figure 14-47 displays six types of reflector antennas.

is

an

infinite flat sheet,

but in

(b) the sheet is

be analyzed by simple image theory, but the

tion theory.
this

In Fig. 14-47a the reflector

The infinite flat sheet can

one requires the addition of diffrac-

of finite extent.
finite

In Fig. 14-47c the reflector has degenerated to a thin linear element, and

arrangement

is

analyzed by simple array and circuit theory.

antennas are shown in Fig. 14-47rfand

e.

In one case the antenna

driven element), and in the other case the corner reflector

reflector

antenna

is

shown

in Fig. 14-47/".

is

Corner
is

reflector

active (contains

passive.

parabolic

ANTENNAS AND RADIATION

sec. 14-20]

Direct-ray

shadow boundary

679

Reflected-ray

shadow boundary
Total power pattern

Direct, reflected,

Direct and

<

and diffracted rays

diffracted rays

FIGURE

14-48

(a) Linear element with finite reflecting sheet analyzed

(b)

power

pattern.

Note

that this pattern

is

by diffraction theory and

on a logarithmic

(decibel) scale in

order to cover the wide range in power.

14-20.1

Infinite-flat-sheet Reflector

In Fig. 14-47 a linear antenna

is

situated a distance

conducting metal sheet, or ground plane.

d from an

infinite flat perfectly

This combination can be analyzed simply

by replacing the ground plane by an image of the antenna at a distance 2d from the
antenna (see Fig. 7-16). The pattern in the 6 plane (H plane)t is that of two equal
isotropic point sources spaced a distance 2d apart and in opposite phase, or
Ej,(9)

If

d=

sin (fid cos 0)

A/4, the pattern is as suggested in Fig. 14-47a

14-20.2

with

(1)

maximum

at

= 0.

Finite-flat-sheet Reflector

When the reflecting sheet or ground plane is reduced in size, the analysis is less simple.
The
(1)

situation

is

illustrated in Fig. 14-48a.

There are three principal angular regions:

region of direct and reflected rays (in front of sheet or to right side); (2) region of
t In the

by

(1).

E plane the

pattern

is

that of an isolated antenna in this plane multiplied

680

ELECTROMAGNETICS

direct but

no

reflected ray

If the
is

[CHAP. 14

reflected ray (contains plane

behind sheet or to

left; i.e.,

of sheet)

(3)

shadow region (no

direct or

only diffracted rays are present).

ground plane

is extensive (say several wavelengths in size) and the antenna

ground plane, image theory accounts adequately for the pattern in
In region 2, the distant field is dominated by the direct ray from the

close to the

region

1.

antenna.

In the shadow region behind the sheet (region 3) the geometrical theory

of diffraction (Sec. 12-9) must be used.

two cardioid-pattern sources

The

pattern in this region

(see Fig. 14- 19c),

one

is

effectively that

at each edge of the sheet.

of

The

beam maximum of the upper-edge pattern is directed upward in the figure, and
beam maximum of the lower edge is directed downward. Calculating the fields

the
for

the three regions as outlined above results in the E^ff) (#-plane) pattern

shown in Fig.
1 4-486 for the case where the sheet dimension D = 1 0A and the antenna spacing d = 1/4.
It is assumed that the antenna is a linear element and that the sheet is infinite (or very
large) perpendicular to the page.
Note that this pattern is on a logarithmic (decibel)
scale in order to show the wide range of fields involved.
The pattern in Fig. 14-48* is
similar to that in Fig. 12-22 except that the power pattern [P (dB) = 20 log| E\ ] is in
polar coordinates here and that the field pattern \E\ is in rectangular coordinates in
Fig. 12-22.
In the shadow region the power is down 40 to 50 dB.

Thin Reflectors and Directors

14-20.3

When

the reflecting sheet degenerates into a thin linear conductor, as in Fig. 14-47c,

the arrangement can be analyzed by circuit and array theory as follows. f

active (driven) element be designated element

element be designated element

2, as in Fig.

and the passive (degenerate

14-49a.

The

Let the
flat

sheet)

circuit relations for this

arrangement are

V = hZ n + I2 Z 12
= 1%Z22 A^21
x

(1)

"I"

F = terminal voltage of element 1

/j = terminal current of element 1
I2 = terminal current of element 2
Zn = self-impedance of element 1
Z22 = self-impedance of element 2 (if it were opened at center and

where

measured in

isolated situation)

Zi2 = Z21 = mutual

t G.

impedance of elements

and 2

H. Brown, Directional Antennas, Proc. IRE, 25: 78-145 (January

1937).

sec. 14-20]

ANTENNAS AND RADIATION

681

Driven
Driven element

element

Reflector

Director

(c)

Maximum >

3 radiation

(a)

\
2

(d)

(6)

FIGURE
and

14-49

Driven element with single parasitic element and

element with reflector and director.
(a)

From

(6)

(d) driven

(1)

h=
The

and

(c)

pattern

field

^- -I
^22

-I

\Z l2
|Z 22

Ltan- 1 (Xl2 /R 12 )
Ltan- 1 (X22 /R 22 )

E9 (4>) (H plane) is then

Ee = kil, + I2 /Pd cos
{(j>)

where I2

as given in (2)

is

and k is a constant.

(2)

(3)

<j> )

In general,

and I2

/j

will

be of different

magnitude, so that the pattern cannot be reduced to the simple forms used in Sec.
14-6.1.
Also note that, in general, Z22 # Z lt
Suppose that the driven element 1 is
.

approximately A/2 long

sarily exactly the

(/,

same.

Xjl)

and that

l2 is

about the same length but not neces-

Referring to Fig. 14-41,

linear element varies rapidly as

we note

positive to negative reactance values as the length

is more than
somewhat shorter (say

element 2

Et

pattern

{<j>)

kjl

> 0.5),
< 0.48), it

{l2 jk

l 2 /X

shows that

as above

element 2 acts as a reflector and

(0

4>

= 0,

reactance

is

Fig. 14-49

negative, the

2 acts as a director.

An

b).

it

l/X

in the

going from

0.5,

Thus,

reduced.

the length of

if

has a positive self-reactance; while

has a negative reactance.

if it is

Calculating the

former case (element 2 reactance positive)

maximum

On

is

radiation from the array

the other hand, if

maximum radiation

is

to the

l2 is

left (0

is

to the right

shorter, so that

(j>

180)

its self-

and element

assumed that the interelement spacing is small, say 0. 1 to 0.2/1.

is to add a third (passive)
14-49c and d. Let element 2 be made inductively reactive (X22
It is

extension of the arrangement of Fig. 14-49a and b

element as in Fig.
positive)

and

that the reactance of a thin

a function of frequency when

and element

acts as a reflector

3 capacitively reactive (^33 negative).

and element

3 as a director.

dimensions are as follows: driven-element length

For thin
/j

In this

way element 2

linear elements typical

= 0.48A

(to

make

Xn = 0),

682

ELECTROMAGNETICS

[CHAP. 14

reflector-element length

d12 d13

spacings

= 0.50A, director-element length l3 = 0.46A, and interelement

The gain of such an array is about 5 ( = 7 dB) as comA/2 dipole and the directivity D = 5 x 1.64 = 8.2 ( = 9 dB =

l2

0.1 A.f

pared with a reference

gain over isotropic source).

maximum

14-49a and b the

(D =

With a

single reflector (or director) element as in Fig.

possible gain

is

about 5

dB

with respect to a A/2 dipole

7 dB).

In arrays like those discussed above, where one element

are passive,

customary to

it is

driven and the others

is

refer to the passive elements as parasitic elements.

Thus, the above reflector and director elements are parasitic elements.

Arrays

may

be constructed with larger numbers of parasitic elements.

For example, YagiJ has

director elements, and arrays with a number of para-

built arrays with

sitic

elements are

to as Yagi or Yagi-Uda antennas.

Corner Reflectors

14-20.4

many parasitic
commonly referred

corner reflector consists of two

Fig. 14-41d.

When

flat reflecting sheets intersecting at

the corner angle

square-corner reflector.

is

90, the sheets

With a driven element placed

effective (active) directional

meet

is

an

effective (passive)

wave

in

forming a

as shown, the arrangement

antenna for a wide range of corner angles (0

Without a driven element a square corner

an angle, as

at right angles,

<9 <

is

an

180). If

reflector over a

wide range of angles of incidence (0 <

< 45). The reflected wave is directed
along the path direction of the incoming wave, as illustrated in Fig. 14-47e\
f

The active corner reflector (with driven element) can be readily analyzed using
method of images for corner angles equal to 180/", where n = 1, 2, 3,
It will
be convenient to illustrate the method for a square corner, as in Fig. 14-50. For this
case there are three images, as indicated. The driven element and all images carry
equal currents. Driven element 1 and image element 4 are in phase, while image
elements 2 and 3 are in opposite phase (with respect to 1 and 4). The field pattern

the

Ejffi)

(H plane)

is

given by
E^ff)

where s

is

direction

2klx

cos(fis cos 9)

the corner-to-driven-element spacing.

= 0

(along corner bisector) for

all

cos(0s sin 6)

The maximum

values of s

(1)

radiation

is

in the

< A.

The exact lengths are a weak function of the element thickness. Thus, compare
solid and dashed curves in the RX diagrams of Figs. 14-41 and 14-44.

H. Yagi,

Beam Transmission of

Ultra-short Waves, Proc. IRE, 16: 715-740 (June

1928).
S. Uda and Y. Mushiake, "Yagi-Uda Antenna," Research
Communication, Tohoku University, Sendai, Japan, 1954.
IT

J.

Institute

of Electrical

D. Kraus, The Corner Reflector Antenna, Proc. IRE, 28: 513-519 (November
J. D. Kraus, U.S. Patent 2',270,314, Jan. 20, 1942.

1940);

ANTENNAS AND RADIATION

sec. 14-20]

683

Corner angle bisector

(a)

Plane of driven
element

*
'Aperture width

(6)

FIGURE

14-50

Square-comer reflector with driven element analyzed as driven element with

three image elements, (b) Path lengths in corner reflector.

(a)

The terminal voltage

for the driven element

V = 7,Z U + AZ 14 x

is

2/,Z 12

= self-impedance of element (driven element)

Z 12 = mutual impedance between elements 1 and 2
Z14 = mutual impedance between elements and 4

where Zj l

(2)

or between

and

Similar expressions can be written for the voltage at the terminals of each of the image
elements.

Then

for a

power P to the driven element (power to each image

is

also

P)t

we have
I,

=
^11

^14

(3)

2/?i2

t But the actual total power supplied to the corner-reflector antenna

to the driven element (=P).

is

that supplied

684

ELECTROMAGNETICS

The pattern

[CHAP. 14

for the isolated driven element (no reflector present)

t(0)
Substituting (3) in (1)

and taking the ratio of (1) to

(driven) element (in direction 9

is

Rn

that

if

gain over the reference

(cos ps -1)

(5)

A/2 long, the image elements are also A/2 long

and for a

0.25A the gain of the square corner-

antenna over a A/2 dipole antenna

The above

(4) yields the

SRu+Ru-

typical driven-element-to-corner spacing s

reflector

(4)

0) as

Gain over reference

If the driven linear element

is

10.2

is

dB

(directivity

analysis assumes reflecting sheets of infinite size.

the side length (in plane of page, Fig. 14-50a)

element-to-corner spacing, the reduction in gain

is

D=

In practice

12.35 dB).
it is

found

at least twice the driven-

is

The

small.

reflector length per-

pendicular to the page should equal or somewhat exceed that of the driven element.!

The corner reflector is very widely used in many television,

and radio-astronomy applications.

point-to-point

communica-

tion,

14-20.5

Parabolic Reflectors and Lens Antennas

Parabolic reflectors and lens antennas

may

be considered together since the design

of both involves the principle of equality of path length.

.Consider, for example, the

We wish to convert the circular wavefront

from the driven element or source at the point O (focus) to a plane wavefront. f From
parabolic reflector

shown

in Fig. 14-51a.

geometric optics the Principle of Equality of Path Length requires that the electrical
length along all ray paths between a pair of wavefronts be equal. This means here
that the electrical distance

OAO = OBC or that

R cos 0,

from whichlf

1L

R=

2L = R +

(2)

cos 6

For a detailed discussion of corner-reflector parameters see Kraus, "Antennas,"

pp. 328-336.

wavefront

is

an equiphase surface;

i.e.,

the phase

is

the

same

at all points

on the

surface.

For more detailed discussion see, for example, S.

Theory and Design," McGraw-Hill Book Company,

f More

explicitly

we should

" Microwave Antenna

York, 1949.

Silver,

New

write

2L +

i=R+Rcos0 + i

(1)

where the distances are expressed in wavelengths. Waves reflected at the parabolic
surface undergo a phase reversal (180 phase change) which is equivalent to a path
length of A/2. Hence the terms with 1/2 in (1).

sec. 14-20]

Parabola

ANTENNAS AND RADIATION

=e

everywhere

.< c

Hyperbola

in dielectric

v>c

Ellipse

in

685

metal

lens

Driven element
or source

Driven element
or source

\
Plane wavefront
Plane
wavefront
-

Circular
Circular wavefront

wavefront

Focal plane

\
Plane wavefront
Parabolic reflector

Dielectric lens

(a)

Metal lens

(6)

FIGURE

(c)

14-51

(a) Parabolic-reflector, (A) dielectric-lens,

and

(c)

metal-lens antennas with path

lengths used in analysis.

This

O.

the equation of the required surface.

It is a parabola with focus at O.
Consider next the dielectric lens of Fig. 14-516 with driven element or source at
The principle of equality of path length requires that OAB = OCD or that R =
is

-L) +L, from

>l(R cos

which

Of

R=
r\

where

r\

is

the index of refraction

{r\

l)L

cos 9

= Jer

(3)
1

for nonmagnetic media).

gives the required shape of the curved lens surface.

We
velocity v

note that with the parabolic reflector,

c.

For the

dielectric lens,

however,

r\

Equation

(3)

It is

a hyperbola.

r\

everywhere and the wave

>

in the lens,

and hence

<

c.

Therefore, the electrical distance, or phase change, is greater per unit distance
in the
lens than in the air outside, and this is accounted for in (3).
general expression for
the electrical length is

Electrical length

fi

[r\

dl

\r\dl

(4)

where A
length.
r\

[CHAP. 14

ELECTROMAGNETICS

686

y/7r

the free-space wavelength

is

In (2) and

(3)

we have

integration

is

over the physical path

applied (4) in incremental form with

The

in the dielectric lens.

and the

integral

form of (4)

is

r\

required only

if

in air
r\

is

and

a con-

tinuously variable function of distance.

The lens consists of a

somewhat more than 2/2 and hence capable of

Finally, let us analyze the metal-plate lens of Fig. 14-5 1 c.

stack of

parallel metal sheets spaced

transmitting a higher-mode

For

Sec. 13-14).f

R + (L -

jR

cos

wave (E perpendicular to page), as in a waveguide (see

we must have OAB = OCD or

equality of electrical path length

= L, from

&)r\

which

*=

-")L

(1

(5)

r\

cos 6

However, in the metal-plate

to keep both numerator
in
order
is
used
form
of
(5)
lens v> c and r\ < 1, and so the
equation of an ellipse.
the
<
is
this
case
(j
1) (5)
and denominator positive. For
metal
plate parallel to the
(and
page
parallel
the
to
Metal-plate lens antennas with E
lens of Fig. 14-51c.
way
as
the
the
same
in
analyzed
page) are also possible. They are
cylindrical geom14-51
both
to
applies
of
Fig.
This discussion of the antennas
This relation

is

identical with (3) for the dielectric lens.

etry (uniform surfaces

and

line

source perpendicular to page) and to spherical

geometry (surfaces given by figure of revolution around axes) with point sources. As
a generalization, all the antennas of Fig. 14-51 may be regarded as wave transformers

which convert a cylindrical or spherical wavefront to a plane wave or vice versa.

14-20.6

The

Some Comments on Corner

Reflectors vs. Parabolic Reflectors

driven element or source for the parabola (and also lens) antennas should be
and for greatest operating efficiency most of its pattern angle should be

directional,

included within the angle subtended by the reflector (or lens) (Fig. 14-51). This
puts an additional requirement on the primary or feed antenna used at the focus of a

This requirement is not present in the corner reflector, where the

a simple linear antenna with omnidirectional pattern (in plane of

parabolic reflector.
driven element

is

page, Fig. 14-50).

should be noted that the principal advantages of the corner reflector are for
When the aperture width (Fig. 14-506) is less than 22, a corner
reflector may be simpler and more practical than a parabolic reflector, but for larger
It

small apertures.

For small apertures the actual shape of the reflector

is better.
becomes of secondary importance since the ray-path differences involved are small.

apertures the parabola

W.

E.

Kock, Metal Lens Antennas, Proc. IRE, 34: 828-836 (November

1946).

sec. 14-21]

ANTENNAS AND RADIATION

It is instructive to

687

apply the principle of equality of path length to the corner

Referring to the square-corner reflector of Fig. 14-506 with driven-element-

reflector.

to-corner spacing of A/2, the path lengths of the four principal rays to a point at an
infinitesimal distance to the right of the plane of the driven element are

= OA
Length (ray 2) = OA + \ = \ + i = 1A
Length (ray 3) = 05+C0 + i-i = i+i + -i=lA
Length (ray 4) = OD + \ = i + i = U
Length (ray

Thus,

14-21

all

1)

four rays are in the same phase.

SLOT

AND COMPLEMENTARY ANTENNAS

If a A/2 slot

is cut in a large flat metal sheet and a transmission line connected to

the
points FF, as in Fig. 14-52a, the arrangement will radiate effectively due to currents
flowing on the sheet. The analysis of such a slot antenna is greatly facilitated by con-

sidering the slot's complementary antenna

.f

ary to the slot of Fig. 14-52a

is

Thus, the antenna which

the dipole of Fig. 14-526.

is

The metal and

complement-

of
According to Booker's theory,t the pattern
identical in shape to that of the dipole of Fig. 14-526
air regions

the slot are interchanged for the dipole.

of the

slot

of Fig. 14-52a

is

except that the electric field will be vertically polarized for the slot and horizontally

Flat conducting
sheet

Current

line

Slot antenna

Transmission

line

Complementary
dipole antenna

Transmission

FIGURE

line

14-52

antenna and (b) complementary dipole antenna.

have a width w. They are fed at the points FF.
(a) Slot

t H.
J.

Both are A/2 long and

G. Booker, Slot Aerials and Their Relation to Complementary Wire

IEE (London),

93, part IIIA, no. 4 (1946).

Aerials,

688

ELECTROMAGNETICS

[CHAP. 14

Further, the terminal impedance

polarized for the dipole.

the terminal impedance

Zt of the dipole

by the

intrinsic

of the

impedance

slot is related to

Z = 376.7Q) of
(

free space as follows:

(1)

Thus, knowing the properties of the dipole enables us to predict the properties of the
complementary slot. For example, let the width of the dipole and slot of Fig. 14-52

be reduced to a very small fraction of a wavelength so that the dipole qualifies as a

thin A/2 linear dipole with Zt = 73 +J42.5 (see Sec. 14-19). According to (1), the
terminal impedance of the complementary slot antenna will be

35 481

'

= 363-;2im

73 +742.5

HORN ANTENNAS

14-22

horn antenna may be regarded as a flared-out (or opened-out) waveguide.

pyramidal horn fed by a rectangular waveguide

of the horn

is

shown

in Fig. 14-53o.

The function

to produce a uniform phase front with a larger aperture than that of the

is

waveguide and hence greater directivity. The principle of equality of path length is
Instead of specifying
applicable to the horn design but with a different emphasis.
phase,
in
this requirement
exactly
mouth
be
horn
plane
the
of
wave
over
the
that the
amount.!
than
a
specified
less
but
by
may
deviate
where
the
phase
one
is relaxed to

From
tan

hfll,

we have

14-536

the geometry of Fig.

cos 6

//(/

<5),

sin 9

h/[2(l

8)],

from which
/

= J-

8 small

(1)

50

_,
= tan _1 - = cos
,

2/

In the
(8

<

E plane of the horn,

it is

A/5) (deviation 36).

say 135 (8

<

E =
t

1+8

(2)

customary to allow no more than 72 phase difference

plane the difference can be larger,
in the

However,

31/8), without undesirable

(boundary condition

effects since

E goes to

zero at the horn edges

satisfied).

path length, where the inequality

t This could be stated as the principle of inequality of

does not exceed a specified amount.

sec. 14-22]

ANTENNAS AND RADIATION

689

Horn mouth

plane

FIGURE
(a)

14-53

Horn antenna showing

field orientation in

mouth and

(b)

dimensions used

in analysis.

EXAMPLE

Determine the length /, width w, and half-angles and

(in E and
planes,
a pyramidal electromagnetic horn as in Fig. 14-53a for which the mouth
10A.
The horn is fed by a rectangular waveguide with TE l0 mode as in the
<j>

respectively) of

height h

figure.

solution

Taking S

= A/5 in the E plane, we have from (1)

(in the E plane) is

that

62.5A and half the flare angle

Taking 8

= 3A/8

in the

= tan- -=tan1

^=

4.6

H plane gives
+

= *-

and the width w of the horn

in the

w=

2/ tan

62.5

H plane
<j>

=2

+ 0.375

= 6r

is

x 62.5 x tan

6.3

= 13.7A

= h z /S8 = 100A/(|) =

[CHAP. 14

ELECTROMAGNETICS

690

APERTURE CONCEPT

14-23

According to (14-5-29), the

tivity

effective aperture

of an antenna

is

related to

its

direc-

D by
471

n=
The

Ae

effective aperture is

antenna

it

may be

(i)

a unique quantity for any antenna.

larger than

its

For a short-dipole
However, for

physical cross section or aperture.

antennas like horns or parabolic reflectors which have large aperture dimensions in
terms of wavelength, the effective aperture is always less than the physical aperture.
Ideally if the field

is

thermal losses, the

these conditions the

The

exactly uniform

effective aperture

maximum

and

Ae

in

is

phase over the aperture and there are no

equal to the physical aperture

Dm

is

4k

An J2

directivity

Ap

Under

achieved, or

(2)

ratio of (1) to (2) gives the aperture efficiency

(3)

uniform

field distribution

has the disadvantage that minor lobes are large

Although minor lobes are

(see Fig. 14-24).

less for

a tapered

field distribution (zero

minor lobes for cosine-squared or gaussian distributions), the effective aperture and
One of the problems of antenna design is to achieve a suitable
directivity are less.
compromise, so that the aperture efficiency is as large as possible for a given side-lobe
An empirical approach to the problem is to postulate a number of aperture
level.
The patfield distributions and to select the best efficiency and side-lobe combination.
can
efficiency
aperture
while
the
(14-8-1),
from
tern with side lobes can be calculated
be evaluated from

(4)

2
)a

E = field at any point in aperture

E = average of E over aperture
2
2
( ) ay = average of E over aperture

where

It is

assumed that

E is exactly in phase over the aperture and that there are no thermal

losses.f
simplified.
t This discussion is highly

" Radio Astronomy," pp. 212-223.

For a more complete treatment

see Kraus,

SEC. 14-23]

ANTENNAS AND RADIATION

EXAMPLE
E

Calculate the aperture efficiency of a rectangular aperture antenna (x by

y

with a uniform

y direction and a linearly tapered field in the x direction going from

x /2) to at both edges (triangular taper as in Fig. 14-246).

field in the

at the centerline (x

We have E(x) = E

solution

[>(*)]..

Hence the aperture

= -L-

[x/(x /2)l

E\x) dx

and E(x)

= -^

= E/2.
x*dx

Also

= %-

efficiency is
CP

691

"

[E{x\tf

[E 2 (.x)]

cosine rather than a linear type of taper

apertures, the field going

from a maximum

is

common

in practice for large

at the center to about 0.3 times the center

value at the edges (down about 10 dB).

The aperture concept is useful in other ways. For example, it can be used to
help decide between a very long end-fire antenna or a broadside array of several
shorter end-fire antennas and, for the latter, to determine what spacing should
be
used between the antennas. This will be illustrated in the following problem.

EXAMPLE

Design a circularly polarized antenna using one or more end-fire elements

(= 20 dB) for operation at a given wavelength A.

so as to produce a directivity of 100

solution From Table 14-4 the highest directivity per unit length for an end-fire array
obtained with the increased-directivity condition as with the helical-beam antenna.
For
this antenna with C = A, the required length is
is

L- B - -6.7A
or about 31 turns for a helix with a pitch angle = 12, as shown in Fig. 14-54^.
A more
compact design results with four shorter helical-beam antennas (quad-helix array). From
(2)
the physical aperture required (with uniform distribution) to produce a directivity

Dividing

Ap

in quarters,

we

D=

get 2A 2 as the required effective aperture for each helix,

100

is

and so

the directivity for each helix must be

D = ~2\>=25
and

its

length

L = f| =

1.7A

or about 8 turns for a helix with a pitch angle a

= 12,

each helix

from

at the center

helices should

be

.4A.

of its aperture area,

we

see

as

shown

in Fig. 14-546.

Fig. 14-54rf that the spacing

Putting

between

692

[CHAP. 14

ELECTROMAGNETICS

Single helical

beam antenna

-L=6.7X

4/ a o o a vawa aooeeeeaa'aa awffwwff

(o)

31 turns

Side view

Ground plane

Quad

helix

Front view

M~i;4X-*J

ffoTTBOTu'

Physical

aperture

Or

(b)

2.8X

8 turns

^BOTuinnr

O^l

Ap A e

*L=1.7X*H
Side view
Front view

Helicone
Conical horn.

Coaxial line

Helical

(e)

beam

antenna

ix19 turns

3.6X
Effective

aperture
Front view

FIGURE

14-54

(a) Single helical-beam antenna,

compared on

(W quad-helix

basis of effective apertures

Ae

array,

and

(c)

helicone antenna

All three antennas have equiv-

alent directivity.

The antennas of Fig. 14-54a and b have equivalent directivity, but the quad
helix array is more compact. On the other hand, the single long helix is simpler to feed.
Other choices are also available to the designer.

For example, a helicone antenna,!

likely candidate.

This antenna consists of a helical-beam

Antenna with Low Side Lobe

a
shown
antenna (fed by coaxial cable) enclosed in a truncated conical horn. A directivity
D = 100 requires a length L = 4.51 and horn-mouth diameter of 3.6A. Note that the

as

in Fig. 14- 54c, is

Circularly Polarized
t K. R. Carver, The Helicone:
Level, Proc. IEEE, 55: 559 (April 1967).

ANTENNAS AND RADIATION

sec. 14-23]

693

100

90

80

70

60

50

'o

E
UJ
40

30

20

-50X(r)c(l-r 2 )

10

n
Highly tapered

Uniform

n=2

n=l

n=0
-15

distribution

distribution

-35

-30

-25

-20

n=3

_J

-40

First sidelobe level (dB)

FIGURE

14-55

Aperture and beam efficiencies as a function of side-lobe

tribution) for a circular aperture of 50A diameter.

about KM. 2

level

(and aperture

dis-

more than the 8A 2 needed for

because the aperture distribution over the mouth of

physical aperture for this antenna

is

a uniform aperture distribution

the helicone antenna is highly tapered

EXAMPLE

Determine the aperture

This

is

(resulting in reduced side lobes).

efficiency

and the beam

efficiency e

a 50A-diameter circular aperture with aperture distribution of the form

at the center

and low

and

at the

rim (n

= constant).

What

(1

r2

)",

Cl/l A for

where

distribution provides high efficiency

side-lobe level ?

solution

The aperture efficiency

is

obtained from

(4).

The beam efficiency is obtain-

ed by integrating the pattern over An for Cl A and over the main beam for Cl using numerical
methods. The results are shown in Fig. 14-55 with efficiency as ordinate vs. the first side-lobe

694

ELECTROMAGNETICS

level as abscissa.

[CHAP.

14.

We note that a highly tapered distribution (n = 3) gives a very low side-lobe

(36 dB) but a low aperture efficiency. However, at a side-lobe level of 20 dB (n i)

both aperture and beam efficiencies are large. Hence, at this side-lobe level and corresponding aperture distribution (see dashed curve) we have a satisfactory compromise.t

level

14-24

COHERENCE

In Chap.
the

Ex

1 1

and

coherence was discussed in connection with wave polarization.

Ey components

Ex = E

(t) sin

wt

all

S(t)]

(2)

the time functions are independent, the resultant

pletely unpolarized or incoherent.

[It is

assumed that

will

now

traveling to the right, as in Fig. 14-566.

As

Coherence

in

if

(1)

Ey = E2 (t) sm[cot +
where

Thus,

of a wave (as in Fig. 14-56a) are given by

another sense

be discussed.
the

wave

wave

<'1 2 (f)>

travels

is

said to be

com-

= <2 2 (0>]

Consider a plane wave

from point

to point 2,

^Wave fronts^

Wave
direction

Bx

Ex

Wave outward (front view)

Wave to right (side view)

(a)

FIGURE

(6)

14-56

(a) Electric fields of

wave

and

with electric

(b) wavefronts

traveling out of page illustrating polarization coherence

fields at

four points illustrating wavefront (or

phase) coherence.

t K. R. Carver, B. Potts, and R. Widner, Beam and Aperture Efficiencies vs. Sidelobe
Level, IEEE/G-AP Symp. Los Angeles, 1971.
See also R. T. Nash, Beam Efficiency

Limitations of Large Antennas,

(December

1964).

IEEE

Trans. Antennas Propag., AP-12: 918-923

ANTENNAS AND RADIATION

SEC. 14-24]

we

expect that the phase at point 2 will be retarded by

phase of point 3 to be identical with that at point

we expect the

(phase difference zero).

The

phase difference of points 2 and 4 should be zero.

Likewise,

fid.

695

Also the

conditions just described

may

be characterized as providing wavefront coherence, in contrast with the situation

involving

polarization.

homogeneous media, wavefront coherence

In

However,

but the wave need not be coherent with respect to polarization.

inhomogeneous media

means

we can no

that

is

it

possible

longer define a wavefront,

Ex components

for example, the

wavefront incoherence

that

of the

field at

i.e.,

prevails,

exists,

an equal-phase surface.

points

and

in

which
Thus,

3 (Fig. 14-566) will be

time-independent

Wavefront incoherence may be apparent

received through the earth's ionosphere

from

at wavelengths of 10 to 20
celestial

radio sources

m for waves

when comparing

This
1 and 3 in Fig. 14-566) separated by several kilometers.
means that an interferometer antenna with elements separated by such distances will
not produce fringes (multilobed pattern, as in Fig. 14-22). The ionosphere has less

phase at points (such as

effect at shorter

wavelengths, and at 100- to

200-mm wavelengths

fringes are obtained

with interferometer base lines (element separation) of thousands of kilometers, thus

This

demonstrating wavefront coherence over large distances.

is

remarkable since

very small (one quantum contains hf J of energy, where

34
6.6 x 10~
J sand/= frequency inHzjat A = 200 mm, A/

the energy of a radio photon

is

h = Planck's constant =
10" 24 J). Hence, it means either that a photon spreads
front or that

all

its

energy over a very wide

the photons in a wavefront are phase-locked.

Although wavefront coherence usually

prevails,

it is

instructive to consider the

For example, conOhio State University

behavior of receiving devices which are completely incoherent.

sider the cosmic-ray telescope planned

Radio Observatory.

The

by Grote Reber

at the

basic element of this telescope

is

a pair of disk-shaped

cosmic-ray sensors arranged as coincidence counters (A and B), as in Fig. 14-57.

Thus, a particle passing through both sensors {A and B)

passing through only one sensor will not.

As a

will

be counted, but one

result the pair

of sensors has a cone

of reception (or beam width) defined by the ray-path geometry (geometrical optics).
By increasing the sensor-separation distance d the cone angle can be decreased. The
of the telescope can be increased by adding more units (C and D) as in

sensitivity

Fig. 14-57.

In general, particles passing through the sensors (A and B) of the

left

unit will be independent or incoherent with respect to those passing through the

sensors (C

and D) of the

sensitivity

but the

right unit, so that

beam width

adding the right unit doubles the telescope

This is in contrast
is unchanged.!

(cone of reception)

t However, if a shower of particles forming a flat disk (transverse to their direction of

travel) impinges on the telescope, the difference in arrival time at the two units (left
and right) could be used to determine the direction of arrival, hence providing quasi
coherence.

696

ELECTROMAGNETICS

[CHAP. 14

This particle
not counted
This particle

counted
of
reception

M
Coincidence
counters

Particle sensors

FIGURE 14-57
Cosmic-ray telescope,

illustrating lack of

coherence between

left

and

right units

for individual particles (After Reber.)

to

an antenna, where adding another element or unit

increase the effective aperture (and directivity)

in a broadside array would both

and make the response pattern (beam

width) narrower.

14-25

FRIIS

FORMULA AND RADAR EQUATION

Consider the communication

antenna of

effective aperture

circuit

of Fig. 14-58a, consisting of a transmitter

A et and

receiver

T with

with antenna of effective aperture

The distance between transmitting and receiving antennas is r. If the transpower P, were radiated by an isotropic source, the power received per unit area
the receiving antenna would be
.

mitter
at

S
and the power

available to the receiver

~4nr 2

(1)

would be

P'

= SA

(2)

ANTENNAS AND RADIATION

sec. 14-25]

697

(o)

(Passive)
Reflecting

object

(b)

FIGURE

14-58

(a) Single transmission path with

used in Friis transmission formula.

(c) Double-path geometry used
in obtaining radar equation.

eters
(6)

-*

paramA.

and

The transmitting antenna has an effective aperture A et and hence a directivity D =

4nA et IA 2 so that with the antenna connected the power available at the receiver is
,

P'4nA et
(3)

Substituting (1) in (2)

and

(2) in (3) gives

r r~
Pr

P.''

A A et
r

P,A 4nA e
4nr 2

k2

(dimensionless)

Pr = received power, W
P, = transmitted power, W
A = effective aperture of receiving antenna, m 2
2
A e = effective aperture of transmitting antenna, m
antennas,
m

transmitter
distance between receiving and
r
X = wavelength, m

where

(4)

698

ELECTROMAGNETICS

This

is

[CHAP. 14

the Friis transmission formula, which gives the ratio of the received to trans-

mitted power for a direct path.

now

Let the problem

be modified, as in Fig. 14-586, so that a passive reflecting

object, such as a flat sheet or a single passive corner reflector, constitutes part of the

For convenience

transmission path.

obtain
7T

where

Applying the

AAA

P
i

the reflecting object be regarded as a pair of

let

antennas connected as in Fig. 14-58c.

formula in two

steps,

(dimensionless)

A e effective aperture of reflecting object, m 2

r distance from transmitting (or receiving)
in (5) are the

same

(5)

antenna to reflecting object, m.

This is a radar equation

as for (4).

for the ratio of the received to transmitted power.

and transmitting

If the receiving

A e = A = A we have
Pr A 2 A e 2
F -~rVS~

antennas are identical so

we

= " *U
r a

The remainder of the symbols

Friis

(6)

The aperture A e

is

reflecting object

is

Ae = Ap

where

Ap

the effective scattering aperture of the reflecting object. t

a large,
is

flat,

conducting sheet oriented for specular

the physical aperture.

For most objects the

scattering aperture

is

much

less

than the cross-sectional

area because the objects tend to scatter or reradiate isotropically.

change

(i.e.,

is

A2a

P,

47tAV

the radar cross section

(m 2 ).J

power (assumed
Thus

ratio of the scattered

power

Making

this

assuming isotropic scattering) we get the radar equation

r

where a

If the

reflection,

density).

t It is

We

cross section {&)

beam

is

equal to the

4nr 2 Sr

power

power

(8)

density

S-,inc

assumed that the angle subtended by the

the antenna
t

The radar

isotropic) to the incident Poynting vector (or

scattered

incident

(7)

reflecting object

is

small compared to

width.

have

P = (scattered
r

power) x (fraction of sphere subtended by receiving antenna)

on object) X a X (fraction of sphere subtended

= (power density incident

by receiving antenna)

= tP,4-nAI4nr 2 A 2 x a x (.4/4* r 2
= i,/4 2 <r/47rAV4
We can also derive (7) from (6) as follows
)

object

but it

Referring to (6) the aperture of the

equal to the radar cross section as regards incident power [A e (inc)
a],
set equal to the aperture of an isotropic antenna as regards scattered power
A 2 /4i7].
Hence, we can put A e 2
A e (mc)A e (sc) <rA 2 /4rr in (6), which
:

is set

is

[A e (sc)

gives (7).

ANTENNAS AND RADIATION

sec. 14-26]

Sr =

where

Sinc =
For a

power density at distance

power density incident on object,
scattered

Wm~

cross section

moon

W m~

large perfectly reflecting metal sphere of radius a, the radar cross section

equal to the physical cross section (na

the

r,

699

is

is

smaller.

about

2
).

For example,

at

is

For imperfectly reflecting spheres, the radar

meter wavelengths, the radar cross section of

0.1 times the physical cross section.

RADIO TELESCOPES, NOISE POWER,

14-26

AND ANTENNA TEMPERATUREf

The noise power per unit bandwidth available at the terminals of a resistor of resistance
R and temperature T (Fig. 14- 59a) is given byj

w = kT

= power per unit bandwidth, W Hz -1

k = Boltzmann's constant = 1.38 x 10~ 23
T = absolute temperature of resistor, K

(1)

where w

The temperature Tin

(1)

may be

If the resistor is replaced

noise

power

will

K"

called the noise temperature.

by a

lossless

impedance presented

(Fig. 14- 59ft), the

resonant antenna of radiation resistance

at the terminals

not be the same unless the antenna

is

is

unchanged.

receiving

However, the

from a region

at the

temperature T.

(o)

w=kT

R<T

Half -power
level

Radio telescope

FIGURE
(a)

14-59

Noise power of resistance

radiation resistance

R at temperature T and (6) noise power of antenna

TA due to distant regions in antenna

at temperature

beam.

t
t

For a more complete treatment see Kraus, " Radio Astronomy," pp. 97-104.
Electric Charge in Conductors, Phys. Reu,, 32:

H. Nyquist, Thermal Agitation of

110-113 (1928).

700

ELECTROMAGNETICS

[CHAP. 14

In a radio-telescope antenna operating at centimeter wavelengths, the beam may

be directed at regions of the sky which are at effective temperatures close to absolute

The noise temperature TA of the antenna is equal to this sky

temperature and not to the physical temperature of the antenna structure. Thus, for
a radio-telescope antenna the noise power is given by
zero (0 K, or -273C).

(WHz H )

w = kTA
where

TA

is

(2)

the antenna (noise) temperature, or temperature of the antenna radiation

by the sky temperature at which the antenna beam is directed.

Thus, a radio-telescope antenna (and receiver) may be regarded as a radiometer (or
temperature-measuring device) for determining the temperature of distant regions of
resistance, determined

space coupled to the system through the radiation resistance of the antenna.

been assumed that the antenna has no thermal losses and that
pattern

is

all

It

has

of the antenna

directed at the sky (negligible side and back lobes).

In receiving with a radio telescope

it is

convenient to express the received power

as a (power) flux density.

the effective aperture

Ae

Thus, dividing the received power per unit bandwidth by

of the antenna (Fig. 14-596) yields the flux density S, or

S= T =
Note that the

units for the flux density

(W

(Wm- 2 Hz-')
nrT 2

Hz -1 ) are the same as

we may regard

vector per unit bandwidth, so that

Most such sources

Fig.

1-7),

celestial sources

of radio

waves which are completely unpolarized; i.e.

Since any antenna (regardless of its polarization

radiate

they have polarization incoherence.

characteristics)

for the Poynting

the flux density as a measure of the

Poynting vector (per unit bandwidth) received from distant

emission.t

(3)

can receive only half the incident power of an unpolarized wave (see

the actual flux density should be twice that given in

(3),

or

(4)

_]
flux density, W m~ 2 Hz
= Boltzmann's constant =1.38 x 1(T 23 J K"
TA antenna (noise) temperature, K
A e = effective aperture of telescope antenna, m 2

where

S=
k

The value S in
celestial

(4) results

source extent
t

from power received over the entire antenna pattern. If the

small compared to the antenna-beam solid angle Q and
x

fi s is

The common
(fu)=10- 26

unit of flux density in radio-astronomy measurements

Wm-

Hz-'.

is

the flux unit

ANTENNAS AND RADIATION

SEC. 14-26]

beam

the antenna

Hence,

(3).

If the

known,

it is

is

as given in

said that this relation applies to point sources.

angular

it is

aligned with the source, the observed flux density

is

701

size

of the source

is

small (compared to

possible to determine the source temperature

QA ) and its magnitude is

T very simply from the
s

relation

ta = ^tTs

or

il A

where

J2X

TsS =

(5)
y
'

antenna beam solid angle,

sr (see Fig. 14-596)

Qs =

source solid angle, sr (see Fig. 14-596)

TA =
7^ =

antenna (noise) temperature,

source temperature,

T
n

EXAMPLE

The antenna temperature for the planet Mars measured with the U.S. Naval
1
Research Laboratoryf 15-m radio telescope at 31.5
wavelength was 0.24 K. Mars
subtended an angle of 0.005 at the time of measurement, and the antenna HPBW = 0.1 16.

mm

Find the temperature of Mars.

solution

From

(5) the

temperature of Mars

This temperature

given by

0.116 2

to*

than the infrared temperature of the sunlit side (250 K), implying

is less

that the radio emission

is

may

originate further below the Martian surface than the infrared

radiation.

The

sensitivity

of a radio-telescope receiver must be

small antenna (noise) temperatures

radio-telescope receiver

is

TA

The minimum

sufficient to

measure very

detectable temperature of a

given by

(6)

= system constant (order of unity), dimensionless

T'sys = system temperature (sum of antenna and receiver equivalent temperatures),

where k

K
A/= predetection bandwidth of receiver, Hz
t = postdetection time constant of receiver,
t

C. H. Mayer, T. P. McCullough, and R. M. Sloanaker, Observations of Mars and

Jupiter at a Wavelength of 3.15 cm, Astrophys. J., 127: 11-16 (January 1958).

702

[CHAP. 14

ELECTROMAGNETICS

EXAMPLE

The Ohio

aperture of 1,200

bandwidth

MHz,

is

State University 110

the output time constant

Find the minimum detectable

solution

From

minimum

= -g= =
VA/7

From

(4) the

minimum

5mln

Smia =

or

By

s,

and the system constant

is

2.2.

," X9

V6

38

is

given by

=0.026K

x 10 6 x 10
is

<10 '" Xa026

= 5.9xl0-"W m - 2 Hz-

>

Ae

59

10

detectable temperature

detectable flux density

= g^ii!= 2

is

flux density.

(6) the

r.

m radio telescope has an effective

K at 1,415 MHz. The radio-frequency

by 21

and a system temperature of 90

1,200

mfu

repeating observations and averaging, the

minimum can be

further reduced

where n = number of observations). In an all-sky survey at 1,415 MHz

Ohio
State University radio telescope it is expected that about 20,000 radio
with the
detected at flux densities above 0.18 fu.
be
sources will
(oc y/l/n,

14-27

ROTATING BAR MAGNET AS A

VERY-LOW-FREQUENCY RADIATOR (PULSAR)t

One of the most remarkable

objects discovered with radio telescopes

These radio sources emit short pulses of radio energy

The

pulse repetition rate for a given pulsar

is

at rates

is

the pulsar.

up to about 30

astonishingly constant.

It is

thought

that these pulsars are rapidly rotating superdenset neutron stars within our
galaxy.

_1
s

own

A pulsar is believed to result from the explosion and subsequent gravitational

Typically a pulsar

collapse of a burned-out star.

core perhaps 10 to 20

may

consist of a dense spherical

km in diameter with a co-rotating disk of plasma extending out to

a distance at which the peripheral velocity equals the velocity of light. It is implied
that a pulsar emitting 30 pulses per second is rotating 30 r s~ * (= 1,800 r min~ *). For
such a pulsar this means a radius

model of such a pulsar

also very intense (millions

Anthony Hewish,
Nature of Pulsars,

Mass of

(r

= c/co)

of 1,600 km.

An

The magnetic
of gauss or hundreds of teslas), and

shown

is

cm 3

in Fig. 14-60.

idealized simplified
field

Pulsars, Sci, Am., 219(4): 25 (October 1968);

ibid., 224(1):

of star

is

of the star

is

this field locks the

J. P.

Ostriker,

48 (January 1971).

approximately equal to the mass of

km3

of rock.

The

ANTENNAS AND RADIATION

sbc. 14-27}

FIGURE
Idealized

703

14-60

model of

pulsar, a superdense neutron star with co-rotating disk of

magnetized plasma, which attains the velocity of light at its perimeter. Verylow-frequency radio waves of enormous power are radiated by the mechanical
rotation of the star (and disk). The radiation is linearly polarized in the equatorial

plane (plane of disk) and circularly polarized in the direction of the axis.

surrounding plasma region to the star (Fig. 14-60).

The magnetized plasma

(Sec.

15-10) rotates with the neutron star at the center.

like

It has been suggested that the pulsar is a huge rotating machine which acts
an antenna to radiate extremely powerful very-low-frequency radio waves (order

of 30

Hz

oscillate

or less).| Usually an antenna structure is stationary, and electric currents

back and forth along the antenna conductor. However, in the pulsar we
t

M.

J.

Rees,

New Interpretation of Extragalactic Radio

(Jan. 29, 1971).

Sources, Nature, 229: 312-317

Discusses pulsars and galactic nuclei.

704

[CHAP. 14

ELECTROMAGNETICS

appear to have a huge bar magnet (several thousand kilometers long) rotating about
its center and producing a changing magnetic field, which generates an electromagnetic
wave radiating in all directions. In the equatorial plane (perpendicular to the axis)
the pulsar radiates with linear polarization (E parallel to axis), while in the axial (or
polar) direction it radiates with circular polarization (E rotating in plane perpendicular

In Sec. 14-3 on the short dipole

to axis).

small antenna
length)
(A

which

of 10

22

In the case of a 30

it

gradually spins

down

W are probably involved.

1

its

'

pulsar,

is

an

is

km

about

energy through the very-low-frequency radio

Radio powers of the order

Assuming that the power is radiated isotropically,

Gm distance

is

this implies

0,000

over millions of years.

S = T-i2 =
4nr

given by

10

22

-T-^Ts
18

W m"

800

47tl0

electric field intensity

E = JSZ = ^800
An

r s

have significant radiation resistance.

sufficient for the structure to

is

the Poynting vector at

and

that the radiation resistance of a

Thus, the diameter (3,200 km) of the rotating magnetized disk

According to Rees, a pulsar loses

waves as

we noted

Hence, systems which are small (compared to the wave-

very low.

may not be effective radiators.

elf).

A/3,

is

x 377

550

V m"

-1

of this intensity lasting for the order of i period (1/60 s for a 30 r s

pulsar) can accelerate electrons to relativistic velocities (v <, c) (see Chap. 15).f This
is

electric field

at a distance

of

Gm.

The magnetic

field

associated with this

E will

charged particles in the direction of the wave, so that the particles

swept along by the waves

large distance

radiation at

(like surfers)

from the pulsar.

much

producing a

shell

of

will

deflect the

tend to be

relativistic particles at

These relativistic particles in turn produce synchrotron

higher radio frequencies, and this

is

the radiation detected by radio

telescopes.

14-28

ANTENNAS FOR POLARIZATION MEASUREMENTS}

In this section

we

describe

an antenna system

of a received wave of unknown

suitable for measuring the polarization

characteristics.

The measurements

will

be made

without measurement of phase.

t In a relatively short (1-km) distance an electron would fall through a potential difference of 550 kV. From (15-1-11.) this is sufficient to attain a velocity e-vc.
X Kraus,

"Radio Astronomy,"

pp. 126-128.

ANTENNAS AND RADIATION

SEC. 14-28]

705

Circularly polarized

Linearly polarized dipole antennas

helical

-J\

beam antennas

A_

135"

Horizontal

FIGURE

Vertical

Slant 45

Slant 135

Left

Right

14-61

Arrangement of

six

wave of unknown

antennas for measurement of polarization parameters of

Four

characteristics.

linearly polarized

and two

circularly

polarized antennas are used.

Let the antenna system consist of four linearly polarized dipole antennas and
two helical-beam antennas (one left- and one right-handed)t deployed as in Fig. 14-614
The linearly polarized antennas are oriented for horizontal, vertical, and two perpendicular slant polarizations.

the page.

The received wave

is

incident

from a direction normal to

Let a separate receiver be connected to each antenna (or a single receiver to

each in turn).

The power received from

the incoming

wave by each antenna may then

be designated by

Px for horizontal dipole

Py for vertical dipole
Px for slant 45 dipole
P'y for slant 135 dipole

PL for left circular helix

PR for right circular helix
assumed that all six antennas have equal effective aperture. This means that all
should give identical power response to a completely unpolarized (incoherent) wave.
Then from (11-5-3) to (11-5-6) we have that the normalized Stokes parameters are
It is

/
S

PX +Py
Px + Py

+ Py'
Px +Py

P.'
'

Pl+Pr
Pl

+ Pr

(1)

_Q_PX ~Py
1

p,

p,

(2)

left-handed helical-beam antenna receives (or transmits) left circular polarization

definition), and a right-handed one receives (or transmits) right circular polar-

(IEEE

ization.

Suzuki and A. Tsuchiya used three pairs of antennas similar to this arrangement
for measuring the polarization characteristics of radio waves emitted by the sun
[A Time Sharing Polarimeter, Proc. IRE, 46: 190-194 (January 1958)].

t S.

706 ELECTROMAGNETICS

These

results are

power pairs (Px

[CHAP. 14

summarized

in

$2

S3

I~

(3)

Px + Py

P*

Table 14-5.

(4)

P*

Assuming

that the

sums of each of the

+ Py Px + Pr Pt + PR ) are unity, the normalized Stokes parameters

,

by the difference of the vertical and horizontal dipoles, the

difference of the two slant dipoles, and the difference of the two helices, respectively,

s lt s 2

s 3 are then given

as suggested

There

by the table.
some redundancy

is

in the

above system,

i.e.,

A minimum of four is

be measured with fewer antennas.

the Stokes parameters can

required in the most general

For example, we can use a horizontal dipole (Px), a slant 45 dipole (Px d, and a
Then the normalized Stokes
left- and right-handed helical-beam antenna (PL and PR )

case.

parameters are given by

Pl + Pk
*o

(5)

Pl+Pk
2PX

-PL -PR

Sl

"

PL +P

S2

2PX - Pl-Pr
PL + Pr

(6)

(7)

Pl-Pr
*3

(8)

P,.+P*

Table 14-5

RELATION OF NORMALIZED STOKES

PARAMETERS TO ANTENNA PAIRS
Two

Vertical and
horizontal
dipoles

Two slant

beam

dipoles

antennas

so

Sum

Sum

Sum

Si

Difference

Stokes
parameters

Si
Si

helical-

Difference
Difference

ANTENNAS AND RADIATION

PROBLEMS

707

PROBLEMS
*

14-1

thin-dipole antenna

(a) directivity

D,

resistance R,

and (/) terminal

*
where

The radiated

14-2

is

A/15 long.

(b) gain G, (c) effective aperture

If

Ae

it

has a loss resistance of 1.5

(d)

beam

solid angle Cl A

find

resistance.

of a short-dipole antenna

field

= length, / = current, r = distance,

and 6

is

given by ||

= orientation

angle.

= 30/(//r)sin 0,

Find the radiation

by integrating the Poynting vector over a large sphere and equating to

resistance

ft,

(e) radiation

PR

r ,

where

= radiation resistance.

R,

Two equal

14-3

and plot the

isotropic in-phase point sources are spaced 1A apart,

pattern

field

and

find the directivity,

(b)

(a) Calculate

Repeat for two sources

in opposite

phase.

Find the point near a short dipole where circular cross field (or elliptical cross
the most nearly circular) occurs.
Make a drawing showing the distance in

14-4
field

which

is

wavelengths and the angle from the dipole to the point.

A linear broadside array consists of four equal isotropic in-phase point sources
= 1A). (a) Calculate and plot the field pattern.

14-5

with A/3 spacing (overall length of array

(b)

Find the

directivity.

14-6

phase

shift

Repeat Prob. 14-5 for the ordinary end-fire

between sources

is

situation,

i.e.,

where the progressive

120.

AM

14-7 (a) An
broadcasting station is to be located south of the area it is to serve.
Design an antenna for this station which gives a broad coverage to the north (from
through N to NE) with reduced field intensity in other directions. However, to obtain

NW

Federal Communications Commission approval the pattern must have a null

N) in order
is

on the same frequency

to protect another station

to consist of

an

in-line array

in that direction.

SE (135 from
The antenna

of A/4 vertical elements oriented along a north-south line

with equal spacing between elements.

The minimum number of elements should be used.

In plan view the problem reduces to a linear array of isotropic point sources, (b)
Repeat part (a) with the additional requirement of another null to the west (90 from N).
Hint :
Hint:

Apply pattern
14-8

Show

multiplication.

that the

HPBW

approximation) by 50.87L A , where

14-9

An

and a cosine

of a long uniform broadside array

L A =L/A = length

antenna with rectangular aperture xiyi has a uniform

field distribution in the

Xi

in

both x and y

given (without

direction (zero at edges,

field in

maximum

the

direction

at center).

If

16A and y t = 8A, calculate (a) aperture efficiency and (b) directivity.
14-10 Repeat Prob. 14-9 for the case where the aperture field has a cosine distribution
14-11

directions.

Assume

that the current

is

of uniform magnitude and in phase along the entire

length of a thin center-fed A/2 dipole antenna,

(b)

is

of array in wavelengths.

Find the radiation

resistance,

(c)

(a) Calculate

and plot the

far-field pattern.

Compare the resistance value of (b) with

center-fed A/2 dipole with sinusoidal in-phase current distribution

* Answers to starred problems are given in Appendix C.

that for a thin

and with the radiation

708

ELECTROMAGNETICS

[CHAP. 14

FIGURE

FIGURE

P14-16
Three sources in triangle.

resistance of

P14-17

Four sources

a A/2 dipole calculated by means of the short-dipole formula,

in square.

(b)

Explain

the reasons for the three different resistance values.

14-12

(a) Calculate

Assume a

antenna.

An

14-13

and

plot the far-field pattern of

sinusoidal current distribution.

(b)

and spaced A/2

apart.

twice the current in the end dipoles (binomial array),

The

is

and plot the far-field

value compare with the HPBW

(c) How does this HPBW

(b) What is the HPBW ?
a binomial array of three isotropic point sources spaced A/2 apart ?

A uniform

14-14

beam

HPBW,

If the

(b) level

and (/)

efficiency, (e) directivity,

14-15
A/4.

linear array consists of 12 isotropic in-phase point sources spaced

Calculate exactly (a)

A/2 apart.

of

current in the center dipole

(a) Calculate

pattern,

id)

resistance.

antenna array consists of three collinear thin A/2 dipole antennas each with

sinusoidal current distribution

for

a thin center-fed 2.5A dipole

Find the radiation

of first side lobe,

(c)

beam

solid angle

QA

effective aperture.

uniform linear array consists of 16 isotropic point sources with a spacing

phase difference 8

= 90

(ordinary end-fire condition), calculate exactly

the six quantities listed in Prob. 14-14.

14-16

Three isotropic point sources of equal amplitude are arranged at the corners

of an equilateral

triangle, as in Fig. P14-16.

If all sources are in phase, determine

and plot

the far-field pattern.

14-17

Four

isotropic point sources of equal amplitude are arranged at the corners

of a square, as in Fig. P14-17.

plot the far-field pattern.

If the

phases are as indicated by the arrows, determine and

A graphical solution may be used for this problem and Prob. 14-16

and Table 14-3).

Apply the Fourier transform method to obtain the

(see Fig. 14-19

14-18

of two equal in-phase isotropic point sources with a separation

to

its

d.

of an array

Reduce the expression

simplest trigonometric form.

14-19

with uniform

on

far-field pattern

An interferometer antenna consists of two square

field distribution,

centers, calculate

and

(a) If the apertures are

plot the far-field pattern to the

broadside in-phase apertures

10A square and are separated 60A

first

null of the aperture pattern.

709

ANTENNAS AND RADIATION

(b)
is

How many

lobes are contained between

the effective aperture ?

does this HPBW compare with the

sources separated 60A?

14-20

(a) Calculate

and

(a) Calculate

(b)

A square loop antenna is

14-23

Find

the following three cases:

14-24

a small loop

(1)

What

(e)

How

the

and

(2)

HPBW?

If the current is

far-field pattern in

D = A/3,

the

antenna with 5A

HPBW?

1A on a side.

plot),

is

of a terminated

uniform and in phase

a plane normal to the plane of the

uniform in-phase current has a diameter D.

(fe)

radiation resistance,

and

(c) directivity for

D = 0.75A, and (3) D = 2A.

Using a Poynting vector integration, show that the radiation resistance of

(a)
is

is

circular loop antenna with

(a) far-field pattern (calculate

What

(b)

far-field pattern

What

around the loop, calculate and plot the

loop and parallel to one side.

(c)

plot the far-field pattern of a continuous in-phase aperture

and plot the

legs and 45 included angle,

14-22

nulls of the aperture pattern?

HPBW for the central lobe of two isotropic in-phase point

field distribution,

20A long with cosine-squared

14-21

first

{d) What is the HPBW of the central interferometer lobe ?

2
equal to 320w*(y4/A)

1,

where

A = area of loop.

(6)

Show

that the effective

2
aperture of an isotropic antenna equals A /4n\
14-25
conductor is wound into a right-handed helix of 10 turns with diameter of

100

mm and

turn spacing of 20

the polarization state

3

MHz and (6)

*
The

14-26
helix

HPBW?
*

is

(c)

14-27
currents,
if

if

is

fed with radio-frequency

power

at a frequency of (a)

GHz.

helical-beam antenna has 30 turns, A/3 diameter, and A/5 turn spacing.

right-handed,

What

Two

what

Calculate and plot the far-field pattern and describe

mm.

the helix

is

is

(a) Calculate

the directivity?

(d)

and

plot the far-field pattern,

What

is

(b)

What

is

the

the polarization state?

A/2 dipoles are crossed at 90.

two dipoles are fed with equal

If the

the polarization of the radiation perpendicular to the plane of the dipoles

the currents are (a) in phase, (b) phase quadrature (90 difference in phase), and (c) phase

octature (45 difference in phase)?

14-28

Two

identical helical-beam antennas,

one left-handed and the other

right-

handed, are arranged as in Fig. P14-28. What is the polarization of the radiation to the
right if the two helices are fed (a) in phase and (ft) in opposite phase ?
14-29

A linear array consists of an in-line configuration of 24 A/2

The dipoles are fed with equal currents but with an arbitrary
dipoles.

What

value of 8

is

dipoles spaced A/2.

progressive phase shift 8 between

required to put the main-lobe

maximum

(a) perpendicular to

the line of the array (broadside condition), (b) 25 from broadside, (c) 50 from broadside,

and (t/)75 from broadside? (e) Calculate and plot the four field patterns in polar coordinates.
(/") Discuss the feasibility of this arrangement for a scanning array by changing feed-line
lengths to change 8 or by keeping the array physically unchanged but changing the fre-

What

practical limits occur in both cases?

Three identical right-handed helical-beam antennas spaced 1.5A apart are
arranged in a broadside array, as in Fig. P14-30. (a) If the outer two helices rotate on
of the
their axes in opposite directions while the center helix is fixed, determine the angle

quency.

14-30

<f>

main lobe with

respect to the broadside direction

and describe how

<f>

varies as the helices

710 ELECTROMAGNETICS

[CHAP. 14

Right-handed helix

fe fe

K=>

Zenith

Radiation

A/ 2 dipole
Left-handed helix

Ground plane

FIGURE

FIGURE

P14-28

Two helices.

rotate.

(6)

function of

/4-i/

What
<j>

is

14-32

scan angle <?

(c)

What

is

FIGURE P14-37
Dipole above ground.

the main-lobe

HPBW

as a

Design a log-periodic antenna of the type

between 50 and 250

maximum

the

P14-30

Three-helix array.

MHz.

Make

in Fig. 14-37 to operate at frequencies

a drawing with dimensions in meters.

Two short-dipole antennas A/10 long with 1

to each other (side by side)

and

situated IX apart.

Cl loss

Find

resistance are oriented parallel

(a) the

mutual impedance and

(b)

the terminal impedance.

14-33

Explain

why

the mutual resistance of two antennas can be both positive

and

negative but the self-resistance of a single antenna can only be positive.

14-34 (a) Calculate the mutual impedance of two parallel A/15 dipoles as a function
of their separation from zero distance to 2A. (b) Plot the mutual resistance
as abscissa
vs. the mutual reactance
m as ordinate and put marks on the resulting impedance spiral

RM

at dipole separation intervals of 0.1A.

14-35

Two

Compare with Fig. 14-44 for two A/2 dipoles.

A/2 dipoles are spaced A/8 apart and fed with equal currents in opposite

phase, producing an end-fire antenna.

The dipoles are oriented parallel to each other and

perpendicular to the line connecting their centers (as in Fig. 14-46). The dipole length-todiameter ratio is 2,000. (a) Calculate and plot the two principal-plane far-field patterns.

What are the HPBWs in the two principal planes? (c) What is the gain in decibels over
a single A/2 dipole (assuming zero losses) ? (d) What is the directivity in decibels ? (e) What
is the gain over an isotropic source if each dipole has a loss resistance R = 1 CI
?
L
(/) Design
a two-conductor feed system for this antenna with a terminal or driving-point impedance of

lb)

400 n.
14-41

For self- and mutual-impedance values needed

and

in this

problem interpolate from Figs

14-44.

14-36

broadside array consists of four in-phase A/2 dipoles spaced A/2 and fed
The dipoles are oriented parallel to each other and perpendicular to

with equal currents.

ANTENNAS AND RADIATION

PROBLEMS

the line connecting their centers.

and plot the two principal-plane

principal planes ?

(c)

711

The dipole length-to-diameter ratio is 2,000. (a) Calculate

(b) What are the HPBWs in the two
far-field patterns,

Find the gain in decibels over a

single A/2 dipole (assuming zero losses),

the gain in decibels over an isotropic source (assuming zero losses) ? (e) What is
the directivity? (/) What is the gain in decibels over an isotropic source if each dipole
1 O?
has a loss resistance R L
(g) Design a two-conductor feed system for this antenna

(rf)

What

is

with a terminal or driving-point impedance of 500 ii. Note that this feed system must supply
equal currents to the four dipoles in spite of unequal dipole feed-point impedances.

14-37

thin A/2 dipole

is

an

parallel to

infinite flat

ground plane and a distance s

(a) Caclulate and plot the gain of the dipole with ground plane

it, as in Fig. P14-37.

with respect to a single A/2 dipole in free space for s values from zero to A/2. Confine
consideration to the zenith direction (perpendicular to ground plane). Assume zero losses.

above

(Z>)

Repeat

(a) for dipole loss resistance

RL =

Ci.

Calculate and plot the far-field pattern of a thin A/2 dipole spaced A/2 from an
Plot the pattern in polar coordinates
infinite conducting strip 8A wide (as in Fig. 14-48a).

14-38

with

field in decibels vs. angle, as in Fig. 14-486.

14-39

Show

a large distance

D 2 = directivity

maximum mutual impedance Zm of two antennas separated by

/
Z =(v ) Z) 2 \/ RiR^jltrr,., where D = directivity of antenna 1,

that the

is

= radiation resistance of antenna 1, R 2 = radiation resisr* = r/A = separation of antennas (wavelengths) (see Fig. 14-38).

of antenna

2,

Rt

2, and
assumed that the receiving antenna is terminated for maximum power transfer.
14-40 (a) Show by means of an equivalent network that at the terminals of a receiving antenna (see Fig. 14-38) the equivalent, or Thevenin, generator has an impedance
Z 22 (.Z m 2 IZu) and an emf CLT,Z/Z 11 where Z ,= self-impedance of transmitting
antenna, Z 22 = self-impedance of receiving antenna, Zm = mutual impedance, and ITi = emf

tance of antenna
It is

applied to terminals of transmitting antenna.

(J))

terminals of the receiving antenna results in the

14-41

Show

that the directivity

D=

12.5 dB.

square-corner reflector has a driven A/2 dipole A/2 from the corner,

Calculate and plot the far-field pattern in both principal planes,

in the

two

principal planes?

(d) Calculate the directivity

and

(2)

from

to the

transfer?

square-corner reflector has a spacing of A/4 between the driven A/2 element

and the corner.

14-42

What load impedance connected

maximum power

(c)

What

two ways:

HPBWs, and compare.

is

(b)

What

are the

(a)

HPBWs

the terminal impedance of the driven element?

(1)

from impedances of driven and image dipoles

Assume

perfectly conducting sheet reflectors of

infinite extent.

14-43

(a)

Show

that the variation of field across the aperture of a paraboloidal

reflector with isotropic source is proportional to 1/[1

from

axis of paraboloid

plane and the feed

is

and

L = focal

distance.

+ (r/2L) 2

(6) If the

isotropic over the hemisphere subtended

],

where

= radial

distance

parabola extends to the focal

by the parabola, calculate the

aperture efficiency.
14-44

The Schmidt telescope employs a spherical

Use the principle of equality of path

as in Fig. P14-44.

thickness of the plate as a function of radius.

reflector with

a correcting plate,

length to determine the required

712 ELECTROMAGNETICS

[CHAP. 14

Rays

FIGURE

P14-44
Schmidt telescope.

14-45

Two

A/2-slot antennas, as in Fig. 14-52a, are arranged

conducting sheet with a spacing of 1A between centers.

end to end

in

a large

If the slots are fed with equal

in-phase voltages, calculate and plot the far-field pattern in the two principal planes.
that the
plane coincides with the line of the slots.

Note

14-46

90 + /10
is

ft.

The complementary
If the slot

antenna

is

dipole of a slot antenna has a terminal impedance

boxed

in so that

the terminal impedance of the slot antenna?

it

radiates only in

Z=

one half-space, what

The box adds no shunt susceptance

at the

terminals.

14-47 (a) Determine the length /, width w, and half angles in E and
planes for a
pyramidal electromagnetic horn for which the mouth height h = 8A. The horn is fed with
a rectangular waveguide with TE 10 mode, as in Fig. 14-53. Take 8 = A/10 in the E plane

and S
is

= A/4

in the

the directivity?

H plane,
(d)

(b)

What

is

What

are the

HPBWs

in

both

E and

H planes ?

(c)

What

the aperture efficiency?

14-48 A radio link from the moon to the earth has a moon-based 5A-long righthanded helical-beam antenna (as in Fig. 14-33a) and a 2-W transmitter operating at 1 .5 GHz.
What should the polarization state and effective aperture be for the earth-based antenna in

order to deliver 10"

seconds.

,4

to the receiver?

Take

the earth-moon distance as 1.27 light-

'

ANTENNAS AND RADIATION

14-49

(a)

Design an earth-based radar system capable of delivering 10

713

W of peak

echo power from Venus to a receiver. The radar is to operate at 2 GHz, and the same antenna
is to be used for both transmitting and receiving.
Specify the effective aperture of the antenna

and the peak transmitter power. Take the earth-Venus distance as 3 light-minutes, the
diameter of Venus as 12.6 Mm, and the radar cross section of Venus as 10 percent of the
physical cross section, (b) If the system of (a) is used to observe the moon, what will the
received power be? Take the moon diameter as 3.5
and the moon radar cross section

Mm

as 10 percent of the physical cross section.

14-50

(a)

Design a two-way radio link to operate over earth-Mars distances for data

MHz bandwidth

power of 1 ~ 9
Hz ~ * is to be delivMars receiver. The Mars antenna must
be no larger than 3 m in diameter. Specify effective aperture of Mars and earth antennas
and transmitter power (total over entire bandwidth) at each end. Take earth-Mars distance
as 6 light-minutes. (6) Repeat (a) for an earth- Jupiter link. Take the earth- Jupiter distance
transmission

at2.5GHzwith5

ered to the earth receiver and

W Hz

'

to the

as 40 light-minutes.

Design a

14-51

CW doppler radar operating at 30 GHz for measuring the velocity of

automobiles at distances up to 1 .5 km. The transmitter power is limited to 500 mW, and the
receiver requires 10" 10 W.
Specify antenna size and the precision with which frequency
shift

must be measured

Take the radar

14-52

which x

= 0,

sheets are

formed.

order to achieve velocity measurements accurate to

in

same

= 0,

and z

made to intersect
To distinguish this

and

their intersection at the origin.

in this

If three flat

manner, a three-dimensional square-corner

reflector

from a corner
(a)

Show

reflector with driven

wave

that a plane

from any direction within the octant bounded by the planes

direction (producing retroreflection).

compared

reflector are large

physical aperture,

Show

that

if

active antenna?

What

reflector

element

incident
is

conducting

on

reflected

it

is

may be

this passive

back

in the

the dimensions of this retro-

how

such a reflector might be used to

on an

axis

limitations

assist

in radar

which makes equal angles with

the x, y, and z axes of a passive square-corner reflector, could

14-53

hr _1 .

to the wavelength, the effective scattering aperture equals the

Discuss

(c)

(6)

(d) If an isotropic source were placed

navigation,

km

Consider the three mutually perpendicular rectangular coordinate planes for

called a passive square-corner reflector,

reflector

cross section of an automobile as 30 percent of the physical cross section.

it

function effectively as an

would there be?

Consider a collinear array of eight A/2 dipole antennas with spacing between

dipole centers of A/2.

and

Let the dipoles be numbered consecutively

through

8.

(a) If each

and 6, and 4 and 5 is connected together by an identical transmission line of the same length, show that a plane wave incident on this array will be reflected
back in the same direction. This passive retroreflecting antenna is called a van Atta array,
(b) What limitation is there on the angle of incidence 6 of the wave?
Measure 6 from the
dipole pair

8,

broadside direction,
is its

2 and

(/)

What

(c)

radar cross section?

navigation.

7,

(e)

is

Compare with

how

such an antenna might be used to

What
radar

the passive square-corner reflector of Prob. 14-52.

elements other than A/2 dipoles be used?

eight be used?

(d)

assist in

the effective scattering aperture of this antenna ?

Discuss

(h)

Can more than

(g)

Can

eight elements or fewer than

714

ELECTROMAGNETICS

[CHAP. 14

The Ohio

14-54

State University radio telescope operates at 2,650

MHz

with the

following parameters:

system temperature 150 K, predetection bandwidth 100 MHz,

2
postdetection time constant 5s system constant 2.2, and effective aperture of antenna 800
.

Find

(a)

minimum

detectable temperature

and

(b)

minimum

detectable flux density,

four records are averaged, what change results in (a) and (6)?
14-55 Design a 30-GHz antenna for an earth-resources observation
earth-surface temperatures with

km 2

(c) If

to measure

satellite

Make a

resolution from a 300-km orbital height.

drawing of the antenna with dimensions.

14-56
at

20

Flux densities of 10

MHz.

What

is

Jupiter distance as 40 light-minutes

14-57

or red,

shift

W m"

~ 20

power per

the

unit

Hz -1

commonly

are

bandwidth radiated

and assume

received from Jupiter

Take the

at the source ?

earth-

that the source radiates isotropically.

Some radio sources have been identified with optical objects and the doppler,
(= AA/A = v/c) measured from an optical spectrum. Assume that the objects

R = v/H where
= 1 Mpc = 3.26 x 10 6 light-years), v = velocity
of recession of object in meters per second, and H = Hubble constant = 75 km s" Mpc~ \
Determine the distance R to the following radio sources
(a) Cygnus A (prototype radio
galaxy), z = 0.06, (b) OQ208 (Seyfert galaxy with peaked radio spectrum), z = 0.08, and
(c) 3C273 (quasi-stellar radio source, or quasar), z = 0.16.
{d) The above sources have
with larger red shift are more distant, according to the Hubble relation

R = distance

in megaparsecs (1

megaparsec

'

flux densities as follows at 3

(1

fu

flux unit

10

-26

GHz:

Cygnus A, 600

W m -2 Hz -1

)-

fu; OQ208, 2 fu; and 3C273, 30 fu

Determine the power per unit bandwidth radi-

Assume that radiation is isotropic.

The peak flux density of a pulsar is 25 fu

ated at the source.

14-58

per pulse

(Jr

Hz"

)?

(b) If the pulsar is at

and

that 25 fu

is

above rate

Assume

if it

has one solar mass

10 30 kg)?

(= 2 x

diameter is 25 km and that its kinetic energy of rotation

power with 5 percent efficiency.
14-59
dipole,

Consider the following four antennas:

is

Assuming

how

Assume

that

long can

it

that the pulsar

transformed to radio frequency

horizontal A/2 dipole, vertical A/2

left-handed 8-turn helical-beam antenna, and right-handed 8-turn helical-beam

antenna with four power response Px ,Py ,P L , and P R , respectively.

for

(c)

the only energy lost by the pulsar,

is

the

is

that the source radiates iso-

the average flux density over the bandwidth,

the above radio-frequency radiation

radiate at the

(a)

a distance of 100 light-years, what

energy radiated per pulse in a 20- MHz bandwidth?

tropically

MHz. The pulsar rate is

What is the energy density

at 150

30 s" 1 with pulse duration one-fourth the pulse period,

maximum

response to waves from a distant source.

All antennas are oriented

Each antenna can be switched

in

Case
1

P*
P,

Pl

P*

0.5

20
20

10
10

0.5

0.5
0.5

10
10

20

0.6

0.3

0.5

0.4

0.4

0.1

0.5

0.5

9
6

15
15

18

ANTENNAS AND RADIATION

PROBLEMS

turn to a power-measuring unit to which

parameters and also the parameters

power responses

to the eight

d,

it is

properly matched,

AR, and

waves are as tabulated above.

and of each helical-beam antenna as

systems are required to produce waves of these types?
14-60

A center-fed dipole is
How

in diameter.
(6)

efficient

would

150

mm long.

this dipole

The

32.

Take the

(b)

dipole

is

What

from

efficiency,

could

the measured

efficient

(g)

of each

type of transmitting

made of copper

antenna
aside

if

directivity

be as transmitting antenna at

MHz, (c) 100 MHz, (d) 1 GHz, and (e) 100 GHz? (/) For
how large should a dipole be (measured in wavelengths)?

Determine the Stokes

t for eight different cases

A/2 dipole as 1.6

(a)

715

wire

(a)

mm

10 kHz,

operation as an

What

other factors,

restrict its usefulness ?

14-61

Prove (14-23-4).

14-62

Show

that the terminal impedance of a center-fed linear dipole

is

twice that

for half of the dipole oriented as a stub perpendicular to a ground plane.

What is the effective aperture of a pocket transistor-radio receiver at 1 MHz ?

(6) At 1 km from a 10-kW 1-MHz broadcast station what power
would be available in this aperture? (c) What power is actually required by the receiver
(approximately 1 /*V in 300 ft)? (d) What is the antenna efficiency? (e) Would the pocket
unit be effective for transmitting at 1 MHz ? (/") Would the above considerations be altered

14-63

Assume

(a)

isotropic pattern.

appreciably

14-64

if

one used the actual patterns instead of assuming isotropic patterns ?

Use a computer to calculate the field patterns of the following arrays of

(a)

equal-amplitude isotropic point sources:

(1)

10 sources broadside (A/2 spacing), (2) 100

sources broadside (A/2 spacing), (3) 10 sources ordinary end fire (A/4 spacing), and (4) 100
sources end fire (A/4 spacing). Plot field patterns through 180 for parts (1) and (2) and

through 360 for parts

angle.

(3)

of five'graphs required,

and

effective aperture

and

(4).

Use

linear scale for field vs. polar (or linear) scale for

Also plot one graph in decibel scale for

14-65

Ae

(b)

Tabulate

field (in

HPBW, BWFN,

addition to linear scale).

directivity

A total

D, beam solid angle

Cl A

for all four cases.

A radio-telescope antenna of 500 m 2 effective aperture is directed at the zenith.

is uniform and equal

Take the ground temperature equal to 300 K and assume that half the minor-lobe
beam area is in the back direction. The wavelength is 200 mm and the beam efficiency is 0.7.

Calculate the antenna temperature assuming that the sky temperature

to 10

K.

15
PARTICLES

15-1

AND PLASMAS

INTRODUCTION

The first part of this chapter treats the motion of nonrelativistic particles in the presence
of electric and magnetic fields in vacuum. The effect of a plasma (a cloud of equal
positively and negatively charged particles; total charge zero) on the progagation of
electromagnetic waves in the presence of a static magnetic

field is

then considered,

and the phenomenon of Faraday rotation is discussed. Finally, a brief treatment is

given on magnetohydrodynamic waves (mechanical waves in a dense magnetized
plasma).

15-2

CHARGED PARTICLE

IN

A STATIC ELECTRIC FIELD

Let a particle of charge ef be placed in a uniform electric field E. Since

per unit charge (newtons per coulomb), the force F on the particle is

F = eE
t

The symbol
since e

is

E is the force

(1)

e is used in this chapter to designate the charge of a particle (instead of g)

usually employed for the charge of a particle, commonly an electron.

SEC. 15-2]

The

force

PARTICLES

is

in the

same

to the field if the charge

the particle

direction as the field if the charge

is

negative.

is

If the particle is at rest

AND PLASMAS

and opposite

positive

and the

717

field is applied,

accelerated uniformly in the direction of the

is

Newton's second law, the force on a

particle

related to

is

its

field.
According to
mass and acceleration by

F = ma

= force, N
m = mass, kg
a = acceleration, m

(2)

where F

-2

Therefore the acceleration of the particle

= -E

The

velocity v

of the particle

is

after a time

/ is

(3)

then
e

a/

= Et

(A\

'

-1

= velocity of particle, m s
= charge of particle, C
:n = mass of particle, kg
E = electric field intensity, V
t = time, s

where v
e

The

field

m _1 orNC

imparts energy to the charged particle.

If (4)

is

reexpressed as

ma = eE
ms the dimensions of force.
tb j energy
acquired. Thus
it

W=mJ

he

line integral

cifference
( 5)

(5)

Integrating this force over the distance

When

and
a

2,

yields

a-dl= -ej E-dl

of E between two points,

between the points.

moved

(6)

may be recognized
and dl

dv/dt

as the potential

\dt are

substituted,

becomes

W=m

(7)

or

where

vdv = eV

f
J

W=im(v2 -v

2
1

= eV

(8)

W = energy acquired, J
= velocity at point 2, or final velocity, m s"
velocity at point 1, or initial velocity, m s"
e = charge on particle, C
V = potential difference between points 1 and 2, V

v2
vt

or J

C" 1

718

[CHAP. 15

ELECTROMAGNETICS

If the particle starts

from

rest,

the

W = eV = imv
where v

is

Equation

the final velocity.

sional relation in SI units

Joules

zero; so

initial velocity is

(9)

(9)

has the dimensions of energy.

The dimen-

is

coulombs x

volts

meters2

= kilograms

seconds 2

the energy acquired by a particle of charge e starting from rest and passing
through a potential drop V is given either by the product of the charge and the poten-

Thus
tial

difference or

by one-half the product of the mass of the

particle

and the square of

the final velocity.

Solving (9) for the velocity gives

lleV
V

The energy acquired by an

potential difference of

is

electron (e

19
1.6 x 10"

For an electron

1.6

KT

19

')

x 1(T 19 C) in

1.6

(10)

falling

This amount of energy

J.

unit for designating the energies of particles

(ms

and

and

is

is

through a

a convenient

called one electron volt (eV).

m = 0.91

x 10

-30

kg, so that (10)

becomes
r

= 5.93xlOyV

(ms

-1

(11)

_1

Mm

It is
s"\ or 590 km s
V, the velocity of the electron is 59
an
electron.
velocity
to
large
very
a
imparts
apparent that a relatively small voltage
s~\ or about one-tenth the velocity of light.
If V = 2.5 kV, the velocity is 30

Thus,

if

V=

Mm

The above relations are based on the assumption that the particle velocity is small
compared with that of light. The mass of a particle approaches an infinite value as
above relations are
the velocity approaches that of light (relativistic effect), whereas the
is of negligible
increase
mass
the
however,
Actually,
based on a constant mass.
that of
percent
least
10
is
at
velocity
the
unless
consequence for most applications
light.

(rest

The relation between the mass m of the particle and its mass m

mass)

is

at

low

velocities

given by

(12)

SEC. 15-3]

PARTICLES

FIGURE

AND PLASMAS

719

15-1

Path of positively charged particle in

electric field.

where v
c

=
=

velocity of particle,
velocity of light

If the velocity

is

m s _1

300

Mm s"

one-tenth that of light, the mass

is

only one-half of

percent greater

than the rest mass.

If the particle has an initial velocity which is not parallel to the
field
assumed above, the particle describes a parabolic path (Fig. 15-1). The
of an electron by a transverse electric field is discussed in Sec. 15-4.

as

CHARGED PARTICLE

15-3

From

(5-9-3) the force

F on a

A STATIC MAGNETIC FIELD

current element of length dl in a magnetic field

d =
where I

IN

direction,

deflection

x B) dl

(I

(N)

is

(1)

= current (vector indicating magnitude and direction

of current),

B = magnetic flux density, T

dl = element length, m
This

is the fundamental motor equation of electrical machinery.

It also applies to
moving charged particles in the absence of any metallic conductor.
The current I in a conductor or in a beam of ions or electrons can be expressed

in

terms of the current density

J, the

charge (volume) density p, the

beam

area A, and

the velocity v by
I

= JA = pvA

(2)

Substituting (2) for I in (1) gives

dF = pAdl(vxtt)
But pA dl

dq, the charge in a length dl

of the beam.

(3)

Thus

dF = dq (v x B)
For a

single particle

of charge

e,

we have

F=
Consider

now

of flux density B.

(4)

for the (Lorentz) force

e(v

x B)

(5 )

the motion of a particle of charge e in a uniform magnetic field

The velocity of the particle is v. From Newton's second law the

720

[CHAP. 15

ELECTROMAGNETICS

FIGURE

15-2

Path of positively charged particle in

magnetic field.

on the particle
(=dvldt). Thus

force

is

equal to the product of

ma = e(y x
a

or

its

mass

and

its

acceleration a

B)

(6)

= -(vxB)

(7)

normal to the plane containing the particle path

and B. If the direction of the particle path (indicated by v) is normal to B, the acceleration is a maximum. If the particle is at rest, the field has no effect. Likewise, if the
According to

particle

path

(7),

is

the acceleration

in the

same

is

direction as B, there

is

no

effect,

the particle continuing

Only when the path or the velocity v has a component normal to B

does the field have an effect.
If a magnetic field of large extent is at right angles to the direction of motion of
undeflected.

a charged particle, the particle

is

deflected into a circular path.

region a positively charged particle

field-free

is

moving to the

Suppose that

in

right, as indicated in

and that when it reaches the point P a magnetic field is applied. The direcAccording to the cross product of v into
is normal to the page (outward).
the acceleration a is downward, so that the particle describes a circle in the

Fig. 15-2,

tion of

in (7),

clockwise direction in the plane of the page.

Radius

Let us determine the radius

(radially inward)

on the

particle

is,

by

of the

circle.

is

also given

(8)

and

(8)

by

F
Equating

the force

(7),

F=ma = evB
This force

The magnitude of

mv

=1r

(9)

(9) yields

mv

evB

(10)

~~R

mv
~eB

(ID

SEC. 15-4]

PARTICLES

R = radius

where

of particle path,

AND PLASMAS

721

m = mass of particle, kg
= velocity of particle, m s _1
e = charge of particle, C
B = flux density, T
v

Thus, the larger the velocity of the particle or the larger its mass, the
greater the radius.
On the other hand, the larger the charge or the flux density, the smaller the radius.

Frequency
path

is

The number of

revolutions per second of the particle in the circular

called the frequency /f of the particle.

eB

This frequency

(FS

~2^R~2^i

is

_1
(12)

EXAMPLE An
flux density.

electron has a velocity of 10 km s' ' normal to a magnetic

field of 0.1 T
Find the radius of the electron path and also its frequency. (See
Table 15-1

or Appendix A-2 for charge and mass of electron.)

solution

From

(11) the radius is

Jl

This

is

a very small

If the particle in the

x IP' 31 x 10*

9.1

~i.6xio-xio-' ==569nm

The frequency

circle.

/=
to

is

2.x5 7xl0-^ 2 8Xl09rS

-

above example had an

as well as perpendicular to B, the particle

"

initial velocity

would move

in

component

helical

parallel

path with the

axis of the helix parallel to B.

15-4

THE CATHODE-RAY TUBE

cathode-ray tube (CRT) is a device for observing rapid voltage

variations. In a
cathode-ray tube (Fig. 15-3) a beam of electrons is emitted from
a cathode and accelerated by an electrode A and impinges on a fluorescent screen.
By means of either a
transverse electric or a transverse magnetic field the beam can
be deflected so that
it strikes the screen at a distance
y from the undeflected position. The spot on the
screen is visible, and the particular usefulness of the cathode-ray
tube is that,

because

of the small

of the electron beam, it can follow very rapid changes in the applied
deflecting field.
This is a somewhat oversimplified description of a cathode-ray
tube
but will suffice for the following brief analysis of some of its
characteristics.
inertia

t This

is

the

same as gyro, or cyclotron, frequency

(see Sees. 15-5

and

15-8).

722

[CHAP. 15

ELECTROMAGNETICS

Electron

gun

Accelerating

Screen

^Deflecting plates

electrode

FIGURE

15-3

gun
Cathode-ray tube with electrostatic deflection. Electron fired by electron
parallel plates.
is deflected by electrostatic field between

The

positive accelerating potential

produces an accelerating

field

Ea

Va

is

This

applied to the electrode A.

From

that imparts a velocity v x to the electrons.

(15-2-10)

l2eV.
(1)

After an electron leaves the accelerating electrode,

it

maintains this velocity v x

differthis case let us consider two plates at a potential

field is a
deflecting
transverse
in
the
electron
an
of
path
ence Vd as in Fig. 15-3. The
is neglected, the elecparabola. When fringing of the field at the edges of the plates

Electrostatic deflection

tron

is

field

Ed

For

Ed for a distance / or for a time t = lfvx

the y direction, which, from (15-2-3), is
in
a
produces an acceleration y

subjected to the deflecting field

The

eVd
y

(2)

md

Thus, the electron acquires a velocity component

v, in

the

direction given

by
(3)

md v.

The

deflection angle a (Fig. 15-3)

is

= tan

then
1

= tan
t).

tan

But from the tube geometry, assuming x

>

(4)

mdv r2

(5)

2Va d
I,

the angle a is also given by a

= tan

'

(y/x) ;

so from this and (5)

Vd lx
'2Vd

(m)

(6)

sec. 15-4]

PARTICLES

-I

AND PLASMAS

723

yk

v___-

(out)

FIGURE

15-4

Cathode-ray tube with magnetic deflection.

where y

= deflection distance

at screen,

Vd = deflecting potential, V
/ = length of deflecting plates,

= distance from deflecting plates

Va = accelerating potential, V
d = spacing of deflecting plates, m
x

to screen,

Solving for the volts per meter of deflection (ratio

VJy),

2Va d

EXAMPLE
V = 1.5 kV,

(Vnr

lx

we have

(7)

A cathode-ray tube with electrostatic deflection has an accelerating voltage

a deflecting-plate spacing rf= 10 mm, a deflecting-plate
length /= 10 mm,
and a distance x = 300
from deflecting plates to the screen. Find the voltage V. required
1

to deflect the spot

solution

mm
mm on the screen.

by 10

From

field.

(7)

10-

vc

or

10

To

Neglect fringing of the

x 1Q- 2
x 30 x 10" 2

1,500
2

lOkVm-

V mnr

increase the sensitivity,

deflection,

Va

or

d should

Magnetic deflection
Fig. 15-3 be replaced

Fig. 15-4.

The

i.e., to decrease the number of volts

per meter of
be decreased or an increase made in / or x.

For

this case let the deflecting plates

by a magnetic

direction of

is

field

of flux density

and electric field Ed of

normal to the page as in

outward from the page.

Now for this case the acceleration due to the magnetic field is not in the
tion but

is

normal to the circular path of the

electron.

Assume, however, that

direc/ is

so

724

[CHAP. 15

ELECTROMAGNETICS

small that as an approximation the acceleration can be taken in the y direction.

the velocity

component

in the

direction

is

= ay t = ay

vy

Then

(8)

Thus from (15-3-7)

_evx B l__em

The deflection

angle a (Fig. 15-4)

vx

(9)

'

is

-- '-S -S "'("^
But we have also a

tan

-1

<10)

and so

0>/x),

1
Kh-^
'P
\2mV

y = xB

(ii)
>

tt

= deflection distance at screen, m

x = distance from magnetic deflecting
B = flux density of deflecting field, T
e = charge on particle, C
m mass of particle, kg
Va = accelerating voltage, V
/ = axial length of deflecting field, m

where y

field to screen,

Solving for the flux density per meter of deflection (ratio B/y),

B
y
where the

ratio

B/y

is

}_

x/V

2mVa

(12)

in teslas per meter of deflection.

we have

3.38x10-'^

For an electron

(12)

becomes
(13)

XI

= distance from deflecting field to screen, m

= axial length of deflecting field, m
Va = accelerating voltage, V

where x

EXAMPLE 2 A cathode-ray tube with magnetic deflection has an accelerating voltage

= 20 mm, and a distance x = 300 mm
V. = 1.5 kV, a magnetic deflecting field axial length /

PARTICLES

sec. 15-5]

from the

of an electron

From

solution

15-5

Find the magnetic

deflecting field to the screen.

deflect the spot

beam

AND PLASMAS

field flux density

725

B required

to

mm on the screen.

(13)

3.38

3.0

X 10" 6 x 1.500" 2
x 10- x 2x 10" 2
1

21.8mTm-

THE CYCLOTRON

The cyclotron

is

for obtaining a

a heavy-particle accelerator, invented in 1929 by Ernest O. Lawrence,

beam of

high-energy ions.

Particles such as protons, deuterons, or

alpha particles are given multiple accelerations in a resonance chamber.

to Fig. 15-5,

two hollow D-shaped copper

electrodes, or dees, are

Referring

between the pole

Target

specimen

Deflecting
electrode

Top view

(a)

Magnet

pole

L_0J

Dees

zr^.

Ion paths

South pole

Side view

(b)
:

Tient<
Filament

North pole

'M

FIGURE
Cyclotron.

15-5

726

ELECTROMAGNETICS

[CHAP. 15

The region between the pole pieces is evacuated.

vacuum chamber are omitted. The two dees are con-

pieces of a large electromagnet.

In Fig. 15-5 the walls of the

nected to a high-frequency source of alternating voltage.

In the case of deuteron

operation, ions in heavy-hydrogen gas at low pressure are produced at the center of

the chamber by electrons emitted from a filament.

on the right-hand

Starting with a deuteron at the point P, a negative potential

dee accelerates the deuteron to the

free of electric field but

that the lower pole

circle in

still

Entering the dee, the deuteron

right.

in the

magnetic

a north pole so that

is

a clockwise direction.

between the pole

field

B is upward.

If the timing

moves

it

By repetition of this process the energy of the deuteron

radius.
until

The deuteron then moves in

when the deuteron

has reversed,

field

Having acquired additional energy,

left.

Here a

reaches the periphery of the dees.

it

a region

Suppose

proper, so that

is

again reaches the gap between the dees the electric

ated to the

in

is

pieces.

it

be acceler-

will

in a circle of larger
is

increased in steps

deflecting electrode at a high

negative potential pulls the deuteron through an opening in the dee so that

it

can im-

pinge on a specimen placed outside the dee.

The frequency
magnetic

field in

in revolutions per second of a particle

a cyclotron

is

given by (15-3-12),

moving normal

eB
/==
Provided the particle velocity
Since e and

is

small

to the

i.e.,

(1)

compared with

light,

is

substantially constant.

B are also constant,/is a constant regardless of the velocity of the particle.

However, from (15-3-11) the radius of the

circular path

is

proportional to the particle

velocity.

EXAMPLE
in

Find the frequency for a deuteron

a cyclotron with a flux density

solution

From

(1) the

B=

(e

= 1.6 x

10-

19

C,

m = 3.34

x 10" 27 kg)

1.5 T.

frequency

is

2ttx 3.34 x 10" 27

This

is

the frequency required of the oscillator connected to the dees.

oscillator frequency

must be

11.4

Accordingly, the

MHz for accelerating deuterons.

The final energy of a particle is determined by the radius R and the flux density B.

From (15-3-11)

R=

eB

or

(2)'
v


PARTICLES

SEC. 15-5]

The energy

W of the particle

AND PLASMAS

is

W=imv 2

or

J2W
/

'

(3)

Equating these relations for v and solving for the energy

W of the particle yields

W-\*2
2
m
EXAMPLE
B = 1.5 T.

727

(4)

A cyclotron has a maximum working radius R = 500 mm and a flux density

Find the energy which can be imparted to deuterons.

solution

From

energy

(4) the final

is

10-"

1.6 x
x 1.5)'
= 216pJ
W = (0.5 x2x3.34x10-"

Since

Energy

..

in electron volts

in joules

= energy
19

this result in electron volts is

or an energy of 13.5
If the voltage

of revolutions

is

MeV.

applied between the gap of the dees

reduce the dispersion of the

The

as possible.

a given energy
revolutions

is

small, a large

required before the particle reaches the periphery.

beam

total energy

it is

desirable to

make n

small and hence

W acquired in n revolutions W= 2nVe.

is

W as determined by

(4)

and for a particular voltage

V,

number n

However, to

the

V as large
Thus, for

number of

is

W
where

is

in joules.

If the energy

expressed in electron volts

WcV

= ^xl.6xl0-"

= charge of particle, C
F=dee voltage, V

where e

is

(r)

(6)

728

ELECTROMAGNETICS

[CHAP. 15

B(out)

Screen

FIGURE

15-6

Mass spectrograph.

15-6

THE MASS SPECTROGRAPH

mass spectrograph

different mass.

is

a device for separating particles of the same charge but

velocity v into a uniform magnetic field of flux density B.

strike the fluorescent screen or

of entry, as suggested.
of masses

known

Referring to Fig. 15-6, the particles are injected with a

and

m2

It

Particles with larger

mass

photographic plate at a greater distance from the point

can be shown that two particles of the same charge e and

injected with the

same

by

velocity v are separated at the screen

a distance d given by

d=
where

15-7

2v/Be

m2 m

(1)
x

m 2 > mf.

TABLE OF CHARGE AND MASS

FOR COMMON PARTICLES

The charge and mass for a number of common particles

The mass given is the rest mass, or mass at zero velocity.

Table 15-1

are listed in Table 15-1.

CHARGE AND MASS OF PARTICLES

Ratio elm,

Mass m, kg

Particle

Charge

Electron
Positron

-1.602 x 10-"
+ 1.602 x 10- 19

Neutron
Proton (hydrogen nucleus)
Deuteron (heavy-hydrogen
nucleus)

Alpha

e,

+ 1.602 x

10

" 19

Ckg-

9.11

x 10- 31

-1.76 x 10 11

9.11

x 10" 31

+1.76X

1.6747
1.6725

x 10" "
x 10 ~ 27

10 u

+9.6

x 10 7
x 10 7

+ 1.6

x 10" 19

3.34

X 10-"

+4.8

+3.2

x 10- 19

6.644

x JO" 27

+4.81 x 10 7

particle (helium

nucleus)

sec. 15-8]

PARTICLES

FIGURE

AND PLASMAS

729

15-7

Relation of direction of wave propagation to coordinate system. The direction of

wave propagation

plane at an angle

is

in the

xz

with respect to the z

axis.
The electric field E of the wave
lies in a plane perpendicular to the direction of wave propagation.
steady
magnetic field B is in the z direction.
<f>

15-8

PLANE WAVES IN AN IONIZED MEDIUM

IN THE PRESENCE OF A MAGNETIC FIELDf

In Chap. 10 wave propagation in lossless and conducting media was discussed.

Although an ionized medium may be classified as a conducting medium, the condition

a > a>e assumed in Chap. 10 may not apply. Furthermore, an ionized medium becomes anisotropic (propagation variable as a function of direction) in the presence of
a steady magnetic

This greatly complicates the situation compared with that

field.

considered in Chap. 10, where the

same

in all directions).

for a plane

wave

in

The medium

medium was assumed

to be isotropic (propagation

Consequently, a more general development will

an ionized medium

to be considered

in the presence

now be made

of a steady magnetic field.

a plasma;

i.e., it consists of equal positive and

would be described as cool (thermal energies
small) and tenuous (collisions unimportant).
In such a plasma only the interaction
of the negative charges (electrons) with the wave will be considered, the positive ions

is

negative charges (total charge zero) but

being too massive to interact appreciably.

Referring to Fig. 15-7,
is in

the z direction; that

t

is,

it

will

\B\

be assumed that the steady magnetic flux density

= Bz

This simplifies the analysis without any loss

This section and Sees. 15-9 and 15-10 are adapted from

sec.

5-4 of

J.

"Radio Astronomy," McGraw-Hill Book Company, New York,

straightforwardness of the development of this section is due to Prof. R. G.
jian, Lecture Notes for Electromagnetics Course, Ohio State University.

D. Kraus,

The
Kouyoum-

1966.

730

[CHAP. 15

ELECTROMAGNETICS

The

of generality.

direction of propagation of the

wave

the angle between this direction and the z axis equal to

ated with the
tion,

i.e.,

wave may have any

may

The

is

taken in the xz plane with

The

tj>.

electric field

E associ-

orientation perpendicular to the direction of propaga-

be polarized in any manner.

force equations involving a charged particle of charge e

presence of a wave and magnetic flux density

Fm =

e(v

and mass

are

x B)

(1)

Fe = eE

(2)

dv

ma

Jb

m in the

==

(3)

dt

where F

= total

force,

F m = magnetic force, N
Fc = electric force, N
e

= charge

m mass

of particle,

of

kg

particle,

= velocity of particle due to the wave, m

B = steady magnetic field, T
-1
E = electric field of wave, V m
2
a = acceleration of particle, m s~
t = time, s
v

In (3) (Newton's second law)

to the velocity of light;

it is

i.e., relativistic

Combining the force

s"

assumed that the

velocity

is

small compared

effects are neglected.f

relations [(3) equals (1) plus (2)] gives

F=

e(E

x B)

=m

dv
(4)

Jt

Dividing by e to obtain the force per unit charge yields

mdv

F
-=E
+v x B=-
e dt

(5)

A more

general statement of (3)

is

dv
-Lr^ = m +
= </(mv)

where mv
to (3).

is

the

momentum

< c, dmjdt = 0;

dm

dt

of the particle.

dt

For

(3a)

dt

so (3d) reduces

PARTICLES

SEC. 15-8]

Expressing this vector equation in terms of

its

AND PLASMAS

731

three (rectangular) scalar components,

we have

Fx

= Ex +

7e =

Vy

Bz =

mdv
-^

< 8>

= 77T
e at

Assuming harmonic motion of the wave and charged

component velocities, we have

(6)

particle

j(mwle)E x + B2 Ey
/"22/_2\
2
BD 2 - (m
m 2 /e 2

or angular frequency

its

V)

"-^
w = radian

for

where

and solving

(1I)

2nf, rad

_1
s

/= frequency, Hz
From
particle to
direction).

these relations

move

it

can be inferred that the wave

and

vy

become

infinite, indicating

at this frequency.!
at

which

a>

= -p-

(12)

a strong interaction of the wave with the particles

This corresponds to gyro resonance.

this occurs tog

Calling the radian frequency

we have

.-*
where

io

is

cause the charged

For the condition

B*

vx

will

in a helical path with axis coincident with the magnetic field (z

the (radian) gyro frequency

The

effect

(13)

= 2nf, rad s _1 ).

of collisions between particles

is

neglected.

732

ELECTROMAGNETICS

This
(o

is

is

[CHAP. 15

the resonance condition in a cyclotron (Sec. 15-5),

also called the cyclotron (radian) frequency.

and for

Introducing (13) into

this
(9)

reason

and

(10)

yields

vx

=-e

is

the

movement of
(16) in terms

e co g Ex +jcoE y

number of charged

(16)

particles per unit volume.

It is

assumed that a
Expressing

x H)x =jcoe Ex + Nevx

(17)

(V x H), =jcoe

Ey + Nevy

(18)

(V x H) z =jcoe

Ez + Nevz

(19)

vy

and

vz

/r;

(V x H),

-co

from

is

much

and

(14), (15),

(11) into these

Ne 2<oga E y

"1

+
- co r)] m(co 2
2

Ne 2

"\_

2
- co T
)\

(21)

(22)

mco 2 )

greater than the gyro frequency

identical in form.

m(co

(20)
v

-co 2 )

Ey [^1 +

+Jcoe

Ne 2

= jcoe Ex [\
\

If the frequency

m(co g

Ne 2 co3g E*x
= 2
2
m(co

(V x H),

Ne
-

(co

>

co ), all
g

three relations

Further, if

Ne 2
2

(23)

or plasma, frequency occurs, for which the permittivity and the index of

refraction of the
(

(15)
v
'

we obtain

critical,

co

of its three rectangular components, we have

,
_ T.
(V x H)* =;coe j 1

co

H = J +jcoD = Ne\ +jcoe

Introducing the values of vx

become

the charged particles constitutes a conduction current.

(V

relations,

(14)'

equation we have

X
where

'
\
co
g

y=
From Maxwell's

E +jcoEx
co i2

co

medium become

zero.

Designating the

critical

frequency as

co

= 2nf ), we have
<o

,
2

2
= Ne

(24)

PARTICLES

sec. 15-8]

AND PLASMAS

(25)

or

where

733

= critical radian frequency = 2nf rad s~

= critical frequency, Hz
JV = density of charged particles, number m"
e = charge of particle, C
m = mass of particle, kg
-1
e = permittivity of medium = 8.85 pF m

oo

For an electron

(25)

becomes
(26)

where f

= critical

frequency,

N = electron density,
At or below

Hz

number

m~ 3
wave

the critical frequency the

typical value of JV for the earth's ionosphere

which from

V x

is

totally reflected

we may

H =jox.

by the medium.

12
about 10 electrons per cubic meter,

(26) corresponds to a critical frequency of 9

Introducing (24) into (20) to (22),

where e

is

is

MHz.

write in general, thatf

(27)

the tensor permittivity, or

en

-Ju

13

J 2 i

22

23

31

32

33.

-j(O
(1

jto

or
w(a>g

r
2

(Q g e

-a> 2 )

2V

a>(co g

/
I,

(28)

CO
g

o2)

<o
co

2
z

(29)

V+aS-atr

('"*>
t

V X H = oE +jaie E we could write V X H = o

E, describing the medium

or we can write VxH=y<ui'E,
describing the medium in terms of an effective tensor permittivity e. In the present
instance, the latter procedure has been chosen, as given by (27).

From

in terms

of an

effective tensor conductivity a,

734

ELECTROMAGNETICS

[CHAP. 15

Equation (29) has the form of a skew-symmetric tensor (elements symmetrical about

main diagonal, upper

magnetic

field

(B

= 0)

left

to lower right, but signs reversed).

or where

a> P- co
g

In the absence of a

(29) reduces to

'(-*>
(-*>

(30)

(-*>
where

(31)

and

(32)

The quantity 1 - (w 2 /co 2 ) is equivalent

The index of refraction is then

to the relative permittivity

c,

of the medium.

<-*4-'-M$
When/>/
.

r\

is less

enter a denser

wave

is

the index of refraction

is

than unity, so that refraction

medium.

When/</

not propagated but

When

is

However,

propagated.

opposite to that which occurs

the index of refraction

is

when waves

imaginary and the

reflected.

a cloud of charged particles

plasma frequency.

and the wave

real
is

(33)

is

perturbed,

it

may

tend to oscillate at the

Certain types of radiation from the solar atmosphere or corona

are believed to be due to such plasma oscillations?

The

electron density in the solar

corona decreases with height, and from (26) the plasma frequency will also decrease.
Hence, at a given frequency/, radiation is received on the earth only from the corona
at heights

above the height for which f=f

Actually the radiation appears to

mostly from a relatively thin layer just above the height for which

When

These comments apply to the quiet sun.

as flares, occur,
jets or

more dynamic phenomena may be

f=f

come

disturbances on the sun, such

involved, in which high-velocity

rise into the solar corona.

Plasma oscillations of electron clouds
mechanisms appear to be responsible for relatively narrow-band

shock waves

excited by such
radiation,

which

drifts

downward

in frequency as the disturbance rises

through the

corona into regions of lower electron density.

Returning to the general situation (with

VxE=

j= 0), let

-/ayiH

us take the curl of

(34)

PARTICLES

sec. 15-8]

for the case

from

\i

a wave equation in E.

is

zero), let a solution

/?

a)

(35)

Neglecting the effect of collisions (attenuation assumed

be assumed of the form

e Km

E=E
where

we have

VxVxE =
This

735

and introduce V x

(ferromagnetic effects assumed negligible)

fi

(27) so that

AND PLASMAS

= phase constant (in

= distance measured

'-

(36)

r direction)

in direction of wave propagation (see Fig. 15-7)

an exercise (straightforward but tedious) to show that (36) is a solution

It is left as

provided
2

= co 2 fi {(e n 2 - e 12 2 - eu e 33 )sin 2 + 2e n e 33 [(en 2 + 4el2 2 e i3 2 cos 2 " 2 }/2[(Cll - e 33 )sin 2 + e 33

cj>

where

n
mg

<j>

e 12

is

the angle between

etc.,

so that

Ci

When
is

(or z)

c 33

<o

the

it

- e u e 33 ) 2 sin4

</>

(37)

When

u> is

much

and where
larger than

but not so large that e 12 (or 21 ) can be neglected, (37) reduces to

wave propagates

P
(39)

direction (see Fig. 15-7),

are as given by comparing (28) and (29).

called longitudinal propagation

From

and wave

e l2

parallel to the

and

magnetic

Wy/fioien

= 0), the condition

(39)

consists of

components of opposite sense of rotation.

When the wave propagates perpendicular to

and

(</>

e lz )

may

can be shown that the wave

transverse propagation

field

(38) reduces for this case to

B ($ =

two

circularly polarized

90), the condition is called

(38) reduces for this case to

(40)

when

the electric field

of the wave

when E

is

magnetic

parallel to B.

field

is lirfcarly

/n

a> y

polarized perpendicular to

were present.

This wave

to as the extraordinary ray.

to

(41)

e 33

In this latter case propagation

is

the

same as though no

sometimes referred to as the ordinary ray,

is

while in the case described by (40), where

B and

is

perpendicular to B, the wave

is

referred

736

ELECTROMAGNETICS

More
and

[CHAP. 15

general relations than the above can be deduced for quasi- longitudinal

For the former

quasi-transverse conditions of propagation.

small enough to permit sin

unity.

For the

latter case

We

terms to be neglected.

and

<j>

<j>

is

sin*

<f>

case,

<t>

is

considered

terms to be neglected although cos

<f>

not

is

considered to be sufficiently near 90 to allow cos

</>

then have for the quasi-longitudinal case that

+ 12 cos

wV/tofei

(42)

<j>)

ierse case that

o>Jlh Kit

- (1 +

These propagation relations are summarized

Table 15-2

in

1)

^- sin

Table

(43)

<f>\

15-2.

PROPAGATION RELATIONSt

Quasi-longitudinal

propagation

(<j>

small)

ft

toV|Llo(ll

jj COS

^)

Quasi-transverse

propagation (^ near 90)

Longitudinal propagation

Transverse propagation

Ordinary ray:

(||

f}=

to

(J_ to

B;

<j>

<j>

to

/fioUn

/?= a>Vju (eu

== 0)

(1 l)^-sin^

e 12 )

= 90)

E B
E B

/?

II

= w V^,

et

Extraordinary ray:

P =*>

J*>('-')

t See text for assumptions involved.

2
Note: , j
[u
{1
/vV " 2 )]}o

= +
= <o <aa2 e /u>(aig2 2)
-1
a) = radian frequency of wave = 2nf, rad s
where /= frequency of wave,
cu, = radian gyro frequency = (e/m)/B
a>0 = radian plasma frequency = eV JV/e ">
B = magnetic flux density, T
e = charge on panicle, C
m mass of particle, kg
~
e = permittivity of vacuum = 8.85 pF m
= phase constant = 2nl A
A = wavelength, m
~'
Ho = permeability of vacuum = 400ir nH m
angle between wave direction and magnetic field B

e, j

to

/3

<f>

Hz

PARTICLES

15-9]

AND PLASMAS 737

(Wave

FIGURE

1)

(Wave 2)

15-8

Resolution of linearly polarized wave

into

two

and

2)

circularly polarized waves (1

of opposite rotation direction.

FARADAY ROTATIONf

15-9

A linearly polarized wave may be regarded as the resultant of two circularly polarized
waves of equal amplitude and opposite sense of rotation (see Chap. 1 1). If the two
circularly polarized waves have different phase constants, the plane of polarization
of the resultant linearly polarized wave rotates as the wave propagates.

Thus,

consider two circularly polarized waves traveling out of the page, as in Fig. 15-8 for
the quasi-longitudinal case
is

(<t>

The elemental angular

small).

rotation of one

wave

given by

d9 l

= p~

dr

(1)

d9 2

= p+

dr

(2)

and of the other wave by

where dr

P~ and
is

is

+
j8

and

the elemental propagation distance perpendicular to the page,

respectively.

Now from

Fig. 15-8

tan0

E
E

Now

and where

are as given by (15-8-42) for the cases where the sign in the parentheses

assuming

sin 0j

cos

E 2 sin
+ E 2 cos

(3)
2

E = E2 we have
t

tan0

The name

sin

cos

sin
+ cos

(4)
2

derives from Faraday's early investigations of the phenomenon, in which

he observed the rotation of the plane of polarization of light passing through a
crystal with an applied magnetic field.

738

ELECTROMAGNETICS

[CHAP. 15

from which

= tan

tan

and the

resultant rotation angle

is

(5)

then
n

01-022

= -L

and from

(1)

and

(6)

(2)

*>-r^*
( 7)

Introducing (15-8-42) and the values of e lt and e l2 from (15-8-28) and (15-8-29) for
co

> cog

co

co

and small

we obtain approximately

</>,

de

The

total

cos

2*
2 3
Sn
c e

cj>

that

dr
(rad)

m2

Faraday rotation for the quasi- longitudinal case

g 3l2
r-j-^3

%n c

= charge

where e

Ne BX

of particle,

m = mass of particle,

(8)

(4>

small)

then

is

m j Jf

N5 cos

<j>

dr

(rad)

(9)

kg

= 2nf= radian frequency, rad s -1

_1
e = permittivity of vacuum = 8.85 pF m
c = velocity of light = 300 Mm s"
N = particle density, number m -3
B = magnetic flux density, T
= angle between B and direction of wave propagation,
r = distance in direction of propagation, m
co

<j>

If

B and

are constant, (9) becomes

<j>

Q=

If

B and

rad or deg

are

<j>

BX 2 cos<b

,.*,
on
c e

known, a measurement of $

Ndr

(rad)

( 10 >

at a wavelength 1 permits a determination

of the total number of charged particles in a column

m2

between the source and

observer given by

N,= J(Ndr

(11)

where N

is

total

number of particles

in a

column of length

and cross

section of

2
.

SBC. 15-9}

PARTICLES

For the

quasi-transverse case

d&

and for the

total

where the symbols are as

rotated during

and

<j>

we have

near 90)

Ne*X 3 B 2 sin 2

<t>

dr
(rad >

3
3 4
,o,3>
32n
c m e

(12)

Faraday rotation

Equations

(<f>

AND PLASMAS 739

(9)

and

^
32n

c4l3
*
3

4.

c*m

^
3

NB 2

sin

</)

dr

(rad)'
v

(13)

listed following (9).

(13) give the angle

by which a linearly-polarized wave

passage through a magnetized plasma for the two cases,

its

<f>

is

small

near 90.

In radio-astronomy observations of the polarization of radio sources, the position

angle
is

of the electric

field

vector of the linearly polarized

component of the radiatibn

usually measured relative to north with 9 increasing counterclockwise, or to the

east.

Since the position angle of the electric field vector varies as the square of the

wavelength [see

(9)],

fl-vs.-A

range of wavelengths.

By

graph should show a

straight-line relationship over

polarization angle at the source, or intrinsic polarization, can be determined.

slope of the line

is

positive

(= 0/A 2 ) has been

when the magnetic

ambiguity of nn (n
are

made

= integer)

sufficiently close in

extrapolating the straight line to zero wavelength the

called the rotation measure.

field direction is

The

toward the observer.

in the position angle,

which

frequency to ensure that

is

resolved

changes

less

The

rotation measure

if

There

is

an

measurements

than n/2 between

measurements.

With the assumption of a constant electron density and angle

the
Faraday rotation A0 for the quasi-longitudinal magnetic case (9) reduces for
<f>

total
elec-

trons tof

A0
where

= 2.6

N = electron density,

x 10" l3 NBX 2 cos

number

(rad)

(14)

m -3

B = magnetic flux density, T

X = wavelength, m
= angle between wave direction
Ar = path length, m
<t>

Ar

and B, rad or deg

For a transmitter

in an artificial earth satellite traveling toward or away from the

observer along a line nearly parallel to B it is convenient to replace A0 by ir and Ar

by v AT, where v is the satellite velocity in meters per second and Ar the time in
seconds between nulls in the signal as measured with a linearly polarized receiving
antenna. Assuming that B is known, the average N can then be determined. It is
assumed that the fading (nulls) is entirely due to Faraday rotation effects and not to
rotation of the satellite (with a linearly polarized antenna).

740

ELECTROMAGNETICS

FIGURE

[CHAP. 15

15-9

Geometry for rectangular slab in a medium of conducting fluid. Motion of

the slab to the right pulls the magnetic
field

with

field line.

it

as suggested by the dashed

(See footnote, p. 741.)

MAGNETOHYDRODYNAMIC WAVES

15-10

In Sec. 15-8 no motion of the

particles

was considered.

medium

Suppose

other than a harmonic motion of the charged

now

that a layer or

column

in a

conducting

moved or displaced. No incident

waves are considered, but let us study the effect of this motion of the medium in generating a disturbance, or wave. This situation has been investigated by Alfven. f The

medium with applied magnetic field

is

mechanically

waves produced by the motion are referred to as magnetohydrodynamic waves

(MHD

waves) or Alfven waves.

It

should be noted that the

from that considered in

are important.

It

Sees. 15-8

medium
and

being considered here differs significantly

15-9, because

might consist of a conducting

it is

fluid,

much denser,

so that collisions

such as mercury; or

it

might

consist of a very dense, hot plasma.

Let a slab in a conducting

the y direction.
field

is

direction,

The

slab

is

fluid,

as in Fig. 15-9, be

infinite in extent in this

applied in the z direction.

Since v

an

electric field

Fig. 15-9.

velocity v in

we have from

E=
that

is

moved with a

y directions. A steady magnetic

in the y direction and B in the z

This

E will

field

vxB

(1)

be induced by the motion in the x direction, as suggested in

produces a current in the slab in the x direction which has

return paths through the

medium above and below

the slab.

its

This current, in turn,

produces forces which tend to impede the motion of the slab but which impart an
t

H. Alfven, "Cosmical Electrodynamics," Oxford University

Press, 1950.

PARTICLES

SEC. 15-10]

AND PLASMAS

741

and below the slab in the y direction. Thus, the original

motion of the fluid, which propagates as a wave in the direction of the

acceleration to the fluid above

motion

initiates

magnetic

field.f

The

pertinent equations are

VxH = J + ?
at

(2)

VxE=4B

(3)

8t

(T(E

v x B)

(4)

B = fiU

(5)

= G + -(JxB-\p)
^
dt
p
= velocity of displacement, m s _1
J = current density, A m~ 2
-3
p = mass density, kg m
-1 -2
p = pressure, kg m s
G = parameter involving nonelectromagnetic forces,

(6)

where v

Equations (2) and

(3) are

Maxwell

equation characterizing the

curl equations, (4)

is

m s -2

a current equation,

(5) is

an

medium (constitutive equation), and (6) is a hydrodynamic

equation with the dimensions of acceleration (Navier-Stokes equation).

Consider the case of plane waves in an incompressible

From

(4)

we have

J_
a

From

the geometry

Jy

= JZ = 0;

Ey = Ez = 0. From

(3)

and

dz

Ex in (9)

(7)

-vy B

(8)

dH
= -"-37y

(9)

"

yields

dHy

vxB

(5)

dEx

Introducing (8) for

E+

so

,=
and

fluid (density constant).

that

Bo

dJ

dvy

dz

a dz

(10)

net result is as though the magnetic field is frozen into the slab and its motion
to the right pulls the magnetic field B with it as suggested by the dashed field line in
Fig. 15-9. This displacement of B then propagates vertically (in z direction).

The

742

ELECTROMAGNETICS

[CHAP 15

we have from

Neglecting G,

(6)

dvy

dt
It is

in (1

--Jx Bn
,

(ii)

assumed that v> can have no component perpendicular to

we have (neglecting dD/dt in comparison with J)

Introducing (2)

z.

1),

dt~p
Eliminating v, between (10) and (12),

we

(12 >

dz

obtain a wave equation in

for a

wave

propagating in the z direction, or

^, = V^, +
dt

For

up

sufficiently large conductivity the last

wave, or Alfven velocity,

dz

y
2

(13)

ficrdtdz

term

may be

neglected.

The

velocity of the

then given by

is

Vw

(14)

= velocity of wave, m s" 1

B = magnetic flux density, T
_1
H = permeability of medium, H m
3
p = density of medium, kg m~

where v

It is

of

interest to evaluate (14) for the case

p =

200mgm~ 3
400n nHm" yields a velocity of 6.3 km s _1
appropriate

values

are

photosphere

of the solar photosphere, where

=100mT.

Taking h = h =
magnetohydrodynamic wave in the
In the solar corona at about one solar radius

and

for a

(visible surface of sun).

above the photosphere the value of B is less but p is much less, and a
magnetohydrodynamic wave velocity approaching the velocity of light is obtained.

PROBLEMS
*

15-1

A particle with a negative charge of 10" l7 C and a mass of 10" " kg is at rest
If a uniform electric field E =
kV m" is applied for l'/ts, find (a) the

in a field-free space.

velocity of the particle

a magnetic

15-2

a magnetic

field .8

and

(b) the radius

of curvature of the particle path

= 2 mT with this velocity moving normal to B.

if

the particle enters

cathode-ray tube has electrostatic-deflection plates as in Fig. 15-3 plus

between these plates. The magnetic field is oriented parallel to the

field

electrostatic-deflection field.

d = 20 mm, and

The tube has an

( and magnetic

a plate length

V = 1 kV, a plate spacing

= 20 mm. Find (a) the deflec-

accelerating voltage
field length) /

PARTICLES

PROBLEMS

tion-plate voltage

and

(6) the

B required

magnetic flux density

AND PLASMAS 743

to deflect the electron

beam

mm in the y direction and 100 mm in the z direction (perpendicular to page) at a screen

placed at a distance x = 500 mm from the deflecting plates.
*
15-3 The cathode-ray-tube screen of a television receiver measures 400 mm (horizonThe electron-beam spot scans 525 lines per frame and 30
tally) by 300 mm (vertically).
50

and

(a) the horizontal spot velocity in millimeters

Find

frames per second.

(b) the total distance traveled

of magnetic flux density

required

by the spot per second,

if

(c)

What

is

per microsecond

the rate of change

the tube parameters are as in Prob. 15-2?

deuteron accelerated through a potential difference V enters a region of

uniform magnetic flux density B moving normal to B. If the radius of the deuteron path
15-4

(a)

mm, what is the required relation between B and VI

20

(a)

but the particle

id)

Compare

is 1

is

a proton, find the radius of

its

path,

(b) If
(c)

V and B are the same as in

Repeat for an alpha

particle.

the kinetic energies of the three particles.

maximum energy (in MeV) for alpha particles in a cyclotron with

maximum usable radius of 450 mm. The flux density B = 1 .2 T in the air gap. (b) Repeat
{d) How many revolutions
(c) Repeat (a) if deuterons are used,
(a) if protons are used,
does each particle make if the peak potential between the dees is 9 kV?
15-6 (a) What is the ratio of the path radius of a proton to that of a deuteron if both
enter a magnetic field B normally with the same kinetic energy ? (b) What is the ratio of the
15-5

Find the

(a)

path radius of an alpha particle to that of a deuteron?

15-7 Find the radius of curvature for the path of a proton with velocity v
*

moving perpendicular
(b)

to a magnetic field with (a)

B = 100 /xT (earth's

15-8

magnetic

5-keV proton

with respect to B.

is

field), (c)

B=

injected into

10

B = 5 mT
nT

Find the following parameters

(solar-corona magnetic

(interstellar

a magnetic

field

magnetic

15-9

The chimney

(a)
1

by 0.5 m.

flue

/*g

and a charge of 10"

precipitate the particle against

assume
(0.25

that

when

B = 500 mT at

an angle of 85

of the resulting helical path:

electrostatically.

15-10

a typical particle has

Take

an electrode?

the

(b)

smoke

it is

velocity v

is

required to

= m
1

s"

and

at the greatest possible distance

Design a precipitator for a 1 -m-diameter

capable of precipitating solid particulate wastes being exhausted

s" *. Assume that the particles have masses
with a velocity of 3

flue

kg min
of lfig each. The flue is 20
226(3) :47 (March 1972).
at the rate of 10

If

C, what voltage between the electrodes

the particle enters the electrode space

chimney

(a) radius,

and meters per

smoke

m) from the electrode to which it will move,

cylindrical steel

field),

of an industrial factory has a rectangular (interior) cross

1
square are placed against the long sides of the

Flat electrodes

flue to precipitate solid particles in the

a mass of

field).

per second
(6) distance between turns of helix, (c) circular velocity in radians
period.
and
(e)
per
second,
meters
in
axial
velocity
second, (d)

section of

= 10 Mms"

'

tall.

See, for example, A.

(a) In Millikan's oil-drop

D. Moore,

Electrostatics, Sci.

experiment a tiny droplet of oil of mass

Am.,

m is suspended
E

between two
in equilibrium in the gravitational field (of acceleration g) by an electric field
magnitude
the
show
that
charges,
electron
droplet
has
the
If
capacitor
plates.
horizontal

of the charge e of an electron

is

given by e

be determined by the experiment?

= mglNE.

(b)

How can the polarity of the droplet

744

ELECTROMAGNETICS

The

15-11

TCHAP 15

alternating electric field of a passing electromagnetic

wave causes an
This oscillation of the electron makes it equivalent
that the ratio of the power scattered per steradian to the incident

electron (initially at rest) to oscillate.

to

a dipole radiator.

Poynting vector
electron
field

Show

e 2 sin 6/4)*, where e and

m are the charge and mass of the

and 6 is the angle of the scattered radiation with respect to the direction of
the electric
given by

is

(fi

of the incident wave.

Such reradiation

15-12

is

This ratio times

Thompson

called

4tt is the

radar cross section of the electron.

scatter.

ground-based vertical-looking radar can be used to determine electron

by means of Thompson scatter (see Prob. 15-11). The

densities in the earth's ionosphere

scattered-power radar return

is proportional to the electron density.

If a short pulse is
transmitted by the radar, the backscattered power as a function
of time is a measure of the
electron density as a function of height.
Design a Thompson scatter radar operating at
430
capable of measuring ionospheric electron densities with 1-km resolution
in height
and horizontal position to heights of 1 Mm. The radar should also
be capable of detecting
a minimum of 100 electrons at a height of 1 Mm. The design should specify
radar peak
power, pulse length, antenna size, and receiver sensitivity. See W.
E.

MHz

Backscatter from the Earth's Ionosphere,

(December

IEEE

Trans. Antennas Propag.,

Gordon, Radar
AP-12:873-876

1964).

15-13

Right-handed helical-beam antennas are used for communication

between
an earth base and a space vehicle to eliminate fading due to vehicle spin and Faraday
rotation
in the earth's ionosphere.
If the space vehicle is transmitting at 1 50 MHz and traveling
without spin toward the earth base at 7 km s" \ what change in frequency
results? The trans-

mission path

is

substantially parallel to the earth's magnetic field.

the electron density

N = 10 m -

B= 100 /iT and

Take

15-14

Layers are said to exist in the earth's ionosphere where the ionization
gradient
waves back to the earth. [Although the wave actually may be
bent gradually along a curved path in an ionized region of considerable
thickness, a
is

sufficient to refract radio

useful

simplification for

some

situations

is

to

assume

that the

wave

is

reflected as

though from a

horizontal perfectly conducting surface situated at a (virtual) height h.]

The highest frequency
at which this layer reflects a vertically incident wave back to
the earth is called the critical
frequency f
Higher frequencies at vertical incidence pass through. For waves at
oblique
.

incidence (<f>>0 in Fig. P15-14) the

communication on the earth

The

frequency f

critical

is

is

maximum

usable frequency

MUF =/ /cos ^,

given by

(MUF)

where

</>

for point-to-point

= angle

a function of the electron density, which in turn

is

of incidence.
a function of

and other factors. Both/ and h vary with time of day, season, latitude,
and phase of the 11-year sunspot cycle. Find the MUF for (a) a distance d=
1.3
by
solar irradiation

F2

= 325

layer (h

distanced
</=

Mm

1.5

Mm

km)

reflection with

Mm by F

by sporadic

F2

layer electron density

N=6 x

10 11

m" 3

(6)

= 275 km) reflection with N = 10 12 m- 3 ,and(c)adistance

{h = 100 km) reflection with N = 8 x 10 11 m" 3
Neglect

layer {h
layer

earth curvature.

15-15

Referring to Prob. 15-14,

angle of incidence

= 300 km

</>

for the

F2

show

layer {h

and the electron density

that

= 300

due to earth curvature the maximum

km)

N = 10 12 m"

is

all

about

74.

If the

F2

layer height

along a great-circle path around the

PARTICLES

To communication

satellite

Ionosphere

Ionospheric

AND PLASMAS 745

layer

Transmitter

N Earth's surface

FIGURE

Earth's surface

FIGURE

PI 5-14

Communication path

via reflection

signal,

(c)

What

is

P15-16

Communication

from

ionospheric layer.

world, find (a) the

Receivers

path

via

stationary

relay satellite.

MUF

and

(6) the

number of hops required

for

an around-the-world

the percent delay of the round-the-world signal with respect to a circular

path at the earth's surface?

(d)

What

is

the transmission time for the round-the-world

signal (in milliseconds) ?

15-16

40

Mm.

Stationary communication (relay) satellites are placed at heights of about

This

is

far

above the ionosphere, so that the transmission path passes completely

Since frequencies of 2 GHz and above are
in Fig. P15-16.

through the ionosphere twice, as

usually used, the ionosphere has

widths.

If the

little effect.

The high frequency

ionosphere consists of a layer 200

with a uniform electron density

usable frequency,

MUF)

N=\0 11 m~ 3

and

minimum

(6) to the

15-17

find the lowest frequency (or

(c)

minimum

satellite (a) for vertical

Discuss the relation of answers to (a)

usable frequency for a ground-based radio telescope.

we were interested in the maximum usable frequency and in

minimum usable frequency. Does the maximum usable frequency of one

In Prob. 15-14

Prob. 15-16 in the

case equal the
15-18

which can be used with a communication

incidence and (6) for paths 30 from the zenith,

also permits wide band-

km thick between heights of 200 and 400 km

minimum

A 7-MHz

Must a receiving station

usable frequency of the other case ?

radio station in San Francisco transmits with vertical polarization.
in

Tokyo use vertical

Explain reasons

polarization for best results?

for your answer.

15-19

Compare communication between two points on the earth using ionospheric

between 3 and 30 MHz) and using transmission through the ionosphere

reflection (frequencies

via relay satellite (frequencies of 2

(d) 10

Mm, and

time, cost, etc.

(e)

20

GHz and up) at distances of (a) 50 km, (6) Mm, (c) 5 Mm,

Mm. Compare

reliability, fading,

bandwidth, power, signal delay

746

ELECTROMAGNETICS

Assume

15-20
density

N=

magnetic

'

m-

[CHAP. 15

the ionosphere consists of a

at heights

14-MHz wave

wavelengths) for a

B=

and that

field is vertical

(e)

What

B=
1

is

further that the earth's

the electrical-path length (in

through the layer at vertical incidence and

Consider both modes of propagation for

km?

200

(d)

How can

a path length

(in

wavelengths) be less in the denser

the Faraday rotation in turns over each path (1 turn

is

100 juT,

B = 5 mT, N =
B = 10 mT; and

'

m~

7V=10 12 m~ 3 medium is fully ionized nitrogen

,

3
,

medium is

fully ionized

{d) liquid mercury, p

15-22

Prove

(15-8-9), (15-8-10),

15-23

Prove

(15-8-29).

15-24

Prove (15-8-37).

15-25

Prove

15-26

Design a

= 2w rad).

Note

(15-9-8).

pump

1.4

and

that

hydrogen

x 10* kg

(c) earth's core,

m~ 3 B =
,

(a) the earth's

corona,
~ 3
0* kg
,

(b) the solar

T.

(15-8-11).

VfTS = + (8/2),
1

when S

<

1.

mercury using no moving parts which can move

of mercury per second against a head of 2 m of mercury.

liter

What

Calculate the velocity of an Alfven wave for the following cases:

15-21

ionosphere,

(a)

layer with uniform electron

Assume

Calculate the ratios of the above values to the path length (in wavelengths)

(c)

over 200 km in free space,

medium?

100 ^T.

(a) traveling

(b) traveling horizontally in the layer for

each case,

200-km thick

between 200 and 400 km.

A plasma
= 30 Mm s -

15-27
velocity v

'

for liquid

travels through a square tube of 100

.

The plasma has an

electron density

by 100

mm cross section

JV=10 17 m" 3

If a

at

magnetic

field

B=lmTis applied transversely to the tube, what emf is generated across the tube ? The

tube

is

curved to correspond to the radius of gyration of the plasma.

15-28
it

will

move

If

in

an electron

is

injected into

a magnetic

the distance between turns will decrease until a point

turns becomes zero and the electron then reverses

toward

its

mirror.
at

relation

must be

satisfied

its

is

If

gradually increases,

reached where the distance between

direction with respect to

B and returns

as though the electron were reflected

by a

between the electron energy and the magnetic

field

starting point (see Fig. PI 5-28).

What

field traveling nearly parallel to B,

a helical path with large distance between turns.

It is

such a mirror point ?

15-29 A charged particle can be contained between two mirror points forming a
magnetic " bottle." The magnetosphere of the earth forms such a bottle for particles which
travel

between northern and southern hemispheres (see Fig. PI 5-29).

of the earth's magnetic

field

100

/iT.

What

particles to be contained if the particles are (a) electrons

particles will be

dumped, producing an auroral

The maximum value

energy-field conditions are required for

and

(b)

protons ?

Higher-energy

display.

15-30

How can

15-31

Discuss similarities and differences between plasma and gyro resonances.

15-32

Consider a horizontal ionized layer (such as in the ionosphere) which

a magnetic "bottle" be employed to contain a plasma?

thick with uniform electron density

N = 10 12 m~

and above

it

is

100

km

another layer of equal thick-

N = 4 x 10 12 m~ 3 The magnetic field is vertical and of magnitude 5 = 100

A circularly polarized 16-MHz wave is incident on the bottom layer from below at an
= 40. Trace the paths of the wave transmitted through or reflected
angle of incidence
ness but with

^iT.

cf>

from the

layers.

Consider

all

modes of propagation.

PARTICLES

AND PLASMAS

Trapped

747

particles

weaker

Earth

Mirror point
N

Particle path

FIGURE

FIGURE

PI 5-29
Magnetic bottle in magnetosphere.

PI 5-28

Magnetic mirror.

15-33

(a)

Field lines in

magnetosphere

Show

that under conditions

where electron

collisions with gas molecules

are significant but the magnetic field can be neglected, the complex relative permittivity of an

medium can be

ionized

expressed as

Ne
1

eo/wt/^

t,

= e' je"
r

where

Ne 2fc

+o

and

mu>(fc 2

+ to 2

= collision frequency
cu = 2-n x wave frequency

where fc
It is

assumed that the wave power lost as heat in the medium is equal to/c times the momentum
the electron. (*) Demonstrate that attenuation due to collisions is greatest when

mv of
fc

=<u.
Calculate the attenuation (in decibels) for a 3-MHz wave which has traveled
= 6x 10 1J1 m~ 3
through a nonmagnetized ionized region with electron density

15-34

100

when fc

= u>.
Find the radius of curvature of a proton with velocity

15-35

traveling normally to the following magnetic fields:

magnetosphere,

The

nT.

B = 50

/xT, (c) solar

particle is relativistic;

(15-2-12)].

Such a

15-36

corona,

i.e., its

particle in space

B = 5 mT,

mass

is

and

The magnetic

that the particle radius

(d) interstellar

is

field

a constant

a synchrotron are that

B and gyro

(= R

of

_1

medium, B = 10
mass [see

),

Show

B = lirfmje

frequency /are both varied in such a

and energy required

same purpose.
N.

that the

and /=

two conditions which must be

iV2tj-.Ro

beam

satis-

Thus, increasing the frequency

must be increased with the frequency

to produce a thrust of

Specify proton

way

so that an annular, or ring-shaped, magnet can

(and particle mass) to keep the particle radius constant.

15-37 (a) Design an electron gun for propulsion of a space vehicle.

the

significantly larger than its rest

increases the particle velocity (and energy) while

current

10 8

T, (*) earth's

could be called a cosmic ray.

be used, at a substantial cost reduction.

beam

=2 x

B=1

A synchrotron is a particle accelerator similar to a cyclotron but with the follow-

ing modifications.

fied in

(a) synchrotron,

N.

(b)

Specify electron

Design a proton gun for

current and energy required to produce a thrust

748

ELECTROMAGNETICS

15-38

Explain

why a charged

[CHAP. 15

particle is " swept along like a surfer "

frequency wave from a pulsar, using F

= eE and F = e(v X B).

by the very-low-

(See Sec. 14-27.)

15-39

Prove

15-40

Why is the deflection of an electron by an electrostatic field as given by (15-4-6)

(15-6-1).

and (15-4-7) independent of its charge and mass but its deflection by a magnetic
by (15-4-11) and (15-4-12) is dependent on its charge and mass?

field as

given

16
MOVING SYSTEMS AND SPACE-TIME

INTRODUCTIONf

16-1

Early thinkers believed a given instant of time had the same meaning for all observers,
moving or stationary. However, a critical analysis by Einstein in 1905 resulted in a

new formulation of

ideas about space

discussed in the next section,

is

and

time. J

The concept of

fundamental to this formulation.

simultaneity,

Space-time relation-

ships based on the concept of simultaneity are developed in subsequent sections.

These are followed by the Lorentz transformation of coordinates, a four-dimensional

formulation, fields of moving systems, and the interrelation of electric and magThese topics culminate in the development of the Maxwell stress tensor.

field

netic fields.

16-2
Is

it

At

SIMULTANEITY
on the earth and an event of Mars are simultaneous ?
might appear to be a simple matter to do so. Mars and the earth,

possible to say that an event

first

glance

it

An

equally good

% Albert Einstein,

Moving

for this chapter is Moving Systems and Relativity.

Zur Elektrodynamik bewegter Korper (The Electrodynamics of
title

Bodies), Ann. Pays., 17: 89 (1905); reprinted in English in

Methuen and Co., Ltd., London, 1923.

Relativity,"

"The

Principle

of

750

[CHAP. 16

ELECTROMAGNETICS

FIGURE
is

or vehicles. Vehicle II
passing vehicle I with a relative ve-

locity

ABC

16-1

Two space ships

vehicle I

A'

B-

JL.

-* PHIZZS
*

to the right.

C7

=0

however, are in motion with respect to each other, and for bodies having such relative

motion

it is

not possible to state that an event

is

simultaneous on both.

Einstein's hypothesis, simultaneity can be invoked only if there

Thus, consider two points

point

A and C on the earth's surface, say

B midway between, or 60 km from A

both A and C simultaneously

120

is

no

According to
relative motion.

km apart, and a third

A radio signal transmitted from B

and C.

A and C can be set to

A and C can be comrun in synchronism. In this way
signal
A and C from B is
to
reach
for
the
radio
pared. Since the time required
-1
5
and
C by 200 /is will
the
clocks
at
A
advancing
200 /xs [=60 km/(3 x 10 km s )].
will arrive at

so that clocks at

the simultaneity of events at

and the simultaneity of events at all three points can be

compared. Likewise the simultaneity of events at two points on the surface of Mars
can be ascertained by sending a radio signal from a third point midway between them.
But the simultaneity of events at points on the earth relative to points on Mars cannot
synchronize

all

three clocks

be determined except in an approximate manner.

two long spaceships, or vehicles, as in Fig. 16-1.

Both vehicles have been constructed in the same factory to identical dimensions and
The length of each is /. Each vehicle is equipped with three clocks,
specifications.
one located at the center and one at each end. Consider that we are aboard vehicle I
(lower one in Fig. 16-1) and that vehicle II is passing us from left to right with a velocity
Assume that we are far out in space and remote from the
v relative to our vehicle.

To illustrate

this let us consider

we are aware is the

shown in Fig. 16-1 vehicle II is
= on clock at B) a radio signal

earth or other celestial bodies, so that the only motion of which

relative

motion of the two

directly alongside
is

vehicles.

At

the instant

our vehicle, and at this instant

emitted from the center of our vehicle.

of the vehicles, the signal

vehicle

II.

is

received at the

For convenience we

(t

Neglecting the transverse separation

same

instant at point B' at the center of

shall say that this also is at

on the clock

at B'.

Wavef ront

of

impulse emitted

from

FIGURE

16-2

Space vehicles in Fig. 16-1 after a time

/

= //2c.

SEC- '*"2 J

MOVING SYSTEMS AND SPACE-TIME

751

All three clocks on vehicle I were previously synchronized with the

clock at B, and
all three clocks on vehicle II were previously synchronized
with the clock
at B'.

When

the impulse emitted from

reaches the ends of our vehicle (points

the clocks on our vehicle will read

=^ =

//2c,

where c

is

and

C),

the velocity of the radio

However, as suggested in Fig. 16-2, it will appear to us on vehicle I that the

have arrived at point A' on vehicle II at an earlier time t and will arrive
2
point
at the other end of vehicle II at a later time t
Thus
3

waves.

impulse

will

at

1/2

(1)

__

,. =
= 1/2 ^
h
^h
c V

and

On

(2)

the other hand, an observer on vehicle II can regard our

vehicle as passing
from right to left in Fig. 16-1, in which case the impulse from
B (or

his vehicle

J?')

arrives at the

C of his vehicle but at an

same time tt at the end points A' and

the right end C of our vehicle and at a

time (:/,) at
of our vehicle.

It is

evident that what

taneous on the other (unless v

A convenient

way of

is

earlier

later time (>/,) at the left end A

simultaneous on one vehicle is not simul-

0).

illustrating these space-time relationships is

by means of

a Minkowski diagram.f

Consider, for example, the expansion of the radio waves

from point B on our space vehicle I. As shown in Fig. 16-3a, the
radio impulse is
transmitted at / =
and radiates in a spherical manner like an expanding soap
bubble. The wavefront reaches points A and C at the time / =
f, = //2c.
In the
Minkowski diagram corresponding to this situation (Fig. 16-36) distance
along the
axis of vehicle I is represented by the x coordinate (abscissa)
while the distance that
the radio waves travel is represented by the y coordinate
(ordinate).

The medium

is

vacuum (//, =

the

same

er

as for light.

1),

and the radio waves

travel with the velocity c

= 300 Mm s " '),

Thus, the ordinate can be expressed in terms of ct, the distance

t.
The position of the wavefront at any time / is

traveled by the waves in the time

given by the two slant lines making a 45 angle with the

reaches the points

Fig. 16-3*,

and the

A and C

at the

axis.

C (x = /). These perpendiculars are

while the slant lines are called light lines.
be called radio lines since our signaling

is

When

the wavefront

shown in
= 0) and
called world lines for the points A and C,
In our case they might more appropriately
being done by radio instead of light.

slant lines intersect perpendiculars to the

at

two ends of our vehicle

I,

ct

1/2,

axis at

as

(x

H. Minkowski, Die Grundgleichungen fur die elektromagnetischen

Vorgange in
bewegten Korpern, Goettingen Nachr. 1908, 53; Space and Time,
address given at
Cologne, Germany, 1908, reprinted in "The Principle of Relativity,"
Methuen and
Co., Ltd., London, 1923.

752

[CHAP. 16

ELECTROMAGNETICS

A = l/2c

t=l/2c

(a)

FIGURE
(a)

16-3

Space vehicle I with expanding radio

wave emitted from point B at the center

of the vehicle; (b) Minkowski diagram
(b)

for this situation.

SPACE-TIME CONCEPT

16-3

now the representation by a Minkowski diagram of the situation for the two
As illustrated in Fig. 16-4,
vehicles I and II which we have been discussing.

Consider
space

the world lines for the points

A and C of vehicle I will be perpendiculars to the x axis,

C of vehicle II are slant

same as before, while the world lines for the points A' and
with respect to the ct axis, where tan
lines making an angle
the velocity of vehicle II with respect to vehicle I and c is the
the

intersection of these

and P'

(ct

world

ct 3 ), as

ously deduced.

makes an angle

The

lines for

shown

A' and

in Fig. 16-4,

v/c,

and where

velocity of light.

is

The

C with the light lines occurs at P (ct = ct2

where

t2

<h

and

t3

> tu

as

we have

previ-

slant line connecting these intersection points (PP' in Fig. 16-4)

with the

axis.

Taking a

PP' as one axis and

on vehicle II, we can

slant line parallel to

the world line for the point A' as the other axis for an observer

Minkowski diagram of Fig. 16-5. Here the space-time axes for an

observer on vehicle I are x and ct, as before, while for an observer on vehicle II the
space-time axes are x' and ct'. For convenience in Fig. 16-5 we also consider that the
(or t' = 0) from the point A (or A") at the left-hand end
radio pulse is emitted at t =
of from the center point B (or B') as in Fig. 16-4.
instead
16-1)
of the vehicles (see Fig.
unprimed axes (for vehicle I) system S and the
for
the
We shall call the coordinates

construct the

sec. 16-3]

MOVING SYSTEMS AND SPACE-TIME

World

753

line

for point

CorC"

FIGURE

16-4

Space-time, or Minkowski, diagram for both space vehicles of Figs. 16-1

and
16-2 with midpoint as reference. The world lines for vehicle I
are vertical, and
those for vehicle II are slanting.

Light line

FIGURE

16-5

Space-time, or Minkowski, diagram for both space vehicles of Fig. 16-1

with point A as reference.

and 16-2

[CHAP. 16

ELECTROMAGNETICS

754

coordinates for the primed axes (for vehicle

II)

system

The

an oblique (nonorthogonal) coordinate system.

S axes

respect to the corresponding system

as given

= tan~

The system S' axes form

make an angle with

S'.

S' axes

by

(1)

=
c =

where v

velocity of vehicle II with respect to vehicle I

velocity of light

As indicated

in Fig. 16-5, the light line

This line

axes).

is

of both systems

on

both

of

sets

ct

(2)

ct'

(3)

x'

said that these light lines are invariant;

To us on

identical (bisects

denned by the relations

x=

It is

is

vehicle I all events

on the x axis

on the

vehicle II all events

x' axis

appear progressive with increasing

The projection of a
also shown in Fig. 16-5.

(t

(t'

i.e.,

they are the same in both systems.

= 0) appear simultaneous. To an observer

= 0) appear simultaneous while to us they

'.

general event

on both the

(x,ct)

and the

(x',ct')

axes

is

now determine the relation between the unprimed coordinates for vehicle I
and the primed coordinates for vehicle II. Following Einstein, we make three

Let us
(ours)

assumptions:

The two systems are at rest or in uniform linear motion with respect to each other.
The velocity of the radio waves (or light) is c as measured in any system.
Measurements cannot disclose any fundamental difference between the two

2
3

systems.

The coordinates
vehicle II

for vehicle

by the system

are given by the system

I (in

system S)

is /

Each system moves with

S' (primed).

so that the vehicle appears at rest in

its

own

coordinate system.

and the length of vehicle

are stationary with respect to each other (v

when

there

is

relative

motion

vehicle I to be less than

observer on vehicle

To fulfill

/ (

II to

(v

=/ -

be

less

S (unprimed) and those

0),

II (in

= 0),

system

S')

its

The

is /'.

length of vehicle

When

the lengths are equal

the length of vehicle

for

respective vehicle,

II will

(/

the vehicles

/').

But

appear to us on

A/) while the length of our vehicle will appear to an

than /'(=/'-

A/').

This

is

illustrated in Fig. 16-6.

assumption 3 above requires a reciprocal relationship between the

two

systems such that

-LLl-M
l'-Al'

(4)

sec. 16-3]

MOVING SYSTEMS AND SPACE-TIME

FIGURE

755

16-6

Space-time relations between the two space vehicles illustrating the length-contraction effect.

From

the geometry of Fig. 16-6

(/'

we have

A/')

I'

= / 2 (1 + tan 2

9)

(5)

= x\l + tan

0)

(6)

the relations

l-Al = x(l -tan

Combining

(6)

and

y/l

- tan 2

quantity

less

than the length

The

factor y/l

factor.

It is

A/

is

the length

of vehicle

(v/c)

lies in

I as

/'

tan

(4),

(7)

= J\-

of vehicle

the range between

/'

(8)

we have

II as

observed from

the factor by which the length

observed from vehicle

It

and the square root of (5) into

/-A/

(9) the

0)

2
tan
= (/- A/) J\ + 2
1

In

(7) gives

/'

Substituting (8)

(v/c)

(9)

observed from vehicle I.

itself

and

of vehicle

II

by the factor ,/l

and

is

It is

(v/c)

2
.

called a contraction

appears to be shortened as

I.

also follows that

^-yr^w

(10)

756

[CHAP. 16

ELECTROMAGNETICS

vL/

Vehicle I

v.

\U

CD

Vehicle!*-

j,

CD
J5L, CD

Vehicle n
Vehiclel

FIGURE

16-7

Space-time relations between the two space vehicles illustrating the time-retardation

In (10) the quantity

/'

than the length

is less

effect.

A/'

/'

is

the length

of vehicle

II as

of vehicle

as observed

observed from

itself

from

vehicle

II.

It

by the same contraction

factor.

From

(8)

Thus, although

and

/' is

(9)

the

This

is

s/T-- We)

/'

{t

be

0)

when

(/ c)

(11)
2

illustrated in Fig. 16-6.

I

related as illustrated in Fig. 16-7.

t'=

at rest

In the unprimed system of vehicle

will

on vehicle II as / is on vehicle I, the length

be longer than / on the Minkowski diagram by the factor

same length

contraction requires that

given in (11).

/'

and the primed system of vehicle II the times

We set the clocks at zero on both vehicles

the two vehicles are opposite each other

[(1) in

Fig. 16-7].

At

this

After the two,

instant a radio signal is transmitted from the left end of both vehicles.
traveled
length / of
the
have
radio
waves
and
the
vehicles have begun to separate
will
have
moved,
vehicle
I
clocks
of
=
on
the
hands
//c and the
vehicle I, the time /
on
vehicle
an
observer
To
in
Fig.
16-7].
at
[lower
clocks
(2)
for example, one-half turn

SEC I* *]

MOVING SYSTEMS AND SPACE-TIME

the clocks on vehicle II at this instant will appear to read

clocks at (2) in Fig. 16-7].

Thus, the clocks on vehicle

vehicle I to run slow, or be retarded.

(I'/c)

The retardation

-At'

t'

= (I'/c) - At'

[upper

appear to an observer on

seen from the geometry to be

[v

Vic

is

II

757

< 12 >

\c)

After the radio waves have traveled the length /' of vehicle II, the time t = I'/c
and the hands on the clocks of vehicle II will have moved one-half turn [upper clocks
at (3) in Fig. 16-7].
To an observer on vehicle II the clocks on vehicle I at this instant
'

/ = (l/c) - At [lower clocks at (3) in Fig. 16-7].

Thus, the clocks
appear to an observer on vehicle II to run slow, or be retarded, in the

appear to read

will

on vehicle

amount

m -"-J>-&
V

lie

03,

\c

same retardation factor by which the clocks on vehicle II appear to run

I.
It is also the same factor by which the lengths appear
to be shortened.
The length contraction and time retardation are necessary conseThis

is

the

slow to an observer on vehicle

quences of the theory of
It is clear that the

the observer.

relativity.

appearance of the space vehicle depends on the viewpoint of

it will appear to all observers requires a formulation

To describe how

involving both space

and time.\

16-4

COORDINATE TRANSFORMATION

Let us

now

consider the transformation of coordinates between a system S (unprimed

is at rest with respect to us and a system S' (primed system)
which is

system) which

moving in the z direction with a velocity v, as in Fig. 16-8. The motion produces no
on the transformation in the x and y directions, so that we have

effect

The

x'

=x

(1)

y'

=y

(2)

length of a meterstick in the stationary or unprimed system is z

However,
of a meterstick moving in the primed system will appear altered in the
.

the length z

unprimed or stationary system

like the lengths

of the space vehicles discussed in the

preceding section.
t

How
term

the vehicle appears

relativity.

is

relative to the viewpoint

of the observer.

Hence, the

758

[CHAP. 16

ELECTROMAGNETICS

FIGURE

16-8

Rectangular coordinate system x', y', z'

moving with a velocity v in the z direction with respect to stationary system x,

The fourth coordinate

involving time must also be considered in the trans-

Thus, a moving meterstick involves space (three dimensions) and time

(fourth dimension); so a four-dimensional transformation is required.
Let a point P be expressed by the coordinates x, y, z, ut in the unprimed, or

formation.

and by x', y' z' ut' in the primed, or moving, system. The factor
m has the dimensions of velocity, so that the product ut (velocity times time) has the
dimensions of distance, like the other three coordinates x, y, z. Later we take u =jc,
stationary, system

where c

is

The time t = t' = is taken as the instant when the

moving and stationary coordinate frames are coincident.

the velocity of light.

origins of the

EXAMPLE Draw the Minkowski

= 0.53c.

diagram for the case where the above moving system

has a velocity v

solution

Minkowski diagram for

this case is

shown

in Fig. 16-9 with rectangular

Here tan = v/c = 0.53, and so the primed axis z' makes an angle
6 = 28 with respect to the unprimed axis z. In Fig. 16-9 we have arbitrarily made z = 3
grid units.
Now from (16-3-11) we have
grid for reference.

Zo

Therefore, zi

It is

=Zo

= 1.33z = 4 grid units,

Vl+OVc) 2
Vi - (vlcY

:1.33z

(3)

as indicated.

necessary to distinguish between the length of a line in millimeters on

the Minkowski diagram and the units of length the line represents in a given system.
Thus, Zq is longer than z on the diagram but Zq and z represent the same length in

each system.

MOVING SYSTEMS AND SPACE-TIME

sec. 16-4]

759

ct

^**
^-^Y

1 grid unit

/*'C

= 28
\

FIGURE

16-9

Space-time diagram illustrating calibration of axes for case where moving system
has a velocity v = 0.53c in z direction. (See worked example.)

The relation of a general event P to the coordinates is illustrated in

Comparing Fig. 16-10 with Fig. 16-6 and noting (16-3-8), we can write

__T

z'

Z-Vt
,

or

l-Al

Vl+(Wc) 2
y 1 _ (/ c)2

+(w<o

yjl-(v/c)

FIGURE

(4)

2-vt

Vl-OVc)

vt
_
^1 (r/c) 2
z

(5)

16-10

Space-time diagram for moving (primed)

and stationary (unprimed) coordinate

systems.

Fig. 16-10

zvt

760

ELECTROMAGNETICS

The

units for z'Q

represent the

[CHAP. 16

and z differ on the Minkowski diagram (Fig. 16-9), but they each
same length in their own system; therefore if we put z' = z
,

vt
(6)

y/l

which

(V[C)

a relation for the transformation of the coordinate z from the unprimed

is

(stationary) to the

primed (moving) system.

In like manner from Figs. 16-10 and 16-6

z'

vt'

_V_

z'+vt
or

yi+(f/c)

z'+vt'
-(V/C)

(8)

(9)

as before, gives

y/l
is

+ vt'
yi (vie) 2

z'

which

\c.

z'

z'o

(7)

y/l+Wcf'Jl
z

IV

Jl-(vlc)

and

z'

we obtain

Putting

+ vt'
~ (Vic?

the relation for the transformation from

z'

(10)

to

z.

Eliminating z from (6) and (10) yields

2

f+

(v/c) z'

y/l

'We) 2

(11)

Eliminating

z'

from

(6)

and

(10) yields

t'

V
Equations (11) and (12) indicate
stationary system and vice versa.

how

(vjc )z

wo

the time

(12)
2

transforms from the moving to the

In summary, the transformations from the primed to the unprimed coordinates

and

vice versa are

=x

y=y

x =x

(13)

(14)

sec. 16-4]

MOVING SYSTEMS AND SPACE-TIME

z'

y/l

+ vt'
~ (VIC) 2
(v/c

y/l

y/l

y/l

(VIC?

761

vt
(v/e)

(15)

(v/c )z

(VIC)

(16)

Taken as a group, these constitute the Lorentz transformation.^ If the wave

is assumed infinite instead of equal to the velocity of light, so that c in the

velocity

above equations

is

replaced by oo,

x=

= 0,

An
follows.

obtain the Galilean transformation

x'

= y'

z'

t'

The Galilean transformation

If v

we

=x

(17)

y'=y

(18)

x'

vt'

z'

=z

vt

(19)

(20)

implies instantaneous transmission

of information.

both transformations reduce to the same form

=x

y
z

= y'
= z'

=f

(21)

alternative (nongeometrical) derivation of the Lorentz transformation

Referring to Fig. 16-8, suppose that a radio wave

is

of the stationary (unprimed) system of coordinates at the instant

(primed) system

is

coincident with the unprimed system.

(x,y,z) in the stationary

is

as

emitted from the origin

t

= when the moving

An

observer at a point
system will detect the passage of the wave after a time / such

that

x2 + y2 + z 2

2 2
t

(22)

An observer moving in the primed system will detect the wave after a time
x'

+ y' 2 + z' 2 -

2
t'

t'

such that

(23)

t H. A. Lorentz, Electromagnetic Phenomena in a System Moving with Any Velocity

Less than That of Light, Proc. Acad. Scl. Amsterdam, 6: 809 (1904); reprinted in

"The Principle of Relativity," Methuen and

Co., Ltd., London, 1923.

[CHAP. 16

ELECTROMAGNETICS

762

The desired transformation must be linear and also must not alter
form of the equations. Thus, the transformation must be of the form

the quadratic

=x
y'=y
z' = a^z + a 2 1
t' = a 3 z + a A t

(24a)

x'

(24*)
(24c)

(24d)

Taking the time derivative of

where aj.a2.a3, a 4 are constants to be determined.

(24c) yields

dz

dz'

Tt=
Now

dzjdt

However,

v,

dz'/dt

respect to

it.

When

this is

a>

>dt

(25)

the velocity of the primed system as seen by a stationary observer.

= 0,

+ y2 +

an observer moving with the primed system

since

Thus, a 2

and the values of x',

x2

,.

= -a

y', t'

(fli

compared with

and

from

z'

a t (z

(24a), (24*),

and

is

Substituting this value of

vt).

it

z'

(24a") into (23) yields

*y _ 2zt(ai 2 v + c 2 a ad = (cW - ai 2 v

(22),

at rest with

)t

(26)

follows that
a,

a4
a

- c2 a 2 =
- a 2 v2 = c2
1

Solving these three equations for a t ,a 3

c a 3 d4.

(27)

a4 we obtain
,

(28)

a3

v
cV 1 - (vie) 2

Substituting these values back in (24c)

and

(24a"),

(29)

we

obtain for the Lorentz trans.

formation

the

same

as in (13) to (16).

x'

=x

z'

= a,(z-

t'

a3 z

vt)

a4

=
=

Vt

(vfc )z

(30)

MOVING SYSTEMS AND SPACE-TIME

sec. 16-5]

FIGURE

763

16-11

General space-time (Minkowski) diagram, relating position and time for different
systems and epochs.

16-5
It

INVARIANCE IN SPACE-TIME

has already been shown that the light lines defined by x

ct (or x'

ct')

are invari-

ant with respect to a stationary coordinate system or one moving in the

x direction.
The Minkowski diagram of Fig. 16-5 represents but one quadrant of the more general
situation of four quadrants shown in Fig. 16-11.
Here the light lines are given by
x

x+

and

More

generally

we may

where p

=q=

in (1)

and

(2).

pq
Equation
equal to

(5) is
2

ct axes.

ct

ct

(1)

(2)

write

p = x-ct

(3)

=x+

(4)

and

ct

Taking the product of p and

= (x-

ct)(x

ct)

= x2 -

c2 t 2

q,

we have
(5)

the equation of a hyperbola for which the product pq is a constant

+ / is the intercept of the four branches of the hyperbola with the x

where

This

is illustrated in Fig.

16-11.

764

[CHAP. 16

ELECTROMAGNETICS

FIGURE

16-12

Diagrams

illustrating the invariant* of the space-time hyperbolas.

>

7
-0

^vv

=28
1

v0

FIGURE

16-13

Space-time diagram showing hyperbola as invariant calibration curve for case

where v = 0.53c in x direction. On the diagram x' is four-thirds the length of x ,
but the observer in each system sees his own length as equal to unity.

MOVING SYSTEMS AND SPACE-TIME

SEC. 16-6]

From

765

the Lorentz transformation

x'

vt'

p'

+ (p/c)V

c%

(6)
It

follows that

pq

= x2 -

c2 t 2

= x' 2 -

2
t'

(7)

Thus, as illustrated in Fig. 16-12, the hyperbolas are the same whether defined
stationary (unprimed) coordinates (x

coordinates (x'
ct axes is

/,

2
t'

= pq

2
I'

).

The

2 2
t

=pq =

2
I

in the

or in the moving (primed)

intercept of the hyperbola with the

whereas the intercept of the hyperbola with the

x'

and

ct '

axes

is /'.

x and

From

= 0.53c, the lengths are related to a hyper= 1. This is illustrated in Fig. 16-13.

the example illustrated in Fig. 16-9, where v

bola for which x

2
I

and

Returning to Fig. 16-1

we

x'l

2
I'

note that the origin represents the point " here-now "

= t = 0),

while points above the x axis represent the future and points below the
For an observer in the primed system points above the x' axis (see Fig. 16-12)
represent the future, and points below represent the past.
It is now clear that the light lines are invariant axes on the world diagram while
(x

past.

the hyperbola

16-6

is

an invariant calibration curve.

FOUR-DIMENSIONAL FIELD FORMULATIONf

In this section Maxwell's equations are given a four-dimensional (or 4-space) formulation.

This prepares the

way

in the next section for the

development of

moving systems by incorporating the Lorentz transformation

in the

fields

of

four-dimensional

formulation using tensors.

Maxwell's curl equation from Ampere's law

is

given by

VxH = J +
Expanding

this in rectangular coordinates,

we have

Moon and D. E. Spencer, " Foundations of Electrodynamics,"

Nostrand Company, Inc., Princeton, N.J., I960.

t P.

(1)

p. 263,

D. Van

766

ELECTROMAGNETICS

[CHAP. 16

This vector equation yields three scalar equations; adding to this the expansion of
Maxwell's divergence equation from Gauss' law (?
=
yields a set of four

p)

equations:

dH*

an
My

fin
dD

dy

dz

dt

dH z

dHx

dDy

dx

dz

dt

dH

dH.

dx

dy

dx

dD x

--\

The terms on

the

left

If

8DZ

Ji x

(3)

= Jy

(4)

J,

(5)

dt

8D y

dD,

dy

dz

=p

(6)

of the above equations have been arranged so that the

and

z,

while the fourth

absent, a zero

is

introduced to indicate this

are derivatives with respect to x, y,

to time.

any term

is

is

first

explicitly.

of the above equations involves current density J or charge density

Recalling (16-4-22), we have

right side

x 2 + y2 +
The equivalent equation

2 2
t

three

a derivative with respect

The

p.

(7)

in four-dimensional notation is

V+x

+ x 3 2 + x* 2 =

(8)

so that
Xl

=x

x2

=y

x3

=z

X4 =jct

(9)

where j

is the complex operator (=^-1).

All four of these quantities have the
dimensions of distance since c (velocity) multiplied by t (time) yields distance.

Substituting

and multiplying the fourth equation by jc to achieve symmetry givesf

dH 3

dH 2

ox 2

dx 3

ox t

dD,

jc = j
.

dx 4

dx 3

<5x 4

8H 2

dHi

dx t

dx 2

dD,

dD

dD

ox t

dx 2

dx 3

dx.4

2
3
Jc-T+jc^+jcl+
.

(10)

dD

-jc3 = J,3

=jcp

Note that

_.JC i__.JC 8D,dr_

dxt.

6t

dXi

_jcd* dt _
~~
~
jc

dt

dt

8DX
dt

MOVING SYSTEMS AND SPACE-TIME

sec. 16-6]

Now

let

Gm* =

Here

767

Gmn

is

-H
H2

H 3 -H 2
H
-H

jcDy

jcD 2

note that

Gmn

when

We

can

now

summation form as

dGnm _

m= 1,2,3,4

=1 oxn

matrix as

= G m Gm =

contains only six independent quantities.

write the four scalar equations in concise

(11)

jcD 3

a skew-symmetric or antisymmetric matrix (G m

m = n). We

where G

-jcD1
-JcD 2
-JcD 3

(12)

in (11)

--jcp

Equation (12)
(V x

=J+

is

a four-dimensional formulation of two of Maxwell's equations

dD/dt and

For each value of

D = p)

in

which the fourth dimension involves time.

we have a scalar equation, and for each value of n we have one

term of that equation.

manner we can express Maxwell's curl equation from Faraday's law

dB/dt) and Maxwell's divergence equation from Gauss' law (V B = 0)

In a like

(V x

E=

in four-dimensional

form as follows

dF
Y ll= =
=

The

Gm

Ei

E2

-Ei

L -jcB,

-JcB 2

and

-E2

-E3

quantities

1,2,3,4

E3

where

m=

dxn

Fmn

and the Faraday-Maxwell

Expanding the continuity

jc

(13)

Bn

jcB 2

(13a)

jcB 3

-JcB3

are sometimes called the

Ampere-Maxwell vector

vector, respectively.
relation

V- J =

dp
(14)
dt

768

ELECTROMAGNETICS

[CHAP. 16

we have
dJ

-H

5J

dx

dy

Let x x
JU

= x, x2 = y, x 3 =

-H

xA =jct,

z,

dJ z

dp

8z

(15)

dt

as before,

and Ji

=Jx

J2

Jy

J3

=J

JCP> so that
dJ,

dJ 2

dJ 3

6Ja

oxj

dx 2

dx 3

ox 4

(16)

dJ

or

(17)

m = 1 C^m

where

is

Likewise from the relation

the four-dimensional divergence of J.

between the scalar and vector potential

we can

(18)
dt

write
4

dV
(19)

m=l OX-

where

Vt
X\

=A x
X

V3

*2=y

x3 = z

SGm
1

dx

JV
K4 =

The above four four-dimensional equations

=A

V2 = A y

in

summary

Xt =jct
are

m= 1,2,3,4

(20a)

W =l,2,3,4

(20b)

0^ =

(20c)

dF
-^-"
=
n=l OX

or
4

57
=
1^
>idxm

(20d)

MOVING SYSTEMS ANTr SPACE-TIME

sec. 16-7]

By

Einstein's

summation convention these four-dimensional

relations

769

can be abbre-

viated as follows:

8Gml

= J

(21a)

dx

Maxwell's equations:

8Fm

(216)

dx n

Continuity relation or equation

(21c)

of conservation of charge

dx m

Potential relation:

(2ld)

dx m
1

According to the Einstein summation convention, the above equations are understood

summation

to have

signs as in the equations of (20a) to (200*).

These four equations of

(20a) to (20d) or (21a) to (21a") contain the essence of electromagnetic theory in highly
concise form.

The four-dimensional formulation

is

especially useful in coordinate

transformations, as demonstrated in the next section.

16-7

We

FIELDS OF A MOVING SYSTEM

have discussed moving systems and four-dimensional

field

formulations.

Let us

now apply this information to the problem of the electromagnetic fields associated with
a moving system. The first full solution of this problem was made by Minkowski
in 1908.

Consider two coordinate systems x t x 2 ,x 3

x4 and x[, x'2 ,x'3 ,x'^. The use of

means that we can specify not just a point in space
,

the four-dimensional formulation

but a point in both space and time.

Let the

first

(unprimed) system be at

rest

with

and the second (primed) system in motion with a velocity v. Without

loss of generality let the x 3 (or z axis) coincide with the direction of v (see Fig. 16-8).
Also assume v to be linear and uniform (no acceleration).
respect to us

According to the Lorentz transformation (16-4-30), we have

X =X

y =y
2

'

vt

Vi - (v/c) 2
2
t - ivlc )z
1'

Vl -

(/cY

(1)

770

But

ELECTROMAGNETICS

[CHAP. 16

in the four- dimensional formulation

= xt

x'

y = x2
z = x3

y'
z'

x
t= -

t'

= x\
= x'2
= x'3

(2)

jc

jc

Hence, in the four-dimensional formulation the Lorentz transformation becomes

Xt

X2

X2

*3

X\

1
/

y/\

{VIC)

\x 3

In tabular or matrix form the transformation

+j - xA \
-

(3)

is

x2

X,

Xl
1

x'2

(3a)

jy-

-jy.

where

=
y/l

In general this transformation

may

dxj

XA

be expressed as
dx\

dx'i

dx,

dx'

dx'

2,

dx'2

8x'2

dxi

dx 2

ax.

6X4

dx

dx'

ox 2

ox 3

x.+

2
4

3x3

x'i

(V/C)

dx 2

dx t
,

_
x
3

ox l

dx'A

= t x, +
3x x

dx'A

8x2

x2

dx'A

dx 3

dx[

Oxa.

3
3
=-~-x +-^-x
2 +-^-x 3
1

(4)

dx

3
+-^x
A
3x

3x1

x3

+ r-= x4
dx A

MOVING SYSTEMS AND SPACE-TIME

.16-7]

771

or more concisely as
4

dx'

m=

xn
x m la
=1 ox n
-.

where

dxm
dx

"dx'i

dx\

dx'i

dx'i

dx i

dx 2

Sx 3

dx 4

dx'2

dx'2

dx'2

dx'2

dx t

dx 2

dx 3

dx A

dx 3

dx'3

dx'i

dx 3

dx t

dx 2

dx 3

dx 4

dx'4

dx^

dx^

dx'i

Sxj

dx 2

dx 3

dXi_

1, 2, 3,

(4a)

a 12
a 22

Ol3
a23

a 14

21
031

a32

33

41

a 42

43

34
a 44

a2i
(5)

or

(6)

Jy

-jyc

where

J\ - (vfc?

A quantity transforming as in

(4) is

employed instead of subscripts

as here.

matrix.

It

describes

how

a tensor of rank

The

1.

coefficient

Commonly superscripts are

dxjdxn is the transformation

a dimension in the unprimed system transforms into a

dimension in the primed system.

The four-dimensional equation incorporating Maxwell's equations

-dB/dt and V

B=

from

(16-6-13)
4

dF

VxE =

is

ro

l,2, 3, 4

(7)

=i dxi

To transform the fields given by Fmn in the stationary or unprimed system to the moving
or primed system we write
4

F'u=

4
Y, a km a m

k,

1, 2, 3,

(8)

where Fkl represents the transformed fields in the moving system and the coefficients akm
and a ln are as given by amn in (6). The field quantities Fmn and F'u which transform as
Each component F'kl of the transformed field is, from (5),
in (8) are tensors of rank 2.
given by the

sum of 1 6 terms.

We note, however, that many elements of the matrix for

772

ELECTROMAGNETICS

[CHAP. 16

the coefficients are zero. Thus,a 12

aA2

= 0-

a 13

= aiA = a21 = a 2i = a24 = a 31 = a 32 = a41 =

= F 22 = F 33 = F 44 = 0.

Also the components of the untransformed field F tl

Thus, for the transformed or primed

field

components we find

*n=0

(9)

F'12

=ait a22 F12 =Fl2

F'13

= a n a 33 F 13 +

(10)

a tl a 3A F lx

= y\F13 +j-F A

FU = an a43 F l3 + a^a^F^ = y\F1A -j-F l3

.

F'l\

F 22

= F21 = _ F12
=

(11)

(12)

F'12

F23 = a 22 a 33 F 23 + a 22 a

(13)
(14)

= y[F 23+j -F^

(15)

(16)

F3 1

(17)

F32

-Fi,

(18)

F 33 =0
F34

(19)

= a 33 "44
F 43 = ? MF34
F34 +
(I44F34.
a 3A "43
a i3 r
-f "34
^ 34 +
\

-p
^2

F
* 43

i
J

(i-^)f 3 4 = f 3

(20)

F = -FU
F42 = -F 24
F43 = -F 34
F44 =

(21)

(22)
(23)
(24)

Writing these in matrix form and, from (16-6-13a), putting

F l2

= E3 F13 = E2
,

F 14 = jcBu etc., we get

-y(E 2 +vB 1 ) jcy^B 1 +

-F 3

y{E l

vB2 )

jcy\B 2

JF

- -^ ,

FL,=

y( 2

+ vB

-tfEi

t)

-W^i + J F j
2

-jc? (b 2

- B 2 )
-

JE

JcB 3

-jcB 3

J
(25)

MOVING SYSTEMS AND SPACE-TIME

sec. 16-7]

FIGURE

773

16-14

E and B and their cross products with v. In the figure we

E, = Bx = 0, and so |E|=.E |B X =.B,, |vXE| =
|vX B| =(vX B)>. (Note that vXE and vXB are always

Components of fields
have

arbitrarily

(vXE), and

perpendicular to

Comparing
etc., it

set

v.)

(25) with (16-6-13a) element

by element and putting

Ex = Eu Ey = E2

follows that

The choice of the

+^E^

E'x

= y(Ex -vBy )

B'x

= y^Bx

E',

= y(Ey + vBx )

B'y

= y(By -

e:

Bl

e.

z direction for v

is

arbitrary.

JE )

(26)

B,
Referring to Fig. 16-14, the above

relations can be expressed more generally by

B'n

i = y|Ex

5i =

where the subscript

subscript

J_

||

+ (vxB)|

electric

field

observer the electric

7 |B x

z)

-(iyxE)

(27)

component parallel to the motion and the

component perpendicular to the motion. The inter-

and magnetic

perpendicular field components.

and magnetic

indicates the field

indicates the field

dependence of

=B (=B

,',=(=,)

fields is illustrated

by the equations for the

Thus, to a stationary observer the electric

appear to be independent quantities.

field

has parts involving both the electric

field

However, to the moving

field

E and the magnetic

774

ELECTROMAGNETICS

field

[CHAP. 16

of the stationary observer, while the magnetic

the magnetic field

B and

the electric field

field

has parts involving both

of the stationary observer.

It is clear

that the appearance

of afield depends on the viewpoint of the observer and that there is

no such thing as a pure electric field or a pure magnetic field which retains its identity
for

all observers.

For small

<

velocities (v

c), (v/c)

approaches zero and

J\ -

(v/c)

(28)

These equations then reduce to

E ~E
I = |Ei + (vxB)|
'\\

\\

Integration of the relation E'y

nj

By

Stokes' theorem

-dB/dt

it

[E d\

|E

=j E

(v

= JJ(V x

'\\

B'L

(29)

= BL

B)| around a closed circuit yields an

+ j (v x

d\

B = -^i

B)

d\

emf
(30)

VxE=

E) ds] and Maxwell's equation

follows that

-"--//*

(31)

so that (30) becomes

cur

rdB

=-fc^s+ f (VXB

)- dl

This is the same equation as appears in (8-5-1). As discussed in Chap.

term represents transformer emf and the second term motional emf.
The four-dimensional divergence of J is, from (16-6-17),
4

j=
To transform

the current densities given by

J'm=

n=l

we

amJn

8,

the

first

SJ

Ip =

(33)

n=l oxn

system to the moving or primed system

(32)

Jm

in the stationary or

unprimed

write

m= 1,2,3,4

(34)

MOVING SYSTEMS AND SPACE-TIME ' 775

SEC. 16-8]

where a mn

is

as given in

(6).

It

follows that

j;

05)

j,

K = yUz - vp)
P'

= y[p-j2 j)

The interdependence of current and charge is illustrated by this result. To a

J and charge density p appear to be independent quantities. However, to the moving observer the current density involves both
the current density J and the charge density p of the stationary observer. Also to
the moving observer the charge density involves both the charge density p and the
current density J of the stationary observer. It is clear that the appearance of the
current and charge density depends on the viewpoint of the observer and that there is no
stationary observer the current density

such thing as a pure electric current or a pure charge which retains

its

identity for all

observers.

16-8
Stress

MAXWELL'S STRESS TENSOR

is

defined as the force acting on a unit area in a material

section a four-dimensional force-energy relation

is

medium.

In this

derived which describes the stress

produced by the electromagnetic field on a body.

Maxwell's equation from Ampere's law is

V x
Assuming harmonic

variation

H=J+
3D

(1)

and taking the cross product with

1

(V x H) x

yields

B = J x B + - (D x

B)

2 at

Maxwell's equation from Faraday's law

(2)

is

V x E= -

(3)

at

Taking the cross product with

D yields

(VxE)xD=-l|(BxD) = l|(DxB)

(4)

776

ELECTROMAGNETICS

Adding

(2)

and

[CHAP 16

(4) gives

(V x E) x

D + (V

x H) x B = J x B

+ j (D x

B)

(5)

Each term of this equation has the dimensions of force per unit volume (M/L2 T 2 ).
We can expand and regroup (V x E) x D so that it takes the form

In like

Adding

(V x E) x

D=V

(V x H) x

B=V

(T4I ).

- E(V D)

(6)

(Tkl ) m

- H(V

(7)

manner we have

(6)

and

B)

(7) yields

D + (V

(V x E) x

x H) x B

= V Tkl - E(V

TtI = (Tt)) e + (Tt m = Maxwell stress tensor, N m"

= stress tensor due to electric field, N m~ 2
2
(Xu)m = stress tensor due to magnetic field, N m~

where

D)

- H(V

B)

(8)

,)

(Tn) e

The

subscripts k,

assume values

UEX + nH,* - UeE + pH

2

T,

+ nH, H
eE*E, + fiHx H
cEx Er

2
)

to 3,

and the matrix for the

cEx E,+ pHx H

+ pH * - i( 2 + pH 2 )
Er E + nHr H,

stress tensor is

+ pHx H
+ /xH H
+ fiH, 2 - i(eE 2 + pH 2 \
*EX EZ

eEy 2

*EZ Ey

eEz z

(9)

Each term of the

has the dimensions of force per unit area,f and may

be a pressure, a tension, or a shear. We note that V B = and V
= p. Substitutstress tensor (9)

ing (8) in (5) then yieldsj

V Tkl = pE + J x B +

(D x

B)

(10)

ot

Each term of

this

equation has the dimensions of force per unit volume.

represents the electric force per unit

volume.

When

the

sum of these two

f

volume and J x
force terms

= V-Tir -(DxB)

For example, the force dFx

dFx

in the

= Txx dy dz +

is

called

f,

(10)

is

pE

becomes

(Nm" 3 )

direction

Thus,

the magnetic force per unit

(11)

given by

Tx , dz dx + Txz dx dy

X See note with Prob. 16-4 regarding the significance

(N)

of the divergence of a tensor.

MOVING SYSTEMS AND SPACE-TIME

SEC. 16-8]

and assuming

and n

/i

f=V-TH -i|(ExH)
c

If

we

xt =

let

x,

x2 =

x3 =

y,

777

z,

x4 jet,

(12)

ot

this relation

can be expressed

in

four

dimensional formulation as

f=V-T -t

(Ex

H)

=1,2,3

k,l

(13)

c ebc 4

f=D-T

or

where

is

stress tensor
-Loth

m,n

the four-dimensional divergence

whose matrix

is

and

1,

2, 3,

(14)

TmB is the four-dimensional

Maxwell

given by

"~

eES + fiHS zE

E1 +

tE.Es

l(eE 2

i.H i

+ (iH 2

cf.fj

eEz 2

+ pHz 2 - UtE 2 + pH 2
eE2 E3

+ nHiHs

+ pH^z

+ fxH2 H3

'

+ fiHtf,

(ExH),

eE3 E1

+ fiHjH2

(ExH) 2

eES + nHS-UtE' + pH 2 )

- (EXH) ;

-(EXH),

eEtEs

;(EXH),

-^(EXH) 3
H<r

+ m// 2 )

(15)

The concise equation of (14) incorporates all of Maxwell's equations. Integrating

(14) over a volume yields the total force on the volume due to electric and magnetic
forces.
T, are

It

is

to be noted that the top three elements of the

equal to

Tkl

first three columns of

Thus, the components for the matrix of Tm can be grouped as

follows:

(16)

Si

Although the elements of all zones have the dimensions of force per unit area, we note
2
that element
+ nH 2 )] may be regarded as representing energy per unit
[ = \{eE
volume, while the elements of groups 5 X and 5 2 involve the Poynting vector ( = E x H),

or power per unit area.

778

ELECTROMAGNETICS

[CHAP. 16

'

^'
1

""?'
l-Hj

*\;-

,--

FIGURE

16-15

Plane wave incident on conducting slab

for calculation of the radiation pressure.
(See worked example.)

Incident

wave

EXAMPLE A plane wave is normally incident from air onto the plane surface of an infinitely
thick perfect conductor.

Calculate each element of the stress tensor

the conductor to find the radiation pressure

solution

Referring to Fig. 16-15,

let

T m

at a point inside

on the conductor.
the incident complex electric field be given by

the equation

% =y
The transmitted

eJ ,+ ' *>
,

(17)

electric field is

fe,

=$TE

e x "eJX ">ei "

(18)

2Z,

where r

Zo

+ Zi

transmission coefficient

= impedance of conductor
Z = impedance of air
8 = 1/e depth of penetration
Zx

Similarly, the transmitted magnetic field

H,

is

iarHo e" V"

V"

(19)

MOVING SYSTEMS AND SPACE-TIME

sec. 16-8]

where a

Thus, in

= Zo/Zt

The instantaneous

fields are

E,

=Ket, =f\r\E

H,

= Re H, = -

i.

given by

+|+ 0)

(20)

H e'" cosj <ot + |\

(21)

e"" cos(mt

ar

779

(15),

,=3

=#,=#2=0

(22)

E2 =E,

(23)

(24)

so that the stress tensor

is

given by

+ ^H, 1 )

i^oE.'

E,H,
i{e E,

-^H,

2
)

-H* a E, 2 -p

H, 2 )

J
-E
H,

He

E, 2

+
(25)

where E, and H, are given in (20) and (21). The volume force density f has a component
the x direction which can be found from (14) as

in

/,=(D-T.),

artl

BTtl

dTl3

8x

8y

dz

= -j-^(coE,* + n
The

force-density

not a function of

j8T1A

H,*)

dt

+ jt (E,H,)

components/, and/, are zero since T2 i

The time average of/, is given by

is

not a function of y and

(26)

T33

is

z.

(/,).v

= ij

T
(27)

/x(x,Orf/

Carrying out the operations indicated in (26) and (27) gives

(/,)..

^-U^Eo^ + ^-^H^Y^-

(Nm->)

(28)

where (/*)., is the average force per unit volume in the negative x direction or the stress on
a unit area (/ by /) of conductor of thickness dx, as shown in Fig. 16-15. Thus the total

780

ELECTROMAGNETICS

average stress in the

[CHAP. 16

on the conductor

direction

finitesimal slices dx, that

(r*).,

r"

we now assume a perfect

is

= -^o^o 2

causes the conduction electrons to

wm

is

given by

(30)

(N m - 2 )

(31)

H,

move

inside the conductor (in the

x H,

in the

produced.

is

ductor

by

of thickness dx

is

given by (J

direction)

This force density in turn

x direction, colliding with stationary atoms

of the conductor, thus producing a net pressure on the conductor.

/

(29)

due to the interaction of the conduction-current density

direction) with the magnetic field

so that a magnetic force of volume density J

element

in-

given by

Physically, this pressure is

(in the

HoHo 2 \

(Nm-

= - (r,)., = mo /To 2

such

conductor (Z, *-0; r -* 0), the

(7-,)

radiation pressure

ZT
Z

E 1+\

Thus the

all

dx

(/,).,

-^
If

found by integrating over

is

is,

H,)/ 2

dx and thus the

The

force

total force

on a square
on the con-

is

F = /2

"(JxH,)*

(32)

J*

(End of worked example.)

See also Probs. 16-30 and 16-31.

PROBLEMS
Show

16-1

that

VxE = -8B/8/

and

V-B =

can both be expressed

dimensional form as

n=l

16-2

Show

8Fm
-7^=0

m = 1,2,

OX,

that

\_dV

VA

c 2 dt

can be expressed as
*

p=
6V

n= I ox m

where K,

= A x V2 = A,, V3 = A
,

=j V/c.

3,4

in four-

MOVING SYSTEMS AND SPACE-TIME

16-3
field

Show

that the force F'

and magnetic

field

is

on a point charge q moving

Show

that

(V x E)

+ (yXB)\

can be expressed as

(T,)

- E(V

D).

Hint:

After expanding (V

tensor such as T* (

defined as

in a stationary electric

given by the components

Fl=qy\Ej_
F',=qE
16-4

781

add and subtract terms like E, 8E,/8x, E, 8E,/8y, E,8E,/8z

from each component and regroup. Note further that the divergence of a nine-component
is

E)

J8TX ,

8Txy

8Txt\

+*> + !*]
+ *(?*!!
^
y
\dx ^ 8y
8z
)

+ *{-8x- + -8j +
i

-8z-)

Thus, the divergence of a nine-component tensor

16-5

Show

that each element of the

is a vector.
Maxwell stress tensor T has the dimensions

of force per unit area or energy per unit volume.

16-6 Show that the electric field E of a charge e moving with a uniform velocity v
is given by

!-?

E = r-^
477<Fr

(l- 2 sin 2 ^) 3

= radius vector from charge to point for

= angle between r and v
= vle
Note that when v = 0, this reduces to E = lel4nc
where

which

'

is

given

$
fi

16-7
is

Show

that the magnetic field

r , as given in Chap. 2 for a static charge.

of a charge e moving with a uniform velocity v

given by

4n
where r
<f>

Note

= radius vector from charge

= angle between r and v

that for v

< c this reduces

(1

- ft* cos

to point for

that a,

is

given

(dv

is

a volume element.)

which

r dv

as given in Chap. 5 for a steady current of density J.

Show

3
<f>)

to

dB = i

16-8

= o4 = y and

a3

= - vy/c 2

in (16-4-28),

where y

1/Vl ~(vfc) 2

fROMAGNETICS

]CHAP. 16

Prove (16-2-1) and

16-9

(16-2-2).

stationary observer determines that a certain region of space contains a

_1
does this field appear to an observer moving
10
static electric field

16-10

Vm

How

through the region in the positive y direction with a velocity of (a) 10

(c)

10

ms-',

m s"

(d) 10"

and

(e) 3

m s"

x 10

m s~

(b) 10*

*,

m s_

',

Repeat Prob. 10-10 for the case where the stationary observer determines that
Ez = 10 sin 207r/(Vm" 1 ), where

16-11

the region contains a time-varying electric field as given by

t

time in seconds.

16-12

magnetic

static

stationary observer determines that a certain region of space contains a

field

Bx =

How

/12n- (T).

/^.

does

this field

appear to an observer moving

through the region in the positive y direction with a velocity of

(c)

10

m s-', (d) 10 m s"

1
,

and

(e) 3

x 10

m s'

(a)

10

ms"

1
,

(fc)

10*

m s"

*,

Repeat Prob. 16-12 for the case where the stationary observer determines that
the region contains a time-varying magnetic field as given by Bx = (/i /12r) sin 20n-/ (T),
16-13

time in seconds.
A stationary observer determines that a certain region of space contains a
How do
electric field E, = 10 V m" and a static magnetic field Bx = /a /12it (T).

where

16-14
static

components appear to an observer moving through the region

these field

m s"

direction with a velocity of (a) 10

(e)3x

10 s

ms"

16-15

Bx = ^oll2ir

the positive
(6)

What

direction at the

16-16

its

would

it

same

make

point charge of

peak

How

(T), respectively.

difference

wave

polarized

stationary observer determines that

and

(6) 10*

m s"

10 7

(c)

*,

m s"

',

in the positive

(d) 10 8

m s" \ and

A plane linearly

(a)

*,

traveling in the positive

is

electric

does

= c,

that

to the observer if the

is

direction.

is,

= 10 V m _1

fields are

wave appear to an observer

this

velocity (v

and magnetic

traveling in

riding along with the wave)?

wave were

How

situated at the origin.

circularly polarized?

does this point charge

appear to a moving observer traveling in the positive y direction at a constant distance

"
7
8
1
s , (d) 10
s ',
s - *, (c) 10
if the observer velocity is (a) 10 ms" , (b) 10*
z = 10

and(e)3 x 10"
this current

An

16-17

s" ?

electric current

of 10

is

flowing in the positive y direction.

appear to a moving observer traveling

in the positive

direction at

How

does

a constant

m from the current the observer velocity (a) 10 m s" (4) 10* m s _1
(d) 10 8 m s"
and (e) 3 x 10 8 m s" ?
(c) 10 m s"
*
incident normally on a perfectly conducting flat
16-18 A plane wave in free space
kW m -2 what the pressure (in kilograms
If the Poynting vector of the wave
sheet.
distance z

10
,

is

if

is

is 1

is

per square meter) exerted by the wave on the sheet ?

(b) If so what
(a) Can an antenna be used for propulsion of a space vehicle?
N?
produce
thrust
of
1
must
radiated
to
a
be
power
16-20 (a) If radio sources all have the same power output and are distributed uniformly throughout space, show that the number JV observed by a radio telescope at a flux
312
(6)
density equal to or greater than S should conform to the $ power relation N oc S~

16-19

* Answers to starred problems are given in Appendix C.

MOVING SYSTEMS AND SPACE-TIME

PROBLEMS

783

Discuss the effect of the following factors in producing a reduction in the observed number

N at low values of flux density 5:

(3) velocity

(1) radio-telescope resolution, (2) radio-telescope sensitivity,

of recession of sources, (4) absorption by the medium, and (5) actual reduction

in source density at large distances.

source increases in proportion to

its

In factor 3 assume that the velocity of recession of a

distance.

At the earth's distance from the sun the sun's electromagnetic radiation amounts
2
(mostly in optical and infrared regions) while the solar wind consists
mainly of protons with velocities of about 1 Mm s _1 and densities of about 2 x 10 7 protons
per cubic meter, (a) Compare the radiation pressure in newtons per square meter of the
16-21

to about 1.5

kW m~

electromagnetic radiation with that of the solar wind on a

flat uncharged conducting sheet

normal to the radiation and wind direction. (b) Can either the radiation or the wind be used
to propel an interplanetary vehicle? (c) If so how much sail (square meters) is required to
develop 100 N thrust downwind, and (d) how much is required cross wind (traveling perpendicular to wind direction) ?

16-22

A plane 5-GHz wave with power flux (Poynting vector) of 10 W m" 2 is travel-

The wave is incident normally on a large flat perfectly conductxz plane). Find the pressure on the sheet if the sheet is
(a) stationary and (b) moving in the positive y direction with a velocity v = c/5.
(c) What is
the Poynting vector of the reflected wave for the two cases ? (d) What is the frequency of the
reflected wave for the two cases?
ing in the positive y direction.

ing sheet (sheet

is

parallel to the

A spacecraft
= 0.8c. After

16-23

velocity v

it

leaves the earth

and

away from the earth

at

reaches a distance 8 light-years from the earth, a radio signal

is

transmitted from the craft back to the earth,

(a)

travels directly

How long after the craft departs is the signal

on earth? (b) How much time has elapsed on the spacecraft since departure when
the signal is sent from it ? (c) From the spacecraft point of view how much time appears to
have elapsed on the earth when the signal is sent, and (d) when the signal is received there?
received

(e)

Draw

a space-time diagram with appropriate lines to

illustrate the

problem and

its

solution.

16-24

A spacecraft receding from an observer at a velocity v emits a radio signal with

Poynting vector
vector

is

5em measured in the rest frame of the craft.

S"- S'"
16-25
craft,

show

If the craft

'

T+Wcj

(Wm

>

of Prob. 16-24 emits the signal at a frequency/, as measured on the

r
=/.

is

given by

Vl- (v/cY
,,.,..
1

is

~Sm

\-{vlcY

that the observed frequency

Jo

where z

Show that the observed Poynting

given by

or

+ (Vic)

the doppler, or red, shift.

_

= 1 -+ (vie)1 = + z
(v/c)
1

f,

Thus, the red shift

l

VT

+ (v/c) _
- (v/c) 2

is

784

[CHAP. 16

ELECTROMAGNETICS

From

16-26

Probs. 16-24 and 16-25

S, m

From

16-27

= S.M + z)

Probs. 16-24 and 16-25

show

show

Km = SiM + z)
16-28

If the radiation

from the

craft

that

(Wm- 2)

that

(Wm-'Hzis

emitted over a continuous spectrum (not

monochromatic) with spectral index a defined by S oc/ *, show that the Poynting vector
_2
the
craft at the emitted frequency/, is
frame
of
in
the
rest
Hz ')
per unit bandwidth (W m

Sio=Sib>(l

Two

16-29

velocity of 0.7c as

+ 2)" +1

(Wm-'Hz"')

spaceships traveling in opposite directions

measured by an earth observer.

What

is

away from the earth have a

the velocity of one spaceship

as observed from the other?

Show that

16-30

ductor
is

is

= pH

of a wave incident on a perfect con-

equal to the force on a unit area caused by the reversal in

reflected.

energy

the radiation pressure

Hint:

= mc 2

16-31

impedance

momentum when the wave

Consider the mass equivalent of a wave column using Einstein's relation

What

Z= Z

is

the radiation pressure of a

so that the

wave is

wave

incident

entirely absorbed?

on a medium with

intrinsic

APPENDIX
UNITS, CONSTANTS,

AND A FEW

MATHEMATICAL RELATIONS

A-l

UNITS

Multiples and submultiples of the basic units are designated by the prefixes listed in
Table
that in the SI system, multiples and submultiples are in steps of 10 3 or 10" 3

Note

or smaller quantities are designated by the exponential form.

and

For a discussion of dimensions

units see Sees. 1-2 to 1-4.

Table

Prefix

Abbreviation

Magnitude

Derivation

tera (ter'a)

IO 12

Greek

giga (ji'ga)
mega (meg'a)

10

Latin gigas, " giant

Greek megas, " great

kilo (kfl'6)

milli (mil'

micro (mi' kr6)

nano (nan' 6)
pico (pe' ko)
femto(fem'to)

atto (at' to)

/*

n
P

10 6
10 3
io- 3
io- 6
io-
io- 12
io- 15
IO" 18

1.

Larger

Greek

(eras,

"monster"

chilioi,

"a thousand"

Latin mille, "a thousand"

Greek mikvos, " small
Greek nanos, " dwarf"
Spanish pico, " small quantity
Danish femten, " fifteen "

Danish

atten,

"eighteen"

786

[APP.

ELECTROMAGNETICS

In Table 2 dimensions or quantities

betically

commonly used in electromagnetics

under the headings Fundamental, Mechanical,

are listed alpha-

and Magnetic. In the first

the second column the common

Electrical,

column the name of the dimension or quantity is given and in

symbol for designating it. In the third column (Description) the dimension is described in
terms of the fundamental dimensions (mass, length, time, and electric current) or other
secondary dimensions. The fourth column (SI unit) lists the SI unit and abbreviation,

column gives equivalent

dimensions by means of the symbols

and the

fifth

units.

The

M (mass), L

last

column

indicates the fundamental

(length), T(time),

and I

(electric current).

Table 2
Fundamental units

SI unit

Name of

and

dimension

abbrevia-

menEquivalent units

sion

ampere

6.25 electron

(A)

charges per

or quantity

Symbol

Description

tion

Current

charge
time

Di-

second

=
s

Length

Mass

L,l

M,m

mm

meter
(m)

1,000

kilogram

1,000 g

100

cm

(kg)

second

Time

T,t

(s)

Amin=

3^00

-86,400^

hr

Table

2(Continued)

Mechanical units

SI unit

Name

of
dimension
or quantity

Symbol

Description

Acceleration

velocity

length

meter

time

time 2

second 2

kg

and

Di-

abbrevia-

men-

tion

Equivalent units

sion

T2

(msi
Area

A,

a, s

length 2

meter 2

L1

(m 2 )

Energy or work

force

length

= power
Energy density

x time

Nm = Ws = VC ML

joule

= 10 7 ergs
= 10 s dynes mm

(J)

T2

energy

joule

tso erg

volume

meter 3

mm 3

LT 2

(Jm- 3 )
F

Force

mass x acceleration

kgm

newton
(N)

~m

= 10*
Frequency

Length

T2

dynes

cycles

hertz

cycles

time

(Hz)

L,l

M,m

Mass

ML

mm

meter
(m)

1,000

kilogram

1,000

kgm2

ML 2

T2

100

cm

(kg)

Moment

(torque)

force

length

newtonmeter

(Nm)
Momentum

rm

mass x velocity

= force

x time

energy

newtonsecond

kg

(Ns)

J s

ML

velocity

Period

Power

force

second

frequency

length

time

(s)

watt

(W)

Nm
s

energy
time

ML 2
T3

kgm 2
s

787

Table 2

Mechanical

units

(Continued)
SI unit

Name

and

of
dimension

Di-

men-

abbrevia-

or quantity

Symbol

Time

T,t

Equivalent units

tion

Description

second
(s)

mm= 3706 hr
1

6O

sion

~
Velocity (velocity

of light in

vacuum

= 300

Mms-

length

meter

time

second

day

L
T

(ms- 1 )

Volume

86,400

L3

meter 3

length 3

(m 3 )
Electrical units

Admittance

mho

impedance
Capacitance

potential

J2 T 3

C2

ML 2

farad
(F)

charge

(U)

C = C2 As
~ V
V
J
= 9 x 10" cm

J1 T 4

ML 2

(cgs esu)

Charge

Q,9

current

x time

coulomb

6.25

10' 8 elec-

= As
=3 x
=
Charge (volume)

density

Charge

volume

V-D

coulomb
meter 3

(C

Conductance

resistance

Conductivity-

IT

tron charges

(C)

10'

cgs esu
0.1 cgs

emu

As

IT

L3

m- 3)

mho

C2

IT 3

(U)

V~ Js

ML 2

mho

resistivity

meter

I 2 T*

CI

ML 3

(Urn" 1 )
Current

hi

charge
time

ampere
(A)

-=3
s

=
t S is the SI abbreviation for Siemens, used outside the United States for

mho.

x 10 9

cgs esu
0.1 cgs emu

(Continued)

Table 2

Electrical units

SI unit

Name

and

Di-

of
dimension
or quantity

Symbol

Description

tion

Current density

current

ampere

area

meter 2

sm2

L2

Asm

LIT

men-

abbrevia-

P(=<70

coulombmeter

charge x length

sion

m- 2)

(A
Dipole moment

Equivalent units

(Cm)

Emf

IT

Energy density

y.

(electric)

jV

rf i

volt

Wb

(V)

~C

t4o

erg

energy

joule

volume

meter 3

mm 3

ML 2
IT 3

M
LT 2

(Jm- 3 )
Field intensity
(E vector)

potential

force

volt

length

charge

meter

"Cm
=

(Vm- 1 )

ML
IT 3

X 10-*

cgsesu

= 10s cgs emu

Flux

>l>

Flux density
(displacement)

(D

charge

ds

coulomb

charge

coulomb

As

area

meter 2

m2

vector)

(C

Impedance

'

potential

P-.

density

ohm

charge

coulomb

length

meter

(CmPermittivity
(dielectric

con-

stant)
(for
<<,

vacuum,

IT

A
m 2 s _1

IT

L2

m- 2)

current

Linear charge

As

(C)

ML2

As"

T3

IT

capacitance

farad

I 2 T*

length

meter

Vm

ML 3

(Fm-

8.85 pF
w 10-'/36ti-

Fm"

789

Table 2

Electrical

units(Continued)
SI unit

Name of

and

Di-

dimension
or quantity

men-

Symbol

Description

abbreviation

Equivalent units

sion

Polarization

dipole

coulomb

As

IT

meter 2

m2

L2

moment

volume

(CmPotential

work

volt

charge

(V)

Nm _ Ws
c

-W

ML 2
IT 3

Wb

A~

= 3^ cgsesu
= 10 8 cgs emu
Poynting vector

power

watt

area

meter 2

sm 2

T3

(Wm- 2
Radiation

intensity

watt
steradian

potential

current

Relative
permittivity

ML 2

power
unit solid angle

(WsrReactance

*T

ratio

T3

ohm

ML 2

(O)

I 2T3
di-

men-

sionless

Resistance

potential

current

ohm

(CI)

A~C

ML 2

Js
2

I 2T3

= ixlO-"
cgsesu

= 10" ' cgsemu

Resistivity

resistance

x length
1

~ conductivity
Sheet-current
density

ohm-

Vm

meter,

ML
I

(Qm)

current

ampere

length

meter

(Am- 1 )

mr x m
2

T3

Table 2

Electrical

units(Continued)

SI unit

Name

and

of
dimension
or quantity

Symbol

Susceptance

Wavelength

Description

moment

(magnetic)

m
(=eo

wm

Flux (magnetic)

A
V

J2J-3

ML 2
L

units

amperemeter 2

pole strength x length

current x area

(Am2)

Cm

II 2

magnetic flux density

energy

joule

rwerg

volume

meter 3

mm 3

+~

(B vector)

mho

meter
(m)

If

Flux density

sion

length

torque

(magnetic)

Equivalent units

(U)

~~

Energy density

men-

tion

reactance

Magnetic

Dipole

Di-

abbrevia-

weber

8 "*

force

current

telsa (T)

force

pole

(Wb)

moment

weber
meter 2

magnetic flux

(Wb

area

flux

turns

IT 2

10"

Mxt

(cgs

emu)

Vs

m2

Am

IT 1

10 4

Gt

(cgs

m- 2 )
Flux linkage

ML 2

nm

T
=

Vs =

emu)

weber-

ML2

turn

IT 2

(Wb
turn)

field

(H vector)

mmf

ampere

length

meter

(Am- 1 )

W
N
Wb~Vm

= 4tt

x lO" 3
Oet (cgs emu)

= 400ir
gammas
t

Mx, G, and Oe are

SI abbreviations for maxwell, gauss,

and

oersted.

791

Table 2

Magnetic units(Continued)

SI unit

Name of

and

dimension
or quantity

abbrevia-

Symbol

Description

tion

Inductance

magnetic flux linkage

henry
(H)

current

Di-

menEquivalent units

Wb _ J
A A2
= Qs
= ixlO-"

sion

ML2
J2J-2

cgsesu

cm

10*
(cgs

Magnetization
(magnetic

magnetic

moment

volume

polarization)

Mmf

Am

ampere

emu)

Am

m3 ~ m2

meter

(Am" 1 )

JH.rt

ampere-

turn

(A turn)
Permeability

/*

vacuum
= 400n-nH
m->)

(for
|Uo

Permeance

inductance

henry

Wb

length

meter

Am Am

J2J-2

Wb
A

J2J-2

(Hm0>

magnetic flux

mmf

ML

Vs

henry
(H)

ML 2

reluctance

Pole density

P-

pole strength

ampere

volume

meter 2

L2

current

(Am" 2)

area

= V-H= - V-M
Pole strength

Qm,t}m

current

length

amperemeter

Cm

IL

(Am)
Potential

(magnetic)

jH.rfl

ampere
(A)

Wb

V ~

C
s

(forH)

(cgs

Ob is

the SI abbreviation for gilbert.

emu)

Table 2

(Continued)

Magnetic units

SI unit

Name of

and

Di-

dimension

abbrevia-

men-

or quantity

Symbol

Relative
permeability

p.

Description

ratio

tion

Equivalent units

sion

Di-

V-

men-

Mo

sionless

Reluctance

St

mmf

magnetic flux
1

henry

A
Wb

J2J.1

ML2

(H- 1 )

permeance
Vector potential

current

x permeability

Weber
meter

HA N
m

ML
IT 2

(Wb

m- 1 )

793

794

[APP.

ELECTROMAGNETICS

CONSTANTS AND CONVERSIONS

A-2

Symbol

More accurate
value*

or abbre-

Quantity

Astronomical unit
Boltzmann's constant
Earth mass
Earth radius (average)
Electron charge

viation

Nominal value

AU

1.5

1.38
6.0

x 10" km
x 10"" J
x 10 24 kg

-1.60 x 10- 19
10- 31

Electron rest mass

Electron charge-to-

9.11

mass ratio
Flux unit

elm

1.76

x 10 11

fu

Foot
Foot squared

ft

ft

x 10" 23

x 10- 24

-1.602
31
9.10956 x 10"

kg

1.758803 x 10'
10" 26 (by definition)
0.3048
2
9.290 x 10"

kg" 1

Wm-'Hz"
0.30 m (1 m = 3.281 ft)
2
2
2
9.3 x 10" m (1 m = 10.76
10~ 26

300Mm(l

Light -second

5.98

Mm

6.37

1.496
1.38062

K- 1

Mm s-12

ft )

Gft)

(1 Gfts" 1 )

299.7925
2
9.4605 x 10'

Light, velocity of
Light-year
Logarithm (natural)

300

LY

9.46

base
Logarithm, reciprocal
of base

2.72

2.71828

0.368

0.36788

= 2.3 log x
log x = 0.43 In x
1.61 km(l km = 0.6214 mi)

In x = 2.3026 log x
log x = 0.4343 In x
1.60935

Ue

In

Logarithm conversion
Mile

mi

(statute)

Moon

distance

x 10

km

Mm
1.738 Mm

(average)

380

Moon mass
Moon radius

6.7

(average)

x 10 22 kg

nHm-

Permeability of vacuum
Permittivity of vacuum

/i.0

1,260

Pi

77

8.85
3.14

Pi squared

IT

Planck's constant

9.87
6.63

rest mass
Radian
Space, impedance of

rad

1.67 x 10- 27
57.3

376.7

Proton

pFm"

4007T (exact value)

8.854185 = l//x c
3.1415927
9.8696044
34
6.62620 x 10"

x 10- 34 Js

27
1.67261 x 10"
57.2958
376.7304
ft c
12.5664
4tt
41,252.96
3.04617 x 10"

kg

(a 120rr)

12.6 sr

Sphere, solid angle

2
41,253 deg

x 10" 4

Square degree

deg 2

3.05

Steradian (square
radian)
Sun, distance

sr

2
3,283 deg

AU

1.5

Sun mass
Sun radius (average)

2.0

sr

(180/tt)

x 10 s km
x 10 30 kg

3,282.81

x 10"
x 10 30

1.496
1.99

Mm

695.3

*o
V2
Vi

700

1.73

1.73205

Vio

3.16

3.16228

1.41421

1.414

note that the values for

Regarding permittivity and space impedance
units as nominal value.
the measured value of c (velocity
these quantities are determined by the exact (definition) value of fi and

Same

of light).

UNITS, CONSTANTS,

SEC. A-4]

A-3

AND A FEW MATHEMATICAL RELATIONS

TRIGONOMETRIC RELATIONS
y) = sin x cos y cos x sin y
y) = cos x cos T sin x sin y
sin(jc + y) + sin(x y) = 2 sin x cos y
cos(x +
+ cos(x y) = 2 cos x cos y
sin(x + y)~ sin(x ^) = 2 cos x sin ^
cos(x + y) cos(x jO = 2 sin x sin ^
sin 2x = 2 sin x cos x
cos 2x = cos 2 x sin 2 x = 2 cos 2 x = 1 2 sin 2 x
cos x = 2 cos 2 ix = 1 2 sin ix
sin x = 2 sin ix cos ix
2
2
sin x + cos x = 1
tan x + tan
sin(x

cos(jc

>>

.y)

-X)

1-tan xtany
tanx- tan y
+tan x tany
-

-*)
1

2tanx

an 2*
1

inx

A-4

tan 2 x

3
= x X

"17

X5
5!"

7!

X*
X'
X"
__
+ ___+...

cosx

=l

tanx

= x+ T

x3

2x
+
+ -r^5+'
5

17x 7

62x

HYPERBOLIC RELATIONS
ex

sinhx

X
X
X _L
-e~' =
X +
+ + +
I

e"

coshx

3!

5!

'

7!

+ e~"
2

sinhx

tanhx

coshx
coshx

cothx

sinh

tanh x

jy) = sinh x cos yj cosh x sin y

/y) = cosh x cos y j sinh x sin
coshfix) = K + Jx + e~ ") = cos x
. de Moivre s theorem
^
sinh(/x) =K
~ e~ y*) =y sin x
e jx _ cos x y sin x
sinh(x

cosh(x
.

j>

'

'

..

795

796

ELECTROMAGNETICS

[APP.

X'

e*

= cosh x + sinh x
= cosh x sinh x

e~*

coshx = cosjx
x %\r\jx

j sinh

2x

sinh

sin

ui
tanh(x

jy)\ = r-^-T^.
J cosh 2* + ycos 2y
cosh 2x + cos 2y

coth(x

sinh 2x
sin 2v
/y) = cosh 2x cos j-cosh 2x cos

2y

A-5

2y

LOGARITHMIC RELATIONS
x = log jc
common logarithm
natural logarithm
x = In x
log, x = 0.4343 log, x = 0.4343 In x
In x = log c x 2.3026 logio x
logio

log e

A-6
(8

is

= 2.71828

APPROXIMATION FORMULAS FOR SMALL QUANTITIES

a small quantity compared with unity)

8) = 28
8)" = nh
Vl + 8 = + J8
2

(1
(1

1
1

= 1-48
Vl+8
e = + 8
ln(l + 8) = 8
1

/.(8)=-^-

n\2

where /

is

Bessel function of order n.

for|8|l

Thus

M8)=l

UNITS, CONSTANTS,

SEC. A-9]

AND A FEW MATHEMATICAL RELATIONS

797

SERIES

A-7

Binomial
(x

+ y)' = x" + nx"'

Taylor's: /(x

A-8
If ax

3!

2!

+ ^)=A*) +- ZrT + - S r- J! +-^-35 +

SOLUTION OF QUADRATIC EQUATION

+ bx + c = 0, then
-b Vb - 4ac
2

2a

A-9

VECTOR IDENTITIES

(F, /, G,

and g are

F
where 6

= cos" '(F

H are vector functions.)

G = FG cos d

scalar (or dot) product

G = hFG sin

vector (or cross) product

G)/FG

FX
where 6

and

scalar functions; F, G,

= sin"'(F X G)/iiFG
= unit vector normal to
V (V x F) =

plane containing

F and

G
0)

V-V/"=V /
2

(2)

VxV/=0
*(/"+*)

(3)

=v/-+v*

V (F + G) = V F + V G
V X (F + G) = V x F + V X G

(5)

(6)

v(fg)=gW+fVs
V.(/G) = G-(V/)+/(V.G)
V x (/G) = W) x G +/(V x G)

< 7)

(8)

(9)

VX(VxF)=V(V-F)-V F
F=VT = iV 2F1 + yV !F,+ zV 2F,
V (F X G) = G (V x F) - F (V x G)
F-(GxH) = G(HXF) = H(FXG)
V X (F X G) = F(V G) - G(V F) + (G V)F - (F V)G
V(F-G) = (F-V)G + (G-v)F + FX(VxG) + GX(vxF)
2

(10)
(11)

(12)
(13)

(14)
(1 5)

798

ELECTROMAGNETICS

fA pp.

In the following three relations the surface

side

(left side

of equation) encloses the volume (right

of equation)

I fds
J

<j>

j>

Vfdv

(16)

JV

ds

V F dv

divergence theorem

&XFds = j V xFdv

(18)

In the following two formulas the line integral (left-side of equation)

which bounds an unclosed surface

(j>

(right side

jfdl

=j

=|

d\

(1 7)

is

along a closed path

of equation)

X Vfds

(19)

(V x F) ds

Stokes' theorem

(20)

RECURRENCE RELATIONS FOR BESSEL FUNCTIONS

A-10

Equations expressing Bessel functions or their derivatives in terms of Bessel functions of the
different order are called recurrence formulas or relations.
A few of these formulas

same or

are listed below for reference.

dJM)

du

u
V

du

is,

the slope of the

(1)

(3)

y- + l(w) + J- -'(M

(4)

/()

= /() - /. _ i()

(5)

= n = 0,

-- = -/,()
That

= /,- i()- 7, + ,()]

/. + 1()

for v

(2)

we have

'
'v

+ /. i()

JM = y

(1)

W"/

dJM)
dJy (u)

From

. ,

()

curve

is

equal to

/i().

(6)

SEC. A-]

A-ll

UNITS, CONS! ANTS,

1]

799

AND A FEW MATHfMATIl'AL RELATIONS

COORDINATE DIAGRAMS

Diagrams

for rectangular, cylindrical,

and

spherical coordinates arc

shown

in

Fig.

A-l.

shown in Fig.
The construction for elemental volumes and lengths in the three systems are
A-2.

See inside back cover for gradient, divergence,

curl,

and laplacian

rectangular,

in

cylindrical, spherical, and general curvilinear coordinates.

Plane z const

aney^const

RECTANGULAR

Cylinder

Plane

Conetf

<f

= const

= const

= const

SPHERICAL

Sphere

FIGURE

A-l

Rectangular, cylindrical, and spherical

coordinate systems.

Plane0 = const

?'

= const

800

ELECTROMAGNETICS

[APP.

dv

dx dy dz

dl= ydx' + dy' + di 2

Rectangular

dv

d;i-dtt>dz

dl= \^+{rd4>)* + d*
Cylindrical

dv=

dv

r2

sing d8 d<b

dl- ydr + (rdi)- +

2

dl

FIGURE

(,-

sin

d<t>)"

sin 8 dtj>

Spherical

A-2

Elemental volumes and lengths

nate systems.

in rectangular, cylindrical,

and spherical coordi-

APPENDIX

BIBLIOGRAPHY

Footnotes throughout the text include

The following list

specific topics.

many

references to important articles or

books on

provides a few selected general references for supplemental

reading.

billingham,

(ed.):

J.

A Design Study of a System for Detecting ExResearch Center (Code LT), MofFett Field, Calif.,

"Project Cyclops:

traterrestrial Life,"

NASA/Ames

1972.

" Ferromagnetism," D. Van Nostrand Company, Inc., Princeton, N.J., 1951.

and e. d. isaac: "Magnetic Domains and Techniques for Their Observation,"
Academic Press Inc., New York, 1966.
elliot, r. s. " Electromagnetics," McGraw-Hill Book Company, New York, 1 966. Provides

bozorth,
cury,

r. m.

r.,

excellent historical insight.

feynman, r. p., r. b. leighton, and m. sands: " Feynman Lectures on Physics," AddisonWesley Publishing Company, Reading, Mass., 1964.
halliday, d., and R. resnick: "Physics," John Wiley & Sons, Inc., New York, 1962.
Harrington, r. f. " Time-harmonic Electromagnetic Fields," McGraw-Hill Book Company,
:

New York,
hayt, w.

h., jr.:

1961.

"Engineering Electromagnetics," 2d

ed.,

McGraw-Hill Book Company,

New York, 1967.

jasik, h. (ed.) :

1961.

"Antenna Engineering Handbook," McGraw-Hill Book Company, New York,

802

ELECTROMAGNETICS

Jordan,

c, and

e.

k. d.

[APP. B

balmain: "Electromagnetic Waves and Radiative Systems," 2d

Englewood Cliffs, N.J., 1966.

"The Theory of Linear Antennas," Harvard University

ed.,

Prentice-Hall, Inc.,

king, r. w.

p.:

Press,

Cambridge,

Mass., 1956.

MAONUsson,
Inc.,

moon,

p.,

c: "Transmission

p.

Lines and

Wave

Propagation," 2d ed., Allyn and Bacon,

Boston, 1970.

and d.

e.

spencer: "Foundations of Electrodynamics," D. Van Nostrand, Company,

Inc., Princeton, N.J., 1960.

moore,

a. d.

"Fundamentals of

Electrical Design,"

McGraw-Hill Book Co.,

New

York,

1927.
r., and r. e. collin: "Principles and Applications of Electromagnetic Fields,"
McGraw-Hill Book Company, New York, 1961.
purcell, e. m. " Berkeley Physics Course," vol. 2, " Electricity and Magnetism," McGraw-

plonsey,

Hill

ramo,

Book Company, New York,

s., J.

r.

whinnery, and

ics,"

3d

ed.,

schelkunoff,

1965.

vanduzer " Fields and Waves

:

John Wiley & Sons, Inc., New York, 1965.

a. "Applied Mathematics for Engineers and

s.

Company,
:

R.

in

Communication Electron-

Scientists,"

D. Van Nostrand

Inc., Princeton, N.J., 1948.

"Electromagnetic Waves," D. Van Nostrand Company, Inc., Princeton, N.J., 1943.

a. " Partial Differential Equations in Physics," Academic Press Inc., New York,

sommerfeld,

1949.

smythe, w.

r.: "Static

New York,
stratton,

j.

watson, g.

a.:

n.:

and Dynamic

Electricity,"

2d

ed.,

McGraw-Hill Book Company,

1950.

"Electromagnetic Theory," McGraw-Hill Book Company,

"Theory of

Bessel Functions,"

Cambridge University

New York,

Press,

1941.

London, 1944.

APPENDIX

ANSWERS TO STARRED PROBLEMS

CHAPTER
1-1

1-2

velocity,

3
second

..

potentml,

ML

2-4

2-7
2-8

TJ
5

ML

volts; constant,

newtons; current density,

j
Xj

CHAPTER

2-3

ratio, dimensionless

ML

2-2

work, -=r- , joules ;

charge, IT, coulombs; potential, -=rf > v0,ts ; electric field intensity,
volts

2-7

amperes

.
;

iiicxcr

F = 900 nN (repulsive)
= 4.0 (angle with x axis)
1*1 = 14.3 nN;
E = 9kVm- ;K = 900V
E = 0; K=l/2irea
E =^-yl2Vm- ;K =
= &0.424 + y0.566 + 20.707
1

fl

meters

^, -j

j-

force,

ML

force, -=j*

newtons

^,

804

ELECTROMAGNETICS

2-10

at (0,0),

2-11

at (0,0),

2-14

(a)

Vm"
2-20

E = -Sco
E = -&12;

E = -foo (0,4), E = -f(0.75) Vm"'

E = -il2 - ?42; (5,0), E = -xl2; (5,3), E = -il2 - $42

(4,0),

App c

(0,3),

y = 303mm; (6) Qy =
when r<R, E = t90r kV m" 1 V = 900 + 45,000(0.01
when r > R, E = r(90/r 2 ) V m" 1 F = (90/r) V
,

-r )V
2

2-34

= 4ne R 3
(a)
G = 32C; (6) G = 8C; (c) G = 24C; (rf) G=164C
(a)
Q, = 518 TC, Q m = 6.04 TC; (6) fi/ = Vg^ / = 2.23 x 10" C m

2-38

(a)

2-25

2-29

<ji

E = 0(r^r

61 + 62

E=
j
4ne r

y = J-(Ql +
47T \r

ri

47re

\r 2

/ e,

K_

47T

2-40

ri

^r<r

2 ),

Ql)

(.

r>r 2

{b)

-Q

),E =

<ri,

J
ri

<r <r 2

r /

+ 6A
r

r>r 2

K=3.22xl0 I5 V

CHAPTER

3-10

= 800mCm- 2 (6) G'=80mCm

= 500mCm- 2 (b) D = 670 mCm" 2 (c) > = 670mCm- 2
(rf)
=18.8GVm = 75.4GVm"
left half: E = 10 kV m"
O = 265 nC m- 2 P = 177 nC m" 2
right half: E = 10 kV m"
D = 88.5 nC m" 2 P =
(a)
= 15kVm- = 5kVm- P = 0,
P = 88.5nCm- 2 Z) = D = 133 nCm- 2
(b)
V. = 25 + lS.OOOO' - 0.005) (V), K = Sy (kV)
(a)
= 1.13kVmE2 =226Vm~
(b)
F=111N
e = 2.57

3-12

in the slab,

3-1

3-2

(a)

i'

(a)

/>

5-i

',

i-4

3-6

in air,

g=

D = f (1

t (1

3-16

3-5

(<?)

C=
(/)

()

4.02

o6

39.3 pF,

nF

+
dt,

- e ;) +
r

eP )

_,

rfe,

(b)

,
'

E = t(l

=- ?(1

44.2 pF,

d r

,
'

P=

<(1

~ <*) + de,

\ + rfe, ,/> =

er)

(c)

<=,)

_,

58.7 pF,

(rf)

87.4 pF,

(e)

171.0 pF,

ANSWERS TO STARRED PROBLEMS

CHAP. 4]

3-20

3-24

(a)

W (before) = 5000 W (after) = 1.000Q J;

(6)

a mechanical energy of 50QQ J was supplied

J,

E = 0,V= Ane

3-26

K=-r=-,ln

3-33

(6)

3-34

4ne

ratio

*m

xi

Xi
= 2.67;

K=15V; (</)
L = 150mm;

(c)

C//

= 70.8 pF m"
K = 51.78 kV,

(b)
I, = 155 mm;
= 52.62 kV; (a ) K = 45.97 kV; (e) K= 59.95 kV
V-F=0; (b) V-G = -acosaz
(a)
2
2
2
2
(<2)
0;
3+12j> 2 +10r; (b) 2y z; (c) 14x yz + z~
(a)
1
2
6- r
2z; (/) 3r-" -r- cos2^; (g) r' + /-'2sin <cos
(e)
Sz + r-'sin" 0sinccos20
(A)

(a)

(c)

F.

5-59
5-44

'

cos

<

*>

G
= - 47TC/J
2

CHAPTER
4-2

=21.93

/xO;

jR

(a*)

/=1.43MAm-

4-5

35.6kWm2=12 A

4-4

(a)

2* 12

(c)

2J b/

= 219 ptl, G 3 * =4.64 MU;

(a)
(c)

average current path longer in L-shaped bar

(a)

2?

(d)

average current path longer for (a) than

2f 3 4

1
,

Jx

= 509 kA m" 2

Ex

= 8.93 mV m"

(ft)

= 13.2^;

(6)

2*

= 12.8^n

= 4 A (flowing toward junction)

R = 107fiO; (6) 2J = 63/xO;

4.5

/4

4-9

(a)

(a)

2?=iOOmVm-';

4-29

2*//=17.6knm-

t-20

< fl)

4-24

(A)

i>

(e)

4-26

= 3.51 ^n, G 12 = 285 kU; (6)

= 1 10 nO, G6/ = 9.12 MU
2
E2 = 3.13 mV m"
>, = 178 kA mratio = 2.86
ratio = 1.13;
2J = 3.95/xQ;

(d)

4-6

(c)
7 = 2.28kA;
=
25.0mVm
(e)
<f) 2>=114W;
^ms
=
71.3
=
(i)
410kJ;
Jf

G= 45.61 kU;

(#)

4-5

(6)

(a)

(a)

/?

<r

||

(2>)

7f

(c)

(b)

= 14.3^0;
but shorter than

(c)

/=171mm

(c)

=910 Vm"

= 6.3 (from normal)

1
,

= 2lb ,n ^
= 2.4Um _1

(W
(*)

R=

<,*(rJ-r >y
= 4 and v = 2 m

(C)

_1

*=

^7l|^)

805

806

ELECTROMAGNETICS

Ap

CHAPTER 5
5-1
(a)
F= 100 mN (repulsive);
5-2
5-4
5-5

(b)

F= 100 mN (attractive)

F//=-i40 + y30Nm-'
F=-y800/^N
(a)
./-,= 9.73 Wb;
(6)
^m = 2nkrV/b
(a)

= 62.5,t*T;

(a)

0;

5-14

7* = 31.4mT;
(6)
(c)
5 = 62.1 MT
r=3.53Nm; (c) m=0.884Am 2
(a)
B = 12.6mT; (A) ff=10kAm-'; (c) *, = 62.8 Jm" 3
{d) K=\0kKm(e)
W = ll.lmJ; (/) L=5.6mH
B=\JyLT
= /* 1/4r (r = focal length)

5-17

(a)

5-18

r=3.77/*Nm

5-7
.5-70
.5-77

(b)

5-13

(b)

Z,

(a)

5-7

r=lmNm
H=-\0kKm(a)
#i = 1.59 MA m1

6-10

= 0.92
5 1 0.5 m south

6-11

31- fri-MoM

*i ln(ju,/^
(J-20

5-27

5-25

5-27

6-3

6-8

= 36Am 2

CHAPTER

6-5

/w

= 40mH
H = 0; (b) 7/ = 41.8Am-; (c) fl =159.0AmV x F = -i2j> + y2*; circle with center at origin

5-22
5-27
5-32

r=144Nm;

1
;

6.7

(b)

TTj

=7.96

kAm"

1 .7 m deep

w = 2.83mAm 2
/T = 326MAmA7 = 2.4 kA turns
(a)
^ =4.29 mH,^ = 233 H-;
1

5-30

1217.5 kg weight

6-33

F=78.3kN

CHAPTER
7-7

(a)

7-2

(a)

(b)
(rf)

<?>

(6)

= 31.5 juH,

3t

= 31,746 H"

=lkVm-';

= 885 nCm"'; (c) C// = 7.08 nFm"

Vmi)i
pCm- 2 (c) 7J, =50nT, B2 = 16.7 nT;
ft=39.8mAm- ,ff1 = 13.3mAm-'; (e) C// = 50.6 pFm"
Z.// = 219.7 nHm= 505 pCm-; (A) ^// = 2.2 nWbm{g)
7i,

>

(6)

=227.5 Vm-S/fj =75.8

=2.01 nCm- 2 , 2 =671
1

(/)

,/-//

CHAp

gi

7-4

ANSWERS TO STARRED PROBLEMS

'

Vm- ,i)i= 1.17 nCm" 1 #i=10 A m- ,Bi = 12.56 /iT,

=
E2 83.2 V m" D 3 = 736 pC m" H2 = 6.67 A m" B 2 = 8.38 /*T,
E = 35.8 V m- D, = 316 pC m" 2 H = 3.33 A m" B =4.19 jtT
= angle from axis; (b) turns per meter (of
sheet current K* oc sin 0, where
(a)
,

(a)

197.8

7-77

continuous winding) should be proportional to sin 6

7-12

7-75

= - f/io fco C = Ho *o 7? 3 /3
K = C, In tan(0/2) + C 2 where C = Ki/In[tan(ff,/2)/tan(tf /2)] and
(a)
- K,/{[ln tan(0,/2)/ln tan(0/2)]- 1} (6) C, = Ki/ln tan(0,/2), C 2 =
=
C,
=
K Fi sin(ira/20) (a is the angle made with the V = sheet)
Ci

7-22
7-23

Q
V = Z

K=

r3

<r<co

_e>-l)r + r

er'^

4weo

(,-iy,

^=4

^^
r^

+ r,

(e r

-l)r 2

r,<r<r 2

rr 2

e,r 2 r 3

+ r3

e,r 2 r 3

47TEO

r2

-r,-|

< r < ri

rir 2 J

+ a 2 )(r 2 - b 2 ) + e
2
e(b 2 + a ) + e
f
2 -q 2
2
-2g 6 6ocosfl(r
)
Kl_
2
2
2
2
"r[c(6 + a + e(6 -a )]
-, cos

7-27

6 e(6 2

Vo-

(6

(b

- a 2 )(r 2 + b 2 )
-a 2
)

7-29

(a)

ps
7-34

K=-[l-(a/r) 3 ]7io rcos0;

=e

(a)

7io[l

- (a/r) 3

= 3f E

cos

(on sphere),

K = (K +K,)g+(F'i->'1 )[|^j
1

cos

li(^% (cos .)--]

5

d = 6.24s
= (po/6Jti)(x, 3 - x 3 )

7-35

{b)

7-42

CHAPTER

c
dB/dt = 66.6 mTs"'
U = 7.5mV;
= 125 turns
c
c
*<0, U=0;0<x<80, U=400/xV;80<x<100, U=0;

8-1

(a)

-2

TV

(/>)

100
-4

(o)

S-5

-07

^ x < 120, nj = -400 /iV

clockwise;

(Z>)

8-7

= 78.5 mV
nj = 2vBVR 2 - r 2
c
U = n-TVB7i 2
(a)

-70

/f=1.27mAm-

8-6

(b)

(on sheet)

-f6 (^)%3(cos0) +

3-3

807

counterclockwise

(6)

T=ilBR 2

(c)

clockwise

808

ELECTROMAGNETICS
[APP.

8-11

8-13
8-15

5-/9
8-23

V = 1.27V; (b) nj=1.27sin(10rt) - 18.86cos(l(W) V

= 3.68 /*V
M = 7.96mH
nj=o
(a)
J = -tpH cos(>t-px); (b) J = xtfH, sin 2*
sinM - ft,)]
-y[2// cos 2* cos(o* - y)] + z[,8//x sin 2x cos(<o/ - fly)]
P = 5.22W
(a)

17

8-26

CHAPTER

9-5

because isolated magnetic charges do not exist

9-8

(a)

J = (/ /7r/f 2 )sin
/<=0forallr;

(6)

9-9

(a)

9-/0

74
Jt

9-11

^=

(A)

for

</<;

/J

f^ cosa>/
2a

= C, r, where G = ie ai 2 B sin ar and r < R

= Cx r, where C2 = ifi e co 2 K sin to/ and r<R

(a)
(rf)

70-2

^ ^ /?, y = for r > R

B = ^I^r for />/{, 5 = poIr/lvR 2
;

(c)

y^f^cos^,.
"

CHAPTER
10-1

cot for

(a)
(rf)

10

v = 173.2 Mm s->;
= 165.4 mWrn"
5
(6)
(c)
Z = 217.7 0; () ffml =27.6mAmy = 150MmsA = 1.5m; (c) Z = 377H;
(6)
= 43.4Vm(e)
77 = 115.1 mA m"
S, =47.7mWm= 319 pirn"
(6)
2

pe . k

5, v

= 82 7 mWm" 2

10-4

(a)
(c)

10-7
70-*

m/v

1.77x

h>

10-"kgm-

= 77.7 mm
5 = 1.4kWmP = 3.96 x 10" W; (c) = 726.5 Vm"
(6)
W) f = 8.33min
(a)
5 = 10pWm- 2 (6) = 61.4 ^Vm"
7> = 2.83TW
(c)
(a)
/> = 4.24X10"W;
/>, = 1.93 x 10" W;
(6)
/ = 1.35 x 10" years
(c)
(a) Z/Zo =0.111;
A/Ao =0.111; (c) /p =0.111; (d) ^ = 90
(6)
(a)
VSWR = 1.5; (b) VSWR = 3.0; (c) VSWR = 3.6;
(d) when = 10mVm(a)
=245^Vm"
ff=650nAm-'; (c) 5 = 159 pW m" 2
(6)
(rf)
w = 2.66 x 10- 19 Jm- J {e) W = 7.25 X 10" 2 J
t

(a)

70-70

70-/J
70-75
70-7*

70-79

70-29

(a)

70-5/

(a)

8(1

10-33

(a)

{d)

/0-5*

(a)

conductor;

(6)

= 4.59 m,

dielectric

= 21.1m; (6) 8(1 /e) = 25.2 mm,

%) = 1 1 5.8 mm
(c)
8(1/*) = 650 pm, 8(1 %) = 2.99 mm
Z/Z =0.80-70.19; (b) A/A =0.033; (c) r/o =0.033;
S(l/e)=4.86mm; (e) a = 1.03Np; (/) p =0.153Z -130
8(l/e)=212mm; (b) 8(1 %) = 977 mm
S(l/e )

8(1%)

ANSWERS TO STARRED PROBLEMS

CHAP. 13]

CHAPTER

11

AR = 1.5;
AR = 6.0;
P = 14.9 kW

11-1

(a)

(A)

t=0'

(c)

clockwise;

11-3

(a)

(A)

t=0

(c)

counterclockwise;

11-6

809

= (a)0%;

11-11

a*

11-13

P = 500 mW

11-19

(1)

AR =

100%;

(b)

1,

LCP;

AR = 1,

(2)

0.50
0.25
0.20

0.50
0.50
0.50

50%;

(c)

left-handed
(d)

Antenna
2

11-8

Wave

(a")

70.7%;

(d)

RCP;

right-handed

0.50
0.35

(e)

AR
t= 90;

100%

oo, horizontally polarized,

(3)

(5) AR = oo, linear, x =45;

(4) AR=
=
=
LEP,
t = 22.5; (8) unpolarized;
=oo,
135;
AR
3.13,
t
linear,
AR
(7)
(6)
= i, AR = 1, RCP; (10) d = i, AR = oo, linear, x = 0; (11) d=h
(9)
AR = oo, linear, t=45; (12) a = 0.47, AR = 2.41, LEP, t = 90. (Note:
LCP = left circularly polarized, RCP = right circularly polarized, LEP = left ellip-

t=0;

oo, vertically polarized,

rf

tically polarized)

CHAPTER

12

12-1

(a)

12-2

(a)

tf 2 /#,

12.96

m;

12-3

u=2.65Um-'

12-4

(a)

= 903/xVm"

(c)

tf=12.5nAm"

12-5

12-6

= 27.5 mm
_ cos 0i
(a)

1
;

= 1.97 m;

x=263mm

(c)

(6)

0.0026% error

(b)

E= 1.05 /*V m"

(c)

= 2.10/*Vm-

Tii

cos

(6)

2
in 0A
+ V8 sin
2
8cos0, + V8-sini*0T7
\
2
cos 0, - V8 - sin 0,

Tx=l +

CHAPTER

0,

-8 cos

0i

cos

13-1

(b)

= 1.99995;

0,+

VS- sin

'

(a )

0, B

13

Z = 15.4 + j 26.9 Q; (6) VSWR = 4.26;

= 1.568 Z.11.3
Z = 32.3-y6.8 0; (b) v = 31.3 Mm s
(a)
(a) Zo = 337ii;
(6) /=60MHz
d/c = 0.61
(A)
(a) Z =45.9n;
d = 0.342A, Z, = 229 O
= 0.195A, d2 = 0.150A

()

= 7O.5

0,

(c)

= 0.62/29.7,

ii-5
13-5

13-6
13-8
ii-9

rfi

(c)

18.6%

>

810

13-10
13-11
7.J-75

ELECTROMAGNETICS

[APP.

= 0.75A, d = 0.25A, R = 245.0 fi, R 2 = 173.2 O

r= -0.715; (6) = 0.244 Z 76; (c) no
Kmax = Vmin = 500 mV (2) f, = 500 mV, Vmla = 250 mV
(1)
d = 0.226A, 2 = 0.104A
(a)
/ = 54.8 mm;
{b)
, = 1.87
di

(a)

13-18
13-19
13-25

rf

{a)

TEM

not passed, TE,

4.19,

TE 21

TE,,

TE 10

1.87 GHz, TE 20 3.75, TEi 3.75, TE 02 7.50,

TE l2 7.73 GHz; (b) v/v = 1.342 for all modes passed
GHz, TE 20 6.0, TE It 4.8, TE 21 7.1, TE 3.7, TE 02 7.5, TM 4.8,
5.30,

13-26

(a)

13-28

TM 21 7.1 GHz; (b) vlv = .237 for all modes

(a)
TM i2.3GHz, TM 02 5.3, TM,, 3.7, TM 12 6.7, TE

3.0

TEn
13-29

1.8,

TE 12

GHz;
GHz, TE 21
5.1

TE 1.17
TE 02 4.46, TE 12

(a)

TM-u
13-30

fc = 0,

13-33

3.39,

TM 02

4.82,

1.97,

TE 22

3.51,

v/v

(b)

TE 31

4.27,

1.565 for

TE 41

2.67,

TM 01

1.52,

3.37,

TM

TM 12 4.46 GHz; (b)

13-52

13-34

7>=10.9nW

(a)

f= 2.65 GHz,

13-60

(a)

d=

CHAPTER
(a)
(e)

1.1

mm,

(b)
(6)

4.05,

113

mm,

(c)

12,413

Q= 5,178

D = l.5; (b) G=1.05; (c) A = 0.083 A

= 3.5Q; (/) R T = 5.0n

R = 790(//A) 2
(o)

14-9

(a)

{d)

C1 A

= 8.38 sr;

fi

(e)

= cos(tt cos 0), D = 2.0 (6) = sin(7r cos 0), D = 2.0

e. B = 81.1%;
Z) = 1,304 (31.2 dB)
(6)
HPBW = 8.4; (b) -13.0 dB; (c) H^ = 1.15 sr; (rf)
7) = 10.8 (10.3 dB);
(/) /f = 0.86A
HPBW = 53.6; (b) -13.2 dB; (c) O^=0.87sr; (rf)
2
7) = 14.4 (11.6 dB);
(/) A, = 1.14A

/\
(a)

E(6)=1

(a)
(e)

74-26

R,

14-2

14-23

modes

14

74-3

7^i/i
74-20

all

2.45,

TM

(a)

14-15

TE 01

3.28,

13-54

74-74

6.7,

13-59

14-1

4.01,

TM 21

1.809 for

a = 8.89/Ao Nprn"
=3m
S=1.33nWm- 2
TEM: a = 5.09 mNp m" TE 10 a = 12.79 mNp m-'
a = 8.42/ dBm"', where r = radius

13-51

TE 51

2.45,

v/v

TE 02

i3.7,

modes

all

(a)

i-/fl\

'-

2w cos

^/.(l.OS sin

(a)

7,

7f,=691 n,7J 2
E(0)

= cos

S '"

4 cos 2

(6)

(a)

74-52

(a)

0.01

^9

where

(ft)

0), 2 =/,(2.36 sin 0),

= 1,836 n,7? 3 = 2,573

sin(yr/2)

74-27

0.01

sin(207r cos 6)

(6)
(a)

= 91%;

7f 3

0;

^2 = 90

<r

= 83%;

HPBW=4.2
=/,(6.28
(c)

(0.2 cos

7),

sin 0);

= 1.5, D 2 = 1.2,

7>,

= 3.6

- 1.217)

HPBW = 23.2; (c) D = 98.7 (19.9 dB); (<7) REP, AR = .02

linear, t= 45;
(b)
CP; (c) elliptically polarized, AR = 2.41, t =45
Z = -6.4+yl.8n (b) 2V =2.5 +/1.8 fi
1

ANSWERS TO STARRED PROBLEMS

CHAP. 15]

14-37

gain

(a)

=2 /__Rn_
V Rll "^12
1

14-46
14-47

sin

*ii*o

811

2ns

Z = 778-/86fi
/ = 80A, w = 12.7A, e E = 2.86, 0 = 4.52;
()
HPBW = 7.0, HPBW = 5.3; (c) Z> = 762 (28.8 dB);
(6)
e p = 59.7%
(</)
A = 121.6 m 2 (RCP)
Arml =0.015 K; (6) ASml = 518fu; (c) A5m ,=259fu
(a)
= 320Mpc; (c) /? = 640Mpc;
/? = 240Mpc;
(6)
(o)
.

14-48

14-54
14-57

W Hz"

x 10"

PJAf= 3.99

(d)

Pb IAf= 2.36

1
,

Hz"
x 10"
2
Hz"
>v = 2.08x 10-" J m(a)
13
years
t = 5.2x 10
(c)
,,=() 0.0006%; (6) 5.5%; (c)

Pc /Af=

1.42

14-58

14-60

/> 0.01 A

(/)

x 10"

W Hz

1
,

for better than

90%

W = 4.68 x

(6)

99.8%;
efficiency;

10 17 J;

100%;

(</)

(e)

100%;

polarization mismatch,

(#)

insufficient gain

14-63

= 7,162 m 2

()

,4,

id)

>10"*%;

(b)

/>,

no;

(e)

= 5.7W;

(/)

(c)

parts (a)

and

P,.
(6)

= 8.3 /xW;

by a few dB but (d) and

(e)

not

significantly

14-65

53.5

CHAPTER
15-1

(a)

15-2

(a)

15-3

()

15-5

(a)

15-7

(a)

15

= Mm s"
R = 500 mm
(b)
K = 1.4kV; (b) fl = 2.82mT
= 6.3km; (c) dBjdl = 177.7 T s"
= 6.3 mm /as(6)
Jf = 7MeV;
(c)
W=14MeV;
(b)
W=14MeV;
= 388 r, p = 777 r, = 388r
=20.8m; (6) /? = 1.04 mm; (c) R = 10.4 Mm
1

i>

15-13

A/= 3.5 kHz

15-14

(a)

15-16

(a)

MUF = 15.6 MHz;

/ =9MHz;

where
15-20

rf

<ma. is

the

(6)

= 10.4 MHz;

(6)

MUF =26.1

maximum

(c)

MHz;

(c)

MUF =

MUF = 12.7 THz

/<>(<

= 0)

cos<

pointing angle from the zenith

2,603.6A
(b)
fir

=
= 0.972, 03r)-/(j3/-) = 0.982, (0r)/<0r) = 0.952;
for horizontal travel, A0 = 0, for vertical incidence,
(e)
if w > <o, but a. < a>o
(rf)
A0 = 380 turns
= 89.2 mm s"
(c)
= 824.7 km s"
v= 109.1 Mms"
(6)
(a)
= 7.54 ms"
id)

(a)

(/3r)

(c)

(j8r)

=2,659.8A

(/&)- =2,687.8A;

+ /03r)

15-21

t>

15-27

1J

= 3kV

t;

t>

8 12

ELECTROMAGNETICS
[

CHAPTER
16-10

(a)

16

El = 10.0 Vm-\BL = 1.1 fT; (b) EL = 10.0 Vm" 1 BL = 1.1 pT;

1 = 10.0 V m" 1 51 = 1.1 nT; id) EL = 10.6 Vm"', BL = 11.7 nT;
1 = oo, 5^ = oo
1 = 333.3 nVm-,B{ = 33.3 nT; (6) EL = 333.3 /uV m" 1 BL = 33.3 nT;
1 = 333.3 mVm- 1 51 = 33.3 nT; {d) EL = 3.5 Vm" 1 BL = 35.3 nT;
1 = oo Bi = oo
l = 10Vm-', 51 = 33.3 nT; (b) EL = 10 Vm- 1 BL = 33.3 nT;
l=9.7V m - 1 51 = 32.2 nT; (rf) 1 = 7.1 Vm" 1 51 = 23.5 nT;
EL = oo, 51 = oo
,

(c)

(e)

/<W2

(a)

(c)
(e)

16-14

(a)

(c)
(e)

Ap p c

/5-iS

^ = 3.33mgm-

7(5-25

(a)

16-29

18 years;

= 0.94c

(b)

6 years;

(c)

3.6 years;

(rf)

6.5 years

INDEX

helical-beam, 656-662
(See also Helical-beam antennas)
helicone, 692-693
horn, 688-689

Abramowitz, M., 466n.

Absolute potential, 22
Absorption of wave, 451

Ac

generator, 315
Aerial (see Antennas)
Alford, A, 441n.
Alfven, H., 740n.

pyramidal, 689
interferometer (see Interferometer)
isotropic,
linear,

AM broadcasting station, 707

molecule,

646-648

loop, 654-655

ALSEP experiment, 660, 662

Alternator, 419
AM, 480 (Prob. 12-10)
Ammonia

623

lens (see Lens antenna)

Alfven velocity, 742

Alfven waves, 740
Algorithm, 273

for,

mutual impedance

(Prob. 14-7)

599 (Prob. 13-58)

Ampere, definition, 3, 147

Ampere-Maxwell vector, 767
Ampere's law, 7
and H, 163-165
Maxwell's equation from, 166

of,

reflector,

678-687

Analog computer, 276-278

Angle of refraction, 453n.

corner, 682-684

Anisotropic materials, 57
Anisotropic medium, 729

infinite-flat-sheet,

Anode, 289
Anomalously dispersive medium, 376
Antarctica, 424 (Prob. 10-36)
Antenna array (see Arrays)
Antenna efficiency, 619
Antenna fields, computer-calculated, 648-649
Antenna impedance, self- and mutual, 670-675
Antenna pattern, 602
Antenna reactance, 625
Antenna resistance, 625
Antenna temperature (see Temperature, noise)
Antenna voltage response, 435
Antennas:

691-692
complementary, 687-688
current distribution on, 646
dipole, 602
directivity of, 622
effective aperture of, 625
Fourier transform relation for, 644-646
frequency-independent, 665-667
active region, 666
equiangular spiral, 666
log-periodic, 666-667
gain of, 623
half-wave dipole (see Half-wave dipole)
circularly-polarized,

673-675

parabolic dish, 642

phase center of, 648
pocket ruler, 665-666
polarization state, 435
prolate spheroid, 648-649
radio telescope, 700
(See alsc Radio telescope)
receiving, 601, 624

679
679
nonspherical, 470
parabolic, 684-685

finite-flat-sheet,

spherical, 711 (Prob. 14-44)

thin,

680-682

resonance of, 648

retroreflector, 713 (Prob. 14-52)
self-impedance of, 670-673
short dipole (*ee Short dipole)
shortwave, 599 (Prob. 13-62)
slot,

687-688

small helical, 655-656

stored energy of, 602
supergain, 659
transmitting, 601, 624
traveling wave, 650-654
V, 602, 672
Yagi-Uda, 682
Antipodal state, 435
Antisymmetric matrix (see Skew-symmetric

matrix)
Aperture, 624
effective,

690

physical, 690
scattering, 698

Aperture concept, 690-694

Aperture distribution, 646
continuous, 642-644

814

INDEX

Aperture efficiency, 690

of large circular aperture, 693-694
Apollo 14 astronauts, 660
Appearance of field, 773-775

Applied field, 46
Approximation formulas, 796
Array factor, 629
Array gain, 677
Array theory, 627-642
Arrays:
beam width of, 662
binomial, 629-631
directivity of, 662
collinear, 708 (Prob. 14-13)
continuous, 642
grid structure, 663-665
of helical beam antennas, 661
with n sources, 631-636
broadside case, 633
end-fire case, 633-635
null directions of, 632
phase center of, 629
scanning, 663-665
of six-point sources, 638

two

isotropic point sources as, 627-629

uniform, 631
van Atta, 713 (Prob. 14-53)
Artificial dielectrics,

63-66

Beam solid angle, 620-621

Beam width between first nulls (BWFN), 620
Bengtsson, N. E., 346 (Prob. 8-34)
Bessel function, 284, 554
recurrence relations for, 798

root of, 554

Betatron, 345 (Prob. 8-28)
Beverage antenna, 577

Binomial array, 629-631

Binomial series, 514n., 631
Biot-Savart law, 144
Blackbody, 422 (Prob. 10-11)
Bonn, E. V., 568n.
Boltzmann's constant, 422 (Prob. 10-11), 700
Booker, H. G., 644n 687n.
Born, M., 432n., 441n.
Bound charge (see Charge polarization)
Bound wave (see Surface wave)

Boundary

relations:
at conducting surface, 50-51
at conductor-conductor boundary, 130
at conductor-insulator boundary, 127-128
electric field, 67-69

magnetic

field,

216-221

tables of, 71, 221

for time-varying fields, 336-337

Boundary-value problems, 262

Bozorth, R. M., 801

Associated Legendre function, 284

Astigmatic tube, 470
Astronauts, Apollo 14, 660
Atomic current loop, 198

Breakdown

Atomic

Broadband matching, 515

dipole:

333
magnetic, 199
Attenuation, 546
electric, 59,

at frequencies above cutoff, 563

at frequencies below cutoff, 561
Attenuation constant, 488, 564, 566

(see Dielectric strength)

Brewster angle, 458, 593 (Prob. 13-20)

Brightness, 422 (Prob. 10-11)
Brillouin,

Leon, 375

Broadside array, 633

Brown, G. H., 680n.
Bushing, capacitor, 106 (Prob. 3-34)
BWFN, 620

for cylindrical guide, 568-569

for infinite-parallel-plane line, 566-567
Attenuation factor, 404, 488
Automobiles, detection of, 347 (Prob. 8-35)
Auxiliary equation, 487
Axial ratio (AR), 427, 428

Candela, definition, 3
Capacitance, 49
of coaxial line, 81
of earth, 55 (Probs. 2-36, 2-39)
of isolated metal sphere, 55 (Prob. 2-35)
of stripline, 296 (Prob. 7-1)
of two concentric shells, 55 (Prob. 2-41)

Backlobe, antenna, 604

Balanced transmission line, 606
Balmain, K. D., 801
Bandwidth, 376, 436, 513
of helical beam antenna, 659
predetection, 701
of transformer, 511-515

Capacitor, 49

units of, 49

Barium

titanate material,

452

C, 441
Barrow, W. L, 532n., 551n.
Base units, 2
Battery, energy of, 105 (Prob. 3-18),
Barkofsky, E.

108 (Prob. 3-50)

Baytron Co., Inc., figure, 595
Beam of ions, 719
Beam direction, 664

Beam

efficiency, 622n.
of large circular aperture, 693-694

analogy to storm cloud, 105 (Prob. 3-21)

dielectric strength of, 73

divergence of
energy in, 74
energy density
field cell, 89

and

in,

in,

99

75

parallel-plate, 49, 61-62, 71-73,

99-100

(See also Parallel-plate capacitor)

two conductor line as, 85
Capacitor bushing, 106 (Prob. 3-34)
Cardioid-pattern antenna, 680
Carrier modulated wave, 377
Carter, P.

S.,

674

Carver, Keith R., 663n., 692., 694n.

Cathode, 289
Cathode-ray oscilloscope, 303 (Prob. 7-39)

INDEX

Cathode-ray tube (CRT), 721-725

electrostatic deflection in, 722
magnetic deflection in, 723
Caustics, 471
Cavity resonator, 583

Collector, 289
Collin,
E., 802
Collisions, effect of, 735

with gas molecules, 747 (Prob. 15-33)

Communication
between earth and space vehicle, 744 (Prob.

Cell:

conductor, 131

15-13)

90

field, 89,

Center-fed dipole antenna, 646

Cesium atom, Q for, 599 (Prob. 13-58)
Characteristic impedance, 382, 488, 573
of coaxial line, 493-495
determination by dc measurement, 493
of two-wire line, 495
Characteristic resistance, 489

Charge:
induced, 46
point, 13
polarization, 99, 1 10
single shell of, 43
density:

linear,

30

polarization, 60-62
surface, 30

volume, 29
particle
in static electric field, 716-719
in static magnetic field, 719-721

Chart:
rectangular impedance, 502
Smith, 505

SPEMP,

11
transmission-line, 502

Chawla, B. R., 303 (Prob. 7-40)

Child, D. C, 292
Child-Langmuir law, 292
Chimney flue, 743 (Prob. 15-9)
Chu, L. J., 532n.
Circularly polarized antenna, 691-692
Circularly polarized wave, 427
left-handed, 429
right-handed, 429
Circulator, three-port, 591 (Prob. 13-11)
P.

Compass needle, 140

Complementary antennas, 687-688
Complementary solution, 288
Complete solution, 282
Completely polarized wave, 437
Completely unpolarized wave, 437

Components, microwave, 594-595 (Prob. 13-36)

Computer:
analog, 276-278
digital: desk-top, 276
programmable, 276
vest-pocket, 276
Computer calculation of antenna pattern, 715
(Prob. 14-64)

Charged

Clemmow,

on moon, 596 (Prob. 13-38)

Communications service, 594 (Prob. 13-32)

Complex conjugate, 391

Complex power, 619
Complex vector, 430

electric, 13

Charge

815

C,

644n.

Closed rectangular pipe, 485

Coaxial line, 284-288
with asymmetrically located inner conductor,
195 (Prob. 5-24)

Computer methods:
for field of charges, 107 (Prob. 3-38)
relation to field mapping, 92

Computer program:
for electric charge distribution, 53-54 (Probs.
2-30, 2-31)
Fortran IV, 273, 275
Condenser (see Capacitor)

Conductance, 116
Conducting bar with notch, figure, 134
Conducting block, rectangular, 301 (Prob.
7-20)

Conducting
Conducting
Conducting
Conducting
Conducting
Conducting
Conducting

fluid,

740

half-plane, 465

hemisphere, 302 (Prob. 7-29)

media, 399
paint,

277

rod, 349
sheet:
with slot, 302 (Probs. 7-28, 7-34)
between two conducting planes, 278-283
Conducting wheel, 342 (Prob. 8-7)
Conduction current, 328

capacitance of, 81

Conduction-current density, 400

characteristic impedance of, 495

cylindrical, 297 (Prob. 7-2)

Conductivity, 116

elliptical,

297 (Prob. 7-3)

172-173

field cells,

128
inductance of, 161
square, 304 (Prob. 7-41)
Coercive force, 229
fields in,

Coercivity, 230

Coherence, 694-696
polarization, 695
wavefront, 695

Coherency matrices, 441, 443

Coherent light, 467
Coil (see Inductor)

equivalent, 334
table of, 117
tensor, 733n.

Conductor, 45, 57, 401

electron motion in, 1 10
over ground plane, 295
with right-angle bend, 303 (Prob. 7-40)
Conductor cell, 131
Conservation of charge relation, 126n.
Conservation of energy, principle of, 389
Constant-phase point, 368
Constants, fundamental, 794
Constitutive relations, 358
Continuity relation, 126, 357, 767

816

INDEX

Continuous aperture distribution, 642-644

field of, 644
Continuous array, 642
Contour map, 604
Contraction factor, 755
Convergence, 283
Conversion factors, table of, 794
Cool plasma, 729
Coordinate diagrams, 799
Coordinate transformation, 757-763
Galilean, 761
Lorentz, 761

Corona, 73n., 734

Cosine integrals, 648
Cosmic ray, 747 (Prob. 15-35)
Cosmic-ray telescope, 695
Coulomb, Charles A. de, 13.
Coulomb's law, 13
Coulomb's Law Committee, 59n.
Counter emf, 324
Coupling factor, wave-to-antenna, 439
Cowan, John D., Jr., 648
Crawling caterpillar, 377
Creeping-wave diffraction, 474
Critical angle, 456
Critical frequency, 732, 744 (Prob. 15-14)
Cross field, 441
Cross product, 152
Cross section, radar, 698
Crossfield ellipse, 577
CRT (see Cathode-ray tube)
CRT display, 519
Crystals, 58n.

Curie point, 222

Curl, 175-179
compared with divergence, 184
examples of, 179-182
Curl meter, 180-181
Current, 109-110
continuity relation for, 126
density of, 118
induced, 306
mapping of, 131-134
retarded, 605, 647
solenoidal, 124
Current density, 118
Current distribution on antenna, 646
Current mapping, 131-134
Current moment, 142
Current reflection coefficient, 497
Current sheet, 338
Curvilinear square, 88, 132
power through, 522
Cury, R, 801
Cutoff frequency, 544
for cylindrical guide, 558
for rectangular guide, 545
Cutoff wavelength, 526
for cylindrical guide, 558
for rectangular guide, 545
doppler radar, 713 (Prob. 14-51)
Cyclotron, 725-727
Cyclotron frequency, 732
Cyclotron magnet, 260 (Prob. 6-29)

CW

Cygnus A, 714 (Prob. 14-57)

Cylindrical conductor, charged, 293

Cylindrical guide, low-loss mode in, 556
(See also Waveguide, cylindrical)

D'Alembert's equation, 366

Damped harmonic
Damping

oscillation,

coefficient,

333

332

dBi, 624

Debye,

P., 332, 582n.

Decibel, relation to neper, 562n.

Dees, 725
Deflection distance, 723, 724
Degree of polarization, 438
Del, operations involving, 183
Del operator, 33
Demagnetization, 236-237

Demagnetization curve, 234

de Moivre's theorem, 281, 611
Deperming, 237
Depth of penetration, 406
Derived units, 3
Deschamps, G. A., 431
Deuteron, 726
Diamagnetic materials, 199
Dielectric, 57, 110, 401
artificial, 63-66
Class A, 57n.
lossy, 334
nonmagnetic, 456
Dielectric coated conductor guide, 580
Dielectric constant, 13., 57n.
Dielectric hysteresis, 332, 599 (Prob. 13-61)
Dielectric rod line, 483
Dielectric strength, 73
Differential permeability (see Infinitesimal permeability)
Diffracted field, total, 476
Diffracted ray, 465
Diffraction of plane wave, 466
Diffraction coefficient, 475
Diffraction curve, half-plane, 467
Digital computer solution to Laplace's equation,

273-276
computer techniques compared with
manual techniques, 276
Digital voltmeter (DVM), 277
Dimensional analysis, 5-6
Dimensions, 2, 786
Diode, 289
Diode current, 292
Digital

Dipole:
atomic, 59
electric (see Electric dipole)

magnetic, 201
short (see Short dipole)
Dipole antenna, 602
center-fed,

646

Dipole moment, 60
Directivity, 622, 627
approximate formula

for,

635

of isotropic antenna, 623

relation to gain, 624
of short dipole, 623
Dish antenna, parabolic,

642

INDEX

of charge, 80
of charge, 79
moving charge, 781 (Prob. 16-6)
octopole, 53 (Prob. 2-16)
quadrupole, 52-53 (Prob. 2-13)

an
an

infinite cylinder

Distortionless line, 489

of
of
of
of
of

Distributed quantities, 48S

Divergence, 94
compared with curl, 184
of
and P in a capacitor, 99-101

relation to electric flux density, 40-41 , 62

relation to electric potential, 33-36
relation to polarization, 62
of several point charges, 18-20

Dispersive medium, 376

waveguide as, 529/1.

Displacement current, 328
Displacement-current density, 330, 400

D
H

example

infinite line

of two equal point charges, 27-29

opposite sign, 27

97-98

of,

817

in toroid, 214
of
of J, 125-126
of tensor, 776.
Divergence operator, 96
Divergence theorem, 98
Dixon, Robert S., 273
Domain, magnetic, 221

same

sign,

28

Electric field strength (see Electric field

intensity)
Electric flux, 39-43
over closed surface, 41-43

tubes of, 39

Dominant mode, 549n.

figure,

for cylindrical guide, 559

Doppler shift, 714 (Prob. 14-57)
Double-stub tuner, 506, 590 (Prob. 13-9), 592
(Prob. 13-17)
Driving force, 332
Drude-Debye relations, 346 (Prob. 8-32)
Dyadic Green's function, 532n.

40

Electric flux density, 39-41, 62

boundary relations for, 51, 68
definition,

40

divergence in capacitor, 99-101

relation to electric field intensity, 40-41, 62
Electric potential, 20-22
(See also Potential)
Electric scalar potential (see Electric potential)

Electric susceptibility, 63

Earth, magnetic field of, 261 (Prob. 6-36)

Earth-Jupiter radio link, 713 (Prob. 14-50)
Earth-Mars radio link, 713 (Prob. 14-50)
Earth-moon radio link, 712 (Prob. 14-48)
Earth resources satellite, 714 (Prob. 14-55)
Easy magnetization, 222, 226
Echo signals (see Ghost signals)
Ecology, 593 (Prob. 13-22)
Eddy currents, 315, 327

Edge

diffraction,

474

Effective aperture, 625,

Efficiency:

690

Electromagnet, 260-261 (Prob. 6-30)

Electromagnetic interference, 593 (Prob. 13-22)
Electromagnetic theory in concise form, 769
Electromechanical analogy, 332
Electromotance, 120n.
Electromotive force (emf), 120
Electron, 116
Electron cloud, 332
Electron current, 289
Electron density, 733
Electrostatic precipitator, 743 (Prob. 15-9)
Electrostatic press, 104 (Prob. 3-15?)

antenna, 619

Element pattern (see Primary pattern)

aperture, 690

Elliott, R. S.,

beam,

Ellipse, 686
crossfield,

622/1.

Eigenvalues, 556, 559

Einstein, Albert, 749
Einstein's energy relation, 784 (Prob. 16-30)
Einstein's relation, 422 (Prob. 10-13)
Einstein's summation convention, 769
Electric and magnetic field relations compared,

340
Electric circuit, 113
Electric dipole, 36
electric field of, 38
figure, 27,

37

potential of, 37
table for, 38
Electric-dipole moment, 36, 60
Electric field intensity:
boundary relations for, 67
at center of circular ring, 105 (Prob. 3-23)
at center of square ring, 105 (Prob. 3-24)
of charged dielectric sphere, 108 (Prob. 3-49)

of charged spherical shell, 43-44

due to charges, 307
of dipole, 38
of a finite line of charge, 78-79

801

577

polarization, 43
Elliptical

conducting loop, magnetic flux density

of, 193 (Prob. 5-16)

Elliptically polarized wave,

427

Emde, F., 466n.

Emf, 306
total induced, 312
Emf-producing field, 119, 307
Emitter, 289
Endfire array, 633-635

Energy:
in capacitor,

74

conservation of, 389

in inductor, 173-174
in

magnet, 232-234

per pulse, 395

stored, 398

Energy density, 388

average, 394
electric, 74-76
instantaneous, 394
magnetic, 174

818

INDEX

peak, 394
total, 389
Energy product, 235

Filter, low-pass,

Energy

relations:
in standing wave, 396
in traveling wave, 388
Energy velocity, 392, 529
Equality of path lengths, principle of, 684

Equation numbering,

407

Flow

Equator-to-great-circle angle, 432

Equiangular spiral antenna, 666

Equipotential contour, 25
Equipotential line, 22
Equipotential surface, 29, 294
Equipotentials, magnetic, 170

Equivalence of magnetized rod and solenoid,

203-205, 215
Equivalent conductivity, 334
Equivalent surface charge distribution, 292
Equivalent units, 786
Exponential line, 591 (Prob. 13-12)
Exponential taper, 515
Extraordinary ray, 735

chart, 273
Fluid, incompressible, 741
Fluorescent screen, 721

Flux:
electric (see Electric flux)

magnetic (see Magnetic flux)

Flux-cutting law (see Motional-induction law)
Flux density:
electric, 39-41, 62
magnetic, 142, 149
of radio source, 700
Flux leakage, 324
Flux linkage, magnetic, 159
Flux tubes, electric, 39-40, 87
Flux unit (fu), 700n.
Fluxmeter, 224

FM, 480

(Prob. 12-10)

Force:
aligning, 140

per charge, 16
coercive, 229
current-carrying conductor, 152
driving, 332

on

electric,

14

Fading, polarization, 659n.

gravitational, 15

due to Faraday rotation, 744 (Prob. 15-13)

Far field, 613
Far-field pattern, 604
Farad (unit), 49

magnetic, 140-142
magnetic gap, 254
on parallel-plates of capacitor, 104 (Prob.

Faraday, Michael, 305, 307

Faraday disk generator, 314, 342 (Prob. 8-11)

Faraday-Maxwell vector, 767
Faraday rotation, 659n., 737-739
position angle, relation to, 739
for quasi-longitudinal case, 738
for quasi-transverse case, 739
total, 739
Faraday's ice pail experiment, 46n.
Faraday's law, 307
Federal Communications Commission, 707
(Prob. 14-7)
Fernald, D. L., 596 (Prob. 13-38)
Ferrite material, 452
Ferrite-titanate medium, 424 (Prob. 10-33)
Ferromagnetic materials, 199-200
Ferrous-dielectric material, 479 (Prob. 12-5)

Field cell, 89
capacitor, 89, 90
magnetic, 243
transmission line, 171
Field components, transformed, 772
Field distributions, 76-77
Field intensity:
electric (see Electric field intensity)
magnetic (see Magnetic field)
Field lines, 87n.
Field mapping, 86-93, 242-250
Field maps:

comparison, 247-248

3-6)

between two parallel conductors, 146-147

per unit volume, 776
Force relations, 357
Fortran IV program, 273, 275
Forward tilt of wave, 574
Four-conductor transmission line, 301 (Prob.
7-25)

Four-dimensional divergence, 768

Four-dimensional field formulation, 765
Fourier sine expansion, 282
Fourier transform method, 708 (Prob. 14-18)
Fourier transform relations, 644-646
Fra Mauro highlands of moon, 660
Frequency-independent antennas, 665-667
Freshwater lake, 480 (Prob. 12-9)
Fresnel cosine integral, 466
Fresnel diffraction pattern, 481 (Prob. 12-14)
Fresnel integrals, 466
Fresnel reflection coefficients, 460
Fresnel sine integral, 466
Friis transmission formula, 698
Fringing field, magnetic, 141
Frozen magnetic field, 741n.
Fundamental dimensions, 2, 786

Fundamental

units,

2-4

G-string guide, 582

Gabor, D., 467.

Gain:

properties of, 91-92

Field pattern, 613
Field relations, general, 338-340
Fields of moving system, 769-775

array, 677
relation to directivity, 624
Gallilean transformation, 761

Filamentary conductor, 295

Gapless

antenna, 623

circuit,

250-252

INDEX

Gauss, Karl Friedrich, 43

Gauss' law, 43-44, 47
Gauss' theorem (see Divergence theorem)
Geometric mean, 509
Geometrical optics, principle of, 468
Geometrical-optics approximation, 468
Geometrical optics concepts, 468
Geometrical theory of diffraction (GTD), 474,

680

Germanium semiconductor,

139 (Probs. 4-24,

4-25)

Ghost

signals, 481 (Prob. 12-10)

Glass envelope, 289
Gordon, W. E., 744 (Prob. 15-12)
Goubau, G., 582n.
Gradient, 33-36

ionization,

744 (Prob. 15-44)

34-36
Mapping,

in rectangular coordinates,
Graphical field mapping (see

graphical)
Gravitational attraction, 55 (Prob. 2-34)
Gravitational field of earth, 55 (Prob. 2-33)
Great-circle angle on Poincare sphere, 432
Greenland, 424 (Prob. 10-36)
Green's function:
dyadic, 532n.

Helical coil (see Solenoid)

Helicone antenna, 692-693
Helmholtz formula, 312
Helmholtz pair of coils, 195 (Prob. 5-23)
Henry, Joseph, 305

Henry (unit), 159

Hertz, H., 357
High-resistance voltmeter, 277
Higher-order mode, 482, 600 (Prob. 13-66)
wavelength of, 529
Holes in semiconductors, 116
Hologram, 467
Holography, 467
Homogeneous equation, 288
Homogeneous medium, 56
Hondros, D., 582/1.
Horn antenna, 642, 688-689
HPBW, 620
Hubble constant, 714 (Prob. 14-57)
Hubble relation, 714 (Prob. 14-57)
Huygens, C., 464n.
Huygen's principle, 464, 642
Hydraulic analogue, 122
Hyperbola, 685
Hyperbolic relations, 795
Hyperbolic wire, magnetic flux density
(Prob. 5-15)

electrostatic, 32
free-space, 608/1.

Hyperfine transition, 599 (Prob. 13-58)

Grid structure array, 663-665

Group velocity, 376
Guide (see Waveguide)

Gyro frequency, 731

Gyro resonance, 731

Hysteresis, 229-231
Hysteresis loop, 230, 325

major, 230
minor, 326

Ice:

Half-cylindrical metal tube, 303 (Prob. 7-38)

Half-cylindrical trough transmission line, 300
(Prob. 7-10)
Half-plane, conducting, 465
Half-plane diffraction curve, 467

Half-power beamwidth
Half-wave dipole:
directivity of, 648
fields of, 648-650

(HPBW), 620

permittivity of, 346 (Prob. 8-32)

thickness of, 425 (Prob. 10-36)
Idaho potato, 346 (Prob. 8-34)

Idealness, 15

Image of conductor, 296

Image charges, 292
Image current, 296
Images, theory of, 85
Imaginary part of c, 334

radiation resistance of, 648

Hall effect, 197 (Prob. 5-34)
Hall voltage, 197 (Prob. 5-34)
Hallen, Erik, 671.
Halliday, D., 801
Hansen-Woodyard increased directivity condition, 658n.
Hard magnetization, 226

Impedance:

Harmonics, 283

Incoherent wave, 435

Incompressible fluid, 741
Increased directivity condition, HansenWoodyard, 658n.
Incremental permeability, 326
Index of refraction, 374
for conducting medium, 407
for plasma, 734

Harrington, R. F.. 580n., 801

Hayt, W. H, Jr., 801
Heaters, RF, 599 (Prob. 13-61)
Heaviside, Oliver, 389
Heaviside's condition, 489

Helical-beam antennas, 429, 656-662

array of, 661-662

bandwidth

of, 659
directivity of, 659
pattern of, 658

phase velocity on, 659

use in space communication, 659

819

antenna, 670-675
characteristic, 382, 488, 573
intrinsic (see Intrinsic impedance)

mutual, 323

670-673
wave, 574
self-,

Incident wave, 383

Induced
Induced
Induced
Induced
Induced

charges, 46
current, 306

emf, 310
field,

46

magnetization, 222

of,

193

S20

INDEX

Inductance, 159
average, 325
of coaxial line, 161
general case, 311
of solenoid, 160
of toroid, 161
of two-wire line, 162
Inductor, 158
coaxial line, 161
energy in, 173-174

'

Kock, W. E., 63n., 686n.

Kouyoumjian, R. G., 729n.
Kraus, John D., 63n., 441n., 465n., 603n.,

two-conductor

line,

162

RF

heating, 332
Infinite-parallel-plane line, 483
Infinite square trough, 269-272
Infinitesimal dipole (see Short dipole)
Infinitesimal permeability, 325
Initial magnetization curve, 226
Injection molding, 599 (Prob. 13-61)
Insulator, 110
Interference, electromagnetic, 593 (Prob. 13-22)
Interferometer:
as radio telescope antenna, 640

two-element, 640-642
Interferometer pattern, 642
International System of Units (SI) 2-3
Interplanetary vehicle, 783 (Prob. 16-21)
Intrinsic impedance, 379, 380, 382, 511, 573
of conducting medium, 409, 410
of copper, 412
,

Intrinsic polarization, 739

Intrinsic resistance, 379

Invariance in space-time, 763-765

Invariant axes, 765
Invariant hyperbola, 763
Invariant lines, 753
Inverse-square law, 13

Ion beam, 719

Ionization gradient, 744 (Prob. 15-14)
Ionized medium, plane waves in, 729
Ionosphere of earth, 55 (Prob. 2-36)

waveguide, 571-573
Iron ring, 210-213
Isaac, E. D., 801
Iris,

Isbell,

D.

E., 666/1.

Isotropic antenna, 623

Isotropic material, 56
Isotropic source (see Isotropic antenna)
Iterative method {see Point-by-point method)

Jahnke, E., 466n.

Jasik, H., 801
Jensen, H. J., 593 (Prob. 13-23)
Jordan,
C, 801
Joule heating, 327, 564
Joule's law, 112

R, 344 (Prob. 8-20)

Lamellar field, 167, 186
Laminations, 327
Langmuir, I., 292
Laithwaite, E.

Laplace's equation, 103, 262

for conducting media, 133-135
in cylindrical coordinates, 283
for parallel-plate capacitor, 264-266
in rectangular coordinates, 263
in spherical coordinates, 284
Laser beam, 593 (Prob. 13-20)
Latitude on Poincari sphere, 431

Launching device, 579

Law:
Ampere's, 7
Biot-Savart, 144
Child-Langmuir, 292
Coulomb's, 13
of edge diffraction, 475
Faraday's, 307
Lenz's, 306, 327
Lawrence, Ernest O., 725

Left circular polarization, 429

Lens antenna:
dielectric,

685-686

metal plate, 686

Lenz's law, 306, 327
Light lines, 751
Light pen, 303 (Prob. 7-39)
Line integral, 23

around closed path, 24

Linear antenna, 646-648
Linear charge density, 30
Linear medium, 56
Linear motor for vehicle, 344 (Prob. 8-20)
Linearly independent solution, 288
Linearly polarized wave, 426
Linkage, magnetic flux, 159
Load impedance, 496
Log-periodic antenna, 666-667
Logarithmic pattern, 604
Logarithmic relations, 796
Longitude on Poincarf sphere, 43
Longitudinal propagation, 735

Loop:
atomic current, 198
105 (Prob. 3-23)
magnetic field of, 147-148
torque on, 154-155
Loop antenna, small, 654-655
Lorentz, H. A., 761n.
Lorentz transformation, 761
static: electric field of,

Keeper, 237
Keller,

J. B.,

475n.

Kelvin, Lord, 2
Kelvin, definition, 3

Kennaugh, Edward M., 648

642/1.,

656n., 659/1., 663/1., 671n., 682n., 690n.,

699n., 704zt 729/1.

solenoidal, 159-160
toroidal, 161
Industrial

Kilogram, definition, 3
Kinetic energy, 290
King,
W. P., 801
KirchoS's current law, 124-125
KirchofFs voltage law, 120
Ko, H. C, 439?!.

INDEX

Magnetic poles, 175

Magnetic potential function (see Magnetostatic

Loss resistance, 619

Loss tangent, 334
Lossless line, 496
Low-pass filter, 407
Lunar communication, 596 (Prob. 13-38)
Lunar Fra Mauro highlands, 660

McCullough, T.

Magic

P.,

701

pencil (see Light pen)

Magnet:
bar, 201

force on, 257 (Prob. 6-4)

magnetic flux density of, 259 (Prob. 6-17)
rotating (see Pulsar)

domains in, 223

effect on current-carrying wire, 141-142
energy in, 232-234
iron ring, 261 (Prob. 6-33)
materials for, 236

lobe, 604
Main-lobe solid angle, 622

of,

701

Mass spectrograph, 728

206
field)

flux, 149,

193 (Prob. 5-16)

hyperbolic wire, 193 (Prob. 5-15)
infinite linear conductor, 145-146
iron ring, 210-212
parabolic wire, 193 (Prob. 5-14)
relation to magnetic field, 163-164, 206
relation to magnetization, 206
of solenoid, 155-157, 214-216
of toroid, 160, 208-209
of two-conductor line, 162
Magnetic Mux linkage, 159
of
of
of
of
of

Main

Marcuvitz, N., 570n.

of solid cylindrical conductor, 164-165

Magnetic
308
over closed surface, 150
tubes of, 150
Magnetic flux density:
of bar magnet, 259 (Prob. 6-17)
boundary relations for, 217
of coaxial line, 161
of current carrying loop, 147-148
single-turn, 259 (Prob. 6-18)
square, 193 (Prob. 5-13)
due to current distribution, 149
definition, 149

Magnetostatic potential, 167

Magnusson, P. C, 802

Mars, temperature
Mass, 332
relativistic, 718
rest, 718

of earth, 261 (Prob. 6-36)

energy density in, 174-175
frozen, 741n.
of moving charge, 781 (Prob. 16-7)
relation to magnetic flux density, 163-164,

field intensity (see

potential)

Magnetic saturation, 222, 226

Magnetic sheet current density, 204-205
Magnetic susceptibility, 207
Magnetic vector potential, 186
Magnetization, 202
induced, 212
permanent, 212
Magnetization curve, 223
Magnetizing force, 223
Magnetohydrodynamic (MHD) waves, 740-742
Magnetomotance (see Magnetomotive force)
Magnetomotive force (mmf), 168

Major lobe (see Main lobe)

Mapping, graphical: current, 131-134
electric field, 86-93
magnetic field, 242-250

permanent (see Permanent magnet)

U-shaped, 260-261 (Prob. 6-30)
Magnetic bottle, 746 (Prob. 15-29)
Magnetic circuit, 238
with air gap, 252-254
Magnetic dipoles, 201
Magnetic field:
in air gap, 252
figure, 244
boundary relations for, 218
definition, 163-164

Magnetic
Magnetic

821

elliptical loop,

Magnetic gap forces, 254

Magnetic levitation system, 258 (Prob. 6-9)
Magnetic moment, 154, 201

Matching:
broadband, 515
polarization, 435
Matching stub, 504
Matrix, 777
coherency, 441
Matthias, B. T 116n.
Maximum usable frequency (MUF), 744
(Prob. 15-14)
Maxwell, James Clerk, 354
Maxwell's equations, 354
from Ampere's law, 330, 355
from Faraday's law, 356

form, 319
309-310
354
from Gauss' law, 356, 357
for harmonically varying fields, 358
involving curl, 182
involving divergence, 97, 150
symmetry of, 361 (Prob. 9-5)
Maxwell's stress tensor, 775-780
Mayer, C. H., 701
Measurable quantity, 59
differential

integral form,

in free space,

Mechanical moment (see Torque)

Mechanical system, equivalent, 332
Megaparsec, 714 (Prob. 14-57)
Mesh, 355
Metal detectors, 346 (Prob. 8-35)
Metal plate lens, 686
Metal trough with partition, 301 (Prob. 7-17)
Meter, definition, 2
(see Magnetohydrodynamics)
Microwave communication circuit, 480

MHD

(Prob. 12-10)

Microwave components, 594-595 (Prob. 13-36)

822

INDEX

Microwave ovens, 599 (Prob. 13-61)

Microwave radiation hazard, 425 (Prob. 10-39)
Millikan's oil-drop experiment, 743 (Prob.
15-10)
Minimum detectable flux density, 702
Minimum detectable temperature, 701
Minimum usable frequency, 745 (Prob. 15-16)
Minkowski, H., 75 In.

Minkowski diagram, 751

Minor lobe, 604
Minor-lobe solid angle, 622
Mirror point, 746 (Prob. 15-28)
Mitchell, Edgar, 662

MKSA

system, 3
-Mobile units, 594 (Prob. 13-32)
Mobility, 116, 139 (Probs. 4-24, 4-25)

Nonmagnetic

materials,

200

Nonmetallic guide, 579

Nonspherical reflector, 470

Nonuniform transmission line, 486

Normal magnetization curve, 231
Normalized admittance, 503
Normalized impedance, 503
Normalized power pattern, 620
Normalized Stokes parameters, 438, 706
Normally dispersive medium, 376
North pole, magnetic, 140, 199
Notation, longhand, in.
Null directions of array, 632
Null principle, 346 (Prob. 8-35)
Numbering, equation, 5
Nyquist, H., 699n.

Mode, -482
higher-order, 482

Moment:
current, 142
electric-dipole (see Electric-dipole

Oblique incidence:

moment)

magnetic, 154

mechanical (see Torque)

Monochromatic

Moon,
Moon,

P., 765n.,

transmitter, 435

802

occultation of, 481 (Prob. 12-14)

Moore, A. D., 86, 743 (Prob. 15-9), 802
Motional induction, 312
Motional-induction law, 311
Motor, linear (see Linear motor for vehicle)
Motor equations, 142, 719
Moving charge:
electric field of, 781 (Prob. 16-6)
magnetic field of, 781 (Prob. 16-7)

Moving conductor, 310

Moving meterstick, 758
Moving system, fields of, 769-775

MUF

(see Maximum usable frequency)

Multilobed pattern, 642
Mumford, W. W., 425 (Prob. 10-39)
Mushiake, Y., 682n.
Mutual impedance, 323
of antenna, 673-675
Mutual inductance, 321

Nabla, 33
Nash, R. T., 694n.
National Aeronautics and Space Administration,
660
Navier-Stokes equation, 741

Near

field,

614

Near-field pattern, 604

Negative magnetic pole (see South pole)

Neumann function, 284, 552
Neumann's law, 307
Neumann's low-frequency inductance formula,
362 (Prob. 9-13)
Neutron star, 703

Newton, definition, 14n.

Newton's second law, 6, 717, 730
Noise power, 700
Noise temperature, 699
Nondispersive media, 376
Nonisotropic (see Anisotropic materials)
Nonlinear elements, 111

of elliptically polarized wave, 460

of linearly polarized wave, 452
Observer, viewpoint of, 757
Occultation of moon, 481 (Prob. 12-14)
Octopole:
distance factor for, 38
electric field intensity of, 53 (Prob. 2-16)
Ohio State University Radio Observatory, 661
Ohm, Georg Simon, 111

Ohmmeter, 493
Ohm's law, 111, 349
at a point, 118, 357
Oil-drop experiment, Millikan's, 743
(Prob. 15-10)
Onnes, H. Kamerlingh, 115
Open-circuited line, 499
Optical spectrum, 714 (Prob. 14-57)
Ordinary permeability, 326

Ordinary ray, 735

Orthogonality, 25
Outwardness, 18
Oven, microwave, 599 (Prob. 13-61)
Ozone, 593 (Prob. 13-22)

Parabola, 685
Parabolic dish antenna, 642
Parabolic line of charge, 108 (Prob. 3-47)
Parabolic wire, magnetic flux density of, 193
(Prob. 5-14)
Parallel-plate capacitor, 49, 61-62, 71-73
analogy with storm cloud, 105 (Prob. 3-21)
charging, 423 (Prob. 10-21)
divergence of
and P in, 99-101
force squeezing plates together, 104
(Prob. 3-6)
fringing field of, 303 (Prob. 7-37)
Laplace's equation for, 264-266
magnetic field in, 362 (Prob. 9-9)
power dissipated in, 345 (Probs. 8-25, 8-26)

problems involving, 103-104

with space charge, 289-292
Parallel polarization, 453, 458-460
Paramagnetic materials, 199
Partial differential equation, 367
Partially polarized wave, 436

INDEX

Particle:

energy of, 718

frequency of, 721
velocity of, 718
Particle accelerator, 725
Particular solution, 288
Particulate wastes, 743 (Prob. 15-9)

Pascal's triangle, 631

Pattern:

Pipe, closed rectangular, 485

Pipeline location, 258 (Prob. 6-10)
Pipelines, buried, 138 (Prob. 4-19)
Planck's constant, 395, 422 (Prob. 10-11), 695
Planck's radiation law, 422 (Prob. 10-11)
Plane of incidence, 453

Plane wave, 364

diffraction of,

466

Plasma, 729

interferometer, 642

Plasma frequency, 732

Plasma oscillations, 734
Plessner, K. W., 58n.
Plonsey, R., 802

mutilobed, 642

Plotter, computer-controlled,

antenna, 603
far-field,
field,

604

613

near-field,

604

Plume

power, 604
principal-plane, 604
rectangular, 604
Pattern multiplication, principle of, 629
Pendulum, wire, 341 (Prob. 8-5)
Permanent magnet, 141, 222

with gap, 255-256

materials for, table, 236
Permeability:
incremental, 326
infinitesimal,

325

226

maximum, 224
200
table, 201
of vacuum, 144, 200
Permeance, 239

intrinsic,

density, 101

Polarization fading, 659n.

Polarization matching, 435
Polarization measurements, antennas for, 704

Permittivity:

57

artificial dielectric,

Poincarf, H., 431n.

Poincare sphere, 431
Point charge, 13
Point-by-point method, 267-269
Point-to-point correspondence, 468
Point relation, 355
Point source (see Isotropic antenna)
Point-to-wave correspondence, 465
Poisson's equation, 103, 262, 288
Polarization:
degree of, 438

739
volume charge
wave, 364

relative,

air, 14,

velocity, 531
Pocket ruler antenna, 665-666
Pocket transistor radio, 715 (Prob. 14-63)

of dielectric, 59-63

ordinary, 326

of
of

276

Plume, water, 529

polar, 604

initial,

823

64-66, 104 (Probs.

3-10, 3-11)

of crystals, 58n.
of ice, 346 (Prob. 8-32)
relative, 57
table, 58
tensor, 733

of titanates, 58.
of vacuum, 13, 57
Perpendicular polarization, 453-457
Peters, Leon, Jr., 468n., 474n., 476n.
Phase center of antenna, 629, 648

Phase constant, 488

Phase factor, 404, 488
Phase shift, 383
Phase velocity, 369, 372
in guide, 559
rectangular, 545
relative, 373
of standing wave, 423 (Prob. 10-23)
on transmission line, 490
Phased array (see Arrays, scanning)
Phasor, 335
Phasor addition of fields, 636-639
Photon, 695
Photosphere, 742
Physical aperture, 690
Physical optics, 465
Physical optics approximation, 468

706
Polarization state, 431, 432
antenna, 435, 439

wave, 439
Polarizing angle, 459

Pole strength, 175

Poles, magnetic, 175
Positive magnetic pole (see North pole)
Positiveness, 17
Potato, Idaho, 346 (Prob. 8-34)

Potential:
absolute, 22
electric scalar: definition, 20n.
of linear charge distribution, 31

of octopole, 53 (Prob. 2-16)

of quadrupole, 52 (Prob. 2-15)
retarded, 604

of surface charge distribution, 31

of volume charge distribution, 32, 102
magnetostatic, 167
vector (see Vector potential)
Potential difference, 21
rise in, 25
Potts, Bing, 694/1.
Pounder, E. R., 346 (Prob. 8-32)

Powdered-iron cores, 327

Power:
complex, 619
through curvilinear square, 522

824

INDEX

dissipated
reactive,

by

dielectric,

335

619

Power
Power
Power
Power

density, surface, 391

divider, waveguide, 570
factor, 334
flow:
reactive, 650
real,

650

on transmission line, 520

Power flow lines, 417
Power-flow-map, 521
pattern, 604
normalized, 620

Power

in polar coordinates, 604

in rectangular coordinates, 604
relation to Poynting vector, 620
Power per unit area:

average, 393
instantaneous, 393

peak, 393
Poynting, J. H., 389

Poynting vector, 389

average, 392, 430
circuit application of, 416
complex, 391, 429
in conducting media, 413
flow lines of, 418
instantaneous, .390
Precipitator, electrostatic, 743 (Prob. 15-9)
Prefixes for units, 4
Press, electrostatic, 104 (Prob. 3-15)
Pressure, 776
radiation, 778-780

Primary pattern, 629

Primary winding, 320
Principal-plane pattern, 604
Principle of superposition:
of electric fields, 18
of electric potentials, 30
Prolate spheroid antenna, 648-649

Propagation, 363
longitudinal, 735
transverse, 735
Propagation constant, 404, 488, 499
Propulsion of space vehicle, 747 (Prob. 15-37),
782 (Prob. 16-19)
Pseudo-vector, 335
Pulsar, 702-704, 714 (Prob. 14-58)
Pulse, square, 517

Q, 587-589
Quadratic equation, solution of, 797
Quadrupole, distance factor for, 38
Quanta, 395
Quarter-wave plate, 509
Quarter-wave transformer, 509-511
double section, 513
single section, 509
three section, 514..592 (Prob. 13-16)
wave reflections on, 515
Quartz crystal, Q for, 599 (Prob. 13-58)
Quartz window, 593 (Prob. 13-20)
Quasar, 714 (Prob. 14-57)
Quasi conductor, 401
Quasi-longitudinal conditions, 736

Quasi-stationary field, 614

Quasi-transverse conditions, 736
Quiet sun, 734

Radar,

CW doppler,

Radar cross

section,

713 (Prob. 14-51)

698

of automobile, 713 (Prob, 14-51)

of electron, 744 (Prob. 15-11)
Radar equation, 698
Radar system, 425 (Prob. 10-36)
design of, 713 (Prob. 14-49)
Radiation field, 615
(See also Far field)
Radiation hazard, microwave, 425 (Prob. 10-39)
Radiation intensity, 620
Radiation pressure, 778-780
on conductor, 784 (Prob. 16-30)
Radiation resistance, 617
Radio, pocket transistor, 715 (Prob. 14-63)
Radio communication, 480 (Prob. 12-9)
Radio lines, 751
Radio source, extragalactic, 421 (Prob. 10-9)

Radio telescope:
of Ohio State University, 661, 702
of University of Texas, 661
Radio-telescope antenna, 700
Radio-telescope receiver, sensitivity of, 701

Radiometer, 700
Ramo, S., 802
Rank of root, 554
Ray paths, 468
Rayleigh, Lord, 259 (Prob. 6-24), 532n.
Reactive energy, 398
Reactive power, 619
Real part of e, 334
Reber, Grote, 695, 696
Receiving antenna, 601, 624
Reciprocity, 603, 668-670
directivity, 669-670
impedance, 670
pattern, 669
Reciprocity theorem, 323, 668
Rectangular impedance chart, 502
Rectangular loop, rotating, 342 (Prob. 8-9)
Rectangular trough transmission line, 298
(Prob. 7-7)

Recurrence relations for Bessel functions, 798

Red shift, 714 (Prob. 14-57), 783 (Prob. 16-25)
Rees, M. J., 703
Reflected wave, 383, 451, 453
Reflection coefficient, 387, 447
for current, 497
for voltage, 497
Reflection coefficients, Fresnel, 460
Reflector (see Antennas, reflector)
Refracted wave, 453n.
Refraction, angle of, 453n.
Relative motion, 750
Relative phase velocity, 373
Relativity,

757

Relaxation time, 409

Reluctance, 239, 327

Remanence, 229
Residual density (see Remanence)
Resistance, 111

INDEX

Semiconductor rod, 304 (Prob. 7-43)

Semiconductors, 109-110
germanium, 139 (Probs. 4-24, 4-25)

antenna, 625
loss,

619

radiation, 617
Resistance-measurement method, 494
Resistance paper, 277, 492
Resistivity, 114
Resnick, R, 801
Resonant frequency, 584
of atomic dipole, 333
Resonant wavelength, 584
Resonator, 517, 583
Response, system, 440
Resultant wave, 525
Retardation time, 605n.
Retarded current, 605, 647
Retarded potentials, 604
Retarded scalar potential, 606
Retarded time, 757
Retarded vector potential, 606

Sensitivity of receiver, 701

Sensors, cosmic-ray, 695

Separation of variables, 263, 278, 284, 286, 537

Septum, waveguide, 570

Series circuit, 352
Series expansions, 797
Series impedance, 485, 534
Seyfert galaxy, 714 (Prob. 14-57)
Shear, 776
Shearing line, 256
Sheet-current density, 415
Shell:

of charge, 43-45
conducting, 46-48
Shepard, Alan, Jr., 662
Shock waves, 734

Retentivity, 230
Retroreflector antenna, 713 (Prob. 14-52)
heaters, 599 (Prob. 13-61)
heating, industrial, 332
Right circular polarization, 429

RF
RF

Right-hand

rule,

Short-circuited line, 499

Short dipole, 606-617
effective aperture of,

near

Right-handedness, 17
Risman, P. O., 346 (Prob. 8-34)
Rotating bar magnet as pulsar, 702-704
Rotation measure, 739

Rowland

ring, 224
Rudduck, R. C, 476n.
Rumsey, V. H., 666n.
of,

402

Salisbury, W. W., 596 (Prob. 13-38)

Saltwater, plane wave on, 479 (Prob. 12-1
Satellite:

earth resources, 714 (Prob. 14-55)

stationary, 745 (Prob. 15-16)
Saturation, magnetic, 222, 226
Scalar potential:
electric, 20-22, 340
(See also Potential, electric scalar)
magnetic, 167
(See also Potential, magnetostatic)
Scalar wave equation, 366
.

Scanning arrays, 663-665

Scarborough, J. B., 267
Scattering from strip, 472-474
Scattering aperture, effective, 698
Schelkunoff, S. A., 802
Schmidt telescope, 711 (Prob. 14-44)
Screen, fluorescent, 721
Search coil, 347 (Prob. 8-35)
Seawater, electrical behavior of, 402
Second, definition, 3
Secondary dimensions, 2, 786
Secondary units, 2-4
Secondary winding, 320
Self-impedance of antenna, 670-673
Self-inductance, 322
Semiconductor junction, 304 (Prob. 7-43)

625

energy flow of, 614

far fields of, 613

612-613
614

fields of,

140

figure, 141

Rural ground, electrical behavior

Russo, P. M., 476n.
Ryan, C. E., Jr., 474n.

825

fields of,

radiation resistance of, 617-619

scalar potential for, 610
vector potential for, 609
Shortwave antenna, 599 (Prob. 13-62)
Shunt admittance, 485, 534
SI (International System of Units), 2-3, 786
Sidelobe, 604
Silver, S., 684n.

Silver paint,

277

Simultaneity, 747-751
Sinclair, G., 435.
Sine integrals, 648
Single-stub tuner, 504
Single-wire line with ground return, 85-86
Singular point, 29
Skew-symmetric matrix, 767
Skew-symmetric tensor, 734
Skilling, H. H., 180n.

Skin effect, 406

Skin resistance, 416
Sky survey, 702
Sky temperature, 700
Slater,

J.

C,

514n.,

570n 572n.

Sloanaker,
M., 701
Slot antennas, 687-688

Slow-wave structure, 665

Small helical antenna, 655-656
Smith, P. H., 505n.
Smith chart, 505

Smythe, W.

R, 802

Snell's law, 455

of reflection, 455n.

Soap bubble, expanding, 751

Solar atmosphere, 734
Solar corona, 742
Solar disturbances, 734
Solar power output. 422 (Prob. 10-13)
Solar wind, 783 (Prob. 16-21)
Solenoid, 155-158
and equivalent rod, 203-205, 215

826

INDEX

inductance of, 160

magnetic field of, 156
magnetic flux density, 192 (Prob. 5-9)
Solenoidal currents, 124
Soleno'idal field, 186
Solenoidal flux tubes, 150
Solid angle, 620

beam, 620-621
main-lobe, 622
minor-lobe, 622
Solid zonal harmonic, 284
Sommerfeld, A., 476n., 579n., 802
Source temperature, 701
South pole, 199
Southworth, G. C, 551n.
Space and time quadrature, 398
Space charge, 288
Space-charge limitation, 289
Space cloth, 449
Space paper, 449, 493
Space-time concept, 752-757
Space-time diagram (see Minkowski diagram)
Space vehicle, propulsion of, 747 (Prob. 15-37)
Spaceship, 422 (Prob. 10-14)

Spangenberg, K. R., 289

Spectrograph, mass, 728
chart, 10-11
Spencer, D. E., 765n., 802
Sphere, Poincar6, 431
Spherical conducting shell, thin, 302 (Prob.
7-34)
Spherical reflector, 711 (Prob. 14-44)
Spherical trigonometry, 432
Spherical wave, 613
Spheroid radiator, 648
Spring constant, 332
Square trough, infinite, 269-272
Square trough transmission line, 297 (Prob. 7-4)

SPEMP

Stabilized magnet,

256

Superconducting solenoid, 345 (Prob. 8-31)

Superconductors, 115, 593 (Prob. 13-23)
Supergain antenna, 659
Suppressed carrier, 377
Surface wave,. 457, 579
Surfers, analogy of particles to, 704
Surge impedance (see Characteristic impedance)
1

Susceptibility:

63
magnetic, 207
Suzuki, $., 705/1.
electric,

Symbol, 786
Symbols, how to read, 4-5
Synchrotron, 747 (Prob. 15-36)
Synchrotron radiation, 704
System temperature, 701

Tables:
array

beam widths and

boundary
electric

directivities,

662

relations: electric fields, 71

and magnetic

magnetic

fields,

fields,

338

221

impedance of lines, 490, 496

charge and mass of particles, 728
comparison of circuit and field quantities, 491
comparison of field relations, 168
comparison of time-varying electric and magnetic field relations, 340
characteristic

conductivities, 117
cylindrical waveguide modes, 557
dielectric strength, 73
electric and magnetic fields compared, 191
field-map quantities, 249
fields of short dipole, 616

input impedance of terminated line, 500

Maxwell's equations: in differential form, 361
in integral form, 360
multiples and submultiples of basic units, 785
normalized Stokes parameters, 438
parameters for six-element array, 639

Standing wave, 385

Standing-wave envelope, 385, 386
Standing-Wave Ratio (SWR), 386
(See also Voltage Standing- Wave Ratio)
State, polarization, 431
Static electric field, 24

Pascal's triangle, 630

penetration depths, etc. for conducting media,

Staticness, 15

permeability, 201

Stationary envelope, 385

Stationary observer, 773
Stationary satellites, 745 (Prob. 15-16)
Stegun, I. A., 466n.
Stokes, G., 436
Stokes parameters, 436
normalized, 438
Stokes' theorem, 319
Stone, J. S., 631n.
Stored energy, 398
Storm cloud, energy in, 105 (Prob. 3-21)
Stratton, J. A., 375, 802
Stress tensor, Maxwell's, 775-780
Strip transmission line (see Stripline)
Stripline, 296 (Prob. 7-1), 590 (Prob. 13-6),
598 (Prob. 13-54)
Stub:

permittivities, 58
polarization and magnetization, 257
potential and field of charge configurations,

matching, 504
short circuited, 503
Substrate, dielectric, 590 (Prob. 13-6)

412
permanent magnetic materials, 236

39
Poynting vector and energy density relations,
395
propagation relations, 736
rectangular guide, relations for, 550
reflection and transmission coefficients, 502
solutions of wave equation, 371
Stokes parameters, relation to antenna pairs,
706
units, fundamental, mechanical, electrical, and
magnetic, 786-793
waveguide parameters, 569
Tai, C-T, 312, 532n., 557n., 674

Tapered

line,

486

Teledeltos paper, 277, 492

Telescope:
cosmic-ray, 695

index

radio (see Radio telescope)

Schmidt, 711 (Prob. 14-44)

Television receiver, 743 (Prob. 15-3)
Temperature:
minimum detectable, 701
noise, 699
sky, 700
source, 701
system, 701

Tension, 776
Tensor:
divergence of, 776n.
of rank 1, 771
skew-symmetric, 734
Tensor conductivity, 733n.
Tensor permittivity, 733
Tenuous plasma, 729
Terminated transmission line, 496

Terminated wave, 499

Terminations, waveguide, 570
Theoretical quantity, 59
Theory of images, 292
Thevenin generator, 625

Thompson scatter radar, 744 (Prob. 15-12)

3 dB beamwidth (see Half-power beam width)
Three-halves relation, 292
Three-port circulator, 591 (Prob. 13-11)

Thumbnail electromagnetics, 6-10

Tilt angle:

of polarization
of wave, 576

ellipse,

432

Time constant, postdetection, 701

Time-domain analysis, 517
Time-phase quadrature, 614
Toroid, 206
inductance of, 161
Toroidal coil, 320
with gap, 208-209

Torque:
aligning, 142
on dipole, 53 (Prob. 2-19)

on loop, 154-155, 195 (Prob. 5-20)

infinite-parallel-plane,

483

TE mode
infinite

in, 523
uniform, 486

496
nonuniform, 486
phase velocity on, 490
power flow on, 520
rectangular trough, 298 (Prob. 7-7)
square trough, 297 (Prob. 7-4)
lossless,

strip (see Stripline)

terminated, 496
triangular trough, 300 (Prob. 7-9)
of two conducting strips, 303 (Prob. 7-36)
two conductor, 495
Transmission line cell, 171, 492
impedance of, 381
Transmission-line charts, 502
Transmission-line equations, 490
Transmitted wave, 383, 451, 453
Transmitting antenna, 601, 624
Transverse electric (TE) mode, 524

Transverse ElectroMagnetic (TEM) mode, 351,

364
Transverse field components, 539
Transverse Magnetic (TM) mode, 351, 547
Transverse propagation, 735
Transverse-wave impedance, 574
Traveling wave on antenna, 651
Traveling-wave antenna, 650-654
terminated, 651
unterminated, 650
Triangular trough transmission line, 300
(Prob. 7-9)
Trigonometric relations, 795
Trigonometry, spherical, 432
Trough carrying brackish water, 342 (Prob.
8-12)
Tsuchiya, A., 705n.
Tube of rays, astigmatic, 470

Tubes:
of current, 123-124
of electric flux, 39-40, 87

magnetic, 140
Total internal reflection, principle of, 457
Total transmission, 458
Trace, 441

Tuner:

Transformer, 324
bandwidth, 511-515
quarter-wave, 509-511
spherical-to-plane wave, 686
Transformer core, 323
Transformer induction, 312
Transformer induction equation, 309
Transitions, waveguide, 570
Transmission coefficient, 447
for current, 501

Tunnel, vehicular, 594 (Prob. 13-32)

TV, 480 (Prob. 12-10)

for voltage, 501

Transmission line, 482

balanced, 606
coaxial, 484
dielectric rod, 483
distortionless, 489

exponential, 591 (Prob. 13-12)

four-conductor, 301 (Prob. 7-25)
half-cylindrical, 300 (Prob. 7-10)
high power, 593 (Prob. 13-22)

827

double-stub, 506, 590 (Prob. 13-9), 592

(Prob. 13-17)
single-stub,

504

Two-conductor line (see Two-wire line)

Two-element interferometer, 640-642
Two-element vacuum tube, 289
Two metal cones, 300 (Prob. 7-16)

Two

metal sheets at angle, 301 (Prob. 7-22)

Two- wire

line:

capacitance of, 85
characteristic

impedance

of,

495

inductance of, 162

potential of, 82-83

Uda, S., 682n.

Uniform array, 631
Uniform transmission line, 485
wave equations for, 486
(See also Transmission line)
solution, 266, 267

Unique

INDEX

828

Uniqueness, 266-267
Unit, 2
U.S. Naval Research Laboratory, 701
University of Texas, radio telescope of, 662
Unpolarized wave, 435
Upatnieks, J., 467n.
Urban ground, electrical behavior of, 402

antenna, 602, 672

perfect, 421 (Prob. 10-3)
Vacuum chamber, 593 (Prob. 13-20)
van Atta array, 713 (Prob. 14-53)
Van Duzer, R., 802
Vector, complex, 430

Wave impedance, 574

Wave path, 525
Wave polarization, 364
Wave-reflection method, 515
Wave reflections on quarter-wave transformer,

Vector identities, 797

Vector potential:

515

magnetic, 186, 188, 340

Wave

retarded, 606

Vehicles:

Waveguide, 483, 524

motor for, 344 (Prob. 8-20)

radio equipped, 424 (Prob. 10-36)

linear

of,

coated conductor, 580

corrugated, 579
cylindrical, 551
dielectric slab, 581
discontinuities in, 543

750

Vehicular tunnel, 594 (Prob. 13-32)

Velocity, 367
electron, 116, 139 (Probs. 4-24, 4-25)
energy, 392-529
group, 376
hole, 116, 139 (Probs. 4-24, 4-25)
of light, 372
phase (see Phase velocity)
relativistic, 784 (Prob. 16-29)
Venus, echo power from, 713 (Prob. 14-49)
Vertically polarized wave, 364
Vibrational modes, 334
Viewpoint of observer. 757

Voltage reflection coefficient, 497

Voltage response of antenna. 435
Voltage Standing- Wave Ratio (VSWR), 386,
500
Voltmeter:
digital (DVM), 277
high-resistance, 277

Von

Hippel, A.

R,

58n.

Walter, C. H., 644n.

Washer, conducting, magnetic flux density

193 (Prob. 5-12)
Water, distilled, 425 (Prob. 10-38)

Water plume, 529

Watson, G. N., 802

Wave:
circularly polarized, 427
elliptically polarized, 426
linearly polarized, 426
partially polarized, 436

propagation of, 363

reflected, 451
resultant, 525
spherical, 613
surface, 457, 579
transmitted, 451
Wave absorption, 451

(388)

;\

transformer, 686

Wavefront, 525
Wavefronts, 468

Vector product (see Cross product)

motion

(see Beverage antenna)

Wave-to-antenna coupling factor, 439
Wave equation, 536
for conductors, 403
for dielectrics, 366
general development, 419
for rectangular waveguide, 536
scalar, 366
for transmission line, 486
Wave equation solutions:

exponential, 371
trigonometric, 371

Vacuum,

relative

Wave antenna

of,

elliptical,

560

G-string, 582

nonmetallic, 579
open surface, 484

phase velocity in, 559

rectangular, 532-550
reentrant, 560
screw in, 548
square, 561
wavelength in, 559
Waveguide devices, 570-571
Waveguide iris, 571-573
Waveguide power dividers, 570
Waveguide terminations, 570
Waveguide transitions, 570
Wavelength, 367, 372
Wavelength in guide, 559
West, R., 58n.
Wheeler, H. A, 656n.
Whinnery, J. R, 802
Whitmer, R. M., 582n.
F., 596 (Prob. 13-38)
Wickersham,'
Widner, Ronald, 694n.
Windows in building, 594 (Prob. 13-33)
Wolf, E 432n., 441n.
World lines, 751

Xerographic machine, 55 (Prob. 2-42)

Yagi, H., 682n.

Yagi-Uda antenna, 682

Zenneck, J., 579.

Zonal harmonics, solid, 284

GRADIENT, DIVERGENCE, CURL AND LAPLACIAN

IN RECTANGULAR, CYLINDRICAL, SPHERICAL,
AND GENERAL CURVILINEAR COORDINATES
Spherical Coordinates

Rectangular Coordinates

V/=r^ + fii^ + <j>-J-^

v
rdO

dr

V .A =

i^
r

dr

Vx A = ?

9 d<p

r sin

ox

+ -L-J-(v /4(|S
in0) +
"

r
'

r sin

9 89

'

h^-%

-^- 5
r sin 9

m, _
M,

i
1

d(f>

a
8_

dr

*)

d i

8f\

0/\

in

V x
r\dr

"

89

A=

*\8y

VA

A2

[dz

dX )

j8A y

8A X \

y
8

8y
Ay

8z

\-8x-~Jy-)

8x

Ax
d

8x

8y

+
8z

Cylindrical Coordinates

8z

r dq>

3 3x 3

cbc 2

hxh2 h3

(^

hiAi

ir

h3hiA2

h h
'

^)

r or

8z

Bq>

8_
1

h2 h
"a3 [lSx 2

h3

A3 -

8
h
dx

t
2

A2

h3 h

\8x 3

8x

h.hjh-.

dx

8 (h 3 h i 8f

8f\

8x 2 \ h 2

hY

hi

h2

8x 3 \ h 3 ox 3

h3

Xi

Rectangular

Cylindrical

Spherical

Mi\
+ A/Mi JMf+
+ A.l'hhi^LX
dxj ^n dxj
8xJ

d th 2 h 3
'__

V 2/ =

dtf>

dr
ox,

8z)

J8A^_8_A1\

General Curvilinear Coordinates

rtj

8z

JS^_dJA

ir

8y

+w
T72,

Bz

^ ^

V-A=
8x

8<j>

\sin

oy

r sin

x2
y
<t>

x3
z
z
<t>

d (^8f\

'

r 8r \

dr

f
+ 1
+
8^
1

'r

f_8 f +
d?~8?
8

I8f
rdr

\_&J_2+

8<f>

8^f
8z

1
r

8~r

8~4>

8~z

"U

Az

Ar
._l
2i

Kraus
and
Carver

OTHER McGRAW-HILL
INTERNATIONAL STUDENT EDITIONS
IN RELATED FIELDS
Angelo: ELECTRONICS-BJT's, FET's, and Microcircuits
Carlson: COMMUNICATION SYSTEMS-An Introduction to
Signals and Noise in Electrical Communication

2/e

FEEDBACK CONTROL SYSTEM ANALYSIS AND SYNTHESIS, 2/e

Elgerd: CONTROL SYSTEMS THEORY
Fitzgerald: BASIC ELECTRICAL ENGINEERING, 4/e
D'Azzo:

Fitzgerald

ELECTRIC MACHINERY, 3/e

Hayt: ENGINEERING

ELECTROMAGNETICS, 3/e

Hayt: ENGINEERING CIRCUIT ANALYSIS, 2/e

Johnson: TRANSMISSION LINES

AND NETWORKS

INTEGRATED ELECTRONICS
Millman: PULSE. DIGITAL, AND SWITCHING WAVEFORMS
Ryder: ENGINEERING ELECTRONICS, 2/e
Schilling: ELECTRONIC CIRCUITS
Schwartz: INFORMATION, TRANSMISSION, MODULATION,
AND NOISE
Taub: PRINCIPLES OF COMMUNICATION SYSTEMS
Terman: ELECTRONIC AND RADIO ENGINEERING, 4/e
Terman: ELECTRONIC MEASUREMENTS, 2/e
Millman:

McGRAW-HILL
I

KOGAKUSHA