^ ^ !

^
Sfetf
•X>

&5& "* .»•«

THE ELECTROMAGNETIC

.yu & uj
JJ
J- u.

Dale M. Grimes & Craig A. Grimes

World Scientific
THE ELECTROMAGNETIC
ORIGIN OF
QUANTUM THEORY
AND IIGHT
This page is intentionally left blank
THE ELECTROMAGNETIC
ORIGIN OF
QUANTUM THEORY
AND LIGHT

Dale M. Grimes & Craig A. Grimes

The Pennsylvania State University

V f e World Scientific
wk New Jersey •* London • Singapore
Sh • Hong Kong
Published by
World Scientific Publishing Co. Pte. Ltd.
P O Box 128, Farrer Road, Singapore 912805
USA office: Suite IB, 1060 Main Street, River Edge, NJ 07661
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.

THE ELECTROMAGNETIC ORIGIN OF QUANTUM THEORY AND LIGHT

Copyright © 2002 by World Scientific Publishing Co. Pte. Ltd.
All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means,
electronic or mechanical, including photocopying, recording or any information storage and retrieval
system now known or to be invented, without written permission from the Publisher.

For photocopying of material in this volume, please pay a copying fee through the Copyright
Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to
photocopy is not required from the publisher.

ISBN 981-02-4785-0

Printed in Singapore by World Scientific Printers (S) Pte Ltd

 Foreword

Suppose that the technical knowledge bases of electromagnetic theory and

the theory of electrons available in the year 2001 were available when the
historic interpretation of quantum theory was constructed. Would the
additional facts affect the outcome? That is, is the historic interpretation
uniquely correct as constructed? Knowledge of several physical properties
that would appear to play critical roles in quantum theory were discovered
after the interpretation was completed. These include {1} the standing
energy that accompanies and encompasses active, electrically small
volumes, {2} the power-frequency relationships in nonlinear systems, {3}
the possible directivity of modal fields, and {4} electron nonlocality. How
could it be that such basic physical effects would not importantly affect the
dynamic interaction between an electron and its nucleus, for both isolated
atoms and atoms immersed in a plane wave?
It is indisputable that the mathematics of quantum theory coupled with
the historic interpretation adequately accounts for observed physical
phenomena. It is also indisputable that, in contrast with other physical
disciplines, the historic interpretation requires special, indeed rather quixotic,
quantum theory axioms. Perhaps the philosophically most significant result
of the historic interpretation is the inherent uncertainty of physical events:
The status of all physical phenomena at any instant does not completely
specify the status an instant later. The inherent uncertainty goes against all
other natural philosophy.
In addition to an inherent uncertainty of physical events, in the absence
of the four physical phenomena listed above the axioms needed to
satisfactorily explain quantum mechanical phenomena require rejecting
selected portions of classical electromagnetism within atoms. In this book,
we incorporate these four phenomena into a new interpretation, and
understanding, of quantum theory. This is primarily based upon showing that
both the magnitudes and the consequences of radiation reaction forces have
been greatly underestimated. We find that such forces are responsible for the

V
VI The Electromagnetic Origin of Quantum Theory and Light

inherent stability of isolated atoms and for the nonlinear, regenerative drive
of transitions between eigenstates.
The interpretation that results from incorporating these physical
phenomena preserves the full applicability of electromagnetic field theory
within atoms and it enables calculating the complete electromagnetic
solution for a photon. Furthermore, it preserves causality in the sense that the
status of all physical phenomena at any instant completely specifies the
status an instant later.
Detailed three-dimensional, time-dependent plots of power, energy, and
electromagnetic stress in the vicinity of a radiating electric dipole, of mixed
electric and magnetic dipoles, and photons near a radiating atom are
maintained on website:
<http://www.ee.psu.edu/grimes/antennas/breakthrough.htm>

Dale M. Grimes
Craig A. Grimes
University Park, PA USA
September 2001
 Table of Contents

Foreword v

Prologue xiii

1. Classical Electrodynamics 1
1.1 Introductory Comments 1
1.2 Space and Time Dependence upon Speed 2
1.3 Four-Dimensional Space Time 5
1.4 Newton's Laws 7
1.5 Electrodynamics 9
1.6 The Field Equations 12
1.7 Accelerating Charges 16
1.8 The Maxwell Stress Tensor 17
1.9 Kinematic Properties of Fields 22
1.10 A Lemma for Calculation of Electromagnetic Fields 24
1.11 The Scalar Differential Equation 26
1.12 Radiation Fields in Spherical Coordinates 30
1.13 Electromagnetic Fields in a Box 34
References 36

2. Selected Boundary Value Problems 37

2.1 Traveling Waves 37
Scattering 40
2.2 Scattering of a Plane Wave by a Sphere 40
2.3 Ideal Spherical Scatterers 47
Biconical Transmitting Antennas 52
2.4 General Comments 52
2.5 Fields 54
2.6 TEMMode 57
2.7 Boundary Conditions 60
2.8 The Defining Integral Equations 65
vii
viii The Electromagnetic Origin of Quantum Theory and Light

2.9 Solution of the Biconical Antenna Problem 68

2.10 Power 75
2.11 Field Expansion for ^-Directed Exponential 78
An Incoming Plane Wave 82
2.12 Incoming TE Fields 82
2.13 Incoming TM Fields 83
2.14 Exterior Fields, Powers, and Forces 86
2.15 The Cross Sections 91
Biconical Receiving Antennas 95
2.16 General Comments 95
2.17 Fields of Receiving Antennas 97
2.18 Boundary Conditions 99
2.19 Zero Degree Solution 103
2.20 Non-Zero Degree Solutions 105
2.21 Surface Current Densities 106
2.22 Power 107
References Ill

3. Antenna Q 113
3.1 Instantaneous and Complex Power in Circuits 113
3.2 Instantaneous and Complex Power in Fields 117
3.3 Time Varying Power in Actual Radiation Fields 119
3.4 Comparison of Complex and Instantaneous Powers 122
3.5 Radiation Q 127
3.6 Chu's Q Analysis, TM Fields 131
3.7 Chu's Q Analysis, Exact for TM Fields 136
3.8 Chu's Q Analysis, TE Field 138
3.9 Chu's Q Analysis, Collocated TM and TE Modes 140
3.10 Q the Easy Way, Electrically Small Antennas 142
3.11 Q on the Basis of Time-Dependent Field Theory 142
3.12 Q of a Radiating Electric Dipole 149
3.13 Surface Pressure on Dipolar Source 154
3.14 Q of Radiating Magnetic Dipoles 158
3.15 Q of Collocated Electric and Magnetic Dipole Pair 159
3.16 Q of Collocated, Perpendicular Electric Dipoles 164
3.17 Four Collocated Electric and Magnetic Dipoles and
Multipoles 165
3.18 Numerical Characterization of Antennas 172
3.19 Experimental Characterization of Antennas 179
Contents ix

3.20 Q of Collocated Electric and Magnetic Dipoles:

Numerical and Experimental Characterizations 183
References 190

4. Quantum Theory 191

4.1 Electrons 192
4.2 Radiation Reaction Force 194
4.3 The Time-Independent Schrodinger Equation 198
4.4 The Uncertainty Principle 203
4.5 The Time-Dependent Schrodinger Equation 205
4.6 Quantum Operator Properties 209
4.7 Orthogonality 210
4.8 Electron Angular Momentum, Central Force Fields 212
4.9 The Coulomb Potential Source 215
4.10 Hydrogen Atom Eigenfunctions 220
4.11 Perturbation Analysis 223
4.12 Non-Ionizing Transitions 225
4.13 Absorption and Emission of Radiation 228
4.14 Electric Dipole Selection Rules for One Electron Atoms 231
4.15 Electron Spin 234
4.16 Many-Electron Problems 236
4.17 Electron Photo Effects 239
References 243

5. Photons 245
5.1 Power-Frequency Relationships 245
5.2 Length of the Wave Train and Radiation Q 251
5.3 Phase and Radial Dependence of Field Magnitude 254
5.4 Gain and Radiation Pattern 258
5.5 Kinematic Values of the Radiation 260
5.6 Telefields and Far Fields 266
5.7 Evaluation of Sum Si2on the Axes 270
5.8 Evaluation of Sums S22 and S32 on the Polar Axes 273
5.9 Evaluation of Sum S 32 inthe Equatorial Plane 279
5.10 Evaluation of Sum S22in the Equatorial Plane 281
5.11 The Axial Fields, Summary 283
5.12 Infinite Radius Radiation Pattern 287
5.13 Self-Consistent Field Analysis 290
5.14 Power and Energy Exchange 295
x The Electromagnetic Origin of Quantum Theory and Light

5.15 The Wave Train 297

5.16 Multipolar Moments 299
5.17 Field Stress on the Active Region 302
5.18 Summary 313
References 315

6. Epilogue 317
6.1 Historic Background 317
6.2 Overview 322
6.3 The Radiation Scenario 324
References 329

Appendix 331
A.l Introduction to Tensors 331
A.2 Tensor Operations 334
A.3 Tensor Symmetry 336
A.4 Differential Operations on Tensor Fields 337
A.5 Green's Function 340
A.6 The Potentials 345
A.7 Equivalent Sources 346
A.8 A Series Resonant Circuit 351
A.9 Q of Time Varying Systems 354
A.10 Bandwidth 357
A.l 1 Instantaneous and Complex Power in Radiation Fields 358
A.12 Conducting Boundary Conditions 361
A. 13 Uniqueness 364
A.14 Spherical Shell Dipole 365
Spherical Harmonics 367
A.15 Gamma Functions 367
A.16 Azimuth Angle Trigonometric Functions 371
A.17 Zenith Angle Legendre Functions 374
A. 18 Legendre Polynomials 379
A. 19 Associated Legendre Functions 383
A.20 Orthogonality 385
A.21 Recursion Relationships 387
A.22 Integrals of Legendre Functions 397
A.23 Integrals of Fractional Order Legendre Functions 402
Spherical Bessel Functions 405
A.24 The First Solution Form 405
Contents xi

A.25 The Second Solution Form 410

A.26 Tables of Spherical Bessel, Neumann, and Hankel Functions
414
A.27 Sums Over Spherical Bessel Functions 423
Multipolar Sources 428
A.28 Static Scalar Potentials 428
A.29 Static Vector Potentials 433
References 440

Index 441
This page is intentionally left blank
 Prologue

A radiating antenna sits in a standing energy field of its own making. Even at
the shortest wavelengths for which antennas have been made, if the antenna
is electrically too small, that is, if the length-to-wavelength ratio is too small,
the amount of standing field energy is so large it essentially shuts off energy
exchange. Yet an atom in the act of exchanging electromagnetic energy may
be scaled as an electrically short antenna, and standing energy is ignored by
quantum theory seemingly without consequence. Why, in one case, the
energy is dominant and, in the other, it plays no role has been a mystery. The
framers of the historic interpretation of quantum theory could not have
accounted for the standing energy, since an analysis of it was first
formulated more than twenty years after the interpretation was
accomplished. Similarly, nonlocality is a significant and essential feature of
eigenstate electrons, yet nonlocality played no historic role since it was
discovered a half-century after the interpretation was accomplished. Similar
statements apply to the power-frequency relationships of nonlinear systems
and to the maximum possible gain of an electromagnetic mode.
In this book, we form a simplified and deterministic interpretation of
quantum theory that accounts for standing energy in the radiation field, field
directivity, the power-frequency relationships, and electron nonlocality. We
find that all play integral and essential roles in atomic stability and energy
exchanges. Together they form a complete electromagnetic field solution of
quantum mechanical exchanges of electromagnetic energy, without the
separate axioms of the historic interpretation.
Stable atoms occupy space measured on the picometer scale of
dimensions and exchange energy during periods measured on the picosecond
scale of time. Since this dimensional combination precludes direct
observation, it is necessary to infer active atomic events from observations
over much larger distances and times. Large-scale measurements led the
framers of the historic interpretation to conclude that the equations of
classical electromagnetism do not fully apply on an atomic scale of

Xlll
XIV The Electromagnetic Origin of Quantum Theory and Light

shows no inherent distance or time scale limitations; derived results are

independent of an observer's size and speed. A primary purpose of this book
is to assist in resolving this enigma.
Although this book is primarily a monograph, early versions were used
as a text for topical courses in electromagnetic theory in the Electrical
Engineering Departments of the University of Michigan and the
Pennsylvania State University. A later version served as a text for a topical
course in theoretical physics in the Department of Physics and Astronomy of
the University of Kentucky; it was after this course that we began systematic
work on the book. Throughout the book, all theorems used were carefully
reexamined, with the emphasis used that best met the needs of the book. For
the same purpose we extract freely and without prejudice from accepted
works of electrical engineering, on the one hand, and physics, on the other.
Too often there is imperfect communication between these two sides of the
same coin. The result is an innovative way of viewing scattering phenomena,
radiation exchanges, and energy transfer by electromagnetic fields.
The equations of classical electromagnetism are derived and developed
in Chapter 1. In Chapter 2 the equations are applied to a series of
increasingly complex boundary value problems. The choice of solved
problems is based on two criteria: First, the solution form is a general one
that, when the modal coefficients are properly chosen, applies to any
electromagnetic problem, and hence to atomic radiation. Second, each
solution is electromagnetically complete; the solution is in the form of an
infinite series of constant coefficients times products of radial, spherical and
harmonic functions. Completeness is required to assure that no solution
forms have been overlooked. The importance of completeness cannot be
overemphasized. For example, historically the character of receiving current
modes on antennas was not correctly estimated; the full complexity and
beauty were not appreciated until an electromagnetically complete solution
to biconical receiving antennas became available. That is to say, our modern
technological culture, with its vast experience with antennas, did not
understand the modes on the simplest of receiving antennas until a complete
mathematical solution became available in 1982. Similarly, we cannot be
sure we fully understand a radiating atom without a complete solution.
Chapter 3 deals with standing energy fields associated with
electromagnetic energy exchanges. To analyze them, it is necessary to re-
examine complex power and energy in radiation fields. The use of complex
power is nearly universal in the analysis of electric circuits and fields.
Although complex power leads to the correct power at any terminal pair,
Prologue xv

expressions for complex power in a radiation field suppress a radius-

dependent phase factor. Hence, no equation that depends upon the phase of
field power versus radius can be solved using only complex power. To avoid
the difficulty, we switch to a time domain description of the fields, and use it
to calculate modal field energies. From them, we calculate the ratio of
source-associated field energy to the average energy per radian radiated
permanently away from the antenna. We confirm earlier work showing that
the ratio increases so rapidly with decreasing electrical size that all modal
combinations, save one, are subject to severe operational limitations. We
then show the multimodal combination to which such limitations do not
apply.
Chapter 4 contains a brief review of quantum theory that is conventional
in most ways, but unconventional in the treatment of atomic stability. We
show that the standing energy of a dipole field generated by an oscillating
point electron creates an expansive radiation reaction force on the electron.
That force is the same order of magnitude as the trapping Coulomb force and
is three orders of magnitude larger than the commonly accepted radiation
reaction force. We hypothesize it forces the electron into a charge density
distributed throughout the eigenstate. As an analogy, consider a drop of oil:
A small quantity of oil in isolation, and under the influence of surface
tension, forms into a spherical drop, yet when placed on a pond it expands to
cover the entire surface. An expanded electron is not small compared with
atomic dimensions and may be formed into a non-radiating array of charge
and current densities. Such an array is inherently stable and interaction
between the intrinsic and orbital magnetic moments produces a continuous
torque and assures continuous motion of the parts. This model and energy
conservation forms an adequate basis for Schrodinger's time-independent
wave equation; his time-dependent equation follows if the system remains in
near-equilibrium. In this way, Schrodinger's equations are the equivalent of
ensemble energy expressions in classical thermodynamics. In both places,
general results are obtained without detailed knowledge of the ensemble.
Schrodinger's time-dependent equation treats state transitions by
describing the initial and final states. Although answers are unquestionably
correct, the approach gives no information about fields that are present
during the emission and absorption process, yet clearly photon near fields
must exist. The classical interpretation of quantum theory supplies no
counterpart to the full field sets obtained in Chapter 2 for the transmission
and reception of energy by biconical antennas.
XVI The Electromagnetic Origin of Quantum Theory and Light

In Chapter 5, we integrate and extend results of the preceding chapters.

We begin by showing that the Manley Rowe equations, which are
meaningful only with nonlinear systems, correctly describe the Ritz power-
frequency relationships of photons; yet, the Schrodinger and Dirac equations
are linear. We next impose the kinematic properties of quantized radiation as
a boundary condition on a general, multimodal field expansion. The results
are field expressions with a form similar to those of Chapter 2, and are
closely related to multipolar expansions for plane waves. In addition, the
resulting modal fields are members of the set of resonant modes for which,
by Chapter 3, standing energy limitations do not apply. Hence, a rapid build-
up of radiated power is expected. These results are combined and used to
determine all radiated fields, near and far, during eigenstate transitions, i.e.,
during photon exchanges. Next, we use the similarity between photon and
plane wave multipolar expansions to re-express the photon fields in an
expansion from infinity inward. This expansion permits the evaluation of the
radiation reaction force of a photon field on its generating electron. We find
that the radiation reaction pressure on the surface of a spherical, radiating
atom is at least many thousands of times larger than the Coulomb attractive
pressure. The reaction pressure is properly directed and phased to drive the
nonlocal electron regeneratively, and nonlinearly, to a rapid buildup of
exchanged power. Hence, radiation in accordance with the Manley Rowe
power frequency relations occurs, and continues until all available energy is
radiated.
The mathematics and the ideas presented here, summarized on a point-
by-point basis in the Epilogue, combine to give a simple,
electromagnetically complete, and deterministic interpretation of quantum
theory.
 1. Classical Electrodynamics

1.1 Introductory Comments

There are two quite disparate approaches to electromagnetic field theory.
One is a deductive approach that begins with a single relativistic source
potential and deduces from it the full slate of classical equations of
electromagnetism. The other is an inductive approach that begins with the
experimentally determined force laws and induces from them, incorporating
new facts as needed, until the Maxwell equations are obtained. Although the
inductive approach is the way in which the theory was developed, it is the
deductive method that shows the full beauty, symmetry, and simplicity of
electromagnetism.
The inductive approach is commonly used in textbooks at all levels.
Coulomb's law is the usual starting point, with other effects included as
needed until the full slate of measurable quantities are obtained. From this
viewpoint, the potentials are but mathematical artifices that simplify force
field calculations. They simplify the calculation necessary to solve for the
force fields but are without intrinsic significance. The deductive approach
begins with a limited axiomatic base and develops a potential theory from
which, in turn, follow the force fields. In 1959 Aharonov and Bohm, using
the premise that potential has a special significance, predicted an effect that
was confirmed in 1960, the Aharonov-Bohm effect: Magnetic field
quantization is affected by a static magnetic potential even in a region void
of force fields. We conclude that the magnetic potential has a physical
significance in its own right and has meaning in a way that extends beyond
the calculation of force fields. There is physical significance contained in the
deductive approach that is not present in the inductive one.
To begin the deductive approach, consider that the universe is totally
empty of condensed matter but does contain light. What is the speed of the
light? Since there is no reference frame by which to measure it, the question
1
2 The Electromagnetic Origin of Quantum Theory and Light

is moot. Therefore, introduce an asteroid large enough to support an observer

and his equipment, which determines the speed of light passing him to be VA-
Since there is nothing else in the universe, a question about the speed of the
asteroid is moot. Next, introduce a second asteroid, identical to the first but
many light years away from it. The asteroids are separated far enough to be
independent of each other by any means of which we are currently aware.
An observer on the second asteroid determines the speed of light passing
him to be vg. Will the measured values be the same? By the cosmological
principle, an experiment run in one local four-space yields the same results
as an identical experiment run in a different local four-space. Therefore we
expect that VA = VB = c.
Next, bring the asteroids into the same local region. Either the speeds
depend upon the magnitude of the local masses or they do not, and if they do
not, there is no change in speed. However, in the local region, a relative
speed between asteroids A and B may be determined. Since there is no way
one asteroid can be preferred over the other in an otherwise empty universe,
the two observers continue to measure the same speed. This condition
requires that the speed of light be independent of the relative speed of the
system on which it is measured. Next, bring in other material, bit by bit, until
the universe is in its present form, and the conclusion remains the same. The
speed of light is independent of the speed of the object on which it is
measured, independently of the speed of other objects.

1.2 Space and Time Dependence upon Speed

Let a pulse of light be emitted from an origin in reference frame F and
observed in reference frame F'. If the speed of light is the same in all
reference frames, if the two frames are in relative motion, and if the origins
coincide at the time the light is emitted, the light positions as measured in the
two frames are:

x1 + y2 + z2-c2p- = j<>2+y1+l2-J2t2 (1.2.1)

If the relative speed is such that F' is moving at speed v in the z-direction
with respect to F, then at low speeds:

j=x, y=y, z>={z-vt); /=/ (1.2.2)

Classical Electrodynamics 3

Since Eq. 1.2.1 isn't satisfied by Eq. 1.2.2, it follows that Eq. 1.2.2 doesn't
extend to speeds that are a significantly large fraction of c. To obtain a
transition that is linear in the independent variables, and that goes to
Eq. 1.2.2 in the low speed limit, consider the linear transformation form:

x* = x; y = y, 2= y(z- vi); /= A/+ Bz (1.2.3)

Parameters y, A, and B are undetermined but independent of both position

and time. Since Eq. 1.2.3 approaches Eq. 1.2.2 in the limit of velocity v
much less than c, in that limit:

Y = l; A = l; B=0 (1.2.4)

Since the coordinates are independent variables, combining Eqs. 1.2.1

and 1.2.3 and solving shows that:

Z 2( Y 2_ 1 _ C 2 B 2J = ()

2
^ +YV-^A ) =0 (1.2.5)

Z/J^2+AB^2) = 0

Solving Eq. 1.2.5 yields:

A = Y = (l-v2/^2)~1/2; B = ~2 (1.2.6)

Combining yields the Lorentz transformation equations:

x'=x; y=y, z,= y(z-vS); /=y t- vz (1.2.7)

7j
This transformation preserves the speed of light in inertial frames.
Equation 1.2.7 is a sufficient basis upon which to determine results if
events in one frame of reference are observed in another one. Let the
observer be in the unprimed frame. A stick of length Lo as determined in the
moving frame, in which it is stationary, lies along the z-axis. It moves past
4 The Electromagnetic Origin of Quantum Theory and Light

the observer in the z-direction at speed v. A flash of light illuminates the

region, during which time the observer determines the positions of the ends
of the moving stick, z\ and z 2 . It follows from Eqs. 1.2.7 that the measured
positions are:

2X = YOi - W0) and *»2 = Y(^2 - Wo) (1-2-8)

The length as measured in the stationary frame is:

L = ( * 2 - * I ) = ( * W I ) / Y = LO/Y (1.2.9)

It follows that:

L=L0(l-v2/f2j <L0 (1.2.10)

The observed length of the stick is less than is measured in the rest frame;
this fractional contraction is the Lorentz contraction.
Next pulses of light are issued at times t'2 and f\, again in the moving
frame. When does a stationary observer see them, and what is the time
interval between them? Using Eqs. 1.2.7:

/2 = Y ( / 2 - ^ 2 / c 2 ) and /I = Y ( ' I - ' * I / ^ ) 0-2-11)

From Eq. 1.2.11 the time difference in the frame at which the two sources
are stationary is:

To = / , 2 - ^ = Y [ ( ^ 2 - ^ l ) - K ^ - ^ l ) A 2 ] = Y T ( l - ^ 2 ) (1-2-12)

T is the time measured in the stationary frame. Solving for T gives:

T I(L
= YT0 = i T T27 ^ T 0 O-*-")
(l-^f
The observer measures the time duration between pulses to be more than is
measured in the rest frame; this time expansion is time dilatation.
Classical Electrodynamics 5

1.3 Four-Dimensional Space Time

The equality of the speed of light in all inertial frames is the basis for a
system of 4-vectors. Let x\,X2,x3 represent the three spatial axes x,y^z of
three dimensions and x4 = ict where /'= V-l.

(xhx2,x3,x4) (1.3.1)

Since three of the axes determine lengths and one determines time, a
three-dimensional rotation represents a change in spatial orientation and a
four-dimensional rotation includes a change in time. Such four-dimensional
rotations are Lorentz transformations. These transformations are usually
simple and contain a high degree of symmetry. Such transformations are
covariant with respect to changes in coordinate systems; that is, an equation
that represents reality in one reference frame has the same form in all other
inertial frames.
The imaginary property of the fourth dimension represents an essential
difference from spatial ones: the squares of the space coefficients and time
coefficients have different signs. For notational purposes we use Roman or
Greek subscripts to indicate, respectively, three- or four-dimensional tensors.
For example, the rotation matrix element in four dimensions is c„ v where,
for velocities v directed along the xi-axis:

Y 0 0 tyvlc^
0 10 0
c _ (1.3.2)
nv
0 0 1 0
-fyvlc 0 0 y

It follows similarly to three-dimensional direction cosines that

C = C c
\X,V M\I ' |xv c np = "vp > " et l c ji.vl ~1 (1-3-3)

The Lorentz direction cosines c^v are:

c
\i ~ \ivxv (1.3.4)
6 The Electromagnetic Origin of Quantum Theory and Light

The proper time interval, Ax, between two events with space-time
coordinates spaced A a apart is defined to be:

? 1
(Ax) = T&xa&xa (1.3.5)
c

Using three-dimensional notation, the proper time difference is

(Ax)2 = ( A / ) 2 - ^ - (1.3.6)
c

Since (Ax)2 can be zero, positive, or negative, Ax may be zero, real, or

imaginary. Since the speed of light is the same in all reference frames, by
Eq. 1.2.1 the proper time is also the same in all reference frames. If it is real,
it is "time-like" and if imaginary, it is "space-like". If time-like, the proper
time is the time separation of the two events in the same frame. If space-like,
there is a frame in which c times the proper time is the spatial separation of
the two events that are simultaneous in that frame.
With x as proper time, consider the 4-vector defined by the expression:

dxu
V>-£ C3.7,

Since both x^ and x are independent of details of the particular inertial frame
in which it is measured, so is U^; U^ is therefore a 4-vector with the four
components:

dr dr d/ TT dy dy d/
U,1 = — = = yv
r x x ; U2 2 = — = — — =, yv; v
dx d/ dx di d/ck , N
. , (1.3.8)
dz dz d/ TT d[ict)
U 3 = — = — — = yvz ; \J4=^-^ = yic
dx d/ dx dx
The three-dimensional velocity components are v, and the 4-velocity
components are U„.
Classical Electrodynamics 7

A particle of mass mo with 4-velocity U^ has 4-momentum given by:

P^ = m 0 U^ (1.3.9)

Combining shows the momentum components to be:

//=ym0v ; j?4 = ym0/c= iW/c ; W-ym^c (1.3.10)

The quantity W, defined by Eq. 1.3.10, is the energy associated with the
moving mass.
The binomial expansion is:

(l±a)n = l±na + ^ ( n - l ) a 2 ± . . . (1.3.11)

This equation combines with the definition of y, see Eq. 1.2.6, to show that:

(1.3.12)
2c2 8c 4

Combining Eqs. 1.3.10 and 1.3.12 shows the total energy of the particle:

S 3v4
/F= m 0 ^ 1 + + ... (1.3.13)
2c2 8c4

In the rest frame mo is the rest mass. The particle energy is:

WQ = m0c2 (1.3.14)

By Eq. 1.3.14, the first term of Eq. 1.3.13 is the self-energy of the mass. The
second term is the kinetic energy at low speeds and the higher order terms
complete the evaluation of the kinetic energy of the mass at any speed.

1.4 Newton's Laws

The Minkowski force is defined to be:
8 The Electromagnetic Origin of Quantum Theory and Light

F =v
»i» (lAl>
This force is a 4-vector with the x-directed component:

Fx=— (m 0 U 1 ) = Y^(Ym 0 v r ) (1.4.2)

The corresponding three-dimensional force component is:

The factor y in Eq. 1.4.3 was known before the full relativistic effect was
understood. Although relativity makes it abundantly clear that the result is a
space-time effect, it was historically interpreted as an increase in mass
whereby the effective mass m is a function of speed:

m = Ym0 (1-4.4)

Even with relativity, the nomenclature remains and the effective mass of a
moving particle, by definition, is equal to Eq. 1.4.4.
Since the 4-momentum is a 4-vector, it is conserved between Lorentz
frames. That is,

wl = W1-p1c1 (1.4.5)

The energy is related to momentum, in any given frame, as:

fF2 = mlc4 + /?2c2 (1.4.6)

Since W is second order in v/c, three-momentum is constant in low speed

inertial frames. Energy is also nearly conserved. However, in high-energy
systems neither energy nor momentum is conserved, only the combination.
This example illustrates a general characteristic of 4-tensors that at low
speeds the real and imaginary parts are separately nearly conserved but at
high speeds only the combined magnitude is conserved.
Classical Electrodynamics 9

1.5 Electrodynamics
The three scalars defined so far are speed, c, time interval between events in
a rest frame, x, and mass, mo. A fourth is electric charge, q; electric charge
can have either sign. Just as an intrinsic part of any mass is the associated
gravitational field, G, an intrinsic part of charge is the associated 4-vector
potential field A„. Consider that the individual charges are much smaller
than other dimensions and that there are many of them. For this case choose
a differential volume, with dimensions (x\^2^'i), m which each dimension is
much less than any macroscopic dimension of interest but contains large
numbers of charges. If both conditions are met, the tools of calculus apply.
Charge density p is defined to be the charge per unit volume at a point.
Charge density po is defined in a frame in which the time average position is
at rest. Observers in fixed and moving frames see the same total charge but,
because of the Lorentz contraction, the moving observer determines the
volume containing it to be smaller by a factor of y. Therefore, the charge
density in a moving frame is increased by the factor:

P=YPo (1-5-1)

If the charge density moves with 4-velocity U„, in a way similar to three
dimensions the 4-current density is defined to be:

Ju = PoU u = {ypQV,y/cp0} = {j,/qp} (1.5.2)

The vector terms within the curly brackets indicate the first three
dimensions, and the scalar term represents the fourth dimension. The 4-
divergence of the current density is:

The first equality of Eq. 1.5.3 follows from definition of terms and the
second is true if and only if net charge is neither created nor destroyed. Pair
production or annihilation may occur but there is no change in the total
charge. The zero 4-divergence shows that the net change in the four-current
10 The Electromagnetic Origin of Quantum Theory and Light

is always equal to zero. Physically a net change in the total charge does not
occur and charges are created and destroyed only in canceling pairs.
The 4-vector potential field A^(Xy) is defined to be the potential that
satisfies the differential equation:

a2Av
(L5 4)
d^r^-^ -
Constant |X is defined to be the permeability of free space; it is a dimension-
determining constant and defined to equal 4rc/107 Henrys/meter.
Taking the 4-divergence of Eq. 1.5.4 then combining with Eq. 1.5.3
gives:

9XV 9Xn9Xg 3Xg3Xg 9XV 3XV

Combining, it follows that:

9A V /9X V = 0 (1.5.5)

Equation 1.5.5 shows that the divergence of A v is zero, from which it

follows that, like charge, the total amount of 4-potential doesn't change. If
transitions are made between different reference frames changes occur in the
components of the potential but not in the sum over all four components.
The four-dimensional Laplacian of Eq. 1.5.4 may be integrated over all
space to obtain an expression for the potential itself. By Eq. A.6.2 the
potential of a moving charge is:

Aa(XY) = f f [ f J ? ( , V , - R / , V (,.5.6)

The integral is over all source-bearing regions, dV is differential volume, Xy

are the 4-coordinates of the field point, X'y are the 4-coordinates at the
Classical Electrodynamics 11

source point, R is the vector from the source point to the field point. At low
speeds Eq. 1.5.6 simplifies to:

Aa(XYy) = - M
JJJ
JJ^^ J-MF (1.5.7)
47t R(XY,Xy)

Substituting in the three-dimensional values of J a results in the three-

dimensional potentials:

H fff j(/V'-R/4

(1.5.8)
*(/•,/)
v = -*A4V4(,,/)
y = J - ffT P t Z l ^ f W
' 4TOJJJ RO-/•',/)

Constant e is defined to be the permittivity of free space; it is a dimension

determining constant and defined to be exactly equal to l/((0.c2)
Farads/meter.
For a point charge, instead of a charge distribution, the corresponding 4-
potential is:

Aa(Xv) = — / hr (1-5-9)

The three-dimensional potentials are:

A(v ) = JL*£^M
' 4n (R-R«>>/c)
/ , x (1.5.10)
V ;
47t(R-R«v/c)

It the charge moves at a speed much less than c Eqs. 1.5.10 are the usual
three-dimensional vector and scalar potential fields of individual charges.
It is apparent from Eqs. 1.5.10 that a charge moving towards or away
from a field point generates potentials with magnitudes respectively larger or
smaller than the low speed value.
12 The Electromagnetic Origin of Quantum Theory and Light

1.6 The Field Equations

If po is the charge density in an inertial reference frame in which the average

speed of the charges is zero, then p = ypo is the charge density in a moving
frame. The charge density and the three-dimensional current density J; were
extended to form the 4-current density, as shown by Eq. 1.5.2, from which
the Laplacian of the 4-potential was defined by Eq. 1.5.4. Other useful 4-
tensors follow from 4-dimensional operations on the 4-potential Aa(Xy,);
some especially important ones follow.
A second rank antisymmetric tensor of interest follows from the potential
by the equation:

Antisymmetric 4-tensors are spatial arrays of six numbers and, in common

with all antisymmetric tensors, the trace is zero:

faa=0 (1.6.2)

Writing out the six values that appear in the upper right portion of the 4-
tensor, and using the result to define function O, gives:

_ BA2 3At = oAy dAx =

12
dX{ dX2 dx dy

dAy dAz_

dz ay

_ dAx dAz _
dz dx

_ 3A4 3At _ / BO dAx _ i

14 x
~l^~l)X^~~c~dx~~tidt~~ c
Classical Electrodynamics 13

f =
24 = ~ ~ - T^T —E?
c dy icdt
(1.6.3)
id® 9AZ_ /
r
34-—H ^37"—b-
C OZ /CO? C

With the deductive approach to electromagnetism, Eqs. 1.6.3 are the

defining terms for field vectors E and B. The result written in tensor form is:

0 Bz -By -iEjc
-Bz 0 Bx -iEy/c
(f) = (1.6.4)
By -Bx 0 -iEz/c
iEx/c feylc iEz/c 0

Differentiating fap with respect to Xp results in the equality chain:

3f,«p 3 (dAP 3A a 32AP 32A

*-=HJ<x (1.6.5)
8Xp 9Xp 3X a 8Xp 9Xp3Xa 8Xp3Xp

Combining terms:

3f,«P _
MJo (1.6.6)
3xP

Evaluating Eq. 1.6.6 results in:

9f,ip _ 3B Z 9B^ 1 3E,

= J*Jj
3Xp dy dz ct df

9f
2p _ dBx dBz 1 5Ey
= MJ^
3Xp dz dx ,2 dt
14 The Electromagnetic Origin of Quantum Theory and Light

3f 3p dQv 3B 1 3E.
y • HJ,
3Xp d* dy <2 dt
(1.6.7)
•3f, 4P dEr 5E P
>
/ 3Xn dx dy dz e

These are the nonhomogeneous Maxwell equations and relate fields to

sources. In three-dimensional notation:

Vx-^f-M, eV«E = p (1.6.8)

The nonhomogeneous Maxwell equations relate force field intensities E and

B to sources p and J. The first order terms of E and B are, respectively,
independent of and proportional to the first power of the speed of the charge.
It follows from the definition of fao that:

3f,vo dfnn
L df.
ocv
3Xa + -3XaaV +•
oXCT
(1.6.9)

Evaluation of Eq. 1.6.9 for each tensor component shows that:

BfiT 3fo-i df:

23 , Ui 31 _
+ •
ax, ax, ax
af24 at41 , 3f 12 = * I
+ 3XT dx.*
dXx (1.6.10)
af
, 42 , df23 _ 1
3*34
3 ax 4
ax 7 ax
9f
14 , 43 , ddf,
9f f
31 _ 1

ax3 axj ax 4 /c
These are the homogeneous Maxwell equations and relate force field vectors
E and B. In three-dimensional notation:
Classical Electrodynamics 15

V x E — - = 0; V«B = 0 (1.6.11)

Another useful 4-vector is the force intensity, defined by the equation

^ = f a pj p (1.6.12)

Evaluation of each component of Eq. 1.6.12 shows that:

F\ = F% =JJJSz-JzBJ,+ pEx

„v CN (1.6.13)

^ = -(EA+V>'+EA)

These equations relate force and power to the interaction of the charges and
the fields. In three-dimensional notation:

/^ v = pE + J x B ; -fc/$=E»J (1.6.14)

To assist in the interpretation of Eq. 1.6.12, consider the 4-scalar formed by

taking the scalar product:

^ J a = fapJ a Jp = 0 (1.6.15)

The second equality of Eq. 1.6.15 follows from the antisymmetric character
of f a p and shows that the 4-vector F^ is perpendicular to the 4-current
density. Since the 4-current density is proportional to the 4-velocity, it
follows that F^ is also perpendicular to the 4-velocity. Consider the
differential with respect to proper time of the square of the 4-velocity:
16 The Electromagnetic Origin of Quantum Theory and Light

Therefore both the 4-acceleration and j£ are perpendicular to the 4-

velocity. This is a necessary but insufficient requirement for j£ to be the
force density.
This approach to the Maxwell equations is based upon the original axiom
relating a charge to its accompanying potential. The form of the source
shows that only charges produce a 4-curvature of the 4-potential field. The
technique is a neat way both to package the electromagnetic equations and to
show that they take the same form in all inertial coordinate systems. The
relationship between fields E and B and the potentials follows from
Eq. 1.6.3. By direct comparison

3AL_aAi=
Bt=>VxA = B
dx-, dx\
(1.6.17)
J d A
r- -=Ei=>- ' ™ :E
ox\ at

1.7 Accelerating Charges

The potentials surrounding electric charges in uniform motion are given by
Eq. 1.5.10 and the force fields are related to the potential by Eqs. 1.6.3. The
partial derivative operations of Eqs. 1.6.3 take place at the field position and
time, (r,t). The position and time at the source, (r',f), do not enter into the
operations. To carry out the operations it is convenient to define S by the
equation:

R«v
S= R- (1.7.1)

Operating upon the potential while keeping terms involving charge

accelerations gives:

E= J > 2 3
R-R-
1
2 3
Rx a
R-R^ x — v
4ne | Y S c) <- S c dt (1.7.2)
B= —RxE
R^
Classical Electrodynamics 17

Keeping only first order terms in powers of vie leads to:

E= R-R- + - r R x Rx —v
4TCR J v c)
(1.7.3)
V4 „ I R3
B=— ^ R x v+——v
4TIR J c dt

The equations show that: A stationary charge produces an electric field

intensity that varies as the inverse square of the radius, but there is no
magnetic field. If the charge is moving, both electric and magnetic field
intensities exist that are proportional to the speed of the charge and that vary
as the inverse square of the radius. If the charge is accelerating, both electric
and magnetic field intensities exist in proportion to the acceleration of the
charge and the inverse radius. Where charge distributions are applicable
Eqs. 1.7.3 take the form of spatial integrals over charge bearing regions.

1.8 The Maxwell Stress Tensor

Another result of four-dimensional field analysis is the Maxwell stress
tensor. It is defined to be the symmetric, second rank 4-tensor T a p:

M<Tap - faKfKP + TSocpfvafva (1.8.1)

A symmetric 4-tensor consists of an array of ten independent numbers. It

may be shown, after some algebra, that the force density 4-vector of
Eq. 1.6.12 is related to the Maxwell stress tensor as:

9T
^x = ap/3xp (1.8.2)

The independent components of T a g follow from Eqs. 1.6.7 and 1.8.1. The
result is:
18 The Electromagnetic Origin of Quantum Theory and Light

T U =|(E2-E2- E 2) + -UB2-BJ-B2)
2^'
T eE 1
12 - . r E ^ + — ®xBy

T
22 = ^[Ey~ E
z~ E jr +
- ) 2Li( B ->'~ B ^ ~ B - r

T23 - E E ^ Z + — B^B^

T33 = | ( E 2 - E 2 - E J)+-|-(B2-B2-B5/)
2JI'
(1.8.3)
T-i
L i — EE r E „ H— B -B,
31 J
Z^J: ' "z"x

2
T
44 - ^ ( EE^ r++ EE^+
> ' + EE^^) ++ T2?L-T( (B
" ^i +' "y
B i +' B*)
"•?

T
14=~(E^-E^BJ
T
24 - " Z T ( E ^ B ^ _ E-fB-^)

T34 =
"^( E -* B >'~ E >' B -'')

The tensor may be written in the form:

-N
[T] = (1.8.4)
-N w
_c

By definition w = T44 is equal to:

T 44 = £- EE 2^ ++ _—L oB2 (1.8.5)

2 2^1
Classical Electrodynamics 19

Ty is the 3-dimensional Maxwell stress tensor:

£-|^b
r c 2 _ cby 2 _ cb 2
x x
2 eE
1
6E x Ey H BxBy x E z + — BxBz
lr> 22 D 22-BD 22]
+ r-[B
-—|B Vx^ -B
—Byy — xB x
2n (1.8.6)
ifc 2 _ n 2 _ n 2
1 2^y "Z "X
£EJ yVEY H B VyB . EEyEz + ByB;,
v- +_ L B v 2 - -z
2^1% B , 2 -"x
a2
E
•}[ z x ky
£E 7 E V H B7BV EE7EVH B7BV ,
* " ^[B. 2 -B, J -V]

N is the 3-dimensional Poynting vector:

N = (ExB)/n (1.8.7)

Symmetric tensors of rank two in three dimensions reduce from six to three
components by transforming to the principal axes and aligning one axis with
the source field intensity. For example, if there is no magnetic field and if
the electric field intensity is directed along the x-axis the tensor reduces to:

0 0
[T] = - 0 -E" 0 (1.8.8)
0 -E"

To interpret the stress tensor, consider the 4-dimensional spatial integral

of Eq. 1.8.2. The equation may be written:

Jj J K a ^ crx1! dx'2 dx'3 dx'4

dT (1.8.9)
= JIJJ c ' 0 cx^- dX, l dX '2dX'3dX'4
ax

Working with the left side:

20 The Electromagnetic Origin of Quantum Theory and Light

JJJJc'TO^dX'1dX,2dX'3dX,4
= ||||^dX'1dX'2QX,3dX'4

= ||||/S,dX1dX2dX3dX4

Working with the right side:

I I I K a ^cD^dX^dX^dX^
•P

= JJJc,<rara4dX'idx'2dX'3

The last equality results since the integral at the limits of the spatial integrals
vanish. Working with the last integral, note that
c
'cxp Tacc = c\p c ' o a c'ty T'aY (1.8.10)

Since c\p c'^y = Spy it follows that c' a pT o a =c' CTa T' a p from which
c
'oa T'a4 = c 'a4 T 0 a • This leaves the equality:

JJJJ/"^dX'1dX'2dX,3dX,4
(1 R 11)

= JJJc,a4Toa<DC1dXI2dXI3

Since c' a 4 = U a jic this may be written:

JJJJ/'^dX,1dX,2dX3dX4=ljJjT(iaUadX1dX2dX3 (1.8.12)

To change the 4-integral into a three dimensional one, differentiate by (ict)

to obtain:
Classical Electrodynamics 21

JJJ/'^dX , ,dX , 2 dX , 3 =^=--i-^jJ/T <ra U a dX 1 dX 2 dX3 (1.8.13)

Since all time integrals are zero at time t = - oo, time integration has a value
only at the present time.
To examine results of these equations, consider a charge moving with
low speed in the z-direction. With the axis in the direction of motion, the
sum TCTaUa takes the form:

T3aUa=|E2» (1.8.14)

Combining:

F
-1^-T\H $fcW. dV\ (1.8.15)

The sign was changed to represent reaction of the field on its source, rather
than vice versa. For a low speed particle undergoing differential acceleration
Eq. 1.8.15 has the form:

F <L816)
"i^'T,
The mass is calculated as:

m=^j(jE2W (1.8.17)

The interpretation accorded these equations is that Eq. 1.8.16 is Newton's

law for electromagnetic mass, confirming that F is a force. The expression
for the mass shows that (eE 12) is the energy density of an electric field.
22 The Electromagnetic Origin of Quantum Theory and Light

1.9 Kinematic Properties of Fields

To further analyze the kinematic properties of fields, begin with the four-
dimensional force equation, Eq. 1.6.14:

i^v=pE + JxB; -/c/^=E«J (1.9.1)

To express this equality in a way that depends upon the fields only, it is
necessary to substitute for p and J from the nonhomogeneous Maxwell
equations, Eq. 1.6.8:

9E
^ v = eE(V«E)-Bx —VxB-e
u\ dt
(1.9.2)
-ICJ tf =E» -VxB-£ —
[i dt

It is helpful to add zero to each equation in the form of terms proportional to

the homogeneous Maxwell equations, Eqs. 1.2.11. The added terms are:

i ( 3B
-B(V«B)-eEx VxE +
dt
(1.9.3)
and - B » | V x E + —
dt

Combining gives:

Z' v =e{E(V»E)-Ex(VxE)} + -{B(V»B)-Bx(VxB)}—y-^

(1.9.4)
/cJ$ = ' £ E 2 + _ L B 2 ' + V»N
dt 2 2u.

Writing the first of Eqs. 1.9.4 in tensor form gives:

EjEj -^SijEkEkj + ^ B j B j - i 8 i j B k B k H - - ^ - N i (1.9.5)

ox-.
Classical Electrodynamics 23

Integrating over a closed three-dimensional volume gives:

i\(m--2> ij k k5 E E
+—
V-
B B _ 5
i j r ijBkBk dS;
(1.9.6)

4 1 3Nj
c2 a/
+J?
T
"
d^

By Eq. 1.8.16 the last term on the right is the rate of change of momentum of
all charges contained within the volume, ^charge- Therefore, the first term on
the right is the rate of change of field momentum, />fleid- It follows that the
left side of the equation is equal to the force on the charges and fields within
the volume of integration. The results may be written as:

/fleld=^-J"Ndr and ^ v = pE + J x B = - A
^charge (1.9.7)
c Cu

Since F is a force density, it follows from Eq. 1.9.7 that the electric field
intensity is a force per unit charge. Since a wave travels at speed c, by the
first of Eqs. 1.9.7 the momentum passing through a planar surface is:

1f
Afield = - J N-cLS- (1.9.8)
cJ
By definition dS is a differential vector area normally outward from the
surface.
Integrating the second of Eqs. 1.9.1 and 1.9.4 over a three dimensional
volume gives:

rc 1 ^
J(E.j)dr=ij ^ E 2 + - L B 2 d^+^N-Ay (1.9.9)
2[i
V J

Since the field intensity is a force per unit charge it follows that the left side
of Eq. 1.9.9 is the rate at which energy enters the volume of integration.
Therefore the volume integral on the right side must be the rate at which
energy increases in the interior, and the surface integral must be the rate at
24 The Electromagnetic Origin of Quantum Theory and Light

which energy exits through the surface. It follows that the energy in the
electromagnetic fields is equal to:

/r = J £E2 + J_B2 dv (1.9.10)

2 2ji

It also follows that the rate at which energy exits the volume through the
surface is:

P^N-dS" (1.9.11)

A different formulation of Eq. 1.9.10 that is sometimes useful is by

rewriting it in terms of the potentials. Combining Eq. 1.9.10 with Eqs. 1.6.8
and 1.6.17 results in:

/T=J"[pO+J»A]d^+<f -e((J)E) + - ( A x B ) • dS
(1.9.12)
' BA A 3E~
+£ - E « — + A» — dV
3/ 3/

For a charge moving at a constant speed, or if the charge acceleration is

small enough so the energy escaping into the far field is negligible, only the
first term of Eq. 1.9.12 is significant. For that case the total field energy may
also be expressed as:

/T=j[pO+J.A]d^ (1.9.13)

1.10 A Lemma for Calculation of Electromagnetic Fields

A lemma is needed to assist in the unrestricted and systematic calculation of
electromagnetic fields about known sources. To obtain it begin with the
general form for fields in a source-free region containing time-dependent
fields:

3E 3B
V x B - eu — = 0 = Vx E+eu—- (1.10.1)
P
3/ *d/
Classical Electrodynamics 25

Taking the curl of Eq. 1.10.1 and substituting as needed gives

32B 32E
Vx(VxB) + E ^ - ^ - = 0 = Vx(VxE)+e^i—T (1.10.2)
dr dr
This shows that, away from sources, E and B satisfy the same partial
differential equation.

V2V-e\id2y¥/dr =0 (1.10.3)

This is useful because of an associated lemma that begins with the vector
field F(r,t), defined by

F = Vx(/*F) (1.10.4)

The lemma is that if *F satisfies Eq. 1.10.3 then F satisfies the differential
equation:

Vx(VxF)+en32F/a/2 = 0 (1.10.5)

To verify that Eq. 1.10.5 is correct, multiply Eq. 1.10.3 by (-r) then take
the curl:

/ \ a2
2
-Vx(/V ¥j + en-2-[Vx(/F)] = 0 (1.10.6)

Comparing Eqs. 1.10.4 through 1.10.6 shows that Eq. 1.10.5 is satisfied if:

V x {V x[Vx (/¥)]} = - V X ( / V 2 X F ) (1.10.7)

To confirm Eq. 1.10.7, begin with the identity for the curl of a scalar-vector
product:

Vx(^¥) = W(Vxr)-rxS/x¥ (1.10.8)

26 The Electromagnetic Origin of Quantum Theory and Light

Since V x r = 0, it follows that:

Vx[Vx(/P)] = - V x ( / - x W ) (1.10.9)

Combining Eqs. 1.10.7 and 1.10.9 gives:

VX[VX(/-XVVP)]-VX(/V2XF) = 0 (1.10.10)

Two identities from vector analysis are:

V(A»B) = A x ( V x B ) + B x ( V x A ) + (B»V)A + (A»V)B

(1.10.11)
V x ( A x B ) s A ( V » B ) - B ( V » A ) + (B«V)A-(A»V)B

Putting A = r and B = Y¥:

V(/-» VY) = (/•• V)V*F + (V*F • V ) r = ( r » V)VXF + W

(1.10.12)
2
V x (v x V ¥ ) = / V T - 2W + (/•• V) W

Combining Eqs. 1.10.10 and 1.10.12:

V x ( / - x V T ) - / V 2 ¥ + VT + V(/-.VY) = 0 (1.10.13)

Since the curl of the gradient vanishes, taking the curl of Eq. 1.10.13 yields
Eq. 1.10.10 and completes the proof.

1.11 The Scalar Differential Equation

To solve Eq. 1.10.3 it is useful to remove the time-dependent portion. For
that purpose use the Fourier integral expansion:
Classical Electrodynamics 27

oo
x
V{r,f)= j v(r,(0)e*,/dfi) (1.11.1)
—co

Substituting Eq. 1.11.1 into Eq. 1.10.3 leads to:

J (V2\|/ + A:2vt;)e/to/dco = 0 (1.11.2)

—oo

By definition k2 = co2E|X. For this equation to be zero for all values of co, the
integrand of Eq. 1.11.2 must equal zero:

V2y + Jt2\\i = 0 (1.11.3)

This is the Helmholtz equation, solutions of which combine with Eqs. 1.10.3
to 1.10.5 to obtain the full solution for vector fields.
Certain helpful vector operations in spherical coordinates are listed in
Table 1.11.1. Using spherical coordinates with 9 the polar angle from the z-
axis, <>| the azimuth angle from the x-axis, and r the radial distance from the
origin, by Table 1.11.1 the Helmholtz equation is given by:

sin 6-— 3 > J__3_ + k2\!f = 0 (1.11.4)

2
r sin0 30 + r2 sin2 9 3<b2 + r.2 dr dr

Dividing the equation by k shows that the radial dependence of the solution
is a function of only the product a = kr,and therefore \|/ may be written as
\|/(a,9,<t>). A theorem applicable to problems using spherical coordinates is
that the complete solution of Eq. 1.11.4 is obtained by summing over all
possible functions iy(a,9,<))) where:
28 The Electromagnetic Origin of Quantum Theory and Light

Orthogonal Line Elements: dr,rdB, /-sin8 d(j)

Divergence of Vector A:
96 '• /-sine 9())

Components of Curl A: 1 a^ineA^,) a A e

(VXA) =
/-sinG 36 3<(i

(VxA) _ _ i _ ^ _ I ^ K )
V ;e
/-sinG 3<|> r dr
3(/-A9) 3A,.
( V * A )* = dr ae

Laplacian of W = V l : iar^ar]+ i af. .a^v I afvl

/-3/-V 3/-J /-2 sin 6 391 sinB-r—
39 Jl + -r 2 sin2 6 3<t>2

Table 1.11.1. Vector operations, spherical coordinates

v(/-,e,<l» = R(a)e(e)<&(<i» (1.11.5)

To obtain \|/(a,0,<|)), it is necessary to begin by solving for the solutions of

Eq. 1.11.5 that involve only one independent variable. After obtaining the
functional forms, all possible products are formed and weighted by a
constant multiplying coefficient. The coefficient is determined by matching
boundary conditions. Finally, all individual product functions with
appropriate coefficients are summed.
Substituting Eq. 1.11.5 into Eq. 1.11.4 and multiplying by r2 gives:

1 d ( . ade\ 1 d24> 1 d ( 2 d j O 2 A
sm6— + s- + —— a — + o z = 0 (1.11.6)
GsinGdeV d0j O d<))2 Rda{ da J

The first two terms are independent of the radius and the last two terms are
independent of the angles, yet the two sets equal each other's negative,
requiring both sets to be constant. The constant is known as the separation
constant. A convenient choice of separation constant is for the radial terms to
equal v(v+l) and the angular terms -v(v+l), and results in the separated,
complete differential equations:
Classical Electrodynamics 29

_L_d_f 2 dR v(v+l)
R=0 (1.11.7)
2
+ 1
rj day do a2 ;

<D d 8 dz<&
sin 6 — (1.11.8)
sinG d9 d0 + —x2 r- + v(v + 1 ) 6 0 = 0
sin 0d(|) 2
The radial equation is a differential equation with one independent variable.
The angular equation may be written as:

sin 8 d d0 1 d2Q
sin6 + v(v+l)sin 2 0+ =0 (1.11.9)
0 d9 d0 O d(|)2

The first two terms of Eq. 1.11.9 are functions of 6 only and the third is a
function of (j> only, yet the terms equal each other's negative. Again, both
sets are constant. Putting the first two terms equal to m2, where m is the
second separation constant, results in two separated equations, each
involving only one independent variable:

1 d(. Qd6. ( m
\
sin6— + v(v+l)- 0 =0 (1.11.10)
sin 2 6
sinOde^ d0

d2<D
+ *ro=o (l.n.ii)

Solutions of the separated differential equations and tabulated functions are

in the Appendix.
Solutions of the radial equation are spherical Bessel, Neumann, and
Hankel functions, respectively, j v (a), y v (a), and h v (a). A particularly
important linear combination is Hankel functions of the second kind and
integer order: h^(a) where "f' represents any integer value of "v". Solutions
of the zenith angle equation are associated Legendre functions; solutions are,
in some instances, of integer order and in others of noninteger order. In all
cases, the orders of the radial and zenith angle solutions are the same.
Trigonometric functions form the solutions of the azimuth angle equation:
30 The Electromagnetic Origin of Quantum Theory and Light

sin<|), cos<|), and exp(±/w<|>). Since all solutions to be considered extend over
the full range of azimuth angle, zero through 2iz, only integer values of
degree m, are present. With exponential notation, the exponent may have
either sign. With symbol z v (a) representing a linear combination of possible
radial solution forms, rather than writing the solution as two separate sums it
is rewritten as:

<(/-,0,(|)) = z v ( 0 ) 0 ^ ( e ) e - ^ (1.11.12)

With this notation, completeness requires m to include the full set of positive
and negative integers, however the degree of the Legendre function is
always positive.

1.12 Radiation Fields in Spherical Coordinates

Replacing B by |iH more closely matches common usage. For what lies
ahead we are concerned only with free space and there \i is merely a unit-
determining parameter that measures the magnetic field in amperes per
meter instead of webers per square meter.
The field calculation procedure is due to Hansen, and begins with the
vector theorem that a field with zero divergence is completely specified by
its curl. It is, therefore, helpful to introduce the two independent field sets:

TiH1 = /-xVxF1 and E2 = rxVx¥2 (1.12.1)

Since the free space divergence of both vectors are zero, solutions of
Eqs. 1.12.1 provide the complete set of possible values for vectors Hi and
E2. The remaining field solutions, H2 and Ei, may be obtained from
Eqs. 1.12.1 using the Maxwell curl equations. The total fields, (E1+E2) and
(H1+H2), are then complete. If the boundary conditions are matched, the
fields are also unique.
In what follows we use the notation that time dependence is exp(z*(ot) and
2 -2
azimuth angle dependence is exp(-jm§), where i =j = - 1 . The reasons for
separate notation are that it permits separation of polarization and time
dependencies and it permits restriction of separation constant m to the field
Classical Electrodynamics 31

of positive integers, without loss of generality. With Hansen's method the

defining terms for phasor fields are, see Eq. 1.10.4:

TIHJ = /-xVv|/j and E 2 = / , x V f 2 (1.12.2)

A tilde over a vector indicates that it is a phasor. It is required that the scalar
functions satisfy the Helmholtz equation, Eq. 1.11.3. For integer modes, the
results are solutions in the form of Eq. 1.11.12:

(1.12.3)
V2 = / 5 ( ^ » ) z / ( a ) e f e - ^

The order is not restricted to integer values and the radial function z((c) may
be any linear combination of spherical Bessel and Neumann functions. The
zenith angle function may be any linear combination of associated Legendre
functions. Both the applicable functions and the constant multiplying
coefficients F(^,m) and G(J.,m) are determined by the boundary conditions.
Applying the operation of Eq. 1.12.2 to Eqs. 1.12.3 gives the result:

/•xVw = -^- + 4> ur- (1.12.4)

sine dty 39

Combining gives:

~m®f 1dOf -Jm§

r\Hl=F(lm)ze(a)
sin6 d0
(1.12.5)
t2 = jG{£,m)ze(c) ,->*)>
J Y
sin0 d0

Taking the curl of the second of Eqs. 1.12.5 then applying the Maxwell curl
equation leads to:

( d0
nH 2 = -Kj(t,my<** \j£{£ + l ) ^ 0 f ; + z\
de
j- * e +sin0^V (1.12.6)
V J
32 The Electromagnetic Origin of Quantum Theory and Light

Carat "A" indicates a unit vector, and:

Z
H°)=-^K(°)] (1.12.7)

Taking the curl of the first of Eqs. 1.12.5 then applying the Maxwell curl
equation leads to:

d 9 f A .rri&m ^
tx = iY(t,m)frjm* £{t + l)^Qf?+z'£
d0 e-/^<!>
J (1.12.8)
sinG
J

The total fields are the sum of Eqs. 1.12.5, 1.12.6 and 1.12.8. They may be
written as:

(=0 m=0
z l
*lHr=-//£ £ rrt-r-Ji,
^ , ^ „\i,li>
) ^ +, i l\ )A^5^?e) fQ( c^ „o„s„0Q L) e- y " < t
C
^=0 m=0

X(l,„)Zt — -G(t,„)Zzr-JL ->«t>

Y{£,m)i del fc{£,m)z\ ->*!>

d0 sin 9
(1.12.9)
oo ^

r def ->*>
E0 = -y'E E sinG
?
"dF
£=0m=0

V[£,m)Z( — - KJ(£, m)Zf —— X-Jmty

v l v
/=0«=0 ' sinG ' ' d6

Without loss of generality, the phases of constants F(£,m) and G(£,m) and the
multiplying factor re have been picked for later convenience. Coefficients
F(£,m) multiply the radial component of the electric field terms and are TM
(transverse magnetic) fields and modes, where "T" indicates transverse to
Classical Electrodynamics 33

the radial direction. Coefficients G(£,m) multiply the radial component of the
magnetic field and are TE (transverse electric) fields and modes. Terms with
I = m = 0 have no radial fields and are the TEM (transverse electric and
magnetic) fields and mode. This result is valid for all possible
electromagnetic field solutions.
Keeping only the real or only the imaginary part with respect to " / '
provides, respectively, x or y polarization of the electric field intensity. The
fields are right or left circularly polarized, respectively, with j = i orj = - i.
Since this result is applicable to all time-dependent outgoing waves, it
follows that it also applies when the rate of change is arbitrarily small.
Hence, it describes fields in the limit as the frequency goes to zero, a static
charge distribution. Because of this general result, it is helpful to obtain a
physical view of what constitutes field sources. The sources of coefficients
¥{t,m) and G{(.,m) for static fields are discussed in the appendix,
Sections A.28 and A.29.
Consider a few special cases of Eqs. 1.12.9. If the described fields are
contained within a source-free region of space, and if that space is loss free,
solutions have positive, integer values of orders and integer values of
degrees. Spherical Bessel functions, which have no singularities, form the
radial portion of the solution; spherical Neumann functions, which have
singularities, are not present. Associated Legendre functions of the first kind,
and of integer order, which have no singularities, form the angular portion of
the solution; fractional order associated Legendre functions and those of the
second kind, which have singularities, are not present.
In the main, if the fields originate at a point and support an outward flow
of energy from that point, the radial portion of the solution consists of
spherical Hankel functions of the second kind. A solution within an enclosed
space that excludes the z-axis, but has rotational symmetry, is described by
associated Legendre functions of both the first and second kind, with
noninteger, positive-real orders and integer degrees.
In all cases, if the medium in which the fields exist is lossy, the
separation constants are complex numbers with a positive real part. Since all
cases of interest in this book concern lossless medium and a full 2n spatial
rotation about the z-axis, both the order and degree are real and degrees have
only integer values.
34 The Electromagnetic Origin of Quantum Theory and Light

1.13 Electromagnetic Fields in a Box

To analyze a later radiation problem, it is helpful to know the number of
independent field solutions that can exist within a prescribed volume. For
example, consider all possible electromagnetic field modes that can exist
inside an otherwise empty, rectangular cavity confined by walls of infinite
conductivity. With sinusoidal field excitation, the Maxwell equations have
the form:

TlH = - V x E E = --VX(T]H) (1.13.1)

' k

Assign axial directions to each side of the box; the cavity length in the JC-
direction is a, in the j-direction is b, and in the z-direction is d.
Application of conducting boundary conditions requires the z-component
of the electric field intensity to equal zero at x = 0 and a, and at y = 0 and b.
If both the electric and magnetic fields at z = 0 and d are equal to zero the
result is the trivial one of no fields at all. Application of the boundary
conditions also requires the z-component of the magnetic field intensity to
equal zero at z = 0 and d and to have zero slope at the x- and >>-directed
boundaries. The result is that the most general forms of normalized field
solutions have the z-components:

r\Hz(x,yz,/)
flnx)
= cos — cos
r nm:y
\ sinf /z7rz^*)/
aJ
v b J d )
(1.13.2)
( /OTZ
Ez{x,y,z,t) = sm irmy .A»/
sin cos
a \
Symbols £, m, and n indicate positive integers.
For the case ofEz(x,y,z,t) = 0, the complete set of remaining fields is:

r\Hz(x,y,z,t) = cod cos —-^ I sinf —— sin((o/)

(
rmy\ ( miz
x^x{x,y,z,t) = - 1 fin mi sin (.nx cos cosl
%y«j
v«) b ) \ d .
rrni frn^ (inx\ . (ntKy\ (mtz^
nH V{x,y,z,t) = --J, cos sin(co/)
k2K b K " J cos V b ) *
/
Classical Electrodynamics 35

/ k ( rmi\
E {x y ZJ)=
x
cos( inx\ s i. n(/my)s l I. (raiz\ , \
cosH
'' l?\j) l—) m t ~7)
f
in) . (lux" itmy^ rmz
Ey(x,y,z,t) =—^ - sin cos sin cos (to/)
a ) V a j V o )

In Eq. 1.13.3 the definition k= (o/c is used, and:

2
/«*,>> /OT
*M^i rtm (1.13.4)
+ +
a
K
in) :
(naO
c + (1.13.5)
T
The field energy is given by:

/F=-j(eE2 + uH 2 W (1.13.6)

Substituting Eqs. 1.13.3 into 1.13.6 gives the result:

In) (mi) fmn} ( mi^

abd bJ (1.13.7)
W=z 1+ a
16
+ rmi
Tj l b

Evaluating the energy of solutions with Hz(x,y,z,t) = 0 yields dual solutions

and the same energy form.
Of interest with a problem to come is the number of possible separate
solutions for electromagnetic fields in a box of dimensions a,b,d. Unit
lengths along the x-, y- and z-axes in £-space are, respectively, nla, nib and
n/d. Since each point in the upper right quadrant of £-space corresponds to a
separate solution to each of Eqs. 1.13.2 the number of separate solutions is
just twice the volume of the upper right quadrant. If the box is very large the
line spacing is vanishingly small.
36 The Electromagnetic Origin of Quantum Theory and Light

References
R. Becker, Electromagnetic Fields and Interactions, Blaisdell Publishing Co. and
Blackie and Son (1964), reprinted by Dover Publications (1982)
W.W. Hansen, "A New Type of Expansion in Radiation Problems," Phys. Rev., vol.
47,pp.l39-143(1935)
J.D. Jackson, Classical Electrodynamics, 2nd ed., John Wiley (1975)
L.D. Landau, E.M. Lifshitz, The Classical Theory of Fields, trans, by H.
Hamermesh, Addison-Wesley (1951)
W.K.H. Panofsky, M. Phillips, Classical Electricity and Magnetism, 2 nd ed.,
Addison-Wesley (1961)
A. Sommerfeld, Electrodynamics, Academic Press (1952)
J.A. Stratton, Electromagnetic Theory, McGraw-Hill (1941)
J.B. Westgard, Electrodynamics: A Concise Introduction, Springer-Verlag (1997)
 2. Selected Boundary Value Problems

2.1 Traveling Waves

It is helpful in many ways to have an expression for a plane wave although,
since they extend infinitely in two directions normal to the direction of
propagation, they do not exist. Plane waves are used to describe the
interaction of a wave with an object much smaller than the region over
which the wave approximates a plane wave. Spherical waves do exist and so
long as the radius of curvature of the sphere is much larger than other
dimensions of interest a plane wave analysis is justified. The criterion is
simply that the non-planar nature of the wave be negligible at the problem of
interest.
The electric and magnetic fields of a unit magnitude, x-polarized, z-
directed plane wave expressed in rectangular coordinates are:

E = ie-/fe and TiH = > T / f e (2.1.1)

The same fields expressed in spherical coordinates are:

E = tTto cosG {sin9cos(|)A+cos9cos(|)0 - s i n # }

(2.1.2)
r]H = e - ^ c o s 6 {sin6sin(|)£ + sinGsin^G+cos#}

The fields of Eqs. 1.12.9 may, of course, be used to describe plane waves; it
is only necessary to obtain appropriate values for the coefficients F(£,m) and
G(£,m), and to evaluate the different mathematical functions. To do so it is
most convenient to work with the radial field components only. The radial
component of the electric field intensity of Eqs. 1.12.9 is:

37
38 The Electromagnetic Origin of Quantum Theory and Light

Er = ' " £ X ^ M A ^ + 1 ) —effcoseK^ (2.1.3)

Since a plane wave has no singularities neither does the radial function of
Eq. 2.1.3; it follows that only spherical Bessel functions form part of the
solution, with z((a) replaced by j^(a). Since the wave occupies all values of
azimuth angles (j), degree m must be an integer. Since the z-axes are included
in the solution only integer order associated Legendre functions of the first
kind, P/^cosB), are present. Applying these conditions and equating
Eq. 2.1.3 with the radial component of Eq. 2.1.2 gives:

sin9cos(t)e-* cose = / £ £ reF{t,m)l(t + l ) ^ P f (coseje-^

(2.1.4)

The azimuth dependence of Eq. 2.1.4 shows that only coefficients of degree
one, F(^,l), are different from zero, as are all imaginary parts with respect to
"/'. This leaves the equality:

e -*cos9 = i y re¥(tMt + l)}e(c)^!!l (2.1.5)

a^Tj sinG

Another expansion for the exponential is listed in the appendix,

Table A.27.1.2:

e -*cos0 = £ y fl,2i+ jx. (o) WfeosGj

a^Tj sinG

Equating the two expressions shows that the coefficient is:

F(*,l) = -^ ( (2.1.7)

Entering these results into Eq. 2.1.3 gives:

Selected Boundary Value Problems 39

°° i (a)
E r = / £ re(2£+ l)^-^P}(cosB)cos^ (2.1.8)
e=i
Working with TE modes in a similar way results in the equalities:

oc.i)=-^4 (2.1.9)

a
r\HT = / £ f\2i M )r>l<
+ l)^^P^(cos0)sin4» (2.1.10)

The angular field components follow from the radial components and the
formofEqs. 1.12.9:

^ ( 2 1 + 1)
JA yd e J A y s i n 9 cos (J)

,dPl
= JA / d e J A y s i n 9 COS(j)
^ , ? / <(/+!)
£(l+l)
(2.1.11)
w (2l+l)
JA ; d e J A ; s i n 0 sin<j)
(/ + 1)
oo
wfo + l)
JA i d e JA 7 s i n 0 sin())
*=1 •1)

Equations 2.1.8, 2.1.10, and 2.1.11 are the electric and magnetic fields of a
unit magnitude, x-polarized, z-directed plane wave expressed in spherical
coordinates using spherical functions.
40 The Electromagnetic Origin of Quantum Theory and Light

Scattering

2.2 Scattering of a Plane Wave by a Sphere

A spherical object of radius a is immersed in the plane wave described by
Eqs. 2.1.8, 2.1.10, and 2.1.11. Analysis of the interaction between the sphere
and the plane wave is done by separating the procedure into steps. In the first
step energy and momentum is extracted from the wave and applied to the
sphere; these are extinction values of energy and momentum. In the second
step the extinction values separate into parts. The scatterer permanently
retains the absorbed energy and the scattered energy goes back into space.
To determine the scattered fields everywhere, note that all possible fields are
expressible in the form of Eqs. 1.12.9. Problem solution is simplified if three
characteristics of the scattered field are noted. First, since scattered fields
exist on the z-axis and since only associated Legendre polynomials of integer
order converge on that axis the zenith angle dependence varies as associated
Legendre polynomials of integer order. Second, the total scattered power is
constant in the limit of infinite radius and constant power requires outgoing
fields to vary with distance as e~ ; °/a. Only spherical Hankel functions of the
second kind have the needed limiting form and satisfy the spherical Bessel
differential equation. Third, the scattered field possesses only the symmetries
of the scatterer and the input field. Therefore, only fields of degree one are
present.
To solve for the magnitudes and phases of the scattered modes it is
convenient to multiply each set of modes by sets of complex constant
coefficients: a£ for TE modes and (3^ for TM modes. With this definition the
radial components of the scattered fields are:

Er = / £ r£{2£+i)MiMp](cose)COs(t)
(2.2.1)

TiHr = / £ r<(2€+l)^H>](cose)sin<|>
a
e=\

Problem solution requires evaluation of each value of <Xg and |3^. With
"•" defined by Eq. 1.12.7, the sum of plane and scattered fields is:
Selected Boundary Value Problems 41

a E r = / £ /^'(2€ + lXJ/ + P/h^ )PJ (cos6)cos<|»

i=\

crnHr = / £ r^(2^ + l)(j£ + a^)p/(cos0)sin(|)

rt 2^ + 1
Ee-S \n vn t> d e u/ / ^ sin0
COS(|>
£i *(*+!) (2.2.2)
ri 2l + \ dPi . , p;
*!&•=£ (j/+P/M-jH"+<j5+«^)- 0
COS(|)

d9 sin 9

d9 sin0 sin(|)

- _ ^ ^ 2^+1 jpl pi
^ft + cc/ru)—^- + ( j / + B,h / ) —l— sin(p

By Poynting's theorem the time-average power, P a v , on a spherical,

virtual surface of radius elk circumscribing the scatterer is equal to the real
part of the surface integral of the radial component of complex Poynting
vector N c r = (ExH) ll, see Eq. A. 11.5:

2 27i n
Pav = - ^ y J d(t>Jsined9Re(Ncr) (2.2.3)
^ o o

Using Eqs. 2.2.2 to evaluate the radial component, and "*" indicating
complex conjugate, the complex Poynting vector is:

^t((2l + l)Y (2«+l)

Ncr=Re 2•S 1 +1
2ti^ ,r, ~ U' )JU"+0
42 The Electromagnetic Origin of Quantum Theory and Light

/. «*, *V.. r,,.\fdP)dPi 2 P)PI 2

J* + P A , J? + P ^ ' J ^ f - ^ cos2 <J> + - ^ s i n 2 <I>
)
\ A ly de de Sin
2
e J
., *. # * \ | dP« dP„ j P<>P„ ?
2
d9 * sin 9 y_
(2.2.4)
*,*\| dPl Pi . 2 P] dP' 2 ^
d6 sine sine d6
+
»*\(.» n ,.\^ dPJ P' 2 P/ dPi . 2 ^
+(j>a*X*)(j;
": +-•-«Mu ' " —L—— cos24> + —^ ^sin <>
(
d6 sine sine d6
Inserting Eq. 2.2.4 into the integral of Eq. 2.2.3 and integrating over the
azimuth angle gives:

no2 ^, ^, n _ / ( 2 l + 7)Y(2/y+l)
Ref^S
2n/f
2. (e+i)){*(x+\)
<=1 n=\

' [(j* + PX)(j* + MJ) - (it + a A)(j* + °X*)]

dP|dPJ,+2*Pi (2.2.5)
Pav=Jsin0d9 de de 2
Sin e
+[(j*+P«I»;)(£+«X*)+0<+«^)(j«+PX)]
1 d(pjpj)
sine d6

Using integrals in Table A.22.1.3 andA.22.1.6 to evaluate the integrals of

Eq. 2.2.5 gives:

(2.2.6)

In the limit as the radius becomes many times larger than either radius a or
wavelength X, Eq. 2.2.6 simplifies to:
Selected Boundary Value Problems 43

CX) .

Tl*" /=1 (2.2.7)

Energy and momentum are transported into the system on the plane
wave; both are transferred to the scatterer. The input power is equal to the
term proportional to Re(oc^ +(3^). The total power first extracted from the
beam is defined as extinction power, and is always positive. Changing the
sign to conform with this usage, the extinction power is:
oo

PEX—^ReX^ + l X a ^ )
W t=\ (2.2.8)

The power scattered back into the field is equal to:

W (=1 (2.2.9)

Absorbed power, the negative of Eq. 2.2.7, does not reappear in the field but
may be calculated by subtracting the scattered power from the extinction
power. Lossless scatterers have no absorbed power and, therefore, for them
Eq. 2.2.7 is equal to zero.
The critical scattering parameters are commonly normalized to a value
that is independent of the magnitude of the plane wave. Define scattering
cross section, C s c , to equal the scattered power-to-incoming power density
ratio. With a unit magnitude electric field intensity the incoming power
density is l/(2r\), see Eqs. 2.1.1 and A. 11.5. Values are sometimes also
normalized with respect to the geometric cross section. Cross section has the
dimensions of an area, and normalization with respect to the geometric cross
sectional area gives a measure of size the scatterer appears to be versus the
size it would appear with zero wavelength optics. Define geometric cross
section, CQE. to be the area the scatterer presents to the plane wave. For
example, the geometric cross sectional area of a spherical scatterer of radius
2
a is CQE = Ka • Combining the definition with Eqs. 2.1.11 and Eq. 2.12.9
shows the scattering-to-geometric cross section ratio to be:
44 The Electromagnetic Origin of Quantum Theory and Light

oo
'SC £(2* + l)[a,a;+p$] (2.2.10)
CGE k2a2 £_

Similarly, the extinction cross section, CEX> is defined to equal the extinction
power-to-incoming power density ratio. Combining the definition with
Eqs. 2.1.11 and Eq. 2.2.8 shows the extinction-to-geometric cross section
ratio to be:

'EX 2
2 2
£ ( 2 * + l)Re(a,+P,)
<-GE k a i=x (2.2.11)

A third cross section that is often of interest is radar cross section. Define the
radar cross section, CRCS> to equal the quotient of the power that would be
scattered if the power density were everywhere equal to its value at 9 = rc
divided by the incoming power density. It is a measure of the power returned
towards a single interrogating radar antenna. By definition, the power
scattered in direction 9 = n is the back-scattered power. To determine the
radar cross section, evaluate Eq. 2.2.4 at 9 = n. The angular functions at that
angle are equal to:

dPi P/1
=^+l)(-l/ (2.2.12)
d9 sin9 2

Carrying out the calculation then normalizing by both the incoming power
density and the geometric cross section results in the normalized radar cross
section:

— - ^ E S(2^l)(2«+l)(-l)^U(£-/?)f(a,-P,)(a;-p;)
-GE k a f-= 1 n=\
(2.2.13)

Function U(£-n) is the step function:

Selected Boundary Value Problems 45

1 £>n
U(*-n) = 1/2 £ = n (2.2.14)
0 £<n

As shown by Eq. 1.9.8 the fields carry momentum as well as energy, and
momentum transfer from the field to the scatterer constitutes an applied
force. The momentum transferred to the scatterer by the extinction energy is
in the direction of the incoming wave and, by Eq. 1.9.8, is equal to the
energy divided by c. The scattered power transfers momentum in proportion
to the cosine of the angle between the incident and scattering directions. The
back-scattered and forward-scattered portions of the power produce
momentum respectively into or away from the direction of the beam. The
sign of the total transferred momentum depends upon which type dominates,
and that depends upon details of the specific scatterer. Although the resulting
force is too small to be significant in most macro-scale applications,
nonetheless it exists and affects all scatterers and receiving antennas.
It is also possible to calculate the force on a scattterer because of the
scattered field, Fsc- The physical origin of the force is that the surface
currents move, at least partially in phase with the incident field, in the x-
direction. The incident magnetic field intensity is j^-directed. It interacts with
the jc-directed current density to form a z-directed force. However, this is not
the way to calculate the force. For purposes of calculation, note that the
momentum density is directly proportional to the power density, differing
only by a factor of c. The force is most easily calculated by taking the z-
component of the scattered power, which is equal to the integral of the
product of the Poynting vector and the cosine of the scattering angle. With
the help of Eq. 1.9.7 the expression for the force in the direction of the plane
wave is:

2 2nn
^C = 2 J Jsin0cosed0Re(N r ) (2.2.15)
2ck Q 0

Substituting the scattered fields of Eqs. 2.2.1 into Eq. 2.2.15 and integrating
46 The Electromagnetic Origin of Quantum Theory and Light

oo oo
TOO n-l (2^ + 1) Y (2* + l)
ReES
2kL ^ + l)A «(/? + !)
dPJ dP* | P>^
F$c = -J sin6d9 COS0 .(2.2.16)
2
de de sin e

1 d(p>j)
+[a/„h^h>a*^h;h;*] COS0
sine de

Inserting the integrals of Table A.22.1.4 and A.22.1.7 into Eq. 2.2.16 gives:

TOCT
„ •£> ( £{£ + 2) Y * ,* i • r. r>* , , .* 1
—— R e.L I "J-—pr [ a ^ + i ^ + i h ^ - p A + 1 h ^ h f + 1 J

^sc - 2
[PA-l h *-l h * - o ^ a ^ h ^ h ^ J
^
2
(2^ + 1)

(2.2.17)

In the far field Eq. 2.2.17 goes to

+ Z
¥ ( * * a a* a*a \
•h^(alae+l+atat+l+$fie+l+Vfit+l)
• (2.2.18)
(2i+l)( Q* *a \
/(m)(a^ + a
^)

Using Eq. 2.2.11, the force due to reception of the extinction power, the
extinction force, FEX is m the direction of the incoming field. Normalizing
F E X by the incoming power density determines the normalized force,/EX-
Normalizing it by the geometric cross section gives:

(2.2.19)
<-GE ck a £=x
Selected Boundary Value Problems 47

Summing Eq. 2.2.18 and Eq. 2.2.19 gives the normalized total force on the
scatterer:

(2* + l)Re(<x / +P / )
(TSC + TEX). i(l + 2)l * * n a* a*a \
,2 2 £*>
C
GE ck a £-\
(2£ + l)
+
^Tlj( a ^ +0C ^)
(2.2.20)

Although the energy absorbed by a lossless scatterer is zero the momentum

transferred is not. Even lossless scatterers are accelerated in the direction of
an incoming plane wave. Although the effect is small enough so that the
effect of sunlight on atmospheric molecules is less significant than normal
thermal unbalance, in other cases it can be significant. For example, the
impulse electromagnetic wave produced by a nuclear blast results in forces
of major significance.

2.3 Ideal Spherical Scatterers

The form of electromagnetic fields produced by a spherical scatterer
immersed in a plane wave depends upon the symmetry of the scatterer, and
the scattering coefficients ocn and p n depend upon its electromagnetic
characteristics. The special case of a perfectly conducting sphere is
important since it approximates many natural objects, it is convenient to
analyze, and yet it demonstrates a full range of solution characteristics. The
solution procedure is to apply the boundary condition that the tangential
component of the total field intensities are equal on either side of the r = a
boundary, see Section A. 12. On the exterior this is the sum of the incident
plane and scattered waves; with an ideally conducting scatterer the tangential
component of the interior field is zero. Therefore, the following sums are
equal to zero:
48 T^e Electromagnetic Origin of Quantum Theory and Light

"E e (a,9,(j))dP] E0(g,9,(|>) p] '

s cos(|) d0 sin(j) sinG
=0
(2.3.1)
E^(g,9,(|)) <fl>]N
OO
E9(g,9,(|)) P^
s cos(j) sinG sin0 d9
=0
«=1

Use of Eq. 2.2.2 shows that Eqs. 2.3.1 may be expressed as:

/[j*(/fo) + |3^h* (/&•)] dp]dpi PJPI

2
d9 d9 S in 9
ywfe + O =0 (2.3.2)
1 d(pjP])
sin6 d9

dPJ dPJ | P/P]

[h{*a)+aehe(£a)]
(21 + 1) de de sin2e
2r t(t + l) 1
=0 (2.3.3)

sin 9 d9

Next, form the integrals:

Ee(a,9,<|>) dP] E4((a,9,<|)) p]

J"sin0d9X cos(j) d9 sincf) sin0
=0
0 n=\
(2.3.4)
71 OO
E9(q,9,<|>) P* E^(q,9,4>) dp]"
J*sin9d9£
n=\
cos(|) sin 9 sin(|) d9

Inserting the needed integral forms from Table A.22.1 into Eq. 2.3.4 gives,
after simplifying:

•&(ka) + $^(ka) =Q
(2.3.5)
]^ka) + OLfh^ka) = 0
Selected Boundary Value Problems 49

Solving for the coefficients:

An important special case is a scatterer with a small radius-to-

wavelength ratio; for this case incorporating the values of the spherical
Bessel functions in the limit as ka => 0 gives:

, v (fe)/3 i{kaf
aAka) = -—'- T= \
3
(ka)l3+ilUaY
V V ;
' (2.3.7)

U
' 2ll-il(taf 3/

The cross sections and normalized forces are:

CEx C S C _ c/EX 10(/fo)4

C C C 3
GE GE GE , . , , ,
\4
^C^ = 9 ( / f a 7 )4. ^SC=4(M.
C J
CGE GE

As always, c represents the speed of light.

Figure 2.3.1 shows the normalized extinction cross section for a
conducting scatterer as a function of ka, for scatterers of any physical size.
Since the incoming power is fully directed, Fig. 2.3.1 also shows the
momentum transferred to the scatterer; Eq. 2.3.8 shows that for small
scatterers it varies as (ka) . The largest normalized extinction cross section
occurs at ka = 1.2, and is equal to 2.28. For larger values of ka the total cross
section oscillates towards a limit of twice the geometric cross section, in the
limit of infinite radius.
In the lossless case the extinction and scattering cross sections are equal
and the total force on the illuminated sphere is the extinction plus scattered
forces, (FEX+FSC)- The force on the scatterer because of the scattered field
is shown in Fig. 2.3.2. Electrically small objects scatter predominantly back
50 The Electromagnetic Origin of Quantum Theory and Light

into the direction from which the wave came, increasing the thrust in the
direction of the wave. Electrically large objects scatter predominantly in the
direction of the incoming wave, decreasing the thrust on the scatterer. The
sign of the scattering force changes at about ka = 1.38. The largest forward
magnitude is about 0.257 and occurs at ka= 1.12.
The extinction momentum is in the direction of the incoming wave. All
interacting energy forms part of the extinction momentum but, upon re-
radiation, it may either add or subtract momentum from the scatterer. Since
the subtracted momentum cannot exceed the extinction momentum, it
follows that:

Absorbed Energy
• < c (2.3.9)
Absorbed Momentum

Scatterers are commonly divided into groupings that depend upon the
radius-to-wavelength ratio. The Rayleigh region is over frequencies for
which ka« 1, the Mie region is over frequencies for which ka is on the
order of one, and the optical region is over frequencies for which ka » 1.

2.5

2 -

1.5

C
GE
1 —

0.5 \-

Figure 2.3.1 Extinction Cross Section Versus ka for a Conducting Sphere of

Radius a
Selected Boundary Value Problems 51

0.4

0.2
r
sc 0
-GE
-0.2

-0.4

-0.6

-0.8

Figure 2.3.2 Normalized Scattering Force Versus ka for a Conducting Sphere of

Radius a

3.5

2.5 —
RCS
2 —
-GE
1.5

0.5

Figure 2.3.3 Radar Cross Section Versus ka for a Conducting Sphere of Radius a

The normalized radar cross section is shown in Fig. 2.3.3. Since

conducting spheres that are the right size to be held in a person's hand have
radar cross sections that are convenient to measure, the curve of Fig. 2.3.3 is
often used as a laboratory calibration standard.
52 The Electromagnetic Origin of Quantum Theory and Light

Optical scattering results determine the optical properties of the sky. A

clear atmosphere of gaseous nitrogen and oxygen, without suspended
particulate matter, scatters a portion of the light that passes through it. Since
the molecules are much smaller than a wavelength of visible light, the
scattering process selectively acts more on the shorter wavelengths than
longer ones and more blue than red light is scattered. When the sun is
directly overhead the sky away from the directly incoming beam is
illuminated by scattered light, which is dominantly blue. Some of that
dominantly blue light is scattered to the earth, giving the sky its
characteristically blue color. The removal of selected wavelengths makes the
sun appear yellow.
At sunrise and sunset the sun's light travels farther through the
atmosphere than it does at noon and more light is scattered. The remaining
direct sunlight, therefore, is dominantly red. If particulate matter, such as
dust, about the size of an optical wavelength is present, scattering is
insensitive to the wavelength and the sky appears to be dark.

Biconical Transmitting Antennas

2.4 General Comments

A biconical antenna is illustrated in Fig. 2.4.1. The input power is applied
across a sphere of radius b, centered at the apices of the cones. The cones
extend from radius b to radius a, the length of the cones, at angle \j/ as
measured from the z-axis. All surfaces are ideal conductors. Source radius b
is much smaller than either a or wavelength X.
Biconical antennas are unique in that they are amenable to a rigorous and
complete electromagnetic analysis and are shaped similarly to many
practical antennas. Any solution with fields that satisfy the Maxwell
equations and for which the fields match the boundary conditions is both a
unique solution, see Section A. 13, and a complete solution. Completeness
assures that all solution terms are present, in contrast with numerical
solutions that begin with an assumed symmetry and obtain an iterative
answer. For those cases, the output solution contains only symmetries
Selected Boundary Value Problems 53

Figure 2.4.1 Schematic Illustration of a Biconical Antenna

The antenna arms are conical sections that extend from b and a, expansion
half angles y/ are measured from the z-axis, and the outer termination of the
cone is capped by a spherical segment of radius a.

present in the initial input and hence the solution is only as complete as the
initial input.
Transmitting antennas include an energy source that applies a sinusoidal
steady state voltage or current to source region b. The two cones, although
oriented in opposite directions from the center sphere, act as a transmission
line and direct the energy through the inner region, radius b to radius a, as a
TEM mode. The energy then passes through the open aperture at r = a and
enters the outer region. All radiation has rotational symmetry about the
antenna axis and many wavelengths from the antenna the electric field
intensity is linearly polarized in the direction of the conical axis. The
impedance that the antenna presents to the source is determined by details of
the antenna structure: cone angles, cone length, and the wavelength of the
radiation. The outgoing waves undergo a discontinuity in the wave
admittance (impedance) at the open aperture that results in infinite sets of
TM modes in both the interior and exterior regions. Both inner and outer
modes support standing energy and a steady state outward energy flow.
Solution of the transmitting antenna problem requires solving for the input
admittance, the coefficients of each of the infinite sets of interior and
exterior TM modes as well as the TEM mode, and the radiation pattern.
Analysis is simplified by dividing space in the following way:
54 The Electromagnetic Origin of Quantum Theory and Light

Source region
r<b\ 0<e<rc; 0 < (|) < 27r (2.4.1)

Interior region
Arms
b<r<a; 0 < 6 < \ | / and7i-\|/<9<7i; 0 < <}) < 271 (2.4.2)

Space
b<r<a; y<Q<n-x\f, 0 <<))<27r (2.4.3)

Exterior region
r>a\ 0<6<n; 0 < <(> < 27t (2.4.4)

Aperture
r=a\ \|/<6<7i-\)/; 0 < (> < 2TT (2.4.5)

2.5 Fields
The first objective is to obtain an expression for all fields. The procedure
bebegins with the general expansion, Eqs. 1.12.9, and imposes boundary
conditions specific to the biconical structure of Fig. 2.4.1. As was the case
for the analysis of scatterers, field determination is greatly simplified by
incorporating general field properties before matching the boundary
conditions. General field properties are: (1) Since the antenna has rotational
symmetry about the z-axis there is no dependence upon azimuth angle (j) and
only functions with degree m equal to zero form part of the solution. All
coefficients F(v,m) and G(v,m) are equal to zero for m greater than zero.
This changes the sums over orders and degrees of Eqs. 1.12.9 to a sum over
orders only. (2) The source drives straight currents that produce no current
loops. Since TE coefficients are generated by current loops all coefficients
G(v,0) are equal to zero. (3) A source located evenly between the two cones
drives surface current density with the symmetry I(r,\j/) = l(r,n-\y) and
surface charge density with the symmetry p(r,\\i) = -p(r,jt-i|/). By
Eqs. 1.12.9, and with v = £ an integer, E e is proportional to dP^(cos8)/d0. It
is shown in Section A. 18 that Legendre functions have either even or odd
Selected Boundary Value Problems 55

symmetry as £ is even or odd. Consider a Legendre function of order £

containing terms with the symmetry of cos 6. For that term:

If P^(cose) = cos / 9 then E e « — ^ ^ = ^ c o s / _ 1 0 s i n 0 (2.5.1)

For £ odd Eq. 2.5.1 shows that E e (a,0) = E e (o,7i-0) and for £ even
E e (a,0) = -E e (a,7t-0). Since the source drives only even symmetry electric
fields, it follows that only odd symmetry Legendre functions appear in the
field solution. Therefore, the coefficients of all even order Legendre
functions are equal to zero.
The exterior region: (4) Since the z-axis is included in the field region all
terms have null coefficients except Legendre functions of the first kind. (5)
In the limit as the radius approaches infinity, energy conservation requires
the radial dependence to be exp[i((ot-a )]/a which, in turn, requires the
coefficients of all radial functions except Hankel functions of the second
kind to be zero.
After incorporating the five constraints into Eqs. 1.12.9 and making the
notational shift:

F, = /-'F(*,O)

The most general possible set of exterior field components is:

(2.5.2)
^ , .dP/(cos0)
n
TiH^-zX F,h,(o) d 6
>
l=ly>

oo

oEr= ^ £{£+\)¥^{a)?e{cosQ)
^=l;o
56 The Electromagnetic Origin of Quantum Theory and Light

The symbol £ = l;o indicates the sum begins with £ = 1 and is over odd
integers only. The constants F^ form an infinite set of unknown but constant
field coefficients. Complete problem solution requires obtaining a solution
for each of them.
The interior region: (6) Since the cones exclude fields from the z-axis
modal orders need not be integers. Since symmetry requirement (3) requires
null coefficients for even functions by Eq. A. 17.26 the coefficients of the
even parity portion of Legendre functions, Lv(cos6), are equal to zero. This
restricts solutions to odd parity Legendre functions, Mv(cos9). It follows in
the same way that the zero order Legendre function, Po(cos0), has a null
coefficient but, by Eqs. A. 18.14 and A. 18.15, zero order Legendre function
of the second kind, Qo(cos0), does not; the derivative of the zero order
Legendre function of the second kind remains finite on cone surfaces. (7)
Both the source voltage and the source current are finite. The voltage and
current are, respectively, proportional to c times the electric and magnetic
field intensity and the radial functions approach zero as j v (a) => a and
y v (a) => a~' v+ , see Eqs. A.24.9 and A.24.1. Therefore the input voltage
and current values remain finite only if the coefficients of all spherical
Neumann functions except v = £ = 0 are equal to zero.
Incorporating these constraints into Eqs. 1.12.9 and separately denoting
the zero order TEM mode shows that the general forms of the interior field
components are:

E,= £ r v v ( v + l ) i i ^ M v ( c o s 9 )
CT
v>0
Be = | o r v ; v ^ + / [ a ( a ) + d „ y - 0 H ] ! ! « M

1 * ~ ' X rvjv^+[cdo(,) + doyo(a)]!!^

d6 d9
v>0

Coefficients of noninteger order modes, F(v,0) of Eq. 1.12.9, are denoted

by Tv and coefficients of zero order spherical Bessel and Neumann functions
respectively by the constants CQ and do-
Selected Boundary Value Problems 57

2.6 T E M M o d e
The TEM mode may be reformulated in terms of measurable antenna
parameters. Consider properties of Qo(cos6), see Section A. 18:

1 + COS0
Q o (cos0) = ln cot — In (2.6.1)
l-cos0 y

Differentiating:

dQ0= 1
(2.6.2)
d9 sin0

The zero order spherical Bessel, Neumann and related functions are:

. , x sin a , , cos a .., , cos a , , , sino

Jo(°") = — ; yo(w) = ; J O ( ° ) = — ; yo(°") = — (2-6-3)
a G O G
Substituting v = 0 and Eqs. 2.6.2 and 2.6.3 into Eq. 1.12.9 give:

Er = 0
1
CTEQ = (coCosCT + dnsina) (2.6.4)
H
/sin9 v u " '
ariH,), = —r~- (do cos a - c 0 sin a)

The voltage difference between equal radii positions on the two antenna
arms is a measurable quantity. It may be calculated from knowledge of the
antenna structure and the electric field intensity using Eqs. 2.6.3:

a V „ .„ 1 , . , . *? d9
V(/*) = — I EQ6Q = — (CA COS a + d 0 sina) | — (2.6.5)
* sin 6
v V

Integrating Eq. 2.6.2 shows that:

58 The Electromagnetic Origin of Quantum Theory and Light

7l-V|/
d9 co
V
sinG
= 21n
«i
It is useful in what lies ahead to define the line admittance of the
transmission line formed by the two antenna arms to be G(\|/) where:

%
G(V) = - (2.6.6)
r|ln cot

Combining Eq. 2.6.5 and 2.6.6 shows voltage V(r) to be:

2%
V(r) = -r (cocosa + dosina) (2.6.7)
#r)G(v|/)

Substituting Eq. 2.6.7 into the TEM component of the electric field intensity,
Eq. 2.5.3, shows the zero order electric field intensity to be:

_ Tl/EV(r)G(x|/)
Efl = (2.6.8)
27iCTsinQ

Since the magnetic field intensity is directed around the cone arms, the
current on the antenna arms is radially directed. Use of Eq. 2.6.3 gives the
relationship:

271
l(r) = — sin8 J H^dtj) = — ( d o cosrj-c 0 sinaj (2.6.9)
llrC
0

Substituting Eq. 2.6.9 into the TEM component of the magnetic field
intensity term of Eq. 2.5.3 shows the zero order magnetic field intensity to
be:

k\(r)
HA - • (2.6.10)
2TICT sin 6
Selected Boundary Value Problems 59

Defining position r = a to be the terminus, the voltage and current there

follow from Eqs. 2.6.7 and Eq. 2.6.10:

V(tf) = r { c o cos(^3,) + do sin(^a')}

Tl vJ/A

(2.6.11)
1(a) = — {d 0 cos(^5?) - c 0 sin(^isr)]
T|A

Inverting Eqs. 2.6.11 to obtain the field coefficients in terms of terminator

voltage and current gives:

c 0 = —\iGV(a)cos(ka)- l(a)sin(fca)}
(2.6.12)
d 0 = — {l(a)cos(ka) + iGV(a)sin(ka))

Next, define Y(a) to be the terminator admittance:

Y{a)=l{a)/V{a)

Rearranging gives the voltage between, and the current on, the cone arms as
a function of Y(a):

V(r) = ^{Gcos[£(a-r)] + lY(a)sm[t(a-r)]}

l(r) = V(a){Y(a)cos[t(a- rj\ + /G sin[/k(a- /•)]}

In terms of the terminator and line admittances, the admittance at each radius
along the cones is:

, . Y(a)cos[t(a-r)] + Xjsin[i( #-/•)]

Y[r)=
Gcos[t{a-r)\+tY{a)sm[£{a-r)} (26]4)

Use Eq. 2.6.14 to define the antenna input admittance Y(0) then put it
equal to Yo, the admittance at r = b in the limit as b approaches zero. Also,
define the input voltage, V(0), and current, 1(0), to be:
60 The Electromagnetic Origin of Quantum Theory and Light

, , Lim . . , , Lim , ,
v(o) v( ; l(o) =
%^o ^ ^ol(^ (2A15)

The input admittance is:

_ j Y((?)cos(-fo)+iGsin(/fo7)1
I Gcos{ka)+iYya)sm.fca
I (2.6.16)

The radial dependence of the admittance as a function of the input and line
admittances is:

w \ „ I Y n cosCT-/GsinaI
Y(a) = G^—^ \ (2.6.17)
[Gcosa-/Y0sinaJ
The line admittance equations have the exact form of admittance transfer
along a TEM transmission line and show that the cone arms jointly act as a
constant admittance line guiding the TEM mode from the source to the
terminus. Quite differently from a parallel wire transmission line in which
the guiding conductors remain equally spaced along the length of the line,
here the guiding conductors are oppositely directed on either side of the
source. Like many transmission lines the line impedance is constant, see
Eq. 2.6.6. Voltage is measured between equal radius points on the cone arms
and the current is measured along each arm.

2.7 Boundary Conditions

Packaging the TEM results of Section 2.6 into the interior field equations
shows that the general form of the interior fields is:

Er=irvv(v+i)i^Mv(cose)
CT
v>0
y r .. dM v (cos8) T)*GV(o)
9 + {2JA)
" v>o de ^in?
_ - dMv(coS9) E{G)
v>0 d9 2TO7sin6
Selected Boundary Value Problems 61

The infinite set of multiplying coefficients T v and the input admittance Y(0)
are unknown and to be determined.
Since the magnetic field is entirely ^-directed, all currents on the cones
are directed along the length of the cones. The total current consists of the
sum of currents associated with the TM modes and the TEM mode. Define
the TM modal current I'(cr) to be the complementary current and the TEM
modal current 1(a) to be the principal current. The total current is the sum:

IT(o)=r(o)+l(o) (2.7.2)

The first term in the expression for HA shows that the complementary
current, in amperes, is:

_,/ A 2no ^ . , vdMv(cos9)i

^vfo""^7 d9 |9=V
(2.7.3)

Since j v (a) varies asCTVfor small radii, where v > 0, it follows that the
complementary current vanishes in that limit:

Lim ,.
r(0) = 0 (2.7.4)

The principal current at the origin follows from Eqs. 2.6.9 and 2.6.12, and is:

1(0) = \{a)cos{ka) + i GV(#)sin(/fo) (2.7.5)

Since only the principal current exists at the source, only it can support the
energy flow away from the source. Since the time average power supported
by the TEM mode does not depend upon the radius, it follows that the time
average power is guided through the region by the principal current.
Application of the conducting boundary conditions to the exterior fields
of Eqs. 2.5.2 shows that the field intensities on the caps are related to the
surface charges and currents as:
62 The Electromagnetic Origin of Quantum Theory and Light

0<6<\|/ and 7t-\|/<0<7i;

CO

eEr(/fo,e,<t>)=- 2 ^+i)F^(/fo7)p^(cose)=p(/^,e,()))
CT
*=l;o
<2 76)
M-M*)- £ F ^ M ^ L o d6
'
/=l;o

H^,e,f)=-- X F ^ ( ^ ) n ' os j=-/e(^,e,^)

d0
^ /=l;o

Symbol p(£a,6,(|)) indicates the surface charge density on the caps in

coulombs per square meter and symbol Ie(£a,9,(|)) indicates surface current
density on the caps in amperes per meter. Application of the conducting
boundary conditions to the interior field components of Eqs. 2.7.1 shows that
the interior field intensities on the arms are subject to the constraints:

kb<G<ka

Er(o,V)$)=X v(v+l)Fv^^Mv(cos\|/) = 0
°
rwtGV(o-) ^ .dM v (cos0)i
eEe(a,v,0) = - — + e2, r vJv—*J7— L \B= V = P(O,V.4>) + P (a.v.W
2 no sin \\i ^Q d0 ' v
^, „ . dMv(cos6)i ti(o)
TiH+(o,v,4>) = - / ' i , rvJv *jr 'e=w + „ .' = Ir(o.V.«l») + I'r (o,V.*)
v>0 d0 ' 2TOT sin 0
(2.7.7)
Symbols with and without the primes indicate, respectively, principal and
complimentary surface charge and current densities on the cone arms.
The null value of the radial field component at the conical surfaces is
only satisfied by a nontrivial solution if for every value of v:

Mv(cosY|/) = 0 (2.7.8)

Equation 2.7.8 determines an infinite and unique set of positive-real

eigenvalues of v. Plots of v versus angle \|/ for which Eq. 2.7.8 is satisfied
Selected Boundary Value Problems 63

are shown in Fig. 2.7.1 for the first through the fifth sequence of roots.
Function Mv(cos0) is plotted versus v in Fig. 2.7.2, showing the first 24
zeros. Plots of Mv(cosG) versus 6 at the first two roots of Mv[cos(5°)] are
illustrated by Figure 2.7.3.

Figure 2.7.1 Root Values, Noninteger Legendre Functions

Plot showing values of V for which the three functions Mv(cosV|/),
M v (cos\|/), and dL v (cos8)/d6\Q=^ , are equal to zero versus cone angle \j/.
Cardinal numbers indicate root order.
64 The Electromagnetic Origin of Quantum Theory and Light

in
en
O 0.5

0 —If

-0.5
16 20 24

Figure 2.7.2 The function Mv[cos(50)] plotted versus v.

On the aperture, virtual boundary conditions apply, see Eqs. A. 12.6, and
all field components are continuous through the boundary. Imposing these
conditions on Eqs. 2.5.2 and 2.7.1 give the constraining equations:

\\f < 6 < n - \ | / ;

OO CO

aEr(/fe7,e,<)>)= £ £(f + l)F^h*(/fo7)P£(cos9) = ^ v(v + l)rvj*(/fe7)Mv(cos6)

l=\v v>0

(2.7.9)
\|/ < 6 < r t - \ | / ;

„ / , „ ,\ v^ . • / , \dP/(cos0) ^. . . / r xdMv(cos6) riKjV(tf)

. , d6 , d0 27iasin0
£=l;o V
(2.7.10)
\|f < 8 < u - i i / ;
/ \ •ri / , \dP/(cos9) •c-i„ . / , \dM v (cos9) £V(a)
v
, , d6 "7 d6 27tasm8
(2.7.11)
Selected Boundary Value Problems 65

Figure 2.7.3 Two lowest order functions Mv(cos0) vs. 0; Mv[cos(5°)] = 0

These are the field values on the interface between interior and exterior
regions. This completes the discussion of the field equations at a point as
boundary conditions.

2.8 The Defining Integral Equations

In the preceding sections, the general forms of the boundary conditions are
obtained as infinite sums over radial and harmonic functions. The exact form
of the functions and the relationships between them is specified. In each case
what remains are sums over an infinite set of modal orders and it remains to
separate out the coefficients, one by one. In all but one case this is
accomplished using the orthogonality of the Legendre functions. Sets of
orthogonal integrals are formed and evaluated that change the equalities
66 The Electromagnetic Origin of Quantum Theory and Light

involving Legendre functions of the preceding sections into linear algebraic

equations. The algebraic equations are used to solve for the coefficients.
The first algebraic equation is obtained without using orthogonality.
Operating on Eq. 2.7.11 to evaluate the line integrals of the expression for
H^ around the periphery of the antenna arm on both sides of the r = a
boundary gives:

7t-l|/
J aH^dij) = - / — ^ F^h^(/fc2,)P^(cos 6) 7 1 - V | /
v ^l=l
(2.8.1)

= -^iryrv.Jv(^)M vM
v (cose)r+ +^
v j v (^)M v (cose)|-V

The condition that Mv(cos\|/) = 0 removes the sum over v. Collecting the
remaining terms and making the substitution that 1(a) = Y(a)V(a) gives the
interior fine admittance at the terminus as a function of the exterior
coefficients:

~ . oo
Y
M T i n 2 F *M^) p /( cos v)
=
(2-8-2)

Under the summation sign of Eq. 2.8.2, symbol ^o;l indicates that t
represents the field of odd integers with the lowest value of one. This
equaiton, the first of the algebraic equations, equates the applied voltage and
the admittance to a sum over odd order, exterior modes.
The next algebraic equation is obtained using the orthogonality of
integer order Legendre functions. Multiplying Eq. 2.7.10 by
sin6d9dP n (cos0)/d9 and integrating over the aperture gives:

(2.8.3)
Selected Boundary Value Problems 67

Although with the problem as stated the limits on both integrals are from y
to 7i- V|/, it follows from Eq. 2.7.6 that the sum on the left side of Eq. 2.8.3 is
equal to zero over the caps. Therefore, the range of integration of the left
side may be extended to the full range 0 to 7t without affecting the value of
the integral. Use the extended angular range and use the definitions of
Tables A.22.1 and A.23.1 that:

lu = and l£v= j ?£(cos0)MV(cos8)sin0d9 (2.8.4)

Symbol T with two subscripts indicates an integral and with one subscript
indicates current. Evaluating Eq. 2.8.3 by incorporating Eq. 2.8.4,
Table A.22.1.6, and Table A.23.1.1 gives:

t(l + l)¥th't{^)ltt = *(* + 1 ) £ rvj;(/fa)l^v - ^ ^ P ^ ( c o s V ) (2.8.5)

%a
v>0

Equation 2.8.5 is the second algebraic expression that equates individual

exterior modal coefficients to a sum over interior modal coefficients.
The third algebraic equation uses the orthogonality of fractional order
Legendre functions. Begin by multiplying Eq. 2.7.11 by
sin0d0dM^(cos0)/d0 and integrating over the aperture:

y «o;l

= *f sin ede^f£r v j v ( fo )^ + 32v(£)l

J de | Z , vJW / de 2nasme\

Evaluating the integrals of Eq. 2.8.6 using integral A.23.1.5 with A.23.1.1
and integral A.23.1.7 with A.23.1.6 gives:

V{» + l)l W i^j | 1 (A7)= X £{l + l)Ftht{Aa)ltVL (2.8.7)

68 The Electromagnetic Origin of Quantum Theory and Light

Equation Eq. 2.8.7 is an algebraic expression that equates individual interior

modal coefficients to a sum over exterior modes.

2.9 Solution of the Biconical Antenna Problem

The result of applying the orthogonality of Legendre functions to point
equations is an expression for individual exterior or interior modal
magnitudes as sums over interior or exterior modes, respectively. The
equations are:

Feht(ta) = —^4-// ' + S rvjv(^h^ (2-9.1)

1 °° I

rvJvOk) = - r - A E F ^ ( ^ ) ^ + l ) - ^ (2.9.2)
V
VV+Vfo;l :
w

Each equation contains an infinite number of linear algebraic equations. The

zero order interior mode is:

Y
M = \T\
W a
E F<M^)p cos
*( \f) (2.9.3)
\) &>;1

It remains to solve the three equations for the individual coefficients.

After some manipulation, including multiplying through by
h ^ ( a ) / h * ( a ) , Eqs. 2.9.1 and 2.9.2 combine to form the equality:

*M*«)- E Fnhn(i*) £ °(n+i)i*i.»&(feMfe)

(2.9.4)
TiG V(a) P^cosy) he(to)
n a t(l + i)leeh't(*a)
Selected Boundary Value Problems 69

Equation 2.9.4 represents an infinite set of linear equations, one for each
coefficient F», and has the form:

^+£N*nXn=B^ (2.9.5)
n=l

Because the magnitudes of coefficients ¥e decrease rapidly with increasing

modal number, and to keep the magnitude within available computer range,
it is helpful to solve the problem with the initial variable x^ equal to Ffh^ka).
After solving for xe and knowing h^ka), solve for F^.
Although an equation of the form of Eq. 2.9.5 may be readily solved
using matrix techniques, doing so requires the series to be truncated and
truncation produces errors. The solution procedure is to: (1) pick an arbitrary
but specific value for the ratio V(a)/a, our choice was one, (2) use the matrix
solution to solve for the product Ffh^ka), (3) divide by h.^ka) to obtain F^.
The procedure determines as many of the previously unknown exterior
coefficients as needed, and is limited only by the capability of available
computers. This completes the calculation of the exterior coefficients.
Knowing F^hfika), Eq. 2.9.2 may be truncated and solved for F v , and

Eq. 2.9.3 may be truncated and solved for the admittance Y(a):

Y(*) = - 7 - w - £ F ^ ( / t o ) (2.9.6)

All quantities on the right side are known. Since for each mode F(he(ka) is
proportional to V(a)/a, the magnitudes in the numerator and denominator of
Eq. 2.9.6 cancel and the value of Y(a) are correct for any applied voltage.
To change the field normalization to the more conveniently determined
value V(0)/a = 1, enter the value into the first of Eqs. 2.6.13 to obtain:
V(tf) G
a Gcos( ka) + /Y( a) sin( ka) (2.9.7)
70 The Electromagnetic Origin of Quantum Theory and Light

Use of Eq. 2.9.7 to re-normalize F^ completes the numerical analysis of

biconical transmitting antennas.
Badii, Tomiyama, and Grimes used The Pennsylvania State University
main frame computer, programmed for quadrupole precision, to do
numerical analyses of several biconical antennas through 12-place accuracy.
The analyses included series truncation with 17 external (maximum modal
number of 33) and 16 internal modes. Table 2.9.1 lists, with six place
accuracy, values of F^h^/fc?) and Tv]v{ka) for an antenna with \|/ = 5° and
ka = 2, external modes one through 17 and internal modes 1.444 through
16.391. Table 2.9.2 lists the first six figures of F^ and T v for the same
antenna. The table values illustrate that the magnitudes of F^ and Tv
respectively decrease and increase rapidly with increasing modal number.
The coefficients and Eq. 2.9.7 determine the terminal admittance, Y(a). Y(a)
and Eq. 2.6.14 determine the antenna input impedance Y(0).

£ F/M^) V rvjv(te)
1 (1.50924-/2.40989)D-01 0
3 (5.42697-/1.70419)D-02 1.444 4840 (4.33823-/17.5963)D-02
5 (21.95628-/6.98844)D-03 3.6094475 (3.68170-/2.37137)D-02
7 (12.5283-/3.81076)D-03 5.754 8721 (2.23379-/1.17634)D-02
9 (8.18028-/2.93358)D-03 7.887 3272 (2.59971-/1.27571)D-02
11 (5.72681-/2.11622)D-03 10.016 937 (-10.8348+/5.10427)D-02
13 (4.17152-/1.57614)D-03 12.143 571 (-8.8961 l+/4.07774)D-03
15 (3.10771-/1.19513)D-03 14.268 228 (-3.57476+/1.60594)D-03
17 (23.4057-/9.13644)D-04 16.391 498 (-19.5095+/8.62755)D-04

Table 2.9.1 Constants for \|/ = 5°, ka = 2

Figure 2.9.1 shows the input resistance and reactance as a function of

arm length for 5° cones. The mark at about -/450 ohms shows the input
impedance of an antenna with ka = 0.5; succeeding marks are spaced at
intervals of cones made longer by A(£a) = 0.5. Figure 2.9.2 shows the input
impedance of an antenna with V|i = 5° as a function of arm length. Note that
the initial resonance is much sharper than succeeding ones; peaks are
centered at about ka = 1.11, 4.06, and 7.14.
Selected Boundary Value Problems 71

I F^ V
rv
1 (-6.00998-/50.5094)D-02 0
3 (-9.9686 W36.9686)D-03 1.444484 (1.32836-/5.3 87966)D-01
5 (-3.75732-/11.81043)D-04 3.609448 (14.3649-/9.25205)D-01
7 (-6.96510-/20.3034)D-O6 5.754872 (3.36507-/1.77209)D+01
9 (-7.74256-/21.5902)D-08 7.887327 (3.04273-/1.49310)D+03
11 (-5.7288-/15.5031)D-10 10.01694 (-16.5306+/7.78756)D+05
13 (-3.01440-/7.97810)D-12 12.14357 (-2.67294+/1.22520)D+07
15 (-1.18075-/3.07030)D-14 14.26823 (-2.98297+/1.34009)D+09
17 (-1.225 86-/9.11817)D-18 16.39150 (-6.07222+/2.68528)D+ll

Table 2.9.2 Constants for \\i = 5°, ka = 2 (cont.)

500 I I

400 -

300 / ^ ^ ^ \

200 - / ^ - ^ \
w 100 - / / sf ^ \ V
u

u
< -100 -J \ V J I
-200 \» y
-300

-400

-500 1
0 100 200 300 400 500 600 700 800 900
RESISTANCE

Figure 2.9.1 Input impedance of a biconical antenna

Input Impedance of a biconical antenna, with \|/ = 5 ° as a function of arm
length. Zero reactance values are at ka = 1.11, 2.59, 4.06, 5.51, 7.14. The
first mark is at ka = 0.5, each succeeding mark increases ka by 0.5.
72 The Electromagnetic Origin of Quantum Theory and Light

By Eq. 2.7.6, each modal contribution to the total magnetic field

intensity at the aperture is equal to F^h^(i»)dP^cos(6)/d9; the magnitude
of Yfh^ka) is listed in Table 2.9.1. From the theory of Legendre
polynomials, see Table A. 18.1:

P f cos(e) ,2= H ) (-)/ 2 J^ (2.9.8)

d6 (/-!)!!

Values of the ratio of the modal contribution to EL to that of the exterior

dipole mode at ()) = nil are listed in Table 2.9.3. At the aperture, the modal
magnitudes decrease so slowly with increasing modal number that a
reasonably accurate description of interface affects requires a large number
of modes. In the far field, on the other hand, only the first few modes
determine the fields.

i Aperture Far Field V Factorial Interior

Ratio Ratio Ratio Ratio
1 1 1
3 2.0005D-01 7.5275D-02 1.4444840 3.2498 2.0713
5 8.1033D-02 2.4366D-03 3.6094475 3.6529 5.6258D-01
7 4.6052B-02 4.2199D-05 5.7548721 3.8350 3.4050D-01
9 3.0563D-02 4.5092D-07 7.8873272 3.9504 4.0232D-01
11 2.1471D-02 3.2493D-09 10.016937 6.8083D-02 2.8677D-02
13 1.5682D-02 1.6767D-11 12.143571 3.8859D-01 1.1845D-03
15 1.1710D-02 6.4671D-14 14.268228 9.6692D-01 1.8995D-04
17 8.8363D-03 1.8087D-17 16.391498 1.6152 3.4457D-03

Table 2.9.3 Magnetic Field Modal Magnitudes

Modal magnitudes of HJka,ii/2) for exterior and interior orders. All ratios
are normalized by the magnitude of the exterior dipole mode. The first three
columns refer to exterior fields: the first is modal number, the second the
modal-to-dipole field ratio at the aperture and the third the modal-to-dipole
field ratio at far field. The next three columns refer to interior fields: the first
is modal number, the second the magnitude of Eq. 2.9.9, and the third
modal-to-aperture field ratio.
Selected Boundary Value Problems 73

In the interior region, modal magnitudes of H^(ka,n/2) at 0 = n/2 are

equal to:

rvJv( (2 9 9)
^(VTi)ii -"
Values of the ratio of factorials and the ratio-magnitude product for each
mode are listed in Table 2.9.3. The ratio is a measure of the rate of
convergence of the field expressions with increasing order. Although the
external modes are monotone decreasing with increasing modal number, the
internal modes are not; the interior modes decrease but not monotonically
with increasing modal number. The difference is because interior-to-exterior
modal coupling depends upon the numerical difference between the interior
modal orders and odd integers, as well as the magnitudes of the modes.
Field values determine the charge and current densities on the antenna
surfaces. Results are summarized as:

Cap: <3=kcr, 0<8<i|/; 7T-i|/<0<7i

p(**,9) = e £ ^ H ^ M M p ^ e o s Q ) ^ (2.9.10)
to meter2
fo;l
l
r 11 a\ ' v c u l r \ dp ^(cos9) amperes
T] £yX 00 meter

Cones: b<r<a, 9 = \|/

P ^ ) = (e I r v j - v ( a ) ^ ^ ) | e = , + ^YW_l^lornbs
dG
v>0 ' 2ncasmy\ m e t e r 2
&(r) i " . . >dMv(cos0)i I amperes
Ir(a,V|/)+rr(/-,\|/) =
meter
(2.9.11)
On the cones the surface current density is radially directed and on the caps
it is zenith angle directed. The two currents have quite different
dependencies upon radius and zenith angle and therefore the current is not
continuous through the cone-cap junction. A loop of charge accumulates at
74 The Electromagnetic Origin of Quantum Theory and Light

the junction with a sign and magnitude that depends upon antenna structural
details and the radiated wavelength. The resulting ring charge is:

V R u ka11. \ ^ ( c o s v ) v „ . /, vdMv(cosi|/)
/CO
L $M ) V——+ L r vJv(^j v- L
cosin\|/
>;1
(2.9.12)

Similarly, charge densities on the cap and cone have quite different
dependencies upon radius and zenith angle. The electric field intensity on the
cone and cap are, respectively 6 and r directed, and just off an ideal 90°
junction the field is directed at an angle of 45° as measured from both the
cone and cap.

0.011
0.010

0.009
p 0.008
0 0.007

w 0.006

E 0.005
R 0.004

0.003

0.002
0.001

0.000

NORMALIZED RADIUS, ka
Figure 2.9.2 Output power versus ka for a 5°, biconical, transmitting antenna.
Antenna with 5 ° cone angles and a constant input voltage, showing radiated
power peaks at ka =1.11, 4.06, 7.14.
Selected Boundary Value Problems 75

2.10 Power
For a transmitting antenna, the time average power in the interior and
exterior regions follow by use of the fields of Eqs. 2.7.1 and 2.5.2,
respectively. In the interior region, the time average real power satisfies the
transmission line rules between radii b and a. The input impedance and the
radiated power are strong functions of the physical location of the standing
energy wave, and it depends upon the antenna arm length.
The time average power produced by an antenna is equal to the integral
of the real part of the radial component of the complex Poynting vector at
the surface of a virtual sphere which, for ease in calculation, is made
concentric with the antenna. The fields of Eqs. 2.5.2 show that the time
average output power is:

2 |27t Jt oo oo HP HP I
P w = — R e Jd4>Jsinede2 £ F , F n V ^ K ( a ) - ^ - - J L (2J0.i)
dG d 9
W U 0 t=l n=l J

Replacing the Hankel functions by their far field values and evaluating the
integral gives:

1 ~ 4*+l)r *i
P [F<F ] ,2J02)
"V,?,(M '
Since all terms on the right side are known, Eq. 2.10.2 is sufficient to
evaluate the output power.

The time average power input, Pin, to the antenna is:

P i n =^Re[v(0)l*(0)] (2.10.3)

By Eq. 2.6.13 the TEM voltage and current in the interior region are:

V(r) = ^-{Gcos[A(a-r))+ /Y(«)sin[^-r)]} ^

!(/-) = V(a){Y(a)cos[t(a- /•)] + Xisin[jf{a- /•)]}

Combining Eqs. 2.10.3 and 2.10.4 gives:

76 The Electromagnetic Origin of Quantum Theory and Light

Pin=^Re[v(tf)V»]Y» (2.10.5)

Since V(a) and Y(a) are known, Eq. 2.10.5 is sufficient to evaluate the input
power.
In a lossless antenna:

Pin=Pav (2.10.6)

The equality serves as a check on all procedures.

The complex power, P c , on a concentric sphere of normalized radius a
is:

kb < a < ka

Pc(a) = Jsin6d0 — X r v r v j* v j v — * + \ 2 V r I (r)

0 2 V a6 ; 2na sin e
(2.10.7)
a>/fez
2 n
• f AX> ^2 °°
Pc(«) = ^ T JJ ^ sinedeE F,F;n;(a)h;(o)
n^ 0Ue; f=l

With biconical antennas, the electric field intensity just off the surface of
the caps has only a radial component. Therefore there is no normally
directed Poynting vector and no energy is exchanged between the cap and
the field. Since all aperture fields are continuous through the aperture, the
total complex power is a continuous function of radius between positions
a-6 and a+6, where 5 is a differential radial length. Since all fields are
continuous through the aperture, so is the energy density. Adjacent to the
caps, the radial component of the electric field intensity and the azimuth
component of the magnetic field intensity are not equal to zero. Therefore
the energy per unit length as a function of radius is discontinuous between
positions a-6 and a+6. The magnitude of the discontinuity increases with
increasing cone angle.
Figures 2.10.1-2.10.3 describe the complex powers about three antennas
with normalized arms lengths of ka = 0.70, 1.28, and 2.00; all have cone
Selected Boundary Value Problems 77

angles of 1°. The antennas are, respectively, electrically short, resonant, and
electrically long. In all cases the real power, Preal, is constant.
The normalized complex power about an electrically short antenna,
ka = 0.7, is shown in Fig. 2.10.1; the real power is small and the terminal
impedance is capacitive. The peak reactive power is capacitive and occurs at
approximately kr = 0.\. From there, the power decreases slowly with
decreasing radius until reaching the terminals, kr = 0. For increasing radius it
decreases more rapidly until reaching kr = ka, where it drops abruptly, then
decreases slowly to zero with increasing radius for kr > ka.
The normalized complex power about a resonant antenna, ka= 1.28 is
shown in Fig. 2.10.2; the real power is large and the terminal impedance is
resistive. The capacitively phased reactive power peak of Fig. 2.10.1 has
moved outward to about kr = 0.8. From there it decreases slowly with
decreasing radius to zero at the terminals. For increasing radius it behaves
very similarly to Fig. 2.10.1.

x lO" 4
k"Tea\

1
. . . . . . .
-2 • /
:
'T^ •' • •'
/ "^ : ^reactive
-3
: : / : : : : :
/
. / . . : : :
-4 ...•.../.: '
:/ :
• /
:
/ : : : : : :
s

0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25
kr

Figure 2.10.1 Normalized real and reactive powers versus kr for a biconical
transmitting antenna.
Applied voltage V(0) = a, cone angle \|/ = 1 ° and ka = 0.70.
78 The Electromagnetic Origin of Quantum Theory and Light

The normalized complex power about an electrically long antenna,

ka = 2.00, is shown in Fig. 2.10.3; the real power is less than that of
Fig. 2.10.2 and the terminal impedance is inductive. The capacitively phased
reactive power peak has moved outward to about kr = 1.6. From there it
decreases slowly with decreasing radius, passes through zero and becomes
inductively phased at the terminals. For increasing radius it behaves very
similarly to Figs. 2.10.1 and 2.10.2.

xl(T

^freal

kP2 IP,reactive
-3

/
V
\J
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
kr
Figure 2.10.2 Real and reactive powers versus kr for a biconical transmitting
antenna.
Applied voltage V(0) = a, cone angle V|/ = 1 ° and ka = 1.28.

2.11 Field Expansion for j-Directed Exponential

For a receiving antenna to function it is necessary that a component of the
electric field intensity be aligned parallel with the antenna axis; optimum
operation is with full alignment. With plane waves, the directions of
polarization and propagation are perpendicular and, in the analysis of
Selected Boundary Value Problems 79

scattering from a sphere, the incoming plane wave propagated in the z-

direction. To analyze scattering from a receiving antenna it is convenient for
the antenna axis to lie along the z-axis. It is necessary, therefore, to analyze a
z-polarized plane wave and, with z-polarization, it must propagate
somewhere in the xy-plane. We make the arbitrary choice that the plane
wave propagates in the ^-direction. It is, therefore, necessary to expand a y-
directed wave in spherical coordinates. Such an expansion is done in a way
similar to the way it was done for a z-directed plane wave, see Section 2.1.
The desired exponential is:
jiiat-ky) _ /(COZ-CT sin0sin<|>)
(2.11.1)
Since Eq. 2.11.1 has no singularities, the spherical coordinate expansion can
contain no spherical Neumann functions, no fractional order Legendre
functions, and no Legendre functions of the second kind. Only spherical
Bessel functions and associated Legendre polynomials remain. The most
general form for the expansion is:

xlO"

S ^Preal

k?

N4- -••*P,
reactive

kr
Figure 2.10.3 Real and reactive powers versus kr for a biconical transmitting
antenna.
Applied voltage V(0) = a, cone angle \y = 1 °, and ka = 2.00.
80 The Electromagnetic Origin of Quantum Theory and Light

^-lasinOsini)) _ t(t + l)
XX (j( cos mty-i (j( smmfy j,(a)Pf(cos0)
^=0 m=0
(2.11.2)

To complete the description it is necessary to evaluate both infinite sets of

coefficients cGf and sGf. The reverse superscripts "c" and "s" indicate
coefficients of cosine and sine, respectively, and the multiplying factor
£(£ + \)/rn is chosen for later convenience. As written Eq. 2.11.2 is in the
form of two sums over infinite sets of coefficients. In order to evaluate the
coefficients it is necessary to reformulate the equation as a doubly infinite
number of separate algebraic equations, each of which provides a definite
value for one coefficient.
From the theory of spherical Bessel functions, Eq. A.24.9:

J/H = (2*+l)! + higher order terms (2.11.3)

Substitute Eq. 2.11.3 into the right side of Eq. 2.11.2; take £ differentials of
both sides with respect to a, then go to the limit as a goes to zero. The result
is the equality:

1(1+1) t\
(-,)WesinV£ t 'Gfcosmty-i s
Gfsmm^
(2m)!!
Pf(cosG)
(=0 m=0
(2.11.4)

The left side of Eq. 2.11.4 involves powers of trigonometric functions.

Identities that convert powers to multiples of the angles are:

neven
n/2-l
n
sin (t) = — lx^ . ^ - k ^ n ) !
V ]
cos[(n_2k) ] + JnL
[K m
2n ^ (n-k)!k! ( n /2)! 2
k=0
(2.11.5)
nodd
(n-l)/2
11
sin <)) = fX ^ '- k J(n-k)!k!
v
i . Sin[(n-2k»l LV ;VJ
2n ^
k=0
Selected Boundary Value Problems 81

Substitute Eqs. 2.11.5 into Eq. 2.11.4 then multiply by cos(q())), where q is an
integer, and integrate over the azimuth angle. Next, multiply by sin(q(|)) and
repeat the procedure. The results, where the 5 represents Kronecker delta
functions and s represents an integer, are:

c
j£^ni„e)M^]*><e (2,i,)
q (2/ + 1)!! (q/2)![(*-q)/2]!
^^Pncos9)J-f;qr)^yii)s^e
q (2/ + 1)! (q/2)![(*-q)/2]!

By Eq. 2.11.6, CG^ is equal to zero if q is odd and SG^ is equal to zero if q
is even. This reduces the total number of nonzero coefficients by half.
Next, multiply the top of Eqs. 2.11.6 by P^(cos0) and integrate over the
zenith angle. Integral forms are Eqs. 1 and 10 of Table A.22.1. Results are:
qeven
2 < + 1 /</ + q)l s(f| 2z)
C
GM^1) 2 (* + q)l (2.11.7)
q
(q/2)![(/-q)/2]!(2/+l)P ' ' q (2* + l)!!(2*+l) (<-q)!

By Eq. 2.11.7, CG^ is equal to zero if £ + q is odd. Since q is even, it follows

that £ is also even. Conducting the same operation on the bottom of
Eqs. 2.11.6 shows that SG^ is also equal to zero if £ + q is odd. Since, for
this case, q is odd, it follows that £ is also odd. Therefore the two groups of
coefficients form non-overlapping sets and the distinction may be dropped:
G^ represent both sets of functions. This reduces the set of non-zero
coefficients to one-fourth of the original number in Eq. 2.11.2.
Simplifying results shows that:

m 2m(2£+lX£-m)\dl£+ m,2q) 2m(2£ + \\£- mW£+ m,2q)

£+ m m\ £(£+ \\£+ m)\\{£-m)\\
2e£(£+l)\
(2.11.8)
82 The Electromagnetic Origin of Quantum Theory and Light

This equation is correct for all modal combinations. Combining

Eqs. 2.11.2 and 2.11.8 shows that the spherical coordinate expansion for a_y-
directed, z-polarized plane wave is:

e-* s i n e s i n *= X £cosH0-/£ £sin(^)}fcilGfj,(a)Pf(cose)

m
U=0;e me (=l;o mo J
(2.11.9)

An Incoming Plane Wave

2.12 Incoming TE Fields

To find the expression for the radial magnetic field component, Hr, of a y-
directed plane wave with the electric field intensity z-directed, begin by
noting that:

riH r -sinecos(j)e- / e s i n e s i n < t > = ^ — e - / i J s i n 9 s i n * (2.12.1)

a d(j)

The first equality in Eq. 2.12.1 is by definition. Using the derivative

operation of Eq. 2.12.1 on the spherical coordinate expression for the
exponential form of Eq. 2.11.9 gives:

mr :+l)Gf^^Pf(cos6)
f=l;o mo i=2;e me
(2.12.2)

Combining Eq. 2.12.2 with the same field component of Eq. 1.12.9
determines the constant field coefficients. Knowledge of the constant field
coefficients and the component forms of Eq. 1.12.9 are sufficient to obtain
the full set of TE modes.
Selected Boundary Value Problems 83

2.13 Incoming TM Fields

A longer procedure is necessary to obtain the coefficients for TM modes. It
is convenient to define E r in a way analogous with Eq. 2.12.2, using
Eq. 2.13.1 to define coefficients C F/ Z and S F ^ , then solve for the constants.
The general form for the field is:

E
r=£ X [ c Ffcos(M|))-/ s Ffsin(/«t))]4^+l)^lpf(cose)(2.13.1)

The radial component of the ^-directed plane wave is:

/
-its sin0sin(|) _ v -ia sin8sin<|)
E r = cos0e (2.13.2)
a sin § 39

Using the derivative operation of Eq. 2.13.2 on the spherical coordinate

expression for the exponential form of Eq. 2.11.9 gives:

E rr = Y, £cos(w<|))-•/£ j)sin(zw|>)
sincj) (e;0 me £o;\ mo (2.13.3)
l(l + lJ ) ^ J i ( a ) d P / ( c o s 9 )
x-* -G\
K
m CT d0
Symbols ^e;0 and ^o;l indicate sums respectively over even integers starting
at zero, and over odd integers starting at one. Symbols we and wo indicate
sums respectively over even and odd integers. A trigonometric identity that
puts Eq. 2.13.3 in a more useful form is:

^ - = 2£sin[(2s+l)(l)] (2.13.4)
sin* s%

Combining Eqs. 2.13.3 and 2.13.4 results in:

84 The Electromagnetic Origin of Quantum Theory and Light

oo £
Er = 2 ^ i ^ ^cos/wt)+^ ^sin/w|) sin[(2s +1)(>]
s=0|_ te;0 me £o;\ mo (2.13.5)
^(£+1) j^dg^cose)
x^ ^G
m a d6

Other useful trigonometric identities are:

2sin /7«|)sin(2s+ l)(j) = cos[(/«-2s-1)(|)]-COS[(OT+ 2s + l)(J)]

r/ I i (2.13.6)
2coszw(|)sin(2s+ l)(()= sin[(#/+ 2s+ l)(|)J - sin[(w- 2 s - 1)<|)J
The procedure is to substitute Eqs. 2.13.6 into Eq. 2.13.5, equate Eqs. 2.13.1
and 2.13.5, differentiate both (^-1) times by a, then go to the limit as a goes
to zero. The result, for I odd, is:

Gf dPf(cosG)
X J ) {cos[( m- 2s -1)<|>] - cos[(z»+2s +1)<|>]}
m d9
s=0 (2.13.7)
= X [CF"cos(z«|>)- /' S F/ sin(yw|))]p/(cos 0)
a=0

Next multiply Eq. 2.13.7 by cos(q<])), where q is an integer, and integrate

over the full range <>| = 0 to 2n. This shows that S F / is equal to zero for odd
values of I; it follows in a similar way that C F / is equal to zero for even
values of £. Therefore the coefficients form non-overlapping sets and, again,
the notation may be simplified by dropping the reverse superscript, with Ff
representing both sets of coefficients. This reduces the set of non-zero
coefficients to one-fourth of the original number in Eq. 2.13.1. The resulting
equality is

£8(q,*)F;p;(cose)
*=o (2.13.8)
= £i{%l-2s-l|)-5(q, W+ 2s + l ) } ^ ^ ^ l
s=0/we m d0
Selected Boundary Value Problems 85

Evaluating the delta functions and collecting terms gives, after some algebra,
and with U(q) representing a step function of q:

, ,('-t1)/2Gr2S+1<lP?+2s+1(cose)
F?P?(coSe)=2q;q) z q;2s+,
a a, ,
S=U n
d; <?•"•»
With the aid of Table A.21.1.1, Eq. 2.13.9 may be rewritten as:

(i-q-\)/2 rq+2s+l
F^P^(cose) = U( q ) X %—•
s=0 q+2s+l (2.13.10)
cl+2s +2s+2
x[(^+q+2s+lX^-q-2s)P / -P; '

Multiply Eq. 2.13.10 by P^q(cos9) and integrate over 6 using the integrals of
Table A.22.1.2. After simplifying, the result is:

(l-q-l)/2 G q+2s+l f_iVTj/ a \

e V 4/
s-e0 (q + 2s + l ) ( * - q - 2 s - l ) !

Combining Eqs. 2.11.9 and 2.13.11, with n equal to any of the full set of
positive integers, gives:

p * - 4 ^ ) , , J~qY)/2 (-l)SUW + q.2n + l)

F
^-"fTT)" ( q ) sf0 (^q-2 S -l)!!(^q + 2s + l)!! (2 13 12)
' "

The sum of Eq. 2.13.12 is listed in Table A. 15.1.8. Incorporating the sum,
replacing q by m to give the same dummy index as Eq. 2.11.9, and letting n
denote any of the full set of possible integers, the two coefficient sets
F/* and Gf are equal to:

_ 2(21+ 1) V(m\l-m)\^£+ m,2n + l)

' ~ e(£+\) (*+z*-l)!!(i-z»-l)!!
(2 13 13)
m_2(2£+I) m(t-m)\d(l+ m,2n)
£
£(£+l) (l+m)»(t-m)V.
86 The Electromagnetic Origin of Quantum Theory and Light

Coefficients Yf and Gm have opposite parity in that F/* is other than zero
only if £ + m is odd and Gm is other than zero only if £ + m is even. At
degree m = 0 coefficients Y™ have the maximum value and coefficients Gm
are equal to zero. Values through the first five orders are listed in
Table 2.13.1.

1 o!=2
2

2 2
6 12

F»-2 3 3
r 33 - 7

3 c 72 48 96
F4> = ^ - r 2_ 3 F43 =-l- 4
160 160 320
F».»
16
5
240
T.2
5
11
240
3 =1920"
5
5
5760
G =
11
11520

Table 2.13.1 Values offieldcoefficients for a y-directed, z-polarized plane wave

2.14 Exterior Fields, Powers, and Forces

The radial field components of a jy-directed, z-polarized plane wave are
given by the combination of Eqs. 2.12.2, 2.13.1, and 2.13.13:

E
rHX S^CHO-'X S^^) ^ + l)Ff^Pf(cos9)
o;l me £e;2 mo

T]Hr = X £COS(H>)-/'E £sm(^) ^+i)Gf^^pf(cose)

fo;l wo £e;2 #?e
(2.14.1)

The angularly directed field components follow from Eqs. 2.14.1.

One of the two differences between the field forms of the incoming
plane wave and the scattered waves is that in the limit of infinite radius the
Selected Boundary Value Problems 87

scattered wave varies with distance as exp[/(co/-cj)]/a. This functional form

requires the radial dependent functions to be spherical Hankel functions of
the second kind. It is, therefore, necessary to replace the spherical Bessel
functions by spherical Hankel functions. Similar to a spherical scatterer,
different modes scatter with different magnitudes and different phases. To
account for these changes introduce two new infinite sets of field constants,
af and fif, as part of the scattered fields. Let af be the coefficient of TE
modes and $f be the coefficient of TM modes. Incorporating these results
into the radial components of the scattered field gives:

oo £-1 oo g-i
Er= X E cos(zw|>) - / J X sin(H>) ^ + l ) P f F f ^ ^ P f ( c o s 9 )
io'X me £e£ mo
oo a oo £
<*+l)a?G?^^Pf(cos8)
£oi mo £e£ me
(2.14.2)

Problem solution requires evaluation of the full parameter sets

af and $f. As with the spherical scatterer, the total field is the sum of the
incoming plane wave fields and the outwardly directed scattered fields.
Summing the radial field components of Eqs. 2.14.1 and 2.14.2 gives the
total radial field components. The angular field components follow directly
from the radial ones, see Eqs. 1.12.9, and are:

E e = Y, X cos(/*|>) - / £ S sin(w<|>) "(jj+PThj)

Joi me le\L ma de

X YJ COS(/*A)>) - / £ X sin( m§) GfL+afh,)^

_?e2 me fo$ mo \ ' si:
sin 9

oo {-] oo e-\
TlH^ cos
X 2 ( ^ ) - ' S S sin(/wj>)
,^o;l me leQ. ma

-/ £ £ cos(/«|>) - /'£ £ sin(H>) G f ^ + a f h ^

leQ. me to'X mo \ / sin
sinf9
88 The Electromagnetic Origin of Quantum Theory and Light

-EA = i YJ X COS(//A|>) - / ' £ £ sm{m§) Fffj'+Pfh^

l&Q. mo 1.0$ me \ sin0
'SI

(2.14.3)
£ £ COS( mfy) - / £ £ sin(/w(j))
dPf
/o;l /wo ie£ « e d9

TlH e = £ £ cos(/w|>) - / £ £ sin(/w(>)

\ / sin0
si
.^e2 OTO &>;3 /we
oo ^ oo ^
cos sin 7 • \d_Pf
(j*+«fh-,)
5) X (^ - 'X X (H>)
_&>;! /wo & 2 me de

Power on a circumscribing virtual sphere of radius greater than a is obtained

from the radial component of the complex Poynting vector. Substituting
Eqs. 2.14.3 into Eq. 2.2.3 and breaking the resulting vector into four parts
gives:
( oo l-\ l-\ n-1 n-1
£ £ cos(zw)>)- / £ £ sin(/*)>) X X cos(p<(>)+ / £ X sin(p<t>)
\lo me (e mo ^ no pe ne po

WdPf dpP
"F„ p (j^pfh-,)(j n+ prh;
Re de de
Nm = 2n
'oo^-l oo t - l oo n - 1 oo n - 1
£ X cos(H))- / £ J ) sin(H>) X X C°S(P<I>)+X X sin(p(t))
V ^e mo to me v ne po no pe

p ,.. *\UPf pPnp

X/F; % (j*+pfh-,)(j n+ 3rh;
sin6 sin6
(2.14.4)
Selected Boundary Value Problems 89

» °o n =<= n
£ Y, cos(m<b) - £ X sin(H>) S X cos(p<|)) + / £ X sin(p<t>)
\ (o mo (e me V^ no po ne pe

dPf dPnP
xX3fGP(j, + <h,)(j- n + a P X )
Re V
de de
Nrl2=-
2TI V °° n °° n '
sin C0S
^ X COS(H>) - ' S S (^*t>) X X (P<I>) + £ X sin(p<|))
te me lo mo /v, ne Pe no
p° j

' ^ p p ^
x/GfGSa/+aJ'h<)a,n + a5*h;*) sin6 sin0
(2.14.5)

<-I *-i
X £ cos(H>) ~ ' X X sinC^) X X cos(p<|>) + / £ £ sin(p(t>)
V (o me (e mo V^ ne pe no po y

.*jpP^dPf
</FfGPn(j-+P7h-)(j; + aPnVn
Re sinG d6
Nr2i= —
2TI l-\ e-i
X X cos(/w|>) - / £ £ s i n l ^ ) X X «>s(p<t>) + / £ X sin(p(|))
\, £e mo £o me V no po ne pe

x*r&s(j-/+P?h- )(J; + <xi^^

A
\ ' sin6 d6
(2.14.6)

f
o° * oo ^ A oo n-1 <» n-1
X X cos(/«|)) - / £ £ sin( #A>) X X c°s(p(|)) + / £ X sin(P<t>)
We* &> mo V, no pe ne po

xfffGE(j /+ a?h,)(j n + P5Vn

Re sin9 d6
Nr22=-
2^ oo n-1 oo n-1
X X cos ( "$) - X X sin ( ^ ) X X cos(p0) + / £ X ^ " ( P ^ )
V ^o /wo te me V. ne po no pe

<,FrGP(j,+a?h<)(jv n + p P Y i ^ ^
\ ' de
'I sin6 d(
(2.14.7)
90 The Electromagnetic Origin of Quantum Theory and Light

The total surface power is equal to the surface integral of Eqs. 2.14.4
to 2.14,7. Integrating over the azimuth angle gives a Kronecker delta
function of m and p, decreasing the number of sums by one. Results are
shown in Eqs. 2.14.8 and 2.14.9:

271
111+111 *?K(h+w*'()(k+wX)
J X N r l l + Nrl2) = ReV to no me te ne mo J

1 1 1 + 1 1 1 U?G-(j£+afh,)(rn + afhr)
V to no mo te ne me J

X —fl+5(/W,0)l 2
2T] L V n
de de sin e
(2.14.8)

271
111+111 lFfG-(j- +Pfh-)(/n + a f h'n*)
le no mo to ne me J
Jd(|)(Nr21+Nr22)=Re
- 111+111 lFfGn-(j,+a^,)(jn + pfh;)
V & no me to ne mo ,

nsinel^M
(2.14.9)

To complete the evaluation it is necessary to integrate Eqs. 2.14.8

and 2.14.9 over the zenith angle. The integral of Eq. 2.14.9 gives a null
result. Evaluating the integral of Eq. 2.14.8 and replacing the coefficients by
the values of Eqs. 2.13.13 gives:

P a v = ^ R e
1\k \
11+11 Z(t+\Yl-m-W{l + m-\)\\ \M ^ f J
A^ Hf
"Ij
to me te mo
°o e °° t
Ana m2(2{+l)(e-m-l)\\(e + m-l)» / Vf-Ki^V*^
t]k2
Re 11+11 ^+l)(^w)! ! ( ^ m ) ! ! /(Jz + ^ h ^ + a , h, j
to mo te me
(2.14.10)

In the limit of infinite radius, Eq. 2.14.10 goes to:

Selected Boundary Value Problems 91

471 «. t-\ ~ ^-1 \]{m\2£+\\£- m)\\{£+ m)\\

P = (RePf + Pftf)
to me te mo 4 * + !)!!(/- /w- 1)!!(£+ m- l)»V
OO / OO ^
4JI
'"'("^^-"'-""('^-""(Reaf.afaf)]
T[f to mo te me, ^^+lX^-/w)!!(^+/w)!!
(2.14.11)

The terms are interpreted similarly to those of Eq. 2.2.7 for scattering
from a sphere: terms proportional to both oc^oc n and P^fP „ describe time-
average power scattered away from the antenna; each term is positive. Terms
proportional to Re a "^ and ReP^ are negative and describe inwardly directed
power; the time integral of Eq. 2.14.11 is the negative of the extinction
(absorbed plus scattered) energy. For an ideal antenna with shorted
terminals, the two sets of terms have equal magnitude and opposite sign and
sum to zero.

2.15 The Cross Sections

The purpose of this section is to compare and contrast the scattering
properties of spheres with those of a biconical receiving antenna. One view
of a receiving antenna is as a lossy scatterer. Cross sections were defined in
Section 2.2. Analogously with Eq. 2.2.10, the scattering cross section Csc is
defined to equal the ratio of scattered power to the input power density. The
geometric cross section is equal to the cross sectional area of the scatterer.
Using Eq. 2.14.11 to determine the scattered power, for a spherical scatterer
of radius a the scattering-to-geometric cross section ratio is:

" oo t - \ oo (-1
-SC _ \5{ml2l+\ll-m)\\{l+m)\\(amam*
CGE i V .to me te mo £{£+\\£-m-\)\\{£+m-\)\^iPi l\
oo £ oo I 2
m (2£+l)(£-m-l)V.(£+m-l)\l, „ J
*V 11-11
to mo te me
£{£+\\£-m)\\(£+ m)\\
(2.15.1)
92 The Electromagnetic Origin of Quantum Theory and Light

The normalized extinction cross section Cgx is equal to the ratio of the
total power extracted from the incoming plane wave to the geometric cross
sectional area of the scatterer. For a spherical scatterer of radius a and using
Eq. 2.14.11 to determine the total power extracted from the wave, the
extinction-to-geometric cross section ratio is:

C E x_ 4 \j(m)(2£ + l)(£ - m)\\(£ + m)\\

(RePf)
Q}E *v 11+11
to me te mo
£(£ + l)(£ - m - l)\l(£ + m-Y)\\

ml (2£ + X){£ -m- \)\\{£ + m - 1)!!

1 1 +1 1 (Reaf)J
*V to mo He me £(£+l)(£-m)ll(£+ m)\\
(2.15.2)

The absorption cross section, CAB> is equal to the absorbed power-to-cross

sectional area of the scatterer ratio. Using Eq. 2.14.11, the value is:

-AB _ (2£+\)(t-m)\\(£ + m)\\

QJE
11+11
* V . to me te mo
l
>.+ l)(e-m-l)W-
l
\(£+ m-\yA
* i\
oo e oo e mz (2£ + l)(£ -m-\)\\{£+ m-\)\\
Reaf-ccfccf)
*V 1 1 + 1 1 !+l)(v-/»)!!(v + /w)!!
to mo te me
(2.15.3)

The thrust on the antenna from the total power absorbed is c times the
value of extinction power, Eq. 2.15.2. The thrust on the antenna from the
scattered power is equal to the component of scattered wave in the direction
of the incoming field integrated over a virtual surface:

r2 2n
sin
^ySC = ^"J J 4>#J s i n 2 9d6Re(Nr) (2.15.4)
2r)/T 0 0

Inserting the scattered field terms of Eqs. 2.14.3 into Eq. 2.15.4 gives:
Selected Boundary Value Problems 93

(FfFripfpr1h*h;)[i+8(^)]
( \ dPf dP,f+1 ^(^? + l) „ O T D O T +I
2TC
TtRe
SSS
to ne me
x|--
d0 d0
^—+ -
sin 6
'Pf'P n
I Nrsin<|)d(j):
4r|
-(FfFripfpr1h>;)[i+5(—i)]
+SSS
V te no mo )
f dPf* d P ^ /»(#z-l) „ ™_i
d0 d9 sin2e
J J

JiRe
'sss ' \ ( n
' n en
^ de de S in
2
e
£o ne OTO
4n
+SSS \ € n
' n
' n \ de de sin2e J
V te no me J

(G- + 1 Ffa- + 1 *p m h-h n *)[l + 5(m)]

7tRe
'sss ' mPf dPnm+1 (m + l)Pnm+1dP^ffl
^o no me
4r\ sinS d9 sin 8 d0
+SSS
V te ne mo y F h
mPfdPr^^-OP'-'dPr
n * «n P<? tK
sine d6 sine d6

(G m F n m - 1 a m P^" 1 *h,h;)[l + S(m-l)]

7iRe
sss£o no mo 'mPf dp™-1 3 l
{™-l)K~
m-l m
Jt)
dP,
4n sin 0 d0 sin0 de
+SSS
V te ne me ^ pBplH + l J O » l t l * L . * < dP m + 1 ( m + l ) P m + 1 dP;
sin6 d6 sine d6
yj
(2.15.5)

Next, integrate Eq. 2.15.4 to find the directed power through a virtual
circumscribing sphere, with the aid of Table A.22.1.8, 22.1.9, 22.1.11,
and 22.1.12. In the limit where the scattered waves extend to infinite radius,
the normalizedy-directed force due to the scattered field is:
94 The Electromagnetic Origin of Quantum Theory and Light

+
^^+1 +PlP +,
' W(<--l)"(<+-.-l)H
1 1 +1 1
V ^o /we le mo ,

' /„«„*+i* . .,,*» « + i \ ' ' » ( ' * + 0 ( i - ' " - i ) ! ! ( i + / » + i ) ! ! '

+ a a
l ' < + i + a ' < V i ) — „ . ,x2 /# _ w , . x„—
rfjSC 4 (t+if(e-m)n(e+ m)\\
=
1 1 +11
CGE ~ *V to mo £e me A„m„m+\* +,
-r-|Ct/>Ct,, .
^ ^ A <»(<»»+ 1 X < - « - 0 » ( < + " + ! ) "
Ct/ Ct„ . I
v l x]
'-' ~ (i)2(e-M-2)»{e +m)\\

(m + l)(2l + lje-m}.\((+ m)\\

ii+ii
tome krno
(ar'V+ar'pf) e (e+i) (e-m-\)\\(e+ m-i)\\^
2 2

m(2t + V{t-m-\)\\{t+ m + 1)!!

EX + SE [aePe + ae v>t ) 2
(e+\) (e-m-2)\\(e +m)\i _
(2.15.6)
The normalized force due to the extinction power is:

»e-\ °° t-i \](mp.l+\\l-m)\\(t+ m)\\

Refi'l
rfjEX £{£+\\£-m-\)\\(£+ m-\)\\
Ho me ie mo
QJE fa1 m2(2£+l\£-m-l)\\(£+ m-\))\\
Reed
£(£+\)(l-m)\\(e+m)\\
to mo (e me
(2.15.7)

To compare results obtained using y- and z-directed incoming plane

waves, consider scattering by an ideally conducting sphere. For a z-directed
wave, the coefficients are given by Eq. 2.3.5 and 2.3.6; for a y-directed
wave, the boundary conditions follow from Eqs. 2.15.3, and are:

*e(fa) + P f h* (fa) = 0 je(fa) + afht(fa) =0 (2.15.8)

These lead to:

$?(fa): af(fa) = - (2.15.9)

h't(fa) he(fa)
Selected Boundary Value Problems 95

These equations show that the coefficients are independent of degree. For
the case where the scatterer radius is much less than a wavelength the dipole
coefficients are:

pO ( A r ) = */3 « 1 a} (fe)/3 Jkaf_

3 2
2/3-/(/fa) 3/ {Aa)/3+i/{£a) 3
(2.15.10)

The cross sections and normalized forces are:

cEx = c s c = rfEX = i<W 4 (2151I)

3
Q}E QJE QJE

These results are equal to those of Eqs. 2.3.7 and 2.3.8.

Biconical Receiving Antennas

2.16 General Comments

Biconical receiving antennas are of special significance for the same reasons
biconical transmitting antennas are. Namely, only biconical and ellipsoidal
shapes closely represent three-dimensional antennas and only for them do
mathematically complete solutions exist. The list of practical antennas that
biconical shapes approach is longer than the list for ellipsoidal ones. The
receiving antenna problem is a scattering problem. The antenna is immersed
in an otherwise steady state plane wave. Some of the incoming energy, the
extinction energy, is transferred to the scatterer and the rest continues
unperturbed; part of the extinction energy is absorbed and the rest is radiated
away as a scattered field. The objective is to analyze a biconical receiving
antenna of arbitrary cone length and half angle and with an arbitrary value of
impedance attached to the terminals. A full analysis requires knowing all
fields at all points in space. This is obtained by matching the full sets of
possible field forms in the external and internal regions, see Fig. 2.4.1, to
conducting boundary conditions at the antenna surfaces and to virtual
96 The Electromagnetic Origin of Quantum Theory and Light

boundary conditions in the open aperture. The extinction energy and

momentum, the scattered energy and momentum, and the surface charge and
current densities may be evaluated once the fields are known. The absorbed
power and the impedance at the antenna terminals follow from the surface
currents and the interior fields.
For transmission, either voltage V(0) from a constant voltage source or
current 1(0) from a constant current source is applied between the terminals
of the cones at r = b. There are no incident fields. For reception, the power
sink at r < b is a passive, isotropic energy absorber. Extinction power is
extracted from an incident, j-directed plane wave; some is scattered and
some is absorbed. It is convenient to break space into regions similar to
those of transmitting antennas. The regions are:

Sink
r<b\ O<0<7t; 0 < <f) < 271 (2.16.1)

Interior
Arms
b<r<a\ 0<6<\|/; 7 t - \ | / < 6 < 7t; 0<<j)<2rc (2.16.2)

Space
b<r<a\ V|/<8<7t-\|/; 0 <§<!% (2.16.3)

Exterior
r>a\ 0<9<7t; 0 <(b<2TX (2.16.4)

The spherical coordinate expansion for a ^--directed plane wave,

Eqs. 2.14.2 and 2.14.3, contains products of trigonometric functions,
harmonic spherical functions, and spherical Bessel functions of integer
order, with orders ranging from one to infinity. Although exterior and
interior modal products of different degrees are orthogonal, exterior and
interior modes of different orders are not: Each exterior order contributes to
all interior orders of the same degree. Interior modes are associated with
surface current and charge densities on the cones. All driven antenna modes
absorb energy and momentum from the plane wave, some of each is
absorbed and some of each is scattered away. The zero degree plane wave
modes excite TM scattering modes and TM and TEM interior modes similar
to the transmitter modes; modes known as transmitter modes. Higher degree
exterior modes excite both TM and TE scattered and interior modes. With
Selected Boundary Value Problems 91

these modes the extinction and scattered energies are equal and result in the
absorption of momentum but not energy from the wave; these are the
receiver modes.
As an example of receiver modes, at low enough frequencies a surface
current flows along the illuminated face of the antenna; it is largest near the
caps. (Detailed sketches are shown in Fig. 2.21.1). Going toward the conical
apices at each differential length some of the current terminates on local
electric charge densities, until it disappears entirely at the terminals. At the
cone-cap junction, some charge is stored and some passes through onto the
cap. Since the currents into and out of the cone-cap junctions are not
necessarily equal, an oscillating ring of charge resides there. A similar
current distribution is repeated, but oppositely directed, on the shadowed
side of the antenna. The current pattern generates a magnetic dipole moment;
the cross sectional area of the dipole is the geometrical cross section of the
cones perpendicular both to the incoming wave and to the antenna axis, in
this case the x-direction. By Lenz's law, the phase of the generated magnetic
moment is opposite that of the incoming magnetic field and results in a
scattered wave.
Just as for transmission, the charge and current densities on the cones are
functions of the interior fields and those on the caps are functions of the
exterior fields. The signs of adjacent arm and cap surface and line charge
densities may or may not be the same. The current that flows from the cone
to the cone-cap junction is not necessarily equal to the current that flows
from the junction to the cap and, as noted, differences result in a ring of
charge at the junction.

2.17 Fields of Receiving Antennas

Combining Eqs. 2.13.13 and 2.14.3 shows the TM and TE modes
respectively to be proportional 8(£ + m,2n+l) and 5(/ + w,2n), and expresses
the condition that the associated Legends polynomials satisfy the symmetry
conditions:

TM modes P f (cos 0) = - P f (- cos9)

(2.17.1)
TE modes P f (cos 9) = P f (cos 9)
98 The Electromagnetic Origin of Quantum Theory and Light

The total exterior field, the modal fields of the plane and scattered waves, are
equal to the sum of Eqs. 2.14.1, 2.14.2 and 2.14.3. A complete field
evaluation requires evaluation of the scattering field coefficients.
The interior modal structure of a receiving antenna depends upon the
symmetries both of the driving field and the antenna. For the antenna axis
parallel with the direction of polarization, the antenna implementation
retains the symmetry of the external fields in the internal region. As was the
case for a transmitting antenna, finite interior fields require the multiplying
coefficient of all functions yv(a) to be equal to zero for v > 0. Coefficients of
the JV(CT) functions are nonzero for both TM and TE modes. The symmetry
of the interior TM and TE modes remains the same as the exterior symmetry,
with undetermined coefficients respectively defined to be T™ and A™. A
full solution requires evaluation of the functional relationships between
internal coefficients T™ and Am and the scattering coefficients af and $f.
Combining all the above for the interior fields, the zero degree terms
have the same form as the transmitter terms, Eqs. 2.7.1, and combine with
the higher degree terms requirements to provide the expanded equation set:

w
2_j cos mty - / ' ^ sin m§ v(v+i)r v ^^M^(cose)
v>0 e mo a

riH r =£ ^ c o s m§ - /2J sin m§ v ( v + l ) A ^ ^ i L ^ ( c o s e )

v>0 mo
r\GW{r)
Ee = 2^ cos mfy - /'2^ sin m§ VJVV / d e
2%rsin 6
v>0
oo oo oo

2^cos m§ - / ^ sin m§
v>0 me sin 6
(2.17.2)

^-
^ =-
'
i 27trsin 9
v>0
^ c o s m§ - i^j sin m<\>rvjv(cr)-^-
mo
oo oo

A
2^ cos m§ - I2J sin m§ vJv(CTj^-7
v>0 sin 9
Selected Boundary Value Problems 99

^ c o s m§ - / ^ sin w<() r v M C T ) - T - 7 -
v>0 sin6
oo oo

^ cos m§ - i^j sin m§ AvJv(°)

VJVV I —
d 0
v>0 mo

^ c o s m§ - / ^ sin m§ vMCT)
v>0 me
sin9
oo oo
aL,
\m+ i „\ v
5^ cos m§ - / ' ^ sin m§
v>0

As was the case for transmission, in the limit as b goes to zero the only
nonzero terms just off the r = b surface are the TEM components of Ee and
H,]). The TEM fields guide the energy through the interior region. Repeating
the procedure of Eq. 2.9.6, evaluate the integral using Eq. 2.17.2:

V
{ dGH^I m=0

The result is the algebraic equation:

V(a) 2/G
£D?(j, + P°h,) (2.17.3)
a Y R (#) t

2.18 Boundary Conditions

Several boundary conditions have been built into field Eqs. 2.14.2
and 2.14.3: Rotational symmetry requires m to be an integer and regularity
of the zenith angle functions on the exterior axes, r > a, requires integer
order Legendre functions of the first kind. The limiting condition as the
radius becomes infinite requires spherical Hankel functions of the second
kind. In the interior, regularity of the functions at r = b requires the
100 The Electromagnetic Origin of Quantum Theory and Light

coefficients of all negative order Bessel functions to be zero. The boundary

conditions still to be applied are:

1. Interior region, b < r < a , 9 = i|/ and 6 = 7t-V|/

On the cone arms the tangential components of the electric field intensity, E r
and E<j), and the normal component of the magnetic field intensity, He, are
zero.

2. Exterior region, r = a, 9 < \|/ and 9 > TI-\|/.

The tangential components of the electric field intensity, Ee and E^, and the
normal component of the magnetic field intensity, Hr, are zero.

3. Boundary, r = a, \|/ < 9 < 7i-V|/

All fields are continuous through the virtual interface between internal and
external regions.

To satisfy the first boundary condition, note that the tangential

component of the electric field and the normal component of the magnetic
field are equal to zero at the surface of the cone. From Eqs. 2.17.2, the sums
are equal to zero for all interior radii only if:

M*(cos\|f) = 0 and dLv(cos9)/d9| e = ¥ =0 (2.18.1)

For each degree, an infinite number of orders satisfy Eq. 2.18.1. Figure 2.7.1
includes plots of the first few values of v and \\i for functions with m = 1 that
satisfy these boundary conditions. In the limit as \|/ approaches zero the first
solutions for the odd and even functions occur respectively at v = 2 and 3.
To satisfy the second boundary condition from Eq. 2.14.3 it is necessary
that:
0<V|/ and 0>7i-\|/
(2 18 2^
Ee(/fo,0,<|)) = 0 and E^(/fo7,9,<j)) = 0

To satisfy the third boundary condition, with 5 a vanishingly small

positive number, it is necessary that:
Selected Boundary Value Problems 101

\|/<6<7t-V|/
E(/fo-5,8,(|>) = E(/fe7+5,e,(t)) (2.18.3)
H(/fez- 5,e,(b) = H(/fcz + 5,e,(b)

The fields on the left side of Eq. 2.18.3 are those of Eqs. 2.17.2. The fields
on the right side of Eq. 2.18.3 are the sum of Eqs. 2.14.1, 2.14.2, and 2.14.3.
Desired algebraic equations are most easily obtained using the second
and third boundary conditions to construct the four integral equalities of
Eqs. 2.18.4 through 2.18.7. In addition to these four equalities, the process is
to be repeated with a similar set of integral equations after replacing sinm(])
by cosm(|) and cosmtj) by -sinm<j). Although the zenith angle limits on both
integrals would be \|/ to n-\\t the second boundary condition permits
changing the limits to 0 to n for the electric field components.

Jt-V|/ 271 f ,„OT p« }

/ sinOdef d(j)|E e —^-cos(^(j))-E ( .^-^-sin(^<j))l
de sine
v o I sJ ,
a=^-8 (2184)
71 2TI I ,pm p/n 1
= J sin8d0 j d<MEe —— cos(,w<|))-E(j) —— sin(/w<Jm
de sine
o o I J
a=/ca+o
m m
n-iif 2JI f A\x yi 1
f sin9d0 f dMHL -cos(/«<t>) + H e — ^ s m { m t y ) \
. d0 sinG
b
*" (2.18.5)
n-y 2n dM" , , „ mtA'
f sinedefd^ilL, ^-cos(/wj)) + H e — ^ s m ( m ^ ) \
n d0 sin0

j sin0d0 J d(J)jEe —— cos(/«)>)-E^ —Z-sinmtyi

Sln9 d6
V 0 I Jc=Af_5
->„ r i (2.18.6)
= J sin0d0J d < M E e — — c o d m § ) - E 6 — — s m m § \
Sm9 d6
n n
102 The Electromagnetic Origin of Quantum Theory and Light

J" sinede|d6jH (|) ^-^cos(/«(|)) + He—^sin(/rc<|>H

Sin 6 J
° ^ r ^ ° = ^ (2.18.7)

= f sin9def d9JH(1)^-^-cos(w<|)) + H e —"sm(m§)\

„ sinG d0

Carrying out the integral operations of Eqs. 2.18.4 through 2.18.7 with the
similar set obtained by replacing sinmij) by cosm§ and cosmfy by -simwcj)
results in the four linear equations:

£(f + l)j-Ffl« + £(£ + l)pfh-Ffl«

m (2-18.8)
= €(£ + l)£r-jX v -^^P,5( W ,0) + 2 W Pf£A-j p L-
7Ttf
VO p

v(v+l)r v w j v K vv =Xn(n+l)F n " j n K nv + £n(n+l)F*Pj:h n K n v (2.18.9)

n n

*(* + l)G7j / I« + *(i + l)G?a2V« =Sp(p + 1 )ApJp^ P ( 2 - 18 - 10 )

p

oo oo

p(p + l)A£j;i pp =p(p + l)X Grj;i r p + p(p + l ) E G ^ a ^ I r p

r r
(2.18.11)
oo oo

n n

All but five terms are known in Eq. 2.18.8 - 2.18.11: V{a), af, Pf,
Ap', and T™. Problem solution requires evaluation of each of them. With
transmission there were but three unknowns: Yj(^),P^, and T v .
Selected Boundary Value Problems 103

2.19 Zero Degree Solution

Since only the zero degree modes carry absorbed power, as discussed in
Sec. 2.7 it is convenient to analyze them first. Equation 2.17.3 and the m = 0
portion of Eqs. 2.18.8 through 2.18.11 are

a V
W)no '

i(e + I K P > ; I « = -t(t + I)F?J*IW - ^ ^ P^ + i(t +1)£ r v jX v

na v

v(v + l)r v j v K v v = £ n(n + l)F n °j n K nv + £ n(n + l)F n °p> n K• •nn v

(2.19.1)

The transmitter coefficients p^ and the receiver products F^ P^ play similar

roles: both sets of coefficients multiply TM fields that emanate from the
antenna. Although Eq. 2.19.1 and Eqs. 2.9.1 to 2.9.3 are similar in form, a
different approach to problem solution is helpful.
A case of special interest is an equated load. For this case, the receiver
antenna load impedance equals the input impedance the transmitter antenna
applies to incoming power. To analyze the case, adjust the driving field so
V(0) = a. The antenna parameters given in Eqs. 2.19.1 are then the same as
those of the transmitter case of Eqs. 2.9.1 to 2.9.3. Since identical equations
give identical solutions:

^ / = J^+P% (2.19.2)

Equation 2.19.2 shows that the relative phases and magnitudes of the
transmitted and scattered fields per mode, £, are not the same. Several
coefficient values are tabulated in Table 2.19.3. Values are calculated using
the numerical results of Table 2.9.1 and Eq. 2.19.2 for the special case
YR(a) = YT(a), ka = 2, \|/ = 5°, and V(0) = a.
104 The Electromagnetic Origin of Quantum Theory and Light

£ real part, (3° imaginary part, 0? x order of magnitude

1 -1.0073 +0.15175

3 -1.3064 -0.14110 10,-2

5 -5.4652 -1.5762 10,-3

7 -1.1887 -3.4642 10_

Table 2.19.1 Values of (3^> for the special case of an equated load, ka = 2,
V = 5°, and V(0) = a

Comparison of the transmitting and receiving equations shows that

r° = r (2.19.3)

That is, the internal field coefficients for the two cases are the same.
Equations 2.19.2 and 2.19.3 contrast the relationships between the
transmission and reception coefficients. Comparing Eqs. 2.9.6 and 2.19.1
shows that the termination admittances Y(a) for the two cases are identical.
Equation 2.6.16 translates the admittance Y(r) to the terminals and confirms
that the terminal impedances of the antenna as a transmitter and as a receiver
are identical.

YR(0)=YT(0) (2.19.4)

For an arbitrary but known load impedance, the solution procedure is to use
Eq. 2.6.17 to solve for Y(a) then combine with Eqs. 2.19.1 to 2.19.3 to
obtain the linear equation

-¥titIu-nl(e+l)YK(a)iF^ +
^ v(v + l)j v K vv
P°V 2
F?h*,L 2r)/G Pi ^ 0fto Y Y o n(n+ l)j*K^vKnv 0
K
M'+1)VMZ^ ?hU v(v+l)jvKvv P h
""
(2.19.5)
Selected Boundary Value Problems 105

Symbols lu and K^v et al. represent integrals listed in Tables A.22.1 and
A.23.1. The equation form is the same as Eq. 2.9.5, and the solution
technique is the same. Since all terms in Eq. 2.19.5 are known except (3$ it
may be solved first for P$h^ and then for P^. Once the P$ are known,
Eq. 2.19.1 may be used to solve for V(a). The value of T® may be obtained
using Eq. 2.19.3. The zero degree solution is then complete.

2.20 Non-Zero Degree Solutions

To find the solution for m > 0 solve for the exterior field parameters af and
P^ using Eqs. 2.18.8 through 2.18.11. This is most easily done by rewriting
them in the forms:

" ( " + l)JnJ*yK nyK.iv

-^+l)Ffj-,I„ + ^ + l ) 2 5 X v(v+l)j v K v v
oo oo : fj#?l oo oo j J (jm\ pm
2m?r 2 : S G f J p y p rp
- 4m2P? S ^ X JpJ 1 P
" [ "
p r Jp^PP P n R P + Ujp^pp
(P?V) =
({e+ l)h*Ffl w
2
+2mPriiG7^^{a7K)-^ prii:K 'p[ p" {KK)
p r Jp"r'pp p n f \ P + IjJp'pp

V
v n V V + UJvK-w V
'
(2.20.1)

-i(+ l)Gfj,I„ + £ £ G ? P ( P + W ' P 1 »

P r Jp*pp

(afhe) = -2^X2>;r JpJl 7 p p n

e(e+\)G?ie P n Jp!pp

P r Jphr'pp p n Jp'pp
(2.20.2)

Equations 2.20.1 and 2.20.2 have the general algebraic form:

106 The Electromagnetic Origin of Quantum Theory and Light

oo oo

r
" (2.20.3)
OO OO

y£ + I X n x n + S R ^ y r + = Al
n r

All needed integrals are listed in Tables A.22.1 andA.23.1 and all other
parameters are known. The sums of Eq. 2.20.1 and 2.20.2 may be truncated
and solved concurrently for coefficients af and $f, from which T™ and
A^ follow directly. Knowledge of these four parameters provides the total
solution of the fields around a receiving antenna.
Although the integral of products of Legendre functions of integer and
noninteger orders, M v (cos0)P^(cos0) for example, are largest if v is nearly
equal to £, none of them vanish. Therefore cross coupling exists between all
modes of the same degree, m. Coupling between af and $f terms show that
all modes, even and odd, of the same degree are coupled. That is, each TM
mode interacts with all other TM and TE modes, and vice versa: Each value
of fif and af depends both upon all values of P^ andaf, but are
independent of Pjj? and a^ for m £ p.
Since for m = 0 and \|/ approaching zero the order approaches an integer
value quite slowly, the integrals of cross modal terms remain significantly
large, and, therefore, coupling is significantly large even for \\i near zero.

2.21 Surface Current Densities

A conducting boundary condition is that the surface current density, I, in
amperes per meter is related to the magnetic field adjacent to it by the
vector-phasor relationship

I = /?xH (2.21.1)

Unit vector n is normal to and outbound from the conductor. Knowledge of

the field coefficients permits the calculation of all antenna currents.
Figure 2.21.1 illustrates surface current patterns for the lowest order exterior
Selected Boundary Value Problems 107

modes, (t,m) = (\,0) and (1,1) and the three lowest order interior ones,
modes (v,m) = (0,0), (l+,0), and (2+,l). The exact interior modal number
depends upon the value of \|/: with 5° cones the modal numbers 1+ and 2+
are respectively 1.444 484 and 2.022 029. The figure depicts current patterns
for a small antenna, a < n/4. A plane wave is incident from the left, with a z-
directed electric field intensity. In the interior the TEM mode fields, see
Eq. 2.7.1, are maximum at r = b. The current of mode (l+,0) has rotational
symmetry around the cones. The current of mode (2+,l) is in phase with
mode (l+,0) on the front face and out of phase on the back face; it is equal to
zero in between. Both are TM modes, both are zero at the origin, both are
large near r = a, and both are zero at the sink. The current of mode (l+,0)
produces a z-directed electric dipole moment. The current of mode (2+,l)
produces a ^-directed magnetic dipole mode, with a magnetic field phased
according to Lenz's law. In the exterior, the currents of the two lowest
modes, (1,0) and (2,1), produce respectively TM and TE fields. The cap
current density of mode (1,0) is 9-directed and zero at the center, and the
current density of mode (2,1) is x-directed and maximum at the center.

2.22 Power
The time average power on the surface of a virtual surface of radius a/k that
circumscribes the antenna is

P = % J d<t>J"sinedeRe(Nr) (2.22.1)
k
0 0
For receiving antennas, the time average received power is equal to the
negative of the real part of Eq. 2.22.1, after inserting the coefficients
evaluated in Sections 2.19 and 2.20. The power and the cross sections were
calculated in Sections 2.14 and 2.15. The normalized absorption cross
section CAB/CGE> see Eq. 2.15.3, is:
108 The Electromagnetic Origin of Quantum Theory and Light

(1,0)
<u,0 Cap
currents

f
(0,0) (2+.1)
Interior
k currents

Figure 2.21.1 Receiving Modal Surface Currents, Wave Incoming from Left.
In the interior region the (0,0) current is maximum and the (l+,0) and
(2+,l) currents are zero at r = 0. The (l+,0) current is unidirectional and
rotationally symmetric around the arms. The (2+,l) current is bi-directional,
creating a magnetic moment directed in accordance with Lenz 's law. In the
exterior, the (1,0) cap currents are 6 directed, rotationally symmetric and
zero at the midpoint. The (2,1) currents are x-directed with a maximum at
the midpoints.

oo t - \ oo l-\
_.
(2e+i)(e-m).\{e+m)\\
-AB
2 2 XX+XX Re(p?) + Pffjf
CGE ka t(t+\){Z- m-\)\\{Z + m-\)\\
.to me te mo
m2'(2l+l)(l-m-i)\\(e + m-\)\\ •
ka 2 2 xx+xx
to mo te me
e(e+i)(t-m)\\(£+m)\\
Re(ocf) + a mnm*

(2.22.2)
Selected Boundary Value Problems 109

As was shown for a sphere, Sections 2.2 and 2.3, the portion of Eq. 2.22.2
proportional to -Relocf+Pf) represents the extinction power extracted
from the wave, and the portion proportional to c c f a f and P ^ f represents
power scattered away from the antenna. The sum, expressed by Eq. 2.22.2, is
the power absorbed by the antenna.
The incoming wave transfers both momentum and energy to the antenna.
Since the incoming plane wave is j-directed, linear momentum is transferred
to the antenna in that direction. The force on the scatterer is related to the
momentum transferred as:

K.y = —V(linear momentum) (2.22.3)

d/ '

The net force applied to the antenna follows from the rate of momentum
absorption and scattering, and is equal to:

a2 2n
F
y = ^ / s i n N r sin z 6d8 (2.22.4)
c/r

By Eq. 2.15.7 the normalized force due to the extinction power is

\j(m)(2( + l)(l - m)\\{i + m)\\

Re(pf)
ff>X_ = __8E_ . to me (e mo t{£ + \){l-m-\)\\{t + m-\)\\
r ,22
m2(2£ + \)(t-m-\)\\(e + m-\))\\
Re(af)
.to mo te me t(i + \)(t-m)\\(l + m)\\

(2.22.5)
By Eq. 2.15.6 the normalized scattering force is:
110 The Electromagnetic Origin of Quantum Theory and Light

*W1\ U ( ^ - ^ ) ! ! ( ^ + ^+2>! 1
v y
( ^ + 1 ) 2 ( ^ - W - 1 ) ! ! ( ^ + OT-1)!
(^+i
1 1 + 1 1
^ . 1 * . Q * W 1I \ U(»X<-">'(< + ^ !
\ £o me £e mo •(CTrr+pfpr, )-' ( ^ - O T - 3 ) ! ! ( ^ + OT-1)!
2

L*w^»* i a ^ n , ^ - i H ^ + 1 ) ^ - ^ - 1 ) ! ! ^ + ^ + 1 ) ! !
a f a ^ + 1 + a f a£+1 I -j- —
f[>SC _ 4
2„2 te me )
CQE fra' + a^a^_i +a^ a^_i -z
V ^ I (2(£-m-2y.\(e + m)\\ )

v
£ o me tefflo ;
\e2(t+i)(e-m-i).\(e+m-i).\j

ii +ii \a^r^^nri^M)^m-xW'm+x}"
U«° temr \e2(e+\)2(,e-m-2)»{e+m)»)
(2.22.6)
For the special case m = 0, the scattering cross section and normalized
force terms are:

Cgc = ^ _ y (2l+l)l!! 2 g0p0*

(2.22.7)
2 l
CGE * V £ ^+1)[(^-1)!!] '

+ 2 ! !
^ROR 1 * 1
w f + i ++ Rw° * RP ^ ^+ i / 7 ^— ^ )
e+ne-w
+(mu+vrvu)r(*-3)!!(*-l)!!
e\r (2.22.8)
r ,2 2 ^

2(21+l)l!!2
^arp?+ajpr) l (l+l) [{£-l)»]\
1*«0 , „ 1 « 0 * \
2 2

If the antenna is electrically small and only the dipole terms are significantly
large, Eq. 2.22.8 shows that:

^«. ' („rpf+B|pr) (2.22.9)

L Ly
CGE ka '
Selected Boundary Value Problems 111

Although the energy density-to-linear momentum density in the

incoming plane wave is c, as discussed in Section 2.3 the received energy-to-
momentum ratio satisfies the relationship:

Received Energy ^ (22210)

Received Momentum

This is in contrast with a transmitting antenna. When transmitting, power is

radiated over a spread of angles and the average value of the cosine of the
angle can never be greater than one. Therefore, the transmitted energy-to-
momentum ratio obeys the relationship:

Transmitted Energy ^ }

Transmited Momentum

References
D.M. Grimes, "Biconical Receiving Antennas," J. Math. Phys., vol. 23, pp. 897-914
(1982)
D.M. Grimes, C. A. Grimes, "Transmission and Reception of Power by Antennas,"
in T. W. Barrett, D. M. Grimes, Advanced Electromagnetism: Foundations,
Theory and Applications, World Scientific Publishing (1995) pp. 763-791
H. Hertz, Electric Waves: Researches on the Propagation of Electric Action with
Finite Velocity Through Space (1893) Translated by D. E. Jones, Dover
Publications (1962)
G. Mie, "A Contribution to Optical Extinction by Metallic Colloidal Suspensions,"
Ann. Physik. vol. 25, p. 377 (1908)
W.K.H. Panofsky, M. Phillips, Classical Electricity and Magnetism, 2 nd ed.,
Addison-Wesley (1962)
S.A. Schelkunoff, Advanced Antenna Theory, John Wiley (1952)
S.A. Schelkunoff, Applied Mathematics for Engineers and Scientists, 2nd ed., Van
Nostrand (1965)
W.R. Smythe, Static and Dynamic Electricity, 3rd ed., McGraw-Hill (1968)
H.C. Van de Hulst, Light Scattering by Small Particles, John Wiley (1957)
This page is intentionally left blank
 3. Antenna Q

3.1 Instantaneous and Complex Power in Circuits

It is commonplace to discuss power and energy in radiation fields using
complex numbers. To make a critical examination of the procedure, since
one-dimensional electrical circuits are simpler systems than three-
dimensional electromagnetic fields, we begin with electrical circuits.
Consider the time-varying power and energy of an electrical circuit that is
driven by a sinusoidal, steady state source. With % and C, representing circuit-
dependent phase constants, the input voltage and current to an electrical
circuit are:

v(/)= V 0 cos(w/-x) and i(/)= I 0 cos(oo/-£) (3.1.1)

With this notation, either Vo or Io can be the independent variable with the
other being the dependent variable. Both are real, time-independent
quantities. The power at the terminals follows from the force laws, and is the
simple product:

p(/) = v(/)i(/)
_1
= r V 0 Io{cos(£- x) + cos(C+ x)cos(2co/)+ sin(C+ x)sin(2co/)}
2
(3.1.2)

Trigonometric identities may be used to transform Eq. 3.1.2 into the more
useful form:

P(') = 2
-z V 0 I 0 {cos(C - x)[l + cos(2co/- 2$)] + sin(C - y) sin(2oo/- 2^)}
2
(3.1.3)
113
114 The Electromagnetic Origin of Quantum Theory and Light

Although either a plus or minus sign could be placed in front of the sin(£-%)
term, a positive sign is convenient and leads to no loss of generality.
It follows from Eq. 3.1.3 that the three numbers needed to characterize
the power are the product Volo, the phase difference (£-%), and the phase
angle £,. The equation also shows that the term proportional to
cos(£-%)[l+cos(2otf-24)] is zero twice each field cycle, it is never negative,
and it describes the time average energy flow into the circuit. The term
proportional to sin(£-%)sin(2GM-2£) is in time-quadrature with the first term
and changes sign twice each power cycle; it describes the lossless,
oscillatory energy flow between the circuit and the energy source and is not
associated with a time average energy flow. In many cases, instantaneous
phase £, is irrelevant and appears only as unwanted clutter. For such cases,
the quantities ^-^and Volo determine the important properties of the power
and no other information is either needed or desired. For such cases, phase
factor % is suppressed and real power, pr(/), and reactive power, px(0> are
defined by the equations:

Pr(') = T V 0 I 0 cos(C- x)[l + cos(2<o/)]

\ (3-1.4)
P x ( ' ) = j V 0 J 0 sin(C- x)sin(2w/)

Since only two pieces of information are included and since complex
numbers have two places available to carry information, this power may be
conveniently described by complex numbers.
To restate the same physical situation using complex numbers, write the
input voltage and current in phasor form:

V(/)= V o e ^ - x ) l(/)= I o e * ^ ) (3.1.5)

Equations 3.1.5 differ from Eqs. 3.1.1 in that virtual terms, the imaginary
parts of Eqs. 3.1.5, have been added to the phase of each variable. The real
parts of Eqs. 3.1.5 are equal to the actual values of Eqs. 3.1.1.
Equations 3.1.5 are used to form the product:
Antenna Q 115

Pc=|v(/)I(/)* (3.1.6)

The real part, Pr, and imaginary part, P,, of Eq. 3.1.6 are:

(3.1.7)
= 2 V 0 I 0 [cos(C- x)+ /'sin(C- x)]

Comparison of Eq. 3.1.3 with Eq. 3.1.7 shows that the latter contains all
information except the suppressed phase factor. The real part is equal to the
magnitude of the time-average input power and the imaginary part is equal
to the magnitude of the oscillating power. In this case, both real and
imaginary parts of the power represent actual quantities. The phase-
quadrature difference between real and reactive powers is indicated by an "/"
in Eq. 3.1.7. Equations 3.1.4 and 3.1.7 are but different notations for the
same physics. Neither contains information about phase angle £,.
By definition the Thevenin circuit input impedance elements are

R = -4-cos(C-x) and X=-^sin(C-x) (3.1.8)

Combining Eqs. 3.1.7 and 3.1.8 shows that the complex power may be
expressed as:

Pc = | l o l ! ( R + > X ) (3.1.9)

With Eq. 3.1.9, Io has been modified to a complex number that includes
factor exp(-z'Q.
Consider next the case of two isolated electrical circuits. It is easy to
show that the total power is the simple sum of the power in each circuit, as
described by Eqs. 3.1.2. The sum is:
116 The Electromagnetic Origin of Quantum Theory and Light

2
p(^)=Svk(/)ikW
k=l
2
1
T
2
E v kIk{cos(^k-Zk) + cos(Ck + Xk)cos(2co/) + sin(Ck + xk)sin(2co/)}
k=l
(3.1.10)

To learn how to express the information contained in Eq. 3.1.10 using

complex notation begin by rewriting it in a form similar to that of Eq. 3.1.3:

1 2
pM = r S [VkIkcos(Ck - Xk )][l + cos(2u)/- £)] + K 12 sin(2to/- £)
2
k=l
(3.1.11)

Insisting that Eqs. 3.1.10 and 3.1.11 be identical and solving for K12 results
in the equality:

K? 2 = V 1 2 I?sin 2 (C 1 -Xi) + V 2 2 lisin 2 (C 2 -X2)

+ 2V 1 V 2 [-sin(xi-X 2 )sin(Ci-C2) + sin(Ci-X 2 )sin(C 2 -Xi)]
(3.1.12)

Consider the special case where one of the two equalities apply:

Cl = £ 2 or 5Ci = X2 (3-1-13)

For either case, Eq. 3.1.12 simplifies to:

K 12 = V 1 I 1 sin(Ci-Xi)+V 2 I 2 sin(C 2 -X 2 ) (3.1.14)

Combining Eqs. 3.1.14 with Eq. 3.1.11 shows that:

2 2
p ( / ) = : S V k I k {cos(C k -Xk)[l + cos(2o)/-^)] + sin(C k -Xk)sin(2a)/-^)}
2
k=l
(3.1.15)
Antenna Q 117

According to Eq. 3.1.15 if either of the two conditions of Eq. 3.1.13 is met
the powers of the two circuits combine by simple addition. Within
interconnected electric circuits, the Kirchhoff circuit laws assure that one of
the conditions of Eq. 3.1.13 is met, either between circuit nodes or along
circuit branches. For these special cases, the complex power is the simple
sum over the power of the different circuit elements:

Pc = k Vkl^"*) (3.1.16)
z
k

As we shall see, different modes of multimodal radiation fields do not meet

the conditions of Eq. 3.1.13 and therefore Eq. 3.1.16 does not apply.

3.2 Instantaneous and Complex Power in Fields

To analyze power and energy about an antenna it is enough to consider only
antennas with rotational symmetry about the z-axis. With this choice the
field solutions are of degree zero and there is no dependence on the azimuth
angle. Since, as will be shown, the essential points of interest depend only
upon the radial field functions, and since the radial field functions are
independent of degree, results are general and apply to a full multipolar
expansion. Written in phasor form, but keeping the retarded time phase
dependence, in terms of the letter functions of Appendix A.26, the general
form of the field expansion terms, Eq. 1.12.9, is:

o%= E F ^ + l)[B £ (a)+/A^(a)]P / (cos0)e-' i

oo

a2TiHr = - £ G ^ + l)[B^(a)+/A^(a)]P^(cose)e~*

aEe=£FjD,(a)+^(a)]^«le-

W^=£FjMa)-*,(a)]^^
e=i d0
118 The Electromagnetic Origin of Quantum Theory and Light

° V £ G , [ A , ( a ) - * , ( a ) ] ^ ^ e" / a

or,He=-£G,[D,(a)+/C,(a)]^!i e ia

Every possible radiating antenna field with rotational symmetry about the z-
axis may be fully described by picking appropriate choices of multiplying
coefficients F^ and G^.
After making use of Eq. 3.2.1 and Table A.22.1.6, the surface integral of
the complex Poynting vector evaluated on a circumscribing, spherical
surface of radius o/k is:

(F,F;+G^;)[A,(a)D,(0)-B,(a)C,(a)]
«•>-*•-£!£§ +'" (P/F/ -GeG*e)[Ae(o)Ce{a) + Bt{a)Dt(a)]
(3.2.2)

The absence of cross product terms between TM and TE modes shows that
the two modal types act independently. The sign of the imaginary term
depends upon whether the field is TE or TM; if both are present and of equal
magnitude the net is zero. Since each modal coefficient is multiplied by its
own complex conjugate, a phase difference between sources has no affect
and all modal phase factors are suppressed.
Examination of Eq. 3.2.2 shows that the two numbers needed to evaluate
the modal power are weighted sums over (A^D^-B^C^) and
(A^Q + B ^ ) . By Table A.26.2.8, ( A | D r B { C ^ ) is equal to one for all
orders. The second term is defined to be:

yt(a) = Ae{a)Ci{o) + B^(rj)D^(a) (3.2.3)

Values of jfip) are listed in Table 3.2.1.

Table 3.2.1 shows that the magnitude of y^(c) increases precipitously
with small and decreasing values of a and with increasing modal number t.
All signs in Table 3.2.1 are the same and y((a) is a monotone decreasing
Antenna Q 119

function of a. Taking Y^(CT) as a

measure of reactive power, the surface
reactance has the same sign for all radii: capacitive for TM modes and
inductive for TE modes. This is in marked contrast with the numerical
analysis of center-driven biconical antennas where the sign of the reactance
of a TM antenna at the input terminals is primarily a function of normalized
cone length. This emphasizes, see Fig. 2.9.1, that changes in the sign of the
input reactance versus antenna radius for TM sources are due to the
transmission line character of the antenna arms and not to intrinsic properties
of the radiating surface.

l
/ \
Yi(<*) = - —
a
i \ 18 3
a a
/ x 675 90 6
Y3(°) = - - 7 ~ - 5 - - 3
0 a 0
i v 44100 4725 270 10
Y4(cr)= 9 7 5 3
a a a a
( v 4465125 396900 18900 630 15
Y
5(°)= 11 Q n7 n5 Ji
,/v 648,336,150 49,116,375 1984500 56700 1260 21
Y6lCTJ- ,3 n 9 7
a 0 a a a5 "r7
Table 3.2.1 Radial Dependence of y/(cy)

3.3 Time Varying Power in Actual Radiation Fields

The actual fields, from which the phasor fields of Eqs. 3.2.1 follow, are
listed in Eq. 3.3.1. The driving source varies with time as COS((JL>0. If m e
constant coefficients of Eqs. 3.2.1 are entirely real or entirely imaginary,
respectively the upper or lower set of terms within the square brackets in
each field component of Eq. 3.3.1 applies. Although the absolute phases of
the elements are not important to our results, the phase differences between
modes are. Since by proper adjustment of the time origin all phase
120 Tfe Electromagnetic Origin of Quantum Theory and Light

relationships are expressible as sums over the upper and lower terms results
of analyzing this set of field equations are general.

B^ cos^/yp) - Kji sin(u)/>)

a 2 E r =£F/0?+l) P^(cos0)
Ae cos(co/^) + B^ sin(co/^)
B^ cos(a%)- Ag sin((o/tf)
a 2 TiH r =-£G^+l) P^(cos0)
A^ cos(co/yp)+ B^ sin(co/^)
D^ cos(co/yp)- C^ su^co/^) dP^(cos9)
aE e =XF,
Q-H cos^/^H- D^> sh^co/^) d0
(3.3.1)
oo A^ 003(0)/^)+ B^ sin(co/^) dP^(cos6)
- B ^ cos(co/^) + At sin(co/tf) de
oo A £ cos((o^) + B^ sin(co/^) dP^(cosG)
- B ^ cos(co/^)+A^ sh^co/^) dG
-D^ 005(0)/^) + Ce sin(co/^) dP^(cos9)
OTlH0=XG
-C^ cos(co/^) - Df sin(co/^) de

Using Eqs. 3.3.1 to evaluate the radial component of the time-dependent

Poynting vector then integrating over a constant radius surface centered at
the origin gives the surface power:

^).K^f,g^ [A^D^[l ± cos^co/^)] - B^Q[l + cos^co/^)] 1

(3.3.2)

The upper or lower signs respectively apply to the upper or lower terms in
the square brackets of Eqs. 3.3.1. The sign choice depends upon the phase of
the modes but does not depend upon the TM or TE character of the modes.
Hence, in contrast with results obtained using phasor fields, Eq. 3.3.2
depends upon the relative phases of the driving modes.
Examination of Eq. 3.3.2 shows that it contains three separate
parameters: weighted sums over A^D^, B^C^, and A^C^-B^D^. For what
follows it is necessary to work with functions with a zero asymptotic limit at
infinity. For that purpose, define o^(o) and (^(o) to be:
Antenna Q 121

oc f (a) = ( A ^ + B ^ ) - ( - l ) f
(3.3.3)
P/(a) = ( A / Q - B / D / )

Combining Eq. 3.3.2 with Eq. 3.3.3 shows that:

2 1± (-1)' cos(2(B/>)
^MM[^ } ±[a/(o)cos(2to/ff)-P/(o)sin(2a)/ff)]
(3.3.4)

Within the curly brackets of Eq. 3.3.4, the envelope of the first term is
independent of distance from the antenna. Functions oc^(a) and P/or) are
represented by alternating series and oscillating functions of distance.
Functional values of ae(a) and P / a ) are listed in Tables 3.3.1 and 3.3.2 for
£ = 1 through 6.
The first term in Eq. 3.3.4 is the real power pr(<V#) where:

Pr(<v*)=A i TZTAW+G^2Il ±
H' cos 2w
( ^) (3.3.5)

This equation describes power that travels ever outward at speed c in the
form of periodic, trigonometric pulses. There is no time-independent, radius-
dependent phase term and the magnitude does not approach a limit at infinite
radius.
The distance dependent power terms in Eq. 3.3.4 are given by Pi(<Vj?)
where:

Pi(^^) = ±^£4^||[F/+G/][a^a)cos(2(o//?)-p/(o)sin(2(o^)]

(3.3.6)

As may be seen from Tables 3.3.1 and 3.3.2, the maximum of the envelope
for each term occurs at the antenna surface and goes asymptotically to zero
at infinite radius.
122 The Electromagnetic Origin of Quantum Theory and Light

«l(o-) = —
o
/ x 36 18
a
2(CT) = - T - 3
/ x 1350 720 72
a
3(°)= „6 „ 4 + 2

/ x 88200 49350 6000 200

a +
4(°)= „6«
a c a4 a2
/ x 8,930,250 5,159,700 699300 31500 450
a + +
5(°)= .0 „8 „6 ~ 4 „2
a a a a o
, x 1,296,672,300 766,215,450 111,370,140 5,900,580 123480 882
a 6 ( 0 ) = + +
„" „10 „8 „6 4 " "a2
a a a cr o

Table 3.3.1 Radial Dependence of oc^(cr)

Pl(G) = _ J 2
cr o
aM ClT )\= _18
_ +
33
_ _ 6
_

a i x 675 1250 276 12

P3(CT)=-—+— r+"^
„ / x 44100 83475 20220 1300 20
a
a a a a
„ / x 4,465,125 8,533,350 2,201,850 169470 4425 30
Ha)= n - + — 9 7—+—5 r+—
„ /x 648,336,150 1,247,555,935 335,975,850 28,797,930 961380 12201 42
+ + +
M°)= in— —zu ~9— ——i zr- ^r-^-a
a a c a o o
Table 3.3.2 Radial Dependence of P^(o)

3.4 Comparison of Complex and Instantaneous Powers

In the discussion to follow only TM modes are analyzed. The result carries
over in the same form with TE modes, only the sign of the imaginary part
changes.
Antenna Q 123

With electric circuits, the complex power form of Eq. 3.1.4 is determined
by Eq. 3.1.3 and, conversely, Eq. 3.1.3 is partially determined by Eq. 3.1.4.
In a similar way, the time-dependent field power of Eq. 3.3.2 leads to the
complex power of Eqs. 3.2.2. Equation 3.3.2 may be put in the form:

Pc(°V/>) = " ^ £ T ^ T T f / i t 1 ± cos

( 2 ^ -2$)] + Y,(<i)sin(2(B/> -2$)}
(3.4.1)

The instantaneous power expression for the identical set of electromagnetic

fields is given by Eq. 3.3.4, and repeated here for TM modes only:

l±(-l)^cos(2(0^)j
(3.4.2)
^ =^£(27^ ±[a^(o)cos(2co/y?)-P^(a)sin(2co/y?)]

With these two descriptions of the same energy flow, Eqs. 3.4.1 and 3.4.2,
the curly brackets of the two equations are multiplied by identical factors but
contain, respectively, two and three time-dependent terms.
The first term of Eq. 3.4.1 is the real part of the complex power. It does
not go to a limit at infinite radius, it is equal to zero twice each field cycle,
and it is never negative; it describes a unidirectional energy flow away from
the source. The gamma power term of Eq. 3.4.1 is in phase quadrature with
the real power and, by definition, is the reactive part of the complex power.
It goes to zero in the limit of infinite radius; at each point it oscillates
between equal negative and positive values and hence describes radially
directed, alternating power. The time-dependent terms contain identical
mode- or radius-dependent, time-independent phase factors.
The first term of Eq. 3.4.2 is the real power. Like its counterpart in
Eq. 3.4.1, it does not go to a limit at infinite radius, it is equal to zero twice
each field cycle, and it is never negative. It, too, describes a unidirectional
energy flow away from the source. The real power and oc^(a) power are in
time phase, and both are in time quadrature with (3^(a) power. Both a £ (a)
and P^(CT) powers go to zero in the limit of infinite radius; at each point both
oscillate between equal negative and positive parts and hence both describe
radially-directed, alternating power. There are no mode- or radius-
dependent, time-independent phase factors.
124 The Electromagnetic Origin of Quantum Theory and Light

The phases of the real part of the complex power and the real power
differ by a radius-dependent phase factor. Since the instantaneous power
represents an actual physical entity, it follows that the real part of the
complex power does not. A quantitative expression for phase angle ^ ( a )

may be obtained by equating Eqs. 3.4.1 and 3.4.2. The result is:

A eB
tan(2^)= A 2 !2 (3.4.3)

It follows from Eq. 3.4.1 that the group velocity of the real part of the
complex power is:

"gp= .. ,, (3-4.4)
8P
l+d£//do
It may be verified using Table A.26.2.20 that:

_d_ ( AtBe ^
<0
da •B,
V (3.4.5)

Combining Eqs. 3.4.3 and 3.4.5 with functional properties of the tangent
gives:

d^/da<0 (3.4.6)

Combining Eq. 3.4.4 with 3.4.6 shows that the real part of the complex
power, Eq. 3.4.1, propagates faster than the speed of light. A basic tenet of
physics is that the speed of electromagnetic energy is never greater than c.
This also suggests that the complex power is not a physical entity and it does
not describe an actual energy flow. In contrast, the first term of Eq. 3.4.2
does travel at the speed of light and does describe actual energy flow.
It follows from Eq. 3.4.1 that if the calculus operations of differentiating
or integrating complex power with respect to the radius is done, the
calculation must include operations on the function ^ ( a ) . Yet with complex
power, knowledge of ^ ( a ) is suppressed and unavailable. Therefore, it is not
possible to carry out such operations from knowledge of only complex
power.
Antenna Q 125

U°) £=l 1= 2 £-3

0 0 0 0
-Tt/2 0.618 0.777 0.785
-71 1 1.414 1.566
-3TI/2 1.618 1.882 2.221
-2n 00 2.449 2.739
-57t/2 4.104 3.289
-371 00 4.310
-7TI/2 7.852
-An oo

Table 3.4.1 Radius for Which Selected Values of Phase Angle Occur, Three Lowest
Modes

If suppressed phase angle ^ ( a ) of mode I is assigned a value of zero at a

vanishingly small radius, the value decreases with increasing radius to equal
-{l+\)ii at infinite radius. Table 3.4.1 lists values of a for which the phase
angle reaches selected values as a function of radius and modal number.
Since the use of complex power is uncompromised in electric circuits,
the complex power expression of Eq. 3.4.1, re-expressed as Eq. 3.4.7,
applies to the driving circuitry, including the input side of the radiating
surface, a = ka:

oo
l(l+l). '[AjD, - B ^ ] [ l ± cos(2a)/> - 2 ^ ) ]
P
c(<V*) = -T2 £
Tl^/ti^+l) +[A £ C £ + B £ D £ ] sin(2©fc - 2%t)
(3.4.7)

The time-dependent power expression of Eq. 3.4.2, re-expressed as

Eq. 3.4.8, applies to the external region, including the output side of the
radiating surface a = ka:

*C,/*} T,* 2 h (2/+1) ' }+[A,Q - B,D,]sin(2a^) j

(3.4.8)
126 The Electromagnetic Origin of Quantum Theory and Light

The mean square value of the time varying portions are respectively given
by:

[A,D, - B ^ Q ] 2 + [AjC, + BPzf = [A f D f + B£Cef + [Afit - B^] 2

(3.4.9)
It follows by inspection that Eq. 3.4.9 is an identity. Therefore, the total
power is continuous through the interface. The left side terms are the
magnitudes of the real plus imaginary parts of the input complex power on
the source side. On the right side, the first term applies to the time variation
of the real power and the in-phase oscillatory power. The second term
represents the out-of-phase oscillatory power.
The first two terms inside the curly brackets of Eq. 3.4.8 may be written
as:

[A,D, - B^Q] ±[AeDe + B,C, ]cos(2©fc)

= [A,D, - B ^ ] [ l ± (-1)' cos(2co/y?)] ± [ A , D , + B^C£ - (-l) / ]cos(2(B^)

(3.4.10)

Comparison of Eq. 3.4.7 and 3.4.8 as modified by Eq. 3.4.10 at a = ka shows

that the real power undergoes a phase discontinuity of 2t,i as it passes
through the antenna. The absolute phase is determined by the phase of the
source and the antenna circuit impedances.
In summary, although the total time-dependent power is continuous
through the interface between the source and field regions, the separation of
that power into constituent parts is different. On the source side, the power
separates into real and reactive parts the time varying portions of which are
in time quadrature. Power that is in phase with the input power represents
power loss from the system. On the field side, power that is in phase with the
real power does not represent power loss; some oscillatory power is in phase
with the real power and some is in phase quadrature.
At the surface of a radiating sphere, it is correct to write the complex
power in the form of Eq. 3.2.2 as:

( F , F ; +GeG*e)[A((ta)De{ta)- Bt(to)Ct(Aa)]

+/{FeF;-GeG*e)[Ae(£a)C((£a)+Be(£a)De(te)]
(3.4.11)
Antenna Q 127

The equation is correct only at radius a. The equality does not extend to
larger radii for the imaginary part.

3.5 Radiation Q
Consider a series electric circuit consisting of all three passive circuit
elements: inductance, capacitance, and resistance. Let the circuit be driven
by time-dependent voltage v(t) that produces current flow i(t). The integro-
differential equation the circuit satisfies is:

L ^ + Ri(/) + ^Ji(/)dt = v(/) (3.5.1)

The homogeneous equation has the form of the harmonic oscillator equation:

d 2 i(/) R di(/) 1 ./ x rt
— \2 1 + ^ + 1(/) = 0 k(3.5.2)
d/ L d/ LC W '

The current as a function of time is equal to:

i(/)=I 0 e s / (3.5.3)

Substituting Eq. 3.5.3 into 3.5.2 shows that:

R JR 2 1

Introduce the notation that:

a = R/2L and co0 = 1/VLC (3.5.5)

Combining shows the homogeneous current to be:

128 The Electromagnetic Origin of Quantum Theory and Light

i(/)=I0e-«/e±/^:^ (3.5.6)

The character of the solution depends upon the relative sizes of a and (Do-
Consider first the special case where:

(D 0 >a (3.5.7)

Combining Eq. 3.5.7 with Eq. 3.5.6 gives:

it/Hoe-"^'*^ (3.5.8)
The energy of the system is proportional to:

W(/) = i(/)i*(/)=I 0 Ioe- 2 a / (3.5.9)

The power out, the rate of energy decay, is:

P(/)-l[i(/)i*(/)]=-2aIoI^-2a/ (3-5.10)

A dimensionless quantity that measures the quality of an oscillating system

is:

<BWpk(/) CO
Q=
Pav(') 2a
(3.5.11)

Q, the quality factor of the oscillating system, measures the rate of the decay
of the envelope of energy W{i), which is equal to the peak value Wp^(t). P(/)
is the time-average rate of energy dissipation.
Bandwidth is inversely proportional to Q. This may be shown by noting
that the input impedance of the RLC circuit is:

Z(w) = R+/wlil-co 0 2 /tO 2 ) (3.5.12)

Antenna Q 129

The bandwidth of any system is defined to be the frequency difference

between half-power points. In this case the lowest impedance occurs for
frequency u) = coo, at which frequency the impedance is purely resistive and
equal to R. Half power points occur when the magnitude of the impedance is
equal to the square root of two times R, and this happens when the real and
reactive parts are equal. If u)) is the frequency at a half power point, it
follows that:

R = to 1 l/l-Q) 0 2 /co 1 2 ) (3.5.13)

Expanding the equation shows that:

/ W N. CfllR
(cDi-0) 0 )(a) 1 +a)o) = - J - (3.5.14)

Bandwidth is particularly useful if it is reasonably small, and if it is small,

Eq. 3.5.14 is approximately equal to:

8(i) = R/2L

The substitution has been made that 8u) = ±(a»i-coo), the frequency
difference between one of the half-power points and the resonance
frequency. With the definitions of Eqs. 3.5.5 and 3.5.11, the total bandwidth,
B, normalized to the actual frequency is:

B = 8cfl/co0 = l/Q (3.5.15)

It follows that in a low-loss, series resonant system Q is a direct measure of

and inversely proportional to the bandwidth.
A special case is the case of a lossy inductor. Although Q follows from
Eq. 3.5.11, because of the importance of the case consider another
viewpoint. The steady state input impedance for a lossy inductor driven at
frequency w is:

Z=R+/toL (3.5.16)
130 The Electromagnetic Origin of Quantum Theory and Light

If the current is Iocos(co/) the energy stored in the inductance and the power
loss in the resistance are:

W(/) = - L I 0 2 [ l + cos(2o)/)]
4
(3.5.17)
P(/) = - R I 0 2 [ l + cos(2(0/)]

Combining the definition of Eq. 3.5.11 with Eq. 3.5.17 shows that:

n ooL „
Q = — = tan£ (3.5.18)
K
Angle C, is the phase angle of the impedance. A similar expression holds for
lossy capacitors. This is a convenient measure of Q when operating far from
the resonant frequency.
Radiation Q is important with antennas since it is often necessary to
radiate a certain amount of time-average power at a given frequency. It
follows that the peak standing energy that must be present in the local fields
about the antenna is:

W pk = ^ Q (3.5.19)
p
CO

The larger the standing energy the larger will be the antenna surface
currents, the ohmic loss, and the amount of energy that returns to the source
twice each field cycle. If Q is large enough, the magnitude of standing
energy required may be more than the source can supply.
Although antenna Q is important, calculation is made difficult because
the energy radiated permanently away from the antenna is not absorbed.
With circuits, energy once absorbed is no longer a factor. With fields, all
energy remains. In the steady state the source has, ideally, been active since
time t = - oo and there is an infinite amount of field energy. Since only
energy that returns to the source affects it, the critical question in Q
calculations is how to separate energy that returns to the source from energy
that does not.
Antenna Q 131

3.6 Chu's Q Analysis, TM Fields

Chu was the first to quantify the relationship between Q and the electric size
of an antenna. In his work, he analyzed zero degree, phasor field equations
with TM sources, the same fields analyzed in Sections 3.2, 3.3, and 3.4. For
this case the phasor field components are:

o2ET = ^£(t+l)Fe(Be + i A^)P £ (cos0)e-/a

aE 0 = £ *v( D * + '' Q ) —P f (cos6)e-' & (3.6.1)

*=i de

™\% = £ F,(A, - / Be)-^P£(cos6)e-'&

de
1=1

He analyzed the frequency behavior of the input impedance of an antenna

producing these fields and, for each field mode, was able to connect the
frequency dependence with radiation Q.
To keep the work general it is necessary to separate the analysis from a
specific antenna. For this purpose he constructed the smallest virtual sphere
that just circumscribed the antenna and replaced the antenna with equivalent
surface sources producing identical external fields, see Section A.7. He
analyzed only the field energy external to the sphere. Since he ignored
interior energies, the calculated Q is the least possible value for any antenna
that can fit inside the virtual sphere. That is, with an antenna of length 2a
Chu's results are based upon exterior fields only. Interior field energy will
add an undetermined amount to Q.
The complex power on the surface of the virtual sphere follows from the
complex Poynting theorem:

271 £, *t(l+l), w x
A
Pc = —J 2 W ]h-i( i + ''*llVt + * Q) (3-6-2)

Chu next introduced voltage V^ and current le, respectively proportional to

E e and HQ, as a generalized force and flow. The complex surface power for
each mode of Eq. 3.6.2 may be written as:
132 7Vie Electromagnetic Origin of Quantum Theory and Light

(A+/BXD+/C) = ^ (3.6.3)

Since, for each mode, the angular electric-to-magnetic field ratio does not
depend upon either zenith or azimuth angle, defining modal impedance
Ze(o) to equal the ratio E e / H ^ gives the result:

V, D^(q)+/Q(a)
Zt(a) = • = T1 (3.6.4)
Hrf
.A/(°)~/B<(°)

Equations 3.6.3 and 3.6.4 are both satisfied if:

_F>_ Unl(£+\), , . /CT

D/ +
^ " T " J 3(2*+l)' ' ^6
(3.6.5)
F, 47C<£+l), N . /a
' Ti/^y 3(2^+ 1) V ' fy

The modal impedance of Eq. 3.6.4 may be used to synthesize equivalent

circuits that simulate the affect of the antenna upon its source. To do so
break the quotient into partial fractions. For the dipole case, £ = 1, the
impedance has the form:

TI r\
—T 2
+ -+*l
/c q il (3.6.6)
Zm(°) = 1
—+/ *+_L+!
a

To evaluate the impedance at the spherical surface r = a, replace k by co/c

and simplify:

1
Z 1 E (ka) = + : (3.6.7)

r| a>\ia
Antenna Q 133

The circuit with the impedance characteristics of Eq. 3.6.7 consists of a

capacitor of (ea) farads in series with a shunt configuration of an inductor of
(|xa) henries and a resistor of Tl ohms. For small values of ka, the input
impedance is large and dominated by the capacitive reactance. Power to the
far field is represented by power dissipated in the resistor.
The quotient of Eq. 3.6.4 using partial fractions is valid for each value of
I. The resulting circuit is shown in Figure 3.6.1. The circuit is a reactive
ladder network with a single terminating resistor. Each additional modal
number adds an additional L-C pair to the circuit ladder. Power in the ladder
network represents power flows made necessary by the changing geometry
of the field as the radius increases. For electrically small antennas the input
impedance is dominated by the first capacitor in series with the first
inductor. The input reactance is dominantly capacitive so long as t(2l-\) »
,2 2
k a .
Turning to field properties, the impedance of a virtual shell of arbitrary
radius, r ~ afk, may be expressed as:

Z/E(°) = R/(°)+*/(<*) (3.6.8)

Inserting the letter functions of the spherical Bessel and Neumann functions
shows that:

M<J)=TI
A^af+B^of
(3.6.9)
Xt{o) = T\
Ae{af + Be(a)2

Values of both numerators and the denominator are listed in Table 3.6.1 for
several modes; there are no resonances and the reactance is negative for all
values of a. No resonances are expected since, as illustrated by biconical
antennas, resonance occurs when the combination of antenna arms, acting as
transmission lines, and the surface impedance resonate. It is not because of
impedance changes on the spherical surface. Rather than evaluate Q
separately for each modal equivalent circuit, Chu stated that the work
involved would be "tedious" and sought approximate values that were easier
134 The Electromagnetic Origin of Quantum Theory and Light

to calculate. Since he was interested in electrically small antennas, he

approximated the equivalent circuit as a series circuit then added a lossless
inductor needed to make the system resonant. Therefore his resonance arises
quite differently from one dependent upon length of the antenna arms.

C = ea C ea
(21-3)

R = r)

Figure 3.6.1 TM Multipolar Equivalent Circuit.

This circuit has the same input impedance as a spherical shell of radius a
radiating an electric multipole mode of order £.

£ AtCt + B^D, A(2 + Bt2

1 1

2
~a3 "7
3 18 1 3 9

c rj a a
3 6 75 675 , 6 45 225
1+ + +
a 3
a a 5 7
? ^ ^
4 10 220 4725 44100 , 10 135 1575 11025
1+ + +
3 5 7 9 „a2 a „8
„a 4 +a a6
a a a a
5 15 630 8900 396900 4465125 15 305 6300 99285 893025
3 5 ~ 7
~ 9
rrU
(T a o a a a a
a a a
Table 3.6.1 Table of Functions Needed for Equivalent Impedances;
Antenna Q 135

Since at resonance the time-average values of electric and magnetic

energy are equal, he took the stored energy to be twice the time-average
stored electric energy. With a series RLC circuit, the relationships between
the input reactance and the reactive elements, Lp and Cg, are:

1 \ dX, U T 1
\
X , = col^- and (3.6.10)
coCU dco • =co^
— coL« +coC^
—
V
Solving for the values of the elements as a function of the reactance gives:

dx i _x 1 dX, XA
c, = - dco co
1 •+ -
2 dco co
(3.6.11)
CO

Chu put the time-average radiated power equal to that dissipated in the
resistor. Combining the above shows that:

1 * x\ *
(3.6.12)
2 2 ( A / + B,2)

Using Eqs. 3.6.8 and 3.6.11, the time average stored electric energy is:

frf- ^2 Hvl~ (3.6.13)

4co C^ 8 dco co

Using Eq. 3.6.7, the calculated modal value of Q is:

dX, X,
th A , 2 + B, 2 ) (3.6.14)
2J) dco co
v

For the special case £ = 1,

A,' + B l '.. 1 + W and Xj=- r\ (3.6.15)

(.*-)' (fa)\\ + (/ta)2
136 The Electromagnetic Origin of Quantum Theory and Light

\ + 2{kaf
QlE = (3.6.16)
(Jtaf l + (Jta)2

In the limit of electrically small antennas:

Lim 1 1
+ •
ka- 0 {kaf {*«) (3.6.17)

Chu stated that this result is adequate for practical antennas.

3.7 Chu's Q Analysis, Exact for TM Fields

Although Chu's technique for approximating the value of Q for each
equivalent circuit is adequate for practical purposes, a more exact analysis is
needed if critical comparisons with other analytical techniques are to be
made. For this purpose we make an exact analysis of the dipole circuit of
Fig. 3.6.1, £ = 1, and from the analysis obtain an exact value of antenna Q.
Let ii(t) and i2(t) be the currents respectively through the capacitor and
the inductor; the current through the resistor is ii(t) - i2(t). Then:

*,(/)= L^4 + Ri2(/) (3.7.1)

The instantaneous energies stored in the capacitor and inductor are:

(
/ \ ll 2 ( / ') J / \ Li
22(') (3.7.2)
Hfc

The power dissipated in the resistor is:

p(/)=R[i 1 (/)-i 2 (/)] 2 (3.7.3)

For sinusoidal steady state operation, introduce:

Antenna Q 137

i 2 (/) = I 2 cos(co/) (3.7.4)

Combining shows the charge on the capacitor to be:

qi(/) = Jil(/)d/=£ sin(co/) + — cos((0/) (3.7.5)

Resulting energies and power are:

2^1
1 coL" 1 o)L"2^ 2L
»c{<)=H4(0 2 cos(2(0/') + sin(2co/H (3.7.6)
0)C• + CR2 coC CR
CR

M^(/) = - ^ - w L [ l + cos(2co/)] (3.7.7)

C02L2
p(/) = - ^ - I 2 2 [ l - c o s ( 2 c o / ) ] (3.7.8)

Use of component values from Fig. 3.6.1 in Eqs. 3.7.6-3.7.8 gives:

wc(() -r-- + {Aa) ~{ka) cos(2co/) + 2sin(2co/)| (3.7.9)

4co

»t (3.7.10)

(/) = T l I 2 2 - ^ - [ l - c o s ( 2 c o / ) ] (3.7.11)

The total reactive energy is the sum of Eqs. 3.7.9 and 3.7.10:

Wx
M = ^ { - ( h ) + 2 H ~ [jhj ~2^) co<2^+2sin(2co')} (3-7-12)
The cyclical peak of stored energy is:
138 The Electromagnetic Origin of Quantum Theory and Light

_TlV
wnpk 2{ka) + l + 4(ka)A (3.7.13)
4co (ka) + [ka)

The time average output power is:

(kaf
Pav = W (3.7.14)

Combining:

wfKP k
Q= l + ^\+4(ka)4 (3.7.15)
2{kaf (ka)

This is the exact expression for the Q of the circuit of Fig. 3.6.1 for the
special case i = 1. In the limit as ka goes to zero Eqs. 3.7.15 is equal to
Eq. 3.6.17.

3.8 Chu's Q Analysis, TE Field

The fields about az-directed magnetic multipole follow from Eq. 3.2.1:

a2r\RT=-YKi+ l
PABli°)+ ' A^(CT)]p<?(cos6)

crriHe = - X G£[D£(G)+ i Q ( a ) ] — P^(cose) (3.8.1)

d 0
*=i
°° H
<*> = £ G£[Ae(c)- i Be(c)] — Pe(cosQ)
«_i at)

Following the procedure used for TM modes, for the TE modes introduce a
generalized force and flow as a voltage and a current, this time proportional
respectively to E^ and -H e . Each mode then satisfies the power equation:
Antenna Q 139

Pc = - ^ G , G * ^ | [ M a ) - / B , ] [ D , ( a ) - / Q ] = l v / , (3.8.2)

Like TM modes, for each TE mode the angular electric-to-magnetic field

ratio depends upon radius, not angle. Defining modal admittance Y^ to equal
the ratio HQ/EA^, it follows that:

Hfl 1L De(a)+iCe{a)
Y/M(CT) = (3.8.3)
V/ Ae(a)-iBt(a)

Both Eqs. 3.8.2 and 3.8.3 are satisfied if:

G^ Un£(e+l)r , . ia
l n n n
k \ 3(2*+l) '
(3.8.4)

T -G« —A-—i[ D / (o)+/Q(o)]e

ip - Ttf'Y
—r 3(2*+l)

Comparison of Eq. 3.6.4 and 3.8.4 shows that:

(3.8.5)

Repeating the procedure used to evaluate the TM equivalent circuits gives

the equivalent circuits for TE modes. The resulting circuit is shown in
Figure 3.8.1; it is the dual of Figure 3.6.1.
Since the circuits are exact duals, each power and energy of Section 3.6
has an exact counterpart in Section 3.8, though what is capacitive becomes
inductive, and vice versa. For example, the input impedance of electrically
small electric dipoles is equal to the large input admittance of electrically
small magnetic dipoles. If the circuit used to calculate Q is a parallel
capacitor, inductor, and resistor the magnitudes of energies are unchanged,
though the forms are reversed. Q therefore is the same:
140 The Electromagnetic Origin of Quantum Theory and Light

ea C= ea
(2*-l) {21-5)

R = T]

Figure 3.8.1 TE Multipolar Equivalent Circuit

This circuit has the same input impedance as a spherical shell of radius a
radiating magnetic multipole mode of order /.

Q ^ 2(/fo)
-JLJU^WK J
(3.8.6)

In the limit of electrically small antennas:

Lim 1 1
XM
ka^ -^f{ka) (3.8.7)

3.9 Chu's Q Analysis, Collocated TM and TE Modes

In addition to analyzing individual moments, Chu also analyzed
superimposed (TE + TM) modes of the same order, phased to produce
circular polarization. A basic difficulty is that to add modal powers it is
necessary to account for phase differences, yet phase information is not
contained in the complex power expressions. There are, however, other ways
to account for the phase difference; Chu did this by requiring circular
polarization. With both modes present and the field circularly polarized, the
standing energy oscillates between the radiation fields of the two dipoles.
Antenna Q 141

The average standing electric energy in the TM mode is the capacitive

energy obtained using Eq. 3.6.10. Since the TE mode is its exact dual, the
magnetic energy stored in the TE mode is equal to the electric energy in the
TM mode. Therefore the total standing electric energy is:

1
fre(o) = 11 t l (3.9.1)
4co 2 Q 4 " 4 dco

The time average radiated power is twice that of Eq. 3.6.12:

\2
(ka)
• = R/i/i/* = n i / i / * - L - ^ A
\ + {ka) (392)

The reactance is given by Eq. 3.6.10. For radiating dipoles the derivative is:

dX^ _ {ka) dXj, _ x\ l + 3(/fo)21

du) co d{ka) CO
{ka)\ + {Jtaf

Combining the above gives the dipole Q:

\2
l + 3[JtaY 1
Q> -+ • (3.9.3)
2{kaf \ + {kaf 2{ka) ka

For electrically small antennas, this Q is approximately half that of either

dipole acting alone. The interpretation is that since the standing energy
simply moves back and forth between reactive elements the total value is
nearly the same as for either dipole acting alone and the radiated energy is
twice that of a single dipole. Therefore, Q is reduced by an approximate
factor of two.
142 The Electromagnetic Origin of Quantum Theory and Light

3.10 Q the Easy Way, Electrically Small Antennas

It is possible to solve for Q from the impedance most easily by use of Chu's
equivalent circuits. For antennas electrically small enough so l{2l-\) »
2 2
k a , the input reactance is dominated by the first reactive element, a
capacitor for TM modes and an inductor for TE modes, and Q is very nearly
equal to:

Q^ = tan[^(a)] (3.10.1)

Combining Eq. 3.10.1, Chu's equivalent circuits, and the impedance results
of Section 3.2 shows that the modal Qs of electrically small antennas are,
very nearly:

Q ^ M (3-10.2)

Values of y( are listed in Table 3.2.1. Keeping only the lead term gives:

4(2^-l)!!f
v _
/ , \2M-1
\Ka) (3.10.3)

3.11 Q on the Basis of Time-Dependent Field Theory

The analytical works that follow are based upon the analysis of an idealized
radiating sphere. Experimental or numerical confirmation, however, requires
actual or numerical embodiments and, in the main, embodiments are made of
straight wires and wire loops, not spheres. The analyses to come assume no
source coupling between modes and therefore the desired modes, and only
those modes, exist.
With this analysis, as with Chu's, only fields at a radius greater than the
radius of the virtual, source-containing sphere of radius a are considered.
Ignoring fields at smaller radii has the great advantage that results are not
specific to a particular antenna. However, although the interior volume for
an electrically small antenna is small, for a fixed moment the field
magnitude increases rapidly enough with decreasing radius so the interior
Antenna Q 143

energy remains a significant portion of the total standing energy. As

examples, for the spherical shell dipole analyzed in Section A. 14 the interior
energy is half that of the exterior energy. For biconical transmitting
antennas, see Eqs. 2.7.1, the TEM mode and an infinite number of TM
modes are included. For biconical receiving antennas, see Eqs. 2.17.2, the
TEM mode and infinite numbers both of TM and TE modes are included.
Nonetheless, to keep the results general it is necessary to ignore interior
fields.
To calculate the Q of a multiport antenna, an antenna driven by more
than one terminal pair, it is necessary to account for suppressed phase
angles. Since phase angles are an integral part of actual fields, we begin with
a general multipolar expansion for phasor fields and then transform phasor
fields into actual fields. The phasor form of the multipolar field expansion is:

Er = X X ^F(^wM^ + l)^^p;(cose)e"*"y>w*
^=0 m=0
(3.11.1)
- t h,(o). -AT-y/wJ)
ti^r = y'X X '' G(e,m)e(e + i)-^-^p^'(cose)e
^=0 OT=0

-iG-jm§
Ee = X X '" ff(e,m)h'e(o) — ?™(cosQ)-G((,m)he(o) P"(cos6)
1=0 07=0 d9 sin 6

-e F(e,m)he(a) -fo-jmty
— ?^(cosQ)-/G(e,m)b.](a) P"(cos9)
d9 sin 9
~ I
E r -i<s-Jm§
* = -yX X iF(e,m)h'e(a) ?"(cosB)-G(e,m)h((a) — pf(cos6)
t=0m=0 sin 9 d9

-/o—jmty
IHQ = JX X i~ F(e,m)ht(a) P™(cosQ)-iG(lm)h'e(o) — P^(cos9)
l=0m=0 sin 9 d9

We examine the Q of different modes and modal combinations by

considering a series of examples.
The first example is the set of TM modes of degree zero. For this case,
Chu's "omnidirectional" case, all coefficients except F(^,0) are equal to zero
and, for simplicity in notation, the arbitrary normalizing equality is made
that F(^,0)/~ =1. After replacing Hankel functions by equivalent letter
144 The Electromagnetic Origin of Quantum Theory and Light

functions, see Appendix A.26, and accounting for the suppressed time
dependence, in terms of retarded time, tR, the actual field terms are:

a 2 E r = £ ^ + l)[B^(o)cos(a)/ J? )-A / (a)sin(co/ J? )]P^(cose)

e=i
CTE
9 =£ [ D ^( CT ) cos (w/ / -)-C^(a)sin(w/ J? )] — P^(cos9)
e=i de
OT H _d_
1 <1)= E [A^(cT)cos((0/^) + B^(CT)sin(co^)] —P^(cose)
^=1
'de
(3.11.2)

The total energy density, wj, at each point in the field is:

w1v = - E « E + - H « H (3.11.3)
2 2

Substituting the field forms of Eq. 3.11.2 into the energy density expression
shows that for each mode:

i
V*1' [(Ae2 + B / ) - ( A / - B , 2 ) c o s ( 2 o > f c ) - 2 A / B , s i n ^ a v ^ P ^ c o s e ) ]
Wf ••
J _ [ ( A , 2 + B(2 + Ce2 + D , 2 ) + ( A , 2 - Bt2 - Ce2 + D,2)cos(2a>fc) dP^(cos9)
o [ + 2 ( A ^ -C^)sin(2a)/>) de
(3.11.4)

For brevity, the dependence of the letter functions upon a is suppressed. The
right side, top row of Eq. 3.11.4 is the energy of the radial component of the
electric field intensity. The remaining terms are the combined energies of the
angular field components. The first and second lines have different parity
with respect to the zenith angle.
The modal components of the Poynting vector are:
Antenna Q 145

1 [(APf - BeCe) + (A f D f + B^)cos(2cofr)|fdP^(cos9)

2TIO- 2 | -(A^-Bp^sin^co^) d0 J

Ne = - ^ ^ { 2 A ^ c o s ( 2 ( 0 / J ? ) - ( A / - B / ) s i n ( 2 ( 0 / J ? ) } p ^ ( c o s e ) ^ ^ ^
2r|a

(3.11.5)

The continuity equation describes energy conservation in the field, and is:

V.N + ^ 1 =0 (3.11.6)
dt

The equality of Eq. 3.11.6 is readily verified by substituting Eqs. 3.11.4

and 3.11.5 and solving.
We seek to separate the total energy density into a part that travels with
the wave on its outbound journey and a part that separates from the wave,
remaining within the local region of the antenna. Separated power may be
calculated by riding with the wave and determining, at each point, the rate at
which energy departs from the wave. For this purpose note that the
divergence of the power at constant retarded time is equal to the negative
rate at which energy per unit volume separates from the wave:

V*.N = - (3.11.7)
'R

Symbol H>S indicates the energy density at each point that separates from the
outbound wave and oscillates over a distance of X/2; we define it to be
standing energy density. Symbol V^ operates at constant retarded time.
The divergence operation of Eq. 3.11.7 is aided by values obtained by
taking the derivatives of Table A.26.2.13 and A.26.2.14. The results are:

(£+l)[P^(cos0)]2
Vz>«N„ = + 0 2A^B^ cos^w/tf) - [ A / - B / ] sin(2oo/>) l2
2TIO-<
d
—p.n (cose);
de
146 The Electromagnetic Origin of Quantum Theory and Light

2
V*«N r —picose)
n ;
2r\(f de

-^-^A/B/-2(A<-D/)(B/+C/) (2w/>)
COS!

CT
(3.11.8)

-(A/-B/)-(A, B/J -~Ce — D/j

sini(2o)/>)
+2(AA + B A )

Summing the components of Eq. 3.11.8 then taking the indefinite integral
with respect to retarded time gives:

e2(£ ++ Y)2
4 * 1(A2-B/)cos(2(0/^)+2A^sin^co/^)|[P^(cos0)]2
eK e
^ 4 4 (A^-D^) 2 -(B^ + C^)2]cos(2co^) dP^(cos9)
+2(A^-D^)(B^ + C^)sin(2co/^) de
(3.11.9)

K is a constant of integration, with dimensions chosen for later convenience.

There are two requirements on K: Since it is an energy density it can never
be negative and it must appear in the wj expression. This is the equivalent of
requiring that both zenith angle parities in Eq. 3.11.9 be everywhere greater
than or equal to zero. Evaluating K and entering it into Eq. 3.11.9 gives:

2 2

' ( ^4 1 } {(A/ + B / ) - ( A / - B/)cos(2co^)-2A^ sin^co/^P^cose)] 2

w$ ; {Ae-D()2 + {Be + C(f

dP^(cos9)
2 2
(A,-D,) -(B, + C,) ]cos(2cofc) de
+2(A^ - Df )(B/ + C^)sin(2co/y?)
(3.11.10)

This is the source-associated energy density. It is separate from the traveling

wave and oscillates about a fixed position in the field. The top row of
Antenna Q 147

Eq. 3.11.4 is the energy density of the radial field component and the other
terms are the energy densities of the angular field components. Comparison
of Eq. 3.11.4 with 3.11.10 shows that the top lines are identical: all energy of
the radial field component remains attached to the source. Some energy of
the angular field components remains attached to the source and the rest
does not.
Subtracting ws from wT gives the energy that remains part of the
traveling wave: the field-associated energy density w5.

"*=i dP^cose)
d0
(3.11.11)

The expressions for N r and vv§ differ by a multiplicative factor equal to the
speed of light, c. A characteristic of traveling energy is that power is equal to
the product of the energy density and the speed of travel. The movement of
w§ produces the radially directed power density. Its value at the generating
surface r = a determines the antenna input impedance. Both wT and ws are
positive real, physical entities, but vt>§ connotes power and hence can be
negative.
For single modes, an alternative and simpler derivation of the standing
energy is to divide the radial component of the Poynting vector by c and
subtract the result, Eq. 3.11.11, from the total energy density expression,
Eq. 3.11.4. The result repeats the source associated energy density,
Eq. 3.11.10. Although the technique is arguably correct for single modes, the
process does not generalize to multi-modal situations.
Consider what happens if the source is suddenly disconnected. Since
nothing travels outward faster than the speed of light, the originally
outbound portion of the field continues without change, and energy W§ is
transported on out into free space. Energy W$ is fixed in position. As the
fields at radius less than r collapse, the energy density exterior to that radius
becomes larger than those nearer, producing an inward pressure on the field.
We presume, therefore, that energy Ws returns to the source.
During steady state operation, it is helpful to determine energies at the
time a given wave is emitted. In a form of the ergodic theorem this is equal
148 The Electromagnetic Origin of Quantum Theory and Light

to energy calculated by summing over the curve in its outward journey. In

such terms the total standing energy is equal to the volume integral of ws'

271 71
W% = — J a 2 d a j d<t>j"sinedew$(a,/>) (3.11.12)
ka 0 0

Substituting Eq. 3.11.10 into 3.11.12 and integrating over the full solid angle
leaves:

4A / (A < -D < ) + — ( A / C / + B / D / +2A / B / )

do
Tte £(£ + 1)
*k = +1 -—(A A -BpAcasfaatg) (3.11.13)
kH2t + \) )i do
~ ( A A + B A -(-l)^)sin(2o)^)

The radial integrals can be done with the assistance of

Table A.26.2.11 through A.26.2.14. Doing the integrals shows the standing
energy to be:

ka
- ( A / C / + B / D / + 2A < B / ) + 4 j d a A < ( A / - D / )
oo
re 1(1 + 1) .(3.11.14)
rK = 3
A {2t + l) +(A^C^ - B^D^)cos(2co/J?)
+(A,D, + B^C, - {-if jsin^co/*)

In Eq. 3.11.14, the letter functions are evaluated at r = a. The peak energy
value is:

/fc?
- ( A / C / + B<D/ + 2 A / B / ) + 4 j o < j d c A < ( A / - D / )
re/(* + !)
k
^ ^ 3 ( 2 / + l ) ' +^|(A? + B$fc] + D ] ) - 2 ( - l ) ' ( A A + B ^ ) +

(3.11.15)
Antenna Q 149

The total output power on the antenna surface is obtained by taking the
surface integral of N r :

2 2n n
? = ^jj d<t>j"sin8deNr(o\/tf)
0 0 (3.11.16)

= If {l + (A<D<+B<C<)cos(2m/ig)-(A<C<-B<D<)sin(2a)/ie)}

By Eq. 3.5.11, the ratio of the peak of Eq. 3.11.15 to the time average of
Eq. 3.11.16 determines Q:

w
^Speak
Q> (3.11.17)
average

Combining Eqs. 3.11.15 with the values of Eq. 3.11.16 and Eq. 3.11.17 to
give an expression for the Q of arbitrary mode £ of radiation:

ka
-(A < C < + B / D / + 2A / B / ) + 4 j d o A / ( A / - D / )
1 (3.11.18)
2
+^(A/ + B^2)(c/ + D / ) - 2 ( - l / ( A ^ + B ^ ) + l

3.12 Q of a Radiating Electric Dipole

The actual fields of a time-varying, z-directed electric dipole follow from
Eq. 3.11.1. With coefficient F(1,0) equal to one and with all others equal to
zero the fields are:

a z E r = -2 cos(co/^) + — sin(a>4p) COS0

CTE0 = cos(co/^) +
(-4) in(a>fc)
sinl sin 6 (3.12.1)

rniH^ - cos(co/J?) + sin(co/^) sin 6

150 The Electromagnetic Origin of Quantum Theory and Light

The energy densities follow from Eqs. 3.11.4, 3.11.10, and 3.11.11:

— [ l + cos(2(D/>)] + —sin^a)/,?) + — [ l -cos^co/tf)] [cos 2 6

a a a6 J
Wr=—-
1
4
„2 „6 L' —V /f/J • 4 ~">V~*"/r/ 3 5

(3.12.2)

— [l + cos(2co/^)] + —-sin(2a)/>) + —-[l-cos^oo/^)] [cos 2 9

^
+ —-[l -cos(2co/^)]sin 2 0

(3.12.3)

F f 1 2 r 2
„3 „5
sin(2co/^)[sin20
\o aJ
(3.12.4)

The energy density described by the first lines of Eqs. 3.12.2 and 3.12.3 is
centered on the antenna axis and is the energy of the radially directed
component of the electric field intensity. The second lines are centered at
0 = 7t/2; in Eq. 3.12.2, it is the energy of the angularly directed field
components and, in Eq. 3.12.3, it is the energy of the nearest radial field
term. Plots of ws at four different times are shown in Fig. 3.12.1 for
ka = 0.1; note the changes of scale.
The components of the Poynting vector follow from Eq. 3.11.5:

N 1 \ 1 2 (2 1
^ IfJ—[ 1 _ C O S ( 2 W ^)] + C~T H 2 C O / * ) - Vc3
r = •^•1 5
OJ
sin(2co/^)[sin20

(3.12.5)

sin^co/tfHsinOcos© (3.12.6)
2T1 a VCT CT J
Antenna Q 151

(a) (b)

0.1

0.05.

0.

-0.05.

(c) (d)

10 4

5 2,

N ° N 0-

-5 -2

Figure 3.12.1 Standing energy density of a z-directed electric dipole.

Plots are shown at the four times 2ti)tR = (a) 0, (b) ii/2, (c) n, and (d) 3it/2 at
the range ka = 0.1. Note change of scale on axes. The standing energy
density is centered on the z-axis and the far field radiation pattern is
centered on the equator. For more detailed plots see:
<http://www. ee.psu. edu/grimes/antennas/breakthrough. htm>

Substituting values of the letter functions into Eq. 3.11.14 gives:

2ne 2 1
m.= [l + cos^w/^)] + 2" sin^co/^) + j [ 1 - cos(2co/^)]
(ka) (kaf (ka)3
(3.12.7)

Similarly evaluating Eq. 3.11.16 gives:

152 The Electromagnetic Origin of Quantum Theory and Light

471 r , X1 2
P= 7 \ \} ~ cos(2(0/^)J + - — —1 cosl( 2 ( 0 ^ ) - sin(2oo/^)>
3nr {/ca) (**) (/fa)3
(3.12.8)

To relate Eq. 3.12.8 to the input impedance, rewrite it in the form of time-
dependent complex power:

471
Pc=" • [l-cos(2o)/>-2£(o-))] + — l — sin(2co^-2^(a)) (3.12.9)
3TI^ ka
I \ ) J
Since the radius of the generating surface is fixed and there is but a single
mode, the value of £,{a) is unimportant. The complex power follows from
Eq. 3.12.9 and is:

471
Pc(°) = 1+- (3.12.10)
3T]/P (Aaf

The antenna input impedance follows from Eq. 3.12.10.

Substituting values of the letter functions into Eq. 3.11.18 gives:

l 1
Q> 1 +^\+A{kaf + (Jta)
2{Jtay (3.12.11)

This is the same value obtained using the exact analysis of Chu's equivalent
circuit, Eq. 3.7.15. In the electrically small limit Q goes to:

(3.12.12)

Analytical results are summarized in Table 3.12.1.

To examine the effect of coordinate rotation, rotate the dipole from the z-
to the x-direction. This illustrates the role of antenna rotations that appear in
the more complicated modal structures. The force fields of an x-directed
electric dipole follow from Eq. 3.11.1 with F(l,l) = 1, all other coefficients
equal zero, and keeping only the real part with respect to j , are given by
Antenna Q 153

Eq. 3.12.13. Analytical results are summarized in Table 3.12.2. Power

maximum occurs at 6 = 7i/2.

a E r = 2[BCOS(CO//? ) - A sin(G)/£)] sin Gcos §

OEQ - [Dcos(a)/f) - Csin(a>/£ )]cos Gcos if
(3.12.13)
cmH,!, =[ACOS(CO/ /? )+ Bsin(co/£)]cos6cos<|)
CSEQ = -[DCO^CD/^) - Csir^tfl/,?)] sin <|)
anH,], = [Acos(co/yp) + Bsin((0/f)] sin <|)

\ —T- [l -coslXatfij] + —-sm(2at#) + — [ l + cos(2co/>)] [cos2 6

v^(/ R ) = - [a a5 a 1
+ — + — [l-cos(2a)^)] + ^ - c o s ( 2 c o ^ ) - — sin(2ffl/>) sin 2 6
J D
I K<r a") cT Vo a ; J J

- [l - cos( 2co/>)] + —- sin( TiSitR) + —j [l + cos( 2(0/^)] [ cos2 9

' a5 a J
"$('*) =
+ \ —- [l - cos( 20)/^)] [ sin 2 6

J1 sin 2(o
^(^)=| ^ [ ! - H ^ ^ ) ] - 4 - - A ( ^)+4"cos(2to^)sir,2e
•'la \a a J a J
Nr(/>) = ^ j ^ - [ l -COS(2UM>)] -f - \ - - ^ ] sin(2aM>) + 4-a*(2aM>)|sin2 6
a cr

Ne(^) = • —cos( 20)/^.) + — - — I sin( 2(0/^) sin 9cos 6

2n
N(/*) = O

m,=3t3 J\(kaf3 [l-cos^CD/^)^ ^sin(2©/^) + -.—r[l + cos(2(0/>)]

(hY
An 2
[l-cos^co/k)]- sin(2(0/J?) + 2"COs(2C0/^)
3r\/t3 (*") {Jtaf ( ^
1 ,4,+. 1 3
Q> 1+* Gain = —
2

Table 3.12.1 Radiating, z-Directed Electric Dipole

The force fields of this table result z/F(l,l) =\,all other coefficients are
equal to zero, and only the real part with respect toj is retained.
154 The Electromagnetic Origin of Quantum Theory and Light

1 9 1 2 2
[l - cos(2co/tf)] + — sin(2(0/^)+—[l + cos(2co/>)] sin Gcos (>
|
La
wr(4?) = - " T + _ 6 P-cos(2a)^)]+4 r cos(2(o/ /? )- — — 5 \sm{2atR)
a CT y a 1.(5 a
Gcos2J<>| + sin2 <|)J
(cos2 Gcos
x Icos

i 2 1 2 2
-r-[l - cos^co/^)] + —^sm(2<s>t/l)+—[l + cos(2co/#)]sin Gcos <>
|
*%('*)= 4 r a o"
+ -j[l-cos(2co/^)](cos 2 Gcos2 <>
j + sin2 0)

w^(tjl) = ^[l-cos(2co^)]- — r sin(2a>/y?) + —cos^co/^) Icos2Gcos2()) + sin20J

2 a VCVCT/ (j J* '
2 2 2
N r ( ^ ) = - —j[l-cos(2co/>)]- —r r sin(2co/J?)+ —j-cos^co/^) Icos Gcos (> + sin Gl

N e (/^) = —I—2- cos (2co^)+ —j j \sin(2(0/ji)[sinGcosGcos <>

|

N
<|)('>?) = ^ - | 4COS(2'°<ff)+ — 5 Uin(2(0^)lsinGsin(jicos(j>
2nl o 4 " " r " ' " ' l ( i 3 a 5 , j

4TI 1
P=
7 T 2 [i-cos( 2w ^)]- n(2o)^)H (2m*)
(*a) (faz)3J (*a) z

>—-!-JTfl + Vl + 4(*a)4) + --r

2(k^) (to)

Table 3.12.2 Summary of Results, x-Directed Electric Dipole.

3.13 Surface Pressure on Dipolar Source

Radiated fields carry with them the kinematic properties of energy,
momentum, and angular momentum. These kinematic properties produce
both a pressure on the radiating surface, and, in certain special cases, a shear.
Consider the pressure on a spherical surface generating the electric dipole
field of Eq. 3.12.1. The calculation technique is based upon the three spatial
dimensions of Eq. 1.8.2. For a resting sphere, the equation is:

V-aTg/axj (3.13.1)
Antenna Q 155

F v , the force per unit volume, is given by Eq. 1.6.14 and the Maxwell stress
tensor, Ty, is given by Eq. 1.8.6. Changing from rectangular to spherical
coordinates may be done directly or by extension. The result is equal to:

(ftE r - E e -E*
(eErEe + uHrHe) (eErE0 + uHrH<,)
^[ H r '-He'-V]
|[ Ee 2 -V-E r 2 ] '
Fur (eEeEr + uHeHr) (eEeE^ + uHeH^)
2 2
+ ^H e -V-H r ]

^ j ^ ~Er _ E 8 j
(eEfEr + |ifyH r ) (eE^Ee + uH^He)
+
T[ H «I> -H
r ~He ]
(3.13.2)

The off-diagonal terms describe surface shear. Matrix element Tn- describes
the surface radiation reaction tension; the net pressure on a radial surface is
equal to the difference between exterior and interior surface values of T rr .
Let the generating source be a conducting sphere of radius a that supports
the surface charge and current densities that generate the exterior fields. The
exterior fields are those of Eq. 3.12.1, and there are no interior fields. For
that case the surface radiation reaction pressure is:

Trr(^) = f [ E r 2 - E e 2 - E ( t ) 2 ] + ^ [ H r 2 - H e 2 - H 0 2 ] = p ( / J ? ) (3.13.3)

The expression for Trr(/#) at the surface is most easily evaluated from the
expression for energy wj(tR) of Table 3.12.1 by reversing the sign of the
angular field terms and replacing a by ka:
156 The Electromagnetic Origin of Quantum Theory and Light

g-[l-cos(2(0/)] + jsin(2(o/)+ 4-[l + cos(2a)/)Hcos28

[(tor) {ka) (&a) J

[l-cos(2co/)] + jcos(2(o/)
p('K {kaf (kaf (ka)
|-sin2e
in(2co/)
sin
{(ka) (ka) )
(3.13.4)

60 90 120 150 180

0 (Degrees)

Figure 3.13.1 Surface pressure on a virtual sphere producing a z-directed electric

dipole radiation field
ka = 7i/2, 0 = 2dX; oilfields at lea < n/2 are equal to zero.

A special case of interest is ka - nil. The pressure on the surface of the

virtual sphere for several phases is shown in Fig. 3.13.1. The figure shows
pressure, p(r), in Pascals as a function of zenith angle at several time phases.
The pressure on the z-axis, 0 = 0 and n, is expansive and fluctuates between
about 0.13 and 0.45 Pa; the pressure in the xy plane, 0 = n/2, is compressive
Antenna Q 157

and fluctuates between about 0.02 and 0.27 Pa. At 9 = rc/4 the pressure
alternates between compressive and expansive. Although the figure
represents actual pressure on a spherical, conducting surface, for virtual
surfaces the pressure is given by Eq. 3.13.1. Commonly, electric dipole
antennas are driven from a point source at the origin through a transmission
line of approximate length X/2; the figure applies to such an antenna with
hemispherical, radiating caps; an example is a wide angle biconical antenna.
Since, as Eq. 3.13.4 shows, for antennas with ka much less than one the
surface pressure varies as the sixth power of the product ka, so does the net
force on the radiator. A plot of radiation reaction pressure versus zenith
angle for ka = 0.1 is shown in Fig. 3.13.2. The extreme expansive pressure
on the z-axis is 2 MPa, and the extreme compressive pressure in the xy-plane
is 0.5 MPa. Whether an electrically small, radiating sphere distorts to
become a needle is determined by the relative sizes of these pressures and
the physical strength of the source.

6 (Degrees)

Figure 3.13.2 Surface pressure on a virtual sphere producing a z-directed electric

dipole radiation field
ka = 0.1, 0 = 2oX; all fields at ka < 0 are equal to zero.
158 The Electromagnetic Origin of Quantum Theory and Light

We conclude that a substantial radiation reaction pressure exists on the

surface of an electrically small sphere generating electric dipole radiation
from surface sources. The pressure acts to extend the length in the direction
of the electric moment and to compress the center region.
Detailed three-dimensional, time-dependent plots of power, energy, and
electromagnetic stress in the vicinity of a radiating electric dipole, and of
mixed electric and magnetic dipoles are maintained on website:
<http://www.ee.psu.edu/grimes/antennas/breakthrough.hhn>.

3.14 Q of Radiating Magnetic Dipoles

Consider zero degree TE modes. Although the electric dipole fields are the
dual of magnetic dipole fields, the sources are physically quite different.
Sources of TM and TE fields, see Sections A.28 and A.29, are respectively
linear currents and current loops. To obtain the fields, put all coefficients of
Eq. 3.11.1 except G(^,0) equal to zero and, for simplicity in notation, make
the arbitrary choice that G ( ^ , 0 ) / = - 1 . After putting j = /, replacing
Hankel functions by letter functions, and accounting for the suppressed time
dependence the actual values of the remaining field terms are:
oo
CT2T H
l r = X Ki +
l)[A^(o)cos(co/>)+B^(o)sin(co/ J? )]P^(cose)

00
d (3.14.1)
CTE
<t> = _ £ [Bt(°)COS(G)/>)-At(a)sin(co/>)] — Pe(cos 0)
d6
e=i

00
H
^iHe = ~ X [C/Hco^fO/f) + D/(o)sin((o/>)] — P^(cos6)
de
e=i
Results of a radiating, z-directed magnetic dipole are tabulated in
Tables 3.14.1; magnitudes are identical with the electric dipole case and only
the phases of the time-dependent terms differ. The force fields of this table
result from Eq. 3.14.1 for case t = \.
Antenna Q 159

—-[l + cos(2a>/>)] r sin(2co/tf) + — [ l - c o s ^ / , ? ) ] [cos2 6

*r('.ff)=4
+ - T + - T [l + cos(2a)/>)]-^rCOs(2co^)+ 4~ "3 " 4 "„)5s i n ( 2 f l » ' i ? ) si
"2e
CT (J CT CT

—[l + cos^/tf)] -sin(2a)//p) + —-[l - c o s ^ / ^ ) ] [cos2 9

M a CT
s(/>)=4
2
— [ l + cos^/tf)] [ sin 6

M
S ( ^ ) = | -T[1+COS(2CO^)]+ 4"~~T s i n ( 2 a ) ^)--r C 0 S ( 2 ( D ^)h i n 2 e
*• CT V CT CT y O

Nr(fc) = —-[l + cos(2a>/i?)] + — sin(2(0/>) jcos(2co/>)[sin 2 6

a2 VCT CT 1 CT

N9('>?) = — 7 cos(2(0/>)+f — - — j s i n ( 2 ( 0 / > ) sinGcose

~ 1 4 " ! 1 -cos(2a)/>)] + 4- c o s ( 2 t 0 / /?) - [ " T " s] M2®'*) [sin e

T l I CT

4TI
[l + cos(2o)/>)] + in(2©/>)-
P=
3n£ 3 {*») {*af sml w -cosl•(2fiWfc)

1
Q2
^l l+ ^wj + (i) Gain = —
2

Table 3.14.1 Radiating, z-Directed Magnetic Dipole

3.15 Q of Collocated Electric and Magnetic Dipole Pair

Since the TE and TM field components of z-directed antennas do not
overlap, when evaluating the energy densities it is necessary to determine the
integration constants for each solution before summing over the vector
fields. The vector field equations for F(1,0) = - G(1,0) = 1, all other
coefficients are equal to zero, and/ = i are shown in Eqs. 3.15.1.
160 The Electromagnetic Origin of Quantum Theory and Light

a 2 E r = -2 cos(a)/^) H— sin(a)/^) COS0

a 2 riH r = 2 — cos(a)/^)+ sin((o/^) cos£

'_ 1
aE e = -cos((£>ffi) + sn^co/tf) sin0 (3.15.1)
c

o-nH^ —cos((0/Jp) + sin(co/yp) sin 6

aE-s = - co^co/^) + — sh^co/^) sin 9

aTiHe = 005(004?) + — sin(co/^) sin 6

The far field is circularly polarized. A table of dynamic values similar to

those of Tables 3.12.1 and 3.14.1 is given in Table 3.15.1. The three energy
densities and the radial component of the Poynting vector are all time-
independent. Since there are two sources, each of which produces the same
average output power as listed in Tables 12.1 and 13.1, the time-average
value for this case is double that of the previous cases. The zenith and
azimuth components of the Poynting vector respectively are and are not
equal to zero. The gain and pattern are the same as for individual dipoles.
Since the total standing energy of both dipoles is nearly equal to the peak
value of either, Q is about half that of an isolated dipole.
If the phase of the magnetic dipole of Table 3.15.1 is shifted by nil, so
G(1,0) = -i, the TM modal terms are the same as listed in Eq. 3.12.1 and the
TE modal terms are given by Eq. 3.15.2:
Antenna Q 161

' 4 2n (2 4 I 1
«t = 2" 2 sin2 9
T + - T |cos e +
VCT a
4_ _4_ cos 8+—7-sin 6k w$ = —jsin 0
=
^ 2 ' VCT ay a I a

N r (fc) = — 7 s i n 2 e ; N e (fr) = 0; N*(fc) = rsin6cos9

87te 1 1 8TC
ffk = J - + - ; p
3^ 2{kaf {ka) 3r\Jt2
1 3
Q^ Gain =
2{Jcaf +
{&*) 2

Table 3.15.1 Collocated, Radiating z-Directed Electric and Magnetic Dipoles

Producing Circularly Polarized Fields

a 2 n H r = 2 cos( (MJI ) + — sin( (o/^ ) COS0

O"

CTE. =- - cos( co/^ ) - sin( (0/^) sin6 (3.15.2)

/ 1
071% = •cos in
sini(»^?) sin0
• ( « > ( * ) -

v'V,
The powers and energies produced by the electric and magnetic moments are
in phase. Time-dependent powers sum to twice the values of Table 3.12.1.
Both gain and Q are equal to those of Table 3.12.1.
Although the integration constant is introduced quite differently in
collocated parallel and crossed moments, the calculated Q is the same.
Next consider collocated x-directed electric dipole and ^-directed
magnetic dipole sources. For this configuration the fields strongly overlap
and the integration constant is determined after summing over the vector
fields. Fields with F(l,l) = 1> G(1>1) = ~h all other coefficients are equal to
zero, and/ = i are listed in Eq. 3.15.3:
162 The Electromagnetic Origin of Quantum Theory and Light

1
CT Er =-2 cos(vytff) + — sin(co/^)
sin 6 cos <)|
O"

CT2nHr = - 2 cos( (0/^ ) H— sin( (0/^) sin G sin 0

—cos(co/#) - sin(o)/#)
a
CTEfi = cos<]) (3.15.3)
1
—cos(u)/£) ! - • in((0/jf) COS0
sin
V 0 J

cos(co/^) - sin(co/^) cos 9

CTEA = - sin 4>
1
+ —cos(co//p)- 1 - sin (©/>)
V. a" J
Vfiq, =-E(j)cot<l); T]He = E 0 tan <>|

Far fields are linearly polarized. Dynamic values are listed in Table 3.15.2.
As in the case of Eq. 3.15.2 the energy and power terms are in phase, the
energy and power terms are doubled, and Q is equal to that of either dipole
radiating in isolation. The radial and zenith portions of the Poynting vector
contain factors that are proportional, respectively, to cos0 and sin9. The
terms are suppressed in the Table since they do not affect the total energies.
If the relative phasing of the dipoles is F(l,l) = 1, G(l,l) = - 1 , all other
coefficients are equal to zero, and/ = / the fields of Eq. 3.15.4 result

CJ E r = 2[Bcos(a)/tf) - Asin(u)/(?)] sin 0cos (|>

(3.15.4)
a 2 T|H r = 2[Acos(co/J?) + Bsin(co/^)] sin 9 sin <|)
OEQ = jJDcos^/tf) - Csu^co/jp)]cos9 - [Bcos(a)/tf) - Asin(co/J?)]}cos§
r\GHfy ={[Acos(cfl/>)+ Bsin(co/J?)]cos0+ [CCOS(G)/ (? )+ Dsin(co/ye)]}cos<t)
oEfy = -{JDCO^CO/^) - Csin(a>/£)] - [Bcos^/tf) - Asin(co/>)]cos 9}sin <\>
r\aRQ = {[Acos^/^) + Bsin(co/#)] + [Ccos(co/^) + Dsir^w/^ )]cos 9} sin 0
Antenna Q 163

—[l-cos(2a)/^)] +—sin(2a)/^) + — [l + cos(2a)/^)Hsin20

Wy
(',)- -a + -6a fc-^K)]
+^ (i+cos2e)
+ —3-cos(2u)/^)- sin<2co/>)

8 4 1
-[l-cos(2(0/y?)] +—sin(2a)/^) + — [ l + cos(2co/^)]lsin26
4
^-[l - cos(2co^)] W1 + cos2 e)

**('*) = \ \ — \-X - c o s ( 2 ^ ) ] + — cos(2co^)-f — J 8^(2(0^) Ml + cos2 e)

1 f 1 2 ^ 2 1^ 1
N
r ( ^ ) = —W[1-cos(2co/yP)]+ — cos^u/^)- — sin(2(D/>)Kl + co S e) :
2n l a 2 a Ka3 5
aJ J

NQ(^) = — cos(2aw> ) - —— — I sin(2aw> ) iin8cos8; N (|) (/ v? ) = 0

.a \a a

*§(') = - ^ 7 1 r[l-cos(2M/>)]+ - s i n ( 2 u ^ + —-[l+cos(2co/>)]

3t [(Jta) {kd) (to)
8rc I, , y, [ 2 1
P= T l U - c o s ^ c o / J J - ——• sin^O)/^) + cos^G)/^)
3TIKJ
(to) (hif (A»)

QS
2(i»y
{.•V l + 4(>to) 4
+
(A»)
Gain = 3

Table 3.15.2 Collocated, Radiating z-Directed Electric and Magnetic Dipoles

Producing Linearly Polarized Fields

Output power is centered on the positive z-axis and is equal to zero on

the negative z-axis. The far field is circularly polarized. Dynamic values are
listed in Table 3.15.3; again power and energy terms proportional to cosG are
ignored. Q is reduced below that of Table 3.15.1. This is because some of
the energy that is source-associated with parallel radiating elements becomes
field-associated with orthogonal elements.
164 The Electromagnetic Origin of Quantum Theory and Light

2 2 | 2
— + — sin 9 + — cos(2co^) + —-sm(2(a^) sin 9cos(2(|))
4 6
a CT a a J a

W
Y{'JI) = -
{T7?h«*^
-HI ~ - ~ + — ^ cos(2co/>)+ — - — ]sin(2o)/'y?))sin29cos(2(())

V
S(^) = H ( — + —jsin2e+/f— cos(2w^) + —sin(2co^)\sin 2 9cos(2<|>)l

1 1
(l + cos 2 e)
—+
H = 2 - 6
t('j{) -
a 2a
+( — - — 1cos^a)/,,,) + — - — ]sin(2co/>))sin2 9 cos(2(|))

N r = ±|J_( 1 + c o s 2 ( e ) ). — - — cos(2co/) + — - - 7 |sin(2co/) sin (e)cos(2<|>)

a a J Va a

N9 = — j — C O S ^ C O / ^ ) - — sin(2co//f)[sin9cos9cos(2<|))
T) l a \<ya5J J
NL =—•{- — cos9 + ( 005(203/^)+ sin(2co/J?))[sin9sin(2(|))
3 4
r| [ a \ a VCT3a V /J
8TO 1 1 8TI

^7_3(>fc) - + ( ^ ) j 3
3TI^:
1 1

3(iaf {*»)

Table 3.15.3 Collocated, Radiating z-Directed Electric and Magnetic Dipole Pair
Producing Circularly Polarized Fields

3.16 Q of Collocated, Perpendicular Electric Dipoles

This example is two collocated, perpendicularly directed electric dipoles
driven nil out of phase. These force fields, Eq. 3.16.1, result if coefficient
F(l,l) = 1, all others are zero, and; = i. Output power is centered at 6 = nil
and circularly polarized. As with the perpendicular electric and magnetic
dipoles of Table 3.15.2, since the field components strongly overlap energy
Antenna Q 165

densities are combined before the integration constant is evaluated. The

result is similar to in-phase electric and magnetic moments in that the
standing energy of the two dipoles peak out of phase, and the peak standing
energy is about equal to that of a single dipole. Since the output power is
twice that of a single dipole, Q is reduced by about a factor of two. This is
similar to, and for the same reason as, the reduction in Q shown in
Table 3.15.1.

cx2Er = 2[Bj cos(u)/> - 0) - Aj sin(wtjf - (())] sin 0

CTEQ = [DJ COS(O>/> -<]>)- Q sin(to/£ - ()))]cos 8
crriH,), = [Aj cos(u)/# — (b) + Bj sn^CD/tf - <j))]cos 0 (3.16.1)
cEfy = [Cj cos(co/tf -<!>)+ D t sin(u)/^ - (J))]
or|H0 = [Bj cos(co/y? - §) - A{ sin(co/£ - §)]

The azimuth energy flow produces a z-directed angular momentum.

Results are listed in Table 3.16.1. Different from the previous electric dipole
cases, but like the counterpart of Table 3.15.3, a term proportional to a~
appears as part of the outgoing energy. It forms part of the standing energy
in other special cases; among the results is that the u>8/Nr ratio is no longer
equal to c. The energy shifts position from standing to traveling energy and
results in Q being further reduced, to the value of Table 3.15.3.

3.17 Four Collocated Electric and Magnetic Dipoles

and Multipoles
Dipoles
Table 3.17.1 tabulates results of four collocated electric and magnetic
dipoles generating equal time-average output powers. The electric dipoles
have the same orientation and phasing as the example of Table 3.16.1. The
magnetic dipoles have the same orientation and phasing as the electric ones,
and produce the dual results of Table 3.16.1. Electric and magnetic dipole
pairs lie along both the x- and y-axes. Two pairs of electric and magnetic
dipoles are formed into two units and driven in phase quadrature. Since the
resulting circularly polarized field components overlap, the energy densities
166 The Electromagnetic Origin of Quantum Theory and Light

—A\ - COS(2G)/£ - 2<|>)] + —r sin(2to/^ - 20) + —j [l + cos(2(0/> - 2(|))] [ sin 2 Q\

la a a J

M<*)~( (Jf + ^)( 1+cos2e )

-J 4+4" cos 2c0/
( >" 2 ( | ) ) + rT ? sui(2<D/*-2<|>) sin 2 9
\a a o ) \a aJ
P r i o A l
«$(/*) = -j-g-[4-3cos(2(afr-2<|>)] + -jsin(2(o/ J? -2<|>)+ — [l + cos(2(0/^-2<|>)] [sin 2 9
4 la a a J

^+2^)(1+ C
°s2e)
»s('/>)=2
-a .T " „T, , cos(2(O4.-20)+ i ^4T "- --5-| s i n ( 2 a w > - ^ ) sin 2 6
rr /r / \ (T (J

Nr(/*) = ^ H - ( l + cos 2 e)+ f ± - A j C O s ( 2 ( 0 / j 7 - 2 ( t . ) + ( ^ - ^ j s i n ( 2 ( 0 / J ? - 2 ( t ) ) sin 2 9

N Q ( / ^ ) = — —7005(2(0/^-2(1))- —r- r sin(2a>{ ? -2<|msin9cos9

*1 l a Va av J
N
*(^) = - ( ^ 3 + ^ 5 J + f 4 - - ^ - | c o s ( 2 ( 0 / J ? - 2 ( | ) ) + 4 r sin(2(o/ J? -2(|)) sin9

8TIE 1 1 871
ffi = r+ -
3^ 3(>fo)3 (*») lr\£2

1 1
Q^ r+- Gain = —
3(^)3 (*») 2
Table 3.16.1 Radiating x- and ^-Directed Electric Dipoles, Circular Polarization

are combined before the integration constant is evaluated. Since no net

energy leaves the traveling wave in its journey from the antenna surface to
infinity, there is no source associated standing energy. The vector fields are:

a Er =2[B, COS((0/ R -<|))-A 1 sin((0/j ? -(|>)]sin6

a 2 TiH r = -2[Ai cos((0/R - <|>) + B, sin((0/R - <)>)]sin9

aEe = { [ 0 ] COS((0/ R -<|>)-Cisin(G)/ fi - ^ ^ c o s B + fAj cos((0/R -(()) + B, sin(oo/R -<|))]}

anH,!, = {[A! COS((0/ R - <|>) + B] sin((o/R - <(>)] cos 9 + [D! COS((0/ R - <|>) -C^ sin((o/R - <)>)]}

aE^ = {-[B| COS((0/ R —<))) — A t sin(a>/R - (|))] cos 9 + [C] cos((D/R - (|>) + D, sin((o/R - <|>)]}

ar]H e = {-[C, cos((D/R -<(>) + D] sin((o/R - (|))] cos 9 + [BJ COS((0/ R - <|>) - A, sin(a>/R - <|>)]}
(3.17.1)
Antenna Q 167

Resonance depends upon the equality of the field coefficients

F(l,l)and-G(l,l). If the criterion F(l,l) = -G(l,l) is met, and j = i, the
input equivalent circuit is a simple resistor and Q is equal to zero.
The pressure on a virtual sphere generating these field modes follows in
the same way as the example of Section 3.13. The radial field component of
the Maxwell stress tensor follows from the expression for wj(tn), and is:

f
T =- — — sin 6 - — + — (l+cos2e)- —cosel (3.17.2)
2 6 X
a a) ' a2 J

Wp
4 4 ^ . 2
sin z 9+
'
A + J_j( 1+cos 2 0 ) + ^ cos eJ
w§ = 0
/ 4 4
«s = - sin 2 0+ 2_ J_ (i+cos2eU^-cosel
Vcr a J Va 2 +
c6;

Nr = - ^ (l + cos 2 e)+ 2 + 4 COS0[

Z v
TICT [ ' V <ry

No = ^sin8
-no

N* = - -COS0 +
' 1 p sin 9
Tl Vcr ay
167r
rar n P =
^ =° 7T2"
3ri^ z
Q>0 Gain = 3

Table 3.17.1 Four superimposed x- andy-directed electric and magnetic dipoles,

circular polarization

This equation shows the surface pressure to be independent of time. In free

space, all fields are continuous across virtual boundaries and the pressure
gradient is given by the spatial directional derivative of Eq. 3.17.2.
Figures 3.17.1-3.17.3 show the radiation reaction pressure, Eq. 3.17.2,
on virtual spheres of radii ka = 5, 1, and 0.1 if all fields at smaller values of
168 The Electromagnetic Origin of Quantum Theory and Light

radius are equal to zero. The radiation reaction pressure is expansive in the
direction of the moments, 0 = JI/2, and compressive along the z-axis. In both
cases the pressure decreases rapidly with increasing radius.

60 90 120 180
G (Degrees)

Figure 3.17.1 Radiation reaction pressure on a virtual, conducting sphere of

electrical radius ka = 5, if all interior fields are equal to zero.

Although the total pressure in the xy-plane is time independent, the fields
that produce it and the sources upon which they act are not. For example, the
radial tensor components due to the electric and magnetic fields are:

1 1
r 1+ + 1+ cos(2co/#) + -—-sin(2co/£)
2 ~'. ^ {kaf {ka)1
IV
2e 1
r 1+ 1+ cos( 200/^) - -—r sin (2co/^)
2 {kaf (kaf {kaf
(3.17.3)

Although the sum of the electric and magnetic field terms, which is equal to
the total pressure, is constant, the electric field acts only on electric charge
densities and the magnetic field acts only on the current densities. On a rigid
surface the difference is not significant. However, on a surface sufficiently
Antenna Q 169

flexible to respond to local pressures, differences may produce source

turbulence.

£ -l -

-2

30 60 90 120 150 180

0 (Degrees)

Figure 3.17.2 Radiation reaction pressure on a virtual, conducting sphere of

electrical radius ka = 1, if all interior fields are equal to zero, see Eq 3.17.2.

8 105

6 105

4 105

2 105

-2 105

-4 105

-6 105
30 60 90 120 150 180
0 (Degrees)

Figure 3.17.3 Radiation reaction pressure on a virtual, conducting sphere of

electrical radius ka = 0.1, if all interior fields are equal to zero, see Eq 3.17.2.
170 The Electromagnetic Origin of Quantum Theory and Light

To examine the significance of a zero Q, note the peak value of standing

energy about any radiator is equal to:

This energy returns to the antenna when, for example, a shift in frequency or
source shutdown occurs. By Eq. 3.17.5 no matter how large an amount of
energy is available for storage in the field and no matter how small is the
acceptable time average power, there is a minimum acceptable value of Q.
Applications that appear to be impractical for dipole antennas because of the
magnitude of required energy include radiative decay of atomic states and
very low frequency communication. Quite differently, the lower limit on Q
tabulated in Tables 3.17.1 and 3.17.2 shows that with Q equal to zero,
energy, once radiated, never returns to the source. There is no lower limit on
the antenna diameter-to-wavelength ratio and the full amount of field energy,
once radiated, ultimately continues on to the far field. This does not imply
that during steady state operation, at a fixed time-average output power,
there is less local field energy; it implies all standing energy will ultimately
travel outward to the far field, and not return to affect the source.

Multipoles
An expanded antenna is obtained by replacing the dipoles of Table 3.17.1 by
omnidirectional modes of arbitrary order. The modes are located, oriented,
and phased similarly to those of Table 3.17.1. The resulting vector fields if
coefficients F(^,l)= 1, G(^,l) = - 1 , all others are equal to zero, a n d / = /,
are:

Let S^ = p](cose)/sin9 and Te = dP](cos9)/d6

a 2 E r = £(£ + \)[Be cos^/jj. -§)-Ae sii^CO/? - §)]Se sin0

a2r\Rr = -£(£ + \)[Ae cos(a>/£ -<)>) + B^ sinfco/^ -ty)]Sesin6
(3.17.5)
Antenna Q 171

oE e = {[Df cc^oo/tf -ty)-Ce sin(co^ - <\>)]Te + [Ae cos(co/> - <|>) + B^ sin(co/^. - 0)]S^}
crqHq = {[Af COS((!)/> -<(>) + B^ sin(mtR - <|))]T^ + [D^ COS(OO/> -<)>)- C^ sin( w/,,. - §)]S(}
aE^ ^ - [ B ^ c o s ^ / ^ -((>)- A^sii^ci)/^ -<t>)}l> + [C^ cos( (1)/^ -<(>) + Df sii^co/^ -<t>)]S;}
anHg ={-[C^ COS((I)/# - <|>) + Df sii^CD/^ - <|>)]T^ +[B^ COS((I)/^ -(]>)- A( sin(co/^ - <t>)]S^}

+ 1
^ ) /»(A/
* 2 l B/) 2 \ c 2„ ; „2
A / ++ BD / S/sin^e

+-T(A/ +B/+c/ +D/k2+T/) +-S/T,(AA-BA)

W;=0

e
1
+^ ( A / + B / +C / + D/)(S/+T/) + —S^(A,D^-B^)

Nr {(A^-B^Q)(S/+T/) + (A/ + B / + C / + D / ) S ^
2

N e = - ^ ^ ( A A + B^)S£2sine

£+1)
N* = - [(A/D< - B^Q)S/I> +(A(2 + B^)s/]sin0
no
... 1671
3r|>tz

Q>0 Gain = - ^ + l) 2

Table 3.17.2 Four superimposed x- and y-oriented electric and magnetic multipoles
Power and energy results, tabulated in Table 3.17.2, show that the zero-Q
aspect extends through all modal orders.

Table 3.17.2 confirms that the zero-Q aspect extends through all orders.
Since modes of different orders operate independently, any combination of
such modes will have a net Q of zero.
172 The Electromagnetic Origin of Quantum Theory and Light

3.18 Numerical Characterization of Antennas

In an effort to confirm the Q model presented within this chapter, the
radiation properties of a radiating spherical surface have been numerically
modeled using finite difference time domain (FDTD). Accurate numerical
modeling requires the properties of the radiating sphere to be retained, while
eliminating properties of the driving network, a network which for the multi-
dipole source includes power splitters, cables, and possibly multiple phase
shifters. Source Q is based upon Eq. 3.5.11:

This equation is applicable to any antenna design, independently of its

complexity.
The technique makes possible determination of the energy that returns
from the standing energy field back to the antenna. The measurement begins
by driving the antenna to steady state. Time average output power P is
numerically obtained by integrating over a virtual sphere that circumscribes
the source. After steady state operation has been reached, the voltage source
is turned off, after which the local standing energy field collapses. The
source-associated portion of the standing energy returns to the antenna from
which, in turn, it is either reflected back into space or absorbed by the
antenna. In the numerical model, the antenna is driven by a 50 Q. source that
absorbs returning energy. The absorbed and reflected energies are summed
to obtain Wpk.
The time domain technique avoids spurious errors due to unwanted
power reflections within the feed network. For example, when two antennas
are driven by a single generator, through a power splitter and a feed network
in which one arm has a nil voltage phase shift, the waves reflected from the
antennas back to the generator are TI out of phase and cancel. No reflected
power is measured. This null result does not mean that the Q of the antenna
is zero, or that the antenna input impedance is purely real. Rather it indicates
that techniques for determining Q are required that separate transmission line
effects from antenna performance.
The analytical techniques used to determine the radiation Q of a source
necessarily solve for the steady state fields external to a virtual sphere
enclosing the radiation source. Doing so ignores standing energy at radii less
than the length of the antenna arms, hence the analytic expressions for the
Antenna Q 173

standing energy are inherently too small. In contrast, the numerical

technique accounts for all standing energy in the near field. It forms a check
on analytic techniques and a guide for experimental implementation. The
total energy returned to the antenna is a simple sum over the energies
returned and reflected by each element.
The numerical method for determining Q begins by determining the
power that passes through a large radius, circumscribing sphere. That power
is put equal to P in Eq. 3.18.1. After source turn-off the voltage across the
source resistor, Vj n , is determined and used to calculate the source current,
Iin. These values are combined and integrated over time to obtain the
returned energy:

Returned = J l i n ( / ) V i n ( / y ^ J [ l i n ( ^ ) ] 2 R s ^ (3-18.2)

Rs is the 50 Q. source resistor. I;n is the current that flows through the
resistor and Vjn is the voltage across the antenna terminals. Both voltage and
current are determined using FDTD, then the time integral of Eq. 3.18.2 is
evaluated and entered as the energy portion of the numerator of Eq. 3.18.1.
The portion of the original (at source turn-off) standing energy that escapes
outward is obtained by calculating the instantaneous power on the surface of
an encompassing virtual sphere of radius R at times t> R/c, integrating over
both the surface and time. Details of the finite difference time domain
(FDTD) code used in this work are given in the two references by Liu et ah
For certain antennas the standing energy varies with time, hence use of
Eq. 3.18.1 to determine Q requires repeating the calculation process over a
range of turn-off phase angles to obtain the peak value PFspk- However for
other antennas, such as a turnstile antenna with the two dipoles driven in
phase quadrature, the source-associated standing energy is time independent.

Biconical Dipole Antennas

A field-based analysis of a single electric dipole is given in Section 3.12; Q
is listed in Table 3.12.1. A terminal-based analysis of the same antenna is
shown in Section 3.7 and listed in Eq. 3.7.15. Results are identical and equal
to:

Q i [i+^? (Jta)
(3.18.3)
174 The Electromagnetic Origin of Quantum Theory and Light

A 5° arm angle biconical dipole antenna was the basis for a numerical
analysis. The dipole was divided into 13 discrete radial segments; for ka = 6
each radial segment length is 0.073 X. To ensure steady state operation the
dipole was driven for eleven time-periods before the source voltage was
turned-off. During steady state operation power on a spherical surface of
radius R about the source was determined and entered as the denominator of
Eq. 3.18.1. For the first standing energy measurement, turn-off was done at
the most negative value of input power, point A of Figure 3.18.1. At each
field point for time t > R/c the fields collapsed. Some of what was standing
energy moves outward away from the antenna and some moves inward
toward it. Some of the collapsing power is dissipated in the source resistor
and some is reflected back into space. The returned energy is measured then
the process is repeated with a different cyclical turn-off phase until the
maximum returned energy is obtained. The maximum energy is substituted
into the numerator of Eq. 3.18.1.
The instantaneous power at the antenna terminals, P;n, is calculated and
plotted as the dashed curve of Figure 3.18.1, where outwardly and inwardly
directed power is respectively positive and negative. It has the form:

p(cr,/) = P[l + Y(CT)COS(2CG/>)] (3.18.4)

The steady state portion of the curve is dominated by the reactive term. For
the first iteration the source is turned off at time t = A, at which time there is
a maximum rate of reactive energy return to the antenna. After turn-off the
terminal power drops abruptly then takes what appears to be an
exponentially damped, oscillatory form. Oscillations occur at a wavelength
less than that of the driven field.
The instantaneous power at the surface of an encompassing, virtual
sphere one wavelength in radius is also calculated and plotted as the solid
curve of Figure 3.18.1. In the steady state regime since magnitude
3
Y(<J) = 1/CJ is a monotone, rapidly decreasing function of radius the peak-to-
peak magnitude is much less at the field point than it is at the terminals.
Antenna Q 175

0.002
;
0.0015

0.001
"—t
\
A Pout
Pin -----
\ / \ , -
0.0005
B
S-H
0 » H i f- ;

-0.0005 -
O
-0.001 \ / \
; ;
-0.0015
; A
-0.002 i
4.5 5.5 6.5

Time (Periods)

Figure 3.18.1 Time analysis of a single biconical TM dipole.

Pin is the instantaneous power at the antenna terminals and Pout is the
power on the surface of the encompassing, virtual sphere one wavelength in
diameter, shown at the same retarded time. Source turn-off occurs at the
most negative value of Pin, time t = A. Some standing energy travels inward
and is absorbed by the input resistor and some travels outward. At time B
the field collapse reaches radius R surrounding the antenna.

After source turn-off the output power remains continuous then becomes
increasingly positive until reaching a positive value larger than the
maximum steady state value, then decays to zero. The figure verifies that
there is continued emission of energy after the source has been discontinued,
and such energy can come only from what was once standing energy.
Since the source-associated standing energy of a single dipole is time
dependent, the measured Q depends upon the phase at which the source is
turned off. The numerically determined variation in source-associated
standing energy of a biconical TM dipole of electrical size ka = 0.6 is shown
in Figure 3.18.2 with t- 0 defined to be when Pj n is at its most negative
point.
Comparative values of Q calculated using Eq. 3.18.1 and obtained
numerically using the described technique are shown in Figure 3.18.3. As
expected, in all cases the numerically calculated values are slightly larger:
the analytic expressions do not consider the standing energy contained at
176 The Electromagnetic Origin of Quantum Theory and Light

radii less than the antenna arm length and the numerical calculations do.
Furthermore for ka > 1.1 octupole moment radiation becomes important.
Such effects are accounted for in the numerical analyses but not in the
analytical curves. The octupolar moment introduces oscillations in the
powers and energies and result in the oscillatory Q behavior of
Figure 3.18.3.

Turnstile Antennas
A turnstile antenna consist of two collocated and spatially orthogonal
electrical dipoles. A turnstile antenna is the simplest multi-element antenna
for which theory shows that Q is dependent upon inter-element phasing.
When the two dipoles are driven in phase the far field is linearly polarized
and Q is the same as for a single electric dipole, Eq. 3.18.3. However when
the dipoles are driven in phase-quadrature, see Section 3.16, the fields are
circularly polarized and Q is given by:

l l
Q= 3
3(^) (*")
(3.18.5)

8 10"'

7 10"'

J2 6 Iff15
•S 5 Iff"
& 4 JO' 3
l-H

C 3 10"13
W
2 10 1 3

1 Iff" 0.1 0.2 0.3 0.4 0.5

Source Cycle Turn-Off Point (Period)

Figure 3.18.2 Source-associated standing energy for a biconical TM dipole
The dipole is of electrical length ka = 0.6 and the plot is a function of source
turn-off point. The time reference zero point is the minimum value ofPjn.
Antenna Q 177

a
c
o
•!-H
• 1—1

Figure 3.18.3 Numerically (open dots) and theoretically (solid dots) determined
radiation Q of a biconical antenna versus electrical length ka.

In the electrically small limit, the relative phasing of the dipoles produces a
factor of three difference in Q. The difference serves as an important test
case for the different models.
To make a comparative numerical analysis each biconical electric dipole
was divided into 13 discrete radial segments; when ka = 6, each radial
segment is equivalent to about 0.073 X. Plots of numerically determined Q
versus relative electrical size, ka, for a turnstile antenna when the dipoles are
in phase and out of phase are shown in Figure 3.18.4. The data show that the
relative phasing between the two dipoles affects Q. Since the analytic
solutions do not account for standing energy within the inner region of the
antenna calculated Q values are expected to be larger than the theoretical
predictions and Figure 3.18.4 shows that to be the case. Also as expected the
largest fractional reduction in Q occurs with a phase difference of 90°. As
with a single biconical antenna when the turnstile antenna supports linear
polarization the standing energy is time varying. It is, therefore, necessary to
determine the source turn-off point that produces the largest calculated Q.
This point was determined to be the same point it was for a single biconical
antenna. A plot of Q reduction in switching from in-phase to phase
quadrature as a function of ka is shown in Figure 3.18.5.
178 The Electromagnetic Origin of Quantum Theory and Light

—•-.. 90° relative phase diff. FDTD

-B— 0° relative phase diff. FDTD
• 0° phase diff. theory
90° phase diff. theory
G
O
1
• 1—I

Figure 3.18.4 Numerical and analytical values of radiation Q versus ka for a turnstile
antenna, when phased to support linear polarization and when phased to support
circular polarization.

0.76
o
O 0.74
@
a
o
0.72
O
Q\ 0.7
<&
O 0.68
o
•i—i
-4->
0.66

0.64

ka
Figure 3.18.5 Fractional reduction in Q versus ka obtained by shifting the relative
phase between dipoles of turnstile antenna from 0° to 90°.
Antenna Q 179

The figures show that the relative phasing between the dipoles affects the
radiation Q of turnstile antennas. This change in Q is due to a change in field
structure that, in turn, affects the fraction of the standing energy that returns
to the radiating source.

3.19 Experimental Characterization of Antennas

Biconical Dipole Antennas
A technique similar to the numerical one may be used to experimentally
determine antenna Q. The block diagram of the experimental system for a
single port antenna is shown in Figure 3.19.1. A wave generator drives a
circulator that in turn drives the antenna. The return from the antenna passes
back to the circulator and from it to an integrating oscilloscope; the portion
of the power reflected from the antenna back into space is unknown. The
experimental procedure is to obtain steady state operation, determine the real
power P using a network analyzer, then switch off the generator. Energy
returned from the antenna after source turn-off is directed by the circulator to
the transient-capturing oscilloscope, put equal to the source-associated
standing energy and entered in the numerator of Eq. 3.18.1. The Q
measurement technique isolates antenna performance from the feed network
and enables characterization of the antenna itself.

Antenna
Circulator
Arbitrary
Wave
Generator

Oscilloscope

Figure 3.19.1 Experimental set-up for determining the radiation Q of a single

antenna.
180 The Electromagnetic Origin of Quantum Theory and Light

As detailed in the C. A. Grimes et al. references, a Tektronix 500 MHz

arbitrary waveform generator (AWG610), which is able to terminate the
waveform virtually without a measurable transient, was used as the source
generator. The antenna was suspended in an anechoic chamber, driven from
the AWG610 through a circulator, with the reflected waveform captured
with an oscilloscope through the other circulator port. A HP 54845A
oscilloscope with a sampling rate of 8 G samples/second and an advanced
triggering option that capture waveforms up to 1.5 GHz was used to capture
the transient signal returning from the antenna after source turn-off. All
components in the experimental setup were 50 Q. devices. The generator
output power in steady state was determined from the measured voltage and
found to be about 7.1 mW (8.5 dBm). The circulator effectively divided the
input and reflected signals so the generator always saw the network as a
50 O load and delivered the same power.
Using the programming capabilities of the AWG610, a waveform of
frequency 450 MHz was generated and delivered to the antenna. The
duration of the source signal was pre-selected, the antenna was driven until it
reached the steady state, and then switched off. There was no detectable
transient response. It was found that signals of 25 ns duration, about 12
periods at 450 MHz, were enough to reach steady state. After the waveform
was turned off the power returned from the antenna to the oscilloscope was
measured and time integrated to obtain the source-associated energy. A
typical reflected power waveform for a wire dipole of length 0.2X is shown
in Figure 3.19.2. All oscillations after turn-off are due to returned power.
The time-average power radiated by the antenna was measured
indirectly. The scattering parameters of the three port network of
Figure 3.19.1 without the generator, oscilloscope, and antenna were
measured with a HP8753D network analyzer. The antenna was then
connected to the system and the network analyzer was used to determine the
input impedance. This was sufficient to permit calculation of the voltage and
currents at the terminals of each port when the generator produces its
measured voltage, the antenna presents its measured impedance, and the
oscilloscope supplies a fifty-ohm load. The calculated real power at the
antenna terminals is equal to the radiated power. The power reflected and
captured with the oscilloscope was also calculated. Calculated and measured
values of reflected power were the same.
Antenna Q 181

3 1
' ' i ' ' ' 1 ' • ' j ' ' ' .

2.5 i A i
: ; j j1 [
j Source turn-off point i
2 1 ! j ' i

A ' '
1.5 i \ : :
t i l !

O 1 i-'l!
OH
j
0.5

0 i :W
20 22
41 In A ilA i
24 26 28 30

Time (ns)
Figure 3.19.2 Measured values of reflected power for a 0.2A -electric dipole antenna.

. . 1 • • , • 1 • • , • 1 . , • • 1 • , • , j . • , , 1 • _
40
;
35
heory : :
a 30
1*

11 \ i >:
y -1
••••--- Q - Experiment : :
G 25
O 1 • | : " "~
•!-H
20
•i-H
T3
15 n ! i :
\* ! ! •

10
5
0
- f^Ld- -. ,
•—
. i , i

0.5 1.5 2.5

ka

Figure 3.19.3 Experimentally determined Q versus electrical length of thin-wire

electric dipoles.
182 The Electromagnetic Origin of Quantum Theory and Light

Turnstile Antennas
The radiation Q of the turnstile antenna is measured in a way that is similar
to antennas with a single input port. The network needed to characterize a
turnstile antenna is shown in Figure 3.19.4. A hybrid 3 dB-splitter forwards
equal power to each dipole, a phase shifter adjusts the phase difference
between the dipole drives, an attenuator compensates for the loss in the
phase shifter, and circulators separate incoming from reflected signals. The
antenna is a two-port system, the scattering parameters of which are
measured by the network analyzer. The scattering parameters of a five-port
network (six with the hybrid port connected to a 50 Q. resistor) were
measured. Then using network theory the power radiated by the turnstile
antenna was determined. This approach accounts for parasitic coupling
between the two dipoles. The oscilloscope captured the reflected waveforms
and the reflected powers were determined from them. The source-associated
standing energy of the turnstile antenna was determined by summing the
time integrals of reflected powers from the two dipoles.

Phase shifter Circulator

Arbitrary
Waveform
Generator <^y—^ a
3 a
o a
o, a
Circulator
^ircula 6 B
o a
<2>
Attenuator
B

RA RJi

Oscilloscope

Figure 3.19.4 Experimental setup for measuring Q of the turnstile antenna.

A turnstile antenna was implemented using thin wire, equal length

dipoles and measured using the setup described in Figure 3.19.4. Q was
determined with both the drives in phase (linear polarization) and the drives
in phase quadrature (circular polarization.) Measured values of Q versus the
electrical length of the lines are plotted in Figure 3.19.5. Results confirm that
the radiation Q of this antenna is a function of the difference of driving
phase between the two dipoles.
Antenna Q 183

40

35

—e— Q - Linear Polarization

o 30

25 L.......!»
.—•—- Q . Circular Polarization
a
o 20
•i—i
•i-H
15
•i \
10
1 V
5

0
0.5 1 1.5 2 2.5 3 3.5

ka
Figure 3.19.5 Q of a thin-wire, turnstile antenna versus electrical length of the
dipoles, for in-phase drives and phase quadrature drives.

3.20 Q of Collocated Electric and Magnetic Dipoles:

Numerical and Experimental Characterizations
The antenna discussed in Section 3.17 consists of four collocated dipoles; an
electric and magnetic dipole pair radiating equal powers is oriented parallel
with the x-axis and an identical pair is oriented along the y-axis. The
configuration is depicted in Figure 3.20.1. The electric moments are
implemented as straight wires and the magnetic moments as rectangular
loops. The magnetic loops are positioned to the side of the electric dipoles in
a way that produces strong coupling between the x-directed electric moment
and the ^-directed magnetic moment, and symmetrically between the other
pair. By the analysis of Section 3.17 if driven with the proper phases the
lower limit on radiation Q is equal to or greater than zero.
Early attempts to characterize this antenna design used separate feeds for
each dipole and were unsuccessful due to unwanted and interfering power
transfer between dipole feeds. This implementation uses two separate dipole-
pair elements, with equal power from the electric and magnetic moments.
With this design rather than interfering, the coupling appears to contribute to
184 The Electromagnetic Origin of Quantum Theory and Light

the desired outputs. The radiation Q of the antenna system is determined

using the numerical and experimental techniques detailed in Section 3.19. Q
values were measured as functions of the phasing between the dipole pairs
and the relative electrical size, ka.
Two things determine the relative phasing of the different radiators: the
driving phase and local coupling. Similarly directed dipoles are driven by the
same set of terminals and by the strong local field interaction between the x-
directed electric moment and v-directed magnetic moment, as well as
between the other two of elements. It is found that if the two driving ports
are in phase the radiation is similar to that of Table 3.15.2 and Q is given by
Eq. 3.18.3. With the driving ports in phase quadrature, the generated
radiation is similar to that of Table 3.17.1, for which Q has no analytical
lower limit.

Figure 3.20.1 Implementation sketch of antenna comprised of two dipole-pair

elements, each element provides equal TE and TM power.
Lines of different thickness differentiate the two sets of dipole-pair elements.

The single dipole pair embodiment is shown in Figure 3.20.2. A Method

of Moments (MoM) analysis was done to ensure that the elements radiate
equal TE and TM power. The fields on the surface of the smallest virtual
sphere that circumscribes the radiating elements were computed using NEC4
MoM. Using the technique described in Section 3.18 the calculated fields
were equated to the equivalent terms in a multipolar field expansion to
Antenna Q 185

determine the TM dipole field coefficient F and the TE dipole field

coefficient G, see Eq. 3.11.1. Figure 3.20.3 shows the calculated TE/TM
power ratio plot for the structure of Figure 3.20.2 with loop sides a = 111 =
12 cm. As shown in Figure 3.20.3 for these dimensions the element radiates
equal TE and TM power at 166.67 MHz. Since the dimensions scale linearly
with frequency for loop side a = £11 = 4 cm the equal power frequency is
500 MHz.
The four-dipole source was modeled numerically. Since straight-wire
elements were used for the antenna implementation, FDTD computations
were made using a rectangular, three-dimensional computer code based on
the Yee cell. The problem space was chosen as 120 x 120 x 120 cells, with
cell dimension Ax = Ay = Az = 5 mm; a perfectly matched absorbing
boundary layer was used to terminate the computational space. Each
radiating element consisted of a square loop and a straight-wire electric
dipole. For the numerical computations, the dimensions of the antenna were
held constant at loop side length 12 cm and electric dipole length 24 cm. The
operational frequency was varied above and below 166.67 MHz, the
frequency at which the TE and TM time-average powers were equal. For
experimental characterization a thin-wire antenna of loop side a = ill = 4 cm
was built and tested in an anechoic chamber.

z
72

Figure 3.20.2 A Single Electric and Magnetic Dipole Pair

186 The Electromagnetic Origin of Quantum Theory and Light

10* • ' ' • ' I ' ' • • I • ' • ' I • ' • I ' ' • ' I ' ' • ' I • • ' ' I ' • ' ' -_

'•& 1000

100

o
10

0.1
50 100 150 200 250 300 350 400

Frequency (MHz)
Figure 3.20.3 The TE/TM power ratio versus frequency from the single dipole-pair
element pair shown in Figure 3.20.2, with dimensions a = 111 = 12 cm.

30 1 1 1 1 , , i , i , i
1
-
—•— 0° Phase difference \»
25 - ---a---90° Phase difference

o ; \
20
o
\
% •

15
•i—i
\
a
10

, . , 1 , 1 , , , ,,
0 0.1 0.2 0.3 0.4 0.5

Source Cycle Turn-Off Point (Voltage Period)

Figure 3.20.4 Numerically determined Q as a function of source-turn off point.

Referenced to the input power minimum, for dipole-pair elements in phase
and phased to support circular polarization; ka = 0.42.
Antenna Q 187

The FDTD-determined radiation Q of the antenna for which ka - 0.42

versus the source turn-off point is shown in Figure 3.20.4, relative to the
minimum input power point. Theory indicates that the source-associated
standing energy is time varying for all relative phases except 90°, when the
dipole pairs support circular polarization. As seen in Figure 3.20.4, Q is
independent of source turn-off point when circular polarization is
maintained. However for other relative phases, Q varies with source turn-off
point; the correct value of Q is the largest value that is determined when the
source-associated standing energy is a maximum.
The numerically and experimentally determined radiation Q of the
antenna at ka - 0.42 versus phase difference between elements is shown in
Figure 3.20.5. In agreement with theory, the radiation Q is dependent upon
relative phasing between the antenna elements. When driven in phase Q is
approximately that of an electric dipole of the same size. When driven out of
phase antenna Q is reduced by an approximate factor of 4.5 from the in-
phase results. For this relative electrical size, the measured Q value is
approximately a factor of three below the minimum Q value determined by
Chu for an antenna of the same electrical size.

30

—*— Numerically determined values

—•&— Experimentally determined values / i
25 V
--
"*\ A' " '.
O
e 20
- V- /
o
• *—i
V\ i Chu Limit ; /
•4—> -
15

\ \ik.. .//
10
-
3
• ^ '• ~ ' ' ' ^ *

40 80 120 160

Relative Phase (degrees)

Figure 3.20.5 Numerically and experimentally determined Q versus relative phase

between dipole-pair elements; ka = 0.42.
188 The Electromagnetic Origin of Quantum Theory and Light

The sensitivity of Q to distance along the z-axis between the elements is

shown in Figure 3.20.6. Using antennas for which ka = 0.42 the antennas
were displaced in steps of 5 mm for the numerical model with dimensions of
a = 1/2 = 12 cm. Steps of 1.67 mm were taken for the experimental work
with dimensions of a = 112 = 4 cm. With 90° relative phasing between the
dipole-pair elements the radiation Q is respectively small and large when the
displacement is small and large. In contrast, when the four dipoles are in
phase the radiation Q is approximately that of an electric dipole of the same
ka independently of the spacing. Experiments moving the dipole-pairs
relative to each other in the other two dimensions showed similar results; the
Q of 90° phased dipole-pair is sensitive to relative location, i.e., modal
coupling, and the Q of the in-phase dipole pairs is not.

o - Relative phase = 0° numerical

--a-—Relative phase = 90° numerical
a 20 - •- - Relative phase = 0° experimental
o -• - - Relative phase = 90" experimental

•!-H
15

.iP-
:f
10 ..-r

5 10 15 20 25 30 35 40

Separation of Dipole-pairs Along Z-axis (step)

Figure 3.20.6 Numerically and experimentally determined Q as a function of spacing

between dipole-pair elements.
Step size was 5 mm with the numerical model and 1.67 mm with the
experimental model, ka = 0.42.
Antenna Q 189

The numerically determined Q versus the relative electrical size of the

antenna is shown for in-phase drives and for phase quadrature drives in
Figure 3.20.7. The trends shown by the numerical work were confirmed
experimentally over the more limited range of ka = 0.37 to ka = 0.42; the
circuit devices, not the antenna, determined the frequency limits. The
circulators imposed the low frequency limit and the Tektronix Arbitrary
Wave Generator AWG610 imposed the high frequency limit.

a
a
•i-H
o
a
•i—i

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Relative Electric Size ka

Figure 3.20.7 Numerically determined Q of antenna.

Dimensions a = £/2 = 12 cm, shows effect of relative phasing between
dipole-pair elements and electrical size ka.

At ka = 0.23 the Q of the circularly polarized antenna is more than a

factor of 20 below Chu's limit. The oscillations seen in the in-phase Q
results are due to higher order modes that cause variations in the outbound
3
real power. The in-phase results show the familiar ll{ka) dependence of Q
as the antenna becomes electrically small. In contrast, with the two dipole-
pair elements phased to support circular polarization Q is relatively
insensitive to frequency. The frequency response seen in Figure 3.20.7 is
indicative that the current and charge distributions on the dipole-pair
elements self-adjust to support radiation fields that minimize source-
associated standing energy and hence Q.
190 The Electromagnetic Origin of Quantum Theory and Light

We conclude that the numerical and experimental results support the

analytical Q results within the limits imposed by our irreducible differences
between among analytical, numerical, and experimental embodiments.

References
L.J. Chu, "Physical Limitations of Omni-Directional Antennas," J. Appl. Phys., vol.
19, 1163-1175(1948)
C.A. Grimes, G. Liu, F. Tefiku, D.M. Grimes, "Time Domain Measurement of
Antenna Q," Microwave and Optical Technology Letters, vol. 25, pp. 95-100
2000.
C.A. Grimes, G. Liu, D.M. Grimes, K.G. Ong, "Characterization of a Wideband,
Low-Q, Electrically Small Antenna," Microwave and Optical Technology
Letters, vol. 27, pp. 53-58 (2000)
D.M. Grimes, C.A. Grimes, "Power in modal radiation fields: Limits of the complex
Poynting theorem and the potential for electrically small antennas," J. Electro.
Waves and Appl., Vol. 11, 1721-1747 (1997)
D.M. Grimes, C.A. Grimes, "Radiation Q of Dipole-Generated Fields," Radio
Science, Vol. 34, 281-296 (1999)
D.M. Grimes, C.A. Grimes, "Minimum Q of Electrically Small Antennas: A Critical
Review," Microwave and Optical Technology Letters, vol. 28, pp. 172-177
(2001)
R.F. Harrington, "Effect of Antenna Size on Gain, Bandwidth, and Efficiency," J.
Research, Nat. Bureau of Standards, Vol. 64D, pp. 1-12, 1960.
G. Liu, C.A. Grimes, D.M. Grimes, "A Time Domain Technique For Determining
Antenna Q," Microwave and Optical Technology Letters, vol. 21, pp. 395-398
(1999)
G. Liu, K. G. Ong, C.A. Grimes, D.M. Grimes, "Comparison of Time and Frequency
Domain Numerical Modeling of Outbound And Local Power From Two
Perpendicularly Oriented, Electrically Small TM Dipoles," Int. J. of Numerical
Modeling, vol. 12, pp. 229-241, (1999)
 4. Quantum Theory

This chapter contains a review of quantum theory that is conventional in

many ways, but dramatically different in key concepts. It has been known
for over seventy years that electrons exhibit both particle and wave
properties, and for twenty years that they are nonlocal. Although wave-
particle duality plays an integral role in conventional quantum theory, the
full affect of nonlocality has not been developed. In this work nonlocality is
an integral part of the theory. Furthermore the full applicability of the
equations of classical electromagnetic theory is retained. It is important to
emphasize that the historic radiation reaction force on radiating charges is
based upon energy radiated permanently away from the charge; the resulting
reaction force acts to slow the motion of the charge and is much smaller than
the Coulomb force. However, an oscillating charge is enmeshed in a
standing electromagnetic energy field of its own making. As we show,
although the radiation reaction of the standing energy does not affect the
velocity of the radiator, it acts to distort and distend it; the magnitude is
3
(l/ka) times larger than the radiation reaction force as it has been
historically considered, and it is not small compared with the Coulomb force.
There is no accounting for this large radiation reaction force in the equations
of quantum theory.
Consider a point-electron as it nears an atomic nucleus. The forces acting
on it include both Coulomb and centrifugal forces. By the classical laws, a
point electron with an appropriate value of angular momentum will form a
temporary elliptical orbit, with the nucleus at one of the foci, then two things
happen to the orbit. First, the intrinsic and orbital magnetic moments interact
in a way that produces a continuous torque, and thus rotation of the orbit.
Second, orbiting objects accelerate and by the laws of classical
electromagnetism, Eqs. 1.7.3, accelerating charges radiate energy
permanently away from the system. Based on these laws, a point charge will
lose energy and spiral into the nucleus, where it will be annihilated.
However, electrons do not spiral into the nucleus. Instead, as Dirac points
out, there is a "remarkable stability of atoms and molecules."
191
192 The Electromagnetic Origin of Quantum Theory and Light

To further assess the "remarkable stability," consider the reactive energy

that accompanies far-field radiation. As discussed in Chapter 3, reactive
energy plays a dominant role in determining the properties of electrically
small antennas. Yet, although the atomic diameter-to-optical wavelength
ratio is on the order of 1/1000, the conventional quantum theory of radiation
ignores reactive energy. Neither does it supply the full and complete set of
electromagnetic fields that support the radiation process, as expressed in
Chapter 2 for scatterers and biconical antennas. The quantum theory
explanation is simply incomplete. This chapter contains the basis for an
examination of quantum radiation that includes effects of reactive energy.
The final explanation requires no special hypotheses and provides the full,
steady state radiation solution, including all fields.
In this chapter, we combine electron nonlocality with the full radiation
reaction force to construct an electron model. The model violates no physical
laws and no experimental facts, yet it permits us to develop a view of
quantum physics based upon electromagnetic field theory and the
conservation of energy. We postulate that an eigenstate electron is not a
particle, but extends throughout the state as a charged cloud. The exclusion
principle results if, on a fine enough scale of dimensions, the cloud consists
of granular units of charge. The units are minute compared with the total
electron charge and are in dynamic equilibrium. Each eigenstate electron is
modeled as an ensemble of such units. The mathematical results are the same
as those of the historically accepted quantum theory, but the philosophical
implications are not. Differences are discussed in this and the next chapter
and reviewed in the epilogue.

4.1 Electrons
An isolated, static array of point electric charges cannot be held in
equilibrium by electrostatic forces alone. Opposite charges collapse upon
themselves and like charges forever repel. For a charge distribution to be
stable something other than electrostatic forces must be at work. Since an
electron contains at least a dominantly negative charge and since it is stable
it follows that something other than electrostatic forces are present.
An electron's physical extent has important repercussions. An early
attempt to determine the size of an electron equated the electrostatic energy
2
to its mass using Eq. 1.3.14, W= XOQC . By classical electrostatics the energy
of a virtual shell of radius R carrying charge e is:
Quantum Theory 193

2
W=— (4.1.1)

The energy relationship results in radius Ri:

Ar = xl0~ 7 =2.82xl0~ 1 5 m (4.1.2)

m
v °v

The result, the Lorentz radius of an electron, is the radius an electron would
have if only electrostatic energies were present.
There have been many experimental and theoretical attempts to
determine electron size. One method is based upon accurately determining
the ratio between intrinsic values of magnetic moment and angular
momentum: g-factor data. Such measurements show the electron radius is
not more than about 10~22 m. Scattering experiments show that electrons
have no internal structure on the smallest scale of dimensions at which
measurements are possible, about 10 - 1 8 m. Theoretical quantum
electrodynamic arguments point to a structureless particle with a vanishingly
small radius, and string theory modifies the quantum electrodynamic result
to the order of 10 - 3 5 m. It is generally agreed, therefore, that whatever size
an electron may be, since atoms are typically about 10 m in diameter, free
electrons are much smaller than atoms.
An electron with a diameter much less than 10 m trapped within the
confines of an atom will either be captured by the nucleus or undergo
acceleration. Classical electromagnetic theory, see Eqs. 1.7.3, requires an
accelerating electron to radiate energy into the far field, yet the energy of
confined atomic electrons is fixed: it does not radiate. Both the requirement
and the absence of radiation are indisputable. It is widely believed, therefore,
that classical electromagnetic theory is not consistent with atomic stability.
Paul A. M. Dirac stated it succinctly: the forces known in classical
electrodynamics are "inadequate for the explanation of the remarkable
stability of atoms and molecules." A primary motivation for this work is to
show that atomic stability, atomic absorption, and atomic emission processes
do not conflict with classical electrodynamics but rather result from
Maxwell's equations and energy and momentum conservation.
194 The Electromagnetic Origin of Quantum Theory and Light

4.2 Radiation Reaction Force

To examine the fields produced by an accelerating, point electron of
charge e being captured as it spirals into an atomic nucleus, turn to the fields
of Eqs. 1.7.3. To keep the model as simple as possible, require the electron
velocity, v, to remain much less than the speed of light, c. In that event, the
radiation fields simplify to:

\ie (dv
E= rx\ rx — H: - —X; (4.2.1)
Anr dt J 4%cr\ dt

The distance from an origin located at the center of mass of the system to the
point electron and its velocity are, respectively, r and v. To further simplify
the algebra, although the acceleration is radially directed consider it to be ±z-
directed; a more realistic radial acceleration model complicates matters and
adds nothing essential. With ±z-directed acceleration the generated force
fields are:

„ u,£? 6 dv . n e 0 dv .
E=- —sin6 H (4.2.2)
4TI/- dt 4ncr dt-sin6

The radial component of the Poynting vector is:

-i2
dv
sin 2 6 (4.2.3)
c 4TIR dt

It follows that the radiated power is:

lie1 dv
P= (4.2.4)
6cn dt

For energy to be conserved it is necessary that the time-average radiated and

generated power be equal. This in turn requires a radiation reaction braking
force, FRR, acting on the electron to satisfy the condition:
Quantum Theory 195

2X
vd/+
\ie' 3v
j^RR d/= 0 (4.2.5)
0
67t<?
"" 0L a/
Time x is the period required for an integer number of rotations of the
electron about the nucleus. Doing the integral by parts leads to:

]ier2 d->2 v^ , lie2 dv

^RR _ • vat=-—— •v (4.2.6)
6C713/2 6CJI 9/

With oscillatory motion, such as an electron orbiting an atomic nucleus, the

right side is equal to zero, leaving the integral equal to zero. With p equal to
the electron momentum, the integrand is equal to zero:

\ie2 d2v \ie d p \ie(i)'

RR | real- 2 (4.2.7)
6CK d/ 6nc dp 6nc

Equation 4.2.7 expresses the time-average value of the radiation reaction

force on the electron due to dipole-radiated energy, as it permanently leaves
the system. This radiation reaction force is a braking force that acts on the
entire charge, and has no affect on the shape of the electron.
Results of Chapter 3 include that a dipole radiating from within an
electrically small region of radius a supports a reactive power that is larger
3
than the real power by a factor of about \l{ka) . In the mid-optical frequency
range and with an atom of radius 0.1 nm the factor is on the order of
3 9 9
il{ka) =10 .It follows that the reactive radiation reaction force is 10 times
3
larger than the braking force of Eq. 4.2.7. Multiplying Eq. 4.2.7 by \l{ka)
shows that the radiation reaction force due to the reactive energy is:

RR| reactive ~ (4.2.8)

6nza

Equation 4.2.8 expresses the time-average value of the reactive radiation

reaction force on the electron. It is the same order of magnitude as the
196 The Electromagnetic Origin of Quantum Theory and Light

Coulomb attractive force, and it is ignored by radiation analyses based upon

the historic interpretation of quantum theory.
A more formal derivation of the reactive radiation reaction force of
Eq. 4.2.8 follows by viewing a radiating electron as an electric dipole and
using the radiation impedance of Eq. 3.6.4. After substituting for the letter
functions and doing the long division, the impedance of Eq. 3.6.4 on a
virtual sphere of radius r = a/k may be written:

Z=n
Ai-'B, _/_j
(4.2.9)
a
11 2 3
- n a 4 + ?t]G5 +
/a

Defining a generalized voltage and current leads directly from Eq. 4.2.9 to
the voltage-current relationship:

T
f l • 2 ^ 3 4 . 5 ^
V,=
— + /r\G+r\a - fr\o - t]o + ir\o +...

Replacing a by (ka) gives, on the radiating surface:

^ 2 2 3 4 5
c . yd . 3Ja Aa . sa Ij (4.2.10)
— +/C0<7+0)
/O)^
/CO —r--u) -^-+/to —r+.
V c

After the second term, each succeeding term is (ka) times the previous one.
Therefore the magnitude of each succeeding term is down by a factor of
about 1000 in the mid-optical frequency range and the series converges
rapidly. Odd and even powers of (ka) respectively describe oscillatory and
outgoing energy. In this model of generalized force and flow the voltage is
proportional to the driving force and the current is proportional to the
magnitude of the dipole moment. To go from generalized parameters to
specific ones introduce unknown constant K by the relationship:
Quantum Theory 197

V! = KFX and Ij = /copj (4.2.11)

By definition, F\ is the total dipolar radiation reaction force and pi is the

electric dipole moment. Combining Eqs. 4.2.10 and 4.2.11, then switching to
time notation by replacing z'co with a time derivative shows that:

,, ( J- A2 J- A* J> A* „*• A^>

/j(/)^_p 1 (/) +tf _p 1 (/)-__ Pl (/) + _- ; i r P l (/)-__p 1 (/) + ...
K
V
(4.2.12)

The third term within the round brackets of Eq. 4.2.12 is the first term that
contributes to energy loss from the oscillator; it is equivalent to Eq. 4.2.7.
Making the equality shows that:

6na2
K=

Substituting K back into Eq. 4.2.12 gives:

e 4
r^l \ ^V) d2 /\ e d3 /N ea d1_4 / \
J z L J J
6nea 6neac &t '' 6jtec dt '' 6mc'* At
(4.2.13)

Equation 4.2.13 is the complete expression for the radiation reaction force.
Terms with an even or odd number of time derivatives, respectively,
represent reactive energy exchange or resistive energy loss. The first term is
a restoring force due to the local standing energy field. The second term is
the mass of the standing energy field. The third term is the first term that
leads to an energy loss from the system, etc.
The lead term of Eq. 4.2.13 is not small compared with other forces. It
may be written as:

2 2
f\V)=- IjPl(') = "T=* "J (4-2.14)
198 The Electromagnetic Origin of Quantum Theory and Light

The last term applies at full extension of the oscillation, where force
magnitude varies as the inverse square of the distance of oscillation. The
force is expansive on the charge itself, and acts to extend it to an ever-larger
size; it vanishes only if the charge is restructured to be nonradiating. If the
charge oscillates through a region on the order of several times the electronic
radii of Sec. 4.1, the expansive force is astronomical. This large force acts
until or unless the electron becomes large enough to encircle the nucleus.
Once that size the force rends the previously small electron into a stable,
dynamic, and evolving ensemble of charge and current densities that is not
small compared with atomic sizes and that is spread over the full eigenstate.
For a hydrogen atom the reactive radiation reaction force and Coulomb force
differ in magnitude by only a factor of 2/3. As an analogy, consider an oil
droplet: If isolated from external forces, surface tension acts to form it into a
spherical, liquid drop. If placed on the surface of a pond, the surface tension
force no longer dominates and the drop distributes itself over the surface.
We arrive at the following picture: as a point electron approaches an
attracting nucleus, it begins to orbit and, under the influence of orbital
acceleration, radiates. Radiation onset produces a radiation reaction force
that converts the point charge into an extended ensemble of charge and
current densities, the smallest size of which is determined by how finely the
electron charge is subdivided. Local forces, within and about the trapped
state, distribute the ensemble throughout the region. Internal forces also
require the ensemble to remain in continual motion. Several studies show
that there are an infinite number of stable arrays, see Kim and Wolf.
Evidence to support this electron model is discussed in both preceding and
coming sections.

4.3 The Time-Independent Schrodinger Equation

An integral part of the science of statistical mechanics is the analysis of large
numbers of identical, interacting particles taken as a single ensemble. The
state of the ensemble is specified by the positions and velocities of the
particles, and is sufficient information to determine the kinetic and potential
energies of the system. With particles modeled as realistically as possible,
there is little or no difficulty interpreting an experiment that measures the
ensemble-average of a kinematic variable. The state of an isolated ensemble
at any instant determines its future values. Since large ensembles contain too
many degrees of freedom to detail, no attempt is made to make precise and
Quantum Theory 199

detailed calculations. Instead, most probable values averaged over all

particles are calculated and assigned as ensemble-average values.
A single electron trapped by the Coulomb force of a positive nucleus
accelerates. By the arguments of Section 4.2, the electron is transformed into
an ensemble of charge and current densities. Since our present knowledge
does not permit solving for the exact array, in common with statistical
mechanics, it is necessary to consider each eigenstate electron on a statistical
basis. Physical properties are calculated by imposing conservation laws on
the ensemble. A primary result of imposing energy conservation on such an
ensemble is the Schrodinger equation. First published in 1926, it is a
mathematical description of the quantum character of electrons. Schrodinger
discovered the usefulness of the differential equation that now bears his
name:

S/lU(r) + A(r)U(r) = ff\J(r) (4.3.1)

2m
In this equation 27i7z is Planck's constant, A(r) is the electrostatic potential,
and U(r) is the wave function. The time average value of electric charge
density at each point in space is:

p(r) = eU*(/-)\J(r) (4.3.2)

Although Schrodinger discovered that solutions of his equation correctly

described the actions of electrons, he was not led to the result by first
principles. The equation is invaluable for describing atomic level phenomena
but solutions are statistical in character and there is no way to directly
determine a unique physical basis for the equation. For example, the
*
equation is no help in determining whether eU(r)U (r) represents an actual
static charge density, the fraction of the time a point electron occupies a
particular differential volume, or something in between. It is only known that
*
solving Eq. 4.3.1 for U(r) then evaluating eU(r)U (r) gives the correct time-
average spatial charge distribution.
Since Schrodinger first presented the equation, it has been shown that
many different postulate sets yield it as a derived result. Since the
interpretation depends upon the nature of the model used to derive it no
single result is a sufficient basis for deciding if a particular model is correct.
This section derives the Schrodinger equation using a thermodynamic
200 The Electromagnetic Origin of Quantum Theory and Light

approach, and results are interpreted accordingly. A precise description of an

extended, moving, bound charge density trapped by an electrostatic force
and coupled to its own magnetic field is beyond our capability; we simply
don't know enough about electrons. Therefore, in a way similar to
thermodynamics, we seek an energy function from which follows general
ensemble properties without detailed knowledge of the ensemble. The
approach is adequate to obtain time-average values of kinematic properties,
i.e., expectation values.
The approach begins by noting that a dynamic charge distribution
supports time-average values of charge and current densities, respectively
p(r) and pip), within the spatial range r and r+dr and the momentum range p
and p+dp. Momentum densities are directly proportional to current densities.
Let an electron occupy a single eigenstate and let the charge density be
everywhere the same sign. The constraint is expressed by introducing
complex functions U(r) and T(p), defined by the relationships

eU*(r)U(r) = p(r) and eT*(p)r(p)=p(p) (4.3.3)

U(r) and T(p) are complex functions and, by definition, are wave functions.
It follows that

Ju*(/-)U(r)d^= 1 = Jr*(/»)r(/>)d^ (4.3.4)

Differentials dFand dVp represent, respectively, differential volume in space

and momentum coordinates.
Since U(r) and T(p) describe the same dynamic charge distribution, they
are relatable. Each position in coordinate space receives contributions from
the full range of momenta in proportion to the value of F(p) at each velocity,
and vice versa. Therefore we seek a linear transformation between the two
coordinate systems that satisfies the conditions:

U(r)=L{r(/>)} and r(/»)=L- 1 {u(r)} (4.3.5)

L is a linear operator and L its inverse. A general linear function that meets
these requirements is the Fourier integral transform pair:
Quantum Theory 201

3/2
1 ir* p
u(,)= 2nh MP) h
d^ p
(4.3.6)
3 / 2
I
1 r* p
/ . . „ > »

JU(')«PI dF
27lfc V * 7

The constant h is a dimension-determining constant; its magnitude must be

determined by experiment. Dropping to one dimension for simplicity,
Eqs. 4.3.6 take the form:

1/2
I
U(*) = Jr(p)exP[^]d/,
_2nh
(4.3.7)
1/2
1
r(p) = 2TI^ >wexp dr

The expectation value of momentum, p, is given by the equation:

(p) = \pT*(p)r(p)dp (4.3.8)

The same value may be calculated using U(x). To do so, substitute F(p)
from the second of Eqs. 4.3.7 into Eq. 4.3.8. The result is:

r p1'2^
</>> = J^(p)Ju(x)exp[fJ«lr (4.3.9)
2nh

Integrating the second integral by parts gives:

1/2 °°
au(*) ( v „ >

<P> =
2%h
J r*W "K'Z.-*!^"*" dx
\foj

Since an acceptable wave function is equal to zero at infinity, the first term
within the brackets vanishes. Substituting the complex conjugate of the first
202 The Electromagnetic Origin of Quantum Theory and Light

of Eqs. 4.3.7 into the second term and reversing the order of integration
gives:

h dU{x)
</>>=Kw dx
dx (4.3.10)

Equation 4.3.10 is an example of the general case: A dynamic variable in

momentum space may be replaced by an operation in dimensional space, and
vice versa. Letting O indicate that the variable is written in operator form, in
three dimensions the momentum operator is:

h.
o(/»)=-v (4.3.11)

It is understood that the operator acts on wave function U(r). Repeating the
above procedure for p shows, after 'n' partial integrations, that the result
generalizes to:

<Mi) *• (4.3.12)

A significant result is that it is not necessary to solve for both U(r) and F(p)
to solve a kinematic problem. It is only necessary to work with one
functional type, typically U(r), and express conjugate variables in operator
form.
The conservation law of primary importance is the low speed energy of
an electron with total energy W. The sum of kinetic plus potential energies
is:

^=^jdrp[/>2r*(/,)r{j,)]+jdP\A(r)lJ*(r)lj{r)] (4.3.13)

An arbitrary constant, such as the self-energy of the electron, may be added

without affecting results to follow. Applying Eq. 4.3.12 to Eq. 4.3.13 gives
the result:
Quantum Theory 203

jdFV*(r) -—V2\j(r) + [A(r)-fr]\j(r)i = 0 (4.3.14)

Although only the integral is required to equal zero, the more stringent
condition that the integrand equal zero at all points within the region may
also be applied. Doing so returns Eq. 4.3.1, the time-independent
Schrodinger wave equation. Function U(r) is a wave function that provides
the time average charge density of the electron of interest at each point.
The above development shows that the Schrodinger equation is a
statement of energy conservation; Planck's constant appears as a phase-
determining normalization constant in the scalar product between velocity
and position vectors, see Eq. 4.3.6. The Schrodinger equation is correct only
at electron speeds much less than c and it does not account for electron spin;
it is necessary to add electron spin separately to the wave equation. In
contrast, Dirac's equations apply in all inertial systems and spin is an
integral part of the whole. Although Dirac's work is of singular importance
to quantum theory, it does not assist in resolving basic issues of photon
exchanges considered here. Therefore, it is not discussed in this work.

4.4 The Uncertainty Principle

By the uncertainty principle, it is not possible to determine simultaneously
the exact value of conjugate variables, such as position and momentum. The
more accurately the position of a point electron is known the less accurately
the momentum can be known, and vice versa. As a simple example, consider
the case of an electron described by a Gaussian wave function. That is, U(x)
2
is proportional to exp(-x IB), where B is undetermined but constrained to be
positive:

0<B<°o (4.4.1)

The electron is confined to position zero only if B increases without limit

and the smaller the value of B the larger the physical extent of the charge
distribution. The system is normalized if the probability density at each point
is:
204 The Electromagnetic Origin of Quantum Theory and Light

2x2
U (x)U(x) = J—exp (4.4.2)
TtB B
V J

The expectation value of x may be calculated using the integrals of

Table 4.4.1:

f . ->\
2xL B
exp - x 2 dx=
^ — (4.4.3)
7lB B
V J

V7C
jexpl-tf x \dx=—
0

\x expl-jr \dx=
0

j e x p l - ^ x \cos(frx)dc= exo -
Aal

Table 4.4.1. Short Table of Gaussian Integrals

Substituting U(x) into the second of Eqs. 4.3.7 results in the momentum
space form of the wave function:

f B MM
B 2
r(/0 = 271^
ex — P
(4.4.4)
Ah

Using Eq. 4.4.4 to calculate the mean-square value of momentum gives:

r B M/2
(P2) = {exp pdp=- (4.4.5)
2nK -oo V
2h2
Quantum Theory 205

Recalculating <p > in coordinate space using operator notation gives, after
some calculation:

1/2 «•
( X2^ dl
{f) = -v TtB dx1
exp 6&r =
B
(4.4.6)

The r.m.s. values of position and momentum satisfy the parabolic

relationship:

i< r x / >
2 2 *i
= -
(4.4.7)

By Eq. 4.4.7 it is not possible to know position and momentum more

accurately than AxAp = h, where Ax and Ap are, respectively, uncertainty in
the measurement of position and momentum. This is a quantitative statement
of the uncertainty principle. It results from the properties of the Fourier
integral transform relationships relating the wave functions in momentum
and coordinate space. The same is true for all conjugate pairs, i.e., pairs
related by Fourier transforms; they satisfy the parabolic uncertainty
relationship of Eq. 4.4.7.
It may be shown that a Gaussian wave function provides the least
possible uncertainty; all other wave functions provide a greater uncertainty
than that of Eq. 4.4.7.
By the electron model of Section 4.2, the uncertainty is due to
incomplete information available about the intra-electron ensemble. Were
the structure adequately known, in principle the exact solution could be used
to obtain full knowledge of any physical result.

4.5 The Time-Dependent Schrodinger Equation

Solutions of the time-independent Schrodinger equation describe time-
average values of the kinematic parameters and, of course, time-average
values are constant. However, the time-average expectation values occur
over time intervals that are long only when compared with changes in the
electron configuration within the state. Time variation of solutions over
206 The Electromagnetic Origin of Quantum Theory and Light

much longer periods of time, but still short compared with events in the
macroscopic world, are often of interest. In this section, we examine changes
during times that are longer than needed for electron configuration changes
but short compared with macroscopic times. The result determines the initial
variation of expectation values away from equilibrium positions by
calculating changes that occur slowly enough so ensemble averages always
remain in near-equilibrium conditions. If the potential changes too rapidly,
or if the potential change is too large, the near-equilibrium condition is
violated and the Schrodinger equation ceases to apply. In summary, the
Schrodinger time-dependent equation applies only if the ensemble remains
in a near-equilibrium condition.
From Eq. 4.3.2 the time-average charge density at a point due to total
charge e is:

p(r)=e\J*(r)\j(r) (4.5.1)

Since the current density is real, it may be written as:

J( / .) = 1 P ( / . ) / , = ^U*(A)VU(/-)^—fu*(/-)VU(/-)-U(/-)VU*(/-)l (4-5.2)

To describe the time dependence introduce notation similar to that of

Eq. 4.5.1:

p[r,t)=0V*(r,f)y(r,t) (4.5.3)

By extension

he
j(/\/) = \|/*(/-,/)V\|/(/-,/)-\i/(/-,/)V\)/* (/•,/)] (4.5.4)
0 im

The rate of change of the charge density is:

3/ = e V*(r,t)—y(r,t)+y{r,t)—y*{r,t) (4.5.5)
Quantum Theory 207

The divergence of the current density is:

fie
V • J(r, /) = - ? - f \|/* ( A, /)V2\|/(/% / ) - v ( r, / ) v V (/•,/)] (4.5.6)

The continuity equation is:

V.J(,/)+^=O (4.5.7)
0/

Substituting Eqs. 4.5.5 and 4.5.6 into Eq. 4.5.7, multiplying by {tilie), and
adding and subtracting potential term A(r) gives:

- — V 2 y( /•,/)+A(r)y( /•,/) + -3-v(/-,/)

2w / at
(4.5.8)
-V(r,/)
2m 1 at

To connect with the time-independent equation, we seek wave function

\\f(r,t) that, as the time dependence becomes vanishingly slow, goes to:

h2 ->2
U » V U( r) + AU(/-) - MJ( r)
(4.5.9)
-u(,) _ 2/7?
1_V 2
U» + A U » - M J »

Since each line of Eq. 4.5.9 is equal to zero, so are the two lines of Eq. 4.5.8
in the low speed limit. Therefore, the satisfactory, time-dependent wave
function is:

J" v*M -—V

2m
2
Y|/(/-, /) + A(/-)\|f(r, /) + - - - y ( r , t) dV=Q
idt
(4.5.10)
208 The Electromagnetic Origin of Quantum Theory and Light

Insisting that not just the integral but also the integrand be equal to zero
results in the equality:

ti d
—.j^{r,t)=B^{r,t) = m^{r,t) (4.5.11)

Equation 4.5.11 is the Schrodinger time-dependent equation. The

Hamiltonian operator, H, is defined to be the operator that acts on the wave
function to produce the time dependence and state energy of Eq. 4.5.11.
From the first equality of Eq. 4.5.11:

/A\f(r,A = V2y(r,t)+ A(r)y(r,t) (4.5.12)

2m

From the second equality of Eq. 4.5.11:

( iW
\|/(/-,/) = y(/%/0)exp •t (4.5.13)
v « j

The initial value of the wave function is the equilibrium value:

y(/%/ 0 )=U(/-) (4.5.14)

An important result of Eq. 4.5.13 is that the frequency of an eigenstate is

related to the energy as:

C0 = /r/ft (4.5.15)

The time-independent Schrodinger equation has the character of a

thermodynamic equation in that only time-average values taken over times
long compared with the periods of possible intra-state electron movements
are known. Detailed charge and current densities remain unknown. Although
Eq. 4.5.11 provides correct average values, it does not imply a time-line of
actual events. It only applies to initial changes at the onset of instability, not
to the full transition.
Quantum Theory 209

4.6 Quantum Operator Properties

An extension of the logic that supported the use of operators to calculate
momentum generalizes to include functions of momentum. To make the
generalization consider the integral:

1= J [lR(/0/*s(/F)]drp (4.6.1)

TR(P) and Is(p) represent two eigenfunction solutions of the same

differential equation. For each function TR(P), there exists a Fourier integral
transform function in coordinate space, UR(T). TO rewrite the integral of
Eq. 4.6.1 using spatial functions, repeat the procedure used going from
Eq. 4.3.8 to Eq. 4.3.10. Taking the gradient in the direction of the
momentum and working with the "S" functions, the integral of Eq. 4.6.1
becomes:

1= J UR(/0 -vu s (/-) dV (4.6.2)

Similarly, working with the "R" functions gives:

I- J Us(/0 -vu R (r) dv (4.6.3)

Since all physical results are real, Eq. 4.6.3 is equal to its own complex
conjugate:

J Uj(r) :VU R (r) dv= J u s (r) - - V U R ( A ) dV (4.6.4)

Combining Eqs. 4.6.2 and 4.6.4 gives:

210 The Electromagnetic Origin of Quantum Theory and Light

j UR(/0 * VU s (r) dV= j U s (r) * VU R (r) &F (4.6.5)

—oo —oo

The result generalizes to:

oo oo

\ U R (r)0[U s (/-)]d^= | U s (/-)(o[U s (/-)])V (4.6.6)

The symbol "O" indicates any quantum mechanical operator. An operator

that satisfies Eq. 4.6.6 is, by definition, a Hermitian operator.

4.7 Orthogonality
To examine the orthogonality properties of wave function \p(r,t), let O be a
quantum theory operator, let \|/R(r,f) and ys(f,t) be time-dependent
eigenfunctions, and let IR and Is be the corresponding state values. That is:
and
0
V R M = I
RVRM Ovs(/-,/) = Is\|rs (/•,/) (4.7.1)

Functions \\iR(r,t) that satisfy this equation are eigenfunctions and constants
IR are state values. Multiplying the left equation by \j/s (r,t), the right
equation by V|/R (r,t), subtracting one from the other, and integrating over the
volume gives:

J[vRlsVs-Vs I RVR] d ^=J VRO\|/S-¥S(OVR)* d r = ( l s - I R ) J"i|/R\)/Sd^

(4.7.2)

It follows from Eq. 4.6.6 that:

J[\|/R(OI|/S)-\|>S(OI|/R)* d^=0 (4.7.3)

Quantum Theory 211

Combining Eqs. 4.7.2 and 4.7.3 gives:

(ls"lR)JvRVsd^=0 (4-7-4)

If a system has more than one eigenfunction with the same state energy, the
system is degenerate; the number of solutions that produce the same state
energy is the degree of degeneracy. A conclusion from Eq. 4.7.4 is that if the
states are not degenerate the functions are orthogonal; if the state energies
are equal the functions are degenerate and may or may not be orthogonal.
Wherever solutions of a single operator result in many eigenfunctions,
Vs(r>0> the physical result is a sum, weighted by constants as, over all
possible eigenfunctions:

oo

¥(/•,/)= Xa s x|/ S (A,/) (4.7.5)

S=l

ysfcO are normalized values of the wave functions. Requiring that the total
wave function be normalized gives:
oo oo oo

JV¥dr= £ £ aRasJ W s d ^ = £ aRaR = 1 (4.7.6)

R=l S=l R=l

Equation 4.7.6 shows that the sum over the magnitudes of all coefficients is
one. This leads to the conclusion that:

oo oo oo

<0> = J V o \ | r d ^ = X X aRasJVsOi|/Rd^= £ IRaRaR (4.7.7)

R=l S=l R=l

In words, the expectation value of any dynamic function "O" is the sum over
the probabilities that the electron occupies a particular state multiplied by the
state value. For any particular measurement, the use of operator <0>
produces only the particular value aRaR*. With a linear system, a single
electron in a single atom forms partial solutions over each wave function.
That is, an electron is distributed in a statistical way over the possible
eigenstates. With nonlinear systems, and if the nonlinearity is required for
212 The Electromagnetic Origin of Quantum Theory and Light

movements between eigenstates, mixtures of nondegenerate states do not

occur within individual atoms. Either way, if the measurements are repeated
enough times, the different state values appear with a probability equal to the
square of the eigenfunction coefficients.
The initial rate of change of an expectation value follows. Differentiating
the first equality of Eq. 4.7.7 with respect to time gives:

6_
£<°>-J
d/ 3/ 3/
diV (4.7.8)

Using Eq. 4.5.11, this may be written as:

^<0> = - ^ j ( / 0 ( ^ ) - ( / f y ) * ( 0 V ) ) d ^ (4.7.9)

Incorporating Eq. 4.6.6:

—(0) = --JV(CW-j5O)vdr (4.7.10)

For the special case where I = r:

(/,) = m±{r) = ^(J7r-r/f) (4.7.11)

d/ n
The bracket on the right side of Eq. 4.7.11 is defined to be the commutator
of the indicated variable. This particular bracket is the commutator of
position.

4.8 Electron Angular Momentum, Central Force Fields

By definition, the angular momentum, /, in kinematic and operator forms is:

/Srx/,= -rxV (4-8-D

Quantum Theory 213

The first equality follows from classical mechanics and the second using
quantum theory operator notation. By Eq. 4.8.1 the operator form of the
angular momentum components about each of the three axes is:

a 3s ft'
4= = - sin<p— + cot9cosq>—-
az ay ^
3 o\ _ti f (4.8.2)
'y = z——x— cos<p-—cotGsinffl^—
c V y
ox az) i V 39 9<|>

_3__ _3_
'z
3y dx

The first set of equalities in Eq. 4.8.2 follow directly from Eq. 4.8.1 and the
second follows after changing to spherical coordinates.
Another quantity of interest is the magnitude of the angular momentum.
The operator form of the square of the angular momentum follows from
Eq. 4.8.2; evaluation gives:

£ + /y = -k2 2
2ad TT + COtfr
~d
£ = -n2
v ae d$2
39 3d)2 J

The sum is:

1 3 >2^
2
I2 = /2 + /2 + £=-n sin9 39
sine^^ + 2
96 J sin 9 3d)2
(4.8.3)

The electrostatic force fields about atomic nuclei have spherical

symmetry. With spherical symmetry the potential is a function only of the
magnitude of the radius and there is no angular dependence. For this case the
Schrodinger equation has the form:

h2
V 2 U(/-)+ A(r)\j(r) = ff\j(r) (4.8.4)

Introducing the Laplacian operator in spherical coordinates and rearranging

terms gives:
214 The Electromagnetic Origin of Quantum Theory and Light

-Li-
sine 30
sin 6
dV(r,Q,<\>) 3 2 U(A,0,(|))
30 sin 2 0 3<|)2 (4.8.5)

2 dU(^M 2m i
^+A(/-)]^u(/-,e,(|))
dr

To solve the Laplacian, break function U(r,0,<j)) into functions of a single

variable; that is, put U(r,B,§) equal to product function R(r)©(0)O((|)) then
sum over all possible solutions:

u^e^X^M0)0^) (4.8.6)

Substituting the single function product form into Eq. 4.8.5 then multiplying
by the inverse results in the equality:

1 d ( . Qd0^ dzO
0sin0 d9
sin6—I+
1
d 9 j Osin 2 9 Rdr ^V^^y
drj h
(4.8.7)

Since the left side of the equation is only a function of angles and the right
side is only a function of radius, each side is constant. It is most convenient
to put the separation constant equal to -£(£ +1). In a similar way, with
separation constant m, the terms on the left side of Eq. 4.8.7 break into
functions of 6 alone and 0 alone. The result is two complete differential
equations:

d23>
+ mzQ> = 0
d<j)2
(4.8.8)
1 de m
sin6 i(e+\)- G=0
sin6 d0 de + sin 2 0

The 0 solutions may be written either as

<&(<!>) = Awcos(z«|>) + Bwsin(z«<|>) or C ^ + V>„firr^ (4.8.9)

Quantum Theory 215

Since solutions that describe physical reality cannot be multivalued, m must

be an integer. If the solution is proportional to either Am or Bm, by the third
of Eqs. 4.8.2 the z-component of angular momentum is zero; if the solution
is proportional to Cm or Dm the z-component of angular momentum, lz, is:

/z = ±mh (4.8.10)
Combining Eqs. 4.8.3 with the 0-dependent part of Eq. 4.8.8 shows that the
angular momentum satisfies the equation:

/2=£(£ + l)h2 (4.8.11)

The 9-equation provides a physically real solution only if £ is an integer and

solutions with integer values of £ are associated Legendre functions.
The two separation constants, £ and m, are both integers and in the
equations of quantum theory are called quantum numbers. The ranges in
which such solutions exist are:
0<£<°° and -£<m<£ (4.8.12)

Comparison of the above shows that under all circumstances:

/2>/z (4.8.13)

Therefore the entire angular momentum is never about a single axis. This
point supports electron configurational aspects discussed in later sections.

4.9 The Coulomb Potential Source

Let point charge +Ze attract an electron of charge -e. The resulting potential
energy is:

Mr) = —A (4.9.1)

Combining Eq. 4.9.1 and the radial portion of Eq. 4.8.7 results in:
216 The Electromagnetic Origin of Quantum Theory and Light

2 , A ( .xo\ Ze2 l(l+l)h2 n m

R +— ij—K = fFR. (4.9.2)
2m *-2 &r\ dr Anzr 2mr 2

The total energy W is less than or greater than zero respectively for bound or
free electrons. This equation is most easily solved by introducing the
parameter a and variable p where, by definition:

p=OV (4.9.3)

It is helpful to introduce the additional definitions:

cc=—jp-1 and n= l-r—, = T (4.9.4)

h2 4nehp\lf] 2mah
2

If the electron energy is negative, substituting p back into Eq. 4.9.2 and
using Eqs. 4.9.4 gives:

_L_d_[ 2dR ? + ! ) „ nR R
R+ =0 (4.9.5)
p 2 dp[ P dp P,
2
p 4

To solve Eq. 4.9.5 begin with the asymptotic limit at infinity. As p increases
without limit the asymptotic differential equation is:

d2R R
2
=0 (4.9.6)
da 4

The solution of Eq. 4.9.6 is:

R(p) = /(p)e-P / 2 + 6(p)eP /2 (4.9.7)

The conditions on Eq. 4.9.6 are that, in the asymptotic limit of large radius,
F(p) and G(p) must change much more slowly with p than the exponential
terms and, since the total charge is finite, function R(p) must vanish at
infinity. It follows that since a non-zero value of function G(p) does not
Quantum Theory 217

represent physical reality it is multiplied by zero. Substituting the remaining

form F[p)e~p'2 into Eq. 4.9.5 results in the differential equation:

ALF dF n-1 £(£ + l)

F=0 (4.9.8)
2-+ dp~
dp vP j

A convenient solution method for Eq. 4.9.8 is a power series expansion. The
solution procedure begins by forming the summation:

^p)=pssy (4.9.9)
j=o

Solution requires that ao + 0 and aj > 0. Substituting Eq. 4.9.9 into Eq. 4.9.8
results in the sum:

[s(s+l)-^+l)]p"
(4.9.10)
+S{[( s + j + 1 )( s +j + 2 )-^ + 1 )] a j + i-( s +j + 1 - n ) a j}p j _ 1 = 0
j=o

Since ao is not equal to zero, the first term of Eq. 4.9.10 requires that either
s = £ or s = -(£+l). Since the latter is singular at the origin, it cannot
represent physical reality. Substituting s = £ into Eq. 4.9.10 results in the
coefficient ratio:

*j+l j+l + ^-n

(4.9.11)
aj (j+l)(j + 2* + 2)

As index ' j ' increases without limit Eq. 4.9.11 goes asymptotically to the
index of the expansion for e p . Combining this with Eq. 4.9.7 shows that the
radial function is proportional to e p in the limit of very large p, aphysically
unacceptable result. Hence a nontrivial solution of F(p) exists if and only if
218 The Electromagnetic Origin of Quantum Theory and Light

the series terminates, and the series terminates only if n is an integer. For
that case, Eq. 4.9.4 shows that state energy Wn is equal to:

=
^ " „ 2 2 2,2 < 4 - 9 - 12 >

This result shows that an infinite number of energy state values exist, that
the energy is independent of quantum numbers £ and m, and that the state
energy varies as the inverse square of quantum number n.
It is helpful to define radius ro as:

2Z , 4neh2
rQ = — hence rQ = r- (4.9.13)
2
not me

Evaluating Eq. 4.9.13 shows that:

/j> = 5.29172xl0 - 1 1 m (4.9.14)

By definition, ro is the Bohr radius. The electrostatic energy may also be

written as:

Ze2
^n=—0M (4-9.15)
87ie

Combining Eqs. 4.9.12,4.9.13, and 4.9.15 shows that:

<l//-) = z/(n 2 /o) (4.9.16)

In the limit of large values of n, the energy goes to zero and the expectation
value of the radius of the electron state becomes infinite: the electron is
distributed over all space.
To solve for the radial function, rewrite Eq. 4.9.11 as:
Quantum Theory 219

(j + ^ + l - n )
a +1 = (4.9.17)
J (j + 2^ + 2 ) ( j + l ) a J

From Eq. 4.9.17, each coefficient gives

(
J ( n - l - l ) ! ( 2 l + l)! ^
*H-0 ( n - j - ^ - l ) ! j ! ( 2 ^ + j+l)!
ao

To put this notation in agreement with common usage, define ao to be:

(n+l)! 2
a
° ~ ~ ( n - ^ - l ) ! ( 2 ^ + l)!

Combining gives:

a
i=(-1) 17 \"/ \\ (4.9.18)

It follows that:

£<n (4.9.19)

Combining all results shows that R(p) depends upon both quantum numbers
n and £, and is equal to:

V
' -To ( n - j - ^ - l ) ! j ! ( 2 ^ + j+l)!

Functional values for n = 1 through 4 are listed in Table 4.9.1.

Wave function normalization follows from Eq. 4.9.20, and the integral
result:
220 The Electromagnetic Origin of Quantum Theory and Light

£= \ = 2 £=3 = 4

R°(CJ) - e ^ 2 -(2)!(2-p)e* / 2 - s ( 6 - 6 p + p 2 ) e•-P/2 4-6p ,-p/2

-24
2 3
+2p -p

-(3)!pe^ / 2 -(4)!(4-p)pe- p / 2 20-10p

Rn(°) -60 rP/2
z
pe
l+3p y
3 2 2„-p/2
tnV* '
-<5)!p -(6)!(6-p)p

-(7)!p 3„-p/2

Table 4.9.1 Values of R „ ( o ) for n = 1-4

,Hp 2 dpR$(pK(p) = ^ ^ 8 ( n , q ) (4.9.21)

The first row of Table 4.9.1 shows that the largest value of the function
occurs at the origin. Including the origin, there are a total of (. maxima as a
function of radius. The remaining rows show that each function is multiplied
by the radius raised to power I. Therefore the value is equal to zero at the
origin for all except £ = 0 and the radius of the region with a negligibly small
value of charge increases with increasing values of £.

4.10 Hydrogen Atom Eigenfunctions

The full expression for an eigenfunction of a spherical potential source is:

Unto(^a<t))-An^^(an/-)Pf(cose)e-^ (4.10.1)

It was shown in Sections 4.8 and 4.9 that both quantum numbers n and £ are
integers. A separate requirement that £ and m be integers follows from the
requirement that the full range of solid angle be available for angular
Quantum Theory 221

solutions. Also since the full range of angles are available the coefficients of
Legendre functions of the second kind are all equal to zero. From the theory
of Legendre polynomials:

Jsi„eae[pr(coSB)r^^

Using Eqs. 4.9.21 and 4.10.2 to solve for the total probability of each state,
then normalizing that value to unity, permits solving for the constant
coefficients of Eq. 4.10.1. The result is:

a\ (n-^-l)!(2^ + l)(^-^)!l°- 5 (41Q3)

A nim\ 3
4rcn ( n +^)! (t + m)\

Each wave function has (n-£) zeros, including infinity, and undergoes (n-1)
nodes (functional maxima) as a function of radius. Several complete
eigenfunctions are listed in Table 4.10.1.
The energy levels of Eq. 4.9.12 show that the energy depends upon
quantum number n but not upon quantum numbers ( and m. In common with
other boundary value problems only eigenfunction solutions can exist.
Parameter ro of Eq. 4.9.13 is a normalizing radial factor that shows atomic
radii to be on the order of 0.1 nm.
Since wave functions with m = 0 have spherical symmetry the charge
density associated with it produces monopole electrostatic fields. There is a
charge density node at the origin and (n-1) others at increasing values of
radius. Wave functions with m = 1 have bilateral symmetry. There is a null
in the charge density at the origin and (n-1) nodes. Wave functions with
m = 2 have quadrilateral symmetry. There is a charge density null at the
origin and (n-2) nodes, etc.
For n = 1, both I and m are equal to zero and there is but one
eigenfunction; there is no degeneracy. For n = 2 there are two types of
solutions: one is I = 1 with an accompanying triplet of state values of m:
m = - 1 , 0, +1. The other is the singlet 1 = 0 with an accompanying singlet
state value of m: m = 0. Since the energy depends only upon n, and since for
222 The Electromagnetic Origin of Quantum Theory and Light

\3/2
-Zrlrn 1
Uioo = - / - ! -
(I//-)

3/2
1 Z/- -Z/72r n 1 4^0
U
200 - 2 <!//•> Z

f
Zr^ -Zrllr, COS0
u 210
*bJ r
Qj

r •y'S
v
e ° sinGe
u 21+1 —
S-yfn r
o) <0J r

1 fz\^( Zr Z V ,-Zrf3r0
2 2^
9^o
U
300 - " 27-18—+ 2 2
8W3JI *>) 'o r <!//•>
o ;

U
310 -
S fzV"6 Zr Z£ -Z/-/3/-„
COS0
8lVn 'a;

1 2> Z/
" | -ZW3*, • Q ±/
6-- — e ° sinQe
u 3i+r
8lVn 'O,

1 ;_z N 3/Yz2r2^
Z/73r
lr n
/,, 2 ,\
U
320 - " e °° (I; 3 COS Q0 - 11
r
81V67I v o, \ r
0 J
3/2 (^i 2^
1 Z r
-Zr/3rn . n n ±A|>
U 32±1 — ~2 e ° sin0cos0e T
8lVn /
o; \ r
O J
3/
Y 7 2 2>
1 Z r -Zr/3rn0 . 2 Q ±/26
U 32±2 sin 0e Y
162Vn , r oJ V ro

Table 4.10.1 Hydrogen atom eigenfunctions for n = 1 through 3

(l/<r>) depends upon quantum number n only.
Quantum Theory 223

each value of I there are (21 + 1) values of m, the result is a (2£ + 1) fold
energy degeneracy. For each value of n there are n-1 values of L Hence the
total energy degeneracy is:

n-1
£ ( 2 ^ + l) = n 2 (4.10.4)
t=\

2
That is, there are n possible solutions for each value of energy.
The degeneracy is lifted if the electron system is immersed in a static
electric or magnetic field. The m = 0 states are more closely tied to the
nucleus than are the m =±1 states, which extend further outward from the
nucleus. Therefore a static electric field affects the different states differently
and removes the degeneracy. This is the Stark effect. The m = 0 states
support no angular momentum and produce no net magnetic moment. The
m = ±1 states do support angular momentum and do produce a magnetic
moment. Hence, the states respond differently to an applied static magnetic
field. The different response energy removes the energy degeneracy and the
result is the Zeeman effect.

4.11 Perturbation Analysis

Consider an atom to be immersed in an external force field. So long as the
applied field and its gradient are much less than those of the trapping
potential the wave functions retain their original character and solutions
entail a re-scrambling of the occupied states. As an example, consider the
special case of an atom immersed in static, externally applied electric field of
a magnitude small compared with the Coulomb field. Modifications result in
a change of occupational probability of the original eigenfunctions.
To show that this is true, let the Hamiltonian operator HQ characterize the
energy of an isolated electron system. The resulting total eigenfunction is a
sum over wave functions that are solutions of the Schrodinger equation with
operator // 0 . Let one possible eigenfunction be U n o. The possible
eigenfunctions and the corresponding energies Wno are known and each
satisfies the relationship:
224 The Electromagnetic Origin of Quantum Theory and Light

J(u;o^bU„o)d^=^oJu^oU„0d^ (4.11.D

A small external force field is applied that modifies the Hamiltonian to the
operational form:

H=H§ + HX (4.11.2)

Since the external field is controllable by external means, for example the
intensity of an applied laser beam, the actual operational form may be
written as:

J/=J/0 + aJfl (4.11.3)

An experimenter may control the value of oc from zero to one. The

eigenfunctions and energies are functions of a. If the applied force is small
enough a power series in powers of a will converge, with the result that:

Un = Uno + a U n l + a 2 U n 2 + ...= X a r T j n r
r
(4.11.4)

The "0" subscripts form the total solution in the absence of the external field,
the " 1 " subscripts describe the first order correction, the "2" subscripts
describe the second order correction, etc. For small fields, only the
correction terms that are first order in a are large enough to be of interest
2
and terms proportional to a may be ignored. The first order terms are:

|{(u*o + «U*nl)(^o + «^iXu n 0 + «U nl ))d^

= /{(<(, + otU*nl)(^n0 + <x/rnlXun0 + ccUnl)}d^
(4.11.5)
= f {u* n o^oU„o + 4 U n l ^ 0 U „ 0 + U ^ U ^ , + U * n 0 ^ 0 U n l ] ] d F

=jfcou>nO +oKoU> n 0 +^ n l U* n 0 U n 0 + ^ n 0 u ; 0 U n l ] } d ^
Quantum Theory 225

Confining attention to the last equality of Eq. 4.11.5, the first terms are equal
and may be subtracted out. The procedure may be repeated for the first terms
within the square brackets. Applying the Hermitian property of quantum
operators to the last terms within the square brackets shows that they too are
equal and they, too, may be subtracted out of the equation. Eliminating these
three terms leaves the center terms within the square brackets:

Ju;o^iUnod^=/rnlJu*noUnod^=^nl (4.11.6)

This is the first order correction term. Equation 4.11.6 shows that it is not
necessary to know the corrected wave function to calculate first order energy
changes. It is only necessary to know how the first order Hamiltonian
correction affects unperturbed eigenfunctions.

4.12 Non-Ionizing Transitions

Let an electron in an unperturbed atom be described by the Hamiltonian
operator HQ, eigenfunctions V|/n, and energies Wn. Schrodinger's equation is:

(4 12J)
^0VnM = ^ V n M -

The total wave function is a weighted sum over all possible wave functions:

¥(/;,/)= 5 > n y n ( / i , / ) (4.12.2)

n

Next, let a second electron be attached to the same atom, affected by the
same Hamiltonian operator, and have the same set of eigenfunctions and
energies but a different set of coefficients:

¥(/- 2 ,/)=$> n Y|/ n (/- 2 ,/) (4.12.3)

n

Make the definitions:

226 The Electromagnetic Origin of Quantum Theory and Light

Vmn('',')=Vm('i> / )Vn('2,')
U m n (/') = u m (/;)u n (A 2 )
T
"mn "m "n

The equation that describes both electrons is:

/'^.n/
c
^('''^X X mnVmnM = X S^nUmn(,i'/'2)ex; (4.12.4)

If a perturbing field is applied that changes the Hamiltonian operator from

Ho to HQ + H\, the wave functions remain unaltered and the probability
coefficients, c m n , become time dependent. To show this, write Eq. 4.12.4 in
the form:

(^o+^i)E £ cmn(/)Vmn(^/) = - | - X £ cmn(/)Vmn(/-,/) (4.12.5)

m n m n

Writing out the equation term by term gives:

mn + c m n mn I

(4.12.6)
= EE
The first terms on either side are equal; subtracting them leaves the equality:

1 1 C m n M ^ r n n M ^ E 1 ^ f ^ V m n M (4-12.7)
m n

Multiplying through by t|/pq (/*,/) and integrating over all space gives:
Quantum Theory 227

^cpqW =
| Z £ C
mnU)j Vpq^lVmnd^ (4.12.8)
m n

The integral is over the volume occupied by both electrons. Make the
definition:

<pq|Ar1|mn> = J u * p q ^ 1 U m n d ^ (4.12.9)

The terms of Eq. 4.12.9 are, by definition, the matrix elements of interaction
potential H\. With the aid of Eq. 4.5.13, Eq. 4.12.8 may be written in the
form:

dc
pq(') i(wm + Wn-Wv-Wq)t
<pq|#l|mn)exp (4.12.10)
At h n

As a special case suppose that at time / = 0 the electrons are in states m and n
and find the probability, as a function of time, that they will occupy states p
and q. The initial condition is that

'i(w?+wq-wm-wn)t
^ H W = |(pq|// 1 |mn)exp (4.12.11)

Coefficient c m n is equal to one at time t = 0, when all other coefficients are

equal to zero. Make the definition that:

Afr=frp+prq-?rm-frn

Restricting analyses to times short enough so that c m n remains nearly equal

to one, the integral over time shows the initial time dependence of the
coefficient to be:
228 The Electromagnetic Origin of Quantum Theory and Light

n
c p q (/) = (pq|^i|mn) ^ ' (4.12.12)

The probability of state (p,q) being occupied is:

* , , , ,2 sin2(AW//2h)
CpqCpq = 4 { p q ^ mn) 2 f -^—^ (4.12.13)
(ArVf

Equation 4.12.13 shows that the probability that a particular transition will
occur is proportional to the square of the matrix element. The magnitude of
the matrix element depends upon both sets of quantum numbers, pq and mn.
Transitions are "forbidden" if the matrix element is equal to zero. Since the
2 2
ratio sin (x)/x has maximum magnitude at x = 0, it follows that energy
conservation requires that the most probable value of A W be zero.

4.13 Absorption and Emission of Radiation

The purpose of this section is to describe the absorption and emission of
radiation by an atom with a full compliment of electrons. The atom is
immersed within an externally applied plane wave of radian frequency co
where the wavelength is much greater than the initial size of an atom.
The calculation begins with the relationships between the Hamiltonian,
H, the state energy written in operator form, and the time dependence of the
applied field as described by the time-dependent Schrodinger equation. The
relationships are, see Eq. 4.5.11:
.f. *^

- V , V n M = ^ n M = ^ n ¥ n M (4-13.1)
/ at

It follows from Eqs. 4.5.13 and 4.5.14 that the time-dependent

eigenfunctions are related to the time-independent ones as:

^n(r,t) = \Jn{ry^'n (4.13.2)

Quantum Theory 229

The complete wave function *¥(r,i) is a weighted sum over all possible
eigenfunctions:

X
FM=5>n(')U n ('-)e A V (4.13.3)

As in Section 4.12, it is possible that the probability coefficients are time

dependent. Substituting the first order perturbation equation, Eq. 4.11.6, into
Eq. 4.13.1 and Eq. 4.13.3 gives the differential equation that describes the
rate of change of the coefficients as a function of the coefficients
themselves:

-£^M')^*'^^M'V**' (4.13.4)
n n

Symbol H\ represents the operator form of the modification to the

Hamiltonian due to the perturbation. This equation shows that the primary
affect of the applied field is to make the state coefficients time dependent.
The electric field intensity in the perturbing plane wave is E(r,t) and it
varies with time as cos(coO- Although E(r,f) is a real function, it is
convenient to rewrite it in complex terms as the sum of complex conjugate
functions:

E(/-,/) = E 0 e / < c o / -'"') + c.c. (4.13.5)

For atoms of diameter much less than a wavelength, the perturbing energy is
approximately equal to:

Jfl=-eE( /)•/•(/) (4.13.6)

Combining Eq. 4.13.6 with 4.13.4 gives:

dt n
n n^ J
(4.13.7)
230 The Electromagnetic Origin of Quantum Theory and Light

To determine the affect of interaction between states n and k, multiply

Eq. 4.13.7 through by Uk(r) and integrate over all space. The result is:

(4.13.8)

Symbol ( U | c | / ' | U n ) is defined in Eq. 4.12.9. To simplify the problem

consider as a boundary condition that at initial time t = 0 only the single state
'n' is occupied by an electron. Therefore only c n is different from zero and it
is equal to one. Doing the time integral of Eq. 4.13.8 under these
circumstances gives the initial solution:

s /(o) n -co k -(fl)/_ j e/(cOn-Q)k+0))/_1

te Jk*r -ilc»r
ckW = - - +e E 0 «<U k k|U n >
((0n-C0k-0)) (con-cok+co)

(4.13.9)

-ik»r
The exponential phase factor may be expanded as [e~ ~\-ik*r+ ...].
The first order value, one, is sufficient since atomic sizes are much less than
a wavelength. Making the replacement shows that the square of the
magnitude of Eq. 4.13.9 may be written as:

sin n k ;
2 2 : 2/T
|c k (;)| = (2tfE 0 ) <U k MU n > (4.13.10)
(fra-fFL±to>y

The term within the curly brackets of Eq. 4.13.10 is significantly different
from zero only if the argument of the sine is equal to zero. This requires that:

u)k = con ±u) (4.13.11)

Quantum Theory 231

Multiplying Eq. 4.13.11 through by ti shows that (Wk-Wn)=± h(Q. If

(Wk> Wn) energy ftco is added to the system and the transition is associated
with energy absorption. If (W^< Wn) energy ha> is removed from the system
and the transition is associated with energy emission. Therefore, it seems
reasonable to ascribe the upper or lower sign of Eq. 4.13.11 respectively to
energy absorption or emission by the electron. Equation 4.13.11 also shows
that the transition probabilities for emission and absorption are the same. .
For simplicity, the development is carried out with the restriction that
Eq. 4.13.10 is valid only over times so small that cn(7) remains nearly equal
to one and all other values c^(t) remain much less than one.
If the radius of the atom is equal to the first Bohr orbit, see Eq. 4.9.14,
and the wavelength is at the center of the optical band, 500 nm:

2/fr- = 6.65xl0" 4 (n 2 /z) (4.13.12)

If n is small enough so the magnitude of Eq. 4.13.12 is much less than one,
the perturbation expansion of Eqs. 4.11.4 converges rapidly and calculated
results may be limited to the first correction term only.

4.14 Electric Dipole Selection Rules for One Electron Atoms

To obtain the probability of an energy exchange by absorption or emission
of radiation it is necessary to evaluate the matrix element. For an atom with
only one electron, the wave functions are given by Eq. 4.10.1 and in
Table 4.10.1. The components of the matrix elements directed along the
three rectangular coordinate axes are:

oo 71 2ft

X = J rdn sin 0d0 J U ^ ^ U ^ / ' s i n 9cos (f>d<))

0
° ° (4.14.1)
°° ft 2?t
Y = j rdrj sin 0d0 J U„<£>m<XJ„£mrsm 0sin 0d<])
0 0 0
232 The Electromagnetic Origin of Quantum Theory and Light

Z = J A 2 dr\ sin0d0 j U„<t<m'1J„tmrcos0d(|)

0 0 0

The total wave function is related to the individual functions as:

U/zte(/•,e,(t)) = A„ £w Rya /^ /•)Pf(cose)e- />w<t, (4.14.2)

Combining the first of Eqs. 4.14.1 with 4.14.2 gives:

X = A ^ A ^ J R / > ^ M P / P f sin2ed6 J e ' < * - ^ cos <j>d<>| (4.14.3)

0 0 0

Evaluation of the azimuth angle integral gives: g

27t
J d§ cos §[cos(m'- m)<\> + isin(/rf-/»)§] = nSim^mt I) (4.14.4)
0

Incorporating Eq. 4.14.4 then using the identity of Table A.21.1.7, gives:

sin2 0 P r lTf = ^ s i n e ^ T i 1
- ^ ) ^ 1

(4.14.5)
2 1
sin e P r l p - = ^_ s in0(pf + 1 - Pf_,)pf
2f+l

Combining Eqs. 4.14.3 and 4.14.5 shows that for both values of m the zenith
angle integral of Eq. 4.14.3 has the form:

n
1
^{sine^Ti'-P^jpr'de
21.. Q

Consider the integral:

Quantum Theory 233

71 271
J p ^ P f sin 6d0| cos^'-OT)*d0 =
2

0 0

;
(2*+l) 1(2^+3) {l-m)\ ^ (2^-l)(^-«-2)P '
(4.14.6)

The integral is different from zero only if:

e=t±l (4.14.7)

Combining the second of Eqs. 4.14.1 with Eq. 4.14.2 gives similar results.
Combining the last of Eqs. 4.14.1 with Eq. 4.14.2 gives:

°° 7t 271
In .

'J
Z = A ^ / W A ^ W J R J R J ^ / / - J P / P / ' s i n 9 c o s 8 d e J e ' ^ ^ d f ) ) (4.14.8)
0 0 0

Table A.21.1.4, shows that:

1
sin ecosOPfPf = — — sin0[(^-m+litfli -(i + m)VP_^ (4.14.9)

Combining with the zenith angle integral of Eq. 4.14.7 leaves the form:

nn
1
— Jsine[(^- zw+l)P£i - ( * + zw)P£i]p*d9 (4.14.10)
2£~
0

Consider the integrals:

234 The Electromagnetic Origin of Quantum Theory and Light

n 2n
1
P ^ J s i n e ^ - * + l ) P i T , -(£ + ^ P f ^ J p f d e j d O e ' ^ ^ =

, ; \'\{t-m+\)) 'T8(?,£ + l)-U + m)) (-6(f,£-l)

(4.14.11)

Like Eqs. 4.14.6, Eq. 4.14.11 is different from zero only if Eq. 4.14.7 is
satisfied. Since the radial integer provides no restrictions on n, electric dipole
transitions occur only if:

A^ = ±l and Am = ± l o r 0 (4.14.12)

The interpretation of Eqs. 4.8.10 and 4.14.12 is that the z-component of

radiated angular momentum is equal to zero or to h. Since the angular
momentum of the source changes by that amount, it must be carried by the
radiation.

4.15 Electron Spin

It was known before Schrodinger's equation was known that a complete
description of electronic events requires four quantum numbers. An integral
part of Dirac's equations is that electron characteristics include more than
charge and mass: there is a permanent angular momentum and a permanent
magnetic dipole moment.
The angular momentum of a particle in terms of its mass and velocity is:

l = m rxv (4.15.1)

By definition / is the angular momentum, r the radius about a fixed point,

and v the velocity of the point mass. If the mass also supports charge q, the
magnetic moment is:

Q = qrx v/2 (4.15.2)

Quantum Theory 235

Comparison of Eqs. 4.15.1 and 4.15.2 shows that

Q=q//2m (4.15.3)

Thus the expected relationship between an electron's magnetic and

mechanical moments, Eq. 4.15.3, is equal to:

Q = -e//2m (4.15.4)

However, the proportionality between spin magnetic moment and angular

momentum is twice as large as the ratio of the intrinsic magnetic moment
and angular momentum.
It is found that the intrinsic angular momentum obeys rules similar to
those of orbital motion about a central force field, Eq. 4.8.10 and 4.8.11,
except that with spin the only allowed quantum numbers are plus and minus
one half. That is, the total angular momentum is:

s»s = s(s+l)h2 (4.15.5)

The component along a particular axis is:

sz-ms% (4.15.6)

Quantum numbers w s are equal to either ±1/2. Combining terms shows that
the magnetic moment can have either of the two values:

Qz = ±eh/2m (4.15.7)

The absolute value of the moment is called the Bohr magneton, and of value:

Q z = 9.274xl(T 2 4 J/T (4.15.8)

These relationships may be put in appropriate quantum theory terms by

considering S to be an eigenfunction, sz an operator, and writing:
236 The Electromagnetic Origin of Quantum Theory and Light

j^S = ±fcS/2 (4.15.9)

The total spin wave function combines the functions with coefficients as:

S(JZ) = C+S+(JZ) + C_S_(JJ) (4.15.10)

The wave functions are orthogonal and normalized.

4.16 Many-Electron Problems

To examine a multi-electron atom note that the Hamiltonian operator of a
system of n electrons may depend in a complicated way on the internal
structure of each electron. Regardless of what the complications may be, a
property of critical importance is that electrons are physically
indistinguishable: all results are invariant upon interchange of electrons.
That is, all energies, including both electron-nucleus and electron-electron
interactions, are symmetrical with respect to an interchange of electrons.
Since the energy of an electron is proportional to the square of its wave
function, symmetric energies occur with both symmetric and antisymmetric
wave functions. To examine the symmetry of a wave function break it into
the sum of symmetric and antisymmetric parts, respectively:

M?^{r)=^(-r) mdyAs(r) = -yAs(-r) (4.16.1)

Any physically real function of time can be expressed as the sum of

symmetric and antisymmetric functions of time. For the case of wave
functions:

\|/(/-)=\|/ Sy (/-)+\|/ As (/-) (4.16.2)

The energy density is proportional to the probability density and it is equal

to:
Quantum Theory 237

y(r)*y(r) = ysy (/•)ySy('')+VAs {r)wM{f)

(4.16.3)
s As As Sy
V * (r)y {r) +y {r)y {r)

The first term of Eq. 4.16.2 is invariant with respect to the interchange of
electrons and the second term is not. It follows that the second term of
Eq. 4.16.3 is equal to zero. Therefore the wave function may be either
symmetric or antisymmetric but it cannot be a mixture.
The simplest possible multi-electron system has two electrons, say
electron "a" and electron "b". Let ?ab be a permutation operator that
interchanges the electrons. It follows that:

PabV(*.£) = V(4*) and PabPab\|/(*,£) = Mf{a,b) (4.16.4)

In turn, it follows from Eq. 4.16.4 that:

PabPab = landP a b = ±l (4.16.5)

It follows that

?abUAs(r) = -VAs(r) PabUSy(r) = USy(r) (4.16.6)

Since taking operations with respect to time in quantum theory does not
affect positional symmetry, time does not affect symmetry: A state that is
initially symmetric or antisymmetric before a quantum mechanical operation
has the same symmetry after the operation. Whatever symmetry the wave
function has at time / = 0, it keeps that symmetry for all time. This argument
generalizes to include an arbitrary number of electrons. The conclusion is
that either there is but one type of symmetry in nature, with all wave
functions of the other symmetry everywhere equal to zero, or there are two
separate types of physical reality. If two types of reality exist, one type of
reality would be constructed of electrons with symmetric wave functions and
the other would be constructed of electrons with antisymmetric wave
functions. Consider a two-electron atom for which the total and individual
wave functions satisfy the relationships:
238 The Electromagnetic Origin of Quantum Theory and Light

¥(/•>")= Va('')Vb('")

(4.16.7)

Introduce the notation that:

^{r'S'Hv^M*-")* VbCW)] 16
,
^ a b ( ' ' ' ^ " ) = [Va('' )v|/ b (/-")-Vb(/'')v l / a (/-")]

Since both electrons occupy all points, examine conditions for /•' = /•":

r ab (/,/)=2 Va (/) Vb (/')

. . (4.16.9)
^(/r)=o

The electrostatic interaction energy between the two electrons is:

e
JK\
ab Vab('V"W('''>''")d^ (4.16.10)
4TO' \r>-r'\

If the electron charge density is a continuous function of position,

Eq. 4.16.10 gives a physically acceptable result with either electron
symmetry. As expected from Eq. 4.16.9, the energy with symmetric
functions is larger than that with antisymmetric functions. However, the
definite integral of Eq. 4.16.10 correctly represents the system energy if and
only if the wave functions are continuous functions of position. If there is a
dimensional scale below which the charge density is granular, on that scale
of dimensions it is necessary to replace the integration by a sum over
interaction energies. With symmetric wave functions, by Eq. 4.16.9 the
granular charges are adjacent or overlapping and the sum is singular; there is
no parallel with antisymmetric wave functions since the overlapping
densities vanish. It follows that if there is a dimensional scale on which the
charge density is granular, only antisymmetric wave functions exist. The
Exclusion Principle, first formulated by Pauli, states that only antisymmetric
Quantum Theory 239

wave functions exist. On the basis of the above argument, it also suggests
that, on an appropriate dimensional scale, electron charge distributions are
granular.

4.17 Electron Photo Effects

Hertz discovered in 1887 that a spark jumps a small gap between conductors
more easily when the conductors are illuminated than when in the dark. He
found that the effect becomes more pronounced as the light spectrum goes
from blue to ultraviolet, is most pronounced with clean and smooth
terminals, and cathodes are more active than anodes. The result is a
photocurrent due to the forcible ejection of electrons from the cathode.
Experimentally determined characteristics of photocurrents are:

a) Define the stopping potential Vo to be the voltage difference

between the two plates that just causes the current to cease. The
electron stream continues so long as the electrons have sufficient
energy to make the transit. It must be, therefore, that the actual
voltage V satisfies the condition

V>V 0 (4.17.1)

b) Each type of metal has a characteristic frequency, con. A

photocurrent exists only with light of that or a higher frequency 0).
That is:

co > coo (4.17.2)

c) Photocurrent magnitude is proportional to the intensity of the light:

I=l(E§) (4.17.3)
The symbol Eo indicates the electric field intensity within the light.

d) Photocurrent magnitude is independent of frequency for frequencies

greater than the characteristic frequency:
240 The Electromagnetic Origin of Quantum Theory and Light

e) Photocurrent onset occurs without a measurable time delay after

onset of illumination.

f) Expressed in terms of the above symbols, the maximum kinetic

energy per electron is:

rV=h((Q-(ti0) (4.17.4)

Einstein analyzed the photocurrent problem by treating light as if it

consisted of particles. His argument was sufficiently convincing to persuade
most physicists of the correctness of his model, and, for the analysis, he was
awarded the 1921 Nobel Prize in Physics.
However, it is not necessary to imagine that light consists of particles to
explain the effect. It can be explained in terms of the interaction between
electrons in quantized energy states and an engulfing plane wave. Let the
source metal be sized much larger than atomic dimensions. Inside the metal
electrons are trapped in quantized energy states. Since the skin depth of a
good conductor in the mid-optical range is on the order of 10 nm, the light
penetrates the metal deeply enough to interact with the conducting band
electrons. Let an electron in eigenstate n with energy Wn interact with an
applied plane wave of frequency co. Define the work function of the metal,
WQ, to be the additional energy an electron must have to exit the metal.
Then:

W§ = tm (4.17.5)

It follows from Section 1.13 that there are electromagnetic cavity

solutions of energy Ey within the containing box. As the size of the box, L,
becomes large, the possible energy levels form a quasi-continuum. The
transition equations of Eq. 4.13.10 are repeated here:

sin 2fl\ n k )
|ck(/)|2 = (2*E0)2<UkHun>2 (4.17.6)
(jrn-K±twf
Quantum Theory 241

Let O be the energy of the first state of the quasi-continuous spectrum.

Photoelectron emission occurs because of a transition from state n to state k.
Let the kinetic energy of the ejected electron be Tk. The relationship between
the energies follows and is equal to:

^ k - ^ n = $ + Tk (4.17.7)

Rearranging and rewriting in terms of frequencies:

Tk = ft((0-a>o) (4.17.8)

The voltage required to stop all emission follows from the above and is
given by:

h
V0 = -(u)-CD 0 ) (4.17.9)

Combining Eq. 4.17.6 with Eqs. 4.17.7 and 4.17.8 gives an expression
for the coefficient magnitude as a function of the electron kinetic energy and
the applied frequency:

•2
sin (T k -ftw)/
2 2 2 2h
|ck(/)| = (2*E 0 ) (U k Hu n > (4.17.10)
(Tk-M

To determine the rate of electron ejection it is necessary to integrate

Eq. 4.17.10 over the full range of kinetic energies which, in turn, requires
summing over the quasi-continuum states. Since it is a quasi-continuum,
replace the summation with the integral shown:

sin (T-fcco)/
2 2 2
P(^) = i:|c k (/)| ^(2.E 0 ) <U k HU n > { dT (4.17.11)
(7-haf
242 The Electromagnetic Origin of Quantum Theory and Light

P(/) is the probability that emission has occurred. The integral may be
rewritten as a Dirac delta function using the relationship:

sin ~(T-M' ( /T/r\

^d = 8(x-x 0 )dx (4.17.12)
( /^ 2
ylhj
(T-ftco)
2/z
\<- n
J

Combining Eq. 4.17.11 with Eq. 4.17.12 and integrating over all possible
kinetic energies gives:

p(,)=^M<uk|Hun>2/ (4.17.13)

This probability of emission, Eq. 4.17.13, is directly proportional to time

and, therefore, with a constant light intensity electrons are ejected at a
constant rate. There is no time delay between onset of the light and the onset
of electron emission and the rate of electron ejection is proportional to the
intensity of the illuminating field atomic states. The analysis of Sections 4.12
and 4.13 contained the approximation that cn(t) remained constant, a result
that served as a preliminary to generalized cases with cn(t) a variable. In this
case, since the active electrons lie initially near the Fermi level and the
conduction band is part of a quasi-continuum, the constancy is a reality.
This analysis shows that the photoelectric effect may result from an
interaction between a classical radiation field and a quantized electron. It
appears to be a historic accident that radiation fields as particles acquired its
strongest early support from Einstein's considerations of the photoelectric
effect.
Quantum Theory 243

References
D. Bohm, Quantum Theory, Prentice Hall, Princeton NJ (1951)
K. Kim, E. Wolf, "Non-Radiating Monochromatic Sources and their Fields," Optics
Communication, vol. 59, pp. 1-6, (1986)
P.A.M. Dirac, The Principles of Quantum Mechanics, 4th ed., Oxford (1958)
G. Greenstein, A.G. Zajonc, The Quantum Challenge: Modern Research on the
Foundations of Quantum Mechanics, Jones and Bartlett Publishers (1997)
W.E. Lamb, M.O. Scully, "The Photoelectric Effect Without Photons," in
Polarization, Matter and Radiation, Presses Universitaires de France (1969) pp.
363-369
L.D. Landau, E.M. Lifshitz, Quantum Mechanics: Non-Relativistic Theory,
Addison-Wesley, Reading MA (1958)
A. Messiah, Quantum Mechanics, vol. 1, trans, by G.M. Temmer, John Wiley, New
York (1966)
R. Omnes, Understanding Quantum Mechanics, Princeton University Press (1999)
L.I. Schiff, Quantum Mechanics, McGraw-Hill Book Co., New York (1949)
F. Schwabl, Quantum Mechanics, 2nd ed., Springer, New York (1995)
R.C. Tolman, The Principles of Statistical Mechanics (1938), reprinted by Dover
Publications (1979)
This page is intentionally left blank
 5. Photons

This chapter discusses an electron transitioning between eigenstates. It is

based upon the atomic model of Chapter 4 in which the Schrodinger
equation applies during periods of electron equilibrium and near-
equilibrium, but not during transitions. During transitions a unique resonant
electromagnetic field set exists that consists of properly phased TE and TM
modes of equal magnitudes. The mixed TE and TM radiation also produces a
regenerative radiation reaction force that drives all higher order modes of the
same degree. This nonlinear and regenerative force drives the radiation until
all available energy is radiated or received.
We first determine that the radiated power-frequency relationships obey
the Manley Rowe equations. Next we examine the radiation kinematics and
find that they are the same as photon kinematics. We use the kinematics as a
boundary condition on a multimodal field expansion to determine the full set
of fields present during steady state photon radiation. Finally, we show that
the radiation reaction pressure of the near fields is many orders of magnitude
larger than the Coulomb trapping pressure and that it is directed and phased
to drive the source regeneratively.

5.1 Power-Frequency Relationships

To address classical power-frequency relationships in a closed system,
consider an equilibrated, charged system that supports an internal
electromagnetic oscillation at radian frequency cos. The system is immersed
in and perturbed by a plane wave of frequency oo. If the system is linear, in
the sense that doubling an applied force doubles the system response, only
the two frequencies oo and cos will exist inside the system. If the system
response is not linear, there are additional responses at difference and sum
frequencies. The system response may be written as:

245
246 The Electromagnetic Origin of Quantum Theory and Light

S ( / ) = [ACOS(CO S /)+BCOS(CO/)] P (5.1.1)

A, B, and p are constant, system-specificparameters. The system response

may be expanded as a polynomial of trigonometric functions. An especially
important example of a nonlinear response is the case of/? = 2, for which:

A 4- R
S(/) = +AB[COS[(COS - co)f] + cos[(o)s + to)?]}
2
(5.1.2)
+ - | A 2 cos(2cos/) + B 2 COS(2Q)/)|

The constant term is unimportant for present purposes. The generated

frequencies, [(cos±co), 2cos, 2co], remain within the system and, being part of
it, also drive it and thereby produce additional frequencies. The ultimate
series of generated frequencies continues and includes all frequencies of the
form (mco+nu)s), where m and n are integers. The result is true for all values
ofp greater than one.
Energy is conserved in lossless systems independently of the degree of
nonlinearity. Let Pm>n represent the time average power out of a system at
frequency (mco+ncos). For lossless systems, energy conservation requires
that:
oo oo
P
£ X m,n=0 (5.1.3)
m=-°° n=-°°

It is helpful for what lies ahead to rewrite Eq. 5.1.3 as:

°° °° mP °° °° nP
CO 2 £ - ^ ^ - + C0S £ Z - ^ n _ = 0 (5.1.4)
mco + na). *~* mco + nco,.
s s
m=-°° n=-°° m=-°o n=-°°
Equation 5.1.4 contains redundant information since with integer pair
(mo,no) the sums are identical to the sums obtained using (-mo, -no). The
redundancy is removed yet all information retained by writing the sums as:
Photons 247

OO OO y-» OO OO

mco + ncOo ^ ~ mco + no)s

s
m=0 n=-°° m=-°° n=0
For example, an ideal, nonlinear capacitor is an example of a lossless,
reactive system; other reactive systems may be analyzed in a parallel way.
The charge on the capacitor may be expressed as:

OO OO

=
*M 2 £ £ Qm,nexp[/(mco + ncos)/] (5.1.6)
m=-°° n=-°°

The value of Q m , n depends upon such parameters as the capacitor size, shape,
permittivity, and the supported voltage but does not depend upon frequency
of operation. Since q{t) is real, the condition Q_m,_n* = Qm,n follows.
Similarly the voltage, v(f), across the capacitor is:

OO OO

^ ) =- X X Vm>nexp[/(ma> + na>s)/] (5.1.7)

m=-°° n=-°°

Like Qm,n> Vm>n depends upon the capacitor parameters of size, shape,
permittivity, and the contained charge but not frequency. The capacitive
current i(f) is equal to the rate of change of charge, and may be written

OO OO

i /
( )=2 X X Vn e x P[/( m w + nws)']
m——°° n=—°°
OO OO (5.1.8)
=2 2 5) (m© + n© s )Q m>n exp[/(m© + na)s)/]
m=-<» n=-°°

The second equality follows by differentiation of Eq. 5.1.6 with respect to

time and shows that, differently from either the charge or voltage, current
Im,n does depend upon the frequency. The time average power into the
capacitor is:

" V n = - r R e ( v m n I m > n *) = -(mfi> + no>s)Re(/ V m n Q m n *) (5.1.9)

248 The Electromagnetic Origin of Quantum Theory and Light

It follows that

^s?Ssj-H'v«Q^*) (5iio)

Since the right side of Eq. 5.1.10 depends upon the product of Q m n and
Vm,n which are not frequency dependent, the right side is independent of
frequency. It follows that the left side is also independent of frequency. The
parallel argument follows if the roles of m and n are reversed.
Hence, Eq. 5.1.5 has the general algebraic form:

c1co + c2cos = 0 (5.1.11)

The coefficients cj and c2 are independent of frequency. Therefore the two

frequencies are independent variables and both ci and c 2 are equal to zero.
Applying this result to Eq. 5.1.5 shows that the sums are separately equal to
zero:

ml
y y m,n _ 0
„ ^ mco + ncOm
m = 0 n=
-°° (5.1.12)
oo oo
nFmp
y y »" =0
^ „ mco + ncos s
m=-oo n =0

To illustrate the use of this equation, consider a system in which one of

two possible atomic states is occupied by an electron. The state frequencies
are (Dmjtiai a n d Wfmal» for initially occupied and initially empty states. The
ensemble is then enmeshed in a plane wave of frequency u) where the
frequency of the applied wave satisfies the relationship

W =
hnita|-Mfinal| (5-U3)

With linear systems only the driven frequency, a), and the system
frequency, C0initiai, are present. With nonlinear systems, all frequencies
Photons 249

(mcOinitiai+nco) are driven and are potentially present. If the system is

somehow restricted to support only to initial and G0fmai, only the three
frequencies to, 0)initial> and 00finai are present. Consider the case of a
an<
nonlinear system that supports only frequencies co,nitial and Ufmal l is
driven at frequency GO.
For that case, if coinitial > to and Winitial> tOfinai, Eq. 5.1.12 is satisfied for
integer pairs (m = 1, n = 0) and (m = 1, n = - 1 ) ; if co initial > to and
tOinitial < tOfmal» Eq. 5.1.12 is satisfied for integer pairs (m = 0, ±n = 1) and
(m = - l , ±n = 1). Changing the power subscripts to match the frequency
ones, for these special cases Eqs. 5.1.12 go to:

JktiaL + ijfinaL = 0 and ijaitkL + J U o (5.1.14)

w w w
initial ^final initial

In the second equation, the sign is respectively positive or negative if

tOinitial is greater or less than C0finai. The energy flows are illustrated in
Fig. 5.1.1.
By definition the energy that goes into the final state is:

^ n a l = J"Pfinal<# (5-1-15)

Combining Eqs. 5.1.14 and 5.1.15 shows that:

"initial ^final W
= = (5.1.16)
w CO
initial t° final

Wfinal is the energy that goes into the final state and W is the energy
exchanged between the remote field and the electron as it undergoes a
change of state. The energy-frequency ratio of Eq. 5.1.16 is independent of
system parameters, therefore of system details and, consequently, the ratio is
constant. For eigenstates that constant is Planck's constant, %. It follows that

^final = -fctOfimu i n i t i a l = ^tO i n i t i a l fF=±h(H (5.1.17)

250 The Electromagnetic Origin of Quantum Theory and Light

The upper sign applies if CO;nitial > Wfinal> anc* v z c e versa.

ATOM

Nonlinear
Initial state interaction Final state

'• A•
Emission V Absorption

REMOTE
FIELD

Figure 5.1.1 Diagram illustrating power flows in a nonlinear source.

Initial and remote field energies interact nonlinearly resulting in an energy
flow from the initial state to the final state and either into or from the remote
field.

The results are that if the energy in the initial eigenstate is quantized into
energyficothe energy exchanged among the radiation field and the initial and
final energy eigenstates all have the same energy-to-frequency ratio.
Although Eqs. 5.1.1 through 5.1.14 apply to both linear and nonlinear
systems, energy flow at the sum or difference frequencies is equal to zero for
linear ones. Therefore, Eqs. 5.1.15 through 5.1.17 have meaning only for
nonlinear systems. We conclude that the power-frequency relationships
accompanying electron transitions are unique to nonlinear transitions, not to
quantum effects.
Photons 251

5.2 Length of the Wave Train and Radiation Q

An important property of photons is the length of the coherent wave train.
The purpose of this section is to estimate that length for optical frequency
photons. We begin by expressing the time varying electromagnetic power in
a plane wave as a function of the magnitude and the frequency dependence
of fields. For this purpose, let the electric and magnetic field intensities of a
plane wave be expressed as integrals over all possible frequencies:

E(/)= J E(co)e/to/'du) and H(/)= j" H(co)e/to/dco (5.2.1)

—oo —oo

The rate at which energy passes through a unit area of surface follows from
the Poynting theorem. With both fields perpendicular to the surface normal:

oo oo oo oo

{ N ( / ) d / = - R e J" E(co)dQ) / H(o)')*dco' { e ' ^ ^ ' d /

—oo -oo —oo —oo

oo oo

= TtRe j E(co)dco j H(a)')*dco'5((o,a)') (5.2.2)

—oo —oo

oo

= 27tRejE(co)H*(co)dco
o

The electric and magnetic frequency dependencies are related by:

"n|H(co)| = |E(co)| (5.2.3)

Combining shows that the power through the surface is:

oo oo

jN(/)d/=-^/|E(co)|2dco (5.2.4)
-oo ' 0

If the plane wave is turned on at time t = 0 and terminated at time t = T,

the result is a wave train of length / = xc. The relationship between the length
252 The Electromagnetic Origin of Quantum Theory and Light

of a wave train and the measured width of the frequency spectrum follows in
a way similar to that used to demonstrate the uncertainty principle. Consider
the special case where frequency co' is turned on at time - T / 2 and off at time
T/2, and let |co-co'| = Aco. The resulting electric field intensity is:

E(/) = E 0 e/to'/ |/|<|T/2|

(5.2.5)
=0 |/|>|T/2|

For this case

T/2
E(co) = ^> 7" e ^ ' - » ) V / = ^
{ AtOT
2
J ACOT
(5.2.6)
-T/2

The first zero of Eq. 5.2.6, half the width of the frequency pulse, occurs
when the argument of the sine term is n. For that case:

TAto = 27t (5.2.7)

Substituting the length of the pulse train, 1 = ex, into Eq. 5.2.7 gives:

/ to
— = n= =Q (5.2.8)
A Aco
By definition n is the number of wavelengths in the wave train and the
ratio Aco/co is the fractional bandwidth. With A equal to the wavelength and
for a fixed value of Q the minimum duration of a pulse is:

2TIQ QA
T = (5.2.9)
co

Feynman's estimate of the Q of a photon begins with the definition of

Eq. 3.5.11, and is repeated here:
Photons 253

Q=^ (5.2.10,

Let a point electron oscillate between positions at ±ZQ, thereby producing

electric dipole radiation. When oscillating at frequency co the maximum
energy of the electron is:

W = -mco 2 zo (5.2.11)

The power output of an electric dipole radiator is listed in Table 3.12.1, with
unit normalization. The normalization factor, -k ez0J^m, follows by
comparing the radial component of the electric field intensity with that listed
in the table. Substituting the actual values shows that:

Hence, the calculated Q, using Eq. 5.2.10, is approximately:

1 = (?<Q r4nV > ^

(5.2.13)
Q 6nemc3 3X
V-"V ^4nemc
2
j

The last bracket in Eq. 5.2.13 is the Lorentz radius of the electron,
2.82 x 10~ m. At the center of the optical spectrum the wavelength is on
the order of X = 530 nm, corresponding to a frequency of about
14
5.7 x 10 Hz and a period of about 1.75 fs. Substituting these values into
Eq. 5.2.13 shows that:
Q = 4.5xl0 7 (5.2.14)
7
Such an oscillator would need to radiate some 4.5 x 10 radians, or some
7 x 1 0 oscillations, before dropping to 1/e of its original intensity. With a
resulting decay time of about 10 ns, it follows that the wave train is about
254 The Electromagnetic Origin of Quantum Theory and Light

3 m long. With wave trains of this length, transient effects are not expected
to be significant.
It was shown earlier that the field energy also contributes to Q. For an
electrically small dipole the approximate Q due to the field energy is:

Q = i/(^)3 (5.2.15)

-11
At the radius of the first Bohr orbit, 5.29 x 10 m, and at the same
frequency:

ka= 6.27xl0~ 4 (5.2.16)

Combining Eqs. 5.2.15 with Eq. 5.2.16:

Q«4.0xl09 (5.2.17)

The Q of the dipole field energy is approximately 100 times larger than the
Q calculated using the kinetic energy of a point electron generating a dipole
mode. Therefore, the radiation Q of atomic radiation is significant, yet it is
ignored in analyses of quantum mechanical transitions.

5.3 Phase and Radial Dependence of Field Magnitude

The equation set that describes a circularly polarized, z-directed plane wave
is given by Eqs. 2.1.8, 2.1.10, and 2.1.11 and repeated here:

/^+i)iMpi(COS0);

w(2l + Q , N PJ(cos9) . . , xdPl(cos6)

E=S +1 CT
( ) . a ' + *M \a
exp(-yi)))
S ^ ) sin 6 d9
. , NdP>(cos6) ,,, NP>(cos9)
-J M CT ) \n + M P ) .
sin A6
TlH = yE (5.3.1)
Photons 255

The primary concern is radiation triggered by the plane wave. In the

scattering cases considered earlier, the phases of the scattered modes are
determined by the phase of the incoming plane wave and the phase angle of
the scattering coefficients.
Table 3.17.1 shows the radiation fields produced by four radiating,
collocated electric and magnetic dipoles, all of degree one. One electric and
one magnetic dipole are oriented along the x-axis and an identical pair is
oriented along the j-axis. As discussed in Section 3.17 the two sets of pairs
radiate n/2 out of phase. The calculation procedure shows there is no source-
associated standing energy and the radiation Q is low, possibly as low as
zero. The phasor multipolar fields that generate Table 3.17.2 are:

^ + 1 )^Mpi( c o s e );

h(q) ji(co g 8l + <h . (g) dPJ(co B e)'

«xp(-y<t>)
sin 9 d8

/ x dP f (cos9) ../ N Pf(cos0)

-J
de sin 6
T)H = yE (5.3.2)

As is the case for Eq. 2.2.1, the modal phases are the same as that of a plane
wave. In accordance with the requirements of Table 3.17.1, but differently
from Eq. 2.2.1, the magnitudes of the TM and TE modes are equal. The
question we address here is: if the fields of Eq. 5.3.1 trigger a metastable
source within the circumscribed sphere, what are the expected values of F^?
The first step in seeking an answer is to examine the phases of the field
components. On the positive z-axis the angular functions, see Table A. 18.1,
are given by:

Pj(l) = 0

PJ^ose), _dP](cos8), l ( l + l)
_le=0 le=0
~^ " de - ~1T
256 The Electromagnetic Origin of Quantum Theory and Light

Pj(-1) = 0
(5.3.3)
>1/
Pi (cos 9), _ dPi(cose), , AMt{i + l)
——^— le=o-
sine '"-" T7
de le=7t-l-ij ur-
From Eqs. A.24.9 and A.24.13, the expressions for the radial spherical
and related functions are:

^.j-t^
^(2s)!!(2^ + 2s + l)!!
s /+2s-l
^0(2s)!!(2^ + 2s-l)!!
s=
(5.3.4)
S e l+2s
( \- V ^ - 2 s - l ) ! ! £ (~1) c~
ydC) +1 2s
sf0(2s)!!^ - sto(2s-l)!!(2^2s)!!

. ,CT, ^ ( 2 ^ - 2 5 - 1 ) ! ! , > ~ (-l) s a ^ 2 + 2 s , ,

y*( ) = I J T - : r mT 2Ts V ^ - 2 s - y / ' 7 ^ V + 2s
^ ' s t o(2s)!!a - ' s t^ 0 (2s-l)!!(2^ + 2s)!! '

Combining Eqs. 5.3.2 through 5.3.4 shows the relative modal phases. With
terms in square brackets indicating phase only and using the zenith angle
electric field component as an example, on the positive z-axis:

E6 - r'{[ht(o)] + {h'e(aj\} = r%(o)+y;(o)]- / W [y,(a)-]}(a)} (5.3.5)

The first terms in each of the square brackets of Eq. 5.3.5 are due to TM
modes. The phases of the first term on the right side of Eq. 5.3.5 may be
written as

,-(\; ^ . „ » / / r V I - .-(/ i\s*e+2s ,-tt i\s_M-2s-2 , rt-t-2+2s

1 + a 1 +/
\it\<*) yt\p)\-' (-iJ -i r ) ° °
£ odd
£ = i;« {-/[„]+ /[o 3 ]- i[o5]+ /[07]...} + {/[o-1]- t[a]+ /[a 3 ]- /[a5]+ /[a7]..] - /cr"3
^3;={/[a 3 ]-/[a 5 ] + /[a 7 ]...} + {-/[a] + /[o 3 ]-/[^] + /[a 7 ]...} + /{[a- 5 ] + [a- 3 ] + [a-']}
Photons 257

^"even

«-*-M-M + ^H + H" a l + M-M + [4-hM t [^l + M*w}

(5.3.6)

The results contained in Eq. 5.3.6 show that for powers of a greater than
or equal to zero the phase of each power of the radius is the same for all
moments. Therefore, along the positive z-axis, and in the near field, driving
one dipole moment drives the corresponding far-field radial components of
all odd, higher order modes. Quite differently for powers of a less than zero,
higher order terms have opposite signs and act to cancel the total near-field
radial field component.
The second term on the right side of Eq. 5.3.5 is due to TE modes, and
may be written as:

' |yH CT )-M a )J = 2 / H)° +/

°
^odd
. = l;,2{[l]-[a 2 ] + [ a 4 ] - [ a 6 ] . . j + [a- 2 ]

* = 3;=2{-[a 2 ] + [a 4 ] - [a6]...} - {[<,"<] + [a" 2 ] + [l]} (5.3.7)

£even
/ =2;-2/^a] + [ o 3 ] - [ a s ] + [a 7 ]...}-/{[o- 3 ] + [a- 1 ]}

. = 4;»2/{[a 3 ]-[a 5 ] + [a7]...} + / { [ a - 5 ] + [a- 3 ] + [a-'] + [a]}

Since the results of Eq. 5.3.7 are the same as those of Eq. 5.3.6, driving
the magnetic dipole moments also drives the corresponding far-field radial
components of all odd, higher order modes. For negative powers of a the
terms have opposite signs and act to cancel the total near-field radial field
component. The relative phases of the two equations show that the dipole
far-field terms of Eq. 5.3.6 produce the same phase, even order terms as does
Eq. 5.3.7, and the dipole far-field terms of Eq. 5.3.7 produce the same phase
odd order terms as does Eq. 5.3.6. In this way, the system is phased so the
dipole terms drive all higher order terms.
258 The Electromagnetic Origin of Quantum Theory and Light

On the negative z-axis both TM and TE modes contain alternate signs of

the expansion modes, with canceling phases and no field buildup.
We conclude that a buildup of a term proportional to a , where n > 0, by
any mode builds the magnitudes of the far-fields for all modes for z > 0 and,
at the same time, reduces the magnitudes of the near-fields. This condition
lends itself to a regenerative buildup of field magnitudes.

5.4 Gain and Radiation Pattern

The gain of an antenna is a power ratio; it is the ratio of the product of the
maximum power density on the surface of a virtual, circumscribing sphere
and the surface area to the average surface power. In mathematical terms:

, . Lim 47TCT2 [ N r ] m a x
Jmax
G(o)= , p (5.4.1)
a ^ o o ki p av
Both Einstein and Planck referred to the "spherical symmetry" of
radiation modes. It was surely this idea that was the basis for Einstein's
comment that the full directivity of quantized radiation made a quantum
theory of radiation "almost unavoidable" Certain radiation modes do,
however, carry linear momentum. Harrington quantified the maximum gain
in 1960, about four decades after both Einstein's 1917 paper on directivity
and Planck's 1920 Nobel prize paper that addressed the issue.
Consider the gain of fields described by Eqs. 5.3.2, after making the
equality / =/. For this case the maximum value of Nr(a,0,(f>) occurs at angle
8 = 0. Making this substitution and using Table A. 18.1 from the Appendix
gives the fields:

E(o,0) = - S ^ F ^ + 1 )[h/(o) + '•h;H]{e-/•*}e-' i , ,

1
£=l (5.4.2)
riH(o,0) = i1 X /»-%l{l + l)[he(o) + t h?(a)]{e-/ fy"*

The radial component of the Poynting vector is:

Photons 259

1 °° °° -.*

T|Nr(of0) = - 2 r%£{£ + lihe(o) + Ai'((c)}x £ / , F n n(n+l)[h n (a) + /h*(o)

'/=! n=l
(5.4.3)

Limiting forms of spherical Hankel functions are given in the Appendix:

Eqs. A.25.17 and A.26.4. Making the substitution gives the maximum value
of the radial component of the Poynting vector:

N r (o,0) = - (5.4.4)
*1<*

Using the fields of Eqs. 5.3.2, the output power on a virtual sphere of
limitlessly large radius is:

2 00 2jl 7t ./(/+!) Pi dPi

P
av=^-2-i: J#Jsin6der^]2- [<M]
2X\k (=i 0 0 sin0 de
(5.4.5)
-*{M)
fi e Pi dP]
x/ [e+4]
sin0 d6

Evaluation gives:

(5.4.6)

Substituting Eqs. 5.4.4 and 5.4.6 into 5.4.1 gives:

"i2

G(c) = /=1 (5.4.7)

SFi 2^+1
f=l
A particularly interesting special case is if the modal coefficients satisfy the
relationship:
260 The Electromagnetic Origin of Quantum Theory and Light

F/ = —. £ (5.4.8)

For this special case the gain is:

oo

G=^{2£ + 1) (5.4.9)

This expression for gain, first derived and published by Harrington in 1960,
vividly demonstrates that the radiation of spherical modes can be arranged to
support power that does not possess circular symmetry. That is, a net transfer
of linear momentum is possible.

5.5 Kinematic Values of the Radiation

Radiation from electrically small sources is dominated, in the main, by the
moment with the lowest power of ka. A primary reason is that the Qs of
electrically small antennas producing single modes increase rapidly with
increasing order, i.e., as |y(o)| of Table 3.2.1. However it was shown in
Section 3.18 that a multimodal source generating the fields of Eqs. 5.3.2
does not necessarily extract a returnable standing energy from the source. By
Eq. A.28.12 and Eq. A.29.18, the magnitudes of electric and magnetic
multipolar fields of order £ are respectively proportional to
{/taf and (to)M. The magnitude of fields scattered by a passive, electrically
small object will, therefore, decrease rapidly with increasing modal order,
see Eq. 2.3.6. For the case of interest here, however, the scatterer is not a
passive object but an excited, eigenstate electron. The host atom is immersed
within a z-directed, circularly polarized plane wave that triggers a nonlinear,
radiating transition to a lower energy eigenstate. We seek to find radiation
details that apply during the transition.
With p z and / z representing respectively linear and angular momentum,
the kinematic properties of atomic radiation, i.e., photons, are:

fr/pz = c /r// z =co Pz//Z =* (5.5.1)

Photons 261

To examine the radiation fields using classical field theory consider the
equation set of Eq. 5.3.2 and examine the rate at which energy, linear
momentum, and angular momentum exit through the surface of a sphere of
radius a/k circumscribing the atomic source. The rates are:
Energy:
Use of Eq. 1.9.11 shows that the rate of energy loss through a spherical shell
is:

AW rr2 2 n n
= _ _ f d<j)fRefNr]sine«e (5.5.2)
d/ IT2- *
* 0 0
Linear Momentum:
2
By Eq. 1.9.7, the momentum contained within a volume is equal to \lc
times the volume integral of the Poynting vector:
1 2n n
/ ^ - y ^ - J V d c j j " d<))jRe[N]sin0de (5.5.3)
c
& 0 0

Since the equality holds for every volume in space the rate at which the z-
component of momentum exits a closed volume is equal to c times the
surface integral of the z-component of momentum:

j 2 2TI n
^ • = - ^ y f d(t)fRe[N r cos0-N 0 sin0]sine^e (5.5.4)
d/ ck 0 0
Angular Momentum:
Angular momentum is related to linear momentum by Eq. 4.8.1; the rate at
which z-directed angular momentum exits a closed volume is:

,. 3 2TX 7i
-Z- = ^ T Jd«|> i fRe[N ( „]sin 2 eflB (5.5.5)
ck 0 0

Using the fields of Eq. 5.3.2, treating the coefficients F^ as unknowns,

and puttingy = / shows the Poynting vector to be:
262 The Electromagnetic Origin of Quantum Theory and Light

CO OO
( h / h ; + hsinex 1d6 ^sinG
A^ d6
Re y
(5.5.6)
Z
'I e=\ n=i Pj P^ | dPJ d P ^
-^h**-h*h*j
sine sine d6 d6 J)

Re r l P'P'
Nfi=- X XF,F n % n
^[n(n + l)h* n h-, + ^ + l)h,hr]^ (5.5.7)
2ot| = 1 n=l

OO OO
ntn+lKh^^-^ + l ^ O i ^ -
(5.5.8)
N,=^S
!*=1 n=l
W<
+ l)+n(n+l)]hX^

Substituting these values into Eqs. 5.5.2, 5.5.4, and 5.5.5, evaluating the
integrals using Table A.22.1, and replacing the spherical radial functions by
letter functions gives:

(5.5.9)

* ^ + l)/ 7 2 2 ?\ *2^2(^ + l ) 2 / -
F F F c +B D
' 'ra 2 2
I ' '"o(M"^ ' ' ^
. l (l + l)(l + 2)
+i +1 _ + B u
w+17277T5f2T+?T^ * *+i ~ B ^+i u ^/
d
/ , 2 = R c y 2TI „ * 2/ 2 (^ + l) 2 (f + 2) 2 .
d/ /TiTiot2 " 2 F
' F / + 1
( 2 ^ + 1 ) ( 2 ^ + 3)G [ A e C M + B ( D M
>

* (£-\ft(e+i)2, ,

~ 2 F A " 1 (it+iyie-iy, (AA-I + B^_,J

(5.5.10)
Photons 263

l 2 ( l + l)

d/, 4na _ ^ 2 2 2 '(A,C,_,+B,D,_,)

R e
„r F r -\('-l) ' (<+l)
=
"H7 ^ " ^ -( F , F w F w F -/(A^D^-B^C^)
' 'llHpH)
(A^C£+1 + B^D^+1)
-0 2 2
fy+lfy -FcF// + 1 ^ ( l + l ) ( l + 2)
2
/ 2(2^ + l)(2^ + 3) -/(A^D^ + ) -B^C^ + 1 )
(5.5.11)

In the limit as the radius becomes infinite:

dW 471 v C C * ^ + 1 ) 2
b b
d/ t i ^ i T2^i t e ( 2 ^ + 1 )

&z 4TI y <('+!) F F *,FF* ^ + 2 f , F F* ( < - l f ( ^ + l) (5-5.12)

** 2
TIC* /ti (2^+1)' (2^ + 3) (2<~1)

d/, 47t l 2 ( l + l) 2
at nco^ /=1 (2£+l)

Equations 5.5.12 show that both the energy-to-angular momentum ratio and
the energy-to-linear momentum ratio depend upon the magnitude of
recursion relation F^. Before solving for F^, it is necessary to consider some
additional factors.
To examine properties of a field described by Eqs. 5.3.2 note that sums
over spherical Neumann functions can be put in closed form only at very
large and at very small radii. For a very large radius note from Eq. A.24.15
that the limiting values of functions are:

Lim , . 1 71,
u CJ) = — cos a-f(l + l)
(5.5.13)
Lim / N 1.
yj0l = — sin o-|(/n, n.
+ l) = —cos
a a-f(/+2)/
The two functions differ in phase by TI/2. Next, multiply the N e u m a n n
functions by (±/), as is done in the formation of spherical Hankel functions.
264 The Electromagnetic Origin of Quantum Theory and Light

This changes the phase by another nil and brings the functional form to
either the same phase or n out of phase, depending upon whether the phase
shifts add or subtract. The result is that in the limit of infinite radius
changing spherical Bessel functions to spherical Hankel functions results in
the far field sum of Bessel and Neumann parts either doubling the field terms
or summing to zero.
In the limit of small radius, the two functions are equal to:

Lim . / \ _ o Lim (21-1)!!

(5.5.14)
k a
3=*0 ^ '~ (2£ + l)V.and yAG) = - —TT—
CT=>(T A ' </+i

Bessel function solutions are continuous through all orders at the origin but
Neumann function solutions undergo an (^+l)-order singularity. Spherical
Neumann functions therefore play an essential role in the description of
scattered and generated fields. In a step we support during the rest of this
chapter we assert that the correct recursion relationship is:

_(2£+l).
F,= (5.5.15)

F is real and independent of t. Substituting the relationship back into field

Eqs. 5.3.2 repeats Eqs. 5.3.1 except spherical Bessel functions are replaced
by spherical Hankel functions; of course with Hankel functions, the limit
values obtained using Eqs. 5.5.13 and 5.5.14 apply.

«(*+i)^pj(cose)r
,(2£+l) , , vPj(cose) ,., .dpj(cose)
A ; n exp(-^)
sine ' d0

. vdP](cos6) . . , vP](cos6)
h g +
-J H ) \a M°) \ a

TiH = yE (5.5.16)
Photons 265

Since the source is much smaller than a wavelength, the rules of geometric
optics apply in the region where the wave interacts with the source. There
the phase of the Neumann function undergoes a step change of m and, in
optical terms, is a caustic. It is expected, therefore, that if the Bessel and
Neumann contributions to the teledistant terms are in phase along the
positive z-axis they will be 7t out of phase along the negative z-axis, and vice
versa.
Substituting Eq. 5.5.16 into Eqs. 5.5.12 gives:

Urn d ^ . ^

Lim d/7z 4TTF2 £ . .

Lin, d4 = 4*Fi£

It follows from Eqs. 5.5.17 that:

Lim AW 1
c
d / / d/
Lim dW / d / z _
CO (5.5.18)
>~d7/ d/
Lim /d4_ : ^
- d / / d/

Since the time-variation of the three kinematic properties is identical,

Eq. 5.5.18 leads directly to Eq. 5.5.1. Therefore, the kinematic properties of
this radiation are the same as the kinematic properties of photons and there is
no dichotomy between the kinematic properties of photons and classical
field theory. All three parameters are proportional to the gain, see Eq. 5.4.9.
Next, consider what requires Eq. 5.5.15 to be uniquely correct. The
requirements of resonance and zero Q are met by Eq. 5.3.2. As discussed in
Section 5.3 the phase portion of Eq. 5.5.16 uniquely defines fields for which
equal powers of a have equal phases in all modal orders, creating
266 The Electromagnetic Origin of Quantum Theory and Light

regenerative feedback. The near fields are properly phased to cancel. To

consider the relationship further it is necessary to examine the full set of
photon fields.

5.6 Telefields and Far Fields

The field intensities of a circularly polarized, z-directed plane wave are
expressed in spherical coordinates in Eq. 5.3.1, and by:

E = e~' b c o s 6 (rsine+ecose-/0)e~AK t|H = / E (5.6.1)

Since the magnitudes of the field terms are independent of distance, define
them to be telefield terms. For comparison, far field terms are proportional to
2
I/a, inverse square terms are proportional to 1/a , and near field terms are
proportional to 1/a where n > 2. Since the electric and magnetic field
intensities are related as described by Eq. 5.6.1, it is sufficient to solve for
the electric field.
The Uniqueness Theorem requires Eqs. 5.6.1 and 5.3.1 to be identical.
By Section 5.5 the external radiation fields produced during state decay is
obtained by inserting Eq. 5.5.15 into Eq. 5.3.2. The result is:

, , ,PJ(cos0) ,., ,dPi(cos0)l-

*(/+l)ijMpJ(co8e)rH
sin 8 d6
-»>
. , N dP](cos9) ,., N P](cos6)
n A
' d9 ' sine
(5.6.2)

To obtain more useful expressions for the near field terms it is helpful to
evaluate the sums of Eq. 5.6.2. For this purpose, introduce special sums over
the spherical functions, defined by:

s
l ( a > e ) = E / J '(2* + l)MCT)P](cose)=S 1 1 (a,8)-/S 1 2 (o,e)
Photons 267

/ \ £ - / ( 2 * + l), dPl(cos6) „ , ft. „ , n\

h ;
s 2 (q,e)=S> j777iy <(°) de -s 2 1 (a,e)-/s 2 2 (a,e)
(5.6.3)
/ \ £ ,./(2^ + l), P](cos0) , > „ / „x

Sums S n ] are over spherical Bessel functions and sums S n 2 are over
spherical Neumann functions. Once the functional form of S3 is known, S 2
follows using the relationship:

S 2 (a,e) = ^[sineS 3 (rj,e)] (5.6.4)

To complete all needed information about the radiated field it is also

necessary to evaluate the forms:

$ 2 (o,e) = - — [ o S 2 ( a , e ) ] $3(cr,0) = -—[aS 3 (rj,6)] (5.6.5)

Combining all the above shows that:

oE^S^o.eJe"*
E e = [$ 2 (a,e)+ S 3 (o,6)]e-^ (5.6.6)
E(t) = -/[s2(CT,e)+$3(0,e)]e-'*1,

Consider the sums over spherical Bessel functions. The first sum follows
by equating the radial components of Eqs. 5.6.1 and 5.6.2; it also follows
from Eq. A.27.6:
00
1
SnM)= £ / " ^ + l)j^(a)p](cose) = asinee- / & c o s e (5.6.7)

To evaluate the other sums, equate the angular components of Eq. 5.6.1 and
Eq. 5.3.1 to obtain:
268 The Electromagnetic Origin of Quantum Theory and Light

cos6e~ s 9 = y ,^(2£+l} , N Pi (cose) . . , C TN dPi(cos6)

e(e+\) u (v ° )' sine + 'J)nM ; ) — —d6
(5.6.8)
-/CTCOS0 _ V , , dPi(cose) . . , v Pi(cos9)
= 1i j-1 ^( 2+1 1+ l)) (g) \Q
d6
+ €<(o)
v \
' sineQ

To evaluate the sums, use the identities from Eqs. A.21.1.1 and A.21.1.5:

-^p;(cose)=i[^+i)p,-p21
_d
de
>1
(5.6.9)
v l e
sin9 2cos6L ' i

Substituting the identities into Eqs. 5.6.8 gives:

X F%U 1)1 J < ( o ) ^ ^ + /C(a)P,(cosG)

-/CTCOS9 1 e=l
cos6e
,y re(2t+l)
J,W%^-^(^ 2 (cose)
h tu+i)
. / \~ / „\ .•/ N P/(cos6)
£r'(2/+i) J< (a)P < (cos9) + j<(o) ^Vos9 >
-/acos9 1

-1 wM+ 1 ( P?(coSe)+^(a)%5)
J* o)
A ^ )
(5.6.10)

The two sums on the top lines of both of Eqs. 5.6.10 sum to a known
polynomial of trigonometric functions, with the assistance of Eq. 5.6.7. The
result is that Eqs. 5.6.10 form two linear, algebraic equations with the sums
on the bottom lines of both Eqs. 5.6.10 as unknowns. The equation forms
are:

( 1
fj(o\cos6) = - /y f 2 (o\cos6) = — x + - (5.6.11)
cos6 cos 6,/
Photons 269

Solving Eq. 5.6.11 and carrying out the details yields the equality:

-/CJCOS0 COS0
-e 1 + 2/
a sin2 6
e
£'''f^'<* >= +/ . COS0
2 cos a -/sin a
(l + cos 2 e)
asin 2 0 COS0

(5.6.12)

Next, sum over each of Eqs. 5.6.9 after multiplying through by the factor:

^ ( 2 1 + 1). , v

In each equation, the sums on the right side are those of Eqs. 5.6.7
and 5.6.12, and give exact values for the sums S21 and S31. From these sums
come exact values for $21 and $31. All sums over spherical Bessel functions
are listed in Table 5.6.1.
The sums of Table 5.6.1 evaluated on the positive and negative z-axes
and in the equatorial plane are listed in Table 5.6.2,

S 11 (a,8) = CTsin0e-id cos 8

S 2 i(a,0) = e-'CTCOse+ l
2
-I'CTCOSG
cos 0 - (cos a cos 8 -i sin a )
a sin 8
S 31 (o,9) = ^2[ e - ' o c o s e - ( c o s a - / s i n a c o s 0 ) ]
a sin 8
$2i(CT,e) = c o s 8 e - i o c o s e + l -icrcosG (cos a - i sin a cos 6)
rjsin 2 8 L
$31(o,e)= -iacos0 c o s 8 - ( c o s a c o s 8 - / s i n a )
asin28L

Table 5.6.1 Closed form solutions for the field sums defined in Eq. 5.6.3, over
spherical Bessel functions.
270 The Electromagnetic Origin of Quantum Theory and Light

1. S n (o,0)=S 1 1 (o,7t) = 0 Sn(a,n/2) =a

-ia sina ,n sina

2. S 21 (o,0) = | S2l(CT,7r) = - e
CT

sina
S21(a,7r/2) = 1-

-to sina ,a sina

3. S 31 (o,0) = | S 3 1 (O,TI) = - - e
a

S 31 (a,7i/2) = — [ l - c o s a ]
-;/aa sina ,w sina
4. $2i(a,0) = | e + hi(w)=-\ e +

$2I(CT,TI/2) = —[l-cosa]
a
-OT sina ,„ sina
5. $3i(a,0) = | e + $3l(M=| e +
a
^ i ,„\ sina
$ 31 (a,7i/2) =

Table 5.6.2 The sums of Table 5.6.1 on the coordinate axes

5.7 Evaluation of Sum S12 on the Axes

Evaluations of sums over spherical Neumann functions can be made on the
axes but not at other angles. The expression for the sum Si2, see Eq. 5.6.3,
is:

S i 2 M ) = X / 1 ^(2^+l)y^(a)PJ(cose) (5.7.1)

To evaluate this sum on the axes note that since the radial component of the
field is proportional to sin9, it is equal to zero on both the positive and
negative z-axes:
Photons 271

S12(cT,0)=S12(a,7i) = 0 (5.7.2)

It remains to evaluate the sum at the equator, 0 = n/2. The series form of the
associated Legendre polynomial at 0 = n/2 is given in Table A.18.1 and
repeated here:

Pi(0)=(-ir 1 " 2 (^|^,2n + 1 ) (5.7.3)

The Kronecker delta function, with n representing all integers equal to or

greater than zero, shows that the function vanishes for even numbered
modes. The series form of the spherical Neumann functions is shown by
Eq. A.24.12 and repeated here:

v (g) - V (2l-2s-l)!l <- 1+2s y (-If a"J-l+2s

£ (5.7.4)
^
s=0 (2s)!! s ~ ( 2 s - l ) ! ! ( 2 ^ + 2s)!!

Combining Eqs. 5.7.1, 5.7.3, and 5.7.4 results in:

^(2l-2s-l)!! , 1+2s
2s !!
>12 <H-2(2M
71 h () (5.7.5)
fo;l (-l) s r/~ 1 + 2 s
+z
fj(2s-l)!!(2^ + 2s)!!

Lower limit 7 o ; l ' indicates the sum is over odd values of t and the lowest
value is one. Since only odd modal orders contribute to the sum, by Eq. 5.7.5
only even powers of a are present and Eq. 5.7.5 has the form:

71
S12 < = 2 ^ (5.7.6)
V l) ne;-{M)

Solving Eq. 5.7.6 to obtain coefficient A n results in Eq. 5.7.7:

272 The Electromagnetic Origin of Quantum Theory and Light

( ^ (2^l)(^)!!(-l) ( 1 - 1 > / 2
(^-l)!!(^ + n+l)!!(n-^)!!
A= f
Lim ^ k (2l + l)(l)!!(l-n-2)H
(5.7.7)

vj j

L is the largest modal number present. Consider the coefficients with n

greater than zero. For the special case of coefficient A2, the terms equal:

[3 7 55 105 1995 5313 243243

=0 (5.7.8)
[8 32 1024 4096 131078 524288 33554452'

The first two terms come from the first sum of Eq. 5.7.7 and the rest from
the second sum. Although the equality is correct only in the limit where L is
infinitely large, the series converges rapidly for positive powers of a, and
with L large but finite the sum approaches zero rapidly.
The coefficients of all other positive powers of a follow in a similar way,
and all of them are equal to zero. Although no individual term is equal to
zero each modal order contributes the proper magnitude and phase for the
sum to equal zero. Since the total field is equal to zero so is the field energy
on the axis. Note that if the magnitudes of all coefficients are changed in a
way that preserves recursion relationship Eq. 5.5.15 the field remains equal
to zero independently of the magnitude of the coefficients. On the other
hand, if any single mode were different from the value of Eq. 5.5.15 the field
of that mode would support a large field energy. Should a variation from that
value occur the generalized gradient of the added energy is a forcing
function that acts until Eq. 5.5.15 is restored; this is a unique characteristic
of recursion relationship Eq. 5.5.15.
For negative powers of a, the series diverges and, therefore, must
terminate. With the coefficients of all higher order modes equal to zero, the
series is equal to:

»12 2 (5.7.9)
) e£-l 'CM" (<-n+l)!!an
For the special case n = 0 the series of Eq. 5.7.9 is:
Photons 273

Jl 21 J^5 2625
l , 2 1 6 + 1 2 8 + 2048 + ' "J
+
(5.7.10)
= -(1.5+1.3125+1.2891+1.2817+..) = -J0

Each term approaches unity as the modal number increases and A o i s

proportional to L. Very nearly, for L much larger than one:

4> = L/2 (5.7.11)

For the special case n = -2, the series is equal to:

f„ 63 2475 165,375 16,967,475 "\ A

- 3 +—+ + ' +— • + .... \ = -A7 (5.7.12)
V 4 64 2304 147,456 )
3
Aj is proportional to L . Since a term-by-term expansion shows that
i t

An ~ L the sum, evaluated at the equator, is equal to:

S12(a,n/2) = - £ CT 4
n=0;e
(5.7.13)
L (2/+l)fl!(/+n-2)t! +1

" ^ , (/-!>! ( / - n + l \ l ! " L

,o;n-l

Each coefficient An contains contributions from all modal orders.

In summary, only negative powers of a are present in the field
expression and energy of the radial field component is localized to the
source region.

5.8 Evaluation of Sums S22 and S32 on the Polar Axes

The expression for sum S22> as defined by Eq. 5.6.3, is:
274 The Electromagnetic Origin of Quantum Theory and Light

c i „\ •? - c ( 2 ' + l ) / >dP<(cose)
SaM= (o) (5iU)
£ ' V>r ~*
The series form of the associated Legendre polynomial at 0 = 0 and n are
shown in Table A. 18.1 and repeated here, see also Eq. 5.3.3:

dFJ (cos e), _ / ( / + !) dPJ(coge). _ tl{t+l) .....

le=0 le= (_1) ( }
de " ~T~ de *~ ~T~
The expression for sum S32 as defined by Eq. 5.6.3 is:

c t n\ £ ^(2^+1) P)(cos6)

The series forms of the associated Legendre polynomial at 9 = 0 and 71 are

shown in Table A. 18.1 and repeated here:

Pjfrose), _ W + Da n d . PJ(cos9), „<+!*(* + 0 .....

— sin9
r-r—e=o-—2 ;— ~ z — e=ji - ( - 1 )
— r sin9 —2 ~— (5-8-4)

Comparing axial values of the two sums shows that:

S 2 2 (CJ,0) = S 3 2 (CT,0) and S22(CT,TI) = -S 32 (a,7i) (5.8.5)

Because of these equalities, it is only necessary to evaluate one sum on the z-

axes. Substituting Eqs. 5.8.4 and the expansion for the spherical Neumann
function, Eq. 5.7.4, into Eq. 5.8.3 gives the series expansion:

1 °°
S32(G,0) = -^re(2£ + l)ye(G)
1
i=\
1 °° 1 1 °°
cos a
= -2 X r£(2l+l)ye(o)--y0(a)2 =2- £ rt(2£ + l)ye(a) + 2cT
£=0 /=o
Photons 275

(-Ds J-l+2s
1 °° s e 0 (2s-l)!!(2/+2s)!! COSCT
•(21 + 1) >+ •
£-1 ~2cT
(2l-2s-Wa-t-i+2*
s=0 (2s)!!
(5.8.6)

For convenience in evaluating the sum, the I = 0 term is left in the expansion
and subtracted from the total. Both even- and odd-numbered modes are
present; even values of I produce odd powers of c and vice versa. Within the
curly brackets on the bottom line of Eq. 5.8.6, the upper term contains
powers of a ranging from (i - 1) to QO, the lower term contains powers of a
ranging from -(£ + 1) to (£-3). The upper term contains only positive
powers of a and the lower term contains both positive and negative powers.
Next, let n be a positive integer and determine the coefficient of o .It is
convenient to separate Eq. 5.8.6 into sets of different parity:
n
( _ 1 ) n /2n+l (2£+l)<J

2 ~ , ( n - * ) ! ! ( n + * + l)H
S
32 = (5.8.7)
©o
(-l) ( ^~ 1)/2 (2^ + l)(^-n-2)!!rj n
(* + n+l)!!
' &)=n+3
n
( _ 1} (n-l)/2 n+1 (2£+l)c

2 £ 0 ( n - * ) ! ! ( n + * + l)!!
J_ y (-l//2(2l + l)(l-n-2)!!an cosrj
" 2 *e=n + 3 ^ + n+1 !!
) + 2
°

The upper sign of the ± sign is to be used at 9 = 0 and the lower sign at
8 = K. Defining S32' to equal the top row of Eq. 5.8.7 and expanding the
series results in:
276 The Electromagnetic Origin of Quantum Theory and Light

n even, i odd.
7 (2n + 3)
n/2 - + ...+(2n + 2)!!
(n-l)!!(n + 2)!! (n-3)!!(n + 4)!! (5.8.8)
, <-D
S32'-
(2n + 7)(l)!! (2n + ll)(3)!! (2n + 15)(5)!!
(2n + 4)!! (2n + 6)!! (2n + 8)!!

The bottom row of Eq. 5.8.8 may be evaluated by writing

(2-M-l) = (^?+n+l)+(^-n) and regrouping the terms as:

f
(1)!! (1)!! 3 ^
K)
, _ . . . . + •(2n+4)!! 3 - ( 2 n + 6 ) (2n+6)
(2n+4) (2n+4)!! 1
V (5.8.9)
(3)" ( 5 (2n+2)!!
K)
"
(2n+6)!! 5 - ( 2 n + 8 ) - + ...
(2n+8),

Inserting Eq. 5.8.9 back into Eq. 5.8.8 results in the equation:

n even, £ odd.

-+•
n/2 (n-l)!!(n+2)!! (n-3)!!(n + 4)!! (5.8.10)
S
, /("I)
32'~
(2n+3) 1
(2n+2)!! (2n+2)!!

Evaluating Eq. 5.8.10 for the special case of n = 0 gives:

/ 3 1
2V2 2, 2

Repeating the process for the special case of n = 2 results in:

Y3 7 p
\,8 48 48, 4

Repeating the process for all even, positive values of n equal zero or more
then summing results in:
Photons 277

n even, I odd.
i (5.8.11)
S 3 2 = —COSCT

Repeating the process for the opposite parity results in:

n odd, I even.
(5.8.12)
S32 = ± —sina

Combining results for all non-negative values of n:

Positive powers of o\
-i„
/a cos a . j* coso (5.8.13)
S 32 (o,0) = ^ /e + S32(a,Tt) = /e
a o- ;
. ,a cos a
S22(a,0) = I[*-*+^ S 22 (a,Jt) = - - /e
v

The related sums are obtained by operating on Eqs. 5.8.13 using Eq. 5.6.5:

Positive powers of o.
-to cosa ( cos a
$32(CT,7t) = - (5.8.14)
$32(<*.0) = d * " /e +-
2V a v
-m cos a ,(j COS O"A
$22«J.0) = - it $ 2 2 (a,n) = /e +
/
The transcendental functions in Eqs. 5.8.10 and 5.8.11 result in a
multimodal sinusoidal wave that propagates away from the source. Each
mode contributes the proper magnitude and phase for the infinite sum to
equal these specific functions. Although the series terminates at modal
number L the series converges rapidly and the results are quite accurate for
finite values of L. Note that, as with results obtained using spherical Bessel
functions, the magnitude of the first term in each sum is independent of a.
These characteristics are unique properties of recursion relationship
Eq. 5.5.15.
278 The Electromagnetic Origin of Quantum Theory and Light

The coefficients of negative powers of o~ appear only in the second sum

of Eq. 5.8.3. Writing negative powers of a as positive values of n leaves the
equation form:

L+l
^ ( 2 f + l)(* + n-2)!!
-i (=n-\
E (€-n+l)!!o
^ 5 ( ^ + n,2q + l)

These sums are most easily evaluated by writing (21+1) = (£-n+\)+(£+n)

then expanding and regrouping. The result is:

(2n-3)!! (2n-l) (2n-l)!! (2n+l)

(2n-l)-(2) (2n+l)-(4)
J-n (0)!! (2) J (2)!! (4) ,
2o n (2n+l)!! rL+n 2(L+n)!!
+ (2n+3)-(6)^±^U...+ "
(4)!! (6) (L-n-1)!!

All terms except the last one are equal to zero, leaving only the series
remainder:

TL_1 (L+n)!!
n
2CT (L-n+1)!!

The full set of solutions for negative powers of o~ are:

rL-l L+l L-l L+l

(L+n)! (L+n)!!
s32M = - L2r-E
n = 1 ( L - n + l)!!a
n MM'-VS-
2 ( L - n + l)!!a n=1
n

(5.8.15)

2 ^ ( L - n + Dlta" 1 '" ' 2 n t 2 ( L - n + l)!!<T""

Although the magnitude of the modal terms increases with increasing modal
number, only the remainder is left and it results from the highest order mode
only. Therefore individual modes contribute nothing to the field and are not
affected by them, so long as the relationship of Eq. 5.5.15 is maintained. If
the relationship is disturbed the disturbing mode will create a field and that
Photons 279

field will generate a radiation reaction force in a direction that reduces the
field, thereby minimizing the field energy and maintaining Eq. 5.5.15.
Results are shown in Table 5.8.1.
- L - l L+l (L + n)!!
s22(c,o) = s 32 (o,o)=4[*-*+^!£
2 ^(L-n+l)!!^
L
if * cosa\ T y (n-l)(L+n)!!
$ 22 (a,0) = $ 32 (a,0) = n+1
• v °" 7 2 nt2(L-n+l)!!a

1 ( .^fa cos a\ L-l L+l (L+n)!!

S22(cf,Jr) = -S 3 2(a,7i) = — r | /e n
2 2 nfi(L-n+l)!!o-

1 ' ./a cosa"l /LI^ (n-l)(L+n)!!

$2 2 (a, 7i) = - $ 3 2 (a, 7t) = n+1
v a J 2 nt2(L-n+l)!!a

Table 5.8.1 Field sums over spherical Neumann functions on both the
positive and negative z-axes.

5.9 Evaluation of Sum S32 in the Equatorial Plane

The expression for sum S32, as defined by Eq. 5.6.3, is:

c / a\ v ^(2^+1) , , Pi (cose) (5.9.1)

We seek to find a simpler expression. At 0 = 7t/2, the value of the Legendre

function is, see Table A. 18.1:

m=^^m^) (5.9.2)

Substituting Eq. 5.9.2 and the spherical Neumann function, Eq. 5.7.4, into
Eq. 5.9.1 gives the expression for the sum at the equator:

-l+2s
(2l+l)(l-2)!! y ( 2 / - 2 s - l ) ! ! /-I+2, (-D s
S
3 2 | °">^ -X (£+1)!! ~ 0 (2s-l)!!(2* + 2s)!!
s=0 (2s)'!
(5.9.3)
280 The Electromagnetic Origin of Quantum Theory and Light

Since only odd orders of £ are present, there are only even powers of a and
the sum over the positive powers of o has the form:

>32 (5.9.4)

Combining Eqs. 5.9.3 and 5.9.4 gives:

n
j =( Dn/2 y^+l)(^-2)!!(-l)(^1)/2 y (2l + l)(£-2)!!(l-n-2)!!
n
^(^+l)!!(n-^)!!(^ + n+l)!!
&>;1
^+3 (^ + l)!!(^ + n+l)!!
(5.9.5)

Consider as special cases the coefficients ^ a n d ^ . Writing out

Eq. 5.9.5 term by term for these cases gives:

3 7 llx32
=1
2x2 4!!x4!! 6!!x6!!
j
(5.9.6)
z
3 7 llx3 )_
4= - + — + ——+.
2x4!! 4!!x6!! 6!!x8!! 6

Extending the evaluation to all positive values of n, then summing gives:

( 7^ sin a
>32 = /- (5.9.7)
l
V Jn>0

To evaluate the coefficients of negative powers of a, define a new set of

coefficients B n and write the expansion as:

>32
(5.9.8)
n<0 &

Combining Eqs. 5.9.1 and 5.9.8 shows that:

Photons 281

£ (2^l)(<-2)M(<+n-2)!!

The coefficients of the first two terms are:

Bj = /1.31125+1.28906+1.28174+1.27853+...) * 5L/8
3 (5.9.10)
£4 = /(36.09375+69.21387+106.58936+165.99072+...) - L3

If n increases by one the power of L increases by two.

Combining equatorial values for both positive and negative exponents
gives:

/ ,«\ sinfj v-i B-n

S 3 2 (a,Jt/2)=/ +/£-f (5.9.11)
° n=2°

The related sum follows by operating on Eq. 5.9.11 using Eq. 5.6.5:
*, / ,~\ cosa ^ (n-\)Bn
V
$32(a,«/2) = — — + £ J / (5.9.12)
° ne;2 O
Each mode contributes just the correct amount so the entire set of positive
powers of a are expressed by the transcendental functions of Eqs. 5.9.11
and 5.9.12; this is another unique property of recursion relationship
Eq. 5.5.15.

5.10 Evaluation of Sum S22M1 the Equatorial Plane

The expression for sum S22 as defined by Eq. 5.6.3 is:

We seek to find a simpler expression for the sum. At 6 = n/2, the value of
the Legendre function is, see Table A. 18.1:
282 The Electromagnetic Origin of Quantum Theory and Light

dPJ(O) /(^+l)!l , x
—£li=/^r^"S(^2q) (5.10.2)
d6 (€-2)!! v '
Since the derivatives of odd order Legendre functions, with respect to 9, are
equal to zero at the equator, only even values of I appear in the summation.
Substituting the spherical Neumann function, in the form of Eq. 5.7.4, and
Eq. 5.10.2 into Eq. 5.10.1 gives the expression for the sum at the equator:

( n) " (2l+l)(l-l)!! fc'(2l-2s-l)!! , 1+2s " ( - l ) s o ^ 1 + 2 s ] cosa

22 + +
r'2j-,to M» \h (2->l sf0(2s-l)!!(27T^)ii) - ^ -
(5.10.3)

For convenience in evaluating the sum, the ^ = 0 term is added and

subtracted. Since only even orders of £ are present, only odd powers of o
appear in the sum. The result is:

( lO n cosrj
>32 CT
'T = E ^ ^ +- (5-10-4)
noil °

Comparison of Eq. 5.10.4 with Eq. 5.10.3 shows that:

\£/2
n [
~ ' L
£e;0 W"
V->" v(n-*)l!(/
,,
~- l /"V^ + n+ n' +V"
l)!-f +3
fen+3 (<>! (/ + n+l)l!
(5.10.5)

Evaluation of the special cases n = 1 and n = 3 gives:

1 5 9x3!!xl!!
Cx = - — + — — + ^ ^ + ... = 0
1
[ 2 2!!x4!! 4!!x6!! (5.10.6)
f 1 5x1!! 9x3!! 13x5!!xl!! . A
Ci5 = -\ + + + ..1 = 0
3!!x4!! 2!!x6!! 4!!x8!! 6!!xl0!!

Extending the evaluation to all positive values of "n" then summing give, for
positive powers of a:
Photons 283

sJo.fl =5ȣ (5,0.7)

The reduction of Eq. 5.10.3 to this simple form is a unique property of

Eq. 5.5.15.
To examine the coefficients for negative powers of a, define the new
coefficients:

S22M/2)n<o = - E ^ - (5-10.8)
no;lCT

Combining Eq. 5.10.8 with Eq. 5.10.3 shows that:

_ L ( 2 l + l ) ( l - l ) ! ! ( / + n-2)!!

Evaluation of the first term gives:

Ci = 5L/8 (5.10.10)

With each succeeding increase in n, the power of L in the approximate

equality increases by two.
The total value of the sum is equal to the sum of Eqs. 5.10.8 and 5.10.9:

cosa ^ C„ , ^ („n)_ ,- s m C T j.,-v ( n ~V^n

>22 CT
'2 a
no;l O l CT
V ) no;l O
(5.10.11)

5.11 The Axial Fields, Summary

Sums of spherical Hankel functions on the +z-axis follow from the values for
spherical Bessel functions of Table 5.6.2 and the spherical Neumann
functions from Sections 5.7 through 5.10. The associated sums follow from
the regular sums and Eqs. 5.6.5:
284 The Electromagnetic Origin of Quantum Theory and Light

S!(a,0) = 0
, - L - l L+l
(L+n)!!
SzCa.O^e - * 1 1 - - (5.11.1)
2a) 2 ^(L-n+\y.\an
S 3 (a,0) = S 2 (a,0)
rL L+l (n-l)(L+n)!!
$ 2 (a,0) = e- / C T 1 + -
2a •V
2 If n 2(L-n+l)!!a
n+1
$3(rj,0) = $ 2 (a,0)

The remainders arise from the highest order mode only, and are valid if the
recursion relationship of Eq. 5.5.15 holds through order L and all higher
order have zero magnitude. Putting these sums into the field forms of
Eq. 5.6.6 and ignoring the remainders gives the electric field intensity on the
positive z-axis:

Er = 0
-'o„
E0 = 2e "~e (5.11.2)
EA=-/2e-/&e

The field is circularly polarized and the magnitude is independent of distance

from the source. The time average Poynting vector is:

N = - R e ( E x H * ) = - — Re(/ExE*)
(5.11.3)
N = £N z = - £
Tl

The axial power density is independent of distance from the source and
totally directed in the positive z-direction.
The remainders in the equatorial plane are quite different from those on
the z-axes; each mode contributes a proportionate share to the whole and
fields exist throughout the region. The sums are:
Photons 285

S 1 (a,rc/2) = a + / 2 - ^
ne;0CT

S2(a,7t/2) = l - - e - A , + /X- i : J a
vl O

S3(a.«/2) = -i-[l-e-*] + 5:%

° ne;2°

% (a..«).^[l-.-]
CT
+ ife^
no;3 <J
(5.11.4)

0
ne;2 °

The letter functions representing the remainders are:

i, (2£ + i)£V.(£ + n-2)\\

L ( 2 l + l)(l-2)!!(l + n-2)!!
(5.11.5)

L ( 2 / + l ) ( ^ - l ) ! ! ( / + n-2)!!

Substituting these results into the field forms gives the field:

E(o,n/2) = f*r
ne;0 ° (5.11.6)

[ne;2CT no;3 CT
J [ no;lCT ne;2 CT
J
286 The Electromagnetic Origin of Quantum Theory and Light

With O representing order, these fields produce the Poynting vector:

f <o
N(o,*i/2) = - | o P2 r+ -1 + 0 6+0 (5.11.7)
va ; va 2 ;

The radial term is an outbound power density proportional to B2/CJ . The

zenith angle term is a power density of magnitude independent of distance
and directed in the positive z-direction. It describes an energy flow from the
lower to the upper hemisphere.
On the negative z-axis, as it was on the positive one, the remainder is
from the highest order mode only and it does not contain contributions from
each mode. The other terms are equal to:

S2(a.n) = - £ e - * S 3 M = £e-to
(5.11.8)
$ 2 M = -^e-<° $3
ia
^=2fJe
Putting these sums into the field forms of Eq. 5.6.6 shows that the electric
field intensity and the Poynting vector on the negative z-axis are equal to
zero:

E(o,n) = 0 (5.11.9)

So long as the recursion relationship of Eq. 5.5.15 holds, any field that exists
on the negative z-axis arises from the remainder of the highest order
moment.
These results show that positive powers of the radial terms result in a
normalized power density of 4/r] along the positive z-axis, z-directed
uniform power density of l/r\ in the equator, and no power at all along the
negative z-axis. The first order, negative power terms show radially
outbound power in the equatorial plane. This result combines with energy
conservation to require energy that exits the generating source in the lower
hemisphere to pass upward through the equator. By Eq. 5.11.3, all energy
ultimately becomes positive z-directed.
Photons 287

5.12 Infinite Radius Radiation Pattern

The teledistant fields of Section 5.11 are coefficients times transcendental
terms; half of the terms arise from spherical Bessel functions and half from
spherical Neumann functions. Only spherical Neumann functions produce
near and far field terms. To extend the axial field expressions over all angles,
first note that the fields proportional to spherical Bessel functions are given
by Eqs. 5.3.1 and 5.6.1. In the limit of infinite radius, using Eq. A.24.16, the
spherical Bessel and Neumann functions satisfy the equality:

Lim , x dj/.(a)
y/(o) = —if-i- (5.12.1)
a=>°o do
The equality permits solving for the full angular expressions for the telefield
portion of the sums over spherical Neumann functions. Using Eq. 5.12.1 and
adding terms proportional to Bessel and Neumann functions gives the full
field over spherical Hankel functions. The functional form is:

Lim - , , ( d ~\ r i
E(o,e,<|>)= l + ' V - {Eq. 5.6.1} (5.12.2)
a=>°° v da;
The operation may be applied both to the field forms of Eqs. 5.6.1 and to the
sums of Table 5.6.1.
Operating on Eq. 5.6.1 shows that the limiting value of the total
teledistant field is:

E = e~^ c o s 8 (l + cose)(^sine+ecose-4); TiH=/E (5.12.3)

The result of operating on the sums of Table 5.6.1 is shown in Table 5.12.1.
The electric field intensity also follows from the definitions of Eqs. 5.6.3
through 5.6.6:

E r = e-*™ 30 sine(l + cose+ / / a ) e ^

E e = e-'CTCOS0 cos 9(1+ 0089)6-^ (5.12.4)
E ^ - z e ^ " ^ ^ ^ ^
288 The Electromagnetic Origin of Quantum Theory and Light

S1(a,e) =CTsinefl+ c o s 6 + - V ' a c o s e

S2(a,9) = e'
-J'CJCOSG,
l + cos6--l + • \e-i°cose_e-ta ](l+cos6)
v) asin 2 6 l
S3(a,0) = Vfe-'' C T C O s e -e-' a |(l + cos0)
asur8L J

$2(a,0) = ( -j'acosS cos 0(1 + cos 0) + — Q -i'acos9 cos0-e T

a asin z 0 hd ](l + cos0)

aCOS9cose e 1+cos9
$3(a,0) = Vk''
L
- ~''1(
J
)
asin 0

Table 5.12.1 Limiting infinite-radius solutions of sums over spherical Hankel

functions, keeping only radial terms with power minus one.
Listed values are correct only in the limit of infinite radius.

Comparison shows that the teledistant axial field of Section 5.11, obtained
by direct summation, is the same as that of Eq. 5.12.4.
These fields may also be used to construct phasor fields for absorption
since the complex conjugate of any electromagnetic phasor field is another
electromagnetic phasor field:

Energy emission
rsin0(l + cos6+//a)
E(a,8,(t)) = e- /i:Icose
(0cos9-/$)(l + cose)
(5.12.5)
Energy absorption
£sin6(l + cos6- //a)
E(a,9,4>) = e zocosQ »*
(§cose+4>)(l + cose)|

The fields are circularly polarized. The emission equations describe a wave
that exits its source at z = 0 and forms a fully z-directed plane wave that
travels to z = +00. The absorption, complex conjugate, equations describe an
oppositely directed plane wave at z = 00 that travels to a sink at z = 0. The
Poynting vectors are equal to:
Photons 289

N c = ± - ( l + cos6) 2 i (5.12.6)
r)

The upper and lower signs respectively apply for emission and absorption.
The radiation pattern is shown in Fig. 5.12.1. The figure is similar to
conventional radiation patterns in that the magnitude of the power density at
each angle is proportional to the distance from the origin. Unlike other
patterns, all energy flows in the direction of the pattern maximum. The result
is fully directed, z-oriented power with magnitude that is independent of the
distance from the source. It remains to be determined how such a condition
is consistent with energy conservation.

Figure 5.12.1 Radiation Pattern for Fully Directed Radiation.

Antenna is located at mid-point dot on bottom of the curve. Similar to other
radiation patterns field values at a particular angle are proportional to
indicated distance from the origin at the lower center of the pattern. Unlike
other field patterns, all power flows in the direction of the pattern maximum
and, at a specific angle, power density is independent of radius.

The absorption and emission discussions of Chapter 2 involve structures

with ideally conducting surfaces, for which, except for a biconical receiving
antenna, the absorbed power is zero. They respond linearly to the incoming
plane wave fields of Eqs. 2.1.11, and the powers are given by Eqs. 2.2.7
and 2.14.11. The extinction power is proportional to the first power of the
290 The Electromagnetic Origin of Quantum Theory and Light

scattering coefficients, ae and |3^, and the scattered power is proportional to

the square of the coefficients. The response of a nonlinear sink is quite
different. When triggered by an incoming wave an active sink does more
than absorb from the plane wave; in the manner described above it generates
near fields that extract energy from the full cross section of the field in the
inverse of emission. After the process has begun, the exterior fields at the
source are those of Eqs. 5.5.16, not Eqs. 2.1.11. Changing from emission
fields to absorption fields is the equivalent of taking the complex conjugate,
including changing from Hankel functions of the second kind to those of the
first kind.

5.13 Self-Consistent Field Analysis

An analytical method that extends the preceding results is the method of
self-consistent fields. Self-consistent fields are to electromagnetic fields
what a Taylor series expansion is to other mathematical functions. If the
fields are continuous through all orders, and if the solution is known exactly
at any point, information at that point may be used to construct the field at
any other point. For this case, the exact expressions are the telefield terms of
Eq. 5.12.4. The method uses continued application of the Maxwell curl
equations:

H=^VxE E = - —VxH (5.13.1)

k k
The electric field intensity of Eq. 5.12.4 is substituted into the first of
Eqs. 5.13.1, which gives a corrected expression for the magnetic field
intensity. The corrected magnetic field intensity is substituted into the
second of Eqs. 5.13.1, which gives a corrected expression for the electric
field intensity. The corrected electric field intensity combines with the first
of Eqs. 5.13.1 to yield a further corrected expression for the magnetic field
intensity that is, in turn, combined with the second of Eqs. 5.13.1 to yield a
further corrected value of the electric field intensity. With each step, the
solution becomes more exact and extends to smaller radii. Although labor
intensive, the iterative process may be continued as many times as desired to
obtain a satisfactory solution. There are inherent difficulties with this
technique. Since it only analyzes field symmetries present in the starting
Photons 291

fields, only those symmetries are present in the final one. Iterative errors
quickly accumulate if the starting fields are inexact.
Applying the technique to the spherical Bessel function terms simply
repeats the terms. The solution is exact and iterations produce no additional
terms. Bessel function terms are the first terms inside the brackets of
Eqs. 5.13.6-5.13.8; all other terms arise from spherical Neumann functions.
By Eq. 5.12.5 the beginning radial electric field component is:

E(°) = e-* , c 0 8 e sine(l + cose)e- J * (5.13.2)

Taking iterations begins the expansion at infinity and works towards smaller
radius solutions. After completing two full iterations the calculated radial
field component is:

sin8(l + cos8)+-sine(l + 3sin 2 e)

(2) _ -i'ccosG
+ — sin6cos6l9 + 6sin 2 e) -A(>
(5.13.3)

—^-sine(l8-30sin 2 e) + — sinGcosG

Continuing, after completing four full iterations and truncating with the
negative fourth power of a, the radial field component is:

sin0(l + cos6) + — sinGH + 7 sin2 0J

+ —sin9 cos e(21 + 54sin2 0)

p M = -iacos9
e-® (5.13.4)
—'- sin 0(174 - 30 sin2 0-300 sin4 0)

—-sin0 cos0(642 - 1620sin2 0 - 840 sin4 0)

292 The Electromagnetic Origin of Quantum Theory and Light

After eight full iterations, and truncated with the negative fourth power of a:

2
sin 0(l + cos0) + - s i n e ( l + 15sin e)

+ — sin 6 cos e(45 + 294 sin2 6

a2 v )
v(8) _ -I'acosO .-*
Er ' = e
— ' - sin e(966 + 2898 sin2 0-5460 sin4 0
)

— sinGcos 0(l4152+ 12462sin 2 0-88O8Osin 4 0)

L a
(5.13.5)

The foregoing series are sufficient to permit generalization to an

arbitrary number of iterations. After n iterations, the three field components
generalize to those of Eqs. 5.13.6 through 5.13.8. The equations are
expressed as a function of the number of iterations used in the self-consistent
field solution. To use the solution, it is necessary to obtain the same
expressions as a function of a known physical quantity, for example the
number of the largest numbered mode. Establishing a relationship between
the fields and those obtained by direct summation is mostly easily done
using Eq. 5.13.6.

l + cos9+-(l + [2n-l]sin 2 e)

(3[2n-l] + 6[n-l] 2 sin 2 eJ

2[l0n 2 -21n+ll]+2[8n 3 -51n 2 + 82n-39]sin 2 8

E^sinee-*0089 -10[2n 3 -9n 2 + 13n-6Jsin 4 e

[224n3 -368n 2 + 1672n- 792]

81n 4 -3750n 3 + 41529n2

-cos6 + sin 2 0
4a" -166950n+233400
-[300n4 - 2760n3 +11460n2 - 29160n + 3648o]sin4 0
(5.13.6)
Photons 293

cos 8 + cos2 8 + — [2n -1] sin 2 0cos 8

a
+ -V(2[n -1] + 6[n - 1 ] 2 sin2 9cos2 e)

'2[4n - 5][n - 1] + 8[2n3 - 9n 2 + 13n - 6]sin 2 0^

T-C0S6
c(n) _ - i a c o s B
Eg - c -,
O -10[2n3-9n2+13n-6]sin48 ,-*

[96n 3 -432n 2 + 648n-312]

+[210n4 - 2524n3 +13254n2 - 37292n +44352 ]sin 2 8
4 3 2 4
4aH -[l99n +2070n -37339n +162750n -232200 ]sin 8
+[300n4 - 2760n3 +11460n2 - 29160n +3648o]sin6 8
(5.13.7)

1 + cos6 + — [2n - 1] sin2 9

a
+ -^-cose([2n - 3] + 6[n - l] 2 sin2 d\

[8n2 - 26n +23] + [l6n3 - 84n2 + 140n - 78]sin2 6^

e(n) _ _ .-rtjcosS - i o [ 2 n 3 - 9 n 2 + 13n-6]sin 4 0 -*

-[96n3 - 576n2 +1200n - 864 ]

220n 4 -3080n 3 +20060n 2
rCOs8 - sin2 9
40* L- 7 0 9 6 0 n + 1 0 4 -400
+[300n4 - 2760n3 +11460n2 - 29160n +3648o]sin4 9
(5.13.8)

Equating terms proportional to the inverse radius yields the approximate

equality:

2/ / T
—n=— L (5.13.9)
CT 2a

The left side of Eq. 5.13.9 comes from Eq. 5.13.6 and the right side from
Eqs. 5.11.6. The equations are equal if:
294 The Electromagnetic Origin of Quantum Theory and Light

L = 4n (5.13.10)

Other term-by-term comparisons between the "L" and "n" expressions yield
a similar relationship. Therefore, for equivalent descriptions the number of
self-consistent field iterations is one-fourth the maximum modal number of
the radiation. This relationship permits writing the full angular dependence
of the fields, Eq. 5.13.6-5.13.8, as functions of the maximum modal number.
Incorporating Eq. 5.13.10 into Eqs. 5.13.6, 5.13.7, and 5.13.8, and
keeping only the leading terms shows the approximate field set through
inverse quartic terms to be:

L/ 3 3L 3
sin8 + sin0cosG + — sin 0 + — -Zs i n 6cos9
Er(o,e,(t>) = e- I C T C O S B 2a 8CT
,3
L4 / \
—r-sin3e(l-5cos2e)- -sin 3 9 cos0181 -300sin 2 8
16o3 v
' 1,024CT H

cos8+cos 0+ — sin 0cosG+—^-sin 0cos 0

2a Saz

Ee(o,0,(|))= e -i'acos9 + -^—rsin

3
2
0cos0(l- 5cos2 0 -At>
16o ^

21Osin20-199sin40 + 3OOsin60)
l,024o^

/L •> 3L 7
l + cos6+—sin 8+—r-sin 9cos8
2o 8az
^ ( c e , ^ ) = -('e -itJcosG -/it.

+-^Tsin2efl-5cos2e) ^- r sin 2 8cose(220-300sin 2 e)

16a 3 " ' 1,024a4 " '.

nH=/E (5.13.11)

The first term in each field component arises from spherical Bessel
functions, all other terms arise from spherical Neumann functions. Although
the transcendental propagation term is correct only if L is infinite, in all of
these cases with positive powers of a the series converge so rapidly that L as
large as ten appears to be adequate to describe source behavior.
Photons 295

5.14 Power and Energy Exchange

To obtain the power and energy in a photon field, begin with the complex
Poynting vector. With the fields of Eq. 5.13.11, it may be written as:

Nc = - R e ( E x H * ) = — - R e f / E x E * ) (5.14.1)

The vector components are:

N c = - R e { ^ / E ^ E e ) + e(/E r Ej) + 0(/EeE*)} (5.14.2)

The surface power on a circumscribing sphere surrounding the source

follows from the angular components of Eqs. 5.13.11 and 5.14.2.
The real part of the radial component of the Poynting vector for L large,
through quartic terms, is:

3L2 L4 r l
nNr = 2cos 0+—Tsin20cos20
2
Tsin
2
6^430-719sin 2 0+600 sin 4 0
4a2 1,024a 4 [ J

(l + cos20) + ^ s i n 2 0 ( l + 2 c o s 2 0 W - ^ s i n 4 0 ( 4 - l l c o s 2 0 )
+ COS0 4a z v
' 64CTH

4
- sin 2 ©[430 - 719 sin 2 0 + 600 sin 4 ©]
1024O-
(5.14.3)

In common with the output powers calculated in Chapters 2 and 3, only

products of one Bessel and one Neumann function term integrate to a
nonzero value over an enclosing surface. Since all Bessel function terms are
teledistant, each surface power term contains at least one teledistant field
term. If a source generates the fields of Eq. 5.13.11 without the Bessel terms,
the result is standing energy. A plane wave impressed on such a field
supplies the leading terms of Eqs. 5.13.11, and may produce an energy
exchange.
The power on the surface of a virtual sphere of radius a/k calculates to
be:
296 The Electromagnetic Origin of Quantum Theory and Light

471 2CT2 L2 1553 L4

Pr = - + 10 8960 2 (5.14.4)
r\/t* a

All terms of Eqs. 5.14.3 and 5.14.4 are proportional to the teledistant Bessel
function. The first term is also proportional to the teledistant Neumann term,
the second term to the inverse square Neumann term, and the third term to
the inverse quartic Neumann term.
In Chapters 2 and 3, the time-average surface power is supported by the
product of field terms each of which is proportional to 1/a, for example
2
Eqs. 3.3.1 and 3.3.2. Since the area increases as a , the product of far field
power density and area is independent of distance. The radiation analyzed
here is dramatically different. The power from the first term of Eq. 5.14.4
2
increases with distance as a , the second term is independent of a, the third
2
term decreases as 1/a etc., through higher powers. Yet, energy conservation
requires the total value to be independent of distance. This, in turn, requires
the energies carried by the higher order terms to transfer to lower order terms
as the energy travels outward from the source.
It follows from Eqs. 5.13.11 and 5.14.2 that, through inverse quartic
terms, the azimuth-directed portion of the Poynting vector is equal to:

sin0(l + cos9) 2 + — 2 s i n 3 e ( 1 + 3cos0+2cos2e

)
TiN0=-
1/ T,
+ 7 sin 8 4 - l l c o s 2 9 cos6(l + cose)(301-600cos 2 e)
4
[ 64a
(5.14.5)

At the equator, this vector component is +z-directed and of magnitude:

L2 L4
-r|Ne = r|Nz= 1 + ^ - + j (5.14.6)
4a 2 16a 4

Equation 5.14.6 shows that energy passes from the lower into the upper
hemisphere. This is consistent with energy conservation only if energy that
exits the source into the lower hemisphere veers and passes upward.
Photons 297

5.15 The Wave Train

The preceding analysis considered only steady state fields, fields that have
existed since time t = -co. The wave train, therefore, is infinitely long.
Actual wave trains, of course, are finite in length; generating sources start
and stop. One estimate of the length of a photon wave train is given in
Section 5.2. With a finite wave train, calculated kinematic results are
meaningful only for fields that are contained within a sphere, centered on the
active region, with a radius equal to the length of the wave train. For a wave
train of length /, if W is the total radiated energy, F is the normalizing field
constant, Eq. 5.14.4 is the expression for the output power, and letting O
indicate order or the variable, Wis equal to:

1553 V
W= +0 (5.15.1)
8960 (Mf

A separate expression for the total output energy follows from

Eq. 5.5.17:

(5.15.2)
ri*2 £ , I T\k2 I
Keeping only the lead term of Eq. 5.15.1 and requiring Eqs. 5.15.1
and 5.15.2 to be equal shows that:

(>f/)2 = 1.5L2 (5.15.3)

This result shows that the normalized wave train length, kl, is on the same
order of magnitude as the maximum modal number of the radiation. For
example, if L were equal to 1000 the wave train would equal about:

/=200A (5.15.4)

The estimated length of Eq. 5.15.4 is much less than that obtained for a
dipole field generated by a point electron, see Sec. 5.2. However that
estimate is based upon the electron as a point charge and oscillations were
298 The Electromagnetic Origin of Quantum Theory and Light

produced by oscillations of the entire electronic mass. In this case, the

oscillations are within the structure of the electron.
To estimate F, note that Eq. 5.15.2 is also equal to h(0. Making the
equality and, to form a definite model, letting F be a constant magnitude
rectangular pulse of width Ax, and equal to zero at all other times:

3fi
(5.15.5)
4eX/3

Energy passing through the equator between radius a, the limit of the
active region, and the length / of the wave train may be obtained by
integrating the Poynting vector of Eq. 5.14.6 over the appropriate portion of
the equatorial surface. The result is:

kl
271
f d/ F 2 — 2
W^= J , L2
T f crda 1 + — - +
L4 j + Or — ! ^
A/ ^ I 4a 2
16a 4
[a {
,
(5.15.6)

271 Lz V
Jd/F2 4
-+ 0
(^) 80(/fo): 80(/fo)

Since the result is positive, energy passes from the lower to the upper
hemisphere.
With a wave train containing 200 or more wavelengths, field effects are
dominated by the steady state solutions and emission is not just a transient
act. The field equations of Eq. 5.13.11 confirm that as the radius increases
the energy paths veer towards the z-direction. When the source is terminated
the outermost portion of the wave train is nearly collimated and the ratio of
transported energy to transported momentum is only slightly greater than c.
Since the transient solution is not available, whether the transient fields
change the near equality with c to an exact one is unknown.
Photons 299

5.16 Multipolar Moments

The purpose of this section is to describe a means by which high-order
modes are generated from an initial charge density. Although the magnitudes
of the high-order fields are too small to produce the coefficients needed to
produce the fields of Eqs. 5.13.11, the needed source symmetry is formed
that is, in turn, driven by forces, to be described in the next section, that
produce high order terms of significant magnitude.
When a passive, perfectly conducting sphere is immersed in a plane
wave, at the surface of the scatterer the magnitudes of the scattered and
incident waves are the same, see Eqs. 2.3.7. The equality requires the ratio of
field coefficients to be:

Incident field magnitude ikd)

& v
Scattered field magnitude '
(it + l)\l(2t - l)!! (5.16.1)

With (ka)« 1 the ratio of Eq. 5.16.1 is so small and decreases so rapidly
with increasing I that the lowest order terms dominate scattering.
When an atom in a high energy state is immersed in a plane wave, at the
surface of the scatterer the magnitude of the scattered wave is much larger
than that of the incident wave. For the incident wave coefficients of Eq. 5.3.1
to produce the scattered wave coefficients of Eq. 5.3.2, it is necessary that
the ratio of incident to scattered field magnitude be:

Incident field magnitude

& ill + \)\\{2£ - l)!!
~v / v
2i+1
• (5.16.2)
Scattered field magnitude (£a)

With (ka)« 1 the ratio of Eq. 5.16.2 is so large and increases so rapidly
with increasing t that the highest order terms dominate emission.
Changing the coefficients Eq. 5.16.1 to those of 5.16.2 is only possible
because electron-generated radiation involves much more than scattering.
The analysis involves atomic states in unstable equilibrium and radiation
reaction forces that far exceed those due to the incident plane wave. To
examine a possible source of the driving fields consider an occupied state
that supports time-average current and charge densities at all points within
the source region. By Eq. 4.5.1 and Eq. 4.5.2 the time-average values of
charge and current densities are respectively eU*(r)\j(r) and
300 The Electromagnetic Origin of Quantum Theory and Light

— U*(/-)VU(/0-U(r)VU*(/0 With biconical receiving antennas, see

linv-
Section 2.14, by Lenz's law the time changing magnetic field induces
currents that generate an opposing field. Quite differently, an intrinsic
magnetic moment interacts with the applied magnetic field to generate a
field in the direction of the incoming magnetic field.

Figure 5.16.1 Source modes resultingfroman electric dipole formed within an

electric charge distribution.
Arrows indicate current densities. Current {1} creates a dipole source,
Currents {1} and {2} together create an electric octupole source, and
Currents {3} create a magnetic quadrupole source. By continuation, this
creates all odd numbered TM field sources and all even number TE field
sources. With a dual source of a magnetic charge distribution, the result is
creation of all even numbered TM field sources and all odd numbered TE
field sources.
Consider a sphere of charge in which the condition V • J = 0 applies. In
Fig. 5.16.1, an applied field drives a dipole mode, as illustrated by the center
arrow. Two current options are illustrated by the smaller arrows. Some
terminate on the edge of the source, leaving a net charge density that
generates an electric octupole field. The interior charge structure drives
continuous current loops that generate a magnetic quadrupole field. So long
as source constraints permit the charge density to be further subdivided in
this way, each current path produces a similar set of higher order modes. The
Photons 301

result is that an oscillating electric dipole moment drives all odd-order

electric multipolar moments, TM fields, of the same degree and all even-
order magnetic multipolar moments, TE fields, of the same degree.
In a dual manner, using magnetic circuit techniques and an effective
magnetic current density, a similar analysis shows that an oscillating
magnetic dipole moment drives all even-order TM fields, and all odd-order
TE fields of the same degree.
As shown by the multipolar field coefficients for quantized radiation, the
field coefficients are proportional to the generating charge times ka raised to
an integer power. The method of calculating moments from source
distributions is discussed in the Appendix, Sections A.28 and A.29. Results
are that the electric and magnetic multipolar moments of order £ are
respectively proportional to (ka) and (ka)
Changing from individual charges to charge density permits changing
calculation techniques from sums over individual charges to volume
integrals over charge distributions. The former is much more difficult than
the latter. Electric charge density may be used in local neighborhoods where
the ratio of contained charge to the volume of containment is continuous.
The discussion associated with Fig. 5.16.1 shows no lower limit to the
possible size of current eddies. However, the approach is valid if and only if
the charge density is continuous in the neighborhood of a source point, see
Section 1.5. We conclude, therefore, that the above arguments apply down to
a scale of dimensions on which the use of charge density to determine
properties breaks down. Since the order of a radiated mode varies inversely
with the physical dimensions of current eddies, the maximum modal
number, L, is determined by lower limit on the physical size of current
eddies. L is determined by the granularity of the electron charge distribution.
(L-l) 999
A fixed array requires 2 units; if L = 1000 then 2 units are necessary.
Yet a single unit that moves rapidly enough can generate the same moment.
A necessary axiom is that the moment is enabled by electron nonlocality.
An electrically small scatterer driven by the field of a plane wave was
treated in Chapter 2, where it was shown that high-order induced moments
and high-order scattering are relatively small. In contrast, the induced
sources produce much larger high-order moments. This approach also shows
an intrinsic upper limit, L, to the maximum modal number that depends upon
302 The Electromagnetic Origin of Quantum Theory and Light

the dimensional scale of the smallest units of charge and current

distributions. If the charge density is continuous ever-larger orders will be
driven. The process ceases only when the dimensional scale is so small that
granules of charge appear as a three-dimensional mosaic.

5.17 Field Stress on the Active Region

As discussed in Sections 1.8, 3.13 and 3.17, Maxwell stress tensor fields
accompany electromagnetic fields. An external field intensity that is parallel
with or normal to a surface exerts a pressure on that surface directed
respectively towards or away from the field. In spherical coordinates, the
radial portion of the diagonal components of the stress tensor is:

T„ = | ( E ? - E § - E J ) + ^ ( H ? - H § - H J ) (5.17.1)

If a field exists on one side of the surface, and if there are no fields on the
other side, a positive or negative value of Tn- denotes respectively a positive
or negative radiation reaction pressure.
To establish stress magnitudes associated with commonly occurring
phenomena, consider the solar radiation field on the surface of the earth. The
intensity at the equator with a dry atmosphere and the sum directly overhead
is about one kW per square meter. The frequency spread is infrared to
ultraviolet and it is randomly polarized. Approximate values of electric field
intensity and pressure are:

E = 27.5 V / m and - E 2 = 3.33 x 10 - 9 Pa (5.17.2)

2
Consider next the pressure on a spherical shell of radius a that contains a
uniformly distributed total inner surface charge of -e centered about a point
charge of +e. By Coulomb's law the electric field intensity just inside the
shell is:

Er = — ^ (5.17.3)
Ama

Using Eq. 5.17.1, the Coulomb pressure is:

Photons 303

T =
J (5.17.4)
2 4
1
rr 32n ea
This equation shows that the pressure varies as the inverse fourth power of
the radius of the radiating shell. The positive sign indicates that the pressure
is directed towards the field, in this case inwardly. If the radius is 'n' times
the radius of the first Bohr radius, 5.29 x 10-11 m, the inward pressure is

T1rr =
— 4 Pa (5.17.5)
u
Next, let the shell contain no sources or fields, but the exterior surface
supports the surface charge and current densities that generate the fields of
Eqs. 5.13.6 through 5.13.8. The total radiation reaction pressure produced by
these fields is constant since, like Eqs. 3.17.3 the electric and magnetic fields
are sized and phased in a way that the time-dependent portions cancel.
However, by Sec. 4.3 the source contains both charge and current densities
and electric fields effect charge densities and magnetic fields effect current
densities. We define the phase of transcendental functions on the surface of a
spherical shell of radius a to be O where:

<D = 2[co/- kacos 8 - <\>] (5.17.6)

Matrix element T^ calculated from Eqs. 5.13.6 through 5.13.8 is:

3V
2cos20(l + cos6)2 + sin 2 6cos 3 6(l + cos6) + -sin4Gcos4e
W
T
—F2 + rsin49cos2efl -5cos 2 0 )
^ka)"
L4
-sin 2 0cose(l + cos6)(5 - lOsin^ - SOOsinM
256(/fo)
304 The Electromagnetic Origin of Quantum Theory and Light

L
\2 .: 6 , 3L .4
2sin 9(l + cosO) -sin n0 + -sin 0cos8(l + cosO)
2{ka) 2{kaf

+v
9V
4
sin6 0 cos2 0 sin6 of(l-5cos 2
1 - 5 cos2 e)6 cosO
32(/ta) 4
8(/fa)

sin2 6cose(l + cos0)(5- lOsin2 0 -300 sin4 o\

256{kay
2L 4O(l + cos0)+ 3L
7^rsin
2
sin60cos0
+iF sinO
4 2
-sin 0(l + cos0)(l-5cos e)
4(ia}
(5.17.7)

The azimuth angle dependence shows that the tensor rotates around the z-
axis twice each field cycle.
2
The magnitude of coefficient eF 12 follows from Eq. 5.15.4 and
Eq. 5.15.5. In the mid-optical range, X = 500 nm, it is:

,2
eF"
= 2 . 3 7 x l 0 - s Pa (5.17.8)

Combining Eqs. 5.17.7 and Eq. 5.17.8 shows the positive z-axis pressure to
be constant, compressive, and more than 100 times the solar pressure at all
radii. There is no pressure on the negative z-axis.
7
Tff(a,0,<l>) = -3.80x10 Pa andTrr(a,7t,(l)) = 0 (5.17.9)

Figure 5.17.1 shows a plot of total radiation pressure versus zenith angle
at the surface of a virtual radiating sphere with radius equal to 1,500,000
Bohr radii, where L/(ka) = 1. This is a radius approximately equal to the
length of the wave train, and here the teledistant field terms are dominant.
The maximum pressure magnitude is on the positive z-axis and equal to
-3.8 x 10 Pa. This pressure is the mechanism by which the radiation exerts
a backward force on the radiator, in accordance with Newton's second law
of motion.
Photons 305

5
°- | ! 1 ! ! ! ! ! ! 1

0- i \ ':•... .^yr^^^i::^:^^ ;

-0.5 \ : .'/.. ; --: \ ; : -..-

Trr "1-5 \ \i J | ! \ |
-2- ; / ; \ : ; ; i -

-2.5 i /..; ; i i i | ; -.-

-3 \.-L : ; : : \ ;. I.-..-

-3.5 -.. •/'•, i .; : i -

-41 i i i i i i i i
0 20 40 60 80 100 120 140 160 180
6 (Degrees)

Figure 5.17.1. U(ka) = 1, total pressure versus zenith angle.

Total pressure on the surface of a radiating sphere with radius equal
1,500,000 Bohr orbits at it generates the fields ofEqs. 5.13.6 through 5.13.8.
—13
At this radius the Coulomb pressure is about 2.3 x 10 Pa. The
radiation pressure is greater than the Coulomb pressure by a factor of about
160,000 and greater than the maximum sunlight pressure by a factor of about
10. We conclude that the radiation reaction pressure dominates the dynamics
of electron behavior at this radius.
A Mercator projection of the surface pressure showing pressure in the
vertical dimension, but limited to half the full azimuth angle for better
viewing, is shown in Fig. 5.17.2.
306 The Electromagnetic Origin of Quantum Theory and Light

Figure 5.17.2. L/(ka) = 1, partial three-dimensional view of total pressure versus

zenith angle.
A third dimensional Mercator projection of total radiation pressure on the
surface of a radiating sphere with radius equal 1,500,000 Bohr orbits due to
the fields ofEqs. 5.13.6 through 5.13.8. Negative values are compressive.

Figure 5.17.3 shows a plot of only the electric field portion of

Eqs. 5.17.6 to 5.17.8. Both the field and this portion of the stress tensor are
time dependent. The maximum negative electric pressure is the z-axis value
_7
of-1.9 x 10 Pa. The maximum positive pressure is at 8 = 67°; it is about
half the maximum negative pressure, and oscillates between equal values of
compression and expansion. This pressure acts only on electric charge
densities, which it drives as part of the coherent generated field.
Figure 5.17.4 shows a three dimensional plot of the pressure magnitudes
calculated using Eq. 5.17.7. The upper lobe is compressive and the two
lower lobes are expansive.
Photons 307

xlO"
I i Msr^i I

Trr

-10

o o=o
-15 A o = 45
• o = 90
• o = 135
0 0=180
-20 i i

0 20 40 60 80 100 120 140 160 180

6 (Degrees)

Figure 5.17.3 L/(ka) = 1, electric pressure versus zenith angle.

Effects on the surface of a radiating sphere with radius equal 1,500,000
Bohr orbits supporting the fields of Eqs. 5.13.6 through 5.13.8. Maximum
magnitude is on the positive z-axis. Phase variation with zenith angle is
suppressed.

Although the self consistent field calculations of Sec. 5.13 were

truncated at the inverse fourth power of the radius, it is clear from the
calculations that continued expansion produces polynomials consisting of
2s
trigonometric functions multiplied by (LIka) , where s is a positive integer.
It is also clear that the multiplying coefficients increase rapidly with
increasing values of s. The series is of the same order as Eq. 5.16.2. Near the
outer edge of the original wave train, where the L/ka ratio is on the order of
one, only the teledistant fields are significantly large. Moving inward, as the
radius decreases terms with larger values of s become dominant until the
actual radius of the radiating region is reached. The largest value of s for
which we have complete information is four.
308 The Electromagnetic Origin of Quantum Theory and Light

All axes x 10

15-

10

5-

Figure 5.17.4 L/(ka) = 1, a three dimensional plot showing the electrical portion of
the same stress tensor magnitude. Phase variation with zenith angle is suppressed.

For the special case s = 2, Figure 5.17.5 shows the total radiation
pressure versus zenith angle at the surface of a virtual radiating sphere with a
radius of just one Bohr radius. The maximum pressure magnitude of
9.0 x 1016 Pa is expansive and occurs at zenith angle 67°. The magnitude so
dominates the Coulomb force that the size of the radiating eigenstate
electron surely undergoes a nova-like expansion of the upper hemisphere
once the radiation process begins. Indeed, the pressure is so large that it
belies the use of perturbation techniques to calculate radiation effects, as
done in Sec. 4.13 and 4.17. The maximum compression is -4.2 x 10 Pa
and occurs at a zenith angle of 112°. Although expansion of the upper
hemisphere is expected, compression is against other atomic forces and
hence distortion is less probable.
Photons 309

16
,x_10

i i ! ! i i i i
40 80 100 120 140 180

6 (Degrees)

Figure 5.17.5. L/(ka) = 1 5 x 1 0 , total pressure versus zenith angle.

Total pressure on the surface of a radiating sphere with radius equal one
Bohr orbit supporting the fields ofEqs. 5.13.6 through 5.13.8.

Calculated ratios of radiation-to-Coulomb pressure and radiation-to-solar

pressure are listed in Table 5.17.1. At 9 = 67° the radiation pressure is about
25
80,000 times the Coulomb pressure and 10 times that of sunlight. The first
ratio shows that the Coulomb force is not significant and the second ratio
shows sunlight pressure to be trivial.

Radius is Pressure ratio Pressure Ratio

one Bohr Field/Coulomb Field/sunlight
radius
67° 77,000 24
27x10
112° -38,000 24
-1.4x10

Table 5.17.1 Ratio of radiation pressure to Coulomb pressure and to sunlight at a

radius of one Bohr orbit.
310 The Electromagnetic Origin of Quantum Theory and Light

Figure 5.17.6 L/(ka) = 1 5 x 1 0 , Mercator projection of total pressure versus zenith

angle.
A third dimensional Mercator projection of the total radiation pressure on
the surface of a radiating sphere with a radius of one Bohr radius due to the
fields ofEqs. 5.13.6 through 5.13.8.

A Mercator projection of the radiation pressure is illustrated in

Fig. 5.17.6. As was the case with Fig. 5.17.2, the z-axis indicates both
magnitude and direction. Positive values represent pressure in the direction
of the radiation and vice versa.
Figure 5.17.7 shows a plot of electric radiation pressure versus zenith
angle at the surface of a radiating virtual sphere the size of the first Bohr
radius; at this radius L/ka equals the number of wavelengths in a wave train.
The maximum swing in positive pressure is between zero and 9.7 x 10 Pa
and occurs at a zenith angle of 67°. The maximum swing in negative
pressure is between zero and -4.2 x 10 Pa and occurs at zenith angle 112°.
Photons 311

0 20 40 60 80 100 120 140 160 180

9 (Degrees)

Figure 5.17.7 L/(ka) = 1.5 X 10 , electric pressure versus zenith angle.

Effects on the surface of a radiating sphere with radius equal one Bohr orbit
supporting the fields of Eqs. 5.13.6 through 5.13.8. Maximum pressure is
about 9.1 x 10 Pa. Phase variation with zenith angle is suppressed.

Figure 5.17.8 is a three dimensional plot of the electrical pressure

magnitude at one Bohr radius. The upper and lower lobes represent
respectively positive and negative pressure. Lobes on either side are identical
and result from the 2§ angular dependence of the pressure.
To illustrate the importance of the radiation reaction force it was
necessary to pick a specific set of examples. Results, however, do not
depend upon the particular set chosen. The large radiation reaction pressure-
to-Coulomb pressure ratios are robust, and similar results are obtained for
any reasonable choice of examples.
It is important to note that the physical origin of the radiation reaction
pressure is the standing energy fields that accompany energy exchanges by
electrically small radiators. The character and magnitude of the terms make
them the primary driving mechanism for quantized transitions. As noted in
312 The Electromagnetic Origin of Quantum Theory and Light

earlier sections, although a plane wave contains all needed field symmetries
and phases, the input magnitudes of high-order modes are far too small to
support the multipolar coefficients of photon radiation. Quite differently, the
radiation reaction pressure of the reactive fields meets all requirements. We
conclude that atomic transitions consist of a regeneratively driven, nonlocal,
eigenstate electron occupying a turbulent region of space that, in the upper
hemisphere, may be orders of magnitude larger than the usual size of nascent
atoms.
Detailed three-dimensional, time-dependent plots of power, energy, and
electromagnetic stress in the vicinity of an atom as it radiates a photon are
maintained on website:
<http://www.ee.psu.edu/grimes/antennas/breakthrough.htm>

All axes x 10

Figure 5.17.8 U(ka) = 1.5 x 10 , three dimensional plot of electric pressure versus
zenith angle. Phase variation with zenith angle is suppressed.
Effects on the surface of a radiating sphere with radius equal one Bohr orbit
supporting the fields of Eqs. 5.13.6 through 5.13.8. Maximum pressure is
about 9.7 xlO16 Pa.
Photons 313

5.18 Summary
Radiation onset by a point-electron as it enters an eigenstate generates a
standing energy field. That energy field produces an expansive force on the
electron in proportion to the inverse square of the electron size, see Sec. 4.2.
This is a large force that has not been considered; we postulate that it
transforms the electron into an extended, nonradiating, eigenstate electron in
dynamic but stable equilibrium. Such an eigenstate electron is an ensemble
of evolving charge and current densities. It signifies that the physical
significance of e\J {r)\J{r) lies somewhere between a static charge density
and the probability density of a point charge. The electron structure
continuously evolves, in response to intrinsic and local forces. For example,
interaction between the electron and orbital magnetic moments results in a
continuous torque on what might otherwise be a fixed orbital circuit of
current.
Energy conservation is applied to the ensemble and results in the
Schrodinger time-independent equation, see Sec. 4.3. The method of
deriving the equation is similar to methods of thermodynamics in that energy
conservation yields ensemble properties although details of a particular
ensemble unit are not known.
Emission from a high-energy eigenstate begins after the structure
evolves to that needed to generate TE and TM dipole radiation fields with
the orientation and phase of Table 3.17.1. Since evolution of the electronic
picostructure is not instantaneous, a time delay is expected after the
application of an external field before any particular atom ejects a photon.
The duration of the delay depends upon the initial structure of that electron
and hence is statistical in nature. At some point, the ensemble forms the
structure necessary to begin the radiation process. Once begun, the radiation
reaction force increases rapidly with modal order and drives all source
modes regeneratively. The resulting electromagnetic fields produce the
standing energy field of Section 3.17, in which all near field terms are
proportional to spherical Neumann functions. When superimposed with an
incoming plane wave, which consists solely of spherical Bessel function
terms, photon emission occurs with no further time delay.
Full directivity is achieved with spherical modes. A three dimensional
plot of the radiation reaction pressure on a rigid, radiating of radius a at the
phase of maximum extensive pressure is shown in Fig. 5.18.1. This
extensive pressure is thousands of times larger than the Coulomb attraction
pressure.
314 The Electromagnetic Origin of Quantum Theory and Light

Figure 5.18.1 Electric radiation reaction pressure on a rigid radiating surface.

The radius of the radiating surface is at mid-range and phased for maximum
upper hemisphere extension. The upper and lower hemispheres show
respectively expansvie and compressive pressures. Phase variation with
zenith angle is suppressed.

Since source build up is nonlinear, the Manley Rowe equations apply

and Eqs. 5.1.14 and 5.1.17 correctly predict the observed power-frequency
relationships. The nonlinearity voids any possibility of describing the
process using equilibrium equations, such as Schrodinger time-independent
equation, Eq. 4.3.14, or the Dirac equations; no linear equation can describe
such nonlinear events. The time-dependent equation, Eq. 4.5.10, describes
near-equilibrium characteristics that are applicable over times long
compared with rates of change of the electronic ensemble. It describes
events leading up to the transition, and it describes events after the transition,
but it does not describe events that occur during the hiatus between periods
of equilibrium. A necessary axiom is that the structure of high order
multimodal moments is enabled by electron nonlocality.
If energy absorption by an atom was a linear process, see Section 1.13,
the incoming energy would be statistically distributed over all available
eigenstates. There would be more energy in the lower energy states and less
energy in the higher energy states, but all would contain some energy. With
Photons 315

a system that responds nonlinearly, the regenerative process puts energy into
a single eigenstate; there is no mixing of nondegenerate atomic states. The
probability of an atom being in a single, nondegenerate entry state is one,
there is no mixture of states, and hence no wave function collapse during a
measurement process.

References
R. Becker, Electromagnetic Fields and Interactions, trans, by F. Sauter, Dover
Publications (1964)
D.M. Grimes, C.A. Grimes, "A Classical Field Theory Explanation of Photons," in
T.W. Barrett, D.M. Grimes, Advanced Electromagnetism: Foundations, Theory
and Applications, World Scientific (1995) pp. 250-277
R.F. Harrington, "Effect of Antenna Size on Gain, Bandwidth, and Efficiency," J.
Res. National Bureau of Standards vol. 64D, 1-12 (1960)
J.D. Jackson, Classical Electrodynamics, 2nd ed., John Wiley (1975)
L.D. Landau, E.M. Lifshitz, The Classical Theory of Fields, Addison-Wesley
(1951), trans. By M. Hamermesh
J.M. Manley, H.E. Rowe, "Some general properties of nonlinear elements, Part I:
General energy relations," Proc. IRE, vol. 44, pp. 904-913 (1956), and "General
energy relations," Proc. IRE vol. 47, pp. 2115-2116 (1959)
W.K.H. Panofsky, M. Phillips, Classical Electricity and Magnetism, 2 nd ed.,
Addison-Wesley (1962)
W.R. Smythe, Static and Dynamic Electricity, 3rd ed., McGraw-Hill (1968)
J.A. Stratton, Electromagnetic Theory, McGraw-Hill (1941)
This page is intentionally left blank
 6. Epilogue

6.1 Historic Background

Particles have played a critical role in the analysis of physical events since
Newton (Newton, 1687) examined them in the latter part of the seventeenth
century. For the next several centuries studies involving combinations of
elastic spheres dominated physics. Consequently electrical problems were
first interpreted by analyzing the behavior of charged particles; Maxwell
(Maxwell, 1891) used the force between "two very small bodies" to discuss
implications of Coulomb's law. Although Maxwell's equations showed that
fields are essential to explain occurrences that particles alone cannot, still the
lore of particles permeated physics at the end of the nineteenth century.
Therefore after Thompson (Thompson, 1897) discovered and measured
particle-like electrons the idea of electrons as particles was widely accepted,
along with the Lorentz (Lorentz, 1909) particle-electron and, later, the Bohr
(Bohr, 1913) atomic model. The finely honed and widespread skills of
classical mechanics were carried over to quantum effects; even the name
"quantum mechanics" is indicative of such an origin.
Building upon his earlier work with particles, Newton (Newton, 1704)
compared the propagation of light with that of projectiles. According to him,
luminous bodies eject light-making projectiles that continue in flight until
acted upon by other objects. Quite differently, Huygens (Huygens, 1678)
compared the propagation of light with the propagation of sound through air
and waves on water. According to him, light was not a thing but a
disturbance that propagated through space. Light is emitted over a spread of
angles by a luminous object. It bounces off objects and the total of all
bounces off all objects that is intercepted by an observer forms his field of
view. After Huygens, more than a century passed before Young in England
and, in an arguably unparalleled technical outpouring, Fresnel in France

317
318 The Electromagnetic Origin of Quantum Theory and Light

confirmed that light is propagated as a wave with transverse vibrations and,

therefore, two polarizations, (see Born and Wolf, 1965)
The modern theoretical basis for the wave theory of light began with
Maxwell (Maxwell, 1891). Hertz (Hertz, 1893), who had earlier discovered
the photoelectric effect by showing that ultraviolet light increases current
emission from a cathode, was the first to construct a transmitting-receiving
pair of electric dipoles. With them, he confirmed that electromagnetic waves
carry energy through space. His dipole radiation has rotational symmetry
about its axis, a symmetry that Einstein referred to as "spherical" symmetry.
In 1900 Planck (Boorse and Motz, 1966, for example) showed that
quantizing electromagnetic energy in units of W = h(d accounts for otherwise
significant discrepancies between observed and calculated radiation laws. He
wrote that a most suitable body for energy exchange seemed to be Hertz's
dipole with its "spherical" waves. He went on to say, "There is one particular
question the answer to which will, in my opinion, lead to an extensive
elucidation of the entire problem. What happens to the energy of a light-
quantum after its emission? Does it pass outwards in all directions,
according to Huygens' wave theory, continually increasing in volume and
tending towards infinite dilution? Alternatively, does it, as in Newton's
emanation theory fly like a projectile in one direction only? In the former
case the quantum would never again be in a position to concentrate its
energy at a spot strongly enough to detach an electron from its atom."
Einstein (Einstein, 1905) used field energy quantization to explain the
photoelectric effect. These events rather conclusively showed that
electromagnetic energy is exchanged between atoms and free space in
quantized units, and this result, in turn, led to a fundamental difficulty.
Einstein wrote that quite differently from results of the Maxwell wave
equation "monochromatic radiation ... behaves in thermodynamic
theoretical relationships as though it consists of distinct independent energy
quanta of magnitude W=h(Q." He later extended Planck's work (Einstein,
1917) to show that the laws of statistical mechanics require quantized radiant
energy exchanges to be accompanied by quantized momentum exchanges,
p = W/c. This, in turn, requires that all of each unit of radiated energy travel
in the same direction. He commented: "... (Atomic) emission in spherical
waves does not occur, the molecule suffers a recoil of magnitude h oVc. This
seems to make a quantum theory of radiation almost unavoidable." In this
way, fully directed emission led to the conclusion light propagates as if it
Epilogue 319

consists solely of waves and exchanges energy as if it consists solely of

particles.
Only a few years after Einstein's "quantized momentum" paper,
electrons were shown to support an intrinsic magnetic moment. There was
an immediate problem with the result: If the moment arises because the
electron charge spins about an axis at a distance equal to the Lorentz electron
radius the necessary circumferential speed is many times the speed of light.
Then Schrodinger (Schrodinger, 1926) published the equation that now bears
his name, followed a few years later by Dirac's equations (see Dirac, 1958).
Both the Schrodinger and the Dirac equations correctly describe the behavior
of electrons in equilibrium. The Schrodinger equation is correct at non-
relativistic electron speeds and the Dirac equations are correct at all speeds;
electron spin is inherent to the Dirac equations but must be added in an ad
hoc way to the Schrodinger equation. Both equations yield the probability
that an electron will enter a transition. When it does, the input and output
energies and the correct power-frequency relationships result. Both
equations treat electrons as waves. Therefore, like light, an electron has
historically been thought to have both wave and particle natures.
Schrodinger developed his equation using an analogy with a known
relationship between classical mechanics and geometric optics. According to
Mehra (Mehra, 1972), Dirac, in the search for his equations: "started playing
with equations rather than trying to introduce the right physical idea. A great
deal of (the) work is just playing with the equations and seeing what they
give." Both Schrodinger's and Dirac's equations, therefore, came without an
inherent and obvious physical interpretation. Although results of solving the
equations undeniably give correct time-average values of measurable
quantities, the question of how to interpret the underlying behavior remains.
For example, accelerating charged particles radiate energy; stability requires
a closed current loop and/or a spherically symmetric region of charge that
pulsates radially. Yet atoms are stable and the Bohr orbit is some 20,000
larger than Lorentz's estimated electron size. Neither of the equations
addresses this issue.
Schrodinger (Schrodinger, 1952) expressed concern that quantum jumps
occurred in systems described by his linear differential equation. Bell (Bell,
1987) quotes Schrodinger as saying: "If we have to go on with these damned
jumps, then I'm sorry that I ever got involved." In an attempt to show that
what came to be the historic (Copenhagen) interpretation of quantum theory
cannot be correct, Schrodinger imagined a cat that might be described by
either of two eigenfunction solutions about a single potential. In one
320 The Electromagnetic Origin of Quantum Theory and Light

eigenstate, the cat was healthy and, in the other, it was dead (see Penrose,
1989). By the accepted interpretation of quantum theory, the cat is partially
in both solutions and thus both dead and alive; only after a measurement is
made is the cat one or the other. Although Bohr made the point that a cat,
being a large object, is not directly subject to the laws of quantum theory, to
many an enigma remained. Dirac (Dirac, 1958) commented on the power-
frequency law: "One would expect to be able to include the various
frequencies in a scheme comprising certain fundamental frequencies and
their harmonics. This is not observed to be the case. Instead, there is
observed a new and unexpected connexion between the frequencies." He
went on to say that this result is "quite unintelligible from the classical
standpoint."
Different opinions about the interpretation of the quantum theory
equations led to highly publicized discussions between Einstein and Bohr.
Einstein argued that the quantum equations do not supply complete
information and hence are incomplete. Bohr argued that the equations are
complete and supply all information there is on any level. Currently most
theoretical physicists support Bohr's view: the linear differential equations
of Schrodinger and Dirac are complete and describe all that can be known
about quantum mechanical events. Results of our Chapters 4 and 5 show this
viewpoint to be incorrect.
Quantum theory is based upon a number of disparate axioms and, in
contrast, electromagnetic field theory rests on only a few. Quantum theory
also requires that the classical electromagnetic laws not fully apply within
atoms. To some, it seems incongruous that nature should require such
disparate and seemingly conflicting bases for such strongly overlapping
sciences. To this end, Einstein (Einstein, 1959) wrote that: "I am, in fact,
firmly convinced that the essentially statistical character of contemporary
quantum theory is solely to be ascribed to the fact that this theory operates
with an incomplete description of physical systems." He also said that he had
devoted more time to thinking about this than any other subject. Although he
believed that the mathematics of quantum theory was uniquely correct, he
was bothered, for example, by the statistical nature of radiation onset from
an atom that is initially in a high-energy state. He argued that either the atom
is stable or it is unstable. If it is stable, it will not spontaneously decay and if
it is unstable, it will begin the decay process without a time delay. Yet,
atoms are stable until spontaneously undergoing a discontinuous energy drop
and emitting a pulse of radiation. He concluded that the wave function
description of these events is incomplete. He said: "Assuming the success of
Epilogue 321

efforts to accomplish a complete physics description, the statistical quantum

theory would, within the frame-work of future physics, take an
approximately analogous position to statistical mechanics within the
framework of classical mechanics. I am rather firmly convinced that the
development of theoretical physics will be of this type; but the path will be
lengthy and difficult."
While attention was focused on discussions of quantum theory, advances
in electromagnetism of ultimate consequence to quantum theory were being
made. Mie (Mie, 1908) used the classical wave theory of light and spherical
functions to analyze scattering of light by electrically small metallic
particles. Forty years later Chu (Chu, 1948) showed emission of
electromagnetic energy is necessarily accompanied by a source-associated
standing energy. As the size-to-wavelength ratio decreases, the standing
energy of radiated mode I increases as the inverse size-to-wavelength ratio
raised to the (2-M-l) power. Harrington (Harrington, 1960) showed that the
maximum possible gain of a single radiated mode is not zero but £(£+l). For
atoms of diameter 0.1 nm and light of wavelength 500 nm, the standing
energy that, by Chu's calculations, is necessary to support a maximum
modal number that, by Harrington's calculations, is necessary to obtain an
apparently infinite gain, is so large that the idea of fully directed energy
emission is untenable.
During the next decade, Manley and Rowe (Manley and Rowe, 1956)
derived the power-frequency relationships produced by nonlinear systems.
The result would satisfactorily explain atomic power-frequency relationships
if the atomic response was a nonlinear function of the driving force, but the
quantum theory equations are linear. During the next decade Bell (Bell,
1964) proved inequalities that subject certain aspects of the philosophical
discussions between Bohr and Einstein, as extended by Bohm, to
experimental test. In 1981 and 1982 such tests were successfully conducted
by Aspect, Grangier, and Roger. Bohm's discussion involves two electrons
that have opposite spin but are otherwise in the same state. Results of the
Aspect et al. experiments are that operating on one electron affects the other
without a measurable time delay even when the electrons are spaced an
arbitrarily large distance apart. The effect is known as electron nonlocality.
Adding nonlocality to wave and particle properties shows electrons are even
more complex than previously thought.
Since nonlocality does not prove that the quantum theory equations are
incomplete, an obvious, widely accepted interpretation is the opposite one:
322 The Electromagnetic Origin of Quantum Theory and Light

the quantum theory equations are complete. In Chapters 4 and 5, we supply a

counter explanation.

6.2 Overview
Because the first analysis of standing energy about an antenna was published
in 1948, during the developmental years between 1910 and 1935 the
interpreters of modern quantum theory could not have known of its critical
importance. Yet theoretical physicists constructed a logically coherent and
complete interpretation of quantum mechanics; their success is testimony to
the ingenuity of the participating individuals. The conceptual framework,
however, comes at a significant cost: it requires rejection of causality in the
sense that the dynamical structure of the universe at a given instant does not
uniquely determine the dynamical structure at the next instant. We suggest
that the non-causal interpretation of quantum theory is required largely
because standing energy is ignored in atomic processes
The axioms upon which electromagnetic theory is based show no
dependence upon the velocity of an observer, see Chapter 1; a conclusion is
that the speed of light in free space is the same in all inertial frames of
reference. In free space, the same axioms show no dependence upon the size
of an observer; the conclusion appears to be that the equations apply equally
well to all sizes. Experimental evidence shows that the axioms upon which
electromagnetic theory are based apply equally well from the nanometer
scale of electronic devices at least through the scale of galaxies. Yet, it is
widely believed that selected parts of electromagnetic field theory break
down within atoms. We suggest that the equations apply at least down to the
picometer scale of dimensions, without restrictions. Belief to the contrary is
caused, in large part, by an insufficient accounting of the affects of standing
electromagnetic energy.
It has been recognized for more than seventy years that electrons act, in
some circumstances, as a wave and, in other circumstances, as a particle. It
has been recognized for nearly twenty years that eigenstate electrons are
nonlocal. Although a detailed characterization of a nonlocal electron is
unknown, understanding eigenstates requires a detailed analysis of the
internal dynamics. Before addressing eigenstate electrons, note that there is a
tendency to think that upon going from the macroscopic to the atomic scale
of dimensions things will simplify. The notion has no logical or
experimental basis. For example, early last century nuclei were considered to
Epilogue 323

consist of protons and neutrons as basic building bocks; we now know

nucleons consist of quarks and gluons. Similarly, an electron is not a simple
object and nonlocality seems to suggest that a trapped, stable, eigenstate
electron be distributed over the full state independent of its size.
We show that as a Lorentz electron enters an eigenstate, it generates
dipole radiation, thereby producing an encompassing electromagnetic energy
field. That standing energy, in turn, produces an expansive radiation reaction
force on the electron, of a magnitude at least as large as the Coulomb
binding force, Eq. 4.2.14. Analyses based upon classical electromagnetism
show that an expanded charge density has arrays of possible stable modal
combinations. We postulate that the expansive radiation reaction force
transforms a Lorentz electron into an eigenstate electron consisting of a
dynamic ensemble of charge and current densities.
Combining energy conservation with the ensemble yields the
Schrodinger wave equation as a statistical descriptor, Section 4.3, but if and
only if the ensemble is in equilibrium. The linear equations of quantum
theory apply to eigenstate electrons that are in equilibrium or in near-
equilibrium, but not otherwise. This use of energy conservation applied to an
ensemble of objects is very similar to the science of thermodynamics.
Although radiation by a Lorentz electron is, in the main, limited to
electric dipole fields, Sections 3.11 and 4.13, radiation from the ensemble is
not. If an ensemble structure generates electric dipole radiation, that
radiation produces a force that drives higher order radiation of the parity
E ( 8 ) = E ( - 0 ) : magnetic quadrupole, electric octupole radiation, etc.
Similarly, if an ensemble structure generates magnetic dipole radiation, that
radiation produces a force driving higher order radiation of parity
E(0) = -E(-9): electric quadrupole, magnetic octupole, etc. Neither set of
parities, alone, is resonant. Quite differently, if the structure drives both
parities, properly phased and oriented, the system is resonant, Section 3.17.
In that case, the radiation reaction force regeneratively builds the radiation
level. The reaction force due to this radiation, see Section 5.17, becomes
thousands of times larger than the Coulomb binding force and dominates all
other forces. One result is a rapid energy change between eigenstates, an
energy jump. Another is that the Manley Rowe power-frequency
relationships, Section 5.1, apply during the nonlinear process.
The radial components of the steady state photon fields are given by
Eqs. 5.6.1 and 5.6.2, and the electric portion of the resulting radiation
reaction pressure is shown by Eqs. 5.17.7 and 5.17.8. This radiation pressure
is due to the local energy field and, even when truncated with inverse quartic
324 The Electromagnetic Origin of Quantum Theory and Light

field terms, is many orders of magnitude larger than the pressure of Coulomb
attraction. The sign of the force alternates with time. We expect it to force
the radiating electron to undergo nova-like, expansive oscillations in the
direction of radiation. Figure 6.2.1 illustrates the time variation of the
radiation reaction pressure on a rigid, radiating spherical surface at a few
Bohr radii. For clarity, phase dependence upon zenith angle is suppressed.
Figure 6.2.1a is drawn for phase zero. The upper hemisphere cap represents
an expansive pressure centered about the direction of telefield radiation and
the lower hemisphere pressure is compressive. Figure 6.2.l.b shows the
pressure at phase n/5. The expansive upper hemisphere pressure is decreased
and the compressive lower hemisphere pressure is increased. Figures 6.2.l.c
and6.2.1.d are at phases 2n/5 and 3rc/5 and entirely compressive.
Figure 6.2.le is at phase 4n/5; the upper hemisphere pressure is small but
again expansive. Phase n returns to Fig. 6.2.1a. The configurational
sequence is reminiscent of the action of a bellows.

6.3 The Radiation Scenario

Classical statistical mechanics analyzes an ensemble of identical
particles. The particles are modeled as realistically as possible and there is
little or no difficulty interpreting ensemble-averaged results. The state of the
ensemble may be specified by the positions and velocities of the component
particles, which are sufficient to determine the system energy. With
statistical mechanics, studies of complicated systems are accomplished with
no knowledge of the precise state of individual particles. The actual state is
assumed the most probable state, and if there is full knowledge of an
ensemble at a particular instant, its value at the next instant is predictable. A
question fundamental to quantum theory is why an individual eigenstate
electron acts as a statistical ensemble.
The radiation reaction force, see Eq. 4.2.14, provides at least a partial
answer; it is an extensive force of magnitude greater than the attractive
Coulomb force. Using an eigenstate ensemble as a physical basis, the Fourier
integral transforms of Eqs. 4.3.7 describe the unknown physical realities. To
match known experimental results, it is necessary for eigenstate electrons to
contain a definite, but unknown, distribution of units of charge and current
density; the smallest unit is determined by the discreteness of the charge
distribution. The uncertainty of calculated results, Eq. 4.4.7, arises because
Epilogue 325

o.

-0.5.

-1.

-1.5,

-2.

Figure 6.2.1 Time variation of radiation reaction pressure on a radiating electron at a

mid-range value of L/(ka). Phase variation with zenith angle is suppressed.
326 The Electromagnetic Origin of Quantum Theory and Light

of incomplete information about the detailed structure of the charge and

current distributions, not because of inherent properties of the distributions
themselves. Causality applies in the sense that the detailed structure and
kinematics of all charge and current densities at one instant uniquely
determine the values at the next instant. Although this difference does not
affect expectation values, it has a profound effect upon the interpretational
philosophy and the characterization of measurable quantities. Causality also
retains Einstein's deterministic view of atoms.
A set of resonant, regeneratively driven electromagnetic field modes, to
which Chu's proof of the limiting value of Q versus electrical size does not
apply, is shown in Section 3.17. The field set and certain properties are listed
in Table 3.2.1. Section 5.5 and Eq. 5.5.18 show that radiation with the
kinematic properties of photons result from imposing energy minimization
as a boundary condition on a general multimodal expansion, the field set of
Eq. 3.17.4. Therefore, fields with the kinematic properties of photons are
resonant and are not subject to Chu's criterion about electrically small
radiators. For this reason, large magnitudes of large order radiation modes
are supported in the regeneratively driven radiation.
The nonlocal electron and classical electromagnetism not only predict
fully directed electromagnetic energy exchanges, but also supply a complete
expression for all fields that exist during the transition process. In this sense,
the analysis is as complete as the energy exchanges by macroscopic
biconical antennas, see Chapter 2. There is no way Einstein could have been
aware these results were possible when he wrote that full directivity makes
"a quantum theory of radiation almost unavoidable." The electromagnetic
background necessary to understand this aspect of the physics was simply
not available. The linear equations of quantum theory do not provide a
description of the complete electromagnetic radiation. They describe only
equilibrated states before and after transitions, Sections 4.3 and 4.5, and the
equilibrium dipole moment of interaction between states that, in turn,
determines the probability of a transition.
Equilibrium periods, during which the linear quantum theory equations
apply, are separated by time hiatuses during which energy exchanges occur.
As shown in Chapter 5, the initial radiation produces a radiation reaction
force on the source that drives it regeneratively and produces more radiation.
The process continues until all available energy is gone and the electron has
entered a new energy state. Because the process is nonlinear the Manley
Rowe equations apply to the radiation, the energy level changes, and only
the initial and final electron states are occupied. There is no mixture of
Epilogue 327

states; that is, Schrodinger's cat is dead or alive, but not both. There is no
way Dirac could know of the time hiatus during which linear equations do
not apply when he wrote: "there is observed a new and unexpected
connexion between the frequencies." Neither could Schrodinger have known
of them when he wrote about "these damned jumps."
A charge with an intrinsic magnetic moment interacts with the closed
circuit currents to produce a continuous torque. A continuous torque assures
a dynamic configuration. Radiation begins when the ensemble configuration
evolves into a configuration set that generates the multipolar moments of
Table 3.17.2. For that set of moments, the reactive reaction force of
Eq. 4.2.14 is nonexistent, only the real reaction force of Eq. 4.2.7 remains,
and hence the regenerative energy exchange of Chapter 5 begins. In
accordance with classical statistical mechanics, the initial details of the intra-
electron structure and the governing laws determine the onset time of
spontaneous decay. With this model, the primary difference between
statistical and quantum mechanics is that statistical mechanics, in the main,
deals with an ensemble of neutral particles and quantum mechanics does not.
If the length of the emitted, steady state wave train and the upper modal
limit of the source fields are both infinite, the field is fully directed and
retains its original shape and size over arbitrarily large values of time and
distance. However, since both are finite, the size of the calculated steady
state wave packet would slowly increase. Although the transient field
solution that builds to and from the steady state value is unavailable, it
necessarily describes flux closures for steady state flux lines. It is unknown
if closures produce a fully directed wave packet; that is, it is unknown if the
energy packet arriving from a distant star has the same physical extent it had
when emitted or is extended over a larger volume. We know only that the
total energy in the coherent wave packet remains constant as it travels
through lossless space.
The descriptive transcendental terms in the field equations of Chapter 5
come, bit-by-bit, from each mode, see Eq. 5.8.11 and 5.8.12 for example,
and are approached only if the maximum modal number is large. An
important question is why the recursion formula of Eq. 5.5.15, the same
recursion relationship that applies to spherical Bessel terms in a plane wave,
is uniquely correct for quantized radiation. The reasons appear to be that
only this particular recursion formula produces a set of z-axis transcendental
fields that supports unidirectional energy flow to or from the source, and that
only it produces a null for the standing energy. A greater or lesser
dependence of modal coefficients upon modal number would result in fields
328 The Electromagnetic Origin of Quantum Theory and Light

that depended upon all modes, not just the highest numbered one, see
Eq. 5.11.1. For such a case, reactive energy that does not contribute to the
regenerative drive would be present and the radiation reaction force of
Eq. 4.2.14 would apply, braking power emission. The recursion relationship
of Eq. 5.5.15 uniquely accomplishes two things: it avoids the radiation
reaction of Eq. 4.2.14 and it meets the requirements of Section 3.17. Only
with Eq. 5.5.15 is the radiation reaction force on an electron merely that
caused by energy escaping from the system, Eq. 4.2.7.
Spherical Bessel functions give rise only to half the teledistant terms of
Eqs. 5.13.6 through 5.13.8, the other half and all other terms come from
spherical Neumann functions. Yet, all Poynting vector terms that describe
unidirectional energy flow are products of a spherical Bessel term and a
spherical Neumann function term. A possible radiation scenario is that the
picostructure of an eigenstate evolves to produce the Neumann function
terms of Eqs. 5.13.6 through 5.13.8, by which the source is immersed in its
own standing energy field with no energy exchange. When an incoming
plane wave, totally described by spherical Bessel functions, is applied to the
system all needed field forms are present for radiation to occur. The
incoming plane wave need only trigger a quantum of energy by providing
terms with the needed phases. Planck could not have been aware of this
when he wrote of extended photons "a quantum would never again be in a
position to concentrate its energy at a spot strongly enough to detach an
electron from its atom."
As described herein, the primary historic obstacle to understanding
atomic-level phenomena was that the principals lacked the tools necessary to
account for the radiation reaction force of the generated fields, and they used
a localized electron model. For these reasons, they were forced to conclude
that electromagnetic field theory is not totally applicable within atoms. We
show that consistent application of the full equations within atoms, without
restrictions, leads to self-consistent results and removes the "strangeness" of
quantum theory required by the historic interpretation.
Epilogue 329

References
Aspect, P. Grangier, G. Roger, "Experimental Tests of Realistic Local Theories via
Bell's Theorem," Phys. Rev. Lett. Vol. 47, 460 (1981) and "Experimental
Realization of Einstein-Podolsky-Rosen-Bohm Gedanken Experiment," Phys.
Rev. Lett. Vol. 49, 92 (1982)
J.S. Bell, "On the Einstein-Podolsky-Rosen Paradox," Physics Vol. 1, pp. 195-200
(1964), see also Speakable and Unspeakable in Quantum Mechanics,
Cambridge Press (1987)
N. Bohr, "On the Constitution of Atoms and Molecules," Phil. Mag. Vol. 26, 1-19
(1913), Also in The World of the Atom, H. A. Boorse and L. Motz, eds., Basic
Books (1966) p. 751-765
M. Born, E. Wolf, Principles of Optics Electromagnetic Theory of Propagation,
Interference and Diffraction of Light, 3rd ed. Pergamon Press (1965)
L.J. Chu, "Physical Limitations of Omni-Directional Antennas," J. Appl. Phys., vol.
19,1163-1175(1948)
P.A.M. Dirac, The Principles of Quantum Mechanics, 4th ed., Oxford Press (1958)
pp. 1-2
A. Einstein, "Concerning a Heuristic Point of View About the Creation and
Transformation of Light," Ann. Phys. vol. 17, 132-148 (1905); also in The
World of the Atom, H. A. Boorse and L. Motz, eds., Basic Books (1966) p. 544-
557
A. Einstein, "The Quantum Theory of Radiation," Phys. Z., vol. 18, 121-128 (1917);
also in The World of the Atom, H. A. Boorse and L. Motz, eds., Basic Books
(1966) p. 888-901
A. Einstein, "Reply to Criticisms" in Albert Einstein, Philosopher and Scientist
(P.A. Schilpp, editor) Harper Torchbooks Science Library (1959) pp. 665-688
R.F. Harrington, "Effect of Antenna Size on Gain, Bandwidth, and Efficiency," J.
Res. National Bureau of Standards vol. 64D, 1-12 (1960)
H. Hertz Electric Waves: Researches on the Propagation of Electric Action with
Finite Velocity Through Space (1893) Translated by D. E. Jones, Dover
Publications (1962)
C. Huygens, "Treatise on Light," (1678) trans, by S.P. Thompson, Dover
Publication (1912); also The World of the Atom, H. A. Boorse and L. Motz, eds.,
Basic Books (1966) pp. 69-85
H.A. Lorentz, The Theory of Electrons, (1909) also Dover Publications 2nd ed.,
(1952); also Pais, A., "The Early History of the theory of the Electron: 1897-
1947," in Aspects of Quantum Theory, A. Salam, P. Wigner, eds., Cambridge
Press (1972) pp. 79-93
J.M. Manley, H.E. Rowe, "Some general properties of nonlinear elements, Part I:
General energy relations," Proc. IRE, vol. 44, pp. 904-913 (1956), and "General
energy relations," Proc. IRE vol. 47, pp. 2115-2116 (1959)
J.C. Maxwell, A Treatise on Electricity and Magnetism, 3rd ed., (1891) reprint Dover
Publications (1954)
330 The Electromagnetic Origin of Quantum Theory and Light

J. Mehra, '"The Golden Age of Theoretical Physics': P. A. M. Dirac's Scientific

Work from 1924-1933," in Aspects of Quantum Theory, Cambridge Press
(1972) p. 45
G. Mie, "A Contribution to Optical Extinction by Metallic Colloidal Suspensions,"
Ann. Physik. Vol. 25, p. 377 (1908)
I. Newton, Mathematical Principles of Natural Philosophy, (1687) Translated by
I.B. Cohen, A. Whitman, Univ. of California Press (1999)
I. Newton, Optiks, (1704); Dover Publications (1952)
R. Penrose, The Emperor's New Mind Concerning Computers, Minds, and the Laws
of Physics, Penguin Books (1901)
M. Planck, "The Origin and Development of the Quantum Theory," Nobel Prize in
Physics Address, 1919, in The World of the Atom, H. A. Boorse and L. Motz,
eds., Basic Books (1966) p. 491-501
E. Schrodinger, Ann. Phys. Vol. 79, p. 489 (1926); also "Wave Mechanics," in The
World of the Atom, H. A. Boorse and L. Motz, eds., Basic Books (1966) pp.
1060-1076
E. Schrodinger, "Are There Quantum Jumps?" Brit. J. for the Phil, of Sci., vol. 3,
pp. 109-123 and 233-247, (1952)
J.J. Thompson, "Cathode Rays," Phil. Mag. Vol. 44, pp. 293-311 (1897), also The
World of the Atom, H. A. Boorse and L. Motz, eds., Basic Books (1966) pp.
416-426
R.C. Tolman, The Principles of Statistical Mechanics (1938), reprinted by Dover
Publications (1979)
 Appendix

A.1 Introduction to Tensors

The application of field concepts to classical physics is made easier by the
use of tensors. Tensor notation simplifies what would otherwise be tedious
notational bookkeeping. The simplest and lowest rank tensor is a scalar, the
next higher ranking tensor is a vector, and higher order tensors are referred
to simply as tensors:

Rank
r=0 Scalar
r=1 Vector
r=2 Tensor

Table A. 1.1 Properties of Tensors

The number of numbers that it takes to construct a tensor, No, depends

upon the rank of the tensor and the number of dimensions. If N and r are,
respectively, the number of dimensions and the rank, the number of numbers
is:

N0 = Nr (A.l.l)

Independently of the number of dimensions, a scalar is fully described by a

single number. Examples are the speed of light, c, and electron charge, q.
Scalars have the same value in all inertial frames.
It takes as many numbers as there are dimensions to describe a vector.
Example vectors are electric field intensity, velocity, and position. Let
vector, A, be known in three dimensions. The three numbers represent
components along each of the three orthogonal coordinate axes, [x\,X2,x-$).
331
332 The Electromagnetic Origin of Quantum Theory and Light

If the same vector is determined using a set of axes rotated to new coordinate
positions yx\ ,x\ ,^'3) the result is the new vector components:

A'j = Aj cos(x'j ,x\) + A 2 COS(JC'2 ,*i) + A 3 cos(x'3 ,xj)

A' 2 = AiCOs(x'i ,X2) + A2COs(x'2,X2) + A3COs(x'3,x2) (A.1.2)
A'3 = A j cos(x'! ,x 3 ) + A 2 cos(x'2 ,^3) + A 3 cos(x'3 ,x^)

The directional cosine of the angle between axis 'i' in the prime coordinates
and axis ' j ' in the unprimed coordinates is signified by cos(x'j ,x.). Making
the definition that the direction cosine c ^ - coslVj,x-. I, Eq. A.1.2 may take
the more compact form:

3
A'j=ScikAk (A.1.3)
k=l

It follows that:

3
Ar=£ckrA'k (A.1.4)
k=l

An example of a second rank tensor is the stress tensor in crystals. Such a

tensor transforms between coordinate systems as:

3 3
T'rs^EX^sjTij (A.1.5)
i=lj=l

The number of direction cosines for a transformation between coordinates

systems is the same as the rank of the tensor.
Like all other vectors, a position vector transforms between coordinate
frames as:

y 1 = c 11 jr 1 +c 12 Jr 2 +c 1 3jr3
^ 2 = C2i-fi+C22-f2 + c23-*'3 ( A - L6 )
y 3 = C31Jr1+C32Jr2+C33jr3
Appendix 333

By definition the rotation matrix is:

cn c 12 c 13
c c
ij = 2 1
c
22 c
23 (A. 1.7)
c31 c32 c33

The length of a differential vector in three dimensions is:

(A/-)2 = (AVj ) 2 + (A^ ) 2 + (A^3 ) 2 s (Ax,)2 + (A*2)2 + (Ax3)2 (A. 1.8)

The transformation equalities derived from Eqs. A. 1.8 are:

c
? 1 + c21 + c31 = 1 = c12 + c22 + c32 = C H + c23 + c33 (A. 1.9)

This may be written as

3 3
£CJJCJJ = 1 and ^ c y c ^ O ; j * k
i=l i=l

Define the Kronecker delta function to be

8 j k = l if j = k; 5 j k = 0 if j * k (A.1.10)

With the definition, the condition on directional cosines may be written more
compactly as:

£ c i j c i k = 5 jk (A.l.ll)

An exercise is to show that

detcy = 1 (A.1.12)
334 The Electromagnetic Origin of Quantum Theory and Light

Solution: Let the volume of cube xi*2*3 be equal to one. The volume in the
transformed coordinates is unchanged by describing it in another frame, so it
too is equal to one. The volume is given by:

V=^.(^2x^3)=l

Writing out cross products in terms of directional cosines gives:

-^2 X-^3 = -^1 ( c 2 2 c 3 3 ~ c 23 c 32 ) + -*"2(c33cl 1 ~ c 31 c 13) + -^3( c l l c 22 ~ c 1 2 c 2 1 )

From which it follows that:

c
ll(c22c33 - c 2 3 c 3 2 ) + C
12(C23C31 -C21C33)
•M-^x-^H
+ c 13( c 21 c 32 - c
22c3l)

The determinant of cjj is:

|cij| = cll(c22C33-c23c32) + c
12(c23c31-c21c33)+c13(c21c32-c22c3l)

Comparison shows that

N=1

A.2 Tensor Operations

A common summation convention that reduces the number of symbols that
would otherwise be required is that if an index occurs twice a summation
over all possible values is required. That is:

c
ijcik=5jk (A- 2 - 1 )

Consider some arithmetic operations on tensor fields. Tensor addition is

defined only for tensors of equal rank; for example, addition of a scalar and a
vector is not defined. Addition of tensors of equal rank is by:

C (A.2.2)
ij-Aij + B
i)
Appendix 335

Proof consists of showing that the sum obeys the coordinate rotation
properties of a second rank tensor. In Eq. A.2.2 indices 'i' and ' j ' appear
only once in each term and, therefore, are running indices. Equation A.2.2
consists of nine separate summations.
Subtraction is accomplished by multiplying By by minus one and
adding.
Multiplication is defined between tensors of arbitrary rank. By
definition,

Ci..jr..s = A L j B r .. s (A.2.3)

The rank of C is the sum of the ranks of A and B. For example the product of
vector Ai and scalar a is aAi, another vector. The product between two
vectors is a second rank tensor, for example the product of Aj and Bj is Qj =
AjBj, where indices 'i' and ' j ' are both running indices; Qj represents nine
numbers.
Division by tensors other than rank zero is not defined.
In addition to these scalar-like arithmetic operations there are operations
confined to tensors.
Tensor contraction is accomplished by equating two indices. Equal
indices signify a summation and summation results in a tensor reduced in
rank by two from the initial one. The process is, therefore, restricted to
tensors of rank r > 2. As an example:
A
'rsm = c ri c sj c tk c u ^A ijk/ (A.2.4)

After equating 's' and 't' and summing:

A
Vssu = c ri c sj c sk c uf A ijkl (A.2.5)

Since:

c
sj c sk= 8 jk and A'rSsu = c ri<W A ip (A.2.6)

This is a tensor of rank two less than the starting one.

A common example is the scalar product between two vectors, A; and
Bj. To evaluate, begin with the product:
336 The Electromagnetic Origin of Quantum Theory and Light

Cij=AiBj (A.2.7)

Equating indices ' i ' and ' j ' and summing over the indices gives:

Cj^AjB^D (A.2.8)

The product forms scalar D.

A.3 Tensor Symmetry

Physically real tensors are either symmetric or antisymmetric. Terms of
different symmetry are defined as:

Symmetric tensor Arstu = Artsu

(A.3.1)
Antisymmetric tensor Arstu = - A r t s u

Symmetric and antisymmetric tensors have, respectively, N(N+l)/2 and

N(N-l)/2 terms.
An important special case is a three dimensional tensor of rank two, say
Ty. Such tensors transform as:

T'ij = c i r c j s T r s (A.3.2)

Equation A.3.2 is short hand notation for the nine terms of T'jj, each of
which contains nine separate numbers. For example:

c
l 1C21T11 + c l 1C22T12 + c l 1C23T13
T ' 1 2 - c l r C 2 S T r s - +c 1 2 c 2 1 T 2 1 + c 1 2 c 2 2 T 2 2 + c 1 2 c 2 3 T 2 3 (A.3.3)
+ C 13 C 21 T 31 +C 13 C 22 T 32+ C 13 C 23 T 33.

If Ty is antisymmetric, Ty = - Tjj and the transformation simplifies to:

T
'23 = C 11 T 23 + C12T31 + C13T12
T
'31 = C21T23 + C22T31 +C 23 T 12 (A-3-4)
T
' l 2 = c 31 T 23 + c 32 T 31 + c 33 T 12
Appendix 337

The proof of Eq. A.3.4 follows by writing out the terms in the form:

° + C11C22TJ2-C11C23T31
T
'l2= -Cl2C21 T 12 + 0 + c
12 c 23 T 23
_ + c 1 3 c 2 1 T 3 1 - c 1 3 c 2 2 T 2 3 + °_
= ( c l 1^22 - C12C21 ) T 12 + (ci2 c 23 ~ c 13 c 22 ) T 23 + ( c 13 c 21 ~ c l 1 C 23) T 31

Similarly:

T
*23 = ( c 21 c 32 ~ c 2 2 c 3 1 ) T 12 + (c22 c 33 ~ c 23 c 32 ) T 23 + ( c 23 c 31 _ c
21 c 33 ) T 31

From the determinant:

c
ll(c22c33-c23c32) + c12(c23c31-c21c33) + c13(c21c32-c22C3l)=1

Combining this result with c^c^ = 5jk results in:

C 1 1 C 1 1 +C 1 2 C 1 2+C 1 3C 1 3 = 1

The latter two equations combine to show that:

C11=(c22C33-C23C32); C12 = ( c 2 3 C 3 1 - C 2 1 C 3 3 ) ; C13 = ( c 2 i C 3 2 - C 2 2 C 3 1 )

Substitution of this result back into the expansion results in Eq. A.3.4.
This result shows that an antisymmetric second rank tensor, Tjj,
transforms like a vector. It is tempting to call it a vector, but if the coordinate
system is switched form a right hand system to a left hand system the
components change sign. It is therefore a pseudovector.

A.4 Differential Operations on Tensor Fields

The gradient operation increases the rank of a tensor by one. As an example,
let a(r) represent a scalar field. Taking the partial differential:
338 The Electromagnetic Origin of Quantum Theory and Light

Mr) „
^ L 2 = Vi (A.4.1)

Make the equality:

da(r) = tejda(r)_c, y
1J
dxx 6x\ dx'; '

Combining gives:

Vi = c , i j V' j (A.4.2)

Since Vj transforms as a vector, it is a vector. This gradient operation may be

conducted on a tensor of any rank. For example,

\.k\ =~ ^ (A.4.3)

The divergence operation decreases the rank of a tensor by one.

0(r)^ (A.4.4)

To show that a(r) is a scalar, write it as:

a(r) = ^-±J^Jl = c- c. ^L = ^ L = &{r) (AA5 )

J
ax\ ox^ dx^ ox\

The divergence operation may be conducted on tensors of any rank.

dR
i ik
dxk

Proof follows in the same way as for Eq. A.4.4.

The curl operation begins with the vector differential operation:
Appendix 339

T b L £ _ °*-i.-n..k (A 4 7)
1 J kn
-- a*n a*j

This increases the rank by one. A particularly useful special case is for
vectors. Let:

T
--rl-f' (AA8)

Note that since Tjj is antisymmetric it has N(N-l)/2 independent numbers. In

three dimensions, it has three, the same as a vector and we already saw that
antisymmetric second rank tensors transform like a vector. Therefore, the
curl of a vector changes the vector to an antisymmetric second rank tensor
that is pseudovector. The pseudovector acts like a vector in any given
coordinate system but changes sign if the systems are changed from a left to
right hand system.

T AX(BXC) = B(A»C)-C(A»B)
2. grad((])V|/) = V (<j)\|/) = (j) V\j/ + \r/Vd>
3. div(<|>A) = V«(<t>A) = (|>V«A+V<j).A
4. cur 1(<|>A) = VX(<|>A) = <|>VXA + V<|>XA
5. V«(AXB) = B»(VXA)-A«(VXB)
6. VX(AXB)=A(V»B)-B(V»A)+(B»V)A-(A«V)B
7. V(A»B) = AX(VXB) + BX(VXA) + (B»V)A + (A«V)B
8. V 2 (l/r) = 0, i f r > 0
9. V2A = V ( V « A ) - V x ( V x A )

Table A.4.1 Table of Vector Properties

|A»cL$'=f(V»A)d^
J J
10.
n
- ^<t)d$,= J(V(t))d^
12
- |AxdJ'=-J(VxA)d^

Table A.4.2 Integrals over closed surfaces

340 The Electromagnetic Origin of Quantum Theory and Light

!3. §<\>d£ = jdSxV<\>

Table A.4.3 Integrals over open surfaces

A.5 Green's Function

The 4-Laplacian of the electromagnetic potential is defined by Eq. 1.5.4, and
repeated here:

a | = * <"•»

We seek to integrate that differential equation in order to obtain a general

expression for the electromagnetic potential itself. For this purpose, it is
helpful to define a similar but simpler function, to integrate that function,
then to use the integral to obtain an expression for the electromagnetic
potential. The function is Green's function G(X a ,X' a ). By definition it is:

In Eq. A.5.2, the four-dimensional delta function indicates the Dirac delta.
By definition the one dimensional Dirac delta function satisfies the integral
relationship:

|f(x)d(^:-y)dr = (A53)
o
The upper or lower solution applies if the range of integration respectively
does or does not include x\ The integrand magnitude of a Dirac delta
function increases without limit and the width Ax decreases without limit in
a way that retains a product value of one.
Construct the equation:
Appendix 341

Jd[g(x)-g(y)]ck=J d[g(^)-g(y)] dg(*)

dg(jr)/djr

It follows from the definition of the delta function that

(A.5.4)
dg(jr) / dr
x=ji

The method used to integrate Eq. A.5.1 is a four-dimensional extension

of a common three-dimensional technique. The procedure begins with the
quadruple integral:

d2G a2A
JJJJA. 9Xn3Xg
G-
3Xg3Xn
: 1 dX 2 dX 3 dX4 = 0 (A.5.5)

The equality results since all integrals are evaluated at ±00 and the integrand
decreases with distance rapidly enough so the integral is zero at the infinite
limits. Substituting Eq. A.5.2 into the first term in the integrand and
substituting Eq. A.5.1 into the second gives:

A a (X p )= |4J7 j y X ' y )G(X'Y ,XY)dX1dX2dX3dX4 (A.5.6)

Since JaQ^'y) is known but G(X'Y,XY) is not, it is necessary to solve for

G(X'Y,X Y ) using Eq. A.5.6 before, in turn, solving for A a (Xp). For this
purpose, consider an aside on the four dimensional Fourier transform pair:

F
(XYH ^ JJljH(KY)e-/XvKVdK1dK2dK3dK4
271
(A.5.7)
H K
( v) = ( i ) |jnF(xY)e/X^dX1dX2dX3dX4

KY and XY are unknown conjugate variables that are to be determined.

Making the definition that:
342 The Electromagnetic Origin of Quantum Theory and Light

F(X Y ) = [ 5 ( X Y - X ' y ) ] 4 (A.5.8)

Combining Eqs. A.5.7 and A.5.8 results in:

( 1 ^2
[ 8 ( X Y - X ' y ) ]l 4 ~ T~ JJJIe"' KY(XY " X ' Y) dK 1 dK 2 dK 3 dK 4
271
2 (A.5.9)
/ 1 \ /K-yK. Y
H(K Y ) =
V27ty

It is convenient to introduce an additional function, g(KY), defined by the

equation:

G(X Y ,X , Y ) = JJJJg(K a )e~ / K Y ( X Y " X ' Y ) dK 1 dK 2 dK 3 dK4 (A.5.10)

To solve for the function g(KY) consider, as an example, the conjugate pair x
and kx to be a single dimension of Eqs. A.5.2 and A.5.9 with the equalities:

d2G(x-x')
= -8(x-jr')
dx2

(A.5.11)

G(x,x') = lg(Jtx)Q~ik'{x~^dJcx

Differentiating G(x,x') twice with respect to x gives

8 ( x - x1) = /tljg(/tx)e~l^(x~^dJtx (A.5.12)

Combining gives:

^g(/y=i- (A.5.13)

Extension to four dimensions gives:

Appendix 343

n2 i
8(K
V ) = . 2* K
aKa (A.5.14)

Substituting Eq. A.5.14 back into Eq. A.5.10 results in:

G(x y ,X' y ) = j ^ j J J J J 6 KK " dK dK

l 2dK3dK4 (A.5.15)

It is convenient to use three-dimensional notation to evaluate Eq. A.5.15. For

this purpose note that the four variable set (x,y,z,ict) is complex and, if the
exponentials are to remain oscillating functions, it is necessary that the
conjugate variable set K a also be complex. Writing it as (kK,ky,kz,i(o/c) and
substituting into Eq. A.5.15 gives:

•/ , \ 4 -(Ky(X Y -X'y)
G(XY,X'Y) = ^ — J f f J I p _ ^ dAto (A.5.16)

where

k2 = kl + ^ + ^l; dA= dJcjMy&kz (A.5.17)

Equation A.5.16 is the sum of two Cauchy integrals, integrals that may be
evaluated by use of the Cauchy integral identity:

27i/f(z') = f - ^ ; d z (A.5.18)

Introducing p as a small, real, positive number used as a construction tool

whose value is eventually put equal to zero, Eq. A.5.16 may be written as:

G(X Y ,X V )^J]fd4 dcoe- /K v (X v- X V

(2TI) 4 (W - ck- 4))(co + c/c+ rp)

Moving the space portion of the exponential out from under the time-
dependent integral results in:
344 The Electromagnetic Origin of Quantum Theory and Light

G(XY ,X' ) = - ^ fJf Ueikri dCde

""°(/"° (A.5.19)
Y Y 4JJJ 7 v
(2TI) (a-ck-ip)((i>+ck+rp)

Restating, the problem is that given an electric charge at (/•'/) to find the
function G(Xy,X'y). The field is zero before the charge is introduced. That is,
with /" = t-f all fields are zero for /" < 0. The last integral of Eq. A.5.19 may
be evaluated first along the real axis and then back around an infinite,
complex co path. For t" < 0 the return path encompasses the lower half-plane,
where no poles are enclosed. For t" > 0 the return path is around the upper
half of the complex plane, where two poles are enclosed. Evaluation of the
integral gives:

e-//"(c£+rp) Q-it"(-ck+ip)
dcoe"^" 2TC
= 271/ = —sm(ckt")
(co - ck- /p)(co + ck+ /p) 2ck 2ck ck
(A.5.20)

Combining Eqs. A.5.20 and A.5.21 gives:

G(X Y ,X' y ) = — l —jjjjdJk;~'* r sm(ckt") (A.5.21)

(2TI) *

Next let R be the space vector from source point r' to field point r, and
choose it to be in the z-direction. Then kr = £Rcos0 where 0 is the polar
angle. Also, replace dk with £2d£sin0d0d(|):

G(XV ,X'V ) = —l-^r fff/fdy{sined0d4)e~/;aicose sin(c*r") (A.5.22)

r r
(2n) 3 JJJ

Evaluating the angular integrals over an enclosing sphere gives:

? ' °° Ak
G(X V ,X' V ) = — ^ - f—sinO»)sin(fifr") (A.5.23)
Y Y
(2n) 2 J0 R
Appendix 345

Since the integral of Eq. A.5.23 is an even function of k, it may, without

changing the value of the integral, be replaced by the equation:

/(a>/
G(XYY ,X'YY ) = - 42 - J f d4e-
L
"~m - e''(t0' "+m (A.5.24)
87t R „

Equation A.5.10 shows that Eq. A.5.24 is the sum of two Dirac delta
functions. The second one is evaluated at advanced time /" < 0 when there
are no charges, and if causality applies all results from it are equal to zero.
Working with the retarded time t" > 0 when there are charges, using
Eq. A.5.3 it follows that:

8(R-cf") = --8(f"-R/c) (A.5.25)

Combining with the first term in the integrand of Eq. A.5.24 gives:

G(XV,X'V) = — — Mf-R/c) (A.5.26)

v
' ' ' 4mRc
This completes the derivation of G(Xy,X'y).

A.6 The Potentials

To obtain potential A v of a moving charge density, substitute Eq. A.5.26 into
Eq. A.5.5. The result is:

Av(Xv) = 0fdrJd, 1 i^(,"-R/,) «A.„,

Distance R(Xy,X') is the distance between the source and field points. Using
Eq. A.5.3 to evaluate Eq. A.6.1 gives:

J(/V)
JJJ
4TI (R-R«V/C)
346 The Electromagnetic Origin of Quantum Theory and Light

o ( ^) = JLrrr P M . ^ . (A . 6 . 2)
4 n J J J ( R - R •»/£•) '

Equation A.6.2 is the final form for the electromagnetic 4-potential of a

moving charge. Distance from the point of field emission to the field point
when the radiation is received. These are the Lienard-Wiechert potentials.
To obtain the potential A v of an oscillating charge density, note that
Eq. A.6.2 remains applicable except, for this case, the average velocity of the
oscillating charge is zero. The resulting equation is:

Aov(/-)e^ = ^ J d r ' J d / ' ^ ^ 5 ( / ' , / - R / , ) (A.6.3)

Subscripts "0" indicate the value is independent of time. Applying Eq. A.6.3
to a differential volume in space gives shows that in three dimensions the
potentials at position r due to a current density are given by:

A(, ) e ^ = Ji- ± 1 ^
4nJ R(/%/-') (A.6.4)
1 roMe^'-11^
47teJ R(r,r')

These are the retarded potentials.

A.7 Equivalent Sources

It is shown in Section 1.10 that the force fields satisfy the partial differential
equations:

V x ( V x E ) + eu,—- = 0 and V x ( V x B ) + en—^- = 0 (A.7.1)

Other helpful relationships are:

Appendix 347

VX(VXE) = V(V«E)-V2E and e\i — = -/t2E (A.7.2)

dr

Combining the two equations shows that:

V 2 E + >f2E = 0 and V 2 B + /f2B = 0 (A.7.3)

These are the Helmholtz equations for the field intensities. In rectangular
coordinates the form of the vector and scalar Laplacian operators are
identical.
The objective is to obtain expressions for the field vectors at any field
point, r(x,y,z), external to a field-generating volume as a function of field
values on the surface of the volume. The development requires three vector
integral equations, the divergence theorem and two related ones. Let dS
represent a scalar differential area on the surface of the volume and n be a
unit vector directed normal to the surface at the same point. At the surface,
fields F and (]) have the continuity properties of electromagnetic fields: they
are continuous with continuous first derivatives.

|di,»F = Jv»Fd^ ^dJ T xF = | V x F d ^ ^dS=jV^>dF (A.7.4)

Next let 0 and \y each represent scalar fields and construct the function
<J>Vij/. Substituting the new function into the divergence equation gives:

^(j)V\|/.d$ , s|v«(<l)V\i;)dr=J[V(|)«V\|/ + <l)V2v)/]dF (A.7.5)

Reversing the roles of <>

j and \\i and subtracting the result from Eq. A.7.5
gives

<f [<])V\|/ - yV<t>] • dS= j"[(|)V2\|/ - \|/V2<|)ld^ (A.7.6)

For the special case of an oscillating charge, the defining equation for
Green's function, Eq. A.5.1, in three-dimensional form satisfies an equation
similar to that of the Helmholtz wave equation. With point r'(x',y',z')
representing the source position:
348 The Electromagnetic Origin of Quantum Theory and Light

V 2 G(/%/)+ £2G(r,r') = -d(r,r') (A.7.7)

In free space, the solution is:

-/#•(/•-/•')
a y) e <A78)
^ '—^^r -'
The objective is to construct a virtual sphere about a source then to
calculate the fields at an arbitrary field point, r(x,y,z), in terms of the fields
that exist on the surface of the virtual sphere. In this way, the fields can be
obtained without knowledge of the source itself. For this purpose, begin by
substituting into Eq. A.7.6 that 0 = G and \|/ is equal to one component of the
electric field intensity. Then repeat twice with \j/ representing the other
electric field components and sum over the three equations. The result is the
vector form of Eq. A.7.6:

| {E(r')[rt • V G(/%/)] - G(r,r')[fi •V]E(r'j\dS'

rr o n (A-7-9)
= j[E(/-')V 2 G(/-,/)- G(/-,/-')V2E(/-')]d^'

Next, let the field point be in the vicinity of the source and construct a
virtual sphere with a radius just large enough to contain both source and field
points. Substituting Eqs. A.7.3 and A.7.7 into the volume integrals of
Eq. A.7.9 results in:

JE(r')b(r,r')Ar'=E(r) (A.7.10)

The second term in the surface integral of Eq. A.7.9 may be written:

G(«'»V')E = ^ » V ' ( G E ) - E ( « ' » V ' G ) (A.7.11)

Combining shows that:

E(/-) = |{2E(/-')[//»V'G(A,/-')]-[«,«V'][G(/-,/'')E(/-')]}a\5" (A.7.12)

Appendix 349

This equation expresses the field intensity at the field point in terms of field
values on the surface of a virtual sphere surrounding both the source and the
field. Although Eq. A.7.12 expresses the electric field intensity at the field
point in terms of values on the surface of an external virtual sphere, it is not
satisfactory since the divergence operation contains derivatives of the
electric field intensity.
Next, consider the field point to be outside the virtual sphere and, for
completeness construct a second virtual sphere. It is concentric with the first
one and of radius large enough to contain both source and field positions,
and it contains no other sources. With no sources, the volume integral of
Eq. A.7.9 is equal to zero. The integral similar to Eq. A.7.12, for this case, is
equal to zero:

<f{2E(/-')[«,«V,G(/-,/-')]-[«'«V'][G(/-,/-,)E(/)]}dS,'=0 (A.7.13)

The integral is taken over both the inner and outer spherical surfaces, with
the normal direction always extending outward from the field containing
volume. Next, let the radius of the outer surface increase without limit, in
which case, as discussed below, the integral over the exterior surface is equal
to zero. Comparing Eqs. A.7.12 and A.7.13 then shows that the surface
integral over the inner surface is equal to -E(r).
To restate Eq. A.7.12 in a way that involves field vectors only note that
the last term of Eq. A.7.13 may be written as a volume integral:

| ri »V (GE)cLS,= j V' 2 (GE)d^

Substitute vector (GE) into the vector identity:

V 2 (GE)=V[V»(GE)]-VX[VX(GE)]

Combining and using the second and third integrals of Eqs. A.7.4 to return to
surface integrals, Eq. A.7.13 goes to:

<f{2E(//'.V'G)-«'[V , «(GE)] + /?'x[V,x(GE)]}(L5,= 0 (A.7.14)

Completing both the divergence and curl operations and using the Maxwell
equations to substitute for vector operations results in:
350 The Electromagnetic Origin of Quantum Theory and Light

§{2E(ri»VG)- ri(E»VG)- AarixB + rix(VGxE)}dS=0 (A.7.15)

Substituting for the triple product and simplifying results in:

|{-/Co(« , xB)G + (//'»E)V'G + (« , xE)xV'G}cLy=0 (A.7.16)

By Eq. A.7.13, the value of the surface integral is equal to -E(r).

However, as defined above the normal is from the field-containing region to
the source-containing region. Reversing the direction so the normal extends
outward gives:

E(r) = j{-Ao[n'xB(r')]G(r,^) + [^»E(r')]VG(r,r') + [^xE(r,)]xVG(r,r')]dS

(A.7.17)

The requirements are that the field point is external to the contained region
and the source is fully contained by it.
This equation, when combined with Eq. A.7.8, is the exact expression
for the exterior electric field intensity in terms of the surface fields on a
source-containing region. It is not necessary to know anything about the
source other than it created the surface fields. Although both the external
electric field intensity and the fields on the surface are unique, the inverse is
not true: the source necessary to produce E(r) is not unique.
The corresponding expression for the magnetic flux density follows in a
similar way. Carrying out the calculation gives the result:

B(r) = | {/to[ie[«,xE(/-')]G(/-, /)+[/?' •B(/)]V G(A, / ) - [/?'xB(/)] x V G(r, SJfa

(A.7.18)

With static sources Green's function decreases as the inverse of the

radius, the electric field intensity decreases as the inverse square of the
radius, and the surface area increases as the square of the radius. Therefore
in the limit as the radius becomes infinite the contribution to field intensities
E(r) and B(r) due to the outer surface goes to zero. On the other hand, for
dynamic sources, Green's function and the electric field intensity both
decrease as the inverse of the radius and the surface area increases as the
square of the radius. From this point of view, contributions of the outer
surface integral to E(r) and B(r) remain constant in the limit of infinite
radius. That the null result remains, however, may be seen by application of
Appendix 351

a dynamic boundary condition: The radius of the outer sphere is greater than
the speed of light, c, times whatever time is of interest in the problem. Even
with dynamic sources, the outer surface integral has no influence on fields
E(r) or B(r).

A.8 A Series Resonant Circuit

An important special case is a series arrangement of inductor L, resistor R,
and capacitor C. The differential equation relating the current and voltage is:

v(/)=L^ + Ri(/)+I/i(/)d/ (A.8.1)

To obtain the steady state solution define the current in the circuit to be:

i(/)=I 0 cos(u>/) (A.8.2)

Combining Eqs. A.8.1 and A.8.2 shows that the voltage across the circuit is:

( 1^
v(/) = I 0 Rcos(a)/)-I 0 coL sin(co/) (A.8.3)

The power into each element is:

p(/) = v s (/)i(/) (A.8.4)

The voltage across the element in question is vs(/). The power into each
element is:

r \T T T^
PL(0 = r-fi-sin(2(B/) pc(0 = ^sin(2cof)
l
^ (A.8.5)
= RI 1+cos
PR(0 -y-[ (2atf)]

The energy stored in each reactive element is:

352 The Electromagnetic Origin of Quantum Theory and Light

W L ( 0 = — [l+cos(2a>*)]
(A.8.6)
T2

W c (W0 = —4-[l-cos(2a>f)l
4co2CL V n

It would be convenient to relate the voltage and current by a

multiplicative constant. Comparing Eqs. A.8.2 and A.8.3 shows that this
cannot be done with trigonometric functions. However, adding an imaginary
term to Eq. A.8.2 gives the exponential function:

1(0/
i(t) = I0[cos(atf)+ i sin(co?)] = I 0 e" (A.8.7)

Equation A.8.7 is the phasor form of the current. The phasor form of the
circuit voltage is:

'(<) = R + i coL--coC I0e'iwt (A.8.8)

By definition, the input impedance, Z, of the circuit is equal to the complex

voltage-to-complex current ratio at the circuit terminals. For this case:

(
Z = R + / CoL- (A.8.9)
V coC

Using phasor notation the exponential is suppressed and the reader is

supposed to know it should be there. Using phasor notation in this case the
current and voltage are:

I=I0 and V=I R+/ 0)L- (A.8.10)

coC

Although phasors provide a constant multiplicative relationship between the

current and voltage, other problems arise. Each variable, i.e. phasor current
and phasor voltage, consists of terms that do and terms that do not represent
physical reality, the actual and virtual terms. Separating actual from virtual
values in the voltage and current expressions can easily be done since the
Appendix 353

actual values appear as real numbers and the virtual values appear as
imaginary ones. However, when products are taken things are not so simple.
Using power as an example, the product of the phasor current and voltage
consists of four types of terms: actual current times actual voltage, actual
current times virtual voltage, virtual current times actual voltage, and virtual
current times virtual voltage. Of these four products, only the first type
represents actuality and only it is desired. The second and third types are
multiplied by '/' and thus may be discarded. The fourth type, however, is a
real, unwanted number. Special multiplication rules are necessary to
eliminate the fourth type of product.
Consider circuit power as an example. From Eq. A.8.5 the time varying
input power is:

1
P(/) = M R [ 1 + COS(2O>/)]- coL- in(2u)/)
sinl (A.8.11)
coC

By way of contrast consider the product phasor P c = VI* 12:

Pc = f { * + c.-oL--wC1 (A.8.12)

The real and reactive powers shown in Eq. A.8.11 are phased in time
quadrature. The real part of P c is equal to the time average power. The
imaginary part of P c is equal to the magnitude of the reactive power. Since
the instantaneous value of the power is, in many cases, of no interest, the
remaining quantities of interest are both contained in Eq. A.8.12: the time
average real power, Pav> <ind the magnitude of the reactive power, Pre.
Because of these relationships, it is common when dealing with power in
electrical circuits, to work with the complex power:

"c — "av "*"l "re (A.8.13)

Although the power is complex, it is not a phasor: both real and imaginary
parts represent physical reality and there is no virtual part.
354 The Electromagnetic Origin of Quantum Theory and Light

A.9 Q of Time Varying Systems

Q is a dimensionless ratio that describes the quality of anything that
oscillates. Although developed for application to a closed system, such as an
electrical circuit, a bouncing ball, or a swaying bridge, Q is also useful for
dealing with the open system of a radiating antenna. For example, antenna Q
is important in communication antennas since modulation is essential,
modulation requires a minimum bandwidth, and there is a direct relationship
between Q and bandwidth. For high power antennas, Q is a measure of how
much energy must be stored about an antenna to obtain the minimum
acceptable power output. This is significant in such diverse applications as
the decay of atomic states and electrically small, high power antennas used
to communicate around the surface of the earth. In nearly all radiation
problems, Q is a critical measure of antenna worthiness.
Since only energy within a half wavelength of an antenna can return to it
during steady state operation, only this near field energy affects an antenna's
input impedance. To the driving terminals of an antenna, energy radiated
permanently away from the system is indistinguishable from energy
absorbed by a resistor; the power loss is therefore measured as an effective
antenna resistance. Since the energy that oscillates to and from an antenna is
indistinguishable from reactive energy, the oscillation results in an effective
radiation reactance. To the driving circuit, this input impedance is
indistinguishable from the input impedance of a properly synthesized closed
circuit. Hence, from the point of view of the driving source, an antenna may
be replaced by and analyzed as if it were an electric circuit.
Anything that oscillates can be assigned a value of Q. With W{t)
representing the energy stored in the system, the magnitude of the Q of any
system is defined to be the dimensionless ratio:

Q= V- (A.9.1)

Q is a measure of how rapidly a system grows or decays. Rewriting

Eq. A.9.1 for a lossy system gives:
Appendix 355

The solution of Eq. A.9.2 is:

^ / ) = ^5e _C0//Q (A.9.3)

WQ is the initial value of energy. As an example, consider a ball bouncing on

a smooth, horizontal, surface. In a uniform gravitational field, the energy is
proportional to the height and, at maximum height the energy is entirely due
to gravity. The maximum height reached by the ball follows the exponential
0)
decay of Eq. A.9.3, and the ratio of heights on successive bounces is e .
Inductors require a current path and current paths, generally speaking,
are also resistive. The ratio of inductance, L, to resistance, R, depends upon
the nature of the path (the wire) and its geometrical arrangement. Since the
peak energy per cycle is equal to the total oscillating energy, a definition
derived from Eq. A.9.1 is the peak cyclic value of stored energy-to-average
energy loss per radian ratio:

(affit) .
Q = -rr*- (A.9.4)

Mav

The current and voltage in an RL circuit may be written:

i(/)=I 0 cos(u)/)
(A.9.5)
v( /) = I 0 [R cos(co/) - coL sin(co/")]

The power dissipated in the resistor is:

P R ( ' ) = - ^ - [ 1 + COS(2CD/)] (A.9.6)

The energy stored in the inductor is:

LT2
^ L ( / ) = - ^ [ l + cos(2(B/)] (A.9.7)

Combining Eqs. A.9.4, A.9.6, and A.9.7 shows that:

356 The Electromagnetic Origin of Quantum Theory and Light

_. coL
Q=— (A.9.8)
K

For an RLC circuit, the instantaneous energy-to-time average power ratio

is equal to:

Ll£ r . ,„ A l . IA 2
[l + cos(2co/)] + -2-[l - cos(2co/)]
^ (A.9.9)
RI20
[l + cos(2u)/)]

From Eq. A.9.9 it follows that, if the inductive energy exceeds the capacitive
energy, Eq. A.9.8 gives Q. If the capacitive energy exceeds the inductive
energy, Q is:

Q = -^7T (A.9.10)

At the resonant frequency, where ' a v ' denotes time average values, a
commonly used formula is:

Lav
Q= (A.9.11)
"av

For these simple circuits, values calculated using Eqs. A.9.9, A.9.10,
and A.9.11 are equal at resonance. In more complicated circuits where the
reactive elements are driven with different phases, Eq. A.9.11 is not exact.
A slightly modified definition that is sometimes used with simple
systems is to equate Q with the tangent of the impedance phase angle. So
long as the system frequency is low enough for capacitive effects to be
negligible the definition reproduces Eq. A.9.8 for the simple case of an RL
circuit. Using all three definitions, results with lossy capacitors are similar to
those with lossy inductors. By all three definitions, a capacitor C in series
with resistor R simply replaces coL by 1/coC.

Q = l/(u)CR) (A.9.12)
Appendix 357

With an antenna radiating in the steady state since time t = - oo there is

an infinite amount of energy in the field. The difficulty with calculating Q is
separate the finite field that returns to the source upon shutdown from that
which does not. During steady state operation, the magnitudes of the field
intensities decrease with increasing radius. The Maxwell stress tensor shows
that radiation fields exert an expansive self-pressure equal to the gradient in
field energy density. In this way, the tensor describes forces acting to drive
the field energy ever outward. However, upon source turnoff, shutdown, the
inverse is true. If the fields vanish near the source the tensor describes
compressive forces that act to drive the field energy back to the source; it is
the returned energy that forms the numerator of the expression for Q, see
Eq.A.9.1.

A.10 Bandwidth
The normalized bandwidth is defined as the ratio of the frequency difference
between the two points at which a resonant circuit drops to half the resonant
power (half power points) divided by the resonance frequency.
By definition, the resonance frequency of a series circuit is that
frequency at which the reactive power vanishes. From Eq. A.8.3 the input
voltage is:

1
v(/)=I 0 Rcos(co/)-I 0 u)L- sin(co/) (A.10.1)
coC

From Eq. A.8.11 the input power is:

cos (2co/)]- CflL--coC sin(2co/) (A. 10.2)

The resonance frequency, coo, is the frequency at which the reactive power
vanishes. It is equal to:

2 1
0 (A. 10.3)
LC
358 The Electromagnetic Origin of Quantum Theory and Light

To begin a dimensionless analysis of bandwidth introduce the expressions:

cOi=Q)o(l + 5)anda) 2 = u)o(l-8) (A.10.4)

Frequencies coi and Gt>2 a r e the half power frequencies. With no loss of rigor,
it is convenient to have the circuit be subject to a constant current input. For
that case, the dissipated power drops by half when the real and reactive
voltage magnitudes are equal. This occurs for:

co0L 5(2+5) _
-^— = -T—^ = 25 (A.10.5)
R (1+8)

The bandwidth is 28coo. A result of combining Eqs. A.10.4 and A.10.5 is:

Q=—= — (A. 10.6)

v
28 (0 1 -co 2 '

For more complicated circuits, the actual circuit may be replaced by its
equivalent Thevenin or Norton circuit and analyzed in a similar way. For
structured circuits in which different passive elements have differently
phased driving currents the inductive and capacitive energies are not in
phase quadrature. The peak value of stored energy contains contributions
from both inductors and capacitors.

A. 11 Instantaneous and Complex Power in Radiation Fields

To compare methods of describing power, for steady-state radiation fields
return to the Maxwell equations, Eqs. 1.6.8 and 1.6.11:
-YD
VxE + — = 0 V«B = 0
' (A.11.1)
VxB-u-E — = u J eV»E = p
at

Replacing B by u,H and integrating over any closed volume gives:

Appendix 359

lN»dJ,+ f ( ^ H « ^ + e E » ^ | d ^ = - [ E » J d ^ (A.11.2)
J 3
\ dt dt)

This is the Poynting theorem. It is a restatement of Eq. 1.9.9 with vector N

defined by Eq. 1.8.5. The term on the right side of Eq. A.l 1.2 is power into
the system. The volume integral on the left is the rate energy enters the field,
and the first term on the left is the rate at which energy leaves through the
surface. This interpretation is independent of the wave shape.
When dealing with sinusoidal, steady state radiation it is often
convenient to use phasor notation. With expO'cttf) time dependence the phasor
version of Maxwell's equations is:

VxE+AoB = 0 V«B = 0
(A.l 1.3)
V x B - /u)[ieE = fAj eV • E = p

The integrated results corresponding to Eq. A.11.2 are:

^N c «dJ , + — j ( j i H 2 - e E 2 ) d F = - - j E « J * d ^ (A.l 1.4)

By definition the complex Poynting vector is:

Nc = E x H * / 2 (A.11.5)

The same volume of integration is chosen for Eq. A. 11.4 as was chosen to
evaluate Eq. A.11.2. Although the complex Poynting vector is a complex
quantity, it is not a phasor, since both real and imaginary parts represent
actual quantities.
The real part of the volume integral on the right of Eq. A.l 1.4 is equal to
the time average power into the region. Since the volume integral on the left
has no real part, the real part of the surface integral must equal the time
average output power through the surface. The volume integral on the left
side is 2co times the difference between the time average electric and
magnetic energies in the volume of integration.
Consider the volume of integration to be spherical, let 8 be a vanishingly
small distance, and place the source currents on the sphere. Consider three
360 The Electromagnetic Origin of Quantum Theory and Light

volumes: {1} Inner concentric, virtual spheres containing all radii less than
a-8. In this region, there are no sources and all fields are ignored. {2}
Concentric spheres of radii within the range a ± 8. In this region are all of
the sources. {3} An exterior region containing all radii larger than a + 8. In
this region, there are fields but no sources.
Since the exterior region contains no currents, within that region the
current-containing integral of Eq. A. 11.4 is equal to zero. Since the volume
integral on the left has no real part the sum of the real part of the surface
integral taken at infinity and at a + 8 is equal to zero. For finite fields, in the
limit of infinite radius the imaginary portion of N c decreases more rapidly
2
than \lr with increasing radius and, hence, the imaginary part of the surface
integral is equal to zero at infinity. Therefore the imaginary part of the
surface integral at radius a + 8 is equal to (co/2) x {the difference between
the time average magnetic and electric field energies} within the volume of
integration:

Im[|Nc«djj =-j{eE«E*-iiH«H*}d^ (A.11.6)

Within the source region, a ± 8, since the volume is proportional to 28

and there are no singularities in the fields the volume integral on the left of
Eq. A. 11.4 goes to zero with 8. This leaves:

Im[^N c »d i $] + - I m [ j E » J * d ^ ] = 0 (A.ll.7)

For simplicity consider the special case where radius a « X, an electrically

small antenna. Sources may then be considered as circuit elements and the
current containing integral of Eq. A.l 1.4 may be written as:

k
lr 1
-jE.J*d^-XVjI* (A.11.8)
L Z
j=l

With more than one current source, unless the sources meet one of the phase
conditions of Eqs. 3.1.13 powers do not combine by simple addition.
Appendix 361

Therefore, unless that condition is met Eq. A. 11.8 does not correctly
describe the complex power. If not, the right side of Eq. A. 11.4 is not the
complex power and, if it is not, neither is the surface integral of Eq. A.l 1.4.

A.12 Conducting Boundary Conditions

Let an electric field intensity exist in the vicinity of a smooth boundary about
a closed volume that is immersed in the field. The volume is arbitrary in size
and shape. Requirements on the volume are that its size be much less than a
wavelength, in all three dimensions, and that it includes regions on both
sides of the boundary. Apply the condition of Eq. 1.6.8, that:

V»E = p/e (A.12.1)

Evaluate the integral of Eq. A.12.1 over the volume in question, with the
result:

Jv»Ed^=|E»d$'=q/e (A. 12.2)

Symbol "q" indicates all charge within the volume. Next, let the dimension
normal to the boundary become vanishingly small on both sides of the
boundary so the shape approaches that of a disc. The contribution to the
surface integral due to electric field intensity normal to the disc is thereby
vanishingly small. Thus, only the normal components of the field intensity
are of interest, and Eq. A. 12.2 goes to:

j>E»dS=p/e (A. 12.3)

Symbol p indicates the charge per unit area at the interface. From
Eq. A. 12.3, since charge density in free space is equal to zero the normal
component of the electric field intensity on any virtual boundary is
continuous.
If an electric field intensity existed inside a nearly ideal conductor, it
would drive a nearly infinite current density that would, in turn, absorb a
nearly infinite amount of power. Therefore an electric field intensity inside
an ideal conductor is zero. In turn, if an electric field intensity is applied
362 The Electromagnetic Origin of Quantum Theory and Light

normal to the surface of an ideal conductor, by Eq. A. 12.3 it is equal to the

charge density on the surface normalized by the permittivity of free space.
For a magnetic field intensity, by Eq. 1.6.11:

VxE = -—- (A.12.4)

at
Consider the surface integral of Eq. A.12.4 over an open area that, like the
volume of Eq. A.12.1, extends on either side of a smooth boundary. The
integral is:

r rdB
*E»dl = - — "cL*1 (A.12.5)
J J
ot
The symbol At indicates differential distance along the periphery of the open
area. Next, let the dimension normal to the boundary become vanishingly
small. In this limit, the open area becomes vanishingly small and since B is
finite, the entire right side of the equation is vanishingly small. Therefore,
the line integral of the electric field intensity around the loop is equal to zero.
Since the length of the loop is the same on either side of the boundary, and
since dl is oppositely directed on either side, the tangential component of the
electric field intensity is continuous through virtual boundaries. At a
boundary between free space and a conductor since the electric field
component inside the conductor is equal to zero it is also equal to zero just
off the conducting surface.
Let n be a unit vector normal to a smooth surface. For a virtual surface,
the boundary separates regions one and two. For a conducting surface, the
field is in the free space region only:

Virtual Surface
/ Z x ( E 1 - E 2 ) = 0; *.(E!-E2) =0
(A. 12.6)
Conducting Surface
« x E = 0; «»E = p s / e

Consider a closed volume that is immersed in a magnetic field. Like the

volume considered for the normal component of the electric field intensity, it
may be arbitrary in size and shape. The requirements are that its size be
Appendix 363

much less than a wavelength, in all three dimensions, and that it includes
regions on both sides of the boundary. Apply the condition of Eq. 1.6.11:

V«B = 0 (A.12.7)

Take the volume integral of Eq. A.12.7 over the volume in question, with the
result:

(fB-di^O (A.12.8)

Since Eq. A.12.8 applies to a closed volume, let the dimension normal to
the boundary become vanishingly small so the shape approaches that of a
disc. It follows that the normal component of B is continuous through the
boundary. Inside a conductor, B is constant since otherwise Eq. A. 12.5
shows that it would produce an electric field intensity there. For time varying
radiation fields, the normal component of the magnetic field intensity is
equal to zero.
ByEq. 1.6.8:

8E
V x B = ne — + n J (A.12.9)
at
Consider the surface integral of Eq. A.12.9 over an open area that includes a
smooth boundary. The integral is:

$B»dt = nej — •dS+^J»dS (A.12.10)

Let the dimension normal to the boundary become vanishingly small. In this
limit, the open area becomes vanishingly small and, since E is finite, the first
term on the right side is vanishingly small, leaving:

JB»<M=ii$J»dS=iils (A.12.11)

Symbol Is is the total electric current I that flows through the open area. In
free space, there is no current and the line integral of the magnetic field
intensity around the loop is equal to zero. Since surface currents may exist
on conductors, the tangential component of a time varying magnetic field
364 The Electromagnetic Origin of Quantum Theory and Light

just off the surface of a conductor is equal in magnitude and perpendicular in

direction to the current per unit length on the surface:

//xB=uIs (A. 12.12)

Summarizing boundary conditions:

Virtual Surface
/ 7 xV( B , - B 2 ); = 0; / ? »V( B
i
, - Bl}2 ) = 0
(A.12.13)
Conducting Surface
nxB-iils; /?»B = 0

Both field vectors are continuous through a virtual surface. On

conducting surfaces the tangential component of the electric field intensity
and the normal component of a time varying magnetic field intensity are
both equal to zero. Just off the conducting surface, the normal component of
the electric field intensity is equal to the surface charge density and the
tangential component of the magnetic field is equal to the surface current
density in amperes per meter.

A.13 Uniqueness
If, within a given boundary, a potential reduces to the correct value on the
boundary, or to the correct normal derivative of the potential on that
boundary, then that potential is unique. This theorem justifies the use of
arbitrary solution methods so long as the resulting solution obeys Laplace's
equation in the charge-free regions. No matter how the solution is obtained,
if it satisfies these conditions the solution is unique.
Taking (|>V(|> to be a vector field and substituting into the divergence
theorem gives:

J" 4>V(|) •&$'=]' V»((()V<l))d^=jf(V(t))2 + (|)V24)|dr (A.13.1)

Appendix 365

Since Laplace's equation is satisfied, the last term is equal to zero. Suppose
§\ and (|>2 are different potentials that have either equal values of potential or
normal derivatives thereof on every conductor in the field:

J > 1 -<hM4>i - < h ) « d ^ = J"[v(<t>i -<t> 2 )fd^= 0 (A.13.2)

Either equality at the conductors requires the surface integral to equal zero.
Since the volume integral is equal to zero it follows that the integrand is
equal to zero everywhere. Therefore the potential and/or the electric field
intensity are equal everywhere in the field and, therefore, the functions are
the same.
For time-dependent solutions, it is only necessary to substitute functions
\|/l and \|f2 of Section 1.12 into the divergence theorem and repeat the above
procedure.

A.14 Spherical Shell Dipole

Consider a spherical shell of radius a that supports a static surface electric
charge density q:

p = —^-r-cosG (A.14.1)
Ancr

The charge creates an electric field intensity in both the interior and exterior
regions about the shell. The form of the external fields follows directly from
Eq. 1.12.8, with all coefficients equal to zero except the coefficient with
order one and degree zero F(1,0). In the limit of zero frequency the external
field components are:

E r = -^F(l,0)cos6 E e =-4-F(l,0)sin6 (A.14.2)

a a
The internal fields follow from the same equation set. To avoid functional
singularities at the origin with interior region fields it is necessary to replace
spherical Hankel functions by spherical Bessel functions. Writing F'(1,0) as
366 The Electromagnetic Origin of Quantum Theory and Light

the internal coefficient, in the limit of zero frequency the internal field
components are:

E r = -F'(l,0)cos8 Ee=--F(l,0)sine (A.14.3)

The boundary conditions on the fields are:

C \
= 2eF(l,0) •|F(I,O)
4na v(M; (A. 14.4)
i
F(1,0) = -F(l,0)-

From Eq. A. 14.4:

(A. 14.6)
\2nza %neaA

Combining shows the fields to be:

Exterior
2«q aq . n (A. 14.7)
Er = r-cosG E
\2mr 9= -sin9
Interior 12ner
q
E r =" 2 cos9 Ee -sinG
\2nea ' Ylvza

The interior field is constant, and equal to:

E=2EirandEz = ^_ (A. 14.8)

\2%Z(T

The exterior field is that produced by a z-directed electric dipole moment of

magnitude:
Appendix 367

pz = aq/3 (A. 14.9)

The energy densities are:

Interior Exterior
2
* 6 " V (l
/, + „3cos22ne)
\ (A 14J0)
-

The total field energies are:

Interior Exterior
2 2 (A.14.11)
W T =— W T =—
216TO^ 1087ie^
Equation A.14.11 shows that respectively one third and two thirds of the
energy is stored interior and exterior to the shell.

Spherical Harmonics

A.15 Gamma Functions

Products in which successive factors differ by one occur frequently in the
formation of power series. If t is an integer, such products may be expressed
as special products of a certain number of integers, beginning with one. The
factorial of integer i is, by definition:

*!=1»2«3»...»* (A.15.1)

It follows that

a=e(t-i). (A. 15.2)

368 The Electromagnetic Origin of Quantum Theory and Light

The same symbolism is useful for noninteger numbers, v. A similar equation

is defined:

v!=v(v-l)! (A.15.3)

Similarly:

V
v(v-l)(v-2)...(v-/g+l)= , ' , (A. 15.4)

(y— m)\

Let f(v) be any function that satisfies the condition

f(v) = v f ( v - l ) (A. 15.5)

Taking the ratio of Eq. A. 15.5 to A.15.3 shows that:

* ( v ) = ^ =| ^ = <Kv-l) (A.15.6)

It follows that 0(v) is a periodic function of period v.

Euler proposed that the definition of a noninteger factorial be:

oo

v!=JVe _/ d/ (A.15.7)
0

This definition is valid over the range

-1<V<0 (A.15.8)

Factorial v! can be evaluated for any value of v using Eq. A.15.3. For
example, if-2 < v < -1 then v! can be written

, (v+1)! (v+l)!(v+2)! (v+l)!(v+2)!...(v + n)!

v!=- - =- — - =- — — (A.15.9)
v+1 (v+l)(v+2) (v+l)(v+2)...(v+n)
Appendix 369

Equation A. 15.9 shows a simple pole exists for v equal to a negative integer.
Other results of Eq. A. 15.7 are that:

0!= 1=1! (A.15.10)

(-1/2)!= Vit (A.15.11)
(-v)!(v-l)!=—f-^ (A.15.12)
' sin(nv)

The Stirling formula for the approximate value of v!, in the limit of large
values of v, is:

v!= - V2nv
ttV (A. 15.13)

A related and frequently recurring product form is with succeeding

numbers that differ by two:

v(v-2)(v-4)(v-6)...

The series is denoted by:

v!!=v(v-2)(v-4)(v-6)... (A.15.14)

It follows that for even and odd integers, respectively:

(2/)!=2'(4 and (2/+l)!!=^±j4 (A.15.15)

The left and right equations of Eq. A.15.15 for £ - 0 show that:

(0)!!=1 and (l)!!=l (A.15.16)

Although Eq. A.15.15 is in indeterminate form for l = -\, evaluating the

identity:
370 The Electromagnetic Origin of Quantum Theory and Light

j-f- = (t-l\l (A.15.17)

(*)!! V '

For t = 0, Eq. A.15.17 shows that

(-l)!!=l (A.15.18)

Table A. 15.1 contains a listing of useful and selected sums involving

factorials.

J, {i + m-\)\{l-m-\)\\ , w N

4
£A (<+»)!!(<-»)!! < ' (<-l)!!

V 47
^ 2 *(* + l)(*+ »)!!(*-/»)!! *!!

4. 4 V ^ £ w r-U(/»)5(*+z»,2q + l) = l
^ 0 < ^ + l X ^ + «-l)!!(^-/w-l)H V ^

5. 4V—^ T7—^- r^-8(^+w,2q) = l

6 8
^0 ,(A 1)(,+^: onci-L- DM 5 ^ + - 2 ^ + ^ u ^ ) - ( 7 ^
v M
^ ^ + 1)(^+W-1)!!(^-OT-1)!! ' ' *!!

8 ^-^')/2 l(-l) s 5(l + ^,2q + l) _ l

s=0 (^ + ^+2s+l)!!(f-w-2s-l)!! {l +m-\)\\{l-m-\)\\

Table A. 15.1 A Table of Sums over Factorials

Appendix 371

A.16 Azimuth Angle Trigonometric Functions

The solutions of Eq. 1.11.11 are trigonometric functions:

<D((t)) = £[C w cos(zw|>)+ D OT sin(H))] (A.16.1)

m
An equally satisfactory solution is:

(A. 16.2)

By definition y = (-1). For cases of interest here, the azimuth angle occupies
the full range of angle from 0 through 2TT. This condition requires the
solution to satisfy the relationship:

$(<)>) = 0(<|> + 27t) (A. 16.3)

Equations A.16.1 through A. 16.3 are jointly satisfied only if m represents the
full range of positive integers, including zero.
The trigonometric functions form an o r t h o g o n a l set. Trigonometric
identities show that:

2% 271

J-
j cos(//«]))cos(/z(|))d(|) = — J d<)>(cos[(z»-/7)<|)] + cos[(/0+#)<|)]) (A. 16.4)
0 "0

Evaluating the integral on the right gives:

2JI
sin[(w- //)(])] sin[(//7+ /?)([)] 2%
J
j cos(/w|))cos(MJ))d(|) =
2{jn- n) 2(m+ n)
J
0
(A.16.5)
0

Since both m and n are positive integers, the second term on the right of
Eq. A.16.5 is always zero; the first term is also positive unless m=n, for
which case the result is indeterminate. Evaluation may be accomplished by
either evaluating the indeterminate or by substituting into the integrand the
identity:
372 The Electromagnetic Origin of Quantum Theory and Light

cos (m<^) = — [\ + cos(2m§)\ (A. 16.6)

Integrating Eq. A. 16.6 gives:

2JI 1 2TC
j d<|)cos (m§) = — j d<|>[l + cos(2ztf(|))] = n (A.16.7)
2
0 0

Combining Eq. A.16.5 through A.16.7:

271
j cos(,w(|))cos(/z<]>)d<|) = 7i5(w,/z) (A. 16.8)
0

The Kronecker delta function is indicated by 8(w,«). By definition:

1 if m= n
8(m,n) = (A. 16.9)
0 if m± n

It follows in a similar way that:

271
J sin(/w)))sin(M)))d(|) = nb(m,/i) (A.16.10)
0

Also

271
j sin(//«|))cos(/2<|))d<|) = 0 (A.16.11)
0

Combining Eqs. A. 16.8 through A. 16.11 gives:

2%
(A. 16.12)
0
Appendix 373

For example, Eq. A. 16.12 may be used to evaluate the product function:

2n 27t
J ^"-"M sin<i>d(j> = — J" UW-»*W -e-Z'^-^D^W
2 y
0 0 (A.16.13)
= n^>{/ri,m+1) - d(m',m-1)]

It is often convenient to express functions in terms of the order of

trigonometric functions. The formulas for going from power to order follow
directly from the geometry; all possible combinations are given by the four
sums:

1 ^ 2(2*)! r , x , (2*)!
cos <|, = -L.J V v
{ cos2(l-k)|) +-^4 (A. 16.14)
2 L l W 2
2 M k f 0 (2^-k)!k! (*!)
• 2^ x
sin
^^27
1
s
k=0
(A. 16.15)

e-1 (2i-l)!
2^-1 A - cos[(2*-2k-l)0] (A. 16.16)
cos • • 32?2Mkfo(2^-k-l)!k!
S
Lk=0
4
• 2^-1 •
sin
••^|SHr-'^^s*««-k-.w (A.16,7)

An expansion for l/(sin(j)) is necessary to accomplish needed

calculations. To form the expansion, note that since l/(sin<|)) is an odd
function of 0 it is expressible as:

1 °°
^—=XAssin(2s-l)<|) (A. 16.18)

To evaluate coefficients A s , multiply Eq. A. 16.18 by sin(2p-l)(j)and

integrate over the full range of the variable. The result is:
374 The Electromagnetic Origin of Quantum Theory and Light

271 ' (n _IVK °° 2JI

j d<$>Sm[P ^ = X A s J d<>sin[(2s-l>|>]sin[(2p- 1>J)] = nA p 5(p,s)
Sm(P
0 s=l o
(A. 16.19)

The left side of Eq. A. 16.19 is a periodic trigonometric function for all terms
except p = 1 and all periodic terms integrate to zero. This leaves:

? ,, sin(2p-l)d) „
d(j)—i-£—£- = 27i (A. 16.20)
J sin(j)

Substituting back into Eq. A. 16.18:

1 °°
= 2Vsin(2s-l)<|) (A. 16.21)
sin(
t> s =l

A similar sum over cosine terms is equal to zero:

oo

£cos[(2s-l)<|>] = 0 (A. 16.22)

s=0
It follows that:

— = 2yYe-^2s+1)* (A.16.23)
sin(() s=0

A. 17 Zenith Angle Legendre Functions

The easiest way to obtain the general solution of Eq. 1.11.10 is to first solve
the special case m = 0, for which case the equation is:

d20 „d0
2
+ cot9— + v(v+l)0 = O (A.17.1)
de de
Appendix 375

The character of separation constant v depends upon the boundary

conditions applicable to the region in which the equation is applied. In the
case of spherical waves in free space, for example, v is an integer, and
denoted by v= L For lossless waves in conical structures v is real and
noninteger. If there is loss, v is imaginary. In this book since only lossless
problems are considered v is real in all cases.
Since Eq. A. 17.1 contains a singularity on the polar axes, the character of
the functions ©(0) in the region away from the axes are of special interest.
Rather than go directly to a solution, consider first the solution form in the
region of interest. A useful substitution is:

1
G:
1/2
G (A. 17.2)
(sine)

Differentiating:

dG 1 dG cos0
Q
dG (sin9) "2 de 2(sin6) 3/2
(A. 17.3)

d2Q d2Q cos6 dG 3cos 2 6

Q
de2 (sine) 1/2 dQ2
2(sin6) 1/2
N.5/2
(sine) 3/2 d e 4(sin0):'/

Combining Eqs. A. 17.1 through A. 17.3 gives:

2
d2Q \
- + v+- +-|i+core (A.17.4)
d9 2

For 6 near nil, cot 0 is much less than one and Eq. A. 17.4 is nearly equal to:

d20 ( IY i
- + v+ + - G =0 (A.17.5)
de^ v 2 y
376 The Electromagnetic Origin of Quantum Theory and Light

Solutions of Eq. A. 17.5 are:

e(9) = rjTj [ A v COS(K V 9)+ B V sin(Kv0)] (A. 17.6)

(sin6)'

By definition:

1/2
n2 i
*v = v+- +— (A.17.7)
2 4

Near the equator the zenith angle functions are trigonometric functions of
(K V 9) normalized by the square root of sin8. The interval over which
Eq. A. 17.6 is valid increases with increasing values of v.
To examine the solution near its singularity, begin near the positive z-
axis, where

cot0=- (A. 17.8)

0

Equation A. 17.1 has the form:

d 2 0„ + 1 d 9 + v ( v + l ) e = 0 (A. 17.9)
de2 ede v ;

Equation A. 17.9 is the cylindrical Bessel equation, for which the solution is:

e(e)=A v j 0 (pe)+B v Y 0 ((39) (A.17.10)

By definition

p = [ v (v+l)] 1 / 2 (A.17.11)
Appendix 377

Jo(P6) and Yo(P9) represent, respectively, cylindrical Bessel and Neumann

functions of zero order. In the limit as 0 goes to zero, Jo(P0) goes to unity
and Yo(|30) becomes logarithmically singular:

G(0) = C v Y 0 (0) (A. 17.12)

By symmetry, near the negative z-axis the function takes the same form. The
local solution is:

e(7i-e)=A v Y 0 [p(7i-e)]+B v Jo[P( 7I - e )] (A. 17.13)

Combining shows that:

G(7i)=D v J 0 (0) + C v Y 0 (0) (A.17.14)

It follows that if one solution of Eq. A.17.1 is Pv(cos9), the other is

Pv(-cos0). The first solution is regular on the positive z-axis and singular on
the negative z-axis and the second solution is singular on the positive z-axis
and regular on the negative z-axis. Both functions are periodic, see
Eq. A. 17.6, in the center region. The full solution is:

0(e) = A v P v (cos8)+B v P v (-cos0) (A. 17.15)

For values of m different from zero it is convenient to rewrite Eq. A.17.1

by introducing the variable % where:

X = - ( l - c o s 6 ) = sin 2 (0/2) (A. 17.16)

Derivatives are:

d6_ d 6 d(cos8) dO
dx d(cos0) dx d(cos0)
? •> r / o2 9/ (A. 17.17)
d20 _ d20 d(cos0) d 9 d 2 (cos6)
dx 2 d(cos0) 2 L d5C J d(cosG) dx 2
378 The Electromagnetic Origin of Quantum Theory and Light

Combining Eqs. A. 17.16 and A. 17.17 with Eq. A. 17.1 gives:

X(1-JC)—5- + (l-2jc)-r- + v(v+1)6 = 0 (A.17.18)

d% dx

A power series expansion results in:

e(x)=SajXj
j=o
a J_1
—0(x)=X
j=oj J X (A.17.19)
2
d j-2
2
©(x)=5>jj(j-i)x
,
dX "" j=o

Combining Eqs. A.17.18 and A.17.19 results in the equality:

X [ ( j + l ) 2 a j + 1 - j ( j + l ) a j + v(v+l)ajjx j =0 (A.17.20)
j=0

Since Eq. A.17.20 is an identity in %, the coefficient of each power of x is

separately equal to zero. It results in the recursion relationship:

a
i+l j(j+ l ) - v ( v + l )
-£L = £±—'- \ '- (A.17.21)
a
J (j+1) 2
If v is an integer, the numerator of Eq. A.17.21 is equal to zero at v = j and
the series terminates. Substituting Eq. A.17.21 in the first series of
Eqs. A.17.19 results in the solution:
Appendix 379

Pv(cose)=£^±llsin2JA (A.17.22)
j=o(j)! 2 (v-j)! 2

Pv(-cos9) is also a solution; changing the sign in Eq. A.17.17 gives:

X = - ( 1 + cosG) = cos 2 (6/2) (A. 17.23)

Combining Eqs. A.17.22 and A.17.23 gives the second solution:

P v (-cos0) = £ ( " 1 ) ? J ( V + J ) ! cos 2 J(6/2) (A.17.24)

j=o(J)! ( v - j ) !

Neither Eq. A.17.22 nor Eq. A.17.24 is a totally even or odd function of 8.
Since it is convenient to work with equations of definite parity, it is
convenient to define the new functions:

L v (cos0) = -{P v (cos6)+ P v (-cos0)}

(A.17.25)
M v (cos 0) = - {Pvv(cos 0) - Pvv(-cos0)}
(cose)-P (-cos 9)}

The functional symmetry is:

L v (cos0)=L v (-cos0); M v (cos6) = - M v ( - c o s 0 ) (A. 17.26)

Since the regions of convergence for Pv(cos0) are -1 < cos8 < 1 or 0 < 0 < n,
the region of convergence for Lv and Mv are 0 < 0 < n.

A.18 Legendre Polynomials

To solve problems with the z-axis is included in the region where a solution
is necessary, all functions must remain bounded at the endpoints: 0 < 0 < n.
This happens only if separation constant n is an integer and the series
380 The Electromagnetic Origin of Quantum Theory and Light

solution of Eq. A.17.21 terminates. For that case, both Eqs. A.17.22
and A. 17.24 remain bounded on both the ±z-axes, but are not independent.
Since the product £{£+\) is the same for £ = n as it is for £ =-(n+l),
solutions are the same for the range of integers respectively from 0 to +oo
and from -1 to -oo. Therefore, only positive values off need be considered.
To characterize Legendre polynomials, it is more convenient to redo the
expansion than to work from the existing solutions. For this purpose, rewrite
Eq. A. 17.1 by replacing noninteger v by integer £ and defining % = cos8, to
obtain:

2
(l- X )^f-2x^ + ^ + l ) 0 =O (A.,8.!)

The power series expansion is:

oo

0
Oc)=Eaj?cJ

Q
5C j = 0

Combining gives:
oo

2 [(J +1)( j + 2)aj_ 2 - j(j + l ) a j + ^(^+ Daj]z j = 0 (A.18.3)

j=0

Combining:

a j + 2 _ , j ( j + l H ( l + l) (A. 18.4)
*j 0+1)0 + 2)
Combining Eq. A. 18.4 with the expansion for 0(%):
Appendix 381

*M , y(-l)JX2iOT'Kt+2)-l)!! ,.,„,,
e tt) <A 185>
' " " I (2J)!tf-W-2j)« '
Examples are:

.2.35 ^
© o ( x ) - a 0 ; e,(x) = a 0 ( l - 3 x 2 ) ; e 2 ( X ) = a 0 l-10% 2 + - x ' (A. 18.6)

ao is an arbitrary constant and is redefined for each value of £ to make

0^(0) = 1. With that definition, functions ©^(%) are defined to be Legendre
polynomials of the first kind, and indicated by P^(cos8). Therefore P^(%) is
given by the series:

P , (x) l [ f H£ W-W x <-» (A , 8 . 7)

2e s=0
s!
(*-s)!(*-2s)r

The symbol [112] indicates the largest integer contained in ill. From
Eq. A. 18.7, it follows that:

P/(l) = l ; P 2 / + i(0) = 0; P 2/ (0) = ( - l ) ' ^ j ^ ; P/(-Jc) = (-l)'P*U)

Values of the important functional combinations on the axes are shown in

Table A. 18.1.

Pf(cos9) dPJ(cos6) PJ(cos6)

d9 sin0
6=0 l
-t{t + l) -t(t + l)

e=* (-i)< I/(/+i)(-i)«+1>/2 l/(/+i)(-i)«-1)/2

^11 /VlMitoA /^^5(2q,£) /-l-W!L8(2q+W)

fl! (*-2)H (^-1)!!

Table A. 18.1 Values of Selected Functions on the Axes

382 The Electromagnetic Origin of Quantum Theory and Light

For even values of £, the expansion may be written:

„ / \ 1 d^ <£ ( - l ) V ) ! 2<?-2s
PfW=i 7EA,;/ \,X (A-18-8)
The binomial expansion is:

(X2-i)' = y (-^'w x2/-2. (A189)

Combining results gives another expression for Legendre polynomials:

'M'Tfi^?-1)' ,A 18,0)
-
Equation A.18.10 is the Rodriques formula for Legendre polynomials.
Comparison using Eqs. A.17.27 and Eq. A.18.9 shows, for integer
orders:

L2^(X) = M % ) ; M 2 , + 1 (x) = P 2 /+i(x) (A.18.11)

The second integer-order solution to Eq. A. 18.1 is commonly defined as:

Qv(%)= * JP v (X)cos(v7t)-P v (-x)] (A.18.12)

2sm(V7t)
QV(X) is obviously a solution of the Legendre differential equation and,
when v is equal to integer I; it is in indeterminate form. Differentiating
numerator and denominator then letting v become an integer:

Lim l
Q/(Z) = -7tsin(vn)Pv(x) + cos(v7t)— - —
v=>£ 2cos(v7t)L dv dv
(A.18.13)
l[dP v (x) dP v (-x)|
2 I dv dv j v = (
Appendix 383

Using Eq. A. 18.13, the functions at the lowest three orders are:

Q o (x) = ln[cot(0/2)]

Q 1 (x) = ( c o s e ) l n f c o t | j - l

3
Q 2 (x) = ^(3cos 2 e-l)ln| c o t9- •— COS0 (A. 18.14)
2
V 2y
Comparison with the noninteger functions, Eq. A. 18.12, shows that:

Q«(x)=^U«
(A.18.15)

Q2t+i{X)—z— v=*2/+l
3v

In this work, we are concerned only with the zero order function Qo(x)-

A.19 Associated Legendre Functions

Associated Legendre functions are solutions for the extended case m > 0.
The Legendre differential equation, Eq. 1.11.10, may be rewritten as:

/, 2\d 6 „ d© m
v(v+l)- G=0 (A.19.1)
V ;
dx 2
dx
R)
Solutions are most easily obtained by starting with the m = 0 equation and
differentiating m times to obtain:

OT+2 w+1
/ 2 \,dd 0
fc> , d fc>
v d 0 r / ,\ i \id 6
2 { m + l ) X + [v{V + l) m{m +l ) ]
' ) ~ ^ ~ ^ - ^= ° (A 19 2)
- -

Introducing construction function W(0) and solving for the first two
derivatives:
384 The Electromagnetic Origin of Quantum Theory and Light

e(e) = w(e)Sinme
— = — sin w e-mWsin m - 2 9cos9
d% d%

—"- = — - s i n m 9 - 2 m sin m 2
9 c o s 9 + mini -2)\Vsin m 4
9cos29-mWsinm 2
0

(A.19.3)

Substituting Eqs. A.19.3 into Eq. A. 19.1 for the special case m = 0 results in:

d2W 7 , x dW r / \ / XT
—^-sin z Q-2(m+l) cos6+\v(v+l)-m(m+\) W = 0 (A. 19.4)
dx 2 dX
For the special case where v = £, an integer, comparison of Eqs. A.19.2
and A. 19.4 shows that W is given by:

n,/M d"Wcos0)
W / (e)= ^ '- (A. 19.5)
A ;
dx w
Since Eq. A. 19.5 satisfies the associated Legendre differential equation, the
solution of that equation is:

~/ \ mi \ m d^P/fcosG)
e(e)=pf (cose)=sin^e—^—'- (A.19.6)
dXW
The equality holds for all integer orders, £, and degrees, m. Combining
Eq. A.19.6 with the Rodriques formula shows that the corresponding
expression for associated Legendre functions is:

ml X 1 / 7\m/2 <l£+m I 7 V
P = 1 £ <A19 7)
'(") (S)iT( -' ) ^ M '
Appendix 385

A.20 Orthogonality

An integral relationship for products of Legendre polynomials follows

directly from the differential equation, Eq. A. 19.1, for integer orders.
Multiplying the differential equation by another Legendre polynomial of the
same degree but unspecified order gives:
d 2 pf
J C(x) _d_
M dX2
/(* + ! ) -
(-4
dx = 0 (A.20.1)

To evaluate the left side, integrate the first term once by parts. The result is:

2^dPfdC m dx=0
+ £[t+i)- (A.20.2)
-1
.-*»)
dX dx
M
Next exchange positions of £ and n, repeat the process, and subtract the
second integral from the first. The result is:

[*(*+l)-n(n+l)] J Pf(x)tf(x)dx = 0 (A.20.3)

The result shows that the associated Legendre polynomials form an

orthogonal set.
To evaluate the integral with £ = n, put Ii equal to the integral:

J>f(x)fdx (A.20.4)

Combining Eqs. A. 19.7 and A.20.4:

(-l)m I de+m i 2 \t\, 7 \m ^+m , , x

(A.20.5)
386 The Electromagnetic Origin of Quantum Theory and Light

Integrating by parts 1 + m times leaves:

At+m
l
1 At+m
I l = 7 T ^z H ( l - X 2
/ d (x2-vm (x2-i)* (A.20.6)
[(2f)!!] j dx t+m dx t+m

Only the highest power of X survives the indicated differentiation operations:

\t+m 2 t+m it+m At+m

dXt+m (* -feM dX dX t+m

y2m_^
x
dX Hm
21

(A.20.7)
_ {2t)\ de+m ,J+m {2i}.(l + m)\
;X
(e-/n)ldx£+m (t-m}.

Combining Eqs. A.20.6 and A.20.7 leaves:

(2£)l(£ + m)\
I, = Z jdx(i-x2)' (A.20.8)
[(2£)V] (1-^)1^

Using the binomial expansion, Eq. A. 18.9, and integrating:

_ 2(-l)'(2l-!)!!(* + /»)! J , (-l) s l!

1_ (A.20.9)
{2l)\\{Z-m)\ s ^ 0 s!(^-s)!(2^-2s + l)

The sum may be written in closed form as:

2 {t + m}.
Ii = (A.20.10)
(2£+l) (£-/»).

Combining Eqs. A.20.3 and A.20.10 gives the orthogonality relationship for
associated Legendre polynomials:

(t + m)\,
(A.20.11)
'.-JtfWFWxfc-^^*.'')
Appendix 387

A similar integral, the value of which follows by a slight extension of the

above, is:

1
(£ + m)\
h = f Pf(x)Pr2S(x)dX=(-l)S, ,,K \ (A.20.12)

A.21 Recursion Relationships

It is helpful to compile a table of identities involving associated Legendre
polynomials. A convenient starting point for determining the recursion
relationships is Eq. A. 18.7. For the case that t is odd the upper limit is
{1-\)I2. Substituting p = (l-\-2s)ll into the equation yields:

n , \ X(-lf-l)/2%)/2 (-l) p (£ + l + 2p)! ,_2s

p
/(x) = Z : L J 7 X ; / , „ _ w^ , .+, l +,,^X
2p^ (A.21.1)
2' % ( 2 p + 1 )(^E
V ^ J

Rewrite the equation as:

Pt(x). (-,)"-'" j a ! ! _ " f <-•>'*'"'«^ w i w - •)••

u/v v / (*-l)!! ~ (2s + l)!W!!(^-l-2s)!!

For the case of f even, the upper limit is ill. In a similar way the equation
goes to:

P<U)-( i) w „ L (2s)! (^ 1 ) !!( ^_ 2s ) !! (A-21-3)

With Eq. A.21.3, replace I by (£+1) and write out the first few terms, then
repeat with t replaced by ii-Y). The resulting series are:
388 The Electromagnetic Origin of Quantum Theory and Light

P (r)-( 1^ + 1 ) / 2 M ! fl X2(^ + 2)!!(^-H)!! y 4 ( / + 4)!!(^ + l)!! 1

*MW K V (e + l)\\[ 2! ($] [£-l)V 4! (£)l\ (l-3)\\ "j

P te)-( tf'-WJ'-tyli I? M" (^-1)",X 4 (^2)!!(£-1)!! 1

t \\h) \ ) (i-\)\\\ 2! {t-ip(i-$p 4\ (e-2).\(t-5)n ""j

These expressions combine to form the indicated sum:

(, - 1 ) / 2 ( 2 ^ l ) ( Q ! ^ 2 ( - l ) V s + 2 ( l + 2s)!!(l-l)!!
1
' (*-l)!!) s t J (2s + l)(*)!!(*-l-2s)!!
(A.21.4)

Combining Eqs. A.21.3 and A.21.5 results in:

( 2 I + 1)XP^(X) = ^ _ I ( X ) + (^ + 1)PM(X) (A.21.5)

Proofs for even values of t follow in a parallel way and give the same result.
Equation A.21.5 is the first recursion relationship.
The same technique with the indicated operations results in the second
recursion relationship:

(«+DPI(X).M.M (A.21.6)

The integral expression of Eq. A.21.7 followsfromEq. A.21.6:

"P/+lQc)-P/-l(%)'
JP^(x)d% = (2£ + \)
(A.21.7)

Differentiating Eq. A.21.5 by X and adding I x Eq. 21.6 gives:

(A.21.8)
Appendix 389

Differentiating Eq. A.21.8 m times, multiplying through by sin m+1 9,and

using Eq. A. 19.7 gives:

P m 1 ( x ) = X P r 1 ( x ) + (^ + ' » + l ) s i n 9 p r ( x ) (A-21.9)

A series of identities follows by mixing and matching. Selected ones are

listed in Table A.21.1.

1. ^-=^+*)(/-«+i)prl-?r1]
2
- wpf i r B « - 1i i n * i -
= -[(t + m){l - m+ DPfl + P£V ]
sinG 2
3 HP W 1
d6 2^ + 1 ' '+1 "
4
cos 0Pf - ^ [ ( * - * + l ) P ^ ! + (^ + z*)P£! ]

5- Pf*1 = 2«cot6Pf - ( * + « X * - >»+ ^ r 1

6
- (*-/»+ 1)P^.! = (2* + l)cos8Pf - (€ + /w)P^!
7
- (2f + l)sin9P/' = P / t t 1 - P £ , j 1
8
- P^Ti1 = cosGP/""1 + (I + m+1) sinGPf

9. p^+j1 = cos ePf^1 -il-rn) sinGPf

^=/rccotepf-pf + 1
de ' '
I I. dpf i
—^- = -mcote?^+(£ + m)(£-m+ l)Pf~'
Table A.21.1 A Table of Identities for Legendre functions

Associated Legendre functions have even or odd parity, respectively, if

the sum {I + m) is even or odd:
390 The Electromagnetic Origin of Quantum Theory and Light

\£+m
Pf(xH-lF>f(-x) (A-21.10)

Let delta be a Kronecker delta and let q equal any of the field of positive
integers, including zero. At 0 = TC/2, the equator, functional values are:

^n0),{-lf-^^±^,(2qj+m) (A.21,1)

— Pf(cos0) „ =??»r{0) (A.21.12)

eK
de' ' 2
In the tables to follow, order is £, degree is m, and % = cos 0. Polynomials are
even or odd functions of % as (I + m) is even or odd, respectively; absence of
a superscript indicates degree 0. Normalization is chosen to make P^(l) = 1.
Table A.21.2 contains values of associated legendre polynomials. Useful
recursion relationships for constructing the table are:

pr(x) =
7^[ (2 '- 1)xP ^ (x) - ( '" 1 + w ) p ^^]

VT{l)-[2{m-\)ootQVr\%)-{l-\+mll+2-m)?r2{%)\

dP,
PJ(X) = - d0

The sum of all coefficients in the expansion for P/*(%) is

{e + m)\/2m(m)\(t -m)\. To normalize, multiply each associated Legendre
function by •yj(2e + l)(e-m)l/2(£ + m)\.
Table A.21.3 contains values of spherical angular function dP/Yd0,
based upon the relationship:

d p f / d 0 = - mcot 0 P f +(£ + m){l - m+ l)Pf~ l = wcot 0 P f - P,m\\

I m = 0, m = l, TO
aa.
Values for P^(l) Values for p|(%)
_ .

1 X sin9
2 l/2 2 \ 3xsin6
;(3X -1)
2'
3

i(35 Z "-30 Z 2 + 3) f ( 7 " 2 - 3 ) SsinG

2 V " /
|(63x 4 -70x 2 +15) ^(21x 4 -14x 2 +l)sin6

-^(231x 6 -315x 4 + 105x 2 -5) ^ ( 3 3

^ ~30^ + 5 sin9
)
7 — (429x6 - 693x4 + 315x2 -35) — (429x6 - 495x4 + 135x2 - 5) sin 6

— (6435x8 - 12012X6 + 6930x4 - 1260X2 + 35) —(715x 6 -lOOlx4 + 385x2 -35)sin6

—(l2155x8 -25740x6 + 18018X4 -4620x 2 +315) — (2431x8 -4004x 6 +2002x4 -308x 2 + 7)sin6

Table A.21.2a Table of Associated Legendre Polynomials, P f (x)

392 The Electromagnetic Origin of Quantum Theory and Light

CD
a

o _c ON
5*C
'co"" CO

CD
'co^ + "3-
+
+
CD a uo
't/a ON
'co ^ 1 1
1 1 • *

I ?<
CD 1 1—I
X X Ov
CO CM
> C*"l ts CO
-a
:5
• < *
CD
co" ro ?<
_g o
in 1 >n CN I/O 1
^ 00
in 0 0 NO V

< — ' 1 <N

O CO co | CO CO

|<N"C
PH
t-i
CD

«S
1
CD
CN M
SJ
CD cS
> <N .gw r~
3sin

IT) ""> 1 i

hv
Appendix 393

£ m = 4, m = 5,
Values for P*(x) Values for pf (x)
4 4
105sin 9
5
945%sin4e 945sin56
6
945(1 l x 2 - l ) s i n 4 0 / 2 10395xsin58
7 3465x(13x 2 -3)sin 4 6/2 10395(13x 2 - l)sin 5 6/2
8
10395(65%4 - 26x 2 + l)sin40/8 135135x(5x 2 -l)sin 5 9/2
9
135135x(17%4 - 10x2 + l)sin 4 9/8 135135(85x 4 -30x 2 + l)sin 5 6/8

Table A.21.2c Table of Associated Legendre Polynomials, P/^X/

I m = 6, m = l,
6
Values for P, (%) Values for p / ( x )
6 6
10395sin 9
7
135135xsin60 135135sin79

8 135135(15x 2 -l)sin 6 0/2 2O27O25xsin70

9 675675x(17x 2 -3)sin 6 0/2 2027025(17x 2 9- l)sin70/2

Table A.21.2d Table of Associated Legendre Polynomials, Pf*(x)

m = 9,
Values for ?$(%) Values for pf(x)
8
2O27O25sin80
9
34459425xsin80 34459425sin90

Table A.21.2e Table of Associated Legendre Polynomials, P ^ ( x )

 m = 0, Values of dP^/d0 = - P ^ are listed above in Table A.21.2.a
£ m = l, m = 2,
dPJ /dG = - cot 9PJ + ^ + l)Pj dpf /d9 = - 2 cot QdPf + (£- !)(£ + 2)dPJ

2
3(2x 2 - 1 ) 6
XsinQ

3 3
z(l5Z2-ll) 15( 3 % 2 -l)si
sinG
hi
4 5
(28%4-27%2+3) 3 0 x ( 7 Z 2 - 4 ) sin 8
2
5
^(l05%4-126%2+29) l^(15x4_12%2+1\sine I

(l98% 6 -285% 4 + 100%2 - 5 ) -(99jc 4 -102x 2 + 19)sin0

a"
66 44 22 6 4 2
— (3003% -5049x +2385x
(3003% -5049x +2385X -275J
-275) — (lOOl* -1265% + 375% -15Jsin0
-15Jsin0 §
8 |
— (5720% 8 -11011x 6 +6545x 4 -1225% 2 +35) -(286% 6 -429% 4 + 176%2 - n ) s i n 6 |
16' ' 8 * ' K
8 6 4 2
—(21879% -47464% + 34034X -8932% +623) (l989v. -3458% + 1820% -294% 2 +7)sin9
8 6 4
|
16 ' ' 16 » ' o
6.
Table A.21.3a Table of Spherical Angular Function dP^ */d9
 m = 4,
dP//d6 = -3cot8P/ + (£- 2)(£ + 3)P/ dP//d9 = - 4 cot 6P 4 + (* - 3 ) ( l + 4)P*
3
45xsin20
4
105(4x2 - l)sin 2 0 420xsin39
5 315x(15x 2 -7)sin 2 e/2 945(5x2 - l)sin 3 9
6 945(22x4 - 15%2 + 1 )sin20/2 945%(33x 2 -13)sin 3 6
7
315x(1001?c 4 -902x 2 + 141)sin29/8 3465(91x4 - 54x 2 + 3)sin30/2

8 10395(104%6 - 117%4 + 30x 2 - l)sin 2 6/8 10395x(130x4 - 104x + 14) sin30/2

9
10395x(663x6 - 897x 4 + 325x 2 - 27) sin20/16 135135(153x6 - 155x4 + 35x 2 - l)sin 3 0/8

Table A.21.3b Table of Spherical Angular Function dP/*/d0

 m=5 m =6
dPf /d6 = -Scot 6P| + (I - 4)(l + 5)P/ dP|/d6 = -6cot9Pf + (£ - 5)(£ + 6)P|
4725xsin40
4725(6x2 - l)sin49 62370xsin5e
10395x(91x2-31)sin49/2 135135(7x2-l)sin56
8
135135(40x4 - 21x2 + l)sin46/2 810810x(10x2-3)sin59
9
675675x(153x4 - 110x2 + 13)sin46/8 2027025(5 lx 4 - 24%2 + l)sin50/2 3

Table A.21.3c Table of Spherical Angular Function dP/yd6

t m = 7, m = 8,
i
dP/ /d9 = - 7 cot GP/ + (t - 6 ) ( l + 7)P| dP|/d9 = - 8 cot 8P* + ( l - 7 ) ( l + 8)P/
7 S'
945945xsin60
to
ca
8 2027025(8x2 - l)sin60 16216200xsin79
3
9 2027025x(153x2 - 41)sin60 34459425(9x2 - l)sin70

Table A.21.3d Table of Spherical Angular Function dPf*/d8

I
3
Appendix 397

I m =9,
dPj/de = -9cot9P| + (£- 8)(l + 9)P/
9 310134825%sin86

Table A.21.3e Table of Spherical Angular Function dPf/d6

A.22 Integrals of Legendre Functions

Integrals of different functional combinations of Legendre polynomials are
needed. Consider, for example, the integral:

i3 = 4^(pfp;r)de=o (A.22.1)
0

The equality follows since the integrand is a perfect differential and

Pf(±l) = 0 f o r » i > 0 .
Consider the integral:

(A.22.2)
[\ de de sin2e j

The first term in the integrand may be written:

dPfdC_ 1 d mKxrn pf d dP w
sinep; sin6^2-
de de sine de de sine d6 de
The second term in the integrand, after using the differential equation, may
be written:

ml?tK Pf d s i n 6 ^ - + n(n+ l)P^Pf

sin 2 e sin6de de
398 The Electromagnetic Origin of Quantum Theory and Light

Combining results in:

1 d dP w
I 4 = Jsined0 s i n 6 P f ^ - + *ZM-l)PfP* (A.22.3)
sin0 d0 ' de
0

The first term is an exact differential that integrates to zero, leaving:

U = n{n+ l)fsin9dePfP*= 2 * ( * + 1)( * + / * ) ! 8(l,*) (A.22.4)

0 (2t +
(Zl + l)(ll)(t-m)\
- m)!

Consider the integral

I 5 = Jcosesin9de (A.22.5)
de de sin 2 e I

The procedure is similar to that for I5. Replace the first term using the
differential equation, sum, and partially integrate once to obtain:

I s = Jsin8d8< + n{n+\) cos 8PfP"

' de
0

Combining with recursion relationships, Table A.21.1.3 and A.21.1.4, shows

that:

JI
n(n- m+ \)(n+ 2) nm
Vm -t-
(n+m)(/T-l)
I 5 = Jsined0P; n-\
(2n+1) (2/?+1)
0

Evaluation using Eq. A.20.11 gives:

2n(n+2){n+m+\)\ (rt-\){n+ m)\

U= 8(i,/i+l) + •8(^/7-1) (A.22.6)
(2n+\)(2n+1)(n- m)\ (2n-\)(2n+X)(n- m-\)

Consider the integral

Appendix 399

I6 = Jsin 2 6de^(pfP-) (A.22.7)

0

Expanding the differential and using recursion relationship Table A.21.1.3

results in:

n(n- m+l) {n+ m){n-\)

(2/7+1) (2n+1) tVl
I6 = Jsin0d0
0
(2f + l) (2^ + 1)

Term by term evaluation shows that:

(A.22.8)

Consider the integral

dPf d P ^ _ Mm+Jl „ ^
I 7 = J"sin20d9 2 (A.22.9)
de de sin e
Substituting the differential equation into the first term, summing, and taking
one partial integration results in:

JsinedGP^^i-cose^-h ^ + l)sin6+ m
de sin©

After using the recursion relationship of Table A.21.1.10 on the first term
then summing the curly bracket becomes:

{[*(* +1) + m] sin0Pf + cos GP?*l}

With the use of the recursion relationships of Tables A.21.1.4 and A.21.1.7
the bracket becomes:
400 77ie Electromagnetic Origin of Quantum Theory and Light

1 + 2) (t-W+l)'
(2^ + 1) (2^ + 1)

Putting the bracket back under the integral sign and integrating gives:

2£(£ + 2)(£ + m+2)\ g, 2(1-!)(*+!)(*+>»)! g, . „

I7 = 0(n,£+\) o(/t,£-l)
7
(2^+l)(2£ + 3)(^-w)! (2£-l)(2£ + l)(£-m-2)\(A.22.10)

Consider the integral

/
,*/dP^m+l
I 8 = Jsin9de ( / w + l ) P / ^ 1 - ^ - + wP^ (A.22.11)
0 V de
This may be rewritten as:

71 (
f Hfll s i n e P1f- f l ^ + W s i n 0 d
d e |sinflPf* ptfZpWfl

0 V de de
Integrating the perfect differential by parts gives:

( "\ 71
rtrt-\ dPi fl m+\
JsinedGP, ^--wcotep' = -Jsin9dePf P^
0 de J 0

The second equality is in the proper form to use Table A.21.1.10. The result
is:

2 (* + />/+l)! g / ,
8 (A.22.12)
(2*+1) ( * - » - ! ) !

Similarly, except using Table A.21.1.11, integrals with (w-1) replacing

(/w+1) may be evaluated, and are listed in Table A.22.1.
Appendix 401

1. 2{(. + m)\
l„ = fprP*sined8= ^m>- 8(/>g)
(2e + W-m)\
2. 2(-l) s (£ + ;»)!
J".fPf +2s
sin6de =
(2^ + 1) (*-z»-2s)!
3. 7C

-J^(p^)de=o
4. f d / . -\ 2m(( + m)\ .
] — PfP* cos9de= (2£
_ + „l)(£-w)!
,, \.8(l,*)
5. 71
)^(pfPr)sin20d6 = 0
de
6. dPf dPOT m P
n , l P* . 2t(i+w+my.*,, ,
de de sin 6 sin6d6 = —
(2( + l)(£-m)\ o{l,ri)
2e(e+2)(e + m+\y.
2
dPf dP^ m PfP™ (2£ + l)(2i + 3)(,e-m)\
cos9sin9d8 =
de ae + s i n 2 e 2(e-l)(i+ !)(£ +m)l
8(/tJ-l)
Q.t-\yii+\Yji-m-\y.
2((e+2)(e+m+2y.
8(»,* + l)
dPf dPr' ,njm+\)YTK sin2 6d6 = Qt + i)(2l + 3)(t-my.
I de de sin26 2(i-ixe + \)(i+my.
5(^-1)
(2f-l)(2f + l)(^-w-2)!
9. 2 ^ + 2)(^ + «)!
X 6(^+1)
dpm dpm-l m{m„Xy?™V™~ (2£ + \)(2e + 3)(i-my.
J sinz0d9 =
de de sin26 2((-\)(t + W+m)\
8(^-1)
' Q.l-\)Q.l + \)(t-my.~
10.
P/ (cos6)sin 6d6=(-l)v ;/
/ v - 8 U + /»,2q)
J V V
' (2*+l)!
11. ,-H-I** . _ D «dP,
OT+l
l+m+\)\
sin6d6 = - -
^ de " de (2^+i) (e-m-iy.
12. OT-1^
2 (t+m)\
\ {m-x)?r
&K
x
^+mr„,dP; sin6d6 8(*,/r)
de de '(2e+i)(e-m)\

Table A.22.1 Table of Integrals of Legendre Polynomials

402 The Electromagnetic Origin of Quantum Theory and Light

A.23 Integrals of Fractional Order Legendre Functions

The Legendre differential equation of fractional order is:

1 d . Q d0ff(cos9) m
+ v ( v + l ) - sin 2 0 0?=o (A.23.1)
sin9 d9 sin 9 — -
V dG

Functions 0^f(cos9) represent L^(cos9), M^(cos9), or any linear

combination thereof. Useful boundary conditions are:

,.mi „\l , dL™(cos9)i

v (A.23.2)
M^(cos9)|e=v=0 and ^ '|9=¥=0

This evaluation of integrals over noninteger order Legendre functions

includes the boundary conditions of Eqs. A.23.2.
Consider the integral

7I-\|/
dM? dP™
I9= J sin9d6 (A.23.3)
d9 d9 sin2 9
V

The evaluation procedure is to use the differential equation and rewrite the
first term as:

dM£dP*_ 1 d M* d s i n 9 ^ _
sin6M
v ?^- (A.23.4)
d0 d9 sin9 d9 d6 sine d9 d9
v
The first term on the right side of Eq. A.23.4 forms a perfect differential and,
after imposing Eq. A.23.2, the integral of that differential is equal to zero.
With the differential equation substituted into the remaining term, the result
is:

7t-V|/
I 9 = n(n + 1) J M^P„ w sin9d9 (A.23.5)
Appendix 403

To evaluate Eq. A.23.5, since M^(cos9) has odd parity and P^(cosG) is
even or odd as (n + m) is even or odd, the integral vanishes if (n + m) is
even. If (n + m) is odd, repeat the procedure used in Eq. A.20.1
through A.20.3. The result is:

It—VJ/
dM dP
sin0
n-y " d6 d8
I 9 = j M^PfsinGde:
/?(/z+l)-v(v + l)

After imposing the boundary condition of Eq. A.23.2

dM^(cosy)
n-y P*(cos V ) :
J" M^P™sin9d6 = -2sin\|/ d9 6(//,2q + l) (A.23.6)
«(/?+!)-v(v + l)

The integer 'q' represents any positive integer, including zero.

The next integral to be evaluated is:

7I-\|/ ^ —m ,~m -J-j/n-am^

z
I10= J sinGdG
dL«dP« ™ ylK
+• (A.23.7)
de de sin2e
The same technique that was applied to Eq. A.23.2 when applied to
Eq. A.23.7 results in:

n-v|/
I 10 = ji0i + 1) j" L^P*sin6de (A.23.8)

The same technique applied to Eq. A.23.4 results in:

404 The Electromagnetic Origin of Quantum Theory and L

dP^(cos\|/)
7I-\|/ L^ (cos nO-
I 10 = J" L^P"sin9de = 2sin\|/ de •5(«,2q) (A.23.9)
n(n+l)-li(H + l)

It follows at once from the parity of the functions that the integral:

Tl-\\l
Ii 1 = sinGdG —+ — -= 0 (A.23.10)
11 2
1 de de sin e
Consider the integral

2
T "7 • nJdM™dMZ m M"M^
I]2 = sinGdG - —+ r—- (A.23.11)
2 2
J \ de de sin e
The same technique applied to Eq. A.23.2 results in:

7t —\|/

I 12 = v(v+1) J" sinGdGM>l™ (A.23.12)

V

The same technique applied to Eq. A.23.4 results in:

71-V

m UdM
lvl
v „ / »; "dM*
"*H
sin 9 M M
n-y dG de
j M^Mj7sin0de =
u.(u. + l)-v(v + l)

The boundary condition shows that the result is zero unless

H(H +1) - v(v+1). For that case, evaluating the indeterminate form gives:
Appendix 405

7t-l)/
2sin\|/ 3My (cosy) 9My (cosy)
I12
12 = Jf Mv> CHs i n e d e = ^ ^ 5(v,n) (A.23.13)
2v+l 3v ae
The delta function indicates a Kronecker delta function with a noninteger
argument.
Consider the integral

r-W Jjr«Jr«
In= sinGdG — - — — + =—— (A.23.14)
13 2
| ^ de de sin e
In a way similar to the earlier integrals:

Tc-y

Il3 = H(H + 1) J sin BdeL^L* (A.23.15)

2siny 92L"(cos\|/)
1.3 = J" L-L-sin9d6 L^(cos\|/) 5(v,u) (A.23.16)
2p, + l dvdQ

Results are summarized in Table A.23.1.

Spherical Bessel Functions

A.24 The First Solution Form

Since the radial differential equation, Eq. 1.11.7, is independent of
separation parameter m, so are the solutions. The spherical Bessel
differential equation with separation parameter v is:

v(v+1)
2 + 1- R=0 (A.24.1)
a day do a2 j
406 The Electromagnetic Origin of Quantum Theory and Light

P-(cos ¥ )^e) | e =
lv„= j M^P^sin0de = -2sinv|/ 8(m+ «,2q + l)
«(«+!)-v(v+l)
-V

iy "*/7 V*/?
+ • sin 6d9 = /*(/?+ l)l,
L
de de sin20 v»
-v v
dP^(cosy)
L£(cos\|/)-
K ^ = J L^P^sined0 = 2sinV|/ d6 5(//,2q)
«(n + l)-H(H + l)

sin0d0= JLi(|J. +1)1^.^

J HA rlfl „ : J2 f l
de de S in e
f ^
\ + 2 sin6de = 0
de de Sin e
7t-\)/
r «/« 2sinvi/ 3M*(cosy) dM^(cosy)
5(v,ji)
Ivv= f M*M*siii8d8 = - 3v ae
vv J v
^ 2v + l
""¥^dM-dMj+«2M^
J de de sin2e
sin6de=v(v+l)I vv

71-ty
2sinV|/ ^ 2 L^(cosx|/)
Kw= J L^sined9 = - Lv; (cos MI) — £ — 8(v,n)
2^ + 1
V

I de de +
sin 2 e
sined0=|a,()LH-l)K^

Table A.23.1 Table of Integrals, Noninteger Order Legendre Functions

Appendix 407

Equation A.24.1 is a second order equation and has two independent

solutions. Because of a singularity at the origin, solutions are satisfied over
2
the region 0 < a < oo. For the range of solutions in which a » v(v+l), the
differential equation goes to:

r dR(g)
2 a' R(a) = 0 (A.24.2)
a da da

Solutions of Eq. A.24.2 are:

R(a)
v = — L cosa+^=-sina (A.24.3)
' a a

C\ and C2 are constants of integration. This is the asymptotic limit for large
radius.
To obtain solutions valid at all radii, use a power series expansion to
solve differential Eq. A.24.1. The series and its first two derivatives are:

00

Rv(°)=2>sas+P
s=0

dR v (a)
£ ( s + p ) a s a s + e -1 (A.24.4)
da s=0

d2Rv(a)
= S(s+PXs+p-l)asas+P-2
dcr s=o

Substituting Eqs. A.24.2 into Eq. A.24.1 and solving leads to

[p(p + l ) - v ( v + l)]a 0 a p - 2 +[(p + l)(p + 2)-v(v + l)] a i a p

00
= 0 (A.24.5)
+X{[(s + P + 2)(s + p + 3)-v(v + l ) ] a s + 2 + a s K + P
I s=0
408 The Electromagnetic Origin of Quantum Theory and Light

Since the series of Eq. A.24.5 is an identity in a, the coefficient of each

power of a is separately equal to zero. There are but two nontrivial ways that
the coefficients of and a can both be equal to zero. One is if ao is
equal to zero and (p+l)(p+2) = v( v +1). The other is if ai is equal to zero
and p(p+l) = v(v+l). Arbitrarily making the second choice, the condition
that p(p+l) = v(v+l) is met either of two ways: p = v or p = -(v+1); these
choices determine the two independent solutions.
For the case p = v the portion of Eq. A.24.5 in the curly brackets is zero,
and results in the recursion relationship:

^±2- = -7 ^ r (A.24.6)
as (s + 2)(2v+s + 3)

This relationship, after redefining the dummy index, leads to the functional
form of the radial function R v (o):

^ , (-l) s (2v+l)!! _ v + 2 s
R v (a) = a 0 > — a (A.24.7)
s
sT02 s!(2v + 2s+l)!!

Making the definition that a 0 = l/(2v+ l)l! the result is the function:

(
JV(°)=£ . "1)S—°v+2s (A-24-8)
s
sT02 s!(2v+2s + l)!!

Functions j v (a) are the spherical Bessel functions of order v. The functional
limit at small values of a follows from Eq. A.24.8, and is equal to:

Lim r , v, 0V
IJv(o)l = 7 \" (A.24.9)
LJvV n
CT=>0 (2v+l)!!
For the case p = -(v + 1), the last term of Eq. A.24.5 results in the
recursion relationship:
Appendix 409

a
s+2 1
(A.24.10)
as (s + 2 ) ( 2 v - s - l )

This relationship leads directly to the series solution:

-2 „4 ,2P
1+
2(2v-l) 2.4(2v-l)(2v-3) + '"
...+ (2p)!!(2v-l)...(2v-2p-l)
- -+...
yv(-)=^r 2(v-l) ,2v
(2v-2)!!(2v-l)!! l«(2v)!!(2v-l)l! 3»(2v+2)!!(2v-l)l!
(A.24.11)

The series is monotone for s less than v and oscillatory for s greater than v.
The combination is readily described by separate sums over the monotone
and oscillatory portions. After using the definition &Q = l/(2v+ l)l! and again
redefining the dummy index:

H) s v-l+2s
,M--%££SM-i
to(2s)!!a -s
v+I 2s
t (2s-l)!!(2v 2s)!! s 0 +
(A.24.12)

The symbol [v] indicates the largest integer less than v. The first sum of
Eq. A.24.12 describes a monotone power series with inverse powers of v,
powers that range upward from -(v + 1) to (v + 1) and a second sum that
represents an alternating series with positive powers of v. The sums are the
spherical Neumann functions; Eq. A.4.12 shows the functional small
argument limit of j v (a) to be:

Lim
[y (g)] = J2v-1)"- (A.24.13)

For integer orders, v equal integer £, and o » 1, the functions are

determined by Eqs. A.24.8 and the second of Eqs. A.24.12:
410 The Electromagnetic Origin of Quantum Theory and Light

Ms
s r 0 (2s)!!(2£ + 2s + l)!!
(A.24.14)
1 —
/ \___V ( 1) ^-l+2s
M = CT
°' s f 0 (2s-l)!!(2^ + 2s)!!

Term-by-term comparison of the series representation of Eqs. A.24.3

and A.24.14 shows that:

Lim , . 1 ft,
n(o) = — cos a - ^ + l)
(A.24.15)
Lim , . l . 71,
y/(a) = - s i n o-£(/ + l)
Using Eq. A.24.15, it follows that the Bessel and Neumann functions are
related as:

(A.24.16)

A.25 The Second Solution Form

If the separation constant is an integer, the spherical Bessel differential
equation, Eq. 1.11.7, is given by:

1 d ( 9 dlO M_M±1) 0
+ |R = (A.25.1)
a 2 da da

Solutions follow that are more convenient to use than Eqs. A.24.8
andA.24.12. Integer solutions, valid over the range 0 < a < oo, may be
obtained by removing the value at infinity before obtaining a detailed
Appendix 411

solution. To remove the value at infinity, note that as the radius becomes
infinite Eq. A.25.1 approaches:

— - ^ + R,(a) = 0 (A.25.2)
da

If Eq. A.25.2 were an exact solution, the result would be exponential with
constant coefficients. Although Eq. A.25.2 is not exact, it is helpful to write
Eq. A.24.3 in the form:

Re(o) = Fe(a)e~'° +Gt(o)eJG (A.25.3)

A requirement is that at large radii F^(a) and G^(a) vary much less rapidly
with increasing radius than do the exponentials. Also since F^(a) and Ge(a)
are complex conjugates it is only necessary to solve for one of them.
A convenient method of obtaining them is with a power series
expansion. The series and the first two derivatives are:

dR£(a) dF^(a) -id

R / ( o ) = F/(o)e" -in. •tf>(a)
da da
(A.25.4)
d z R^(a) _ -/a
da 2

Substituting Eqs. A.25.4 into Eq. A.25.1 results in the differential equation:

d ^ o ) + 2 r 1 WCa) 2/ 1(1+1)'
F^(a) = 0 (A.25.5)
-+
da^ Va J da
The most convenient method of solving Eq. A.25.5 is with a power series
expansion. The series and the first two derivatives are:

M°)=£a s a s+ P
s=0
412 The Electromagnetic Origin of Quantum Theory and Light

dry(q)
= S ( s + P )a s a s + P- 1
da
s=0
(A.25.6)
2
d Fi(q) s+ 2
= S ( s + p)(s + p - l ) a s a P -
do" s=0

Inserting Eqs. A.25.6 into Eq. A.25.5 and gathering similar powers of a
result in:

P(p+l)-l(l+l) y s+p+,

CT
S=0 (A.25.7)

x{as+1[(s + p + l)(s + p + 2 ) - ^ + l)]-2/a s (s + p + l)}

Since the series is an identity, the coefficient of each power of a is

separately equal to zero. It follows from the cr term that either p = I or
p = -(£ +1) and it follows from the square brackets that:

a
s+l 2/(s + p + l)
(A.25.8)
as (s + p + l)(s + p + 2 ) - ^ + l )

With the option p = t, in the limit as a becomes infinite Eq. A.25.8 goes to:

Lim f a s V 2/
(A.25.9)
s=*°°{ a s j s

Equation A.25.9 is also the limiting form for a series expansion of exp(2/a).
-ia
Therefore, since F^(a) varies more slowly with a than e , the recursion
relationship of Eq. A.25.8 is not an acceptable solution. Returning to the
option that p = -(v + 1), Eq. A.25.7 goes to:

l 2J(1-S)
*s+ (A.25.10)
as (s + l)(2*-s)
Appendix 413

Since the progression of Eq. A.25.10 terminates at s = I the power series

truncates to a polynomial of highest order £. The general term is:

a s+1 (2/) s l!(2l-s)!

(A.25.11)
as s\(2t}(t-s)H

The series results in solution R / a ) where:

e-^^ (2/) s l!(2l-s)!

R/H= a s!(2/)l(/-s)! a °
(A.25.12)
sf0

To characterize the solution substitute p = £-s, rewrite Eq. A.25.12 as a

sum over p, then change the dummy index back to s. The result is:

-ta
CT = a
l!(2/f 4, (l + s)l ( 1 ^
M ) o' (A.25.13)
a (2/> s ? 0 s!(£-s)! V2zay

With the definition that ao = i(2£-l)\\, the full solution is the function he(a)
where:

M C T ) = —a - — <£-
S s!(*-s)
- (A.25.14)
V2/(Jy

Function h^(a) is a spherical Hankel function of the second kind. The real
part is a spherical Bessel function and the negative of the imaginary part is a
spherical Neumann function. By definition:

h
t(°) = 3t{o)-fye(a) (A.25.15)

The complex conjugate of Eq. A.25.14 is the second independent solution of

the equation. It is a spherical Hankel function of the first kind. In this work,
we shall be concerned only with Hankel functions of the second kind.
For vanishingly small values of a, the dominant term in Eq. A.25.14 is:
414 The Electromagnetic Origin of Quantum Theory and Light

Lim r , i(2l-\)\\

a^o^ ( a ) ] = " ^ r ^ (A 25 16)

- -
As the radius increases without limit, the dominant term is:

Lim r , ,n / + 1 _,„
\hf(a)\ = e ^ (A.25.17)
Equation A.25.17 shows that as the radius increases without limit the
function [ah^(a)] does not approach a limit. For those cases where it is
necessary to impose a limit condition, it is necessary to use the solutions of
Section A.24. For all other cases, the above form is convenient and
applicable.

A.26 Tables of Spherical Bessel, Neumann, and Hankel

Functions
To evaluate spherical Bessel, Neumann, and Hankel functions, it is helpful
to factor each function into rational and transcendental parts. We introduce
rational functions Ae(a) and B^(a). In these terms the functions of
Eqs. A.25.14 and A.25.15 are:

Jl(o) = — {B^(rj)cosa + A^(a)sincj}

y^(a) = — {-A ^ (a) cos a + B^ (a) sin a} (A.26.1)

M a ) = i{B,(a)+fl\,(a)}e-*

Comparison of the equations shows, with q equal to any integer, that:

Appendix 415

/ \ vi (^ + s ) ! £-s)/2
( - l ) ^ - s ^ 8 ( ^ + s,2q)
(A.26.2)
!
B,(a)= V (i+s)
. x r i r / A^-S+D/2 5(^ + s,2q+l)

Comparison of the equations shows values of the letter functions. The

primary recursion relationship used to develop the table follows from
Eq. A.24.6:

J^+2(CT) = — — 'h+l{a)-h{°) (A.26.3)

Important related functions are obtained by operating on the radial

function to obtain the special function:

h (A.26.4)
'(°)=ad^CTh/^

Similarly to Eq. A.26.1, the related functions factor into rational and
transcendental parts:

jH CT ) = - { D K C T ) C O S C T + Q( C T ) s i n C T }
y*(a) = — {-C^(a)cosa + D^(a)sina} (A.26.5)

h*(a) = -{D^(CT)+ ^ ( a ) } e ~ *

Term-by-term comparison shows that:

-JU. = Bt(o) + Ct{o) — * L i = D,(o)-A,(o) (A.26.6)

It follows upon combining Eq. A.26.4 with Eqs. A.24.9 and A.24.13 that in
the limit of a vanishingly small radius:
416 The Electromagnetic Origin of Quantum Theory and Light

Lim . . , . (£+l)a£ 1 Lim . . . €(2^-l)!!

]A0)=—/ —r ^ny<(a)=AiM- (A 26 7)
- -
rj^OJn ' (2^ + 1)!! rj=>0

It follows similarly upon combining Eq. A.26.4 with Eqs. A.24.15 that in the
limit of an infinitely large radius:

Lim
n,
MCT) = — s i n a-f(* +l)
(A.26.8)
Lim 1 n,
y/(a) = — cos a-£(/
+ l)

^ A/O) B/(a) C,(a) D,(a)

1 0 0 1
0
1 l/o -1 -l/o^+l 1/a
6/a 2 -l
2 3/o 2 -l -3/CT -W+S/a
3 2
3 15/o- -67a -15/o +l -45/a 4 +21/a 2 -l 45/a 3 -6/a
4 105/oM5/o 2 +l -105/o 3 +10/a -420/o 5 +195/a 3 -10/a 420/a 4 -55/a 2 +l

Table A.26.1 Table of Values of the Radial Letter Functions

A 5 = 945/cr5-420/o-3+15/o-
B 5 =-945/o- 4 +105/a 2 -l
C 5 = -5(9!!)/CT 6 +2205/O 4 -120/CT 2 +1

D 5 = 5(9!!)/a 5 -630/a 3 +15/a

Appendix 41

A 6 = (11!!)/O 6 -5(9!!)/CT 4 +210O 2 -1

B 6 = -{11 !!)/o 5 +I260/o 3 -21/o

C 6 = -6(n!!)/cT7+31(9!!)/CT5-1680/o3+21/a
D 6 = 6(11 !!)/a 6 -8505/o 4 +231/o 2 -l

A 7 = (13!!)/o 7 -6(ll!!)/o 5 +3150/a 3 -28/a

B 7 =-(13!!)/o 6 +17,325/o 4 -378/o 2 +l
C 7 = -7(13!!)/o8+43(l 1 !!)/CT 6 -26,775/O 4 +4O6/0 2 -1
D 7 =7(13! !)/a 7 -l 31,670/a 5 +4662/o 3 -28/c

A 8 = (15!!)/a 8 -7(13!!)/a 6 +5(ll!!)/a 4 -630/o 2 +l

Bg = -(15!!)/o 7 +2(13!!)/o 5 -6930/a 3 +36/o
C 8 =-8(15!!)/CT 9 +57(13!!)/CT 7 -46(11!!)/CT 5 +8190/(T 3 -36/CT

D 8 =8(15!!)/o 8 -17(13!!)/o 6 +7(ll!!)/o 4 -666/a 2 +l

A 9 = (17!!)/o9-8(15!!)/o7+7(13!!)/o5-13,860/o3+45/a
B 9 =-<17! !)/a8+35(13! !)/o 6 -(13! !)/cr4+990/a2-l
C 9 = -9(17!!)/o 10 +9(17!!)/o 8 -70(13!!)/o 6 +17(ll!!)/a 4 -1035/a 2 +l
D 9 = 9(17!!)/a9-22(15!!)/rj7+l 1(13!!)/CT 5 -15,840/CT 3 +45/O-

Table A.26.1 Table of Values of the Radial Letter Functions (cont.)

418 The Electromagnetic Origin of Quantum Theory and Light

dA^/da = Q + B^ dB £ /da = D £ - A ^
2 i + l
B +B - B

£Ai_x-(£ + \)AM = (2£ + \)

_ da a *

ffl/_1-(^ + l)B / + 1 = (2^ + l) ^ - ^ + A ;

da a *
^ £
dA£/dc = A^+A^+B, dB^/da = B^ + B ^ j - A ,
a a
5. ^+1 £+1
dA^/da = A£-A^+1+B^ dB^/da = B^-B/+1-A^
a a
6. £+1
C
* = A
^ + A^-l = •A/-A M-l

a a

De - B f + B^_j B^ - B f + 1
7. 1(1 + 1)
dC//do = - A^ + D^
a2 .
1(1 + 1)'
6Dt/da = - 1 - B* — Ct
2
a
AeDe-Bfie =l
9. f „2^
A A =1 -C D
A - i ~ e-fie ^•e^i-i ^ - i / ~~
CT
v y
£+1 £+1
A
A
^+iD^ ~B/+iC' -• ^ + i -&ece+i ~'
a
10. 2^ + 3
A
^+2 B ^ A
A+2

l + 3[ (1 + 1)1 + 2)
C
^(+2®? ^+2

Table A.26.2 Radial Function Identities

Appendix 419

11. A , 2 - B / - C / +D/)

= Jda|4(A / -D / )(B / +C < )-2(A^-Bp / )^illJ

A/ +B / +C/ +D/)

= jdo^2(AeCe + BeDe)^^j
12. (A^-Cp,)
= -Jdaj[(A,-D,)2+(B,+C,)2 + (A,D, + B A ) ^ - 2 |

(A^+C^)

= J d a ( - A / + B / - c / + D,2) + (A,D, + B A ) : ^ - ^
' <7

13. (¥rBA)
= /da|[-(A,-D,) 2 +(B^+C,) 2 .(.i-<^) )
( A ^ + Bp,)

.lda{(-A 2 -B 2 + C 2 + D 2 ) + (A 2 + B 2 )(^)|

14. AA + B^-l-lf]
= |da|-2(A,-D^)(B^+C^) + 2 A , B i ^2

15. d
— (AA + B ^ + C ^ + Dp,,)

= 4 [ ^ + D(AA + B p n ) + n(n + l)(AnC^ + BnD,)]

a
Table A.26.2 Radial Function Identities (cont.)
420 The Electromagnetic Origin of Quantum Theory and Light

16.
—(AA-BA+CA-DA)
da
+1
=4[^ ) ( A P n - B , C n ) - n ( n + l)(A n D,-B n C,)]

17.
(A,Cn-C,An+Bpn-DfBn)
da
= \[n(n + \)-l{l + \)}(AiAn + B£Bn)
a
(A^Cn + C^A n -B^D n -D^B n )
da
—[n(n + l)-*(* + l)](A,An + B,B n )
a
+2(A/Dn + AnD^ + B<Cn + B n C / ) +
-2(A^A n -B £ B n -C^C n + D£Dn)

19.
- ( A A + B/An-QDn-DA)
da
\[l(l + D(AfDn + B^Cn) + n(n + 1)( A n D, + BnC£)]
a
= • +2(A,Dn + AnD^ + B£Cn + BnC^) +
- 2 ( A / A n - B / B n - C / C n + DPn)

20. /
A,B, (A/2 + B,2)
_d_ - ( A / +B / )
da v A ^ ~Be J ( A / - B ^

Table A.26.2 Radial Function Identities (cont.)

Appendix 421

A,D, + B,C,-(-l)'

1. 2 / 0 2 ; 2. 3 6 / a 4 - 1 8 / a 2 ; 3. 135o/a 6 - 7 2 o / a 4 + 7 2 / V

4. 88200/o 8 - 49350/o 6 + 6000/a 4 + 200/o 2

5. 8930250/V 0 -515970o/a 8 +69930o/a 6 - 31500/V + 4 5 o / a 2

6. 1296672300/a 12 - 76621545o/o 10 +11137014o/o 8 - 590058o/a 6

+ 123480/a 4 - 882/a 2

Table A.26.3 Radial Dependence of [ A ^ D ^ + B ^ - ( - l ) ]

i [A^-B^D,]
1
- l/o3 + 21a

2
- 1 8 / C T 5 + 33/(T 3 -6/O

3
- 6151a1 + 1250/O5 - 276/CJ 3 + 12/CT

4
- 44100/a 9 + 83475/c 7 - 20220/a5 + 13OO/03- 20/O

5
-4465125/a' 1 +8533350/O 9 -2201850/o 7
+ 169470/a5 - 4425/o3 + 30/a
6
- 648336150/c13 + 1247555935/a1'
-335975850/a 9 + 28797930/a?
- 961380/cr5 + 1220l/o 3 - 42/a

Table A.26.4 Radial Dependence of [A£C^ - B^D^]

422 The Electromagnetic Origin of Quantum Theory and Light

I [ A ^ + BjD,]
1 -l/o3
2
-18/o 5 -3/o 3
3
-675/o 7 -90/a 5 -6/a
4
-44100/a 9 -4725/a 7 -270/a 5 -10/a 3
5
-4465125/c 1 l
-396900/a9-\^900/a7-630/a5-\5/a3
6
-648336150/a13-49116375/c11-1984500/a9-56700/cr7-1260/a5-21/CT3

Table A.26.5 Radial Dependence of [A^C^ + B^D^]

£ 2(A<-D<XB<+C<)
1 0
2
36/a 5
3
2700/a 7 -360/a 5
4
264000/a 9 -65100/a ? + 1800/a5
5 9
35271500/CT 1 '-11510100/a + 642600/a 7 -6300/o 5
6
6483361500/a13-2436172200/a1' + 189162540/a9
-3969000/o 7 +17640/o 5

Table A.26.6 Radial Dependence of 2( A^ - D^ )(B^ + C^)

1
{A£-Def-(B£ + Ce)2
1 -1/a 4
9 6 4
2
-36/0-9/0
3
-2025/cr 8 +1440/a 6 -36/o 4
4
- 176400/O10 + 174825/o8-14400/o6+100/o4
5
- 22325625/a12 + 26195400/a 10 -3316950/a 8 + 81900/o 6 -225/o 4
6 -3890016900/a'4 + 5058986625/o12-802531800/O10+ 32345224/a8-335160/o6+441/o4

Table A.26.7 Radial Dependence of (A^ - D() - (B^ + C^)

Appendix 423

( A , - D , ) 2 + (B £ + Q ) 2
1 l/o 4
o 6 4
2
36/o +9/a
3
2025/CT8 +360/CT 6 + 36/a 4
4
176400/a10 + 23625/a8 + 1800/a6+100/a4
5
22325625/a12 + 2381400/O10 + 141750/a8 + 6300/o6+225/a4
6
3890016900/O14 + 343814625/a12 + 16669800/a'0
+ 593224/a8 + 17640/o6+441/o4
Table A.26.8 Radial Dependence of (A, - D( f + (Be + Ce f

A.27 Sums Over Spherical Bessel Functions

Any electromagnetic field may be expressed as the product of spherical
Bessel, Neumann, Hankel functions of a, or linear combinations thereof,
times linear combinations of Legendre functions of 9, times linear
combinations of trigonometric functions of azimuth angle $.
A particularly useful function is a z-directed plane wave:
e - / = e_/CTC0S . It follows that functions with m = 0 are present, and there
is no dependence upon (j). Since the function is regular on the z-axis, only
spherical Bessel functions are present. The result, expressed using spherical
coordinates, is the general solution form expressed as a sum over the single
product:

e -*c O S 0 = £ a ^( C T )p^ c o s e ) (A271}

The objective is to evaluate each of the infinite number of constants a^. To

do so, multiply both sides by Pn(cos9) and integrate over the full range of
zenith angle. The result is:
424 The Electromagnetic Origin of Quantum Theory and Light

^ \ * e = |sin6deP^(cose)e- /i:icose (A.27.2)

Differentiating both sides £ times with respect to athen going to the limit of
vanishing small radius, see Eq. A.24.7, gives:

,&i\-, L
T ~ = f sin edGP. (cos9) cos ^6 (A.27.3)
(2*+l) (2* + l)H J

The integral is listed in Table A.22.1.10. Doing the integration and solving
for ae gives:

&e = rt(2t + l) (A.27.4)

Combining Eq. A.27.1 with A.27.4 gives:

oo

e -/acos9 = £ r ' ( 2 ^ + l)j < (o)P / (cos9) (A.27.5)

Other related sums follow from Eq. A.27.5. Differentiating both sides of
Eq. A.27.5 with respect to 6 and using Table A.21.1.10 gives:

CTe-*cos8 = y / - ^ + 1)j Jc)^2^1 (A.27.6)

£f] sinG

Evaluating Eq. A.27.1 on the positive z-axis gives the three series:

oo

e-*=2r<(2*+l)j,(a)
' (A.27.7)
OO

(£_1)/2
sinrj=X(-l) ( 2 ^ + l)j^(a)
Appendix 425

c o s a = X ( - l / / 2 ( 2 f + l)JKa)
£e;0

Subscripts "o;l" and "e;0" indicate respectively odd integers beginning with
one and even integer beginning with zero.
Evaluation of Eqs. A.27.5 and A.27.6 at 9 = TC/2, see Table A.18.1, gives:

l ( 2 + ) j ( a )
^teO ' ' ^ '
(A.27.8)

£o;\ (^-l)!! 2

Application of Eq. A.26.4 to Eqs. A.27.8 results in:

a
ie-,0 (W.)
(A.27.9)

i V (2£±1)_«1_-v ^
£o;l

Use of Eq. A.7.6 to integrate Eq. A.27.5 over 6 gives:

-/acosG 8,
= £ ' J/(^+l(cose)-P<_1(cose)]g2 (A.27.10)
8, ^=0

Evaluation of Eq. A.27.10 between limits 0 = TC and rc/2 gives:

f
• m \ °°
1-e*
= X '^ - 1 J/(°)[P/ + i(o)-P,-i(o)Ho(a) (A.27.11)
^ ° J £=0
426 The Electromagnetic Origin of Quantum Theory and Light

Evaluation of Eq. A.27.11 between limits 6 = n/2 and 0 gives:

(.
= -(Jo(o)- X ' J/(<Ofr+l(0)-P*-i(0)] (A.27.12)
£=0

Subtracting Eq. A.27.11 from Eq. A.27.12 gives:

COSCT-1 ^ ,„. „ (^-1)! • , x

= V (2^ + 1) i/(a) (A.27.13)
a /T 0 (*-l)!!(* + l)!! J /
The operation of Eq. A.26.4 results in:

(
— = E (2^+1), t~/)! , j'(a) (A.27.14)
CT y
/= 0 (^-l)!!(^ + l)!!

Evaluating Eq. A.27.6 on the positive z-axis gives

1 °°
oe
~ * = ^ E / 1 - ' ( 2 ^ + 1 ^ + l)j/(a) (A.27.15)
The exponential breaks into the two equations:

1 °°
c coso = -JJ(-lf-l),2(2£+ !)£(£+1) i£(a)

(A.27.16)

osmo = ^(-l)m(2l + iy(e+l))e(G)

2
£e;2

Table A.27.1 contains a listing of sums over different functions of spherical

Bessel functions.
Appendix 427

L
e--cos9=£^(2^l)j,(a)P,(cose)
1=0
oo
2. cose
asin6e-* =£ /"'(^^((^(cose)

1-cos
a

a 4 ; '(* + l)!!J* >

6.
° ^e;0 V /"

a = y (V2 ^ + l ) 7 i * - j iMa )
^ Vl)!! '
"(2/+1) (l)!!
fi ^ (<-l).J<(0)

2
f=i
10.
(2- *)e~ t o = ± £ /-'(2/+1)(^ +1)^(0)
2
^=1

Table A.27.1 Table of Sums over Spherical Bessel Functions

428 The Electromagnetic Origin of Quantum Theory and Light

Multipolar Sources

A.28 Static Scalar Potentials

To examine physical sources of TM modal coefficients ¥(£,m), it is
convenient to start with the static scalar potential field, <&(#•). By Eq. 1.5.4
the differential equation that governs the scalar potential is:

V 2 0 = p/e (A.28.1)

By this equation the scalar potential has a spatial curvature only where
electric charges exist and a field function exists only if the curvature is other
than zero. It follows that static scalar potentials arise only from electric
charges.
To characterize such potentials we establish an origin near or in a region
that contains static electric charges. Based upon that origin, the coordinates
within the charged region are r(x,y,z). The field points at which the potential
is to be evaluated are r(x,y,z). The field point may be either interior or
exterior to the charge-containing region. The distance from the source to the
field point is R where, by definition:

1/2
R(,,r)= ( ^ x ) 2 + (^-y)2 + ( , - z ) 2 (A.28.2)

The potential at a field point was previously calculated, see Eq. 1.5.9,
and is given by:

*('') = T - f ^ d ^ (A.28.3)
47teJ R
It is convenient to work with distance from the origin to the field point, r, but
the function in the denominator of Eq. A.28.3 is the distance from the
differential source point to the field point, R. To replace 1/R by a function of
1/r, use the Taylor series expansion:
Appendix 429

1 a (\\ a a r1 a a a ^ i ^»

+ XiXj XiXjXk
r ox;i V R y V 2 a Xi a Xj U J / 6 aXj a Xj aXk vRy
R I a a a a f o
+
24 XiXjXkXm
a x i a x j a x k a x m \^Jr + ...
(A.28.4)

Placing the expansion of Eq. A.28.4 into Eq. A.28.3 results in the desired
form:

* ( / • ) =
4ne

(A.28.5)

The first term of Eq. A.28.5 has a first order singularity at the origin:

i q = Jp(r)dr
<&o(/-,e,4>)= 4ne r
where (A.28.6)

Succeeding terms have successively higher order singularities. Fields

associated with succeeding singularities are generated by equal values of
positive and negative charge, spaced incremental distances apart. The total
potential is the sum of that from each charge and, in the limit as the inter-
charge spacing goes to zero, the mathematical affect on the potential is the
same as obtained by differentiating with respect to x;. For example, for the
special case of charge separation zo all differentials in Eq. A.28.5 are z-
directed.
Consider the multipolar electric moment of a charge distribution with £
displacements. The moment is of order £ and degree m:

p f = Jp(r)x i x j ...x k dr (A.28.7)

430 The Electromagnetic Origin of Quantum Theory and Light

Moments are designated by the number of charges involved; moment pf,

where £ is the order, is formed by 2 charges. Charges are arrayed according
to the coefficients of a binomial expansion of the same order. Let a total of
{l-m) displacements be z-directed, let s of them be x-directed, and of them
(m-s) be v-directed.
Easily verified equalities satisfied by the Legendre polynomials are:

P/(cos0) =
VR; r
(A.28.8)
P^(cos8)sin(|) = i se f n
(t-tydydz (-1 vRy
Consider, as examples, structures that generate the lowest order
multipolar moments. A dipole consists of two discrete charges: charge q at
zo/2 and charge -q at -zo/2; the volume integral, Eq. A.28.7, over order one
is qzo and over any even order is zero. A linear quadrupole consists of four
discrete charges: charges q at +zo and -zo and charge -2q at the origin; the
2
volume integral over order two is 2qzo and over any odd order is zero. A
linear octupole consists of eight discrete charges: charge q at 3zo/2, -3q at
zo/2, 3q at-zo/2, -q at -3zo/2. For a source of order three, the volume
3
integral is qzo /4. The same integral over any even order is zero. The volume
integral over order one is zero but the volume integral for odd orders greater
than three is not zero.
In all cases, the volume integral of Eq. A.28.7 is zero if the charge
distribution and the displacements have opposite parity. It is also equal to
zero if there are fewer charges than the number of displacements. In all other
cases, the integral is non-zero. Tables A.28.1 and A.28.2 show some basic
features of common multipolar electric moments. In each table column one
shows the order of the source. Column two shows the discrete charge
distribution that generates that order of singularity. Column three shows the
volume integral of Eq. A.28.7. Column four shows the lowest non-vanishing
order of the potential. With the aid of the static portion of Eq. A.28.6,
column five shows the radial components of the generated electric field
intensity. Table A.28.1 is for only z-directed displacements and Table A.28.2
Appendix 431

is for one y- and (^-1) z-directed displacements. Table A.28.1 uses scalar
charge q as the unit cell and Table A.28.2 uses a y-directed dipole as the unit
cell.
The scalar potential of an arbitrary charge distribution is the simple sum
of values obtained from each moment:

*(/•,&•) = - T - X S ^TPf(cose)e-^ (A.28.9)

4TO r
£=o m=-l

This is the static scalar potential at an arbitrary, exterior field point due to the
charge distribution.
The radial component of the electric field intensity follows from
Eq. A.28.9 with the aid of the static portion of Eq. 1.6.3, and is equal to:

£ Charge Sites Vt 47te0>^(/-,8) 47ieE/<(/-,e)

0 +q at 0 q
r P-
1
-q at -zo/2 Pi = q^O ^PlfcosG) %P,(cose)
+q at zo/2 r
2 2r
*• +q at -zo P2= Pl- O ^ 2 (cos9) %2(cos0)
-2q at 0 r r
= 2q*§
+q at zo
3
-qat-3zo/2; P3 = 3P2Z0 p
|p 3 (cose) ^p3(cose)
+3q at -zo/2; = 6qzo r
-3q at zo/2
+q at 3zo/2
4 p 4 = 4p 3 z 0
+q at -2zo ^4-p4(cose) %P 4 (cos0)
-4q at -zo = 24q4
+6q at 0
-4q at z0;
+q at 2zp

Table A.28.1 Electrostatic Potentials of a Linear Array of Sources, I Charges Spaced

Distance ZQ Apart
432 The Electromagnetic Origin of Quantum Theory and Light

£ Charge and Sites

p{ 47ieE^(/-,9,(())
sine)) sin<j)

l +q at y0/2 P! =zWo
4P/(COS6) 4P/(COS0)
-q at-yo/2 r r
2 +q at (y0+z0)/2 A- = Pl^O 4P21(COS6) 4p](cose)
-q at (-yo+zo)/2 r r
+qat-<yo+zo)/2
-q at (yo-zo)/2
3 +q at (y0±2z0)/2 v\-= 2 P2^0 4P31(COS6) ^pj(cose)
-2q at yo/2 r
+2q at -yo/2
-q at (-yo±2zo)/2
4 +qat±(yo+3z 0 )/2 P\- = 3 P3^0 4P4(COS6) 4p4,(cose)
-3qat±(y 0 +z 0 )/2 r
-qat±(yo-3z 0 )/2
3qat±(y 0 -z 0 )/2

Table A.28.2 Electrostatic Source Potentials, One yo and (^-1) zo Charge Spacings

EArM) = -±-5, i ^Mpf(cos0)e-^ (A.28.10)

4718 r
^=0 m=-l

Connection between the coefficients of Eqs. 1.12.7 and those of

Eqs. A.28.10 is by direct comparison. In the limit as the frequency goes to
zero, combining Eqs. 1.12.7 and A.24.13 gives:

E,(^e,<t>)=-£ £ r^, w )^+i)%^pf(cose)e-^

£+2 CT
t=0 m=0
(A.28.11)
Direct comparison of Eq. A.28.10 with A.28.11 shows that:
Appendix 433

1 / ti+2
Vn
m

F(t,m) = — - , < (A.28.12)

For the special case of a z-directed electric dipole of moment mi, it

follows from the third row of Table A.28.1 that:

piCos6 , ^ 2picos0
<D,=-^
— — r- and
and E rr =
E = -- ^ —
- =- (A.28.13)
4ntr 4mr

A.29 Static Vector Potentials

To examine physical sources of TE modal coefficients G(£,m), it is
convenient to start with the vector scalar potential field, A(r). By Eq. 1.5.4
the differential equation that governs the vector potential is:

V2A(r)=\\j(r) (A.29.1)

By this equation the vector potential has a non-zero spatial curvature only
where electric currents exist. Since a field function cannot exist unless,
somewhere, the curvature is not zero, it follows that static vector potentials
arise only from electric currents.
Although the formal descriptions of the scalar and vector potential are
similar, the sources are not. Static scalar potentials arise from stationary
electric charges and vector potentials arise from moving ones. The integrated
form of Eq. A.29.1 follows from Eq. 1.5.8. With J(r) representing a
continuum charge distribution it is:

A r J(r) f r (A 29 2)
( )=r-f^rV --
4nJ R{r,r)
Combining the Taylor distance expansion of Eq. A.28.4 with Eq. A.29.2
gives:
434 The Electromagnetic Origin of Quantum Theory and Light

fj(r)d^+ fx i j(r)dJ' + fxiXij(r)d^

*HJ i
+—
(
d a a I * s
JxiXjxkj(r)dr
6 . dxj 3x: 9x k R i
(A.29.3)

Consider a filamentary current of magnitude I that is located in the xy

plane at z = 0. The current is at radius a and flows in the § direction. By
Eq. A.29.1, the potential component in a particular direction is proportional
to the current in that direction and there is no z-component of the current.
Therefore, there is no z-component of the potential. Since the current has
rotational symmetry about the z-axis, there is no loss of generality in placing
the field point in the xz-plane, (xo,0^o)- With the differential source at
position a(xcosty + _psinty) = xx+y}> the source to field distance is

R=[U-x) 2 +y 2 + ^ ] A . By Eq. A.29.2 the zero order vector potential is:

T 27t
/ = (A.29.4)
Ao( ') ~;— I <s{-5'sin<|) + -^cos<|))d(|) = 0

To evaluate the first order vector potential of the current loop, note the
partial derivatives:

1 1
dx
— =-2-sin9; —
i) -*
The first order vector potential is:

/ N (i sin 9 r
\I sino i „ x
ulna sin 9 -
Al{r) = 4 n J-— I #cos(|>»I(-.rsin<p + _j'cos<|>)»tfa<|) = •-Z- 2
r <{.(A.29.5)
0 471 r

Since the field point is in the xz-plane, the j-direction of Eq. A.29.5
generalizes to the § direction.
Appendix 435

To evaluate the second order vector potential, note the partial

derivatives:

= -1^ /3 .s .i n2z-9 - l .\) ; d2 ' ^ = 0 a2ro

dx2U j rr rr ~ ' d*3y vRy 3y^ V R
R
7,- /•

The second order vector potential is:

.271
1
I3sin 0 - l j j t f cos (|)«l(-.rsin<|> + >'cos<|))»«d<l)
0
A =0
»M-# 2TC
—j\a sin <|)»l(-.J-sin(|) + j'cos<|))»tfd<|)

(A.29.6)

Since all integrals vanish, so does the potential of this and all other even
orders.
To evaluate the third order vector potential, note the partial derivatives:

aV o 2
i3 ( p
= —rsin e(5sin 6-3); = 0;
dxJ vRy 3x 2 a y vRy

= -3
sin 9 a3ro =0
_4 '
3x8y2lR > / •
ar v R ;
The third order vector potential is:

3 —sin9|5sin 9-3] J (-sin <().*•+cos <|)j>) cos <>|c<

lp>
r
o
A 3 (,) = ^ - (A.29.7)
4 2«
-9—sin 9 I ( - s i n ^ + c o s ^ c o s ^ s i n <|)d(|)

Evaluation of the integrals gives:

436 The Electromagnetic Origin of Quantum Theory and Light

A 3(/.)=M^sme(5Sin2e_4) (A.29.8)

The radial component of the magnetic field follows by adding Eqs. 1.2.17,
A.29.5, and A.29.8:

Brl=
VT=V7PlM (A 29 9)
--
B,3 = ^cose(5c0s2e-3) = » ^ ^ ) (A.29.10)

The magnetic dipole moment of the loop is, by definition:

mx = m.\z=na Ii (A.29.11)

For non-circular loops, with S representing the planar area of the closed
current, the definition generalizes to:

m!=IS (A.29.12)

It follows from Eqs. A.29.5 and A.29.11 that the dipole potential is
expressible as:

A i1V(/•); = — m l, x (A.29.13)
An Kr2J

For the special case of a z-directed magnetic dipole of moment mi ;

combining Eqs. A.29.9, A.29.12 and A.29.13 gives:

. umisinG- 2nmicos0 ,»„„,,„

Al = r l (j) and Br= ^ l — (A.29.14)
4nr 4nr
The electric field of Eq. A.28.13 and the magnetic field of Eq. A.29.9 are
identical in form. Therefore, field form cannot be used to determine whether
a magnetic source consists of separated magnetic monopoles or a current
loop. In other words, the role played by a current loop in determining the
Appendix 437

vector potential is the same as that played by separated charges in

determining scalar potential.
Extension to higher order moments follows similarly to the dipole case.
An electric quadrupole consists of two superimposed sets of separated
electric charges and a magnetic quadrupole consists of two superimposed
sets of separated current loops. With linear displacements, current loops
follow the same rule as electric moments with one lateral and (^-1) linear
displacements. Results for several orders are compiled in Tables A.29.1
and A.29.2.
Tables A.29.1 and A.29.2 display features of multipolar magnetic
moments. In each case, column one shows the order of the source. Column
two shows the discrete current distributions that generate that order of
singularity. Column three shows the value of the volume integrals that
appear in Eq. A.29.3. Column four shows the lowest order derived
potentials. Column five shows the radial components of the generated
electric field intensity. Table A.29.1 is based upon a z-directed loop and
Table A.28.2 is based upon a j-directed loop. In both cases, all
displacements are z-directed.
Equating the field value of Eq. A.29.10 with the octupole moment of
Table A.29.1 shows that the octupole moment of a circular current loop is:

3l7I<74
m3 = (A.29.15)
4

The value of odd, higher order moments follow similarly.

Contrasting Tables A.28.1, A.28.2, A.29.1, and A.29.2 shows that
Tables A.28.1 and A.29.1 have identical structures, quite different potentials,
and identical force fields; the same is true of Tables A.28.2 and A.29.2.
Since there are electric monopoles but not magnetic monopoles, the £ = 0
case occurs only in Table A.28.1. For all higher order modes, Tables A.28.1
and A.29.1 have z-directed dipoles as the unit cell. Tables A.28.2 and A.29.2
have ^-directed dipoles as the unit cell, with identical structures except the
source structure of Table A.28.1, order £, is the same as that of Table A.29.1,
order (*+l).
All orders and degrees of any source may be described similarly to those
of the tables. The result is the radial component of the static magnetic field
component of arbitrary order:
438 The Electromagnetic Origin of Quantum Theory and Light

I Loop m£ 4nA¥(r,Q)/\i 4nBri(r,Q)/\x

Direction,
Sites
1 TtatO mj = I S 0 m
l • a
^^(cose)
^-sinG
r
2 Uat-zo/2 m2=2miz0 m-j
—^-3sin8cos6 ^p2(cose)
It at +zo/2 2r3 r
3 It at -ZQ m 3 = 3m 2 z 0 ^|sine(5cos2e-t) 4
?P3(cose)
2llat0 r
It at -zn
4 Uat-3z0/2 m 4 = 4m3Z 0 •^•sine(7cos 3 e-3cose) ^ - P ( c o s 0 )
4
3lt at -zo/2
3-Ll at z0/2
ft at 3ZQ/2

Table A.29.1 Magnetostatic Vector Potentials and Radial Fields of Sources, z-

Directed Current Loops of Area So with (f-l)zo Separations.
Arrows indicate direction of the unit cell magnetic dipole moment.

B,M) = f £ X fcihLpf(cos0)e-^ (A.29.16)

Similarly with the case of the electric field, in the limit as the frequency goes
to zero the radial component of the magnetic field of Eq. 1.12.7 goes to:

B
r = z £ %r£G(t,m)e(e+i)£l^p™(C0SQ)e-s>^ (A.29.i7)
c
£=0 m=Q CT

Combining Eqs. A.29.15 and A.29.17 gives:

G(*,M) = - ^ - , ^ (A.29.18)
Appendix 439

£ Loop ni/ 47iA^(r,e,<|>)/V 4rcB/<(/-,9,<|>)

Direction, Hsin(J)
Sites
1 atO mJ = I S 0 ptmj 2m
(0cos())-$cos6sin())J —^-P^cosB)
2
4n/-

<=at-zo/2; m
2 = m! z 0 ml
Asin9 + 26cos0jcos([>
^ip^cose)
=> at +zo/2
4 3cos 8-lsin<|)
2

=> at zo/2 m 3 = 2m2Z 0 -2/sin9cos6cos<|> ' ^Pi(cose)

2<=at0 3m^ 2
+e/5cos 6-l]cos<|>
=> at zo/2
-<|>l5cos 0-3lcos8sin<|)

«= at -3zo/2 rri4 = 3m3Zo

^ipi(cose)
3/11 - 5cos 6]sin6cos<|)
3 =* at -zo/2
3<= at +zo/2 +4e/5cos29 - 3)cos9cos(|)
$
=*• at +3ZQ/2
sin<|)
-30cos26+37 )

Table A.29.2 Magnetostatic Vector Potentials and Radial Fields of Sources, y-

Directed Current Loops of Area So, (£-l)zo Separations.
Arrows indicate direction of the unit cell magnetic dipole moment.
440 The Electromagnetic Origin of Quantum Theory and Light

References
W. Kaplan, Advanced Calculus, Addison-Wesley (1952)
J.D. Jackson, Classical Electrodynamics, 2nd ed., John Wiley (1975)
T.W. Korrner, Fourier Analysis, Cambridge University Press (1988)
R.B. Leighton, Principles of Modern Physics, McGraw-Hill (1959)
P.M. Morse, H. Feshbach, Methods of Theoretical Physics, McGraw-Hill (1953)
W.K.H. Panofsky, M. Phillips, Classical Electricity and Magnetism, 2 nd ed.,
Addison-Wesley, (1962)
S.A. Schelkunoff, Advanced Antenna Theory, John Wiley (1952)
S.A. Schelkunoff, Applied Mathematics for Engineers and Scientists, 2nd ed., Van
Nostrand (1965)
R.. Scott, Linear Circuits, 2nd ed, Addison-Wesley (1962)
W.R. Smythe, Static and Dynamic Electricity, 3 rd ed., McGraw-Hill (1968)
M.E. VanValkenburg, Network Analysis, 3 ed., Prentice Hall (1974)
J.R. Wait, Electromagnetic Wave Theory, Harper & Row (1985)
 Index

296, 313, 327, 328, 365, 376, 377, 405,

408, 410, 413, 414, 423, 426, 427
A biconical antenna, xv, 52, 53, 68, 70, 71,
absorbed power, 43, 92, 96, 103, 289 76, 119, 133, 157, 177, 192, 326
absorption cross section, 92, 107 biconical receiving antenna, xiv, 91, 95,
accelerating charge, 16, 191, 319 143, 289, 300
acceleration, 16, 17, 21, 24,193, 194,198 biconical transmitting antenna, 52, 70, 77,
active region, 297, 298, 302 78, 79, 95, 143
Aharonov-Bohm effect, 1 Bohm, 1, 321
anechoic chamber, 180,185 Bohr, 317, 320, 321
angular momentum, 154, 165, 191, 193, Bohr magneton, 235
212, 213, 215, 223, 234, 235, 260, 261, Bohr orbit, 231, 254, 305, 306, 307, 309,
263 311,312,319
antisymmetric, 12, 15, 236, 237, 238, Bohr radius, 218, 303, 304, 308, 309,
336, 337, 339 310,311
antisymmetric wave function, 236, 237, boundary condition, xvi, 28, 30, 31, 34,
238, 239 47, 52, 54, 60, 61, 62, 64, 65, 94, 95,
aperture, 53, 54, 64, 66, 67, 72, 76, 96 96,99,100,101,106, 230, 245, 326,
Aspect, 321 351, 361, 364, 366, 375, 402, 403, 404
associated Legendre function, 29, 31, 33, boundary value, xiv, 37, 221
38,215, 383, 384, 389, 390
asteroid, 2
atomic stability, xiii, xv, 193
c
attenuator, 182 cap, 73, 74, 76, 97, 107, 108
axial field, 283, 288 cap current, 107, 108
cap currents, 108
cat, 319, 320, 327
B Cauchy integral, 343
back-scatter, 44, 45 causality, vi, 322, 326, 345
bandwidth, 128, 129, 252, 315, 354, 357, caustic, 265
358 centrifugal force, 191
Bell, 319, 321 characteristic frequency, 239
Bessel differential equation, 40, 405, 410 characterization, 172, 179, 185, 190, 322,
Bessel function, 29, 31, 33, 38,49, 56, 326
57, 79, 80, 87, 96, 100,133, 264, 265, charge density, xv, 9, 12, 54, 62, 199,
267, 269, 277, 283, 287, 291, 294, 295, 200, 203, 206, 221, 238, 299, 300, 301,
302, 323, 345, 346, 361, 362, 364, 365
442 Electromagnetic Energy Exchanges: Continuous and Quantized

Chu, 131, 133, 135, 136, 138, 140, 142, difference and sum frequencies, 245
143, 152, 187, 189, 321, 326 dipole antenna, 157, 170, 174, 179, 181
Chu's limit, 189 dipole ntenna, 173
circularly polarized, 33, 140, 160, 161, dipole selection rules, 231
163, 164, 165, 176, 189, 254, 260, 266, Dirac, xvi, 191, 193, 203, 234, 242, 243,
284, 288 314, 319, 320, 327, 340, 345
circulator, 179, 180, 182, 189 Dirac delta function, 242, 340, 345
classical mechanics, 213, 317, 319, 321 directed power, 91, 93, 147, 174
closed system, 245, 354 direction cosine, 5, 332
coherent wave train, 251 directivity, v, xiii, 258, 326
complementary current, 61 driving phase, 182,184
complete, v, xiv, xvi, 27, 28, 30, 34, 52, dual, 35, 139, 141, 158, 165, 191, 300,
56, 95, 105, 192, 197, 205, 214, 221, 301
229, 234, 320, 321, 322, 326, 349 dynamic equilibrium, 192
complex power, xiv, xv, 76, 77, 78, 113, dynamic variable, 202
115, 117, 123, 124, 125, 126, 131, 140,
152, 353, 358, 361
complex Poynting vector, 41, 75, 88, 118,
E
295,359 eigenfunction, 209, 210, 211, 212, 220,
conducting boundary conditions, 34, 61, 221, 222, 223, 224, 225, 228, 229, 235,
62, 95, 361 319
conducting surface, 157, 289, 362, 364 eigenstate, vi, xiii, xv, xvi, 192, 198, 199,
cone-cap junction, 73, 97 200, 208, 211, 212, 240, 245, 249, 260,
conjugate variable, 202, 203, 341, 343 313, 314, 320, 322, 323, 324, 328
constant current source, 96 eigenstate electron, xiii, 192, 199, 260,
constant voltage source, 96 308,313,322,323,324
Coulomb field, 223 eigenvalues, 62
Coulomb force, xv, 191, 196, 198,199, Einstein, 240, 242, 258, 318, 319, 320,
308, 309, 323, 324 321, 326
Coulomb pressure, xvi, 302, 305, 309, electric moment, 158, 183, 184, 429, 430,
311 437
Coulomb trapping pressure, 245 electrically long antenna, 78
Coulomb's law, 1, 302, 317 electrically short antenna, xiii, 77
current densities, xv, 62, 73, 96, 97, 106, electrodynamics, 1, 9, 193
155,168,198,199, 200,208, 299, 300, electromagnetic mass, 21
303, 313, 323, 326 electromagnetic potential, 340
current density, 9, 12, 15, 45, 54, 62, 73, electron model, 192, 198, 205, 328
106,107,206, 207, 301, 324, 346, 361, electron spin, 203, 234, 319
364 electron stream, 239
current eddies, 301 electrostatic force, 192, 200, 213
current loop, 54,158, 300, 319,434,436, electrostatic potential, 199, 431
437, 438, 439 elliptical orbit, 191
embodiment, 142, 184, 190
energy absorption, 231, 314
D energy conservation, xv, 55, 145, 199,
deductive approach, 1, 13 203, 228, 246, 286, 289, 296, 313, 323
degenerate, 211, 212, 315 energy emission, 231, 321
Index 443

energy quanta, 318 Grangier, 321

energy-to-momentum ratio, 111 Green's function, 340, 347, 350
equivalent circuit, 132,133, 134, 136,
139, 140, 142, 152, 167
exclusion principle, 192, 238
H
expectation value, 200, 201, 204, 205, half power point, 129, 357
206, 326 Hamiltonian operator, 208, 223, 225, 226,
exterior coefficient, 69 236
exterior mode, 66, 68, 96, 107 Hankel function, 29, 33, 40, 55, 75, 87,
exterior region, 53, 54, 55, 65, 75, 100, 99, 143, 158, 259, 263, 264, 283, 287,
360, 365 288,290,365,413,414,423
extinction cross section, 44, 49, 50, 92 Hansen, 30, 31
extinction momentum, 50 Hansen's method, 31
extinction power, 43, 44, 46, 92, 94, 96, harmonic oscillator equation, 127
109, 289 Harrington, 258, 260, 321
Helmholtz equation, 27, 31, 347
F Hertz, 239, 318
historic interpretation, v, xiii, 196, 319,
FDTD, 172, 173, 185, 187 328
feed network, 172, 179 homogeneous equation, 127
field coefficient, 56, 59, 82, 86, 98, 104, homogeneous Maxwell equations, 14, 22
106,167,185, 299, 301 Huygens, 317, 318
field energy, xiii, xv, 24, 35, 130, 131, hydrogen atom eigenfunction, 220, 222
170, 254, 272, 279, 318, 354, 357
field stress, 302
field-associated energy density, 147
I
forbidden, 228 incomplete, 192, 205, 320, 321, 326
forward-scatter, 45 input admittance, 53, 59, 60, 61, 139
Fourier integral, 26, 200, 205, 209, 324 input impedance, 70, 71, 75, 103, 115,
Fourier integral transform, 200, 205, 209, 128, 129, 131, 133, 134, 139, 140, 147,
324 152, 172, 180, 352, 354
Fourier transform pair, 341 input reactance, 119, 133, 135, 142
Fresnel, 317 instantaneous power, 122, 123, 124, 173,
174, 175
G integral equation, 65, 101, 347
integrating oscilloscope, 179
gain, xiii, 160, 161, 258, 260, 265, 321 inter-element phasing, 176
gamma function, 367 Interior currents, 108
gamma power, 123 interior mode, 68, 73, 96
Gaussian wave function, 203, 205 interior region, 54, 56, 73, 75, 99, 100,
generalized force, 131 108, 365
generalized force and flow, 131,138,196
generator, 172, 179, 180, 189
geometric cross section, 43,44, 46,49,
K
91,92 kinematic properties, xvi, 22, 154, 200,
g-factor data, 193 260, 265, 326
gluons, 323
444 Electromagnetic Energy Exchanges: Continuous and Quantized

kinematic variable, 198 Mehra, 319

Kirchhoff circuit laws, 117 Method of Moments, 184
Kronecker delta function, 81, 90, 271, Mie, 321
333, 372, 405 Mie region, 50
Minkowski force, 7
modal coefficient, xiv, 67, 68, 118, 259,
L 327,428,433
Laplacian, 10, 12, 28, 213, 214, 340, 347 modal coupling, 73, 188
Laplacian operator, 213, 347 modal impedance, 132
Legendre function, 29, 30, 31, 33, 38, 54, MoM, 184
55,56, 63, 65, 66, 67, 68,79, 99,106, momentum densities, 200
215, 221, 279, 281, 282, 374, 383, 384, momentum space, 202, 204
389, 390, 397, 402, 406, 423 monochromatic radiation, 318
Legendre polynomial, 40, 72, 79, 97, 221, monopole, 221,436, 437
271, 274, 379, 380, 381, 382, 385, 386, multi-element antenna, 176
387, 393, 397, 401, 430 multipolar electric moment, 429, 430
Lenz's law, 97, 107, 108, 300 multipolar expansion, xvi, 117, 143
Lienard-Wiechert potentials, 346 multipolar moments, 299, 301, 327, 430
line admittance, 58, 59, 60 multipolar sources, 428
linearly polarized, 53,162,163,176
local coupling, 184 N
Lorentz, 5, 317, 319, 323
Lorentz contraction, 4, 9 network analyzer, 179, 180, 182
Lorentz frame, 8 Neumann function, 29, 31, 33, 56, 57, 79,
Lorentz radius, 193, 253, 319 133, 263, 264, 265, 267, 270, 271, 274,
Lorentz transformation, 3, 5 279, 282, 283, 287, 291, 294, 295, 296,
313, 328, 377, 409, 410, 413, 414, 423
Newton, 317, 318
M Newton's law, 7, 21,304
magnetic current density, 301 nonhomogeneous Maxwell, 14
magnetic moment, xv, 97, 108, 161, 165, nonhomogeneous Maxwell equations, 14,
183, 184, 191, 193, 223, 234, 235, 300, 22
313,319,327,437 non-ionizing transition, 225
magnetic potential, 1 nonlinear, v, xiii, xvi, 211, 246, 247, 248,
magnetostatic vector potential, 438, 439 250, 260, 290, 314, 321, 323, 326
Manley, xvi, 321, 323 nonlinear source, 250
Manley Rowe equations, xvi, 245, 314, nonlocal, v, xiii, xvi, 191, 321, 322, 323,
326 326
many-electron problem, 236 normalized bandwidth, 357
matrix element, 5,155, 227, 228, 231, Norton circuit, 358
303 nova-like expansion, 308
Maxwell, 317, 318 nucleons, 323
Maxwell curl equation, 30, 290 numerical solution, 52
Maxwell equations, 1, 14, 16, 22, 34, 52,
193, 349,358, 359
Maxwell stress tensor, 17, 19, 155, 167,
o
302, 357 occupied state, 223, 299
Index 445

oil droplet, 198 quasi-continuum, 240, 241, 242

omnidirectional, 143, 170
operator form, 202, 212, 213, 228, 229
optical region, 50
R
orthogonal, 28, 65,66, 67, 68, 96,163, radar cross section, 44, 51
176, 210, 211, 236, 331, 371, 385, 386 radiation pattern, 53, 151, 258, 287, 289
radiation Q, 127, 130, 131, 172, 177, 178,
P 179, 182, 183, 184, 187, 188, 251, 254,
255
parasitic coupling, 182 radiation reaction force, v, xv, xvi, 191,
parity, 56, 86, 144, 275, 277, 323, 379, 194, 195, 196, 197, 198, 245, 279, 299,
389, 403, 404, 430 313, 323, 324, 326, 328
permeability, 10 radiation reaction pressure, xvi, 155, 157,
permittivity, 11, 247, 362 158,167,169,302,303,311
perturbation, 223, 229, 231, 308 rank, 12, 17, 19, 331, 332, 334, 335, 336,
perturbation equation, 229 337, 338, 339
phase shifter, 172, 182 Rayleigh region, 50
phase-quadrature, 115, 176 reactive energy, 137, 174, 192, 195, 197,
photo effect, 239 328, 354
photocurrent, 239, 240 reactive radiation reaction force, 195,
photoelectric effect, 242, 318 196,198
photon, vi, xv, xvi, 245,251,252,260, recursion relation, 263, 264, 272, 277,
265, 295, 297, 313, 326, 328 281, 284, 286, 327, 328, 378, 387, 388,
Planck, 258, 318, 328 390, 398, 399, 408, 412, 415
Planck's constant, 199, 203, 249 reflected waveform, 180, 182
point mass, 234 related functions, 57, 256, 415
power splitter, 172 remainder, 278, 284, 285, 286
power-frequency relationships, v, xiii, remarkable stability, 191, 192, 193
xvi, 245, 250, 314, 319, 321, 323 requires, v
Poynting vector, 19, 41, 45, 75, 76, 88, resonance, 70, 129, 133, 134, 135, 167,
118, 120, 144, 147, 150, 160, 162, 194, 265, 356, 357
258, 259, 261, 284, 286, 288, 295, 296, resonant antenna, 77
298, 328, 359 retarded potentials, 346
Poynting's theorem, 41 rigid surface, 168
principal current, 61 ring charge, 74
probability coefficient, 226, 229 Ritz power-frequency, xvi
probability coefficients, 229 Rodriques formula, 382, 384
probability density, 203, 236, 313 Roger, 321
projectile, 317, 318 Rowe, xvi, 321, 323
proper time, 6, 15
s
Q scattered field, 40, 45, 49, 87, 92, 93, 95,
quadrilateral symmetry, 221 103, 299
quantum electrodynamic, 193 scattered power, 40, 43, 44, 45, 91, 92,
quantum theory operator, 210, 213 290
quarks, 323 scattering cross section, 43, 49, 91, 110
446 Electromagnetic Energy Exchanges: Continuous and Quantized

scattering parameters, 43,180,182 T

SchrOdinger, xv, xvi, 198, 199, 203, 205,
206, 208, 213, 223, 225, 228, 234, 245, Taylor series expansion, 290, 428
313,314,319,320,323,327 TE coefficient, 54
Schrodinger equation, 198, 199, 203, 205, TE mode, 39, 40, 82, 87, 97, 98, 106,
206, 208, 213, 223, 228, 245, 319 118,119,122,138,139, 140,141,142,
self-consistent field analysis, 290 143, 158, 255, 257, 258, 298
separation constant, 28, 29, 30, 33, 214, telefield, 266, 287, 290
215, 375, 379, 410 TEM mode, 53, 56, 57, 60, 61, 107, 143
singlet, 221 TEM transmission line, 60
solar radiation, 302 tensor contraction, 335
source region, 53, 54, 273, 299, 360 tensor field, 302, 334, 337
source turn-off, 173, 175, 176, 177, 179, tensor introduction, 331
180, 187 tensor operations, 334
source-associated energy density, 146 tensor symmetry, 336
source-free region, 24, 33 terminal pair, xiv, 143
spherical scatter, 43,47, 87, 91,92 terminator admittance, 59
spherical wave, 318 thermodynamic, xv, 199, 200, 208, 313,
spin wave function, 236 318,323
stable arrays, 198 Thevenin circuit, 115, 358
standing energy density, 145, 151 Thompson, 317
Stark effect, 223 thrust, 50, 92
state value, 175, 210, 211, 212, 218, 221, time dilatation, 4
327 time-dependent equation, xv, 206, 208,
static scalar potentials, 428,433 314
static vector potential, 433, 438, 439 time-dependent Schrodinger equation,
statistical mechanics, 198, 199, 318, 321, 205, 228
324, 327 time-independent equation, 207, 313, 314
Stirling formula, 369 time-independent Schrodinger equation,
stopping potential, 239 198, 205, 208
stored energy, 135, 137, 355, 358 TM mode, 40, 53, 61, 83, 87, 106, 107,
stress tensor, 17, 19, 155, 167, 302, 306, 119, 122, 123, 138, 139, 141, 142, 143,
308, 332, 357 245, 256
string theory, 193 torque, xv, 191, 313, 327
surface current density, 54, 62, 73,106, transient-capturing oscilloscope, 179
364 transition, vi, xv, xvi, 3, 10, 208, 225,
surface currents, 45, 96,108,130, 363 228, 231, 234, 240, 241, 245, 250, 254,
surface power, 90,120, 131, 258, 295, 260,311,314,319,326
296 transmitting antenna, 52, 53, 70, 74, 75,
surface tension, xv, 198 77,78,79,95,96,98,111,143
symmetric, 17, 19, 108, 183, 236, 237, trapping potential, 223
238, 319, 336 trigonometric function, 29, 80, 96, 246,
symmetric wave function, 236, 237, 238 268, 307, 352, 371, 373, 374, 376, 423
symmetry, 1, 5, 33,47, 52, 53, 54, 55, 56, triplet, 221
98, 99, 107, 117, 118, 213, 221, 236, truncation, 69, 70
237, 238, 258, 260, 299, 318, 336, 377, turnstile antenna, 173, 176, 177, 178, 179,
379,434 182, 183
Index 447

u w
Uncertainty Principle, 203, 205, 252 wave equation, xv, 203, 318, 323, 347
uniqueness, 266, 364 wave train, 251,252, 253, 254, 297, 298,
Uniqueness Theorem, 266 307, 310, 327
waveform generator, 180
V
Y
virtual shell, 133, 192
virtual sphere, 75, 88, 131, 156,157, 167, Yeecell, 185
172, 173, 174, 175, 184, 196, 259, 295, Young, 317
310, 348, 349, 360
virtual surface, 41, 92, 107, 157, 362, 364
z
Zeeman effect, 223
 This book presents a rigorous application of modern electromagnetic field
theory to atomic theory. T h e historical view of quantum theory was
developed before four major physical principles were known, or understood.
These are (1) the standing energy that accompanies and encompasses
electromagnetically active, electrically small volumes, (2) the
power-frequency relationships in nonlinear systems, (3) the possible
directivity of modal fields, and (4) electron nonlocality. The inclusion of

THE ELECTROMAGNETIC ORIGIN OF

QUANTUM THEORY AND LIGHT
these four effects yields a deterministic interpretation of quantum theory
that is consistent with those of other sciences; the quixotic axioms of the
historically accepted view of quantum theory are not needed. T h e new
interpretation preserves the full applicability of electromagnetic field
theory within atoms, showing that the status of all physical phenomena
— including that within atoms — at any instant does completely specify
the status an instant later.

ISBN 981-02-4785-0

www. worldscientific.com
4841 he