This action might not be possible to undo. Are you sure you want to continue?
IEEE PRESS SERIES ON ELECTROMAGNETIC WAVE THEORY
The IEEE Press Series on Elcctromagnetic Wave Theory consists of new lilIes as well as reprints and revisions of recognized classics that maintain longterm archival significance in ela:tromagnelic waves and applications.
Series Edilor
Donald G. Dudley
Universit)' ojArizona
Advisory Board
Robef1 Il Collin
Case Western Re;fe/w Uniw!rsity
Akira lshimaru
Unil'ersio' of Washing/on
D. S. Jones
UnillfrsiIYQ!D,llldee
Associatt Edilors
ElECTROM ...ONnIC THEORV, ScATIEIUNC.
ANI) DWFIlACTION
INTEGRAL EQUATION MF.THODS
Donald R. Wilton
University of Hou£tol1
Ehud Heyman
TelAviv Ulliw!r$ily
DIFFEREr'TIAL EQUATIO:" METHOOS
ANTENNAS. PROl'AGATION. AND MI(:ROWAVIOS
Andreas C. Cangellaris
David R. Jackson
University ofAri:ona
Unil'ersity ofHouslOn
BOOKS IN THE IEEE PRESS SERIES ON ELECTROMAGNETIC WAVE THEORY
Christopoulos. C. The TransmissionUne MQdeling Me/hads: TLM Clcmmow. P. c.. The Plane WaveSpeclrnm Represemmion o[£leclromagnetic Field~' Collin. R. B.. FieM Theoryo[Guided Waves. Second Edition Collin. R. E.. Foundations [Qr MicrQwave ElJgilJeering Dudley, D. G.. Mathematical FQundatiollS[or Elec/romagne/ic Thl'oJy Elliot, R. S., Electromagnetics: lIis/Qty. TheQry. and Applicll/iQns Felscn. L. 11.. and Marcuvit"z. N.• Radiation and Scallerillg o[Wal't's Harrington. R. F.• Field ComplllariQII by MOIllI'Il/ Metho<ls Hansen I:t at, Plalle· Wave Theory (I[TimeDomain Fields: NearField $calllling ApplicaliQIIS lshimaru, A.. Wave Propag(J/ion and Scullering ill Rani/om Media Jones. D. S., Me/hads in Electromaglletic Wm't' Propagation. Second Edition Lindel1. l. Y., Methoos[or Elecrromagm:tic Field Analysis Peterson el at. Computational Merhads[or Electromagnerics Ta;, C. T.. Generalized VeclOr ami Dyadic Analysis: Applied Mathematics in Field TheOlY TaL C. T.. Dyadic Green FIIIIClions ill Elec/romagnetic TlleD/Y. Second Edition Van madel, J.. Singular Elec/romagnetic Fields and Sources Volakis et al.. Fini/e Element Method[ol' Elecrromagnetic:;: Allt(!lIIlas. Micro....ave Cirellits. and $collering ApplicaliQlls Wait, J., ElectromagnetiC Waves ill S/ra/j!ed Media
An IEEE Press Classic Reissue
TIMEHARMONIC ELECTROMAGNETIC FIELDS
Roger F. Harrington
Professor ofElectrical Engineering (retiraJ) Syracuse Uni''€rsity
IEEE Antennas & Propagation SocielY. Sponsor IEEE Microwave Theory and Techniques Society, Sponsor
+IEEE
IEEE Press
rnWILEY
~INTERSCIENCE
JOHN WILEY & SONS, INC. New York • Chichester • Weinheim • Brisbane • Singapore • ToronlO
This text is printed on
acid~free
paper.
®
Copyright (() 2001 by the Institute of Electrical and Electronics Engineers, Inc. All rights reserved.
No part of this pUblication may be reproduced, stored in a retrieval system or transmitted in any fann or by any means. electronic. mechanical, photocopying, recording. scanning or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act. without either the prior written permission ofthe Publisher, or authorization through payment of the appropriate percopy fee to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, (978) 7508400, falC (978) 7504744. Requests to the Publisher for pennission should be addressed to the Pennissions Department, John Wiley & Sons, Inc., 605 Third Avenue, New York, NY 101580012, (212) 8506011, fax (212) 8506008, EMail: PERMREQ@ WILEY,COM,
For ordering and customer service. call 1800CALLWILEY.
Library ofCongress Cataloging in Publication Data is available.
ISBN 047120806X
Printed in the United States of America.
10987654321
IEEE Press 445 Hoes Lane, P.O. Box 1331 Piscataway, NJ 088551331 IEEE I'ress Edilorial Board Stamatios V. Kartalopoulos, EditfJr in Chief
M. Akay
J. B. Anderson R. J. Baker J. E. Brewer
M. Eden
M. S. Newman
M. Padgen W. D. Rcevc C. Zobrist
M. E. ElHawary
R. F. Herrick
R. F. Hoyt D. Kirk.
Kenneth Moore, Direc/or qf Books & In/ormation Se,.vice.~ Catherine Faduska, Seniol' Acquisirion.~ Editor
IEEE Antennas & Propagation Society, Sponsor AP·S Liaison to IEEE Press, Robert Mailloux IEEE Microwave Theory and Techniques Society, Sponsor MTIS Liaison 10 IEEE Press. Karl Varian Cover design: William T. Donnelly, WT Design
FOREWORD TO THE REISSUED EDITION
The purpose of the IEEE Press Series on Electromagnetic Wave Theory is to publish books oflongtcnn archival significance in electromagnetics. Included are new
titles as well as reprints and revisions of recognized classics. TimeHarmonic Elec· lromagrtetic Fields. by Roger F. Harrington, is one of the most significant works in electromagnetic theory and applications. The book has been one oflhe principallcxls in the first graduate course on electromagnetic theory for the past fony years; many would say lhe principal text. This classic volume contains a complete coverage of dynamic fields and is as fresh today as it was when originally published in 1961. TimeHamlOnic Electromagnetic Fields has proved to be popular over the past 40 years with students, professors, researchers and engineers who require a compre~ensive, indepth treatment of the subject. Indeed, a colleague of mine, Dr. Kendall F. Casey of SRI writes, as follows: "When I begin a new research project, I clear my desk and put away all texts and reference books. Invariably, Harrington's book is the first book to find its way back to my desk. My copy is so worn that it is falling apart." Another colleague, Professor Chalmers M. Butler of Clemson University adds the following: "In the opinion of our faculty, there is no other book available which serves as well as Professor Harrington's does as an introduction to advanced elcctromagnctic theory and 10 classical solution methods in electromagnetics." Professor Harrington has been an internationally wellknown contribUlor to eleetromagnetics for many years. He is universally regarded as the "father" of the Method of Moments. His book on the subject, Field Computation by Moment Methods, was added to the series in 1993. Professor Harrington is a Fellow of the IEEE. Prior to his retirement from active teaching, he was a Distinguished Professor at Syracuse University. Among his many awards and honors, he was awarded the IEEE Centennial Mcdal in 1984, the IEEE Antennas and Propagation Society Distinguished Achievement Award in 1989, the URSI Van der Pol Medal in 1996, the Jubilee Tesla Medal in 1998, the IEEE Electromagnctics Field Award in 2000, and an IEEE Third Millenium Medal in 2000. It is with pleasure that I welcome this classic book into the series.
DONALD G. DUDLEY University ofArizona Series Edito,' IEEE Press Series on Electromagnetic Wave Theory
vii
PREFACE
This book W8..'l written primarily as a graduatelevel text, but it should also be useful lIS a reference book. The organization is somewhat different from that normally found in engineering books. The material is arranged according to similarity of mathematical techniques instead of according to devices (antennas, waveguides, cavities, etc.). This organization reflects the main purpose of the bookto present mathematical techniques for handling electromagnetic engineering problems. In the sense that theorems are proved and formulas derived, the book is theoretical. However, numerous practical examples illustrate the theory, and in this sense the book is practical. The e.'qlerimental lISpect of the subject is not considered explicitly. The term timeharmonw has heen used in the title to indicate that only sinusoidally timevarying fields are considered. To describe such fields, the adjective ae (alternatingcurrent) haa heen borrowed from the corresponding specialization of circuit theory. Actually, much of the theory can easily be extended to arbitrarily timevarying fields by means of the Fourier or Laplace transformations. The nomenclature and symbolism used is essentially the same lI.8 that of the author's earlier text, "Introduction to Electromagnetic Engineering," except for the following ehange. Boldface script letters denote instantaneous vector quantities and boldface block letters denote complex vectors. This is a departure from the confusing convention of using the same symbol for the two different quantities, instantaneous and complex. Also, the complex quantities are chosen to have rms (roo~meanl!quare) amplitudes, which corresponds to the usual ac circuit theory convention. The many examples treated in the text are intended to be simple treatments of practical problems. Most of the complicated formulas are illustrated by numerical calculations or graphs. To augment the examples, there is an extensive set or problema at the end of each chapter. Many of these problems are of theoretical or practical significance, and are therefore listed in the index. Answers are given for most of the problems. Some of the material of the text appears in book form for the first time. References are given to the original sources when they are known.
ix
x
PREFACE
However, it has not been possible to trace each concept back to its original inventor; hence many references have probably been omitted. For this the author offers his apologies. Credit has also been given to persons responsible for the origina.l calculations of curves whenever possible. A bibliography of books for supplemental reading is given at the end of the text. The book has been used for a course directly following an introductory course and also for a course following an intermediate one. On the former level, the progress was slower than on the latter, but the organization of the book seemed satisfactory in both eases. There is more than enough material for a year's work, and the teacher will probably want to make his own choice of topics. The author expresses his sincere appreciation to everyone who in any way contributed to the creation of this book. Thanks to W. R. LePage, whose love for learning and teaching inspired the author; to V. H. Rumsey, from whom the author learned many of his viewpoints; to B. Gruenberg, who read the gaUeys; to colleo.gues and students, for their many valuable comments and criticisms; and, finally, to the several secretaries who so expertly typed the manuscript.
Roger F. H arringtqn
CONTENTS
Foreword to the Revised Edition.
vii
Preface
Chapter 1.
Fundamental Concepts
11. 12. 13. 14. 15. 16. 17. 18. 1·9. 110. 111. 1_12. 113. 114.
Introduction. Basic Equations. Const.itutive Relationships. The Generalized Current Concept Energy and Power . Circuit Concepts Complox Quantities. Complex Equations. Complex Conatitutive Parametera Complex Power AC Characteristica of Matter A DilKlU8$ion of Current . AC Behavior of Circuit Elements Singularities of the Field . Chapter 2. Introduction to Waves
• 7
9 12
1 1
13
16
18
19 23
26 29
32
21.
The Wave Equation
37
41
22. Waves in Perfect Dielectrica 23. lutnusia Wave Constanta . 24. Waves in Lossy Matter
25. Reflection of Waves 26. Transmissionline Concepts 27. Waveguide Concepts 28. Resonator Concepts 29. Radiation 210. Antenna Concepts . 211. On Waves in General
48 51
54
61 66
74 77
81 85
Chapter 8.
31. 32. 33.
34.
Some Theorems and Concepts
35. 36.
The Source Concept Duality . Uniquenesa . Image Theory The Equivalence Principle.
95
98 100
103 106
110
Fields in Halfspace
2.. The Rect. DipOle and Conducting Sphere . 45. 311. . Other Radial WaveguidC8 .ngular Cavity Partially Filled Waveguide The Dielcctricslab Guide . Orthogonality Relationships Space as a Waveguide . Wave Transformations. Surfaceguided Waves Modal Expansions of Fields Currents in Waveguide8 Apertures in Ground Planes Plane Current Sheela Chapter 6. 43. . .. The Circular Cavity Other Guided Waves Sources of Cylindrical Waves.4... PilUle Wave FunctioDs 113 116 120 123 126 129 132 . 67. 66. 511. 69. Apertures in Wedges Chapter 6.. 1·6. 313.. 63. 66. Construction of Solutions The Radiation Field Chapter •. Scattering by Spheres .. Twodimensional Radiation Wave Transformations. 312... 66. Spherical Wave Functions 198 20' 208 213 216 223 228 230 232 238 2'2 245 260 61. 513. 68. 62. 39. Other Resonators 67. 38. 410. Scattering by Cylinders Scattering by Wedges .9. 66. 512. The Rectangular Waveguide Alternative Mode Seu.. ~IO. 26' 269 273 276 279 283 286 289 292 298 .ion Theorem Reciprocit. Bources of Spherical Waves 68.xii 37. .... 510. The Wave Functions The Circular Waveguide RAdial WaveguidC8.1. 4t l. 310. ..7. The Spherical Cavity 63.a. 64. CONTENTS The rnduct. The Wave Functions 62. 69.y Groon's Functions Tensor Green's Functions. Cylindrical Wave Functions 143 1'6 "8 162 166 158 163 168 171 177 ISO 186 61. The Wave Functions Plane Waves .. 4_12. 48. Integral Equation8 . Threedimensional Radiation Apertures in Cylinders.
811. g. The Riu Procedure. Oneporl Networn. C.roduction. 78. 77. 88. Aperture Coupling to Cavities A. Obstaclea in Waveguides Posts in Waveguides Small Obstacles in Waveguides Diaphragms in Waveguides Waveguide Junctions . 7+6. B. 79. ~2. The Reaction Concept. 85. D. Stationary Formulas for Waveguides. 813. Waveguide Feeds . E. Legendre Functions 381 389 391 393 398 402 406 411 414 420 425 428 431 434 436 447 451 Appendix Appendix Appendix Appendix Appendix 456 460 465 471 Bibliography . Veetor Analysis Complex Permittivities Fourier Series and Integrals Bessel Functions . 72. 83. 812. Stationary Formulas for Cavities. Apertures in Spheres .ximum Antenna Gain Chapter 7. 89. Excitation of Apertures Modal Expansions in Cavitics. Stationary Formulas for Impedance . Perturbational and Variational Techniques 71.t. Twoport Networks. 73. Cylindrical Waveguides Modal Expansions in W&Veguides The Network Concept. 810.. Probes in Cavities . 74. 710. Chapter 8.. 75. Perturbations of Cavity Walls Cavitymaterial Perturbations Waveguide Perlurbations . Ficlds External to Cones Ma. 711. 815. 613. 86. 87.. 612.CONTENTS xiii 301 303 307 ~ll.. Microwave Networb 317 317 321 326 331 338 340 345 348 355 362 365 g. 814. Stationary Formulaa for Scattering Scatlering by Dielectric Obstacles Transmiaaion through Apertures. .1. 473 . Int. lndu . 712.
We consider these units to be fundamental units. called the magnetw intensity .l to those used in practice. $.CHAPTER 1 FUNDAMENTAL CONCEPTS 11. caUed the electric intensity (volts per meter) :Il..units of all other quantities depend upon this choice of fundamental units. and are called secondary units. more convenient for our purposes. (amperes per meter) called the electric flux density (coulombs per square meter) called the magnetic flux den8ity (webers per square meter) called the electric current density (amperes per square meter) q. The topic of this book is the theory and analysis of electromagnetic phenomena that vary sinusoidally in time. The concepts necessary for our study are but a few of the many electro~ magnetic field concepts. The usual electromagnetic field equations are expressed in terms of six quantities. Basic Equations. Introduction. these innovations arc extensions of a.c circuit concepts. In this system the unit of length is the meter. The fundamental concepts which form the basis of our study are presented in this chapter. The . the unit of mass is the kilogram. that is. 12. henceforth called ac (alternatingcurrent) phenomena. The mksc system of units is particularly convenient because the electrical units are identica. to our needs. We shall start with the familiar Maxwell equations and specialize them. New notation and nomenclature. The vector analysis concepts that we shall need are summarized in Appendix A. and the unit of charge is the coulomb. We shall view electromagnetic phenomena from the "macroscopic" standpoint. The rationalized mksc system of units is used throughout. For the moSt part. linear dimensions are large compared to atomic dimensions and charge magnitudes are large compared to atomic charges. No treatment of the mechanical forces associated with the electromagnetic field is given. This allows us to neglect the granular structure of matter and charge. a. will be introduced. These are &. the unit of time is the second. We assume all matter to be stationary with respect to the observer. called the electric charge density (coulombs per cubic meter) 1 x. It is assumed that the reader already has some acquaintance with electromagnetic field theory and with electric circuit theory.
» """ (II) q~ These equations include the inIormation contained in the equation of conJ. s"_ _. 1·1. 12. 17. We shall call a quantity wellbehaved wherever it is a continuous function and has continuous derivatives.d. In the equations of the first column.inuity a/ aq. d1 and ds on an open surface.FIo. as shown in Fig. Wherever the above quantities are weUbehaved. The circle on eo line integral denotes a closed con1.A1WONlC ELECl'ROMAGNl7I'1C FlELDS dl c Flo. they obey the Maxwell equations VX&= il<ll a:n vx:re+. introduced in Sec. we use the convention that ds points outward from a closed surface.0 v .2 TnIEB. In the equations of the last column. These are actually more general than Eqs. (11) because it is no longer required that the various Quantities be well·beh80ved. ds on a closed eurfaoe.!l at at v·m . (12) which expresses the conservation of charge. Corresponding to each of Eqs. since we wish to reserve the usual boldface roman letters for complex quantities. 11. (11) are the integral forms of Maxwell's equations 1!>CIl ds ~ 0 o 1!>:n ds o = ffl (13) q. The integrallorm 01 Eq.ourj the circle on a 5urface integral denotes a closed 5urface.0 the righthand rule of Fig. we employ the usual convention tha. (12) is (14) .t ell encircles ds according 1. 12. Note that we have used boldface script letters for the various vector quantities.
is positive reference at the start of the path of integration.. Charge is IIr "netamount" quantity. This is illustrated by Fig.. en~n~. This is the statement of conservation of charge as it applies to a region..PUNDA.ion for • Reference convention for .ld equations. I '\ J_ I \ \ . since all qunntities appearing in them are field quantities. Our convention for a usurfaecintcgrnl" quantity. Reference convent. 14. such 88 current. called the electric charge (coulombs) !/t. This is shown in Fig.. Fla. (11) to (14) jU. . The explicit relationships of the field quantities to the circuit quantities can be summarized as follows: v= f E·d1 iII g·ds q .Ilf q. or integral quantity.. (volts) i. We sh&11 use the name field quantity to describe the quantities discussed above. is positive reference in the direction of ds. ~_r.".J£EN'1'AL CONCEPTS 3 where the samc convention applies. called the magnetomotive force (amperes) .. Our convention for a fflineintegral" quantity. such as voltage. Associated with each field quantity there is a circuit quantity. FIG.ds +' . Corresponding equations written in tcrJ1l8 of circuit quantities we shall describe as ciTcuit equation&.S) AU the circuit quantities are algebraic quantities and require reference conditions when designating them..dr +~ffCB. Equa + 4l I \ • A. 13. We shall call Eqs. being the amount of positive charge minus the amount of nega· tive charge.If D·ds u=fx. I .cn (J. 14. voltap. These circuit quantities are " called tbe oolIa{}. called the electric current (amperes) q.""" . called the electric flux (coulombs) u. called the magnetic flux (webers) "". 13.
Finally. (17). They state that charge .4 TIMEHARMONIC ELECTROMAGNETIC FIELDS tiona (13) are commonly written in mixed field and circuit form as . the cir 1> 1>~~ = (18) ddt' +i (19) and the circuit form of Eq. we shall use the notation that :z denotes summation over a closed contour for a lineintegral quantity.ds = _ 'Jr dq dt (17) Finally.kt. (11). It states that a changing magnetic flux induces a voltage in a path surrounding it. and summation over a closed surface for a surfaceintegral quantity. While an extensive exposition of these concepts properly belongs in an introductory textbook. let us here summarize them. For this. (17) is Note that the first of Eqs.1J. extended to the timevarying case. It is essentially a partial definition of electric flux. An understanding of the concepts is an invaluable aid to remembering the equations. (16) are In this notation. and (19) are all forms of the law of conservation of charge. The first equation in each sct i. The fourth equation in each set is Ga'U88' law and states that lines of ~ begin and end on electric charge.d1 _ _ d>/ 'Y dt f d>/' :JC·d1+i dt (16) Similarly. Eqs." that is. It is a partial definition of magnetic intensity and magnetomotive force. and Eq. lines of <B can have no beginning or end. the equation of continuity in mixed field and circuit form is 4. essentially Faraday's law of induction. The third equation of each set states that magnetic flux haa no "flux source. It is apparent from the preceding summary that many mathematical forms can be used to present a single physical concept. (13). (19) is a generalized form of Kirchhoff's current law. and (18). (14). (12). Consider the sets 01 Eqs. the various equations can be written entirely in terms of circuit quantities. (1~8) is a generalized form of Kirchhoff's voltage law. cuit forms of Eqs. The second equation in each set is essentially Ampere's circuital law. (16).
Lines of current must begin and end at points of increasing or decreasing charge density. (112) that 8.FUNDAMENTAL CONCEPTS 5 can be neither created nor destroyed. Measurements have established that I c = .= 2.3C) 11 . fO ell (113) It then follows from Eq.. merely transported. In addition to the equations of Sec.99790 X 10' ~ v 3 X 10' meters per second (112) foJlo The choice of either fO or Jlo determines a system of electromagnetic units according to our equatioDs. 1~3. (Bl and n in terms of & and :re. By international agreement. The constitutive relationships assume the particularly simple for.s in free space (111) where fO is the capacitivity or permiUivity of vacuum.m. We shall consider the domain of E and ac as the electromagnetic field and express :D. Under certain conditions. We say that such matter is linear . Light is electromagnetic in nature. It is a mathematical conaequence of the field equations that (foJ10)~ is the velocity of propagation of an electromagnetic disturbance in free space. and this velocity is called the velocity of light c. 12 we need equations specifying the characteristics of the medium in which the field exists. the constitutive relationships become simple proportionalities for many materials.854 X laIt ~ 3~ X lOg farad per meter (114) for the mksc system of units. The term free space will be used to denote vacuum or any other medium having essentially the same characteristics as vacuum (such as air).3<:) (110) are called constitutive relationships.3<:) III ~ 1Il(8. the value of Jlo has been chosen as lJo = 4r X 107 henry per meter for the mksc system of units.11(8. Explicit forms for these can be found by experimentation or deduced from atomic considerations. Equations of the general form :D ~ :D(8. and 110 is the inductivity or permeability of vacuum. Constitutive Relationships. .
. :1)=. (115) hold "under certain conditions.o. called diamagnetic. = p/po is called the rekUive inductivity or rel4tive permeability and is.be order of 0.Po for them." They may not hold if & or 3C are very large.oUer for short. There is a class of materials. characterized by u = 00 I and to approximate good. The capacitivity f of any material is never less than that of vacuum to. We originally made the qualifying statement that Eqs. Materials having large values of 11 are called conduaora and those baving small values of u are called insul4toT3 or dUlulria. can be overcome by extending tbe definition of linearity. The dielectric constant of a good conductor is hard to measure but appears to be unity. The ratio p. has values of Jl much larger than p. the inductivity p.6 TIMEHARMONIC ELJ:crROMAGNJ::TIC PfELDS in the simple sense. as in the freespace case. in simple matter where.Jl~ lJ = uE Thus (115. or if time derivatives of & or:Ie are very large. Quite often the restriction on the time rate of change of the field. but t. A third class of materials. and call it limple m. dielectrics by perfed dielectrica. and p. There is a clagg of materials. For our purposes. For analyses. it is often convenient to approximate good conductors by perfect conductors. we shan call all materials except the ferromagnetic ones nonmagnetic and tnkc II . is approximately that of free space lAo.& } (B =.hese materials arc often nonlinear. called paramagmtic. when the constitutive relationships are the following lincar differential equations: in linear matter (116) Even more complicated formulas for the constitutive relationships may . The parameter u is called the conducliuity of the medium. E. and call it linear matter. of course. For most linear matter. Matter is often classified according to its values of tT. caUedferromagmtic. = tlto is called the dielutric constant or relative capacitivity.o (of t. characterized by rT = O. The ratio t. (115). We say that matter is linear in the general sense. f is called the capacitivity of the medium and p is called the inductivity of the medium. for which Jl is slightly greater than 110 (again of the order of 0. for which IA is slightly less than P. essentially unity for nonmagnetic matter. made on the validity of Eqs.01 per cent).01 per cent).
In dielectrics. The symbols i and k will be used to denote net electric and magnetic currents and the same superscripts will indicate the type. The physical significance of the extended definition of linearity is 88 follows. In view of the symmetry of Maxwell's equations.a:D at ol<II + . A similar picture holds for (B except that the magnetic moment of the electron is the contributing quantity. As discussed above we define total currents j j j j j .!l' . He visualized this displacement current in free space as a motion of bound charge in an Uether/' an ideal weightless Buid permeating all space. It holds for:D if the electron is a bound electron. The Generalized Current Concept. These are the currents we view as the cause of t. He amended the law to include an elutr* di. part of the term aD/at is a motion of the bound particles and is thus a current in the true sense of the word. conduction. Such a picture holds for 9 if the electron is a free electron. We have since discarded the concept of an et. 14. but Eqs. V X X . (116) are the most general that we shall consider. (1~16) contribute to an "admittivitylJ and an II impedivity >l of a material in the timeharmonic case.he direction of g changes. j .. it also is convenient to consider the term (}(B/at as a magnetic di8ploament current.ter have mass as well as charge. For example suppose an electron has been accelerated by the field. There will be a time lag before the electron can change direction because of its momentum. It will be shown in Sec.3plaament currem O'D/at in addition to the conduction current.!l' + . and then t. was incomplete for timevarying fields.. Note that Eqs. it is convenient to consider the entire O'D/at term as a current. to represent sources. at where the superscripts t. (117) m<' . Finally. We shall not attempt to give significance to each term of Eqs. We shall see in the next section that the impressed currents represent energy sources. we amend the field equations to include impre88ed currents. 19 that all terms of Eqs. :1. and i denote total. The symbols .her for it bas proved undetectable and even somewhat illogical in view of the theory of relativity.l with superscripts indicating the type of current.g and ml will be used to denote electric and ma. electric and magnetic. and impressed currents. SO when the field changes rapidly t.+m<.FUNDAMENTAL CONCEPTS 7 be necessary in some cases. However. (115) when the time derivatives of & and 3C become sufficiently small. c.gnetic currents in genera. (116). The atomic particles of mat.he particles cannot U follow" the field. It was Maxwell who first noted that AmpAre's law for statics.he field. (116) reduce to Eqs.!l.
Actually. In terms of the generalized current concept. However. the totol currents are solenoidal. (120) to (123) are well suited to our purposes. (l17) is i.8 TIJlE. (121) applied to (125) 1P OIl' . PSdl.=d¥.ds  0 Thus. the mixed fietd~ircuit p:redl ~ II a'ds (121) form is (122) PSdl~k' and the circuit form is (123) Note that these look simpler than tbe equations of Sec. the circuit (orm corresponding to Eqs. For example.OKAONETlC nELDS Thus.g' (120) and in the integral form. related to the a and ml by iII ads v X & = _:nt' k  II mHs (119) where these apply to any of the various types of current.H. we obtain VOIl'O VlJ'O (124) for V· V X a "'" 0 is an identity. . of course. We have done so to emphasize that this information is included in the above sets. Lines of total current have no beginning or end but must be continuous. in the differential (orm.+i. we have merely included many concepts in the functions:nt' and a'j so tome of the information contained in the original Maxwell equations bas become hidden. Note that we have omitted the H divergence equations" of Maxwell from our above sets of equations. closed surfaces became Similarly. the basic equations of electromagnetism become. 12. taking the divergence of each of Eqs. VX:JC=. Eqs.II ""'ds Also. and the forms of Eqs. our study comprises only a small portion of the general theory of electromagnetism..A1WONIC ELEcrB.+i' dl (118) k'=#+k' dl The i and k &re. (120).
namely.FUNDAliEl\I"TAL CONCEPT8 9 As an illustration oC the generalized turrent concept.=ea. Figures 15 and 1. In Fig. the Hvoltage source" mz. The field oheys the Maxwell . current.. Types of magnetic current. the impressed currents are those currents we view as sources. large SlI' FlO.gnetic core. as suggested Flo.oiDg hy Fig.. consider the circuils 01 Figs.5ll' i . Consider a region oC electromagnetic field. A region contai. 16. Our understanding of the concept will grow as we learn to use it. the Ucurrent source" I)' produces a conduction current I)c through the resistor and a displace ment current a'  iYD/dl through ~urce l_1P_ _j~ l the capa. a doubleheaded arrow represents a magnetic current. Type. In these pictures we have used the convention that a singleheaded arrow represents an electric current. In another prof>. 15.6 anticipate one application. a polarization or magnetization current might be considered as the source current. which in generalized current . 1.C produces an Fro. or impressed. 17. electric current in the wire which in turn causes the magnetic displacement current 9n" "'" iJ(B/iJt in the ma.citor.. In Fig. 17. l5 and 16.a of electrio current. Source <. Energy and Power. In one problem. that of representing sources.. lem. equations.. In a sense. 15. It is not possible at this time to give the reader a complete picture oC the usefulness of impressed currents. More generally. 16.6. a conduction current might be considered as the source. the impressed currents are those currents in terms of which the field is expressed. p.
 (131) where a' and mt' represent possible source currents.Ji't6 1 J.!.. it e&n be shown that a product t· tJ is a power density. (115). = v . This suggests a scalar multiplication of the second of Eqs. (126) can then be interpreted as the rate of increase in energy density at a point. IlXt ) + :JC • mti at 2 The terms W.LECTROKAGNETlC FIELDS notation are Eqs. The point relationship (129) P. as defined by Eqs. The Poynting vector s=&XX (128) is postulated to be a densityofpowcr flux.E . Also. Eq. S = V . (120) by s. (120). the other terms of Eq. there results (127) We shall interpret these as equations for the conun:ation of eneTfl1I.g' + :Ie. and the integral cP. (126) being the differential form and Eq. (! . V X :Ie a Bcalar multiplication of the first of Eqs.E') + .g' 3C • ml' = . The other terms of Eq. The generally accepted interpretation of Eqs. in view of the vector identity v . For media linear in the simple sense. (127) being the integral form.s the fate of increase in energy within the region. w. Further identification of this energy can be made in particular cases. (127) can be interpreted n. As an extension of circuit concepts. (E X :Ie) is then a volume density of power leaving the point. and the divergence theorem applied to the first term. V X E .10 TIKE!lARMONIC J:. difference of the l'C8ulting two equations is The (126) v· (E X :Ie) + E·. (120) by ac is suggested. "'" . the last two terms of Eq. (E X :Ie) = :Ie .i'1J3C1 (132) . = 1PSods = 1/>& X :JCods (130) is the total power leaving the region bounded by the surface of integration. (126) and (127) is as follows.&' + E 'Il' at 2 ! (. .0 If this equation is integrated throughout a region. (126) become E· .'!. Similarly.
FUND.ulENTAL CONCEPTS
11
are identified as the electric a.nd magnetic energy densities of static fields, and this interpretation is retained for dynamic fields. The term
(133)
is identified as the density of power converted to heat energy, called diuipaled power. Fin&1ly, tbe density of power supplied by the source currents is defined as
(134)
The reference direction for /SOurce power is opposite to that for dissipated power, as evidenced by the minus sign of Eq. (134). In terms of the abovedefined quantities) we can rewrite Eq. (126) as
p.  P,
a + p, + at (w. + w.)
(135)
A word statement of this equation is: At any point, the density of power . supplied by the sources must equal that leaving the point plus that dissipated plus the rate of increase in stored electric and magnetic energy densities. A more common statement of the conservation of energy is that which refers to an entire region. Corresponding to the densities of Eqs. (132), we define the net electric and magnetic energies within a region as
(136)
Corresponding to Eq. (133), we define the net power converted to heat energy as
ll', =
JJJ.&' d,
(137)
Finally, corresponding to Eq. (134), we define the net power supplied by sources within the region llS
(138)
In terms of these definitions, Eq. (127) can be written as
ll'. = ll',
d + ll', + cit (W. + w.)
(139)
Thus, the power supplied by the sources within 8 region must equal that leaving the region plus that dissipated within the region plus the rate of incrcuse in electric and magnetic energies stored within the region. If we proceed to the general definition of linearity, Eqs. (l16), the sepa.ration of power into a reversible energy change (stor!"ge) and ao
12
TIHEHARMONIC ELECTROMAGNETIC FIELDS
irreversible energy change (dissipation) is no longer easy. Contributions to energy storage and to energy dissipation may originate from both conduction and displacement currents. However, Eqs. (135) and (139) still apply to media linear in the general sense. We merely cannot identify the various terms. In See. 1·10 we shall see that for ae fields the division of energy into stored and dissipated components again assumes a simple form. 16. Circuit Concepts. The usual equations of circuit theory are specializations of the field equations. Our knowledge of circuit concepts can therefore be of help to us in understanding field concepts. In this section we shall quickly review this relationship of circuits to fields. Kirchhoff's current law for circuits is an application of the equa.tion of conservation of charge to surfaces enclosing wire junctions. To demonstrate, consider the parallel RLC circuit of Fig. 18. Let the letter 0 denote the junction, and the letters a, b, c, d denote the upper terminals of the elements. We apply Eq. (17) to a surface enclosing the junction, as represented by the dotted line in Fig. 18. The result is
~OG
. +. +' +. +,+dq = ~<>lI ho I."" ~J dt
0
where the i... are the currents in the wires, i l is the leakage current crossing the surface outside of the wires, and q is the charge on the junction. The term dqjdt can be thought of 88 the current through the stray capaci~ tance between the top and bottom junctions. In most circuit applications both if and dq/dt are negligible, and the above equation reduces to
i ... +i<>ll+ioc+i",,=O
This is the usual expression of the Kirchhoff current law for the circuit of Fig. 18. Kirchhoff's voltage law for circuits is an application of the first Max. well equation to closed contours following the connecting wires of the circuit and closing across the terminals of the elements. To demonstrate, consider the series RLC circuit of Fig. 19. Let the letters a to h. denote

.
R
 " S
,
FIo. 1.8. A parallel RLC circuit.
,
c
a)
L
FUND~TAL
CONCEPTS
13
b
a
FlO. 19. A Bel'ies RLC
 c
d
I
R
I
circuit.
I
I I
I
I
h
L
•
B
I
c
f
the t.erminnls of the elements as shown. We apply the first of Eqs. (16) to the contour abcdefgha, following the dotted lines between terminals. TJ;Us gives
~+~+v~+~+v~+~,+~+~+~O
dy,
where the are tbe voltage drops along the contour and !/I is the magnetic flux enclosed. The voltages v.., v.... v~, and Vfl are due to tbe resistance of tbe wire. The term dl/I/dl is tbe voltage of the stray inductance of the loop. When the wire resistance and the stray inductance can be neglected. the above equation reduces to
v._
This is the usual form of Kirchhoff's voltage law for the circuit of Fig. 19. In addition to I{irchhoff's laws. circuit theory uses a number of l'elcment laws." Ohm's law for resistors. v  Ri, is a specialization of the constitutive relationship .11 = u6. The law for capacitors, q  Cv. expresses the same concept as ~  t6. We have from the equation of continuity i .. dq/dt. so the capa.citor lnw ellD also be written as i .,. C dv/dt. The law for inductors, '"  Li. expresses the same concept as m  ",3<:. From the first Maxwell equation we have v  d"'/dt. 80 the inductor law can also be written 8B v  L dijd,t. Finally. the various energy relationships for circuit theory can be considered as specializations of those for field theory. Detailed expositions of the various specializations mentioned above can be found in elementary textbooks. Table 11 summarizes the various correspondences between field concepts and circuit concepts. 1·7. Complex Quantities. When the fields are ae, that is. when the time variation is harmonie, the mathematical analysis can be simplified
14
TIMEHARMONIC ELECTROMAGNETIC FIELDS
TABLil 11. CoRRESPONDENCES BETWEEN
CmcUlT CoNCIlPT8 AND FuLO CoNC.E1'1'8
Circuit concepts Voltage
II
Field concepts
Eleetric intensity 8
Electric current. deMity or magnetic intensity Ie
Current i
a
Magnet.ic Bux ""
Cb&rgc q
Magnetic ftux density i13
Charge density q. or electric Awe density n
Kirchhoff's voltage lAw (generalized) MaxwellFaraday equiltion ilGl 4
Kirchhoff's current law (generalized) Equation of continuity
dq dl
L·Lt·•
dt
VX&
at
v·$Jc.a,
Coo!5Litutive relationships (linear in the simple
..noe)
aq.
Element laws (linear) 1 Resistors i   D
R Capacitors q  Gil
.,
Conductors.9<  crE
Dielectrics :D<I;
O<
d. iCdl
jJ"e
.s
a'
(B 
Inductors '"  Li
0<
Magnetic properties
~:JC
vLdt
di
0<
~"JJ.
ax at
Power flow PI  vi
Power dissipation in resistors 1 \Pol _ til. _ _ pt
Power Bow B  & X :JC
Power d.issipo.Lion Pol  & '.g'  erst
R
Energy in capacitors
'W. 
Electric energy
w.~:D·6
liqo  He.,

~«:I
<
Energy in induetore
'W. 
"Magnetic energy
w.. ~~m':fC ~1o'3C1
li..  liLi'
FUNDAMENTAL CONCEPl"8
15
by using complex quantities.
The basis for this is Euler's identity
%E
ei·
cosa+jsina
where j = v'=1. This gives us a relationship between real sinusoid&1 functions and the complex exponential function. Any ae quantity can be represented by a complex quantity. A scalar quantity is interpreted according to l
v =
v'21V1 co, (wt + a)
where" is called the instantaneoulJ is called the complex quantity. The notation Re ( ) stands for lithe real part of/' that is, the part not associated with j. Other names for V are "phaaor quantity" and flvector quantity," the laat name causing confusion with space vectors. In our notation v represents a voltage, hence V is a complex lJol14ge. Equation (140) with IJ replaced by i and V replaced by I would define a complex current, and so on. Note that the complex quantity is not a fUDction of time but it may be a function of position. Note also that the magnitude of the complex quantity is the effective (rootmeansquare) value of the instantaneous quantity. We have chosen it so because (1) a.c quantities are usually specified or measured in effective values in practice, and (2) equations for complex power and energy retain the same proportionality factors as do their instantaneous counterparts. For example, in circuit theory the instantaneous power is p  vi, and complex power is P = V I·. A factor of ~ appears in the equation for complex power if peak. values of v and i are used for IVI and IJI. Complex notation can readily be extended to vectors having sinusoidal time variation. A complex E is defined as related to fln instantaneoui t according to
v'2 Re (Vel~) quantity and V "'"' lYle""

(140)
s
~
v'2 Re (Ee'·')
(141)
This means that the spatial components of E are related to the spatial components of t by Eq. (140). For example, the z components of E and t arc related by
S. =
v'2 Re (E.d·')

v'2 IE.I co, (wi + a.)
where E", = IE",lei"'.. Similar equations relate the y and: components of E and t. The phase of each component may be different from the phases of the other two components, that is, a"" a., and a", are not necessarily equal. In our notation & is an electric intensity, hence E is called the compia eled.rU; intetuity. Equation (141) with E replaced by H and 8 by X
I The convent.ion " _ "the imaginary part. or." peak value of II.
~
1m (Vel.') can &lao be used, where 1m ( ) atancla fOT The factar v"2 can be omitted if it is deeircd that IVI be Lb.
16
'I'IJrOioItAJUlONlC ELECTROMAGNETIC nELDS
defines a comple:e magnetic intenrily H, representing the instantaneous magnetic intensity X, and 80 aD. Note tha.t the magnitude of a component of the complex vector is the effective value of the corresponding component of the instantaneous vector. This choice corresponds to that taken for complex scalars and has essentially the same advantages. A real v~tor, sucb as S or :re, can be thought of as a triplet of real scalar (unctions, ne.mely, the %, Y, and t components. At any instant of time, the vector has a definite magnitude and direction at every point in space and can be represented in three dimensions by arrows. A complex vector, such as E or H, is a group of six real scalar functions, namely, the real and imaginary parts of the X, 1/t and z components. It cannot be represented by arrows in threedimensional space except in special cases. One such s~ia1 case is that for which 0::", = a.  a., 80 that the vector has a real direction in space. In this case the instantaneous vector always points in the same direction (or opposite direction), at a point in spa.ee, changing only in amplitude. We could define a ucomplex magnitude" and a "complex direction" for lJ, complex vector as extensions of the corresponding definitions for real vectors, but these would ha.ve little use. Throughout this book we shall use the following notation. Instantaneous quantities are denoted by script letters or lowere&se letters. Complex quantities which represent the instantaneous quantities are denoted by the corresponding capital letter. Vectors are denoted by boldface type. 18. Complex Equations. The symbol He ~ ) can be considered as a mathematical operator which selects the real part of a complex quantity. A set of rules for manipulating the operator Re ~ ) can be formulated from the properties of complex functions. The following are the rules we shall need. Let a capital letter denote a complex quantity and a lower~ase letter denote a real quantity. Then
Re (A)
+ Re (B)

Re (A
0
+ B)
(142)
Re (oA) =
Re (A)
f Re (A) dz 
:x
Re (A) 
Re(~~) Re (f Adz)
The proof of these is left to the reader. In addition to the above equations we shall need the following lemma. If A and B are complex quanlitiu, and He (AeJ ' )  Re (Bei"'l) for aU I, then A = B. We can readily show this by first taking t = 0, obtaining Re (A) ~ Re (B), and then taking wt  T/2, obtaining 1m (A) = 1m (B). Thus, A  B, for the above two equalities are the definition of this. To illustrate the derivation of an equation for complex quantities from
FUNDAMENTAL CONCEPTS
17
one for instantaneous quantities, consider
Expressing v and £ in terms of their complex counterparts, we have
V2 Re (v"·') 
f V2 Re (E';·') . dl
By steps justifiable by Eqs. (142), this reduces to
Cancella.tion of the "\I'2's and application of the above lemma then gives
Note tha.t this is of the same form as the original instantaneous equation. We have illustrated the procedure with a scalar equation, but the same steps apply to the components of a vector equation. From our rules for manipulation of the Re ( ) operator, it should be apparent that any equation linearly relating instantaneous quantities and not involving time differentiation takes the same fonn for complex quantities. Thus, the complex circuit quantities V, I, U, and K are related to the complex field quantities E, H, J, and M according to
VfE.dl IffJ·d'
U=fH'dl K=ffM'd'
(143)
There is no time differentiation explicit in the field equations written in generalized current notation. The complex forms of these must therefore also be the same as the instantaneous forms. For example, the complex form of Eqs. (120) is
v X H ... JI
(144)
Even though these complex equations look the Bame as the corresponding instantaneous equations, we should always keep in mind the difference in meaning. As an illustration of the procedure when the instantaneous equation exhibits a time differentiation, consider the equation
VX£=aT
alB
Again we express the instantaneous quantities in terms of the complex
18
TUlE1lARJ40NIC ELECTROMAGNETlC FIELDS
quantities, and obtain
The time variation is explicit, Bnd the differentiation can be performed. By steps justifiable by Eqs. (142), the above equation becomes
y'2 Re (V X Eel") =  y'2 Re (jwBeI"')
By the foregoing lemma, this reduces to
V X E  jwB
It should now be apparent that each time derivative in a linear instantaneous equation is replaced by ajc.J multiplier in the corresponding complex equation. For example, the Maxwell equations in complex (orm corresponding to Eqs. (11) are
V X E jwB
vXH~iwD+J
v·B  0 v· D  Q.
(145)
The other forms of these can be obtained in a similar fashion. 19. Complex Constitutive Parameters. The constitutive relationships for matter linear in the general sense can be specialized to the ae case by the procedure of the preceding section. To illustrate, consider the first of Eqs. (116), which is
:O_(f+fl'!+f,a + ...)& at at'
t
The complex (orm of this equation is readily found as
D = (f
+;Wft 
Wlf:
+ ..
')E
The quantity (t + jWEl  Wlft which we shall denote by i(w).
+ ...) is just
D  .(w)E
8 compler. (unction of w, Thus, the complex equation
which looks like the form for simple media, is actually valid for media linear in the general sense. The other two of Eqs. (116) simplify in a similar manner; so we have the ac comlitulivt relatiomhip!
D  '(w)E B ~ p(w)H J'  '(w)E
(146)
for linear media. We call l the complex permiUiuily of the medium, P the compkx permeability of the medium, and 4 the complex tXlnductivity
FUNDAMENTAL CONCEPTS
19
of the medium. Remember that these parameters are not necessarily the de parameters, but
I:(w), A(w), 8'(w)
_.
+ E, lA,
CT
The de parameters may apply over a wide range of frequencies for some materials but never over all frequencies (vacuum excepted). In terms of tbe generalized current concept, the induced currents (caused by the field) are
(8 jw.)E  g(w)E M  jwpJI  f(w)H
J
+
(147)
The parameter ti(w) bas the dimensions of admittance per length and will be called the admiUivity of the medium. The parameter ~(w) has the dimensions of impedance per length and will be called the impedivity of the medium. Note that fj is a combination of the 8' and ~ parameters. A measurement of fj is relatively simple, but it is difficult to separate 4 from i. The distinction is primarily philosophical. If the current is due to free charge, we include its effect in a. If the current is due to bound charge, we include its effect in i. Thus, when talking of conductors, the usual convention is to let fj = 6' + jWf.o. When discussing dielectrics, it is common to let '0 = jWf. To represent sources, impressed currents are added to the induced currents of Eqs. (147). Thus, the general form of the ae field equations
IS
 V X E
~
f(w)H
V X H ~ g(w)E
+ M' + J'
(148)
The z(w) and 'O(w) specify the characteristics of the media. The J' and Mi represent the sources. Equations (148) are therefore two equations for determining the complex field E, H. Solutions to these equations are the principal topic of this book. 110. Complex Power. In Sec. 15 we considered expressions for instantaneous power and energy in terms of the instantaneous field vectors. We shall show now that similar expressions in terms of the complex field vectors represent timeaverage power and energy in a.c fields. For this, we shall need the concept of complex conjugate quantities, denoted by·, and defined as follows. If A  a' + jalf = IAlei", the conjugate of A is A· "'" a'  jalf = jAlcJ... It follows from this that AA' ~ IAI', Let us first consider any two ac quantities a. a.nd CB, which may be scalars or components of vectors. These are in general of the form
a  01A I cos (wt + a)
(II 
01BI cos (wI
0 Re (A,"') + P)  0 Re (Be i.,)

 BI/:JC.P) + cos (2".H· .  B.20 TlMJ)oIIABJ. + H' .:JC.r over that quantity.IAIe'. M' The lefthand term is ..) It there This is a sum of terms.t + P) .) + u.Xy) + uvCB. (149).os (a .P) 80 + i sin ~a .(&. V X E "'" E • JI. we 8calarly multiply the first by H· and the conjugate of the second by E. .:JC..IAIIBI(.0 r" The integral form of this is obtained by integrating throughout 8. region .:JC.t + a) 01BI cos (. + H· . We can obtain an equation in which S appears by operating on the complex field equations in a manner similar to that used in the instantaneous case. The time average of the above expression i. fore foUows that s= & X "" = Re (E X H') In view of this we define a complex Poynting veet4r S = E X U· (151) whose real part is the time average of the instantaneous Poynting vector. (E X H·) by a mathematical identity.p») (150) it is evident that Cl<ll = Re (AB') This identity forms the basis of definitions of complex power. The product of two such quantities is (149) 01AI cos (". (E X H') + E . The instantaneous Poynting vector [Eq. V X H· .. The difference of the resulting two equations is E . so we have (153) v .e Cl<ll  IAIIBI cos (a . Starting from Eqs. each of which is the form of Eq.B. or SRe(S) (I52) We shall interpret the imaginary part of Slater..(8.t + a + P») We shall denote the time average of a quantity by a ba. (144). M' .V .3C.P) We also note that AB' . (l28)} can be expanded in rectangular coordinates as S = u.[ONlC ELECTROMAGNETIC nELDS where A .and B Cl<ll  ~ IBI.IAllBllcos (a .
pqwer.jWIIIHI' where lEi' means E . (140) and (141)).. (153). and 19 are not sinusoidal quantities but are Cormed of products of sinusoidal quantities.fH = jWIIH so E ..(. or Re (fJJ) . As suggested by Eqs.1ft E X H'· ds It is evident from Eqs. (130) and (152) that the real part of this is the timeaverage power flow.  (159) V •S + P' + j2". (129). M' . (153) as 'fJ. This results in 0 (I54) 1ft E X H' .) . ds + fff (E· J" + H' . we have ~'IEI' ~"IHI' We can now write Eq.j""IEI' H.(w. (I53) and (154) expressions for the conurvation of complex.V .. (132) and (133). or (I58) Note that these relntionsrnps are quite different Crom those used to inter. pret most complex: quantities IEqs.. We shan can Eqs.IEI' } in simple media w. (126) and (127). we define a compla volume cWuity of power kaui1U} a point as iiI . S = V • (E X H') (155) The real part of this is a timeaverage volume density of power leaving a point. w.  .0 (160) . (129) and (152). H*. This is because 5. P. + j". let us first specialize to the ease oC a sourceCree field in media linear in the simple sense.FUNDAKENTAL CONC&P1'8 21 and applying the divergence theorem.  Ill. The various terms of the above equations are interpreted as foUo\\1J. J" = . In terms of the instantaneous energy and power definit. M') dT  Comp.and IHI' means H . .PI (156) where PI is defined by Eq.IEI' . To interpret the other terms of Eq. We then have J' = OE . po'IOeT kauing a region as Similarly.e these with Eqs..ions of Eqs. we define the complex (157) PI = 1ft S· ds .)E Mf . E. the former applying at a point and the latter applying to an entire region..
in conformity with circuit theory nomenclature. We can. is called reactive power. however. M') dr He (P. and. Note that this interpretation of complex power is precisely that chosen in circuit theory. IC sources arc prescnt. the imaginary part of 'PI as defined by Eq. (154) to sourcefree simple media is therefore 1ft S· ds + <1'. Note that we have never defined <Pd. for media linear in the general sense. continue to use Eq. or He (ft. (134). !!! (E . The integral relationships corresponding to Eqs. P. OW. .) .) = <1'. J" + H' .. (168) for the . (153) in general as w.PI (164) We can write Eq.0 (162) corresponding to the point relntionsbip of Eq. (I55) is 2w times the difference between the timeaverage electric and magnetic energy densities. (138). complex power density supplied by tM sources ca. 'W. arc defined by Eqs.22 TIMEHARMONIC ELECTROMAGNETIC FIELDS Thus.n be defined as (163) ft. or '\\'. . (160).. . where p. (166) where. The imaginary part is related to timeaverage energies. is defined by Eq. a.. . it is evident that (167) Then the form of Eq...(E· J" + H'· M') The real part of this is the timeaverage power density supplied by the sources.  where all terms have been identified for simple media. + j2w('W. from Eq.) (165) + P' + j2w(w. (159) arc in simple media (161) where C9d.) (168) The real part of this represents a timeaverage power balance. + j2w('W. (165) applicable to an entire region is P. (136) and (137).) . the total complex power 8upplied by sources within a region can be defined as P. The specialization of Eq. and '9.'11'.  'W. .P.. Similarly.PI + <1'.
In metals. !I:le_. ! and y assume their simplest forms. The timeavernge electric and magnetic energies are defined in general as (170) which reduce to the last two of Eqs. The first term of the integrand represents both conduction and dielect·ric losses. In soureefree regions. More discussion of this concept is given in the next section.V X E . (170) includes kinetic energy stored by free charges as well as the usual field and polarization energies. Thus. The permittivity of metals is hard to measure but appeara to be approximately that of vacuum. AC Characteristics of Matter. Von Hipple. We shall consider this case later_ In good dielectrics. the conductivity remains very close to the dc value for all radio frequencies. 1954. The timeaverage power dissipntion is defined in general as fJ'. up to the infrared frequency spectrum. ~~:~ : jw~ jWfO } in nonmagnetic metals (172) In ferromagnetic metals. 111. it is common practice to neglect" and express 9 entirely in terms of l. Thus. This is done as follows. being O(w) = jWfo lew) = jWJAO 1 in free space (171) These hold for all frequencies and all ficld intensities. . New York.ti(w)E In free space. (161) in simple media.jw.l(w)H V X H . JAo would be replaced by fl. l l We can express l in both ree A. lew) = jWJAo I in nonmagnetic dielectrics (173) Let us now consider l(w) in more detai1. that is. The first of Eqs. ti(w) . (161) in simple media." John Wiley & BoWl. the complex field equatiolU5 read .FUNDAMENTAL CONCEPTS 23 general case of linear media by extending our definitions. "Dielectric MaterWa and ApplicatioWl. and tbe second term represents magnetic losses. = Re [ffl (tilEI' + flHI') dT] (169) which reduces to the first of Eqs.
A U good dielectric II is defined to be one for which l remains almost constant at all radio frequencies and for which . and /. and wt" contributes to power dissipation (acts like u in simple matter). 113 we shall see that they are related to the capacitance. are real quantities.jt" = llieil (174) where I." characterized by a varying land a large l' in the radiofrequency range. ~~ (175) Thus. The only perfect dielectric is vacuum. paraffin.0012 ~ / 0. respectively... ~. A <I perfect dielectric" would be one for which I' = O. Measured values of ~(w) are usually expressed in terms of I and tan 0. porcelain. Examples of good dielectrics are polystyrene.24 TIMEHARMONIC ELECTROMAGNETIC FIELDS tangular and polar form as ~(w) = l . 110. We shall use the latter representation. We call l the ac capacitiuity. I'. and 0 the dielectric loss angle. There is no group of dielcctrics which havc unusually high dielectric constants.. . and Bakelite.1) "'" f'  if" frequency for polystyrene at 25"C. The titanate and ferrite ceramics fall into this 3 "/f<J 2 ~ 0. There is also a group of Illossy dielectrics.0008 0. cycles per sec verllus . or in terms of t' and (". In Sec. . resistance.. and loss angle.(1.' and E" versus frequency for Plexiglas to illustrate the characteristics of a lossy dielectric. Examples of lossy dielectrics are Plexiglas. Figure 111 shows . In terms of power and energy. of an ideal circuit capacitor. Figure 110 shows E' and . (169) and (170) that W. I contributes to stored energy (acts like t in simple matter)./' versus frequency for polystyrene to illustrate the characteristics of a good dielectric. and Teflon. l' the dielectric loss /Mtor.  1/ FIG. Iff <'IEI'd.0:.0004 1 o o '''/eo 10 102 103 104 10S!Q6 107 lQ8 10' 10 10 Frequency.~ IfI w<"IEI'd. we have from Eqs." is very small.
15 3 K \ . Thus. in both rectangular and polar form as pew) ~ "' . Thus. both conduction and dielectric losses may be significant. magnetic losses become important. and loss angle. A table of i for some common dieleckics is given in Appendix B. when it can be considered linear../ . cycles per sec .je" versus frequency for PlcxiglM at 25°0. respectively. and 6.. 0. In terms of power and energy.05 " o o Frequency..·· (177) where /Jo'. In Sec. and 6..(w) can be treated in a manner analogous to the treatment of lew). ~u+jw'l f. of an ideal circuit inductor.. but we shall view i as simply a measured parameter. We call p' the ac inductivity.  10 102 103 104 105 106 107 loa 109 10 10 FlO. are real quantities. Such dielectrics are usually lossy. ~ 0. 113 we shall see that they are related to the inductance. In addition to these. :C. p" the magnetic loS8 factoT. ~ <P.10 f..FUNDAMENTAL CONCEPTS 25 0.j"" . = jwp. we express jJ.Iple. resistance. we have from Eqs. (169) and {170) that \1>.  H "'!HI' ff fff W""IHI' dr (178) dr . ~". III.20 4 \ \ 0. In ferromagnetic matter.2 1 ("/'0 . in ferromagnetic matter (176) The parameter p.e' . A qualitative explanation of the behavior of i can be made in terms of atomic concepts. the magnetic loss angle.. class (the latter also being ferromagnetic).) .
January. However. to which must be added the conduction and dielectric losscs for the total power ~ipation. in which case f! assumes t. 0 FIG.ized by n dc magnet. at.he form of an asymmetrical tensor. Originally. Magnetized ferrites can be used to build II nonreciprocal" devices. Ferromagnetic metals are extremely lossy materials (primarily due to u). Thus. These materials become even more usefu1 when magnet.ic field. but ilie concept of displacement current has been retained. no. 1952. lQ2 lQJ 10" la' lQ6 Freqllency.. to illustrate the characteristics of ferrite ceramics. 31.26 28 TUrnHARM:ONIC ELECTRO)(AON"ETIC FIELDS 24 l. A Discussion of Current. L. which was visualized as the displacement of bound charge in matter and in an "ether. We shall use the latter representation." The existence of an ether has been disproved. Microwave Frequencies. cycles per sec 10. or in terms of 1£' and p.'. . the ferromagnetic ceramics CaD be profitably used at radio frequencies to obtain high values of p. and also quite nonlinear with respect to fl. in that they also have appreciable 1J". . such as U isolators" and <l circulators. Measured values of p(<4) are usually expressed in terms of p.12. They are seldom intentionally used at radio frequencies. They are lossy in the magnetic sense. "1 112. 1. Hogan.'  hi' vel'8Wl frequency for Fcrramte A at ZSC. Figure 112 shows /J' and /J" versus frequency for Ferramic A. t vol. where the above lJ'" is only the timeaverage magnetic power loss. J. 1. t C.". the term current meant the flow of free charges in conductors.~ flO • 16 c / I • • f 12 8 : I "I" I 4 \ o .~' contributes to stored energy and Jl" to power dissipation. Bell Sf/Item Tteh.'/1'0 . /. '< k }()8 1()9 10 . This concept was extended to include displacement current.. The Ferromagnetic Effect.{.' and tan ~.'(w) . The concept of current has broadened considerably since its inception."...
ble 12. called the reactive current. For our purposes. we have a. but it is usually small. In ma.tter. component in phase with E. In absence oC matter. Cor this discussion. motion of charge. the current is. In Cree space there is no motion of charges at aU. A further genera.lly a generalization of the circuit concept of current. All the currents mentioned are classificd as indueed currents.) Even in dielectrics there is some conduction current. real quantity. In this sense. Thus. both electric and magnetic. due entirely to the motion oC free electrons. When matter is present. J = jwe'E. and we have only a freespace di8placement current. This involves viewing the current in terms of a. magnetic freespace diaplacement current. This is usually true at radio frequencies. This is essentia.FUNDAMENTAL CONCEPTS 27 even though it is not entirely a. have been introduced to represent sources. it is called the displacement current. that is. many different phenomena are included. component out of phase with E.nother division of the electric current is convenient. We shall summarize the notation and concepts in complex form in this section. still a. Both the nomenclature and the concepts of complex magnetic currents Qre similar to those for electric currents. they are independent oC the field and are said to cause the field. giving rise to an induced magnetic current in addition to the freespace displacement current. The nomenclature used for electric currents is summarized in the first column of Ta. for all practical purposes. given by J = jWEoE. Such current is called the conduction current and is expressed mathematically by J = aE. J = (a + w~If)E. Because the term J = jw~E is of the same mathematical form as the freespace displacement current. (We shall consider u = a.nalogous to the electric case.lization was made to include magnetic displacement current as a. We call this the magnetic polarization . are caused by tbe field.. and the nomenclature used is somewhat lengthy. Consider the complex electric current density. Finally. Internal to conductors. called the di8sipative current. a. The one essential difference in the two concepts is the nonexistence of magnetic II charges II in nature. Because of the breadth of the concept of current. we have a. impressed currents. in addition to the conduction current and the freespace displacement current. we have magnetic effects due to the motion of the atomic particles. Impre88ed currents arc used to represent sources or known quantities. there is no free magnetic charge and no magnetic conduction current.EO)E. The total electric eurrent is the sum of the induced currents plus the impressed currents. This is called the polarization current and is expressed mathematically by J = jw(~ . current due to the motion oC bound charges. where the dissipative current produces the power loss and the reactive current gives rise to the stored energy. . and a. M = jWlloH. a. dual" concept of the electric displacement current.
..pt)H . We find it convenient to divide the magnetic current into a component in phase with H...p"B .101.wi1J..(.enl current. To represent sources or known quantities.(p .. This is defined as magnitude of reactive current density Q ... nnd the reactive magnetic current contributes to the stored energy..a}E + W.. expressed by M = jw(p.y Conduction Freeepace dillplacement.~ Po1Ariul.E jWttE i.)H.. a..p" gE .j...<E Disaipative Reactive Induced (.JoI. M .(0' + .tion Displacement.Aaa1J'ICATlOS OJ' ELunuc AND MAGNZTlC CtmuN'T11 Type Complex electric cunent density Complex magnetic current.jiH . The dissipative magnetic current contributes to the power loss. the induced magnetic current is simply the freespace displacement current. densit.)E J' ~E + . The nomenclature for magnetic currents is summarized in the second column of Table 12.p. The term M = jwpH is called the magnetic dilplacem..ft)E .!' + j . .(l . that is. and the polarization current. "... and a component out of phase with HI called the magnetic reactive current. caused by the field. called the magnetic di38ipative current. + """)E ...')8 + MI Impressed Total M' J'  + Ji M' _ fa CUrTent. A convenient classification of matter from the elcctric current standpoint can be made in terInB of a quality factor Q..:.:+i":~=" In nonmagnetic matter.'H JIB .. In nonmagnetic matter. this involvcs a ratio of stored electric energy to ~' (179) . ..iH .wp. being the sum of the (reespaec displacement current.. All the aforementioned magnetic currents arc indtu:ed currents. CI.magnitude of dlSSipative current density :IIC . j. we use impTC88ed currents.j"'lloH. M ..28 TnlEIlARMONIC ELECTROllAONETIC nILD6 TAJlLIl 12.(..w. ."H.. reactive current.. M jWIA'H.
According to the concepts of Sec. . the concept of Q in nonmagnetic matter can be considered as an extension of the concept of Q for capacitors in circuit theory. 2. given in Sec. an element is calJed an impedor. the Q defined above would be called the electric Q.) (181) In terms of circuit concepts.W. or wavelength. aa we shtJl aee in Chap. The complex notation used for ac fields is the extension of the complex notation used for a<l circuits.FUNDAMENTAL CONCEPTS 29 power dissipated. there is an additional power dissipation due to magnetic hysteresis loss. In general. When P is primarily real.IVI'Y' where Z and Yare the impedance and admittance of the element. since it includes only the power loss due to el~ctric effects. The complex field equations bear a relationship to the complex circuit equations which is similar to that for the timevarying case.(a _ W _ + ". (180) must be modified. the complex power supplied to a circuit element is p ~ il'. and inductors) are merely configurations of matter and thus have characteristics which depend upon the properties of matter. (179) can be written n. When magnetic matter is considered. and an analogous magnetic quality factor Q~ could be defined. + j2w(W. Eq. 113.")[E]' (180) ""lEI' peak density oC electric energy average density oC power dissipated = 211' peak density of electr\c energy density oC energy dissipated in onc cycle Thus. while conductors have an extremely low Q. AC Behavior of Circuit Elements. A good dielectric is a highQ material. Insight into the interpretation of the impedivity and a. In this case. . 110. capacitors. the I The smallness of an element depends upon the frequency. The basic elements of circuit theory arc small I twoterminal structures whose fields are largely confined internal to the clements. we shall not expand this concept further. Circuit elements (resistors.I/I'Z . In terms oC the energy and power densities .s Q . Since we deal principally with nonmagnetic materials. The interpretation given to Eq. the power supplied to an clement also can be written as (182) p .dmittivity functions of field theory can be gained by considering their relationship to the more familiar characteristics of impedance and admittance of circuit elements. 16.
IV.30 TUlEH. Some authors define P . 113c. (c) complex diagram. G t <bJ 1. We usually classify elements according to their lowfrequency behavior. consider the parallelplate capacitor of Fig.ARM. in which case the lign of react. dependent upon frequency. to somo degree. It should be noted that p.ive power is opposite to that. circuit. A reactor is called an inductor or capacitor according as 1m (Z) is positive or negative. respectively. Furthermore. For an explicit discussion. or capacitor is. 1~13a. equivalent. A capacitor according to circuit concepts.he convention P . complex power to the element is l IVI'(G  jwC) For a H good" capacitor (wC» G) the current leads the voltage by almost 90°. . The element in this case could be classified as a resistor.. the element is called a reactor. (b) element is called a resistor. where the conductance G accounts for energy dissipation (Lnd the capacitance C accounts for energy storage.ONIC ELECI'ROllAONETlC FIELDS + V ~I +i t1. }·13. For a "poor" capacitor (0)> wC) the current and voltage are almost in phase. is a function of frequency. The angle between Ie and I is called the loss angle 8. I. + I. and hence Z. = P = YV = (G + jwC)V (183) The Figure 113c shows the complex diagram representing this equation. which we get.VI. inductor. The relationship or complex terminal current I to complex terminal voltage V is I = I. we shall approximate the field by E=(f V J=A I 1 We are using t.' <oj . the designa~ tion of an element as a resistor. Thus. and the power is principally dissipative. I. and when P is primarily imaginary. The lowfrequency equivalent circuit of tbis element is shown in Fig. and the power is principally reactive. 1·13b. Let US idealize the problem to a capacitor with perfectly conducting plates. 1 • <OJ Flo. as shown in Fig. I I I I V (a) Physical capacitor.
equations. 1 FIc. (c) complex diagram. To demonstrate this. lUb. 114.a. also can be demonstrated. +J V. to ueircuit power.. . (183) shows that Y  fI d A G .fiE where we have taken 4 = u. + V. The equivalency of l<field power.')E + we" + j~') ~ + wi') d A Substituting for E and J from the preceding I = ~~ V = (u V A eompnrison of this with Eqs. The lowfrequency equivalent circuit of this element is shown in Fig. + wi' + i". An inductor aeeording to circuit concepti.. (182). (181).. consider the toroidal inductor of Fig. we have ~ (. of the plates and d is their separation.l d Thus. For our idealized element P III fI'IEI' d. (6) . for our idealized circuit element. where the resistance R accounts for energy dissipation and the inductance L accounts for energy storage. The magnetic properties of matter are similarly related to the circuit behavior of an inductor." Eq. = ZI = (R + jwL)I The complex diagram representing this.(. 114c. = fI'IEI'Ad ~ IVI'Y' We can use this result to define the admittance of a cube and then view admittivity 9 as the admittance of a unit cube. equivalent circuit.." Eq.. The relationship of c!omplex terminal voltage V to complex terminal current I is (184) v = V. The constitutive relationship for the field between the capacitor plates is J . the admittance is proportional to the admittivity of the matter between the plates. (0) Toroidal inductor. v v (a) /{ (b) L V.FUNDAMENTAL CONCEPTS 31 &0 where A is the area. 114. equation is shown in Fig. A C .
The magnetic constitutive relationship for the field in tbe core is M = !H = (wp. (l84). (l82).. l is the average circumference. We now idealize the problem to an inductor of perfectly conducting wire and approximate the field by where N is the number of turns. we can think of impedivity as the impedance of a unit cube. ') 1N'A JWIJ I Comparing this with Eq. and the power is principally reactive. The angle between VI and V is called the magnetic loss angle (I. Using this result to define the impedance of a cube. we see that Z _ I N'A 1 R = WIJ"N2 A 1 L=IJI N2A 1 Thus." + jwp.. (182).l N2A I.I . 114c. Most of our discussion so far has been about wcll·bchaved fields. A summary of the various concepts is given in Table 13. 114. and the power is princi~ ~pally dissipative. and A is the crosssectional area. as shown on Fig. but we have meant to include by implication certain types of aUowable singularities. Singularities of the Field. the power supplied to the inductor is p  III IIHl'dT IIH!'AI.32 TIMEHARMONIC ELECTROM~GNETIC FIELDS The complex power to the element is p ~ JII'(R + jwL) For a good inductor (wL» R) the current lngs the voltage by almost 90°.c circuit concepts and ac field concepts..')11 A substitution for 11 and M from the preceding equations gives .III'Z which is consistent with Eq.• 1.. From Eq. . A field is said to be singular at a point for which the function or its derivatives are discontinuous. For a poor inductor (R» wL) the current and voltage are almost in phase. . the impedance is proportional to the impedivity of the matter. This development serves to illustrate the close correspondences between a. for the idealized inductor.. The element in this case could be classified as a resrstor.0 (" WIJ +.
) Admittivity D(. (144).D CONCEM1I 33 A~C circuit conccpts AC field concepts Complex electric intensity E CompleK magnctic currcnt density M Complex electric current density Complex magnetic intensity H Density of complex powcr now E Impedivity i(". y{".) .f_): JtTE Ii 1 Admittivity. COIUtESPONDENCES BETWEEN AC CIRCUIT CONCEPTS AND AC Y.(R + jwL)l Stored energy.... 9{".'E· E· Density of power dissipll. tT Current density. A13 evidenced by Eqs. 115.C R + j". I Stored energy (~ + j""C) V ~CVV J .')H Density of stored energy." + jWjJ.) Admittance Y("") Resistors: Admittance. as represented by Fig. H..) Conductors (IT Complcx voltage V Complex current 1 Complex power flow VIImpedance Z(". wjJ.ce distribution of currents J_ and M.FUNDAMENTAL CONCEPT8 TABLE 13.) Current. respectively. M . V .. ("'f" Density of stored energy. E· Magnetic propertics: lmpedivity.. !I(". oE .&" »tT): Admittivity.) .. Ii 1 V 1 R VV· Powcr dissipation Capacitors: DeJUlity of power dissipation.'U:I.."'jJ.) .." + jwjJ...e')E Current. the total electric and magnetic currents are vortices of Hand E.R + j".tion.l~ Voltage. By applying (11l5) . 1 . EDielectrics (".(wp. Z(". Admittancc.. Rl1· These can occur at material boundaries (discontinuous ~ and '0) and at singular source distributions.) ... 9(. "'f" Current density. H.. "'f"E ...' Magnetic currcnt. HL1P Power dissipation."H . Suppose we have a surfa.'H' HDeJUlity of power dissipation. such as sheets and filaments of currents.l + j".) +. a Power di!l8ipation ~ VV· InductOIll: Impedance. 1  J x a » ".. y{".
Surface currenU. Whenever J. (186) arc tbe Uboundary conditions" to be satisfied at the source. the boundary condition for H is 'fa "'H. say region 2. no induced surface current can result. Thus.. M." McGrawBill Book Company.crial for which the tangential components of Hare ?cro at its surface. the "boundary condition" at a perfect electric conductor is vanishing tangential components of E. at a filament of electric current I. we obtain 1 (l86) where n is the unit vector normal to the surface and pointing into region (1). For example. and M. are zero. If J. 74. Inc. since y "'" a is infinite. purely a mathematical concept. at a filament of current. "Introduction to Electromagnetic EnginC<!ring. (185) must be satisfied at a filament of magnetic current. a surface conduction current J. Harrington. (186) state that the tangential components of E and H are continuous across the surface. can exist even though E is zero. no matler how small the contour. . are impressed currents. 1_15. Equations (18(.1 (l88) A similar limit of the second of Eqs. however. This is.d! radiul of 00 . F. tangential compomnts of E and Hare e<>nlinuom aero" any material boundary. p. Eq•. (144) express at volume distributions of currents. Eqs. If z and fJ are finite in both regions 1 and 2. They express at current sheets the same concept as Eqs. (185) yield the current enclosed. the field must be singular such that Eqs. Eqs. Equations (186) arc essentially the field equations at sheets of currents.34 Region (1) TllIElLUUolONIC ELEcrROMAGNETIC FIELDS n J. (186) reduce 10 n X H = nXEO J'I at a perfect conductor (187) where n points into the region of field. The superscripts (1) and (2) denote the side of S on which E or H is evaluated. The necessary" magnetic conduction current" on its surface has no physical significance. to rectangular paths enclosing a portion of the surface currents. I R. In this CllSC. If onQ side of S is a perfect electric conductor.) apply regardless of whether or not 8. Finally. 1958. Thus. discontinuity in media exists on S. and M. Region (2) s FlO. perfect conductors excepted. The perfut magnetic e<>nductor is defined to be So mat.
16. 13. Given H ... 19.0 (complex charge density vanishes) in a sourcefree region of homogeneous matter.l.5 X 10. 116. Starting from Ma. 14. (c) H .+. (13). X (B where h is the Hall constant.10 + j5. justify that 1I . Determine i' and k' through the disk z . Show that Q. Show that Eq. determine the (B {or which the second term of the above equation is 1 per cent of the first term. Given g .C dvld4. Unit cube for Prob.. (11) are equivalent to Eqs.11 ). 17. show that Eqs. we can think of an abrupt material boundary aa the limit of a continuous.. x' + y' .0. 16. i . 111. 13. Prove Eqs. but rapid. Determine the instantaneous quantities corresponding to (a) 1 .e/r'l sin (ry) z 11"'<' 1 X y FlO. 110. Using Stokes' theorem and the divergence theorem. . For copper (h _ 5.(2 + j3). frequency of (a) 1 megacycle. determine the Poynting vector.(u. sin y in a 9Ourcefree region of Plexigla. (126) is satis6ed for this field.yl sin wt and :lC . . (b) 100 megacycles. I. 18.useS + j3) + u.xwell's equations. + u. determine n' and ::mI.0 covered by perfect conductors. For the field of Frob. PROBLEMS 11. 1/ ..u. Similarly. . determine E And g at a. a sheet of current can be thought of as a volume distribution of current having a large magnitude and confmed to a thin shell. and the circuit law for inductors. Consider the unit cube shown in Fig. Using an atomic model. (142). For example. linear in the general sense. If B• . change in '0 and!. By such expediencies we can avoid much tedium in the exposition of the theory. derive the circuit law for capacitors.L dildt. (b) E . u"x cos wl.. The conduction current in conductors is affected by the magnetic field 1\8 well as by the electric field (Hall effect). u. 100 sin (ry) and 11.)elt. Show that the instantaneous Poynting vector is given by 5  Re (8 +E X Hel"I) Why is 5 not related to S by Eq.FUNDAMENTAL CONCEPTS 35 It is often convenient for mathematical and discussional purposes to consider the various singular Quantities aa limits of nonsingular quantities. 111..e. 116 which has sl1 sides except the face z .. tr8 + tr'kE. (141)7 111.
U B . it is found to have an impedance of (500 . Derive Eqs.10\ a.u. A small capa. (186). Suppose we have a 10megacycle field B . For a prnctical toroidal induetor of the type shown in Fig.I' ..5. determine (4) tbe timeaverage power d.1.jlOI).jl" is an 8nalytie function of '" and !Ihow that + .J. within tbe cube. and j. Ia. + ..)110' at tbe operating frequency.eitor has a de capacitance of 300 micromicrofaradll when airfilled. neglecting conductor IOIllle8.35 TlKEHARMONIC ELECTROMAGNETIC FIELDS over the open face aDd no sources exist. 10[1(10)  fll dw w' r}o w' (Equations of this type are valid for any analytic function regular in the lower b&1f plane. and ~"of tho oil. in a material having ~ . 113. at tome point. 112.f· 2 __ ~ 101'(10) dw rOw'".u. .(8 . Determine fl. H . IIfi.o. Assume that ~ . ..I4a. Suppose a filament of zdiroctcd electric current Ii .) 117. When it is oillilled.(14 .2.' ( . 1.(l determine the complex power and the timeaverage power supplied by this source.issipated within the cube.10 is impressed along the 2: axis from z os 0 to z = 1. I(w) . Determine each type of current (except impressed) listed in Table 12. (b) the difference betwccn the timeaverage electric and magnetic energies witbin the cube.u. t'.11 X 10' at 0lJ 10'. show that the power IOflll in the wire will usually be much larger than that in a core of 10w·loSli ferromagnetic materiaL 116.
linear (£ and 1) independent of lEI and 181). and substitute from the first equation. homogeneous (z and 1) independent of position). and isotropic (z and y are scalar). A field that is a function of both time and space coordinates can be called a wave. we obtain VXVXHk'H=O (24) Thus.£gE .gE The curl of the first equation is V X V X E = zV X H (21) which. 37 . becomes V X V X E . equation becomes In terms of k. For the present. Electromagnetic fields obey wave equations.te various ac electromagnetic phenomena. called a wave equation. H is a solution to the same complex wave equation as is E. be 11 bit more restrictive in our definition and use the term wave to denote a solution to a particular type of equation. In this chapter we shall consider a number of simple wave solutions to introduce and iIIustra. The frequently encountered parameter k ~ 'Ii £0 (22) is called the wave number of the medium. We shall. the preceding (23) v X V X E .CHAPTER 2 INTRODUCfION TO WAVES 21. take the curl of the second equation. let us consider fields in regions which are sourcefree Ui = Mi = 0). however. (21). The Wave Equation. so the terms wave and field are synonymous for timevarying electromagnetism. upon substitution for V X H from the second equation. The complex field equations are then v X E .£H V X H .k'E = 0 which we shall call the complex vector wave equation. If we return to Eqs.
and u. VI where u.V X V X A In rectangular components.. + u. . Eqs. arc the rectangularcoordinate unit vectors and is the Laplacian operator. (25). such that Eqs. or H:r:... let us consider a solution (29) This satisfies V . I We shall usc the symbol'" to denote "wave functiODs. It is implicit in the wave equations that v·H . (26) and Eqs. (27). Ell. solutions to Eq. (28) Also.. and v'" + k'" k  ". for they do not imply Eqs. the rectangular components of E and H satisfy the complex 8calar wave equation or Helmholtz equation I (27) 0 We can construct electromagnetic fields by choosing solutions to Eq. + u. first of Eqs. this reduces to vIA _ u"V 2A. and H. (26) then reduces to The d'E.VIA.y. In particular. however.0 (25) shown by taking the divergence of Eqs. we can write Eqs. (25) are also satisfied.. A) .38 TIMEHARMONIC ELECTROllLAGNETIC FIELDS The wave equation is often written in another (orm by defining an operation V'A ~ '1('1 . (25) and the operation defined above. . Do not confuse these ".. (25) arc equivalent to Eqs. Solutions to this are linear combinations of eih and ei *'. and E. Thus. (23) and (24). in which case fJ "'" jWf:. H II . take E to have only an x component independent of x and y. They arc not. let us construct a simple solution. Take the medium to be n perfect dielectric. Using Eqs. In other words. E = 0 and is therefore a possible electromagnetic field. (23) and (24) as V'E+ k'E=0 V~H+k2H=O (26) These we shall also call vector wave equations. (27) for E:r:. To illustrate the wave behavior of electromagnetic fields.! "'" jwSl.IVIA." that i9.'s with magnetic flux." U II . (23) and (24). 2 dz + k'E • ~0 which is the oncdimcnsional Helmholtz equation. so general as the previous forms.
and the length of an arrow represents the magnitude of a vector. using Eq. (141). The velocity at which an equiphase surface travels is called the phau " . 120. "'" 377 ohms We shall see later that the intrinsic impedance of a medium enters into wave transmission and reflection problems in the same manner as the characteristic impedance of transmission lines. the & lines are always parallel to the :r: axis.  :JC. 21 illustrates sand 3C along the z axis at t = O. At some specific timc.INTRODUCTION TO WAVES 39 The associated magnetic field is found according to iWIAH = .""" V. The direction of an arrow represents the direction of a vector.ikE~ which.ve. If we take a slightly later instant of time. the picture of Fig. The wave impedance associated with our present solution. 21 will be shifted in the +z direction. is called tbe intrinsic impedance of the medium. The instantaneous fields are found as E.. cos (wt  kz) kz) (213) This is called a plam wave because the phase (kz) of Sand :JC is constant over a set of planes (defined by z = constant) called equiphaae aurfaces. can be written as E~ = ~Hr (210) Ratios of components of E to components of H have thc dimensions of impedance and arc cRUed wave impedames. (212) [.. In this wa. B and X are sinusoidal functions of z. Thc term polarization is used to specify the beha. cos (wt V2 E.. The vector picture of Fig.  V2 E. Jlo = In vacuum. We say that the wave is traveling in the +z direction and call it a. let Eo be real and determine Sand 3C according to Eq.V X E = u. It is called a uniform plane wave because the amplitudes (Eo and EO/JI) of Sand 3C are constant over the equiphase planes. To interpret this solution.vior of & lines. traveling tDaQe. and the wave is said to be limorly polarized in the :r: direction. S and X are said to be in phaae because they have the same phase at any point. (28).
It is given by k~ "'" Zr.. a. x Direction of travel ~ z y Fro. measure of whether a.• dz. distance is long or short. A linearly polarized uniform plane traveling wave.. 22. the value of z.mple.. 1 (214) In vacuum. dt w k v. The range of wavelengths encountered in electromagnetic engineering is large.p must also increase to maintain this constancy.40 TUIEIlARldONIC ELECTROMAGNETIC FiELDS velocity of the wave.he phase velocity dz . according to the above equation.!'O dz d' The phase velocity of this wave is called the intrinsic phase veWdty the dielectric and is.. whereas the freespace wavelength of a lOoomegacycle wave is only 30 centimeters. or X = 2r . The wavelength is often used as a. This is illustrated by Fig. This distance is shown on Fig. the freespace wavelength of a 6(k:ycle wave is 5000 kilometers.. lip of . = v. which is a plot of 8 for several instants of time. 2TV. distance of 1 kilometer is very short at 60 cycles.. and tbe plane z = z. (213) is called the intrinsic wavelength ~ of the medium. 22. this is the velocity of light: 3 X 10' meters per second. k w f (215) where f is the frequency in cycles per second. will move in the +z direction. differentiate the above equation.. (213) is constant. the argument of the cosine functions of Eq. The wavelength of the particular wave of Eqs.1. Thus. To obtain t. . 2. For exa. = zp is defined by wt  kzl' = constant that is. This gives wk./dt. An cquiphase plane z. The wavelength of a wave is defined as the distance in which the phase increases by 2'11'" at any instant. As t increases.
22.z). Waves in Perfect Dielectrics.jh (217) lh H.INTRODUCTION TO WAVES Glt = 41 0 "'t = ". of which free space ia the most common example. This procedure.. other uniform planewave solutions can be obtained by rotations of the coordinate axes./4 "'t".. so we shall consider only the transformations (x. Because of the symmetry of the rectangular ~coordina. = 2'11" = eiJ:. Yo z). z). The usual cireuit theory is based on the assumption that distances arc much shorter than a wavelength. & at several instants of time in alincarly polarized uniform planc traveling but very long at 1000 megacycles. wave. +z traveling wave.e.. In this section we shall consider the properties of uniform plane waves in perfeet dielectrics. corresponding to cyclic interchanges of coordinate variables.te system.2. ~ v. E z = Eoe.= D • e/h .x. w/2 z. To summarize. where k = w.y'.z) to (Y.x.. 2.. We have already given a special case of the uniform plane wave in the preceding section... gives us the four waves E. (216) ~=~ " It is an :vpolarized./h E H w = .>1 FIG... We wish to restrict consideration to +z and z traveling waves. and to (y.y. together with our original solution. z 1 . = Ce ih • • 0 H= __ e • • H~+ = H w = A eih + B . to (x.
.42 TUlEIL\RlIONlC ELECTROMAGNETIC FIELD8 where the preVlously used Eo has been replaced by A. (217) in Sec.. but the velocity of propagation of energy is never greater than the velocity of light. 21. as Thus. ~ H " =JCOS"'" obtained by combining the first and third waves of Eqs. (217).nd (219) we find v.' S = & X ~ = g = •E. or D. the phase velocity may be greater or less than the velocity of light. We have not yet mentioned power and energy considerations. In gen· eral. so let us do 80 now. hall of the We can define eo power flow densit. (217) with • (220) . = (219) For the uniform plane traveling wave from Eqs.' cos' (loll 2 kz) w. These two velocities are not necessarily equal for other types of electromagnetic wa. the electric and magnetic energy densities are equal. w. :JeWcity of propageui<Jn of energy v.denotes a z traveling wave. = ~ X' = fE o' cos' (wt . and the superscript.he wave being electric and hall magnetic. e' "'" fE. ... We have already interpreted the first wave of Eqs.. B J C.ves. + w.. The superscript + denotes So +t traveling wave. (214)]. 1 which is also the phase velocity [Eq. . (218) a.kz) (218) E.y S energy density . j v. Another property of waves can be illustrated by the uanding ~ E IIl == Eosin kz ..E. Given the traveling wave we evaluate the various energy and power quantities as w• .V.• u.' E X H* u. This also constitutes an interpretation of the other three waves if the appropriate interchanges of coordinates are made.  cos' (loll  kz) energy of t. The most general uniform plan~ wave is a superposition of Eqs.
= ~ X' = fE o2 cos 2 kz sin 2 wl S = & X:JC = u. The corresponding instantaneous fields are 3C~ _M = v2Eosinkzcoswt = V2coskzsinwt Eo • Note that the phase is now independent of 1" there being no tra. & and 3C are 900 out of phase. The energy and power quantities associated with this wave are w. and vice versa. with & reaching its peak value when 3C is zero. The timeaverage Poynting vector S = Re (5) is zero. The planes of zero & and 3C are fixed in spa<:e.INTRODUCTION TO WAVES 43 A =  C &". The wave is still a plane wave. It is still linearly polarized. A picture of energy x z y FlO. for equipbase surfaces arc planes. . for its amplitude is constant over equiphase surfaces. = i &2 ". The electric energy density is a maximum when the magnetic energy density is zero. A picture of E and 3C at some instant of time is sho'm in Fig. The field oscillates in amplitude. In other words.veling motion i hence the name standing walle.. It is still a uniform wave.. 23.. for E always points in the same direction (or opposite direction when 8 is negative). 2~ sin2kzsin2wt S = E X H$ = UIU sin2kz E2 'E' 2. 23. 23. the zeros of & being displaced a quarterwavelength from the zeros of X. showing no power flow on the average. A linearly polarized uniform plane standing wave. fE o2 sin 2 kz cos' wt (221) w. <E jE o/2. and vice versa. as shown on Fig. Successive zeros of 8 or of 3C are separated by a half· wavclength.
and if IAI = ICI. perfect electric conductors can be placed over one or more of these planes.u n b A+C vAt AC ) (223) The rms amplitude of E is + C2 + 2AC cos 2kz which is called the 8tandingwave pattern of the field.) H II = :Ell ! (Aeib  (222) If A = 0 or C 0. = nr. For A ¢ C. or E. = Aeih + CeJ'h Ceil. = • VAS + CS + 2AC cos 2kz e_jtanl ( . 24. for it corresponde to a judicious choice of 2 and t origins. (217). Standingwave pattern of two oppositely traveling waves of unequal ampli tudes... n an integer. (220) represent the solution to the problem of reflection of a uniform plane wave normally incident on this conductor. Thus. If an electric conductor covers the plane z "'" 0. we have a pure traveling wave. (220) represent the solution of a oncdimcnsional "resonator.44 TUllEHARMONIC ELECTROMAGNETIC FIELDS A+C AC FlO. This gives E. If two electric conductors cover the planes kz = th'F and kz = ntr. This is illustrated by Fig. The voltage output of a small probe (receiving antenna) connected to a detector would essentially follow this standingwave patI This is actually no restriction on the generality of our interpretation. This is a superposition of the first and third of Eqs. let us take A and C reaP and express the field in terms of an amplitude and phase. Note that we have planes of zero electric intensity at kz. Eqs. Eqs. .." A more general xpolarized field is one consisting of waves traveling in opposite directions with unequal amplitudes. Z 24. we have a pure standing wave. oscillating between the electric and magnetic forms can be used for this wave.
.nt. the standingwave pattern is a const. a "rectified" sine wave. (217). From Fig..8)..lt = 5'11'/4 _ X I /'  rotates in this direction e ~ ". If A = 0. Letting A = IAj&"' and B =. that of a pure standing wave is infinite.A H = + u. with the axis of polarization inclined at an angle tan. we have the instany y e vibrates In this direction . This is illustrated by Fig. ".'l. that is. For a pure traveling wave.. . This is a superposition of the first and second of Eqs. 25.. it is evident that SWR~A+C A C (224) because the two travelingwave components lEqs.. The standingwave ratio of a pure traveling wave is unity. E = (u.t.t"" '11'/2 FlO. traveling wave in which both E~ and Ell exist. t will no longer point in a single spatil11 direction. I t. (b) elliptical polarization. and for a pure standing wave.t = '11'..lt = / X 0 . with SWR's between one and infinity. the wave is linearly polarized in the y direction. Ii A and Bare complex with different phase angles. If A and B are both " real (or complex with equal phases). it is of the form Icos kzj.. Polaritation of a uniform plane traveling wave.'. The ratio of the maximum of the standingwave pattern to tbe minimum is called the standingwave ratio (SWR). 25a. that is.. (u~B + uIIA) !e (225) jh If B = 0./2. (a) Linear polarization.INTRODUcrtON TO WAVES 45 tern. The distance between successive minima is >. Let us now consider a.1 (B/ A) with respect to the x axis.. (222)] add in phase at some points and add 1800 out of pbase at other points.IBleitl . 24. we again have a linearly polarized wave. Plane traveling waves reflected by dielectric or imperfectly conducting boundaries will result in partial standing waves. . the wave is linearly polarized in the x direction.t "" 3"/4 /' t = TT/4 .
giving E .ju. there is no change in energy and power densities with time or S = E X H· = UJ  2 S "'" t X :JC = u. 21 for this wave would show 8 and:JC in the form of two corkscrews. The specialization of Eq.kz + a) + b) A vector picture of S Cor various instants of time changes in both amplitude and direction. and the field is said to be circularly polarized. this ellipse caD be of arbitrary orientation in the xy plane and of arbitrary axial ratio. The polariza. going through this variation once each cycle. a = 0.julI)j Eo eiA. For example. (225) to righthanded circular polarization is obtained by setting A = jB = Eo.. The polarization is said to be lefthanded if 8 rotates in the opposite direction. If the axial ratio is unity. .kz (wt . Depending upon . 25b.tion is said to be righthanded if 8 rotates in the direction of the fingers of the right hand when the thumb points in the direction of propagation.J'· (226) H = (u". :or ~ ac' "'" fE o' (227) "2 Eo' " Thus. and b = 1r/2. A plot of & for various values of t in the plane z = 0 is shown in Fig. . Linear polarization can be considered as the special C!\8e of elliptic polarization for which the axial ratio is infinite.  &" = v'2 [EI cos v'2IA[ cos (wt .(u. . let IAI = 21BI. The tip of the arrow in the vector picture traces out an ellipse. Circular polarization gives a steady power flow J analogous to circuittbeory power transmission in a twophase system.)E.Eo' space.46 TIMEHARMONIC ELECTROMAGNETlC FIELDS taneous electric intensity given by &. Tho various energy and power quantities associated with this wave are " to. the tip of the arrow traces out a circle. with 8 perpendicular to :JC at each point. As time increases.Ii and B. A vector picture of the type of Fig. this picture would rotate giving a corkscrew type of motion in the z direction. and the field is said to be elliptically polarized.. .
26. This field can represent resonance between two perfectly conducting planes situated where E is zero. giving a standingwave pattern in the z direction. This is the superposition of Eqs. A vector picture of t and :Ie at t = 0 is shown in Fig. 26.. D = B = E o/2. = .INTRODUCTION TO WAVES 47 As a final example. It thus seems that the picture of energy oscillating between the electric and magnetic Corms (229) . consider the circularly polarized standingwave field specified by E = (U. x z y FIo. As time progresses. the amplitudes of t and :re being independent of time.u ll sin wt) :JC = (u" cos wt  v'2 Eo sin kz u ll sin wt) V2 Eo cos kz . + julI)Eo sin kz (228) H = (U. It is only the direction of t and :JC which changes with time. The corresponding instantaneous fields are t = (U. cos wt . this picture rotates about the z axis. X' = tEo' cos' kz S=tX3C=O 5 = u l l Eo'sin 2kz It is interesting to note that the instantaneous energy and power densities are independent of time. A circularly polarized uniform plane standing wave. &' = f. (217) for which A = 0 = jEo/2. + JUII) Eo cos kz . The energy and power densities associated with this wave are w... however. = . a function of z. The amplitudes of t and :re are..E o' sin' kz w. Note that t and :re are always paraUel to each other. .
We have already seen that when k = k'. We actually have two coincident resonances (called a degenerate case)." + wE" + jwE' + jwJJ. the circularly polarized standing wave is the sum of two linearly polarized waves which can exist independently of each other. and may be written as k=k'jk" (232) where k' is the intrinsic phase constant and kIf is the intrinsic attenuation constant. 27." are always positive in sourcefree media. = u Z = wJJ. and JJ.' Re FIo. However. Intrinsic Wave Constants. for they account for energy dissipation. The behavior of k can be illustrated by a complex diagram rehting k to ! and y. The product zQ then usually lies in the bottom hall of the complex . Complex diagram relating k to I and 1). in general. We shall see in the next section that kIf causes an exponential attenuation of the wave amplitude. (211) when i and g are not specialized to the case of a perfect dielectric. as shown in Fig. it enters into the phase function of the wave. and the picture of energy oscillating between electric and magnetic forms applies to each linearly polarized resonance. The second equation is a generalization of Eq. and hence specifies the characteristics of the medium. The wave number is. 27.' are usually positive but may be negative for certain types of atomic resonance. 23. This is shown in Fig. f". obtaining (231 ) A knowledge of k and" is equivalent to a knowledge of ! and y. (230) for z and y. the wave Dumber k and the intrinsic impedance 7Jt given by k = yzg (230) play an important role. We can solve Eqs. Thus. When the wave aspects of electromagnetism arc emphasized. complex. The parameters E' andJJ. 27. (211).48 TIMEHARMONICELECTUOMAGNETIC FIELDS is not generally valid (or resonance. In the 1m expressIons fJ 0'. z and f) usually lie in the first quadrant of the complex plane. obtained in the same manner as Eq.
and land p. v.' are usually positive." are always positive. Expressing" in rectangular componenUl.I!IlDI the last equality following from Eqs. When " or p. consider the case of no magnetic losses. The ratio I/g therefore usually lies in the right half plane and " in the sector ±45° with respect to the positive real axis. 28. and k is real. In losslcss media. 28. lies in the fourth quadrant. From the first of Eqs. k" is positivej it is only k' that could conceivably be negative.' is negative. • . (234) no mBgnetic losses . The intrinsic wave impedance can be considered in an analogous manner. The complex diagram relating" to 9 and I in general is shown in Fig. the wave impedance is real. we have Now for: = jwp. k "'" V £0. and p.Jp. In lossless media. but at is never negative. In sourcefree regions. There are several special cases of particular interest to us. I'..jkk' ~ . we have (233) where at is the intrinsic wave resistance and X is the intrinsic wave reactance.jk . FIG. First. we bAve I ik· jk·. We shall see in Sec. showing that k' and k" arc usually positive. Even when l or p.INTRODUCTION TO WAVES 49 1m to' and g. g = jWf. Thws t usually lies in the first quadrant and l/g in the fourth quadrant. that X introduces a phase difference between & and x. (231). I = jr. Ccmp1ex diagram relating" x 1/9 plane. The principal square root. " may lie anywhere in the right half plane. """ jl~l. " is purely resistive and is there~ fore the ratio of the amplitude of & to X. 24.' is negative. (230). For a wave in a perfect dielectric.
ble 21 shows these parameters separated into real and imaginary parts. In this C8. "~ (' "~ Good conducLOr ~ if .:") AND tJl j!r. In this case.NCI: (If  (t .R .. with ~ »Wf. (2) good dielectrics. The perfect dielectric case is that for which k=w~ f/=~ This is summarized in row 3 of Table 21.50 TIlLEHARMONIC ELEC!'ROAUGNETIC FlELDS TaLE 2.'(lj.R '" '" Separation into real and imaginary parts is shown explicitly in row 2 of Table 21.. 213.) "'" V jW.•" ) 'J7 I + ) 2<' V k· which is summarized in row 4 of Table 21..I J ~ 2t' Good dielectric 2 . Y = f1 + jWf. a good conductor is characterized by ~ = jWJ.g 0 k" Ifl Perfect dielect..1.I. fJ = we" + jwl..'» l'. A good dielectric is chaneterized by! = jWj. we have kw~""(ljj)~wv..t.. V..f Im~ 1m v::t9 ""R k' lm~ Ifl 0 '" No magnetic IOI!lell n. . v. with f..' • The last row of Ta.1. A similar simplification can be made for the ease of DO electric losses.. we have k = jwp(u + jWf. and (3) good conductors.. ..I.l.ric .r.lej~ + INTRlN81C k' k" til General "" v' . WAVE NUlI8ER bPEO. ..SC. .) " .. (See Pooh.:.l' ... ( .uT "IUl~ WJ.) Three special cases of materials with no magnetic losses arc (1) perfect dielectrics.IUl k' ~ r. Finally.
INTRODUCTION TO WAVES
51
24. Waves in Lossy Matter. The only difference between the wave equation, Eq. (27), for lossy media and lossfree media is that k is complex in lossy media and real in lossfree media. Thus, Eq. (29) is still a solution in lossy media. In terms of the real and imaginary parts of k, it is Also, H is still given by Eq. (210), except that "l is now complex. the H associated with the E of Eq. (235) is Thus,
(236)
H w = Eo ell. = Eo eHel".cjl'.
•
1.1
where "l = II'JIeir. and (236) are
The instantaneous fields corresponding to Eqs. (235)
&.. =
V2 EoC
k
".
cos (wt  k'z) cos (wt  k'z 
J<;, =
v2 ~ e
k
".
n
(237)
Thus, in lossy matter, a traveling wave is attenuated in the direction of travel according to c k "', and X is no longer in phase with S. A sketch of & and 3C versus z at some instant of time would be similar to Fig. 21 except that the amplitudes of & and JC would decrease exponentially with z, and 3C would not be in phase with & (:Je usually lags e). A sketch of &:0 versus z for several instants of time is shown in Fig. 29 for a case of fairly large attenuation. A sketch of :Jew. versus z would be similar in form.
c.
Direction of travel _ ~ Envelope  e II".
~

 z
(lit = ../2



<>It =
"'/4
"'t 
a
FIG. 29. & at sevClal instanUJ of time in a linearly polarized uniform plane traveling wave in dissipative matter.
52
TJMEHAR~[QNtC
ELECTROMAGNETIC FIELDS
The wave of Eq. (237) is still uniform, still plane, and still linearly polarized. So that our definitions of phase velocity and wavelength will be unchanged for lossy media, we should replace k and k' in the lossfree formulas, or
A=2'1'=~
k'
f
(238)
Then v. is still the velocity of a. plane of constant phase, and X is still the distance in whicb the phase increases by 2,... Two cases of particular interest are (1) good dielectrics (low·1055), and (2) good conductors (highloss). For the first case, we have (see Table
21)
in good dielectrics (til
« i)
(239)
Thus, the attenuation is very small, and e and JC arc nearly in phase. The wave is almost the same as in a lossfree dielectric. For example, in polystyrene (see Fig. 110), a lomegacycle wave is attenuated only 0.5 per cent per kilometer, and thc phase difference between Sand 3C is only 0.003 0 • The intrinsic impedance of a dielectric is usually less than that of free space, since usually l > fll and jj = jja. The intrinsic phase velocity and wavelength in a dielectric arc also less than those of free space. In the high·loss case (see Table 21), we have
k'=~
k" =
~w;o.
1,1
~
"1/.
T
fo;j,
in good conductors (<1 » Wf)
(240)
!~4
Thus, the attenuation is very large, and 3C lags e by 45°. The intrinsic impedance of a good conductor is extremely small at radio frequencies, having a magnitude of 1.16 X 1O~ ohm for copper at 10 megacycles. The wavelength is also very small compared to the freespace wavelength. For example, at 10 megacycles the free·spacc wavelength is 30 meters, while in copper the wavelength is only 0.131 millimeter. The attenuation
INTRODUCTION TO WAVES
53
in a good conductor is very rapid. For the abovementioned lOmego.cycle wave in copper the attenuation is 99.81 per cent in 0.131 millimeter of travel. Thus, waves do not penetrate metals very deeply. A metal ncts as a shield against electromagnetic waves. A wave starting at the surface of a. good conductor and propagating inward is very quickly damped to insignificant values. The field is localized in a thin surface layer, this phenomenon being known as skin effect. The distance in which a wave is attenuated to lie (36.8 per cent) of its initial value is called the skin depth or depth oj penetration 3. This is defined by k"3 = I, or
! _ ~ 2
WP.d
_,!" _ ~.
k'
(241)
2r
where A.. is the wavelength in the metal. The skin depth is very small for good conductors at radio frequencies, for A... is very small. For example, the depth of penetration into copper at 10 megacycles is only 0.021 millimeter. The density of power flow into the conductor, which must also be that dissipated within the conductor, is given by
S == E X H* = u,lllo!217..
where H 0 is the amplitude of H at the surface. The timeaverage power dissipation per unit area of surface cross section is the real part of the above power flow, or watts per square meter
(242)
where <R ",. Re ('1...) is the intrinsic resistance of the metal. <R is also called the surjace resistance and 11... the surface impedance of the metal. Eq. (242) is strictly true only when the wave propagates normally into the conductor. In the next section we shall see that this is usually so. In most problems Eq. (242) can be used to calculate power losses in conducting boundaries. (An important exception to this occurs at sharp points and corners extending outward from conductors.) More general waves can be constructed by superposition of waves of the above type with various polarizations and directions of propagation. For waves uniform in the xy plane, the four basic waves, corresponding to Eqs. (217). are
H + ",. A e1"'ejlt'
" l{~+ B e'''~ef''~ "C H, " H., = D e1"'ei1' "
,
= = e!""eJ1"
(243)
54
Tl'fEHARMONIC ELECTROl'llAGNETIC FIELDS
The preceding discussion of this section applies to each of these waves if the appropriate interchange of coordinates is made. A superposition of waves traveling in opposite directions, for example
E,. ...
H II =
Aek"leit'.
+ CeJ<"'ei1h
 Cef"'gil")
! (Ae.t"'ei1"
(244)
•
gives us standingwave phenomena. However, it is no longer possible to have two "equal" waves traveling in opposite directions. One wave is attenuated in the +z direction, the other in the z direction; hence they can be equal only at one plane. Suppose that the wave componenta are equal at z = 0, that is, A = C ,in Eq. (244). There will then be standing waves in the vicinity of z = 0, which will die out in both the +z and z directions. This is illustrated by Fig. 210 for a material having fairly large losses. Far in the +z direc~ion ~he +z traveling wave has died out, leaving only the z traveling wave. Similarly, far in the z direction we have only the +z traveling wave. The standingwave ratio is now a function of z, being large in the vicinity of z = 0 and approaching unity as Iz] becomes large. For very small amounts of dissipation, say in a good dielectric, the attenuation of the wave is small, and standingwave patterns are almost the same as for the dissipationless case. Other superpositions of Eqs. (243) can be formed to give elliptically and circularly polarized waves. In a picture of a circularly polarized wave traveling in dissipative media, the If corkscrews" for E and :JC would be attenuated in the direction of propagation. Also, & would be somewhat out of phase with:re. A circularly polarized standing wave would be a localized phenomenon in dissipative media, just as a linearly polarized sta.nding wave is localized. 26. Reflection of Waves. We saw in Sec. 114 that the tangential components of E and H must be continuous across a material boundary.
 __
Z
e""a
Flo. 2·10. Standing.wave pattern or two oppositely traveling waves in dissipative matter.
INTRODUCTION TO WAVES
55
Region (2)
A ratio of a component of E to a component of H is called the wave impedance in the direction defined by the crossproduct rule applied to the two components. Thus, continuity of tangential E and H requires that wave impedance. normal to a material bO'Und
Region (I)
Incident
Transmitted
ary must be continuoua.
•
Reflected
..
The simplest reflection problem ill that of a uniform plane wave normally incident upon a plane boundary FIG. 211. Reflection at a plane dibetween two media. Thill is illustrated electric inte.rfa.ce, norma.l incidenoe. by Fig. 211. In region I the field will be the sum of an incident wave plus a reflected wave. The ratio of the reflected electric intensity to the incident electric intensity at the interface ill defined to be the reflection coefficient r. Hence, for region 1 E a (1) c:z Eo{Cik,. + reiA:,.) H (I) = E, (eJkl.l  reik,l) _
II
~1
In region 2 there will be a transmitted wave. The ratio of the transmitted electric intensity to the incident electric intensity at the interface is defined to be the transmission coefficient T. Hence, for region 2
E.(t.) = EoTe jlt••
H
•
(2)
:lO:iI
Eo Tefl"
~.
For continuity of wave impedance at the interface, we have
Z.
I_0 =
E.(I) H (1)
1/
I
.0 = ~1 I
~.
I
+r _ r
= ~.
where '/1 and ~. are the intrinsic wave impedances of media I and 2. Solving for the reflection coefficient, we have
r =
From the continuity of E. at z given by
=
'/, + ~1
~
2'/.

'/1
(245)
0, we have the transmission coefficient
T ~ I
+r
~.
+ '/1
(246)
If region 1 is a perfect dielectric, the standingwave ratio is
SWR
E~ _ 1
=
E~.  1  lrl
+ Irl
(247)
56
TDmHARMONIC ELECTROHAGN·ETIC FlELDS
y'
z
Fro. 212. A plane wave propagating at an angle
~
,, ,
with respect to the
z
%0%
plane.
because t.he incident and reflected waves add in phase at some points and add 1800 out of phase at other points. The density of power transmitted across the interface is
lio  Re E X H'·
u.1._0  s..,(l 
Irl')
(248)
where Si..  E.I/'lJl is the incident power density. The difference between the incident and transmit.ted power must be that reflected, or
(249)
We have used an xpolarized wave for the analysis, but the results are
valid for arbitrary polarization, since the :z: axis may be in any direction tangential to the boundary. Those of us familiar with transmissionline theory should note the complete analogy between tho above planewave problem and the transmissionline problem. Another reflection problem of considerable interest is tha.t of 8. plane wa.ve incident at an angle upon a plane dielectric boundary. Before considering this problem, let us express the uniform plane wave in coordinates rotated with respect to the direction of propagation. Let Fig. 212 represent a. pla.ne wave propaga.ting at an angle E with respect to the xz plane. An equiphase plane z' in terms of the unprimed coordinates is
z' "'" zcosE+ysin;
and the unit vector in the v' direction in terms of the unprimed coordinate unit vectors is Uy' . u.. cos ~  u.sin ~
U1TRODUCTION TO WAVES
57
The expression for a uniform plane wave with E parallel to the z "'" 0 plane jg the 6rst of Eqs. (217) with all coordinates primed. Substituting from the above two equations, we have
E• .... Eoeitbol..f+'_ fl
H = (u.,cos
~

u.sin~) ~Orfl(r"f+_fl
(250)
The wave impedance in the z direction for this wave is
Z. "" E. _ _ ._
HI/
cos
t
(251)
In a similar manncr, from the second of Eqs. (217), the exprcssion for a uniform plane wave with H parallel to the z = 0 plane is found to be
E  (u"cos t  u.sin E)Eoe1t (l1oJ. H .o H. = _ Eo eit(l1ol.t1(+._o
(2.52)
•
The wave impedance in the z direction for this wave is
E. Z.=:a::17C08t
H.
(253)
Thus, the zdireet.ed wave impedance for E parallel to the z  0 plane is always greater than the intrinsic impedance, and for H pnrallel to the z = 0 plane it is always less than the intrinsic impedance of the medium.. Now suppose that a uniform plane wave is incident at an angle t ... 8; upon a dielectric interface at z  0, as shown in Fig. 213. Part of the wave win be reflccted at an angle E......  8r , and part transmitted at an angle E  8,. Each of these partial fields will be of the form of Eqs. (250) if E is parallel to the interface or of the form of Eqs. (252) if H is parallel to the interface. (Arbitrary polarization is a superposition of these two
RegIon (1) RegIon (2)
FlO. 213. Reflection at a plane dielectdc interface, arbitrary angle of incidence.
z
58
T1MEHARMONIC ELECTROMAGNETIC FIELDS
cases.) For continuity of tangential E and H over the entire interface, the y variation of all three partial fields must be the same. This is so if
k l sin 8, = k 1 sin Or = k! sin 6,
From the first equality, we have
(254)
that is, the angle of reflection is equal to the angle of incidence. second equality, we have
l sin 8,  =  = V, = k sin 8; k, VI
From the
.j§'"'
~2~!

(255)
where v is the phase velocity. Equation (255) is known as Snell's law of refraction. The direction of propagation of the transmitted wave is thus different from that of the incident wave unless IiWt = fWt. In practically all lowloss dielectrics, 1J.l = Ilt = /le. If medium 2 is free space and medium 1 is a nonmagnetic dielectric, the righthand side of Eq. (255) becomes VEl/EO = V;;, which is called the index of refraction of the dielectric. The magnitudes of the reflected and transmitted fields depend upon the polariza.tion. For E parallel to the interface, we have in region 1
E~(I)
= A(gil,._., =
+ rei
l ,._ ••)
H~(I)
",
~ cos 8;(ell,..; 
reil,..,)
wbereA includes the y dependence. in region 1 at the interface is
Thus, thezdirected wave impedance
Z~(t) = EP) = !! 1 + r H~(I) cos 8,1  r
This must be equal to the zdirected wave impedance in region 2 at the interface, which is Eq. (251) with t .... 8,. Thus,
r =
711 112
sec 8, sec U, +
111 111
sec U; sec 8;
(256)
Note that this is of the same form as the corresponding equation for normal incidence, Eq. (245). The intrinsic impedances are merely replaced by the zdirect.ed wave impedances o{ single traveling waves. It should be apparent {rom the form of the equations that, for R parallel to the interface, the reflection coefficient is given by
r ...
112 cos U l fit cos 0,
+ 'Ill cos 0;

'11
cos 8;
(257)
INTRODUCTION TO WAVES
59
In both cases we have standing waves in the z direction, the standingwaye ratio being given by Eq. (247). Two cases of special interest are (1) that of total transmission and (2) that of total reflection. The flf'S(. case occurs when r = O. For E parallel to the interface, we see from Eq: (256) that r = 0 when
Substituting for 9, froro Eq. (255) and for the '1'5 from Eq. (211) we obtain
(258)
as the angle at which no reflection occurs.
real solution for
(Ji.
This does not always have a
In fact, sin
(Jj _
...,,..
co
For nonmagnetic dielectrics (PI = PS = po) there is no angle of total transmission when E is parallel to the boundary_For the case of H parallel to the boundary, we find from Eq. (257) that r """ 0 when sin 8; =
f.JEI 
Ill/P.I
Es/Et 
EdEs
(259)
E.
Again this does not always have a real solution for arbitrary po and But in the nonmagnetic case
(260)
There is usually an angle of total transmission when H is parallel to the boundary. The angle specified by Eq. (260) is called the polarizing angle or Brewster angle. If an arbitrarily polarized wave is incident upon a nonmagnetic boundary at this angle, the reflected wa.ve will be polarized with E parallel to the boundary. The ease of total reflection occurs when Irl = 1. We are considering lossless media; so the 'l'S are real. It is apparent from Eqs. (256) and (257) that Irl ~ 1 for real values of 8, and 81• However, whon flJ.l1 > EtJ.lS, Eq. (255) says that sin 8, can be greater than unity. What does this mean? Our initial assumption was that the transmitted wave was a uniform plane wave. But Eqs. (250) specify a solut.ion to Maxwell's equations relYlrdless of the value of sin E. It can be real or complex. All that is changed is our interpretation of the field. To illustrate, sup
Thus. (255) becomes RjX . called a react1've field or an evanescent field. if EI > ft. from Eq. in the nonmagnetic case. = . Eq. Note the 90° phase difference between E~ and Hili so the wave impedance in the z direction is imaginary.I}J!!fl}Jl' Thus. (250) become (262) which is a field exponentially attenuated in the z direction.. (263) is called the critical angle. Remember that the field in region 2 is not zero when total reflection occurs.ve impedance imaginary in region 2) is smB. (wa. The reflection coefficient.. becomes of the form r = R +jX when total reflection occurs. When region 1 is a nonmagnetic dielectric and region 2 is a nonmagnetic conductor. Of particular interest is the case of So plane wave incident upon a good conductor at an angle 8.ll > EU'! or. (255) it is evident that sin 0.'s and 8's are allowed to be complex.. Optical prisms make use of the phenomenon of total reflection. (252) when sin t > 1. Eq. .60 pose sin Tl1ofEHARliONIC ELECTROMAGNETIC FIELDS e> 1 in Eqs. total reflection occurs only if the wave passes from a "dense" material into a C1less dense" material. Returning now to our reflection problem. and there is no power flow in the z direction.l. Eqs. A wave incident upon the boundary at an angle equal to or greater than the critical angle will be totally reflected. is greater than unity when sin 8. All the theory of this section can be applied to dissipative media if the . > Yf. It is an exponentially decaying field."" EIJ. A similar interpretation applies to Eqs. (257). Note that there is a real critical angle only if EJJ. (2~56) or Eq. the point of transition from real values of 81 (wave impedance real in region 2) to imaginary values of 8. (250) and let = ksin!~~ keos! kv'l sin'!:= ±ja (261) If we choose the minus sign for a.l (263) The angle specified by Eq.§. It is evident in this case that IrJ is unity.
It is also postulated t. we have where Y is a shunt admiUance per unit length. Let us review the circuit concept of a transmission line and then show its relationship to the field concept.dz i'1 Co) FtG.n ~ r. the wave can be considered to propagate normally into the conductor regardless of the angle of incidence. Thus. Let Fig. Thus. Taking the derivative of the first of Eqs. dVIZdz where Z is a series impedan.ion line accordi. 214a represent a. equivalent circuit. i}'+V+d~ ~r. For each incremental length of line dz there is a series voltage drop dV and a shunt current dI.. 26.0 dz' d'l .ZYl. .. thc ac transm.ng to circuit concepts. (264) and substituting from the second. Transmissionline Concepts. The general solution .0 (:HIS) which are onedimcnsion:ll Helmholtz equations. The circuit theory postulate is that the voltage drop is proportional to the line current I. Dividing by dz.ce per unit length. For most practical purposes.IZ dl dz~ VY (264) Implicit in this development are the assumptions that (1) no mutual impedance exists between incremental sections of line and (2) the shunt current dI flows in planes transverse to z.. we obtain d'V dz' .4 l+dI l}+v ===j:======ll===". A lnmrro.dz I (b) (a) Phyaicalline. current is proportional to the line voltage V. dl ~ VYd. 214.i8sionline equation8 dV d. (6) This is an extremely small quantity for good conductors.lNTROD'OCI'ION TO WAVES 61 . The transmission line is said to be uniform if Z and Y are independent of z.hat the shunt.ZYV . twoeonductor transmission line.
"iio H.+ Vj)~ Z. COMPARISON OF TRANSMISSIONLINE WAVES TO UNIFOIW PLANE WAVES .+. _ 0 Transmission line Uniform plane wave d.y'ij 8.._ ~ H..+.66).Voer' we have from Eqs.I t +. .' + k'E" II d'/ .0 d. (2.. The reader haa probably already noted the complete analogy between the linearly polarized plane wave and the transmission line. .y p .' . with propagation constant (266) Choosing the +z traveling wave V+ =. we assumed no mutual coupling . _ H. _ ik .""" + 1.. and it is common practice to write Z=R+iwL Y G+jwC (268) The equivalent circuit of the transmission line is then as shown in Fig..'V  d'V 0 d'E" d.VP " _ Eo+ __ 8.  V+ ]+  V1' rz (267) which is called the characteri8tic impedance of the transmission line.U: is a sum of a +z traveling wave and a z traveling wave..' 0 V _ VZY V.+ 8 0fJ S.64) that /+ = I atr'" V+ Z 'Y 1+ = :y = y Substituting for 'Y from Eq..E../10 + E.[0. (2.. In the circuit theory development."i l• V.+e/~' 1 . 214b. we have z...62 TIMEHARMONIC ELECTROMAGNETrc FIELDS TABLE 22..rtf . This analogy is summarized by Table 22.e"· + H.11+ . The imaginary parts of Z and Yare usually positive.' d'H I +k'H __ _ a.r· + Voe'" T E.
or in the vicinity of a discontinuity on the line. Let us therefore specialize the problem to that of perfect conductors immersed in a homogeneous medium. In this section we shall restrict consideration to transmissionline. then gives 'YE" = JH. the zdirected wave impedance of any TEM wave is the intrinsic wave impedance of the medium. We assume E. 1 Hu. ax ay = 0 It follows from these equations that 'Y = jk (269) The propagation constant of any TEA! wave is the intrimic propagation constant of the medium. which must have an E. and a TEM wave is called a trammissionline mode. mode. = fJEI/ ax ay aH" _ all. We can summarize this by defining a transverse Laplacian operator V. = H. From the field theory point of view.INTRODUCTION TO WAVES 63 between adjacent elements of the transmission line. or H. 'YE. is required to support the zdirected current. For the TEM mode to exist exactly. or an H. The higherorder modes are usually important only in the vicinity of the feed point. are called higheroroer mode8. the conductors must be perfect.'E = 0 The boundary conditions for the problem are E. Finally. modes. the boundaryvalue problem for E is the same as the electrostatic . The proportionality of components of E to those of Ii: expressed by the above equations can be written concisely as • Thus. or TEM. AU other waves. (21). manipulation of the original six equations shows that each component of E and H satisfies the twodimensional Laplace equation. = H. Eqs.. for Maxwell's equations show that infinitely many wave types can exist. abbreviated TEM. This is not the only wave possible on a transmission line.HXu. = JH1/ aE" _ aE~ = 0 'YH. or both.I = E=. this is equivalent to assuming that rio E. exists.XE (270) at a' ax' + ay1 (271) and writing V. = flE~ 'YH. Expansion of tbe field equations. or else an E. = 0 and z dependence of the form er. Each possible wave is called a. Such a wave is called transverse electromagnetic. = 0 0) at the conductors (272) Thus.
c. V ( E.ve V~'/. The electrostatic and magnetostatic problems h:ve E and H everywhere orthogonal to each other and are called conjugate problems. Hxu"dl~. the characteristic impedance of the line magnetostatic inductance per unit length by Z. the second of Eqs..nce of the transmission line is related to the electrostatic capacitance per unit length by Zo = T V = 7/ C • (274) Similarly.. the line voltage and current are related to the fields by V ~ /.dl I 1 }e.ted to each other through Eqs. The boundaryvalue problem for H is the same as the magnetostatic problem having "a. H. To show the relationship of the statio L's and C's to the Za of the transmission line. (270) we have From the second of these and I  ~ ( '1 le. Cross sedion of a. dl But in the corresponding electrostatic problem the capacitance is c _ .c. consider a cross section Flo./.64 TrnEHARMONIC ELECTROMAGNETIC FIELDS problem having the same conducting boundaries.15. 2.. ( H.0.dl In the corresponding magnetostatic problem we have L=!.c. It is for thiS reMon that II static" capaci tances and inductances caD beAused for transmission lines even though the field is timeharmonic. .!. (273) and (270) we ha./.  IS related to the V L T .) boundaries. u. In the transmissionline problem. (275) Note also that Land Care rela. .'L _. E. of the line as represented by Fig. the characteristic impeda. dl V}c. transmission Iino. 215. Therefore.nticonducting" (no H . Thus. 215. from the first of Eqs. (274) and (275). H·d! (273) where C1 and Ct are as shown on Fig. X E· d! _! ( 'I le._!!. E·dl 1.
. d h»d Shielded pair l@ n . When the dielectric is lossy but the conductors still assumed perfect. log 2.w 1D »b CoUinear plate >i'. (I+~ " b + Vb' . (275). 1D h»d w»d Once the electrostatic C or the magnctostatic L is known. (274) or Eq.. all of our equa.t) D' D' D »d • »d Wire in trough . T d ~ Z . . Zo (proportional to .h) Z. .. __ 101_ • 2D ~'o wire 0 tD+l ~ D »d • d Coaxw @ ~ Iwl. • Confocal elliptic Z.y T w _h"j/d z.1+  • 4D Z...(:. ... _log• w D» 1D Wire above ground plane h hi°.tions still apply. the Zo of the corresponding transmission line is given by Eq. _ . . Iog• 4h 2..jk) ..: log d " C' +.. Table 23 lists the characteristic impedances of some common tra..' . • ..!. Parallel plate b b Z..!..L T I+D~ Z.INTRODUCTION TO WAVES TAllLE 65 23. Iog• b 2. log (4W """dtanh2.) and 'Y ( . CRARACf£R16TJC bIP£DANCE8 OJ' Bolla CoUKON TRAl\I6Jo11SSl0N LINES Line Geometry Characteristic impedanco Z.nsmission lines..
) . The attenuation constant in this case is the intrinsic attenuation constant of the dielectric (Table 21. IJ>. This concept of wave guid/ ance is quite general and applies to / many configurations of m. and the power flow is given by p. For example. Z. row 4).ARMONIC ELEC!'ROMAGNETIC FlELDS become complex. Thus. the most com· monly used waveguide is the reclan· y b gular waveguide. 21J>.. (276) While this equation is exact if d'>" and ~I are determined exactly. Hence a +ztraveling wave will be of the form V 1.66 TlMEB. The rectangular waveguide...lt. The waves on a transmission line can be x z viewed as being guided by the conductors. It is a hollow conducting tube . FlO. illustrated by Fig. the waves will still be characterized by a propagation constant "y = a + ifJ. .atter. . = 2a(f>. in terms of timeaverage powers. the field is no longer exactly TEM. d. column 2. IV. We shall carry out such a calculation for the rectangular waveguide in the next section.. and exact solutions are usually impractical. its greatest use lies in approximating a by approximating <P". The most important effect of this is that the wave is attenuated in the direction of travel. = V I· . 2·16..Re (P.I' e.. • Poef. (242) holds at their surface.d<J>. However. Z· .. IJ>. When the conductors arc imperfect. or <J>" . In general. . or. 216.~· The fate of decrease in ~I versus z equals the timeaverage power dissipated per unit length (1)01. 27. the attenuation constant is given by .Re (P... systems which guide waves are called waveguides. attenuation due to losses in imperfect conductors can be approximated by assuming that EQ.). Apart from transmission lines. Waveguide Concepts.
and the subscript n denoting the choice by Eq.. the subscript 0 denoting no variation with x. so there is no tangential component of E at the wall y = O. There remains the condition that E" = 0 at <::: y = b. . The modes in a waveguide nrc usually classified according to the existence of z components of the field. is said to be a tramverse electric (TE) mode.INTRODUCTION TO WAVES 67 of rectangular cross section..t _ e. the parameters (277) (278) " . are called eigenvalues. For k real (lossfree dielectric). . A mode baving no E. . sin (k. which is satisfied if n = 1. Each choice of n in Eq.~ with respect to the xz plane (see Fig. the propagation constant 'Y can be expressed as k >T b n. (279) determines a possible field.kllalnE)eih_t = 2jA sin (ky sin~) ejh . (250) and write E" = A(cikllth.nd k. There is only an E. The complete system of modes will be considered in Sec.k' The above field can now be written as E. = E. are related by + cos 2 E = 1. (280) k < n1r .2j A) and define kc=ksin~ 1'=jkcosE ~ In view of the trigonometric identity sin i l' a. The modes represented by Eqs. (279) These permissible values of k. If the waves are xpolarized.. E" """ 0 at y = 0. (278) and (279) have DO E. All the modes in the rectangular waveguide fall into one of these two classes. we use Eq. The particular modes that we are considering are TEon modes.2. or charaeteri8tic values of the problem. is said to be a transverse magnetic (TM) mode.. 43. 212). so no component of E is tangential to the conductors x = 0 and x a.3. E Let Eo denote (.y) r>' let us see if this field can exist within the rectangular waveguide.k.' . and arc therefore TE modes. One having no fl. Also. or mode. (279). Fields existing within this tube must be characterized by zero tangential components of E at the conducting walls_ Consider two uniform plane waves traveling at the angles ~ and ..
At frequencies less than f. mode. and the mode is called a propagating mode. the mode propagates. Using fJ from Eq. and (282).. = VI (Uf)' (2&5) showing that the guide wavelength is always greater than the intrinsic wavelength of the dielectric.). obtaining I.68 'Y TUllEHARMONIC ELECTROMAGNETIC FIELDS where a and fJ are real. approaching k~ as f + O.. When 'Y = a. In particula. the field decays exponentially with z. it is eviden t that (283) Using the last equality and k as CI 27:/ W in Eq. n (281) This is called the cutoff frequency of the TEo. the mode is called a nonpropagating mod€. the phase constant fJ of a propagating mode is always less than the intrinsic phase constant k of the dielectric. The guide phase velocity v. In this case. When a mode propagates. we have wave propagation in the 2: direction. The attenuation constant of a nonpropagating mode is always less than k~. intrinsic wavelength The corresponding (282) x. This follows from Eqs. (wavelengths greater than A~).y. approaching k as / + co. (284) 1 <I. When = ifJ. Thus. we can solve for the transition frequency. {JAg = z. from Eqs. (284). The transition from one type of behavior to the other occurs at a = 0 or k "'" n1r/b. the mode is nonpropagating. Letting k = 2rf v'~p. we have X. and there is no wave propagation. that is. At frequencies greater than Ie (wavelengths less than >". _ 2b n is called the cuwff wavelength of the TEon mode. or an evanescent mode. (279). (281). Thus.. the concepts of wavelength and phase velocity can be applied to the mode field as a whole. is defined as the . A knowledge of /~ or A~ is equivalent to a knowledge of k~i so they also arc eigenvalues.... = 2b .r. the guide wavelength A is defined as the distance in which g the phase of E increases by z. (280). (277) and (279).. we can express 'Y I> I.
standingwave ratios.Ii .. we find '. etc.ve impedance in (288) H.. _ j". ~ This is ca. . _ E. The characteristic impedance of a nonpropagating mode is reactive and approaches zero as f . a z traveling wave is possible.E. All our discussion SO far has dealt with waves traveling in the +z direction.newave reflection. in a manner analogous to that used to derive Eq. If we substitute into the above equation for "y from Eq. The concepts of reflection coefficients. (This we have not shown.lled the characteristic impedance of the mode and plays the same role in reflection problems as does the Zo of transmission lines. for . (289) v(un' 1 1 <I. The dominant mode in a rectangular waveguide. The guide phase. The simultaneous existence of +z and z traveling waves in the same mode gives rise to standing waves. = J"" . '. (278) according to V X E "'" jwJJI..y) r<' H" = H. el" (287) J"'. The wa.. assuming b > a. Thus. we find I> I. propagating mode is always greater than the intrinsic impedance of the dielectric.Vi . The result is E. Thus. also apply to waveguide problems. For each +z traveling wave. used in the case of uniform pla. velocity is therefore greater than the intrinsic phase velocity. (Un' (286) where v.. (287). ~ Eo cos (keY) where E. is the intrinsic phase velocity or the dielectric.INTRODUCTION TO WAVES 69 velocity at which a point of constant phase of & travels.. let us find H (rom the E of Eq. the characteristic impedance of a TEo.. is the TEOl mode. To show this.L Eo sin (keY) tr". The mode with the lowest cutoff frequency in a particular guide is called the dominant mode. (214). obtained by replacing "y by 'Y in Eqs. approaching '1 as f . has been repeated for convenience.O. the z direction is z. (284). sin (k. Another important property of waveguide modes is the existence of a characteristic wave impedance. 00. .
. l . ' . 217. 217. I!. is real and is therefore the timeaverage power transmitted. This gives PI'" foG fob E~H: dx dy = IEol! 2i: which. Table 24 specializes our preceding equations to this mode and includes some additional parameters which we shall now consider. I I . If only one mode propagates in a. The rectangular waveguide is usually operated so that only the TE ol mode propagates. I... (282) with n = 1. . . Below cutoff. AB time progresses.70 y )lltplt I I 'I TIMEHARMONIC ELEcrnoMAGNETlC FIELDS . this will be the only mode of appreciable magnitude except near sources or discontinuities. the mode pattern moves in the z direction. we see that the cutoff wavelength of the TEo! mode is Xc :>:: 2b. .. The mode pattern of the TEo! mode in the propagating state is shown in Fig. however. waveguide. This figure is obtained by determining E and JC from the E and H of Eqs. so bad as it seems at first. indicating no timeaverage I We are referring to the intrinaic wlLvelengt. . . . Because of the importance of the TED I mode.rie filling the waveguide. above cutoff.. It is admittedly confusing to learn that many modes exist on a given guiding system. .. This is therefore the only wave of significant amplitude along the guide except near sources and discontinuities. wave propagation can take place in a rectangular waveguide only when its widest side is greater than a halfwavelength. '. The power transmitted along the waveguide can be found by integrating the axial component of the Poynting vector over a guide cross section.! A sketch oC the instantaneous field lines at some instant is called a mode pattern. Mode pattern for the TE ol wAveguido mode. It is not. . .. the power is imaginary.. 9 { . which is usually free space. (287) and specializing the result to some instant of time.. x e. let us consider it in a little more detail. Thus.h of the dieleet.) From Eq. we have not considered all modes. lines into paper Lines out of paper • •• FIG.
INTRODUCI'ION TO WAVES 71 TABi. '.ge and current associated with a waveguide mode. SUMMARY OF WAVEOUIDE PARAMETERS FOR THE DO). TJI H cose"'" • ... Characteristic impedallce . there is no unique volta. the amplitude of a modal traveling wave (Eo in Table 24) enters into waveguide reflection problems in the same manner as V in transmissionline problem.. (The preceding equation applies only at z = 0 below cutoff unless the factor cta.2b Propagation constant {j~ "1'  a . Attenuation due to lossy dielectric a.a" VI m U<ff)t [2..  VI VI ..[INANT MODE (TE ol ) IN A RECTANGULAR WAVEOUIDE Complex field b Eo .n' (fll. However. . Eo sin TY e"f' Cutoff frequency Cutoff wavelength I. I <I. 232).E 24.(f. V1 u." I b E~ .  '....WII z. (f<ff)1 Guide phase velocity '.)! f >1. ~)'] +b J 1 power transmitted.s. jkVI 2.. (f<!f)' Power transmitted p _ IEoltab 2Z.( i"/V1(f<lfP 1) .) It is also interesting to note that the timeaverage electric and magnetic energies per unit length of guide are equal above cutoff (see Prob. Guide wavelength A.ff)' 1>1. . . In contrast to the transmission·line mode.  1 2b V III ). is added.  WI" 2"VI (f<!f)' Attenuation due to imperfect conductor a.IV . TY 1 H .. I <I.8m e ' v Zo b Eo I. .
+z traveling wave. .ture. In the z traveling wave. mode current. Other definitions of mode voltage. mode current] such that Zo = T v P = VI .a. Our treatment has 80 far been confined to the ideal lossfree guide.V /Zo. except that most parameters become complex. we can continue to use the relationship provided f is Dot too close to f~. the ratio V/ I is a. case shown dashed). I = . 218. the characteristic impedance is complex at all frequencies. it is common to define a mode voltage V and a. There is no longer a real cutoff frequency. Remember that we are dealing with only 0. These alternative definitions will always be proportional to our definitions (see Prob. function of z. The behavior of'Y = a + jfJ in the lowloss case is sketched in Fig. 218.nt (or & l066y waveguide (lossfree. 234). it is evident that V = EO~eT. for 'Y never goes to zero. When waves in both directions are present. o f . In the lowloss case. The behavior of 'Y for the lossfree case is shown dashed. The most important effect of dissipation is the existence of an attenuation constant at all frequencies. Also. When losses are present in the dielectric but not in the conductor. Propagation const.72 TIMEHARMONIC ELECTROMAGNETIC FIELDS To emphasize this correspondence.jk" and referring to • Flo. (290) From Table 24. and characteristic impedance can be found in the litera. Letting k = k' . (291) satisfy this definition. all our equations still apply.
or "perturbed.(!lIE. we have a.I' (:i)' J..1_. 6lZ. The power per unit length dissipated in the wall y = 0 is lP.I') dy (. 7). and Eq.' d% The power per unit (!lIE.I' [i. . The tangential component of E is now not quite zero at the conductor.)..I' = <Ix ~ (!lIE.. The total power dissipated per unit length is the sum of that for the four walls. and the field is only slightly changed.(!l f. because the boundary conditions are cbarrged.If)' [I + b (I.)' (2a + b) ] ab Zo' .I' (. (242) is used to approximate the power dissipated in the conductor. L.I'a (~)' and an equal amount is dissipated in the wall y "" b. or lP. for good conductors. Even more important is the attenuation due to imperfectly conducting guide walls.). When both . + (:i)' (20 + b)] Equation (276) is valid for any traveling wave.[b + (I. However.. ~. vI (!l V. we find (292) This is t!le attenuation constant due to a lossy dielectric in the guide. Such a procedure is called a perturbational mdJwd (see Chap. + Gi)' ~] b dy and an equal amount is dissipated in the wall x "" a. the tangential component of E is very small.] 7 This is the attenuation constant due to conductor losses. = (!l jo (lH." from the lossfree solution. jo (ry/b) Zit [sin' + (I. ." IH. The lossfree solution is used to approximate H at the conductor.I' [2.I' + IH.1 2a a.lNTRonVC'I'lON TO WAVES 73 Table 21. Our solution is no Jonger exact in this case. length dissipated in the wall x = 0 is lP.(!lIE. cos' ry] "1/1 . so using the above and ~I "*' P of Table 24. _ (!lIE.
. we choose pc = 24.ei/l. 219. When losses are present. Resonator Concepts. in the vicinity of a resonant frequency. Resonators can therefore be used for the same purposes at high frequencies as LC resonators nrc used at lower frequencies. The TElll waveguide mode already satisfies this condition over four of the walls. 28. We recall that standing waves have planes of zero field. In Sec. consider the "rectangular cavity" of Fig. is Solving {or the resonant frequency I = In we have f _ r2bc .clec~ troroagnetic fields can exist within Y a sourcefree region enclosed by a FIG. called resonant frequencies. These fields can exist only at specific frequencies. perfect conductor. the total attenuation constant is z a""'a<J+a~ (294) a b for by Eq. In the lossfrceca. we choose E" = E~+ = + E~. dielectric. 22 we noted a similarity between standing waves and circuit theory resonance. like the impedance of an LC circuit.) Eosin (7) sin pz 'Jr. this is the resonant frequency of . The rectangular cavity. This consists of a conductor enclosing a.!. 219.sc. both of which we will assume to be perfect at present. To illustrate resonator concepts. a source must exist to sustain oscillations. (276) we merely add the two losses. For E" to be zero at z = 0.= A sin 'Jr: (ei'" . according to Table For E~ to be zero at z = c.74 x TIMEHARMONIC ELECTROMAGNETIC FIELDS dielectric losses and conductor losses need to be considered. which. which sugges~ trying the standingwave TEol field. We desire to find solutions to the field equations having zero tangential components of E over the entire boundary. The input impedance seen by the source behaves. ~b' + c' tJl (295) When a is the smallest cavity dimension.
rz C08..2'W. FlO. 220.. we have for the TEo II mode it ·ry·r? BIt = E .I .. From the conservation of complex power..I'abc cavit. •• ...SIDbSlD C II II = jbE. we know that 'W. and W..!J(~ e~. The additional subscript 1 indicates that we have cbosen the first zero of sin f3z. The higher zeros give higherorder modes.IN'TRODUC'I'ION TO W AYES 75 the dominant mode. .~. \. • . (139).. _ _ ._. " Vbt _/ + ct ..~' • I I I I II<. will be maximum and twice its average value. smt t "V b + c C c out of phase... / .. ~ ."y . so & is maximum when 3C is Note that E and Jl are 90 minimum and vice versa. Also of interest is the energy stored within the cavity..' • • • • ""\ I I .• • • • • • • • • • • • J L. Eq.. Therefore.. ' ) .l · 1 b /// . w . ••• •..z Slncosb c (296) jcEo . that is. = i IEol abc t (298) is the total energy stored within the cavity.. . Thus.. will be zero. (168).• \ . '\'1. modes with higher resonant frequencies.... . Setting f3 c: rIc in the above expression for E and determining H from the Maxwell equations. 220.... ]( we choose a time for which 3C is zero.. A sketch of the instantaneous field lines at some time when both & and 3C exist is given in Fig. . ~_ . "'" 'W. ~I I • . ~ i IB.. Mode pattern for the TE'11 cavity mode.. I '\ I I I \ \ JI. r1J .. Eq. =  '\\l... called the TEolI mode.I·····I·11 I I . the timeaverage electric and magnetic energies are ll. that the total energy within the resonator is independent of time.. ~ ~ fff lEI' d.\.)' (297) We also know from conservation of energy. ~ '\\l.
but if a > b we no longer have the dominant mode.~y IHI' d.76 TIMEHARMONIC ELECTROMAGNETIC FIELDS When the resonator has losses. We assume H at walls to be tha. we have 1. and Eq. which is evident from Eq. (179). (298).+Q.2m bc(b' + ct)H + c') + 2a(b' + c') a(bt (2101) maxi~ From the symmetry of Qc in band c. however.~I~~'c') [bc(b' + c') + 2a(b' + c')] ••n. .11'1 Q. and by corrosion of the metal.m(1 + b/2a) (2102) The Q also increases as a increases. Usually more important in determining the Q is the loss due to imperfect conductors. (242). This idealized Q will. consider a cubic cavity constructed of copper. As an example of the Q's obtainable. we have 11'"'1 Q. To summarize. When both conductor losses and dielectric losses are considered. . the Q of the cavity becomes Q~Q. it is evident that b = c for mum Q.v. For a II squarebase " cavity (b = c).t of the lossfree mode and calculate &d by Eq. This is determined to the same approximation as we used for waveguide attenuation. (299). by imperfections in the construction. we have (2100) so the Q of the resonator is that of the dielectric. we define its quality factor as X energy stored w'W Q . This is valid for any mode in a cavity of arbitrary shape. If the losses are dielectric losses. (295) into Eq. Eq. c.1. gives Q's of several thousand. I I I (2104) . ~ m 1f. be lowered in practice by the introduction of a feed system. Eq. at microwave frequencies. ~ 2". In this case we have Q.average power dissipated = lYd _ w (299) by analogy to the Q of an LC circuit.07 X 10'/0 (2103) which. Sub8tituting this. (299). . 19.
In homogeneous media i·he two potentials are in the ratio".A) ... or impressed.klA = ] . a constant.V(V . (2110) are ca. (2105). by a vector identity.>tV4> (2108) which. In homogeneous media. (2107) + jw.lled wave potential~. Consider the field equations v XE ~ jw. To obtain the equation for A. In terms of the magnetic wave potential. substitute Eqs. but the extension to lossy media is effected by replacing jw~ by ! and jlNt by g. J .Voj> where ¢I is an electric acalar potential. .j". Radiation.1ectromagnetie theory it is more oommon to let A be the vector paten\ial of B.voj> Only V X A was specified by Eq.k'A = V • A. Solutions to Eq. (2105).INTRODUcrlON TO WAVES 77 29.A = .V'A . we have E = jw. This gives v X V X A . We are still free to choose If we let v·A = jWt~ (2109) the equation for A simplifies to V'A + k'A  J (2110) This is the Helmholtz equation..V XA (2106) where A is called a magnetic vector 'Potential. becomes v(v . the divergence of the first equation is v·H = 0 Any divergcnceless vector is the curl of some other vector i so H .0 Any curlfree vector is the gradient of some scalar. we have V X (E Substituting Eq.jc. (2106) + jw.l into the first of Eqs. E Hence.A H~VXA + J. or complex wave equation. A) J'" (211I) I In general e. These equations apply explicitly to a perfect dielectric. current. We shall now show that a source in unbounded space is characterized by a radiation of energy. (2106) and (2107) into the second 01 Eqs. (2J06).. A) .JI v XH jW<E+J (2105) where] is the source.
o . . (2106). k . clement. Bence. and take where C is a constant.78 TIJrLEIL\RMONIC ELEcrROMAGNETIC FIELDS obtained from Eqs. 221. (In dissipative media. and the above equation reduces t. and (2109). .!.. forming a current elermnl or x electric dipole of moment Il. cur~ rent I extending over an incremental II. but the solution must also reduce to the static field u r .dA. and the Becond an inwardtraveling wave.) + k'A • _0 dr This has the two independent solutions the first of which represents an outwardtraveling wave. as shown in Fig.k' . (2107).r<'. r dr 2 (r. so we take A to have only a z component. for which the solution is A._E 4".ed and situated at the coordinate origin. Thus.y length l. let A.O.(r). the coordinate origin. = A. satisfying z . C might be a function of k.) We therefore choose the first solution. A zdirect. this current element to be zdircct. 221.. The current is zdirected. (2lJOY reduces to Poisson's equation. + ktA~ = 0 everywhere except at the origin. Let us first determine A for a. The principal advantages of using A instead of E or H are (1) rectangular components of A have corresponding rectangular com~ ponents of J as their sources and (2) A need not be divergenccless.'!. VIA. Eq. and the first solution vanishes as T + co. at.ed eurrcnt.ik". Take Fra. and the second solution becomes infinite. I As k+ 0. The scalar quantity AI has a point source Il and should therefore be spherically symmetric. C is not Go function of k. 1 To be preciee.
. I~ll' [1 . Eqs. 10" d8 r 2 sin 0 E.(!. The outwardtraveling wave represented by Eq.H: (2115) = • 2. sm  ' r r2 8 Very close to the current element.~ ~ll' The timeaverage power radiated is the real part of PI. or I (2116) This is independent of r and can be most simply obtained from the racliation field. (2112) is called a.(iw"+"+ JWfCr.8m8 1). At intermediate values of r the field is called the induction field. and the field is said to be quasistatic.. (2111). The result is E = 2'It1'" 1 r eib (!L + _.Il 2XT 2XT 11 eib sin 8 H. which is negative.INTRODUcrrON TO WAVES 79 Our constant C must therefore be and hence (2112) is the desired solution for the current element of Fig. since surfaces of constant phase are spheres. Eq. The electromagnetic field of the current element is obtained by substituting Eq. 221. (2114). .] ~I = . = '11 L . indicates that there is an excess of electric energy over magnetic energy in the near field. the H reduces to that of a constant current element. = L eib sin 8 l T» X (2114) which is called the radiation field. The outwarddirected complex power over a sphere of radius r is PI = t:ffi E X H* .). (2112) into Eqs. the E reduces to that of a static charge dipole. The reactive power. r: JWfCr ' (2113) eib (ik + 1)._1_) cos 8 r2 Jl '. ds = 10'2" dq.=_e~r E 41r r Jl H. spherical wave. = 4'11" . (2113) reduce to E. Far from the current clement.
4ilr . 222. r is the distance from the source to the field point. (2112) can be generalized to a current element of arbitrary orientation by replacing Il by /1 and A. and a superposition over all such elements is (2118) .y'.r'l To emphasize that A is evaluated at the field point (x. for a current distribution J. for the equations are linear.y'p + (z  ZI)Z Note the direction of the vector potential is that. 222. we shall use the notation A(r) and fl(r'). Thus. x To obtain the field of an arbitrary distribution of electric currents. The above equation then becomes Il(r')c. Radius vector Dotation. (2112).x'p + (y . r should be replaced by jr  t'l = V(x .y.z') . so Eq. The ufield coordinates" arc specified by r=u~+u. by A. we need only superimpose the solutions for each element.y+u.iilrr'l A(r)  4rlr _ <'I (2117) Finally. A superposition of vector potentials is usually the most convenient one.z) and II is situated at the source point (x'. Jl not at the coordinate origin.z and the Hsource coordinates II by r' = U~' + llllY' + u. the vector potential from current element of arbitrary location and orientation is A _ II eilIrr'1 .z' For In Eq. the current element contained in a volume element dT is J dT. For this purpose. of the current. we shall usc the radius vector notation illustrated by Fig.80 TIMEHARMONIC ELECTROMAGNETIC FIELDS z FlO.
we must know the current on the antenna. (2122) into Eqs. The medium may be dissipative if k is considered to be complex. The linear antenna.INTRODUCTION TO WAVES 81 The prime on dT' emphasizes that the integration is over the source coordinates. (2121) must be retained in the «phase term" ei11rf'I. An exact determination of the current requires the solution to a boundaryvalue problem. the radiation field is relatively insensitive to minor changes in current distribution. (2114) for all elements of current. Equation (2118) is called the magnetic vector potential integral. . we shall consider the linear antenna of Fig.e'l = r .E. We therefore have a formal solution for any problem characterized by electric currents in an unbounded homogeneous medium. 223. it is called a dipole antenna. The magnetic vector potential. substitute Eq.··ILI2 L/2 l(z')&"I:o'' dz' r»L Note that the second term of Eq. = . A device whose primary purpose is to radiate or receive electromagnetic energy is called an antenna. Fortunately. . but not in the llamplitude term" Ir . (2118). for this particular problem is _ 1 A. Eq. . 210. When it is energized at the center. = J1'W/l sin 0 A. and mueh use • r y X L/2 I(r') FIo. To evaluate the radiation field. It is intended to include the cases of surface currents and filamentary currents by implication. ~ T» Z' (2121) (2122) 41rr . L12 1 r 'I r 2rz'c080 (2119) (2120) Ir r'l = vrt+z't The radiation field (r large) is of primary interest.} z T H. It consists of a straight wire carrying a current I(z). This gives E. large rr' • (2123) This result is equivalent to superimposing Eqs. To illustrate antenna concepts.e'II..b where ILI2 I(z')eJilrr'l d. in which case Ir . 223. Antenna Concepts.z' cos 0 and A. (2111) and retain only the l/r terms. To obtain the field components...
11. 2111'" rib[COS(k~COS~) . We have already seen that on transmission lines the current is a harmonic function of kz. = IaeibjLI':J. llI' (2128) where I is some arbitrary reference current.1rT L/2 2 =~ I.(2. =jT/I.1' [cos E. and continuous at the source (z = 0). This is also true for the principal mode on a single thin wire. = E..eit./'1.cos H) sm B r d6 (2127) The radiation resistance R r of an antenna is defined as iJ>. 2 [cos (k~COS 0) . For the dipole antenna. sin[k(~lzll)]ejh. Hencc. or 2 ~I = .. ~ cos 6) . Note" that the radiation field is linearly polarized._tdz' 4.cos H)]' (2126) The total power radiated is obtained by integrating S.  • _ ... (2129) . over a.11.lly picked as J.1' r H Jo [cos R.. .COS(k~)] sm 8 (2125) with H. for there is only an E. (2123).H.)' H cos sin 6 6) . the reference current is usua.82 TIMEHARMONIC ELECTROMAGNETIC FIELDS Cui information can be obtained from an approximate current distribution. The density of power radiated is the r component of the Poynting vector s.cos (k~)] ksin 2 8 From EQ. The current on the dipole antenna must be zero at the ends of the wire. symmetrical in <:. large sphere.. Sr r 2 sin BdB dtIJ 10 10" 2'lr ~ .. we choose (2124) The vector potential in the radiation zone can now be evaluated as A. Thus. the radiation field is E.
is an alternative method of showing radiation characteristics. 2. 244).24. 120 80 40 antenna. the radiation field pattern is essentially the bracketed term of Eq.25. o This integral can be evaluated in terms of tabulated functions (see Prob. . as it is for the dipole antenna. 225 for kL small (short dipole). A graph of Rr versus L is given in Fig.. Radiation field paUerm ror the dipole antenna. defined as a plot of ISrl at constant r.. This is shown in Fig. the power patt./2 210. Radiation reIlist.. 2. For a. When the radiation field is linearly polarized. (2125)..ern is the square of the field pattern.ance or the dipole / R 160 . The radiation power pattern. (halfwavelength dipole). The radiation field pattern of an antenna is a plot of lEI at constant r in the radiation zone. (fullwavelength dipole). 224. kL = .INTRODUCTION TO WAVES 83 280 240 200 Flo./ '/2 L 310. The gain g of an antenna in a given direction is defined as the ratio of the power required from an omnidirectional antenna to the power FIG. dipole antenna. 1/ . and kL = 2.
. a "loss resistance" must be added to Eq.sin'lk(L/2)J In the limit as kL is mnde small. A knowledge of the reactive power. 8m 2 and the input resistance is R. 2. We define the input resistance of a lossfree antenna as R.84 TIllEDARMONIC ELEcrROltAGNETIC FIELDS required from the actual antenna. Similarly. C082" EO: ..ted and I. _ • . (2132) to obtain the input resistance. 2 (2131) In the limit kL . the maximum gain is 2.64. The input resistance accounts for the radiated power (and dissipated power if losses are present). if L = X/10. for a fullwave dipole. we find (2133) R. kD I i"'" .1 ohms (2135) .(kL)' 24. . 224 and Eq. (2130) For L :S X. . we have g{7T/2) = 1. we use Fig.. . g( 2 . which cannot be obtained from radiation zone fields. is the input current. For the dipole antenna. we can use Fig. the ratio of the complex terminal voltage to the complex terminal current. (2132) where ~I is the power radia. (2126) and (2128)./2.11. For example.5j so the maximum gain of a short dipole is 1. .. the maximum gain of a dipole antenna occurs at 8 From Eqs. assuming equal power densities in the given direction.. we have ~).0.[% ( 1~19. fIJi ~. that is.5. . the input resistance is about 2 ohms. (2134) The short dipole therefore has a very SDlali input resistance.41.73. For a hallwave dipole.C08 kL)' ~R.. (2133) and find R. is needed to evaluate the input reactance. _ R.R. For the haUwavelength dipole.24 and calculate a maximum gain of 1. Thus. . If losses are present. kL)' = 11 ( 1. The input impedance of an antenna is the impedance seen by the source.
Surfaces over which the phase is constant (instantaneous function vibrates in phase) are called equiphase 8'Urfacu.z)ei 4l (··r••) (2136) where A and ~ are renl. The instantaneous phase of a wave is the argument of the cosine function of Eq. in general. The phase ~ of the complex function is thc init. or spheres..y. On Waves in General. p~ • a<i> ay P. (The term phase constant is used even though it is not. but not infinite. The input resistance of the fullwavelength dipole is actually large. (2137). 211.y.ial phase of the instantaneous function. (140) is called a tooVt function. The rate at which the phase decreases in some direction is called the phase constant in that direction. may be expressed as '" = A (x.) cos [wI + "(z. These are.Y. 713). cylinders. of course.. = (2139) These may be considered as components of a vector pha8e constant defined (2140) The maximum phase constant is therefore along the wave normal and is of magnitude IV~I. The corresponding instantaneous function is V2 A (x. in the direction of V~ and are the curves along which the phase changes most rapidly. a constant. (2133) shows R o '"'" co.y. Waves are called 'Uniform when the amplitude A is constant over the equiphase surfaces.')] (2137) The magnitude A of the complex function is the rms amplitude of the instantaneous function. the phase constants in the cartesian coordinate directions arc p .iNTRODUCTION TO WAVES 85 For the fullwavelength dipole.ax • by a. or spherical according as their equiphase surfaces are planes.) For example. which has a null at the source.z) ~ constant (2141) . This incorrect result is due to our initial choice of current. and depends markedly on the wiro diameter (see Fig. Eq. cylindrical.re called plane. A complex function of coordinates representing an instantaneous function according to Eq. A wave function y" which may be either n scalar field or the component of a vector field. Perpendiculars to the equiphasc surfaces are called walle normals. A surface of constant phase is defined as wt + "(x.. These are defined by (2138) Waves a.
At any instant. along cartesian coordinates arc v" = ~ For example. ax 0<%1 04> + . (2136). Phase velocity is not a vector quantity.dy + . l) must decrease to maintain the constancy of Eq. the sur~ faces of constant phase coincide with the equiphasc surfaces. the phase velocities ~ v" = . (2141) must vanish. The dircction of a wave impedance is defined according to the rightband <I crossproduct" rule of comyonent E .86 TIMEHARMONIC ELECTROMAGNETIC FIELDS that is. As time increases. the total differential of Eq. A vector propagation constant can be defmed in terms of the rate of change of e as r . is V4>· ds >:::  i» d... iJy iJz To keep the instantaneous phase constant for an incremental increase in time. ds = 0 That is. complex function whose imaginary part is the phase 1'. the instantaneous phase is constant. In the electromagnetic field. as (2144) where e is a. The components of IX arc the logaritbmic rates of change of the magnitude of y. (2140) and a: is the vector attenuation conatant. =  a()/az ~ = i. ratios of components of E to components of Hare callcd wave impedances. we must have (oj dt + V<}l . For any incrementds the change in ifl.va . and the surfaces of constant phase move in space. The phase velocity of a wave in a given direction is defined as the velocity of surfaces of constant phase in that direction.• + j~ (21'15) where (J is the phase constant of Eq.. w The phase velocity along a wave normal (ds in the direction of v<P) is (2143) which is the smaUeat phase velocity for the wave. We can also express the wave function.d. in the various directions. (2141). Eq. fJll!jay = fJM (2142) v.
ve impedance in the z direction is Z. the meaning of the word "intrinsic. so the vector phase constant is ~ . E.E. a. fl. since k' is constant. and H. H.. . = Z*"+ . is Eaet'. 80 the vector attenuation constant is u . is a wave impedance in the z direction. = k tl ..u.ialized to the uniform plane traveling wave are all intrinsic parameters. Consider thc xpolarized ztraveling wave in lossy mntter... _ E~·"·c"Jr'· H . is constant over each equiphase surface... = fJ. u. so the wave is uniform. The cart. """ 0. E. . The cartesian components of the phase constant are fl. .." E H.I = 0. by z = constant.= Z_. a.= u.esian components of the attenuation constant are a. the z component is s. while (2l47) E 1/ = Z. The wave normals all point in the z direction.= constant. by definition.u. Note that the various parameters sJ>C(.. The amplitude of E. = k'.INTRODUCT10N TO WAVES 87 rotated into component B. For example.= E.k'. is The wave impedance in the __ """ Z. (2149) The conccpt of wave impedance is most useful when the wave impedances arc constant ovcr equiphase surfaces." . is II.+ = Z.+ = Z".. so the wave is a plane wave. Equiphase surfaces are defined by k't.. The vector propagation constant is y ~ • + 1~ . The phase velocity in the direction of the wave normals is ~JI = wlk'.k". or. +z direct. c''''cit'. The amplitude of E. (2148) The Poynting vector cnn be expressed in terms of wave impedances.· and its phase is _k'z.(k" + jk') . Let us illustrate the various concepts by specializing them to the uniform plane wave.I H w "" 71.ion involving E. = Z. For example. This is. (2146) wave impedance in the +z direction.jk Tbe wa. These arc planes. .
.V. show that the wave is circularly polarized if E. 22. show that the phase velocity i.t the velocity of propagation of energy 8.e tI" . (2. (222). For the field of Eqs. Given a uniform plane wave traveling in the +z direction. + .  v.88 TIMEHARMONIC ELECTROr. Show that the quantities of Eqa. (23)..20)... . Show tha. ). show tha. .fAGNETIC FIELDS PROBLEMS 21. . 1 (A+C cost kz + A + C kz AC. Show that the uniform plane traveling wave of Eq.9 defined by Eq. (227).±.1ectromagnctic field. 0 + Are these valid for nonisotropie mcdia1 Do Eqs.0. For the field of Eqs. 2·4. 1 sin 2kt sin 2wt < _. possible c. Show that for any lossless nonmagnetic dielectric . 215. (221). (219) is II. 29._ cos 2h: cos ~ . (25) hold for inhomogeneous mediar 23.. . where .~ + to be true for al1ll. and.. fields? 28. show that the zdireeted wave impedances are Would you e.(. (2~28).) A _ C Sln 1 27.. and (220). This ill nol II. Show that the uniform plane traveling wave of Eq.\:pect Z~.. For the field of Eqs. vacuum. Derive the "wave equatioDs" for inhomogeneous media VX(f1VXE)+tiEO V X (f)IV X B) !H .t it does not Hlllisfy Eq. 210. .(El + jE1)C /h . . (225) can be expressed as E . (Z5).35). being righthanded if the ratio is +j and lefthandcd if thc ratio is j. t..... Show that E. . i$ the dielectric constant and k o '10. (26) but not Eq.  E. &peat {orEqa.v.Er[/b satisfies Eq." _ ~ . (2J 8) satisfy Eq. . (1.c are the intrinsic parameters of .. .Z. 26. (225) can be expressed as the sum of a righthand circularly polarized wave and a lefthand circularly polarized wave..v.
' ~ ~(I +~) .. iHi.Re (E) of Prob. and (d) copper.(1 +~) h~(I~) where Q is defined by Eq."Tl\ODOCl'JON TO WA YES 89 where B l and E t are real vecton lying in the %1J plane. Show that for metals Q «1 "'"'" lJI. (b) Plexit!u. + 2101. (179).7 VI 3. 216.rary complex E i 1m (E) and usc the rcsulta traccs out an elliptl6 in space.e" .5. Fig. Show that the tip of the arrow representing t for 80 arbit.01 X 107 Vi VI where I is the frequency in eycles per lICCond.10. Fig.yrene. " . 211." (I ~) lJI. . 211.52 X 107 VI 2. . Show that. (c::) Fernunic A. + iOC for (a) polyst. when all1088ell are of the maplctie type (" . and" is the conductivit.0). 112. Show th8t for nonm8gnetic dielectrics Q»I where Q is defined by Eq.k' . III.10.n. Show that (or nonmagnetic conducton . and 1000 megacycles. determine k . Fig.1 where Gl is the surface resistance. 2. _ lJI. II is the skin depth. ReJate E 1 and E z to A and B.8 X 107• 21S.\ 2HJ." ~ ". Derive the following formulu (Jl Gl (silver) Gl (copper) Gl (gold) (aluminum) Gt (braaa)  2. f. For the frequencies 10.61 X 10.12 X 10.26 X 10' 5.y.7 v1 3. . (179). [Hint: let E . 100.ik" and.(1 +11 I '(1" • "'1 . 110.
Find t. and calculate the percentAge of power reflected and tranamitted wben a plane wave is normally incident on a calm lake.~ 227. Show that the field . show that for the lwowire line of Table 23 R .. U8ing results of Prob..81.56.index of refraction wbere . (b) highdensity gla8ll.he conductora i. . .2. 26 for Il. 2. 228. Calculate the two polarizing angles (internal and external) and the critical angle for a plane interface between air and (a) water. to .he power per square meter dissipated in II. Gl4 + b 2. f< ... 2215. defined by Eq..0 Ilod 11 . circularly polarized standing wave in dissipative media.1. 225.a the value of a at the critical an&le? 2f:!.26. j. (261)J for the three caee8 of Prob.2. 2Gt and that for tbe couialline . (6) 1000 mcgacyeletl.. 228. What i... Show that '1 &ad C of a transmission line are related by ~" "'. Calculate the distance from the boundary in which the field ia attenuated to 1/_ (36. Given a uniform plane wave normally incident. upon a plMC airtodielcctric interface. 221.. 222. Take the index of refraction of water to be 9.. From Eq8.•• . is the dielectric constant of the dielectric (ILSSUmed nonmaloetic and 100000free). 3.'" O:rCZ • • when the dielectric II homogcneous. (270). Show that R of a tranamisaion line is appro:!imatelyequal to tbe doC n:cistance pet' unit length of hollow conduclora haviol thickneu .19.rc whcre ")' . (U)6) and (268).. (b) 1 megacycle. 8how that when R «wI.teS a plane dielectric. . U1»b " . Give a verbal deacription of 8 and :te.. and G «we a2~ R +GVLiC 2 fJ . f< . . toair interface. CoDJIider a parallelplate waveguide formed by conductor8 covering the planes y . 220. and (e) polystyrene.hat the standingwave ratio is SWR  V.8 per ceot) of ita "lue at the boundary. f<. and that for the parallelpiate line R.J .d D d»a »d 4»1 R .g. (akin depth) provided H is approximately conatant over each conductor and the radiu8 of curvature of t. 222.or + jfJ. Calculate the attenuation constant in the air ta .e lIurfae6 is 1 ampere per meter at (a) 60 cyclCtl.a large compared. abow t. copper aheet if the rma magnetic intensity at it.90 TlllEHAlWONlC ELECI'aOJolAONETIC FIELDS 218.b. Verify Eql. Make a sketch similar to Fig. 2.. Suppose a uniform plane wave in I!Io dielectric just gra.
). Show that the power transmitted per unit width (z direction) of the parallel· plate waveguide of Prob.I' >o}1 . i Compare this with a obtained by using the results of Probs.(J)' for the TM. 228 ia actually a. .28. modea (n P! 0). modes. COST er· n. mode. 228 ill Pfor the TE.ies per unit. modes are Show that Eqs.oW V. length are w. and for the TM.. 231. modes (n >L 0). for tbe TE. modes.IN'I'ItoDUCTION TO WAVES 91 defines a set ot TE" modes and the field H. •. 229. defines a aet ot TM. Show that the timeaverage velocity of propagation of energy down a rectangular waveguide ill g. modes.. ahow that the timeaverage electric and magnetic energ.. and TM. Show that.  ~  lEt/lab Can this equality of W. For the parallelplate waveguide of Prob. _k' I. and '\9. . For thc TEfl reclangWar waveguide mode.H. for thia mode the atteouation due to con· ductor loaaea is a.. (283) to (286) apply to the parallelplate waveguide modes. . . I  G')' . 232. Show that the TM. 1.. n in both CMClI. and bl:.b.. TEM mode. 2. Show that the cutoff trequeneies of the TE. (162)1 2SS. where y _ \l(n.2hV.. be predicted from Eq... n {I. Bhow thaL Lbe attenuation due to conductor 1088ca is for the TE. 2·30. 1 p. 2·26 and 224..2. mode of the parallelplate waveguide 1\8 defined in Prob.
This result is valid for any TE mode. Show that these are Show that P ¢: V /. 2·36. For the TED. PI for z < 0 and 'z./lll' and show tbat R _ r'<Rjbe(b' + e Z) + 211(b' 32(b' + c')' + c.ZOI Zn + ZII > 0. fl. 238.8 reduces to 1<1 L _ ~'I + 'I U Compare this to Eq. Let a rectangular waveguide have 8. voltage Vaa fE' dl acro88 the center of the guide and 8. 1 as the total z.t < O.IU Jollfl) where /<1 is the cutoff frequcncy . respectively. Calculate the Q in each ClLllC.directcd current in the guide wall :r _ O. For the rectangular cavity of Fig. diseontinuity in dielectric at t _ 0. 237. 219. III for z > 0. Take a parallelplate waveguide with 'I. Why? Define a characteristic impedance ZYI .V / I and show that it is proportional to Zo of Table 24.. Show that v EtII iii~/IVII Defino a mode conductance G 88 G  and show that G _ lR(bc(b Z el) 2II{b' 2'l'a' (b' e') + + + + e') I Define a mode resistance R &8 R . 236.1 I PI ) /" Joll(s. p.)1 . Show that there ill no rollected wavo for the TE ol mode in Prob. that is. 239.iii. Thue rCllu!ta are valid for any TM mode. Show that the reflection and transmission coefficients for a TEll wave incident from z < 0 arc r _ where ZOI Z . thi. Show that there is no reflected wave for a TM mode incident from z < 0 when For nonmagnetic dielectrics. (260). 235 when / ~  11(/01. Note that we cannot have a reflectionless interface when both dieJectrics are nonmagnetic.92 TIME~HARMONIC ELECTROMAGNETIC FIELDS 234. rectangular waveguide mode. . and Zot are the characteristio impedances z < 0 and z These results are valid for Bny waveguide mode. for z > O. define 8. current. Design a squarebase cavity with height onehalf the width of the base to reaonate at 1000 meg&cyc!ea (II) when it is air!illed and (b) whcn it is polystyrenefilled. III Cor z < 0 and fl. dcfine a voltage V lUI that between midpoints of the top and bottom walls and a current I &8 the total :H:Iirected cur· tent in tbe side walls.
. 24.+ ik + I) sin 8 .0 and show that Ira! A • _e'. 2.yr1+ a l yr + a! l 2ra sin 8 cos . In terms of the tabulated functioDs Si(.• .. 226. 411" r r! Show that the radiation resistance of the small loop referred to I is R''l~e:r 248...r loop of current. 2~21 and the current loop of Fig. z . Consider the small loop of constant current I that the magnctic vector potential is all A. 226 to exist simultaneously.. A circula. 226. I y 24. .! .. a .p' ! _ exp (ik .04.. Show that the radiation field is everywhere circularly polarized if II .}o . (2123).. .t) ..ISis called the magnetic moment of the loop.. Derive Eqa.e'~' (ik 'L. (irk + ') rl SID 8 .r 411" r r! l k 'lIS E• e'.2..t dx ._0 f" 0 !c08q/d.1 is IS H.US 244..• (k! . Show 240.1.t Oi(x) __ r~dx ]. 24. Show that the field of the small current loop of Prob. rl li.. e  I) +l cos 8 r IS e'l< ( k ."') 2ra sin 8 cos .:.p' Expand I in a Maclaurin scrica about a .INTRODUCTION TO WAVES 93 shown in Fig.i ) SID . FlO. Consider the currcnt element of Fig. ( '" ain . The quantity hal . where A~ I __ Ia 411" .
SikL) coskL( C + log k~ + Ci2kL .. is H. If the linear a..2L/". the current will assume the form regardless of the position of the feed 8.I itr ''C. is Euler's constant. c 2. . . resistance for a lossfree antenna with feed point at :: .0. and Ci is as defined in Prob.. is Show that the input R(._.5772. a .Ci(2nr)J where 11. 244.2L/>. . _ . . current nulL Such &n antenna is said to be of resonant length. . C .". . (2129) can be expressed as Rr  i [ C + Jog kL  CikL + sin kL(~8i2kL +H . For an antenna of resonant length (Prob.0 (the halfwave dipole) and show that R. .L 21fT '':c. ) COB n odd '" even 8 where n . show that the radiation resistance referred to I .his result to L .8 long as it is not neM 8.>"/2.5772 . 246.ntenna. Show that the radiation field of the antenna is E.4: Ie + log 2n.. of Fig. . is an integer.0.2CikL)] where C . 223 is an integral number of baUwavelcngthll long.~ e/tr .a).73 ohms.. 246. (7 0) 8 ..94 TIMEHARMONIC ELECTROMAGNETIC FIELDS show that Eq. .. 245). . (n.'"< 8m 810 COB I SID 2"c088 E. SIn 2r(a R~ + n/4} Specialize t..
We have purposely omitted superscripts on J and M because their interpretations vary from problem to problem. and so all. 9' . (a) Current Bouree.l interpretation. It then follows from the conservation of charge that the current in the leads is equal to the impressed current. The Source Concept. Circuit sources in terms nf impressed currents. J might represent a conduction current that we wish to keep separate from the 1]E term. 31. Eq. That it has the characteristics of the current source of circuit theory can be demonstrated as follows. In stln another problem. This is shown in Fig. We make the usual circuit assumption that the displacement current through the surrounding medium is negligible. M might represent a magnetic polarization current that we wish to keep separate from the ~H term. For our first illustration. (166). for power. independent of the load. We can think of J and M as If ma. In one problem. In terms of field concepts it can be pictured as a short filament of impressed electric current in series with a perfectly conducting wire. In another problem.CHAPTER 3 SOME THEOREMS AND CONCEPTS 31. Jet us show how to represent 1/ circuit sources II in terms of the "field sources" J and M. they might represent actual sources. The current source of circuit theory is defined as one whose current is independent of the load. reduces to FlO.the· matical sources/' regardless of their physica. The complex field equations for linear media arc (31) v X E . Cb) voltage I I' Ca) I + V • + V K'~ D (b) source.£H+M vxHgE+J where J and M are sources in the most general sense. 3la. in which case we would call them impressed currents. The field formula.
96 TllLEBAJUlONIC . We can use the circuit sources in field problems when the source and input region are of "circuit dimensions. we approximate the current in the wire. To excite the antenna. (166). we postulate that this current should produce only the TEol waveguide mode. In terms of field concepts it can be pictured as a small loop of impressed magnetic current encircling a perfectly conducting wire. removal of the impressed current leaves an open circuit. K' "" . consider the linear antenna of Fig. of dimensions small compared to a wavelength. over the cross section of a rectangular waveguide. we can apply the current source of Fig. Furthermore. . The field formula for power. 223. We then use this current.V. that is. We have only electric currents. 114. we can apply the voltage source of Fig. Eq.tion K "'" :J'E· d1 to a path coincident with the wire and closing across the terminals. aIb. plus the current source across the gap. we neglect displacement current and apply the field equa. We shall find much use for the concept of current sheets. we can place a current source (a short filament of electric current) across the gap. . To show that it has the characteristics of the voltage source of circuit theory. Given a pair of terminals close together. An exact solution to the problem involves a determination of the resulting current in the wire. 3lb. This is difficult to do. p. 8 short filament of impressed electric currcnt.VI' The "internal impedance" of the source is infinite. considered in Sec.. 32. Given a conductor of 5Inall cross section. and therefore the terminal voltage. The impressed current. a small loop of impressed magnetic current. since a removal of the impressed current leaves a short circuit. is independent of load. The internal impedance of the source is zero. 80 the line integral is merely the terminal voltage. since 8. Instead. As an example of the use of a circuit source. as shown in Fig. The voltage &aurce of circuit theory is defined as onc whose voltage is independent of the load. ala. that is. drawing on qualitative and experimental knowledge. =  III H· MfdT  K't H·dl = VI which is tbe usual circuit formula. in the potential integral formula to give us an approximation to the field. suppose we have a J." that is. As an example. The E is zero in the wire. that is.£LEcraoMAGNETIC nELDS the circuit Cormula (artius source. The geometry of the physical antenna is two sections of wire separated by a small gap at the input. This is illustrated by Fig. reduces in this case to p. hence I" III E'J"dT  I E·d1. which causes a current in the antenna wire.
so that n = U" and obtain uAH."1/ Zo sm b = A .= . H. I I . and E£ from above reduces these equations to u~ A+B.COS(7nf b where B is the mode amplitude of the z traveling wave. ry .J_o = J. (186) must be satisfied..+  H. Eqs.  u.E£l_o =0 Substitution for H. . z <0' B.J. + I 1. Table 24. A sheet of current in a rectangular waveguide.. sin 7: . The z traveling wave is of the same form with {J replaced by {J and Zo by Zoo Thus. (E~+ .SOME THEOREMS AND CONCEI"I"a 97 x /1 /L __ // / • J. Take the (1) side to be z > 0. .•• H •..= .. which propagates outward from the current sheet."1/. J. 32. H~+ = ~ cos 1r: ei6• f where the constant A specifies the mode amplitude.L __ /: / z /' FlO.B c: 0 (32) Let J.Zosmb(7~' Bf. E£ = B sin i: eJt'. we have the wave Abstracting from E£+ = A sin 1r: ci4~ z>O H"+=~sin7rY_i6~ Zo b <> . At z = 0.
Duality.sin ..' . (32) exists over the guide cross section z = 0. The first two rows of Table 31 givo the field equations in each case.ter.. obtained by systematically interohanging symbols. A duality of importance to us is that between a problem for which all sources are of the electric type and a problem for which all sources are of the magnetic type. Eqs. are duals of eaoh other.. DUAL . lid.IAONETIC SoURCES EXIST (I) Electric sourcell (2) Magnetic sources vXHfJE+J v X E . Note that the field equationa. Note that our approach in this problem was to assume the field and find the current. The last two formulas of column (1) were derived in Se<:."'" r'l d.'\lS IN Wmell (I) ONJ.v X F F == I 4r If! v""'" ... Quantities occupying the same position in dual equations are oalled dual quantitie8. but this is not of concern at present.IH +M v XE . Two equations of the same mathematical form are called dual cqualion8. 29 for homogeneous space. Thus. If the equations describing two different phenomena are of the same mathematical form. if the current of Eq. then JoZ.e"" z>o E.fJE E ... •y • . ~ 2 b (33) • JoZ o sln 'TO'!J ~ z<O 1)e' 1r It would admittedly be difficult to obtain the current of EQ.. We shall learn how to treat more practical problema la. The reader should oheok for himseU that a replacement of the symbols of TAllLE 31. The formal recognition of this is called the concept of duality. 32. r r' Ii! Ir M.y ELECTRIC SOURCES ExiST MW (2) ONLY \I.. (3~2) in practice. A systematic interchange of symbols ohanges the first equation into the second. and viceversa. solutions to them will take the same mathematical form.98 TIMEHARMONIC ELECTROMAGNETIC FIELDS The preceding equations then have the solution A = B = JoZo/2. The corresponding equations for the magnetic source case are evidently the last two formulas of column (2). The particular interchange of symbols is summarized by Table 32.EQU"TIO~S ron PaOBLE. (31). This we shall find to be a very powerful concept. . fH HvXA I A 4r  v X H .
lfS IN WHICH (I) ONLY ELECTRIC SoURCES ExIST. AND (2) ONLY MAONETtC SoURCES EXIST (I) Elearic IOUrcu (2) AIagnetic =. thus cutting the mathematical labor in half. so the total field can be considered as the sum of two parts. The concept of duality is important for several reasons. in analogy to A. of coursc. The concept of duality is based wholly on the mathematical symmetry of equations. For example.SOME THEOREMS AND CONCEPTS 99 TAIlLE 32. It is often convenient to divide a. the picture of elect.M) (34) . Such a picture can serve as a guide to the mathematical development but cannot. interchange symbols. column (1) of Table 32 by those of column (2) in the equations of column (1) of Table 31 results in the equations of column (2). we can picture magnetic charge in motion as giving rise to magnetic current. (31). or E ~ . The quantity F of these tables is called an electric vector potential. and obtain the solution to another type of problem. To be explicit. II magnetic vector potential.= H E H E M J A F 9 k • .ric charge in motion giving rise to an electric current can also be used for magnetic case. It is an aid to remembering equations. The field equations. one produced by J and the other by M.V X A + r'(v X V X F . serve to argue for the existence of magnetic charges in nature. suppose we have both electric and magnetic sources in a homogeneous medium of infinite extent. Eqs. We can also use a physical or intuitive picture that applies to one type of problem and carry it over to the dual problem. • 9 k 1/.V X E" = iH" + M We have the solution for each of these partial problems in Table 31. DUAL QUANTITIES FOR PR08LE. are linear.J) II .V X F + y'(v X V X A . That is. since almost half of them are duals of other equations. let where and E = E' + E" H = H' + H" V X H' = yE' + J V X E' = zH' V X H" = yE" . For example. The complete solution is therefore just thc superposition of the two partial solutions. single problem into dual parts. It shows us how to take the solution to one type of problem.
oJ~IS. We might suspect this from the circuit source representations of Fig. A zdirected magnetic current dipole of moment Kl at the coordinate origin is the dual problem to the electric current dipole (Fig.. 33..·. 221). These two sourees radiate the same field if Ki ../.~.100 where TLMEHAnMONIC ELECTROMAGNETIC FIELDS Iff If _ ell dr' 1 rrf M(r'). let us consider the fields explicitly. 33. (a) Magnetic currcnt element. The above formulas arc meant to include by implication sheets and filaments of currents..' 1 A(r) = 411" J(r'). M. sm(J The small loop of electric current is considered in Probs. For example.::>lS (oj (bj FlO. according to Table 32.. we have the electric intensity given by E. Uniqueness. Thus.(k' L sin 'k) it • _ r r2 (J A comparison of the above two equations shows that they are identical ir Kl = jwp. 212. 31.'·' F(r) ~ 4rJJ [ r "I d. r r I). An interchange of symbols. However. S encloses linear matter !lnd sources]. Abstracting from Prob. the electric intensity is E. effect of an clement of magnctic currcnt can be realized in practice by a loop of electric current. ratber than rely on this argument.'1. J (35) We thus have the formal solution (or any problem consisting of electric and magnetic currents in an unbounded homogeneous region. 33. 34. = ' 41r IS e. 241 and 242 and is pictured in Fig. 226. (2113) will give us the field of the magnetic current element. (b) electric current loop.~ Kl __ eib (jk +2 4. FtO.IS (36) This equality is illustrated by Fig. A solution is said to be unique whcn it is the only onc possiblc among a given class of solutions. . in Eqs.. It is important to have n C:. It is instructive to show that an infinitesimal dipole of magnetic current is indistinguishable from an infiniteBimalloop of electric current.
) We form the difference field oE. they tell us what information is needed to obtain the solution. H" and Eb. (2) The field is unique among a class E.1m (Ul!IEI'J dT ~ 0 (38) For dissipative media. it is eomforting to know that a solution is the only solution.tisfied. First of all. ds + III (11!T/1' + UO!IEI'l dT Wheneve' effi (IE X mO) . (These can be thought of as the fields when the sources Qutside of S are different. then Eqs. Hb. (3·1). as well as the more usual reverse procedure. E". The conditions (or uniqueness are those for which oE = oH = 0 everywhere within S. Finally. H having n X E specified over part of Sand n X H specified over the rest of S.0 ovcr S. are as follows. (3) The field is unique among a class E. Secondly. (38) nrc satisfied only if oE = oH = 0 everywhere within S. as suggested by Fig. oE. however slight. We now apply Eq. These possibilities ean be summarized by the following uniqueness theorem.HH v X IH ~ tPE I within S Thus. Eqs. (1) The field is unique among a class E. 34. (154) to the difference field and obtain effi (m X mO) . we obtain v X IE . ds . for then n X. the volume integrnl must also vanish.SOM. (37) is true. Consider two possible solutions. III [Re (1)1!Ii[' + Re (g)laEI'J dT ~ 0 III [1m (zll!HI' . If we assume somc dissipation everywhere.= 0 over S. oR according to oE=E"_Eb Subtracting Eqs. if Eq. uniqueness theorems establish conditions for a onetoone correspondence of a field to its sources. (37) is sa. A field in a lossy region is uniquely . Any field within S must satisfy the complex field equations. H having n X H specified on S.sons. Re (z) and Re (t/) are always positivc. Suppose we have a set of sources J and M acting in a region of linear matter bounded by the surface S. Cor then n X oR = 0 over S.E THEOREMS AND CONCEPTS 101 precise theorems on uniqueness for several rea. the difference field satisfies the source·free field equations within S. (31) for the a field from those for the b field. for then E" = Eb and H" = Hb. H having n X E spccified on S. This allows us to calculate the sourees from a field. and thereCore uniqueness is obtained in lossy regions. Some of the more important cases fOr which Eq. then 0 (37) Thus.
so that all space is included.\r&r . " . Our solution at large r is Eq. not inward. or j l l " rsln H • • = e' . The difference field in this case is 6H.\T It can be shown that the inwardtraveling wave .o . we threw out this second solution by reasoning that waves must travel outward from the source. consider the current element of Fig. In Sec.' 1 .Hl"" j :~coskrSill 6 'E II . can be thought of as limiting cases of volume distributions and therefore are included by implication. If the sources are of finite extent. We have explicitly considered only volume distributions of sources and closed surfaces in OUf development. 29. To obtain uniqueness in this case. Note that our uniqueness proof breaks down for dissipationlcss media. The lossfree case call be treated as the limit of the lossy case as dissipation vanishes. we cO'Mider the field in a dissipationleS8 medium to be the limit of the corresponding field in a lossy medium as the dissipation goes w zero. = H.' H "' is also a solution to the equations at large r. 221. sm H . any solution 8ati8fying Eq. Surfaces of infinite extent can be thought of as closed at infinity and can be included by appropriate limiting procedures. 8 ' E' = .j l l " . H satisfying Eq.=2. 8 2. Singular sources. Let us now consider these two solutions in the light of the uniqueness theorem. (39) fnWlt be identically equal W the potential integral 80lution. such as current sheets and current filaments.\r81n krSIO 8 . To illustrate the above concepts. (34) and (35) vanishes exponentially as eJ:"r. "'" E . (39). Of particular importance is the case for which the bounding surface is a sphere of radius r + llQ. .102 TIMEHARMONIC ELECTROMAGNETIC FIELDS specified by the sources within the region plus the tangential components of E over the boundary.0 . . According to our uniqueness proof this must be the only solution for a class E. or the former over part of the boundary and the laUer over the rest of the boundary. or the tangential components of H over the boundary. l' + llQ. (2114). the vector potential solution of Eqs. Let this be the a solution of our uniqueness proof.E' "'" ". given sources of finite extent in an unbounded lossy region. but the results are much more general than this. Thus. We therefore have (39) for this solution (in lossy media). .
(2114). The rectangular waveguide of Sec. the solution to this problem is also the solution for a current element adjacent to a plane conductor. An element of source plus an "image" element of source. r. if a. The procedure also applies to magnetic conductors in a dual sense. vanishes. (37) can be satisfied without obtaining uniqueness of the solution. 35. only the a solution vanishes as r + 00. 34. The radiation field of a single clement is given by Eq. sin kr and cor kr have no zeros T > 0. The necessary orientation and excitation of image elements is summarized by Fig. Let us determine the radiation field. This must produce the same field above the ground plane as do the two elements of Fig. . 36b. 36a. As an example of image theory. (37) cannot be satisfied for any r. radiating in free space. 27 is an example of a boundaryvalue problem. respectively. conducting sphere is adjacent to the plane conductor in the original problem. In other words. 36b is the superposition r • j II H. or oE. The radius vector from each current element is then parallel to that from the origin and given by ro=rdcosO) r. It is therefore the dp.sm 8 eil:"). Image Theory. Problems for which the field in a given region of space is determined from a knowledge of the field over the boundary of the region are called boundary~value problema. The procedure is known as image theory. The application of image theory in ac fields is much more restricted than in dc fields. cos(kd cos 8) sin 0 (310) mary of image theory.SOME THEOREMS AND CONCEPTS 103 In dissipationless media (k real). we must maintain symmetry in the image problem. we can pick a sphere r = constant such that either oH. However. produce zero tangential components of E over the plane bisecting the line joining the two elements. = 2A To ~ (efl: + . as shown in Fig. consider a current element normal to the ground (conducting) plane. and Eq..sired solution in lossfree media. A sum "" j Il ~r rile. Matter also can be imaged. We shall now consider a class of boundaryvalue problems for which the boundary surface is a perfectly conducting plane. Thus. For example. The boundary conditions at a perfect electric conductor are vanishing tangential components of E. then two conducting spheres at image points are necessary in the image problem. In this case. 35. Eq. = r + d cos 0 r»d _II where subscripts Q and i refer to original and image elements. in lossy media. so the radiation field of the two elements of Fig. FlO. It is exact only when the plane conductor is perfect. According to uniqueness concepts.
As led + 0. . n (a) n (6) (a) Original problem. a gain of more than four times that of the isolated element (1. 2 = _'I_=C=OS'"2ki</. 3. This is g = 3 at /cd = 0.chieved. = 7JH~. and E. I Ill' [1:3  cos 2kd (2kd)' + sin 2kd] (2kd)' (3l1) the power radiated is equal to that radiated by an isolated element [Eq. for which 9 = 6. Thus. the power radiated is double that radiated by an isolated clement.B: ds . The total power radiated by the system is ~... According to image theory.t 00. 36a represents the antenna system of a short dipole antenna adjacent to a ground plane. r image problem. The gain of the antenna system over an omnidirectional radiator. (310) and integrating.57. and g 6 as kd + 00. according to Eq.:kd 2 J 3 (2kd)' (2kd)' =0 (312) along the ground plane. (2130). we have Substi is'. 36a above the ground plane.+'''i=n<i.lt r t sin () dO > 0.phere hemi_ = 2'l1'"7J fo· n 11i. Figure 37 shows the radiation field patterns for the cases . The maximum gain occurs at Jed 0= 2.{t A current clement adjacent to a ground plane.5) can be i. (b) FlO. = JJ B.1rT 211I H40l t is'.88.104 TIMEDARMONIC ELECTROMAGNETIC FIELDS z e r Z e II '" r. is g 4. (2116)]. this must also be the solution to Fig.2""i: As kd . r + 00 where integration is over the large hemisphere z tuting from Eq. The problem of Fig.
3&. clement of Fig.SOME THEOREMS AND CONCEPTS 105 FlO. Problema involving multiple images.ment. (0) Current ele. (b) current element in a conducting wedge. in a conducting tube. 38. • II Fro. 3·7. . Radiation field patterns for the current.
This is shown in Fig. (r. E. (maximum gain). (186). 3·6b produces the same field above the plane z = 0 as do the currents on the conductor of Fig. prod1J~ the 5&ffie field c. These currents are therefore J. 38b). A simple application of the equivalence principle is illustrated by Fig. H are the original fields over S.xlernal to S lill do the . 39. 35. M.106 TIMEHARMONIC ELEC'rnOMAGNETIC FIELDS d = 0 (element at the gound plane surface) and d = 0. H external to S and zero internal to S. with free space everywhere. From the uniqueness theorem. In the case of a conducting wedge (Fig. (34) and (35). The Equivalence Principle.459>. we know that the field so calculated will be the originally postulated field. 38. For example. a finite set of images results. We can set up a problem equivalent to the original problem external to S as follows. When we are interested in the field in a given region of space. Equivalent sources will serve as well.. Two such cases are illustrated by Fig. 39a represent a source (perhaps a transmitter and antenna) internal to S and free space external to S. In the case of a conducting tube (Fig. 36a.  n X H M. on S according to Eqs. the image current element of Fig. Let Fig. we do not need to know the actual sources. Since the currents act in unbounded free space. To support this field. that is. there must exist surface currents J. and the null field internal to S. we can determine the field from them by Eqs. 3·9. The final result of this procedure is a formula for E and R everywhere external to S in terms of the tangential componentS'" of E and H on S. Let the original field exist external to 8. Image theory can be used for conducting wedges when the wedge angle is 180o/n (n an integer). Many source distributions outside a given region can produce the samo field inside the region. Image theory also can be applied in certain problems involving more than one conducting plane. 3Sa). 39b. an infinite lattice of images is needed. The equivalent currents original eources. Two Rources producing the same field within a region of space are said to be equivalent within that region. ~ E Xn (313) where n points outward and E.) Flo.
M. II' is the field of the b problem. Internal to S. there must be surface currents J.. We have two original problems consisting of currents in linear media. E'. Ha is the field of the a problem and E'. 310.(Eo .. In this case the necessary ~urface currents are the negative of Eqs. External to S. Note also that we cannot use Eqs.he cquivalcnce prhu::iple... 31Oc..H') M. as shown in Fig.J. these are given by J.S AND CONCEPTS 107 E'H' . Note that in each case we must keep the original sources and media.H' I (0) (b) \.. 31Od.. \ / !n . To suP"' port this field. ___ I\ 1. and sources remain the same as in the b problem. (34) and (35) to . . We were overly restrictive in specifying the null field internal to S in the preceding example. 310. and M. (e) equivalent.. We can set up a problem equivalent to a external to S and equivalent to b internal to S as follows.. as shown in Fig. (d) equivalent to b exwmal to Saud to a internal to S. and sources remain the Barne as in the a problem. to a external to S and to b internal to S.Jn (. Flo. /~__.E') X D (314) where Ea. we specify that the field. giving us infinitely many equivalent currents as far as the external region is eoncerned. We can also set up a problem equivalent to b external to S and to (J internal to S in I\n Analogous manner. This general formulation of the equivalence prineiple is represented by Fig. = D X (Ho .. . (186). we specify that the field. (b) original b problem. medium. This equivalent problem is shown in Fig... 31Oa and b. According to Eqs.. ~ S' (0) J. medium. (a) Original 4 problem. on S.. in the region for which we keep the field. " \ \ E"H". ' / s:.SOME TIlEOREJd.. (314). A general formulation of t. Any other field would serve as well.
problems can be found in terms of only magnetie eurrents (tangential E) or only electric currents (tangenlial H)... . Since the tangential components of E are zero on the conductor (just behind M. 311a and bj so according to our uniqueness theorem the field outside of S must be the same in both cases.. We can derive the alternative equivalent problem of Fig.(a) \ Zero field Electric conductor (0) ~ D " J _/ s .. The field external to S is the Mme in (a). (186) that MI = E X n (315) We now have the same tangential components of E over S in both Fig.. From uniqueness concepta.. So far. We shall now show that. perfect electric conductor. we know that the tangential components of only E or H arc needed to de~rmine the field. note that the restricted form of the equivalence principle (Fig. We set up the equivalent problem of Fig. We then find that the elcctric current sheet ]. For this we need the perfect magnetic conductor.. that is..EXn Flo.. we have used the tangential components of both E and H in setting up our equivalent problems. and on top of this we place a sheet of magnetic current MI' External to S we specify the same field and medium as in the original problem.. 311b as foUowa. (e) eledric current backed by a magnetic conductor.ng 8 produces the same field external to S 88 do the original sources. and (e)... 311c in an analogous manner. By now.. .. determine the field of the currents unless the equivalent currents radiate into an unbounded homogeneous region. equivalent. the general philosophy of the equivalence principle should be .. Ca) Original problem.). (b) magnetic current backed by all electric conductor..108 E.B D D E. it follows from Eqs. ! E.B { I Sources \/ S . _. Finally. D X H (316) over a perfect magnetic conductor coveri. a boundary of zero tangential components of H. 39) is the special case of the general form for which all a sources and matter lie inside S and all b sources are zero. as shown in Fig. Over S we place 0.:ftI (e) M. 3110. and equal to the original field components just in front of M I .B / TIMEHARMONIC ELECTROMAONETIC FIEL06 . (b).. Consider a problem for which all sources lie within S. 311..
312a.l to S is zero.311b. A circuit theory analogue to thQ equivalence principle. in the problem of Fig. in the presence of the electric conductor. the partial fields produced by J. the field is produced by the magnetic currents alone. Perhaps it would help us to understand the equivalence principle if we considered the analogous concept in circuit theory. It is based upon the onetoone correspondence bet\"een fields and sources when uniqueness conditions are met.rent. Alternatively. we could back the equivalent currents of Fig. 39b.nce connected. alone will change external to 8. We could just as well introduce a perfect electric conductor to back the current sheets of Fig. We have previously assumed that free space existed within S. 312. When matter is placed within S in Fig.nged. 39b with a. we can determine all sources. (c) source impedance replaced by a short circuit. the field interna. The original source is switched off. 39b. as shown in Fig. equal to the terminal current in the original problem. 38) that an electric current just in front of an electric current conductor produces no field. A voltage I ~ Source tv Passive network Passive network Source Impedance (b) Passive network (0) I t"''_~_"'_tw_O_'_k_. so that the potential integral solution could be applied. . which is Fig. (b) equivlilent sources. perfect magnetic conductor and obtain the equivalent problem of Fig. Consider a source (active network) connected to a passive network. (0) Original problem. is placed across the terminals. as follows. A current source 1. (We can think of the conductor as shorting out the current. If we specify the field and matter everywhere in space. We derived our various equivalences in this manner. For example. 39b.) Therefore.J Passive (d) FIG. but the total field must remain uncha. alone and M. 3He. It can be shown by reciprocity (Sec. It therefore makes no difference what matter is within S as far as the field external to S is concerned. (d) source impedance replaced by an open circuit. Considerable physical interpretation can be given to the equivalence principle.SOME THEOREMS AND CONCEPTS 109 appa. We can set up a problem equivalent to this as (ar as the passive network is concerned. leaving the source impeda.
A combina. 313. let the original problem consist of matter and sources z < 0. It is evident from the usual circuit concepts that there is no excitation of the source impedance from these equivalent sources. " 8 'C C ~ u ~ M. 3lIb. 312b.] 11 n (b) (e) n FIo. 3lIb yields the equivalent problem of Fig. This consists of the magnetic currents of Eq. S8. we may replace it by an arbitrary impedance without affecting the excitation of the passive network. whereas the excitation of the passive network is unchanged. This leaves only the current source exciting the network. as shown in Fig. Fig. as illustrated by Fig. This is analogous to the field problem of Fig. and free space z > 0. 39b.H Zero field z~o I I E. (317).0 E.. To illustrate. .2El<n i. as shown in Fig. produces the same excitation of the passive network as does the original source.H fie d I I I I ~n (a) . (315) adjacent to an infinite zo E. 313b. .. 312b is the circuit analogue to Fig. This is analogous to the field problem of Fig. Since there is no excitation of the source impedance in Fig. = El<n M. 312<1. This is illustrated by Fig. 312b.uo TIMEHARMON'IC ELECTROMAGNETIC FIELDS source V. 3·13a. In particular. Thus. the voltage source alone. Illustration 01 the otcps uoed to establish Eq. is placed in series with the interconnection.H matter @ I I I Ima~e E. Thus.H Sources and • . This is analogous to the arbitrary placement of matter within S in the field equivalence of Fig. Now consider the source impedance of Fig. 39b. An application of the equivalence concepts of Fig. let the source impedance be replaced by a short circuit. 312b replaced by an open circuit. 311e. 312c. Fields in Halfspace.tion of the equivalence principle and image theory can be used to obtain solutions to boundaryvalue problems for which the field in halfspace is to be determined from its tangential components over the bounding plane. equal to the terminal voltage in the original problem. This shortcircuits the current source and leaves only the voltage source exciting the network (recall circuit theory superposition).
313c. according to Fig. (317). as pictured in Fig. 313b. so the abovc identity must also be valid for E replaced by H. 35.. 314a). which is identical to Eq. z . AB an example.V X II 2'. radiating into unbounded space produce the same field z > 0 as do the original sources. A coaxial line opening onto a grcund plane. This cao be summarized mathematically by E(r) . The im3gCS are equ31 in magnitude to. the M. which is of no interest to us. (23).1 E(r') X ds' (317) This is a mathematical identity valid for a. the field must be the same as that produced by Fig. The field of Fig. suppose we have a coaxial transmission line opening into a ground plane (Fig. and essentially coincident with. 314b. (a) Original problemj (b) ground plane. According to the above discussion.son THEOREMS AND CONCEPTS z . for tangential E is ECro over the ground plane. 314. (24). We now image the magnetic currents in the ground plane. (23). They produce an image field z < 0."'. equivalent problem. of Fig. We can show this by reasoning dual to that used to establish Eq. • 111 • ~4y x y x p FIG.ny field E satisfying Eq. Thus. the magnetic currents 2M. that is E•  p log (bfa) v . Note that M. 313c is then calculated according to Eqs. Let us asume that the field over the aperture is the transmissionline mode of the coax. (34) and (35) with A = O. exists only over the aperture (coax opening). The above result is particul:JJ'ly useful for problems involving apertures in conducting ground planes. The H field satisfies Eq.
The gain of the antenna system is g = 3. 37. rent in Fig.a') . Hence. (311) with kd = 0 and II given by Eqs. acts as an electric dipole (dual to Fig. we have the radiation field given by . (36). or _V(b' .J. Note also t. 3& with kd = O.hat our answers are referred to a volt.0 curve of Fig. the radiation conductance is 4r' [ b' .a') 2 ~.. The power radiated is Eq.112 TIMEH.V(b' .:·rETIC FIELDS where V is the line voltage.nas. the radiation field patted is the d . X'iog (bla) I I I' I' (321) Note that the power radiated varies inversely as )._ Thus. characteristic of aperture antennas.a')V 3. . dp.a' ]' G.KS (319) We have now reduced the problem to that of Fig. = 2r. From Eq. The total moment of tbe source is then (318) The equivalent electric current clement must satisfy the equation dual to Eq.. 2X log (bla) :I = 4r r'(b' . = flJI.. 314b is M To this approximation. = 2Ar log (bfa) e. (318) and (319). the magnetic cur• = V p log (bla) This is a loop of magnetic current which. if b «A.. 33). (310) and the above equalitics. Visualize this current as a continuous distribution of magnetic current filaments of strength dK = M.AJUlONJC ELECI'aOMAG. or II = jw. For aperture antennas we define a radiation conductance according to (322) where V is an arbitrary reference voltage.4.j" X' log (bla) (323) . (320) H. This is in contrast to answers referred to current for wire anten.. 314 it is logical to pick this V to be the coaxial V at the aperture. In the coaxial radiator of Fig.age." 510 8 and E.
SOME THEOREMS AND CONCEPTS
113
n
t
E_E'+gE' n
Source Obstacle Obstacle
;/
/ J .... Hi Xn 
'c~ M.
(a)
Flo. 315.lilustration of the induction thoorcm. equivalent.
(b)
nxE'
(a) Original problem; (b) induction
For the usual coaxial line, Gr is small, and the coaxial line sees nearly an open circuit. As a and b are made larger, the radiation becomes more pronounced, but our formulas must then be modified.! 37. The Induction Theorem. We now consider a theorem closely related in concept to the equivalence principle. Consider a problem in which a set of sources are radiating in the presence of an obstacle (material body). This is illustrated by Fig. 315a. Define the incident field E', Hi as the field of the sources with the obstacle absent. Define the scattered field E, H' as the difference between the field with the obstacle present (E, H) and the incident field, that is,
E'  E  E' H'  H  H' (324)
This scattered field can be thought of as the field produced by the cur· rents (conduction and polarization) on the obstacle. External to the obstacle, both E, Hand E;, Hi have the same sources. The scattered field E', H' is therefore a sourcefree field external to the obstacle. We now construct a second problem as follows. Retain the obstacle, and postulate that the original field E, H exists internal to it and that the scattered field E', H exists external to it. Both. these fields are sourcefree in their respective regions. To support these fields, there must be surface currents on S according to Eqs. (186), that is,
J. 
n X (H'  H)
M.(E·E)xn
where n points outward from S.
According to Eqs. (324), these reduce to
M.  n X E' (325)
J. 
H' X n
It follows from the uniqueness theorem that these currents, radiating in the presence of the obstacle, produce the postulated field (E, H internal to S, and E', H' external to S). This is the illdmlion theorem, illustrated by Fig. 3l5b. It is instructive to compare the induction theorem with the equival H. Levine snd C. B. Papas, Theory of the Circular Diffraction Antenna, J. Appl. Ph"., voL 22, no. 1, pp. 2H3, January, 1951.
114
TIMEHARMONIC ELECTROMAGNETIC FIELDS
lence theorem. The latter postulates E, H internal to S and zero fieM external to S, which must. be supported by currents
J. 
H X
0
M. '"'" n X E
on S. These currents can be considered as radiating into an unbounded medium baving constitutive parameters equal to those of the obstacle. Thus, we can use Eqs. (34) and (35) to calculate tbe field of the above currents. However, we do not know J. and M. until we know E, H on S, that is, until we have the solution to the problem of Fig. 315a. We can, however, approximate J. and M. and from tbese calculate an approximation to E, H within S. In contrast to the above, the induction theorem yields known currents [Eqs. (325». (This assumes that EI, Hl is known.) We cannot, however, use Eqs. (34) and (35) to calculate the field from J., M" for they radiate in the presence of the obatacle. A determination of this field is & boundaryvalue problem of the same order of complexity as tho original problem (Fig. 315a). We can, however, approximate the field of I., M, and thereby obtain an approximate formula for E, H internal to Sand E', H" external to S. A simplification of the induction theorem occurs when the obstacle is a perfect conductor. This situation is represented by Fig. 3160. The solution E must satisfy the boundary condition n X E .... 0 on S (zero tangential E). It then follows from the first of Eqs. (324) that
n
X E'
= n
X E'
ooS
(326)
We now know the tangential components of E' over S; so we can construct t.he induct.ion represent.ation of Fig. 31Gb as follows. We keep the perfect.ly conducting obstacle and specify that external t.o S the field E', H' exists. To support this field, there must be magnetic currents on S given by (327) MI=E'Xn=nXEi
to jump from zero at the conductor to those of E' just outside M"
We can visualize this current as causing the tangential components of E The
;/
t
E_EI+E'
n
o
conductor
Perfect
Source
p,rloct
conductor
~~ M. (n) (b)
nXE'
(a)
FIo. 316. The induction theorem a.e applied to a perfeetly conducting obstacle. Original problem; (b) induction eQuivalent
80ME THEOREMS AND CONCEPTS
115
Ei+ E'
Incident wave
..
M. . . .
Conducting
Conducting
plate
Ca) FlO. 317. Scattering by a conducting plate. cquiva.lent.
Cb)
plate
(a) Original problem; (b) induction
tangential components of E in Fig. 316b therefore have been forced to be E'. Thus, according to uniqueness concepts, the currents of Eq. (327) radiating in the presence of the conducting obstacle must produce E', H' external to S. It is interesting to compare this result with the previous one (Fig. 31Sh). We found that, in general, both electric and magnetic currents exist on S in the induction representation. How, then, can both Fig. 31M and Fig. 3lab be correct for a perfectly conducting obstacle? The answer must be that an electric current impressed along a perfect electric conductor produces no field. If the conductor is plane, this is evident from image theory. We can prove it, in general, by using the reciprocity concepts of the next section. To iIlustru.te an application of the induction theorem, consider the problem of determining the back scattering, or radar echo, from a large conducting plate. This problem is suggested by Fig. 317a. For normal incidence, let the plate lie in the z = 0 plane and let the incident field be specified by
(328)
According to the induction theorem, the scattered field is produced by the currents 11{1/ = Eo on the side facing the source and M 1/ =  Eo on the side away from the source. These currents radiate in the presence of the original conducting plate, as represented by Fig. 317b. Let the field from each clement of current be approximated by the field from an element adjacent to a ground plane. According to image theory, this means that each element of !If1/ seen by the receiver radiates as 2M1/ = 2E o in free space. Hence, far from the plate, it contributes
116
TIM.EHARMO:'lo.C ELECTROMAGN1o."T[C FIELDS
in the back«atter direction. Each clement Dot seen by the receiver contributes nothing to the backscattered field. Summing over the entire plate, we ha.ve the distant backscattered field given by E,'
if
dE,' =
j;~,A etk
(329)
where A i the area of the plate. The ed&o area or radar cross aeclion of an obstacle is defined a.s the area for which the incident \vave contains sufficient power to produce, by omnidirectional radiation, the same backscattered power density. In mathemfltical form, the echo area is
A. "'" lim (4Jrr! ~'.) ........ S'
(330)
where Sl is the incident power density and S· is the scattered power density. For our problem, = IE,lt/" and, from Eq. (329),
gl
S' _ !
•
I I'
kE,A
2rr
The echo area of eo conducting piMo is lherefore
A, =     r ~,
k'J'P
hAt
(3.11)
valid for large plates and normal incidence. 38. Reciprocity. In its simplest sense, a reciprocity theorem stat.es that a response of 8. system to a source is unchanged when source and measurer are interchanged. In a more general sense, reciprocity theorems relate a response at one source due to a second source to the response at the second source due to the first source. We shull establish this type of reciprocity relationship for ac fields. The reciprocity theorem of circuit theory is a special case of this reciprocity theorem for fields. Consider two sets of ac sources, J", M nnd Jb, Mb, of the same frequency, existing in the same Iinenr medium. Denote the field produced by the a sources alone by E, H·, and the field produced by the b sources alone by Eb, Hb. The field equations are then
v X HG
 .. X E'
=
~
m' + M"
tiE
+ J
"XH'~E'+J'
v X Eb= iH'+Mb
We mulliply the first equation scalariy by Eb and the last equation by Hand add the resulting equations. This gives

~.
(E' X H')
~ ~E'
. E' +
m" . H' + E' . J' + H' . M'
SOME THEOREMS AND CONCEPTS
117
where the lefthand term has been simplified by the identity
V . (A X B)
~
B.V X A  A.V X B
An interchange of a and b in this result gives
 V . (Eo X H')
~
OEo . E'
+ SHo . H' + Eo . ]' + H' . Mo
+ H' . Mo
 E' . ]0  Ho . M'
(332)
=
A subtraction of the former equation from the la.tter yields
 V •
(Eo
X
H'  E'
X
Ho) _ Eo . ]'
At any point for which the fields are sourcefree (J reduces to
V .
M
=
0), this
(333)
(Eo X H'  E' X Hp)
~
0
which is called the Lorentz reciprocity theorem. If Eq. (333) is integrated throughout a. sourcefree region and the divergence theorem applied, we have
effi (Eo
X H'  E' X Ho) . ds ~ 0
(334)
which is the integral form of the Lorentz reciprocity theorem for a sourcefree region. For 0. region containing sources, integra.tion of Eq. (332) throughout the region gives
 effi (Eo
X H'  E' X Ho) . ds
 III (Eo.]' 
Ho. M'  E'·]o
+ H'· Mo) dT
(335)
Let us now postulate that all sources and matter arc of fInite extent. Distant from thc sources and matter, we have (see Sec. 313)
E, = TjN..
r +
E~ =
TIll,
The Icftrhand term of Eq. (335), integrated over a sphere of radius <:0, is then
'Il1/> (lJ,GH,b + H.alI. b Equation (335) now reduces to
11,bH,a  JJ(>bH.") ds = 0
III (Eo.]' 
Ho. M') dT
~
III (E'·]o 
H'. Mo) dT
(336)
where the integration extends over all space. This is the most useful form of the reciprocity theorem for our purposes. Equation (336) also applies to regions of finite extent whenever Eq. (334) is satisfied, For
118
TIMERAlUIONIC ELEctROMAGNETIC FIELDS
example, fields in a region bounded by a perfect electric conductor satisfy Eq. (334) i hence Eq. (336) applies in this case. The integrals appearing in Eq. (336) do not in general represent power, since no conjugates appear. They have been given the name reaction. I By definition, the reaction of field a on source b is
(a,b) ~
fff (E· J' (a,b)  (b,a)
H' . MO) dT
(3117)
In this notation, the reeiprocity theorem is
(3118)
that is, the reaction of field a on source b is equal to tho reaction of field b on source a. Reaction is a useful quantity primarily because of this conservative property. For example, reaction can be used as a measure of equivalency, since a source must have the same reaction with all fields equivalent over its extent. This equality of reaction is a necessary. but not sufficient, test of equivalence as defined in Sec. 3·5. We shall use the term &tllreaction to denote the react.ion of a field on its own sources, that is, (a,a). A valuable tool for expositional purposes can be obtained by using the circuit sources of Fig. 31 in the reaction concept. For 0. current source (Fig. 31a), we have
(a,b) 
f E'· I'd! I' f E·d! ~ V'I'
where V is the voltage across the b source due to some (as yet unspecified) a source. For a voltage source (Fig. 31b), we have K' =  V', and
where 1 is the current through the b source due to some a source. summarize, the "circuit reactions" are
To
(a,b) 
l + V'I
J
V'l'
b a current source b a voltage source
(3119)
If we use a unit current source (I' = I), then (a,b) is a measure of V(the voltage at b due to another source a). If we use a unit voltage source (V' = I), then (a,b) is a measure of J (the current at b due to aoother source a). To relate our reciprocity theorem to the usual circuit theory stMement of reciprocity, consider the twoport (Courterminal) network oC
IV. H. Rum.scy. The Reaction Concept in Electromagnetic Theory, PAy,. Rn.,
Bel'.
2, vol. 94, no. 6, pp. 14831491, June 15, 1954.
SOllE THEOREMS AND CONCEPTS
119
Fig. 318. The characteristics of a. linear network can be described by the impedance matrix [z} defined by
[V,] _[,,, ,,,] [I,] V, I,
Zu Zu
(340)
Supposc we apply a. current source II at port 1 and a. current source 1, at port 2. Let the partial response Vi} be the voltagc at port i due to sourcc I} at port j. Each current source sees the other port openeircuitcd (sec Fig. 31a)j hence
ti}  
V'J
IJ
In terms of the circuit rcactions [Eq. (339)J, (j,tl =  VijI.; honce
z·· 0,') 1)
/;/}
(341)
Thus, the clernent<s of the impedance matrix are thc various reactions among two unit current sources. The reciprocity theorem [Eq. (338)], applied to Eq. (341), shows that
(342)
which is the usua.l statement of reciprocity in circuit theory. Equations (341) and (34.2) also apply to an Nport network. The use of voltage sources instead of current sources gives reactions proportional to the clements of the admittance matrix {yI, and reciprocity then states that
Yi} = Vit'·
The proofs of many other theorems can be based on the reciprocity theorem. For ex~mple, the preceding paragraph is a proof that any nelwork constru.cted 0/ linear isotropic matter has a symmetrical impedame matrix. This llnetwork" might be the two antennas of Fig. 319. Rcciprocity in this case caD be stated as: The voltage at b due to a current source at a is equal to the voltage at a due to the same current source at b. If the b antenna is infinitely remote from the a antenna, its field will be a plane wave in the vicinity of a, and vice versa. The receiving pattern of an a.ntenna is defined as the voltage at the antenna.
v~{~
\
(1)
(b)
(2)
Network
\
Flo. 319. Two antennas.
Fla. 318. A twoport network.
l20
TIMEHARMONIC ELECTROMAGNETIC FIELDS
terminals due to a plane wave incident upon the antenna. The reciprocity theorem Cor antennas can thus be stated as: The receiving patlern of any antenna constructed of linear i80tropic matter i. identical U> it& tranamilling patlem. In Sees. 35 and 37, we used the fact that an electric current impressed along the surface of a perfect electric conductor radiated no field. The reciprocity theorem proves this, in general, as follows. Visualize a set of terminals a on the conductor and another set of terminals b in space away from the conductor. A current clement at b produces no tangential component of E along the conductor; so V06 (Vat a due to 16) is zero. By reciprocity, Vk (Vat b due to 10 ) is zero. The terminals b ata arbitrary; BO the current element along the conductor (at a) produces no V between any two points in space; hence it produces no E. We can think of I. l\S inducing currents on the conductor such that these currents produce a freespace ficld equal and opposite to the freespace field of I •. 39. Green's Functions. Our reciprocity relationships are formulas symmetrical in two fieldsource pairs. Mathematical statements of reciprocity (symmetrical in two functions) are called Green's theorems. The difference between a Green's theorem and a reciprocity theorem is that no physical interpretation is given to the funcHons in the former. The scalar Greeo's theorem is based on the identity
V· (fV~)  fV'~
+. vf· v~
When this is integrated throughout a region and the divergence theorem applied to the lefthand term, we obtain Green's first identity
(343)
Interchanging y. and ¢ in this identity nnd subtracting the interchanged equation from the original equation, we obtain Green's second identity or Green's theorem
(344)
This is a statement of reciprocity (or scalar fields y. and ,p. The vector analogue to Grecn's theorem is based on t.he identity
V· (A X V X B) = V x A . V x B  A· V X V X B
An integration of this throughout. a region and an application of the divergence theorem yields the vector analogue to Green's first. identity
1ft (A X v
X B) . cis
~
III (V X A· V x B 
A· V X V X B) d. (345)
SOM.E THEOUEAiS AND CONCEP1'S
121
d.
FlO. 3·20. Region to which Green's thoorem is applied.
o..c:..4:~
r
We can interchange A and B and subtract the resulting equation from the original equation. This gives the vector analogue to Green's second identity, or the vector Green's theorem,
~(A X V X B  B X V X A) ·ds
=
OUf
fff (B • V X V X A 
A • V X V X B) dT
(346)
reciprocity theorem [Eq. (33S)}, for a. homogeneous medium, is essentially Eq. (346) with A = Ell and B = Ell. For an inhomogeneous medium, still another vector Green's theorem corresponds to our reciprocity theorem (sec Prob. 328). Green's theorems have been used extensively in the literature as foUows. Suppose we desire the field E at a point r' in a. region. Instead of solving this problem directly, a point source is placed at r , a.nd its field is called a Green's function G. We then substitute E = A and G = B in Eq. (346). This gives a formula for E at r' , as we shall discuss below. What we have done is solve the reciprocal problem (source at the field point of the original problem) and then apply reciprocity. The equivalence principle gives the solution more directly. Let us summarize the various Green's functions used in the literature. Stratton chooses l
(347) whe'e
~ _
CJk1rr'[
I'
"1
(348)
and c is a constant vector. A comparison of Eq. (347) with Eq. (2117) shows that G 1 is the vector potential of a current clement II = 4rc. Hence, G 1 is a solution to Eq. (2108), or
V X V X G1

k 2G 1 = V(V . G 1)
r F r'
(349)
Now suppose we wish to find E at r' in a sourcefree region enclosed by S. The source of G 1 is placed at r' and surrounded by an infinitesimal sphere 8, as shown in Fig. 320. Equation (346) with A = E and B = G 1 is now
I J. A. Stratton, "Electromagnetic Theory," p. 464, McGrawHill Book Company, Inc., New York, 1941,
Furthermore.e . An application of Eq. This is proportional to the electric field of an elect. is (E X V X G. . 1955.. X V X E s E X V X G. R. in Fig.ric current elementi 80 G. r' (352) We now apply Eq. We can choose other G's such that t. All of t. formula.) . G t is a solution to r .\Q) which is a. for calculating E at r' in terms of n X E.G. it is required that E be continuous and have continuous first derivatives on S. (354) where ¢ is given by Eq. 320." p. (353) is a substantial improvement over Eq. E  t $. 14. (348). In fact. Thus. I The lefthand aide of this equation is a function only of the primed coordinates.)· ds (3. The result. Eq. The result. This is a severe restriction on the usefwness of Eq. Hence. n X V X E. A choice of Grecn's function which overcomes some of the disudvantagcs of Eq. V' X E ~ 1ft (G. "Scat. they. (353). P@tgamon Press. (346) with A = E and B = G t to the region enclosed by Sand.are fields of sources radiating into unbounded space. (353) can be shown to be identical to the formula obtained from the equivalence principle of Fig. E on S. is' 4. a prime is placed on v' to indicate operation on r' instead of r. (352). 39. . nor do we need to know n . Kew York. (348). (3SO). Menber. I J. Eq. (346) would yield a formula for E at r' .122 TI3oIEllARMQXIC ELEcrROMAGNETIC FIELDS applied to the region enclosed by Sand 4rc. (350).ion of Radio Waves. applied to a homogeneous medium. (350) is l O 2 = v X ttl> (351) where tP is given by Eq. E on S.. Another useful Green's function is G.. also satisfies Eq.he G's considered 80 far are tlfrcespace" Green's functions. Bence.hey satisfy boundary conditions on S.=vxvxcq. similar in form to Eq. and n . that is. This is evidently the magnetic field of a current clement 11 = be. ds (353) This is a formula for Vi X E (bence for H) at r in terms of n X E and n X V X E on S. X V X E + E v· G.tering and Diffract. Equation (353) does not require E to be continuous on S. although it can be amended to admit singular E'8 on S.
.] E. The most general linear relationship between two vector quantities can be represented by a tensor. All these various formulas. as it applies to a homogeneous region. because of Eq. be· v· X E ~ 1ft (G. (346) with A = E and B = G 4 results in Eq." Suppose we have a current clement II at r' and we wish to evaluate the field E at r. We have discussed the Green's function approach mel'cly bec:luse it has been used extensively in the literature. r. 310. (353) with the last term zero. X V X E) .iating in thc presence of a perfect electric conductor over S. r. and the G t ' is the scattered field..] [II. 311. and ma. ill which case . = [r. (356).. Eq. can be directly obtained from the equivalence principle.] Il" Il. We shall henceforth usc the term /lOreen's function" to meau "field of a point source. Hence. I'lIz 1'1/11' I'.j is the ith component of E due to n unit jdirected electric current clement. (361) r az rill r. (352) and n X V X G4 "'" on S (356) The physical interpretation of Gt is that it is the magnetic field of a current element II = hc rad. the field E is related to the source 11 by E ~ Ir]!l (360) In rectangular components where IfJ is called a tensor Green's function.trix notation.SOME THEORE~IS AND CONCEPTS 123 (355) For example. Similarly. The E might be the freespace field of 11. defining a G~ such that n X G6 = 0 on S (358) we can obtain a formula 41fc • V' X E = 1} (E X V X G 6 ) • ds (359) and so on. Tensor Green's Functions. The G 2 is the incident field. The same formula can be obtained from the equivalence principle of Fig. Thus..l' r . (360) becomes Elf [E. Thus. Application of Eq. and many more. ds s (357) which is a formula for V' X E in terms of only n X V X E over S. let Gt = G2 +G 0 t' such that G 4 satisfies Eq.
A) . v(v .124 TIMEHARMONIC ELECTROMAGNETIC FIELDS [rJ would be the "freespace Green's function. represents the solution of Eq. (360).1. it should prove instructive to flnd an explicit IrJ. E '" = Jwp. (362) is symbolic of the solution. (363) Equation (362) also represents the field of currents in the vicinity of a material body if {rl represents the appropriate Green's function.4. (2111).. (361) for 1"" = Il. = OJ we sec that r""" = ." Still other Green's functions are those relating H to 11. If II is zdirected. A + 1 a'A.. which is E A  jw. 1 ( JwJl+·. and [1') would then be called the uGrean's function subject to boundary conditions. Let us take [rJ to be the freespace GreenJs function defined by Eq. For example. "I . a x' E __1_ alA". and so on. Our principal use of tensor Green's functions will be for concise mathematical expression.r I =:. (360). E __1_ alA".rdT . tI jwe ayax jWf. those relating E to Kl. f JWf. '. E might be the field of 11 radiating in the presence of some matter. and so on. a X a') 1 a r "'" jWf.'"'''' ':t'lrlf  + Jw.A Iff J.dZ ax iJ"Yt eJl:1rr'1 f  4rI' _ <'I (364) . In other words.!. Eq. '" JWt . the equation (362) where [r] is tbe freespace Green's function defined by Eq.ay 1/1 ax r"" where = JWt J. regardless of whether or not we can find [r]. . Even though we shall not use tensor Grecn's functions to find explicit solutions. I az ax Comparing this with Eq." Alternatively. I LeJl:lrr'l and A .
(2118) is essentially an integral equation when J is unknown. (327). _ (i w• +. the potential integral of Eq. An integral equation is one for which the unknown quantity appears in an integrand. To illustrate the formulation of an integral equation. consider the induction theorem of Fig.r')) dM(r') The total scattered field for the problem is then the summation E'(r) 1} [r(r.'~J' r r01 1CI ia" jWf ai aj (3ll5) with 1/J given by Eq. For example. Errors in some elements of the summation may be compensated for by errors in otber elements./s. The result is theref9re r. There is good reason for this. (364).he conductor over S. It is much more important that tbe elements contributing most to the summation be correct than that the elements of minor contribution be correct. We already have the concepts needed to construct integral equations. and it is not necessary that each element of the summation be correct. From symmetry considerations.S$uming the current on the linear antenna of Fig.a~2 a')~ JWf . When exact solutions nre desired. 317b. This is why we were able to obtain useful results by B.will differ only by a cyclic interchange of (x. Integral Equations. this is dE(r) ~ [rer. all elements do not contribute equally to a summation.(r') do' When r is on S. the differential equation approach is usually the simpler one. In equation form. 311. Integration is a summation process. symmetry The reciprocity theorem is reflected in the (366) which can be proved for r's subject to boundary conditions as well. (326) must where M. Eq. 223.r')]M. the other r . is given by Eq.z). by assuming the field of each element of magnet. and so on...r')J be the tensor relating the E field at r due to an element of M at [' radiating in the presence of t. .. Let [r(r.!.SOME THEOREMS AND CONCEPTS 125 The other elements of [r] are found by taking n to be ydire9ted' and then zdirected. An important use of integral equations is to obtain approximate solutions. Most problems can be formulated either in terms of integral equations or in terms of differential equations.y. 316. Also.ic current in Fig.
tH + M' vxHliE+J' where ~ and y arc functions of position. We can now rewrite the field equations as v X E = t. Since J and M are functions of E and H. the field satisfies v X E . (367) would be difficult even for the simplest specialization.(Ii . (Recall the linear antenna problem. We can define normal values of iropedivity and admittivitY. 223 and the obstacle of Fig.E + J' + M' (369) The effective currents in excess of the true sources (M' and J') are now just those due to the motion of atomic particles in vacuum. is assumed to be known. Problems involving a region homogeneous except for small uislands" of matter are commonly encountered. if ~ = . possibly containing sources Ji and M'..€1 and '01. we can consider the total field to be the potential integral solution of Eqs.)E + M' + J' These effective turrents can then be treated as source currents in a homogeneous region.126 TU. a solution in terms of them will lead to an integral equation. Let us reconsider the problem of scattering by an obstacle in the light of the above concepts. (369).) Note that.)H J jw('  'olE +. which may be any convenient constants (usually the most common .fEHA.€l and '0 = 01 except in small subregions.)H J . where we assumed I on tbe antenna wire.n X 1ft Ir(r. Examples of such problems are the linear antenna of Fig. 3t5a. so Eq. when the normal z and yare taken as the freespace parameters. Eqs.H where the effective currents are +M v X H = Ii.1.R~lONlC ELECTROMAGNETIC FIELDS also be true. (367) is an integral equation for determining Irl. with J and M given by Eqs.1i.r')]E'(r') s X ds' ron S 1367) The incident field E. hence n X E'(r) . an exact solution to Eq. To illustrate the general concepts involved. The incident field is that produced . suppOSe we have an inhomogeneous region. we can assume J and M in the su~ regiolls and obtain approximate expressions for E and H elsewhere. 315a. (34) and (35). Given the problem of Fig. However. Within this region. (368) reduce to M ~ jw(P .".€ and '0 in the region). As we mentioned earlier.E +J (368) M = (I .
. and this current resides primarily on the surface of the obstacle.)H (370) J = jw(' .'I~el dT . then Eqs.E throughout the obstacle.400 l F k  _ 1 oblto. (371) and (372) will give us an approximate solution. If we assume the obstacle perfectly conducting. The solution in this case reduces to E' = 1 114 r"JWfo V X V X effi JI'e/l:1r'I 'l ds' t r r (3i3) 8 If wc specialize this equation to S.he surface current on the conductor is given by J. r . 321.c1e iff 1.. we assume that thc total field is negligible in n Incident wave • Ca) Cb) (a) Fro. M. To be explicit. represent a perfectly conducting obst~cle illuminated by some source. . will be zero. known as the physical optics approximat1'on.el "I dr .pproximation. is as follows.r' i (3i2) obn.. In terms of the t.<o)E + . Let Fig. An approximation to J. (326) must be met.jw(ll . 3210.. t. then Eq. M. Thc physical optics approximation. J iff l + JWfo v J X V X A (3il) .SOME THEOREMS AND CONCEPTS 12i by Ji and Mi' outside of the obstacle. then J becomes a true surface current.otal field. outside of the obstacle E' .ele with J and M given by Eq. = n X H When the obstacle is large. Original problem. and consequently F. J reduces to uE. (370)... For a good conductor. If we can guess J and M with reasonable accuracy... For a nonmagnetic obstacle.v X F where A . and the scattered field is that produced by M . and we have an ibtcgral equation for determining J•. (b) the .
(329) and (376)) is no coincidence. IRE Tram. 150153. April. pp. Harrington.~w If dE' . ..Ilu e' The total distant backscattered field is thereCore . "'" 2n x HI over Sf (374) where S' is the illuminated portion of S.t28 TUlEBARllONlC ELECTROMAGXETIC FJEbDs the II shadow OJ region. The physical optics a. Furthermore. hence fl' . 321b. 2. again consider the large conducting plate of Fig. On Scattering by Large Conducting Bodies. The incident E is given by Eq. _ 2Eo • Each element of this radiates as a current element in Cree spnce.. no. 1959.. It can be shown that the two a.. According to image theory. in Sec.. each element of surface behaves similarly to an element of a ground plane.jkE. As an explicit application of the physical optics approximation. (376) which is identical to Eq.: The physical optics approxi 1 .. if the obstacle is smooth and gently curved. the approximation obtained from the induction theorem. (329). We therefore approximate the current on the obstaclj) by I. (331). I I R..:=. (374») is therefore J. 29. = Eo eih . the tangential components of H at a ground plane are just twice those from the same source in unbounded space. This equality of the two approximations to back scattering [Eqs. The contribution to the radiation field in the backscatter direction from each J. 2." .pproaches always give the snme back scattering but do not give the sume scattering in other directions.Jwto V X v X If 8' (n X H~c"l..pproximation to the echo area of the plate is therefore that of Eq. vol.do' If 1 I (375) This approximation is illustrated by Fig.. AP7. 317a. ds is • jkE o ds dE• = 211'T . 8B analyzed. The physical optics approximation to the obstacle current [Eq. F..'" .. mation to the scattered field is therefore E' ~ .A .. (328).
. More important. There is a great deat of arbitrariness in the choice of vector potentials. For instance.SOME THEOREMS AND CONCEPTS 129 312. The electromagnetic field in terms of A and F is given by Eqs. (378) to V2A V2F 2 v·F =w (380) +k A = 0 + k F '. the 6cld aa. (380) are satisfied. we shall learn how to construct many other solutions. we can alternatively write Eqs. (34) with J = M = 0.<l2) Also. we have explicitly considered ~nly two types of solutions to the field equations. we can express the field in terms of a magnetic vector potential A or in terms of an electric vector potential F. we can employ superposition and express part of the field in terms of A and part in terms of F.tion. and the F a solution to the dual equation. Note that the rectangular components of the wave potentials satisfy the scalar wave equa. The general equations for vector potentials aro therefore v X V X A V X V X F  k 2A = 9v~ k 2F = zV<t>' (378) where cit" and ¥ arc arbitrary scalars. namely. A = U4>" This reduces Eqs.0 2 (381) SolutioDs to these equations arc called wave potentials. A general method of obtaining these solutions is considered here. (J. (2108) with J = 0. or E=VXF+~VXVXA H~VXA+~VXVXF (379) Equations (378) and (379) arc the general form for fields and potentials in homogeneous sourcefree regions. The A must be a solution to Eq.ZH X H ~ ~E (377) In view of the divergcncelcss character of E and H.tis6ca v v X E . we can choose the arbitrary *'s according to v . when Eqs. In a homogeneous sourcefree region. uniform plane waves and thc potential integrals. or Helmholtz cquation. So far. Construction of Solutions. In the next three chapters.
130 (379) .gF + v (v . = 0 (386) 1 E.M).. where we considered t.+k') '" dZ' A field with no IJ.v X F 1 H . = 0 H. = ~(::s + kS)Vt A field with no E. u. In the dual sense. Let us now consider some particulo. We afe now concerned with regions of finite extent. This happened to be our choice Cor the potential integral solutioo.fA + ~ V(V . is called transverse magnttic W z (T. fI=_iJt • iix H. A) H . TIYEDARMONIC ELECTROMAGNETIC FIELDS 1 E = v X FlA+v(V·A) 9 1 H . If we take F = 0 and A = u. is called tromt'trse electric to z (TE). if we choose A=O and F .1/t (387) (388) then E . regardless of its actual sourcc. F) (383) We have yet to decide how to divide the field between A and F. As a word of caution.r choices of potentials.. We shall find it possible to choose Vt sufficiently general to express any TE field in a homogenoous sourcefree region according to the above formulas. (384) then E . F) f 1 ii'~ Expanded in rectangular coordinates. = • 'i ax oz fayot E _ iJt • AX: 1I _ ~ ii"l (389) E. and we ctln represent a field in terms of A or F or both.gF + j v(v. We shall find it possible to choose y. . this is E~ =  a".Y (ii' .v X A .he sources everywhere. do not make the mistake of thinking oC A as due to J and F as due to M. oy H.. .y.v X A (3&) This can be expanded in rectangular coordinates as E~ = gaxaz yayaz 1 ii"l 1I _ ii'" • iiy E ~ ~ ii"l . sufficiently general to express an arbitrary TM field in a homogeneous sourcefree region according to the above formulas..
Explicit expressions for the field would be superposition of Eqs. our starting equations are V X E = ZH + M (392) vxH=yE+J instead of Eqs. We can think of the particular solution as the field due to . We already have a particular solution.E~ + E" H ~ H•• + H" (393) where the particular solution (pa) is formed according to Eqs.' _ az' 'f' !H • Thus. (379). called a complementary solution. 32. is found from 8'" + k'. We must therefore study solutions to the scalar Helmholtz equation to Jearn how to pick the y. ThereCore. where the y. This TM field will have the same E. (389). as does the original field. 'It according to 8"/1' az' + k'. with superscripts a and f added to the y. v X X (c>il) +~v y X V X (c>/l') (391) H = V (c>/l') 1 + 'j V X V X (C~f) where the y." '~ "E' 'f 1I which will generate a field TM to z according to Eqs. an arbitrary field in a homogeneous sourcefree region can be expressed as the sum of a TM field and a TE field. the po~ential in~e gral solu~ion of Sec.'s. We cnn therefore determine this difference field according to Eqs. (392) call be constructed as the sum oC any possible solution. and the complementary solution (cs) is constructed according to Eqs. (34) and (35).'s are solutions to Eq. (382). namely. so the difference between the two will be no TE field. solu~ions in a homogeneous region containing sources are given by E .c>/l' where c is a constant vector. Since the z direction is arbitrary. called a particular solution. If the region is not sourceCree but is still homogeneous.SOME THEOREMS AND CONCEPTS 131 We can dctermine Now suppose we have a field neither TE nor TM. (386). which become E  The field is then given by Eqs.'s to distinguish between them. we can express this independent of the coordinate system by defining (390) A .'. (391). (377). (386) and (389). a. General solutions to Eqs. plus a solution to the sourcefree equations.
(See. be neglected in the flphase factors. unless r~ «".r' cos t (394) » where t is tbe angle between rand r'. by duality. 322. however. If we specialize to the radiation zone (r r:.['1+ r . Note tha..) F  in the radiation zone. (394) can be neglected in the Umagnitudc factors/' Ir .he field due to sources outside the region. Geometry for evaluating the radiation field. we have Ir . 29 and 210. (34) and (35). Consider a distribution of currents in the vicinity of the coordinate origin.).. The complete 80lution to the problem is represented by Eqs. The same is true of a ma. we shall formalize the procedure for specializing solutions to the ra. The Radiation Field. of Eqs..t we now ha.iation zone. Hence. for example.d. the second term of Eq. In Sec. Rather thnn blindly expanding Eqs.ell).tions or Eqs. Furthermore. Thus. Eqs. Many or the opera. (35)..) In this section. immersed in a homogeneous region of infinite extent. as suggested by Fig. It is easier to evaluate tbe radiation (distant) field from sources of finite extent than to evaluate the near field. 322. (35) reduce to A ~ : ' JJJ J(r')&"'dT' ~ JJJ M(r')&"''dT' (39.r/lI.ve the T dependence shown explicitly.. (34) can therefore be performed. 313.1I exp (jkjr . the z To distant field r . let us draw upon some previous conclusions. 29 it was shown that the distant field or an electric current clement was esscntinUy outwardtraveling plane waves. POi"1 r Source r' Fro.gnetic current element. Sees. y x . It cannot. (34).132 TIMEBARlIONIC ELECTROMAGNETIC FIELDS sources inside the region and the complementary solution as t.
for the partial E field due to M.jkA.LA.(V X Flo . .l H field due to J according to H' =. we have. let us determine r' cos ~ as a function of the source coordinates. we ha. E. with E' given by Eqs. in the radiation zone. y y x radiation zone must be characterized by  (396) since it is a superposition of the fields from many current elements. (397) E. . cylindrical. (396).SOME THEOREMS AND CONCEPTS 133 z Flo. we have the transformations x=rsin8cost/J y = rain 8ain t/J z=rcos8 To obtain r' cos ~.ve H~ =. in the radiation zone. The three coordinate systems of primary interest are the rectangular.V X A (see Sec. Simila.LA. or E.rly. The total field is the sum of these partial fields. We can evaluate the partia. = jwj. . =jkF. as illustrated by Fig. (396). Also.W oricutu. Conventional tlounIiull. = jwj.tioll. 323. H. E~ = (v X F).(V X A).' ~ . no differentiation of the vector potentials is necessary to obtain the radiation field. with H given by Eqs. and spherical.jkF. ~ (V x A). (396). for future reference. with R" given by Eqs. 3~23. = jkA. Retaining only the dominant terms (r 1 variation). Thus. For the conventional orientation shown. r' = xx' + yy' + zz' (399) .jkF. + jkF. 32). X=PCOSt/J y=paint/J z= z (398) we form rr' cos ~ = r .
(398).134 TWEHARMONlC ELECTROMAGNETIC FiELDS X.J. Show that a current sheet J u. substituting into Eq.0 in the waveguide of Fig. Substituting into Eq. Substituting for Y.0 plane produces the out.. we obtain ~ = r' cos (x' cos tP + v' sin 4» ein (J + %' C08 8 (311lO) which is the desired form when rectangular coordinates are chosen for the source. Instead of the electric eurrent sheet. we obtain r' cos ~ = p'ain Seos (41 . Suppose now that the two current sheets J.p')] (3102) wruch is the desired form when spherical coordinates arc chosen for the source. (3100) for x'. y'. . Vi. '7 . sin 8 sin 8' cos (I/> . z' from the second set of Eqs.. over the: . suppose that the magnetic current sheet exi!lt. (398). .>0 . _. Finally..u·Z.. z' from the first set of Eqs. (3100) for x'.u r A sin A .SIn"b M.. 32.<0 ss.s over the ctOSll section: .q/) + z' cos (J (3101) which is the desired form when cylindrical coordin:ltes are chosen for the source. PROBLEMS S1.~"'{ Tel'u .> 0 J <0 in an infinite homogeneous medium.. (398). z from the first set of Eqs. . .wardtraveling plane waves HI<  . 32.. we have r cos E = r'[cos 8 cos 0' +. magnetic current :>roducea a field Show that thiJl.
hat wben d i.' { E "'~ I'.. 37. .. 32. Show that this produces the same field as the electric current clement of Fig. 228) arc almost dual to each other. an odd number of guide quarterwavelengths.z. 228 is the TM. suppolle that a "shorting plate" (conductor) is placed over the croaa section I: . Show that the field dual to the TM. By integrating along the eontours shown in Fig. 324b. (0) Twin3lot line. Figure 3240 aho. Obtain the field of an infinitc.> 0 d { jJ. 221 if I 31.. (33)1.d. 36." the cfOlSllleCtion of a "twin·~o1ot" transmission line. 228 is the TE modo of this new waveguide.. Figures for Prob. mode of Prob. '('J H (b) Flo. determine the line voltages and TT 1 (oj lW E ~ H c. mode for the parallelplate guide having conductors over the planes 11 _ b/2 and 11 .llimalloop of magnetic current having zdireeted moment KS. _ J.EOREMS AND CONCEPTS ~ 135 Show that. Show t.. (c) into . Show t. 36. (b) collinear plate line. Iration contoUr!.b/2. no HI erists {or I: > O. In Fig. for Ii > 0 is twice that for the current sheet alone [aee Eq.hat the field distribution is dual to that of the eollinur plate line of Fig.(1 e. but when d is an integral number of guide haHwavelengths.ei~ sin Tsin 1tJ(d + %)1 < I: < 0 Note t. D H cd \ c.z. 32. thcaa pro exist.hat the current abeet of Eq.ame''2 b .. . modo of Prob. 324.•• oJ' rv _. (32) now produces a field H.SOME TU.. E. The TE and TM modes o{ a parallelplate wavcguido (prob. simultaneously over the Cf'06l leetion duce a field I:  0 of Fig. Show that the field dual to the TE. • >0 •<0 This BOurce is a "directional coupler. 32k." 34.
(Z')NolUM . 712). +. almost vanishes at the resonant frequency (Eq. a small M. ~++. show that the currente J ~E!VXVXE 1 will support this and only t. Show that the radiation field ia given by E..sin 'Z sinh 'YC at the wall z . Show that the current sheela II ] . tilaI'.m' a at Ja l 1). (2. {295)J..v X &.ric. 89. H having identieal lange. _ 'Ill . given by Eq. (32).. . From Table 23. 310. Show that the same H. principle. ) .hat 7i>3~'~':. Seo.eU jh (iwp.'lI.. JoZ• sm 2 To! I~ . What do our uniqueness theorems say about tbia l5CCOnd IlOlution? WhaL ean we say about it on physical grounds? Give a couple of other possible solutions to tbe problem.. (l produce the field or Eqa. {Z.SUl' o· over the sphere r . Show that thia field supported by the lJOurce V M.Bln"bBln 'YJ: where eAD be 'Y  (rib)' . . and if fH ..d(ZIpta_U.he rectangular cavity of Fig. that iB. M. 38.ne. 811.> 0 . 4.. log (WIlD) D » UI The two tranamission linea are said to be complementary struetUJ'C8 (Bee Babinct'..  { J:.u.0 pla.)ot. produces a large E.rY'h E .4... M. Show t. If E is wellbehaved ill a homogenoous region bounded by S. and interpret them physically.hi& field among a claas E." 1).I ia also Ii mathematical801ution to the problem of Fig.c. but different H. 3U1.113) r > a and !loro field r < 4. it foUowa that. can be obtained within this (ll&sfl if magnetic sources K are allowed in addition to J. 325. 219 a field ..AI' ain 6 lin (kd sin • ain 6) " . Show that the field E.tl and 1 is complelt (lOll8Y dielectric).136 TIMEHARMONIC ELEcrROMAGNETIC FIELDS currents of both the 810t line and the plate line. 32 with J. Consider a ~rected current element II a distance It in front of a ground plane covering the l' .nti&! components of E on S.<0 ain ~ . Show tbat for a low~loS6 dielect.E. Suppose there exiBte within t. . (ik (l II u. aa shown in Fig..E . .
2>01). For large d. Current element parallel to a ground plaoe.. 325.SOME THEOREMS AND CONCEPTS 137 and 1/110 _ H.5 for d small.Thus. E. ('8)' [1 + cos 2kd "3 (2kd)l <' sin 2kd] (2kd)' .. and that the gain is 7. (kd cos ' ) .a. instead of the current element. O Air R.15 for d . referred to / is Find the power radiated and show that the radiation resist.. referred to I is Find tbe power radiated and show that the radiation resistance R.. _ I/ktlS ei~' sin 8 cos (kd sin".:O 15 ('8kd)' . id.IS '1 . • FlO. suppose we have a small loop of electric current. sin 8) 2~ .ximum radiation is at 6 _ 45" for small d.e 6.'lj2. In Fig. instead of the current element. z . the ffilH'imurn radiation is in the 1/ direction.. 4. In Fig. and the gain i. Show that the radiation field is given by E • . 3&. the maximum radiation lies close to tbe ground plane. ma. Show that the radiation field i~ given by E. II y 313. .. with ~irccted moment IS..A/4. The gain at small d is 15. _ E•. 314. and approximately 6 for d large.nce R _ '1""11 [~ • A' 3 _ sin 2kd _ cos 2A:d + sin 2kd] 2kd (2kd)l. suppose we have a small loop of electric current with ~irected moment IS.. 3·25. ~ i'1""ISkd e/" sm 28 u . ' ~e"8m Sln and ?lB. (2kd)l Show that For d . For small d.s: k/4.
fed at. (0) • 010' dl . y>O mn8 H. Assume n X II is C8S6ntially zero over the entire surface and that tangential E is that of the transmissionline mode over the open end... as shown in Fig.1. Denne the radiation conductance of this antenna Il8 O. 81ri. 317. fed at the ground plane. it ill 7.. 88 shown in Fig.. •The slot and ground plane can be viewed as flo transmission line.. the center. A slot antenna consist" of a slot in a.ne z direction. 32G. for d . Consider an openended e08Jl:ial line (Fig..)'rindlpol• . z. It is coiled a dipole alot antenna when fed by a voltage impressed across the center of the slot. y<O 1. the radiation resistance of the ). and obtain the magnetic current equivalent of the form of Fig. Uaing duality. (320) and that the radiation conductance is onehalf that of Eq. 210.6 ohms.0.23). 3·16.. and the field in the slot will be essentially a harmonic function of kz.' H. show that this equivalent representation is the dual problem to the dipole antenna of See. TIMEHARMONIC ELECTROMAGNETIC FIELDS Show 'I'll th~t tho radiation resistance referred to / is R. 327. AS8ume in the slot. l<o  (}IdlY..) FIG. Show that the field is the same 118 that from the dipole antenna (Fig./4. For example.e_weoS(k~COS~) ~ C08(k~) _{ 4(R. it is 6. 314a without tho ground plane) of small radii a and b.. 36. abow that the radiation field is jv. R'u::QT X 4. (323).H. Show that the gain of the monopole is twice that of the corresponding dipole and that the radiation resist.138 aod . and for d. Treat the problem according to the equivalence principle as applied to a surface just enclosing the coax. conducting grolWd plane. E~./4 monopole is f(z. The monopole antenna consists of a straight wire perpendicular to a ground plane."". 31:k." (OS)' which is twice that for the isolated loop. For d . the gain is 3. Show that to this approximation the radiated field is onehalf that of Eq.1 and show that . in t. The monopole antenna.ance is onehalf. For w small. _ (OS)' [~ + sin 2kd + 008 2kd _sin 2M] h 3 2kd (2kd)' (2kd)' For smaJl kd The maximum radiation is lLlong the ground plane.). 326.
y . 317.n (k ~ ".i . 80 the input conductance is given by 818.. by Vi . in the aperture is that of the TE'1 z . oJ . lUlaumo E~ in the slot the same as in Prnb. FIG. A rectangular waveguide opening onto a ground plane. 327. Flo. 0'in. 327.SOME THEOREMS AND CONCEPTS 139 z . Figure 328 shows an'aperture o. 328..V.) [00' H 00") . antenna. ground plane. Bin (kL/2). For the antenna of Fig. and show that for arbitrary to JVMe.. is as plotted in Fig.. Assume that E.~)  {He lit II> 0 11 <0 319. A slot.. 224. /r I I I I where R.00' (k ~)] k 2 cos<f>s1II9 w.' 'Iff where {I'.nUmna consistillg of a rectangular waveguide opening onto 8. The input voltage Vi is related to V . 61116 f«(J.
. jiB .140 TUdEHARMONIC ELECTROMAGNETIC FIELDS waveguide mode. (336). Scattering by a rectangular plate.] It 1111 lin VI satisfies the reciprocity relationship Un ..he radiation field of the dipole antenna of Sec. sin 4Jo)J j21fT k(a/2)(Bio tP sin .. By applying voltage lIOurces t. 2jbEte/ b sin ( k ~ COB'.. T the . and show that.o the network of Fig... ~pcat Prob.scattered field in the zy plane is kED/lblllv sin (k(a/2)(sin . and show that the radiation field is H• .. sin 8) cos ( k ~ C08 8) 'IT C09 4> (. 317. 323.hogonal polarization.. 210. of Fig.. 318. (338) is valid. obtaining Eq. Usc rcdprocit.jlr sin Ik{a/2) (sin 4> sin 4>0) J 2. and &how that at large E . Figure 3·29 represents a rectangular condullting plate of width a in the direction and b in the .. (2125). 322. show that the admittance matril( luI defined by [ I. 329..~u when Eq.. at large T t. 3·20. place n 9direeted current element at.. sin fla) cos "'a + + Show that the eeho area is the same as obtained in Prob. . 4:.[ab eos 4>0 si~ (ia sin .abe.e t.] _["" "u] [v.ho scattered field in the :cy plane is H " .. that is.. 320 for the ort..po) cos 4> + + Show that the echo STelL is A• .po)]' XkaSln 4>0 321.y to evaluat.. k(a/2) (sin.'. To do this. tIond apply Eq.1 (kb COS 8)11 320. large r.li direction. Let the incident plane wave be specified by ~ Use the induction theorem with the aame appr~ximation as waa used in the problem r r FlG.
v X <tov X B) dT Let A . (350). Repeat. 326.(0. (316) I..0) p where (0. define the input impedance of the sheet of current as z. 4".ering.. <to .iI in an inhomogeneous region..A . Let VI' and V.E·. that.B X v X A)' ds fff . v X <tov X B and derive the modified vector Green's theorem 1ft <to(A X V X B . is.' _ V t·. V X B . Differential Obstacle 3·24. Prob.<tov X A . 329. and show that it reduces to Eq.I be the corresponding voltages when the obstacle is absent. 325 lor the current sheet and field of Prob.0) is the selfreaction of the currents and I is the total current of the sheet. let A_E· and B . (335).O .acle. 32. Define the scattered voltages as and show that V.E~ in a homogeneous isot. B _ E·. lIClatt. Derive the lefthand term of Eq.. (A X <tov X B) . be the voltage received at antenna 2 when & unit current source is applied at antenna 1.f{.SOME THEOREMS AND CONCEPTS 141 Flo. and show that the above theorem reduces to Eq. (E X V X G I (B' v X <tov X A . 330. Let Fjg. 3·2li.. Let VI be the voltage received at antenna 1 when a unit current source is applied at antenna 2 and V. 31. show 'JY •  G t X V X E +EV'GI)'ds If f'I. 330 represent two antennas in the presence of an obst. Evaluate Z when the field is given by Eqs. For the problem of Fig. 328. (335).ropie region.heorem (Eq. In the vector Green's t.c·E .A . Use the vector identity v . (33). 327.
be the magnetic field of a zdirected current element situated 11 > 0 and radiating in the prescnce of a perfect electric conductor covering the 11 "" 0 plane. je. (3~86). . Show that this gives the same answer as obtained in Prob. and evaluate the electromagnetic field. and evaluate the magnetic tenoor Green's function 11') defined by H . 328. S33. (365).00.. Let G. Let". Let of .8 you can. "" .[rlKI in free space. Show that where fl ClO ft  vex Vex + (11 y')' + ell x')! + (11' + Ill' + (z x'P il' )1 z')' 331. In other words. "'. 321 by an interchange of 4J and 4Jo.' differs from that of Prob." differs from that of Prob..8 many waye B.u. (357) to the problem of Fig. Apply duality to Eqs. 3~20 by an interchange of 4i and 4Jo. (395). etc. and evaluate the electromagnetic field. 836. 387. Classify this field in B. 330 to fl ...142 TIMEHARMONIC ELECTROMAGNETIC FIELDS 330. of Prob. Classify this field in as many ways as you can. 835. and show that the answer for E. Repeat Prob. Let c . Repeat Prob. 319.[rill 334. and apply Eq.. . Evaluate the I'll for the freespace tensor Green's function defined by H . let c . 3~21 using the physical optics approximation. (391).ii" and evaluate Eqs. 332.r il. e. 838. S39. 3~20.U z . Show that the echo area is identical to that of Prob. and show that the &nswer for H . polarization. 3·2l. (3~97) by c. Classify this field in as many ways as you can (wavetype. (3·89). Derive Eqs." e. Specialize the G.'\:panding Eqs.ib in Eqs. .). (34) with A and F as given by Eqs./iw in Eqs. 320 using the physical optics approximation. Show that the echo area is identical to that of Prob. and S be the y co 0 plane.
The Wave Functions. kll . Eq. Substitution of Eq.) lIt hM been shown by Eisenhart (Ann. and (2) those solvable by using one or more uniform plnne waves. (The choice of minus a constant squared is taken for later convenience. and z. By a method called 8sparation of variabw8. yields (43) Each term can depend. general solutions to the Helmholtz equation can be constructed in certain coordinate systems. let where k". The Helmholtz equation in rectangular coordinates is (41) The method of separation of variables seeks to fInd solutions of the form f . 'l'he problems that we have considered 80 far are of two types: (1) those reducible to sources in an unbounded homogeneous region. (42) into Eq. and z. solutions which are the product of three functions of one coordinate each. and k. (41). we use the method of scparation of variables to obtain solutions for the rectangular coordinate system. on only one coordinate.X(x) Y(y)Z(.p. 284. that is. 35.) (42) that is.. are constants. Since each coordinate can be varied independently. I In this section. Y. (43) can sum to zero for all coordinate values only if each term is independcnt of x. vol. 1934) that the Hclmholh equation is separable in 11 threedimensional orthogonal coordinate systems. Mmh. p. 143 . Thus. and division by. are independcnt of x. Equations (391) show us how to construct general solutions to the field equations in homogeneous regions once we have general solutions to the scalar Helmholtz equation. y.CHAPTER 4 PLANE WAVE FUNCfIONS 41. at most.
2 + k..1:.. needed for any particular problem are determined by the boundary conditions of the problem and are called eigenvalues or characteristic values. only two of the lei may be chosen independently. From Eqs. I The term harmonic function also is used to denote a rrolut. t = kt (45) This last equation is called the separation equation. is a solution to the Helmholtz equation. We can therefore construct more genera.) (47) are solutions to the Helmholtz equation when the Ie. For example. the separation parameters must satisfy k. satisfy Eq.2 + k. Linear combinations of the elementary wave functions must also be solutions to the Helmholtz equa. As evidenced by Eq. .1:.l wave functions by summing over possible choices for one or two separa.•••• . They will be called harmonic equatiom... A sum of harmonic functions is still a harmonic function. (45).)h(k.tion parameters.tion.y)h(k. A constant times a harmonic function is still a harmonic function.z) (48) where the B.)h(k.j are constants. in general.. Commonly used harmonic functions arc (46) Any two of these arc linearly independent.'Y _ + k!Z ' dZ dz 2 = 0 where. (42) and (44) it is evident that f •. f ~ LLB••••f .1:. (41) separated into the trio of equations d'X dx t d'Y + k 'X _ Z 0 0 (44) dy' 2 + k. The elementary wave functions corresponding to specific eigenvalues are called eigenfunctions. . . Any solution to the harmonic equation we shall call a harmonic junction.ion to Laplace's equat.. The values of the k. Equations (44) arc all of the same form. by Eq.y)h(k. (43).. R••••h(k. k.144 We DOW TnIERARMONIC ELEcrnoMAGNETIC FIELDS have Eq.. (45).h(k. by h(kzx).! and denote it.. These solutions llre called elementary wave functions.ion.LL . . This is not the present meaning of the term.
)h(k.Ie. they represent localized standing waves. say k = .k. The choice of the proper harmonic functions in any particular case is largely a matter of experience..) Note that the degenerate caso k = 0 has the harmonic functions h(Ox) = 1..ves which are attenuated or augmented according as 1m (k) is negative or positive.. 01:. Solutions of the form h(kx) = r ih (k positive real) represent waves traveling unattenuated in the +z direction. The trigonometric and hyperbolic functions are then just specializations of the complex harmonic functions. IIf(k. the above two harmonic functions represent evanescent fields. we have +z traveling wa. (Re (k) > OJ represent z traveling waves. We should get used to thinking of the various functions as defined over the entire complex kx plane. and the integr3. Similarly.) dk. solutions of the form h(kx) = fib. according to Eq.tion is over any path in the complex k". 42. We shall see that solutions for finite regions (waveguides and cavities) a.NE WAVE FUNcrION8 145 StiU morc general wave functions can be constructed by integrating • over one or two of the k a For example.)~.1ues. and we say that there exists a continuous spectrum of eigenvalues. If k is complex and Re (k) > 0. If k is purely imaginary.0).) is an analytic function.. (49) where f(k".•••• dk. Consider an elementary wave function of the form (410) .ja with a real.y)h(k. a solution to the Helmholtz equation is ~~ !l k.riation of the separation parameters. while solutions for unbounded regions (antennas) often require continuous spectra. and Ie" domains. Equation (49) exhibits a continuous va.ja with a and fJ real is used. (49) are most commonly used to construct Fourier integrals.PLA. If k is purely imaginary. and facility in this respect will be gained as we use them.k. Wave functions of the form of Eq.<)h(k.. If k is complex. (The convention k = fJ . dk.re characterized by discrete spectra of eigenv3. then the" trigonometric functions" sin kx and cos kx can be expressed as Hhyperbolic functions" sinh ax and cosh ax. Table 41 summarizes the above discussion. attenuated or augmented if k is complcx. dk.. Solutions of the form h(kz) = sin kz and h(kx) = cos kx with k real represent pure standing waves. We should be familiar with the mathematical properties and with the physical interpretations of the various harmonic functions so that we can properly choose them for particular problems. Plane Waves. (14.%. f(k. Keep in mind that wave functions represent instantaneous quantities.
\: complex ...I: rn.z (412) (413) we can express Eq.ink... . lin 11% Evanucent licld AtteDulUed trlveli". PaoPERTIES or TIn: HARMONIC FUNCTIONS' M. Note that in terms of k and the radius vector r=u. We now have equiphase surfaces perpendicular to lJ and equiamplitude .. ::0: ...h: ....... b . Loc&li..I/f..Vel COlI .ity. This gives y = v(jk.. ted traveling . Equation (413) therefore represents a scalar uniform plane wave propagating in the direction of k..a• .j.....in fb lotIh . ~his columD &ive... 0..Ie. For k complex. i _ ." SUndinl ... with itself.. the harmonic (unetlollll t For I" _"ti. + u. b(n+H)r k" .b\ Ze~t I!lfi .% COlI It" cosh ... must satisfy Eq.inh .lil.".. For k real.j .. J . .. The amplitude of the wave is constant (unity)... . 8% co. we define two real vectors k = ~ . ~h" ..ula. .ympwu" behavior..e Two lI".... ·.. . Bln. (45).. we apply Eq..ARMONIC ELECTROMAGNETIC FIELDS TABLE 41...l:rnl k imlll'inary k\lOmplu eo.. ."... .. (414) and determine the vector propagation constant a.. .IlNCC'" field.. t.. ±j.. ... " .x+uwY+u.l: complex . + uwk...l: real k Im. k. (4lO) as y.eent field AtUnu ...I'. n . the equiphase surfaces are planes perpendicular to k..y .. which is of the form of the scalar product of a vector (411) k = u"k".146 TIMEH.i (OfIB.k...ed atandin.\:. (2 140) and determine the vector phase constant ~ ~ V( k· r) ~ k Hence.. The Ie.flo Ie !"(lal i" . tions of .. . ainh ...61<> Speei_\ l"CJ)r<:eentatlollll Ph)'. j ttaveLin.ifJ..t ." StIDdi.. f_ A: Im'fri""Y . (2145).. ltl~t Sped. . ave Two ev...neacen~ fielde LocalilO8d 'landi"l WIIVeI • For ... Figure 41 illustrates this interpretation.II".....h .ccording to Eq..i~a1 interpretatiDa .:II +i...'" " .....r) =jk +j~ =. ±i 'inll oJ: .....in(jr. +z tT.. h . M = eft·.vclinli: wave Evane.. &l"$ 1I(0z) ..1 1 imacinary k eomplu ...
be positive real. \} must.p. k "'"' {J'  a'  j2a . When a: and 0 are mut.ik.  + u. (386). H is perpendicular to k by Eq. which.. (48) and (49) are linear combinations of the elementary wave functions. Fields TM to z are obtained if 'It is interpreted according to A = ua.vc.uJk'" = v'It X Ua + u.k')~ For k real. Thus. for lossfree media. Flo.{uJk. Eq. since sinusoidal wave functions are linear combinations of the exponential wave functions. (410) or Eq. (413) are quite general. k' = k . either 0: = 0 or a: • I) = O. (415). X k (415) (416) and UE . In fact.kJ< + u. We therefore conjecture that all wave functions can be expressed as superpositions of plane waves. since Uk· E . + uJk. become H .ik'" = Nu.. A uniform plaDe w .)~ + n.ja and k "'"' It .k' + kJ<')~  0 .ually orthogonal we have an evanescent field.jk" do not imply that {J equals k' or that a equals k" in general. When a: "'" 0 we have the uniform plane wave discussed above.PLANE WAVE FUNCTIONS 147 Equiphase surface Direction of propagation Z . such as was encountered in tot"l reflection [Eq. lIenee. and E is perpendicular to k. (262)J.s a plane wave propagating in the direction of IJ and attenua. This choice results in Eqs. (410). when k is complex. 41. Wave functions of the type of Eqs.jk. (4~1O). It is a uniform plane wave only if ~ and a: are in the same direction.ting in the direction of a. Let us now consider the electromagnetic fields that we can construct from the wave functions of Eq. for the 'It of Eq. The elementary wave functions of Eq. Note that definitions k """ lJ .k'>/(. (413) represcnt. / / y x surfaces perpendicular to a.(k.
fJ = k'. the wave is TEM to the dirootion of propagation to z). It will be TEM to {. trary electroma.')f The boundary conditions 011 the problem are that tangential components of E vanish at the conducting walls.k. the wave is not necessarily TEM to the direction of propagation. Eqs. n are the direction cosines. (415) and (416).rmonic functions having two or more zeros are the sinusoidal functions with k i real. For k complex. For k real. An arbi.2.i'. 43. a = kif. m = 1. and y = b. It is conventional to classify the modes in a rectangular waveguide as TM to z (no H a) and TE to z (no E. .x ~ k a .2. IH ~ (k. that is.x)h(k. m.veguide provides a good illustration of the use of elementary wave functions.3.\ by Eq. n real.\ are in tion. In this case. Hence. m. sin k.ly k=~ • b n = 1. Modes TM to z are expressible in terms of an A having only a z component ¥t. (414). h(k. (389) apply."'..ja = (uJ + ulJm + u. For k complex.y)..  In particular.x) ~ siu k. We wish to consider traveling wavesj hence we consider wave functions of the form f ~ h(k. . this is n. In this section we shall consider the complete mode spectrum. Its polarization is orthogonal to the corresponding TMtoz wave. The problem of determining modes in a rectangular wa. E. and l.148 TIMEHARMONIC ELECI'ROMAGNETIC FIELDS Thus. 27 we considered only the dominant mode.ly) .k + u.). E. if k = l} . that the wave is not necessarily TEM to the direction (that of (. 216. ~ (k' . The Rectangular Waveguide.n)k (as well as TM It then follows of propagation the same direc with l. In this case. The geometry of the rectangular wa. (386). define a: and (.\). In Sec. x = a.veguide is illustrated by Fig.gnetic field in a homogcneous region can be considered as a superposition of these plane waves.\ only if a and (.· (418) The electromagnetic field is given by Eqs.k')f (417) which are dual to Eqs. wave TEM to k and TE to z. must be zero at x = 0. Thus. .. All these fields are plane waves. The dual procedure applies when'" is interpreted according to F = u. The only ha. Y = 0. choose h(k. giving E~#kxu.3.
2. . k ~ n k _ n.. .1.. n = 0. Again.. y '"'" b. . .. modo fields are obtained by substituting the l/t. Hence.• a.... The TE.co. (45)] becomes (m. Equation (420) determines the mode propa.. (389). the TE. 3... The separation parameter equation remains the same as in the TM MSC [Eq..nd n "'" I. or mode. I. 1. ." mode functions are If. are satisfied.2. so the If must be of the form of Eqs..)' = (k. Interpretation of each mode is similar to that of the dominant TE ol mode. . For k real.. specifies a mode. . . considered in Sec. the propagation constant vanishes when k is ~("'aT)' + (".. (418)... Each integer m and " specifies a.and the separation pal"11m~ eler equation {Eq.) . mode fields are obtained by substituting the If.. 1. In particular. "TB into Eq•. b Each inl.2. m=O.' _ k' (386). (420) The TM. possible field.2. (389).. m = n "'" 0 excepted. .)' + k. The TM.gntion const.nnt y "" jk. E~ • ay %  the first of which must vanish at y "'" 0. nry a 81D b e iIt " (419) with m . .1. "TM into Eqs.I.be' (421) with m "'" 0.. Modes TE to z are expressible in terms of an F having only a z component J/I.)' + (n.nTE = cos m'lr% a cos nry . 27..cgcr fn and n. (420)]. 2. and the second at % = 0. .PLANE WAVE J'UNC'I'lONS 149 that the boundary conditions on E.. a.2. 7rWCU funclicm are therefore 50 . .'nI: = sm  mn . The electromagnetic field this time will be given by Eq•. except m = n "'" 0 (in whieh case E vanishes identically). Harmonic functions SiLtisfying these boundary conditions are h(k... k.... 3•. n=0.)•• (422) . we wish to find traveling waves...
Waveguides are usually constructed so that only one mode propagates.. We do not wish to operate too close w I~t since the conductor losses are then large (see Table 24). suppose we wish to design an airfilled waveguide to propagate the TEol mode at 10. It is apparent that "Y = jk. TM... foursome of degeneracy. TE... for k > k.:7:'W'F'E:iill • y(mla)' (nib)' (425) In terms of the cutoff frequencies... ~'" . It is the physical size "(compared to wavelength) of the waveguide that determines which modes propagate.) To illustrate the use of Table 42. "'"' jk. (2S6). since. (285) and mode phase velocities by Eq.2 ~ ~(~)' + (~)' + (424) and the cutoff wavelengths 2. = where mode indices mn are implied. .)' I> I• (426) 1 <I.ISO TlM'EBAlUlONIC ELECTROMAGNETIC FIELDS The (k..) •• . (422) we determine the cutoff frequencies (j. where mode indices are again implied. tabulation of some of the smaller eigenvalues for various ratios bla. From Eq. the mode is nonpropagating (evanescent). In the sqU::Lre guide (bla = 1).)•• is called the ctdoff wave number of the mn mode. the mode is propagating. We can also define mode wavelengths (or each mode by Eq.. ~ I j~ . hence b/a > 1 usually.(J... values of k... If we take l j~ jk 1  a"" kc (J)' ..yo: :c1 . we can rooxpress the mode propagation constants as ... and for k < k....). and TM. Table 42 gives &...." .j yk' a = (k.. The corresponding TE. we have a 2:1 frequency range of singlemode operation.2. if the guide is too thin. 2 (Xc). they are said to be degenerate modes. (The breakdown power is proportional to for fixed b.. and TM. for each mode has the same interpretation as "Y for the TEol mode. and this is the most common practical geometry..k' Thus. modes nrc always degenerate in the rectangular guide (but not in othershaped guides). For b/a .000 megacycles (>" 3 centimeters). •• For otber (423) ~ jk. Whenever two or more modes have the same cutoff frequency. arcing may occur.2. = k = /T.)••' (k. we have . modes form &. the TE. It is undesirable to make b/a greater than 2 for highpower operation.
Rectangular waveguide mode patterns.) Figure 42 shows sketches of crosssectional mode patterns for some of the lowCI''Oroer modes.606 TEll TM l1 2...325 • • • • b = 2 centimelers. b (k... (J~) . Flo.803 2.472 TEll 3 3 3 3 3 1. In addition. (386) and (419) or Eqs..1 (/. . 217.. . we determine E and H from Eqs.). '' ' . .tterns (field lines) a.. _ (>.236 3.162 4. The mode pattern is a plot of lines of t and 3C at some instant.123 6.. :JC from Eq. (")... (A more direct procedure for obtaining the mode patterns is considered in Sec. at.).. The TEll and TM l l modes become propagating at 16.4 centimeters for the TEol mode.!J( . and we are operating welt above cutoff. it actually loops down the guide. TEll FOR TIlE RrCUNOULAlt >c  b • 1 TE•• I 1 1 1 1 TEll 1 l. (e) TEI2 e )Io.606 4.).. When a line appears to end in space in these patterns.000 megacycles. 42..083 TEn TM n 2..5 2 3 2 3 • • • ..re also of interest. a frequency of 15.236 3.162 TE H 2 2 2 2 2 TEll 2 3 4 TE u TM II 2.. and then determine S. 6 6.828 3. (d) TEo. 81..770 megacycles. '/ .h.500 2. (389) and (421). . The mode pa.. A more complete picture is shown for the TEo l mode in Fig... each mode is chnrocterized by a constant (with respect to (a) TEo1 (b) TEn (c) TMIl ~~' ..5 TM II 1. The next modes to become propagating are the TE IO and TEn: modes..828 3. (1. For this.. (141). and so on..PLANE WAVE FtiNCTJONS TABLE 42.\VEQUlDE.' . ~"r:f ". _ 151 W.. then}..236 2.414 1.
hat the product (Zo}••T&(Zo}. (428) f < f. . for the TM•• modes. H" E. Figure 43 illustrates this behavior. instead of Eq. we h:lve from Eqa. (386) and (419) j~. Thus.ve from Eq•. 11. and the TM characteristic impedances are capacitive.. The classification of waveguide modes into sets TE or TM to z is important because it applies also to guides of nonrectanguln.  jk.ve impedance.herefore f > f. If..nd (421) y) zdirectcd wa.N. we h. However.. (384). jwtElI = jk. Alternative Mode Sets. 44.. the TM. we choose (429) . more convenient e1a. ]WIJ. and the TM characteristic wave impedances are always less than '1.ll. TIMEHARMONIC ELEC"rROMAONETIC FIELDS For the TE•• modes in lossfree media. (427) f < f. Attenuation of the higherorder modes due to dielectric losses is given by the same formula as for the dominant mode (see Table 24).152 2:. = 1·kI 8~ == J·kE fJy •• The TEa.r cross section. k'l~ .0. It is interesting to Dote t. P 8• JWIJ ... :t = jl. P < k for propagating: modesj so the TE characteristic wave impedances are always greater than 'I. the TE characteristic impedances aro inductive.n1 = 'It at all fr~ quencies. Attenuation due to conductor losses is given in Prob. characteristic wave impedances are t. ·H '" "'" 1·k8~ = 1·k. For non propagating modes. (389) .ssifications can be made.: = jk. 44. By Eq.. characteristic wave impedances are (Z) ~ 0 •• Wf E. H .. (426). Similarly. for many rectangular waveguide problems.a JW' f > f. We now consider these alternative sets of modes.
we have an electromagnetic field given by a set of equations differing from Eqs. Similarly.. 43. the field is given by H.. Hr=. "jXo \RoTE . we chooso (431) we have an electromagnetic field given by E __ .0 II ~ . az (432) ay . a. / _Ko TAt / " X.• ay This field is 'I'M to x.vcguide modes. Chnracteristic impedance of wfl.. E _ • a. To be specific. ~ {Ro tt < t. a. instead of Eq.TE/ " ':. (387).TM \ o 1/ / 1/ I 2 3 FlO. a. ..PLANE WAVE FUNCTIONS 153 \ Zo 1\ > t. y. if. z. (430) a. (386) by a cyclic interchange of x.
modes) we have (434) where m . Note that our present set of modes have both an E. (435) f <f. while tbe cutoff TMx. (430) and (432). and H.modes. m ~ 0. n = 0. . The mode patterns of these hybrid modes can be determined in the usual manner. ' . modes. (432). The electromagnetic field is found by substituting Eq.YONIC ELECTROMAGNETIC FIELDS This field is TE to x. to illustrate the character of the higherorder mode patterns. (426). and the TMxo. the cutoff TMx.2. 312. H. The characteristic impedances of the hybrid modes are also of interest. and k. 216) is relatively simple. .2.154 TUdEHAR. :JC. then S. For modes TM to z (TMx•• modes) we have (433) where m .. . (430). The choice of ¥'s to satisfy the boundary conditions for the rectangular waveguide (Fig. is given by Eq. mode patterns are those of the TEo.. (except for the Oorder modes). mOOes are the TEo. mOOes of Sec. Note that {or (l small. (434) into Eqs.3.. . . . (426). The field is obtsinOO by substituting Eq. . modes have inductive Zo's.". {rom . and the TEx.0. n = 1. is again given by Eq. and k.o modes. For modes TE to x (TEx.. Similarly. .. . . According to the concepts of Sec. and specialize to some instant of time. we have from Eqs. o mode patterns are those of the TE.jk. Note that the TAu. 1. 2.) The TEa. All other modes of Eqs. . For the TM% modes. (430) and (433) H. (Determine E.. have capacitive Zo's.~ Hence. .1.a: u and TMxll modes. o modes are the TE. 3. (433) into Eqs. 43. . aD arbitrary field can be C!'nstructed as a superpositioo of Eqs... Figure 44 shows the mode patterns for the TF. 2. modes. Such modes are called hybrid. .. (433) and (434) are linear combinations of the degenerate sets of TE and TM modes. the zdirected wave impedances are f > f. 1.
(m.. /a)' T. and then the cavity modes arc standing wa. II%.~~ .. '! I \ . ! ~. The geometry of the rectangular cavity is illustrated by Fig. (436) f < f..'./a ond ny/b interchanged. z properly interchanged.E. It is conventiona. Eqs.30) and (432) with x. Hybrid mode pat. 44.terns..ves of the usual TE and TM waveguide modes./a)' = Note that for a small. (432) and (4.  .. ~~_ / Iif h I' 11 . We shall now consider the complete mode spectrum. (4. The Rectangular Cavity.. We considered the dominant mode of the rectangular cavity in Sec.z kt ~."..tes. the cutoff TEx. %j SO we can express the fields as TE or: TM to anyone of these coordina. . 28.. (419) . y. The fields would be given by equations similar to Eqs. kt (m.'34) we find Z) ~ ( 0. Y..a. \ 1:0 A I I . The TMy and TEy mode functions would be given by Eqs.~I>:~=. Sets of modes TM a.P k' (m. (b) TAbu PSG. The wave functions of Eq.a)' f> f. 4'S.k. 219. :.".. (433) and (434) with .nd TE to Y can be determined by letting A = U~Vt and F = U1/1/I. The problem is symmetrical in x.\ x P~E WAVE FUNCfIONS 155 x t ~::: (a) TExu rTTTT' 111 '~II bY'/'l>'1 ~~~S§i ~ 1 1 \ \..l to choose the z coordinate.... respectively." modes all have inductive characteristic impedances.
(438) for the resonant frequencies (440) For a < b < c1 the dominant mode is the TE oJ1 mode. .00 3. The separation equation remains Eq.00 2. the TE••JI' mode fUDctioDS Brc given by . m'll'Z .91 3.. and Eq.I . we solve Eq.. 2.2.84 1. AB indicated by Eq.0 excepted. ..1.2.00 2.58 1.. (386).58 1.05 3. 1 • • 1 TEln TE II1 TM no 1 TM III TEl..19 3.58 1. .00 2.12 2.3.26 1.'.p/(J. The TE.01 1. j n = 1.00 2. .82 4.35 TE llt 1. .91 7.55 2. .SID . This is evidently accomplished if ~TK = "'''1' SID . (420) becomes .2.58 1. (438).. I.84 1. . a  <b <C  b . each mode can exist at only a single k.53 3. im = n .26 1.16 3.76 1.2.88 4.58 1.58 1.20 7. b.00 2.26 1.z) to satisfy this condition at the remaining two walls. Setting k = 2r/ W. .156 TIMEHARMONIC ELECI'ROMAGNETIC f'TELD8 satisfy the boundary condition of zero tangential E at four of the walls. It is merely necessary to rcpick h(k. nry abc COB  prz (437) with m = 1. Note that (J. Table 43 gives the ratio (j.25 1.00 1.91 2.03 4. p .COS  mrX nry. 43.1. . prrz a bsm c (439) with m = 0.3. (389). .58 .58 1. 73 1. (438) The field of the TM••.71 7. .Till: "" • •1' = C08 .00 3.91 3.3.84 2."VITT. (437) into Eqs.66 4.ulT8tll TEl" TE III TM I11 TM m TM m 1.84 2..73 2.. mode field is given by substitution of Eq. p "'" 1. .22 1.24 3.84 1.68 5.58 2.83 ..) .65 3..08 1. TE.16 4. in = 0. . c.62 3.2.91 3. given a.. TABLE (fr)tlI Fon TilE R ECTANC':t!LAn C .)".60 5..65 3.34 1.96 1.71 7. 1.)oll for cavities of various side lengths. Similarly.0.58 2. mode is given by substitution of Eq.26 1. (439) into Eqs. (438)...13 2 2 4 4 4 2 2 4 4 8 16 1 1 1 1 I 1 I 1 1 1.58 1.58 1.26 2.
000 000 I o " ". Fig. still other degeneracies occur. When two or more sides of the cavity are of equal length.. The greatest separation between the dominant mode and the next lowestorder mode is obtained for a. The Q's due to conductor losses for the various modes are given in Prob. the second resonance is = 1. given by Eq.5. The Q varies rougWy as the ratio of volume to surface area of the cavity. (2100).• . The Q due to dielectric losses is the same for all modes. z = constant plane. The most significant difference between the waveguide patterns and the cavity patterns is that E is shifted from JC by 'A Q/4 in the latter case. The mode patterns of the rectangular cavity are similar to those of the TE or TM waveguide modes in a. are always degenerate. r a " .. Section A . and vice versa.. Also. The TEoll mode pattern is shown in Fig. 410. 10 modes. 4. Rectangular ca. To illustrate higherorder mode patterns. and similar to the hybrid mode patterns in the other two cross sections..PLANE WAVE FUNCTIONS 157 the TE"'''10 and TM. 45 shows the TE 123 mode pattern. 220.. " . 8 and :JC are 90° out of phase in a cavity. In this case. mnp all nonzero. Note that the Q increases as the mode order increases.vity mode pattern for the TE iU mode. squarebase cavity (b = c) with height onehalf or less of the base length (b/a ~ 2). since the energy is v% . 28 for the dominant mode. so E is zero when :JC is maximum. The quality factor Q of each cavity mode can be determined by the method used in Sec.. " 0 0 0 L go Section C Section B FIG.58 times the first resonance.
t  (442) ktt = ".. A pa. Phy•.. The problem contains two homogeneous regions. . (429).1". L. An attempt to find modes either TE to % or TM to % will prove unsuccessful.. . 46. . region 2 is x > d) to represent the x component of A. 1944.. Pincberle.y eit •• :IC with n I. . The separation parameter equations in the two regions are k£l t t + (nT)' + k. as in Eq. as we now show.tEJJ. n'lrtJ 'f'1 (441) 1/It "'" Ct eos [kd(a CII . (430). 46.o case. 4·6. Electromagnet.x)] sin n. An attempt to find modes TE or TM to % will prove successful. ". It has been anticipated that Ie" nff/b and k.. Most modes arc therefore hybrid.•..ll k ot nT)' + (b + k. .rtially di· electriofilled rcclanlU1ar waveguide. d (or has two dielectrics).. < d and d < z < a. 5. o b y stored in the dielectric and the losses are dissipated in the conducting walls. 66.Ii " . Such problems are solved by finding solutions in each region such that tangential components of E and H are continuous across the common boundary. This is illustrated by Fig.AONETIC FIELDS x z .158 TIW. and B. must be the same in each region for matching tangential E and H at x = d. o < :r. pp.. 3. t b  kit"...ic Waves in Metal Tubes ruled Longitudinally with 00.1X sIn b. = C1 cos k. Partially Filled Waveguide.( d Flo. we take . 118130. l Consider a waveguide that is dielectric filled between x = 0 and:z.tEU't 1 Two Dielectrics.EllAlU:IONIC ELECl'RO!ol. vol. having both E. To satisfy the boundary conditions at the conducting walls..". For fields TM to x.. except for the TE. 2.. The field in terms of the vis is then given by Eqs. RG. we choose ""s in each region (region 1 is x < d.
(442). is found.'s in each region to represent the z component of F. Bin k d % sin nb1r'Y trfk·· .. b and E. is given by Eq.x)] cos b r JWf. k.• Continuity of 11~ and H. by Eqs.. (442)i so the above is a transcendental equation for determining possible k. at x ...!.)] sin n"!/ ".d tl . we choose "".wIl) snd (441) we calculste 1 nT.d») Division of Eq. (430) and (441) we calculate 1I~1 = jk.k..d)] tl Et (445) Both kd and kill are functions of k.. d requires that .v e (446) ik .. sin [k•• (a tt a)] (443) Similarly.t' r i4 . To satisfy the boundary conditions at the conducting walls. C"'.I<. k.v r ik .tX cos b "". sin lk. ..C 1 cos klll% sin n. we take ill =..." ~T Ct cos {kdCa .. (444) gives (444) kill tan kll1d """ _ kill tan [kd(a ..y = . C1kdk.t ll •• E' l = Ed .'s (modepropagation constants). at x = d requires that CI cos kdd = Ct cos (kd(a .. nq E~l = .z)J cos n~ rPo.. kd and kd are given by Eqs. Once the desired.! Csk. snd the ratio C.jC. For fields TE to z. ..%)1 sin n~ rfl·· Hd  ~T Ct cos k%l% cos n.v ell•• H~t .C 1 sin kdx "'t  cos n.Ctkat 5 8m [kat(a . J...Continuity of E~ "'" 1.jk..(a ~ . C. (443) or Eq. 'IWE I E~t 1 nT . sin k. (443) by Eq. from Eqs. n. (444)..PLANE WAVE FUNCTION8 159 From Eqs.(a .. Ct sin [klll(a  z)] cos n...C 1k lO1 b SlD. (.Ct cos {k". Hd ..
. the next modo by m = 1. (445) can be approximated by ~ k. as follows. 1 a. (given Cal).. (426)..(o III d)J 111 (4471 The kd and kd arc functions of k.. Uaing Eq. The field is calculated from the by Eqa. A matching of tangential E and H at % . emptyguide modes. lIenee. is zero in an empty guide).r and k.. For b > ". The mode with the lowest cutoff frequency is denoted by m = 0.lly filled rectangular waveguide arc distorted versions of the TEa and TM% modes of Sec. The cutoff frequency is obtained hy oetting k. (442). =.re given by Eqs.k". then Eq.0) of tbe various modes will always lie between those for tho corre· sponding modes of a. Note that when kd is real. and 50 on. In the lossless case. (442).. . 44. The dominant mode of the partially filled guide is then the TMxol mode when b > a.451'1). and vice versa. When k l is DOt very different from k l .'s for the modes TE to :t.0_d:=!) ( 1'1 (448) With this explicit relationship between kd and kd . whicb is also the TE u mode of the empty guide. Tbis numbering system is chosen so that the TMx. (445) has a. the cutoff frequencies of the corresponding TE:t and TMx modes will be different. The modes of the partia. Of special interest is the dominant mode of a partially filled guide.d _ _ k. (442). The separation parameter equations are again Eqs. at otber frequencies by Eq.. denumerably infinite set of solutions. (445) and (~7) at each frequency. cot[k. the cutoff frequencies (k.. guide filled with a material El.0 in Eqa. (445) when the k. (442) with n = 1. For a given n. we should expect k. we can solve Eqs. 44. a. Eq. (432).scendental equation for determining k. If this is 50.) In contrast to the filled guide. cot k. . Also.'s a.'d 1'1 _:. this is the mode corresponding to the TJ\uol mode of the empty guide. (442) simultaneously for k. I. (448)..z is imaginary. and thoseof a guide filled with a material E" PI. 74. so tho above is a tra.. 2. by Eqs.d yields.'". Fign. The modo patterns are similar to those of Fig. knowledge of the cutoff frequencies of the partially filled guide is not sufficient to determine k.re 47 shows some calculations for the case E ". we have for the .z to be small (k. Ill...~ tional procedure of Sec. partially filled waveguide modes correspond to the TMx.the characteristic equation ""s k.nd k. 2. We have to solve Eqs. .. (This can be shown by the pcrturba..160 TIMEHARMONIC ELECI'ROMAGNETIC FlELDS with n "'" 0. except that tho field tends to concentrate in the material of higher E and Il.. the propagation constant of the dominant mode is given by the lowestordcr solution to Eq. k. We shall let m denote the order of these solutions.
rtially filled guide will a. or TE lo mode.0.45.2 0. :«i bl / T. .6 Fla. When J11 == J11 == this reduces to (450) Note that thi. ~ 0. valid when Eq..45. and baving traDS~ L == J10 ~ == EIES fl(a .O.lso be a .4 0. shorted at each end. A waveguide partially filled in tbe opposite manner (dielectrie boundary parallel to the narrow side of the guide) is tbe same problem with a > b. The dominant mode of tbe pa..1.3 0/).d) + Etd per unit width.) dominant mode kd ' l + (i)' = WlflJll fl(~~d) k + (i)' "" zl l W ftJll These we solve for the cutoff frequency r ColC"=b fl Col == We.1 .2 ~ • 1 t.2. alb . 47.rtially filled with dielectric. obtaining J flea d) ( d ) fUll a  + + ftd E2UfIJlI (449) J1.5 0.s is the equation for resonance of a parallelplate mission line. All cylindrical (cross section independent of z) waveguides at cutoff are tw<Hlimensional resonators.6 1.0 0. 0. (Alta Prank.8 / 0. • . dla .4 o 0. The dominant mode of the empty guide is then the TEz u mode. PLANE WAVE FUNCTIONS 161 /' . (448) applies.SO. Propagation constant for a rectangular waveguide pll.
die Ie 1  toj ~l r7k:: . mode of (0) ia the reIIOnAot frequency of (b).6 o{>. 0.2. 0. For the cutoff frequency of the t:= dj T 1 b (oj 0 "I (2. (447) could then be found by perturbing k.8 fl.4 0. and 50 on. • . (447) with n "'" O. 49.. When k 1 is not too different.V V  . This numbering system corresponds to that (or the empty guide.'12 13 :: k2 I Id 1 (bj 0 dl I FIo. 1.5 0.162 TWEHARMON1C ELECTROHAONETIC FIELDS TEz modej SO the eigenvalues are found from Eq.451.0 FIo.?=' ~ ::. that with the next lowest by m = 2. the dominant mode being the TEx lO mode. from k" we might expect k Z1 and kd to be close to the emptyguide value k.6 0. 1 and kd about 7/a.2 0.1'2 Zo  'II I {j '""' kl I "I' Zo . An approximate solution to Eq. o ~I 0.8 .) . The cutoff frequency of t.8 1..he dominant. / . = ria..6 12 Idi . Propagation constant for a rectangular waveguide partially 6UOO with (After PraM.. (b) tran. That with the lowest cutoff frequency is denoted by m = 1.375 0. We shall order the modes by m 88 followa.amission·line resonator.167 0 0... /V f/ '/1# ~ Iff l.280 0. 48.4 o dielectric. (a) Partially filled waveguide.
Eqs. For modes TM to 2. Such phenomena also occur in inhomogeneous dielectrics. 49 for the case «: . The simplest illustration of this is the guidance of waves by a dielectric slab.2.p an odd function of x. 419. 1 Some calculated propagation constants for the dominant mode are shown in Fig.!."'" ax (452) We shall consider separately the two cases: (1) . rj!~ variation. It is not necessary to have conductors for the guidance or localization of waves. (442) becomo kd and Eq.cot k1cd = . 48. 410. MIT Rad. at eutoll. It is desired to find ztrllveling waves. This viewpoint has been used extensively by Frank. the TEz lo mode reduces to the parallelplate transmissionline mode that propagates in the z direction. For case IN.45«:0. (k' _ k. Rept. Similar results for a centered dielectric slab arc shown in Fig. denoted by 1/t'. H. dlJ (451) It is interesting to note that this is the equation for resonance of two shortcircuited transmission lines having Z. L4b. 410. The Dielectricslab Guide. The socalled slab waveguide is illustrated by Fig. _ )we .'s of '11 and 'Ill and P's of k 1c and k'k. dominant mode.')" H.. The reason for this is. and we shall choose the latter representation.. a. allowing no variation with the y coordinate.nd (2) '" an even function of x..PLANE WAVE FUNCTIONS 163 x Flo. Frank.7.1942. . that is. denoted by ~.(a  I I ". The dielectriclliab waveguide. . . _ ... Eqs. 9. We 2e 2/JI ". Modes TE and TM to either x or z can be found. (386) reduce to E = k. Wave Guide Handbook. as illustrated by Fig. '" WE a" ax E. and the characteristic equation for that case is given in Prob. 4. 7·10. (447) becomes k"'ll l = kl~2 = wc2el~1 = k 2c I .cot (k. We shall consider the problem to be twodimensional.
. 2 = k ol = 1 1 W (dJJd ". (455). [xl < ~ 2 11" = . coupled with Eqs.1£0#010 (455) Evaluating the field components tangential to the airdielectric interface. .Au cos u:c e E. = JWEd ~ u ' sin tLX ep..v1e""eil .s ]WED Continuity of E. we have E. = kd t = k. (It will be seen later that u and v are real for unattenuatcd wave propagation. = . } jk . is the characteristic equation for determining k/s and eutoff frequencies of the odd TM modes.. we choose in the dielectric region [xl and in the air region <2 ~ a a (453) x>'2 x (454) < 2 We have chosen kuj = u and k&O = jv for simplicity of notation. and HI/at x = ± a/2 requires that The ratio of the first equation to the second gives ~tan~ = ~va 2 2 EO 2 (456) This.164 TIMEHAUMONIC ELECTROMAGNETIC FIELDS (I). JWEo 8 x>2 a E..) The separation parameter equations in each region become u1 _Vi + + k.B v1e""eik.. .
Dielectric guides operated in a radiating mode (below cutoff) are used as antennas. obtained from the +'s by equations dual to Eqs.± 0{2 yields . a~ • WI! ax fI. tbat is. of course. The fields are.... In this case. there is attenu~ atoo propa. which are. _ . For the TE modes with odd 1/t we have ua ua tan 2 2 ~ Il4va PO 2 ( 459 ) as the characteristic equation. explicitly. and the even wa.. (452). is real). (455).0. and H. The odd wave functions generating the TE modes are those of Eqs.ttenuation must be accounted for by radiation of energy as the wave progresses. "'" fJ . we ehoose (457) The separation parameter equations are still Eqs. (389). (k' JWP _ k.')~ E~ • ax (461) These are specializations of Eqs.ve functions generating the TE modes are those of Eqs. . at :t . The field components are still given by Eqs.tisfy Eqs.cot.k.'s and cutoff frcquencies of the even TM modes..PLANE WAVE FUNCTIONS 165 For TM modes which are even functions of %.t different interpretation than for metal guides. The concept of cutoff frequency for dielectric waveguides is given a somewha. (457).2 2 t6 2 ua ua E.gation (k. matching E. and for the TE modes with even 'It we have _ ~cot 2 ua = 2 1A1l~ lAO 2 (460) as the characteristic equation. Since tbe dielectric is loss free... this ll. Below the cutoff frequency. Above the cutoff frequency.!. as we define it. the dielectric guide propagates a mode unattenuated (k. H _ . The u's and v's still s:l.).= . (452). The phase constant of an unattenuatcd mode lies between tbe intrinsic phase constant of tbe dielectric and that of air. (455).va (458) This is the characteristic equation for determining the k. (453) and (454). There is complete duality between the TM and TE modes of the slab waveguidej so the characteristic equations must be dual.
This we solve for the cutoff wavelengths >. Simple graphical solutions of the characteristic equations exist to determine k. This characteristic is considered further in the next section. I. Equations (455) require that u and v be either real or imaginary when k. The lowest frequency for which unattenuated propagation exists is called the cutoff frequency.~ = 2a n '\I fO~O n IfdJld _ 1 n = 0. (462) and the cutoff frequencies f~ = 2a V fd~d fO~O n "'" 0. When v is real and positive the characteristic equations have solutions only when 'U is also real. is real. The characteristic equations have solutions only when v is real. . else the field will increase with distance from the slab [see Eqs. and it follows from Eqs. Furthermore. Let us demonstrate this . . 2. when These equations are satisfied n = 0.s k. k. the cutoffs occur when the guide width is approximately an integral number of halfwavelengths in the dielectric. at any frequency above cutoff. (462) and (463). Finally. Note thatf~ for the TEo and TM omodes is zero.166 TUfEBARMONIC ELECTROMAGNETIC FIELDS This can be shown as follows. and TE" according to the choice of n in Eqs. in which case v + O. This is a general property of cylindrical dielectric waveguides.. both u and II are real. it is evident that cutoff occurs a. The result is tan (. (463) The modes are ordered as TM. (462) that when fdJld» EQJto. observe from Eq. In other words. as the slab becomes very thin. zero half~wavelengtb included.2. so the field extends great distances from the slab. II must be positive. 2. the cutoff frequency of the dominant mode (or modes) is zero. ylkd ' ko') = 0 cot (~ylkd' kO') = 0 which apply to both TE and TM modes. . The cutoff frequencies are therefore obtained from the characteristic equations by setting u = ylkd' ko' and v = O. < ka. (454) or (457)J. This result is So property of cylindrical dielectric waveguides in general. 1. 1. the lowestorder 7'E and TM modes propagate unattenuated no maUer how thin the slab. From the above discussion. Hence. However. + ko and t1 + 0. . (455) that ko < k. ko.
\ I'. \ 1/ .''" . ou l Elimination of k.) (U~. Grp.\\ "2 .'ua cot ua .. As w or e" is varied.u.. These can also be interpreted as the mode patterns of the TM o and TEl modes if g and 3C are intercba. from Eqs. I ~ ~2 cot 2/~ ..ll.)  (ua)' "2 2 Values of uaj2 for the various modes are the intersections of the plot of the lefthand terms with the circle specified by the righthand term. . / I I I I I / / I I I [ I I I /i I . The graphical solution for the TM mode eigenvalues is similar. + + . fOf there is complete duality between the TE and TM cases. Sketches of the mode patterns afe also of interest. r • • ~ f  ~ " I [ (UB) (UB) . A representative plot of the righthand term is shown dashed.) Fio.) If Pd F Po.. (U~. only the radius of the circle changes.V"2 ('m . Figure 411 shows a plot of the lefthand terms for Jld = .u. we can write tho TE characteristic equations at'! ua \ I(wa)' 2 _. (455) gives + Vi = kill  l ko = WI(E. the solid curves must be redrawn." 2 ~uatan . Figure 412 shows the patterns of the TEo and TM 1 modes.PLANE WAVE FUNCTIONS 167 for the TE modes. 411.ngcd.. (For the case shown. only three TE modes arc above cutoff.phical solution of the characteristic equation for the Blab waveguide." 2 .iPd  ~llPll) Using this relationship.u.
• I • • • 0 • • • • (b) • (a) TEo mode (X lines FlO. more loops appear within the dielectric. 4~8. This is shown in Fig. 410 can be shown to be reactive.) and the TE. The modes of the dielectriccoated conductor arc those of the dielectric 0 plane. . A simple way of obtaining a single reactive surface is to coat a conductor with a dielectric layer. These are the TM.. n = 1. 4. 412. The wa...s for the dieJectric6lab wll.1. Surfaceguided Waves. We shall show that any "reactive boundary" will tend to produce wave guidance along that boundary.. As the mode number increases.Vcguidc. 2.IAGNETIC FIELDS • 0 • • ~~'4 COCO~ It)D1.1:\1 Q . .3. Mode pfl. ..• modea =0 =0 . .modes (odd if. x It ~~~ • • !t1.ttern.168 TIMEHARMONIC ELECTRO?.ve impedances normal to the dielectrictoair interfaces of the slab guide of Fig. dashed).5. . 0)8) Q'\~r2:.i >t jJ~~ 0 It 0 00 ". but not in the air region. slab having zero tangential E over the x n 0.1 14 l\~t1 (ltZ.. 413. (b) TM 1 mode (t lines IlOlid).
413. A dielcctrieeoated conductor.56E o. if the coating is polystyrene (to" = 2. . k4 . for thin coatings. (456) with 'a/2 replaced by t (coating tl}. The characteristic equations for the 'I'M modes of the coated conductor are therefore Eq.1. (460) with a/2 replaced by t..) . . The dominant mode is the 'I'M o mode.. 414. becomes n f. . (455). and for TE modes n .ickness). For thick coatings. Let us consider in more detail the manner in which the dominant mode decays with distance from the boundary.. 2. We shall retain the same mode designations for the coated conductor.1'0 Z Pta. which. The mode pattern of the 'I'M.. the field in 0.8 per cent of its value at the surface.. (463). and. JI.PLANE WAVE FUNCTIONS 169 (even '/I) of the slab. from Eq. (464) where for TM modes n = 0. "'_k. 3. However..hI the air space. k.he coated conduetor (s lines 1lO1id.  41 V""" <0.. The TM. 4. . 5.4 . x ". "EOJ. The characteristic equation for the TE modes is Eq."'0). . mode is sketched in Fig. o o 0 • • • •• • • • • 0 • • 0 0 • 0 0 0 Flo.IO 1~1 I W&fI (465) This attenuation is quite large for most dielectrics. 414. the' field attenuates as r·%. mode pattern for t. The cutoff frequencies are specified by Eq. For example. . .12~ has decayed to 36. which propagates uoattenuatcd at all frequencies. for the coated conductor. .
ve x>d where The wave impedance looking into the corrugated surface is Z . the field decays slowly.. A corrugated conductor.0001 wavelength thick. + k~. W~O (468) Note that this is inductively reactive. = ju 1I.8 per cent of its value at the surfaco." E. we assume that the parallelplate transmissionline mode . 47.. (The TE fields of Sec. 415. Another way of obtaining a reactive surface is to flcorrugate" a conducting surface. and E.2r o . capacitively reactive surface.. For a simple treatment of the problem. 47 require a. The teeth will essentially short out any E III permitting only E. We say that the field is It tightly bound" to a thick dielectric coating and Uloosely bound" to a thin dielectric coating..)t (466) If the polystyrene coating were 0. we would have to go 40 wavelengths from the surface before the field decays to 36. k. at the surface. 415.. In this case. and that there are many slots per wavelength. as suggested by Fig. the interface must be an inductively reactive surface. hence we shall assume that thia field exists in the air region. let us assume that the" teeth JJ are infinitely thin. we ha. The TM fields of the dielectricslab guide are of this type.flmall Ilo f.) In the slots of the corrugation.. so to support such a field. and v:.}. (467) .170 T!KEJlARI(ONIC ELECTROMAGNETIC FIELDS ]{ z FlO.. Extracting from Sec. k("' .
Consider the rectangular waveguide of Fig.J For kod < T/2. Culler. Electromagnetic Waves Guided by Corrugated Conducting Surfaces. Let E~ = 0 and E. This we shall illustrate for the rectangular waveguide. we can determine the amplitudes of the various waveguide modes. (469) Equating Eqs. we have k. These are then shorkircuited transmission lines. While at this time we lack the concepts for estimating the accuracy of the above solution./ ~ where g = width of gaps and t := width of teeth.) When the teeth are considered to be of finite width. from Eq.iv.)' tan' k. 44 have no Ea.J and. The result is' k.. 49. the wave is loosely bound for very small kod.. it has been found to be satisfactory for small kdl. ~1 + G . the input wave impedance is Z_. except in tho vicinity of the corrugations. Tho mode pattern of the w&ve is similar to that for the TM o coatedeonductor mode (Fig. The TEx modes of Sec. (468) and (470) Vi + tan' k. 414). = k. tan k. (467).tched (only outwardtraveling waves exist). 216. This is found by assuming Eq. Rept. (We should expect E a to terminate on the edges of the ~th. the fields must differ from those assumed in the vicinity of x = d. Note that.PLANE WAVE FUNCTIONS 171 exists. Modal Expansions of Fields.y) be known over the z = 0 cross section. C. for we have only approximated the wave impedance at x "'" d. Given the tangential components of E (or of H) over a waveguide cross section. In the true solution.ting modes are of appreciable magnitude only in the vicinity of sources or discontinuities. (470). of characteristic wave impedance 'l0. MM44160218. The modes existing in a waveguide depend upon the excitation of the guide. (469) to hold over the gaps. an approximate solution can be obtained by replacing Eq. 50 let uS Ie. tan k.. . . (469) by the average wave impedance. BeU TdqMru Lab. "" J(x. and by assuming zero impedance over the region occupied by the teeth. from Eq. we have • = k. Hence. (469). becoming more tightly bound as kod becomes larger (but still less than 11/2). We wish to determine the field z > 0. this is inductively reactive. The nonpropaga.J (471) It should be pointed out that this solution is approximate. assuming that the guide is ma. October. ~ k.. 1944.
The solution for E. A.....E•• = 2t ab ll )0 {" dx r .. (472)._1 . = /(x. 0 is P•• Because of the orthogonality relationships for the sinusoidal {unctions. given by Eq.. m. (423). specifying a field according to Eqs. each mode transmits energy as if it existed alone. 0 and E.. . = O. the zdireetoo complex power at z . when many modes exist simultaneously. '\' L. It is thus evident that 'Y. E..Y) and E.. .= 0 cross section is a superposition of the two cases E• . A . are given over the z .. at z = 0 is given by E~ ..t '\' ... This is A •• sin m. For a large class of waveguides.. The general case for which both E~ and E... I sin mn cos n b ry (473) where f .... the field is given by Eqs. a nry cos b ".z cos n~ rr' (472) where A•• are mode amplitudes and the r•• are the modepropagation constants. 0 given over the z .. givCD over the z = 0 cross section can be obtained in a dual manner..0 cross section can be obtained from the above solut. The solution for the ease H ~ and H. . . = 2 for n > 0 (Neumann'& number). (432).EHAR~(ONIC flLECI'ROKAGNETlC FIELDS take a superposition of these modes..A••• and hence the field.. The ._0 a io dy E. sre the Fourier coefficients of E" or "roo. In particular.... are now evaluated.A .172 TI:M. We shall show that the rectangular waveguide has this property. 0 Note that this is in the form of a double Fourier series: a sine series in :t and a cosine aeries in y (sec Appendix C). Given the wave function of Eq. (432). In terms of !fi._I .. am .ion by a rotation of axes. . = I for n = 0 and E..0 2: 2: = . Lt I 0 "r. 'I' = .
.(tr/a)' "" J(. The above equation is simply a. only the m = 1 term of the m summation remains in Eq. = 'YhAh =  5ln . The dimensions are such that only the dominant mode (TE IO) propagates in each section. the complex power at z = 0 is p E 1. Let us usc this solution to obtain an "aperture admittance" for the junction.c/b)]'j nTc/b k' . nll'c (476) ~ a<:' 2b (Y)' 0 10 + 2 '\' (Y)' ~ • where.. 4. and let the larger guide be matched.. the power for a propagating mode is real and that for a non propagating mode is imaginary. (474) and (476)..veguide junction.. y<c y> c (475) From Eq. For an approximate solution. To illustrate the above theory.. the only nonzero mode amplitudes are E lo = 'Y10A IO =b 2 n~ C b Thus.dmittances. 416.(j ". from Eqs.PLANE WAVE FUNCTIONS 173 Incident wave L z FIG. summation of the powers for the individual modes. given by the reciprocal of Eqs.) .. (436).. A capacitive wa. 0 1>1 ['in (n. (472)..0 sma 0 . assume that Ell at the junction is tbat of the incident wave Ell I { "'" . consider the waveguide junction of Fig. vn' . = k . this reduces to (474) where (yo) ..(~/a)' 2  "" Vi f/ (Uf)' j2b(Y o hG (2b/}". Let there be a wave incident on the junction from the smaller guide. In a lossless guide.JJla }.. (436). are the TEx wave a.~16. (Yoho = wp. (Yoh. (473). From Eqs.
0.< sin! (nrc/b) (nrc/b)' n' (2b/~.656. For small c/b.1 f.3 I' • 'I 2 f I. 2b + 1 ~. Calculated values for B..)' (478) where A. (4_78) to IlIlall clb yields a numerical factor of O.0 c/b· Flo. The aperture admittance is then y • = lVI' = ( p. are those of the dominant mode..Z."~ 0 N I o 02 0.)' ] (477) The imaginary part of this is the aperture suscepta.37\) instead of 0. llre those of the TE lo mode.6 0.]l (479) 1 This equation ill a quuistatic result.174 3 TUIEHARMONIC ELECl'ROYAQNETIC FIELDS +~o~ 1\ 0. are shown in Fig.~. y [a .4 '" 0.20 \ ' sin! (nrc/b) (2b/~. = log I0656~ [1 + 1 (~~). The Ie and X. The direct 15peeialization of Eq._1 (nrc/b)' In' f.he aperture. 417. 0 " "~ ". .nce B • . we have' ~2:' B. We shall refer the aperture admittance to the voltage across tho center of t.8 1. 1\ ~' tE E .\ 1'\\ ~ 1 \. Susceptance of a capacitive aperture. 2a \ '  .2 n. .< • . and Z. 417.sin(1l' % fa} rrt X =r . whieh is V "'" c.J.
Again we can find an aperture admittance for the junction.Tall (me/a)'J (481) Thus.r solution is explicit only for When a second mode propagates. su. the complex power at z = 0 is • p _ 2bc' \ ' (Y)' .ive waveguide junction. 1. Take a wave incident on the junction from the smaller guide... Lt.0 [sin 1 (mTc/a) ]' (me/a)! .PLANE WAVE FUNCfION8 175 tx Incident W<M! tx z Flo. Note that the susceptance is capacitivc (positive) i so the original junction is called a capacitive waveguide jurn. (478) would be summed from 11. 418.·0 ~ { 0 . TZ 81D c x<c Z >c (480) From Eqs. hence ou.. ment of microwave networks in Chap.' since we assumed E in the aperture. 0. From Eqs. only the 11. and Eq. and so on. For an approximate solution. Again we assume only the dominant mode propagates in each section.rger guide be matched. it contributes to the aperture conductance.tion. = 0 term of the 11.'a . 418. . 89 that the true susceptance cannot be greater than our present solution. (472). An induct..) We have assumed that only onc mode propagates in the guide.mmation remains in Eq. Another problem of practical interest is that of the waveguide junction of Fig.. we determine the only nonzero mode amplitudes as E 2c sin (mrc/a) . we assume Ell in the aperture to be that of the incident wave E. and let the la. = 2 to (lO. 8. (We shall see in Sec. (474) and (481). The aperture susceptance is a quantity that will be useful for the treat. Remember that our solution is only approximate. (473).
 .. admittance referred to this voltage is therefore The aperture Y.U'/f)' 'IJ (Yo). we must.2 0. have a < X. 419.8 0. The susceptance is inductive (negative). . " 0.4 0.<...t:y _ 1 m> I The voltage across the center of the a. ~ 1::3 o 0. .11 I ..1 h E 0:: t.0. t.2 ~ ~"=i1 F: af. 419. [sin (moefa)]' 'VI(mx)' .. For singlemode propagation.perture is V .8 ~ 1..6 sin (yx/c) 1 II b X ' " .5~7 '" '" 0. == k l __ y.~~ ([t~«.wp..fJ  VI .. 90 the original junction is called an inductive waveguide junction. ::::.""" 0.9 ~ " " ".0 cIa Fro.. (436).:a)t • = ~j ~(. "\' "1. b..(_fa)' 2a (482) The imaginary part of this is the aperture susceptance (483) which is plotted in Fig. so our explicit interpre I I I 0. from Eqs. Y) _ k' _ ( o 10 .176 TIMEHARMONIC ELECTROMAGNETIC FIELDS where.r (Yo)" _ 1. Susceptance of an inductive aperture...::?.
which is 21'. and H ~ continuous. consider a rectangular waveguide with a sheet of zdirected electric currents over the z = 0 cross section.p's is given by Eqs. thllt is.refer to the regions z > 0 and z < 0. Continuity of E~ nnd Ell at z = 0 requires that B. 32.cos 7 sin n. Moreover./2 if it is airfilled. we must have c > ). (485) The remaining boundary condition is the discontinuity in H ~ caused by J s.. We shall assume that only waves traveling outward from the current are present. E~. However..= B. At z = 0 we must have E~. and vice versa. the aperture susceptance is defined only in terms of E~ in the aperture and has significance independent of the manner in which this E~ is obtained. we can have wave propagation in it when c < >. This is illustrated by Fig... 44. + (484) B. Currents in Waveguides. ."'0 .... We shall now consider" current excitation" of waveguides. The problems of the preceding section might be called II aperture excitation >1 of waveguides. respectively. To illustrate the solution... (430)./2... nTy a sm T .. 1 . For wave propaga. Hz must also be antisymmetric about z = 0..) Superpositions of the TMx modes are L 2: "...n aperturetype problem into a currenttype problem...y e"'' z <0 where superscripts and ... hence it must be identically zero. (Note that J and its images are xdirectcdi so it is to be expected that an xdirected A is sufficient for representing the field. This involves the determination of modal expansions in terms of current sheets over a guide cross section. if the smaller guide is dielectricfilled...r .. 410. The equivalence principle plus duality can be used to transform a. cos mr.y) is now arbitrary.tion in the smaller guide. where J..8 ..PLANE WAVE FUNCTIONS 177 tation of the solution is restricted to this range.... the guide is matched in both directions. = uz!(z. The field in terms of the .. and it is convenient to use the TMx modes of Sec. The only difference between aperture excitation and current excitation is that the former assumes a knowledge of the tangential electric field and the l::Ltter assumes a knowledge of the discontinuity in the tangential magnetic field.+ = B.
J: d. we have the formal solution for all possible cases of currents in a rectangular waveguide. consider the coax to waveguide junction of Fig.. We therefore assume a current 00 the probe I(x) R< cos k(a . given by Eqs. In a lossless guide." are the TMx wave impedances. The solution for a ydirected current corresponds to a.c) (489) ._o We express J.. This is 8. (435). the power reduces to (487) where (Zo). are the Fourier coefficients of J~. fa [b m1rX ... When both J" and J II exist. A zdirected electric current can be treated as a loop of magnetic current in the crosssectional plane.B.. This is a summation of the powers that each J '"II alone would produce in the guide. as shown in Fig.)o(y . It is also of interest to find the power supplied by the currents in a waveguide.B... If the probe is thin. As an example of the above theory. = 2t". the current maximum is at the joint x = a.. This is most simply obtained from p  J J . by Eqs..z) (488) The current sheet approximating this probe is J.. and that associated with a oonpropagating mode is imaginary.178 TIMEHARMONIC ELECTROMAGNETIC FIELDS This is a Fourier eosine series in x and a Fourier sine series in y. according to Fig.. (430) applied to the above solution... the power associated with each propagating mode is real. nry fib}o ax io dyJ"cOSaSlllT (486) This completes the determination of the field.I(. . A zdirected magnetic current is the dual problem. the solution is a. rotation of axes in the above solution. With the probe joined to the opposite wa. . superposition of the two cases J II = 0 and J" = O.veguide wall. 33. the current on it will have approximately a sinusoidal distribution. The solution for a magnetic current sheet in the waveguide is obtained in a dual manner. 420. that is.. It is evident that 2y.. Because of the orthogonality relationships.._0 E . 2'Y.. Thus." the probe being the center conductor of the coax. waveguide llprobe feed.J: dx J: dy J: E.. = Jill.. as on the linear antenna.. 420.!. in its Fourier series and evaluate E.
The Fourier coefficients for the current arc then obtained from Eq.ka sin ka sin nrc/b b[(ka)' (mr)'l (490) This.coska where. the series for X. . However. Equation (491) gives reliable input resistances only when we are somewhat. where !(y . The power supplied by the stub is given by Eq. Then only the m "'" 0. + iX. Feeds in waveguides with arbitrary terminations are considered in Sec. This is incorrect for an actual junction. (487). (491) says that the input resistance is infinite at this resonance. In the vicinity of a "'" "A/4.1I'"C m b (491) All other terms of the summation of Eq. the reactance will be capacitive. 210»).. 420.(Z) b (tanka ka)' . removed from resonant points. and R. Ii . diverges.. A coax to wAveguide junction. 1 5.c to X. (487) is renl.c) is the impulse function.PL. and the error ties in our assumed current. coupled with our earlier formulas. let us consider the input impedance seen by the coaxial line. For small a.~"'E WAVE FUNCTlONS 179 Matched load Matched load I Coax Flo. To obtain a. (488). determines the field.  a 001 = . [This is similar to our linear antenna solution (Sec.  1:"'. since we assumed a filamentary current. or delta function (sec Appendix C). (487) contribut. In terms of this solution.. 1 term of Eq.R. Note that Eq. =  OOIJ"I' (ZO)o1 4 I. 811. The impedance seen by the coax is then Z. from Eq. the input current is Assume that the waveguide dimensions arc such that only the TEol mode propagates. finite X•• we must consider a conductor of finite radius. above which the reactance is inductive. we have a resonance. n . (486) as J •• = 2e".
There is a onetoone correspondence between a function and its transform.180 TIMEHARMONIC ELECTROMAGNETIC FIELDS x x Incldent =fl1 T~.~ Flo. For our present problem. the two [orms of solution must be equal.. illustrated by Fig. we shall take an alternative approach and obtain a different form of solution.. Let us demonstrate the theory for an aperture in the ground plane y = 0. there being only an E. in Sec. Taking a clue from our waveguide solution (Sec. The transform pair for E% over the y 0 plane is :IE: (492) where a bar over a symbol denotes transform. (49).). By the uniqueness theorem. At this time. 49). (389). 36. (493) as representing a field TE to z.. We further restrict consideration to the case E. let us consider Fourier transforms (see Appendix C). We have already solved the problem of determining the field from apertures in ground planes. l z 411. according to Eqs. and the other form for other calculations.. A parallelplate guide radiating into halfflpace. The form of the transformation suggests that we choose as a wave function (493) which is a superposition of the form of Eq. in the aperture.• (494) . 4·21. 421. we take Eq. however.. One form may be convenient for some calculations. Apertures in Ground Planes. = 0. hence it is evident that the transform of '" is if ~ J(k.k.
k. (492).. is n. we must choose k' k.)ei'. For Eq.ppa. As a word of caution.rent from symmetry that H.2 .' > v'k.) A comparison of this with Eqs.PLANE WAVE FUNCl'IONB 181 We also can rewrite Eqs. lW~ 1"" (495) R. (492) shows that 1 !(k. .he transform of H.J.' (497) The minus sign on the lower equality is necessary to remain on the sarno branch as designated by the upper equality. The case of H. ifR _ k. exist over the y=:O plane can be effected by adding the appropriate TE to x field to the above TE to z field.· dk. and E. This completes the soluk. will be the only component of H. and we must choose the correct root. (491) to remain finite as Y"''+ 00. 421 represent a parallelplate waveguide opening onto a ground plane.k. Let Fig. Let us therefore take H. k. ' is double~valued.' tion.k...!(k. = ± Vk l k.k.(k. it. _ kJc.)&>.I.jk.' k k + k.' + k.. as the scalar wave function and construct I H. or to y. is evident that t. It can also be obtained as the sum of fields TE and TM to z.' < v'k. if. For simplicity. (389) in terms of transforms as n.) 1 .·ei'. and HII specified over the y = 0 plane is the dual problem and can be obtained by an interchange of symbols. = !(k. we have D.) . (498) From this.icb both E. it is a.k. _ k' :. !(k. (496) where E.. If the incident wave is in the transmissionline mode (TEM to V).' if 1"" Specializing the above to the y = 0 plane.' (499) . we shall choose our illustrative problems to be twodimensional ones._o = .~k D.. is given by the second of Eqs. The extension of this solution to problems in wh. or to x.
E.)  f1 __ E. kll Al1 f _.I . it is evident that this root is k < k> Ik... which is f J(x)O'(x) dz  i. kllk z ..O). 421 for y > 0. (499). in the aperture to be of the fOfm of the incident mode.:/2) dk.) ~ k2 sin (k. sin'Y. f J(k)U*(k) dk We can express the power per unit width (z direction) transmitted by the aperture as From Eqs. = .I 110 WE k.182 TIMEHARMONIC ELECTROMAGNETIC FIELDS From the field equations. To evaluate this. as (4100) Ea Specializing E a to y = =  0. (4100) and (4102).1 (4104) The fields are found from the transforms by inversion.) may be found. we assume E. From Eq.1 Ik. A parameter of interest to us in future work is the aperture admittance. J(k. J(k. For an approximate solution to Fig.1' dk.! _. _0 ~ WE k... that is. we must also choose the root of kv for proper behavior as y _ 00. 2") :r (4103) To complete the solution."'· dz (4101) from which f(k. we have  k'll • E. this becomes p ~ !:!. we find E. 111. we shall make use of the integral form of ParsevaPs theorem (Appendix C). 1.. ~ ~ Ixl < 2 " Ixl > " 2 1 0 (4102) Using this in Eq. 2T f I. we relate the transform of E to R. (4101).(x.
.1 > k.a/ 2) dk • IVII ).o b/2 ainl tD dtD w' w' v' (ka/2)' sin' w dw (4105) b/2 w' v'w' (ka/2)l x.1 < k and imaginary for lk. 2 .s) to small ka gives a numerical facwr of 4. Aperture admit. 422.l The above integrals can be simplified to give X. .8 1.k.135 .a1 >t"1 f' 4 B• .."1(11 (f' + J. the aperture conductance and susceptance are plotted in Fig. J.' k' dk.PLANE WAVE Jl'UNC!'IONS 183 4 3 FJO. ".135...') k. into ita real and imaginary parts as G. t>:: ('.' v'k.2 0. ) I r0. these 0.!.6 ai' We now define the aperture admittance referred to the aperture voltage V . I" T1 E .. I.1 a (4106) 3..p...1'] } I: < 0. 422. = .ll ka.11(11 _ _ k.l .1 ~~ r.loG.l v'k1 k.2 log ka For intermediate ka. ill a qU&!iet. The direct specialUation of the seeond of Eq•. "B.a/2) dk.. tbe above integrand is res! rnr Ik. (4104).00 1 = 2 f. .). For large w. I r. k. = r [I >.. For srna..a as y _ p' = 4 sin' (k. we have I The formula for B..2 )."B.ic result..4 0. f" Note that.Lance of a capacitive dot radiator.at.."1B.p. by Eq.a/2) __ J: J: sin' (k.232 instead of 3.0 o 0. sin' (k. We can therefore separate Y. (4lo..
) . tbat is.' ~"o.· dk. ~ ~ ! ~ > 1 (4107) The aperture is capacitive. 421 to be that of the incident TE mode. Analogous to the preceding problem.) is evaluated by specializing Eq. Let us again calcula.} _ 2T' cos (k. The power transmitted by the aperture is P  f.)eJk.af2) w. = 2.H:)~. ~ (:.184 TIME~BAlWONlC ELECTROllAGNETIC VJELDS ~. (4110) and (4113). in the aperture of Fig.0 = f(k.t .p.}  f_'.n:). will be the only component of E. is the same as in the preceding example.~. since B. (4109) to y "'" 0. 421 when the incident wave is in the dominant TE mode (TE to y).(k. where we have used Parseval's theorem. we find E. Another problem of practical interest is that of Fig. this is 2. [E. E. E.•• dk. [E.0 ~ E. f(k.. :::s (4109) From the field equations."·· dx TX (411 I} For an approximate BOlution..I. (4104). as our scalar wave fundion.. WJl H. is always positive.O). This completes the formal solution. (4108) In terms of Fourier transforms. R.1'.te the aperture admittance.0 ~ ! 0 COS a (4112) Substituting this into the preceding equation. 2. we construct 1 E. we assume the E... which gives 2. . we find the trans(orm of H to be k. From Eqs.(x. and we shall take E...faC08(~' + DT] j f" f(k.B. I.S..z.)' [1 .. In this case. f(k. _ k. _. given by Eq. WJl (4110) Thef(k.a)t (4113) The choice of the root for k.. dx  ~ f.
. ~ 2.5 0.ure admittance of an inductive Blot radiator.a)'l' " kll 2 (k~aj2) dk • The integrand is real for Ik~! < k and imaginary for Ik.PLANE WAVE FUNCTIONS 185 0. ~ B... 423..2 t (. Apert. into real and imaginary parts is therefore accomplished in the same manner as in the preceding example.' 1/ WIJ f" _ . • a ~ G." vlw% (kaj2)tcos2ww B = d }... 1r).4 E """ cos ('lIx/a) V V V 0. < 0.I.. 21fwIJ f" _. f.8 f0... Aseparation of Y.a/2) dk.l > k.6 f. (k.5 ai' FIo.lJB~ V o 1..{l<JG.a)2J2 We shall refer the aperture admittance to the voltage per unit length or the aperture.194 l a ). which is V = 1. 1~2  k: cos' (k. k*IE.. > a 1.0 1. )..5 (4116) . this becomes p ~ ..I' dk.1 (4115) For intermediate ka.\ 0. This gives y . rvr _ = _2~a2 Wa f" [r'cos (k.. G.. we have ~G ~ ~(~)' }.  2 }o vi (ka/2)2 [(r/2)' w 2 cost W w'J' dw • If.. ~ 0.. For large ka.=!..I I I I tX _E \L a IT ( . The result is '1 1 (0/2 ). 423. 2 h/2 [(orr /2)2 w 2J2 (4114) For small ka. the aperture conductance and susceptance are plotted in Fig. = 2>. p. .
let us follow the more circuitous path of constructing thc solution from basic concepts. y FIo. . We DOW reconsider the problem from the alternative approach of constructing transforms. Plane Current Sheets. (386).ve function representing the zcomponent of magnetic vector potential. The aperture is inductive since B" is always negative. ~ jk. such as Eq. Suppose we have a sheet of zdirectcd electric currents over a portion of the y = 0 plane.186 TIMEHARMONIC ELECTROMAGNETIC FIELDS z J. type. be determined by the potentia. rather than taking this short cut. if the equivalence principle plus image theory is applied to the results of the preceding section. 424. x and B. We anticipate the wave functions to be of the transform type. requiring continuous distributions of eigenvalues. The field of plane sheets of current can. The procedure is similar to that used in the preceding section for apertures. (493)..jk.l integral method of Sec. 412.) The problem is of the radiation. (This we know from the potential integral solution. In fact. A sheet.0 .~ D. of course. However. given by n. we have the transforms of the field components for the TM to z field. of zdi~ rccted currents in the y  0 plane. as suggested by Fig. From Eqs. 29. we have complete duality between apertures (magnetic current sheets) and electric current sheets. is negligible. .~ (4117) D. 424. The field can be expressed in terms of a wa.
as •• · if!'" = f+(k. (4123) where k" = k. the transform of J. Hencel for y = 1 ... we must choose k. In this example.. .. k" > 0 we have A jll '" = . = II _. J ~~·~eii .• dx dz = .eJA··~. is J... 8..) = J".hr r = V x""+'y""+'z::. J.. The field is given by the inverse transformation. (497)..k..•.. For example.!Jt as well as the . (495). J". That is.··dzdz (4121) This completes tbe determination of the field transforms. .J/! where I _.z). 221. Ell. as in Eq. and kll.II '(x) '(z) J. The boundary condition 011 Es and Ell leads to 1+ = 1. 41r f' f" _. The potential in~ gral solution is A := u. J. consider the current element of Fig. and a discontinuity in Our boundary conditions at the current sheet are continuity of E s and Il~.PLANE WAVE FUNCTIONS 187 These are dual to Eqs.' (4122) For the transform solution. (186).. 4r u..) y>O y<O (4118) For the proper beha.+ is given by Eq.lIeJA·· dk~ dk. Our two solutions (potential integral and transform) plus the uniqueness theorem can be used to establish mathematical identities...(k••k...k. ../.as the other root.(x.!Jt where lle'h '" ..vior of the fields at large Iyj. according to Eq. f" f" _.. and the boundary condition on H ~ then leads to (4120) where J.) f(k.~ We construct the transform of y. (4119).
Finally.·o<Jimensional problem to which we shall have occasion to refer in the next chapter is that of a ribbon of axially directed current.. the field y > 0 is unchanged if the aperture is replaced by a magnetic current ribbon K = 2V. This is shown in Fig.pe.and zdirected magnetic cunent sheet (or a solution.l.. (4105).. Z"M'" ~ 2 ~ J(ka) 1 11  11' "2 Y. we have 1 21fj _. This ribbon radiates into whole space. equating Eqs. l. The parameter of interest to us is the If impedance per unit length.' ks t _.. According to the concepts of Sec. if the sheet contains ydirected electric currents.. (4122) and (4123).••.1:.. A ttl. so the power per unit length is twice that from the aperture. 422.~" = ..s apply duality to the aperture problem of Fig.~ jP:pott = 2Vlt lL)" 7::: . The solution for xdirected current can be obtained by a rotation of coordinates. 425. uniformly distributed.X J(k4) By duality.lb = p.. t . since kll changes sign as y changes sign. The admittance of the magnetic current ribbon is thus z y ."'~ where the aperture admittance J.. the solution is a superposition of the xdirected case and the zdirect." defined by (4125) where P is the complex power per unit length and I is the total current.··dk. TKft Y. we can convert to the equivalent x. y "'"' G.' . 36. which we can represent by 1 Y. we have the radiation impedance of the electric current ribbon given by x FIa.. r =  f" f ei. We have considered explicitly only sheets of zdirccted current. When the current sheet bas both x and z components. yk 2 k. and vice versa for ~irecLcd magnetic currents.ed case. the identity rib Hence. The solution for magnetic current sheets is dual to that for electric current sheets. Rather than work through the details. 425.dk • (4124) This bolds for all y... is given by Eq. + jB. A ribbon or current.188 TIMEHARMONIC ELECTROMAGNETIC nELDS field is unique.. let u. M " (4126) .
135 Sec. The fo. and R  1.a t\ solution to the !Calar Behnholh equation.1.3.)' bm' + an' ] . . For Ie .ctor<lftwo diffcrence arises because the aperture of Fig. It. (49) i.000 megacycles.ain (Jz coah G% COlI b: . 2. are the mode functions generating the twodimenaional TE. 31. show that lin b . oycles. Derive Eq!!l.) For narrow ribbons..·8. Show that. (417) . (293).2.b planca. (293) tor aU TE.ionB generating t. modes according to Eq•.(I.... An airfilled rectangular waveguide ia needed for operation at 10. modes and by (a) _ 2m [ " (a • ..he 11 and 11 . ~ 2~ IT + j(3..ltLANlf: WAVE FUNQTIONS 180 (Compare this with Prob. (4106) and (4126) z. with center frequency 10.. • 2. (389).. ° nI. (386).. modes. parallelpllt. we have {rom Eqs. Show that Eq. Following the method UlIed to establisb Eq. It is desired to have singlemode operation over a 2: 1 froquency range. Determino the waveguide dimensions and the attenuat. . m and n nonzero. show that.ion constant ot the propagating mode for copper wa. sbow that the attenuation con. .lIs.COIJ (Jz eoah lEt  + J Bin fJz sinh cr% i C08 (Jz Binh GZ 4.000 meg./J)1 + ~ I Vi I. ab vI .( 7 6l m l + ain l tor TEo. modCll aocording to Eq•. is also desired to havo maximum poWOl'handling capacity under these conditionll.. 4.he twodimensional TM.vc~dc formed by conduotors coverins t....ln'm'b + nla + t t (or TM•• modes.a..te wa./J)' + b)U. .. For 8. PROBLEMS 2 log ka)J (4127) This we shall compare to the corresponding Z for a cylinder of current in . arc the mode funct.. 3.22 radiatcs into halfspaco and the twin~ slot line sees allapace. . 56.fJ . the TEM mode ill generated by .. .atant duo to conductor loasca in a rectangular waveguide is giveD by Eq. and by ( ) Qc • •  IfQb mfb' lI'a' vi26l (J..3• . • 6. 4..
. that is..' ] qabck. (445) with and for modes TE to x the characteristic equation is Eq. mode" according to Eqa. Show that for modes TM to % the characteristic equation is Eq._ mr k pr ·'7 k•• = .' + 2b<:k. E TE.') "aock. and n = 1.1 + a(b + clk.)".fERARMONIC ELECTROMAGNETIC FIELDS "·7...0. 238. .0...) .plate waveguide of Prob...d .4lR{b(a + C)k.. ' + k. C.. Consider the twodimensional parallelplate waveguide formed by conductors over the z .'k. (10.8. Note that n . 43. (447).) .l fr.' + 2ack.'k. which generate the TM~ modes Recording to Eq8. 11 . (12.tRlbc(k••• Til f/abcku'k.) + DH./k. and y _ b planes arc the cutoff frequencies of the rectangular waveguide.1 + k. (430)..0 and x _ a planes. Show that the resonant frequencies of the twodimensional (no: variation) resonator formed by conducting plates over the x ..'Bl. Note that no mode . 46 are n .a. .. il. 1. f/abck? Til: (Q.0 in the above TEx mode function gives the TEM mode. (2100.' .. _ TIoh _ A (EO'.2.0. (Q TK 2Gl. Following the method used to establish Eq. TEM to z (the direction of propagation) is possible.') + abk.3. which generate the TM:t. _ Vk. 49... (432). . l ) + eu:(k.. 44 are linear combinations of the TE and TM modes of Sec..2lR(abk. '1abck..4..2..l) a k nr 'T where k.' + 2abk.  + 2ack/ + 2abk.(bck.TE H C(H. .1 (Q')..' 2lR(ack.. Show that the TEa: and TM:z: modes of Sec.l) + k"k. f.0. Show that an alternative sct of mode functions for the parallel.' + k. :e . Calcull1tc the first tcn higherorder rcsonatlt frequencies for the rectangular cavity of Prob.' 411.l + k.T)I) Determine A...') (Q... and D.. show that the Q due to conductor losses for the various modes in a rectangular cavity arc Q Til: ( <l. and dielectrics ~I for 0 < x < d and ~I for d < x < a.' k.' Til ' + 2bck.190 TU...T£ + BE.
46) for a Denote the emptyguide propagation constant (d . that for Bmall d 416. t l and ~I .. Show that the static inductance and capacitance per unit widt.... 417.. the field is almost TEM. p . _ Col . 46 with additional conductors covering the z . 0) hy > b. (448) about d _ 0 and k. Denote the emptyguide propagation constant. _ tJ.ssionlinc formula k..a a rIFI + .0) by and show.. that for small d k. ~1 or as d ..JL{j thereforc applies if {I is small. FI~tfJ. 4:14. Show that thc resonant frequencies of a partially filled rectangular cavity (Fig.. .nd k. show that if c resonant frequency of the dominant mode for smll. .. 0. .. . the . Considcr the dominant mode of t.F1 . Show that.. (445) and (447) with whcre n . WIlen d is small. from the Taylor expansion of Eq. . n _ p _ 0 excopted.2.. 416. .. Cd .2. Also. .tJ.+ . 1. Show that the lowestordcr TM to % mode of Prob.tJ.18. .d 3ptfJ..)1 .10 for the dominant mode. For the partially filled cavity of Prob.(. (448) for the dominant mode.he partially filled guide (Fig... Consider the dominant mode of the partially filled guide (Fig.h and length of the transmission line are The usual transmi. (445) can be approximated by Eq. . 1.. 46) for b > a.. and show.(kit  k. (447) about d _ 0 o. O.PLANE WAVE FUNCTIONS 191 a« . from the Taylor expansion nf the reciprocal of Eq. if ).c planes) are solutions to Eqs.') (d)' a 4:16. 412 reducC6 to the transmissionline mode either as ' I .. 0 and z .11 d is given by > b > a. . 0.. Eq. .
56) or (468). Using Fig.20.18.. . Bow can the cutoff frequenciee of corresponding TE and TM modee be the same. aa shown in the insert of Fig..4CYcles. show that the dominant TE mode is characterized by 1I . For the plU'tially filled cavityof Prob. the 4. 4.. By a Taylor expansion of Eq. A plane slab of polystyrene (" . tan (. 414.. 710. 4. show that if c resonant fr~ucney of the dominant mode for IilUall d i8 given by "'I [1 _~ a + c' ~ . Who. yet the propagation conlJtants be different. 4.ad) . (456) about II ._ a ".000 megacyclea? Calculate the cutoff frequencies of these modes.192 where cq TDtlEBARIoIONIC ELECTROHAGNETIC FIELDS is the r'e«>Dant frequency of the empty ea.I 'Ull 1) (~)'] a wbere All is the reaonant frequency of the empty cavity Him: Use the roeults of Prob.k DI ) " 2 ..tcred dielectric slab.('. determine the prop&gation constanta of the propagating TE modes at 30.~o2 . d) •• co l\.J + (. T .0.21. &bow that the dominant TM mode of the lllAb guide (Fig. 415. Derive Eq.cd at a frequency of 30.12" k .000 mer.. 416.19. and for modes TM to :t it is where The dominant mode is the lowestorder TE mode (smallest root for n .(k4' .2. Determine t. l '1"1'1 (tlJl.0). Similarly.56) is ~ centimeter thick.3._. (4. Col _ (.0.)' > a> b. k Hint: Usc the results of Prob. It . "'.he propagation constanta of the propagating TM modes by numerical solution of Eq. Consider a rectangular waveguide with a cco.2~. (4S8). 411.. _.vity. .. 410) is characleriud by for small /I. Show that the characteristic equation for determining tho propagation constants of modes TE to :t is k.t lllabguide modes will propagate unattenuat. .
.~t5.se.. 0 eoe. The centered capacitive waveguide junction is shown i. (478).coa!Trt~. It is UlIumed that E.+2k.. Determine the minimum depth of slot needed.) : < 0 and beight b lor. /.he aperture sWlCCptance per unit width relerred to the aperture voltage is is that of the eecond drawing of Fig.·23. 42'.megaeyde field. . 416.n }o'ig. Suppo8e that the elota of the corrugated eonductor of Fig.. .. 426.. Compare this with Eq... ~I + ~ tan where k 4 . k.()()().ee vaJue at one wlwclength from the eurfaee. !! k 4 tan k... in a 3O.>0 wh~ A.0) to & thickness of O. (28.. 426. has been assumed in the aperture.k.For the corrugated conductor of Fig. it is desired that the field be attenuated to 36. (The C10M fJeCtion Using the formulation or Prob.' I 0 E. A plane conductor baa been coated witb sbellac (eo. "'. l k#1 ~l A. where a constant E.. • . Will any tightly bound surface wave be possible! Calcuhte the attenuation constant in the direction perpendicular to the coated conductor.. . In each ea... 426. the propagation constant is given by k. > 0. of ita eurfa..k.b planes. (478) with>" replaced by 2>.. Use the TE: mode functioD. B • O¥ 4 \' sin l (n . is given by Eqs..3. Conslder the junction of two parallelplate tran&niaeion linel of height c for with the bottom plate continuous. Show that a field baving DO E.. .8 per cent.c!b) II>' "" (O"C/b)1 VOl (26/>')1 • " ..ll of Prob.4. A centered capacitive waveguide junctioD . 415. 47 for the parallelplate waveguide 10rmed by conductors covering the ¥ _ 0 and ¥ .. '. orJI d 0127. ...OOS inch. show that t. (432) with . It is to be U8ed..d . '~6. 415 are filled with a dielectric charaeterhed by f4J ~4' Show that. in the aperture is that of the incident mode.. ..PLANE WAVE FUNCTIOSS 193 for Iml&l1 /I. for this case II  . 4. Show that tbe aperture llU8Ceptanee referred to the maximum aperture voltage fa given by Eq. FlO.
7.. Uso the analogy rM. lol . t.....b p1aOell. A"  " • >0 • <0 dy where 2b Jo I" (... 3..:. dll I D. 6eld produeed by this Zodirect. ...0 and" . •so.29. ~ (~)' ". 427. '32.. Consider the centered inductive 1I"&veguide junction of Fir.ln T . (483). Let.. or Prob. no..161' f." p... 0 B..1..ed eurrent.194 TWEHABllONlC ELECI'ROllAGNETIC I'lELDS fx Incident tx w. ••.'0 a " 3.ob Jo J....d% 0%11 /../(!!!)' _(~)' 2 ).be aperture is that of the incident mode.31.W>  I FlO. Let there be ••beet of .. /.egration. _ b... '/.. table 161.d1l .. z and cia "'" th to show that 8. Halner Publishing Company. over the z . A centered inductive waveguide junction..directed current J. \ k' [""" (m. lQ39 (reprint). In Eq.''')' I ). 8i(2 c/a_O sin 11 r. in t.. b JJ. Show that the 6eld produced by the c\U"rent sheet is . The guide ill matched in both tbe +z and z direetiotlll. Bierel18 de Baan.0 plane of a parallel· plate waveguide formed by conductors over the 11 .. J. 225..hat. ZI %d% Integrate by pArliI.2.T. cos n. New York. note that M c/a . &how that the apertl1l'6 IU8CePt.0 the summation becomes similar to an int. )._1 0 _ b..r~)]' 1 (rne/a)! "V . 43l be %directed imteAd cf vdirected.  B" .226 2.11.. 4./a ..hat B. n. sheet. is Show B..Si(Z.) ) _ 0. ~ ~ A.w) eos T n . . the eurrent.. 4·27.(y) 8. 4.."(".y . . sheet...1 _ { ll. Aasumin& t..... "Nouvelles tables d'intkgrales definiCll./0_0 .he muimum apertuJ't! voltage is giVeD by _ B. anee referred to t. and use the identity' toahow that • sin ~ . " 2: B" T ain .
429.. (4105) becomea 0. this last conwur integral rOO.6 (ta)' + 60"2 I (to)' 2" Eqs. + C.. Re Flo.36. L Y (aj T (b) FlO.B...'i I Ir' Coa' Ir::. resistance Ren by the coax ia R _ ~ (Z) .l ch&raeteristic wave impedance. Show that as ta/2 becomes l&rge.. By expanding (sin VJ/w)l in a Taylor aeries about of Eqll. w' v' w' (I . 42&. the input. " • c. Xib I. Show tbat.86. sbow that the first ).. 436. 429.). Assume that the aurren\ on the wire varies a8 cos (ll). . aDd C. CootoW1l for Prob. (4107). Co where C1 ia shown in Fig. .ueee to the eeoond of Eqa. I 1008 2" (ta)' + . Cooaider the closed contour C1 + C. 428. Coosider the cou to waveguide junction of Fig. Co . where I is the diatance from the end of the wire.ions.PLANE WAVE FUNCTIONS 195 '83. + C" &lid eJ:press M. ia the TE. 4286. ] 1m 4.'1'/1. in terma of a contour integral Over C. Coax to waveguide junctions.. B w plane Ile [/. which is matched in both direct. [Bihleoal(c+d) n (re!b) ain id]' where (Z. Consider the second of (4105) as the contour integral . Show that the input resiatance Ren by the coax is now R _ ~ (z) I Co '11 lain (.N)dw ] (14/2)1 c. Only the TE'I mode propagatell in the waveguide.. 433 ia changed to that of Fig.  'If I [ 1 .a/bUrin k(c + d) kacosk(c+d) sin b:1}' tD  f. 'K Suppc>ee that the coax to waveguide junction of Prob.
ka_!o"8 Re le.. \' b. 430. 430) C!\8e a . integrate by partll.. about that the first of Eq•.002083 _ 4088.0. (4114) to the and U8C the identity (8ee l"rob. (4ll4) reduce.196 TIl&EHAB140NJC ELECTROMAGNETIC FlELDe tD _ 4.012182 0.E' Co C._4a). Evaluate this Ill. and ahow that • '0 0.11.0 .0.e phase and equal magnitude. _02./2)1 where C..St. I> 1.will in a Taylor seric. 4. and CXPre&8 G. abow . 439. By expanding coal tD/[(1I'/2)· ..+0. (4114) bccomCll • 2 ~ ).194 ~ ).0. Conloul1l for Prob.bo_ 1m x ED pltlne c+.8t Conllider the clOlled contour C. • (0)" +1.0.. b... 431. Fro.I . in .e the second of Eq•. . bl 0. Two parallelplate trR~ sion lines radiating into halfspllcc. . + C.!!' B. FlO. contour integral. jt to show that f. to the contour integral ~G. [(.+0. Show that the 6n. Assume E.O..05M13 .t of Eq•. SiC. 6. in terml of a conlour integral over CJ and C..) (11'"/2)' :r:' 1I'}o II 1/ 11' r . Two parallelplate trans.to. lUI sbown in Fig.189108 . Specialir. 431.} . • 'G [f (I + ''''')". + C. >: bl b... dz _ ~ rain 1/ d _ ~ SiC.4. lIin 2z. _/2 Cl r h R.67401 . ia 8hown in Fig..misaion lines opening onto & conducting plane are excited in oppoeit. 4·30.If! dW] + C. b..
Specialize the potential integral801ution to . 423.k.(k. 4:4:2.. and show that f + 4~.\" J. 424) by (0) tbe potential integral method and (b) tbe transform method. is given by Eq.p1 are equal. .C08a z.t the impedance per unit length. ei~ .directed. k COB 8) where J.to to l vw sin 4 w dID l (ka) I 4:·4:1.directed rather than z. Construct the vector potential A .. where Y.b 0 WI V (bip 8in' to dto ttll B. _ 10. (4125).0 plane (Fig. Supposo that tho current in Fig. and show that the aperture susceptance referred to the aperture voltage of one line Us G •  8 . J...stant lor each line. where I is the current per unit length. (4~126). _ ! (. . 425 Bnd of magnitude J. Show tha.PLANE WAVE FUNCTIONS 197 the aperture is a con.p"r' is now the llperture admittance of Fig.. (344)] that the twO.. .\" J. (4121).UN {or & sheet of tditeeted currents over tbe 11 ..) is given by Eq.( k cos q. sin ~ 8. defined by Eq. Show by use of Grcen'a second identity [Eq.
a.lso be independent oC z if the equation is to sum to zero Cor all p. The Wave FUDctions.! d'~ R dp Now the second term is independent of p _ k ') . we can construct electromagnetic fields according to Eqs." + a. (391). Following the method of separation of variables.." 198 . dp ( p dR) + ~ d~' + (k' ~ . 1 We shall usually orient the cylindrical coordinate system as shown in Fig. 51."2 + k'f p up p up p! oq. _ 0 • p and z. and the other terms are I The term "cylindrical" is often used in flo more general sense to include cylinders of arbitrary cross section. Once we have these scalar wave functions.l. .e. CHAPTER 5 CYLINDRICAL WAVE FUNCTIONS 61.t dz _ 0 (51) which is Eq. (27) with the Laplacian expressed in cylindrical coordinates.. We are at present using the term to mean "circularly cylindrical. q" z. We first consider solutions to the scalar Helmholtz equation. Hence...0 The third term is explicitly independent of p and q. The scalar Helmholtz equation in cylindrical coordinates is ! ~ ( af) + . . is a constant. It must a. we seck to find solutions of the form f  R(p)~(¢)Z(z) (52) Substitution of Eq.C'. (52) into Eq. (51) and division by '" yields 1 d (dR) pR dp p dp + p'4> d¢' + Z dz' + k 1 d'4> 1 d'Z . Problems having boundaries which coin· cide with cylindrical coordinate surfaces are usually solved in cylindrical coordinates. Substitution of this into the preceding equation and multiplication by pi gives .! d'Z t Z dz _ k' • (53) where k.
.p) .••  B.ions [Eqs.p) is the Hankel function of the second kind.(k_p).0 0= (57) 0 The cia and Z equations are harmonic equations.k' and write the separated equa.•.z) (59) lIt iI more usual to denote solutions w Bessel'.(k.0 (55) x Flo. (54). so B.(k. in general.p) ~ J.p) is the Bessel function of the first kind.1 Commonly used solutions to Bessel's equation ar. Id'4> . equation of order n.p) is the Bessel function of the second kind.(k. (52). (k. B.' . but we wiah to avoid conIuaion with our Z(z) function and with impedances.p)."'(k. These functions are considered in some detail in Appendix D. H.(k. 61. According to Eq. a linear combination of any two of them.p)h(n~)h(k. The R equation is Buw'.llP). The wave equation is now separated.z). in general.<ll(k. define k_ as (56) + k. H.' To summarize. we can now form solutions to the Helmholtz equation as of••.p) is the Hankel function of the first kind. and H. (53).ell. N.p) is. k. Any two of the functions of Eq. which is an equation in p only.p)'  n')R .0 d'''' + n'4> d~' ~~ + k. Cylindrical coordioat.(k. The preceding equation then becomes dR) . z (54) p n' where n is a constant. solutions of which we shall denote in general by B. (58) arc linearly independent solutions.. N."'(k.p) (58) where J . equation by Z. sud (55)1 as d p dp ( p dR) dp + [(k. These we denote.n l p d ( p Rdp dp • y + (k l  k~l)pl . H. giving rise to harmonic functions. by h(n¢) and h(k.IZ .(I)(k.p).(k.= '" d~' Hence. and we shall discuss them later in this section.CYLINDRICAL WAVE FUNCI'IONS 199 independent of ¢.(k. .
)h(k.) if "it is to be singlevalued.•• are constants.  (511) ..(k. possible solutions to Lhe Helmholtz equation are (512) • where the integrations are ovcr any contour in the complcx planc and !. and k.B. although n is usually discrete (this is dieeu!!800 below).p) .p)h(n. consider the various solutions to Bessel's equation. Me interrelated by Eq.LLO•. Graphs of the lowerorder Bessel functions are given in Appendix D.. have occllsion to integrate over either k.0. We can also integrate over the separation constants. For example. however.p from 0 to 2. and trI·· are more descriptive.p) must be J. This means that h(1I.. and the elementary L1..p) must be periodic in . so the sa. "/!  ~ f O.) and g.(. In most cases.. is a solution to the Helmholtz equation.(k. The.(k.(k. We dillCu8ged the interpretation of the harmonic functions in See./·(k.(k.(k.)B. • Y.."it.)B.p) and n.) are functions to be determined from boundary conditions. 4. y.(.l coordinate system is also one of the rectangular coordinates. for they are interrelatoo). places restrictions on the choice of h(n. = LI. if we desire the field in a cylindrical region containing all .(k. Equation (512) is used to construct FourierBessel integrals. Linear combinations of the elementary wave functions are also solutions to the Helmholtz equlltioD.p + 2. in which case n must be an integer. We note that only the J . we choose sin (n. and k. or easier to deal with analytically. (but not over k. a summary being given in Table 41. For example. or k. We shs.p.Il use Eq.z) die. 4 apply at present. We call these '" ~lenumI<Jry wave Junction. although in some caBell the exponentisls eI·. We can sum over pO!!llible values (eigenvalues) of 11 and 10" or of 11 and k.. the 11 summations of Eqs. Now. it is necessary that y.. as such. The.(k..p)h(k••) ••• (510) where the C•.p.me considcrations that dictated the choice of h(k~) in Chap.p)h(n. Thus.(k. as we did in Chap.p). g. the B. if a field is to be finite at p .p) or a linear combination of the two.200 TDUHAIUlONlC J:LECl'IlO¥AOnrrtC ""LOS where Ie. (lHl). (510) to (512) are usually Fourier serics on . (511) to construct Fourier integrals.Jo. Thus..z) dk.p) or cos (n...p)h(nq. 41.. We shall. coordinate of the cylindriea. Hence.p)h(k.•..p) funotions are nonsingular at p = O. ..p coordinate is an angle coordinate and.
Il. as do the sinusoidal functions... Re (k. (2)(kp) analogous to e. exhibit oscilla. 2 If k is complex.e fl •• p = 0 included We have written the harmonic functions in exponential form.. The H.. Insight into the behavior of solutions to Bessel's equation can be gained by noting their similarities to harmonic functions. J" and N .+IH.p)el". is real. choose the root according to the limit 1m (k) . = ± yk' Ie. that is.1. = J. Our convention will be to choose the root whose real part is positive.( jap) K.. which is still general since sines and cosines are linear combinations of them.. co included (514) Other choices ~f cylinder functions are convenient in certain cases. as we shall see when we apply them.CYLINDRICAL WAVE FUNCTIONS 201 (513) wave functions are of the form y.. When k is imaginary (k = ja). They represent outwardtraveling waves if k. (kp) a. and K".j'J.. as do the exponential functions. it is conventional to use the modifu:d Bessel functions J.ik ..Cap) ... w representing inwardtraveling waves and H. then our interpretation of JI"lIl &nd ll"m would be reversed.nalogous to cos kp N . H . Cil and H .... is imaginary.p) are the only solutions which vanish for large p if k.. the elementary wave {unctions become p .tory behavior for real k. the B. I Ilf k. Therefore...(tl(k.. This direction of wave travel ilJ a consequence of our choice of ...{2)( jap) (516) + O... except for an attenuation of l/ykp.p) if p + co is to be included. Hence.(ap) = i (j) .' is indeterminate with respect to sign.. (Dll) and (D13)]. (2 1 (k.) > 0. They therefore represent cylindrical traveling waves. It is evident from the asymptotic formulas of Appendix D that.. Hence. Note that H . 1 Now consider the asymptotic expressions for the various solutions to Bessel's equation [Eqs. is complex. (56) that k... the traveling waves arc attenuated or augmented in the direction of travel (in addition to the l/Ykp factor).(k.(kp) analogouB to sin kp (515) H .W(kp) analogous to ell. if there are no sources at infinity.i'" time variation.p) must be H. these solutions represent cylindrical standing waves.(k. C2l representing outwardtraveling wavcs. defined by I . Note "'from Eq. If we had initially chosen ./.m functions represent traveling waves for k real. the following qualitative analogies can be made: J . For example..
will increase as we apply the various functions to specific problems. and expanding Eqs. = fjp a"" a¢ az H. .. When k = 0.(Op) . Table 51 summarizes the properties of solutions to Bessel's equation. That the various analogies of EqB. the method of Sec.. and N •. '"'"' 0 .. while the wind gives rjse to "harmonic Cunction" waves.p."". Pn .) field existing in a homogeneous sourcefree region.y. we have the degenerate Bessel functions B. . log p 8. it is evident that we have the qualitative analogies I. a dropped stone would give rise to "Bessel function" waves. Both Bessel's equation and the harmonic equation arc specializations of the wave equation.(ap) analogous to eo' K. (515) and (517) e. (385) in cylindrical coordinates.202 Tllol&HARMONIC ELECTROMAGNETIC FIELDS These are real when ap is real. 0 Note that these are essentially the smallargument expressions Cor J.. To express an electromagnetic field in terms of the wave functions 1/1. given in the last column. In the case of waves on water. _! a"" yap az 1 . (D19). Our understanding of the physical interpretation. Similarly..~(::. and expanding Eqs. From their asymptotic behavior. The unit z~oordinate vector is a constant vector. . _ a~ ap (519) E. 312 can be used. The result is E.(Op) """1.+k')~ which are sufficicntly general to cxprcss any TM (no H.0 (518) E. 80 we can obtain a field TM to z by letting A = u. we can obtain a field TE to z by letting F = u.'tist is no coincidence. Eqs. (388) in cylindrical coordinates.(ap) analogous to e' (517) From these it is apparent that the modified Bessel functions are used to represent evanescenttype fields.P a~ 1 a~ E. The result is E E .
I: real........(k... 2'(" .) I..l .·.(4"....)] ...''''''di".p t" . n"".1:.. r I) I "_ 0 .. H[II.)" " >0 t complexattenu...:p .b"dinC . _r&\!II>ent formulas (It...1:.j. PnOPERTIEs OF SOLUTfO:<.".~ In1iniliea PhylllCl&llnterpl'flUlion .I: ima&illllrytwo eva.u"di.. t floroplnloc..I:. .(t.. p...lIl{.jN.) flOll ..... 1:. k" ."di.inary_v.I)! . ve " _ 0  2'(" ..8LE 51.I.10.. t comp\eorloc&lired .l:P" 0 . 0) Lar..rdU'a. 2·{II 2"..(t.) ("pl" +.(OI{kp) + H.(..(k.ve . (. i~na~van~nt fi~d i:l J. ted travtllnc wa"e t rul_.{k. ..1.1:.livo .. TI:P [H....(t"l ... "_ 0 .III 2". I..I: realinwardtravellng w"ve H.) Aher ...1:" (..) r..~..ee"t 5eldt tpo i j .I.. ....:......o nuroher . 1 " >0 I"finil.. j... w.....)· " >0 axit ..).........) ..... .II}{.. .: lm&&.(t"l + jN ..: rulout ...!.....[± ( ..:.:p) " .vtllnc wave 1:. ~IO"(~) ." .0 j"si'p .eent field I: complcx_twnuated travelin. j ... 0 l+j.. " >0 ..I:...2 4 20 .) lj~IOg(. SmaUac&\!II>ent formul ...: kp . .ve H......781)·t B.I:....h. ....) _ HoU)(...I:"lJ (.Hre<! . . 'hfl ru.).)  2i ......I:..... waves r(.. . wave . 0 __ j•• i. . inarytwo flvall_lIt fi""d.(»(.....IOCC:.....)' ".).1 axi! kpo :tj I: lm.I:..l:"ain Infinite number alon.S TO BESSEL'S EQUATTON Coy  1..)' (1:""2...) z. .. ) H..n~t.l...) r )'.I'I(..)'i '" ..I.... .lon.l. 1 ..TA.
The circulu waveguide. Dl. the x z Flo.p ISin n~) . called tho circular waveguide. 80 the wave (unctions must be of the form of Eqs.(k ) . the phenomenon is Bimilar to wave propagation in the rectangular waveguide. . In particularJ E. may be determined.0 Hence. hence '" _ J. The TM field is found from Eqs. (520) is the desired form of the mode functions. For modes TM to %. mode degeneracy except for the cases n "'" O.(k'  1 9 k. Qualitatively. y . The coordina. we may express the field in terms of an A havin« only & z component 'It. It is conventional to express the 4J variation by sinusoidal functions. The propagation of waves in a. An arbitrary field (one having both an E. provides a good illustration of the use of cylindrical wave functions. and an H. 52. These are ordered a.0. Note that for each n there are a denumerably infinite number of zeros.. Either sin nq.(k.nd designated by X"JO.f'" cos nq. 43.. (518) applied to the above y.204 which are sufficiently general to express any TE (no E. we must have (521) from which eigenvalues for k. considered in See. (513).tes to be used are shown in Fig. hollow conducting tube of circular cross flection.a) . 52. may be chosen. The Circular Waveguide. p .) can be expressed 8S a superposition of Eqs. 80 we have 8. The functions J M(Z) afC shown in Fig. The field is finite at.·)~ p /I.) field existing in a homogeneous sourcefree region. or COB nq. (518) and (519). 62. which must vanish at the conductjng walls J .
ORDERED ZEROS ~..969 4.520 8.796 6. with the field determined by Eqs. .)' + k. The electromagnetic field is then determined from Eqs.340 12.536 11.'  k' (524) Subscripts np on the k. The J~(%) have a dcnumerably infinite number of zeros.761 13. hence.5 5.40.~. 2. .'(pressed in terms of an F having only a % component J/I. Equation (521) is now satisfied if we choose k .173 13.417 11. and p = I.201 8.832 7.416 10.832 7.. (. . component is 81/I/iJp.324 1.620 14. (520).771 12.. 0' J . 3. .620 13.331 8.) The lowcrorder zeros are tabulated in Table 53. is determined according to Eq.. .a) ~ 0 (525) must be satisfied.173 13.339 2 3 • 2. that is.170 . (518) with the above y. The lower order %•• are tabulated in Table 52.841 5.688 11. .'" a (522) Substituting this into Eq.706 3.987 3 • 3.317 9. The E. 2. ~ ~ I' J.136 8. (519). 1. the J~ also are oscillatory functions. Modes TE to % are e. lcosn¢o (523) where n "'" 0.015 11.380 9. (520). The mode phase constant k. This wave function must be of the form of Eq. (56). are sometimes used to indicate explicitly that it depends on the mode number.).682 13.054 6.016 10. x 1 2 0 1 2 3 • 5. which must vanish at p . .372 • 8. J~ "'" J l .a. which we order as x~Jl' (The prime is used to avoid confusion with the zeros of the Bessel function itself.015 first subscript referring to the order of the Bessel function and the second to the order of the zero.324 6. are oscillatory fUDctions.054 J1 . hence the condition J:(k.".282 • 6.016 10. The J . we have the TM •• mode functions ~. ( ••• a p) J sin n¢) . (For example.706 9.7fi2 3.005 14.CYLrNDRlCAL WAVE nJNCI'JONS 205 x I TABLE &2.(:) 0 I 2 3 • 7.
we note that the zeros in ascending order of magnitude are X~l. .k' (528) This completes our determination of the mode spectrum for the circular waveguide. 3. (519) with the above if. 'hI. . (526) becomes (~)' + k. (520). (527) where n = 0. setting = 'br/X. we obtain the cutoff wavelengths ') TE ( A. . and to the x~p for the TE modes.. :Ira v EIJ _I (I) "p TE_ •  x" P 211'"a VEIJ . TM I1 . is the same for all cylindrical guides of arbitrary cross section if the dielectric is homogeneous. 2... Circular waveguides are used in applications where rotational symmetry is needed. we have (k) ~"p TM _  x"p a (k. The mode propagation constant is dotermined by Eq. degeneracy). Y.. (524) and (528). the mode propagates. and if k < k. a (529) Letting If k k~ = > k.erpretation of the mode propagation constants is the same as for those of the rectangular guide and. Hence. . Referring to Tables 52 and 53.. TE u ..p mode functions ~ TO ~ 11)1 J "a = (~) I cos nq.ve function of Eq. and TEo I (a. we obtain the cutoff frequencies (I) .) . 2. The electromagnetic field is given by Eqs. in fact. (530) Alternatively. and X~I' etc. which with Eq. X01. "p = 2'1fa x' •• (531) Thus.. hence there is no frequency .. from Eqs. The int. pTE = ~ . we have the TE. (525) by choosing k• =~ . Hence. tho cutoff frequencies are proportional to the X"p for TM modes. etc. (This we show in Sec. the modes in order of ascending cutoff frequencies are TEll. TM o1 .) The cutoff wave number of a mode is that for which the mode propagation constant vanishes.and p 1."p ~ =2 k~ x. The dominant TEn umode" is actually a pair of degenerate modes (sin 4> and cos 4> variation). a (526) Using this in the wa.. 1... . X~h Xu. (56). sin n¢J "I'.206 TIMEHAn~rONIC ELECTROMAGNETIC FIELDS We now satisfy Eq. the mode is cutoff. 81.' .
Note that all or these . (Recall that singlemode operation over a 2: 1 frequency range is possible in the rectangular waveguide. 9£ (f) 7M21 FIG. which is plotted in Fig. (532) which is the same as Eq. These can be determined in the usual manner (find £ and :JC. The behavior of the Zo's is therefore the samc as in the rectangular waveguide. (Z. 53. The mode patterns for some of the lowerorder modes are shown in Fig. in the same manner as they did in the rectangular waveguide. (427). range for singlemode propagation. and specialize to some instant of time).tions arc shown in Fig. Hp k. _ ~ H. 59. 43.CYLINDRICAL WAVE FUNC'I'rONS 207 (a) TEll (b) TMOl (c) TBn (d) TMll e~. in a TE mode. Circular wtlveguidc mode patterns. For example.)" _ E.~ (e) 2"Eo'l. except for the degeneracies betwcen TE op and TM 1p modes. The modes of the circular waveguide have HJirected wave impedances of the same form as we found in the rectangular waveguide. 82. Representative cross scc. Modal expansions in circular waveguides can be obtained by the general treatment of Sec.) Note that. TE and TM modes have different cutoff frequencies and hence different propagation constants. ~ _ E. 53. Field lines ending in the crosssectional plane loop down the guide. 54. Attenuation of waves in circular waveguides due to conduction losses in the walls is given in Frob. Solutions for cylindrical waveguides of other cross sections also can be expressed in terms of elementary cylindrical wave functions.
ves can be supported by parallel conducting plates.z). We shall call them radial waves.p)h(k. l In this section some simple waveguides capable of guiding radial waves will be considered. 54. (a) Coaxial. Radial Waveguides.ve functions. Wave functions of the form '" . (d) semicircular.ves.H.. (f) sectoral. (b) coaxial with baffle. (e) circular with bame.z) and h(nq. Wave functions for the guides of Fig.z)h(n~) (ll. 211.p) . but we are using the term "cylindrical wavo function" to mean "a wave function in the cylindrical coordinate system. Radial wa. Examples are given in Prob. 55 to 57.p)j with h(k. the cquiphase surfaces arc parallel planes. These waves have cylindrical cquiphase sur· faces (p = consta. Some waveguide cross sections for which the mode functions arc elementary w&. are formed by conductors covering complete p = constant and 4> = constant coordinate surfaces. that is.) real."'(k.l:l cquiphllJle surfaces.Ol(k.208 TIMEHARMONIC ELECTROMAGNETIC FIELDS b Ca) Cb) (e) (d) (.(kpp) and h(k~z) real. have equiphase surfaces which arc int. . = constant surfaces).±i·· with B. and we shall call them circulating waves.nt).B. DependI These arc true cylindrical waves as defined in Sec. 54 are given in Probs. Such waves travel in the circumferential direction.(k. 53. (e) wedge.) (I) FIo. 510.crsecting planes (the q. we might have wave func· tions of the form '" ~ h(k." regardll$S of il. and travel in the radial direction. In the circular waveguide we have plane wa. Finally.
TE 51n (mT ) cosn411H. Radial waveguides. (56).p) N..CYLINDRICAL WA VI'J FUNCTIONS 209 z y (b) (0) (a) PBrBl1cl plate. and..nd n =. .. and the IJ. Figure 55a shows the coordina.2.5. m (k_p) k.'(k.. . (518) with the above l{!.p) represent outwardtraveling waves. 1. 1.(k.'(k.p») (535) where m '""' 1. we call the guiding plates a parallelplate radial waveguide. (b) wedge. we have the phase constants for the above ~'s given by p ~~  N. cIJ(k.and n .te sy&tern we shall use.r)' The electromagnetic field i!5 given by Eqs.. 3. _ k' _ (533) where m = 0.p) ap [tan' + (536) . . :z 0 at Z = 0 and z . The .(I)(k. aZ JH""(k.p)] J . For a complete set of modes.0. In both the TM and TE cases.. a... . (534) still applies.. 2. by Eq.0. . I. .p») aZ IJ. . Following the general definition of Sec.. iog upon the excitation. Radial waves are characterized by a phase constant which is a function of radial distance. . . FIa. 2. a arc l{!•• ~ . (534) ("'. 2. those with sin nq. 211.TpJ. . variation must also be included. . The electromagnetic field for the TE modes is found from Eqs. . .O'(k. u1 (k. (e) hom.(k. When the waves are of the radial type.p) represent inwardtraveling wa. . TE wave functions satisfying the boundary conditions arc .p) 2 1 . (&19) with the above!/t..cos _ (mT) cosn41 IH. The TM wave functions satisfying the boundary conditions E_ """ E. and Eq.ves (toward the Z axis).p) _ . 5. . waves between the plates may be either plane or radial. the 11.
for . is the modified Bessel function (see Appendix D).p) ~ H.(l)'(k. only the TM o modes propagate. = . H. "" ja. and where K. In this case.I*.l. Using Eqs. H• .lAGSETIC FIELDS Using asymptotic formulas (or the Bessel functions.p) (540) B .. the modes m > 0 are nonpropagating (evanescent). if k..lll(k... Z +.(2)'(jap) Wf K~(ap) (541) which are always capaeitivcly reactive... It is seen from Eq. . Note that the TE wave admittances are dual to the TM wave impedances... Z .>E H.jWf. _ H.·anitraveling TM modes Z +" ~ __ E.. (I)I(k.UI'(k.w~ H. (534) that k. (~) and (518). (537) This is to be expected.p) E. from Eq.(II'(kPJ) Z_. Hence. (538).l2>(k. Note that the phase constant of Eq.TE "" .."'(jap) ~ ja K. ~ = 11.p} E..I. for TE modes Z TIl: +" = E.. ... The radial wave impedances become imaginary.ja.(ap) j<M. we find for outv.H • .. we find that (or real k. The mode functions are now everywhere in phase.p} Note that for real k. ~ H. where the first equation applies to outwardtraveling waves and the second equation to inwardtraveling waves. ~ ~ ja H..t large radii the waves should be similar to plane waves on the parallelplate guide. For example. jWIl T.. is positive and K~ is negative.! k. H. Components of E and H transverse to p are not generally in phase.Ol(k.TI. we find (539) Similarly. indicating no power flow. let k. (536) is that of tbe mode function and not that for the field... and there is no wave propagation.. H .p) (538) while for inwardtraveling TM modes k. is imaginary if mr/a > k. Each mode of the radial waveguide is also characterized by a single radially directed wave impedance. For small 0. H .TW = Z+. They become in phase at large radii. fJI' II.210 TW}:HARJ. since K.p} . whenever a < >'/2. we have Z_.lONIC ELECI'RO. because a."."'(k.
Ol(k p .t· =:. For this mode we have (544) representing inwardtraveling waves. Figure 56 illustrates this behavior by showing XI R.(kp) ~ IH. We shall call kp = n the point of gradual cutoff.nd a net radially directed current on one of 3 \ 1\ 1 0 K1 2 kp o 3 4 5 FIG. 56.CYLINDRICAL WAVE FUNCTIONS 211 which Eq. where Z+~Ttd = R + jX.. H. From the above discussion it is evident that the TM oo mode is dominant. the mode being TEM to p.ip  + N. . (&33) reduces to ~. H.+ = .cos n~ lH.u'(kp) )! (542) From Eqs. propagates energy effectively at smaller radii than any other mode.p < n)...HoW(kp) k' (545) 4 3 2 n 4 which represent outwardtraveling waves. For example.(kp)]! (543) A consideration of the behavior of the Bessel functions (Figs. because of its similarity with plane transmissionline modes. It is called the transmissionline mode of the parallelplate radial guide. JY] . Ratios of wave reactance to wave resistance for the TM Oto radial modes on the parallelplate waveguide. ~ _ /H. at a given radius we can calculate a unique voltage between the plates a.n..'''(kp) H. that is. Z_. Hence. the wave impedances being predominantly resistive when kp > n and predominantly reactive when kp < n." the wave impedances become predominantly reactive.U)'(kp) jIJ. (or the first five TMo" modes.(kp)J. Note that these gradual cutoffs occur when the circumference of the radial waveguide is an integral number of wavelengths. (538) and (539) we have the wave impedances for these modes given by Z+~TM =:.+ = kH 1(2'(kp) E.. Dl and D2) reveals that {or arguments kp < n the N" functions and their derivatives become large in magnitude. Note that there are no p components of E or H. and JW. when 2..(kp)N.7(kp)i' l.
and (549) for outwardtraveling waves. called a wedge radial waveguide. The dominant mode is evidently the TE g mode. 1. .te guide [Eqs. but if numerical results are desired we would be hampered by a lack of tables for functions of arbitrary fractional order.po ~) IHr~/ H. This is as we should expect from our knowledge of plane waves between p$l. the radial transmission line can be a.(kp») (546) where p = 1. Again the transitional point is that for which the argument and order are equal. 3. Also. parallelplate modes. chamc· .. = cos (1'cPo ~) IHf~/··(kp) I < Hp. TE wave functions satisfying the boundary condition E~ = 0 at </J = 0 and t/> = tPo are l/!pT". and the elccLromagnetic field is given by Eqs. (518). '[J7f/q. from Eqs. We shall assume no z variation of the field. that is. 55b.. by k.212 TIMEHARMONIC ELEcrnoMAoNETIC FIELDS the plates. These wave impedances exhibit the same characteristic of gradual cutoff for fractionalorder Hankel func~ tions as they do for integralorder Hankel functions. .g = kp. "'" 0 at t/> = 0 and t/> = cPo are Vtt)'rM "'" sin (pr . except that nonintegral orders of Hankel functions appear. We nccd only replace n by 1J7r/q... The radii so determined correspond to those for which the arc subtending the wedge is an integral number of halfwavelengths long.nalyzed by the classical transmissiollline equations with Land C a function of p (Prob.'4>. (538) to (540)].o and k. 513).0).r~lIel plates (the limiting ease rJ'O'. Radial waves also can be supported by inclined conducting planes.••(kp) (547) where p = 0. The interpretation of the modes is essentially the same as that for the TM o. in which case.(kp) •. The radial wave impedances for the wedgeguide modes are of the same form as for the parallelpla. we have (519). TM wave functions satisfying the boundary condition E. (547) and (519).and the electromagnetic field is given by Eqs. . as shown in Fig. 2. This introduces no conceptual difficulties. This is a transmissionline mode. 2. considering the problem as twodimensional. __ . . . (548) for inwardtraveling waves.
modes propagate if >. It is a simple matter X to modify the circular waveguide mode functdona to satisfy the addiFlo. 44). (551) The TE modes are specified by the mode functions fo. tws being the radius z at which the guide cross section is about the same size as a rectangular waveguide at cutoff. 1. U a 1d eeetion of circular waveguide is closed by conductors over two cross sec.. and so on.o H •. The TM modes are specified by the wave functions 1/I. The field is given by k. 55c. . simple radial waves can be supported by the homshaped ~e of Fig... (544) Illd (545»). no transmissionline mode. the Circular Cavity.0. The field is given by (~19). (If a > >. . ~) (H~2. Finally.'.2) where m ..2. 2. sin(~%)C<ls(""'~) (H~/. these plus the TM 1.TM _ cos (~z) sin ("". or H.qs. VE Pt1NCTIONB 213 terised by no E.3. and TEl. There is. This mode also can be analyzed by the classical teall&missionline equatdons for nonunifonn lines (L and C a function of pl." •• (k. = ~k' _ (m:)' .. 2. Each propagating mode has a radius of gradual cutoff./2 one might argue that the TE lo mode is dominant at small radii. (518)./2.. 3. Eq•. The TMo l mode is usually considered as the dominant mode. 1. 57. and .. of course. These modes are qualitatively similar to the hybrid modes of the rectangular waveguide (Sec. . and p ". we have a resonator known &8 y \he circular cavity. Not< tbat tbe field is dual to tbat of the pamllelplate line (Eqs. This is shown p in Fig. circulAr uvity. and possessing a unique voltage and cuceent at aD1 given radii. ~7.p) (k"')1 a q" H • .T< _ F. Only the TMOJo modes propagate if a < >.2. T •  .(k. 1.CYLINDRICAL Wit.1. The.p) (550) where m .p)1 a q. because of the single conducting boundary." (5. . . tions. by Eq. called a sectoral horn w<MJ6fluiM.0.) 64. . J and p . (551)..··(k. ../2 < a < X. . and k..
FOR THE .13 2.' 1. 72 1.29 2.59 ~ TM ul TE t " ~ TE I l I ~ TM lIO TMoio 0 0. . .31 2.57 1. The X Il .27 2.0 1.96 2.)' .)"""". (56)] becomes j and the 1.\VITT OF RADIUS a AND ENOTJl d eL rnCULAJ'l. .00 3. I. ..80 1. =. .. j'P = field is given by Eqs. .42 1.05 1. . ~x~" + (~. ~ J (x~.211 2.13 1. . . we can 1 2rG  (j. .3.3. 2.0 1.13 2. . 54. The set of modes TE to z is specified by ~.59 1.31 2.p) ·)If II Icos n~) 1sin nib sin ('1d" .Ct".80 1. .63 1.0.24 1.08 1. The separation constant equation IEq. .0 4. 2.u~1 e.78 2.66 3. for the TM and TE modes.q = 1.. and q = 0.52 1.08 5.19 1..13 2. p =: 1.56 2.20 1.. (519).00 .00 2. TABLJ. 2.y.59 1. 2. and x~p are given in Tables 52 and 53.0 2.0 3.29 2.50 1.)do_.0 2.)':. The field is given by Eqs. .72 1.72 1. 1.0 ~ 1.41 2. pair of degenerate modes (cos n¢ or sin nIP variation).:'ted by = 0 and Z E: d.0 1. respectively.87 1. TEsli ~ d  • TM Q1t TE IlI ~ TM ut TM oII 1.) (&54) wheren 0= 0.0 1.214 TIMEHARMONIC ELECTROMAGNETIC FIELDS ? tional boundary conditions of zero tangential E at The result is a set of modes TM to z.. (518).27 1.29 2.00 1.2.80 1.20 1. .13 2.91 2.2. Vt~:q = J ft (X.0 1.. solve Cor the resonant frequencies {fl Tlil r IIPO _ Setting k = 2701 V. 3. 1.U. speci.p) {:~::} cos (~ z) a (553) where n .50 2.59 1.0 1. ~X"p 2+ ()' d qra (&55) Each n except n = 0 denotes a. .32 1..0 1. The resonant frequencies for various ratios of d/a are tabulated in Table 54. ~ 2•• 'v..0 1.60 3. .
~ Fla. 68. Mode pattern (or the TM1 • 1 mode (dominant when dlo :$ 2).t... To . t 0. (553) and (518) we determine the field components of the mode as Following the procedure of Sec./ / / ~ . e. < 1. The TE IIl mode corresponds to the first resonance of a shortcircuited circular waveguide operating in the TEn mode... The TM OIO mode corresponds to the first resonance of a shortcircuited radial transmission line. If dlo. Note that this is very similar to the squarebase rectangular cavity of small height (the mode separation is 1.0 corresponds to that of a twodimensional circular resonator... while for dlo. which is dominant for small d. Its mode pattern is thus that of a standing wave in a circular wnveguidc ../ / 0"\' ••••• ••••• ••• •• \ • ° • • • • • • • • • • • • • . "" .:. _ !5: 2. we calculate thc stored energy in the cavity as w ~.. 58.. 530. ~ ..' (... From Eqs.. is shown in Fig..d ""f}O (. The last row of Table 54 therefore is also the cutoff frequency spectrum of the circular waveguide. the sceond resonance is 1.!J(. 28.::::" .. 1 / /~" /. The field pattern of this mode.\ . ". . similar to Fig. especially the Q of the TM OIO mode (dominant for small d).p) dp a ffl IEI'dT . The Q'S of the circulnr cavity are also of interest. Note that for dla < 2 the TM tII mode is dominant..nt frequency.. . 0\ .59 times the first resona. ~ 2 the TE IIl mode is dominant. . "fl ':0) r \\ ... The case dlo.58 in that ease). for which the resonant frequencies are the cutoff frequencies of the circular waveguide. .CYLINDRICAL WAVE FUNCTIONS 215 / I . pJ.
X'')' adJ (a 211" [ ( l t %01) + 2 J.. Jahnke and F.. Dover Publications. ..z)J cos nlf> H . For the partially filled radial waveguide of Fig.' the result being rk 4 cIa' W = 2 J 1'(XOI) The power dissipated in the conducting walls is approximately "" (556) rf>.l and we obtain iJ>d = at (X~l Y 2Ta(d + a)J 1'(XOI) ~ (557) The Q of the cavity is therefore wOW Q. 146. The Q's for the other modes of the circular cavity are given in Prob.c + a) Recalling that the condition for resonance is ka = simplify this to 1.216 TIMEHARMONIC ELECTROMAGNETIC FIELDS This is a. as well as their qualitative behavior.p! = C.' pJl'((X"p) dP] 0 a The above where /R is the intrinsic wave resistance of the metal walls. The methods of solution for the systems of Figs. The geometries of some other cylindrical systems capable of supporting guided waves are shown in Figs. known integral. 4." p. we can (558) where f/ is tbe intrinsic impedance of the dielectric.. for the same heighttodiameter ratio.{!l(kpp) = c! cos Ik. 59a. 55.s to be expected.(a ..'(d k 4 da' n=::"".<R = <R 1f> IHI' d. New York.o.3 per cent higher Q than the rectangular cavity. cos k. since the volumctoarea ratio is higher for a circular cylinder than for a square cylinder. This i. Q . 59 and 510. 516. We treated the analogous planewave systems in Chap.ll.(!l(kpp) (559) 1 E.<R(1 + a/d) X01 = 2. 4.'. Other Guided Waves. 1945 (reprint). integral is again known..405. = 0 at z = 0 and z = a by choosing ::II Vtl . the circular cavity has an 8. cJl. we can obtain fields TM to z which satisfy the conditions E p E. (2102»). . are similar to those of Chap.202.z cos n~ H . 2wt(]b.. "Tables of FunctioD. Erode. This can be compared to the Q of a squarebase rectangular cavity [Eq. 59 and 510. It is seen that..
].(!) replaced by 8.. . (c) coated conductor. Equations (559) represent outwardtraveling waves. fl ~1 1[ (1 1)] . . We have anticipated that the p and 4J variations must be the same in both regions to satisfy boundary conditions at z . (559). kit CI sin kold = k.1)] tJI! ~ up uZ «=l fl fl .. we ha. = :which reduces to JW 1[ a' (1 tJll . iJ4J 81. (6) dielectric slab. 2. l .(1). of course.1/11 Jwi/.. To evaluate the G's and k" we must satisfy the conditions that E" E. H" and H.a' tJlt ..! k.. C sin k.CYLINDRICAL WAVE FUNcrlONS 217 z Conductor z z (c) FlO.. (d) corrugated conductor. using the y. where n = 0.. satisfy the separation relationships k. (518). . we have (E.Ii) ft (561) = For E. .. Some radial waveguides.:I = . 69.1 0 .E"J.I . (a . 1.. The k's in each region must. respectively.ve [B.'s or Eqs. The subscripts 1 and 2 refer to the regions z < d and z > d. (a) Partially filled. . Inwardtraveling waves would be of the same form but with H .! + + kIll kl!! DO kl! = k!! = W!~1PI W 1t!}l1 (500) The field vectors themselves arc obtained from Eqs.1 = 0 .. be continuous at z = d. For E.E. d.
for TE modes...ll J.)]..d») ts The kd and kd are fun~tions of k. .("'.. (563) is a.'s..U)(k.2.4 ~0 Division of Eq.)] ap which a. yields k. as shown by Eqs. (561) by Eq. (563) and (565)] are of the same form as those for the partially filled rectangular waveguide {Eqs.l2 (565) as the equation for determining k.[!. . For H. for H. since at large p the Ha.(a ."'... (562) (563) ~ _ k. is evaluated.. The Matching field components are found from these ""s by Eqs. it can be analyzed by conventional transmissionline concepts.t(a . The modes of the partially filled radial guide can be ordered in the Mme manner as were the modes of the partially filled rectangular waveguide. (D13).nkel functions reduce to plane waves. 1zcosnq..H. For fields TE to z we can satisfy the condition E.. . 1.d) Finally. tan Ik. = Ot sin k.218 TIMEHARMONIC ELECTROMAGNETIC FIELDS which also reduces to Eq.]. the ratio CI/Ct may be obtained from either Eq. It reduces to the radial transmissionline mode in the empty guide and has no cutoff frequency. (561) or Eq.p) where n = 0. It should be apparent from our treatment of the waveguide of Fig. . tangential components of E and H at z = d yields k d cot k. (562). c. (562). ".z) cos nq. (560).gain reduces to Eq. 59a that the characteristic equations for the radial waveguides of Fig.. (~)(k. . (519). . It is interesting to note that the characteristic equations for the partially filled radial waveguide lEqs. (560) must again be satisfied.d)] J. we have IH. Once k.1d __ kd cot [k 12(a . The dominant mode is the lowestorder TM mode (logically designated the TM oo mode). = 0 at z = a by choosing 1/11 = C1 sink. according to Eq. H . we have . transcendental equation for determining possible k. tan k"d tl . 59b. j and Eqs. so Eq. For a « 11. which reduces to H.. (561)...p) (5<H) tit.. . (445) and (447)}. ~ 0 (562) 0 1 cos kdd = C S COB kd(o . = E. This we could have anticipated. [aa~ ("'.
because. b. by (~) for modes TE to z.. (471) in the planewave case. 410. except for rota.CYLINDRICAL WAVE FUNCTIONS 219 and d will be of the same form as those for the plane waveguides of Figs. and 415. Hence.. in region 1.2" cot 2" k~t un + k.'s. Finally. = 0 over the midplane of the slab..·t(k... We can consider them all at oncc.lP) sin n.t = . cit.. I(k. 59b. 59c are those of the slab waveguide having E.'s by k. The modes of the coated~nductor radial waveguide of Fig. the modes arc neither TE nor TM to any cylindrical coordinate.413. We then choose electric and magnetic ""s ~ ". efk.lP) cos nq. and c have the common property that they are If twodielectric" problems. which has no cutoff frequency. (463). the lowest TE and The cutoff frequCDcies of the mades in general are given by Eq. 411.. 5100. cj1l.tionally symmetric fields. = wtf~ (568) kot wtE~O Possible solutions to these equations can be obtained graphically by the method of Fig. = ~' = (570) (571) CB. We need only to repla. and ~ ut "2 taD "2 "" "" 2"cot2" "" "" "" to k.. The circular waveguide systems of Fig. The dominant mode is the lowest TM mode. 510 are interesting. The systems of Fig.. r i " •• .t _ _Ill + k. The cutoff frequencies of the modes in general are given by Eq.. BB•• I(k. DB.. as follows. = E. the characteristic equations are for modes TM to z. for the dielectricslab radial waveguide of Fig. the characteristic equation for the dominant mode is TM modes have no cutoff frequencies... (4&1). = k o VI + tan' kod (569) This is analogous to Eq..l . Let region 1 be the inner dielectric cylinder in each case and region 2 the outer one."'(k. y. I :z A B. Just as in the planewave case.tP) cos nq. k. for the corrugatedeonductor radial line of Fig. and ". 59d. The u and v are related l ~taD~ 2 2 (567) .tP) sin nq.ce the k.
(e) coated conductor.sa) WEJ<! Ak. 1B.220 X TLME1lAIWONIC ELECl'ROHAGNETIC FIELDS Z X Z f~"2 / .oI'(k.t' + k.tB.ta) => a lead to pJk"IDB. (d) corrugated conductor. (518) and the 1ft' determine partial fields according to Eqs.n B.I'BB...determine partial fields according to Eqs. ~10_ Some ci«:ular waveguides."'(k.. The "'....t(k..k. (6) dielectric alab.sa) Ck.. The B .·t(k...1/(k.. be continuous at p E.) (d) Pto.d(k.' "'" k k. in region 2.1B"d'(k. 1a) WEta + Bk..ry conditions except those at the interface p = a.. 1a) = wp..1(k. The total field is the sum of the two partial fields in each region. sa) ClIp.. and E.n B.l(k'la) = E1k"tCB.. chosen 80 as to satisfy all bounda. H.. 1a) _ + Dkln B. k k.t a) AKin B.p) denote appropriate solutions to Bessel'l:! equation of order n._l'(k.la + Bk.I' t  1' 1' "'" W El/ll ".p'S must satisfy the separation relationships + k.sa) These equations have a nontrivial solution only if the determinant of the ..(k.d(k.1B.l'AB.~ + Dk. (4) Partially filled. cU'" y Conductor (a) y (b) Z X Z y y (. (519).sa> IAtk.'I(k.. In each region the .'ElIlt t = (572) = The requirements that H" E..za) Ck.
(573) F s "'" B .(k.IFiF. =0 (574)  .sk. "'" 0 at p "'" b. the field must decay exponentially above the cutoff frequency and represent outwardtraveling waves below the cutoff frequency.(k.la) S(k. we choose (579) F. . which reduces to the TEll mode 01 the empty guide. = JII(k.n .IF:  k.tFIF~ .4). the field must again be finite at p = 0. and its cutoff . EI = Eo. However.ta) Hence.. B. 510a).kplF.. the dominant mode is the lowest n = 1 mode.(k. external to the rod. = 0 (575) lor TM modes (n . 511 for the case EI "" lOEo.la) (577) b.  . When n = 0. For the partially filled circular waveguide (Fig. For the dielectricrod waveguide (Fig.I(k.(k"b) Furthermore.. F I "" B Fa = B I(k.sF.t'F I 0 k.F'.0.(k"a)J:(k.CYLINDRICAL WAVE FUNcrlON8 221 eoefficien ts of A.ob) The dominant mode is the lowest<lrder n = 1 mode. defining. and k.)N.a f. to satisfy E.sP~ k'IF~ W'" . and D vanishes.ta) F." 0). ~ J. the field must be finite at p = 0.sF s 0 0 0 Illk..)J.sa) The characteristic equation in determinantal form is f..sF tF~ .n F k. so Eqs. "". C.lk. and the characteristic equation is much simpler. = 0 at p = = P. = B. A solution for the k. we choose (578) F. = 0 (576) for TE modes (n = 0)..N.)N. the field separates into modes TE and TM to z. hence FI To satisfy E... III = "'" ra Ilo.J.F .(k.k.n F k.' k.ob) .ob) . W""Ia W". Hence. of this dominant mode is plotted in Fig.I'Fs k..ot(k.n F k. b = 0.N. It is k.. we choose (580) Once again. (5~77) still apply. We must now pick the proper F functions for the various cases.(k. 5lOb).s2F4 IlJc.
222 TIMEHARMONIC ELECI'nOMAGNETIC FIELOS 3 l/ 1 / 0.ially filled circular waveguide." pp. (After H. . N. < k l . C.8 1. Van Nostrand Company.4 0. ° 0.2 0. of the dominant mode are shown in Fig. 5lOc we must again have exponential decay of the field as p + CIO. 512. Inc.1. For the coated conductor of Fig. 1943..4 0. so Eqs. Seidel. Phase constant for the circular dielectric rod. A. which is the same relationship that applies to the dielectricslab guide of Sec.6 al>' FIo. frequency is zero. f  10 '0 3 V I V 00:4(0 • I 2. Princeton.0 (After M. SchelkunofT.) 10£1.0 alb II  Flo.. However.4).. b _ 0.6 @J " " . Note that ko < k. 47. 1 Some solutions for the k.5 4:1 ~G 0. 512 for the case E1 = EO and ~l = Jl2 = po.) ° 02 0. GrG1I. "Electromagnetic Waves.J. 511.0.8 1. PhlLSC constant (or the part. (580) still apply. 425428. to 1 S.
.CYLINDRICAL WAVE FUNcrIONS 223 satisfy the condition E. In this section we shall consider twodimensional sources of cylindrical waves.he coated wire. no. . = 0 at p = b.t for the corrugated plane. '" should be independent 1 G. the solution approaches tha.a)J. 513. From the theory of Sec. 511).hat used for the corrugated plane (Fig. (Compare it with the dominant mode of the plane coated conductor of Seo. Suppose we h. . Proc. 29.lb) F1  J.N. Sources of Cylindrical Waves. Surfacewave Tra. IRE. that is. 51Od can be analyzed in a mAnner similar to t. As the radius of tbe corrugated cylinder becomes large.(k. The extension to three dimensions can be effected by a Fourier transformation with respect to z (see Sec..(k. I Finally. The characteristic equation is obtained by matching wave impedances at the corrugated surface. Goubau.(k. which bas no cutoff frequency. From symmetry..la)N~(k..nlfmission Linee.la)J~(k . m..".b) and.b) . sources independent of the z coordinate. to satisfy E• ... 0 at P .Q current along the z axis.8 shown in Fig.a)N. 619624.lb) (582) For this guide the dominant mode is the lowest n r: 0 TM mode.N.(k. fi·6. we should expect the field to be TM to z. the corrugated wire of Fig. 39. expressible in terms of an A having only a z component "'. 6.(k. June. = J. 0. 513a.. The field extel'nal to the corrugated wire will be essentially the dominant TM (n . we should choose (581) F.) Copper wire with an enamel coating can be used &8 an efficient waveguide for some applications. pp. 1951. 0) mode of t. vol. The field in the corrugations will be essentially that of tho shorted parallelplate radial transmission line.(k.&ve an infinitely long filament of constant &. 415).". An infinite filament of collltaDt ae current (0) along the I" axis aDd (b) placed parallel to the I" axis. 48. b. z I y 'I" Y P' . X X p Ca) (b) FIo.
4j H.pd¢ = I ~o'f Evaluating H := V X A. I A. lines of electric intensity run parallel to the current. The outwarddirectcd complex power crossing a cylinder of unit length and radius p is P. [H. using the 1ft of Eq.""(kp)]' X = (SJ" (586) The real part of this is the timeaverage power flow which.! Ikl!'H.rp rT H."1 (585) which is essentially an outwardtraveling plane wave.)] ~ j2C The preceding equation then yields C=4j Hence.224 TIM."'(k. _ ."'(k.H:pd4J _". and lines of magnetic intensity encircle it. The result is (584) Thus.kp gli. The electromagnetic field is obtained from Eqs. j '. . Equiphase surfaces are cylinders. (583).kl ~s. we choose A.  ."'(kp) where C is a constant to be determined according to lim '+'H. in contrast to the r 1 variation in the threedimensional case. The amplitude of the wave decreases as p~t.imensional source.) (583) I is the desired solution. The line current is the elemental twod. by virtue . To represent outwardtraveling waves. (518). but E and H are not in general in phase. = 1PE H··ds 102'r E. we find 8p dp t_O rp H.Of _ C 2. = kl ~8.. However.CH. at large distances we have E. .~ . 29) is the elemental threedimensional source.~ . just as the current element (Sec.EBARMONIC ELECTROMAGNETIC FIELDS of ¢ and z."'(kp)[H..
(&85). as we should expect.ted at '1 by writing A.. as represented by Fig. we can extend Eq.'1 .y' as shown in Fig."'lO""'+:I.directed current can be represented by a summation of the A/a from each current clement.) ~ l~. (017)1. 514a. (0) Dipole aource./ H.'X"" (aj Flo. (6) quadrupole aoUl'Ce. (583) by replacing p by the distance (rom the current to the field point..(p) and that J is located at p' by writing I(ri). . the timeavemge power is independent of the distance from the source. .. (583) to read A. In radius vector notation.L WAVE FUNcrIONB 225 of the Wronskian [Eq. As the separation 8+ 0 and the magnitude 1_ 00 such that 18 remains constant."'(kl. Sources of higherorder wave. we specify the field point by puzX+UI/Y and the source point (current filament) by '1' . If the current filament is not along the z axis but parallel to it.. It could be more simply obtained from Eqs. The solution for two or more filaments o( z. We can now generalize Eq..(. Suppose we have two filaments of equal magnitude but opposite phase. 513b. point is then The distance (rom the source point to the field x')' I..' + u. .V(x = . is evalua. reducee to /PI  Re (PI) _ . ~I4.k Ill' 4 (587) Hence.Uz:z.OYLINDRICA.'1) (588) This is our (re&space Green's (unction (or twodimensional fields.. we have a twodimensional dipole y y I j.pI + p'l + (y 2pp' if)' cos (I/> 1/>') We emphasize that A.
. by reasoning similar to that above.y) . the vector potential of a quadrupole line SOurce is a. 529) that. The dual analysis applies to the case of magnetic current filaments. from Eq. .A. the vector potential of a dipole line source is a cylindrical wave function of order n """ 1._ _0 8 (JA. 4.X'. (588). 514b we have. wave function of order 11. Hence. Note that A.l is that due to a single current filament at the origin [Eq.p'[) (591) Using both electric and magnetic multipoles.. For the quadrupole BOUTee of Fig. \ kIsts! a A. = which reduces to k2~?8t 1:l 2(2l(kp) sin 2q.. (590) Thus. oy [H."'(kp) cos ~J A. Hence. A.'''(kip .0). (583)].O) is the same as A. we can generate an arbitrary sourcefree field in homogeneous space (P > 0). 514a. at (x . where (589). the electric vector potential at p due to a magnetic current filament at p' is F. We shall call such a SOurce a multipole source of order n.. In the limit 8 + 0 the above equation becomes A. the vector potential is A.226 TIMEHARMONIC ELECTROMAGNETIC FIELDS source.'(Z +~.(2) is the vector potential of the dipole 8Ouroe.coo on an infinitesim. a possible souree consists of 211. when A. A. for Fig.al cylinder. It is merely necessary to replace I by K and A by F in the various vectorpotential formulas of this section.A+ ~. given by Eq. It can be shown (Prob. = 2. This procedure can be continued to obtain sources for the higherorder wa.l ax = .(p} = K~r) H ..Y) due to a current filament at (0. . 4j H 1 (2 J(kp) cos ¢ Thus. at a point (xtY) due to a current filament at (x'.(H o (2)(kp)] 4J ax (589) 18 a The differentiation yields kl. current filaments equispa.ve functions.y) where A. is a wave function of order 11. For example.
(The general case is considered in Prob. .l A.jl[i where P is the complex power per unit length p = . (518) with the above ¥to and satisfying the boundary conditions.kaJ.kaJ. we obtain .(kp) E..(t'(kp) p<a p>a The boundary conditions to be satisfied are where J.._ ..J:ad4l = r 2raJ:E.H. A cylinder of uniform cmrcnt.+ ~ C.J.J. is the density of the zdirccted current sheet. 515.) The geometry of the problem is illuSotrated by Fig.2 .Jo'J E.CYLINDRICAL WAVE PUNCI'lON8 z Fro. 530.I. Let us calculate an impedance per unit length for this source.. " 1. By definition.'''(ka)J."'(kp) T p <a (592) p>a as the only component of E. 515. as we did for the ribbon of current in Sec. Using Eqs.(kp) ~ CI H. p Z . Because of the rotational symmetry. y The field due to a cylinder of Currents can be obtained quite simply by treating the problem as a boundaryvalue problem.(ka)fl. We shall consider here only a cylinder of uniform zdirected surface current. we choose of I A. 412.
Tho resistances (real parts) are identical."'(ka) 4 Using smallargument formulas for J a and H 0(2) I we obtain (593) . 67. M88S" 1956. Since the equations for A z due to J~ and for Al' due to Jl' a. it CaD be shown l by a quasistatic approximation that the impedance per unit length of a small elliptic cylinder of minor axis a and major axis b is the same as that of a circula.j II J. W. . the above equation also applies for z replaced by x or y. (4127)].k J. "The Theory of Linear Antennas. The reactance of a cylinder of current of small diameter d is approximately equal to the reactance of a ribbon of current of width w = 2d.. Harvard University PreM. we have A.781. Compare this with the Z of a ribbon of current [Eq..re of the same form as those for A~ due to J o.228 TIMEHARMONIC ELECTROMAGNETIC FIELDS a. vector potential 0 dA.j II J(e')H. King. P. Twodimensional Radiation. We can construct the solution {or an arbitrary twodimensional distribution of currcnts by dividing the source into elemental filaments of current and summing the fields {rom all elements. we have the vector equation A(e)  .. Hence.f(a + b) ) A ribbon is the special case a = 0 and b = w. (594) where"Y = 1. . ~ . More generally.(e')H. each element J ds' produces &. if' we have a J~. ( "J21og. For example. 1620.a d~ ~ 2~aJ.nd I is the total zdirected current J =}o f 2. r ka z __ . J.(ka)H. independent of z. cross section of thc source.. Combining components.) O2"2 la."'(kle . entire sourcc. the impedance per unit length is z ~ ."'(kl.e'l) d8' (595) I R. ~J418' Ho"'(kle e'l) Summing over the where ds' is an element of area perpendicular to z." pp. Ca.r cylinder of diameter d ~ ).e'l) dB' where the integration extends over a.mbridge.
as usual.&16.f"I+ P  p' cos (~  ~') (597) The second term must be retained in the phase factor.fl'l large. The electric vector potential due to twodimensional magnetic currents M is given by the formula dual to Eq.£1'1). Hence. y Fio. (395) in the threedimensional ca.V X A. When the field point is distant from the source. the Hankel function can be represented by the asymptotic formula Furthennore. we have If' . Geometry (or determining the radia.. exp (jkle . or F(p)  i ff j M(p')H.p'l) M (596) 'the electromagnetic field in this case is given by E = .. .V X F..ce currents and current filaments are included by implication. The electromagnetic field is obtained. as shown in Fig. the vector potential. from H . 516. So". 10 . form similar to those for threedimensional radiation (Sec.sc. our formulas simplify to a. of Eq•. The cases of surfa. (595)."'(klp .CYLINDRICAL WAVE PUNcrlONB 229 representing the solution for an arbitrary twodimensional distribution of electric currents. x . (595) and (596) reduce to (598) provided p »p~w These arc the radiationzone formulas corresponding to Eqs. 313). when p» p'.(l'l~. but not in the magnitude factor. For klo . tion field.
he same.. H~'. Suppose we have the plane wave eiz .. The corresponding E~.e""..A. except for the contrasting pJ1 and r. (599). Note that. or ) E.. so the superposition of fields from all current elements should also be of this type. the radiation fields are of similar mathematical forms in two and three dimensions.. given by H' = V X A. It is often convenient to express the elementary wave functions of one coordinate system in terms of those of another coordinate system. are constants. rip . (585).:o:: • L. in the radiation zone. . 51 is assumed.jwp. The total field is simply the sum of the primed and doubloprimcd components.. E. the distant field of a single current filament is essentially an outwardtraveling plane wave. To obtain the field components.jkF. . . Hence. (599) which can be verified by direct expansion of Eqs. given by E" = . Some representative wave transformations are derived in this section. E~. To evaluate the a. . Wave Transformations.A. A13 evidenced by Eq. . .. (5100) in the radiation zone. which we wish to express in terms of cylindrical waves.230 TIMEHARMONIC ELECTROMAGNETIC FIELDS We now have the p variation explicitly shown in Eqs. and simplified formulas for the radiation field can be obtained. (The conventional coordinate orientation of Fig.H. let us again divide the field into that due to I. even though they may be geometrically t. E. a.V X F. (34). with H given by Eqs. we obtain H. Others will be derived as they arc needed. This gives 10 2 . = jwp.. it muet be expreesible 80S rf:r< = ei. E. (599). E~' = H~ = jkA. in the radiation zone. Hence. (598). jkF. (p)e"'· where the a. E': = jkF.. + jkF. These formulas correspond to Eqs... (598).. d</J = 2'11"a""/. and that due to M.. Retaining only the dominant terms (p~ variation).. using Eqs.te systems are considered to be distinct if their origins or orienta· tions are different. and H:' can be determined from Eqs.' We refer to expressions of this type as wave transformations.J . (397) in the threedimensional case.(p) I Two coordina. = jkA.1 dependences.) This wave is finite at the origin and periodic in 2'11' on 1/>. multiply each side by r~ and integrate from 0 to 2'11" on </J. 68.H.
141. "Fourier Series and &undary Value Problems.. .. The mth derivative of the lefthand side with respect to p evaluated at p .= l j... for'" must represent outwardtraveling waves.(p')H•.. Also.. MeGrawHill Book Company. New York. In the region p < p'... and use the asymptotic formulas for the Hankel functions. ...0 p ". Churcbill. (I)(p)e../2.0.... In the region p > pi. f is of the form • b. and we have shown that eP and also that = e1.(P) (frI02) Equation (5101) is the wave transformation expressing the plane wave ria in terms of cylindrical wave functions. Rcnce. but we need not recognize this. 1941." p. are constants. 0 is 2". .0 is j J. '" must be symmetric in primed and unprimcd coordina. O.(p)ei".tes (reciprocity). Our original I R.. We shall reexpress '" in terms of wave functions referred \:.....te thc b. Consider the wave function '" ~ H... It is closely related to the soealled "generating function" of Bessel functions. let p' + 00 and 1/>' .(I)(I(l  (I'D = Ha(ll(Vpl + p" 2pp' cos (41 41')] where p and p' are as defined in Fig. . Tbe mth derivative of the righthand side evaluated at p ..CYLINDRICAL WAVE FUNcrlONB 231 The lefthand side is actually a wellknown integral.'(p)ei".! Another wa... on 41.... p> p' wbere the b. for'" is finite at p "'" 0 and periodic in 2. To cvalua. • (p)e iro• (5101) J.. Hence."·. V.. We can think of '" as the field of a line source at p' in terms of a cylindrical wave function having its origin at the source.H.J .J. permissible wave functions are J.0.ve transformation of interest is that which corresponds to a tmnslation of cylindrical coordinate origin. 2 jcos· <p e~ dt/J __r_ _ o 2 . Inc. permissible wave functions are ll. n an integcr. n an integer. 513b."'(p')J.(p)ei".. p < p' It _ _ • It _ _ • L • L b.
. we can express the incident field as • E..(1.. that is. An addition theorem for Bessel functions of the second kind is obtained by subtracting that for H O(I) from that for HoOl. (5105) Using the wave transformation or Eq... (5101).(p)""f'" J. i·J .... 2: ."'(p')J. Extension to threedimensional cases can be efleeted by the method 01 Sec. Thus.. .(kp). It is also valid for superscripts (2) replaced by superscripts (1). and. . • (5104) which is the addition theorem for Bessel fUDctions of the first kind.(P')J. as represented by Fig. Adding the addition theorem for BoU) to that for Bo(l).(!)·.. 2: 2: • • H ... 59...232 TIM... 512. We shall at present consider only twodimensional cases..MONIC ELECTROMAGNETIC FIELDS and our constructed expression for !f becomes These are now representations of a plane wave. 1.. ... F < p' (5103) p> p' This equation is known as the addition theorem for Hankel functioDs. 517. we obtain J .. it follows that b. Take the incident wave to be zpolarized.  . (5101). A source radiating in the presence of a conducting cylinder is onc of the simplest "wavescatter" problems (or which an exact solution can be obtained. Let us first consider a plane wave incident upon a conducting cylinder.&RAR.(p')H.Ul = H. since H.' = E."'(p) ....(p).. Scattering by Cylinders.. from Eq.'1)  2: J .
we .'''(ka) J. form Eo' . A plane wave incident upon a conduct. (5107) and (5108). 617...) (5107) At the cylinder the boundary condition E• .(k. ing cylinder. and simplifying the result by Eq.t is.. (017).H..ves. E.. and we have essentially a filament of current. .a must be met. Using the sma. tha. the scattered field must be of the Eo • • 2: j"a. The surface current on the cylinder may be obtained frOID Using Eqs.."'(kp»)e"'· ... '.llargument formula for H o(2). 0 at p . ~ E. m(kp)ejr... E.~ (5106) hence the total field is E...H .CYLINDRICAL WAVE FUNCTIONS 233 y Flo.. ~ H.' + Eo' To represent outwardtraveling wa. we obtain (5109) In a thin wire the n = 0 term becomes dominant..(kp) + a. p = Incident wave x The total field with the conducting cylinder present is the sum of the incident and scattered fields. It is evident from the above equation that this condition is met if (5108) which completes the solution. l j[J.
the total field is considered as the sum of the incident and reflected fields. The pattern of the scattered field is also of interest. the scattered field is of the form no' = H.' + H.' = Hrril:s "'" H o L jJ.111) This is the scatteredfield pattern.TDdERA.'''(kp»)''''· (1. dominant and For small ka.'" (ka) ""'1 (1. the current in a thin wire is 900 out of phase with the incident field.113) Again.R.112) The scatteredfield pattern for a thin wire is therefore a circle. (1. a __ • where the a.(kp) + b. f... At large distances from the cylinder we can use the asymptotic formulas for H . .o J..H. L.. The magnitude of the ratio of the scattered field to the incident field is therefore ~ .H.'1 _(21 .kp [2Jk ~.H.("") H...110) Hence. . (1. Iog "" JWp."'(kp)""· 6Dd the total field is given by H. 1.kp IE. and Eq.. it can be expressed as • H. H.V:.  l • j·[J.(kp)e""· (1. a __ • l • jb.10ti) becomes • i eft' ~ a.ad~~.. (5108). that is. whicb is to be expected..' fo represent outwardtraveling waves. the n = 0 term becomes (1. arc given by Eq.(I). When the incident field is polarized transversely to %. = H.'\' '< • J .tI.E I ' + Eo "\j.IolONIC ELEcraOKAGNETIC FIELDS find the total current as h 2rE.114) . since the wire is essentially a filament of current.
.(kp) + b.·' (5115) and the boundary condi tion is met if J.0 • j'[J. '' \' b"eJlof with b" given by Eq.+ H 0 . L: .. the n = ± 1 terms radiate more efficiently and cannot be neglected. the n = 0 term becomes dominant.H.CYLINDRICAL WAVE FUNCTIONS 235 This time our boundary condition is E.""(kp))..."1 Inl > 1 Hence.< '\' 4 4 • "00" J.'I  "JTkp I 1. .(ka) H. (5105) and (5113).H. J~ = 0 at p = a..""(ka) ' I (5117) For small ka we find jT(ka)' J. (5115). .' .4 . V1fkp /2j eI".). as we shall now show. At large distances from the cylinder..(ka) M H.. equations 1 E.. for thin wires the scatteredfield pattern is ]H." p 11  2 eos ~I (5118) The n "" 0 term of Eq. to incident field is thus The magnitude of the ratio of the scattered ~ _ I2 IH. the scattered field becomes • H. = H"C2l'(ka) An incident wave of arbitrary polarization can be treated as a superposition of Eqs. (5116) is equivalent to a ~directcd magnetic . ..'l T(ka)' ~2 k "aoO IH... From the field = ~ H.(ka) b. However....(ll'(ka) = n=O jT(ka)' Inl 1) 1 = 1 InJl(ln  jT(ka(2)... the surface current on the cylinder is (5116) For small ka.."I.. (V X a. When the incident wave is polarized transversely to z....
parallel to a conducting cylinder.. ...U'(ka) (5121) satisfies the boundary condition E.H... A current filament.. co..·) To this we must add a scattered field of the same form..l solution is • kif 4w.. For p (5119) < p' we have. Thus.(2)(kp')]e"'(...236 y TlMEHARMON1C ELECTROMAGNETIC FIELDS p Current p' filament Fto..= L: ~:I 2: .irected electric dipole.' = O. .H"Ul(kp')H. as shown in Fig. • Il.i = by the addition theorem (Eq.. E.·) p < pi > p' H.(kp') p (5122) .H. while the n = ± 1 terms are equivalent to a y.(t)(kp)[J . 518.(2)(kp')J . our final E.(kp)ei"<. = 4:12: c...' = ..(kp) + c.(ka) c. namely• • E.'''(kp')[J. (Planewave incidence is the special case p' . the incident field is ktl E. 5l"18.u'(kp)]e!"'''" + c. A more general problem is that of a current filament parallel to So can.) When the filament is an electric current I.. + E. but with the J.. ~:I l: • H.'''(kl.H.. Conductor x current filament. (5103»)... replaced by H II (2)..<J.H. '"" E.{2)(kp)ejro(·") (5120) From the preceding two equations it is evident that J.'1) 4w.. dueting cylinder...
75). tP and pi. (5121). If the line source of Fig. (5121) are equal to those of Eq.(kp) + b. (5115) instead of those of Eq.H.. ~' (reciprocity). 518 is a magnetic current filament K... (5122). in· dependent of the incident field.  • j. except that the reflection coefficients at the conducting cylinder must be those of Eq."''''J B .: I(p) _ 2: .'''(kp)[J. (~122) to the far zone. (5108) and are. ~~:K 2: . . Specializing the second of Eqs. (plane of Fig."'(kp»)e'·'''" + b. elements of finite length as long as the reflector is of infinite extent.(kp') p> p' (5123) According to the equivalence where the b.25" in front of a plane reflector is Fla. ."'''' p < pi H. in general.25). 519 are also valid for current reflector case shown dllllbed). Note also that the "reflection coefficients" of Eq. we have instead of Eq. The radiation pattern of a current filament 0. away from a cylinshown for comparison. (5115). or H. 0.CYLINDRICAL WAVE FUNCTIONS 237 Note that our answer is symmetrical in P. Therefore. The problem is dual to the electric current case.. are given by Eq. cur· rent filament 0. Figure 519 shows the radiation pattern of a current filament O. Radiation pattern for a. The patterns drical reflector of radius 3.:.U'(kp')[. the final solution will be dual to Eq.~~:K • • H. we have E.U'(kp')(J. (5119).= 2: .25~ away from a conducting cylinder of radius 3.(kp') J.75A.(ka) H '''(k ')] . [ J . 519.H.Ul(ka)· p The magnitude of this is the radiation field pattern.
.J.constant. p + co.211'" . In this cue the wedge covel'lJ Problema involving conductol"l over tol'l1phk coordinate surfaces are usually easy t. since this is sufficient to satisfy the boundary conditions. ground plane shown dashed). coordinate surfaces. the field of a nATrOW slot in conducting cylinder is the same as the field of a magnetic current on the surface of a conducting cylinder. Consider first the case or a filament of electric current at P'.o . (5119) and has only a t component of E.0.""(1<0) ~ 8. The magni tude of this is the radiation pattern of a "stitted cylinder. A wedge of vanishingly small angle is the classical conducting halfplane problem. (slit..cent to a conducting wedge defined by til .tern for a a1itted cylinder of radius 2). (&123) to the case p' "'" a.0) sin . This is shown in Fig. The total field also will have only n.sin . . 520.(~ 0) 0) 0) p < p' p> p' • (&124) I to solve."'(kP')J. 520 are also valid for slits of finite length as long as the conductor is of infinite extent. slittedcylinder pat. The pattern for a slit in an infinite ground plane is shown for comparison.. S<attering hy Wedges. The incident field is given by Eq.' adja.tern for the ease a = 2>. We construct E.'''(kp) sin ." Figure 520 shows a. . = l • I aJl. Specializing the second of Eqs. Radiation pat.(kp')H.(~.(~ I a. in a source radiating in the presence of a. 0: and til .. 1 We again restrict consideration to the twodimensional case at this time.. ~' . A FIG. The patterns of Fig.0: (wedge angle . conducting wedge is also a relatively simple problem.(kp) sin . We shall solve for the field of current filaments in the vicinity of wedges and obtain solutions for planewave illumination and aperture radiation as special cases..20:). z component of E. . 521. q..w . 510.!(P) n __ • 11. we have • \' j"e i'" II. .238 TIMEHARKONIC ELECTROMAGNETIC PlELDS principle.
.1 • This is simply a Fourier series for the current on p = p'.a) sin v(r/l ..(P'+) .H. we view the current element as a.. \ ' a. To evaluate the a.(p') Using the field equations and Eq. use the asymptotic formula for H.a) p Thus.U)(kp')J~(kp) sin v(¢' Lt ~ Lt a.~ I(p) 2>"J.(2)(kp) in the second of Eqs. This. The Fourier sine series for an impulsive current of strength I at ~ = r/l' on p = p' is J~ = ('II" I a)p" 2: • sin v(¢' . = WIJ. (5124). sin v(¢' .a.(kp')1l.(1l'(kp) sin v(¢' " • • a) sin v(¢ . ." a. . = W/. Ilre determined by the nature of the source.0:) sin v(¢ . (D17)J. gives pi ~' Current filament Condudor • x E.3. v = 2('lf a) The a. To obtain the radiation pattern of a current I near a wedge.2. 521. . To satisfy the boundary conditions E.0:) By comparison of the preceding two equations it is evident that y a.a) Flo.7rI 2(r .a) sin v(¢ .(kp') • sin v(r/l' . at p = p'.a and ¢ = 211' .a) sin v(¢ .H. we find JWIJ H.0:) 1.l1rp 2. using the Wronskian [Eq..J. A current filament adjacent to a conducting wedge. (5124).a) p < p' > p' JWIJ . The boundary condition to be satisfied at a current sheet is J.CYLINDRICAL WAVE FUNCTIONS 239 which satisfies reciprocity and insures continuity of E. with Eq. = !:. we have the surface current given by J.n impulse of current on the surface p = p'. = 0 over ¢ =.Il.a) (5126) This completes the solution. (5126). we choose mr (5125) m = 1.
) where v is given by Eq.' . This gives E..:kpi eItII The total field in the vicinity of the wedge is obtained by specializing the first of Eqs. p' _ a.__ l ~T"%' ei"'" L: • a"jJ.Cent to a.a) sin v(.') (5127) . Wail. Radiation patteroe for an electric current filament adjll.diation patterns for the special case a = 0 (the conducting half plane).p . This is obtained by letting the source recede to infinity.' where E = ~ Eoei t . (Ajur J.. the incident field becomes This is recognized as the pla. 4. Another special case of interest is that of planewave illumination. . f2T V. In this case. Figure 522 shows some ra. kp. R.fJJJJI  c.(kp) sin v(rI.240 TIMEHARMONIC ELEcrRQMAGNETIC FIELDS FIG. conducting hal! plane.a) ..newave field E. (5125).' . (5124) to Ia.rge p'. 522. ~/4.
0 at q.a) + " . . The radiation pattern of a magnetic current K near a wedge is obtained from the second of Eqs.. are determined by the nature of the source. ~ r . 4J n:' sin n 2 (5129) which is the solution for a plane wave incident on a conducting half plane.125). The result is H. substituting fnr a" from Eq.(kp')H.J. 521. we obtain E. a) 111. "" a and 4J = 2r . (5126) and for I from Eq.CYLINDRICAL WAVE FUNCTIONS 241 Finally.. The Halmost dual" problem (dual except for boundary conditions) is that of a magnetiecurrent filament K at p'.a) p < p' (5130) 2: b.4 j'J.j'J. . = 2Eo ~ jf'/'Jf'/!(kp) sin . .  0.a) (5128) where 1/ is given by Eq. v(. . (So... We construct a solution HI"" ! • L . ..2.a) co. This is the solution Cor a plane zpolarized wave incident at the angle 41' on a wedge of angle 2a.(kp) cos vW . v(."'(kp')J.. in a manner analogouB to that used to obtain Eq. b.0.. .H. The result is v ~ 0 (5132) v>O whicb completes the solution. . 1.a) sin v(.(t)(kp) COBV(41'  a) cosv(4J .a) p> p' which is similar to Eq. E.a) co. 1(P) L.(kp) sin vW 2rE.(k. (5130) by using the asymptotic expression for H.. (5127).. . (5131) The coefficients b. .. (5124) except for the sines replaced by cosines.a can now be satisfied by choosing 1/  mr 2(. . 4J' in Fig. The boundary conditions E~ . " .') cos vW .I!)(kp). (5126). For 0 we have • 0. .. .
'\' ••j'J../4. (5130) to the case p' + co. Finally.fUr J. Radiation patterM for a magnetic current filament adjacent. 611. 523. When q. A thrrodimensionnl problem having cylindrical boundaries can be reduced to a twodimensional problem by applying a Fourier transformation with respect to z (the cylinder Flo. ' _ . for planewave incidence we can specialize the first oC Eqs. Wait) . and the result is H.s 1 (or".242 TIMEHARMONIC ELEC'l'ROYAGN"ETIC FlELDS where Neumann's number t.(kp) cosvW . (5128). = TH. p' . The procedure is analogous to that used to establish Eq. R.' .a) 7f  a 1::1 • cosv(~  a) (5133) This is the field due to a plane wave polarized orthogonally to z incident at an angle tIJ' on a wedge of angle 2~ The case a = 0 gives (5134) which is the solution for a plane wave incident on a conducting half plane.0 and 2 for 1/ > O. to a conducting half plane. . (A.II. . i. Figure 523 shows some radiation patterns (or tbe special case a = O.a we have the solution (or a radiating slit in a conducting wedge. Threedimensional Radiation.
w)"" dw This is usually a diffioult operation. 524. + :..0 f _*". the threedimensional solution is obtained from the inversion 1 ~(x.(x. Fortunately. we construct a solution HVXA (5135) A = u.xis.f(w)H . we construct ~ .. in the radiation zone the inversion becomes quite simple. is a wave function independent of ¢ and representing outwardtraveling waves at large p.z)ei¥· dz y will be a solution to the twodimensional wave equation (:.i.) J. Anticipating the need for Fourier transforms.'''(p yk' w'). In the usual way. A filamCDt of curreDt aloDg where «' = k' . The only restriction placed on the current l(z) is that it be Fouriertransformable. ..y. R. tho z a. J"" f(w)H.z) is a solution to the threedimensional wave equation z l(z) . 105·)1).y.w'. = 0 x FlO. if t/. f" _" ~(x.y.t/. where t/. according to 10'1.p r' a' a' a' ( ax' + ay' + az' + k' ) '" then J..w) = . is solved.• dw Tne Fourier transform 0/ jP is w') The I(w) is determined by the nature of the source. as illustrated by Fig.y."'(p yk' which is of thegenersJ form of Eq..(x. Consider the problem of a filament of zdirected current along the z axis. evidently ~ . 524.(x.. p d¢ . 1 For example. Once the twodimensional problem for J.::t lew) I Thill applies to cylinders of arbitrary cross sectioD as well as to circular cylinders.CYLINDRICAL WAVE FUNCTIONS 243 axis). We shall now obtain this farzone inversion formula.2. t/.) . + «.y..
.en. For example.v'k' lew)  f"" I(z')clw'dz' The field is obtained from !/I according to Eqs.. . f ~ 1 wt)ei"'dw and Eq..0 and the preceding equation yields few)  !If 2· It') w').he threedimensional problem are of the same form as the equations in the twCHiimensional problem. 4. Another solution to the problem of Fig. which . (5136) becomes !/I gO .J"/~+<. 1 _ II . In particular. Compare the equations of t..his paragraph to those of the second paragraph of Sec. (5138) to the radiation zone. (~l22). if 1(:) is & short current..dw (5139) Many other identities can be established in a similar fashion..TIMEHARMONIC ELECTROWAGNETIC FIELDS where D. (5135). Hence. 524 is f where 1.< V p' + (. 524 is the "potential integral solution II oC Sec. • ~ f" _. the specialization is given by Eq. 56. l(w)H.. I(z') .p is unique in this problem. we have From the smallargu B. f" IJoUl(p Vk 8"1 .(t)(p v'k t r 2J" f" w 1 )ei"'. 210. This is J. (5136) and (5138) are equal. The transformed equations in t. ~9.. Eqs.. (5138) becomes Ile/!r Equating these two !/I's we have the identity elk' = . giving us a mathematical identity. ment formula for a. and we did so in Sec. For example. and I are tDe transforms of HI and I. It can be shown that the. 8rJ f" _ .f(w) iJp . (5135)... the "transform solution" to the problem of Fig. it is simple to specialize Eq. then 1(10) ".1.. Il and Eq.+ rp ... It is convenient to have two forma for 'f because some operations are easier to perform on one form than on the other."'I..• dw (5136) (5137) Hence.!i. element of moment Il... o'l' ")' dz' (5138) with the field given by Eqs..
__ ~. vol.k cos 8) ep.·dw~2jrl(kcooB) (5142) which holds for any function 1(10).. I Finally.U'(P y'k' w·). Erde1yi. More important. Ph". t Silver &nd Saunders. 612. Appl..(I) (:x:) AAlong as" ¢ 0 or 7'.." pp. By Eq.· (5141) Hence. 5. . (5142) valid for Hankel functions of arbitrary order. the specialization of Eq. then . if k is complex (some dissipation assumed). February. the radiation field is simply related to the transform of the source evaluated at 10 .U}(:x:) + from which it is evident that f2i 'Vn j"~ 8 . The desired generalization can be effected by considering the asymptotic expression H .(1).: jw~ 4Tr sin 8 1( ..(2)(%) + j"B .. 21. We therefore have the identity f _. rI'O' r (5143) We shall have use for this formula in the radiation problems that follow. w·) . J..' Consider a conducting cylinder of infinite length in which one or more apertures exist. . jw~A. The result is  f·.l(w)H''''(py'k·  rik..... using the method of steepest descent. (5136). pp. we shall need a formula similar to Eq.. New York.k cos 8. (397) we have ri" (5140) E. 1956. Dover Publieationa. Also.·dw+ 2j'+'1(k cos B) ~.10' is never zero on the path of integration.. The External Field Produced by a Blot in an Infinite Cir· cular Cylinder. I l(w)H.CYLINDRICAL WAVE PUNcrION8 can be written as "'~ 4n l(k COO B) where lew) is given hy Eq. The geometry is A. we have p + 00 as r + co I since P = r sin 8. "Aaymprotic Expanaiona. We are then justified in using the asymptotic formula for Hankel functions and can replace the HoC') of Eq. 153158.:::. (5142) by . 00.y k' . Equation (5142) can also be estal>lished by contour integration. (5137). =jw~sin8'" or E. 1950. 2627. (5140) must also be the corresponding specialization or Eq. Apertures in Cylinders.8 .
z)r~rjw· The inverse transformation is = 2T E. dz E.(. shown in Fig.z)  Note that these nre Fourier series on 41 and Fourier integrals on %. We seck a solution for the field external to the cylinder in terms of the tangential components of E over the apertures. . 312.w"A +...). 1 " E.uJ'. y p x . Anticipating that we shall use transforms of the fields.1D) = 2 1 r T 0 .")&· d. dz E.246 TDUiHAlWONIC ELECTROMAGNETIC FIELDS z r FlO. 2~ 2: . 525.. let us define the "cylindrical transCorms l l of the tangential components of E on the cylinder as J. vv· F JF . An aperture in a condu. According to the concepts of Sec."" V X A .'0 X F  2: "." dljl f" 1 B..~.(a.(a. . f."' !".tD) = 2 J.".·" dw (5145) .jwiF where + JW. (5146) A "" u.(n... '0'0' A JW' H .A..!. E. (5147) ..¢. the field is given by E . The field external to the cylinder can be expressed as the sum of a TE component and TM component. ." d~ f" E.(n..(n.z)c~e/W· (5144) 0 _.ctiog cylinder.~. &25. (n..
L.~w f . as " (5148) eM F. (5145). Hence.tP. in which case we ean use Eq. let us calculate E.(n. To determine the f. and F. We choose the Bessel functions as H. \' ~ . (w) v'lcl p = w 1 1I.nw w') E.(w) _ (k 1 _ jw.w')e'" dw . (w) in Eqs.(n. 2r \' " . __ (w)H . according to Eqs..~.(p. we f.. __ w 4 f" g.(w)H."'(p Vk' .  ei •• f" [.(w) and g.E.r) = vk! _ fI.. (5145).( tJ·..(P."'(p Vk' w')t!· dw "". =1.W)] This completes the solution. (5143).CYLINDRICAL WAVE FUNCTIONS 247 We now construct the wave functions A.. (~l'(a Vlcl _ 2.. The result is E. (5146).. We choose the rp and z functions such that the field will be of the same form as Eqs.(w)H.z) = 2 \' 1 rJWf " 1.(I)'(p Vk 1  Wi)] eiw' dw Since these equations specialized to have a must equal Eqs..w) w 1)H"U)(a Vk2 Wi Wi) Wi) [ g.. (511). ) which are of the form of Eq. (5148) are difficult except for the tar zone. (ll(p Vlc 1  Wi) + g.!. The inversions of Eqs.. and E.. we have (5150) .(n. (5148). f" (k' _ JWf w')!..w) (5149) + a(k.) E.. (t...(I) to represent outwardtraveling waves.
For the axial Blot we shall assume in the aperture E..) For a very narrow slot (a + 0) the transforms of Eq. = O. ~. 526.~I' I 11·a1 Flo. To illustrate the theory. (3 97) apply j hence R E. (This approximates the case of excitation by a rectangular waveguide. 2 (LtD)! . in the radiation zone Eqs. I I ~. (b) a circumferential slot..j"+!j. cosL rz and E. rib ain rr 8 n" 1t _ _ OO Lt \ ' eJrl. The only difficulty is that the number of significant terms in the summation becomes veri)' large for cylinders of large diameter. let us consider the thin rectangular slot in the two orientations shown in Fig..I slot. (a) (b) Finally.. 526... = a.w .. V J!J.. the radiation pattern of apertures in cylinders is relatively easy to calculate..ia. + jwp ..248 TWEHARMONIC ELECI'ROKAGNETIC FIELDS z Z a3 I fa'::: I v II 'f. A conducting cylinder and (a) an a.. (5144) become ! ~<z<~ <~< a a (5152) 2 2 '" ( ) _ VL cos (wL/2) n. ( k cos 6) (&151)  Thus.a . J.
the crosspolarized component E.... no. in other directions.(w) .( ) _ Va cos (na(2) E• n. by Eqs.r /2 and ~ . However.nd (5151) we can calculate the radiation field as kVaeiir " E. the field is entirely Bpolarized. . 128137...w) . (5144) become . The Radiation Field Produced by a Slot... (5151) we have the radiation field given by E.... 526b.(n..ka sin 8 Lt [r (na)')HII<t)I(ka sin B) Lt [rt 11 (5155) 11 In the principal planes 8 . we assume in the aperture ~.(tl~(1uJ sin 9) (kL)t ~ .. . may be appreciable. = 0 and E.July. Then Cram Eqs.0. (nc:r)']H... The radiation pattern in the plane B = 90° is identical to that of the stitted cylinder....) For a narrow slot (W + 0) the transforms of Eq.<"2 a a (5154) • j" cos (na/2) t/". =.0 and (~. g. in a Large Cireular Cylinder... t • .w) = O.. O. AP3..(n. 520. e JrT'sm 1 ~<'<~ "2<..cot'8 ~ nj' cos (na(2) . 0 plane is almost indistinguishable from the radiation pattern of the same slot in an infinite ground plane. pp. the pattern is given in Fig.O. (5149) we then have /.Vr.. C08 E. 1955. (5149) . vol. (t)(1uJsin 8) E __ Va".. IRE Trani. 3.Ul'(a yk' w') Finally. The radiation patterns for circumferential slots in reason~ ably large cylinders are very close to the radiation patterns for the same I L.(w)  Frnm Eqs... VL cos (wL(2) (LtD)']a yk' _ w' H. l For the circumferential slot of Fig. The "vertical" pattern in the . cos9 rJar 1~ (5153) which can be further simplified to a cosine series in'..1D r1 (na)' and B.CYLINDRICAL WAVE FUNC"I'lON& 249 .nd B. (Again this approximates excitation by a rectangular waveguide. .[COS(¥COS8)] 11 ''_H. Bailin. 80 for a 2>. W v a and E• . L.
his.. Figure 528 shows the reciprocal problems of (a) a current element and a conducting wedge and (b) an aperture in a conducting wedge.tern for a circumferential slot of length 0.. (5156) .i. Fig.on pat. we shall consider only the case of a distant current element and the radiation field of the aperture. &27. in the twodimensional problem of Sec.. Apertures in Wedges.. By this. We shall illustrate the procedure Cor the case of conducting wedges.(.. The incident field is e1l1rr'1 "" . To keep the theory simple. we mean that a solution to one of these problems is readily converted to a solution to the other by using the reciprocity theorem. reduces to "'''' "" "'" Ilei". <'I which.1.Il 4. For the zdirected elcctriccurrent clement of Fig.  _·tJi.TIMEElAlWOllo'lC ELEcraOllAGNETIC FIELDS Fla. This is simply a plane wave incident upon the wedge. 528a the field will be TM to z. The problem of diffraction by a conductor is reciprocal to the problem of radiation by apertures in the conductor.65>" in a conducting eylinder of diameter 3>" (same slot in a ground plane shown dashed). To illustrate t.. when r» r'.~. expressible in terms of an A = u.·.65X long in a cylinder 3X in diameter. The radiation pattern for the same slot in an infinite ground plane is shown dashed. 513. The'" in this threedimensional problem is subject to the same boundary eondition (w = 0) on the wedge as is E. 527 shows the radiation pattern in the plane 0 = 11/2 for a circumCerential slot O. slot in an infinite ground plane... Radiat...
To obtain the solution to Fig. 528b.CYLINDRICAL WAVE FUNCTIONS 251 510. Eq. 528a and b. (335) reduces to .&"1:1'000' ejh hr = "2(". where (5157) "4'0 V = I l . Hence the solution must be of the same (orm as Eq.0) . (335» to the region bounded by the conducting wedge.=""a") m.2<>1.. (518) and (5157) we calculate H: = p. .in v(~ . The reciprocal problems of (a) a current clement and a conducting wedge and (b) an nperture in a conducting wedge. (518).bH: ds .a) sin v(4) . Because of the bound· ary conditions on E at the conductor.JJ E. we apply reciprocity (Eq.'7f'/!o a) L • vjoJo(kp' sin 9) cos v(4)' .(. From Eqs. 528a.0) ?< a 4 . This completes the solution to Fig. that f .." !lE. respectively. is.· (5159) where the superscripts a and b refer to the fields of Figs.a) z n z y y x x (a) (6) FlO. (5128).Ckp' sin 0) sin vW . (5158) In terms of y" the field is given by Eqs. "\' j'J. 528.
a) J..p . (5156) with 1 replaced by K.z) 0 P Note that f. From Eqs. 52& with Il replaced by Kl. '' (5160) (5161) . 528b.'lI. (335)}.. This reduces to . >/ ~ T ~." dp' __ 0 p r ~T/ooE. (519) and (5162) we 11." p. McGrawHili Book Company. (&133). (5158). 5284 with Il replaced by Kl.. "Fourier Tranaforma.(kp' Bin B) coo . and in the radiation E.p' = a.252 TlMEHA. 6.(up) dp 1 E. &v'dz 10" J . .(10. e Jl a .. ~ KIH. Again the threedimensional problem is essentially the lS&IDe as t.) a \' '' vj'J.he twodimensional problem of Sec. I.E. we can reduce Eq. 1951. (5159) to IlE. Inc. respectively. 8m sa .ic field is found from'" according to Eqs.2 ( r". _· hr .BMONlC ELECTROMAGNETIC FlELOS Specializing this to the surface .l In a similar manner. component of the radiation field caD be obtained by applying reciprocity to Fig.. and of Fig. N.". the E. (519).(tlJ. Sneddon.9. 510. the 8 component of E in the radiation zone is given by E.a) (5162) where .E. SlD ZODB 9 Hence.(~' Lt .(kp' ain B) sin . The incident field is then specified by Eq.. ex) m = . 2. '.a) COB . k sin 0) where 1. fJ \ ' vj" sin v(.. New York.. __ .) f" ...(k cos 8. ff (E. The electromagnet. This zdirected magneticeurrent element gives rise to a field TE to z. To relate this solution to the field from an aperture in a conducting wedge.o.·) d. .eft.' (5164) where superscripts a and b refer to the fields of Fig. f" dz' f..'0 = K l .u) is of the form of a Fourier transform on z and a FourierBessel (or Hankel) transform on p. 1/10 is given by Eq.. we again apply reciprocity [Eq. . expressible according to F = u. ..(~ .(~  a) Finally.(p. t.11. ex \ ' •• j'J.hat is.'H: . The solution is then of the form of Eq.'" (5163) v = 2(.
. where v =~. and from Eq. (5160) we see that E.(p.0) cas L . Q'. L: ¥t.(w. J.CYLINDBJCAL WAVE FUNctIONS 253 calculate H: . h. . . k sin 8)] dz fo . (5161) and (5166)] are then found to be I. Sin 8 "'~. dz / .(kosin 8) (5168) Plots of FlO. The result is = ~H.~O 8.(k cos 8..a) f. 1fkl sin /1 cos /1 ( ) wp.(p. .. as shown in Fig..(~ • I() T 8m · 8 caslk(L/2) cas 8) .j·cos. k sin 8) (&165) + j sin 8 h. 1f a L:.z) We now have a complete solution for the radiation field from apertures in conducting wedges.a)(cos 8 g.(k cos 8. . e it. The I. and h functions [EQs.a. 1 (kL cos 8)1 . 529. • f. 629. cos v(t/>' . 1f a l H. (5165) we have E• "" L: .4r(1f a) '\' L.1fk ( sin I 8) 1f1J JWIJ. we evaluate Eq. (5167) is the only tangential component of E.J·J~(kP' Stu 8) ..a) J .z) (5166) where Q'. 1.J~(up) dp E.. _ "fB.u) ~ /_"" . ~ VI(p . .(kP' sin 8) cos v(t/>' . . e.0 7'1 h• ~ 2.a.a) cos vet/> . %. We shall assume that in the slot E.j·J.' = .j· cos vet/> . As an example...(up) dp E.(w..u) = J..VL cas (wL/2) (Lw)! J ( ) • ua From Eq. . (5164) and use the radiationzone relationship E..a) Finally. let us calculate the radiation from a narrow axial slot of length L..a) cos vet/> .0. A narrow axial slot in a conductine halr plane. . E.
(518).Ila of two lICalar wave fuoctioDa.. For twodimensional fields (no :e variation) ahow that an arbitrary field in a llOurcefrce homogeneous region can be eJl:prCMC'd in ten. 6S. phue constant ia given by The . PROBLEMS 61. ihis corresponds to choosing * j.96>.90 0 arc shown in Fig.p.254 TIMEHARMONIC ELEctROMAGNETIC Fl. (379) whcro A .ELDS Flo. Repeat for the case f.Oog p)e'k ia a 1I01ution to the sealar llelmholu equation. 66.ll in a conducting half plaoe (the slot in fLn infinite ground plane is shown dashed). (512) is a IlOlution to the ecalar Bclmholtl equation.B. Show that". with the infinite groundplane pattern shown dashed for comparison.. Radiation patterns for axial slot. &.)eslA'•• where TM modes are detenninoo by Eqs.p)h(n. the radiation pattern in the plane 8 . .. (380). . are sbown.erized by wave functions of the form 'I. and lJ1. 0. What ia ita inside diameter if it.4 are characl.c.constant plane. Note that.(. 530 for the case ex .. 62.u"py. A circular waveguide has & dominant mode cutoff frequency of 9000 megacycles.16A and a = 0. f). What pbysicalllystem IIUpporla this wave? Repcat for the TE case. (&18) and TE IDOdea by Eqs. Sketch the t and :Je Jines in a ~ .30. All the waveguides whose CtOSll eeetiona are shown in Fig. 64.0 (half plane).P(~~) ~ _ _j~(F:) illlltead of Eqs. Show that Eq. Determine the TM field generated by this" according to Eqa. _ 4. is airfilled! Determine the cutoll frequencies for the next ten loweJ~rder modeJ. The cases a . (&19). according to Eqs.
&5).edge waveguide of Fig. 27.here n _ 0. 54c and d are the special cases .) COIn.(k". is a root of Show that for TE modes B.. or cos n • .here the baffle iI at • .. 64b) are ehal'$Cterizcd by the lAme B.(k".:. or COlI n• . . &lid for TE modes n  0.>:'. waveguide of Fig.·· . The dominant mode is !.(k.(i"p) h(n.:. cua.(l".o)J. .. of J..O.. 1.) fWlction. M•.. .(i. 5411. 54c) iI . 6k supportA TlI..) _ &in n.··· ' and kpa is a tero of J~(k.(i"p) ..0.. and T.. 1'.J:(k. G8.(i". .(lllS).he lowen TE mode with 5·7.... Show that the modea of the coaxial waveguide with a bame (Fig.a).. :w Show that it aupporta TE modes specified by I/TZ _ J... u the coaxial guide (prob. • h"" • 2• nO..and k. Show that the .2. ..~. .T 2..tn' ..2.p) h(n.. Show that for TM modes B...' + GY] . and by .) . • )i. . 1.2.l modes specified by . &lid k.2.J .a)N.IJ is a KJ'O ~na n. but for Tl\f modes n ..(i".•..) _ N . Show that the cutoff wavelengtb for tbe dominant mode of the circular waveguide witb bame (Fig. a denote the inner ndius and b the outer radius of the eoaxiaJ. .here n . respectively. e*ia.tn' [(>:. }i. '" "'1 V.a)N... ahow tbat tbe attenuation eonIl&nta due to conductor loaaes in a circular waveguide arc given by ~ .. is a root of G·8.. .p) ..7i~:7:ff..(k.. .) .(i. _ J .iii for all TM modes. ..CTLTh"DRICAL WAVE FUNCTIONS 255 Let.. • a .) ain "'. and i.N~(kptJ)J. Using the perturbational method of Sec. The guides of Fip... 1. 1.'" VI" V. . H.sin n..
.I]. i.. (619) if n ill a root of = N~(ko) Tp B . that for luge radii Z+.ELDS (or all TE modes. c _ 2rtp • Why.nd p . (549»).nI _ Z_.. IJ'. Cooaidenhe wedge guide of Fig.0 • >0 where y _ 1.p ft. { ~k"(d. is given by Eq.he dominant mode (Eq.. For t.(}.should we expect circuit conecpta to apply for this mode? 61'. (b36).I"+) »~ J'J k [(2• )'(k.781. Show that Eq.. Conaider the radial parallelplate w&veguide of Fig..)!' 1.(k.256 TU.. ShOw that the wave function specifies circulating modes TM to I: according to Eqa. 550. (518) if n is a root.~·~ .. + IN.static" parametera L _..H~ . LJ dl jwCV dp  where Land C are the ".nt. (538) and (539).0) the attenua. ". and that for small radii Z+.Z_..A J~(ko) J:(kb) N .»)' k r.! "'2 + n .iDg cylinders. ..RARMONlC ELEC'I'ROMAGN'ETlC Fl. lUI V(P) _ oB. (545)J. 618. of Show that tbe above wave function specifics modes TE to z according to Eqa.![l(~)']..tion decreases without limit as f .b. 610...e 2. show that the radial phaae constant of E. one can define a voltage and current. (637) is also valid for this phase constant.. For the TM radial wave specified by Eq.. while tbe radial phase constant of H.(kb) 611....TN _ 11 Show . as 1(P) .CJ a. (533). 5Sb. Consider the TM radial wave impedances of Eqs. IJ. Note that for the "circular electric" modes (n . one can define a voltage and eurre.~ . . For the transmissionlino mode IEq8.. 612.. Consider the twodimensional "cireulatinc waveguide" formed of concentric conduct. Show that V &nd I sati8fy the transmiMionline equatiool!l  dV dp J(jI .
510a). Following the perturbational metbod used to derive Eq. The circular cavity of Fig.. Consider the dominant.3 centimetera.lo)N.(k.. 613) with c~ ••• 616.41. (558).t t. Which model can propagate unattcnuated in the a1ab? Rcpeat the problem for the coatcd~onduetorguide of Fig.0 A . Let II .(k. Show tbat for small a tbe cbaracteriatic equation beeomes (. Detennine the first ten resonant. frequeDciei and tbe Q of the dominant mode if the . . 61 are 117.(PXl) Note tha. . approximations (~)I. S9b.(k..10)J1(k_lO) .1 + .la) .(k. and . Show tbat for small a and large ..a) 621.d .1) mode of the dielectricrod waveguide of Fig.alls are copper.aN'. and 0 . 620.bere ia no cutoff frequency.f)(11 + If) 2.J'I(k_IO)J1(k_1 0) B ..tatic L _ lAid + "1(0 2r.0/2.k. 618.. . (574)] for the n . 618.k. using the . &90.tb)l .. Show that the re8Qnant frequenciel of the twodimensional cylindrical cavity (no I: variation..d) PI  619. For the partially IDled circular waveguide (Fig. (n .(k. sbow tbat the Q due to conductor lOIlllel for tbe various modes in the circular cavity of Fig.l!) where + BJI(k_.).51Ob.k.K.CYLINDRICAL WAVE FUNCTIONS 257 Show that V and 1 aatiafy tbe tranamissionliDe equation (prob.. Consider tbe dominant mode of the partially filled radial waveguide of Fig.l!)IlAN.IJ..0) arc equal to the cutoff frequencies of tbe circular waveguide. Conlider tbe dieleetricelab radial guide of Fig.ib) + BJI(k. conductor over p ..tO)J 1 (k. show tbat tbe characteristic equation (Eq.1 modell reduces to IANI(A:_.k_IJ. the phase constant is Compare tbill to the uniform transmiasionline formula IEq... 59c witb t . 67 baa dimensions II .
Consider the circular cavity with a dielectric slab. and and ~ .. Plirti~ly filled cavities.:::.405 according to where I..(ktcJ] "Jo(ke) 1'/0 N. &310. Show that the dominant resonant frequency is the smallest root of ~ J. given in Prob. show that resonant frequency"" is related to theemptyeavity resonll.~. Consider the circular cavity with concentric dielectric rod.#1) rod I 2 1 ~l + t l og:.h where. 623.nce %" :1:01  2.0.va .  where lola is the cmptycavity reaooant frequency.1. Using the results of Prob. 624.L 781..b • k.1). show that for a small (4 « A.(koc) For small c/o.. as shown in Fig. 523.).t _ kt _ (X.(kl$:) . Take I I .1" . I ~a I 1 r b L l ~ .IY lola • V~l + ~ lr. 531b.  . (k1o)l h .258 TIMEHARMONIC ELECTROMAGNETIC FIELDS 622. ~tank. Show that the characteristic equation for tbe resonant frequency of the dominant mode is where Show that when both d and b are small Iolr .Jo(kotl)N.(kc) "" ~ [No(ktIJ)J.(koa)Jo(koe) Jo{koa)N.J(~l::::JIZ/..9'1 and a .l. as shown in Fig. . nonmagnetic &01 . .'J . and calculate the distance from the lUis for which the field is 10 per cent of its value at the surface of the rod. 631. 521. d a • (b) (a) FlO..... The field external to a dielectricrod waveguide varies as K1{vp)../JI/J •./to.)~bd I l)bjd f.
V X A where Show that t.ed magnet.hat.0 3.0 and '1 _ O. /(2. Show that the field is given by H . in the limit a .9 1. s33b in the limit .14 0.be field is idcntical to that.2 0.. Consider the quadrupole 80uree of Fig.7 3.ic currents +K at 11 .0 and values of ware 11  . s33c in the limit '1 .83 126.. Consider the circular cavity "dth a conductiDg wedge. shows a linear den.5 3.V X where u. .._ 0 is given by E _ v X u. Show that the field of the magneticdipole source of Fig. the resoDant frequency of the dominant mode is given by whcre 1D is the firet root of J.. Show that the field is given by H . Somc representative • 0. 0. on a cylinder of radius p _ a. Wedge jn A circular cavity.'/2 and K at 11 . Figure 5330." a2ft Figure &33d represents a 50urce of 2n current filaments.56 0. _ O. &27.4.8 3. produced by t.28 0.(w) . 126.. as showD in Fig.6 3.. 532./2 in tbo limit.'" where 628.70 3.he magnetic dipole formed of adirect. Show t.e. .CYLINDRICAL WAVE FUNClIONS 259 FlO. for daman.ity of zdirectcd current elementa alODg the r axil. 532 Show that.. equal in amplitude but alternating in sigo.
.(ll(ka)J.directed current (Fig. +1 A (e) . the field is given by H ...V X u. Show that the radiation field from a ribbon of uniform z.:.(t)(kp)e i " . 680.. wh"" 1 0 A.I' + I' .I 1)1 r2j(n (kG)" II •'''(k) 2" p sinn'.260 TIMEHARMONIC ELECl'nOMAGNETIC FIELDS y y J.'''' <.I' .G f.'" where .l • CG) .ed magnetic currents is dual to this problem.I I +I X '+1 'I •'+1 I (d) X FIo. Let the cylinder of current in Fig.<fimensional sources.I. p A. • _ K (b) T + X Y Y . + I.. • AJ.+1 I . .. X • . 425) i. .H. .33.j . S. but still independent of z. 631.1. &15 be an arbitrary function of 41.(ka)H. Some two.V X u. Show that the field is given by H .y.e given by E. with •• 2j +2j I 11.>. • •• I I I " .(kp). ..2. .." J. +1 . .. /.A .../'" d41 A cylinder of zdircct.
show thllt the scattered ficld is given by Eq. .J(J) H. Note that thiB problem ia completely dual to Prob.(k4P)ei'" witb c. and make the Mn1l'Opiion that tangential E in the a10t is ~B .. For a TM incident plane wave fEq. Derive the following wave tra08formations: .!!in 41)  l • I. 534 in the nonmagnetic csse reducea to E. li83.(ka)/laJ.ltJ(ka) Ind that the field internal to the cylinder ill given hy • B. (ka)l(fo.' . . J.) H .. 1""Iae ' E • 2rjkp 0 (ka/2) COR • And E.(k4C) fJ. Show that the radiation field is sin cos .(ka) + a. Consider the a10t antenna of Fig.TiH •..VE FUNCTIONS 261 632. _ ]W~P :8.O.Ul'(ka)/kaH.CYLINDRICAL WA.I4. (5129). liSli. 421. .~/I" Repeat for the opposite polarization.(k. The IOlution is given by Eq./IJ.polariud plane wave of m~nitude B.. . 'L\ • nj.CII(ka) ~J:(k. _1)H. 517 be dieloctric witb parameters '4. that ill. . .(ka) ] 14J~(kllG)/lk4aJ.(ka) [ a.U'l(k4)) Note that tbis solution reduces to the solution for the conducting cylinder when foI .. 00. Show that the solution of Prob.. using the result of Prob. Repeat Prob. (5105»). Show that./1(kp) sin n:' . (5113)...0 surface and a . when the incident field is given by Eq.nt. Let the cylinder of Fig. the current on the half plane is J.. 535.) where Il6~7.J IoocP) cos 2n• • 0 lin cPain 41) ..2 L ••0 • JIoo'tlcP)ain (tzn + 1) • li34. .J'. 534.. 534 for the opposite polarization. 4 jrE.JI.CIl(k. an angle Q'. Noto that the solution reduces to the solution for a conducting cylinder as Jl4 .·_ . 1_ (""2 COR cP. 1 l j~.ra)/Jc4aJ.J. 80 the solution is obtllinable by using the interchange of symbols of Table 32.J. 1536.. incident at. a cooaia.1J. Con&idcr a conducting half plAne eovering the. (5106) "itb J.(k4a) .B. .
'~/" H... (&134) Show that the currcDt on the half plano is J.. E. ill Neumann'.~ \' • J .ral characteristic of knife odgea. The solution is given by Eq.. ..262 TIMEllMUIONIC ELECTROMAGNETIC FIELDS Show that near the edge of the half plane 1. . Figuro 5344 show. Show that t. 534.0. 1/1 2" Il1D "2 Bence.~) . &0 d p =aio..2H.O. number.. T' Show that.IIJ. . and J.lIy pointing electric dipole a dilltance b from the axill of II. while E..bl ..t.!II(a) . Z . This is a geoc.~. • .8 shown in Fig.. IUD . _ _ 8.0. E.' ... Conducting cylinder with to) axial magnetic dipole on its surface. z KI z Z ~ ~ f!'.0 .:O "llo 2 1/1'. vanishes lUI VfP.) CO!! . 537 with the incident plano wave polarized tral1llverse to %. ~ .ka sin 8 and fJ .. 8. 1/1 jrkpC08"2Sln"2 where . . 638. '\' • B • . . conducting cylinder of radius G. (b) uial eleetric dipole a distance b from t. Consider the uill. Show that the radiation field is given by where r.he AXis.' . bceomCll infinite as 1/v'kP...N . >II n X X X r.(a)N. li4. (b) . _/(. Nole that J.._2H.II (I:.) sin' where a .he radiation field is given by E. ColUlider the half plane of Prob. a conducting cylinder with an aria11y pointing magnetic dipole Kl on ita surface at.~) . 534b.(a)1. (e) (a) Flo.0.. DeAl" the knife edge J.. This is abo a gencc!l1 charlot teristic of knife edges.:0' 2£ J2jkp ... ~ • 1 .. and (e) radial electric dipolc on its surface.. is finite at p . bceoroCll infinite as l/.' i$ the angle of incidence and 4> the angle to the field point. 639.k.1b ain 8.
.j .itO/IJ. I• n.. the pattern is a cardioid with a.O. Figure 535a shows a conducting half plane with a magnetic dipole parallel to the edge.. 536. nj"ItJ"u(ka sin 8) sin2 eil' sin 8 2: ~. M shown in Fig. 534<:. Electric current element on the edge of a conducting wedge..lr sin 9 l• •• 0 '. 535a points in tho x direction instead of the % direction. Consider the zdirectcd electric dipole on the edge of a. M shown in Fig. and on the side 4> . components. eonducting wedge. A eonducting half plane with a magnetie dipole on the side 4> .0 a distance a from the edge.t"II(k4 sin 9) cos n2</J z x p II x FlO. Fro.. Consider the radially pointing electric dipole on a conducting cylinder of radius a. Show that the radiation field is then given by K/. The field in other directions hM both 9 and q. null in the 4> .  j:t . Suppose that the magnetic dipole of Fig. Show that in the z _ 0 plane (in which Illics) the radiation field is given by E.e/lr E. 643.0 • .4r . a distAnce a from it.11(ka sin 8) eos n: is Neumann's number.CYLINDRICAL WAVE FUNCTIONS 263 641. 6U. Show that the radiation field is Eo where~. .0 direction. 536.f(P) ~ H ~(I)'(ka) • I ~ • nj" sin nq. . Show that in the plane of the element the radiation field is given by For a half plane. j~~1 e. 642. 9 ar 81n E. 535.I'.
H sin Od9 SID 0 d8 + 1) (64) because the properties of the H functions depend upon whether or not n . we obtain ~ 2. The spherical coordinate system is the simplest one for which a coordinate surface (r = constant) is of finite extent.. l 0 (61) Again . ( r dr R dr .. 61.nd multiplying by r* sin' 9.y construct electromagnetic fields. :r (r ~) + r* ~n 9 :9 (sin 9 ~) + r* s. and we let (63) where m is a constant. In spherical coordinates the Helmholtz equation is f..CHAPTER 6 SPHERICAL WAVE FUNCTIONS 61.e use the method of separation of variables and let (62) Substituting this into Eq.. a. The usual definition of spherical coordinates is shown in Fig.ted out.dR) + Hsin9d9 ( sm9 dH) sin'9 + k'r' "'" 0 1 d . II d9 d9 SID Tbe ¢ dependence is now separa. (61). 8 dH) + ell d¢* + k'r' sm '8 = 0 ! d~ .. Substitution of this into the preceding equation and division by sin l 9 yields 1 d ( Rdr T dr . (. An apparently strange choice of separation constant n is made according to 1 d (.dR) + sin 9!!.n* 9 . . Once again we must determine solutions to the scalar Helmholtz equation. m' d9 dH) m' sin' 8 = n(n This scparatal the rand 0 dependence. dividing by 'It.. The Wave Functions.. from which we ma.~ + kty.
_ 1'2_ n(n+l)+k 2r 2 . We shall denote solutioDS in general by L... aH) m' sin 8 d8 am edi + [n(n + 1) . (cos 8) (68) where P . The R equation is closely related to Bessel's equation. With this choice the preceding equation becomes 1 . giving rise to solutions h(m41).."(C08 8).. ~L The spherical • x I I I I coordinate system. The ~ equation is the familiar harmonic equation.. 0 .O R dr dr a( aR) (65) which completes the separation procedure..(C09 8) are the associated Legendre functions oC the second kind. Collecting the above results.r'd1' dr (66) a"l> + m"l> a~' ~ 0 Note that there is now no interrelationship between separation constants.."(cos 8) are the associated Legendre functions of the first kind and Q.f:in2 e] H". These are considered in some detail in Appendix E... (C08 0) "...0 1 a(.(k1')..SPHERICAL WAVE FUNCTIONS 265 z r Flo. we have the trio of separated equations a( aR) + (kr)' . denoted b. which are related to ordinary Bessel functions by (&7) (see AppendiX" D). Its solutions are called spherical Bessel functions. and ita solutions are called a3sociaied Legerulre functiona. Commonly used solutions are L. The 8 equation is related to Legendre's equation." P . (C08 8). We caD DOW form ..n(n + I)IR . y is an integer.. Q..
the elementary wave functions are (&12) r 0 included "'_.. (kr) represent standing waves. h.(cos 9).t 6 ..(II(kr) represents au inwardtraveling wave.•b.. 9)h(m~) (&9) These are the elementary wave functions (or the spherical coordinate Again we can . with m an integer.construct more general solutions to the Helmholtz equation by forming linear combinations of the elementary wave functions. for k real..(l)(kr) represents an outwnrdtraveling wave.b.. sin kr Jo(k r ) . .(Jrr).i..(cos 6) with n an integer. Thus. if oJ. (610) where the C. :IZ . it turns out that the spherical Bessel functions are simpler in form than thc cylindrical Bessel functions. but such forms are not needed (or OUf purposes. then n must also be an integer and L...• .is to be finite in the range 0 to 'I' on 6. The only spherical Bessel functions finite at r = 0 are the i. .(cos system... Oil 4J is desired.kr no(kr) . which can be readily obtained from the recurrence formula.p). the zeroorder functions are .(kr)L. The ho.(C08 6) must be P.. 41. If a singlcvalued y.. i.(kr)L.(kr)P.tions over m and n are also solu· tions to the Helmholtz equation.266 TIMEJiARMONIC ELECTROMAGNETIC FIELDS product solutions to the Helmholtz equation as f . For examplc. in the range 0 to 2. _ cos kr kr (&11) C06 The higherorder functions are polynomials in l/kr times sin (kr) and (kr). or of ~ and r~.rities a.(1cr) and n .rmonic functions h(mlj) have already been considered in Sec. IncidentaUy. are constants. Thus... and h. Thus. The most general form that we shall have occasion to use is a summation over possible values of m and n ~ II C_. A study of solutions to the associated Legendre equation shows that aU solutions have singula.. we must choose h(m.) to be a linear combination of sin (mQ) and cos (m. to represent a finite field inside a sphere..(cos 9). The spherical Bessel functions behave qualitatively in the same manner 88 do the corresponding cylindrical Bessel functions.0 or 6 = 'I' except the P.. Integro.(cos 9)h(m~) • • ..
. and somewhat simpler. (378). 312.y. (379). For the magnetic vector potential we let A = u. (378). An alternative. we must choose outwardtraveling waves (proper behavior at infinity). '!'". 61.'" = ur'!' cos 8 U6'" sin 8 (615) which generates a field TE to z. The 0 and q...sin fJ (6~14) which generates a field TM to z. To represent a finite field outside of a sphere. The A.!" we can use the method of Sec. representation of an arbitrary electromagnetic field is also possible in spherical coordinates. To represent electromagnetic fields in terms of the wave functions . Note that the above two equations are satisfied identically jf we choose (616) Substituting this into the rcomponcntequation obtained from the expansion of Eq.. and F r must satisfy. 1 2 sin 00 a¢' . one TM to r and the other TE to r. (378)]. (617) . and F = urF'" with the field being given by Eq.s letting y. arc the desired elementary wave functions. because 'V 1 A.nd F.!. + k'A. Hence. respectively.. a.0 ~ . are not solutions to the scalar Helmholtz equation. For this we choose A = urA. ¢ ('V 2A). = Ury.uey. or' + r_1_0 00 ( SID (J CA') + rl sin' 8 c'A. Suppose we attempt to construct the field as a superposition of two parts. r + !Xl included (613) with m and n integers. Explicit expressions for the field components in terms of. we return to the general equations for vector potentials [Eqs. cos fJ . The dual choice is F = u. Explicit expressions for the field components are given in Prob.SPHERICAL WAVE FUNCTIONS 267 with m and n integers.. = h"U)(kr)P.A. To determine the equations that A. be a rectangular component of A or F."'(cos 8)ei"'. 61. we have c'A. are given in Prob. components of the resulting equation are. where <Il" is an arbitrary scalar. This involve. hence the logical choice is A = u. and expand the first of Eqs. An arbitrary electromagnetic field in terms of spherical wave functions can be constructed as a superposition of its 'I'M and TE parts. The z component is most simply related to spherical components.
N.. (379). Bchelkunoff.=0 r (620) is the equation for Fr.. it is convenient to introduce another type of spherical Bessel function.V g 1 X V X fOP (622) H = V X r. defined as ' ". Inc. ar (619) we find that (v.. To be explicit." pp. 5152. _ 0 (624) which can be obtained by substituting for b.+M(kr) (623) These arc the spherical Bessel functions used by Schelkunoff.. ...p +!V X V X 'of' These we shall find sufficiently general to express any ae field in a sourcefree homogeneous region of space. A dual develop ment applies to the electric vector potential..krb.J. _ 0 r (618) so Ar/r is a solution to the scalar Helmholtz equation.'s Brc solutions to the Helmholtz equation. by choosing Thus. Van Nostrand Company.'!: + k' [ dr2 _ n(n r2 + 1)] h. (66). "'. in the first of Eqs. "Electromagnetic Waves. which is explicitly E ~ V X "" 1 +. General forms for the AT and FT in terms of the spherical lB.(kr) A r.(kr) . +k')F. l Their qualitative behavior is the same as the corresponding cylindrical Bessel function. and choose zif>'> ~ aF.. The 1//a of Eqs. Princeton.2 B. (378). A. (622) arc always multiplied by T.268 TruEHARMONIC ELECTROMAGNETIC FIELDS It readily caD be shown that this equation is (V' + k') A. 1943. D. electromagnetic fields can be constructed (621) where r = Urr is the radius vector from the origin and the y. The field is found from the above vector potentials by Eq. The differential equation that they satisfy is . if we take F = urFn substitute into the second of Eqs. because of this. and. in terms of lJ.
= 0. (379). when only A. = E. + k )F. we obtain E. p. + . as given above. ! aA. a2P. • rsin 8 aq.(kr)L. The J.$.. 62. ::u 0.•ll.) are the same as those used in Eqs. ir sin 8 aT aq... are constants. Note X that E.F'" and expanding Eqa. = 0 and F.SPHERICAL WAVE FUNCTIONS 269 Bessel functions of Eq.. exists. . _ J. The field components are tben found from Eq... The considerations involved in choosing specific forms for 13.+. (626) H. = 1 1i ar2 + kt r sin (a.'" is chosen because the field must be finite at 8 = 0 and 'lI'.aF~ + l 0 a¢ 1 t a2 A. 62.(cos 9) J 1cos m~) smm. The spherical cavity. Letting A = u~A. and h(mq. and F.. aA.. we have a field TM to r. For future reference. = 0 at r "'" a if y J . that is.. formed of a conducting sphere of radius a enclosing a homogeneous dielectric t. = E • =!. We shall find it possible to satisfy the boundary conditions (tangential components of E vanish at r = a) using single wave functions. f)r ar ao E. _! of..bove equations represent a field TE to r. Similarly. = '1 ar 2 £ (a. + r ao f)r sin 0 ar aep H.. The Spherical Cavity. When F.."'(cos 0).(kr).(kr)P. ) A. (626) with A. (627) z wbere m and n are integers..(kG) .(cos 9)h(m~) m•• (625) where the e. (612) and (613). the P. L. is chosen because the field must be finite at r "'" 0.!..0 (628) FlO. For modes TE to T we choose F. Figure 62 shows the spherical cavity. (623) are I C_. the . when A. 1 ~rara8 H __ a'F. + r ao = _1_ a' A. let us tabulate explicit formulas for finding the field components in terms of A. and F = u.
II ?Lot.&54 8. We now satisfy the boundary conditions by choosing k = u"" a (629) which is the condition for resonance. The denumerably infinite set of zeros of j~(1L') are ordered as U~PI and the lowerorder ones are given in Table 62.924 ZO. . the TE to r mode functions "'(cos 9) are (Fr ) .183 11. 2.221 5.524 21. ..714 6..281 • 8.128 9. .323 15.122 4 • 9. n = 1.513 14. 2. TAllLE 62. The denumecably infinite set of zeros of J.371 Hence 1m must be a zero of the spherical Bessel fUDction.753 18. p • Table 61 givC8 the lowerorder zeros. 1. = E.103 20. . .275 13.967 16.921 17.698 16.744 6.973 8.270 T(MEHARMONlC ELECTROMAGNETJC FIELDS TABLE 61.006 17.713 13.904 14.796 21.211 12. . Hence.915 23.062 9.190 20.653 7 11.182 3 • 6 4 4. Jcosm¢) l sin m¢ (630) where m = 0.314 4 5 7. .64.648 20.670 18.705 15.905 20.009 21. (626) with A.725 10. The boundary conditions E. 946 3.983 22. .791 16.968 13.1•• 0' J. (U) 6 ~n 1 2 1 2 3 6.425 21. The field is given by Eqs. 3.095 12.674 19.301 21.. (u) are ordered as U.515 18.525 10.272 18. ORDERED ZEROS u~" Of' J~(u') >.. 1 2 1 2 3 4.317 12. ORDERED ZEnas 1.870 7.939 19.064 15. we generate a field TM to T..321 22.355 19.763 9.040 18.493 7.626 7 8 10.207 17.189 14.140 11. = O.431 18..615 .988 10.641 20.. 2. .657 15. .443 10. and p = 1.894 3 • 7 4 5 2.356 12. .722 12..117 9.486 15.380 16.579 17.923 8 12. . If an A.689 21. = 0 at T = a arc then satisfied if J~(ka) ~ 0 (631) so ka must be a zero of the derivative of the apherical Bessel function for TM modes. is chosen of the form of Eq.335 14. (627). 3.417 13.4 18." = ') J( a P.391 15.295 22.
.t = where superscripts "even" and "odd" have been a.2. (634) that the resonant frequencies are proportional to tbe u . Hence.1) even. the modes being ordered (0.493~)sin8cosq. from Tables 61 and 62 it is evident that the modes in order of ascending resonant frequencies are TM""I. 2. 3.e:n by Eqs..)iTl = JI(4.. We see by Eqs. respectively. (1.2.1." and u~p.. fJJ (634) ')TM (J. I m~) ~ The where m = 0.1... Jl (4.2. . and (2.8PHERICAL WAVE FUNCTIONS 271 Our boundary conditions are now satisfied by choosing ."'(cos8) 1sinmq.11 and so on. = The TM to r mode functions are cos (633) ." 21fa _ v~ fjJ. we have TE (f. The resonant frequencies of the TE and TM modes are found from Eqs.... "'. J . The next higher TE resonance has a fivefold degeneracy.1 = J t ( 4.1) odd. =2 u..(cos 8) exists only for m S.1).. since fr is independent of m. .. The lowestorder mode~ ar~ then. and p = 1.... 2.. Note that there are numerous degeneracies (same resonant frequencies) among the modes.1) even. . 3. In this case there are two characteristic mode patterns. field is giv. ... a r) P . (626) with F r = O. TM.. Letting k = Ztrfr V. . TE. respectively.dded to denote the choice cos mq. 2..1.l. the three lowestorder TE modes are defined by (Fr ) 0. . For each integer increase in n.. . .2. ( u. . TM. r . since P. therefore (A r ) . the degeneracy increases by two.." a (632) which is the condition for resonance. The situation for TM modes is analogous..493 ~) cos 8 (F. and sin m¢. in"" 1. These three modes have the same mode patterns except that they are rotated 900 in space from each other.l.) ..1) odd.. For example.. k = u ..r 'll'a v . 1. TE". (629) and (632). (1.. . ..493~) sin 8 sin ¢ (FrH~t. (2.l. n. ... 1.2..I. .2. = 'U ..
these three modes havo the same mode pattern. 9 ( . 1.2 [J.. consider the TM u . Mode pattern for the dominant. The magneLic field is given by H. New York.  Illlff> 11l1'd.. 63.01 (1)36) Hence. J. 28.14) The power dissipated in the conducting walls is approximately tl'. For this calcula." p.744~) sin IJ 1 Following the procedure of Sec.1. = ill 8. (ore the three TM".'(2. =JJ Jo2·d~ for dB !o"drH.. . 146.: (1.744~)dr Thus. w _ 8.I modes.(ka)J. The Q of the lowestorder modes is also of interest.>) This last integral is evaluated as l (. Except for a rotation in space.' (2.'(kT) dr .2rtsinIJ The 8 and q. modes of the spherical cavity.14) = 1. the Q of the cavity is _ Q _ ~W _ ~ l! ill kIllJ .UONETIC FIELDS e • . is numerically equal to 1. for ka energy is II: 2...  Flo.I.'(ka) .(ka)] which. the stored (1)3.14/k. Jahnke and F.. a J. 1945 (reprint).  ~ 1 (2." 1.744. 63.. giving w  S. 1 mode. we calculate the stored energy as w = 2W. integrations are easily performed. " III IHl'd.744) (1)37) I E.'(2.. which is sketched in Fig. Dover Publications.744) ~"(1. Emde.'J. "Tablea of Functiona.272 TIMEllAR~tONlC ELECTRm.tion.
dT (638) The righthand side vanishes if 1ft and 1f.. 64.. and P. 0 (640) fo" P ."'(cos 0) sin 1»4'.. Orthogonality Relationships... The Q's of higherorder modes are given in Prob. (ElO) and integrating by parts. In many ways the Legendre polynomials are qualitatively similar to sinusoidal functions. we have rl 10 d¢ fa" dO sin 0 ( 1fl 0..(cos e) which are solutions to the Helmholtz equation. An arbitrary function can therefore be expanded in a series of Legendre polynomials in this interval. (cos 8). (344)].pq sin 0 dO if n .~. derive the necessary orthogonality relationships.(kT)P . For example.) ds 2 . form a complete orthogonal set in the interval 0 to 11" on 8. becomes Equation (639) then = 0 27rkT'(j.(cos O)Pq(cos 0) sin 0 dO = When n "'" '1.(cos e) j.SPHERICAL WAVE FUNCTIONS 273 Comparing this with Eqs. (638) to a sphere of radius r.jqj~) This must be valid for all Hence.V"". We shall. the P. the integral itself must vanish. Hence.. . 63. The functions P. choose '" ~ j.V"". an arbitrary function defined over the surface of a sphere can be expanded in a series of tessel'al harmonics.. similar to the Fourier series in sinusoidal functions.. Assuming this to be the case and applying Eq.1 ~..j~ . we have for [P. 10" p.t is 25 per cent higher than the Q of a circular cavity of height equal to its diameter and 35 per cent higher than the Q of a cubic cavity. are well behaved solutions to the same Helmholtz equation. .) 1f. For our proof it is convenient to use Green's theorem [Eq. (cos O)p sin 0 dO = 2n ~ 1 (641) which can be obtained by using Eq. sometimes called zonal harmonics. form a complete orthogonal set on the surface of a sphere. ~:. ~ at) = 0 (639) In particular. Tj 80.e q. sometimes called te88eral harmonics. Q th3. we see that the spherical cavity has a. (558) and (2102).(kr)P. which is 1ft (~. (cos 8) cos mq. ~~ . Iff (~.~. in this section.
except the one n . (641).l a.' 2 • f(B)P . n"q For the t/I integration.q. (640).J~ ..~) ..(cos 8) sin 8 and integrate from 0 to on B . we have the known orthogonality relationships 102" sin m</1 sin p</1 dt/l 0 r"· r" )0 8lDmt/lsinp</1dt/l = Jo cosmt/lcospt/ldt/l .. hence Eq.(COB e) 'If (642) Multiply each side by P ." e)p.(c." e) sin ede .·1'.P.P.(cos e) COB m~ T.274 TIM:EBAltMONIC ELECTROMAGNETIC FIELDS o to 'It: on 6..(c.(0 ...." e) sin ede .ivj~) 102• d¢ 10" dB T. For a more general result. (cos 8) sin mt/J and assume two solutions to the HelmholLz equation as (644) These are well behaved within a sphere of radius r.(c.  2. The result is a./ sin Bde "" 0 T The term outside the integral vanishes for arbitrary hence only when n .. ::::II m"p mp¢O (645) ..~(81t/J) = P. define tbe tesseml harmonics as T••·(e...+11... It converges in the same sense as the usual Fourier series. which is given by Eq. l• a. (&39) applies and reduces to kr2(j..." B) sin B de (643) Equation (642) with the coefficients determined by Eq.. • Each integral on the right vanishes by Eq.(c. (643) is called a FourierLegendre &erie.P... J: f(e)p. J: P. we assume To obtain a Legendre polynomial representation of a function /(8) in fee)  .
.m) I ('..') (648) l L (a .S••' + b.46). 2..'(9. All terms except those ha. (n m)!}o d~}o d9f(9. dq.... the final orthogonality can he expressed 88 I:' d~ I. + 1 }o (" d~}o d9 f(9.~)T. ·0 . For this we assume f(9. 0 . o  2n 4.'(9..(coa 9) (.. (3. = b..S.q..ving tn. 2n + 1 (n .~)T ..) on n spherical surface.~)8m 9 + (649) 2n 2. 0 multiply each side by T 1/) sin 0.~)8m 9 .. + 1 (n + m)l}o d~}o d9f(9.'(9...~)T. d~ f.... (n The series Eq..~) P. Eq. q vanish by Eqs. which is 1P (Eo X H' I E' X Ho) . + b. (646). ds ... cos mq.. _0 . (334)]. A twodimensional FourierLegcndrc series can now be obtained for a... \Vhen m. (' . l l • • (a.~) = . and by Eqs.0 (650) valid when no sources are within the surface of integration.m)! (2'" r'" . and integrate over 0 to 2r on rP and 0 to 'lr on 8.i=e + m)l m)! + 1 (n  m '" 0 (647) which can be obtained by using Eq. (649) converges in the same sense as the usual Fourier series. e or o. (648) with coefficients Eqs. To establish the desired relationship. (647) a. function !(O. .'(8. (E16) for P . . /0" d8 T ..! We could just as well use the vector Green's thoorem.SPHERICAL WAVE FUNOl'IONS 275 Hence.. .... d9 T.. .~)J 8m 9 t' i 2n+1 2r !n 2n m=O. and integrating on 8 by parts.) 0 0 (646) sin 8 = m.n ¢ p. a.) T 2 pq l(8.g where i ..(cos 0) . Still another orthogonality relationship is of interest when dealing with vector fields. n = p.q. d9 [T•• '(9.. n "" p.. we have 4T 10 2. l<'(lr the ... sin m41)P....~) sin 9 10 . <1.q. ..'(9.. we start from the Lorcntz reciprocity theorem (Eq.
. ni and i = e or o.. For arbitrary T and n '" q this equation CRn be satisfied only if the integral vanishes.'.)' + do e~. Also.(kr)T. Space as a Waveguide. 2. (626) with F r 0 and At respectively. n.(n + I) (n + m)! 2n + 1 (n m)! which can be obtained by integrating once by parts and using Eq. q '''}o )0 sm ao a8 sin 8 84> oq. (644). 64..0 ::.(n + 1) 2n + 1 ~ 2". There exists a set of modes TM to T.276 TIMEHARMONIC ELECfROMAGNETIC FIELDS """ a and b fields. i (651) When m.. by the orthogonality relationships of Eqs. we obtain !(J'J J'J) ("d~ ('dB(' BaT. ' ..(B. i .' +_1_aT••'aT".'O)(h)! (653) where n = I.. f" dB (Sin 8 aT"'. generated by ( A) "'''. . even though there is no material guiding the waves.)_o o . 61. o.. (647).. we have r d~ j.(21(kr) r .)'] "". (645) the integra../) ~ 0 ao ao smO iJq..(B.. 1. . We have seen that in a complete sphericalshell region (0 ::. (650) to a sphere of radius r. n.7'. The spherical coordinate system is defined in Fig.jaT.' iJT p.l vanishes if m ¢ p and i ¢ j. }o}o ('1'" dq.. .'(B) J 11.. q. ::. m = 0. i"'" e (652) n. q./ aTpl + _.. the space is often called a 8pherical waveguide.I! P._1_ aT.. . In the dual sense there exists a set of a. 2..~) A.q. 8 ::. 211") only spherical wave functions of integral m and n give a finite field.'  J. T functions are defined by Eqs. . (2).p m. 3. Thus.ve functions can be thought of as the Hmodes of free space. ~ J. and the field is given by The (654) Inwardtraveling waves are represented by the (!) and outwardtraveling waves by the 11. j.~) Applying Eq.::>I p. \ l m = 0. choose those obtained from Eqs. .' dB [Sin B(a~. . 71." When viewed in this manner.(kr)T•• .. The fields specified by these wa. q. . i :.
so an interchange of E by H and H by E in Fig. and i = e or o.+. The spherical modes are qualitatively similar to the radial modes of Sec.(2)(kr) E. genera. 3.'" = H. 1.." To illustrate.+ = . .. Mode patterns for the TM ol and TE ot modes are sketched in Fig. n. For 9(_ (b) Fro. .(I)(kr) (657) where the superscripts + and .T. . 64. . a summation of them cRn be used to represent an arbitrary field in a sourcefree region. . Mode patterns lor the (a) TM ol and (b) TE ol modes ollrce space.E. (e. m = 0. consider the radially directed wave impeda.)•• .+ E. In.. . respectively.H.m(kr) HT£.VXE~~i !w..and inwardtraveling waves.. 64.= JY] O. 2. C21 / (kr) E. There is no welldefined cutoff wavelength but rather a Ir cutoff radius..= H. B.{l)/(kr) = . . Note that. 64 gives the TEo I and TM ot mode patterns.denote outward. B .nd k. field is given by The •• 1 = . (656) Thc set of TM plus TE modes is complete. .. Z_/'M = (Z+r™)*..SPHERICAL WAVE FUNCTIONS 277 modes TE to r. for real Y] a. (655) where n = 1. .nces for the TM modes Z+. The 'fM and TE modes are dual to each other.ted by (F..+ "'" 1'1 B..H. that is.~) . 2. 53.W(Ier») n.
aod predominantly resistive when Icr > n. The frequency derivative of the various wave impedances is of interest for determining the bandwidth of various devices (see Sec. pp. ":. Chu. which also illustrates the above cutoff phenomenon l was devised by ProCessor Chu. as shown in Fig. The dissipation in the resistive element at the end of the network represents the transmitted power in the field problem."' 11 . 1 He took the wave impedances and. December.(l)'(kr) (658) The behavior of these wave impedances is qualitatively similar to the behavior of the twtrdimensional wave impedances. I L. 65b. Appl. The value kr = n is the point of gradual cutoff. 56. 194. 11631175. 19. using the recurrence formulus for spherical Bessel functions. "'" 1'1 {j~ + 2n _ jkr 1 +l 1 2.." B.(t)CJ... For example.E. the wave impedances of Eqs. modes is shown in Fig.3+ jkr (6. for fixed r.11~("ik::rl H. It is therefore apparent that.J o .r) lJ. for the TM impedance of outwardtraveling waves Z+rTV. .+ '"'" _j.+ H. J.. the higher the mode number n the less easily power is transmitted by a spherical waveguide mode. Nate that this cutoff is independent of the mode number m.\9) 1 +31 +~ jkr j~ + 1 This can be interpreted as a ladder network of series capacitances and shunt inductances.278 TDlEHARMONIC ELEC'rROYAGNETlC FIELDS the TE modes the radially directed wa. Physical Limitations of Omnidirectional Antennas. Phl/. ~5a. illustrated by }~ig.. obtained a partial fraction expansion.Ull(kr) E. (657) and (658) are predominantly reactive when kr < n. In other words. 613).ve impedances are Z n: +r 0:: ZTE  E. The equivalent circuit (or the TE. Those of us familiar with filter theory will recognize thc equivalent circuits as highpass filters.H. A Dovel way of representing this frequency derivative. vol. J.+ O.8..+ __ E.
.. 66.rn in Fig..ves 'W. An approximate calculation of the Q's for Q > 1 is shown in Fig. stant and ~ .ho·... We can have waves outside or inside eo single conducting cone. In TM waves 'W.... Note that for kr > n the wave impedances are low Q and for kr < n they arc high Q_ This again illustrates the cutoff phenomenon that occurs at kr .f. Equivalent circuits for the (0) TM. and W.. A number of structures capable of supporting radially traveling waves can be obtained by covering 8 = con.. so the Q's of TM waves are equal to the Q's of the corresponding TE waves....ZTII _ _ •• L 2nI ' "' la.... 65.. Other Radial Waveguides. > OW.i"f. while in TE wa.. the two eases are dual to each other. &70 and b... . on 1/1 and .n.constant surfacee with conductors.8PHERICAL WAVE F'ONcrION8 279 2n3 " .. 66... z'£ __ •• FIG. C· 2n1 " 2n5 " ~. 88 shown in Fig.""'H". And (b) TE. These two cases are actually a single problem with two different values of 81• The fields must be periodic in 2. and (J' is the power dissipated in the resistance. > 'W.. for modes of order n can now be defined 88 (660) 'W. are the average electric and magnetic energies stored in the C's and L's. where W. '" A quality factor Q. Such II radial waveguides" are e. &7.i'f. However. model of free apaee.. > OW...
. 2.ELDS  kr 10'10.'(eos 8 )cosm..'~(lT) (.280 TWERARMONIC ELECTROWAGNE'rIC I'l.pl B. Quality factor'll Q.er v must be a solution to P . .ee. finite at 6 = O.P. • I I '" am m". .(cos 9.) . r mode functions (661) where m = 0.. smmep (663) where m = 0.66.2..0 (662) Also.. . = 0 at 8 = 81.P. we choose the TE to T mode functions '" (F. 1. n. . the parameter" must be a solution to (664) .. modes of free llP:t. _ _ To satisfy the boundary condition E. Hence. the parnmet.)" .'(eos 9) cos mq. 1. for the TM•• and TE. . .). we choose the TM to (A."'(kr) . To satisfy tho boundary condition E r = E.0 at () = 01.
)P.2. 1. Choosing the latter two sohfiions. (b) conical (waves internal). by Eqa.'(eoa 8) p. or P.(cos 8)."( eoa 8) P.p. of course."( eoa 8.(..coe 8). I I fJ. . Because of a scarcity of tables for the eigenvalues v. we find modes TM to T defined by (A.( .) = 0 . The field components are.) .)p.(eos e.cos e. and F. P. .cos e. . 67.l::(kf') (665) and the v are determined by the rootB of (666) P. (e) wedge. The biconical and coaxial guides of Fig.(cos 8) and Q.). 67c and d are again a single mathematical problem. it is difficult to obtain numerical values.(cos 8) and P.) (f) Flo. (c) biconical.) ."(eoa 8.)] lc?S m4» 18lD m4J where m = 0. (f) born.. Now both 8 = 0 and 8 = 'If are excluded kom the region of 6eldj so two Legendre solutions. obtained from the A. are needed.(eos e.P. = [P. (d) couial. (a) Conical (wavC8 external).(. Borne spherically radial waveguides. (1)26).SPHERICAL WAVE FUNCTIONS 281 zl ~ (0) " (0) (e) (d) (.
of Eqs. The wave impedance in the direction of travel is (611) which is the same as for TEM waves on ordinary transmission lines. . .(COS 8) dP. ..and the 1:1 are determined by the dP._( _ cos 8) dP. At a given r.0 satisfy both Eqs. where m = n. 1."'(kr) .cos 8.( .". ./2) . (666) and (665).)_.(cos 8.) P. E d ' 1 cot (8.) . or transmissionline.)] lc?Sm4>\ sm mq. .cos 8. The characteristic impedance defined in terms of voltage and current.al(kr) TOOts (~7) of (665) 0.0 exists. mode. TEM. and F. _ [p. e±iAr uxr 810 8 'f = + 1'Sl08 e±i i (610) kr where the upper signs refer to inwardtraveling waves and the lower signs to outwardtraveling waves. ' r 8 = 311 og cot (8J2) e (672) and the current as 1 f02~ H. t.je±/i'r (613) .. We could redefine Eq.Q.) dP.( . = 'h I.~~~s 8. The dominant mode of the biconical and coaxial guides is a. sin 8 d.(cos 8. (665) such that the limit v .( ~8'cOS 8. but instead let us separately define the TEM mode 88 a TM oo mode defined by (A.) _ 0 de! d(h dB.(cos 8)B. dB I Again the field components are found from 'the A r and F r of Eqs..+21r.log cot 2 ('l'J)~tt (II B (669) The field components of this mode. (626). for the modes TE to r we have (F. (665) and (667) vanish. The eigenvalues m = 0. but the A.p . v .282 TW&IIARMONIC ELECTROMAGNETIC :rtELDS Similarly. (665) and (667) according to Eqa. is of greater interest.) dP.2.he volLage is defined as V 1:3 J. determined from Eqs. (626).) _ dP.. are ( Eif' _ H • 'k J.
The values of to are those of Eq. 67 by one or two conducting spheres. The TE modes are defined by (F.'''(kr) '" (677) where n . the dominant mode is the dominant TM to T mode of the complete spherical cavity. except that the travelingwave functions 11. functions allowed. to is given by Eq."(cos fJ) of nonint.. For this reason the biconical and coaxial radial lines are called uniform radial transmission lines. Again. cot (8. the TE mode. . (548) and (549). Hence. 2. . For the hemispherical cavity of Fig. The characteristic impedance is v+ V. 3. . . 68.. 2. . be no TEM mode. Some examples are shown in Fig. ./2) (674) Z.(I)(kr) are replaced by standingwave functions J. Spherical waves on the wedge waveguide of Fig.lUCkr) where n = 1. . . The fields in each case can be expressed in terms of mode functioll8 which are the same as for the radial waveguides of the preceding section. the TEM mode being a cylindrical wave defined by Eq.'CC08 8) 'in w~ n.and w= (675) pr ~.)_ = P . . the wave functions will contain only the PIl"(cos 0) with n an integer and to determined by the boundary conditions. (kr).[+ = . 680. 2.SPHJ!:RICAL WAVE FUNCTIONS 283 At small r these are the usual circuit quantities. 3. (665) and (666) with m changed to wand only the sin wq.12) Note that the various equations are the same as for the usual uniform transmission lines. Other Resonators.egral v and w. (676) with p "'" 0. . 1.. .[_ = 2r log cot (8. 67/ will require Legendre functionsL. There will. and to is given by Eq. There is no TEM spherical mode. Similarly... the sphericalhorn waveguide of Fig. 68.(kT) and IV. .(l)(kr) and 11. 3. of course. (676). Resonators having modes expressible in terms of single spherical wave functions can be obtained by closing each of the radial waveguides of Fig. 2.I. 6·6. functions allowed. The TM modcs can be defined by Eqs. (667) and (668) witb m changed to w and only the cos wq. Finally. Numerical calculations are hampered by a scarcity of tables of eigenvalues. = P. We then find TM modes defined by CA. (676). can be defined by Eq. .'(C08 8) coo w~ l1.).. 67e exist for all fJ but only for restricted fjI. Let us calculate the Q's for the dominant modes of the first three cavitics of Fig. C676) with p = 1.
571) I..ve functions.13) The power dissipated in the plane wall is (el'..<R2r o H. (635»).  ~ JI (2.744~) sin 8 w 4<.14) The power dissipated in the hemispherical part of the walls is onehalf that dissipated in the walls of the complete spherical cavity. considered in Sec...284 TWEBARMONIC ELECTROMAGNETIC FIELDS z " zl ~aol (a) NJ fal (b) zrt (0) ~!J 'r Z! . hence 4r (el'. (6) hemisphere with cone. = <liS (1./2 r dr = <lI2r(O. 62. (f) !legmen"...) •••.." .. (d) (.~ ....~1.)_ . Cd} conical. hence = 3k (1. (a) Hemispherical. Some cavities having modes expressiblo in terms of singlc spherical wft..) f (f) FlO. and the stored energy is onehalf that for the complete spherical cavity [Eq. The magnetic field is ll. &8. (e) biconical. (e) wedge.
5 per cent lower than the circular cavity Q. .SPHERICAL WAVE FUNCTIONS 285 Thus. 3. When small losses are present.573 (ii ~'W . = jZo tan ka I . (678) If we compare this with the Q of a rectangular cavity (Eq.r. A sin ken .41 times the lowest resonant frequency for the hemispherical cavity.nces occur when ka = n7:/2. In other words.r) r sin 8 _ . 68b and c arc theoretically important because they have circuit terminals available. . A cos ka Hence. . 2. but we have removed the mode degeneracy. where n = 1. The hemispherical cavity Q is 54 per cent less than the spherical cavity Q.r) E :. (674)}. the hemispherical cavity Q is only 3. (2102») and with the Q of a circular cavity IEq.2 per cent higher than the rectangular cavity Q. the input impedance is infinite for n odd (antiresonance) and zero for n even. for the same heighttodiameter ratios. (We saw in the preceding section that the TEM mode of the bieonieal guide is a uniform transmissionline mode. = 0 2.) The resona. tbe input impedance is large for n odd and Bmall for n even.2r H..58 for the rectangular and circular cavities. secn by the source is VI. The dominant mode H ...  _ A cos k(n .. = V.. the input impedance seen by the source is ZI. . = lim ~O The current at j. From Tables 61 and 62 we find that the second resonant frequency is 1. compared to approximately 1.'3"1 rsin8 will be excited if the cavity is fed across the cone tips. a voltage and current calculated at the cone tips have the usual circuit theory interpretation. la (679) which is the usual formula for the input impedance of a shortcircuited uniform transmission line. the source is II. or w.0. (558)1 we sec that. r d8 = 2'WjAZ o sin lea ro}" The voltage where Zo is the characteristic impedance [Eq. = lim [" E. =2 y. .i$>. and 4. the Q of the resonator is Q .. The cavities of Fig.=" "n"' nr (680) In the lossfree casc. r dq.
D.8.4Z. N.824 log cot (0'.0.. 288290. in which case Q .ance is minimum when the cone angles are 81 "'" T . =. where Z. \.1) in more detail.. because Zo is also a function of 8. The input conductance [Eq.2°.4<ll T' [ 1 + 0. 24. """ Iff T·Z. (681» is not minimum when Q is maximum. is given by Eq. T  csc 8 1 + := CSC 8. For exam· • B. the minimum conductance is obtained when 81 _ 7.Ql vl. "Electromagnet.4<ll 1 T' I+ 0. 29." pp.824 iOi!{cot (0. the input conduct.2) 1 + CSC 8 1 ]' (633) This Q is maximum when 8• .TIllEHARMONIC ELEcrnOYAONETIC nELDS Let US consider the lowest resonance (n =. 6&).Q . because of the feed system.276. 1943.  rv:fi . This is a lower Q than that for the hemispherical cavity without the cone [Eq. For the conefed hemispherical cavity (Fig. 6&.350.J. and 8. 33.. The sources of the lowest. For the biconical resonator (Fig.)' . Note that this is smaller than the Q's of other cavities that . Inc.order spherical waves are current elements.72) tan (0..0.ic Waves.1°.I' IHI'd. Van rJ08trand Company. in which case Q .9. A.. Princeton. T (631) <T>" wOW The energy stored 'W is simply calculated as VI =. (678»). Schclkunorr..90°. The input conductance at resonance can be determined from the power losses as G I. 67.5°.we have considered because of the introduction of the biconical feed system. = ~ IAI'Z. we have the conefed hemispherical cavity of Fig. =. (674) and Q can be calculated in the usual manner as 1 Q ..5°. treated in Sec. for which Q ./2)] . Sources of Spherical Waves. (632) This Q is maximum when 81 8. 68b). Thus Gla Q(2rZ. . In the special case 8.
.. the field can be represented by a radially directed A given by (685) The field of the current element is discussed in detail in Sec.SPHERJCAL WAVE FUNcrJON6 287 Z Z ..U> is the spherical Hankel function of Eq. can be obtained by tbe same metbod as used in Sec. X n Y X II Y .t . The dual source is the magnetieeurrent element of Fig.. the electriccurrent element of Fig. or F~ is the same as AI or A.j • t y (d) FlO. 29. .) llt). with I replaced by K. 69c.1. Tbe fields of the dipole and highermulti pole sources. Alternatively. 69a radiates a field given by H . pte. z r Xl r y X y X (0) X • n In (0) y (b) Z Z z n 01 (. Some 8OU~ of spherical wavcs. rcpresented by Fig. 69c to I. 56. The field of this source is given by E . For example. . V X A with (684) where 1i. &9. X (f) \0" .V X F where F. for the dipole source of Fig.. 69b. (611).
m is for the dipole of Fig. Hence for the dipole of Fig. jklls a 4. we have A• = 81".288 TIMEHARMONIC ELECrROMAGNETIC FIELDS where A. alA.. we find A.j h. (684»).(cos 8) (686) and H = V X A. for the quadrupole source of Fig. given by Eq. 69d.(t) 8. 6ge. We also . for the kllla 4.'(cos 8) sm ~ (&88) Thus. we have which caD be written as This is a firstorder wave function of n dipole source of Fig. the vector potential is a firstorder spherical wave function. For the dipole source of Fig.! all dz clA. A • + _0 where r = 8 .iJz ho(t)(kr) v!x t + yl + Z'. 1.I is the potential from a single current element [Eq.  co. 69c. For example.iJy where A.. . AI! the separation 8 is made small.  k'Ils 4.j h.1. Thus. 69/. Similarly. all wave functions of order one can be interpreted as the A.U'(kr)p.'  aA' IJz . This procedure can be extended to highermullipole sources in & straightforward manner. of dipole sources. m . (686).U'(kT)P. Also. 69c A..
Let w first consider the plane wave eJ.'"{kr)P.SPHERICAL WAVE l'UNm'IONS 289 MV. j" 2. [h. P .(cos 9) sm 9d9 = (2n + 1)1 The nth derivative of the righthand side evaluated at r . A convenient method of obtaining the desired results is that of Sec. Now that we have wave functions in three basic coordinate geometries available." c" _. equo. 61 for tbe coordina. .ting the preceding two expressions..te orientation).. of a multi pole source of 2n zdirected current elements. the number of possible wave transformations becomes very large. 12:11" ht(t kr)Ptl(cos 6) sm ifJ (689) In this manner we can identify each wavc function of order n with the A. [h..(r)P. (641) we have 1. (640)].{cos 9») ~ = ~!.+t(n!)l j..and express it in terms of spherical wave functions. This wave is finite at the origin and independent of +i hence an expansion of the form • e> = e>' a. kh..0 is 2"+I(n!)I (2n 1)(2n + + 1)1" + 1) Hence.(cos 9) sin 9 d9 r 2:."'(kr)P. multiply each side by P. ('r .<"{kr) T ~] rar ky. and by Eq.(cos 8) sin 6 and integrate from 0 to T on 6. = jk l IlJh8t l( . all terms except q "" n vanish. To evaluate the a. 69/ is A. 58.. Wave Transformations. Because of orthogonality (Eq..=Ois .j'(2n . !.'(cos 9) sin ~ Hence tbe vector potential of the quadrupole of Fig.  ~ h.(r) The nth derivative of the lefthand side with respect to r evaluated at .. 68. cos' 9 p.(I)(k1') sin B cos B sin ~ . we obtain a. r' ht(t'(kr) . We shall here establish only a few representative transformations involving spherical wave functions.~ 1 j.(cos 9) ••• I must be possible (see Fig.j..
p.. we differentiate each side 2n times with respect to r and set r = O.(r) = i..I. substituted back into our starting equation.. Now let us consider wave transformations corresponding to changes from one spherical coordinate system to another..)P. + •• 0 (692) Note also that the two equations preceding Eq.l n!(n Hence the desired wave transformation is • .(." h" eir . As before.J. (692) establish an integral formula for iz.'''(I' .290 TIMEHARMONIC ELECTROMAGNETIC FIELDS which.(r) b _ (I)·(4n . (cos 8) sin 6 dB (691) Equation (690) is the desired transformation expressing a plane wave in terms of spherical wave functions.(cos 8) o to 11' on 8.)P. For example.. Hencc. ..('8>08) ~ 2'" 'nl(n 1)1 ]. and symmetrical about 8 = 7f/2.  I• • •0 j·(2n + l)j.(cos 8) sin 6 dB = 4..(cos8) L. Transformations from cylindrical waves to spherical waves can be obtained in a similar fashion.~ 1 .(cos8) (690) Note that we have also established the identity ..(..'I eJlrr'l . consider the field of a point source at r' h.. which is finite at T = 0. sin 8) ~ I• •• 0 b. independent of . / To determine the b". (r). "\' (I)·(4n 1)(2n .r'1) . there exists an expansion J. 11' .(.I) I ..j.o ' P . This gives + 1)(2n 1) I 1)1 2 1 . gives eP = .(p) ~ J..)P. To illustrate. .. J.(.. consider the cylindrical wave Jo(p). we multiply each side by P a(cas 8) sin 8 and integrate from The result 18 for Jll(r sin 8)P"..(p) ..
ble wa.(cos Furthermore. If we let the source recede to infinity. r ..(. >" where the c..c2l(r)P.".(r)P.SPHERICAL WAVE FUNCTIONS 291 z source FlO.ve functions r > r' are h. are constants. hOU1(jr  r'D = ••• ••• L c."'(")j..jP... where rand r' are defined in Fig. the field in the vicinity of the origin is a plane wave.je~ir' r'o. 610.(cos t). the field is symmetric in rand r' j hence we construct n.. We desire to express this field in terms of wave functions referred to r = O..j"(r)P ..(")h. and allowa. The field bas rotational symmetry about the r' axisj so let us express the wave functions in terms of the angle t where cos t = cos (J cos 8' + sin 8 sin 8' cos (<p  <p') (693) Allowable wave functions in the region r < r' are j.<" .(.(cos I) • L c"j.)P.rl') jeir' I' . Using the asymptotic formula we have for the leCtohand side of the preceding equation holU (1r . L c... 610. r'_oo and for the righthand side .h. r eir' .(cos I) • .'_0 >.(cos 9) . Spherical coor· dinates or r and r'.
"'(r')j._0 (2. + l)j. Finally. 8') co.(co. 8)p.wHiIl Book Company. Since ha(l) = ham. 406414._0  r < r' (694) r> r' I This is the addition theorem for spherical Hankel functions.t'l). Figure 611 represents a conducting sphere illuminated by an incident plane wave. (690) shows that c" = 211.292 TutEHAlUlONlC ELECTROMAGNETIC FIELDS z r 1 .. plus some ot.(cos!l L: (2. (694) is nn addition theorem for io(lr . The real part of Eq. 1 Equation (695) is an addition theorem for Legendre polynomials.(r)P."'(r)P. 69. In other words. one can express the zonal harmonics P . "Elec:l.(r')h. . (694) is also valid for superscripts (2) replaced by (1). The identity is (n . New York.(cos t) in terms of the tesseraJ harmonics P. .ton. 610 cnn be expressed in terms of wave functions referred to the 8 = 0 axis. can be found in Stratton's book. Take the incident wave I J. 1941.r'J) = L: . (695).her wave transformations that we have not treated explicitly.rD.(cos 8)h(m4».(co. + (695) where too is Neumann's number (1 for m = 0 and 2 for m > 0).I I y Flo. Eq.romagnetic Theory.. + 1j hence h.heorem for nD(lr .(cos n ."'(lr . &. m(~ . A. x t Incident plane wave J A comparison of these two expressions with Eq. + l)h. a wave function referred to the t = 0 axis of Fig.m)' P . and the imaginary part is an addition t. Inc. !) ~ __ I '_ (n '' m)! P . Slrat. MeGra.(co. The proof of Eq.11. A plane wave incident on a conducting sphere. Scattering by Spheres." pp.~') \' .
we express t.' ~  jEtk:~: ~ L . • .his incident field as the sum of components TM and TE to r. the magnetic vector potential as Ai = E. (624).I. nnd Fr from Hr. E~. ej~r~ 2>'(2n + l)j.(kr)P.(kr)P.J._0 Finally.ring this expression with the preceding formula for Er'. Noting the form of E r \ we construct.. .'(eDs 8)  aP. = j'(20 + I) n(n + I) (698) A similar procedure using H.' by Eqs. . (626) we see that AT can be obtnined from Er.n(n + I)J. we see that a.: ~ 2>'(20 + I)J... sin 9 E~' _ Eo c~:/ :9 (ei l'r_') Using Eq.SPHERICAL WAVE FUNCTIONS 293 to be xpolarized and ztraveling. (698).. we can write this as Ei = E.jEt:... t. L • a.'(cDs 8) Compa. E. in terms of an Fr and an Ar _ From Eqs. (623) and the relationship E. a. . is. (690). I ~.. H.cos q.J.(eDs 8) • . = Eo c11ir_' (696) Er./ao = P.(kr)P.hat. (697) and evaluate E.'(cos 8) " .. we obtain • Simplifying the result by Eq. CDS w. ~ L. that is.(kr)P. using Eq.' and F r ' gives Pi = where the a. _ EI1~~flir_' Eo eft. _ O. For convenience in applying boundary conditions. sin ~ . (626). = = Ecl.a. The r component of E' is '.(kr) :0 P.' .he II. t. we obtain l .'(CDS 8) (699) Me again given by Eq.. 0 term of the lIummation drops out because ptl  Note that.
The surface current on the sphere can be found according to J.m(ka) where the a.(kr) + b. = ~ cos ~ E W" F.<11. .re given by Eq.. Therefore E and H arc given by Eqs.B. the rest of the solution parallels the cylinder problem (Sec..  . of the same form as the incident field with J. The result is J. B. the sum of the incident a. a.(t"(ka) .. 64) with many modes superimposed. I(cos 6) F r" = ~o sin q.Ul(ka) This completes the solution.J ..(kT) + c.I(C08 8) ] sin 8 O. of course.. The scattered field will be generated by an A."'(kT)jP. . • [a. Note that the problem call be viewed as a.(~'(ka) . (6·98). which require that . s. replaced by 0 . I(cos 8) The total field is. and P....1 • (6100) c"fl. • 4· . we construct scattered potentials as • A r" == Eo cos W" q. Hence. '< ..m(kr)]p.I'(COS 8)] iB. (S)(kr)P. 59). • s. shortcirouited radial transmission line (Sec.I'(COS 6) 0 ka O.  ~. _1 The boundary conditions are E.. + jP.'(cos 6) b  ..  a .• ka 8iD412: a [P. 2:.J. B"u)(kr)P.i E . '\' b.. IW(ka) J. B..'(cos 6) (6101) I 2: [a. "'" E.nd scattered fields. "'" 0 at r = a.. I(C088) .(ka) J:(ka) (6102) c a . = iE 11 • cos Q \ ' a [Sin 6 P.J.(t)(ka) (6103) . The distant scattered field can be . (626) where A..B..n 6 B. .n ~ 2: ... = U r X H at r = a. .294 TIMEHARUmnC ELECTROUAGNETIC FIELDS Now that the incident field is expressed in terms of radially TE and TM modcs..sin 8P. .
For small ka.. • j~o eib cos. Echo Ar'e& of a ooDdue...4 and was a/' FlO._. B)] (6104) . (6105) becomes dominant and .  E•. of small spheres varies as >.' I '_r .''''(ka) (6105) '" ~ 1 / A plot of A.l is shown in Fig. [b P B)] where the b" and c" are given by Eqs. \' (I)'(2n + 1) .. 612.. I'{C08 The result is e. B B) ..I>.. . ..+te1u and retaining only the terms varying 88 1/r. j" sin P.SPHERICAL WAVE FUNCTIONS 295 found from the general expressions by using the asymptotic formula O. (330). we find A _ ~.1.0 1. P. which is .. • • o eikr sin..._r • __ r/2 From this we can calculate the echo area according to Eq...c.U>{kt) ~ j .4 V V Lr A • + 9X' (1m)' . 0."(cos 0) + (21)' n(n and the Wronskian of the spherical Bessel functions.'j') ·IE l .8 1.1 f1 o 02 0. \ ' j" jE P"l{COS 8) _ c sin 8 !'(cos kr 4 " sin B " " . (6102).IE..'(cos 0) ~ (I)' n(n 8m B Ir 2 (..10 . the backscattered field Of particular interest is E. + 1) • '< . 4.0 1 0. B.' ~ E"I . It states that the echo area. ol + 1) sin 0 P ."'(Im)B..2 which is a good approximation when a/~ < 0. 612.:~~o.._. Equation (6106) is known as the Rayleigh scattering law.E.. tien shown dashed)..tooo 4...6 0.ting apbere of radiua a (optical approrima.. .. the n = 1 term of Eq. (6106) 0. A • = ~~ Making use of the relationships P. 2: [b...
.+1 ~) (6108) so the n = I terms of Eqs. Hence. (b) magnetic moroent. (6104) become dominant for small /ro. (a) Electric moment. Components of surface current giving rise to the dipole moments of a conducting sphere. )...~ n laO n +1 [2'(n  (2n)! 1)']' (ko)"+' . (6102) aod (698) that b. jk' (ka)' (6111) The ratio of the magnetic to electric dipole moments is IKif Ill.k (ka)! sin 4> r cos 0 . The region between the Rayleigh and optical approximations is called the resonance region and is charac· terized by oscillations of the echo area. (6110) plus the field of a ydirected magnetic dipole Kl . A surface z J.+ Eo . .. (6109) (~ E. _ _ la_O c.in the blueness of the sky..' . A.O r rrlJ. = '1'//2. kG.E.. z x (aJ (bJ x FlO.jkr B. 613. Let us now look at the field scattered by the small conducting sphere. we find from Eq. Using smallargument formulas for the spherical Bessel fUllctions.. J For large spheres (6107) lI'a 2 which is the physical optics solution.k (ka)l cos 4> (cos 8 k.296 TIMEHARMONIC ELECTROPoL\ONETIC FlELDS first used to expla.1) A comparison of this result with the radiation field of dipoles shows that the scattered field is the field of an xdirectcd electric dipole Il = Eo ~~{ (ka)' 2.' laO 10 Eo . Figure 613 illustrates the origin of these two dipole moments.. e.. at large distances from small spheres.
. fl.a) + .. let the region r < a of Fig.. co.j vi.(k. while a circulating current gives rise to the magnetic moment.. fl..(k."'(koa)J~(k. H..'(cos 0) WJlO F. lSd....+ = E. In addition to the field externnl to the sphere. fl. in contrast to staticfield problems..a) + ."'(koalJ~(k. we find b.. and imposing the above boundary conditions.vi"...(k.a) .. ~: sin ~ .al v'.. d.. (6101). Ei . 2: .''''(koa)J.a) ..v. J. "'" ."'(koa)J~(k.(k.....''''(koa)J."S 297 current in the same direction on each side of the sphere gives rise to the electric moment.a) a.a) a • c.J.. Boundary conditions to be met at r :."'(koalJ~(k. the scattered field of any small body can be expressed in terms of an electric dipole and a magnetic dipole..(k.. cos ~ \ ' d...''''' n. J~(koa)J. fl. In the special case of a small dielectric sphere. n.ld _ 0.t .vi..r. In general... (&112) for T < 0.. but the electric moment must always exist.+ = H.'(cos 0) The superscripts ..n'(koa)J.. fl...a) _ . and superscripts + denote the region r > a.(koa)J~(k. For a conducting body. there will be a field internal to the sphere.... fl. J~(koa)J.denote the region r < 0.al • where a" is given by Eq. 0" vi"". Now consider the case of a dielectric sphere. (698)..a) . • (6112) J...(koa)J~(k. tangential components of E and H must be continuous.. ote that.: a are E. (626)....(k. Determining the field components by Eqs.. 1010. (6113) e _ • i~ 0 v.)P. = vi.''''(koalJ .lP.E.. (6101) for r > a and Eqs.vi.+ = lJ.. . that is. using Eqs.. The conducting sphere can be obtained as the specialization J.al v.lJ. such that k" remains finite. J. the magnetic moment may vanish.+ = E.(k. and the region r > a by to. the n = 1 coefficients . that is. specified by • A.. specified by potentials of the form of Eqs..I.a) .co is not sufficient to specialize to a conductor..SPHERICAL WAVE FUNCTIOr.. t... 611 be characterized by t. L...r...~ E.I.(k.
the specialization represented by Eqs.+2 Note tbat the magnetic dipole vanishes if the dielectric is nonmagnetic.k' (ka) <. . if IJr "'" 1. _ 47rj 1(. Dipole and Conducting Sphere. &14. The conducting sphere and a radially directed dipole. Figure 614a shows a radially directed electric dipole near a conducting sphere. a magnetic material with IT "'" 1 would scatter no electric dipole field. "'" IldllJo.) where f r cc fd!fG and /J.(2 + p. (0) Oripnal prob lem. (b) reciprocal problem. A calculation of the scattered field reveals that it is the field of the two dipoles II .298 TIM"LHARYOSIC ELECrROaLAGNETIC FIELDS are dominant nod reduce to (IH14) tlt ~ 2jl. 614a in the following sense.1 2 + (IHI5) 4rj KI = u~ E tF (k a}'"' . The component of E. The field internal to the sphere is uniform in both E and H for tbe small sphere. Similarly.u. that is.Eo .1 11. In fact. (6114) is the lfquasi·static" solution'! It can be obtained by taking the de electric and magnetic polarizations and assuming that they vibrate in phase Quadrature with the incident field. Figure 614b shows a problem reciprocal to that of Fig. 610.in the direction of Ilb equals the component of Eb in z II z x x (a) FIG.
/l cJkr 4n' (6116) we have the wave of Eq.'(cos U') 'If  Finally.(cos U) " (6118) This is the radiation field of a radially directed electric dipole on the surface of lL conducting sphere. we have the planewave scatter problem treat.nd (6116). The ficld in the entire region r > b can be determined from the radiation .. E.(kb) + b... a. • n(n + 1)[.J.. we need the r' component of E. (Superscripts reter to Fig..SPHERICAL WAVE FUNCl'IONS 299 the direction of /la.. . in the vicinity of the conducting sphere we have (E%. Letting jwp. To relate this solution to that of Fig. 614a lLnd b. b and Eo are given by Eqs. q.fl.". respectively.) If the II of Fig.) evaluated at r' = b. L: ll .]Wt ur" . (6102).t o jwjJIl e/trcitr'_r ro_ 4rr which is a plane wave. Lt 0.. by reciprocity..i)t .' '=' 0 j:'. 614b is specified by Eqs. (6117) reduces to • E _ . The pattern for the very small sphere is the usual dipole pattern.. q. Eo'  8..'(cos U) (6117) where all. 614b recedes to infinity. that is. can be obtained in a similar manner. and a = 2>. 69. In the special case b ". Eq. (696). (698)... the radiation field of Fig. 8.'''(kb)]P. \ ' j'(2n thjkr . Hence. For a very large sphere it approaches the pattern of a dipole on a ground plane but always with some diffraction around the sphere.(kb) + b. which is 1 Er' . the field of Fig.. Hence. 614a.J. when the current clement is on the surface of the sphere. In particular. and also for magnetic dipoles.ed in the preceding section.(1)'(100) + 1) P . 4.. . 614a can be simply obtained from the results of Sec.fl. • n(n + 1)[a. 8' = equals Efa at r. . Figure 615 shows the radiation patterns for spheres of radii 4 = >"/4. Hence."'(kb»)(I)·P. The radiation field for dipoles of other orientations. • (0'  + k' )Ar' ~: cos ¢' L:. (6101) with coordinates primed.II .
and represent outward traveling waves. field as follows. by Eqs. Also.. A.. r>b (6119) From this we can calculate E. ~ • L.. fl. 614a) we conclude that H = u.. hence A. must be independent of q. 6Hi.H+.'''(kr)P. obtaining (6120) The a. and therefore the field can be expressed in terms of an A = urA r.300 TIMEHARMONIC ELECTROMAGNETIC FIELDS I FlO. (cos 0) . are then evaluated by equating this expression to the radiation . (626). From symmetry considerations (Fig. Radiation patterna for the radially directed dipole on a conducting Bphere of radius a..
P"l(COS 8) sm 8 d8 . For example. cross section of the guide. one having only an H...('I'(kr) 2T1l(n + 1) 2n+l . (6123) Noting iJP. . that is. we obtain the field r > G. By the orthogonality relntionship [Eqs. = • k \' a jWf. . (6120) to Eq.'''(h)P.G. there are no incomir. 616 gives rise to such a field if there exists only an E. nud (6121)./a8 = p.(2)'(ka) z + • • Tbe field everywhere can now be obtained from Eq. The slotted. The field is expressible in terms of nn A r of the form A.(t)I(lcr) a8 P. only the tangential components of E over the apertures.. When apertures exists in a conducting sphere of radius r = a.a. 49 we saw how to express the field in a matched rectangular waveguide in terms of the field over a FlO. (6119). Slotted conducting sphere. x 611.(cos 8) . we obtain ...l. A general treatment of the problem is messy.. A given sphere r = a is a cross section of the spherical guide. (626). (6118) and obtain a" _ 1l(2n 1) (6121) 41fkB . If r > a contains only free space.n.. conducting sphere of Fig.2:''. In Sec.. ".:'. so let us restrict consideration to the rotationally symmetric TM case. Apertures in Spheres.g waves.)r 4.(co. Our f9rmulas for the field r > a then reduce to ones involvinl?.T' L. a"J1.. that is. in the special case b = a we equate Eq....n.SPHERICAL WAVE FUNCTIONS 301 field previously determined.I(COS 8) sin 8 and integrate from 0 to 'II' on 8. we multiply each side of the above equation by P. (626) we calculate E. tbe tangential components of E are zero except in the apertures. 64 we saw that space could be viewed as a spherical waveguide. 616. In Sec. independent of ~ in the slot. 0) • (6122) From Eqs.  l .::. (646) nnd (647»). then the guide is matched.. By writing the general expansion for outwardtraveling waves and specializing to r .=' fao• E.. "" .
m in Eq.. 616. r Using the ". (cos 8) sm 8 de .. 8D .. Radiation patterns for the slotted sphere. 11:/2. Specializing this to r = a. we assume (6126) .. (6123).J (6125) This result could also be obtained from the planewave scatter result of Sec. For tho slotted sphere of Fig. (IH24) The field simplifies to some extent in the radiation zone. let us assume a small slot width. using reciprocity. i. is essentially an impulse fundiOIl at r = a. 617. Hence. 69. so that E.302 TIMEHARMONIC ELECTROMAGNETIC FIELDS I FIG... determined as ja(2n 1) aft = 172'1m(n + l)fl.. we have the coefficients a.(z)'(ka»)o E. + r' • I ' r_a P .j"P"I(cos 0) kro .. . we obtain E. ! l eII:< '\' a. asymptotic forms for fl..
Eqs. l We shall here l"C3trict consideration to the e.. (6124) reduces to 0) 0 + l)P. 618.. (661) to (664). 515. the conical surface gives the field of X a slotted cone. rotationally symmetric case of "ringsource" excitation of a conducting y cone.. The modes of the "conical waveguide" are considered in Sec./4 and 2" are shown. I L. Very small spheres produce a dipole pattern. l 612." . Bailin and B. The geometry of the problem Current filament is shown in Fig.uJI (ka) ... Fields External to Cones. Consider first the case of an electric current ring. cos (6127) . AP4.. which is to be expected in view of the equivalence of a small magnetic current loop and an electric current element. 1956. ing cone is also messy but can be found in the literature..2rn(n + I)O.SPHERICAL WAVE PUNcrlONS 303 where V is the voltage across the slot. while in the region r > a we have outwardtraveling waves. 0 we obtain tho patterns of Fig.) P '( 9) E. The special case of a magnetic current ring on e. the tip.  Then Eq. 65. (6125)] becomes • jV. IRE TroBl. 9 \ ' . The limit as the magnetic current ring approaches the cone tip gives the field of an axially directed electric current element on FlO. In the limit 80 """. 6018. The general treatment of the probz lem of sources external to a..'(20 + 1)P. that is. while very large spheres produce an almost omnidirectional pattern with severe interference phenomena in the 8 "'" 0 and 8 = r directions.. jV(2n G. I.. In the region r < a we have standing waves. conduct.""(ka) and the radiation field [Eq. From symmetry considerations. Exterior Electromagnetic Boundary Value Problems for Spheres and Cones. Silver. when the conductor is divid~d into hemispheres. no. Patterns for sphcres of radii >. vol.'(cos 9. The general problem of finding the field in terms of arbitrary tangential components of E over a sphere is treated in the literature. it is evident that E will have only a 4> componentj so the field is TE to T. January. 2S1" SID 0 Lt n(n + l)fl. Ring excitation of a conductiog oooe. ~p. 615. . J(C08 8 sin 8 . Figure 617 shows radiation patterns for the case 80 = r/2.
.0.P. we construct F~ 1a. on 8.(eas o)B. .ln(kT) "'" I• b. For the current filament.(cos 8) sin 8 and integrating from 0 to 6.JW1JB r • a '0 P..) a and Eq.n 0 dO (6134) This completes the solution (or an arbitrary 4Kiirected current sheet at r = Q..B.(eo.P._" 21' + 1 am iJ8 iJu 1~.  ~~ J. (6134) reduces to I (6135) .(kT) 1 • [fop. 63 the following orthogonality relationship can be derived: f.B. Thus.(cas 0. (6131) By the methods of Sec. using Eq.""(ka) .N. ao IP. we obtain .(ka) . J. . a'p.(ka)!' v which.(ka) .) (6136) .J.J..(ka) sin 0.. multiplying each side of Eq.nOdO = where N. "'"  . becomes J. = 0 (6l21l) Continuity of E."(:op·)(:op·). O)la.a J f" iJ j.(eo.] .(eas 0) r>a (6128) r<a where the" are ordered solutions to t.ln(ka) = b. .304 TIMEHARMONIC ELECrROllAGNETIC FIELDS Hence. 8 P. (6131) by P.+1)[. Il. P. = .(e<)rl) J .(eas 0)) s.b. O)J. at T = a requires thaf a. J. . k J."" Ira • a '0 P. a~.(. (6130) and the Wronskian of the spherical Bessel functions. at r = a must be discontinuous by an amount equal to the eurfaceeurrcnt density (in our case it is an impulse function).(ka) (6130) Finally.6(0 .. (6132) w= " (6133) Hence.
re ordered solutions to P . J .(ka) At T = a we have E. P. Multiplying each side of Eq. sheet at r .(kr) r<a wbere the u a.P.SPHERICAL WAVE FUNCTIONS 305 Numerical calculations are difficult because of the problem of obtaining the eigenvalues '" and the eigenfunctions P•. we construct \l c..) . (6138) is j. . 618 is a magnetic current. Thus.(cn. analogous 1. _ u(u 2u + 1) [. ll. at T = a requires that Continuity of (6139) c. When the ring source of Fig.P.(ka) sin Ot aO P.0(0 .."(:Op·)(:Op·)·inOdO where M. (6140) by P . wu (6141) (6142) ao au '_'.] +1 l~f. (6143) rednces to J c..d.(cos 0) sin 0 and integrating from 0 to 0 1 on 0.. .2 \' :.].. O)J.) K a (6144) and Eq.0.(ka) (6140) The orthogonality relationship for the eigenvalues defined by Eq.(cos 0. we have M. except for boundary conditions. • c. Henee.u'(kr) A. .0 (6138) in contrast to the v which were solutions to Eq.. For the magnetic current filament. ap.a. M.'''(ka) . we obtain (6143) This completes the solution for an arbitrary ~irected magnetic current.in 0aP.(cos Ot) t K a (6145) .(cos 0) a ./ I • r>a (6137) d.(cos O)J1. (6129). (6131).J1..0 Eq. the problem is dual to the eleetriecurrent case. = '11 M . discontinuous by an amount equal to the surfacecurrent density.
)/QuJ Some radiation patterns for slotted cones with cone angle 30° are shown in Fig.(ka)Il. O.306 TWEHAn~ONIC ELECTROMAGNETIC FiELDB • FlO.'" \ ' j·(2u Jr u(u + l)(. New York. &19. A discussion of the problem of planewave scattering by a cone is given by Mentzer. calculation of the eigenvalues u and the eigenfunctions P. If we now let 82 = 91 and set K = V in the magnetic current solution. For T > a Eq. (6137) becomes A r = : sin! Ih • Using the asymptotic form for we find for the radiation field LA}u p~ L< • (cos 81)P. 1955. R._. is difficult. (6146) E.(l)(kr) 10 8 m and evaluating E. 8193.P.. Inc. (626).) (ilfler Bailin and Again a. Radiation patterns for the slotled conducting cone.. _ V . "Scattering and Diffraction of Radio Waves. by Eq.oJ J.(eo.P. 619.. I I J. Pergamon Press. we have the case of a cone slotted at r = a with a voltage V across the slot." pp. SUVff. (cos 8)J. Mentzer...(cos O)/.(ka) + 1)(. .
is defined by Eq. (626) with A. Let us orient our spherical coorclinatc system so that maximum radiation is in the tJ = 0 direction... The gain of an antenna. ..o(kr)P. 2: a••fl. In practice. it appears that arbitrarily high gain can be obtained. the field can be projected backward toward tbe origin as far as desired. This apparent discrepancy betwoon theory and practice can be resolved if the concepts of cutoff and Q of spherical waves are properly applied.sed to describe antcnnss for which the &Duree (primary or aeeondary) is COl1Iltant in 3mplitudc and phase over a given area on .t n(n + l)j"(l1 alii cos "'Ill  h. At some sphere T . 35).' (&147) Given an arbitrary field at T = Ti. The radially directed power flux in this direction is then (&149) (8. ~.SPHERICAL WAVE PUNCTIONS 307 613.'(cos 8) cos (m~ + P••) ..)..a we can determine sources by the equivalence principle (Sec.H: From Eqs. which will support this field. = E. sin (hI!) • (&150) I The term "uniformly illuminated aperture" ill u. (&147) and (&26) we find E~ = 2jr eik' \' L. (2130) in general.·(cos 8) cas (m~ + a••) 2: b••fl.).'O(kr)P. plane..E. The general form of the field in a spherical spnce external to aU sources is Eqs. however. By the discussion of the preceding paragraph. We shall here consider the largest gain g  4n"(S. is the maximum power density in the radiation zone and ~I is the power radiated. Hence._ ~I (&148) where (8. A uniformly illuminated aperture l type of antenna is found to give the highest practical gain.. regardless of antenna size.H: . and zero e1aewhere.). the gain of a directive antenna is found to be related to its size. it appea. F. Maximum Antenna Gain.rs that sources on an arbitrarily small sphere can support any desired radiation field.
As long as n is unrestricted. and B. The result is ~ _ . =0 Also..'.'. contains only wave functions of order n ~ N.. real.HARMONIC ELECTROBLAONETIC FIELDS in tbe 8 = 0 direction of the radiation zone.308 TU.£E. (6148) involves only the aIR and bhl coefficients.(2n + 1)(n m)! ~ . Equations (6148) to (6151) give a. .... = real (&155) The maximum gain thorefore will be found among those specified by (&156) • where A . so they may be chosen for convenience without loss of generality. Furthermore1 g is symmetrie in A.. and B. (6147). as we anticipated earlier.. hence the maximum exists when A ..I') • where (6153) (&154) The denominator of Eq. """ B . by setting a.1 for m . maximum. (6153) is independent of the phases of A.2 for m > O. this g is unbounded. the denominator caD be decreased without changing the numerator. so we ean maximize the numerator by choosing A..1'+!lb I') . We used the ortbogonality relationships of Eqs.ves. and B..'\' n(n+ I)(n+m)! (I ~/. specified by Eqs... The total radiated power is found by inwgrating tbe Poynting vector over a. Note that the Ilumerator of Eq.. (651) in the derivation of Eq.~ '' . (&151) where f ..=. In particular1 let al. .. We shall now consider under what conditions g is a. +E. Hence. . .. (6151). is real.0 and t. = T and Pl.tor of 9 arc independent of (&152) 0'1" and PI. = b. then an upper limit to g exists.I' + IB. If the field..a. both numerator and denomina.) [' g . large sphere..22: 2n ~ 1 (IA. general formula for gain in terms of spherical wa. = T/2 1 and the gain formula reduces to 12: (A.'.. Setting iJgjaA i = 0 for .
We shall use the Q concept of Sec. a circular. t S. uniformly illuminated aperture of radius a h3S the same gain as the abovedefincd normal gain. iohn Wiley &: Sons.. Hence.SPHERICAL WAVE FUNCTIONS 309 all A i. we find U_  . for large 1M. Antennas having higher gain are a distinct possibility and will be called supergain antennQ. and Efficiency. vol. Bandwidth. F. "Fields and Waves in M. (6151). it is reasonable to assume that modes of order n > ka are not normally ·present to any significant extent in the field oC an antenna of radius a. J." 2d ed. The normal gain is not an absolute upper limit to the gain of an antenna. January. 64 to show that (1) supcrgain a. I.>'11. (6160) I R. p. no. We saw in Sec. Hence.ntennas must necessarily be narrowband devices.e. pp. Effect of A. . A similar limitation to the nearzone gain also exists. ReuGrcA NBS." It is thereCore not surprising that the uniformly illuminated aperture gives the highest antenna gn. 640.. structure and (4) they tend to have excessive power loss in the antenna structure. and (2) supergain techniques yield only a smaU increase in gain over normal gain for large antennas. Inc.in in practice. the normal gain is maximum gain obtainable when only uncutoff modes are present. 64 that spherical modes of order n were rapidly cut 01T when ka < n. New York. l N (2n and also A .odern Radio..$. > W. l To relate gain to antenna si7. Whinnery.. we define the radius a of an antenna M the radius of the smallest sphere that can contain the antenna. w. 533. 1960.N' + 2N 2n + 1 A 3 I (6157) (6158) Equation (6151) represents the highest possible gain using spherical waveguide modes of order n 5" N. We define the normal gain of an antenna oC radius aas (6159) u_••• = (ka)' + 2ka which is obtained by setting N "'"' ka in Eq. Other characteristics which we shall not demonstrate here are (3) supcrgain antennll8 have high field intensities at the antenna.ntenna Siu on Gain. Ramo and J. 1953. Harrington. It is interesting to note that. 112. The Q of a lossfree antenna is defined as '11. = + I) . R.
2: A. is the power radiated. If the Q of an antenna is large. N.. the antenna has broadband potentialities..' (2n ~ I) Q. If the Q is small. it can be interPreted as the reciprocal of the fractional bandwidth of the input impedance. must be less than or equal to the Q of any other lossfree antenna of radius a having the same field r > a. Quality factofft for ideal lossfree anten· nas adjusted for mlUimum gain using modes of order n ::.. Antennas adjusted for maximum gain according to Eq.. Hence. Because of the orthogonality of energy and power in the spherical modes.". 620. 'Vc shall define an ideal lossfree antenna of radius a as ODC having no energy storage r < a.. defined by Eq. .310 Ill' TlllEHARMOXIC ELECTROI. and OW.m . (660) and plotted in Fig. for 'I'M modes slid "XI". The Q. of spherical modes.='=o'='i:=:::t. the Q of our ideal lossfree antenna is Q~ 2: p. The Q of this ideal antenna. since fields r < a can only add to energy storage.'\ o 5 10 \.. for 1'E modes.nd the rJJ is twice that of the TM II mode alone.. the tolal encrgy and power in any field is the sum of the modal energies and powers. When QII < I..LAGNETIC FIELDS 10' Ill' 0' e 0 30 25 \ 20 Ill' 10 10 i~5 \ \ 25 FlO. and it is convenient to deal with Q's for equal TM and TE modes. are the timeaverage electric and magnetic energies Bnd {j>. The Q for equal TM. I \'. 2: p 22:A"(2n~l) .Q. we take it as unity. involve OW.. mode alone a. is essentially that of the TM. 66. (5158) have equal excitation of 'I'M and TE modes. \ \ 1\ 15 20 ka where 'W. and TE" modes is ka < N (6161) because the 'W. We need Q's defined in terms of the same energy for aU modes..
1 N 1) Q. modes. work.~ ~ 8 7 6 " Ii ~ g ~ 5 . Ma... The problems of narrow bandwidth and high losses associated with small antennas are wellknown in practical antenna. of radius a. 20 " 0 10 30 40 50 60 70 80 90 Fro. Figure 621 shows the ratio of this upper bound to the normal gain. Note that for large ka the increase in gain over normal gain possihle by supergain techniques is small. are given in Fig.. All very small antennas are supergain antennas by our definition. thie becomes Using Q . lO'" '. is the transmitted power in the TM..!! 0 ~ . 66.. The Q of Fig. '" Q = 10· 2 1 . (6158).ximum poasible iocrea. we can calculate an upper bound to the gain of an antenna of radius a.. ns ka + 0 the supergain condition is unavoidable..+.. l (2n + ..'\ 3 .SPHERICAL WAVE FUNCTIONS 311 where P.·.(ka) (6162) where the Q.::::2"'N""·. Note that the Q rises sharply for ka < N. . showing that supergain antennas must necessarily be high Q.. Curves of antenna Q for several N nre shown in Fig.. 4 3 \" ~l~ '.. Eq. or frequency sensitive. For small ka supergain can give considerable improvement over normal gain. 10 9 II .4"N. By picking a Q...ee in gain over normal gain tor a given Q. 620 is a lower bound to the Q of any lossfree antenna.. 621. and TE. In fact. 620.
.')] . .8 iJ ") .ylUn • + Jw€1'aD [COS.  ~ 1 [ sin .".744)' "'I <ol. E.sm' 6) ] :.y.jw6' sin 8 iJ(/I . (Q. 63.).b. .(>/Sin 8 r Sin 8 iJ(J 1 a [COS // B...." "" 2m T& "U A" where the u. H.744/b. _ 1(2. r r (I . 66.l~ • [COBt 8 .[.. (3·85) p.1fsJJJy.J. In the concentric6phcre cavity of Prob.. 65 let a «b. _ 2..a. . _ cot 8 ay. Consider the cavity lying between concentric conducting spheres r "" a and r .744) J" n. cos (I 1 a ..nd t.. For the spherical cavity of Fig. _! a".) 1 1 a .I. for TE modes. 1 a . {Hint: El'press the characteristic equation in the form :a 0. with b > a. r sin 8 a8 (y.. Show that the characteristic equation for modes TM to r is And for modes TE to r it is flo(kb) Jo(kb) Jo(. %(2. E . 1 H. 62.i} .) . 64. and. (614) to show that a general expression for fields TM to t: is . COB 9) ] where'" is a solution to the scalar Helmholtz equation.Ur ..iJ9 (V.. JfsJj. san' 8) a .) 66.(r". (J : . Consider an airfilled ..(rti) laln... frequencies and the Q of the dominant mode. show that the Q due to conductor losses is.a) where <ol.he wave potential of Eq. (617) and (618) are identical.j . ." are given in Table 61. + 1<. Verify that Eqa. Use Eqs..312 TIMEHARMONIC ELECTROMAGNETIC FIELDS PROBLEMS 61..iJr . 62. Determine the first ten resonant. and el'pand in s.fI. for TM modes.1 iJr a ar (r:l. .A"  ' In . (rof) + aag (".(. TO (Q').:t R. . u." + 1)] where the u~" are given in Table 62."" v.". and show that the resonant frequency w is related to the empty cavity resonant frequency w.(2. .744) (a)' b f(k...spherical resonator of radius 5 centimeters bounded by copper walls. 'l'aylor series about kl . by w  . UA " n(n _Ut .
. I. 1'1 for r < a and a dielectric III 1'1 for a < r < b. 610. Show that the field of an electric current elemcnt II is the dominnot TM spherical mode of space. I C 2AB COB 611..ce field of a zdirected magnctie. Show that the characteristic equation for the dominant mode is h:(k.. '"' ..b)J I (k.) { : • B and determine the coefficients 4.. Show also that the change from primarily resistive to primarily reactive wave impedances occurs at kr .a) n:(k.. Consider the characteristic impedances of the spherical modC$ of f:Ipa. 67c and d) ill given by .744) . Compare this with the answer to Prob. where 6<l. In the partially filled spherical cavity of Prob. 2" I0 0 . (648). Taylor series about the empj.. and b•• for the twodimensional FouricrLegendre scriea of the form of Eg.current element.. Consider the partially Iilled spherical cavity formed by 8. By expanding the characteristic equation in 8.B and show that. .. Show that the dominant spherical TE mode of the wedge guide (Fig. n. show that the resonant frequency"..J~(ktb)n. 612.744) + 2 b f.744)' ~~(2. (657)1. let 11« b and II .1 (~)' 11 (2..Ycavity resonant frequency ""0.. for B > A.1'1.cot 01 /2 6101.<01 ~ and k l . show that the attenuation COtl8tant due to conductor losses for the TEM mode of the biconical or coaxial radial guide (Fig. 67e) is the freerspa. 68.744/b V. 69. '11/ZTM. and the field of a magncticcurrcnt clement Kl is the dominant TE mode.a) .(k.SPHERICAL WAVE FUNCI'IONS 313 8'1'. 61S.a) J..lR cae 01 + esc 8.. 67..la and III .ce IEqs. Define C _ A . Let A and B be two vectors and B be the angle between the1ll..Bal2 cot . conductor covering r _ b and containing: a dielectric II.b)J~(k. Using the usual perturbational method.(k.<01 V.d•• and <010  2. Consider the function [('. Bhow that and zrE ..b)n I(ktll) where k l . is given by ~ WI ItS 3i(2.
(Ol'o'lC ELECTROMAGNETIC J"IELDS 61fi. 61'1..' +i' t~ J. Conaidcr the lleCOnd rCllOoance In . 616. is the pth I:cro of J:(Ul) and II is a solution to Eq.. in the limit. 819. mode.2 in Eq.horn guide of Fig. (~2). frequeocie.(I)(b')P1(eoa. radiWi for which the Cf'OlllI aection is t.. ().. k'Kla 4... mately equal to Uat. to iustify the statement...) .be angle between rand 621. ll. Coll8idcr the conical cavity of Fig..:(C04 I) COl m<ll J.el"8.. II . __ 0 (2" Irr. 1" ellrrJ 1 l• + 1). 820. &Sa) constructed of copper with a ."'(coe 8) coe "'«IJ.P .A. elemcntll of Fig... variation mUllt allO be included. "here f is t. Derive tho following wave transformat. the field is given by B . G9c be replaced by magneticcurrent. show where 10:. For a complete lIet of modes tho sin m'.8d. (4"1 2r..__ 0 A. .. 6·21.A.}.·(eoa. Derive the following wave transformation: J.(m + 2ft + 1}(2m}1 + .(cos U r. (r}P. v X u.HAIU. Caleulate the Q of the mode and the input resistance seen at t.. that modes TE to r are given by E ...~ .he \lame aa II.j 1a.!.he CODe tipa. and airfilled. Considtr a hemispherical eavity (Fig... 622 where tric current.ion: ~ . 6&. (Ul•• /i) where ia the pth zero of J. Uae the qualitative behavior of the tpbtrieal Hankel funetioJ\l..v X uJ.10 eentimet.. .jr.... rect.)1"11 . 622.. (10:... the ficld it! given by H .hat the apherieal. y the quadrupole tIOW'Ce of each element is an elecShow that. elements Kl. and the Q of the dominant. 618.. t... cutoff.(P) .1 is a 101utiofl to Eq. in the limit 0.anlUlar guide at. tho eumlDt..p.p. (664). where P.lfJ(r)P.) • L wh"" A • _ (1)·.. (680)! of the biconieal cavity of Fig. 0. Consider Fir.314 TUn. z T ...v X u. Show that modes TM to r arc given by H _ v X u. Show that.. 67} has a "cutoff radiw" approxi.. ~) Similarly. where (A. .' ..} to.. Determine the fint ten resonant.(tlJ) aod 1. Let. '\"'" 0 and " . where (F. where x FIG. A quadrupole source.(r')B.
226.. Consider a loop of uniform current I of radius 0. 611)... .A. is given by Eq. Consider a radially directed magnetic dipole Kl adjacent to a conducting sphere (Fig. FlO. and t • where a. (626).. as shown in Fig. (698). .ered field is plane pol/lri1ed in t.. Con8ider a radially dircc:tcd eleetric dipole adjacent to a dielectric sphere (Fig. . A plane wave incident on a and obtain exprcllaionll for b. G14 with n replaced by Kt). e jlrt 2: .. Show thnt the radiation field is given by B. y e. . 614 with the sphere DOW dielectric).b... d.polariled. by Eq. 2:<:' ~ l .. • A. (698) and c.IG~(k(1)J. .J~(ka)R_(kr)JP. 623. and coaled conducting sphere.. llCatt.'(COlI B) p. o where for r > b the A.'looo') ~ I W<lve Impose boundary conditions 00 the tanzential componcnts of E at H at r . Show that the distant. and F. z traveling plane wave incident on a conducting sphere encased in llo concentric dielectric coating. "'I" COli <It \ ' . Incident .I.)G. Considcr an z. Derive the following formula: r' >r where ~ is the aogle between rand r'. IUl shown in Fig.. (6102). (GlOl). (6102).  i. c. 628. . 626.(u) ~·!in • . in terma of a. 623.lb)[P. 626. eo.. 6·24.J . • d. I x '.Cm) "H. is given by Eq.he direction 8 _ 60'. givcn by Eq. 627.(ka)J.i.P_ ' (O)P. Show that the radiation field ~ given by E.l . _ E. < b z . . and for (1 < .. CoIlBider the scattering of a planepolariled wave by a smrill conducting sphere (Fig.[G. Show that the radiation field i8 then given by Eq.(u) .8•.fill.J.SPHERICAL WAVE FUNCfIONS 315 623. and A.ICCOlI B) U wbu. are given by Eqs. (6117) if b. Show that the 6eld ill given by Eqll. (6113) instead of Eq.
O.7 77' 0. &5 and the definition of Eq.''''(ka) _ J.he equivalent circuit of Fig. the minimum possible Q for a smalll088free antc.(ka) 8. . 624.9 90' 1. (660) for Q.2 24' 0. A small antenna (say ia < 1) will have minimum Q if only the n . 625.3 37' 0. A conducting sphere with a concentric ring of electric current. 630. By considering t. Figure 624 shows & conducting sphere of radius R concentric with a loop of uniform current 1 of radius (I.I(r) sin 8 ~(C08 8) where u is the first root of P .(ka) .5 60' 0. show that the Q of the n . 628 aa R . this reduces to the answer for Prob.316 TWEHARMONIC ELECI'nOMAGNETIC FIELDS 629.6 69' 0.1 modes are present in ita field. Figure 625 shows a current element II at tbe tip of a conducting cone.0 631.. FlO.(ka) .nnll is where a is the radius of the smallest. Show that the radiation field is of the SlUllO form as given in Prob.l mode is If equal TE and TM waves are present.(kR)R. Current element at the tip of a conducting cone.(ka) Show that.")(ka) J. Hence. 628 except that .1 epherica.(kR)J. r '.(kR)R.l '0' 0. the total Q is approximately onehalf this value.O.(C08 81) . ephere that can contain the antenna.1.(kR)J. Show that. z z e r y x FIa.8 84' 0.R.' _ 8. the radiation field is given by E. .R.4 49' 0. Some approximate eigcnvalue8 are u I"I I I I I I I I I O.
we shall consider two techniques uscful for integral equations arising in electromagnetic theory. In Cact. The perturbational methods arc useful {or calcula. iC an assumed field is expressed as a series oC functions with undetermined coefficients. at least in principle. if a complete set of Cunctions is used for the assumed field. Figure 71a represents a resonant cavity formed by a conductor covering S and enclosing the lossfree region T. The variational methods are useful for determining characteristic quantities. and the It perturbed" problem. 311 that electromagnetic field problems caD be expressed in integral equation form. 76). In contrast to the perturbational procedure. The word "perturb" means to disturb or to change slightly. In this chapter. We have already used perturbational methods for calculating resonator quality factors and waveguide attenuation constants. Further uses are given in Sees. This means that the formula is relatively insensitive to variations in an assumed field about the correct field.ting changes in some quantity due to small changes in tbe problem. Perturbational Methods. This form is particularly useful for (1) obtaining approximate solutions and (2) for general expositions of theory. Usually two problems are involved: the "unperturbed" problem. rather than to changcs in the quantity. The differential equation approach of the preceding three chapters leads to an exact solution of the mathematical problem. the exact solution can sometimes be obtained. Introduction. then the coefficients can be adjusted by the Ritz procedure (Sec. 72 to 74. many problems cannot be treated by this method. Perturbations of Cavity Walls.CltAPTER 7 PERTURBATIONAL AND VARIATIONAL TECHNIQUES 71. impedances. such 85 resonant frequencies. Variational Methods. Figure 71b represents a deformation or the original cavity 317 . which is slightly different from the unperturbed one. the variational Cormula may be an upper or lower bound to the quantity. for which the solution is known. The variational procedure differs from other approximation methods in that the formuJa is Ustationary" about thc correct solution. Furthermore. and SO on. If the dcsired quantity is real. However. We saw in Sec. the variational procedure gives an approximation to the desired quantity itself. 72.
In both cases the field equations must be satisfied.lHS· H By analogous operations on the second and third of Eqs.AT. The resulting equation is t H X Et· ds = j(w . H (b) (a) Original cavity. 71. H.jwotE: . we have c/fHXEt. (b) perturbed cavity. (H: X E) = jwp. H: . The divergence theorem is applied to the lefthand terms. We wish to determine the change in the resonant frequency due to the change of the cavity wall. w represent the corresponding quantities of the perturbed cavity. FIG. . since n X Eo = 0 on S. Pert.H . E These last two equations arc now added. V X B v· (H X Et) = jWfOE· ES  jWaJ.wo) JJJ (fOE· Et + p. such that the conductor covers 8' = S .318 TIMEHAR~{oNIC ELECTRO~[AGNETJC FIELDS D D s S' Eo. H o.V X Eo = jWoJlH o V X HI) = jWflEE o v X E =jwlJH vXH=jwfE (71) We sealarly multiply the last equation by E~ and the conjugate of the first equation by H.lJ.S and encloses r' = 'T . and let E. H o (a) E. (71).A. that is.dS=O .H . one of which vanishes. and the sum integrated throughout the volume of the perturbed cavity. because n X E = 0 on 8'. HS) dT (72) Finally. The resulting two equations aTC Et·V X H =jwEE·Et H· v X Et = jwo.lHci· H Adding these and applying the identity V • (A X B) we have ~ B •V X A .urbation of cavity walls. WI) represent the field and resonant frequency of the original cavity. Let Eo. we obtain V .l.
that is. are timeaverage electric and magnetic energies is the total energy stored in the original originally contained in AT and . while the denominator is proportional to the total energy stored. to conform to the convention that ds points outward.I') dT III (_lll. H in the denominator. ds "'" = 1f> .111 •• (H o X E~) ._lll. ds (.IE. and also substituting Eo. Note that our development assumes that E and}J are real.. With this approximation the integral in the numerator of Eq. Hence. (162)].. Problem 71 gives the gcneral formulation in the lossy case. H o. (73)..IE.IE..I' .s (73) This is an exact formula for the cha. The crudest approximation to be made in Eq. (73) becomes effi . (72) as ieffiH X Et·ds .I' . H by the unperturbed field Eo.nge in resonant frequency due to an inward perturbation of the cavity walls. H X E~ .·ds = 1ftH X E.ds d The last term is taken as negative. (74) can be written as (75) where tlv>.I' + . we have WWDR<~" w. III WI. iw.v .I') dT The last equality follows from the conservation of complex power [Eq.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 319 and the lefthand side of Eq. Substituting this into Eq.I') dT • (74) Note that the terms in the numerator are proportional to the electric and magnetic energies II removed" by the perturbation. H o for E. For small perturbations this is certainly reasonable in the denominator and should be valid in the numerator if the deformation is shallow and smooth.. We can now rcwrite Eq. wc have assumed no losses. (72) cau be written as 1ft r H X E:·ds = ss 1ft H X E. and a'W. Eq. (73) is that of replacing E.
T If the perturba. and 6oW. (298).tion occurs at the midpoint of the longer side wall (maximum H).. (75) we find ~ :lIE.. (tV. Furthermore.Wo ::::: 2 tl:r w. spoceaverage energy density Ul.320 cavity. = 0.I' + c'/b') 60T 2117 (78) ::::: 1 + (c/bp7 Note that for a squarebase cavity (b = c) the cbange in resonant frequency due to 117 at maximum H is only onehalf as great (and in the opposite direction) as that due to a7 at maximum E. position of maximum E and zero H J or vice versa. we have 11"1». and will lower the resonant frequency if it is made at a point of large E (high We)... because we calculated W when we determined the Q'5.. Thus.I' 60T (77)  w . from Eq. It is evident Crom the preceding equations that an inward perturbation will raise the resonant frequency if it is made at a point of large H (high U>. (74) to (76) are easy for the cavities treated previously. Numerical calculations using Eqs. (75) can be written as W 000 000 . The opposite behavior results from an outward pertur batiOn.I'T 4  For tJ..vity of Fig. W is given by Eq. TIMErfiARMONIC ELECTROl\lAQNETIC FIELDS Finally. W caD be written as T times a.T located at the midpoint of the base (maximum E) we usc Eqs. . ~ Hence. (296) to find aW.. For the dominant mode of the rectangular ca. or 'II> ~ IE.). from Eq. Eq. we caD approximate the AW'S by aT times the energy densities at the position of liT.. It is also evident that the greatest changes in resonant frequency will occur when the perturbation is at a. = 0 and 6oW. (75) we fInd wwo Wo 2(1 IE.) Wr AT = C AT T (H) where C depends only on the cavity geometry and the position of the perturbation. Hence. 219. = 10. if AT is of small extent.
O. Figure 72a represents the original cavity containing matter E.680 Hemispherical ':U (a) 2. In general.L AND VARlATIONAL TECHNIQUES 321 'fABLE 71... Let us now investigate the change in the resonant frequency of a cavity due to a perturbation of the material within the cavity. .'~jI ~(b) (a) 2 tb!'" b) ( 1 + (c/b)l 2 Short cyliDder (d < Za) (a) 1.. IJ + All. located at (a) maximum E and (b) maximum H.02 (b) 0. Figure 72b represents the same cavity but with the matter changed to E + AE . (73).PERTURDATIONA. TUE PAB.843 (b) 1 + (1. / (4) 0.80 Spherical tE"(b) 'a. shallow deformations.714/d)' 2. H by Ee.' Long cylinder (d ~ Za) d~ "(a> (a) 0.3Gl (b> 0.. These values have been obtained using the crude approximations of replacing E. They are therefore valid only for smooth.L /" (a) /' 1A1 ~(. IJ. (76) FOR DEFOnMATIONS (a) AT MAXJllU1ll H 011' THE DOMiNANT MODE Cavity Geometry C Rectangular (a S b . Cavitymaterial Perturbations. (76) for cavities of several geometries for a. the frequency shift depends on the shape of the deformation as well as on the shape of the cavity. H o in Eq. The formulas for deformations of the form of small spheres or small cylinders caD be obtained from the results of the next section by letting E _ 00 and 11.' (a> .680 Table 71 gives the value of C in Eqs.A)lETER C OP E AND (b) AT MAXlloIU){ EQ.85 (b) 0.5 c) I (O!>L. 73.
y. Ht) d. (79) gives CD: X E) = jwCp + dp)R· H: . that is. this can be rearranged as .pB: . we can approximale E. w by Eo.jW(ltE: . E The sum of the preceding two equations is integrated throughout the cavity. (oj represent the corresponding quantities of the perturbed cavity. as ~t + 0 and IIp + 0.322 TDrERAIUIONIC ELECTROMAGNETIC FIELDS n n '.. and add the resulting two equations. H Analogous operation on the second and third of Eqs. 75.w. Perturbatwn of mat. X E:) = jW(E + AE)E • E: .. This gives v .. The lefthand terms then vanish. due to a change in t and/or p within a cavity. that is. E: + ~pH· Hn dT + pH .') dT III (·IE. (H V. E and p are real. + 6. (710) This is an exact formula for the change in resonant frequency.III III (llEE· Et (tE. In the limit.. Let E Ot HOI Wo represent the field and resonant frequency of the original cavity.V X Eo = jwopHo V X H o "'" jwotE o . and let E.l') dT (711) .1p)H (79) AB in the preceding section. (a) s (6) Flo.Wo w w"jE· E: + [wu. + 6.j(J}. The general formulation when losses are present is given in Prob. we 8calarly multiply tbe last equation by E: and the conjugate of the first equation by H. Within S the field equations apply.' + .y. The result is 0= III IIw(. .6.I' + 6.V X E = jw(p V X H = jW(E + . and the divergence theorem is applied to the lefthand terms. H. (4) Original cavity.E)E + . because both n X E = 0 on Sand n X Eo = 0 on S. Wo and obtain III (6·IE. H o. 72.) w . .IH. Once again our development bas assumed the lossfree case.IH. H.) .ter in a cavit.. (b) per· turbed cavit.IH· H:I dT Finally.
PERTURBATIONAL AND VARL\TIONAL TECHNIQUES 323 This slates that any small increau in E and/or ~ can only decrease the resonant frequency.emy. change the resonant frequency. while for a material perturbation at (b) of Table 71 we have C 1 = 0 and C! = C. For the thin slab wit.he various terms of Eq.r.he normal com .red to wavelengt. the Helmholtz equation can be approximated by Laplace's equation.. These are shown in Fig. (713) wit. it is merely necessary to replace E by Hand E by p. The procedure is justifiable. Now if the change in E and ~ occupies only a small region AT. small change in E at a point of zero E or a small change in p. Any large change in E .)d 10. we can further approximato Eq.T. Henco. C . (712) by (713) where tb is the space average of W. in which case C! = 0..and/or ~ can be considered as n succession of mnny small changes. If we compare Eq. (711) as energy expressions and rewrite it as w w. The preceding approximations require that AE. it is evident that. The parameters C1 and C! depend only on the cavity geometry and the position of aT.+ d" .h E normal to it (Fig. We shall now cODsider a procedure for removing these restrictions on AE and Ap.urbat.ional solution is very simply accomplished. 73a). because. and AT all be small. 73 for the dielectric case. for a material perturbation at (a) of Table 71 we have C1 "'" C and C! = 0. quasistatic approximation to the field internal to tl. (76). The modification is accomplished by using a. or at a point of . There are Cour types oC samples for which this quasi~static modification to the pert.h. For the cases considered. For the magnetic case. This introduces the further complication that the change in frequency depends on the shape of tl.h Eq.C! . To be explicit. as well as on its location. at a point of zero IJ does not.. ::::s  w. we must have continuity of t.tero E. '" 1 fl! (d' . E "'_ ~ T (712) where W is the total energy contained in the original cavity. Ap. Note that 0. We can recognize t.CJ.in Table 71. This assumes that the ficld internal to AT is related to the field external to AT in the same manner as for static fields. any imrease in f and/or ~ within a cavity can only decroou the reMnant frequ. aT' is either at a point of zero H. in which case CI = O. in a region small compa.
l0NIC ELECTROMAGNETIC FIELDS O'_.:. of the preceding equations. . 1950. Some small dielectric objects {or which the quasi. 73. New York. 6768. Stratton. because the contribution from AT is small compared to that from the rest of T.JD (cj (d) FlO. A. we can use the static solution. "Electromagnetic Theory. McGrawHili Book Com. (710) by E 1D . New York. 73c). "Static and Dynamic Electricity. 1941. Smyth.static solutions are simple. for E normal to a small sphere (Fig. I J. I which is (716) Finally. ponent of D. we can use the static solution. we must have continuity of the tangential component of E. so that (715) Again this approximation is independent of the crosssectional shape of the cylinder. For the long thin cylinder with E tangential to it (Fig. so that (714) This approximation is valid regardless of the crosssectional shape of the cylinder. Hence. 7311).324 TIMEHARr. 205213..' To use the above quasistatic approximations. Inc. 73d). Inc. McGrawHill Book Company.I'dT (718) I W. pany.IE.. R. we approximate E (and H in the magnetic case) in the numerator of Eq." pp. For E normal to a long thin circular cyliuder (Fig. our quasistatic correction to the perturbational formula is (oj  WD ~ JII AtE IDI • Eci dT 2jjj.' which is (717) The static solution for a dielectric ellipsoid In a uniform field is also known but is not very simple in form. In the denominator we can stilJ use the approximations E = Eo and H = H o." pp.
However. lE.nd we treat it to the same degree of approxima. to 'W•. These are illustrated in Fig. for the case t1p. of course. For a dielectric slab on the base of a rectangular cavity (Fig. 744).o illustrate t. but with E replaced by H and E by p. let us apply Eq.o and EI = EO shows that our answer is identical to the fU'St term of the expansion for cal in powers of dla. (711). CAvities used t. (715) should apply. given by Eq. 1 dW=a. so Eq. we again get the correct first term of the expansion.. (718) then yields ..tion (match tangential H). 74.l = P. Equation (718) is.j (b) (e) (a) FIo. 417 for P. (718) and (7Jl) give . 74. (The denominator bas been simplified by equating W.. Note that Eqs.I "'" p.. 28. that is. "'" O. I.ld (719) where d is the slab thickness and a is the cavity height. Application of Eq.71 . The field and energy expressions for the unperturbed cavity are given in Sec. (718) to problems for which we have the exact solution. ~n fact.2. 74b) has but little effect on the resonant frequency. because E is zero at the wall. throughout. which is I) ~ o Er F:S It is apparent that the above formula is accurate only {or when tu is small.. To illustrate the improvcment obtained by using the quasistatic field. is also nonzero a.PERTURBATIOYAL AND VARIATIOXAL TECHNIQUES 325 A . (719) to the result obtained from Eq. (714). most valuable for problems for which the exact solution is not known. if tip. A comparison of this with the result of Prob. we can compare Eq. A nonmagnetic dielectric slab at a. side wall of the rectangular cavity (Fig. In this case E is tangential to the airdielectric interrace.be pcrturbatKlnal fonnulas. so that we may gain confidence in the results as well as pr:lCuce in the procedure. . ) The corresponding formula for the frequency shift due to a magnetic material would be of same form. we have E l .
74.ions of eylindrieaJ waveguides. 75b represents a wall perturbation. that is. _ 3' 1) (~)' • (720) A comparison of this with the answer to Prob. . fOr 1 (2. This we can compare to the exact solution (Peob.291 where a is the radius oC the small dielectric sphere and b is the radius of the conductor.(t~l)f.Bo E.ve the correct first term of the expansion when tip. =~JI(2. n n • E. "'" O. of course.wo In particular. ". consider the spherical cavity with a concentric dielectric sphere (Fig. H Eo.744&)Sin9 Applying Eq. (718).. we obtain 0 a Wo ". As a final example.326 TIMEHARMONIC ELECTROMAGXE'I'lC FIELDS identical approximations in this case. 740). (635). Fig. Figure 75a represents a cross section of the unperturbed wavegUide. all z: = constant cross sections are identical. The guide boundary is taken as perfectly conducting in aU cases. (c) material perturbation.+2 b and tbe stored energy is given by Eq. . 75c repre. All perturbations must. (0) S C (b) S . (b) wall per~urbalion.c C' (. This sbift in resonant frequency caused by the introduction of & dielectric sample into a resonant cavity can be used to measure the constitutive parameters of matter. 75. and Fig. we obtain w "'0 Wo === 0. The field of the unperturbed cavity is defined by H. which is the same. The perturbational method used in conjunction with the quasistatic approximation gives excellent a. 418 shows that we again ha. be independent of z.744 ~)' t. <a) Original cross section.t'll'"Xd sm :t: a ~ _ T' (. Perturbat.J.j c Flo. (717). 6·8).sents a material perturbation.H '. Waveguide Perturbations. .ecurncy when properly used. using the quasistatic Eq. We shall now consider waveguides cylindrical in the general sense.
which reduces to the length of the segment of thecylindrical waveguide. shallow perturbations.PERTURBATIONAL AND VARIATtoNA. The crude approximation of replacing the perturbed fields E.[') _ d. we can obtain a better approximation to .[' . the lefthand side results from the integral 1P(H X E: +H: X E)·ds taken over the perturbed surface. For example. an inward perturbation of the waveguide walls at a position of high E will lower the cutoff frequency.u (721) where 6C is the contour about the volume of the perturbation and S' is the cross section of the perturbed waveguide (see Fig.c ~ w.. This leads to  6"." formed from the rectangular waveguide by . we can apply the cavity derivations directly to the region formed by the cyundrical waveguide bounded by two z "'"' constant planes. from Eq. If (. (74).[E..{. H by the unperturbed fields Eo. changing only some of the explanations. This leaves only the surface integral on the lefthand side of Eq. the ficlcle are independent of Zj 80 the surface integrals over the two z = constant cross sections cancel each other. (73) applies directly for calculating the change in waveguide cutoff frequency.'Y".. II (.. (721).IH. For lL length of a cylindrical waveguide at cutoff. we find that Eq.7 8 (722) which is analogous to Eq. H X E:·ndl _"l._.c "'" fJ B (oE • E: + .6. An example of the perturbation of waveguide wallE is the "ridge waveguide.IE. Hence. Following the derivation further. 75b). We should therefore expect formulas similar to those for perturbations of cavities to apply to waveguides at cutoff. while one at a position of high H will raise the cutoff frequency. (72).L TECHNIQUES 327 At the cutoff frequency a cylindrical waveguide is a twodimensional resonator.wc by using a quasistatic approximation for H in the nwner· ator of Eq. H o in Eq. For perturbations not shallow and smooth.!'). (73) we obtain the change in cutoff frequency 6wc due to an inward perturbation of the waveguide wall as j. (72) taken over the wall of the waveguide.H· H:> .IH. (721) gives good results for smooth. Hence. But both numerator and denominator involve an integration with respect to z.c'__'_ 6". in deriving Eq. In fact.u 11~~.['+.
analogous to Eq.328 TIM~HARMONIC ELECTROAlAGNETIC FIELD8 adding ridges along the center of the top and bottom walls. The ridges also decrease the characteristic impedance of the guide. (718) we bave in the nonmagnetic case ~ A"" !:J. If at' and OjJ are large.I' d.. Note that an increase in either f or ~ can only decrease the cutoff frequency of a waveguide.E and !:J. (711). 783788. B. we can replace E.I' + "IH. from Eq.H. It is A". IRE. AB long as the perturbed guide is homogeneous in f and P. 73. (:')' (726) w 1 .. (714) to (716).o..)' < CAl.. pp.IE. we can determine the propagation constant at any frequency from the cutoff frequency according to ~ IS.I') d.· "" ff .la. H:> d.. hence. Propertiea of Ridge Waveguide.k. For example. Hence.a. H by Eo. (710) we can obtain tbe exact formula for the change in cutoff frequency due to a change of matter with the waveguide. 35.. Cohn. are small. if (alE· E: + d~' H:> dB ff (oE· E: + . "'. l j~ jk~la .IE. J f 2 (723) where the integrals are taken over the guide cross section. August.I') d' ff (. no. they arc used for impedance matching. (724) This is analogous to Eq. H o and obtain ~ A". 1 Such ridges will lower the cutoff frequency of the dominant mode and will raise the cutoff frequency of the next higher mode (sce Prob.I' + A"IH. S. If. ff (A. "'. 1947.IE. The formulas for material perturbations in cavities can also be specialized to the case of material perturbations in waveguides at cutoff. . a greater fange of singlemode operation caD be obtained. Hence. PNX. E: dB (725) where Elu is given by the appropriate one of Eqs. vol. we caD improve our approximation by using the quasistatic method of Sec.. but of small spatial extent. The reasoning is essentially the same as that used for the wallperturbation case. 712)..(..
(H X E: + H: X E) ~ jw(AEE • E: + ApR· H:) Integrating this equation throughout So region and applying the divergence theorem to the lefthand term. d8 (730) in the case of a material perturbation. we express the fields according to Eqs. 7·6.. j). u.:.l slice of Fig. The unperturbed and perturbed fields satisfy Eqs. In the lossfree case we can express the unperturbed fields as Eo "'" to(z. 76.j). consider a material perturbation.: X 1l:) • n dl = ) . 1l::> d.) If the perturbed guide is inhomogeneous.:. Il are changed to E + At and p + Ap.in 1ft (H X E: + H: X E) ·ds = jw IfI (AEE·E: +ApH·H:)dT (731) This is an identity for any two fields of the same frequency in a region for which E and .y)ei~a (727) (728) The perturbational formulas are then QQ' ~  . In such cases we can obtain perturbational formulas for the change in 'Y.: X 1l: + j)._::r f! s • (j). (731) to FlO. and ~  ff (6. (727) and (728) and apply Eq. For material perturbations in a cylindrical wav&.dzI guide.s~::. and. no such simple relationship exists.0 j. The perturbational formulas in the lossy case are given in Probs.• . . ~. X ll::) ..r. their sum becomes V .y)ei~•• and the perturbed fields as E = t(x.rential slice or a the differentia.. u.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 329 (This is proved in Sec. (79) with w = wo. On the cylinde.: X 1l: + j). d8 (729) in the case of a wall perturbation..: + 6.. 715 and 716. for the frequency is kept unchanged..y)ei6a H = :a(x. with Wo = w.ll: .. ~ w T.. 81. DilTe. _ X 1l::> .y)ei~" K o = a:o(x. (79) are still valid.''Y=6C"::.. The two equations following Eqs. we obta. (j).: ff (j). To illustrate the derivation of the above formulas..
Using the conservation of complex power (Eq. Equations (729) and (730) as they stand are exact. we can obtain better results by using the quasistatic approximation for the fields within .o..o. (729). II (li: ao+ li. o in Eq. hence Eq.: . In the derivation of Eq. To usc them.: .: . so this part of the surface integral vanishes. an· ds • X X (733) For large .:. in the case of shallow. we have the approximation for material perturbations fJ . " ' ..0.... (730). we arrive at the result a [f (P[O. II (6·12.a:) d• p • Rearrangement of this equation gives Eq. ft by to... since the thickness of the slice is a differential distance.I' .330 TIMEHARMONIC ELECTROMAGNETIC FIELDS waveguide walls both n X E and n X Eo vanish. (731) is zero.o. (731) reduces to j(ji  ~. . As an example of the perturbational approo.p.. smooth deformations of waveguide walls.) II (a x ~: + a: x E). for small . For example.1') ds fJ .) If the perturbation is not shallow and smooth.:. just as we did in the cavity problems of Sees.e and . (729). we can approximate t.1' + 6pl0.jw II (6&:' li: + 6 a . consider a circular waveguide of radius b containing a concentric dielectric rod of radius a.o. u.o. Similarly.[') d. better results can be obtained using a quasistatic modification.p. (731) caD be expressed as the integral over the cross section t. The rightband side of Eq.II (li: ao+ li..fJ.. 1l::) .E and . 73 and 74. == Cal " .fJo """ Cal ". (162)]. d• (732) • X X (The denominator is twice the timeaverage power flow in the unperturbed guide.. The exact solution to this problem was . u. we must make various approximations for :2 and H. the righthand side of Eq. and the lefthand side equated to zero leads to the desired result.[2. Also. d• • .imes dz.o.E and .p.'''.ch applied to a waveguide problem.
• b ~~/ / / . ko "'" vi 2.7892 ~ 1J where We is the cutoff frequency. 75. For the perturbational SoluLion we shall use Eq. = O. PP.9  ~ 0./' L?'Perturbation solution o 0..'v X E ~IEE = 0 V X r v 1 X H  ". 511. (532)]. The {<wave equations" arc v X p./w)1 1 (734) Figure 77 compares this solution to the exact solution of Fig." 1.44 0. 511.2 03 o/b considered in Sec.8 0.841 £) sin rP H. J I ( 1. Comparison of the perturbational 801u· tion ""iUt the eu.)' (w. (734) may be taken as that of the perturbed guide. the We in Eq.146 •. 77. OUf approximations give good results for small a/b. = ~: The denominator E.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 331 1.:: ~~ EO( 1. (730) reduces to iY E. = H.841 and Eq.' d41 0 0 dp p(E.ct lJOoo lution for the partially filled circular waveguide.l + E. + b 1(. In thc numerator we make the quasistatic approximation of Eq. The numerator is easily evaluated as .0 FJO.~l~ = 0 (735) . b .. 55. a.nd a numerical example is shown in Fig. Suppose we have a resonant cavity formed by a perfect conductor enclosing a dielectric. .7 0. I _ lOc. and in the denominator we approximate t.:U J~ (1. (730) with tip.6 @ . (716)..l) = 0.841~) cOS ¢ ~: where Zo is the characteristic impedance [Eq.. = l.1 0. by ~lll 11 0 • The unperturbed field of the dominant TEll mode for the circular waveguide is :a E. At frequencies near the unperturbed cutoff frequency.0. of Eq. (730) tben becomes 2 Zo 1. possibly inhomogeneous. Stationary Formulas for Cavities.
21lPwt (741) By definition. (740). but. 130. (741) equals the corresponding term of Eq.t is called the second l!ariaiwn. we obtain 6. The Maclaurin expan ".. Englewood Cliff" N.t(O) Note that the first term is the true resonant frequency. Hildebrand. If the first of Eqs. we obtain III H·v X .. . Metbodaof Applied Mathemat.J. .iea.(E+pe)d... is the resonant frequency.. (736) and substitute for the true fitl4 E a trial field (738) where 'P is an arbitrary parameter. where we show sion of ".t(p) = wrt + p&.E'd' III E· V X . (736) Similarly. B.. 0 +t up 21 up 2 t I . w'(p)  This procedure gives (739) ''''77. by assuming field distributions in a cavity.d (737) Equations (736) and (737) arc identities. each term of Eq.(E+pe). They are particularly wellsuitcd for this latter application because of their Ustationary" character. A formula for w t is said to be IF. These reduce to the usual Helmholb equations when f and ~ are constants.332 TIMEBARMONlC ELECTROMAGNETIC FIELDS where w.. multiplying tbe second of Eqs. because "'rt.'V X Hd." p.L!/"''/I'. even more important. function of 7J for fixed e. In the variational notation l the above expansion is written ::::I as ".Inc. .."t + p' + . (740) ".t is wt as a. 1952. PrenticeHall. III (E+pe)·V X .2(P)=Wr t ow'l p2 u w +p.. w. 1II. tbe term 8t ".. (735) is scalarly multiplied by E and the resulting equation integrated throughout the cavity... The term 8".H' . they are useful for approximating w. 0 + .t is called the first variation of ""t.. which we shall now discuss.. sDd so on. (735) scalarly by H and integrating throughout the cavity.' = '''/'/'I'. We take Eq.1.·v xEd..' ~ !.·V X (E+pe)d...
 fffoE'd. (7·39) is. A similar identity It is a vector identity that 111 E· V X pIV X e dr  The last term vanishes. The above equation vanishes if n X e = 0 on 8._0 _ D(O)N'(O)D'(O)N(O)D'(O) 1ft [(. We now wish to show that Eq. The derivative of the numerator N(p) evaluated at p .e·Ed. (736). 1ft [(.. Using these two identities and the fi.n one p parameter is straightforward. "'" 0 on 8. We then obtain aW'1 ap . fff .2 vanishes.0.·V X E) X eJ· d.IV X EdT 1ft [(. using Eq. for the resonant frequency if the tangential components of the trial E vanish on the cavity walle. This is equivalent to (742) The extension to more tha. which requires n X EbloJ. (739) is stationary.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 333 stationary if the first variation of ". for p . ff f .e· E d. Hence. (736) is a stationary formula. D'(O) .tor D(P) of Eq.rst of Eqs. (735). V X EdT"'" IfJ e· V X ".·v X e) X E)· d.'v X E) X e)· d. because n X E "'" 0 on S. we obtain N'(O)  2w.IV X E) X eJ· d.. Eq. (743) which has been simplified. .0 is N'(O) = JJJ (E·V X pIV X e +e·v X pIV X E)d'1 JJI pIV X e· V X E d'1 + 1ft [(. states JJf ptv X !.2 The derivative of the denomina.
provided n X E lrial =: 0 on S. All such modifica.LLf"'f". Further modifications to account for discontinuities in n X H or n X (f.in follow directly from the methods of Sec. ds The last term vanishes. The Hfield formula corresponding to Eq.'V X EdT .·(v X E)' dT + 2# [I. 77. especially if the geometry is complicated.·V X E) X E]· ds is added to the numerator.' _ f".1V X H) over surfaces within the cavity can be made. Still further modifications in OUf formulas are required if n X E or n X (/r1v X E) are discontinuous over some surface within the cavity. (736).fl'dT (746) which turns out to be stationary subject to 00 boundary conditions on S. X E) X E] .E'dT which is stationary..tions can be quite simply effected by the reaction concept of Sec. (744) is . we obtain Substituting this w.tiooary formula.334 the identity TIMEIlARMONlC ELECTROMAGNETIC F1ELD8 Equation (736) can be put into a more symmetrical form by applying III E· V X .III .·V 0 on S. in terms of H. it is evident that the requirement n X E kW = 0 on S caD be relaxed if the term 2# [(. even if n X £"1&1 yf 0 on S. . provided n X (ciV X H) = 0 on S. 77. (744). These modifications aga.c'v X E) X E!· ds lll. because it is not always easy to find a trial field with vanish· ing tangential components on the cavity walls.·(V X E)'dT l.I<"7'7(V_X_H_)_'_dT JJJ . . (737) is a st8.f".E='d:TI (744) This formula proves to be stationary.. A similar procedure shows that EQ. If we look carefully at the first variation of Eq. because n X E identity into Eq.'  This gives (740) III . This is an important modification.'  III .·V X E· V X EdT = + # l<.
the method of establishing stationary formulas is to construct formulas of the proper form and then separate the stationary ones from the nonstationary ones by testing the first variation. 77 we shall give a general procedure which leads directly to the various stationary formulas. say . llS shown in Fig. . One characteristic of all such formulas is that the numerator and denominator contain squares of the trial field. smaller error in 61' than does the nonstationary formula.tage. 1 This is shown in Fig. It is evident that for small PI the stationary formula gives a. saddle point at 'P . as we shall show later.O. The stationary formula then gives upper or lower bounds to the parameter.' o (a) p. In Sec. Our formulas for 61 ' give upper bounds.dE ." An crror of the order of 10 per cent in the assumed field gives an error of the order of only 1 per cent in the parameter. In some cases the true field can be shown to yield an absolute minimum or maximum for the parameter. This property is sometimes summarized as follows: fl A paramctcr determined by a stationary formula is insensitive to small variations of the field about the true field. Figure 78 shows pictoraUy the primary ndvao. I A complex parameter would have 0. p Flo.. Illustration of Wi Ver1!US 'P for (0) a stationary formula and (b) a nOlUltationary formula. Classically.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 335 . the corresponding error in the resonant frequency is WI' . (738). The parameter 61 ' determined from a nonstationary formula must have some definite slope at P =* 0.'. 78a. p o (b) p. the parameter w' (p) determined from a stationary formula such as Eq. Let us now briefly consider the advantages of a stationary formula over a nonsLationary one. 78. For a givcn error in the assumed ficld. 78b.pie. This insures that amplitude of the trial field will have no effect on the calculation. Given a class of trial fields of the form of Eq.w. (739) will have a minimum or maximum at p = 0. We might also inquire about the general procedure of establishing stationary formulas.
58 and is given mathematically by E. give us Eq. of course. (747) The field is sketched in Fig.405~) H~ .2 (1  ~) Equation which is chosen to satisfy the condition n X E = 0 on S..2 Equation (746) then becomes wi = foG 4pdp EIJ. Suppose we first try a formula that requires no boundary conditions [Eq. ('146)1. The TMo.336 TIMEHARMONIC ELECTROMAGNETIC FIELDS OUf Now let us apply some of stationary formulas to a problem Cor which we have an exact answEl):. = a VEP.. = j J o (2. Assume as a trial field VXH=u.818 a~ (748) This is 16 per cent too high. (746) then yield. so that we may get an idea of the accu racy obtainable. This suggests that our trial field was too crude an approximation.o mode is dominant and the exact resonant frequency is 2.4048 w. for the case d < 2a. We can improve our trial field by assuming H = 14 (p  %) 4(1  V X H = u.. 57). 1012 = l" fll p j. Consider the dominant mode of the circular cavity (Fig.410 _ /a v EjJ.~05 J (2. 2.o' ( 23a')' pdp p ~)' pdp 180 ElI __ . 8 =_ Epa' 10 4 pVJP and our approximation is w r ~ 2. which is a relatively poor result.nd oW' approximation is now Wr "'= 2.31a' 4. (749) . (747).405~) 1 Substitution of this true field into any of our stationary formulas must.
 Ilf pe' (v X IlIV X pe . since n X E "'" 0 on S."i('=~)"'. McGraw~HiI1 Book Compa. we would have had to use Eq. (745). If the same trial field is used in Eq.. 0 on S. (744.(E+pe)':a. Even though a formula is stationary. the fields of the various modes.'L'j'j'j.w. say Eq. dp 6 2. Note that all our approximations are too high. Chap. (739)..)." part I. (737). (i51). that is... the eigenfunctions..' . "Methods of Theoretical Physics. Now consider a stationary Efield formula. (744). for this will hclp to obtain a trial field close to the true field.449 _l_ (750) OVE".8 per cent too high. Substituting the above equation into Eq. making use of the wave I Philip M.PERTURBATIONAL AND VARIATIONAL TECUN1QUES 337 This is only 0..t. henee we choose VXEu. It is advisable to meet the physical boundary conditions 88 closely as possible. This suggests that the true resonant frequency is an absolute minimum. Inc. take Eq. by means of various identities.2 per cent in error. form a complete set of orthogonal fUDctions in the cavity space. This formula requires n X E = 0 on S.. For example.ny. .W/Epe) dT (751) It is known thR. and.. we again get Eq.a Substituting this into Eq.a' Ell Jo l~ pdp Our approximation is therefore Wrt::S Jo a l {o!. the error field pe can be expanded in a series where the Ali are constants and the Eli are the various mode fields. which is 1. New York. If we had chosen a trial E field not satisfying n X E . Morse and Herman Fcshbach. (749). 1953. which we shall now sbow. 6. put it into the form w' . we obtain 1 w' ~ r"'.! Hence. we must use care in choosing trial fields..
' JJ! .· pH the preceding equation becomes This requirement is met for all U by the conditions of Eq. Eq.Jf. A further advantage of the variational formulation is that one can choose the best approximation to a stationary quantity obtainable from a given class of trial fields.. Since we ha.t• lI • = va. (753) The desired where. Any trial field satisfying . which is seldom the case for complicated geometries. The Ritz Procedure. Hence. .'f"!"!'. we obtain l wI  (~..')A. since DO boundary conditions are required.ve chosen W r as the lowest eigenvalue.'dT (752) where the arc the resonant frequencies of the ith modes. This is done by . any w calculated from Eq.OonS is orthogonal to aU static fields. (752) is always positive. requires that tbe dominant modo be known exactly. (747»).ve an upper bound in this case? The answer lies in the fact that we have overlooked the "static mode.E.. Why do we not ha. of course. it is easy to insure that our trial field is orthogonal to all static fields. The result is W r "'" 0. so we obtained upper bounds to the dominant TM oio mode. This.' w r ' = ~. (736) will be an upper bound to the true resonant frequency. if we choose a trial field orthogonal to the field of the lowest mode. Also." A static magnetic field (wr = 0) can exist in a cavity bounded by a perfect electric conductor. in general.(:E"'")'dT . H .H) . (U. "U + U.338 TIMEHARMONIC ELECTROMAGNETIC FIELDS equation and the orthogonality relationshipl5. By virtue of the identity ". 76. we have an upper bound to the next higher resonant frequency.H. thereby obtaining an upper bound to the dominant ae mode. Our choices for H in the foregoing examples satisfied Eq. (753). and so on. Look now at Eq. which is less than the true resonant frequency IEq. Fortunately. as desired. The trial field H = constant vector is a permissible trial field. as we shall now prove. (746).. (753).w. orthogonality is w.
we obtain (2 ".. . we obtain (755) The best approximation to w. (746). the method may. it is also necessary that some choice of the Ai will give a reasonably close approximation to the true field. Substituting the trial field into Eq. For example.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 339 including adjustable constants.' i ::::0 1. 2 will be the minimum value of can be chosen by requiring =0 aA." . so we choose Eq. in principle.2.n This general method is known as the Ritz procedure.+ 3Ap)'p dp _ f: = I + A p2)!p dp 15 8 + 1Ma + 9(Aa)' a"~ 15 + 24Aa + lO(Aa)' E~ foG (p (759) The method is also referred to as the "RayleighRitz procedure. if we choose (754) where the Ai are variational parameters.:""".. and substitute into the stationary formula Eq. (757). is used. (p + Ap') v X H ~ u. When a complete set of functions E. let us again consider the circular cavity of Fig. However. I The most common way to include variational parameters is to express the trial field as a linear combination of functions (757) Since the labor of the calculations increases approximately as the square of the number of terms in EQ. For an example of the Ritz method. w\ which (756) a". or variational parameters. (746) as the stationary formula.' _ J. It is also sometimes convenient to choose the E. 57 and trial fields of tbe form H = .. lead to an exact solution. it is desirable to keep n small. (736). .(2 + 3Ap) (758) where A is a variational parameter_ Note that H satisfies no boundary conditions on S.. as an orthogonal set. in the definition of the trial field and then choosing those parameters which will give a minimum or maximum of the stationary quantity.
191. RumllCy. . using the concept of reaction 88 defined in Sec. To reiterate. The Ritz procedure also gives us an approximation to the true field.31 gives ka = 7. Hence.520.b)  f (Eo. Many of the parameters of interest in electromagnetio engineering are proportional to reactions.b) . H. Reo. The Reaction Concept in Electromngootic Theory.' This can be solved (or Aa as A a _ 55 ± v'ffi ~ 56 ( 1. pp. The Reaction Concept.Ab) (764) where the notation Aa means the a field and source are multiplied by the number A. we set 3i'.0 aA 24 + 55Aa + 28(Aa)' .b) . 94.340 TU!EHAIW:ON1C ELECTROllAGNETIC FIELDS Note that the approximation of Eq. and 80 00. we obtain approximations to the lowest three eigenvalues. (747)J. PhlJ•.a) Ho .c) (Aa. the reciprocity theorem (Eq. which is an approximation to the next higher eigenvalue 5. lief. (336)1 .b) . To determine A by the Ritz method. 14831491.A{a. dM') (762) H all sources can be contained in a finite volume.b) + (a.0 a". no. 38. The solution Aa . but it is difficult to esta. 6. (761) is the desired "best" approximation to t.(a.he true resonant frequency {Eq. dJ' (a. (763) The linearity of the field equations is reflected in the identities (a. the reaction of field a on source b is (a.1. If the trial field bas two variational parameters.. and obtain . 2. Eq.4087 (761) which is smaller than what the first of Eq. For example.(a. (749) is the special case Aa ". 77.blish the nature of tbe approximation. (760) gives.6543 (760) A substitution of the second of these values into Eq. 1954.b + c) .. (759) gives '" ~ aw 2.(b. 1 A general procedure for establishing stationary formulas can be obtained.3100 0. vol. the impedance parameters of a IV.. June 15.
The reaction (a.)  + P.) + p... (765) is also stationary for small variations of a and b about c. It is then argued that tbe best approximation to a desired reaction is that obtained by equating reactions between trial fields to the corresponding reactions between trial and true fields.0 We can think of this as stating that the resonant frequencies are zeros of the input impedance.c. we have the three relationships (a.b) ..) (a.) .a) ....) Using the last two equations in the first equation.. so the reaction of any field with the true source is zero... if we let a = b represent a trial field and associated source.) + p.") The approximation (a..) = (c. (766) are satisfied.« .c. To be specific. we obtain It is now evident that Eqs. proving the stationary character of (a.. (767).(c. To apply Eq.b) = (c.. (341).p..p.(c..b) ap... For example.b).. as shown by Eqs.b) .e.) (71i5) because we have imposed all possible constraihts. Approximations to the desired reactions can be obtained by a.JlIloO  a(a.p. Hence.«. Eq. an assumed E field can be sup .. (765).00) Substituting for a and b ioto Eqs.(c... We have a slightly different case when the reaction concept is used to determine resonant frequencies of cavities.ssuming trial fields (or sources) to approximate the true fields (or sources).. (The symbol c stands for II correct.c. and c.b) aP/l I ""'PlO 0 = (7.b) obtained from Eq.) (c.b) is then best if we 15ubject it to (a.) + PIJ. .«.(a.<.c..(c. we assume a trial field and determine ita sources irom the field equations. This we can prove by letting and showing that a(a. (765) reduces to (7f>7) (a.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 341 multiport Unetwork" are proportional to reactions..e... The true field at resonance is a sourcefree field.. suppose we want an approximation to the reaction (c. I P.)..) + P.b) . Equation (765) caD be thought of as the statement that all trial sources look the same to themselves as to the correct sources..(e.
Statiomiry formulas in terms of both E and H are also possible. This is illustrated by Fig.E  }w ~V X (.V X E M. If 8. C~. We now substitute from Eqs.. (767) now lea.""nxC~VXH) onS as represented by Fig. (768) and (769) into Eq. If n X E ~ 0 on S.~ 1P (n X E) .'V X E) • (768) However. 79a. (737) if M. this reduces directly to Eq. .nd magnetic currents.'V X E) dT . (736). 1P Iff HII ~ E dd . trial H field. we need the additional magnetic surface currents M. stationary formula. Application of Eq. (b) a trial H field.O.II  }w ~ V X (.E·Edd M.'V X E) d' If n X E = 0 on S. ported by the electric currents ] ~ jw. (746).a) = jw III J. 7·9.jw. V X E) ds E· V X (.'V X II) (770) M. They are found from the trial fields according to J = jW<E +VXH (771) M = jwllH . Sources needed to support (a) a trial E field. This time we consider both electric a..342 TIMEHARMONIC ELECTROMAGNETIC FIELDS '8 (a) (b) (c) FIG.ds to Eq. (767) and obtain o (a. nxE ODS . as shown in Fig. in terms of the H field is desired. = n X E on S (769) to support the discontinuity in E at S. and (el both a trial E field and a.. (745). 79b. . or to Eq. 79c. (. the above equation roouces to Eq. if the trial field docs not satisfy n X E "" 0 on S. we consider the trial field to be supported by the sources M .
(772) This is sometimes called a "mixedfield 1/ stationary formula. (765) were stationary. In the usual manner. if the trial fields have discontinuities in n X E or n X Hover surfaces within the cavity. da XH (jW<E' + E· v + H· v X E + j".a) to vanishj hence.w + a(a.a) ~ iJw. <jp H .a) is a function of both wand p. but it. po . Earlier we showed that reactions constrained according to Eq. we have a(a. parameter times an error field. <jpE X H·cls which can be rearranged to " ..E') d. we must add the appropriate surface currents to support the discontinuities.H') d. the first variation of w vanishes..O. o The second term of this equation vanishes because (a. as shown in Probs. 7Zl and 728. We shall now prove that the w so determined is stationary about the true resonant frequency.PERTURBATIONAL AND VABlATIONAlo TECHNIQUES 343 Equation (767) then gives 0..a) ~ iJp p _ 0 .. M) . 1P E X H· ds fff (....a) is stationary about p . But in the above cavity formulas we calculated w by forcing the reaction to vanish. 84). The coefficient of the first term is not in general zerOj SO ow = 0 Thus..H' . This procedure leads to additional surface integrals in the stationary formulas. J  H . We . the denominator is twice the stored energy in the cavity..fII (E· V X H + H· V X E) d. Finally. p. we let the trial field be the true field plus 8. and all formulas for w derived from Eq. as wand p are varied.fII (E.. (767) are stationary. represented by a=c+pe For fixed e the reaction (a. Henco. is easily shown that E and lJ are 90° out of phase in the lossfree case (see Sec. The reaction concept also provides US with an alternative way of viewing the Ritz procedure for improving the trial field or source. Equation (767) constrains (a. The minus sign in the denominator might seem strange.Iff d.1 . M.
represented by a"'" Uu + Vv + .v) = (v. (775) reduce to U(u. According to the reaction concept.v) + V(v.u) (a. 57). V. (c. all trial fields should look the same to themsel ves as to the true source.v) +  ~ (c.v) = 211'da t ( jwp. (773) where U. To illustrate. because the true field is sourcefree. = 3ia (777) w.1 w.v) = . Our trial field was Eq. Calculating the various reactions according to Eq. (762). V.v) + V(v. according to Eq.344 TIMEHARMONIC ELECTROMAGNETIC FIELDS assume the trial field or source to be a linear combination of functions..v) (775) which can be solved for the parameters U.u) + + V(v. 76. which was the Ritz procedure applied to the circular cavity (Fig. we obtain the set of equations U(u." =.u) U(u. .u) + V(v. arc numbers to be determined. .u) .(c."4 2 + jWf:) a' (u.(c.u) = 211'da 3 ( jwlJ.v) . The solution 80 obtained is identical to that obtained by the Ritz procedure. Bence.. a' '6 + jwri 9) 0 0 2 + jWf) (778) All reactions with the correct source Ilre zero. we obtain (u. hence we should enforce the conditions (a...lt) = (c. so for the same approximation by the reaction concept we choose (776) The sources of these fields. let us reconsider the example of Sec.v) (774) Substituting from Eq. 5" (v. M . are 2· M.u) := a' 211'da 2 ( jwp..u) U(u. (773). (770). (758).u) (c.v) = 0 and Eqs..
'v II tE'da X E) X E] .Sln a .ll) . let. yields Eq. Corrections for discontinuous trial fields can be made as outlined in the preceding section. The solution of Eq. n dl (780) where n is the outwardpointing unit vector normal to the waveguide walls. (U. (761).. As an example. (767). if we take a slice of the waveguide. so we should expect stationary formulas for the cutoff frequencies to be of the same form as those for the resonant frequencies of cavities. At cutoff. II .'  II . Renee. "" U.nce.lOntrivial solution only if the determinant of the coefficients of U and V vanishes. 484.'(V X E)' ds + 2 f lc.fI (E· V X H +IH· V X E) d.UXll. (7412) None of the above formulas require boundary conditions on the trial fields. The Hfield formula corresponding to Eq.(U. Starting from Eq. the Efield formula corresponding to Eq.f E X H· n dl I (. . In See. (780) and a trial field E  . us use Eq. . 46 we obtained a transcendental equation for the cutoff frequency [Eq.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 345 These equations can have a . because the sources of our trial fields are not of finite extent. (779). (746) is w '  . we arrive at stationary formulas differing from our cavity formulas only in that volume integrals are replaced by surface integrals and surface integrals by line integrals.'(V X H)' d. The height of the slice is common to all terms and therefore cancels. (778). II·H'd. as was done in Sec. 78. (451»).E') d. (745) is w. surface integrals over the top and bottom just cancel at resona. For a variational solution..II' . with the reactions 01 Eq.J.J . (7411) and the mixedfield formula corresponding to Eq. However. (772) is w. consider the partially filled rectangular waveguide of Fig. 74. a waveguide is a twodimensional resonator. of course. Stationary Formulas for Waveguides. be careful in applying the reciprocity theorem. Hence. We must.ll)1 = 0 (779) ia the equation from which w is to be found.
y)eJ~' == Ca. and the derivation proceeds as follows. in contrast to the exact equation. 2. (726). . (1 2r~]" SIna (783) Noto that this is an explicit fonnula for We. and It .} [nc.352 0. In all of the previous examples.. t B. We therefore have need of stationary fonnulas for propagation constants."(da 2r 1.85 0.350 1. I For inhomogencously filled waveguides.319 A knowledge of the cutolT frequency of a waveguide homogeneous in and ~ is sufficient to determine the propagation constant at any other frcqucncy according to Eq. as for example the abovetreated rectangular waveguide with dielectric slab.B. IRE Trana. J3erk.346 TIMEHA1U£ONIC ELECTROMAGNETIC FIELDS which is the emptyguide field. there is no simple relationship betwecn the cutoff frequency and the propagation constant.VEoumE WIOTlt TO CU'J'()I'P WAVELENGTH FOR TIll: Rr:cTANOOLA"lt WAVEGUIDE WITH DII~LECTaIC SLAB ("ElI:aet" valuce read from curvell by Frank) dfa a~ (cnct) al>. It turns out that the operator for waves traveling in the z direction is the adjoint of the operator (or waves traveling in the +z direction. TABLE 72.= o Vt!'ut r [1 +" . pp. 44. 1956. 1956. D. We should expect the approximation to become \Vorse as tdEt becomes larger. (approximate) 0 0.)eii• H+ "'" ft+(x.286 0.50 0.O. p.000 0. Friedman.)ei '. the field equations lead to an operator which is not selfadjoint Hence. vol. since the field then tends to concentrate more in the dielectric.'1. the field equations were given by an operator which was selfadjoint with respect to the desired integration. u. no. If the guide is inhomogeneously filled.500 0. 104110.. an appropriate adjoint operator must be fowid and the derivation of the stationary formulas suitably modified..4.167 0.319 0. Variational Principles for Electromagnetic RCl'Ionatotli and WllV~ guides. :::I + (784) I A. April.383 0. .45(.375 0.4sa 0. Eo. New York. 2. AP4. Define +2: traveling waves 88 t E+ :t+(::.486 O. RATIO OF W . which is transcendental." John Wiley "nd Son.600 0..500 0. The result is l w... Table 72 compares the above result with the exact solution for the case fl ..500 0.y)eJ6' == (~I + u. "Principles and Techniques of Applied Mathematics.
. fl+ .= jpu.w.= jpu.B.t+ . and proceed as in the derivation of Eq.2 II~.. r.f!+  jPU.y)e'" = (fl. (V X ~) pIEI' vEl w'.tt+ . and add the two resultant equations.)e"· (785) where the t/. and t. t] d8 = 0 (787) stationary if n X E = 0 on C.. we find a.. n a .. as shown in the insert uf Fig.... X iI By direct substitution.Uwsm . take . for which ilJ should be replaced by 'Y = (I + iP· For an example of the calculation of propagation constants... Ji..jw'£+ . consider the centered dielectric slab in eo rectangular waveguide.fl. V X ~ + ~+ . X ~jw.£+ .P:. ~ = 2iP~.+ Jll. + jw. The resultant formula is pi II lAIE/Ids . and 9 . (786). Equations (786) and (787) remain stationary in the lossy case. E replaced by P. V X ~ + J"t+ . (tl . ~. x ft+ X ~ Using analogous definitions for z traveling waves.u. tt. stationary if n X E = 0 on C. eliminate it from the +2 and 2 wave equations.j21J JJ dB + II I.. of Eqs. lA.. . it can be shown that for any +z traveling wave solution there exists a 2 traveling wave solution given by E..PERTURBATIONAL Al'Ii"D VAlUATIONA. ~(x.'(v X t+) . we obtain V X ~+ V X it+ V X ~V X + jW. This gives e. when integrated over the guide cross section and rearranged.jW&:+ "'" ilJu. fl. Now multiply the first of the +z wave equations scalarly by :A. Tho Hfield formula is given by Eq.+ jw.)e»· H..y)e~" ..L TECHNIQUES 347 Substituting these into the field equations. AJ< a trial field. (784) and (785) are the same functions..he second of the z wave equations by ~+.fl. V X a) dB P = <.= f!(x. (787) with E. + u. X II a. which. H. For the Efield formulation.7"""C. 710. X a l • u.<. E. and it is stationary with no boundary conditions on H..2.dB (786) This is a mixed4field formula. u. V X iI. yields (w..
451:0. 710.2. Fla. (788) with the exact values for pjk a is shown in Fig.348 1. The antenna.45 ~ /.1 ~O 0. A formula for impedance in terms of reaction is given by Eq. 419 and requires the solution of a transcendental equation. cU..4 ExactApproximate  a 1. Figure 711 represents a perfectly conducting antenna excited by a current source.0_0.8 1//' / ( 02 0.3 l=: 0.I:.ngential components of the total electric field vanish on the conductor. (After BeT. 79. 710 for the case E = 2. 0. ~ kG [1 + ~ a '+! Sin!'!) . (787).6 TIMEHARMOllo'lC ELECTROMAGNETIC FIELDS y+ Idl+0 1.. A comparison of a values obtained (rom Eq... If a trialeurrent distribution J.4 0.) and use Eq.0 1.454.(.4 BEI 1.8 V ~:  d/o.!. lJp. Stationary Formulas for Impedance.ion col1$lanta for the rectangular waveguide with centered dielectric slab.6 ~~ 0. . terminals are close togetherj so the reaction of any field with the current source is of the form . t . Compariaon of approximate and enet. The result is l p. propagat.1. koa (788) The exact solution is given in Prob. the formula for input imped· ance [Eq.5 0. when constrained according to Eq. is a stationary formula for impedance.2 of>.VI.1 • 12 • 2.)']~ (~ l l " a E. Such a formula. (765). (341)J is (78lJ) I Berk.is assumed on the antenna.. (341). The resultant current on the antenna will distribute itself so that ta.
Circuit Relations iD Radiating Systems and Applications to Antenna. (765) are met is similar to that for selfimpedance. Problems. (c. On the antenna surface. 712. 10041041. (789) and (790).l solutions to the same problems. (791) involves the assumption of currents due to both sources. Let the z axis lie I P. . (765) have been met. respectively. The calculation of mutual impedance is usually simpler than the calcu~ lation of selfimpedance because the source and field points are separated. 7·11. the tangential components of tbe true field E~ arc zero except at the input. as. for example. If we have two sets of input terminals. Equation (790) is not stationary unless both the true current and the trial current are real. When the trial current is MSumed real.. 6.a) = (a. An antenna excited by a current &(lurcc. Hence. 712. as long as the antenna diameters are small compared to wavelength and compared to antenna separation. as we shall now show.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 349 where I is the input current. The extension to N sets of terminals is straightforward.c) by reciprocity. the variational formula for mutual impedance is (791) where I .a) lC:  Vel = . 20. since E" is the field of J. we get tbe same answer from Eqs.. The demonstration that the constraints of Eq. 50 the constraints of Eq.. Carter. (789) is stationary about the true current. Note that Eq. DO. the input impedances for waveguide feeds calculated in Sec. 1932.a) I Also.re also variationa. June. vol.' This method should not be confused with the induced emf method (790) FIG. B. and 16 arc the input currents at terminals a and b. No appreciable error will be incurred by assuming the currents as filamentary. Consider the parallel linear antennas of length >. Let us therefore take a mutualimpedance problem as our first example. Equation (789) was used to calculate impedance before its stationary character was noticed./2 as shown in the insert of Fig. IRE. in the case of the two linear antennas shown in Fig.PZI • = (a. PrO(.. which is based on the conservation of complex power. (789) is a stationary formula.. The impedance as calculated by Eq. and Eq. hence (c. pp. 410 a.
.1/ / •• fV' ). _ R. A .. cos T 2".. 1\ we obtain Vol G(. at antenna. o 20 \ \ K /1. + ..350 BO TIMEHARMONIC ELECTROMAGNETIC FIELDS 60 i\ \ R.' _ }WE v3 + k') A.z') 1 (a.i1. By the usunl vectorpotential method we have E.• 4.'jI . I jW4 'A/4.. free space.5 "" 40 1". cosX (792) Our formula for mutual impedance [Eq.j4 E.j" dz' cos "'" cos ~ G(z."I'd. (791») become8. 712. 40 o '/l~di / \~/ V'' c§ 20 1\ \ x. 2". and assume /./ FlO. .. ["(t') ei1v"'+(0/)' V d' + (z 2rz 1\ %')2 dz' Substituting for E. between parallel 1+. ) e..and I' in our expression for Z... b... Zn = . P ...X. where '" f Z..') I _).'d'+<0./2 linear anlennu in along antenna a.I. where. in this case.0 ~..+ k' kjwE az' V d' + (z z')' dz f)"/4 2rz' (793) (794) . I.... Mutual impedance Z.' ~(:'.
High Frequency Models in Antenna Inveatigations.u) + V(v. we obtain (a a) = {(c u) (cu») . Proc. (795).v) .v) Solving where (e. = . we obtain for the result R.' + .U(u. "(11. 22.'11 Silv'(kd)' +.Cilv'(kd)' X. Let the current on the antenna be represented by two functions. S.s The evaluation of the selfimpedance of a linear antenna is more difficult because oC the singular integrands encountered. TRE. Our trial current is then a surface current of the form J. +iX. pp. for U and V. 4. Circuit Relations in Radiating Systems and Applications to Antenna Problems. as we shailialer show. vol. (v. vol.'11 (795) where Ci(x) nnd Si(x) are as defined in Prob.ND VARIATIONAL TECRNlQUES 351 The integrations of Eq... The geometry of the centerfed linear antenna is shown in the insert of Fig. Proe. April.1') [(U. I G.U) (v.v) + V(v. 6. the trial functions should look the same to the assumed current as to the true current..u) ~ U(u.l. King." (796) where U and V are adjustable parameters. .' .' . Figure 712 shows 0. no."'!2Si(kd)  + . 1934.U)] (e..(c. The mutual impedance between linear antennas of otber lengths and orientations can be found in the literature. 457480. 1932.PERTURBATIONAL A. pp.v) ~ (c. according to Eq. June.u) . 1004104. hence we enforce the conditions (a. According to the reaction concept. (773). 20. (793) can be expressed in terms of sine integrals and cosine integrals.. we have in matrix notation Substituting for U and V into Eq. 713. I Letting Z. The details of the integration can be found in the literature. plot of Eqs.v) (798) I P.' + . (796) and calculating the selfreaction.') Silv'(kd)' +.UJ.'J  + .u)].u) and (e. Let us use this problem to illustrate the use of adjustable parameters in the trial current. no. Brown and R. IRE. .u) (a.1." + VJ..v) can be calculated.. Cilv'(kd)' R. 244. = 4: (2 Ci(kd) .. Camt.v) [(C.
Variational solution for the input impeda.2400 I V o 2 4 6 kL (b) B 10 12 FlO.000 1600 800 ~ / '"\ E o V L/' ~Lla = 1800 I#" £V o BOO ~ Ljo ~ 150\ I/j. Hu. o 24oo ~ Lla"" 150 V rl II "~ (oj / /" I Lla = 22. (Afttr Y.352 5000 TIMEHARMONIC ELECTROMAGNETIC FIELDS z f4000 f T L . Y.. .) (a) Input resistance.nce of the symmetrical cylindrical antenna.7'// / IV II 1// / 1600 .000 1+i ~2a 3OO0 J 2000 Lla = 1800 1000 Ah. 713. %=0 Lla a:: 22. (b) input reactance. .
. (798). A precise evalua. we obtain ( ) (c."  2 l"(.v) + ! .) u.v)'(u. where a is the antenna radius. are the values of the u and v trial currents at the input. is the input voltage and [" is the z current at the input. Note the singular nature of the Grecn's function [Eq.1.VI. By the potential integral method we can calculate the field of the current J.21.v)t Now note that n x E· = 0 on the antenna surface except at the feed.ctioos of Eq.S ' ( kS 8 1t Jwt a') L/2 0 + a ILI2 dz' 10'" dfj/J.s as 1 E. (7102») a..x) "'" . and I..u}(v.tion of Eq.v) which can be rearranged to read (799) where I. Let us now look at the form of the reactions. where VI. The field of the current is approximated by the field of a filamentary current of the same magnitude. Expanding Eq. The currents will be rotationally symmetric %dircctcd surface currents on the cylinder p = a. V.vXu.!" (or any z.a . we obtain Z I.'" "" . Using the above two relationships in Eq.. t 1. This is . """ Z I.. These currents can be expressed as r. using the reciprocity condition (u."'"J. 80 (c.v} (u.' 0 (7103) where Eo'" is given by Eq. ..v) = (v..v) .1 (11.y) = I£/2 L/2 dz 10'" adq..(u.u)'(v.a. (799) are then given by (x.uXv.PERTURBATIONAL AND VARIATIONAL TECHNlQUES 353 Equations (7+97) and (798) also apply to the case of N adjustable constants if the various matrices are extended to N rows and/or columns. (7103) would be difficultj SO the following approximation is usually used..v}S (u.u).v) + (c. (789).t p .S(v.rG 2 % (7101) (7102) where The various ren. (7101) with p =* a. 11.uXc.E.a 1 (7100) where I" is the total current and x .v) (u.'1') .2(c.u) a. (u.
For trial functions in the second~order solution. Storer. pp.ttering Cross Sections of a. Centerloaded Cylindrical Antenna. It is said to be antiresonant when X is zero and kL Rl n'll'". and formulas in terms of sine integrals and cosine integrals have been given by Storer' and Hu.. Cruft LaO. fLI' L/2 d' fLI' L/2 liz' 1. I Y.) G uZ (7105) where G is given by Eq. as they will be for thin antennas.v) is long and involved. 1952. (7103) (x. Variational Solution to the Problem or the Symmetrical Cylindrical Antenna. Eqs. Note that./2 linear antenna. Eqs.354 TIMEHARMONIC ELECTROMAGNETIC FIELDS p = equivalent to replacing Eq.co.. = 42. vol AP~6. . The resultant current is essentially that of Eq. (7100). I J. no..u}.. note that for very small a = d.(. As long as the equivalent magnetic currents are negligible. we obtain from Eq. (7102) for a by (7104) For thin antennas. Y.) (k' + :'. I. Mass. BackllCa. n odd. Cambridge. The field of the filament of current is a sourcefree field in tbe region external to the linear antenna. Hence. t Numerical values of the input impedance are given in Fig. (7104). January. the selfreaction is equal to the mutual reaction between two identical antennas fed in phase and separated by a distance B. 712. Using the above approximation for G.. k (~  (7107) 1'1) have been used in the literature.!.')1(. IRE TraM./2. Note that.v}. The evaluation of Eq. and (v. (u. We can therefore assume that this field exists and calculate the equivalent currents on the surface of the antenna..y) ~ "X~JWE . Hu.5 (7106) as is evident from Fig. I"  Sill k(~ 1'1) I' .. (7~105) for (x. In particular. Rep. 140148. we can take the equivalent electric currents for our trial currents. 0:. 713. (795) reduce to R. as caD be shown by the following argument. 1958. to this approximation. n even. Resonance (X = 0) occurs for L slightly less than >.1 X. The antenna is said to be resonant when X is zero and kL n'lf. = 73.1 . the error introduced by Eq.y} = (u. (795) with d replaced by a give the firstorder (one trial function) variational solution for the input impedance of a >. E. TR 101. (7104) is negligible.
If .PERTURBATIONAL AND VARIATIONAL TECHNIQUE8 355 D in the vicinity of resonance.ve EI' (S. or n X E. The more general case of dielectric obstacles is considered in Sec. Perhaps a better choice for the v current would be 00. The reaction of the sca.acle. R is inE _ El + E~ sensitive to antenna thickness.. For simplicity.s) Echo .cle hy the source. Let us first treat the ba.c) 1 Af.(II)' 'Jr E' . be denoted by Ei. calculations have not been made for this choice. U correct" currents induced . Stationary Formulas for Scattering.st.to n.ckscattering. .ering by an the input impedance calculated there.. or radar echo. . using reciprocity.. J.(ll)' (7111) where (c. is the current induced on the perfectly conducting obstacle. Then.t.n be written as (c. The problem is represented by Fig. Wave seat. the obstacle will be considered a perfect conductor and the source a current element n. Hence. d. Both trial currents of Eqs. 210 gives good results.i) ~ liE. I"  ~  Izi which is finite at z = 0 for all L > O.n X E' (7110) Hence. 711. Let the incident field. the freespace field of the source alone. we ha. It consists of a source and one or more obstacles. 714. (7107) are zero at the input for kL = 41r. 7·10. from cannot be valid in the vicinity of kL = 47. on S (7109) where J.' ~ IV' (7108) where V· is the scattered voltage appearing across l. Let the echo be defined as the ratio of E.e) stands for the selfreaction of the on the obsta. The boundary condition at the obstacle is n X E = 0. Eq.(II)' ~ (II)'  (I~)' 1P E. Echo . (7109) ca. . that is. However. type of problem by the variational method. and we wish to determine the field scattered back to the source. It / Obstacle is in these regions that the analysis tSource of Sec. J. 714. d.ttered field on the current element is (8. The total field E with the obstacle present is then the sum of the incident field Eo plus the scattered field E·. FIG.1) (i.
(7113) (TI)' 1f Eo..BoeJb .1f d. we take (i.on S and approximate (c. (r(r. where E. The tensor Green's functions alSee. 714. we have (i a)' Ecbo = (Tl) '(a. Note the close similarity of the echo problem to the impedance problem of the preceding section. defined by Eq.a) (7112) The last equality results from Eq. This is the variational formulation of the problem.' J"(r) . we have. 310 can be used to put Eq. in the vicinity of the obstacle..r») dJo(r) Then Eq. Define [r(r. dEo(r) _ (r(r.( a. we let TZ be z4rected and located on the.a ) and. .)' JO d. (7110). we assume a current J.r')] as the tensor of proper· tionality between a current element dJe at r' and the field dEe that it produces at r.356 TI~RARMONIC ELEC1'ROMAONETIC FIELDS For a stationary formula.:z: axis. subject to the constraint (a.a} in Eq. For linearly polarized fields. (330)..a). To express tbis constraint in a form for which (ala) is insensitive to the amplitude of J.::="" .(e. Jo(r) ds r E' = u.a) . The impedance problem is essentially an echo problem for which the source is at the obstacle. replacing (c.(i.a) .a) .. the echo area is given by (7114) If.a) = (1fE" Jo d. A commonly calculated parameter is the echo area. and then let r = ::c: + co.c) by {a. (7113) can be written as Echo """ 1f d. in Fig.[i11f E'(r) . (7113) into a more descriptive form.is the field produced by the assumed currents Ja.'7. A more general formulation of the echo problem can be made by replacing I l with an arbitrary source. that is. i"ll tJh 2~T = u.r')] Jo(r') This equation is in a form characteristic of variational solutions in general.c) by (a. (7111).a)' (a.
·oi" d... i" Therefore. J'd.LI2 fL12 dz' [(z)1'(zI) (k' + ~) G L/2 a.. L~o = 8 2f. The integral in the denominator of Eq.. approximated by Eq. As an example. Y....L: \\: 2 4 \ 6 .ide echo .OOO 12 10 Flo. Broad. = . we have A = A E. (7114). our stationary formula for echo area is A. by definition.fL/2 do . we have echo = E.PERTURBATIONAL AND VAlUATIONAL TECHNIQUES 357 2. wire. (7113) E J• = _"il..0 1.5 ~\Lla . 715. of .. hence from Eq. J. as represented by the insert of Fig. . 7]5. by Eq.. ~E." (1f> J.X .1600 o ~ . (7115) is just the selfreaction of the assumed current on the wire. m 1 1'"'20 •• ( \ Llo  " 150 1. Bu. Defining A as the selfreaction. (7105).) Also.'fll.polarized and x traveling. 1P E'· J'd.. This is the same type of reaction that we encountered in the linearantenna problem.rea 1L A. (1f> u.'  (7116\ .dB 'it' = <kj"'E ~. consider the scattering of a plane wave by a thin conducting wire. (7115) when the incident plane wave is . 1:. .5 .)' . (Aflu Y.. J'e d8)' '1'21\r 1f> E' .0 r'\ 0.
in kL) Si (kL) + log 2ykL . no. Y. IRE TraM.358 TIMEHARMONIC ELECTROMAGNETIC FIELDS where G is given by Eq. I.715 shows a plot of A.B [(z) dz = I ( k L . kL 2'in kL) [ Ci (kL) + log . vol.ves incident at an arbitrary angle are given by Tai. Electromagnetic Backscatlering [rom Cylindrical Wires. AP6. (7115) evalukL) ..Ci (2kL) + Si (2kL) 'in' (kL)) (7118) 1m (A)  ~: 1 + kL co. corresponds to the echo area of the threedimensional problems.kL cos 2 sin k 2 2 which defines B. .014. the quantity echo width L. 8.diation. t The results for plane wa. The integro.~: (kL + kL cos kL ./). In twodimensional problems. Appl. . Figurc. the same backscattered power density.catlering CroM Section of a Centerloaded Cylindrical Antenna.~n] (kL .. J. This solution gives good accuracy out to about kL = 8. no. 14. 1958.. (7118) and (7119). Y.! for the secondorder solution (two trial functions). Hu. I Ll' Ll2 k Hencc.' He also shows the effect of choosing different trial functions. I C. Ph". pp.cos k '2 (7117) Equation (7116) can then be evaluated as Re (A) .(I + cos kL) .. Tai. In equation form. by cylindrically omnidirectional ra. the echo area is A. the current should be practically zerOj hence we assume for our trial current j = cos kz L . (7104). At the ends of the wire.1' (7120) with A and B given by Eqs. Bu._ (2ro gl lim ~') (7121) 1 Y.8. H152. August. The echo width is defined as the width of incident wave which carries sufficient power to produce. T. vol.2 . pp.l in the numerator of Eq. the ceho width is L. Back. 23.in kL I (7119) where y ates to ZE 1. as calculated by Y.. For the current on the wire we should expect a constant current "forced II by the incident field plus a" naturalmode" sinusoidal current. = I~ I:. 009916. January.781.
.tionj L c = ~~ 8 .oJ. of a conduct. if the incident field is zpolarized and :J: traveling./ aI' 0.' = 1 (7125) Because the current is real."J.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 359 orj for linear polariza.. J.fJ$ dy = P where P is the complex power per unjt length supplied by J .. the result being P =. Going through a development similar to that used for Eq../2 E.fJ dy = J'/2 .lI respecti vely. Assume that the current induced on the ribbon is uniform j that is. . (7123) is J'/2 . 716.. 422.0 o 0. SimilarlYj (7124) (f J. the integral in the denominator of Eq. In both Eqs. if the incident field is ypolarized and x traveling. (7123) FIG. For an example of a twodimensional problemj consider a zpolarized plane wave normally incident on a conducting ribbon of width a.4 dl)' ..!. _. ing ribbon of width 4./2 E.. (7123) and (7124). J' dl .~rt is plotted in Fig."h dl)' • fE'.8 1. The echo width j according to Eq. (7115). This is illustrated by the insert of Fig.6 0.'dl "h  . ( IE'I') 2rp E' (7122) 6 E' tHi IT " It' / >• where superscripts 8 and i stand for llscattered U and H incident.1[21Z = at I• Y_~rt where Y. From symmetry. except that a line source is used j we obtain 7 0.. J L.2 ( J. 716. J" in Eq.. it is assumed that thc scatterers are cylinders generated by elements parallel to the z axis and the line integrals are in a transverse (z = constant) plane. 412. Echo width L. But we have already analyzed the ribbon oC uniform current in Sec. • 'f • 2X f EtJ. (7124) should have no z component.
scattering. be ealculatcd exactly. Hencc. Our problem is to obtain the variational formula for V r ./ FIG. 717. Then. The total signal received is the superposition of the inci· dent field. and obtain For large a we can use Eq. due to the currents c on the obstaclc.r) involves the freespace field of the currents on the obstacle. The more general case of differential scattering. plus the scattered field. We shaH here consider only the simple case of a pcrfectly conducting obstacle. the general case being considered in Sec. (339). the voltage across the receiving current due to the transmitting current is (7128) where t and r refer to the source or field of the transmitter and receiver. The problem consists of a transmitter. and a receiver at which we wish to evaluate the scattered signal. Transmitter Differential (7126) A plot of this is shown in Fig. which illuminates the obstacle. so V r' can. let us consider both the source and receiver to be unit electric currents. while a.r) is calculated with the obstacle absent and (c. due to the transmitter alonc. The transmitter and receiver currents are assumed to be known (they are current clements in our simplified case). in principle. Applying reciI A traDsmission problem involves the evaluation of the total field at the receiver. scattering problem involves the evaluation of only the scattered field. 711. .360 Receiver TIMEHARMONIC ELECTROMAGNETIC FIELDS . (7129) where (t. respectively. 321). (4107) (7127) which is also the physical optics approximation (see Fig.' is represented by Fig. 716. or transmission. according to Eq. 717. For simplicity.
(r. For our stationary formula.a.. ..}(t.c. we have V. we approximate (c"c. (7133) become (a"a. (> = [1ft (E~' . a:)' d8 (~lM) where (Ea)r is the field due to the assumed currents (J~.a. by Eqs.c) 1ft (E')· a. .. a:>. d8] [1ft (E')' . (7135) reduces to the formula for backscattering [Eq.' 1ft (E')· (J. Our boundary conditions on the various true fields are n X E = 0 at the obstacle boundaryj hence n X (E~.dB] 1ft (EO)' .c.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 361 procity.) = (CI.a.) = (Cr.).)".) . which is (a.. we have.V. where the a's denote assumed currents on the obstacle.) by (a. (7131) and reciprocity... Hence.4.o). (7131) where superscripts i and 8 refer to incident and scattered components. Note also that Eq.n X (E') .n X (E')' n X (E')' .) Gr. (7132). (765).a.' ~ (c.. which approximate the currents induced by the receiver.) = (a"c..(r .r) .) = (t. we have for our variational formula a) V .) = (c. and constrain the latter according to Eq.) (7133) In the language of t·he reaction concept.4. (7135) involves the assumption of currents on tho obstacle due to sources at both the transmitter and receiver.' = (r. (J. Eq.) (7132) where (er. By Eqs. and t and r refer to transmitter and receiver sources.) stands for the reaction between the field of the II correct " currents induced on the obstacle by the receiver and the" correct" currents induced by the transmitter...) (7134) Substituting from Eqs.4.c.')' d8  (c. Note that Eq. = (a r.) = . (7133) says that the assumed currents look the same to each other as to their respective true currents. (7113)J when the transmitter and receiver coincide.4.. (7130) and (7131). (7134) into Eq.. for the scattered voltage at the receiver. Eqs.a.(.')' tU (7130) where 0/)' is the surface current induced on the obstacle by the transmitter and (Ei)r is the field of the receiver current calculated with the obstacle absent (the incident field).) (a"a.
. no.) .a.. 742...)  JJJ (E~'. .<o)E . 19M.) = G(a"c.) Such con(7140) and we find G(c"a. I The problem of differential scattering by a. 193199.) = G(c"a. The extension to magnetic obstacles is given in Prob. Application of the Reaction Concept to Scattering ProbJelll8. R.(ar. (7137) as .. Using the relationship E' = E .c)  JJJ (E~'· 0')' dr (7137) where the notation is the same as in the preceding section.. there will be induced in it polarization currents given by J' . October.V.(E' + E') (7136) SuperscripUi tor r will be added to the various quantities to indicate that the exciting source is at the transmitter or receiver. respectively.s a dielectric body.) ~ (r. AP3.jw(. The treatment of differential scattering of the preceding section made DO assumptions about the nature of the obstacle in the derivation of Eq. (7·130) j hence for unit currents at t and r .. Note that F is symmetrical in Cr and and is actually the reaction between Er and aC)1 with the obstacle prescnt. dielectric obstacle is represented by Fig. we approximate the true currents c by trial currents a and set ..) Cl which defines the functional F.  JJJ .E .362 TIMEHARMONIC ELECl'RQAlAGNETIC FIELDS 711. 717 if the obstacle is now considered a. pp. We shall assume it to be nonmagnetic (}ol = 1£0). Scattering by Dielectric Obstacles.a..) . To obtain a stationary formula for the scattered voltage at the receiver. Cohen.) = G(a.c.. (765) applied to G. O'}' dT (7138) = F(c"c.al) (7139) subject to constraints of the form of Eq.(r. .Vr ' ~ F(a"a. but it may be lossy if E is complex. vol.'0')" 0')' dT  JJJ (E')'.E' and Eq. straints are G(""a. When the obstacle is excited by a source. 4.V. we can rewrite Eq. .. IRE TraM.(c"c. 0')' dr J (E')'· 0')' dr JJ (7141) Combining the preceding equations to render Vr ' insensitive to the ampliI M. (7136).)  G(.) ~ (t.
we have the variational formula _ V' _ (r. is found to be .')' d.a) = JJJ G JJJ E'· J'dr)' ''(J')'dr  JJJ (7143) E'·J'dr The echo area. . the echo width. . II .00 [fJI (E')'· (J')' dr] [fJI G')' dr  (Ery" (J')' dr] (7142) III (E')'· G')' dr For the lossy case.a)/l)' F(a.III. (7146) liE" J'd.a. if the incident wave is z traveling and zpolarized. (7~145) and (7146) are over the cross section of the oootacle in a z = constant plane. 'Inc surface integrals in Eqs.II ENN. (7114). (7143) by letting the source recede to infinity.F(a"a. For a zpolarized. we obtain (7144) In twodimensional problems. Using the definition of Eq. we have the backscattering problem. defined by Eq.'(J. (7115).. Ie = .'(J')' d. (7109) for echo.) • .PERTURBATIONAL AND VARIATIONAL TECHNIQUES 363 tudes of the trial functions.e' + 0 and Eq.a.a) (a.aj)(t.) (a.)' h . For a perfectly conducting obstacle.'e/ d. (7142) reduces to Eq."WE + fT. i hence . (7135). 'G')" t ! . defined by Eq. (7122). . we have Echo _  v/ _ I' «i. The steps parallel those used to obtain Eq. if the incident wave is :t traveling and vpolarized.  "(JJ J. 2: traveling incident wave. can be obtained from Eq. when the source is a unit current.. and (7145) L 7 . When the transmitter and receiver are represented by the same source.2X II .) .
28 Flo.24 1.00012 ~ hi .iollAl solution. 1.00008 ~  .) . I I +l". let us consider a problem for which the exact solution is available. The incident wave is z. . 718.(oJ V. (7147). take (7148) where A is a variational parameter to be determined either by the Ritz procedure or by the reaction concept. 71. r. The resulting series converged fairly rapidly.rder varia· ~ionalllOlution. r. (7148) is a better approximation than Eq. The integrations occurring in the various reactions were accomplished by expressing the exponentials and Hankel functions as Bessel function series.. .20 1.. ..364 TIllE~BAJU.polarized.tering by a dielectric cylinder (0) exact.04 1. 718. (a) Ie. (AfUr' Cokm. This very crude assumption yields curve (b) of Fig. the circular dielectric cylinder. aolut.12 '" .(b) V 0. While Eq. let us take (7147) where k .) . Scat. For our first approximation. according to Sec.. it is still crude. . 718.<Jrder variat.IONIC EL:ECTROlLAGNETIC FIELDS To illustrate the accuracy that we might expect from the variational formulas.1+/. as shown in the insert of Fig./ 1'. which yields curve (c) of Fig. and (c) aec:ond.. fj ~ " . An alternative procedure for treating dielectric obstacles can be given 0.08 1.00004 / o1....16 1. i ~ / ~ 0.00 / 1/ 1. is the wave number of the dielectric. For a better approximation.8. and the cylinder is defined by p "'" a = >"0/2.ion. 58. (b) Jirsw.
H"') for the cases (a).PERTURBATIONAL AND VARIATIONAL TECRNlQUES 365 in terms of cQuivalent currents over the surface of the obstacle. The currents induced on the screen thus produce no tangential H over the V .ion concept. we find nXE"'=nXEi over S.. no. For the complementary magnetic screen. n X HI over S. By the above four equations.= 0 over S. vol. Let S. Rumsey.. B. the sum E' + E. as shown in Fig. and (e). Transmission through Apertures. IV. be the screen surface of Fig. The precise interrelationship is shown by the following extension of Babinet's principle for optics.. On the screen itself we havc the boundary condition n X E. June 15. 712. and (e) radiating in the presence of a magnetically conducting screen. The Reaction Concept in Electromagnetic Theory. 719b. (The apert.+ H. An element of electric current produces no components of H tangential to any plane containing the element (see Sec. 719b.han one formula for the desired parameter. Consider the three cases of a given source (a) radiating in free space. following similar reasoning.. 29). and Rumsey discusses how to choose the best approximation according to the react. 148&1491.ure of one is identical to the obstacle of the other. (Eo.) Let the fields V > 0 be designated (EI. (b) radiating in the presence of an electrically conducting screen.. perfectly conducting plane is closely related to the problem of scattering by plane obstacles. Then Babinet'a principle for complementary screens states that H' + H" = H' (7149) proved as follows. 94. 719."" 0 over S. pp. respectively.satisfies nx(E'+E)=nxE' n X (H' + H) "" n X Hi 2 aer. hence n X H. The problem of transmission through apertures in an infinitely thin. n X H. overS. The electric and magnetic screens are said to be eomplemen/.0 plane.aTY if the two screens superimposed cover the entire V = 0 plane with no overlapping. 6. I This method leads t. over S. The total field in each case is the incident field E' plus the scattered field E' produced by the currents on the screen. H. be the aperture surface of Fig.Ho). and S. and (E"'. 1954.o more t.H'). PAil" Rtl1. . (b).
719d. . Illustration of Babinet'a principle. field has the same n X E as the incident field over part of the y . An alternative statement of Babinet's principle can be given in terms of the dual problem to Fig.). 719. ~n I Flo.n . so Babinet's principle [Eq. 719c.ced by K).EHARldO~'lC ELECTROMAGNE'l'IC FIELDS Hence. I Magnetic conductor * Dual source I I I IS. (7149») follows.0 (d) Is. Electric conductor (6) I I I Is. the magnetic screen source is replaced by its dual replaced by an electric screen.0 (a) 1 r+. ~n . If the original repla. He.0 plane and the same n X H over the rest of the y .and H numerically equal to E'" (see Table 32). If the field of this dual problem is + a I I I I I Electric conductor S. 33). I I I . 0 plane. "10 Is.. shown in Fig. the e m. EM.0 (0) ..366 Tl!ld.. and the medium replaced by its "reciprocal" (11 by 1/. then E will be numerically equal to H. These conditions are sufficient to determine E. I I t Source I S. H in the region y > 0 according to the uniqueness theorem (Sec. Hili. EJ.. '10 t I I I I I I Source Is. HI 1JO E.~O ~n S.
Transmitter ./ Receiver Flo.. (7135).g"t. we have at the receiver H. The trnll8mitted field E' of (a) ill equal to the scattered field E' of (b).E" = Hi (7·150) The problem of Fig. Babinet'e principle [Eq. =  [fJ (H')'· (Moo)' dS] ff (Ho).E) X n = 2E' X n (7152) where E+ and E.·1.denote E in the regions y > 0 and y < 0. stationary formulas for. = (E+ . and 1. dual to Eq. 719d is more easily approximated physically than is the problem of Fig.the signal at a receiver on the shadow side of a screen are of the same form as the stationary formulas for the scattered signal at a receiver in the complementary problem. 720a and b can be expressed as . respectively. It approximates the true magnetic current M. ". In Fig. R". (7149)} becomes H'" . The direct application of Babinct's principle to the problems of Fig. 720b. 719c.c cood"to.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 367 Electric conductor E' M. let the sources at the transmitter and receiver be magnetic currents across the "terminals" I. I (6) E'+& 1. and n = U lt• The interrelationships between Fig. 720. denoted by E".. Then.I.1.. Hence. 720a and b shows that the field transmitted by an aperture in a planeconducting screen is equal to thc negative of the field scattered by the complementary obstacle.. (Mt)' ds [fJ (H')'· (MtV dS] (7151) where Moll denotes the assumed magnetic current on the obstacle. Receiver (0) Transmitter "I..
that is.perture. we must assume an n X E in the aperture due to (H')' alone and due to (H')' alone. If ~ and 1. (7154) Electric conductor Incident Complete electric conductor plane wave ~ M.d. 721. Another quantity sometimes defined is the trammisrion area. We define the tranfmission coefficient T of an aperture as the ratio of the power transmitted through an aperture to tho power incident on the aperture. .ic current elements across I. are images of each other. to apply Eq. 721a. from Eq. as shown in Fig. Sometimes it is the total power transmitted through the aperture that is of interest. {a) Transmission through an aperture. (H~" (n X Eo).is the magnetic field calculated from the E. and (b) equivalent problem for the > o. 720. D . We shall explicitly consider uniform plane waves normally incident on an aperture in a plane screen. Note that T depends on both the nature of the source and the geometry of the a. (n X Eo). Let the incident 0 Hi.A. and.0 'a) region 'V .'"ETIC FlELDS Hence.. [JJ II <k] [I1 (7153) where E. ] H'· t.368 TDlEH. and l~. (7151). (7153).IU10NIC ELEcraOMAGr. then the aperture problem becomes the same as an echo problem.(Ho). T _ Re If E' X H'··ds Re If E' X ds apeR ~.• = ~. which is the transmission coefficient times the area of the aperture. as they appear in Fig. (n X Eo)' d. we obtain for the aperture problem (H~'.0 (b) FIG.is an assumed field in the aperture and H. . The sources of Hi are magnet. because of the symmetry of the plane screens about y = o.
per~ . our stationary formula for (c. Hence.' dS)' II H' .c) by (a. which for M.a)' _ . 721a in the region y > 0..a) according to Eqs. 721b. M. so by Eq. (7156) n X Hi is real in the aperture. We have (a.. in the equivalent (7159) II M•.' ds (7160) where H" is the field of the assumed current M. (c.PERTURBATIONAL AND VARiATIONAL TECHNIQUES 369 (7155) wave be specified by H' = Ueih E' = '7H' X UI/ where u is any unit vector orthogonal to principle..a) . (765). we noted that in the aperture "1/' In the proof of Babinet's (7156) n X H' . = Re ff E' X H'· .c) where nIl sources radiate in the presence of the conducting plane.a)  (fI H'· M. Hence.a) .. (7161L we have T = Hence. we approximate (c. For a variational formulation.pen Now consider the problem of Fig..=E'xn is equivalent to Fig.a) and constrain (a.c) is (eel . (7155).Re (7158) lIenee.n X H' because the currents on the conducting screen produce no tangential components of H in the y = 0 plane. ds ~ Re (c.c) = (c.c) ~ (a. H' .(c.A (7161) where A is the area of the aperture. (c. (7154) to 17A "'!Re[W ff u·n XN8)'] (7162) H4. we have the power incident on the aperture given by <!" .(a.c) where (c. <l>. and (c. that is. <!" ~ .(a.c) is the selfreaction of the correct magnetic currents radiating in the presence of an electric conducwr covering the entire y=O plane.. ds = Re ff EI X H' • ds (7157) .. problem. n X Eo dB .a) by reciprocity. combining Eqs.a) can be calculated because we know n X He o:z n X Hi. Equation (7155) chooses Hi to be real in the y "'" 0 plane. For the incident field of Eq.
Re r. incident.. 422.ve is polarized transverse to the slot.6 where E· is the assumed tangential electric field in the aperture and H· is the magnetic field calculated from E· by the methods of Sec.~. (Alter llfiJu.. 'y' I'l~_ "a [U dJ)'] (7163) where a is the width of the slot.2 E. &ll. for a slot. 722. 422.0 0.. (7163) reduces to T = (7164) where Y. (7165) From Eqs. 3~... I u·Eo X 7 . 411 we defined the admittance of an aperture as y.:::. Hence.""•• = G. (7162) to a unit length of our twodimensional slot. (7162) is II HO • n X Eo dB = (ff Eo X Ho' • ds)' 10 Sec. 7·22.. 0.5 h 1. hence we take Eo _ I in the slot. If we assume E· in the slot to be real.8 1. When the incident wa. have for smaH a + jB" is shown in ~ "a He (+) y.5 o 0. I~I' ff EX H' ·ds and calculated it for a slot for particular assumed E's. (7167) . wave polarized traosvcnro to slot. we have the case of Fig. TransmissiOll coefficient. we have . As an example.4 0.X H·)* 01" FIG. kolog ko r' and from Eqs.ted conductor.0 TI1lEHARMONIC ELECTROMAGNETIC FIELDS \.r.U!.370 2. u. applying Eq. (4106) we (7166) To.s shown in the insert of Fig.""•• Fig. then E" X H"* = (E. (4107) we have for large a T_I .V .t 1 "'act Variational 1 I 0. let us consider tbe twodimensional problem of transmission through a slot....0 1.. Now Eq.) and the denominator of Eq.
corresponding to &t. August. (73) is IT. defined by nXEZHj I Morsc and Rubenstcin. "'"'0·. 1938. 54. . so to make use of the analysis of Sec. T 423.tion (7163) then reduces to T = 4a 1r. 814). we should not expect the trial field of Eq. Equa.. Show thlLt thc pertllJ'batiolllll formula.. J?etJ. PhYI.85 (~y For large a we should expect the field in the aperture to be uniform.t = G have for small a d + jB. 7~14) from onc having perrectly conducting walls to one having a wall impedaoce Z. The case of a.. 760770. December. can be interprcted according to whcre . 20. which can be obtained by solving the wave equation in elliptic coordinates l (Fig. we have the case of Fig. 00. t. t If the incident wave is polarized parallel to the axis of the slot. jdbH X Eo·ds _ Note that both wand <. vol. J. 1949.. say a > X. From Eqs. Equation (7169) actually approaches 0.E' E. The Diffraction of Wavcs by Ribbons and Slits. 722).. Appl. plane wave at an arbitrary angle of incidence is considered by Miles. 8. pp. instead of the expected value 1. fI (7169) where Y. 71 cOlltaiu lossy material characterized by /l." Re(_1) Y:. is shown in Fig. pp. ... . The variational solution is compared to the exact solution. Hence.. PROBLEMS and 71. Suppose the cavities of Fig. 411 we would assume Ed = cos .. vol. (7168) to give good results for large a. tJ. a (7168) in the slot. 895898. 11. Consider the perturbation of a cavity (say Fig.PERTURBATIONAL AND VARIATIONAL TECHNIQUES 371 This last result is the geometrical optics approximation. ///I..to must be complex. no. On the Diffraction of an Elcetromagnetic Wave through a Plane Screen. Miles.81 for large a..H· Hold. 423.. (4115). we (7170) ~ 6. A complex resonaoce ill the lowloss case ~CT~lf~. Phy. W. is the rClll resonant frequency and Q is the qunJity factor (see &e. 72..
E.. 75 for the ease IT .o.Ho .lB./2Q. perturbational formula is j 'LB. and let E .nee duo to metal walls is (ft  "'r . +. _ and f + 6t.372 at the wan.l. 1" Use the results of Prob.'.K'!f. Col ..0. and that Eq. dt 7i'T. .. to show that •• . '16. • H.0.b we and Col are complex if" and IT + 4lr are not identieaUy zero. and show that the perturbational formula gives Note that the formula for Q is identical to the one that we have been uain& if Re ('f). Show that the perturbational formula corresponding to Eq.ld. (7ll) still applies witb Col changed to """ .jlH.I dr • JJ .Jdr ColO .is real but (If iJ complex if Z haa a real part.j. + .v in addition to I.  [(..•• WI 1 2Q where "'I is the resonant frequency of the cavity with perfcctly conducting walla. . it the intrinsic impedance of the conducting walls. 73. Jl + 6/10. TlUliHAIDJONIC ELEC1'ROlolAGNETIC FIELDS Show that the exact..JiB· H. Suppose the cavities of Fig. JJf f [Co jtM!Col)E·E.ed by IT and <T + t. and show Lhat the fractional change in reftOn. where the aubecript 0 denotes unperturbed qtw1tities. where .!L 1ft "  WI  fff (oEE.. (710) is then W . Use the resuIt. 72 and the approximatioIlll to show that Use the relatiollShipa Z lSI.!J.JiB·B.s of Prob.Col. 72 &ro chart\Cteri1./w)E· E.HH. Note that "'.'I_B.0 Bot.... Q " fff "lBoi'd. H .l. Use the result of Prob. 76.. 73.I_' / d..
Show tbat tbe change in the resonant frequency of the domi· nant mode due to the introduction of the dielect.. &..BATIONAL AND VARIATIONAL TECHNIQUES 373 7~7.4IS W. Suppose the wall is s1ir. &8 .PERTUR.hown in Pir.l . . 110.. 2A (1  ~ ) wbere A is the crol5lJfJC(ltional al'(l& of the cylinder. Conaider t... For tbe ypolariloo dominant mode.. and Q... .polariJ:ed (E in the center pointe in the % direction) dominant mode is Aw.." is the complex permiLlivity of the sample and w.. Show that the chanr.42 .ngular cavity with a Imall centered dicleet. 723b.. Considcr a rectangular cavity with a Imall centered dielectric cylinder. Area A Use a quasiatatie approximation.841/a". .. A1 ..e in cutoff ffflquency for the z.btly fl&tt.6Or 2(/f" '" where l .  1If. 79.he rect:t.a. respectively.'.1. CollJlider the circular waveguide of Fig.Q. 7·23. &2. I I I 1 (0) fo I· (b) Flo. as Show that tbe cb.. _ 90. shown in Fig..be Col  "'. 7230. is the perturbed telIOnant frequency. Suppose that a small sample of 10ll8Y dielectric is introduced into a..rie is 1 1 I . .obc.nge in the re800AIlt frequency of the domi· nant mode due to "the intro<luctioQ of the dielectric is WWG d'. [f the 101l8ell of the unperturbed cavity are significant.y whose unperturbed resonant frequency is 101..+2 . ". is the unperturbed cutoff (requeney. 101 where d is the diameter of the sphere. then QQ. Show that. are the Q's of the cavity with and without thc samplo. clwit. where Q.1... _ O.. Use a quasietatic approximation.0 wbere A ill the erofl8600tional aroa of the deformation and w.ened at the point. 78.. A .0' Bence. the mode degeneracy has been removed.ric sphere. Rectangular cavity with (0) dielectric cylinder and (b) dielectric sphere.
•• where We _ rlb. small centered dielectric cylinder in II. T 1 0 I' b 'I jd}(0) 0 T 0 I' b ~ 'I 1 ~ Idl(b) FlO. Consider t. Consider the shown in Fig.od a qutulis1. 724b. Consider tho rectangular waveguide with smaJ.y.ional method and quasistat. according to ••• •• ~ 2ob . given in Frob .he bottom covered by fI.2r/b Hence.t. .. where can be taken .. the same 88 the first given in Prob.atic approximation to show that the dominantmode cutoff frequency differs from the TEo! rcetangu1:l.. Rectangular waveguide with (4) dielectric cylinder nnd (b) conducting ridges. 724.=)'" rdle.=. perturbational method I!l. 713. 88 the cutoff frequency of the perturbed guide. as shown in Fig. Usc a perturbational method and a quasistatic approximation. show that v'l (J. ~ "d' "'e 4D..'il.urba. /6 y".....r guide cutoff. rectangula. Show that the change in cutoff frequenc)' of the dominant mode from that for the empty guide is ~ . rectangular waveguide. Show that the ncxt higher mode (b ~ 2a) cutoff frequency differs from the TEo...In' is the emptyguide phase constant.k. Use a. all Use a perturbational method and quasi6tatic approximation to (1fJ.~ 200.r guide cutoff.ion is increased. sa rectangulnr waveguide with 8..ic approximatiou to show that the phase constant is w.1 semicircular ridges. 46 wit.+. centered dielectric cylinder. the mode separat.d' where l<l. Note that this is term of an expansion of the exact characteristio equation.610>. Figure 7240 shows 8.hin dieleetric slab (Fig..h d Sa). :v'~"l=~(. . where (10 .l 1 til. according to .. t. 414. 712. 7240.374 TIMEHARMONIC ELECTROMAGNETIC FIELDS 711.b where "'. 114. U8C 8 pert..he rectangular waveguide with t..
.... . E ..urbed guide. and comrider .do Show that..\U'batKJD of illl pcrfect.his reduees to Eq..11 .urbation of material in a lossy waveguide from I. Consider t. (730) ia . .. E: and H .he resultant.. and show t.. (276). ..PERTURBATIONAL AND VAIUATIONAL TECHNIQUES 375 711. H. t.(J( + j~.. Ae'" Note t. .) 'u.h 108lIY material.u cutoff frequency /J .urbed TE.ly conducting walls.. if u il cloac to u 4 • Sbow thA\ at the unpert....e directions of propagation..  II s' .hat. Consider the perturbation of the walla of a waveguide from ductor to an impedance sheeL Z.+ 6p.fJ(••• .2 xli. to Eq. (7. )2. and let..he fielda lUI in Prob.1 do IJ (t. 110 t.hat. .H perlcet con Represent the unperturbed and perturbed fields a. Show that.. let Z .. approximation for /J is Eq. (729) in the losefree case. to . of the pert..) 'u.. 8uch t.(...he propagation constant. t. + two Represent t.11. and show that 719.. 715.he pert. H _ E .. .s in Prob. ..hat the formula corTCllponding 'Y .. "'" a + jfJ.d4 D.  toe. ReprC8eot the unperturbed 6eldl (lUbecript OJ and tbe perturbed fiddll (no lubscripto) by E. Suppose t. 716. 1/. t.  1 .. k.:30) in the lOllSfree ease. . ~... In the pert. Uee lobe reeulta of Prob.he oppoeit.a + j/J..t xli...~ Alii. X 11..hat a waveguide ill filled wit.c 1:o Xft'ndl ct. Note that this is an Approximate form of Eq.urbed guide by .. I/. (733) aod a " _~IJT"_III. pert. and let E . 718 and let the unperturbed guide be lOllSfree.do Show that thia reduces to Eq.he unperturbed guide be loafree. 716. 715.hAt. Use the reeults of Prob.. to I + 111.x:a:.. 711..£e. t.I'_d'_ 2Refft./J.. nXEZ.. 718.. 716. Denote t. 'f ..... (729) ia Show t. X 11.u...&:.he formula corresponding to Eq..
722. the above formula for a is the approximation that we have been using to calculate attenuation in metal wAveguides.lll. Consider a arnaU ddormation of the walls of a cavity.17.'" If Z .376 TIMEHARMONIC ELEGraOMAGNETIC PIELDS H . Vfij ~Ol + cl be ... as a trial field. Compare with the results of Prob.IBI·d... provided n X E _ 0 on S.. (295)1. Figure &31b shows a partially lilled circular cavity. ~z(~ .'Iv ... Show that.. ."': •: .... __ . X EI' d. but is not stationary if 108!le8 a~ pre&ent.  ~/ /"/. and abow that (J  fJ.y field E. /// (. f "'19. 1. . Show tluLt. X a: ·u. (744)).. (74).... trial field H.IB. f 1ll19. (746) and . such all represented by Fig.I'd. In the exact 1lO1ution (Eq.u. with no boundary conditions required on H. CoOllider the rectangular cavity (Fig. .') to show that the dominant. F~. Show that thia formula is essentially the same as Eq. Use a trial field E . 71.c) and abow that.. (746) is a stationary formula for . is a.01) gives w. 219) and the stationAry fonnula IEq. UIMl Eq. which requires no boundary conditions on E.I') d. (74.. a:. '~''''77/ / / . the numerical factor is 1T inatead of V'iO. 721.b)(z .. Tako tho variational formula IEq.. Show that / / / .l· . (745)).. 2. 723..u. Eq. '1111. 7110. Eq. 624.405~) 1 _ ~(..J .. stationary formula for the resonant frequency of a lOIllru cavity...da 2Re / / t. mode re8On&nce ia V.1' JJ t. the intrinsic impedance of metal walls.11. (745) reduces to Wi .1· •• A... and take the unperturbed cavit.405 a 1 (2..dl dl 2Re Xa:'u.IB.
Consider the partially filled rectangular waveguide of Fig.IIV X E) discontinuous at 8. Determine Al and AI by the Ritz method.." 3o. (744). (78)1. cutoff frequency. and . Show tbat no boundary conditions e. that the stationary Eficld formula is where subscripts 1 and 2 refer to regions] and 2 (Fig. Considcr the rectangular cavity (Fig.11 u~smbsmc . H o as a trial field.U r A ISln1)COS ·. on Sand n X (plV X E) continuous at a.I. Repeat Prob. Consider a perturbation of material in a cavity. Use variational formulas to approximate the lowest. given by replacing E by H. • by .y'.l.he unperturbed cavity field Eo.ional Connula lEq. auch as represented by Fig. '" u. Show also that a variational solution in terms of trial fields satisfying n X E . 725).. 727 is of the same form 9.. Consider a waveguide whose cross section is an equilateral triangle of aide length 11." u'lcosb'Bln. . a.+A c ry.. is given by Eq. the surface S represents a perfect electric conductor enclosing a cavity. using the reaction concept of Sec.. (772)]. (772) then reduces to Eq.r. and show that the resultant formula for "'. 731.SID'B . A variational solution is desired in terms of a trial field aatisfying n X E _ 0 n I n "'I (1) I (2) t.nd take t. Show that the variational Hfield formula.I by t. Use the Eficld variational formula [Eq. Trial fields are disconLiJJuous over a. Sbow that Eq. 77. but with n X E discontinuous at a. 729. but with n X (. . 726. Z c where AI and AI are variational parameters.. In Fig. Choose a trial ficld E H . 725. 72. and the trial field E . for Prob.0 on Sand n X E continuous at a. 725. 219) and the mhedfield variat.y ". is the exact formula IEq. s Show FlO.. (772)1. Why do we get an exact solution in thia case? 727.. (295)1. Take the mixedBcld variational formula fEq. 4&. (711).PERTURBATIONAL AND VARIATIONAL TECHNIQUES 377 7·2l5. The exact solution ia . 728.l.l. y .t S are required in the Hficld formula. 726. 730.8 the above EBeld formula.
. 732 reducetl to Eq.. represented by Fig.!. Show by the variational method that. 734. No boundary conditions are required in the above formula.e waveguide feed 8yst. 75« and b. p 736. and show that Eq.I od ~Ioad'1111 matched ~! . (785) and a perturbation of matter in a waveguide..1 Note that only one new reaction is needed in addition to those obtained in Probl.z) Obtain the input impedance secn by the coax by the variational method.log . 735 and 736. Use the reaction concept to derive the mixedfield variational formula ror waveguide phase constants which corresponds to Eq.781.". Figure 726 shows a coaxial stub to parnllcl·plat. H.. 726. .2 7.wave scattering by a conducting ribbon. ____=j=. 736 for the sccondorder variationalllOlution. 700 and c. assuming trial currents /.em. remove the restriction on . 735. &8 illustrated by Fig. Speeialhie the result to II _ "/4. 49). Repeat Prob. II matched load FlO. 73li.eo that fL rensonablo trial current is a uniform current. Aasume II « ".. Use the unperturbed field E.tub and assume a trial current on the I . _ ! a [1 + ~ (!! 2 r a .cos 1(4 . 737. 782. lUI a trial field.he coax is • Z _ ka 4 (1 .:I) /. Consider the variational formula of Eq.. Consider tho twodimensional problem of planc. (732). but with the opposite polariZl\tion. T38. (185) reduces to Eq. (785) if n X E _ 0 on C. 753.) . the impedance !een by t. Coax to parallelplate feed. and ahow that the formula of Prob. 4 where l'  1. Bin 2lt'd)J~ flo Compare BOrne calculated points with the exact eolution (Fig. 716.. Use the unperturbed field E" H. In Prob. 73'2 and II perturbation of waveguide walls.COli 1:(11 . (733). shown in the insert of Fig. as a trial field. Consider the variational formula of Prob.378 and show that TWEBARMONlC ELECTROMAGNETIC FIELDS Col.
.d. 0.. Hi ia pa. 7tO. 73 The imaginary put of (0... )... Repeat Prob. Why should we expect the above formula to be inaccurate for large ka1 1·59. 14.) ao. show that the backecattcring area is A • . and we know that.. (7114) with E' evaluated in the I' direction.. 423.86)../2). r (to T L 1 I Flo..how that the variational solution ia L • 'Z _ 1 32"1 "Y. 740 for the ease of differential ecattering./2). showing that the differential echo area ia A• . At the first resonnnce (L .. Using the approximations of Prob. Show that lLS la . (lJOO Fig. 224) cO. I' r~ where" Y. 0.". Figure 727 represents a re8Onaot length of wire illuminated by a uniform plnne wave at the angle 8. at resonance the echo area is A• . (1115). ill defined by Eq.'ilQ1JE8 379 'In other words.86).66 times t...1 This is relatively inaensitive to the diameter of the win.)']' ooa !co. 715. .' [ ~. (~ao. represented by Fig. the current ilJ 1.1 [ 'o..cd for naonanee.! b (0. (~ao. u~ cos and .he physical optics solution. ~ receiver) r"(to transmitter) Apin this is relatively inaeollitive to the diameter of the wire. polarized in the rz plane. ")]' SID " lIln ' " whl!re . 739.0) . Consider planewave scattering by a wire. ).. Using Eq. show that. Ulle the trial current ] ..PERTURBATIONAL AND VARIATIONAL TEClL.0) is BelO becaWle the length is adjUllt.86).1. zl ~'" L. 0. 727..pu~ ia given in Fig. Scattering by a resonant wire (L .rallel to the aria of tho ribbon.. this answer reduces to 0.
. across cd. The aperture antenna. M') d. (a) A sheetmetal antenna.'. fed . 739...'(J')' . Figure 72& represents a metlLl. Consider differential scattering by a magnetic obsf.· J' . Let Z". (a) (b) FlO. MG.ntenna. (7..(a.a) (a. M") dr In the above formulas. be the input admittance to the slot antenna.a) ('.a) FCa. '(M')') d. . similar to those or Prob. J fJ JJJ (E. and use duality. Hinl.aele (Fig. antenna cut from a plane conductor and fed across the slot abo Figure 728b represents tho aperture formed by the remainder of the metal plane left lLftcr the metal antenna was cut. Show that the transmission coefficicnt ill T "" 0.. and (b) its complementary aperture antenna.a) . 717) and define Show tbat. and show that Hint: Consider line integmls of E BOd H from a to band c to a.: Use the result of Prob. 728. is said to be complementary to the metal a. JJJ (. be the input impedance or the mctalllntenna and Y.380 TIMEHARMONIC ELECTROMAGNETIC FIELDS 7i2. H' is the incident.a) . 7'U" Consider a narrow resonant slot of approximate length "'/2 in a conducting 8creen.a}jl)1 F(a.143). EI. field. 743 and IUIsumptioD8 . 743. Jft and M· are the assumed electric and magnetic polarization currents on the obstacle.KG.52w where to is the width of the slot. instead of Eq. and E". (E·' 1'> . we have Echo _ where «i. H" is the field from J".H' .
1951. general solutions for the field in a homogeneous region can be constructed from solutions to the Helmholtz equation (81) In cylindrical coordinates. sec. Several special cases of the cylindrical waveguide. McGrawHill Book Company. Marcuvitz. "Waveguide Handbook. 381 . and k. Cylindrical Waveguides. We now wish to give a general treatment of cylindrical x s (cross section independent of z) waveguides ''c consisting of a homogeneous isotropic dielectric bounded by a perfect electric conductor.i'(x. vol.' = kt (85) and 'Vr is the twodimensional (transverse to z) del operator l". 10. arc related by + k. Our formula.t'!' + k.. 0 dtZ + k tZ = 0 dz t • ke t (83) (84) where the separation constants k. 312.y)Z(z) (82) The resultant pair of equations are 'v'.CHAPTER 8 MICROWAVE NETWORKS y 8 1. Ino.. New York.l"u a . ' '!' . 12. already have been considered. this equation can be partially separated by taking '" . Cross section of a cylindrical waveguide.tion of the problem will be similar to that given by Marcuvitz. Figure 81 represents the cross section of one such waveguide. 1 A1J shown in Sec. n such as the rectangular and circular guides.az (86) IN." MIT Radiation Laboratory Series. 81. FIG..
 (87) which.'Jt'" (superscript m denotes TM) and. and rewrite the above as I dZH. we determine (812) Defining the transverse electric field vector E. The k. by Eq. The boundary is perfectly conducting.'" = .' = . it is evident that for E.i") d JWIJ Z (810) (811) It is evident from Eqs. 1 dZ' E. dual to Eq.. = H .'1")Z' where 1is the unit tangent to C and n is the unit normal to C (sec Fig. we have. are determined from Eq. we take A = u. to vanish on C we must meet the boundary condition on C (814) . hence E l = 0 on C and a".(V. For TE modes... (811).II..382 TJi\lEHARMONIC ELECTROMAGNETIC FIELDS Solutions to Eq. we define a tran8verse field vector as H.k.+ u . (88).' _ 0 an on C (89) The associated magnetic field is given by 1 1 H" = . (84) are of the general form Z(z) . we take F = u•.JWp.Aef'. is a superposition of +z and z traveling waves. + BeI'. (85) aIter the k e (cutoff wave numbers) are found by solving the boundaryvalue problem.p (superscript determine 6 denotes TE) and (88) The component of E tangential to the waveguide boundary C is E 1" = 1· (u.. X VI'!") = (n ." u. dual to Eq. 81).u. are everywhere perpendicular to each other. real. iJxiJz ayaz W For morc concise notation. For TM modes. for k.v x E' = .~1f·) a'". JW~ (a'f. V.+ u.(V1'I'''') d JWE Z (813) From the second of these equations. (810) with H replaced by E. (88) and (811) that lines of 8 and :Je.
]. comparing Eqs. the necessary boundary condition is '1'. It is convenient to introduce mode !urn.·Ad.· A "'"' V. and mode currents I(z) according to E' = eoVo H.i'u. e.. . It should be kept in mind that Eq." =. I_dZ· (816) JWIt dz foc TE modes. E. I (818) I where the integration extends over the guide cross section. II V. (812) and (813). II (.= b. = O. and let A .k. all amplitude factors are included in the V's and l's.i" = h' X u.dz  h. These modes can be suitably ordered. z. (83) subject to boundary conditions is an eigenvalue problem. = 0 on C. to :Furthermore. X V.constant on each conductor. The corresponding ficld is TEM to z and is a transmissionline mode.V + i'·Vli' d. (815) with Eqs.  Iv. we normalize the mode veetors according II (o')' d. we see that we may choose Vo .. (814) also satisfies the condition 1 .tions e(x.u. We shall now show that all eigenvalues aTe real. and.'I'I'I') ~ '1" : : dl .'I'I' . it is possible to have k.\{I.'I'·· V.X u.II (h)'d..)'d. In this case.k/j"irl l and the divergence theorem becomes ff (IV. Hence. giving rise to a discrete set of mades.hOI" (815) Comparing Eqs. and the various equations of this section then apply to each mode. (815) with Eqs.i'. When the waveguide cross section is multiply connected. = z (817) for TM modes. mode voltages V(z).V.'!'11 . . such as in coaxial lines. 1 v= jwt dZ.i'·v. Then. eo "" u. (88) and (811). Note that EQ.ndl "'" v. Consider the tW(r dimensional divergence theorem ~A.Xe 1.y) and h(x.MJCROWAVE NETWORKS 383 provided Icc ':F O.y).XV.II (h')' d.
caD differ only by a cODstant.P .0 or (Jllfjan = 0 on C. (820) i  121l"Vf~ (821) Then. Then the Helmholtz equation is V. There is also no loss of generality if we take all eigenfunctions .384 TIllEHARMONIC ELECTROi\(AGNETIC FIELDS But the boundary conditions on the eigenfunction "if are either Ilf =. Hence. !! IM'I'd. u and v for a particular k. and '1' is in phase over a. These are. ~ . (822) < j..l real.VI • . The boundary conditions. !!I<I'ds (819) The eigenvalue kef is therefore positive real. of course. . and let '1t = u + jv. 2'1' + kc 2'l' "'" Vb' + kctu + j(V.tv + k/v) onC = 0 which. To justify this statement.. ~ (823) . just the relationships that we previously established for the rectangular and circular waveguides.. . the righthand term vanishes a.r. For f and J.y to be real. suppose ir is not real. guide cross section.nd k' . we have the propagation constant given by j j > j.''. Hence. we have a cutoff wavelength ~. either 'I'=u+juO or an a'1' = oU+j{Jv_ O an an on C are satisfied independently by 'U and v. Figure 218 illustrates the behavior of a and {:J versus f. represents two Helmholtz equations for the real functions 'U and v. Let us now look at the propagation constant 'Y = jk. so u and v arc solutions to the same boundaryvalue problem. the concepts of guide wavelength. When the mode is propagating (f > fo). and a cutoff frequency t k.=iTjj'<i (Mf)' 2. We caD take it to be real and include any phase in the V and I functions. (85). from Eq. since kc l is real.
the analogy . . and guide phau velocity. "'" .. Figure 43 illustrates the behavior of the Za'S versus frequency.k./fl' JW~ f > f. (826) where the cMraderidic impedance Zo is. are useful. and. W _. and (817) it i.). v. _ JWE JWE (J)' k.JoLICROWAVE NETWORKS 385 where X is the intrinsic wavelength in the dielectric. Z. (87). we can show that V and 1 also satisfy the tram missWlIrlim eqootiom (829) where Yo=. from Eqs. from Eqs..Y ' + Ve Y ' 1(z) . Turning now to the mode voltagea and currents. Hence. in general they are of the form of Eq.s apparent that V+ [+ =0 Zo V 1'" Z. J+er· + IeY ' (825) where superscripts + and . (84). we see from their definitions IEqs. Hence. /1 _ L)' V Ie f < f.. y l w.denote positively Bod negatively traveling (or attenuating) wave components. f > f. and (817). for TE modes. for TM modes. 27. :L JW' 1 ~. Note that these are just the characteristic wave impedances that we previously defined for rectangular and circular waveguides. ·~I . (84). (816). Also.1/Zo is the charaderi'lic admittance. or V(z) "'" V+e. (827) Q" ./fl' (824) where tI.Vi JW~ •U. Finally. P Vi (f. (816). VI  (fIf..2. These parameters are discussed in Sec.' _ jw. (84). (816) nnd (817)] that V and I satisfy Eq. f < f. (828) Zo. is the intrinsic phase velocity.
386 TIKF.UAJUlONIC E. see Wilbur LePage and Samuel Seely. New York.v'ZY . with transmission lines is complete. 82. Such an equivalent circuit may help us to visualize waveguide behavior by presenting it in terms of the more familiar transmissionline behavior. .. the transmission line equivalent to a TE mode is as shown in Fig. . labeled in ohms. transmission lines Cor waveguide modes (IICries element. "General Network Analysis. we obtain jyXB Equating the above Zo and 'Y to those of a TE waveguide jX "" jWIJ ) 'B =)~ . + . and all of the techniques for analyzing transmission lines caD be applied to each waveguide mode. and Zo are t.LECTROYAGNETIC FIELDS I joI' i VI. (0) TE modes. Inc. For a dissipationless transmission line. Thus.' I I I I I~ dz (0) ~/jOH I I I I I I I I I jOJI£ I I I I I iI dz (b) 1I I' ·1 I FlO. 26).. k~' 1"" jB = jWt (831) I For u&mple..9 and 10.k. . (6) TM modes. we have Zo ~ = /Z IX VY = VB (see Sec. mode. l We may define an equivaknt trammiuion lim for each waveguide mode as one (or which." Chape. McGraw_Hill Book Company.' JW~ (830) &2a. 'bunt elements in mhos). 1952. Similarly.he same as those of the waveguide mode. for a TM mode we obtain ]X=]Wp+. Equivalent.
H the dielectric is lossy. in Eqs.. Keep in mind that all of the equations apply to each mode and tha.'I". ._. . of course. V(mb)' + (nap SID a:Z: SID bY 't" . and lines of instantaneous 8. The result is 1 'l'••C = _ T _. Those equations common to both TE and TM modes are written centered in the table. Recognizing that the gcneral exposition of cylindrical waveguides has been quite lengthy. Lines of instantaneous :re. is proportional to v.. For the rectangular waveguide of Fig.  VI· = Vl· IJ e'da JJ e X h··ulda (832) (833) = VI and the timeaverage power transmitted is /P. Similarly. for TM modcs. . Hence. SimilarLy. we can recognize the equivalent t. pictures of lines of Sand :JC at some instant. let us summarize the results. It is therefore quite easy to sketch the mode patterns directly [rom the eigenfunctions 'It.. let us tabulate the normalized eigenfunctions for the special cases already treated. for future reference. lines 0/ constant '1" are auo linea of instantaneous S. (419) and (421) and normalize them according to Eq. (818).YICROWAVE NETWORKS 387 The transmission line equivalent to a. for the . obtained by replo. Finally. the equivalent transmission will also have resistances. and E is perpendicular to H. (mb)l ab. obtained by integrating the Poynting vector over the guide cross section. can be obtained directly from the v's. are everywhere perpendicular to lines of instantaneous 8. For TE modes. for the +z direction. It is also worthwhile to note that the modc patterns.Re (V[') Hence. lines of constant 'It. power is calculated by the usual circuittheory formulas.cing jWE by (1 + jWf. + (na)' cos (rnT) cos (nT) a x bY (834) where m. 82b. that is. = JJ E X H··uld.. TM mode is therefore 88 shown in Fig. The power transmitted along tlie wavcguide is. Hence. H. we can pick the w's from Eqs. p. Table 81 lists the more important relationships that we have derived. (830) and (831). .. aTe aUo linea of instanlaneoua :re. are everywhere perpendicular to lines of instantaneous:JC. .t many modes may exist simultaneously in any given waveguide. 1. (m = n = 0 excepted)..2/ ab _(rnT)_(nT) = ..ransmission lines as highpass filters.. 2. n = 0. In the light of filter theory. in terDl5 of the mode voltage and current.. 216.
Y.u. X e Normalization Propagation constant .RY OF EQUATION15 "OR TUE CTLINDBlCAL WAVEGtllDZ (TEM MODES NOT INCLUDED) TE modes Transverse Helmholtz equation Boundary relations Vt 1 '1' I 1 TM modes + k c i1t 0 qi'''  an 0 a." dz + .  vr .ed H.'do  JJ "dol j~ jkVl . Vi UII.' on C 0 00 C e' .' dV jfoli' Z.l 0 Transmissionline equations Mode voltage and current Transverse field dl . X'V."  'Vt'l''' u..'t'V' )W_ k...(V+el" . I <I.388 TABLE TIMEHARMONIC ELECTROMAGNETIC FIELDS 81. Y.If)' a ..+ ." .Z.'Vj'l" Mode vectors h"  .j . .. 1 Characteristic Z and Y Z.u.Y..VO liz V _ V+e'Y' + Ve'Y' 1 _ . X'Vt'lt· h' .Ve7') I Z.k.+r~ power P.eV 1!.' ." k.jk.. I JJ .. .h1 Longitudinal field z~ect. E.'I:' ehXu. Smoo. .U.)' 1 I> I.. . h .." .' E.
chap.veguide. we shall use the subscript i to denote the mode number. and p 1.~ + ~VN) ds  ¢ ~ :: dl and Green's second identity in two dimensions. 81 apply to each mode.. 44 for the special case of the rectangular wa.) 1cos n¢ J sin n~ I (835) where n . . (523) and (527) and normalize them.. . (sin n~l cos n¢ J . (x~. Modal Expansions in Waveguides. Normalized eigenfunctions for the coaxial and elliptic waveguides arc given by Marcuvitz.v." V1wt '1'._.. . 52.. we obtain _t. 2. Marcuvitz.." MIT Radiation Laboratory Series.p = "'Ir[(x~p)2 I t..~ . The equations in Sec. (x~pp/a) '!' • _ "p V..MICROWAVE NETWORKS 389 circular waveguide of Fig. 1 82. p are given by Table 52. J .2. 2.)' IJ '!'t'!'. 0. The result is ':lr' . vol. Normalized eigenfunctions for the parallelplate guide are given in Prob. Henceforth. is orthogonal to all other mode vectors.'ds Using the same substitution in Green's second identity.. ff (v. For this. we can pick the if's from Eqs.p/a) X . The X. 3.' and ¢ = iI!. . I. This concept was used in Sec. An arbitrary field inside a section of waveguide can be expanded as a sum 'over all possible modes.(•• .) co. . If (~VN II  ~VN) ds ~ ¢ (~:: . "Waveguide Handbook. n'] J .' g J.' d.:)' .' in Green's first identity. we have [(k. and the x~J> arc given by Table 53. 10... 1951.~ :~) dl First.')'1 IJ'!'t'!'. Let us first show that each mode vector e.(k. ~ 0 IN... New York.. . to identify a particular mode. . . We now wish to consider such expansions for cylindrical waveguides in general. 81.. +l(x. Illc. McGrawHill Book Company. we shall use the divergence theorem in two dimensions. .(k.. consider two TE modes and form the product Letting If = et· et = hi" hi" = V1W. Green's first identity in two dimensions..d.
. j (838) The orthogonality relationships (Eqs. we obtain II eiPds ff Hjoh. Noting that the mode vectors are normalized.. At any cross section along a cylindrical waveguide. we need two u crOS8I A discrete spectrum of eigenvalues i8 a88umOO. Lbtl.") .'lt/. = ViP (840) where p = e or m.pds= liP El . orthogonal sets of mode . X Vl'lt. we must consider the TETM cross products = . and we obtain If V. (836) to (838)} also arc valid for the c's replaced by the h's. (839) Because of the orthogonality of the mode vectors.V. and we have If c.' + h.. if kd ' becomes' ¢ k~t..··erd. (839) by an arbitrary mode vector and integrating over the guide cross section.·e/"ds=O et· er = h:· hI" i¢j (837) Finally. . (825) and (826). VI 'It/" If we let A = 'It/'u.I. the contour integral vanishes because of the boundary conditions. functions for degenerate case8 can also be found. as shown by Eqs. and the preceding equation If et· ej'ds = 0 (836) A dual analysis applies to the TM modes. we see that II e. Since there are two independent constants in V and I for each mode..' u.0 for all i... the integral must vanish. the field can be expressed as a summation over all possi ble modes: E.390 'l'IMEHARi\IONIC ELECTRO&IACNETlC FIELDS Rence.(u. However. X VI'!': in the divergence theorem. = HI = Lei'Yt + e.". we can determine the mode voltages and/or mode currents at any cross section by multiplying each side of Eqs. . X V/I'tds =0 Comparing the preceding two equations.
Furthermore. (1) . Suppose that only one mode propagates PIO. over waveguide cross sections.  II ij E X H'. (2) matched waveguide and H. 83.ds  . (4) H. 83.s. 82. It is also of interest to note that. over one cross section. a system for which a closed surface separating the network from the rest of space can be found such that n X E = 0 on the surface except over one or more waveguide cross sections.) The energy stored per unit length in a waveguide is also the 8um of the energies stored in each mode (see Prob." that is. the usual circuittheory reciprocity applied if the medium was isotropic.V. A microwave network. The Network Concept. the equivalent circuit of a section of waveguide in which N modes exist is N separate transmission lines of the form of Fig.1j II ei·ei ds  IIO:e. we saw that. instead of circuit voltages and current. and/or H. 2: V. when many modes exist simultaneously in a cylindrical waveguide. given N sets of "circuit" terminals.nce matrix was shown to be symmetrical. Let Fig. we can obtain additional cases involving discontinuities in E.1f i (841) which is a summation of the powers carried by each mode. In Sec. the modal voltages and currents of waveguide "ports" a.) L V. 83). each nwde propagates energy as if it exi813 alone. (We have used the indices i and j to order both TE and 1'1\1 modes in the above proof.. These may be (1) matched waveguide and E . over another eross section. when we have currents in a waveguide. We shall now show that the same network formulation applies if. the voltages at the terminals werc related to the currents by an impedance matrix. 49 are examples of case (1). .JdlCROWAVE NETWORKS 391 sectional" boundary conditions. This impeda. 38. over one cross section and H. Hence. we calculate the zdirected complex power P. and (5) E . To show this power orthogonality. over two cross sections. (3) E .re used. 410 are examples of this situation. over one cross section. The solutions of Sec.l U microwave network. The solutions of Sec. 83 represent a genero. over two cross sections. that is.
then that guide will be represented by N porn on the equivalent network.h. 1948. We visualize a surface surrounding an Nport microwave network such that E I = 0 on S except ovor the waveguide cross sections. assuming we are far enough along each waveguide for higherorder modes to die out. the threeport network of Fig. for isotropic media.[. . where V . can be similarly written for any Nport network. Furthermore. where (E. Hence. knowledge of E t or H" respectively. (334)]. Yij = Yj. H. Dicke.392 TIM&HARMONIC ELECTROMAGNETIC FIELDS in each wa. MIT Radiation Laboratory Serics. (H. microwave networks arc reciprocal in the same sense as are the usuallumpedelement networks. "Principles of Microwave Circuits. A knowledge of the mode V or I in the guide is equivalent to a.. and In are the mode voltage and current in the nth waveguide. the relationship must be linear if the medium is linear. according to the uniqueness concepts of Sec.] [2 (842) I. let us apply the Lorentz reciprocity theorem [Eq." Chap. Now that we have established these linear sets of equations.] V.veguide. Montgomery. To prove this."2 It is also of interest to show that. 83 but. [II. Zu Zu Zu Va = Z21 Zn Zu [I.l. vol. It states that <ffi E" X a·· ds "'" <ffi E6 X a"· ds for two fields E".1 Then. R.] [YU Yn yu][V. we can use all the usual techniques for solving linear equations. I For example. II{ N modes propagate in a single waveguide.J = %j. and an impedance matrix [z] is defined by V. H· in linear. New York. M.. .] [zu Zn zn] [V. and E. 4.•• V. %.a = Y21 Yn Yn Va Un Un Yu (843) Equations (842) and (843) have been written explicitly for. only the dominant mode exists in each guide. a" and E6. 33. (842) defines an admittance matrix (y] according to I.). McGrawHill Book C<Jmpany. (844) that is. . The inverse relationship to EQ.). 8. Inc. see C. The electrical engineer knows these techniques by the name of U network theory. of course. a knowledge of V (or I) in all guides is sufficient to determine I (or V) in all guides. since the mode vectors depend only on the geometry. Purcell (eds. D. isotropic media.
84.· .·I.Jt and Vf .' = 0 except 11. We then have P l" = effiE X H··ds = VI*1Pe X h·ds .I. and Eq. Similarly.MICROWAVE NETWORKS 393 (The n here refers to the nth waveguide. = dr [i!'. and The input impedance to the network is therefore Gf..' N N and the Lorentz reciprocity theorem reduces to Il_l L: V.i!'. Oneport Networks. the FIG.  'N.p •• .. as shown in Fig.._1 N N (845) To show that Eq. OW. Visualize a surface enclosing the network such that the field is zero on the surface exccpt where it crosses the input guide. z is the power dissipated. is the magnetic encrgy stored. + j2w(W.. the cj>EX H'·ds = ~ I .) (846) w.)J (847) which is well known for lumpedelement network theory.L: V."I. is the electric energy stored in the network. Similarly. Then Vt = zJ. A oneport network is characterized by a single impedance or admittance element. Because of the conservation of complex power [Eq. surface en· closing it._1 V." = 0 except It and (2) all f. (162)]. we have V 1° . it is merely necessary to consider the special CMes (1) all f..0 except V tin Eq." = 0 except VI". %tifl. work and a. (845) establisbes y" ~ Yii. 84. taking all V. A oneport net. not the nth mode. and all V." that is.'cj>e"Xh"'ds= . ~s .. 84.. (845) is equivalent to Eqs.  W.(W. (845) reduces to 'Z(j = zJi. . where ~~ + j2".' .'I. (844).Vf* where Vand f are the mode voltage and current !lot the II reference plane.) desired surface integrals become Hence. at the cross section cut by the surface enclosing the network.1 V.
I it must be the only solution. and hence V must be 900 out of phase with I. tion for ita proof. 84). a. any field baving n X E ..) and hence nand G Brc even functions of wand X and B are odd functions In the losslcss case. The frequency derivatives of Eqs.G . Suppose we choose our reference plane such that V is real. (850) are V X oE ow = JJl "H" . E is 900 out of phase with H.O.('».n in the lossfree ease. Bence. (847) to (849) we (1) A dissipationless network has be negative in the lossy case. can draw the following conclusions.394 TtM. as well as the field equations v X E = jwpoH VXH=jWfE (850) can be satisfied by assuming E rcal and H imaginary. R . 33 required some di!sipa.11) aw If we scalarly multiply the first of these by H* and the conjugate of the 1 It.&HARMOl\'1C ELECTROMAGNETIC FIELDS input admittance is Y  1 !Vi' . Then n X E is rcal over the reference cross section of the input guide and zero over the rest of the enclosing surface (sec Fig. p~ 1'2".G +1'B (849) From Eqs. . V 1* is imaginary. eve. This is therefore a possible solution. and the real and imaginary parts of Y to be conductance and susceptance. (3) and magnetic energies arc equal.)] (848) As usual. We shall now show that everywhere within the network E is in phase with V and B is in phase with I.JWJl oH aw VXaw=J( oH "E+" oE JWE (8.0 over the entire boundary would be uncoupled to the input ports. we define the real and imaginary parts of Z to be resistance and reactance.'».rvr (<I'. Let us now considcr the effect of a change in frequency. respectively. assuming uniqueness. and would have no influence on the external behavior of tbe network. may be recalled that the uniqueDel!Ill theorem of See. Z("') Y(". . Hence.nd. z~ R +1'X Y . (2) The Rand G cannot At resonance (X = B = 0) the electric (4) The Z and Y satisfy Y'( "') ~ z·("') of w. our eonclusxms apply only if we visualize some slight loa However. These boundary conditions.
1 B fheir frequency derivatives are therefore dX_ ck" . (aE ~ X H.2j(~.lEI') + ~. (850) by dE/ow.. and the divergence theorem applied to the lefthand term.. aE ~ ~ Similarly. (aE X H' _ ow X E') ~ j"IHI' .j. 7 aw javi dB j aI dw . this equation is integrated throughout a region of space. if we scalarly multiply the second of Eqs.•.) = jpIHI" _ jwp aH.IEI' (852) Finally. and subtract. ds (I' av + v.) dT . aw aH iJw _ jw. Equation (853) is now applied to the oneport network (Fig.. + jw.JI•• iJH aw We now subtract the above equation from the preceding one and obtain V. (853) becomes ow aw 11 (av T' + al v.V ow I le. H.) e X h. 84). Hence. (oR X E.M1CROWAVE NETWORKS 395 second of Eqs. E. The field vanishes over the enclosing surface except where it crosses the input port. we can write Eq. we obtain v .) aw = jelEI" + jWf: aE.here V and I are the mode voltage and current at the input reference ?lane. we obtain v. (854) rhe input reactance X and susceptance B are given by jX f = _. al) aw ow ". (851) by E· and the conjugate of the first of Eqs.J.a:a\ (855) VCOIlltu>\ . effi(~~ X H'  ~~ X E')'dS .j fff ~ _ ("IHI'+'IEI')dT (853) Note that the righthand side is proportional to the total electromagnetic ~nergy contained within tbe region. (850) by aH/aw. (853) as l ' : : + V· :~  j fff (MI' + . and the lefLhand side of Eq. . and BuLtract.
) (8.[If' (w..396 TIMEHARMONIC ELECTnOMAGNETIC FIELDS Hence. (856) and (857) is that all polu and zeros of the reactance or 8U8Ceptance function for a lonfree oneporl network are simple. IWlationships (8. B 2w 2w . and X'(wo) = ai.)' + . BdZ Svdem Tw. pp.\9) that is. This is known as Foster's reactance thtwrem. hence the poles of X are simple.ylor series about is then w. Furthermore. I An important consequence of Eqs. 1924. X has a simple zero at Wo and B "'" l/X bas a simple pole at '4 Similar reasoning shows that the zeros of B are simple.sceptance for a l08sfru oneport network is always pontive.\7) fiil' (''''. + w. A Reactance Theorem.) 2 (856) Equations (856) state that the slope oj lM reactam:e or su. X(w) = al(w . M. (856) and (857) for the energies. 3. we obtain (856) Because the energies are positive.) + a2(w  w. it follows that dX dw . Foster.) Solving Eqs.<». from Eq. it follows that >dw w dX X dB B >dw w (8.\6) to (859) were first establisbed in lumpedelemenl network theory. From Eqa. which must be positive hy Foster's reactance theorem.) fiil' (w. the slope of the reactance or susceptance is always greater than the slope of a straight line from the origin to the point of consideration. To prove this. suppose X vanishes at a resonant frequency WOo The Ta. Hence.. + w. J" vol. (847) and (848) we also have for lossfree networks X . the poles and zeros for the rea.ctance or susceptance function of a lossfree oneport network must alternate along I R.[If' dB dw  2 (w.. . 259267. . (854). April.w.w.
Other equivalent circuits of the Cauer type. 86. 86 reveals that it would behave somewhat like the dashed curve of Fig. always be greater than zero. the w axis. Figure g·5a illustrates the general behavior of a reactance or susceptance function. Inc. An exa. only an approximation to physical networks. Englewood Clift's. ~Dtico ." Chap. Also. 86. Equivalent circuits for reactance functions of the Foster typel arc illus~ tratcd by Fig.. else X'(w) will not. 713. An important difference between microwave networks (distributed elements) and lumpedelement networks is that the former have infinitely many resonances. the reactance (imaginary part of Z) of a slightly dissipative network would not become infinite for any real Co) but would be somewhat like that shown in Fig. and (e) a Foeler equivale. while the latter have a finite number of resonances.J. A study of the resistance corresponding to the reactance of Fig. g. Van Valkenburg. 12. The lossfree network is. It is known from the usual network theory that a slight amount of dissipation shifts the poles and zeros of the impedance function from the Co) axis to points above it. I or of mixed FosterCauer type. of course. Hall. (a) Typical reactance or susceptance function. It is therefore desirable to know how the behavior of networks with small losses differs from the behavior of lossfree networks. for which the power ulossJl is actually radiated power.MICROWAVE NETWORKS 397 x 0' B • (0) ~ o (b) (0) ne~ Flo.mple of a lossy oneport network is the linear antenna. can be found. see M.. of Fig. N. The effect of small losses cnn be shown in the equiva. 8. since Z(w) is an analytic function of WJ the resistance (real part of Z) is not independent of X. "Network Anal)'si5.lent circuits by adding I For 8luunple.5b and c. (b) a Foster equivalent work of the first type.5. 1956. Hence..nlo network of the eecond type.
convenient from a mathematical viewpoint. However. [1 to VI'.. VI = V1 II = '+ VI' II' + I{:E _1_ (V. Mathematica. it is usually convenient to make the natural choice that Zlll is the characteristic impedance of the waveguide connected to port 1. (862) are merely a linear traMformation from the two quantities VI. . Figure 87 suggests this travclingwave concept. V{) (862) and similar equations apply to port 2.lly. Network (2) z. • ". The port voltages and currents can be considered to be tbe superposition of incident and reflected components. microwave network.0. Another choice.. 7. . V {. Eqs.. is to normalize the characteristic impedance by choosing all Zo's equal to unity. a possible matrix for describing +It:. • R .veling waves for a twoport network... . While much of the theory can be presented in terms of the impedance matrix [%1.398 TumHARMONIC ELECTROMAGSETlC J'IELDS •. . From the travelingwave viewpoint. 8. and it is apparent that ZOl can be arbitrarily chosen. We shall make the former choice. 86. 85c. of small losses on the another and (2) filtering of signals impedance or tl. {or port 1.. . Hence. large resistances in parallel with the LC resonators of Fig. Twoport Networks. ()l _1. The effect. Tra. defined by (860) or in terms of tbe admittance matrix (8(H) it is often more convenient to use other matrices which emphasize the waveguide character of the ports. from one another. FlO. .5b aDd by adding small resistances in series with the LC resonators of Fig. ~7. 85.i Z.. The primary uses of twoport networks in microwave theory are (1) transmission of energy from one place to FlO.
&. along with other relationships among the various matrices.] V. of course.l [V. (866) and (867).8.. Defining the matrix ["~I . The various matrices defined for a twoport network are. as illustrated by Fig. the transmission matrix is related to the scattering matrix by [T) = su _ SuE" su] Sa Sa Sl1 1 [ . related to one another. (864) Another matrix commonly used to describe microwave networks is the scattering matrix [8] defined by S.[~:: ~::][ t::] (863) This matrix is particularly convenient when microwave networks are cascaded. The relationship of fSJ to IzJ is more compUcated.J we have [S] ~ [z  '.Sa S12 (867) The derivation of Eqs. that is.] _[S" Sn [Vi] V.J' (866) Similarly. It can also be easily extended to the case of multiport networks. 88. N twoport networks cascaded. twoport microwave networks is the transmission matrix (T]. Note that SII is the reflection coefficient seen at port 1 when port 2 is matched and Sn is the reflection coefficient seen at port 2 when port 1 is matched.MlCROWAVE NETWORKS 399 Flo. (861). 8 of the Radiation . the T matrix of the overall network is the product of the T matrices of the individual networks.[go> ~. defined by [~::] . The incident and reflected waves at the input of network 7t + 1 are the reflected and incident waves. Hence. can be found in vol. at the output of network n. [yJ is the inverse of (z].r 8 21 (865) This matrix is convenient for considerations of jmpedance matching. as stated by Eq.1I' + '. respectively. For example.
TuTu = zZ" (Sll9) in the transmission matrix. IDe. For any particular network. by Eq. the reciprocity relationships (Eqs.) ~ 0 (871) More generally. and E.Re (za) Re (Ztl) Re (VII) Re (Yn) . then til is the input impedance. that is. we obtain from Eq.quires From Eq. 19M. R. (848) that Re (Yu) ~ 0 Re (y. There are realizability conditions imposed on the matrices by the con~ servation of energy theorem. fp. PrenticeBall..) ~ 0 (870) Similarly. lne. McGrawBiD Book Company. "Network ADalYliJI. (847) we know " Re (zu) ~ 0 Re (z. From Eq. (847) and (848) must be valid for any passive termination. since Eqs.. natural" equivalent circuit. and restrictions on them can be obtained from the corresponding restrictions in the oneport case. "Principles or Microwave Circuita.R). vol. 8. For example. Hence. when the network is lossfrec. (867).ion Laboratory Series. we can show that Re (Zl1) Re (zu) . For example. M. Zu is the input impedance looking from port 2.0). l Our principal concern for thc remainder of this chapter win be to obtain equivalent circuits for microwave networks. it follows that reciprocity TuTu . (844») apply.400 TDlEIlA. the clements of the impedance and admittance matrices become imaginary. One of oW' tasks will be to choose a ... • M. I For networks constructed of linear isotropic matter. Montgomery. if port 2 is opencircuited (II .J. it is evident that reciprocity requires (lHl8) in the scattering matrix. 1945. Englewood CIiBI.[ONIC ELEcrRO»AGNETIC FIELDS Laboratory Series. whon port 1 is opcoi:ircuited. Similarly. Equations (866) and (868) also apply to multipart networks. Such considerations are particularly useful in the theory of filters. Van Valkenburg. (800). N. These conditions can be obtained from the corresponding oneport conditions by terminating the twoport network in various ways to form a oneport." Chap.." MIT Radiat. New York. D. PuN:cll (eda. an infinite number of equivalent circuits will exist.. .)." Chap. ]3. H. using the y matrix and short circuits on the ports.Re (Ya) Re (Y21) ~ 0 ~ 0 (872) Finally. Dieke. one which suggests the physical nature of the network. a section of I C.
It should be emphasized that it is only the 8ign of a reactance or susceptance that dictates whet.n jX P08iLivc reactance Inductor . n:l Change in impedance level Transmission line .. The reactance or susceptance does not. in general. jB Negative S1. TABU. have the simple frequency dependence of a lumpedelement inductor or capacitor. USED IN EQutVAL£ST C1RCutTS or Lo&&FREE NETWOU8 Symbol Represents .ypical equivalent circuit for a lossfree twoport network. . For lossfree networks.l8Ccptance 4~ Capacitor Negative reactance 4ri!!Ideal transformer Positive lJusceptal)ce ~C Z. we shall use the symbolism of Table 82 in equivalent circuits. since t. Figure 89 illustrates a t. ~ll Waveguidc section . waveguide would not be represented by an equivalent tee or pi circuit. 82. A typical equivalent cireuit tOt a 1000·tree t.WCROWAVE NETWORKS 401 FIo_ 89. .her an inductor or capaoitor is ohosen. SUlllOLl&W: Element.his would hide the transmissionline character of the guide.woport mierowave network.
. Obstacles in Waveguides. and (b) an equivalent circuit.(e. Fig. the characteristic impedances and propagation constants of the equivnlent trans· mission lines can be assumed complex to account for losses.il" + 61")e  = 2A cos (fJz) e l ') (873) Hll) . fJ is the phase constant. where e and h are the mode vectors. in a waveguide terminated at z = 0 by a magnetic conductor and matched as z + . the two Z6'S would probably be different from each other. as. in a waveguide. and the small losses introduce only secondorder corrections to the reactances calculated on a lossfree basis.s')b =. all of the dominant mode (see Table 81).uJi . In region (2) there will be only a wave in the z. Figure 811 shows the electric source ].. A 2A ei. the Z's will all be jX's. In the case of dissipative networks. Before considering the obstacle problem. Let superscripts (I) denote the region I < z < 0.402 TIMEHARMONIC ELECTROMAGNETIC FIELDS T T Z. hence Elm = Be. direction. and superscripts (2) denote the region z < l. In the more general case of an unsymmetrical object. Then in region 1 there will be an incident wave plus a reflected wave such that HI = 0 at z. 31 for rectangular guides.sin (ftz) b JZ. for example. Similarly. 810.I. Continuity of E l at z ". and Zo is the characteristic impedance. I requires that 2A cos fJl =. 32. Z. Z. . T ~ (0) Z. symmetric about the cross section T. (b) Zo FIG.. Figure 8lOa shows an obstacle.. Ej(l) = A (e. Most of the networks used in microwave practice are DIlly slightly lossy.'~e HI(I) _  8 el" h Z. (a) A 8ymmetrical obstacle in a cylindrical waveguide. resistors in series with X or in parallel with B can be used to represent the losses. = O. Figure 8lOb shows a possible equivalent circuit. and it might even be desirable to choose two reference planes T. 86. let us consider "dominantmode sources" in cylindrical waveguides.Be. An object in a cylindrical waveguide can be represented as a twoport network.00. In the lossfree case. Hencc.. The method of treating this problem is that used in Sec..
d. the H. ff E . 812b. If J. By symmetry arguments. Then the dominant mode will be a pure standing wave in the region . This divides the problem into two isolated parts. The boundary condition on H at l is uJ X tRw .==. = 0. is Reaction where E' is the field of .l < z < 0 of Fig. which leads to J. The equivalent circuit of Fig.. then by the usual tro. The excitation is provided by the dominantmode source J.H(ll] = J. + 2Z. is located where E .==l.) J. Fig. and the J. J. = Zo jtanpl (875) the source of arbitrary l. and E' is the field of the current on the ~ Matched guide IIi:. . which we have just analyzed. 812a is shown in Fig.I.) + (c.nsmissionline formulas Z Zo ~'or = Z. Now return to the original problem. 810a.11.~~ ej~le A quantity of interest to us is the selfreaction of the current sheet (8. J. We define even excitation of the waveguide as the case of equal incident waves from both z < 0 and z > 0. 2A' Z. the total reaction on J. 812a. S. (1 + &~') (874) We shall use dominantmode current sheets as mathematical II waveguide probes" to determine the equivalent circuit im"pedances. J·r Magnetic conductor ..) ~ If E . is zero at z = O. is equivalent to a shunt current source I.:. is maximum and H. scattered by the obstacle will also be zero in the z = 0 cross section i so a magnetic conductor can be placed over the z = 0 plane without changing the field. phased so that E . do = ff (E' + E') .812a.flCROWAVE NETWORKS 403 t: "'" which determines B in terms of A. d. J. one of which is shown in Fig.) We now further restrict the problem to the lossfree case. .. alone. A domiuantmode source in a waveguide terminated by a magnetio conductor.(".17 Fto. (The magnetic conductor is equivalent to an open circuit..
(s. using the identity (I tan 2 = 1 + COS a (876) sin a Eq.c)  . + 2X. and Z. (875) becomes X.'.(8... (1 + cos 2~1) 2A' Z. (873).z (a) t 11. E'· dJ' (877) where E· is the incident field. (.1 z 1+1.8) . 8100 is rcpre8entcd by (e).1'< (d) Flo.. we have 2A' Z. given by Eq. because only the lossfree case is being considered.1 J • T M• Mag..404 TIMEIIARMONIC ELECTROMAGNETIC FIELDS T T T T I Matched ~ guide .  We have replnced the Z._.= 0. (1 + "") Taking A as real. (C.._. Odd excitation of Fig... . .... which bas an CqUivlllCll~ network (6). by jX. Z. .1 (b) T V T Elect. Re (C. Even excitation of Fig.. By reciprocity. condo T Matched ~ guide .. H l is adjusted to a cross section for which E. which bu an equivaIcnt network (d).) . (e) Za ..in 2~1) and. and J'" is the current OD .' + 1 I. 812...8) where the last equality is Eq.J. obstacle alone..8) 1m (c. 8IOa is represented by (a). (874). both radiating in the waveguide terminated by the magnetic conductor at z = O. condo Za 1.)  2A' Z. then the reaction vanishes and the above equation becomes (C... andjX...
By Eqs.Re (c.c) (see Sec.a) .a) Hence.a) is real because E' is real. (c.. (a.(a. since n X E' = .a) . or a magnetic dielectric (.a) This formula applies only when J.' Note that the problem is now identical to the echo problems of Sees.a) . (879) becomes ~ ~~. ffE" r." on the obstacle and calculate (a. The change of sign in going from Eq. 77). The last equality is met by reciprocity.c) and Eq. For a stationary formula. (879).(s.. n X E' = 0 X E" (878) and (s. then (s. In the hUer case the term .a) subject to the constraints (a.c) . ~ "0). For the case of a perfectly conducting object. but is usually at some constant phase.MlCnOWAVE NETWORKS 405 the obstacle.') (879) Our problem is now one of finding the selfreaction of the currents induced by the incident field of Eq.(c. 1m (c. we therefore have Xb + 2X" Z.'.c) . (879) to Eq.a) (880) This.0 X E~ on the obstacle surface. represents the variational solution to the problem. (~77). (873) with A real. except that all currents radiate in the environment of the waveguide plus the magnetic conductor.~ ~ I I' (a. which is usually the case. . coupled with Eq. Zo 1m (a. and.\ and n X E = 0 on its boundary. 710 and 711.a)' (881) X. the obstacle current is eo surface current J.(c. Equation (880) can then be written as (c. Hence.a) Re (a.e) represents the selfreaction of the currents induced on the obstacle. our stationary formula for (c. we assume currents J.c) . + 2X. (876) to (878).fH"' dM< must be added to the righthand 8ide of Eq. I The obstacle may be a conductor.c) is (s a)' (cc)"" ." is real. (881) can be explained by noting that J.' ds where (c. a nonmagnetic dielectric. If the trial current is taken as real.~ is not real for the given E'.
LlIL.. 8lOa) as the case of equal incident waves from both z < 0 and z > 0.YoX.406 TIMEHARMONIC ELECTROMAGNETIC FIELDS We define odd excitation of the waveguide (Fig. 812d.. 8IL The equivalent circuit of Fig.. (The electric conductor at z = 0 is equivalent to a short circuit. II. 87.. and their selfreaction {a.. Dual to Eq.  1m (c.. rectangular waveguide. Hence. 812c is shown in Fig. together with the electric conductor covering the z = 0 plane. (879) we have j 1 YoZ. Posts in Waveguides. in the region . Finally.) The analysis of Fig. e (882) where Yo = I/Z o is the characteristic admittance of the dominant mode.2C cos (~z) h El' =  ~C sin «(3z) jY. By symmetry. is dual to Fig.c) Re (c. curt:ents J. (881) we have ~ 1 YoX. Some variational solutions for circular posts in rectangular waveguides can be carried out relatively simply. If the J. and the M.a) is calculated with an electric conductor over the z = 0 plane. phased so that E l = 0 and H j is maximum at z = O.c} is the selfreaction of the obstacle currents radiating in the presence of an electric conductor over thc z = 0 cross section (sec Fig. 812c). 1m (0..0) Rc (0.I X (a) FIG. is equivalent to a series voltage source V. The excitation is provided by a dominantmode magnetic source M.. and so an electric conductor caD be placed over the z = 0 plane without changing the field. (873). which. one of which is shown in Fig. then dual to Eq. the E l scattered by the obstacle must also be zero in the z = 0 cross section." are assumed on the obstacle. Posts in and (c) otherwise.l < z < 0 we have a source field Hi . .. dual to Eqs. (a) cylindrical to y. 812c.  (~ X 00 X . Figure 813 illustrates three classes of obstacles: (1) those cylindrical to y.0) (884) where (a.. (b) cylindrical to %." is real. 812a. This divides the problem into two isolated parts. for a variational solution. 812c is dual to that used for Fig.a} is calculated.c) (883) where {c. 813.
E. For even excitation (Fig.(lCROWAVE NETWORKS 407 J.. o o J.14. ..). The field of case (2) will be TE to x. we can write where the first term is the freespace field of J.J.l_) drp (1lll6) . o o J. 8124).SC is the dominant mode with E parallel to y and HI parallel to x.. Type (3) problems require two scalar wave functions to express the field (see Sec. Then the field of case (1) will be TM to y. ljI. r 1fd The field produced by J. Image system for the circular post in a rectangular waveguide.2}0 dy}o 2 d~ (J. Hence.in the waveguide with magnetic conductor is onehalf that for the complete post in a waveguide. 815. (2) those cylindrical to x.) = 41r}o 1 (. as shown in the insert of Fig. We shall consider only the centered circular post. expressible in terms of a single wave function Ail ". '" (see Sec. expressible in terms of a single wave function F" ". 312). [The c)·tinders are not necessarily eircular.. hence (. 44).. f (" (Ev""" + E.and the second term is the frccspace field from all its images. OX FlO.) . o J...in the waveguide closed by the magnetic conductor will be the same as the freespace field from the image system of Fig. and case (1) is different from case (2) only because of the excitation. The selfreaction of J. ("d .. assume a constant current on the post e _ U _I (885) ] .] It is assumed that the incident wave in each C3. 814. J. and (3) all other eases. g.
. (k..··I. we shall now show that it can be transformed to .a) for the assumed current of Eq. (887) becomes (a. the Hankel function summation in Eq. (886) reduces to (a. single cylinder of constant current was calculated in Sec. Hence.a) ..at [ Er L~ + J.781 and S is the rapidly convergent summation 1 I] (2b/X)' n (890) . (885).) E.2rJ.'·.. with p replaced by the distance to the image. Eq.J.a) .) . \ ' (I)'H.'''(nkb) ~ ~~[ ~ ~(2b/X)' 1 1 4 b I +iG10g2r +8)] (889) where"Y "'" 1. (592). l "00 A . K=  §ka[!Jo(k~) L: .) 2 .·d¢ .K [ H 0'" (k . 5. • (I)"H. E."'(nkb)] (888) is an unimportant constant.. 56.) E. . convenient for computation... Equation (888) is an exact evaluation of (a.. (see Sec. (888) converges slowly and is not..408 ¢.8).... TUlfr. (k..o (887) and Eq. r Thus." is independent of The "image " term is a sourceiree field in the vicinity of the post and = can therefore be expressed as ElIlm.cting Crom Eq.. (kp)e f . (k. Lo] ~kIJo(k~)Ho(2)(kp) p The field of a. Abstra.'..) where + J. However.2< A oJ'(k.. Unfortunately.ElARMONIC ELECTROMAGNETIC FIELDS Now the upost" term is independent of 4> since the I. we have E II I1OH"".'·.  ~~ The field from each image is also of the above form. ..
For large 11.+ I)] .. 1. Returning now to the selfreaction..2: _ Re • • C05no .Re ( 2 10g tan (1/2) .. we have the righthand side of Eq. by the method of Sec. (888) and obtain Re(aa}C 2 . \' sin •• 2 sin where only the first term is real bccause it is assumed that 1 < (2b/X) < 2. and p set equal to zero.(kd/2) 1m (a. 814 with J.. we can find the total field in the z = 0 cross section due to the above f....2 + 2S 2.3..b ] (892) .5. (889) is the E" from all images of the filament 1 2 .k across the center of the original waveguide.. V (2b/X)' 1 C' b X .. (889). It is • E ow ~ ~ [ (n/b) (n. (584). \' (e")" _ Re 11.3.)sinH." replaced by the above f.5. . 410. • .2 log tan 2 1 Hence. • V(2b/X)' IlL.a) = C [ . letting x = (b/2) + p in Eq. _1.MICROWAVE NETWORKS 409 The freespacc field of a fllament of current is given by Eq. ~This problem is Fig. ) 2:~sin('. Hence.. Vn' (2b/X)' sin +.2J . (891) and 8 = 7rp/b in the above identity. the lefthand side of Eq.. n 4 . we can add and subtract the latter from the former and obtain E " 00' _ _0 r ~[ 1 V(2b/X)' 1 +j (! log 2b _ 1 + S)] 2 rp The freespace Ell from the same filament I is When this is subtracted from the total Ell.(kd/2) + log T .N... (! log 1I + eei./2) (nn/b)] (891) . the above summation has terms equal to those of .) Then. (889) into Eq.') _ Re (! log I j sin I ) 2 2 coso 1 j) _ I I . wc substitute Eq.
0 0.4 1. C .l Top view IB.. because of the crudeness of OUf initial trial current..2 O.~  2 + 28) (893) Now...d ). However. (892) and (893) into Eq. 11 0..O Equivalent circuit 0...4 0. End view L. (890).J4>. The centered circular inductive poet. . 2 "'" '" " R . + 2X..'.0 ~ 1. ~ L0. (884). 812c).. Hence. T X'. lot . 'X• Zo J 'fl R::: ~I.... For odd excitation (Fig./Zoh o I 0.6 "lb=2.15 ~ ::::: . for small d/>./Zob 1.. (881)."'" u"sin ~ (894) where S is given by Eq. ~ ~ [lOg 46 _ 2 + 28 (~)] Zo ). in a rectAngular waveguide.8 7> ~ 0.) ..410 TIMEHARMONIC ELECI'ROMAGNETIC FiELDS where C is the unimportant constant.j jXII l'.05 I ~ ~ I'... 12 The appropriate variational formula is Eq. J.ka l'J"(k~) __ 4r 2 Equation (892) is still exact for the current assumed in Eq. we llSSume a current (895) induced on the post.10  g.a) ~ C (lOg .ve X. substituting from Eqs. we caD expect our result to be valid onll:.20 (Afkr dth FlO. we use smallargument formulas for the Bessel functions and obtain 1m (a. Zo jX. 815.ld . we ha. (885).. Mamaih.
McGrawBill Book Company.15. (894). If the obstacle is symmetrical about a transverse plane. However. will now be discussed.a) (897) where (a. The result is (896) f6 I FlO. In a rectangular waveguide. For real current. 812. because the real part of the reaction is independent of the size and shape of the obstacle.a)/reo . 10.!lucnoWAVE NETWORKS 411 the exact evaluation of which follows steps similar to those used to derive Eq. S16. pp. independent of the size of the obstacle. the Z's arejX's. Consider even excitation of the guide (Fig. Hence. Small Obstacles in Waveguides.a) = Re (Il.. as the obstacle becomes small. A small obstacle Figure 815 shows X . Let us first make some qualitative observations. the imaginary part of the freespace selfreaction becomes extremely large as the obstacle becomes smaU. "Waveguide Handbook. The effect of a small obstacle is small. 88. hence Z& is small and Z. Marcuvitz. Formulas and calculations for offcentered posts are also available. is large. An approximate evaluation of tho reactions. Equation (881) is then X. New York.a) "'" 1m (a. The formulation of the problem for a conducting obstacle is that of Sec. 1951." MIT Radiation Laboratory Series.. If the obstacle is lossfree. we can let 1m (a. made possible because the obstacles are small and not too near the guide walls. Figure 816 represents a small obstacle in a waveguide of arbitrary cross section. 812a). the reaction (a.a) is the selfreaction of the assumed currents in the waveguide._ (898) In contrast to this. the real part oC the Creespace reaction approaches a constant.. we can calculate the dipole moment Il of the freespace obstacle and let Re (a. vol. 86. . and X& as calculated in a waveguide.a) is the freespace selfreaction of the obstacle plus the mutual reaction with all its images. The mutual reaction between the obstacle and its images therefore cannot be neglected. for sufficiently small obstacles.pacitive post (Fig.a) Zo = 2 Re (a. 8lOb. 257263. 813b) is given in Prob.. I A solution for the circular ca. the error being of the order of 10 per cent for d/b = 0. the equivalent circuit is as shown in Fig. Inc.. from a secondorder variational solution.Il) (899) IN. (894) and (896)] is accurate for small d/b. 1 1m (a. 1 Our solution [Eqs..
A small conducting sphere centered in!L rectangular waveguide. Eq. we can set Re (!l. assume J. 410 for a current sheet J. 817.~.. As a." is that which produces the dipole field external to the sphere.. even though we shall not need it explicitly. (2115) evaluated at r = c. consider the small sphere of radius c in the ccnter of a rectangular waveguide.Z..Re (P) of Eq.(yn Because the current is real. .. Equation (898) is therefore 1m (a.a) "'" 12rJ c .If)  Jo 1 =Il ab Hence. . (II)' (8102) ." = U'2 28m 8 . 23~ " (~l)' (L)' 1\ /Ii(. (Il. ab (If)' =  ab). (486).n example. . As our trial current.Il+~).bj FIG. This current. the ima. (8100) where (J is measured from the x direction. Hence.~(Il)' ' (8101) For the real part of (a. 1m (a. (487) and obtain Re where.412 TIMEHARMONIC ELECTROMAGNETIC FIELDS 'Side view xt J. from Eq. as shown in Fig.a). to z II ~c L~~ End view T~ 4. we can use the analysis of Sec. Because the above current produces the same field as an xdirected element of moment Il. (899) becomes 2 Re (a a) ~ .a)f'" . = ..ginary part of the freespace selfreaction is the imaginary part of Eq. 817.. is approximately J. The righthand term represents the selfreaction of a current element Il in the waveguide.ll) .
24~'l.a) ~ Re (KI. (8101) aod (8102) into Eq. Our freespace reaction is the Rayleigh approxi· mation (Eq.acle approximation Cor a centered sphere in 8. and Eq. . Atab (8103) This is the 8ma1l~bst.c' x. we evaluate the righthand side of Eq. (8101). (8102). there can be no net electric dipole moment. (897).1 and c« a/2.1. (899). (6106)]. we should expact Eq. Equivalent Circuits for Small Symmetrical Longitudinal ApertW1!l aDd Obstacles. which can be calculated from the MSumed current. (8104) by methods dual to those used to establish Eq. (884).. given by Eq. Hence.KI) . < 0. Then. assume that which produces the magnetic dipole field external to the sphere.a) ~ Re (KI. 1960. for odd exeitation.. ". (8103).a) 1':: 1m {a.Kf) (8104) where the righthand term represents the self reaction of a magnetic current element Kl in the waveguide. no. we have Zo ahA. which is valid for c/). For the trial current.KS 413 Sub. ~ .12. 1.. can then be made according to Eq.tituting from Eq.. MTT8. It is evident from symmetry that. we obtain 1m (a. Taking the current as real.a)I . A. Hence. 812c). (Kl)' (S.a) according to the freespace approximation [Eq.'. I I A. IRE TraM.MICROWAVE NETWOR. we evaluate the imaginary part of (a. January. For the centered ydirected magnetic current element iu tbe rectangular guide. There will be a magnetic moment KI (unless the obstacle has zero axial thickness). 817). (884). The freespaee sclf~rcaction of this current is then just the dual of that for the electric dipole. Return now to the specific problem of a conducting sphere in a rectangular guide (Fig. X. the resultant magnetic dipole will be ydirectcd.ab (Kf)'  ab. (898)]. Re (a. 01'_ = ~(Kl)' Y. Oliner. rectangular waveguide. we use the approximation Re (a.. vol. (899) does Dot apply.106) ~ Substituting from this and from Eq. (8105) into Eq.2e' The accuracy of this formula is at least as good as that of Eq. (8103) to be accurate wheo c/~ < 0. However.. because of the symmetry of the obstacle and of the excitation.a).. we bave Z. 12 2 I (8105) "re For the real part of (a. The evaluation of other smallobstacle equivalent circuits can be found in the literature. Now consider odd excitation of the guide (Fig. The evaluation of X. . analogous to Eq.
The equivalent circuit is shown in Fig.a) Re (a. condo ~+. and the opening in it is called a.414 TlMEHAlUIONIC ELECTROMAGNETIC FIELDS ['. Diaphragms in Waveguides..on the dia guide IJ· I. To evaluate the shunt susceptance. The exact equivalent circuit is just a shunt element. B 1m (a. negative (inductive). as shown in Fig. which h. 819.0 plane. which reduces to ~ 2Y. and (6) an equivalent circuit. Symmetrical excitation of Fig. (a) Mag.1 FlO.1 ~guide Matched M·t I..a) (&107) where (a. the susceptance may be p<l6itive (capacitive). condo Elect. . ~lOa ill also represented by (e). which has an equivalent circuit (d). Figure 8184 represents a cylindrical waveguide of arbitrary cross section with an infinitely thin electric conductor covering part of the z . 87.!. or change from positive to negative 88 the frequency is varied (resonant when B "'" 0). Depending upon the shape of the diaphragm or window.. (0) Mag. 81&. + B/2¢ I (d) I . 819a. 819b. Symmetrical excitation of Fig. condo I . we can use the method of Sec. Taking the case of even excitation (Fig. This conductor is csJled a diaphragm. 8IOa is represented by (a). . window.. the diaphragm problem reduces to Fig. 89.a) is the selfreaction of the assumed current Diaphragm ~Matched J. all equivalent circuit (6).1 z I. 818. 812a). The appropriate stationary formula is EQ. (881).1 z + $ PI I. (a) ~ (b) T T FIG. The diaphragm pill! the window cover the entire: = 0 cross section.. I (b) B/2¢ . (0) A diaphragm in a waveguide.
.. so that complete duality is preserved. 0 ~ r' (b) Is y.o is related to the tangential E in the window of the original problem according to M. d.o on the window. (8108) can also he viewed as a specialization of Eq. """ Re {a. 0 a FlO. To illustrate the theory. 820).' ds .. ~ (8109) Because the M. because E is real.ff E X H . We can think of Fig. 8l3b. M.(.ruct. consider a capacitive diaphragm in a rectangular waveguide (Fig. (884).  ff H'· M. (Note that it must be capacitive. 819a. (8107).ion of the window by placing pieces of magnetic conductor on top of an electric conductor.ic current sheet.a). we have the alternative representat.' d. (8108) where the subscripts m are added to emphasize that (a. 820. . instead of the electric current sheet of Fig. Then. the problem is the same as tbrnsc treated in . (a) Capacitive diaphragm. (8107) which is an obsladecurrentformulalion. ff H' . we have B 1m (a. dual to Eq.. because it is a special case of Fig.a).MICROWAVE NETWORKS 415 phragm. 819a as being constructed by placing pieces of electric conductor on top of a magnetic conductor. 2Y.) Take the Efield formula {Eq. 819c. This can be viewed as a const.a). Because the diaphragm problem is selfdual. Note that Eq. noto that (a.ff E X H'· d8)' .a)". Hence. is the selfreaction of assumed magnetic currents M.p' U.T T b Side view yt a 1 (0) I~x End view ·1 T a T y. (8108) is known as an aperturefield formulation of the problem. This is in contrast to Eq. (a. . that is. and (b) an equivalent circuit.' u.a). (8108)] and. X E (8110) Eq. u. The source has been changed to a magnet.ion of Fig.
. .416 Sec.h. ~x Y <c (8111) Hence. (473)..P' ~ "2 1.I' • . .. A more general treatment of the problem proceeds as follows. . 1 2Y. y>c in t. the Fourier coefficients E I . = 8b (~. above into Eq.. we have (a. 813b that the ficld must be TE to x.0 where.a). are E..!!.B... if we assume Ell' I. 49.: (Y. ~ ..IE.. we calculate • ab \ ' 1 (a.I"' lIY. 0 r {f(Y)'in : 0 Y <c (8113) y> c Then. We know from the discussion of F'ig. = f [e n1rY t )0 fey) cos Ii"" dy j2bY.IVI'Y. substituting from the (8112) B. is the aperture admittance.he problem of Fig.) where Y. ~ 4b lVI' (Y. which is the only mode having real characteristic impedance. 417. = X. (2bjX.IE.  P' .he window. from Eq.. by Eq. and so the most general form for the tangential field in the window is E:l _ . TIMEHARMONIC ELECTROMAGNETIC FIELDS In particular.p and the characteristic admittances of the TEXt" modes are (Y) o 1. by the methods of Sec. because of our assumption that only the dominant mode propagates. ~ ="'2"Y'. we then have precisely t. (8108) becomes B . (477)."IE". 49. Hence. V nt (8114) The Yo and X. . (8108).0 { = sma 0 .b + jB.'.) X. .a).h. pertain to the dominant mode.I' . 417. 2aY o where the quantity in parentheses is plotted in Fig. Eq. we obtain Yo a Yo Finally.
/(y) cos nTl/]' b dy (8115) B 8b • \' Yo ~ >. or by applying the Ritz procedure. • I _. given by Eqa. The x component then adjusts itself to make the field TE to x. If we assume a current J. (432) witb • !J' = sin ~ \ ' A cos nry e''''~ aLi" b .0 _abjrt cos~\' nA"sin nry b wp a '' J" . Better approximations to B/Yo can be obtained by using a better choice for fey).. The field is TE to x. becomes 1 « f1. 8194) is then J. [ J.. upon substitution from the preceding equations. =7[l'.).0 where The current on a diaphragm backed by a magnctic conductor (Fig.><J(Y)dY~]"' Equation (8112) represents the special case fey) = 1. JWP •• 0 k' sin "" a 2: • A" cos n' _"_V b .'("'2b"'/"'~. the current has both x and y componenta. =V7n:C'C=. The stationary formula.0 Hence.0 = ( T/a)' .""'HlIl.WCROWAVE NETWORKS 417 which. (8107)] is specialized to the capacitive diaphragm as follows. but the A" can be determined from the y component alone. in terms of obstacle current [Eq.T.H. "" and define Fourier coefficienta g(y) sin ~ a (8116) (8117) then .
. 821.2 0. . where the characteristic impedances (Zo)" are the reciprocals of Eqs. 1951.' o 0.. ab \ ' 1 .1 0. ~ CO) ' Cd) Also.1 .~. .' 2Y. New York. at z = O. '. and (d) crude qU8llistatic solution. (8115) with f(y) ~ 1. we find the sclfreaction of J.1 Curve (b) is the crude aperturefield variational solution.).phragm with c . which is also Eq.. (8117) and for (Zo) . Flo. (e) is • crud. [nc. (8115). 10... 4 .3 0. Eq." as • (a. ""A" cos nry ~ b Hcnce.sm . 35 and 51. Eq. (8112). (474) was derived.a) ~ "2 L. vol.J. McGrawHill Book Company. (8107) reduces to • (Z. This solution is obtained by finding" quasistatic field and then using it in the variational formula.4 b/'A. = l/(Ya)l./ ~~ . the tangential electric intensity is given by E" = 0... only the n = 0 term of the summation is real..J. we finally have (8118) This is the stationary formula in terms of obstacle current for the capacitive diaphragm of Fig. and • Ell = . (e) crude obstAclecurrcnt variational solution.. I N.. (a) Exact solution.. Because only the dominant mode propagates. The capacitive dia. Curve (0) is called the exact solution because the estimated error is less than the accuracy of the graph. Figure 821 compares various solutions to the capacitive diaphragm problem for the case of a diaphragm covering half the guide cross section. .r:. Cu'v. in the same manner as Eq. from Eq.~1=~~_ B"" 2ZJo" L: Substituting for In from Eq. 820." MlT Radiation Laboratory Scries.418 5 TIMEBAnMONIC ELECTROMAGNETIC FIELDS (b~/ .2b. Marcuvit. (b) crude aperturefield variational solution. rx \ ' a . and Eq.. secs.). (8114). "Waveguide Handbook. (Z. (8114).
Since Eq. For a crude aperturefield solution. The treatment of the inductive diaphragm (Fig. McGraw· Hill Book Company. This is evident. 823) is similar to that of the capacitive diaphragm. our approximation [(V) = 1 was an exceedingly crude choice.s that. (475) for E. "Static and Dynamic Electricity.MICROWAVE NETWORKS 419 obstaclecurrent variational solution.t in tbe aperture. is above the true solution. 1510. If we were to use Eq. because the boundary condition that the current vanishes at 11 = c is violated. .""" logcscYo X 2b Q (8120) In practice. curve (c). (8121) in Eq. lne.25. I W. Eq." 2d ed. so this last solution is a good approximation for most purposes. the former yields upper bounds and the latter yields lower bounds to the true B/Yo. (Ty(2b) y ""in' (Tc(2b) 'in' ("Y(2b) (8121) which actually has a singularity at y = c. Note that the aperturefield variational solution. is below the true solution. That this is so for any trial functions /(y) and g(y) follows from the fact that Eqs. 822 all have the same equivalent circuits. Smythe. yet it led to usable results. Our crude variational solutions give an error of the order of 20 per cent. the solution diverges. (8115) and (8118) are positive definite and hence are an absolute minimum for the true fields. (8118) gives Yo/B. A quasisLatic solution to the problem is f( ) _ co. 'K(y . firstorder quasistatic solution to the problem 1 B 8b 'KC .) Curve (d) is a.. 814 and 815. (8118). New York. but it is remarkable that they are as close n. we ~an assume Eq. Hence. curve (b). R. 1950. The general variational formulas for upper and lower bounds are given in Probs.. (8119)] is equally crude.. (8115). Our approximation to g(y) [Eq. and the obstaclecurrent variational solution. the result would be very close to the true solution.c) g(y) . because the image systems for all three cases are identical. It is interesting to note that the three diaphragms shown in Fig. with . waveguides are usually operated with b/XQ < 0.m 2(b _ c) (8119) (If the case g = 1 is tried. Sec. The existence of variational formulas for both upper and lower bounds is not very common and is a consequence of the selfduality of the problem plus the positivedefinite nature of the resulting variational formulaa. (8115) gives B/Yo and Eq.
a. New York._ _l. The problem can also be treated by quasistatic methods) a firstorder solution being! B Yo t::>< _ Xli' (1 + esc' 2a cot' 2a ~) 'Ire a (8123) A combination of the quasistatic and variational methods can be used to obtain solutions of high accurney. We shall now consider waveguide junctions formed by butting two cylindrical guides together.(7fe/a) (. These three diaphragms give rise to the same shunt. &2. 822. New York. Inc. IN. !leC.. The values of BIY.ive diaphragm.he. &6 do not apply directly. 60 the methods of Sec.nd Cb) an equivalent circuit. "Statie and Dynamic Electricity. is the aperture susceptance plotted in Fig. _ _ _T'_ _ I Stde view (a) xt. Inc. capacitance. [~ 1. p. R. We there1 W. (0) Induct.:: = 0 cross section. (6) FIo.' 810. Figure 824 represents the general problem. ." 2d cd. calculated from Eq.!. (6) (c) .6 f Jl ~ ~ Yo C 1T T T jB End view Y 1I y..420 TDlEHARAlONJC ELECTROMAGNETIC FIELDS IriI (aJ L... 1950. 419. 555. Mareuvitz." M1T Radiation Laboratory Series. vol. McGrawBill Book Company. Smyt. No longer is there symmetry about the:. (8122) will be higher tban the true values (of the order of 20 per cent higher). 10.:: = 0 cross section. Waveguide Junctions. McGrawHiU Book Company. 1951.. "Waveguide Handbook. possibly with a diaphragm covering part of the:. &23. This procedure gives _X. Flo.6 B) (c/a)']' a 5m X (8122) where B. Y.
MIcaOWAVE NE'TWORXB 421 (ore take the more fundamental a. respectively. If the characteristic admittance of the rightband transmission line were chosen as n t times the characteristic wave admittance of the guide. (0) A waveguide junction.refer to regions z > 0 and z < 0. are the cutoff modeattenuation constants.gQ ..(. . &24. are the characteristic admittances. and r is the reflection coefficient for the dominant mode. We shall usc Eq. because an electric conductor placed across the entire z = 0 cross section presents a short circuit to both waveguides.£1 E.If E+ X H+· ds .. The characteristic admittances of the equivalent transmission lines are taken to be the characteristic wave admittances of the guides.e""'~h.pproach of const. we caD think of Eq.(ei~~  rei·''') 1 ~.. (8124) to obtain stationary formulas for Band n t • It is assumed that the excitation is at z = .. a.: . L: YIV. are the mode vectors. and the ideal transformer represents the change in admittance level. then the transformer would not be needed.. (8125) where ei. The subscripts 0 denote dominantmode parameters. An equivalent network for the junction is shown in Fig.".X H..... In terms of the reaction concept. (8124) as stating that the reaction is conserved at the junction. + y. and (b) an equivalcnt. + ref'") 1 :. ds (8124) where superscripts + and . rho  LV.. Y. circuit. It is evident that only a shunt element is required to represent the junction. 824b. r . h.ructing complete solutions in each region and enforcing . hence in the region . Matched conditions are assumed at z = gQ i hence in T I __1Side view (a) End view Y~C: (b) T l:n T FIG. <0 E.~" H ..
we have real and the Yi. dB . (8131).+ "'" Y o +9"oei8 'h. where the carets distinguish the various parameters from their z < 0 counterparts. with Vi and 1".'·b. B (8128) OJ are imaginary.V.~ .. because Ee . The applicat.. 1'.... arc formulas stationary with respect to small variations . i Y. given by Eq.. to obtain the Vi and 9'" we need only specialize Eqs.~ Y. obtain Vi = JJE. for Remembering that the Yo real Vi and 1'"...+ = ~oei~zeo + H.t (8129) Y.0 on the conductor..e'. · c.9". E. ~ ~ ~ • YoV. with matched conditions at evident that <Xl.t  L: i Y.' + l: 1'. it is hence (8130) Finally.•.  !! ._~~~_ l: Y.1 .ELDS the region z > 0 E. 82..V. EquatioDs (8129) and (8130).ion of Eq.~r"I'ei + 2: . (8124) to the above field expresSiODS yields Yo +9'o' + I 1".9". l: ~.422 TIMEHARMONIC ELECTROYAGNETIC }O'l..1'" Yo ¥o' Yo +9'. Y G + J Y.. d• (8131) Note that the integration extends only over the aperture. = YoYo' Z ::I From OUT equivalent circuit. using the methods of Sec. (8125) and (8126) to z . 0 and.G jB _ '.% (8127) Now the relative admittance seen from the lefthand guide is lr 1 r + aTC Y. (8126) f. . .
Our approximate answer (Eq. and for the . For the capacitive junction (Fig. But. because the field in the aperture is less singular at the edge of a step than at a knife edge.1.ff< . Alternative stationary formulas in terms of current on the conducting wall at % =.!! = Yo 2c B. "" E. we have V.HIcnOWAVE NETWORKS 423 in the aperture E j about the correct field. However. (8129)] is positive definite in our particular casej so Eq. the field in the aperture is of the form E. (8134) where the quantity in parentheses is plotted in Fig. In Sec. (8129) then vanishes. .~.  Uif slnac a . by Eq. 49. (8129) specialized to the case of two identical guides is the diaphragm solution of the preceding section. and the second summation is related to the aperture susceptance of Eq. This can be compared to the 20 per cent accuracy in the corresponding diaphragm prolr lern. In general. . (8134) gives values of B/Y 0 higher than the true values. To illustrate the theory. (478) by ll'. 818). 417. for I(y) c:  I.t "" ac/2j hence. 418 is treated in a similar manner. regardless of our assumed tangential E in the aperture (8132) we bave by Eqs. "'" /(x). (8130) and (8131) (8133) This is therefore the exact transformation ratio of the ideal transformer. consider the rectangular waveguide junctions of Sec. = t acYo 4c ("'o2a B") ZO X. 821. (8134)] gives an accuracy of the order of 10 per cent.. . The first summation in the numerator of Eq. the dominantmode vectors are e. 416). (8129). we calculated the aperture susceptance corresponding to the crude choice 1(Y) . we should expect the 868umption /(V) = 1 to give better results in the junction problem than in the corresponding diaphragm problem.ie'B. The general expression IEq. as illustrated by Table 83.' ilVl'B. The inductive junction of Fig. 49.0 can also be obtained (see Prob. illustrated by Fig. Note that Eq.d _ ulff(y) sin ~ a Hence.
424 TABLE TIMEHARMONIC ELECI'ROMAGNETIC FIELDS 83.(cia)' (8136) a. because of the positive definiteness of the variational formula. by Eq.4 t 1. to.in Eq. 524.' 0. 1951. 824a.. z < 0 and z > 0.~~~ LY.84 2.. (8136») depends on the assumed aperture field and is therefore approximate. Finally. by Eq. McGrawBill Book Company.b (8137) Yo C A" where the quantity in parentheses is plotted in Fig.63 1. _2X. (8134) 'to THE ExACT c/b .(_ s) jn Yo = .5 SoLUTION I FOR THE CAeE Exact Approximate o 0. 825. (8129)." MIT RncliationLaboratory Series. vol.57 1. the transformation ratio [Eq. sec. which. (8130).69 1. in contrast to Eq.has been retained on Y o. (8129). (8133). 825 illustrates a very useful way of viewing the waveguide junction T 1 :n T of Fig. Inc. Note also that the characteristic wave impedances of thc two guides.10 2. The alternative equivalent circuit of Fig. ~ ic '/r'a [Bin (7fe/a) ]' 1 .. can be identified as J!.67 1. (8~137) will be larger than the true solution. (8137). CoMPARISON OP EQ. "Waveguide Handbook. the normalized shunt susceptance as :IDCl: FIG. we then find the transformation ratio of the ideal transformer as n. are now differentj so the superscript .. We bave separated the shunt susceptanee into two parts.nd.93 2.44 N. 49 we assumed E. ~ . Marcuvib:. solution of Sec. 824a. New York.0. Alternative equivalent circuit lor Fig. Note that. the value of B/Y o obtained from Eq. 419.3 0.V.' YoVo t jB+ Y o+ = .G = u~f(%) = Uwsin ~ c (8135) By Eq.
and (b) an equivalent circuit. 826b. B+ is approximately B/2 for the diaphragm problem corresponding to the guide z > O. our formulas are stationary. in the region z > 0. each coupling to one mode. incorrect. When morc than one mode propagates. there will be N ideal transformers in series.MICROWAVE NETWORKS 425 n : 1 T T Yo. The analysis will be exact only for zerothickness junctions. as illustrated by Fig. Bay N modes.depends only on guide z < 0. 811. Let the feed be viewed as l\ sheet of current J.) Then. An equivalent circuit when only one mode propagates is shown in Fig. (a) A thin coax. II Side view (0) r End \liN r (b) Fla. By thin. 826. we effectively divide the problem into two parts. The justification for this equivalent circuit will be found in the analysis. This assumption is. of course. we have (8139) where r. T Yo. we meaD that the dimension in the axjal (z) direction is small. (8131). Note that B. such as Figs. so B. each part relatively insensitive to the other.in the junction problem is approximately B/2 in the corresponding diaphragm problem.perture susceptances according to Eqs. (8138).+ is the +z reflection coefficient of the ith mode referred to . An aperture susceptance calculated for the aperture and ODe guide. We shall DOW consider thin coaxtowaveguide feeds. where the Vi and V"i are given by Eq. which is usually small. 8260. (This neglects the effeet of the gap. in the z = 0 cross section. by defining a. Waveguide Feeds. 417 and 419. Similarly. assuming E t in the aperture is unchanged. thereby becomes useful for a wide variety of problems. and in particular is onehalf tbe shunt susceptance of a diaphragm.towaveguide feed. but. Hence.
for z < 0... We now use the stationary formula of Eq.. is the characteristic impedance of the ith mode. and Eq.426 Z  TIMEHARMONIC ELECTROMAGNETIC FIELDS O. = II"s ~ V. are known.. .)' + (1 z. This is a sta.H .)' r. (1  r. (8139) for E. (8139) and (8140).s9) to determine the impedance seen by the coax. 1  r. 2: v. We have ensured continuity of E...' L. we have v... substituting for Vi from Eq. (8141) for ].. .y. X h . d.+) Finally. we have 1 '\' Z. (8140) where rr is the z reflection coefficient of the ith mode referred to z = O. and integrating over the guide cross section.+ 1 + r. (~ ~ ~:= + ~ ~ ~::) u. We can put it into a slightly different form by noting that .. = u. we obtain ZI.+)(l + r. where the integration extends over the z = 0 guide cross section and II.) 1 0 .I. X (H.' II EJ. The boundary con~ dition on H at z = 0 is J. (ff J•. ZOo e.. is the current at the reference plane T'. Using the first of Eqs.+ r. )(1 + r. (8142). (8141) Multiplying each side bye. Similarly. d8 (8142) The field is then completely determined if the r's and J. Y i 1 1 '\' • (1  • + r.. e..+)' (8143) where Z. at z = 0 by choosing coeflia cients Vi the same in both Eqs. Y.+ . This formula is ~  I~..d..II J.tionary formula for the input impedance of a zerothickness COQxtowaveguide feed. (7. G~ ~:= + : ~ ~::) .
may be taken as zero unless some obstacle is close to the feed. Eq. sin ked  x) '(y . (8148)] divergcs. ~ Zl • 1_ + r.x) a(y . "11 b Equation (8147) is therefore n "'" giving ~ foci dx 10 dy sin ked .SlOtC tan t kk'ab b 2 The summation for X [Eq. As an example. Vab sm 6 /2 .?tfiCROWAVE NETWORKS 427 the wave impedance of an ith mode referred to z = 0 is b..c) sin ? T (8150) n' . because the current was (d) . = except i = 0. consider a probe in a rectangular guide (Fig. (8144) Hence.  1~.c) x<d x>d (8149) where k = 2r/X is the wave number of free space. of course. then all Z. r. ~. (8143) can also be written as (8145) This shows that the guides z > 0 and z < 0 appear in parallel for each mode. and all r. Equation (8143) or (8145) can then be written as ° (8146) where (8147) (8148) Equation (8146) is. 827). Nonpropagating modes decay exponentially from the junction and their r.2 . If we assume that only one mode propagates. provided the terminations are not too close to the feed. Assume J. just that for the equivalent circuit of Fig.826b. are imaginary except i = 0.. vector is eo = The dominantmode u.
as in Fig. 410.llobstacleanalysis that X is capacitive (negative) for a short probe and is of the order of magnitude of X for a probe over a conducting ground plane. rectangular obstacle approximation of Sec. ~ II (E X H . 812. ds). Indeed. . As far as the waveguide is concerned. if the probe is very thick.u:t. it. A variational solution to the problem can be obtained by assuming tangential E in the aperture. How~ ever. (491) n=b~ . we shall have to modify the equivalent circuit of Fig. The general problem is represented by Fig. 627. We now wish to consider conducting bodies containing apertures excited by waveguides.. 420.. when the probe is connected to the opposite wall of the waveguide.. From our equivalent circuit (Fig. The reactance of a short probe can be estimated by the smallFlO. 82Gb. Note that our present solution [Eqs. ds). we can think of our present problem as a junction between the waveguide and external space. calculating the resultant fields on each side of the aperture. 428.. is evident from the sn. specialized to a rectangular waveguide matched in both directions.. (8146) to (8148)]. For example. n! ~o 2 n Z. (tan lea)! . 828a. ~ 'll"C SInb" (8152) Other possible feeds are shown in Fig. 88. 410. ! 2R I • (8151) where R l • is the quantity calculated in Sec.TIMEHARMONIC ELECTROMAGNETIC FIELDS taken as filamentary. is the same problem treated in Sec. 810 for the waveguide junction. Hence. 2. It wAveguide. pe~t (8153) This is the same approach that we took in Sec. 826).pen . If the probe is taken as circular in cross section.•• . = under matched conditions. Probe in fl. Excitation of Apertures. the reactance can be evaluated by methods similar to those used in Sec. the aperture appcars simply as a load across the reference plane T. is evident that the coax sees R . 87. and then conserving the flux of reaction by II (E X H . we have from Eq.
' X H'· d...V. (8156) where V is some reference voltage and H.. Let us therefore abstract from Eq.Side view Ca) Conductor . For tbe external problem. d& Y. is given by Eq. (8154) (8155) These formulas give the internal shunt susceptance B in terms of an 8.. Aperture I End view (b) I FlO. they are the same as variational solutions.is the external magnetic field calculated from the assumed E 1. S28.~. : I . 411.was assumed real. .. 828b. .' jB where V. the external admittance of the aperture." E. we dofine an aperture admittance as  JJ Eo' ... (8138) LY..& in the aperture. We have anticipated this separation by taking the equivalent circuit as shown in Fig. (a) An aperture excited by a waveguide. (These calculations were made on a conservation of power basis. The internal problem is identical to onehalf of the waveguidejunction problem.  Y . .. The ideal transformer accounts for possible differences of impedance reference in the internal and external problems. but. beeause E. (8157) T l' n .S8umed E.)(ICROWAVE NETWORKS Once the tangential E in the aperture is assumed. but we have referred the aperture admittance to V j hence "I V' n'. ff ... Y. where jB represents the internal susceptance of the diaphragm and Yap••. external and internal.V. and (6) an equivalent circuit. jS I y.) To determine we note that the dominantmode voltage coupled to the aperture is V. V where V. ~ ~.. the problem separates into two parts.. (8155) applied to the dominant mode.. Examples of some apertureadmittance calculatioll8 are given in Sec. e.I ..
ngula. Univ.8 alb . .! Figure 829 shows the aperture admittances for a square aperture and for a rectangular aperture with Bides in the ratio 1 to 1 and 2..25 1/ 0. 1 1 Cohen. 26. then n . waveguide of the same dimensions as the aperture. 22.r aperture in a conducting ground plane.. and Levi.) . /i' . Crowley.0 0.· in the aperture can be found by the methods of Sec.'Ib . Aperture admittance for rectangular apertures in ground planes. . as shown in the insert of Fig. . Qhw Stal~ UnilJ. g. 1957. The field due to E.:. Additional calculations have been made hy R. Hence. referred to the dominantmode voltage of a rectangula. a/~~ .1. Rept.002 o C 0. The aperture admittance has been calculated for the assumed field E •• = . The Aperture Admittance of a Rectangular Waveguide Radiating into Half6paee.r waveguide of the same dimcDsiona. The mathematical details are tedious but can be found in the literature. Crowley. . referred to the voltage v=jJ (8159) which is the dominantmode voltage for a. . / 'V .004 . The Cavitybacked Slot Antenna.6 ]. U 1I 810 7' a (8158) in the aperture. ac 21114 SR no. (8156).I' . 36. ~G 0. Anten1UJ Lab. (Ailer Cohen..~I " " CEJI Iaol 0. H153.29.25 to 1. IUiMi4 Antenna Lab.006 FIG.k.J 1.430 TIMEHARMONIC ELEC'l'ROMAGNETIC FIELD8 0. and the aperture admittance calculated by Eq.:JL 0. Tector.004 I alb 1. &pt.002 I I 0. J. when the aperture is simply the flanged open end of a rectangular waveguide. 829.2 / .. and Levis. . An aperture of practical importance is the recta.
.s m  • • Hence. 830.)... Consider a cavity formed by a perfect conductor enclosing a dielectric medium. 830.25Y.r aperture in ground plane. where Y. (8157) and (8159). giving  B Sa rb a R:: Iogcsc.2. b by 012.54 :. An approximation to B is therefore given by Eq. A square waveguide £eediDg a reeuDgula..25. 829. suppose we have a square waveguide of height and width 0 feeding 8. and c by b/2..25 curves of Fig.(JCROWAVE NETWORKS 431 As an example. hy Eqa. 2a X... for which V2 .54Y. . 813. Modal Expansions in Cavities. Each mode must y Side view End view & Fla. rectangular aperture with sides in the ratio alb .•• is given hy the alb""" 2. we have V o . 01" a 0 dz 1~ dy sin' ~ 0 0 = _b_ V2 and eo.!.25 • The shunt susceptance B is onehalf that for the diaphragm of Fig.. n' IC1 b . as shown in Fig. 822b.= 3. (8155) and (8158). the terminating admittance seen by the waveguide is y ~ j3.. "" eo = u.2. + 2. The waveguide is excited in the dominant ypolarized mode. Hence.. (8120) with B replaced by B/2. by Eqs.
is chosen real. we wish to normalize the mode vectors." chap.0 (8161) valid even if E and p are functions of position. 1953. Either Eo or Hi may be eliminated {rom the above pair of equations.jWiJlH.:> .432 TIMEHARMONIC ELEC'1'ROMAGNETIC FIELDS satisfy the field equations V X Eo' "'" . Furthermore. l Hencc. coupled with the boundary condition n X E.nner. Now suppose that electric sources exist within the cavity. Hi form a complete orthogonal set in the Hermitian sense. corresponding to the orthonormal Ei are i¢j i=j (8164) We have alrea. then the cor· responding Hi is imaginary.tions are then v X E = jwpH and the wave equation is v XH ~jw.dy shown in Sec. Morse and Herman Feshbach. the eigenvalues Wi (resonant frequencies) are real. part I. and the eigenfunctions Eo.w'EE "'" jwJ (8165) I Philip M. the timea. New York.IJlH.". . .verage electric and magnetic energies are equal. = 0 V X (elv X Hi) . tions. so that the orthogonality relationships are ir&j i = j (8163) Normal which can be derived from Eqs. 6. as suggested by Fig. . is an eigenvalue problem in the classical sense. also normalizes the Hi. Hence. and vice versa.1V X E. The field equa.) ~ Each of these wave equa0 on S .E+J V X (plV X E) ..'"E. giving the wave equations V X (1l.w. 84 that if E. (8160) in the usual ma. 831a.<8162) where n is the unit normal directed outward from the cavity boundary S. McGrawHili Book Company. because tha. the orthogonality relationships for the H. for f and Jl real (no dissipation). "Methods of Theoretical Physica.n X (c'v X H. ization of the E.t is.. (8160) where i is a mode index.
' . which. A.jw fff J. caD be written as L . In fact. but finite. we can let (8166) SUbstituting this into Eq. represented by Fig. are a complete set.HICROWAVE NETWORKS 433 D Flo.jwJ . the field becomes infinite at a resonant frequency in the lossfree case. in any physical problem there will always be some dissipation. The dual problem is that oC magnetic sources in a cavity.w'. by Eq. (8161). the wave equation in H is (8170) . and we have (w. (8165).E. = jwJ If each side is now multiplied Bcalarly by Ej and integrated over the volume of the cavity.pproaches Bome resonant frequency.] . obtained from the field equations.[V X (. (0) (h) Because the E. (8163)].) . A cavity containing (a) electric sources. Actually. the field is large. E~ dT (8167) which determines the A" Hence.2  W 2 )fE. 831. (8166) becomes (8168) and the corresponding H. Hence. and (b) magnetic sources. is (8169) Note that the field becomes extremely large as W a.use of orthogonality [Eq. 83Ib.'V X E. In this case. . Eq. which is to be expected. &11 terms except i = j va. at all real resonant frequencies. so the Wi are complex.(w. we h&ve ~ A.nish beca.w')A.
are given by (w. (8167). the effect of dissipaI The eigenvalue Wf . Phll. the expansion of H due to magnetic currents M is l H _ '\' Lt w' ". the B.J W' ..E.0 must be included in both EqIJ. 814. AB long as there is no dissipation. as shown in Fig.. Mathema. (8175).t jwH.he cavity is lossy but high Q. See. March. rrr M. Probes in Cavities..434 TIMEB. we cnn represent a probe in a cavity in terms of electric currents in the cavity. The impedance seen at the input terminals to the probe can then be calculated by the varia.H~dT (8173) and the corresponding E field is E. J. 831a. Teichmann and Wigner. (~168) and (8173). (8172) Hence. However. and 1 is the corresponding input current. is the assumed current distribution on the probe. Appl. vol. 24. This effect is usually negligible and can be taken into account by the methods of the next section. (8168) with the' dropped.0 account for the irrotlltional parte of E and H.ARMONIC ELECTROMAGNETIC FlELOS \Ve then expand H in terms of the orthonormal mode vectors a as (8171) where. we obtain (8176) where (8177) The analysis neglects the effect of the aperture through which the probe is fed.tically.'\' i.. for example. if t. the input impedance will be purely reactive.' Jll . will be chosen realj 80 the field produced hy JO is given hy Eq. w.( 1 )B. Jll rrr M.. we can superimpose Eqs. dual to Eq. The modee a.tional formula (8175) where J.' . All mode vectors E.ssociated with Coli . lOSt . = jw Iff M· Ht d.H~dT (8174) If both electric and magnetic sources exist within the cavity. (8168) and (8174) for a solution. jw. Substituting this equation into Eq.
)' where ao is obtained from Eq." p. we need only compare the formula for the impedance of the parallel RLC circuit jW/C wo'(1 wo! = LC I + j/Q) Q~ R wL R woL (8181) with the last term of Eq. 1955. (8176) by ZIII "'" jX  w2 jw(ao/ f)' wo 2(1 j/Q) + (8179) where X is the reactance due to all modes except the i = 0 mode (8180) The effect of dissipation in modes not near resonance is negligiblej hence. N. To determine the values of R.!..1fz Sill  (8182) For the J". (8179) is shown in Fig. we assume Eo = U". (8179). current on the probe. it is not included in Eq. L. In the probered An equivalent circuit ror or cavity in the vicinity vicinity of a resonant frequency.'(l +~) (8178) FIG. "Network Analysis. according to l w. a where Q is the quality factor. An equivalent circuit which represents Eq. The normalized mode vector of the dominant mode is VEabc b c where the normalization factor was obtained from Eq. we can approximate Eq. lne. 832.x)!( _ b')!( _ c') smkd Y y x<d x>d (8183) o I M. It is then evident that R _ Q wo (afo )' L _ (. consider a probe in a rectangular cavity (Fig. (8177).)' fwo C= U.o= { I sin k(d .!1. . To illustrate the theory.J. 833). 364. Wo (not necessarily the dominant resonant frequency). Van Valkenberg. PrenticeHall. say resonance. Englewood Cliffs.' ~ W.1I'y. 832. (297).. E. (8180). and C. = Sin 2.raaCROWAVB NETWORKS 435 tion can be taken into account by letting the resonant frequencies be complex.
(8177).' " II FlO.!L__...~. The treatment or loops in cavities is essentially the same as the treatment of stubs. once a current is assumed on the loop." and conserve reaction according to . We cannot. (8185) ." X H"· ds). we have = a. 815.ff (E. of course. X will be of the same order of magnitude as for a stub over a ground plane. 824 and 826. r oL . 834a. (2101). Probe in a rectangular cavity. (295). &33. I kYtabc 2 tan (kV'(b)'("') 2 b slnwsmc (8184) The other parameters needed to evaluate R.. given by Eq. I I 1 J" '' x .436 TIMEHARMONIC ELECTROMAGNETIC FIELDS rI+. and C are the resonant frequency fr = wo/2r.. for which it is capacitive. (8183) to evaluate X..iT . The actual diameter of the stub must be considered. since the resulting reactance would be infinite. . use the filamentary current of Eq. the reactance is capacitive. for short stubs. Some explicit loop feeds are considered in Probs. given by Eq.· X H"· ds). by Eq. L... To a very rough approximation.1 tx . and the Quality factor Q. Then. apert apert u.i~T TO b' Z ty 11=. The general problem of coupling a cavity to a waveguide through an aperture is represented by Fig.==:=l. For a variational treatm~nt or the probleml we assume an aperture field E. The evaluation of the series reactance X is a much morc difficult problem.".rf (E. When the stub is bent into a small loop and joined to the cavity walll the system is often called a loop feed.. The series reactance X for small loops is inductive l in can· trast to the smallstub case. Hence. Aperture Coupling to Cavities.
T (a) (6) ~ r~""N'C. (b) the cavity..Y. circuit in tbo vicinity of re8Clnanee is Abown in (d). ri'7 (8187) .wo parts.•• = _pert. Hence.1. 834b and c... And (e) the waveguide. Y. The waveguide part of the problem is identical to the problems treated in Secs. we can rewrite :Eq.. as shown in Fig.V.. each side of t.' X H··ds) ••.o t. the Y II are the modecharacteristic admittances.)' V o (8188) .. which amounts to dividing the original problem int.. (8186) in the cavity part is the negative of the terminating current in the waveguide part...his equation can be considered separately.IOCROWAVE NETWORKS 437 For a given E. 810 and 811. The equivalent current M.J f (E l. and is therefore of the form JJ (E. and Y is the admittance seen by the dominant mode. = 1 . . ••• (v.... T 1 :n R .~ot where Yo is the characteristic admittance of the dominant mode and B. YVo' + l ... An equivalent.=nxE.. (8185) as ~o = jB.X H. dS).L: Y....L (el (d) FIo_ 834. (a) Aperture coupling from a wavelUidt! to a cavity caD be divided into two p&rts.' where the V II are the various mode voltagcs.._. 0 Y..
was treated in Sec. = '' .: \' jwb._..' Wi! c V' (8192) and.' w2 W.. Taking the mode vectors Hi as real.= w' wo'(1 Wa' = LC + j/Q) 1 1 Q = wCR. (8187) can be written as The first term in the brackets represents the susceptance due to all nonresona.rt of the problem. as shown in Fig. we caD determine the field by Eq." X HG • ds). The formula for admittance of a series RLC circuit is 1 y = .= w. For the cavity part of the problem. we again take losses into account by Eq. The.t.. we have introduced the ideal transformer •• n'  (:... and Eq. (8191) as a series RLC circuit.y (8193) Finally.calculation of B.ods (8l90) In the vicinity of a. 810.2 (8189) where b. The above equation caD therefore be written as Yo ~J y oS + n' [oS II Yo J 0  w1 jw(bo/V)'] wo'(1 +ilQ) (8191) where the . we obtain The righthand side of Eq. to account for an arbitrary referencevoltage V. and the second term gives the resonant..°w"'/L:.Jr. resonant frequency. (8186). mode effect. ..CR . (8173) with the current given by Eq. (8185) is then given by .. (8178).nt modes in the cavity.. we can represont the last term of Eq..susceptance due to nonresonant cavity modes is S~~ w' b.pet~ 1f (E . 834d.438 TIMEHARMONIC ELECTRO~tAGNETIC FIELDS is the shunt susceptance introduced by the waveguide pa..  If Eo" X H.
MICROWAVE NETWORKS 439 Comparing thi:5 with the last term of Eq.f. Also. 4a' 1l'd Yo A. Cavity T ·1 ~ b Top view ~ . ~ R Q "'0 (bVo)' c . we see that . (8192) directly.rticular problem. let us view the aperture as the junction between two waveguides of height a' and a. 835. (8196) if the waveguide were the same height as the aperture (n 1 would be 1 in that case). ~ • d ±I Side view II .te the above theory by a. B. The waveguide part of the problem is identical to problems previously considered. 835. 2a' (8195) For the cavity part of the problem. The susccptT Waveguide FIG. let us make our oftenused assumption Et = U. (8197) which is the waveguide dominantmode voltage that would be excited by Eq. treatment of the rectangular waveguide to rectangular cavity junction. (8190). Aperture coupiing from a rectangular waveguide to & rectangular cavity. hence n1 = d a' (8198) Rather than calculating Eq. In OUf pa. shown in Fig.y (8194) where bo is obtained from Eq. In particular. Let us illustra. will be approximately onehalf of Eq. or ~Iogesc B.. (8191). let us refer the eavity admittances to v.SlD1J .!. ry (8196) in the aperture. (8120) with the appropriate interchange of symbols.(~)' Vwo L ~ (:. the waveguide dominantmode voltage is V o = v'ba'/2.
.. .Jif.. . ") ( u~slnbcosu.. L. Consider an 2:·directed current element II at the point z'. (8195) with af replaced by B. v(mbP + (nap cos m. (2101). z' in II.. PROBLEMS 81. (834) and.mn< b +. Show that the field is given by formulas of Table 81 where 'It's are given by Eqa.. all parameters of the equivalent circuit (Fig. . nwy +.ac(1 2d c/b)' V + (1)201) The resonant frequency f~ = wo/21r is given by Eq.• . 2. ..x' sin flry' 7_10.. normalited on a per unit width basis. (295) and the quality factor Q by Eq.L Vb v'2b . •y .). which is Ho = vi J1abc(b 2 + 0 1) 2 .. and C.b planes. we obtain (b... v. _ [1(Z) .. But we wish to refer it to the V of Eq..... y'. 216).. and (8197).~f•• where v. rectangular waveguide (Fig. arc 1J . ~ . for 11.0. 3. f . log esc 2a (1)199) Finally.y. _ . 834d) have been evaluated. so we should multiply by d/a and obtain B~ R: 4d .. Consider the parallel. to determine the R. (8196).0 and y .. ...' = where 11. . (8190). . .d X. (8197).. . +. m 'po! 0. .. we need the normalized dominantmode vector.. V2b nrll '0'b n< 1. from EqB. 82.ccoob6InC C (8200) Hence.platc waveguide formed by eonductors covering the Show that the eigenfunctions. Hence. for m . .'1 0 fllb'+na' abe and.440 TIMEHARMONIC ELECTROMAGNETIC FIELDS ance Be referred to the mode voltage of a waveguide of height a could then be approximated by Eq.
'W. for Z.V~/Vf. 44) andin circular waveguides (prob. and. For the general cylindrical waveguide (Fig.(1 + ZII)(ZU .1) 1 '" 1Iu + . show that the titneaverage electric energy per unit length of guide is and the timeaverage magnetic energy per unit length of guide is Note that these are just the sum of the energies in each mode alone.(1 . for general cylindrical waveguidCil. . . show tbat. Wit S 1. the attenuation constant due to eonduct<lr 10Sll is m.i.'" 2'. 87. 59). r is real.III)(:U ... <JI. and characteristic impedance Zoo Show that Z.1) . Show that the transmission matrix IT] is related to the impedance matrix [zl by 2T Il 2T 1t   2Tu  2T. denotes inVin.'1' '" I Zu + .7'. real. 2"fl" for TM modes. [. Derive Eqs.sic resistance of tbe metal walls.{.Z")(:II + 1) '" + I (1 + + 1) '" Zll ZU)(:II 1111 + I (1 .{. 88. propagation constantjJ3. Use the above formulas to determine the attenuation in rectangular waveguides (Prob. • Bence.MICROWAVE NETWORKS 441 8S. "intrinsic impedance of the dieleetric. Let the characteristic impedances of ports (1) and (2) of Fig. and 1'" k ¢(''')' 'an t  dl 1 III ~ j a. 86. and the other symbols correspond to their usage in Table 81. "y (+"Pdt ] flt . 81). 84.. (1 1 lrtt)IVflt/Z. ('''')' dt iff + "'. tan 81/2 815. Let the T equivalent circuit of Fig.T:~ and Tit . Using the usual perturbational method. for TE modes.  Show that in the lossfree case Til ..ne the reflection coefficient r . . Consider a oneport network. at resonance. (872). Show that.jZ. and dc6._lV f lt 1m (r)/z. 80. 87 bo normalized to unity. 89. in a BOurcefree network. . and ~ . 81Ob represent a section of waveguide of length I.jZo esc fJl z.
812. such t..1 a '''·· In the aperture. aperture by Approrimating B. Derive Eq.. 8If.ennioe M... of the magnetic conductor terminatioa t. nctwork parameters &nl B. 823. The approximations are good for die 14  Y.3 And at>. tangential E must be of the form Show that is a variational (onnu1& (Of the ahunt IU8Ceptance.he 6eId aa H .2. Note that it gives upper bouads to B/Y..fdl 2011./e).. Show that t. Y. Centered capacitive post in a rectangular waveguide. 811. Add a magnetic current. 811. . . sheet. Expreee t. .L . of Fig.hat.ej(%) . (8122) is a crude variationallOlution (or the shunt IU8Oeptance. Consider the indlJt:t. 823) and the variational fonnulaill terms of obstacle current {Eq.beet M.ional dominan~ mode lIOurce.442 TIJdEliARMONlC ELECI'ROHAQNETlC FIELDS 810. tj. 836. < 0. On the diaphragm. 816..sin (77. The inductive diaphragm (Fig. the entire field must be TM to Jj'. with the electric current. Problem 8J3 is the special cu. the t:urrent is of tbeform ]. Det. (. mode ill TM to 1/. 813.IID . coineident. f Side view End view Equivalent circuit Flo. (896). 823) baa boundaries cylindrical to r.) . Conaidu the inductive diaphragm of Fi&. they are & uni~t. T . (8107»).]. The incident. . sending waves in the +~ direction only.where • '\' A . in the ehow that Eq. 836. . < 0. and ]."" . llhown in Fig.. hence.he euide. Determine the Idfreaetioa of this source in the presenu. Consider the centered capacitive poet in a rectangular w&velUide.he equivalent.ive diaphragm (Fig.V X uri.
ltIICROW AVE NETWORKS
443
Show that
•
Y.
a
/:12 v(n/2)'1
'I;'
 11  2>.
[!.c [!.,. a d:l:
0
(4/X)'
,,(x) sm
]'
•
n.z
4 en
J'
c g(x) Sin
TZ
is the variationallormulo.lor lower bounds to B/Y•. 8·16. Show that the shunt susceptance of the capacitive diaphragm of Fig. 837 is given by the same formula 88 applies to Fig. 8228.
FIG. 837. A capacitive diaphragm (metaJ shown dashed).
817. Consider the capacitive junction of Fig. 838. Show that the parameters of the equivalent circuit are B+ 4b+ ...c   log CIlCYo ~ 2b+ B4b...c I,.,~ Yo ""' 2b

n
l
b
 b' 
Use the approximation of Eq. (8120).
1 {~;;I j
Side view End view Equivalent circuit
FlO. 838. A capacitive junction. 8·18. Considc{ the waveguide junction of Fig. 824a and the equivalcnt circuit of Fig. 825. Show that, Analogous to Eqs. (8138),
Y,jBand n l

101/1,t. The mode current.e arc given by
1, 
ff H," !i,do
where H,+ and H, denote tangential H on the +z and z sides of the junction, retlpcct.ively. Variational formul!l8 are obtained by M5uming H.+ IUld H, subject to the restriction H, +  H, in the aperture.
444
TIMEHARMONIC ELECTROMAGNETIC FIELDS
819. Let 0#(Z,1I)  !(P,.) be a 8OIution to the twodimensional sourcefree Helmholtz equation p < a. Prove that
where
,i..e is
an operator defined by
Sin
. D
..,.....
,.
}k
ax
COllD..,....}k
1 •
ax
and ,i.D.J(O) means e'·D.p(z,y) evaluated at value theorem.
:t 
0, II  O.
This is a. kind of mean
820. Consider the eoa'll: to waveguide feed of Fig. 4ZO. Let d denote the diameter of the coa.xialstub, and let a «>.. Show that, for the equivalent cireuit of Fig. 826b,
1 "r<: n""l),nn"b a "l'kd X AI ,,)og• 2
2a
where 'Y ... 1.781. 821. Let the rcctangular aperture of Fig. 829 be thin (b« A) and of resonant lengt.h (0 ... ),,/2). Show that
Him: Usc the duality concept of Prob. 743 and the approximat.ioD8 of Prob. 739.
Note that the aperture radiates only into halfspace. 822. Figure 839 represents a parallelplate transmi$8ion line radiating through & slot into balf.space. Let Fig. 828b represent the equivalent circuit, and evaluate the parameters, using the aperture admit.tance of Fig. 422.
b ',~
•
.
FlO. 839. A parallel. plate t.ransmi88ion line radiating into half~pace.
MICROWAVE NETWORKS
445
823. Figure 840 represents a rectangular waveguide having aides a, b radiating into halfapace through a narrow resonant slot. Uaing the results of Frob. 821, show that reflectionless tranamisaioD through the slot occurs when
;; ~ HI  (~)']jl
When blA
4
0.54 coa' (..A/4b)
< 0.7, the above formul& can
be approximated by
~~~..)l(~)· )" 3 2b
The waveguide is excited in the dominant mode.
FIG. 840. A rectangular
waveguide radiating into half·space through a resonant slot.
r b 'I rr..,
U ..
t1
I I
! I
J
824. Considcr the loopfed rectangular cavity of Fig. 841. AMlumc that the loop is arnall, 80 that the current on it may be flS8umed constant. Show that the eJements of the equivalent circuit (Fig. 832) are given hy Eqs. (8181) where
a, 2• . T  .v;a:bC SiD
Whcn c' «c, this reduces to
ao
(b').(") b sm .. c
..
/ ...
cv~c
2...A SID ( b') . _r=c "b
where A  c'd is the area of the loop.
FlO. 841. A loopfed rectangular cavity.
446
TUtEHARMONIC ELECTROJl,lAGNETIC FJELDS
826. Show that the normalized mode vector for the dominant mode of the circular
cavity (Fig. s..42) ill
where
ZOI 
2.405.
o
(bl
FlO. 842. A circular cavity with (4) probe feed, and (6) loop feed. 826. Figure 842a represents a probefed circular cavity. AMume sinusoidal de. tribution of current on the probe, and show tha.t the elements of the equivalent circuit (Fig. 832) are given by Eqa. (8181) where
~_
I
ka y;;:b 1 1 (%01)
1
tan(k~)J,(zo,~)
2 a
and Xu  2.405. 827. Figure 842b represents a loopfed circular cavity. Assume unifonn CUm!llt on the loop. and show that the e1ementll of the equivalent circuit (Fig. 832) are given by Eq8. (8181) where
7  a v;;bd J o (ZIIE) 1,(ZOI) a
Show that, when c "" a, this reduces to a. 
1
~
A:tol
al~
where A  (4  c)d is the area of the loop. 828. Recollllider Fig. 841 for the case of a Bmallioop. Represent the loop by • magncticcurrcnt element Ki, according to Fig. 33, aDd evaluate
R _Re{a,a} .. Kl·B
[I [I
The result i.e. the asme as the small·loop answer in Prob. 824. 829. Reconsider Fig. 842b by the method outlined in Frob. 828. Show that the result is the 8I1me 1UJ the smallloop answer of Prob. 826. 8~SO. Show that the normalized H mode vector for the dominant mode of the spherical cavity (Fig. 62) is
H  U+
'va,.
O:fJ~ J,
(2.744!) Bin 8
a
APPENDIX A.
VEcrOR ANALYSIS
We shall normally orient rectangular (:r;,'Y,z), cylindrical (p,<p,z), and spherical (r,8,<p) coordinates as shown in Fig. Al. Coordinate transformations are then given by x=peos<prsin6cos<p 'Y ". P sin tP . r !!lin 6 sin <p
p "'"
z=rcos6 y:r;1 tit _ rsin 8
+
tP
z:
tan t
+ tit + z' t t 8 """ tanI Vx + 1/ _
r _ y:r;'
•
1!
(AI)
vpt
•
+ z! tan1 e
•
Transformations of the coordinate components or a vector among the three coordinate systems are given by
A.  A" cos tP  A. sin tP  Arsln8 cos tP A,l =0:0 All SiD tP + A. cos q,
+ A,eos8c08tP
A.sintP
"'" Arsin 8 SiD tP A,cos8sin tP A, "'" A r cos 8  A, sin 8 A" "'" A.C08¢+ A,sin¢ =0:0 A,sin8+A,eos8
+
+ A. cos q,
(A2)
A. =  A. sin tP + A, cos cfJ A. = A.sin 8eGS ¢ A,lain 8ain tP
A, 
+ + A,cosS  All sin 8 + A. cos B A. cos 8 cos tP + A. cos 8 sin tP  A, sin 6
"'" A" cos 8  A, sin 8
The coordina~unitvectors in the three systems are denoted by (u.,u.,u.), (u",u.,u.), aDd (ur,u"u.). Differential elements or volume are
d,  dzdytb  pdpd~dz  "siD qd,dqd~
447
(A3)
TIJLE.ElARMONIC
ELECTRO~QNET1C
FIELDS
z
z
...........................
r
Flo. AI. Nomul coordinate orientation.
X
z
I /' Pl..../
y
differential elements of vector area. are
ds  u.dydz
+ u.dzdz + u.dzdy  u.P d~ dz + u. dp dz + u.p dp d~  uyr t sin e de dt/J + u,r sin 6 dr d4> + U ..1' dr de
d1 = u.dz = u.dp = u.,.. dr
(A4)
and differential elements of vector length are
+ u,dy + u.dz + ..pd~ + u.dz + u,r de + u,.r sin e d,p
(A5)
The elementary algebraic operations are the same in all righthanded orthogonal coordinate systems. Letting (Ut,Ut,u,) denote the unit vectors and (AI,At,At) the corresponding vector components, we have addition defined by
A
+B
= u.{A.
+ B.) + u,{A, + B,) + u,(A. + B.)
+ AtB t + AlB,
(A6)
scalar multiplication defined by
A· B = A IB I
(A7)
and vector multiplication defined by
(A8)
The above formula is a determinant, to be expanded in the usual manner. The differential operators that we have occasion to use are the gradient (vw), divergence (V· A), curl (V X A), and Laplacian (Vito). In rectangular coordinates we can think of del (v) as the vector operator
(Ag)
¥ECI'OR ANALYSIS
449
and the various operations are
U.
Ur
u.
VXA=!..
()x
ay iiz A. A. A.
d
d
(AlO)
V'w ~ d'w ax l
+ d'w + d'w
alii
az l
In cylindrical coordinates we have
In spherical coordinates we have
law I aw + U. + ur.SID.0a¢ r ao la 1 a .) 1 dA. • v, A = (r S A) +. (A ,smO +.
vw
=
u.iir
T10T
iiw
VXA=u.. A.sm8  r 8m e ae of/>
1 [d ( . ) dA.]
T SID 0 00
r 510 0 o¢
1 dA. + U, [ I.    ar (rA.) ] a T 8m8 o¢
(Al2)
Vlw ==
,1
.!:. ~ (TS ()w)
ar
1 ~ 1 ar + ,1 sin 0 as (Sin 0 iJw) + r l sin l 0 alvJ ao a¢1
1 +u. [a ( r A . )aA.] riJr a8
A number of useful vector identities, which are independent of the choice of coordinate system, are as follows. For addition and multiplica
450
TIM&HA1UtONlC ELECTROMAGNETIC FlELDS
tiOD we have
A'  A· A
IAI'
A
= A·A'
A+B=B+A A·B  B·A
x
B
~
B
x
A
(A13)
(A + B) . C = A· C + B . C (A + B) X C ~ A x C + B X C A·BxC=B·CXAC·AxB A X (B X C) = (A· C)B  (A· B)C
For differentiation we have
V(. + w) ~ V. + Vw V . (A + B) = V • A + V . B V X (A + B) = V X A + V X B v(vw)  • Vw + w v. V· (wA) = wV . A + A· Vw V X (wA)  wV X A  A X Vw V . (A X B)  B . V X A  A . V X B V'A ~ V(V· A)  V X V X A v X (v Vto) "'" Vv X Vw v X Vw = 0 V·VxAO
For integration we have
(A14)
JJJ V·Ad<  effi A . ds JJ V X A·ds  ¢A.dl JJJ V X A dr   effi A X ds JJJ Vwdr = effiwds JJ n X vwd. = ¢wdl
Finally, we have the Helmholtz identity
4rA V
(AIS)
( rl f'J.y Irv'·A
dr'
+V
X
f'(yv' X A
J. 1r
r[dr'
(A16)
valid if A is wellbehaved in all space and vanishes at least as rapidly as ,.1 at infinity.
APPENDIX B
COMPLEX PERMITTlVITIES
The following is a. table of relative ac capacitivities dielectric loss factors E~' where
~, "
E~
and relative
= 
"~
EO
E' = EO
J
. E"
Eo
= E,,
 JE,
. 11
is the relative complex permittivity. The measurements, along with many others, were reported in II Tables of Dielectric Materials" (vol. IV, Mass. Inst. Technol., Research Lab. Insulation, Tech. Rept.). They also appear in Part V of U Dielectric Materials and Applications," Technology Press, M.LT., Cambridge, Mass., 1954.
... 24 23 " 10"..... . 7.9 360 2.7 150 2..17 205 2.. 4.48 148 5......' 2.41 ..25 .. ..... .37 365 5.17 . 3 120 2..25 ... " lO'e:.17 ...17 16 5.10j~' lOtf~' . .25 25 25 24 25 360 2....17 1 165 2.17 6 145 2. (2 per cent iron oxide) ...17 2.56 680 2.. .. 9.....1?' Beeswax (white) .63 310 2..... .15 585 2. Glass.6 234 3.73 570 17 3..7 116 5..23 130 13 6... cycles per second Material TOO 10' Amber (fOSflil resin) ..3 24 5.00 210 115 95 85 80 75 5.43 2..25 220 5.' ....17 ::...9 52 5..2 115 5..25 10 6..39 2.. . . ....25 85 600 4..17 .5 2.52 6....4 320 2.. ..65 2.5 4.. ..M Fibergla.. 5.8 255 5.5 410 2..' ...9 3..24 105 240 .......0 223 3 X 10' tOtO . . " Carbon tetrachloride .. . 14..' .7 49 7.65 180 4.....0 17 5.38 48 22. 130 2... Clay soil (dry) .35 3.27 35 2.27 1 2..7 84 10' 10' 2. _.16 " lO'e:..... .64 190 2.7 34 8.57 470 390 280 170 24.17 470 2.35 113 2. Ethyl alcohol (absolute) .. 10'f.79 0 8 2..40 28 1.. .... . .1 .Frequency. lO'f~' 2.3 34 6.4 340 2.. Bakelite (no filler) .....2 . 130 4...5 165 4... pho8phate ..' .65 10' 10' a x 10' 2.... 25 10' 10' 2.a BK 174 (laminated) .7 10 " ... ~ 10"... 23.44 98 24. .94 0.25 80 5..9 330 2.. 2.... 4.....8 12.2 1100 2..
... .25 3. . 2.25 1600 1500 2.. " 10·~' 2...... 10".54 6600 . . 25 25 Loamy soil (dry) . 160 34 2... Neoprene compound ... . .. .58 175 . . gl!U1S) . .40 2050 5 2.22 360 2. . .25 4...53 460 2...24 2700 " 10'':'' " 10'':' .02 5 2.....47 95 95 5.40 2. . 4.47 2.60 840 725 3....42 120 6.60 760 2. .38 300 270 2...84 570 4 2.59 175 .. . ..88 560 3......66 165 81 Plexiglas . .' 4 2.' 3..4 3.. (38 per cent ON) 970 3...03 4.57 126 120 2.....75 2..68 2.48 360 2.lex 400 (micA..58 23 2... ..5 1050 2...83 1400 2... 2.29 3..02 6 .12 1450 5 2..02 6 2.47 865 7.......02 2..... .24 700 2..... 5 2...00 lOA 27 .99 1150 2.69 950 2.Gut.. 220 6..... ..25 3. ..45 2.39 . 104t~' .84 1250 7..60 150 2..25 2.61 2.. lOt.50 200 2.86 1600 2.12 235 10":' 25 24 25 25 25 3....44 2.40 2.percha . ..... ..75 300 2.33 860 4.. 10·':" .5 4050 3.10 620 2......16 660 2.60 10 13 3.. .70 1070 140 6.62 1600 2..95 885 5 2. ..... .. .. .00 1050 ~ Nylon 66 ..... .57 82 7... .....71 270 .....70 Paper {Royalgrcy) ..... .24 5 Paraffin 132 0 ASTM ..... . " 10":' ...25 2... " lOt.02 10 3..45 880 6...~5 ....60 730 3.47 145 2.... " 100 2..26 2400 3.44 27 2...06 2100 2...' .so 7.20 2000 7... 2..... 23 Myca......38 .63 360 7..76 385 6 2...02 1 2....77 4.55 s< 2.25 250 2.. 2..53 105 2..ta.30 3.75 105 6.54 750 . 2....45 2.60 260 7.00 1350 3..... Lucite RM1l9 .....
. .. ...08 800 5..78 7.53 145 .. ••• 77 2.." 250 2.......• 10''. ...5 10 1t 2.78 • 2..' 440 . Pyranol 1478. .4 2........• 10'.78 • • .58 215 2....78 28 32 3..59 3.54 160 2.' 43 2." 410 65 10''..47 2.50 1700 4.42 6700 3.. 25 25 25 26 25 25 25 25 " 10<..:' " 10'':' " 10''..... Porcelain.08 380 '. ruled .' 34 2. dry process .65 530 240 2...75 805 5....53 55 390 '90 '....50 2.87 530 630 5.80 8800 3..15 2.Frequency.:' lO'~' ..14 .. 3. 10 1. 90S..61 .. Sandy soil (dry) ......75 940 2.... pale crepe (Revea) ...36 750 '.' • • • 2. ...56 1..50 1200 5...25 3500 2.53 92 .. ..• .. ..53 9 5." 6.. ....4 43 2.56 1...3 6.23 550 4. .' 34 2.3 Polystyrene Cahoot 8tock) .... ..' .56 5 6..78 2.fi4.8 10' 2.....72 8. ." 8....56 10' 10' 2. 11 2.53 64 440 '. "'et process ....24 1100 10' 2... ax 9 ID a x 10' 2.....3 e.•.78 3.02 1. ...55 160 2.80 780 5.82 670 1i04 10' 2.55 3 5.. 3.80 770 2.5 2..4 67 3.04..74 740 .78 15 2... .98 1800 6.. 0 3..3 6...56 1.94 1450 2.78 23 S...51 850 Porcelain...' 120 2.....53 23 5.. cyclee per lCCond Material T'O 10' 10' 2.. G80 5.56 1....3 5.64 300 2. 2.•... ~ QuartJ..• .91 2300 2. .. Rcaio No...78 3..53 9 350 '... Rubber.
....69 3.' IO'....76 480 tLIO 3.10 4.. .. · .10 2 87. .5 2. 1. .... ..62 Styrofoam 103. .1 3 2..Shellac....' 2.58 6..1 7 2.. (3..•.r.6 600 46 38 3000 Wflter . .......:' .03 3.. Sulfur.1 4 2. 1.16 . ... 6... .. .. 2.. 10'.6 2500 76.. .. 2 3.. natural XL" .•.2 310 68...69 1I 2...08 8 2. ....... 86...69 8 2 3. .. . .25 68 63 55 3000 60 2200 . Tellon .. . 87 61 78 39 · .:' ..03 1 3300 1...16 8 2. ... .10 19 80.. 10 1... .16 ....10 2 2.3 2. . .. . .03 1I50 4.45 .6 280 77..80 2700 .1 1200 67....:' 10'... 10''.......1 7 2....6 2 2 87.... ... . 2. ..60 2100 1.. . 25 80 10'.2 .7 ...03 1 3.86 730 3. 1650 78.47 3....... .....03 1 3...1 3 1.16 2500 1.:' 3100 36 . ..1 11 3..... 1I00 4.10 7 9 .... . .5 2.. ... ....0 2 2......... .. 10 t:' . ..81 280 3. . 87 17 78..03 1 3...33 1700 1.. ...10 4 2... ...16 2....~' 3.•.. .... .03 930 3.5 .....65 4850 1.. .:' . ... 4.26 3.2 490 25 55 " 10 1 f .1 7 2.1 4 2... ... 2... .. .6 per cent wax) 28 70 .... ...10 ' .16 6. ........69 8 2...10 ..16 14 2..69 8 2. ·.50 6800 1..16 3...00 2200 1. ..03 2 3.... .1 4 2.....69 8 2. 25 25 22 lO'f~ lO'f~' I: Vaaelioe .. 78........ .86 200 6... ...... lublimed .. 2.0 .... .......18 22 2.. '" 34 7.66 82.... .2 16. . . .... ..
ten as • fCZ) = " where .. and Eq. . series converges to f(x) at ench point of continuity and to the midpoint of each discontinuity.APPENDIX C FOURIER SERIES AND INTEGRALS A periodic function f(x) with period a and aatisfying the Dirichlet conditions can be expanded in a Fourier series • f(x) . (el). Alternatively. outside the given interval. elb.. ~o + cos + b" sin (01) 2:. e:1l" x) e:1l" x)] where a" = b. However. say 0 < x < a. (Cl) with Eq.. the Fourier series can be writ.)~ dx (04) A comparison of Eq. ~ foG f(x) sin (2:'" x) d. Also.. In a given interval. but the series will not be unique until we specify the manner of extending the function beyond x = a. Moreover. a finite Fourier series (n ~ N) is a leastmeansquare error approximation to l(x).. fo" f(x)ei(2.jb. 1 a . (Q. the series does not equal f(x) . .. we can represent f(x) in the interval o < x < a in terms of a Fourier series of arbitrary period b ~ a....ei(Z. the fUDction caD be represented by Eq. In particularJ if we choose a period 2a and take 456 ... (e3) reveals that 2c.. L: C. 01a. . = ~ ~G f(x) cos J. '. as represented by Fig.. (03) the exponential farm of the Fourier series. Now consider a nonperiodic function. [a. (2:'" x) dx (02) Such a.. /G)~ (C3) e. but instead the series gives a periodic extension of l(x) . "....I) Equation (C1) is called the trigonometric form. as represented by Fig. = a.
 o (Ol • 2• .) where ~ = L:. = f(x) cos (':>) (n.. we have the Fourier cosine series f(. • A. sin B. a}o f(x) sin (n.nd (d) &.. (e) a Fourier cosine series. . cos where A. • B. peciod 2a and take the odd extension of f(x) from a to 2a.. (0) A function can be repreeented in the interval 0 < :I: < 0 hy (b) a "com· plete Fourier &eries.. 2• • Flo_ C1.. if we choose n. x) lb.) (nr ) a Z (C6) (09) dz . as shown in Fig. we have the Fourier sine series f(.. Fourier sine aeries. (C6) (C7) Similarly. (d) It ~_. Clc.FOURIER SERIES AND INTEGRALS 457 . Cld. .)  ~. the even extension of f(x) from a to 2a. + ~ foil L: . as shown in Fig. 1 [. a. .
" dw (010) (011) f(x). 1 (0 If(x)I'dx  la. bll . is aJ. At x = 0 and x = a.!(wW(w) dw f_. Similarly. the summation must be replnccd by an integration. (C8) converges to f(x) on the open interval o < x < a. l(w). (Cl) also exist. All crossproduct terms drop out because of orthogonality.. OId).I') 1 '\' • .21 r where lew)  ff _. which is the midpoint of the discontinuity in the extended function (see Fig. (C8) converges to zero. (013) This is readily proved by substituting for f(x) in tbe lefthand term from Eq. for the Fourier integral." dx The J(w) is called the Fourier tram/arm of f(x).. but we shall not consider them here. A useful relationship between the Fourier coefficients all.!(x)g'(x) dx . more generally. (la...OQ < X < 00.2r f. we have a Parseval theorem __ f__ If(x)I'dx . while Eq.2r f(015) . known as Parseval's theorem. and we have f(x) .458 TIMEHARMONIC ELECTROMAGNETIC FIELDS o ::5 x The representation of Eq. (CIO}J then converges to I(x) at all points of continuity and to the midpoint of points of discontinuity. The inversion {Eq.I' + Ib. and Eq. (C1O) is called the inverse trans/ormation. Fourier integrals corresponding to the trigoDometrio series of Eq. (C6) converges to f(x) on the closed interval ::5 a.I' + 2 L. Eq. 0" and the integral of If(xW over its period. In this case. 1 _.. Sufficient conditions on I(x) for lew) to exist arc f__lf(x)ldx <~ (C12) and f(x) satisfies the Dirichlet conditions. Equation (Cll) is called the direct transformation. (Cl) or (C3) and interchanging summation and integration. say . A function f(x) can also be represented as a superposition of sinusoidal functions in an infinite interval.II(w)I'dw 1 (014) or.
Schwart. _. Paris.·· dx = 1 _. j'TMorie deB dist. Such a picture gives an intuitive justification of Eq.. (C13) can also be given. J(w) J(w)e.' It is helpful to visualize the impulse function as <x<2 2 c c Ixl > 2 where c is an appropriately small number.ions. will be 1 L.·· dX] dw A similar generalization of Eq.. .z..!. f(x)u*(x) dx = = f. . It is evident that ~(x) is not a function in the usual sense. 1091 and 1122.ribut. [2~ f.tritlk. The inverse of Eq. 19501951.FOURIER SERIES AND INTEGRALS 459 The proof of Eq. u*(x)e. (016). the impulse function (delta function) is useful in Fourier analysis. Finally. it follows that a(w) c = 1.·· dX] U*(x) dx _. Hermano' & Cie. the transform of the 6 function contains all frequencies in equal amounts. but its use can be justified by rigorous means. (018) is 1 211" f· elv'dw = o(x) (C19) Our use of ~(x) which is a particularly simple and useful result. that is. From Eqs.cienti~uelJ et indu. By definition. the impulse function ~(x) satisfies the integral equation J: f(x) o(x . o(x)e.x') dx = ~(X') I a < x' < b x'<aorx'>b for all f(x). (011) and (016). (017).. f· [f· 2. (015) is summarized as follows f. primarily as shorthand notation for Eq.'· Actualitie' . nos.
for integral v "" n. (D4) where.!!. The J. we have (D3) and Eqs.(x) "'" Lt m!(m + v)! 2 ". these are two independent solutions to Bessel's equation. (D2) a.lorder.(x) co. N. In this case a second solution may be obtained by a limiting procedure.re no longer two independent solutions. _.J_. when v = n is an integer.0 2 (I)(x)'" (D2) where ml = rem + 1) in general. 460 .(x) . the result being J.APPENDIX D BESSEL FUNcrroNs Bessel's equation of order v is y x .blishes a second solution to Bessel's equation of integra. and the N.(x) ~. As long as v ia not nn integer.lim N.(x) are called Bessel functions of the second kind of order v.+.(x) are called Bessel functions of the first kind of order v. "" . It is conventional to define another solution to Bessel's equation as N •() ""' x J. (D5) This limit exists and esta. However.(x a ) ax ax + (x • '\' 2  v')y = 0 (DI) Solutions may be obtained by the method of Frobenius.(x) am V1r .
1 + M + M + . ..!.6 . (X)'+'ml (10 . Bes8eJ fUDctioDl of the firtt kind. )< ~ l'6 1/ ~ 1/ I\. (D2) and (D5). Figure Dl shows 1..781 ~(m) . (I)H (ml)' (z)'" ~(m) 2 (06) for the zeroorder functions... 0.. J.(z) ..(z) .o g .5772 (Euler's constant) .x 'II" 21 '" 2 ••• (z)'" 2 12: _l 1 ~ (1).. ~ m!(m + 10) I for n > 0.8 0.. x . 2 0. .I)!  (2)'x (07) 2 [~(m) + ~(m + 10)] (08) 'Y = 0. ~ ~ • . • .. 02 / 1/ / .m ..BESSEL FUNCTIONS 461 Of particular interest are integral orders of Bessel functions./ / 1\ 1\ 1'\ 1/ h I o 0. ) 4 V 1"17!'J f' 17 rs:.~ (ml)' 2 N .1. 'z 2 • (z)'2 +. log "2 J . 4 K \ \ 2 lX.0 0..(z) = ~ m!(m + 10)1 . m The Bessel functions have been tabulated over an extensive range of orders and arguments. one can determine From ~ (1)J. .\j.. J () N . and tables are available. Eq. and ~ (1)J. 01...(z) .6 o 8 16 Flo. +.() = .4 V\ i')/r. !10 12 14 "\ 1/ l\ I>< / K I/< 0. .. J. where log .
D2 shows those for the second kind..(x) . Bes8el functiona of the llCCond kind. '\j.(x) + /2 _.(x) H. J\ /' / 1\ o 0./ / / I 1/ X.(x) jN. [2 . For small arguments.Ol(X) ~ J.X ''j'\.(x) _ . N. (or v (D9) > 0.8 1. For the expression of wave phenomena.2.6 0.4 TIlLEHARMONIC ELECTROMAGNETIC FI. asymptotic series exist. J.(x) .jN. For large arguments. 0.V IV h ~ / '\.0 1. \ .462 0. t' " 0.(x) + _ (v . we have from the series Jo(x) __ 1 _0 2 'Y" No{:z:) . .... it is convenient to define linear combinations of the Bessel functions H.2 0. curves for the lowesk>rder functions of the first kind.(x) + (D12) called Hankel functions of the first and second kinds.4 1/ X 1\ )\ ..r log2 _0 and. v.Ol(X) . 02 / ~. ')< ~ '" V / / 1/ o 2 4 6 8 10 12 14 16 Fto. Smallargument . and Fig.6 0.in (x _! _ or) _. the leading terms of which are cos (x _! _ or) 4 2 N.Z 4 2 J.::t _0 ~ (~y 'Ir N.1)1 (~)' :z: (DlO) provided He (tI) > O.J.ELDS No ~ \ N.2 '" ~ / )< / '\.x (Dll) provided Iphase (x)1 < r..
and N.. V 2'11"'U e" K.. modi. 0. (J)o+'H. and H. in the special case v .(u) ~ /72 .) . from a knowledge of Bo(x) and 8 1(x). General formulas for I... Derivative formulas and recurrence formulas can be obtained by differentiation of Eqs._ 2 (D16) from which Wronskians for other pairs of solutions cnn be easily obtained. Figure 03 shows curves of the zero.(. n > I..(x) = B_ 1  8. the largeargument formulas become In (DIa) whicb place into evidence the wave character of the Hankel functions. particular._ 1  B. When x = ju is imaginary.. (02). The Wronskian of Bessel's equation is often encountered in problem solving. The largeargument formulas.. (D11) and (D12).(._ .(u) ~ j'J..fied Bessel functions of the first and second kind can be defined M I. Letting B.. (014) yields the recurrence formula B. . can be obtained from the' corresponding formulas for J.(u) .(ju) K."'( ju) (DIS) These are real functions for real u..B..I) which is useful for calculating 8 .. we have B. become B.(2)...(x) denote an arbitrary solution to Bessel's equation.(x) . and K.BESSEL FUNCI10NS 463 and largeargument fonnulas are obtained from those for J.) (DIS) The difference of Eqs. V2U T (D19) .(u)~ _. obtained from Eqa.(x) "'" 2(. 8+ 1 +! • ! B• (D14) B• • which. This is (D17) • 8.(x).and 6rstorder modified Bessel functions.. I.
• Note tbat which is of interest in radiation problems.\K. 03. is the spherical Bessel function oC the first kind.+~. it is conventional to define spherical Bessel functions as (D20) Modified The b.(x) .j·ID. • 2 • Bessel illustrate the evanescent character of the modified Bessel functions.(x) _ C.(x) cos (x _ C. Derivative fonnulas and recurrence formulas ca.(x) . functions./ " V <:/ 'I o Fro.(x) + jC. etc.(x) sin (x  $ .(x)jr'· :s:.j·ID.) (D22) n. The various fonnulas (or b.464 TUlEBARMONIC ELECTROMAGNETIC PlELDS . • • 2 / . \'''' I~ . are given the Damc and letter as the corresponding B.) _ (i)o(n D..."..D. is given the same name and symbol as the corresponding B+ K .i' are used in the solution of the Helmholtz equation in spherical coordinates. it is convenient to define the alternative spherical Bessel functions (O2i) where 8.. where C ( ) • x D. (D24) ...n be readily obtained {rom Eq. giving j .Hi become exact..) In ac electromagnetic field problems. (For example.+ H .(x) co. In scalarwave problems.jC.(x) sin (x  n.(x)  '\' 00 '< (2m) I(n + 2m)! 2m) !(2x)" (D23) '02:<'' 00 (2m + i) I(n (i)o(n + 2m + i)! 2m 1) 1(2:<)·. Bessel functions of order n + J.) + n.. j . or particular interest is the fact that asymptotic expansions for B.n(x) . (x _ n. h.. and B.) n. can be obtained from the corresponding formulas for B.(x)]"· n.ll) is the spherical Hankel function of the second kind.... (Di4) to (Dt6).(x) .
0 _ • + m)1 (I m)!  2 u)' (E5) si~.0.cosS m' ] y _ 0 (E2) in Eq. As long as v is not an integer.(V + 1) _ 1 u' (E3) When m =. If. In this case.( 1£) are two independent solutions to Legendre's equation [Eq.(u) and P. (m!)'(v .. ~ . P. P. Inparticu]ar.) d'y _ 2u dy du' du + [. s ~ 1r..u. so we shall be interested In the spherical coordinate system. (E5) becomes a liniu series called the Legtndre polynomial of degree n. 0 in solutions over the range 1 ~ u..n is ao inuge" Eq..forI1.v)!(". (E4)].) d'y _ 2u dy du' du + 0(0 + I)y ~ 0 (E4) We shall first consider solutions to this special case and later generalize to the associated Legendre equation. The result is (I ... the Legendre function of the first kind caD be expressed as s: N P ( ) _ " (1)'(v •u L. +1 • (m .(u) .(u) . S 1.1.APPENDIX E LEGENDRE FUNCTIONS The associated Legendre equation is Si~ O:O(siO O~~) + [V(V + I) u Si:: 0] y  0 (EI) This can be put into another common (orm by using the substitution =. (El). the associated Legendre equation reduces to the ordinary Legendre equation (I .u. . + v)! (I 2 u)' (m!)' where N is the nearest integer N ~ v. ~) .ul < 2.(I)·P.
(cos 8) .(u) ~.' 3) P.I)' 2"n! du" (EIO) Some of the lowerdegree polynomials are P.I P. whichever is an integer.(  moO l: (I)(2n .P.466 TIMEHARMONIC ELECTROMAONE'l'IC FIELDS and we no longer have two independent solutions.(u) [ }<i log . because these are the only solutions finite over the entire range 0 ~ 0 ::s:." I d' (u' . ~~ ¢(n)] (I)(n (m!)'(n + \' '< m_' • + m)! ¢(m) (I m)! .tion in Eq.1)/2.0 and 8 = r. in terms of 6..2M (E9) where M = n/2 or (11. (u ) When v = = !p. which caD be rearranged to M p u) .(u) .3u) + (EII) or.± 1. Po(cos 8) .(u) .}<i(35u' .2m)! 2'm!(n m)!(n 2m)! U .(u) ~ }<i(5u' .}<i(3u' . The Legendre functions of the second kind for integral II "" n are infinite at 8 . and sometimes more convenient.(U) 2 .(u) . or at u . Another solution. (E5) remains. sm '" ~ (E7) n is an integer. In this case. expression for the Legendre polynomials is given by Rodrigues' formula P (u) . only the first summa.(u) Hm Q. (ES) exists and defines a second solution to Legendre's equation. .u)2 (E13) .(u) .. called the Legendre function of the second kind.)i4(35 cos 48 20 cos 28 + (E12) + + 9) Figure E1 shows curves of the Legendre polynomials plotted against 8. 'If.u P.)i(5 cos 38 + 3 cos 8) P .(cos 8) .(U) cos".1) P.(cos 8) ~ ~(3 cos 28 I) P. The Legendre polynomials arc of particular interest. An alternative. is defined as Q.30.(u) . They can be expressed as Q.1 Pl(COS 6) = cos 6 P. . the limit Q.P.
'£ + (n .4 .0. Eq.£ t 3u . P.6 ~~ nO 1/7 \\ .\ 4 ~ 2 0..l t J dy/du" in the above equation.4 0.0... ..2 .u 2 l+u l+u (E14) or.0 0..0. Qo(C08 8) "'" log cot '2 ~(3 9 Q1(C08 8) "" cos 8 log cot '2 9  1 (E15) Q. times. in terms of 8. (D8). (E3) with y replaced by w.2u(m + 1).2 ~\3\ 1\ j / 1/ 'r..on the righthand side of these definitioD..LEGENDRE FUNCTIONS 467 1.(cos 6).U) . Hence.% cos 9 Figure B2 shows curves of these functions plotted against 8. o . Now consider the associated Legendre equation. If Eq.m)(n + m + I)] du" .1. .6 \ \I 1 \ 1\/ ' i\ \ 1\V 11\ 1 \ "J '.£ 2 11. ~ Some of the lowerorder functions 1.log .I log 1 + u _ 3u 4 1. (E4) is differentiated 111.8 . El. where are ~(m) is defined by Eq.0 1"" ~~ Fla. (E3). there results d"'Y [ (I .u').1 11.Q.0 du' du Letting w = (1 . For simplicity. Legendre functions of the first kind.(u) Qt(u) = log . we obtain Eq. V \ \ '\ 1\ _/2 V \/ / /\ /  .0.I) log cot 2 .'!. solutions to the associated Legendre equation are 1 I Smythe and others omit the factor (1).£ Q.8 0.(u) . we first take 111.(cos 9)  9 co.l. to be an integer.. 9 .
5u') P.. A standard formula. Some of the lowerordcr associated P.u')~ P...'(u) ...15(1 . / V .'(u) .. 2 N f".3(1 ..15(1 ...u')~ P..%:~u) Q.u'2) 1 3U'] u'p + 5u _ (l while the Q. r.. (u) are given by Eq..(u) ~ (1)(1 .Q. for Legendre functions of the first kind.'(u) .(u) are given by Eq. 1\ Q~(C08 f). V '/y i' I'.(u) _ (1)(1 _ u')" d. .'(u) .u')" d.'(u) . Some of the lowcrorder associated Legendre functions of the second kind aTC = while the P"O(u) Q11= (1U2)~(HI0g~+:+1 u u')~ (%U log 1 + u ')~ n Iog 1 + 'U u 1_ u') (E18) Q. (Ell)..u') (E17) p .u + 3u'.. 5 4 3 2 \\ \\ ~n_O 1 '\ !'1 .' _ (1 _ Q2' _ (1 _ 'U [u 1.J:~U) (E16) Note that aU p . t. (u) .%(1 .u')~ P." 1\' • \ 3 4 5 FIG.3(1 .'(u) . (E14).'(u) . 0 for m Legendre polynomials afC > n.468 TIMEHARMONIC ELECTROMAGNETIC FIELDS P.u')~(1 .. E2.u')u P..(1 . When m is not an integer.. the situation becomes even more complicated... Legendre functioIl9 of the second kind.
) . Finally.1 < 1.(u) (E25) which is a useful special case.. we have L.ny formulas for derivatives also exist.11.I)L:.. which arc Ma."'~1 "'" 0 (E23) for the range 111.v v + l l .... some of I L.. 2 sin (v + w)~ (E21) The solutions P .n .· + (m + n)(n  m + t)J. + (2n + l)uL. we have (m . In this latter case.~.(I ... 2 . ". is then sin"".. (u) denote an arbitrary solution to the associated Legendre equation.) and P. in m also exists and is L.. (~. some specializations of the argument will be of interest to us..ted Legendre function of the second kind is defined Q "(u) _ ! p.. + p"(u)(wI)I .I I + (n + m)L:.LEGENDRE FUNCTIONS 469 valid for 11 . '\' (a " + m)!(~ + m)! m!(~ + m)! _H (E20) the associa..• 'II" 'ul (u I)"" ( F u) ...utpi .ul < 2. Letting L.'(u) . .+1 + (1 2m:tp~ L"."(u) cos (0 + w)~ .m + I) L _. ....(m + n)L:."( u) . Perhaps the simplest way to calculate the Legendre functions is through the recurrcnce formulas.... (n .. ~ 0 (E22) A recurrence formula..."'(11.p.w .1)1 I)!(P 1)1 _0 L..'(u) = '1''u''2 InuL.] (E24) = ~ L _ + (n + m)(n .u')>>L."'( u) arc linearly independent.... (1 . the limit of Eq.m 1] u' (n + l)uL. 1 .(E19) "'·2 I' where F is the hypergeometric function F(a.1+ (a For real as 11. (E21) provides a second solution. except when u + w is an integer. 1 mu ut (l 1 L"'+t u2»)4i" If m = 0 in the last formula...~.L"''" + I)L:.
P nm(O) = (E26) { (_I)(n+m>l2 1 . 5 .m) +m  1) + meven n + m odd n + m even n + m odd n (E27) Some specializations involving derivatives are (E28) .470 TIMEHARMONIC ELECTROMAGNETIC FIELDS At 8 = 0. 3 . 5 0 2 . that is. at u = 1. that is.4 .3 (n (n .6 (n +m  1) (n . at u = 0.m) Qnm(O) = {~_1)(n+m+ll/£2'14·6 . the Qnm functions are infinite and m = 0 m>O At 8 = 1r/2.
10. 1950 (reprint). Son.. E. A. 12." MeGraw." Dover Publications. New York. New York.BWLlOGRAPHY A. Spence. W." McGrawHill Book Company. and R. Jeans. 0. Inc.. New York. S. 1949. F. Princeton.: "Electric and Magnetic Fields. 1931. Seely. Galbraith: "Fundamentals of Electrical Engineering. 1946. 1953. 9. Inc.: "Electricity and Magnetism... Inc.: "Fundamentals of Electromagnetic Waves." 3d ed. D. J. Becker: "The Classical Theory of Electricity. New York. Van Nostrand Company.. New York. New York." McGrawHill Book Company. 1954 (reprint). 15. lnlrodudory Book8 1... R.: "Introduction to Electromagnetic Engineering. John Wiley & Sons. New York." Dover Publications. Rogers.: "Electromagnetics. Clauical Books 1." Cambridge University Press. E. Inc. N. 1959. 1958. New York.: "Electricity and Magnetism. Skilling. Neal.: "Engineering Electromagnetics. 1958.. and N. A'I and R." John Wiley & Sons..: uAn Approach to Electrical Science. New York. Reading. R. H. P. 4" . Inc.. W. 4." AddisonWesley Publishing Company. 1933. New York. London.. Inc. Sears. Inc.:" A Treatise on Electricity and Magnetism.Hill Book Company.. H.. 8. Englewood Cliffs." McGrawHill Dook Company. 1950.: "Electrical Engineering Fundamentals. Page. Ltd. 11...J. 1954." MeGraw~HilI Dook Company.: "Electromagnetic Theory." The Ronald Press Company. F. 13. Heaviside. 1953. S. New York." Pitman Publishing Corporation. Inc. 1955. Kraus. 1948. C. Maxwell. 3. 7. New York.: "Electric and Magnetic Fields. W. 14.." PrenticeHall. Glasgow. Abraha. New York. 2.: "Introduction to Electric Fields. 1932. Attwood. 1949. Peck.: "Fundamentals of Electric Waves. N. Inc. L. New York. H. 16. J. J. P. C.. Shedd. H. 3. Inc. 1960. J. 6.: "Electromagnetic Fields.: "Introduction to Electromagnetic Fields... Hayt. G. Harrington. L." Blackie &. 4. Inc. 1958. 1955. D. Weber." 2d ed.." McGrawHill Book Company." D. Mass. Inc." McGrawHili Book Company.: "Elements of Electromagnetic Waves. New York.. Ware. E. 5.J. Inc. 2." McGrawHill Book Company. Adams: "Principles of Electricity. B. Booker. John Wiley & Sons.m.
1952. Books on Special Topics 1. Moreno. Dicke.. 10. Beam: /fTheory and Application of Microwaves. McGrawHill Book Company.: "Static and Dynamic Electricity.: fI Electromagnetic Radiation From Cylindrical Structures. A.: /fTopics in Electromagnetic Theory. J. R. J. 1943. 13.: "Waveguide Handbook'" MIT Radiation Laboratory Series.. Chicago. Kraus. New York. Wait.... N. Smythe. ft." Dover Publications. Aharoni. Oxford. vol.. New York. J." Pergamon Press. J. N..: "Scattering and Diffraction of Radio Waves. vol. 7. 1947. 1951." D.: "Electromagnetic Engineering. lnUr'mtdiate and Advanced Boo~ 1. 1959. 4. W." John Wiley & Sons. 2." PrenticeHall. R. Montgomery. J." Radio Research Laboratory.. New York. D. 11.. New York. M. Princeton. King. New York. 1950. New York. 7. Slater. Van Nostrand Company. J. W. R." McGrawHill Book Company. Whinnery: "Fields and Waves in Modern Radio. London. and E.. New York." Pergamon Press. 10. 1953.: "Antennas. New York. New York.ton.. C." McGrawHiU Book Company. 3. S. G. 6.J. Mentzer. Inc.: HMicrowave Transmission Design Data. Inc. H." Clarendon Press. N." McGrawHill Book Company. John Wiley & Sons." D. 2. Purcell (eds. A. Inc. Inc. 1950. M. J. Princeton.J. New York. Englewood CliiTs. Inc. 1950. 1951. Ramo.: IIAdvanced Theory of Waveguides. Inc." Univenity of Chicago Press. N. Watkins. Silver. Marcuvitz. New York. Inc.: "Electromagnetic Theory. (ed.: HAntennae. McGrawHill Book Company. Schclkunoff and Friis: l<Antennas. Inc. . 1950.. McGrawHill Book Company.472 TIMEHARMONIC ELECTROMAGNETIC FIELDS C. Mason. Van Nostrand Company.: "Microwave Electronics. Inc. Weaver: "The Electromagnetic Field. and R. 1947. C." 2d ed. Inc. 1950. 1941. 1958. 1946. McGrawHill Book Company. 1948.): "Very High Frequency Techniques." John Wiley & Sons. Inc. McGrawHill Book Company. L. 6. P. A. Scbelkunoff.. New York." Radiation Laboratory Series. 14. 1955. T. H. D. 12...: "Electromagnetic Waves and Radia...: uMicrowave Antenna Theory and Design. S. D." 2d ed.): "Principles of Microwave Circuits." Illiffc and Sons. R. 5. Inc. Strat. vol. Theory and Practice.J. 12. Inc. Jordan. and J. Reich. E. and W. New York. 1929.. 3... 5.. E.: "Electromagnetic Waves. 9. Inc. 8." MIT Radiation Laboratory Series.." McGrawHill Book Company. Bronwell. 1953. R. Inc. 4.ting Systems... S. Lewin. 1958 (reprint). 8. New York.
3 Circular cavity. 24 relative. 366371 in waveguides. 313(13) of ciroulargttides. I&H86. 66. 189(') of tranamission lines. 144 Chu. spherical. 5 ae. 69. 152.2326 Ampl!ro's law. 19). 173. 365367 Bailin. S. 205. 349. 258(23. 103 Bounds. 119. 307311 Apertures. 91(80) of rectangular guides. A. 385 Characteristic values. 335 Brewster angle. 139(18. 26 473 . 1 Addition theorems. 29 Circuit quantities. 436440 in conell. 86 of biconical guides. 65 of waveguides. 48. 255(6. 13. 11 admittance of. 270 Bibliography. 208. 34. 67. 255(9) of guides in general. 313(13) Bierens de Haan. G" 351 Babinet's principle. 351 Cavities (set Resonators) Characteristic impedance. 8) plUtially filled. 389 with bafBe. 83 maximum. 321. 201. 265. 463 Capacitivity. D. 174. 301303 transmission through. 73. 392 Admittivity. 42S431. 11. 348 Bessel functions. 204208. 257(20). 306 in plane conductors. 176 in wedges. 292 Admittance matrix. 366371. D. 55 Boundaryvalue problems. 6 Capacitor. 86. L.6 BiconicnJ wa. 59 Browo.. 154. 30 Carter.veguide. 28428. V. 213216 plUtially filled. 46. 441(6) intrinsic. 306 Belk. 232. 250254 Associated Legendre functions. 48 of parallel~plate guides. 428431 in cavities. R. 220. 90(24) Bessel functions. 13. 8185 Antenna gain. P. 208. 231 Circuit elements. 4~64 modified. L. 331 Circulating waves. 259(26) Circular polarization..444(2123) in spheres. J" 278 Churchill. 265. 256(10) Circulator. upper and lower. 62. 268. 464 zeros of. 199203.. 471472 BiconicnJ cavity. 303. L' J 249. 10.. 4 Antenna concepte. 376(19). 88(8) Circular waveguides.INDEX Boldface numbers in parenthcsC8 refer to problems Ae phenomena. 346. 138(17). 281283. Z') with wedge. 194 Boundary conditions. 468470 Attenuation constant.
170. 150. III spherical. 34). 179. 179.219 Coaxial line. 6 Diaphragms. 431434 conduction. 7881. 1. 150. 100. 227. 177179. 459 Depth of penetration. 262(3941) displacement. 263(42. 390 Delta function. 24 . B. 118 near spheres. 434436.. 7. 68. 95. 279281 Conjugate problems. 195(~. 442(1316) Dicke. 440(2) near wedges. 381391 Cylindrical waves. 169. 365Complex quantities. 166. H.384 Cutoff wavelength. 444(20) opening onto plane conductor. 445(2427) with waveguide.. 2 Coated conductor. 6. 2 Closed surface. 131 Complementary structures. of charge. 430 Cohn.194(31.219 Critical angle. 6. 451455 power. 34 Conical cavity. 430 Current. with cavity. 136(7). 97. 260(31) sheets. 4 of complex power. 292297 apertures in. 232238 and current elements. 254(5) Cohen. 321(6. 85 Degenerate modes.27. 34 in cavities. 106. 400 Dielectric. 6 perfect. 27 Conductivity. 34 source. 263(") Cutler. 24. 279281 Conducting cylinder. 188. 136(1214) reactive. 315(28) near planes. 10. 315(26) Conduction current. 198. 68. 425428. 301303 and current element. 261 (34. T. 262(3943) Conducting sphere.32). R. 316(30) 88 waveguide. 384 Cylinder of currents. 171 Cutoff frequency. 103. 28 ribbon of. 206. C. 206. 53 Diamagnetism. 18. 6 complex. 18 Conductors. 20 Concentric spheres. 316(29) with dielectric coating. 93(41. 451455 Dielectric cylinders. 287 filament (set Filament of current) near halfplane. 11 Constitutive relationships. 316(30) near cylinders. C. 48.S complex. 43) impressed. 447 Cylindrical waveguides. 18 pcrmittivities. S. 33 in waveguides. 316(29) surface. 1923 Poynting vector.42).. 27). 364. 27 elements. 6. 7. 380(43) Complementary solutions. 21 of energy. 36) DieJeetric loss angle. 314(18) Conical waveguide. 303.134(14). 7. 2. 150.425428. 284. 27 loops. 168. 298. 220. 64 Conservation. 223 radial. 260(30) Cylindrical coordinates. M.. 27. 306 current element on. 281 waveguide modes of. 65 junction. 13 constitutive relationships. 315(26.474 TIUEBARMONIC ELECTROMAGNETIC FIELDS Closed contour. 315(26) and current loop. 414420. 18 Corrugated conductor. 279 speratures in. 193(26) circular. 392. H" 362. 298301. 6) Conducting cone.. 60 Crowley. 328 Complementary antennas. 27 near cones.
383 Electric quantities. 200. 141(28) in two dimensions. 80 Field quantities. 257(21) Dielectric slab guide. 81. 116. 9. 85 Equivalence principle. 105 ncar ground plane. 386 Erdclyi. I. 1. 62 of waveguides. 1 flux. 219 Dielectric spheres. 309 Gauss' law. 357 Echo width. 458459 Fourier·Legendre series. 10. 241242 near wedges. charge.309 of dipole near ground plane. 396 Fourier series. 15 scalar potential. 3. 123125 Greeo's identities.. 360 Dipole. 259(26. 42 Equation of continuity. 3 Filament of current. 456458 Fourier transforms. 272 Energy. 363 Echo area. 362365 Dielectric rod guide. 358. 98100 Equivalent circuit.INDEX 475 Dielectric obstacles. 75 Dominantmode source. 83 antenna. 39. 221. maximum. of obstacles in waveguides. 437 of spherical w&ves. 223. 192(22). 425 of microwave networks. 243 near cylinder. 279 of transmission lines. 402 Duality. 297 Differential scattering.. A.. 77 vector potential.384 Eigenvalues. 225 Directional coupler. 99. 27 Dissipated power. 129 Elementary wave functions... 78 Faraday's law. 104.. 34. 128. 432 Field coordinates. 180. 68 Echo. 7. 8185 in conducting wedge. 50. 435. 145. 3 current. 135(3) Displacement current. 163168. H. 4 Good dielectric. 163 Free space. 84 supergain. 120123 tensor. 104 magnetic. 15 Eigeniunctions. 7. 11 Dissipative current. 337. 106110 Equivalent circuit. 236238 near haU·plane. 78 antenna. 144. 144. 4 Ferromagnctism. 46 Emde. 6. N. 364 Effective value. 11 velocity of. 5 Fundamental units. 26 Euler's identity. 137(1214) of dipoles. 307311 normal. 15. Gain.27 dipole. F. 27 Dominant mode. 147 Evanescent mode. 67. 245 Ether. 120 modified vector analogue. 23 conservation of. 24 Goubau. 2 Equiphase surfaces. 238242 Foster's reactance theorem. 62 of coaxtowaveguide feeds. 144. 25 Feshbach. G. 21. C. 1. 15 Evanescent field. 355. 401 . 359. 275 Frank. 222 Grecn's functions. M. 402 of resonant cavities. 389 266 Elliptical polarization. 10. 3 intensity. H. 69. 27) twodimensional. 223 Gray.
443(17. 26 Isotropic matter. 42). 42lH25. 113 Index of refraction.34).. 384 Hall effect. 152. 444(20) waveguide. 7.. 195(33. 144.. 199 Instantaneous quantity. 84 Instantaneous phase. 39. 94(0U46). 93(41. 30 of dielectric. 317 Intrinsic parameters. 15 Impedance. 179. 1 Magnetic quantities. 446(2729) near conducting cone.309 Helmholtz equation.18) HOgflD. 86.ONIC ELECTROMAGNETIC FIELDS Green's identities. 58 Induced emf metbod. 82. of capacitor. 154. 53. L. 193(2729). 386 Levine. 199203. 119. 125128. 39. 272 Junctions.. 37 Harrington. R.. 15 Inaulatof'li.. 468470 LePage. 69.. Y. 385 of circuit elements. H. 315(28. 332 Jahnke. 284 Hildebrand. 39. 266. R. 198 in rectangular coordinates. B.476 TIMEHARM. C. 170 Lorentz reciprocity tbeorem.t 26 Homogeneous matter. 445(2427) coax·towaveguide. 24 of inductor. 25 Lossy dielectric. 349355 Linear matter. 24 Maeroscopic standpoint. 68.100 in cavity. 428431 characteristic. 113115 Inductivity. 45 Loop of current. 85 Legendre functions. 179.. vector analogue. 352. 425428. 5. E. 351 Kirchhoff's laws. 6. F. 69. 7. 84 intrinsic. 48. F. 77 in spherical coordinates. 25 Inductor. 462 spherical. 354. 62. 35(2) Hankel functions. 19. 265. 79 Induction theorem. 459 Incident field. 37 Hu. 18 Linear polarization. 32 magnetic. 39. 265. 6 Integral equations. 464 Harmonic functions. 111 Loss angle. 38 in cylindrical coordinates. 68. 8185. 31 Input impedance. 371(2). 113 Levis. 65. 12 1m operator. 349 Induction field. 27 . 352 matrix. 436440 waveguidctowaveguide. 87 of linear antenna. 445(24). 375(18) wave.toeavity. 158 King. 29) Loosely bound wave. 398 surface. 4. 40. 465470 associated. 392. 450 Hemispherical cavity. conductor. coaxtoeavity. 128. 27 Impulse function.. 93(42) input. 48. 434436. 55. 121 Guide phase velocity. 303306 near conducting sphere. 430 Linear antenna. 87 Isolator. 154. 152 Impedivity. 29 of current loop. 13. 264 Helmholtz identity. Y. 6. 94(4446). of apertures. 358 Hybrid modes. 385 Guide wavelength. R. 34. 2325 Impressed current. 172177. W. C. 34 current. A.
463 Monopole antenna. 15 1068 angle. 432 Multipoles. 209. 5. 400 Morse.383 Oliner. 13 Papas. 25 vector potential. 63. 381.. 129 Magnetomotive foree. 182. 75. 8). 44 Perfect conductor. 2. 40. 431434 Mode. L. 374375 rectangular. 189(6. 158163. 16 Obstacles in waveguides. 190(12. 85 Marcuvits. 100 flux. 169 for free space. 314(19) Neumann's number. 1 Modal expansions. 392. 410. 215 for circular guide. 122. 413 Orthogonality. dipole. 376377 Partially 6lIed waveguides. 371. 85 Phase velocity. 85 Phase constant. 72. 371377 Phase. 383. 39l402 Mksc unit8. 119 Maxwell's equations. 402418 Ohm's law. 6. 432 Notation. 48. 73. circular. 259(29) spherical. 191(1618). 345. mode. 119 stl1nding. A.424 Matrix impedance. 34 Perfect dielectric. 138(15) Montgomery. 85. 277 in general. cylindrical. 18 Mentzer. 158 Plano waves.348 variational formulas for. 157 for rectangular guide. A. 13). 273. I.. 181186 p. 3 intensity. 420. circular. 143. 371373 rectangular. 325 spherical.. 145148 . 383 Modified Bessel functioJ1!.306 Microwave networks. R. 18 Perturbational methoda. 28&289. 127 Pincherle. 331345. 131 Pattern. 309 Normalization. 18 Permittivity.. 3 Magnitude. 220. 321326. 76. 418.ti. 5. 272 Mode voltage. 39. C. 90(28). 151. 72. 326331. 70 radiation field. 68.INDEX 477 Magnetic quantities.. 387 for rectangular cavity. 326 variational formulas (or. 440(1) and coaxial feed. 86. 83 receiving. 168 for spherical. P. 201. 15 Physical optics method.. 172 Nonpropagating mode.321. 70 for circular cavity. 458 Partfally filled cavities. 337. 256(13) Paramagnetism. 377378 Particular solution. 24 Permeability. 257(18) md. 191(1~16).ally. 389. 313(7. J.331 perturbational formulas for. H. 317331. 171177. cavity. 390.257(20). M. 70. 378(3637) opening onto ground plane.lIy filled. 68. G. 258(23. 68 Normal gain. I.wave. 155 for slab waveguide. 2fYl for coated conductor. 432 Mode patterns. 411. 6 Parseval's theorem. 75 Mode current. 2') perturbational formulas for. 385 Phasor. 345348. N. 77. 113 Parallel·plate waveguide.. 69.389391 in cavitics. C. 39. 226.. 383 ~{odefunction. 7).
. 7) of circular cavit. 76.325. 298. 446(26) in waveguides. 285 of cavities in general. 148155. 140(20. 232236. V. 118. 118. 378(32) Purcell. 314(20) Quality factor. 45. 444(21). 190(10) of spherical cavity.y. 339 Re operator. 44 rectangular cavity. 279 Quasistntic. 309 Rayleigh scattering. 79. 128. 420 Radar cross section. 112 Radiation field.. 116120 (or antennas. 132134 R&diation resistance. 316(31) of rectangular cavity. 93(42. 158163. 4()() Receiving pattern. 4. 442(1~) Potentials. 88(8) Polarizing angle.6. E. 335 Saunders.373 Rectangular waveguide. 372(3. 94(46) Ramo. 20 Probes. 19.) Scattering. 340 Rcacti ve currcn4 27. 371 Rumsey. 355361 by cylinders. 116 Radar echo. 288.ability conditions.. 202. of biconical cavity. 3 Re8ection of waves. 466 Rubenstein. 338. 391 Posts in waveguides. 119 Reciprocity. 120 for circuits. J. 99. 6674. 129 Power. 261 (34. 141(24. 355 Radial waveguides. IS. in cavities.~78 TIMEHARMONIC ELEcrROMAONETIC FIELDS Polarization. 365 Saddle point. 113 reciprocity for. 82. M. 22 Realiz. 28 of hemispherical cavity.387 partially filled.6) Resonant slots. 68. circular cavity. 272. 86.. 257(16) dielectric. 55. W. 2 Ritz procedure. 115. 374 Reference conventions. 431434 spherical cavity. 68 Propagation constant. 191(1618). 10. 115. 29 minimum antenna Q. 340. 400 Q (au Quality factor) Quadrupole. 213216 concentric sphercs. " 6. 226. 242245 twodimensional. 44). 348. H. 295 RayleighRitz procedure. 279283 ~ation. 327. 155 sources in. 79. 421 Resonance. 36). 216. definitions. 191(1416). 7781. 208213. 269272 spherical secklr. 39. 94(4. 406411. 81. 74. 312(4) of spherical waves. 364 . 188. 392 Rectangular cavity. 77. 5461 Reflection coefficient. 28 Reactive power. 16 Reaction. 178. 9. by conducting plate. 119 for microwave networks. 434436. 192(19). 59 Porta. 346348. 7476. 285 of lossfree antenna. 425428 Propagating mode. 260(31) Ridge waveguide.. 374(12) Rightrhand rule.21) by conducklrs. 419. 155157 partially filled. 445(23) Resonaklrs. 259(28) 8pherical. S. 344 Rodrigucs' formula. 216219. 74 Resonant antennas. of matter. 27 of waves. 310.22 Poynting vector. 284 Ribbon of current. 309 magnetic.. 245 Scattered field. 284 onedimcnsional. K. 228230 Radiation condoctance. P. 62. cylindrical. 384 stationary formulas for.
. by dielectrics. W. N. 58 Source coordinatcs. 303. 238 Slotted sphere. 18). 202. 170 Total reflection. 268. 268. 85 Surface currents.219 Surface impedance.430. T. 59 Transmission. 313(7.. 79. 332 Twin51ot line.444(21. 324. 399 Transverse 6eld vector. TM. 459 Secondary units. 6. 63. S.375(18) 381 Silver. 380(22) by ribbons. 67. 4247.> . 265. 286 Schwa. 8). 135(.63 parallel·plate. L. 45. 88) by magnetic obstacles.261(32).. 261(37. J. 198. 202 Traveling waves. 434 Tensor Green's functions. 357. 420. 39 Trial field. 55. J. 85. 440(1) radial. 445(23) Slotted cone. 399 SehelkunofI. 53 Slot in ground plane. 121. 434440 for impedance. 350.. 181186. TEM. 7. 138(17. 382 Transvcrse fields. 355365 for transmission. 267. R. 317. 53. H. 130.. 313(13) equivalent. 286289 Standing wave. 12. 222. 419.96 Spherical Bessel functions. 345348 Storer. R.. 130. 430 Teichmann. 464 Spherical cavity. 284 Seidel. 326 Spherical coordinates. 69 Standingwave pattcrn. 7).370. 354 Stratton. A. 265.19. 18 Singular 6eld. S. A.. 467 Sneddon. 90(28). 276. 309 Surface of constant phase. 302 Smythe. S. 33 Surface guided waves. 238242 by wires. J.. 365371 for waveguide feeds. 379(3941) Scattering matrix. 360 by haIrplanes. 428431 for cavities. 1. 382 Tector.371(2). 368 Transmission coefficient.95. 32 Skin depth. 123125. 292298 stationary formulas for. 211 twinslot.1 for aperture admitlance. 269273 partiaUy611ed. 55 Static mode. 44 Standing·wave ratio. 1 Secklral horn.. 368 Transmission lincs.356 Tesseral harmonics.. 264. 355365 by wedges. 34.22).. 213 8«>ly. 447 Spherical waves. 306 Simple matter. 168171. 80 Sourcefree regions. 143. 362365 differential. T. 348355 for obstacles in waveguides. 284286. 306 Slotted cylinder. 360 Transmission area. 245. 189(6.. 386 Segmental cavity. 425428 for waveguide junctions. 67. 378(88) by spheres. 386 modC8.INDEX 479 Scattering. 338 Tai. C. 01 (31). 402406 for scattering. 63. 324 Supcrgnin antennas. 135(7) wedge. E. 212 Transmission matrix. 6166 biconical. liS Separation of variables. 358 TE. 241242. 222 SeIr·reaction. 273 Tightly bound wave. Stationary formulaa.rtz. 420425 for waveguides. 37 Sources. 331345 for cavity feeds. 252 Snell's l3W.
381391 psrallelplate (see Parallelplate waveguide) posts in. 39. 5154 Wedge cavity. 179. 100103 Units. 332 Variational methods. 68. E. 86 characteristic. 270 Zonal harmonics. 192(22) in general. 223 dielectric slab. 163. 3). 255(7). 143145 spherical. 279 partially filled. 216 rectangular (see Rectangular waveguide) Wavelength. 384 intrinsic. 289292 Waveguide feeds. 96. 118 Von Hipple. 375(18) Wave equation. 170. 85 Uniqueness. 152 Wave number. 208. of energy. 3. 313(13) circular (see Circular waveguides) corrugsted conduclor.. 121. R. 178. 205 or spherical Bessel functions. 129 Wave transformations. 406411. 99 corrugated wire. 85 cylindrical. 256(14) Whinnery. 77. 150. 425428. 86. 331380 Vector analysis. 273 . 447450 Vector Green's theorems. 88. 147 Uniform waves. A" 23 Wait. 420425. 400. 77. or Bessel functions. 385 Voltage. 309 Wave impeda.nce. 88(2) Wave functions. R.. 42 of light.34). 39. 444(20) Wigner. in dielectrics. 18) Waveguides. 446(26) radial. 230232. 40. E. J. J. 414 Zeros. 371(2. 199204 plane. 37 for inhomogeneous matter. 40 cutoff. 195(33. 69. 240. 193(25) Variation. 443(17.384 Velocity.425428.206. P" 434 Windows. 317. 284 waveguide. 8587 in lossy matter. 172177. 55. 264269 guide. 40 Waves. M. 1 Van Valkenburg. 208.480 TIMEHARMONIC ELECTROMAGNETIC FIELDS Uniform plane wave. 141(28) Veclor potential. 88. 284286. 4148 in general. 5 of phase. 15 Voltage source. 397. 39. 442(12) probes in. 242 Wall impedance. 435 Waveguide junctions. 37 Wave potentials. 66 biconical. 193 (2729).
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.