You are on page 1of 736


Maxweil:s equations:

V·D=p \l·H=O

v x E ='II.H - in \1 x If = jr..,'f E + J

Surface resistance and skill depth :

b~ =




lnput impedance uf lermiw:lled lossless transmission lilies:

z. _ Z Zl. + jZo tanfJC

In - 0 Z L Z ·::.Ie

'o-r} L tan .I'

Z;n = .i Zo tan pl

Zin = -J 20 COL p.t

(arbitrary load)

(short-circuited line) (open-circuited line)

Relations between 10m! impedance OJU/ refiection coefficienr:

L'= Zl.-ZO ZL. + Zo

DefinitirJIPs of return loss. insertion loss and SW/?:

lL = -20 log ITI·

RL = -201ogln·

Conversion /u!:/wcen dB and nepers:

I neper = 8.686 riB Elements of the ferrite permeability tensor:

( .v(IW:m)

/1, = lIn I +., ?

""0 - W'


1i=1.l.1,!"J ?

..... 0- ,J)-

Conversion between some values of rejtection coefficient, SWR. and return 108s:

(or 2.8 l\;fHz/Oer -ted)

WI 0.024 0.032 0.048 0.050 0.056 0.10 0.178 0.200 0.316 0.33

SWR 1.05 1.07 1.10 1.1 J 1.12 1.22 1.43 1.50 1.92 2.00

RL (dB) 32.3 30.0 16.4 26.0 25.0 2D.() 15.0 14.0 10.0 9.6

.~ = I C-;U

[J=L LJ = I

The ABeD Parameters of Some Useful Two-Port Circuits,

c 0 0 ,Il,fj cv P ararneie r,

: (] 0
.\ = I B=(1
\=y 0=1
~ 4" ,'U' ~I

C = 11"1 'In 111

II ~ IZo'5ll1l3l I) =- ,0' 131

4=N C=!I

8;0 IJ=j_


fl = J_


D=-I-_!j_ I;

.\= I -I ~ ,(3

( = .L l,

Microwave Engineering



Microwave Engineering

Second Edition

David M. Pozar

University of Massachusetts at Amherst

New York • Chichester • Weinheim Brisbane • Singapore • Toronto

Librar; ,if ['IJIIRrFS.~ Cwulo[!.illg iii Pubhvuuo» /)"/(1 Puzar, l)avill i\t

Microwave englnecring I Om IU T\'I, P'lI..llr, - ,~'ltJ ed

p ern,

ISBN n-47"1 -17090-8 Cc!u\h , alk, paper)

! Microwaves. -, 11.111.:['0" ave devieex, 3, I"'Jj"mwnw circuks. '- lItle.

TK7S76, P6<) 1'198

r,:!! JS ]'J-·dC'1U

EXECl1TIVE: \:'.l)!T()R Clnffily R"h~> EDITORiAL ';'SSISTANT Su,aml': Dwyer MARI",:ETI.\l(; MA:'iAGER Harper MOl'~

'iEj\'lOR PRODtl(T10N l'-L\..'\AOER LucIIJ e BUonocore SEmD.R PRODUCTION EOnnR Moniuue (:'1k-110 rOVER DESLG:>;'ER D[\\tiJ levy

lLLLTSTR.\TIOi\ STUDIOS Wcllrngron '" Vanl~"\gr An " [ l'STRATInt-: CUORDl.\fATOR Gem: ,\]eJlII vli'l..;\/lJF/\.CrURl'IG MA~,'\GER \1()J1iyuc Caleflo

Ihrs bnok wn~ sct rn I nil.:! Timev Ruman by ETP HARRISO.\l and prrrued and hound D_I RR. D,Jrllleliev & Son, Cumpuny, Cf!lwf,'rrlsdUe, 11"" cover ~\a., prinlC'o ny Tue Lemgl) 1'1'1"'" Inc

R~c<lgnlzin~ the imporlunce of pre-erving what hll, been written. II IS it

f\\\\\q Ilf l""11 Wiley & Sm",. \!1~, l\:l h~w b\l\>\(,. III el'll.lIlTl1lgc ",\\Il\' j1llblh\\~tl in tile, LT ~i!eJ Sl!1 IE' prillte..! on acid- free paper, and We. exert our be"

"from w Ihell end.

('he paper on thi, h"Oh II,,, manufactured b1 ,~, mill \\'hll~~ fl.lre,~1 management prngrnmv lnchuh,u:.'lllincd Y'E!U harvc-ring of jte, I irnbc r lands. Sustained Yield harvesting pnlll'lple- ensure IhOlI the number 01 [rec., CUi each ~'ear 0(lc" JIll! exceed the <1nJlIun! uf 11 e " growth

Cnpyrigh! @ 1'1%. by John Wife! & .s,m.,. Inc

."J.\ ml~TS fC!i~IYe,L Pllbli~lJed ~lnltlllaneol!;1) ln C anada,

Reprnducnon or uunsluuon (If an,\' IJ<Ht "r

l111' "I'r" beyond Ihill perrniued b.' Sections ii)7und LIIS "r ihe 1')76 United States COp)r1ghi Act" ilhOlH 1!1~ pernussion or the copynghr owner i' unlawful Kequ~'ih I'or pL'nni,,,icHl

OJ. mnher rnformaunn \huu\,\ be "dtln:c""tl to

the Permis-iuns Department. J(\JJn Wiley & Spn~, UK'.

Pri.ll!<,d in the United SUUt!5 of America

14 1] 12 11


Because education should be the accumulation of understanding. not just an accumulation of facts, I have tried to write a textbook that emphasizes the (undamentat concept of Maxwell's equations, wave propagation. network analysis. and design principles as applied tn modem microwave engineering. Allhough I have avoided the handbook uppreach, 'in which a large number of results are presented with little or no explanation or context, a considerable amount of material in this book is related to the design of specific microwave circuits and components. for both practical and motivational value. T have tried to present the analysis and logic behind these designs so lim! the reader can see and understand the process of applying fundamental concepts to arrive at useful results. The engineer who bas a firm grasp of the basic concepts and principles of microwave engineering. and has seen ho\\ these can be applied toward a specific design objet> live, is tile engineer who is most likely to be rewarded wili1 a creative and productive career.

Modem microwave engineering involve-, prcdorninanrly circuit analysis and design. i11 contrast to the lield theory orientation ofa generation ago. This ma~ come as 11 surprise. possibly mixed with regret, to those of us teachers for whom microwave engineering meant applying sophi ticated mathematics to the solution of Maxwells equations for waveguiu; components. But the majority 01 practicing microwave engineers now design planar components and monolithic integrated circuits With no direct recourse to field theory analysis. Microwave computer-aided design (C.t\D) software and the network analy-zer ure the essential tools (If today'. microwaveengineer. Microwave engineering education must respond 10 this shif] in emphasis to network analy: is, planar circuits and components, and active circuit de ign. Microwave engineering will always involve electromagnetics (many of the more sophisticated microwave CAD packages implement ri gorous Ii eld theory sol UI ion s). tim] students wi II 51 ill henefi r from ex pas u re tu subjects such as waveguide modes and coupling through apertures. hut the: change in emphasis to microwave circuit analy IS and design is dear.

There are other changes as well. Several generations of sruderus were taught that open-circuit STUhs should not be used for tuning because of spurious radiation. but this is not a proh lem with miniature microwave cireui try. in which ope n-circuit lines are. of tell used fOT matching in transistor circuits. Technology advancements ill planar traosrnission Jines, bipolar and field effect transistors, dielectric resonators, low-noise amplifiers. tran-

sister oscillators. PIN diode control circuits. and monolithic integrated circuits similarly v



require reconsideration of many of the assumption." underlying the trudltional study of the subject. Applications of microwave engineering: arc also changing, with im .. Teasing emphasis on commercial use of microwave technology for per onal communications sy~terns. wireless local area networks, millimeter wave collision avoidance vehicle radars, radio frequency (RF) identificetion lagging, direct broadcast satellite television, and muny other systems related [0 the information infrastructure. These developrnents suggest that there will be DO shortage of challenging problems in RF and microwave engineering and a clear need for engineers having creati vity and an understanding or the FUlldamel1taL<. (1]' microwave engineering.

The success of the lirst edition of Microwave Engilleenll,1!. has been gratlfyrng. Feedback from students and teachers dearly supported coverage of the i1nil.lysi~ and design of impedance matching networks, resonators, filters. coupler". ampiillcrs, and oxcillarors but also supporting topics such as transmission line theory. microwave network. i.ll1.lIy .. h. and aperture coupling of waveguide . 111e most consi: tent call wu" for more coverage of active circuit design. Because of iii constraint On the length of this edition. new material could not be added without deletion of some existing material-a task that is more difficult than might be imagined. To this end, and with the ubjecrlve (If 11m eliminating critical fund amenia I material. I have limited major deletions from the first edition to the topics of plane wave propagation III anisotropic media, the tra nsrnission line analogy fur plane wave reflection. and transients on transmission lines, Giber reductions include topics that. could be useful but are nOL ncccssruy for the study or the rest or the hook. such as Mason's rule, Fosters reactance theorem. and parallel connected networks. A few other sections in the beginning of the book were combined 01' otherwise consolidated. Several examples were eliminated. but most of these have been com cried In problems in an attempt to retain at least orne mention of the topic. that were cut, Although most teachers appreciated it. students seldom read the historical materia) in Chapter 1 of the first edition, so thls chapter was eliminated and the material largely moved to the beginning of the relevant chapters and sections throughout the book. We have also expanded the coverage of acti ve circuit design from one to LWO chapters in this edition. Chapter I () covers noise, detectors. mixers, PIN diode circui ts. microwave uuegraied circuits, and n review of sources. Chapter I I is devoted to transistor amplifier and osci llator design, with new materia] on the phy: ical con rruotiou 01' transistor: and their small-signal equiv .. alent circuit model , balanced amplifiers. distributed amplifiers. and dielectric resonator oscillators. We have also added a discussion of the Thru-Rcllcct-Line (TRLJ method of network analyzer calibration. Other changes include revised analyses of the Wilkinson POWET divider and multisectlon quarter wave uunsformers, the addition of titles to the examples, new or modified e.xample.;; ~md pr~)h\ems, ami the CCrrenll.)Il oC ~e'Je.ra\ typographical errors ..

This text was written for use in a two-semester course in microwave engineering. fur senior or first-year graduate tudems, If students have had a good course. in undcrgrad .. uate elecrromagnetics. the material In Chapters 1 and '2 can be rev iewed (air!) quickly.

tudents with less background should study this material in more detail, The chapters, are organized in the sequence in which our course at the University of Massachusetts at Amherst is taught, but it 1S certainly possible to pick and choose from the text to suit the needs of a one-semester course or an advanced l,;OUTSC in active circuit design or to cover toplcs in a different order.


Two irnportant things That will be included in a successful course on rnicrowave engineering are a microwaveluhoratory experience and rhe use of computer-aided design (CAD) software for microwave circuit analysis, A hands-on laboratory is expensive to equip but provides the best way for students to develop an intuition and physical feeling for microwave phenomenon. A laboratorv with the first semester of the course should cover the measurement of microwave power. frequency. standing wave ratio (S"WR) .. impedance, and S-paramele.r;,. as well as the characterization of basic microwave components such as tuners .. couplers, resonamrs, loads, circultuors, and tilters. Important practical knowledge about connectors. waveguides, and microwave test equipment will also be acquired in this way. If avai lable, a more advanced laboratory session can consider topics such as noise figure,. amplifier characterization, intermodulatltm products. and microwave mixers. Naturally, the type of experiments that can be offered i~ heavily dependent on the test equipment thai is available.

There are several commercially available CAD packages for microwave circuit analysis, for both personal computers and mainframe computers, Providing students with access to such software allows them ro verify the results or the design-onemed problems in the text, gi ving immediate feedback that bui Ids confidence anJ makes the effort more rewarding. Because the drudgery of repetitive calculation is eliminated. students can easily try alternative approaches and explore problems in II more detailed. way. The effect af losses on the re5pun.~e of a filter. [or example, would he pmclicaliy impossible to evaluate by hand calculation, but it i~ eusy to do on the computer. And because c.w software is used extensively in [he microwave industry. classroom experience with such tools will be useful upon graduation.


I would like 10 thank many people for their hdp in completing this book, but my foremost appreciation goes (0 rne many stuce n f~ who 11<1 ve llscd the first edi don of Mit'1'OWCII'f! Engineering, I also would like to Thank my colleagues in microwave engineering at the University of Massachusens for their support and collegiality through the years. In particular. Bob Mcintosh Mel Keith Carver made many helpful suggestions based on their experiences with the bonk in their classes. I [hank my friends in industry and uni versifies for supplying photographs: Dr. Naresh Den uf Mi llitech Corp .. Dr. John Bryant of the University of Michigan. Me Harry Syri.gos of Alpha Industries. Professor cal Swift of [he University of Massachusetts. Dr. Mike AdlcrsleiJl of Raj theon Co .. Mr. Hugo Vifian of Hewlett-Packard Co., Mr. Mark Russell of Raytheon 0:) .. and Dr. M. Abouzahra of Lincoln Laboratory. Finally. I would like to thank the staff of John Wiley & Sons for their helpful efforts during this project. The cheerful prnfessionalism of Monique Calello (Prouucrion Editor) was especially appreciated.

David M. Pozar Amherst, MA



Electromagnetic Theory

We begin our study of microwave engineering with a brief historical overview of t11<: field. followed by a review of topics in electromagnetic. that we will need throughout the book.



The term microwaves refers to alternating current signals with frequencies between 300 MHz (3 x W~ Hz) and 300 GH.z (3 x 1011), with a corresponding electrical wavelength between>' = c] f = 1 ill and>. = 1 mm. respectively. Signals with wavelengths on the order of millimeter are C,'1iled millimeter WOI·('J'. Figure 1.1 show rhe Jocmion of the microwave frequency band in the electromagnetic spectrum. Because of the high frequencies (and short wavelengths). standard circuit theory generally cannot be used directly to solve microwave network problems. In a sense. standard Circuit theory is an approximation or special use of the broader theory of clcctromagneucsas described by Maxwell's equations. This is due to the fad thar, in general. the lumped circuit element approximations of circuit theory are not valid at microwave frequencies. Microwave components me often distributed elements, where the phase of a voltage or current challges signinc[{mly OVer [he physical extent of lht: device, because the device dimensions are on the order of [be microwave wavelength. At much lower frequencies, the wavelength is large enough that there is In ignificanl phase variation across the dimensions of a component. The other extreme of frequency can he identified as optical engineering, in which the wavelength is much ihorter than the: dirnensiuus of the component. In this case Maxwell's equations can be simplified Lo the geometrical optics regime, and optical systems can he designed with the theory of geometrical optics. Such techniques are sometimes applicable to millimeter wave );_ sterns, where they are referred to as quasioptical.

In microwave engineering, then, one must often begin with Maxwell's equations and their solutions .. It is in the nature of these equations that mathematical complexity arises, since Maxwell's equations involve vecror differential or integral operations all vector field quantities, and these fields are functions of spatial coordinates. One of the goals of this book, however, is to try to reduce the complexity of a field theory solution to a result





I.l J ntroduction to Microwave Engineering Applications of Microwave Engineering

Engineering 3


A Short History of Microwave

Maxwell's Equations 5

Fields in Media and Boundary Conditions 9

Fields at '" General Material Interface 12 • Fields at a Dielectric

Interface 14 • Fields at the interface with a Perfect Conductor (Electric

Wall) 14 • The Magnetic Wall Boundary Condition L5 •

The Radiation Condition 15

1.4 111.e Wave Equation and Basic Plane Wave Solutions 16

The Helmholtz Equation 16 • Plane Waves in a Lossless

Medium 16 • Plane Waves in a General Lossy Medium 18 •

Plane Waves in ~ Good Conductor 19

L5 General Plane Wave Solutions 21

Circularly Polarized Plane Waves 25

1.6 Energy and Power 26

1.2 1.3

Power Absorbed by " Good Conductor 29

1.7 Plane Wave Reflection "from a Media Interface 30

General Medium 31 • Lossless Medium 32



Perfect Conductor



Impedance Concept 36

1. 8 Obi i q u e Incidence at a Dielectric Interface 19

Parallel Pciarization 40 • Perpendicular Polarization

Total Reflection and Surface Waves; 43

1.9 Some Useful Theorems 45


The Reciprocity Theorem The Uniqueness. Theorem


image Theory

45 49


Wave. Propagation OIl a Trans ltd ssiou Line

The L()s~ks~ Line







2.1 The Lurnpcd-Elerneru Circuit Model for H Transmisxion Line 56

2.2 Field Analysis or Transmission Lines 59

Transmission Line Parameters 60. The Telegrapher Equations Derived

from Pleid Analysis of a Coaxial Line 63 • Propagation Constant,

impedance. ami Power Flow for the Lossless, Coaxial Line 6~

2.3 The Terminated Lossless Line 65

Special Cases of LOf.<~less Terminated r.nlc;; and Nepers n

2.4 The Smith OU.\I1 73

The Comhlncd lmpedance-Admiuunce Smith Chart

Line 79



111<' Slotted

2.5 The Quarter- Wave Transformer

The lrnpedance Viewpoint 83


• The Multiple Reflection Viewpoint

2.6 GenL;'nltor and Load Mismatches 87

l.oad Matched to Line 89 • Generator Matched W I .oaded





Lossy Transmission Lines

Tht" Low-Loss Line 90

T erminatcd Lossy Li Ill! 93

Attenuation 9..J •


1111' Disrorrionless Line.



• The Perturbation Method I'OT Calculating

The Wheeler Incremental lrrducrance Rule 96





General S{)luLi0l1S 1m Tf'M, TE, and TM Waves TE:YI Waves 107 • TE Waves 109


TNI Wave,


Attenuation Due tll Dielectric Loss 111

3..2 ParaHd Plme Waveguide

TEM Mode~ 111 •


TM Modes


TE Modes


3.3 Rectangular Waveguide 120

TE Modes I::!Q • T~' Modes J15 • TE",,, Modes of a Partially

Loaded Wa veguide 130 • Poiru oj Interest: Wm -eguide Flanges J 31

3.4 Circular Waveguide 132

TE Modes Ll3 • TM Modes 137


Coaxial Line TEM Modes

141 141

Higher-Order Modes /46


Point of

Interest: Coaxial Connectors



• Point o] lnterest:

3-6 Surface Wave:, on a Grounded Dielectric Slab

TM Modes 147 • TE Mode ['iO

Rnm·Filldlll,r: A1KoriTilll1S ! 52

3.1 Stripline 133

Formulas for Propagation C__(jn~"tJ:]L Characterise-ic Impedance. aad

Attenuation 154 • An Approximate Electrostatic Solution 157

3.8 Microsrrip 160

Formula for Erfective Dielectric Constant, Characteristic Impedance.rand

Attenuation 162 • An Approximate Electrostauc Solution 164

3.9 The Transverse Resonance Technique 167


Ti'vt Modes lor the Parallel Plate Waveguide

Partially Loaded Rectangular Waveguide 169

3.ICl Wave Velocities and Dispersion 170

Group Velocity 170

3,11 Summary of Transmission Lines and Waveguides 173

Other Types of Lines linn Guides 1/4 • Point ((fill/crest: Power Copacity

uf Tmll 'ill? i"si(m Lines ! 76




4.1 Impedance and Equivalent Voltages and Currents 18::1

Equivalent Voltages and Currents 183 • The Concept

nf lmpedance 187 • Even ~lIId Odd Properties (lr Z(wJ and nw) 190

4,::! Impedance and Adnurtance Matrices 191

Redprol:1.I1 Nerworks 19.' • Lossless Networks [95

4.3 The Scattering Matrix 196

Reciprocal Networks and l.ossless Networks 199 • A Shift in Reference

Planes 201 • Generalized Scauering Parameters 204 • Point of

Interest: The l'l" lor NI'I\L'or!.. Atlllfy:er 2U5

4.4 The Transmission (ABeD) Matrix 2()6

.. k)

Relation 10 Impedance Matrix

. etworks ::'10

Signal fJow Graphs 2 t3

Decnmposltion of SiglltJl now Graphs

Anal Y7;C[ C21i brsuon ::2 17 •

Equrvalcru Circuits for Two-Port



Apphcation 10 TRL Network

Poiru 0/ InIV rest: Computer-Aided Design for

Mil.'I'IJ!\'I1\'e Cirruits :?22

4.6 Discontinuities and Modal Analysis 22::1

Modal Analy j, l1f an /I-Plane Step In Rectangular Waveguide 225 • Poim

o] Interest: Micmsl1'.ip Discontiuuitv Compensation :?29

4,7 Excitation of Waveguides-Electric and Magnetic Currents 230

Current Sheets Ihat Excite Only One Waveguide Mode 230 • Mode

Exciration from an Arbitrary Elecuic or Magnetic: Current Source 232


6.2 Transmission Line Resonators

Short-Circuited A 12 Line 3U6

Open-Circuited -\/2 Line 31 I

6.3 Rectangular Waveguide Cavit.ic~ 3 J 3

R.e'ion~ nt F'reqllenc ies 3 J 3 • Q of the TF:I u. Mode J 15

6.4 Circular Waveguide Cavities 3J 8

Resonant Frequencies 318 • Q of the TE""d Mode 320.

0.5 Dielectric Resonators 323


• Short-Circuited ),/4




-1-.8 Excitation of Waveguides-e-Aperture Coupling 237

Coupling Through an Aperture in a Transverse Waveguide Wall Coupling Through an Aperture in the Broad Wall 01 fl W:lvegul(Je






5.1 Marching with Lumped Elements (T. Networks) 25~

Analyuc Solutions 253. Smitl: Chart Soiutious "34

11111'("('.1"1: Lumped Elements jor MicI"II["(ll'e Integrated Circuits '357

5.2 Single-Stub Tuning 258

Point fI/

Shunt Stu bs 25\)

5.3 Double-Stub Tuning


5.4 5.5

Smith Chart Solution 266 • Analytic Solution ~.7ll

The Quarter- Wave Transformer 271

The Theory of Small Reflections 175

Single-Section Transformer n6 • Multisection Trnnsformer 277

Binomial Mulrisecrion Matching Transformers 27R

Cheb, shev Multiscction Matching Transformers :::82

Chebyshev PolymlmlaJ:.> 283 • Design t)( Chebyshev 'Transformers

Tapered Lines 28'S


5.6 5_7


Exponential Taper ~9()

Tuper :291

5.9 The Bodt'-Fano Criterion

• Triangular Taper







6. J Series and Parallel Resonant Ci rcu its 300

Sene" Resonant Circuit JOO • Parallel Resonam Circuit

Loaded and Ull10aded Q 306


Resonant Frequencies of TE.)lt Mode JJ..:t

6..6 F abry- Perot Resonators 328.

Stability of Open Resonators JJO



Excrtarioo of Rexonators

Critical Coupling 332-

Resonator 334 •

Cavity Perrurbarions Material Perturbations


• A Gap-Coupled Microsuip

An Apcnure-Coupled Caviry 337




Shape Perturbations





Basic Properties of Dividers ami Couplers 35 [

Three-Port Networks I T-jL1nctions) 351 • Four-Port Networks (Direcrional

Couplers) 354 • ['(lilif of Interest: .l1eas1IIil1g Coupler Directivity 357

7.2 The T-Iunction Power Divider 359

Lossless Divider 360 • Resisuve Divider 361


7.3 The Wilki.nsnn Power Divider 363

Even-Odd Mode Analysis

Wilkinson Dividers 367

7.4 Waveguide Directional Couplers 368

Unequal Power Division and Y-WRY

Bcthe [ IDle Coupler 369 • De-ign of Multihole Coupler, 37.:1

7.5 The Quadrature 190~) Hybrid 379

Even-Odd Mode Anllly,i:; 380

7.6 Coupled Line Di rcctional Couplers 383

Coupled Line Theory 384 • Design of Coupled l.rne

Couplers 389 • Design (11 Muuisection Coupled Line Couplers 394

7.7 The Lange Coupler 7.8 The 180" Hybrid

3yg 401

Even-O£Jcl Jl'fwJe i'lnalysi., of liu.· Ring I-lyt>nd Analysis of the Tapered Coupled Lme Hybrid

NIagic-T 41 I

7.9 Other Coupler" 411

Poin: of lnterest: 71I1! Rejit'<"l(!Irut/l'r .J 14

hen-Oeld ,\.1ode Waveguide:

+OJ 407




H.I Periodic Structures 413

Analyxi« of Infinite Periodic Structures 4:!-1- • Termmated Periodic

Structures 427 • k- 1 Diagrams and Wal e Veluciues -l-2t1

8.2 Filter Design by the Image Parameter Method -+31

Image Impedances und Transfer Functions for Two-Port Nctwnrke, 431 •

Constant-I Filter Sections 433 •• m-l)~rived Filter Sections .J.3ti •

Composite Fillers 4-W


xiv Conients

8.-3 Filler Design by tne Insertion Loss Method 443

Characterization by Power Loss Ratio 444 • Maximally Flat Low-Pass Filter

Protorypc 447 • Equal-Ripple. Low-Pass Filter Prototype 450 •

Linear Phase Low-Pass Filter Prototypes 451

8.4 Filter Transformations 452

Impedance and Frequency Scaling

Transformatirms 457


Bandpass and Bandstop

a.s Filter Implementation 461

Richard's Transformation


Kuroda's Identities



Impedance and Admittance lnverters 461!i

8.6 Stepped-Impedance Low-Pass Filters 470

Approximate Equivalent Circuits for Short Transmission Line Sections 470

8.7 Coupled Line Filters 474

Filter Properties of a Coupled Line Section 4- 74 • Design of Coupled Line

Bandpass Filters 477

Filters Using Coupled Resonators 486

Bandstop and Bandpass Filters Using Quarter-Wave Resonators 486 •

Bandpass Filters Using Capacitively Coupled Resonators 490 •

Direct-Coupled Waveguide Cavity Filters 493




Basic Properties of Ferrimagnetic Materials 498

The Permeability Tensor 498 • Circularly Polarized Fields

Effecl of Loss 506 • Demagnetization Factors 50S •

Interest; Permanent Mug/Jets 570

9-2 Plane Wave Propagation in a Ferrite Medium 511


504 •

Point of

Prcpagation in Direction of Bias (Faraday Rotation)

Transverse LO Bias (Birefringence) 515

9.3 Propagation in a Ferrite-Loaded Rectangular Waveguide

TE",,, Modes of Wa l'IoglJide wim 11 Single Ferrile Slab 518

of Waveguide with Two Symmetrical Ferrite Slabs 52 J

9.4 Ferrite Isolators 523




• TE,,,,,, Modes

Resonance Isolators 523 .'

9.5 Ferrite Phase Shifters 530

Nonreciprocal Latching Pbase Shifter 530 •

Shifters 533 • The Gyrator 535

The Field Displacement lS(},laIQT


Other Types of Ferrite Phase

9,6 Ferrite Circulators 535

Properties (If a Mismatched Circulator


function Circulator







Noise in Microwave Circuits Dynamic Range anti Source, of

Noise Temp~rah:Jre 550

Y -faczor Method 553 •

C<L~caded System 557

Detectors and Mixers 559

Diode Rectifiers and Detectors 559 • Single. Ended Mixer 565 Balanced

M~eT 558 • Other Types of Mixer, 571 • Inrerrnodul ali on

Products 574 • Poin! oj Interest: The Spectrum A_naly,,~r 575

10.3 PIN Diod Control Circuits 576

Single-Pole Switches 577 • PIN Diode Phase Shifters 580

IO.4 Microwave Integrated Circuirs 583



Noise Power and Equivalent

• Measurement of Noise Temperature by the

oise Figure of a



Hybrid Microwave Integrated Circuits

Integrated Circuits 584

Overview of Microwave Sources

Solid-State Sources 589 •

584 •

Monolithic Microwave


Microwave Tubes





11.1 Characteristics of Microwave Transistors 60 I


MicTOWllve Bipolar

Microwave' Field Iiffecl Tnm.\-i.<iwrs i FEJ_'s)

Transistors 604

11.2 Gain and Stability 606

TWQ-Pun Power Gains 606 • Stability () L2

Single-Stage Tran istor Arnplifler Design 618

Design for Maximum Gain tCurrjl1g_atc Marching) 618 •

Circles and Design for Specified Gain (Unilateral Device) 622

Amplifier Design 628

Broadband Transistor Amplifier Design 632

Balanced Amplifiers 632 • Distributed Arnplrfiers

Oscillator Design 641


Constant Gain

• Low-Noise




One-Pun Negative Resistance oscntI1tors M 1 • Transistor

Oscillaum, 644 • Dielectric Resonator Oscillators 648





System Aspects of Antennas 655

Definite of Important Antenna Parameters 655 •

An/ennas 656 • Antenna Parrcru Charnc(o;:nstics

Efficiency, Gain. and. Teruperarure 661


Basic Types of

658 •


The Radar Equation

Radar 677 •

12.4 Radiometry 679

Theory and Applications uf Radiometry 679 •

Radiometer fl81 • The Dicke Radiometer


Pulse Radar



xvl Contents

12.2 Microwave Communication Systems 662

Types of Cornmurucauon Systems 662 • The Friis Power Transmission

Formula 663 • Microwave Transmitters and Receivers 666 •

Noise Characterization of II Microwave Receiver

Prequency-Multiplexed Sy terns 670

12.3 Radar Systems 672


Radar Cross Section 678

Tetal Power 6X4

12.5 Microwave Propagation 685
Atmospheric Effects 685 • Ground Effects 687 • Plasma
Efects 688
12.6 Other Applications and Topics 689
Microwave Healing 689 • Energy Transfer 690 • Electronic
Warfare 69L • Biological Effects and Safety 694 APPENDICES 697


Prefixes 698 Vector Analysis Bessel Functions

698 700

Other Mathematical Results Physical Constants 7()4


F Conductivities for Some MaLcrial~ 704

G Dielectric Constants and Loss Tangents for Some Materials 705

H Properties or Some Microwave Ferrite Materials 705

Standard Rectangular Waveguide Data 706

J Standard Coaxial Cable Data 707


F requency (fu)

s x IO~ ))( 106 Jx 107 J:x 108 s x 109 3}( J01Q

3x JOII 3x lOlz 3x Hil~ 3 x rol4
1 I I
1 g I :51
1 .", I ~l
~ ~
I ~ I ~I
I ,;: I £1
I 8 .5 I .!:! 1
I u.. I >1
! I I
10-3 lO-4 [I)-~ 2

Chapter 1: Eleelromagnetic Theory



J r.s: I I lEI I I I ~I I 1~151 I 1'p..I.",1 I 12: I e I I 1 .... 1.01 t I I~I I

1 l!:!:ol!




Wavelength (m)

Typical Fr~uencies

AM broadcast band Shortwave radio FM broadcast band VHF TV (1.-4) VHF TV (5-6) URFTV (7-13) UHF TV (14-:83) Microwave ovens

AppFoximate BIlnd Designations

535-1605 kHz 3-30 MHz 8&--108 MHz 54--72 MHz 7("",88 MHz 174-216 MHz 47Q.-c89U Mfu 2.45 GHz

lrband S-band C-band X-band Ku-band K-band Ka-band U·band

1.-2GHz 2--4 GHz 4--S GHz &--12 GHz. 12-L8GH1; 18-26GJf:>. 26-40Gll~ 40--60 GHz

FIGURE 1.1 The electromagnetic. spectrum.

that can be expressed in tenus of simpler circuit theory. A field theory solution generally provides a complete description of the electromagnetic field at every point in space, which is usually much more information than we really need for most practical purposes. We are typically more interested in terminal quantities such as power, impedance, voltage, and current, which can often be expressed in terms of circuit theory concepts. It is this complexity that adds to the challenge, as well as the rewards, of microwave engineering.

Applications of Microwave Engineering

Just as the high frequencies and short wavelengths of microwave energy make for difficulties in analysis and design of microwave components and systems. these same factors provide unique oppornmkies for the application of microwave systems. This is because of the following considerations:

• Antenna gain is proportional to the electrical size of !he antenna. At higher frequencies. more antenna gain is therefore possible for a given physical antenna size, which has important consequences for implementing miniaturized microwave systems.

• More bandwidth (information-carrying capacity) Can be realized at higher frequencies. A 1 % bandwidth at 600 MHz is 6 MHz (the bandwidth of a single television channel), and at 60 GHz a 1 % bandwidth i5600 MHz (100 television channels). Bandwidth is critically important because available frequency bands in the electromagnetic spectrum are being rapidly depleted.

1.1 Introduction to Microwave Engineering

• Microwave signals travel by Iine of sight and are not bent by the ionosphere as are lower frequency signals. Satellite and terrestrial corrununication links with very high capacities are thus possible. wuh frequency reuse at minimally distant locations.

• The effective reflection area (r-adar cross section) of a radar target is usually proportional to the targets electrical size. This fact. coupled with the frequency characteristics of antenna gain. generally makes microwave frequencies preferred for radar systems.

• Various molecular, atomic. and nuclear resonances occur at microwave frequencies, creating a variety ().f unique applications III the areas of basic science. remote sensing. medical diagnostics and treatment, and heating methods.

Today, the majority of applications of microwaves are related to radar and Communication systems. Radar systems are used for detecting and locating air, ground, or seagoing targets and for air-traffic control systems. missile tracking radars, automobile collision-avoidance systems, weather prediction, motion detectors, and a wide variety of remote sensing systems. Microwave communication systems handle a large fraction of the world's irnemmional and other tnag-118uI telephone. data, and televisioIl fransmissions. And most or the currently developing wireless telecommunications systems. such as direct broadcast satellite (DBS) television, personal communications systems (PCSs), wireless local area computer networks (WLANS). cellular video (CY) systems, and global positioning satellite (OPS) systems, operate in the frequency range 1.5 to 94 GHz, and thus rely heavily on microwave technology.

A Short History of Microwave Engineering

The field of microwave engineering is often considered a fairly mature discipline because the fundamental concepts of elecrromagnctics were developed over 100 years ago, and prohahty because radar, being the first major application of microwave techHo)ogy, was intensively dcv eloped as far back as Work] War [L But even though microwave engineering had its beginnings in the. la;-,t century. significant developments in high-frequency solid-state devices, microwave integrated circuits, and the ever-widening applications of modem microsy rerns have kept the field active and vibrant.

The foundations of modem electromagnetic weary were formulated in 1873 by James Clerk Maxwell [1]. who hypothesized solely from mathematical considerations, electromagnetic wave propagation and the notion that light was a [01111 of electromagnetic energy. Maxwell'j, formulation was cast in its modern form hy Olive; Heaviside. during the period from 1885 to 181n. Heaviside was a reclusive genius whose efforts removed many of the mathematicalcomplexities of Maxwell's theory. introduced vector notation, and provided a foundation for practical applications of guided waves and transmissron lines. Heinrich Hertz, a German professor of physics and a gifted experimentalist who also understood the theory published by Maxwell, carried out a set of experiments during the period 1887-1891 that completely validated Maxwell' s theory of electromagnetic waves. Figure 1.2 shows a photograph of the original equipment used b. Hem in his experiments. It is interesting CO observe thai this is an instance of a discovery occurring after a prediction has been made on theoretical grounds-a characteristic of many of the



Chapter 1: Electromagnetic Theory

FIGURE 1.2 Original apparatus used by Hertz. fOJ his elecrrornagneries experiments. (I) 50 MHz transmitter spark gap and loaded dipole antenna. (2) Pan a llel wire grid tor polarizution experiments. t3) Vacuum apparaLus rorcalhode ray experiments, (4) Hot-wire galvanometer. (5) Reiss or Knochenhauer spirals. (6) Rolled-paper galvanometer, (7) Metal sphere probe. l8) Reiss spark micrometer, (9) Coaxial transmlssion Iine. (l0-12) Equipment (0 demonstrate dielectric polarization effects, (l3) Mercury induetioucoil interrupter. (4) Meidinger cell. (15) Vacuum bell jar, (l6) High-voltage induction coil. (17) Bunsen cells. (18) Large-area conductor for charge storage. (19) Circular loop receiving antenna. (20) Eightsi ded recei ver detee tor, (2 I) R otarin g mirror and m ercury interru pter. (22) Square loop receiving antenna. (23) Equipment fOJ refraction and dielectric constant measurement. (24'1 Two square loop receiving antennas. (25) Square loop receiving antenna, (26) Transmitter dipole. (27) High-voltage induction coil. (28) Coaxial line .. (29) High-voltage discharger, (30) Cylindrical parabolic reflector/receiver. (31) Cylindrical parabolic reflector/transru iuer, (32) Circular loop receiving antenna, (33J Planar reflector. (34, 35) Battery of accumulators, Photographed on October I, 1913 at the Bavarian Academy of Science, Munich. Germany. with H ertz' s asslstam, Julius Amman.

Photograph ami identificaticn courMsy of I. H. Bryant. Universiry o'f Micnigan,

major discoveries throughout the history of science. All of the practical applications of eleerromagneric theory, including radio, television, and radar, owe their existence to the theoretical work of Maxwell.

Because of the lack of reliable microwave sources and other components, the rapid gI0wth of radio technology in the early 19DOs occurred primarilyin the high frequency


12 MaxweH's EqL/atJOns

lHFl to very high frequency (VHF) runge. II was 1]1)1 until tnt: lY-1-Ds und the advent of radar development during World ~(aJ U that microwave rheory and technology received substantial lnrerest. In the United Slate s, the Radiation Laboratory was established at the MassachusettS Institute of Technology l~llT) to ti('\clop radar theory and practice. A nurnner of top sci<!ru.ists, indudJ!lg N, I\jarcul itl. 1. L R:1bi. 1. S. S,:h"~'iilger_ H_ A. Bellie, E. M. Purcell. C G. Montgomery, and R. H. Dicke. among others. were gathered for what turned (lUI 10 he. a very intensive period of development in tile microwave field. Their work indmk-d the thcoreticill and expcmnenlal Ireatml!-nt of \HI\ egulue components. microwave antennas, small aperture coupling theory. and [hoe beginnings of microwave network theory Many of these researchers were pllJ sicists who wenr bad, La plJ)'sics research after the war ImAmy later n-ceived NobeJ Prizes), I1ll1 their microwave work is summarized in Li1C classic ~8-\'oILll11e Railiation Laboratory Series of boob that still finds. application today.

CommullicfltloJ1<'; "yslem" using microwave technology began In be developed 5110n after the birth or radar. benefitting: from much of the work that was original I y done for radar systems. The advantage" offered hy microwave systems, including wide bandwidths and line-of-sight propagation. have proved lo be critical 1'01' hoth terrestrial and satellite communications systems and have thus provided an impetus for the continuing development of low-cos! minuuurized microwave tompoll<.:ms. We refer the imeresied reader to the <peciul Crntenniul lssue of the IEEE Transactions 011 Mic/"IIlI'LlI'f' Tlieor; Wid Techniques l21 lor further blsroricul perspectives on the field of microwuve engineering-


Electric and magnetic phenomena at the macroscopic level are de-scribed hy 1ftxwells equations. a:. published by IVlaxwel1 tn It-!7:1 Ill- Tlus work summarized the stale of electromagnetic science at th.u wne anJ hyporhesized from theoretical considerations the existence or the electrical displacement current, which leu [0 ihe discovery by l lerrz and Marcon i of electromagnetic v. ave propugution. M~lX well' -; wnrk wa~ based on a large body of empirical and theoretical knowledge developed 11) Gauss, Ampere. Faraday, urnl others. A Ilrst course in elcctrornagnctics usually follows this hlviorical (or deductlvc) approach. and it j;;, assumed that the reader has had such a course a:. a prerequisite to the pTescm material. Several hooks are available, 131-191. that provide a g011U treatment or electromagnetic theory aj_ the undergnuluate or graduate level,

This chapter-will outline the lundumental concepts 11[ elecrrornagneuc theory [11m we wi 11 requi rc for the re $1 nf thc book. Max \l, ell' ~ equal 1011" W1 LI be presented. and boundar) conditions amJ the effecr of dielectric and magnetic matcrial« \\ iii be ui~CU"S~lL Wave phenomenon is of essential importance in microwave engineering, so much of rhe chapter is spent on plane v. ave topics. Plane \\ ave~ are tilt: simples; Iorm Dr electrermngncuc waves and so sen-e ro illustrate a nurnhcr or ba."il· properties as<;oeillteu with WilV€! propagation, Although It IS assumed that the reader hu studied plane waves before. the present material should help to reinforce many at' the basic principle- in the readers mind and perhaps to introduce some concepts thar the reader ha:,. not seen previously_ This material will <11-;0 serve as a useful reference lor later .haptcrs.


1.1 b


Chapter 1: Electromagnetic Theory

With an awareness of the historical perspective, il is usually advantageous from a pedagogical point of view LO present electromagnetic theory from the "inductive," or axiomatic, approach by beginning with Maxwell's equations. The general form of time-varying Maxwell equations, then, can be written in "point," or differential. form as

- -8B'V x E = fit - M,

- 815 -

'V x 'H = at + 3,

'V. fJ = p, 'V,B=O.,


J.Ic l.1d

The :MKS system of units is used throughout this book. The script quantities represent time-varying vector fields and are real functions of spatial coordinates x,'I).zo. and the time coordinate t. These quantities are defined as follows:

t is the electric field intensity, in VIm.

1t is the magnetic field intensity, in Aim. 1) is the electric flux density, in Coul/m''. B is the magnetic flux density, in Wb/rn2,

Mis the (fictitious) magnetic current density, in V/m2 . .J is the electric current density, in Alm2,

pis the electric charge density, in Ceul/nr'.

111e sources of the electromagnetic field are the currents M and .:J, and the electric charge p. The magnetic current M is a fictitious source in the sense that it is only a. mathematical convenience: [he real source of a magnetic current is always a loop of electric current or some similar type of magnetic dipole, as opposed to the flow of an actual magnetic charge (magnetic monopole charges are not known to exist). The magnetic current is included here for completeness, as we will have occasion to use it in Chapter 4 when dealing with apertures. Since electric current is really the flow of charge. it can be said that the electric charge density p is the ultimate source of the electromagnetic field.

In free-space, the following simple relations hold between the electric and magnetic field intensities and flux densities:

B = MH, ij = fot,



where Ji.{J = 470 X 10-7 henry/m is file permeability of free-space, and EO = 8.854 x 10-12 faradJm is the permittivity of free-space. We will see in the next section how media other than free-space affect these constitutive relations.

Equations ({ 1.1 d) are linear but are not Independent of each other. For instance, consider the divergence of ( Since the divergence oftbe cud of any vector is zero

1.2 Maxwell's Equations

[vector identity (B.12), from Appendix Bj, we have

- 0 - -

v" V x £ = 0 = - ar; . B) - '\ .. 1\.,1.

Since there is no free magnetic charge. V .. M = 0, which leads to v ·8 = O. or (1.10). The continuity equation can be similarly derived by taking the divergence of (1.1 b). giving


where (l.Ic) was used. This equation states that charge i~ conserved, or that current is continuous, since '\1. j represents the outflow of CUITem at a poi nt. and ap! at represents the charge buildup with time at the same point. It is this result thai led Maxwell to the conclusion that the displacement current density i)f)! Dt was necessary in 11.1 h). which can be • ceo by lHking. (he dh·crgellce or Lhis equation.

TIle foregoing differential equations call be converted to integral form through the use of various vector inregral theorems. Thus. applying the divergence theorem (B.IS) to (1.1 c) and ( 1.1 d) yields

J V· d§ = j f/ an = Q,

Js \


r. 13· cD,; - O.



where Q in (1.4) represents the total charge contained in the dosed volume l' (enclosed by a closed surface 5'). Applying Stokes' theorem {B.16 \ to (1.1 al give.

,,[ t dl = . l'c.

o /. -

- B·d.,;

of ..


r, JV! . as, Is

which, without the .M term, is the usual form of Faraday's law and forms the basis for Kirchhoffs voltage law. In (1.6), C represents a dosed contour around rbe surface S. as shOWn in Figure 1.3. Ampere's law can be derived by applying Stokes theorem to n.u»

J n dT = .D I' f) . as -f i s . d'5 = .0 l tr. tis + T.

J(' Of . s .Is cJt ./8

where I. = Is ... 1· dli is the total electric current Row through the surface S. Equations (1.4)-(1.7) constitute the integral forms of Maxwells equations.

The foregoing equations are valid for arbitrary time dependence. but most of our work will be involved with fields having a sinusoidal, or harmonic, time dependence. with steady- tale conditions assumed. In this case phasor notation is very convenient. and so all field quantities will be assumed to be complex vectors with an implied I:J"-'I lime dependence and written with roman (rather than scrip I ) letters. Thus. 1:1. sinusoidal electric field in tbc J- direction uf [be foml




E(x,y, z) = 5:A(x,-y,z)eN.



Chapter 1; ErectromagneNc Theory

FIGURE 1.3 The closed contour C and surface S associated with Paraday's law,

where A is the (real) amplitude, ta is the radian frequency. and ¢ is the phase reference of the wave at t = 0, has the phasor form

We will assume cosine-based phasors in this book, so the conversion from phasor quantities to real time-varying quantities is accomplished by multiplying the phasor by eiwl and taking the real part:

t(x, v, Z, t) = Re[E'(x, v. z)e1wt],


as substituting (1.9) into (LlO) to obtain (1,8) demonstrates. When working in phasor notation, it is customary to suppress the common &,wt. factor on all terms,

When dealing with power and energy, we will often be interested in the time average of a quadratic quantity. This can be found very easily for time harmonic fields. FOI example, the average of the square ofthe magnitude of an electric field given by

t = :LEI cos(wt + <PJ) + fj~ cos(wt + rP2) + 2Ez cos(wt + <p-]), 1,11

which has the phasor form

E == xE1ej¢1 + fJ~ei<h + 2E3ej¢J 1


can be calculated as

lEI;" = ~ loT E· Edt

= ~ iT [E; cos2(wt + <PI) + E? cos2(wt + r/h) + E5 cos~(wt + <b)] dt

1 E22 2 11 -12 1._ ,.. ~

= 2:( I + Ez + B,) = 2: E = 1,E' E .

Then the root-mean-square (nus) value is IElnns = IEI/.J2.

Assuming an ejw1 time dependence, the time derivatives in (Lla)--(1.1d) can be replaced by jw, Maxwell's equations in phasor form then become


'V x E = -jwB -M, 'V x it = jwD + .1,

L14a 1.14tJ


'.3 F!elds in Media and Boundary Conditions 9

\J . D = o. 1.14r.:

V . 11 = O. 1.l4d

The Fourier transform can be used to convert a solution to Maxwell's equations for an arbitrary frequency w to a solution for arbitrary time dependence.

The electric and magnetic current sources, J and if, in (1.14) are volume current densities with units Alm1 and V/tnI, respectively. In many cases, however. the actua] currents will be in the form of a current sheet. a line CUITenl. or an infinitesimal dipole current. These special types of current distributions tan always be written as volume current densities through the use of delta functions, Figure 1.4 shows examples of thi~ procedure for electric and magnetic currents.


In the preceding section it was assumed that the electric and magnetic fields were in free-space, with no material bodies present. In practice, material bodies are often present; this complicates the analysis but also allows. the useful application of material properties to microwave componems. When electromagnetic fields exist in material media. the field vectors are related to each other by the constitutive relations.

For a dielectric material, an applied electric. field E causes-the polarization of the atoms or molecules of the material to create electric dipole moments that augment the total displacement flux, D. This additional polarization vector is called t: the electric polarization. where


In a J'inear medium, the electric polarization is linearly related to the applied electric field as


where Xc. which may be complex. is called the electric susceptihility. Then.

I.! 7



is the complex permittivity of the medium. The imaginary par! of f accounts for loss in the medium (beat) due to damping of the vibrating dipole moments. (Free-Sp<lce. having a rea] t;, is Iossless.) DUe-!O energy COnservation .. as we will see 1./1 Sec/ion J .6. the imaginary part of (: must be negati ve (~/I positive l. The loss of a dierectric materi <11 may also be considered as an equivalent conductor loss. In a material with conductivity IT. a conduction current density will exist:

.J = uS,



Chapter 1 :Electf0magnetiC Theo.ry






L-====~ 1,.«(, y) Aim

lex. y, l) = -'rlx. Y) 6(<. - zo) Alm~


M,.cx. y) V 1m - - 1 M(x . .\', ~) == M/x, y) 6(z - ~J Vim


y .;(~(x) A



.rV,,(x) V



M(x. s. ~) ==.l:Vix} 8(" - .?;J li~ -""l V!ml


f(_lt . .1'. z) == _"[,,(x) IiU' - )'0) .I:l!~ - "") Alrn2



!I A-Ill ...............


VI V-m

- [x". y". 2"Q)


x M(x,j'.;:) ==.~VlB(.T-X,,) 6(y- )'6) 6(z -<:,,) Vim?


J(.T, y. z) =;flO(x-x,,) 8(y- Ya) 6(;:;:- z~) Aiml

FIGURE 1.4 Arbitrary volume, surface. and line currents. (a) Arbitrary electric and magnetic volume current densities. (b) Arbitrary electric and magnetic surface current densities in !be z = zo plane. (c) Arbitrary electric and magnetic line currents. (d) Infinitesimal electric and magnetic dipoles parallel to the z-axis.

which is Ohm's law from an electromagnetic field's point of view. Maxwell's curl equation for H (1.14b) then becomes

\l x H = jwD + J = jwdt+ rrE

1.3 Fields in Media and Boundary Condi11ons 11

= jWf! E t fWf" + rr JE:

= .i<.<>-(rc' - .1/1 • j ~ i= 1.20


where it is seen that loss due LO dielectric damping (vJ(") is indistinguishable from conducuviry loss (0-). The term .ae" + IT can then be considered us the total effeerive conductiviry. A related quantity of interest is the loss tangent. defined as

which is seen to be the ratio of [he real to ihc imaginary part of the total displacernem current. Microwave materials are usually charucterizeo b) specifying the real pcnnntivity, f' = ~rfl\. and the loss tangent at <I certain frequency. These constants arc listed in Appendix G for several types of marerials, lt is useful lrJ note tbal. ilf!er a problem ha.~ been solved assuming a lossless dielectric, loss can easily be introduced by replacing the real ( with a complex ( = t' - Jr1' = f'(l ~ j tan oJ.

In the p reccdi ng discussion it was i\S5U rued that P, was a vector i n the '\~ me direction U~ E. Such mall:rials are caBed isotropic materials. out not all materials ila vc tllL ... property, Some materials are anisotropic and art' characterized by a more complicated relation between p,. and E. or lJ and E. The most general linear relatitm between these vectors lakes the form of a tensor i)f ran I... two i a dyad). \\ h ich can be wriuen ill rnatri x fann as


It is rhus seen that a given vectnr component of E gives rise. in general, to three components of D. Crystal structures and. ionized gases are examples of anisotropic dlelectrjc.~. Por /1 linear isolwpic material. the main). of I J .~:?) would reiJuce m a di,lgDllal matrix with elements L

An analogous siuration occurs for magnetic materials, An applied rnagnetic Reid may al i g n magnetic dipole moments i 11 11 rnagneri c rnaten OIl to prod u Coo' a III agrietic pn larization i or magnet izau 0111 Pm . Then,


For :1 linear magnetic ma/erial . .Pm i ... linearly related to tJ as


where 1m i!\ a complex magnetic susceptihility. From (1.23) and (l.:?:4).


where /1 = p.()(1 + '(mJ = !/ - Jill! is the permeability of the medium. Again, the i muginary part of "7" or /I accou n is for loss due to damping forces; there u. 11(l magnetic conductivity, since there is no real magnetic current, A~ in the electric case, magnetic



Chapter 1: Electromagnelie Theory

materials may be anisotropic, in which case a tensor permeability can be written as

An important example of anisotropic magnetic marerials in microwave engineering is the class of ferrirnagnetic materials IlIlOIVIl as ferrites; these materials and their applications will be discussed further ill Chapter 9.

[f linear media ate assumed (('I.~ not depending on E or [f). then Maxwell's equalions can be written in phasor form as

\I X E = -jwft.fI - j~I, 'V x R = jWfE + .I,

V fJ = p,

\1. B = O.

1.27a L27b 1.27c 1.27d

The constitutive relations are

fJ =f-E;, B=I.LH.



where e and J1 may be complex and rna}' be tensors. Note [hat relations like (1.28a) and (1.2Sb) generally cannot be wriuen in time domain form, even [or linear media, because of the possible phase shift between fJ and E. or Band fro The phasor representation accounts for this phase shift by the complex form of t and 11.

Maxwell's equations (1.27a)-O.27d) in differential form require known boundary values for a complete and unique solution. A general method used throughoutthis book is to solve the source-free Max well's equations ina certa i n region to obtain solutions with unknown coefficients. and then apply boundary conditions to solve for these coefficients. A number of specific Cases or boundary conditions arise, as discussed in the following.

'Fields at a GeneraJ Material Interface

Consider a plane interface between two media, as shown ill Figure Lfi. Maxwell's equations in integral form can be used to deduce conditions involving the normal and tangential fields at thi s interface. The rime-harmonic versi on of (l A), where S is the closed "pillbox't-shaped surface shown in Figure 1.6. can be written as

1 jj. tis = f p dv.

Js i:

In the limit as h -l 0, the contribution of Dtan through the sidewalls goes to zero, so (1.29) reduces to


l:!.SD1"-!:::,,SDI~' = t::..Sp~, D2n - DIn. = PD'


'.3 Fields in Media and Boundary Conditions

0,,", P.,


FIGllitl: 1.5 FidJl,. cwrenh. ;:mJ surface charge at L! ~ncrill imcrface between twn media..

~ktliuIrl I


where p.< is the surface charge tlenSllJ· on the interface. In vector Ierrn. we can write


A similar argiuneru Ior jj ]eD(JJ, 11:) 111(' resLJlllhlll


since there ix 1]0 free mag-nelil:: charge.

F0r the iangenua! components Dr the cleCirlC field \\ e use the phasor forru of 0.6 I.

/' .ff· d,ii"


in connection with the closed contour C shown in Figure 1.7, In the limit a~ II - O. the surface Integral of f3 vanishes (since S = II j.( varu-Jres). l'he contrihutiou from the surface integral lIC a, however. 11111) be nonzcra if a magnetic surface current density

vkdillm I


FTGFRE I. 7 Clo~cil contour C fur equauon (1.331.






Chapter 1: Electromagnetic Theory

!VIs exists on the surface. The Dirac delta function can then be used [0 write


where h is a coordinate.measured normal from the interface. Equation (1.33) then gives

which can be generalized in vector form as

(Bh - Ed x n = his.


A similar argument for the magnetic field leads to


where Js is an electric surface current density that may possibly exist at the interface. Equations (1.31), (1.32), (1.36). and (1.37) are the most general expressions for the boundary conditions at an arbitrary interface of materials and/or surface currents,

Fields at a Dielectric Interface

At an interface between (WO lossless dielectric materials. no charge or surface current densities will ordinarily exist. Equations (1.31), (1.32). (1.36), and (1.37) then reduce to

n· b, = fi· [hn· HJ = fi· Fh, f1, x El = n x Eq, n x tIL = fi X H2.


1.38b I.3ac 1.38d

In words. these equations state that the normal components of [) and B are continuous across the interface, and the tangential components of E and fI are equal across the interface. Because Maxwell's equations are nOL all linearly independent the six boundary conditions contained in the above equations are not all linearly independent. Thus, the enforcement of (1.38c) and (l.38d) for the four tangential field components. for example will automatically force the atisfaction of the equations for the continuity of the normal components.

Fields at the Interface with a Perfect Conductor (Electric Wall)

Many problems in microwave engineering involve boundaries with good conductors (e.g., metals), which can often be assumed as lossless (u -J (0). In this case of a perfect conductor, aJ] field components must be zero inside the conducting region. This result can be seen by considering a conductor with finite conductivity (u < (0) and noting that the skin depth the depth to which most of the microwave power penetrates) goes to zero as (T -J 00. (Such an anal ysi. will be peIfo:r:me.d in Section 1.7.) If we also assume here that iY[" = O. which would be the case if the perfect conductor fined all the space 011

1.4 FreId:; in Medra and Boundary Conditions

one side of the boundary, then (LJ!). (1.32), (1.36), and /1.37) reduce LO the following:

fI_ ._D = p". fl· B =0: ft x t = O. ft,;.,; il = }8'

l.39a ! .390 1.39(' U9d

where ps and .7£ are the electric surface charge density and current density, respectively, on the interface. and il is the normal unit vector pointing out of the perfect conductor. Such a boundary is also known as an electric wall, because the tangential components of E are "shorted out," as seen from (1.39c), and must vanish at the surface of the conductor.

The Magnetic Wall Boundary Condition

Dual to the preceding boundary condition IS the magnetic wall boundary condition. where tbe tangential components of B must vanish. Such a boundary does 1111L feally exist in practice, but may be approximated by a corrugated surface, or in certain planar (nmsmis,,'ioll line problems. In addilio[l, the ideali2ati0l1 that fi. x JJ = 0 at ill! interface is often a convenient simplification. as we will see in later chapters. We will also see that me magneliL' wall boundary condition is analogous 10 the relations bel ween the voltage and current at the end of ill] open-ci rcuited transmission line, while the electric wall boundary condition is analogous to rhe voltage and current atthe em] of a short-circuited transmission line. The magnetic wall condition, then, provide" a d.egree of completeness in our formulatlon of boundar}' conditions and is a useful approximation in several easel> of practical interest.

The fields at a magnetic wall satisfy the following conditions:

fi· tJ = O. il' B = O.

ft x E -= -Xfo" h x f1 = O.

IAOa lAOI-I lADe L40d

where '11 is the normal UOLt vector pointing out of the magnetic wall region.

The Radiation Condition

When dealing with problems that have one or more infinite bonndaries, such as plane waves in an infinitl! medium, or infinitely long transmission lines, a condition on the fields at i I1n n ity must be enforced. This boundary condition is known 1:"\5. the radiation condition, and is essentially a statement of energy conservation. It states thai, at an infinite distance from a source. the Gelds must either be vanishingly small (i.e., zero) at propagating in an outward direction. This result can easily be seen by allowing the infinite medium to C(Jntain a small loss faCtor (a~ any physical medium wou1.d have).



1 .. 4

Chapter 1: Electmmagnetic Theory

Incoming waves (from infinity) of finite amplitude wouJd then require an infinitesource at infinity, find so are disallowed.


The Helmholtz Equation

In a source-free. linear, isotropic, homogeneous region, Maxwell's curl equations in phasor form are

\l x E = -jwMH, V x if = jwcE,



and constitute two equations for the two unknowns, E and iI. As such, they 'can be solved for either E or H. Thus, taking the curl of (Ld la) and using {L41b) gives

- . - 2-

'V x V x E = - jwp V x H = uJ pEE,

which is an equation for E. This result can be simplified through the use of vector identity (B.14), V x V x A = V(V . ;:1) - \72.4, which is valid for the rectangular components of an arbitrary vector A Then,


since V . E = 0 in a source-free region. Equation (1.42) is [he wave equation, or Helmholtz equation, for E. An identical equation for fr can be derived in the same manner:


A constant k = w.jiii; is defined and called the wavenumber. or propagation constant, of the medium; its units are 11m.

As a way of introducing wave behavior, we will next study the solutions to the above wave equations in their simplest fOIIllS, first fora Iossless medium and then fora lossy (conducting) medium,

Plane Waves ina lossless Medium

In a loss less medium. E and f1. are real numbers, so k is real. A basic plane wave solution to the above wave equations can be found by considering an electric field wtth only an X component and uniform (no variation) in the x and )' directions. Then, al8x = alay = 0, and the Helmholtz equation of (1.42) reduces to

a2Ez;2 .

8z2 + k Ex =0.


1.4 The Wave Equation and Basic Plane Wave Solutions

The solutions [0 this equation are easily seen. by substitution, to be- of the form E.r(::j =: ETt-Jf:< +E-e!h.


where £-t and E': are arbitrary amplitude C0l1S1an T,S.

The above solution is for the lime harmonic case at frequency ,», In the time domain.

this result is written as


where we have assumed that E" and E- arc real constants. Consider the first term in ll.46). This tenn reprt:senrs a wave traveling in the ,<: direction, since, to maintain a fixed point on the Wave (u,JI - I.:: = constant). one must move in the +z direction as rime increases. Similarly. the second term in (1.46) represents a wave traveling in the negative z direction: hence the notation 0 I and E- for these wave amplitudes. The velocity of the wave in this sense is called the phase velocity, because it is the velocity at which a fixed phase point 011 the wave travels, and it is given by

(lp = dz = .!!_ (""I - constant) = ~ = __ 1_

at ill k k"fiLE


In free-space. we have up = I! V 11'0 eO = c = 2.998 x lOR rn/sec. which is the speed of ligbt-

The- wavelength. >.. is defined as the distance between two successive maxima (or

minima. Or any other reference points) on the wave, at a fixed instant of time. Thus

[ .... 1 - trot] - [ .. ,)/ - k(.~ + I\ll = 2IT.


1~ "l-n /'

A == ~ == ... nll P == ---f!-.

J; ..,J .I


A complete specification uf the plane wave electromagnetic field must include the magnetic field. Tn general. whenever E or fT i~ known. the other field vector can be readily found by using one of Maxwell's curl equations, Thus. applying. (1.41.a) to the electric field of (1.45) gives HI = H: = O. and

H - ~[EI -J"'=_E--1A-"1

'J- j(' {.-.

, '1/


where T] = uJf,.~/k = //1/" i:, the wave impedance for the plane wave. defined as the ratio of the E and H fields. For plane waves, this impedance is also the intrinsic impedance of the medium. In free-space we have ')11 = J IlLi / to == 377 n. Note that the E and H Vectors are orthogonal to each other and orthogonal to the direction of propagation (_:l_;;_l; this is a characteristic of transverse electromagnetic (TEM) wave ..

~ EX..-\,MPLE 1.1 iB A plane wave with a frequencj material with e, = 7 and /1" = J, wave impedance: for this ...... ave,

Basic Plane ~'ave Parameters

of ? GHz i propagating in an unbounded Compute the wavelength. phase velocity. and


I¥'. ~

'I] = = 1]0 - = 377 - = 246.Sfl.

f'r 7


18 Chapter 1: Electr0magnetic Theory


From (1.47) the phase velocity is

1 C 3x108 _ 7

'V = -- = -- =. = 6.5) x IO m/sec,

1> "filE ..jJl,.Er J(7)(3)

This is slower than the speed of light ill free. space by a factor of J2l = 4.58. From (l.48) the wavelength is

). = 'Up = 6.55 X 107

f 3 X 109 = 0.0218 m.

The wave impedance is

Plane Waves in a General Lossy Medium

Now consider the effect of a lossy medium. If the medium is conductive, with a conductivity, a Maxwell's curl equations can be written, from (lAla) and (1.20) as

'V x E = ~.iwJ.kH

'V x n = jwtE + aE.

1.50a 1.50b

The resulting wave equation for E then becomes

'V'2j£ +W2jt€ (1 -.1 :JE = 0, where we see a similarity with (1.42), the wave equation for E in the lossless case. The difference is that the wavenumber k'2 = w2 p,£. of (l.42) i replaced by w2 J.kE[ 1 - j (rr / W€)] in (1.51). We then define a complex propagation constant for the medium as


l' = Q + .1f3 = .1W.../iiiV 1 _ j a



If we again assume an electric field with only an :i: component and uniform in x and y, the wave equation of (1.51) reduces to


f)z2 .... ~TE'" =0,


which has solutions


The positive traveling wave then has a propagation factor of the form

1.4 The Wave Equation and Basic Plane Wave Solutions

which in the time domain is of the form

We see then that this represents a wave traveling in the +z direction with a pha e velocity tip = uJ lB. a wavelength ), = 271 /.d, and an exponential damping factor. The rate of decay with distance is given by the artenuation constant, ri, The negative traveling wave term of (1.54) is similarly damped along the -z axis. I f the Loss is removed. IT = 0, and we have '1 = j k and (1 = 0, Ii = k,

As ui~cussed in Section 1.3. Loss can also be treated through the use of a complex. permittivity. From (1-20) Wit11 (1 = 0 bur I: = (.I - ,it" complex, we have (ba(


where tan ~ = E" I e' is the loss tangent of the material.

Next, the associated magnetic field can be calculated as

. DE -'"

J[ - J 'x _ J) (E+ -'F E- "P)

.1---,------ (- - f" .

,wp 8~ ,JJJI

As with the lossless case, a wave impedance can be defined to relate the ele uric and


magnetic fields:

I_ip '/=--.



Then (1.56) can be rewritten as


Note III at 1]' is. in general, complex and red uces 10 the lossless case of 17 = ~ when 'I = Jk = j...v";/--IL

Plane Waves in a Good Conductor

Many problems of practical interest involve loss or attenuation due to good (bUI not perfect) conductors, A good conductor is a special case of the preceding analysis. where the conductive current is much greater than the. displacement current, which means [J -::?'> we Most metals can be categorized as good conductors, In terms of a complex e, rather than conductivity. this condition is equivalent LO (il »(.1. The propagation constant of U.5 2) can then be adcquatel y approximated by ignori ng the displacement current term, to give

r = n + j(j ~jw/JK ~ = (J + ))\!"';/J.-O.

V y::;; ! :2

The skin depth. or characteristic depth of penetration, is defined as



0< = - = --.

. CI WftlT



jwp. . flWp. . 1

7/ = - ':::: (1 + J) - = (I + J)~.

") 2(T (T<)"

Notice that the phase angle of this Impedance is 45°, a characteristic of good conductors. The phase angle of the impedance for a lossless maleria! is OQ, and the phase angle of the impedance of an arbitrary lossy medium is somewhere between 0° and 450•

Table 1.1 summarizes the results for plane wave propagation in lossless and lossy homogeneous media.



Chapter 1: Electromagnetic Theory

Then the amplitude of the fields in the conductor decay by an amount 1/ e. or 36.8%, after traveling a distance of one skin depth, since e-"" = e-alJ• = e-I. At microwave frequencies, for a good conductor, this disraace is very small. The practical importance of this result is that only a thin plating of a good conductor (e.g., silver OJ gold) is necessary for low-loss microwave components.

EXAMPLE 1.2 Skin Depth at M.icrowave Freqnencles

Compute the skin depth of aluminum, copper, gold, and silver at a frequency of 10 GHz.


The conductivities for these metals are listed in Appendix F. Equation (I.60) gives the skin depths as

D .• = {d; = j 7fl~()(T = j 7f(J010)(4~ X 10-1) ff

= 5.03 X 1O-3/f.

For copper:

> . 0-3 V 1 -1

b,,=5.03xl .. ;]=8.14xlO m.

3.616 x 10 .

OB = 5.03 X 10-3. /~ 8 I .07 = 6 .. 60 x 10-7 111.

V ). 13 x 1

s, = 5.03 x lO-3/ 9 1 01 = 7.86 xl.O-i m,

V 4.0 8 x 1

For aluminum:

For gold:

For silver:

s, = 5.03 x 10-3) 6.1731X 107 = 6.40 X 10-7 m.

These results show that most of the current ftow in a good conductor occurs

in. an extremely thin region near the surface of the conductor. 0

111e wave impedance inside a good conductor can he Obtained from (1.57) and (I_59).

The result is

1.5 General Plane Wave Solutions


Type or Medium

TABLE 1.1 Summary of Results rOT Plane Wavl'! Prnpagatlon in Various Media


Good Conductor


Lossless (fir = (J = 0)


-. = II +.JJ j _1.)/10"/2

Complex propagation constant

'1 = jwJfif

= jUhjp.f' J(I - .itr)/ ... -e'

Phase con slant (wavenurnben


Attenuation constant

11 =0

I] = Vp./€ = "'·,l/k

1,1' = Ret"/) r{ = J;.,)/-h 1\, = l/n ,\ = 2,,/.3 I'p = ..,;13


Skill depth

Wavelength Phase velocity

A = 211/,j I'p = ..2/.)

", J

e »f ort1»w~

.9 = lrn(l) = ,ju;pr:r/'J. fl = Re(~I) = v' .... '1'(1" /'1 '7 = (1 + J) j_')~112a (.s = J2/11i'WT ,\=27,/.1

"p = ..,;/3



Some specific features of plane waves were discussed in Section 1.4. Here we will look at plane waves again. from a more general point of view. and solve the wave equation by the method of separation 01" variables. This technique will find application in succeeding chapters, We will also discuss circularly polarized plane. waves, which will be importarrt for the discusS"ion of ferrites in Chapter 9.

In free space. the Helmholtz. equation for E can be writtenas

.,_ ,_ ()2E; a~E 8~E~-

v- E + k:5E = ~ + ~ + -:--) , + kiiE = O.

0;"- 1.'1/- ( 2-

and this vector wave equation holds for each rectangular component of E:

[yl E, 82 E, ft· t: .,_

~ +: ,ecl ~ - ~ + koEi = O.

oe: uy- UZ"

when~ the i..n.UI'!X i = 1C, iJ. or 6. 'I11'1(', e.qumiull wH.1 now 'De ~olverl by the method of separation of variables. a standard technique for treating such partial differential equations. The method begins by assuming that the solution 10 (1.63) for. S,ty E.r' can be written as a product of three functions for each of the three coordinates:

E",(;r, /j. z ) = f(·1·) .rJ(y) h(t).

Su bstituti ng this r ann into (I. 63) and dividing by f 9 h gives

/,1 /I 1,1

_ + [!_ + ~ + ~,2 = 0, f 9 h n





1 .. 69


Chapter 1: Electromagnetic Theory

wherethe double primes denote the second derivative. Now the key step in the argument is to recognize thateach of the leans in (1.65) must be equal to a constant, since they are independent of each. other. Thai is, r If is onlya function of .7:, and the remaining terms in (1.65) do not depend on :/:, so !"I f must be a constant, and similarly for the G.ther terms in (1.65). Thus, we de.fine !h.ree separation constants, k"" ky and k;, • .'web that

r If = - k!; gil I 9 = - k;;

rP f 2 rfg 2

or dx2 + k!];f = 0; dy"l + ky9 = O~

Combining (1.65) and (1.66) showsthat

k1 .' 1 •• 1 + k.2 = k2 '" T fl,y .z 0'

h"ll! = -12·


d!h ·2

-d 2 +kzh=O. z



The partial differential equation of (1.63) has now been reduced to three separate ordinary differential equations in (1.66). Solutions to these equations are of the form e±jk""" e±jl<"y, and e±ik.,z, respectively. As we have seen in the previous section. the terms with + signs result in waves traveling in the negative z, y, or z direction, while the terms with =signs result in waves traveling in. the positive direction, BoLh solutions are possible and are valid; the amount to which these various terms are excited is dependent on the source of rbe fields, For our present discussion, we will select a plane wave tra,veling in the positive direction for each coordinate, and write the complete solution for Ex. as


where A is an arbitrary amplitude constant. Now define a wavenumber vector k as

Then from (1..67) If I = Au. and so n is a unit vector in the direction of propagation, Also define a position vector as



then (1.68) can be written as

E ( '. a) - A -j/i;." '" x,,!{Z - e .


Solutions to (1.63) for Ey and E; are, of course, similar in form to Ex. of (1.11), but with different amplitude constants:

E· ( ) B -jk·"

11 x,y,z, = e ,


E ( ) - C -jk·"

z ;1::, y,z - e .


The :x, y, and z dependences of the three components of E ill (1.71 rO. 73) must be the same (same kia;, ky, k~), because the divergence condition that

. - 8Ex. BEy ss;

\,J . E = -.-.-. + - .. -. - + ~. ~ = 0

ax ay 8:;:

1 .5 General Plane Wave Solutions

must also be applied ill order to satisfy Maxwell's equation". which lrnplies that Ex. E.I;I' and E; must" each have the same variation in .c,!/. and z , (Note that the solutions in the preceding section automatically satisfied tile divergence condition. since E.1i was the only component of E. and Eo-: did not vary with .1'.) This condition also imposes a constraint on the amplitudes A. B. and C, since if

En = AJ - B.O + ('5:

we have


t: = E;(le-.ikf:

'V' F; -= \'. ('Eo!' Jkr) = Eo' \'t.-Jkr = +j]: .'£\/t::-j!;.r = o.

where vector identity (B.7l was used. Thu . we must haw:

i En = 0,


which means thai [he electric 'field amplitude vector En must he perpendicular to the direction of propagation, A:. This condition is a general result for plane waves and implies that only two of ihc three aruplitude constants. A. B. and (", tan be chosen independentl y.

The magnetic field can be found from Maxwell equation,


to give

- -.I i':' x" -J~""

- 1':"() \ ('


-j ~ k-

= --E,l (-jk)r-i'r


~·(I.~ E- - fIC· "

= --Il X llf'

...... 110 .

I - E~' -j~' r = -n x 1)1':


- -II x E:,


where vector identity !B.91 was used in -obtaining the second line. This result shows that the magnetic field intensity vector H lies III a plane normal £0 1.:, [he direction of propagation. and that li is perpendicular to E. See Figure 1.8 for an illustration or these vector relations. The quantity lJ() = J /-in)".!) = 377 n in (1.76) is the intrinsic impedance of free-space.





Chapter 1: Electromagnetic Theory


FlHURE 1.8 Orientation of me E, II, and k = A'{Jii vectors fora plane wave.

The time-domain expression for the electric field can be found as

£(:z;, v. .it, t) = Re{ E(x, v; z)e,1WC!} = Re{ Eoe~ jk·f e/'-Jt} = Eo cos(k . l~ - wt),


assuming that the amplitude constants A,B, and G contained in Eo are real. If these constants are not real, their phases should be included inside the cosine term of (1.77), From (1.77) we see that a wavelength, defined asthe distance the wave must travel to undergo a phase shift of 2)T,can be found as

IklAn = kOA(J = 2:n-,

or Ao = 2'1[/ ko, which is identical to the result obtained in Section J.4. The phase velocity, or the speed at which one would have; to travel Lo maintain a constant phase point on the wave, can be found from the condition that

k . f - wt = constant.

so that

ko1Jp =w,

'Up' = w/ko = _1_ = c = 2_998 X 108 m/sec,


Taking the derivative with respect to time gives dr

kOdt -I'll = 0,


as in Section 1.4.

EXAMPLE 1_3 Current Sheets as of Plane Waves

An infinite sheet ~f surface current can. be _conside:ed _as a source for plane waves. If anelecrric surface current density J" = Jox exists 011 the z: = 0 plane in free-space, find the resulting fields by assuming plane waves on either side

of the CUITenl sheet and enforcing the boundary conditions.

1.5 General Plane Wave Solutions


Since the source does not vary with x or y, the fields will not vary with z: or Y but will propagate away from the source in the ±z direction. The boundary 1:ond~tious to be satisfied at z = 0 an:

fl x (Eh - Ell = i x (Eh - E1) = O.

fj. x (fh - HI) = i X (Hz - jll) = lox,

where £1. HI are the fields for z < 0, and £2, R2 are the fields for z> O. To satisfy tbe second condition. H must have a fj component Then for E t6 be orthogonal to H and £, E must have an i component, Thus the fields will have the following form:

£1 = xAT/l]e)kllZ, HI = _YAe.ik<'r". £1 = .'i:HI/oe-.iI"J?,

fl.l = YBF-.iku".

where A. and B are arbitrary amplitude constants. The first boundary condition, that E~, is continuous at z = 0, yields A = 8. while the boundary condition Jor rr yields the equation

for -z < 0,

for ~ » 0,

-B -.4 = L;

Solving for .4, B gives

which completes the solution.


Circularly Polarized Plane Waves

The plane waves discussed above all had their electric field vector poinung in a fixed direction and so are called I inearly polarized waves. In general, the polarization of a plane wave refers to the orientation of the electric field vector, which may be in a fixed direction or may change with lime.

Consider the superposition of an :t. linearly polarized wave with amplitude EJ• and a y linearly polarized wave with amplitude E,,_. both traveling in the positive z direction. The total electric field can he written as


A number of possibilities W1W arise. If E, :j:. 0 and E2 = 0, we have a plane wave linearly polarized ill the ,i- direction. Similarly. if E, = 0 and E2 #- 0, \>.'C have a plane wave linearly polanzed in the fj direction. ffC, and E2 are both real and nonzero, we have a plane wave linearly polarized at the angle


£(z, i:) = Eo{x cos(w,t - koz) + Y cos(wt - koz - 7l/2)}.



Chapter 1: Electromagnetic Theory

For example. if Bt = ~ = Eu. we have

£ = Eo(1: + fj)e-:ikuz.

which represen tsan electric field vector ,in a 45°angle from the .r-axis.

Now consider the case in which EI = .iFh = Eo, where Eo is real, so that

E = Eo(i: - j1/)e-JI;~".

TI1e time domain form of this field is


This expression shows that the electric field vector changes with time or, equivalently, with distance along the .a-axis, To see this, pick a fixed position, say z = O. Equation (1.80) then reduces to

teo, t) = Ekl{.r cos wt + fj sin wt},


SQ as wt increases from zero. the electric field vector rotates, counterclockwise flam the z-axis, The resulting angle from the z-axis of the electric field vector at time i; at z = 0, is then

(sin wt)

rP = lan-I -- =wt.

cos wt

which shows that the polarization rotates at the uniform angular velocity w. Since the fingers of the right hand point in the direction of rotation when the thumb points in the direction of propagation, this type of wave is referred to as a right hand circularly polarized (RHCP) wave. Similarly, a field of the fonn

E = Eo(:1t + jfj)e- jk1>" 1.82

constitutes a left hand circularly polarized (LHCP) wave, where the electric field vector rotates in the opposite direction. See Figure 1.9 for a sketch of the polarization vectors For RHCP and LHCP plane waves.

The magnetic field associated witha circularly polarized wave may be found from Maxwell's equations, or by using the wave impedance applied to each component of the electric field. For example, applying 0.76) to the electric field of a RHCP wave as given in (1.79) yields

- EOk £Ok jEo .r..-

E = ~i x (i - jfj)e-N" = -('0 + ji)e-) .,,~ = -' -(i - jfj)e-J....,Z,

1]0 '1'/0 ''70

which is also seen to be a circularly polarized wave with a polarization vector of the RHCP sense.



In general, a source of electromagnetic energy sets up 'fields that store electric and magnetic en(l;l"gy and carry power that may be transmitted or dlssipatedas loss. In me

1 ,6 Energy and Power




FIGURE 1.:9 Electric field polarization for (a) RHCP and (b) LHCP plane waves.

sinusoidal steady-state case, the time-average stored electric energy In a volume Il lS given by',

1 j' _ _

1'1' e = -Re E, D: dv,

4 t

which in the case of simple [ass less isotropic, homogeneous, linear media. where F is a real scalar constant, reduces to

We = i IE, E;* d».

Similarly. the time-average magnetic energy stored In the volume V is

I j' _ -

IF", = -Re . H, IY dv,

4 v

which becomes

Wm = * ! fJ· n: d»,

. ,

fora real, constant. scalar Ii.

We can now derive Poynting's theorem. which leads LO energy conservation for electromagnetic fieldxand sources, If we have an electric source current, .18' and a <xlllduclion current (TE, as defined in (1.19), then the total electric current densrty is J = Js + a E. Then multiplying (1.27a) by tt- . and multiplying the conjugate of O,27b) by E. yields

II . (\7 x E) = -j...,,'tL1Hf - Ir, Ms.

F:. ('V x £[*) = E: JV - jWf"i"IE11 = E, 1: + 01£'12 - .jwf:IEf',

where 1vJ~ is the magnetic source current. Using these two results in vector identity (B.B)


1 33





28 Chapter 1: Electromagnetic Theory

'V . (E x .fI~) = IT' . (\7 x E) - it . (\7 x fl"')

= -0-1E12 + jW(E~ IEI2 - p1H12) - (8:. J: + fl* .. IiiIg).

Now integrate over a volume V and use the divergence theorem:

1 'V. {E x Fr)dv = Is Ex Fr - ds

= -(1 [ 11W dv + _'-w [ (e" IEI2 - J.LIH-r~)dv - l (E· J~. + lr . lC.(,) (lv, 1.87

Jv t. t.

where S is a dosed surface eI1c!O&ing the volume V, as shown in FigH]1C l.lO. to = fi' - j(_11 and 11, = fil' - j/-l/' to be complex to allow for loss, and rewriting (1.87) gives

This result is known as Poynting's theorem. after the physicist J. H. Poynting (1852- J 914), and is basically a power balance equation. Thus, the Integral on the left-hand side represents the complex power, PSI delivered by the sources ],. and Ms,inside S:

Pg = -~ r (E. J; + fr ·l\1s)d'l.I. .Iv

The first integral on tile right-hand side of (1.88) represents complex power flow out of the closed surface S. lfwe define a quantity called the Poynting vector, tI, as



then this power can be expressed as

11~ - lje;

Po- "'" "2 Is E x H' . d.s = 2 Is" . ds.


FIGURE 1.10 A volume V, enclosed by tbe closed surface S, containing fields it, {f, ~d cunen t sources J5, ]1;[,..

1.6 Energy and Power

ll"l.e.S\.lii"-dCC. S In l!.I;}! \ mu"l he. a closed surlace in order fOT ilii;; i.merpr-eU'ttlQfI to be valid, The real parts of P, and flu in (1.8.9) and (I _91 J represent ti me-a verage powers.

The second and th i I'd i ntegral sin (I. 8E) are real quail tn ies represemi ng the limeaverage power di ssipated in the volume V due to conducri vi ty. dielectric. an d magnetic losses. If we define this power ;to, PI' We have that


which is sometimes referred \0 :1:0. jul.l k· slaw. The \ ::lSI in,egral i 11 \ 1 .8'6) can be xeen to be related to [he stored electric and magnetic energies, as def ned in 0.84) lind (1.86).

With (he above definitions. Poynting's theorem can be rewritten as


III words. this complex power balance equation states that the power delivered by the sources (P.~) is equal tu the sum of tile [lower transmitted through [he surface (Po), the power lost to heat in the volume IPiJ. and 2....,' times the net reactive energy stored m the volume,

Power Absorbed bya Good Conductor

To calculate artcnuatlon and Joss due to animperfect conductor. One must find tnt: power dissipated in the conductor. We will show that this can be done using only [be fields at the surface of me conductor, which is a very hel pful simpl i lication when calculating attenuation.

Consider the geometry of Figure 1.1 I, which shows the interface hetween a losslcss medium and a good conductor. We assume thru a field is incident Irorn ;:; < 0 am] that the field penerratcs into the conducting region z > 0, The real average power cruering the conductor volume defined by [he cross-sectional surface 811 at the interface and the


/Lo E


"" ...




\ I I


J;~I I





S~ I


_./ ----------\--


FlGURE 1.11 An interface between :J lossless medium and ,1 good conducrnr with a closed surface Sr, + S fer computing the- power uissipatcd in the conductor.



Ch~pler 1: Electromagnetic Theory

surface S is given from (1.91) as

] 1 - - .. h

Pay = +Re Ex H . nds,

2 So+8

where n is a unit normal vector pointing into the closed surface So + S, and E. if are the fields at this surface. The contribution to the integral in (1.94) from the surface S can be made zero by proper selection of this surface . For example, if the field is a normally incident plane wave. the Poynting vector S = E x n- will be in the z direction, and so tangential to the top, bottom. front, and back of 5, if these walls are made parallel to the z-axi . If the wave is obliquely incident, these walls can be slanted to obtain the same result, And, if the conductor is good, the decay of the fields from the interface at z = 0 will be very rapid, so that the right-hand end of S can be made far enough away from z = 0 so that there is negligible contribution to the integral from this part of the urface S. The power entering the conductor through SO C~ then be written as

1 1 - -. ~

p .. v = -Re E x H . z ds.

2 SD

From vector identity (B.3) we have

ince H = n x Ej7]. as generalized from (1.76) for conductive media, where 1} is the intrinsic wave impedance of the conductor. Equation (1.95) can then be written as


Rgll- 7

Pay = 2 HI-ds,


[ .~f.t fgp.. 1

R~ = Re(ry) = Re (1 + J) - = - = ---:c

2rr 20' rru~

is called the surface resistivity of the conductor. The magnetic field ff in (1.97) is tangential to the conductor surface and needs only to be evaluated at the surface of the conductor; since H; is continuous at z = 0, it doesn't matter whether this field is evaluated just ont.tide the condl1ctor or just inside me conductor. In me next section we will show how (1.97) can be evaluated in terms of a surface current density flowing on the surface of the conductor, where the conductor can be assumed to be perfect



A number of problems to be considered in later chapters involve the behavior of electromagnetic fields at the interface of a lossy or conducting medium. and so it is beneficial at this time to study the reflection of a plane wave normally incident from freespace onto the surface of a conducting half-space. The geometry is shown in Figure 1.12 where the lossy half-space .z > 0 is characterized by the parameters €, u, and (J.






1.7 PI.ane Wave Reflection from a Media Interface


If. p.. fT

FlGURE 1.12 Plane wave reflection from a lossy medium; normal incidence.

General Medium

With no loss of generality. we can assume that the incident plane wave has an electric field vector oriented along the z-axis and is propagating along the positive z-axis. The incident fields "an then be written, for z < CL as

E-· - x~-E"e-Jk""

1- JIJ •


J .990

where 1]0 is tbe wave impedance of free-space. and Eo is an arbitrary amplitude, Abo in the region z < 0, a reflected wave may exist with the form


1.1 DOb

where r is the unknown reflection coefficient of the electric field, Note that in (L IDOl. the sign in the exponential terms has been chosen as positive. to represent waves traveling in the -z direction of propagation, as derived in (1.49). This is also consistent with the Poynting vector Sr = E,. x E1; = -WI1IEoI21/1]O, which shows power [0 be traveling in the -z direction,

As shown in Section l.4., from equation (1.54) and (1.58), the transmitted fields for z > 0 in the lossy medium can be written as

EI = :tTEoe--r'. tt, = JjTEOp_1~, 7/


1.10 I b

where T is the transmission coefficient of the electric field and /} is. the intrinsic impedance


7 = a + .i!3 = jw..jjii VI - i« !w£..



Chapter 1: Electromagnetic Theory

of the lossy medium in the region z > O. From (1.57) and (1.52) the intrinsic impedance is


and the propagation constant i'S

We now have a boundary value problem where the general form of the fields are known via (1.99)-{l.JOl) on either side of the material discontinuity at z = O. The 11.'10 unknown constants, r and T, are found by applying two boundary conditions on E" and H lJ .31 z = O. Since these tangential field components must be continuous at z = 0, we arrive at tile following two equations:

(I + f) = T, l-r T

710 11

Solving these equations for (be reflection and transmission coefficients gives




r=--, fl + Tlo


T= I +f = __l!L.

T/ + l70

This is a general solution for reflection and transmission of a normally incident wave at the interface of a lossy material, wheren is the impedance of the material. We now consider three special cases of the above result



If the region for z > 0 is a lossless dielectric. then a = O,and J1. and f_ are real quantities. The propagation constant in this case is purely imaginary and can be written as

'Y = j{J = jwJlIi = jko JI.trEr,


where ko = w~ is tire wavenumber of a plane w:ave in free-space. Tbe- wavelength In the dielectric is


which is seen [0 be shorter than the wavelength in free-space (AO)' The corresponding phase velocity is.

w 1 e

vp = - = -.-. = -----,.

i1 ,ffif ~


1.7 Plane Wave Reflection Irom a Media Interface

which is slower than the speed of light in free-space fe). The wave impedance of the dielectric is

. jwp fjf.l JI!;

rl = -- = - = l]() -.

~I . f (,_


which may be _greater or lesser than the impedance of free-space (1]0), depending 0]] whether It,.. is greater or lesser than "r- In the Iossless case. 1') is real, so both rand T from (I. ! 05 J are real, and E and n are. in phase with each other in both medi urns.

Power conservation for the incident. reflected, and uansmitted waves can be demonstrated by computing the Poynting vectors in the two regions. Thus, for e < 0, the complex Poynting vector is

s- = t x.rr = (E; + Er) x (Hi + fiT.)

= zIEl)f~(f-JIq,Z + I'ejk"")(e-j"'oz - I'eJko'T 170

= zIE~ll_2._(1 -lfl2 + fp2J"'o.t _ r~c-ljf,:"-,) I/O

= iIBol'2_!_(1 -1[12 + 2jf sin 2kO;;), '70

t. 110«

since I' is real For z > 0, the complex Poynting vector is

which can be rewritten, using (1.105). as

c+ ~IE 11 41/ "IE 11 I (1 11"12)

v = Z CIt) = Z 0 - -. '.

(77 + r]Q)~ 110


Now observe that at z = 0, S = g+. so that comple-x power flow is conserved across the interface. Now consider the time-average power flow in the two regions. For z < o. the time-average power flow through a I-m 2 cross section is

. - . 1 -_ ~ I 1 I 1

P = -Re(S z) = -IEol -(1 -1r! ).

2 2 7]0


and for z > 0, the time-average power flow through a l-m 2 cross- section is

+ I - i, 1. .~

P =-2Rc(S~. £) = -IEol~-(1 -Ifl~) = P-,

,. 2 rIo

1.111 b

so real power Bow is conserved.

We now note a subtle point. When computing the complex. Poynting vector for z < 0 in (1.1 lOa). we used the total E and fI fields. If we compute separately the



34 Chapter 1: Electromagnetic Theory

Poynting vectors for the incident and reflected waves, we obtain

5-= E.; x H~ = £\Eo\2

'f .. 1. )


Sr = Er x B; = -2 IEol21rl2 .


and we see that s. + S r =F s: of (1.11 Oa), The missing cross-product terms account for stored reactive energy in tbe standing wave in the z < 0 region. Thus, the decomposition of a Poynting vector into incident and reflected components is not, in general. meaningful. Some books define a time-average Poynting vector as (1/2)Re(E x H*), and in this case such a definition applied to the individual incident and reflected components will give the correct result, since Pi = O/2)ReIEo 121TJo, and P; = -(1/2)IEoI2 jrp~ Ii/O, so P; + P; = P-. But even this definition will fail to provide meaningful results when the medium for z < 0 is lossy.



Good Conductor

If the region for z > 0 is a good (but not perfect) conductor, the propagation constant can be written as discussed in Section 1.4:


Similarly, the intrinsic impedance of the conductor simplifies LO

Now the impedance is complex. with a phase angle of 45° so E and H will be 45" out of phase, and rand T will be complex. In (1.113) and (1.114), 08 = I/o. is the skin depth, as defined in (1.60).

Let us check the complex power balance in thi case. For z < 0, the complex Poynting vector is

S- = Ex H" = CEo + EI') x (Hi +HrY = z\EO\l_!_(l - \r\2 + 2j Im(re2:ikll%»), i/o

1. 115 a

which can be evaluated at 2 = 0 to give

For z > O. the complex Poynting vector is

s: = s; x fI; = iIEo\2ITI2_!,e-2i>i, '1'/

1.7 Plane Wave Reflection from a Media Interface


and using (I.! 05) fur T and r gives

/~+ = ~I Eoll 47/ 1 c -2.,;~ = 21Eo 12 J_(! - 1f12 + r - r*)e-zo".

171 + 'in 1- 'IJo

So at tbe interface at z = 0, 5'- = $+. and complex power is conserved.

Observe that if we were to compute the separate incident and reflected Poynting


vectors for J! < 0 as

.,-- -_JEol~

S; = £1 X H'i ;= Z--,.


S- -E- H-· __ ~IE(]12In'-

r - ~,x - N ~

r /70

we do not obtain $, + S,.. = ,9- of ( l.Il5a), even for z = O. It is possible, however, to consider real power flow in terms of the individual traveling wave components. Thus, the time-average power flows through a I-m~ cross, section are

] .IIOn


_ 1 -_ _ I I IZ 1 ~Iz

P = 'JRe(S . 2) = ?,Eo, -0-11 ).

- - 1)0

r= = ~Re(S+. z) = ~IEoI2J_(1 - I rf)C';';nz .

2 2 'rio

which shows power balance at z = O. In addition .. Pi = I Eull/}I]o. and PI" = -IEoI2 IfI2/2t/D so that P; ---t- P; = P . showing that the real power flow for <: < 0 can be decomposed into incident and reflected wave components.

Now notice that s-. the power density in the lossy conductor, decays exponentially according to the e-la.z atrenuarion factor. This means that power is being dissipated in the lossy material as the wave propagates into the medium in the +z direction. The power, and also the fields, decay [0 a negligibly small value within .a few skin depths of the material, which for a reasonably good conductor is an extremely small distance at microwave frequencies.

The electric volume current density flowing in the conducting region is givenas



tt 18

and so the average power dissipated in (or transmitted into) a I m.?: cross-sectional volume of the conductor can be c-alculated from the conductor loss term of ( 1 ,92) (J ou les law) as

• 1 J - - 1 11 1,1 J''''-

P' = ? E. J~ tiu = ? (:'l'EIJTe--")· (/i:aRoTe-O,,,), dz dy dx

- II - x=[) y=e ;:",(j

= ~ulEol2 ITI2 j.= c-2<u dz = u1Eol~IT12 1.119

2 z=D 4~

Since 1/1) = a6~/(l + j) = (a j2a){l- j). the feci power entering the conductor through a 1-012 cross section (as given by (I j2)Re( 5" . i) at -z = 0) can be expressed using (Ll 15b) as pl. = lEo I~ ITI2(O' !4n), which is in agreement with (1.119).


Chapter 1: Electromagnetic Theory

Perfect Conductor

Now assume that the region z :> 0 contains a perfect conductor. The above results can be specialized to this case by allowing o _, 00. Then, from (1. 113) Q _, 00; from (Ll 14) 'I ..... 0; from (1.60) b .• _, 0: and from (l.105a,b) T _, O. and r _, -1. The fields for z > ° thus decay infinitely fast and are identically zero in the perfect conductor. The perfect conductor can be thought of as "shorting out" the incident electric .field. For .Z < 0, from ( 1.99) and (1.100), the total E and fI fields are, since r = - J,

E = Ei + B, .. = xEo(e-jko~ - dko;::) = -.';;2jEo sin koz.

- - J Of. -k 2

H = HI + HI' = 'fj-Eo{e-J '9Z + e' <)2) = y-Eo cos koz-

I}o 170

Observe that at 1: = 0, E = 0 and 11 = f;{2/'I}o)Eo. The Poynting vector for z < 0 is


- - - 4 I?

S- = Ex H* = 2j-IEo - sin koz. cos koz.




which has a zero real part and thus indicates that no real power is delivered to the perfect conductor.

The volume current density of (1_118) for the lossy conductor reduces to an infinitely thin sheet of surface current in the limit of infinite conductivity:

pt = _! ( E. J"d1! = ___!__ r iJj2dv.

2 Jv 2u lv

For a 1 m2 area of conductor surface, the power transmitted through this surface and dissipated as heat is, from (1.119),


- - ( 2 ) 2

J8 = f~ x H = -2 X f)-Eo cos k():: 1=0 = .'i-Eo A/m. 1.122

ryu ryo

The Surface Impedance Concept

La many problems, particu larly those in which the effect of attenuation or conductor loss is needed, the presence of an imperfectconductor must be taken into account. The surface impedance concept allows us to do this in a very convenient way. We will develop this method from the theory presented in the previous sections,

Consider a good conductor in the region z :> O. As we have seen, a plane wave normally incident on this conductor is mostly reflected, and the power that is transmitted into the conductor is dissipated as beat within a very short distance from the surface. There are three ways to compute this power.

First, we can use Joule's law, as in (t .119):

pt = a-IEoI2ITI2 40:

Usin$ (1.1501)) for T. (1.114) for 'T}. and the fact that a: = 1/08"gives the following

1.7 Plane Wave Reflection from a Media Interface


0-6,,,41771:= .~ ~ _ 8R."

il1 + TJO 11 - 0-0,,116 '7fi'

where we have assumed "/ « Tlu, which is true for a good conductor. Then the above power can be written as,


? •




( I + ) ) 1 I.;;p..

R_, = Re(n) = Re -- = - = V-:;-

an" rIO~ _0"


is the surface resistance of the metal.

Another way lo find the pmvcr loss is to compute [he power flow into the conductor using tl:ie Poyrrti ng vector, since all power entering the conductor at s: = 0 is dissipated, As in (l.llSh). we have

which for large: conductivity becomes, since 'I « 'Iu.


which agrees with (1.124).

A third method uses an effective surface current density and the surface impedance. without the need for the fields inside the conductor. From ( 1.1 J 8). the volume current density in the conductor is


so the total current flow per unit width lL1 the .r direction is

and taking the limit of o'T Ii for large (T gi es

rrl5s 2'1 ~ ~ 2(1 + j)

(I - j) (1) + 1")0) (.1 +.iJ (1(\,7111

- ~ Un

J~ =:1:- A/m. 110

If the. conductivity were infinite, a true surface current density of

qY ')





- ~ I - - J 2Eo

1.. = fi x H :=0 = -z X (Hi +H,.)I"=,, = .rEu-11 - I") = J'--' Afm

770 770

would flow, which is identical LO the toja) current in (1.128).



Chapter 1: Electromagnetic Theory

Now replace the exponentially decaying volume current of 0-127) in Joule's law of (1,123) with a uniform volume current extending a distance of one skin depth. Thus. let


so that the total current flow is the same. Then use (l.123) to find the power lost:

where J~ denotes a surface integral over the conductor surface, in this case cho en as 1 m2. The result of (1. 130) agrees with our previous results for pt in (1.126) and (1.124) and shows mat me power loss can be calculated as

pt = R'~lIJ 12d = Rsj'lfI 12d

2 s s 2 t S

s s


in terms of the surface resistance R" and the surface current JI" or tangential magnetic field Ht. It is important LO realize that the surface current can be found from h = 11, x H, as if the metal were a perfect conductor. This method i. very general. applying to fields other than plane waves, and to conductors of arbitrary shape, as long as bends or comer. have radii on the order of a skin depth or larger. The method i also quite accurate as the onlyapproximation in the above was that 'T] <K 'r/o. which is a good approximation, As an example, copper at 1 OHz has I'r/I = 0.012 n, which is indeed much less than 1]11 = 377 O.

EXAMPLE 1.4 Plane Wave Reflection from a Conductor

Consider a plane wave normally incident on a half-space of copper. If f = 1 GHz. compute the propagation constant, impedance, and skin depth for the

conductor. Also compute the reflection and transmission coefficients.


For copper, a = 5.813 X l07S/m, so from (1.60) me skin depth is

I3s = (2 = 2.088 x 10-6 m.

V~ ,

and the propagation constant is, from (L I 13),


I = ~ = (4.789 + j4.789) x 105 m-J•


1 .. 8


1.8 Oblique Incidence at a Dielectric Interface

The intrin ic impedance is, From (1.1J4),

l +] 1

·'1 = -,- = (8,239 + )8.239) x 10- n. u?J5

which is quite mall relative to the impedance of free-space (1]0 = 377 fiJ. The reflection eocfficicnt is then

r = 'I - 'Tin = l/179.99<> 11+ 11n

(practically tnar of an Ideal short circuit), and the transmission coefficient h



We. continue ,1ur discussion uf plane W,WCR by cnm;\de.ring the prob~em of a plaI\e wave obliquely incident on a plane interface between (WO lossless dielectric regions. as shown in Figure 1.13. There are two canonical cases of this problem: tile electric field is either in the r : plane (parallel polarization), or normal to the .r:: plane (perpendicular polarization). An arbitrary incident plane wave, of course, may have a polarization that is neither of these, but it can be expressed as a linear combination of these two individual cases.

The general method of solution is similar to the problem of normal incidence: we will write expressions for the incident, rerlected, and transrniued fields in each region and match boundar)' conditions to find the unknown amplitude coefficients and angles.

£1,11-1 Region [

FIGURE L13 Geometry Ior a plane wave obliquely incident at the interface between two dielectric regions.




Chapler 1; Electromagnetic Theory

Parallel Polarization

In this case, the electric field vector lies in the xz plane, and the 'incident fields can be written as

E; = Ell(x GOS ()i - i sin 8~)e-jkl(xs;nlf.+zooseil, Hi = EOye-jkl(XSlnlh+z<Xlsllil,



where Itl = WVIlOEL, and 171 = J/kJ/~1 are the wavenumber and wave impedance of region 1. The reflected and transmitted fields can be written as

ET = Eor(xcos8,. + Z sin e,.)e-j-k,(..";,, fi,,-"OO'lllr), Hr =-Eof f)e-jkIC:z; sin tJ.-zcos6'd,


Et = EoTe§; COS ()t - z sin Bt)e-jk,(x sln I.!,,+z c.;,; 6'1),

fIt = EoT ye-jl'2lrss;nBt+zcose,).





1. 134b

In Ute above, rand T are !he refiecaonand transrmssron coefficients, and /VZ, '12 are !he wavenumber and wave impedance of region 2. defined as

kz = WJItOf.2,

A( this point, we have I', T, 8r·, andOt as unknowns.

We can obtain two complex equations for these unknowns by enforcing the continuity of E", and Hy, the tangential field components, at the interface at z = D. We then obtain

cos (J,e-,ikl", sui 0; + r cos Or e-jk, x "in 1/" = T COS 9t e-jl..'1.Z .inO" 1 .~-jl,,,x.ine; r e-jk,o: <in 0:,- __ T ~-ji0""in8t

-e - - . ~

1)1 ·'11 172

Both sides of (1.l35a) and (L135b) are functions of the ·coormnate x. If E", and Hy are to be continuous at the interface z = 0 for all $, then this x variation must be the same on both sides of the equations, leading to the following condition:



which results in the well-known Snell's laws of reflection and refraction:

0i = Or,

k: sin (Ji = Ii;,. sin (}t.

1.136a 1.136b

The above argumentenaures that the phase terms in (1.135) vary with :tat the same rate on both sides of the interface, and so is often called the phase matching contif)ion.

1.8 Oblique Incidence at a Dielectric Interface

Using (1.136) in (1.135) allows us to solve for the reflection and transmission coefficients a .

T' _ 712 cos flt - '1)1 COs B, - 1'/;1 cos (it + 1/1 cos a, J


l')2 cos i3t + 111 cos e,.·

Observe that for normal incidence, we have {)i = (j.,. = Ot = O, so then



I' = 17] -T/I 'I'/?,+771


T= ,

7)2 +""'1


in agreement with the results of Section 1.7.

For this polarization. a special angle of incidence, (h. called the Brew ter angle. exists where r = O. This occurs when the numerator of (1.l37a) goes to zero Wi = Bb):

which can be reduced using

to give


Perpendicular Polarization

In this case, the electric field 'lector is perpendicular to the x,z plane. The incident field can be written as

E- . - 1:' y~e- J Ie, \'" SIn 11.+ zoos eI.)

)- GO . ,

it = EO(-;fcoS()i + z sin (},)e-Jk.f:rsme,+Hostl". 1)1



where kl = WJliOfl' and 171 = v;.;;r;; are the wavenumber and wave impedance for region 1, as before. The reflected and transmitted fields can be expressed as



Et = EQTfje-Jk'lT. sm IIt+zcoS(lll,

- EDT "I(;' • (} ()

H, = --(-x cos (Jt + z sin fJde-J 2,X.5ll1 "Z~D" d.


i .141a





Chapter 1: Electromagnetic Theory

with k2 = W.jltOf.2 and Til = ../110/f2 being the wavenumber and wave impedance in region 2.

Equating the tangential field components Ey and HI: at z = 0 gives

e-jl;;lx,inll, +Pe-j&lxsio9r = Te-jk-~:r'mej, L142a

-J cos ();e-jkl3Jsin9; + r cos BI_e-jklX sin Or = -T cos ete-jk,xsin8,. 1. 142b

'171 111 1/2

By the same phase matching argument that was used in the paralle1 case, we obtain Snell's laws

identical to (I.] 36),

Using (1.136) in (1.142) allows us to solve for the reflection and transmission coefficients as

r = 172 cos (h -1]J cos 8t, 1] 1 cos e, + '1]1 cos f)t

T = 21)2 COS Hi

"72 cos Of + 'fJJ cOS BI

Again, for the normally incident case, these results reduce to those of Section 1.7.

For this polarization no Brewster angle exists where r = O. as we can see by examining the numerator of (1.143a),


and using Snell's. Jaw to give

k~(ry~ - 'fIi) = (k~T)~ - ki11i)sin2 OJ.

But this leads to a contradiction, since the term in parentheses on the right-hand side is identically zero for dielectric media. Thus, no Brewster angle exists for perpendicular polarization, for dielectric media.


Plot the renection coefficients for parallel anel perpenaicuiar polarized plane waves incident from free-space onto a dielectric region with lOr = 2.55, versus.

incidence angle,


The wave impedances are

1]1 = 3770.

1/0 377

'TJ2 = - = -- = 236n .

.;e; vU5

1.8 Oblique Incidence at a Dielectric Interlace







10 20

~o 50 60 70 80 90

Incidence angle 61

FIGURE 1.14 Reflection coefficient magnitude for parallel and perpendicular polarizations of a plane wave obliquely incident on a dielectric half-space.

We then evaluate ( I _J 37a) and ( I .143a) \Tr~US i ncidence angle: the results are shown in Figure 1.L4. 0

Total Reflection and Surface Waves Snell's law of (1.136.b) can be rewritten as

. 0 /f;1. II

Sin "/ = -Sillv,.



Now consider the Case (for either parallel or perpendicular polarization), where "I > f:2. As (J, increases from 0" to 90°. the refraction angle ()t will increase frnrn OC' to lJO°. but at a faster rate than (J'i increases. The incidence angle e, for which Bj = 90<> i~ called the critical angle, B<" thus

Fi sin e~ = V;;-'


At this angle and beyond, the incident wave will be totally reflected, as [he transmitted wave is not propagating into region 2. Let u look at this situation more closely. for the case of (), > Be with parallel polarization.

When e'i > Be (l.I44) shows that sin Of:> 1, so that cos et = VI ~ sin] (}1 must be imaginary, and so the angle fit loses its physical signi ficance At this point, it is better



Chapter 1: Electromagnetic Theory

to replace the expressions for the transmitted fields. in region 2 with the following:

- (jQ. f3.) .'" .._

E = EaT -. x - -.2 e.-j""'e-o ..

-t kq k2 '

Ht = EoT fje-iI'j"'e-[]lz.


The form of these fields are derived from (1.134) after noting that -jk2 sin (}t is still imaginary for sin Bt> L but -jk".!. cos (}t is real, so we can replace sin 8t by (3/k", and cos (jt by ja:! kz. Substituting (1.146b) into the Helmholtz wave equation for fI gives




Matching E:r; and By of (1.146) with the~· and fi Components of the incident and reflected fields of (1.l32) and (U33) at z = 0 gives


...!:..e-jkt:rsinO; _ £_e- jk,x sin s; = T e-i{i'·.

~ ijl ~2


To obtain phase matching at the z = 0 boundary, we must have

which leads again to Snell's law fur reflection. B; = Br> and to {3 = kl sin e,. Then D' is determined from (L147) as


which is seen to be a positive real number. The reflection and transmission coefficients can be obtained from (1.148) as

(ja/k"l)7]2 - Til cos B. r = (jn! k2)172 + 'II cos @i·

T = 2111 case. .. (jU/k2)T/2 + 7]1 cos ei

Since r IS of the form (a - jb)!(a, + jb), its magnitude is unity, indicating that all incident power is reflected,

The rransm itted fields of (\.146) show propagation in the x direction, along the interface, but exponential decay in the z direction. Such a field is known as a sUiface wave. '" since iL is tightly bound to the interface, A surface wave is an example of a



" Some authers !lTgue ilia! theterm "surface wave" should not be used for. ficld of tbis t)!P~. since it exists only when plane wave fie) ds C'xist in rh e z < 0 regie D. and so preie r 10 call it a "surface wa ve-tike" fieltl, or a· "forced s urface W:IVe."


1.9 Some Useful Theorems

nonuniform plant: wave. so called because it has an amplitndc variation in lim z direction. apart from the prop'l.gation factor in the .1" direction,

Finally. it is of interest to calculate the complex Pcynting vector tor (he surface wave fields of (1.l46):

C> _ E::: . H-' _ IEnI2ITI~ (~jll + .... !!__) -~n::

'N - t ,l< t - - I .1 I_ ( •

'II ~ .. ~ If1


This shows that no real power flow occurs in the : direction. The real power flm\ in the x direction is that of" the surface wave field. and decays exponentially with di!-.lanr.:e into region 2. So even though no real power i.., transmitted into region 1. a nonzero field does exist there. ill order to satisfy the boundary conditioru- at the interface.


Finally, we discuss everal theorems in elcClTtll11agnetics that we wiLl find useful for later discussions.

The Reciprocity Theorem

Reciprocity is a general concept tha: occurs in many ureas of physicx and engineering, and the reader may already be familiar with [he reciprocity theorem of circuir theory. Here we will derive the Lorentz reciprocity theorem Ior elecrromagnetic fields in LwO different forms. This theorem will be used later in the book to obtain general properties of network matrices representing microwave ci rcuits lind to eval uate the coupling of waveguides from current probes and loops, ami the coupling of waveguides through aperture .. There are a number of other important uses of thi~ powerful concept.

Consider the two separate sets of sources, .II' ,\11 and .l~ .. Ch which generate the fields £1. HI. and respectively. In the volume '\' enclosed hy U1C closed surface S, as shown in Figure 1.15, Maxwell's equations are satisfied individually for these two


,\1 j '"!;_l '!_I f.·2· 1'2


Geometry for (lie Lorentz reciprocity theorem.



46 Chapter 1: Electromagnellc Theory

sets of sources and fields, so we can write

\l X £:1 = -jWJLHl - MI. \l X i'lJ = jWf.E1 + 31,

\1 x 1h = -jw!lfl2 - M2;, \l x ih = jWf.E2 + J2·

I.152a 1.152b



Now consider the the quantity \l. (E\ x .th. -Jh x fIll. which can be expanded using vector identity (B.B) to give

Integrating over the volume V, and applying the divergence theorem (B.15), gives

f \7. {El X H2 - ih X fld dy = i (El X H2 - E2 X HI) . ds l.l55

h h

= r (~. J1 -.it1 . J2 + HI . M2 - fh . Ml)dv Jv

Equation (1.155) represents a general form of the reciprocity theorem, but in practice a number of special situations often occur leading to some simplification. We will consider three cases.

S encloses no sources. Then J1 = 12 = IIi!) = All = 0, and the fields El J HI, and E~I H2 are source-free fields. In this case, the right-hand side of (1.155) vanishes with the result that

Thi result will be used in Chapter 4, when demonstrating the symmetry of the impedance matrix for a reciprocal microwave network .

.'3 bounds a perfect conductor. For example. S' may be the inner surface of a closed, perfectly conducting cavity. Tben the surface integral of (1. 155) vanishes, since E] x HJ ·11 = (11, X EJ) . B2 (by vector identity B.3), and it x n; is zero on the surface of a perfect conductor (similarly for Eh.). The result is


This result is analogous to the reciprocity theorem of circuit theory. In words, this result states that the system response EJ or E2 is not changed when the source and observation points are interchanged. That is, Ez (caused by J2) at 1J is the same as EI (caused by JJ) at J2.


, . 9 Some Usetu\ Theorems

S is a sphere at infinity. In this case, the fields evaluated on S are very far from the sources and so can bl': considered locally as plafle Wilves. Then the impecilllu;e relation iI =-n x E/17 applies (0 (L155) to give

(£1 x fh - E. x HJ) • fl. = (fi. X £1)' B1 - [n, x £}) . tt,

J - - I ~ .

= -HI .H2 - -rl~·HJ =0.

1] 1"/

so thai. \he re<;u It of (1. \ 57) is ag'11u ohtained. This result Ul.!:\ also be ()b\ai.n~J h,w the case of a dosed surface S where the surface impedance boundary condition applies.

Image Theory

In many problems a current SOUTCe is located in the vicinity of a conducting ground -plane. lmageLhemy pennia; the renlClval of the ground plane by placing a virtual image source of {he other side of the ground plane. The reader should be familiar with Ibis concept from electrostatics. so we will prove the result for an infini re current sheet next to an infinite ground plane and then summarize the. nther possible cases,

Consider the surface curre nt dens i ty i: = J Btl:?· parallel to aground pi ane, 9..~ shown ill Figure I. J 6a. Because the current source is of infinite extent and is uniform in the z , y directions, it will excite plane waves traveling outward from it. TIle negatively

(JruUlJd __ -r Plane •

i_--S("lUrLe t J, = '!sf).1


~ .... __ Image
-d 0
\bl t ......... Suurce

ti, r


Illustration of image theory as applied to an electric "'(lITe)].1 .'(lurce next to a ground plane. raj An electric surface current density parallel 10 a ground plane. {bj The gmnnd plaoc of (a) noplac~{j witb image <'.Unent at;; = -d.



Chapter 1: Electromagnetic Theory

forO < z < d,

traveling wave will reflect from the ground plane at z = 0, and then travel in the- positive direction. Thus, there will be a standing wave field in the region 0 < z: < d and a positively traveling wave for z > d. The forms of the fields in these two regions can thus be wrltten as

E! = A(ejl<;~:; - e-jk{)",), H<J = -A (ejk",,, + e-jkO.4),

lJ 'lJQ


for 0 < z < d,


for z > d,


for z > d,


where T}o is the wave impedance of free-space. Note (nat the standing wave fields of (U58) have been constructed to satisfy the boundary condition that Eo; = 0 at z = o. The remaining boundary conditions to satisfy are continuity of E at z = d, and the discontinuity in the I! field at It = d due to the current sheet. From (1.36),since Ms=O,

.J~ = z x f)(li;; - H~)I;;=()· Using (1.158) and (1.159) then gives

2jA sin kod = Be-jk;,d, B -jktJd 2A

J80 = --e - - cos 1t'fJd,

'170 170


.1:- Jl; =0,

while fr-om (1.37) we have


which can be solved for A and B:

A = -J",OTJoe-Jkod,

2 .

B = -jJIJo1Josinkod.

So the total fields are

E~ = - j Jso1loe - j k;,d. sin I4Jz.,. H; = .lsoe-ik{)d CQSk.()z,

Et = -jJsoiJO sin kode-jk,.z, H: = -jJ",usinkodg-.1l.:oz.



for 0 < s < d, for 0 < z < d,

I..I61a 1.1610

for z > d, for z > d.

1.162a 1.1620

Now consider the application of image theory to this problem, As shown ill Figure 1.16b, the ground plane is removed and an image source of - J" is placed at J* = -d.

1 ,9 Some Useful Theorems

By superposition, the total fields for .:::. > 0 can be round by combining the 6eJd~ from the two sources individually. These fields can be deli ved by a procedure similar to that above, with the following results:

Fields due [0 source at z = d:

{ -J,,()t/(I -jl'~i;;-dl ---e


E:r = -

-.l ,(\1/(1 r;Ji.:n{Z- (I)


{ -:Nlle ikjl~-li)

fly =

J;D t IA"lz-dl



for ~ < d.

for z > r1

J. I 63b

for z < rl.

Fields due to source at z = =d:

for z > -d


for z <:: -(t,

for 2: > rt

l.! 6411

for ~ < -d.

The reader can. verify that the sol u lion is Iden ti cal to thai of ( 1.161) for 0 < ;: < ti, and 10 (1.162) for z > d. thus verifying: the validity of the image theory solution. Note rhat image theory on Iy gives the correct fields to the right of the conducting plane. Pi gnre 1.1 7 shows more general image theory results for electric and magnetic dipoles.

The Uniqueness Theorem

Once we haw round a solution to Maxwell's equations and the appropriate boundary conditions, the uniqueness theorem assures us that. under rhe proper conditions, this sQlution 15 the only possible solution. This is a particularly useful result when, as in Chapter 4. we can find The fields due to' a source by postulati ng the form of the fields. and then enforce boundary conditions by adjusting some amplitude constants. The uniqueness theorem then guarantees that this procedure gives the correct and unique solution.

Although il can be expressed in various ways. we wiU state the uniqueness theorem in the follow i ng form: I n a region bounded by a c! osed surface .9 and completely f Jl ed with dissipative media. the tiekl £, fI is uniquely determined by the source currents in the region and the tangential components of E or H on S.

This result can be proved by assuming I WQ solutions 10 Maxwell's equations, EU, tr- . . and EO. Bo, and showing that they must be identical. Thus, if E;Q, H" and E/'. iflJ satisfy



Chapter 1; Electromagnetic Theb.ry

Original Gwmetry

Image Eqwvalel'll



~ ! t




I i I

--- 1----












FIGURE 1.17 Electric and magnetic cerrenr images. (ll) An electric current parallel to II ground plane. (b) An electric current normal to a ground plane. (c) A magnetic current parallelto a ground plane. (d) A magnetic current normal [Q a ground plane.

Maxwell's equations in S, then the difference fields, E"" - ED, f:Jf1_ fib, must also satisfy Maxwell's equations inside S. Furthermore, these difference fields must be source-free fields, as substitution into Maxwell's equations (1.27a,b) readily shows. Then Poynting's theorem of (1.88) (with 0" = 0) gives for these difference fields the following reslJit

Now, if E and/or II is prescribed on the surface S in any of the following ways, or in any combination of these conditions, then the firstintegral in (1.165) will vanish:

L it x E =0, electric walls. 2. n x H = 0; magnetic walls,


3. ft x E = Et, a fixed tangential electric field.

4. f1 x Ff = n.. a fixed tangential magnetic field.

Now assume a small loss in the medium. so that E = e' - »" and /1 = ti - j/1" are complex. Theil the real part of ( 1. L65) becomes

r (l/'IH'" - Hj'12 +E"IEQ - Ebll)d~l = O. .Iv


Since all of these terms are nonnegative. the equation can be satisfied only if 13(1 = _Eo and fIn = 1)1', which shows that only one solution is possible.

Note thai it was necessary to introduce loss to achieve this result: loss can always be introduced into a problem by making f and/or /.1 complex. nus result aha suggests that problems involving lossless materials may not have unique solutions. This is indeed the case as the following examples point out,

Resonant modes i11 a lossless cavity. Such source-free fields are defined as having equal time-average electric and magnetic stored energies. as given by (1.84) and (1.86). If there is no loss. p, anti f are real. and the second integral in (1.165) is identically zero for a difference field equal [0 the field of such a resonant moue. Such source-free fields would rapidly dissipate in the presence of loss. however.

Plane wave incident 011 a lossless dielectric slab. In Problem 1.7 we will solve the problem of plane wave transmission and reflection from a dielectric slab of finite thickness. This solution will satisfy Maxwell's equation. and the boundary conditions, but it may not be the only solution. This is because a surface wave field can also exist all the dielectric slab. This fi eld is source- free, and also satisfies Max well' s equations and the boundary crmdirions. For a lossy slab, such a field would quickly be dissipated.

Plane \vaves in free-space. til infinite lossless free-space. source-free plane wave solutions are possible for any polarization and any direction of propagation. If loss is introduced, such field would quickly decay to zero.


[IJ 1. C. Maxwell. A Treatise Oil Electricity and M'agJleti m, Dover, N.Y .. 1954.

[2] A. A. Oliner, "Historical Perspectives on Microwave Field Theory:' IEEE Trans. Microwave Theory UIId Techniques, vol. MlT-32 , pp. 1022--l045, September 1984 (LIm special issue contains other articles on the history of microw ave engineering).

[3] S. V. Marshall and G. G. Skitek, Electromagnetic Crmcepts and Applications, Third Edition, Prentice-HalJ, N.J .. L990.



Chapter 1: Electromagnetic Theory

141 c. A. Balanis. Advanced Engineering Electromagnetics, 101m Wiley & Sons. N.Y .• 1989.

[51 R. E. Collin. Foundations for Microwave Engineering, SecOJ1d Edition. McGraw-Hill. . Y., 1992. [6] D. K. Cheng. Field and Wave Electromagnetics, Second Edition. Addison-Wesley, Reading, Mass .•


l?] S. Ramo. T, R. Whinnery, and T. 'Ian Duzer, Fields {wd Waves in CmrmumicatiQrl Electronics, Third Edition. John Wiley & Sons, N.Y., 1994.

[8] C. G. Montgomery. R H. Dicke, and E. M. Purcell, Principles of Microwave Circuits. vol, 8 of MIT Rad. Lab. Series, McGraw-Hill, N.Y .. 1948.

[9J R.F. Hanington, Time-Harmonic Electromagnetic Fields, McGraw-liill, N.Y., 1961.


1.1 Assume thai an infinite sheet of electric surface current density J. = J"ff; Aim is placed on the z es 0 plane between free-space for z < O. and a dielectric with ( = (:rEO for z > 0, as shown below. Find the resulting E and fJ fields in the two regions, HJN1': A sume plane wave solutions propagating away from the current sheet, and match boundary conditions LO find the amplitudes, as in Example 1.3.

J =i!. AIm

., Il


L2 Let E = Epp+ E",$+ Ezz be an electric field vector in cylindrical coordinates. Demonstrate that it is incorrect to interpret the expression V]. E in cylindrical coordinates as pVz Ep +$V2 E¢+zV2 E; by evaluating both sides of the vector identity 'V x V x E = V (V . E) - V' E for the given electric field.

1.3 An anisotropic material bas a tensor "permittivity [E] as given below, and a permeability of 4p,().

At a certain point in the material. the electric field is known to be E = 3i + 4fj + 62. What is D at this point?

-2.i 0] 3 a

a 4

1.4 Consider a permanent magnet with a steady magnetic field II = Hoy, and a parallel plate capacitor with an electric field E = Eijx, arranged as shown on the next page. Calculate the Poynting vector ata point between both the magnet poles and the capacitor plates. TIlls nonzero result seems 10 imply real power flow in the z direction. but dearly there is no wave propagation or power deJivemd from the sources. Bow do you explain this apparent paradox?



LS Show that a linearly polarized plane wave of the form E = Ro(f I·· 2y)e-;kpr can be represented as the sum of art RHCP and an LHCP wave.

1.6 Compute the Poynting vector for the general plane wave field of ( 1.76).

1.7 A plane wave is normally incident Oil a dielectric slab of permittivity f,. and thickness d. where d = An/(4y7;). and ,\u is the free-space wavelength of the incident wave. as shown below. If free-space ex ists on both sides. of the slab. find the reflection coefficient of the wave reflected From the front of the slab.




1.8 Consider an RHCP plane wave normally incident lmrn tree- space (z < 0) On103 half-space Ii:: > 0) consisting nf a good conductor. Let the incident electric field be of the fOJ1Il

E = f:n(:f: - .if)lr,-Jk,."

and find the electric and magnetic fields in the region z > O. Compute {he Poynting vectors for Z < 0 and z > 0, and show that complex power is conserved, What is (he polarization of [he reflecred wave?

1.9 Consider a plane wave propagating in a los-sy dielectric medium for z < 0, with <1 perfectly conducting plate at il = U. Assume thai the lossy medium is characterized by f = (5 - j2)fQ. !J, = 1111. and ilia! the frequency of the plane wave is j.O GH;[, and let the amplitude of the incident electric field be -+ "'lim at z = U. Find the relkneci electric lield fur = < 0, <'Inti plOl th~ mllgnirude of the total electric field for -0.5 S z ::::: ().


54 Chapter 1: Electromagnetic Theory

1 .. 10 A plane wave in free-space is normally incident on a thin copper sheet of thickness t. \Vbat is Jhe approximate required thickness :if the copper sheet ·is to be used asa shield to reduce tim level of the transmitted elecrric fleld by 90 dB? Do this calculatioa for f = 1 MHz, I GHz, and 100 GJiz. HINT: Simplify the problem by ignoring reflections at the interfaces.

1.11 A uniform lossy medium with fr = 3.0, tan 8 = 0.1, and J1 = po tillsthe region between it = 0 arid z = 20 em, with a ground plane at z = 20 cm, as shown below. An incident plane wave with an electri cfiel d,

Et = :i:1OOe-T" V 1m,

is present at z = 0 and propagates in the +2: direction. TIle frequency is f = 3.0 GHz.

la) Compute Pi, the power density Qfthe ineident wave, and. P,., the power density ofthe reflected wave, at s = O.

(b) Compute the input power density, FIn, 301 Z = 0, &Gm the total fields at Z = O. Does Pin = Pi ~ P,.?


1=20cm z

1,12 Redo Problem. 1.1, bUl with an electric surface current density of J. = JI).ie-iP" Aim, where {J < Iva.

1.13 A parallel polarized plane wave is obliquely incident from free-space onto a magnetic material with permlttlvity EO and permeability P.Opr. Find the reflection and transmission coefficients. Does. a Brewster angle exist for thiscase, where the reflection coefficient vanishes fora particular angle of ineidence?

1.14 Repeat Problem 1.13 for the perpendicularly polarized case. 1.1,5 Consider the gyrotropic permittivity tensor shown below:

The lJ and E fields are relaredss

Show that the transformations

E+ = E", ~ jE~,. E_ =E"+jE,,,

D+ = D", ~jD~, D_ = Do; + jD""


allow the relation between E and n to be written as

where [F' J is now a diagonal matrix. What arc the elements ("Ii" If']'! Using this result, derive wave equations for E... and E_. and lind the resulting propagation constants.

L 16 Show that the reciprocity theorem ex pressed in (I. 157) also applies to a region enclosed hy a closed surface S. where a surface impedance boundary ondition applies.

L17 Consider an electric surface current density of .1. = {j.Joc-ch A/Ill. located 011 the ;: = d plane. If a perfectly conducting ground plane is placed at z = 0, use image theory to find the total fields for z > O.



Transmission Line Theory

In many ways transmission line theory bridges the gap between field analysis and basic circuit theory, and. so is of significant importance in microwave network analysis .. As we will see, the phenomenon of wave propagation on transmission lines can be approached froman extension of CIfCllit theory Dr from a. specialization of Maxwell's equations: we shall present both viewpoints and show how this wave propagation is described by equations very similar [0 those used in Chapter 1 for plane wave propagation.



The key difference between circuit theory and transmission line theory is electrical size. Circuit analysis assumes that the physical dimensions of a network are much smaller than the electrical wavelength, while transmission lines may be a considerable fraction of a wavelength, or many wavelengths, in size, Thus a transmission line is a distributedparameter network, where voltages and currents can vary in magnitude and phase over its length.

As shown in Figure 2.)a, a. transmission line is often schematically represented as a two-wire line, since transmission lines (for TEM wave propagation) always nave at least two conductors. The short piece of line of length I::!.z of Figure 2.1 a can be modeled as a lumped-element circuit, as shown in Figure, where R, L, G, C are per unit length quantities defined as follows:

R = series resistance per unit length, for both canducl,ors, in f2/ ill. L = series inductance per unit length, 'for both conductors. in Him. e = shunt conductance per unit length, in Slm.

e = shunt capacitance per unit length, in Ffm.

The series inductance L representsthe rota} self-inductanee of the two. conductors. and the shunt capacitance C is due to the close proximity of the two conductors. The series resistance R represents. the resistance due to the finite conductivity of the conduc-

56 tors,and theshunt conductance G is due to dielectric loss in the marerial between. !be

2.1 The Lumped-Element Circuit Model for a Tranernlsslon Une

I ( •• I) ---

+ .-I~. r)

------ .1.::-----+-


;(~, I)


"(~ + ;;, rJ --


C.: ,,(,:+£":;, r)



FIGURE 2.1 Voltage and current definitions and equivalent circuit far an incremenral length of transmis ion line. (a) Voltage and current definitions. (b) Lumped-element equi vale nt circui t.

conductors. R and C. therefore, represent loss. A finite length of transmission line can be viewed as a cascade of sections of the form of Figure 2.1 b.

From the circuit of Figure 2.~b, Kirchhoff's voltage law can be applied to give


and Kirchhoff's current law leads to

. . 81J(z + c.z. t) .

1,(Z.t)-G~,c7'(1+D,~.tJ-C~z -1(.Z+.1Z.t) =0.


Dividing (2.l a) and (2. I b) by .6..:: 3J1d taking the Ii mi t as .6. z -..;. 0 gives the following differential equations:

fJv(z, t) '. 8i(z, t)

()z = - Rdz. t) - L------a;-.

8i(z. t) = _P •• ( t) Cov(z. t)

az u-vr;, at'

These equations are the time-domain fonn of the transmission line, Or telegrapher, equations.

POT the sinusoidal steady-state condition. with cosine-based phasors, (2.2) simplify to




dz = -(R + j..;L}J(z).

dIe;;;) .,..

~ = -(G + .1wC)VCz).




58 Chapter 2: Transmission Une Theory

Note the similarity in the form of (2.3) and Maxwell's curl equations of (lAla) and (1.41b).

Wave Propagation on a Transmission Line

The two equations of (2.3) can be solved simultaneously to give wave equations for V(z) and I(z):


d2V(z) 2

dz1 . - '7 V(.z) = 0,

,p I(z) 2

~ - 'Y J(z) = 0,

,= a+ j(3 = j(R+ jwL)(G + jwC)

is the complex propagation constant, which is a function of frequency. Traveling wave solutions to (2.4) can be found as

V(z) = V/e-rz + v;,-e,('2, I(z) = r: e.-1'z + I;; e ",

where the e-")'Z term represents wave propagation in the +z direction, and the elz term represents wave propagation in the -3 direction. Applying (2.3a) to the voltage of (2.6a) gives the current on the line:

Comparison with (2.6b) shows that a characteristic impedance, ZQ, can be defined as

Zo = R+jwL == 'Y

to relate the voltage and current on the line as


10 l,,-

Then (2.6b) can be rewritten in the following form:

v+ v-

fez) = ~e-'F - _o_e-Y;z.

z, z,

Converting back to the time domain, the voltage waveform can be expressed as

v(z t) = IV,'+I cos(w,t - !3z + ¢J+)e-nz + lVo-1 cos(wt + fJz + rjJ -)eO"",









2.2 Field Analysis of Transmission Lines

where 1jJ± is the phase angle of the complex voltage Vo±' Using arguments similar to. those in Section 1.4. we find that the wavelength on the line is

>. = 271"



and the phase velocity is

2:1 I

The Lossless Line

The above solution was for a general transmission line. including loss effects, and it was seen thai the propagation constant and characteristic impedance were complex. In many practical cases, however, the loss of the line is very small and so can be neglected, resulting in a simplification of The above results. Setting R = G = 0 in (2.5) gives the propagation constant as


l' = a + jl3 = jwVLG. (3 =wVLC,

2.12a 2.12b


As expected for the lossless case, the attenuation constant a is zero. The characteristic impedance of (2.7) reduces to



which is now a real number. The general solutions for voltage and current ()I1 a lossless transmission. line can then be written as

V(z) = V,,+e-f1h + 1io-eJ.B-", 1ft .' \f-

I(z) = ....£_e-Jf3~ _ ....2.....ejP4.

Zo Zo



The wavelength is


and the phase velocity is

w 1

vI) = 7i = .v'YC'




In this section we will rederive the time-harmonic form of the telegrapher's equations, starting with Maxwell's equations. We will begin by deriving the transmission line



Chapter 2: Transmission Line Theory

parameters (R. L, G, C) in terms of the electric and magnetic fields of the transmission line and then derive the telegrapher equations using these parameters for the specific case of a coaxial line.

Transmission Line Parameters

Consider aIm section of a uniform transmission line with fields E and fl, as shown in Figure 2.2, where S is the cross-sectional surface area of the line. Let the voltage between the conductors be 'i1oe±Hh and the current be l"e:i;ipz. The time-average stored magnetic energy for this 1 m section of line tan be written, from (1.86), as

M J - -,.

W,." = - H·H ds,

4 s

and circuit theory gives Wm. = Lll.,12/4, in terms of the current on the line. We can thus identify the self-inductance per unit length as

L = l~j2 is fI . fr ds Him.


Similarly, the time-average stored electric energy per unit length can be found from (1.84) as

w: = ~_ f E E* ds

" 4 1s '

and circuit theory gives We = elY" 12/4, resulting in (he. following expression for the capacitance per unit length:

C = 1\~12 is E· E* de F/m.

From (1.130), the power loss per unit length due to (he finite condactivityof the metallic conductors is




FIGURE 2.2 Field lines on ail arbitrary TEM transmission line.

22 Field Analysis of Transmission Lines

(assuming Jj is tangential to S). and circuit theory gives P,. = RIJpI2j2. so the series resistance R per unit length of line is

R" i --

R = -, -I' TJ· u til. n/1l1.

III - . n-tc:


In (2.19). R., = 1/ ab; is the surface resistance of the conductors, and C1 + ('2 represent intcgratiOJl paths over the conductor boundaries. From (1.92). the lillie-average power djssipated per unit length in a lossy dielectric is

wi' ~ - _

P" = - £. Et ds.

2 ."

where (' is the lmaginarj part of the complex dielectric constant ( = e' - .Jt'li = .' (l - j tan b). Circuit thenry gives I~, == GJ 1 .l' /2, so the shun I conductance pe r U nil length call be written ax

.oc" /' - _

G = - ,_ 12 t,' . E' tis S/m.

11" . s

EXAMPLE 2.1 Transmissiun Line Parameters of a Coaxial Line

The fields of a traveling TEM wave inside the coaxial line shown in Figure 2.3 can be expressed as


t: = I' :'f, t-r~. p [lUj(L

,- To//; _,~

II z. --I" , '27fP

where -{ is the propagation constant of the Line. The conductor are assumed [0 have a surface rcsisxivily I?". and the material tiHtng the space hO't\veen


Gcometry of a coaxial line with surface resistance R; un the inner and outer conductors.



Chapter 2: Transmission Une Theory

the conductors is assumed to have a complex permittivity t: = f'_' - jf" and a permeability J1. = fMjJ.Lr. Determine the transmission line parameters.


From (2.17)-(:2.20) and the above fields the parameters of the coaxial line can be calculated as

It 12." 1" 1 . po

L =- .. - 2Pdpd¢ =-10 b/u Rjm,

(211")2 ¢=¢ p=a P 211"

f' fir r 1 hE'

C = (In bla)2 )<1>=0 }p=a p2Pdpdq; = In bla F/m,

R = ~ {121O _!__adrb + 12,.- ]_bd4>} = Rs (2. +!_) n/m

(21£-)" ¢=o a2 • q>=o lP- 27f a b

WE" r r 1 21TW€"

G = (In bja)2 }",=o }p=a p2Pdpd<p = In bla S/m. 0

Table 2.1 summarizes the parameters for coaxial, two-wire, and parallel plate lines.

As we will see in the next chapter, the propagation constant, characteristic impedance. and attenuation of most transmission lines are derived directly from a field theory solution; the approach here of first finding the equivalent circuit parameters (L, C, R G) is useful only for relatively simple lines. Nevertheless, it provides a helpful intuitive concept and relates a line to its equivalent circuit model.

TABLE 2.1 Transmission LItI.e Parameters for Some Common Lines
@ tG r----"'-j
D tl
lG __i
L p.. b ~ wstc1( D) ud
2r. a .". 2a IV
C 2.".€' -rr:er E'W
Lnbla cosh-I(D/2a) d
R s, (1 I) R, 2N,
27T ;- + b .".0 W
2'1J"WE11 "lTWElt w~nw
G Lnbla cosh-1{D12a) d 2.25

2 .. 2 Field Analysis- of Transmission Unes 63

The Telegrapher Equations Derived from Field Analysis of a Coaxia.lline

We now show lhat the telegrapher equations of (1.3). derived using circuit theory, can also be obtained from Maxwell's equations. We will consider the specific geometry of the coaxial line of Figure 2.3. Although we will treat T6M wave propagation more generally in the next chapter .. the present discussion should provide some insight into the relationship of circuit and field quantities,

A TEM wave on the coaxial line of Figure 2.3 will be charactedzed by Ez = Hz = 0; furthermore. due to azimuthal symmetry. the fields will have no Q)-variation. and so 8j8rp = O. The fields inside the coaxial tine will satisfy Maxwell's curl equations.

\l x E = -.i",_'pf[, \l X H = jwr:E,


where t = r/ - Jr" may be complex to allow for a lossy dielectric filling. Conductor loss will he ignored here. A rigorous field analysis of conductor loss can be carried out, but at this point would tend to obscure our purpose: the interested reader is referred to Ramo, Winnery. and Van Duzer [I] or Stratton [Z].

Expanding (2.2Ia) and (2.'21 h) then gives the following vector equations:



Since. the 2 components of these two equations must vanish. it is seen that E<f, and Hm must have the forms

ErJ) = f(2). p

s; = g(z:). p



To satisfy the boundary condition that EifJ = 0 at p = a. b, we must have E,~ = 0 everywhere, due to the form of Ed! in (2.23a). Then from the p component of l2.22a), it is- seen that Hp = O. With these results, (2.22) can be reduced to



'" 'E

-a = -JWf. p'



From the form of H¢ in (2.2:lb) and (2.24a). Ep must be of the form

64 Chapter 2: Transmission Une Theory

Using (2.23b) and (2.25) in (2.24) gives 811(;;)

----az = -jWJ.1,FJ{z),.

8g(;;) . ch ) -a:;- = - JW (z-.



Now the voltage between the two conductors can be evaluated as

lb .. t d . b

V(z) = Ep(p, a) dp = h(z) _E = he;;) In -,

p==n (l=a P a


and the total current 0]] the inner conductor at p = .(t can be evaluated using (2.23b) as


1(z) = H ¢(a, ;;)a drb = 271"g(z).


Then h{z) and g(z) can be eliminated from (2.26) by using (2.27}to give

81l(z) = _jwj.t in b/a l(z) ,

OZ 21f

81(:;;.) __ . (I _. II) 2:nV(z) f)z - Jwr;: Jf lnb/a'


Finally, using the results for L, G, and C for a coaxial line as derived above, we obtain the telegrapher equations as

av{z) .

---a;;- = -JwLI(;t),

81(z) .

~ = -(G + JWC)V(z)



(excluding R, the series resistance, since the conductors were assumed to have perfect conductivity). A similaranalysis can be carried out for other simple transmission lines.

Propagation ccnstant, Impedance, and Power Flow for the Lossless Coaxia.1 Line

Equations (2.24a) and (2.24b) for Ep and H¢ can be simultaneously solved to yield a wave equation for Ep (or H",):


from which it is seen that the propagation constant is 71 = -W2I1f, which, for Iossless media. reduces to



2.3 Ths Terminated Lossless Transmission Une

where the last result is from (2.1 :2), Observe that this. propagation constant is or the same form as {hat for plane waves in a lossless dielectric medium. This is it general result for TEM transrrussion tine)'.

The wave impedance is defined as Z" = l."i/H<J;. which can be calculated from (:2.24a) assuming an e- IfYZ dependence 10 give

This wave impedance is then seen In be identical to the intrinsic impedance of the medium, q, and again is it general result for TE1\1 transmissionIinns.

The characteristic impedance of the coaxial line is defined as

"" I/~ E,J 1:1'1 (Jill Zo = - = -----'.---

1" };r H..,

1) In b/o ,I I-L In I!/o

:!'" = V-;:-~'

w here the forms for i.'1-' and H'i' from Exa tuple ~.I ha ve been used, The characteristic impedance is g,eomelry tle.pendem and win b~ J.ifkr-en\ fQI- otnel' uam;mis"ion line con figurations.

Finally, the power fiow (in lbe ::; direction I Oil the coaxial line may he computed from the Poynting vector .IS

a result that is in clear agreement with circuit theory. This shows thai the flow of power in a transmission line takes place entirely via the electric and magnetic ~oJds between the LWO cnnduutors; power is not transmitted thrnugh rile conductors themselves. As we will see later, for rhe case of finite conductivity. power may enter the conductors, but this power i::; then lOSL as hca; and is n01 delivered to the load,


Figure 2.4 shows a lossl e55 trans mission h nc terminated in an arbi I rar)' J (lad Impedance ZL. This problem will illustrate wave reflection 011 transmission lines. a fundamental property of distributed systems.

Assume That an inc ideru wave of the fonn 1/-+ (] .i.n is zerie ra ted from a source at

(, ~

Z < 0, We have seen thai the ratio of voltage 10 CUlTCnL for such a traveling Wave IS

VI~I, /1;:1 --.. !

---------___,~ + I

_____ Z_~I_,8_' \_I~~



r = V",- = Z [. - Zo

V,,+ ZL +Zo

A current reflection coefficient, giving the normalized amplitude of the reflected current wave, can also be defined. But because such a current reflection coefficient is just the negative of the voltage reflection coefficient (as seen from (2.34)), we will avoid confusion by using Dilly the voltage reflection coefficient in this book.

The total voltage and current waves on the line can then. be wrilten as



Chapter 2: Transmission Line Theory

Zoo the characteristic impedance. But when the Line is terminated in an arbitrary load ZL #- Zo, the ratio of voltage to current at the load mu t be 21,. Thus, a reflected wave must be excited with the appropriate amplitude to satisfy this condition. The total voltage on the line can mel] be written as in (2. J 4a), as a sum of incident and reflected waves:

V(z) = V,,+e-j/3,,- + v:,-eJ{1,z.

Similarly, the total current on the line is described by (2.14b):

l(z) = v;,+ e-j{h. _ V,,- ejth

Zo Zo

The total voltage and current at the load are related by the load impedance, so at z = 0



we must have

Solving for V",- give.

v.- = Z L - Zo v+

o ZL + Zo a '

The amplitude of the reflected voltage wave normalized to the amplitude of the incident voltage wave is known as the voltage reflection coefficient, I':

V(z) = V/ [e-itJz + reii1z] I I(z) = V,,+ [e-j(h -reiJ}2J.




From these equations it is seen that tbe voltage and current on the line consist of a superposition of an incident and reflected wave; such waves are called standing waves. Only when r = 0 is there no reflected wave. To obtain I' = 0, the load impedance ZL must be equal to the characteristic impedance Zo of the transmission line, as seen from (2.35). Such a load is then said to be marched to the line, since there is no reflection of the incident wave.

Now consider the time-average power How along the line at the point z:

2.3 The Terminated Lossless Transmission Una

where (2.36) bas been used, The middle two terms in the brackets are of the form A - A~ = 2jlm(A) and SQ are purely imaginary. This simplifies the result to

P. = !1\!~:-12 (I -Ifl:'.)

~\ 2 Zn '

which shows that the average 'power flow is constant at any point on tile line, and chill the total power deuvered tome load tp';l\") is equal to tnc inciuem power (1'i·iJ+I·')2Zo1. minus the reflected power (1IIcf·1f I! 1 2Z,l!. If r = O. max irnum power is deli vered tu the load, while no power is delivered for If 1 = 1. The above discussion assumes that the generator is matched, so that there is no reretlecuon of the [efletted wave from z < o.

When the load is mismatched. then, nol all of the available power from the generator is delivered to the load. TIllS "loss" is culled return foss (RL), <HId is defined (in dB) as


RL = -20 log 11"1 dB:


so that a matched load (I' = 0) has a return lo~s of 'x' dB (no reflected power). whereas a total reflection (I I'I = I) 11<1<' a ret lim loss of 0 dB (aIL inc ideru power i ~ ref ected).

ff the load is matched EO the line. f,=O and the magnitude of the voltage on the line is ,"Vez)] = IV;;+-I, which is a constam. Such a line Is sometimes said LO be "nUL·· When the load is mismatched. however, the presence of a reflected wave leads to standing waves where the magnitude. or the voltage on [he line is not constant. Thus, from (2.36a),

IF(z)1 = II';;,+III +fe2.ui21 = 1\'",+111 + fe 2J311 = IV,:-Ill + Irl(,J("-~·~'iI.


whered = - z is the posi ti ve distance measured from the load at ;;: = O. and f)' i s the phase. of the reflection coefficient If = IfI,.,(I). This result shows that the voltage magnitude oscijlates with position "[along the line, The maximum value OCCI,t!"!; when rile phase term ej(fI-1BO = 1. and is given i:ly


the minimum val ue occu rs when the phase term ejl (I 1/111 = - l. and is gi \ en by


As [T 1 increases. the rari 0 of 1t,n ... x to 1. ~in mcreases, so a measure or the misrn arch of a line, called the standing Wal' e ratio (S\VR), can be defined us

SVv'R = \ ;,,~, 1 ~l1in

1 +Irl

I - If .


This quantity is also known as the raftage standing wave ratio. sometimes identified as VSWR. From (2.41.1 it is seen that S\VR is a real number such (hat 1 :5 S"WR -::;: oc.. where SWR = I implies a matched load .

. From (2,39), it is seen that the d istance between two successive voltage max fill a (or minima) is f = 2x 12ff = 7r A/2 Tf = )../2, whi 1 e the di s ranee between El. max i mum and a minimum is t' = '7i" /28 = )../4. where ;\ is the wavelength on the transmission line,



Chapter 2: Transmission Line Theory

The reflection coefficient of (235) was defined as the ratio of the reflected [0 the incident voltage wave amplitudes at Ihe load ce = 0), but this quantity can be geueralized to any point f all the line as follows. Prom (2.34a), with 'l = -€, the ratio of the reflected component to the incident component is

V- -jtJl

r(C - "e - r 0 -2jf3t

) - _L. '~l - (. )e ,

VQ eJ,-


where [(0) is the reflection coefficient at z = 0" as given by l2.35). This form is useful when transforming the effect of a load mismatch down the line.

We have seen that the real power flow on the tine is a constant but that the voltage amplitude, at least for a mismatched line. is oscillatory with position on the line .. The perceptive reader may therefore have concluded that the impedance seen looking inro the. line mUSL vary with position, and this is indeed the ease. At a distance f = -z from the load, the input impedance seen looking toward the load is

Z = V( -() = \~; [e=itu + re-jp~] Z = I + re-li.m z In u-r, Vo+ [eJP€ _ re-jj3I'].o 1- re-1.j,sr. 0,

where (2.36a,b) have been used for V(z) and I(z). A more usable form may be obtained by using (2.35) for I' in (2..43):


(Zr. + Zo)ej(:Jc + (Zr; - Zo)e-JBe z., = ZO(Zr. +- Z()ej~f - {ZL - Zo)e.-jjJf

_ Z ZL cos(3e + jZo sin (3£

- 0 ZIJ cos (J.e + jZr. sin (3f

ZJ." + j211 tan i3P =Zo---=-----

Zo + JZL tan (3J!.


This is an important result giving the input impedance of a length of transmission line with an arbitrary load impedance. We will refer (0 this result as the transmission line impedance equation; some special cases will be considered next,

Special Cases of Lossless Terminated Lines

A number of special cases of lossless terminated transmission Jines will frequently appear in our work, so it is appropriate to consider the properties of such cases here.

Consider first tbe transmission line circuit shown. in Figure 2.5, where a line is terminated in a short circuit, ZL = O. From (2 . .35) it is seen that the reflection coefficient for il short circuit load is T = -1; iL then follows from (2.41) thai tile standing wave ratio is infinite. From (2.36) the voltage and current on the line are

11(2) = V,,+ [e-jih - ellh] = -2jV,,+ sin {3z ..



2.3 The Terminated Lossless Transmission Line

FIGURE 2.5 A transmission line terminated 111 a short circuit.

which shows that V = 0 at the load (as it should. for a short circuit), while tbecurren[ is a maximum there. From (2.44). Of the ratio 1f( t)/J(-t), (he inpur impedance is

z, = j Zo tan .u:


which is Seen LO be purely imaginary for any length, f. and to lake all all values between +joo and =jix». For example, when f = 0 we have Zm = Q. but for f. = Aj4 we have Zifl = 00 (open circuit). Equation 12.45c} also shows thai the impedance is periodic in t, repeating. for multiples of ,\/2. The voltage. current. and input reactance for the short-circuited line are plotted in Figure 2_6.

Next consi derrhe open-circuited line shown in Figu re 2.7. where Z 1.. = oc. Di viding toe numerator and denominator of (2.35) by LL and allowing 21.. ~ QC shows that the reflection coefficlent r or this case is r = I, and the standing wa ve ratio is again i nfi nj te. From (2.36) the voltage and current on the line are


whicb shows [hat now 1 = 0 at the load, as expected for an open circuit, while the voltage is a maximum. Theinput impe-dance is

Zin = - jZ\I cor 3i.


which is also purely imaginary for any length, t, The voltage, current. and input reactance of the open-circuited line are plotted in Figure 2.8.

Now consider terminated transmission lines with some speclal leugths. If r = ,\/2, (2_44) shows [hal


mean:mg [hat a half-wavelength line (or any multiple of -"/2) does not alter or transform the load impedance, regardless of the characteristic impedance.

If the line is a quarter-wavelength long or. more generally, f = )../4 f- n)..12, for n = 1,2 .. 3 ..... (2.44) shows that the input impedance is given by

Z'" = Z11.





Chapter 2: Transmission Line Theory




FIGURE 2.6 (a) Voltage, (b) current, and (c) impedance (Rio = 0 or 00) variation along a short-circuited tlWlSmiSSiOll line.

Such a line is known as a quarter-wave transformer because it has the effect of transforming the load impedance, in an inverse manner, depending on the characteristic impedance of me line. We will study this case more thoroughly in Section 2.5.

Now consider a transmission line of characteristic impedance Zo feeding a line of different characteristic impedance, Zh as shown in Figure 2.9, If the load line is



FIGURE 2.7 A transmission line terminated in an open circuit.

2.3 The Terrnlnated Lossless Transmission Line



/1;:)2(, -2jV,,-



FIGURE 2.8 (a) Voltage. (b) CUJreIlL and Ie) impedance (R'n = (I or ::>c,l vuriauon along an open-circuited transmission tine.

infinitely long. or if it is terminated in its own characteristic impedance. su that there are no reflections from its end. then the input impedance seen by the feed line i ZI. so that the reflection coefficient r L~

r = Zl - Z~.

ZI +Z()

Not all of the incident wave is reflected; ollie. of it is transmiued onto the second line witb a voltage amplitude given hy a transrnission coefficient. T.

From (2.36aJ. rhe voltage Cor ~ < 0 i.


;:; <0.


where 11/ is the amplitude of the incident voltage wave On the feed line. The voltage wave for z > U, ill the absence ofreflectious. it> outgoing only. and can be written as

\r~z) = \ ;,1 T'r-J'<" .

for: > u.




Ghapter'~: Transmission Line Theo,ry


FIGURE 2.9 Reflection and transmission at the junction of two transmission lines with different characteristic impedances.

Equating these voltages at z = 0 gives the transmission coefficient, T, as

Zl - Zo 2Z1

T=l+r=}+' ... =---

Zl + Zo Zl + Zo'


The transmission coefficient between two points in a circuit is often expressed in dB as the insertion loss, I L,

IL = -20 log 11'1 dB.


POINT OF INtEREST: Decibels and Nepers

Often the ratio. of two power levels, PI and P2, in a microwave system is expressed in decibels ( ) as


10 tog A dB .

Thus, a power ratio of 2 is equivalentto 3 dB, while a power ratio of O. l is equivalent 10 - lOdE. Using power ratios in dB makes it easy to calculate power loss Or gain through a series of components, since multiplicative loss Of gain factors can be accounted for by adding the loss or gain in dB for each stage. For example, a signal passing through a 6 dB artenuator followed by a 23 dB amplifier will have an overall gain of 23 - 6 = 17 dB.

Decibels are used only to represent power ratios, but if PI = v? / RI and P2 = 1122/ EW., then the result in terms of voltage ratios is

10 log 1Ii: Rl = 20 log VI" {R;. dB,

ViRI . 1:1 V Ii;

where RI, R2 are the load resistances and VI, V1 are the voltages appearing across these loads. fi the load resistances are equal, then this formula simplifies to




The ratio of voltages across equal load resistances can also be expressed in terms of nepers (Np) as


2.4 The Smith Chart

The corresponding expression in terms of powers is

1 PI N

"2 In p"! I p.

since voltage L~ proportional to the square root of power. Transmission line attenuation is often e.x.pressed in nepers, Since 1 Np corresponds [0 a power ratio of e", [he conversion between nepers and decibels is

INp = 10 log f"!. = 8.686 dB.

Absolute. powers can also be expressed in decibel notation if a reference power level IS assumed. If we let p:! = 1 mW, then the power PI can he expressed in dflrnas


10 log --' - dBm.


Thus a power of 1 mW is 0 dBm. while a power nf I W is 30 dBm, etc.


The Smith chart. shown in Figure 2. J 0 is a graphical aid that is very useful when solving transmission line problems. Although there are a number of other impedance and reflection coefficient charts that can be used for such problems [3J, the Smith chart is probably the best known and most ... videly used. It was deveJoped in ]CJ39 by P. Smith at the Bell Telephone Laboratories [4]. The reader may feel that. in this day of scientific calculators and powerful computers, graphical solutions have no place in modern engineering. TIle Smith chart. however. is more chan just a graphical technique. Besides being all integral pan. of much of the current computer-aided design (CAD) software and test equipment for microwave design. the Smith chart provides an extremely useful way of visualizing transmission line phenomenon and so is also important for pedagogical reasons. A microwave engineer can develop intuition about transmission line and impedance-matching problems by learning to think in terms of [he Smith chart,

At first glance the Smith chart may seem intimidating, but the key to its understanding is to realize thai it is essentially 11 polar plot of the voltage reflection coefficient. T. Let the reflection coefficient be expressed in magnitude and phase (polar) form as F = If'leJti. Then the magnitude III is plotted as a radius nI'l S I) from the center of the chart. and the angle (}(-180" ~ B $ !SOb) is measured from {he right-hand :;idc of (be horizontal diameter. Any passively realizable (Ir! S 1) reflection coefficient can then be plotted as a unique point on the Smith chart.

The real utility of the Smith chart, however, lies in the fact that it can be used to convert from reflection coefficients to normalized impedances (or admittances), and vice versa, using the impedance (or admittance) circles printed on the chart. When dealing with impedances on a Smith chart normalized quantitites are generally used, which we will denote hy lowercase letters. The normalization constant IS usually the characleri<;tic impedance of the line. Thus, z = Z / Zu represents the normalized version of the im pedance Z_



Chapter 2: Transmission Une Theory

FIGURE 2.10 The Smith chart.

Tf a lossless line of characteristic impedance Z(I is terminated with a load impedance ZL-, the-reflection coefficient at the load can be written from (2.35) as

ZL - I .

r = --. = [rleJIl, ZL + j

where ZL = ZL/ZO is the normalized load impedance. This relation can be-solved for ZL in terms of r to give (or, from (2.43) with e = 0)

1 + Irle1fJ Zl. = 1 _ Irlel""

This complex equation can be reduced to two real equations by writing rand 2;L in terms of their real and imaginary parts. Let r = r,. + .W;, and ZL = J'L + jXL. Then,

. (l + r r) + jr i

T[. + ):tL = (1 _ r,..) - jr,'



2-4 The Smith Chart

The-real and imaginary parts of this equation can be found by multiplying me numerator and denominator by the complex conjugate of the denominator to give

1- r2 - r!

r I



Rearranging (2.55) gives

( T)2 (1)2

r L_ ..L['2 = _

r 1+ rt. " 1+ rL


~ ( 1)2 (l)~

(rr- 1)- + I', - -. = -,.

.'1L :r./,


which are seen to represent two families of circles in the r r , T, plane. Res i stance e ire I es are defined by (2.568), and reactance circles are defined by (2.S6b). For e X:J;IJ1P le, the t: L = 1 circle has its center at T r = 0.5, T, = O. and has a radius of 0.5, and so passes through the center of the Smith chart. All of the resistance circles of (2.56u) have centers on the horizontal T', = 0 axis. and pass through the r = 1 point on the right-hand side of the chart. The centers of an of the reactance circles of 2.56b) lie on the vertical r r = I line (off the chart), and these circles also pass through the r = 1 poim. The re isiancc and reactance circles are orthogonal.

The Smith chart can also be used to graphically solve the transmission line impedance equation of (2.44). since this can be written in terms of the generalized reflection coefficient as


where I' is the reflection coefficient at the load. and P is the (positive) length of transmission line. We then see that (2.57) is of the same form as (2.54). differing only by the phase angles of the r terms. Thus. if we have plotted the reflection coefficient Iqr./q at the load, the normalized input impedance seen looking into a length f. of transmission line terminated with ZL can he found by rotating the point clockwise an amount 2,3£ (subtracting 2f3f! from OJ around the center of the chan. The radius stays the same, since the magnitude of r does no! change with position along the line.

To facilitate such rotations. the Smith chart has scales around Its periphery c-alibrated in electrical wavelengths, toward and away from the "generator" (which just means the d:irectiO[l away from the load), These scales are relative, so only the difference in wavelengths between two points on the Smith chart is meaningful. The scales cover a range of 0 to 0.5 wavelengths, which reflects the fact that the Smith chart automatically includes the periodicity of transmission line phenomenon. Thus, a line of length ),,/2 (or any multiple) requires a rotation of 2(i/! = 2" around the center of the chart. bringing the point back to its Original position, showing that tile input impedance of a load seen through a ),,/2 line is unchanged,



Chapter 2: Transmission Line Theory

We will now illustrate the use of the Smith chart for a variety of typical ttansmission line problems through examples.

EXAMPLE 2.2 Basic Smith Chart Operations

A load impedance of 130+ j90 n terminates a 50 0 transmission line that is 0.3), long. Find the reflection coefficient at the load, the reflection coefficient at the input to the line, the input impedance, the SWR on the line, and die return loss.


The normalized load impedance is

ZL . :

ZL = Z = 2.60+.71.80,


which can be plotted On the Smith chart as shown in Figure 2.1 L. Using a compass and the voltage coefficient scale below the chart, the reflection coefficient magnitude at the load can be read ft.') WI = 0.60. This same compass setting can then be applied to the standing wave ratio (SWR) scale to read SWR = 3.98, and to the return. loss (in dB) scale to read RL = 4.4 dB. Now draw a radial line through the: load impedance point, and read the angle of the reflection coefficient al the load from the outer scale of the chart as 21.80.

We now draw an SWR circle through the load impedance point. Reading the reference positi on of the load on the wavelengths-toward-generator (WTG) scale gives. a value of 0.220>.. Moving down the line 0.3), toward the generator brings us to 0.520). on the \VTG scale, which is equivalent to 0.020>.. Drawing a radial line at this position gives the normalized input impedance at the intersection with SWR circle of Zin = 0.255 + jO.117. Then the input impedance of the line is

z., = ZOZin = 12.7 + j5.8 O.

The reflection coefficient at the load still has a magnitude of Irl = 0.60; the phase is read from the radial line at the input and the phase scale as 165.8". 0

The Combined Impedance-Admittance Smith Chart

The Smith chan can be used for normalized admittance in the same waythat it is used for normalized impedances, and it can be used toconvert between impedance and admittance. The latter technique is based on the fact that, in normalized fonn, the input impedance of a load ZL connected to a )./4 line is, from (2.44),

Zin = I/ZL.

which has the effect of converting a normalized impedance to a normalized admittance.

Since a complete revolution around the Smith chart corresponds to a length of )../2, a ),/4 transformation is equivalent to rotating the chart by J 80"; this is also eq.uiv.aJent to

2_4 The Smith Chart

F1GURE 2.11 Smith chart for Example :U_.

imaging a given impedance (or udrninancej point across the center of the chart to obtain the corresporuling admittance (Or impedance) point.

Thus, the arne Smith chart can he uxcd I·Of both Impedance ant! admittance calculations during the solution of a given prchlern. AI di fferenr stages of the solution. [hell. the chart may be either an impedance Smul: dum or an udusiuunce Sm;lh chnrt. This procedure can be made less confusing by usl ng a Smith chart rhat has a superposition of the scales for H regular Smith chart and the scales 0]' a Smith chan which has been rotated 180". as shown in Figure :2.12. Such a chan is referred 111 as an impedance ani] udmituno:« Smith. chart and usual Iy has different-colored scules for .impedance and admittance,

~ EXAMPLE 2.3 Smith Chart Operutious lYsing Admittances

iCI A load of Z, - 100 - j50 n termmatcs a 50 [l line. Whal are the load admittance iU1U the input admittance i r the line is 0.15), long?



Chapter 2: Transmission Une Theory


The normalized load impedance is 2.L = 1 + j 1. A standard Smith chart can be used for this problem by initially considering tt a an impedance chart and plotting ZL and the SWR circle. Conversion to admittance can be accomplished with a ),,/4 rotation of ZL (easily obtained by drawing a straight line through si: and the center of the chart to intersect the SWR circle). The chart can now be considered as an admittance chart, and the input admittance can be found by rotating 0.15..\ from YL.

Alternatively, we can use the combined Z1} chart of Figure 1.12, where conversion between impedance and admittance is accomplished merely by reading the appropriate scales. Plotting XL on the impedances 'cales and reading the admittance scales at this same point give lIL = 0.40 - jO.20. The actual load

FIGURE 2.12 ZY Smith chart with solurion for Example 2.3.

2.4 The Smith Chart

admittance is then

Yr = ,1/LYn = ~~ = 0.OU80 - jO.0040 S.

Then. on the WTG scale, the load admittance is seen to have a reference position of 0.214/\. Moving O. L5..\ past till point brings us to U.364..\, A radial line at this poirn on the WTG scale Intersects the SWR circle at an admittance of y = 0.6 I + jO. 66. He acru aJ input udrn i trance is then 1· = O. () 1?:2 + )0.0132 S. ()

The Slotted Line

A slotted line is a transmission line ccufiguration (usualJy wa veguide or coax) thai allows the sampling of the electric field amplitude ora standing wave Oil a terminated Line, With this device the SWR and the distance of the first voltage minimum from the load can be measured, and from this data the load impedance can be determined, Note that because the load impedance is in general a complex number (Will] two degrees of freedom), two distinct quantities mUSL be measured with the <Iotted line to uniquely determine this impedance. A typical waveguide sloucd line i. shown ill rigure 2.13.

Although [be slotted line used to be the principal way of measuring an unknown impedance at ruicrowave frequencies. it has been largely superseded by the modem vector

~ A .. IlX-band waveguide sloued llne.

Courtesv <It' Hcwletl·.Pad:a«! C(lmp'~!l'f. S:mto. R{),\'.. Ctl!l.f.



Chapter 2: Tranamtssfon Line Theory

network analyzer in terms of accuracy, versatility, and convenience. The slotted line is still of some use, however, in certain applications such as high-millimeter wave frequencies, or where i! is desired to avoid connector mismatches by connecting the unknown load directly to the slotted line, thus the use of imperfect transitions. Another reason for studying the slotted line is that il provides an unexcelled tool for learning basic concepts of standing waves and mismatched transmission lines. We will derive expressions for finding [he unknown Load impedance from slotted line measurements and also show how the Smith chart can be used for the same purpose .

. Assume that, for a certain terminated line, we have measured the SWR on the line and l:'mJn, the distance From [he load to the fiTSL VOltage minimum on the line. TIle load impedance ZL can then be determined as follows. From (2.40 the magnitude of the reflection coefficient on the line is found from the standing wave ratio as

rrl- SWR-J - -SW-· -R-+-] .

From Section 2.3, weknow that a voltage minimum occurs when ej({i-2:Jl) = -1, where B is the phase angle of the reflection coefficient, r = rrlejll. The phase of the reflection coefficient is then

where emin is the distance from the load to the first voltage minimum. Actually, since the voltage minimums repeatevery ),/2, where )., is the wavelength on the line, any multiple of ),/2 can be added to emi:n without changing the result in (2.59), because this just amounts to adding 2Bn;\/2 = 21T1l. to (j, which will not change r. Thus. the two quanrities SWR and Rrttltt canbe used to find the complex reflection coefficient I' at the load" It is then straightforward to use (2.43) with e = 0 to find the load impedance from T:

i-r r ZL = Zo---.


The use of the Smith chartia solving Ibis problem is best illustrated by an example.

EXAMPLE 2.4 Impedance Measurement with a SJo!tw Line

The follo,:,",ing two-step pIOced~re has been carried out with a 50 n slotted line to determine an unknown load impedance:

1. A short circuit is placed at the load plane, resulting in a standing wave on the line-with infinite SWR, and sharply defined voltage minima, as shown in Figure 2. L4a .. On the arbitrarily positioned scale 0[1 the slotted line, voltage minima are recorded at

z = 02 em, 2.2 em , 4.2 em.




2.4 The Smith Chart

... 11'1


Shari circuit



11!11, III t(!:!!)(\!!!!l

nil .3 .t 7\

l'nknown load


FIG RE 2.14 Voltage standing wave patterns lor Example. 2,4. (a) Standing wave for short-circuit load. (b) Standing wave For unknown load.

2. The short circuit lS removed, and replaced with the unknown load. The standing wave ratio is measured a, SWR = 1.5, and voltage minima. which an>: not as sharply defined as those in step I, are recorded at

~ = n.n cm.2.72 cm.-t.72 em.

as shown in Figure 2.14b. Find the load impedance, Solution

Knowing that vohage rmrurrra repeat every 1\12. we have from the data of step 1 above that A = 4.0 em. in addition. because the reflection coefficient and input impedance also repeat every ,\/2. we cao consider the toad terminals to be effectively located at any of the voltage minima locations listed in step L Thus, if we say the load is at 4.2 ern. then the data from step 1 shows that the next voltage minimum away from the load occurs at 2.71 em, giving {mm = 4.2 - 2.72 = 1.48 em = 0.37".

Applying (2.58)-(2.60) to this data gives


1.5 - I

WI = 1.5 + 1 = 0.1.

0= iT + 4i1 (1.481 - 86.r. -1-_0

r = n.2e-,1Ui-l° = 0.0126 + jO.1996.

The load impedance f~ then

(I +r\

Zi. =50 ~r J = ..J.7.3 +}19.7 n- 1- ;



Chapter 2: Transmission Una Theory

For the Smith chart version of the solution, we begin by d:rawing the SWR circle for SWR = 1_ 5, as shown in Figure 2.15; the unknow n normalized load impedance must lie on this circle. The reference tbat we have is that the load is 0.37.-\ away from the first voltage minimum, On the Smith chart, the position of a voltage minimum corresponds 10 the minimum impedance point (minimum voltage. maximum current), which is the horizontal axis (zero reactance) to [he left of the ori gin. Thus, we begi n at the voltage minimum point and move 0.37 ), toward the load (counterclockwise), to the normalized load impedance point, st: = 0.95 + jO.4, as shown in Figure 2.15. The acmal load impedance is then Z L =: 47.5 + j20 n. in close agreement with the above result using the eq uations.

Note that, in principle, voltage maxima locations could be used as well as voltage minima positions, but that voltage minima are more sharply defined than voltage maxima, and so usually result in greater accuracy. 0

li'IGURE 2.15 Smithchart for Example 2.4.