USEFUL INFORMATION

T .. hl,,1 Phv<;",1 Crmstantc

.1'1. c, ... ~I ,,_
-0
Velocity of light in vacuum c 2.9979 X 108 m/sec
Permittivity of vacuum EO 8.8542 X 10-12 Farad/m
Permeability of vacuum !-Lo 47T X 10-7 Henry/m
Electron charge e 1.60219 X 10-19 C
Electron mass me 9.1095 X 10-31 kg
Planck's constant h 6.6262 X 10-34 J-sec
n. .~, .~ '" 27
. JU 'H~"" P ,.v,~v~ 'V "6
Hnhr r-arljn c n. o <;?l<!A

Avogaoro s numoer iVA O.U'::':: ».
n ~o In 23 r nr
,.~v, 'v ,,~

Table 2 Energy Conversion Factors
leV = 1.602 X 10-19 J
= 2.42 X 1014 Hz
~ 07 "X 103 (,ITI-1

_,
"VUJ\. . ".J7 / lV ,,. :"
TV




Table 3 The Electromagnetic Spectrum
Typical Wavelength Frequency Photon Energy
(em) (Hz) (eV)
AM radio 3 X 104 106 4 X 10-9
FM radio 3 X 102 108 4 X 10-7
Ka ar j 10 4 X 10

1111 I<1l "U " IV IV v ...
Vi<ihl" (,"X 5 <; 1014 '")

travio et x j X .L
V" ~ " 8 ,,,18 uaco.
-,~. 'v .vvv



L I I'- 1'1\, "II L I'- 1'1\, I '" I \L
.I. \...1 ~.I.....'. ~ .Y.I..I...I. ~ .l..I~.I...I1...J \,,-'.1..

OPTOELECTRONICS









CLIFFORD R. POLLOCK
School of Electrical Engineering
Cornell University





















r ..
--------
rot. . D D D, roo
~V&V'" ~v".vu ~u"uv" L 1U",,, '-'u,,"'''''
London • Madrid . Mexico City • Sydney • Toronto

""""''--C1.rUUJ U. '''''Y'' "', .,"'--., 177,J

flU . '5"" ,c,,,, 'cu .. 'v p,u, v, u", may o c
<torpli in " retr'ieva svstem or
1lI allY lUIIll V1 uY allY rueaus , , ,
photocopying, recording, or otherwise, without the prior
written permission of the publisher.
Publisher: Tom Casson
Editorial assistant: Carrie Berkshire
Editorial coordinator: Christine Wright
Prrsier-t editor- Jlit" Mc-Mnllpn
Prodtrctton manager. Lara Feinberg

n,,<i<Jnpr' S"nln.<
v.over uesigller: lim ",a"ge
Art Tn~n~n"'r' T<iTn
0
"'''Ul''1 , UI\ ...
Typeface: 1011 2 Times Roman
Printer: R R Oonnelley & Sons Company
Library of Congress Cataloging-in-Publication Data
Pollock, C. R. (Clifford R)
Fundamentals of optoelectronics / Clifford R Pollock.
n pm
Hl'-'UU'" lHU,,"A.
TSRN 0- )';(i- 04-1
L 1. 1111".
TA 7';fP(i4 qq,;
V"",1..)OI ~'+- l')')'~

Printed in the United States of America
234567890001098765 UUl l (\.1 l

nbers started a revolution m the commurucations mdustrv about n .vears.azo,

Fibers iouay all: UI>I:U HUl oiuv lUI Inrrg=distarrce reiecommu ica ion u or con-

~;~,., ~;n,.,~lc t~ th", ,.,"';<Th lc ;,., th", ";t;,,, , ~,.,rl ."

-0= , 'O~'~U

in medium- to large-size offices. As demand for information technology increases, the demand for large bandwidth communication links will increase, ensuring that optical fibers will be increasingly used in neighborhoods, homes, offices, and even

instrumentation and automobiles. . .

There are some major technological problems, and indecisions, inhibiting the

npyt ot"oP of fihpr_ontir' _1 For PY"rnnlp "ho111i! infnrrn"tion hp "pnt

-0 -c -c •

usma time IVlsIOn ormat mnarv DU ses ase mo u anon communi-

cam , or waveienz OUUIaI 0 m dtvtston mu iPlexing re uires a eavv

II_' . L • --'-

tive filters and tunable sources which can be economically produced. Coherent detection requires accessible frequency standards in the infrared region. None of these technologies is impossible, but presently none of them are cheap to implement, either. Until significant hardware breakthroughs occur, it is unlikely that we will see fibers connected to the home or office.

This book is directed at the hardware aspects of optoelectronics. Six major top-

ICS are coverec: n tamental nrmcmres iacnenc theorv: ontical

. . . . ~ . .

waveguiues; .j) coupung ana numencai anaiysis or waveguiues: 4) _j!_e_lec_lors ana

r"tinn ",i· hp h""pi! on n"""ivp u,,,,,ponii!p rOl1nlpi! ",ith np", tTl"tpri"lc

c -0 --c

and structures. The text, pitched at the senior/first-year-graduate course level, requires only a basic familiarity with electromagnetic waves, and the ability to solve differential equations with boundary conditions.

Part 1 of the text explores the basis for optical propagation and establishes the use of the MKS system, discussing the wave equation and the properties of materials ~uch as attenuation and disnerxion Somp of thp material i~ redundant with

w al SLUe en s s ouio ave seen as nmtors u i as een mv ex enence t at verv

~ " .. .

ICW MUUClltl> i ceu ry Wall\. away WIll. a WVIl\.lll~ l\.llVWICUgC VI ivraxwc. I> cljual 0 II>

waveguides, then systematically advance to more complicated structures, such as

v

vi

Preface

,1. A .._, .1 ••

0' va, " .. " .... a. va , "uu • vv"Ve~UIUC~.

the review of the WKB approximation in the graded-index chapters. These sections can be skipped if the reader is willing to take certain results on faith later in the book.

From the beginning, this text introduces numerical techniques for studying nonanalytic structures. Most chapters have numerical problems designed for solution using a computational program such as Matlab or Mathematica. Any course taught rrom trus nook ~holl h~vp ~rrp~~ tn ~ ~nrf i~tp <:of """" Tn

ran _j or me OOOK an entire chapter is devoted to one ot the numenc sImulation

, , , ,

.1. I u"" bei II~ useu HI UIlI l UIU~ UeM~l1 \ rue UCi:UH nel 00 ,

',1. c, , .1 .1

"U~ 1" ~ • ,~ ... ~ ~Pl' -.urrro:y -ror 'HUU"UL" LV ... J\.l'.V." "VIU" UVY"l Vl'U""1 ~

too ..... ""h .. f'f'"rt h,,,, .. t .. , .. rl t" ,n"111rl,,, c e-rt o l n' .. r» .. c "f' "nA.. ._~ ~ .

1"" 'Pt'

priate to get a student started on the numeric work. As I am not a numeric type,

the coding is often crude, so don't be afraid to explore better or simpler numeric methods. Mathematica was used on a personal computer for most numeric problems in this book. There is nothing exclusive about Mathematica, as many students in my courses have used Matlab, Maple, and even programmable calculators to solve thp rrrrihlern

rsecause a rna or Issue 10 0 ica svs em esizn IS noise, t-art 4 conra ns two

cuapter ~ l

ano noise. v.na ne IL 1 ro uces De une amentai noise con-

Part 5 of the text covers the generation of light in terms of stimulated and spontaneous emission, followed by a description of optical amplifiers, lasers, and semiconductor devices. A thorough description of these devices could easily fill several volumes. In this text the material is discussed at a general level, with the goal of establishing a fundamental understanding of first principles.

Part 6 of the text discusses optoelectronic devices such as modulators and ~en-

in signincance astrrevrrranrre;

r._ • • • .1 r-

IVI " "VUU, ... UCVVU;U lV 1'"""IVO;:; WaV' , UIO;:; 111"l 11 's WUUIU

"J "0· ,

Chapter 4, on graded slab waveguides, and portions of Chapter 6 on graded index circular fibers, could be skipped without suffering in later chapters. The coverage on the rectangular waveguide (Chapter 8) is especially useful for students interested in designing semiconductor lasers; otherwise it could be skipped without too much pain. Similarly, Chapter 9 on the beam propagation method is self-

contained and rOlllrl h", d,innprl ~<: rlp<:irprl A nv <:prioll" rli<:rll <:"ion of ",,"nl,no ",.

e evices w' re uire COUDle mo e eorv so l a ers u an are stronzlv

1

detection processes. It nas oeen my expenence that most students do not have a

Preface vii



...I .+ .J., .+ .. 1 ...I' ,I • ...I,
o~~~ ~_ .. ~_ ~~ .. ~.~_, ~~ ... _~_ "'__'J .. ~_~ ....... ~ -""1''_'~ .~ ...
., I",r-tnr", r-nnr"", . 11" l'nr dnrl",nt" ,,,hn ",1.,n tn r-nnt;nn", ;ntn th", """t"' ...... c ·;rI",
, . , r JJ
of optical commurucation.
Optical devices such as lasers and amplifiers are discussed in the next three
chapters. Because of all the activity in new optical devices, such as high-speed
lasers, quantum wires, surface-emitting lasers, strained layers, etc., it was decided
to discuss the fundamental principles of lasers and semiconductor devices in these
chapters.
Advancer omcs are not eoverec as IS could easuv m anotner tex: Hased on
me rnaterral in v.napters 1+-10, a stucent COUla reac acvanceu papers in mis new.

1 .L· .c .L _L
"" l""'<Ul"'l llll"llL ""IL <V", <Ill" pVl UVIl Vl Ul" "'VUl "" VY lUl "U"11 1
'T' ;11. rI. .J., rI. J., .J.,
... _ ... ~ •• ,~ _""1' ,_. ~ u ... ~ .... ._ ~~ ... _ ~_, .-_~ ..... , _ .... ~_ ..... ~_ ~ .. ~_~ ~ ..... _
lrnn\ll] pli OP '.Ii in thp nrpvinm: I h Thp." ~ ni('P. ~nn]i('Minn
~ . . . , . ..
of the knowledge, and form a motivating summary to any course in this subject.
I would like to acknowledge the help of several graduate students on this mate-
rial. Alphan Sennaroglu was especially helpful in reviewing mathematical deriva-
tions. Duane Barber and Tim Carrig both provided real data for parts of the text,
and Martin Jaspan was instrumental in helping me with computer problems, and
-" full crerlit for develonino the Mntlab in Chanter q T would like
. .
0 anx om I.. asson ano om UCKer 0 rwin or a e eeuoacx an all vice an
~. . 'U
IVI !-,ULllll)o!, UI-' W IllJ Iy Utoli:IY1> allU WVllYlll~, rIlli:1l1Y, 1 WallL LV UlaI11\. Illy
:+, 1\.1',,1: e. :.J., .J.,'. .++, +, .J., +< rI +, ,11 ~
.. ~_, H~_"~~", .~. 1' ....... 0 .. 1' ......... ~ _~~~ •• ~~ .... _ 1'''~ .' J- ... ~, .... ~ .~ ....... _ •
'''nrlr nn th;" t"'vt U'h'l", T h.,,,,,, tr;",rI tn ...... ., .", "nr'" ' HT ,n th", t",vt ;"
'J '0
or properly referenced, I am sure there are mistakes and oversights. I apologize III
advance for these,
Clifford Pollock



1 'I \ I\,. I I' I\,. . I ..... I I\,. UIJII I
\. ,\.J , .1 .I. "II I l..') I , Ut.'.1.L_j1


('~fA D'i 'in) Q
0 .. 0 .... ~ 'A lI.U AD TY 'Ir
&~n& & DI:<'
T'T T ... T ......... 11. KT"'''' T...... T .... . '-'...., LJ' lL&'<o. ~ ... .... .o.'<o."'''_'
.1 Ul~.LIrtlV.lJ...jl~.l ~ T W~~ IP"': 'J I';;


CHAPTER 1

THE FUNDAMENTAL TOOLS PART 3
OF OPTOELECTRONICS: COUPLING AND NUMERIC
MAXWELL'S EQUATIONS 11 ANALYSIS 241

r ADTVD"
YAK 1 ~
'T'T rT ~ ... K nnr'!.nA"" A ...... Tr'!..,..T
TH h (JPTll :AI -wTRF. 4T .I. .I.-rx:.- ~ . .1.'\...1.1."1
l\,ft:1TUr\,"'\ t:1r\D A 10.:. A V7 ·1\.Tf" (""\'D'T
~. ~ ~ &&-
-wLi\.' 1·.(';'1 II ... \,; 'JLl'~
CHAPTER 2
THE PLANAR SLAB CHAPTER 10
WAVEGUIDE 49 COUPLED MODE THEORY AND
CHAPTER 3 APPLICATION 271
DISPERSION IN WAVEGUIDES 75
'" lAYTEK II
OTTPTTl'IITTRFTW hEN ()PTI( AT
l:HAI'TEK 4
.... ~~y-~-~ -y- ~~y.~.- ~
-<I< A 1 nq J-I N 1 JI-'. X :\Ul Kl ~ A'll J YVAV nl JI"' :'\ LYY

YVA.V t'A.JUIUb;:) TIIT

CHAPTERS PART 4
STEP-INDEX CIRCULAR NOISE AND DETECTION 325
WAVEGUIDES 125

CHAPTER 6 CHAPTER 12
nI"':0t:10 ~Tr\1\.T A lI.T1"'\ no A 1"'\t:11"'\_ I1\.Tnt:1V
.~ Mr\~Thl D"II A I
... H"'_I<~ I'" '-'& ..... .._" lL~
IPTP( 'T IR"': ~ 7

rllAPTI<'.R 7
ATT1:'lI. ·A'l"' rr~"" A .,..-~ lI. AD (,HAPTER 11
'lL& ... & ... U ...." lL ..... '-'& "T'>TI'L"" ' ... H '.L.J.L" , ...... lL.L,"
EFl'bCT~ IN WAVEGUIDES 175 UPTICAI.:'TIETECTORS 357
Ix Contents in Brief

yAKI !)

Arr~I'I L1\ A

H.J' Ir' A .... ' 1« . ........~X")

CHAPTER 14

OPTICAL RADIATION AND AMPLIFICATION 387

APPENDIXB

BESSEL FUNCTIONS 559

'" LAy 1 ~K I!)

"' .... _EN x r

IP I c8. AIVIt' .Lr I"',~ I-\I"IIIJ

(l~ u' I\. T D()'X' ~i~ 'I\,;T7 (ll A

CHAPTER 16

SEMICONDUCTOR LASERS 453

PART 6

Cl1Arl~

WA\. t<,l1'1

'1"', Ml II" i\' II( ~ LlI,,\

CHAPTER 18

FIBER-OPTIC SENSORS 533





, " \ I\,. I L I\,. I ."
, ..... '-.J ,. .I. .I. ,. .I. U



JJ"I.L · .... nnTT( TTnN A._Nn l.WL 'Iota Internal Isino

n, ,u'U i\x ~ Kay racing .)'"
- ~~ ~ wave J,lescrtDtlOn ot 'iotai Internal
1.11
A D_;aJ' ;p<. ,_" ~.f 'T1 ;n ;~ 'l ~ n n
OJ ,)0
Development of the Optical Waveguide 5 1.12 Phase Shift upon Reflection 39
Types of Optical Communication Systems 7 1.13 Summary 42
Opportunities in Optoelectronics 7
Suggested Readings and References 8
PART 2
H t<, ut" II ,AI W. K.t'. 4

eftAI I
r~ _c,
C 1"1 JAIVI l"lA o
(:HAPTI<:K 2
THE PLANAR Sf.A R
CHAPTER 1 WAVEGUIDE 49
THE FUNDAMENTAL TOOLS OF 2.1 Introduction 49
OPTOELECTRONICS: MAXWELL'S 2.2 The Infinite Slab Waveguide 50
EQUATIONS 11 2.3 Electromagnetic Analysis of the
1 1 T. -' • 'T'I- rt" __ 1 _C d. 'T'. -" 11 r anar vvave2:ui e JU
b~ • ~u~ n V~ u", u ... u~ ~~
.... • E '" ... _ ine _1 onar udmal Wavevector: J:i 52
L~ "L~I£" L.4
" ~ ~ '1') J::<.l2:enva IIP!'< tor the Siah irlp "4
i.J ~UllMllUU V~ 1'+
~ 1 ne wave cqua!j_on 10 ?n OnticalMode Confinement .si

1.:) :SolUtIOnS to me wave bquatIOn 1~ 2.7 The Symmetric Waveguide 61
1.6 Transverse Electromagnetic Waves and the 2.8 Intuitive Picture of the Mode 64
Poynting Vector 21 2.8.1 Structure of the Mode 64
1.7 Phase Velocity 22 2.8.2 Why Is f3 Limited to Certain
1.8 Group Velocity 24 Values? 65
1 Q Rnl1nrl"n, ronrlition~ for n;plp('tr;(' Intar; ,.,,, n. . .c ,.... -' LL
-_ <-./ ~ .V_!'~~ .. ~~ lYiVU..," vv
aces: xe ection an xe ra tion Lt .... '" .... T. _L r-. ._. .... ~
..:...~v • ,,""uv..,. V~ \.J U1U~U lYiVU"" H1 U
rot" lnrerna Reflection 1 Uh '.:1. L."'7
'w'~_.~_ v
R"v l'raein, 17 ,., 1 1 'T'I- 1I.T, • _, A t::o
<.. ~ ~ ~ U~ •. _uw .. ,,<U £>'1-'''' 'U." V/
..... _ " ~
c; i":" ._, U111111U1 J '1
xl xii Contents


'T'1. H1_ ~ "-.1.' •. 1
CHAPl'RR3 ,_'l_
~.- ~H~ 'W '-''iUUUVH
""':1-1 ,R"'I IN I" WA\' .... ·~ '\1-<'''' ' '" r~~.-rl;"n.o~ 1 '"'It:.
Introduction 75 5.3 Solution of the Wave Equation
3.1 for E, 127
3.2 Three Types of Dispersion 76 5.4 Field Distributions in the Step-Index
3.3 Material Dispersion 77 Fiber 130
3.3.1 Frequency Dependence of the 5.5 Boundary Conditions for the Step-Index
Permittivity, E 77 "'_ i.-lp. 112
.1 r.rOll" r nrl"v "nrl Group Th" IO:n'>t; c l\Aorl,,~ ",f" th", ~t"'n_ .rlo~
"h
Uelav ~2 '" -r
wavezur e U4
'~ ~ ~ : +rour Vp.lo~1tv i<:np.r\:inn X~ Fransverse Rlprtri(' and Tr"n,,_
' 1.- . '1.h
11.:1 Th", "n.-l verse IV. azne ic lVI0 es J4
Ettectrve Bandwidth 'lSI 5.6.2 The Hybrid Modes 136
3.4 Modal Dispersion 88 5.6.3 The Linearly Polarized Modes
3.5 Waveguide Dispersion 92 (LP Modes) 137
3.6 Normalized Propagation Parameters 96 5.6.4 Summary of Mode Shapes 141
3.7 Simultaneous Effect of Material and 5.7 The Normalized Frequency (V Number)
1\,{",,,1<>1 n;, ._, . 00 ana LU oTI 141
T
Sl C .rnrn'>r" (10 1 e r-un amenta tiE Morte 14~

'J
'Ill Total Nl rmher of M~ in " Sten.:ri1.-1py
LHAYlJi,;K" waveguice 14:

"U A 1'\1·. 11_1 N 1 )J-< X ontinement 10 a Step-index
):TU Power
WAVEGUIDES 105 Fiber 146
4.1 Introduction 105 5.11 Summary 147
4.2 The Ray Picture of Propagation in a
Graded-Index Material 106 CHAPTER 6
4.2.1 The Eikonal Equation 107 DISPERSION AND GRADED-INDEX
"'TDDDe' 1 C 1
4 . ..!. • ..!. _lis ersion Ke uc ion wi a .. ' J.JJL<.I:" .... .. .JJ.
ro. _. T, -" 1" 1 1\ r ntrorl,.rot;"," 1<;
'-'1. uu~u - UIU"A 1. I VillI; 1. • V

ci.ut c Ul LIlC Uli:lUt:Ol - HUC)\. c.,., T. ,1 TY .. 0, T -'
't • .J ivnxnu ""'P~1.~'VH ... -..--.:owl' u".~
n 1 11\ Fihp.r 1<;
,u'~5U"'~ 1.1.V
4.4 Approximations Using the WKB 6.3 Single-Mode Waveguide Dispersion 156
Method 113 6.4 The WKB Analysis of the Graded-Index
4.5 Direct Numerical Solution of the Wave Fiber 158
Equation 118 6.5 Dispersion in the Graded-Index
4.6 Summary 121 Profile 164
h'i 1 POU1"r_ T ,,,w Prnfil,,~
:"1\" 'F.R 5 rcv I-'rofllp" Ih4

~THD_ '1'\.: I)l Y D( AD Tnt,,] Nllm~ nf Mnrl~ m "
n.'i,2
WA 1-<1 il '1"' •. "" /.., Ura<le - ndex waveguide ():')


,).1 TmroaucIion lL') xiii

~ ~ " -' TV ...., _, T.

V.V .lHUU<LL ""l"P"l "lUll UI UI<O -U",''''A

n: 1£0

n _Ll , ..,.."

6.7 Single-Mode Fiber Design 170

6.8 Summary 172

8.2.4 Solutions Near Cutoff 225

8.3 Perturbation Approach to Correcting f3 225

8.4 Effective Index Method 229

8.5 Summary 235

CHAPTER 7

ATTENUATION AND NONLINEAR PPPJU"'T.C;: TN WavpnTTTnpc;: 17,\

7 T lntrodnction 175

PAK ~

7

,m!" 171'

A

.

('( IDI NfT ANI)" 1\11 I-<'R I(

7 1 Ah"orntion nllP ~prtronir

A I'll A I.Y ."\ L"+

7.3

'Iransitions ITl

7.2.2 Absorption Due to Vibrational Transitions 179

7.2.3 Attenuation Due to Impurities 180

Ravleizh Scattering 180

CHAPTER 9

THE BEAM PROPAGATION METHOD FOR ANALYZING OPTICAL WAVEGUIDES 243

7 4. Ontical r-rner

IXi

01

CI 1 A 1'I.T, "'T.

--:>. • ••• ,~.~ ~..,~" .•• =...,

) .d.L1

.

it irm Clf W<lVP" ").1,,

•• ~ HUV' '0

Nonlinear Effects in Dielectrics 188 Stimulated Raman Scattering (SRS) 191 Amplification Using Stimulated Raman Scattering (SRS) 195

7.9 Stimulated Brillouin Scattering

tC'T,! C'\ 1 on

.- .

The Fourier Transform in Guided-Wave Optics 245

9.3.1 The Discrete Fourier Transform (DFT) 247

9.3.2 Wavefront Curvature and

...., --.

.l L" L..Jl

7.6

9.3

7.7 7.8

--near JJ Jill (1~U U II L..J L.

~. . .-

.IV ,,,-,,,~~v' \""1'''' ~

~ ~ .., ~A

1.11 Vpll-..al kU

~ _ ~ ~AA

I.IL. ~U 11< kV7

tTll .,1\ "'';:'7

~"'J ~~

A 1\,1, I"h +~r £'\~a

CHAPTERS

RECTANGULAR DIELECTRIC WAVEGUIDES 215

-0'

Dimensional BPM 260

Waveguide Coupler 265 Summary 267

9.7 9.8

8.1 Introduction 215

R'J -W;vp Frmation An<lh",i,,--;::;:t!.I

CHAPTER 10

~~YT~Y ~ .......... .r" ..... ~ ..... TY~~~~r .... T ......

\....~J lVlVUC lOCVKTru"llJ

nl::!r

.

Wavesuide '11 h

101 liiiffidiwtion ?71

Porrnll1<1tirm of th .. 'R

')priVllt1on of thp, 'ollnlina Enuation

Value Problems LI7

-USlllg weal Modes L/ I

xiv rcm'pn'o


I f\ ') lIo.T. _, r» L ~.
• v ..... . r c .. ," ~VUl'UllC; U,"L v .... 11 1nVUI;" PART 4.
UT. '1'71::
'b'''~~ ~ v l\.T{"\TC'I A 11. '""' :r"'T'T"lIo. T ""'1:
10.3.1 Coupling Due to a Dielectric '- "' ........ ....- ... ~ rL'~ --;:F .... -,
Perturbation 276
10.3.2 Intermode Coupling 279 CHAPTER 12
10.3.3 Some Warnings about Coupled NOISE IN OPTICAL
Mode Theory 284 DETECTORS 327
1(1 A n l\,f. ~ r», .J.' '10C
-0' W "A"~~ ~VU1'''''6 ~v .... 'J ':n
1 f\ C ...... ...... _.
.v ..... "0 W Hll111 a ~VII uc;aLI;U 1 'J 'J nptp~nn ~
W"",,, c..,; £I", 'Jon
·0' '1 l'.T~;~a 'l'2r
10 I:: C' '10')
AV.V ~uuuu,uJ' 40/ .... 1'1 A C'Il'I.TD~.;~ ~_ t=\. :In.
'1' ,._u •• v ''_' UllU
CHAPTER 11 Signal Currents 334
COUPLING BETWEEN OPTICAL 12.5 Other Sources of Noise 337
SOURCES AND WAVEGUIDES 299 12.5.1 Background Radiation 338
12.5.2 Johnson Noise 340
11.1 Introduction 299 12.5.3 Dark Current Shot Noise
340
11.2 Counlinz of Modes between TTrT"IUise' --yr.y
.,,- '-' ~,...,... lk.;)~
,aVI;C;UIUI;" £..77 l om nine 1::'.1 ec
inQ: 0 an unrica Source L.. • o A t e
\. ou jU4
~ --:JZIZt
.. unuca oource . . ane iji
k, ulgnal error xa es "4,
LIliiriicu;rlZiiLlur "u.:
11.3.2 Coupling an LED to a Step-Index ILl LoherenTUeTecTlon 'TIU
Waveguide 307 12,8 Summary 353
11.3.3 Coupling an LED to a Graded- CHAPTER 13
Index Waveguide 308
11.3.4 Using a Lens to Improve OPTICAL DETECTORS 357
r'. _.' ')f\O '''' 1 T. ~~~
V .... vv '.1,' JJ
I 1 " _£. ...... .. ' ~ ~ ..
.. , ,-,Ullat-I; , <l DI;i:U1I Lv a ~1aU 1.J .s: 1 rrerrxterrrarr' llU ";)0
W" ,"',.. ;£1", '200 l''Iegar'Ve t!l c ron f\ nicv
~b .z:
, , n' roo -,' ,.,nn ~ --~
••• 01 • ""111 "--'VUl'"11C; .1V 7 ~JUU
1 1 ., "1""_ ro _1' ...... r ~' ~ - - --
~ ...... • 11" -r- '0 ,-,Vll"L<ll1L IVI r 11M1I J..J.k.k LIliif r eu aruc LI;'l "
Coupling 312 of the External Photoeffect
11.4.3 Disadvantages of Prism Coupling Detector 361
314 13.3 Photoconductive Detectors 362
11,5 Grating Couplers 314 13.4 pn-Junction Detectors 367
11.5.1 Basic Grating Physics 315 13.4.1 Quantum Efficiency 'T1 371
1 1 .,. ., r-.. ro ...... 1 ue pin Di ooe
LJ. • .J .... ~ULl'UL '-'VUl'1111C; WIUI <l J..J."t.k 14
r». .1 ,)''7 Noise in on an r: in
..... j.4.j e ectors
11 C ') 'T't r=. .,-' r« .C£! • ,.,.,n ~~~
'- ........... ... ._ '0 .1"'V .JrJ
r: ... ~-,...
II.V Y J£..V Contents XV



,'"> ,,~ ., ,,-,,~ '1"70 1 I: A A. ~~ ~
~J. "l' .... Ua.L _,'v ~_,. ~p"""-' ~ • ... v
~~" ~ ~ r> r> .~ ,~
.Lt. r ,)7 ~.J • .J Vllll;l VpUI,;(1l ,is tJ ~
13.5 Commercial Optical Detectors 381 15.6 Noise in an Optical Amplifier 433
13.6 Summary 382 15.7 Lasers 435
15.8 Fabry-Perot Cavities 437
15.9 Gain Saturation and the Laser 442
PART 5 15.10 Output Power from a Laser 444
( II-' It AI xr n 1« F~ iX", " 1 ~n,....,....~¥" 11111
J

L:UAriJj;K re
("11,\ P"FR 1.:1
SEM Il:UN Dill T()k! I AXF.RX 4.1.i
flP' I( '<1 1l<1 ") <1'1 I( 11\:
AND AMPLIFICATION 387 16.1 Introduction 453
14.1 Introduction 387 16.2 Recombination in the
14.2 Blackbody Radiation and Planck's Semiconductor 454
Law 388 16.3 Material Properties of III-V
14.2.1 Densitv of Modes in a One- Semiconductors 456
....... ., ,,-, ,">O{l 10.Lt ~I vens~ LtJI
~Al"ll: _,v/
JCCU a ion rro aDili:v 40U
lLt.L:.L: LJenSHV or ivioues a ~WI -
.,-.,.: .1 ,,-, . '101 16.6 Uotical Emission 40T
v"Hay 'J ~
~ ~ ~ . If, 7 Inrir-al A ~nrl ·~~in all"
~ '+.L.J LlClIiSUY VI lVIVUCiS lIi a ~ 111 .... - L
Dimensional Cavity 394 16.8 Gain Coefficient 468
14.2.4 Blackbody Emission 397 16.9 Quasi-Fermi Levels and Inversion 472
14.3 The Einstein Picture of Absorption and 16.10 The Junction Diode Laser 474
Emission of Radiation 399 16.10.1 The Homojunction Diode
14.4 Lineshapes 404 Laser 474
~. ~- 1t:10'"lTT T 11'7t:
l Lt. VatlI '+v ~VdV ... £"""""-'J ........ ,,'"
"'''~''.' ,~
14.0 1 nree- ana r otrt=rzversvsrems LtV, ~ U d U. J 1 nUUI; zrrnrxnrrrcrrr
uam sa ura ion £"" ... 41 I1Q{
14.1 4
1L1r.A A " T 101
14.X A mnlitication m a TIIstnnuteo .<v • .<v. ~ 'v~
t\mpliner 1 L '" ~ ,,-,1.. ,,-, r.
Lt1l ~V.~V.J '0' vUl"UH' HI
14.9 Summary 419 Well Lasers 485
16.11 Modulation Rates in Semiconductor
CHAPTER 15 Lasers 486
OPTICAL AMPLIFIERS AND 16.12 Summary 490
LASERS 423
1" 1 T, 11'1'1

1 C '"I A, .1 A '"' A"A
~_, .... '-'P""= T Cir rzs-
, '" '"> n. . en. ...... r 11')1:
~,J.J ~ • VP'" """ V' "'" 'P"U ,~~~ . _ _, __lI_\fi

C'HAPTFR_lR_
PA.KT I> D D C illo.rClr\D CI .c::'l'l
nliYT Ir AT T'\"11V :;'1;; A n: ........................ '-'~ ~ ~'-' ............ .l 'u'"' .......... "oJ oJ
~~ . ~'"-',~ ...... ~~ ~~~~ r- _,,,
18.1 Introduction 533
18.2 Intensity-Modulation Sensors 534
CHAPTER 17 18.3 Phase-Modulation Sensors 537
WAVEGUIDE MODULATORS 495 18.4 Multiplexing of Phase Sensors 542
17.1 Introduction 495 18.5 Other Types of Sensors 542
1 '7 " u: ,+ 1111" •• +, 11. K • .l •• 1 AOt; rec uer CV-IV, 0 U atec
4 ._ ~ '6''''''~ ~4 L.A", .. , L~' r/~ LO ••
__1_"Z_2_ ..., • lI.L " ""'"
4 .~ , ...... t o «"U L .. '" .r ....
~FF"'rot AOO ,0 r- " ,,'- ~ ..
~r ~o.J.", ;U -Jt;llMUVt;
1"7 '1 1 'rho D-"<ll1'1 '.~ ~+ ~lorot~. o .- <::A'l
.. .r- '" ._ ~
magnetic Waves in an 18.5.3 Polarization Modulation 544
Anisotropic Medium 499 18.6 Fiber-Optic Gyroscopes 546
17.3.2 Pockel's Effect 503 18.7 Summary 551
17.4 Phase Modulators 508
17.5 Power Required to Drive a Phase
Moiln btnr "00
ftCCE"l'U.,.A ~
.j, P lI.Jf"il"l.,torc '" " ~TT~ ............ "'" "-1" _ T~ ~_T ~ ,y
-r n r; \.Jvv':)-'[ l'l \ n r L'I o 1
1"7 "'7 T. -P • ,11K .r., 1 <::1 A
4 • ~UL"" ...... '" ~~
I.., "'7 1 n~L n " ... <::1,j
.l • d ~ «v' -~ "'.~. ,~ -'.l
__ A '7 ,. .. r
1/.I.k IVHll,ll-L.I;UUl;l .e JIU AYY~l'IIVIAD
17.7.3 Coupled Waveguide BESSEL FUNCTIONS 559
Modulators 517
17.8 Electroabsorption Modulators 520

17.9 Acoustooptic Modulators 522 APPENDIXC
17.9.1 The Photoelastic Effect 523 (lPTU 'AT _l)()W_ER T .TMTT OF A
WAY nu' lun I JI t'. l ~ IVll I.A t'. J
.lJ.'.l Raman-Nat Moc1 ators ~1.4
..... ,. Jr ...... T _flL"!_ ''T'''T ................. Tr< ~
_lJ'~ Hra!J£ MOGt 11It()r~ YLh .l'\..~IV.l~l ..... ':H.,..I'1.1 .1 D.l'\. .. '1'- .JU
Annl;"<otinn" nf A in",
-rr -r -'"
uevi es :)1.
17.11 Summary 528 INDEX 563



T"""T T .... T~ ... 1t. IT""" .... Tr"n'" T rI
1"1 I '-I I JA IVI 1"1" I A I ., g II g T
~ _ ~ ~4 ~~ • ~~~ ~4~~_ . ~ .. •
1




































1



Y"" Try-,.....,."T""Io.Y T""-'ry-,y".,.. T "- .,.. TT""Io. '" TT""""""" TYT"""" T
I '-I I 1'<1 II II I I II 1'-1 A '-II I I IV ...... I'< V I ...... VV
~~ ~~-----~~-~ ~ ~~ - - ~~- ~~









CHAPTER OUTLINE
A Brief History of Telecommunications
Development of the Optical Waveguide
Types of Optical Communication Systems
l In Uotoel lCS


A DDICC UI~TnDV nc TCI IMII'ATlnM~
~ _ ... - ... ,,_._ .. _ . --- __ ...... _ .. _ ._ w_

In 1837, the era of electrical communication began with the demonstration of the
telegraph by Samuel Morse. The telegraph passed information at a data rate, in
modern terminology, of a few bits per second, but the speed of propagation was
essentially infinite compared to the message itself. The transmission medium was
wire cable. The telegraph was followed by Alexander Graham Bell's invention of
the telephone. The first telephone exchange was operated in New Haven, Connect-
rent. J~A In XIX. A imarerv 4 KHz bandwidtn. the teiennone
a major increase in me errecTive oanowimn or a mooerare-oismnce communicaTion
"y,,"vll1.
A .1. -, ,t.. . • + .t.. 1 fl • T.
'1>6uu""),, LU'-' 6''-'"L'-'~L '" "'-"U'-' ''-'U''"'''L ~. L"'-' • J L" '-''-'''LU'), '"~ ~"U''-'~
( M"YUTPI '" pllll"irhtinn nf "M"V\llP' '" Fr111<1tinn,," in 151."151. rh",,,,,,
T L .• T
mathematically describe the propagation of electromagnetic waves. Maxwell's
equations, and applications derived from them, are the foundation of electrical
machines and electronic devices which form what we now characterize as "high
technology." Based on the predictions of Maxwell, Heinrich Hertz demonstrated
long radio waves [2] in 1888, and in 1895, Guglielmo Marconi demonstrated radio
(r-ornrrmnir-ation wirhrmt wiT"""!) h~"prl on _1 . w~v".."
~
;:SInCe ese roneennz e or s scien is s an engineers ave mao e s eaov
'J','v';~ lUWiHU ueuer auc LdMt:! • 1 lIt: L1 t:IlU lId:> LJt:C! .
. ~t.. + • '.t.. • 1 ..1 .. 1, •
L~ '"n .. '''6'''-''-'''-'':IU'-''''-' ,-,u".,-,. '""'-'~ "LU i" ~l:'~'
h~nrh,,;r1thc £~ .. ,.., ;n£~ .. rn~t;~n i., .. h, .. .,r1;~ ,,,h;,..h ,..., .... ;"..1 ,~;,.." c;on.,l" ,,,;th
'J At::> 'J , '0'
15 kHz bandwidth, operated In the 0.5-2 MHz range. TeleVISIOn, which requires
3 4 and



C.1I.6ll, .•. ·.h'. _] .c ''''''ITT ..... ,~
'UJV~' V HU£~ V<4 flUUl, 1 ru"vu ,-,alllvi 11 vy LV IVV lYlII£. lJUlllIJ;; llv l'1"1-US,
.. ~rI~ .... es o c. .t. -' I. , ..
'u H,",'i .v "''"' eLL \HU'-'lV .a .. ,,,, • IVU~lU-
wave technology has been widely used to create terrestrial and satellite communi-
cation links. The push toward higher frequency continues today with the
application of optical frequencies to carry information.
The advent of the laser [3] in 1960 gave system designers visions of transmis-
sion channels with huge bandwidths. The first laser operated at a wavelength of
694 nm, which corresponds to a carrier frequency of approximately 5 X 1014 Hz.
t'.Ven 1 nercen of this ro nlel hp. rr-al I 7 p.cI I n ::I i r::lti "... cuda .....
-I
iI WOUIO represerrr a slgnar enarmer with,) ·1 Hz bandwidth (a teranertz IS rrr=-R7.Y
C' •• L . . ",;
.... U'"'H " "]"L'-'IH WVUIU l,,,U aflfll' tv IV V IUt:V raru eis aT
lI.,fll. .1, .1 1 ... 9 .1. ,t. • 11. c: T7TT . TT.
••• ~ t'~. , -or -.:-u- 'r ,","H" ". -' .n.~~L p'-'I ""H. ~LV ,v 'VL, ~
toward ncino this: trplYlpnrlr\1lC h"nrlll"rlth ll'''C 1'lYI,tarl h •. hun 1'""tnrc:
~ OJ
1. Electronic components do not operate at such frequencies or speeds. Since
most information today is ultimately converted to electronic form, the speed
of the electronics determines the realizable bandwidth of any communication
link.
2. There was no suitable transmission media for light.

rnrl" ..... n .. a .J.,n~ .J.,~~~ .-I~n ,.-In~ n+"~ ,I, h ,C .t.
-J' LU~' , ...... v .. VL "LV'" , Lvla-
h prt 7 r'ntTl .. .,ta~ ~ .. a ct, ..... nu' ...... a Inc' "a TJ., a a ;r- -' .1
r- '0 -1' "'''''"''
neck IS e rna or concern. 'I he inve or taster svstems has 1~c1 to taster p]pt'tron,t'
wnicr nave swi c Ig mes a roac ing severa icoseconas. IS ::11-
ways dangerous to define limits, but many feel that electronic devices have just
about reached their ultimate speed. This has motivated the study of integrated op-
tics, where light, which has a much higher implicit frequency limit, is used to
control light. Also, there has been creative research into optical communication
systems where, for example, information is carried by many wavelengths, or by
ultrashort optical pulses called solitons. Without doubt, the biggest research task in
• J., -'. .L -' ,I .C .1 -' .
.~ .. ~ ~_' ,~~~ 'Ul v", 'H'" U'" v, vP""'" "VV ""H"''' "HU U"'Y''-''-''', "HU Ut:LLt:l
cornmunicati on
e act< 0 S:l itable transrnission mertia has hp.f'.n hi o""r th"
J
past jU years. TITe TIellf or oPtical waVegUideS IS wen estabnsned and IS nOW:!
d. .• • _]. ~'r' ~. • • •
,HII 'IHIS IllUU"Ll y. 111t: HI ~L pi:l1 L V1 Ull~ Lt:.II.L Will Ut:Vt:lUP LIlt: LHeury i:UIU appncanon
of the optical waveguide, both for long-distance communication (optical fibers) and
for integrated optic applications.
To address these problems, a new discipline has emerged called optoelectronics.
Using light to convey information requires special technologies. Information must
be put on the light beam using extremely high-speed modulation. Once modulated,
.t. . ~t. . .. .. .. ..
.u,", "6'" ""L." ~-un;r i>l£aUlv allU VI "Wlll,llvU-uTLIIe"
rlp~'rprl rpt'P' \X.'ha ... ;. n .. ~"a~ ~. tho .. ar-o;"a ..• ha ; ... +~ ..... ,.'
, 'W , ..... ~u V~ 1 ""."'u
from thp. h ohr ( \n~,rI"r thp . h]ot'lr rI,,,or<l", 01' " po", ~"~'o~.
~ -r -0' "7
er counies in orma IOn on 0 a ransmission
1. e ransmt nanne 10 hp form
. .
VI " ~U1L"Ult:--srgrrar.
2. The channel is a medium bridging the distance between the transmitter and
receiver. For electromagnetic signals, the channel might be a wire, a Development of the O_illical Wav=ide_ !i



",~ .. ao,,;rla ~r Fra ~~~na A ~.h n' ,n~1 fro ,I •• J.. ~J.. .J..
-0' , 'I' '0' ,. '~O" ••• ~ , •• 'U
• h, <ltt",," <It,,iI <I"rI rli ~t"rt"iI
r- '0' -J
3, The receiver extracts a weakened signal from the channel, and amplifies it. A
semblance of the original information (audio, video, etc.) is generated from
the modulated signal.
.. , Infnrm~tinn .

&
• _l
T.~ v--... »<: I"'h~ L.----. /' D
~ .... ,~ ~ .~~~, .~

Figure 0.1 A communication channel consists of a transmitter, a transmission medium,
and a receiver. The transmitter converts information into an energy form appropriate for
the transmission channel. The channel carries the energy, but also distorts the signal and
adds noise, Following detection, the receiver regenerates the information in (hopefully) a
nearly identical form to the original signal.


-_ .. _. _ ..... _ ... _ -- _ .. - .. .... ,,-_ ...... -
1.11:: Y I::LVr IV'I::'" vr nl:: vr I '''-,",L YY,",Y I::UU'L11::

D ',11 ,'I' ,..l .~ 4'. .4' 1
• -~ -~ _ .. "'. u "'y'~' J ."'~ ........ "'''' uy""'''' ~ .... ~"'. v"' ..... o ... o u~,
e, 't..,hla f"r ra ,.,hla ;n.,t;~n I'nlre Dr~hla",e ~f h ,rI , .... ~tJ.. {. , ... ' , ,\
'r -r'/'
flying objects that interrupted the beam, and the need for line-of-sight links com-
plicated reliable transmissions. Laser beams are also distorted and randomly aber-
rated by propagation through turbulent air. There are a few specialized applications
for free space communication, such as communication between satellites in orbit,
but most terrestrial applications require a protected transmission channel.
A _1 to thp.~p. _1 i~ to ltP. 1 i ohr _L ~ w '..l. whir-h
. . . .. .. .. ..
0 oro ec s e beam rom In erru IOns and counters 1 racnon the tendencv
~ , VIlC "'~IlCIIle LU<1L is C<1",Y LO <11. a .yze anu visu-
VI <1 W<1VC LV "IJIC<1U <1", 1L
.1: .L ' -l: 1 ;n: rl""· 'T'L 1 11. c.
"ULv .ro 'Uv pv"VU'v 'vH" '"YvOU'Uv V 'OU'v V .... , •• Hv 'vH"'-''' Ly lvlV,,"U"
th" I, oht <'''' ,1-'''''- .,nrl .,Ie" "11,,n, lioht t" t .. .., ,al .,r""nd__a n±L _h d
-0' , ~ , -0' 0' ,_, ,
It IS fairly straightforward to select a focal length and lens spacing that will provide
a stable beam path [4]. Manufacturing difficulties and surface reflection losses rule
out such a structure as a practical solution for waveguiding. An extension of this
idea is the continuous lens, Consider making each lens continually weaker in fo-
cusing power, while increasing the number of lenses so that the light stays con-
fined, In the limit of weaker yet more frequent lenses, one gets a continuous lens,
t'umre U,2 illustrates this idea,



Ci~ --a-- n:!~ M t\
fLit--- 1m --- Jfi3--"'" Fill / [(I hi HI·----..I!t --- .......
VI VI VI VI V V V \1W"-
Lens Waveguide Continuous Lens
Figure 0.2 The optical waveguide is a natural extension of the lens waveguide, 6 Introduction and Oy,;:;:vj"w



T. a.t. . £" .. . . ..
~" '''''' IU" "~'11", ll1v I". V~ "0111 1" "'Vll"la11U, Uv11l ua",,,- lvwalU Ult;; ~t;;!llt;;f
~+ tho "'~",.~ ._, 'T'1-. 1. • 4' _1 . l':1. ',1. 1.' •
-0 ."''''~. U O'U"" "v",. .. ,"" "'0" ll1UCA
of refraction at the core than at the outer perimeter. This waveguide structure
solves the problem of surface reflections and also allows the waveguide to be bent.
We will develop an understanding of these guiding structures in the next chapters.
What about transmission losses? Is it possible to send light through kilometers
of dense material without excess attenuation? Glass is an obvious candidate for
making an optical waveguide; it is commonly available, easy to draw into fibers,
ann 0010; tran III rp.nt HowPvP.r 011r cornmon pxnp.ripnrp with olll<;:<;: 11<;:"" i"_
. ~
vorves lOoKing If reumrnrates no thiCker than a lew munmeters. W nat nannens if
•• . L ~ . n~
"01lL pa""v" a "' .. VII.",,, VI ~la1>1>. Dillly lllt;;a1SUICIHCHl1> mal e
.1' .. c. . : _1 . _L ,{\{\{\..1 ..... .. " ~ ~
~. w ~ '" 0'"'''' '"'' "UVUl p'-'I • -n:TTUIIY
illtp thi" "" ..... ha .. ra,..., th"t" rlR i~ rI",fi"ari c c ., I",~., .. ith ..... i,.. .. .,ti",:
TI' '0'
Pout (0.1)
Power (dB) = 10l0glo-
Pin
A 1000dB loss represents Pout = lO-IOOPin! Considering that there are approxi-
matelv 1019 ohotons ner watt of visible lizht this remarkable nnmher irnnlies th"t
.~() i~Ill
II WvUIU UC .v IU iauncr anproximatel Y 1 V vv 0 o ~e o on
~ .. ~
u"vU5" "v11,-,-"-1lvHIClCl JCH)o;UI vJ )0;1(11)1> • .Lv PUl UH1> HI pCI1>pC\'-U VC, lHC 1>UH
","1 (!\ ':l.Q v ~
-J Mn~~1-. ,~~
In vipw of rhi c '",.. .. "'rI,h1'" <ltt"'""<lt,r", ,t,~ 'tnnr",,,,,,,,,,, th<lt f<lr~ilTht",rI
-z- ~
pursued optlcaflioer waveguides. In 1l}66, K. Charles Kao or Standard lerecom-
munications Laboratories in Harlow, England, suggested that the high loss was due
to impurities, and not an intrinsic property of glass. Kao went on to propose using
fibers as a transmission medium, describing many of the fundamental modes
that such a fiber would carry in his classic paper [5]. In 1970, Kapron, Keck,
and Maurer [6] of Corning Glass Works confirmed this prediction by making an
u ra ure ~. ass usinl! c emica va or ue osi ion ec ni ues t at disotaveo onlv
"'" ..... n Te-ss-
cs» Ull/"-HI . -nmay; ullfapUft;; glaSS !lUW uisprays
,h "(\ ') .-lll n. ,_ -. •• 1. .4'. e
~ 'U ""'''o'''~' "-,,,,,,".
()""'" th", "an11n>;~n n~~hla ..... 11m~ "'~~'a~a.-l ~nt; ~n hI- L
" , '1' --.::>"V
came a rna or rar nr in imr-armn svcternc for hoth lono-rlil:tllnre ,,"rI 1",,,.,1
~
area networks. Motivations for using optical communication include:
1. Optical communication links have a wider bandwidth than copper or
microwave links, so more information can be carried on a given link. The
effective bandwidth of current optical fibers is approximately 25 THz. Think
about this-25 THz information bandwidth on one single fiber!
,.., A ,. c: •. ... -' .
<". 111 0Ia"" HUvl1> I" Iv"" ll1all 111 "Vppvl VI uicr owave
v' wt. D" •• _ _, . 1. ..
'J u.'" • , U11U '~110"" """'"""'''''' "''''' v,"
more rOl:t~"
':l ()nt',..<1 ota",~ n~a v",n la~ nn.-l li~hta~ ~; in~ than ,n ~ mnt, in _,
'r -J '0' , o· "0 "0 ''''''''''
ducts or aIrcnift. Opportunities 'in Optoelectronics 7



. . .. ~ OL irn osstbte LO a or moni or
-to V~LlU:U Wi:lVC~UIUe1; are ~UUL so
-'.
Ui:l'i:l """""UU'J ." ",,,,, ......
5. Optical waveguides are immune from electromagnetic interference, ground
loops, induced cross talk, etc.
6. Finally, and perhaps most important, semiconductor technology has
developed a family of lasers, detectors, and other integrated optical devices
that are compatible with optical fibers in power, wavelength, and size.


,.vn~1I:' nlf: nnT.I" A. I"n ........... I" AT.n ... ' •• 11:'
.. ..~ "'. "' ... "'~ .. ...,"' .......... " ... ...,~ ."' .. ~. ~ ..... ~

T. .f' ,1 J., rI, t.r: 1.
•. ~ 'J l:'~~ ~. ~l:'''''''' ~J~'~'" .... ~ -r . ... ~.~ .. ~ .... J, '~"6-
ni~t~n('p. tp.1 nmir-at ion (tp.lp.('om· UJ~~ thp. fir-er onti('~ I ('ommnni('Minn ~nnl'_
• -rr
cation. Telecom primarily involves point-to-point links, such as a long-distance
telephone link, which carry vast numbers of multiplexed signals between two cit-
ies. Optical fiber has become the standard for telecommunication links for the
reasons listed above, not the least of which is cost. All of the advanced technology
that appeals to engineers and scientists would never be installed were it not the
mos t cos t - .",,- solution to most _1 nnicat ions nrohlems The maior cn~t
. J
com onen in e ecom is urcnasing ano ins a ing e omtca er.
r
llC 1;Cl;UllU Ly~C Ul U~Ul;i:ll 11; ueua
T"\ rl l' ,I, ..; c. rI, ,J., J.,
m .. ~~~ M ..... ~ .... ~ ........ ~ .. ~~, ._~~ ~ .. _ .... ~ ,~, "'-"'~'J " ..... ~~,
"nil "Tnrl tn.,.",th",r in " In"" "r",,, n",t"TnrL th"t rn'> <'n,.,.,
, ~ -J 'r
thousands of connections and hundreds of kilometers. Datacom applications are
different from telecom in that each user requires an optical link. The cost pressure
is no longer on the carrier medium (the optical fiber), but on the associated hard-
ware such as transmitters, connectors, switches, filters, and receivers. The need
for cost-effective devices for datacom applications is driving much of today's re-
search in optoelectronics. Visionaries in the field see the day when every worksta-
non w De lllKe to a H!I - ata-rate 0 nca network t at a ows interactive VIdeo.
. . . . .
massive UlSLHOUleU compuung, anu rapiu access lU every otner noue on a nerworx.
r". ~1. • 1. -' .. _1. _1.
'-''-'., ........ ~l:' .. '-' .... uv,-" u ... o 'U'-' V"'UU '''HU ,~ ""'pp~" ""''-'u ... u,-,,' 'vu"., VUL U .......... ,,'
,f' f'¥. ,¥. ·rI,
" 'J .. " .. , , • 'J ~"l:'l:'~' • ... ~ ..... ~ .... ~J~w." ... _ ... , ... _, ...... , .~
sril not np.vp..1onp..n to thp.. noint of ('o~t-p.ffp.('tivp. . 1p..ntMion 'hi~ i c " arpM
. L . 0
opportunity!
OPPORTUNITIES IN OPTOELECTRONICS
To transfer information from one point to another, whether between two work-
stations or two cities communications svstems will reouire switches
am i ers LH' ers, e C. vv e sue II evices are we ueve 0 e or copper- ase ano
, .. , r .,.
LV'" lU11\.", luau, UI UI"'''''' l;UIUI UU HUL CAlM lUI U~Ll-
~",I li,.,l", Q .. ¥ '" .1 th .. ¥, .~,." ~,., .C' t, 1-,,,, "".¥L > 1
, ..... .~.~
mh,('h <It fir-er ol<ln('p ",ioht lnnlr triui<l' In f<l('t ('nnnp,..tino <In nnti,..<ll inlr "Tith
0 '0' -0 'r
good repeatability, low loss, negligible backscattenng, and rmmmum skill IS a 8

Introduction and Overview

,I ',].' ~ r..

'I:' _. ~ .. ~ ••• ~ ~,~ ••••• ~.~ ~ .... ~~ ... '-'11",", a

hardware problems m optical commurucations are needed.

Unlike electronic communication systems that have a limited bandwidth, optical systems can essentially be treated as having infinite bandwidth (well, in truth it is only 25 THz, but that is close enough to infinity for most of us in today's world). System designers need no longer be constrained by the communication theories developed by Shannon and others concerning limited bandwidth systems. A new

mentautv IS reautrea to take TU aavantaee 0 tnis ricn new tecnnotoev.

~1 or; lOr, ,r! ~1- r, ,11" .... "

'.. '0 -a 'J .~ I:' \"-"~~ J.

'hp>!p nttp.r . in siorra -tn-nni~p. ratios 'hp>!p "npw" nnti"<l tp"h_

. ~ .

niques are really tried-and-true modulation schemes that are presently used in modern radio communication systems. This new technology has also opened the door to new phenomena, such as optical solltons, which exploit nonlinear properties of optical fibers to make potentially better communication systems.

A system designer today will ask, "Should the huge bandwidth available in ontical links be exnloited throuzh time division multiolexins usino exrrernelv ~hnrt

empora uses. vr S ouiu wave eng . dvision mu ipiexmg e use were eac

. . ~

.. "" .,.. .

M~l1i:ll II> Ui:ll1l>l1l1ncou C:lL u" VVV 11 ". .L Ill"

. VI

nas me

......

Noting that this book IS primarily devoted to describing the hardware aspects of optoelectronics, readers interested in the systems aspect of optical communications should also refer to the excellent book by Paul Green [7] on fiber optic networks (datacom) and the books by Stuart Personik [8] on telecommunications. These books serve as excellent companions to this one in terms of complementary material and concept development.

_ .. _------ .- - -- ,- - -

1. 1. C. Maxwell, A Treatise on Electricity and Magnetism, unabridged 3rd ed., 2 vols. (New York: Dover Publications, 1954).

2. Heinrich Rudolph Hertz, Electric Waves, 2nd ed. (London, New York: Macmillan and Co., 1900).

3 .... T. Maimann "Stimulated Ootical Radiation in Ruby," Nature 187 (1960), nn. 493-94.

... J. ver eyen, ~ay racing 10 an opuca oys em cnan. L- 10 aser tuectromcs zn

(n. ..J ,..."cc. "TT. n. TT_ll T. 1 non

.e:; D n v.

T'\ D V,

.-1 D T'\ l\o6. A' -I DI. T I.., (1 ",..,In

423-25.

Suggested Readings and References 9



~ ., . . '~ ~ ··r 'T; ~
raUl D. ,-,1<0<011, H., "'YO::. VP"" on WUf fI,~ ~ '-llH~, n •• r 1<;11""<; llall,
100'1\
8. Stuart Personik, Optical Fiber Transmission Systems (New York: Plenum Press,
1981), and Fiber Optics Technology and Applications (New York: Plenum Press,
1985).



rT"IT TT""" T"""T T .... TT"'o. ... It. IT""" .... TrT"I'" T rT"I""" ,..... T rot
• M 1""1 1""1. '-I • IA IV. 1""1 " • A • • I II ... , r» ... /lDT[I;;D 1
~ ~~~ ~ - ~ ~~ -~ ~~~ ~~-~ ~--~- _ .. ...,. .. - ...
OF OPTOELECTRONICS:

MAXWELL'S EQUATIONS








CHAPTER OUTLINE
1.1 Introduction: The Tools of the Trade
1.2 Maxwell's Equations
1.3 Constitutive Relations
"1"" nT. n
, . , uv ,u"v ~':IUULL~"
oo u ions 0 e vvave cu ua ion

.h I ransverse hlectroma!Jnetlc Waves and .ne Vector

7 DJ..~~, ".. .~;h'
OJ
1.8 Group Velocity
1.9 Boundary Conditions for Dielectric Interfaces: Reflection and Refraction
1.10 Total Internal Reflection
1.11 Wave Description of Total Internal Reflection
1 ,.. ... ,,'- .,,- .... .<1
~. ~4. I UU"'-' >JUl< l UpVll
l.U oummarv





1.1 INTRODUCTION: THE TOOLS OF THE TRADE
There are many tools available to analyze or design an optical device. Imagine
trying to characterize the optical behavior of a simple magnifying lens. We might
first project an image of an object onto a screen. Using a ray picture, we could
rlp"l'r;hp thp. m:lonHil':lt;on fOl':l1 lpnoth .1 nhnp" >Inri "0 forth of thp lpn"
~ . , .
. .t'umre a ms c aractenzation IS ca e ~eor.netrlc ODtlCS we were
vl::fy IJ<:: L ""'... I V'-', WI:: h!,IlL nonce mu CUIUIS ur i mtgl::s at SI igl LY di -
c.: c: 1. In: 1 1/1..\ "1"', .cc
,v,.., .. , HV ... lU'" .V .. o v 'o'''V '."VI r : .LV .. , ,",V""'LLVV LL .. o V"V'-",
"" '''~nlr1 J..~, t~ ,,.,rI, ·vt~,.,rI ,,.,rI .. v,.,·I~,.", tJ.. ~,~t .. ,.;~ ~ .... rI rI;v,., .. ,.~:,~ .... ., .. , yf
'1' '1' '1'
the lens [2].
11 12 Chapter I The Fundamental TOOISOf



7J\ H
...:/./ -- <, \:/./ '""""=- ~
~ ---~]!JS
...-.. ~
Blue Red
(a) (b)
\ \ \ ( ( I -n 'P: -- - ,.., --
n --
- ...... -......J.
I ["j_kJ
VI VI

-"
IV, lUJ

t-Igure our 0 manv wavs to desert e the 0 tica behavior 0 a lens are a an imaze is
formed through ray tracing; (b) the image position may vary with wavelength, due to dispersion in
the lens; (c) diffraction can blur the image of a small object; and (d) an image is formed one photon
at a time,
If we could shrink the diameter of the lens to dimensions on the order of the
wavelength of light, we would notice that the image begins to blur. Fringes beain
t~ <u''''<3<lr th <It O~d ~ ,. :~ .1.d ~h:d '1'1. .J uu:
+r-r- '1"t' VJ' ,_ ~'JJ ,~~ •• ~ r .~ .. ~ a.~
rmvct nnt"h1" .,t ~...-.., j o ,,,h<3r<3 th<3 ;"" OT'\t;,,~ .,~~" •• :~_ .1..
0 -r or
Iizht travels in stf3i!mTTules beams to fail The nroner descrinrion of th~
1 eq ui n:::; wave S eoua ions
It:O ~v oasec 0 iviaxwe an IS ca e tmvstca optics
r~'
V, YVLt ve VI" ,'-.> PT.
If we repeat our measurements in very dim illumination, we will reach a point
where the corpuscular nature of light is evident. For sufficiently low intensity, one
photon at a time passes through the lens. The individual photons arrive with cer-
tain statistical patterns (such as a Poisson distribution) [4]. Different sources of
light can have different statistical properties, and these properties will affect the
~.,.,I' ,f' ',1 '.-IO_ .1- I- ...1-10..1- '.' T _'.'
"1' oJ '1' 'U"' ""U OJ" HVH' U>" ,up"" "'5H<". HI rurs
nhotrm T",,1", m.'~~t,,~ n~t;~" ;c t~ rl<3c"r;h", th", ""ct<3 re:
r- T =r 'J l" J'
hac 0 tnese onnca I 1<: IS rare within a certain domain Whpn
eafi l~ wi ue mo es 0 a wavemn e we use nvsica onncs, W en eann with
• 1 r • . T
vpu"<,, VI ~l~llal~, WC U"~ \.IU<llllUlll V]JLI""1 "Vll'-C]Jl1>. HI , we resor
to the technique that leads to the most direct solution of a given problem. In this
chapter, we will introduce Maxwell's equations, which are the optical tools needed
to describe the propagation of light in optical waveguides. Using Maxwell's equa-
tions, we will derive and solve the wave equation in an isotropic media. Following
solution of the wave equation, we will explore refraction and reflection. Total inter-
nal reflection is addressed in the following chapter.



1 ? 1\lIAYW:1 I ,~ enllA' InM~
-- .,,_
Maxwe 's ec 11" 1 ms ~re aronahlv the most "i(Jnitir.~nt scienrifir- develonmenr of'

~
me 1 'jm cermrryrr is impressive to realize tnat tne same equations can be apprleQ
from 0 Hz (DC) to frequencies exceeding 1018 Hz. (In fact, an upper frequency
limit on the validity of the equations has never been shown.) The fou~ equations 1.2 Maxwell's Eauations, U



~n~ h~ ~~ -'!CC. ',1 . , .. ,.nl +" 'T'1- 1: ..1 _j, .,.]_ oJ"
-0' " .... ~ .. ~J ~. .. ~--~ ~~~~ ••• vv ...
t' .... r ..... c

-(JB f E· dl = -fr J B' dS
VXE=- (1.1)
at
loop area
aD f H ·dl = J J'dS+fr J D·dS
VXH=J+- (1.2)
at
loop area area


Y'H - U I H'd:S-U _il3)_

surface

r
rT ..,. ..,. rn .r.
II p J II U.;:J I.tenclosed ~L"')
surface
where S is the unit normal to a surface, and the area integrals cover only the area
enclosed by the loop. In this text we will use MKS units, with the exception that
most physical distances will be related in centimeters where it is convenient and
obvious. The units that describe an optical field in MKS units are listed in Table
1 1 h,l~ ..



Table 1.1 Electromagnetic units in the MKS format
Hnlts
E Electric field amplitude Volts/meter (VIm)
H Magnetic field amplitude Amps/meter (AIm)
D Electric flux density Coulombs/meter- (C/m2)
B Magnetic flux density Webers/meter'' (Wb/m2)
J Current density Amps/meter? (Alm2)
p Charge density Coulombs/meter' (C/m3)
n (,horn. r: {('\


N .... t" th.,t 1<' .,nrl J:I c r- .. n.-.."I;t,. '/no ",h;"h rI .. ~"r;h .. th .. ~tr .. n{Tth .... t' ., fi .. lrI c t o
or '0
o;,,,"n nnint j n "n~('f" 11 ~nn R ~rf" HlJY'p~ I Intortlln~tf"h nllf" tn thf"
'" r . J -J
o en use or e ec romaznenc uan Illes t e istmcnon etween am olitude anc
flux is sometimes lost. This is especially true for the magnetic field. Amplitudes
and fluxes are both vectorial in nature, which means that direction and magnitude
are important.
Within a continuous medium, these quantities are continuous functions of space
and time, with continuous derivatives. All real solutions are bounded; there are no
infinities allowed in physical situations. Any valid description of the amplitude and
Q. !1 _1 A .c. J:_
HUh "'" U" "'''0'''- "uu"u '" u" pv"u". ru "uuu""", '''''''' .. ," u,,,u,uuuv,, v. ""U'o"
,. ~ .. rr~w ~n ~.~rl;+;~rl h • "rlnr' ."..1' . .• ~ ~1,
, 'J ..... ••• ~~J~
(,f"nt
e mtezra orms 0 tee uation are IS e more or re erence rna tor Doten-
rial aPr In LnIS text, 1 ney are, uvwevta, LO establish oounuary con-

ditions (see Problem 1.7). The integral forms of the curl equations are readily
derived from the differential forms by application of Stokes' theorem. This theo- 14 Chanter I The Tools of



_1 .C C. .. 1.' c
."' ... ",v .. '''' •• o ... '" "' .... v. " ,,,,,,,.v. , ... , v,"'. "H a.",,, UlLV a UH" V.
£ • rI ,J., ,C ,J.,
.... ~~ .. " .. " .. ~~. ,~ .. ~ ... ~ t'~., .... ~,~. J~ ... ~ ~.~~.
J (V X A)· dS = f A·dl (1.5)
area loop
where dS and dl are unit vectors oriented normal to the surface, or tangential to
the loop, respectively. For the divergence equations, Gauss's divergence theorem
or a gt:I1t:lii Vt:C 0

r r
I
·a~ v . 'av .t
J
closed surface volume enclosed

relates the two forms. The integral forms are often used to describe fields with a
high degree of symmetry. In optical waveguide situations where field values at an
individual point are required, there is rarely sufficient symmetry to use the integral
form, so the differential forms of Maxwell's equations must be used.
1 ':l ~nl\l~TITIITIVI= ~I=I ATlnl\l~


The IX densities U and H. are related to the nero amnntudes ~; and H hv the
consTiTUflve retations. 1 ne nature or me meaium uennes me runcrional rorrn or me

". _I' LL
., '1J •• ·v. UH"a., 111"'Yla, LU'" .' a.~ "uu!'.,)' "I Y~ll u:
B = ILH, D= EE, (1.7)
where E is called the electric permittivity of the medium, with units of Faradsl
meter, and IL is the magnetic permeability of the medium, with units of Henrysl
meter. A linear medium is one where the permittivity, E, and permeability, IL, are
independent of field strengths. The assumption of linearity is valid only for low
'if''' ",hf'rf' ]?, i c for f'Y"rnnlf' rnnrh If''''' th,," thf' r0111rn"h fif'1,k th"t
.
hmd electrons to the central nucleus. Smce these hmdmo nerds are on the order ot
10 non inear o~JLici:l e using igl-in ensilY ig
u v cm ec s are 0 Y 0 serve
, T .cc : O. c . .0'
~H~"L" "all U~ .v, YallUU" , allU we W III UCal W HI.
,1- 1 • ¥ ~L " . ;~ ~ I;no~¥ ,J., J.,' ..1"
.... J .... ... ~.~ ~.~ .. ~
fields to be distorted by intense fields. (However, under extreme intensities, e.g.,
> 1023W/cm2, it is possible to spontaneously create an electron-positron pair from
vacuum, so even vacuum cannot be considered strictly linear. Such field strengths
are well beyond the realm of interest for present electrooptical devices.) The
vacuum values of E and IL are symbolically denoted as EO and /La, respectively, and
have values
.,'>
Eo IU~;)4 x r-ara s/m 0
,LU,
·7
fA{) ,+'11 A IV racru y I>/lli

I ...... , ...... or!; <> tho ... o ... or<> ov"'ro~~;nn~ £nr ;,,; .. , " ... ..1 "''''r ...... '''"h .. I; ... "r",
, or r -J r -J
not necessarily scalar quantines. Since the field quantities are vectorial, the consti- J.3 Cons titutize, 1§



•• ~ ~~1 .• n,~· ,. J.. ~ '1"'1, .1. . Q LL _Ill •
u 'J:' ~~ _"v~~ vJ ~ ~"~V" ... ~ ~.~v .. &V .. ~~ ~~"~".' ~, w
n .. ~n" .. l" rI"~(' .. ;h,,rI h •.
r -r -J "J
n, = EijEj = ( ± EijEj) (1.9)
j=!
Einstein's notation of repeated indices is used, and Eij is the permittivity tensor. A
similar tensor expression exists that relates the magnetic flux, B, to H [6]. The
.c ~ ~ ,1, . + .~ • ,1 ,., . ,1 ',1. 1
v. ~IJ v ..... ~ p.vp"" U~~ v .... ~ .. &~.~ ...... '-'. r o .... o , ..... v .. ..,-
..... ,,"v ~F """,, .. ,,11. h""" .,,"~~ .. ;., ;,,; .. U~ ... hh o ....
o· JJ 'J '" -oJ r- -J' "0' -J ~J '!_U __
struc nres suer as Na or ~1 con an l~ such as o lass are
-
iso ro ic in ermi tvttv uruess e svmme rv 0 eir s rue ure IS er ur e

Lll1UU~ll <l "U<llll.

Example 1.1 Permittivity Tensor of Two Crystalline Materials
Two optical crystals are shown below, GaAs and calcite. GaAs has the diamond structure,
"nn h,,~ "n isotronic ner mittj vitv Thp thr .... ,.,rv~t"lIinp ax ... ~ "rp ,.,ho~pn to Ii .. alono th .. thr ....
-
\IUU/ axes VI Ult: crysun, r uc permitttvtty leUSOI is gtven uy

" CL f\ f\
~ b-'V V V
C c. { 51> ( 1(1)
-v , .
'In
If the GaAs crystal were rotated 90°, it would look and act exactly the same as before the
rotation, so it is understandable that the different polarizations all see the same permittivity.
~
r-. "', <010>
~~ "
" ,,::---..1 "
'\. \ ~~
'" "- " 7R " \
<, <, '\l "
<, <, -- "
",<lUU> ..a:.:::::::...;
... nn1,> ....... l&.
GaAs Calcite
Figure 1.2 GaAs and calcite have different regular structures. GaAs is
a cube which looks identical along each axis. Calcite has a different-
length crystal axis along each direction, and the axes meet at non-right
angles.
r>~1 .• 1. • .1. l'; T+ ,1.' o{\O
JJ ,~ , _. .... ... 0"_" • " V_&V''_ OJ OV,_.V&& JV
"hf\l1t th .. r axis the crvsta wi' look {mit .. different to a hp"m of Ii oht . thrrmoh it
-. -
rVI 11UI:'<111. VVl<111LI:'U 11gUL VI 111 UII:' z in \'o<1I\.OiLI:', LUI:' <'1<,\.OuIC nero can
~ln' • A .l' 'T't.. • l' .l'. l' •
J t' , _J' '~J OU
identical, and IS called the ordinary index of refraction. For polanzed light along the z 1A Chanter 1 The Fundamental Tools of Ontoel



• ,.I, .e .& • ,.I '",," __TJ. .. .. ~
~'" .... > .. ~ ~> >~>.- ,..~ .... ,. ... ~~ .... ~""~'" ... ' J '"~~''' •• ._ t'~> ....... ''';] lVl
""lritp i~

e·75 0 2U
E = EO ~ 2.75 (1.11)
0
Calcite is called a negative uniaxial crystal because the extraordinary index is less than the
ordinary index. We will deal with crystal optics in later chapters, especially when dealing
-, .., _, ~ .. -,
WIUl "l",",UVVl'U'"' L~ •• Vl a 5VVU 1""'" V. '"'. r~'aL Vl'Ll"~, ~"" UI IilHV
<>nrl V",h r7
•• 1




In an isotropic medium, the permittivity is independent of orientation and is de-
scribed accurately by the scalar relation D = €E. But beware! Isotropic does not
necessarily mean homogeneous. The permittivity can be a function of position,
E(r). In an inhomogeneous medium, the electric field will encounter a different
permittivity, E, depending upon spatial location in the material. A graded-index
" .,,'..l ..l' ,,,,,~,,,rI;n rh ,."'r, "~n..l 7 . rI _1 ,-F
~o , "f- , ~ . __ --" ~ ..... -0
t-or most onnca ie ectnc matena s IS e ectrve ,., We can IQnnrE" rna o.,
~
neue Cl1C~l~ except W lIt;I W III ~I-'t;~ldl rc opttca materia s suc as
('vu", • _1 • 1. ..
y 1" JUJU HVll eaJ 11'" .. ,'-',, U~"U cro all vl'""a •• "v.a.v. "'''1''''''''11 va dllU
sources. Unless otherwise stated, it is safe to assume that the permeability, JL, is
that of free space, JLo. We will discuss the frequency dependence of JL and E in
Chapter 4.
1.4 THE WAVE EQUATION
eiec romaglleLi~ wave equa io tree LV s e ua ions.
comes rom maxwe er-
.c: :c . . .l .
.,aUUll .ao " .. ' .. an> 11 '" cs.o o "'"" ... a. a." .',. UI-'Ll~dl
.r. ,.: T rI;.: c.
' .. l' _U' ~. . .. ~ ~ ... _ ~1'~'''''''o ..... u~ 'l.X. J '<-Co
(n o 1 0) lin ea r (~ arid "rE" . _nf__R_and_ _H1_ _and_ ;~ntr'lrl;"
> , or
medium. Equations 1.1-1.4 become
aB
VXE=-- (1.12)
at
aD
VXH=- (1.13)
at

v 'U U 14\

U.J! 11 1'"
\ ... ."
r!-,p"", "irnnl",_lnnvino ",rm"t;nn" rnrnnl",t",l" il",,,rrih,,, th", "'~ ;" fi",lrI ;n
-r "0 -a- -z- -J --
time and positron. Are the assumptions reasonable'! Sure. At high frequencies (e.g., 1.4 The Wave Eauation 17



<, 1 n13 U~\ +~oo ~1. ,~. o rI ,~"o ,. ,1J, .... .s: -
"J' '0' " ~.~ ... " ... ~ u~~.vv ~.
npt;" pnpr,,, 'hp tun;",> nf nnt;",> pnpr"u '>rp pjp"tr;" nr ...,'>"npt;" r1;n 10
-OJ 'J r or 'OJ -0 'r
formed by atoms and molecules undergoing transitions. Maxwell's equations ac-
count for these sources through the bulk permeability and permittivity constants.
Equations 1.12-1.15 are strongly coupled first-order differential equations. To
decouple the two curl equations, we follow the usual technique of creating a single
second-order differential equation. First take the curl of both sides of Equation
1.12

-dH -d,cH
V x IV x~) - V x -vx .u.uo,
o, dt

Assuming LOaL J.L~T, l) IS . or Lime anu prrsitiorr; cquadon _!-!Q_[)_ecomes

V X V X E = -IL( V X a:) (1.17)
Since the functions are continuous, the order of the curl and time derivative oper-
ators can be reversed:
~
VXVXR -u-(V X H) .u.is,
~t


Substitutmz V x H - iJU/iJt into Euuanon 1.11 and assummz E IS tune _ .mxanant

I
all
V.x.v.x.~ f-Lai\at)
(1.19)
a2E
= -ILE-
a?
Now we have a second-order differential equation with only one variable, E. The
(V X V X) ooerator is usually simolified usinz a vector identity

.,,),
v.x.v.x..r, v~ v • .r,) v-~ _U_.Q,Jl_

.,.., ..,2 -, ~ .c. , T ,
• U<O Y Ul'<O"UVl "UVUIU llU. LJ<O ~UIHU"<OU vv lUI UI" " .... UT LUp.UUUfL ul'<Ola.uI •• 11<0
u2 ;n P nuo';"'" 1 HI ;~ .1.0 T~· .1~. 'm ,1.0'
'I:' -r - ,~ -".,
in this case E. For a rectangular coordinate system, the vector Laplacian can be
written in terms of the scalar Laplacian as
V2E = V2 E x + V2 E Y + V2 E Z (1.21 )
x y z
where x, y, and z represent unit vectors along the three axes. The V2,s on the right-
hand side of Equation 1.21 are scalar, given by

ri'" ri'" ri'"
UL (1 ')')\
:;.,2 :;_ ,L :;.;1. , .
JJ .~
in (,>Irtp<;:i>ln rates of thp vP('tor W>lVP ermarion rermires that WP fir-er
( . .
break the equation into the orthogonal vector components, WhICh IS sometimes 18 Chanter I The Fundamental Tools of Ontoelectronics



:F;:: .1, .1. -1 ,1.. • -1: • ;:: .1-1 .1,
VY'"' '":'" , "U~ "'~u .u'" un .. , "",y", '''''''.v ... "'" .. "v.yuv .. o
t, ,,,,th,,,r
What about the term, V . E? It IS not necessarily equal to zero, as is often
assumed. We know only that V . D = O. Simple calculus leads to an expression for
V'E:
V'D= 0
= V·€E (1.23)
-u",.1<' ",u . 1<'


Solve or V • ~;-


- VEO
V .~ .~. 1.,(,"1-,
e
Plugging this value into the linear wave equation for electromagnetic waves yields:
2 a2E (VE) (1.25)
V E - f.L€ af = - V E'-;
Th .. r;oht_h<lnrl ~;rI .. rI .. c .. r" .. c "n .... i<l1 t> nn"irl .. r<ltinn It j c nnn_'7 .. rn ,"h .. n th .. r .. i c <I
'0' <r-
!!ra, uem III the lVItVOr ie me rum. Suer zrac tents are not m
lZui e -wave 0 ics. vvi e exce ion 0 S e -in(]ex waveznt es mos zm e -
_J . . C' L .I .•• ·.L ."
""av", "u ... "' ..... "''' ... ,,'" a 5.a",,,,,,, 1''''.HH«. 'u s : UV HUVV ",v '''' u",a. ""UH LlU" ",,,,ua ''''HU.
n U. tJ.. t. • :1.. .11 ID • 1 "
,~U, ,~ 'Eouv,,,, n .• v. ,,,vU' uu Y' "Y'~U, ••• '" ...... u u ... y ... ".vv."' ... ,._
explores the limits of V€/€, showing that it is almost always negligible.) Neglecting
this term, the wave equation reduces to its homogeneous form:
a2E
V2E - f.L€- = 0 (1.26)
at2
Had we started with Equation 1.13 instead of Equation 1.12, we could have derived
.F ;:: I. D.
y" ...... y. '" '''' ~':IY""vu U •• "'.ULU V •• u~ u~.~ ".ut'U'Y~~ 'u"'~ ... vv'~'"
1.lm,
,.,
crH
V~H - U€· -0 (} 27)
ilt~
"
1.5 SOLUTIONS TO THE WAVE EQUATION
Consider the units of each term in Equations 1.26 and 1.27. The V2 term has units
of L'(distancej/. The second-order time derivative clearly has units of 1/(sec)2. In
order to make nhvsical sense. the unit" of u.e must he ("ecim)2 We will "how in a
. r;-;-- .
ater sec ion a v €JL IS e ase ve ocilY 0 ig in a meo ium. rvo ice aL
~ . , , . _L . , . '"
UIC 1>l1ccu UI 11> UC'CI Ullll<OU uy 1I1<O 11<11<1111<0'<011>. HI 11<0<0 1>11<1'-<0,
crtt. - "OOQ v lOS .I~"'P , .. r» th", ~~"''''~ yf F ,ht ~n ,P11 I~ f'T'h ,~ .f
, r-u-u , , -t' -0'
ioht j" nnUl rI .. fin .. rI tn h .... Y<I .. th, ?OC 70? LI.,\R ral « .. " rh .. .,.., .. t .. r i c rI .. fin .. rI in t .. r.,..,,,
'0' 'J
of the speed of light, being the distance hght travels III 1I21)1),7n,45~ sec. ThIS 1.5 Solutions to the W_aye_E~

HI

section.

Equations 1.26 and 1.27 are vector equations. They can be simplified by rewriting them in terms of the components of the field. In rectangular coordinates, the vector Laplacian breaks into three uncoupled components. The scalar component equations become:

01

Here e su senor in icates e 1 com onen were 1 stan s or x v. or 7. ana

...,? . T....· T"""I'. ..

v U; lilt; 1>~,U<11 ~l Vt;II III D4U<111UII 1.4-4-. o uice LIlt; symuo or e vec or

and scalar Laplacian look the same, we rely on context to distinguish the operators.

The choice of coordinate system is critical to solving the wave equation. For example, choosing rectangular coordinates to describe a wave in a cylinder leads to component coupling upon reflection at the cylindrical surface. The cartesian components are inseparable in such a system. When a coordinate system can be found with no coupling between the orthogonal components, the individual equa-

-r'P --, , -, .~ ••• - ... ~ ... ~--, -':l~~-

t;nnQ .,Q thp ",...,I.,r '"<H'P prll • .,t;nnQ In ,...,rtp,,;.,n thp crvc l o r- ",.,,,p

... ,

p(lI1>1tinn is written as'

(1.29)

where '" stands for anyone of the orthogonal amplitude components. To find a valid solution, we use the separation of variables technique to get:

IjJ(r, t) = IjJ(r)</J(t)

0.30)

III" exnr TK' rl exnt ox: + c.c.

'u -a- -r-

(m units ot rads . and to is the aneular treouencv of the wave__Lin___units_o

raos/sec). we Will use me wavevector as me pnmary vananre in most waveguide calculations. The magnitude of the wavevector is defined in terms of the angular frequency and the phase velocity:

Ikl=w~=k

(1.31 )

The wavevector k noints in the direction of travel for the nlane wave The mao ni-

LUUe;: 01 IK uescrrnes HOW muc pnase accumu ares as a DIane wave

rave sa tmtr

"",'-' "'. ,.

1.3 shows the two cases.

20 Chanter I The Fundamental Tools of Optoelectronics



I \ I \ I \ I \
\
\ \
V V V V z
e Urot + jkz) e Urot - jkz)
Figure 1.3 The general solution to the wave equation in a linear
homogeneous medium leads to plane waves. Depending on the relative
sign, the wave will travel left or right.




n 0 tics It IS common to escn e onuca eic S DV t eir wave enztn. I.. onsi er
me wave ill Ftgure 1."t.


i- A -i

Q) (\ f\
'0
.2
0.. r \ I \ I r
t::: I \ I
<l \ \
I \ I
,/ ,/



Figure 1.4 The basic description of the wavelength is that the
wave accumulates 2'IT of phase after traveling one wavelength.
The waveform in Figure 1.4 shows the real part of the spatial component of the
plane wave, I/J(r) = I/Joejkr. The distance between two adjacent peaks in amplitude
i" (''' lIprl " ,m ,10'1oth ,\ Thp arrmlirnde of thp w"vp "t thp fir"t np"lc pjkrJ i" thp
u , '.
same as the amnntude at the neak ocate one wave enztn away. er= "'.


e' 'i e- "' 'I
_ pjKrJplKA


This equality holds only if ejk)' = 1, which requires that kA = 27T. Solving for k:
k = 27T/A (1.33)
This is the expression for wavevector k in terms of wavelength, A .


.... ... . ". ... ..... ..... , ... '" ...... ...
................ ..... . .... ~. I....... ... ..... .................. .... . ............... ~ ...

,.-'"nciripr " nl"np "'''''''' "f liaht ",ith - 11m 'hp lioht i. riirp("'tpri in thp Y. nbnp "t Ll'i0
r -0· -0· _ r
Descnbe the magnituoe ana cirecuon or the K vector. I 6 Transverse Electomasnetic Waves and the Povntinz Vector

21

T'he rnao nimde of k. i s:

27T k=A

27T

= ---,--

10-4 em

L"I 0'> 1

0.34)

Nntp. thE'. rmits- rarlia ns npr

.

e K vee or can e roxen own In 0 1 scorn onen S In

e c osen coor ina e svs em

U_A ~O A ICO

= k 0.707(i + y) (1.35)

= 44,421(i + y) rads/cm

a .- .- a -. .- , •... __ ._

1.0 I ,yt:n.:u: t:Lt: ... 1 n' 1I .. t: II ... VV",Yt:~ "'I'll'" I nt:

.. " .". II""" " ... "'." ••

fjplrl i.: nnl:lri7prl :llnno thE'. r :lyi.: I{' r t rF. rn.:( (olf - b·\ In rornnlpy nntMion

.r- ~ v ,

ttus would be described as

E(r, t) = i Eo ~ (e-j(kz-wt) + e+j(kz-wt»

(1.36)

= x EOe-}(kZ-wt) + c.c. 2

t e unctiona orm. we must e care u to take t e reat oart 0 ex ressions IKe

::'L ua '0 .-'1: wnen we want 0 II esc] iue a PHysicai wave.

plitude using Maxwell's equations. Plug the electric amplitude (Equation 1.36) into Equation 1.1, and use Equations 1.7 and 1.30 to show:

k Eo 'kz'

H(r, t) = y - -e =) e}wt + C.c.

/LW 2

A wv-;;:€ Eo +ik: 'wt

Y "I "" l..L.

\ i.J 'J

- v e J"~eJ T c.c.

11 L.

wnere TJ is caueu me cnaracrerisric impeoance or me meoium,

22 Chanter 1 The Fundamental Tools of Optoelectronics



r=

,.,.,- .. f1 'Hl\
\I \ . '/
-
In vacuum, the characteristic impedance is 7]0 = 377ft Thus we see the magni-
tude of the magnetic amplitude is directly proportional to the magnitude of the
electric amplitude. Note that E is perpendicular to H.
A useful concept for characterizing electromagnetic waves is the measure of
power flowing through a surface. This quantity is called the Poynting vector, de-
mec as

~ v II '1 '1{h
~ ~ .... \,.-' /,
S thp ,I: mrensitv ( W Im~\ of thr- w~vp 'hp vprtor
L J ~
DOimS 10 e irec IOn 0 ower ow w IC IS e en(]ICU ar 0 0 (I e J'., an It
fields. The time average intensity for a harmonic field (i.e., sinusoidal waveform)
is often given using phasor notation
1
(S) = - Re[E X H*] (1.40)
2
where Re is the real part, and H* is the complex conjugate of H. The total electro-
_.I, ,c. c .L
pvvv"" LUV"UL5 uuv" ,Vu,uu", a o U"'."'UU'U"'U v. ""u, '"'''' uU"'5'''' v, .u"
D, ,0, 01, .f' .l- .1. r\f'o, .1,
'~J"""o '~~.~. ~ ~ .... ~ ~ ......... ". J< ... ~ ~ ...... ~. --.~~ .. ~ ... ~ ~"'J • ... ~.~~w~ ...
thp, ~vpr~{Jf'. p,np,r{J" flow in one rlirer-t ion f'.{J thp, riower rrol:I:in{J ~ tiiplprtrir in_
~ ~ ~ . ~
terrace, n sucn cases ne ot nro ct 0 the Yovn 102' vector wan me 011
vector must ne evaiuareu, e.g ..
1
(Sz) = '2 Re[E X H* . z] (1.41)
This value of (Sz) is valid at only one point in space. To calculate the total power
flow in a waveguide of finite extent, where the values of E and H vary in position,
... • L 1" ••.
n 1" y LV L11" I V.'11U11l;; '''''LVI VV"1 Ul" "IV"" """UV11 VI Ul" l;;UIUC.

.. .., ... LJ A ........ , ..... ,..." ...... "
• r ."' •• ""~o;;;. y t •
'1". ,l- . ,.: .1. :.' rI, ,~,:h oh .r, .f' ,1
. .. ~ .. ... ,~~ .
These are the phase velocity and the group velocity. We will consider phase veloc-
ity first. Consider the sinusoidal electromagnetic wave plotted in Figure 1.5, trav-
eling in the z direction. A point is attached to the top of one ofthe amplitude crests,
as shown in Figure 1.5. How fast must this point move to stay on the crest of the
wave? Since this crest represents a specific phase of the wave, the point must move
at a speed such that:
;flr .of\
e cons an l .Lf-,L

,1,' , .f;, rI ~f' To ' • f.\ ' of' ••
.... ~ ... ~ .. ,." ~o ••• <U ~"uJ ~~~ "\0/ '''''~o ~""~'J'

wt
7( t - T constant .4i
k 1.7 Phase Velocity 23


',-
.... \', J
"......._ •
/ \ /
\ I
~ II \ / \
:E
c.. V V
E t
<
'IIY'" ... • "" l'"a,,, ."'v,",,,), " u"'''''''''',",u u)' "''"' 'l''"''"'U
necessary or a point to ride the crest ot a wave .



" s i cc, a ..I.L .1
.. " "'all "V) VV lUll ""}I""" ,V <1111.., ,V lll1U Ul'" }llla"", 'J' 'V)·

dz w
dt = k = vp (1.44)
The phase velocity relates the angular frequency to the magnitude of the wave-
vector. Also, recall from Equation 1.31 that w = k/~, so:
v, = llV~ (l.45)

'T'L' .1 ..I, ..I' n 1 'JL .L ..I ,£ 1.'
1111" .ro <11v ",""v '~'~~", "un vvv UV11'VU III ""'iUUUVll 1 .JV, "V U"" "1''''''''"' Vl U6'"
.l, .4' .l, • .l, ,l, ,I, H .. ....
'uu, '"'~.u~~ ~ ... ~ •• u~ 'U'~ . .~ .u'"' t'uu~,", ,~" "'J' •• t'~. uu ... '''J ~ -u-
thpn i\: Ip\:\: thnn r thp, \:nppr! .... f Iiohr in ~ V~('I111m Fxppnt tor unusual cirr-nrn ,
l' . ._, .
stances sucn as nronasation In oiasmas most matena s have a nerrmttrvrtv, E hat
is greater in magnitude than EO' Do not be concerned if the phase velocity exceeds
c in certain situations. Such instances are results of collective action by an oscillat-
ing medium.
We define the index of refraction, n, of a medium as the ratio of the phase
velocity of light in a vacuum to the velocity in the medium. Using Equation 1.45:
c
n .'1-1:
v"

Ul, leI Ul1> Ul Ule luau:;! 1(U ~l oper ue~ UI Ule

",/
n = .!:::'.._~
~ (1.47)
=~ when IL = ILo
The index of refraction is an imnortant narameter in ootical desizn and material
cnaracterization. we wif exmore ill; ue encence 0 wave engtn i rater cnapte s.

"" . .11 ..I "'L . £ £ .
.1 ue lauv 0 q 1" ",all"U Ul" "''''''L, 'L LUlt,HUlt, • .1 llv l11Uv'" VI Ivll""LlVll, It, 1" UH"
'rn' '''0 ,.~ ,t ~f' tho rliolo"t .. i" "~nct.,nt
....
Wp .... ftpn ",ritp thp I< in tprm\: .... f thp uar-n ,m wnllPlIPrtnr k, ~nr! thp
v'
Index ot retraction. The vacuum wavevector IS the magnitude ot the wavevector In 24

Chaoter 1 The FurirlameniJil ThOlsof Ontoelectronics

k=W~=w~~=w~n=kon

(l.48)

Once we know the vacuum wavevector, we can define the magnitude of the wavevector in all media based on the index of refraction .

..., . ~ " . .

1.V 1LCO, uie 11 ell ueucy, W, Vi a piaue wave i~ 1 e ItIca In a

..I' 'T'~' c. c. _ £ "" _ " _ 1 ~.

o e plane wave. e k vee or sea es as k k"n III a me ium WIth I~I~( tru-

constant n.

1.8 GROUP VELOCITY

Except in regions of high attenuation, energy in an electromagnetic wave travels at

,I, .1. T. .£. .L • _~ • • ..I

,uv 6' v ~r ,~.v~ '7, g'~u "UvU'" " • .u .. "u vy vu a~

expression

or If e group ve ocilv rouz a sune osi IOn 0 two waves WI

1 -

r- c.

'CO'C01" .. coy

.. " ~

n llil ure

WI = W + ilw,

W]_ = W - ilw,

(1.49)

the two associated wavevectors will have values

kl = k + ilk,

(1.50)

• • .J .J

V"U vv uucou

1 -~v -,

.~

cos W i.l.W

.« ~k'Z

Using the trigonometric identity

2 cos x cos y = cos (x + y) + cos (x - y)

(1.52)

the electric field superposition can be rewritten as

..." • • • ~ • ~~ r

1. 11 .rc ' "upco1. Vi t'VV WiiVCOI1 iil 11 COli lCOiiUI1 LV a ':;1 IJVliil oear

• - J >t' t- '0" JU<> ,~.. "uP'"

.. (\f thE' hu", u,"" .. c rh .... n" .. I",n .. ",f th .. "n"lnl't"rI .. "I .. "rh, .J...~ h~n'

trequency.

1.8 Group Velocity

25

\' I

/

v

1111: I:IIYI:IVPI: VI 'Ill: near rraveis a, 'Ill: glUup '1:1V"Uy.

The group velocity is the speed at which a pulse, or in this case, the envelope, travels. The envelope is described by the cos (dwt - dkz) term of Equation 1.53. We can again attach a point to the crest of the envelope, and ask what speed, v(t), is required to stay on the crest of the envelope. Following the arguments used to derive the phase velocity, we set the phase argument dwt - dkz = constant. Solv-

irio for 7(t\

~wt

zt t) - T constant

(1.)4)

lilt: gruup verocuy I~ ure uei r vall VI;; UI uus: d: dw

V g = dt = dk becomes

dw dta

lim -=-=v

A~O dk dk g

(1.55)

The group velocity, vg, depends on the first derivative of the angular frequency with respect to the wavevector. In free space, where w = kc, the relation is simple

aw

an eads to - c. n a vacuum t e Dr ase ano 2TOUD vetoer ies are 1 en rca .

dk

e re a ion is more com ica eo in 0 e meo ta: I e 1;0 rstitu i ve 1;0 ista is, es-

." "

LJCI;IC:lJ l V to, U~U<111y VlI 1.

_1

(1.56)

c

=---

dn n-AdA

e as re ation can e con rmeo wi a sim e ca cu attorr; e g ou velOcitY is

,.,_ ._ ~ '-

lIC<111Y CYU<1I LV lU<O }111""<O Y<OIV'-llY, UUL 1" Vi UI'-iv""vU uy " "iuau l<Oi Ui

velocity IS less than the phase velocity, c/n. Optical materials display regular

26 Chanter I The Fundamental Tools of Ontoelectronics



-l: .L .L .l .l
v,,,1-'''''''V'' .... v ... "uv ..... u"u ""U"1-'"'''''' ''',,'V'''', "V """.". ""V\_.,, ".v ,,_. UI<lll UIC
~h ,~o n;~~ 'r~;' '.~ ,r, ..1. ,nr;h ,..1 • • r. ;1 • r~ .•.
I"' ~I"' r" -I"' - ...
1.9 BOUNDARY CONDITIONS FOR DIELECTRIC INTERFACES:
REFLECTION AND REFRACTION
When two different media are adjacent to one another, the wave solutions in the
two regions must be connected at the interface. The rules for connecting solutions
are a lee rv cona. tions. In zenera 1" there IS an mdex
IWO meaia, mere will oe a renecrion. 1 nis is cal!t!<.! a _! resn~ r~nec_!ion alter the
". _L T " _. ",",00 10"',""
J.·J.vll"ll ""J.vllU"', '"". J. J.J.v"ll"'J. lJ. 'UU·-J.U ..
ro. .+. . c: 1 '7 Tl I, .+ ,I
.,," un" •• """ ""v ........ " .... " •. . ••• " .. '''''LV' V. "" ".,,"
\1'"'''' rvr .... n<oo<>tP"" fr .... rn .... ru= inl .... (<>rr .... rnn<>ni"rI h" <0 n<or.'" ro~on
.r- - roc> T -j .r-
tion back into the originating media). The wave has frequency w, and is incident
on the interface from region 1 at an angle of incidence, (}i' The two regions have
indices of refraction nl and n2, respectively. We want to determine the amplitudes
of the transmitted and reflected waves, E, and Ep and their respective wavevectors,
kt and kr.


K
""" . L
<, <II r
,,/ n ......
._, .. '".
c: " It
r ~ ~
.;Y at ~
t
k
C,' .r
/"-
"
1-1.
J

I"'gure 1.1 A ray IOClOem on an IOtef!_ace at allgl_e
flo will r .. A .. ct :lnll r .. fr:l"t into two diff .. r .. nt ravs rh ...
electric field in this figure is directed out of the page
for all waves.
We must first solve the wave equation (Equation 1.26) in each region. It is
straightforward to write down general solutions to the wave equation on either side
of the interface

... ... ,ilk,· r' rut) --
LI , 'J LIt:: ,1.,) J

uThoro 1<'.'~ tho <1"",nl'tnrlo. Tho ~nh~"r'nt 1 ~~h"lrln .h,t
-, . -r" , ,
u,i· <>ri" .. 'h" tmroh n<>rt .... f th .. nr .... hl"rn is r(lnn .. r-t ino th .. " .. " .... hu i .... n" <>1 tho intor_
~ . . ~
tace. The boundary conditions that apply to this situation can be derrved from the 1.9 Boundarv Conditions for Dielectric Interfaces: Reflection_and_Refractio

27

,1 ~_ .~ llK, ,11' D.

HU"L ~L HL"~ ,vu ~ v':I""U~u~ •• ~L LV <LV , L" "

s X (E2 - EI) = 0 s X (H2 - HI) = 0 s· (B2 - B1) = 0 s . (02 - 01) = 0

tangential components ofE are continuous; (1.58) tangential components ofH are continuous; (1.59)

normal component ofB is continuous; (1.60)

normal component of 0 is continuous, (1.61)

nere are wo OSSIOJe orientations or t e electric eJG WIt resnect me

me: race, Ilt ieu ca ue .uia or para er e tuane OT inciaence. e

perpendicular to the plane of incidence, it is called a transverse electric (TE) wave. Figure 1.7 shows the specific case of a TE wave incident on an interface at an angle 0i'

Inspection of Figure 1.7 shows there are six field amplitudes (E;, Et, E,., Hi, Ht, Hr), three wavevectors (ki, kt, kr), and three angles (0;, (Jt, and Or)' Some of these, like E, and 0i, are initial conditions of the problem, while the others are dependent

v<>ri"h]p" It i<: tn fir-er rphtp thp "n(J]p nf in('irtpn('p ~__anale___nf_

TP.1iP.( inn'

tI; o;

. ,~

J IS SU -,,,r .waru: we can apply r-erruai S PllIl(;lple ~c IVW<Olll l.<t), or

conservation of photon momentum (Problem 1.5).

The general description of the x-polarized incident electric field is:

(1.63)

The wavevector k, is described in terms of its vector components,

K· [2' cos fj; V SIll fj; Iknn

.b4

wuere Ko is uie vacuum wavevecwr, oe eo as K. WlC. osi ion r is a so escri e

... .

HI V<O~lVI IVllll.

r=.x.x+yY+zi

(1.65)

Substituting these terms into Equation 1.63, the complete description of the incident field is

E;(x, y, z, t) = xEie -jkonl(Z cos IJj-Y sin IJj)' (x.HyY+zi)~wt

A ;kH,(, r- n' {I, ,in {I,\ ',.t

(1.66)

.,., .~ C_1.1· _1 _1 .• C .C C,

111<0 l1"'lU '" "'Vll1Pl<Ol<Ol; 111 l"'"11" Vl Ull"''''UV11, """:I.u"'""'. , UllU pv-

• .,. -.I " -rr 'r

it is thp. ""rnp in" rp(Jinn" 'hf' nthf'T f'1f'C'tTiC' np]rI" in Fi(JuTp. "rp

simuarty described.

28 Chanter 1 The Fundamental Tools of Optoelectronics



.... r . at: -jk,· r
~t\~, J, "I ~~1" h'

- ..,1<., ]KOn2 z cos tI, ysm tI,
ErCx, y, Z) = xE,e -jkr• r (1.68)
= xE e-jkonlC-zcoS Or-ysin Or)
r
. We have assumed that the electric field will continue to point out of the page for
each component. This mayor may not be true: in some cases the phase of the field
.1. , OAO ., T" "' .. .L
au vall""" VJ ~uv , allU lll" UU'-'''UVll 'vu~u .~".'''''v. U Ull" , '1110011 WV 11aVe
...I ~1, .40 :11 .1.' _1: ...I
~. ~~. , ~ .. '" ~. ...'" "'~"'Y~""""~ , ... u'" T U J .. ""'0 .... ,,-
"ion Thi" mp"n" f hnt UlP c1icl nnt nirk thp rnffPrt nfi, initi"lh, Rnt it h"" nr.
~ c -J
conseouen ~e so 10m' as we consistenuv annr tne nounrtarv conrtmons and oeo-
._.
meuic JJlo'ecdons in ou so trttorr; ;,0 o nor ue 00 concerne a ou c oosing e
proper orientations initially, as these problems will solve themselves.
We will need to describe the magnetic fields for the three waves. The appropri-
ate k vector for each magnetic field is the same as for the electric field. Note that
for TE waves, the H fields have two vector components, a z component and a y
component. The magnitude of the magnetic field is related to the electric field
through the impedance, YJ, of the medium (Equation 1.37).

H II( I'YI f.Q\

r=r:
wnere TJi v /-Ufoi' U sing rr cy, uie l,vue"t eAI-J1e1>Mvl. lor me tt nein com-
YU11 ...... o ......
Hi - (Ei iYJl)(Z sm (Ji + Y cos (Ji)e -;"l"o\~ w. r71 -ysin VI)
(1.70)
H, = (E, /YJ2)(i sin Ot + .9 cos Ot)e - jn2kO(Z cos 0, -y sin 0,) (1.71)
H, = (Er/YJl)(Z sin Or - .9 cos OJe-jnlko(-zcos Or-ysin Or) (1.72)
With a complete description of the field in all regions (Equations 1.68-1.72), we
can connect the solutions at the interface yielding formulae for transmission and
n ~. , .. "' ~
leHe"UVIl, rll1>t, al1IJlY Ule tHat Ule VI L IU:.t ue
,l- .c:
,~u ~_.~uu •••• "'w'" .",.
;IX 1(·+ I( - '/ X I{ '1
.~

1 ne tangennai z new ar me interrace IS me l'..x componem. rsxpancmg trns at z - u,
and using the fact that i X x = .9 and 0i = On yields:
.9Eie(jkon1Y sin Oi) + .9Ere(jkon1Y sin OJ) = .9Ere(jkon2Y sin 0,) (1.74)
Combining terms of equal phase:
y(Ei + Er)e(jkonIY sin 0i) = yEte(jkon2Y sin Ot) (1.75)

ror t IS e uanon 0 0 It must oe true or a va ues 0 v. At v - tee uanon
S1I pry:

1<.+ 1< 1< r, r.f "".' .\ ( 7f..\
I -J "0'
Sutrstttutmg mis imo tquaIion I.t:J ana canceling common terms yielOs:
1.9 Boundary Conditions for D~

land

konlY sin OJ = kon2Y sin Or

Canceling common terms on both sides we arrive at Snell's Law:

(1.78)

(1.79)

VI ure rs; UVUUUi:UY conui ion rovi es e irec ion 0 e ransmu-

_. 'T''-' • _. "'- ~ ~.. TT

I"'U wan;. ~ HI" Ivavv" UlllY IHv ,---r:t,---r;,., LIt, i:UIU £Ir lV ue ueierruuieu. lU

th", ., ",F 1< ;" ~+ r;o .. " _ •• n •• _+ +. +1.

r " VVUHU .....

"nnclitinn" ~hp "nntinllit" nf

(1.80)

In this case, H, has both z and Y components, so we must be careful to carry only the Y component through the cross product. Using Equations 1.70-1.72, Z X Y = -x, and Or = OJ:

Z A OJ

( . n.

tUB)

'r/

2

+ X E COS 0· tk. ( . )

z X H, = r 'e-} on, -ysm (Ji

111

where Ej, Ep and E, represent magnitudes, not vectors. Adding the terms according to Equation 1.80, and applying Snell's law (Equation 1.79), we get

(1.83)

(E- - K) cos O-/T1, - E cos ° tTl"

(1.R4)

omce ti, ti, + ts., we can replace J:!." inTerms ortne otner variables

E; _ (112 cos OJ - 111 cos 0t)

-

E, (112 cos OJ + 111 cos 0t)

(1.86)

Similarly, we could eliminate E; from Equation 1.85 and solve for the ratio E/Ej:

f

fl Q'7\

Q\

\ .v, I

\ IL

tJ

, I' ~.t'"

rewritten as:

n· or a materia rne caret to rnsnno rich 'YI frnrn..,) Ifu thpn 'YI "<In hp

30 Chanter 1 The Fundamental Tools of Optoelectronics



r=: -=-
,~, 'Th
'YO. ~ .. ~ ~ {I 10\
I \I \ ' . 'I
I I-U I
Substituting this expression into Equations 1.86 and 1.87 generates the more fa-
miliar forms of the amplitude transmission and reflection formulae for a transverse
electric field. In these formulae, the field is incident from the nl side, entering into
the n2 side.
tr (M ",,~ tl ".,~ a \
r I -"- t II o/n
17 {. II ll\ '~.U7 J
.... , '''I ~v~ vt .~ ~v~ vtl_

F '11 ro~ H.
{I Q{\\
F. (n ('n" fl. +_1L ~ , . 'I
L ~ ."
The expressions for transmission and reflection of a wave which has the mag-
netic field H perpendicular to the plane of incidence (the so-called transverse mag-
netic or TM wave) are significantly different. Their derivation is left as an exercise
to show:
Er nl cos ()t - n2 cos ()i Et 2 cos ()i - (I nl \
1< ".-.~ tl -+ .. ".-.~ tl r:;' .c, _b_ __)_~ ,,-,,-A A ,~ .. ~,
I -1 t -L. I I -"-' 1/ I t

')np UT.-.r~ of ",mt;nn "h.-.nt thp Prp,,"p] f.-.r1Tlllhp· thp, ~p"('r;hp thp non ,1;'''~n nf'
-, ·-r
t e transmitte an re ecte eu nott e ower 0 t e neu s .me nust ne caretul
STIle, S HI e c rcurusta ces , Ilt: eo LIle ans nitre e ec ric eu can
be larger than that of the incident electric field. This dilemma is resolved when
total power is accounted for in the solution. In such cases, one can either rigor-
ously solve for the z component of the Poynting vector for the transmitted and
incident waves, or account for the geometric change of area between the incident
and transmitted beams. Problem 1.8 explores the power issues of these formulae.
We could also develop expressions for the H components, but these can be found
.1. _" .1. .J., ;I,
~uUP'J ~.~ ~"~~"J .... ~~ou ... ~ . .u .


1 ':l Nnrm .. 1 frnm .. r.1 .. "tc,

The simplest example of Fresnel reflection is that which occurs when light strikes a glass-
air interface. We experience this effect on a daily basis. Let's apply the reflection formula
to this problem to illustrate the magnitude of the effect, and the phase shift which occurs.
A beam of light is incident normally on a glass-air interface as shown in Figure 1.8.
What is the intensity of the reflected light if the glass has an index of refraction of n = 1.S?
<> .. 1, •• ' :

Plllaa;na nIlTnhf"r<: ;ntl\ Q( nof irro that "I\<.R i n th; .. "" .. P WP. apt
.,., ., . .,

t: It:.· .L. .1.J'l.
1 -t- 1.) 1.10 Thtal Internal Reflection 31




n 1 n - 1.1;


Reflected
Light -
Incident Light ...



Flaure 1.8 A beam of lisht strikes a !>Class
.~, .11. ,11
,~ J> '~"J' ,~, "'0 H"~"

The reflected amplitude is 20 percent of the incident amplitude. The negative sign indicates
that the reflected wave is 1800 out of phase with the incident wave. In general, when light
strikes a surface with a higher index, the phase of the reflected wave will be reversed. Now,
what is the intensity? Using the Poynting vector and the fact that I HI = IE "1], we find the
incident intensity is:
~?
'" ~ DO "',.,
"'inc ~ ,L7J.
'"

",h;1. th .. . n ~ ;nt .. nc;h, ;c "nh,·
-,/ .

II ·lH.)~
:::ire! 2 U.U4:::iinc (1.94)
YJ
Thus, only 4 percent of the incident power is reflected by the glass interface. This reflection
can become a significant loss in certain applications. For example, a camera lens often will
consist of three or more separate lenses, representing six glass-air interfaces. The total
transmission for such a system would be T = (0.96)6 = 0.78 if the lenses are not modified .
'T'~ . :<:. 1. ,f' ,1; ,J., • • J., ,11. .n;
-r ~ r '0'
ciencv IS critic a Not onlv would the reflections reouire arzer anertures and longer exno-
,"UH;; UI11<O", UUl l11<O'y <U"U "UUIU "U11UIUUl<O lU !511U"l 111la!5<O" U11 UI<O 111111. ~ 11<0"<0 PIUUI<Ol1l1> al<O
.fI. fAD .rf'~ 'T'h .D
'J r- '0 "0 '0
basicallv a stack of A14-thlck lavers of dielectnc matenal which interferometricallv reduce

UI<O lUl41 l<011<O"l1U11
1.10 TOTAL INTERNAL REFLECTION
An imnortant nhvsical nrocess in zuided wave ontics is total internal reflection.
." " r . .
ryc W 11~ lUU1\. al lUl,U l' BUIll lWU VCl"VC\"llYC". lay lla\"lllg euru ll.e
n. . " .. , .L .L _" . " ., ., .
,av", "'':I ... auvH. n.a; "a""H6'" ... ,,"'..... 'H"'H LH'" V. LH'" vi'''''''''
c r- .. Io r-o .. "roTnn<lr .. rI tro th .. uT<I" .. I .. ncrth rof lioht. R<I" tr<l"ino i" 11" .. fll1 fror
-0' -z- -0 -0' "'/ -0 -r
such lII: thp. 011mp.TWlI II 'hp. WlIVP. nic.tnrp. '.1: II romnlptp. npl:rrintion
. . . . .
01 the phase srurts ana evanescent nerds that accompany total internal renection. 32

Chanter 1 The

1 in 1 C.:a"

~ "" -'" -'" <r

The only action of an optical element is to redirect the ray. The angle of incidence

of the ray and the properties of the optical element establish the degree to which the ray is redirected.

The important operational rules for ray tracing are Snell's law

(1.95)

.~ th T .~ r: _J3

H.

( qt\,

tJ. .

-r =c» -'" 'r 'J 'J .""

lent references [8-10] elaborate on the application of ray tracing to optical design.

Powerful matrix techniques have been developed based on these simple laws which allow the engineer to design linear optical systems. The ray tracing analysis is usually of limited use for guided wave optical design, however. Its most common application is to describe graded index waveguides, and to define the numerical

.

" n.

y

,,,, racuor (onen s Law.

equa ,...n~le nenec Ion

.f't f;,

angle OJ rettection.

1.10.2 Total Internal Reflection Using Ray Tracing

Total internal reflection (TIR) is the phenomenon where light is completely reflected at a dielectric interface without the help of reflective coatings. TIR is often

_1 • -d to make ",ffir;pnt "rhrntTl"t;r·£1 For _1 r;oht_,,~nlQ nr;"~"

are 0 en use 0 recirec i!!llI rom rmaamz svs ems sue as mnocuiars lr

. ~. . . .

I>Cl VC <11> lul;;l;;cU IIlIITUUS ior rngn-powereu rasers. lilt: a ipncat 0 a in e es s us

~o ."' .• 'OL"'" .• v ."u~u"w~ L"" '''J P'"'U''' v, " "6,ll-alll;;lC

nri"tTl <>nrl nf " :rlQ 'hQ irQ' l'~r "T rD ;" th~ th . yl- 1.

r -0" J '1' '0'" 'U," LI'-'

incident on a dll~rectnc Illtenace from the high-index side. Thus an optical waveguide must consist of a layer of high-index dielectric surrounded by material with a lower index.

I 10 10tal Internal as



'\.
"'-
"'-
.• ~ ~n~/
r
Figure 1.10 Total internal reflection can be implemented in many ways. The right-
angle prism and the optical waveguide both use total internal reflection to redirect or
trao liaht. resoectivelv. Note that the lisht is incident from the hizh-index side of the
'., ~''T'TD




T'~tn 1 ' .t .• r, .n ~ ,r, ~, .r ~"rt, ' ,n .fe 71, 'Q: 11
_. _ ... , ,v ,- ••• ~"I:>' , ... ··o·_v. '0'''- •••• va." 'v
~ w~vp, inr.irlpnt at an ~nO'Ip. H, on ~ rlipJpNrir. intprf~r.p from thp. hiO'h_intfpy l;lirlp
~ . ~
The refracted ray in the low-index medium, n2, exits at angle °2, The exit angle is:
, 1 (nl. ) (1.97)
(h = sin" n2 sm OJ
As the angle of incidence, OJ, increases, the angle of refraction, O2, must also in-
• co " ... , co A'
\;ICdljt; LU IjdllljlY LIlC ''1''~'" ,DUL "llfl2 ,/ I, UIC I Cll ac UUIl dlll;llC, 172, WIll
.1, • + O£\o l. .fe, Ll A . 'T'l. '
• ___ a. .. , ..... _ OJ' J" "_.,,._ V _,,_v •• a.w ,, __ ... v , .. _ ..

n,
,,: .. tl - .:..: II OQ\
'I " . -,

This value of (}1 is known as the critical angle. For angles of incidence larger than
the critical angle, O2 must be a complex number to satisfy Snell's law. A complex
angle in the expressions for transmission (for example, Equation 1.90) leads di-
rectly to complex amplitudes in the low-index region. Complex amplitudes simply
means that a phase shift occurs. As with all simple reflections, the angle of reflec-
t:~ .. ;" "ron., 1 tn th" .,,, ... 1,, ",fe • 'A. ~fe th r.".
"]' '0' ~J'

..


f'1 II? "?
~
..... T
.•• ~ ~2
~

t:l1< n ioa AnCie tl1 - nuca Anale tl1> \. n lea AnOle

• L ,~ . , L
,·If"·" ., 'OH~~ 'ouvw "u~.~ u.~ auo ..... v. .0 u .... vw, a., au ..
aDOVe me cnnca angle res ecuvery,

Total internal reflection is the key to optical waveguiding, Consider the dielec-
tric structure shown in Figure 1.12. A dielectric slab of index n, is surrounded by of incidence at the interface exceeds the critical angle. This is a simplified picture, as the actual ray picture of a waveguide is more subtle in terms of allowed directions for the rays. (This will be fully developed in the next chapter.) However, the essential idea behind the optical waveguide is that light is trapped in a high-index media through total internal reflection.

Figure 1.12 A waveguide can be formed when total internal reflection traps a wave between two surfaces.

electromagnetic field. This is called the optical Kerr effect. The Kerr effect modifies the index of refraction in the following way:

n(l) = no + n2 • I

(1.99)

where n2 is a small constant. For this example, let's assume that n2 = 10 - 10 cm2tw. (This

1.10 Total Intemal.Reflection !II;



~,,~ . . .
u. , .. 'u .. ,~ LHaL Hhillh .r , ..1 .VI IVW • allu low • Uo;;JUIIU UIo;; • rUl
UQru 1"" . . . . Q "Q"IQ"t th, n"rt;,," ,,4' ~ \ th . .
'J -0' r -mgnr» ,_.
angle, "e, tor a ray incident trom nhigh to n/ow'!
b. If the angle of incidence is 0.5° smaller than the low-intensity value for 8e• the wave
~ will have a finite transmission into the lower index media for low intensities. What is
the power of the transmitted field?
c. If the angle of incidence is 0.5° smaller than the low-intensity value for 8e• at what
intensity will the interface cease to transmit light?


a. l'he en teal anzte IS ziven bv sinH. n" .In ' or
~'O
/
sin II 1 \ ,n
Ur <to • V\.
\ 1.;) /
b. If 8 is set to be 41.31°. the ray will be partially transmitted at low intensity because
this angle is below the critical angle. The amplitude of the transmitted field, Et• is
given by Equation 1.89. The refracted angle. 82, is needed to evaluate the transmitted
term. From Snell's law. 82 = 81.97°, and it follows that
,., il
ti: /t: -'-high ~~~ v
t I I a il\
"-high vue> v "low v"" V:!.I Ull

., • 1 '\. r()~ A. 1 ~ 1
7R
" • "0" II 1 0 """ !l Q7°

Using this result with Equations 1.41 and 1.69, the power of the transmitted beam
(recall that 111 = 1101nl. YJ2 = YJoIn2) is given by
S = nIE;. Z
z 2710 (1.102)
(1.778Ei)2 Q1 ()'"70 () OtV\C() r;'2
,., ",..,,.., """~ "b~ ~.V~~J~ ~I
... ..1
'hp. . nower is

_..,
nh;oh~i
.), ·z
LTJo (1.103)
1.5E[ 2
= --. COS 41.310 = 0.001494 E:
2. 377 I
Dividing Equation 1.102 by the result of Equation 1.103, we see that approximately 39
percent of the power is transmitted through the interface .
'T'. _1 • • 1 ,£I,
L. ." ,",un y",. .v .u ......... "' .. ,," • ,\.1.1.,"" 1.1.1.'-'''''.1'1, high' ,"U,,. , .v
n., .r l\ _L]_ .-1£1fi
<'1$'" ,,;n..11 ':t 0


The critical index Will be reached when nhigh U) - 1.~141S. The necessary intensity IS 36 Chaoter 1 The Fundamental Tools of Ootoelectronics



/T\ 1 <:: 1 (I-lOL _j_~ / 'l"\&!'
"mgn'" "~ .~ , •• ~. c~ ,1dVJ

,~ ?
0UI vc lUI 1 LU ~e;L 1 1. .. 0 1 V Y' '''111 •
Is this a practical switch? No. First, we have chosen an unrealistically large value for
the nonlinear index term. Second, the amount of light actually transmitted in the "on"
state is a small fraction of the incident light. A practical switch for systems application
would require a more efficient transmission factor.



1 11 YVA-VI: II-Sf ;HIt"IIUN Ut' I U tAl _IN_l_I:HI'lAl ~L~I~ luN

vve cnu ua uie lay LULa LY euecte; lUI allg es evonc e critir; ana e,
_1 .+"+". _, • _il ~
,'-" .. ,.., vU'. .., .. v..,u".., ,.., VU"""'U LV .".PPV'L .... o ..,." ..... o "'"L LU",
identity is impossible to rationalize using real angles. We can put the description
on a more physical basis by examining total internal reflection using electromag-
netic waves. The wave picture provides a physical explanation of the reflection,
and yields information on the phase shift caused by reflection.
Consider a TE plane wave, polarized along the x-axis with amplitude Eo inci-
dent on a dielectric interface, as shown in Figure 1.14. The angle of incidence is
lp~~ than R Onlv the sriat ial oftherwowavesnre ,,;n{'p
L . .
t e time enavior IS 1 entica or ot :

E' r. _\ ~D - ikon,(z cos (J -=-y_sin (J,j_
~1\., <., -"~o~ L.L. (If!

E?(v. Z - TXl!-f\e J"U""" ~" .r ~"' -t- c.c.
where T is the amplitude transmission coefficient (from Equation 1.90). The angles
(h and O2 are related by Snell's law.
. nl .
sin lh = - sm 01
n2 (1.107)
n,"
N'" - .. ..!..,,;n L.
'L. V ~ , ,
..
s i u ins ese va ues in 0 b( ua ion
~u we ze an ex ression or e rans-
,. TO' .c. .rd.
llllLle;u arrrpru.uuo, D;, LllaL 1" a V1 LIIe; all;:,le; v ,
E2 = TxEo exp { -jkon2( Z 2 )}
nl . 2 nl . (1.108)
1 - 2 sm 01 - Y - sin 01
n2 n2
Physically, we can understand refraction by considering what happens to the
wavefronts at the interface. On the incident side, the wavefront strikes the interface
~nrl ;v n~ .. t;~ll" .. "R",-t"rI "nrl ",,, .. t;.,l1,, t .. "nvn?;H"rI If a. ~ t:I th" ,,, .+"_~nt, ~,,~.
l' OJ l' OJ C'
he II~ across the mterrace he none where these two
rave s a onz e in er ace WI ave ociev v as sown m rnzure . :).1 e ve-
, c .... _,
IV'-'llY VI LIU" 1ULC;I"C;"'LlVI~, "node» 1" "11111'1,

!,,;n /1 H\(}\
nuu"l 'PI .. 1.11 Wave Descriotion of Total Internal

___ .. ronsse :

I ----

on a ie ectnc mter ace at anste (j,

I

_.

----

'CL;UIIU JUCUIUJU. Ill" l"llClOICU Ht, .,

not shown for clarity.

'2

Figure 1.15 The plane waves on either side of the interface must connect as they cross the interface. These connecting nodes travel along the interface at a velocity that depends on the angle of

identical to that of the mcident field. -Smce llie phase velocity in medium n2 is different, the only way the node velocities can be matched is if the direction of the transmitted field refracts to angle (h such that

This is simply a restatement of Snell's law.

(1.110)

_, .J:' • _J Il .,

L"t.i> Uiv alle"" VI , Vb 111,,1 Ci:l1SC1S, rue U,

Wi:lVC1S IlIU1Sl IlI~C a

comnonent:

At R. R (y,,, R, onp" tn 7prn ",.£1 tho t .. ",.,~ ...... ;tto£1 +;01£1

Ie< .'" cr

(1.111)

This is the description of a plane wave traveling parallel to the interface in the y direction. This direction will yield a node velocity that is as slow as can be achieved in medium n2' In the ray picture, we would say that the transmitted ray is parallel to the plane of incidence. Figure 1.16 shows this condition.

Plituae is aescrioea as

~1.11L)

38 Chaoter I The Fundamental Tools of Ootoelectronics



'<,
"-
-, <, 01 0"
"- -,
~


.~ "-
"-
<,


Fiaure 1.16 At the
•• ~1
,oc:
rransmmeu piane waves
travel parallel to the
interface.
where we choose the proper sign of the radical to ensure that the amplitude decays
as distance from the interface increases. This cumbersome form is often written as

... ~{;fi' Y Z J{Jy /1 1 I'}
-L- -u' \~"~.JJ
uThpt'P '\J thp a t to n n a t irrn .c.c 1Llu.nit<c: .rzrn, :_1_l
or -JJ

~ ~nl . 2 0·114)
']I = kon2 ""2 sm 0, - 1
n2
and f3 represents the propagation coefficient (units: rads/cm)
f3 = kon, sin 0, (1.115)
ns ec ion 0 coua ion -' s ows a e e am itu e ecavs exnonentia Iv
away Hum uie interface: 1 His ueiu is caueu ~ e_llfl_fl_esce1'!l_!J_e_!(.l_. _!_!l_e evanescent

c. _LJ .,,~ II" .1. (l{\O 0" .1. '.1. . ..
.. ",n.' ''''"' '"'U"'~~' ~ "HU ~~, OJU. 'H"'. "'''' /v V'H v, pH .. "'" ,H" coa~ll UlllCOl .
'T'h "" 4';. ° .~ ,1 r,
.w y~ ,"'" .-"._!'_~ ,_ .. ~H ,-"_ .. ,- l' ''''', .. v
wor k is rlnnp hnt enero v 1" storart This pv~np>:~pnt tiplrl is verv . .nt tn rip",,.,,
~ - ,
applications. It is possible to tap some of the energy away using special structures.
We will see many such devices in later chapters concerning switches, modulators,
and couplers.
Returning to the physical picture, when 01 is increased beyond the critical angle,
as in Figure 1.17, the node velocity in n, is slower than the minimum possible
velocitv of nodes in medium n~. In medium n« the nhase fronts advance hpvontl
LI. eir gellt:Jalillg counterparts ransrrrittec wave
n . I'\.S we rons rave a ea
_. c. _1.' .1.
LllCO, I Ull up Vll VH' V "'UVIH" H Vlli "'UIll"" llVU"'''. ru U "'COllaIll Uli>lall~t:, LIJt:
fr"nt~ 'n., ,,,,11 h" 1 Q()O nnt "f nh.,~p ,,,,th~_o:f_ -<>..n.d ~. ,;. ;~.
'L- r- 'I'
fpt'pnrp \Xli orrll1' Thp 1~1'(Jp1' thp ~n(Jlp of inrirlpnrp R, thp "lo\l!P1' thp ... ,rip ""1,,,.
c» c» .
tty m n, will be. Destructrve mterterence will occur sooner, Ieading to mcreased 1.12 Phase Shift unon 'iQ



_ .. ~n .. n';~n UTa 0 J'¥~~ P , .. 1 11-:1:L :L rro •
-a ~. LU~ LUV T
'"' thp ,,"olp of ;n~;...Ian~, i" .
-'"
~ n2
\:------
" -,
" ~ --------_
~
-, /"
"\,/ <,
/" ------
k '\.. T
-, "
-, IVY'" I - pfli:ll:it:

Figure 1.17 Beyond the critical
angle, the plane waves on the
low-index side of the interface
travel faster than the nodes due to
the incident field. They get ahead
VL LH~U ov UL~' --uum:;;>,-..uu-mcIr
react aCK against them.


- .. --
I. I" I""nl-\;:)I: ;:)nll- I UI""UI'f r1l::I'"Lt:.1.. IUN

J-\ more suuue Yt:ll:lilil:i:luy impuhalH~S in 11K is me pnase snin

of the light upon reflection. These phase shifts help determine which modes propa-
gate in a waveguide, After reflection, the optical signal slightly lags in phase com-
pared to the incident wave. One can view this phase shift as being due to the extra
distance the light travels when going into and returning from the low-index media
during its evanescent phase (the Goos-Hanchen shift, described in Problem 1.22
_J: A \ '.' _. 'r '.' ••
auy '1" V. VU'"' VUU <Lv' LU,", pU<1"e; MIllL <1" V""UUllI1'O uue; LV Lue; llUAlll1'O
~J' .. ,,~ un, w .hn. .J' .L '.L .£1
~, ~"O""J .~,~. y" ~ UBY,", 'U'V/'
How hio is thp nh",:p xhif't? For" '1-. H'''''''' th", nh""", "hift """ h", ~",t"'rrn;n"'~
~ L r-
tree Iv nv wn mz the amm nude renee ion orrm I" Emmtl()n l)( In no]"r form'
L
.., " ~
Dr _ \"1 "V" vI "2 ~u:s (12) = I rl el2¢> (1.116)
- -
E; (nl cos 01 + n2 cos (2)
The reflection coefficient is described in terms of its magnitude, Irl, and phase
shift, 2<1>. Beyond the critical angle, cos O2 becomes purely imaginary
(cos O2 = VI - nt/n~ sin2 (1), Letting a = nl cos OJ, andjf3 = n2 cos °2, Equation
1.116 can be rewritten as:

J;' :D
r J /111-"
17 ." .,.,' 'I
LJ, -= 7P

SI ns nc the WI II" CO" H~ trom Fflll"tion th" nh""" of thi" tr"ndpr
-a r
!Unction is: 40 Chaoter 1 The Fundamental Tools of Ootoelectronics



/ a\ / a\
'"1m -1 I r-: -1 f'J
<..-rTE ...... ...... \
Lt / \U
= 2 tan-1 (-:) (1.118)
-Yn2 sin? () - n2
=2tan-1 1 1 2
n1 cos (}1
...... _. -., -,'_' c. " " ...... - -' " _.
.1 HI" ~y'uaUVH 1" VIll] vauu IVI V1 - Vcr' .1 H~ VI 1" VUV IVU"I, UHUY·
u; .1. • 1. ,0 • ,0. 'T'l .. ,
'" 1" ... ''' ..... o ..... .~ ., .. ~ .......... " '"'~ .. "'"'. 1~'"'''1''' .V. ..... • .... ,," 1" 51"""
h,,'
-01'
I -J17 Vn1 ~in" fj, - n£\
m, t.,n '1 ~ ( 1
l1Vl \ n~ n1 COS (}1 J ' . "
Figure 1.18 shows the dependence of $TE as a function of the angle of incidence
(}1 for two ratios n/n2' The ratios 0.3 and 0.7 correspond to the approximate values
of a GaAs-air and glass-air interface, respectively. The phase shift for the TM
case is similar.

",I':>
..,
A ~
.;::: n. t n. .r ~ ./ /'
s: -Co ,,/ V
V) ....... V /n2tn1 =0.7
Q)
Cf)
cu
.s::. /' (
a,
0
0 30 60 90
Angle of Incidence
." ,.. T. ,~ • _1. . " :", ....
.... . ,_ t' >U M. M
runcuon or me angle or mciuence, tJ, note mat me pnase
shift helnw the critical ancle js 7".0




For angles of incidence below the critical angle, there is no phase shift upon
reflection (actually, the phase shift can be 0 or 1T, depending on the relative
indices).

s:: 1. J:: S::i .. I.t .. c:., ••• ,.., ... .ti .. ,. .. T,..t .. 1 I, .........
. ."
thp. situation shown in Fi p"nrp. IQ A 'F. W:lVP. is . from ":lA~ with an
irrdex Vi J.'+ VlllV rue Ud"""-dii mtertace; dl <tll 'Ulblt:> VI Vi '+J • rue
,], • 1':. A •• .11 _" • ,or
• __ ....... ,~u ..... '-b" -0 .... - 'b' M
length of A - IM-m. 1.12 Phase Shift upon Reflection

41

I~

z

~ ,.. _'lA

,.. 1

';": _

""",n

v

1=1" .. r .. 11 Q A wave

-0

die ectnc interface trom the

-'" .

PUl""LCU, "llU "". a

wavelength of I p.m.

Solution:

For this dielectric interface, the critical angle is:

_ . -I 1 _ 0

(Jerit - sm '1 A - 17.1

(1.120)

~.

wnere x., "7T/lf.Llll o ... O;) A IV ,~lIU

2

nl . 2

l' = kon2 2: sm (JI - 1 n2

= 6.283 X 104.1. Y3.42(0.707)2 - 1 = 1.37 X 105cm-1 The value of the transmission coefficient, T, is directly derived from Equation 1.89.

(1.122)

7 t1.1'(''»)

n cos Ul n, cos u,

n this case H~ IS imazmarv,

COS (J2 = - .. 11 - n~ sin2 (JI = =i 2.19 (1.124)

'J n2

where j stands for v'=l. We chose the negative sign of the radical in order to assure that the field in Equation 1.108 decays with increasing z, Having chosen the sign, we must use it consistently throughout the rest of the calculation. Plugging cos (h into Equation 1.123 yields

" 1'1 A' In '7n'7\

~ v-« 'I_ .v- ~ _'-

11 1 ,..,,,,

~J. v.v J,,-.~7

stmply 1I11:;'"1~ rna llllOllO is " retatt VIO 1'""~10

T, ... .~h. •

t'u, ,

T, is equal to

42 Chapter 1 The =;;:;: ~ ThO[, of '"\n'"",1



r / "",\1
'A ... 17
'j I L- lall \V, lall
L..Lf I . 7h
= 1.48L42.38°
The transmitted electric field in the air (z > 0) is
E(z, y) = 1.48Eje -1.37x 105zej1.51 x 105yej(wt+42.38°) (1.127)
where mixed units (radians and degrees) are used in the last term, but the meaning should
be clear.
"'., ,", ~ ~ I, aHU LIIC fJlHtSC SHill is ztven )v
'''v llCIU can UC IUUllU llULlll~ LHilL
111I\tn~'
.
vn sin tJ n;;;
~ tan ~
n. cos H,

-1 -V3.42(0.707)2 - 1 (1.128)
= tan
3.4· 0.707
= -42.29°
Thus the reflected field (z < 0) will be given by:
, ~ A • {\ 7{\7, _ 0, .. 0
--r;(Y, ZT l!.;e 'U' eJ TITI9T

Noh' t l-mt th .. ~ ,~l;hul", "f th", ... ",l",,,'r;,, t;",lA ;< ;A"'nt;"~ t" th .:: IA
,
mere 1S a pnase eray oTZWDetween the two waves.


1.13 SUMMARY
This chapter reviewed Maxwell's equations, using them to establish a set of units
(MKS) and several important quantities and concepts. We derived the wave equa-
tion, and solved it in homogeneous media. From the solution, we developed ex-
ressions or ase an zroun ve 0 ltv The concent ot the 'rtor U,<I"
imroouceo arnrreraretrnrtrre anzuiar rrequencv or a wave. using DoundaiV con01-
, r- n r-
"V"", '''' u'" ''''IVfJCU IVI u rc l' allU ICl1dl,;UVll VI CICI,;I
J'. '_1 ~
'~,~~ .. ~u. ~ ~, =-rrIUOr'''V'''.
W .. th .. n ov ... l"ro,-l t",t,,1 ;,.,tor,.,.,1 ~,.,ol "c I.,,,, "'''0 110"',-1 t" ;1 .~t_~.~ .1.,
~'
ray picture of total internal reflection. While Snell's law, if used with complex
angles, can give a total description of the evanescent fields associated with these
reflections, the wave description based on Maxwell's equations provides a clearer
picture. Using the wave picture, we used the Fresnel formulae for reflection and
transmission at a dielectric interface to develop expressions for phase shift associ-
ared with TTR T'Iric nh"" .. "h;ft <llm<lv" .n;p" TT"R <Inri nhv" <I 11 ;""0 r"la
-. -r r-J -,
in es aOlis ing w ic ravs w e a owe msi e an 0 nca wavezur e.
.,., .. ue e mi iug e action
l11C 11ldLCl1d parameters, J.I" aBU to, playa crruca roie
,J' ~ .:1 ~ nT. ,11, A. .1- J'. ."
" - ~. ""'v "U"" ''''_v. ''-' " ..... '-''-'~ LV LU'-' "'-''i'''v~''-, . VI lHCSC
.".,t",r;., ... "r"."o.orv;~ .1.0 A;v "J' ~r" , ... "01,,,,;.. '1'1.' _1
~ b' -1' OJ' .. ~ .. - ... , '-''"''''''
oped m Chapter "J. Problems 43



DCCCDCt..lI"'C~
.. _. _ .. _,.- -

r .. nlrin~ "nil \l 'hit .. P"ndnJnpntnl. nf nnt;r. 'N .. u. Vnrlr· M,..(;r"u._ 0<:;7\
-J
2. E. Hecht, Optics (Reading, MA: Addison-Wesley Publishing Company, 1990).
3. S. G. Lipson and H. Lipson, Optical Physics (New York: Cambridge University
Press, 1981).
4. W. Louisell, Radiation and Noise in Quantum Electronics (New York: McGraw-Hill,
1964).
" D T n. TI. .1' T • 'l. rl In,.I', J:<, .1. rl. ("'1, D.
, ",,"' _, -J '0' , , ,
Il}XT
fl. ~ee or examoie W. aoor . Mae:neto- JotIC Materia s In aser Handbook,
VUl. c, CUI>. r. L illlU D. V. J\,;lIUlL.-UUUUl1> \UICilL Dl11ill11. l'Ullll-IIUllilllU
P"hlichinCT c. 10TH
=r 'J'
7. A. Yariv and P. Yeh, Optical Waves in Crystals (New York: John Wiley and Sons,
1984).
8. F. Jenkins and H. White, Fundamentals of Optics (New York: McGraw-Hill, 1957).
9. A. Nussbaum and R. Phillips, Contemporary Optics for Scientists and Engineers
(Englewood Cliffs, NJ: Prentice Hall, 1976).
10. Warren Smith, Modern Optical Engineering, 2nd ed. (New York: McGraw-Hill,
lQQO)




"'H(JML~MS


1. Derrve the Fresnel amplitude renecnon and transmission coemcients tor an
electromagnetic wave that is polarized with the electric field in the plane of
incidence (TM wave).
2. We simplified Equation 1.24 by assuming that the term - V(E • VEle) is negligible.
Determine how small Ve must be for this assumption to be reasonable. Starting from
the exact wave equation (with the above term included), use separation of variables
to ~olvp, for th .. on .. -ilimpn~ion"l W"VP (i.. F - Z( '7' ( t Solv .. for T( t) in t .. r mc of
I~ . r-rom me resuuing e
se ara Ion cons an Kana l,u,e ua Ion or £tZ , n a rouzn
e. t7_ .. : .... .e ,:;, u,
'u.u~ .~. ,~~~. U '''U' ~. ~ .. ~ 'u -co' ~ •••• ~ .. ~'u ••• ~ ~ ... u ..
1\"
rnucr h .. to mnk .. it npl7lil7ihlp, (~"v lp,,,,,,, than np,r" .. nt in rna to
e ~ ~ . ,

LU" VLU'" L"UU" U, LU" wa." •
. k k
3. Show that for a harmonic wave, the average value (8) = (E X H) = Ii -- E6,
k 2wfJ-
for a wave with wavevector k and electric amplitude Eo.
4. Fermat's principle states that if a light ray travels between two points, it follows the
path that takes the least time. Use Fermat's principle to a) verify that the angle of
.~ ,1, r ,-" c. .l '-
"'1 ua,,, L1'" ,n,lS'" v, " ,v, a "uu1"" 1Hau", UUBV" auu UJ u" .. ''''
Sn .. l '" law for" rav crossj no " di .. l .. ctric .
" T~", of "nil th .. facr that a nhoton has mom .. ntnm o iv .. n hv
p n ... n"l\(], WHCIC "'0 II> LHC Vil\,;UUHl Ul UlC l'HULUll, LU U, MIUW 11lilL
..... ,1. +l. .1. ,I' ,-" 1' • .1.
... ~ U"o" ~'1' 'b'~ ,~, ' .. .. ,
b) derive Snell's law for a ray crossing a dielectric interface. U C'h"nt"r 1 Th" 11 Tnol. off



unuer w at conuitions wtl LI e pnase 0 II e re
o. ecteo e ec nc lie e Op~OSI e a
• j: .... ·A . .'l
.. ' ,_ .
7. Derive the four boundary conditions, Equations 1.58-1.61, relating E, H, D, and B
across a dielectric boundary. Use the integral form of Maxwell's equations.
8. Consider the situation shown in Figure 1.20 below. A TE wave (polarized along :9)
with a wavelength of 1 f.Lm is incident from air onto the GaAs-air interface at an
angle of incidence of 45°. The index of refraction of GaAs equals 3.6 at If.Lm.
Describe the electric fields in all regions surrounding the interface.


v

""
~ 1'1.11 \.;lI1;lI'\l:j
'\
'\
-, ....... ""'"
H~
. ~~ .. '~ •• v ~ HOY'"' ",,",,u,",,,, a •
.. ;, on a Ole ectnc mter ace
fr-orn thE> Jnw_;nn"y <in,. ThE>

w a Y<O " 'L pv,aH"'<OU a"u "a, a
wave engt 0 ltJ,ffi.


9. Using a computer and a program like Mathematica, plot the amplitude and power
reflection, and determine the amplitude and power transmission coefficients as a
function of angle, for both TE and TM waves for a) a glass-air interface (nglass =
1.5), and b) a GaAs-air interface (nGaAs = 3.6). Assume the light is incident from
the air side

Beeinnins with Equation . j. derive the homogeneous wave eouation in terms of
• .c .I .I ...
un;; u".u ·u .
~ ... .. ~ " .
II, l'lVUl un;; VI Lll';; r V:'llLl1l~ V';;"LVl, ""'luauvlI 1.J7, "'llVW llHU D'IuaUUIl J,"U
f"l1"ur~ urh .. n th .. Ti' fi .. lei ;" " ~;nl1"o;ei,,1 fl1~
12. It is easy to understand trigonometric identities as (J becomes complex: use the Euler
identity, cos (J = 1I2(eilJ + e-j~, and let (J become a complex number, a + jf3. Show
that cos2(J + sin2(J = 1.
13. Confirm Equation 1.119 for the phase shift that occurs for a TM wave upon total
internal reflection.
14. Consider a TE wave with amplitude Eo incident on an air-dielectric interface from
me dielectrtc siue. 1 ne dretectrtc nas an inuex _lJl_ rt!!£ac_!ion, '!t, O!__!_.!). _!!l_e angie 01
. ,0 .~. .~, .~;.'.~ ~ ,1 . 1
-e' "e ,
0. What IS the cntica angle tor this interface?
~ • .c .I .0' J'. .ll. . . .
U. LU" "'';;''lH'' U';;lU lV' au pUU"" Ull lU" all "'lU';; Vl tIle
"f~.
c. What IS the phase shirt, 2<1>, tor the reflected light'! Problems 45



." .... L -"., . .C!. L n • 0.., L ..1_
'oJ. • u,", U ",",uu,",u UJ '-''1u"uvu •. U " "U"''-',",I
mh .. n th .. "no-I .. of . e-er- .... rlc th .. "rili"" aoo l .. Wh"t rlo .. c it rro .. "n nhvoi""ll,
~ '-0 r-J -J
wnen me transmission coerncient becomes complex.
16. Use a computer and programming language such as Pascal or Mathematica to
generate and plot phase shift curves for the TM wave, similar to those in Figure
1.18. Generate curves for dielectric-air interfaces, where the dielectrics have indices
of 3.6 (GaAs) and 1.45 (glass).
17. It is possible to convert linearly polarized light into circularly polarized light by
th ~h o~ ;f't .h .1 . • +1 r' •
'0 r <r- . .. _
rhombus shown In Fizure 21 inear nolllri7",rI liohr lit 4'\° to the horizontal ax i e i"
~, , ~l1"~Ue LUe
Vll Lll<O bL""" I11UIUU. 1 U<O 11<;l1L l11i:ll\.<O" L WV I11L<OI11al
hl~~J..- ~nrl ~ ~hn oha-. .r, ,h .+1 f"r J. .'1'1:1 '1'11...-
, 'r
comnonents of the Iizht. ( ircular nolarization is achieved when one.ot.the
r
"r- \-" UI Y J VI L11<O 11l1<OaI 11lpUL U<O",VIU<O UY '111'<' laUlall1> VI pua",,;.
Use a computer to calculate the total relative phase shift for an optical beam
incident on the slab as shown, and determine what incident angle is necessary to
achieve circular polarized output. Assume the index of refraction of the glass is
n = 1.54.
1"-...
/\"
<c 0\ <,
""'-). <,
"S
.......
<,
Figure 1.21 A Fresnel rhomb.
18. Show that the continuity of the transverse electric field is satisfied at the dielectric
interface used in Example 1.5 for the field shown.
19. Repeat Example 1.4 for a TM wave.
zu. nuw lal away llUI11 LUC UI,",I,",""I" uuerrace UI>CU HI 1.J mUM one <;0 oe ore
.:. "_11 . tA c, ... ''''
T ,~ ,_, y .. " ..... _ ,~.u_"' ••• _ ••• __ •• ~ __
17

n a hot avon the hizhwav the distant road sometimes annears to he a shining
n . 1' ~ .. _11 ,.
I<OU<o"U/" pVVI Vl wa''''I. lUI" ro I"""y "11 Vl LVLal """I11al
reflection. The air directly above the surface of the road has a lower index of
refraction than that of the surrounding layers. Assume the index of refraction of air
at 273° is n = 1.0003, and the index is directly proportional to the density of the air.
If air follows the ideal gas law (PV = NRT), and the surface layer of air on the
highway is 30° higher than the surrounding layer, what is the critical angle of
incidence for total internal reflection at this interface?
"" T d. ... _ .. .n .c .
<0<0. 111 1I1'" ,a ,\.0 PI"'UI\.o, ,\.0 1\.l1V' 1Iial upv" lv.al 111l\.ol11al al all IlIl\.oI1<''''', lIl'"
o-nirlp.il WlIV .. " nh""p. "hift ?<t in th .. rll" ni{'tnrp w .. {'lin
1 J •
in e re a pnase s i as a aera isplacemen 0 ere ece wave. ms IS Known
n, .~ r'.. Ur ,I. ..... , 'I' .1
-JJ -r ._, '" , __ "_'oJ
aornal lv h .. vrmrl thp .. inm th .. ]AWP.r inil e. " m .. ilill h .. for e. trrrrri no
~
arouno. Corrsirter rigure 1.-"" snowing a ray penetrating a surrace. 48 Chaoter 1 The Fundamental Tools of Ootoelectronics



~ Z.,.-------.j
-r-
/' <,
I, . 'to /"-.. '>0 _...,...._ JI
./ " ,
/ .~.
Figure 1.22 Schematic representation for
Problem 1.22.


.L • L .. -I- . ,0
,.~~~ ... ~ ."a. <1'" . la) ", ,,.11lU"5 <1'<0 ",,""a,,<o a. all a"5'''' "'Aa,",U) • ,aI5"'1
.h~n .h~ ~r;.;~~1 ~n"J. ·'n A A ",h,t;o th rI, •• " .f' f'r
"0' • '0 .. , 'r r
this interrace? W hat IS the lateral displacement 2z.·! How does the depth comoare to
I- c: I- n
<11'" U"'p.ll Ul <1'" ~ ""'U 11Ull' <1"" "<1U"'Ul<O.
23. A waveguide has a core index of 1.457 and a cladding index of 1.454.
a. What is the critical angle for this interface?
b. How far does the field extend into the cladding if excited by 1 ~m light at 88°?
Find the characteristic length, 11".
24. Confirm that the reflection coefficient for the electric field, ErlEj is unity for all
"no I .. c or .. "t .. r th"n th .. f'riti",,1 "nol .. for TP "'''" .. "
-0 0- -0-
7,\ Rp.np.:lt _'\ for :I TM W:lVP._ findino the fill of thp. p.v:lnp.""p.nt
. . . .
em me uumg re a ive pnase sni s ano con rm rna me con muilY 0 e ec fie
.1! e , Co. _, .h .£_ - f'r ." th ~al,



rT"'I T T T"" ,.....,. T'lI. rT"'I T,...., A T ~~ TT T'lI. T""
I r1 r I II""" I II ~ I VVIKf""I g A g T
- - - ---~ . .--. .. •
2




































47



P'T"'IY Yy-. ""Y '\lIo.T,\..,. rlY ,\ ..,. ........ T ,\.... T y-. ,...., Y T T T'"'o. y-.
I r1 t r- I I-\. "'-I I-\. K . , I I-\. n vv I-\. V ...... YI I I J ..... r» ... " DYE:~ '"
~~~~- ~ ~ ~--~-~ _ .......... _I I ...










CHAPTER OUTLINE
2.1 Introduction
2.2 The Infinite Slab Waveguide
2.3 Electromagnetic Analysis of the Planar Waveguide
~ ~. T . , .. ~
~. .lut:: '" .p
7" Eiaenvalues tor the SllIh Waveouide
~
I'. On.;"" l\1f",~", {'",nh
'r-

z: I 1 ne ;:)J ,",,,';UIUt::
2.8 Intuitive Picture of the Mode
2.9 Properties of Modes
2.10 Number of Guided Modes in a Waveguide
2.11 The Numerical Aperture
~ 1~ C'''~~''r''
'J






:ii!:.l IN I HUUU~ IIUN
In this chapter we establish the fundamental concepts of guided waves. It is per-
haps the most important chapter in the book, as almost everything else will build
on these concepts. The homework problems offer many opportunities to test your
understanding of the concepts. Due to the transcendental nature of the eigenvalue
ermar irms 1I with . .• analvtical soft wara is rermirerl tn mlllcp. thp..
.
comtm a ions. ou are s ronzrv a vise 0 eveiot a se 0 s an arc rograms ase
_. _ ,. _, . ,
UU Ult:: rueuoi reu 111 lUI" "U Uli:tl YUU \_,i:tU YUIl;l\.IY Ult:: Ui:tMl; Wi:t'
",4' " ~;, .n ~'r' on Vr ;11 h· rI .~ .4' .. 1 _.
t- o" .... ~ .".~~~ ~~~ ...... ~ ,~
~;ff"'r"'n' hlnp~ nf "...",~p ",mn1;no .,n~ ~p";,,p "",ne 1"tpr ;n
-.I r '0 , =r '0'
the text.
49 Cha ter 2 The Planar Slab Wave uide

The slab waveguide, shown in Figure 2.1, consists of a high-index dielectric layer surrounded on either side by lower-index material. The slab is infinite in extent in the yz plane, but finite in the x direction. The index of refraction of the guiding slab, nfi must be larger than that of the cover material, nc, or the substrate material, ns, in order for total internal reflection to occur at the interfaces. If the cover and

waveguides are not infinite in width. However, the one-dimensional analysis we will develop is directly applicable to many real problems, and the techniques form the foundation for further understanding. We will begin by solving the wave equation using boundary conditions for the slab waveguide structure. We will always choose the direction of propagation to be along the z axis. This will lead naturally to the concept of modes. We will then develop formal mode concepts such as or-

2.3 ELECTROMAGNETIC ANALYSIS OF THE PLANAR WAVEGUIDE

Consider the waveguide structure shown in Figure 2.1. The three indices are chosen such that nf > n, > nc, and the guiding layer has a thickness h. The choice of the coordinate s stem is critical in makin the roblem as sim Ie as os sible

e vector 0 t e gut e wave w zigzag own t e z aXIS, stn mg t e mterfaces at angles greater than the critical angle. The field can be transverse electric (TE) or transverse magnetic (TM), depending on the orientation of the electric field. The TE case has no longitudinal component along the z axis; the electric field is transverse to the plane of incidence established by the normal to the interface, and by the k vector. Because of the different boundary conditions that control both

Transverse Electric

Transverse Magnetic

Figure 2.2 Transverse electric (TE) and transverse magnetic (TM) configurations.

. .

vacuum wavevector of magnitude ko, where Ikol = Wale. To find the allowed modes of the waveguide, we must first solve the wave equation in each dielectric region, and then use the boundary conditions to connect these solutions. For a sinusoidal wave with angular frequency Wo, the wave equation (Equation 1.26) for the electric field components in each region can be put in the scalar form:

not expect the amplitude to vary along the z axis, but we do expect that the phase varies. We write a trial solution to Equation 2.1 in the form:

(2.2)

{3 is a propagation coefficient along the z direction, but we do not know its Plugging this trial solution into Equation 2.1, and noting that

52 Chaoter 2 The Planar Slab Waveguide



D <, 1, T}' ·r. • r. o . . , ... fC. " '1\ ;11 }, .1 ~l
r-: '1)' ,. ~ ••• ~. '/ ,~ ... u ,~ ~ o~ .. ~.~. ~,,~ .. ~ ..
"'1th" ron, OV"".,o.,t,'n' fortn·
=r

E (X) = E e±V f32 - k6n~ X for f3 > "oni (2.4)
y o.
where Eo is the field amplitude at x = O. To be physically reasonable, we always
choose the negatively decaying branch of Equation 2.4. This solution should re-
mind you of the evanescent field of a total internally reflected (TIR) wave at an
interface.
~
~J UIC case WIlCIC P ......_ K{ ;. u ic Ilill> ill . Y ~UI
. ~
F ·yl _ /<:',p.Cj ."0' i f-' A tor H < k..n . 'I)
J'

~o e en mz on eva ue 0 0. e so u Ion can e el er OSCll atorv or exoonen-
tially decaying. If f3 > kOni we define an attenuation coefficient, y, as;
y = v'f32 - k6nr (2.6)
and describe the field as Ey(x) = Eoe-Yx• Compare this to Equation 1.114 for the
evanescent field of a TIR wave. If f3 < kon;, then we define a transverse wavevec-
tor. K. as'

K v Kon· 0 k.

r- r» +iKX TL' r. ""7 .L Q .1 L
;,v LoV\"', '-"OC' • '-';"H6 L.J':!"""VH ... , '" """ 'H'" I-' "HY n. ""H "" ""<LH.
A 1, 1, • A; hI·
.~.~.~~ .~ ... ~ .~.~. ,~_ ,_ "~" "U' , .... .......
L k~K
z 13
.," .. ., a
."'.
respecnveiy, or me wavevector K.



n .11. • L I .
I-' <LHY n. < .. " "" .. "u u," .... u ...... o '''''''' '''''' '''".v, ", '"''1''''''' ''''),
inside the guiding film. These terms will be used extensively to characterize many
types of waveguide mode, so become familiar with the relation shown in
Figure 2.3.
2.4 THE LONGITUDINAL WAVEVECTOR: fJ
'T'I- .1: ,I Q' • A. :& • • --1: '--1, ,I --1 • un:l.
_. .. " , ... " ''''''v. I-' .~ ~~~u .v ... u. '''.~ ...... vu,,~. .... " .... ;
... ,~h.r ,I' ~~.'Irl ho ~h~co ... t~ hI th;e- r~lo D ~c rl.,h ... orl o o th, ~.-. ....
r- ,,... 0",'
v~111f~ of the mode Fionre 2.4 nlots the transverse electric field dis . in ~
s an wavezui e or vanous va ues 01 0 j • nlllls 0 wea ow e an!:!. e etween
/(. ano z 10 vary nom ':IV 10 V • 1 his in eriect varies ure value U1 p, which is simply
2.4 The LOIwitudinaJ ·R C'i



rh ... 7 r- ... ",n ... n",nt ... f th4 ~~~ (\ •. D t. 'T'l :1 .S: J'l'
-r , r-rmax • ~W V<UU,," V' ,.., 1» P1U",",,,
alono thp hortzonta <lvi~

/
•• T r. :.~ I
4 I ~ <, I I
/ /' 'lIIIo.


~
-
II" --- '-
•. <:::; ::::;:::::> . .......",_
n, ... ~ .•.. '? I .. ~
ns :? T ~ I ( I
I I I
I I I
I 13
I ,. ....
n
v "()c "0 5 0 ,

Figure 2.4 IS rav an wave nicture shows the electromaanetic fields ~< ~
Fn"~t;nn nF a



I ne rap sxetcn Ole iguTe-z:zt snows the ray picture ot the nero, wlliTe TIle Tower

sketch shows the wave picture (solutions to Equations 2.4 and 2.5). There are three
special points on the f3 axis, and the first one is at f3 = kone- For f3 < kone'
solutions to the wave equation in all regions of space are oscillatory (Equation 2.5).
The ray picture shows that when f3 "'" 0, the wave travels nearly perpendicular to
the z axis of the waveguide. Like light going through a sheet of glass, the ray
,s:. .... ~ . 'V wave is
a< <11" ll,," 11ILCI1i1~C:>, UUL I:> 11UL II i1IJIJCU • .M.11 U: res-
• .1. .1. ~
.u u.'" un",,,, ,...."... ,. rrc '0'
'h", ~prnnel "npr;'" ~ ... ; ....... r»r» .... ~ .,. 1-. fl 1=1..... 1- fl .-- --- -.
'r 1" 'V' S' 'V' e -O s' 'B' =s P"''''''''
SOWS a ray otal-mte n<ll.:r;>ffprtino <It thp fil",_rn"",r in~ h, .... a4' .. n".':"'~ n'
-= '0
lUC IUWC!' suostrate-rum in er ace. n e wave IC ure e e evnnas-
cent in the cover region. The field will still be oscillatory in the film and substrate
regions. This condition is called a substrate mode.
As f3 increases beyond kons' the evanescent conditions are satisfied in both the
cover and substrate region, while oscillatory solutions are found in the film itself.
Such solutions are, in fact, the guided modes of the film. The ray picture depicts a
ray trapped between the two interfaces.
14' D ". ... ".:" • .- ... I. f~l _L
U"]VUU O"t PU]""_UB. 1L 1:> 11UL ~lCi11 11UW
thi~ " ... 11lel ",,,pr h", .-1"'"4 ~;"".a D;~ p;~~I" .h4 7 oS: I. , .L 'r-
, ·t,"J -r-r- "U"W, U'''B .LJ'fUUUVll
1.4 IS satisned eve ~ rpCJi ... n~ muct }'<l"P ~n p_~;_"n
-r-
ne 0 IV wav 0 sa is:v oun arv con 1 Ions IS to c oose ex i:l]]" inrrp:l~ino
r- • '"
HCIUS III LIlC S Ull '0 -uierecuic- egions; '0 me IIeIITTo expTooe Toward 54 ( hanter 2 The Plan-;lr ~,jp



'.f: • . \. _,. c. .::, c- i :» .. .
UH.U"J ~u 'H~ ~ HaH\.-\.- HVH, <1'\.- lilli' • ,",au",.' 1115 L1H" VUUIU i c-
el11ir" inhnit" .~~~. n_ ,l.~ ~~,;~~ ~~~~;n _" ~I .\. . ~I
'"1' '0." I:' .... w "Ht'''JU'~~'.
A guided wave must satisfy the condition that
kons < f3 < kont (2.8)
where it is assumed that nc ~ n ; This is a universal condition for any dielectric
waveguide, regardless of geometry,


... po .. ,-- ••••••• ____ .... _ •• _ .....• _ , ••••• __ •.. _-
,.~ 1;;1\.;111;;' ... vru .. UE:o;:) run nE: VV,",Y E:UUILlE:
'T'. .::. .L 1''' _, .1. .1
'-V l111U L11\'- ,alU\.- ~ LV (111V' IIOU LV ll1C wa VC , we-IIIUST
<11"11"11" th" hn11nrl~~. ~~~rI;,;~~P t~ tl.~ ~~~~r~ p~I ... i~np rI ... re: I~n. Q, ,., A
'or r'J J 0' 'l" "1' ...
and 2.S. Assume that f3 satisfies Equation 2.8. Then the transverse portions of the
electric field amplitudes in the three regions are
Ey(x) = Ae-Ycx O<x
Ey(x) = B cos(Kfx) + C sin(Kfx) -h <x < 0 (2.9)
D7::\ n~(x+h) _...- .L
yV'I L

n I' .1 ~ .1'r- ' 1 .11'.
'11\'-'\'- £>., '-', '-", ullu~l'"LUUC lV UC UClCllllll1CU llVll1--un;uuunu-
Clr •. l"nnrlitinnc ~ ~nrl ~ ~~"~r t~ th" ~""n11.,tinn ;"ntp ;n '" ~. ,r -J
'J , C IS 'LOU
"tr~tf' rf'''"f'rti""l '1=/ £111 <lti on sr Clnrl U' ic th" trClnc""rc" N"\rnnnn"nt ," IT in tl."
guiding mm (from Eq~ation 2.6). The b~undary conditions tha~r connect the solu-
tions at the interfaces are:
1. Tangential E is continuous.
2. Tangential H is continuous.
we rarerv worry a ou con munv 0 e norma com onents ot U anr R hf'._
.. wnen we satrsfv me ransverse
cause uiese ----an; almost always
-J: • c: '" . r- _, . n ,.,n ~
. ~" .. LV' a" "' ....7, .ro uall" ''''' "" LV LIn;; 111lt:lli:ll:t:,
.nrl .• r. I:,' 1 • _, .1, n; .J., • J., . . " .
'J 0 'u, .. M ut'I:"J' "u, uu~ .. , '''~ """IUr
conttmutv of maoner ic fi"lrl 1-l? .li:hnlllrl \"" writf' rln"," ~ ""t nf f'{1Il<ltinn" cin>i1.,r .~
'0 ,
Equations 2.9 that describe the magnetic field as a function of position? Indeed,
we could do that, but there is usually a simpler way to derive expressions for the
magnetic field. If we assume that the fields are harmonic, then we can describe the
magnetic intensity in terms of the electric intensity, and derive a simple boundary
condition for the magnetic terms. Recall that
:lR
t:vl<' f,., 1 ("\\
::.. \ 'VI


Q.,~ 'rI -,
~ .,," ,
B(t) = J.LH(t) = J.LHoe'wt (2.11)
so 2.5 Eigenvalues for the Slab Waveguide 55



t7 v];',+\ .• U,,\ ''"I 1'"1'
'J f""J~~~"J ,~. ~~I
an ex ression or e anzen ia com onen
we nee e z com onem m 1S case
of H. Expanding the V x term of Equation 2.10 into its individual components,
and taking the z component, we get
A( eu ilEx) .
z :::.::::.J: - - = -fJ-JwHz (2.13)
ilx ily
Since there is no Ex component to the field (it would not vary with y even if it did
pyi>:t rllIP to thp, mfinire nl annr >:trll('tllrp) wp opt ~n • p(1l1~tion for thp t<ln_
, ~ , ,
aential comnonent ot the magnetic held H

;11;'
t.r .s. :::.::::.J: ''"I 1 A'
"z :;" ,~.~ 'I
r:
The tangential component of H, Hz, is defined in terms of the electric field quan-
tities. Since u. and ware identical in all the media, the continuity of the tangential
magnetic field is guaranteed if ilEy/ilx is made continuous across the interface. We
can now find the amplitude coefficients, A, B, C, and D.
At the x = 0 interface, the condition that Ey be continuous requires that
A -YeO D f\\ ". f\\ ''''I ''''
H<- ~ ,",V" '''"; v ) "-' ""'''''; vI V' •• ~.J'
ic is sa is ell OnLY i .Ii £..lVU1Kin}!; e ma}!;neLic con inuous a x
w e
.. r- r~ ,~ rv
le;'"IUlle;~ U1<ll Ule; 111 ~l ue;u vall vc Vi uy, Uuy/UA ue; al A V,
A
'A'Y.e c - '11K.sm K.l + K.COS K.lJ1 /,..
J J \k.iV,
-AYe = +CKf
yielding
C= _AYe (2.17)
Kf
All coefficients are written in terms of A. Using these coefficients and annlvinz
ue corrditfo 'n In is a ositive tum yie
nat (I" 01:: com ous aI x e s
r- .,
'V
A ('n>:, -U',h) - -"!: "in(-U',h) ')0 Ys\l, -ft) IR)
J J
L "} _J
.,.,. c. n
~ u", ,-"au u'-" "V, ,'-"'" .V. LJ ~ "u,\ +) "U,\-", auu '-"v,,\ -", ,-"u"\-",,
D = A[ cos(Kfh) + :; sin(Kfh) ] (2.19)
Putting all the terms together,
s, = Ae-Ycx x>o
I I
D rc • J, "\ I • n ''"I '"In
"'v H '-"V", "f-", "U,\"",.-", " ~-" v \~.~v,
"I


l7 - A ,{ e .,( .\ Ys(X+h) .-- ,1.
~y .. ,~ "''J' u,u"'J 'J
L ""; _j
where A is the amplitude at the x = 0 interface. Equation 2.20 describes the am-
plitude of the electric field in all regions of the problem. Note that negative values 56 Chanter 2 The Planar Slab Wavezuide



_I' -J' .1- • -J' -J 0 .. 1..0 '_n In. __ ~ :t: ~ ,I~ ,;11
~~ <uu~. OJ~ u~~u <" .,,~ 'OJ ~J ,~ ,~ <~
.... ,"'" ... ~ ... o c, ... c'",,, ".,I .. ",c rJ"c ,c ''''''., h., ... A. f"~ .. .,I~tt''''o ~ .. t th .. tnnrl"
0 '"1' OJ -J 1" "0
profiles of guided modes. We will apply It to many proDfems III ffiis iliapter and
text.
Having found the amplitude coefficients in Equation 2.20, is this description of
the transverse electric field complete? No! The propagation and decay constants,
'Yo 'Ys' and Kf all depend on 13, which is still undefined. The fourth and final bound-
ary condition, namely the continuity of aEylax at x = - h, gives an equation for 13.


:.IT<'
:.::::.l: ,1 ~_c'n~_'" " ,..nc{ ~_"y {-F.l." t",,..,, \
;:}. . " Ie 'J "
.u f"l "11 \
\~'~~I
I
= A l cos(Kfh) + ~; sin(Kfh) J 'Ys (substrate term)
Divide both sides of the equation by cos(Kfh) to get the eigenvalue equation for 13.
tan(hKf) = 'Yc + 'Ys (2.22)
r ,
I 1'r)l, I
KFI """"
IV

'h,~ ,~ " I"nt" th<lt rrm ct h" "nh,,,rI nl1rr>"r',." ror Iv
-a- < J
A tprm" rlpnp.nrl on thp v"II1P of H is r"llp.rl thp. characteristic o n urt t trvn tor the
<
TE modes of a slab waveguide. Solution of this equation will yield the eigenvalues,
f3TE that correspond to allowed TE modes in the waveguide.
Had we set up our initial problem with transverse magnetic fields, as opposed
to transverse electric fields, we would have arrived at a different characteristic
equation for the eigenvalues, f3TM' We leave it as an exercise (see Problem 2.1) to
confirm that for the TM case the eisenvalue eauation for B is


1 .. 4: ~ I
l ... l ...
~'l. s ~'l. c
s c
tan(hKl'l tL.Lj)
n;
Kf -----:f:2 'Y c 'Y s
ncns
Each waveguide structure that we explore in this text will have a characteristic
equation that must be solved to find the eigenvalues of the modes.
The transcendental equation can be solved routinely with a numerical package
on a nersonal comnuter or it can be solved zraohicallv, The amnlitude distribution
,A " , r r- " /r.
':'.':'U) i1l1U ure II ,lUI UUUIllg rne t:lgt:U valut: ,L<y'ua-
" "" " ",." .c .L .rt
UVW,) ....... "l1U "'''"') CU" LiU"'''' VL ,u", "'Y. ... "UVll" ,11", JV'" "-ll U,,'" , ~~ u<~J
~hn ,IA h", 'ntn " rn;-;tin" ~ ,...,n h" ,...,~ "nA morlifiprl
J:' -0'
"" rlp"irprl for rI' "ih,,,tion,, rn nrovirlp i nci obf intn thp ei cenermatirm Ex-
< -,
ample LI below shows the graprncai solution. 2.5 Eizenvalues for the_Slah_ 57



." ".. ,.., • L U . •• ~,.I, •• • n .,,: .", ...
........ '1""-' ... '-I''' .~-. _ ................. ~-. • ~ ..... p ... ~ .. ,,, ......... ..
Consider the planar dielectric structure shown in Figure 2.5. The guiding index has value
1.50, the substrate index is 1.45, and the cover index is 1.40. This is an asymmetric wave-
guide. The thickness of the guiding layer is 5p.m. We want to determine the allowed values
of f3 using Equation 2.22 for this structure. Assume that light with wavelength of 1 p.m is
used to excite the waveguide.
•••••• •••••••••••••• . .......................
B ... ;::.I.'l-U"· .••••...••••. ............. ................. <i .
-
A= !.II





Figure 2.5 The planar slab waveguide configuration used in
Example 2.1.
Solution:
We will use Kj as the variable for plotting all the terms of the equation. This choice is
,~h;.~,,~, 1",4 ,n"IA h .. 4 ~h, '04" Q\ h".; .... ')..40 th4 ,f" th4 ,n .l, ..... J;
oJ 0 '2
A va: rabies ius be uennec In terms or K,:

'Yr = V fj - KiiUAJ

'r., vp /l,o\i.'1"J) \~.~'1")

{3 = Vk6(1.5? - Kj
Using these values, both sides of the TE characteristic equation (Equation 2.22) are plotted
as a function of KJ on the graph in Figure 2.6. The variable KJ ranges from a value of 0
(when f3 = konJ)' to Kmax = Vk'6nJ - k'6n;. The tan(KJh) term generates the typical pattern

1
\


r ar ,K(,
""" \
o
a -) ) ) )
.r' K 20000 -
roo 'V7V K
-
-"i
\

\

in \
v
rlgure ".0 1 lie V'Ul Ul "'ljuaUUll .:. • .:..:. lUI rue MIUWlIlH
Figure 2.5 shows four allowed K values. 58 Chaoter 2 The Planar Slab Waveguide



~f a r"","at;",o fnn"t;~n "vt"n.-1;no fr~", -ro t~ -I-m, Th" r;oht_han.-1 0;.-1" ~f Pnnat;~n ",""
-r -0 ~ -0' -a-
vieic s a sower tuncnon that rverzes towar 00 around K - LJUL em • and then comes

III llVll1 <«, "'L ure l'VlllL~ WIl<OI<O Lll<O LWV "UI V<O~ "IV~~, -a £..£.£. l~ • r uese
points represent allowed values of K for this waveguide. From the plot we see that the
allowed K values are approximately 5,500, 12,000, 16,500, and 21,500 em -1.
This plot was generated using Mathematica, although there are several other suitable
numeric packages that can perform these calculations and plots. To serve as a guide, the
Mathematica code is listed below:
n-f~1."n·
,
nc:: ,1 A.~.

ne .4(;'

-u , VVvo,
, n~ A ;
~~".~~~ v
k=2 Pi/lambda;
beta=Sqrt[ k~2 nf~2- kappa~2] ;
kappamax=Sqrt[k~2 nf~2 - k~2 ns~2];
gammas=Sqrt[beta~2-k~2 ns~2];
gammae=Sqrt[beta~2-kA2 neA2];
Plot[{Tan[kappa hI, (gammas + gammae) /
\Kappa\ 1 - gammas gammal;{Kappa<!))j, tKappa, I, KappamaXj,
Pln-tR"nno _""> ,1 n 1 n 1
w
1 ne transcenuentai characteristtc IIlUS( De soiveu numericallY, wnicn is a reia-
.1. F, '.n. ,1 A
'J '0' "J '0' '0
Math.emat ica the tollowino command was IlSP.eI IIv to find each root of rhe
equation.
FindRoot[Tan[kappa h) -- (gammas + gammae)/ (kappa(1- gammas
gammae/kappa~2)), {kappa, 5000}]
The last bracket of the command tells Mathematica to begin its search around a value
of K = 5000, The program returned the first K value of 5497.16. To find higher roots, we
,1. ,l- F, th ,l- I.
0' &" '0 r ,
accurate value .11 the eioenvalnes for I< were found to hp. 'i4' lfi OQfi,7
I VJJ 1, '"IU s: 1 J .. J "Ill . '11<0 <O~I:\<011V'UU<O~ ~11 L<OII11" VI P "all V<O IVUUU U~I<OCLlY 110111 Ul<O
, 'rl. ,1 ,1. t:< '} "J, OAIl!!'7 O'lt;.Il!! o'}!!! 0 OI'7"'} 1
, "0 .,' , , , , ,
resnectivelv,
. .
'111~ r ~I1VW~ ~Vlllv Lypl'_allvaLulv~ VI vpu._al wavvI:\UIUv~. ,'11M, Lll<O I:\UIUllll:\l11111
need not be very thick. It is generally on the order of a few wavelengths, Second, the index
difference required to achieve a guiding structure is small. In this case, Lln = 0.05 between
the core and substrate. This is actually a huge difference compared to many practical de-
vices which have index differences as small as 0,001. Finally, inspection of Figure 2.6
shows that if the waveguide is made too thin, so that the argument Kh does not extend
beyond approximately 7T12, it is possible that the two sets of lines will never cross, and
thprE', Ulill hp no moilp "lIoUlE',il in thE', ~trll{'tl1rp M"tPri,,1 oroUlth ,Ii~t~
0 .
get cauan t III t IS roniem a ter ISCOVenng t at an exotic new 0 tica matena t ev ave
-, . , rv ,_"
'_all Vllly "" l5'VWll 111 lay"'~ Lllll111'" Lllall v., /-,,11' ""'V," 111L"'"a' ~uall' 'UIll~ Ulv
hVE',r 2.5 Ei , for the Slab 59



Th" ,,;,,1£1,,£1 f', ,~ f'r a +r ,11 rl un
T J .~. 1-" ~ •• ~ .... ""V ''''u ."vu"'". "". uv"'''
thi" tTlP"n? P"f'h tTl"tip h"" thp ""mp 'nth nf' 1;"ht· th"" ,~"h ;"vt t~.,,,,, ;n .,
'0' '0' , 'J ' J'
Slightly dltterent direction within the waveguide, In the ray picture, the modes
would be shown as four discrete rays traveling at slightly different angles, as shown
in Figure 2.7. Notice that only a few discrete rays actually propagate in the
waveguide.
.......r
............. ...- ;:s"
-- _____..., ?
~;.-..-
... 1
"f
ns
Figure 2.7 In this ray depiction of the four
allowed modes in the waveguide, each ray has
the same magnitude of k vector. They are simply
oriented slightly differently with respect to the
z axis.

'T'. .L c, .,. L _L .._ ,L ... ~
~u LllU"'" vv ILlI ua"." , Lll" 1)1 UU1"HI HI DA-
.1. "I 1 ,1. 1, • 1, + • :1: 'T't. • .1. . _ .. .+. C. .J
"'"t''''' ~ .• ""v ... 'u .vv,~ '''''] . ~ .,," O' T ." V~ w" ... ,,"'u .v " .. u
thp "l1""'Pti pnpro-" pio-pnu,,111P" "f " n"rtif'1p in " finitp n"tpnti" '"P1 r Ll 'hp wrrc l ,
'0. -0 .r- r L ••
Of1V th~ nar-ncte-m-n-hov and the onrrr-al Irl~ nrohlern I!'< verv
~ . . ~ .
strong: both situations describe waves which are confined between two reflecting
boundaries. In both cases, the waves partially tunnel into the surrounding potential
barrier before turning around. Only certain allowed energies, in the case of the
particle, or transverse propagation coefficients (K), in the case of the optical wave,
are found to create a standing wave in the one-dimensional system.
To complete the solution the coefficient A should be related to a physical na-
Li:llllClca. 11 IJli:lI,;UI,;C,.tt L:; LCli:llCl lU lHC IJUWCI ca. ICU 11 lHC Wi:lVCgUIUC. IlC IJUWcr
.1, .+.1. n. .1.
w "'"."' .... "."'u OJ] .... "'0." .. "0.,,"' .. V~ .,,'" ~ V] """0 '''''''.v. v,"' .... '" ",.v",,-
v""t;, .n., .,~"., ,f' th" "";£1,,.
o·


S, - '2Re(E X H . i) (2.25)
Note that we are using the time-averaged power. The average power in a
TE mode is
r, .: r EvHxdx = (~) r IEvl2dx (2.26)
L- J 00 \L-WI-'-O/ J' 00

C': .1. ,I .1, rl: 1. .+
~ .. ""'" .. ,'" .... "'0·"' "1-'""" ~"'J ~"'" U""''''''~''' ... '" .... "'0·"· .. "" ....... " v. 1-'V ''''. 1-''''' ..... H
lpno-th (in thp " tiirpf'ti"nl,
-0
It IS milch more enhonrernno to S~~ the aetna mode that
':' . .
to eacn vaiue or p. using me rouowmg taatnemauca commanas to Implement 60

Chllntpr? Th» Planar Shh W"vP.l'lIirlp

th~ t.,t. A. ,hI. +r .11.

t'''r ... ~ .". r~' "

wave[x 1:= Exp[-gammac xl Ij x>O

wave[x_l:= Cos[kappa xl -(gammac/kappa) Sin [kappa xl/j (x<=O) && (x>-h)

wave[x 1:= (Cos[kappa hI + (gammac/kappa) Sin[kappa hl)Exp[gammas (x+h j ] Ij x<=-hj

amplitude=11 Sqrt[beta/(2 omega mu) * NIntegrate[(wave[xl)A2, {x,

-u.uu , U.uuu.::

n, _. ., . .••. n n

thp

.

Exarnnle 2.1.

13 = 91752

,,_ " ,/ \ _L_ ~

""YV ~ _/I \ I \

9t!819

_..., \ I <,

\

\ I

\. /

Mnno1 / \.

/ ~

~ ;

Mods O :7'\

13 = 94087

_R _7 _fl _" _4 _~ _':l -1 () 1 ':l

urn'

"!Jure ".0 1 lit: ur si tour 1 r. ~ ()I_LII_e

wavezuide described in Examole 2.1 are shown. The

vertical lines represent the location of the dielectric interfaces.

Since the waveguide is asymmetric, the modes are slightly asymmetric, although it is not obvious to the casual glance. Notice that the modes have alternating even and odd Symmetry. and that the evanescent tails of the hizher-order modes extend

he "IE mode has one zero :rOSSlnQ In the wavesuide, theE mode has ? n ... rlp"

etc, 1 nere wtll aiso De a set or 1M mooes wim simitar oesignauons,

carried in the evanescent tail? Power in the guiding layer is found by integrating the Poynting vector over the area of the waveguide structure. The fraction of the power contained in the core is simply

Pc/ad = 1 _ Pcore Ptotal Ptotal

(2.28)

There is not a simple, general, closed-form expression for these integrals, although they are straightforward to evaluate numerically. In general, higher-order modes

pendence of the mode c on SIze waveguide. Mode confinement is an important property for waveguide designs. A mode that is loosely confined will be more affected by bends and neighboring structures than will a tightly bound mode.

62 Chanter 2 The Planar Slab Wavezuide



u; I. 4', +~ A, _£"_..1
.'" '''' .. ,"'., .. ~ .. u "'~""""~'" ,~, .. ,'" '''' .. ~'''' ,~ ~u~ 'u .. , .. ,'" 6"'U"""' H""~ U"''' .. llp-
t;",n ",j' <> "",,,,rI,,, ",;th;,., th", ~""",,,,,, .. tr;" dr, ,,,tur .. (j'nr th
'J v- ._ .~ ... - ~]~."'.u
shown III FIgure 2.9) IS:
E = Ae-y(x-hI2) for x;;:>: hl2
y
cos KX A ~in KX
E =A or for -h12 < x < hl2 (2.29)
. y cos Khl2 sin Khl2
E = + Ae Y (x+hl2) for x< -h12

'T'~ A, .4' 'T'l"" A, 1. _1 ,1. """.
>u'" -cs T ~, 'U'" > .... u~~'" "' .. u u'" • .111<01'" aH;
tuT .... f .... r thp .... f thp fiplrl in thp onirlino l<1vpr ... nrl;n~ n,.,
-r ~ ~ -,; , 'r .''''
whether ~ (cosme) or ann tr: I~me mode I~ p-yc.jtP-ri 'hp f<1pt
a e mo es can e unk UeLY c arac erizeu in erms 0 even or 00( QTOU S IS a
natural consequence of the even symmetry of the index structure. The sign of the
field in the lower substrate is positive for the even modes, and negative for the odd
modes. The characteristic eigenvalue equation for the TE modes in a symmetric
waveguide is:
'Y for even (cos) modes
tan Khl2 =-
K (10)

I<
for .... clrI ("in Tn .... rlp<:
'\

ue cnaracterrsuc LIlt: IV llUUCS IS
tan Khl2= (~rl for even (cos) modes
n, K (2.31)
= -(:;r; for odd (sin) modes
r
'"\. UlUlIUC lcalUlC Ul UIC ". " ..... " ... " '" 1" urai 11 ca atways su (JO a eas
" . _'. .L £ .
Vll'-' IHVU'-'. '11" 5'<1PH'''<11 'v, <1 ".1. w'
;,., =<'v<>"",n1 .. '1 '1 1, .. 1",,,,
T . .


Example 2.2 The Symmetric Waveguide
Suppose the waveguide shown in Figure 2.9 has a film index of nf = 1.49, and cladding
index equal to n, = 1.485, The difference in index between the two layers is very small.
Let's calculate what the allowed values of f3 are for this structure. Let the wavelength be
O.8Mm. We will use the graphical solution, as it best demonstrates why the symmetric
.:11 _1. '1". .:11 ,
'u ''''6"'U'"' H .. u, 'UJ~ ~"yyv' u. ''"'u~, v.w ... vu,",. , 'v , .. , u'"' , "
·~Il.m ann h Ij,Lm
-


rh .. ~ .. ~~ .. nnl hun "~r;~hl .. c ;n ,h;o nrnhl .. ",: >n.-l 11 "I. :11
~J . . , l'
functions In terms ot K. 2.7 The Svmm:;;tfiC

63

2

T s I-' "'O"s 0 "f "s

,.. 'u'"]

Plugging numbers into these expressions, using ko = 27T'/A = 7.853 X 104cm-l, yields

'Y = v'9.176 X 107 - ,(2 (2.33)

f3 = v'1.3694 X 1010 - K2

To find the eigenvalues of the TE modes, we must solve Equation 2.30. Graphically, the

,~-I./') »t , ,~,-I A, ,r'" ... 1" .. ",,-1 ~~ .h", On~'" nrn ... h nO n f"~M;~~ ~f 'rh","",

" r 0 -r

lire nlotted azamst K for the case where h - ",ll.m m Fipure 2.

"

IV \

-

K

o

-5 \

- KI'Y ----+\

,

I

-1 n I

.". . "c' "

.. , .. on D,

auowen moae, wmcn occurs near K ouuucm

The top curve, which corresponds to the even mode, begins at +00, and terminates with a value of O. Notice that the tan Khl2 starts at zero and increases. It is inevitable that the two curves will cross, so there must be at least one mode. In fact, we can generalize this statement: a symmetric waveguide will always carry at least one guided mode.

As the wavezuide is made thicker more modes are allowed. Consider the zraohical nlot

-0 •

. ~ .

UI LIlt; lUi Ult; (;Il~t; wuou LUt; ,a'''5u<u" MIlU I~ IJlLlll UU(;l\., as ~I1UWll 111

IU

0 ..

\ I ..

r"n ll'h I? ..

__ y/K

'" 1-"

odd modes.

( I"lgure z. I e miCK waveguiae supports nom even anu

2.8 INTUITIVE PICTURE OF THE MODE

guiding layer of an even mode in a symmetric waveguide has the form EyCx) = A cos KX e-j/3z

Since cos KX = (ejKX + e-jKX)/2, we can rewrite Equation 2.34 as

(2.34)

components vector wave

with it. These two plane waves zigzag down the waveguide, continuously crossing each other's paths as they travel along. Being excited by the same source, the waves are coherent with one another at a given plane in the waveguide, and so they form stable interference patterns. The structure of the mode is a result of this interference. When the modes constructively interfere, the electric field is maximum, and

occurs, IS a mirumum. IS

sketched which conforms with the interference pattern that would result from the two interfering plane waves within the waveguide.

2.8 Intuitive Picture of the Mode 65


~ A. ~ Wh" 1!Il R. I . tn r.l!3lrtAin VAil II!3IC?
'1

IWO nronerties 0 the eizenvai es have become lear- onrv rnscrere values ot Ii are
allowed, and the magnitude of {3 must lie within a special range for a mode to be
guided.
Why is {3 limited to values lying between kanf and kans? Let's use the ray pic-
ture to explore these extremes of operation. The plane waves of a mode can be
represented as rays traveling at an angle
!j
an H- 'L-ih'
K


10 e manar wavezurce. n e rav mcture e uorer mu on the va ue 0 Ii
Le;ple;~e;IH~ i:l wave Ui:lVe;llll~ Mli:ll~llL UUWll Llle; 4. i:lll.l~, wruuu L;UI LU K
in Equation 2.7 ({3 = kanf)' Physically, there is no way that {3 can be any larger
than this value.
The lower limit on the magnitude of {3 is established by the critical angle of the
film-substrate interface. At the lower limit, {3 = kans' so K = ka V n} - n;. Plug
this value into Equation 2.7 to find the angle of the ray at the lower limit of {3:
K.ans
Li:lU U
«s v n: n~ /,.. ",."
J ~~.J


\, n:: n.~
J
Compare this to the critical angle as defined by sinOe = nslnf' The tanOe is then
sin Oe
tan Oe =--
cos Oe
nJnf "
~~.JO,
V n~ln;:
J

n.
'v n: n~

Note that the final form of Equation 2.38 is the same as that of Equation 2.37,
indicating that waveguiding ceases once the incident angle drops below the critical
angle for the dielectric interface. This answers in a ray picture why there are upper
and lower limits on the value of {3.
The discrete nature of {3 can be found using this same model of interfering
waves. To avoid decav of enerzv due to destructive interference as the waves travel
. .c ,.
Lue; W<lVe;~UIUe;, Lue; LULi:l1 pUi:l~e; ~ui:lu~e; lUI i:l PUlllL uu Lue; Wi:lVe;UUllL ruai
c. .c. {y • L I . .l L
""Vv'" HVU' V"V UlLv""vv V' v J LV LUv "vAL \'" , "UU LJ"v~ "5'"", UlL'''L UV
o rnllltinl" nf '),.,.. H"nr ., u,,,,,,, in,-.ir!pnt o+ "nolp II " nh"c" chift nf lr ... J, "nc II lc
. '0 , r
::Irrllml l::ItprI on thp fir-st tr::ln\:vP.r\:p n::l\:\:::ICfP rh rrm ob thp film ::Inrl::l nh::l"p "hift of
. '" '" c
'LCVe occurs at the h1m-cover mtertace, Another knfh cos fJ ot phase IS accumu-
lated traveling back down, and finally there is a -2<1>s phase shift at the film-
substrate interface. The transverse resonance condition requires that 66

Chanter 2 The Planar Slab Waveguide

'0 T -" <r- -a-

waveguide. We will use it in the next chapter to develop a generalized dispersion relation for slab waveguides of any construction.

2.9 PROPERTIES OF MODES

Once R is ined for a wavecuide the field amnlitudes can he -l. ,r! in ,,11

reztons 0 e wavezt we US10 b ua IOnS 1,.1. or tee urva ent or 'M wave"

.. .,. ...

VYC 11i:lVC UCC11 lClCl111l~ LU UIC1>C 11CIU UIMl1UUUOll1> i:l" moaes . .1 e conce JL 0 e

-' ~ -' .c_ . •. -' TT

lUVU'" ", ''''1 pv ''''' .ur -auu p,",ulap" ainu", ,",vu"""'u.s 'V "I'" . nClC WC

The general expression for the electric field solution in all space is

E(x, y, z) = E(x, y)e-i/3z (2.40)

There is a corresponding magnetic field distribution for each mode, given by H(x, y, z). The term E(x, y) describes the transverse structure of the field, and is usually called the mode structure or mode shape. The modes have the following

every P l:UlleSpU11US a UlS 111;1 OUC 0 e svs em. every ~

2. Most modes will not be guided. Most values of {3 will lead to unguided, or radiation, modes. The spectrum of {3 for unguided modes is continuous, meaning there are an infinite number of unguided modes.

3. Only a finite number of modes will be guided. These are associated with the solutions to the eigenvalue equation for {3. The spectrum of {3 for guided modes is discrete.

4. A mo es are or ozonar. IS IS a very im or ant Oint. r-or QUIGe modes

UI . y HOY UIH01> Llli:lL

r

o

I ru {. .\" U r. ..ll_-L

/" . "

~l 0

,,_.,

where 8ij is the Kroenecker delta function, the area of the integral is the infinite xy plane at a particular value of z, and where Ei(X, y, z) and

HJ<x, y, z) represent normalized modes of the system. For radiation modes, the formal relation is

r

o

/" .,,'

',_.,

r •

rhrw mtmite nlanr- wave" Each mode 1<; nnirme and cannot hp

.. .

uescnbec 10 terms or otner modes.

2.10 Number of Guided Modes in a F.7



r- n . ...... rate IllUUt:S wit snare LIlt: same va ue 0
..). ..:lVlll!;; llIVUC" ill C LalC •
n .n -l." .1 • r- .. . .. T
,." ........ 'il' HU '" ... o .. H5 "'''' .Ule- ll!;;IU Ul"U . III "Ue-H
degeneracies, field solutions can be found which are orthogonal, and they will
satisfy Equation 2.40. A good example of such a degeneracy is the
fundamental mode in a circular dielectric fiber. The mode can have two
different electric field polarizations, Ex and Ey, each of which has the same
spatial energy distribution.
6. The modes of a given system form a complete set. Completeness means that
.L • L .r .1 .
'H" " .. V '''U ULVU"" "I:""H LU" "HLH" '1' VI lllC ". r"lCl11. rJ.". '"
rI;~.~;h"';~n nf ala~'~;~ fip1r1 ""n -:i -:i ~~ ~ . ;~n ~ .. ~,,~ ~+'
-I:" '1:" , ,
thp. iatetv wero htert mortas nf thp ,rip
'0

5~<U
E(x, y, z) = ~ aiEi(X, y, Z) + J, . . d(f3)E(x, y, z, 13) df3 (2.43)
; radiation
where E;(x, y, z) are the discrete modes of the system, the a, are weighting
coefficients for each mode, and the E(f3) are the radiation modes of the
system, with their respective weighting factor, a(f3).
1 he power or me moae concept res In commeteness and the use ot SUDer-
.. ........ . . ~ ing oe-
ptrsu. tun. ~ HI" "UjJ!;;1 e-Vlle-!;;jJl 1" verv poweri Ul IUr l:OU
-l:CC. UT. . !! .. .. ... ,., T
• 'vvH • ,V U .. '''LvH. "J "L"LU". '" v .... -.:;oupn:rrg- HI U!;;lall 111 '--Hal'l!;;1 11. III
o",n", .. o ., oi"",n 1Tlorl", in nnp e"d"'1Tl . ha ~~ ~ .. ,J' ~ .rI,
~ , 0 -.; "I:" Lt-
in " "p"nnrl ~v"tpm nf whirh "omp. rna hp ol1irl",rI "nrl e01TlP 1Tl"', h" .. "rli"ti"",
J -= -.;
2.10 NUMBER OF GUIDED MODES IN A WAVEGUIDE
We often use terms like single-mode or multimode to characterize a waveguide.
The importance of this distinction will become apparent when we discuss infor-
1Tl.,tion h"nrl",irlth in th", n",vt "h"nt" .. ~nrl ",han ",a rI;~~"c~ .1' .r. hahuaon rI
-z- , T -0
vi rp. " l n rhic ~prtinn WP. lip.vp1on "n--;:n;;- ,nne of th", nll1Tlh"r of 0'1 irl"rI
L -r+r- o·
mo es in a anar wavezuio e.
~ rue piauar waveguide ucsc: . oea ne wavezuide su -
L,L .
-l • pp ,J.. • J' ...I, .. TC L _,
1:"" 'U UUL,",L'_UO ~L L"~U"~ U'"'t Vll Hi> ,,,.U 'v Hau
the relative indices between the layers, we would also have found that the mode
number varied. The lowest order mode has a k vector that is nearly parallel to the
z axis.
f3lowest order"'" knf (2.44)
The highest order mode will have a wavevector at nearly the critical angle.

D a 1. f" 'C
r-mgnest oraer "f"V cnuca ""s v .... T-' I

T'h» rp"t nf thp. Tnnrlp." will h~vp pjop ... ".,l11Pc ~a th.,t f.,l1 th"c" h,,~
-0
extremes 'In !Jp.t an dea ot the number nf Tnnrlp" in thp ilip. rpr,,1 thp
~
general etgenvattre equation ror me 11:, mooes, 68 ~ ~M Slah Wavecnide




C 1s
,all\l'" IL.4t
'V" \
K\1 K2)
Graphically, the two sides of the equation can be plotted against K to create a plot
such as Figure 2.13, which is a replot of Figure 2.6.
,
in
I

\
\
\
0 \
I Tt: ~ T~ I T~
I (A-- -u ----- --- -;,: I~ ~;j
0 ~I _/I ../ 1 ./1"-- ...
:ruu .-iI? 150(10 K 20000 I ( K
1 I 1 1 I
-5
\
I
." \
v
-iyun. "'.1'> • lit; 'UlUL;Vll to we e equa Ion ror an asymmetric
waveguide shows that ever v time the arsument ". h in{"rp~<p< hv -tr ~n",hpr ~ ria
• • • TL ~ •• -~
~6' • ~v'-JaJ JJJVU.-, all;; ,HV VJJ JV
the structure described in Example 2.1.
.'
The right-hand side of the equation starts at zero, slowly diverges to -00, then
comes in from + 00 and terminates at a value somewhere above zero at Kmax' The
left-hand side of Equation 2.46 is a periodic function (tan(Kh» that goes from -00
to +00 pvprv timp ... h inrrpl'l"p" h\! ,... Nnti"" th<>t if th" ,,,,I,,,, nf U' h:n ~_~~.~_
J max: <5'
TIran 1T7T.TIfen we are zuaranteea 01100 at Teast one TE mode in the 'l/l'I~
TP ~ ,~ ..
11 Kmaxft -- .J 7f1 c, urer WC ill C guararueea lC no al eas wo mo es 10 e
• 'T''_ .1..J: -;- p, 'time
r«' - b . ~"'"' 'U'Uv" U' "-max -an:;~TUTUJJ CUftUtttUftS . .DYCI
,~ l. i"" .. a~0av h" ,... ""nth" .. ,~A~ : n ~. I~.,,~A "f'J-. .L .£
'max' OJ , "I:'r H' ,'-',",, u,
modes, m, can be found from
[hKmax]
m = Int ----:;;- (2.47)
rhk(n2 - n2)1~1
= Int L s
I =tr I

'Y" • • L . _._ T"' _. ..
--r-rITI"J "P P' a o v,",,, .... ..., .. '''~oV llUIllUv1. "VI ,H 11'
~'P <> 'f'j:l rnnrl" p""n llhpn I; in :1' lanv th~ ... _/'1 {vaa n. _. ~
-v max '0' 'J -J.
'hp nnrnbar of monp" m is l'In annrox irnatirm <>nrl i c n c l l o chn"lr1 h" _"n ... AaA un
• ..- OJ "p
To TIre nexTTarger integer to get me accuraTe numoer of TE modes that wJITl)e 2.11 'Ihe.Numerical.enenure,

o 'J 4J:'J:' ~ •••• , "'U'

i" not l<nouln Not", th<lt th", mort", ('ollnt in('r"'<l""''' ",ith th", th;('l<n",,,,, h

ot the guide, with the ditterence m index (nJ n;) between the core and cladding, and as the wavelength of the guided light gets shorter, Also note that the point at K = 0 is not considered to be an allowed mode, even though it appears on the graph that the two equations are crossing at that point.

We usually characterize a waveguide by its normalized frequency, defined as

17 .22,112

In terms the norma izert t rermencv the annroxirnate number of modes m in a

waveQuic e is m V 'TT'. e mo e cu 0 con 1 IOns are usua LV escnoe m terms

r $'..., '$" • .• 't.

Vi UIC HVI ' llCYUCH..,Y. i'VI , II II I" UCMICU LV UUllU a wa LHaL

only carries the first three TE modes, what should the dimensions and index difference be? (See Problem 2.19.) Since the cutoff depends only on the V number, one can trade thickness for index difference to achieve the desired result. The normalized frequency helps establish the relationship between the parameters that influence the number of modes carried by a waveguide.

Why do we worry about the number of modes in a waveguide? The answer is

r . .r- '0 ,

in a rlioit<ll cornrmmir-at ion svstern thf' nntse enf'rov wi di str-ihurerl over

all the allowed modes ot the wavezuide. 11 eacn mode travels at a sllllli11Y__illllerent

• r.

VCiV~lLY, as I" ViLCH LHC ..,al>C, LHC rerupm eu iVllU Vi Lue pu rsc w ..,uau~e al> L ie

,.1 .,..L· , 1 .1.l .$'$'. .l. ••. • _1.

PUH'" l: .r- ~ •• U .. , '''U5LlI''U", lU" PUI"" UUU ""'-'''Ll'''IY ,lU" IUl'-' al VVIU"U

pulses can be sent. Different types of waveguides can be built to minimize this effect. However, before we study them, we need to understand dispersion. Waveguide and modal dispersion in the planar waveguide are discussed in the next chapter .

., 11 T .... ~ NIIU~DI"'AI AD~DTIID~

A v",rv n<l' for ('h, iz irro w<l"",ollirt",>: i" th", .. ,mpr;rn, a no r:

ure. ne COnCeOI IS ase on rav racine an re rae Ion so 1 IS 0 LV ant ica e t

multimode waveguides. Consider the optical waveguide shown in Figure 2.14, where a high-index layer with index n] is surrounded symmetrically by a lowerindex medium, n2' The thickness, d, of the guiding layer is not critical so long as it is many times greater than the wavelength of the light being carried. We want to explore how light (in the form of rays) can couple into the end of such a structure. A ray is shown entering the edge of the waveguide, where we assume the index of

J'¥nn<4A • OR t~ ~n, 'r Inn, h4nA;nn t,,,,,n¥A ,1' 'T'~

'0 , -0 '0' '~"'.~J

travel" until it strikes one of the dieler-tric interfaces Tf the rav strikesthe .

a an ans e excee ina e cntica am' e tJ.

sm .n-Jn 1 wII De totauv inter-

nauy r

ana zigzag irs way oown me waveguioe.

Figure 2.14 A ray incident on the end of a waveguide is refracted toward the axis, and subsequently undergoes total internal reflection at the dielectric interfaces.

face. As it bouIl(;~~ down the waveguide, a fraction of its energy leaks out at each reflection, and the guided energy is attenuated. Such rays are unguided.

Figure 2.15 A ray incident on the waveguide at too steep an angle win not satisfy the condition for TIR inside the guide. The wave partially reflects and partially transmits at each reflection.

face of the waveguide, we can determine the maximum incident angle, Omax:

the refracted ray will strike the interface at the critical angle, and thus be totally internally reflected. (Jinc is the maximum angle for a guided ray.

2.12 Summarv 71



tl - ~.'n{O(\ tl \
me "1 'V v cnticatr

n cos. (1, .. >11

= nl Y 1 - sin2 ()critieal (2.49)
= n1YI - n~/nI
= YnI - n~
The numerical aperture is defined as the sine of the acceptance angle:
(Vii sin!1. vn n;; L.)l
A" Wp will "pp hltpr thp NA i" ~ n c .. fnl for lnr-o .. _rorp, rrmlr irmvl .. \lJ<"'''_
. ~
curdes where nronasanon or !Jh 1 S rather rntncu to exnress In terms ot electro-
ma~nedc neios . .t"ra~maricallY, 10 couple li~m imo a wave~uioe if is essenuai mat

1.' I.'. -' . • 1. d .
Lll'" "bllL U'" III "U"'ll a vvay LllaL all lay" "'" vv lLIllll Llll" aUblv Vma,x.
Example 2.3 The Numerical Aperture for a Symmetric Waveguide
Consider a symmetric waveguide with a guiding film index n, = 1.5, and surrounding
indices n2 = 1.4. The numerical aperture is directly found from Equation 2.49:
NA - "\/152 - 1.,12 - f'I."1.Q (')"1\

'T'L'
•• u~ ~V"~~pVU~~ LV <UpUL "Ue'~'
·1 'n n .0
Umnx 1S111 ~ V.J.J:J') .J~ l~· ,~.

so the uu WI tn 0 the accentance anzre IS about 04".

n=1.4
n=1.4
Fiaura 2.17 'he a anzle of
• Le.O ,,'-
.... 0 O""HU'~'O v' •• w .. ~~ ",y, "'~
numencai aper ure uoes no . uepenu
on th .. rlim"n.inn. nf thf' wavao uid ..

vOUJ LH~ .~,"u,~ ... u.~~o.


According to ray analysis, any ray incident on the waveguide within the numer-
ical aperture will be guided. The NA effectively defines the cone of acceptance.
Only for large structures (where the guiding film thickness is on the order of 50-
lOU A) IS the ray mcture reasonably accurate.


I) 11) c:.IIUIIJIADV
_ .. ___ ..........

Wp have introdnr-ert rnanv rmnortant rnn~p.nt" nf ontir-a wave.ouirt .. " in rhi s rh~ntpr
- . . . ~ .
We usee formal eiectromagnetic analysis to solve for tne allowed nero structure
inside a dielectric waveguide. We found that boundary conditions establish the con-
necting formulae that define the shape of a mode. We derived a characteristic 72 Chanter 2 The~·'~n,.;rI~



+" .L 'T'r. -' 'T''' If 'T'L' -'
.~. ~~ '" .... all\.1 '" ~n ~"V\.1,",". '" Ill" ,",':!uaUVll 1" u , "V 1L 1 t:1I Uln::S
nn",,,,r;,..., nr ~_1 .. +;_~ n -'h +'. .+' _1
b' -r ,~.u • n .u~ ~. ...... ll
We distinguished asymmetric guides from symmetric guides, and noted that a sym-
metric guide will always carry at least one mode. We reviewed the mathematical
details of the mode concept, stating but not proving the important properties such
as orthogonality and completeness that a mode solution will satisfy. These proper-
ties will be used in later chapters when calculating coupling. Finally, we described
the numerical aperture of a waveguide, based on a ray tracing analysis of the total
mterna re ection cone 1 on tor ~ waves 111(1",



H~t-~H~Nl :I-~


c. UlKOV c anter ':J III Mathematical rnvsics kea ina MA: AGe ison-wes ev
Publishing, 1968).
2. J. D. Jackson, chapter 8 in Classical Electrodynamics, 2nd ed. (New York: John
Wiley & Sons, 1975).
3. H. Kogelnik, in Integrated Optics, 2nd ed., Vol. 7 of Topics in Applied Physics, ed.
T. Tamir (Berlin: Springer-Verlag, 1982).
4. For example, see S. Gasiorowicz, Quantum Physics (New York: John Wilev & Sons
lQ'1.1\,

<; A II.,n U (' ,,-1 .. _ ~_rl T.-1.._ 1"\. T ~1..~_ 11 . fl, UT_ '.1. 'T'L '''T
"J , -r- ~6' .~ •• ~v V'"
York: Chapman aiidHall TID. 19!B)'



t"HU~Lt:M::;
An asterisk indicates a problem that is best done with a computer,
1. Using the boundary conditions for a TM wave, confirm Equation 2.23 for the
eigenvalue equation of an asymmetric waveguide.
2. Develop the eigenvalue equation for the TE mode in a symmetric waveguide.
3. Develon thp . 1111" for the TM modes in " ir- w"vPPllinp
~
"'4. For an asvmmetnc nlanar inp with n <; n 4- "nn n (\
" ,-!_ .
determine ure i1UUWt;U Vi11Ut;~ Ul P lUI uie 1 n UlOUe~ in a waveguiue Willi trnckness
7 ,~ "",,_ +1.. .~',~ . +1..' 1
'0' 'I'"
*<; POT thp "" ""'nn~",~",rI in .1 +inrl th", ,..nt~H' ""'. ,I •• nth, 1', ~ 'T't:)
'0 , '0' -o
ana 1 tl modes.
*6. Which is the lowest-order mode in an asymmetric waveguide, the TEo mode or the
TMo mode? Prove your case graphically.
7. Use the Fresnel reflection formula for TE waves to determine the attenuation
coefficient for a leaky ray in a waveguide. Assume that the waveguide is symmetric,
and is SOp,m thick. Assume the power lost per reflection is given by 'fl = 1 - 'r2"
~ AA ~ ...
.. llv' ~ .~ 6 .-.;:;rr-uy "'1 1.7V. LAp.""" w" all""". III '''1111'' Ul lH;;p"''''u.= .
Assume the 0,,;1i;"0 J,;"'PT h,,~ ;"lipv <; thp ~"TTO,,"Ii;"o l<1vPT~ h<l"p ;"lipv A<; 0_,.1
-OJ • ,
e ray is inciaen a
rsvaruate urc size or a connneu 1 n moue as a unc ion 0 me gur mg rum
'0.
,.., ._, .. ~"''' <'
U ~J 'U'V6un." HLU 6un.uH6 wu"" v. -'.J UHU
surrounding indices of 3.49. The wavelength is Ip,m. Problems 73



n. \X, • ~n ml;~:' rl. ;~. ,1' th h .• lrl 1', th Tn ~ .rl,
'r 'r -u
h ")I"finp thp mntlp .;71' "f thp fiplrl hv thp rI;"t"n('p thp n";nt,, ",hprp thp
,
amputuce IS reuuceu to IlL or me peak mtensity, rmo me IUIl wmtn (at nau
maximum) of the TEo mode for film thickness h, ranging from 0.1 to 10Mm.
Evaluate at least five thicknesses.
c. Approximately what film thickness leads to the smallest mode?
d. Calculate the mode confinement for h = 10Mm, h = IMm, and h = O.1Mm.
9. For the waveguide described in Example 2.1, at what distance into the substrate is
.l. h.l,.1 (I (1(1 I r:;' r:;' • ,L .1: .J' .')
... - .. , 1--' ,~ -u, ~u r ~ ........ ~--.
nl"tl"rminl" thi" for I""rh m"rlp. At what . int" the r"vp.r i" t he fip.lrl thp. ""mp.
value. ' nat noes thrs teu you aUOUl uow luicl<c to make rne clauuin~ iayers on a
.o
''''-0'
If) (,"" ;rl. ~ 'th ~ n ;,.1; hl~ " thO ,t h thO ,t hi,
-0 o: '0 , -J
with index 1.4:>. The wavelength IS iMm.
a. If the guiding film index is 1.5, calculate the numerical aperture of the
waveguide.
b. What is the mode confinement for TEo mode in this waveguide?
c. If the guiding film is made to be IMm, what is the mode confinement of the
TEo mode?
'11. Corrsider a planar sian waveguiae or innnire extent in we y anu z drrectrorrs: 1 ne
.. ,n hI • A, " th ,l. "~, . rl. • I All ~. ,,.1 th . I (I
o· '0 . , , ..
'hp. thickness is h. - :Il.m 'he W3VP.P'lIiOp. is exciter] with 3 1.'i r.m source

101 1 D 11IUU"'''.
nTL . C ..l n .1 . C_ .I •• .r.
~. 'a", .~ ,a~ ."a6~ v. "av~~u i'-' '"'U~~ ,v> "a~ '"'~6u,u~.
b. What is the numerical aperture for this waveguide?
c. Numerically or graphically, determine the allowed values of f3 and K for
h = 2Mm.
d. How many modes will this waveguide carry if the excitation wavelength is
0.600Mm?
12. Show that the eigenvalue equation for f3 for the TE mode in an asymmetric
wavezui e re uces to t at 0 t e svmmetnc wavezur e wnen n n.

ts
>flOW a flr !C.V or a wave.
f.LW
ror an asymmelric planar waveguiue wim nf IHO, anu a thtckness or
1 q. I.;}, ns
II Tn 1'. I - I tho 1'.
, 'J '0
conditions:
a. Covered by air (nc = 1).
b. Covered by water (nc = 1.33).
c. Covered by another substrate (nc = 1.45).
d. Explain in words why the number of modes did or did not change in these three
cases.
ickness waveguiue is neeueu LO increase we num
e. n .'+~, W a er 0
..l. .')
a.vu~~ 'v ,~a.
f U"", urill thp "fTM ",,,rip,, ""r,, in n, b, ,inri r?
- -J
*1" r' >n ., o"'-nrn"'ri~ ",ith H 1 " H • .17 1 ,,-n, <onrl j, ~ ,rn
"J "0' '] . , " ,
determine the tracuon ot power carried m the cladding tor the TEo and TEl modes. 74 Chanter 2 The Planar Slab Wavezuide



~ r,. .. ,. ".,~ ror LIU:; "ymmelric
iV. iJlOil VIO LlllO 1I1VUIO "ULVii iVI LlllO i '""v Il1VUIO 111 LlOilll" Oi
~
*17. Repeat Example 2.1 for the TM mode case. Make plots of the allowed modes similar
to those shown in Figure 2.8.
*18. Plot the mode profiles for the TEo and TEl modes in a slab waveguide with a core
index nf = 1.5, ns = 1.49, and nc = 1. The film thickness is lO~m thick, and the
guided wavelength is Ium.
*19. Consider an asymmetric planar waveguide with a film index nf = 1.50, a substrate
in ex n .'+~:, an a cover III ex n; .'+~ .
v c .L .s: "
.... L1 Ul" ...... UUIU ' ...... 1"'''0''1 VI LU" OUl" ... " ~ 1" 1 ~ WlI<U 1" L1110 llli1l
thp o nirli no laver can hp to snrmorr thp 'E rnorle?
, '.
D. ncrease e micKness 0 me gui e oun III ar a ov ercent an etermille
,I D ,+.1. Tn ,r1,
'0' r-:
c. Calculate the confinement factor for this mode which is near cutoff.
d. Plot the mode profile for this mode. What can you generalize about modes near
cutoff?
*20. A planar waveguide is made with nf = 1.48, ns = 1.46, and nc = 1.44. The
thickness of the guiding film is 10~m. What is the longest wavelength that can be
carried in a TE mode in this waveguide?


nFSI(.N ~~[)HLI, .. MS


L . wnte a genera rozram usmz any convenient so tware ackaze e.z. matnematica
n ',," -".
VI ' \"'0" L a""al Vi""'" WIll . L11" ' Vi a "'au vvavlO~UlUt:,
find the normalized amplitude of the TE and TM modes, and plot the mode
structures. For data input, provide the three material indices, the guiding layer
thickness (the cover and substrate are assumed to extend an infinite distance beyond
the layers), and the operating wavelength. Such a program will be very useful for
many of the problems in this chapter, and in those to follow.
22. Repeat Design Problem 21 for a symmetric waveguide.
L.~. uxtenu we ueve opmen 0 we moca eigene4ua ion 0 a our- ayer s rue ure wit
c, • • 11 :rr • . .. : +L' _1 •
IVUl I:'VL"'''''''''J U<11 ....... "L 111Ul ...... o ""U L' '" , .. J .... ~ "'''''''1 .. ,," "uv~LIaL"
~r~ to extend an infinite distance bevond the lavers) Perhans the ~i",nlp~t
'.
way 10 uo III is rs to wri e lilt: oounuary co rdt ions in a rna rix orm, an use mear
~In"hm t .. ~hn·n .. , 1'0, ,;I th, n,.~ ""'- .c, ,,£_
'0' -r- r J P Jo'",,,
WIth some SImple structures.



T"""o.TC"'I"""'T""""""'C"'ITr.....lIo.T TlIo.T ~~TA~ TT""",...,TTTT"""o.T"""C"'I
IJI' .... r K'IIJI~ I ~ \IV~vr.ITI I I Jr.' rW.llul ~
---~--~ _ •• ~ •• _, • v

.








CHAPTER OUTLINE
3.1 Introduction
3.2 Three Types of Dispersion
3.3 Material Dispersion I
1.4 Modal OJ
~'\ n;<.:npr"ion
-0 -r
~h Norrn!>li7prl Pron!>o!>tion
-C'O
'1'7 c· c++ .. ~ ~+ 1\/I"o"r;o 1\/1" .rI,1 TY
,~~ .,uy~' U'~H
3.8 Summary
3.1 INTRODUCTION
One of the strongest motivations for using optical waveguides is the large infor-
mat on canacirv or an nntrcal lmk Much ot rrns canacirv comes trom the hioh
, , ~
carrier rrequencv or me ligm itsen (on me oraer or W' HZ). II even .i.nercent 0.1
, , . .,.,- -,-, .. . '"
un; lVli:U U<111UWIUlll <1V<1U<1Ulv '-VUIU UC , Ulv llllVll11<1UVll VI UHe
_1. ..,11 AS ,1. nT1. :1.
vy,n".<. u._...... 'VLUU v,", ,",uVLLISU LV UL .. 'U',", V,,",, ,V "UU". Utl" H '" LLU-
1.1..-.>1.. thprp "ii p"pr hI' rlo~.o .rI for th!>t rn!>n" "irnlllt!>n",oll" t",l",nhon", ,.!> I" on on",
-J -J ,
line, applications such as video require large bandwidths and can take advantage of
such capacity.
The system designer must be aware of the fundamental bandwidth limitations
in optical waveguides. The most prominent limitation is dispersion. Dispersion de-
scribes the spreading of a signal in time. Figure 3.1 illustrates how dispersion limits
thp infor mnrion of ~ cornrnnnir-af ion {'h~nnp.1 At thp ;nnnt ~ "pr;p" of
. ,-
uses re resen ing er a s inarv in orma IOn are aunc e on 0 an onnca
~.
waveguiue. l.Jl1!1J.1Cl MVI l;aU1!It:1!I eaci. Ul L11C1!IC J.lUl1!1C1!I u, spreau ue, yy e t. ev
.1. .1. .1 1. ..,1 .1. .1. _u, __j,
u." ''"' U, LUV VLLLYU', ... v pu."v" UU'V LV LUV pVU" 'U". V ... v. VV6U' LV
~or;~,,~l. ~ •• >rlq~ ~"l ~o~ 'rho ~~r"qrl~"n >++,,~t' ,h . oJ.,
'J -r- --J r- -r- -r -0 -J _u
the maximum data rate for a communication link.
75 76 ( haoter 3 Disoersion in Waveguides



vlylli:l.llIl
'"
-0 ;:,iqnal uu
;j
JUU~I Dispersive Channel I----+-
..
Time Time p
Figure 3.1 Short optical pulses enter the dispersive waveguide. Due to dispersion, they exit with much longer
temporal distributions. If adjacent pulses are too close, it will become impossible to distinguish one from
another.



T. .t. . ,t. .t.. ..I ..I:
u, .... o ~"Up'~', ,~ '''~ ,."" .. "~,, ,~u~ ,~ ~'''P~' "'~" .. , ,u,,,,-
. rI. A ~ ........ rI n, ith thie ,rli,.,o '''Q ".,,., tho e ..... ,,,i., ;7. ,rI n,.,
o· , -rr -r
omrtes sucn as the craded index anc smote- node waveomde that are rtescrrhert in
~
later chapters. There are many clever equalization techniques, both optical and
electronic, which are being developed to extend the useful bandwidth of optical
fibers. These topics are not covered in this chapter, but are referenced where
appropriate.
.
_, ._-- _ .. ,-_ ....... - - , ... __ ...............
.l,':: I nn~~ I 'r'~;;:' vr LlI;;:,r·~n;;:,IVI'll

n., ,1 .t. .c .. 1 -1: .J.
'-Jp"~"' VYU' '''''P'u. ..,,~~ ,;, P"''' ~. ~'''P~' "'~". ..... c- , ... "'''y<-'",v", 'w .....
rI;cnorc'nl'J "nrl ,riD rI;~norcinl'J. Tn .... m dispersion; u,,,,,p_
, '0 =r -r
lr-norhc of l i ohr travel at velocmes WIthin ~ orven ~
~ ~ ~
purse that has a nrute spectral oancwictn, LlA. 1I me purse IS iauncnec III a disper-
sive material, each wavelength component of the pulse will travel at a different
velocity. The pulse effectively spreads out (or disperses) in time and space. "Aha!"
you might be thinking, "why not just make the pulse have only one wavelength?"
Nice idea, but it won't work. We will show that all finite temporal pulses must also
have a finite frequency bandwidth. Dispersion is a fundamental issue with system
UClilg .
'", I '.t.
l"UUU£ u£"pc:,."u. all""" HI wa' '''" 11IVI" Lllall Vll" 11IVUv.
T T,.,'lacc thQ ',rlQ h.,c h",,,,n c ..... "'",., I" rI",c; O',.,,,,rI (fAr th", O'r.,rlQrI i,.,rlav
'0 -r OJ ~ ,. , , o·
rlf'<;:t'rihpel in thf' npyt t'h~ntpr p~t'h ~llowpel moelp in thp ielp wil
'" . ~
rave WI a I eren arou ve OCHY. e use energy III a wavezurce wu e
distributed among the various allowed modes, either through the initial excitation,
or through mode coupling that occurs within the waveguide. The modes arrive at
the end of the waveguide slightly delayed relative to each other. This effectively
spreads the temporal duration of the pulse, which again limits the bandwidth.
Waveguide dispersion is a more subtle effect. The propagation constant f3 de-
pends on the wavelength, so even within a single mode different wavelengths will
• 1: ,t..1, ..I:~~. r' . ' ..I.
u, ""o""J ~ .. '''''~''' "P"''''~''' "'~ ,~ "'ULV"U' u,,~ '''~~u, ~'"P'''''''~'''
,." . rI. rI';e..... 'c';"n ic ncn.,' thQ o ..... ., Act ;,., ..... ., . L[""'Q' ,a~ ;n tho "',,,;nit,
-0' 't' oJ , 'J
of thp ~o-c~ll~cI zero disn .~c:,"n noint for materials ide c~n h~
. . .
e ormnant e ec III a sinai e- mo e svs em. waveaui e IS ersion can e usee to
. .. . . ....
cancel marenai utspersron, alluwmg uie Ut:SlgIl Ul spt:Clal urspersion-snu teu wave- 3.3 MATERIAL DISPERSION

A general understanding of the optical properties of dielectrics is essential in optoelectronics. The index of refraction, n, is the most widely used material parameter for In this we will the index of refraction's

We will assume that the dielectric tensor is a simple scalar, i.e., the dielectric material is isotropic. The index of refraction is defined in terms of the relative permittivity of a medium:

Consider the simple atomic model consisting of a positively charged nucleus and a surrounding negatively charged cloud. When there is no electric field present, the two charges are centered upon one another.

No Field

E -field

Figure 3.2 Pictorial depiction shows the dipole moment induced in a neutral atom by an external field.

78 ( haoter 3 DisDers~




... '1lL t.) . .)J
thM .h~
where a IS the rh~ro", ~nrt r j~ th", r"'l~t v", rtj~t~nrp f'rrrrn PflI ilihri
T
charges move.
The constitutive equation (Equation 3.2) can be written in terms of this dipole
moment.
D = EE = EoE + P (Clm2) (3.4)
where P is the bulk polarization of the material, defined by

I" Nn Nnr ''1 <;\
.
71.T • .1. . C .l: _ . _ ..
~ .. ~. w ... v ..... UJ. "J.PUH_" P"> uuu 'UJ.U>"".
r .. ~,,,,,>I rv ,,, ,~ n~nl".: nn rl .~, .1' :11 .L
't' "J '1' 1'v ........ , .. J, 'v .... "",",-.:rrv-.:JII<:O"
nens ona Clrp1 mClrt",] of th", atrrrn .nrpnt7 mClrtp]prt thp ~tClm '>0 tn", "',>r_
r
ticles bound together by a spring. The model is shown in Figure 3.3. The position
Xo represents the equilibrium position of the two charges when there are no external
forces. Xo could easily be zero.
Nucleus Spring Electron
7 '" /' <, -: <,
( ./
\ \ \I
-, ./ y -, ./ -, /


y. _v· ....
-v
Figure 3.3 The Lorentz model of the atom consists of
a heavy nucleus bound to a light electron through a
spring. Xo represents the equilibrium distance between
the charges when no external forces are present.
x represents the displacement from equilibrium.
LlL e snrinz 0 mooe e a rae ion
UMllg a SI e ween an e ec ron an a nuc eus
. - .. rrrre
uUbUL ,",COCOUl arrrrrecrunertnrrrr 11'> i:t~IUi:tHy Ui:tSCU Ull SUUIlU pllyM~i:t1 ['
h;n~;n<T ~n~r' ~1' .~ In •• ,_~~ .. rl ,I c, ..
'oJ l' 'J v .... ·6v~ .... v.v .. " .... " ... ~ .. u " LV
f hnt "hClUln in ; 0'11'-'" '1 II

V(X) '\
/
--r--
..
"( r- 051 ion

Fiaure 3.4 A rouzh model shows what the
-:~ ..
'0 t" '_t"
--nucJelIS' rrugnt lOOK liKe. 1 ne exact snape is
-, generally unknown, but it must have a
smooth minimum at some point,
designated Xo. 3.3 Material Disnersion 79



'T'L '" .. . ., •• 1 .1' "'T • .L
.111" "l""'UVIl 'ill I ""'IU" at tIl" '''''''.n'' ~L ... ~LL, A~. L'O~UL AU, LUV
• ,1 -J ',_ "1"'. 'a";,.v >v~..,"v;~" ..,v
vuu OJ ~p OJ T
dV 1 d2V 2 1 d3V 3 (3.6)
Vex) = V(xo) + -(x - xo) + --2 (X - xo) + --3 (X - xo) + ...
~ 2~ 6~
At the minimum, Xo, the first derivative, dV(xo)/~, is zero, so it can be dropped
from the expansion. To second order, the potential can be described as
1 -I2l!
T T/. T 7/. \2
\A) \""0) ~ J_ \A ~UJ
...,UA t.J.IT


v., + -« W en x"
L

where k is a constant. This is the simple Hooke's law potential for a spring, where
k is the spring constant (not to be confused with the wavevector; this is an unfor-
tunate example of identical notation used to describe completely different phenom-
ena). This expansion is only accurate for small values of x, but if the applied
electric fields are small compared to the binding potential of the electron (on the
order of lOlDV/cm) the annroximations are reasonable. (For comparison, what is
.L ., .c..' -' ·.L "L ~ rI, ,1-' .1- .L'
lIl" "."'''' lll'" l1"'U 'llli Ltl" ~pLL~"'L H~L~ ~uv L~ U, LH~ VUL LH o
. £_ .'l c . n. ,J,. 'l c: \
"UL LUV' • ~vv • 'J
Tn thp nrp~pn('p of >In p~ >I charoe a exneriences the T .orenrz force
r

~' aLE v J<.. nj rxew ons
where v is the charge's velocity. Only at relativistic speeds or under conditions of
strong DC bias, as in Hall measurements, is the magnetic term significant. We can
ignore the magnetic force in optical interactions, simplifying the driving force on
the charges to be
F= qE (3.9)

nT. :11 .- ,.,_ ~ n> .. _. :n ,,,h:,.h.., th", ~~t;~" ,...,n ha ..,Hr:hn.",rI .~
'LLL U" LH' "F ,
thp plp('tron \llith >In pffp('ti\lP rn>l~~ of


memn .n
m mp l.J.IU)
me T mn
The net forces acting on the electron determine its motion. These forces include
the external force from the applied field, acceleration, friction, and spring restoring
forces. Summing these forces together (recall q = -e)
d2x dx (3.11 )
m- + 'V11l- + kx = -eE(t)
lIF tit

, L' £, • .....
VYI: nave nLlLJUUl:I:U a uecav term, vm U.MU', WIll"'ll a"'l" a" a lll"'llVll rcr ru. LlU llUl
.11 .•• .1- .1. .. ..,1 ~..,nnar .,~_ ana_~ rI:v
, 'v UU" LV'LU ,n LU " 'oJ
t1..". ~"""rv rlna .~ _..,rI;"":~n ~_ nhnnnn "'1'l'1;~~;nn
L
Dividing both sides oTille equation by mass, m, and recalling that tor a ~-
and-mass system, the resonant frequency of oscillation is given by COo = kim,
yields 80

( haoter 3 Disoersion in Waveguides

2

I') 1 "1\

".. ...

fl.

To solve for a particular solution, x (t), we must specify the driving term. Assume that E(t) has a harmonic dependence, E(t) = Eoejwt (where it is understood that the actual electric field is given by the real part of the expression). A trial solution, x(t) = xoejwt, leads to the following solution:

y(i\

= e/m

,.,. jiat

•• L ; ....

'U

J I

were tne istuacemen IS ara e 0 e appne e ec ric e . 1'I..nowing XII we

.~ ~ r- ~. ,

ca: uu ccu y reraie Ult:: t::Ht::I,;L VI i:t1I 1It::IU LV LIlt:: JJVli:tll£i:tLJUII VI Ult::

means that there will be a phase shift between the applied field and the response of the medium. Using Equation 3.4, we can determine the electric displacement vector

D = EE = EoE + Nqx = EoE - Nex Ne21m

a.: 1<' -+ -1<'.

n lLl)

V J

( 161+

«evm \

I~

or + 1'V(d

1 ue 1. T ;y~Ut::pt::llUt::lllIVllll UI ure Ult::lt::I,;Llll,; I,;VIIM<lIILI,;<l1I ut:: t::ALl<lI,;Lt::U llUlllUlI1>

expression

E( w)

-- = E'(W) + ji'(w)

Eo

(3.15)

&.1 [ w;; w + V w

EI\I : w;; w + V w

TL .c r- .~

.I. UV UIUvA V, 1""U""Vll '"

n(w) = -v~~l_

(3.16)

(Ne2Im)(% - (2) . (Ne2Im)('Yw)

1 + - J----;;.----;;-'-~....:.......,;:-::;-

Eo[(% - (2)2 + yw2] EO[(% - (2)2 + yw2]

The index of refraction increases with the densitv of dinoles (atoms in this case)

"...... ~ ••• '11 L '.1.", ."'. ......L·

1 Y. 1 Ill;; UI;;Il"l;;l LIlCO llICOUIU, L11CO IaJ ","'1 VV III U'" LIllO' IllUIO'A VI • 1. Ill"

• L '..1. I. ..I..r 1 nnn'}\ .L:l. ~1:..I. L '..1:

,u; uu llU" "IV .11 ...... A ,Vll .. , ... v, ...... , V • .I..vvv-, "1.1.,,,, """ ... " """ ... 1.1., ... , ...... "

-0' • ••

'hp. itTl~oin~rv nart of thp. index nf ret racrron leads tn inn nr oain rlp.-

J 1 • ~

pending on the SIgn. In regions or transparency, me imagmary component or me

3,3 Material Disoersion

W" \IIi "nn""ntr~t" nn tha t .. ., ... C',..." .. a ... ' .. ani" ... i .... h;v ~h"",,'a .. v.,,,; ... ~

-.- -r 'C'T' '0

tecnmcat descnption ot attenuation tor Lllapter 7.

The index of refraction slowly increases with frequency. As the frequency approaches a resonance, Wo, the real part rapidly decreases. A resonance occurs when the frequency of the applied field is identical with a transition frequency between two energy states of the system. As the frequency rises beyond the resonance, the index again begins to increase. The region where the index of refraction is decreas-

19 as e aOPlie rec uencv increases IS ca e anomalous ats erston. e anom-

alons Ul1!ipt:rMu11 lCglU1111!i LV 11uL--auserurregior in wnicn ro worx, oecause

,J.,. 1. l.' .l. 'T'~ _1 .C .L .CC_

.. ,~~ ~.~ ... ~.~. ~ U' lv'" V~ Ulv IIlYv'" V~ lvlH'''UV11

... ' 'J r: 'E" •

1.6 f-

_j

/

I

I

I

1 A

I

I

I

I

V

I I I

I

5

6

7

8

9

10

11

12

Frequency (x1014)

Finu ... 3.5 T'he real nar+ of th" ;ml .. y "f , 00 I h" th ..

ana the damping constant 15 'Y - J nU!S7sec.~ote that the mdex n

in"rp"<p~ in mao riitnde f"y(,pnt in ,p' ";~;";t,, "I'thp

What happens it' there are many electrons, each with its individual resonant frequency? We simply add the effect of each resonance together

~ fiNe2/m

E'=Eo+L" 2 2 .

i=1 Wi - W + TYW

(3.17)

where (; is called the oscillator strenetn of each resonance. This oscillator strength

_I, "..,.,."

LUAv" lllLV Ulv LL. Ul<1L C(11.,l1 CICl;UU11 W Ull rue

'Y"'~ ".

Tn nnti"" it ill mnr" tn ila;]~ ,,,,,,,a1a ... nthC' th., ... ,,,ith frarolla ... ~,

terms, 'Y, are negligiOlY small, one can rewrITe LITe expression tor the index ot

isma the retarion « 'L'TTrl It. In Eo )~tlnf 1 ~nrl assuming th~t th" rI

82 Chaoter 3 Disoersion in Waveguides



.c, .J' ,1, ,J., TJ.,· J'nr, .11. A .J., (' ,,11,
• ~ .. ~~ .. ~u .u W. um ~. ~ ~.~uo' ... , ~ ... ... ~ ~~, .~ -'1 ,~
t:" ;~
,
n2 - A = 2: GkA2 (3.18)
k A2 - A~
where the terms A, Ak, and Gk are called the Sellmeier coefficients, and represent
resonant wavelengths and oscillator strengths, respectively, for a given system.
a es 0 oe meier coe ctents can e OUlll many rvstcs or optics anc OOKS
~ ,., ,. n " ""r . r-
lJ--UJ. n.:-. (UI , BIUII;; J. ~ 11M" llll;; ~I;;llllll;;ll;;~ "UI;;~ll"l1;;llL:-' ~Ul "I;;VI;;1<l1 ....
,J., •• ,1. TJ.,· A, ,~ ·;1 A'
.. ~u~t'~~u •• u~w •• ~m .u •••• ,~ -0' . ..
o:>nrl nr"hlprn~
-r r

Table 3.1 Sell meier coefficients for several dielectrics
NaCI (A = 1.00055) SiOz (A = 1) Alz03 (A = 1) CaFz (A = 1)
X.k(lLm) Gk X.k(lLm) Gk X.k(lLm) Gk X.t<lLm) Gk
0.05 0.198
U.I V.'W.:J70 v.vuo· V.U7Ul U.UUIJ 1.U"'JO V.VJVJ V.JVIJO
n.l?!! n "<!!I>Q

• :515 .L:JYY15 oL ',4l'/':I4 . )15j . UU4 1,4 U':I
~ rv "O~AL A OA""I A OA~ '0 ,.., A"IL C "IOAO 'lA LA" '10,'0 ,..,
'V·oJ V.VVV • v·v_ J •
I>nQ!! "< 17
Note that NaCI has resonances at several wavelengths, such as 50nm, lOOnm,
158nm, and 40.5p,m. The short wavelength resonances are due to electronic tran-
sitions within the NaCI crystal structure. The resonances at 40.5p,m and 60.98p,m
are due to ionic vibrations. These vibrations occur at lower frequencies than the
e ectr me resonances necause t e masses 0 t e atoms are arzer t an t ose 0 t e
eiecrron. we witt use rnese lU anaiyze me Ul:-'Pl;;lMUI, U1 me

lua,,, .. al».



3.3.2 Group Index and Group Delay
A useful term is the group delay, Tg• Group delay is defined as the time it takes
for a pulse of light to travel a unit distance. For example, it takes a light pulse
about 3.336 nsec to travel one meter in vacuum. Thus the group delay for vacuum
• '1 'l'lJ': .1. D, A .. JO;· . . .r • A .. l" • ,J., ;n' ,J' ,J,. ,1.
.~ ~.~~~ u~~~,. ••• ~J , 0' ~ "1:' ~J o· "1:' 'J'


To [IVo UKIUW .'::1

r" rp1 <ltp thi" tn thp innpy nf inn .,( ".) <mh>:ti tntp v .",./,. intn

Equation J.IY. 3.3 Material Dispersion

83

'R .t

uw

dn w n dio =--+-dio c c doi

(3.20)

=-----

thp. index virh rp.'mp.d to +rennencv tnverr mo T_ we O"P.t an exnressmn tor \) rhar

• • ~ <> ~. 6

nas a form simuar to me nnase veiocnv, v., - ctn

(3.21)

= clNg

where the term Ng is called the group index and is defined to be

1\; -

g

UT, :11 +:. .+..Co

'v ' ....... ~ ... v 6'~"t-' ... ~v~ '''~'V .. uv.... '"V" ••• --

dn N =n-A-

g dA

(3.23)

Keep in mind the dnld): is negative in most regions. The group index, Ng, is always larger than the regular index of refraction, n, except in regions of anomalous

rsnersion.

':II ':II ':II n ............ \I. : ..

-r _,_.,

11 a signal or pulse contains more than one wavelength, the mdividual components of this signal will travel at different group velocities. These components will reach the receiver at different times, effectively stretching out the time it takes for a signal to arrive. This effect is called group velocity dispersion (GVD).

Consider an optical pulse with a finite spectral bandwidth, aA, traveling through a dispersive medium. The time required to travel a distance L is called the

tatencv. and is me Product or me group delay, T wim me propagation distance L.

L

T - -Nn(AI (j.L41

c

The spectral WIdth ot the pulse spans trom Al to A2 (i.e., LlA - I Al A2 I).

84 Chaoter 3 Disoersion in Waveguides





~
n / '\
:J _)-&-\
:1::
0..
E
-c
..
Al A2
Wavelength (A)
Finu .... 3_6 An nntica nnlse wil h"v" a finit ..

0P'~"U '" ,a'~'~"6""



P~"h ,1. ,,.th ",;11 nrnn",."t .. "t o ~l;,.htl" ~;££"'r .. nt cn"' .. ~ Tn th ..
'b' 'J:' J:' , 0 "J:'
time domain, the pulse spread will be
L
AT = -(NiAl) - NiA2»
c
=~AN (3.25)
c g

LaN.
- --"-8..A
r ti


So what IS atv ias: We can denve an exnression or IS Tom h uat on _i [i:
b
dNg = ~( n _ A dn )
dA dA dA
dn dn d2n
=----A- (3.26)
dA dA dA2
d2n
- A
r1.A~

ining D( ua ion ".L. wi coua ion view s an ex ression
om .L or e s rea

HI UII:: llUI1;t; i:1111 Vi:11 UHlt;.

L d2n
AT= --A-AA (3.27)
c dA2
We see that pulse spreading depends on the second derivative of the material dis-
persion. The term
.\ ti2",
n---- n ?R)
r til"


i~ ~~llp.cI thp. material rlisne .. ~i~", Onp. often finds material disner sion Iisred in units
. .
or nsectt nm km t.e. e numner 0 icosecon s t e nu se w s rea as It trave s
one kilometer per nanometer Ul spectra; . Aiso, oe aware uiar sornetimes

the minus sign is carried in the material dispersion term, D, as we have done, and
sometimes it is carried in the latency expression as in Equation 3.27. ~
3.3 Material Disnersion



.::v ......... I .. '1 1 I":! .... ,_ ,,_. __ •••• 1'"\. _____ , __ • ,.
"7 , .. 'w'w ,-

The Sell meier coefficients for sapphire are listed in Table 3.1. What is the index of refrac-
tion, group velocity, and group velocity dispersion for sapphire over the range from
0.5 to 2.5M-m?
1.77 f\
e
0 <,
- 1 7t::: "
til """=
- <,
,:;. <,
1 7r:; <,
0 <,
>< <,
~ 1.74 I- ~
.£
I I I
1.0 1.5 2.0 ~.5
Wavelength (urn)
0 ,..,.. ./ <,
,.... .V<1 / <,
7 <,
x 7 <,
o I <,
~ I ......
-.. --, "
~ 1.68 I- -.
~
·u
0
Q)
>
a. 1.67 I-
::J
0 I I I I
(rj
1 n 1 r::. ..,,, ?r:;

--vIT- .... {"rn\
." . \r



0 2.5
1.5 2.0
-200 f- Wavelength (pm)
E
:r.
E -400 l-
e
o 7
!!1
r; ,..,.."
-vvv




Figure 3.7 The index, group velocity, and group velocity dispersion of
sapphire are plotted as a function of wavelength.