2007 Book IntegratedRingResonators PDF

optical sciences
D. G. Rabus
Integrated Ring
Resonators
The Compendium
฀
Springer Series in
optical sciences 127
founded by H.K.V. Lotsch
Editor-in-Chief: W. T. Rhodes, Atlanta

Editorial Board: A. Adibi, Atlanta
T. Asakura, Sapporo
T. W. Hänsch, Garching
T. Kamiya, Tokyo
F. Krausz, Garching
B. Monemar, Linköping
H. Venghaus, Berlin
H. Weber, Berlin
H. Weinfurter, München
Springer Series in
optical sciences
The Springer Series in Optical Sciences, under the leadership of Editor-in-Chief William T. Rhodes, Georgia
Institute of Technology, USA, provides an expanding selection of research monographs in all major areas of
optics: lasers and quantum optics, ultrafast phenomena, optical spectroscopy techniques, optoelectronics,
quantum information, information optics, applied laser technology, industrial applications, and other
topics of contemporary interest.
With this broad coverage of topics, the series is of use to all research scientists and engineers who need
up-to-date reference books.
The editors encourage prospective authors to correspond with them in advance of submitting a manu-
script. Submission of manuscripts should be made to the Editor-in-Chief or one of the Editors. See also
www.springeronline.com/series/624
Editor-in-Chief
William T. Rhodes
Georgia Institute of Technology
School of Electrical and Computer Engineering
Atlanta, GA 30332-0250, USA
E-mail: bill.rhodes@ece.gatech.edu
Editorial Board
Ali Adibi Bo Monemar
Georgia Institute of Technology Department of Physics
School of Electrical and Computer Engineering and Measurement Technology
Atlanta, GA 30332-0250, USA Materials Science Division
E-mail: adibi@ee.gatech.edu Linköping University
Toshimitsu Asakura 58183 Linköping, Sweden
E-mail: bom@ifm.liu.se
Hokkai-Gakuen University
Faculty of Engineering Herbert Venghaus
1-1, Minami-26, Nishi 11, Chuo-ku
Sapporo, Hokkaido 064-0926, Japan Fraunhofer Institut für Nachrichtentechnik
E-mail: asakura@eli.hokkai-s-u.ac.jp Heinrich-Hertz-Institut
Einsteinufer 37
Theodor W. Hänsch 10587 Berlin, Germany
Max-Planck-Institut für Quantenoptik E-mail: venghaus@hhi.de
Hans-Kopfermann-Straße 1
85748 Garching, Germany Horst Weber
E-mail: t.w.haensch@physik.uni-muenchen.de Technische Universität Berlin
Takeshi Kamiya Optisches Institut
Ministry of Education, Culture, Sports Straße des 17. Juni 135
Science and Technology 10623 Berlin, Germany
National Institution for Academic Degrees E-mail: weber@physik.tu-berlin.de
3-29-1 Otsuka, Bunkyo-ku
Tokyo 112-0012, Japan Harald Weinfurter
E-mail: kamiyatk@niad.ac.jp Ludwig-Maximilians-Universität München
Ferenc Krausz Sektion Physik
Ludwig-Maximilians-Universität München Schellingstraße 4/III
Lehrstuhl für Experimentelle Physik 80799 München, Germany
Am Coulombwall 1 E-mail: harald.weinfurter@physik.uni-muenchen.de
85748 Garching, Germany
and
Max-Planck-Institut für Quantenoptik
Hans-Kopfermann-Straße 1
85748 Garching, Germany
E-mail: ferenc.krausz@mpq.mpg.de
D.G. Rabus
Integrated Ring Resonators

The Compendium
With 243 Figures and 8 Tables
123
Dr. Dominik G. Rabus
Feodor-Lynen-Fellow
Baskin School of Engineering
University of California Santa Cruz
1156 High Street
Santa Cruz, CA 95064-1077, USA
E-mail: book@ringresonator.com
www.ringresonator.com
ISSN 0342-4111
ISBN-10 3-540-68786-6 Springer Berlin Heidelberg New York
ISBN-13 978-3-540-68786-3 Springer Berlin Heidelberg New York
Library of Congress Control Number: 2006940350

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specif ically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microf ilm or in any other way, and storage in data banks. Duplication of this publication or
parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its
current version, and permission for use must always be obtained from Springer-Verlag. Violations are liable
to prosecution under the German Copyright Law.
Springer is a part of Springer Science+Business Media.
springer.com
© Springer-Verlag Berlin Heidelberg 2007
The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective laws and
regulations and therefore free for general use.
Typesetting by SPi using a Springer LT
A X macro package
E
Cover concept by eStudio Calamar Steinen using a background picture from The Optics Project. Courtesy of
John T. Foley, Professor, Department of Physics and Astronomy, Mississippi State University, USA.
Cover design: WMX Design GmbH, Heidelberg
Printed on acid-free paper SPIN: 11600657 57/3100/SPi 543210
Dedication
To my dad
If you want to build a ship, don’t drum up people together to collect wood
and don’t assign them tasks and work, but rather teach them to long for the
endless immensity of the sea.
Antoine-Marie-Roger de Saint-Exupéry
Foreword
This book is all about integrated ring resonators made out of optical
waveguides in various materials for different applications. Integrated ring
resonators are used to build optical filters, lasers, sensors, modulators, dis-
persion compensators, just to name a few. The book covers the range from
simulation of various ring resonator configurations to the fabrication processes
currently used, to the description of the application of integrated ring reso-
nator in various fields. The book presents a summary of the activities in this
field form the past to the current state of the art giving several examples. The
book is written for students, graduates, professionals, academics and industry
working in this exciting field.
The examples in literature used in this book have been collected over four
years, categorizing the different examples into what are now the chapters of
this book.
Integrated ring resonators have been and still are a fascinating class of
devices and have interested me since the beginning of my PhD in 1999.
I am therefore grateful to the entire integrated ring resonator community
for providing me with exciting and cutting edge research material which ena-
bled me to write this book. Due to the enormous amount of publications in
this attractive field, I hope, that I was able to refer to most of the work done
so far and would be thankful if I was made aware of any “undiscovered” work.
As research on integrated ring resonators is ongoing, I would also be happy
to receive material for my database which can then of course be included in
further publications.
I would like to take this opportunity of expressing my sincere thanks to
Dr. Venghaus, one of the editors of the Springer Series in Optical Sciences
and Springer for making the publication of this book possible.
Furthermore, I would like to thank the Alexander von Humboldt foun-
dation and the Baskin School of Engineering of the University of California,
Santa Cruz, for supporting my research.
My special thanks go to my wife Regina and my two children, Anouk and
Tizian, for giving me the freedom and time to write this book.
List of Abbreviations
ω angular frequency
η coupling efficiency to the fundamental waveguide mode
κ coupling parameter
Γ gain of the filter
Λ length of waveguide segment
α loss coefficient of the ring (zero loss: α = 1)
αn absorption coefficient
θ phase
β propagation constant
λ wavelength
γφ complex angular propagation constant
∆λ free spectral range (see also FSR)
Φ(ω) phase shift
∆ωn complex frequency deviation
λ0 center wavelength of the passband
ε0 permittivity of free space
αcoupling fiber–chip coupling losses
λg bandgap wavelength
αinsertion insertion loss
φk coupling angle
ψk phase shift
αpropagation intrinsic losses
ωRn resonant frequency of ring resonator n
ϕt phase of the coupler
αth coefficient of thermal expansion
1/τ amplitude decay time-constant
2δλ 3 dB bandwidth
A(z) polynomial
AFM atomic force microscope
X List of Abbreviations
APF all pass filter

ASE amplified spontaneous emission
B intensity enhancement or buildup factor
B(z) polynomial
BCB benzocyclobutene
BH buried heterostructure
BHF buffered hydrofluoric acid
BPF bandpass filter
BW modulation bandwidth
c phase velocity of the ring mode (c = c0 /neff )
c0 vacuum speed of light
CAIBE chemically assisted ion beam etching
Cn auger recombination coefficient
Cp auger recombination coefficient
CROW coupled-resonator optical waveguides
CTE coefficient of thermal expansion
CVD chemical vapor deposition
cw continuous wave
DEMUX demultiplexer
DFB distributed feedback
DGD differential group delay
DI deionized
dk length of the ring waveguides
DR-RCL double ring resonator coupled lasers
DTFT discrete-time Fourier transform
DUT device under test
E complex mode amplitude
e(t) energy amplitude
E(t) time dependent mode amplitude
E-beam electron beam
ECL external cavity laser
ECR-CVD electron cyclotron resonance chemical vapor deposition
EIM effective index method
EO electro-optic
F finesse
FHD flame hydrolysis deposition
FSK frequency-shift keying
FSR free spectral range
FWHM full width at half maximum
FWM four wave mixing
GRIN gradient index
h(n) impulse response of a filter
HEMT high electron mobility transistors
I intensity/current
ICP-RIE inductively-coupled-plasma reactive ion etching
IPA Iso-propanol or 2-propanol
K contrast/confinement factor
k vacuum wavenumber
L circumference of the ring/optical output
List of Abbreviations XI
LC lattice constant
lk length of the coupler (note: this is not the coupling length)
LPCVD low-pressure chemical-vapor-deposition
m integer
MBE molecular beam epitaxy
MEMS microelectro mechanical system
mh effective hole masses
ml effective electron masses
MIBK methyl iso butyl ketone
mlp effective hole masses
MMI multimode interference
mn mode number
mo mode order
MOVPE metal organic vapor phase epitaxy
MQW multiple quantum well
mt effective electron masses
MUX multiplexer
MZI Mach–Zehnder interferometer
neff effective refractive index
ng group refractive index
NMP n-methyl-2-pyrrolidone or C5 H9 NO
NSOM near-field scanning optical microscopy
OEC optimum end-fire coupling
OLCR optical low-coherence reflectometry
OSA optical spectrum analyzer
OSP optical signal processing
P transmission power
P (t) time dependent power
p(t) total energy stored in the ring
PBS polarization beam splitter
PD photo diode
PECVD plasma enhanced chemical vapor deposition
PhC photonic crystal
PIC photonic integrated circuit
PLC polarization controller
PMD polarization mode dispersion
PMMA polymethylmethacrylate
pn poles of H(z)
PRRM pulse repetition rate multiplication
PSTM photon scanning tunneling microscope
Q quality factor
QC quantum cascade
QCSE quantum confined Stark effect
QW quantum wells
R effective radius of curvature/end-face reflectivity
r radius
RF radiofrequency
RIBE reactive ion beam etching
RIE reactive ion etching
RPM rounds per minute
XII List of Abbreviations
RTFT real-time Fourier transformation

RW ridge waveguide
RZ return-to-zero
S optical path length
SAMOVPE selective area metal organic vapor phase epitaxy
SCCM standard cubic centimeters per minute
SCISSORS side-coupled, integrated, space sequence of resonators
SEBL scanning-electron-beam lithography
SEM scanning electron microscope
SMSR side mode suppression ratio
SNOM scanning near-field optical microscopy
SOA semiconductor optical amplifier
SOG spin on glass
SOI silicon on insulator
SPM scanning probe microscopy
STM scanning tunneling microscopy
STP standard temperature and pressure
T characteristic matrix/temperature/tuning enhancement factor
t coupling parameter/time
TE transverse electric
Tg glass transition temperature
TIR total internal reflection
TM transverse magnetic
TN (x) N th-order Chebyshev polynomial
TO thermo-optic
UMatrix unit matrix
vg group velocity
w width
WDM wavelength division multiplexing
WGM whispering gallery mode
x0 3 dB cutoff frequency
xC center frequency
xn vector
xP changeover region on the passband
xS changeover region on the stopband
zm zeros of H(z)
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Ring Resonators: Theory and Modeling . . . . . . . . . . . . . . . . . . . . 3

2.1 Single Ring Resonators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.1 Ring Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.2 Racetrack-Shaped Resonators . . . . . . . . . . . . . . . . . . . . . . . 16
2.2 Double Ring Resonators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.1 Serially Coupled Double Ring Resonator . . . . . . . . . . . . . 18
2.2.2 Parallel Coupled Double Ring Resonator . . . . . . . . . . . . . 21
2.3 Multiple Coupled Resonators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3.1 Serially Coupled Ring Resonators . . . . . . . . . . . . . . . . . . . 26
2.3.2 Parallel Coupled Ring Resonators . . . . . . . . . . . . . . . . . . . 34
3 Materials, Fabrication, and Characterization Methods . . . . . 41

3.1 Wafer Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.1.1 Bonding with Intermediate Layer . . . . . . . . . . . . . . . . . . . . 42
3.1.2 Bonding Without Intermediate Layer . . . . . . . . . . . . . . . . 43
3.1.3 Benzocyclobutene Wafer Bonding . . . . . . . . . . . . . . . . . . . 44
3.2 Dry Etching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3 Si-Based Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3.1 Ring Resonators Based on Si–SiO2 . . . . . . . . . . . . . . . . . . 47
3.3.2 Ring Resonators Based on Ta2 O5 –SiO2 . . . . . . . . . . . . . . 52
3.3.3 Ring Resonators Based on SiN, SiON, and Si3 N4 . . . . . . 54
3.3.4 Ring Resonators Based on SiO2 –GeO2 . . . . . . . . . . . . . . . 57
3.4 III–V Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.4.1 The Quaternary Semiconductor
Compound GaInAsP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.4.2 The Semiconductor Compound AlGaAs . . . . . . . . . . . . . . 59
3.4.3 Lateral Coupling in GaInAsP/InP . . . . . . . . . . . . . . . . . . . 60
3.4.4 Vertical Coupling in GaInAsP/InP . . . . . . . . . . . . . . . . . . 67
XIV Contents
3.4.5 Lateral Coupling in AlGaAs/GaAs . . . . . . . . . . . . . . . . . . 71

3.4.6 Vertical Coupling in AlGaAs/GaAs . . . . . . . . . . . . . . . . . . 73
3.4.7 Implementation of Gain in Ring Resonators . . . . . . . . . . . 74
3.5 Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.5.1 Conventional Fabrication Techniques . . . . . . . . . . . . . . . . . 83
3.5.2 Replication and Nanoimprinting . . . . . . . . . . . . . . . . . . . . . 89
3.6 Temperature Insensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
3.7 Polarization Independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
3.8 Characterization Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.8.1 Conventional Characterization . . . . . . . . . . . . . . . . . . . . . . 100
3.8.2 Optical Low Coherence Reflectometry . . . . . . . . . . . . . . . . 103
3.8.3 Evanescent Field Measurement Methods . . . . . . . . . . . . . . 106
4 Building Blocks of Ring Resonator Devices . . . . . . . . . . . . . . . . 111

4.1 Couplers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.1.1 Directional Couplers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.1.2 Multimode Interference Couplers . . . . . . . . . . . . . . . . . . . . 114
4.1.3 Y-Couplers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.2 Bends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
4.3 Spot Size Converters for Light In- and Outcoupling . . . . . . . . . . 123
5 Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.1 Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.1.1 Passive Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.1.2 Devices with Gain Section . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.2 Tunability Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
5.2.1 Wavelength Tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
5.2.2 Center Wavelength Trimming . . . . . . . . . . . . . . . . . . . . . . . 159
5.2.3 Tunable Couplers in Ring Resonators . . . . . . . . . . . . . . . . 164
5.3 Dispersion Compensators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5.4 Mach–Zehnder Interferometers Combined
with Ring Resonators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
5.5 Modulators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
5.6 Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
5.6.1 All Active Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5.6.2 Devices with Gain Section . . . . . . . . . . . . . . . . . . . . . . . . . . 187
5.6.3 Passive Ring Resonator Coupled Lasers . . . . . . . . . . . . . . 191
5.7 Wavelength Converters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
5.8 Optical Signal Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
5.8.1 Logic Gates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
5.8.2 Switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
5.9 Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Contents XV
6 Whispering Gallery Mode Devices . . . . . . . . . . . . . . . . . . . . . . . . . 215

6.1 Whispering Gallery Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
6.2 WGM Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
6.3 WGM Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
7 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
1
Introduction
Integrated ring resonators have emerged in the last few years in integrated
optics and have found their way into many applications. Integrated ring res-
onators do not require facets or gratings for optical feedback and are thus
particularly suited for monolithic integration with other components. The
response from coupled ring resonators can be custom designed by the use of
different coupling configurations. In this way, the response from ring resonator
filters can be designed to have a flat top and steep roll of. Other coupling con-
figurations lead to a ring resonator reflector device, which can be used on one
side or on both sides of an integrated semiconductor optical amplifier serving
as a laser.
Ring resonators do not only find application in optical networks, they have
recently been demonstrated to be used as sensors and biosensors.
Several PhD theses were and are still devoted to the study, fabrication and
characterization of integrated ring resonators in several material systems. One
of the first thesis on integrated ring and disk resonator filters is written by
Rafizadeh (1997), where devices with diameters down to 10 µm in the material
system AlGaAs/GaAs are fabricated and characterized. One of the first the-
oretical theses is written by Hagness (1998), which presents a finite difference
time domain (FDTD) analysis of ring and disk resonators. A thesis which
followed in 2000 based, on the material system AlGaAs/GaAs is written by
Absil (2000), where vertical and lateral multiple coupled ring resonator con-
figurations are fabricated and characterized. In this work, nonlinear effects
are used in ring resonators to demonstrate four wave mixing.
These examples of fabricated ring resonators are all passive and do not con-
tain any gain to modify the filter performance. Ring lasers have of course been
realized consisting of all active material, but have not been used as optical
filters. Gain is introduced into ring resonator filters in the theses of Djordjev
(2002) and Rabus (2002) based on the InP material system. In Djordjev (2002)
vertical coupled disk and ring resonator filters with a radius down to 10 µm are
demonstrated which find application as modulators, routers, and switches. In
Rabus (2002) lateral multiple coupled racetrack shaped ring resonators with
2 1 Introduction
integrated semiconductor optical amplifiers (SOA) are presented, leading to

modified filter functions. Ring resonators in InP with one of the smallest cir-
cumference of only 20 µm are fabricated and characterized in the thesis of
Grover (2003). One of the first investigations of the nonlinear properties of
InP-based ring resonator devices is presented in the thesis of Ibrahim (2003).
These examples of ring resonator devices have all been realized in semicon-
ductor material. Ring resonators in electro optic polymers used as modulators
are fabricated in the thesis of Rabiei (2003). Polymer ring resonator modu-
lators are also demonstrated in Leinse (2005). Ring resonator research at the
University of Twente has recently led to a number of theses. Silicon oxyni-
tride and silicon nitride based devices are fabricated and described in Tan
(2004) and Geuzebroek (2005). Wafer bonding for fabricating ring resonator
devices is described in detail in the thesis of Christiaens (2005). A coupled
mode theory is used in Hiremath (2005) to analyze the filter characteristics
of ring resonators.
These theses are examples of ongoing research on ring resonator devices,
showing their potential and their diversity of application which are described
in the following chapters of this book.
The book is organized as follows: Chap. 2 covers the main aspects of
ring resonator theory. Chapter 3 gives an overview of the various fabrica-
tion methods and material systems. Chapter 4 serves as a brief introduction
into the building blocks of ring resonator devices like couplers and bends.
Chapter 5 describes the numerous applications of fabricated ring resonator
devices. Chapter 6 provides a short introduction into whispering gallery mode
resonators. The book ends with an outlook given in Chap. 7.
2
Ring Resonators: Theory and Modeling
One of the first papers dealing with the simulation of an integrated ring
resonator for a bandpass filter has been published in 1969 by Marcatili. The
layout of the channel dropping filter which he proposed is shown in Fig. 2.1.
This can be regarded as the standard configuration for an integrated ring
resonator channel dropping filter. Two straight waveguides also known as the
bus or the port waveguides are coupled either by directional couplers through
the evanescent field or by multimode interference (MMI) couplers to the ring
resonator. A simpler configuration is obtained, when the second bus or port
waveguide is removed. Then the filter is typically referred to as “notch” filter
because of the unique filter characteristic. In the following chapter, the ring
resonator simulation model is described beginning with the basic notch con-
figuration and adding more bus waveguides and ring resonators to eventually
build a multiple coupled ring resonator filter. Different types of ring resonator
simulation models will be explained, so as to be able to chose from a range of
models which best suit the need.
2.1 Single Ring Resonators

2.1.1 Ring Structure
The Basic Configuration
The basic configuration, which consists of unidirectional coupling between a

ring resonator with radius r and a waveguide, is described in Fig. 2.2, based
on Yariv (2002a, b). Defining that a single unidirectional mode of the res-
onator is excited, the coupling is lossless, single polarization is considered,
none of the waveguide segments and coupler elements couple waves of differ-
ent polarization, the various kinds of losses occurring along the propagation of
4 2 Ring Resonators: Theory and Modeling
Fig. 2.1. Ring resonator channel dropping filter
Fig. 2.2. Model of a single ring resonator with one waveguide
light in the ring resonator filter are incorporated in the attenuation constant,
the interaction can be described by the matrix relation:

Et1 t κ Ei1
= . (2.1)
Et2 −κ∗ t∗ Ei2
The complex mode amplitudes E are normalized, so that their squared magni-
tude corresponds to the modal power. The coupler parameters t and κ depend
on the specific coupling mechanism used. The ∗ denotes the conjugated com-
plex value of t and κ, respectively.
The matrix is symmetric because the networks under consideration are
reciprocal. Therefore 2 2
κ + t = 1. (2.2)
In order to further simplify the model, Ei1 is chosen to be equal to 1. Then
the round trip in the ring is given by
Ei2 = α · ejθ Et2 , (2.3)

2.1 Single Ring Resonators 5
where α is the loss coefficient of the ring (zero loss: α = 1) and θ = ωL/c,
L being the circumference of the ring which is given by L = 2πr, r being the
radius of the ring measured from the center of the ring to the center of the
waveguide, c the phase velocity of the ring mode (c = c0 /neff ) and the fixed
angular frequency ω = kc0 , c0 refers to the vacuum speed of light. The vacuum
wavenumber k is related to the wavelength λ through: k = 2π/λ. Using the
vacuum wavenumber, the effective refractive index neff can be introduced
easily into the ring coupling relations by
2π · neff
β = k · neff = , (2.4)
λ
where β is the propagation constant. This leads to
ωL kc0 L 2π · neff · 2πr r
θ= = = k · neff · 2πr = = 4π 2 neff . (2.5)
c c λ λ
From (2.1) and (2.3) we obtain
−α + t · e−jθ
Et1 = , (2.6)
−αt∗ + e−jθ
−ακ∗
Ei2 = , (2.7)
−αt∗ + e−jθ
−κ∗
Et2 = . (2.8)
1 − αt∗ ejθ
This leads to the transmission power Pt1 in the output waveguide, which is
2
2 α2 + |t| − 2α |t| cos (θ + ϕt )
Pt1 = |Et1 | = 2 , (2.9)
1 + α2 |t| − 2α |t| cos (θ + ϕt )
where t = |t| exp (jϕt ), |t| representing the coupling losses and ϕt the phase
of the coupler.
The circulating power Pi2 in the ring is given by
2
2 α2 (1 − |t| )
Pi2 = |Ei2 | = 2 . (2.10)
1 + α2 |t| − 2α |t| cos (θ + ϕt )
On resonance, (θ+ϕt ) = 2πm, where m is an integer, the following is obtained:

2
2 (α − |t|)
Pt1 = |Et1 | = 2 (2.11)
(1 − α |t|)
and
2
2 α2 (1 − |t| )
Pi2 = |Ei2 | = 2 . (2.12)
(1 − α |t|)
A special case happens when α = |t| in (2.11), when the internal losses are
equal to the coupling losses. The transmitted power becomes 0. This is known
in literature as critical coupling, which is due to destructive interference.
In using the above equations, it is possible to get a good idea of the behav-
ior of a simplified basic ring resonator filter configuration consisting of only
one waveguide and one ring. The wavelength-dependent filter characteristic
for a ring resonator configuration with a radius of r = 148 µm with matched
coupling and loss coefficient, derived using (2.1)–(2.11), is shown in Fig. 2.3.
This model can be extended to suit the requirement of various types of ring
resonator configurations.
The next configuration which is discussed is the basic ring resonator add–
drop configuration, consisting of one input, one output waveguide and the ring
resonator. The four ports of the ring resonator are referred to in the following
as input port, throughput port, drop port and add port (Fig. 2.4).
The ring resonator simulation model has been updated according to
Fig. 2.4. For simplification Ei1 is as defined before equal to 1. The throughput
mode amplitude in the first waveguide is given by
−κ1 κ∗1 t∗2 α1/2 ejθ|t1 | +|κ1 |2 =1

t1 − t∗2 α1/2
2
2 2
t1 − t∗2 αejθ ejθ
Et1 = t1 + = . =
1− t∗1 t∗2 α1/2
2 ejθ1− t∗1 t∗2 α1/2
1 − t∗1 t∗2 αejθ 2 ejθ
(2.13)
In this calculation, α1/2 and θ1/2 are used which are the half round trip loss
2
and phase, respectively. It is α = α1/2 and θ = 2θ1/2 .
Now, the mode amplitude in the ring has to pass the second coupler as
can be seen from the schematic to become the new dropped mode amplitude
Et2 . The dropped mode amplitude in the second waveguide is then given by:
−κ∗1 κ2 α1/2 ejθ1/2

Et2 = . (2.14)
1 − t∗1 t∗2 αejθ
Fig. 2.3. Notch type ring resonator filter characteristic

t1
Ei1 Et1
κ1 −κ1
t1
Er
t2
−κ2 κ2
Et2 Ei2
t2
Fig. 2.4. Model of a basic add–drop single ring resonator filter
At resonance, the output power from the drop port is given by

2 2
1 − |t1 | · 1 − |t2 | · α
2
Pt2−Resonance = |Et2−Resonance | = (2.15)
(1 − α |t1 t2 |)2
The throughput port mode amplitude Et1 (2.13) will be zero at resonance for
identical symmetrical couplers t1 = t2 if α = 1, which indicates that the wave-
length on resonance is fully extracted by the resonator. The value of α = 1
can only be achieved by the implementation of gain incorporated in the ring
resonator to compensate the waveguide losses. The value of the loss coeffi-
cient α is fixed in a purely passive ring resonator. A possibility of achieving
minimum intensity (Pt1 = 0) at resonance of the output transmission Pt1 at
the throughput port is to adjust the coupling parameters t1 , t2 to the loss
coefficient α. From (2.13) we obtain

t1
α = . (2.16)
t2
If the ring resonator is lossless (α = 1), then the couplers have to be symmetric
in order to achieve minimum intensity. The transmission of a lossless ring
resonator add drop filter with radius of r = 148 µm is shown in Fig. 2.5.
There are different kinds of requirements on the simulation of various
kinds of ring resonator configurations. Starting of with the given equations
satisfies most basic models. The ring model can for example be divided
Fig. 2.5. Add–drop ring resonator filter characteristic
into more segments to account for different materials or modified waveguide

paths. Examples of calculated models can be found in Rabus (2002) and
Michelotti et al. (2004).
Ring Resonator Parameters
Ring resonator filters can be described by certain figures of merit which are
also generally used to describe optical filters. One important figure is the dis-
tance between resonance peaks, which is called the free spectral range (FSR).
A simple approximation can be obtained for the FSR by using the propaga-
tion constant from (2.4), neglecting the wavelength dependency of the effective
refractive index
∂β β ∂neff β
=− +k ≈− . (2.17)
∂λ λ ∂λ λ
This leads to the FSR ∆λ, which is the difference between the vacuum wave-
lengths corresponding to two resonant conditions.
−1
2π ∂β λ2
FSR = ∆λ = − ≈ . (2.18)
L ∂λ neff L
Note that (2.18) is for the resonant condition next to a resonance found for
the used propagation constant.
If the wavelength dependence of the effective index can not be neglected,
it can be incorporated in the following way to obtain a modified version of
(2.17).
∂β k
= − ng , (2.19)
∂λ λ
where ng is the group refractive index, which is defined as:
∂neff
ng = neff − λ . (2.20)
∂λ
The group refractive index can be used instead of the effective index whenever
appropriate avoiding the approximation and obtaining more accurate values.
The modified FSR ∆λ is then given by
λ2
FSR = ∆λ = . (2.21)
ng L
The next parameter of importance is the resonance width which is defined

as the full width at half maximum (FWHM) or 3 dB bandwidth 2δλ of the
resonance lineshape. Using the expressions for the drop port (2.14) and (2.15)

−κ∗ κ α ejθ1/2 2 2 2
1 |κ1 | |κ2 | α
1 2 1/2
= . (2.22)
1 − t∗1 t∗2 αejθ 2 (1 − α |t1 t2 |)2
Assuming that the coupling coefficients are real, lossless, and without a phase
term, (2.22) can be written as
2 2
κ1 κ2 α1/2 1 κ1 κ2 α1/2
2 = . (2.23)
1 − 2t1 t2 α cos (θ) + (t1 t2 α) 2 (1 − t1 t2 α)2
Then
2 2
2 (1 − t1 t2 α) = 1 − 2t1 t2 α cos (θ) + (t1 t2 α) . (2.24)
For small θ, using the real part of the series expansion of the Euler formula
θ2
cos (θ) = 1 − . (2.25)
2
Therefore
2
(1 − t1 t2 α)
θ2 = . (2.26)
t1 t2 α
This equation can further be simplified if the loss in the ring is negligible and
the coupling is symmetric (t = t1 = t2 ) to

2
(1 − t2 ) 1 − t2
θ= 2
= . (2.27)
t t
Using (2.5) and (2.17) to translate into the wavelength domain
λ 2 1 − t2
2δλ = . (2.28)
πLneff t
The expression which is commonly used can be obtained by assuming weak
coupling and λ δλ
κ2 λ 2
FWHM = 2δλ = . (2.29)
πLneff
Another parameter which can now be directly calculated from the parameters
in the previous chapter is the finesse F of the ring resonator filter. It is defined
as the ratio of the FSR and the width of a resonance for a specific wavelength
(FWHM):
FSR ∆λ t κ1 π
F = = =π ≈ 2. (2.30)
FWHM 2δλ 1 − t2 κ
A parameter which is closely related to the finesse is the quality factor Q of a
resonator, which is a measure of the sharpness of the resonance. It is defined
as the ratio of the operation wavelength and the resonance width
λ neff L t neff L
Q= =π = F. (2.31)
2δλ λ 1 − t2 λ
The quality factor can also be regarded as the stored energy divided by the
power lost per optical cycle.
The intensity in the ring resonator can be much higher than that in the
bus waveguides, as the traveling wave in the ring resonator interferes construc-
tively at resonance with the input wave and thus the amplitude builds up. In
addition to this intensity increase, the field also experiences a phase-shift of
an integral multiple of 2π in one round trip. The intensity enhancement or
buildup factor B is given by (for a configuration shown in Fig. 2.2)
2
Ei2 2 −ακ∗

B=
= . (2.32)
Ei1 −αt∗ + e−jθ
For a configuration like Fig. 2.4, the buildup factor is given by

2
Er 2 −κ∗1
B= = . (2.33)
Ei1 1 − t∗1 t∗2 αejθ
On resonance, the intensity enhancement factor is

−κ∗1 2
B= . (2.34)
1 − t∗1 t∗2 α
For a lossless resonator and setting κ1 = κ2 = κ which is 1, B can be

written as
1 (2.30) F
B = 2 −−−−→ . (2.35)
κ π
This equation directly relates the intensity enhancement factor B to the
finesse F .
For the all-pass configuration and on resonance, the buildup factor is given
by (t1 = t2 = t)
1+t
B= . (2.36)
1−t
Ring resonators can be used for nonlinear optical devices since the intensity
in the resonator can be much higher than in the bus waveguide. Examples of
devices utilizing nonlinearities in ring resonators are given in Sects. 5.7 and 5.8.
The Time-Dependent Relations
Before going on to extend the previous model incorporating additional ring

resonators, the time-dependent relations are adapted from Little et al. (1997b)
for the basic ring resonator add–drop configuration. Like in the previous
model, it is assumed that the ring supports a traveling wave of amplitude
E(t), P (t) represents the total power flowing through any cross section of
the ring waveguide at time t. The ring is regarded as an oscillator of energy
amplitude e(t), normalized so that p(t) represents the total energy stored in
the ring. Energy and power in the ring are related through
2 2πr
p (t) = |e (t)| = P (t) , (2.37)
vg
where vg is the group velocity. The ring resonator has a resonant frequency
of ωR and amplitude decay time-constant of 1/τ . The decay rate and the
power exciting the ring resonator are related to each other. The decay rate
includes the power coupled to the transmitted wave 1/τtr , the power lost
due to intrinsic effects 1/τie , and power coupled to the output waveguide
1/τt2 . This leads to
1 1 1 1
= + + . (2.38)
τ τtr τie τt2
The time rate of change in ring energy can then be written as

d 1
e = jωR − e − κ∗ · Ei1 . (2.39)
dt τ
The relation between the coupling parameter κ and the decay rates of the
transmitted wave 1/τtr and the output waveguide 1/τt2 , is determined by
power conservation. The case is considered, when the ring resonator is excited
to an energy of |e0 |2 , no output waveguide is present and no input wave Ei1 .
The ring energy, with no intrinsic loss, then decays as follow:

2 2 −2t
|e (t)| = |e0 | exp . (2.40)
τtr
From these set of eqs. (2.37–2.40), the power transfer characteristic for the
drop port can be calculated with the input wave Ei1 being proportional to
exp(jωt)
2
−j
τtr
e= E . (2.41)
1 i1
j (ω − ωR ) +
τ
From this equation, we can calculate the transmitted wave at the throughput
port
1 ∗ 2
j (ω − ωR ) + + jκ
τ τtr
Et1 = Ei1 − κ∗ e = Ei1 . (2.42)
1
j (ω − ωR ) +
τ
Finally the drop port power transfer characteristic is obtained by using the
equation for power conservation (note, Ei2 = 0):
4
2 2 2 2 2 τt2 τtr 2
|Et2 | = |Ei1 | − |Et1 | = |e| = |Ei1 | (2.43)
τt2 2 1
(ω − ωR ) +
τ
This equation can be simplified further if both waveguides couple equally to
the ring, then τt2 = τtr .
The Z-Transform
Another approach to simulate ring resonator filters is by using the

Z-transformation to describe the spectral and temporal response of ring
resonator filters, described in detail in Madsen and Zhao (1999). Z-transform
relationships for basic optical elements were first developed for fiber optic
filters in Moslehi et al. (1984). This technique is used in pole-zero diagrams
to design ring resonator filters in Kaalund and Peng (2004). In the following
chapter, the Z-transform for the basic add–drop configuration (consisting
of one ring resonator and two waveguides) is given which shall serve as a
starting point for the calculation of more complex devices. Z-transforms
are discussed extensively in the aforementioned literature on digital signal
processing, therefore only a brief introduction is given here.
The Z-transform is an analytic extension of the discrete-time Fourier trans-
form (DTFT) which converts a discrete time signal into a complex-variable
frequency signal

∞
H (z) = h (n) z −n , (2.44)
n=−∞
where z is a complex variable and h(n) is the impulse response of a filter or

the values of a discrete signal. Each term z −n represents a delay. Of particular
interest is when the absolute value of |z| = 1. This is called the unit circle in
the complex plane where pole and zero locations of the function H(z), which
is evaluated along z = exp(jω), are plotted. In the case of ring resonator fil-
ters, |z| = 1 corresponds to resonant frequencies. A complete roundtrip of the
unit circle corresponds to the FSR of the filter. Poles and zeros are related
to the filter’s frequency spectrum by their position on the complex plane.
A zero positioned on the unit circle results in zero transmission at the fre-
quency corresponding to the angle of this zero. A pole on the other hand on
the unit circle causes unity transmission at the corresponding frequency. As

poles and zeros move away from the unit circle, their effect on the magnitude
spectrum is reduced.
A linear discrete system with input signals x has the following output
signal:
y (n) = b0 x (n) + b1 x (n − 1) + · · · + bM x (n − M ) − a1 y (n − 1) ,
− · · · −aN y (n − N ) . (2.45)
The Z-transform for this type of filter is then given by

M
−m
bm z
m=0 B (z)
H (z) = = . (2.46)
N A (z)
1+ an z −n
n=1
A(z) and B(z) are M th and N th-order polynomials. The zeros zm and poles
pn of H(z) can be derived from the roots of the polynomials as follows:

M
Γ z N −M (z − zm )
m=1 B (z)
H (z) = = , (2.47)

N A (z)
(z − pn )
n=1
where Γ is the gain of the filter. In passive filters, the transfer function
can never be greater than 1, so Γ has a maximum value determined by
max {|H(z)|z=exp(jω) } = 1 for these types of filters.
A ring resonator has a response which can be expressed in the form
∞

1
H (z) = an z −n = . (2.48)
n=0
1 − az −1
The basic configuration of an add–drop ring resonator filter (Fig. 2.4) is the
simplest filter with a single pole response.
The sum of all optical paths for the drop port is given by
√
Et2 (z) = −κ1 κ2 αz −1 1 + t1 t2 αz −1 + · · · Ei1 (z) . (2.49)
Using the Taylor series expansion, the equation can be simplified to give the
drop port transfer function:
√
Et2 (z) −κ1 κ2 αz −1
H21 (z) = = . (2.50)
Ei1 (z) 1 − t1 t2 αz −1
There is a single pole at t1 t2 α. As the matrix is symmetric, the transfer

function H21 (z) is equal to H12 (z).
The sum of all optical paths for the throughput port resulting in its transfer
function H11 (z) is given by

Et1 (z) = t1 − κ21 t2 αz −1 1 + t1 t2 αz −1 + · · · Ei1 (z) (2.51)
2 −1

κ1 t2 αz
= t1 − Ei1 (z)
1 − t1 t2 αz −1

t1 − t2 αz −1
= Ei1 (z) ,
1 − t1 t2 αz −1
Et1 (z) t1 − t2 αz −1
⇒ H11 (z) = = .
Ei1 (z) 1 − t1 t2 αz −1
Similarly, the transfer function H22 (z) can be derived to be
Et2 (z) t2 − t1 αz −1
H22 (z) = = . (2.52)
Ei2 (z) 1 − t1 t2 αz −1
The obtained results for each transfer function can be expressed in two differ-
ent matrix forms. The first form relates the input ports to the output ports
and is called the scattering matrix (see also (2.1)) which is given by

Et1 (z) Ei1 (z)
= SRR (z) , (2.53)
Et2 (z) Ei2 (z)
⎛ √ ⎞
t1 − t2 αz −1 −κ1 κ2 αz −1
⎜ 1 − t1 t2 αz −1 ,
1 − t1 t2 αz −1 ⎟
SRR (z) = ⎜⎝ −κ κ √αz −1
⎟.
1 2 t2 − t1 αz −1 ⎠
1 − t1 t2 αz −1 1 − t1 t2 αz −1
The second form relates the quantities in one plane to the ones in another
plane as can be seen in Fig. 2.6. This type of matrix is called the transfer
matrix which is given by

Ei1 (z) Ei2 (z)
= ΦRR (z) , (2.54)
Et1 (z) Et2 (z)

1 1 − t1 t2 αz −1 −t2 + t1 αz −1
ΦRR (z) = √ −1 .
−κ1 κ2 αz −1 t1 − t2 αz 1 + κ1 κ2 αz −1
The scattering matrix form is used to express the implication of power con-
servation and reciprocity. The transfer matrix form is used for describing
multistage filters. It is therefore also referred to as the chain matrix. This
type of transfer matrix is suitable to describe multiple serially coupled ring
resonators. The transfer matrix to be used for describing multiple parallel
coupled resonators has the form (Grover et al. 2002)

Ei1 (z) Et1 (z)
= ΦRR (z) , (2.55)
Et2 (z) Ei2 (z)
√
−1
1 1 − t1 t√
2 αz κ1 κ2 αz −1
ΦRR (z) = .
t1 − t2 αz −1 −κ1 κ2 αz −1 t1 t2 − αz −1
Fig. 2.6. Z-transform layout of an add–drop ring resonator filter
Using the different simulation techniques described in this chapter it is

possible to build a simulation model for all relevant lateral and vertically
coupled ring resonator filter configurations.
Incorporating Loss in Ring Resonator Filters
In the previous simulation models, loss is basically included by insertion of the

parameter α in the formulas. Loss in ring resonator filters scales the spectral
curves without changing them significantly if the loss is not too large. Different
types of losses have been considered and described in literature to account for
various fabrication methods and tolerances. A few examples will be given
in the following paragraph to give a starting point for the calculation and
simulation of losses in ring resonators.
A loss which is most obvious is the radiation loss which occurs in the curved
section of ring resonators (Chin and Ho 1998). The radiation loss will become
large for a large surface roughness. The surface roughness on the other hand
can induce contradirectional coupling which can degrade the performance of a
ring resonator filter and can even cause a splitting of the resonant peak (Little
et al. 1997a). The calculation of the radiation and scattering losses in ring
resonators is presented in detail in Rabiei (2005). Here a perturbative method
for the calculation of losses due to an arbitrary variation in the refractive index
of the core or due to an arbitrary variation in the shape of ring resonator is
described.
Loss in ring resonator devices can also be used advantageously as was
demonstrated using (2.16). An analysis of how loss and gain can be used to
tune, trim, or compensate the wavelength response of ring resonator filters is
described in Little and Chu (2000).
In the following chapter the basic model of a ring resonator is extended to
analyze a racetrack shaped ring resonator.
2.1.2 Racetrack-Shaped Resonators
The equations in the previous chapter described the general behavior of a

basic ring resonator filter. However to simulate certain ring resonator config-
urations, this set of equations has to be extended. In a waveguide-coupled
ring resonator filter, the coupling gap size is determined by the amount of
coupling required and the coupling length available. Using lateral coupling, a
larger gap is desirable as it increases the fabrication tolerance. In this case,
for a given coupling coefficient, the gap size can be enlarged if the coupling
distance is increased to obtain a racetrack-shaped ring resonator (Fig. 2.7).
The coupling distance can be increased by using a lateral or vertical geom-
etry (Chin and Ho 1998). In the case of the racetrack shaped ring resonator,
the coupling distance is approximately the length of the straight sections (if no
coupling occurs before the actual coupler region starts), which are tangential
to the circular arcs at output and input ports. At the transitions between the
curved and the straight sections, the mode will change adiabatically between
the radial mode in a curved waveguide and the normal mode in a straight
waveguide. As the straight sections are lengthened the radius of the curved
sections must be reduced if the total cavity length is to remain constant. This
cannot always be achieved in practical applications.
In the case of racetrack resonators, it is important to consider the variation
of phase difference ϕ which occurs in the coupling region in both couplers
between t and κ. The phase difference is length-dependent and affects the
output characteristics, not only in the magnitude but also in the resonant
conditions. This phase difference can be implemented in (2.13) and (2.14)
using the method described in Lee et al. (2004).
Fig. 2.7. A racetrack-shaped ring resonator filter with integrated platinum resistors
on top of the curved sections and the coupler (contact pads visible)
2.2 Double Ring Resonators 17
For the throughput port:

t1 ejϕt1 − t∗2 α1/2
2
ejθ t1 t∗1 ej2ϕt1 − κ1 κ∗1 e−j2ϕκ2 ejϕt2
Et1 = . (2.56)
1 − t∗1 t∗2 α1/2
2 ejθ ejϕt1 ejϕt2
For the drop port

−κ∗1 κ2 α1/2 ejθ1/2 e−jϕκ1 e−jϕκ2
Et2 = . (2.57)
1 − t∗1 t∗2 αejθ ejϕt1 ejϕt2
The resonant condition (θ + ϕt1 + ϕt2 ) = 2πm has changed slightly com-
pared to (2.9). The output transmission at resonance is no longer indepen-
dent of t and κ which can be seen comparing the relevant terms in (2.50)
and (2.51). This leads to a significant performance change and confirms the
importance of taking into account the phase difference when analyzing race-
track shaped ring resonator configurations. The phase difference is not only of
importance in directional couplers, but also in MMI couplers. An analysis of
a racetrack shaped ring resonator add–drop filter having two MMI couplers is
given in Caruso and Montrosset (2003). The MMI length should be selected in
order to satisfy the π/2 relation between the phases of the transmission and
the coupling terms to obtain a symmetric transmission characteristic of the
throughput and of the drop port. If the π/2 relation is not satisfied, a strong
asymmetry in the transmission characteristic will be obtained.
A further description of the implementation of directional and MMI cou-
plers in ring resonators is given in Sect. 4.1.
2.2 Double Ring Resonators

Double ring resonators offer the possibility of realizing a “box-like” filter char-
acteristic which is favorably used in optical networks. This is not the only
advantage, but also from the point of characterization, two rings if coupled
in series have the drop port in the same direction as the input port, which
is also convenient for interconnection of many 2 × 2 devices. Serially and
parallel coupled ring resonator configurations have been described in detail
in Chu et al. (1999a); Little et al. (1997b, 2000a); Melloni (2001) and Emelett
and Soref (2005).
In the serially coupled configuration, each ring resonator is coupled to one
another, and a signal that is to be dropped from the input port to the drop
port must pass sequentially through each resonator. Because of this sequen-
tial power transfer, all resonators must be precisely resonant at a common
wavelength. The resulting resonant line shape in the series configuration is
determined physically by the separations between the ring resonators. In the
parallel-coupled configuration, all resonators are coupled to both the input
and drop port waveguides, but usually not directly to one another (the res-
onators can also be coupled to one another resulting in a wavelength selective
reflector as is described in Sects. 2.2.2 and 2.3.1). The resonators are instead
indirectly coupled to each other by the optical path lengths along the input
and output waveguides that interconnect them. These lengths determine the
details of the resonant line shapes. An optical signal in the parallel config-
uration passes through all ring resonators simultaneously. This softens the
requirement that the resonances of each ring have to be precisely identical.
Nonaligned resonant frequencies instead lead to multiple peaks, or ripple in
the lineshape.
The ready to use transfer functions of serially and parallel coupled double
ring resonators will be described in the following sections.
2.2.1 Serially Coupled Double Ring Resonator
The schematic of a serially coupled double ring resonator is depicted in

Fig. 2.8.
From this model and using the same procedure as in Sect. 2.1.1 the fields
depicted in Fig. 2.8 can be calculated as follows:
θ1
E1a = −κ∗1 Ei1 + t∗1 α1 ej 2 E1b , (2.58)
θ θ
E1b = t∗2 α1 e j 21
E1a − κ∗2 α2 e j 21
E2b , (2.59)
θ θ
j 22 j 22
E2a = κ2 α1 e E1a + t2 α2 e E2b , (2.60)
θ2
E2b = −κ∗3 Ei2 + t∗3 α2 ej 2 E2a (2.61)
Fig. 2.8. Two ring resonators coupled in series

θ1
Et1 = t1 Ei1 + κ1 α1 ej 2 E1b , (2.62)
θ
j 22
Et2 = t3 Ei2 + κ3 α2 e E2a , (2.63)
where α1,2 = αR11/2 ,R21/2 represent the half round trip loss coefficients of
ring resonator one and two respectively. From (2.58) to (2.63) the general
expressions for the transfer functions for the throughput and the drop port
can be derived. A simplified form can be obtained by assuming a coupler
without losses and symmetric coupling behavior, setting t = t∗ and κ = −κ∗
(note that the phase factor −j has not been introduced into the assumption
and can be added if required) which gives the ready to use amplitude forms
for the throughput port (Ei2 = 0)

Et1 −t1 κ21 α1 ejθ1 t3 α2 ejθ2 − t2
= (2.64)
Ei1 1 − t3 t2 α2 ejθ2 − t2 t1 α1 ejθ1 + t3 t1 α1 α2 ejθ1 ejθ2
and for the drop port:
θ1 θ2
Et2 κ3 κ2 κ1 α1 α2 ej 2 ej 2
= . (2.65)
Ei1 1 − t3 t2 α2 ejθ2 − t2 t1 α1 ejθ1 + t3 t1 α1 α2 ejθ1 ejθ2
For realizing a double ring resonator with maximally flat response, first the
input/output waveguide ring coupling coefficient κ1 (κ3 ) has to be determined.
To simplify the model further, it is defined that the input/output waveguide-
ring coupling coefficients κ1 = κ3 . The calculation of the coupling coefficients
to obtain the appropriate coupling values between the two ring resonators in
order to achieve maximally flat response can be made according to Emelett
and Soref (2005), Little et al. (1997b) by using the geometry and index profile
shown in Fig. 2.9, where two coupled waveguides of width 2wp and 2wq with
indexes of np and nq , surrounded by a cladding of n0 , at the plane of smallest
separation 2s0 , defined as the center to center gap are shown.
The approximate coupling coefficient is then given by

ωε0 cos kxp,q wq 2 πR [αq (wq −2s0 )]
κ= np − n02
e
2 Pp Pq kx2p + αq2 αq

× αq cos kxp wp sinh (αq wp ) + kxp sin kxp wp cosh (αq wp ) . (2.66)
Fig. 2.9. Index profile and geometry of coupled waveguides

Using

βp,q 1
Pp,q = wp,q + , (2.67)
2ωµ0 αp,q

kxp,q = n2p,q k 2 − βp,q
2 , (2.68)

αp,q = βp,q2 − n2 k 2 , (2.69)
0
where Pp,q is the mode power, kxp,q is the transverse propagation constant of
the core, and αp,q is the decay constant in the cladding, βp,q is the propagation
constant, ω is the circular frequency, ε0 is the permittivity of free space, all
within waveguide p or q. The refractive index n0 of the surrounding media
is set equal to 1 (air). The coefficient R is defined as the effective radius of
curvature of the ring and is given by:
r1 r2
R= , (2.70)
r1 + r2
where r1,2 represents the radius of ring one and two respectively. The radii of
the rings are chosen to satisfy the 2π phase shift condition with the completion
of one round trip in the ring resonator which is given by:
mλm = 2πr1,2 neff . (2.71)
In order to realize a flat passband, the analysis of the power loss ratio,
which is the ratio of total input power to power present at the detected port,
is required. The power loss ratio in polynomial form is given by
4 2
Ei 2 1 µ1 µ41
PLR = = 1 + 4 4 ∆ω + 4
− 2µ2 ∆ω + µ2 −
2 2 2
,
Et2 µ1 µ2 2 4
(2.72)
where ∆ω is the frequency deviation which is related to the resonant frequency
ωm by
∆ω = ω − ωm . (2.73)
The coefficient µ is the fractional power coupled. It is given by
κ21 vg1 κ2 vg1 vg2

µ21 = and µ22 = 2 2 . (2.74)
2πr1 4π r1 r2
The goal of realizing a maximally flat response is obtained when the power
loss ratio is zero at the defined resonance. This is the case for
µ22 = 0.250µ41 . (2.75)
Assuming identical rings, this leads to the coupling coefficient
κ22 = 0.250κ41 . (2.76)

Using these equations, it is possible to design a double ring resonator filter

with maximally flat response for the drop port.
A serially coupled double ring resonator opens up the possibility of
expanding the FSR to the least common multiple of the FSR of individual
ring resonators. This is done by choosing different radii in the double ring
resonator. In the case of different radii, the light passing through the double
ring resonator is launched from the drop port when the resonant conditions of
both single ring resonators are satisfied. The FSR of the double ring resonator
with two different radii is expressed by
FSR = N · FSR1 = M · FSR2 , (2.77)
which leads to
FSR1 · FSR2
FSR = |M − N| , (2.78)
|FSR1 − FSR2 |
where N and M are natural and coprime numbers.
The use of two ring resonators with different radii opens the possibility
to realize a larger FSR than would be achieved using only a single ring res-
onator. The transmission characteristic of the throughput port has mainly a
Lorentzian shape. A box-like filter response using two different radii can only
be realized by using two parallel coupled double ring resonators (R1 = R2 ).
The use of such configurations as optical filters is limited by unwanted addi-
tional resonant peaks. Investigations on these types of filters have been per-
formed in Suzuki et al. (1995), Sorel et al. (1999). Different types of waveguide
coupled ring resonator configurations to expand the free spectral range have
been analyzed in Hidayat et al. (2003).
2.2.2 Parallel Coupled Double Ring Resonator
The schematic of a parallel coupled double ring resonator is shown in Fig. 2.10.
From this model, the fields in Fig. 2.10 can be calculated as follows:
θ1
E1a = −κ∗1 Ei1 + t∗1 α1 ej 2 E1b , (2.79)
θ1
θ2

E1b = t∗2 α1 ej 2 E1a − κ∗2 ejθW κ4 α2 ej 2 E2a − t4 Ei2 , (2.80)
θ1
θ2
E2a = −κ∗3 ejθW t1 Ei1 + κ1 α1 ej 2 E1b + t∗3 α2 ej 2 E2b , (2.81)
θ2
E2b = −κ∗4 Ei2 + t∗4 α2 ej 2 E2a , (2.82)
θ1
θ2
Et1 = t3 ejθW t1 Ei1 + κ1 α1 ej 2 E1b + κ3 α2 ej 2 E2b , (2.83)
θ2
θ1
Et2 = t2 ejθW t4 Ei2 + κ4 α2 ej 2 E2a + κ2 α1 ej 2 E1a (2.84)
θW = kW · nWeff · Λ, (2.85)
Fig. 2.10. Parallel coupled double ring resonator
where θW is the phase shift introduced by the segment of length Λ with effec-
tive refractive index nWeff of the input waveguide joining both ring resonators.
From (2.79) to (2.85) the general expressions for the transfer functions
for the throughput and the drop port can be derived. A simplified form can
again be obtained by assuming couplers and bus waveguides without losses
and symmetric coupling behavior, setting t = t∗ and κ = −κ∗ (note that
the phase factor −j has not been introduced into the assumption and can be
added if required) which gives the amplitude forms for the throughput port
(Ei2 = 0)

Et1 jθW
θ
jθW 21 f (a + bc) a + bc
= t3 t1 e + t3 κ1 α1 e h+ + κ3 t4 α22 ejθ2 . (2.86)
Ei1 1−d 1−d
And for the drop port

Et2 θ
j 21 2 jθ1 f (a + bc) θ2 a + bc
= κ1 κ2 α1 e + κ2 t1 α1 e h+ + t2 κ4 α2 ejθW 2 ,
Ei1 1−d 1−d
(2.87)
where
κ3 t1 ejθW
a= , (2.88)
1 − t3 t4 α22 ejθ2
θ1
κ1 α1 ej 2
b= , (2.89)
1 − t3 t4 α22 ejθ2
θ1
κ1 t2 α1 ej 2
c= (2.90)
1 − t1 t2 α12 ejθ1
θ2
κ2 κ4 α2 ejθW 2
f= , (2.91)
1 − t1 t2 α12 ejθ1
θ1
κ1 t2 α1 ej 2
h= . (2.92)
1 − t1 t2 α12 ejθ1
d=b·f (2.93)
Similar to the serially coupled double ring resonator, it is possible to increase

the overall FSR of this configuration by adjusting the length of the waveguide
joining the two ring resonators.
The parallel configuration can be treated as a grating. Constructive inter-
ference of the reflected waves from each ring resonator is obtained by choosing
Λ to be equal to an odd multiple of a quarter wavelength.
λ0
Λ = (2m + 1) , (2.94)
4nWeff
where λ0 is the center wavelength of the passband.
If the length Λ is chosen such that the FSR of the ring resonators and the
FSR of the “grating” obey the following condition:
FSR = NRing · FSRRing = MGrating · FSRGrating . (2.95)
The Vernier effect (Griffel 2000b) causes the transmission peaks of the ring
resonators within the overall obtained FSR to be suppressed, which results in
a larger FSR than would be achieved for a single ring resonator. The distance
between the resonators can be calculated in this case using:
Mgrating neff
Λ= πr. (2.96)
NRing nWeff
These set of equations can of course be transferred to multiple coupled parallel
ring resonator configurations as will be shown in Sect. 2.3.
Parallel Coupled Double Ring Resonator with Coupling Between

the Two Ring Resonators
A special double ring resonator configuration is obtained, when coupling

between the two ring resonators is allowed (Fig. 2.11). The reflective prop-
erties of such a device have been analyzed in Chremmos and Uzunoglu (2005)
and the following transfer functions are based on this reference. A similar
wavelength reflective filter consisting of four interacting ring resonators has
been presented in Poon, Scheuer, Yariv (2004e) which is going to be described
in Sect. 2.3.1.
Using the fields and coupling coefficient in Fig. 2.11, it is possible to derive
the general expressions for describing the transfer characteristics of this type
of configuration.
Fig. 2.11. Double ring resonator with inter ring coupling
A simplified expression for the reflectivity at port Ei1 (denoted by the

symbol ← as the superscript) is obtained for the lossless case and symmetric
coupling coefficients between the ring resonator and the bus waveguide (κ =
κ1 = κ3 ):
2 2
κ2 κ2 t2 1 + t2
← 2 4 cos θ −
2t 2t
E i1 = ⎡ 2 ⎤ . (2.97)
2
t2 1 + t2 2 2
⎣ cos θ − κ2 κ ⎦
+
2t 2t
The phase factor for the distance Λ between the rings does not appear and
therefore does not have an influence on the transfer characteristic. This is
a special property of the double coupled configuration and does not apply
for multiple serially coupled ring resonators. Different reflectivity profiles can
be realized using appropriate coupling coefficients. Weakly coupled ring res-
onators lead to single reflection peaks in the reflectivity function where the
height of the peak depend on the value of the coupling coefficients. In order to
realize a maximally flat response with a single peak, the coupling coefficients
have to obey the following equation:
κ2
κ2 = √ √ . (2.98)
2 1 + t2 + 2t
The corresponding FWHM is given by

κ2
FWHM = 4 sin−1 κ 3 . (2.99)
22 t
The expression for the reflectivity at port Ei1 incorporating loss parameter α
is given by
2.3 Multiple Coupled Resonators 25
2
κ2 κ2 α 2 + 1
← 2 2αt

E i1 = 2 2 2 2 (2.100)
t2 1 + α 2 t2 κ2 1 − α 2 t 2
cos θ − +
2αt 2αt
⎡ 2 ⎤
t2 1 + α 2 t2 t2 1 − α 1 − α2 t2
2
⎢ cos θ − + ⎥
⎢ 2αt 2αt 1 + α2 ⎥
×⎢ ⎥.
⎣ 1−α 2
2
⎦
+ sin θ
1 + α2
So far ready to use transfer functions for single and double ring resonator
configuration have been presented. In the following chapter, different calcula-
tion methods are presented to derive the transfer function of different types
of multiple coupled ring resonators.
2.3 Multiple Coupled Resonators

The use of multiple vertical or lateral coupled ring resonator configurations
opens up the possibility to realize custom designed transmission functions
and thus is suitable for a wide variety of device implementations (Schwelb
and Frigyes 2001). Vertical and lateral ring resonator architectures where the
rings are either coupled in series or in parallel have been implemented in
the past. A detailed theoretical analysis of a vertically stacked multiple cou-
pled ring resonator, where the ring resonators are on top of each other is
presented in Sumetsky (2005). The use of multiple coupled ring resonator
configurations together with other photonic devices like grating couplers or
Mach–Zehnder interferometers increases the functionality and transmission
characteristic even further (Weiershausen and Zengerle 1996). One of the main
targets of realizing optical filters is to tailor the passband shape. In analogy to
electronic filter design, Butterworth (1930) and Chebyshev type of optical fil-
ters are preferred. Butterworth filters are maximally flat and do not have any
ripples in the passband. The shape of the transfer function does not change for
higher order filters except that the roll off becomes steeper as the order of the
filter increases. Chebyshev filters have a steeper roll off than Butterworth type
of filters and have ripples either in the passband or stopband which distin-
guishes Chebyshev type I and type II filters. The square magnitude response
of a Butterworth filter has the form (Madsen and Zhao 1999)
2 1
|HN (x)| = 2N , (2.101)
x
1+
x0
where x0 is the 3 dB cutoff frequency.

The response of Chebyshev type I and II filters have the form (Madsen and
Zhao 1999)
Type I
1
|HN (x)| =
2
(2.102)
x
1 + y 2 TN2
xC
$
cos N cos−1 x for |x| ≤ 1
TN (x) = ,
cosh N cosh−1 x for |x| > 1
Type II
1
|HN (x)| =
2
⎡ ⎤, (2.103)
2 xS
T
⎢ N xP ⎥
1 + y2 ⎢ ⎥
⎣ 2 xS ⎦
TN
x
where TN (x) is the N th-order Chebyshev polynomial, y determines the pass-

band ripples, xC is the center frequency of the filter, xP and xS define the
changeover region on the passband and stopband, respectively.
2.3.1 Serially Coupled Ring Resonators
One of the first papers to present a calculation method for serially coupled
ring resonator synthesis is Orta et al. (1995). The method is based on the Z-
transformation using the transfer matrix (see (2.54)). The Z-transformation
has also been used in Madsen and Zhao (1996) to simulate and fabricate a
serially coupled ring resonator filter. The transfer matrix method has also
been used in Melloni and Martinelli (2002) and Poon et al. (2004c) for
simulating and designing various serially coupled ring resonator configura-
tions. Serially coupled ring resonators are also referred to as coupled-resonator
optical waveguides (CROW) Poon et al. (2004d). A model for deriving the
transfer functions of serially coupled ring resonators based on the time-
dependent calculation as described in Sect. 2.1.1 is presented in Little et al.
(1998b). Another method of simulating the transfer function of serially coup-
led ring resonators is by using a characteristic matrix approach presented in
Chen et al. (2004a). These different methods for analyzing serially coupled
ring resonators (Fig. 2.12) will be used in the following to derive the transfer
functions.
The Z-transformation is used to start with in the beginning. In order to
describe a serially coupled ring resonator, the filter can be broken down into
two components, a symmetrical directional coupler and a pair of uncoupled
guides as shown in Fig. 2.13 (Orta et al. 1995). Another way to describe the
transfer function of a directional coupler is to use a chain matrix, which is
given by
Fig. 2.12. Photograph of a serially coupled triple ring resonator
Fig. 2.13. Components of a serially coupled ring resonator filter

csc (φk ) −e−jβlk cot (φk )
HkC = je jβlk
, (2.104)
e−jβlk cot (φk ) −e−2jβlk csc (φk )
where φk is referred to as the coupling angle and lk is the length of the coupler
(note: this is not the coupling length). The chain matrix of the uncoupled
guides is given by
R jβdk 1 0
Hk = e , (2.105)
0 e−2jβdk
where dk is the length of the ring waveguides. The chain matrix for the entire
system is then expressed by
% csc (φk ) z −1 e−jψk cot (φk ) csc (φ0 ) cot (φ0 )
H= , (2.106)
cot (φk ) z −1 e−jψk csc (φk ) cot (φ0 ) csc (φ0 )
k=N,1
where
zejψk = −ejβ(lk +lk−1 +2dk ) . (2.107)
As previously described in Sect. 2.1.1 the elements of the matrix H are N
degree polynomials. The chain matrix can also be compared to the chain
matrix derived in (2.54). Of interest is again the transfer function for the
throughput and the drop port, both related to the signal at the input port.
The polynomials are given by
Et1
N
= H11 (z) = ak z −k (2.108)
Ei1
k=0
and
Et2
N
= H21 (z) = bk z −k . (2.109)
Ei1
k=0
The transfer functions of the filter are now based on the definition of these
polynomials. Assuming, that the filter is lossless, the scattering matrix is
unitary for |z| = 1. On |z| = 1, z ∗ = z −1 , then the relationship between
(2.108) and (2.109) can be written as follows:

∗ 1 ∗ 1
H11 (z) H11 ∗
= 1 + H21 (z) H21 . (2.110)
z z∗
Using this equation, it is possible to calculate H11 (z) for a given H21 (z).
The coupling angles φk (k = 0 . . . N ) and the phase shift ψk (k = 1 . . . N ) are
calculated as follows. First, a superscript is introduced which denotes elements
belonging to the structure formed by the first n+1 coupler. Then the elements
[N −1]
of the chain matrix relative to the first (N − 1) ring resonators, H11 and
[N −1] [N ] [N ]
H21 are related to H11 and H21 through
[N −1] [N ]
csc (φN ) z −1 e−jψN cot (φN ) H11 H11
cot (φN ) z −1 e−jψN csc (φN ) H21 [N −1] = [N ] . (2.111)
H21
This leads to the equations for the coefficients of the polynomials (2.108) and
(2.109), which are given by
[N −1] [N ] [N ]
ak = ak csc (φN ) − bk cot (φN ) (2.112)
and & '

[N −1] [N ] [N ]
bk = −ak+1 cot (φN ) + bk+1 csc (φN ) ejψN , (2.113)
where k = 1. . .(N − 1) and

[N −1] [N −1]
aN = b−1 = 0. (2.114)
Then the coupling angle of the N th coupler is given by

[N ] [N ]
b0 aN
cos (φN ) = [N ]
= [N ]
. (2.115)
a0 bN
Equation (2.115) is satisfied if the ratio is real. Next step is to determine the
phase shift ψk using (2.115) and (2.113). All other coefficients follow the same
procedure.
In order to start this calculation method and derive a transfer filter char-
[N ]
acteristic, the polynomial H21 has to be determined. As stated before in
Sect. 2.3, a bandpass filter characteristic is preferred. When using the Z-
transform, the resonant frequencies of the ring resonator filter are all placed
on the circumference |z| = 1. For a Butterworth type filter, the “zeros” are
located on z = −1. The zeros for a Chebyshev type of filter are given by

z0k = e
2j arccos sin( FWHM
4 ) cos (2k−1)π
2N
,
k = 1 . . . N. (2.116)
This is an intuitive approach to design ring resonator filters, however it is

challenging to extract the poles of the frequency response.
The time-dependent relations which have been presented for a single ring
resonator can be extended to multiple serially coupled resonators. The
response of the first ring (counted from the bottom of the ring filter) can be
written as (Little et al. 1998b)
−κ∗ Ei1
e1 = , (2.117)
−κ∗1
2
j∆ωR1 +
−κ∗N −1
2
j∆ωR2 . . . +
j∆ωRN
⎧ ⎫
⎪
⎪ − −
1
⎪
⎪
⎪
⎪ ω ω Rn j , n = 1, N ⎪
⎪
⎪
⎪ τ n ⎪
⎪
⎨ 1 1 ⎬
∆ωn = ω − ωRn − j − j , n = 1, N ,
⎪
⎪ τn τtr ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ 1 2 ⎪
⎪
⎩ ω − ωRn − j − j , N = 1 ⎭
τn τtr
where ω is the optical frequency of the input wave, ωRn is the resonant fre-
quency of ring resonator n and ∆ωn is the complex frequency deviation. The
energy decay rate of each ring is given by τn . The response of resonator N
can be expressed by a product of continued fractions:
%
N
eN = Ei1 Tn , (2.118)
n=1
where
−κ∗n−1
TN = ,
−κ∗n−1
2
j∆ωn +
−κ∗n−2
2
j∆ωn−1 +
−κ∗1
2
j∆ωn−2 . . . +
j∆ωR1
−κ∗
T1 = .
j∆ωR1
The coefficients for maximally flat and Chebyshev filter characteristics are
given in the following table for two to six serially coupled ring resonators
(Little et al. 1997b):
Table 2.1. Coefficients for maximally flat and Chebyshev filter characteristics
N maximally flat Chebyshev

2
2 κ∗1 = 0.250κ ∗4
κ∗2
1 = 0.250κ
∗2
(1 + 2y)

2
3 κ∗2 ∗2
1 = κ2 = 0.125κ
∗4
κ∗2 ∗2
1 = κ2 = 0.125κ
∗4
1 + 1.5y 3
4 κ∗2
1 = κ∗2
3 = 0.100κ
∗4
κ∗2
2 = 0.040κ ∗4
5 κ∗2
1 = κ∗2
4 = 0.0955κ
∗4
κ∗2
2 = κ∗2
3 = 0.040κ ∗4
6 κ∗2
1 = κ∗2
5 = 0.0915κ
∗4
κ∗2
2 = κ4 = 0.0245κ∗4
∗2
κ∗2
3 = 0.0179κ∗4
y determines the passband ripples of the Chebyshev filter.
Another method of calculating the transfer functions of serially coupled

ring resonator filters is by using a so called characteristic matrix approach,
which has been presented in Chen et al. (2004a). The method is based on
defining the optical paths which an optical signal travels form the input port
to the drop port. There are several paths and the shortest path is called the
common path or the zeroth-order path which is defined to be unity. Looking at
a serially coupled triple ring resonator, the light using the first order paths for
example has three choices, one additional ring resonator path and the common
path. The mth order transfer function is related to the m + 1th order transfer
function by
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
→
H1,m+1 T11 T12 T13 H1,m →
H m+1 = ⎣H2,m+1 ⎦ = ⎣T21 T22 T23 ⎦ · ⎣H2,m ⎦ = T · H m , (2.119)
H3,m+1 T31 T32 T33 H3,m
tij are the transfer functions of the additional path in ring resonator i, with
previous path ends in ring resonator j. T is called the characteristic matrix.
Using the coefficients for the ring resonators as described previously for sym-
metric coupling (t = t∗ and κ = −κ∗ ), T is given by
⎡ ⎤
jθ1 −κ22
⎢t1 t2 α1 e 1 2 0 ⎥
⎢ t 2 2 ⎥
TDropPort = ⎢⎢ −κ3 ⎥
t t α e jθ2
1 1 ⎥. (2.120)
⎣ 2 3 2
t 2 ⎦
3
t3 t4 α3 ejθ3 1 1 1
The general expression for the characteristic matrix for a filter consisting of
n ring resonators is given by
⎡ ⎤
jθ1 −κ22
t t
⎢ 1 2 1 α e 1 0 . . . 0 ⎥
⎢ t22 ⎥
⎢ ⎥
⎢t2 t3 α2 ejθ2 1 1 −κ3 0 . . . 0 ⎥
2
⎢ 2 ⎥
TDropPort = ⎢⎢
t3 ⎥. (2.121)
⎢ ... 1 ... 1 ... 0 ⎥ ⎥
⎢ ⎥
⎢ . . . 1 . . . . . . 1 −κn
2
⎥
⎣ t2n ⎦
jθn

tn tn+1 αn e 1 ... ... 1 1
The general expression for the transfer function for the drop port is then
given by

/ / j θk
κk αk e k
k k
HDropPort = ,
|UM atrix − TDropPort |
k = 1 . . . n, (2.122)
where UMatrix is the unit matrix.

Filter synthesis using (2.121) and (2.122) can be done by solving a matrix
eigenvalues equation. The poles of a filter function correspond to unity minus
the eigenvalues of TDropPort . The eigenvalues and power coupling ratios for
a maximally flat transfer function are given in Table 2.2 assuming lossless
devices.
Using the aforementioned techniques, it is possible to find the transfer
functions for arbitrary serially coupled ring resonators. Each method has its
pros and cons and therefore it is left to the designer to choose the right
method which seems convenient to solve the problem. When designing ring
resonator filters, it is appropriate to evaluate the technological feasibility of
certain parameters like coupling coefficients and loss in order to easily transfer
the design to fabrication.
A special serially coupled ring resonator design (Fig. 2.14) has been pro-
posed and analyzed in Poon, Scheuer, Yariv (2004e). The following calculation
method is based on this literature.
Table 2.2. Eigenvalues and power coupling ratios for a maximally flat transfer
function
N eigenvalues (= 1 – poles) power coupling ratios
2 0.63 + 0.32i; 0.63 − 0.32i 0.5; 0.2; 0.5
3 0.73 + 0.42; 0.71; 0.73 − 0.42i 0.5; 0.14; 0.14; 0.5
4 0.78 + 0.45i; 0.77 + 0.17i; 0.77 − 0.17i; 0.5; 0.13; 0.09; 0.13; 0.5
0.78 − 0.45i
5 0.81 + 0.46i; 0.8 + 0.26i; 0.81; 0.8 − 0.5; 0.13; 0.08; 0.08; 0.13; 0.5
0.26i; 0.81 − 0.46i
10 0.86 + 0.48i; 0.85 + 0.42i; 0.86 + 0.33i; 0.5; 0.12; 0.07; 0.07; 0.07; 0.07; 0.07;
0.88 + 0.21i; 0.9 + 0.07i; 0.9 − 0.07i; 0.07; 0.07; 0.12; 0.5
0.88 − 0.21i; 0.86 − 0.33i; 0.85 − 0.42i;
0.86 − 0.48i
(Chen et al. 2004a)
Eiin Etin
Eiin Etin
Eiin Etin
Eiin Etin
EtN E⬘t0
EtN−1 EtN Et0 E
t1
EtN−1 EiN E⬘i0

Et1
EiN Ei0
EiN−1 Ei1 Ei1

Eiin
Ei1
Ei1 Et1
Et1
Fig. 2.14. Ring resonator configuration with interring coupling
Depending on the number of rings used the filter can either be reflecting or
nonreflecting. The filter acts as a reflector for an odd number of rings (N ≥ 3),
whereas a nonreflecting filter is obtained for an even number of rings (N ≥ 4).
The transfer functions are derived by using a transfer matrix method, starting
in defining a vector xn , which represents the field component in ring resonator
n − 1: & '
← ← → → T
xn = E t E i E i E t (2.123)
n
The arrows used as superscripts are as described earlier for the double ring the
direction of the propagating field, clockwise or anticlockwise without mixing
of waves between the two directions. The coupling between the ring resonators
can be represented by the following 4 × 4 matrix (n ≥ 0):
⎡ ⎤
−t 1
⎢ κ κ 0 0 ⎥
⎢ ⎥
⎢ 1 t∗ ⎥
⎢− 0 0⎥
⎢ κ κ ⎥
xn+1 = ⎢ · x = MP · xn .
−t 1 ⎥
(2.124)
⎢ ⎥ n
⎢ 0 0 ⎥
⎢ κ κ⎥
⎣ 1 t∗ ⎦
0 0 −
κ κ
Assuming only the phase matched waves are being coupled (LCoupler λ).
The vector xn is related to xn by the following propagation matrix:
⎡ ⎤
0 0 0 e−jβrθ
⎢ 0 0 ejβr(2π−θ) 0 ⎥
xn = ⎢
⎣ 0 ejβr(2π−θ)
⎥ · x = MQ · xn , (2.125)
0 0 ⎦ n
e−jβrθ 0 0 0
π (N − 2)
θ = 2π − .
N
Combining (2.124) and (2.125)
xn+1 = MP · xn = MP · MQ · xn = MT · xn . (2.126)
For an N type ring resonator configuration with N > 2, (2.126) yields:
xN = MT N −1 · MP · x0 = MA · x0 . (2.127)
Only the components of vector xin hold the transfer functions, therefore the
transfer functions of the ring resonator configuration are derived using the
relation for the coupling to the external waveguide and the phase relations in
the first resonator:
xin = MPin · xin , (2.128)
→ → ← ←
E t0 = E tin e−jβr 2 E t0 = E tin ejβr 2 ,
θ θ
→ → ← ←
E i0 = E iN ejβr(2π−θ) E i0 = E iN e−jβr(2π−θ) ,
← ← → →
E tN = E iin e−jβr 2
θ θ
E tN = E iin ejβr 2 .
Using (2.127) and (2.128), xN and x0 can be expressed by elements of xin .
Equation (2.127) can then be rewritten as
mMPin xin = MTN −1 MP wMPin xin (2.129)

⇒ xin = MB xin ,
← → →

where m and w represent matrices which express the terms E in , E in and Ei0 ,
→

Ei0 , using components of vector xin respectively. The matrix MB can then be
expressed by
−1 −1
MB = MPin m MTN −1 MP wMPin . (2.130)
If only one input is considered as in the previous ring resonator examples, xin
can be expressed by & → '
← T
xin = 0 Ei 1 Et . (2.131)
in in
→ →
The reflection and transmission function Eiin and Etin can then be derived
using elements of matrix MB and solving the following matrix equation:
⎡ ⎤ ⎡ ⎤
MB4,2 ⎡→ ⎤ MB4,3
⎢ 1 − MB 1 ⎥ E iin ⎢ ⎥
⎢ 4,4 ⎥⎣ ⎦ = ⎢ 1 − MB4,4 ⎥ . (2.132)
⎣ MB2,4 ⎦ ← ⎣ MB2,3 ⎦
1 − E tin
1 − MB2,2 1 − MB2,2
The following equations are also valid for the solutions:
→ ←
MB3,2 E iin + MB3,3 + MB3,4 E tin = 1, (2.133)
→ ←

MB1,2 E iin + MB1,3 + MB1,4 E tin = 0.
In using the above equations, it is possible to calculate the reflectance and the
transmittance spectra of this type of serially interring coupled ring resonator.
2.3.2 Parallel Coupled Ring Resonators
Parallel coupled ring resonators have been addressed in literature and

the transfer functions have been derived by several methods. One of the
advantages of parallel coupled ring resonators over serially coupled ring
resonators is that their transfer functions are less sensitive to fabrication
tolerances, as each ring resonator can compensate errors in any of the others.
Two configurations are discussed, one where the ring resonators are coupled to
two input/output waveguides and the other, where all share one input/output
waveguide.
Coupled to Two Input/Output Waveguides
As can be seen from Fig. 2.15, parallel coupled ring resonators coupled to two
waveguides share the same input, throughput, drop, and add port.
The synthesis of the transfer functions of parallel coupled ring resonators
using a recursive algorithm is presented in Little et al. (2000a). A com-
plex matrix formalism employing racetrack ring resonator filters is derived in
Griffel (2000a). A technique using simple closed-form formulas to determine
Fig. 2.15. Photograph of a parallel coupled triple ring resonator
the Q factor of each involved ring resonator which leads to the coupling
coefficients is demonstrated in Melloni (2001). These different methods for
analyzing the transfer functions will be used in the following.
The model which has been presented in Fig. 2.10 can be easily extended for
multiple parallel coupled ring resonators. The distance between the uncoupled
ring resonators Λ is chosen so that the transfer functions of each ring resonator
add in phase. The transfer function derived by the recursive algorithm used
in Little et al. (2000a) is given by
Et Eti n Eto n
Tn = = Rn − −1 j2θW n−1 , (2.134)
Ei Rn − Tn−1 e
µin µon
Rn = −
1 2 1 2 ,
j∆ω + µi,o
n + µi,o
n
2 2
i,o 2
µn
Eti,o n = Rn −
1 i,o 2 1 i,o 2
,
j∆ω + µn + µn
2 2
where the indices i, o correspond to the through responses of each ring res-
onator n in the waveguides joining the ring resonators and ∆ω is the frequency
deviation away from resonance. The term µ is related to the coupling coeffi-
cient κ (2.74):
i,o i,o vg n
µn = κn . (2.135)
2πrn
The recursion is started with T1 = R1 . Loss can also be incorporated into this
algorithm by substituting j∆ω with j∆ω + 1/τ .
In this type of synthesis, the transfer function is directly related to the cou-
pling coefficient to obtain any desired filter shape. In the methods described
earlier, the transfer functions are rational polynomials related to frequency,
wavelength or z where the coefficients of the polynomials are adjusted for
any specific filter function. In using the recursive algorithm, it is possible to
suppress unwanted sidelobes by apodization, which is realized by adjusting

the coupling coefficients properly.
A straight forward method of designing parallel coupled ring resonator
filters is presented in Melloni (2001). The filter shape can be calculated by
using the given specifications for bandwidth, FSR, out of band rejection and
selectivity. The formulas for a parallel N coupled ring resonator filter are as
follows:
FSR
Qn = (2.136)
gn F W HM

√ 2n − 1
gn = 2 sin π .
2N
The coupling coefficients are given by

π2 4Qn 2
κn = 1+ −1 . (2.137)
2Qn 2 π2
A specific filter characteristic can be derived by using this method and the
relations described in Sect. 2.1.1.
For deriving the transfer functions for parallel coupled racetrack shaped
ring resonators, the matrix formalism described in Griffel (2000a) is used.
The parallel coupled filter can be broken down into two components, similar
to the method shown for the serially coupled ring resonator filter, into a ring
resonator coupled to two waveguides and into a pair of uncoupled guides as
shown in Fig. 2.16.
The transfer matrix for the ring resonator coupled to two waveguides is
given by
⎛ ⎞
a1n a2n − b2n bn
⎛← ⎞ 0 0 − ⎛← ⎞
⎜ a2n a2n ⎟
Ei ⎜ ⎟ Ei
⎜→ ⎟ ⎜ 1 bn ⎟
⎜E i ⎟ ⎜ 0 0 ⎟ ⎜ → ⎟
⎜← ⎟ = ⎜ a1n a1n ⎜
⎟ ⎜E i ⎟ ⎟
⎜ ⎟ ⎜ ⎟ · . (2.138)
⎟ ⎜ ⎟
←
⎝E t ⎠ ⎜ bn a1n a2n − b2n ⎝E t ⎠
⎜ 0 − 0 ⎟
→ ⎜ a1n a1n ⎟ →
Et n ⎝ bn 1 ⎠ E t n+1
0 0
a2n a2n
Fig. 2.16. Schematic of a parallel coupled ring resonator filter

The transfer matrix for the pair of uncoupled waveguides is given by

⎛← ⎞ ⎛← ⎞
⎛ jβ L ⎞
Ei e W Wn
0 0 0 E
⎜→ ⎟ ⎜ →i ⎟
⎜E i ⎟ ⎜ 0 e −jβW LW n
0 0 ⎟ ⎜ ⎟
⎜← ⎟ = ⎜ ⎟ · ⎜E i⎟
.
⎜ ⎟ ⎝ 0 0 e jβW LW n
0 ⎠ ⎜ ← ⎟
⎝E t ⎠ ⎝ E t⎠
→ 0 0 0 e−jβW LW n →
Et n E t n+1
(2.139)
The coefficients a1n , a2n and bn are given by
κ1n κ2n e−j(2βW LCn +βR πrn )

bn = , (2.140)
1 − t1n t2n e−2j(βW LCn +βR πrn )
t2n e−jβW LCn − t1n e−j(3βW LCn +2βR πrn )
a1n = ,
t1n e−jβW LCn − t2n e−j(3βW LCn +2βR πrn )
a2n = .
The arrows used as the superscripts indicate whether the field is propagating
to the left or the right direction. Multiplying the transfer matrices in an alter-
nating rhythm, the transfer functions for a parallel N coupled ring resonator
filter can be derived. Note that in the calculation the propagation constant
for the uncoupled waveguides and for the ring resonators is assumed to be the
← →
same for each one, respectively. The drop port E t1 and the throughput E iN
responses from the resulting matrix H, where h represent the elements of the
→
matrix, are then given by (assuming only one input, E i1 = 1):
← h3,2 → 1
E t1 = E iN = , (2.141)
h2,2 h2,2
→ ←
where E tN = E i1 = 0.
As in the recursive algorithm mentioned earlier in this chapter, the cou-
pling coefficients can also be varied to control the bandwidth of the filter and
thus reducing the sidelobes by apodization.
Several configurations to obtain a box-like filter response using an array
of parallel coupled ring resonators are presented in Ma et al. (2005). A special
configuration consisting of a parallel coupled double ring resonator (with two
waveguides), serially coupled with a single ring resonator has been analyzed
in Okamoto et al. (2003).
Coupled to One Input/Output Waveguide
The basic ring resonator filter configuration consisting of a ring and a wave-
guide is used to realize a multiple parallel coupled filter. Another term for
parallel coupled ring resonators with only one input/output waveguides is

side-coupled, integrated, space sequence of resonators (SCISSORS) (Heebner
et al. 2002). The transfer functions of these types of filters can be derived using
the methods and formulas described earlier in this chapter. In this chapter
the focus lies on the optical properties of this kind of device which have been
analyzed in detail in Heebner et al. (2002) and Pereira et al. (2002). The
following expressions are based on this literature.
In order to describe linear propagation effects, a pulse is considered, which
travels through a single ring resonator, obtaining a frequency-dependent phase
shift that results in a delay or distorts the pulse shape (see also section “The
Time-Dependent Relations”). The effective propagation constant for a res-
onator spacing of Λ is then given by
neff ω Φ (ω)
β0 (ω) = + , (2.142)
c0 Λ
where Φ(ω) is the phase shift which a field acquires when traversing a ring
resonator. It is derived from the transfer function of a single ring resonator
(Fig. 2.2) where the throughput field is related to the input field (2.9).
Et1 (ω) = ejΦ(ω) Ei1 (ω) , (2.143)

t sin θ (ω)
Φ (ω) = π + θ (ω) + 2 arctan .
1 − t cos θ (ω)
The transfer function of a single ring resonator can be expanded into two terms
using the Taylor series, representing the transmitted (expanded about the
normalized detuning θ0 ) and the exponential phase shift (expanded about the
transmitted phase shift of the carrier Φ0 ). The transfer function is given by
$ ∞
∞ n 0

j n
1 dm Φ
H (ω) = ejΦ = ejΦ0 1 + × (θ − θ0 )m
. (2.144)
n!
n=1
m! dθm
m=1 θ0
Using this equation, the field at some point zl+1 expressed by the field at
another point zl which is only a small distance δz away is given by (the trans-
mitted phase shift induced by each ring resonator is assumed to be distributed
over the separation Λ, leading to an effective propagation constant which is
independent of propagation distance):
& '
neff ω0 Φ0
El+1 (ω) = exp j + δz
c0 Λ
$ n 0

∞
jn neff
∞
1 δz dm Φ
× 1+ ∆ωδz + (θ − θ0 )m El (ω) .
n! c0 m! Λ dθm θ
n=1 m=1 0
(2.145)
The Fourier transform of (2.145) leads to a difference equation for the transfer
function expanded to two Tailor series (transmitted and exponential phase
shift)
Al+1 (t) =
∞
∞ m n

j n neff ∂
1 δz dm Φ j ∂
Al (t) + j δz + Al (t)
n=1
n! c0 ∂t m=1 m! Λ dω m θ0 FSR ∂t
(2.146)
For weak coupling, (2.146) can be rewritten using
Al+1 (t) − Al (t) δz→0 dA

−→ (2.147)
δz dz
leading to
∞ m
dA neff ∂
1 1 dm Φ j ∂
= − +j A. (2.148)
dz c0 ∂t m=1
m! Λ ω m θ0 FSR ∂t
This equation yields the group-velocity reduction and group-velocity disper-

sion including higher order dispersion. The group-velocity reduction, which is
proportional to the inverse of the first frequency derivative of the propagation
constant, is given by the term:

dβeff neff 1 dΦ θ =0,t≈1 neff 4r
βeff = = + 0 1+ F , (2.149)
dω c0 Λ dω −−−−−→ c0 Λ
where F is the finesse and r the radius of the ring resonator (see (2.30)).
The group-velocity dispersion, which is proportional to the second frequency
derivative of the effective propagation constant, is given by:
2
√
dβeff 1 d2 Φ θ0 =± F π√3 3 3F 2
βeff = = −−−−−−−→ ∓ 2 . (2.150)
dω 2 Λ dω 2 4π ΛFSR2
The dispersion maxima are obtained at a detuning of θ0 = ± F π√3 .
Higher-order dispersion is derived in a similar way as described before.
The expression for the third-order dispersion is given by the term:
1 d3 Φ θ0 =0,t≈1 4 F3
βeff = , −
− −−−→ − . (2.151)
Λ dω 3 π 3 ΛFSR3
All orders of dispersion have to be taken into account if the pulse bandwidth
corresponds approximately to the resonance bandwidth given by the FWHM.
Ring resonators are versatile devices and do not only exhibit linear proper-
ties, but can also be used to realize nonlinear effects. This can be accomplished
by using a material system like InGaAsP or GaAs for example. The effective
nonlinear propagation constant is given by (assuming that the nonlinearity of
the bus waveguide does not contribute significantly and can be neglected):
2
1 dΦ 1 dΦ dθ d |Ei2 | θ0 =0,t≈1 nl 8r 2
β nl eff = 2 = −−−−−→ β F . (2.152)
Λ d |Ei1 | Λ dθ d |Ei2 |2 d |Ei1 |2 πΛ
The linear propagation (2.148) can then be modified to:

∞
dA neff ∂
1 1 dm Φ
= − +j
dz c0 ∂t m=1
m! Λ ω m θ0
m
2 j ∂
× β nl 2πrB |A| + A, (2.153)
F SR ∂t
where B is the intensity enhancement or buildup factor (see (2.32)–(2.36)).

This configuration of side-coupled ring resonators exhibits strong disper-
sive and nonlinear properties and also supports solitons. As fabrication tech-
nologies mature, integration of these kinds of devices becomes feasible.
The following chapter highlights the realization and characterization of
ring resonator devices in different material systems.
3
Materials, Fabrication, and Characterization
Methods
Recently, advancing fabrication technologies enabled the realization of ring

resonators in many material systems with excellent optical properties. In this
chapter ring resonators made of different materials using corresponding man-
ufacturing processes will be presented based on the current state-of-the-art in
literature. Device performance details will be given in Chap. 5.
There are principally two configurations for the coupling between the ring
or disk resonator and the bus waveguides, namely, a vertical coupling configu-
ration, where the bus waveguides are either on top or beneath the ring or disk
resonator and the lateral coupling configuration, where the bus waveguides
and the ring or disk resonator are in the same plane.
The vertical coupling method allows the precise control of the coupling
distance and thus the coupling coefficient by using epitaxial growth. In this
configuration it is also possible to fabricate the ring and the bus waveguides
out of separate materials which opens up the possibility of creating active
ring or disk structures when using an appropriate material. Realizing ver-
tically coupled ring or disk resonators increases the fabrication complexity
because processes like wafer bonding, regrowth or re-deposition are required.
The lateral coupling method is useful for realizing ring resonators with larger
radius (>100 µm) utilizing a weak guiding waveguide layout where, however,
the confinement of the waveguide mode is still high enough to fabricate ring
resonators with negligible bending loss. Ring resonators with small radii have
also been fabricated utilizing the lateral coupling scheme as will be shown
in Chap. 5. Choosing a suitable waveguide design reduces the tight require-
ments of fabricating a very small coupling gap which is usually much less
than 1 µm in the case of very strong guiding waveguides. The lateral cou-
pling approach allows the use of multimode interference (MMI) couplers with
higher processing tolerances with respect to the power coupling coefficient
which eliminates the fabrication of small (<1 µm) and deep coupling gaps.
The incorporation of gain sections inside the ring resonator using a suitable
material can be accomplished by selective area metal organic vapor phase
epitaxy (SAMOVPE).
42 3 Materials, Fabrication, and Characterization Methods
Both coupling schemes have advantages and disadvantages and it is up

to the device designer to decide which one is more suitable. In the following
chapters, dry etching and wafer bonding techniques for realizing laterally and
vertically coupled ring resonator devices will be briefly explained.
3.1 Wafer Bonding

Wafer bonding is referred to a process in which two similar or dissimilar
semiconductor wafers are bonded with various crystallographic orientations
and lattice constants to form a single substrate which then features specific
properties. The bonding process is activated by temperature, force, chemi-
cals or plasma. An overview on wafer bonding can be found in Alexe and
Gösele (2004). Wafer bonding techniques can be divided into two groups,
one using an intermediate layer (e.g., adhesion bonding, eutectic bonding)
and another without using an intermediate layer (e.g., fusion bonding, anodic
bonding) to realize the bond. Wafer bonding allows using the advantages of
different materials, while avoiding their disadvantages, allowing combinations
of materials otherwise unachievable with other techniques. Wafer bonding is
commonly applied to form silicon-on-insulator (SOI) substrates. Bonding of
wafers of different materials has attracted great interest lately, e.g., GaAs on
Si (Georgakilas et al. 2002), or SiC on Si (Kim and Carpenter 2003).
Different methods exist for bonding various materials and it is again up
to the realized device which process is more suitable. In the following sections
common wafer bonding methods are described.
3.1.1 Bonding with Intermediate Layer
Adhesion Bonding
Adhesive bonding uses photoresist, spin on glasses or polymers to deposit a

planarizing material between two wafers. Such materials can be annealed at
low temperature to provide a low-stress wafer stack. It is a relatively simple
and inexpensive process. The use of polymers has the advantage of bonding
samples which have been structured already which is the case for vertically
coupled ring resonators, where the bus waveguides have been fabricated before
the bonding takes place.
Eutectic Bonding
Eutectic bonding is a special kind of metal bonding, where the low melting
temperature of the eutectic alloy is exploited to achieve low temperature
bonding. Eutectic is a mixture of two or more elements and has a lower
melting point than any of its constituents. Eutectic comes from the Greek
word “eutektos” which means “easily melted.” Eutectic bonds are used when a
3.1 Wafer Bonding 43
hermetic or vacuum seal is required. Eutectic alloys which are used for example
are gold–silicon, gold–tin, or lead–silicon. The metal is usually deposited by
plating, while the silicon source can be the wafer or chemical vapor deposition
(CVD). Solid–liquid mixing occurs at temperatures slightly above the eutec-
tic point and a high contact force (40 kN). A gold–silicon eutectic bond for
example is realized by coating a silicon wafer with ∼500 nm gold. The Au–Si
eutectic point is about 370◦ C. Au atoms diffuse into Si and the eutectic alloy
is formed. The bonding temperature is higher than the eutectic point which
is due to solid–liquid inter-diffusion at the interface.
3.1.2 Bonding Without Intermediate Layer
The requirements of the surface quality (cleanliness and roughness) and flat-
ness of the used wafers are very tight when it comes to direct wafer bonding.
There are basically three types of primary interactions which are responsible
for forming a first bond, Van der Waals, electrostatic Columbic, and capillary
forces.
Van der Waals forces are, however, quite small and decrease very rapidly as
the distance between two molecules increases. Covalent bonds are established,
when the wafers come into very close contact.
Electrostatic Columbic forces between two surfaces are usually stronger
than the interaction between two contacted samples if one of the surfaces has
been electrically charged by adsorbing or desorbing electrons or ions.
Capillary forces come into play when water is used in the bonding process.
Water helps to activate the bonding process at room temperature. Due to
the surface tension, the water between the two surfaces pulls the samples
together and smoothens out any surface roughness present, thus increasing the
contact area. During the following heating process, the water reacts with the
surrounding oxide or diffuses away from the surface leading to the formation
of covalent bonds. True crystal-to-crystal bonding is achieved by using hydro-
phobic bonding which requires a removal of the naturally formed oxide present
at the surface of the samples.
Fusion Bonding
Fusion bonding uses applied temperature and pressure and is used for example
to bond silicon to silicon without depositing an intermediate SiO2 layer
using the native oxide formation on the surface of the Si wafers. Fusion
bonding is commonly used in mass production of bulk micromachined
micro-electromechanical systems (MEMS) packaging. As the same mate-
rials are used in this bonding process, no thermal stress is induced, which
results in a very strong bond strength. Initial bonding is performed in a
temperature range between 300◦ C and 800◦ C, annealing is done between
800◦ C and 1,100◦ C which strengthens the bond. This high annealing temper-
ature is also a key drawback to fusion bonding. Alternatively work is being
conducted by using plasma processing to reduce the annealing temperatures

from ∼1,000◦ C to the range between 200◦ C and 300◦ C. Regardless of this
plasma pretreatment, most of the work performed on direct bonding of III/V
materials uses hydrophobic surfaces, where the oxide has been removed, to
assure a true crystal-to-crystal bond and good electrical contacting through
the bond.
Anodic Bonding
Anodic bonding or electrostatic bonding is performed by applying a high

electric field across the two substrates. Temperatures used in this process are
∼400◦ C. An example of this type of bonding is the bonding of glass/Si. Here
a voltage of ∼1.2 kV is applied. Silicon is attached to the positive, glass to
the negative electrode. The positive ions in glass are attracted to the interface
between Si/glass where a permanent chemical bond is formed. The used glass
has to be of course electrically conductive.
The Si/Si anodic bonding process, is of course also used, when an inter-
mediate glass layer (0.5–4 µm) is present. Sputtering or evaporation of for
example PyrexTM or spin on glass (SOG) is used in this bonding process.
The advantage is the use of a low voltage which is between 30 and 60 V. The
temperature is in the range of 350–450◦ C.
3.1.3 Benzocyclobutene Wafer Bonding
As described in the previous sections, several bonding techniques exist. Due

to the fact that benzocyclobutene (BCB) wafer bonding is widely used to
fabricate vertically coupled ring and disk resonators, the BCB wafer bonding
process will be briefly explained in this section. BCB wafer bonding was first
presented in Sakamoto et al. (1998) for realizing electro-optic modulators on
GaAs.
A detailed description of a BCB bonding process for realizing vertically
coupled ring resonators is presented in Christiaens et al. (2004), Christiaens
(2005) which will be used in the following.
BCB is distributed by The Dow Chemical Company under the trade name
Cyclotene∗1 . There are two varieties: the Cyclotene 3000 series resins (dry etch
type) and the Cyclotene 4000 series resins (photosensitive type). Wafer bond-
ing has so far been only demonstrated with the dry etch variety. The resins
are derived from B-staged bisbenzocyclobutene-based (BCB) monomers. Typ-
ical properties of Cyclotene are: low curing temperature (∼250◦ C), excellent
planarization properties, a dielectric constant between 2.65 and 2.5 (1 MHz–
10 GHz) and low moisture uptake (<0.2%). One of the downsides of BCB is its
1

R
TM∗
Trademark of The Dow Chemical Company (“Dow”) or an affiliated
company of Dow
3.1 Wafer Bonding 45
1.62
λ = 0.63 λ = 1.31 λ = 1.55
1.60
Index of Refraction
1.58
1.561
1.56
1.552 1.544 1.543

1.54
3022 (Hard Cure)
4024 (Hard Cure) 1.537 1.535
1.52
0.2 0.6 1.0 1.4 1.8
Wavelength (µm)
Fig. 3.1. Index of refraction as a function of wavelength for photo Cyclotene (4024–
40) and nonphoto Cyclotene (3022–46) resins after curing at 250◦ C for 60 min2
Fig. 3.2. BCB wafer bonding process
insulating nature like all common polymers and the low thermal conductivity,
which is not suitable for realizing active devices. The refractive index of two
types of resins is shown in Fig. 3.1.
A schematic of a BCB bonding process is presented in Fig. 3.2. Principally
the process flow for realizing vertically coupled devices is started by growing
the waveguiding layers on a suitable substrate (e.g., InP). An etch stop layer
2
http://www.dow.com/cyclotene/solution/refwave.htm
is grown between the substrate and the waveguiding layers which is necessary
for a later removal of the substrate after bonding. After the growth of the
layers, the bus waveguides are defined using photolithographic techniques and
etched using a suitable dry etching recipe. The sample is then spin coated
with BCB using an adhesion promoter.
The BCB sample is then placed on a hotplate for outgassing which is
followed by the bonding and a curing process in a nitrogen environment.
As transfer substrate GaAs can be used for example which is chemically inert
during the selective chemical etching if an InP substrate is being removed. The
bonding process can also be performed in a vacuum chamber. The substrate
is removed by a combination of mechanical polishing and selective chemical
wet etching to the etch stop layer. Then the etch stop layer can be removed
by using a selective chemical wet etching which does not etch the ring or disk
layers. Finally the ring or disk resonator is fabricated using dry etching and
the samples are cleaved. The samples can be antireflection (AR) coated in
order to avoid Fabry–Perot resonances in the bus waveguides.
Several BCB bonded vertically coupled ring and disk resonators will be
presented in Sect. 3.4.
3.2 Dry Etching
Dry etching is a technology widely used in semiconductor processing and

has been developed in recent years to obtain reproducible etching results.
In reactive ion etching systems, ions are accelerated towards the material to
be etched, and the etching reaction is enhanced in the direction of travel of
the ion. Different dry etching systems are available on the market with recipes
for etching various material systems. Examples of systems are conventional
reactive ion etching (RIE), inductively coupled-plasma reactive ion etching
(ICP-RIE), reactive ion beam etching (RIBE), or chemically assisted ion beam
etching (CAIBE). Gases which are commonly used (combining several) in
dry etching processes are methane (CH4 ), hydrogen (H2 ), hydrogen bromide
(HBr), nitrogen (N2 ), chlorine (Cl2 ), silicon tetrachloride (SiCl4 ), argon (Ar),
and xenon (Xe).
Despite the existence of numerous recipes, it is important to note that
same conditions and used gases in etching processes can differ from machine
to machine leading to different etching results. The reasons for this are mul-
tiple and can depend on the mixture and quality of gases and samples being
etched regularly in the dry etching equipment which determine the chemical
composition of the sidewalls in the reaction chamber. The wanted specific
chemical deposition of the sidewalls is also referred to as conditioning or sea-
soning of the chamber and is usually done before etching of devices to obtain
stable etch parameters. This conditioning affects the plasma in the etch cham-
ber physically and/or chemically resulting in changes in the current and power
density in the discharge and reaction rate.
3.3 Si-Based Materials 47
Two ways of etching are distinguished, isotropic and anisotropic etching.

RIE is an anisotropic etching technique. The etch rate which defines the rate
of removal of the semiconductor material is the same in all directions in the
case of isotropic etching. The etch rate is direction dependent in the case of
anisotropic etching, which is important in the fabrication of ring resonators,
where deeply etched waveguides are needed to obtain a strong mode confine-
ment of the field which in turn results in a small radius of curvature with
acceptable loss.
The pattern of a ring resonator layout is transferred to the material used
by choosing the right masking technique. The unmasked regions like the cou-
pling gap for example are “etched” away. There are three broad categories of
masking materials: metals (nickel (Ni), chrome (Cr), nickel chrome (NiCr),
titanium (Ti), aluminum (Al), wolfram (W), and platinum (Pt)), dielectrics
(SiOx , SiNx ) and polymers (Photoresist, E-beam resist). Metal masks are
extremely etch resistant and good choices for deep etching, however, metal
masks are difficult to remove and the metal grains result in transferred stria-
tions. Dielectric masks provide good etch resistance, and can be used to etch
nanostructures. Dielectric masks are regarded as “soft” which results in very
little transferred striations to the underlying material. Polymers are highly
etch resistant when fully cured and can be used as masks for deep etching.
However, polymers as etch resistance are not sufficient to etch the depths
(>1 µm) required for realizing waveguides used for ring resonator circuits and
there is a risk of contaminating the etch chamber.
3.3 Si-Based Materials

Silicon-based devices in photonic integrated circuits are known for quite some
time. Pure silicon has an absorption loss much smaller than 0.1 dB cm−1
at the telecom wavelength of 1.55 µm which makes this material interesting
for telecommunication devices. The processing and fabrication technologies
for this material system are mature also to the fact that the electronics indus-
try is a driving force for these enabling technologies. Another advantage in
using silicon is the possibility of combining photonic devices with electronic
circuits, enabling pre- and postprocessing of optical/electrical signals.
A summary of the properties of silicon can be found in Table 3.1.
3.3.1 Ring Resonators Based on Si–SiO2
One of the first and smallest ring resonators in the Si–SiO2 material system
which is also known as SOI has been presented in Little et al. (1998a) with
radii of 3, 4, and 5 µm. The devices have quality factors up to 250, and an FSR
of 20, 24, and 30 nm for a wavelength of 1,550 nm, respectively. The intensity
difference between the on-resonance and the off-resonance state is measured
to be more than 15 dB. The layout of a typical SOI waveguide is shown in
Table 3.1. Properties of silicon at 300 K
symbol
lattice constant 5.431 Å
density 2.329 g cm−3
dielectric constant 11.7
auger recombination coefficient Cn 1.1 · 10−30 cm6 s−1
auger recombination coefficient Cp 3 · 10−31 cm6 s−1
Debye temperature 640 K
effective electron masses ml 0.98 mo
effective electron masses mt 0.19 mo
effective hole masses mh 0.49 mo
effective hole masses mlp 0.16 mo
electron affinity 4.05 eV
optical phonon energy 0.063 eV
infrared refractive index n(λ) 3.42
3.38(1 + 3.9 · 10−5 · T)
77 K < T < 400 K
absorption coefficient αn 10−18 · no · λ2
λ ≥ 5 µm
http://www.ioffe.rssi.ru/SVA/NSM/Semicond/Si/basic.html
Fig. 3.3. Typical SOI waveguide layout
Fig. 3.3. The width of the waveguide should be less than 600 nm in order to
achieve single mode operation.
The refractive index of SiO2 plotted against the photon energy is shown
in Fig. 3.43 .
The fabrication process is made up of a deposition of a 1 µm thick buffer
layer of SiO2 , followed by a deposition of a waveguiding core layer of amor-
phous Si at 560◦ C with a thickness of 0.2 µm. The amorphous Si was annealed
3
http://www.ioffe.ru/SVA/NSM/nk/Oxides/Gif/sio2.gif
Fig. 3.4. Refractive index of SiO2 vs. photon energy
Fig. 3.5. SEM of a filter incorporating a 3 µm radius ring, with 0.5 µm wide by
0.2 µm-thick waveguides.
Reprinted with permission from IEEE Photonics Technology Letters, vol. 10,
pp. 549–551 1998
c IEEE
at 900◦ C for 2 h to convert it to polycrystalline silicon (polySi) out of which

the ring and bus waveguides (width: 0.5 µm, height: 0.2 µm) were patterned
using an HBr–Cl plasma etch process. In order to increase the fiber–chip cou-
pling efficiency, the bus waveguides are flared to an 8 µm width. The achieved
index contrast is either 3.5:1.5 or 3.5:1.0, depending on whether the material
external to the Si waveguide is SiO2 or air, respectively. An SEM photograph
of a fabricated device is shown in Fig. 3.5.
Racetrack shaped ring resonators based on a similar fabrication process
have been presented in Vörckel et al. (2003) where the radius is 3 µm and
the length of the straight coupling region is 1 µm, obtaining an FSR of about
24 nm and a Q-factor of 1,800. The device exhibits asymmetrical coupling

to achieve minimum intensity at the throughput port on resonance (see
Sect. 2.1.1 (2.16)).
The thickness of the silicon layer and the buried oxide layer of the SOI
substrate are 0.4 µm, respectively. The racetrack shaped ring resonators are
defined by electron beam lithography. The waveguides are etched down to
the buried oxide layer by an HBr inductive coupled plasma reactive ion etch-
ing process. The width of the waveguides is 0.4 µm. The bus waveguides are
tapered to a 10 µm width to increase the fiber–chip coupling efficiency. Two
types of coupling distance variations have been used. In the first one, the
fixed coupling distance is 0.2 µm and the other coupling separation is varied
between 0.1 and 0.2 µm. The second configuration used a fixed coupling dis-
tance of 0.25 µm and varying the other coupling separation between 0.125 and
0.25 µm. The throughput port minimum on resonance has been increased from
−7.2 to −12.8 dB. Due to the unreproducibility of the waveguide loss and the
coupling coefficients, an exact adjustment of the critical coupling cannot be
possible.
Ring resonators with similar dimensions and figures of merit have been
realized in Dumon et al. (2004) by deep UV lithography with an illumination
wavelength of 248 nm. An SEM photograph of a fabricated racetrack shaped
ring resonator is shown in Fig. 3.6.
Racetrack shaped ring resonators based on the design in Fig. 3.3 but having
an SiO2 cladding with radii down to 20 µm are presented in Kiyat et al. (2005).
A vertically coupled disk resonator fabrication process based on a separa-
tion by implantation of oxygen (SIMOX) has been presented in Koonath et al.
Fig. 3.6. Racetrack resonator in SOI. The isolated wire width is 500 nm, the gap
width is 230 nm. Note that when the wires get closer together, the wire width decr-
eases to 450 nm due to optical proximity effects during lithography.
pp. 1328–1330 2004
c IEEE
(2003); Koonath et al. (2004, 2005). The SIMOX process involves the implan-
tation of oxygen ions into a silicon substrate, followed by a high temperature
(∼1,300◦ C) annealing of the substrate in order to cure the implantation dam-
age and to effect SiO2 formation. The thickness and the depth of the buried
oxide layer are determined by the implantation dose and energy. The implan-
tation dose should be in the range of 1 × 1017 –9 × 1017 ions per cm2 , with
implantation energies in the range of 40–200 keV to keep the defect densities
below 105 cm−2 . The implantation of oxygen ions is performed on an SOI sub-
strate which is patterned with thermally grown oxide. The thickness of the
oxide mask is chosen such that the oxygen ions that penetrate into the area
underneath the mask are decelerated. After the high temperature annealing,
angled side-walls of the buried rib waveguide are obtained due to the lat-
eral spread out of the implanted oxygen ions. After annealing, rib waveguides
can be defined on the top layer using a conventional lithography and etching
process.
Disks with radii of 20, 20.5, and 21 µm have been fabricated. The free
spectral rang of these devices is ∼5 nm. In order to fabricate the devices,
an SOI wafer with 0.6 µm of silicon on top of a buried oxide layer of 0.4 µm
thickness was oxidized and patterned using a reactive ion etching process
to form oxide stripes of thickness 0.06 µm, with widths varying from 2 to
12 µm. The patterned wafer was then implanted with oxygen ions with a dose
of 5 × 1017 ions cm−2 , at energy of 150 keV. The implanted wafers were then
annealed at 1, 320◦ C for 7.5 h in an ambient of argon, with 1% oxygen, to cure
the implantation damage. To form the disks, a silicon nitride layer of thickness
0.1 µm was deposited on top, and patterned using standard lithography and
reactive ion etching. The substrate was then oxidized to remove the silicon
on the top layer, everywhere except underneath the disks, realizing vertical
coupling to the buried bus waveguides. An SEM photograph of a vertically
coupled disk is shown in Fig. 3.7.
Fig. 3.7. SEM picture of the fabricated microdisk resonators on the top silicon layer
with bus waveguides underneath.
Reused with permission from P. Koonath, Applied Physics Letters, 86, 091102 (2005)
2005
c American Institute of Physics
3.3.2 Ring Resonators Based on Ta2 O5 –SiO2
The Ta2 O5 –SiO2 material system has also been used very successfully to rea-
lize ring resonator devices. The advantage of using this material system is the
ability to control the refractive index by adjusting the ratio of Ta2 O5 to SiO2 .
Vertically coupled ring resonators with a radius of 10 µm have been presented
in Little et al. (1999). Here alternating layers of SiO2 and Ta2 O5 –SiO2 are
deposited by radiofrequency (RF) sputtering. The waveguides for the ring and
the bus are made using a composition of 30 mol.% of Ta2 O5 and 70 mol.% of
SiO2 leading to a refractive index of 1.7825 at a wavelength of 1,550 nm.
The bottom cladding layer is made up of 10 µm SiO2 . Then a 0.5 µm layer
of Ta2 O5 –SiO2 material is deposited on top. In order to fabricate the bus
waveguides, a 150-nm-thick Cr mask was evaporated on to the sample and
patterned by photolithography and dry etched to a depth of 0.5 µm. While
the Cr on top of the bus waveguides remained intact, 0.5 µm of SiO2 was
sputtered onto the sample to fill in the amount of bus material etched away.
In the Cr liftoff process, the sample was first soaked in buffered hydrofluoric
acid (BHF) for 10 s to remove the SiO2 deposited on the vertical walls of the
Cr. This resulted in a liftoff, of both the Cr and the SiO2 on top of it. For the
vertical separation of the bus and ring waveguides, a layer of SiO2 is sputtered
onto the sample. The thickness of this layer defines the coupling coefficient
between the ring and the bus waveguide. The ring resonator which is on top
of the bus waveguide is made by deposition of a Ta2 O5 layer with following
Cr mask patterning process. The dimensions and the layer sequence of the
device are sketched in Fig. 3.8.
A similar process realizing an eight channel add–drop filter is described
in Chu et al. (1999c). Here the ring and bus waveguide cores are composed
of Ta2 O5 –SiO2 [17:83] mol% which gives a refractive index of 1.6532. SiO2
is used for the cladding and the buffer layers. The bus waveguides has been
slightly offset from the ring resonator waveguide. The dimensions and the
waveguide layout are shown in Fig. 3.9.
The same waveguide and ring resonator layout has been used in Chu et al.
(1999b) to realize a double ring resonator configuration. Here the composition
Fig. 3.8. Dimensions and layer sequence of a vertically coupled Ta2 O5 –SiO2 ring
resonator
Fig. 3.9. Waveguide parameters of the eight channel add–drop filter
of Ta2 O5 –SiO2 has been chosen to give a refractive index of 1.539. The ring
and bus waveguide width are both 2 µm. The SiO2 pedestal height for the ring
waveguide is 0.5 µm. The bus waveguide are buried 0.5 µm below the surface
with an offset of 0.8 µm.
A vertical coupled ring resonator configuration using the Ta2 O5 –SiO2 com-
pound has been presented in Hatakeyama et al. (2004) and Kokubun et al.
(2005). The ring resonators are coupled to the bus waveguides using multilevel
crossings which means that there is a lower and an upper bus waveguide. Sin-
gle, double and quadruple coupled ring resonators have been demonstrated.
The advantage of using a multilevel structure is to eliminate the scattering
loss at bus waveguide crossings. The fabrication process is similar to the pre-
viously described methods, involving an RF sputtering and photolithography
technique. The patterns of the waveguides are transferred using a Cr mask
and a reactive ion etching (RIE) process with C2 F6 gas. The waveguide core
layers are made of Ta2 O5 –SiO2 compound glass (Ta2 O5 30 mol.%, n = 1.785
at the wavelength of 1,550 nm), and the cladding and separation layers are
SiO2 (n = 1.451 at the wavelength of 1,550 nm).
The fabrication of the vertically coupled ring resonator with multilevel
bus waveguide crossing is started by patterning and RIE etching of the lower
bus waveguide. Then the bus waveguide is covered by depositing SiO2 . The
SiO2 layer which has been deposited onto the lower bus waveguide is elim-
inated by a liftoff process through chemical etching of the Cr mask using a
solution of Cerium (IV) diammondium nitrate 13.3 wt%, perchloric acid 70%
and water. In order to realize a perfect planarized surface for the coupling to
the ring resonator, SOG is used. SOG is spin coated and baked at 400◦ C for
40 min under N2 gas, realizing a height of only 0.2 µm. SOG is made up of
ethyl–silicate-polymer 9% (Si 5.9%), ethanol 73%, and methyl acetate 18%.
The width and thickness of a ring waveguide is 1.2 and 0.7 µm, respectively.
The thickness of the separation layer between the ring resonator and bus
waveguides is 0.6 µm. The ring radius is 5 µm, obtaining a record 37 nm FSR.
Other configurations with different dimensions have also been realized.
3.3.3 Ring Resonators Based on SiN, SiON, and Si3 N4
Serially coupled triple ring resonators in SiN have been demonstrated in

Barwicz et al. (2004). The ring and bus waveguides are made of a single layer
of SiN. The high index contrast which is achieved by using this material
combination induces strong polarization dependence. Therefore the device is
optimized for transversal electric (TE) polarization operation. A photograph
of a triple ring resonator is shown in Fig. 3.10.
The fabrication of the device is started by thermally oxidizing a silicon
wafer to obtain a bottom SiO2 layer. Then, a SiN layer is deposited by
low-pressure chemical vapor deposition (LPCVD) in a vertical thermal reactor
using a gas mixture of SiH2 Cl2 and NH3 . Next, 200 nm of polymethyl-
methacrylate (PMMA) and 40 nm of Aquasave were spun on. PMMA is a
positive E-beam resist while Aquasave is a watersoluble conductive polymer
from Mitsubishi Rayon used to prevent charging during direct-write scanning-
electron-beam lithography (SEBL). PMMA is exposed at 30 keV using a Raith
150 SEBL system. The Aquasave is removed, and the PMMA is developed.
Next, 50 nm of Ni is evaporated on the structure, and a following liftoff is per-
formed by removing the nonexposed PMMA. The waveguides are defined by
conventional reactive ion etching (RIE) using Ni as a hardmask. A gas mixture
of CHF3 –O2 in a 16:3 flow-ratio was used during the dry etching process to
obtain vertical and smooth sidewalls. The accuracy of the etch depth is mon-
itored using several etch steps and is measured with a profilometer. Finally,
the Ni is removed using a nitric-acid-based commercial wet Ni etchant.
The design and the dimensions of the used waveguide are shown in
Fig. 3.11. The device is designed to have an FSR of 24 nm.
Fig. 3.10. Third-order add–drop filter based on series-coupled microring resonators.

The rings’ outer radius is 7.3 µm. The ring-to-bus gap is 60 nm and the ring-to-ring
gap is 268 nm.
Reprinted with permission from T. Barwicz, Optics Express, vol. 12, no. 7, pp. 1437–
1442 2004
c Optical Society of America
Fig. 3.11. Waveguide dimensions. The refractive index is given for λ = 1,550 nm
Fig. 3.12. Cross-section and layer sequence of SiON waveguide
A similar design is presented in Wang et al. (2005) with a radius of 25 µm

resulting in an FSR of about 8 nm.
Ring resonators based on SiON are analyzed in Melloni et al. (2003a). The
ring radius is 300 µm. The SiON waveguides are made by deposition using
standard plasma-enhanced chemical vapor deposition (PECVD) techniques
and dry etching. The cross-section and the dimensions of the used waveguides
are sketched in Fig. 3.12.
Lateral and vertically coupled ring and disk resonators in Si3 N4 –SiO2 and
SiON are presented in Klunder et al. (2001) and Klunder et al. (2003b),
respectively. The refractive index of Si3 N4 plotted against the photon energy
is shown in Fig. 3.134 .
The dimensions and the layout of the lateral coupled configuration are
shown in Fig. 3.14. Laterally coupled devices with a radius of 25 µm with and
without a cladding of PMMA have been fabricated. The radius of the realized
disk resonator is 15 µm.
The fabrication of the device is performed using standard lithography and
reactive ion etching. Buffered hydrofluoric acid (BHF) etching is done after
RIE to reduce the surface roughness and to obtain smoother sidewalls. The
layer sequence of a vertically coupled ring resonator is shown in Fig. 3.15.
Here an additional SiO2 layer is grown by PECVD which defines the
coupling strength between the ring and the bus waveguides.
A high-index contrast silica on silicon technology with ∆neff = 0.02 was
developed at the IBM Zurich Research Lab5 based on PECVD deposited
4
http://www.ioffe.ru/SVA/NSM/nk/Nitrides/Gif/si3n4.gif
5
www.ibm.zurich.com
Fig. 3.13. Refractive index of Si3 N4 vs. photon energy
Fig. 3.14. Cross-section of a laterally coupled ring resonator
Fig. 3.15. Cross-section of a vertically coupled ring resonator
silicon oxynitride (SiON) films for realizing integrated ring resonators (Horst
et al. 1998). The minimum bending radius is 1.5 mm. The waveguide layout
is shown in Fig. 3.16.
3.4 III–V Materials 57
Fig. 3.16. Waveguide layout developed at IBM Zurich Research Lab
3.3.4 Ring Resonators Based on SiO2 –GeO2

Ring resonators using GeO2 doped silica waveguides and vertical coupling
have been demonstrated in Suzuki et al. (1992). The GeO2 doped silica
waveguides are fabricated by a combination of flame hydrolysis deposition
(FHD), electron cyclotron resonance CVD (ECR-CVD) with an applied RF
bias (bias-CVD), and RIE. The waveguides for the ring resonator are realized
in a first step unlike the configurations previously discussed, where the ring
resonator is on top of the bus waveguides. The ring waveguides are made of
GeO2 doped silica with a 2% refractive index difference. The dimension of
the ring waveguide etched into an under cladding layer of thickness 30 µm
is 3.5 µm × 4 µm. The coupling distance is adjusted to a height of 2.5 µm
between the ring and the bus waveguides by deposition. The dimension of
the bus waveguides with a 0.75% refractive index difference is 5 µm × 6.5 µm.
Finally the bus waveguides are covered by a cladding layer deposited again
using FHD. The radius of the ring is 1.5 mm.
The advantage of using this kind of material is the low propagation
loss which is smaller than 0.1 dB cm−1 at a wavelength of 1,550 nm. Several
integrated devices based on this material are presented in Suzuki et al. (1994)
using different refractive index differences for the employed waveguides. The
idea which has been followed is the use of two types of waveguides, weak and
strong guiding. Weak guiding waveguides are used for in and out coupling to
reduce the fiber to chip coupling loss.
3.4 III–V Materials

In this chapter the most commonly used III–V material systems GaInAsP on
InP and AlGaAs on GaAs are treated. Ring resonators both vertically and
laterally coupled have been demonstrated in these material systems.
3.4.1 The Quaternary Semiconductor Compound GaInAsP

III/V semiconductors on the basis of InP with a direct bandgap are used
for a variety of components in the all-optical network. The composition of
the quaternary (III–V) semiconductor compound GaInAsP lattice matched

to InP, can be changed, so that the bandgap can be adjusted in the range
between 0.97 and 1.65 µm. The choice of the appropriate bandgap, which is
smaller than the signal wavelength, enables the fabrication of passive, trans-
parent waveguides with low loss (<1 dB cm−1 ). This material system is also
used for the realization of lasers and optical amplifiers in the spectral window
around 1.55 and 1.3 µm. A summary on optoelectronic and photonic inte-
grated circuits on InP can be found in Kaiser and Heidrich (2002).
Electronic components can also be fabricated with this material system,
by incorporating Si, Be, Zn, and realizing p- or n-doped areas. The proper-
ties of the semiconductor compound (Gax In1−x Asy P1−y ) can be described by
Vegard’s law. Using this law and the known semiconductor compounds GaAs,
GaP, InAs, and InP it is possible to determine the coefficients x and y (Fiedler
and Schlachetzki 1987).
LC (Gax In1−x Asy P1−y ) = xyLC (GaAs) + x (1 − y) LC (GaP)

+ y (1 − x) LC (InAs) + (1 − x) (1 − y) LC (InP ) ,
where LC is the lattice constant. The composition of the semiconductor com-

pound is displayed over the bandgap energy or the bandgap wavelength at
room temperature in most diagrams and not x and y. The bandgap of com-
mon semiconductors is displayed in Fig. 3.17.
Physical properties of InP and of lattice matched Gax In1−x Asy P1−y at
room temperature (300 K) are presented in Table 3.2 (Fiedler and Schlachetzki
1987, Della Corte et al. 2000).
Fig. 3.17. Bandgap of III–V semiconductor compounds in eV and wavelength

Table 3.2. Physical properties of InP and of lattice matched Gax In1−x Asy P1−y at
room temperature (300 K)
symbol unit InP Gax In1−x Asy P1−y

lattice constant LC nm 0.58688
lattice match to InP y = 2.202x/(1 +
0.0659y)
bandgap energy Wg eV 1.35 1.35 − 0.72y +
0.12y 2
electron mass/m0 me /m0 0.077 0.07 − 0.0308y
heavy hole mhh /m0 0.6 0.6 − 0.218y +
mass/m0 0.07y 2
light hole mass/m0 mlh /m0 0.12 0.12 − 0.078y +
0.002y 2
refractive index neff 3.169
(λ = 1.55 µm)
dielectric constant εs 12.35 12.35 + 1.62y −
(static) 0.055y 2
dielectric constant ε∞ 9.52 9.52 + 2.06y −
(high frequency) 0.205y 2
density ρ g cm−3 4.81 4.81 + 0.74y
∆Wg /∆T 10−4 −2.67 −2.67 + 0.102y +
eV K−1 0.073y 2
dn
temperature K−1 2.01 × 10−5
dT
dependence of the
refractive index at
300 K
dn
temperature K−1 −2.17×10−10 ×
dT
dependence of the T2 + 3.5 ×
refractive index 10−7 × T +
(λ = 1.523 µm) 1.15 × 10−4
1 dl
length increase due · K−1 0.475 × 10−5
l dT
to temperature
The refractive index of InP plotted against the photon energy is shown in
Fig. 3.186 .
3.4.2 The Semiconductor Compound AlGaAs
The AlGaAs material system is used in the same context as the GaInAsP
material system and is a major optical material used for lasers and passive
waveguide based devices. This material system is used as a barrier material
in GaAs-based heterostructure devices. It is favorably used to fabricate field
6
http://www.ioffe.ru/SVA/NSM/nk/A3B5/Gif/inp.gif
Fig. 3.18. Refractive index of InP vs. photon energy
Fig. 3.19. Refractive index of GaAs versus photon energy
effect transistors like for example high electron mobility transistors (HEMTs).
Some common physical properties of GaAs and AlGaAs are listed in Table 3.3.
The refractive index of GaAs plotted against the photon energy is shown
in Fig. 3.197 .
3.4.3 Lateral Coupling in GaInAsP/InP

One of the first passive ring resonator filters in this material system are pre-
sented in Vanderhagen et al. (1999). Racetrack shaped ring resonators with a
7
http://www.ioffe.ru/SVA/NSM/nk/A3B5/Gif/gaas.gif
Table 3.3. Physical properties of GaAs and of Alx Ga1−x As at room temperature
(300 K)
symbol unit GaAs Alx Ga1−x As

lattice constant LC nm 0.565325 0.56533 +
0.00078x
bandgap energy x < 0.45:
1.424+1.247x
Wg eV 1.424 x > 0.45:
1.9 + 0.125x+
0.143x2
electron mass/m0 me /m0 0.063 0.063+0.083x
(x < 0.45)
heavy hole mhh /m0 0.51 0.51 + 0.25x
mass/m0
light hole mass/m0 mlh /m0 0.082 0.082+0.068x
refractive index neff 3.3 3.3 − 0.53x+
(λ = 1.55 µm) 0.09x2
dielectric constant εs 12.9 12.90 −2.84x
(static)
dielectric constant ε∞ 10.89 10.89 − 2.73x
(high frequency)
density ρ g cm−3 5.32 5.32 − 1.56x
temperature T eV 1.519−
dependence of the 5.405 · 10−4 ×
energy gap Eg (0 < T2 /(T + 204)
T < 103 in K).
dn
temperature K−1 −1.86 ×
dT
dependence of the 10−10 × T 2 +
refractive index 3.49 × 10−7 ×
(λ = 1.523 µm) T +1.47×10−4
http://www.ioffe.rssi.ru/SVA/NSM/Semicond/GaAs/index.html
http://www.ioffe.rssi.ru/SVA/NSM/Semicond/AlGaAs/index.html
Della Corte et al. 2000
radius of 125 µm and coupler lengths of 80 µm are demonstrated. The layout

and the layer sequence of the waveguide are shown in Fig. 3.20. The waveguide
layers are grown by metal organic vapor phase epitaxy (MOVPE). A 100 nm
thick SiN layer is used as the etching mask which is grown by PECVD. The
waveguide patterns are transferred to the SiN layer by photolithography and
a CHF3 based RIE process. The waveguides are etched using a CH4 /H2 and
O2 process.
A bi-level etching technique that permits the use of standard photolitho-
graphy for lateral coupling to deeply etched racetrack shaped ring resonator
structures is presented in Griffel et al. (2000), where a ring laser is fabricated
and characterized. Here, deep etching is used to reduce the bending loss at the
curved sections, but is avoided in the coupling region where the straight sections
Fig. 3.20. Waveguide structure and composition
Fig. 3.21. Layer sequence of the waveguides in the coupling region
of the racetrack shaped ring resonator and the input/output waveguides form
a gap whose distance defines the power coupling coefficient (Fig. 3.21). This
is a way of eliminating the drawback of lateral coupling and realizing defined
coupling parameters. A radius of curvature of 150 µm is combined with four
different coupler lengths of 50 µm, 100 µm, 150 µm, and 200 µm, positioned at
the center of the straight section of the racetrack, which is 50 µm longer. The
width of the used ridge waveguides is 2.5 µm and the gap between resonator
and coupling waveguides is 2 µm. The active waveguide consists of three com-
pressively strained GaInAsP quantum wells imbedded in a 70 nm waveguide
structure of two compositions with bandgap energy Eg = 1.13 and 1.00 eV.
The layer sequence of the used waveguide is shown in Fig. 3.21.
A similar method is used in Rabus and Hamacher (2001) where a 3 dB
MMI coupler of length 150 µm is used for coupling in and out of single and
serially double coupled racetrack shaped ring resonators. Here the rings were
deeply etched along the outer wall of their curved sections (Fig. 3.22) leading
to a stronger optical confinement.
Passive single and double racetrack shaped ring resonator with a radius of
100 and 200 µm, respectively, are presented. The layer sequence of the used
waveguide is shown in Fig. 3.23. The GaInAsP waveguide material used for the
realization of the devices has a bandgap wavelength of λg = 1.06 µm at room
temperature which can also be written as Q(1.06) referring to quaternary
material.
The fabrication of the waveguides starts with depositing the silicon nitride
(SiNx ) mask, realized by PECVD. It serves as the etching mask for the
Fig. 3.22. Scanning electron microscope (SEM) photograph of the input region of
an MMI coupler with deep etching along the outer walls of the curved sections
Fig. 3.23. Layer sequence of the waveguide
waveguides. The thickness of the SiNx is ≈200 nm and is deposited at a

temperature of 370◦ C. The structuring of the SiNx layer is performed using
standard photolithography. The photoresist used is AZ5214. An exposure
dose of 12 mW cm−2 is used. The exposing time is 24 s and the developing
time is between 40 and 50 s for a positive exposure. The developer used is
MIF724. The developed photoresist serves as the etching mask for the follow-
ing RIE process. The etching is done using the gases CHF3 (22 sccm) and O2
(2.2 sccm) at a pressure of 12 µbar and 50 W. The photoresist is removed after
this RIE step and the structured SiNx layer serves as the mask for the next
RIE step. The gases now used for etching the waveguide are CH4 and H2 (6,
40 ml min−1 ). In order to reduce the formation of polymers during dry etching
and to minimize the sidewall roughness, a small fraction of O2 (0.3 ml min−1 )
is added. The power used is 150 W at a pressure of 0.02 mbar. The etching is
controlled by an ellipsometer and a mass spectrometer. The etching is stopped
when the InP etch stop layer is detected. The next dry etching step is the
realization of the deeply etched section on the outer wall of the waveguides in
the curvatures. The SiNx layer from the previous step serves as the etching
mask which is structured again with standard lithography to open an etching
window for this process. The photoresist which is used to structure the SiNx
layer therefore covers only a part of the waveguide, ideally only half of it. This
self aligning process assures that the width of the waveguide is not changed
by this processing step. The etching is again performed using RIE, but this
time without the portion of oxygen which could partly remove the photoresist
mask and cause etching errors. The photoresist is removed after dry etching by
the use of an oxygen plasma (15 min, power = 500 W, T ≤ 200◦ C). An SEM
photograph of a waveguide in the curvature is shown in Fig. 3.24. The samples
are AR coated to avoid Fabry–Perot resonances in the bus waveguides.
In Rommel et al. (2002) a Cl2 /Ar/H2 chemistry is used to etch InP-based
racetrack ring resonators using inductively coupled plasma reactive ion etch-
ing. An SEM photograph of a fabricated racetrack shaped ring resonator notch
filter is shown in Fig. 3.25.
The composition of Cl2 /Ar/H2 has a strong influence on the degree of
undercut in the profile. A nearly perfect anisotropic profile is achieved for
ratios between 2/3/1 and 2/3/2. Waveguide losses of 2 dB cm−1 are obtained.
The straight length of the coupling region is 90 µm. The ridge width is
0.9 µm, and the coupling gap has a width of 275 nm. A radius of 15 µm
is used in the curved regions. The etching depth is 4 µm. Three distinct
etching profiles are obtained by adding H2 . For the Cl2 dominated region
(Cl2 /Ar/H2 ratios between 2/3/0 and 2/3/1.5), high etch rates: 1.8 µm/min
and large undercuts of InP and GaInAsP layers are reported. The second
region (Cl2 /Ar/H2 ratios between 2/3/1.5 and 2/3/2.5) has balanced chem-
istry yielding low but uniform etch rates between 0.5 and 0.6 µm min−1 and
Fig. 3.24. SEM photograph of a waveguide in the curvature

Fig. 3.25. The ridge is etched to a depth of 5 µm, and has a width of 0.9 µm and
gap of 275 nm.
Reused with permission from S.L. Rommel, Journal of Vacuum Science and Tech-
nology B, 20, 1327 (2002) 2002
c AVS The Science and Technology Society
Fig. 3.26. Layer structure of laterally coupled ring resonator in InP
highly anisotropic InP/GaInAsP profiles. A H2 dominated region is reported

for Cl2 /Ar/H2 ratios above 2/3/2.5. In this region, InP layers exhibit large
undercuts, whereas GaInAsP layers showed little lateral etching. The authors
suggest that the selectivity of GaInAsP layers relative to InP may be increased
by increasing the As concentration.
One of the smallest racetrack shaped ring resonators in InP with a radius
in the curved sections of 2.25 µm is presented in Grover et al. (2003) where
an FSR of 19 nm near 1,550 nm, with an extinction of more than 25 dB is
reported. The length of the straight sections is 10 µm and the coupling gap is
only 200 nm. The bus waveguide is tapered to a width of 3 µm away from the
coupling region to improve the fiber to chip coupling. The layer sequence of
the device is shown in Fig. 3.26.
The refractive index of the used Ga0.31 In0.69 As0.71 P0.29 compound is 3.39.
The fabrication details are taken from Grover (2003). The process is started
by depositing a wafer with 900 nm of silicon dioxide by PECVD. A Leica VB6
e-beam lithography machine is used to pattern a bilayer film of polymethyl-

methacrylate (PMMA). The patterning is started by spinning on 100 K MW
5% PMMA in Anisole at 4,000 rounds per minute (RPM) for 60 s, which gives
a 113 nm thick film. A bake follows for 15 min at 170◦ C. Next 495 K MW
PMMA A4 in Anisole is spin-coated at 2,000 RPM for 60 s, which leads to a
204 nm thick film. This is followed by a bake for 15 min. at 170◦ C. Finally
the pattern is written using a dosage of 736 µC/cm2 for waveguides and
ring resonators, and 610 µC cm−2 for device labels. The PMMA-based resist
is developed with 2-methyl-2-pentanone (methyl isobutyl ketone or MIBK)
diluted with 2-propanol as 1:3 for 2 min 30 s. In a final step the wafer is rinsed
with 2-propanol and nitrogen is used to blow-dry.
The reason for choosing two layers of PMMA is to have the lower resolution
PMMA on the bottom and the higher resolution PMMA on top, so that there
is an overhang of the higher resolution PMMA where the pattern is written.
This also enables a better liftoff in the following process. Next, a 50 nm thick
layer of chromium is deposited and patterned by liftoff in warm 1-methyl-2-
pyrrolidone (C5 H9 NO) also referred to as NMP. The pattern is transferred
to the underlying silicon dioxide by dry etching in trifluoromethane oxygen
plasma. Then the InP and quaternary layers are etched by a methane hydrogen
plasma to a depth of 3.2 µm using a Cr–SiO2 mask. The methane based dry
etching process evaluating several mask (Ni, NiCr, Ti, SiO2 , and Ti–SiO2 ) is
presented in Grover et al. (2001c). Then in a final step, the mask is stripped
with buffered hydrofluoric acid.
InP based devices are usually thinned to a thickness between 100 and
200 µm by a bromine methanol based lapping process. This ensures a better
cleavage of the samples and reduces the contact resistance if active devices
with top side p and backside n contact are realized.
Buried heterostructure ring resonators are presented in Choi et al. (2004b).
Ring resonators with a radius of 200 µm are fabricated by first growing a
0.4 µm thick GaInAsP (λg = 1.25 µm) waveguide layer on a (001) oriented
InP substrate by metal organic chemical vapor deposition (MOCVD). The
waveguide layer is then covered by a SiNx layer which serves as the mask
for the following CH4 based dry etching process. The bus waveguides are
aligned parallel to the [110] direction in the crystal lattice. The width of
the buried waveguides is between 0.8 and 0.9 µm. The polymers generated
on the sidewalls during the dry etching process are removed by an oxygen
plasma treatment and the SiNx mask is removed by a buffered oxide etchant.
The waveguides are covered by a 1.5 µm thick InP layer. Finally the sample
is thinned, cleaved, and AR coated. The waveguide loss of the straight bus
sections is measured to be between 0.2 and 0.3 cm−1 . The dimensions and
layout of the buried waveguide are shown in Fig. 3.27.
SEM photographs of a waveguide section and a cross-section of a buried
waveguide are shown in Fig. 3.28.
Fig. 3.27. Dimensions and layout of the buried waveguide
Fig. 3.28. (a) SEM image of a dry-etched ring resonator mesa. (b) SEM cross-
sectional view of a buried waveguide.
pp. 2266–2268 2004
c IEEE
3.4.4 Vertical Coupling in GaInAsP/InP
One of the first vertically coupled disk resonators with a diameter of 44 µm,
both in the GaAs and InP materials systems, fabricated by wafer bonding
is presented in Tishinin et al. (1999). The device design in both material
systems incorporates an etch-stop layer (0.2 µm thick layer of InGaAs/InP
and Al0.95 Ga0.05 As/GaAs) and two waveguiding layers (waveguide layers
0.5 µm thick and disk layers 0.3 µm thick) with low index material layer
(thickness 0.5 µm) in between. In the GaAs sample, Alx O is used for lateral
confinement of the light in the disk structure. The bus waveguides and disk
are fabricated by standard optical lithography and chlorine based ECR-RIE.
First, the waveguides are etched down through both the top cladding layer
and the waveguiding layer and etching is stopped within the separation layer.
The bus waveguides are multimode and have a width of 2 µm. After etching
the bus waveguides, this processed side is wafer bonded to another substrate
and the original substrate is removed after it has been lapped down to a
thickness of 100–150 µm by selective chemical wet etching. GaAs samples are
bonded to GaAs and InP samples to InP substrates. A bonding temperature
for GaAs wafers of 750◦ C and of 400◦ C for InP-based materials is used. It is
important to mention that the transfer substrate need not be the same mate-
rial. After the removal of the original substrate, the disk is fabricated by back
side alignment optical lithography and ECR-RIE.
A vertically coupled GaInAsP thin film disk resonator with a diameter of
10 µm resulting in an FSR of 20 nm using polymer wafer bonding with benzo-
cyclobutene (BCB) is presented in Ma et al. (2000c). The wafer structure with
a 0.4 µm thick GaInAsP (λg = 1.2 µm) guiding layer is grown on an InP sub-
strate by molecular beam epitaxy (MBE). A 400 nm thick SiO2 layer is used
as a hard mask and is deposited by PECVD. The wafer is patterned using
electron beam lithography with a 180 nm thick 2% polymethylmethacrylate
(PMMA) resist. The structured PMMA resist is transferred to the underly-
ing SiO2 layer using RIE. Inductively coupled plasma (ICP) RIE with a gas
mixture of Cl:Ar (2:3) is used to transfer the pattern onto the epitaxy wafer
through the SiO2 hard mask at a temperature of 250◦ C. The waveguides are
etched to a depth of about 1.3 µm. The SiO2 hard mask is removed after RIE
using buffered HF. The BCB bonding process is started by spin coating BCB
onto both the patterned wafer and the transfer wafer (here GaAs is used).
Then the wafers are brought into contact with their BCB side facing each
other using an appropriate weight. The wafer stack is put into a nitrogen
filled furnace at 250◦ C for 1 h. BCB becomes fully cured and the two wafers
are glued together. The overall thickness of BCB is 3 µm. The InP substrate
is removed after bonding using selective wet etching (HCl : H3 PO4 = 2 : 3).
The input and output waveguides are 2 µm wide and are tapered down (taper
length = 500 µm) to 0.4 µm in the disk region. The bus waveguides and the
disk are separated by a gap of approximately 180 nm.
Vertically coupled single mode ring resonators also using BCB wafer bond-
ing with a radius of 5 and 10 µm are presented in Grover et al. (2001a) and
Grover (2003). The layer sequence of the used device is shown in Fig. 3.29.
The patterns are transferred by photolithography to a 400 nm thick SiO2
layer using a 10× i-line stepper and a CHF3 /O2 plasma RIE process. The
waveguides and ring resonator in GaInAsP–InP are etched using SiO2 as the
mask in a reactive ion etching system with a CH4 /H2 /Ar plasma.
The fabrication process starts with etching of alignment keys to a depth
of 2 µm down to the GaInAs layer. The use of Alignment keys eliminates the
use of infrared (IR) backside alignment for processing the other side. Then
Fig. 3.29. Layer structure used for BCB bonded vertically coupled ring resonators
Fig. 3.30. Design of vertically coupled disk resonator
the waveguide layer is etched to a depth of 0.9 µm. Benzocyclobutene (BCB;

n = 1.5) is used for bonding to a GaAs transfer substrate. A thickness of
1–2 µm is chosen for the BCB layer. The growth substrate is thinned to a
height of about 100 µm by chemo-mechanical polishing. The remaining growth
substrate is selectively etched and removed using H3 PO4 :HCl (1:1) which
stops at the GaInAs etch stop layer. This etch stop layer is then selectively
etched using H2 O2 :H2 SO4 :H2 O (1:1:10) which stops at the InP layer opening
the second GaInAsP–InP layers for fabricating the ring resonator. The top
layer is etched to a depth of 0.9 µm, leaving a gap of 200 nm to decrease losses
caused by leakage to the substrate. Finally the whole device is encapsulated
with BCB.
This fabrication process uses the advantages of both materials combining
GaInAsP and GaAs. As appropriate, either layers of GaInAsP or GaAs are
removed leading to a vertically coupled device.
A thermal wafer bonding process is used in Djordjev et al. (2002c) to realize
vertically coupled disk resonators with radii of 4, 6, 8, 10, 12 µm. The layer
sequence and design are shown in Fig. 3.30.
A novel feature which is incorporated in the design of the disk resonator
is a circular post below the resonator. It improves the mechanical stability
of the structure, improves current/field uniformity in active devices and sup-
presses higher order modes improving the transmission characteristics. A disk
supports whispering gallery modes, which propagate by means of total internal
reflection from the disk/air interface (see Chap. 6). The first order mode has
the smallest volume and occupies the outmost region of the disk in the radial
direction. All of the higher order modes have larger volume and are therefore
closer to the center of the disk which results in a leakage into the post. The
post and the waveguides are separated by 1.2 µm which is large enough to
prevent coupling between them.
The fabrication process is described using (Djordjev 2002). The layers
are grown by low pressure MOCVD on a (001) InP substrate at 655◦ C.
A 100 nm thick SiNx is used as a hard mask for the bus waveguide and disk
etching process which is deposited at 275◦ C with a power of 30 W using a
pressure of 450 mTorr8 and the gases SiH4 /NH3 /N2 at a ratio of 40/20/60.
This silicon nitride layer is structured using Shipley S1813 photoresist and a
low pressure (10 mTorr) ECR-RIE CF4 (30 sccm) plasma at 20◦ C, 500 W for
300 s. Before spin coating the photoresist, Silicon Resources AP405 is used to
increase the adhesion of the photoresist on the silicon nitride. The photore-
sist is developed using Microposit MF321 developer. The bus waveguides are
etched using a CH4 /H2 /Ar RIE process. The process used is described in detail
in Choi et al. (2002b). The width of the waveguides is between 0.7 and 0.9 µm.
The waveguide width is tapered adiabatically to 2.5 µm to the edge of the
chip to enhance fiber chip coupling. After the etching of the bus waveguides,
the sample, which is usually 1 cm2 , is bonded to a transfer substrate of the
same size. The cleanliness of the surface of the samples is vital for the fol-
lowing bonding process and many cleaning steps are performed (deionized
(DI) water, trichloroethylene (TCE), acetone, methanol). After a final DI
water rinse, both samples are brought into contact under water and are “wet
bonded.” This prevents oxidation and contamination of the surfaces. The sam-
ples are then loaded into a chamber which is evacuated and heated to 100◦ C
for 10 min. Then H2 is led into the chamber and the samples are heated to a
bonding temperature of 505◦ C for 30 min. In a final bonding step, the samples
are cooled down to room temperature in 15 min. The fabrication of the disk
starts by removing the InP substrate which is done using mechanical polishing
and a selective chemical wet etch (HCL:H2 O – 3:1). The etching is automati-
cally stopped, when the InGaAs layer is reached. The etch stop layer is then
removed using H2 SO4 :H2 O2 :H2 O (1:1:3) which opens the waveguide layers of
the disk. In order to align the bus and the disk waveguides, the sample is
etched down to the waveguide pattern on the edge in a small area making
the alignment marks visible which have been patterned together with the bus
waveguides. The disk is etched after the opening of the alignment marks using
the same processing steps as for etching the bus waveguides.
One of the first ring resonators coupled to buried heterostructure (BH)
bus waveguides in GaInAsP are presented in Choi et al. (2004a). A two step
regrowth process is used in fabricating the devices. The radius of the ring
resonators is 10 µm. The buried bus waveguides are made by growing a 0.4 µm
8
760 Torr = 1 atm = 101, 325 kPa = 1013, 25 hPa = 1, 01325 bar
Fig. 3.31. Cross-section of a planar vertically stacked waveguide produced on BH

mesas.
pp. 828–830 2004
c IEEE
thick layer of GaInAsP with a bandgap wavelength of 1.1 µm on to an InP

substrate followed by an SiNx masking and etching step. The width of the
bus waveguides is 1 µm. The area between the bus waveguides is filled with
epitaxially grown InP. The silicon nitride masking layer is removed and the
second growth step is initiated leading to the ring resonator. The material
for the waveguide layer which has a thickness of 0.4 µm is GaInAsP with a
bandgap wavelength of 1.4 µm. The ring waveguide and the bus waveguides
are separated by a 0.8 µm thick InP buffer layer. The InP cladding layer for
the ring waveguides has a thickness of 1 µm. An SEM photograph of a cross
section of the ring resonator device showing the buried bus waveguides and the
layer for fabricating the ring resonator is shown in Fig. 3.31. The fabricated
ring resonator is shown in Fig. 3.32. The bus waveguides are AR coated to
avoid Fabry–Perot resonances.
3.4.5 Lateral Coupling in AlGaAs/GaAs
One of the first ring and disk resonators based on this material system are
presented in Rafizadeh et al. (1997b). Ring and disk resonators with diameters
of 10.5 µm and 20.5 µm have been fabricated and characterized. The layer
sequence of the used device is shown in Fig. 3.33.
The layers are grown by MBE. The coupling gap used is only 100 nm
wide. A waveguide loss of 3.2 cm−1 is measured for TM polarized light at
a wavelength of 1.55 µm. The devices are created by writing patterns upon
a PMMA resist layer by electron-beam lithography (JEOL JBX 5DII). The
AlGaAs/GaAs is etched using chemically assisted ion-beam etching (CAIBE).
Fig. 3.32. Ring resonator coupled to buried heterostructure bus waveguides.

pp. 828–830 2004
c IEEE
Fig. 3.33. Layer structure of AlGaAs/GaAs based ring and disk resonator
The electron-beam lithography exposure conditions are a 300 pA probe cur-

rent, a 50 eV acceleration potential, an 80 µm × 80 µm field, and an 11 mm
working distance. The PMMA mask is transferred to an underlying SiO2 layer
by RIE which provides a robust mask capable of withstanding the CAIBE
process. The CAIBE parameters used, are a beam voltage of 500 V, a beam
current density of 0.14 mA cm−2 , a chlorine flow rate of 15 standard cubic
centimeters per minute (SCCM) at STP9 , an argon flow rate of 2 SCCM,
and an elevated substrate temperature of 100◦ C. The SiO2 mask is removed
after the CAIBE process. The etch rate achieved is 0.1 µm min−1 . The sidewall
roughness (vertical striations) observed under the SEM is 10–20 nm.
A similar layer sequence with waveguide parameters in the same range is
presented in Absil (2000) where racetrack shaped ring resonators with radii
between 1.2 and 10 µm have been realized. The layer sequence and dimensions
of the used waveguide is shown in Fig. 3.34.
9
Standard temperature and pressure. The volume measurement is made at or
adjusted to a temperature of 0◦ C and a pressure of 1 atmosphere, or 101.325 kPa
Fig. 3.34. Layer sequence and dimensions of AlGaAs/GaAs waveguide
A coupling gap between 0.1 and 0.3 µm is envisaged. The fabrication of the
laterally coupled ring resonator devices starts by depositing a 420 nm thick
SiO2 layer with a nominal deposition rate of 13.6 nm min−1 by PECVD at
300◦ C which serves as the mask for the CAIBE process. Before deposition
of the SiO2 layer, the sample is rinsed in acetone, methanol and isopropanol
followed by a 20 min soak in OPD 4262 (a photoresist developer) to increase
oxide adhesion on the surface of the wafer. After deposition of the oxide mask
layer, the sample is solvent cleaned and dried in a vacuum oven at 120◦ C for
1 h. This step is used to increase the adhesion of the electron beam PMMA
resist on the oxide. The PMMA resist is spin coated onto the sample leading
to a thickness of 500 nm. The resist is cured on a hotplate at 195◦ C for 150 s.
A Leica Cambridge EBMF 10.5 is used at 50 keV to transfer the ring resonator
patterns to the resist. The sample is developed using a solution of MIBK:IPA
(1:3) at 25◦ C for 150 s and rinsing with isopropanol. The patterns are trans-
ferred from the PMMA to the underlying oxide mask layer by RIE using
CHF3 and O2 (18:2) with a pressure of 40 mT and an RF power of 175 W.
The etch rate of the oxide is 32.5 nm min−1 . The waveguides are etched using
a CAIBE process in an ultrahigh vacuum system with 200 eV argon ions and
chlorine gas. The etch rate for GaAs is 0.08 µm min−1 . The etching process is
performed in three etching runs in order for the sample to cool down. After
etching, the oxide mask is removed in an HF dip. Then the sample is spin
coated with a 4 µm thick layer of cyclotene or BCB which serves as a protec-
tion and planarization layer. The polymer is cured at 200◦ C for 2 hours. The
chip is thinned down in a final step to a thickness of 100 µm which is similar
to the InP based devices, also using a bromine/methanol chemo-mechanical
process. In order to avoid Fabry–Perot resonances in the bus waveguides, the
facets are AR coated using a 250 nm Al2 O3 layer which has a refractive index
of n = 1.62.
3.4.6 Vertical Coupling in AlGaAs/GaAs
One of the first BCB bonded vertically coupled ring resonators in this material
system is presented in Absil et al. (2001), Grover et al. (2001b). Here a single
Fig. 3.35. Layer sequence of vertically coupled GaAs–AlGaAs ring resonators
ring resonator channel dropping filter and a 1 × 4 multiplexer/demultiplexer

(MUX/DEMUX) crossbar array with double ring resonator filters is demon-
strated. The radius of the fabricated ring resonators is 2.5 µm, 5 µm, and
10 µm. The waveguide width is 0.7 µm. The waveguides are tapered to 2.5 µm
towards the facet of the chip to improve the fiber to chip coupling efficiency.
The layer sequence of the used device is shown in Fig. 3.35. The layers are
grown by MBE. A 10× i-line stepper is used for transferring the patterns
by photolithography. ICP etching is used for fabricating the bus and ring
waveguides. BCB with a thickness between 1 and 2 µm is used to bond
the sample to a GaAs transfer substrate and cured at 200◦ C. After removing
the original substrate, alignment of the ring resonator layer and etching the
same, the sample is encapsulated with BCB. This ensures a refractive index
symmetry profile, which is also necessary when identical ring resonator struc-
tures are realized on both layers for multiring resonators. In order to reduce
the effect of layer misalignment a thin layer (∼0.2 µm) of AlGaAs is being
unetched.
3.4.7 Implementation of Gain in Ring Resonators
Introducing gain inside the ring resonator adds additional functionality and
device features. One of the first ring resonators with integrated semiconduc-
tor optical amplifiers (SOA) in the GaInAsP material system is demonstrated
in Rabus et al. (2002b). Single, double and triple parallel and serially cou-
pled ring resonators are presented in Rabus (2002). A photograph of a ring
resonator with integrated SOA is shown in Fig. 3.36.
The SOA integration is based on a selective area MOVPE process which
has also been used in Hamacher et al. (2000) to integrate lasers, photodi-
odes, waveguides, and spot size converters. The cross section of the used ridge
waveguide (RW) semiconductor optical amplifier (SOA) structure is shown in
Fig. 3.37. The dimensions are given in Fig. 3.38.
The layer sequence is given in Table 3.4. The layer sequence and compo-
sition of the used SOA is as follows (from bottom to top).
Fig. 3.36. Photograph of a ring resonator with integrated SOA
Fig. 3.37. SEM of cross-section of ridge waveguide SOA, covered with metal contact
layer
Fig. 3.38. Device structure of the SOA ridge waveguide

Table 3.4. Layer stack of the used SOA
no. material λg (nm) thickness dopand (cm−3 ) function

1. InP–Sn-sub. 360 µm substrate
2. InP-buffer 500 nm Si : 3 × 1018 buffer
3. n-GaInAsP 1,150 250 nm Si : 3 × 1018 n-contact
4. n-GaInAsP 1,290 10 nm undoped barrier
5. Q 1% compressive strained 6 nm undoped 6 × QW
6. Q-1.29 1,290 10 nm undoped 6 × barrier
7. GaInAsP 1,150 180 nm undoped LD-WG
8. InP 1,500 nm Zn : 5 × 1017 p-top
9. GaInAsP 1,300 50 nm Zn : 1 × 1018 p-top
10. InGaAs 200 nm Zn : 1 × 1019 p-contact
Fig. 3.39. Mode profile of the SOA (TE polarization)
The calculated mode field profile of the RW structure is given in Fig. 3.39.
The quantum wells (QWs) and the barrier layers have been considered as one
layer with an effective refractive index of neff = 3.4. Due to the multiquan-
tum well structure, the SOA was designed for favoring TE polarization. The
effective index of the SOA is determined to be neff = 3.238. The center of
the guided mode in the RW structure is located lower compared with the pas-
sive waveguide structure. Thus, it is necessary for the integration process to
adjust the height of the active–passive transition to assure minimal coupling
losses.
The butt coupling losses at the passive–active waveguide interface have
been calculated by the finite difference method (Fig. 3.40). The calculated
vertical and lateral offset between the active and passive waveguide results in
a minimum theoretical coupling loss of <1 dB.
The starting point for the calculation of the horizontal shift has been
chosen so that the passive structure is located symmetrical in the center of
the active section. The calculation of the vertical shift starts with the passive–
active section, butt coupled at the position where the rib starts for the passive
Fig. 3.40. Calculated butt-joint losses
Fig. 3.41. Lateral dimensions of the active–passive waveguide interface
structure and in the middle of the active layers for the SOA section. The
practical realization of the active–passive transition is done in using a special
taper structure at the interface. The lateral design is shown in Fig. 3.41.
The use of the tapered structure enables the propagating wave coming
from the passive waveguide to laterally broaden, interfere, and propagate into
the active waveguide with a different refractive index and scale down to the
dimension of the active waveguide.
Another technological challenge is the fabrication of a low resistance con-
tact of the active section. As directly contacting the SOA on the ridge with
a metal needle would destroy the SOA in the worst case, so called “support
mesa” are designed at a distance of a few micrometer away from the SOA
ridge on either side. The p-contact of the SOA is then located on these mesa
for securely placing a measurement needle and for later wire bonding. The
support mesas have the same active material layer stack as the SOA. The
technological challenge lies in creating a low metal–semiconductor p-contact
on the SOA. The isolation of the metal layer on top of the support mesa
is performed with a SiNx layer, which is removed only on top of the SOA
waveguide. A sketch of the cross section of the SOA with support mesa is
shown in Fig. 3.42. The n – contact is realized by metal coating the backside
of the wafer.
Fig. 3.42. Cross-section of the SOA section with support mesa
The fabrication starts with the epitaxial growth of the SOA layers by
using MOVPE. The first CVD processing step is the deposition of an SiNx
layer which is ≈230 nm thick at a temperature of 370◦ C for the fabrication
of the etching mask. The lateral active mesa structures are defined in a resist
pattern by using standard photolithography (positive process, photoresist
AZ5214, exposing time ≈24 s, developer MFI724, developing time ≈50 s).
The resist pattern serves as an etching mask for the following reactive ion
etching step (CHF3 – 22 SCCM and O2 – 2.2 SCCM, pressure = 0.012 mbar,
power = 50 W). The photoresist is removed with an O2 plasma (10 min,
power = 500 W, T < 250◦ C) after this process. The p-top mesa are etched
in the following step by using RIE (CH4 – 8 ml min−1 and H2 – 20 ml min−1 ,
pressure = 0.006 mbar, power = 200 W). In order to remove the poly-
mer which is formed during reactive ion etching, an O2 plasma (20 min,
power = 550 W, T < 250◦ C) and a solution of KOH (20%) are used. The
p-top is etched down to a distance of about 1.1–1.2 µm followed by a wet etch-
ing step by using a solution of 20 H2 O:5 HBr:H2 O2 . This wet etching process
is the “undercut” etching. In this step, part of the sidewall is removed. A layer
of InP protects the active sections underneath. The etching of the undercut is
indispensable for the second epitaxial growth step of the passive waveguide.
The undercut reduces the formation of high “rabbit ears” during selective
area MOVPE at the vicinity of the mask and enables adequate control of
the layer thickness in order to achieve the necessary vertical alignment within
the active – passive transition. For further improvement of the SAMOVPE
growth step, the “undercut” is covered by a layer of SiNx which is realized
by depositing the entire wafer with SiNx and removing the material on the
bottom of the wafer by reactive ion etching, leaving the sidewalls covered. The
next step is the final etching of the p-top and the active layers using RIE (CH4
– 8 ml min−1 and H2 – 20 ml min−1 , pressure = 0.006 mbar, power = 200 W).
The control of the etch depth is very important in order to adjust the re-
growth of the passive material. The following step is the regrowth of the
passive material using selective area MOVPE (SAMOVPE). The correct
height of the passive material is defined by the thickness of an InP buffer
layer. The regrowth process is followed by the removal of the entire SiNx at the
“undercut” using hydrofluoric acid (HF – 5%). After this step, the entire wafer
Fig. 3.43. Active–passive transition
is covered again with a SiNx layer (≈190 nm) at a temperature of 370◦ C. An

SEM photograph of the active – passive transition is shown in Fig. 3.43.
The SiNx is removed at the top of the active mesa using a photolitho-
graphic step, followed by a dry etching step (CHF3 – 22 SCCM and O2 –
2.2 SCCM, pressure = 0.012 mbar, power = 50 W). The photoresist is removed
after the dry etching step using KOH (20%). The following diffusion process
(T = 575◦ C, 18 s) is carried out using zinc arsenide as the dopand. In
the next step, the waveguide is etched together with the laser ridge in a
so called self-aligning process, where the active–passive interface is gener-
ated using RIE (CH4 – 6 ml min−1 , H2 – 40 ml min−1 , O2 – 0.3 ml min−1 ,
pressure = 0.02 mbar, power = 150 W). Due to the larger height of the laser
ridge, the remaining material is etched by selective wet chemical etching using
a solution of HCl and H3 PO4 (1:4). The wet etching is automatically stopped,
when the quaternary layers are reached. The deep etching of the passive
waveguide, which is necessary in order to achieve sufficient optical confine-
ment in small bent waveguides (radii <200 µm), is the next processing step.
The etching mask material is again a SiNx layer structured in a photolitho-
graphic and dry etching step. The SiNx layer also serves as a protection layer
for the laser ridge. The deep etching of the passive waveguide is done by reac-
tive ion etching using CH4 (8 ml min−1 ) and H2 (20 ml min−1 ) with a pressure
of 0.006 mbar and a power of 200 W. The SiNx which was deposited for the
protection of the SOA ridge has to be removed only at the very top of it,
to be able to place a metal layer for realizing an electric contact. Therefore,
the entire wafer is covered with photoresist and is exposed for a few seconds
(≈14 s) at the areas of the SOA ridge which results in opening the tip of the
laser ridge. The photograph of the top view of an SOA ridge with structured
photoresist is shown in Fig. 3.44.
In this process the exposing time of the photoresist is the key figure. If the
time is too high, the photoresist will become very thin between the SOA ridge
and the support mesa and might uncover the insulating SiNx layer which will
be etched in the next fabrication step and so cause unwanted short cuts. On
the other hand, if the exposing time is too short, only part of the photoresist
on top of the SOA ridge is removed and the SiNx layer can not entirely be
Fig. 3.44. Top view of an SOA ridge with structured photoresist
Fig. 3.45. SEM photograph of the cross section of a developed SOA ridge
etched away in the following process. The SEM photograph of the cross-section
of a developed SOA ridge is shown in Fig. 3.45.
The SiNx layer at the top of the SOA ridge is completely free of pho-
toresist. The area in between the SOA ridge and the support mesa and also
the entire wafer are covered by photoresist with a thickness of about 800 nm
which is a sufficient amount for protecting the SiNx layer from being etched
in the next step. The SiNx is removed from the p-top in a dry etching process
(CHF3 − 24 SCCM, H2 – 1 SCCM, pressure = 0.012 mbar, power = 50 W).
Now, the metal contacts are fabricated starting with a photolithographic step
for the definition of the contacts. The p-top of the SOA, the area in between
the ridge and the support mesa and the top of the support mesa is deposited
with titanium, platinum, and gold and the remaining metal is removed in a
“liftoff” process using NMP (n-methyl-2-pyrrolidon, C5 H9 NO).
A vertically stacked asymmetric twin waveguide structure is used in Menon
et al. (2004) to integrate an SOA into an MMI coupled single ring resonator
notch filter. The asymmetric twin waveguide structure is grown on a (100)
Fig. 3.46. Layer sequence of asymmetric twin waveguide structure
n+ InP substrate using gas source MBE. The layer sequence is shown in
Fig. 3.46. The width and height of the single mode passive waveguide is 4 µm
and 1.9 µm, respectively. The SOA is formed by etching a 500 µm long stripe,
and is coupled optically to the lower passive guide via two-section adiabatic
lateral tapers. The width of the taper varies linearly from 3.5 to 2 µm over
a 25 µm length followed by a second linear section that varies from 2 µm to
0.5 µm over a 150 µm length, resulting in a coupling loss of <0.5 dB per taper.
The length and width of the MMI is 685 µm and 12 µm, respectively. The
input and output waveguides are angled at 7◦ C to the cleaved edge of the
chip in order to avoid Fabry–Perot resonances in the bus waveguide. The
wafer is planarized using silicon nitride (SiNx ) followed by a short etch back
to open the SOA ridges and the taper. Electron beam evaporation was used
to deposit Ti–Ni–Au p-contacts on the active ridge including the tapers. The
p-contacts are annealed at 415◦ C for 30 s. The wafer is thinned to 130 µm,
and a Ge–Au–Ni–Au n-contact is deposited on the back of the substrate, and
annealed at 360◦ C for 90 s.
Total internal reflection (TIR) mirror based GaInAsP ring resonator notch
filters integrated with SOAs are presented in Kim et al. (2005b). The ring
resonators are made out of four TIR mirrors, an MMI coupler with a length
of 113 µm and the SOA (see also Fig. 5.36). The layer sequence of the active
and passive waveguide is shown in Fig. 3.47.
The passive waveguide A which leads to the facet and the TIR mirrors
is tapered down (taper length 15 µm) from a width of 1.5 µm to a width
of 3 µm. The edge depth is 1.8 µm for this type of waveguide as well as for
the active waveguide. The passive waveguide B is connected to waveguide
A and differs only from type A by a deeper etch depth which is 4.5 µm.
The fabrication of the device is started by growing the active layers on a n+
InP substrate. The multiquantum wells consist of seven quantum wells, each
6.5 nm thick, separated by barrier layers with a thickness of 8 nm. The active
layers are etched away in unwanted areas in the next step, which is followed
by a regrowth step with passive material. Standard RIE and metallization
techniques are used for realizing the device.
Fig. 3.47. Layer sequence of active and passive waveguide used in TIR ring
resonator
Table 3.5. Layer sequence of active disk (Djordjev 2002)
layer no. material bandgap λ thickness (nm) doping

17 InP 300 n+ , 106
16 InP 700 n, 5 × 1017
15 GaInAsP 1.1 400 n, 3 × 1017
14 InP 800 n, 3 × 1017
13 GaInAsP 1.25 170
12 GaInAsP 1.55; 1% strain 4×8
11 GaInAsP 1.25 3 × 10
10 GaInAsP 1.25 170
9 Inp 100
8 InP 100 p, 3 × 1017
7 InP 100 p, 5 × 1017
6 InP 600 p, 1 × 1018
5 InGaAs 200 p+ , 1 × 1019
4 InP 100
3 InGaAs 200
2 InP 100 n, 3 × 1018
1 InP n+
Vertically coupled disks with radii of 10 µm employing gain are demon-

strated in Djordjev et al. (2002b). The fabrication of the vertically coupled
disk is based on the passive disk explained earlier (see Fig. 3.30). The layer
sequence of the active disk is listed in Table 3.5. In order to limit the inter-
action of the optical field with the metal bond pads, a 3 µm thick layer of
polyimide is applied between the surface of the substrate and the metal layer
by using a standard liftoff process which leaves the active regions polyimide
free for creating the metal contact.
Finally the sample is thinned down to a thickness of 100 µm, the n-metal
contact is deposited and contact annealing is performed. Then the samples
are cleaved into bars and deposition of AR coating on both facets is applied.
3.5 Polymers 83
Fig. 3.48. Cross-sectional scanning electron microscope photograph of an active

microdisk device.
Reused with permission from K. Djordjev, Applied Physics Letters, 80, 3467 (2002)
2002
A SEM photograph of the cross section of a vertically coupled disk is shown

in Fig. 3.48.
These active disk resonators are the first of its kind employing gain to
change the resonant filter characteristic.
A detailed theoretical analysis of vertical coupled ring resonators employ-
ing active and passive layers is presented in Tee et al. (2006). A design layout
based on the calculations is developed.
3.5 Polymers
Polymer photonic integrated circuits have always been playing an important
role in literature (Ma et al. 2002). Polymers have recently attracted again a
lot of attention due to the lower cost of the material (if standard polymers
are used), easier processibility and integration than semiconductor materials.
Polymers can be doped with appropriate materials for engineering desired
properties. Different polymer material systems exist on the market today for
realizing not only purely passive devices, but also electro-optic and thermo-
optic devices. In the following chapter polymer based ring resonators made
by using different fabrication technologies will be presented.
3.5.1 Conventional Fabrication Techniques

In order to fabricate polymer waveguide based components, mainly two
processing methods are used, which are well known from the semiconductor
Fig. 3.49. Photoresist (left) and direct lithographic (right) patterning for realizing
polymer waveguides
Fig. 3.50. Polymer waveguide realized by RIE and spin coating
field, namely photoresist based patterning and direct lithographic patterning

(laser direct writing, photobleaching, electron beam direct writing). Both
methods are briefly sketched in Fig. 3.49.
A polymer waveguide generally consist of a core and a cladding polymer
where the refractive index of the core is higher than the cladding to enable
the propagation of optical waves. Of course the cladding can also be left out
and air can be used instead, which depends on the application. Figure 3.50
shows a typical photograph of the facet of a polymer waveguide, realized by
RIE and spin coating in poly(perfluorocyclobutylaryl ether) (PFCB). As can
be seen from the photograph the core waveguide has a nearly rectangular
shape. RIE processes are aimed for obtaining smooth and vertical sidewalls
to reduce scattering losses.
3.5 Polymers 85
One of the first integrated polymer ring resonator add drop filters with
a radius of 4.5 cm is presented in Haavisto and Pajer (1980). The ring res-
onator is fabricated by photopolymerization of doped polymethylmethacry-
late (PMMA) films spun onto a silicon wafer. The patterns are written by a
focused 325 nm He–Cd laser. The width of the achieved waveguides is 10 µm.
The coupling gap is 17.5 µm, measured from the waveguide center to center.
A prism coupling arrangement is used for measuring the device.
A ring resonator notch filter with a radius of 15.9 mm using polymers syn-
thesized from deuterated10 methacrylate and deuterated fluoromethacrylate
are described in Hida et al. (1992). The ring resonator is fabricated on a sili-
con substrate by spin coating, photolithography and dry etching. The height
and width of the waveguide core is 6.4 and 7.0 µm, respectively. The relative
refractive index difference between the core and the cladding is set to 0.5% by
controlling the copolymerization ratio of the two monomers. A coupling gap
of 3 µm is realized between the bus waveguide and the ring.
Polymer ring resonator filters and modulators with vertically coupled input
and output waveguides are demonstrated in Rabiei et al. (2002a). Two differ-
ent kinds of ring resonator filters are fabricated, one with an index difference
between core and cladding of 0.1 and a radius greater than 220 µm, and a sec-
ond one with an index difference of 0.3 between core and cladding and radius
greater than 25 µm. The layer sequence and the dimension of the waveguides
used in the passive ring resonator filter are shown in Fig. 3.51. A photograph
of a passive ring resonator filter is shown in Fig. 3.52.
The fabrication of the device with an index difference of 0.3 between core
and cladding starts by spin coating a 2.8 µm Teflon film (11% solution of Teflon
AF 1600 (Dupont) in 3M FC-40 solvent) on a silicon substrate. The Teflon
layer is etched for 2 minutes in oxygen by RIE to improve the adhesion of the
Teflon to the next layer. A 1.5 µm SU-8 (negative photoresist) layer is spin
coated onto the etched Teflon layer and patterned by lithography to form the
ring resonator. The ring resonator waveguides are planarized by spin coating a
4.5 µm thick Teflon layer. The Teflon layer is again etched by RIE in oxygen to
increase the adhesion of the following polymer layer. The bus waveguides are
Fig. 3.51. Layer sequence and dimensions of waveguides (refractive indices are given
for a wavelength of 1,550 nm)
10
Introduction of deuterium into the chemical compound
Fig. 3.52. Photograph of a passive polymer ring resonator add–drop filter.

Reprinted with permission from IEEE Journal of Lightwave Technology, vol. 20,
pp. 1968–1975 2002
c IEEE
Fig. 3.53. Layer sequence and dimensions of waveguides used in the modulator
realized by lithographic patterning, RIE etching of the Teflon cladding layer,

spin coating a UFC 170A (URAY Co., Korea) layer, UV curing, and baking.
Remaining UFC 170A which is not part of the waveguide is removed using
RIE. For the fabrication of the ring resonator filter with an index difference
of 0.1 between core and cladding layers the cladding material is UFC 170A
and the vertically coupled bus waveguides are made out of NOA61 (Norland
Products). The fabrication of the polymer ring resonator modulator is in
principal the same as for the passive device. The layer sequence of the ring
resonator modulator is shown in Fig. 3.53.
The CLD1/APC polymer is aligned using corona polling for 30 min at a
polling temperature of 145◦ C and an applied voltage of 10 kV. An upper gold
electrode is deposited and patterned on top of the device to cover the ring
resonator modulator. A photograph of a fabricated ring resonator modulator
3.5 Polymers 87
Fig. 3.54. Fabricated polymer ring resonator modulator.

pp. 1968–1975 2002
c IEEE
A free standing polymer ring resonator notch filter is presented in Paloczi

et al. (2003a), Huang et al. (2004). The polymer waveguides are embedded in
a lower cladding of UV-15 (n = 1.5, λ = 1.55 µm) and an upper cladding of
OG-125 (n = 1.46, λ = 1.55 µm). The waveguide cores which are made of SU-8
(n = 1.57, λ = 1.55 µm) are 2 µm wide and 1.6 µm thick. The SU-8 waveguides
are patterned by direct electron beam lithography and following development
process. The upper cladding is spin coated and UV cured. The free standing
feature is achieved by a thin gold layer deposited on to the silicon substrate
onto which the UV-15 layer is spin coated. Due to the weak adhesion of the
UV-15 layer, it is possible to peal of the polymer layer (thickness ∼10 µm) with
the integrated ring resonators. Optical quality facets are realized by simple
cutting with a scalpel.
One of the smallest ring resonator add drop filters with a radius of 10 µm is
demonstrated in Chen et al. (2004b) in the polymer BCB (Cyclotene 3022-46
from the Dow Chemical Company). An SEM photograph of a fabricated ring
resonator is shown in Fig. 3.55.
Racetrack shaped ring resonators with bending radius of 10 µm and 4 µm
straight coupling sections are also presented. The layer sequence of the device
Fig. 3.55. BCB microracetrack resonator. The filter section has 0.9 µm wide
waveguides, which are tapered to 2.5 µm at the input – output sections.
pp. 470–472 2004
c IEEE
Fig. 3.56. Layer sequence of BCB ring resonator
BCB is cured at a temperature of 250◦ C. SiO2 is used as an etch mask

to transfer the patterns of the waveguides from photoresist to the used BCB.
SiO2 is etched using a trifluoromethane and oxygen plasma, BCB is etched
using a sulfur hexafluoride and oxygen plasma. The etch mask is removed
by buffered hydrofluoric acid. The width of the waveguides is 0.9 µm at the
coupling gap and is tapered to 2.5 µm at the facets.
Laterally coupled organic–inorganic hybrid polymer ring resonator add
drop filters with a radius of 50 µm are presented in Rezzonico et al. (2006). A
photograph of a fabricated ring resonator and a cross-section of a waveguide
are shown in Fig. 3.57.
The waveguides are made out of Ormocore, from the family of
Ormocer R 11
a commercially available inorganic-organic hybrid polymer
(negative photoresist, n = 1.536, 0.6 dB cm−1 at a wavelength of 1.55 µm, Tg >
270◦ C). A two step lithography process is used to realize the laterally coupled
ring resonator add drop filter. The layer sequence is similar to the one shown
in Fig. 3.56 for the BCB ring resonator, substituting Ormocore for BCB. The
SiO2 layer used has a thickness of 4 µm. In a first step, the port waveguides
11
Trademark of the Fraunhofer–Gesellschaft zur Förderung der angewandten
Forschung e.V. München
3.5 Polymers 89
Fig. 3.57. Left: Microscopic front view of a cleaved sample showing the Si substrate,
the SiO barrier, and the cross-section of a port waveguide. The structures are by
about 10% thinner than the films before development since the oxygen at the surface
reduces the polymerization during the exposure. Right: Microscopic top view of a
laterally coupled ring resonator of Ormocore. The external radius of the ring is
50 µm with a cross-section of 3.8 µm width and 4.5 µm height; the port waveguides
are 2.7 µm wide and 3.6 µm high. The gaps at the couplers are 0.3 µm wide.
Reprinted with permission from IEEE Photonics Technology Letters, vol. 18, pp.
865–867 2006
c IEEE
are defined using an exposure in oxygen poor atmosphere through a negative

chrome mask. The sample is developed in methylpentanone. Then the sample
is spin coated again with Ormocore, patterned and developed for realizing the
ring resonator. The gap between the ring and the bus waveguides is aimed
to be 0.5 µm (measured 0.3 µm). The sample is hard baked at 150◦ C in inert
atmosphere. The width and height of the bus waveguides is 2.7 and 3.6 µm,
respectively. The width and height of the ring waveguide is 3.8 and 4.5 µm.
3.5.2 Replication and Nanoimprinting
One of the first ring resonators realized using a nanoimprint technique is

presented in Chao and Guo (2002). Here waveguides made out of PMMA,
polystyrene (PS) and polycarbonate (PC) are demonstrated. SEM photo-
graphs of directly imprinted polycarbonate racetrack and disk resonators are
shown in Fig. 3.58.
Two fabrication methods are demonstrated, direct imprinting (Fig. 3.59)
and template filling (Fig. 3.60).
The fabrication (Fig. 3.59) of the molding tool is performed in two steps.
First, a silicon substrate is etched using RIE, realizing the first molding tool
for imprinting the device pattern into PMMA which is used as a resist for
the following Ti/Ni liftoff process. The remaining PMMA on the surface of
the sample is etched using an oxygen RIE process, thus only leaving the
structured PMMA in place. Ti/Ni which is deposited on to the sample is used
as a hard mask for the following RIE process to etch the SiO2 layer, realizing
Fig. 3.58. Directly imprinted polycarbonate racetrack and disk resonators.

Reused with permission from Chung-yen Chao, Journal of Vacuum Science and
Technology B, 20, 2862 (2002) 2002
Fig. 3.59. Process flow for directly imprinting into a polymer
Fig. 3.60. Process flow for polymer template filling

3.5 Polymers 91
the second and final molding tool for imprinting the waveguides again in
PMMA with a molecular weight of 15,000 (imprinting time = 10 min). In
order to increase the optical confinement of the mode in the waveguides, a
buffered hydrofluoric acid (BHF) etch is performed, resulting in an undercut
of the SiO2 layer underneath the waveguides forming a pedestal. In order to
separate the molding tool from the polymer waveguides, the molding tool is
coated with a surfactant to provide a low surface energy. SEM photographs
of PMMA waveguides sitting on top of a thermally grown SiO2 pedestal are
shown in Fig. 3.61.
In the template filling method (Fig. 3.60), a second layer of SiO2 is grown
by PECVD which is structured using the same processing steps as in the
direct imprinting method for obtaining the molding tool, spin coating PMMA,
imprinting and removal of residual PMMA by dry etching, deposition, and
liftoff of Ti/Ni for the hard mask fabrication, dry etching of the PECVD
SiO2 layer and removal of the metal layer. The molding tool is filled with
polymer. Planarization is done with a flat molding tool. The residual polymer
is removed by RIE and the sample is heated to remove unwanted air bubbles in
the polymer. Finally the sample is again etched with a buffered HF realizing
the pedestal for the polymer waveguides. To reduce the surface roughness
induced scattering loss and therefore increase the performance of the polymer
ring resonator devices, the authors propose a thermal reflow technique in Chao
and Guo (2004). Polymer ring resonator devices are heated with controlled
time duration (60–150 s) in a temperature range of 10–20◦ C below the glass
transition temperature of the polymer. This lowers the viscosity and enhances
the fluidity of the polymer, resulting in a reflow which reduces the surface
roughness significantly under the action of surface tension. The variance of
Fig. 3.61. Cross-section SEM picture of (a) rectangular PMMA waveguide sitting
on top of a SiO2 pedestal. (b) PMMA waveguides in the coupling region of a micro-
ring device.
Reused with permission from Chung-yen Chao, Journal of Vacuum Science and
Technology B, 20, 2862 (2002) 2002
surface roughness is estimated to decrease by 35–40 nm in the ring and by

15–20 nm in straight waveguides.
Replication of polymer ring resonators using a soft lithography method is
demonstrated in Huang et al. (2003). The fabrication process is illustrated in
Fig. 3.62. Ring resonators made out of SU-8 are presented.
The samples are cleaved after molding and curing to achieve optical quality
facets. A photograph of a ring resonator is shown in Fig. 3.63.
This soft lithography method has also been used in Poon et al. (2004a) to
fabricate ring resonator devices without cladding and cladding in polystyrene
(PS) and SU-8, respectively. Here a 3 µm thick layer of OG-125 (n = 1.456,
Fig. 3.62. Replication of polymer ring resonators by soft lithography
Fig. 3.63. Optical microscope image of microring optical resonator fabricated by

soft lithography. The inset shows the detail of coupling region. The ring diameter is
200 µm, the waveguide width is 2 µm, and the gap between the straight waveguide
and the microring is 250 nm.
Reprinted with permission from Y. Huang, Optics Express, vol. 11, no. 20, pp. 2452–
2458 2003
3.5 Polymers 93
a UV curable epoxy) is used as the under cladding instead of growing a SiO2

layer with PECVD as previously done which is UV cured at 80◦ C for 2 min.
The ring resonators without cladding are fabricated by deposition of 10 µl of
PS solution (4 wt% in toluene) on the sample and pressing the mold against
the chip with a force of 25 N for 20 min. The toluene evaporates through the
PDMS mold, leaving behind a 200 nm residual film which does not influence
the characteristic of the ring resonator. Finally, the chip is baked at 80◦ C
for 3–5 min. The SU-8 ring resonator is fabricated in a similar way. The only
difference is that SU-8 is UV cured and the sample is baked at 80◦ C for 3 min.
Another layer of OG-125 with a thickness of 3 µm is spin coated, UV cured and
baked at 80◦ C for 3–5 minutes to form the upper cladding. All samples are
cleaved for device separation. The ring radius is 200 µm and the height and
width of the waveguide is 1.5 and 1.9 µm, respectively.
Polymer racetrack shaped ring resonator notch filters with radii of 100 µm
and 200 µm and a coupling length of 100 µm and 40 µm, respectively, fabri-
cated using an imprinting technique with a smoothing buffer layer are pre-
sented in Kim et al. (2005a). The width of the coupling gap used is 200 nm.
The layer sequence and dimensions of the used waveguide is shown in Fig. 3.64.
The stamp fabrication starts by patterning a quartz glass wafer coated
with chromium using photoresist. After UV exposure, the chromium is etched
and the so achieved hard mask is used to transfer the pattern to the substrate
by RIE. The smoothing buffer layer is made by depositing an SiN film using
LPCVD. The pattern on the stamp is transferred to the cladding of the sample
by general imprinting, where the stamp is pressed against the polymer layer.
The cladding is UV cured and the stamp is removed, leaving behind the trench
for the core waveguide. The core polymer is spin coated onto the sample and
cured. Finally the sample is cleaved and the devices are separated.
Imprinting technologies are an attractive way of realizing ring resonators,
especially due to the fact that the manufactured stamp can be used several
times. This aspect is important, when it comes to producing a large num-
ber of devices which is a requirement for realizing low cost products for the
telecommunication and sensor market. The rise of these fabrication technolo-
gies enables a renaissance of polymer materials and opens up the way for new
materials with appropriate manufacturing properties suitable for replication.
Fig. 3.64. Layer sequence and dimensions of waveguide

3.6 Temperature Insensitivity
Temperature stability of photonic integrated circuits is a key challenge which

applies to all devices, especially resonant based devices like ring resonator
filters. In ring resonator filters, the wavelength to be filtered out for example
has to exactly match the rings resonance wavelength. Temperature shifts the
resonance spectra of ring resonator filters depending on the materials used for
fabricating the device. For stabilizing the resonance wavelength of ring res-
onator filters, a temperature control unit with a heater or a Peltier cooler is
needed. This requires a permanent power consumption of a few watts and
electronic control circuits. The major goal is therefore to eliminate these
additional expenses and sources of defect. For this reason, temperature insen-
sitive or athermal waveguide based devices have been investigated (Kokubun
et al. 1993, Chu et al. 1999e, Keil et al. 2001). The temperature dependence
of photonic integrated circuits (PICs) is mainly caused by the temperature
dependence of the optical path length. In the case of ring resonator add/drop
filters, the center wavelength of the passband is determined only by the opti-
cal path length of the ring resonator. The temperature induced effective index
shift in a ring waveguide cannot be compensated by a similar induced effective
index change of the bus waveguides.
The basic principle of an athermal waveguide is derived in the following
using (Kokubun et al. 1997). The temperature dependence of the center wave-
length of a device can be expressed by

dλ0 1 dS λ0
= · , (3.1)
dT L dT neff
where λ0 is the central wavelength, L is the waveguide length of the device,

T is the absolute temperature and neff is the effective index of the waveguide
(see also (2.4)).
S is the optical path length and is given by
S = neff · L. (3.2)
The temperature dependence of the optical path length of an optical

waveguide on a substrate is given by differentiating (3.2) with T , divid-
ing both sides by L
1 dS dneff 1 dL dneff
= + neff = + neff αth , (3.3)
L dT dT L dT dT
where αth is the coefficient of thermal expansion.
In the case of a waveguide, the coefficient of thermal expansion is the one
of the substrate, since the thickness of the substrate is much larger than the
waveguide layer. Table 3.6 gives the coefficient of thermal expansion for some
bulk materials.
3.6 Temperature Insensitivity 95
Table 3.6. Coefficients of thermal expansion of some materials
material coefficient of thermal expansion αth (×10−6 /◦ C)

BCB 42
MMA 68
Si 2.6
InP 4.6
GaAs 5.8
Au 14
Al 24
Ag 18
Pt 9
Fig. 3.65. Dimensions of athermal waveguide design
In order to achieve an athermal condition, the following equation has to

be satisfied.
dneff 1 dneff
+ neff αth = 0 ⇒ αth = − . (3.4)
dT neff dT
From (3.4), two possibilities exist of designing an athermal waveguide based
device. Either a substrate with a negative value of αth or a waveguide material
with a negative value of dneff /dT can be chosen.
One of the first temperature independent ring resonator filters based on
a strip-loaded athermal waveguide is presented in Kokubun et al. (1997).
Here, a transparent material having a negative value of dneff /dT is cho-
sen for the design of the waveguide. This gives more freedom in designing
the waveguide, as the parameters of used bulk materials obviously cannot
be changed. The structure and dimensions of the athermal waveguide used
are shown in Fig. 3.65. The design makes use of the negative value of dneff /dT
for PMMA which is −105 × 10−6 at λ = 633 nm for bulk material.
The radius of the ring is 5 mm resulting in a free spectral range of 0.037 nm.
The finesse is measured to be 2.6. The athermal design leads to a temperature
coefficient for the central wavelength of 7 × 10−4 nm K−1 .
Another ring resonator filter where the temperature dependence of the cen-
ter wavelength is compensated by application of a polymer overlay is demon-
strated in Chu et al. (1999e). The design of the waveguide is similar to the
one described in Figs. 3.8 and 3.9. The material which is used as the polymer
overlay is a copolymer made by a combination of PMMA and tri-fluoroethyl-

methacrylate (TFMA). The refractive index at room temperature is 1.44 and
the value of dneff /dT is −1.44 × 10−4 /◦ C. The material is deposited onto the
device using spin coating realizing a height of more than 3 µm. The center
wavelength shift due to rising temperature increases linearly without com-
pensation by 0.0137 nm ◦ C−1 from room temperature to 103◦ C. Using the
polymer overlay, this value is decreased in the temperature range between
25◦ C and 55◦ C resulting in a wavelength shift of −0.0025 nm ◦ C−1 . Above
55◦ C this value increases to −0.03 nm ◦ C−1 mainly due to the low glass tran-
sition temperature (Tg ) of the polymer which is around 80◦ C. Polymers with
a higher Tg like polymides are preferred in utilizing this method to realize an
athermal behavior.
A similar method employing a birefringent polymer overlay can also be
used to realize polarization independent ring resonator filters (Kokubun et al.
2001).
A different approach for realizing athermal waveguide based devices is pre-
sented in Keil et al. (2001). Here an all-polymer approach is used. Polymer
materials are used for both the waveguide and the substrate. The substrate is
chosen to have a positive coefficient of thermal expansion (CTE) and the poly-
mer used for the waveguides is chosen to have a negative thermo-optic (TO)
coefficient (dneff /dT ). An athermal (measured between 25◦ C and 65◦ C) and
polarization independent arrayed waveguide grating (AWG) is demonstrated.
The CTE of the substrate is 80 ppm K−1 . The temperature dependent drift
of the center wavelength of the device is ±0.05 nm over the entire tempera-
ture range between 25 and 65◦ C. This approach can readily be adapted for
realizing athermal ring resonator devices.
3.7 Polarization Independence
Handling polarization has been and is an issue when developing photonic

integrated circuits especially for telecommunication systems. A polarization
insensitive ring resonator filter is desirable as different polarization states lead
to undesired shifts in the filter spectrum of the resonator and degrade the
optical performance of the filter.
A well-know method for realizing polarization insensitive devices is by
using a so called polarization diversity architecture. The idea behind this
method is to split the optical input signal into two signals having an orthog-
onal polarization state, like for example one arm is for TE polarized signals,
the other arm is for TM polarized signals. In order to perform integrated
polarization handling of the input signals, polarization splitters, and rota-
tors are needed. Polarization converters using sharp bends are demonstrated
for example in van Dam et al. (1996). An example of a polarization insen-
sitive heterodyne receiver in InP employing polarization rotator and splitter
elements is presented in Kaiser et al. (1996).
3.7 Polarization Independence 97
Fig. 3.66. Layout of a polarization independent filter
A theoretical analysis of a filter comprising two identical ring reso-

nators, two identical polarization converters, straight and curved sections
which is based on the polarization-dependent architecture is presented in
Klunder et al. (2002b). The layout of the filter is sketched in Fig. 3.66.
The ring resonators are assumed to be ideal polarization splitters which
drop the TE polarized light with wavelengths within the pass band completely
and the TM polarized light and the remaining TE polarized light not in the
passband, pass the ring resonator unaffected. The polarization converters are
also assumed to ideally convert from one polarization state into the other
without losses. The calculated polarization dependence of the filter for the
drop port is 0.7 dB, using a polarization converter with a conversion efficiency
of 98%.
A tunable ring resonator filter utilizing a Vernier architecture with polar-
ization diversity is presented in Van et al. (2004). Here a polarization beam
splitter separates TE and TM polarized light into a waveguide which leads
to two third order ring resonator filters with two different FSRs, 575 and
650 GHz. The two waveguide arms are joined using a polarization beam com-
biner after the light passes the ring resonators. A polarization dependent loss
of 0.4 dB and a differential group delay of better than 5 ps are obtained. A
similar architecture is realized in Watts et al. (2005) using integrated polar-
ization splitters and rotators and third-order ring resonators with a diameter
of 16 µm.
Polarization converters with periodical spectral response based on ring
resonators are theoretically proposed and discussed in detail in Morichetti
and Melloni (2006). Several device layouts of parallel coupled and directly
coupled ring resonator configurations are presented.
Polarization independent ring resonator filters are also analytically an-
alyzed in Chin et al. (2004). Two approaches are being considered. In the
first approach, laterally strong confined single mode waveguides with a large
etch depth are analyzed regarding the effective refractive index for TE and
TM polarization depending on the width of the waveguide. The effective in-
dices become equal at the critical width where the mode is circular. Another
issue despite achieving a critical polarization independent waveguide width
is the critical bending radius below which no etch-depth can meet the dif-
ferential loss criterion, as losses for the fundamental modes also become to
high for achieving an acceptable filter performance. In this case, this radius
should not be smaller than 30 µm. The next approach is to realize polarization
independent coupling. MMI couplers (see also Sect. 4.1.2) are proposed, which
can be designed for polarization insensitive operation. Finally in adding pol-
arization independent waveguides and couplers, polarization insensitive ring
resonators can be fabricated.
The idea of polarization independent waveguides and couplers is used in
Headley et al. (2004) for fabricating a racetrack shaped ring resonator notch
filter in SOI. A layout and layer sequence of the waveguide used is shown in
Fig. 3.67.
A coupling length of 500 µm and a bending radius of 400 µm are used in
realizing polarization insensitivity within a shift of the resonance peaks of
2 pm between TE and TM polarized input over a spectral range of four times
the FSR of 192 pm. The devices are covered with a passivating surface oxide
after etching of the waveguides and the coupler, which helps to protect the
waveguides during end facet preparation. The wafers are diced and the end
facets are polished and AR coated in order to minimize Fresnel reflections and
to reduce Fabry–Perot resonances inside the device.
One of the first polarization independent filters using a birefringent fluori-
nated polyimide overlay (birefringence = 0.11) is presented in Kokubun et al.
(2001). Here a vertical coupling scheme is used together with single mode rib
waveguides for achieving polarization independent operation. The geometry
and layer sequence of the used waveguides is shown in Fig. 3.68.
All glass layers are deposited by RF sputtering and patterned by RIE.
The polyimide cladding layer is spin coated onto the sample. The RIE para-
meters are: RF power (30 W), gas pressure (0.67 Pa), flow rate of CF4 and
O2 (10:3 SCCM). An advantage in using a polymer overlay is that remaining
polarization dependence can be eliminated by UV trimming of the polyimide
film, due to the fact that the thickness of the film can be reduced by more
Fig. 3.67. Layer sequence and dimension of used waveguides

3.8 Characterization Methods 99
Fig. 3.68. Layer sequence and waveguide dimensions (the refractive indices are
given for a wavelength of 1.55 µm)
than 20% by UV irradiation while keeping the refractive index and the bire-
fringence constant (see also Sect. 5.2).
Several methods exist for realizing polarization insensitive devices and as
can be seen, the chosen method again depends on the material system and
the device layout for appropriate polarization compensation.
Now that ring resonator theory and device fabrication have been add-
ressed, the focus in the following chapter lies on determining the performance
of the devices, therefore possible characterization methods will be explained
in Sect. 3.8.
3.8 Characterization Methods
The standard characterization of waveguide based devices usually involves

fibers for in and out coupling of the device, a light source and a detector. Due
to the advance in current fabrication methods, analyzing the performance
of ring resonator filters with regard to where the loss occurs is necessary for
improving the manufacturing process and thus improving the characteristic of
the filter. Besides the standard optical fiber characterization setup, microscope
examination, SEM and atomic force microscopy (AFM) measurements, several
methods exist to analyze the performance of waveguide based devices such as
cut-back (destructive), Fabry–Perot (requires single mode lasers) (Park et al.
1995), prism coupling method (requires a minimum length of the waveguide of
several cm), optimum end-fire coupling (OEC) method (Haruna et al. 1992)
and many others. The description of some of the mentioned characterization
methods will be done in the following paragraphs.
The main goal of each method is to determine the losses of the device. In
order to compare the loss parameters, a definition for the standard parameters
is given in the following paragraph.
The insertion loss αinsertion of the device is the minimum transmission for
a specific wavelength range for all polarization states. It represents the worst
possible loss through the device. The insertion loss uniformity of a device
is the difference between the insertion loss of the best-case and worst-case
channels. The insertion loss αinsertion is defined for an input intensity Iin and
output intensity Iout as
Iout
αinsertion = −10 · log . (3.5)
Iin
The total insertion losses αinsertion includes the intrinsic losses αpropagation and
the fiber–chip coupling losses αcoupling

αinsertion = αpropagation + αcoupling dB cm−1 . (3.6)
The intrinsic losses can be described by the following equation:

1 2
1 Iout
αpropagation = − 10 · log + αcoupling dB cm−1 , (3.7)
L Iin
where L is the total length of the measured waveguide.
3.8.1 Conventional Characterization
One of the standard setups for the measurement of ring resonator devices is
shown in Fig. 3.69. Here an external cavity laser (ECL) is used as the light
source which is connected via a polarization controller (PLC) and a tapered
or lensed fiber to the input waveguide.
The transmitted signal is detected using for example the lock-in technique
employing a photo diode and a lock-in amplifier. The ECL signal is coupled
to the input waveguide by using for example a tapered fiber, which can be
adjusted by a three axis piezo drive. Tapered fibers are usually used to couple
into strong guiding waveguides and reduce the fiber to chip coupling loss
considerably compared to a standard butt fiber due to the reduction of the
mode field diameter which suits the high index contrast waveguides better
than the large field diameter of the butt fiber. The near field of the output
Fig. 3.69. Example of a standard optical measurement setup

waveguide can then be focused on a photodiode (PD) by using a microscope

lens which has a sufficient aperture to guarantee correct power measurement.
The device under test (DUT) is placed on a Peltier – cooler/heater so that
all measurements are performed at a definite temperature. Instead of using
a microscope lens on the output side, a tapered fiber can again be used to
couple light from the drop or throughput port which is fed to an optical power
meter or an optical spectrum analyzer.
This setup can now be used to extract various device parameters including
the filter characteristic. The coupling losses can for example be calculated from
measuring the loss of different waveguide lengths of the same device. This
internationally recognized reference test method is known as the cut-back
technique. The measured insertion loss is plotted against different lengths of
the measured devices. The slope of the obtained curve is determined by the
propagation losses of the waveguide. The coupling losses are obtained from
the point, where the extrapolated measurement curve crosses the Y -axis at
a device length of zero (unit length). The cut-back method is a destructive
method to obtain the propagation and coupling losses of straight waveguides.
A nondestructive method which is commonly used is the Fabry–Perot
method. This method has first been introduced in 1978 by Kaminow and
Stulz. The chip is regarded as a Fabry–Perot resonator for waveguide losses
<1 dB cm−1 , where the facets of the chip serve as the mirrors of the resonator.
The optical wave is reflected back and forth within the waveguide. The result-
ing “filter” spectrum depends on the intrinsic losses and the reflection factor
of the facets. The optical length is changed by varying the temperature of the
whole chip or the wavelength exploiting the group velocity dispersion. The
following calculations closely follow (Regener and Sohler 1985, Feuchter and
Thirstrup 1994). The transmission spectrum of a Fabry–Perot resonator is
given by
IT T 2 e−αinsertion L
= 2 , (3.8)
I0 η φ
1 − R̃ + 4R̃ sin2
2
R̃ = R · e−αinsertion L .
The equation for describing the Fabry–Perot resonator is very similar to

the one describing a single ring resonator notch filter. Here IT is the trans-
mitted intensity, I0 is the intensity of the input laser signal, η is the coupling
efficiency to the fundamental waveguide mode, φ = 2βL is the internal phase
difference, R is the end-face reflectivity, L the length of the resonator and
αinsertion the insertion losses of the waveguide. The contrast K of the Fabry–
Perot resonances, which is independent of I0 and η, is given by
Imax − Imin
K= . (3.9)
Imax + Imin
The transmitted maximum and minimum intensities can be described by
(1 − R) · e−αinsertion L
2
Imin = 2 (3.10)
(1 + R · e−αinsertion L )
(1 − R) · e−αinsertion L
2
Imax = 2. (3.11)
(1 − R · e−αinsertion L )
From (3.10) and (3.11) K can be rewritten as
2R̃
K= . (3.12)
1 + R̃2
This leads to
1
R̃ = 1 − 1 − K2 . (3.13)
K
The value of K is taken from the measurement of the Fabry–Perot
resonances. The maximum value of K is obtained for lossless waveguides.
Increasing losses on the other hand reduces the contrast.
If the end-face reflectivity is known, the insertion loss of the device is
given by
4.34 3 4
αinsertion = − ln R̃ + ln R dB cm−1 . (3.14)
L
Equation (3.14) can also be written using Imin and Imax directly
1 √ 2
1 1+ u −1
αinsertion = ln √ + ln R cm (3.15)
L 1− u
Imin
u= .
Imax
When the reflection factor of the chip is unknown, it can be calculated from the
insertion loss measurements plotted for different device lengths using (3.14)
or (3.15). The measurement is approximated by a straight line from which
the reflection factor can directly be taken. The slope of the curve is again the
value for the intrinsic losses αpropagation of the waveguide, which is given by
√
1+ u
4.34 · ln √ or − 4.34 · ln R̃ dB cm−1 . (3.16)
1− u
Another way of calculating the reflectivity of the end faces, which is valid for
weak guiding waveguide structures, is by using:
2
neff − 1
R= . (3.17)
neff + 1
When the reflection factor is known, the intrinsic losses of the chip are
easily determined without the necessity to cut the waveguide several times for
characterization.
Fig. 3.70. Basic OLCR measurement setup in reflection mode
3.8.2 Optical Low Coherence Reflectometry
In optical low-coherence reflectometry (OLCR), the output of a low-coherence

source is coupled into an optical fiber and split in a 3-dB coupler as sketched
in Fig. 3.70. It is basically a white light Michelson’s interferometer.
Half of the input signal travels to the DUT, while the other half is launched
into free space towards a mirror on a scanning translation stage. When the
optical path length from the coupler to the mirror equals the optical path
length from the coupler to a reflection in the DUT, the signals from the two
arms add coherently to produce an interference pattern which is measured
in the detector arm of the coupler. When the optical path length difference
becomes larger than the coherence length of the source, this interference signal
fades and is lost eventually. Therefore the spatial resolution critically depends
on the coherence length of the light source. Reflections with a spatial res-
olution better than 10 µm can be localized. Refractive index discontinuities
less than 10−4 (dynamic range of ≈−80 dB) in the device can be detected.
The amplitude of the interference signal is proportional to the magnitude of
the reflection from the DUT. Movement of the mirror allows the reflectiv-
ity of the DUT to be mapped as a function of distance, which allows the
localization of waveguide discontinuities and imperfections. To achieve pre-
cise control of the fiber chip coupling and to record the transmission data,
the output end of the setup can be coupled to an infrared-sensitive camera
and an optical spectrum analyzer (OSA), respectively. This setup is referred
to as the reflection mode. The reflection mode setup is used to characterize
for example an active–passive butt joint, but can also be used to measure ring
resonator devices without AR coated facets. Due to the AR coating of the
facets, which is applied to most devices incorporating ring resonator filters, a
setup called the transmission mode (Fig. 3.71) is used for the characterization
of ring resonator devices.
A detailed analysis of ring resonator filters using OLCR is presented
in Gottesman et al. (2004). An example of a so called reflectogram of a
ring resonator notch filter without AR coating with a codirectional coupler
(length = 150 µm, coupling gap = 0.9 µm) and radius r = 100 µm is shown
Fig. 3.71. OLCR measurement setup in transmission mode
Fig. 3.72. Reflectogram of ring resonator with r = 100 µm without AR, measured
in reflection mode
in Fig. 3.72. The first peak is measured at the input facet of the device. The
second peak (out1 ) is measured at the output facet. There are no signifi-
cant reflection peaks in between those peaks coming from the ring resonator,
which shows that the input signal has passed the resonator undisturbed. The
peaks in between peak out1 , peak out2 , and peak out3 result from multiple
roundtrips in the ring resonator. The distance between the peaks out1–2 is
2 mm which corresponds to the chip length. The distance between two peaks
coming from the ring resonator is equal to half of the cavity length which
is measured to be ≈466 µm (physical value 464 µm).
A reflectogram of a ring resonator notch filter measured in transmission
mode is shown in Fig. 3.73.
The first peak is the transmission passing the coupler without entering the
ring. The second peak is the transmission from one roundtrip in the ring. The
third peak is the transmission from two roundtrips and so on. The roundtrip
loss is evaluated from the transmission data beginning from the second peak.
Fig. 3.73. Reflectogram of ring resonator with r = 100 µm without AR, measured
in transmission mode
Fig. 3.74. Reflectogram of a straight waveguide with integrated SOA
In this calculation the loss in the straight waveguide is not considered and is
contained in the coupling factor and the roundtrip loss. The length between
the peaks is equal to half of the optical length of the ring resonator. The aver-
age optical length (in the ring) is taken from the measurement (e.g., distance
between peak 2 and 3) and is equal to 6171.86 µm. The measured ring has a
resonator length of 1776.64 µm. The group index is calculated from this data
to be ngr = 3.47. This is the average group index which the traveling light
wave experiences when it passes the coupling region, the curved sections and
the straight section in the resonator.
The active–passive transition can also be analyzed using the OLCR mea-
surement. The result for a straight waveguide with a 500 µm long SOA section,
measured in reflection mode is shown in Fig. 3.74.
The facets are not AR coated which is shown in the high reflection peaks
at the input and output facets. The return loss from the first butt joint is
Fig. 3.75. Reflectogram of a ring resonator with integrated SOA
about −32 dB. The length of the SOA section can also be taken from the
OLCR measurement and is proven to be 500 µm.
The reflectogram of a single ring resonator with SOA (see inset) is shown
in Fig. 3.75.
The ring resonator has a radius of r = 200 µm, a coupler length of 150 µm
(coupling gap = 1 µm) and a gain length of 100 µm. The return loss resulting
from the butt joint is only −50 dB which indicates a lower reflection than
from the straight waveguide with SOA. This is due to the bending loss in the
ring resonator, where part of the reflected intensity from the butt joint is lost,
leading to a lower value for the return loss. There are again no reflection peaks
due to imperfections in the resonator or straight waveguide which indicates a
high quality cavity.
The OLCR method shows only a qualitative indication of the reflection
coefficient, because OLCR takes into account the unknown absorption during
propagation. A technique to evaluate the coupling efficiencies and reflection
coefficients simultaneously at the butt-joint interface is presented in Song et al.
(2005). The technique described here is based on the Hakki–Paoli method
(Hakki and Paoli 1975), which is widely used to extract the gain spectra
of semiconductor lasers and the reflectivity of the AR coating by measuring
amplified spontaneous emission (ASE) spectra.
In summary, the OLCR measurement method enables an insight study
of the quality of the ring cavity and has the potential to efficiently extract
all necessary ring parameters to describe the spectral behavior of the ring
resonator configuration.
3.8.3 Evanescent Field Measurement Methods
Near field techniques have become attractive for optical characterization.

Probing the near field enables to overcome the diffraction limit of conventional
microscopy. Spatial resolutions much smaller than the optical wavelength can
be obtained, approaching λ/50.
Near-Field Scanning Optical Microscopy
Near-field scanning optical microscopy (NSOM) or scanning near-field optical

microscopy (SNOM) which is used as an alternative name is another scan-
ning probe microscopy (SPM) related technique. An optic fiber is stretched
and thinned such that it has a very sharp point. The fiber is coated with
metal such that a small aperture (approximately 25 nm diameter) is formed
at the tip of the fiber. This aperture serves to illuminate a small spot on the
sample which is much smaller than the conventional diffraction limit. The
sample is then scanned beneath the tip and the image is formed in the same
fashion that a dot matrix printer prints a picture. The tip is brought very
close to the surface. The tip–surface distance can be measured via the force
acting between the two. The tip is scanned across the surface as in scanning
tunneling microscopy (STM). NSOM (or SNOM), can detect objects below
the diffraction limit (down to the nm regime) and thus represents a power-
ful new technique in optical microscopy. The significance of NSOM is that it
allows spatial resolution with more than an order of magnitude improvement
over the best conventional optical methods, including laser scanning confocal
microscopy.
Although optical characterization is the most widespread method to
analyze materials from biology to the semiconductor industry, it suffers from
one inherent problem: the diffraction limit provides a spatial resolution limit
of about half of the wavelength of light. Thus, features smaller than 250 nm
can not be imaged or spectrally characterized with visible light. NSOM com-
bines scanning probe microscopy instrumentation with optical microscopy and
spectroscopy to provide optical characterization with, in some cases, 15 nm
resolution in using visible light.
One of the first experiments using NSOM to measure the internal optical
modes inside ring resonators with a radius of 25 µm is presented in Vander
Rhodes et al. (2000). The ring resonators which are analyzed are described
in detail in Chu et al. (1999c) and have also been described in Sect. 3.3.2.
The setup used in the NSOM experiment consists of a standard optical mea-
surement setup as described in Sect. 3.8.1 using a lensed fiber for coupling
into the ring resonator, an ECL and a photodiode at the output port and in
addition the NSOM probe, which is fed into an InGaAs detector to collect
the light from the sample. A multimode fiber optic is used at the output facet
of the ring resonator chip which enables the simultaneous monitoring of both
the through and drop port of the ring resonator when performing the NSOM
scans. The NSOM probe is scanned over the surface of the waveguide with a
tip–surface separation of 10 nm using an interferometrically calibrated trans-
lation stage to scan the tip with an accuracy of ±0.3 nm. Results of NSOM
scans of a ring resonator are shown in Fig. 3.76.
(a) (b)
Fig. 3.76. Multiple NSOM scans of the ring resonator, stitched together for an
image of almost the entire ring on resonance. Entire scan is 58 µm square: (a) shows
the optical modes, clearly showing the light coming into the resonator in the upper
right, with about half of that light leaving the resonator at the drop port in the
upper left. The launch orientation used here is not especially efficient in coupling to
the drop port, which is why very little light is seen at the drop port in the upper
right and (b) is topography, with a height difference of 1.45 µm from black to white.
Reprinted with permission from IEEE Journal of Selected Topics in Quantum
Electronics, vol. 6, pp. 46–53 2000
c IEEE
Photon Scanning Tunneling Microscope
Photon Scanning Tunneling Microscope (PSTM) is another configuration of

a scanning near field system described before. A PSTM is a device where the
object is illuminated by an evanescent wave generated at the face of a prism
or slide and the field is detected via a fiber probe in the near-zone of the
sample (as in NSOM). PSTM measurements are inherently holographic since
the sample is illuminated by an evanescent field and that same field serves as
reference.
The PSTM tips are obtained in a similar way as NSOM probes, by pulling
an optical fiber, coating it with about 7 nm of Cr, which results in no aperture
at its very extremity. The PSTM tip scans at constant height. In this mode
of operation, the signal detected by a PSTM is related to the spatial distri-
bution of the intensity of the optical electric field in the near-field zone. Ring
resonators have been analyzed using photon scanning tunneling microscopy
in Quidant et al. (2004). A photograph of a fabricated ring resonator is shown
in Fig. 3.77.
The used ring resonators are made of a 150 nm thick TiO2 layer (residual
roughness of 3 nm) coated on a glass substrate with a residual roughness
Fig. 3.77. Scanning electron microscopy picture of a TiO2 ring coupled with a
linear waveguide ridge. (Photograph courtesy of Dr. Romain Quidant12 )
0.6
4
2
Intensity (arb. units)

0.4
x (µm)
0
0.2
−2
−4
−4 −2 0 2 4
y (µm)
Fig. 3.78. PSTM image recorded above the microring. The probe tip scans in
plane parallel to the sample located at 50 nm from the TiO2 structures. (Photograph
courtesy of Dr. Romain Quidant)
of 1 nm by ion assisted deposition. The waveguides are fabricated using RIE.

The dimensions of the ring resonators are 6 µm in diameter, 250 nm wide and
150 nm high.
12
ICFO – Institut de Ciències Fotòniques, Parc Mediterrani de la Tecnologia, Av.
del Canal Olı́mpic s/n, 08860 Castelldefels, (Barcelona), Spain
In this PSTM experiment, an He–Ne laser (633 nm) injected in a lensed

single mode fiber is used to illuminate the ring resonator. The end of the fiber
is oriented in such a way that the beam can be focused at the interface between
the dielectric ridge and the glass substrate. Only TM polarized incident light
is being considered. The Gaussian beam is focused at the right end of the
straight bus waveguide in such a way that the components of the incident
wave vector parallel to the surface of the substrate align along the longitudinal
axis of the bus waveguide. The tip is scanned at a constant height above the
sample surface while monitoring the light intensity level. A photograph of the
resulting image is shown in Fig. 3.78.
Several state-of-the-art characterization techniques have been described.
Due to the decreasing dimensions of ring and disk resonator devices and the
increase in integration density, a combination of several characterization meth-
ods is inevitable in the future to guarantee optimum device performance.
The following chapter deals with the building blocks of ring resonator
devices which include bends and couplers.
4
Building Blocks of Ring Resonator Devices
Ring resonator devices comprise a bus waveguide and a ring which is made out
of bend waveguides in the basic configuration. The material used for realizing
the ring resonator filter already imposes restrictions onto the waveguide type
and thus on the performance of the ring resonator. It is therefore important
to analyze the building blocks of the envisaged ring resonator with respect to
the wanted performance of the filter in advance once the material has been
chosen.
The part where the bus and the ring come close to one another can be
regarded as a coupler, also in the basic configuration, where the interaction
path (coupler length) of the optical field in ring and coupler is extremely
short. The performance of the ring resonator depends on the power coupling
factor as one parameter besides several other factors already discussed in pre-
vious chapters. Several coupling schemes can be applied including vertical
couplers, directional couplers, MMI couplers and Y-couplers. A detailed sum-
mary of the simulation and design of directional couplers and MMI couplers
also with respect to polarization dependence is presented in Darmawan et al.
(2005).
In order to fabricate ring resonators with low loss, it is essential to know the
bending loss of the waveguides which depends directly on the used bending
radius, which supplies the FSR of the ring resonator filter if this type of
waveguide is used for fabricating the ring.
A fact which should be taken into account in the designing stage is the
coupling of the bus waveguides to the outside world. Here several concepts
exist like tapering the waveguides in vertical, horizontal (or both) directions
to match the mode field profile of the bus waveguides to an optical fiber.
This chapter provides an introduction into the various building blocks
making up a ring resonator filter.
112 4 Building Blocks of Ring Resonator Devices
4.1 Couplers
4.1.1 Directional Couplers
An important device used to couple light into and from ring resonators is the
directional coupler. This type of coupler is preferably used to realize split-
ting ratios other than 3 dB. One of the first papers to deal with directional
integrated optical couplers is presented in Marcatili (1969b).
Directional couplers have been studied extensively in the past and have
found their way into many integrated optical devices like for example opti-
cal switches (Forber and Marom 1986) and vertical coupler beam splitters
(Raburn et al. 2000).
The well known equations which describe the input and output field of the
coupler shown in Fig. 4.1 are given by
√ √
E3 1−κ √ −j κ E
√
1
= (1 − γ) 2 · 1 . (4.1)
E4 −j κ 1 − κ E2
The relation between the in- and output intensity of a symmetrical directional
coupler is given using

I3 1−κ κ I
= (1 − γ) · 1 , (4.2)
I4 κ 1−κ I2
where γ is the power loss coefficient, κ is the power coupling factor. If both
waveguides have the same propagation constant which is the case in this
configuration and the light is inserted into input port 1, then after a length
Lc the energy will have coupled into the other waveguide and can be detected
at output port 4.
The coupling behavior dependent on Lc for light inserted at input port 1
is expressed by

1 π
E3 (x) = E1 (1 − γ) 2 cos x , (4.3)
2Lc

1 π
E4 (x) = jE1 (1 − γ) sin
2
x , (4.4)
2Lc
I3 (x) = E3 (x) · E3 (x) and I4 (x) = E4 (x) · E4 (x). (4.5)
Fig. 4.1. The directional coupler

4.1 Couplers 113
Fig. 4.2. Coupling behavior of a symmetrical directional coupler
The coupling behavior for a coupler with Lc = 450 µm, E1 = 1, E2 = 0 is

shown in Fig. 4.2. The intensity at both output ports is normalized to the
sum of the output intensities
I3
Intensity at output port 3 = (4.6)
I3 + I4
I4
Intensity at output port 4 = (4.7)
I3 + I4
The power coupling factor κ is taken from the diagram (Fig. 4.2) for a specific
coupling length x.
For example, the power coupling factor κ is equal to κ = 0.9 for a coupler
with a length of 360 µm.
The separation between the waveguides in the directional coupler is the
critical element regarding the fabrication. The resolution of the photolithogra-
phy used (standard photolithography using a quartz/chromium mask, electron
beam lithography, etc.) defines the minimum coupling gap. The fabrication
of the gap depends mainly on the waveguide width and the etch depth in
the gap. An example of a fabricated coupler is shown in Fig. 4.3. The etch
depth in the gap is lower in this example than on the outer side of the coupler
waveguides, which is due to the lower etch rate in the gap. The dry etching
process is strongly dependent on the etch gases and conditions used and has
to be modified if performed with other RIE systems to achieve similar results.
Integrated optical devices require directional couplers to provide the
desired power splitting ratio, independent of wavelength or polarization. In
wavelength division multiplexing systems, where ring resonator add drop fil-
ters are deployed, it is necessary to achieve a constant splitting ratio for
all used wavelength channels. It is often impossible to achieve broadband
polarization-insensitive performance, especially in planar integrated devices,
unless several design consideration are incorporated and tradeoffs taken into
account as is explained in Sect. 3.7. A wavelength- and polarization-insensitive
directional coupler configuration is demonstrated in Little and Murphy (1997)
Fig. 4.3. SEM photograph of the input region of a directional coupler, gap = 0.8 µm
Fig. 4.4. Top view of an MMI coupler
and Murphy et al. (1999). Design rules towards polarization insensitive

directional couplers employed in ring resonator filters are discussed and
presented in Chin (2003).
4.1.2 Multimode Interference Couplers
A multimode interference (MMI) coupler consists of a broad center waveguide

which supports several modes depending on the width and the layer sequence
of the waveguide (Fig. 4.4). The self-imaging principle is used in MMI couplers
for realizing various coupler configurations including a 2×2 coupler to be used
in ring resonator devices. Self-imaging is a property of multimode waveguides
by which an input field profile is reproduced in single or multiple images at
periodic intervals along the propagation direction of the guide (Soldano and
Pennings 1995).
Between these intervals, the input field is reproduced with only a fraction
of the input intensity, divided symmetrically in the multimode section. A SEM
photograph of the input region of a 2 × 2 MMI coupler is shown in Fig. 4.5.
An atomic force microscope (AFM) photograph of the input region of an
MMI coupler is shown in Fig. 4.6 (Rabus 2002). The gap between the two
input waveguides is 2.4 µm.
The main advantage of MMI couplers compared to directional couplers
is the fabrication tolerance with respect to the splitting ratio which can be
for example 3-dB in the case of 2 × 2 couplers used in ring resonators. If
other coupling ratios are required, directional couplers are favored. The mul-
timode section should support more than three modes to assure an appropriate
4.1 Couplers 115
Fig. 4.5. SEM photograph of the input region of a 2 × 2 coupler
Fig. 4.6. AFM picture of the input region of an MMI coupler (Rabus 2002)
interference signal of the input mode. Finite difference time domain or beam
propagation tools (Weinert and Agrawal 1995) are needed to calculate the
behavior of MMI couplers which are commercially available.
The fundamental beat length of MMI is related to the waveguide parame-
ters (step-index MMI section) by Soldano and Pennings (1995):
π ∼ 4neff · w2
Lπ = = , (4.8)
β0 − β1 3λ0
where β0 and β1 are the propagation constants of the fundamental and first
order lateral modes, respectively, λ0 is the wavelength, neff is the effective
index and w is the width of the multimode section.
The 3 dB length of MMI couplers can also be approximated by the formula
given in van Roijen et al. (1994):
3 λ0 neff · w2
L3dB = ≈2 , (4.9)
4 neff0 − neff1 λ0
where λ0 is the wavelength, neff0 and neff1 are the effective indices of the
fundamental and first order mode.
The usage of short couplers with defined splitting ratio is essential in ring
resonators for achieving a high FSR due to a short roundtrip length. One of
the first fabricated MMI couplers on InP having the shortest length at that
time of only 107 µm for a 3 dB coupler are demonstrated in Spiekman et al.
(1994). The width of the MMI is only 9 µm. A ridge waveguide design is used
with an InGaAsP core with a bandgap wavelength of 1.3 µm and a thickness
of 600 nm. An InP cap having a thickness of 300 nm is grown on top.
Extremely short InP–InGaAsP MMI 3-dB couplers with lengths between
15–50 µm are presented in Ma et al. (2000b).
Ring resonators ideally require a tunable coupler in order to adjust the
coupling factor to the roundtrip loss in the case of passive resonators and
to realize specific filter characteristics in the case of multiple coupled ring
resonators. MMI couplers with a tunable splitting ratio are demonstrated
in Nagai et al. (1999); Leuthold and Joyner (2000, 2001) and Jiang et al.
(2005). MMI couplers are usually fabricated in semiconductor materials which
enable the realization of strong guiding waveguides. Recently MMI couplers
are proposed using weak guiding structures (West and Honkanen 2005) and
fabricated in polymer materials (Mule’ et al. 2004, Rabus et al. 2005b).
4.1.3 Y-Couplers
Y-couplers are a well-known class of integrated optical devices and have been
used in several integrated systems to guide and manipulate light. Quite a lot
of research has been performed on digital optical switches (DOS) in polymers
using cascaded Y-couplers to create integrated 1×N splitters (Bernhard 2002,
Hauffe 2002). An example of a Y-coupler is shown in the schematic in Fig. 4.7.
This example of a Y-coupler comprises single mode waveguides for in- and
out coupling with width w, a taper and a coupling region. The taper region is
designed in such a way, that the width at the end is slightly less than 2 times
the width of a single mode waveguide. This assures, that only the fundamental
mode is allowed to propagate in the beginning of the taper and in the end
into the two output waveguides and not the next higher mode which would
cause additional losses, as it is not supported by the geometry of the single
mode waveguides. In the coupling region, both output waveguides have to
4.1 Couplers 117
Fig. 4.7. Schematic of a symmetric Y-coupler
Fig. 4.8. Polymer Y-coupler
be treated using coupled mode theory to describe their behavior. This region
is of importance when it comes to design a symmetric Y-coupler with equal
power splitting in both output arms. If the angle α is chosen too small, power
from one arm is able to couple to the other one being the cause for a different
splitting ratio than intended. On the other hand making the angle too large
causes additional losses. Therefore, a tradeoff has to be found for realizing
a specific Y-coupler. Usually S-bends are used to separate the output arms
and realize a certain distance between them. Due to manufacturing tolerances
like proximity effect and etching, a sharp branching angle is not achieved in
most cases and a compromise has to be made to find the optimum angle.
An example of a fabricated Y-coupler in polymers is shown in Fig. 4.8 (Rabus
et al. 2005b).
Y-couplers have found their way into ring resonators for example in ring
lasers, where they have been specially designed to increase the output power.
An example of an improved Y-junction splitter used in ring resonator lasers
can be found in Burton et al. (1994). The design uses only rectangular sections
which make it similar to an MMI coupler avoiding sharp tips at waveguide
intersections, which are a challenge to fabricate. For more details on ring res-
onator lasers, see Sect. 5.6. Due to the fact that Y-couplers can be regarded
as standard optical components which can be simulated with several commer-
cially available BPM and FDTD software, further treatment of these devices
will not be given.
4.2 Bends
Bends are inevitable building blocks of state-of-the-art integrated optical cir-
cuits and have therefore been always of interest in the optics community.
One of the first papers dealing with the analysis of bends is described in
Marcatili (1969a). Since then, many methods have been proposed and devel-
oped to describe the behavior of the mode field and lower the losses in bends
for example, Neumann and Richter (1983), Bienstman et al. (2002), Melloni
et al. (2001).
One method of describing a waveguide bend is to use a conformal trans-
formation which is presented in Heiblum and Harris (1975). The curved
waveguide is translated into an equivalent straight waveguide with a trans-
formed index profile.
By using a cylindrical coordinate system (x, ρ, φ), propagation of light in
circularly curved waveguides can be described in terms of modal propagation
(Fig. 4.9). A mode in such a structure is given by Smit et al. (1993):
ECirc = Uγ (r) e±γφ φ . (4.10)
This mode has as an equi-phase front which propagates in the angular

direction φ. γφ is regarded as a complex angular propagation constant and is
described as:
γφ = αφ + jβφ , (4.11)
where αφ is the attenuation coefficient, this time in angular direction, and βφ
is the real angular propagation constant in rad−1 .
Using a conformal mapping technique (see also Chin and Ho 1998), the
problem of solving the wave equation in cylindrical coordinates can be trans-
formed to rectangular coordinates by performing the following steps (Heiblum
and Harris (1975), Smit et al. (1993)):

r
u = Rt ln (4.12)
Rt
Fig. 4.9. Curved waveguide transferred to a straight waveguide using conformal

transformation. Curved waveguide with real index profile (left). Equivalent straight
waveguide with transformed index profile (right)
4.2 Bends 119
with Rt a freely selectable reference radius. Equation (4.10) can then be

written as 2
∂ 2 2
+ k n t (u) − γt
2
Ut (u) = 0, (4.13)
∂u2
where
u
nt (u) = n {r (u)} e Rt (4.14)
γφ
γt = (4.15)
Rt
u
r (u) = Rt e Rt (4.16)
The angular propagation and attenuation constants can now be derived

using (4.15) as follows:
βφ = βt Rt , (4.17)
αφ = αt Rt . (4.18)
The radiation loss can then be expressed by

1 −α π 2
φ
Aφ = −20 log10 e 2 = 10παφ log10 e. (4.19)
With this transformation an ordinary straight waveguide problem is

obtained, which can be solved with the effective index method by using a
staircase approximation for the transformed index profile in the u-direction
(Fig. 4.9 right). The index profile n(r) is transformed into an index profile
nt (u). The shape of the transformed profile can be qualitatively understood
by considering that the longer path length of the light travelling in the out-
side bend has to turn up in the straight waveguide problem as a slower phase
velocity, hence the higher index there. The center of gravity of the mode profile
in a curved waveguide moves outward, an effect that increases by decreasing
the radius of the bend.
An approximation can be made when Rt is chosen to be equal to the outer
edge of the waveguide and u/Rt 1 in the vicinity of the waveguide, then
r∼
= Rt + u, (4.20)
which leads to a transformed index profile

u
nt (u) ∼
= n (R t + u) 1 + . (4.21)
Rt
A simple method for estimating bend radii of deeply etched rib waveguides
(Fig. 4.9) is presented in Pennings et al. (1991). This method can be used for
rib waveguides, where the rib is etched completely through the guiding layer.
The effective index method (EIM) is used in this reference to describe the
rib waveguide by a one dimensional refractive index profile with an effective

index nrib under the rib and an effective index of 1 next to the rib (air). This
one dimensional refractive index profile can now be treated as a straight slab
waveguide with a propagation constant of
β = knstraight . (4.22)
All modes with a refractive index nstraigth smaller than the refractive index of
the substrate nsubstrate are treated as below the cut of wavelength and are lost
into the substrate. The straight waveguide problem is solved again by applying
the conformal transformation method as discussed before with the refractive
index of the bend nbend related to the angular propagation constant γ by
Re (γφ ) αφ
nbend = = . (4.23)
kR kR
The imaginary part of the angular propagation has been neglected due to the
high index contrast between the waveguide and the surrounding air, which
leads to a strong confinement of the mode in the waveguide and results in low
radiation losses in the bend. The radius R is defined as being the length up
to the outer side of the waveguide bend. Using the cutoff condition R = Rco
when nbend = nsubstrate and the approximation
1 w
nstraight − nbend = nrib (4.24)
2 R
leads to the equation for the cutoff radius, which depends only on parameters
of the straight waveguide
w · nrib
Rco = . (4.25)
2 (nstraight − nbend )
The main cause for losses in bends is of course the bending losses them-
selves (4.19). One way to minimize the bending loss is by increasing the
radius. This is of course limited when designing ring resonators with certain
parameters and a specified filter performance. Therefore one way to reduce the
bending loss is by changing the refractive index of the material on the outer
side of the waveguide. This can also mean the complete removal of the material
on the outer side of the waveguide and forming a trench. An analysis of this
idea is presented in (Neumann 1982b, Seo and Chen 1996). Schematics of
proposed waveguide geometries are shown in Fig. 4.10. Seo et al. not only sug-
gested to use trenches on the outer side of the curved waveguide to reduce
the bending losses, but analyzed also the slope of the rib with regard to the
transition losses which can be lowered this way.
Another type of loss which is especially important when it comes to the
design of racetrack shaped ring resonators is the transition loss which occurs
at the interface between a straight waveguide and a bend. One of the first
publications dealing with the reduction of these transition losses is (Neumann
4.2 Bends 121
Fig. 4.10. Architectures of waveguide bends
Fig. 4.11. Schematic of a racetrack shaped ring resonator with offsets between the
curved and straight waveguide junctions
1982a). As described earlier, the center of the mode field in the curvature
shifts with increasing radius to the outer side of the waveguide in the bend.
This distance d of the mode shift in a waveguide having a width w1 can be
described by the following equation:
π 2 n2eff w
d= . (4.26)
λ2 R
One method of reducing the transition losses is therefore to apply an offset
between the curved and the straight waveguide sections which is equal to a
distance d. A schematic of a racetrack shaped ring resonator employing offsets
at the respective interfaces is shown in Fig. 4.11. A fabricated racetrack shaped
ring resonator in the material system GaAs–AlGaAs (Fig. 4.12) employing
such offsets is presented in Van et al. (2001). Here, fabricated racetrack shaped
ring resonators with offsets up to 0.08 µm are demonstrated. The parameters
for the racetrack shaped ring resonators are a waveguide width of 0.52 µm,
a coupling gap of 0.2 µm, an average bending radius of 4.7 µm, and straight-
section length of 10.24 µm.
1
Note that w can either be the actual width of the waveguide or the width of the
mode, depending on the type of waveguide used.
Fig. 4.12. Racetrack shaped ring resonator with lateral offsets between curved and
straight waveguide sections.
pp. 1734–1739 2004
c IEEE
The transmission from a straight to a curved waveguide for the case of rib
channel waveguides is given by Subramaniam et al. (1997)
2
2
−d2 (k · n · w) d
e 2w 1 +
2
2R
TSC (d) = 4
, (4.27)
(k · n · w) w2
1+
2R2
where w represents the mode width and n the effective refractive index of the
rib in the center of the waveguide. Minimum transition loss is achieved when:
(a) The amplitude distributions of the two modes are matched
(b) The phase distributions are matched
Condition (b) is satisfied, condition (a) is a cause for the transition losses
which can be evaluated using (4.27) or can be approximated by Neumann
(1982a)
2
d
αSC -losses = 4.343 (dB)
w
where w is either the width of the waveguide or the width of the mode.
These transition losses can be lowered by choosing the appropriate offset,
trench or different material with a lower refractive index on the outer side of
the waveguide in the bend and in addition selecting the right widths of both
the straight and the bend waveguides (Smit et al. 1993).
Bends are not only used as “connecting” elements in integrated optics, but
can also serve themselves as polarization rotators. An analysis of polarization
rotation in bends is presented in Lui et al. (1998), Obayya et al. (2001).
A compact polarization converter based on bends is demonstrated in van
Dam et al. (1996). Polarization rotation can be of importance when designing
ring resonator filters and depends on several influencing factors like type of
material used, waveguide geometry and ring radius.
4.3 Spot Size Converters for Light In- and Outcoupling 123
4.3 Spot Size Converters for Light In- and Outcoupling

Spot size converters are used to reduce the fiber–chip coupling losses when
single mode fibers are butt coupled to an integrated waveguide structure or to
a laser diode. The efficiency of the spot size converter depends again on the
type of material used for realizing the integrated waveguide based device which
results in possible waveguide parameters. Therefore in choosing an appropriate
material for the fabrication of a ring resonator filter for example and with it
a waveguide geometry, already sets the boundary conditions for the spot size
converter. It is therefore important to design a photonic device taking into
account not only fabrication aspects, but also the connections to the outside
world.
An established way to realize a spot size converter is by incorporating
tapered structures operating close to the mode cutoff, which expands their
spot size area to enhance the optical coupling efficiency without deteriorating
alignment tolerances. The use of these designs has improved the coupling
efficiency and at the same time has enabled a more relaxed alignment tolerance
of the fiber to the chip. Several taper designs have been demonstrated, such
as lateral tapers (Kasaya et al. 1993), vertical tapers, or combinations of
either lateral and vertical tapers (Yan et al. 2002) or waveguide configurations
(Hamacher et al. 2000). An example of a spot size converter consisting of a
lateral taper and a vertical coupling to a waveguide underneath is shown in
Fig. 4.13.
Here an approach is applied which relies on an adiabatic coupling between
a strong confined waveguide mode and a weak waveguide mode in a guiding
layer 1 µm beneath the actual waveguide. The rib of the waveguide is laterally
tapered down from 2 µm to 0.5 µm forcing the light to couple between the two
modes. For a better lateral confinement of the weakly guided mode a second
rib is etched in the tapered region only. An example of the integration of this
spot size converter with a ring resonator is shown in Fig. 4.14.
Fig. 4.13. Schematic of a spot size converter (Hamacher et al. 2000)

Fig. 4.14. Ring resonator with integrated SOA and spot size converter
A tradeoff between overall chip size and length of the taper region on both
sides has to be found. A longer taper with a gradual enough taper slope will of
course exhibit a lower loss. An optimization of compact lateral, vertical, and
combined tapered spot size converters using the beam-propagation method is
presented in Haxha et al. (2006).
In the previous chapters, simulation, fabrication, and building blocks of
ring resonator devices have been described. The following chapter presents
applications of ring resonator devices fabricated in different material systems
using different architectures and building blocks.
5
Devices
This chapter is entirely devoted to highlighting the various applications of

ring resonator devices developed to date using examples presented by the ring
resonator community. Some ring resonators have already been described in
Chap. 3, therefore only performance parameters like the free spectral range
(FSR), the full width at half maximum (FWHM), the quality factor Q and
the finesse F will be given in this chapter for selected devices.
Ring resonators have been primarily investigated to be implemented as
optical add–drop filters in optical networks. Several optical filter designs have
been proposed and realized in semiconductors and polymers, purely passive
devices, ring resonators with integrated gain sections and devices made out
of active material. Optical filters need to be tunable in order to be deployed
in optical networks. This fact is even more important, when it comes to ring
resonators, as the resonances of the ring resonator filter have to be aligned
exactly to the required optical channel. Wavelength tuning and center wave-
length trimming methods will be explained in this chapter.
Ring resonators can be used for flexible dispersion compensation and sev-
eral fabricated devices will be explained.
In wavelength division multiplexing (WDM) applications, a square spec-
trum response is desirable. Periodic multiplexers and demultiplexers, which
have a flat passband and a wide rejection region, can be realized integrating
a modified Mach–Zehnder interferometer (MZI) with a ring resonator.
The application of ring resonators as lasers has been demonstrated al-
ready some time ago. It has gained attention again recently with the fact that
combining ring resonators can enhance the tunability of ring lasers. Novel con-
cepts have been proposed and realized using all active ring lasers and passive
ring resonator coupled configurations with integrated semiconductor optical
amplifiers (SOAs).
Due to the advancing state of the art in fabrication technology, ring res-
onators with very small radii have been built which opens up the possibility
to use these elements as modulators and in optical signal processing systems
as logic gates and as switching devices.
126 5 Devices
As equally important as creating a wavelength with ring lasers and filtering

out a specific wavelength, wavelength conversion has been found to be possible
using ring resonator devices.
Finally as a new technology trend is emerging which is currently known as
biophotonics, ring resonators have again found their way into this new field
in the form of sensors.
5.1 Filters
Ring resonator filters have been fabricated in many material systems as is de-
scribed in Chap. 3. State-of-the-art manufacturing technologies have enabled
the demonstration of filter characteristics and parameters useful for optical
networks. The advantages of using ring resonator filters are obvious. Ring
resonators can be integrated with other components such as lasers or SOAs
to form complex photonic integrated circuits. Ring resonators do not require
facets or gratings for optical feedback and can have very small dimensions
which enables the realization of large scale photonic integrated circuits. Sev-
eral examples will be presented in this chapter highlighting various results
achieved for ring resonator filters with different geometries and materials.
5.1.1 Passive Devices
Single Ring Filters
One of the first passive single ring resonator add–drop filter with integrated
throughput and drop–port is presented in Haavisto and Pajer (1980). A thin
film of PMMA is used to fabricate the waveguides. The radius of the ring
is 4.5 cm. The waveguide widths are 10 µm, with a 17.5 µm separation from
center to center. A coupling strength of 2% is achieved. Prism coupling is
used to couple in and out of the bus waveguides. A finesse of 16 is measured.
The loss in the ring is found to be 0.05 dB cm−1 , which includes coupling
and radiation losses. Due to the large circumference of the ring, these ring
resonator filters are limited in their use as channel dropping filters.
Improvement of fabrication methods and the coming up of new material
systems (for example III–Vs) has made it possible to fabricate ring resonators
with radii far below 100 µm. One of these first ring resonators is demonstrated
in Rafizadeh et al. (1997b). Ring and disk resonators in the material system
AlGaAs/GaAs with diameters of 10.5 µm (Fig. 5.1) and 20.5 µm are presented.
An FSR of 21.6 nm and an FWHM of 0.18 nm is measured for the disk
resonator with a diameter of 10.5 µm, leading to a finesse of 120, and a cavity
quality factor Q greater than 8,500. In the case of the ring resonator with a
diameter of 10.5 µm, an FSR of 20.6 nm and an FWHM of 0.43 nm is measured,
leading to a finesse of 48 and a Q-factor greater than 3,500. The FSR and
the FWHM for the ring and disk resonators with a diameter of 20.5 µm are
5.1 Filters 127
Fig. 5.1. SEM images of a 10.5-mm-diameter (a) disk and (b) ring.
Reprinted with permission from R. Rafizadeh, Optics Letters, vol. 22, no. 16,
pp. 1244–1246 1997
10.2 nm/10.7 nm and 0.54 nm/0.12 nm, respectively. The FWHM is broader in
the case of the ring resonator due to the loss caused by the additional sidewall.
Similar ring resonators in glass, vertically coupled to the bus waveguides
with radii of 10 µm are presented in Little et al. (1999). An FSR of 20 nm
and an FWHM of 5.2 nm are measured, leading to a finesse of 4 and a qual-
ity factor of 300. The low Q factor compared to the high Q factor for the
similar AlGaAs/GaAs is due to a stronger coupling, leading to a broader
FWHM and therefore to a lower Q factor. The transmission spectrum from
both the throughput and the drop–port is shown in Fig. 5.2.
These ring resonators are used as building blocks in an 8 channel add–
drop filter (Chu et al. 1999c). The ring resonators are connected in parallel to
form a multichannel add–drop filter. Each ring resonator in the device has a
different radius and thus a different resonance wavelength. In the fabricated
device, the first ring resonator has a radius of 10.35 µm, while the remaining
ring resonators decrease in radius by increments of 50 nm. The parallel drop–
port waveguides are separated by 250 µm. A schematic of the 8-channel ring
resonator add–drop filter is depicted in Fig. 5.3.
The transmission spectrum of the eight channel add–drop filter is polariza-
tion dependent. The FWHMs of the resonances for TE and TM polarization,
from shortest to longest wavelengths, are 0.53, 0.63, 0.64, 0.71, and 0.92, 1,
1.06, 1.16 nm, respectively, leading to Q factors of 2,500 and 1,500. The aver-
age FSRs of the array is 20.1 nm and 20.2 nm for TE and TM, respectively.
A junction induced crosstalk lower than −30 dB is achieved. The channel
spacing achieved with this architecture is approximately 5.7 nm. The channel
spacing can be adjusted to the required parameter by either changing the ring
radius or the width of the waveguide, which results in a different propagation
constant.
128 5 Devices
1.2 Drop Port

Through Port
1
0.8
Power (a. u.)
0.6
0.4
0.2
0
1540 1560 1580 1600 1620 1640 1660
Wavelength (mm)
Fig. 5.2. TE polarized wavelength response of a 10 µm radius ring resonator.

Reprinted with permission from B.E. Little, IEEE Photonics Technology Letters,
vol. 11, pp. 215–217 1999
c IEEE
Fig. 5.3. 8-Channel add–drop filter
Vertically coupled ring resonators have not only been used in a parallel
cascade to form a channel dropping filter, but have also been used in a 2 × 2
cross grid array for crosstalk reduction and spectrum cleanup in Chu et al.
(1999a). A schematic of the 2 × 2 array consisting of three ring resonators is
shown in Fig. 5.4.
The three ring resonators have the same parameters and therefore filter
out the same wavelength. Resonator 1 is the first filter along the input bus
waveguide. Resonator 2 having the same resonance wavelength as resonator
1 sends the filtered wavelength to the drop–port which increases the out-
of-band signal rejection. This ring resonator configuration corresponds to a
double ring resonator leading to a steeper roll off of the filter spectrum and
consequently to a larger out-of-band rejection. Resonator 2 therefore, cleans
up the drop–port spectrum. Signal power at the resonance wavelength that
is not completely filtered out by resonator 1 is filtered out a second time
by resonator 3. Resonator 3 therefore, cleans up the throughput spectrum.
5.1 Filters 129
Fig. 5.4. 2 × 2 ring resonator add–drop node
Fig. 5.5. Photograph of a portion of the 8 × 8 cross-grid vertically coupled glass

ring resonator array.
323–325 2000
c IEEE
The add port which does not have a path to the drop–port is cross connected
to the input and throughput bus waveguide by resonator 3.
Finally, the vertically coupled glass ring resonators have been used to
realize an 8 × 8 cross grid array (Little et al. 2000b). A photograph of the
array is shown in Fig. 5.5. The height of the waveguide in the ring resonators
is 1.5 µm, and the waveguide widths vary from 1 to 1.7 µm in increments of
100 nm. The ring resonators are identical along upward directed diagonals in
the array. The spacing between the bus waveguides is 250 µm.
Vertically coupled ring resonators consisting of a dielectric material core
(Ta2 O5 –SiO2 compound glass with n = 1.8, or SiN with n = 2) and an
air cladding have been proposed, designed and fabricated by the group of
Kokubun. A summary of the work can be found in Kokubun (2005). Ring
130 5 Devices
resonators with radii between 10 and 20 µm, resulting in an FSR of 10–25 nm

and an FWHM 0.1–1.0 nm have been demonstrated. These device parameters
lead to high Q values between 1,500 and 15,000. Kokubun’s group have
pushed ring resonator technology not only in terms of ring size, but have also
demonstrated polarization insensitive and tunable (center wavelength tuning
and resonance tuning) ring resonators which are addressed in the relevant
chapters of this book.
Laterally coupled ring resonators and vertically coupled disk resonators,
with PMMA cladding and without cladding have been designed and realized
in Si3 N4 on SiO2 in Klunder et al. (2001). A finesse of 136 has been achieved
for a laterally coupled ring resonator without cladding and with a radius of
25 µm with a coupling gap of 750 nm. A finesse of 61 has been measured for
a ring resonator with PMMA cladding. TE polarized light is used for the
characterization of the devices at a wavelength of 1.55 µm. Vertically coupled
disk resonators with a radius of 15 µm have been fabricated. The vertically
coupled disk resonators are characterized using a wavelength of 640 nm. An
FSR of 2.2 nm is measured which results in a finesse of 31. The finesse is limited
by the linewidth of the laser used for characterizing the devices. A detailed
analysis of the device performance of laterally coupled ring resonators in the
same material system is presented by the group in Klunder et al. (2003b).
Here ring resonators with coupling gaps of 0.5 µm (not completely open, due
to fabrication imperfections), 0.75 µm and 1 µm have been fabricated and
characterized. A photograph of a ring resonator with a coupling gap of 0.75 µm
Fig. 5.6. Laterally coupled ring resonator in SiON with a radius of 25 µm and a
coupling gap of 0.75 µm.
pp. 1099–1110 2003
c IEEE
5.1 Filters 131
An FSR of 8 nm at a wavelength of 1,530 nm and a finesse of 120 are

obtained for a ring resonator with a coupling gap of 0.75 µm and air cladding.
A finesse of 22 is measured for the resonator with a coupling gap of 0.5 µm
with air cladding. The values for the finesse of ring resonators with PMMA
cladding are 61 and 25, respectively.
As has been mentioned in Chap. 3, recently several groups have demon-
strated ring resonators not only in silicon, but also in GaAs and InP. One
of the first racetrack shaped laterally coupled ring resonators in the material
system AlGaAs with a cavity length of 31 µm (length of the straight sections is
6 µm and 4 µm) is presented in Chin et al. (1999). A photograph of a racetrack
shaped ring resonator is shown in Fig. 5.7.
A gap between the bus waveguides and the racetrack of 200 nm is used.
An FSR of 20 nm and an FWHM of 1 nm is achieved in the case of the 6 µm
long straight sections, leading to a finesse of 20. An FWHM of 0.2 nm and a
finesse of 100 are obtained in the case of the 4 µm long straight sections. The
racetrack shaped ring resonators are surrounded by etched trenches which are
2.5 µm deep and 1 µm wide. The waveguide width is 0.4 µm and is tapered to
2 µm at the facets for easier fiber chip coupling.
Disk resonators with diameters of 5 and 10 µm in the material system
GaInAsP are presented in Ma et al. 2000a. The disk resonators are vertically
coupled using polymer (BCB) wafer bonding. The measured FSR of the disk
resonators is nearly the same as for the devices described in this chapter,
namely 38 nm for the disk with a diameter of 5 µm and 21 nm/20 nm for the
disk with a diameter of 10 µm. The two values for the FSR for the 10 µm disk
are due to the polarization dependence in this disk which is also visible in the
transmission spectrum having two sets of resonance peaks. The FWHM is,
Fig. 5.7. Racetrack shaped ring resonator.

pp. 1620–1622 1999
c IEEE
132 5 Devices
as expected, therefore also polarization dependent which leads to values of

0.25 and 0.22 nm. In the case of the 5 µm disk, the FWHM is measured to be
0.6 nm.
Vertically coupled disk resonators also in the material system GaInAsP
fabricated by wafer bonding with radii ranging from 4 to 12 µm are presented
in Dapkus et al. (2001) and Djordjev (2002). The parameters for a disk with
radius of 12 µm are: an FSR of 10 nm, an FWHM of 0.22 nm leading to a Q
factor of more than 7,000. The details of the fabrication method have been
described in Chap. 3. Photographs of bus waveguides and a disk resonator are
shown in Figs. 5.8 and 5.9.
One of the smallest racetrack shaped ring resonator notch filters in InP
is demonstrated in Grover et al. (2003). A radius of 2.25 µm is realized. The
Fig. 5.8. Bus waveguides at the bonded interface and the remaining transparent
InP layer after the etch of the microdisk mesa.
331–333 2002
c IEEE
Fig. 5.9. A smooth 2.3 µm deep microdisk mesa formed by CH4 based RIE etch.
The bus waveguides are below the remaining thin membrane of InP.
331–333 2002
c IEEE
5.1 Filters 133
straight waveguides have a length of 10 µm. An FSR of 19 nm at a wavelength

of 1,550 nm is measured. A photograph of a ring resonator notch filter is
shown in Fig. 5.10. The values for the finesse and the Q factor are 6 and 500.
A transmission spectrum of a notch filter is shown in Fig. 5.11.
Fig. 5.10. One of the smallest ring resonator notch filters in InP.
pp. 1082–1084 2003
c IEEE
Fig. 5.11. Spectral behavior of InP racetrack shaped ring resonator notch filter.
The solid line is experimental data, and the dashed line is a curve-fit.
1082–1084 2003
c IEEE
134 5 Devices
A racetrack shaped ring resonator notch filter employing multimode in-

terference couplers is presented in Rabus and Hamacher (2001). The curved
sections have a rather large radius of 100 µm when compared to the previous
examples. The length of the used MMI coupler is 150 µm. A photograph of
a ring resonator and the transmission spectrum are shown in Figs. 5.12 and
5.13. An FSR of 0.8 nm (corresponding to 100 GHz in the frequency domain
at a wavelength of 1.55 µm) and a finesse of 6 are measured.
Single ring resonator notch or add–drop filters have been studied exten-
sively in the past and have been realized in major material systems as was
demonstrated in this section. The use of single ring resonator devices as filters
in optical networks is limited by the Lorentzian like function of the transfer
function. Filters employed in optical networks require a rectangular, box-like
passband shape. Multiple coupled ring resonator devices fulfill this request
and have been fabricated in several material systems.
Fig. 5.12. MMI coupled racetrack shaped ring resonators
Fig. 5.13. Transmission spectrum of an MMI coupled racetrack shaped ring res-
onator with a radius of 100 µm
5.1 Filters 135
Double Ring Filters
One of the first integrated double ring resonator filters in silicon which is also
used in a frequency-shift keying (FSK) optical transmission experiment is
presented in Oda et al. (1994). The group had previously demonstrated a
double ring resonator with an overall FSR of 40 GHz in Oda et al. (1991).
In Oda et al. (1994) a double ring resonator consisting of ring resonators with
different radii is used. The use of two ring resonators with different radius (3.33
and 3.82 mm in this case) opens up the possibility to increase the FSR (see
(2.77)). The FSRs of the ring resonators are 14.29 and 12.50 GHz at a wave-
length of 1,550 nm, leading to an overall FSR of 100 GHz or 0.8 nm. The finesse
achieved is 200 for this configuration. A 10 GHz spaced, 8-channel, 622 MB s−1
FSK direct detection distribution experiment over 40 km is demonstrated us-
ing this double ring resonator configuration. A similar double ring resonator
consisting again of two ring resonators with different radius is presented by
the same group in Suzuki et al. (1995). The double ring resonators consist
of two ring resonators with FSR’s of 12.5 and 14.29 GHz, whose ring radii
are 1.75 and 2.0 mm, respectively. The designed overall achieved FSR is again
100 GHz (98 GHz measured). The finesse is 138. In these two ring resonator
experiments thin film heaters are integrated to be able to match the reso-
nances of the two rings. Matching of the resonances in multiple coupled ring
resonators is of utmost importance to achieve an optimum device performance
and to realize the designed filter response. The tuning of the resonances and
the center wavelength of ring resonator devices will be addressed in Sect. 5.2.
A double ring resonator where the ring resonators and the bus waveguides
are coupled to one another by multimode interference couplers is presented in
Rabus and Hamacher (2001). The radius of the curved sections is 100 µm. The
length pf the MMI coupler is 150 µm, leading to an FSR of 0.8 nm at a wave-
length of 1,550 nm. A photograph of an MMI coupled double ring resonator
Vertically coupled double ring resonators in the material system
Ta2 O5 –SiO2 are presented in Kokubun et al. (2002). Double ring resonators
with a finesse of 39.8, an FSR of 16.7 nm and an FWHM of 0.42 nm have
been designed, fabricated and characterized. Due to the double ring resonator
structure, a box-like filter response is obtained.
The design and characterization of a polarization independent double ring
resonator filter with a designed FSR of 50 GHz (47.6 GHz measured) at a
wavelength of 1,550 nm by use of SiOx Ny technology is presented in Melloni
et al. (2003a). The photograph of a double ring resonator is shown in Fig. 5.15.
The coupling between the ring resonators is realized by directional couplers
with coupling coefficients of 0.4, 0.28, and 0.4 for realizing a Butterworth type
filter response. A 15 dB extinction ratio and a bandwidth at −1 dB of 7.4 GHz
are measured. The transmission spectrum of a double ring resonator filter is
shown in Fig. 5.16.
136 5 Devices
Fig. 5.14. MMI coupled double ring resonator
Fig. 5.15. Photograph of a double ring resonator filter.

Reprinted with permission from A. Melloni, Optics Letters, vol. 28, no. 17, pp. 1567–
1569 2003
Fig. 5.16. Drop–port TE (solid curves) and TM (dashed curves) measured spectral
responses of the double-ring filter. (a) Response over a 4-nm span and (b) detail
of the response centered at 1551.2 nm. The dotted curve in (b) is the theoretical
response.
Reprinted with permission from A. Melloni, Optics Letters, vol. 28, no. 17, pp. 1567–
1569 2003
5.1 Filters 137
Double ring resonator filters are a first step to enable the realization of
box-like filter functions. Using double ring resonators as was demonstrated by
the examples in this chapter also makes enlargement of the FSR possible. In
order to improve the roll-off and out-of-band rejection, multiple coupled ring
resonator filters are required.
Multiple Coupled Ring Filters
Serially coupled, racetrack shaped, AlGaAs–GaAs triple ring resonators with

radii of 4.5 µm and straight sections with lengths between 0 and 10 µm are
demonstrated in Hryniewicz et al. (2000). A photograph of a triple ring res-
onator is shown in Fig. 5.17.
The waveguide width, bus waveguide – ring resonator gap and inter ring
resonator coupling gap is between 0.18 and 0.32 µm. The box-like filter re-
sponse of a triple ring resonator is shown in Fig. 5.18.
Parallel cascaded third- and fifth-order disk resonators in the material
system GaInAsP/InP are presented in Ma et al. (2001). Photographs of the
realized devices are shown in Fig. 5.19.
The disks have diameters of 10 µm and are only coupled to the bus
waveguides and not to each other. The coupling gap between the bus
waveguide and the disk resonator is about 0.15 µm. The distance between
two adjacent resonators is 23.85 µm, leading to a π/2 phase shift. The chosen
distance between the resonators introduces in-phase responses among the
parallel cascaded disk resonators. A box like filter response is obtained due
to the constructive interference formed at the drop–port which can be seen
in the spectral response of a fifth-order disk filter in the measurement of the
drop–port in Fig. 5.20. The FSR of the fifth-order filter is about 20 nm and
the FWHM is approximately 3.5 nm.
Parallel, vertically coupled triple ring resonators in the material systems
GaAs–AlGaAs and GaInAsP–InP are demonstrated in Grover et al. (2002).
A photograph of a triple ring resonator is shown in Fig. 5.21.
Fig. 5.17. Serially coupled triple ring resonator.

pp. 320–322 2000
c IEEE
138 5 Devices
Fig. 5.18. Transmission spectrum at drop–port of triple resonator filter.

pp. 320–322 2000
c IEEE
Fig. 5.19. Photograph of third- and fifth-order disk resonators.

Reprinted with permission from Y. Ma, Electronics Letters, vol. 37, no. 9, pp. 564–
565 2001
c IEEE
The width of the single mode waveguides is 0.8 µm. The waveguides are
tapered to a width of 2 µm at the facets to improve fiber chip coupling.
The radius of the rings is 9.55 µm. The distance between the rings is chosen
to be 20 µm (center-to-center). The achieved overall FSR of the triple ring
5.1 Filters 139
Fig. 5.20. Measured filter responses of fifth-order cascaded microdisk resonator

filter for TM polarization.
Reprinted with permission from Y. Ma, Electronics Letters, vol. 37, no. 9, pp. 564–
565 2001
c IEEE
Fig. 5.21. Photograph of a vertically coupled triple ring resonator filter. The ring
resonators are on the lower level.
pp. 900–905 2004
c IEEE
resonator is 34 nm. The transmission spectrum of the GaAs–AlGaAs ring

The increase in the FSR in the case of the third-order ring resonator
filter compared to a single ring resonator device is due to the Vernier effect
(see Sect. 2.2.2). The cavities which interfere with one another are each of
the ring resonators and part of the ring resonators with the bus waveguides.
Fabrication tolerances can introduce an unwanted Vernier effect also and have
an influence on the device performance. A difference in the radius of the rings
of 0.04% is reported for this device.
140 5 Devices
Fig. 5.22. Response of the drop–port of a parallel, vertically coupled triple ring
resonator.
pp. 900–905 2004
c IEEE
Fig. 5.23. Vertically coupled triple ring resonator.

pp. 1525–1529 2002
c IEEE
Vertically, serially coupled triple ring resonators are presented in

Yanagase et al. (2002a). The ring resonators are fabricated in the mater-
ial system Ta2 O5 –SiO2 . A photograph of a triple ring resonator is shown in
Fig. 5.23.
The bus waveguides and ring resonator 2 are in the lower layer and ring res-
onators 1 and 3 are located in the upper layer. The coupling strength between
5.1 Filters 141
the resonators and the resonators with the bus waveguides is controlled by
the thickness of the buffer layer. This is one of the advantages of using a
vertically coupled ring resonator configuration. One triple ring resonator con-
figuration presented by the authors has the following parameters: the radius
of ring resonator 1 and 3 is 22.78 µm resulting in an FSR of 10 nm, the ra-
dius of ring resonator 2 is 39.32 µm leading to an FSR of 6 nm. The measured
overall FSR expanded to 25.8 nm (theoretical value: 30 nm), which is larger
than the FSR of each of the ring resonators. In determining the value for the
overall FSR, one has to take into account different values for the propagation
constant in the buried ring resonator and the ring resonators on top. A trans-
mission spectrum of the response of the drop–port of a single resonance is
shown in Fig. 5.24. The mentioned shape factor is defined here as the width
of the bandwidth at −1 dB divided by the width of the bandwidth at −10 dB.
One of the first and so far only demonstration of 6 and 11 serially cou-
pled ring resonator filters using the material HydexTM1 which is a glass based
material, is presented in Little et al. (2004). The ring resonators are later-
ally coupled to each other. The outer rings are vertically coupled to the bus
waveguides. A photograph of a serially coupled sixth-order ring resonator
is shown in Fig. 5.25. The width and thickness of the cores of the waveguides
are approximately 1.5 µm. The vertical gap between the bus waveguides and
the outer ring resonators varies from 200 to over 1,000 nm depending on the
desired filter linewidth. The gaps between the laterally coupled ring resonators
vary from 500 to 1,400 nm depending again on the application. The drop–port
Fig. 5.24. Measured filter response of a triple ring resonator with radii of 28.52 µm
for resonators 1 and 3 and 39.32 µm for resonator 2.
pp. 1525–1529 2002
c IEEE
1
See www.nomadics.com for more information.
142 5 Devices
Fig. 5.25. Example of a sixth-order ring resonator filter fabricated with 17% index
contrast Hydex.
pp. 2263–2265 2004
c IEEE
Fig. 5.26. Measured response for first-, third-, and 11th-order ring resonator fil-
ters similar in fabrication to the device shown in Fig. 5.25. The responses have been
normalized to their 3-dB bandwidths. The dashed curve is the theoretical fit to the
11th-order filter.
pp. 2263–2265 2004
c IEEE
response of first-, third-, and 11th-order ring resonator filters is shown in

Fig. 5.26. Filters with high quality factor and large out of band rejection
have been realized. As can be seen, maximal flatness of the filters response
is obtained by increasing the number of coupled ring resonators. Polarization
5.1 Filters 143
independent resonances can be achieved by this technology by choosing the

correct aspect ratio for the waveguides.
The examples in this section have demonstrated that fabrication methods
of ring resonator filters in several material systems are available for realizing
devices with adjustable performance parameters suitable for implementation
in future all-optical networks. In order to increase the device functionality of
ring resonator filters, gain can be introduced into the cavity. Examples of de-
vices with gain inside the ring resonator are presented in the following section.
5.1.2 Devices with Gain Section
Ring resonator lasers have of course been around quite some time and ex-
amples will be presented in Sect. 5.6. This chapter focuses on ring resonator
filters with integrated gain sections for realizing improved and loss reduced
filter functions.
One of the first devices with gain inside the resonator is presented in
Djordjev et al. (2002b). Disk resonators with gain region, vertically coupled
to the bus waveguides are used to demonstrate gain trimming of the resonant
response. The fabrication of the devices has been described in Sect. 3.4.7. The
response from the throughput port of a disk with a radius of 10 µm is shown
in Fig. 5.27.
The device has an FSR of 10 nm. A quality factor of 5,700 and a finesse
of 40 are measured. The coupling coefficient is estimated to be 3.6%. As can
be seen in Fig. 5.27 it is possible with active devices to adjust the response of
the disk resonator. When the applied current supplies enough gain inside the
Fig. 5.27. Gain trimming of the transmission response of a disk with radius 10 µm.
Reused with permission from K. Djordjev, Applied Physics Letters, 80, 3467 (2002)
2002
144 5 Devices
resonator to equal the internal losses, the resonator becomes lossless and the
obtained filter spectrum is ideal.
Racetrack shaped ring resonators with integrated SOAs are presented in
Rabus et al. (2002c). A minimum radius of 100 µm which is one magnitude
larger than the ones presented in the previous paragraph is used. Single, dou-
ble and triple serially coupled ring resonator add–drop filters with integrated
SOAs are demonstrated. The fabrication details and the layer sequence of the
devices have been described in Sect. 3.4.7. A photograph of a serially cou-
pled triple ring resonator with SOAs is shown in Fig. 5.28. The output port
is transferred to the right side of the chip via another half circle for a better
accessibility when measuring the transfer response. The devices realized have
FSRs of 12.5, 25, and 50 GHz at a wavelength of 1,550 nm.
The lossless filter response of a single ring resonator with an FSR of 25 GHz
The radius of the ring resonator is 363 µm and the length of the SOA is
400 µm. The length of the couplers is 200 µm (gap = 0.9 µm), which results in
power coupling factors of 0.17. The y-axis has been normalized to the fiber–
chip coupling loss. The on–off ratio is measured to be more than 22 dB. The
FWHM is determined to be 0.012 nm. The finesse of the ring resonator is
F = 17, leading to a Q factor of more than 130,000. The filter response is
measured by sweeping the external cavity laser (ECL) signal with a resolution
of 4 pm. The SOA is operated at a current of 90 mA which provides lossless
operation of the ring resonator as is seen in the transmission spectrum. The
response from the drop–port on resonance is as high as the response from the
throughput port off-resonance.
Fig. 5.28. Serially coupled triple ring resonator with integrated SOAs
5.1 Filters 145
Fig. 5.29. Single ring resonator with two straight input/output waveguides and
directional couplers (length = 200 µm, gap = 0.9 µm), r = 363 µm, gain length =
400 µm, FSR = 25 GHz
Fig. 5.30. Drop–port response from a double ring resonator with two straight
input/output waveguides, two MMI couplers and a directional coupler (length =
150 µm, gap = 1 µm), r = 347 µm, SOA length = 500 µm, FSR = 25 GHz
The response from the drop–port of a double ring resonator with inte-
grated SOAs is shown in Fig. 5.30. The double ring resonator is made up of
two straight input/output waveguides, two 3 dB MMI couplers for the outer
couplers (length = 150 µm, width = 6 µm) and a directional coupler in the
146 5 Devices
Fig. 5.31. Filter response of the drop–port of a triple ring resonator with r =
323 µm, length of the couplers = 325 µm, gain length = 400 µm, FSR = 25 GHz
center (length = 150 µm, gap = 1 µm), r = 347 µm, SOA length = 500 µm
which results in an FSR of 25 GHz.
The driving current for each SOA is 50 mA. The shape factor which is
defined in this case as L1 divided by L2 , achieved for this configuration is 0.5.
The FWHM is measured to be 0.04 nm, leading to a finesse of 5.
The drop–port response of a triple serially coupled ring resonator is shown
in Fig. 5.31.
The measured triple ring resonator has a radius of 323 µm. The length of an
SOA is 400 µm. The length of the couplers is 325 µm with a gap of 0.8 µm for
the outer couplers and 1 µm for the couplers in the center. The achieved FSR
is 25 GHz. The driving current for each of the three SOAs is 50 mA. The steep
roll-off can be seen from the measurement. The two ripples result from a slight
resonance mismatch between the three ring resonators. The on–off ratio for
the fabricated triple ring resonator is more than 18 dB including the ripples.
Ring resonators integrated with SOAs have not only been fabricated in the
serially coupled configuration, but have also been demonstrated in the parallel
architecture. The photograph of a parallel coupled triple ring resonator is
shown in Fig. 5.32.
The ring resonators have a radius of 117 µm, a coupler length of 200 µm
(gap = 0.9 µm) and a gain length of 300 µm. An FSR of 50 GHz is realized. The
driving current for each SOA is 50 mA. The distance between the resonators
was chosen to be equal to half of the circumference of a single ring resonator.
The filter characteristic of the drop–port of the parallel, triple coupled ring
5.1 Filters 147
Fig. 5.32. Photograph of a fabricated triple coupled ring resonator in a parallel

configuration
Fig. 5.33. Drop–port response of a parallel coupled triple ring resonator with
integrated SOA, length of the couplers = 200 µm, r = 117 µm, gain length =
300 µm, FSR = 50 GHz
The ripples are due to a slight resonance mismatch due to fabrication

tolerances. The resonances can be matched when integrated platinum resistors
(see Sect. 5.2) in the ring resonators or adjusted SOA currents are used.
Another example of the integration of ring resonators with SOAs is pre-
sented in Menon et al. (2004). In this example, a vertically stacked asymmet-
ric twin waveguide structure is used to integrate the SOA with the passive
waveguides. The ring resonator is racetrack shaped and is designed as a notch
filter with only one input/output bus waveguide. A photograph of the fabri-
cated chip as well as a schematic of the design is shown in Fig. 5.34. In addition
to the SOA in the resonator, another SOA has been integrated on the side of
the input port. Details on the fabrication have been described in Sect. 3.4.7.
148 5 Devices
Fig. 5.34. Schematic of the microring resonator with integrated SOAs at the input
and in the ring. Upper left inset: Scanning electron microscope image of the fabri-
cated devices with and without an SOA in the ring. Lower right inset: Refractive
index profile and layer structure of the ring resonator with integrated SOA.
pp. 1343–1345 2004
c IEEE
The racetrack shaped ring resonator has an FSR of 0.25 nm. A multimode
interference coupler is used as in the previous example to couple into and
from the ring. The length of the MMI is 685 µm and the width is 12 µm. The
length of SOA 1 is 1,000 µm and the length of SOA 2 is 500 µm. The overall
length of the resonator is 3,020 µm. The availability of an SOA inside a ring
cavity enables the tuning of parameters of the resonator like the FWHM, the
extinction ratio and thus also the Q factor and the finesse. The tuning of these
parameters is shown in Fig. 5.35.
A fabricated ring resonator employing an SOA, an MMI and total internal
reflection (TIR) mirrors is demonstrated in Kim et al. (2005b). Details on the
layout and a photograph of a TIR based ring resonator are shown in Fig. 5.36.
An FSR of 2 nm at a wavelength of 1,568 nm is obtained. The on–off ratio
achieved is 14 dB, an FWHM of 0.3 nm is measured leading to a finesse of
more than 6, and a Q factor of more than 4,900. The length achieved for the
MMI coupler is one of the shortest reported. The SOA current dependable
tunability of the resonance characteristic of the device is shown in Fig. 5.37.
All integration schemes have in common that they require temperature
stabilization. Another effect which takes place is the resonance shift due to inc-
rease in SOA current. This is due to the current dependency of the refractive
5.1 Filters 149
Fig. 5.35. Measured extinction ratio and FWHM as a function of the ring SOA 2
current.
pp. 1343–1345 2004
c IEEE
Fig. 5.36. Total internal reflection mirror-based GaInAsP ring resonators integrated
with optical amplifiers.
pp. 1899–1901 2005
c IEEE
index of the SOA. Despite these unwanted effects, SOAs inside ring resonators
offer tunability of essential parameters and enable lossless switching devices,
which can also be used for lasers, wavelength converters and nonlinear oper-
ation. Examples will be given later on in this chapter.
150 5 Devices
Fig. 5.37. Tuning of the resonator extinction ratio via the SOA current.
pp. 1899–1901 2005
c IEEE
5.2 Tunability Methods
Tunability is essential for the system application of optical filters. In the case
of periodic filters, like ring resonators, it is important to fit the transmis-
sion curve to the defined channel spacing (e.g., ITU-Grid). Several tunability
methods exist either for trimming the center wavelength of the filter after
processing or for tuning the resonance wavelength when the ring resonator
add–drop filter is in operation. These methods including examples are briefly
explained in the following chapters. The physical mechanisms behind these
methods are:
Thermo-optic tuning. Thermo-optic tuning is realized by heating the com-
ponent for example by fabricating a metal heater on top of the waveguide in
the ring resonator. The refractive index of the material is changed through the
heat which in turn shifts the resonances. The heating process is fairly slow
but can be used to trim resonances in for example a double ring resonator
because it is very difficult to fabricate ring resonator devices with exactly the
same resonant wavelength. By adding heating elements and using the same,
negligible additional loss is obtained. On the other hand, using heaters, limits
the density of devices to maintain thermal isolation, and requires significant
power consumption to maintain the tuned status.
Another used tuning mechanism is electro-optic tuning. Here an electric
field is applied to change the properties of the material, which means either
the refractive index or the absorption. Changes in the refractive index are
realized by using a bandgap wavelength of the material which is far away
from the signal wavelength so as not to interfere with it. This effect is quite
small compared to the effect produced by the heating elements. The refractive
index increase is maximized when ring resonators with narrow peaks and low
losses are used. The change in the refractive index by using the electro-optic
5.2 Tunability Methods 151
effect is weak in III–V materials like InP, but has the advantage of being fast
compared to the thermo-optic effect described before. One has to be careful
when applying higher voltages which lead to significant losses. Changes in
the absorption of the material are obtained, when the bandgap wavelength
is closer to the signal wavelength. Only a small electric field is necessary to
induce a large loss. This can be used advantageously in switching devices
which shall be explained in the following chapter.
Carrier injection (or free carrier injection for example in InP) is another
means of changing the refractive index, absorption or gain. Introducing gain
in ring resonators is an advantage of using carrier injection. As is and will
be demonstrated in the following chapters, gain can be used to change the
response of ring resonators, the extinction ratio, the Q factor and the max-
imum transmission. Ring resonators can be switched between dropping and
transmitting a signal by changing the loss/gain inside the cavity. When gain is
introduced into the ring resonator, attention has to be paid to design and fab-
ricate a heat sink which is necessary to suppress the unwanted thermo-optic
effect.
5.2.1 Wavelength Tuning
Wavelength tuning is essential for realizing optical filters to be implemented

in optical network. Tunability offers the network suppliers flexibility when
designing and operating a network.
Thermo-Optic Tuning
A common and well-known method of tuning semiconductor and polymer

based integrated optical devices is by using temperature. The simplest way is
to heat the entire device and thus change the resonance spectrum in the case of
ring resonators. This method is of course only useful for materials with a large
temperature coefficient. A more effective way is to use local heating elements
on top of the waveguides in the ring resonator for changing the refractive
index and so for changing the resonance wavelength of the device. A detailed
theoretical analysis with implementation on thermal tuning in ring resonators
is described in Geuzebroek (2005).
One of the first papers dealing with the integration of such heating ele-
ments in ring resonators made in silicon is presented in Katila et al. (1996)
and Heimala et al. (1996). The ring resonator is made out of silicon nitride
(Si3 N4 ) rib waveguides with a height of 40 nm. The silica undercladding layer
and the silicon nitride layer thicknesses are 3 µm and 300 nm, respectively.
The diameter of the ring is 2 mm. The resistors are made out of a phosphorous
doped polysilicon layer. Aluminum with a thickness of 900 nm is used to create
the bondbads. The layout and the layer sequence are shown in Fig. 5.38.
152 5 Devices
Fig. 5.38. (a) The schematic layout and (b) the cross-section (A–A) of the
temperature-controlled ring resonator.
pp. 2260–2267 1996
c IEEE
Temperature tuning of the entire chip is used in Rafizadeh et al. (1997a)

to tune a disk resonator. A temperature tuning of 1.3 nm per 10◦C is obtai-
ned. An overall tuning range of 6.3 nm over a temperature range of 50◦C is
demonstrated. The tunability is mainly due to physical thermal expansion of
the cavity.
Triple ring resonators incorporating thermo-optic phase shifters for wave-
length and passband shape tuning are demonstrated in Madsen and Zhao
(1997) and Coppinger et al. (1999). In Madsen and Zhao (1997), double and
triple serially coupled ring resonator filters in silicon with chromium heating
elements inside the ring resonators are presented. A minimum bend radius of
4 mm is achieved leading to an FSR of 0.066 nm. The response of the filter
is adjusted by the integrated heating elements and is analyzed by a devel-
oped algorithm. Triple serially coupled ring resonators integrated in a silica
waveguide technology with diameters of 8.5 mm are presented in Coppinger
et al. (1999). The FSR of the filter is 7.5 GHz and a bandwidth of 1.5 GHz
is measured. The application of this rather large ring resonator filter is the
filtering of microwave signals to obtain a microwave transfer function. The

integrated phase shifters (heating elements) cover part of the waveguide in
the ring. As will be explained in the following paragraph in more detail using
another ring resonator example, these heating elements are used to match
the resonances of each ring to one another which leads to an improved pass-
band of the filter. Once the desired passband characteristic is achieved, the
power of the heating elements is increased by the same amount which enables
wavelength tuning of the entire filter.
Another method applied successfully to achieve wavelength tuning is
by using platinum as a heating element inside a GaInAsP/InP based ring
resonator (Rabus et al. (2002b)). The platinum layer which enables the local
heating of waveguide segments of the ring resonators via the thermo-optic
effect is deposited directly on the waveguides. Before the platinum is
deposited, the entire device is covered with a SiNx layer. This layer assures
that no metal will diffuse into the waveguide material and no waveguide mode
couples to the metal and increases the propagation losses. A photograph of
a part of a Pt-resistor in a ring resonator is shown in Fig. 5.39. The pads for
contacting the Pt-resistor are composed of layers of titanium, platinum and
gold. A triple ring resonator with integrated Pt-resistors is shown in Fig. 5.40.
The Pt-resistors have two functions. The first function is the tuning to spe-
cific wavelengths. The second function is the resonance matching of multiple
coupled ring resonators to each other. A demonstration of resonance matching
of a double ring resonator is shown in Rabus et al. (2002a).
The tunability of a single ring resonator notch filter is shown in Fig. 5.41.
The Pt-resistors enable a finer tuning to a specific wavelength and so an
improved on–off ratio is obtained.
The second feature of the Pt-resistors, the resonance matching, is demon-
strated with a double ring resonator (r = 200 µm, upper and lower coupling
factors κ0,2 = 0.5, middle coupling factor κ1 = 0.055, coupler lengths =
150 µm). The unmatched transmission spectrum of the double ring resonator
is shown in Fig. 5.42. The upper curve shows the response of the throughput
port and the lower curve that of the drop–port. This device has not been
antireflection coated, which leads to Fabry–Perot resonances in the straight
Fig. 5.39. SEM picture of waveguide with platinum layer

154 5 Devices
Fig. 5.40. Triple ring resonator with integrated platinum heaters on top of the
waveguides in the curvatures and adjacent contact pads
Fig. 5.41. Tunability of a single ring resonator notch filter with Pt resistors
waveguides. The resonances are visible in the transmission spectrum of the

throughput port. The typical filter characteristic of the double ring resonator
is not disturbed despite the Fabry–Perot resonances. It depends on the app-
lication, if antireflection coating is required. The contrast of the throughput
port is 16 dB and that of the drop–port 13 dB. The simulated filter response
reveals a difference of the effective refractive index of about 0.0003, which
Fig. 5.42. Unmatched transmission spectrum of a double ring resonator. Dotted

line – simulation, solid line – measurement
−
−
−
−
−
−
−
Fig. 5.43. Resonance matched double ring resonator. Dotted line – simulation, solid
line – measurement
is also the origin of the different contrast values. According to the calcula-
tion, the refractive index is lower in the upper ring. This is mainly due to
fabrication tolerances which occur during the deep etching on the outer side
of the waveguide in the curvatures. The physical lengths of the resonators are
the same. In order to match the resonance frequency in both ring resonators,
the upper Pt-resistors are used. A voltage of 0.5 V is applied to match the
resonance frequency. The temperature coefficient of InP can be approximated
in this example to (see Sect. 3.4):
dn
≈ 0.0001 K−1 (5.1)
dT
The local temperature increase is therefore approximately 3 K. The result
of the resonance matched double ring resonator is shown in Fig. 5.43.
156 5 Devices
Fig. 5.44. Schematic views showing critical dimensions. (a) Top view of SOI res-
onator and (b) cross-sectional view of SOI rib waveguide.
pp. 364–366 2006
c IEEE
The tuning to a specific wavelength can be performed after the frequency

matching in such a way, that both of the Pt-resistors in the two rings are used.
The driving voltage of the Pt-resistors is different in both ring resonators due
to the previous frequency matching but has to be increased by the same
amount in order to shift to a specific wavelength.
A tunable racetrack shaped ring resonator notch filter in SOI is demon-
strated in Kiyat et al. (2006). Details of the ring resonator and the used
waveguide are shown in Fig. 5.44. A directional coupler with a coupling
gap of 0.8 µm and a length of 48 µm is used, which results in a coupling
factor of 0.45. The FSR of the ring resonator is determined to be 248 pm.
The FWHM is measured to be 40 pm leading to a Q-factor of 38,000 and
a finesse of 6.2. A heater electrode made out of nickel is used for tuning
the resonance wavelengths of the filter. The electrode covers part of the
waveguide in the curvature as in the previous example where platinum is
used as the electrode material. The resistance of the heating element includ-
ing the bond pads is 170 Ω. The tuning of the resonance wavelength against
power consumption is shown in Fig. 5.45.
Electro-Optic Tuning
One of the first electro-optic tunable, racetrack-shaped ring resonator notch

filters in GaInAsP/InP is presented in Grover et al. (2004). The racetrack
shaped ring resonator filter is also one of the smallest reported, with a radius of
curvature of 2.25 µm and straight sections with a length of 1 µm leading to an
FSR larger than 30 nm. The electrodes are made on top of a BCB layer which
is used to planarize the device after the waveguides have been etched. The top
Fig. 5.45. Shift in resonance wavelength as function of applied power (dotted line
with arrows indicate the tuning parameters for reaching the FSR).
pp. 364–366 2006
c IEEE
Fig. 5.46. Effective refractive index vs. electric field, assuming all the bias appears
across the core. Corresponding value of the resonance wavelength is shown on the
right side. Refractive index in the absence of electric field is obtained from sim-
ulations. Quadratic fit provides us the quadratic electro-optic coefficient for the
waveguide. Value of the coefficient for the waveguide is a lower bound for the
quadratic electro-optic coefficient of the core.
pp. 467–469 2004
c IEEE
contact (p-side) is made out of deposited Ti–Pt–Au. The bottom contact is

made out of deposited Pd–Sn–Au–Pd–Au after the chip has been thinned.
The change of the refractive index and the tuning of the resonance wavelength
of the ring resonator filter with applied electric field are shown in Fig. 5.46.
158 5 Devices
The quadratic electro-optic coefficient can be calculated from the mea-

surement and is determined to be −9.8 × 10−15 cm2 V−2 . A tuning range
of 0.8 nm which corresponds to 100 GHz at a wavelength of 1,550 nm is
realized.
A narrowband, widely electro-optically tunable filter based on a whisper-
ing gallery mode cavity using LiNbO3 is demonstrated in Savchenkov et al.
(2003). The device has an FSR of 10 GHz and can be used in microwave
photonics communication systems. The obtained tuning speed of the filter is
approximately 10 ns. The actual spectrum shifting time is determined by the
30 MHz bandwidth of the filter and does not exceed 30 ms. A channel crosstalk
suppression of −20 dB for a 50 MHz channel spacing is measured.
Carrier Injection Tuning
One of the first vertically coupled disk resonators with a radius of 10 µm using
carrier injection for increasing the refractive index and tuning the resonance
of the resonator is demonstrated in Djordjev et al. (2002d). The device has an
FSR of 10 nm and a Q value of 5,500. The fabrication of the device has been
explained in Sect. 3.4.7. The tunability of the device is 1.1 nm (mA)−1 . The
refractive index change is ∆n = −2.1 × 10−3 . The linewidth of the resonances
is 0.28 nm. A transmission measurement of the tuning of the disk resonator is
shown in Fig. 5.47.
The vertically coupled disk resonator can also be used as an on/off switch.
A demonstration of the switching behavior is shown in Fig. 5.48. The switch
Fig. 5.47. The free carriers injected into the disk resonator change the modal re-
fractive index, which blue-shifts the resonant wavelengths.
pp. 828–830 2002
c IEEE
Fig. 5.48. Switching behavior of a disk at a wavelength of 1,598 nm. Change in the
drive current by ∆I = 200 µA is enough to toggle the switch from the off to the on
state. The voltage change required for this transition is ∆V = 0 : 1 V.
pp. 828–830 2002
c IEEE
is operated by changing the drive current from 200 µA to 400 µA requiring

a voltage of only 0.1 V. A contrast between the on and off state of 10 dB is
achieved.
Two different types of tunable disk resonators vertically coupled to bus
waveguides are presented in Choi et al. (2004d): a disk resonator coupled to
air-guided rib bus waveguides and another one fabricated on buried hetero-
structure bus waveguides. Diagrams and SEM photographs of both types of
resonators are shown in Fig. 5.49. The disk resonators have a radius of 12 µm.
The two types of disk resonators have Q values of 8,000 and 3,000, respectively.
A tunability of 1.5 nm and 2.1 nm is demonstrated at a current of 3 mA due
to free carrier injection.
5.2.2 Center Wavelength Trimming
In the previous chapter, examples of tuning mechanisms have been presented

which enable the selection of resonance frequencies inside a ring resonator
filter. Postfabrication methods which enable the permanent fitting of the
entire resonance spectrum to a desired resonance frequency grid are explained
in this chapter.
Ring resonators once fabricated offer practically no way to change the
resonance spectrum besides heating the entire sample and thus changing the
refractive index which in turn shifts the resonance spectrum. Therefore it is
highly desirable to have methods which can be applied after the ring resonator
filters have been fabricated to change the resonance spectrum. One way is to
160 5 Devices
Fig. 5.49. Schematic drawings and SEM images of: (a) a disk vertically coupled
to a single air-guided bus waveguide, type 1, and (b) a disk built on a buried bus
line, type 2, respectively. The SEM image in (b) shows a stain-etched cross-sectional
view of a buried bus waveguide section.
Reused with permission from S.J. Choi, Applied Physics Letters, 84, 651 (2004)
2004
use a UV sensitive polymer overlay to change the resonance spectrum (Chu

et al. 1999d). Polysilane is used as a polymer overlay to change the resonance
spectrum of vertically coupled glass ring resonators with a radius of 19 µm.
The FSR of the device before UV exposure for TE and TM polarized input
light is 10.54 and 10.56 nm, respectively. Dip coating of the resonator is used
which results in a thickness of the polymer of 2 µm. An Hg-lamp UV source
(Toshiba Lightec TOSCURE 251), having a center wavelength of 370 nm and
an irradiance of 38 mW cm−2 is used for the UV exposure. The trimming
of one resonance using TM polarized input light is shown in Fig. 5.50. The
wavelength shifts obtained for the resonances are 7.62 and 9.14 nm for TE and
TM polarized input light. No change has been observed to a device after UV
exposure which was measured again after one month.
A similar polysilane UV trimming technique has been used in
Suzuki et al. (2002) to trim the resonance wavelengths of a vertically coupled
1 × 8 filter array. A radius of 15 µm is used for all rings in the array. The ring
resonators are made on a SiO2 substrate. The ring and bus waveguides are
made out of Ta2 O5 /SiO2 compound glass. The cores of the ring resonator
waveguides are covered by polysilane with a thickness of 10 µm using dip
Fig. 5.50. The detailed TM spectra around one resonance after three different UV
time exposures.
pp. 688–690 1999
c IEEE
Fig. 5.51. Measured refractive index change of polysilane vs. UV exposure time.
pp. 745–750 2002
c IEEE
coating. The change of the refractive index of polysilane versus UV exposure

time is shown in Fig. 5.51.
The authors report, that the center wavelength change which is caused by
the index change of the polysilane overcladding is due to the effect of oxidation
or humidity absorption during UV exposure of the polysilane. The authors also
162 5 Devices
report on an aging effect which makes hermetic packaging of the device after
UV trimming necessary.
A polysilane overcladding has also been used in Sparacin et al. (2005)
to trim the resonance spectrum of Si3 N4 ring resonators fabricated on a
3 µm SiO2 under-cladding layer on a (100) silicon substrate with a diameter
of 100 µm. The ring resonators are designed for a wavelength of λ = 1,550 nm.
The difference to the aforementioned polysilane trimming method which
utilizes dip coating is that the used polysilane is brought on to the ring
resonators using a vapor phase technique. The polysilane used is (PECVD
6M2S). The UV exposure is carried out with a MINERALIGHT R
handheld
lamp (model UVGL-25) emitting at a wavelength of λ = 254 nm at an inten-
sity of 1.7 µW cm−2 . The refractive index decreased exponentially by ∼4 %,
from 1.52 to 1.46 at a wavelength of λ = 1,550 nm. The overall spectrum shift
obtained is 12.8 nm for the TE mode and 23.5 nm for the TM mode.
Another UV trimming method is presented in Haeiwa et al. (2004) where
the refractive index of the waveguide material SiN is exposed to UV light
which imposes a refractive index change in the material. The refractive index
change of an SiN film versus UV exposure time is shown in Fig. 5.52. The
UV light source used is an Ar ion second harmonic generation laser with a
wavelength of 244 nm and an output power of 100 mW.
Details on the used ring resonator can be taken from the inset in Fig. 5.52.
The maximum refractive index change is ∆n = −1.3×10−2 , leading to a center
wavelength shift of −11.4 nm. This UV induced index change is reported to
be stable for more than 1,200 h without any further treatment. A side effect
of the UV induced index change is that coupling strength and loss inside the
Fig. 5.52. Refractive index change of SiN film by UV irradiation and calculated
center wavelength shift.
pp. 135–137 2004
c IEEE
ring resonator are affected due to the absorption of UV radiation, changing

the performance of the ring resonator filter. The FWHM in the presented
example changed from an initial value of 0.34 to 1.02 nm after UV irradiation
for 24 h.
A different approach of trimming the resonance spectrum of SiN ring res-
onators is demonstrated in Wang et al. (2005). Oxygen plasma treatment is
used to convert the ring resonator material into SiOx Ny changing the refrac-
tive index and thus affecting the resonance spectrum. The refractive index
of SiOx Ny depends on the amount of oxygen incorporated in the SiN layer
and can be tuned over a large range between 1.45 (SiO2 ) and 2.0 (Si3 N4 ).
A schematic of the used ring resonators is shown in Fig. 5.53. The radius of
the ring resonator filter is r = 25 µm. The length of the zero-gap coupler is
4 µm. The waveguide width is 1 µm for the bus waveguides and 2.5 µm for the
ring resonator waveguides. An overall change in the resonance wavelength of
8.9 nm is demonstrated.
In the previous examples, tuning of semiconductor based ring resonators
has been addressed. In the following example, a polymer ring resonator notch
filter fabricated by soft lithography (Poon et al. 2004b) is tuned via a photo-
bleaching technique of CLD-1 chromophores. The radius of the ring resonator
is 207 µm, and the height and width of the waveguides are approximately
1.6 and 1.4 µm, respectively. The coupling gap is 430 nm wide. The trimming
is performed using a visible broadband light source which is focused on one
part of the ring resonator. The wavelength shift achieved is 8.73 nm after an
exposure time of 1.75 h. The stability of chromophores, as it is well known,
Fig. 5.53. Ring resonator filter. (Inset: (upper-left) SiN waveguide structure;
(upper-right) photograph of fabricated ring resonator filter).
pp. 582–584 2005
c IEEE
164 5 Devices
can degrade the performance of this material by the presence of oxygen. There-
fore hermetic sealing and an inert gas atmosphere are needed for practical
applications.
Tunability of the resonance spectrum of ring resonators by changing the
refractive index with different methods has been presented, highlighting sev-
eral practical examples. These methods have only affected the FSR and have
had sometimes as a side effect, changed the shape of the resonances which is
seen by an increase/decrease of the FWHM. A way of definite tuning of the
resonance shape is by implementation of tunable couplers.
5.2.3 Tunable Couplers in Ring Resonators
Tunable couplers have been fabricated and demonstrated as has been des-
cribed in Sect. 4.1. Their implementation in ring resonators is so far limited
by the size and technologies used. A widely used technique for realizing tun-
able couplers in ring resonator filters is by applying heaters on top of the
waveguides in the coupling region. This also affects the resonance spectrum
which is also shifted and has to be compensated with another heater inside
the ring cavity.
Tunable couplers enable the exact matching of the coupling coefficient
to the loss inside ring resonator filters realizing a critically coupled state
with maximum on–off ratio. One of the first disk resonators on an SOI plat-
form, employing tunable couplers is presented in Lee and Wu (2005). Micro-
electromechanical system (MEMS) actuated deformable waveguides are used
to tune the shape of the resonance. The layout and an SEM photograph of
the MEMS disk is shown in Figs. 5.54 and 5.55, respectively.
The disk has a radius of 10 µm and the gap is 1.4 µm wide when no volt-
age is applied. A voltage of 70 V is needed to obtain a zero-gap coupler. One
advantage of using MEMS couplers is the elimination of electron beam litho-
graphy to define the coupling gaps smaller than 500 nm. In this example a
single standard lithographic step is required to pattern both the disk and the
bus waveguides simultaneously.
Another MEMS approach is described and demonstrated in Nielsen et al.
(2004), where a MEMS parallel plate structure is placed above the ring res-
onator enabling switching by lowering the plate and elevating it thus creating
a change in the effective refractive index and a change in resonance wave-
length. A detailed theoretical analysis of tunable couplers in ring resonators
is presented in Kaplan (2006). Tunable couplers are one possibility of realiz-
ing perfectly matched ring resonator filters apart from gain implementation
inside the ring cavity as was described in Sect. 5.1.2. Due to the advancing
state-of-the-art of manufacturing technologies, it will only be a matter of time
until tunable couplers are widely incorporated in ring resonator filters realiz-
ing various types of devices.
5.3 Dispersion Compensators 165
Fig. 5.54. Schematic structure of the disk resonator with deformable waveguides.
(a) At zero bias, the disk is completely uncoupled since the suspended waveguides are
far away from the disk. (b) With voltage applied on the electrodes, the waveguides
are bent toward the disk, increasing the coupling between the waveguides and the
disk.
pp. 1034–1036 2005
c IEEE
Fig. 5.55. SEM photograph of MEMS actuated disk resonator.

pp. 1034–1036 2005
c IEEE
5.3 Dispersion Compensators
In classical optics, “dispersion” is used to denote the wavelength dependence

of the refractive index in matter, (dn/dλ, where n is the refractive index
and λ is the wavelength) caused by interaction between matter and light.
In communication technology, “dispersion” is used to describe any process by
which an electromagnetic signal propagating in a physical medium is degraded
because the various wave components (i.e., frequencies) of the signal have
different propagation velocities within the physical medium.
166 5 Devices
Material dispersion causes different wavelengths to travel at different

speeds due to the variation of the refractive index of the fiber core with wave-
length. However, part of the light also travels in the cladding of the fiber,
which has a different refractive index. This portion of the light propagates
therefore right through the cladding at different speeds to the core – an eff-
ect known as waveguide dispersion. Material and waveguide dispersion are
combined to give an overall effect called “chromatic dispersion.”
Chromatic dispersion is caused by a variation in the group velocity of
light traveling within a fiber with changes in optical frequency. A data pulse
always contains a spectrum of wavelengths. As the pulse travels along the
fiber, the shorter wavelength components travel faster (negative dispersion)
than the longer wavelength components (positive dispersion). This effect
broadens the pulse and causes it to interfere with neighboring pulses and
distort the transmission signal. The allowed chromatic dispersion in an opti-
cal network is inversely proportional to the square of the transmitted bit rate,
so as data rates increase, dispersion tolerances decrease dramatically. While
2.5 Gbit s−1 networks can tolerate 16,000 ps nm−1 of dispersion, 10 Gbit s−1
networks can tolerate only 1,200 ps nm−1 and in 40 Gbit s−1 networks, the
tolerance drops to only 60 ps nm−1 of dispersion. In addition, 40 Gbit s−1 sys-
tems require a wider modulation spectrum. A 12 GHz wide spectrum is typical
for 10 Gbit s−1 , for 40 Gbit s−1 , it can be as high as 50 GHz, which means the
total dispersion per channel is higher. Chromatic dispersion has a fixed, sta-
ble component in addition to a dynamic one. Most of the fixed dispersion is
caused by the fiber and is predictable as a function of the type of fiber and
the distance. In addition, the optical components present in the optical path
add a smaller, fixed contribution. A dynamic contribution must be added to
the fixed one. Since many passive components do not have a simple flat or
linear dispersion curve, laser drift (caused by aging or temperature change)
can lead to dispersion fluctuations that are nonlinear and hard to predict.
Another important type of dispersion is the polarization mode dispersion
(PMD). PMD is caused by light traveling faster in one polarization plane
compared to another. Fundamentally, it is caused by the core of the fiber
not being perfectly round in cross-section and by birefringence which is intro-
duced by mechanical forces. As a result, the optical thickness is not absolutely
identical on every possible axis. PMD is a special case for modal dispersion
or intermodal dispersion, which is caused by multimode propagation of light.
Each mode travels with a different group velocity. The modes can be higher
order modes of the same polarization, or modes of different polarization, as
is it is the case for PMD. Intermodal dispersion is of no concern under single
mode operation.
Finally, there is structural dispersion also known as quadratic dispersion
which is effectively the second derivative of the transmission phase with res-
pect to frequency. Structural dispersion is determined by the architecture and
layout of the photonic integrated circuit.
5.3 Dispersion Compensators 167
Ring resonators can be used as dispersion compensators and examples

presented in literature will be highlighted in this chapter. Practical dispersion
compensators should have a limited tunability, uniform insertion loss upon
tuning of dispersion and should enable the operation at multiple wavelengths.
Ring resonators can be used to enhance the physical length by forcing the light
to traverse the physical distance many times. Resonant enhancement, however,
comes at the price of finite bandwidth. A large enhancement results in a
narrower bandwidth. Ring resonators allow large compensation for multiple
wavelengths simultaneously in a so far limited frequency range.
The building block of a multistage dispersion compensator using ring res-
onators is shown in Fig. 5.56. The group delay of a ring resonator follows
a periodic curve. Cascading multiple loops enables synthesization of various
delay curves.
A tutorial overview on ring resonators used as dispersion compensators
highlighting various configurations is given in Schwelb (2004).
Dispersion compensators utilizing ring resonator architectures have been
investigated by Madsen and coworkers extensively and some examples will
be given in the following section. One filter type is of special interest, the
optical all pass filter (APF). APF’s are devices that allow phase correction or
equalization without introducing any amplitude distortion (Lenz and Madsen
1999). One of the first experimental demonstration of a multistage all pass
filter dispersion compensator that uses ring resonators in planar waveguides
is demonstrated in Madsen et al. (1999b). A two stage optical filter consist-
ing of two parallel coupled ring resonator notch filters with a bend radius of
2.2 mm is realized. Ge-doped silica waveguides on a silicon substrate are used
for the waveguides. A coupling gap of 3 µm is used for the directional couplers.
The phase shifter inside the ring resonators is made out of chrome with a resis-
tance of 55 Ω, which is deposited on the upper cladding of each ring. The FSR
of the device is 12.5 GHz. The parameters of the filter are: D = 4,200 ps nm−1 ,
±5 ps group delay ripple, <3 dB loss, and a 4.5 GHz passband width. The
filter is used in a transmission experiment using 320 km of fiber. The fil-
ter is capable of compensating negative and positive dispersion. Due to
the birefringence of the used waveguides, the filter is polarization depen-
dent and therefore to avoid PMD, either TE or TM polarized input light
is used.
Fig. 5.56. Single ring resonator filter building block of multistage dispersion com-
pensator
168 5 Devices
Four stage ring resonator dispersion compensating filters are presented

in Horst et al. (2000). Ring resonators with an FSR of 25, 33, and 50 GHz
are fabricated in SiON technology. Metal heaters are deposited on top of the
waveguides to allow for tunable phase shifting. A dispersion of −415 ps nm−1
over a usable bandwidth of 22 GHz for a filter with an FSR of 50 GHz is
obtained. The filter is as the previous example polarization dependent which
means that accurate polarization control has to be maintained.
A new architecture combining ring resonators and MZIs to be used for
dispersion compensation is presented in Madsen et al. (1999b). A schematic
of a single stage ring resonator filter with an MZI design is shown in Fig. 5.57.
The MZI design enables a tunable APF by incorporating phase shifters in the
ring and the MZI. The phase shifter in the MZI enables a tunable coupling
ratio while the phase shifter in the ring resonator is for resonance tuning.
Four stages of the ring resonator/Mach–Zehnder combination are cascaded
yielding a dispersion of 1,800 ps nm−1 and a passband of 14 GHz, which is
greater than 50% of the FSR. A low group delay ripple of 15 ps peak is
achieved. The device is realized in Ge-doped silica planar waveguides. The
directional coupler lengths are 477 µm. The radius used is 1 mm, and the FSR
of the filter is 25 GHz (at 1,490 nm).
A device consisting of ring resonators coupled to arms of an MZI are used
in Madsen et al. (2004) to compensate and emulate PMD. A photograph of
a similar realized device is shown in Fig. 5.58. The device which is used for
Fig. 5.57. Ring resonator with asymmetric Mach–Zehnder interferometer
Fig. 5.58. A two-section planar waveguide chip with three-stage all-pass filters in
the upper and lower arms of each section.
pp. 1041–1050 2004
c IEEE
5.4 Mach–Zehnder Interferometers Combined with Ring Resonators 169
Fig. 5.59. Experimental device.

pp. 1041–1050 2004
c IEEE
PMD measurements consists of a directional coupler, which is fed into an

MZI which has ring resonators attached at each arm. A schematic of the used
configuration is shown in Fig. 5.59.
The experimental device based on Ge-doped silica waveguides is a four
stage all pass filter. The radius of the ring resonators used is 350 µm, leading
to an FSR of 74.4 GHz. A discrete polarization beam splitter (PBS) on the
output side and a linearly polarized input signal with the polarization aligned
to either the TE or TM axis is used in the measurement setup. The authors
demonstrate a differential group delay (DGD) tuning range over 100 ps for
10 Gbit s−1 and 25 ps for 40 Gbit s−1 data. Second order PMD with a tuning
range of 255 ps2 and third order PMD with a range of 2,430 ps3 are also
demonstrated.
Ring resonators combined with MZIs can not only be used for dispersion
compensating optical devices. Other applications will be addressed in the
following chapter.
5.4 Mach–Zehnder Interferometers Combined

with Ring Resonators
MZIs are well known and have found their way into integrated optics. A basic
MZI consist of two waveguides of different lengths, which are joined at both
ends using 3 dB couplers. A schematic is shown in Fig. 5.60.
The transfer characteristic has a periodic sin2 shape over wavelength as-
suming that the two 3 dB couplers are wavelength independent regarding their
coupling behavior. The distance between the minima and maxima in terms
of wavelength depend on the length difference in both arms of the MZI. The
3 dB bandwidth of the filter in the frequency domain is given by
c0
∆f = (5.2)
2neff ∆L
where ∆L is the length difference between both arms.
170 5 Devices
Fig. 5.60. Layout of a basic MZI
Fig. 5.61. Configuration of the periodic multi/demultiplexer with a ring resonator.

pp. 1174–1176 2000
c IEEE
Optical filters deployed in optical networks are required to have a “box-

like” filter response as was described earlier. One way to do this is by using
multiple coupled ring resonators. MZIs can also be modified to realize a more
rectangular transfer function by introducing a ring resonator into one of the
arms. A ring resonator has one 2π phase shift per FSR. In order to obtain
passband flattening of the transfer characteristic of the MZI, two ring res-
onator phase shifts per FSR of the MZI are required. This means that the
FSR of the ring resonator has to be half of the FSR of the MZI. Translated
into length requires the circumference of the ring to be twice as large as the
delay length of the MZI.
A periodic multi/demultiplexer with a square spectrum response having
an FSR of 200 GHz is presented in Kohtoku et al. (2000). A schematic of the
device layout including design parameters is shown in Fig. 5.61. The device is
fabricated on an InP substrate. MMI couplers are used in the device which
enables lower manufacturing tolerances regarding the splitting ratio. The
circumference of the ring resonator is 876.6 µm leading to an FSR of 100 GHz
which is half of the FSR of the MZI. A suitable coupling ratio for the cou-
pler in the ring resonator has to be chosen to obtain a rectangular transfer
characteristic with a large rejection bandwidth and a high extinction ratio.
The MMI coupler in the ring resonator has a splitting ratio of 15:85. The
transmission characteristic of the device is shown in Fig. 5.62.
5.4 Mach–Zehnder Interferometers Combined with Ring Resonators 171
Fig. 5.62. Transmission spectra of the fabricated device with 200 GHz FSR. The
solid line indicates ports 1 –3. The dashed line indicates ports 1 –4. The dotted line
indicates the reference waveguide.
pp. 1174–1176 2000
c IEEE
In the following example, a balanced MZI with one arm coupled to a ring
resonator in AlGaAs–GaAs is presented (Absil et al. 2000b). The motivation
in using a balanced MZI is by realizing a notch filter and thus introduc-
ing a zero pole filter when speaking in terms of the Z-transformation (see
Chap. 2). When the round-trip loss in the ring resonator is small compared to
the waveguide coupling strength, the ring resonator introduces a phase shift
of π at resonance. The resonant phase shift is independent of the coupling
coefficient, and minimum transmission occurs. A photograph of the device is
shown in Fig. 5.63. The amplitude transfer response of this kind of device is
given by Absil et al. (2000b):

1 t − e−(α+jβ)2πR −jφ
T = 1+ e , (5.3)
2 1 − te−(α+jβ)2πR

t = 1 − κ2 ,
where φ is the phase imbalance between the two arms of the MZI. The
Y-splitters are assumed of having a 3 dB splitting ratio with a wavelength
independent loss. The device has only one input and one output port coupled
to 3 dB Y-splitters. The transfer spectrum obtained is slightly deviated from
the theoretical notch filter response due to an imbalance in the arms of the
MZI of only 110 nm. The spectrum is shown in Fig. 5.64.
These two examples show how different transfer functions can be realized
with a ring resonator coupled to one arm of an MZI. The response depends on
the length of the arms (balanced/unbalanced), on the type of response desired
and on the coupling coefficient from one arm to the ring resonator.
172 5 Devices
Fig. 5.63. SEM micrograph of the fabricated MZI in GaAs–AlGaAs.

pp. 398–400 2000
c IEEE
Fig. 5.64. Measured Mach–Zehnder transmission spectrum and analytical fit show-

ing imbalanced arms.
pp. 398–400 2000
c IEEE
Recently a novel geometry for ring resonators coupled to an MZI has been
demonstrated in polymer in Paloczi et al. (2003b). A photograph of the pro-
posed and fabricated device is shown in Fig. 5.65. The device is realized in
SU-8. The ring resonators are racetrack shaped and have a bending radius of
100 µm. The straight coupling sections have a length of 50 µm. The designed
coupling gap is 750 nm. The ring resonators have an FSR of 2.2 nm for TE
polarized light at a wavelength of 1,550 nm. Complicated phase properties and
5.5 Modulators 173
Fig. 5.65. Novel design of ring resonators coupled to an MZI. The size of the device
is approximately 1.2 mm.
Reprinted with permission from G.T. Paloczi, Optics Express, vol. 11, no. 21,
pp. 2666–2671 2003
Fig. 5.66. Experimental data (dotted ) and the theoretical fit (line). The fitting
parameters used for the fit were: polarization 93% TE and 7% TM, effective indices
1.48475 for TE and 1.48555 for TM, power coupling coefficients 0.46 for TE and
0.85 for TM, and waveguide loss of 30 dB cm−1 .
Reprinted with permission from G.T. Paloczi, Optics Express, vol. 11, no. 21,
pp. 2666–2671 2003
a periodic spectral response of the absolute field amplitude are observed. The
transmission spectrum of the device is shown in Fig. 5.66.
MZIs with ring resonators have also been demonstrated to be used for all
optical switching (Heebner et al. 2004).
Optical switching in ring resonators will be described in Sect. 5.8. Ring
resonators coupled to MZIs have also been investigated for implementation as
modulators Tazawa and Steier (2005).
The properties of ring resonator modulators are addressed in the following
chapter.
5.5 Modulators
Wavelength division multiplexing (WDM) and dense WDM systems require

modulators to generate the required optical signals to be transmitted. Electro-
optic (EO) modulators based on an MZI are used for fulfilling this task so far.
174 5 Devices
Resonator based modulators are of increasing interest as they can be oper-

ated in principle with a lower driving voltage and have a more compact de-
sign which is a prerequisite for integrated devices. One of the earliest resonator
based modulators is of course the Fabry–Perot resonator. Ring resonator mod-
ulators have recently gained attention due to the improvement of fabrication
technologies which enable very small devices as was demonstrated in previ-
ous chapters, and due to improved and novel material properties. In contrast
to MZI based EO devices, ring resonator modulators have a rather small
bandwidth which limits their applications so far to systems requiring a mod-
est radio frequency (RF) bandwidth but operate at high carrier frequencies.
A single wavelength is used as the input signal for a ring resonator modulator.
The operating wavelength of the ring resonator modulator is tuned, so that
the steepest slope of one of the resonances in the throughput port spectrum
is used. Applying a voltage through an electrical field in the ring resonator
induces a shift in the position of the resonance peak, resulting in an intensity
change at the throughput port.
The main bandwidth limitation is the time which is needed to build up the
optical field in the ring resonator. The time of one roundtrip in the resonator
is given by (see also Chap. 2):
2πRng
τRoundtrip = (5.4)
c0
The total time needed for a ring resonator to build up a resonance (also
referred to as cavity ring-down time) is given by
2πRng
τCav = m · (5.5)
c0
The cavity ring-down time can be related to the quality factor Q of a ring
resonator by Leinse (2005)
λQ
τCav = (5.6)
2πc0
The cavity ring-down time can also be written in terms of the finesse F of the
ring resonator using (2.31) to:
F Rng
τCav = (5.7)
c0
Instead of using the group refractive index, the effective refractive index can
also be used. The group refractive index takes into account the wavelength
dependency of the effective refractive index and therefore a more accurate
value for the cavity ring-down time is obtained. The relationship between the
bandwidth of the filter spectrum response and the maximum transmission
capacity is analyzed theoretically and experimentally in Suzuki et al. (2001).
A vertically coupled ring resonator add–drop filter with a radius of 20 µm is
used for the experiment. An FWHM of 0.75 nm is obtained. A pulsewidth of
5.5 Modulators 175
3.4 ps is measured at the output port of the ring resonator, when an input
pulse of 1.2 ps with a sech pulse shape from a fiber laser is inserted at the input
port. A maximum transmission capacity of 50 Gbit s−1 is estimated for this
type of ring resonator filter. Linear and nonlinear pulse propagation in ring
resonator filters is analyzed theoretically in Melloni et al. 2003b. In particular,
the so-called slow wave structures consisting of a chain of directly coupled ring
resonators are investigated.
In the following paragraph a relationship between the bandwidth and the
driving voltage based on Rabiei et al. (2002a) is given.
Since the traditional switching voltage Vπ in MZI based modulators can-
not be used for a ring resonator modulator a switching voltage of VFWHM is
defined, which is the voltage required to shift the resonant wavelength by the
FWHM. Sometimes other definitions are used which define the voltage needed
for a specific extinction ratio (e.g., V10 dB or V20 dB ).
The modulation bandwidth BW of a ring resonator is given by
1
BW = , (5.8)
2πτLifetime
where τLifetime is the finite lifetime for the photons in the resonator. Using the
frequency relationship for the FWHM
1
∆ωFWHM = . (5.9)
τLifetime
The bandwidth can be rewritten as
BW = ∆υFWHM (5.10)
The index change ∆neff required to shift the resonance by ∆υFWHM is given by
∆neff ∆υFWHM
= , (5.11)
neff υ
where neff is the effective refractive index of the ring resonator and υ is the
optical frequency. The index change relation for EO materials is given by
1 V
∆nef = Kn3EO r33 , (5.12)
2 d
where K is the confinement factor, V is the applied voltage, nEO is the re-
fractive index of the EO material, and d is the distance between the elec-
trodes. Using these equations, the relationship between the bandwidth and
the driving voltage is obtained, yielding the fundamental limitations of these
kind of modulators:
BW Kυn3EO r33
= . (5.13)
VFWHM 2neff d
176 5 Devices
The polymer based ring resonator modulator presented in Rabiei et al. (2002a)
has a bandwidth of 2 GHz and a transmission experiment has been carried
out demonstrating an open eye diagram at 1 Gbit s−1 . Detailed expressions
for the modulation bandwidth and drive voltage for optical resonator based
modulators can be found in Gheorma and Osgood (2002). Ring resonator-
based travelling wave modulators are also analyzed theoretically in Tazawa
and Steier (2006). Theory on ring resonator modulators, device design, and
characterization is presented in Rabiei (2003) and Leinse (2005). In Leinse
(2005), modulators consisting of a ring resonator coupled to an MZI are also
demonstrated.
Polymer ring resonator modulators with diameters of 50, 70, and 150 µm
are demonstrated in Rabiei et al. (2002b). The modulators have FSRs of
1,100, 770, and 300 GHz at a wavelength of 1,330 nm. The fabrication and
layer sequence of the ring resonators has been described in Sect. 3.5.1. The
bandwidth of the devices is 18, 16, and 12 GHz, respectively. The authors
propose a multiwavelength ring resonator modulator and filter configuration
for implementation in WDM systems shown in Fig. 5.67.
In order for the system to be realized, the FSRs of each of the ring resonator
modulators and filters have to be greater than the wavelength spectrum used
in the WDM system. Each ring resonator has to be frequency hence wave-
length stabilized to modulate or filter out the correct wavelength with high
extinction ratio. The device consisting of several ring resonators is feasible
with current state-of-the-art manufacturing methods and as was described in
a previous chapter, wavelength stabilization can be performed with several
methods.
Modulators based on ring resonators on silicon on insulator are presented
in Pradhan et al. (2005), Xu et al. (2005). A ring resonator notch filter is used
with a rib waveguide structure. Photographs of the ring resonator and the
metal contacts are shown in Fig. 5.68.
Electro-optic modulation of the refractive index is achieved by a p–i–n
junction integrated with the ring resonator. The p- and n-doped regions are
on the inner and outer side of the ring waveguide having a concentration of
1019 cm−3 . The width and height of the ring and bus waveguides are 450 and
Fig. 5.67. Integrated WDM system using microring modulators and filters.
pp. 1968–1975 2002
c IEEE
5.5 Modulators 177
Fig. 5.68. SEM and microscope images of the fabricated device. (a) Top view SEM
image of the ring coupled to the waveguide with a close-up view of the coupling re-
gion. (b) Top view microscope image of the ring resonator after the metal contacts
are formed. The metal contact on the central p doped region of the ring goes over
the ring with a 1 µm thick silicon dioxide layer between the metal and the ring.
Reprinted by permission from Macmillan Publishers Ltd: Nature, Xu, vol. 435,
pp. 325–327 2005
c
250 nm, respectively. A coupling gap of 200 nm is realized. The diameter of

the ring resonator is 12 µm leading to an FSR of 15 nm. The FWHM of a
resonance around 1,550 nm is measured to be 0.04 nm. The cavity ring down
time for this type of modulator using (5.6) is calculated to be τCav = 33 ps.
The performance of the device is demonstrated using a nonreturn-to-zero (RZ)
signal at 0.4 Gbit s−1 and an RZ signal at 1.5 Gbit s−1 .
Vertically coupled disk modulators on InP are demonstrated in Sadogopan
et al. (2005). The operation of the device is based on a depletion width trans-
lation mechanism which involves the transfer of optical energy between two
waveguides at the resonance of a ring or disk resonator that couples the two
waveguides. A photograph of a fabricated ring modulator as well as a sketch
of the cross-section is shown in Fig. 5.69.
The fabricated device has an outer radius of 12 µm and an inner radius
of 8.5 µm. The 3 dB modulation BW of the device is 8 GHz. A wavelength
shift of 0.2 nm at a driving voltage of 0.9 V leads to an extinction of the signal
wavelength of 10 dB. The device shows open eye diagrams at 10 Gbit s−1 which
demonstrates the functionality at these speeds.
Ring and disk resonators with a high Q-factor enhance the coupling be-
tween waveguides, thus a low voltage modulation of the resonator Q can modu-
late the power transfer, enabling high speed operation. Examples of techniques
for fabricating ring and disk resonators modulator circuits have been demon-
strated in this chapter and ongoing research will enable faster and smaller
devices.
178 5 Devices
Fig. 5.69. SEM photograph of fabricated ring modulator. The cross-section of the
device shows the three-dimensional coupling which is necessary for the depletion
mechanism.
pp. 567–569 2005
c IEEE
5.6 Lasers
Semiconductor ring lasers have been investigated in the past and have recently
gained attention again because they can be readily integrated with other op-
toelectronic devices and do not require gratings or facets for optical feedback.
Ring lasers avoid spatial hole burning by traveling wave operation, which
results in single mode operation and reduced sensitivity to feedback. Different
types of active and passive coupling configurations have been used in ring res-
onator lasers including Y-splitters, directional couplers and multimode inter-
ference couplers. MMI couplers have attracted widespread attention, because
they allow increased fabrication tolerances compared to directional couplers
as has been described in Chap. 4.
5.6 Lasers 179
Ring lasers have been described in detail in literature, therefore only brief
highlights will be given and different types of laser configuration will be dis-
cussed in the following chapter.
5.6.1 All Active Lasers
In this chapter ring resonator-based lasers fabricated using gain material for
realizing ring and bus waveguides are described. These ring lasers will be
referred to as all active ring lasers.
One of the first demonstrations of the principle of a ring resonator laser is
presented in Weber and Ulrich (1971). A rhodamine 6G2 doped polyurethane
film which serves as the waveguide is coated on the surface of a cylindrical
glass rod with a diameter of 5 mm providing the ring resonator geometry. This
film is pumped with a N2 laser having a wavelength of 337 nm emitting a beam
with a sheet like output characteristic, enabling the pumping of only a short
width of approximately 0.2 mm of the film coated glass rod. In using this pump
method, a ring resonator laser is obtained. A prism coupler is used to couple
light out of the ring resonator. One of the first integrated ring resonator lasers
is presented in Liao and Wang (1980). The ring laser is made in the material
system GaAs–GaAlAs and has a straight branching waveguide serving as the
output. The output waveguide is in principle a Y-splitter. A ring radius of
100 µm is realized. The lasing wavelength observed is around 873 nm under
pulse operation. The threshold of the laser is above 300 mA. In order to prevent
the output waveguide from acting as a second cavity, it was made 3 mm long
and the facets have been scratched to lower or eliminate the reflectivity.
One of the first continuous wave (cw) operating integrated disk resonator
lasers in the GaAs material system with integrated Y-splitter for output cou-
pling is presented in Krauss et al. (1990). The disk has a diameter of 84 µm
and a threshold current of 24 mA. The low threshold current is achieved by
only depositing a metal contact layer on the outer side of the disk with a
width of 4 µm, forming a metal ring and isolating the inner disk with a poly-
imide layer. One of the first integrated ring resonator lasers in GaAs/AlGaAs
with a radius of 150 µm and a threshold current of 72 mA is demonstrated in
Hohimer et al. (1991). A Y-splitter is used to couple out of the ring resonator.
The width of the used waveguide is 8 µm. A side mode rejection ratio of 22 dB
is measured for current values greater than 50 mA, which was one of the high-
est side mode rejection ratios at that time. The laser emits at a wavelength
of 852 nm. The output power has been increased significantly for similar ring
laser devices in Hohimer et al. (1992) by tilting the output waveguide (tilt
angle = 5◦ , 10◦ ), eliminating a back reflection of modes which compete with
the laser mode for gain. Controlled back reflection is also investigated for uni-
directional lasing in ring resonator lasers suppressing the counter propagat-
ing wave. Ring lasers in GaAs–AlGaAs having Y-splitters as output couplers
2
Formula: C28 H31 N2 O3 Cl
180 5 Devices
with tilted output waveguides are also presented in Tsai et al. (1998). Tilt an-
gles from 0◦ to 15◦ between the waveguide orientation and the normal of the
cleaved facet are realized. Ring lasers having a radius of 200 µm and 300 µm are
demonstrated.
Unidirectional ring lasers using racetrack geometry are presented in
Hohimer and Vawter (1993). Inside the racetrack cavity is an equal-radius
crossover S-waveguide that is independently electrically contacted. This S
bend waveguide enables unidirectional lasing. Single mode operation at wave-
lengths around 870 nm with an output power greater than 10 mW is reported.
This group around Hohimer has performed extensive research on ring res-
onator lasers and has published numerous papers on the subject. Therefore
only an excerpt of relevant publications from this group is presented in this
chapter.
One of the first ring resonator lasers in the material system GaInAsP is
demonstrated in Hansen et al. (1992). The ring resonator laser has a diameter
of 3 mm and emits at a wavelength of 1.54 µm. The threshold current is mea-
sured to be 157 mA. The ring laser is coupled to a straight waveguide with a
gap of 1.5 µm. The ring laser is fabricated using buried passive rib waveguides
and two buried heterostructure active sections with multiple quantum wells.
The length of the active sections is 500 µm each, placed on opposite sites
of the ring laser. Continuous wave operation is achieved with a side mode
suppression of 28 dB.
InP-based ring lasers employing MMI couplers are presented and analyzed
in detail in Pennings et al. (1993, 1994) and van Roijen et al. (1994). An output
power of 9 mW under cw operation at a wavelength around 1.6 µm with a side
mode suppression of 35 dB is obtained. Two types of ring resonator geometries
are analyzed and fabricated. A schematic of the devices is shown in Fig. 5.70.
Completely active ring resonators are fabricated. Two types of MMI couplers
Fig. 5.70. Ring laser employing MMI couplers – design

5.6 Lasers 181
are used, output couplers with a width of 7 µm with lengths between 185 and
233 µm and combiners with a width of 8 µm and lengths between 61 and 75 µm.
The use of MMI couplers in ring resonator lasers has also been investi-
gated and demonstrated in Krauss et al. (1994, 1995b). Racetrack shaped
ring resonators in the GaAs/AlGaAs material system with a radius of 400 µm
and a length of 320 µm for the straight section and the 3 dB MMI coupler are
used. The device has a threshold current of 140 mA and emits at wavelengths
around 870 nm. Ring lasers having directional couplers, MMI couplers and
Y-splitters as output configurations are analyzed, fabricated and compared in
Krauss et al. (1995a).
A square ring resonator laser in GaAs–AlGaAs with an MMI output cou-
pler is presented in Kim et al. (1997). A schematic of the laser is shown in
Fig. 5.71. The round-trip length of the square ring resonator laser is deter-
mined by the width of the used waveguide. A waveguide width of 6 µm results
in a 3 dB coupling length of 283 µm and for a waveguide width of 8 µm, a 3 dB
coupling length of 509 µm is determined. The width of the coupling waveguide
is 2 µm for the 6 µm ring waveguide and 3 µm for the 8 µm ring waveguide.
Threshold currents of 200 mA and 250 mA are measured at lasing wavelengths
around 800 nm.
A ring laser with a triangular shape in InP–GaInAsP is presented in Ji
et al. (1997). A photograph of the device is shown in Fig. 5.72. The lasing
wavelength of a presented device is around 1.26 µm with a threshold current
of 64 mA (θ = 25◦ ) under pulsed operation (100 ns pulsewidth, and 20 µs
period). A side mode suppression ratio of 25 dB is achieved. The width of
the waveguide is 4 µm and the cavity length is 300 µm. The Q of the cavity
is determined by the so called structural angle θ which defines the output
angle φ. The dependence of these angles on the output power is investigated
experimentally.
One of the first photonic integrated circuits in GaAs–AlGaAs comprising
a mode-locked ring laser, an optical amplifier, and a high speed photodetector
is fabricated and demonstrated in Vawter et al. (1997). A schematic of the
device is shown in Fig. 5.73.
Fig. 5.71. Layout of square ring resonator laser

182 5 Devices
Fig. 5.72. SEM fabricated triangular ring laser, cavity length 300 µm, structural
angle θ and light output angle φ are defined in the diagram.
pp. 1469–1471 1997
c IEEE
Fig. 5.73. Schematic of photonic integrated circuit.

Reprinted with permission from IEEE Photonics Technology Letters vol. 9, pp. 1634–
1636 1997
c IEEE
The device is used to generate mm wave signals at frequencies of 30, 60,

and 90 GHz depending on the radius of the ring laser. The radii of the used
ring lasers are 860, 430, 290 µm, respectively. Passive mode locking of the ring
laser is realized by a saturable absorber, generating pulses between 1 and 10 ps.
Mode locking in ring lasers is investigated both theoretically and experi-
mentally in Yu et al. (1998) with respect to different configurations of intra-
cavity saturable absorbers. A schematic of the designs analyzed is shown in
Fig. 5.74.
Low-threshold GaInAsP all active ring lasers are presented in Griffel et al.
(2000). The layer sequence has already been described in Chap. 3. The thresh-
old current of the cw operated ring laser is 66 mA with a side mode suppression
ratio of 26 dB at wavelengths around 1.6 µm.
5.6 Lasers 183
Fig. 5.74. Configurations: (a) symmetric, (b) asymmetric and (c) dual absorber
Unidirectional bistability in circular waveguide InGaAs/GaInAsP semi-

conductor lasers has been observed and demonstrated in Sorel et al. (2002).
A ring diameter of 1.2 mm is used for the cw ring laser having a threshold
current of 125 mA. The FSR obtained is 0.166 nm. Three operating regimes
regarding the clockwise and counter clockwise propagating modes are ob-
served. Operation regime one is between 125 and 135 mA, and both modes
oscillate. Operation regime two is between 135 and 220 mA, unidirectional
operation is observed, either in one or the other direction, depending on the
driving current. The extinction ratio between the lasing and nonlasing output
is measured to be more than 30 dB. Operation regime three is above 220 mA,
here the modes randomly switch to the output ports. Bistability is imposed by
two electrodes (see Fig. 5.75), each on one side of the output waveguide. For
example by applying a short pulse to either of the two electrodes when the ring
laser is operated in regime two, changes the direction of the mode in the ring
laser. Operating regimes of ring laser in the material system GaAs–AlGaAs
are also investigated by this group in Sorel et al. (2003). A photograph of a
ring laser where the contact layout can be seen which was also used for the
InP based devices is shown in Fig. 5.75. The presented ring laser is integrated
together with three photodiodes (PD) realized in a simple manner by placing
contacts on top of the active waveguides. Their position can be taken from
the photograph.
The radius of the ring lasers investigated is 1 mm. A directional coupler
with a coupling gap of 1 µm is used to couple light out of the ring. The output
waveguide is tilted at an angle of 5◦ to reduce back reflection of the light from
the cleaved facets below a reflectivity of 0.5%. Different operating regimes are
analyzed both theoretically and experimentally. In addition to the operating
regimes described previously, a bidirectional regime is observed, where the
two counterpropagating modes undergo harmonic alternate oscillations.
Defined mixing of clockwise and counterclockwise ring laser modes of two
ring lasers fabricated in InGaAs/GaAs/AlGaAs is used in Cao et al. (2005)
to obtain frequency beating. The application of such integrated circuits is
in ring laser gyros and optical rotation sensors. The ring lasers in the form
of racetracks have a radius of 1 mm and a straight section length of 2 mm.
Clockwise and counterclockwise operation is achieved by integrated absorbing
spiral elements coupled to the ring waveguide located inside one ring and on
184 5 Devices
Fig. 5.75. Ring laser with contact pad layout.

Reprinted with permission from IEEE Journal of Quantum Electronics, vol. 39,
pp. 1187–1195 2003
c IEEE
the outside of the other ring. Tunability of the lasing wavelength of each ring
is realized by integrated heaters.
Single mode ring resonator laser operation using two coupled active ring
lasers is demonstrated in Hamacher et al. (2003). The ring lasers are fabricated
in the GaInAsP/InP material system. A photograph of the device is shown in
Fig. 5.76.
The ring laser consists of two coupled ring resonators having a radius of
50 µm. The ring resonators, the couplers and the bus waveguides can be driven
independently by separate electrical contacts. Each ring resonator laser gen-
erates an individual slightly shifted comb like laser spectrum with respect to
the driving current. Wavelength tuning to specific single mode emission is
achieved using the Vernier effect. A measured single mode spectrum of the
double ring laser configuration is shown in Fig. 5.77. A high side mode sup-
pression ratio of >40 dB and a maximum output power of 6.5 mW per facet
at a driving current of 250 mA is obtained.
A single mode ridge waveguide (RW) laser coupled to two ring resonators
using MMI couplers also in the material system GaInAsP/InP has been
realized and demonstrated in Bach et al. (2003). The layout of the device
The diameter D1 and D2 of the ring resonators is 50 µm and 60 µm, respec-
tively, leading to an overall FSR of 18 nm. The FSR of the ridge waveguide
5.6 Lasers 185
Fig. 5.76. Single mode operating double ring laser
Fig. 5.77. Single mode output spectrum of a double ring laser
Fabry–Perot cavity is 0.5 nm and is small compared to the FSR of each of

the ring resonators. Therefore only the FSRs of the ring resonators will con-
tribute to the mode selection. The width w of the ridge waveguide is 4 µm.
The facet width e inside the ring resonators is given by
√
e = 2 · w. (5.14)
The ring resonators are not metal contacted and are only pumped through
the straight ridge waveguide by a lateral carrier flow over the cap layer.
A measured spectrum of such a double ring resonator laser is given in Fig. 5.79.
Ring lasers have been made with deeply etched structures to provide a
stronger confinement of the optical mode in the waveguide. This, however,
comes at the price of higher threshold currents due to surface recombination
and scattering losses. Another drawback is the fact that deeply etched struc-
tures lase in multiple lateral modes because of the large index step. Other
types of waveguide designs used for realizing ring lasers like ridge waveguides
or buried heterostructures with small index steps exhibit a better lateral mode
confinement and have lower threshold currents. A detailed analysis of the lat-
eral and longitudinal mode discrimination of ring lasers is presented in Nabiev
et al. (1993).
186 5 Devices
Fig. 5.78. Layout of double ring laser. The ring resonators are also covered with
BCB which is left out for illustration purposes.
pp. 377–379 2003
c IEEE
Fig. 5.79. cw laser characteristics at room temperature of a 700 µm long device

with 50 and 60 µm ring resonators. The inset shows the emission spectrum at a
drive current of 450 mA.
pp. 377–379 2003
c IEEE
5.6 Lasers 187
Ring lasers with different architectures (circular, square, and triangular

ring cavities) and waveguide designs having a radius down to 50 µm have
been described in this chapter. A theoretical comparison of photon recycling
in circular, square, and triangular ring laser cavities is given in Numai (2000).
Smaller devices have been fabricated by several groups and examples are
given in Sect. 6.3.
5.6.2 Devices with Gain Section
Merits of integrated ring resonators have been pointed out throughout this
book. This chapter demonstrates again some of them, like the integration
capability with active sections to enhance the characteristics of ring lasers.
One of the first integration of an SOA with passive silica waveguides and
coupler is demonstrated in Hibino et al. (1992). The gain in the ring resonator
consists of a GaInAsP buried heterostructure SOA. The facets of the SOA
are coated with antireflecting films made from SiO. Both facets of the
SOA are coupled to lensed optical fibers with a lens radius of 7 µm. The
SOA module is inserted into the passive integrated silica ring resonator using
etched fiber grooves. The size of the module is 1.2 mm × 1.2 mm × 3 mm. The
fibers are fixed to the passive waveguides using an UV curable resin. The to-
tal resonator length is 10.2 cm leading to an FSR of 2 GHz at a wavelength of
1.3 µm. A comb like lasing spectrum is obtained with a threshold current of
only 17 mA. The device is temperature stabilized with a Peltier cooler.
An active resonator with passive Y-coupler and outcoupling waveguides
has been realized in the material system InGaAs-GaAs in Cockerill et al.
(1994). The ring laser is made out of strained-layer InGaAs–GaAs buried
heterostructure waveguides. The integration of the passive Y-coupler is made
by selective area growth. The radius of the ring laser is 200 µm. The facets
are antireflection coated to avoid back reflections. The laser operates at a
wavelength around 1 µm with a threshold current of 25 mA. A side mode
suppression of 24 dB is achieved.
Passive circular sections, an SOA and a saturable absorber in the material
system GaInAsP are used in Ohno et al. (2002) to fabricate a modelocked ring
laser. The radius of the curved sections is 100 µm. Modelocking at frequen-
cies of 29.5, 41.2, and 61.7 GHz is obtained for three different devices with
corresponding SOA lengths of 1,000, 650, and 300 µm and saturable absorber
lengths of 30, 20, and 20 µm, respectively. An MMI coupler is used for output
coupling.
Ring lasers with MMI output couplers have been showed to have superior
operation characteristics to devices with Y-junction couplers (examples have
been given in this chapter and in the previous chapter for both coupling
versions). A further improvement in performance of MMI coupled ring lasers
as suggested in Pennings et al. (1994) is by the use of a single active gain
section in the ring, and making the MMI coupler and outcoupling waveguides
of passive waveguide material.
188 5 Devices
The realization of a racetrack shaped ring resonator laser employing an

SOA as the gain element, where the input and output waveguides, the MMI
coupler and the curved ring sections are made out of passive material is
described in the following paragraph. The passive waveguide is made from
quaternary GaInAsP with a bandgap wavelength of 1.06 µm. The ring radius
is 117 µm. The length of the MMI coupler is 150 µm. The fabrication details
and the layer sequence of the passive waveguide have already been described
in Chap. 3. A photograph of the ring laser with integrated SOA is shown in
Fig. 5.80.
A standard ridge waveguide laser structure was used for the 500 µm long
SOA section, which required an additional epitaxial growth step. The SOA
is butt-coupled to the passive waveguide using a self-alignment process. The
detail of the SOA as well as the fabrication process has been described in detail
in Chap. 3.4.7. The butt coupling losses have been determined to be <2 dB
each. The facets of the input and output waveguides have been antireflection
coated. The ring laser has been designed to have an FSR of 50 GHz. The ring
laser is operated at T = 17◦ C. The lasing spectrum was measured using a
tapered fiber and an optical spectrum analyzer. The relationship between the
lasing power in the fiber and the injection current in the SOA is shown in
Fig. 5.81.
The threshold current is as can be seen from the diagram as low as 25 mA.
The kinks in the L–I curve of the ring laser are accompanied by mode-hops
and can be explained by coupled cavity effects, the second cavity being formed
by the output path which includes the MMI coupler (Krauss et al. 1995a). The
kinks have also been observed for fabricated ring lasers which do not have an
antireflection coating on the facets of the outcoupling waveguides. The SOA
does not form a cavity with the butt-joints. This has been verified by using
a straight waveguide with integrated SOA (Fig. 5.82) where the input and
output waveguides have been antireflection coated.
No lasing has been observed in such a device with a 500 µm SOA. An
interesting feature observed was that the same threshold current of 25 mA
Fig. 5.80. Photograph of a ring resonator laser with integrated SOA

5.6 Lasers 189
Fig. 5.81. Output power characteristic of a racetrack laser
Fig. 5.82. Photograph of a straight passive waveguide with integrated SOA
is obtained for a device which was not antireflection coated (Fig. 5.83). The
measurement has been performed monitoring the output power directly using
a photodiode. If Fig. 5.83 is compared to the output power characteristic of the
ring laser (Fig. 5.81), the fiber–chip coupling loss which has been determined
to be ∼5 dB, has to be improved. This can be done for example by using a
vertical taper.
In the output power characteristic there are no kinks visible which leads
to the assumption, that the kinks in the ring laser L–I curve are mainly
attributed to the output path which includes the passive waveguides and the
MMI coupler. This also indicates that the threshold current is only dependent
on the butt-joint losses. The curved waveguide sections have negligible bending
losses.
The lasing spectrum of the racetrack laser at a drive current of 162 mA
is shown in Fig. 5.84. An Advantest optical spectrum analyzer Q8384 with a
maximum resolution of 10 pm in this wavelength range is used. The measured
FWHM is therefore smaller than 1.25 GHz. The total power in the fiber is
0.6 mW per facet. As can bee seen from the lasing spectra, the side mode
suppression ratio (SMSR) is 45 dB. The spectrum shows the designed FSR of
50 GHz of the ring resonator.
190 5 Devices
Fig. 5.83. Output power characteristic of a straight waveguide with integrated SOA
without antireflection coated input and output passive waveguides
Fig. 5.84. Lasing spectra of a racetrack laser at a drive current of 162 mA, showing
single mode operation with SMSR = 45 dB
The lasing spectra covering the entire 50 nm gain spectrum is shown in

Fig. 5.85. It can be clearly seen that there is only one lasing mode.
Ring lasers fabricated by a combination of active and passive waveguide
sections enable the lowering of the threshold current as compared to all active
devices, where the whole structure needs to be electrically contacted. Special
attention has to be made when joining active and passive elements. Mode field
profiles of both waveguides have to match as close as possible to minimize
transition losses.
5.6 Lasers 191
Fig. 5.85. Lasing spectrum of a racetrack laser showing single mode operation over
a 50 nm gain spectrum
Fig. 5.86. Examples of passive ring resonator coupled lasers
5.6.3 Passive Ring Resonator Coupled Lasers
Ring resonator coupled lasers are different from the ring lasers in the previous
chapter as the gain medium is placed outside the ring resonator cavity. Passive
ring resonators serve as optical filters which enable single mode operation and
tunability of the laser. Several examples including different architectures are
described in this chapter.
Ring resonator coupled lasers offer many promising advantages over
conventional tunable lasers, including ultra wide wavelength tuning range,
high side mode suppression ratio, uniform threshold and efficiency, narrow
linewidth, and low frequency chirp. The integration of active gain sections,
passive waveguides and ring resonators forming a ring resonator coupled laser
has recently been proposed and analyzed in Liu et al. (2001). A schematic of
the two proposed architectures is shown in Fig. 5.86.
A detailed analysis of a single passive ring resonator coupled laser having
a design like the one shown in Fig. 5.86b, with only one gain region as the
output waveguide (the other waveguide is also passive like the ring) is given
in Bian et al. (2003).
192 5 Devices
Fig. 5.87. Ring resonator coupled laser
One of the first fabricated passive single ring resonator coupled lasers is
presented in Park et al. (2002). A schematic of the layout of the device is
shown in Fig. 5.87. One material is used for the ring coupled laser. The ring
is operated in a transparent mode. Only the waveguides at the facets intro-
duce gain into the device. The waveguides not leading to the facets act as
absorbers. The ring coupled laser is made using the material system GaInAsP
on InP. Three types of devices having a radius of 5, 10 and 20 µm are fab-
ricated. The width of the waveguides is 0.4 µm. Directional couplers with a
coupling gap of 0.2 µm are used. The bus waveguides are tapered to a width
of 2 µm having a taper length of 200 µm. The FSR of the ring resonator is
chosen large enough that a resonance is in the middle of the gain spectrum
to achieve single mode operation. The threshold of the laser is 70 mA under
pulsed operation. The lasing wavelength is measured to be 1.549 µm.
In order to achieve single mode operation, the radius of the ring resonator
has to be small, in the order of 10–20 µm. Another way to achieve single
mode operation using passive coupled ring resonators is by using a double
ring architecture. The ring resonators have slightly different radii which leads
to an increase of the overall FSR as was described in Chap. 2. Passive double
ring resonator coupled lasers (DR-RCLs) are analyzed theoretically in Liu
et al. (2002a).
A double ring coupled laser is demonstrated in Rabiei and Steier (2003)
using an integrated tunable polymer double ring filter, fiber coupled to an
erbium-doped fiber amplifier, serving as the gain medium. Thermo-optic and
electro-optic tunable devices are presented. The layer sequence of the device
as well as the fabrication is similar to the ones described in Chap. 3. The
radius of the ring resonators is 240 and 246 µm. The overall obtained FSR is
approximately 39 nm at a wavelength of 1,550 nm. A layout of the double ring
resonator structure is shown in Fig. 5.88.
One of several advantages of using a double ring resonator is the possibility
of a higher tuning enhancement factor using the Vernier effect. The tuning
enhancement factor is given by
1 R2
T = = , (5.15)
L1 R2 − R1
1−
L2
5.6 Lasers 193
Fig. 5.88. Layout of double ring resonator structure
Fig. 5.89. Photograph of a DR-RCL
where L1 and L2 are the cavity lengths of each ring and R1 and R2 the
radius, respectively. The tuning range of a double ring resonator is T times the
tuning range of a single ring. The polymer double ring resonator configuration
has a tuning enhancement factor of 40. The transmission experiments are
performed using TM polarized input light, fed to the input port of the double
ring resonator from an erbium doped fiber amplifier which serves as the gain
medium. An output power of 1 mW and a side mode suppression ratio of 30 dB
are measured for both the thermo-optic and the electro-optic device.
One of the first integrated double ring resonator coupled lasers in GaInAsP
combining passive ring resonators with integrated SOAs is presented in Rabus
et al. (2005a). A photograph of a DR-RCL is shown in Fig. 5.89. The layer
sequence and the dimensions of the waveguides are the same as is described
in Chap. 3 for the passive ring resonator and the integration of an SOA into
the ring cavity.
Two configurations of DR-RCLs are investigated and the radii of the rings
are 100 and 108 µm for the first configuration and 100 and 104 µm for the
second configuration. An FSR of approximately 15 and 30 nm is obtained res-
pectively. The coupling between the bus waveguides and the ring resonators is
realized by a codirectional coupler with a length of 500 µm and a coupling gap
of 1 µm. The coupling from the bus waveguide to the ring depends critically on
the separation. The achieved splitting ratio is 3 dB. The laser cavity consists
of SOA1 and SOA2 (see Fig. 5.89) which have a length of 500 µm each and
194 5 Devices
the two ring resonators. The chip length is 2 mm. The remaining SOAs are
not biased and they are used as absorbers to suppress the lasing of sub cavity
modes. The end facets of the chip are as cleaved and are uncoated, which
could be improved to increase the performance in the future. The output of
the DR-RCL is collected using a tapered fiber at SOA2.
The spectrum of the DR-RCL with an FSR of 15 nm is shown in Fig. 5.90,
when the driving currents for SOA1 and SOA2 are 110 and 90 mA, respec-
tively. Due to the fact that the FSR of the double ring resonator does not cover
the entire gain spectrum, another lasing mode is present. The power differ-
ence between the two lasing modes is 13 dB. The threshold of the DR-RCL is
about 30 mA when SOA1 is biased above 40 mA.
The spectrum of the DR-RCL with an FSR of 30 nm is shown in Fig. 5.91.
A side mode suppression ratio of >35 dB is obtained which is limited by the
dynamic range (−35 dB) of the used optical spectrum analyzer (OSA). The
linewidth of the lasing wavelength is also limited by the bandwidth of the OSA
which is 0.06 nm. From calculations, the linewidth is estimated to be <2 MHz.
The driving currents for SOA1 and SOA2 are 100 and 90 mA, respectively.
The output power varies with the currents supplied to SOA1 and SOA2
(Fig. 5.92). The threshold of the DR-RCL is about 25 mA when SOA1 is biased
above 40 mA.
An output power exceeding 1 mW is demonstrated in Bian et al. (2006).
The resonance wavelength of the rings can be tuned by the integrated platinum
resistors on top of the passive waveguides in the ring resonators. Varying
the current of one heater results in a shift of resonance wavelength of the
corresponding ring resonator due to the thermo-optic effect and the lasing
can jump from one mode to the next one where the transmission peaks of
the double ring resonators overlap. Figure 5.93 shows a series of lasing spectra
when varying the heater current between 0 and 30 mA. When the current of
Fig. 5.90. Spectrum of a DR-RCL with an FSR of 15 nm

5.6 Lasers 195
−
−
−
−
−
−
−
−
−
−
Fig. 5.91. Spectrum of a DR-RCL with an FSR of 30 nm
Fig. 5.92. PI-curve of the DR-RCL with an FSR of 30 nm changing the current at
SOA2
the integrated heater is changed, the laser produces 20 different modes, which
are spread uniformly over a 17 nm wavelength range.
Another architecture for a double ring resonator coupled laser is analyzed
theoretically in Chung et al. (2005). The device consists of a coupled ring
reflector and an active straight waveguide providing lasing activity. A sche-
matic of the proposed device is shown in Fig. 5.94.
The properties and formulas describing the behavior of coupled ring re-
flectors have been addressed in Chap. 2. The radius of the ring resonators is
chosen to be different enabling a larger FSR for the coupled device utilizing
the Vernier effect. Two different ring radii enable an enhanced tuning factor as
is described in this chapter. The reflective properties are obtained by allowing
inter ring coupling.
196 5 Devices
−
−
−
−
−
−
−
−
−
Fig. 5.93. Wavelength tuning by applying a current between 0 and 30 mA to the

integrated Platinum heater
Fig. 5.94. Schematic of a ring resonator reflector coupled laser
Fig. 5.95. Photograph of a fabricated coupled ring reflector laser. (Photograph

courtesy of Prof. Dapkus3 )
A tunable laser based on a coupled ring reflector laser is fabricated and

demonstrated in Choi et al. (2005). The device is fabricated by using a
buried heterostructure waveguide design in the material system GaInAsP/InP.
A photograph of the device is shown in Fig. 5.95.
3
University of Southern California, Department of Electrical Engineering – Elec-
trophysics, VHE 314, Vivian Hall of Engineering, Los Angeles, CA 90089, USA
5.7 Wavelength Converters 197
Fig. 5.96. Measured tuning characteristics. (Photograph courtesy of Prof. Dapkus)
The radius of the used ring resonators is 130 and 140 µm, giving an FSR for
each of the rings of 0.8 nm and 0.75 nm. The overall FSR is determined to be
12 nm. Tuning of the device is realized through the free carrier injection (FCI)
method which shifts the resonance wavelengths by the free carrier plasma
effect. The tunability of the device is shown in Fig. 5.96. Carriers are injected
into the ring resonator having an FSR of 0.75 nm leading to a shift of the
resonance wavelength of 0.8 nm which corresponds to 100 GHz at a wavelength
of 1,550 nm.
Examples of ring resonator coupled lasers with different architectures have
been presented in this chapter and due to this relative new research field,
an improvement of the device parameters is expected in the coming years.
Technological progress will also add to advancement in device performance.
Ring resonator based devices both active and passive can be used for
wavelength conversion as will be demonstrated in the following section.
5.7 Wavelength Converters

The four wave mixing (FWM) effect can be applied to passive and active ring
resonators, enabling efficient wavelength conversion for all-optical communica-
tion systems. Ring resonators based on GaAs or InP combine both advantages
of a high nonlinear susceptibility and a high linear gain. In contrast to other
active resonant structures like distributed feedback (DFB) lasers or SOAs,
ring resonators enhance both, the optical power of the signal wavelength at
frequency ωS and that of the converted wavelength at ωC . Maximum conver-
sion efficiency is achieved, when ωS and ωC fit one of the resonance wave-
lengths of the ring resonator. No stop-band has to be taken into account.
Operating at wavelengths not being one of the resonance wavelengths of the
ring resonator results in a reduced conversion efficiency. The internal optical
198 5 Devices
field enhancement and the FSR can be designed by the radius and the power
coupling ratio between the ring resonator and the bus waveguides.
A passive GaAs racetrack shaped ring resonator is used in Absil et al.
(2000a) to demonstrate wavelength conversion by four wave mixing with an
improved conversion efficiency compared to a straight waveguide. A photo-
graph of the device is shown in Fig. 5.97. The increase in efficiency is attributed
to the increase of the interaction length and the enhancement of optical
power inside the ring resonator cavity. Optical propagation losses limit so
far the overall performance. A spectrum of the FWM measurement is shown
in Fig. 5.98. The ring resonator used in the experiment has a radius of 4.2 µm.
A coupling strength of 0.27 is obtained for the directional coupler.
Fig. 5.97. Optical micrograph of a racetrack-shaped resonator with defined ports.

Reprinted with permission from P.P. Absil, Optics Letters, vol. 25, no. 8, pp. 554–556
2003
Fig. 5.98. Optical spectra for on-resonance (solid curve) and off-resonance (dotted
curve) cases.
Reprinted with permission from P.P. Absil, Optics Letters, vol. 25, no. 8, pp. 554–556
2003
The properties of an all-optical wavelength converter with reshaping char-

acteristics, based on FWM in a passive GaAs–AlGaAs ring resonator notch
filter are investigated in Mikroulis et al. (2004). A detailed theoretical analysis
of the static and dynamic reshaping characteristics of the wavelength converter
is carried out.
Improved conversion efficiency is expected in ring resonators fabricated as
all-active ring resonators. An active ring resonator architecture used for wave-
length conversion is demonstrated in Hamacher et al. (2003). A photograph
of an electrically contacted ring resonator is shown in Fig. 5.99. The ring res-
onator has a radius of 182 µm including two 160 µm long MMI couplers which
can be electrically driven separately to couple the light from the straight bus
waveguide into the ring and vice versa.
In analogy to FWM in DFB lasers the ring cavity itself can be used as
the pump signal source. A proper choice of the driving currents provides a
sufficient side mode suppression ratio of greater than 35 dB. An ECL acting
as a cw signal source is used in the FWM experiment. The result of this ex-
periment with a cw output power of 2 dBm measured in the feeding fiber is
shown in Fig. 5.100. A conversion efficiency of −5 dB is derived from the mea-
surement of the output power. Taking into account that the input signal will
be increased due to the high power enhancement of the active ring structure,
the conversion efficiency can be corrected by a factor of 8 dB, resulting in an
“effective” conversion efficiency of 3 dB.
Instead of selecting a single mode laser spectrum by using a ring resonator
with a small radius and/or a proper choice of driving currents as mentioned in
the previous example, it is also feasible to obtain a mode selection by injection
locking. If the frequency of the master laser (in the test set-up a second ECL,
amplified by an erbium-doped fiber amplifier) is close enough to the ring
resonator’s free-running frequency and the signal amplitude is large enough,
the device is locked to the master frequency over a certain bandwidth. This
Fig. 5.99. All active ring resonator

200 5 Devices
Fig. 5.100. Optical cw spectrum measured at the throughput port exhibiting a

conversion efficiency of −5 dB (relation of the intensities in the feeding fibers)
approach may be favored for practical use in a real system in order to realize a
well controlled resonator of specified FSR for FWM and to control the pump
signal by injection locking in a separate device (master laser).
To demonstrate the FWM capabilities of an all active device, an all-optical
cw channel switching experiment is performed. In this experiment the cw
signal at ωS is kept fixed at a resonance wavelength of the ring resonator
of 1567.9 nm. For maximum conversion efficiency the injection locked pump
wavelength ωP is chosen to fit one of the resonances of the ring resonator.
When the detuning of the pump wavelength is changed stepwise by multiples
of the FSR, the converted signals at ωC = 2ωP − ωS fit the resonances of the
ring resonator as well. In this way the conversion efficiency can be kept at a
maximum level over a tuning range of multiples of the FSR, which is verified
in the experiment. Due to a high coupling factor of 0.7, the enhancement effect
of the ring is limited and a conversion efficiency of approximately −20 dB is
obtained. The result of the measurement is shown in Fig. 5.101, where λC is
the converted wavelength, λP is the pump wavelength and λS is the signal
wavelength.
Wavelength conversion can also be performed using passive ring resonators
with integrated SOAs (Fig. 5.80) as the pump laser and inserting a signal
wavelength into the bus waveguide placed on a resonance of the ring resonator.
A spectrum of the ring resonator operating as a laser without signal and
pump wavelength is shown in Fig. 5.102. The result of an FWM experiment
using this racetrack shaped ring resonator with integrated SOA is shown in
Fig. 5.103. The ring resonator has a radius of 117 µm. A directional coupler
with a gap of 1 µm and a length of 150 µm is used for coupling into and from
the resonator. The overall FSR is 0.4 nm which corresponds to 50 GHz at a
Fig. 5.101. All optical channel switching experiment in cw operation using injec-
tion locking. The signal wavelength is fixed and the pump wavelength routes the
conjugated signal
Fig. 5.102. Passive ring resonator with integrated SOA in laser operation
202 5 Devices
Fig. 5.103. FWM in a ring resonator with integrated SOA
wavelength of 1,550 nm. A conversion efficiency of approximately −6 dBm is

achieved. The driving current for the SOA is 119 mA. The signal wavelength is
inserted into the ring resonator from an ECL which is amplified by an erbium
doped fiber amplifier (EDFA) to a total signal power of 3 mW.
Wavelength conversion using integrated ring resonators is a relative novel
research direction as it requires either low loss passive ring resonators or active
devices to achieve comparable conversion efficiencies as if using straight SOAs
for example. This chapter proves again the versatility of ring resonator devices.
In the following chapter, an emerging field will be introduced with several
practical examples: optical signal processing using ring resonators.
5.8 Optical Signal Processing

Optical signal processing (OSP) is one of the research areas which are receiv-
ing an increasing amount of interest from the all-optical network community.
In OSP, light is used to influence, switch and control light. OSP has several
advantages compared to electrical signal processing. One of them is the higher
processing speed, which has been demonstrated to be as high as 100 Gbit s−1 .
Elements used for OSP can be fabricated with current state-of-the-art tech-
nologies and are easier to be realized than compared to similar solutions in
electrical signal processing. Several books can be found dealing with the fun-
damentals of OSP.
The focus in this chapter is on integrated ring resonators in optical sig-
nal processing. Ring resonators can be used for realizing all-optical switches,
multiplexers, and logic gates.
5.8 Optical Signal Processing 203
5.8.1 Logic Gates
All-optical logic gates have evolved in the last couple of years and have received
considerable attention. Examples of potential applications in optical signal
processing systems are optical bit-pattern recognition, bit-error rate monitor-
ing, payload separation and all-optical packet addressing.
One of the most obvious applications of ring resonators in connection with
logic gates is the use as optical delay lines. In optical logic gates, control signals
need to be overlapped in time with incoming data signals. As was shown in an
earlier chapter, ring resonators used as all pass filters can address this appli-
cation. Optical delay lines can also serve as an optical buffer to delay signals
until the system is ready to process them. A detailed analysis of all pass filters
used as optical delay lines is given in Lenz et al. (2001). Multiple coupled all
pass filters are investigated theoretically in Azaña and Chen (2002) for optical
signal processing applications, especially for real-time Fourier transformation
(RTFT) and pulse repetition rate multiplication (PRRM).
Optical signal processing using fabricated nonlinear integrated ring reso-
nators is presented in Van et al. (2002a). Applications of the ring resonators
to switching, time-division demultiplexing, pulse routing, and wavelength con-
version (see Sect. 5.7) using four wave mixing are demonstrated. The authors
also propose architectures for realizing OR, AND, and XNOR optical logic
gates with ring resonators. A schematic of the proposed logic gates is shown
in Fig. 5.104.
The used input signals A, B, and C are tuned to resonances of the ring
resonator. Signal C can be regarded as a constant bias to the gate in the case
of asynchronous operation, or as a clock in the case of synchronous operation.
In the case of the OR and AND gate, the signal wavelength C is not detected
at the output D when signals A and B are not inserted. In the XNOR con-
figuration, the MZI is balanced only when signal C is present, leading to a
signal at output D. In all three configurations, a bandpass filter (BPF) only
allows the wavelength λout to pass through.
In the OR configuration, signal C is brought off-resonance by either signal
A or B which can then be detected at output port D. In the AND configura-
tion, both signals A and B need to be present for signal C to be off-resonance
so that it can be detected at output port D. In the case of the XNOR gate,
signal C will be off-resonance either by the presence of signal A or B, induc-
ing a π phase shift in the MZI and thus leading to destructive interference at
output port D which means no signal is detected. A signal at output port D is
detected on the other hand if both signals A and B are inserted into the gate,
signal C is off-resonance, the MZI is balanced and signal C can be detected
at output D. As described in Sect. 5.5, the speed of operation depends on the
cavity ring-down time of the ring resonators.
All optical AND and NAND logic gates are demonstrated experimentally
in Ibrahim et al. (2003b). The nonlinear effect used in the experiment is
the change in refractive index from free carriers generated by two-photon
204 5 Devices
Fig. 5.104. Optical logic gates using ring resonators. Signals A and B are tuned to
resonance wavelength λin ; signals C and D are at resonance wavelength λout ; BPF:
bandpass filter.
Reprinted with permission from IEEE Journal of Selected Topics in Quantum Elec-
tronics, vol. 8, pp. 705–713 2002
c IEEE
Fig. 5.105. Signal paths leading to AND/NAND operation in a racetrack shaped

ring resonator filter
absorption. InP and GaAs-based racetrack shaped notch filters are used. The
InP device has a radius of 10 µm and a 3 µm long straight section. A schematic
of the signal paths is shown in Fig. 5.105. The intensity of the pump beams
A and B are alone not strong enough to induce switching of the probe beam.
The probe signal which is initially tuned to one of the resonance wavelengths
of the ring resonator leading to a “0,” is detected at the receiver when both
input signals A and B are fed into the ring resonator corresponding to a “1.”
When the probe beam is slightly “blue” shifted and out of resonance obtaining
a “1,” NAND gate operation is achieved when both input signals A and B
are fed into the ring resonator leading to a shift of the probe beam into a
resonance wavelength resulting in a “0” at the receiver.
The experiment using the InP device is performed with a mode-locked
laser at 5.6 GHz, externally modulated at 140 MHz. The pump pulse has
a pulsewidth of 35 ps and energy of 20 pJ. The result of the measurement
demonstrating AND logic gate operation is shown in Fig. 5.106.
The GaAs sample, similar to the InP device, exhibits an FSR of 11 nm.
A gain-switched laser diode at 8.4 GHz and externally modulated at 140 MHz,
is used for the data source. The pump pulse energies of 80 pJ per pulse are
inserted in to the input port of the device. The measurement of the device
showing AND/NAND operation is given in Fig. 5.107.
An all optical AND logic gate based on the four wave mixing (FWM) effect
is theoretically analyzed in Mikroulis et al. (2005). A GaInAsP ring resonator
notch filter is used for the calculation with a cavity length of 69 µm.
Optical logic gates realized with ring resonator filters are a promising
approach which has room for improvement in the future as smaller devices
and material configurations are fabricated and investigated.
Fig. 5.106. Time traces showing an AND logic gate using the InP ring resonator:
(a) “A” and (b) “B” are the two input pumps tuned to the resonance at 1,550 nm
and (c) “F = A · B” is the output probe signal tuned to the next higher resonance
at 1,560 nm.
pp. 1422–1424 2003
c IEEE
206 5 Devices
Fig. 5.107. Time traces showing logic operation using the GaAs resonator: (a) and
(b) are the inputs “A” and “B” tuned to the resonance at 1,548 nm; (c) output “F”
when the probe was initially in resonance at 1,559 nm; and (d) output “F” when
the probe was initially blue tuned out of resonance at 1558.6 nm.
pp. 1422–1424 2003
c IEEE
5.8.2 Switching
All optical switches rely on the control of light by light (the principle of optical
signal processing) as was shown in the previous section for logic gates. In order
to induce switching, an optical control signal is used to change the optical
properties of a nonlinear medium. The device, in this case the ring resonator,
then switches the data signal, which experiences the changed transmission
properties when it passes through. In all-optical switches MZIs with SOAs
have been used so far where the SOAs provide the nonlinear media. In the
case of ring resonators, material systems providing nonlinearity are required
which does not necessarily mean to introduce electrically contacted gain into
the cavity. Ring resonators are ideal for optical switching as they have a high
Q-factor and can be monolithically integrated. A high field enhancement and
finesse in the ring can lower the switching threshold significantly.
One of the first experimental demonstrations of all optical switching in
GaAs ring resonators is given in Van et al. (2002b). Switching is accomplished
by a refractive index change in the ring resonator due to free carriers generated
by two-photon absorption. This switches the probe wavelength in and out of

resonance. The radius of the used vertically coupled ring resonator notch filter
is 10 µm. The parameters of the device are as follows: FWHM = 0.16 nm,
Q-factor = 9, 800, F = 70, calculated field enhancement factor = 7.3, calcu-
lated cavity lifetime = 40 ps. The switching experiment is carried out using an
externally modulated laser diode, producing a 300 ps pump pulse signal with
a 20 MHz repetition rate with an average input power of 10 mW. A 3 mW cw
signal is used for the probe wavelength. The pump wavelength is slightly out
of resonance whereas the probe signal is placed on a resonance of the ring
resonator. The result from the experiment is shown in Fig. 5.108. Both pump
and probe beams are inserted at the input port of the ring resonator via a
fiber coupler. After passing through the ring resonator notch filter, the pump
and probe beams are amplified and separated by a bandpass filter. Switching
times of the probe beam of approximately 100 ps is observed.
A similar ring resonator geometry is used in Ibrahim et al. (2002) to
demonstrate all-optical time division demultiplexing and spatial pulse rout-
ing. The radius of the ring resonator add–drop filter is 10 µm. The cavity
lifetime of the ring is measured to be approximately 30 ps. Switching is again
accomplished by a refractive index change induced by two-photon absorption.
Fig. 5.108. Measured time traces of the pump-and-probe beam intensities with
(a) probe beam initially on resonance (low transmission) and (b) Probe beam ini-
tially off-resonance (high transmission). The transmitted probe intensity is normal-
ized with respect to the maximum transmittance at resonance.
pp. 74–76 2002
c IEEE
208 5 Devices
A data pulse inserted into the ring resonator is demultiplexed and collected
at the drop–port of the ring resonator filter. An on–off contrast ratio of 8 dB
is achieved.
A detailed theoretical analysis of the nonlinear transfer functions of ring
resonators is given in Blair et al. (2002). Ring resonators are compared to
straight nonlinear sections of different lengths. Ring resonators are demon-
strated to have a higher nonlinear phase shift induced by linear and two-
photon absorption when compared to nonlinear straight devices. Based on
the simulated results, a design for an MZI with integrated ring resonator is
presented.
One of the first disk resonators with an electro-absorptive active region is
presented in Djordjev et al. (2002f). The disk uses the quantum confined Stark
effect (QCSE) to loss trim the frequency response of the resonator cavity. The
disk vertically coupled to two bus waveguides has a radius of 10 µm, an FSR
of 10.5 nm, a Q-factor of 5,700 and an F of 40. The transmission and dropped
coefficients can be changed by introducing additional loss into the cavity at a
particular wavelength. One advantage of using electro-absorptive ring or disk
resonators is that the light is not absorbed as strongly as it would be in an
MZI configuration. Switching in ring or disk resonators is done by changing the
resonance characteristics which directs the light to the drop–port for example.
All-optical switching in a racetrack shaped laterally coupled ring resonator
add–drop filter by carrier injection induced by single photon absorption is
demonstrated in Ibrahim et al. (2003a). Single photon absorption is achieved
by optically pumping the used material above its bandgap energy, generat-
ing electron–hole pairs, which in turn alter the gain–loss coefficient and the
refractive index. In comparison to two photon absorption, the pump beam
does not need to be located on a resonance of the ring resonator, reducing
the cavity ring-down time leading to a reduction in the switching time. The
parameters of the used ring resonator add–drop filter is: FWHM = 1.3 nm,
Q = 1, 200, FSR = 18 nm, and a finesse of 14. In the experiment, the ring
resonator is pumped from the top with a Ti:Sapphire laser with a wavelength
of 800 nm. A switching window of 20 ps is obtained. Two switching schemes
are demonstrated. In the first, the inserted wavelength is located near a res-
onance causing it to be on a resonance when the ring resonator is pumped.
The signal is then switched to the drop–port. In the other scheme the signal
is tuned to a resonance and by pumping the ring, the signal is switched to
the throughput port. In this way, a modulator is implemented. A switching
contrast of approximately 7 dB is achieved.
A switching time of only 20 ps is demonstrated in Hill et al. (2004) using
integrated ring lasers in GaInAsP/InP having radii of 8 µm. A photograph of
the fabricated device is shown in Fig. 5.109. The width of the waveguides is
2 µm. The lasers are operated at 281 K in pulsed mode, with current pulses of
80 ns duration every 13.5 ms. The threshold current is approximately 30 mA.
The double ring laser device can have two stable states. In one state, light from
ring laser A injection locks ring laser B, forcing it to lase only in the clockwise
5.9 Sensors 209
Fig. 5.109. A memory element formed by two 16 mm diameter microring lasers

coupled via a waveguide on an InP/GaInAsP photonic integrated circuit. Scale bar,
10 mm. The microring lasers were fabricated in active areas of the integrated circuit
containing bulk 1.55 mm bandgap GaInAsP in the light guiding layer. Separate
electrical contacts allowed each laser’s wavelength to be individually tuned by adj-
usting the laser current. Passive waveguides connected the microring lasers to the
integrated circuit edges.
Reprinted by permission from Macmillan Publishers Ltd: T.M. Hill, Nature 432:206–
209 2004
c
direction. In the second state, ring laser B injection locks ring laser A, forcing
it to lase only in the anticlockwise direction. Pulses of light at the chosen
input can set the system in the corresponding state.
Optical signal processing using ring resonators is a useful alternative to
for example MZIs and progress in research on the switching behavior of these
devices will eventually lead to an implementation in all-optical networks.
5.9 Sensors
Integrated optical sensors are and have been investigated since a long period
of time and have found their way into many applications. Ring resonator
can be used as sensing elements for example by measuring the resonance
frequency shift, which is induced by a change in the effective refractive index.
Integrated waveguide based sensors have gained attention since the downturn
of the telecommunication industry left several research groups looking for
other fields of application. But this is not the only cause. The focus of research
has turned more and more towards biology where single molecule and single
cell detection are the envisaged goal.
This chapter provides an overview of sensors based on integrated ring
resonators including biosensors.
A silicon based integrated ring resonator notch filter with a radius of 3 cm
is used in Adar et al. (1991) to measure the loss in phosphorus-doped silica on
silicon waveguides. The finesse of the resonator is 45. The measuring principle
used is the fact that waveguide loss is related to the width of the resonances
210 5 Devices
(see Chap. 2). The ring resonator is placed on a temperature controlled surface.
An ECL kept at a constant wavelength, with a linewidth smaller than the
width of the resonances of the ring resonator is used as the input source. The
temperature is changed and the response of the ring resonator is measured,
revealing the filter function characteristic for ring resonator notch filters. As
the FSR is known from the parameters of the ring resonator, the width of the
resonances can be taken directly from the transmission experiment leading to
the waveguide loss.
An integrated ring resonator sensor is analyzed theoretically in Kiyat et al.
(2004) to be implemented in a setup to measure the displacement in scanning
probe microscopy. A schematic of the proposed device is shown in Fig. 5.110.
The displacement of the cantilever changes the refractive index of the ring
resonator which changes the transmission spectrum. Several designs including
racetrack-shaped ring resonators are simulated and compared. The highest
sensitivity is obtained for a racetrack-shaped ring resonator. Sensitivities as
−1
high as 10−4 Å are calculated.
Waveguide birefringence is measured in Carriere et al. (2004) by using an
integrated ring resonator with a radius of 14 mm, fabricated by ion exchange
in glass. The waveguide birefringence is defined as the difference between
the effective refractive indexes of transverse electric and transverse magnetic
modes. The birefringence ∆nBi is calculated using the FSR of a ring resonator
as follows (Carriere et al. 2004). Using (2.21) where the group refractive index
is used, the shift in wavelength between the two polarization states seen in
the FSRs is given by

λ2 1 1
∆FSRBi = FSRTM − FSRTE = − . (5.16)
L ng TM ng TE
The waveguide birefringence is defined as
∆nBi = neff,TE − neff, TM . (5.17)
Fig. 5.110. Integrated microring resonator displacement sensor for scanning probe
microscopies
5.9 Sensors 211
Using (2.20), the waveguide birefringence in terms of the FSRs is given by

λ2 ∆FSRBi
∆nBi = · . (5.18)
L FSR (FSR − ∆FSRBi )
This equation is straight forward and the birefringence is easily calculated
from the measurement of the FSRs. In order to differentiate the birefringence
induced by the curved sections of a ring resonator, racetrack-shaped and full
circle resonators can be compared in terms of their birefringence. The differ-
ence is only due to the contribution of the straight waveguide sections.
Ring resonators are not only used for measuring birefringence in wave-
guides, but can also be applied to determine strain. In Bhola et al. (2005)
a polymer ring resonator strain sensor is demonstrated. A photograph and a
cross-section of the fabricated device are shown in Fig. 5.111. A sketch and
a photograph of the strain sensor attached to a micrometer stage are shown
in Fig. 5.112. An advantage of using ring resonators as strain sensors is that
the direction of the strain does not matter due to the ring geometry. The
sensitivity can be increased by using a racetrack shaped ring resonator. In
this case two racetrack-shaped ring resonators perpendicular to one another
can be used to determine the direction of the strain.
These examples demonstrate a small variety of sensors realized with inte-
grated ring resonators in different materials. Lately biosensors have attracted
a lot of attention and ring resonators are again focus of research in this new
emerging field.
A biosensor based on a disk resonator is proposed and analyzed in Boyd
and Heebner (2001). The disk is weakly coupled to a bus waveguide which
allows the field to build up to high values in the disk resonator which is
necessary for sensing. The device designed has the following parameters:
a waveguide width of 0.4 µm, a coupling gap of 0.25 µm and a disk radius
of 2.54 µm. The device is operated in TM polarization. A dielectric constant
of the device of 4 is assumed. The achieved simulated sensitivity is expected
to be approximately 100 molecules.
Fig. 5.111. Cross-section and photograph of the polymer ring resonator used for
the strain sensor. Dimensions are in µm.
pp. 867–869 2005
c IEEE
212 5 Devices
Fig. 5.112. Application of controlled strain using a micrometer stage. ε is the strain
and R the radius of the bend.
pp. 867–869 2005
c IEEE
A detailed analysis of resonant-enhanced evanescent-wave fluorescence

biosensing in ring and disk resonators is presented in Blair and Chen (2001).
Design layouts are given and the sensitivity and the enhancement factor are
derived.
A sensor using a disk resonator vertically coupled to a bus waveguides is
fabricated and demonstrated in Krioukov et al. (2002). The layer sequence
and layout of the device is shown in Fig. 5.113. Similar to the previous simu-
lated example, the disk resonator is weakly coupled to the underlying bus
waveguide, allowing modes in resonance with the disk to build up. The re-
sponse from the disk is measured using water, ethanol and a mixture of water
and glucose in different concentrations as cladding. A shift in the resonance
wavelength indicates the different solutions.
A glucose concentration sensor is realized in Chao and Guo (2003) us-
ing polymer nanoimprinting. The racetrack shaped ring resonator notch fil-
ter includes partially reflecting junctions in the bus waveguide creating so
called Fano resonances (Fan 2002). A schematic of the devices is shown in
Fig. 5.114. Details on the fabrication have been described in Chap. 3 where
a similar device is realized. The measurement is performed by tuning a laser
near one of the resonances of the ring and detecting the change in the spectra.
The ring resonator is placed in different glucose concentrations. A minimum
detectable concentration change of glucose solution is estimated to be 0.024%,
or 24 mg (dl)−1 . Since the Q-factor is proportional to the sensitivity, an in-
crease in the Q-factor of the ring resonator will also enhance the sensitivity.
The shift in the resonance wavelength is linear with increasing glucose concen-
tration.
So far solids or liquids are detected with ring resonators. A gas sensor
is fabricated and demonstrated in Ksendzov et al. (2004) which utilizes a
polymer sensing layer. The racetrack-shaped ring resonator add–drop filter
has a radius of 1.5 mm. The straight sections have a length of 0.1 mm with a
5.9 Sensors 213
Fig. 5.113. Cross-section and view of an integrated optical MC sensor. Light from a
tunable laser is coupled into the excitation waveguide, and the scattering is measured
from the top of the MC as a function of the wavelength with various solutions
in the sensing area. Dotted line, localization of a high-Q whispering-gallery mode;
its propagation direction is shown by an arrow around the circumference. Device
parameters: R = 15 µm, hc = 255 nm, he = 150 mm.
Reprinted with permission from E. Krioukov, Optics Letters, vol. 27, no. 7, pp. 512–
514 2002
Fig. 5.114. Racetrack-shaped ring resonator with partially reflecting waveguide

junctions
214 5 Devices
Fig. 5.115. Layer sequence of waveguide in a ring resonator gas sensor
coupling gap of 2 µm. The layer sequence of the waveguide used is shown in
Fig. 5.115.
The polymer swells in the presence of the analyte and changes its refrac-
tive index due to chemical permeating. The resonance wavelength is monitored
with a tunable laser and the shift is detected in dependence of the concen-
tration of the analyte. The fabricated ring resonator is brought into a sealed
chamber and a controlled mixture of air, water vapor, and isopropanol is in-
serted. Sensitivity to isopropanol of 50 ppm is shown experimentally.
One of the first demonstrations of a ring resonator used for sensing bio-
molecules is given in Yalçin et al. (2006). The functionality of the device is
proven by using avidin–biotin binding on the surface of a vertically coupled
glass ring resonator. The radius of the ring resonator is 60 µm. The penetra-
tion depth of the mode into the medium (deionized water) is calculated to
be approximately 360 nm at a wavelength of 1,550 nm assuming an effective
refractive index of the waveguide of 1.5. The ring resonator parameters are
as follows: FSR = 4.2 nm, FWHM = 0.126 nm, F = 33, Q = 12, 000. The
measurement principle is again the monitoring of the resonance shift due to
a change in the refractive index caused by the binding of biomolecules. The
refractive index sensitivity of the system is 1.8 × 10−5 .
Ring resonators are ideal for use in sensing biomolecules or even cells,
as only a few molecules are required to induce a change in the resonance
wavelength. Going into the direction of single molecule detection requires so
called whispering gallery mode resonators which will be briefly described in
the following chapter.
6
Whispering Gallery Mode Devices
Whispering gallery mode (WGM) resonators and lately photonic crystal

resonator based devices are attractive due to their potential applications in
photonics and recently in biophotonics as is described in Chap. 5. Applications
which have been realized include for example lasers, filters, optical switches,
and biosensors. This chapter introduces WGM devices, although some exam-
ples which have been given in this book have dimensions of whispering gallery
mode resonators. This chapter serves as a brief introduction into the topic as
a significant number of books and other publications exist on this subject (for
example, Vahala 2003, 2004).
6.1 Whispering Gallery Modes

WGMs are named after the whispering gallery at St. Paul’s Cathedral in
London and have been studied by Lord Rayleigh in 1910 (Rayleigh 1910,
1912). WGMs occur at particular resonant wavelengths depending on the
size of the resonator. At these particular wavelengths, light undergoes total
internal reflection at the surface of the resonator creating resonances in the
transmission spectrum of the resonator. Whispering gallery resonators are
dielectric structures in which light waves are confined by continuous total
internal reflection. WGMs can be observed in droplets and glass spheres for
example. In this chapter, examples are limited to planar integrated WGM
resonators.
The light in WGM resonators is concentrated near the circumference of the
device and is assigned a mode number and a mode order. The mode number,
mn , reveals the number of wavelengths around the circumference, and the
mode order, mo , reveals the number of maxima in the radial dependence of
the electromagnetic field within the resonator. These modes can have very
high Q factors if the losses due to reflection at the surface of the resonator
are low. In addition to high Q-values, a high finesse is also obtained. The
high Q-factor makes these resonators very attractive to be used as lasers.
216 6 Whispering Gallery Mode Devices
In order to couple light in or out of these resonators, the evanescent field

of the WGMs has to overlap with the evanescent field of a phase-matched
optical bus waveguide. This in-and-out coupling of WGM resonators is one of
the major technological challenges and was only recently achieved due to the
advance of fabrication technology, as was demonstrated in several examples
in this book.
A review of the basic properties of WGM resonators is presented in Matsko
and Ilchenko (2006a).
6.2 WGM Filters

Optical filters are one of the first applications when considering optical dev-
ices having wavelength dependent resonances. As in ring resonators, WGM
resonators have a wavelength dependent resonance spectrum. Due to the small
size of WGM resonators, they have large FSRs, high Q-factors and high finesse.
A review paper presenting applications of both WGM and ring resonators is
given in Matsko and Ilchenko (2006b).
One of the challenging research tasks is to achieve in and output coupling of
these very small (diameter<10 µm) WGM resonators. Several groups have and
are working on this subject. Before being able to couple in and out of a WGM
device, WGM modes must circulate in the device to build up a resonance
spectrum. This is only achieved when the WGM device has sufficiently low
internal losses, especially a low surface roughness. Both of these aspects are
theoretically investigated in Little and Chu (1996). The coupling between a
WGM disk resonator and a bus waveguide is simulated in Dehmubed and
Diament (1998).
Design and fabrication of laterally coupled ring and WGM disk resonators
in the material system AlGaAs/GaAs is presented in Hagness et al. (1997),
Rafizadeh et al. (1998). The layer sequence of the device has been described
in Chap. 3. Photographs of the fabricated ring and WGM disk resonator are
shown in Fig. 6.1. The diameter of the ring and disk resonator is 10.5 µm.
A waveguide width of 0.5 µm is used. The coupling gap between the bus
waveguide and the ring and disk is 0.1 µm. A trench is fabricated around the
waveguides with a width of 1 µm.
A detailed description of cylindrical WGM multimode resonators verified
by experimental results realized in silicon oxynitride technology is presented
in Klunder et al. (2002a, 2003a). The layer sequence of the fabricated cylin-
drical WGM resonator is shown in Fig. 6.2. The intensity distribution inside
the WGM resonator is analyzed using photon scanning tunneling microscope
(PSTM) and effects like polarization conversion and counter propagation is
detected. A simulation model is derived for calculating the finesse, bending
losses and FSRs of fundamental quasi guided modes in this cylindrical WGM
resonator.
6.2 WGM Filters 217
Fig. 6.1. SEM images of a 10.5 µm diameter GaAs/AlGaAs microcavity ring and
disk resonator coupled to 0.5 µm-wide waveguides.
pp. 2154–2165 1997
c IEEE
Fig. 6.2. Layer sequence of cylindrical WGM resonator
Another form of WGM resonator is designed and analyzed in detail in

Maolatou et al. (1999) and Hammer (2002). The WGM resonator is in the
shape of a rectangle coupled to bus waveguides. A schematic of the rectan-
gular WGM resonator is shown in Fig. 6.3. Rectangular-shaped resonators
are slightly different from ring or disk resonators treated so far in this book
with respect to the intensity distribution at the output ports. As can be
seen from Fig. 6.3, some of the input power is reflected back to the same
port and dropped at all the remaining ports. This is in contrast to ring or
disk resonators, where there is a preferred direction of the light passing the
resonator in the on or off-resonance state, leading to only one drop and one
throughput port.
Fig. 6.3. Rectangular WGM resonator
Fig. 6.4. Grating assisted rectangular WGM resonator
Several configurations have been simulated in Hammer (2002) using an

operation wavelength of 1.55 µm. The length L and width W of the rectangu-
lar WGM resonator is varied between 0.81–6.902 µm and 0.81 µm–3.696 µm,
respectively. The width of the bus waveguides and the coupling gap is varied
between 0.035–0.13 µm and 0.32–0.49 µm, respectively. The refractive index
of the WGM resonator and the bus waveguides is assumed to be 3.4, whereas
the surrounding refractive index is defined to be 1.45.
This type of resonator is coupled to gratings on either side of the WGM
resonator which leads to a superior add–drop filter characteristic. The device
design is presented in Hammer et al. (2004). A schematic of the proposed
device is shown in Fig. 6.4.
The fabrication of WGM resonators has only been possible recently due
to an advance in nanofabrication technologies. Especially in the case of WGM
resonator filters, in and out coupling is the greatest technological challenge.
When these tasks are overcome, tunability of WGM resonators is the next
goal.
Tuning of WGM resonators is addressed in Boriskina et al. (2003). Filter
tuning is proposed by changing the shape of the resonator from circular to
elliptic, which reduces the dependence of the coupling efficiency on the width
of the air gap. The elliptic shape can be compared to a racetrack shaped ring
resonator, where the effect in increasing the coupling length is the same. Two
alternative coupling designs are suggested for monolithic and hybrid integra-
tion technologies. The first one is based on increasing the coupling length. The
other makes use of the field concentration at the increased curvature portion
of the deformed WGM resonator.
Another method of tunability (either wavelength or filter shape) in res-
onators is by introducing gain into the resonator as described in Chaps. 3
and 5. Active disk resonators are simulated and designed in Djordjev et al.
(2002a) for use as modulators, switches and wavelength routers.
6.3 WGM Lasers 219
Totally active devices can of course be used as lasers. WGM lasers have
been investigated and demonstrated earlier than WGM resonator filters, which
is obvious, because light only needs to be collected from the WGM resonator
and not coupled into it for detecting the spectrum. In the following chapter,
a brief introduction is given with examples of fabricated WGM lasers.
6.3 WGM Lasers

WGM lasers are expected to be the light sources of future photonic inte-
grated circuits and optical computer boards due to their extremely small size.
Other advantages are similar to ring and disk resonators and include their
compactness and their integration ability. Further more, no grating or facet
is required for optical feedback which is the main advantage when thinking
of several WGM lasers on a chip placed at different locations. Some examples
described in Chap. 5.6.1 can be regarded as WGM resonator lasers and are
therefore not addressed in this chapter.
WGM lasers having diameters of 3, 5, and 10 µm are fabricated in the mate-
rial system GaInAsP/InP and demonstrated in McCall et al. (1992). Optically
pumped InGaAs quantum wells are used to achieve lasing at wavelengths of
1.3 and 1.5 µm. The WGM lasers have the shape of a “mushroom” with an
InP pedestal holding the disk with the quantum well layers. The thickness of
the InGaAs quantum wells is 10 nm, the thickness of the GaInAsP barriers is
10 nm as well. The GaInAsP end caps have a thickness of 20 nm. The mush-
room like shape is achieved by selective chemical etching of the InP below and
around the quantum wells, resulting in a disk on a pedestal.
Even smaller WGM disk lasers are presented in Slusher et al. (1993). Disks
with a diameter of 2.2 µm and 5 µm in the same material as the previous ones
are fabricated. Six quantum wells with a thickness of 10 nm of the material
In0.53 Ga0.47 As are used together with five 10 nm thick GaInAsP barriers. The
bandgap wavelength of the GaInAsP material is 1.1 eV at room temperature.
The disks are optically pumped with an HeNe laser at a wavelength of 633 nm.
A mode spacing of 87 nm is obtained for the small disk and 37 nm for the
larger disk. Q-values of 260 and 500 are measured at pump powers below
threshold.
A “submicron” WGM disk laser is presented in Levi (1994) having a radius
of 0.8 µm and a thickness of 0.18 µm. The disk consists of six 12 nm thick
InGaAs quantum wells separated by five 12 nm thick GaInAsP barriers. The
disk is optically pumped with a diode laser having a wavelength of 850 nm.
Pulses with a length of 8 ns and a repetition period of 100 ns are used for
the pumping. A conformal mapping technique is used in Frateschi and Levi
(1995) to determine the radial and azimuthal eigenvalues and eigenvectors of
leaky optical modes of this WGM disk laser.
Quantum cascade (QC) WGM disk lasers with an emission wavelength at
5 µm are fabricated and demonstrated in Faist et al. (1996). A mushroom like
device structure is again used. A photograph of a fabricated QC disk laser is

shown in Fig. 6.5. The layer sequence of the device is shown in Fig. 6.6.
Disks with diameters of 17, 25, 45, 80, and 125 µm are fabricated. The
measurement of the devices is done at a temperature of 15 K. The disks are
contacted from top and are operated using 50 ns pulses. A threshold cur-
rent of 2.85 mA is measured for the QC disk with a diameter of 17 µm. QC
WGM disk lasers in the material system InGaAs/AlInAs, emitting at a wave-
length of 16 µm, being one of the longest wavelengths reported for disk lasers,
are presented in Tredicucci et al. (2000). The disk lasers are electrically con-
tacted and operated in pulsed mode. Measurements are performed at temper-
atures of 10 K.
Electrically pumped GaInAsP–InP strained quantum well WGM disk inj-
ection lasers with diameter between 2 and 10 µm are presented in Baba (1997)
and Baba et al. (1997). A photograph of WGM disk lasers with diameters of
8, 3, and 2 µm are shown in Fig. 6.7.
The disks are made of compressive strained multiple quantum wells. Four
quantum well layers with a thickness of 3–4 nm sandwiched between barrier
layers with a bandgap wavelength of 1.2 µm having a thickness of 10 nm. The
Fig. 6.5. Scanning electron microscope image of a 25 µm diameter QC disk laser

with total height 15.5 µm.
Reused with permission from J. Faist, Applied Physics Letters, 69, 2456 (1996)
1996
Fig. 6.6. Layer sequence of the QC WGM disk laser

6.3 WGM Lasers 221
Fig. 6.7. WGM disk injection lasers with diameters of 8, 3, and 2 µm (from left to
right).
tronics, vol. 3, pp. 808–830 1997
c IEEE
active layers include 50 nm thick guide layers with a bandgap wavelength of

1.2 µm and 30 nm thick gradient index (GRIN) layers with a bandgap wave-
length of 1.1 µm. The total thickness of the active layers is 0.2 µm. A thresh-
old current of 0.2 mA is measured for the disk with a diameter of 2 µm under
pulsed operation at a temperature of 286 K. Lasing is achieved at wavelengths
around 1.55 µm.
Optically pumped disk and ring lasers in InGaAs with diameters between
3.5 and 9 µm, fabricated using substrate removal and material transfer to a
glass substrate by liftoff processing are presented in Corbett et al. (1996).
This process is used in Corbett (1998) to fabricate ring and disk lasers
with diameters between 6 and 28 µm having a single compressively strained
3 nm, In0.7 Ga0.3 As quantum well cladded by a 20 nm GaInAsP layer hav-
ing a bandgap wavelength of 1.25 µm and a 100 nm GaInAsP layer with a
bandgap wavelength of 1.15 µm, lattice matched to InP with a total thickness
of 245 nm. The devices are optically pumped continuous wave (cw) with the
514 nm line of an Ar-ion laser at liquid nitrogen temperatures. An optically
pumped threshold of 2 µW is obtained. The lasing wavelength observed is
between 1.4 and 1.55 µm. Multimode operation of the lasers is observed at
wavelengths corresponding to whispering gallery modes.
Determining the optical modes in WGM resonators requires a different
approach as the one presented in Chap. 2 for calculating the transmission
spectrum of ring resonator filters. As these WGM resonators can be extremely
small as is demonstrated in this chapter, the three dimensional Maxwell equa-
tions can not be reduced to a one dimensional problem. A method for calcu-
lating optical modes in WGM disk resonators giving an eigenvalue equation
for optical modes and providing the distribution of the optical field in exact
forms is described in detail in Wang and Dumitrescu (1998). Based on the
results of the optical modes, the emitted optical modes are determined. Disks
with a radius between 0.5 and 1.5 µm having a thickness of 50, 100 nm and
150 nm are simulated.
Fig. 6.8. Schematic of a WGM disk injection laser.

tronics, vol. 5, pp. 673–681 1999
c IEEE
One of the first cw operating, current injection WGM disk lasers in the
material system GaInAsP/InP is presented in Fujita et al. (1999). A schematic
of a WGM disk injection laser is shown in Fig. 6.8. Whispering gallery modes
travel round near the edge of the disk by total internal reflection at the semi-
conductor/air boundary. The current is flowed from the top contact. Carriers
are injected into the active layer from the posts.
A record threshold value of 150 µA is measured for a disk with a diameter
of 3 µm at a temperature of 298 K. The measurement curve is shown in Fig. 6.9.
Two resonant modes are seen at a wavelength of 1.63 µm. The lasing mode
has an FWHM of 0.5 nm. The nonlasing mode has an FWHM of 0.7 nm. The
calculated Q factors for both of them are 3,300 and 2,300, respectively.
WGM resonators have not only been realized in GaAs or GaInAsP, but
have also been demonstrated in the material system InGaN/GaN using mul-
tiple quantum wells (MQW). An example of realized WGM ring and disk
resonator devices is described in Zeng et al. (1999). Both realized devices are
optically pumped. A schematic diagram of the InGaN/GaN MQW WGM ring
resonators is shown in Fig. 6.10. The response from a WGM disk resonator is
shown in Fig. 6.11.
Different fabrication techniques and material systems have been used to
fabricate WGM disk and ring resonators. A wafer fusion method is described
in Song et al. (2000) to fabricate optically pumped GaInAsP disk lasers on
Alx Oy . Continuous wave room temperature operation is achieved for devices
with a diameter of 2.2 µm. The output power and the spectrum of an optically
pumped disk are shown in Fig. 6.12. The FSR of adjacent whispering gallery
modes is 110 nm. The threshold pump power is 1.13 mW, as can be seen in
the measurement.
6.3 WGM Lasers 223
Fig. 6.9. Laser mode peak intensity versus current characteristic and lasing spectra
observed under cw condition at 298 K for a 3 µm diameter device.
tronics, vol. 5, pp. 673–681 1999
c IEEE
Fig. 6.10. Schematic diagram of representative InGaN/GaN MQW microring fab-

ricated by photolithography patterning and dry etching.
Reused with permission from K.C. Zeng, Applied Physics Letters, 75, 2563 (1999)
1999
So far, WGM disk and ring lasers are measured by placing an optical
fiber near the edge of the disk or ring to collect the emission which is then
fed to an optical spectrum analyzer. In addition to the collection of the out-
put light, cooling of some devices and measuring in pulsed mode is required.
One of the first disk lasers with a radius of 10 µm vertically coupled to
underlying bus waveguides in the material system GaInAsP/InP is presented
Fig. 6.11. Optical emission spectrum of an InGaN/GaN MQW microdisk with a

diameter of 9.4 mm. Both the WG and radial modes are observed in the microdisks.
Reused with permission from K.C. Zeng, Applied Physics Letters, 75, 2563 (1999)
1999
Fig. 6.12. Output power against incident pump power for the 2.2 µm diameter
microdisk laser pumped by 1,064 nm Nd:YAG laser. The spectra are measured by
optical spectrum analyzer (OSA) with 0.24 nm resolution.
pp. 954–956 2000
c IEEE
6.3 WGM Lasers 225
Fig. 6.13. SEM image (left) and schematic cross-sectional view of the microdisk
vertically coupled to a bus waveguide. Picture was taken after the disk formation.
A thin InP membrane remains along the edges.
pp. 1330–1332 2003
c IEEE
Fig. 6.14. Lasing spectrum for an 8 µm radius microdisk laser measured at

I = 14 mA.
pp. 1330–1332 2003
c IEEE
in Choi et al. (2002a). The device operates cw at room temperature, has a side
mode suppression ratio of 20 dB and emits at a wavelength around 1.52 µm.
The layer sequence and the fabrication of the device are similar to the ones
described in Chap. 3. A disk laser based on the same fabrication process,
vertically coupled to a single bus waveguide with a radius of 8 µm is demon-
strated in Choi et al. (2003). A photograph of a vertically coupled disk and a
schematic of the cross-section are shown in Fig. 6.13.
The width of the bus waveguide is 0.7 µm. The waveguide is antireflection
coated to avoid Fabry–Perot resonances. A spectrum of the disk laser at an
operating current of 14 mA is shown in Fig. 6.14. The disk laser has a threshold
current of 4.5 mA. The FSR of the disk is 13 nm and the side mode suppression
ratio is greater then 30 dB as can be seen in the diagram. Q-factors between
5,000 and 6,000 are measured for this device. These disk lasers are used in
Fig. 6.15. Micrograph showing the top view of the fabricated eight-channel micro-
disk array vertically coupled to a single bus waveguide.
pp. 356–358 2004
c IEEE
Fig. 6.16. The microdisks for channels 6–8 are pumped at I = 13 mA, while the
current injection levels for the other disks are maintained at I = 10 mA.
pp. 356–358 2004
c IEEE
Choi et al. (2004c) to fabricate an eight channel laser array. A photograph of

the device is shown in Fig. 6.15.
The radius of the disks varies from 10.6 to 10.95 µm with 0.05 µm difference
leading to spectral spacing of the channels of 1.6 nm. The FSR of the disks is
approximately 9 nm. The bus waveguide is AR coated. The output spectrum
of the laser array is shown in Fig. 6.16. An FWHM of 0.25 nm and a side mode
6.3 WGM Lasers 227
suppression ratio of 20 dB are obtained. Different current levels are used to

tune the resonance wavelength to realize a channel spacing of 1.6 nm.
With increasing advance in nanofabrication and modeling, photonic crys-
tal based devices have made it into real devices. Many applications of ring
resonators presented in this book can be transferred to photonic crystal struc-
tures and some have already been demonstrated and described in several books
and journal publications. Therefore only a few examples of photonic crystal
based lasers will be described in the following paragraph to show the potential
and possible future direction.
An optically pumped two-dimensional (2D) photonic crystal (PhC) laser in
the material system GaInAsP/InP is fabricated and demonstrated in Painter
et al. (1999). A schematic of the cross section of the 2D PhC laser is shown
in Fig. 6.17. Four 0.85 % compressively strained quaternary quantum wells
with an emission wavelength at 1.55 µm are used as the gain source. The
membrane has a thickness of 211 nm and a width of approximately 8 µm.
The lattice spacing is 500 nm. The holes have a radius of 160 nm. An SEM
photograph showing a top view of the fabricated device is shown in Fig. 6.18.
For measuring the properties of the cavity, a semiconductor laser with an
emission wavelength of 830 nm is used. The measurement is performed under
pulsed operation of the pump source. Lasing at a wavelength of 1,580 nm is
observed. The absorbed threshold pump power of the PhC laser is estimated to
be 500 µW. An upper limit for the Q-factor is derived to be approximately 600.
A similar layout is used in Kim et al. (2002) for fabricating an optically
pumped hexagonal PhC ring laser in GaInAsP/InP emitting around 1.55 µm.
A photograph of the device is shown in Fig. 6.19.
Fig. 6.17. Illustration of a cross-section through the middle of the photonic crystal
microcavity. Photons are localized to the defect region by TIR at the air/slab inter-
face and by Bragg reflection from the 2-D photonic crystal.
pp. 2082–2088 2004
c IEEE
Fig. 6.18. Top view of a microfabricated 2D hexagonal array of air holes with a
single central hole missing. The interhole spacing, a, is 500 nm, and the radius of
the holes are approximately 160 nm.
pp. 2082–2088 2004
c IEEE
Fig. 6.19. The scanning electron microscope image of a fabricated laser sample.
The photonic crystal ring cavity consists of six waveguides and six 120◦ bends.
Thus the resonant mode has similar properties to the guided mode of photonic
crystal waveguides. In this picture, lattice constant a is ∼ 0.57 mm and the air hole
radius r is ∼ 0.36a.
Reused with permission from Se-Heon Kim, Applied Physics Letters, 81, 2499 (2002)
2002
6.3 WGM Lasers 229
Fig. 6.20. Experimentally measured laser mode using a CCD. This laser sample
has 15 air holes along the ring diameter. Lattice constant a of this laser sample is
∼0.58 mm and the air hole radius r is ∼0.41a. A hexagonal ring mode pattern shows
that the laser is emitted mainly from the ring region.
Reused with permission from Se-Heon Kim, Applied Physics Letters, 81, 2499 (2002)
2002
The diameter of the device is 8 µm. The measurement of the device is per-
formed using a pump laser with a wavelength of 980 nm at room temperature
with pulses having a width of 10 ns. A near field photograph of the standing
wave mode in the PhC ring laser is shown in Fig. 6.20.
The threshold pump power of the laser is measured to be 3 mW. The
wavelength of the lasing mode observed is ∼1, 625 nm. A Q-factor of more
than 2,000 is obtained for this PhC ring laser. A detailed analysis of the
resonant modes in 2D PhC lasers is presented in Park et al. (2002). Simulations
are compared with two types of fabricated devices, air-based free standing
(similar to the one in Fig. 6.17) and SiO2 -based epoxy bonded structures.
These exemplary examples demonstrate novel types of devices which have a
great potential for the future in photonic integrated circuits. Compared to the
history of disk or ring resonator structures, active PhC resonator structures
are on the way to being electrically contacted with the ability to connect to
the outside fiber world using integrated PhC bus waveguides.
7
Outlook
Ring resonators are ideal candidates for realizing all devices ranging from
optical filters to lasers for future all optical telecommunication networks.
They can be integrated with other semiconductor based devices even elec-
tronics and are thus being in the focus of current and future research.
The increasing knowledge of ring resonator devices and the fast improv-
ing state-of-the-art in manufacturing technologies lead to novel devices with
greater flexibility and higher level of integration. Miniaturization and nanofab-
rication technologies have already led to multiple coupled ring resonators and
integrated gain. Enhancement of device performance will be the next step
in ring resonator research, as functionality has been demonstrated as shown
by several examples in Chap. 5. Parallel to semiconductor processing, polymer
materials, and polymer manufacturing technologies like embossing or nanoim-
printing will play an increasing role also due to the “bio” component of the
current worldwide research direction.
Regarding the future of optical devices, WGM resonators as is pointed
out in Chap. 6, have been developed and characterized since quite some time
and their performance parameters are improving with every new fabricated
device. These kinds of miniature devices will eventually be able to form the
key devices for all-optical computers or even quantum computers.
A key question which remains and which is of a general nature in every
optical device is the connection to the outside world, meaning the optical fiber.
Here new waveguide designs and fabrication technologies will be the answer.
Ring resonators will certainly not loose their attractiveness as their versa-
tility will increase rather than decrease which is also demonstrated by several
examples given in this book. Recently ring resonators have proven their ver-
satility in sensors and even biosensors as have been described in this book
opening up new fields of applications. Ring resonators used as biosensors are
an emerging class of new devices with increasing interest.
In summary, one could say that ring resonators are ever attracting devices
which never loose their fascination once beginning resonating.
References
Absil, P.P.: Microring resonators for wavelength division multiplexing and

integrated photonics applications. Ph.D. Thesis, University of Maryland,
College Park (2000)
Absil, P.P., Hryniewicz, J.V., Little, B.E., Cho, P.S., Wilson, R.A.,
Joneckis, G., Ho, P.T.: Wavelength conversion in GaAs micro-ring reso-
nators. Opt. Lett. 25, 554–556 (2000a)
Absil, P.P., Hryniewicz, J.V., Little, B.E., Wilson, R.A., Joneckis, L.G.,
Ho, P.T.: Compact microring notch filters. IEEE Photon. Technol. Lett.
12, 398–400 (2000b)
Absil, P.P., Hryniewicz, J.V., Little, B.E., Johnson, F.G., Ritter, K.J.,
Ho, P.T.: Vertically coupled microring resonators using polymer wafer bond-
ing. IEEE Photon. Technol. Lett. 13, 49–51 (2001)
Adar, R., Shani, Y., Henry, C.H., Kistler, R.C., Blonder, G.E., Olsson, N.A.:
Measurement of very low-loss silica on silicon waveguides with a ring reso-
nator. Appl. Phys. Lett. 58, 444–445 (1991)
Alexe, M., Gösele, U. (eds.): Wafer Bonding. Applications and Technology.
Springer, Berlin Heidelberg New York (2004)
Azaña, J., Chen, L.R.: Multiwavelength optical signal processing using multi-
stage ring resonators. IEEE Photon. Technol. Lett. 14, 654–656 (2002)
Baba, T.: Photonic crystals and microdisk cavities based on GaInAsP–InP
system. IEEE J. Select. Top. Quantum Electron. 3, 808–830 (1997)
Baba, T., Fujita, M., Sakai, A., Kihara, M., Watanabe, R.: Lasing characteris-
tics of GaInAsP–InP strained quantum-well microdisk injection lasers with
diameter of 2–10 µm. IEEE Photon. Technol. Lett. 9, 878–880 (1997)
Bach, L., Reithmaier, J.P., Forchel, A., Gentner, J.L., Goldstein, L.: Wave-
length stabilized single-mode lasers by coupled micro-square resonators.
IEEE Photon. Technol. Lett. 15, 377–379 (2003)
Barwicz, T., Popovic, M.A., Rakich, P.T., Watts, M.R., Haus, H.A., Ippen,
E.P., Smith, H.I.: Microring-resonator-based add–drop filters in SiN: fabri-
cation and analysis. Opt. Express 12, 1437–1442 (2004)
234 References
Bernhard, W.: Digitale thermo-optische Schalter mit integrierten Modula-

toren zur Erhoehung der Uebersprechdaempfung. Ph.D. Thesis, Technical
University, Berlin (2002)
Bhola, B., Song, H.C., Tazawa, H., Steier, W.H.: Polymer microresonator
strain sensors. IEEE Photon. Technol. Lett. 14, 867–869 (2005)
Bian, Z., Liu, B., Shakouri, A.: InP-based passive ring-resonator-coupled
lasers. IEEE J. Quantum Electron. 39, 859–865 (2003)
Bian, Z., He, W.J., Rabus, D.G., Shakouri, A.: A wavelength-tunable mono-
lithically integrated double ring resonator coupled laser. Proc. CLEO (2006)
Bienstman, P., Six, E., Roelens, A., Vanwolleghem, M., Baets, R.: Calculation
of bending losses in dielectric waveguides using eigenmode expansion and
perfectly matched layers. IEEE Photon. Technol. Lett. 14, 164–166 (2002)
Blair, S., Chen, Y.: Resonant-enhanced evanescent-wave fluorescence biosens-
ing with cylindrical optical cavities. Appl. Opt. 40, 570–582 (2001)
Blair, S., Heebner, J.E., Boyd, R.W.: Beyond the absorption-limited nonlinear
phase shift with microring resonators. Opt. Lett. 27, 357–359 (2002)
Boriskina, S.V., Benson, T.M., Sewell, P., Nosich, A.I.: Tuning of elliptic
whispering-gallery-mode microdisk waveguide filters. IEEE J. Lightwave
Technol. 21, 1987–1995 (2003)
Boyd, R.W., Heebner, J.E.: Sensitive disk resonator photonic biosensor. Appl.
Opt. 40, 5742–5747 (2001)
Burton, R.S., Schlesinger, T.E., Munowitz, M.: Improved Y-junction splitter
for ring waveguides. Electron. Lett. 30, 956–957 (1994)
Butterworth, S.: On the Theory of Filter Amplifiers. Experimental Wireless
and the Radio Engineer 7, 536–541 (1930)
Cao, H., Liu, C., Ling, H., Deng, H., Benavidez, M., Smagley, V.A., Caldwell,
R.B., Peake, G.M., Smolyakov, G.A., Eliseev, P.G., Osiñski, M.: Frequency
beating between monolithically integrated semiconductor ring lasers. Appl.
Phys. Lett. 86, 041101 (2005)
Carriere, J.T.A., Frantz, J.A., Youmans, B.R., Honkanen, S., Kostuk, R.K.:
Measurement of waveguide birefringence using a ring resonator. IEEE Pho-
ton. Technol. Lett. 16, 1134–1136 (2004)
Caruso, L., Montrosset, I.: Analysis of a racetrack Microring resonator with
MMI coupler. IEEE J. Lightwave Technol. 21, 206–210 (2003)
Chao, C.Y., Guo, L.J.: Polymer microring resonators fabricated by nanoim-
print technique. J. Vac. Sci. Technol. B 20, 2862–2866 (2002)
Chao, C.Y., Guo, L.J.: Biochemical sensors based on polymer microrings with
sharp asymmetrical resonance. Appl. Phys. Lett. 83, 1527–1529 (2003)
Chao, C.Y., Guo, L.J.: Reduction of surface scattering loss in polymer micro-
rings using thermal-reflow technique. IEEE Photon. Technol. Lett. 16,
1498–1500 (2004)
Chen, W., Chen, W., Chen, Y.J.: A characteristic matrix approach for analyz-
ing resonant ring lattice devices. IEEE Photon. Technol. Lett. 16, 458–460
(2004a)
References 235
Chen, W.Y., Grover, R., Ibrahim, T.A., Van, V., Herman, W.N., Ho, P.T.:
High-finesse laterally coupled single-mode benzocyclobutene microring reso-
nators. IEEE Photon. Technol. Lett. 16, 470–472 (2004b)
Chin, M.K.: Polarization dependence in waveguide-coupled micro-resonators.
Opt. Express 11, 1724–1730 (2003)
Chin, M.K., Ho, S.T.: Design and modeling of waveguide-coupled single-mode
microring resonators. IEEE J. Lightwave Technol. 16, 1433–1446 (1998)
Chin, M.K., Xu, C., Huang, W.: Theoretical approach to a polarization insen-
sitive single-mode microring resonator. Opt. Express 12, 3240–3250 (2004)
Chin, M.K., Youtsey, C., Zhao, W., Pierson, T., Ren, Z., Wu, S.L., Wang, L.,
Zhao, Y.G., Ho, S.T.: GaAs microcavity channel-dropping filter based on a
race-track resonator. IEEE Photon. Technol. Lett. 11, 1620–1622 (1999)
Choi, S.J., Djordjev, K., Choi, S.J., Coleman, J.J., Dapkus, P.D.: Microdisk
laser vertically coupled to output waveguides. Proc. ISLC 2002, Paper ThB6
(2002a)
Choi, S.J., Djordjev, K., Choi, S.J., Dapkus, P.D.: CH4-based dry etching of
high Q InP microdisks. J. Vac. Sci. Technol. B 20, 301–305 (2002b)
Choi, S.J., Djordjev, K., Choi, S.J., Dapkus, P.D.: Microdisk lasers vertically
coupled to output waveguides. IEEE Photon. Technol. Lett. 15, 1330–1332
(2003)
Choi, S.J., Djordjev, K., Choi, S.J., Dapkus, P.D., Lin, W., Griffel, G.,
Menna, R., Connolly, J.: Microring resonators vertically coupled to buried
heterostructure bus waveguides. IEEE Photon. Technol. Lett. 16, 828–830
(2004a)
Choi, S.J., Djordjev, K., Peng, Z., Yang, Q., Choi, S.J., Dapkus, P.D.: Later-
ally coupled buried heterostructure high-Q ring resonators. IEEE Photon.
Technol. Lett. 16, 2266–2268 (2004b)
Choi, S.J., Peng, Z., Yang, Q., Choi, S.J., Dapkus, P.D.: Eight-channel
microdisk cw laser arrays vertically coupled to common output bus
waveguides. IEEE Photon. Technol. Lett. 16, 356–358 (2004c)
Choi, S.J., Peng, Z., Yang, Q., Choi, S.J., Dapkus, P.D.: Tunable microdisk
resonators vertically coupled to bus waveguides using epitaxial regrowth
and wafer bonding techniques. Appl. Phys. Lett. 84, 651–653 (2004d)
Choi, S.J., Peng, Z., Yang, Q., Hwang, E.H., Dapkus, P.D.: A semiconductor
tunable laser using a wavelength selective reflector based on ring resonators.
Proc. OFC/NFOEC 2005, Paper PDP20 (2005)
Chremmos, I., Uzunoglu, N.: Reflective properties of double-ring resonator
system coupled to a waveguide. IEEE Photon. Technol. 17, 2110–2112
(2005)
Christiaens, I.: Vertically coupled microring resonators fabricated with wafer
bonding. Ph.D. Thesis, Ghent University, ISBN 90-8578-020-9 (2005)
Christiaens, I., Roelkens, G., De Mesel, K., Van Thourhout, D., Baets, R.:
Thin-film devices fabricated with benzocyclobutene adhesive wafer bonding.
IEEE J. Lightwave Technol. 23, 517–523 (2004)
236 References
Chu, S.T., Little, B.E., Pan, W., Kaneko, T., Kokubun, Y.: Cascaded micro-
ring resonators for crosstalk reduction and spectrum cleanup in add–drop
filters. IEEE Photon. Technol. Lett. 11, 1423–1425 (1999a)
Chu, S.T., Little, B.E., Pan, W., Kaneko, T., Kokubun, Y.: Second-order filter
response from parallel coupled glass microring resonators. IEEE Photon.
Technol. Lett. 11, 1426–1428 (1999b)
Chu, S.T., Little, B.E., Pan, W., Kaneko, T., Sato, S., Kokubun, Y.: An eight-
channel add–drop filter using vertically coupled microring resonators over
a cross grid. IEEE Photon. Technol. Lett. 11, 691–693 (1999c)
Chu, S.T., Pan, W., Sato, S., Kaneko, T., Little, B.E., Kokubun, Y.: Wave-
length trimming of a microring resonator filter by means of a UV sensitive
polymer overlay. IEEE Photon. Technol. Lett. 11, 688–690 (1999d)
Chu, S.T., Pan, W., Suzuki, S., Little, B.E., Sato, S., Kokubun, Y.: Tempera-
ture insensitive vertically coupled microring resonator add/drop filters by
means of a polymer overlay. IEEE Photon. Technol. Lett. 11, 1138–1140
(1999e)
Chung, Y., Kim, D.G., Dagli, N.: Widely tunable coupled-ring reflector laser
diode. IEEE Photon. Technol. Lett. 17, 1773–1775 (2005)
Cockerill, T.M., Osowski, M.L., Lammert, R.M., Coleman, J.J.: A strained-
layer InGaAs–GaAs buried heterostructure circular ring laser with inte-
grated y-coupled passive waveguide by selective-area metalorganic chemical
vapor deposition. Proc. ISLC 195–196 (1994)
Coppinger, F., Madsen, C.K., Jalali, B.: Photonic microwave filtering using co-
herently coupled integrated ring resonators. Microwave Opt. Technol. Lett.
21, 90–93 (1999)
Corbett, B.: Spectral characteristics of single-InGaAs quantum-well microring
lasers. IEEE Photon. Technol. Lett. 10, 3–5 (1998)
Corbett, B., Justice, J., Considine, L., Walsh, S., Kelly, W.M.: Low-threshold
lasing in novel microdisk geometries. IEEE Photon. Technol. Lett. 8, 855–
857 (1996)
Dapkus, P.D., Choi, S.J., Choi, S.J., Djordjev, K., Sadagopan, T., Tishinin, D.:
Microresonator devices for DWDM systems. Proc. OFC Paper WK 1–3
(2001)
Darmawan, S., Lee, S.Y., Lee, C.W., Chin, M.K.: A rigorous comparative
analysis of directional couplers and multimode Interferometers based on
ridge waveguides. IEEE J. Select. Top. Quantum Electron. 11, 466–475
(2005)
Dehmubed, R.S., Diament, P.: Coupling of a planar waveguide and a dielectric
disk. IEEE Trans. Microwave Theory Tech. 46, 1461–1470 (1998)
Della Corte, F.G., Cocorullo, G., Iodice, M., Rendina, I.: Temperature depen-
dence of the thermo-optic coefficient of InP, GaAs, and SiC from room
temperature to 600 K at the wavelength of 1.5 µm. Appl. Phys. Lett. 77,
1614–1616 (2000)
References 237
Djordjev, K.: Active microdisk resonant devices and semiconductor optical

equalizers as building blocks for future photonic circuitry. Ph.D. Thesis,
University of Southern California (2002)
Djordjev, K., Choi, S.J., Choi, S.J., Dapkus, P.D.: Active semiconductor
microdisk devices. IEEE J. Lightwave Technol. 20, 105–113 (2002a)
Djordjev, K., Choi, S.J., Choi, S.J., Dapkus, P.D.: Gain trimming of the
resonant characteristics in vertically coupled InP microdisk switches. Appl.
Phys. Lett. 80, 3467–3469 (2002b)
Djordjev, K., Choi, S.J., Choi, S.J., Dapkus, P.D.: High-Q vertically coupled
InP microdisk resonators. IEEE Photon. Technol. Lett. 14, 331–333 (2002c)
Djordjev, K., Choi, S.J., Choi, S.J., Dapkus, P.D.: Microdisk tunable resonant
filters and switches. IEEE Photon. Technol. Lett. 14, 828–830 (2002d)
Djordjev, K., Choi, S.J., Choi, S.J., Dapkus, P.D.: Study of the effects of
the geometry on the performance of vertically coupled InP microdisk reso-
nators. IEEE J. Lightwave Technol. 20, 1485–1492 (2002e)
Djordjev, K., Choi, S.J., Choi, S.J., Dapkus, P.D.: Vertically coupled InP
microdisk switching devices with electroabsorptive active regions. IEEE
Photon. Technol. Lett. 14, 1115–1117 (2002f)
Dumon, P., Bogaerts, W., Wiaux, V., Wouters, J., Beckx, S.,
Van Campenhout, J., Taillaert, D., Luyssaert, B., Bienstman, P.,
Van Thourhout, D., Baets, R.: Low-loss SOI photonic wires and ring
resonators fabricated with deep UV lithography. IEEE Photon. Technol.
Lett. 16, 1328–1330 (2004)
Emelett, S.J., Soref, R.: Design and simulation of silicon microring optical
routing switches. IEEE J. Lightwave Technol. 23, 1800–1807 (2005)
Faist, J., Gmachl, C., Striccoli, M., Sirtori, C., Capasso, F., Sivco, D.L., Cho,
A.Y.: Quantum cascade disk lasers. Appl. Phys. Lett. 69, 2456–2458 (1996)
Fan, S.: Sharp asymmetric line shapes in side-coupled waveguide-cavity sys-
tems. Appl. Phys. Lett. 80, 908–910 (2002)
Feuchter, T., Thirstrup, C.: High precision planar waveguide propagation loss
measurement technique using a Fabry–Perot cavity. IEEE Photon. Technol.
Lett. 6, 1244–1247 (1994)
Fiedler, F., Schlachetzki, A.: Optical parameters of InP-based waveguides.
Solid-State Electron. 30, 73–83 (1987)
Forber, R., Marom, E.: Symmetric directional coupler switches. IEEE J. Quan-
tum Electron. 22, 911–919 (1986)
Frateschi, N.C., Levi, A.F.J.: Resonant modes and laser spectrum of microdisk
lasers. Appl. Phys. Lett. 66, 2932–2934 (1995)
Fujita, M., Sakai, A., Baba, T.: Ultrasmall and ultralow threshold GaInAsP–
InP microdisk injection lasers: Design, fabrication, lasing characteristics,
and spontaneous emission factor. IEEE J. Select Topics Quantum Electron.
5, 673–681 (1999)
Georgakilas, A., Deligeorgis, G., Aperathitis, E., Cengher, D., Hatzopoulos, Z.,
Alexe, M., Dragoi. V., Gösele, U., Kyriakis-Bitzaros, E.D., Minoglou, K.,
238 References
Halkias, G.: Wafer-scale integration of GaAs optoelectronic devices with

standard Si integrated circuits using a low-temperature bonding procedure.
Appl. Phys. Lett. 81, 5099–5101 (2002)
Geuzebroek, D.: Flexible optical network components based on densely inte-
grated microring resonators. Ph.D. Thesis, University of Twente, ISBN
90-365-2258-7 (2005)
Gheorma, I.L., Osgood, R.M.: Fundamental limitations of optical resonator
based high-speed EO modulators. IEEE Photon. Technol. Lett. 14, 795–797
(2002)
Gottesman, Y., Rao, E.V.K., Rabus, D.G.: New methodology to evaluate the
performance of ring resonators using optical low-coherence reflectometry.
Griffel, G.: Synthesis of optical filters using ring resonator arrays. IEEE Pho-
ton. Technol. Lett. 12, 810–812 (2000a)
Griffel, G.: Vernier effect in asymmetrical ring resonator arrays. IEEE Photon.
Technol. Lett. 12, 1642–1644 (2000b)
Griffel, G., Abeles, J.H., Menna, R.J., Braun, A.M., Connolly, J.C., King. M.:
Low-threshold InGaAsP ring lasers fabricated using bi-level dry etching.
Grover, R.: Indium phosphide based optical micro-ring resonators.
Ph.D. Thesis, University of Maryland, College Park (2003)
Grover, R., Absil, P.P., Van, V., Hryniewicz, J., Little, B.E., King, O.,
Calhoun, L.C., Johnson, F.G., Ho, P.T.: Vertically coupled GaInAsP–InP
microring resonators. Opt. Lett. 26, 506–508 (2001a)
Grover, R., Absil, P.P., Van, V., Hryniewicz, J.V., Little, B.E., King, O.,
Johnson, F.G., Calhoun, L.C., Ho, P.T.: Vertically coupled GaAs–AlGaAs
and GaInAsP–InP microring resonators. Proc. OFC Paper WK2-1 (2001b)
Grover, R., Hryniewicz, J.V., King, O.S., Van, V.: Process development of
methane–hydrogen–argon-based deep dry etching of InP for high aspect-
ratio structures with vertical facet-quality sidewalls. J. Vac Sci. Technol.
B 19, 1694–1698 (2001c)
Grover, R., Van, V., Ibrahim, T.A., Absil, P.P., Calhoun, L.C., Johnson, F.G.,
Hryniewicz, J.V., Ho, P.T.: Parallel-cascaded semiconductor microring reso-
nators for high-order and wide-FSR filters. IEEE J. Lightwave Technol. 20,
900–905 (2002)
Grover, R., Ibrahim, T.A., Ding, T.N., Leng, Y., Kuo, L.C., Kanakaraju, S.,
Amarnath, K., Calhoun, L.C., Ho, P.T.: Laterally coupled InP-based single-
mode microracetrack notch filter. IEEE Photon. Technol. Lett. 15, 1082–
1084 (2003)
Grover, R., Ibrahim, T.A., Kanakaraju, S., Lucas, L., Calhoun, L.C., Ho, P.T.:
A Tunable GaInAsP–InP Optical Microring Notch Filter. IEEE Photon.
Technol. Lett. 16, 467–469 (2004)
Haavisto, J., Pajer, G.A.: Resonance effects in low-loss ring waveguides.
Opt. Lett. 5, 510–512 (1980)
References 239
Haeiwa, H., Naganawa, T., Kokubun, Y.: Wide range center wavelength trim-
ming of vertically coupled microring resonator filter by direct UV irradiation
to SiN ring core. IEEE Photon. Technol. Lett. 16, 135–137 (2004)
Hagness, S.: FDTD computational electromagnetics modeling of microcavity
lasers and resonant optical structures. Ph.D. Thesis, Northwestern
University (1998)
Hagness, S.C., Rafizadeh, D., Ho, S.T., Taflove, A.: FDTD Microcavity simu-
lations: Design and experimental realization of waveguide-coupled single-
mode ring and whispering-gallery-mode disk resonators. IEEE J. Lightwave
Technol. 15, 2154–2165 (1997)
Hakki, B.W., Paoli, T.L.: Gain spectra in GaAs double-heterostructure injec-
tion lasers. J. Appl. Phys 46, 1299–1306 (1975)
Hamacher, M., Kaiser, R., Heidrich, H., Albrecht, P., Borchert, B., Janiak, K.,
Löffler, R., Malchow, S., Rehbein, W., Schroeter-Janßen, H.: Monolithic
integration of lasers, photodiodes, waveguides and spot size converters on
GaInAsP/InP for photonic IC applications. Proc. IPRM 21–24 (2000)
Hamacher, M., Troppenz, U., Heidrich, H., Rabus, D.G.: Active ring reso-
nators based on GaInAsP/InP. Proc. of SPIE 4947, 212–222 (2003)
Hammer, M.: Resonant coupling of dielectric optical waveguides via rectan-
gular microcavities: The coupled guided mode perspective. Opt. Commun.
214, 155–170 (2002)
Hammer, M., Yudistira, D., Stoffer, R.: Modeling of grating assisted stand-
ing wave microresonators for filter applications in integrated optics. Opt.
Quantum Electron. 36, 25–42 (2004)
Hansen, P.B., Raybon, G., Chien, M.D., Koren, U., Miller, B.I., Young, M.G.,
Verdiell, J.M., Burrus, C.A.: A 1.54 µm monolithic semiconductor ring laser:
CW and mode-locked operation. IEEE Photon. Technol. Lett. 4, 411–413
(1992)
Haruna, M., Segawa, Y., Nishihara, H.: Nondestructive and simple method of
optical-waveguide loss measurement with optimization of end-fire coupling.
Electron. Lett. 28, 1612–1613 (1992)
Hatakeyama, Y., Hanai, T., Suzuki, S., Kokubun, Y.: Loss-Less multilevel
crossing of busline waveguide in vertically coupled microring resonator filter.
Hauffe, R.: Integrated optical switching matrices constructed from digital opti-
cal switches based on polymeric rib waveguides. Ph.D. Thesis, Technical
University Berlin (2002)
Haxha, S., Ladely, E.O., Mjeku, M., AbdelMalek, F., Rahman, B.M.A.:
Optimization of compact lateral, vertical, and combined tapered spot size
converters by use of the beam-propagation method. Appl. Opt. 45, 288–296
(2006)
Headley, W.R., Reed, G.T., Howe, S., Liu, A., Paniccia, M.: Polarization-
independent optical racetrack resonators using rib waveguides on
silicon-on-insulator. Appl. Phys. Lett. 85, 5523–5525 (2004)
240 References
Heebner, J.E., Boyd, R.W., Park, Q.H.: SCISSOR solitons and other novel
propagation effects in microresonator-modified waveguides. J. Opt. Soc.
Am. B 19, 722–731 (2002)
Heebner, J.E., Lepeshkin, N.N., Schweinsberg, A., Wicks, G.W., Boyd, R.W.,
Grover, R., Ho, P.T.: Enhanced linear and nonlinear optical phase response
of AlGaAs microring resonators. Opt. Lett. 29, 769–771 (2004)
Heiblum, M., Harris, J.: Analysis of curved optical waveguides by conformal
transformation. IEEE J. Quantum Electron. 11, 75–83 (1975)
Heimala, P., Katila, P., Aarnio, J., Heinämäki, A.: Thermally tunable integ-
rated optical ring resonator with poly-Si thermistor. IEEE J. Lightwave
Technol. 14, 2260–2267 (1996)
Hibino, Y., Terui, H., Sugita, A., Ohmori, Y.: Silica-based optical waveguide
ring laser integrated with semiconductor laser amplifier on Si substrate.
Electron. Lett. 28, 1932–1933 (1992)
Hida, Y., Imamura, S., Izawa, T.: Ring resonator composed of low loss polymer
waveguides at 1.3 µm. Electron. Lett. 28, 1314–1316 (1992)
Hidayat, I.S., Toyota, Y., Torigoe, O., Wada, O., Koga, R.: Multipath struc-
ture for FSR expansion in waveguide-based optical ring resonator. Electron.
Lett. 39 (2003)
Hill, T.M., Dorren, H.J.S., de Vries, T., Leijtens, X.J.M., den Besten, J.H.,
Smalbrugge, B., Oei, J.S., Binsma, H., Khoe, G.D., Smit, M.K.: A fast
low-power optical memory based on coupled micro-ring lasers. Nature 432,
206–209 (2004)
Hiremath, K.R.: Coupled mode theory based modelling and analysis of circular
optical microresonators. Ph.D. Thesis, University of Twente, ISBN 90-365-
2267-6 (2005)
Hohimer, J.P., Vawter, G.A.: Unidirectional semiconductor ring lasers with
racetrack cavities. Appl. Phys. Lett. 63, 2457–2459 (1993)
Hohimer, J.P., Craft, D.C., Hadley, G.R., Vawter, G.A., Warren, M.E.: Single-
frequency continuous-wave operation of ring resonator diode lasers. Appl.
Phys. Lett. 59, 3360–3362 (1991)
Hohimer, J.P., Vawter, G.A., Craft, D.C., Hadley, G.R.: Improving the perfor-
mance of semiconductor ring lasers by controlled reflection feedback. Appl.
Phys. Lett. 61, 1013–1015 (1992)
Horst, F., Salemink, H.W.M., Germann, R., Offrein, B.J., Bona, G.L.: High
quality ring resonators in high refractive index contrast SiON waveguides.
Proc. 1998 IEEE/LEOS Symp. Benelux Chapter:33–36 (1998)
Horst, F., Berendsen, C., Beyeler, R., Bona, G.L., Germann, R., Salemink,
H.W.M., Wiesmann, D.: Tunable ring resonator dispersion compen-
sators realized in high-refractive-index contrast SiON technology. Proc.
ECOC’2000, Post Deadline Paper PD 2.2 (2000)
Hryniewicz, J.V., Absil, P.P., Little, B.E., Wilson, R.A., Ho, P.T.: Higher
order filter response in coupled microring resonators. IEEE Photon. Tech-
nol. Lett. 12, 320–322 (2000)
References 241
Huang, Y., Paloczi, G.T., Scheuer, J., Yariv, A.: Soft lithography replication of
polymeric microring optical resonators. Opt. Express 11, 2452–2458 (2003)
Huang, Y., Paloczi, G.T., Poon, J.K.S., Yariv, A.: Demonstration of flexi-
ble freestanding all-polymer integrated optical ring resonator devices. Adv.
Mater. 16, 44–48 (2004)
Ibrahim, T.A.: Nonlinear optical semiconductor micro ring resoantors. Ph.D.
Thesis, University of Maryland, College Park (2003)
Ibrahim, T.A., Van, V., Ho, P.T.: All-optical time-division demultiplexing and
spatial pulse routing with a GaAs/AlGaAs microring resonator. Opt. Lett.
27, 803–805 (2002)
Ibrahim, T.A., Cao, W., Kim, Y., Li, J., Goldhar, J., Ho, P.T., Lee, C.H.:
All-optical switching in a laterally coupled microring resonator by carrier
injection. IEEE Photon. Technol. Lett. 15, 36–38 (2003a)
Ibrahim, T.A., Grover, R., Kuo, L.C., Kanakaraju, S., Calhoun, L.C., Ho,
P.T.: All-optical AND/NAND logic gates using semiconductor microreso-
nators. IEEE Photon. Technol. Lett. 15, 1422–1424 (2003b)
Ibrahim, T.A., Cao, W., Kim, Y., Li, J., Goldhar, J., Ho, P.T., Lee, C.H.:
Lightwave switching in semiconductor microring devices by free carrier
injection. IEEE J. Lightwave Technol. 21, 2997–3003 (2003c)
Ji, C., Leary, M.H., Ballantyne, J.M.: Long-wavelength triangular ring laser.
Jiang, X., Li, X., Zhou, H., Yang, J., Wang, M., Wu, Y., Ishikawa, S.: Compact
variable optical attenuator based on multimode interference coupler. IEEE
Photon. Technol. Lett. 17, 2361–2363 (2005)
Kaalund, C.J., Peng, G.D.: Pole-zero diagram approach to the design of ring
resonator-based filters for photonic applications. IEEE J. Lightwave Tech-
nol. 22, 1548–1559 (2004)
Kaiser, R., Heidrich, H.: Optoelectronic/Photonic integrated circuits on InP
between technological feasibility and commercial success. IEICE Transac-
tions on Electronics E85-C, 970–981 (2002)
Kaiser, R., Trommer, D., Heidrich, H., Fidorra, F., Hamacher, M.: Hetero-
dyne receiver Pies as the first monolithically integrated tunable receivers for
OFDM system applications. Opt. Quantum Electron. 28, 565–573 (1996)
Kaminow, I.P., Stulz, L.W.: Loss in cleaved Ti-diffused LiNbO3 waveguides.
Appl. Phys. Lett. 33, 62–64 (1978)
Kaplan, A.: Modeling of ring resonators with tunable couplers. IEEE J. Select.
Top. Quantum Electron. 12, 86–95 (2006)
Kasaya, K., Mitomi, O., Naganuma, M., Kondo, Y., Noguchi, Y.: A sim-
ple laterally tapered waveguide for low-loss coupling to single-mode fibers.
Katila, P., Heimala, P., Aarnio, J.: Thermo-optically controlled ring resonator
on silicon. Electron. Lett. 32, 1005–1006 (1996)
Keil, N., Yao, H.H., Zawadzki, C., Bauer, J., Bauer, M., Dreyer, C.,
Schneider, J.: Athermal all-polymer arrayed-waveguide grating multiplexer.
Electron. Lett. 37, 579–580 (2001)
242 References
Kim, M.J., Carpenter, R.W.: Heterogeneous silicon integration by ultra-high

vacuum wafer bonding. J. Electron. Mater. 32, 849–854 (2003)
Kim, S.H., Ryu, H.Y., Park, H.G., Kim, G.H., Choi, Y.S., Lee, Y.H., Kim, J.S.:
Two-dimensional photonic crystal hexagonal waveguide ring laser. Appl.
Phys. Lett. 81, 2499–2501 (2002)
Kim, D.H., Im, J.G., Lee, S.S., Ahn, S.W., Lee, K.D.: Polymeric microring
resonator using nanoimprint technique based on a stamp incorporating a
smoothing buffer layer. IEEE Photon. Technol. Lett. 17, 2352–2354 (2005a)
Kim, D.G., Shin, J.H., Ozturk, C., Yi, J.C., Chung, Y., Dagli, N.: Total inter-
nal reflection mirror-based InGaAsP ring resonators integrated with optical
amplifiers. IEEE Photon. Technol. Lett. 17, 1899–1901 (2005b)
Kim, H.S., Kwon, Y.S., Hong, S.: Square ring laser diode with MMI coupler
cavity. IEEE Photon. Technol. Lett. 9, 584–586 (1997)
Kiyat, I., Kocabas, C., Aydinli, A.: Integrated micro ring resonator displace-
ment sensor for scanning probe microscopies. J. Micromechanics Microengi-
neering 14, 374–381 (2004)
Kiyat, I., Aydinli, A., Dagli, N.: Polarization characteristics of compact SOI
rib waveguide racetrack resonators. IEEE Photon. Technol. Lett. 17, 2098–
2100 (2005)
Kiyat, I., Aydinli, A., Dagli, N.: Low-power thermooptical tuning of SOI reso-
nator switch. IEEE Photon. Technol. Lett. 18, 364–366 (2006)
Klunder, D.J.W., Krioukov, E., Tan, F.S., Van der Veen, T., Bulthuis, H.F.,
Sengo, G., Otto, C., Hoekstra, H.J.W.M., Driessen, A.: Vertically and later-
ally waveguide-coupled cylindrical microresonators in Si3 N4 on SiO2 tech-
nology. Appl. Phys. B: Laser Opt. 73, 603–608 (2001)
Klunder, D.J.W., Balistreri, M.L.M., Blom, F.C., Hoekstra, H.J.W.M.,
Driessen, A., Kuipers, L., van Hulst, N.F.: Detailed analysis of the intra-
cavity phenomena inside a cylindrical microresonator. IEEE J. Lightwave
Technol. 20, 519–529 (2002a)
Klunder, D.J.W., Roeloffzen, C.G.H., Driessen, A.: A novel polarization-
independent wavelength-division-multiplexing filter based on cylindrical
microresonators. IEEE J. Select. Top. Quantum Electron. 8, 1294–1299
(2002b)
Klunder, D.J.W., Hoekstra, H.J.W.M., Driessen, A.: A normalized approach
for designing cylindrical microresonators. IEEE J. Lightwave Technol. 21,
1405–1410 (2003a)
Klunder, D.J.W., Tan, F.S., van der Veen, T., Bulthuis, H.F., Sengo, G.,
Docter, B., Hoekstra, H.J.W.M., Driessen, A.: Experimental and numerical
study of SiON microresonators with air and polymer cladding. IEEE J.
Lightwave Technol. 21, 1099–1110 (2003b)
Kohtoku, M., Oku, S., Kadota, Y., Shibata, Y., Yoshikuni, Y.: 200-GHz
FSR Periodic multi/demultiplexer with flattened transmission and rejec-
tion band by using a Mach–Zehnder interferometer with a ring resonator.
References 243
Kokubun, Y.: Vertically coupled microring resonator filter for integrated

add/drop node. IEICE Trans. Electron. 3, 349–362 (2005)
Kokubun, Y., Funato, N., Takizawa, M.: Athermal waveguides for
temperature-independent lightwave devices. IEEE Photon. Technol. Lett.
5, 1297–1300 (1993)
Kokubun, Y., Yoneda, S., Tanaka, H.: Temperature-independent narrow-band
optical filter by an athermal waveguide. IEICE Trans. Electron. E80-C, 632–
639 (1997)
Kokubun, Y., Kubota, S., Chu, S.T.: Polarization-independent vertically cou-
pled microring resonator filter. Electron. Lett. 37, 90–92 (2001)
Kokubun, Y., Kato, T., Chu, S.T.: Box-like response of microring filter by
stacked double-ring geometry. IEICE Trans. Electron. E85-C, 1018–1024
(2002)
Kokubun, Y., Hatakeyama, Y., Ogata, M., Suzuki, S., Zaizen, N.: Fabrica-
tion technologies for vertically coupled microring resonator with multilevel
crossing busline and ultracompact-ring radius. J. Select. Top. Quantum
Electron. 11, 4–10 (2005)
Koonath, P., Kishima, K., Indukuri, T., Jalali, B.: Sculpting of three-
dimensional nano-optical structures in silicon. Appl. Phys. Lett. 83, 4909–
4911 (2003)
Koonath, P., Indukuri, T., Jalali, B.: Vertically-coupled micro-resonators rea-
lized using three-dimensional sculpting in silicon. Appl. Phys. Lett. 85,
1018–1020 (2004)
Koonath, P., Indukuri, T., Jalalib, B.: Add–drop filters utilizing vertically
coupled microdisk resonators in silicon. Appl. Phys. Lett. 86, 091102 (2005)
Krauss, T.F., Laybourn, P.J.R., Roberts, J.: CW operation of semiconductor
ring lasers. Electron. Lett. 26, 2095–2097 (1990)
Krauss, T., DeLaRue, R.M., Gontijo, I., Laybourn, P.J.R.: Strip-loaded semi-
conductor ring lasers employing multimode interference output couplers.
Appl. Phys. Lett. 64, 2788–2790 (1994)
Krauss, T.F., De La Rue, R.M., Laybourn, P.J.R.: Impact of output cou-
pler configuration on operating characteristics of semiconductor ring lasers.
IEEE J. Lightwave Technol. 13, 1500–1507 (1995a)
Krauss, T.F., De La Rue, R.M., Laybourn, P.J.R., Vögele, B., Stanley, C.R.:
Efficient semiconductor ring lasers made by a simple self-aligned fabrication
process. IEEE J. Select. Top. Quantum Electron. 1, 757–761 (1995b)
Krioukov, E., Klunder, D.J.W., Driessen, A., Greve, J., Otto, C.: Sensor based
on an integrated optical microcavity. Opt. Lett. 27, 512–514 (2002)
Ksendzov, A., Homer, M.L., Manfreda, A.M.: Integrated optics ring-resonator
chemical sensor with polymer transduction layer. Electron. Lett. 40, 63–65
(2004)
Lee, H.S., Choi, C.H., Beom-Hoan, O., Park, D.G., Kang, B.G., Kim, S.H.,
Lee, S.G., Lee, E.H.: A nonunitary transfer matrix method for practical
analysis of racetrack microresonator waveguide. IEEE Photon. Technol.
Lett. 16, 1086–1088 (2004)
244 References
Lee, M.C.M., Wu, M.C.: MEMS-actuated microdisk resonators with variable

power coupling ratios. IEEE Photon. Technol. Lett. 17, 1034–1036 (2005)
Leinse, A.: Polymeric microring resonator based electro optic modulator.
Ph.D. Thesis, University of Twente, ISBN 90-365-2168-8 (2005)
Lenz, G., Madsen, C.K.: General optical all-pass filter structures for dispersion
control in WDM systems. IEEE J. Lightwave Technol. 17, 1248–1254 (1999)
Lenz, G., Eggleton, B.J., Madsen, C.K., Slusher, R.E.: Optical delay lines
based on optical filters. IEEE J. Quantum Electron. 37, 525–532 (2001)
Leuthold, J., Joyner, C.H.: Multimode interference couplers with tuneable
splitting ratios. Proc. ECOC 3, 177–178 (2000)
Leuthold, J., Joyner, C.H.: Multimode interference couplers with tunable
power splitting ratios. IEEE J. Lightwave Technol. 19, 700–707 (2001)
Levi, A.F.J.: Microdisk lasers. Solid-State Electron. 37, 1297–1302 (1994)
Liao, A.S.H., Wang, S.: Semiconductor injection lasers with a circular reso-
nator. Appl. Phys. Lett. 36, 801–803 (1980)
Little, B.E., Chu, S.T.: Estimating surface-roughness loss and output coupling
in microdisk resonators. Opt. Lett. 21, 1390–1392 (1996)
Little, B.E., Murphy, T.: Design rules for maximally flat wavelength-
insensitive optical power dividers using Mach–Zehnder structures. IEEE
Little, B.E., Laine, J.P., Chu, S.T.: Surface-roughness-induced contradirec-
tional coupling in ring and disk resonators. Opt. Lett. 22, 4–6 (1997a)
Little, B.E., Chu, S.T., Haus, H.A., Foresi, J., Laine, J.P.: Microring resonator
channel dropping filters. IEEE J. Lightwave Technol. 15, 998–1005 (1997b)
Little, B.E., Foresi, J.S., Steinmeyer, G., Thoen, E.R., Chu, S.T., Haus, H.A.,
Ippen, E.P., Kimerling, L.C., Greene, W.: Ultra-compact Si–SiO microring
resonator optical channel dropping filters. IEEE Photon. Technol. Lett. 10,
549–551 (1998a)
Little, B.E., Haus, H.A., Foresi, J.S., Kimerling, L.C., Ippen, E.P., Ripin, D.J.:
Wavelength switching and routing using absorption and resonance. IEEE
Photon. Technol. Lett. 10, 816–818 (1998b)
Little, B.E., Chu, S.T., Pan, W., Ripin, D., Kaneko, T., Kokubun, Y., Ippen,
E.: Vertically coupled glass microring resonator channel dropping filters.
Little, B.E., Chu, S.T.: Theory of loss and gain trimming of resonator-type
filters. IEEE Photon. Technol. Lett. 12, 636–638 (2000)
Little, B.E., Chu, S.T., Hryniewicz, J.V., Absil, P.P.: Filter synthesis for perio-
dically coupled microring resonators. Opt. Lett. 25, 344–346 (2000a)
Little, B.E., Chu, S.T., Pan, W., Kokubun, Y.: Microring resonator arrays for
VLSI photonics. IEEE Photon. Technol. Lett. 12, 323–325 (2000b)
Little, B.E., Chu, S.T., Absil, P.P., Hryniewicz, J.V., Johnson, F.G.,
Seiferth, F., Gill, D., Van, V., King, O., Trakalo, M.: Very high-order micro-
ring resonator filters for WDM applications. IEEE Photon. Technol. Lett.
16, 2263–2265 (2004)
References 245
Liu, B., Shakouri, A., Bowers, J.E.: Passive microring-resonator-coupled

lasers. Appl. Phys. Lett. 79, 3561–3563 (2001)
Liu, B., Shakouri, A., Bowers, J.E.: Wide tunable double ring resonator cou-
pled lasers. IEEE Photon. Technol. Lett. 14, 600–602 (2002a)
Liu, B., Shakouri, A., Rabus, D.G.: Passive microring resonator coupled lasers.
Proc. IEEE LEOS Annual Meeting, 742–743 (2002b)
Lui, W.W., Hirono, T., Yokoyama, K., Huang, W.P.: Polarization rotation
in semiconductor bending waveguides: A coupled-mode theory formulation.
Ma, Y., Chang, G., Chang, S., Ho, S.T.: Optical filter responses of InGaAsP
microdisk resonators fabricated by polymer wafer bonding. Proc. OFC
Paper WY4 (2000a)
Ma, Y., Park, S., Wang, L., Ho, S.T.: Ultracompact multimode interference
3 dB coupler with strong lateral confinement by deep dry etching. IEEE
Photon. Technol. Lett. 12, 492–494 (2000b)
Ma, Y., Chang, G., Park, S., Wang, L., Ho, S.T.: InGaAsP thin-film microdisk
resonators fabricated by polymer wafer bonding for wavelength add–drop
filters. IEEE Photon. Technol. Lett. 12, 1495–1497 (2000c)
Ma, Y., Chang, S.H., Chang, S.S., Ho, S.T.: Improved optical filter responses
in cascaded InGaAsP/InP microdisk resonators. Electron. Lett. 37, 564–
565 (2001)
Ma, H., Jen, A.K.Y., Dalton, L.R.: Polymer-based optical waveguides: Mate-
rials, processing, and devices. Adv. Mater. 14, 1339–1365 (2002)
Ma, C., Xu, Y., Yan, X., Qin, Z., Wang, X.: Effect of ring spacing on spectral
response of parallel-cascaded microring resonator arrays. Opt. Quantum
Electron. 37, 561–574 (2005)
Madsen, C.K., Zhao, J.H.: A general planar waveguide autoregressive optical
filter. IEEE J. Lightwave Technol. 14, 437–447 (1996)
Madsen, C.K., Zhao, J.H.: Postfabrication optimization of an autoregressive
planar waveguide lattice filter. Appl. Opt. 36, 642–647 (1997)
Madsen, C.K., Zhao, J.H.: Optical filter design and analysis, Wiley, New York
(1999)
Madsen, C.K., Lenz, G., Bruce, A.J., Capuzzo, M.A., Gomez, L.T., Nielsen,
T.N., Brener, I.: Multistage dispersion compensator using ring resonators.
Opt. Lett. 24, 1555–1557 (1999a)
Madsen, C.K., Lenz, G., Bruce, A.J., Cappuzzo, M.A., Gomez, L.T., Scotti,
R.E.: Integrated all-pass filters for tunable dispersion and dispersion slope
compensation. IEEE Photon. Technol. Lett. 11, 1623–1625 (1999b)
Madsen, C.K., Cappuzzo, M., Laskowski, E.J., Chen, E., Gomez, L.,
Griffin, A., Wong-Foy, A., Chandrasekhar, S., Stulz, L., Buhl, L.: Versatile
integrated PMD emulation and compensation elements. IEEE J. Lightwave
Technol. 22, 1041–1050 (2004)
Maolatou, C., Khan, M.J., Fan, S., Villeneuve, P.R., Haus, H.A.,
Joannopolous, J.D.: Coupling of modes analysis of resonant channel
add-drop filters. IEEE J. Quantum Electron. 35, 1322–1331 (1999)
246 References
Marcatili, E.A.J.: Bends in optical dielectric guides. Bell Syst. Techn. J. 48,
2103–2132 (1969a)
Marcatili, E.A.J.: Dielectric rectangular waveguide and directional coupler for
integrated optics. Bell Syst. Techn. Journal 48, 2071–2101 (1969b)
Matsko, A.B., Ilchenko, V.S.: Optical resonators with whispering gallery
Modes – Part I: Basics. IEEE J. Select. Top. Quantum Electron. 12,
3–14 (2006a)
Matsko, A.B., Ilchenko, V.S.: Optical resonators with whispering gallery
Modes – Part II: Applications. IEEE J. Select. Top. Quantum Electron.
12, 15–32 (2006b)
McCall, S.L., Levi, A.F.J., Slusher, R.E., Pearton, S.J., Logan, R.A.:
Whispering-gallery mode microdisk lasers. Appl. Phys. Lett. 60, 289–291
(1992)
Melloni, A.: Synthesis of a parallel-coupled ring-resonator filter. Opt. Lett.
26, 917–919 (2001)
Melloni, A., Carniel, F., Costa, R., Martinelli, M.: Determination of bend
mode characteristics in dielectric waveguides. IEEE J. Lightwave Technol.
19, 571–577 (2001)
Melloni, A., Martinelli, M.: Synthesis of direct-coupled-resonators bandpass
filters for WDM systems. IEEE J. Lightwave Technol. 20, 296–303 (2002)
Melloni, A., Costa, R., Monguzzi, P., Martinelli, M.: Ring-resonator filters
in silicon oxynitride technology for dense wavelength-division multiplexing
systems. Opt. Lett. 28, 1567–1569 (2003a)
Melloni, A., Morichetti, F., Martinelli, M.: Linear and nonlinear pulse prop-
agation in coupled resonator slow-wave optical structures. Opt. Quantum
Electron. 35, 365–379 (2003b)
Menon, V.M., Tong, W., Forrest, S.R.: Control of quality factor and critical
coupling in microring resonators through integration of a semiconductor
optical amplifier. IEEE Photon. Technol. Lett. 16, 1343–1345 (2004)
Michelotti, F., Driessen, A., Bertolotti, M. (eds.): Microresonators as building
blocks for VLSI photonics. AIP Conf. Proc. 709, Melville, New York (2004)
Mikroulis, S., Bogris, A., Roditi, E., Syvridis, D.: Investigation of an all-optical
wavelength converter with reshaping properties based on four-wave mixing
in passive microring resonators. IEEE J. Lightwave Technol. 22, 2743–2746
(2004)
Mikroulis, S., Simos, H., Roditi, E., Syvridis, D.: Ultrafast all-optical and logic
operation based on four-wave mixing in a passive InGaAsP–InP microring
resonator. IEEE Photon. Technol. Lett. 17, 1878–1880 (2005)
Morichetti, F., Melloni, A.: Polarization converters based on ring resonator
phase shifters. IEEE Photon. Technol. Lett. 18, 923–925 (2006)
Moslehi, B., Goodman, J.W., Tur, M., Shaw, H.J.: Fiber-optic lattice signal
processing. Proc. IEEE 72, 909–930 (1984)
Mule’, A.V., Villalaz, R., Gaylord, T.K., Meindl, J.D.: Photopolymer-based
diffractive and MMI waveguide couplers. IEEE Photon. Technol. Lett. 16,
2490–2492 (2004)
References 247
Murphy, T.E., Little, B.E., Smith, H.I.: Wavelength- and polarization-

insensitive integrated directional couplers using Mach-Zehnder structures.
Proc. IPR Paper RTuA5 (1999)
Nabiev, R.F., Francis, D., Chang-Hasnain, C.J.: Lateral and longitudinal
mode discrimination in index-guided circular ring semiconductor lasers.
Nagai, S., Morishima, G., Yagi, M., Utaka, K.: InGaAsP/InP Multi-mode
interference photonic switches for monolithic photonic integrated circuits.
Japan. J. Appl. Phys. 38, 1269–1272 (1999)
Neumann, E.G.: Curved dielectric optical waveguides with reduced transition
losses. IEEE Proc. H: Microwaves Opt. Antennas 129, 278–279 (1982a)
Neumann, E.G.: Low loss dielectric optical waveguide bends. Fiber Integrated
Opt. 4, 203–211 (1982b)
Neumann, E.G., Richter, W.: Sharp bends with low losses in dielectric optical
waveguides. Appl. Opt. 22, 1016–1022 (1983)
Nielson, G.N., Seneviratne, D., Lopez-Royo, F., Rakich, P.T., Giacometti, F.,
Tuller, H.L., Barbastathis, G.: MEMS based wavelength selective optical
switching for integrated photonic circuits. Proc. CLEO (2004)
Numai, T.: Analysis of photon recycling in semiconductor ring lasers. Jpn. J.
Appl. Phys. 39, 6535–6538 (2000)
Obayya, S.S.A., Rahman, B.M.A., Grattan, K.T.V., El-Mikati, H.A.: Beam
propagation modeling of polarization rotation in deeply etched semiconduc-
tor bent waveguides. IEEE Photon. Technol. Lett. 13, 681–683 (2001)
Oda, K., Takato, N., Toba, H.: A wide-FSR waveguide double-ring resonator
for optical FDM transmission systems. IEEE J. Lightwave Technol. 9, 728–
736 (1991)
Oda, K., Suzuki, S., Takahashi, H., Toba, H.: An optical FDM distribution
experiment using a high finesse waveguide-type double ring resonator. IEEE
Ohno, T., Ishii, H., Matsuo, S., Okamoto, H., Kawaguchi, Y., Kondo, Y.,
Furuta, T., Ito, H., Yoshikuni, Y.: Hybrid modelocking of semiconductor
ring lasers incorporating passive deep-ridge waveguides. Electron. Lett. 38,
884–886 (2002)
Okamoto, H., Haraguchi, M., Fukui, M., Okamoto, T.: Optical filtering by
microring resonators. Jpn. J. Appl. Phys. 42, 2692–2698 (2003)
Orta, R., Savi, P., Tascone, R., Trinchero, D.: Synthesis of multiple-ring-
resonator filters for optical systems. IEEE Photon. Technol. Lett. 7, 1447–
1449 (1995)
Painter, O.J., Husain, A., Scherer, A., O’Brien, J.D., Kim, I., Dapkus, P.D.:
Room temperature photonic crystal defect lasers at near-infrared wave-
lengths in InGaAsP. IEEE J. Lightwave Technol. 17, 2082–2088 (1999)
Paloczi, G.T., Huang, Y., Yariv, A.: Free-standing all-polymer microring reso-
nator optical filter. Electron. Lett. 39, 1650–1651 (2003a)
248 References
Paloczi, G.T., Huang, Y., Yariv, A., Mookherjea, S.: Polymeric Mach–Zehnder
interferometer using serially coupled microring resonators. Opt. Express 11,
2666–2671 (2003b)
Park, H.G., Hwang, J.K., Huh, J., Ryu, H.J., Kim, S.H., Kim, J.S., Lee, Y.H.:
Characteristics of modified single-defect two-dimensional photonic crystal
lasers. IEEE J. Quantum Electron. 38, 1353–1365 (2002)
Park, K.H., Kim, M.W., Byun, Y.T., Woo, D., Kim, S.H., Choi, S.S., Chung,
Y., Cho, W.R., Park, S.H., Kim, U.: Nondestructive propagation loss and
facet reflectance measurements of GaAs/AlGaAs strip-loaded waveguides.
J. Appl. Phys. 78, 6318–6320 (1995)
Park, S., Kim, S.S., Wang, L., Ho, S.T.: Single-mode lasing operation using a
microring resonator as a wavelength selector. IEEE J. Quantum Electron.
38, 270–273 (2002)
Pennings, E.C.M., Deri, R.J., Hawkins, R.J.: Simple method for estimating
usable bend radii of deeply etched optical rib waveguides. Electron. Lett.
27, 1532–1534 (1991)
Pennings, E.C.M., van Roijen, R., van Stralen, M.J.N., Verbeek, B.H., van
der Heijden, J.M.M., van Dongen, T.: Spectral behaviour of InP-based ring
lasers employing multimode interference devices. Proc. IEEE LEOS Annual
Meeting, 617–618 (1993)
Pennings, E.C.M., van Roijen, R., van Stralen, M.J.N., de Waard, P.J.,
Koumans, R.G.M.P., Verbeek, B.H.: Reflection properties of multimode
interference devices. IEEE Photon. Technol. Lett. 6, 715–718 (1994)
Pereira, S., Sipe, J.E., Heebner, J.E., Boyd, R.W.: Gap solitons in a two-
channel microresonator structure. Opt. Lett. 27, 536–538 (2002)
Poon, J.K.S., Huang, Y., Paloczi, G.T., Yariv, A.: Soft lithography replica
molding of critically coupled polymer microring resonators. IEEE Photon.
Technol. Lett. 16, 2496–2498 (2004a)
Poon, J.K.S., Huang, Y., Paloczi, G.T., Yariv, A., Zhang, C., Dalton, L.R.:
Wide-range tuning of polymer microring resonators by the photobleaching
of CLD-1 chromophores. Opt. Lett. 29, 2584–2586 (2004b)
Poon, J.K.S., Scheuer, J., Mookherjea, S., Paloczi, G.T., Huang, Y., Yariv,
A.: Matrix analysis of microring coupled-resonator optical waveguides. Opt.
Express 12, 90–103 (2004c)
Poon, J.K.S., Scheuer, J., Xu, Y., Yariv, A.: Designing coupled-resonator opt-
ical waveguide delay lines. J. Opt. Soc. Am. B 21, 1665–1673 (2004d)
Poon, J.K.S. Scheuer, J., Yariv, A.: Wavelength-selective reflector based on
a circular array of coupled microring resonators. IEEE Photon. Technol.
Lett. 16, 1331–1333 (2004e)
Pradhan, S., Schmidt, B., Martinez, L., Xu, Q., Almeida, V.R., Barrios,
C., Lipson, M.: Electro-optic modulator on silicon-on-insulator substrates
using ring resonators. Proc. CLEO 1, 107–109 (2005)
Quidant, R., Weeber, J.C., Dereux, A., Leveque, G., Weiner, J., Girard, C.:
Addressing and imaging microring resonators with optical evanescent light.
Phys. Rev. B 69, 081402(R) (2004)
References 249
Rabiei, P., Steier, W.H., Zhang, C., Dalton, L.R.: Polymer Micro-ring filters
and modulators. IEEE J. Lightwave Technol. 20, 1968–1975 (2002a)
Rabiei, P., Steier, W.H., Zhang, C., Wang, C.G., Lee, H.J.: Polymer micro-ring
modulator with 1 THz FSR. Proc. CLEO Post Deadline Paper (2002b)
Rabiei, P.: Electro-optic and thermo-optic polymer micro-ring resonators and
their applications. Ph.D. Thesis, University of Southern California (2003)
Rabiei, P., Steier, W.H.: Tunable polymer double micro-ring filters. IEEE
Rabiei, P.: Calculation of losses in micro-ring resonators with arbitrary refrac-
tive index or shape profile and its applications. IEEE J. Lightwave Technol.
23, 1295–1301 (2005)
Raburn, M., Liu, B., Abraham, P., Bowers, J.E.: Double-bonded InP–
InGaAsP vertical coupler 1:8 beam splitter. IEEE Photon. Technol. Lett.
12, 1639–1641 (2000)
Rabus, D.G.: Realization of optical filters using ring resonators with integ-
rated semiconductor optical amplifiers in GaInAsP/InP. Der Andere Verlag
(2002)
Rabus, D.G., Hamacher, M.: MMI-coupled ring resonators in GaInAsP–InP,
Rabus, D.G., Hamacher, M., Heidrich, H.: Resonance frequency tuning of a
double ring resonator on GaInAsP/InP: Experiment and simulation. Japan.
J. Appl. Phys. 41, 1186–1189 (2002a)
Rabus, D.G., Hamacher, M., Heidrich, H., Troppenz, U.: High Q channel
dropping filters using ring resonators with integrated SOAs. IEEE Photon.
Technol. Lett. 10, 1442–1444 (2002b)
Rabus, D.G., Hamacher, M., Troppenz, U., Heidrich, H.: Optical filters based
on ring resonators with integrated semiconductor optical amplifiers in
GaInAsP/InP. IEEE J. Select. Topics Quantum Electron. 6, 1405–1411
(2002c)
Rabus, D.G., Bian, Z., Shakouri, A.: A GaInAsP/InP double ring resonator
coupled laser. IEEE Photon. Technol. Lett. 9, 1770–1772 (2005a)
Rabus, D.G., Henzi, P., Mohr, J.: Photonic integrated circuits by DUV-
induced modification of polymers. IEEE Photon. Technol. Lett. 3, 591–593
(2005b)
Rafizadeh, D.: Experimental realization of nanofabricated semiconduc-
tor waveguide-coupled microcavity ring and disk optical resources.
Ph.D. Thesis, Northwestern University (1997)
Rafizadeh, D., Zhang, J.P., Hagness, S.C., Taflove, A., Stair, K.A., Ho, S.T.:
Temperature tuning of microcavity ring and disk resonators at 1.5 µm. Proc.
IEEE LEOS Ann. Meet. 162–163 (1997a)
Rafizadeh, R., Zhang, J.P., Hagness, S.C., Taflove, A., Stair, K.A., Ho, S.T.,
Tiberio, R.C.: Waveguide-coupled AlGaAs/GaAs microcavity ring and disk
resonators with high finesse and 21.6 nm free spectral range. Opt. Lett. 22,
1244–1246 (1997b)
250 References
Rafizadeh, D., Zhang, J.P., Tiberio, R.C., Ho, S.T.: Propagation loss mea-
surements in semiconductor microcavity ring and disk resonators. IEEE J.
Lightwave Technol. 16, 1308–1314 (1998)
Rayleigh: The problem of the whispering gallery. Cambridge University Scien-
tific Papers 5, 617–620 (1912). Also published in the Philos. Mag. 20, 1001–
1004, (1910)
Regener, R., Sohler, W.: Loss in low finesse Ti : LiNbO3 optical waveguide
resonators. Appl. Phys. B 36, 143–147 (1985)
Rezzonico, D., Guarino, A., Herzog, C., Jazbinsek, M., Gunter, P.: High-finesse
laterally coupled organic–inorganic hybrid polymer microring resonators for
VLSI photonics. IEEE Photon. Technol. Lett. 18, 865–867 (2006)
Rommel, S.L., Jang, J.H., Lu, W., Cueva, G., Zhou, L., Adesida, I., Pajer,
G., Whaley, R., Lepore, A., Schellanbarger, Z., Abeles, J.H.: Effect of H2
on the etch profile of InP/InGaAsP alloys in Cl2 /Ar/H2 inductively cou-
pled plasma reactive ion etching chemistriesfor photonic device fabrication.
J. Vac. Sci. Technol. B 20, 1327–1330 (2002)
Sadagopan, T., Choi, S.J., Choi, S.J., Dapkus, P.D., Bond, A.E.: Optical
modulators based on depletion width translation in semiconductor micro-
disk resonators. IEEE Photon. Technol. Lett. 17, 567–569 (2005)
Sakamoto, S.R., Ozturk, C., Byun, Y.T., Ko, J., Dagli, N.: Low-loss substrate-
removed (SURE) optical waveguides in GaAs–AlGaAs epitaxial layers
embedded in organic polymers. IEEE Photon. Technol. Lett. 10, 985–987
(1998)
Savchenkov, A.A., Ilchenko, V.S., Matsko, A.B., Maleki, L.: Tunable filter
based on whispering gallery modes. Electron. Lett. 39, 389–391 (2003)
Schwelb, O.: Transmission, group delay, and dispersion in single-ring optical
resonators and add/drop filters – a tutorial overview. IEEE J. Lightwave
Technol. 22, 1380–1394 (2004)
Schwelb, O., Frigyes, I.: All-optical tunable filters built with discontinuity-
assisted ring resonators. IEEE J. Lightwave Technol. 19, 380–386 (2001)
Seo, C., Chen, J.C.: Low transition losses in bent rib waveguides. IEEE J.
Lightwave Technol. 14, 2255–2259 (1996)
Slusher, R.E., Levi, A.F.J., Mohideen, U., McCall, S.L., Pearton, S.J., Logan,
R.A.: Threshold characteristics of semiconductor microdisk lasers. Appl.
Phys. Lett. 63, 1310–1312 (1993)
Smit, M.K., Pennings, E.C.M., Blok, H.: A normalized approach to the design
of low-loss optical waveguide bends. IEEE J. Lightwave Technol. 11, 1737–
1742 (1993)
Soldano, L.B., Pennings, E.C.M.: Optical multi-mode interference devices
based on self-imaging: Principles and applications. IEEE J. Lightwave Tech-
nol. 13, 615–627 (1995)
Song, D.S., Hwang, J.K., Kim, C.K., Han, I.Y., Jang, D.H., Lee, Y.H.:
InGaAsP microdisk lasers on AlxOy. IEEE Photon. Technol. Lett. 12, 954–
956 (2000)
References 251
Song, J.H., Park, J.W., Sim, E.D., Baek, Y.: Measurements of coupling and
reflection characteristics of butt-joints in passive waveguide integrated laser
diodes. IEEE Photon. Technol. Lett. 17, 1791–1793 (2005)
Sorel, M., Gluck, S., Laybourn, P.J.R.: Semiconductor double ring waveguide
resonators. Electron. Lett. 35, 1551–1552 (1999)
Sorel, M., Laybourn, P.J.R., Giuliani, G., Donati, S.: Unidirectional bistability
in semiconductor waveguide ring lasers. Appl. Phys. Lett. 80, 3051–3053
(2002)
Sorel, M., Giuliani, G., Scirè, A., Miglierina, R., Donati, S., Laybourn, P.J.R.:
Operating regimes of GaAs–AlGaAs semiconductor ring lasers: Experiment
and model. IEEE J. Quantum Electron. 39, 1187–1195 (2003)
Sparacin, D.K., Lock, J.P., Hong, C.Y., Gleason, K.K., Kimerling, L.C.,
Michel, J.: Trimming of silicon nitride microring resonators with a polysi-
lane top cladding. Proc. Group IV Photonics, 117–119 (2005)
Spiekman, L.H., Oei, Y.S., Metaal, E.G., Groen, F.H., Moerman, I.,
Smit, M.K.: Extremely small multimode interference couplers and ultra-
short bends on InP by deep etching. IEEE Photon. Technol. Lett. 6, 1008–
1010 (1994)
Subramaniam, V., De Brabander, G.N., Naghski, D.H., Boyd, J.T.: Measure-
ment of mode field profiles and bending and transition losses in curved
optical channel waveguides. IEEE J. Lightwave Technol. 15, 990–997 (1997)
Sumetsky, M.: Vertically-stacked multi-ring resonator. Opt. Express 13, 6354–
6375 (2005)
Suzuki, S., Shuto, K., Hibino, Y.: Integrated-optic ring resonators with two
stacked layers of silica waveguide on Si. IEEE Photon. Technol. Lett. 4,
1256–1258 (1992)
Suzuki, S., Yanagisawa, M., Hibino, Y., Oda, K.: High-density integrated pla-
nar lightwave circuits using SiO2 –GeO2 waveguides with a high refractive
index difference. IEEE J. Lightwave Technol. 12, 790–796 (1994)
Suzuki, S., Oda, K., Hibino, Y.: Integrated-optic double-ring resonators with
a wide free spectral range of 100 GHz. IEEE J. Lightwave Technol. 13,
1766–1771 (1995)
Suzuki, S., Kokubun, Y., Nakazawa, M., Yamamoto, T., Chu, S.T.: Ultrashort
optical pulse transmission characteristics of vertically coupled microring
resonator add/drop filter. IEEE Photon. Technol. Lett. 19, 266–271 (2001)
Suzuki, S., Hatakeyama, Y., Kokubun, Y., Chu, S.T.: Precise control of wave-
length channel spacing of microring resonator add–drop filter array. IEEE
J. Lightwave Technol. 20, 745–750 (2002)
Tan, F.: Integrated optical filters based on microring resonators. Ph.D. Thesis,
University of Twente, ISBN 90-365-2023-1 (2004)
Tazawa, H., Steier, W.H.: Bandwidth of linearized ring resonator assisted
Mach–Zehnder modulator. IEEE Photon. Technol. Lett. 17, 1851–1853
(2005)
252 References
Tazawa, H., Steier, W.H.: Analysis of ring resonator-based traveling-wave

modulators. IEEE Photon. Technol. Lett. 18, 211–213 (2006)
Tee, C.W., Williams, K.A., Penty, R.V., White, I.H.: Fabrication-tolerant
active–passive integration scheme for vertically coupled microring resonator.
IEEE J. Select. Topics Quantum Electron. 12, 108–116 (2006)
Tishinin, D.V., Dapkus, P.D., Bond, A.E., Kim, I., Lin, C.K., O’Brien, J.:
Vertical resonant couplers with precise coupling efficiency control fabricated
by wafer bonding. IEEE Photon. Technol. Lett. 11, 1003–1005 (1999)
Tredicucci, A., Gmachl, C., Capasso, F., Sivco, D.L., Chu, S.N.G., Hutchinson,
A.L., Cho, A.Y.: Very long wavelength (λ = 16 µm) whispering gallery mode
microdisk lasers. Electron. Lett. 36, 328–330 (2000)
Tsai, C.H., Lee, J.H., Chiang, H.J., Yang, C.C., Shih, H.C., Chen, B.C.,
Chuang, T.J., Chang. Y.: Semiconductor circular ring lasers fabricated with
the cryo-etching technique. IEEE Photon. Technol. Lett. 10, 751–753 (1998)
Vahala, K.J.: Optical microcavities. Nature 424, 839–846 (2003)
Vahala, K.J.: Optical microcavities (advanced series in applied physics). World
Scientific Publishing (2004)
van Damy Spiekman, L.H., van Ham, F.P.G.M., Groen, F.H., van der
Tol. J.J.G.M., Moerman, I., Pascher, W.W., Hamacher, M., Heidrich, H.,
Weinert, C.M., Smit, M.K.: Novel compact polarization converters based
on ultra short bends. IEEE Photon. Technol. Lett. 8, 1346–1348 (1996)
van Roijen, R., Pennings, E.C.M., van Stalen, M.J.N., van Dongen, T.,
Verbeek, B.H., van der Heijden, J.M.M.: Compact InP-based ring lasers
employing multimode interference couplers and combiners. Appl. Phys.
Lett. 64, 1753–1755 (1994)
Van, V., Absil, P.P., Hryniewicz, J.V., Ho, P.T.: Propagation loss in single-
mode GaAs–AlGaAs microring resonators: Measurement and model. IEEE
J. Lightwave Technol. 19, 1734–1739 (2001)
Van, V., Ibrahim, T.A., Absil, P.P., Johnson, F.G., Grover, R., Ho, P.T.: Opti-
cal signal processing using nonlinear semiconductor microring resonators.
IEEE J. Select. Topics Quantum Electron. 8, 705–713 (2002a)
Van, V., Ibrahim, T.A., Ritter, K., Absil, P.P., Johnson, F.G., Grover, R.,
Goldhar, J., Ho, P.T.: All-optical nonlinear switching in GaAs–AlGaAs
microring resonators. IEEE Photon. Technol. Lett. 14, 74–76 (2002b)
Van, V., Little, B.E., Chu, S.T., Hryniewicz, J.V.: Micro ring resonator filters.
Proc. LEOS Ann. Meet. 2, 571–572 (2004)
Vander Rhodes, G.H., Goldberg, B.B., Ünlü, M.S., Chu, S.T., Little, B.E.:
Internal spatial modes in glass microring resonators. IEEE J. Select. Topics
Quantum Electron. 6, 46–53 (2000)
Vanderhaegen, B., Van Thourhout, D., De Merlier, J., Sarlet, G.,
Vanwassenhove, L., Moerman, I., Van Daele, P., Baets, R.: High Q InGaAsP
ring resonator filters. Proc. 9th Eur. Conf. Integrated Opt. Tech. Exhibition
(ECIO 99), Italy, 381–384 (1999)
References 253
Vawter, G.A., Mar, A., Hietala, V., Zolper, J., Hohimer, J.: All optical
millimeter-wave electrical signal generation using an integrated mode-locked
semiconductor ring laser and photodiode. IEEE Photon. Technol. Lett. 9,
1634–1636 (1997)
Vörckel, A., Mönster, M., Henschel, W., Bolivar, P.H., Kurz, H.: Asymmetri-
cally coupled silicon-on-insulator microring resonators for compact add–
drop multiplexers. IEEE Photon. Technol. Lett. 15, 921–923 (2003)
Wang, R.P., Dumitrescu, M.M.: Optical Modes in Semiconductor Microdisk
Lasers. IEEE J. Quantum Electron. 34, 1933–1937 (1998)
Wang, T.J., Huang, Y.H., Chen, H.L.: Resonant-wavelength tuning of micro-
ring filters by oxygen plasma treatment. IEEE Photon. Technol. Lett. 17,
582–584 (2005)
Watts, M.R., Qi, M., Barwicz, T., Popovic, M., Rakich, P., Socci, L., Ippen,
E.P., Kaertner, F., Smith, H.I.: Towards polarization independent high-
index contrast microphotonics. Proc. 28th URSI General Assembly (2005)
Weber, H.P., Ulrich, R.: A thin film ring laser. Appl. Phys. Lett. 19, 38–40
(1971)
Weiershausen, W., Zengerle, R.: Photonic highway switches based on ring
resonators used as frequency-selective components. Appl. Opt. 35, 5967–
5978 (1996)
Weinert, C.M., Agrawal, N.: Three-dimensional finite difference simulation of
coupling behavior and loss in multimode interference devices. IEEE Photon.
Technol. Lett. 7, 529–531 (1995)
West, B.R., Honkanen, S.: MMI devices with weak guiding designed in three
dimensions using a genetic algorithm. Opt. Express 12, 2716–2722 (2005)
Xu, Q., Schmidt, B., Pradhan, S., Lipson, M.: Micrometre-scale silicon electro-
optic modulator. Nature 435, 325–327 (2005)
Yalçin, A., Popat, K.C., Aldridge, J.C., Desai, T.A., Hryniewicz, J, Chbouki, N.,
Little, B.E., King, O., Van, V., Chu, S.T., Gill, D., Anthes-Washburn, M.,
Ünlü, M.S., Goldberg, B.B.: Optical sensing of biomolecules using microring
resonators. IEEE J. Select. Topics Quantum Electron. 12, 148–155 (2006)
Yan, X., Masanovic, M.L., Skogen, E.J., Hu, Z., Blumenthal, D.J., Coldren,
L.A.: Optical mode converter integration with InP-InGaAsP active and
passive waveguides using a single regrowth process. IEEE Photon. Technol.
Lett. 14, 1249–1251 (2002)
Yanagase, Y., Suzuki, S., Kokubun, Y., Chu, S.T.: Box-likefilter response and
expansion of FSR by a vertically triple coupled microring resonator filter.
IEEE J. Lightwave Technol. 20, 1525–1529 (2002a)
Yanagase, Y., Suzuki, S., Kokubun, Y., Chu, S.T.: Vertical triple series-
coupled microring resonator filter for passband flattening and expansion
of free spectral range. Japan. J. Appl. Phys. 41, L141–L143 (2002b)
Yariv, A.: Critical coupling and its control in optical waveguide-ring resonator
systems. IEEE Photon. Technol. Lett. 14, 483–485 (2002a)
254 References
Yariv, A.: Universal relations for coupling of optical power between microreso-
nators and dielectric waveguides. Electron. Lett. 36, 321–322 (2002b)
Yu, S., Krauss, T.F., Laybourn, P.J.R.: Mode locking in large monolithic
semiconductor ring lasers. SPIE Opt. Eng. 37, 1164–1168 (1998)
Zeng, K.C., Dai, L., Lin, J.Y., Jiang, H.X.: Optical resonance modes in
InGaN/GaN multiple-quantum-well microring cavities. Appl. Phys. Lett.
75, 2563–2565 (1999)
Index
add–drop filter, 6 quadratic dispersion, 166

apodization, 36, 37 structural dispersion, 166
waveguide dispersion, 166
bandpass, 29
birefringence, 210 etching, 46
box-like, 17 anisotropic, 47
buildup factor, 10 isotropic, 47
buried heterostructure, 187, 196
Fabry–Perot method, 101
Butterworth, 25, 29
filters, 126
cavity ring-down time, 174, 203 finesse F , 10
chain matrix, 14, 26 four wave mixing, 197
characteristic matrix, 26, 30 free carrier plasma effect, 197
Chebyshev, 25, 29, 30 free spectral range (FSR), 8
Chebyshev type I, 25 frequency beating, 183
Chebyshev type II, 25 full width at half maximum (FWHM), 9
conformal transformation, 120
grating, 23
coupling
group velocity, 11
asymmetrical coupling, 50
group-velocity dispersion, 39
critical coupling, 6
group-velocity reduction, 39
symmetric coupling, 7
cut-back, 99, 101 intensity enhancement factor, 10
delay lines, 203 lateral coupling, 41

differential group delay, 169 Lorentzian function, 134
dispersion, 165
chromatic dispersion, 166 notch filter, 3
dispersion compensation, 125
polarization, 96
intermodal dispersion, 166
material dispersion, 166 quality factor Q, 10
modal dispersion, 166 quantum cascade laser, 219
negative dispersion, 166, 167 quantum confined Stark effect, 208
polarization mode dispersion, 166
positive dispersion, 166, 167 racetrack, 16
256 Index
saturable absorber, 187 tuning enhancement factor, 192

scattering matrix, 14 two-photon absorption, 204, 207
shape factor, 141
silicon on insulator, 47 Vernier effect, 139, 184, 192, 195
Single photon absorption, 208 vertical coupling, 41
Taylor series expansion, 13 wafer bonding, 42

time-dependent relations, 11 adhesion bonding, 42
decay time-constant, 11 anodic bonding, 44
total internal reflection, 81, 148, 215 electrostatic bonding, 44
transfer matrix, 14, 26 eutectic bonding, 42
tuning fusion bonding, 43
carrier injection, 151, 158 hydrophobic bonding, 43
electro-optic tuning, 150, 156 wavelength reflective filter, 23
free carrier injection, 151, 197 reflector, 32, 195
photobleaching technique, 163
thermo-optic tuning, 150 Z-transformation, 12
Springer Series in
optical sciences
Volume 1
1 Solid-State Laser Engineering
By W. Koechner, 5th revised and updated ed. 1999, 472 figs., 55 tabs., XII, 746 pages
Published titles since volume 90

100 Quantum Interference and Coherence
Theory and Experiments
By Z. Ficek and S. Swain, 2005, 178 figs., XV, 418 pages
101 Polarization Optics in Telecommunications
By J. Damask, 2005, 110 figs., XVI, 528 pages
102 Lidar
Range-Resolved Optical Remote Sensing of the Atmosphere
By C. Weitkamp (Ed.), 161 figs., XX, 416 pages
103 Optical Fiber Fusion Splicing
By A.D. Yablon, 2005, 137 figs., XIII, 306 pages
104 Optoelectronics of Molecules and Polymers
By A. Moliton, 2005, 229 figs., 592 pages
105 Solid-State Random Lasers
By M. Noginov, 2005, 131 figs., XII, 238 pages
106 Coherent Sources of XUV Radiation
Soft X-Ray Lasers and High-Order Harmonic Generation
By P. Jaeglé, 2006, 332 figs., XIII, 416 pages
107 Optical Frequency-Modulated Continuous-Wave (FMCW) Interferometry
By J. Zheng, 2005, 137 figs., XVIII, 254 pages
108 Laser Resonators and Beam Propagation
Fundamentals, Advanced Concepts and Applications
By N. Hodgson and H. Weber, 2005, 587 figs., XXVI, 794 pages
109 Progress in Nano-Electro Optics IV
Characterization of Nano-Optical Materials and Optical Near-Field Interactions
By M. Ohtsu (Ed.), 2005, 123 figs., XIV, 206 pages
110 Kramers–Kronig Relations in Optical Materials Research
By V. Lucarini, J.J. Saarinen, K.-E. Peiponen, E.M. Vartiainen, 2005,
37 figs., X, 162 pages
111 Semiconductor Lasers
Stability, Instability and Chaos
By J. Ohtsubo, 2005, 169 figs., XII, 438 pages
112 Photovoltaic Solar Energy Generation
By A. Goetzberger and V.U. Hoffmann, 2005, 139 figs., XII, 234 pages
113 Photorefractive Materials and Their Applications 1
Basic Effects
By P. Günter and J.P. Huignard, 2006, 169 figs., XIV, 421 pages
Materials
By P. Günter and J.P. Huignard, 2005, 100 figs., approx. XII, 300 pages
Applications
By P. Günter and J.P. Huignard, 2005, 100 figs., approx. XII, 300 pages
Springer Series in
optical sciences
116 Spatial Filtering Velocimetry
Fundamentals and Applications
By Y. Aizu and T. Asakura, 2006, 112 figs., XII, 212 pages
117 Progress in Nano-Electro-Optics V
Nanophotonic Fabrications, Devices, Systems, and Their Theoretical Bases
By M. Ohtsu (Ed.), 2006, 122 figs., XIV, 188 pages
118 Mid-infrared Semiconductor Optoelectronics
By A. Krier (Ed.), 2006, 443 figs., XVIII, 751 pages
119 Optical Interconnects
The Silicon Approach
By L. Pavesi and G. Guillot (Eds.), 2006, 265 figs., XXII, 389 pages
120 Relativistic Nonlinear Electrodynamics
Interaction of Charged Particles with Strong and Super Strong Laser Fields
By H.K. Avetissian, 2006, 23 figs., XIII, 333 pages
121 Thermal Processes Using Attosecond Laser Pulses
When Time Matters
By M. Kozlowski and J. Marciak-Kozlowska, 2006, approx. 46 figs., XII, 232 pages
122 Modeling and Analysis of Transient Processes in Open Resonant Structures
New Methods and Techniques
By Y.K. Sirenko, N.P. Yashina, and S. Ström, 2007, approx. 110 figs., XIV, 346 pages
123 Wavelength Filters in Fibre Optics
By H. Venghaus (Ed.), 2007, approx. 210 figs., XXIV, 451 pages
124 Light Scattering by Systems of Particles
Null-Field Method with Discrete Sources: Theory and Programs
By A. Doicu, T. Wriedt, and Y.A. Eremin, 2006, 123 figs., XIII, 318 pages
125 Electromagnetic and Optical Pulse Propagation 1
Spectral Representations in Temporally Dispersive Media
By K.E. Oughstun, 2007, approx 74 figs., 452 pages
126 Quantum Well Infrared Photodetectors
Physics and Applications
By H. Schneider and H.C. Liu, 2006, 153 figs., XVI, 254 pages
127 Integrated Ring Resonators
The Compendium
By D.G. Rabus, 2007, approx. 243 figs., XVII, 302 pages
128 High Power Diode Lasers
Technology and Applications
By R. Poprawe, P. Loosen, and F. Bachmann (Eds.) 2007, approx. 535 figs., VI, 546 pages
129 Laser Ablation and its Applications
By C.R. Phipps (Ed.) 2007, approx. 300 figs., XX, 588 pages
130 Concentrator Photovoltaics
By A. Luque and V. Andreev (Eds.) 2007, approx. 242 figs., XV, 298 pages

2007 Book IntegratedRingResonators PDF

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

2007 Book IntegratedRingResonators PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2007 Book IntegratedRingResonators PDF

Uploaded by

Copyright:

Available Formats

optical sciences

Editor-in-Chief: W. T. Rhodes, Atlanta

Integrated Ring Resonators

With 243 Figures and 8 Tables

Library of Congress Control Number: 2006940350

APF all pass ﬁlter

RTFT real-time Fourier transformation

2 Ring Resonators: Theory and Modeling . . . . . . . . . . . . . . . . . . . . 3

3 Materials, Fabrication, and Characterization Methods . . . . . 41

3.4.5 Lateral Coupling in AlGaAs/GaAs . . . . . . . . . . . . . . . . . . 71

4 Building Blocks of Ring Resonator Devices . . . . . . . . . . . . . . . . 111

6 Whispering Gallery Mode Devices . . . . . . . . . . . . . . . . . . . . . . . . . 215

integrated semiconductor optical ampliﬁers (SOA) are presented, leading to

2.1 Single Ring Resonators

The Basic Conﬁguration

The basic conﬁguration, which consists of unidirectional coupling between a

Fig. 2.1. Ring resonator channel dropping ﬁlter

Fig. 2.2. Model of a single ring resonator with one waveguide

Ei2 = α · ejθ Et2 , (2.3)

On resonance, (θ+ϕt ) = 2πm, where m is an integer, the following is obtained:

−κ1 κ∗1 t∗2 α1/2 ejθ|t1 | +|κ1 |2 =1

−κ∗1 κ2 α1/2 ejθ1/2

Fig. 2.3. Notch type ring resonator ﬁlter characteristic

Fig. 2.4. Model of a basic add–drop single ring resonator ﬁlter

At resonance, the output power from the drop port is given by

Fig. 2.5. Add–drop ring resonator ﬁlter characteristic

into more segments to account for diﬀerent materials or modiﬁed waveguide

Ring Resonator Parameters

The next parameter of importance is the resonance width which is deﬁned

Using (2.5) and (2.17) to translate into the wavelength domain

For a conﬁguration like Fig. 2.4, the buildup factor is given by

On resonance, the intensity enhancement factor is

For a lossless resonator and setting κ1 = κ2 = κ which is 1, B can be

The Time-Dependent Relations

Before going on to extend the previous model incorporating additional ring

Another approach to simulate ring resonator ﬁlters is by using the

where z is a complex variable and h(n) is the impulse response of a ﬁlter or

the unit circle causes unity transmission at the corresponding frequency. As

The Z-transform for this type of ﬁlter is then given by

There is a single pole at t1 t2 α. As the matrix is symmetric, the transfer

Fig. 2.6. Z-transform layout of an add–drop ring resonator ﬁlter

Using the diﬀerent simulation techniques described in this chapter it is

Incorporating Loss in Ring Resonator Filters

In the previous simulation models, loss is basically included by insertion of the

2.1.2 Racetrack-Shaped Resonators

The equations in the previous chapter described the general behavior of a

For the throughput port:

For the drop port

2.2 Double Ring Resonators

2.2.1 Serially Coupled Double Ring Resonator

The schematic of a serially coupled double ring resonator is depicted in

Fig. 2.8. Two ring resonators coupled in series

Fig. 2.9. Index proﬁle and geometry of coupled waveguides

mλm = 2πr1,2 neﬀ . (2.71)

κ21 vg1 κ2 vg1 vg2

µ22 = 0.250µ41 . (2.75)

Assuming identical rings, this leads to the coupling coeﬃcient