Introduction To Non-Kerr Optical Solitons

P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000

P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
Chapman & Hall/CRC

Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2007 by Taylor & Francis Group, LLC
Chapman & Hall/CRC is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works

Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4 3 2 1
International Standard Book Number-10: 1-58488-638-2 (Hardcover)

International Standard Book Number-13: 978-1-58488-638-9 (Hardcover)
This book contains information obtained from authentic and highly regarded sources. Reprinted
material is quoted with permission, and sources are indicated. A wide variety of references are
listed. Reasonable efforts have been made to publish reliable data and information, but the author
and the publisher cannot assume responsibility for the validity of all materials or for the conse-
quences of their use.
No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any
electronic, mechanical, or other means, now known or hereafter invented, including photocopying,
microfilming, and recording, or in any information storage or retrieval system, without written
permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.
copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC)
222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that
provides licenses and registration for a variety of users. For organizations that have been granted a
photocopy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and
are used only for identification and explanation without intent to infringe.
Library of Congress Cataloging-in-Publication Data
Biswas, Anjan, Dr.

Introduction to non-Kerr law optical solitons / by Anjan Biswas and Swapan
Konar.
p. cm. -- (Chapman & Hall/CRC applied mathematics and nonlinear
science series)
Includes bibliographical references and index.
ISBN 1-58488-638-2 (alk. paper)
1. Solitons. 2. Nonlinerar waves. 3. Nonlinear optics. 4. Optical
communications. I. Konar, Swapan. II. Title. III. Series.
QC174.26.W28B56 2006
530.12’4--dc22 2006049558
Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
Preface
Recent years have witnessed an explosion of research activities in the field of

soliton propagation in nonlinear optical media. These activities are motivated
by the fact that optical solitons, both temporal and spatial variety, do have
practical relevance in the latest communication technology based on gener-
ation and transportation of localized optical pulses. To date, these optical
pulses have already made tera bit/s transmission through optical fibers fea-
sible in the laboratory. In addition, they also play significant roles in several
other technologically relevant aspects, such as optical switching and signal
processing.
Early investigations in these areas began with Kerr law nonlinearity in
which the refractive index of the medium is proportional to the light intensity.
With further development of the subject, several forms of nonlinearity have
come under investigation. Notable among these forms of nonlinearity are
parabolic or cubic-quintic, power law, dual-power law, and saturating non-
linearities. These nonlinearities reveal many new and interesting behaviors
hitherto unknown in Kerr law of nonlinearity. However, in spite of tremen-
dous progress in these areas in the last 2 decades, and despite the fact that
several important and extremely well-written books are now available that
deal with soliton propagation in optical communication systems and allied
areas, no textbook of worth deals exclusively with soliton propagation in
media that possess non-Kerr law nonlinearities. Thus, to bridge the gap be-
tween availability and nonavailability, we felt the need to bring out a book
exclusively dealing with optical soliton propagation in non-Kerr law media.
This book is organized as follows: Chapter 1 presents an introduction to
the field of fiber optics and basic features of fiber-optic communications. The
nonlinear Schrödinger’s equation (NLSE) has been introduced and mathe-
matical aspects, including conserved quantities, have been outlined in
chapter 2. In this chapter, we have also introduced the perturbation to the
NLSE. Adiabatic dynamics of soliton parameters have also been introduced.
Finally, we have discussed the concept of quasi-stationary solitons and their
influences in this chapter. In chapter 3, we have derived the NLSE for Kerr law
nonlinearity from basic principle. The inverse scattering transform has been
outlined and, using this principle, the 1-soliton solution has been obtained
in this chapter. In addition, we have explained the variational principle and
Lie transform, which are used to integrate the NLSE with Hamiltonian-type
perturbation. The non-Kerr law solitons have been discussed in chapters 4
through 7. Chapters 4, 5, 6, and 7 are respectively devoted to the study of
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
solitons with power law, parabolic law, dual-power law, and saturable law
nonlinearities. In each case, we have developed the soliton dynamics, eval-
uated integrals of motion, and devoted enough space to develop adiabatic
dynamics of perturbed quantities based on multiple-scale perturbation the-
ory. In addition, the existence of bistable soliton is discussed in chapter 7.
Chapter 8 is devoted to intrachannel collision of optical solitons in the pres-
ence of perturbation terms. Both Hamiltonian as well as non-Hamiltonian
type perturbations have been considered. The nonlinearities that are studied
in this chapter are Kerr, power, parabolic, and dual-power laws. In chapter 9,
the stochastic perturbation of optical solitons has been studied. The corre-
sponding Langevin equations are derived and analyzed for each of the laws
of nonlinearity, namely Kerr, power, parabolic, and dual-power laws. Optical
couplers are introduced in chapter 10. Twin core and multiple-core couplers
have been discussed. At the end of this chapter, we have briefly discussed
solitons in magneto-optic waveguides. The book concludes with chapter 11,
which treats an introduction to optical bullets.
This book is intended for graduate students at the master’s and doctoral
levels in applied mathematics, physics, and engineering. Undergraduate stu-
dents with senior standing in applied mathematics, physics, and engineering
will also benefit from this book. The prerequisite of this book is a knowledge
of partial differential equations, perturbation theory, and elementary physics.
Anjan Biswas, is extremely thankful to Dr. Michael Busby, director of
the Center of Excellence in Information Systems Engineering and Manage-
ment of Tennessee State University in Nashville, Tennessee, with which this
author was previously affiliated. Without constant encouragement and fi-
nancial support from Dr. Busby, this project would not have been possi-
ble. The first author is also extremely thankful and grateful to Dr. Tommy
Frederick, vice provost of research at Delaware State University, with which
this author is presently affiliated, for his constant encouragement. Without
these two persons’ blessings, this project would not have been possible.
Finally, the author is extremely grateful to his parents for all their uncondi-
tional love in his upbringing, blessings, education, support, encouragement,
and sacrifices throughout his life, till today. This author is deeply saddened
by the sudden death of his mother after a massive heart attack in Calcutta,
India, which occurred during the course of writing this book.
Swapan Konar is grateful to Prof. H. C. Pande, vice chancellor emeritus,
Birla Institute of Technology, India; Prof. S. K. Mukherjee, vice chancellor,
Birla Institute of Technology, and Prof. P. K. Barhai, head of the Department
of Applied Physics, Birla Institute of Technology, for encouragement and con-
stant support. Finally, he sincerely thanks his wife Tapati for her tolerance and
encouragement and his little son Argho for sacrificing his playtime.
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
Authors
Dr. Anjan Biswas obtained his BSc (Honors) in mathematics from St. Xavier’s
College, Calcutta, and subsequently earned his MSc and MPhil degrees in
applied mathematics from the University of Calcutta, India. After that, he ob-
tained his MA and PhD degrees in applied mathematics from the University
of New Mexico, Albuquerque. His current research interests are in nonlin-
ear optics, theory of solitons, plasma physics, and fluid dynamics. He is the
author of 100 refereed journal papers and also serves as an editorial board
member for three journals. Currently, he is an associate professor in the De-
partment of Applied Mathematics and Theoretical Physics of Delaware State
University in Dover.
Dr. S. Konar received his MSc degree in Nuclear Physics from University
of Kalyani, India, in 1982. He earned an MTech in energy management.
Dr. Konar has been awarded MPhil and PhD respectively in 1987 and 1990
by Jawaharlal Nehru University, New Delhi, India. At present, he is work-
ing as a professor of applied physics at Birla Institute of Technology, Mesra,
Ranchi, India. His current research interest is in the field of photonics and op-
toelectronics, particularly classical solitons, soliton propagation in dispersion-
managed optical communication systems, nonlinear optical waveguide,
induced focusing, self-focusing and all optical switching. He has published
64 research papers in international journals and presented 30 research papers
in conferences.
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Optical Waveguides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Types of Optical Fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.2 Advantages of Fiber-Optic Communications . . . . . . . . . . . . . . . 5
2 The Nonlinear Schrödinger’s Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.1 Nonlinearity Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Traveling Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Integrals of Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.4 Parameter Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4.1 Perturbation Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Quasi-Stationary Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5.1 Mathematical Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.5.2 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3 Kerr Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.1.1 The Nonlinear Schrödinger’s Equation . . . . . . . . . . . . . . . . . . . 29
3.2 Traveling Wave Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3 Inverse Scattering Transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3.1 1-Soliton Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.4.1 Hamiltonian Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.5 Variational Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.7 Lie Transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.7.2 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4 Power Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
5 Parabolic Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6 Dual-Power Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
7 Saturable Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.2 The NLSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
7.2.1 Conserved Quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.3 Bistable Solitons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.4 Arbitrary Pulse Propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
7.4.1 Lossless Uniform Media ( = 0) . . . . . . . . . . . . . . . . . . . . . . . . . . 94
7.4.2 Stationary Pulse Propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.4.3 Lossy Media ( = 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
8 Soliton–Soliton Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
8.2 Mathematical Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
8.2.1 Kerr Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.2.2 Power Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
8.2.3 Parabolic Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
8.2.4 Dual-Power Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
8.3 Quasi-Particle Theory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109
8.3.1 Kerr Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
8.3.1.1 Non-Hamiltonian Perturbations . . . . . . . . . . . . . . . . 114
8.3.1.2 Hamiltonian Perturbations . . . . . . . . . . . . . . . . . . . . . . 115
8.3.2 Power Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
8.3.2.2 Hamiltonian Perturbations . . . . . . . . . . . . . . . . . . . . . 121
8.3.3 Parabolic Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
8.3.4 Dual-Power Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
9 Stochastic Perturbation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
9.2 Kerr Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
9.3 Power Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
9.4 Parabolic Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
9.5 Dual-Power Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
10 Optical Couplers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .145

10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
10.1.1 Types of Couplers and Their Functions . . . . . . . . . . . . . . 146
10.1.1.1 Three- and Four-Port Couplers . . . . . . . . . . . . . 146
10.1.1.2 Star Coupler or Multiport Couplers . . . . . . . . 146
10.1.2 Optical Switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
10.2 Twin-Core Couplers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
10.3 Multiple-Core Couplers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
10.3.1 Coupling with Nearest Neighbors . . . . . . . . . . . . . . . . . . . 153
10.3.2 Coupling with All Neighbors . . . . . . . . . . . . . . . . . . . . . . . . 155
10.4 Magneto-Optic Waveguides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
10.4.1 Mathematical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
11 Optical Bullets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
11.2 1 + 3 Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
11.2.1 Integrals of Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
11.2.2 Parameter Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
12 Epilogue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Hints and Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
P1: Binaya Dash
October 6, 2006 18:2 C6382 C6382˙C000
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
1
Introduction
This introductory chapter is intended to provide a general overview of fiber

optics. It starts with a history of and current developments in fiber optics in
Section 1.1. Section 1.2 provides a brief account of types of optical waveguides
and the issues of fiber-optic communications.
1.1 History
The propagation of optical pulses, or solitons, through optical fibers has been a
major area of study given its potential applicability in optical communication
systems. The field of telecommunications has undergone a substantial evolu-
tion in the last couple of decades due to the impressive progress in the devel-
opment of optical fibers, optical amplifiers, and transmitters and receivers. In
a modern optical communication system, the transmission link is composed
of optical fibers and amplifiers that replace the electrical regenerators. How-
ever, the amplifiers introduce some noise and signal distortion that limit the
system capacity. Presently, the optical systems that show the best characteris-
tics in terms of simplicity, cost, and robustness against the degrading effects of
a link are those based on intensity modulation with direct detection (IM-DD).
Conventional IM-DD systems are based on the non-return-to-zero (NRZ) for-
mat, but for soliton-based transmission at higher data rates the return-to-zero
(RZ) format is used. Soliton-based transmission allows the exploitation of the
fiber capacity much more. [9].
The performance of optical system is limited by several effects that are
present in optical fibers and amplifiers. Signal propagation through optical
fibers can be affected by group velocity dispersion (GVD), polarization mode
dispersion (PMD), and nonlinear effects. The chromatic dispersion that is es-
sentially the GVD when waveguide dispersion is negligible is a linear effect
that introduces pulse broadening and generates intersymbol interference. The
PMD arises due to the fact that optical fibers for telecommunications have two
polarization modes, in spite of the fact that they are called monomode fibers.
These modes have two different group velocities that induce pulse broadening
1
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
2 Introduction to Non-Kerr Solitons
depending on the input signal state of polarization. The transmission impair-

ment due to PMD looks similar to that caused by GVD. However, PMD is
random whereas GVD is a deterministic process. So, PMD cannot be con-
trolled at the receiver. Newly√ installed optical fibers have quite low values of
PMD that are about 0.1 ps/ km.
The main nonlinear effects that arise in monomode fibers are Brillouin
scattering, Raman scattering, and the Kerr effect. Brillouin is a backward scat-
tering that arises from acoustic waves and can generate noise at the receiver.
Raman scattering takes place in both forward and backward directions from
silica molecules. The Raman gain response is characterized by low gain and
wide bandwidth, namely about 30 THz. The Raman threshold in conventional
fibers is of the order of 500 mW for copolarized pump and Stokes’ wave (that
is about 1 W for random polarization), thus making Raman effect negligible
for a single-channel signal. However, it becomes important for multichannel
wavelength-division-multiplexed (WDM) signals due to an extremely wide
band of wide gain curve.
The Kerr effect of nonlinearity is due to the dependence of the fiber refrac-
tive index on the field intensity. The intensity dependence of the refractive
index leads to a larger number of interesting nonlinear effects. Notable among
them, which have been studied widely, are self-phase modulation (SPM)
and cross-phase modulation (XPM). SPM refers to the self-induced nonlinear
phase shift experienced by an optical field during its propagation through an
optical fiber. SPM is responsible for spectral broadening. The SPM-induced
chirp combines with the linear chirp generated by the chromatic dispersion.
If the fiber dispersion coefficient is positive, namely in the normal disper-
sion regime, linear and nonlinear chirps have the same sign, whereas in an
anomalous dispersion regime, they are of opposite signs. In the former case,
pulse broadening is enhanced by SPM, while in the later case it is reduced. In
the anomalous dispersion case, the Kerr nonlinearity induces a chirp that can
compensate the degradation induced by GVD. Such a compensation is total
if soliton signals are used.
If multichannel WDM signals are considered, the Kerr effect can be more
degrading since it induces nonlinear cross-talk among the channels that are
known as XPM. In addition, WDM generates new frequencies called four-
wave mixing (FWM). The other issue in the WDM system is the collision-
induced timing jitter that is introduced due to the collision of solitons in
different channels. The XPM causes further nonlinear chirp that interacts
with the fiber GVD as in the case of SPM. The FWM is a parametric interaction
among waves that satisfies a particular relationship called phase-matching that
leads to power transfer among different channels.
To limit the FWM effect in a WDM, it is preferable to operate with a local
high GVD that is periodically compensated by devices having an opposite
GVD sign. One such device is a simple optical fiber with appropriate GVD,
and the method is commonly known as dispersion management. With this ap-
proach, the accumulated GVD can be very low and, at the same time, FWM
is strongly limited. Through dispersion management it is possible to achieve
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Introduction 3
the highest capacity for both RZ and NRZ signals. In that case, the overall
link dispersion has to be kept very close to zero, while a small amount of
chromatic anomalous dispersion is useful for the efficient propagation of a
soliton signal. It has been demonstrated with soliton signals that dispersion
management is very useful since it reduces collision-induced timing jitter and
also pulse interactions. It thus permits the achievement of higher capacities
than those allowed by the link having constant chromatic dispersion.
1.2 Optical Waveguides

One of the most promising applications of soliton theory is in the field of
optical communications. In optical communications systems, information is
encoded into light pulses and transmitted through optical fibers over long
distances. Commercial systems have been in operation since 1977 and a
transatlantic undersea optical cable has been developed. In 1973, Hasegawa
and Tappert [183, 184] proposed that soliton pulses could be used in optical
communications. However, the technology was not available until 7 years
later, at which time researchers at Bell Laboratories had experimentally
demonstrated the propagation of solitons in optical fibers.
Rapid developments in communications technology have occurred—
for example, the change from the use of wires to send signals (wire tele-
graphy) to wireless or radio telegraphy—leading to enormously increased
communication rates, measured by bits per second by a factor of 1 billion.
The latest in this series of advances is the optical fiber system in which large
amounts of information, coded as light pulses, pass along silica fibers. The
first transoceanic links, namely along the Atlantic and Pacific oceans, have
been established. As marvelous as these advances have been, the present
system still uses only a tiny fraction of the information-carrying capacity of
optical fibers.
Taking a look at waveguides, in particular an optical fiber, one can see
how solitons promise to revolutionize the field of telecommunications. The
main idea of a waveguide is to guide a beam of light by employing a varia-
tion of refractive index in the transverse direction so as to cause the light to
travel along a well-defined channel. The dependence of refractive index on
the transverse direction, the direction perpendicular to that in which the wave
propagates, can be continous or discontinous. The essential feature, however,
is that the refractive index is maximal in the channel along which one wishes
the light to be guided.
Figures 1.1(a) and 1.1(b) show the cross-section of an optical fiber. The
inner core consists of a special form of silica glass with very low absorption
and is between 10 to 60 µm in diameter. This core is surrounded by a glass
cladding whose refractive index, n2 , is very close to but slightly less than n1 ,
the linear refractive index of the inner core. This ensures that the wave is
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Jacket
b
Cladding
a
Core
n1
n2
a
b
Index
n0
Radial Distance
FIGURE 1.1
(a) & (b) Cross section of an optical fiber.
guided, namely its intensity is largely confined, to the inner core by virtue of
total internal reflection.
Two parameters characterize an optical fiber, namely the core-cladding
index difference () that is defined as:
n1 − n2
= (1.1)
n1
and the normalized frequency (V) that is defined as:

2πa
V= n21 − n22 (1.2)
λ
where a is the radius of the fiber core as shown in Figure 1.2 and λ is the
wavelength of light. The parameter V determines the number of modes sup-
ported by the fiber. For a V less than 2.405 the fibers support a single mode
and so the fibers that are designed to satisfy such conditions are known as
Protective Jacket
Cladding
(125 µm diam.)
Fiber Core
(10–50 µm diam.)
FIGURE 1.2
Structure of an optical fiber.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Introduction 5
1 0 1
Detection
Window
Pulse Width
Distance
Between Pulses
FIGURE 1.3
Soliton detection window and soliton train.
single-mode fibers. A typical multimode fiber would have the core radius as
a = 25–30 µm. However, for a single-mode fiber, a typical value of is
∼ 3 × 10−3 and requires a to be in the range of 2–4 µm. The value of the
outer radius b is less critical as long as it is large enough to confine the modes
entirely. Typically, b = 50–60 µm for both single-mode and multimode fibers
(Figure 1.2).
The basic idea of using optical fibers for communications is relatively sim-
ple. The message is coded in binary by representing one as a pulse-like modu-
lation of a carrier wave whose wavelength is in the micrometer (10−6 m) range
and whose frequency is in the terahertz (1014 Hz) range and representing zero
by the absence of such a pulse. The arrangement is shown in Figure 1.3. The
pulses are approximately 10–25 picoseconds (10−12 s) wide and the average
distance between them is four times that amount. Experimentally, fibers have
managed effective transmission rates in the gigabit range (109 bits/s).
1.2.1 Types of Optical Fibers

Based on the refractive index profile there are two types of optical fibers:
1. Step index fiber: In a step index fiber, the refractive index of the core
is uniform throughout and undergoes an abrupt or a step change at
the core-cladding boundary.
2. Graded index fiber: In a graded index fiber, the refractive index of the
core is made to vary in a parabolic manner such that the maximum
value of the refractive index is at the center of the core (Figure 1.4).
Propagating rays in the fiber can be classified as meridional and skew rays.
Meridional rays are confined to the meridional plane of the fiber, which are
planes that contain the axis of symmetry of the fiber. Skew rays are not con-
fined to a single plane. They propagate along the fiber.
1.2.2 Advantages of Fiber-Optic Communications

The various advantages of soliton communication through optical fibers are
enumerated here:
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Cladding
Cladding Core
Core
(a) (b)
FIGURE 1.4
(a) Step-index fiber and (b) graded-index fiber.
1. Wider bandwidth: The information-carrying capacity of a transmis-

sion system is directly proportional to the carrier frequency of the
transmitted signals. The optical carrier frequency is in the range of
1013 −1015 Hz while the radio wave frequency is about 106 Hz and
the microwave frequency is about 1010 Hz. Thus, the optical fiber
yields greater transmission bandwidth than conventional commu-
nication systems and the data rate or number of bits per second
is increased to a greater extent in the optical fiber communication
system.
2. Low transmission loss: Due to the usage of ultra-low-loss fibers and
erbium-doped silica fibers as optical amplifiers, one can achieve
almost lossless transmission. In modern optical fiber telecommuni-
cation systems, the fibers having a transmission loss of 0.2 dB/km
are used. Furthermore, using erbium-doped silica fibers over a short
length in transmission path selective points, appropiate optical am-
plification can be achieved. Thus, the repeater spacing is more than
100 km. Since the amplification is done in the optical domain itself,
the distortion produced during the strengthening of the signal is
almost negligible.
3. Dielectric waveguide: Optical fibers are made from silica, which is an
electrical insulator. Therefore, they do not pick up any electromag-
netic waves or any high-current lightning. They are also suitable
in explosive environments. Furthermore, the optical fibers are not
affected by any interference originating from power cables, railway
power lines, and radio waves. There is no cross-talk between the
fibers in a cable because of the absence of optical interference be-
tween the fibers.
4. Signal security: Signals transmitted through the fibers do not radiate.
In addition, signals cannot be tapped from a fiber in an easy manner.
Therefore, optical communication provides 100% signal security.
5. Small size and weight: Fiber-optic cables are developed with small
radii and are flexible, compact, and lightweight. They can be bent
or twisted without any damage. Optical fiber cables are superior to
copper cables in terms of storage, handling, installation, and trans-
portation, maintaining comparable strength and durability.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
2
The Nonlinear Schrödinger’s Equation
This chapter will talk about the mathematical aspects of the nonlinear
Schrödinger’s equation (NLSE) that governs the propagation of solitons
through an optical fiber. Section 2.1 is an introduction to NLSE. In Section 2.2,
the conserved quantities of the NLSE will be derived. In Section 2.3 the soliton
parameters will be introduced and the formulae for the adiabatic dynamics
of these parameters in the presence of the perturbation terms will be given.
Finally, in Section 2.4, the concept of quasi-stationarity will be introduced.
2.1 Introduction
The NLSE plays a vital role in various areas of physical, biological, and engi-
neering sciences. It appears in many applied fields, including fluid dynamics,
nonlinear optics, plasma physics, and protein chemistry. The NLSE that is go-
ing to be studied in this book is given by [86, 108, 399]
1
iq t + q xx + F (|q |2 )q = 0 (2.1)
2
In (2.1), F is a real-valued algebraic function and one needs to have
the smoothness of the complex function F (|q |2 )q : C → C. Considering
the complex plane C as a two-dimensional linear space R2 , it can be said
that the function F (|q |2 )q is k times continuously differentiable so that one
can write
∞

F (|q |2 )q ∈ C k ((−n, n) × (−m, m);R2 )
m,n=1
In equation (2.1), q is the dependent variable, x and t are the independent

variables, and the subscripts represent the partial derivative of q with respect
to that variable. So, q t stands for ∂q /∂t while q xx stands for ∂2 q /∂x 2 . The first
term in (2.1) represents the time evolution term, while the second term is due
to the group velocity dispersion and the third term accounts for nonlinearity.
Thus, equations of these types are sometimes known as nonlinear evolution
7
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
FIGURE 2.1
Profile of a soliton.
equations. This is a nonlinear partial differential equation that is not integrable,

in general. The nonintegrability is not necessarily related to the nonlinear
term in (2.1). Higher order dispersion, for example, can also make the system
nonintegrable while it still remains Hamiltonian.
Equation (2.1) has been shown to govern the evolution of a wave packet
in a weakly nonlinear and dispersive medium and has thus arisen insuch di-
verse fields as water waves, plasma physics, and nonlinear optics. One other
application of this equation is in pattern formation, where it has been used to
model some nonequilibrium pattern forming systems. In particular, this equa-
tion is now widely used in the optics field as a good model for optical pulse
propagation in nonlinear fibers. Equation (2.1) is known to support solitons
or soliton solutions for various kinds of nonlinearity that will be discussed in
the upcoming chapters. The term soliton refers to a nonlinear wave that prop-
agates without changing properties and is stable against mutual collisions
with other solitons that retain their identities. Solitons have been studied ex-
tensively in various areas of mathematical physics. In the context of optical
fibers, solitons are not only of fundamental interest but also have potential
applications in the field of optical fiber communications. Figure 2.1 shows an
illustration of a soliton. This text is devoted to the study of the propagation
of such solitons through optical fibers with emphasis on the various kinds of
the function F (s) in equation (2.1).
2.1.1 Nonlinearity Classification

There are various kinds of nonlinearities of the function F in (2.1) that are
known so far. They are as follows:
1. Kerr law: F (s) = s

This is also known as cubic nonlinearity and is the simplest known
form of the law of nonlinearity. In this case, the NLSE is integrable by
a method called the Inverse Scattering Transform. This method will
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
The Nonlinear Schrödinger’s Equation 9
be discussed in the next chapter. Most optical fibers that are com-
mercially available nowadays obey this Kerr law of nonlinearity.
2. Power law: F (s) = s p
In this case, it is necessary to have 0 < p < 2 to avoid wave collapse.
In fact, it is mandatory that p = 2 to avoid self-focusing singularity.
This law of nonlinearity arises in nonlinear plasmas and solves the
problem of small K -condensation in weak turbulence theory. It also
arises in the context of nonlinear optics. Physically, various mate-
rials, including semiconductors, exhibit power law nonlinearities.
This case of nonlinearity has been studied, including the pertur-
bation term by multiple-scale analysis. The case where p = 12 is
studied in the context of soliton turbulence.
3. Parabolic law: F (s) = s + νs 2
This law is commonly known as the cubic-quintic nonlinearity. The
second term is large for the case of p-toluene sulfonate crystals. This
law arises in the nonlinear interaction between Langmuir waves and
electrons. It describes the nonlinear interaction between the high-
frequency Langmuir waves and the ion-acoustic waves by pon-
dermotive forces. This case of cubic-quintic nonlinearity was also
studied by multiple-scale analysis.
4. Dual-power law: F (s) = s p + ξ s 2 p
This model is used to describe the saturation of the nonlinear refrac-
tive index, and its exact soliton solutions are known. The effective
GNLSE with this dual-power law nonlinearity serves as a basic
model to describe spatial solitons in photovoltaic-photorefractive
materials such as lithium niobate. Optical nonlinearities in many
organic and polymer materials can be modelled using this form of
nonlinearity. The solitons of this model become unstable and decay
in the unstable region 1 ≤ p < 2, while for p ≥ 2, the solitons col-
lapse in a finite time.
λs
5. Saturating law: F (s) = 1+λs
This law with λ > 0 accurately describes the variation of the di-
electric constant of gas vapors through which a laser beam propa-
gates [30]. Optical nonlinearity saturates at a finite value of optical
intensity in most materials. F (s) in those materials can be mod-
eled using the above form, which is known as the saturating form of
nonlinearity. In semiconductor-doped fibers, the soliton propaga-
tion has been modeled using saturable nonlinearity rather than the
usual Kerr nonlinearity. The main motivation behind such attempts
is the observation of such nonlinearity at not too high intensities
in semiconductor-doped glass and other composite materials. This
case was studied numerically.
6. Exponential law: F (s) = 2λ
1
(1 − e −2λs )
This case of exponential nonlinearity serves as useful model in
homogenous, unmagnetized plasmas and laser-produced plasmas.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
When the phase velocity of slow plasma oscillation is much smaller

than the ion thermal velocity, one can obtain the adiabatic or quasi-
static electron density under the quasi-neutral approximation. Now,
combining the coupling equation that exhibits the slowly varying
complex amplitude with the low frequency plasma motion, one ob-
tains the saturable law of nonlinearity.
7. Log law: F (s) = a ln(b 2 s)
This law arises in various fields of contemporary physics. It allows
closed form exact expressions for stationary Gaussian beams (gaus-
sons) as well as for periodic and quasi-periodic regimes of beam
evolution. The advantage of this model is that the radiation from
the periodic soliton is absent because the linearized problem has a
discrete spectrum only.
8. Higher order polynomial law: F (s) = s + νs 2 + γ s 3
This law is an extension of the parabolic law that is given in #3. The
Hamiltonian energy diagrams are studied in [35]. This law is also
observed in various physical systems [31, 35].
9. Triple-power law: F (s) = s p + νs 2 p + γ s 3 p
This law is an extension of the dual-power law and is a generaliza-
tion of the higher order polynomial law. In this case, the Hamiltonian
energy diagrams are also studied.
10. Threshold law:

n1 : s < I0
F (s) =
n2 : s ≥ I0
A smooth transition of this kind can be modeled as
F (s) = As{1 + α tanh[γ (s 2 − I02 )]}
where for s I0 , F (s) ≈ n1 s, where n1 = A[1 − α tanh2 (γ I02 )],

and for s
I0 , F (s) ≈ n2 s, where n2 = 1 + α. Although examples
of nonlinear optical materials with this law are not yet known, the
bistable solitons have interesting properties that can be useful for
future applications in all-optical logic and switching devices.
Of these ten forms of the function F , the first five laws of nonlinearity will be
studied in this text in the upcoming chapters. Of these first five laws, the first
four forms permit exact soliton solutions; however, for the saturable law, an
exact soliton solution is not known and so this form of nonlinearity will be
studied by means of the variational principle that will be introduced in the
next chapter.
To date, no closed-form soliton solutions are known besides the first four
forms of nonlinearity. However, a considerable amount of research is on go-
ing with higher order polynomial law and triples-power law nonlinearity to
obtain the closed form of optical soliton solution.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
2.2 Traveling Waves

A traveling wave is a solution of the NLSE that represents a wave of perma-
nent form that does not change its shape during propagation and moves with
a constant speed. The wave may be localized or periodic. In order to seek
a traveling wave solution of the NLSE that is given by (2.1), introduce the
ansatz
q (x, t) = Ag[B(x − x̄)]e i(−κ x+ωt+σ0 ) (2.2)
where the function g represents the shape of the soliton described by (2.1) and
depends on the type of nonlinearity in it. Also, in (2.2) A and B respectively
represent the amplitude and width of the soliton, κ is the soliton frequency,
ω is the soliton wave number, σ0 is the center of phase of the soliton, and
x̄ gives the mean position of the soliton so that the velocity of the soliton is
given by
d x̄
v= (2.3)
dt
The soliton width and amplitude are related as B = ( A), where the func-
tional form λ depends on the type of nonlinearity in (2.1). Thus, from (2.2),
one easily gets
q t = −ABvg (τ )e iφ − iω Ag(τ )e iφ (2.4)
where, the phase φ is
φ(x, t) = −κ x + ωt + σo (2.5)
and
τ = B(x − x̄) (2.6)
Also, from (2.2) one can get
q xx = AB 2 g (τ ) − 2iκ ABg (τ ) − κ 2 Ag(τ ) (2.7)
Substituting (2.4) and (2.7) into (2.1) simplifies it to

1 2
−i Bvg + ωg + B g − iκ Bg − 1/2κ 2 g + g F ( A2 g 2 ) = 0 (2.8)
2
From (2.8), equating the real and imaginary parts yields
ABg (κ + v) = 0 (2.9)
and
B 2 g − (κ 2 − 2ω)g + 2g F ( A2 g 2 ) = 0 (2.10)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
so that from (2.9), one obtains
κ = −v (2.11)
Now, multiplying both sides of (2.10) by g , then integrating and choosing the
integration constant to be zero since the wave profile is such that q , q x , and
q xx approach zero as |x| → ∞ gives

B 2 (g ) 2 − (κ 2 − 2ω)g 2 + 2 (g 2 ) F ( A2 g 2 )dg = 0 (2.12)
On separation of variables and integration once more, equation (2.12) leads to

dg
x − vt = 1 (2.13)
[(κ 2 − 2ω)g 2 − 2 (g 2 ) F ( A2 g 2 )] 2
Equation (2.13) can be integrated only if the law of nonlinearity is known. As

discussed in subsequent chapters, this equation is possible to integrate for the
cases of Kerr law, power law, parabolic law, and dual-power law only. In all
other cases of nonlinearity, the closed-form soliton solution is not yet known
and remains to be established. This equation will be integrated completely
in the subsequent chapters and the solitons corresponding to the respective
laws of nonlinearity will be studied in detail.
2.3 Integrals of Motion

An important property of nonlinear evolution equations is that they have
conserved quantities known as integrals of motion. In this section, the con-
served quantities of the NLSE, given by (2.1), will be derived. Conservation
laws are a common feature in mathematical physics, where they describe the
conservation of fundamental physical quantities. Rewriting the NLSE in the
form [137]
∂T ∂X
+ =0 (2.14)
∂t ∂x
represents the conservation law. Here T is known as the density while X is

known as the flux. Neither the density nor the flux involve derivatives with
respect to t. Thus T and X may depend upon x, t, q , q x , q xx , . . ., but not q t .
Now, if both T and Xx are integrable on (−∞, ∞), so that X −→ constant as
|x| −→ ∞, then equation (2.1) can be integrated to yield
∞
d
Td x =0 (2.15)
dt −∞
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
so that
∞
Td x = consta nt (2.16)
−∞
The integral of T, over all x, is therefore called the constant of motion or the
integral of motion. For a dynamical system with a finite number of degrees
of freedom to be integrable, the system needs to have as many conserved
quantities as the degrees of freedom. The first conserved quantity for the
NLSE will now be derived.
Performing the operation (2.1) × q ∗ yields
1
iq ∗ q t + q ∗ q xx + F (|q |2 )|q |2 = 0 (2.17)
2
The complex conjugate of equation (2.17) is
1 ∗
−iq q t∗ + q q xx + F (|q |2 )|q |2 = 0 (2.18)
2
Operating (2.17)–(2.18) gives
1
i(q ∗ q t + q q t∗ ) + (q ∗ q xx − q q xx
∗
)=0 (2.19)
2
which can be rewritten as
1
i(|q |2 ) t + (q ∗ q x − q q x∗ ) x = 0 (2.20)
2
so that the flux is q ∗ q x − q q x∗ . Integrating equation (2.20) with respect to x
yields
∞
d
|q |2 d x = 0 (2.21)
dt −∞
since, for a soliton, q , q x , q xx , . . . approach zero as |x| −→ ∞, as previously

mentioned. Thus, the first conserved quantity for the NLSE is given by
∞
E= |q |2 d x = constant (2.22)
−∞
This conserved quantity is known as the wave energy or the mass, wave ac-
tion, or plasmon number; in optics, however, it is called the wave power, while
mathematically, it is known as the L 2 norm.
Now the second conserved quantity will be derived. The complex conju-
gate of (2.1) is given by
1 ∗
−iq t∗ + q xx + F (|q |2 )q ∗ = 0 (2.23)
2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Performing the operation q x∗ × (2.1) + q x × (2.23) gives

1
i(q t q x∗ − q x q t∗ ) + (q x∗ q xx + q x q xx
∗
) + (q q x∗ + q ∗ q x ) F (|q |2 ) = 0 (2.24)
2
Now performing the operation q ∗ × ∂∂x (2.1) + q × ∂∂x (2.23) gives
1
i(q ∗ q xt − q q xt
∗
) + (q ∗ q xxx + q q xxx
∗
)
2

∗ ∂ ∂ 2 ∗
+ q {F (|q | )q } + q {F (|q | )q } = 0
2
(2.25)
∂x ∂x
Then (2.24)–(2.25) gives

∂ 1 ∂ 1
i (q q x∗ − q ∗ q x ) + (q x q x∗ ) − (q ∗ q xxx + q q xxx
∗
)
∂t 2 ∂x 2
+ (|q |2 ) x F (|q |2 ) − [(|q |2 ) x F (|q |2 ) + 2|q |2 F (|q |2 )] = 0 (2.26)
Equation (2.26) can be rewritten as

∂ 1 ∂
i (q q x∗ − q ∗ q x ) + (q x q x∗ )
∂t 2 ∂x
1 ∂ ∗ ∗ 1 ∂
− (q q xx + q q xx )+ (q x q x∗ ) − 2|q |2 F (|q |2 ) = 0 (2.27)
2 ∂x 2 ∂x
which simplifies to
∂ 1 ∂
i (q q x∗ − q ∗ q x ) + [2|q x |2 − (q ∗ q xx + q q xx
∗
) − 4F (|q |2 )] = 0 (2.28)
∂t 2 ∂x
Integrating (2.28) with respect to x gives
∞
d
i (q ∗ q x − q q x∗ )d x = 0 (2.29)
dt −∞
so that
∞
M=i (q ∗ q x − q q x∗ )d x = consta nt (2.30)
−∞
which is the second conserved quantity, also known as the linear momentum.
Equation (2.1) also has a third conserved quantity that is given by [86, 90]
∞
1
H= |q x | − f ( I ) d x
2
(2.31)
−∞ 2
where
I
f (I) = F (ξ )dξ (2.32)
0
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
with the intensity I as I = |q |2 . This integral of motion is known as the

Hamiltonian. The derivation of this conserved quantity is left to the reader as
an exercise. Note that, for the energy conservation law, there are no contribu-
tions from the nonlinear and the dispersion terms of the NLSE. Also, for the
conservation of linear momentum, there are no contributions from the evolu-
tion term and the nonlinear term. Finally, for the Hamiltonian conservation,
there is no contribution from the evolution term.
The Hamiltonian is one of the most fundamental notions in mechanics and,
more generally, in the theory of conservative dynamical systems with finite or
even infinite degrees of freedom. The Hamiltonian formalism has turned out
to be one of the most universal in the theory of integrable system and nonlinear
waves in general. In the case of nonintegrable systems, the Hamiltonian exists
whenever the system is conservative and it is useful for stability analysis.
The most useful approach in the soliton theory of conservative nonintegrable
Hamiltonian system is a representation on the plane of conserved quantities,
namely the Hamiltonian-versus-energy diagrams [35]. In a two-parameter
family of solitons, the Hamiltonian-momentum-energy diagrams are quite
informative.
The Hamiltonian-versus-energy diagrams have been effectively used to
study the various properties of solitons, namely their range of existence, sta-
bility, and general dynamics. These properties were studied in the context
of scalar non–Kerr law solitons, vector solitons in birefringent waveguides,
radiation phenomena from unstable soliton branches, optical couplers, the
theory of Bose-Einstein condensates, and many more. These Hamiltonian-
versus-energy curves are also useful for analyzing the stability of bound
states, whose definition is introduced in chapter 3.
One can see that (2.1) can now be written in canonical forms
δH
iq t = (2.33)
δq ∗
δH
iq t∗ = − (2.34)
δq
where in (2.33) and (2.34) the right sides denote the Fréchet derivative δ F /δq
that is defined as
∞ ∞
δF
v d x = lim F (q + v)d x (2.35)
−∞ δq →0 −∞
for all continuous v. Thus (2.33) and (2.34) define a Hamiltonian dynamical
system on an infinite dimensional phase space. This system can be analyzed
using the theory of Hamiltonian systems that the behavior of the solution is
defined, to a large extent, by the singular points of the system, namely the
stationary solutions of (2.1), and depends on the nature of these points as
determined by the stability of its stationary solutions.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
The soliton solution of (2.1), although not integrable, is assumed to be

given in the form, as introduced in the traveling wave ansatz in (2.2)
q (x, t) = A(t)g[B(t){x − x̄(t)}]e iφ(x,t) (2.36)
where φ(x, t) is the phase of the soliton that is given by
φ(x, t) = −κ x + ωt + σ0 (2.37)
so that
∂φ
= −κ = v (2.38)
∂x
and
∞
∂φ B 2 I0,0,2,0 κ2 1
=ω= − + g 2 (s) F ( A2 g 2 (s))ds (2.39)
∂t 2 I0,2,0,0 2 I0,2,0,0 −∞
where
∞ γ ν
α β dg d2g
Iα,β,γ ,ν = τ g (τ ) dτ (2.40)
−∞ dτ dτ 2
for nonnegative integers α, β, γ , and ν with τ = B(t)(x − x̄(t)). Also, the

velocity (v) of the soliton is defined in (2.3). Here, (2.39) is obtained by dif-
ferentiating (2.36) with respect to t and subtracting from its conjugate while
using (2.1). For such a general form of the soliton given by (2.36), the integrals
of motion, from (2.22), (2.30), and (2.31) respectively, reduce to
∞
A2
E= |q |2 d x = I0,2,0,0 (2.41)
−∞ B
∞
i A2
M= (q q x∗ − q ∗ q x )d x = −κ I0,2,0,0 (2.42)
2 −∞ B
∞
1
H= |q x | − f (|q | ) d x
2 2
−∞ 2
∞
A2 B κ 2 A2 I
= I0,0,2,0 + I0,2,0,0 − F (s)dsd x (2.43)
2 2B −∞ 0
The exact forms of these conserved quantities will be derived in the

following chapters for the various forms of nonlinearity that are going to be
considered.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
2.4 Parameter Evolution

The parameters for the soliton given by (2.36) are now defined as follows
[86, 185, 191]
∞
∂φ i −∞ (q q x∗ − q ∗ q x )d x
κ(t) = − = ∞ (2.44)
−∞ |q | d x
∂x 2 2
∞
x|q |2 d x 1 ∞
x̄(t) = −∞
∞ = x|q |2 d x (2.45)
−∞ |q |2 dx E −∞
where E in (2.45) represents the energy given by (2.41). To obtain the evolution
of the soliton parameters, differentiate κ(t) and x̄(t) given by (2.44) and (2.45)
respectively with respect to t, keeping in mind that E is a constant of motion.
This leads to the following evolution equations for the soliton frequency
dκ
=0 (2.46)
dt
d x̄
= −κ (2.47)
dt
The velocity of the soliton is obtained. Again, the fact that E is a conserved
quantity and B(t) = λ( A(t)) are true leads to the result
dA dB
= =0 (2.48)
dt dt
from (2.41) so that the soliton amplitude and width stay constant during
propagation.
2.4.1 Perturbation Terms

In this book, the NLSE is going to be considered along with its perturbation
terms. Perturbation terms do arise in the context of optics and cannot be
avoided, as will be seen. The NLSE, along with its perturbation terms, is
given by
1
iq t + q xx + F (|q |2 )q = i R[q , q ∗ ] (2.49)
2
where R is a spatio-differential operator while the perturbation parameter
with 0 < 1 is called the relative width of the spectrum that arises due to
quasi-monochromaticity [145] in nonlinear fiber optics, namely [185]
ω0 ω − ω0
= = (2.50)
ω0 ω0
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
In (2.50), ω0 is the carrier frequency of the light wave, while ω0 represents
the departure from the carrier frequency. By quasi-monochromaticity, one
means that the spectrum centered at ω0 has a spectral width ω such that
ω/ω0 1. Since ω0 ∼ 1015 s −1 , quasi-monochromaticity is valid for pulses
whose widths are ≥ 0.1 ps (ω ≤ 1013 s −1 ).
In the presence of perturbations, the conserved quantities no longer exist.
Instead, they are modified integrals of motion. Now differentiate E(t), M(t), and
H(t) that are respectively given by equations (2.41), (2.42), and (2.43) using
(2.49). The following adiabatic evolution equations for the integrals of motion
are obtained
∞
dE
= (q ∗ R + q R∗ )d x (2.51)
dt −∞
∞
dM
= i (q x∗ R − q x R∗ )d x (2.52)
dt −∞
∞
dH 1 ∗ 1
= 2 q xx + F (|q |2 )q ∗ R + q xx + F (|q |2 )q R∗ d x (2.53)
dt −∞ 2 2
Equations (2.51)–(2.53) are known as the modified integrals of motion. In

(2.51)–(2.53), setting = 0 on the right side–the integrals of motion recovers.
Now, using (2.51), (2.52), and (2.42) one can derive that
∞ ∞
dκ B ∗ ∗ ∗ ∗
= i (q x R − q x R )d x − κ (q R + q R )d x (2.54)
dt I0,2,0,0 A2 −∞ −∞
Again, differentiating (2.45) with respect to t and using (2.49) yields

∞
d x̄ B
v= = −κ + x(q ∗ R + q R∗ )d x (2.55)
dt I0,2,0,0 A2 −∞
Finally, differentiating (2.49) with respect to t and subtracting from its conju-
gate, the following equation for the soliton wave number is obtained in the
presence of perturbation terms
∞
∂φ B 2 I0,0,2,0 κ2 1
= − + g 2 (s) F ( A2 g 2 (s))ds
∂t 2 I0,2,0,0 2 I0,2,0,0 −∞
∞
i B
+ (q R∗ − q ∗ R)d x (2.56)
I0,2,0,0 2A2 −∞
Now, equations (2.51), (2.52), and (2.54)–(2.56) can also be rewritten in the
following alternative forms
∞
dE
= 2 g(τ )[Re −iφ ]dτ (2.57)
dt −∞
∞
dM dg
= − κg(τ )[Re −iφ ] + 2A [Re −iφ ]dτ (2.58)
dt −∞ dτ
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
∞
dκ 2 κ
=− g(τ )[Re −iφ ]dτ
dt I0,2,0,0 A −∞
∞
2 1 dg
− {κg(τ )[Re −iφ ] − B [Re −iφ ]}dτ (2.59)
I0,2,0,0 A −∞ dτ
∞
d x̄ 2 1
= −κ + xg(τ )[Re −iφ ]dτ (2.60)
dt I0,2,0,0 A −∞
∞
∂φ B 2 I0,0,2,0 κ2 1
= − − g 2 (s) F ( A2 g 2 (s)) ds
∂t 2 I0,2,0,0 2 I0,2,0,0 −∞

1 ∞
+ g(τ )[Re −iφ ]dτ (2.61)
I0,2,0,0 A −∞
where, once again, τ = B(t)(x − x̄), while φ = −κ x + ωt + σ0 . These alterna-

tive forms of the adiabatic evolution equations for the soliton parameters are
known as the amplitude-phase format. Equations (2.57)–(2.61) give the adiabatic
dynamics of the soliton parameters in the presence of perturbation terms.
In this book, a particular form of the perturbation term R will be studied,
for various kinds of nonlinearities, that is given by [86]
x
R = δ|q |2m q + σ q |q |2 ds (2.62)
−∞
In (2.62), δ is the coefficient of nonlinear damping or amplification, depending

on its sign, and m could be 0, 1, or 2. For m = 0, δ is the linear amplification or
attenuation according to δ being positive or negative. For m = 1, δ represents
the two-photon absorption (or a nonlinear gain if δ > 0). If m = 2, δ gives
a higher-order correction (saturation or loss) to the nonlinear amplification-
absorption. Also, σ is the coefficient of saturable amplifiers that is introduced
to compensate for the losses in optical fibers. A model with a saturation term
included is more satisfactory from a physical point of view, since then the
stable soliton propagation is ensured, in principle, over an indefinite propa-
gation distance, including transoceanic distances.
Several factors contribute to nonlinear damping or attenuation. One is the
material absorption. Rayleigh scattering is a fundamental loss mechanism
that rises from random density fluctuations frozen into the fused silica in the
manufacturing process and is a dominating factor. Pure silica absorbs either
in the ultraviolet region or in the far-infrared region beyond 2 µm. However,
even a relatively small amount of impurities can lead to significant absorption
in the wavelength window 0.5–2 µm. One of the most important impurity
factors affecting fiber loss is the hydroxide (OH) ion. Special precautions are
taken during the fiber fabrication process to ensure that the OH impurity level
is less than one part per 100 million.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
So, the perturbed NLSE that is going to be considered is given by

x
1
iq t + q xx + F (|q | )q = i δ|q | q + σ q
2 2m
|q | ds
2
(2.63)
2 −∞
To obtain the parameter dynamics of the solitons caused by the particular

form of perturbations given in (2.63), proceed as follows. Substitute R given
by (2.63) into (2.51)–(2.56) and carry out the integration using the notation
that was introduced in (2.40). The following equations are obtained [86]
∞ τ
dE 2
= 2 δ A2m+2 B I0,2m+2,0,0 + σ g 2 (τ ) g 2 (s)ds dτ (2.64)
dt B −∞ −∞
∞ τ
dM 2 κ
= 2 δA 2m+2
B I0,2m+2,0,0 + σ 2
g (τ ) 2
g (s)ds dτ (2.65)
dt B −∞ −∞
dκ
=0 (2.66)
dt
∞ τ
d x̄ 2 σ A2
v= = −κ + τ g (τ )
2 2
g (s)ds dτ (2.67)
dt I0,2,0,0 B 2 −∞ −∞
∞
∂φ B 2 I0,0,2,0 κ2 1
= − + g 2 (s) F ( A2 g 2 (s))ds (2.68)
∂t 2 I0,2,0,0 2 I0,2,0,0 −∞
The integrals in (2.64)–(2.68) will be explicitly calculated in the upcoming

chapters for various laws of nonlinearity where the functions F (ζ ) and g(s)
are exactly known. This will give the laws of adiabatic dynamics of the solitons
in the presence of perturbation terms for various laws of nonlinearity.
2.5 Quasi-Stationary Solution

The idea of quasi-stationarity for solving the nonlinear evolution equations
was first introduced in 1981 by Kodama and Ablowitz. Later, this idea was
extended to study the NLSE, for Kerr law nonlinearity, with Hamiltonian
and non-Hamiltonian type perturbations. Subsequently, it was also used to
study the perturbed NLSE with various kinds of non-Kerr law nonlinearities.
Studying the dynamics of solitons in the presence of perturbation terms gives
more information than was obtained in the previous section by virtue of the
soliton perturbation theory. For example, in order to integrate the perturbed
NLSE up to O( ), one needs to study the perturbed solitons by means of
quasi-stationarity since the soliton perturbation theory fails to integrate the
perturbed NLSE. The basic idea of a quasi-stationary (QS) method is explained
in the following subsection [5, 228].
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
2.5.1 Mathematical Theory∗

In a general setting, the solution of a perturbed nonlinear dispersive wave
equation that is studied is of the type
K (q , q t , q x , · · ·) = F (q , q x , · · ·) (2.69)
Here, K and F are nonlinear functions of q , q x , · · ·, while 0 < 1. The

unperturbed equation (for = 0)
K (q (0) , q t(0) , q x(0) , · · ·) = 0 (2.70)
has a solution q (0) that is taken as a solitary wave or a soliton solution. This
solution in terms of certain natural fast and slow variables is written as
q (0) = q̂ (0) (θ1 , θ2 , · · · θm , T, X : P1 , P2 , · · · , PN ) (2.71)
where, θi for 1 ≤ i ≤ m are the so-called fast variables while T = t and X = x

are the slow variables, and Pl for 1 ≤ l ≤ N is the parameter that depends on the
slow variables. In many problems, only one fast variable, namely θ = x− P1 t in
the unperturbed problem, is needed. One can generalize θ to satisfy ∂θ/∂x = 1
and ∂θ/∂t = −P1 and can use P1 = P1 ( X, T) to remove the secular terms. With
this, such a solution, given by (2.71), is a quasi-stationary solution and one can
write q = q̂ (θ, X, T, ). It is necessary to develop equations for the parameters
P1 , · · · , PN by using appropiate conditions, such as the secularity conditions.
There must be N such conditions. Some of these conditions are formed from
Green’s identity, as follows. Assuming an expression for q̂ of the form
q̂ = q̂ (0) + q̂ (1) + · · · (2.72)
(after introducing the appropiate variables θi , X, T, etc.). Then (2.70) is the

leading-order problem and, assuming that K has only first order in time
derivatives
∂K
L( ∂θi , q̂ (0) ) q̂ (1) = F ( q̂ (0) ) − · qT = F (2.73)
∂q t
is the first-order equation. Here, L( ∂θi , q̂ (0) ), u = 0 is the linearized equation

of K (q , q t , q x , · · ·) = 0 after (x, t) is transformed to the appropiate coordinate
θi . Calling vi the M solutions of the homogenous adjoint problem satisfying
the necessary boundary conditions (e.g., vi → 0, as |θi | → ∞)
L Avi = 0 (2.74)
∗ Portions of Section 2.5.1 are reprinted with permission from Ablowitz, M.J. & H. Segur, Soli-
tons and the Inverse Scattering Transform. 1981. Copyright

c Society for Industrial and Applied
Mathematics.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
for 1 ≤ i ≤ M and M ≤ N with L A being the adjoint operator to L, one forms
L q̂ (1) · vi − (L Avi ) · q̂ (1) = F̂ vi (2.75)
The left side of (2.73) is always a divergence (Green’s theorem). It may be

integrated to give the secularity conditions, also known as the Fredholm’s
alternative (FA). These secularity conditions enable us to compute a solution
q̂ (1) to (2.73) that satisfies suitable boundary conditions (e.g., q̂ (1) is bounded
as |θ| → ∞). However, as is standard in perturbation problems, there is still
freedom in the solution. This is due to the fact that some terms in the solu-
tion q̂ (1) can be absorbed in the leading-order solution q̂ (0) by shifting other
parameters. The solution q̂ (1) can be made unique by imposing additional
conditions that reflect specific initial conditions or other normalizations. Con-
tinuation to higher order q̂ ( N) is straightforward.
2.5.2 Application
In this book, a quasi-stationary solution to (2.63) will be obtained using the
method that was discussed in the previous subsection. The main part of this
work is to implement a perturbation scheme to solve (2.63) as follows [86]:
q = q̂ (θ, T, X; )e ρ(T, X; )
i
(2.76)
where
∂θ ∂θ
= 1, =0
∂x ∂t
and
T = t X = x
Here, as mentioned, θ is a fast variable while X and T are slow variables in

space and time, respectively. When the perturbation terms of the NLSE are
turned on, the soliton parameters A and B are then slowly varying functions,
namely A = A( X, T) and B = B( X, T).
Substituting (2.76) into (2.63) and expanding
q̂ = q̂ (0) + q̂ (1) + 2 q̂ (2) + · · ·
ρ = ρ (0) + ρ (1) + 2 ρ (2) + · · ·
v = v(0) + v(1) + 2 v(2) · · ·
gives, at the leading order

1
(0) 2 1 ∂2 q̂ (0)
− ρT(0) + ρX q̂ (0) + + q̂ (0) F [( q̂ (0) ) 2 ] = 0 (2.77)
2 2 ∂θ 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and

∂q̂ (0)
ρ X(0) − v(0) =0 (2.78)
∂θ
from the real and imaginary parts. Now, (2.78) implies
ρ X(0) = v(0) (2.79)
Setting
1
(0) 2 1
h( B 2 ) = ρT(0) + ρ = ρT(0) + (v(0) ) 2 (2.80)
2 X 2
where the function h depends on the nonlinearity F . Thus, (2.77) changes to
1 ∂2 q̂ (0)
−h( B 2 ) q̂ (0) + + q̂ (0) F [( q̂ (0) ) 2 ] = 0 (2.81)
2 ∂θ 2
whose solution is (on comparing with [2.35])
q̂ (0) = Ag[B(θ − θ̄)] (2.82)
where
d θ̄
=v (2.83)
dt
At O( ), decompose q̂ (1) = φ̂ (1) + i ψ̂ (1) into its real and imaginary parts. Now,
the equations for φ̂ (1) and ψ̂ (1) , by virtue of (2.77), are respectively
1 ∂2 φ̂ (1)
−h( B 2 ) φ̂ (1) + + 2( q̂ (0) ) 2 φ̂ (1) F [( q̂ (0) ) 2 ] + φ̂ (1) F [( q̂ (0) ) 2 ]
2 ∂θ 2
∂2 q̂ (0)
= ρT(1) + v(0) ρ X(1) q̂ (0) − (2.84)
∂θ ∂ X
and
1 ∂2 ψ̂ (1)
−h( B 2 ) ψ̂ (1) + + F [( q̂ (0) ) 2 ]ψ̂ (1)
2 ∂θ 2
∂q̂ (0) ∂q̂ (0) ∂q̂ (0)
=− − v(0) − ρ X(1) − v(1) + σ
∂T ∂X ∂θ
x
(0) (0)
− ρ XX q̂ − δ( q̂ (0) ) 2m+1 + σ q̂ (0) ( q̂ (0) ) 2 ds (2.85)
−∞
Equations (2.84) and (2.85) contain secular terms that are also known as reso-
nances. In order to eliminate these secular terms, to avoid unbounded growth,
one needs to use the FA, which states that the right side of the ODEs in (2.84)
and (2.85) should be orthogonal to the null space of the adjoint operator of the
left side. But for the ODEs given by (2.84) and (2.85), the left side is self-adjoint.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Thus, FA applied to (2.84) yields

∞
(1) (0) (1)
(0) ∂2 q̂ (0) ∂q̂ (0)
ρT + v ρ X q̂ − dθ = 0 (2.86)
−∞ ∂θ ∂ X ∂θ
which leads to
∂A ∂B ∂B
B I0,0,2,0 + A I0,0,2,0 + A I1,0,0,1 = 0 (2.87)
∂X ∂X ∂X
Similarly, with the second homogenous solution of (2.73), one recovers
ρT(1) + v(0) ρ X(1) = 0 (2.88)
whereas, FA applied to (2.85) gives

∞ (0)
∂q̂ ∂q̂ (0) ∂q̂ (0)
+ v(0) + ρ X(1) − v(1) + σ
−∞ ∂T ∂X ∂θ
x
(0) (0)
+ ρ XX q̂ + δ( q̂ (0) ) 2m+1 − σ q̂ (0) ( q̂ (0) ) 2 ds q̂ (0) dθ = 0 (2.89)
−∞
which yields
∂A ∂B ∂A ∂B
B I0,2,0,0 + A I1,1,1,0 + v(0) B I0,0,2,0 + v(0) A I1,1,1,0
∂T ∂T ∂X ∂X
x
(0)
= δ A2m+1 B I0,2m+2,0,0 − ρ XX AB I0,0,2,0 + σ q̂ (0) ( q̂ (0) ) 2 ds (2.90)
−∞
Similarly, with the second homogenous solution of (2.85), one gets
ρ X(1) = v(1) − σ (2.91)
Since A(t) and B(t) are related, depending on the functional form of F (s),
(2.87) leads to the conclusion
∂A ∂B
= =0 (2.92)
∂X ∂X
so that A and B are functions of T only.
Also, in an ideal soliton-based, fiber-optic communication system, input
pulses launched into the fiber should be unchirped in order to avoid shedding
part of the pulse energy as a dispersive tail during the process of soliton
(0)
formation [22]. So, in (2.90), using (2.92) and setting ρ XX = 0 to eliminate
frequency chirp gives
x
dA dB
B I0,2,0,0 + A I1,1,1,0 = δ A2m+1 B I0,2m+2,0,0 + σ q̂ (0) ( q̂ (0) ) 2 ds
dT dT −∞
(2.93)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
It is to be noted here that (2.93) can be recovered by the usual soliton perturba-
tion theory (SPT), provided the type of nonlinearity in F is known. However,
the relations (2.87), (2.88), and (2.91) cannot be obtained by the SPT. This is
where the SPT fails. Thus, a more general approach to study optical soliton
perturbation is the QS method. Now (2.84), by virtue of (2.87) and (2.90),
reduces to
1 ∂2 φ̂ (1)
−h( B 2 ) φ̂ (1) + + 2( q̂ (0) ) 2 φ̂ (1) F [( q̂ (0) ) 2 ] + φ̂ (1) F [( q̂ (0) ) 2 ] = 0 (2.94)
2 ∂θ 2
while (2.85), by virtue of (2.87) and (2.91), gives
1 ∂2 ψ̂ (1)
−h( B 2 ) ψ̂ (1) + + F [( q̂ (0) ) 2 ]ψ̂ (1)
2 ∂θ 2
x
∂q̂ (0)
=− + δ( q̂ )
(0) 2m+1
+ σ q̂ (0)
( q̂ (0) ) 2 ds (2.95)
∂T −∞
The solutions to (2.94) and (2.95) are respectively
φ̂ (1) = 0 (2.96)
and
τ
2A ∂θ̄ 1 s2

ψ̂ (1)
= g(τ ) g(s1 )g (s1 )ds1 ds2
B ∂T 2
g (s2 )
τ s2
2 dA 1
− g(τ ) g 2 (s1 )ds1 ds2
B 2 dT 2
g (s2 )
τ s2
2A d B 1
− g(τ ) s1 g(s1 )g (s1 )ds1 ds2
B 3 dT g 2 (s2 )
τ s2
A2m+1 1
+ 2δ g(τ ) g 2m+2
(s1 )ds1 ds2
B2 g 2 (s2 )
τ s3 s2
A3 1
+ 2σ 2 g(τ ) g 2
(s 2 ) g 2
(s1 )ds 1 ds 2 ds3 (2.97)
B g 2 (s3 )
The QS solution of (2.63) finally is
q ≈ Pe iψ (2.98)
where
P = q̂ (0) (2.99)
and
1
ψ = Q(θ) + ρ( X, T) (2.100)

P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
with
Q(θ) = ψ̂ (1) /q̂ (0) (2.101)
Equation (2.98) is, thus, the “formal” solution to the perturbed NLSE that is
given by (2.63). Note that this solution is a general one in the sense that the
law of nonlinearity given by F (s) is not yet known. However, in the follow-
ing chapters, the special cases of this law of nonlinearity will be considered.
Once the functional form of F (s) is known, it will be possible to evaluate the
integrals seen in (2.97).
Exercises
1. Prove that the Hamiltonian given by (2.31) is a conserved quantity.

2. Derive the wave number for the soliton that is given by (2.39).
3. Use the laws of adiabatic deformation of energy and linear momen-
tum of the soliton that are given by (2.51) and (2.52), respectively, to
derive the adiabatic dynamics of the soliton frequency that is given
by (2.54).
4. Use the expression for the velocity of the soliton that is given by
(2.55) to derive the expression in (2.67) corresponding to the pertur-
bation terms given by (2.62).
5. Use the adiabatic dynamics for the soliton wave number given by
(2.56) to derive the expression in (2.68) corresponding to the pertur-
bation terms that are given in (2.62).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
3
Kerr Law Nonlinearity
This chapter talks about the detailed aspects of optical solitons that are
governed by the nonlinear Schrödinger’s equation (NLSE) with Kerr law
nonlinearity, which is also commonly known as the cubic Schrödinger’s
equation. Section 3.1 talks about the physics of the origin of Kerr law nonlin-
earity. Section 3.2 contains an introductory discussion about the technique of
inverse scattering transform (IST) that is used to integrate the NLSE with Kerr
law nonlinearity. Moreover, the comparison between the IST and the Fourier
transform (FT) is discussed here. Also, the 1-soliton solution is given. The
infinitely many conserved quantities of the NLSE are talked about in
Section 3.3 along with the Hamiltonian structure, while Section 3.4 leads
to a discussion about another technique of solving the NLSE, namely the
variational principle using the Lagrangian. Section 3.5 is about the quasi-
stationarity applied to the case of Kerr law nonlinearity, and Section 3.6 intro-
duces the Lie transform technique that can be used to integrate the perturbed
NLSE, for Kerr law, that contains Hamiltonian perturbation terms only.
3.1 Introduction
In this chapter, equation (2.1) will be studied for a special case, namely when
F (s) = s, which is known as the Kerr law of nonlinearity. In this case, (2.1) is
integrable by the method of the IST that was first applied to solve (2.1) by
Zakharov and Shabat [296] in 1972. The IST is the nonlinear analog of the
Fourier transform that is used for solving linear partial differential equations.
The IST will be discussed in the next section of this chapter.
The Kerr law of nonlinearity originates from the fact that a light wave
in an optical fiber faces nonlinear responses from nonharmonic motion of
electrons bound in molecules, caused by an external electric field. Even though
the nonlinear responses are extremely weak, their effects appear in various
ways over long distances of propagation that are measured in terms of light
wavelength.
27
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
The origin of nonlinear response is related to the nonharmonic motion

of bound electrons under the influence of an applied field. As a result, the
induced polarization (P) is not linear in the electric field (E) but involves
higher order terms in the electric field amplitude as [185, 192, 311]
∞
P = 0 χ (1) (t − τ )E(τ )dτ
−∞
∞ ∞
+ χ (2) (t − τ1 )χ (2) (t − τ2 )E(τ1 )E(τ2 )dτ1 dτ2
−∞ −∞
∞ ∞ ∞
+ χ (3) (t − τ1 )χ (3) (t − τ2 )χ (3) (t − τ3 )E(τ1 )E(τ2 )E(τ3 )
−∞ −∞ −∞

dτ1 dτ2 dτ3 + · · · (3.1)
where 0 is the permittivity of the vacuum and χ ( j) , for j = 1, 2, . . . is the jth

order susceptibility tensor of rank j + 1, which accounts for light polarization
effects. The linear susceptibility, χ (1) , gives the linear refractive index n0 and
the attenuation coefficient. The second order susceptibility, χ (2) , is responsible
for the second-harmonic generation. However, it is generally negligible for sil-
icon dioxide, which has an inversion symmetry at the molecular level, unless
χ (2) contains some resonant effects. As a result, optical fibers do not normally
exhibit second-order nonlinear effects, although the electric-quadrupole and
magnetic-dipole moments can generate weak second-order nonlinear effects.
The lowest nonlinear effects in nonlinear fibers originate primarily from the
third order susceptibility, χ (3) . A light wave with frequency ω sees nonlinear
response of the χ (3) term through the interaction of ω, −ω, and ω compo-
nents. The response contributes to the nonlinear modification of the index of
refraction [22, 192, 227]
n(ω, |E|2 ) = n0 (ω) + n2 (ω)|E|2 (3.2)
where E is the amplitude of the wave electric field and n2 is related to χ (3)
through
3 (3)
n2 (ω) = χ (3.3)
8n0 xxxx
for a linearly polarized wave in the x direction. Here n0 is the linear refrac-
tive index of the medium and the nonlinear index of refraction n2 is called
the Kerr coefficient and has a value of approximately 10−22 m2 /W for silica
fibers. Even though the fiber nonlinearity is small, the nonlinear effects ac-
cumulate over long distances and can have a significant impact due to the
high intensity of the light wave over a small fiber cross-section. Since the
Kerr effect here originates from the nonharmonic motion of electrons bound
in molecules, the instantaneous response (3.3) is satisfied. In a fiber, the elec-
tric field magnitude varies in its cross-section, thus a proper averaging of |E|2
should be taken into account when evaluating the response. By itself, the Kerr
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Kerr Law Nonlinearity 29
nonlinearity produces an intensity-dependent phase shift that results in

spectral broadening during propagation. For the Kerr law of nonlinearity
in equation (2.1), F (s) = s and so f (s) = s 2 /2. In the following subsection,
the derivation of the NLSE with Kerr law of nonlinearity is given.
3.1.1 The Nonlinear Schr ödinger’s Equation

In this subsection, the derivation of the dynamical equation of an optical pulse
propagating through an optical fiber is given for the case of the Kerr law of
r , t) of the pulse is governed by the wave
nonlinearity. The electric field E(
equation
2 2
×∇
∇ r , t) + µ0 ∂ P(r , t) + 1 ∂ E(r, t) = 0
× E( (3.4)
∂t 2 c 2 ∂t 2
r , t) is the induced polarization in the media, µ0 is the vacuum
where P(
permeability, and c is the velocity of light. The induced polarization consists
of two parts: the linear part P L (= 0 χ (1) E)
and the nonlinear part PnL ; 0 is
the vacuum permittivity and χ is the linear susceptibility. The optical pulse
(1)
is taken to be of the form
r , t) = 1 x [ E(
E( r , t)e iω0 t + cc] (3.5)
2
1
PL (r , t) = x [ PL (r , t)e iω0 t + cc] (3.6)
2
nL (r , t) = 1 x [ P
P nL (r , t)e iω0 t + cc] (3.7)
2
where ω0 is the carrier frequency of the light wave, x is the polarization unit
r , t), PL (r , t), and P
vector, and E( nL (r , t) are slowly varying functions of time.
In (3.5)–(3.7), cc represents the complex conjugate. For optical fibers, the first
term of (3.4) can be approximated as
×∇
∇ r , t) = ∇(
× E( ∇ r , t)) − ∇ 2 E(
· E( r , t) ≈ −∇ 2 E(
r , t) (3.8)
so that (3.4) reduces to

r , t)
∂2 P( t)
1 ∂2 E(r,
r , t) − µ0
∇ 2 E( − =0 (3.9)
∂t 2 c 2 ∂t 2
At this stage, it is necessary to make one more simplification. It is assumed
that the linear part of the polarization depends on the frequency, while the
nonlinear part is independent of the frequency. In other words, only those
media for which the response of nonlinearity is instantaneous are considered.
The nonlinear polarization for optical fibers, which are oscillating at ω0 , is of
the form
nL (r , t) = 3 0 χ (3) |E|2 E(

P r , t) (3.10)
4
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
r , t) is now defined as
The Fourier transform of E(
∞
Ẽ(r , ω) = r , t)e i(ω−ω0 )t dt
E( (3.11)
−∞
Treating the nonlinear part of susceptibility as a constant, the Fourier trans-

form of (3.9) gives
∇ 2 Ẽ(r, ω − ω0 )
ω2
+ {1 + χ (1) (ω − ω0 )} Ẽ(r, ω − ω0 )
c2

ω2 3 (3) 2
+ 2 χ |E| Ẽ(r, ω − ω0 ) = 0 (3.12)
c 4
The frequency-dependent wave number k is defined through the relation
ω2
k 2 (ω) = {1 + χ (1) (ω − ω0 )} (3.13)
c2
From (3.13), expanding k 2 in a Taylor series about the carrier frequency ω0
gives
k 2 = k02 + 2(ω − ω0 )k0 k0 + (ω − ω0 ) 2 (k0 ) 2 + (ω − ω0 ) 2 k0 k0 (3.14)
where the prime in (3.14) denotes a derivative with respect to ω and evaluated
at ω0 and
ω0
k0 = 1 + χ (1) (ω0 ) (3.15)
c
Also k0 is the inverse of group velocity vg , while k0 is the group velocity
dispersion. For a normal dispersion regime k0 > 0, while in the anomalous
dispersion regime k0 < 0. Inserting (3.14) into (3.12) and then transforming
the equation in the time domain gives
∂E ∂2 E ∂2 E 3 ω2
∇ 2 E + k02 E + 2ik0 k0 − (k0 ) 2 − k0 k0 + χ (3) 20 |E|2 E = 0 (3.16)
∂t ∂t 2 ∂t 2 4 c
Choosing the ansatz E(r , t) = A(r , t)e ik0 z , where the scalar envelope A(r, t) of
the optical pulse varies slowly over one optical period, one can introduce the
slowly varying envelope approximation as
∂2 A ∂A
< ω0 (3.17)
∂t 2 ∂t
and
∂2 A ∂A
< k0 (3.18)
∂z2 ∂z
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Substituting the ansatz for E(r, t) and employing the slowly varying envelope
approximation gives
∂A
∇⊥2 A + 2ik0
∂z
∂2 A ∂A ∂2 A ∂2 A 3 (3) ω02 2
+ + 2ik0 k0 − k0 k0 − (k0 ) 2 + χ |A| A = 0 (3.19)
∂z2 ∂t ∂t 2 ∂t 2 4 c2
where ∇⊥2 is the perpendicular Laplacian. Now, changing to a frame that is
moving with a group velocity vg of the pulse and introducing the transfor-
mations
z = z (3.20)
and
z
t = t − = t − zk0 (3.21)
vg
gives
∂2 A ∂A ∂2 A 3 (3) ω02 2
∇⊥2 A + + 2ik0 − k0 k0 + χ |A| = 0 (3.22)
∂z2 ∂z ∂t 2 4 c2
In writing the above equation, the primes over the variables are dropped
simply for convenience. The second term is much smaller due to the slowly
varying envelope. Thus (3.22) can be modified to
∂A ∂2 A 3 (3) ω02 2
∇⊥2 A + 2ik0 − k0 k0 + χ |A| A = 0 (3.23)
∂z ∂t 2 4 c2
Now consider only the linear part of equation (3.23), neglecting the fourth
term. It is assumed that
A(r, t) = R(r ) Q(z, t)e −iδz (3.24)
where R(r ) is the transverse mode profile and δ is the corresponding eigen-
value satisfying
∇⊥2 R(r ) + 2δk0 R = 0 (3.25)
For the nonlinear equation (3.23), it is assumed that (3.25) still holds for the
transverse mode structure of the propagating optical pulse. Therefore, insert-
ing (3.24) in (3.23) and employing (3.25), one can finally arrive at
∂Q k0 ∂2 Q ω
i − + n2 |R|2 |Q|2 Q = 0 (3.26)
∂z 2 ∂t 2 c
Averaging this equation over a fiber cross-section gives
∂Q k0 ∂2 Q ω
i − + n2 α|Q|2 Q = 0 (3.27)
∂z 2 ∂t 2 c
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where

|R|4 dr
α= (3.28)
|R|2 dr
The numerical value of α depends on the concrete form of R(r ). For a Gaus-
sian field profile in optical fibers, α = 0.5. In subsequent discussions, only
anomalous media are considered; k0 < 0 and so
t
τ = (3.29)
t0

Q= P0 q (3.30)
t02
Ld = (3.31)
|k0 |
and
|k0 |c
P0 = (3.32)
ω0 n2 αt02
which transforms (3.27) to
∂q 1 ∂2 q
i + + |q |2 q = 0 (3.33)
∂z 2 ∂τ 2
This equation is known as the NLSE and has been widely used for the propa-
gation of solitons through optical fiber with Kerr law of nonlinearity. In order
to maintain the uniformity of the notation in this book, the changes z → t and
τ → x are utilized so that the NLSE can be written in the standard notation as
1
iq t + q xx + |q |2 q = 0 (3.34)
2
3.2 Traveling Wave Solution

For the case of Kerr law nonlinearity, equation (2.1) reduces to
1
iq t + q xx + |q |2 q = 0 (3.35)
2
so that equation (2.10) simplifies to
B 2 g − (κ 2 − 2ω)g + 2A2 g 3 = 0 (3.36)
Multiplying equation (3.36) by g and integrating yields
g 2 (κ 2 − 2ω − A2 g 2 )
(g ) 2 = (3.37)
B2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Separating variables and integrating yields, on choosing the integration con-

stant to be zero

dg
x − vt = (3.38)
g κ − 2ω − A2 g 2
2
On substituting
κ 2 − 2ω
g2 = (3.39)
A2 cosh2 θ
leads to the soliton solution
A
q (x, t) = e i(−κ x+ωt+σ0 ) (3.40)
cosh[B(x − x̄(t))]
where
κ = −v (3.41)
and
B2 − κ2
ω= (3.42)
2
while
A= B (3.43)
This is the 1-soliton solution of the NLSE with Kerr law nonlinearity that can
be obtained by traveling wave ansatz.
3.3 Inverse Scattering Transform

In this section, the IST, a powerful method of integrating the NLSE with Kerr
law nonlinearity that was used for solving (3.34), will be briefly presented.
Subsequently, the 1-soliton solution of (3.34) will be derived. A detailed study
of IST can be found in various books. Among the best books is the famous
book by Ablowitz and Clarkson [6]. Basically, the IST is schematically similar
to the FT that is used for solving linear differential equations. The FT can be
described schematically as
FT
q (x, 0) −→ q̂ (ω, 0)
time evolution ↓ time evolution ↓
inverse F T
q (x, t) ←− q̂ (ω, t)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
while the IST, schematically, is
sca ttering
q (x, 0) −→
(x, 0)
time evolution ↓ time evolution ↓
I ST
q (x, t) ←−
(x, t)
The main idea of the IST consists in recognizing that equation (3.34) can be
expressed as the compatibility conditions of the linear equations for the wave
function ψ(x, t;ζ ) [384]
L(t)ψ = ζ ψ (3.44)
and
∂ψ
= M(t)ψ (3.45)
∂t
where L and M are the differential operators that formulate the Lax pair of
the NLSE that are given by
⎡ ⎤
i ∂∂x q
L=⎣ ⎦ (3.46)
−q ∗ −i ∂∂x
and
⎡ ⎤
i 2 1 ∂q
⎢ −iζ + 2 |q | −iq ζ −
2
2 ∂x ⎥
M=⎢
⎣
⎥ (3.47)
1 ∂q ∗ i 2 ⎦
−iq ∗ ζ + iζ − |q |
2
2 ∂x 2
Equation (3.44) is an eigenvalue problem with eigenvalue ζ . This equation
can also be rewritten as the x-evolution of the wave function, namely
∂ψ
= Pψ (3.48)
∂x
where

−iζ iq
P= (3.49)
iq ∗ iζ
On the other hand, equations (3.45) determine the time evolution of the wave
function ψ. Also, the eigenvalue ζ is isospectral, namely
dζ
=0 (3.50)
dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
For consistency between (3.45) and (3.48), the compatibility condition
∂2 ψ ∂2 ψ
= (3.51)
∂ x ∂t ∂t ∂x
is equivalent to
∂P ∂M
− + [P, M] = 0 (3.52)
∂t ∂x
where in (3.52) the notations mean
[P, M] = P M − MP (3.53)
is the commutator of P and M. This equation gives the NLSE with Kerr law
nonlinearity. The advantage of this representation is that one can solve the
nonlinear equation (3.34) by virtue of two linear problems. In fact, one asso-
ciates the initial pulse profile q (x, 0), the scattering data (similar to the Fourier
coefficients) that are obtained, as follows, from the solution of equation (3.44).
The wave functions are defined as

ψ1
= (3.54)
ψ2
and

φ1
= (3.55)
φ2
with the asymptotic values for (ζ ) = ξ

1
(x;ξ ) → e −iξ x a s x → −∞ (3.56)
0

0
(x;ξ ) → e iξ x a s x → ∞ (3.57)
1
where (ζ ) represents the real part of ζ . It can be shown that (x;ξ ) and
(x;ξ ) can be analytically extended to the upper half plane of ξ . The pair of
solutions , ˜ forms a complete system of solutions to (3.44) where

ψ2∗
=
˜ (3.58)
−ψ1∗
Thus, any other solution can be expressed as its linear combination, namely
(x;ξ ) = a (ξ )
˜ + b(ξ ) (x;ξ ) (3.59)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
with the limit as x → ∞, one obtains from this equation

1 −iξ x 0
(x;ξ ) = a (ξ ) → e + b(ξ ) → e iξ x (3.60)
0 1
From this couple of equations, one can see that 1/a (ξ ) and b(ξ )/a (ξ )
respectively represent the transmission and reflection coefficient. Also, from
equation (3.60), one can write
a (ξ ) = W(φ, ψ)(ξ ) (3.61)
and
a (ξ ) = W(φ, ψ ∗ )(ξ ) (3.62)
where the x-dependence of the Wronskian W( f, g) = f 1 g2 − f 2 g1 is used. It

is to be noted that a (ζ ) can be analytically extended to the upper half plane
of ζ and also asymptotically a (ζ ) → 1 as |ζ | → ∞.
The discrete eigenvalues of (3.44) corresponding to the points ζ = ζn for
n = 1 . . . N; (ζ ) > 0, where a (ζ ) = 0. Then the eigenfunctions obey
(x;ζn ) = b n (x;ζn ) (3.63)
This equation shows that both and approach zero as in the limit as
|x| → ∞. In other words, the discrete eigenvalues correspond to the bound
states of the scattering problem (3.44) or solitons.
For any given potential function q (x, 0), one can then compute the solution
of the direct scattering problem (3.45) with the scattering data

(t = 0) = [r (ξ, 0), (ξ ) = 0, ζn , Cn (0), n = 1 . . . N] (3.64)
where
b(ξ, 0)
r (ξ, 0) = (3.65)
a (ξ, 0)
is the reflection coefficient, while
b n (0)
Cn (0) = (3.66)
a n (0)
is the normalization constant for the nth bound state, and

∂a
a n (0) = (ζn , 0) (3.67)
∂ζ
Since there is a one-to-one correspondence between the scattering data and

the potential functions, one can calculate the time evolution of
and then
invert the transform to obtain the field q (x, t). From (3.46) the time evolution
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
of the scattering is given by
r (ξ, t) = r (ξ, 0)e −2iξ

2
t
(3.68)
and
ζn (t) = ζn (0)e −2iζn t

2
(3.69)
Finally, the field q (x, t) is obtained. First, divide equation (3.60) by a (ξ )(ξ − ζ )
and then take its FT with respect to ξ . Using the property a (ζ ) → 1 as ζ → ∞
for (ζ ) ≥ 0, one obtains the following set of linear integral equations, also
known as the Gelfand-Levitan-Marchenko (GLM) integral equations, for the
eigenfunctions and the field at time t.
∞
1 r ∗ (ξ, t)ψ2∗ (x, t, ξ ) −iξ x
ψ1 (x, t, ζ )e −iζ x = e dξ
2πi −∞ ξ −ζ
N ∗
Cn∗ (t)ψn2 (x, t) −iζn∗ x
+ ∗
e (3.70)
n=1
ζn − ζ
∞ ∗
1 r (ξ , t)ψ1∗ (x, t, ξ ) −iξ x
ψ2 (x, t, ζ )e −iζ x = 1− e dξ
2πi −∞ ξ −ζ
N ∗
Cn∗ (t)ψn1 (x, t) −iζn∗ x
− ∗
e (3.71)
n=1
ζn − ζ
∞ ∗
i r (ξ, t)ψ2∗ (x, t, ξ ) −iξ x
q (x, t) = e dξ
π −∞ ξ −ζ

N
∗
−2 Cn∗ (t)ψn2
∗
(x, t)e −iζn x (3.72)
n=1
Here, ψnm = ψm (x, t, ζn ) for m = 1, 2 and n = 1, 2, . . . N. Equation (3.72)

shows that the general solution for q (x, t) is expressed as the sum of the N-
soliton (second term) and the nonsoliton part that is due to the first term,
which decays in time as dispersive waves. Also known as soliton radiation,
this is commonly referred to as ripples in fluid dynamics.
3.3.1 1-Soliton Solution

In the particular case of an exact N-soliton solution of the Kerr law, one obtains
analytic expressions of the field q (x, t) from the general formula (3.72). In fact,
in this case, the scattering data r (ξ ) vanishes, and a (ζ ) in (3.61) reads
N
ζ − ζn
a (ζ ) = (3.73)
n=1
ζ − ζn∗
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where N represents the number of discrete eigenvalues. With r (ξ ) = 0, the

GLM equations (3.70) and (3.71) reduce to the case of integral equations with
degenerate kernels and hence formulate algebraic equations for the column
vectors
⎛ ⎞
fl1
⎜ f ⎟
⎜ l2 ⎟
Fl = ⎜ ⎟ (3.74)
⎝ ··· ⎠
fl N
√
where fln = Cn ψl (x;ζn ) for l = 1, 2,
( I + M∗ M) F2∗ = E ∗ (3.75)
and
F1 = −MF2∗ (3.76)
where I is the N × N identity matrix, and the N × N matrix M is defined as

∗
e nem
Mnm = (3.77)
ζn − ζm∗
with

en = Cn e iζn x (3.78)
while E represents the column vector

⎛ ⎞
e1
⎜e ⎟
⎜ 2⎟
E =⎜ ⎟ (3.79)
⎝···⎠
en
From (3.72) the N-soliton solution can be written as

N
q (x, t) = −2 e n∗ f 2n
∗
(3.80)
n=1
In the case of a 1-soliton solution, one obtains

∗
(ζ1∗ − ζ1 )C1∗ (t)e −iζ1 x
ψ11 (x, t) = ∗ ∗ ∗ (3.81)
(ζ1 − ζ1 ) e i(ζ1 −ζ1 )x − |C1 (t)|2 e −i(ζ1 −ζ1 )x
2
∗
(ζ1∗ − ζ1 ) 2 e −iζ1 x
ψ12 (x, t) = ∗ ∗ ∗ (3.82)
(ζ1 − ζ1 ) 2 e 1 −ζ1 )x − |C1 (t)|2 e −i(ζ1 −ζ1 )x
i(ζ
∗
(ζ1∗ − ζ1 ) 2 e i(ζ1 +ζ1 )x
q (x, t) = −2C1∗ (x) ∗ ∗ (3.83)
(ζ1∗ − ζ1 ) 2 e −i(ζ1 −ζ1 )x − |C1 (t)|2 e −i(ζ1 −ζ1 )x
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
One can now express the scattering data as

1
ζ1 = (κ + i A) (3.84)
2
C1 (t) = −Ae σ1 (t)−iθ1 (t) (3.85)
where σ1 (t) = Aκt + σ10 and θ1 (t) = ( A2 − κ 2 )t/2 + θ10 . One obtains from
(3.83) the 1-soliton solution to (3.34) as given by (3.40). On comparing with
the notation that was introduced in chapter 2, the function g in (2.36) for Kerr
law is therefore given by
1 1
g[B(x − x̄(t))] = = (3.86)
cosh[B(x − x̄(t))] cosh τ
where κ is the frequency of the soliton, ω is the wave number, and σ0 is the
center of phase of the soliton. Also, the corresponding parameter dynamics are
dA
=0 (3.87)
dt
dB
=0 (3.88)
dt
dκ
=0 (3.89)
dt
and
d x̄
= −κ (3.90)
dt

An important property of the NLSE with the Kerr law of nonlinearity is that
it has an infinite number of integrals of motion. This characteristic is closely
related to the properties of the soliton solutions that are seen in the previ-
ous section. In fact, each conserved quantity gives a restriction, or control,
on the evolution of the solitons. By Noether’s theorem, each conserved law
corresponds to a symmetry that leaves the NLSE invariant. Examples of these
symmetries are the translational invariance and phase invariance.
Basically, in order to derive these conserved quantities, one exploits the fact
that the inverse transmission coefficient a (ζ ) in equation (3.73) is invariant,
namely ∂a /∂t = 0. As noted in the previous section, whenever |ζ | 1 and
(ζ ) ≥ 0, the function ln a (ζ ) is analytic on the upper half plane (ζ ) ≥ 0 and
ln a (ζ ) → 0 as |ζ | → ∞. Therefore, one may write the expansion [5]
∞
Cn
ln a (ζ ) = (3.91)
n=1
ζn
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
By using (3.73) (for (ζ ) > 0 and x → ∞, where ψ → 0) and equation (3.84),

one can obtain the conserved quantities of the NLSE as follows
∞
E= |q |2 d x = 2A (3.92)
−∞
∞
i
M= (q q x∗ − q ∗ q x )d x = −2κ A (3.93)
2 −∞
∞
1 2
H= (|q x |2 − |q |4 )d x = A(3κ 2 − A2 ) (3.94)
2 −∞ 3
The values of the energy, momentum, and the Hamiltonian are obtained by
using the 1-soliton solution of the NLSE given in (3.40). One additional con-
served quantity is

1 ∞
H1 = {q xx q x∗ − q x q xx
∗
+ 3|q |2 (q ∗ q x − q q x∗ )}d x = 2κ A(κ 2 − A2 ) (3.95)
2 −∞
which represents the next hierarchy of the NLSE.
For the perturbed NLSE with Kerr law given by
1
iq t + q xx + |q |2 q = i R[q , q ∗ ] (3.96)
2
the adiabatic parameter dynamics of the solitons are governed by

dA dB ∞ ∗
= = (q R + q R∗ )d x (3.97)
dt dt 2 −∞
and
∞ ∞
dκ ∗ ∗ ∗ ∗
dt 2A −∞ −∞
which can be obtained from (2.51) and (2.54), respectively, since E = 2A for
Kerr law nonlinearity. For Kerr law, with
x
R = δ|q | q + σ q
2m
|q |2 ds (3.99)
−∞
equations (2.51) and (2.52) reduce to
dE ( 1 )(m + 1)
= 2δ A2m+1 2 + 4σ A2 (3.100)
dt (m + 32 )
dM ( 1 )(m + 1)
= 2δκ A2m+1 2 + 4σ κ A2 (3.101)
dt (m + 32 )
Note that for the case of Kerr law, A = B. Considering the particular per-
turbation terms in (3.99), equations (3.97) and (3.98) integrate, on using the
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
1-soliton solution given in (3.40) to
dA dB ( 1 )(m + 1)
= = δ A2m+1 2 + 2σ A2 (3.102)
dt dt (m + 32 )
dκ
=0 (3.103)
dt
Finally, to obtain the velocity of the soliton for Kerr law, use equation (2.55)
and the perturbation terms given by (3.99). This integrates out to
∞
d x̄
v= = −κ + x(q ∗ R + q R∗ )d x = −κ + σ (3.104)
dt E −∞
where the energy E is given by (3.92).
3.4.1 Hamiltonian Structure

It is known that equations solvable by the IST, such as the NLSE with Kerr
law nonlinearity, are completely integrable Hamiltonian systems and the IST
amounts to a canonical transformation from physical variables to an infinite
set of action-angle variables. The phase space M0 is an infinite dimensional real
linear space with complex coordinates defined by a pair of functions q (x), r (x)
for which the variable x may be thought of as a coordinate label [153].
On the algebra of smooth functionals on the phase space M0 , a Poisson
structure is defined by the following Poisson brackets
∞
δ F δG δ F δG
{F, G} = i − dx (3.105)
−∞ δq (x) δr (x) δr (x) δq (x)
where the variational derivative is defined as
δ F (q , r ) = F (q + δq , r + δr ) − F (q , r )
∞
δF δF
= δq (x) + δr (x) d x (3.106)
−∞ δq (x) δr (x)
The bracket in (3.105) possesses the basic properties of Poisson brackets,

namely
1. skew-symmetry: {F, G} = −{G, F }

2. linearity: {a F +bG, H} = a {F, H}+b{G, H} for any constants a and b
3. Jacobi identity: {F, {G, H}} + {H, {F, G}} + {G, {H, F }} = 0
It is to be noted that the bracket in (3.105) is the infinite dimensional generaliza-

tion of the usual Poisson bracket in the phase space R2n with real coordinates
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
pk , q k for 1 ≤ k ≤ n, namely
n
∂ f ∂g ∂ f ∂g
{ f, g} = − (3.107)
∂q k ∂ pk ∂ pk ∂q k
k=1
The coordinates q (x), r (x) themselves may be considered as functional on M0 .

Their variational derivatives are the generalized functions
δq (x)
= δ(x − y) (3.108)
δq ( y)
δr (x)
= δ(x − y) (3.109)
δq ( y)
and
δq (x) δr (x)
= = δ(x − y) (3.110)
δr ( y) δq ( y)
where δ(x − y) is the Dirac δ-function. Substituting (3.108) and (3.109) into
(3.105) gives
{q (x), q ( y)} = {r (x), r ( y)} = 0 (3.111)

{q (x), r ( y)} = iδ(x − y) (3.112)
These results also give
δF
= −i{F, r (x)} (3.113)
δq (x)
δF
= −i{F, q (x)} (3.114)
δr (x)
A dynamical system is said to be Hamiltonian if one can identify general-

ized coordinates q , momenta p, and a Hamiltonian H( p, q , t) such that the
equations of motion of the system can be written as
∂q
= {q , H} (3.115)
∂t
∂p
= { p, H} (3.116)
∂t
where {, } denotes the Poisson bracket. Equations (3.115) and (3.116) are called
Hamilton’s equations of motion, and the variables ( p, q ) are called the conjugates.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
3.5 Variational Principle

Linear evolution equations that are studied in mathematical physics have
a variety of analytical techniques to solve them. Most of these methods—
namely the Fourier series method, integral transform techniques, and Green’s
function techniques—make direct or indirect use of the superposition princi-
ple. However, for nonlinear evolution equations, the superposition technique
does not hold and most of the analytical solution techniques become obsolete.
Although numerical experiments give an understanding of the behavior of
these equations, they serve as inspiration and guidance for theoretical efforts,
and significant efforts were subsequently made in the analytical techniques.
One such outstanding technique is the IST previously discussed. However,
it should be emphasized that the number of nonlinear equations that can be
solved by the IST is very limited, for example, to the NLSE with Kerr law
nonlinearity, the derivative NLSE, the Korteweg–de Vries equation, the Sin-
Gordon equation, and a few others. Furthermore, even in situations when
the method can be applied, the explicit information that can be extracted
from the solution is rather limited. This situation has prompted an effort to
complement the exact analytical solution methods by approximate methods
that sacrifice exactness in order to obtain explicit results and a clear physi-
cal picture of the properties of the solution. One such method that has been
found very useful in many investigations in nonlinear optics is a direct vari-
ational method based on trial functions and Rayleigh-Ritz optimization. This
method has several advantages for studying solitons in the nonlinear optics
community. Some of these advantages are [38]
1. This method is applicable to a perturbation problem for which the

unperturbed system may not be integrable. Also, this method only
requires that the unperturbed system admits a well-defined solu-
tion, such as a soliton, although this method has limitations.
2. It is a universal method that is suitable for equations in any dimen-
sions with external forces and potentials.
3. It often gives results that are quite similar to numerical simulations.
4. The method is simple and can be learned in a very short time.
5. This approach leads to a simplified, finite-dimensional dynamical
system with interesting and rich properties. Consequently, it is eas-
ier to study the reduced system than the original problem.
For a finite dimensional problem of a single particle, the temporal develop-

ment of its position is given by Hamilton’s principle of least action, which
states that the action given by the time integral of the Lagrangian is an ex-
tremum, namely
t2
δ L(x, ẋ)dt = 0 (3.117)
t1
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where x is the position of the particle and ẋ = d x/dt. The variational prob-
lem (3.117) then leads to the familiar Euler-Lagrange (EL) equation

∂L d ∂L
− =0 (3.118)
∂x dt ∂ẋ
The derivation of this EL equation can be found in the familiar work by
Goldstein. Here, it is assumed that the variations of x and ẋ, namely δx and
δ ẋ = d(δx)/dt, vanish at the boundary of the integration. To obtain the Hamil-
tonian formulation of (3.118), one needs to define the canonical momentum
p to the position x as
∂L
p= (3.119)
∂ẋ
The Hamiltonian H is defined by the Legendre transformation
H = p ẋ − L(x, ẋ) (3.120)
so that (3.119) and (3.120) lead to the Hamilton’s equations

dx ∂H
= (3.121)
dt ∂p
and
dp ∂H
=− (3.122)
dt ∂x
For the case of an infinite dimensional problem, such as the NLSE with Kerr
law of nonlinearity, Hamilton’s principle is extended to the extremum of the
action given by the integral of the Lagrangian, which is a real function of the
fields q , q ∗ and their derivatives

δ L(q , q ∗ , q t , q t∗ , q x , q x∗ , q xx , . . .)d xdt = 0 (3.123)
where L is given by
∞
1
L= [i(q ∗ q t − q q t∗ ) + |q |4 − |q x |2 ]d x (3.124)
2 −∞
The variation of (3.124) is then defined as

δ L(q , q ∗ , . . .)d xdt

1
= lim {L(q + (δq ), q ∗ + (δq ∗ ), . . .) − L(q , q ∗ , . . .)}d xdt (3.125)
→0
where the variations δq and δq ∗ are assumed to vanish at the boundary of the
integration. One has, as usual, the definitions δq t = ∂(δq )/∂t, δq t∗ = ∂(δq ∗ )/∂t,
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
δq x = ∂(δq )/∂x, and so on. Then, after integration by parts, (3.125) reduces to
∞ n

n ∂ ∂L ∂ ∂L ∗
(−1) ∗
− δq + {cc} d xdt = 0 (3.126)
∂ x n ∂q nx ∂t ∂q t∗
n=0
∗
where q nx = ∂n q ∗ /∂x n and cc represents the complex conjugate, while it is to
be noted that ( ∂ L/∂q nx ) ∗ = ∂ L/∂q nx
∗
. Since the variations δq and δq ∗ are taken
to be arbitrary and independent, one has
∞
∂n ∂ L ∂ ∂L
(−1) n ∗
− =0 (3.127)
∂x n ∂q nx ∂t ∂q t∗
n=0
and its complex conjugate. The function that makes the variational functional
stationary is also a solution of the corresponding nonlinear evolution equa-
tion. In this process, an intelligent guess is made for the evolution of q (x, t)
in the sense that the form of q as a function of x is modeled in terms of cer-
tain parameter functions, l, that characterize the crucial features of the solu-
tions, namely the amplitude, spatial width, phase variations, and others. The
parameters of this trial function are allowed to be functions of time (i.e.
l = l(t)). Inserting the trial function into the variational integral, the spa-
tial integration can be performed and a reduced variational problem for the
parameter functions l(t) is obtained. The EL equation of the reduced varia-
tional problem becomes

∂L d ∂L
− =0 (3.128)
∂l dt ∂lt
On using the ansatz for the soliton solution of the NLSE [185]
A(t)
q (x, t) = e i{−κ(t)(x−x̄(t))+δ(t)} (3.129)
cosh [B(t)(x − x̄(t))]
the Lagrangian that is given by (3.124) reduces to

d x̄ dδ 1
L = −2A κ + + A3 − Aκ 2 (3.130)
dt dt 3
Substituting A, B, κ, x̄, and δ for l in (3.128) yields the following set of
equations
dA
=0 (3.131)
dt
dB
=0 (3.132)
dt
dκ
=0 (3.133)
dt
d x̄
= −κ (3.134)
dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
dδ 1
= ( A2 + κ 2 ) (3.135)
dt 2
In the presence of the perturbation terms, the EL equation for (3.128) can be
extended to
∞
∂L d ∂L ∂q ∗ ∂q
− = i R − R∗ dx (3.136)
∂l dt ∂lt −∞ ∂l ∂l
where l represents the five soliton parameters. Once again, substituting A,

B, κ, t̄, and δ for l in (3.136), the following adiabatic evolution equations are
obtained
∞
dA dB 1
= = [Re −iφ ] dτ (3.137)
dt dt −∞ cosh τ
∞
dκ tanh τ
= − [Re −iφ ] dτ (3.138)
dt −∞ cosh 2τ
∞
d x̄ τ
= −κ + 2 [Re −iφ ] dτ (3.139)
dt A −∞ cosh τ

dδ 1 ∞ 1
= (κ 2 + A2 ) + [Re −iφ ] (1 − tanh τ )dτ
dt 2 A −∞ cosh τ
∞
κ τ
+ 2 [Re −iφ ] dτ (3.140)
A −∞ cosh τ
where, once again
τ = B(t)(x − x̄(t))
while
φ = −κ(t)(x − x̄(t)) − δ(t)
Equations (3.137)–(3.140) give the adiabatic dynamics of the soliton parame-

ters in the presence of perturbation terms.

In this section, in order to obtain the quasi-stationary (QS) solution to (3.96)
for R given by (3.99), the technique that was developed in section 2.5.1 will
be followed. So, using the ansatz given by (2.76), and keeping in mind that
F (s) = s, equation (2.77) reduces to

(0) 1 (0) 2 1 ∂2 q̂ (0)
− ρT + ρ X q̂ (0) + + ( q̂ (0) ) 3 = 0 (3.141)
2 2 ∂θ 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
∂q̂ (0)
ρ X(0) − v(0) =0 (3.142)
∂θ
Now, (3.142) implies
ρ X(0) = v(0) (3.143)
Setting
B2 1 2 1
= ρT(0) + ρ X(0) = ρT(0) + (v(0) ) 2 (3.144)
2 2 2
equation (3.141) changes to
B 2 (0) 1 ∂2 q̂ (0)
− q̂ + + ( q̂ (0) ) 3 = 0 (3.145)
2 2 ∂θ 2
whose solution is
A
q̂ (0) = (3.146)
cosh τ
where
τ = B(θ − θ̄) (3.147)
while
A= B (3.148)
and
d θ̄
=v (3.149)
dt
At O() level, decomposing q̂ (1) = φ̂ (1) + i ψ̂ (1) yields
B 2 (1) 1 ∂2 φ̂ (1) ! ∂2 q̂ (0)

− φ̂ + + 3( q̂ (0) ) 2 φ̂ (1) = ρT(1) + v(0) ρ X(1) q̂ (0) − (3.150)
2 2 ∂θ 2 ∂θ ∂ X
and
B 2 (1) 1 ∂2 ψ̂ (1) ∂q̂ (0) ∂q̂ (0)
− ψ̂ + + ( q̂ (0) ) 2 ψ̂ (1) = − − v(0)
2 2 ∂θ 2 ∂T ∂X
! ∂q̂ (0)
− ρ X(1) − v(1) + σ
∂θ
x
(0) (0)
− ρ XX q̂ − δ( q̂ )
(0) 2m+1
+ σ q̂ (0)
( q̂ (0) ) 2 ds (3.151)
−∞
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
(0)
Here, as discussed in the previous chapter, setting ρ XX = 0 in (3.151) to
eliminate frequency chirp gives
B 2 (1) 1 ∂2 ψ̂ (1) ∂q̂ (0) ∂q̂ (0)

− ψ̂ + + ( q̂ (0) ) 2 ψ̂ (1) = − − v(0)
2 2 ∂θ 2 ∂T ∂X
! ∂q̂ (0)
− ρ X(1) − v(1) + σ
∂θ
x
− δ( q̂ )
(0) 2m+1
+ σ q̂ (0)
( q̂ (0) ) 2 ds (3.152)
−∞
Fredholm’s alternative, when applied to (3.151), yields

∂B
=0 (3.153)
∂X
and
ρT(1) + v(0) ρ X(1) = 0 (3.154)
whereas applied to (3.152) yields
1
dB 2m+1 2 (m + 1)
= δB + 2σ B 2 (3.155)
dT m + 32
and
ρ X(1) = v(1) − σ (3.156)
Equation (3.153) shows that B is a function of T only and so is A, since A = B
for Kerr law. Thus, these O() equations reduce to
B 2 (1) 1 ∂2 φ̂ (1)
− φ̂ + + 3( q̂ (0) ) 2 φ̂ (1) = 0 (3.157)
2 2 ∂θ 2
and
B 2 (1) 1 ∂2 ψ̂ (1) ∂q̂ (0)
− ψ̂ + + ( q̂ (0) ) 2 ψ̂ (1) = − − δ( q̂ (0) ) 2m+1
2 2 ∂θ 2 ∂T
x
+ σ q̂ (0) ( q̂ (0) ) 2 ds (3.158)
−∞
whose solutions are respectively
φ̂ (1) = 0 (3.159)
and

∂θ̄ τ 1 dB τ2
ψ̂ (1)
= − + cosh τ
∂T cosh τ 2B 2 dT cosh τ
s2
B 2m−1 τ 2 1
+ σ cosh τ + 2δ cosh s2 ds1 ds2 (3.160)
cosh τ cosh2m+2 s1
which leads to the QS solution (2.98) for the Kerr law of nonlinearity.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Now, the dynamical system in (3.102) and (3.103) has a stable, fixed point,
called a sink, that is given by (A, κ) = ( Ā, 0) or (B, κ) = ( B̄, 0), where
2m−1
1
2σ (m + 32 )
Ā = B̄ = (3.161)
δ ( 12 )(m + 1)
Thus, there exists a steady-state soliton whose amplitude and frequency

(velocity) get locked to this fixed value in a medium where the energy growth
rate and energy losses fully compensate for each other. The QS soliton given
by (2.98) for Kerr law travels with the velocity given by (3.104) through an
optical fiber with a fixed amplitude and width that are given by (3.161).
3.7 Lie Transform

In this section, the lie transform (LT) technique will be used to analyze the
perturbed NLSE that contains only Hamiltonian-type perturbations. This is
because the LT technique is restricted to dispersive-type perturbations. So,
equation (3.96) will be studied with R given by [83, 185, 241]
R = λ(|q |2 q ) x + µ(|q |2 ) x q − γ q xxx (3.162)
Thus, the perturbed NLSE for Kerr law nonlinearity that will be studied
using the LT is
1
iq t + q xx + |q |2 q = i{λ(|q |2 q ) x + µ(|q |2 ) x q − γ q xxx } (3.163)
2
In (3.163), λ is the self-steepening coefficient for short pulses (typically ≤ 100
femto seconds), µ is the nonlinear dispersion coefficient, and γ is the coeffi-
cient of the third order dispersion [83, 185]. The self-steepening term creates
an optical shock on the trailing edge of the pulse in the absence of group
velocity dispersion effects. This phenomenon is due to the intensity depen-
dence of the group velocity that makes the peak of the pulse move slower than
the wings. The group velocity dissipates the shock and smooths the trailing
edge considerably. However, self-steepening still manifests through a shift of
the pulse center.
The NLSE, as given by (3.34), does not give correct prediction for pulse
widths smaller than 1 picosecond. For example, in solid-state solitary lasers,
where pulses as short as 10 femtoseconds are generated, the approxima-
tion breaks down. Thus, quasi-monochromaticity is no longer valid and,
consequently, higher order dispersion terms creep in. If group velocity dis-
persion is close to zero, one needs to consider the third-order dispersion
for performance enhancement along transoceanic distances. Also, for short
pulse widths where the group velocity dispersion changes within the spectral
bandwidth of the signal, one needs to take into account the presence of third-
order dispersion.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
The purpose of the LT is to transform the perturbed NLSE given by (3.163)

to a simple one, commonly called the normal form of the original equation,
that is integrable by the IST method. In the following subsection, the normal
form theory will be introduced and, finally, the LT technique will be utilized
to integrate it.
3.7.1 Introduction
In this subsection, first, the definition of the infinite-dimensional vector spaces
of differential polynomials of (q , q ∗ , q x , q x∗ , q xx , . . .) will be given, and then
several linear operators in these spaces will be given. The LT and the
Hamiltonian formalism will then be defined. The degree, Deg, of the dif-
ferential monomial, say X[q , q ∗ ], of (q , q ∗ , q x , q x∗ , . . .) is as follows
Deg( X) = (# of q s in X) + (# of q ∗ s in X)
+ (# of derivatives ∂/∂x in X) (3.164)
For example,
Deg(|q |2 q x ) = Deg(q xxx ) = 4
Now, let χ N [[q , q ∗ ]] be the set of all differential polynomials of degree N

that satisfy the symmetry property X[e iθ q , e −iθ q ∗ ] = e iθ X[q , q ∗ ] for any real
number θ. In addition, let χ N∗ [[q , q ∗ ]] be the complex conjugate of χ N [[q , q ∗ ]].
For example, χ3 [[q , q ∗ ]] = q xx , |q |2 q . These spaces form a finite dimensional
vector space over the field of complex numbers. Define the infinite dimen-
sional space of differential polynomials χ [[q , q ∗ ]] as the following direct sum
of the vector spaces χ N
∞
"
∗
χ[[q , q ]] = χ N [[q , q ∗ ]] = {q , q x , q xx , |q |2 q , . . .} (3.165)
N=1
The space χ ∗ [[q , q ∗ ]] denotes the complex conjugate of χ [[q , q ∗ ]]. In this space,
define the derivative as
∞
d ∂ ∗ ∂
= q (n+1)x + q (n+1)x ∗
(3.166)
dx ∂q nx ∂q nx
n=0
where q nx = ∂n q /∂x n . The directional derivative, or the Lie derivative, which

is denoted by ψ · ∇ with respect to ψ = ψ[q , q ∗ ] ∈ (χ × χ ∗ )[[q , q ∗ ]] is given by
∞

∂ ∗ ∂
ψ ·∇ = ψnx + ψnx ∗
(3.167)
∂q nx ∂q nx
n=0
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where ψnx = ∂n ψ/∂x n . Note that d/d x = q x . ∇ so that, in view of (3.167), one
can define the derivative d/dt as

d dq
= ·∇ (3.168)
dt dt
where t is a variable in the function q (x, t) while
dq
= X[q , q ∗ ] ∈ χ [[q , q ∗ ]] (3.169)
dt
for example, X[q , q ∗ ] = (i/2)q xx + i|q |2 q . The Lie bracket [X, Y] for X and
Y ∈ χ [[q , q ∗ ]] is defined as
[X, Y] = X · ∇Y − Y · ∇ X (3.170)
With (3.170), the space χ [[q , q ∗ ]] forms an infinite dimensional Lie algebra,
namely the space χ is closed under the product of (3.170).
Also, define the space χ (0) [[q , q ∗ ]] as the set of polynomials that satisfy the
relation X[e iθ q , e −iθ q ∗ ] = e iθ X[q , q ∗ ]. The conserved quantities of the NLSE
that are given by (3.92)–(3.94) belong to χ (0) [[q , q ∗ ]]. Now, define the LT of
the equation
∞
dq
= X [q , q ∗ ] = n Xn [q , q ∗ ] (3.171)
dt n=0
where Xn is a homogenous polynomial of (q , q ∗ , q x , q x∗ . . .) and Deg(Xn+1 ) =

Deg( Xn ) + 1. The LT or the Lie exponential transform q = q [q , q ∗ ] ∈ χ [q , q ∗ ]
is defined as
q = exp(φ · ∇) Q (3.172)
where the exponential of the derivative reads

1
exp(φ · ∇) = 1 + φ · ∇ + φ · ∇(φ · ∇) + · · · (3.173)
2!
with φ · ∇ is the directional derivative as defined in (3.167), 1 is the identity
operator, and the function φ [Q, Q∗ ] is called the Lie-generating function.
Here, the transformed equation for Q is given by
∞
dQ
= Y [Q, Q∗ ] = n Yn [Q, Q∗ ] (3.174)
dt n=0
where Y0 = X0 and Yn [Q, Q∗ ] ∈ χ [[Q, Q∗ ]] with Deg(Yn ) = Deg(Xn ), being

defined from the solvability condition for the equation of φ , as will be shown
below. From (3.171), (3.172), and (3.174), the following relation between X ,
Y , and φ can be established
(e φ ·∇ X [Q, Q∗ ]) · ∇ = Y [Q, Q∗ ] · ∇e φ ·∇ (3.175)

P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
This is the transformation rule for the vector fields (3.171) and (3.174), with
the change of coordinates given by (3.172). With the definition of the Lie
bracket given by (3.170), the relation given by (3.175) can be written as the
Campbell-Baker-Hausdroff (CBH) formula
∞
1
Y = [φ , [φ , . . . [φ , X ] . . .]]
n=0
n!
1
= X + [φ , X ] + [φ , [φ , X ]] + · · · (3.176)
2!
Assuming φ in (3.172) has a power series in , namely
∞

∗
φ [Q, Q ] = n φn [Q, Q∗ ] (3.177)
n=1
the CBH formula given by (3.176) gives the following equations for φn at each
order
[X0 , φ1 ] = X1 − Y1 (3.178)
1
[X0 , φ2 ] = X2 + [φ1 , X1 ] + [φ1 , [φ1 , X0 ]] − Y2
2
1
= X2 − Y2 + [φ1 , X1 + Y1 ] (3.179)
2
1 1 1
[X0 , φ3 ] = X3 − Y3 + [φ1 , 2X2 + Y2 ] + [φ2 , X1 + 2Y1 ] + [φ1 , [φ1 , X1 ]]
3 3 6
(3.180)
and so on. In general, the equation for φn is of the form
[X0 , φn ] ≡ a d X0 φn = Fn − Yn (3.181)
where the linear operator a d X 0 = [X0 , ·] is called the adjoint representation

of X0 , while Fn is determined from ( X1 , . . . , Xn , Y1 , . . . Yn−1 , φ1 , . . . , φn−1 ). For
the NLSE X0 = 1/2Qxx + i|Q|2 Q
a d X0 : χ [[Q, Q∗ ]] → χ N+2 [[Q, Q∗ ]] (3.182)
and the polynomial parts of the kernel of the operator a d X0 , denoted by

ker (a d X0 ), are
ker (a d X0 ) ∩ χ1 = {i Q} (3.183)
ker (a d X0 ) ∩ χ2 = {Qx } (3.184)

i
ker (a d X0 ) ∩ χ3 = Qxx + i|Q| Q = X0
2
(3.185)
2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
The solution of (3.181) is then obtained by decomposing Fn ∈ χn+3 [[Q, Q∗ ]]

into the following direct sum
"
{im(a d X0 ) (im(a d X0 )) c } ∩ Fn (3.186)
where (im(a d X0 )) c is the complement of the space im(a d X0 ). Then Yn is chosen

to be
Yn = (im(a d X0 )) c } ∩ Fn (3.187)
which gives (a d X0 φn ) = im(a d X0 ) ∩ Fn and thus φn can be obtained. In partic-

ular, in the choice of Yn in (3.187), note that the linear part of Yn , ∂n+2 Q/∂x n+2
is equal to that of Fn , since im(a d X0 ) contains only nonlinear terms. The linear
part appearing in Yn is a consequence of the linear resonances caused by the
higher order corrections in the linear dispersion relation.
The solution space φn in (3.181) is denoted by ξn [[Q, Q∗ ]], namely
ξn [[Q, Q∗ ]] = {S[[Q, Q∗ ]] | a d X0 S ∈ χn+3 } (3.188)
From (3.188)
χn+1 [[Q, Q∗ ]] ⊂ ξn [[Q, Q∗ ]] (3.189)
In general, the nonpolynomial part of ψn , denoted by ξn \χn+1 , cannot be com-

pletely determined because of the infinite dimensional nature of the operator
a d X0 . However, since the dimension of χn+3 is finite, only a finite number of
independent solutions of a d X0 φn ∈ χn+3 is necessary. In a direct calculation
for n = 1, one finds the nonpolynomial part of ξ1 as
x
∗
ξ1 \χ2 [[Q, Q ]] = i Q |Q(x)| ds
2
(3.190)
−∞
while for n = 2, one obtains

x x
∗ ∗ ∗
ξ2 \χ3 [[Q, Q ]] = Q ( Qx Q − QQx )ds, Qx |Q(x)| ds 2
(3.191)
−∞ −∞
Similar results occur for higher degrees. However, in higher order problems
for (3.181), Fn , depending on the previous solution φn−1 , may not be a polyno-
mial, and the algebra of ξ = ⊕n=1 ξn becomes complicated. In many practical
problems, however, an asymptotic result, including only the first few orders,
might be necessary. Thus, one can assume the spaces of the Lie-generating
functions to be
x
φ1 ∈ ξ1 = i Qx , i Q |Q(x)|2 ds (3.192)
−∞
x x
∗
φ2 ∈ ξ2 = Qxx , |Q| Q, Q
2
( Qx Q − QQ∗x )ds, iQ |Q(x)| ds
2
−∞ −∞
(3.193)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Note from (3.191) that [ξ1 , χ4 ] ⊂ χ5 , which leads to F2 ∈ χ5 . Thus by virtue of

(3.172), the original equation (3.171) gets Lie transformed into (3.174), which is
characterized only by the operator a d X0 . Equation (3.174) is called the normal
form of (3.171).
3.7.2 Application
In this subsection, the theory of LT that was developed in the last section will
be utilized to study (3.163). Note that one can rewrite equation (3.163) in the
following form
q t = X0 [q , q ∗ ] + X1 [q , q ∗ ] + O( 2 ) (3.194)
where X0 is the NLSE part and X1 represents the Hamiltonian perturbation

terms, namely
i
X0 [q , q ∗ ] = q xx + i|q |2 q (3.195)
2
and
X1 [q , q ∗ ] = {−γ q xxx + λ(|q |2 q ) x + µ(|q |2 ) x q } (3.196)
Now look for the LT for the perturbed (3.163) to a simple system that is known
as the normal form. By virtue of the previous subsection, the transformation
is defined by
q = e ·∇ = Q + φ1 [Q, Q∗ ] + O( 2 ) (3.197)
where φ = φ1 + O( 2 ) satisfies (3.178), namely
[X0 , φ1 ] = X1 − Y1 (3.198)
The function Y1 = Y1 [Q, Q∗ ] is the first-order term in the normal form
Qt = X0 [Q, Q∗ ] + Y1 [Q, Q∗ ] + O( 2 ) (3.199)
Following the procedure that is given in the previous subsection, the solution
to (3.198) is
x
i
φ1 [Q, Q∗ ] = − (λ + 6γ ) − i(6γ + 2λ + µ) Q |Q(s)|2 ds (3.200)
2 −∞
with
Y1 [Q, Q∗ ] = −γ ( Qxxx + 6|Q|2 Qx ) (3.201)
The normal form given by (3.199) with (3.195) and (3.201) is known as
the Schrödinger-Hirota equation and is integrable by the IST. The 1-soliton
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
solution is given by
A
Q(x, t) = e i(−κ x+ωt+σ0 ) (3.202)
cosh [B(x − vt − x0 )]
where the velocity and the wave number of the soliton are respectively
v = −κ + γ ( B 2 − 3κ 2 ) (3.203)
and
B2 − κ2
ω= + γ κ(3B 2 − κ 2 ) (3.204)
2
This is, thus, an alternative method of solving the perturbed NLSE with Kerr
law nonlinearity up to O(), with Hamiltonian-type perturbations only. These
solutions can also be recovered by the QS method.
Based on the discussions in this section, one can conclude that, in the
regime of non-Hamiltonian-type perturbation, one is compelled to use the QS
method since the LT fails in presence of non-Hamiltonian-type perturbation.
Also, note that, in both methods, one can obtain a closed form of the perturbed
soliton at the O() level; however, in the QS method, one can also obtain the
adiabatic parameter dynamics of optical solitons that are also recoverable
using the soliton perturbation theory. Once again, the LT fails to do that. So
the question arises, why would anyone want to use the LT? The QS method
is widely used in various other types of perturbations, including the nonlocal
type, while the LT has been very sparingly used since it came into existence
in 1994 [241].
In this context, it needs to be pointed out that the QS method has already
been extended beyond Kerr law nonlinearity, namely to power law, parabolic
law, and dual-power law, as will be seen in subsequent chapters. However,
the question of extending the method of LT beyond Kerr law nonlinearity still
remains open at this stage.
Exercises
1. In the perturbed NLSE with Raman scattering and a nonlocal per-

turbation term given by
x
1
iq t + q xx + |q |2 q = [iν(|q |2 ) x q + χq t |q |2 dτ ]
2 −∞
prove that the adiabatic parameter dynamics of the soliton ampli-

tude, width, and frequency are respectively given by
dA dB 2
= = − χ A3
dt dt 3
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
dκ 2
= A2 (5κχ − 4ν A2 )
dt 15
while the velocity is given by
v = −κ − χ A
2. Consider the NLSE with the perturbations that are given by

1
iq t + q xx + |q |2 q
2
#
= i λ(|q |2 q ) x + ν(|q |2 ) x q − γ q xxx − iβ1 (q 2 q t∗ ) x
$
−iβ2 q x2 q ∗ − iβ3 q ∗ (q 2 ) xx
Prove that for these perturbations the adiabatic parameter dynamics

for the solitons are given by
dA dB dκ
= = =0
dt dt dt
Also prove that, in this case, the velocity of the soliton is given by
A2
v = −κ − 3γ κ 2 + {2κ(4β3 + β2 − β1 ) − (3λ + 2ν + 3γ )}
3
Such types of perturbations where there is no adiabatic deforma-
tion of soliton amplitude and frequency are known as Hamiltonian
perturbations.
3. For the stable fixed point derived for quasi-stationary solitons,
establish the fixed value of the amplitude given by (3.161)
4. Prove the Jacobi identity for Poisson bracket that is given after
equation (3.106).
5. Obtain the traveling wave solution of the Schrodinger-Hirota
equation
1
iq t + q xx + |q |2 q = iγ q xxx + 6|q |2 q x
2
Your wave number should match with the one that is given by (3.204).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
4
Power Law Nonlinearity
In this chapter, the dynamics of optical solitons with power law nonlinearity,
a generalization of Kerr law nonlinearity, are studied. Section 4.1 briefly dis-
cusses the physics of power law nonlinearity and the mathematical issues of
the nonlinear Schrödinger’s equation (NLSE) with power law nonlinearity.
Section 4.2 reviews the three conserved quantities of the NLSE with power
law nonlinearity. It also discusses the adiabatic parameter dynamics of the
solitons with power law due to the presence of perturbation terms. Finally,
Section 4.3 discusses the quasi-stationarity (QS) aspect of the perturbed NLSE
with power law nonlinearity.
4.1 Introduction
Power law nonlinearity is exhibited in various materials, including semicon-
ductors. This law also occurs in media for which higher order photon pro-
cesses dominate at different intensities. Moreover, in nonlinear plasmas, the
power law solves the problem of small-K condensation in weak turbulence
theory. This law is also treated as a generalization to Kerr law nonlinearity.
For power law, the refractive index is given by [74, 81, 86]
n = n0 + n2 |E|2 p (4.1)
where n0 is the linear refractive index of the medium, n2 is the higher order
nonlinear coefficient, and E is the electric field of the light wave. The dimen-
sionless form of the NLSE with power law is
1
iq t + q xx + |q |2 p q = 0 (4.2)
2
Here, one needs to have 0 < p < 2 to prevent wave collapse. In particular, it is
necessary to have the restriction p = 2 to avoid the self-focusing singularity
issue. This aspect of self-focusing singularity is discussed in detail in the book
by Ablowitz and Segur [5]. Also, for p = 1 in (4.2) one recovers the NLSE with
57
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
the Kerr law of nonlinearity. It is to be noted that (4.2) is not integrable by the
inverse scattering transform (IST) unless p = 1, which was already discussed
in the previous chapter. For the case where p = 1, the 1-soliton solution can
be derived by the traveling wave ansatz that was introduced in chapter 2.

For the case of power law nonlinearity, (2.1) reduces to
1
iq t + q xx + |q |2 p q = 0 (4.3)
2
so that (2.10) simplifies to
B 2 g − (κ 2 − 2ω)g + 2A2 p g 2 p+1 = 0 (4.4)
Multiplying (4.4) by g and integrating yields
2g 2
(g ) 2 = [( p + 1)(κ 2 − ω) − A2 p g 2 p ] (4.5)
B 2 ( p + 1)
Separating variables and integrating yields
1
( p + 1) 2 dg
x − vt = √ 1 (4.6)
2 g[( p + 1)(κ 2 − ω) − A2 p g 2 p ] 2
Substituting
( p + 1)(κ 2 − ω)
g2 p = (4.7)
A2 p cosh2 θ
leads to the 1-soliton solution
A
q (x, t) = 1 e (−iκ x+iωt+iσ0 ) (4.8)
cosh [B(x − vt − x̄)]
p
where
κ = −v (4.9)
and
B 2 − p2 κ 2
ω= (4.10)
2 p2
with
12
2 p2
B=A p
(4.11)
1+ p
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Power Law Nonlinearity 59
Here, A is the amplitude of the soliton, B is its width, v is its velocity, κ is the
soliton frequency, ω is the wave number, and x̄ and σ0 are the center of the
soliton and the center of the soliton phase, respectively. Note that in this case
the width of the soliton given by B is related to the amplitude A as seen in
(4.11), unlike in the Kerr law case where A = B, although the relation (4.11)
reduces to A = B on setting p = 1.
The corresponding parameter dynamics for the solitons are given by
dA
=0 (4.12)
dt
dB
=0 (4.13)
dt
dκ
=0 (4.14)
dt
and
d x̄
= −κ (4.15)
dt

The NLSE for power law nonlinearity has three conserved quantities, which
can be obtained from (2.41)–(2.43) using the fact that F (s) = s p so that f (s) =
s p+1 / p + 1. The three integrals of motion for the power law, respectively, are
[81, 86]
1 1 1
∞
1+ p 2 2 p
E= |q | d x = A
2 2− p

−∞ 2 p2 1p + 12
1p 1 1
2− p 1 + p 2 p
=B p (4.16)
2 p2 1p + 12

i ∞ ∗
M= (q q x − q q x∗ )d x
2 −∞
1p 1 1
1 + p 2 p
= 2κ A2− p
2 p2 1p + 12
1p 1 1
2− p 1 + p 2 p
= 2κ B p (4.17)
2p 2
1p + 12
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
∞
1 1
H= |q x | −
2
|q | 2 p+2
dx
−∞ 2 p+1
2 1 2 1 1
Bp 1+ p p B + κ 2 p2 2 p 12 p+1
p
= − 2B p+1 1
2 p2 2 p2 B 1p + 12 p +2

1p 1 1 1
A2 κ 2 p2 1 + p 2 2 p
2 p2
= A p
+
2 p2 1+ pAp 2 p2 1p + 12

2 2 1 p+1
1
2 p 2 p
−2Ap (4.18)
1+ p p+1
p
+ 12
In this chapter, the perturbed NLSE with power law nonlinearity that is going
to be studied is
1
iq t + q xx + |q |2 p q = i R[q , q ∗ ] (4.19)
2
where
x
R = δ|q |2m q + σ q |q |2 ds (4.20)
−∞
In the presence of the perturbation term in (4.20), one recovers the modified
integrals of motion. Using the first two conserved quantities, (4.16) and (4.17),
the adiabatic variation of the soliton parameters are

p−1
dA 2 p2 1p + 12 ∞ ∗
2p
= Ap−1 (q R + q R∗ )d x (4.21)
dt 2− p 1+ p 12 1p −∞
1 1
2 p2 p p + 2
1 ∞
dB p 2 p−2
= B p 1 1 (q ∗ R + q R∗ )d x (4.22)
dt 2− p 1+ p 2 p −∞
and

2 p 1 + 1
1
dκ p−2 2 p p 2
= B p
dt 1+ p 12 1p
∞ ∞
∗ ∗ ∗ ∗
. i (q x R − q x R )d x − κ (q R + q R )d x (4.23)
−∞ −∞
The adiabatic variations of the energy and momentum of the soliton can
be obtained from equations (2.51) and (2.52), respectively, for power law
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
nonlinearity. They are

δ A2m+2 ( 2 )( p )
1 m+1
dE
= 2
dt B ( m+1p
+ 12 )

∞ τ
σ A4 1 1
+ 2 2 2 ds dτ (4.24)
B −∞ cosh p τ −∞ cosh p s
and

δ A2m+2 ( 2 )( p )
1 m+1
dM
= 2κ
dt B ( m+1p
+ 12 )

∞ τ
σ A4 1 1
+ 2 2 2 ds dτ (4.25)
B −∞ cosh p τ −∞ cosh p s
Substituting the perturbation term R from (4.20) and carrying out the inte-
grations in (4.21), (4.22), and (4.23) for the soliton given by (4.8), one obtains
21p
dA 2δ 2m+1 1+ p ( 1p + 12 ) ( m+1
p
)
= A
dt 2− p 2 p2 ( 1p ) ( m+1
p
+ 12 )
2p
1 + p p+1 ( p + 2 )
1 1
σ
+ A3− p
2− p 2 p2 ( 1p )( 12 )
∞ τ
1 1
. 2 2 ds dτ (4.26)
−∞ cosh p τ −∞ cosh p s
m
2δp 1 + p p ( p + 2 ) ( p )
1 1 m+1
dB 2m+ p
= B( p )
dt 2− p 2p 2
( p ) ( p + 2 )
1 m+1 1
1
1 + p p ( p + 2 ) 2p
1 1
σ p
+ B
2− p 2 p2 ( 1p )( 12 )
∞ τ
1 1
. 2 2 ds dτ (4.27)
dκ
=0 (4.28)
dt
Once again, in (4.26) and (4.27) one can observe that it is necessary to have
p=
2 due to self-focusing singularity.
The velocity of the soliton, as an evolution of the center of mass, is given by
∞
d x̄
v= = −κ + x(q ∗ R + q R∗ )d x (4.29)
dt E −∞
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
which, after integration using (4.8) and (4.20), reduces to

∞ τ
A4 τ 1
v = −κ + 2σ 3 ds dτ (4.30)
B E −∞ cosh 2p τ 2
−∞ cosh p s

In order to obtain the QS solution to (4.19) with R given by (4.20), one can, as
in the previous chapter, follow the developments of section 2.5.1. Keeping in
mind that F (s) = s p , one gets at the leading order [81, 86]

(0) 1 (0) 2 1 ∂2 q̂ (0)
− ρT + (ρ X ) q̂ (0) + + ( q̂ (0) ) 2 p+1 = 0 (4.31)
2 2 ∂θ 2
and
∂q̂ (0)
ρ X(0) − v(0) =0 (4.32)
∂θ
Now, (4.32) implies
ρ X(0) = v(0) (4.33)
On setting
B2 1 2 1
= ρT(0) + ρ X(0) = ρT(0) + (v(0) ) 2 (4.34)
2 p2 2 2
(4.31) changes to
B 2 (0) 1 ∂2 q̂ (0)
− q̂ + + ( q̂ (0) ) 2 p+1 = 0 (4.35)
2 p2 2 ∂θ 2
whose solution is
A
q̂ (0) = 1 (4.36)
cosh [B(θ − θ̄)]
p
where
12
2 p2
B=A p
(4.37)
1+ p
and
d θ̄
=v (4.38)
dt
Thus, the soliton frequency and wave number at the first order, as depicted
in (4.33) and (4.34), agree with (4.9) and (4.10). At O(), decompose
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
q̂ (1) = φ̂ (1) + i ψ̂ (1) into its real and imaginary parts. Now the equations for
φ̂ (1) and ψ̂ (1) are, respectively,
B 2 (1) 1 ∂2 φ̂ (1) (1) (0) (1)

(0) ∂2 q̂ (0)
− φ̂ + + (2 p + 1)( q̂ φ̂
(0) 2 p (1)
) = ρ T + v ρ X q̂ −
2 p2 2 ∂θ 2 ∂θ ∂ X
(4.39)
and
B 2 (1) 1 ∂2 ψ̂ (1)
− ψ̂ + + ( q̂ (0) ) 2 p ψ̂ (1)
2 p2 2 ∂θ 2
∂q̂ (0)
∂q̂ (0) ∂q̂ (0)
=− − v(0) − ρ X(1) − v(1) + σ
∂T ∂X ∂θ
x
(0) (0)
− ρ XX q̂ + δ( q̂ (0) ) 2m+1 + σ q̂ (0) ( q̂ (0) ) 2 ds (4.40)
−∞
(0)
In (4.40), set ρ XX = 0 to eliminate frequency chirp to obtain
B 2 (1) 1 ∂2 ψ̂ (1)
− ψ̂ + + ( q̂ (0) ) 2 p ψ̂ (1)
2 p2 2 ∂θ 2
∂q̂ (0)
∂q̂ (0) ∂q̂ (0)
=− − v(0) − ρ X(1) − v(1) + σ
∂T ∂X ∂θ
x
+ δ( q̂ (0) ) 2m+1 + σ q̂ (0) ( q̂ (0) ) 2 ds (4.41)
−∞
Fredholm’s alternative, applied to (4.39), yields

∂B
=0 (4.42)
∂X
and
ρT(1) + v(0) ρ X(1) = 0 (4.43)
whereas, if applied to (4.41), implies

mp
dB 2δp 1+ p ( 1p + 12 ) ( m+1
p
) 2m+ p
= B( p )
dT 2− p 2 p2 ( 1
p
) ( m+1
p
+ 1
2
)
1p
σp 1+ p ( 1p + 12 ) 2
∞
1 τ
1
+ Bp ds dτ
2− p 2 p2 ( 1p )( 12 ) −∞ cosh p τ
2
−∞
2
cosh p s
(4.44)
and
ρ X(1) = v(1) − σ (4.45)

P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Equation (4.42) shows that B is a function of T alone and so is A due to (4.37).

Thus, by virtue of (4.37), one obtains
21p
dA 2δ 1+ p ( 1p + 12 ) ( m+1
p
)
= A2m+1
dT 2− p 2 p2 ( 1p ) ( m+1
p
+ 12 )
p+1
( 1p + 12 )
2p
∞ τ
σ 1+ p 1 1
+ A3− p ds dτ
2− p 2 p2 ( 1p )( 12 ) −∞ cosh p τ
2
−∞ cosh p s
2
(4.46)
Equations (4.44) and (4.46) represent the adiabatic parameter dynamics of the
soliton width and amplitude, respectively, in the presence of perturbation
terms. Also, note that from these two equations, the perturbation scheme
breaks down if p = 2. Now (4.39), by virtue of (4.42) and (4.43), reduces to
B 2 (1) 1 ∂2 φ̂ (1)
− φ̂ + + (2 p + 1)( q̂ (0) ) 2 p φ̂ (1) = 0 (4.47)
2 p2 2 ∂θ 2
while (4.41), by virtue of (4.42) and (4.45), simplifies to
B 2 (1) 1 ∂2 ψ̂ (1)
− ψ̂ + + ( q̂ (0) ) 2 p ψ̂ (1)
2 p2 2 ∂θ 2
x
∂q̂ (0)
=− + δ( q̂ (0) ) 2m+1 + σ q̂ (0) ( q̂ (0) ) 2 ds (4.48)
∂T −∞
Finally, the solutions of (4.42) and (4.43) are, respectively,
φ̂ (1) = 0 (4.49)
and
1 1
2B p 1 + p 2 p
ψ̂ (1)
=−
p 2 p2
φ
∂θ̄ φ 2
s2
tanh s 1
. cosh p s2 ds1 ds2
∂T cosh p φ
1 2
cosh p s1

1 dB φ 2
s2
tanh s 1
− 2 cosh p s2 2 ds1 ds2
B dT cosh p s1

1 dB φ 2
s2
tanh s 1
+ 3 cosh p s2 2 ds1 ds2
B dT cosh p s1
2m+1 φ
2m−2 p+1 1 + p 2p 2
s2
tanh s 1
+ 2δ B p cosh p s2 2m+2 ds1 ds2
2 p2 cosh( p ) s1
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
φ
A3 1 2
+ 2σ 3 cosh p s3
B cosh 1p φ

s3 s2
1 1
. 2 2 ds1 ds2 ds3 (4.50)
cosh p s2 −∞ cosh p s1
which leads to the QS solution given by (2.98).

The fixed point of the dynamical system given by (4.26) and (4.28) (or by
(4.27) and (4.28)) is
12 2m+1p−2
σ Ip 1+ p m+1
p
+ 12
Ā = 1 (4.51)
2δ 2 p2 m+1
p
2
and
m−1 2m+pp−2
σ Ip 2 p2 p m+1
p
+ 12
B̄ = 1 (4.52)
2δ 1+ p m+1
p
2
with
∞ τ
1 1
Ip = 2 2 ds dτ (4.53)
This fixed point is a sink. This physically means that the QS soliton for power
law nonlinearity travels through an optical fiber with the velocity given by
(4.30) at a fixed amplitude and width as in (4.51) and (4.52).
Exercises
1. Consider the following NLSE, with filters as perturbation terms

1
iq t + q xx + |q |2 p q = iβq xx
2
Prove that the adiabatic variation of the amplitude and frequency
are respectively given by
p−1 1
2p + 12
dA 2β Ap−1 2 p2 p
=
dt 2− p B p+1 1p
p+1
A2 B 2 p A2 2 2 1p
. − 2 (κ p + B ) 1
2

p 2 p+1
p
+ 12 p p + 12
and
12
dκ 4β 2 p−2 2 p2 p−2
= 2 κ A2 B p
dt p p+1 p+2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
2. For the power law nonlinearity with Hamiltonian perturbations

given by
1
iq t + q xx + |q |2 p q = i[λ(|q |2 q ) x + ν(|q |2 ) x q − γ q xxx ]
2
prove that the velocity of the soliton is given by

12 + 1p 2p
v = −κ − (3λ + 2ν) 1
4 2 + 2p 1p
p−1 1
3γ p 2 + p
1
3γ
− 2 + 2
4p p 1p 32 + 1p
3. Prove that the fixed point of the dynamical system given by (4.26)
and (4.28) is given by ( Ā, 0) where
12 2m+1p−2
σ Ip 1+ p m+1
p
+ 12
Ā = 1
2δ 2 p2 m+1
p
2
4. Consider the moment of inertia of the soliton that is given by

∞
J = x 2 |q |2 d x
−∞
Prove that J satisfies the relation

d2 J
= 8H
dt 2
where H is the Hamiltonian.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
5
Parabolic Law Nonlinearity
This chapter talks about the optical solitons of the nonlinear Schrödinger’s
equation (NLSE) with parabolic law nonlinearity, commonly known as cubic-
quintic nonlinearity. Section 5.1 contains a detailed discussion of the physics
of parabolic law and the mathematical aspects of the equation. Section 5.2
reviews the three conserved quantities of the NLSE with parabolic law non-
linearity. Also discussed here are the adiabatic dynamics of the optical soliton
parameters due to perturbation terms. Finally, Section 5.3 discusses the quasi-
stationary solution of the perturbed NLSE with parabolic law.
5.1 Introduction
In this chapter, the NLSE with parabolic law nonlinearity will be studied. It
has been known for a long time that optical beams can self-focus in both space
and time while propagating in a nonlinear medium. The collapse of two- and
three-dimensional optical beams in a Kerr law medium was considered as
a means of producing high electric field strengths. It was observed that the
inclusion of a saturable nonlinearity could halt the singular collapse, thus
causing the formation of an optical beam that propagates without changing
its temporal or spatial shape, held together by nonlinear effects.
To obtain some knowledge of the diameter of the self-trapping beam, it is
necessary to consider nonlinearities higher than the third order. It was recog-
nized in 1960s and 1970s that saturation of the nonlinear refractive index plays
a fundamental role in the self-trapping phenomenon. Higher order nonlin-
earities arise by retaining the higher order terms in the nonlinear polarization
tensor. For a fifth order nonlinearity, the refractive index is given by [82]
n = n0 + n2 |E|2 + n4 |E|4 (5.1)
where n0 is the linear refractive index of the medium and |E|2 is the electric
field intensity of the light wave, while n2 = 3χ (3) /8n0 and n4 = 5χ (5) /16n0
with n0 > n2 |E|2 > n4 |E|4 . Here, n2 and n4 respectively represent the
67
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
third- and fifth-order nonlinear coefficients. In general, the coefficients n2 and

n4 could be positive or negative, depending on the medium and the frequency
selected.
Little attention was paid to the propagation of optical beams in fifth-order
nonlinear media because no analytic solutions were known and it seemed
that the chances of finding any material with a significant fifth-order term
were slim. However, recent developments have rekindled interest in this area.
The optical susceptibility of CdSx Se1−x -doped glasses was experimentally
shown to have a considerable χ (5) , or fifth-order susceptibility. It was also
demonstrated that a significant χ (5) nonlinearity effect exists in a transparent
glass in intense femtosecond pulses at 620 nm.
Besides these saturation effects, it was also proposed that the doping of
silica glass with two appropriate semiconductor particles leads (in a region
far from saturation) exactly to the parabolic form of the refractive index with
an effective increased value of n4 and a reduced value of n2 . In other materials,
the values of n2 and n4 may also depend on the doping.
The organic nonlinear material polydiacetylene para-toluene sulfonate
(PTS) is another material that can be described exactly by the parabolic law.
Measurements of PTS indicate that n2 > 0 and n4 > 0 at 1600 nm, with low
loss and negligible multisoliton absorption. The spectrum of PTS reveals three
two-photon excited states that can account for the observed value of n2 . The
absence of nearby states that can saturate can be interpreted as n4 arising
from the Stark shift of the two-photon states. This leads to the conclusion that
the parabolic law model can be considered as an exact model for PTS and
similar materials and this material yields solitons at lower power than has
previously been achieved. The experimental confirmation places PTS in the
unique class of solid-state materials with positive third-order and negative
fifth-order nonlinearity.
A recent study of the nonlinear self-phase modulation of a fundamental
beam involved in a process of third-harmonic generation showed that the
beam experienced an additional higher order phase shift as a consequence of
the n4 coefficient being a sum of two terms. The first is due to the inherent
χ (5) susceptibility of the medium and the second is due to the cascading of
the third-order nonlinearity, namely
n4 = ndir
4 + n4
casc
with the cascading term ncasc (3) 2

4 being proportional to (χ ) . The cascading term
casc dir
n4 can always be made to exceed n4 by manipulating the sample length,
wave-vector mismatch, and beam intensity.
These developments and sustained theoretical interest in optical beam
propagation motivated the study of solitons in a parabolic medium. The
dimensionless form of the NLSE with parabolic law nonlinearity is given
by [82, 86]
1
iq t + q xx + (|q |2 + ν|q |4 )q = 0 (5.2)
2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Parabolic Law Nonlinearity 69
where ν is a constant. For ν = 0, the Kerr law of nonlinearity is recovered.

Equation (5.2) is not integrable by the inverse scattering transform. However,
one can solve it by the traveling wave ansatz.

For the case of parabolic law nonlinearity, equation (2.1) reduces to
1
iq t + q xx + (|q |2 + ν|q |4 )q = 0 (5.3)
2
so that equation (2.10) simplifies to
B 2 g − (κ 2 − 2ω)g + 2A2 g 3 + 2ν A4 g 5 = 0 (5.4)
g2
(g ) 2 = [3(κ 2 − 2ω) − 3A3 g 2 − 2ν A4 g 4 ] (5.5)
3B 2
√ dg
x − vt = 3 1 (5.6)
g[3(κ 2 − 2ω) − 3A3 g 2 − 2ν A4 g 4 ] 2
which leads to the 1-soliton solution
A
q (x, t) = 1 e i(−κ x+ωt+σ0 ) (5.7)
[1 + a cosh {B(x − x̄(t))}] 2
where
√
B(t) = 2A(t) (5.8)
κ = −v (5.9)
A2 κ2
ω= − (5.10)
4 2
and

4
a = 1 + ν A2 (5.11)
3
soliton and the center of the soliton phase, respectively. Equation (5.8) gives
the relation between the amplitude and width of the soliton for the case of
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
power law nonlinearity. Thus, the soliton parameter dynamics are
dA
=0 (5.12)
dt
dB
=0 (5.13)
dt
dκ
=0 (5.14)
dt
and
d x̄
= −κ (5.15)
dt

The NLSE with parabolic law nonlinearity has three integrals of motion. These
are obtained from (2.41)–(2.43) by setting F (s) = s + νs 2 so that f (s) =
s 2 /2 + νs 3 /3. Thus the three integrals of motion are
⎧
∞
⎪
⎨ A2 3
tan−1 A νω :0 < ν < ∞
B 2ν 3
E= |q | dx =
2
νω
(5.16)
−∞ ⎪
⎩ A2 −1
B
− 2ν3 tanh A − 3 : − 16ω
3
<ν<0
∞
i
M= (q q x∗ − q ∗ q x )dx
2 −∞
⎧
⎪
⎨ − κ2 AB 2ν3 tan−1 A νω
2
:0 < ν < ∞
3
= (5.17)
⎩ − κ A2 − 3 tanh−1 A − νω
: − 3 < ν < 0
⎪
2 B 2ν 3 16ω
and
∞
1 1 4 ν 6
H= |q x | − |q | − |q | dx
2
−∞ 2 2 3
⎧ √ √ √3+16νω
⎪
⎨ − 3 8ν2ω + 8ν3 2ν3 tan−1 − 3+4√νω :0 < ν < ∞
=
⎩ − 3 2ω − 3 − 3 tanh−1 − 3+√ 3−16νω
: − 3 < ν < 0
√ √ √
⎪
8ν 8ν 2ν 4 −νω 16ω
(5.18)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
In this chapter, the perturbed NLSE with parabolic law nonlinearity that is
going to be studied is
1
iq t + q xx + (|q |2 + ν|q |4 )q = i R[q , q ∗ ] (5.19)
2
where
x
R = δ|q | q + σ q
2m
|q |2 ds (5.20)
−∞
In the presence of the perturbation term in (5.20), the soliton parameters

deform adiabatically. These adiabatic variations are given by
√ ∞
dA 2 2
= a (q ∗ R + q R∗ )dx (5.21)
dt 2 −∞
∞
dB
= a 2 (q ∗ R + q R∗ )dx (5.22)
dt −∞
and
∞ ∞
dκ B
= 2 i (q x∗ R − q x R∗ )dx − κ (q ∗ R + q R∗ )dx (5.23)
dt AE −∞ −∞
where E is the energy of the soliton given by (5.16). From (2.51) and (2.52),
the adiabatic variations of the energy and the linear momentum of the soliton
due to parabolic law nonlinearity take the forms
√
dE δ 2A2m+1 3 a −1 1
= F m + 1, m + 1, m + ; B m + 1,
dt 2m a m+1 2 2a 2
τ
2σ A4 ∞ 1 1
+ 2 ds dτ (5.24)
B −∞ 1 + a cosh τ −∞ 1 + a cosh s
and
√
dM δ 2A2m+1 3 a −1 1
= κ F m + 1, m + 1, m + ; B m + 1,
dt 2m a m+1 2 2a 2
τ
2σ A4 ∞ 1 1
+ ds dτ (5.25)
B 2 −∞ 1 + a cosh τ −∞ 1 + a cosh s
where F (α, β; γ ; z) is the Gauss’ hypergeometric function defined as

∞
(γ ) (α + n)(β + n) zn
F (α, β; γ ; z) = (5.26)
(α)(β) n=0 (γ + n) n!
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and B(l, m) is the usual beta function. Substituting the perturbation term R
from (5.20) and carrying out the integrations in (5.21)–(5.23) yields

dA δ A2m+1 3 a −1 1
= m m−1 F m + 1, m + 1, m + ; B m + 1,
dt 2 a 2 2a 2
√ ∞ τ
2 1 1
+ σ a 2 ds dτ (5.27)
2 −∞ 1 + a cosh τ −∞ 1 + a cosh s
√
dB δ 2A2m+1 3 a −1 1
= F m + 1, m + 1, m + ; B m + 1,
dt 2m a m−1 2 2a 2
∞ τ
1 1
+ σ a 2 ds dτ (5.28)
−∞ 1 + a cosh τ −∞ 1 + a cosh s
and
dκ
=0 (5.29)
dt
The velocity of the soliton, as the evolution of the center of mass, is given by

d x̄ ∞
v= = −κ + x(q ∗ R + q R∗ )dx (5.30)
dt E −∞
which, after integration, or by directly using (2.67), reduces to
∞ τ
A4 τ 1
v = −κ + 2σ ds dτ (5.31)
B E −∞ 1 + a cosh τ −∞ 1 + a cosh s

In this section, the QS solution to the perturbed NLSE with parabolic law
nonlinearity will be obtained. Following the method that was developed in
section 2.5, one gets at the leading order

1 2 1 ∂2 q̂ (0)
− ρT(0) + ρ X(0) q̂ (0) + + ( q̂ (0) ) 3 + ν( q̂ (0) ) 5 = 0 (5.32)
2 2 ∂θ 2
and
∂q̂ (0)
ρ X(0) − v(0) =0 (5.33)
∂θ
Now, (5.33) implies
ρ X(0) = v(0) (5.34)
On setting
B2 1 2 1
= ρT(0) + ρ X(0) = ρT(0) + (v(0) ) 2 (5.35)
8 2 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
(5.32) changes to
B 2 (0) 1 ∂2 q̂ (0)
− q̂ + + ( q̂ (0) ) 3 + ν( q̂ (0) ) 5 = 0 (5.36)
8 2 ∂θ 2
whose solution is
q̂ (0) = Ag[B(θ − θ̄)] (5.37)
where
1
g(τ ) = 12 (5.38)
1+ 1+ 4
3
ν A2 cosh τ
√
B=A 2 (5.39)
τ = B(θ − θ̄) (5.40)
and
d θ̄
=v (5.41)
dt
At O() level, decomposing q̂ (1) = φ̂ (1) + i ψ̂ (1) into its real and imaginary
parts, the equations for φ̂ (1) and ψ̂ (1) , by virtue of (5.36), are, respectively,
B 2 (1) 1 ∂2 φ̂ (1)
− φ̂ + + {3( q̂ (0) ) 2 + 5ν( q̂ (0) ) 4 }φ̂ (1)
8 2 ∂θ 2
∂2 q̂ (0)
= ρT(1) + v(0) ρ X(1) q̂ (0) − (5.42)
∂θ ∂ X
and
B 2 (1) 1 ∂2 ψ̂ (1) ∂q̂ (0) (0) ∂q̂

(0)
− ψ̂ + + {( q̂ (0) 2
) + ν( q̂ (0) 4
) }ψ̂ (1)
= − − v
8 2 ∂θ 2 ∂T ∂X
∂q̂ (0)
− ρ X(1) − v(1) + σ (0) (0)
− ρ XX q̂ − δ( q̂ (0) ) 2m+1
∂θ
x
+ σ q̂ (0) ( q̂ (0) ) 2 ds (5.43)
−∞
(0)
In (5.43), set ρ XX = 0 to eliminate frequency chirp to obtain
B 2 (1) 1 ∂2 ψ̂ (1) ∂q̂ (0) ∂q̂ (0)

− ψ̂ + + {( q̂ (0) ) 2 + ν( q̂ (0) ) 4 }ψ̂ (1) = − − v(0)
8 2 ∂θ 2 ∂T ∂X
x
∂q̂ (0)
− ρ X(1) − v(1) + σ − δ( q̂ (0) ) 2m+1 + σ q̂ (0) ( q̂ (0) ) 2 ds (5.44)
∂θ −∞
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Fredholm’s alternative, when applied to (5.42) yields
∂B
=0 (5.45)
∂X
and
ρT(1) + v(0) ρ X(1) = 0 (5.46)
whereas if applied to (5.44) implies

√
dB δ 2A2m+1 3 a −1 1
= F m + 1, m + 1, m + ; B m + 1,
dT 2m a m−1 2 2a 2
∞ τ
1 1
+ σa2 ds dτ (5.47)
−∞ 1 + a cosh τ −∞ 1 + a cosh s
and
ρ X(1) = v(1) − σ (5.48)
Equation (5.45) shows that B is a function of T only, and so is A by virtue of

(5.39) so that

dA δ A2m+1 3 a −1 1
= m m−1 F m + 1, m + 1, m + ; B m + 1,
dT 2 a 2 2a 2
√ ∞ τ
2 2 1 1
+ σa ds dτ (5.49)
2 −∞ 1 + a cosh τ −∞ 1 + a cosh s
Again, note that (5.49) and (5.47) are the same as (5.27) and (5.28), respec-
tively, while relations (5.46) and (5.48) cannot be recovered by the soliton
perturbation theory. The O() equations now reduce to
B 2 (1) 1 ∂2 φ̂ (1)
− φ̂ + + {3( q̂ (0) ) 2 + 5ν( q̂ (0) ) 4 }φ̂ (1) = 0 (5.50)
8 2 ∂θ 2
and
B 2 (1) 1 ∂2 ψ̂ (1)
− ψ̂ + + {( q̂ (0) ) 2 + ν( q̂ (0) ) 4 }ψ̂ (1)
8 2 ∂θ 2
x
∂q̂ (0)
=− − δ( q̂ )
(0) 2m+1
+ σ q̂ (0)
( q̂ (0) ) 2 ds (5.51)
∂T −∞
whose solutions are, respectively,
φ̂ (1) = 0 (5.52)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
τ
1 ∂θ̄ 1 s2
a sinh s1
ψ̂ (1)
= −√ (1 + a cosh s2 ) ds1 ds2
2 ∂ T 1
(1 + a cosh τ ) 2 1 + a cosh s1
τ
1 dA 1
+ (1 + a cosh s2 )
2A2 dT (1 + a cosh τ ) 12
s2
a s1 sinh s1 1 dA 1
ds1 ds2 −
(1 + a cosh s1 ) 2 2A2 dT (1 + a cosh τ ) 12
τ s2
1 1
. (1 + a cosh s2 ) ds1 ds2 − δ A2m−1
1 + a cosh s1 1
(1 + a cosh τ ) 2
τ s2
a s1 sinh s1
. (1 + a cosh s2 ) ds1 ds2
(1 + a cosh s1 ) m+1
φ
A3 1
+ 2σ 3 (1 + a cosh s3 )
B (1 + a cosh τ ) 12
s3 s2
1 1
. ds1 ds2 ds3 (5.53)
1 + a cosh s2 −∞ 1 + a cosh s1
which lead to the QS soliton that is given by (2.98).

The dynamical system in (5.27) and (5.29) or (5.28) and (5.29) has a stable
fixed point, so that the QS soliton for parabolic law nonlinearity propagates
through an optical fiber with a velocity given by (5.31).
Exercises
1. Prove that as ν approaches zero, the energy of the parabolic law soli-
ton approaches the energy of the Kerr law soliton—in other words,
lim E = 2A
ν→0
Remember to consider the right-hand and left-hand limits separately.

2. Prove that the parabolic law soliton that is given by (5.7) is equiva-
lent to the Kerr law soliton with ν = 0.
3. Prove that for the parabolic law NLSE with filters as perturbation
terms
1
iq t + q xx + (|q |2 + ν|q |4 )q = iβq xx
2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
the adiabatic parameter dynamics of the soliton amplitude and

frequency are given by

dA √ a 2 B2 −1 a −1
= −4 2β √ tan
dt a −1
2 a +1

a 2
−1 a − 1 1
− βa 2 A3 3 tan −
(a 2 − 1) 2 a + 1 2(a 2 − 1)
and
√
dκ 2A3 B a2 −1 a −1 1
= −β tan −
dt E 3
(a 2 − 1) 2 a + 1 2(a − 1)
2
where E is the energy of the soliton.

4. Consider the Hamiltonian-type perturbations for the NLSE with
parabolic law nonlinearity
1
iq t + q xx + (|q |2 + ν|q |4 )q = i[λ(|q |2 q ) x + µ(|q |2 ) x q − γ q xxx ]
2
Prove that the velocity is given by
√
2 A3 a − 1
−1
v = −κ − (3λ + 2µ) a 2 − 1 − 2 tan
2E (a 2 − 1) 32 a +1

√ γ A a 4 A2 2κ 2 −1 a − 1 a 2 A2 2
−3 2 +√ tan − a −1
E 2 a2 − 1 a +1 4
5. Prove that the series for Gauss’ hypergeometric function F (a , b; c; x)

that is defined in (5.26) is convergent for c > a + b.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
6
Dual-Power Law Nonlinearity
This chapter talks about the optical solitons of the nonlinear Schrödinger’s
equation (NLSE) with dual-power law nonlinearity, a generalization of the
parabolic law of nonlinearity. Section 6.1 contains a brief discussion of the
physics of the dual-power law and the mathematical results of the optical
solitons with dual-power law. Section 6.2 talks about the three conserved
quantities of the NLSE with dual-power law. Also in this section are the results
of the adiabatic parameter dynamics of the solitons in presence of perturbation
terms. Finally, Section 6.3 talks about the quasi-stationary solution (QS) of the
perturbed NLSE.
6.1 Introduction
The NLSE with the dual-power law of nonlinearity appears in various areas of
mathematical physics and nonlinear optics. This model is used to describe the
saturation of the nonlinear refractive index. It also serves as a basic model to
describe the solitons in photovoltaic-photorefractive materials such as lithium
niobate. The propagation of ultrashort optical pulses in a nonlinear medium
can be characterized by the nonlinear refractive index that is given by [88]
n = n0 + n2 |E|2 p + n4 |E|4 p (6.1)
Here, also as in the parabolic law case, n0 > n2 |E|2 p > n4 |E|4 p . The dimen-
sionless form of the NLSE with dual-power law of nonlinearity is
1
iqt + q xx + (|q |2 p + ν|q |4 p )q = 0 (6.2)
2
It should be noted that in (6.2), ν = 0 reduces to the case of power law nonlin-
earity, and if in addition p = 1, the case of Kerr law nonlinearity is recovered.
If, however, ν = 0 and p = 1, one falls back to the case of parabolic law nonlin-
earity that was studied in the previous chapter. Thus, the case of dual-power
law is the most generalized case whose exact soliton solution is known that
77
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
is studied here. In the next chapter, a different kind of nonlinearity will be

studied whose exact soliton solution is not known. Although the dual-power
law and parabolic law have been further extended to higher order polyno-
mial law and triple-power law nonlinearity, their exact soliton solutions are
not yet known. Studies are under way to obtain an exact soliton solution for
these laws of nonlinearity.
Although (6.2) is not integrable by the inverse scattering transform, one
can apply the traveling wave technique to obtain the 1-soliton solution of the
NLSE with dual-power law nonlinearity.

For the case of dual-power law nonlinearity, (2.1) reduces to
1
iqt + q xx + (|q |2 p + |q |4 p )q = 0 (6.3)
2
so that (2.10) simplifies to
B 2 g − (κ 2 − 2ω)g + 2A2 p g 2 p+1 + 2ν A4 p g 4 p+1 = 0 (6.4)
g2
(g ) 2 = [( p + 1)(2 p + 1)(κ 2 − 2ω) − 2(2 p + 1) A2 p g 2 p
B 2 ( p + 1)(2 p + 1)
− 2ν( p + 1) A4 p g 4 p ] (6.5)
x − vt
√
( p + 1)(2 p + 1)dg
= 1
[( p + 1)(2 p + 1)(κ 2 − 2ω) − 2(2 p + 1) A2 p g 2 p − 2ν( p + 1) A4 p g 4 p ] 2
(6.6)
which leads to the 1-soliton solution

A
q (x, t) = 1 e i{−κ x+ωt+σ0 } (6.7)
[1 + a cosh{B(x − x̄(t))}] 2p
where
21p
4 p2
B=A p
(6.8)
1+ p
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Dual-Power Law Nonlinearity 79
with
κ = −v (6.9)
A 2p
κ 2
ω= − (6.10)
2p + 2 2
and

ν B 2 (1 + p) 2
a= 1+ (6.11)
2 p2 1 + 2 p
soliton and the center of the soliton phase, respectively. For dual-power law
nonlinearity, solitons exist for
2 p2 1 + 2 p
− <ν<0 (6.12)
B 2 (1 + p) 2
The corresponding parameter dynamics for the solitons are given by
dA
=0 (6.13)
dt
dB
=0 (6.14)
dt
dκ
=0 (6.15)
dt
and
d x̄
= −κ (6.16)
dt

For dual-power law nonlinearity, setting F (s) = s p + νs 2 p , so that f (s) =
s p+1 / p+1+νs 2 p+1 /2 p+1, (2.41)–(2.43) gives the three conserved quantities as
∞
2A2 1 1 1 1 a −1 1 1
E= |q | d x =
2
1 1 F , ; + ; B , (6.17)
−∞ B2 p a p p p 2 p 2a p 2

i ∞
M= (q q x∗ − q ∗ q x )d x
2 −∞

2κ A2 1 1 1 1 a −1 1 1
=− 1 1 F , ; + ; B , (6.18)
B2 p a p p p 2 p 2a p 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
∞
1 |q |2 p+2 |q |4 p+2
H= |q x | −
2
−ν dx
−∞ 2 p+1 2p + 1

A2 B 1 1 3 1 a −1 1 3
= 1 1 2
F 2+ , ; + ; B ,
2 p a p 4p p p 2 p 2a p 2

κ 1 1 1 1 a −1 1 1
+ F , ; + ; B ,
B p p 2 p 2a p 2

A2 p 1 1 3 1 a −1 p+1 1
− F 1+ ,1+ ; + ; B ,
a B( p + 1) p p 2 p 2a p 2

νA 4p
1 1 5 1 a −1 2p + 1 1
− 2 F 2+ ,2+ ; + ; B , (6.19)
2a B(2 p + 1) p p 2 p 2a p 2
In this chapter, the perturbed NLSE with dual-power law nonlinearity that is
going to be studied is
1
iqt + q xx + (|q |2 p + ν|q |4 p )q = i R[q , q ∗ ] (6.20)
2
where
x
R = δ|q |2m q + σ q |q |2 ds (6.21)
−∞
Now, in the presence of the perturbation term in (6.21), the integrals of

motion are modified to give the modified integrals of motion. Using the first
two integrals of motion, namely (6.17) and (6.18), the adiabatic variation of
the soliton parameters are
1
dA p+1 2 ∞ ∗
= (q R + q R∗ )d x (6.22)
dt pL Ap−1 2 p2 −∞

dB ∞ ∗
= (q R + q R∗ )d x (6.23)
dt L −∞
and
∞ ∞
dκ ∗ ∗ ∗ ∗
dt E −∞ −∞
where E is the energy as given by (6.17) while

1
( 12 )( 1p ) (a − 1)(2 p + 1) p
L=
( 1p + 12 ) 2ν(1 + p)
2
2ν ( p + 1) 3 1 1 1 1 1−a
F , ; + ;
a p 3 (a − 1)(2 p + 1) 2 2 p 2 p 1+a
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book

2 1 1 1 1 1−a
− F , ; + ;
B2 2 p 2 p 1+a

2ν ( p + 1) 2 1 1 1 1 1−a
− F , ; + ; (6.25)
a p 2 (a − 1) 2 ( p + 2)(2 p + 1) 2 p 2 p 1+a
The adiabatic evolution of the energy and linear momentum, in the presence
of perturbation terms, by virtue of (2.51) and (2.52) are

dE 2δ A2m+1 m+1 m+1 m+1 1 a −1 m+1 1
= 2 m+1 m+1 F , , + ; B ,
dt Ba p 2 p p p p 2 2a p 2

σ A4 ∞ 1 τ
1
+ 2 1 1 ds dτ (6.26)
B −∞ (1 + a cosh τ ) p −∞ (1 + a cosh s) p
and

dM m+1 m+1 m+1 1 a −1
2δ A2m+1 m+1 1
= 2κ m+1 m+1 F , , + ; B ,
dt Ba p 2 p p p p 2 2a p 2

τ
σ A4 ∞ 1 1
+ 2 1 1 ds dτ (6.27)
In order to obtain the adiabatic evolution of the soliton parameters, one can
use (6.22)–(6.24) with the perturbation terms given by (6.21) and the form of
the dual-power law soliton given by (6.7) to get
1
dA 2 A3− p p+1 2
=
dt pL B 2 p2

δ A2m m+1 m+1 m+1 1 a −1 m+1 1
m+1 m+1 F , , + ; B ,
a p 2 p p p p 2 2a p 2

τ
A2 ∞ 1 1
+σ 1 ds dτ (6.28)
B −∞ (1 + a cosh τ ) 1p −∞ (1 + a cosh s) p

dB 2 A2 δ A2m m+1 m+1 m+1 1 a −1 m+1 1
= m+1 F , , + ; B ,
dt B L a m+1 p 2 p p p p 2 2a p 2

τ
A2 ∞ 1 1
+σ 1 1 ds dτ (6.29)
and
dκ
=0 (6.30)
dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
The velocity of the soliton, as an evolution of the center of mass, is given by

d x̄ ∞
v= = −κ + x(q ∗ R + q R∗ )d x (6.31)
dt E −∞
which, after integration, or by directly using (2.67), reduces to
∞

τ
d x̄ A4 τ 1
v= = −κ + 2σ 3 1 ds dτ
dt B E −∞ (1 + a cosh τ ) 1p −∞ (1 + a cosh s) p
(6.32)

For the QS solution to the NLSE with dual-power law nonlinearity, follow the
developments of section 2.5. One gets at the leading order

(0) 1 (0) 2 1 ∂2 q̂ (0)
− ρT + ρ X q̂ (0) + + ( q̂ (0) ) 2 p+1 + ν( q̂ (0) ) 4 p+1 = 0
2 2 ∂θ 2
(6.33)
and
∂q̂ (0)
ρ X(0) − v(0) =0 (6.34)
∂θ
Now, (6.34) implies
ρ X(0) = v(0) (6.35)
On setting
B2 1 2 1
= ρT(0) + ρ X(0) = ρT(0) + (v(0) ) 2 (6.36)
4 p2 2 2
(6.33) changes to
B 2 (0) 1 ∂2 q̂ (0)
− q̂ + + ( q̂ (0) ) 2 p+1 + ν( q̂ (0) ) 4 p+1 = 0 (6.37)
4 p2 2 ∂θ 2
whose solution is
q̂ (0) = Ag[B(θ − θ̄)] (6.38)
where
1
g(τ ) = 1 (6.39)
(1 + a cosh τ ) 2 p
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
21p
4 p2
B=A p
(6.40)
1+ p
along with
τ = B(θ − θ̄) (6.41)
while
d θ̄
=v (6.42)
dt
At O() level, decomposing q̂ (1) = φ̂ (1) + i ψ̂ (1) into its real and imaginary
parts, the equations for φ̂ (1) and ψ̂ (1) , by virtue of (6.37), are, respectively,
B 2 (1) 1 ∂2 φ̂ (1)
− φ̂ + + φ̂ (1) {(2 p + 1)( q̂ (0) ) 2 p + ν(4 p + 1)( q̂ (0) ) 4 p }
4 p2 2 ∂θ 2
∂2 q̂ (0)
= ρT(1) + v(0) ρ X(1) q̂ (0) − (6.43)
∂θ ∂ X
and
B 2 (1) 1 ∂2 ψ̂ (1) ∂q̂ (0) (0) ∂q̂

(0)
− ψ̂ + + {( q̂ (0) 2 p
) + ν( q̂ (0) 4 p
) }ψ̂ (1)
= − − v
4 p2 2 ∂θ 2 ∂T ∂X
∂q̂ (0) 2m+1
− ρ X(1) − v(1) + σ (0) (0)
− ρ XX q̂ + δ q̂ (0)
∂θ
x
+ σ q̂ (0) ( q̂ (0) ) 2 ds (6.44)
−∞
(0)
In (6.44), setting ρ XX = 0 to eliminate frequency chirp gives
B 2 (1) 1 ∂2 ψ̂ (1) ∂q̂ (0) (0) ∂q̂

(0)
− ψ̂ + + {( q̂ (0) 2 p
) + ν( q̂ (0) 4 p
) }ψ̂ (1)
= − − v
4 p2 2 ∂θ 2 ∂T ∂X
x
∂q̂ (0)
− ρ X(1) − v(1) + σ + δ( q̂ (0) ) 2m+1 + σ q̂ (0) ( q̂ (0) ) 2 ds (6.45)
∂θ −∞
Fredholm’s alternative applied to (6.43) yields
∂B
=0 (6.46)
∂X
and
ρT(1) + v(0) ρ X(1) = 0 (6.47)

P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
whereas, if applied to (6.45), implies

dB 2A2 δ A2m m+1 m+1 m+1 1 a −1 m+1 1
= m+1 m+1 F , , + ; B ,
dT BL a p 2 p p p p 2 2a p 2

∞ τ
A2 1 1
+σ 1 1 ds dτ (6.48)
and
ρ X(1) = v(1) − σ (6.49)
Equation (6.46) shows that B is a function of T only and so is A by virtue of

(6.40) so that
12
dA 2 A3− p p+1 δ A2m m+1 m+1 m+1 1 a −1
= m+1 m+1 F , , + ;
dT pL B 2 p2 a p 2 p p p p 2 2a

m+1 1 A2 ∞ 1
.B , +σ
p 2 B −∞ (1 + a cosh τ ) 1p

τ
1
. 1 ds dτ (6.50)
−∞ (1 + a cosh s) p
Although (6.48) and (6.50) were obtained before by the soliton perturbation
theory (SPT), relations (6.46), (6.47), and (6.49) cannot be recovered by the
SPT. Thus, these O() equations reduce to
B 2 (1) 1 ∂2 φ̂ (1)
− φ̂ + + φ̂ (1) {(2 p + 1)( q̂ (0) ) 2 p + ν(4 p + 1)( q̂ (0) ) 4 p } = 0 (6.51)
4 p2 2 ∂θ 2
and
B 2 (1) 1 ∂2 ψ̂ (1)
− ψ̂ + + {( q̂ (0) ) 2 p + ν( q̂ (0) ) 4 p }ψ̂ (1)
4 p2 2 ∂θ 2
x
∂q̂ (0)
=− + δ( q̂ (0) ) 2m+1 + σ q̂ (0) ( q̂ (0) ) 2 ds (6.52)
∂T −∞
whose solutions are, respectively,
φ̂ (1) = 0 (6.53)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
τ
A ∂θ̄ 1 1
ψ̂ (1)
=− (1 + a cosh s2 ) p
Bp ∂T (1 + a cosh τ ) 21p

s2
a sinh s1
. p+1 ds1 ds2
(1 + a cosh s1 ) p
τ
A dB 1 1
+ 3 1 (1 + a cosh s2 ) p
B p dT (1 + a cosh τ ) 2 p

s2
a s1 sinh s1 2A dA 1
. p+1 ds1 ds2 − 2
(1 + a cosh s1 ) p B dT (1 + a cosh τ ) 21p
τ

s2
1 1
. (1 + a cosh s2 ) p 1 ds1 ds2
(1 + a cosh s1 ) p
τ
A2m 1 1
+ 2δ 2 1 (1 + a cosh s2 ) p
B (1 + a cosh τ ) 2 p

s2
1
. m+1 ds1 ds2
(1 + a cosh s1 ) p
φ
A2 1 1
+ 2σ 3 (1 + a cosh s3 ) p
B (1 + a cosh τ ) 21p

s3 s2
1 1
. 1 1 ds1 ds2 ds3 (6.54)
(1 + a cosh s2 ) p −∞ (1 + a cosh s1 ) p
which lead to the QS solution given by (2.98).

The dynamical system in (6.28) and (6.30) or (6.29) and (6.30) has a sta-
ble fixed point so that the QS soliton for dual-power law nonlinearity that
propagates through an optical fiber with a velocity given by (6.32).
Exercises
1. Prove that for ν −→ 0, the soliton given by (6.7) approaches the

soliton for power law nonlinearity.
2. Prove that for ν −→ 0, the energy of the unperturbed dual-power
law soliton reduces to the case of power law solitons—in other
words
1
1 + p 2 ( 2 )( p )
1 1
lim E = A2− p
ν→0 2 p2 ( 1p + 12 )
3. Prove that for p −→ 1, the energy of the unperturbed dual-power

law soliton reduces to the case of parabolic law solitons—in other
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
words
⎧
⎪
⎨ 3
tan−1 [2A ν3 ] : 0 < ν < ∞
2ν
lim E =
p→1 ⎪
⎩
− 2ν3 tanh−1 [2A − ν3 ] : − 4A
3
2 < ν < 0
4. Consider the NLSE for dual-power law nonlinearity with filters

1
iq t + q xx + (|q |2 p + ν|q |4 p )q = iβq xx
2
Prove that the adiabatic variation of the amplitude and frequency

dA 4β A2 1 1 1 1 a −1 1 1
=− 1 1 κ F
2
, , + ; B ,
dt Ba p 2 p p p 2 p 2a p 2

1 1 3 1 a −1 1 3
+ B2 F 2 + , , + ; B ,
p p 2 p 2a p 2
and
1 a −1

βκ B 2 F 2 + p , 1 + p , 2 + p ; 2a B 1 + p , 1
1 1 1
dκ
=− 2 2 −1

dt 4p A F 1p , 1p , 12 + 1p ; a2a B 1p , 12
5. For the NLSE with Hamiltonian-type perturbations given by

1
iq t + q xx + (|q |2 p + ν|q |4 p )q = i[λ(|q |2 q ) x + µ(|q |2 ) x q − γ q xxx ]
2
prove that the velocity is given by
3 a −1
1
γ B 2 F 2 + p , p , p + 2 ; 2a
1 1 1
B ,3
v = −κ − (µ + 3γ B ) − 3 2
−1
1p 21
2 p2 F 1p , 1p , 12 + 1p ; a2a B p, 2
−1
2 1
A2 F 2p , 2p , 2p + 12 ; a2a B ,
− 3λ + 2µ 1 1 1 1 1 1 a −1
1p 21
2 p a p F p , p , 2 + p ; 2a B p , 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
7
Saturable Law Nonlinearity
An important class of optical nonlinearity that has recently drawn much

attention is the saturable law nonlinearity, also known as saturating nonlin-
earity. This chapter is devoted to the study of soliton propagation in saturat-
ing media. Section 7.1 explains the basic features of saturating nonlinearity.
The derivation of the nonlinear Schrödinger’s equation (NLSE) with sat-
urable law is given in Section 7.2. The conserved quantities are identified in
Section 7.3. Saturating media admit bistable solitons, which are discussed in
Section 7.4. Finally, Section 7.5 is devoted to the discussion of arbitrary pulse
propagation.
7.1 Introduction
The practical interest in the investigation of optical pulse propagation in
nonlinear waveguides is concentrated on low-loss materials with nonreso-
nant nonlinearities. At low intensity of optical field nonresonant nonlinearity
in materials of practical interest resembles Kerr nonlinearity. Therefore, a very
large number of investigations on temporal soliton propagation in fibers and
waveguides have been carried out employing Kerr law nonlinearity. In Kerr
media, the magnitude of the induced nonlinearity increases linearly with an
increase in the value of the intensity of the optical field. However, as the inci-
dent field becomes stronger, for optical fields whose frequencies approach
a resonant frequency of the material, non-Kerr higher order nonlinearity
comes into play, essentially changing the physical features and stability of
optical soliton propagation. Particularly for short pulses and high-input peak
power, the field-induced change of the refractive index cannot be described
by a Kerr-type nonlinearity because it is influenced by higher order non-
linearities. As a consequence, the optically induced refractive index change
becomes saturated at higher field strength. This is especially important in ma-
terials with higher nonlinear coefficients, for example, semiconductor-doped
glasses and organic polymers in which the saturation of nonlinear refractive
index changes come to play at moderately high intensities and should be
87
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
1.2
Second Model
n2–n20
First Model
2n0n2lS
0.6
0
0 2 4 6 8 10
[E]2/IS
FIGURE 7.1
Behavior of two different form of saturating nonlinearity.
taken into account. The saturating nonlinearity can be modeled by different

equations. Two known models expressing the nonlinear refractive index n are
2n0 n2 |E|2
n2 = n20 + (7.1)
1 + |E|2 /Is
and
2

− |E|
n =
2
n20 + 2n0 n2 Is 1 − e Is (7.2)
where n0 is, as usual, the linear refractive index; E is the electric field; and
n2 and Is are respectively the third-order nonlinear coefficients and the char-
acteristic saturation intensity. The parameter n2 , as seen in chapter 3, is also
known as the Kerr coefficient. For small intensities, namely when |E|2 Is ,
both (7.1) and (7.2) reduce to Kerr law nonlinearity. However, for large inten-
sities, namely when |E|2 Is , the refractive index saturates and approaches
its maximum value, 2n0 n2 Is . The variation of the induced refractive index
with |E|2 is shown in Figure 7.1. Both models signify saturation at large in-
tensity. Since both models produce identical qualitative features of solitons,
discussion using either one of them will be sufficient. Hence, one may choose
only the first one for further discussion.
7.2 The NLSE

Consider the propagation of an optical pulse in a medium possessing the
higher order nonlinear refractive index given by (7.1). The wave equation
that describes the propagation of slowly varying envelope A(τ, t) of the pulse
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Saturable Law Nonlinearity 89
along t-direction may be written as

∂A ∂A kωω ∂2 A n2 k |A|2 iα A
i + kω − + + =0 (7.3)
∂t ∂τ 2 ∂τ 2 n0 1 + |A| /Is
2 2
where k is the propagation constant, kω = ∂k/∂ω represents the reciprocal
of group velocity, and kωω = ∂2 k/∂ω2 is the group velocity dispersion at the
carrier frequency ω of the pulse. The parameter α is the attenuation coefficient.
To normalize (7.3), the following transformations are introduced
τ − tkω
x= (7.4)
x0

A = u P0 (7.5)
|kωω (0)|n0
P0 = (7.6)
n2 kx0
x0
LD = (7.7)
|kωω (0)|
P0
s= (7.8)
Is
where x0 is the characteristic time scale. In terms of these normalized variables,
(7.3) reduces to
∂u σ ∂2 u |u|2 u
i + + + iu = 0 (7.9)
∂t 2 ∂x2 1 + s|u|2
√
where = αn0 /2n2 k P0 and σ = ±1 depending on the sign of the parameter
kωω . A positive sign corresponds to the anomalous dispersion regime kωω < 0,
while a negative sign corresponds to the normal dispersion regime kωω > 0.
Finally, introducing the transformation u = Ue −t , one obtains
∂U σ ∂2 U f (t)|U|2 U
i + + =0 (7.10)
∂t 2 ∂x 2 1 + s f (t)|U|2
where f (t) = e −2t . Equation (7.10) is the modified NLSE (MNLSE) due to the
saturable law of nonlinearity. This equation is not integrable by the IST. This
implies that any input beam propagating in this medium cannot be decom-
posed into stable stationary waves and radiation. This fact has a profound
effect on the properties of solitons. Particularly, the basic notion that solitons
are pulses whose properties remain invariant with propagation is no longer
valid. In addition, the contention that solitons undergo elastic collision—that
is, if two solitons collide, there is no loss of energy to radiation fields—is no
longer valid. In saturable media, soliton shape may change periodically or
evolve monotonically and the collision may become inelastic. However, the
essential properties of localized propagation are still applicable in saturable
media and commonly provide a richer and more useful range of propagation
characteristics than Kerr law nonlinearity.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
7.2.1 Conserved Quantities

The MNLSE with saturable law nonlinearity has three integrals of motion,
similar to the other laws of nonlinearity seen in previous chapters. They are
respectively the energy (E), momentum (M), and Hamiltonian (H). For f (t) =
1, the three conserved quantities are as follows
∞
E= |U|2 d x (7.11)
−∞

i ∞ ∗
M= U Ux − UUx∗ d x (7.12)
2 −∞
and
∞ σ
H= |Ux |2 − f ( I ) d x (7.13)
−∞ 2
where
I
z
f (I) = dz (7.14)
0 1 + sz
and the intensity I is given by I = |U|2 .
7.3 Bistable Solitons

Here and henceforth, only the anomalous dispersive media (namely σ = 1)
will be considered. Moreover, in this section, the lossless media will be studied
by setting f (t) = 1. Since (7.10) is not integrable by the IST and no solutions by
the traveling wave ansatz are known, one may search its soliton solution by
numerical technique. Thus, one may seek solutions of the form [261]

U(x, t) = (x)e iβt (7.15)
where β physically represents the nonlinear propagation constant shift. For

a bright soliton, one can impose the boundary conditions
lim (x) = 0 (7.16)

x→±∞
and
d
lim =0 (7.17)
x→±∞ dx
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Using ansatz (7.15) in (7.10), one obtains the following ordinary differential
equation (ODE)
2
1 d 2 1 d
− + −β =0 (7.18)
4 d x 2 8 2 dx 1 + s
This ODE can be solved numerically to obtain the stationary pulse profile for
a given nonlinear propagation constant β. However, arbitrary values of β do
not permit stationary pulses. Depending on the peak power of the soliton,
only certain values of β are permissible. Now (7.18) can be rearranged as
2
1 d d 1
√ + −β =0 (7.19)
8 d dx 1 + s
which, on integration, yields
2
1 d 1 1
√ + [s − ln(1 + s)] − β = 0 (7.20)
8 dx s2
Since for a bright soliton d/d x = 0 at the peak of the soliton amplitude
(namely at = 0 ), one can immediately obtain

1 ln (1 + s0 )
β= 2 s− (7.21)
s 0
Using the above relationship, one can easily determine the permissible values
of β for the corresponding given soliton’s peak amplitude 0 . Figure 7.2(a)
depicts 0 as a function of β.
Soliton energy can be estimated using
∞ ∞
E= |U|2 d x = |0 |2 |g(x)|2 d x (7.22)
−∞ −∞
where g(x) is the soliton shape function. Equation (7.18) can be solved numer-
ically to find out the shape of the soliton, its energy, and its temporal width τ0 .
A typical behavior is shown in Figures 7.2(b) and 7.2(c). An interesting feature
to note is that E is a multivalued function of τ0 , thus admitting bistable [261]
solitons, namely the existence of two solitons having same τ0 but different
energy.
7.4 Arbitrary Pulse Propagation

When a pulse shape is not in the form of a stationary pulse, such as those
admissible by (7.18), it can still propagate in a media without radiating out.
However, depending on the peak power and width of the pulse, when the
pulse propagates, it might breathe or its width might increase monotonically.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
s=0.1
3
β 2
s=0.5
1
0
0 5 10 15
Ψ0
(a)
s=0.7
Ψ 0.5
s=0.1 s=0.4
0
–5 0 5
x
(b)
1.5
1.4
τ0
1.3
1.2
1.1
0 5 10 15
E
(c)
FIGURE 7.2
(a) Variation of β with ψ0 , (b) shape of soliton with different s, and (c) variation of the temporal
width with soliton energy E.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
The behavior of the pulse depends on the relative magnitude of nonlinearity

induced self-phase modulation and fiber dispersion. When the pulse propa-
gates in a medium, it acquires a chirp. The anomalous dispersion produces a
chirp of opposite sign. The propagation characteristics of the pulse result as
an interplay between these two opposite chirps.
In order to investigate the dynamic behavior, the variational principle that
was introduced in chapter 3 will be exploited here. The Lagrangian can be
written as

∂U ∗ ∗ ∂U
∂U 2 2
L =i U −U
+ − |U|2 + 2 ln 1 + s f (t)|U|2
∂t ∂t ∂x s 2
s f (t)
(7.23)
By Hamilton’s principle of least action, one can write

δ Ldt = 0 (7.24)
In general, the temporal profile of an optical pulse emitted by a present-day

semiconductor laser is very close to sech or Gaussian. Thus, it is reasonable
that one would be curious to know the behavior of such pulses in saturating
nonlinear media. Therefore, for further investigation, one can choose the trial
function
B(t)
e iφ(t)x +ψ(t)
2
U(x, t) = (7.25)
x
cosh a (t)
where B(t), a (t), φ(t), and ψ(t) are the amplitude, duration, chirp, and longi-
tudinal phase of the pulse, respectively. By using the trial function (7.25), the
averaged Lagrangian is given by

dψ 2B 2 π2 dφ 4B 2 a
L = 4B 2 a + + 4φ 2 + 2 B 2a 3 −
dt 3a 6 dt s
∞
2a s f (t) B 2
+ ln 1 + dy (7.26)
f (t)s 2 −∞ cosh2 y
Employing the Raleigh-Ritz optimization procedure, one arrives at the fol-

lowing set of equations
B 2 (t)a (t) = B 2 (0)a (0) = consta nt = N2 (7.27)

1 da
φ(t) = (7.28)
2a (t) dt
d 2a 4 6 d
= 2 3− 2 2 {a G(a )} (7.29)
dt 2 π a π N f (t)s da
2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and

dψ 1 a d 1 a ∂G
=− 2 + {a G(a )} + + a −G
dt 3a 2 2
2N s f (t) da s 4N s 2 f (t)
2 ∂a
(7.30)
where
√
2
s f (t)
G(a ) = 2 sinh−1 √ N (7.31)
a
Equations (7.27)–(7.30) can be used to investigate the propagation of a sech

optical pulse in lossless and lossy medium with saturating nonlinearity.
7.4.1 Lossless Uniform Media (Γ = 0)

In this section, the propagation characteristics of an optical pulse in a lossless
medium are investigated. For such a medium, f (t) = 1. So (7.29) reduces to
d 2a 4 6 d
= 2 3− 2 2 2 {a G(a )} (7.32)
dt 2 π a π N s da
which can be integrated at once to give
2
1 da
+ V(a ) = 0 (7.33)
2 dt
where

2 1 1 6 a 02 C02
V(a ) = − 2 + [a G(a ) − a 0 G(a 0 )] − (7.34)
π2 a2 a0 π 2 N2 s 2 2
with a 0 = a (t = 0) and C0 = (da/dt) t=0 , which can be identified as initial

chirp at t = 0. The left side of (7.33) can be identified as total energy and
is identical to that of an oscillator of unit mass that is executing its motion
under a potential V(a ) and, hence, (7.32) is derivable from the gradient of
V(a ), namely
d 2a ∂ V(a )
=− (7.35)
dt 2 ∂a
Although the analytical solution of (7.35) is not obvious, a lot of valuable

information about a propagating pulse can be extracted by examining the
potential V(a ). √ √
For small saturation parameter s → 0, and for small N, N s/ a 1;
therefore
√
2
−1 s N2 s 8 N6 s 3 N4 s 2
sinh √ N ≈ + − (7.36)
a a 45 a 3 3a 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and the potential reduces to
2 4N2 32 N4 s 12
V(a ) ≈ − + + 2 (7.37)
π 2a 2 π 2a 15 π 2 a 2 π s
Using (7.32), one recovers the governing equation of a sech soliton propagat-
ing through a parabolic nonlinear medium
d 2a 4 4N2 64 N4 s
= − + (7.38)
dt 2 π 2a 3 π 2a 15 π 2 a 3
and, on setting s = 0 in (7.38), one recovers the result of a sech soliton propa-
gating through a Kerr nonlinear medium.
The potential V(a ) given by (7.34) can be examined for saturating nonlin-
earity. It is clear from (7.34) that V(a ) ≤ 0 is the allowed propagation region.
Thus, for a given initial chirp C0 , the value of N for which bound motion is
possible can be obtained from the potential V(a ). For very large a , if V(a )
remains negative, unbounded motion occurs. Physically, this corresponds to
a situation in which the temporal width of the soliton is very large—in other
words, the pulse loses its solitonic character. Thus, unbounded motion results
in the region where the following condition holds
√
2
s2 6a 0 −1 N s π 2 C02 s 2 a 02
F ( N) = 6s − 2 − 2 sinh √ − <0 (7.39)
a0 N a0 4
For further discussion, one can assume a 0 = 1, without any loss of generality.
For initial chirpless pulse, C0 = 0, the variation of F ( N) with N is shown in
Figure 7.3(a) for different values of the saturation parameter s. F ( N) < 0 is the
region of unbounded motion and F ( N) > 0 is the region of bounded motion.
The threshold value of N (Nth ) above which the bound motion is expected
can be estimated by noting down the value of N for which F ( N) = 0. From
Figure 7.3(a) it is evident that Nth increases with increases in the value of the
saturation parameter s. In Figure 7.3(b), F ( N) is plotted against N for different
chirp. It is evident from this figure that Nth increases with increases in the value
of the initial chirp C0 . Thus, one can conclude that indefinite increases in the
value of the chirp may lead to destruction of the soliton of given energy N2 .
For example, for C0 = 0, N = 1 gives a stable soliton propagation. However,
when C0 = 0.9, F ( N) < 0 for N < 1.2. Thus the soliton is unstable and would
disintegrate.
For a chirpless pulse, variation of V(a ) against a for different N and s = 0.3
is shown in Figure (7.4). One can identify four distinctly different types of
behaviors of the pulse width for different values of the parameter N. These are
1. Stationary propagation with a = 1 for N = 1.226

2. Oscillatory bounded motion with 0 < a < 1 for N > 1.226
3. Oscillatory bounded motion with 0.73 < N < 1.226
4. Unbounded motion with a → ∞ for N < 0.73.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
0.04
s=0.6 s=0.45 s=0.3
0.02
s=0.15
F (N)
–0.02
0.6 0.7 0.8 0.9 1
N
(a)
1
C0=0
C0=0.3
C0=0.6
C0=0.9
0.5 C0=1.2
F (N)
–0.5
0 1 2 3 4 5
N
(b)
FIGURE 7.3
(a) Variation of F with N for different s, (b) variation of F with N for different initial chirp.
7.4.2 Stationary Pulse Propagation

For a chirpless, stationary pulse, the condition can be obtained by using
∂ V/∂a = 0. Thus, the value of N for which a stationary pulse is obtained
can be found from the relation
⎡ √ ⎤
√ 2 √ sinh−1 N√ s
⎢ N s N s a ⎥
N2 s 2 − 3a 3 ⎣a sinh−1 √ − √ 12 ⎦ = 0
a a 2
1 + Na s
(7.40)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
0.1
N=1.5 N=1.226
N=1
V (a) 0
–0.1
N=0.5
–0.2
0 1 2 3 4 5
a
FIGURE 7.4
Variation of potential V with a for different N.
This equation can be solved numerically to find out the value of N for a given
a for stationary propagation. Figure (7.5) depicts the variation of a with N
for stationary pulse propagation. The existence of two-state soliton is clearly
evident from the figure. The stability of these solitons can be found from the
sign of ∂2 V/∂a 2 . It was verified that ∂2 V/∂a 2 > 0, so that these stationary states
are stable.
0)
7.4.3 Lossy Media (Γ =
For a lossy medium, = 0 so that f (t) =
1. Since the right side of (7.29)
depends on t explicitly, the potential formalism that was developed in the
previous section cannot be used here. Particularly, d 2 a /dt 2 is not derivable
2 s=0.60
a
s=0.45
s=0.30
1
s=0.15
0
0 5 10 15 20
N
FIGURE 7.5
Variation of soliton width with N for different s.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
s=0.15
1.2 s=0.3
s=0.45
1.1
0.9
a
0.8
0.7
0.6
0.5
0 5 10 15 20 25
t
FIGURE 7.6
Variation of beam width with distance of propagation.
from the gradient of a scalar potential. Therefore, one can solve (7.29) numer-
ically to study the influence of dissipation. Equation (7.29) is solved numeri-
cally using a chirpless input pulse of width a 0 = 1 and normalized dissipation
= 0.001. For a typical value of N = 1.5, variation of pulse width is displayed
in Figure 7.6. It is evident from the figure that, as the pulse propagates, its
temporal width oscillates. As a consequence of energy loss due to dissipa-
tion, net broadening of width takes place over many cycles of oscillations.
The net broadening decreases with the increase in the value of s. One can
conclude from Figure 7.6 that the pulse is more robust under dissipation loss
with increasing value of saturation.
Exercises
1. Consider the Lagrangian that is given by (7.23). Now, use the

relationship
δL δL
= =0
δU ∗ δU
to recover the NLSE that is given by (7.10).

2. Show that in (7.29) G(a ) is given by
∞
s N2 f (t)
G(a ) = ln 1 + dy
−∞ a cosh2 y
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
which reduces to
√
2
s f (t)
G(a ) = 2 sinh−1 √ N
a
that is given by (7.31).
3. Set up a coupled NLSE for two incoherently coupled solitons in
saturating nonlinear media. Show that the Lagrangian for such a
system is given by
2
2
∂ψ ∗j ∂ψ ∂ψ
+ − 2 ψ j 2
− ψ∗
j
L= i ψj
∂t ∂t ∂x
j=1
⎛ ⎞
2
2
+ 2 ln ⎝1 + ψ j ⎠
j=1
4. Use Euler’s equation to derive (7.29) and (7.30).

5. Recall that the NLSE for a parabolic nonlinear media is given by
∂ψ 1 ∂2 ψ 2
i + + |ψ| + ν |ψ|4 ψ = 0
∂t 2 ∂x 2
Consider the input pulse that is given by

A(t)
ψ(x, t) =
cosh Bx
Now, prove that the width B(t) will vary according to
d2 B 4 4N2 64ν N4
= 2 2− 2 +
dt 2 π B π B 15π 2 B 3
where N2 = A2 B = constant.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
8
Soliton–Soliton Interaction
The intrachannel collision of optical solitons in the presence of perturbation

terms is studied in this chapter by the aid of quasi-particle theory (QPT) that
will be developed from basic principles. The nonlinearities that are studied
in this chapter are the Kerr, power, parabolic, and dual-power laws. The per-
turbation terms that are considered in this chapter are both Hamiltonian and
non-Hamiltonian type.
8.1 Introduction
In a soliton communication system, it is necessary to launch the solitons close
to each other to enhance the information-carrying capacity of the fiber. Two
solitons that are placed too close to each other can mutually interact, thus pro-
viding a very serious hindrance to the performance of the soliton transmis-
sion system. However, the presence of the perturbation terms of the nonlinear
Schrödinger’s equation (NLSE) can lead to the suppression of the two-soliton
interaction, thus solving this problem [62, 185].
A considerable amount of research effort has been spent on the reduction
of interaction of solitons. For example, the use of Gaussian-shaped pulses
reduces the interaction because of the steep slope, but this is achieved at
the expense of creating larger oscillatory tails. It has also been shown that
introducing a phase difference between neighboring solitons can lead to a
reduction in their interaction. Incoherent interaction of solitons has also been
analyzed. The third-order dispersion of an optical fiber can also be used to
reduce mutual interactions, but this results in the breakup of the bound state
of solitons. A more realistic way to reduce these interactions is to launch
adjacent pulses with unequal amplitudes. In this case, solitons form a bound
system and effectively maintain their initial pulse separation. The higher order
nonlinear effect is also shown to break up the bound state of the solitons. A
detailed study of soliton–soliton interaction in the presence of fiber loss and
periodic amplification has also been carried out [185].
101
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
In this chapter, the QPT will be introduced and its use as a mechanism
for suppressing the interaction of solitons due to the presence of perturbation
terms will be studied. The QPT will be formulated for Kerr law, power law,
parabolic law, and dual-power law. Numerical simulations will also support
the theory.
8.2 Mathematical Formulation

To study the intrachannel collision of optical solitons, the NLSE with the
following perturbation terms is considered
1
iq t + q xx + F (|q |2 )q = i R[q , q ∗ ] (8.1)
2
where
R = δ|q |2m q − αq x + βq xx − γ q xxx

x
+ σq |q |2 dξ + λ(|q |2 q ) x + µ(|q |2 ) x q − iρq xxxx (8.2)
−∞
The coefficient of β is called the bandpass filtering term. Also, in (8.2), α is the
frequency separation between the soliton carrier and the frequency at the peak
of EDFA gain. Moreover, λ is the self-steepening coefficient for short pulses
(typically ≤ 100 femto seconds), µ is the nonlinear dispersion coefficient,
and γ is the coefficient of the third-order dispersion [241], while ρ represents
the coefficient of fourth-order dispersion [64]. It needs to be noted that the
coefficients δ, β, and σ represent non-Hamiltonian perturbations, while the
remaining terms in (8.2) represent Hamiltonian-type perturbations [86].
The soliton solution of (8.1) for = 0, although not integrable, is assumed
to be given in the form
q (x, t) = η(t)g[ζ (t)(x − vt − x0 )]e (−iκ x+iωt+iσ0 ) (8.3)
where
κ = −v (8.4)
ζ (t) = χ (η(t)) (8.5)
and
ω(t) = ψ(η(t), κ(t)) (8.6)
In (8.5), g represents the shape of the soliton described by the NLSE and
depends on the type of nonlinearity in (8.1). The parameters η(t) and ζ (t) in
(8.5) respectively represent the soliton amplitude and width, while κ(t) and
ω(t) are the frequency and wave number of the soliton, respectively, and v
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Soliton–Soliton Interaction 103
is the velocity. The explicit functional form of ψ in (8.6) is actually given in

(2.39). Also, x0 and σ0 respectively represent the center of the soliton and the
center of the soliton phase. In (8.5) and (8.6), the functional forms χ and ψ
depend on the type of nonlinearity in (8.1).
The 2-soliton solution of the NLSE (8.1) takes the asymptotic form

2
q (x, t) = ηl (t)g [ζl (t) (x − vl t − xl )] e (−iκl x+iωl t+iσl ) (8.7)
l=1
with
κl = −vl (8.8)
ζl (t) = χ (ηl (t)) (8.9)
ωl (t) = ψ(ηl (t), κl (t)) (8.10)
where l = 1, 2. In the study of soliton–soliton interaction (SSI), the initial

pulse waveform is taken to be of the form
x0 iφ1 x0 iφ2
q (x, 0) = η1 g ζ1 x − e + η2 g ζ2 x + e (8.11)
2 2
which represents the injection of 2-soliton-like pulses into a fiber. Here, x0
represents the initial separation of the solitons, namely the center-to-center
soliton separation. It is to be noted that for x0 → ∞, (8.11) represents exact
soliton solutions, but for x0 ∼ O(1), it does not represent an exact 2-soliton
solution. The initial pulse form is modified depending on the type of pertur-
bation considered, as seen below.
1. Non-Hamiltonian Perturbations: For studying the SSI with non-
Hamiltonian perturbations, the case of in-phase injection of soli-
tons with equal amplitudes will be considered. So, without any loss
of generality, η1 = η2 = 1 and φ1 = φ2 = 0 are chosen so that (8.11)
modifies to
x0 x0
q (x, 0) = g ζ1 x − + g ζ2 x + (8.12)
2 2
2. Hamiltonian Perturbations: For studying SSI with Hamiltonian per-
turbations, the case of in-phase injection of solitons with unequal
amplitudes will be considered. So, without any loss of generality,
η1 = η0 , η2 = 1 and φ1 = φ2 = 0 are chosen so that (8.11) modifies to
x0 x0
q (0, T) = η0 g ζ0 x − +g ζ x+ (8.13)
2 2
where
ζ0 = χ (η0 ) (8.14)
and ζ is given by (8.5).

P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
The special cases with regards to the four laws of nonlinearity will now be
individually discussed in the following four subsections.
8.2.1 Kerr Law

For the case of Kerr law nonlinearity, F (s) = s so that (8.1) becomes
1
iq t + q xx + |q |2 q = i R[q , q ∗ ] (8.15)
2
The 1-soliton solution of (8.15) for = 0 that can be obtained by the inverse
scattering transform (IST) has the form
η
q (x, t) = e i(−κ x+ωt+σ0 ) (8.16)
cosh [ζ (x − vt − x0 )]
where
ζ ≡ χ (η) = η (8.17)
and
η2 − κ 2
ω ≡ ψ(η, κ) = (8.18)
2
Also, the 2-soliton solution of the NLSE (8.1) takes the asymptotic form

2
ηl
q (x, t) = e i ( −κl x+ωl t+σ0l ) (8.19)
l=1
cosh ηl (x − vl t − x0l )
where
ζl ≡ χ (ηl ) = ηl (8.20)
and
ηl2 − κl2
ωl ≡ ψ(ηl , κl ) = (8.21)
2
and l = 1, 2. In the study of SSI with non-Hamiltonian perturbations, the
initial pulse form is taken to be
η η
q (x, 0) = 1
e iφ1 + 2
e iφ2 (8.22)
cosh η1 x − x0
2
cosh η2 x + x0
2
For non-Hamiltonian perturbations, the choice η1 = η2 = 1 and φ1 = φ2 = 0

gives
1 1
q (x, 0) =
+
(8.23)
cosh x− x0
2
cosh x + x0
2
which represents an in-phase injection of pulses with equal amplitudes.

P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
For Hamiltonian-type perturbations, the choice η1 = η0 , η2 = 1, and φ1 =

φ2 = 0 gives
η 1
q (x, 0) = 0
+
(8.24)
cosh η0 x − x0
2
cosh x + x0
2
which represents an in-phase injection of pulses with unequal amplitudes.
8.2.2 Power Law

For the case of power law nonlinearity, the NLSE is given by
1
iq t + q xx + |q |2 p q = i R[q , q ∗ ] (8.25)
2
In this case, (8.25) with = 0 is not integrable, as seen before, by the IST.
However, (8.25) for = 0 supports solitons of the form
η
q (x, t) = 1 e i(−κ x+ωt+σ0 ) (8.26)
cosh [ζ (x − vt − x0 )]
p
where
12
2 p2
ζ ≡ χ(η) = η p
(8.27)
1+ p
and
ζ2 κ2
ω ≡ ψ(η, κ) = − (8.28)
2 p2 2
The 2-soliton solution of the NLSE takes the asymptotic form

2
ηl i ( −κl x+ωl t+σ0l )
q (x, t) = 1 e (8.29)
l=1 cosh ζl (x − vl t − x0l )
p
where
12
p 2 p2
ζl ≡ χ (ηl ) = ηl (8.30)
1+ p
and
ζl2 κ2
ωl ≡ ψ(ηl , κl ) = 2
− l (8.31)
2p 2
In the study of SSI for power law, the initial pulse waveform is assumed to be
η1 η2
q (x, 0) =
1
e
iφ1
+
1
e
iφ2
(8.32)
cosh ζ1 x −
p x0
2
cosh ζ2 x +
p x0
2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book

gives
1 1
q (x, 0) =
1
+
1
(8.33)
cosh ζ x −
p x0
2
cosh ζ x +
p x0
2
where

2 p2
ζ = (8.34)
1+ p
which represents an in-phase injection of pulses with equal amplitudes.

φ2 = 0 gives
η0 1
q (x, 0) =
1
+
1
(8.35)
cosh ζ0 x −
p x0
2
cosh ζ x +
p x0
2
where
12
p 2 p2
ζ0 = η0 (8.36)
1+ p
which represents in-phase injection of solitons with unequal amplitudes.
8.2.3 Parabolic Law

For parabolic law nonlinearity, F (s) = s + νs 2 , so that the NLSE is
1
iq t + q xx + (|q |2 + ν|q |4 )q = i R[q , q ∗ ] (8.37)
2
Equation (8.37) is not integrable by the IST for = 0, as discussed in chapter 5.
However, (8.37) for = 0 supports solitons of the form
η
q (x, t) = 1 e i(−κ x+ωt+σ0 ) (8.38)
[1 + a cosh{ζ (x − vt − x0 )}]2
where
√
ζ ≡ χ (η) = η 2 (8.39)
η2
κ 2
ω ≡ ψ(η, κ) = − (8.40)
4 2
and

4
a= 1 + νη2 (8.41)
3
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
In this chapter, it will be assumed that ν > 0, although ν could be negative as

well, as seen in chapter 5. In fact, recall from chapter 5 that for = 0, solitons
exist for ν ∈ (−3/4A2 , ∞). Now, the 2-soliton solution of the parabolic law
takes the asymptotic form

2
ηl
q (x, t) = 1 e i ( −κl x+ωl t+σ0l ) (8.42)
l=1 [1 + a l cosh {ζl (x − vt − xl )}] 2
with
√
ζl ≡ χ(ηl ) = ηl 2 (8.43)
ηl2 − 2κl2
ωl ≡ ψ(ηl , κl ) = (8.44)
4
and

4
al = 1 + νηl2 (8.45)
3
In the study of SSI with parabolic law nonlinearity, the initial pulse waveform
is taken to be of the form
η1 η2
q (x, 0) =
12 e +
iφ1

12 e
iφ2
1 + a 1 cosh ζ1 x − 2
x0
1 + a 2 cosh ζ2 x + x20
(8.46)

gives
1
q (x, 0) = √
12
1+ 1 + 43 ν cosh 2 x− x0
2
1
+ √ (8.47)

12
1+ 1+ 4
3
ν cosh 2 x+ x0
2

φ2 = 0 gives
η0
q (x, 0) = √
12
1+ 1+ 4
3
νη0 cosh η0 2 x − x0
2
1
+ √ (8.48)

12
1+ 1+ 4
3
ν cosh 2 x+ x0
2
which represents an in-phase injection of pulses with unequal amplitudes.

P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
8.2.4 Dual-Power Law

For dual-power law nonlinearity, F (s) = s p + νs 2 p where ν < 0, so that the
NLSE is
1
iq t + q tt + |q |2 p + ν|q |4 p q = i R[q , q ∗ ] (8.49)

2
Equation (8.49) for = 0, although not integrable by the IST, supports solitary
waves of the form
η
q (x, t) = 1 e i(−κ x+ωt+σ0 ) (8.50)
[1 + a cosh {ζ (x − vt − x0 )}] 2p
where
21p
4 p2
ζ ≡ χ (η) = η p
(8.51)
1+ p
η2 p κ2
ω ≡ ψ(η, κ) = − (8.52)
2p + 2 2
and

νζ 2 (1 + p) 2
a= 1+ (8.53)
2 p2 1 + 2 p
For dual-power law nonlinearity, solitons exist for
2 p2 1 + 2 p
− <ν<0 (8.54)
ζ 2 (1 + p) 2
In this case, the 2-soliton solution of the NLSE (1) takes the asymptotic form

2
ηl
q (x, t) = 1 e i ( −κl x+ωl t+σ0l ) (8.55)
l=1 [1 + a l cosh {ζl (x − vt − xl )}] 2p
with
21p
p 4 p2
ζl ≡ χ (ηl ) = ηl (8.56)
1+ p
2p
ηl κ2
ωl ≡ ψ(ηl , κl ) = − l (8.57)
2p + 2 2
and

νζl2 (1 + p) 2
al = 1+ (8.58)
2 p2 1 + 2 p
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
In the study of SSI with dual-power law nonlinearity, the initial pulse wave-
form is taken to be of the form
η1
q (x, 0) =
21p e
iφ1
1 + a 1 cosh ζ1 x − x20
η2
+
21p e
iφ2
(8.59)
1 + a 2 cosh ζ2 x + 2 x0
For non-Hamiltonian perturbations, the choice η1 = η2 = 1 and φ1 = φ2 = 0,

gives
1 1
q (x, 0) =
1 +
21p
1 + a 1 cosh ζ x − x0
2
2p
1 + a 2 cosh ζ x + x0
2
(8.60)
where ζ is given by (8.34). For Hamiltonian-type perturbations, the choice

η1 = η0 , η2 = 1, and φ1 = φ2 = 0 gives
η0 1
q (x, 0) =
1 +
21p
1 + a 1 cosh ζ0 x − x0
2
2p
1 + a 2 cosh ζ x + x0
2
(8.61)
where

p 4 p2
ζ0 = η0 (8.62)
1+ p
which represents in-phase injection of pulses with unequal amplitudes.
8.3 Quasi-Particle Theory

The QPT dates back to the 1981 appearance of the paper by Karpman and
Solov’ev [213]. The mathematical approach to SSI will be studied using the
QPT. In this theory, solitons are treated as particles. If two pulses are separated
and each is close to a soliton, they can be written as the linear superposition
of two soliton-like pulses as [185]
q (x, t) = q 1 (x, t) + q 2 (x, t) (8.63)
with
ql (x, t) = Al g[Dl (x − xl )]e −i{Bl (x−xl )−δl } (8.64)
where
Dl = χ ( Al ) (8.65)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and l = 1, 2, while Al , Bl , Dl , Tl , and δl are functions of t. Here, Al , Dl , and Bl do

not represent the amplitude, width, and the frequency of the full wave form.
However, they approach the amplitude, width, and frequency, respectively,
for large separation, namely if x = x1 − x2 → ∞, then Al → ηl , Dl → ζl ,
and Bl → κl . Since the waveform is assumed to remain in the form of two
pulses, the method is called the quasi-particle approach.
First, the equations for Al , Bl , Tl , and δl using the soliton perturbation
theory (SPT) will be derived. Substituting (8.63) into (8.1) yields
∂q l 1 ∂2 ql
i + = i R[ql , ql∗ ] − F (|ql + ql̄ |2 )|ql + ql̄ | (8.66)
∂t 2 ∂x2
where l = 1, 2 and l̄ = 3 − l. By the SPT, the evolution equations are
d Al
= F1(l) ( A, x, φ) + Ml (8.67)
dt
d Bl
= F2(l) ( A, x, φ) + Nl (8.68)
dt
d xl
= −Bl − F3 ( A, x, φ) + Ql (8.69)
dt
dδl
= ψ( Al , Bl ) + F4 ( A, x, φ) + Pl (8.70)
dt
where
∞
Ml = h 1 ( Al ) R̂[ql , ql∗ ]e −iφl g(τl )dτl (8.71)
−∞
∞
Nl = h 2 ( Al )
R̂[ql , ql∗ ]e −iφl g (τl )dτl (8.72)
−∞
∞
Ql = h 3 ( Al ) R̂[ql , ql∗ ]e −iφl τl g(τl )dτl (8.73)
−∞
∞
Pl = h 4 ( Al )
R̂[ql , ql∗ ]e −iφl [g(τl ) − τl g (τl )]dτl (8.74)
−∞
and the functions F1(l) , F2(l) , F3 , and F4 evolve on using the SPT in (8.66), with
the right side being treated as perturbation terms. The exact form of these
functions can be obtained when a specific law of nonlinearity is considered.
In (8.71)–(8.74), h j ( Al ) for 1 ≤ j ≤ 4 are by virtue of (8.65) and the type of
nonlinearity that is considered. Also, in (8.71)–(8.74), and
stand for the
real and imaginary parts, respectively. Moreover, the following notations are
used [185]
R̂[ql , ql∗ ] = R[ql , ql∗ ] − F (|ql + ql̄ |2 )|ql + ql̄ | (8.75)

τl = Dl (x − xl ) (8.76)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
which represent the argument of the functional form of the lth soliton for
various laws of nonlinearity, while the phase of the lth soliton in a fiber is
given by
φl = Bl (x − xl ) − δl (8.77)
The difference between the phases of the two solitons, namely φ1 −φ2 , is given
by
φ = Bx + δ (8.78)
while the separation of the solitons that is given by the distance between the
centers of the solitons is represented as
x = x1 − x2 (8.79)
δ = δ1 − δ2 (8.80)
The mean values of the amplitude and width of the two solitons are respec-
tively defined as
1
A= ( A1 + A2 ) (8.81)
2
and
1
B= ( B1 + B2 ) (8.82)
2
Now, the difference between the amplitudes and widths of the two solitons
are respectively given as
A = A1 − A2 (8.83)
and
B = B1 − B2 (8.84)
Finally, it is assumed that
|A| A (8.85)
|B| 1 (8.86)
|D| D (8.87)
Ax 1 (8.88)
Dx 1 (8.89)
|A|x 1 (8.90)
and
|D|x 1 (8.91)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
From (8.67) to (8.70), one can now obtain
dA
= M (8.92)
dt
dB
= N (8.93)
dt
d(A)
= F1(1) ( A, t, φ) − F1(2) ( A, x, φ) + M (8.94)
dt
d(B)
= F2(1) ( A, x, φ) − F2(2) ( A, x, φ) + N (8.95)
dt
d(T)
= −B + Q (8.96)
dt
d(φ)
= ψ( A1 , B1 ) − ψ( A2 , B2 ) − BB
dt
x (1)
+ F2 + F2(2) + P + BQ (8.97)

2
where
1
M= ( M1 + M2 ) (8.98)
2
1
N = ( N1 + N2 ) (8.99)
2
and M, N, Q, and P are the variations of M, N, Q, and P, which are
written as, for example
∂M ∂M
M = A + B (8.100)
∂A ∂B
assuming that they are functions of A and B only, which is, in fact, true for
most of the cases of interest; otherwise, the equations for
1
x= (x1 + x2 ) (8.101)
2
and
1
φ= (φ1 + φ2 ) (8.102)
2
would have been necessary. The results derived in this section will now be
utilized to show that SSI can indeed be suppressed in the presence of the
perturbation terms given by (8.2) for the four cases of nonlinearity. In all
four types of nonlinearity, corresponding to the initial waveform (8.13) are
A = 1, B = 0, A0 = 0, B0 = 0, T0 = T0 , and φ0 = 0. The four laws of
nonlinearity are studied in the following subsections.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
8.3.1 Kerr Law

In this case, (8.64) reduces to
Al
ql (x, t) = e −i Bl (x−xl )+iδl (8.103)
cosh[Dl (x − xl )]
where
Dl ≡ χ ( Al ) = Al (8.104)
so that (8.66) transforms to
∂q l 1 ∂2 ql
i + + |ql |2 ql = i R[ql , ql∗ ] − (ql2 ql̄∗ + 2|ql |2 ql̄ ) (8.105)
∂t 2 ∂x2
where l = 1, 2 and l̄ = 3 − l, and the separation
|q |2 q = (|q 1 |2 q 1 + q 12 q 2∗ + 2|q 1 |2 q 2 ) + (|q 2 |2 q 2 + q 22 q 1∗ + 2|q 2 |2 q 1 ) (8.106)
was used based on the degree of overlapping. By the SPT, the evolution equa-
tions are
d Al
= (−1)l+1 4A3 e −Ax sin(φ) + Ml (8.107)
dt
d Bl
= (−1)l+1 4A3 e −Ax cos(φ) + Nl (8.108)
dx
d xl
= −Bl − 2Ae −Ax sin(φ) + Ql (8.109)
dt
and
dδl 1
= ( A2l + Bl2 ) − 2ABe −Ax sin(φ) + 6A2 e −Ax cos(φ) + Pl (8.110)
dt 2
where
∞
1
Ml = { R̂[ql , ql∗ ]e −iφl } dτl (8.111)
−∞ cosh τl
∞
tanh τl
Nl = −
{ R̂[ql , ql∗ ]e −iφl } dτl (8.112)
−∞ cosh τl
∞
1 τl
Ql = { R̂[ql , ql∗ ]e −iφl } dτl (8.113)
A2l −∞ cosh τl
∞
1 (1 − τl tanh τl )
Pl =
{ R̂[ql , ql∗ ]e −iφl } dτl (8.114)
Al −∞ cosh τl
and
R̂[ql , ql∗ ] = R[ql , ql∗ ] − i(ql2 ql̄∗ + 2|ql |2 ql̄∗ ) (8.115)

P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
The study for the Kerr law case will now be split into the following two
subsections.
8.3.1.1 Non-Hamiltonian Perturbations

In the presence of non-Hamiltonian perturbation terms, as discussed after
(8.2), the dynamical system of the soliton parameters by virtue of the SPT are

dA 12 (m + 1) 2m+1 2
= δ
A + 2σ A2 − β A( A2 + 3B 2 ) (8.116)
dt 2m+3
2
3
dB 4
= − β A2 B (8.117)
dt 3
so that by virtue of (8.83), (8.84), (8.79), and (8.80)
d(A)
= 8A3 e −Ax sin(φ)
dt

2m+1
δ 12 (m + 1) 2m + 1
+ 2m 2m+3
(2A) 2r +1 (A) 2m−2r
2 2 r =0
2r + 1
2
+ 4σ AA − 2β A + B 2 A − 4β ABB (8.118)

d(B) 8 4
= 8A3 e −Ax cos(φ) − β ABA − β A2 B (8.119)
dt 3 3
d(x)
= −B + σ (8.120)
dt
d(φ) 4
= AA − β A2 Bx (8.121)
dt 3
where in (8.118)

n n(n − 1) . . . (n − r + 1)
= (8.122)
r r (r − 1) . . . 3.2.1
For the fixed point of the dynamical system, given by (8.116) and (8.117), with
A = 1 and B = 0, one recovers

3δ 12 (m + 1)
β = 3σ +
(8.123)
2 2m+3
2
From (8.120) and (8.121), one has the coupled system of equations for φ, the
phase difference, x, and the soliton separation, with the fixed point A = 1
and B = 0 as follows
d 2 (x) 4 d(x)
2
+ β + 8e −x cos(φ) = 0 (8.124)
dt 3 dt
d 2 (φ) d(φ)
+ 2(β − 2σ )
dt 2 dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book

2m+1
δ 12 (m + 1) 2m + 1 d(φ) 2m−2r
− 2m 2m+3
2 2 r =0
2r + 1 dt
− 8e −x sin(φ) = 0 (8.125)
where in (8.124) and (8.125) β is given by (8.123). Equations (8.124) and (8.125)
show that inserting filters produces a damping in both pulse separation and
phase difference, as seen in Figures 8.1(a)–8.1(c).
8.3.1.2 Hamiltonian Perturbations

In the case of Hamiltonian perturbations, the dynamical system of the soliton
parameters is
dA
=0 (8.126)
dt
dB
=0 (8.127)
dt
and
d x0
= −B − 3γ B 2 − {A2 (3λ + 2µ + 3γ ) + 3α} (8.128)
dt 3
d(A)
= 8A3 e −Ax sin(φ) (8.129)
dt
d(B)
= 8A3 e −Ax cos(φ) (8.130)
dt
d(x) 3
= −B − γ BB − (3λ + 2µ + 3γ ) AA (8.131)
dt 2 6
and
d(φ)
= AA (8.132)
dt
Now
1
A= ( A0 + 1) (8.133)
2
B=0 (8.134)
A0 = A0 − 1 (8.135)
B0 = 0 (8.136)
T0 = T0 (8.137)
φ0 = 0 (8.138)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
60
50
40
30
t
20
10
0
–30 –20 –10 0 10 20 30
x
(a)
60
50
40
30
t
20
10
0
–30 –20 –10 0 10 20 30
x
(b)
FIGURE 8.1
(a) SSI for m = 0, σ = δ = 0.005, (b) SSI for m = 1, σ = δ = 0.005.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
60
50
40
30
t
20
10
0
–30 –20 –10 0 10 20 30
x
(c)
FIGURE 8.1 (Continued)

(c) SSI for m = 2, σ = δ = 0.005.
and
φ = δ (8.139)
so that
d(x) d(φ)
+ (3λ + 2µ + 3γ ) = −B (8.140)
dt 6 dt
For B = 0

x = x0 − (3λ + 2µ + 3γ )δ (8.141)
6
Since, x0 ∼ O(1), x → 0 and thus the pulses do not collide during the
transmission. This is observed in the numerical simulation in Figure 8.2.
8.3.2 Power Law

In this case, (8.64) is
Al
ql (x, t) = 1 e −i Bl (x−xl )+iδl (8.142)
cosh [Dl (x − xl )]
p
where
12
p 2 p2
Dl ≡ χ ( Al ) = Al (8.143)
1+ p
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
50
45
40
35
30
25
t
20
15
10
0
–25 –20 –15 –10 –5 0 5 10 15 20
x
FIGURE 8.2
SSI with γ = 0.14.
so that using (8.63) in (8.1) yields

p
∂q l 1 ∂2 ql p p−r
∗
i + = i R[ql , ql ] − q q2 r
∂t 2 ∂x2 r =0
r 1
p
p
p−r
r
∗ ∗
q1 q2 (q 1 + q 2 ) (8.144)
r =0
r
Here, the separation

p p
p p−r p
p−r
r
∗ ∗
|q | q =
2p r
q q2 q1 q2 (q 1 + q 2 ) (8.145)
r =0
r 1 r =0
r
was used based on the degree of overlapping. By the SPT, the evolution equa-
tions are
d Al
= F1(l) ( A, x, φ; p) + Ml (8.146)
dt
d Bl
= F2(l) ( A, x, φ; p) + Nl (8.147)
dt
dTl
= −Bl − F3 ( A, x, φ; p) + Ql (8.148)
dt
and
2p
dδl B2 Al
= l + + F4 ( A, x, φ; p) + Pl (8.149)
dt 2 p+1
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where, for power law

p−3 1p + 12 ∞
1 2 p2 2p 1
Ml =
R̂[ql , ql∗ ]e −iφl dτl
2− p 1+ p 1p 12 −∞
1
cosh p τl
(8.150)

p−1
1p + 12 ∞
2 p−1 2 p2 2p tanh τl
Nl = A

R̂[ql , ql∗ ]e −iφl dτl
p l 1+ p 1p 12 −∞
1
cosh p τl
(8.151)

p+2
1p + 12 ∞
1 p+1 2p τl
Ql = p+1

R̂[ql , ql∗ ]e −iφl 1 dτl
Al 2 p2 1p 12 −∞ cosh p τl
(8.152)
and

p+1
2 1
+ 1
1 2p 2p
p 2
Pl =
p+1
Al p+1 p 12
1

∞ (1 − τl tanh τl )
.
R̂[ql , ql∗ ]e −iφl 1 dτl (8.153)
−∞ cosh p τl
In addition, the following notations are used
p p
p p−r p
p−r
r
∗ ∗ ∗ ∗
R̂[ql , ql ] = R[ql , ql ] − i q q2 r
q1 q2
r =0
r 1 r =0
r
(q 1 + q 2 ) + i|ql |2 p ql (8.154)
For power law, the study will now be split into the following two cases.

In the presence of non-Hamiltonian perturbation terms, as mentioned after
(8.2), the dynamical system of the soliton parameters by virtue of the SPT are
1 1

2δ 1 + p 2 p p + 2 m+1
1
dA p
=

A2m+1
dt 2 − p 2 p2 1p m+1 p
+ 1
2
p+1
2p 1
τ
σ 1+ p p
+ 1
2
∞
1 ds
+

dτ A3− p
2 − p 2 p2 1p 12 −∞ cosh p τ
2 2
−∞ cosh p s
2 p−1

2β 2 p2 2p 1p + 12 p+1 p
+ 2

A2 p+1
p (2 − p) p + 1 1p p+1p
+ 1
2
2 2 p
1 2 2p
2 p−1
2β 2p 2β 2p
− AB − 2
2
A2 p+1 (8.155)
2− p p+1 p (2 − p) p + 1
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
3 p−2
dB 4β 2 p2 2p
p−2
= 2 B A2 p (8.156)
dt p p+1 p+2
1
d (A) (1) (2) δ 1 + p 2p
= F1 ( A, x, φ) − F1 ( A, x, φ) +
dt 2− p 2 p2

1p + 12 m+1 p 2m + 1
2m+1
. (2A) 2r +1 (A) 2m−2r
p 1
p + 2 r =0
m+1 1 2r + 1

p+1
2p
1
+ 1
σ 1+ p p 2
+
2− p 2 p2 p 12
1
∞ τ
1 ds
. 2 2 dτ

3 − p 2 2p +
2 p−1 1 1
2m+1
β 2p p 2
. (2A) 2− p+2r (A) 2r +1 + 2
r =0
2r + 1 p (2 − p) p + 1

p+1 2m + 1
2m+1
1 p
. (2A) 2r +1 (A) 2m−2r
p p+1 + 1 r =0 2r + 1
p 2
21p 2 p−1
2β 2 p2 β 2 p2 2p
− ABB − 2
2− p p+1 p (2 − p) p+1

2m + 1
2m+1
. (2A) 2r +1 (A) 2m−2r (8.157)
r =0
2r + 1
d (B)
= F2(1) ( A, x, φ) − F2(2) ( A, x, φ)
dt
3 p−2
4β 2 p2 2p
p−2
+ 2 [B A2 p + 2 p B A2 p+1 A] (8.158)
p p+1 p+2
p+2 p+2
1
d(x) A1 A2 2 p2 2
= −B + 2σ −
dt D13 D23 1+ p
1
τ
p + 12 ∞ τ ds
1
1
dτ (8.159)
p 2 −∞ cosh p τ
2 2
−∞ cosh p s
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
2p 2p 3 p−2
d(φ) A − A2 4β 2 p2 2p
p−2
= 1 − 2 A2 p Bx (8.160)
dt p+1 p p+1 p+2
For the fixed point of the dynamical system given by (8.155) and (8.156), with
A = 1 and B = 0, one gets

1 2p 1p + 12 m+1 p
β=

δ( p + 1)

2 2p − 1p + 12 1p m+1p
+ 12
(3 p−1)(2

τ
2p 1p + 12 ∞
p+1)
2 p+1 1 ds
2 p( p+1)
+ σp

dτ
1p 12
2 2
2 p2 −∞ cosh p τ −∞ cosh p s
(8.161)
Thus, by (8.159) and (8.160) one gets the coupled system of equations in x
and φ for the fixed point with A = 1 and B = 0 as
3 p−2
d 2 (x) 4β 2 p2 2p
p−2 d(x)
− 2 + F2(1) − F2(1) = 0 (8.162)
dt 2 p p+1 p+2 dt
and

d 2 (φ) 2p 2 p−1 d A1 2 p−1 d A2
= A − A (8.163)
dt 2 p+1 1
dt 2
dt
where in (8.162) β is given by (8.161). From (8.162) one can observe that a
damping is introduced in the soliton separation and the coefficient of the
damping term is positive as 0 < p < 2. In Figures 8.3(a) and 8.3(b), numerical
simulations show that the suppression of the SSI is achieved for power law
as proved in the QPT.

In the presence of Hamiltonian perturbation terms, as given by (8.2), the
dynamical system of the soliton parameters by virtue of the SPT are
dA
=0 (8.164)
dt
dB
=0 (8.165)
dt
and

d x0 2
12 + 1p 2p
= −B − A 3λ + 2µ 1

dt 2 2 + 2p 1p
p−1

3γ D3 p 12 + 1p
− µ + 3γ B +
2

+1 (8.166)
p2 1p 32 + 2p
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
60
50
40
30
t
20
10
0
–40 –30 –20 –10 0 10 20 30 40
x
(a)
60
50
40
30
t
20
10
0
–40 –30 –20 –10 0 10 20 30 40
x
(b)
FIGURE 8.3
(a) SSI for m = 1, p = 1/2, δ = 0.001, and (b) SSI for m = 2, p = 1/2, δ = 0.001.
d(A)
= F1(1) ( A, x, φ) − F1(2) ( A, x, φ) (8.167)
dt
d(B)
= F2(1) ( A, x, φ) − F2(2) ( A, x, φ) (8.168)
dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
d(x) 12 + 1p 2p
= −B − AA(3λ + 2µ) 1

dt 4 2 + 2p 1p
3γ
2
− 2 D 12D2 + D
4p
p−1
1
3γ
2 p
2
+ 1
p
+ 2 D 12D2 + D

(8.169)
p 1p 32 + 1p
and
2p 2p
d(φ) A − A2
= 1 (8.170)
dt p+1
Now (8.169) can be rewritten as

1
2
d(φ) 2 + p p
1
d(x)
= −B − (3λ + 2µ)g1

dt 4 dt 12 + 2p 1p
p−1
1
d(φ) p 2 + p
1
3γ d(φ) 3γ
− 2 g2 + 2 g2

(8.171)
4p dt p dt 1p 32 + 1p
For in-phase injection of solitons with unequal amplitudes
1
A= ( A0 + 1) (8.172)
2
B=0 (8.173)
A0 = A0 − 1 (8.174)
B0 = 0 (8.175)
T0 = T0 (8.176)
φ0 = 0 (8.177)
and
φ = δ (8.178)
so that for B = 0

x = x0 − (3λ + 2µ)
6
1
2
d(φ) 2 + p p
1

− (3λ + 2µ)h 1

4 dt 12 + 2p 1p
p−1
1
d(φ) p 2 + p
1
3γ d(φ) 3γ
− 2 h2 + 2 h2

(8.179)
4p dt p dt 1p 32 − 1p
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
50
45
40
35
30
25
t
20
15
10
0
–25 –20 –15 –10 –5 0 5 10 15 20
x
FIGURE 8.4
SSI with p = 1/2, γ = 0.14, λ = ν = 0.5.
where

h j (s) = g j (s)ds (8.180)
for j = 1, 2. Thus
x = x0 + O() (8.181)
Now, x0 ∼ O(1) so that x → 0 and thus the pulses do not collide during the
transmission. This can be observed in the numerical simulation in Figure 8.4.
8.3.3 Parabolic Law

For parabolic law nonlinearity
Al
ql (x, t) = 1 e −i Bl (x−xl )+iδl (8.182)
[1 + a l cosh{Dl (x − xl )}]2
where
√
Dl ≡ χ ( Al ) = Al 2 (8.183)
Substituting (8.63) into (8.1) gives
∂q l 1 ∂2 ql
i + + (|ql |2 + ν|ql |4 )ql = i R[ql , ql∗ ] − ql2 ql̄∗ + 2|ql |2 ql̄

∂t 2 ∂x 2

2
− ν ql3 ql̄∗ + 2|ql |2 ql2 ql̄∗ + 3|ql |4 ql̄ + 3|ql |2 ql∗ ql̄2 + 6|ql |2 |ql̄ |2 ql (8.184)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where l = 1, 2 and l̄ = 3 − l. Here, the separation
|q |2 q + ν|q |4 q = (|q 1 |2 q 1 + q 12 q 2∗ + 2|q 1 |2 q 2 ) + (|q 2 |2 q 2 + q 22 q 1∗ + 2|q 2 |2 q 1 )
+ ν[|q 1 |4 q 1 + q 13 (q 2∗ ) 2 + 2|q 1 |2 q 12 q 2∗ + 3|q 1 |4 q 2
+ 3|q 1 |2 q 1∗ q 22 + 6|q 1 |2 |q 2 |2 q 1 ]
+ ν[|q 2 |4 q 2 + q 23 (q 1∗ ) 2 + 2|q 2 |2 q 22 q 1∗ + 3|q 2 |4 q 1 + 3|q 2 |2 q 2∗ q 12 + 6|q 1 |2 |q 2 |2 q 2 ]
(8.185)
was used based on the degree of overlapping. By virtue of the SPT, the evo-
lution equations are
d Al
= F1(l) ( A, x, φ; ν) + Ml (8.186)
dt
d Bl
= F2(l) ( A, x, φ; ν) + Nl (8.187)
dt
dTl
= −Bl − F3 ( A, x, φ; ν) + Ql (8.188)
dt
and
dδl A2 B2
= l + l + F4 ( A, x, φ; ν) + Pl (8.189)
dt 4 2
where
∞ dτl
Ml = h (1)
1 ( Al ) R̂[ql , ql∗ ]e −iφl 1 dτl (8.190)
−∞ (1 + a l cosh τl ) 2
∞ sinh τl
Nl = h (1)
2 ( Al )
R̂[ql , ql∗ ]e −iφl 1 dτl (8.191)
−∞ (1 + a l cosh τl ) 2
∞ τl
Ql = h (1)
3 ( Al ) R̂[ql , ql∗ ]e −iφl 1 dτl (8.192)
−∞ (1 + a l cosh τl ) 2
and
∞ (1 − a l τl sinh τl )
Pl = h (1)
4 ( Al )
R̂[ql , ql∗ ]e −iφl 1 dτl (8.193)
−∞ (1 + a l cosh τl ) 2
Also, the following notations are used

2
R̂[ql , ql∗ ] = R[ql , ql∗ ] − ql2 ql̄∗ + 2|ql |2 ql̄ − ν ql3 ql̄∗ + 2|ql |2 ql2 ql̄∗ + 3|ql |4 ql̄

+ 3|ql |2 ql∗ ql̄2 + 6|ql |2 |ql̄ |2 ql (8.194)
For the parabolic law case, the study will be split into the following two
subsections.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book

In the presence of non-Hamiltonian perturbation terms, as given by (8.2), the
dynamical system of the soliton parameters by virtue of the SPT is

dA δ A2m+1 3 a −1 m+1 1
= m m−1 F m + 1, m + 1, m + , B ,
dt 2 a 2 2a p 2
√ ∞ τ
2 1 ds
+ σ a 2 dτ
2 −∞ 1 + a cosh τ −∞ 1 + a cosh s
⎡ ⎤
a 2
a − 1 1
− βa 2 A3 ⎣
3 tan
−1
−
⎦
a2 − 1 2 a + 1 2 a2 − 1

√ a 2 B2 −1 a −1
− 4 2β √ tan (8.195)
a −1
2 a +1
and
√ ⎡ ⎤
dB 3
2A B ⎣ a 2
a −1 1
= −β
32 tan
−1
−
⎦ (8.196)
dt E a −1
2 a + 1 2 a2 − 1
where E is the energy of the soliton given by

∞ √
2 2A a −1
E= |q |2 d x = √ tan−1 (8.197)
−∞ a −1
2 a +1
√ ∞ 1
τ ds

σ 2 −∞ 1+a cosh τ −∞ 1+a cosh s dτ

β=
2 a2 −1 a −1
− 2(a 21−1)
3 tan a +1
(a 2 −1) 2
−1

m+1 1
δ F m + 1, m + 1, m + 32 ; a2a B p ,2
+ m m+1 (8.198)
2 a a 2 −1 a −1
− 1
3 tan a +1 2(a −1)
2
(a 2 −1) 2
Thus, using (8.79)–(8.80), one can obtain
d 2 (x) d(x)
2
+ βG + F2(1) − F2(2) = 0 (8.199)
dt dt
where β is given by (8.198) and G > 0 represents the coefficient of −βB
in d(B)/dt = d B1 /dt − d B2 /dt. Now, (8.199) shows a damping in the sepa-
ration of solitons, thus proving that there will be a suppression of SSI in the
presence of the perturbation terms given by (8.2). The numerical simulations
in Figure 8.5 show that the suppression of SSI is achieved for parabolic law
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
60
50
40
30
t
20
10
0
–40 –30 –20 –10 0 10 20 30 40
x
FIGURE 8.5
SSI with m = 0, δ = 0.001, σ = 0.001.

In the presence of Hamiltonian perturbation terms, as given by (8.2), the
dynamical system of the soliton parameters by virtue of the SPT is
dA
=0 (8.200)
dt
dB
=0 (8.201)
dt
and
√ √
d x0 α 2 2A −1 a −1 2 A3
= −B − √ tan − (3λ + 2µ)
dt E a2 − 1 a +1
2E a 2 − 1
32
4 2
! a −1 √ γ A a A 2κ 2
−1
. a − 1 − 2 tan
2 −3 2 +√
a +1 E 2 a2 − 1

−1 a − 1 a 2 A2 ! 2
. tan − a −1 (8.202)
a +1 4
From (8.79), (8.80), (8.83), and (8.84), one can now conclude that
d (A)
= F1(1) ( A, x, φ) − F1(2) ( A, x, φ) (8.203)
dt
d (B)
= F2(1) ( A, x, φ) − F2(2) ( A, x, φ) (8.204)
dt
d(x) d x1 d x2
= − (8.205)
dt dt dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
d(φ) 1
= AA (8.206)
dt 2
From (8.202), one can get

d(x) d(φ)
= −B + G (α, λ, µ, γ , σ ) g (8.207)
dt dt
where G is the functional form that depends on the said parameters. For
in-phase injection of solitons with unequal amplitudes
1
A= ( A0 + 1) (8.208)
2
B=0 (8.209)
A0 = A0 − 1 (8.210)
B0 = 0 (8.211)
T0 = T0 (8.212)
φ0 = 0 (8.213)
and
φ = δ (8.214)
so that one can obtain from (8.207) for B = 0

d(φ)
x = x0 + G (α, λ, µ, γ , σ ) h (8.215)
dt
where

h j (s) = g j (s)ds (8.216)
for j = 1, 2. Thus
x = x0 + O() (8.217)
Now, x0 ∼ O(1) so that x → 0 and thus the pulses do not collide during the
transmission, as shown in the numerical simulations in Figure 8.6.
8.3.4 Dual-Power Law

For dual-power law nonlinearity
Al
ql (x, t) = 1 e −i Bl (x−xl )+iδl (8.218)
[1 + a l cosh {Dl (x − xl )}] 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
60
50
40
30
t
20
10
0
–30 –20 –10 0 10 20 30
x
FIGURE 8.6
SSI with ν = 0.3, α = 0.8, γ = −0.35, σ = 0.03.
where
21p
p 4 p2
Dl ≡ χ( Al ) = Al (8.219)
1+ p
Here (8.66) for dual-power law modifies to
∂q l 1 ∂2 q l
i + = i R[ql , ql∗ ]
∂t 2 ∂x2
p p
p p−r p
p−r
r
∗ ∗
− q q2 r
q1 q2 (q 1 + q 2 )
r =0
r 1 r =0
r
2p 2p
2 p 2 p−r 2 p
2 p−r
r
−ν q 1 q 2r q 1∗ q 2∗ (q 1 + q 2 ) (8.220)
r =0
r r =0
r
where the separation
p
p
p p−r r p ∗
p−r ∗
r
|q | q + ν|q | q =
2p 4p
q q2 q1 q2 (q 1 + q 2 )
r =0
r 1 r =0
r
2p 2p
2 p 2 p−r 2p ∗
r
∗ 2 p−r
+ν q1 q2 r
(q 1 ) q2 (q 1 + q 2 ) (8.221)
r =0
r r =0
r
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
was used based on the degree of overlapping. By virtue of the SPT, the evo-
lution equations of the soliton parameters are
d Al
= F1(l) ( A, x, φ; ν, p) + Ml (8.222)
dt
d Bl
= F2(l) ( A, x, φ; ν, p) + Nl (8.223)
dt
dTl
= −Bl − F3 ( A, x, φ; ν, p) + Ql (8.224)
dt
and
2p
dδl Al B2
= + l + F4 ( A, x, φ; ν, p) + Pl (8.225)
dt 2p + 2 2
where for the case of dual-power law

∞ dτl
Ml = h (2)
1 ( Al ) R̂[ql , ql∗ ]e −iφl 1 dτl (8.226)
−∞ (1 + a l cosh τl )2 p
∞ sinh τl
Nl = h (2)
2 ( Al )
R̂[ql , ql∗ ]e −iφl 1 dτl (8.227)
−∞ (1 + a l cosh τl )2 p
∞ τl
Ql = h (2)
3 ( Al ) R̂[ql , ql∗ ]e −iφl 1 dτl (8.228)
−∞ (1 + a l cosh τl )2 p
and

∞ (1 − a l τl sinh τl )
Pl = h (2)
4 ( Al )
R̂[ql , ql∗ ]e −iφl 1 dτl (8.229)
−∞ (1 + a l cosh τl ) 2 p
In addition, the following notations are used
R̂[ql , ql∗ ] = R[ql , ql∗ ] + i|ql |2 p ql + iν|ql |4 p ql

p p
p p−r p
p−r
r
∗ ∗
−i q q2 r
q1 q2 (q 1 + q 2 )
r =0
r 1 r =0
r

2p 2p
2 p 2 p−r r 2 p ∗
2 p−r ∗
r
−i q1 q2 q1 q2 (q 1 + q 2 )
r =0
r r =0
r
(8.230)
For the case of dual-power law nonlinearity, the study is split into two sub-
sections.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book

In the presence of nonperturbation terms, as given by (8.2), the dynamical
system of the soliton parameters by virtue of the SPT are

dA 4δ A2m+2 m+1 m+1 m+1 1 a −1 m+1 1
= m+1 m+1 F , , + ; B ,
dt 2 p a p D p p p 2 2a p 2

2σ A4 ∞ 1 τ
ds
+ 2 1 1 dτ
D −∞ (1 + a cosh τ ) p −∞ (1 + a cosh s) p

4β A2 1 1 3 1 a −1 1 3
− 1 1 D 2
F 2 + , , + ; B ,
Da p 2 p p p 2 p 2a p 2

1 1 1 1 a −1 1 1
+ B2 F , , + ; B , (8.231)
p p 2 p 2a p 2
1 a −1

β B D2 F 2 + p , 1 + p , 2 + p ; 2a B 1 + p , 1
1 1 1
dB
=− 2 2 −1

(8.232)
dt 4p A F 1p , 1p , 12 + 1p ; a2a B 1p , 12
1 m+1 m+1 m+1 1 a −1
m+1 3
1 3
δ 1 + p p F p , p , p + 2 ; 2a B p , 2 σ a p 1 + p 2p
β= m m −1

+
F 2 + 1p , 1p , 32 + 1p ; a2a B 1p , 32
p−1
2p a p 2 p2 2p 2 p2
∞
1 1
1 3
1 a −1

B p , 2 F 2 + p , p , 2 + p ; 2a
1 1 3 1
−∞ (1 + a cosh τ ) p

τ
ds
1 dτ (8.233)
−∞ (1 + a cosh s) p
Thus, using (8.79)–(8.80), one can obtain
d 2 (x) d(x)
+ βG + F2(1) − F2(2) = 0 (8.234)
dt 2 dt
where β is given by (8.233) and G > 0 represents the coefficient of −βB in
d(B)/dt = d B1 /dt − d B2 /dt. Now, (8.234) shows damping in the separation
of solitons, thus proving that there will be a suppression of SSI in the presence
of the perturbation terms given by (8.2). Thus, in Figure 8.7, the numerical
simulations show that the suppression of SSI is achieved for dual-power law

In the presence of the perturbation terms given by (8.4), the dynamical system
of the soliton parameters by virtue of the SPT are
dA
=0 (8.235)
dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
60
50
40
30
t
20
10
0
–40 –30 –20 –10 0 10 20 30 40
x
FIGURE 8.7
SSI with m = 1, δ = σ = 0.001.
dB
=0 (8.236)
dt
and
d x0
= −B − (µ + 3γ B 2 )
dt
3 a −1
1
γ D2 F (2 + p , p , p + 2 ; 2a ) B p ,
1 1 1 1
2
−3 −1

2 p2 F 1p , 1p , 12 + 1p ; a2a B 1p , 3
2
−1

1

A2 F 2p , 2p , 2p + 12 ; a2a B ,1
− 3λ + 2µ 1 1 1 1 1 1 a −1

2p 21
(8.237)
2 p a p F p , p , 2 + p ; 2a B p, 2
From (8.79), (8.80), (8.83), and (8.84), one can now conclude that
d (A)
= F1(1) ( A, x, φ) − F1(2) ( A, x, φ) (8.238)
dt
d (B)
= F2(1) ( A, T, φ) − F2(2) ( A, x, φ) (8.239)
dt
d(x) d x1 d x2
= − (8.240)
dt dt dt
and
d(φ) 1
= AA (8.241)
dt 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
For in-phase injection of solitons with unequal amplitudes
1
A= ( A0 + 1) (8.242)
2
B=0 (8.243)
A0 = A0 − 1 (8.244)
B0 = 0 (8.245)
T0 = T0 (8.246)
φ0 = 0 (8.247)
and
φ = δ (8.248)
so that from (8.237), one can get

d(x) d(φ)
= −B + G (α, λ, µ, γ , σ ) g (8.249)
dt dt
where G is the functional form that depends on the said parameters. For
in-phase injection of solitons with unequal amplitudes, where B = 0

d(φ)
x = x0 + G (α, λ, µ, γ , σ ) h (8.250)
dt
60
50
40
30
t
20
10
0
–30 –20 –10 0 10 20 30
x
FIGURE 8.8
SSI with ν = 0.5, p = 1/2, α = 0.8, γ = −0.25, σ = 0.05, µ = 0.5.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where

h j (s) = g j (s)ds (8.251)
for j = 1, 2. Thus
x = x0 + O() (8.252)
Now, x0 ∼ O(1) so that x → 0 and, thus, the pulses do not collide during
the transmission, as observed in the numerical simulations in Figure 8.8.
Exercises
1. Use the adiabatic parameter dynamics given by (8.116), (8.117),

(8.109), and (8.110) to obtain the relations given by (8.118)–(8.121).
2. Use the adiabatic parameter dynamics given by (8.155), (8.156),
(8.148), and (8.149) to obtain the relations given by (8.157)–(8.160).
3. For the fixed point of the dynamical system given by (8.195) and
(8.196), prove that β is given by (8.198) for A = 1 and B = 0.
4. For the fixed point of the dynamical system given by (8.231) and
(8.232), prove that β is given by (8.233) for A = 1 and B = 0.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
9
Stochastic Perturbation
The use of optical amplifiers affects the evolution of solitons considerably. The
reason is that, although amplifiers are necessary to restore the soliton energy,
they add noise originating from amplified spontaneous emission (ASE). The
effect of ASE is to change the soliton parameters randomly. Variances of such
fluctuations can be calculated by treating ASE as a perturbation. In this chap-
ter, such calculations will be carried out for Kerr law, power law, parabolic
law, and dual-power law fibers.
9.1 Introduction
Besides deterministic-type perturbations, one also needs to take into account,
for practical considerations, stochastic-type perturbations. These effects can
be classified into three basic types:
1. Stochasticity associated with the chaotic nature of the initial pulse
due to partial coherence of the laser-generated radiation.
2. Stochasticity due to random nonuniformities in the optical fibers,
such as fluctuations in the values of the dielectric constant, random
variations of the fiber diameter, and more.
3. The chaotic field caused by dynamic stochasticity might arise from
a periodic modulation of the system parameters or when a periodic
array of pulses propagates in a fiber-optic resonator.
Thus, stochasticity is inevitable in optical soliton communications.
Stochasticity is basically of two types, namely homogenous and
nonhomogenous [89]:
• In the homogenous case, stochasticity originates due to random
perturbation of the fiber, such as the density fluctuation of the fiber
material or random variations in the fiber diameter.
• In the inhomogenous case, stochasticity is present in the input pulse
of the fiber. So the parameter dynamics are deterministic, although
the initial values are random.
135
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Considering the effects of perturbation [89] on the propagation of solitons

through optical fibers for the 1 + 1 dimensional case, the perturbed nonlinear
Schrödinger’s equation is
1
iqt + q xx + F (|q |2 )q = i R (9.1)
2
with
R = δ|q |2m q + βq xx + σ (x, t) (9.2)
where σ (x, t) is the random stochastic perturbation term.
The amplifiers that are placed along soliton transmission lines to restore
soliton energy introduce noise that originates from ASE. To study the impact
of noise on soliton evolution, the evolution of the mean free velocity of the
soliton due to ASE will be studied in this chapter. In the case of lumped am-
plification, solitons are perturbed by ASE in a discrete fashion at the location
of the amplifiers. It can be assumed that noise is distributed all along the fiber
length since the amplifier spacing satisfies za 1. In (9.2), σ (x, t) represents
the Markovian stochastic process with Gaussian statistics and assumes that
σ (x, t) [227] is a function of t only so that σ (x, t) = σ (t). Now, the complex
stochastic term
σ (t) = σ1 (t) + iσ2 (t) (9.3)
is further assumed to be independently delta correlated in both σ1 (t) and σ2 (t)
with
σ1 (t) = σ2 (t) = σ1 (t)σ2 (t ) = 0 (9.4)

σ1 (t)σ1 (t ) = 2D1 δ(t − t ) (9.5)

σ2 (t)σ2 (t ) = 2D2 δ(t − t ) (9.6)
Now, in (9.5) and (9.6), if it is assumed that D1 = D2 = D, then

σ (t) = 0 (9.7)
and
σ (t)σ (t ) = 2Dδ(t − t ) (9.8)
In soliton units, one gets

Fn F G
D= (9.9)
Nph za
where Fn is the amplifier noise figure, while
(G − 1) 2
FG = (9.10)
G ln G
is related to the amplifier gain G and, finally, Nph is the average number of
photons in the pulse propagating as a fundamental soliton.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Stochastic Perturbation 137
In the presence of these perturbation terms, the adiabatic dynamics of

the conserved quantities and the soliton parameter κ can be computed from
(1.34), (1.35), and (1.37) using the R that is given by (9.2). Also, the general
form of the soliton given by (2.24) is used so that
dE 2
= [δ A2m+2 I0,2m+2,0,0 + β A2 B( B 2 I0,1,0,1 − κ 2 I0,2,0,0 )]
dt β
∞
+2A g(τ ) (σ1 cos φ + σ2 sin φ) d x (9.11)
−∞
and
dM 2
= [κ A2 (2β AB 3 I0,0,3,0 − δ A2m I0,2m+2,0,0 ) + βκ A2 (κ 2 I0,2,0,0 − B 2 I0,1,0,1 )]
dt B
∞
dg
+ 2 A g(τ ) (σ1 cos φ + σ2 sin φ) − B (σ2 cos φ − σ1 sin φ) d x
−∞ dτ
(9.12)
From (9.11) and (9.12), it is possible to derive
dκ 4βκ B 2
= (κ I0,2,0,0 + B 2 I0,0,2,0 − B 2 I0,1,0,1 )
dt I0,2,0,0
∞
2 B dg
− B (σ2 cos φ − σ1 sin φ) + 2κg(τ ) (σ1 cos φ + σ2 sin φ) d x
AI0,2,0,0 −∞ dτ
(9.13)
The study of solitons in the presence of stochastic perturbation will now

be split into the following four sections: Kerr law nonlinearity, power law,
parabolic law, and dual-power law.
9.2 Kerr Law

For the Kerr law nonlinearity equations (9.11) and (9.13), the perturbation
terms given by (9.2) respectively modify to
1
dA 2m 2 (m + 1) 2 2
= A δA − β A − 3βκ 2
dt m + 32 3
⎧ ⎫⎤
⎨ σ cos(ωt + σ ) − σ sin(ωt + σ ) ⎬
+π
1 0
2 0 ⎦ (9.14)
⎩ cosh πκ ⎭
2η
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and

dκ 4 π
= − β A2 κ + (2A2 + 2κ 2 − κ A)
dt 3 2η

σ1 cos(ωt + σ0 ) − σ2 sin(ωt + σ0 )
(9.15)
cosh π2Aκ
Equations (9.14) and (9.15) are difficult to analyze. If the terms with σ1 and
σ2 are suppressed, the resulting dynamical system has a stable fixed point,
namely a sink given by ( Ā, κ̄) = ( Ā, 0) where
2m−2
1
2β m + 32
Ā = (9.16)
3δ 12 (m + 1)
Now, linearizing the dynamical system about this fixed point and simplifying
gives

dA ζ
= − A 2m+1
− (9.17)
dt Ā
and
dκ
= −[κ − ζ (1 + A − κ)] (9.18)
dt
where

σ1 cos (ωt + σ0 ) − σ2 sin (ωt + σ0 )
ζ =π (9.19)
cosh π2Aκ
Equations (9.17) and (9.18) are called the Langevin equations and will now be
analyzed to compute the soliton mean drift velocity. If soliton parameters are
chosen such that ζ A is small, then (9.18) yields
dκ
= −[κ − ζ (1 − κ)] (9.20)
dt
One can solve (9.20) for κ, and eventually the mean drift velocity of the soliton
can be obtained. The stochastic phase factor of the soliton is defined by
t
ψ(t, y) = ζ (s)ds (9.21)
y
where t > y. Assuming that σ is a Gaussian stochastic variable, one arrives at
e ψ(t, y) = e D(t−y) (9.22)

[ψ(t, y)+ψ(t , y )]
e =e Dθ
(9.23)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where
θ = 2(t + t − y − y ) − |t − t | − |y − y | (9.24)
∂
ζ ( y)e −ψ(t, y) = e −ψ(t, y) = De D(t−y) (9.25)
∂y
and
∂2
ζ ( y)ζ ( y )e [−ψ(t, y)−ψ(t , y )] = 2Dδ( y − y )e Dθ + e Dθ (9.26)
∂ y∂ y
Now, solving (9.20) with the initial condition as κ(0) = 0 and using equations
(9.21)–(9.26), the soliton mean drift velocity is given by
D
κ(t) = − 1 − e −(1−D)t (9.27)
1− D
From (9.27), it follows that
D
lim κ(t) = − (9.28)
1− D
t→∞
so that, in the limiting case, the mean free velocity of the soliton is given as
D
v = (9.29)
1− D
Thus, for large t, v(t) approaches a constant value provided D < 1. For
D > 1, κ(t) becomes unbounded for large t.
9.3 Power Law

For the case of power law nonlinearity, equations (9.11) and (9.13), respec-
tively, modify to
1 1
2 δ 2m+1 1 + p 2 p p + 2 m+1
1
dA p
= A
dt 2− p 2 p2 1p m+1 p
+ 12
p−1 1
2p + 12
2β Ap−1 2 p2 p
+
2− p B p+1 1p
p+1
A2 B 2 p A2 2 2 1p
− 2 (κ p + B ) 1 2

p 2 p+1p
+ 12 p p + 12
p−1
2 A p−1 2 p 2 2p 1p + 12
+ A
2− p 1+ p 12 1p
∞ ∞
cos φ sin φ
σ1 1 d x + σ2 1 dx (9.30)
−∞ cosh p τ −∞ cosh p τ
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
1
dκ 4 β 2 2 p−2 2 p2 2 p − 2
= κA B p
dt p2 p+1 p+2
1 1
2 p2 p p + 2
1 ∞
p−2 B tanh τ
+ 4 κ AB p 1 1 (σ2 cos φ − σ1 sin φ)
1+ p 2 p −∞ p cosh p τ
1

2κ
+ 1 (σ1 cos φ + σ2 sin φ) d x (9.31)
cosh p τ
Equations (9.30) and (9.31) are difficult to analyze. If the terms with σ1 and
σ2 are suppressed, the resulting dynamical system has a stable fixed point,
namely a sink given by ( Ā, κ̄) = ( Ā, 0) where
2(m−1 p)
2β m+1
p
+ 12 p+1p
1p
Ā = − 1 (9.32)
δ( p + 1) m+1p p+1
p
+ 12 p + 12
After simplification, linearizing the dynamical system about this fixed point
gives

(1)
dA ζ
= − A2m+1 − 1 (9.33)
dt Ā
and
dκ
= − κ − ζ2(1) (1 + A − κ) (9.34)
dt
where
∞ ∞
cos φ sin φ
ζ1(1) = σ1 1 d x + σ2 1 dx (9.35)
−∞ cosh τ
p −∞ cosh p τ
and
∞
B tanh τ 2κ
ζ2(1) = (σ 2 cos φ − σ 1 sin φ) + (σ 1 cos φ + σ 2 sin φ) dx
−∞ p cosh 1p τ 1
cosh p τ
(9.36)
Similarly, as in the case of Kerr law nonlinearity, these Langevin equations

lead to the mean drift velocity of the soliton as
D
v = (9.37)
1− D
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
9.4 Parabolic Law

For parabolic law nonlinearity, (9.11) and (9.13) respectively modify to

dA δ A2m+1 3 a −1 1
= m m+1 F m + 1, m + 1, m + ; B m + 1,
dt 2 a 2 2a 2
√
2β A 2 3 a −1 1
+ κ F 1, 1, ; B 1,
a 2 2a 2

5 a − 1 3
− 2A2 F 3, 1, ; B 1,
2 2a 2

√ ∞
cos φ
+ a 2 A 2 σ1 1 dx
−∞ (1 + a cosh τ ) 2
∞
sin φ
+ σ2 1 dx (9.38)
−∞ (1 + a cosh τ ) 2
and
−1

dκ βκ A2 F 3, 2, 3; a2a B (2, 1)
=− 3 a −1
1
dt 2 F 1, 1, 2 ; 2a B 1, 2
√ ∞
2 a B (σ2 cos φ − σ1 sin φ) sinh τ 2κ (σ1 cos φ + σ2 sin φ)
− 3 − 1 dx
AE −∞ (1 + a cosh τ ) 2 (1 + a cosh τ ) 2
(9.39)
gives

dA ζ1(2)
= − A 2m+1
− (9.40)
dt Ā
and
dκ
= − κ − ζ2(2) (1 + A − κ) (9.41)
dt
where Ā is the fixed point of the amplitude, while
∞ ∞
(2) cos φ sin φ
ζ1 = σ1 1 d x + σ2 1 dx (9.42)
−∞ (1 + a cosh τ ) 2 −∞ (1 + a cosh τ ) 2
and
∞
a B(σ2 cos φ − σ1 sin φ) sinh τ 2κ(σ1 cos φ + σ2 sin φ)
ζ2(2) = 3 − 1 dx
−∞ (1 + a cosh τ ) 2 (1 + a cosh τ ) 2
(9.43)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Again, as in the case of Kerr law nonlinearity, these Langevin equations lead
to the mean drift velocity of the parabolic law soliton as
D
v = (9.44)
1− D
9.5 Dual-Power Law

In the case of dual-power law nonlinearity, (9.11) and (9.13), respectively,
modify to

dA 4δ A2m+2 m+1 m+1 m+1 1 b−1 m+1 1
= m+1 m+1 F , , + ; B ,
dt Bb 2p p p p p 2 2b p 2

4β A2 1 1 3 1 b−1 1 3
− 1 1 B F 2+ , , + ;
2
B ,
Bb p 2 p p p 2 p 2b p 2

1 1 1 1 b−1 1 1
+ κ2 F , , + ; B ,
p p 2 p 2b p 2
21p ∞
2 p+1 cos φ
+ p−2 2
σ1 1 dx
pL A 2p −∞ (1 + a cosh τ ) 2 p
∞
sin φ
+ σ2 1 dx (9.45)
−∞ (1 + a cosh τ ) 2 p
and

βκ B 2 F 2 + p , 1 + p , 2 + p ; 2b B 1 + p , 1
1 1 1 b−1 1
dκ
=− 2 2
dt 4p A F 1p , 1p , 12 + 1p ; b−1
2b
B 1p , 12

∞ a B (σ2 cos φ − σ1 sin φ) sinh τ 2κ(σ1 cos φ + σ2 sin φ)
− 2 p+1 − 1 dx
E −∞ 2 p (1 + a cosh τ ) 2 p (1 + a cosh τ ) 2 p
(9.46)
gives

dA ζ1(3)
= − A 2m+1
− (9.47)
dt Ā
and
dκ
= − κ − ζ2(3) (1 + A − κ) (9.48)
dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where Ā is the fixed point of the amplitude, while

∞ ∞
cos φ sin φ
ζ1(3) = σ1 1 d x + σ 2 1 dx (9.49)
−∞ (1 + a cosh τ ) 2 p −∞ (1 + a cosh τ ) 2 p
and

∞
a B (σ2 cos φ − σ1 sin φ) sinh τ 2κ (σ1 cos φ + σ2 sin φ)
ζ2(3) = 2 p+1 − 1 dx
−∞ 2p (1 + a cosh τ ) 2 p (1 + a cosh τ ) 2 p
(9.50)
Once again, as in the case of Kerr law nonlinearity, one can derive the mean
drift velocity of the soliton as
D
v = (9.51)
1− D
Exercises
1. Obtain the adiabatic parameter dynamics for the soliton amplitude
and frequency that are given by (9.14) and (9.15).
2. For Kerr law nonlinearity, obtain the two-point correlation for
κ(t)κ(t )
3. For power law nonlinearity, establish the fixed point of the ampli-
tude from (9.30) and (9.31) for σ1 and σ2 suppressed.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
10
Optical Couplers
This chapter provides an overview of optical couplers. In Section 10.1, an

introduction to optical couplers is given. Also discussed in this section are
the various kinds of couplers and their functions. In Section 10.2, the concept
of a twin-core coupler is discussed. The various parameters of the solitons
are defined and the corresponding parameter dynamics of the solitons are
obtained from the corresponding nonlinear Schrödinger’s equation (NLSE).
In Section 10.3, multiple-core couplers are discussed. Finally, Section 10.4
contains a discussion on solitons in magneto-optic waveguides.
10.1 Introduction
In this chapter, the study of the NLSE will be extended to the case of optical
couplers. Optical nonlinear couplers are very useful devices that distribute
light from a main fiber into one or more branch fibers. Couplers also have
applications as intensity-dependent switches and as limiters. They can be
used to multiplex two incoming bit streams onto a fiber and also to demul-
tiplex a single-bit stream. Optical couplers can be made as planar devices
using semiconductor material or as dual-core, single-mode fibers with soli-
tons propagating in each core. The coupling of energy from one guide to the
other can occur due to the overlap of evanescent fields between the cores.
There are core, interaction-type couplers as well as surface-interaction-
type couplers. In core-interaction-type couplers, the light energy transfer
takes place through the core cross-section concatenating fibers or by using
some form of imaging optics between the fibers (i.e., using lensing schemes
such as rounded end fibers, a spherical lens to image the core of one fiber onto
the core area of the other fiber, and a taper-ended fiber). In surface-interaction-
type couplers, the light energy transfer takes place through the fiber surface
and normal to the axis of the fiber by converting the guided core modes to
cladding and refracted modes.
145
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
(a)
(b)
FIGURE 10.1
(a) Three port coupler as splitter, (b) three port coupler as combiner.
10.1.1 Types of Couplers and Their Functions

10.1.1.1 Three- and Four-Port Couplers
Figures 10.1(a) and 10.1(b) show the uses of a three-port coupler as splitter and
combiner of signals. Light from the input fiber is coupled to the output fibers
as shown in Figure 10.1(a), or the light from the branch fibers are combined
to form a single input to the output fiber. In one method of splitting, called
the lateral offset method, a single input fiber core is situated between the cores
of two output fibers. In this method, the input power can be distributed in a
well-defined proportion by appropriate control of the amount of lateral offset
between the fibers.
Figure 10.2 shows the directional coupler, which is a four-port coupler.
In this coupler, the fibers are generally twisted together and then spot fused
under tension such that the fused section is elongated to form a biconical
tapered structure. It can act as a three-port coupler (T coupler) if one of the
input ends or one of the output ends is closed. As shown in the figure, each
port is meant for a different function.
This type of coupler is based on the transfer of energy by surface interac-
tion between the fibers. The amount of power taken from the main fiber or
given to the main fiber depends on the length of the fused section of the fiber
and the distance between the cores of the fused fibers. This can also act as a
fiber laser amplifier.
10.1.1.2 Star Coupler or Multiport Couplers

A star coupler is used to distribute an optical signal from a single input fiber
to multiple output fibers. Here, many fibers are bundled, twisted, heated, and
pulled at the twisted area to get fiber fused biconical tapered star coupler.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Optical Couplers 147
FIGURE 10.2
Four-port coupler.
10.1.2 Optical Switching

Solitons in nonlinear couplers have attracted much attention in recent years.
The special interest is in optical switching because of the possibility of very
short switching times in the femtosecond range. Switching is the process of
energy redistribution between the cores for a given input. The problem of
switching, although involved, can be accomplished when the stability of soli-
ton states is known. Another practical application is in logic device or in high-
bit rate communications in local networks. However, an optimal design of the
fiber arrays is still awaited. Since solitons are used in long-distance fiber-optic
communications it would be ideal if intensity-dependent switching could be
performed in all optical devices, instead of converting from optical to electri-
cal signals for electronic switching and then regenerating the optical signals.
Moreover, fast optical computers can be designed using soliton switching and
logic components involving couplers.
The switching of solitons can be realized in terms of all optical logic gates.
Logic gate operations such as OR, AND, EX-OR; and NAND have been re-
alized for soliton switching. It has been realized that these components will
find useful applications in the field of optical information processing, permit-
ting the realization of logic operations at speeds unattainable by conventional
electronic systems. The ultimate aim should be the realization of all-optical
digital computers interfaced, without the need of optoelectronic converters,
of soliton communication systems in which the data rates are so high that
conventional logic gates are too slow.
When a soliton pulse is transmitted through a birefringent fiber, an
exchange of energy is possible between the pulses propagating in two ortho-
gonal modes. The two orthogonally polarized solitons can trap one another
and move at a common group velocity, in spite of their different modal indices
(polarization dispersion). This phenomenon is known as soliton trapping and
is quite important for optical soliton switching. This type of switching in fiber
is intensity dependent. Depending upon its intensity, the pulse itself induces
switching. The control of data flow in a fiber-optic communication system has
been accomplished by switching of optical signals between fibers. Switching
depends on the power and phase of the pulse.
In practice, an optical pulse is used for nonlinear switching. However, the
switched pulse is severely distorted because only certain parts of the pulse
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
have the right power for switching. In particular, pulse wings are too weak
for switching to occur. Solitons can avoid pulse distortion because of their
extraordinary stable property that the optical phase remains uniform across
the entire pulse in spite of fiber nonlinearity. It has been predicted that the
soliton interaction itself acts as a switch to suppress or enhance the switching
dynamics.
10.2 Twin-Core Couplers

For twin-core couplers, wave propagation at relatively high field intensities
is described by coupled nonlinear equations. In the dimensionless form, they
are
1
iqt + q xx + F (|q |2 )q = Kr (10.1)
2
1
irt + rxx + F (|r |2 )r = Kq (10.2)
2
The constant K represents the coupling coefficient between the cores of the
fiber. In (10.1) and (10.2), q and r represent dimensionless forms of the optical
fields in the respective cores of the optical fibers. This system of equations
models various applications, such as intensity-dependent switches and de-
vices for separating a compressed soliton from its broad “pedestal.”
Equations (10.1) and (10.2) are not integrable due to the presence of the
arbitrary functional form given by F and due to the coupling term. There-
fore, the propagation of solitons through these twin-core optical fibers will be
studied approximately. Also, equations (10.1) and (10.2), as it appears, do not
have infinitely many integrals of motion. In fact, they have as few as three.
They are respectively given by energy (E), linear momentum (M), and the
Hamiltonian (H) that are respectively given by [31, 79]
∞
E= (|q |2 + |r |2 )dx (10.3)
−∞

i ∞ ∗
M= q q x − q ∗ q x + rr x∗ − r ∗r x dx (10.4)
2 −∞
and

∞
1
H= |q x |2 + |r x |2 − f (|q |2 ) − f (|r |2 ) − K (qr ∗ + rq ∗ ) dx (10.5)
−∞ 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
where
I
f (I) = F (ξ )dξ
0
and the intensity I is given by I = |q |2 or I = |r |2 depending on the core.

For K = 0, the pulses in the two fibers are assumed to be of the following
functional form
q (x, t) = A1 (t)g [B1 (t) (x − x1 (t))] exp [−iκ1 (t)(x − x1 (t)) + iθ1 (t)] (10.6)
and
r (x, t) = A2 (t)g [B2 (t) (x − x2 (t))] exp [−iκ1 (t)(x − x2 (t)) + iθ2 (t)] (10.7)
Here, Aj and B j represent the amplitude and width of the soliton respectively
in the two cores. Also, κ j and θ j respectively represent the frequency and the
phase of the soliton in the two cores for j = 1, 2, while g represents the
functional form of the soliton. It depends on the type of nonlinearity in F .
Finally, x j (t) represents the center position of the soliton in the two cores. For
convenience, the following few integrals are defined that will be in use
∞ c
dg
Ia(l),b,c = τla g b (τl ) dτl (10.8)
−∞ dτl
∞
Pa ,b,c = x a g b [B1 (x − x1 (t)] g c [B2 (x − x2 (t)]
−∞
cos [(κ1 − κ2 )x − (κ1 x1 − κ2 x2 ) − (θ1 − θ2 )] dx (10.9)

∞
Qa ,b,c = x a g b [B1 (x − x1 (t)] g c [B2 (x − x2 (t)]
−∞
sin [(κ1 − κ2 )x − (κ1 x1 − κ2 x2 ) − (θ1 − θ2 )] dx (10.10)
for nonnegative integers a , b, and c and l = 1, 2. Here, τ1 = B1 (t)(x − x1 (t))

and τ2 = B2 (t)(x − x2 (t)). The integrals of motion in (10.3), (10.4), and (10.5),
on using the soliton form in (10.6) and (10.7), respectively, reduce to [79]
∞ A2 (1) A2 (2)
E= |q |2 + |r |2 dx = 1 I0,2,0 + 2 I0,2,0 (10.11)
−∞ B1 B2
∞
i A2 (1) A2 (2)
M= {(q q x∗ − q ∗ q x ) + (rr x∗ − r ∗r x )}dx = − κ1 1 I0,2,0 + κ2 2 I0,2,0
2 −∞ B1 B2
(10.12)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and

1 ∞ 2 2 ∗ ∗

H= |q x | + |r x | − f |q | − f |r | − K qr + rq
2 2
dx
−∞ 2
1
2 1 κ 2 A2 κ22 A22 (2)

(1) (2) 1 1 (1)
= A B1 I0,0,2 + A2 B2 I0,0,2 +
2
I + I
2 1 2 B1 0,2,0 B2 0,2,0
∞ I
− 2K A1 A2 P0,1,1 − 2 F (s)ds (10.13)
−∞ 0
The soliton parameters for the first core are now defined as
(1) ∞ 12
−∞ |q | dx
4
I0,2,0
A1 (t) = (1) ∞ (10.14)
−∞ |q | dx
I0,4,0 2
(1) ∞
12
I2,2,0 −∞ |q | 2
dx
B1 (t) = (1) ∞
(10.15)
−∞ x |q | dx
I0,2,0 2 2
∞
i −∞ q q x∗ − q ∗ q x dx
κ1 (t) = ∞ (10.16)
−∞ |q | dx
2 2
and
∞
x|q |2 dx
x1 (t) = −∞
∞ (10.17)
−∞ |q |2 dx
Treating the coupling terms as perturbation terms in (10.1) and differentiat-

ing these soliton parameters while using the soliton given by (10.6) as well
as the modified parameter dynamics to obtain the evolution of the soliton
parameters in presence of the coupling terms yields
dE
=0 (10.18)
dt
dA1 Q0,3,1
= −2K A2 B1 (1) (10.19)
dt I0,4,0

dB1 2 A2 Q0,1,1 Q0,3,1
= 2K B1 (1)
− (1) (10.20)
dt A1 I0,2,0 I0,4,0
dκ1
=0 (10.21)
dt
dx1 B1 Q1,1,1
= −κ1 + 2K 2 (1) (10.22)
dt A1 I0,2,0
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
dθ1 κ2 B1 A2 P0,1,1 κ1 B1 Q1,1,1

=− 1 +K (1)
− 2K 2 (1)
dt 2 A1 I0,2,0 A1 I0,2,0
∞
1
+ (1) F A21 g 2 (τ1 ) g 2 (τ1 )dτ1 (10.23)
I0,2,0 −∞
Similarly, the parameters for the solitons in the second core, as described by
(10.2), are defined as
(2) ∞ 12
I0,2,0 −∞ |r |4 dx
A2 (t) = (2) ∞
(10.24)
I0,4,0 −∞ |r |2 dx
(2) ∞
12
−∞ |r | dx
2
I2,2,0
B2 (t) = (2) ∞
(10.25)
−∞ x |r | dx
I0,2,0 2 2
∞
i −∞ (rr x∗ − r ∗r x )dx
κ2 (t) = ∞ (10.26)
−∞ |r | dx
2 2
∞
x|r |2 dx
x2 (t) = −∞∞ (10.27)
−∞ |r | dx
2
These definitions also then lead to the following parameter dynamics for the
solitons in twin-core couplers:
dA2 Q0,3,1
= −2K A1 B2 (2) (10.28)
dt I0,4,0

dB2 2 A1 Q0,1,1 Q0,3,1
= 2K B2 (2)
− (2) (10.29)
dt A2 I0,2,0 I0,4,0
dκ2
=0 (10.30)
dt
dx2 B2 Q1,1,1
= −κ2 + 2K 2 (2) (10.31)
dt A2 I0,2,0
dθ2 κ2 B2 A1 P0,1,1 κ2 B2 Q1,1,1

=− 2 +K (2)
− 2K 2 (2)
dt 2 A2 I0,2,0 A2 I0,2,0
∞
1
+ (2) F A22 g 2 (τ2 ) g 2 (τ2 )dτ2 (10.32)
I0,2,0 −∞
Here, (10.23) and (10.32) are obtained by differentiating (10.6) and (10.7)
with respect to t and subtracting from its conjugate. Thus, from (10.18) one
can see that the total energy in the two couplers remains a constant. How-
ever, the amplitude, width, frequency, center of mass, and phase of the soli-
tons undergo a change as governed by (10.19)–(10.23) and (10.28)–(10.32),
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
respectively. One can apply the new results developed in this section to ob-
tain the parameter dynamics of the solitons in the cores with a particular kind
of nonlinearity, namely F, and the corresponding form of the soliton g.
10.3 Multiple-Core Couplers

The arrays of active or passive coupled waveguides can be utilized for several
practical applications. Multiple-core optical fibers are an area of current re-
search interest. The use of these fibers for high-powered lasers is driving
this interest. One possibility of applications would be all-optical switching
between the cores. The equations derived in this chapter allow us to explore
this possibility. This is one of the major reasons why multiple-core fibers are
studied in optics. Multiple-core couplers can be arranged in rib or ring geom-
etry as shown in Figures 10.3(a) and 10.3(b).
Figure 10.3(a) shows the rib geometry while Figure 10.3(b) shows the ring
geometry for multiple-core fibers. In this section, (10.8), which is now valid
for 1 ≤ l ≤ N, shall be reused, and the following integrals will also be utilized
∞
(µ,ν)
Ia ,b,c = x a g b [Bµ (x − xµ (t)]g c [Bν (x − xν (t)]
−∞
cos[(κµ − κν )x − (κµ xµ − κν xν ) − (θµ − θν )] dx (10.33)

∞
(µ,ν)
J a ,b,c = x a g b [Bµ (x − xµ (t)]g c [Bν (x − xν (t)]
−∞
sin[(κµ − κν )x − (κµ xµ − κν xν ) − (θµ − θν )] dx (10.34)
for 1 ≤ µ, ν ≤ N and nonnegative integers a , b, and c and τl = Bl (t)(x − xl (t)).
(a) (b)
FIGURE 10.3
(a) and (b) Rib and ring geometry.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
10.3.1 Coupling with Nearest Neighbors

This situation can be described by the N-coupled NLSE with the nearest
neighbor linear coupling. They can be used as a more elaborate switching
device. For multiple-core nonlinear fiber arrays, the governing equations for
the ring geometry are given by [79]
1 (l)
t + q xx + F (|q | )q
iq(l) = K (q (l−1) + q (l+1) − 2q (l) )
(l) 2 (l)
(10.35)
2
where 2 ≤ l ≤ N q (0) = q ( N) , q ( N+1) = q (1) . Here, once again, K represents
the coupling coefficient between the neighboring cores and q (l) represents the
optical field in the lth core. Also, (10.35) does not have infinitely many con-
servation laws. In this case, there are also as few as three integrals of motion.
They are energy (E), linear momentum (M), and the Hamiltonian (H) that
N
∞ (l) 2
E= q dx (10.36)
l=1 −∞

i
N ∞ (l) (l)∗
M= q q x − q (l)∗ q x(l) dx (10.37)
2 l=1 −∞
and
N

∞
1 (l) 2 (l) 2

(l)

(l−1) 2
H= q − f q − K q −q dx (10.38)
l=1 −∞ 2 x
Physically, E in (10.36) represents the sum of the individual energies in the N

cores of the fiber. Since (10.35) is not solvable by the inverse scattering trans-
form (IST) or any other method, it will be studied from a different perspective.
For K = 0, the pulse in the lth core is written as
q (l) (x, t) = Al (t)g [Bl (t) (x − xl (t))] exp [−iκl (t)(x − xl (t)) + iθl (t)] (10.39)
where 1 ≤ l ≤ N. Here also, just as in the case of twin-core couplers, Al

represents the amplitude of the soliton, Bl represents the width of the soliton,
xl is the center position, κl is the frequency, θl is the phase, and g represents
the functional form of the soliton, which depends on the type of non–Kerr law
nonlinearity. The notation τl = Bl (x − xl (t)) for 1 ≤ l ≤ N will be used. For
such a general form of the soliton, the three integrals of motion are respectively
given by
N

∞ (l) 2 N
Al2 (l)
E= q dx = I (10.40)
l=1 −∞ l=1
Bl 0,2,0

i
N N
∞ (l) (l)∗ A2 (l)
M= q q x − q (l)∗ q x(l) dx = − κl l I0,2,0 (10.41)
2 l=1 −∞ l=1
Bl
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
and
N

∞
1 (l) 2 (l) 2

(l)

(l−1) 2
H= q − f q − K q −q dx
l=1 −∞ 2 x
N ∞ I
1 2 (l) κ 2 A2 (l)
= Al Bl J 0,2,0 + l l I0,2,0 −N F (s)dsdx
2 l=1 Bl −∞ 0
N 2

A (l) A2 (l−1) (l,l−1)
−K l
I0,2,0 + l−1 I0,2,0 + 4Al Al−1 I0,1,1 (10.42)
l=1
Bl Bl−1
For the case of multiple cores, the definitions of the soliton parameters are as
follows:
∞ (l) 4 12
(l)
I0,2,0 dx
−∞ q
Al (t) = ∞ 2 (10.43)
(l)
I0,4,0 q (l) dx
−∞

(l) ∞ (l) 2
12
I2,2,0 −∞ q dx
Bl (t) = (l) ∞
(10.44)
I0,2,0 2 (l) 2 dx
−∞ x q
∞
i −∞ q (l) q x(l)∗ − q (l)∗ q x(l) dx
κl (t) = ∞ 2 (10.45)
2 q (l) dx
−∞
and
∞ (l) 2
dx
−∞ x q
xl (t) = ∞ 2 (10.46)
(l) dx
−∞ q
These definitions are valid for 1 ≤ l ≤ N. Now, differentiating these para-

meters with respect to t and using (10.39) while treating the coupling terms
as perturbation terms, the following evolution equations for the soliton para-
meters in multiple cores are obtained
dE
=0 (10.47)
dt
dAl 2K Bl (l,l−1) (l,l+1)

= − (l) Al−1 J 0,3,1 + Al+1 J 0,3,1 (10.48)
dt I0,4,0
dBl 2K Bl2 1 (l,l−1) (l,l+1)

= (l)
A l−1 J 0,1,1 + A l+1 J 0,1,1
dt Al I0,4,0

(l,l−1) (l,l+1)
− 2 Al−1 J 0,3,1 + Al+1 J 0,3,1 (10.49)
dκl
=0 (10.50)
dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
dxl 2K Bl
(l,l−1) (l,l+1)

= −κ1 − (l) Al−1 J 1,1,1 + Al+1 J 1,1,1 (10.51)
dt I0,2,0 Al
∞
dθl κ2 I (l) Bl2 1
= − 1 + 0,0,2(l)
− (l)
F Al2 g 2 (τl ) g 2 (τl ) dτl
dt 2 I0,2,0 2 I0,2,0 −∞
2K κl Bl

(l,l−1) (l,l+1)
+ (l) Al−1 J 1,1,1 + Al+1 J 1,1,1
I0,2,0 Al
K Bl
(l,l−1) (l,l+1)

− (l) Al−1 I0,1,1 + Al+1 I0,1,1 (10.52)
I0,2,0 Al
where 1 ≤ l ≤ N. From (10.47), one can conclude that the total energy in
the couplers remains a constant. However, the amplitude, width, frequency,
center of mass, and phase of the solitons in the couplers experience a change as
governed by (10.48)–(10.52). If, however, the law of coupling with the nearest
neighbor is
1 (l) (l) 2 (l)
q = K q (l−1) + q (l+1)
t + q xx + F q
iq(l) (10.53)
2
the same parameter dynamics appear and other features also stay the same.
However, the Hamiltonian, in this case, is given by
N

∞
1 (l) 2 2
H= q x − f q (l)
l=1 −∞
2

− K q (l) q (l−1)∗ + q (l)∗ q (l−1) − K q (l) q (l+1)∗ + q (l)∗ q (l+1) dx

1 2 1 κl2 Al2 (l)
N N ∞ I
(l)
= Al Bl I0,0,2 + I −N F (s)dsdx
2 l=1 2 l=1 Bl 0,0,2 −∞ 0

N
(l,l−1) (l,l+1)
+ 2K Al Al−1 I0,1,1 + Al+1 I0,1,1 (10.54)
l=1
10.3.2 Coupling with All Neighbors

For pulses propagating through N coupled nonlinear fiber arrays, when cou-
pling with all neighbors, the equation is given in the following dimensionless
form
1 (l) (l) 2 (l) N
iq(l) + q + F q q = λln q (n) (10.55)
t
2 xx n=1
where 1 ≤ l ≤ N. In (10.55), the right-hand side is due to the linear coupling

term between the fiber arrays and λi j is the linear coupling coefficient for 1 ≤
i, j ≤ N. The coupling coefficients form a symmetric matrix of N rows and
N columns. So, = (λi j ) N×N where λii = 0 for 1 ≤ i ≤ N and λi j = λ ji ∀ i, j.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
For example, in a twin-core fiber array

0 λ12
= (10.56)
λ12 0
while for a triple-core fiber array
⎡ ⎤
0 λ12 λ13
⎢ ⎥
= ⎣ λ12 0 λ23 ⎦ (10.57)
λ13 λ23 0
and so on. The functional differential-difference equation (10.55) is not inte-

grable by the IST due to the presence of the coupling coefficients and because
of the arbitrary function F . For λln = 0, the pulse in the lth core is given by
(10.39). Moreover, (10.55) does not have infinitely many conserved quanti-
ties. In fact, in this case also, it has as few as three integrals of motion. They
are energy (E), linear momentum (M), and the Hamiltonian (H). Although
the energy and momentum stay the same as in the previous subsection, the
Hamiltonian here is given by
N ∞

1 (l) 2 (l) 2 N
(n) (n−1)∗
H= q
− f q − λln q q −q q
(n)∗ (n−1)
dx
l=1 −∞
2 x n=1
N
1 2 N N
(l) κ 2 A2 (l) (n,n−1)
= Al Bl J 0,2,0 + l l I0,2,0 −2 λln An An−1 I0,1,1
2 l=1 Bl l=1 n=1
∞ I
−N F (s)dsdx (10.58)
−∞ 0
where q (0) = q ( N) . The soliton parameters are defined as in (10.43)–(10.46).

Again, differentiating these parameters with respect to t and using (10.39)
while treating the coupling terms as perturbation terms, the following evo-
lution equations for the soliton parameters in multiple cores are obtained
dE N N
(l,n)
=2 Al An J 0,1,1 (10.59)
dt l=1 n=1
2Bl
N
dAl (n,l)
= (l) λln An J 0,3,1 (10.60)
dt I0,4,0 n=1
2 1
N
dBl (n,l)
= (l)
λln An J 0,1,1 (10.61)
dt Al I0,2,0 n=1
dκl
=0 (10.62)
dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
2 Bl
N
dxl (n,l)
= −κ1 + (l) λln An J 1,1,1 (10.63)
dt I0,2,0 Al n=1
J (l) B 2 1 Bl
N
dθl κ2
= − 1 − 0,2,0 l − (l) (l,n)
λln An I0,1,1
dt 2 I0,2,0 2 I0,2 Al n=1

1 ∞
+ (l)
F Al2 g 2 (τl ) g 2 (τl )dτl (10.64)
I0,2,0 −∞
where 1 ≤ l ≤ N. Thus, from (10.59)–(10.64), it is seen that the energy,

amplitude, width, and phase of the solitons in each array changes in the
presence of the coupling terms; however, the frequency in each array stays
constant.
10.4 Magneto-Optic Waveguides

Optical waveguides, which are made from thin garnet films, have been of
interest since they were first demonstrated in 1972. In addition, it is often
argued that the existence of magnetostatic waves implies devices that have
considerable advantages over acousto-optic applications since garnet films
operate well into the high gigahertz frequency range. This is an important
microwave frequency range, and garnet films offer the added flexibility of
magnetic tunability. Indeed, a whole range of microwave signal processing
devices, such as filters, correlators, spectrum analyzer switches, modulator
frequency shifters, and tunable filters, are either in use or appear to be on the
horizon. Such devices will be even more useful if power can be added in as
another degree of flexibility. In other words, the study of nonlinear magneto-
optical interactions is of prime importance. It is now easy to make the fun-
damental integrated optical building block, which is the channel waveguide.
It is also possible to envision the future integration of magneto-optic devices
with semiconducting substrates containing active devices, such as laser de-
tectors and amplifiers. It would appear, then, that there has never been a
better time to pursue the combination of nonlinearity and magneto-optics.
This is true even though magneto-optics have recently been described as a
“stepchild” of integrated optics. Some of this impression originates from a de-
sire to “insert” magneto-optics into known designs rather than address and
control the fascinating complexity of the materials. The manner in which an
external magnetic field, applied to a waveguide containing third-order opti-
cally nonlinear material (χ (3) ) and magneto-optical elements can force bright
solitons from a state of attraction to a state of isolation from each other [84].
Until recently, any discussion on envelope solitons in nonlinear optics was
based upon the usage of third-order nonlinearity. The general form of the
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
polarization induced by a relatively high-power electromagnetic wave, dur-

ing its passage through a dielectric material, however apart from its
obvious dependence on χ (1) , depends also on χ (2) as well as χ (3) . Truncation
of polarization at the third order is appropriate for most materials and, fur-
thermore, many elements of these tensors are usually zero. This means that
tensors are often reducible to a single, independent parameter or, at most, a
small number of independent ones, by application of crystal symmetry oper-
ations. Indeed the materials needed for real applications are often isotropic
amorphous or possess a rather high crystal symmetry. In an χ (2) material,
however, two field components can mix together to produce a third one that
is, once again, at the fundamental frequency—in other words, back-mixing
(cascading) occurs involving second-harmonic waves and the complex con-
jugate of the fundamental wave. Furthermore, experimental evidence shows
that the back-mixing process is clearly observable, even for large linear phase
mismatching.
10.4.1 Mathematical Analysis

The dimensionless form of the vector NLSE in a magneto-optic waveguide,
due to the polarization control by the magnetic field, is given by [84]
1
iut + uxx + {F (|u|2 ) + α F (|v|2 )}u = Qv (10.65)
2
1
ivt + vxx + {F (|v|2 ) + α F (|u|2 )}v = Qu (10.66)
2
Here Q = Q(x) is called the magneto-optical parameter and α is a constant.

These equations are derived for the case of Kerr law nonlinearity in a χ (3)
medium [298]. Equations (10.65) and (10.66) are not integrable. Therefore,
these equations will be studied by the aid of soliton perturbation theory.
Assume that the solitons in the two waveguides are respectively given by
(10.6) and (10.7). For convenience, the following integrals are defined
∞
G l,m,n = Q(x)xl g m (τ1 )g n (τ2 ) sin{κ1 (x − x1 ) − κ2 (x − x2 ) + (θ1 − θ2 )}dx
−∞
(10.67)
∞
Hl,m,n = Q(x)xl g m (τ1 )g n (τ2 ) cos{κ1 (x − x1 ) − κ2 (x − x2 ) + (θ1 − θ2 )}dx
−∞
(10.68)
for nonnegative integers l, m, and n, while k = 1, 2, where τk = Bk (t)(x −

xk (t)). For such general forms of the solitons given by (10.6) and (10.7), the
energy is given by (10.11). With the same definition of the soliton parameters
as in (10.14)–(10.17), the parameter dynamics, on treating the cross-phase
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
modulation term and the magneto-optical parameter term as perturbation

terms, are given by
dE
=0 (10.69)
dt
dA1 G 0,3,1
= 2A2 B1 (1) (10.70)
dt I0,4,0
dB1 A2 G 0,1,1
=2 (1)
(10.71)
dt A1 I0,2,0
∞
dκ1 2α
= (1) g 2 (τ1 )g(τ2 )g (τ2 ) F ( A22 g 2 (τ2 )) dx (10.72)
dt I0,2,0 −∞
dx1 A2 B1 G 1,1,1
=2 (1)
(10.73)
dt A1 I0,2,0
and
dθ1 κ2 A2 B1 G 1,1,1
= − 1 − 2κ1 (1)
+ α F A22 g 2 (τ2 )
dt 2 A1 I0,2,0
∞
1 A2 B1 H0,1,1
+ (1) g 2 (τ1 ) F A21 g 2 (τ1 ) dτ1 − (1)
(10.74)
I0,2,0 −∞ A1 I0,2,0
Similarly, using the definitions given by (10.24)–(10.27), the evolution

equations of the soliton parameters are
dA2 G 0,3,1
= 2A1 B2 (2) (10.75)
dt I0,4,0
dB2 A1 G 0,1,1
=2 (2)
(10.76)
dt A2 I0,2,0
∞
dκ2 2α
= (2) g 2 (τ2 )g(τ1 )g (τ1 ) F A21 g 2 (τ1 ) dx (10.77)
dt I0,2,0 −∞
dx2 A1 B2 G 1,1,1
=2 (2)
(10.78)
dt A2 I0,2,0
and
dθ2 κ2 A1 B2 G 1,1,1
= − 2 − 2κ2 (2)
+ α F ( A21 g 2 (τ1 ))
dt 2 A2 I0,2,0
∞
1 A1 B2 H0,1,1
+ (2) g 2 (τ2 ) F A22 g 2 (τ2 ) dτ2 − (2)
(10.79)
I0,2,0 −∞ A2 I0,2,0
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Thus, from (10.69), the energy of the soliton remains constant as it is a con-
served quantity for the set of equations in (10.65) and (10.66). However, the
amplitude, width, frequency, center of mass, and phase of the solitons un-
dergo a change as governed by (10.70)–(10.74) and (10.75)–(10.79), respec-
tively. These are the parameter dynamics for the solitons due to non–Kerr
law nonlinearity in a magneto-optic waveguide.
Exercises
1. For twin-core couplers, prove that the three integrals of motion
given by (10.3), (10.4), and (10.5) respectively reduce to (10.11),
(10.12), and (10.13) for the pulses given by (10.6) and (10.7) in the
two couplers.
2. For multiple-core couplers, with nearest neighbor coupling, prove
that the three integrals of motion given by (10.36), (10.37), and (10.38)
respectively reduce to (10.40), (10.41), and (10.42) for pulse in the
lth core given by (10.39).
3. For multiple-core couplers, with all neighbors coupling, prove that
the Hamiltonian is given by (10.58).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
11
Optical Bullets
This chapter gives a detailed overview of the theory of optical bullets that
is due the nonlinear Schrödinger’s equation (NLSE) in multidimensions.
Section 11.1 presents an introduction to optical bullets, including their physics.
Section 11.2 deals with the mathematical issues of optical bullets, including
the structure of the NLSE and its conserved quantities.
11.1 Introduction
First conceived by Silberberg in 1990, an optical bullet is an optical pulse
localized in space and time that maintains its spatiotemporal shape while
propagating. Formation of optical bullets requires an anomalous dispersive
medium where it is possible to achieve exact balance of diffraction and disper-
sion with optical nonlinearity. Dispersion and diffraction attempt to spread
the pulse in longitudinal and transverse directions, respectively. Conversely,
self-focusing effects attempt to compress the pulse. Consequently, a self-
trapped optical soliton could be formed when these opposing forces balance
each other. It has been realized that the solitary wave solutions of the cubic
NLSE are unstable in more than one transverse dimension, hence standard
nonlinear optical materials wherein the refractive index is strictly propor-
tional to the intensity of light do not allow stable light bullets. Therefore, all
theoretical and experimental investigations have been carried out in media
that deviate from Kerr nonlinearity; particularly it has been shown theoret-
ically that optical bullets are stable in saturable media. The mathematical
model that can describe an optical bullet is either the two- or three-dimensional
modified NLSE (mNLSE), depending on whether diffraction is limited to one
or two transverse dimensions. Although the mNLSE permits the propagation
of stable light bullets, these are not true solitons in a strict mathematical sense.
The reason is that, unlike 1 + 1 dimension solitons in Kerr media, which sur-
vive collisions with no loss of energy, these light bullets are found to possess
less energy after collisions. Moreover, in favorable situations, they may lead
to a new phenomenon, such as soliton fusion, fission, soliton tunneling, and
161
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
formation of spiraling light bullets. Nevertheless, in physics literature and

despite protestation of a few pursuits, the pragmatic use of the term solitons
for these robust pulses has become common.
11.2 1 + 3 Dimensions
The dynamics of optical bullets are governed by the NLSE in 1+3 dimensions.
For general non-Kerr law, the NLSE in 1 + 3 dimensions in the dimensionless
form is given by
1
iqt + ∇ 2 q + F (|q |2 )q = 0 (11.1)
2
where
∂2 ∂2 ∂2
∇2 ≡ + +
∂x2 ∂ y2 ∂z2
and F is a real valued algebraic function. In general, (11.1) is not integrable.

This nonintegrability is not necessarily related to the nonlinear term in (11.1).
11.2.1 Integrals of Motion

Equation (11.1) has several symmetries in 1 + 2 dimensions, including rota-
tion, dilatation, the Galelian transformations, and the Talanov lens transfor-
mations. On using the symmetry reductions, the exact solutions to (11.1) in
1 + 2 dimensions were found for Kerr law case, in terms of Jacobi’s elliptic
functions. In addition, there also exists exact solutions for the self-focusing
and the self-defocusing cases that are multivalued at each point of real space.
Moreover, nonstationary singular solutions were also obtained to (11.1).
In fact, for (11.1) there exists at least four integrals of motion. They are
∞ ∞ ∞
E= |q |2 d xd ydz (11.2)
−∞ −∞ −∞
∞ ∞ ∞
i
P= (q ∇q ∗ − q ∗ ∇q )d xd ydz (11.3)
2 −∞ −∞ −∞
∞ ∞ ∞
i
M= r × (q ∇q ∗ − q ∗ ∇q )d xd ydz (11.4)
2 −∞ −∞ −∞
and
∞ ∞ ∞
1
H= (∇q ) · (∇q ∗ ) − f (|q |2 ) d xd ydz (11.5)
−∞ −∞ −∞ 2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Optical Bullets 163
In (11.4), r represents the position vector, namely r = (x, y, z). The integrals
of motion are respectively known as energy (E), linear momentum (P), angu-
lar momentum (M), and the Hamiltonian (H). Also, the conserved quantity
related to the Talanov lens transformation was found in 1985. It is important
to remark here that in the special case of power law nonlinearity, an additional
conserved quantity exists that is given by [361]
∞ ∞ ∞
4t 2
C= |rq + 2it∇q | − 2
|q | 2 p+2
dx dy dz (11.6)
−∞ −∞ −∞ p+1
In fact, it was first pointed out by Kuznetsov and Turitsyn in 1985 that the
invariant (C) in (11.6) is a consequence of Noether’s theorem.
Since there is no inverse scattering solution to (11.1), the dynamics of
parameters of an optical bullet will be derived from their corresponding
definitions. For this, it will be assumed that the solution of (11.1), although
not integrable, is given in the form [78]
q (x, y, z; t) = A(t)g[B1 (t){x − x̄(t)}, B2 (t){y − ȳ(t)}, B3 (t){z − z̄(t)}]
exp[−iκ1 (t){x − x̄(t)} − iκ2 (t){y − ȳ(t)} − iκ3 (t){z − z̄(t)} + iθ(t)] (11.7)
where g represents the functional form of the shape described by the NLSE
and depends on the type of nonlinearity in (11.1). Also in (11.7), A(t) represents
the amplitude, while B j (t) for j = 1, 2, 3 represents the width in the x, y,
and z directions, respectively. Then, κ j (t) ( j = 1, 2, 3) as the frequency of
the soliton in the x, y, and z directions, respectively. Finally, ( x̄(t), ȳ(t), z̄(t))
is the coordinate of the center of mass and θ(t) represents the phase. For
convenience, the following integral is defined
∞ ∞ ∞
α,β,γ
Ia ,b,c, p = (τ1 ) a (τ2 ) b (τ3 ) c g p (τ1 , τ2 , τ3 )
−∞ −∞ −∞
α β γ
∂g ∂g ∂g
· dτ1 dτ2 dτ3 (11.8)
∂τ1 ∂τ2 ∂τ3
with nonnegative integers a , b, c, p, α, β, and γ where τ1 = B1 (t)(x − x̄(t)),

τ2 = B2 (t)( y − ȳ(t)) and τ3 = B3 (t)(z − z̄(t)). For such a general form of the
bullet, given by (11.7), the integrals of motion from (11.2), (11.3), (11.4), and
(11.5) respectively simplify to
∞ ∞ ∞
A2
E= |q |2 d xd ydz = I 0,0,0 (11.9)
−∞ −∞ −∞ B1 B2 B3 0,0,0,2
∞ ∞ ∞
i A2
P= (q ∇q ∗ − q ∗ ∇q ) d xd ydz = − I 0,0,0 (κ1 , κ2 , κ3 )
2 −∞ −∞ −∞ B1 B2 B3 0,0,0,2
(11.10)
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book

i ∞ ∞ ∞
M= r × q ∇q ∗ − q ∗ ∇q d xd ydz
2 −∞ −∞ −∞

A 2
κ2 0,0,0 κ3 0,0,0
= I − I ,
B1 B2 B3 B3 0,0,1,2 B2 0,1,0,2

κ3 0,0,0 κ1 0,0,0 κ1 0,0,0 κ2 0,0,0
I1,0,0,2 − I0,0,1,2 , I0,1,0,2 − I
B1 B3 B2 B1 1,0,0,2
(11.11)
and

∞ ∞ ∞
1
H= (∇q ) · ∇q ∗ − f (|q |2 ) d xd ydz
−∞ −∞ −∞ 2
A 2
2,0,0 0,2,0 0,0,2 0,0,0
= B12 I0,0,0,0 + B22 I0,0,0,0 + B32 I0,0,0,0 + κ12 + κ22 + κ32 I0,0,0,2
2B1 B2 B3
∞ ∞ ∞ I
− f (s) dsd xdydz (11.12)
−∞ −∞ −∞ 0
where the intensity I is given by |q |2 .
11.2.2 Parameter Evolution

There are ten parameters for the case of optical bullets. They are the amplitude
A(t), the width B j (t) ( j = 1, 2, 3) in the x, y, and z directions, the frequencies
κ j (t) in the three directions, and the coordinates of the center of mass of the
soliton that is given by (x̄, ȳ, z̄). These parameters are defined as follows

0,0,0 ∞ ∞ ∞ 12
I0,0,0,2 −∞ −∞ −∞ |q |4 d xd ydz
A(t) = 0,0,0 ∞ ∞ ∞
(11.13)
I0,0,0,4 −∞ −∞ −∞ |q |2 d xd ydz
The definition of the width of the bullet in the x, y, and z directions are
respectively defined as

0,0,0 ∞ ∞ ∞
12
−∞ −∞ −∞ |q | d xd ydz
2
I2,0,0,2
B1 (t) = 0,0,0 ∞ ∞ ∞
(11.14)
−∞ −∞ −∞ τ1 |q | d xd ydz
2 2
I0,0,0,2

0,0,0 ∞ ∞ ∞
12
−∞ −∞ −∞ |q | d xd ydz
2
I0,2,0,2
B2 (t) = 0,0,0 ∞ ∞ ∞
(11.15)
−∞ −∞ −∞ τ2 |q | d xd ydz
2 2
I0,0,0,2
and

0,0,0 ∞ ∞ ∞
12
−∞ −∞ −∞ |q | d xd ydz
2
I0,0,2,2
B3 (t) = 0,0,0 ∞ ∞ ∞
(11.16)
−∞ −∞ −∞ τ3 |q | d xd ydz
2 2
I0,0,0,2
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Optical Bullets 165
The frequency components in the x, y, and z directions are respectively

defined as
∞ ∞ ∞
i −∞ −∞ −∞ q q x∗ − q ∗ q x d xd ydz
κ1 (t) = ∞ ∞ ∞ (11.17)
−∞ −∞ −∞ |q | d xd ydz
2 2
∞ ∞ ∞
i −∞ −∞ −∞ q q y∗ − q ∗ q y d xd ydz
κ2 (t) = ∞ ∞ ∞ (11.18)
−∞ −∞ −∞ |q | d xd ydz
2 2
and
∞ ∞ ∞ ∗ ∗
i −∞ −∞ −∞ (q q z − q q z ) d xd ydz
κ3 (t) = ∞ ∞ ∞ (11.19)
−∞ −∞ −∞ |q | d xd ydz
2 2
Finally, the coordinates of the center of mass of the soliton (x̄, ȳ, z̄) are
defined as
∞ ∞ ∞
−∞ −∞ x|q | d xd ydz
2
x̄(t) = −∞
∞ ∞ ∞ (11.20)
−∞ −∞ −∞ |q | d xd ydz
2
∞ ∞ ∞
−∞ −∞ y|q | d xd ydz
2
ȳ(t) = −∞
∞ ∞ ∞ (11.21)
−∞ −∞ −∞ |q | d xd ydz
2
and
∞ ∞ ∞
z|q |2 d xd ydz
z̄(t) = −∞ −∞ −∞
∞ ∞ ∞ (11.22)
−∞ −∞ −∞ |q |2 d xd ydz
Now, differentiating these parameters with respect to t and using (11.7), the
following evolution equations for the soliton parameters are obtained
dE
=0 (11.23)
dt
dA
=0 (11.24)
dt
dB1 d B2 d B3
= = =0 (11.25)
dt dt dt
dκ1 dκ2 dκ3
= = =0 (11.26)
dt dt dt
d x̄
= −κ1 (11.27)
dt
d ȳ
= −κ2 (11.28)
dt
and
d z̄
= −κ3 (11.29)
dt
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Again, differentiating q given by (11.7), partially with respect to t, and sub-

tracting from its conjugate while utilizing (11.1) yields the evolution of the
phase of the optical bullet that is given by
dθ κ 2 + κ22 + κ32 1 2,0,0 0,2,0 0,0,2 1

=− 1 + 0,0,0 B12 I0,0,0,0 + B22 I0,0,0,0 + B32 I0,0,0,0 − 0,0,0
dt 2 2I0,0,0,2 I0,0,0,2
∞ ∞ ∞
. g 2 (τ1 , τ2 , τ3 ) F ( A2 g 2 (τ1 , τ2 , τ3 )) dτ1 dτ2 dτ3 (11.30)
−∞ −∞ −∞
Thus, from (11.23)–(11.26), the energy, amplitude, width, and frequency of the
bullet remain constant. However, the center of mass and phase of the bullet
undergo a change as governed by (11.27)–(11.30).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
12
Epilogue
The history of solitons is very long. It began one day in August 1834, when
Scottish engineer John Scott Russell first observed, albeit accidentally, a large
solitary water wave in a canal near Edinburgh. This unusual solitary wave,
which is a localized large mass of water, traveled more than 2 km without
noticeable decay in height and change of shape. Ten years later, Russell re-
ported this unusual observation to the British Association for the Advance-
ment of Science. The contemporary scientific community did not take note
of it until 1895, when Dutch mathematicians Korteweg and de Vries theo-
retically proved through the Korteweg–de Vries (KdV) equation that solitary
waves were indeed possible. Seven decades then passed, without anything
memorable being reported. Meanwhile, Martin Kruskal and Normal Zabusky
extended the investigation and, in 1965, came out with the surprising result
that interaction of two solitary waves is identical to that of two colliding el-
ementary particles. Thus, they coined the name soliton for such waves. Soon
afterward, using inverse scattering transform theory, Zakharov and Shabat of
the Soviet Union showed that the nonlinear Schrödinger’s equation (NLSE)
also supports the existence of solitons.
Meanwhile, in 1972, A. Hasegawa and F. D. Tappert showed that the NLSE
is the appropriate equation to describe nonlinear optical light pulse prop-
agation through optical fibers. At Bell Labs, L. Mollenaure, R. Stolen, and
J. Gordon were able to observe optical soliton propagation in 1980. At this
stage, a remarkable suggestion was put forward by Hasegawa and Tappert:
the possibility of a soliton-based all-optical communication system. This idea
stimulated many researchers, and numerous research papers in this field
started appearing in different journals. Meanwhile, optoelectronics and op-
tical communication technology progressed rapidly, largely as a result of the
information revolution. Due to rapidly growing demand from the indus-
try, business, commerce, education, entertainment, and government sectors,
larger and larger quantities of information were being gathered, transmitted,
and processed. One important and essential ingredient in this information
revolution was optical fiber communication systems, which became more
and more sophisticated. Coherent systems and soliton-based systems will
167
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
quite soon be contributing to these developments. In order to design such

systems, knowledge in mathematical formalism of solitons is essential.
In view of this background, chapter 1 introduced a brief history of optical
solitons. An abridged description of optical waveguides and fibers seems
to be helpful for physical understanding of the fascinating world of optical
solitons. This is done in a lucid way and requires no prerequisite knowledge of
the topic. The enormous advantages of optical communication where optical
solitons find wide applications was also highlighted.
In this book, fundamental properties of optical solitons in non-Kerr law
media have been described. It is now well established that different types
of optical nonlinearities can be used to prevent longitudinal and transverse
spreading of light pulses leading to optical solitons, which can be used as bits
of information in sequential or parallel processing configurations. Therefore,
in the beginning of chapter 2, the different types of optical nonlinearities
were briefly discussed and their respective analytic forms were introduced.
The most important nonlinearity is Kerr law nonlinearity, in which the refrac-
tive index is proportional to the first power of light intensity. Obviously, this
nonlinearity has drawn wide attention and many researchers have investi-
gated optical solitons in this medium extensively during the last 2 12 decades.
In the hierarchy of order of attention, parabolic nonlinearity, also known as
cubic-quintic nonlinearity, comes next. Whereas Kerr law nonlinearity grows
with intensity, quintic nonlinearity executes slow saturation in case the quintic
term is negative; on the other hand, it shows much faster growth with inten-
sity when this term is positive. However, parabolic nonlinearity with a neg-
ative quintic term, also known as defocusing quintic nonlinearity, is physically
more meaningful. Another important nonlinearity introduced in this chapter
is power law nonlinearity, in which the refractive index is proportional to the
pth power of light intensity. For a special case of p = 1, this type reduces to
Kerr nonlinearity. Dual type power law is another form of nonlinearity that,
in special cases, reduces to power law or parabolic law nonlinearity. The last
form of nonlinearity discussed in this chapter is saturating law nonlinearity.
Obviously, this law leads to saturation of the refractive index at large inten-
sities. This law exhibits bistable solitons that can have different peak powers
but the same widths. The NLSE and its modified form, which describe op-
tical soliton propagation, possess several integrals of motion. A method to
identify a few of them has been identified. The general evolution equations
of four soliton parameters—amplitude (or width), frequency, soliton center,
and phase—have been introduced. The modified evolution equations due
to small perturbations have subsequently been introduced. At this stage, a
method to obtain the quasi-stationary (QS) solution has been described. As
an example, an application of this method was illustrated.
The concept of optical solitons is based on the integrability of the NLSE.
The formation of solitons is a result of balance between group velocity disper-
sion and optical nonlinearity-induced self-phase modulation. Starting from
the first principle, the NLSE was derived for optical fibers in chapter 3.
The integrability of the NLSE was demonstrated by Zhakarov and Shabat
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Epilogue 169
using the famous inverse scattering transform technique. In a lucid way, this
method was introduced and the method for obtaining the one-soliton solu-
tion of the NLSE was discussed. An important property of the NLSE with
Kerr law nonlinearity is that it has an infinite number of degrees of freedom
and, consequently, an infinite number of conserved quantities. In fact, this is
considered to be one of the definitions of complete integrability. Four impor-
tant conserved quantities were outlined: energy(E), linear momentum(M),
the Hamiltonian (H), and the Hamiltonian for the first integrable hierarchy
of the NLSE (H1 ). It should be pointed out that even though the ideal NLSE,
in which dispersion is balanced by Kerr law nonlinearity, is integrable, the
explicit information that can be obtained from the solution is often rather lim-
ited. This situation has prompted an effort to complement the exact analytical
solution method with approximate methods that sacrifice exactness in order
to obtain explicit results and a clearer picture of the properties of the solution.
One such method is the direct variational method based on trial functions.
Variational formalism has been used successfully and extensively by several
authors to address different nonlinear optical problems involving the NLSE
and its modified form. The main advantages of the variational method are
its simplicity and capacity to provide a clear, qualitative picture and good
qualitative results. This powerful method was introduced in chapter 3. In a
real-world situation, in an optical communication system, for example, soli-
ton propagation is affected by different types of perturbation. As a matter of
fact, these perturbations could result in loss of soliton amplitude due to at-
tenuation, higher-order dispersion, self-steepening, and other factors. Due to
the presence of these perturbations, the parameters that characterize a stable
soliton do not remain constant. These soliton parameters are soliton ampli-
tude as well as pulse width, speed as well as frequency, soliton position, and
the soliton phase. Therefore, a model to deal with these perturbations is es-
sential. The modified variational formalism or perturbed variational method
is in many cases a very powerful method to investigate modified NLSEs
(mNLSEs). Thus, a model to deal with these mNLSEs has been introduced at
the end of the discussion on variational formalism. This method enables one
to derive the evolution equations of the soliton parameters outlined earlier.
These evolution equations can be further investigated to reveal soliton prop-
erties. In addition, QS solitons have been identified using the multiple-scales
perturbation method. This technique was first used in the NLSE by Kodama
and Ablowitz in 1981 [228] and was further studied by Biswas, who extended
it to cases of non-Kerr law nonlinearities [81, 82, 86, 88]. When the nature of
a perturbation is Hamiltonian, an additional powerful method to deal with
such problems is the Lie transform. This technique was first applied to inte-
grate the NLSE with Hamiltonian-type perturbations in 1994 by Kodama and
colleagues [241]. This method was briefly introduced at the end of the chapter
and a comparative study of the multiple-scale techniques and Lie transform
technique were succintly discussed.
The dynamics of soliton propagation in power law, parabolic law,
and dual-power law were respectively addressed in chapters 4, 5, and 6.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Mathematical formalism of these chapters is identical, though the results are

different. In each chapter, the propagation dynamics without perturbation
terms was considered first. As usual, several conserved quantities were eval-
uated. In order to offset loss due to attenuation, linear amplification was
introduced. However, since linear amplification leads to unstable growth of
soliton amplitude, saturable amplifiers were introduced to compensate for the
loss. A model with saturation terms included in the NLSE is more satisfactory
from a physical point of view, since stable soliton propagation is ensured in
principle over an infinite propagation distance, including transoceanic dis-
tances. The QS soliton theory of soliton dynamics was invoked to locate fixed
point.
An important class of optical nonlinearity that has drawn much attention
recently is saturable law nonlinearity. Two different forms of such nonlin-
earity are known and studied widely. Both forms reduce to Kerr nonlinear-
ity at low light intensity, whereas at large light intensity, both saturate to a
constant value. In chapter 7, the mNLSE was derived to describe stationary
wave propagation in such a medium and conserved quantities were evalu-
ated. The existence of bistable solitons, which is an exciting feature of such
mNLSEs, was discussed. The mNLSE in saturating media is not integrable,
hence variational formalism was invoked. An example of an arbitrary pulse
was considered and its dynamics were investigated. A second order nonlin-
ear ordinary differential equation for soliton width was derived. At this stage,
a potential formalism to show equivalency of pulse dynamics with that of a
nonlinear oscillator was introduced. The potential has been examined in detail
to extract important features of solitary waves. Conditions for stationary state
were identified, and regions of stable and unstable propagation located. The
theoretical treatment was extended to both lossless and lossy media.
In a real communication system, soliton pulses never travel as lone enti-
ties. In fact, a continuous stream of optical pulses travels along the line. Since
a soliton solution ideally allows only one soliton in (−∞, ∞), an adjacent soli-
ton always modifies the ideal soliton solution. This modification induces an
interaction force between neighboring solitons. Therefore, interaction among
pulses of the same channel, as well as among pulses of different channels, is
extremely important. In fact, strength of this interaction limits transmission
capacity. In order to avoid soliton interaction, two adjacent solitons should
be separated. This requirement is a drawback for soliton communication sys-
tems. In chapter 8, following the footsteps of Karpman and Solovev formalism
that was first introduced in 1981 [213], the quasi-particle theory was devel-
oped to study the soliton–soliton interaction of two neighboring pulses in the
same channel. Both Hamiltonian and non-Hamiltonian perturbations were
considered. For example, the perturbation terms due to linear damping or
gain, saturable amplification, third and fourth order dispersion, band pass
filter, nonlinear dispersion, and self-steeping were incorporated. In order to
avoid repetition of identical mathematical development, instead of consider-
ing three types of nonlinearity in three different chapters, Kerr law, power law,
parabolic law, and dual-power law nonlinearities are treated in this chapter in
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Epilogue 171
a compact form. After developing appropriate dynamical equations for four

important pulse parameters, direct numerical simulation of the NLSE was
carried out. Numerical simulation results have identified certain perturbation
terms that are helpful for enhancement of collision distance. For example, the
introduction of a filter in the NLSE plays this role quite efficiently.
The focus of chapter 9 was the effect of stochastic perturbation of soliton
propagation. In many cases, random noise affects the performance of soliton
propagation. This random noise may arise due to amplified spontaneous
emission from an amplifier. Moreover, random nonuniformities in optical
fibers, such as fluctuations in the values of the dielectric constant, may also
act as sources of random noise. Keeping these factors in mind, the perturbed
NLSE was introduced to incorporate the random effects discussed in this
chapter. The perturbed NLSE was studied using soliton perturbation theory
to derive the corresponding Langevin equations for four types of nonlinearity:
Kerr, power, parabolic- and dual-power law. These Langevin equations were
further pursued to extract system information.
Optical couplers and switches are essential components of optical commu-
nication systems. In chapter 10, such couplers and switches were described.
The coupled NLSE for twin-core fibers was introduced at the beginning of the
chapter. A couple of conserved quantities were identified, parameter dynam-
ics were worked out. Next, multiple cores were introduced. Here, two differ-
ent cases were considered. In one, coupling with nearest neighbors was dis-
cussed, while in the second case, coupling with all neighbors was discussed.
At the end of the chapter, a mathematical theory of the coupled magneto-optic
waveguides was introduced.
The last chapter, chapter 11, talks very briefly about the existence of optical
bullets, which are optical solitons in 1 + 3 dimensions. The NLSE in 1 + 3
dimensions was introduced, its conserved quantities were studied, and the
basic parameter dynamics were obtained.
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Hints and Solutions
Chapter 2
1. Perform the operation q x∗ × ∂∂x (2.1) − q x × ∂∂x (2.23).
2. Differentiate q given by (2.36) with respect to t and subtract it from
its conjugate. Make use of the NLSE given by (2.1).
Chapter 3
1. Use the soliton perturbation theory formulae given by (3.97) and
(3.98). For velocity, use the formula given by (3.104).
Chapter 4
1. Use the formulae for soliton perturbation theory given by (4.21) and
(4.23).
2. Use the formula for the velocity of the soliton given by (4.29).
3. The fixed point is obtained by setting the right-hand sides of the
dynamical system to zero.
Chapter 5
1. Use l’Hospitals’ rule to evaluate the limits.
2
2. For ν = 0, a = 1. Use the fact that
√ 1 + cosh φ = 2 cosh √ φ/2. Also,
make use of the fact that B = A 2. Finally, define A/ 2 as A of the
Kerr law soliton.
173
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Epilogue 174
3. Use the formulae for soliton perturbation theory that are given by
(5.21) and (5.23).
4. Use the formula for the velocity that is given by (5.30).
5. Use Rabbe’s test for the convergence of a series from Real Analysis.
Chapter 6
1. If ν −→ 0, then a −→ 1 in (6.11). Now, use the double angle formula
for the cosh function, as in exercise 5.2.
2. Expand Gauss’ hypergeometric function and take the approxima-
tion as ν −→ 0.
3. Consider both right-hand as well as left-hand limits.
4. Use the soliton perturbation theory formulae given by (6.22)
and (6.24).
5. Use the formula for the velocity given by (6.31).
Chapter 7
1. Use the definition of the Frechet derivative that is given in (2.35).
Chapter 9
1. Use the adiabatic parameter dynamics formulae given by (9.11) and
(9.13), respectively. Note that for Kerr law nonlinearity, E = 2A.
2.
D [−(1−D)|t−t |]
κ(t)κ(t ) = − e − e [−(1−2D)(t+t )−D|t−t |] +O( D2 )
1 − 2D
174
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography
1. F. Abdullaev, S. Darmanyan & P. Khabibullaev. Optical Solitons. Springer Verlag,

New York, NY. (1993).
2. F. Abdullaev. Theory of Solitons in Inhomogenous Media. John Wiley and Sons,
New York, NY. (1994).
3. F. K. Abdullaev & J. Garnier. “Solitons in media with random dispersive pertur-
bations,” Physica D. Vol 134, Issue 3, 303–315. (1999).
4. F. K. Abdullaev, J. C. Bronski & G. Papanicolaou. “Soliton perturbations and the
random Kepler problem,” Physica D. Vol 135, Issue 3-4, 369–386. (2000).
5. M. J. Ablowitz & H. Segur. Solitons and the Inverse Scattering Transform. SIAM,
Philadelphia, PA. (1981).
6. M. J. Ablowitz & P. A. Clarkson. Solitons, Nonlinear Evolution Equations and Inverse
Scattering. Cambridge University Press, Cambridge. (1993).
7. M. J. Ablowitz, B. Prinari & D. Trubatch. Discrete and Continuous Nonlinear
Schrödinger Systems. Cambridge University Press, Cambridge. (2004).
8. M. J. Ablowitz, G. Biondini & L. A. Ostrovsky. “Optical solitons: Perspectives
and applications,” Chaos. Vol 10, Issue 3, 471–474. (2000).
9. M. J. Ablowitz, G. Biondini & S. Blair. “Localized multi-dimensional optical
pulses in non-resonant quadratic materials,” Mathematics and Computers in Sim-
ulation. Vol 56, Issue 6, 511–519. (2001).
10. A. B. Aceves, C. De Angelis, A. M. Rubenchik & S. K. Turitsyn. “Multidimen-
sional soliton in fiber arrays,” Optics Letters. Vol 19, Issue 5, 329–331. (1994).
11. A. B. Aceves, C. D. Angelis, G. Nalesso & M. Santagiustina. “Higher-order effects
in bandwidth-limited soliton propagation in optical fibers,” Optics Letters. Vol 19,
2104–2106. (1994).
12. A. B. Aceves, G. G. Luther, C. D. Angelis, A. M. Rubenchik & S. K. Turitsyn.
“Optical pulse compression using fiber arrays,” Optical Fiber Technology. Vol 1,
Issue 3, 244–246. (1995).
13. A. B. Aceves, C. D. Angelis, G. G. Luther, A. M. Rubenchik & S. K. Turitsyn.
“All-optical-switching and pulse amplification and steering in nonlinear fiber
arrays,” Physica D. Vol 87, 262–272. (1995).
14. A. B. Aceves, G. G. Luther, C. D. Angelis, A. M. Rubenchik & S. K. Turitsyn.
“Energy localization in nonlinear fiber arrays: Collapse-effect compressor,”
Physical Review Letters. Vol 75, Issue 1, 73–76. (1995).
15. A. B. Aceves, C. D. Angelis, T. Peschel, R. Muschall, F. Lederer, S. Trillo &
S. Wabnitz. “Discrete self-trapping, soliton interactions and beam steering in
nonlinear waveguide arrays,” Physical Review E. Vol 53, 1172–1189. (1996).
16. A. B. Aceves. “Optical gap solitons: Past, present and future; theory and exper-
iments,” Chaos. Vol 10, Issue 3, 584–589. (2000).
17. V. V. Afanasjev. “Interpretation of the effect of reduction of soliton interaction
by bandwidth-limited amplification,” Optics Letters. Vol 18, 790–792. (1993).
18. V. V. Afanasjev, J. S. Aitchison & Y. S. Kivshar. “Splitting of high-order spatial
solitons under the action of two-photon absorption,” Optics Communications.
Vol 116, 331–338. (1995).
175
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
19. V. V. Afanasjev, N. N. Akhmediev & J. M. Soto-Crespo. “Three forms of localized

solutions of the quintic complex Ginzburg-Landau equation,” Physical Review
E. Vol 53, Issue 2, 1931–1939. (1996).
20. V. V. Afanasjev & N. N. Akhmediev. “Soliton interaction in nonequilibrium
dynamical systems,” Physical Review E. Vol 53, 6471–6475. (1996).
21. V. V. Afanasjev, P. L. Chu & Y. S. Kivshar. “Breathing spatial solitons in non-Kerr
media,” Optics Letters. Vol 22, Issue 18, 1388–1390. (1997).
22. G. P. Agrawal. Nonlinear Fiber Optics. Elsevier Science, North Holland. (2001).
23. G. P. Agrawal. Applications of Nonlinear Fiber Optics. Elsevier Science, North
Holland. (2001).
24. G. P. Agrawal & R. W. Boyd. Contemporary Nonlinear Optics. Academic Press,
San Diego. (1992).
25. N. N. Akhmediev, V. I. Korneev & R. F. Nabiev. “Modulation instability of the
ground state of the nonlinear wave equation: Optical machine gun,” Optics
Letters. Vol 15, 393–395. (1992).
26. N. N. Akhmediev, V. V. Afanasjev & J. M. Soto-Crespo. “Singularities and spe-
cial soliton solutions of the cubic-quintic complex Ginzburg-Landau equation,”
Physical Review E. Vol 53, 1190–1201. (1993).
27. N. N. Akhmediev & J. M. Soto-Crespo. “Generation of a train of three-
dimensional nonlinear Schrodinger equation,” Physical Review A. Vol 47, Issue 2,
1358–1364. (1993).
28. N. N. Akhmediev & A. Ankiewicz. “Novel soliton states and bifurcation phe-
nomena in nonlinear fiber optics,” Physical Review Letters. Vol 70, Issue 16, 2395–
2398. (1993).
29. N. N. Akhmediev & V. V. Afanasjev. “Novel arbitrary amplitude soliton solu-
tions of the cubic-quintic complex Ginzburg-Landau equation,” Physical Review
Letters. Vol 75, Issue 12, 2320–2323. (1995).
30. N. N. Akhmediev, V. V. Afanasjev & J. M. Soto-Crespo. “Singularities and spe-
cial soliton solutions of the cubic-quintic complex Ginzburg-Landau equation,”
Physical Review E. Vol 53, Issue 1, 1190–1201. (1996).
31. N. N. Akhmediev & A. Ankiewicz. Solitons: Nonlinear Pulses and Beams.
Chapman and Hall, UK. (1997).
32. N. N. Akhmediev, A. Ankiewicz & J. M. Soto-Crespo. “Multisoliton solutions of
the complex Ginzburg-Landau equation,” Physical Review Letters. Vol 79, 4047–
4050. (1997).
33. N. N. Akhmediev. “Spatial solitons in Kerr and Kerr-like media,” Optical and
Quantum Electronics. Vol 30, 535–569. (1998).
34. N. N. Akhmediev, A. Ankiewicz & J. M. Soto-Crespo. “Stable soliton pairs in
optical transmission lines and fiber lasers,” Journal of Optical Society of America
B. Vol 15, 515–523. (1998).
35. N. N. Akhmediev, A. Ankiewicz & R. Grimshaw. “Hamiltonian-versus-energy
diagrams in soliton theory,” Physical Review E. Vol 59, Issue 5, 6088–6096.
(1999).
36. N. N. Akhmediev & A. Ankiewicz. “Multi-soliton complexes,” Chaos. Vol 10,
Issue 3, 600–612. (2000).
37. N. N. Akhmediev, A. Rodrigues & G. Townes. “Interaction of dual-frequency
pulses in passively mode-locked lasers,” Optics Communications. Vol 187, 419–
426. (2001).
38. D. Anderson. “Variational approach to nonlinear pulse propagation in optical
fibers,” Physical Review A. Vol 27, Issue 6, 3135–3145. (1983).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography 177
39. D. Anderson & M. Lisak. “Bandwidth limits due to mutual pulse interaction
in optical soliton communication systems,” Optics Letters. Vol 11, Issue 3, 174–
176. (1986).
40. D. Anderson, M. Lisak, B. Malomed & M. Quiroga-Teixeiro. “Tunneling of an
optical soliton through a fiber junction,” Journal of Optical Society of America B.
Vol 11, 2380–2384. (1994).
41. D. Anderson, M. Lisak & B. A. Malomed. “Three-wave interaction solitons
in a dispersive medium with quadratic nonlinearity,” Optics Communications.
Vol 126, 251–254. (1996).
42. D. Anderson, M. Lisak & A. Berntson. “A variational approach to nonlin-
ear evolution equations in optics,” Pramana. Vol 57, Issue 5 & 6, 917–936.
(2001).
43. R. L. Anderson & N. H. Ibragimov. Lie-Bäcklund Transformations in Applications.
SIAM, Philadelphia, PA. (1979).
44. A. Ankiewicz, N. N. Akhmediev & G. D. Peng. “Stationary soliton states in
couplers with saturable nonlinearity,” Optical and Quantum Electronics. Vol 27,
193–200. (1995).
45. A. Ankiewicz & N. N. Akhmediev. “Analysis of bifurcations for parabolic
nonlinearity optical couplers,” Optics Communications. Vol 124, 95–102.
(1996).
46. D. Artigas, L. Torner & N. N. Akhmediev. “Asymmetrical splitting of higher-
order solitons induced by quintic nonlinearity,” Optics Communications. Vol 143,
Issue 4–6, 322–328. (1997).
47. M. Arumugam. “Optical fiber communication - An overview.” Pramana. Vol 57,
Issue 5 & 6, 849–869. (2001).
48. A. Arraf, H. He & C. M. de Sterke “Deep gratings with a χ (2) nonlinearity:
Analysis and solutions,” Optical Fiber Technology. Vol 5, 223–234. (1999).
49. I. B. Bakholdin. “Solitary waves and the structures of discontinuities in non-
dissipative models with complex dispersion,” Journal of Applied Mathematics and
Mechanics. Vol 67, Issue 1, 43–56. (2003).
50. G. Baldwin. An Introduction to Nonlinear Optics. Kluwer Academic Publishers,
Boston. (1969).
51. O. Bang, J. J. Rasmussen & P. L. Christiansen. “Subcritical localization in the
discrete nonlinear Schrödinger equation with arbitrary power nonlinearity,”
Nonlinearity. Vol 7, Issue 1, 205–218. (1994).
52. O. Bang, P. L. Christiansen, F. If, K. O. Rasmussen & Y. B. Gaididei. “White noise
in the two-dimensional nonlinear Schrödinger equation,” Applicable Analysis.
Vol 57, 3–15. (1995).
53. O. Bang, Y. S. Kivshar & A. V. Buryak. “Bright spatial solitons in defocusing
Kerr media supported by cascaded nonlinearities,” Optics Letters. Vol 22, Issue
22, 1680–1682. (1997).
54. O. Bang, L. Berge & J. J. Rasmussen. “Fusion, collapse and stationary bound
states of incoherently coupled waves in bulk cubic media,” Physical Review E.
Vol 59, Issue 4, 4600–4613. (1999).
55. O. Bang, W. Krolikowski, J. Wyller & J. J. Rasmussen. “Collapse arrest and soliton
stabilization in nonlocal nonlinear media,” Physical Review E. Vol 66, 046619,
4 pages. (2002).
56. B. Basu-Mallick & T. Bhattacharyya. “Algebraic Bethe ansatz for a quantum
integrable derivative nonlinear Schrödinger model,” Nuclear Physics B. Vol 634,
Issue 3, 611–627. (2002).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
57. B. Basu-Mallick & T. Bhattacharyya. “Jost solutions and quantum conserved

quantities of an integrable derivative nonlinear Schrödinger model,” Nuclear
Physics B. Vol 668, Issue 3, 415–446. (2003).
58. T. B. Benjamin. “The stability of solitary waves,” Proceedings of Royal Society of
London A. Vol 328, 153–183. (1972).
59. L. Berge, O. Bang, J. J. Rasmussen & V. K. Mezentsev. “Self-focusing and
solitonlike structures in materials with competing quadratic and cubic
nonlinearities,” Physical Review E. Vol 55, Issue 3, 3555–3570. (1997).
60. X. Bingzhen & W. Wenzheng. “Traveling-wave method for solving the
modified nonlinear Schrödinger equation describing soliton propagation
along optical fibers,” Physical Review E. Vol 51, Issue 2, 1493–1498.
(1995).
61. A. Biswas. “Optical soliton perturbation with bandwidth limited amplification
and saturable amplifiers,” Journal of Nonlinear Optical Physics and Applications.
Vol 8, Issue 2, 277–288. (1999).
62. A. Biswas. “Soliton-soliton interaction in optical fibers,” Journal of Nonlinear
Optical Physics and Materials. Vol 8, Issue 4, 483–495. (1999).
63. A. Biswas. “Integro-differential perturbation of optical solitons,” Journal of Optics
A. Vol 2, Issue 5, 380–388. (2000).
64. A. Biswas. “Perturbation of optical solitons and quasi-solitons,” Journal of Elec-
tromagnetic Waves and Applications. Vol 14, Issue 1, 95–114. (2000).
65. A. Biswas. “Solitons in nonlinear fiber arrays,” Journal of Electromagnetic Waves
and Applications. Vol 15, Issue 9, 1189–1196. (2001).
66. A. Biswas. “Dispersion-managed solitons in multiple-core nonlinear fiber
arrays,” Fiber and Integrated Optics. Vol 20, Issue 6, 571–579. (2001).
67. A. Biswas. “Solitons in multiple-core couplers,” Journal of Nonlinear Optical
Physics and Materials. Vol 10, Issue 3, 329–336. (2001).
68. A. Biswas. “Optical soliton perturbation with nonlinear damping and saturable
amplifiers,” Mathematics and Computers in Simulation. Vol 56, Issue 3, 521–537.
(2001).
69. A. Biswas. “Optical soliton perturbation with higher-order dispersions,” Fiber
and Integrated Optics. Vol 20, Issue 2, 171–189. (2001).
70. A. Biswas. “Optical soliton perturbation with Raman scattering and saturable
amplifiers,” Optical and Quantum Electronics. Vol 33, 289–304. (2001).
71. A. Biswas. “Dynamically stable solitons in optical fibers,” Fiber and Integrated
Optics. Vol 20, Issue 6, 617–624. (2001).
72. A. Biswas. “Perturbation of solitons due to power law nonlinearity,” Chaos,
Solitons and Fractals. Vol 12, Issue 3, 579–588. (2001).
73. A. Biswas & A. B. Aceves. “Dynamics of solitons in optical fibers,” Journal of
Modern Optics. Vol 48, Issue 7, 1135–1150. (2001).
74. A. Biswas. “Perturbation of solitons with non-Kerr law nonlinearity,” Chaos,
75. A. Biswas. “Integro-differential perturbation of non-Kerr law solitons,” Chaos,
76. A. Biswas. “Optical soliton perturbation with Raman scattering and nonlinear
damping,” Fiber and Integrated Optics. Vol 21, 125–143. (2002).
77. A. Biswas. “Multiple-scale analysis for non-Kerr law solitons,” International
Mathematical Journal. Vol 2, Issue 12, 1157–1197. (2002).
78. A. Biswas. “Theory of optical bullets,” Progress in Electromagnetic Research. Vol 36,
21–59. (2002).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography 179
79. A. Biswas. “Theory of optical couplers,” Optical and Quantum Electronics. Vol 35,
Issue 3, 221–235. (2003).
80. A. Biswas. “Dispersion-managed solitons in optical couplers,” Journal of Non-
linear Optical Physics and Applications. Vol 12, Issue 1, 45–74. (2003).
81. A. Biswas. “Quasi-stationary optical solitons with power law nonlinearity,” Jour-
nal of Physics A. Vol 36, Issue 16, 4581–4589. (2003).
82. A. Biswas. “Quasi-stationary optical solitons with parabolic law nonlinearity,”
Optics Communications. Vol 216, Issue 4–6, 427–437. (2003).
83. A. Biswas. “Optical solitons: Quasi-stationarity versus Lie transform,” Optical
and Quantum Electronics. Vol 35, Issue 10, 979–998. (2003).
84. A. Biswas. “Solitons in magneto-optic waveguides,” Applied Mathematics and
Computation. Vol 153, Issue 2, 387–393. (2004).
85. A. Biswas. “Theory of non-Kerr law solitons,” Applied Mathematics and Compu-
tation. Vol 153, Issue 2, 369–385. (2004).
86. A. Biswas. “Quasi-stationary non-Kerr law optical solitons,” Optical Fiber Tech-
nology. Vol 9, Issue 4, 224–259. (2003).
87. A. Biswas. “Adiabatic dynamics of non-Kerr law solitons,” Applied Mathematics
and Computation. Vol 151, Issue 1, 41–52. (2004).
88. A. Biswas. “Quasi-stationary optical solitons with dual-power law nonlinear-
ity,” Optics Communications. Vol 235, Issue 1–3, 183–194. (2004).
89. A. Biswas. “Stochastic perturbation of optical solitons in Schrödinger-Hirota
equation,” Optics Communications. Vol 239, Issue 4–6, 461–466. (2004).
90. A. Biswas. “Optical soliton perturbation with non-Kerr law nonlinearities,”
Progress in Electromagnetic Research. Vol 50, 231–266. (2005).
91. A. B. Blagoeva, S. G. Dinev, A. A. Dreischuh & A. Naidenov. “Light bullets
formation in a bulk media,” IEEE Journal of Quantum Electronics. Vol QE-27,
2060–2062. (1991).
92. N. Bloembergen. Nonlinear Optics. World Scientific Publishing Company, River
Edge, NJ. (1996).
93. K. J. Blow & N. J. Doran. “Bandwidth limits of nonlinear (soliton) optical com-
munication systems,” Electronics Letters. Vol 19, 429–430. (1983).
94. K. J. Blow & N. J. Doran. “Nonlinear limits on bandwidth at the minimum
dispersion in optical fibers,” Optics Communications. Vol 48, 181–184. (1983).
95. K. J. Blow, N. J. Doran & D. Wood. “Suppression of the soliton self-frequency
shift by bandwidth-limited amplification,” Journal of Opt Soc Am B. Vol 5, Issue 6,
1301–1304. (1988).
96. A. D. Boardman, M. Bertolotti & T. Twardowski. Nonlinear Waves in Solid State
Physics. Kluwer Academic Publishers, Boston. (1991).
97. A. D. Boardman, L. Pavlov & S. Tanev. Advanced Photonics with Second-Order
Optically Nonlinear Processes. Kluwer Academic Publishers, Boston. (1998).
98. A. D. Boardman & A. P. Suukhorukov. Soliton-Driven Photonics. Kluwer Aca-
demic Publishers, Boston. (2001).
99. A. D. Boardman & K. Xie “Magneto-optic spatial solitons,” Journal of Optical
Society of America B. Vol 14, Issue 11, 3102–3109. (1995).
100. A. D. Boardman & K. Xie “Spatial bright-dark soliton steering through waveg-
uide coupling,” Optical and Quantum Electronics. Vol 30, Issue 7, 783–794. (1998).
101. C. M. Bowden & C. D. Cantrell. Nonlinear Optics and Materials. SPIE International
Society for Optical Engineering, Bellingham, WA. (1991).
102. J. P. Boyd. Weakly Nonlocal Solitary Waves and Beyond-All-Orders Asymptotics.
Kluwer Academic Publishers, Boston. (1998).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
103. R. W. Boyd. Nonlinear Optics. Academic Press, San Diego, CA. (2003).
104. H. E. Brandt. Selected Papers in Nonlinear Optics. SPIE International Society for
Optical Engineering, Bellingham, WA. (1991).
105. J. C. Bronski. “Nonlinear scattering and analyticity properties of solitons,” Jour-
nal of Nonlinear Science. Vol 8, 161–182. (1998).
106. N. Burq, P. Gerard & N. Tzvedkov. “The Cauchy problem for the nonlinear
Schrödinger equation on a compact manifold,” Journal of Nonlinear Mathematical
Physics. Vol 10, Supplement 1, 12–27. (2003).
107. S. Burtsev, D. J. Kaup & B. A. Malomed. “Interactions of solitons with a strong
inhomogeneity in a nonlinear optical fiber.” Physical Review E. Vol 52, Issue 4,
4474–4481. (1995).
108. V. S. Busalev & V. E. Grikurov. “Simulation of instability of bright solitons for NLS
with saturating nonlinearity,” Mathematics and Computers in Simulation. Vol 56,
Issue 6, 539–546. (2001).
109. V. S. Busalev & C. Sulem. “On asymptotic analysis of solitary waves for nonlinear
Schrödinger’s equation,” Nonlinear Analysis. Vol 20, Issue 3, 419–475. (2003).
110. P. N. Butcher & D. Cotter. The Elements of Nonlinear Optics. Cambridge University
Press, Cambridge. (1991).
111. G. Chaohao. Soliton Theory and its Applications. Springer Verlag, New York, NY.
(1995).
112. D. Cai, A. R. Bishop, N. Groenbech-Jensen & B. A. Malomed. “Stabilizing a
breather in the damped nonlinear Schrödinger equation driven by two frequen-
cies.” Physical Review E. Vol 49, Issue 2, R1000-R1002. (1994).
113. S. Chávez-Cerda, M. A. Meneses-Nava & J. J. Sánchez-Mondragón. “Breathing
pulses and beams in non-linear media,” Optical and Quantum Electronics. Vol 30,
Issue 7–10. (1998).
114. X. J. Chen & J. Yang. “Direct perturbation theory for solitons of the derivative
nonlinear Schrödinger equation and the modified nonlinear Schrödinger equa-
tion,” Physical Review E. Vol 65, 066608. (2002).
115. Y. Chen. “Self-trapped light in saturable nonlinear media,” Optics Letters. Vol 16,
Issue 1, 4–6. (1991).
116. P. L. Christiansen, N. Groenbech-Jensen, P. S. Lomdahl & B. A. Malomed.
“Oscillations of eccentric pulsons,” Physics Scripta. Vol 55, 131–134. (1997).
117. P. L. Chu & C. Desem. “Mutual interaction between solitons of unequal ampli-
tudes in optical fibre,” Electronics Letters. Vol 21, 1133–1134. (1985).
118. P. L. Chu & C. Desem. “Effect of third order dispersion of optical fibre on soliton
interaction,” Electronics Letters. Vol 21, 228–229. (1985).
119. P. L. Chu & B. Wu. “Optical switching in twin-core erbium-doped fibers,” Optics
Letters. Vol 17, 255–257. (1992).
120. P. L. Chu, B. A. Malomed & G. D. Peng. “Soliton switching and propagation in
nonlinear fiber couplers: Analytical results,” Journal of Optical Society of America
B. Vol 10, 1379–1385. (1993).
121. P. L. Chu, B. A. Malomed, G. D. Peng & I. Skinner. “Soliton dynamics in
periodically modulated directional couplers,” Physical Review E. Vol 49, Issue 6,
5763–5767. (1994).
122. P. L. Chu, Y. S. Kivshar, B. A. Malomed, G. D. Peng & M. L. Quiroga-Teixeiro.
“Soliton controlling, switching and splitting in fused nonlinear couplers,” Jour-
nal of Optical Society of America B. Vol 12, Issue 5, 898–904. (1995).
123. P. L. Chu, B. A. Malomed & G. D. Peng. “Soliton amplification and reshaping
in optical fibers with variable dispersion,” Journal of Optical Society of America B.
Vol 13, Issue 8, 1794–1802. (1996).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography 181
124. C. B. Clausen, O. Bang & Y. S. Kivshar. “Spatial solitons and induced Kerr effects
in quasi-phase-matched quadratic media,” Physical Review Letters. Vol 78, 4749–
4752. (1997).
125. F. Cooper, C. Lucheroni & H. Shepard. “Variational method for studying self-
focusing in a class of nonlinear Schrödinger equations,” Physics Letters A. Vol 170,
Issue 3, 184–188. (1992).
126. J. F. Corney & O. Bang. “Solitons in quadratic nonlinear photonic crystals,”
Physical Review E. Vol 64, 047601. (2001).
127. J. F. Corney & O. Bang. “Modulational instability in periodic quadratic nonlinear
materials,” Physical Review Letters. 133901. (2001).
128. J. F. Corney & O. Bang. “Plane waves in periodic, quadratically nonlinear slab
waveguides: Stability and exact Fourier structure,” Journal of Optical Society of
America B. Vol 19, Issue 4, 812–821. (2002).
129. L. C. Crasovan, B. A. Malomed & D. Mihalache. “Spinning solitons in cubic-
quintic nonlinear media,” Pramana, Vol 57, Issue 5 & 6, 1041–1059. (2001).
130. A. Das. Integrable Models. World Scientific Publishing Company, Teaneck, NJ.
(1989).
131. C. De Angelis. “Self-trapped propagation in the nonlinear cubic-quintic
Schrödinger equation: A variational approach,” IEEE Journal of Quantum Elec-
tronics. Vol 30, Issue 3, 818–821. (1994).
132. C. Desem & P. L. Chu. “Soliton interactions in the presence of loss and pe-
riodic amplification in optical fibers,” Optics Letters. Vol 12, Issue 5, 349–351.
(1987).
133. J. A. DeSanto. Mathematical and Numerical Aspects of Wave Propagation. SIAM,
Philadelphia, PA. (1998).
134. A. Desyatnikov, A. Maimistov & B. Malomed. “Three-dimensional spinning
solitons in dispersive media with cubic-quintic nonlinearity,” Physical Review E.
Vol 61, No 3, 3107–3113. (2000).
135. Q. Ding & Z. Zhu. “On the Gauge equivalent structure of the modified nonlinear
Schrödinger equation,” Physics Letters A. Vol 295, 192–197. (2002).
136. E. V. Doktorov. “The modified nonlinear Schrödinger equation: Facts and arte-
facts,” European Physical Journal B. Vol 29, 227–231. (2002).
137. P. G. Drazin & R. S. Johnson. Solitons: An Introduction. Cambridge University
Press, Cambridge. (1992).
138. P. D. Drummond, K. V. Kheruntsyan & H. He. “Coherent molecular solitons in
Bose-Einstein condensates,” Physical Review Letters. Vol 81, Issue 15, 3055–3058.
(1998).
139. P. D. Drummond & H. He. “Optical mesons,” Physical Review A. Vol 56, Issue 2,
1107–1110. (1997).
140. H. S. Eisenberg, Y. Silberberg, R. Morandotti, A. R. Boyd & J. S. Aitchison. “Dis-
crete spatial optical solitons in waveguide arrays,” Physical Review Letters. Vol 81,
Issue 16, 3383–3386. (1998).
141. H. S. Eisenberg, R. Morandotti, Y. Silberberg, J. M. Arnold, G. Penneli & J. S.
Aitchison. “Optical discrete solitons in waveguide arrays. I. Soliton formation,”
Journal of Optical Society of America B. Vol 19, Issue 12, 2938–2944. (2002).
142. J. N. Elgin. “Stochastic perturbations of optical solitons,” Optics Letters. Vol 18,
Issue 1, 10–12. (1993).
143. D. E. Emundson & R. H. Enns. “The particle-like nature of colliding light bullets,”
Physical Review A. Vol 51, Issue 3, 2491–2498. (1995).
144. D. E. Emundson & R. H. Enns. “Bistable light bullets,” Optics Letters. Vol 17,
586–588. (1992).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
145. R. H. Enns & D. E. Emundson. “Guide to fabricating bistable-soliton-supporting

media,” Physical Review A. Vol 47, Issue 5, 4524–4527. (1993).
146. R. H. Enns & S. S. Rangnekar. “Bistable spheroidal optical solitons,” Physical
Review A. Vol 45, Issue 5, 3354–3357. (1992).
147. R. H. Enns & S. S. Rangnekar. “Variational approach to bistable solitary
waves in d dimensions,” Physical Review E. Vol 48, Issue 5, 3998–4007.
(1993).
148. R. H. Enns, S. S. Rangnekar & A. E. Kaplan. “Optical switching between bistable
soliton states: A theoretical review,” Optical and Quantum Electronics. Vol 24,
1295–1314. (1992).
149. C. Etrich, U. Peschel, F. Lederer, B. A. Malomed & Y. S. Kivshar. “Origin of the
persistent oscillations of solitary waves in nonlinear quadratic media,” Physical
Review E. Vol 54, Issue 4, 4321–4324. (1996).
150. M. Evans & S. Kielich. Modern Nonlinear Optics, Vol 1. John Wiley & Sons,
Hoboken, NJ. (1997).
153. L. D. Faddeev & L. A. Takhtajan Hamiltonian Methods in the Theory of Solitons.
Springer Verlag, New York. (1987).
154. R. Fedele, H. Schamel, V. I. Karpman & P. K. Shukla. “Envelope solitons of non-
linear Schrödinger equation with an anti-cubic nonlinearity,” Journal of Physics
A. Vol 36, 1169–1173. (2003).
155. M. F. Ferreira, M. V. Facao & S. V. Latas. “Soliton-like pulses in a system with
nonlinear gain,” Photonics and Optoelectronics. Vol 5, 147–153. (1999).
156. M. F. Ferreira, M. V. Facao & S. V. Latas. “Stable soliton propagation in a system
with spectral filtering and nonlinear gain,” Fiber and Integrated Optics. Vol 19,
Issue 1, 31–41. (2000).
157. M. F. Ferreira & S. V. Latas. “Timing jitter in soliton transmission with up and
down sliding frequency guiding filters,” Journal of Lightwave Technology. Vol 19,
332–335. (2001).
158. M. F. Ferreira & S. V. Latas. “Soliton stability and compression in a system with
nonlinear gain,” Optical Engineering. Vol 41, 1696–1703. (2002).
159. R. A. Fisher & J. F. Reintjes. Nonlinear Optics III. SPIE International Society for
Optical Engineering, Bellingham, WA. (1992).
160. A. S. Fokas & V. E. Zakharov. Important Developments in Soliton Theory. Springer
Verlag, New York. (1993).
161. M. G. Forest, D. W. McLaughlin, D. J. Muraki & O. C. Wright. “Nonfocus-
ing instabilities in coupled, integrable nonlinear Schrödinger pdes,” Journal of
Nonlinear Science. Vol 10, 291–331. (2000).
162. P. L. Francois & T. Georges. “Reduction of averaged soliton interaction forces
by amplitude modulation,” Optics Letters. Vol 18, 583–585. (1993).
163. H. Frauenkron, Y. S. Kivshar & B. A. Malomed. “Multisoliton collisions in nearly
integrable systems,” Physical Review E. Vol 54, R2244–R2247. (1996).
164. L. Gagnon & P. A. Belanger. “Adiabatic amplification of optical solitons,” Physical
Review A. Vol 43, Issue 11, 6187–6193. (1991).
165. S. Gangopadhyay & S. N. Sarkar. “Variational analysis of spatial solitons of
power-law nonlinearity,” Fiber and Integrated Optics. Vol 20, Issue 2, 191–195.
(2001).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography 183
166. S. Gatz & J. Herrman. “Soliton propagation and soliton collision in double-
doped fibers with a non-Kerr-like nonlinear refractive-index change,” Optics
167. T. Georges & F. Favre. “Influence of soliton interaction on amplifier noise-
induced jitter: A first-order analytical solution,” Optics Letters. Vol 16, 1656–1658.
(1991).
168. T. Georges & F. Favre. “Modulation, filtering and initial phase control of inter-
acting solitons,” Journal of Optical Society of America B. Vol 10, 1880–1889. (1993).
169. V. S. Gerdjikov, E. V. Doktorov & J. Yang.”Adiabatic interaction of N ultrashort
solitons: Universality of the complex Toda chain model,” Physical Review E.
Vol 64, 056617. (2001).
170. J. M. Ghidaglia & J. C. Saut. “Nonexistence of travelling wave solutions to nonel-
liptic nonlinear Schrödinger equations,” Journal of Nonlinear Science. Vol 6, Issue
2, 139–145. (1996).
171. C. G. Goedde, W. L. Kath & P. Kumar. “Controlling soliton perturbations with
phase-sensitive amplification,” Journal of Optical Society of America B. Vol 14,
Issue 6, 1371–1379. (1997).
172. P. M. Goorjian & Y. Silberberg. “Numerical simulation of light bullets using
the full vector time dependent nonlinear Maxwell equations,” Journal of Optical
173. J. P. Gordon. “Interaction forces among solitons in optical fibers,” Optics Letters.
Vol 8, 596–598. (1983).
174. J.P. Gordon & H. A. Haus. “Random walk of coherently amplified solitons in
optical fibre transmission,” Optics Letters. Vol 11, 665–667. (1986).
175. J.P. Gordon & L. F. Mollenaeur. “Effects of fiber nonlinearities and amplifier
spacing on ultra long distance transmission,” Journal of Lightwave Technology.
Vol 9, 170–173. (1991).
176. I. S. Gradshteyn & I. M. Ryzhik. Table of Integrals, Series, and Products. Academic
Press, New York. (2000).
177. V. S. Grigoryan. “Autosoliton in a fiber with distributed saturable amplifiers,”
Optics Letters. Vol 21, Issue 23, 1882–1884. (1996).
178. V. S. Grigor’yan, A. I. Maimistov & Y.M. Skylarov. “Evolution of light pulses in
a nonlinear amplifying medium,” Journal of Experimental and Theoretical Physics.
Vol 67, Issue 3, 530–534. (1988).
179. E. M. Gromov. “Short optical solitons in fibers,” Chaos. Vol 10, Issue 3, 551–558.
(2000).
180. Z. Grujic & H. Kalisch. “The derivative nonlinear Schrödinger equation in an-
alytic classes,” Journal of Nonlinear Mathematical Physics. Vol 10, Supplement 1,
67–71. (2003).
181. Y. Guo, C. K. Kao, E. H. Li & K. S. Chiang. Nonlinear Photonics: Nonlinearities
in Optics, Optoelectronics, and Fiber Communications. Springer Verlag, New York.
(2002).
182. E. Hanamurs, Y. Kawabe & A. Yamanaka. Quantum Nonlinear Optics. Springer
Verlag, New York. (2002).
183. A. Hasegawa & F. D. Tappert. “Transmission of stationary nonlinear optical
pulses in dispersive dielectric fibers. I. Anomalous dispersion,” Applied Physics
Letters. Vol 23, 142–144. (1973).
184. A. Hasegawa & F. D. Tappert. “Transmission of stationary nonlinear optical
pulses in dispersive dielectric fibers. I. Normal dispersion,” Applied Physics
Letters. Vol 23, 171–172. (1973).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
185. A. Hasegawa & Y. Kodama. Solitons in Optical Communications, Oxford Univer-

sity Press, Oxford. (1995).
186. A. Hasegawa. Physics and Applications of Optical Solitons in Fibers. Kluwer
Academic Publishers, Boston. (1996).
187. A. Hasegawa. New Trends in Optical Soliton Transmission Systems. Kluwer
188. A. Hasegawa. Massive WDM and TDM Soliton Transmission Sytems. Kluwer
189. A. Hasegawa. “A historical review of application of optical solitons for high
speed communications,” Chaos. Vol 10, Issue 3, 475–485. (2000).
190. A. Hasegawa. “Soliton-based ultra-high-speed optical communications,”
Pramana. Vol 57, Issue 5 & 6, 1097–1127. (2001).
191. A. Hasegawa & M. Matsumoto. Optical Solitons. Springer Verlag, New York.
(2003).
192. A. Hasegawa. “Theory of information transfer in optical fibers: A tutorial
review,” Optical Fiber Technology. Vol 10, Issue 2, 150–170. (2004).
193. H. Hatami-Hansa, P. L. Chu, B. A. Malomed & G. D. Peng. “Soliton compression
and splitting in double-core nonlinear optical fibers,” Optics Communications.
Vol 134, 59–65. (1997).
194. K. Hayata & M. Koshiba. “Solution of self-trapped multidimensional optical
beams by Galerkin’s method,” Optics Letters. Vol 17, 841–843. (1992).
195. K. Hayata & M. Koshiba. “Bright-dark solitary-wave solutions of a multi-
dimensional nonlinear Schrödinger’s equation,” Physical Review E. Vol 48, Issue
3, 2312–2315. (1993).
196. K. Hayata & M. Koshiba. “Algebraic solitary-wave solutions of a nonlinear
Schrodinger’s equation,” Physical Review E. Vol 51, Issue 2, 1499–1502. (1995).
197. G. S. He & S. H. Liu. Physics of Nonlinear Optics. World Scientific Publishing
Company, River Edge, NJ (1998).
198. H. He, M. Friese, N. R. Heckenberg & H. Rubinsztein-Dunlop. “Direct observa-
tion of transfer of angular momentum to absorptive particles from laser beam
with a phase singularity,” Physical Review Letters. Vol 75, Issue 5, 826–829. (1995).
199. H. He, N. R. Heckenberg & H. Rubinsztein-Dunlop. “Optical pulse trapping
with higher-order doughnut beams produced using high-efficiency computer
generated phase holograms,” Journal of Modern Optics. Vol 42, Issue 1, 217–223.
(1995).
200. H. He, M. J. Werner & P. D. Drummond. “Simultaneous solitary-wave solutions
in a nonlinear parametric waveguide,” Physical Review E. Vol 54, 896–899. (1996).
201. H. He, P. D. Drummond & B. A. Malomed. “Modulational stability in dispersive
optical systems with cascaded nonlinearity,” Optics Communications. Vol 123,
395–402. (1996).
202. H. He & P. D. Drummond. “Ideal soliton environment using parametric band-
gaps,” Physical Review Letters. Vol 78, Issue 23, 4311–4314. (1997).
203. H. He, A. Arraf, C. M. de Sterke, P. D. Drummond & B. A. Malomed. “Theory
of modulational instability in Bragg gratings with quadratic nonlinearities,”
204. J. Herrman. “Propagation of ultrashot light pulses in fibers with saturable non-
linearity in the normal-dispersion region,” Journal of Optical Society of America B.
Vol 8, Issue 7, 1507–1511. (1991).
205. B. Hermansson & D. Yevick. “Numerical investigation of soliton interaction in
optical fibers,” Electronics Letters. Vol 19, 570–571. (1983).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography 185
206. E. J. Hinch. Perturbation Methods. Cambridge University Press, Cambridge.

(1991).
207. T. Iizuka & Y. S. Kivshar. “Optical gap solitons in nonresonant quadratic media,”
Physical Review E. Vol 59, 7148–7151. (1999).
208. E. Infield & G. Rowlands. Nonlinear Waves, Solitons and Chaos. Cambridge Uni-
versity Press, Cambridge. (1990).
209. S. Jana & S. Konar “Induced focussing of two laser beams in cubic quintic non-
linear media,” Physica Scripta. Vol 70, 354–360. (2004).
210. Z. Jovanoski & R. A. Sammut. “Propagation of Gaussian beams in a nonlinear
saturable medium,” Physical Review E. Vol 50, Issue 5, 4087–4093. (1994).
211. Z. Jovanoski & D. R. Rowland. “Variational analysis of solitary waves in a ho-
mogeneous cubic-quintic nonlinear medium,” Journal of Modern Optics. Vol 48,
Issue 7, 1179–1193. (2001).
212. V. I. Karpman & E. M. Maslov. “Perturbation theory for solitons,” Journal of
Experimental and Theoretical Physics. Vol 46, 281–291. (1977).
213. V. I. Karpman & V. V. Solov’ev. “A perturbational approach to the two-soliton
systems,” Physica D. Vol 3, Issue 3, 487–502. (1981).
214. V. I. Karpman. “Radiation of solitons described by a high-order cubic non-
linear Schrödinger equation,” Physical Review E. Vol 62, Issue 4, 5678–5687.
(2000).
215. V. I. Karpman & A. G. Shagalov. “Stability of solitons described nonlinear
Schrödinger-type equations with higher-order dispersion,” Physica D. Vol 144,
Issue 1–2, 194–210. (2000).
216. D. J. Kaup. “Perturbation theory for solitons in optical fibers,” Physical Review
A. Vol 42, Issue 9, 5689–5694. (1990).
217. D. J. Kaup & A. C. Newell. “An exact soliton for a derivative nonlin-
ear Schrödinger equation,” Journal of Mathematical Physics. Vol 19, 798–801.
(1978).
218. D. J. Kaup, J. El-Reedy & B. A. Malomed. “Effect of chirp on soliton production,”
219. O. Keller. Notions and Perspectives of Nonlinear Optics. World Scientific Publishing
Company, River Edge, NJ. (1996).
220. I. C. Khoo. Nonlinear Optics and Optical Physics. World Scientific Publishing
221. Y. S. Kivshar & A. M. Kosevich. “Evolution of a soliton under the action of small
perturbations,” JETP Letters. Vol 37, 648–651. (1983).
222. Y. S. Kivshar & B. A. Malomed. “Many-particle effects in nearly integrable sys-
tems,” Physica D. Vol 24, 125–154. (1987).
223. Y. S. Kivshar & V. V. Konotop. “Solitons in fiber-optic wave-guides with slowly
varying parameters,” Kvantovaya Elektron. Vol 16, 868–871. (1989).
224. Y. S. Kivshar & B. A. Malomed. “Dynamics of solitons in nearly integrable sys-
tems” Reviews of Modern Physics. Vol 61, Issue 4, 763–915. (1989).
225. Y. S. Kivshar. “Bright and dark spatial solitons in non-Kerr media,” Optical and
Quantum Electronics. Vol 30, 535–569. (1998).
226. Y. S. Kivshar & B. Luther-Davis. “Dark optical solitons: Physics and applica-
tions,” Physics Reports. Vol 298, 81–197. (1998).
227. Y. S. Kivshar & G. P. Agrawal. Optical Solitons: From Fibers to Photonic Crystals.
Academic Press, Boston. (2003).
228. Y. Kodama & M. J. Ablowitz. “Perturbations of solitons and solitary waves,”
Studies in Applied Mathematics. Vol 64, 225–245. (1981).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
229. Y. Kodama & A. Hasegawa. “Amplification and reshaping of optical solitons

in glass fiber-III. Amplifiers with random gain,” Optics Letters. Vol 8, Issue 6,
342–344. (1983).
230. Y. Kodama & K. Nozaki. “Soliton interaction in optical fibers,” Optics Letters.
Vol 12, Issue 12, 1038–1040. (1987).
231. Y. Kodama & A. Hasegawa. “Nonlinear pulse propagation in a monomode
dielectric guide,” IEEE Journal of Quantum Electronics. Vol 23, 510–524. (1987).
232. Y. Kodama & S. Wabnitz. “Reduction of soliton interaction forces by bandwidth
limited amplification,” Electronics Letters. Vol 27, Issue 21, 1931–1933. (1991).
233. Y. Kodama & A. Hasegawa. “Generation of asymptotically stable optical soli-
tons and suppression of the Gordon-Haus effect,” Optics Letters. Vol 17, 31–33.
(1992).
234. Y. Kodama, M. Romagnoli & S. Wabnitz. “Soliton stability and interaction in
fiber lasers,” Electronics Letters. Vol 28, Issue 21, 1981–1983. (1992).
235. Y. Kodama & A. Hasegawa. “Theoretical foundation of optical-soliton concept
in fibers,” Progress in Optics. Vol XXX, No IV, 205–259. (1992).
236. Y. Kodama & A. Hasegawa. “Generation of asymptotically stable optical solitons
and suppression of the Gordon-Haus effect.” Optics Letters Vol 17, 31–33. (1992).
237. Y. Kodama & S. Wabnitz. “Physical interpretation of reduction of soliton in-
teraction forces by bandwidth limited amplification,” Electronics Letters. Vol 29,
Issue 2, 226–227. (1993).
238. Y. Kodama & S. Wabnitz. “Reduction and suppression of soliton interaction by
bandpass filters,” Optics Letters. Vol 18, Issue 16, 1311–1313. (1993).
239. Y. Kodama & S. Wabnitz. “Comment: Physical interpretation of reduction of soli-
ton interaction forces by bandwidth limited amplification,” Electronics Letters.
Vol 29, 226–227. (1993).
240. Y. Kodama & S. Wabnitz. “Analysis of soliton stability and interactions with
sliding filters,” Optics Letters. Vol 19, Issue 3, 162–164. (1994).
241. Y. Kodama, M. Ramagnoli, S. Wabnitz & M. Midrio. “Role of third-order dis-
persion on soliton instabilities and interactions in optical fibers,” Optics Letters.
Vol 19, 165–167. (1994).
242. S. Konar. “Supergaussian optical beams in higher order nonlinear media,”
Nonlinear Optics. Vol 23, 9–22. (1999).
243. S. Konar & A. Biswas. “Intra-channel collision of Kerr law optical solitons,”
Progress in Electromagnetic Research. Vol 52. (2005).
244. S. Konar & A. Sengupta. “Propagation of an elliptic Gaussian laser beam in a
medium with saturable nonlinearity,” Journal of Optical Society of America. Vol 11,
144–147. (1994).
245. S. Konar & A. Sengupta. “Self-focusing of elliptic Gaussian laser beams in a
saturable nonlinear medium,” Indian Journal of Pure and Applied Physics. Vol 32,
660–666. (1994).
246. S. Konar & A. Sengupta. “Self-focusing of elliptic Gaussian laser beams:
Saturable nonlinearity,” Pramana. Vol 42, 223–228. (1994).
247. S. Konar, P. K. Sen & J. Kumar. “Propagation of elliptic Gaussian laser beams in
a cubic quintic nonlinear medium,” Nonlinear Optics. Vol 19, 291–308. (1999).
248. S. Konar, J. Kumar & P. K. Sen. “Suppression of soliton instability by higher-order
nonlinearity in long haul optical communication systems,” Journal of Nonlinear
Optical Physics and Materials. Vol 8, Issue 4, 497–502. (1999).
249. S. Konar. “Decay of periodically amplified solitons and suppression of this decay
by fifth order nonlinearity,” Nonlinear Optics. Vol 24, 277–288. (2000).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography 187
250. S. Konar & R. Jain. “Chirped solitons in semiconductor doped glass fibers,”
Nonlinear Optics. Vol 24, 289–309. (2000).
251. S. Konar, P. K. Barhai & S. Medhekar. “Displacement and deflection of optical
beams by nonlinear planar waveguide,” Journal of Nonlinear Optical Physics &
Materials. Vol 12, Issue 1, 1–12. (2003).
252. S. Konar & A. Biswas. “Chirped optical pulse propagation in saturating nonlin-
ear media,” Optical and Quantum Electronics. Vol 36, Issue 10, 905–918. (2004).
253. V. V. Konotop & L. Vazquez. Nonlinear Random Waves. World Scientific Publishing
254. Y. Kosmann-Schwarzbach, B. Grammaticos & K. M. Tamizhmani. Integrability of
Nonlinear Systems. Springer Verlag, New York. (2004).
255. W. Krolikowski & Y. Kivshar. “Soliton-based optical switching in waveguide
arrays,” Journal of Optical Society of America B. Vol 13, Issue 5, 876–887. (1996).
256. W. Krolikowski, D. Edmundson & O. Bang. “Unified model for partially coherent
solitons in logarithmically nonlinear media,” Physical Review E. Vol 61, Issue 3,
3122–3126. (2000).
257. W. Krolikowski, O. Bang, J. J. Rasmussen & J. Wyller. “Modulational instability
in nonlocal nonlinear Kerr media,” Physical Review E. Vol 64, 016612. (2001).
258. W. Krolikowski & O. Bang. “Solitons in nonlocal nonlinear media: Exact solu-
tions,” Physical Review E. Vol 63, No 016610. (2001).
259. W. Krolikowski, O. Bang, J. Wyller & J. J. Rasmussen. “Optical beams in nonlocal
nonlinear media,” Acta Physica Polonica A. Vol 103, Issue 2–3, 133–148. (2003).
260. A. Kumar. “Bistable soliton states and switching in doubly inhomogeneously
doped fiber couplers,” Pramana. Vol 57, Issue 5 & 6, 969–979. (2001).
261. A. Kumar, T. Kurz & W. Lauternborn. “Two-state bright solitons in doped fibers
with saturating nonlinearity,” Physical Review E. Vol 53, Issue 1, 1166–1171. (1996).
262. A. Kundu. Classical and Quantum Nonlinear Integrable Systems: Theory and Appli-
cation. Institute of Physics Publishing, Bristol. (2003).
263. T. I. Lakoba & D. J. Kaup. “Perturbation theory for the Manakov soliton and
its application to pulse propagation in randomly birefringent fibers,” Physical
Review E. Vol 56, 6147–6165. (1997).
264. M. Lakshmanan & T. Kanna. “Shape changing collision of optical solitons,
universal logic gates, and partially coherent solitons in coupled nonlinear
Schrödinger equations,” Pramana. Vol 57, Issue 5 & 6, 885–916. (2001).
265. G. L. Lamb. Elements of Soliton Theory. John Wiley & Sons, New York. (1980).
266. B. B. Laud. Lasers and Nonlinear Optics. John Wiley & Sons, New York. (1996).
267. J. H. Lee, Y. C. Lee & C. C. Lin. “Exact solutions of DNLS and derivative reaction-
diffusion systems,” Journal of Nonlinear Mathematical Physics. Vol 9, Supplement
1, 87–97. (2002).
268. N. Litchinitser, W. Krolikowski, N. Akhmediev & G. Agrawal. “Asymmetric
partially coherent solitons in saturable media,” Physical Review E. Vol 60, 2377–
2380. (1999).
269. C. Lizarraga & A. B. Aceves. “Analysis of switching phenomenon in a dense
medium of two level atoms,” Optics Letters. Vol 18, 687–689. (1993).
270. R. MacKenzie, M. B. Paranjape & W. J. Zakrzewski. Solitons: Properties, Dynamics,
Interactions, Applications. Springer Verlag, New York. (1999).
271. M. F. Mahmood. “Evolution of optical solitons in nonlinear dispersive lossy
fibers,” Optics & Laser Technology. Vol 33, Issue 6, 379–381. (2001).
272. A. I. Maimistov. “Completely integrable models of nonlinear optics,” Pramana.
Vol 57, Issue 5 & 6, 953–968. (2001).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
273. B. A. Malomed. “Waves and solitary pulses in weakly inhomogeneous

Ginzburg-Landau equations,” Physical Review E. Vol 50, 4249–4252. (1994).
274. B. A. Malomed. “Soliton stability in a bimodal optical fiber in the presence of
Raman effect,” Physical Review E. Vol 50, Issue 6, 5142–5144. (1994).
275. B. A. Malomed. “Strong periodic amplification of solitons in a lossy optical
fiber: Analytical results,” Journal of Optical Society of America B. Vol 11, Issue 7,
1261–1266. (1994).
276. B. A. Malomed. “Bound solitons in a nonlinear optical coupler,” Physical Review
E. Vol 51, Issue 2, R864–R866. (1995).
277. B. A. Malomed. “Bound states of envelope solitons,” Physical Review E. Vol 47,
2874–2880. (1993).
278. B. A. Malomed, G. D. Peng & P. L. Chu. “A nonlinear optical amplifier based on
dual core fiber,” Optics Letters. Vol 21, Issue 5, 330–332. (1996).
279. B. A. Malomed & R. S. Tasgal. “The Raman effect and solitons in an optical fiber
with general ellipticity,” Pure and Applied Optics. Vol 5, 947–965. (1996).
280. B. A. Malomed. “Pulse propagation in nonlinear optical fiber with periodically
modulated dispersion: Variational approach,” Optics Communications. Vol 136,
313–319. (1997).
281. B. A. Malomed, P. Drummond, H. He, D. Anderson, A. Berntson & M. Lisak.
“Spatiotemporal solitons in multidimensional optical media with a quadratic
nonlinearity,” Physical Review E. Vol 56, Issue 4, 4725–4735. (1997).
282. B. A. Malomed, M. Goelles, I. M. Uzunov & F. Lederer. “Stability and interaction
of pulses in simplified Ginzburg-Landau equation,” Physica Scripta. Vol 55, 73–
79. (1997).
283. B. A. Malomed, G. D. Peng, P. L. Chu, I. Towers, A. V. Buryak & R. A. Sammut.
“Stable helical solitons in optical media,” Pramana. Vol 57, Issue 5 & 6, 1061–1078.
(2001).
284. J. T. Manassah, P. L. Baldeck & R. R. Alfano. “Self-focusing, self-phase mod-
ulation and diffraction in bulk homogeneous material,” Optics Letters. Vol 13,
1090–1092. (1988).
285. M. Matsumoto, H. Ikeda, T. Udea & A. Hasegawa. “Stable soliton transmission
in the system with nonlinear gain,” Journal of Lightwave Technology. Vol 13, 658–
665. (1995).
286. A. Mecozzi, J. D. Moores, H. A. Haus & Y. Lai. “Soliton transmission control,”
Optics Letters. Vol 16, 1841–1843. (1991).
287. R. Mcleod, K. Wagner & S. Blair. “(3 + 1)-dimensional optical soliton dragging
logic.” Physical Review A. Vol 52, Issue 4, 3254–3278. (1995).
288. S. Medhekar, S. Konar & Rajkamal. “Self tapering and untapering of a self guided
laser beam in an absorbing gain medium with nonlinearity,” Pramana. Vol 44,
Issue 3, 249–256. (1995).
289. S. Medhekar, S. Konar & Rajkamal. “Successive untapering and stationary self-
trapped propagation of a laser beam in a saturating nonlinear medium,” Laser
and Particle Physics. Vol 13, 559–564. (1995).
290. R. W. Micallaef, V. V. Afanasjev, Y. S. Kivshar & J. D. Love. “Optical solitons with
power law asymptotics,” Physical Review E. Vol 54, 2936–2942. (1996).
291. D. Mihalache, D. Mazilu, L. C. Crasovan, B. A. Malomed & F. Lederer. “Three-
dimensional spinning solitons in the cubic-quintic nonlinear medium,” Physical
Review E. Vol 61, Issue 6, 7142–7145. (2000).
292. D. Mihalache, M. Bertolotti & C. Cibilia. “Nonlinear wave propagation in planar
structures,” Progress in Optics. Vol XXVII. 228–309. (1989).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography 189
293. D. Mihalache & N. C. Panoiu. “Exact solutions of nonlinear Schrödinger equa-

tion for positive group velocity dispersion,” Journal of Mathematical Physics.
Vol 33, Issue 6, 2323–2328. (1992).
294. D. Mihalache & N. C. Panoiu. “Analytic method for solving the nonlinear
Schrödinger equation describing pulse propagation in dispersive optic fibers,”
Journal of Physics A. Vol 26, Issue 11, 2679–2697. (1993).
295. D. L. Mills. Nonlinear Optics: Basic Concepts. Springer Verlag, New York. (1998).
296. V. P. Mineev. Topologically Stable Defects and Solitons in Ordered Media. Taylor and
Francis, UK. (1998).
297. L. F. Mollenauer, J. P. Gordon & S. G. Evangelides. “The sliding frequency guid-
ing filter: An improved form of soliton jitter control,” Optics Letters. Vol 17,
1575–1577. (1992).
298. J. V. Moloney. Nonlinear Optical Materials. Springer Verlag, New York. (1998).
299. J. D. Moores. “On the Ginzburg-Landau laser mode-locking model with fifth-
order saturable absorber term,” Optics Communications. Vol 96, Issues 1–3, 65–70.
(1993).
300. J. D. Moores, W. S. Wong & H. A. Haus. “Stability and timing maintenance
in soliton transmission and storage rings,” Optics Communications. Vol 113,
Issues 1–3, 153–175. (1994).
301. R. Morandotti, U. Peschell, J. S. Aitchison, H. S. Eisenberg & Y. Silberberg.
“Dynamics of discrete solitons in optical waveguide arrays,” Physical Review
302. R. Morandotti, H. S. Eisenberg, Y. Silberberg, M. Sorel & J. S. Aitchison. “Self-
focusing and defocusing in waveguide arrays,” Physical Review Letters. Vol 86,
Issue 15, 3296–3299. (2001).
303. J. A. Murdock. Perturbations: Theory and Methods. John Wiley & Sons. New York.
(1991).
304. R. Muschall, C. Schmidt-Hattenberger & F. Lederer. “Spatially solitary waves in
arrays of nonlinear waveguides,” Optics Letters. Vol 19, Issue 5, 323–325. (1994).
305. M. Nakazawa & H. Kubota. “Physical interpretation of reduction of soliton
interaction forces by bandwidth limited amplification,” Electronics Letters. Vol 28,
958–960. (1992).
306. M. Nakazawa & H. Kubota. “Reply to Kodama and Wabnitz,” Electronics Letters.
Vol 29, 226–227. (1993).
307. M. Nakazawa, H. Kubota, K. Suzuki, E. Yamada & A. Sahara. “Recent progress
in soliton transmission technology,” Chaos. Vol 10, Issue 3, 486–514. (2000).
308. K. Nakkeeran & K. Porsezian. “Solitons in an erbium-doped nonlinear fibre
medium with stimulated inelastic scattering,” Journal of Physics A. Vol 28, Issue
13, 3817–3823. (1995).
309. K. Nakkeeran & K. Porsezian. “Co-existence of a self-induced transparency
soliton and a higher order nonlinear Schrödinger soliton in an erbium-doped
fiber,” Optics Communications. Vol 123, Issues 1–3, 169–174. (1996).
310. A. C. Newell. Solitons in Mathematics and Physics. SIAM Publishers, Philadelphia.
(1985).
311. A. C. Newell & J. V. Moloney. Nonlinear Optics. Addison Wesley-Redum City,
CA. (1989).
312. P. J. Olver & D. H. Sattinger. Solitons in Physics, Mathematics, and Nonlinear Optics.
Springer Verlag, New York. (1990).
313. A. Orlowski & K. Sobczyk. “Solitons and shock waves under random external
noise,” Reports on Mathematical Physics. Vol 27, Issue 1, 59–71. (1989).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
314. E. A. Ostrovskaya & Y. S. Kivshar. “Multi-hump optical solitons in a saturable

medium,” Journal of Optics B. Vol 1, 77–83. (1999).
315. T. Ozawa. “On the nonlinear Schrödinger equations of derivative type,” Indiana
University Math Journal. Vol 45, 137–163. (1996).
316. S. L. Palacios. “Optical solitons in highly dispersive media with a dual-power
nonlinearity law,” Journal of Optics A. Vol 5, Issue 3, 180–182. (2003).
317. S. L. Palacios. “Two simple ansatze for obtaining exact solutions of high disper-
sive nonlinear Schrödinger equations,” Chaos Solitons & Fractals. Vol 19, Issue 1,
203–207. (2004).
318. N-C. Panoiu, I. V. Melnikov, D. Mihalache, C. Etrich & F. Lederer. “Soliton gener-
ation from a multi-frequency optical signal,” Journal of Optics B. Vol 4, R53–R68.
(2002).
319. D. E. Pelinovsky, A. V. Buryak & Y. S. Kivshar. “Instability of solitons governed
by quadratic nonlinearities,” Physical Review Letters. Vol 75, 591–595. (1995).
320. D. E. Pelinovsky, V. V. Afanasjev & Y. S. Kivshar. “Nonlinear theory of oscillating,
decaying and collapsing solitons in the generalized nonlinear Schrödinger’s
equation,” Physical Review E. Vol 53, Issue 2, 1940–1953. (1996).
321. U. Peschel, R. Morandotti, J. M. Arnold, J. S. Aitchison, H. S. Eisenberg,
Y. Silberberg, T. Pertsch & F. Lederer. “Optical discrete solitons in waveguide ar-
rays. 2. Dynamic properties,” Journal of Optical Society of America. Vol 19, Issue 11,
2637–2644. (2002).
322. K. Porsezian. “Soliton propagation in semiconductor-doped glass fibers with
higher-order dispersions,” Pure and Applied Optics. Vol 5, Issue 4, 345–348. (1996).
323. K. Porsezian. “Soliton models in resonant and nonresonant optical fibers,”
Pramana. Vol 57, Issue 5 & 6, 1003–1009. (2001).
324. K. Porsezian, P. Shanmugha Sundaram & A. Mahalingam. “Complete integra-
bility of N-coupled higher-order nonlinear Schrödinger equations in nonlinear
optics,” Journal of Physics A. Vol 32, Issue 49, 8731–8737. (1999).
325. K. Porsezian & V. C. Kuraikose. Optical Solitons: Theoretical and Experimental
Challenges. Springer Verlag, New York. (2003).
326. K. Porsezian & K. Nakkeeran. “Solitons in random nonuniform erbium doped
nonlinear fiber media,” Physics Letters A. Vol 206, Issue 3–4, 183–186. (1995).
327. K. Porsezian & D. Vijay Emanuel Muthiah. “Soliton pulse compression in
nonuniform birefringent fibers,” Journal of Optics A. Vol 4, Issue 2, 202–207.
(2002).
328. M. J. Potasek. “Novel femtosecond solitons in optical fibers, photonic switching,
and computing,” Journal of Physics A Vol 65, Issue 3, 941–953. (1989).
329. D. Pushkarov & S. Tanev. “Bright and dark solitary wave propagation and bista-
bility in the anomalous dispersion region of optical waveguides with third- and
fifth-order nonlinearities,” Optics Communications. Vol 124, 354–364. (1996).
330. K. I. Pushkarov & D. I. Pushkarov. “Soliton solutions in some nonlinear
Schrödinger-like equations,” Reports on Mathematical Physics. Vol 17, Issue 1,
37–40. (1980).
331. S. Raghavan & G. P. Agrawal. “Switching and self-trapping dynamics of Bose-
Einstein solitons,” Journal of Modern Optics. Vol 47, Issue 7, 1155–1169. (2000).
332. P. M. Ramos & C. R. Pavia. “Self-routing switching of solitonlike pulses in
multiple-core nonlinear fiber arrays,” Journal of Optical Society of America B.
Vol 17, Issue 7, 1125–1133. (2000).
333. K. O. Rasmussen, Y. B. Gaididei, O. Bang & P. L. Christiansen. “The influence of
noise on critical collapse in the nonlinear Schrödinger equation,” Physics Letters
A. Vol 204, 121–127. (1995).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography 191
334. S. G. Rautian. Nonlinear Optics. Nova Science. (1992).

335. M. Remoissenet. Waves Called Solitons: Concepts and Experiments. Springer Verlag,
New York. (1994).
336. P. Roussignol, D. Ricard, J. Lukasik & C. Flytzanis. “New results on optical
phase conjugation in semiconductor-doped glasses,” Journal of Optical Society of
America B. Vol 4, Issue 1, 5–13. (1987).
337. K. Rypdal & J. J. Rasmussen. “Blow-up in nonlinear Schrödinger equations II,”
Physica Scripta. Vol 33, 498–504. (1986).
338. K. Rypdal & J. J. Rasmussen. “Stability of solitary structures in nonlinear
Schrödinger equations,” Physica Scripta. Vol 40, 192–201. (1989).
339. K. Rypdal, J. J. Rasmussen & K. Thomsen. “Singularity structure of wave col-
lapse,” Physica D. Vol 16, 339–359. (1985).
340. M. Schubert & B. Wilhelmi. Nonlinear Optics and Quantum Electronics. John Wiley
& Sons, New York. (1996).
341. E. G. Sauter. Nonlinear Optics. John Wiley & Sons, New York. (1996).
342. M. Segev. “Optical spatial solitons,” Optical and Quantum Electronics. Vol 30,
Issue 7/10, 503–533. (1998).
343. K. Senthilnathan & K. Porsezian. “Evolution of polarization of a nonlinear pulse
in birefringent fiber with quintic effects,” Physics Letters A. Vol 301, Issues 5–6,
433–441. (2002).
344. I. V. Shadrivov, A. A. Sukhorukov & Y. S. Kivshar. “Beam shaping by a periodic
structure with negative refraction,” Applied Physics Letters. Vol 82, 3820–3822.
(2003).
345. J. Shatah. “Global existence of small solutions to nonlinear evolution equations,”
Journal of Differential Equations. Vol 46, 409–425. (1982).
346. V. S. Shchesnovich & E. V. Doktorov. “Perturbation theory for the modified
nonlinear Schrödinger solitons,” Physica D. Vol 129, Issue 1–2, 115–129. (1999).
347. Y. R. Shen. The Principles of Nonlinear Optics. John Wiley & Sons, New York.
(2002).
348. E. Shiojiri & Y. Fuji. “Transmission capability of an optical fibre communication
system using index nonlinearity,” Applied Optics. Vol 24, 358–360. (1985).
349. Y. Silberberg. “Collapse of optical pulses,” Optics Letters. Vol 15, 1282–1284.
(1990).
350. L. Singh, S. Konar & A. K. Sharma. “Resonant cross modulation of two laser
beams in a semiconductor slab,” Journal of Physics D. Vol 34, 2237–2239. (2001).
351. L. Singh, S. N. Rai, S. Konar & A. K. Sharma. “Information exchange between
Gaussian laser beams in n-InSb,” Physica Scripta. Vol 65, 1–4. (2002).
352. V. Singh, B. Prasad & S. P. Ojha. “A comparative study of the modal charac-
teristics and waveguide dispersion of optical waveguides with three different
closed loop cross-sectional boundaries,” Optik. Vol 115, Issue 6, 281–288. (2004).
353. D. V. Skryabin & W. J. Firth. “Dynamics of self-trapped beams with phase dis-
location in saturable Kerr and quadratic nonlinear media,” Physical Review E.
Vol 58, 3916–3930. (1998).
354. A. W. Snyder & D. J. Mitchell. “Spatial solitons of the power-law nonlinearity,”
Optics Letters. Vol 18, Issue 2, 101–103. (1993).
355. M. S. Sodha, S. Medhekar, S. Konar, A. Saxena & Rajkamal. “Absorption/
amplification induced self tapering and untapering of a laser beam in a saturable
nonlinear medium; large nonlinearity,” Optics Letters. Vol 19, 15–21. (1994).
356. J. M. Soto-Crespo, N. Akhmediev & A. Ankiewicz. “Soliton propagation in opti-
cal devices with two-component fields: A comparative study,” Journal of Optical
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
357. J. M. Soto-Crespo, N. Akhmediev & V. V. Afanasjev. “Algebraic pulse-like so-

lutions of the quintic complex Ginzburg-Landau equation,” Optics Communica-
tions. Vol 118, 587–593. (1995).
358. J. M. Soto-Crespo, V. Afanasjev, N. Akhmediev & G. Town. “Dual-frequency
pulses in fiber lasers,” Optics Communications. Vol 130, 245–248. (1996).
359. G. I. Stegeman & R. H. Stolen. “Waveguides and fibers for nonlinear optics,”
Journal of Optical Society of America B. Vol 6, Issue 4, 652–662. (1989).
360. C. Sulem & P. L. Sulem. The Nonlinear Schrödinger’s Equation. Springer Verlag,
New York. (1999).
361. A. A. Sukhorukov, Y. S. Kivshar, H. S. Eisenberg & Y. Silberberg. “Spatial optical
solitons in waveguide arrays,” IEEE Journal of Quantum Electronics. Vol 39, 31–50.
(2003).
362. S. Tanev & D. I. Pushkarov. “Solitary wave propagation and bistability in the
normal dispersion region of highly nonlinear optical fibers and waveguides,”
Optics Communications. Vol 141, Issue 5–6, 322–328. (1997).
363. S. Trillo, S. Wabnitz, E. M. Wright & G. I. Stegeman. “Soliton switching in fiber
nonlinear directional couplers,” Optics Letters. Vol 13, 672–674. (1988).
364. R. L. Sutherland. Handbook of Nonlinear Optics. Marcel Dekker, New York.
(1996).
365. J. M. Soto-Crespo, N. Akhmediev & A. Ankiewicz. “Soliton propagation in opti-
cal devices with two-component fields: A comparative study,” Journal of Optical
366. A. W. Snyder & D. J. Mitchell. “Spatial solitons of the power-law nonlinearity,”
Optics Letters, Vol 18, Issue 2, 101–103, (1993).
367. J. M. Soto-Crespo, N. N. Akhmediev & V. V. Afanasjev. “Stability of pulse-like
solutions of the quintic complex Ginzburg-Landau equation,” Journal of Optical
Society of America B. Vol 13, 1439–1448. (1996).
368. A. A. Sukhorukov, Y. S. Kivshar, O. Bang & C. M. Soukoulis. “Parametric lo-
calized modes in quadratic nonlinear photonic structures,” Physical Review E.
Vol 63, 016615. (2001).
369. V. I. Talanov. “Focusing of light in cubic media,” JETP Letters. Vol 11, 199–201.
(1970).
370. X. Tang & P. Ye. “Calculation of the bit error rate for optical soliton communica-
tion systems with lumped amplifiers,” Optics Letters. Vol 18, 1156–1158. (1993).
371. J. R. Taylor. Optical Solitons: Theory and Experiment. Cambridge University Press,
Cambridge. (1992).
372. V. Tikhonenko, Y. S. Kivshar, V. V. Steblina & A. A. Zozulya. “Vortex solitons in
a saturable optical medium,” Journal of Optical Society of America B. Vol 15, 79–86.
(1998).
373. L. Torner, J. P. Torres, D. V. Petrov & J. M. Soto-Crespo. “From topological charge
information to sets of solitons in quadratic nonlinear media,” Optical and Quan-
tum Electronics. Vol 30, Issue 7/10, 809–827. (1998).
374. L. Torner, D. Mihalache, M. C. Santos & N. N. Akhmediev. “Spatial walking
solitons in quadratic nonlinear crystals,” Journal of Optical Society of America B.
Vol 15, Issue 5, 1476–1487. (1998).
375. H. Torress-Silvia & M. Zamorano. “Chiral effects on optical solitons,” Mathemat-
ics and Computers in Simulations. Vol 62, Issue 1–2, 149–161. (2003).
376. S. Trillo & W. Torruellas. Spatial Solitons. Springer Verlag, New York. (2001).
377. I. M. Uzunov, V. D. Stoev & T. I. Tsoleva. “N-soliton interaction in trains of
unequal soliton pulses in optical fibers,” Optics Letters. Vol 17, 1417–1419. (1992).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
Bibliography 193
378. I. M. Uzunov, V. D. Stoev & T. I. Tsoleva. “Influence of the initial phase

difference between pulses on the N-soliton interaction in trains of unequal
solitons in optical fibers,” Optics Communications. Vol 97, Issue 5–6, 307–311.
(1993).
379. I. M. Uzunov, M. Gölles, L. Leine & F. Lederer. “The effect of bandwidth limited
amplification on soliton interaction,” Optics Communications. Vol 110, Issue 3–4,
465–474. (1994).
380. I. M. Uzunov, R. Muschall, M. Gölles, F. Lederer & S. Wabnitz. “Effect of non-
linear gain and filtering on soliton interaction,” Optics Communications. Vol 118,
Issue 5–6, 577–580. (1995).
381. I. Uzunov, R. Muschall, M. Golles, Y. Kivshar, B. Malomed & F. Lederer. “Pulse
switching in nonlinear fiber directional couplers,” Physical Review E. Vol 51,
2527–2537. (1995).
382. I. M. Uzunov, V. Gerdjikov, M. Gölles & F. Lederer. “On the description of N-
soliton interactions in optical fibers,” Optics Communications. Vol 125, Issue 4–6,
237–242. (1996).
383. I. M. Uzunov, M. Golles & F. Lederer. “Enhanced sideband instability in soli-
ton transmissions with semiconductor optical amplifiers,” Optics Letters. Vol 22,
Issue 18, 1406–1408. (1997).
384. S. Wabnitz, Y. Kodama & A. B. Aceves. “Control of optical soliton interactions,”
Optical Fiber Technology. Vol 1, 187–217. (1995).
385. S. Wabnitz. “Control of soliton train transmission, storage, and clock recovery
by cw light injection,” Journal of Optical Society of America B. Vol 13, Issue 12,
2739–2749. (1996).
386. M. Wadati. “Introduction to solitons,” Pramana. Vol 57, Issue 5 & 6, 841–847.
(2001).
387. C. Weilnau, W. Krolikowski, E. A. Ostrovskaya, M. Ahles, M. Geisser,
G. McCarthy, C. Denz, Y. S. Kivshar & B. Luther-Davis. “Composite spatial soli-
tons in a saturable nonlinear bulk medium,” Applied Physics B. Vol 72, 723–727.
(2001).
388. M. Weinstein. “Existence and dynamic stability of solitary wave solutions of
equations arising in long wave propagations,” Communications in Partial Differ-
ential Equations. Vol 12, Issue 10, 1133–1173. (1987).
389. G. B. Whitham. Linear and Nonlinear Waves. John Wiley and Sons, New York.
(1999).
390. T. R. Wolinski. “Polarimetric optical fibers and sensors,” Progress in Optics.
Vol XL, 1–75. (2000).
391. J. Wyller & E. Mjolhus. “A perturbation theory for Alfven solitons,” Physica D.
Vol 13, 234–246. (1984).
392. J. Wyller, W. Krolikowski, O. Bang & J. J. Rasmussen. “Generic features of mod-
ulational instability in nonlocal Kerr media,” Physical Review E. Vol 66, 066615.
(2002).
393. Z. Xu, L. Li, Z. Li & G. Zhou. “Soliton interaction under the influence of higher-
order effects,” Optics Communications. Vol 210, Issue 3–6, 375–384. (2002).
394. J. Yang & D. J. Kaup. “Stability and evolution of solitary waves in perturbed
generalized nonlinear Schrödinger’s equation,” SIAM Journal of Applied Mathe-
matics. Vol 60, Issue 3, 967–989. (2000).
395. V. E. Zakharov & A. B. Shabat. “Exact theory of two-dimensional self-focusing
and one-dimensional self-modulation of waves in nonlinear media,” Journal of
Experimental and Theoretical Physics. Vol 34, 62–69. (1972).
P1: Binaya Dash
October 5, 2006 12:36 C6382 C6382˙Book
396. V. E. Zakharov & V. S. Synakh. “The nature of self-focusing singularity,” Soviet

Physics JETP. Vol 41, 465–468. (1976).
397. V. E. Zakharov & E. A. Kuznetsov. “Optical solitons and quasisolitons,” Journal
of Experimental and Theoretical Physics. Vol 86, Issue 5, 1035–1046. (1998).
398. P. E. Zhidkov. “Existence of solutions to the Cauchy problem and stability of
kink-solutions of the nonlinear Schrödinger equation,” Siberian Mathematical
Journal. Vol 33, Issue 2, 239–246. (1992).
399. P. E. Zhidkov. Korteweg–de Vries and Nonlinear Schrödinger Equations: Qualitative
Theory. Springer Verlag, New York. (2001).
400. C. Zhou, X. T. He & T. Cai. “Pattern structures on generalized nonlinear
Schrödinger’s equations with various nonlinear terms,” Physical Review E. Vol 50,
Issue 5, 4136–4155. (1994).
P1: Naresh Chandra
October 5, 2006 12:46 C6382 C6382˙C014
Index
A Hamiltonian perturbations, 131–134

non-Hamiltonian perturbations, 131
Advantages of fiber-optic soliton-soliton interaction, 108–109
communications, 5–6 stochastic perturbation, 142–143
Arbitrary pulse propagation, saturable Dual-power law nonlinearity, 77–86
law nonlinearity, 91–98 integrals of motion, 79–82
lossless uniform media, 94–96 quasi-stationary solution, 82–85
lossy media, 97–98 traveling wave solution, 78–79
stationary pulse propagation, 96–97
B E
Bistable solitons, saturable law Exponential law, 9–10

nonlinearity, 90–91
Bullets, optical, 161–166
one plus three dimensions, 162–166 F
integrals of motion, 162–164
parameter evolution, 164–166 Four-port couplers, 146
C G
Couplers, optical, 145–160
Graded-index fiber, 6
four-port couplers, 146
functions of, 146–147
magneto-optic waveguides, 157–160
mathematical analysis, 158–160 H
multiple-core couplers, 152–157
Hamiltonian perturbation, 170
coupling with all neighbors, 155–157
Hamiltonian perturbations, soliton-soliton
coupling with nearest neighbors,
interaction, 103
153–155
Hamiltonian structure, integrals of
optical switching, 147–148
motion, 41–42
star coupler, 146
Hasegawa, A., 167
three-port couplers, 146
High order polynomial law, 10
twin-core couplers, 148–152
types, 146–147
Cubic-quintic nonlinearity. See Parabolic
nonlinearity I
Integrals of motion
D dual-power law nonlinearity, 79–82
Kerr law nonlinearity, 39–42
Dielectric waveguide, in fiber-optic parabolic law nonlinearity, 70–72
communications, 6 power law nonlinearity, 59–62
Dual-power law, 9 Inverse scattering transform, Kerr law
quasi-particle theory, soliton-soliton nonlinearity, 33–39
interaction, 128–134 1-soliton solution, 37–39
195
P1: Naresh Chandra
October 5, 2006 12:46 C6382 C6382˙C014
K N
Kerr coefficient, 28 NLSE. See Nonlinear Schrödinger’s
Kerr law, 8–9 equation
quasi-particle theory, soliton-soliton Non-Hamiltonian perturbation, 170
interaction, 113–117 Non-Hamiltonian perturbations,
Hamiltonian perturbations, 115–117 soliton-soliton interaction, 103
non-Hamiltonian perturbations, Nonlinear index of refraction. See Kerr
114–115 coefficient
soliton-soliton interaction, 104–105 Nonlinear Schrödinger’s equation, 7–26,
stochastic perturbation, 137–139 29–32, 57
Kerr law nonlinearity, 27–56 dual-power law, 9
application of, 54–55 exponential law, 9–10
integrals of motion, 39–42 high order polynomial law, 10
Hamiltonian structure, 41–42 Kerr law, 8–9
inverse scattering transform, 33–39 log law, 10
1-soliton solution, 37–39 motion integrals, 12–16
lie transform, 49–55 nonlinearities of function F, 8
nonlinear Schrödinger’s equation, nonlinearity classification, 8–10
29–32 parabolic law, 9
quasi-stationary solution, 46–49 parameter evolution, 17–20
traveling wave solution, 32–33 perturbation terms, 17–20
variational principle, 43–46 power law, 9
Korteweg-de Vries equation, 167 quasi-stationary solution, 20–26
Kruskal, Martin, 167 application of, 22–26
mathematical theory, 21–22
saturable law nonlinearity, 88–90
saturating law, 9
L soliton, profile of, 8
threshold law, 10
Lie transform, Kerr law nonlinearity,
traveling waves, 11–12
49–55
triple-power law, 10
Log law, 10
Nonlinearity
Lossless uniform media, arbitrary pulse
dual-power law, 77–86
propagation, saturable law
quasi-stationary solution, 82–85
Lossy media, arbitrary pulse propagation,
traveling wave solution, 78–79
saturable law nonlinearity, 97–98
parabolic law, 67–76
M traveling wave solution, 69–70
power law, 57–66
Magneto-optic waveguides, 157–160 integrals of motion, 59–62
mathematical analysis, 158–160 nonlinear Schrödinger’s equation,
Motion integrals 57
dual-power law nonlinearity, 79–82 quasi-stationary solution, 62–65
Kerr law nonlinearity, 39–42 traveling wave solution, 58–59
nonlinear Schrödinger’s equation, saturable law, 87–99
12–16 arbitrary pulse propagation, 91–98
parabolic law nonlinearity, 70–72 lossless uniform media, 94–96
power law nonlinearity, 59–62 lossy media, 97–98
Multiple-core couplers, 152–157 stationary pulse propagation, 96–97
coupling with all neighbors, 155–157 bistable solitons, 90–91
coupling with nearest neighbors, conserved quantities, 90
153–155 nonlinear Schrödinger’s equation,
Multiport couplers, 146 88–90
P1: Naresh Chandra
October 5, 2006 12:46 C6382 C6382˙C014
Index 197
O Power law nonlinearity, 57–66

One plus three dimension optical bullets, nonlinear Schrödinger’s equation, 57
162–166 quasi-stationary solution, 62–65
integrals of motion, 162–164 traveling wave solution, 58–59
parameter evolution, 164–166
Optical bullets, 161–166
one plus three dimensions, 162–166 Q
parameter evolution, 164–166 Quasi-particle theory, soliton-soliton
Optical couplers, 145–160 interaction, 109–134
four-port couplers, 146 dual-power law, 128–134
functions of, 146–147 Hamiltonian perturbations, 131–134
magneto-optic waveguides, 157–160 non-Hamiltonian perturbations, 131
mathematical analysis, 158–160 Kerr law, 113–117
multiple-core couplers, 152–157 Hamiltonian perturbations, 115–117
coupling with all neighbors, 155–157 non-Hamiltonian perturbations,
coupling with nearest neighbors, 114–115
153–155 parabolic law, 124–128
optical switching, 147–148 Hamiltonian perturbations, 127–128
star coupler, 146 non-Hamiltonian perturbations,
three-port couplers, 146 126–127
twin-core couplers, 148–152 power law, 117–124
types, 146–147 Hamiltonian perturbations, 121–124
Optical fiber types, 5 non-Hamiltonian perturbations,
Optical switching, 147–148 119–121
Optical waveguides, 3–6 Quasi-stationary solution
Kerr law nonlinearity, 46–49
nonlinear Schrödinger’s equation,
20–26
P
application of, 22–26
Parabolic law, 9 mathematical theory, 21–22
quasi-particle theory, soliton-soliton Quasi-stationary solution
interaction, 124–128 dual-power law nonlinearity, 82–85
Hamiltonian perturbations, 127–128 parabolic law nonlinearity, 72–75
non-Hamiltonian perturbations,
126–127
soliton-soliton interaction, 106–107 R
stochastic perturbation, 141–142
Refraction, nonlinear index. See Kerr
Parabolic law nonlinearity, 67–76
coefficient
Russell, John Scott, 167
traveling wave solution, 69–70
Parabolic nonlinearity, 168 S
Parameter evolution, nonlinear
Schrödinger’s equation, 17–20 Saturable law nonlinearity, 87–99
Perturbation terms, nonlinear arbitrary pulse propagation, 91–98
Schrödinger’s equation, 17–20 lossless uniform media, 94–96
Power law, 9 lossy media, 97–98
quasi-particle theory, soliton-soliton stationary pulse propagation, 96–97
interaction, 117–124 bistable solitons, 90–91
Hamiltonian perturbations, 121–124 conserved quantities, 90
non-Hamiltonian perturbations, nonlinear Schrödinger’s equation, 88–90
119–121 Saturating law, 9
soliton-soliton interaction, 105–106 Scattering transform, inverse, Kerr law
stochastic perturbation, 139–140 nonlinearity, 33–39
P1: Naresh Chandra
October 5, 2006 12:46 C6382 C6382˙C014
Schrödinger’s equation, nonlinear, 7–26, non-Hamiltonian perturbations,

29–32, 57 126–127
dual-power law, 9 power law, 117–124
exponential law, 9–10 Hamiltonian perturbations,
high order polynomial law, 10 121–124
Kerr law, 8–9 non-Hamiltonian perturbations,
log law, 10 119–121
motion integrals, 12–16 Soliton train, 5
nonlinearities of function F, 8 Star coupler, 146
nonlinearity classification, 8–10 Stationary pulse propagation, arbitrary
parabolic law, 9 pulse propagation, saturable law
parameter evolution, 17–20 nonlinearity, 96–97
perturbation terms, 17–20 Step-index fiber, 6
power law, 9 Stochastic perturbation, 135–143
quasi-stationary solution, 20–26 dual-power law, 142–143
application of, 22–26 Kerr law, 137–139
mathematical theory, 21–22 parabolic law, 141–142
saturable law nonlinearity, 88–90 power law, 139–140
saturating law, 9
soliton, profile of, 8
T
threshold law, 10
traveling waves, 11–12 Tappert, F.D., 167
triple-power law, 10 Three-port couplers, 146
Signal security, in fiber-optic Threshold law, 10
communications, 6 Transmission loss, in fiber-optic
Soliton detection window, 5 communications, 6
Soliton profile, nonlinear Schrödinger’s Traveling wave solution
equation, 8 dual-power law nonlinearity, 78–79
Soliton-soliton interaction, 101–134 Kerr law nonlinearity, 32–33
dual-power law, 108–109 parabolic law nonlinearity, 69–70
Hamiltonian perturbations, 103 power law nonlinearity, 58–59
Kerr law, 104–105 Traveling waves, nonlinear Schrödinger’s
mathematical formulation, 102–109 equation, 11–12
non-Hamiltonian perturbations, 103 Triple-power law, 10
parabolic law, 106–107 Twin-core couplers, 148–152
power law, 105–106
quasi-particle theory, 109–134
dual-power law, 128–134 V
Hamiltonian perturbations,
Variational principle, Kerr law
131–134
non-Hamiltonian perturbations,
131
Kerr law, 113–117 W
115–117 Width bandwidth, in fiber-optic
non-Hamiltonian perturbations, communication, 6
114–115
parabolic law, 124–128 Z
127–128 Zabusky, Norma, 167

Introduction To Non-Kerr Optical Solitons

Uploaded by

Copyright:

Available Formats

Introduction To Non-Kerr Optical Solitons

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

Introduction To Non-Kerr Optical Solitons

Uploaded by

Copyright:

Available Formats

P1: Binaya Dash

October 6, 2006 18:2 C6382 C6382˙C000

Chapman & Hall/CRC

No claim to original U.S. Government works

International Standard Book Number-10: 1-58488-638-2 (Hardcover)

Library of Congress Cataloging-in-Publication Data

Biswas, Anjan, Dr.

Visit the Taylor & Francis Web site at

Recent years have witnessed an explosion of research activities in the field of

2 The Nonlinear Schrödinger’s Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3 Kerr Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4 Power Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

5 Parabolic Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

6 Dual-Power Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

7 Saturable Law Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

8 Soliton–Soliton Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

9 Stochastic Perturbation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

10 Optical Couplers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .145

11 Optical Bullets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

This introductory chapter is intended to provide a general overview of fiber

2 Introduction to Non-Kerr Solitons

depending on the input signal state of polarization. The transmission impair-

1.2 Optical Waveguides

4 Introduction to Non-Kerr Solitons

1.2.1 Types of Optical Fibers

1.2.2 Advantages of Fiber-Optic Communications

6 Introduction to Non-Kerr Solitons

1. Wider bandwidth: The information-carrying capacity of a transmis-

In equation (2.1), q is the dependent variable, x and t are the independent

8 Introduction to Non-Kerr Solitons

equations. This is a nonlinear partial differential equation that is not integrable,

2.1.1 Nonlinearity Classification

1. Kerr law: F (s) = s

The Nonlinear Schrödinger’s Equation 9

10 Introduction to Non-Kerr Solitons

When the phase velocity of slow plasma oscillation is much smaller

A smooth transition of this kind can be modeled as

F (s) = As{1 + α tanh[γ (s 2 − I02 )]}

where for s  I0 , F (s) ≈ n1 s, where n1 = A[1 − α tanh2 (γ I02 )],

The Nonlinear Schrödinger’s Equation 11

2.2 Traveling Waves

q (x, t) = Ag[B(x − x̄)]e i(−κ x+ωt+σ0 ) (2.2)

q t = −ABvg (τ )e iφ − iω Ag(τ )e iφ (2.4)

where, the phase φ is

τ = B(x − x̄) (2.6)

Also, from (2.2) one can get

q xx = AB 2 g (τ ) − 2iκ ABg (τ ) − κ 2 Ag(τ ) (2.7)

Substituting (2.4) and (2.7) into (2.1) simplifies it to

12 Introduction to Non-Kerr Solitons

so that from (2.9), one obtains

On separation of variables and integration once more, equation (2.12) leads to

Equation (2.13) can be integrated only if the law of nonlinearity is known. As

2.3 Integrals of Motion

represents the conservation law. Here T is known as the density while X is

The Nonlinear Schrödinger’s Equation 13

since, for a soliton, q , q x , q xx , . . . approach zero as |x| −→ ∞, as previously

14 Introduction to Non-Kerr Solitons

Performing the operation q x∗ × (2.1) + q x × (2.23) gives

Now performing the operation q ∗ × ∂∂x (2.1) + q × ∂∂x (2.23) gives

where for s I0 , F (s) ≈ n1 s, where n1 = A[1 − α tanh2 (γ I02 )],

Here, K and F are nonlinear functions of q , q x , · · ·, while 0 < 1. The

tons and the Inverse Scattering Transform. 1981. Copyright