Download the full version of the ebook now at ebookultra.

com

Computational Methods for Electromagnetic and

Optical Systems Second Edition Banerjee

https://ebookultra.com/download/computational-
methods-for-electromagnetic-and-optical-systems-
second-edition-banerjee/

Explore and download more ebook at https://ebookultra.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Computational Methods for Electric Power Systems 2nd

Edition Mariesa L. Crow

https://ebookultra.com/download/computational-methods-for-electric-
power-systems-2nd-edition-mariesa-l-crow/

ebookultra.com

Computational Methods for Electric Power Systems 3rd

Edition Mariesa L. Crow (Author)

https://ebookultra.com/download/computational-methods-for-electric-
power-systems-3rd-edition-mariesa-l-crow-author/

ebookultra.com

Advanced Modeling in Computational Electromagnetic

Compatibility 1st Edition Dragan Poljak

https://ebookultra.com/download/advanced-modeling-in-computational-
electromagnetic-compatibility-1st-edition-dragan-poljak/

ebookultra.com

Computational modeling methods for neuroscientists 1st

Edition Erik De Schutter

https://ebookultra.com/download/computational-modeling-methods-for-
neuroscientists-1st-edition-erik-de-schutter/

ebookultra.com
Computational methods for mass spectrometry proteomics 1st
Edition Ingvar Eidhammer

https://ebookultra.com/download/computational-methods-for-mass-
spectrometry-proteomics-1st-edition-ingvar-eidhammer/

ebookultra.com

Low Frequency Electromagnetic Modeling for Electrical and

Biological Systems Using MATLAB 1st Edition Sergey N.
Makarov
https://ebookultra.com/download/low-frequency-electromagnetic-
modeling-for-electrical-and-biological-systems-using-matlab-1st-
edition-sergey-n-makarov/
ebookultra.com

Measurement Instrumentation and Sensors Handbook

Electromagnetic Optical Radiation Chemical and Biomedical
Measurement 2nd Edition John G. Webster
https://ebookultra.com/download/measurement-instrumentation-and-
sensors-handbook-electromagnetic-optical-radiation-chemical-and-
biomedical-measurement-2nd-edition-john-g-webster/
ebookultra.com

Digital Signal Processing Mathematical and Computational

Methods Software Development and Applications Second
Edition Jonathan M Blackledge
https://ebookultra.com/download/digital-signal-processing-
mathematical-and-computational-methods-software-development-and-
applications-second-edition-jonathan-m-blackledge/
ebookultra.com

Computational Methods for Two Phase Flow and Particle

Transport Wen Ho Lee

https://ebookultra.com/download/computational-methods-for-two-phase-
flow-and-particle-transport-wen-ho-lee/

ebookultra.com
Computational Methods for Electromagnetic and Optical
Systems Second Edition Banerjee Digital Instant
Download
Author(s): Banerjee, Partha P.; Jarem, John M
ISBN(s): 9781439891285, 1439891281
Edition: 2nd ed
File Details: PDF, 55.22 MB
Year: 2014
Language: english
Computational Methods
for Electromagnetic
and Optical Systems
SECOnd EditiOn
 OPTICAL SCIENCE AND ENGINEERING

Founding Editor
Brian J. Thompson
University of Rochester
Rochester, New York

RECENTLY PUBLISHED

Computational Methods for Electromagnetic and Optical Systems, Second Edition, John M. Jarem,
and Partha P. Banerjee
Optical Methods of Measurement: Wholefield Techniques, Second Edition, Rajpal S. Sirohi
Optoelectronics: Infrared-Visible-Ultraviolet Devices and Applications, Second Edition,
edited by Dave Birtalan and William Nunley
Photoacoustic Imaging and Spectroscopy, edited by Lihong V. Wang
Polarimetric Radar Imaging: From Basics to Applications, Jong-Sen Lee and Eric Pottier
Near-Earth Laser Communications, edited by Hamid Hemmati
Laser Safety: Tools and Training, edited by Ken Barat
Slow Light: Science and Applications, edited by Jacob B. Khurgin and Rodney S. Tucker
Dynamic Laser Speckle and Applications, edited by Hector J. Rabal and Roberto A. Braga Jr.
Biochemical Applications of Nonlinear Optical Spectroscopy, edited by Vladislav Yakovlev
Tunable Laser Applications, Second Edition, edited by F. J. Duarte
Optical and Photonic MEMS Devices: Design, Fabrication and Control, edited by Ai-Qun Liu
The Nature of Light: What Is a Photon?, edited by Chandrasekhar Roychoudhuri, A. F. Kracklauer,
and Katherine Creath
Introduction to Nonimaging Optics, Julio Chaves
Introduction to Organic Electronic and Optoelectronic Materials and Devices, edited by
Sam-Shajing Sun and Larry R. Dalton
Fiber Optic Sensors, Second Edition, edited by Shizhuo Yin, Paul B. Ruffin, and Francis T. S. Yu
Terahertz Spectroscopy: Principles and Applications, edited by Susan L. Dexheimer
Photonic Signal Processing: Techniques and Applications, Le Nguyen Binh
Smart CMOS Image Sensors and Applications, Jun Ohta
Organic Field-Effect Transistors, Zhenan Bao and Jason Locklin
Coarse Wavelength Division Multiplexing: Technologies and Applications, edited by Hans Joerg
Thiele and Marcus Nebeling
Microlithography: Science and Technology, Second Edition, edited by Kazuaki Suzuki and
Bruce W. Smith
Physical Properties and Data of Optical Materials, Moriaki Wakaki, Keiei Kudo, and Takehisa
Shibuya
Microwave Photonics, edited by Chi H. Lee
Photonics: Principles and Practices, Abdul Al-Azzawi
Lens Design, Fourth Edition, Milton Laikin
Gas Lasers, edited by Masamori Endo and Robert F. Walker
Optical Waveguides: From Theory to Applied Technologies, edited by Maria L. Calvo and
Vasudevan Lakshiminarayanan
Computational Methods
for Electromagnetic
and Optical Systems
SECOnd EditiOn

John M. Jarem
Partha P. Banerjee

Boca Raton London New York

CRC Press is an imprint of the

Taylor & Francis Group, an informa business
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the accuracy
of the text or exercises in this book. This book’s use or discussion of MATLAB® software or related products does not consti-
tute endorsement or sponsorship by The MathWorks of a particular pedagogical approach or particular use of the MATLAB®
software.

CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2011 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Version Date: 20110720

International Standard Book Number-13: 978-1-4398-9128-5 (eBook - PDF)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to
publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials
or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material repro-
duced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any
copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, 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.copy-
right.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 identifica-
tion and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com
I (John M. Jarem) would like to dedicate the book to Margaret
and Francis Shea, and my parents, Sarah and John Jarem.
I (Partha P. Banerjee) would like to dedicate the book to
my mom and dad, Prativa and Abani Banerjee.
Contents
Preface............................................................................................................................................ xiii
Authors.............................................................................................................................................. xv

Chapter 1 Mathematical Preliminaries..........................................................................................1

1.1 Introduction........................................................................................................ 1
1.2 Fourier Series and Its Properties........................................................................1
1.3 Fourier Transform..............................................................................................3
1.4 Hankel Transform............................................................................................... 4
1.5 Discrete Fourier Transform................................................................................4
1.6 Review of Eigenanalysis.....................................................................................5
Problems...................................................................................................................... 10
References................................................................................................................... 10

Chapter 2 Scalar EM Beam Propagation in Inhomogeneous Media........................................... 11

2.1 Introduction...................................................................................................... 11
2.2 Transfer Function for Propagation.................................................................... 11
2.3 Split-Step Beam Propagation Method.............................................................. 12
2.4 Beam Propagation in Linear Media................................................................. 14
2.4.1 Linear Free-Space Beam Propagation................................................ 14
2.4.2 Propagation of Gaussian Beam through Graded Index Medium........ 16
2.5 Beam Propagation through Diffraction Gratings:
Acoustooptic Diffraction............................................................................. 18
2.6 Beam Propagation in Kerr-Type Nonlinear Media.......................................... 19
2.6.1 Nonlinear Schrodinger Equation........................................................ 19
2.6.2 Simulation of Self-Focusing Using Adaptive Fourier
and Fourier–Hankel Transform Methods............................................ 22
2.7 Beam Propagation and Coupling in Photorefractive Media.............................28
2.7.1 Basic Photorefractive Physics..............................................................28
2.7.2 Induced Transmission Gratings........................................................... 29
2.7.3 Induced Reflection Gratings and Bidirectional Beam
Propagation Method............................................................................ 35
2.8 z-Scan Method.................................................................................................. 39
2.8.1 Model for Beam Propagation through PR Lithium Niobate............... 41
2.8.2 z-Scan: Analytical Results, Simulations, and Sample Experiments..... 43
Problems......................................................................................................................46
References................................................................................................................... 47

Chapter 3 EM Wave Propagation in Linear Media...................................................................... 51

3.1 Introduction...................................................................................................... 51
3.2 Maxwell’s Equations........................................................................................ 51
3.3 Constitutive Relations: Frequency Dependence and Chirality......................... 53
3.3.1 Constitutive Relations and Frequency Dependence............................ 53
3.3.2 Constitutive Relations for Chiral Media............................................. 55

vii
viii Contents

3.4 Plane Wave Propagation through Linear Homogeneous Isotropic Media....... 56

3.4.1 Dispersive Media................................................................................. 57
3.4.2 Chiral Media....................................................................................... 61
3.5 Power Flow, Stored Energy, Energy Velocity, Group Velocity,
and Phase Velocity........................................................................................... 62
3.6 Metamaterials and Negative Index Media....................................................... 65
3.6.1 Beam Propagation in NIMs................................................................ 68
3.7 Propagation through Photonic Band Gap Structures: The Transfer
Matrix Method................................................................................................. 71
3.7.1 Periodic PIM–NIM Structures............................................................ 75
3.7.2 EM Propagation in Complex Structures............................................. 75
Problems...................................................................................................................... 78
References...................................................................................................................80

Chapter 4 Spectral State Variable Formulation for Planar Systems............................................ 83

4.1 Introduction...................................................................................................... 83
4.2 State Variable Analysis of an Isotropic Layer.................................................. 86
4.2.1 Introduction......................................................................................... 86
4.2.2 Analysis............................................................................................... 87
4.2.3 Complex Poynting Theorem................................................................90
4.2.4 State Variable Analysis of an Isotropic Layer in Free Space.............. 93
4.2.5 State Variable Analysis of a Radar Absorbing Layer.......................... 98
4.2.6 State Variable Analysis of a Source in Isotropic Layered Media..... 100
4.3 State Variable Analysis of an Anisotropic Layer........................................... 104
4.3.1 Introduction....................................................................................... 104
4.3.2 Basic Equations................................................................................. 104
4.3.3 Numerical Results............................................................................. 108
4.4 One-Dimensional k-Space State Variable Solution........................................ 112
4.4.1 Introduction....................................................................................... 112
4.4.2 k-Space Formulation......................................................................... 112
4.4.3 Ground Plane Slot Waveguide System.............................................. 113
4.4.4 Ground Plane Slot Waveguide System, Numerical Results.............. 120
Problems.................................................................................................................... 122
References................................................................................................................. 126

Chapter 5 Planar Diffraction Gratings....................................................................................... 129

5.1 Introduction.................................................................................................... 129
5.2 H-Mode Planar Diffraction Grating Analysis................................................ 132
5.2.1 Full-Field Formulation...................................................................... 133
5.2.2 Differential Equation Method........................................................... 139
5.2.3 Numerical Results............................................................................. 145
5.2.4 Diffraction Grating Mirror................................................................ 150
5.3 Application of RCWA and the Complex Poynting Theorem
to E-Mode Planar Diffraction Grating Analysis............................................ 153
5.3.1 E-Mode RCWA Formulation............................................................ 155
5.3.2 Complex Poynting Theorem.............................................................. 158
u
5.3.2.1 Sample Calculation of PWE ................................................ 159
Contents ix

5.3.2.2 Other Poynting Theorem Integrals.................................... 160

5.3.2.3 Simplification of Results and Normalization.................... 160
5.3.3 Numerical Results............................................................................. 165
5.4 Multilayer Analysis of E-Mode Diffraction Gratings.................................... 172
5.4.1 E-Mode Formulation......................................................................... 173
5.4.2 Numerical Results............................................................................. 178
5.5 Crossed Diffraction Grating........................................................................... 180
5.5.1 Crossed Diffraction Grating Formulation......................................... 180
5.5.2 Numerical Results............................................................................. 194
Problems....................................................................................................................202
References.................................................................................................................205

Chapter 6 Application of RCWA to Analysis of Induced Photorefractive Gratings.................209

6.1 Introduction to Photorefractive Materials......................................................209
6.2 Dynamic Nonlinear Model for Diffusion-Controlled PR Materials.............. 211
6.3 Approximate Analysis.................................................................................... 212
6.3.1 Numerical Algorithm........................................................................ 215
6.3.2 TE Numerical Simulation Results..................................................... 215
6.3.3 TM Numerical Simulation Results....................................................224
6.3.4 Discussion of Results from Approximate Analysis.......................... 226
6.4 Exact Analysis................................................................................................ 227
6.4.1 Finite Difference Kukhtarev Analysis.............................................. 229
6.4.2 TM Numerical Simulation Results.................................................... 231
6.5 Reflection Gratings.........................................................................................240
6.5.1 RCWA Optical Field Analysis..........................................................240
6.5.2 Material Analysis.............................................................................. 241
6.5.3 Numerical Results.............................................................................244
6.6 Conclusion......................................................................................................246
Problems....................................................................................................................248
References................................................................................................................. 249

Chapter 7 Rigorous Coupled Wave Analysis of Inhomogeneous Cylindrical

and Spherical Systems............................................................................................... 253
7.1 Introduction.................................................................................................... 253
7.2 Rigorous Coupled Wave Analysis Circular Cylindrical Systems.................. 254
7.3 Rigorous Coupled Wave Analysis Mathematical Formulation...................... 255
7.3.1 Introduction....................................................................................... 255
7.3.2 Basic Equations................................................................................. 256
7.3.3 Numerical Results............................................................................. 261
7.4 Anisotropic Cylindrical Scattering................................................................264
7.4.1 Introduction.......................................................................................264
7.4.2 State Variable Analysis..................................................................... 265
7.4.3 Numerical Results............................................................................. 269
7.5 Spherical Inhomogeneous Analysis............................................................... 275
7.5.1 Introduction....................................................................................... 275
7.5.2 Rigorous Coupled Wave Theory Formulation.................................. 275
7.5.3 Numerical Results............................................................................. 283
Problems.................................................................................................................... 289
References................................................................................................................. 291
x Contents

Chapter 8 Rigorous Coupled Wave Analysis of Inhomogeneous Bipolar

Cylindrical Systems.................................................................................................. 293
8.1 Introduction.................................................................................................... 293
8.2 RCWA Bipolar Coordinate Formulation........................................................ 296
8.2.1 Bipolar and Eccentric Circular Cylindrical, Scattering
Region Coordinate Description......................................................... 296
8.2.2 Bipolar RCWA State Variable Formulation...................................... 297
8.2.3 Second-Order Differential Matrix Formulation................................ 298
8.2.4 Thin-Layer, Bipolar Coordinate Eigenfunction Solution.................. 299
8.3 Bessel Function Solutions in Homogeneous Regions
of Scattering System................................................................................... 301
8.4 Thin-Layer SV Solution in the Inhomogeneous Region
of the Scattering System................................................................................. 301
8.5 Matching of EM Boundary Conditions at Interior–Exterior
Interfaces of the Scattering System................................................................302
8.5.1 Bipolar and Circular Cylindrical Coordinate Relations....................302
8.5.2 Details of Region 2 (Inhomogenous Region)–Region 3
(Homogenous Interior Region) EM Boundary Value Matching....... 303
8.5.3 Region 0 (Homogenous Exterior Region)–Region 2
(Inhomogenous Region) EM Boundary Value Matching.................306
8.5.4 Details of Layer-to-Layer EM Boundary Value Matching
in the Inhomogeneous Region...........................................................306
8.5.5 Inhomogeneous Region Ladder-Matrix............................................308
8.6 Region 1–Region 3 Bessel–Fourier Coefficient Transfer Matrix...................309
8.7 Overall System Matrix................................................................................... 312
8.8 Alternate Forms of the Bessel–Fourier Coefficient Transfer Matrix............. 313
8.9 Bistatic Scattering Width................................................................................ 314
8.10 Validation of Numerical Results.................................................................... 315
8.11 Numerical Results—Examples of Scattering from Homogeneous
and Inhomogeneous Material Objects............................................................ 315
8.12 Error and Convergence Analysis.................................................................... 321
8.13 Summary, Conclusions, and Future Work...................................................... 327
Problems.................................................................................................................... 328
Appendix 8.A............................................................................................................ 333
Appendix 8.B............................................................................................................ 336
References................................................................................................................. 338

Chapter 9 Bipolar Coordinate RCWA Computational Examples and Case Studies................. 341
9.1 Introduction.................................................................................................... 341
9.2 Case Study I: Fourier Series Expansion, Eigenvalue and Eigenfunction
Analysis, and Transfer Matrix Analysis......................................................... 343
9.3 Case Study II: Comparison of KPE BA, BC Validation Methods,
and SV Methods for Relatively Small Diameter Scattering Objects ............ 353
9.4 Case Study III: Comparison of BA, BC, and SV Methods for Gradually,
Stepped-Up, Index Profile Scattering Objects................................................ 356
9.5 Case Study IV: Comparison of BA, BC, and SV Methods for
Mismatched, Index Profile, Scattering Objects.............................................. 368
9.6 Case Study V: Comparison of BA, BC, and SV Methods for Gradually,
Stepped-Up, Index Scattering Objects with High Index Core....................... 377
Contents xi

9.7 Case Study VI: Calculation and Convergence Analysis of EM Fields

of an Inhomogeneous Region Material Object Using the SV Method,
Δε = 1, α = 5.5, Λ = 0, Example .................................................................... 384
9.8 Case Study VII: Calculation and Convergence Analysis of EM Fields
of an Inhomogeneous Region Material Object Using the SV Method,
Δε = 0.4, α = 5.5, Λ = 0, Example.................................................................. 388
9.9 Case Study VIII: Comparison of Homogeneous and Inhomogeneous
Region Bistatic Line Widths........................................................................... 393
9.10 Case Study IX: Conservation of Power Analysis........................................... 396
Appendix 9.A: Interpolation Equations....................................................................405
References.................................................................................................................407
Preface
Finding exact solutions of problems in electromagnetics (EM) and optics has become an increas-
ingly important area of research. The analysis and design of modern applications in optics and
those in traditional EM demand increasingly similar numerical computations due to a reduction of
feature sizes in optics. In EM, a large amount of research concentrates on numerical analysis tech-
niques such as the method of moments, finite element analysis, and finite difference analysis. In the
field of optics (a part of EM), a large amount of research has been conducted on beam propagation
methods (BPMs), analysis of thin and thick diffraction gratings, photorefractive (PR) materials and
PR-induced diffraction gratings, chiral materials, and photonic band gap (PBG) structures. Also,
over the last few years, extensive research in microwaves, terahertz, and optics has been conducted
in the area of metamaterials, i.e., materials that show unusual properties such as negative refraction,
amplification of nonpropagating waves, contradirectionality of power flow and phase velocities, etc.
From the late 1970s to the present, an extremely important technique called the state variable
method has been developed by various researchers for the analysis of planar diffraction gratings;
this technique is known as the rigorous coupled wave analysis (RCWA). It is based on expanding
Maxwell’s equations in periodic media in a set of Floquet harmonics, from which the unknown
expansion variable is arranged in state variable form by which all unknowns of the system may be
solved. For planar diffraction gratings, this technique has proved to be very effective, providing a
fast, accurate solution and involving only a small matrix and eigenvalue equation to find the solu-
tion. A variant of the RCWA, called the transfer matrix method (TMM), is often used in the analysis
of periodic layered structures, such as PBGs.
In the area of control theory and applications, the state variable method has been widely applied
and in fact forms the very foundation for modern control. In EM, the state variable method, although
a powerful analysis tool, has seen much less application. When used, it is applied in conjunction
with other methods (e.g., the spectral domain method, transmission ladder techniques, k-space anal-
ysis techniques, and the spectral matrix method) and is rarely listed as a state variable method. In
this book, we illustrate the application of the state variable method to various problems in EM and
optics, with special emphasis on the analysis of diffraction gratings using RCWA.
This book aims to introduce students and researchers to a variety of spectral computational tech-
niques including k-space theory, Floquet theory, TMM, and BPM, which are then used to analyze
a variety of EM and optical systems. Examples include analysis of radiation through isotropic and
anisotropic material slabs, planar diffraction gratings in isotropic and anisotropic media, propa-
gation through inhomogeneous linear and nonlinear media, periodic layered structures including
positive index material (PIM)–negative index material (NIM) stacks, radiation and scattering from
three-dimensionally inhomogeneous cylindrical and spherical structures, and diffraction from PR
materials. Many of the above cases are analyzed using a full-field vector approach to solve Maxwell’s
equations in anisotropic media where a standard wave equation approach is intractable. Efficient
and rapidly convergent numerical algorithms are presented. The problems in this book have been
analyzed using computational tools such as Fortran, MATLAB®, COMSOL®, and RSoft.
Since the publication of the first edition, several interesting new developments, such as meta-
materials, nanostructures, photonic band gap materials, chiral materials, biomaterials, etc., have
taken place in the fields of EM and optics. The second edition introduces these novel materials
and outlines EM analysis and wave propagation through these materials. To this end, we have
added sections to existing chapters, rearranged the chapters, and added new chapters. For instance,
scalar wave propagation and analysis techniques now appear in Chapter 2, with new applica-
tions to induced transmission and reflection gratings in PR materials, analyzed using BPM, both

xiii
xiv Preface

unidirectional and bidirectional. In Chapter 3, we introduce concepts of EM as applied to chiral

materials, metamaterials (NIMs), and photonic band gap structures. In this regard, the concepts of
Poynting vector and stored energy, and energy, group, and phase velocities are clearly developed. In
Chapter 4, we provide a concise review of k-space state variable formulation, including applications
to anisotropic planar systems. In Chapter 5, we present the full-field rigorous coupled wave analysis
of planar diffraction gratings, including application to H-mode, E-mode, crossed gratings; single-
and multilayer diffraction grating analysis; and diffraction from anisotropic gratings. In Chapter 6,
spectral techniques and RCWA are applied to the analysis of dynamic wave-mixing in PR materials
with induced transmission and reflection gratings. In Chapters 7 and 8, we use the RCWA algo-
rithm to analyze cylindrical and spherical systems using circular, bipolar cylindrical, and spherical
coordinates. Finally, in Chapter 9, we present several RCWA computational case studies involving
scattering from spatially inhomogeneous eccentric circular cylinders, solved in bipolar coordinates.
In many examples throughout this book, we apply the complex Poynting theorem or the forward-
scattering (optical) theorem to validate numerical solutions by verifying power conservation.
This book is intended primarily for senior and graduate students in electrical engineering, opti-
cal engineering, and physics. It should be useful for researchers in optics specializing in holography,
gratings, nonlinear optics, PR materials, metamaterials, chiral media, etc., and researchers in EM
working on antennas, propagation and scattering theory, or EM numerical methods. It should also
be of interest to the military, industry, and academia, and to all interested in finding exact solutions
to various types of EM problems. The book should be ideal for either classroom adoption or as an
ancillary reference in graduate level courses like numerical methods in EM, diffractive optics, and
EM scattering theory to name a few.
We would also like to acknowledge the support and encouragement of our wives, Elizabeth
Jarem and Noriko Banerjee, and our parents and families during the writing of this book. PPB
acknowledges the assistance of Drs. G. Nehmetallah and R. Aylo with proofreading, and with the
solutions for some of the problems.

For MATLAB® and Simulink® product information, please contact:

The MathWorks, Inc.

3 Apple Hill Drive
Natick, MA, 01760-2098 USA
Tel: 508-647-7000
Fax: 508-647-7001
E-mail: [email protected]
Web: www.mathworks.com
Authors
John M. Jarem received his BSEE, MSEE, and PhD from Drexel University, Philadelphia,
Pennsylvania, in 1971, 1972, and 1975, respectively. His research interests include optical and
microwave diffraction theory, numerical electromagnetics, and Green’s function theory. Jarem has
been a professor of electrical and computer engineering at the University of Alabama in Huntsville,
Huntsville, Alabama, since 1987. Prior to this, he was a professor of electrical engineering in the
Department of Electrical Engineering at the University of Texas at El Paso (1986–1987), an asso-
ciate professor of electrical engineering at the University of Texas at El Paso (1981–1986), and an
assistant professor of electrical engineering at the University of Petroleum and Minerals, Dhahran,
Saudi Arabia (1975–1981). Jarem is the coauthor of one book and one book chapter, and is the
author/coauthor of 51 refereed journal articles and 36 conference papers. He has been the PhD
thesis advisor for four PhD students and the master’s thesis advisor for 18 master’s students. He is
also a senior member of the IEEE.

Partha P. Banerjee is a professor of electro-optics and electrical and computer engineering (ECE)
at the University of Dayton, where he was chair of the ECE department from 2000 to 2005. Prior
to that, he was with the ECE department at the University of Alabama in Huntsville from 1991 to
2000 and at Syracuse University from 1984 to 2001. He received his BTech from the Indian Institute
of Technology in 1979, and his MS and PhD from the University of Iowa in 1980 and 1983, respec-
tively. His research interests include optical processing, nonlinear optics, photorefractive materials,
and acousto-optics. He has authored/coauthored four books, several book chapters, over 100 refer-
eed journal articles, and many conference papers. He is fellow of the Optical Society of America
(OSA) and of the Society of Photo-optical Instrumentation Engineers (SPIE), and is a senior mem-
ber of the Institute of Electrical and Electronics Engineers (IEEE). He received the National Science
Foundation (NSF) Presidential Young Investigator award in 1987.

xv
 1 Mathematical Preliminaries
1.1 INTRODUCTION
Popular t-shirts advertising Maxwell’s equations do not go beyond merely stating them. In this
book, we enter into a little more depth and solve these equations for analyzing various electromag-
netic (EM) and optical problems, e.g., diffraction gratings, radiation and scattering from dielectric
objects, photorefractive materials, metamaterials, and chiral materials. The emphasis on finding the
solutions in our text concerns the use of Fourier and state variable analyses, and beam propagation
methods. In this chapter, we briefly review some mathematical techniques pertinent to the analyses
presented in later chapters.

1.2 FOURIER SERIES AND ITS PROPERTIES

It is easy to show that a set of exponential functions {exp jnKx}, n = 0, ±1, ±2, etc., where j = −1,
is orthogonal over an interval (x0, x0 + 2π/K) for any value of x0. The orthogonality can be demon-
strated by considering the integral

x0 + 2 π /K
⎛ 2π ⎞
I=
∫x0
exp( jnKx ) exp( − jmKx )dx ≡ ⎜ ⎟ δ m,n
⎝ K⎠
(1.1)

where δm,n is the Kronecker delta function:

⎧1, m = n
δ m,n = ⎨ (1.2)
⎩0, m ≠ n

Using this, a function f (x) can be expanded in a Fourier series over an interval (x0, x0 + 2π/K) as

f ( x) = ∑ F exp( jnKx)
n = −∞
n (1.3)

Multiplying Equation 1.3 by exp(−jmKx), integrating over the interval (x0, x0 + 2π/K), interchanging
the summation and the integral, and using Equations 1.1 and 1.2, we get

x0 + 2π /K ∞ x0 + 2 π /K

∫
x0
f ( x ) exp( − jmKx )dx = ∑F ∫
n = −∞
n
x0
exp( jnKx ) exp( − jmKx )dx

∞
⎡ 2π ⎤ ⎛ 2π ⎞
= ∑ F ⎢⎣⎛⎜⎝ K ⎞⎟⎠ δ
n = −∞
n m,n ⎥ = ⎜⎝ K ⎟⎠ Fm
⎦

1
2 Computational Methods for Electromagnetic and Optical Systems

Now, replacing m by n,
x0 + 2 π /K
⎛ K⎞
Fn = ⎜ ⎟
⎝ 2π ⎠ ∫
x0
f ( x ) exp( − jnKx )dx (1.4)

Note that if a function fe(x) is defined as fe(x) = f(x) exp(−jαx), where α is a constant, then over the
interval (x0, x0 + 2π/K), it can be written as
∞

fe ( x ) = ∑ F exp(− jk
n = −∞
n xn x ), k xn = α − nK (1.5)

Li refers to this expansion in Equation 1.5 as a pseudo-Fourier series of fe(x) [1].

If two functions f (x) and g(x) having Fourier series expansions

∞ ∞

f ( x) = ∑
n = −∞
Fn exp( jnKx ), g( x ) = ∑ G exp( jnKx)
n = −∞
n (1.6)

over the same interval are multiplied, the product function h(x) has a Fourier series expansion
∞

h( x ) = ∑ H exp( jnKx)
n = −∞
n (1.7)

over the same interval. We can find the Fourier coefficients of h(x) in the following way:

∞ ∞

h( x ) = f ( x )g( x ) = ∑
n = −∞
Fn exp( jnKx ) ∑G
m = −∞
m exp( jmKx)

∞ ∞ ∞ ∞

= ∑∑
n = −∞ m = −∞
FnGm exp( j (n + m)Kx ) = ∑∑F
l = −∞ m = −∞
l −m Gm exp( jlKx )

≡ ∑ H exp( jlKx)
l = −∞
l (1.8)

The limits on l are −∞ to +∞ since l = m + n, and m and n each have limits −∞ to +∞. Hence the
Fourier coefficients Hl of h(x) can be expressed as

Hl = ∑F
m = −∞
l −m Gm (1.9)

Equation 1.7 is sometimes referred to as the Laurent rule [1]. To be more precise, Equations 1.8 and
1.9 should be understood in the following sense [1]:

N L M
⎛ ⎞
h( x ) = lim
N →∞ ∑
l=−N
Hl exp( jlKx ) = lim
L →∞ ∑ ⎜
l=−L ⎝
lim
M →∞
m= − M
∑
Fl − mGm ⎟ exp( jlKx )
⎠
(1.10)
Mathematical Preliminaries 3

The above equation, in the way it is written, emphasizes two important points. First, the two limits
L and M are independent of each other, and the inner limit has to be taken first. Second, the upper
and lower bounds in each sum should tend to infinity simultaneously [1].

1.3 FOURIER TRANSFORM

The one-dimensional temporal Fourier transform of a square-integrable function f (t) is given as [2]

fˆ (ω) =
∫ f (t ) exp(− jωt )dt
−∞
(1.11)

The inverse Fourier transform is

∞
1
f (t ) =
2π ∫ fˆ (ω) exp(+ jωt )dω
−∞
(1.12)

The one-dimensional spatial Fourier transform of a square-integrable function f(x) is given as [2]

∞
f (k x ) =
∫ f ( x) exp( jk x)dx
−∞
x (1.13)

The inverse Fourier transform is

∞
1
f ( x) =
2π ∫ f (k ) exp(− jk x)dx
−∞
x x (1.14)

The definitions for the forward and backward transforms are consistent with the engineering con-
vention for a traveling wave, as explained in [2]. If f (x) denotes a phasor EM field quantity, multipli-
cation by exp jωt gives a collection or spectrum of forward traveling plane waves.
The two-dimensional extensions of Equations 1.13 and 1.14 are

∞ ∞
f (k x , k y ) =
∫ ∫ f ( x, y) exp( jk x + jk y)dx dy
−∞ −∞
x y (1.15)

∞ ∞
1
f ( x, y ) =
(2 π )2 ∫ ∫ f (k , k ) exp(− jk x − jk y)dx dy
−∞ −∞
x y x y (1.16)

In many EM applications, the function f (x, y) represents the transverse profile of an EM wave at a
plane z. Hence in (1.15) and (1.16), f (x, y) and f̃ (k x, k y) have z as a parameter. For instance (1.16)
becomes

∞ ∞
1
f ( x, y, z ) =
(2 π )2 ∫ ∫ f (k , k ; z) exp(− jk x − jk y)dx dy
−∞ −∞
x y x y (1.17)
4 Computational Methods for Electromagnetic and Optical Systems

The usefulness of this transform lies in the fact that when substituted into Maxwell’s equations,
one can reduce the set of three-dimensional PDEs to a set of one-dimensional differential equa-
tions (ODEs) for the spectral amplitudes f̃ (k x, k y; z).

1.4 HANKEL TRANSFORM

In the special case where the function f (x, y) in Equation 1.15 is circularly symmetric, i.e.,
f ( x, y) = g(ρ), ρ = x 2 + y 2 , we can introduce new variables to change from rectangular to polar
coordinates, in both the spatial and the spatial frequency domains:

x = ρ cos φ, y = ρ sin φ; k x = kρ cos φʹ, k y = kρ sin φʹ (1.18)

Substituting (1.18) into (1.15), we get after some algebra:

∞ 2π

g(kρ , φʹ ) =
∫ ∫ g(ρ) exp{ jk ρ cos(φ − φʹ)}ρ dρ dφ
0 0
ρ (1.19)

It turns out that the result of the integration w.r.t. ϕ is independent of ϕ′ since [3]

2π
1
J 0 (kρρ) =
2π ∫ exp{ jk ρ cos(φ − φʹ)}dφ
0
ρ (1.20)

Substituting (1.20) into (1.19),

∞
1
g(kρ ) =
2π ∫
g(ρ)J 0 (kρρ)ρ dρ
0
(1.21)

Equation 1.21 is called the Fourier–Bessel transform or the Hankel transform of the circularly sym-
metric function g(ρ) [4]. The inverse transform is given by

∞
1
g(ρ) =
2π ∫
g(kρ )J 0 (kρρ)kρ dkρ
0
(1.22)

1.5 DISCRETE FOURIER TRANSFORM

Given a discrete function f (nΛ), n = 0,… N − 1, a corresponding periodic function fp(nΛ) with period
NΛ can be formed as [5]
∞

f p ( nΛ ) = ∑ f (nΛ + rN Λ)
r = −∞
(1.23)
Mathematical Preliminaries 5

The discrete function f (nΛ) may be formed by the discrete values of a continuous function f (x) evalu-
ated at the points x = nΛ.
The discrete Fourier Transform (DFT) of fp(nΛ) is defined as

N −1
2π
f˜p (mK ) = ∑ f (nΛ) exp( jmnK Λ),
n=0
p K=
NΛ
(1.24)

The inverse DFT is defined as

N −1
1
f p ( nΛ ) =
N ∑ f˜ (mK ) exp(− jmnK Λ)
n=0
p (1.25)

For properties of the DFT, e.g., linearity, symmetry, and periodicity, as well as relationship to the
z-transform, the Fourier transform, and the Fourier series, the readers are referred to any standard
book on digital signal processing [5].
For the purposes of this book, the DFT is a way of numerically approximating the continuous
Fourier transform of a function. The DFT is of interest because it can be efficiently and rapidly eval-
uated by using standard Fast Fourier Transform (FFT) packages. The direct connection between
the continuous Fourier transform and the DFT is given in the following. For a function f (x) and its
continuous Fourier transform ƒ̃(k x ),

1 mK < π
f˜p (mK ) ≅ f˜(mK ), (1.26)
Λ Λ

In Equation 1.26, ƒ˜p(mK) is defined, as in Equation 1.24, to be the DFT of fp(nΛ). The equality holds
for the fictitious case when the function is both space and spatial frequency limited.

1.6 REVIEW OF EIGENANALYSIS

A large amount of the computations in this book are based on determining the eigenvalues and
eigenvectors of a matrix A . Therefore, this section will briefly review the methods and techniques
that are associated with numerically solving this problem [6,7]. The matrix A, which is a square
matrix in general, transforms a column vector –x into a matrix y = A x . We are interested in those
vectors x that transform into themselves and satisfy

Ax = qx (1.27)

These column vectors are called the eigenvectors of the system. The values of q which satisfy (1.27)
are known as the eigenvalues, the characteristic values, or the latent roots of the matrix A. Equation
1.27 may be written as a linear set of equations as

(a11 − q) x1 + a12 x2 + a13 x3 + ⋅⋅⋅ + a1n xn = 0

a12 x1 + (a22 − q) x2 + a23 x3 + ⋅⋅⋅ + a2 n xn = 0

(1.28)


an1 x1 + an 2 x2 + an3 x3 + ⋅⋅⋅ + (ann − q ) xn = 0

6 Computational Methods for Electromagnetic and Optical Systems

A nontrivial solution exists for the above equations if and only if

P(q) ≡ det (q I − A) = 0 (1.29)

where I is the identity matrix. The result of Equation 1.29 is an nth-order polynomial called the
characteristic equation or eigenvalue equation. The equation is given by

P(q) = q n + a1q n −1 + a2q n − 2 + ⋅⋅⋅ + an −1q + an (1.30)

The roots of this equation are the eigenvalues of the matrix A . When the roots are all unequal to
one another, the roots or eigenvalues are called distinct. When the eigenvalue occurs m times, the
eigenvalue is a repeated value of order m. When the root has a real and nonzero imaginary part, the
roots occur in complex conjugate pairs. In factored form, Equation 1.30 may be written as

P(q) = (q − q1 )(q − q2 ) ⋅⋅⋅ (q − qn ) (1.31)

The coefficients of the eigenvalue equation may be found directly from the matrix A. For
instance, setting q to zero in (1.31), we find

P(0) = an = det(− A) = ( −1)n det( A) = ( −1)n q1q2 ⋅⋅⋅ qn (1.32)

and thus from Equation 1.32,

det( A) = q1q2 ⋅⋅⋅ qn (1.33)

The coefficient a1 may be found by expanding the factored characteristic equation and comparing
the polynomial coefficients of the resulting equation. For example, if n = 2,

P(q) = q 2 + a1q + a2 = (q − q1 )(q − q2 ) = q 2 − (q1 + q2 ) q + q1 q2 (1.34)

and thus

a1 = −(q1 + q2 ) (1.35)

after equating coefficients. For general n,

a1 = −(q1 + q2 + ⋅⋅⋅ + qn ) (1.36)

If the determinant is expanded, we also find that the determinant is the negative sum of diagonal
coefficients, that is,

a1 = −(a11 + a22 + ⋅⋅⋅ + ann ) (1.37)

The quantity in parentheses is an important quantity and is called the trace of A:

Tr( A) = a11 + a22 + ⋅⋅⋅ + ann = q1 + q2 + ⋅⋅⋅ + qn (1.38)

Mathematical Preliminaries 7

k
Let Tk = Tr ( A ). Then a useful formula for the coefficient an of the characteristic equation is

a1 = −T1

1
a2 = − (a1T1 + T2 )
2
1
a3 = − (a2T1 + a1T2 + T3 ) (1.39)
3

1
an = − (an −1T1 + an − 2T2 + ⋅⋅⋅ + a1Tn −1 + Tn )
n

For the case when the roots of P(q) are distinct, a nontrivial vector x i can be found for each root
which satisfies

(qi I − A) xi = 0 (1.40)

The matrix formed of the columns of –x i is called the modal matrix M. The name modal matrix
comes from control theory where a dynamical system can be decomposed into dynamic modes of
operation. For EM diffraction grating problems and also for EM problems that use k-space tech-
niques, the EM field solutions associated with a state variable analysis may be decoupled into spatial
mode solutions. These modes are analogous to the dynamical modes of operation encountered in
control systems.
If the eigenvalues are distinct, which is mainly the case under consideration in this text, the
modal matrix is nonsingular and therefore its inverse exists. Letting M be the modal matrix, we
may write

MQ = AM (1.41)

where Q is a diagonal matrix holding the eigenvalues qi on the diagonal. It can be shown that the
inverse of M exists, hence, Equation 1.41 gives

−1
Q = M AM (1.42)

If Q is squared, we have

2 −1 −1 −1 2
Q = ( M A M )( M A M ) = ( M A M ) (1.43)
−1
and if we further pre- and post-multiply by M and M , respectively, we have

2 2 −1
A = MQ M (1.44)

Similarly, if A is raised to the pth power, we have

p p −1
A = MQ M (1.45)
8 Computational Methods for Electromagnetic and Optical Systems

p
where Q is the diagonal matrix formed by raising each eigenvalue qi to the pth power. A matrix
polynomial N ( A) can be conveniently evaluated as

−1
N ( A) = MN (Q) M (1.46)

where linear combinations of powers of A as given by Equation 1.45 have been used. N (Q) is the
diagonal matrix formed by placing in each diagonal matrix entry the polynomial N(qi). Thus the
modal matrix provides a convenient way to quickly and accurately evaluate powers and polynomials
of the matrix A.
In this text, we will be greatly concerned with calculating the exponential function of the
matrix A. The exponential function of the matrix A, namely exp( A) , is defined as [7]

⎛ 1⎞ 1
exp( A) = I + A + ⎜ ⎟ ( A)2 + ⋅⋅⋅ + ( A)k +  (1.47)
⎝ 2⎠ k!

which is the same infinite series expansion as that used to define the exponential function exp(a).
We now review two important aids that help in the solution and evaluation of exponential matrix
and in fact any function of the matrix A. These are called the Cayley–Hamilton theorem and the
Cayley–Hamilton technique. These aids will be presented only for the cases of matrices with dis-
tinct eigenvalues.
The first theorem to be reviewed is the Cayley–Hamilton theorem. If we have a polynomial
N(q) = qn + c1qn−1 + · · · + cn−1 q + cn, then using Equation 1.46 we have

⎡ N (q1 ) 0 0 ⎤
N ( A) = M ⎢⎢ 0 N (q2 ) ⎥ M −1 (1.48)
⎥
⎢⎣ 0 N (q3 )⎥⎦

where M is the modal matrix. If the polynomial N(q) is chosen to the characteristic equation, that
is, N(q) = P(q), then N(qi ) = P(qi ) = 0; i = 1, 2, …, n and thus

⎡0 0 0 ⎤
⎢0 0 0 ⎥
P( A) = M ⎢ ⎥ M −1 = 0 (1.49)
⎢0 0 0 ⎥
⎢ ⎥
⎣ ⋅⋅⋅⎦

We thus see that the matrix A satisfies it own characteristic equation.

Another important aid to evaluating a function of a matrix, where the function is analytic over
a given range of interest, is provided by the Cayley–Hamilton technique [7]. We first consider the
case where the analytic function is a polynomial of higher degree than the characteristic polyno-
mial P(q) of order n. Let the polynomial be N(q). We consider the case where the roots (or eigen-
values) of P(q) are distinct. In this case,

N (q ) R (q )
= Q(q ) + (1.50)
P (q ) P (q )

where Q(q) is a polynomial, and R(q) is a polynomial of order n − 1 or less.

Mathematical Preliminaries 9

Multiplying by P(q), we have

N (q ) = Q(q ) P (q ) + R (q ) (1.51)

If q is an eigenvalue or root of P(q), then P(q) = 0 and N(q) = R(q). If we substitute A for q in Equation
1.51, we have

N ( A) = Q( A)P( A) + R( A) (1.52)

Since by the Cayley–Hamilton theorem the matrix P( A) = 0, it follows that

N ( A) = R( A) (1.53)

Thus a higher order polynomial matrix may be represented and evaluated using an n−1 polynomial
expression.
Consider next the case where the matrix function is a general analytic function over a region
of interest, for example, F ( A) = exp ( A) . In this case, F(q) may be expanded in an infinite power
series over the analytic region of interest. As in the case when F(q) was a polynomial, F(q) may be
written as

F (q ) = Q(q ) P (q ) + R (q ) (1.54)

where R(q) is a polynomial of order n − 1 given by

R(q) = α 0 + α1q + α 2q 2 +  + α n −1q n −1 (1.55)

Let q = q1, q2, …, qn be the distinct roots of P(qi) = 0, i = 1,…, n. We have, after evaluating
Equation 1.54,

F (q1 ) = R(q1 )

F (q2 ) = R(q2 )
(1.56)


F ( qn ) = R ( qn )

This defines a set of n × n linear equations from which the coefficients αi, i = 1,…, n may be
determined. At this point, we would like to show that the function Q(q) is analytic. To do this,
we write Q(q) as

F (q ) − R (q )
Q(q ) = (1.57)
P (q )

In this expression, we note that over the region of interest the numerator and denominator of Equation
1.57 have the same zeros. Since in Equation 1.57 all functions F(q), Q(q), P(q), and R(q) are analytic
over the range where F(q) is, we may replace q by the matrix A. We therefore have

F ( A) = Q( A)P( A) + R( A) (1.58)
10 Computational Methods for Electromagnetic and Optical Systems

Since by the Cayley–Hamilton theorem P ( A) = 0, we have

F ( A) = R( A) (1.59)

Thus we have shown that the analytic matrix function F ( A) may be evaluated by using a polynomial
matrix expression of order n − 1 as given by R ( A) in Equation 1.55.

PROBLEMS
1.1 Derive the wave equation for the electric and magnetic fields starting from Maxwell’s equations
in a homogeneous, isotropic, source-free region (see Chapter 3). How does this change if the
material is anisotropic?
1.2 Find from first principles the Fourier series coefficients for a periodic square wave s(x) of unit
amplitude and 50% duty cycle. Now find the Fourier series coefficients of s2(x): (a) from first
principles and (b) using the Laurent rule. Plot s2(x) versus x by employing the Fourier series
coefficients you found using (b). Use 5, 10, and 100 Fourier coefficients. Describe the general
trend(s).
1.3 Find the two-dimensional Fourier transform of a rectangle (rect) function of unit height and width
a in each dimension.
1.4 Show that the two-dimensional Fourier transform of a Gaussian function of width w is another
Gaussian function. Functions like this are called self-Fourier transformable. Find its width
in the spatial frequency domain. Can you think of any other functions that are self-Fourier
transformable?
1.5 Find the Hankel transform of (a) a circular function defined as g(ρ) = circ(ρ/ρ0), which has
a value of 1 within a circle of radius ρ0 and is 0 otherwise; (b) a Gaussian function given by
g(ρ) = exp−(ρ/ρ0)2.
1.6 Find the DFT of a square wave function using a software of your choice. Comment on the
nature of the spectrum numerically computed as the width of the square wave changes.
⎛ 1 20 0⎞
1.7 Find sin –A, where –A is a matrix given by ⎜ −1 7 1 ⎟ , using the Cayley–Hamilton ­theorem [7].
⎜ ⎟
⎝3 0 −2⎠

REFERENCES
1. L. Li, Use of Fourier series in the analysis of discontinuous periodic structures, J. Opt. Soc. Am. A 15,
1808–1816, 1996.
2. P.P. Banerjee and T.-C. Poon, Principles of Applied Optics, Irwin, New York, 1991.
3. M. Abramowitz and I. Stegun, Handbook of Mathematical Functions, Dover, New York, 1965.
4. I. Sneddon, The Use of Integral Transforms, McGraw-Hill, New York, 1972.
5. A. Antoniou, Digital Filters: Analysis and Design, McGraw-Hill, New York, 1979.
6. P.M. Deruso, R.J. Roy, and C.M. Close, State Variables for Engineers, John Wiley & Sons, Inc.,
New York, 1967.
7. L.A. Pipes and L.R. Harvill, Advanced Mathematics for Engineers and Scientists, McGraw-Hill,
New York, 1970.
 2 Scalar EM Beam Propagation
in Inhomogeneous Media

2.1 INTRODUCTION
In the previous chapter, we reviewed some of the mathematical preliminaries that will be useful
later on in the text. In this chapter, we discuss some of the basic concepts of scalar wave propaga-
tion, and discuss an important numerical method, called the beam propagation method (BPM), to
study propagation in linear media and in media with induced nonlinearities. Furthermore, we also
discuss propagation through induced gratings, both transmission and reflection type, in order to
assess energy coupling between participating waves. Finally, we introduce readers to an important
characterization method, called the z-scan method, which is often used to determine the focal length
of an induced lens.

2.2 TRANSFER FUNCTION FOR PROPAGATION

For simplicity, we consider the scalar wave equation

∂2 E
− v 2∇2 E = 0 (2.1)
∂t 2
and substitute
E ( x, y, z, t ) = Re{Ee ( x, y, z ) exp[ j (ω 0t − k0 z)]} (2.2)

with ω0/k0 = v. The quantity Ee is related to the phasor Ep according to

E p ( x, y, z ) = Ee ( x, y, z ) exp( − jk0 z ) (2.3)

and we will use one or the other notation according to convenience. Substituting Equation 2.2 into
Equation 2.1 and assuming that Ee is a slowly varying function of z (the direction of propagation) in
the sense that ∂ 2 Ee / ∂z 2 ∂Ee / ∂z << k0, we obtain the paraxial wave equation [1]

∂ Ee
2 jk0 = ∇2⊥ Ee (2.4)
∂z

where ∇2⊥ denotes the transverse Laplacian. Equation 2.4 describes the propagation of the envelope
Ee(x, y, z) starting from the initial profile Ee│z= 0 = Ee0(x, y).
Equation 2.4 can be solved readily using Fourier transform techniques. Assuming Ee to be
Fourier transformable, we can employ the definition of the Fourier transform as in Equation 1.15

Ee (k x , k y ; z) = ℑx , y{Ee ( x, y, z)} =

∫ E ( x, y, z) exp( jk x + jk y)dx dy
−∞
e x y (2.5)

11
12 Computational Methods for Electromagnetic and Optical Systems

and its properties to transform Equation 2.4 into the ordinary differential equation (ODE)

dEe j
= (k x2 + k y2 )Ee (2.6)
dz 2k0

We can easily solve Equation 2.6 to give

⎧ j (k 2 + k y2 )z ⎫
Ee (k x , k y ; z) = Ee 0 (k x , k y ) exp ⎨ x ⎬ (2.7)
⎩ 2k0 ⎭

where Ẽe0(k x, k y) is the Fourier transform of Ee0(x, y). We can interpret Equation 2.7 in the following
way: Consider a linear system with an input spectrum of Ẽe0(k x, k y) at z = 0 where the output spec-
trum at z is given by Ẽe(k x, k y; z). The spatial frequency response of the system, which we will call
the paraxial transfer function for propagation is then given by

Ee ⎧ j (k x2 + k y2 )z ⎫
≡ H ( k , k ; z ) = exp ⎨ ⎬ (2.8)
Ee 0
x y
⎩ 2k0 ⎭

As we will show later, in the split-step BPM, we model propagational diffraction by means of the
transfer function for propagation derived above. For more exact calculations, the nonparaxial trans-
fer function can be used. This may be derived starting from the nonparaxial wave equation, but will
not be presented here for the sake of simplicity.
Incidentally, the inverse Fourier transform of the transfer function for propagation yields the
impulse response for propagation. Starting from the paraxial transfer function for propagation
which resembles a complex Gaussian, the inverse Fourier transform is a complex Gaussian as well,
and has the form

jk0 ⎡ − j ( x 2 + y 2 )k0 ⎤
h( x, y, z ) = exp ⎢ ⎥ (2.9)
2πz ⎣ 2z ⎦

This, when convolved with the initial beam profile, yields the profile of the diffracted beam in the
spatial domain directly. This convolution integral is in fact the Fresnel diffraction formula.

2.3 SPLIT-STEP BEAM PROPAGATION METHOD

If we wish to consider propagation in a material where the propagation constant or equivalently
the refractive index is a function of position, either due to profiling of the material itself (such as
a graded index fiber or a grating) or due to induced effects such as third order nonlinearities, the
paraxial wave equation changes to

∂Ee 1
= ∇2⊥ Ee − j Δnk0 Ee (2.10)
∂z 2 jk0
Scalar EM Beam Propagation in Inhomogeneous Media 13

The quantity Δn is the change in the refractive index over the ambient refractive index n 0 = c/v,
where c is the velocity of light in vacuum. Equation 2.10 is a modification of Equation 2.4 and
can be derived from the scalar wave equation when the propagation constant or equivalently the
velocity of the wave is a function of (x, y, z) explicitly as in gratings or fibers, or implicitly such as
through the intensity dependent refractive index. An alternative, though heuristic, way to justify
the presence of the additional term on the RHS of Equation 2.10 is to note that in the absence of
diffraction (the first term on the RHS), the solution of the equation is of the form Ee ∝ exp[−jΔnk0z],
which explicitly shows the additional phase change due to propagation in the perturbed refractive
index.
The paraxial propagation equation (2.10) is a partial differential equation (PDE) that does not
always lend itself to analytical solutions, except for some very special cases involving special spatial
variations of Δn, or when as in nonlinear optics, one looks for particular soliton solution of the result-
ing nonlinear PDE using exact integration or inverse scattering methods. Numerical approaches are
often sought for to analyze beam (and pulse) propagation in complex systems such as optical fibers,
volume diffraction gratings, Kerr and photorefractive (PR) media, etc. A large number of numerical
methods can be used for this purpose. The pseudospectral methods are often favored over finite dif-
ference methods due to their speed advantage. The split-step BPM is an example of a pseudospectral
method.
To understand the philosophy behind the BPM, it is useful to rewrite Equation 2.10 in the
form [2,3]

∂Ee
= ( Dˆ + Sˆ )Ee (2.11)
∂z

where D̂, Ŝ are a linear differential operator and a space dependent or nonlinear operator respectively
(see, for instance, the structure of Equation 2.10). Thus, in general, the solution of Equation 2.11 can
be symbolically written as

Ee ( x, y, z + Δz) = exp[( Dˆ + Sˆ )Δz ]Ee ( x, y, z ) (2.12)

if D̂ , Ŝ are assumed to be z-independent. Now for two noncommuting operators D̂ , Ŝ,

⎛ ⎛ 1⎞ ⎞
exp( Dˆ Δz) exp(SˆΔz ) = exp ⎜ Dˆ Δz + SˆΔz + ⎜ ⎟ [ Dˆ , Sˆ ]( Δz)2 + ⎟ (2.13)
⎝ ⎝ 2 ⎠ ⎠

according to the Baker–Hausdorff formula [2], where [D̂, Ŝ] represents the commutation of D̂, Ŝ.
Thus, up to second order in Δz,

exp( Dˆ Δz + Sˆ Δz ) ≅ exp( Dˆ Δz ) exp(Sˆ Δz ) (2.14)

which implies that in Equation 2.13 the diffraction and the inhomogeneous operators can be treated
independent of each other.
The action of the first operator on the RHS of Equation 2.14 is better understood in the spectral
domain. Note that this is the propagation operator that takes into account the effect of diffraction
between planes z and z + Δz. Propagation is readily handled in the spectral or spatial frequency
domain using the transfer function for propagation written in Equation 2.8 with z replaced by Δz.
The second operator describes the effect of propagation in the absence of diffraction and in the
presence of medium inhomogeneities, either intrinsic or induced, and is incorporated in the spatial
14 Computational Methods for Electromagnetic and Optical Systems

Initial profile
Ee (x, y; z = 0)

~
Ee (kx , ky , z) = x,y {Ee (x, y, z)}

~ ~
Ee (kx , ky ; z + Δz) = Ee0 (kx , ky , z)exp{ j(k x2 + k 2y )Δz/2k0}

–1 ~
E΄e (x, y, z + Δz) = x,y {Ee (kx , ky ; z + Δz)}

ˆ
Ee (x, y, z + Δz) = E΄e (x, y, z + Δz)exp{SΔz}

FIGURE 2.1 Flow diagram for the BPM split-step method.

domain. A schematic block diagram of the BPM method in its simplest form is shown in Figure
2.1. There are other modifications to the simple scheme, viz., the symmetrized split-step Fourier
method, and the leap-frog techniques; these are discussed in detail elsewhere [2].

2.4 BEAM PROPAGATION IN LINEAR MEDIA

In this section, we will illustrate various cases where the BPM can be used to analyze propagation in
inhomogeneous media. While most of the examples will be connected with beam propagation, we
must point out to readers that the method can be used to analyze pulse propagation as well, simply
by replacing z in Equation 2.11 with t (time), and making the linear spatial transverse differential
operator a similar differential operator in z. With this modification, Equation 2.11 can model the
propagation of one dimensional longitudinal pulse through an optical fiber with arbitrary group
velocity dispersion. For details, we refer the readers to Agrawal [2].

2.4.1 Linear Free-Space Beam Propagation

The propagation of Gaussian beams through free space can be readily analyzed analytically using
the transfer function of propagation. For a traditional scalar (or x- or y-polarized) Gaussian beam,
Ee(x, y; z = 0) is a Gaussian function of the form

Ee ( x, y; 0) = E0 exp[ −α( x 2 + y 2 )] (2.15)

with a Fourier transform

⎛ π⎞ ⎡ (k 2 + k y2) ⎤
Ee (k x , k y ; 0) = E0 ⎜ ⎟ exp ⎢ − x ⎥ (2.16)
⎝ α⎠ ⎣ 4α ⎦
Scalar EM Beam Propagation in Inhomogeneous Media 15

where α may be complex. Hence, using Equation 2.8,

Ee (k x , k y ; z) = Ee (k x , k y ; 0) H (k x , k y ; z)

⎛ π⎞ ⎡ (k 2 + k y2) ⎤ ⎡ j (k x2 + k y2 )z ⎤
= E0 ⎜ ⎟ exp ⎢ − x ⎥ exp ⎢ ⎥
⎝ α⎠ ⎣ 4α ⎦ ⎣ 2k0 ⎦

⎛ π⎞ ⎡ j (k 2 + k y2 )q ⎤
≡ E0 ⎜ ⎟ exp ⎢ x ⎥ (2.17)
⎝ α⎠ ⎣ 2k0 ⎦

whose inverse transform is a Gaussian in transverse spatial dimensions:

jk0 ⎛ π⎞ ⎡ jk ( x 2 + y 2 ) ⎤
Ee ( x, y; z) = E0 ⎜ ⎟ exp ⎢ − 0 ⎥ (2.18)
2πq ⎝ α ⎠ ⎣ 2q ⎦

Equation 2.17 defines the well-known q-parameter of the Gaussian beam:

jk0
q = z+ (2.19)
2α

If the initial Gaussian beam has waist w0 and plane wavefronts, α = 1/w02 and q = z + jzR , where
zR = k0 w02 /2 is commonly referred to as the Rayleigh range of the Gaussian beam. Upon simplify-
ing Equation 2.18 using Equation 2.19, we get the standard expression for the diffracted Gaussian
beam profile as

w0 ⎡ ( x 2 + y2 ) ⎤ ⎡ jk0 ( x 2 + y 2 ) ⎤ − jφ( z )
Ee ( x, y; z) = E0 exp ⎢ − 2 ⎥ exp ⎢ − ⎥e (2.20)
w( z ) ⎣ w ( z) ⎦ ⎣ 2 R( z ) ⎦

where

⎡ ⎛ z ⎞2⎤ ( z 2 + z R2 ) ⎛ z⎞
2 2
w ( z) = w ⎢1 + ⎜ ⎟ ⎥ , R( z) =
0 , φ( z ) = − tan −1 ⎜ ⎟ (2.21)
⎢⎣ ⎝ z R ⎠ ⎥⎦ z ⎝ zR ⎠

Incidentally, if a Gaussian beam with initially plane wavefronts passes through a (thin) lens of focal
length f, the latter introduces a quadratic phase curvature which can be accommodated by multiply-
ing the initial Gaussian with a complex exponential exp[ jk0(x2 + y2)/2f ]. In this case, the scalar optical
field immediately behind the lens (z = 0+) can be written in the form as in Equation 2.15 above, but
where α is redefined as

1 jk ⎛ 1 ⎞ ⎛ 1 − jz R ⎞
α= − 0 = ⎜ 2⎟⎜ (2.22)
w0 2 f ⎝ w0 ⎠ ⎝ f ⎟⎠
2
16 Computational Methods for Electromagnetic and Optical Systems

1
0.8 1
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0
80 0
60 80 80
60 60 80
40 40 60
20 40 40
20 20 20
(a) 0 0 (b) 0 0

FIGURE 2.2 Diffraction of a Gaussian beam during free-space propagation: (a) profile at z = 0 (plane
wavefronts assumed) and (b) profile at z = z R, where z R is the Rayleigh range of the original Gaussian beam.

The paraxial propagation of a Gaussian beam through an optical system involving translation and
lensing can be equivalently studied in terms of the variation of the q-parameter. For instance, due
to propagation in air through a distance z alone, the change in the q-parameter can be readily found
from Equation 2.17 as

qnew = qold + z (2.23a)

On the other hand, if a Gaussian beam is incident on a lens of focal length f, the change in the q-parameter
can be readily found from Equation 2.18 and the phase transformation by the lens given above as

1 1 1
= − (2.23b)
qnew qold f

Equation 2.23a and b will be later used to analyze propagation of Gaussian beams through media
which have a distributed lensing effect, such as a nonlinear medium.
We now demonstrate the application of BPM to Gaussian beam propagation through free space.
In this case, the inhomogeneous operator is zero, and propagation from a plane z  =  0 to arbitrary
z can indeed be performed in one step in this case. However, in the example we provide, we use the
split-step method to convince readers that the result is identical to what one would get if the propaga-
tion was covered in one step. In Figure 2.2, we show the profile of a diffracted Gaussian beam after
propagation through free space, and the results agree with the physical intuition of increased width
and decreased on-axis amplitude during propagation as well as the analytical results in Equations
2.20 and 2.21.

2.4.2 Propagation of Gaussian Beam through Graded Index Medium

A graded index medium has a refractive index variation of the form

n = n0 + n( 2 ) ( x 2 + y 2 ) (2.24)

where n0 denotes the intrinsic refractive index of the medium, and n(2) is a measure of the gradation
in the refractive index.
In this case, the operator Ŝ becomes

Sˆ = − jk0 n( 2 ) ( x 2 + y 2 ) (2.25)
Scalar EM Beam Propagation in Inhomogeneous Media 17

60 60

50 50

40 40

30 30

20 20

10 10

10 20 30 40 50 60 10 20 30 40 50 60

60

50

40

30

20

10

10 20 30 40 50 60

FIGURE 2.3 Contour plots showing periodic focusing of an initial Gaussian profile.

Propagation of a Gaussian beam in a medium with a graded index profile is shown in Figure 2.3.
The contour plots show the initial (Gaussian) beam profile, the beam profile where the initial
Gaussian attains its minimum waist during propagation before returning back to its original shape
again, due to periodic focusing by the graded index distribution. Note that there exists a specific
eigenmode (a Gaussian of a specific width, related to the refractive index gradient) for which the
beam propagates through the material without a change in shape as a result of a balance between
the diffraction of the beam and the guiding due to the parabolic gradient index profile. The con-
tour plot of such a beam is shown in Figure 2.4. Analytical expressions for the Gaussian beam

60

50

40

30

20

10

10 20 30 40 50 60

FIGURE 2.4 Fundamental mode in a graded index fiber.

18 Computational Methods for Electromagnetic and Optical Systems

during propagation through a graded index medium and for the eigenmode can be derived using
the q-parameter of the Gaussian beam and the ABCD laws of q-transformation, but will not be
pursued here for the sake of brevity.

2.5 BEAM PROPAGATION THROUGH DIFFRACTION

GRATINGS: ACOUSTOOPTIC DIFFRACTION
The beam propagation algorithm has been applied to the propagation of a beam through a grating,
and can be also used to analyze the case where the grating is a sound field. In what follows, we give
an example of the use of BPM to analyze diffraction of light by an acoustooptic cell in which a trav-
eling wave of sound causes a change in the refractive index using a modified split-step technique.
The modification consists of the fact that the inhomogeneity due to the refractive index grating is
accomodated for in the spatial frequency domain as well.
The perturbation Δn in the case of sound induced gratings is a function of time and space:

Δn( x, z, t ) = Cs( x, z, t ) (2.26)

where C is an interaction constant (for details, see Korpel [4]) and s(x, z, t) is the real sound ampli-
tude given by,

1
s( x, z, t ) = [ Se ( x, z ) exp( − jKx ) exp( jΩt ) + c.c.] (2.27)
2

where Se is the complex amplitude of the sound field that interacts with the light beam and is travel-
ing in the x direction, and c.c. denotes the complex conjugate.
The quantities K and Ω are the propagation constant and the angular frequency, respectively, of
the sound field. Following Refs. [4,5], a snapshot of the sound field is used at t = 0, so that using
2.26 and 2.27,

exp(Sˆ Δz) = exp( − jk0 ΔnΔz ) ≈ 1 − jk0 ΔnΔz

⎛ 1⎞
= 1 − ⎜ ⎟ jk0 ΔzC [ Se ( x, z ) exp( − jKx ) + Se * ( x, z ) exp(+ jKx )] (2.28)
⎝ 2⎠

In the modified split-step technique, we take the Fourier transform of the above operator operating
on the optical field Ee(x, z), taking care to note from the property of Fourier transforms that

ℑx [ f ( x ) exp(± jKx )] = f (k x ± K ) (2.29)

The main propagation loop of the algorithm is modified from Figure 2.1 and is shown in Figure
2.5. The boxes marked “Shift ±K” are used to facilitate the operation shown in 229 in the spatial
frequency domain.
Figure 2.6 shows problem geometry of a Gaussian beam incident nominally at Bragg angle on
a sound column of width z = L. The simulated evolution of the Gaussian beam is shown in Figure
2.7. The peak phase delay α of the light traveling through the acoustooptic cell is taken equal
to π, and the Klein–Cook parameter Q = K 2 L /k 0 = 13.1. We would like to point out that the same
answers could be derived by using the transfer function for acoustooptic interaction, as given in
Refs. [4,6].
Scalar EM Beam Propagation in Inhomogeneous Media 19

Initial profile
Ee (x, z = 0)

~
Ee (kx , z) = x {Ee (x, z)}

~ ~
Ee (kx , z + Δz) = Ee0 (kx , z) exp{ jk x2 Δz/2k0}

–1 ~
E΄e (x, z + Δz) = x {Ee (kx ; z + Δz)}

× (–1/2) jk0CΔzSe(x, z) × (–1/2) jk0CΔzS e*(x, z)

x x

Shift-K Shift-K

–1
x

FIGURE 2.5 Flow diagram for the modified split-step technique to analyze acoustooptic interaction.

Acoustic grating

Incident 0 order
Gaussian
beam

–1 order

Z=0 Z=L

FIGURE 2.6 Geometry of acoustooptic interaction with a Gaussian beam at nominal Bragg incidence.

2.6 BEAM PROPAGATION IN KERR-TYPE NONLINEAR MEDIA

2.6.1 Nonlinear Schrodinger Equation
The nonlinear propagation of beams through a cubically nonlinear material is modeled by the non-
linear PDE also called the nonlinear Schrodinger (NLS) equation [7]

∂ Ee 2
(2.30)
2 jk0 = ∇2⊥ Ee + 2n2 k02 Ee Ee
∂z

where n2 is the nonlinear refractive index coefficient defined by the functional dependence of the
total refractive index n on the intensity [7]:
20 Computational Methods for Electromagnetic and Optical Systems

Zeroth order

First order

z=0
In φ
te
rac
tio
n φB
len
gt
h –φB Z
z=L

FIGURE 2.7 Simulation plot of the intensity of the angular spectrum of the total field at different positions
along interaction length. (From Venzke, C. et al., Appl. Opt., 31, 656, 1992. With permission.)

n = n0 + δn Ee ( )= n +n
2
0 2 Ee
2
(2.31)

In writing (2.30), we have taken the linear refractive index n0 equal to unity for the sake of simplic-
ity. For a medium with n2 > 0, one can observe self-focusing of a Gaussian beam traveling through a
medium, while self-defocusing is observed for a medium with n2 < 0. The physical reasoning behind
self-focusing is as follows. The Gaussian beam induces a positive lens in the nonlinear material for
n2 > 0 due to the fact that where the beam intensity is high (e.g., on-axis), the induced refractive index
is higher as well, amounting to larger slowing down of the wavefronts. The wavefronts are therefore
bent similar to the action of a positive lens, resulting in initial focusing of the beam. This process
continues till the beam width is small enough for the diffraction effects to take over, leading to an
increase in the beam width. The converse is true for the case of n2 < 0. In this case, the beam spreads
more than in the linear diffraction limited case. A stable nonspreading solution in one transverse
dimension can be analytically found from the NLS equation for n2 > 0 and has the form
1/ 2
⎛ 8κ ⎞ x
Ee ( x; z) = ⎜ sech (2.32)
⎝ n2 k0 ⎟⎠ (2κk0 )1/ 2

where κ is a free parameter. The phase of Ee in the nonspreading solution is linearly proportional to
the propagation distance z.
As discussed above, self-focusing results in increase in the on-axis intensity and the narrowing
of the beam width. For powers above a certain critical power Pc [7,8], the beam may theoretically
collapse with an intensity so high that it can either cause breakdown in the material, triggering
some other physical effects, such as saturation of the index of refraction or failure of the assump-
tions about slowly varying amplitude and paraxial approximation. Zakharov and Shabat [9] have
pointed out that if the nonlinearity is strong enough, this results in higher index of refraction toward
the center of the beam and on-axis rays undergo total internal reflection and are thereby trapped.
As shown above, nonlinearity can balance diffraction of a beam in one dimension, resulting in the
formation of first order spatial solitons, see Equation 2.32. Also, if the nonlinear effect is higher
Scalar EM Beam Propagation in Inhomogeneous Media 21

than diffraction, periodic focusing occurs, or may result in higher order solitons. This may not be
the case in two or three dimensions where spatial collapse may occur.
In normalized form, and assuming n2 > 0, the NLS equation in Equation 2.30 can be rewritten as

∂ue
j + ∇2⊥ue + ue2 ue = 0 (2.33)
∂z

In a system with cylindrical symmetry, Equation 2.33 becomes

∂ue 1 ∂ue ∂ 2ue

j + + 2 + ue2 ue = 0 (2.34)
∂z r ∂r ∂r

The NLS equation as written in Equation 2.33 can also be modified to model pulse propagation
through a nonlinear fiber and in the presence of group velocity dispersion. This is possible due to the
fact that the interchange x → t in the NLS equation (2.33) with a suitable coefficient in front of the
second order derivative term, signifying (anomalous) material dispersion and subsequent renormaliza-
tion transforms the equation to one that can model the propagation of pulses in time τ along a fiber [2]:

∂ue 1 ∂ 2ue 2
j + 2
+ γ ʹ ue ue = 0 (2.35)
∂z 2 ∂τ

Analogous to Equation 2.32, the first order soliton solution of (2.35) can be expressed as

(
ue (τ; z) = K sech K γ ʹ τ ) (2.36)

where K is a free parameter. This profile is called a temporal soliton and can be regarded as a non-
(
linear eigenmode of the NLS system. The propagation of an initial profile ue (τ; z = 0) = sech γ ʹ τ )
with γ ′ = 1 using the BPM to model the NLS equation as in Equation 2.35 is shown in Figure 2.8.
The linear term in Equation 2.35 can be handled in the temporal frequency domain by using Fourier
transforms, similar to the case of propagational diffraction. Note that the nonlinear operator is

~ = sech(τ), Δz = 5e – 005, Level = 3, Time = 3

u e0

1
~ |2

0.5
|u e

0
3
2 10
1 0
z
0 –10 τ

FIGURE 2.8 Evolution of a first order 1-D soliton. (From Nehmetallah, G. and Banerjee, P.P., Nonlinear
Optics and Applications, H.A. Abdeldayem and D.O. Frazier, eds., Research Signpost, Trivandrum, India,
2007. With permission.)
22 Computational Methods for Electromagnetic and Optical Systems

ue0 = 2 * sech(τ), Maxtime = 2, Δz = 5e – 005, Level = 3, Time = 2

|ue| 0

–2
2
10
1
0
z τ
0 –10

FIGURE 2.9 Evolution of a second order 1-D soliton. (From Nehmetallah, G. and Banerjee, P.P., Nonlinear
Optics and Applications, H.A. Abdeldayem and D.O. Frazier, eds., Research Signpost, Trivandrum, India,
2007. With permission.)

exp{Sˆ Δz} = exp − j γ ʹ ue Δz . As expected, the “pulse” remains unchanged with propagation.
2

A similar result is obtained if we program the propagation of an initial beam profile and use
Equation 2.30 in one transverse dimension (x). The result is a spatial soliton.
The second order soliton input ũe(τ, 0) = 2 sech(τ) and its evolution in time is shown in Figure 2.9.
The split-step technique has also been applied to analyze propagation of profiles in two transverse
dimensions [11], and also to analyze propagation of optical fields that are pulsed in time and have a
spatial profile in the transverse dimension [12].

2.6.2 Simulation of Self-Focusing Using Adaptive Fourier

and Fourier–Hankel Transform Methods

The NLS equation for beams using the paraxial approximation has been previously derived in
Section 2.6.1 as
∂ue
j + ∇2⊥ue + ue2 ue = 0 (2.37)
∂z

During the last stages of self-focusing, the assumptions about slowly varying amplitude and the
paraxial approximation may not be valid for large focusing angles. It has been proposed that there is
no singularity if one accounts for nonparaxiality [13]. The paraxial and nonparaxial NLS equations
which are classically used to model the self-focusing phenomenon can be written in the general
operator form:

∂ue ∂ 2u 
− j ε 2e = jLr ue + jN nl (ue ) ue , r ∈ℜ D , z ≥ 0; ue (r, z = 0) = ue0 (r ) (2.38)
∂z ∂z

where D is the transverse dimension in space. The parameter ε = (λ/4πr0)2, where r0 is the initial beam
2σ
radius is often referred to as the nonparaxiality parameter. Also, Lr ue = ∇⊥2 ue, N nl (ue ) = ue . For most
2
cases, σ = 1 and hence N nl (ue ) = ue . The nonlinear operator Nnl(ũe) may be also conveniently modi-
fied to reflect any saturation in the change of refractive index:

δn ue( ) = n u
2
2 e
2 4
+ n4 ue , n4 < 0 (2.39)
Scalar EM Beam Propagation in Inhomogeneous Media 23

5
Mode 1
4 Mode 2
Mode 3
3

2
~a
1

–1

–2
0 2 4 6 8 10
~r

FIGURE 2.10 Townes soliton profiles for the first three “modes.”

Incidentally, for the paraxial case, two transverse dimensions and assuming radial symmetry, and
σ = 1, a special solution of (2.38), called the Townes soliton [7], can be found as the solution of the
normalized differential equation

∇2⊥ R(r ) − R + R 3 = 0, Rʹ(0) = 0, R(∞) = 0 (2.40)

Typical solutions for three different initial conditions, called modes, are plotted in Figure 2.10. These
modes, named ã(r̃ ), are characterized by the fact that the solutions tend to zero for r̃ → ∞. The solu-
tions of Equation 2.40 are very sensitive to initial conditions: for other initial conditions, the solutions
do not converge to zero as r → ∞.
The basic concept behind adaptive numerical algorithms is rather simple: imagine that a Gaussian
beam is spreading during linear propagation due to diffraction. If a numerical solution is being com-
puted, it is clear that after some distance of propagation, the size of the beam will become comparable to
the total transverse grid size, which will result in errors in the numerical computation, such as aliasing if
fast Fourier transform (FFT) techniques are used. The problem can be alleviated if the transverse profile
is re-sampled using a coarser grid size. The step size for propagation can also be increased, based on
the presumption that no sudden changes in the beam would occur at distances larger than the Rayleigh
range. The converse should be true for beams that are focusing: the transverse grid size should be made
finer and the propagational step size smaller.
In the case of the NLS equation, for the case D = 2, and assuming radial symmetry, McLaughlin
et al. [14] predicted similarity solutions of the form

1 ⎛ r ⎞
ue (r , z ) ∝ F (2.41)
( zr − z )1/ 2 ⎜⎝ ( zr − z )1/ 2 ⎟⎠

where F is an arbitrary function. We can use this knowledge to adaptively vary the longitudinal
stepping Δz and the transverse grid size for the beam.
A few more technical details must be mentioned before commencing the discussion on our adaptive
numerical techniques. Weinstein [15] has shown that if the initial beam power P0 is less than the lower
2
bound for the critical power for blowup Pclb = (λ 24πn 0 n2 ) N c, where N c =
∫R r dr = 1.86225, there
24 Computational Methods for Electromagnetic and Optical Systems

is no collapse of the paraxial NLS equation. Also, for an arbitrary initial profile ũe0 ≠ R, Fibich [13]
proved that there is an upper bound for the critical power Pcub = (λ 24πn 0 n2 ) G(ue 0 ) where

⎛ 2 u 2 dr ∇u 2 dr ⎞
G (ue 0) =
⎝ ∫
e0 e0
∫
⎠
⎛ u 4 dr ⎞
⎝ e0 ∫ ⎠

for which blowup will occur for this initial profile if it has sufficient high power. There are few
known integral invariants for the paraxial NLS equation above when σD ≥ 2. These invariants
are based on the symmetry of the NLS equation under gauge, space, and time transformations,
and may be derived from a Lagrangian density of Equation 2.37. Three of these invariants are the
Hamiltonian and the variance, defined as [14,16]

2
N=
∫ u e dr (2.42a)

⎛ 2 1 4⎞
⎝ ∫
H = ⎜ ∇ue − ue ⎟ dr
2 ⎠
(2.42b)

1 d2 2 2 D−2 4

8 dz 2 ∫r ue dr = H −
4 ∫ u
e dr (2.42c)

where the term |∇ũe|2 results from diffraction, and |ũe|4 from the nonlinear effect.
Note that from Equation 2.42c blowup occurs only if D ≥ 2. The “variance” is given as
2 2
∫
V ( z) = r ue dr = 4 Hz 2 + (dV (0) /dz )z + V (0) for D = 2 [17]. Thus, for H < 0, the function V(z)
vanishes at a distance zr = [V0 /−4H0]1/2 > 0. A sufficient condition for blowup is H < 0, i.e., when
the nonlinear effect is stronger than diffraction, the beam self-focuses and collapse occurs at a
distance zc ≤ zr.
− r 2 / 2 r02 1/ 2
For an input Gaussian of the form ue 0 = ce , zr = (1 ( pN c 2 − 1)) , zc = 0.317(p − 1)−0.6346 for
r0 = 1, and zc = 0.1585(p − 1) −0.6346 for r0 = 1/ 2 , where p = N0/Nc and N 0 = c 2r0D [18]. Note that the
ub
condition H = 0, which leads to Pc , leads to an overestimate of the actual critical power.
We now outline our numerical adaptive spectral technique called the adaptive split-step fast
Hankel transform (AFHTSS) used to track the solution of the NLS equation for σ = 1, D = 2 and vari-
able ε. Our scheme is based on the combination of the standard split-step fast Fourier transform
(SSFFT) and the Hankel transform, which exploits the cylindrical symmetry of the problem. This
enhances the computation time and precision appreciably. In addition, we also use the concepts
from the similarity solution developed by McLaughlin et al. [14] and apply them to our split-step
spectral method mentioned above, so that the grid transverse spatial range and the longitudinal
spatial step are adaptively updated. As its name indicates, we use the Hankel transform (see Chapter 1)
instead of the usual Fourier transform, relying on algorithms already developed in the literature [19].
In cylindrical coordinates, Equation 2.38 becomes

∂ue ∂ 2u ⎛ ∂2 1 ∂ ⎞ 2
− j ε 2e = j ⎜ 2 + ⎟ ue + j ue ue ; ue (r , z = 0) = ue0 (r ) = c exp − ar 2 (2.43)
∂z ∂z ⎝ ∂r r ∂r ⎠
Scalar EM Beam Propagation in Inhomogeneous Media 25

We use the definition of the lth order Hankel Transform pair [19,20]:
∞ ∞

∫
Ψ(ρ, z ) = 2π ψ (r , z )J l (2πrρ)rdr ,
0
∫
ψ (r , z ) = 2π Ψ(ρ, z )J l (2πrρ)ρdρ
0
(2.44a)

and the property:

⎛ ∂2 1 ∂ ⎞
j⎜ 2 +
⎝ ∂r r ∂r ⎟⎠
ψ (r , z ) → − j 4π ρ Ψ(ρ, z)
HT 2 2
(2.44b)

The AFHTSS algorithm in Figure 2.11 resembles the symmetrized Fourier split-step technique,
where we change the longitudinal spatial step Δz ∝ (1/c(z1)2 − 1/c(z2)2) ≈ 1/c(z1)2 when c(z2) ≫ c(z1),
adaptively using McLaughlin’s similarity formula (2.41) and the grid spatial range Δrmax ∝ 1/c(z) in
order to track the varying amplitude of the focusing beam when ε = 0.
There are several numerical approaches for implementing the Hankel transform [19,21]. The
importance of Siegman’s method [19] resides in the fact that, depending on the parameters, one can
employ a non-uniform sampling that is denser near the focusing region, which has advantages over
uniform sampling. Yu et al.’s method [21] is based on the expansion of the function and its transform
by a zero order Bessel series that can be written as

N N
ˆ (m) =
Ψ ∑n =1
Cmn ψˆ (n), ψˆ (n) = ∑C
m =1
nm
ˆ (m)
Ψ (2.45a)

2 ⎛ jn jm ⎞ −1
Cmn = J0 J1 ( jn ) J1−1 ( jm ) (2.45b)
S ⎜⎝ S ⎟⎠

ˆ (n) = ψ ⎛⎜ jn ⎞⎟ J1−1 ( jn ) R1, Ψ

ψ ˆ (n) = Ψ ⎛ jn ⎞ J −1 ( j ) R
⎜⎝ 2 πR ⎟⎠ 1 n 2 (2.45c)
⎝ 2πR2 ⎠ 1

The jn’s are the positive roots of the zero order Bessel function J 0, J1 is the first order Bessel
function, and R1, R2 are the spatial and transform ranges respectively, with S = 2R1R2. For rn ∈ r ≥ R1
and ρm ∈ ρ ≥ R2, we have ψ(rn ) = Ψ(ρm ) = 0, where rn = jn /2πR2 and ρm = jm / 2πR1.
We now show sample simulation results using the AFHTSS method that uses Yu et al.’s Hankel
transform technique [21], as well as a novel adaptive version of a split-step fast fourier transform
2
(AFFTSS) technique. For the test function ψ 0 = 4e − r / 2 , self-focusing and collapse is expected
at zc = 0.1487 for ε = 0. Note that zr = 0.288, based on the study above, which is obviously overes-
timated. Figure 2.12 shows the maximal focusing as a function of grid size h = Δr, which proves
the convergence of our method to the numerical focusing point when Δr decreases by varying the
S parameter defined above [22]. Although the trend is similar, this is a considerable improvement
over the convergence test results in Fibich and Ilan [23]. Figure 2.13 shows the growth of the on-
axis intensity using AFHTSS technique for the test function above [22]. Also, in this figure, we take
S = 2πR1R2 = 2π × 2000 (corresponding to approximately 4000 cylindrical samples), which permits
computation for the paraxial case till z = zc = 0.14817 corresponding to the expected zc = 0.1487 writ-
ten above. Figure 2.14 shows the corresponding AFFTSS technique for 1024 × 1024 grid where we
get z = zc = 0.1581 [22]. This implies that the critical distance zc is a little bit over estimated and the
maximum intensity reached is also less than the AFHTSS method. It is instructive to note that
with increase in sampling points, the AFFTSS approaches the results from AFHTSS, but on the
26 Computational Methods for Electromagnetic and Optical Systems

Initial profile
2 2
ψ0 = ce–r /2r0

K=K+1

Ψ(ρ, z) = FHT(ψ(r, z))

j 1/2 Δz
1– 1 – 16επ2ρ2
Δz Ψ(ρ, z)e – 2ε 2
Ψ ρ, z + =
2

Δz Δz
ψ r, z + = IFHT Ψ ρ, z +
2 2

j 2 1/2 Δz
1– 1 + 4ε|ψ|
Δz – 2ε
ψ r, z + 3 Δz = ψ r, z + e 2
2 2

3Δz 3Δz
Ψ ρ, z + = FHT Ψ r, z +
2 2

j 2 2 1/2 Δz
1– 1 – 16επ ρ
3Δz – 2ε 2
Ψ(ρ, z + 2Δz) = Ψ ρ, z + e
2

ψ(r, z + 2Δz) = IFHT (Ψ(ρ, z + 2Δz))

1 1
Δz and Δrmax
c(z)2 c(z)

Yes If K < M No End

FIGURE 2.11 The AFHTSS algorithm, a symmetrized version of the split-step FFT using Hankel transform
instead, and using adaptive longitudinal stepping and transverse grid management. FHT: fast Hankel trans-
form, IFHT: inverse fast Hankel transform.
Scalar EM Beam Propagation in Inhomogeneous Media 27

150
ψ = 4 exp(–r 2)
ψ = 4e(–r 2/2)

100

ψ2max/ψ2max0

50

0
0 0.005 0.01 0.015 0.02 0.025
Grid size: Δr = h

FIGURE 2.12 Maximal focusing as a function of grid size. (From Banerjee, P.P. et al., Opt. Commun., 249,
293, 2005. With permission.)

ψ 2max: Plot versus propagation dist z, AFHTSS

106
(1) Adap ε,ε0 = 5 * 10–3, λ = 0.89 ψ = 4e(–r2/2)
(2) Adap ε,ε0 = 10–3, λ = 0.4 (8) S = 2 * π * 2000
(3) Adap ε,ε0 = 10–4, λ = 0.1256 zc = 0.14817
105
(4) Adap ε,ε0 = 10–5, λ = 0.0375 M = 203
(5) Adap ε,ε0 = 10–8, λ = 0.001257
delz = 0.01
(6) ε = 10–3
ε = (λ/4 * π * r)2
104 (7) ε = 10–4
Beam intensity

(8) ε = 0 (5)

(7)
103
(4)
(6)
(3)
102 (2)
(1)

101
0 0.05 0.1 0.15 0.2 0.25
z
2
FIGURE 2.13 On-axis intensity of ψ 0 = 4e − r / 2 as a function of propagation for fixed values of ε rang-
ing from 10 −2 to 10 −8 where z = zc = 0.1481 for ε = 10 −8, and for an adaptive ε varying as ε = (λ/4πr)2, using
AFHTSS with S = 2πR1R2 = 2π × 2000 (4000 cylindrical samples). (From Banerjee, P.P. et al., Opt. Commun.,
249, 293, 2005. With permission.)

expense of calculation time. Also we note that changing the number of samples in both techniques
does not affect the value of the critical distance drastically, but the maximum intensity reached
at that point will be less or more depending on the number of samples, which is in agreement
with the convergence test mentioned above. We stress that in all computations using AFHTSS and
AFFTSS, energy is always conserved. Consequently, AFHTSS permits us to track peak intensi-
ties higher, faster, and more accurate than what is achievable by AFFTSS. Also, by using adaptive
28 Computational Methods for Electromagnetic and Optical Systems

ψ 2max: Plot versus propagation dist z

106
(1) Adap ε,ε0 = 5 * 10–3, λ = 0.89 ψ = 4 exp(–(x2+y2)/2)
(2) Adap ε,ε0 = 10–3, λ = 0.4
N = 10242
5 (3) Adap ε,ε0 = 10–4, λ = 0.1256
10 (8) zc = 0.15811
(4) Adap ε,ε0 = 10–5, λ = 0.0375
M = 161
(5) Adap ε,ε0 = 10–8, λ = 0.001257
(6) ε = 10–3 delz = 0.01
104 (7) ε = 10–4 ε = (λ/4 * π * r)2
Beam intensity

(8) ε = 0 (5)

(7)
103
(4)
(6)
(3)
102 (2)
(1)

101
0 0.05 0.1 0.15 0.2 0.25
z
2
FIGURE 2.14 On-axis intensity of ψ 0 = 4e − r / 2 as a function of propagation for fixed values of ε rang-
ing from 10 −2 to 10 −8 where z = zc = 0.1581 for ε = 10 −8, and for an adaptive ε varying as ε = (λ/4πr)2, using
AFFTSS with 10242 samples. (From Banerjee, P.P. et al., Opt. Commun., 249, 293, 2005. With permission.)

non-paraxiality parameter ε in the scalar nonparaxial equation, we obtain results similar to those of
the more complex vector method [24] and superior to those when ε is constant [13].
Finally, we compare computation speeds of the AFHTSS and AFFTSS. It can be shown that for
Siegman’s method, the number of computations is proportional to 4N log2 2N + 2N, compared with
2N 2 log2 N computations for the two dimensional FFT as in AFFTSS. Although at first glance, the
number of computations in Yu et al.’s method is proportional to N2, we can make the number of com-
putations comparable to Siegman’s method by a priori computing and storing the zeros of the Bessel
function. The advantages of Yu et al.’s method over Siegman’s are the accuracy for the sampled points
and a simple retrieval expression. For more comparison between Yu et al.’s and Siegman’s method, we
refer readers to table 1 in Yu et al. [21]. Also, we note that the use of the adaptive variation of the longi-
tudinal propagation stepping size Δz and the transverse spatial sampling size according to 1/c2(z) and
1/c(z) allow us to track on-axis amplitudes, for the paraxial case, up to two orders of magnitude more
than what is achievable without the adaptive algorithm, for both the AFHTSS and AFFTSS methods.
Without the adaptive variation, the numerical methods become unstable, and we witness oscillatory
focusing and defocusing of the beam from numerical instability. Typical run times on a Pentium IV
2.4 GHz processor with a 2 GB RAM are around 1 min for AFHTSS when S = 2πR1R2 = 2π × 2000,
10 min for AFFTSS when the mesh size is N 2 = (210)2. Note that the Hankel transform based method,
which exploits the cylindrical symmetry, is one dimensional, and therefore is expected to be faster
than other two dimensional FFT based numerical methods.

2.7 BEAM PROPAGATION AND COUPLING IN PHOTOREFRACTIVE MEDIA

2.7.1 Basic Photorefractive Physics
In this section, a model for beam propagation through a nonlinear PR material that takes into
account inhomogeneous induced refractive index changes due to the nonlinearity is first developed.
In some cases a focused Gaussian beam asymmetrically distorts due to passage through the
nonlinear material.
Other documents randomly have
different content
 census of, 92, 96 f.;
post-Servian, 93-7;
funds for, 93 f.;
opened to plebeians, 94;
equo privato, 94 f.;
equo publico, 95 f., 209;
in comitia centuriata, 209 f.;
prerogative, 211;
after reform, 212, 215, 220, 224, 226 f.;
given seats at theatre, 357, 428;
liable to law against bribery, 378;
made superior to senators, 381;
desert C. Gracchus, 384;
associate with senators in courts, 402, 427 f., 455.
Esquilina (tribus), 50, 220.
Eupyridae, Attic gens, 28.
Exercitus urbanus, 203.
Exile, voluntary, legalized by comitia tributa, 249, 256, 257, n. 5,
267, 446.
Extortion, see Repetundae.

Fabius, Q., trial of (389), 246, 288.

Fabius Buteo, censor (241), 213.
Fabius Gurges, Q., consul (292), resolution on imperium of, 289,
306.
Fabius Maximus, Q., consul (215), and curiate law, 197.
Fabius Maximus Servilianus, Q., trial of, for murder, 257, n. 5.
Fabius Pictor, sources of for early Rome, 26;
on Servian tribes, 51, 52-4;
centuriate system, 67, 85.
Fabius Pictor, Q., praetor (189), trial of, 327 f.
Fabius Rullianus, Q., alters tribes, 64.
Fabri (mechanics, sappers, workmen), 66, 67, n. 3, 68, 81;
assigned to classes, 205 f.;
after reform, 226.
Family law, changes in, 339 f., 352.
Fasti, read in comitia calata, 154 f.;
dies, 471 f.;
Clodian law on, 445.
Faucia (curia), 11, n. 7;
ill-omened, 112.
Ferentarii, 80, n. 5.
Festivals, regulated by law, 340 f.
Fetialis, 176, 265.
Finance, legislation on, 297 f., 310 f., 335-7, 351 f., 392, 403,
422, 438.
Fines, appealed to tribes, 259, 269, 286 f., 292, 317 ff., 344.
Flamen, curial, 10;
Dialis, 203, n. 7.
Flaminian, era, 333-46;
Circus, 465;
Meadow, 465.
Flaminius, C., and curiate law, 191;
 monetary law of, 191 f., 336;
censor, 213;
era of, 333-46;
agrarian law, 334 f.;
supports Claudian law, 335;
influences legislation, 337 f., 343;
assigns libertini to city tribes, 355;
energizes comitia, 343, 475.
Flavius, M., trial of, 291.
Fordicidia, 9.
Foreign affairs, administered by senate, 273;
then fell partly to comitia tributa, 303;
laws on, 349 f.
Forgery, 420.
Foriensis (curia), 11.
Formiani, Fundani, etc., receive suffrage, 352.
Formulae, legal, 464.
Fornacalia, 9, 11, n. 8.
Forum, assembly in, 267, 327, 431, 439, n. 15, 465.
Fowler, W. W., on lex Scantinia, 357, n. 13;
reëlection of tribune, 369, n. 4;
Sempronian lex iudiciaria, 374, n. 7.
Freedmen, see Libertini.
Fregellae, revolt of, 255.
Fröhlich, on Sulpicius, 405, n. 2;
Cornelian-Pompeian law, 406, n. 6;
 lex Cornelia de tribunicia potestate, 414.
Frumentations, 372 f., 395, 398, 401;
abolished by Sulla, 422;
restored by Lepidus, 423, n. 8;
further legislation on, 424, n. 5, 444 f.;
curtailed by Caesar, 453;
under lex municipalis, 456.
Fulvius, Cn., praetor (212), trial of, 249 f.
Fundus populus factus, 401, n. 8.
Furius, L., past consul, trial of, 268.
Furius, P., tribune (98), 257, n. 5, 323.
Furius Camillus, M., and equestrian fund, 94;
dictator, 202;
trial of, 244 f., 288, 290.
Furtum (theft), 339, n. 5;
prosecution for, 321;
under lex Hostilia, 337, n. 5;
Plautia, 424;
see Peculatus.

Gabinius, A., tribune (67), 429 f., 432 f.

Gabinius, Q., tribune (139), 359.
Gabinus ager, 108.
Gades, receives citizenship, 454.
Galeria iuniorum, 217.
Genera, identified with gentes, 12.
Gens, meaning family, lineage, 30 f., 102.
Gentes, 11-13;
unconnected with curiate system, 13;
social composition of, 28-31;
defined by Scaevola, 28, n. 7;
maiores et minores, 35 f.;
origin of patrician, 37, n. 4;
relatively late, 48, n. 2;
common land of, 49;
relation to rural tribes, 35, 50, 55, n. 1;
in war, 78, n. 6.
Gentiles, Gentilitas, 28, n. 7, 29, 30.
Gifts, leges Publicia and Cincia on, 338 f.
Governors, provincial, of the Spains, 346 f.;
under Porcian laws, 349;
Sempronian, 374, 381 f.;
Acilian, 376 f.;
Julian, 442, 456.
Gracchi, see Sempronius.
Grain, see Frumentations.
Greenidge, on social classes, 38, n. 2.
Guilds, see Collegia.

Hackel, on lex Iulia municipalis, 457, n. 5.

Heredium, 49.
Herennius, tribune (60), 162, 438.
Hernicans, receive civitas sine suffragio, 305.
Herzog, on curiate law, 183, n. 5;
Sulpicius, 405, n. 2.
Ἑταιρεία, 8, n. 6.
Holidays, non-comitial, 116.
Horatius, trial of for perduellio, 121.
Hornblowers, in centuriate system, 66.
Horsemen, see Equites.
Hortensius, Q., dictator (287), 313.
Hosticus ager, 108.
Hostilius Tubulus, L., trial of (141), 255, n. 1.
Huschke, on Servian tribes, 51;
ratings, 86;
reformed comitia centuriata, 219.

Ihne, on trial of Opimius, 256 f.;

popular interference with censors, 351, n. 5;
policy of Marius, 389;
Sulpicius, 405, n. 2.
Imperium, true (iustum), 102 f., 187, n. 7;
confirmed by curiate law, 188;
granted by comitia, 188, n. 2;
by senate, 191, 284;
transition of without curiate law, 196 f.;
promagisterial, 305;
abrogated, 324, n. 1, 342, 360, 367, 390, 404, 409;
 limited by Porcian laws, 349;
regulated by Sulla, 417.
Impetrativa, impetrita (auspicia), 100, 103-11;
relation of to oblativa, 112.
Inaugurare sacerdotes, 106.
Inaugurations, in comitia calata, 155 f.
Incertus ager, 108.
Incest, prosecution for, 326.
Index legis, 462.
Ingenuus, 20 f., 36;
son of libertinus becomes, 355.
Instauraticius dies, creation of by law, 308 f.
Intercession, see Veto.
Interdict, decreed by tribes, 249, 256, 257, n. 5, 267, 446.
Interregnum, 183.
Interrex, appointment of, 102;
auspices, 103;
presidency of contio, 140;
right of public speech, 145, n. 4;
nominates king, 183;
lacks curiate sanction, 191;
presides over curiae and centuries, 236, 412, 468, 469.
Italians, benefit by Sempronian agrarian law, 364;
revolt of, 397, 401;
receive citizenship, 401 f.;
dissatisfied, 403;
 equalized with Romans, 409.
Iubere, in legislation, 179.
Iudices (jurors), originally from senate, 345, 358, 374;
from knights under
leges Sempronia and Acilia, 374 f.;
qualifications of under lex Cornelia, 419;
Aurelia, 427;
Licinia and Pompeia, 448;
Antonia, 458;
punished for bribery, 442.
Ius agendi cum populo, 465.
Ius gentium, violation of, 246.
Ius pontificum, 181.
Ius sententiae dicendae, 391.
Iussus populi, 180, n. 7.
Iustitium, 401;
defined, 404, n. 6.
Iustum auspicium, imperium, 102 f.

Janiculum, garrison and flag on, 203, n. 2, 258, 469;

secession to, 313.
Judicial process, in contio, 142, 143;
in comitia, 259 f.;
choice as to assembly, 287;
ballot in, 359.
Jugurtha, 390.
Julius Caesar, C., usage as to comitia and concilium, 125;
creates patricians, 164, 456;
uses centuriate and tribal assemblies, 236;
threatened with prosecution, 324;
supports Licinius Macer, 426;
Manilian rogation, 434;
consul (59), 438-44;
affected by Pompeian laws, 449;
dictator (49-44), 451-7;
adds 10 days to year, 471.
Julius Caesar, C., consul (64), usage as to contio, 125 f.
Julius Caesar Octavianus, creates patricians, 164, 460;
triumvir, 459 f.
Juniors, in centuriate system, 66, 68, 81 ff.;
number of, 84, 205;
after reform, 216.
Junius, L., consul (249), trial of, 248.
Junius Silanus, M., prosecution of (103), 323.
Juno, Curis, 8, n. 5, 9;
Moneta, 2, n. 6.
Junonia, colonization of, 383, 385.
Jupiter, auspices of, 100, 103;
victim to, 264, 274;
feast of, 347;
oath by, 380.
Jurisdiction, of king, 182;
 comitia centuriata, 239-61, 315;
tributa, 264-9, 280, 286-92, 317-29.
Jurors, see Iudices.
Juventus Thalna, M., tribal lex de bello indicendo of, 231.

Kalumniator, 400.
Kaput legis, 463, n. 6.
Κήρυκες, 153, n. 3.
King, auspices of, 103;
presidency of contio, 140;
of comitia calata, 154;
curiata, 173 ff.;
right to address people, 145, 173;
as legislator, 177 f.;
irresponsible, 180;
powers of, 181;
jurisdiction, 182;
election, 182-4, 189 f.;
declares war, 175 f., 181, 230.
Klebs, on reformed comitia centuriata, 223, 225.
Knights, see Equites.
Kornemann, on lex Scantinia, 357, n. 13.

Laelius Felix, defines comitia and concilium, 119;

not in accord with Livy, 119-25;
view of rejected, 131;
 error explained, 137.
Laelius Sapiens, C., prosecution of, 322;
agrarian rogation of, 360 f.
Laetorius Mergus, L. or M., trial of, 247.
Land, see Ager.
Lange, on obnuntiatio, 117;
early legislation, 181, n. 9;
transitio imperii, 183, n. 5, 197, n. 4;
comitia centuriata, 201, n. 4;
reform of, 224 f.;
validity of plebiscite, 278, n. 2;
right of dedication, 309;
lex Appuleia de maiestate, 394, n. 5;
lex Antia, 428;
principium, 466, n. 3.
Lanuvium, curiae in, 8, n. 5.
Latins, rights of, 63;
benefit by Sempronian agrarian law, 364;
proposal to grant citizenship to, 383;
receive citizenship, 401 f.;
limited suffrage, 466, n. 2.
Lator legis, 462, n. 2.
Laurentum, 2, n. 6, 3, n. 1.
Lauretum, 2, n. 6.
Lavinium, Tities in, 4, n. 3.
Law, divine, 177;
 human, 178;
sovereignty of, 308;
see Legislation, Lex.
Legion, instituted, 68, 84;
early republican, 75 ff.
Leges, composition and preservation of, 462-5;
imperfectae, etc., 463;
centuriate, consular, etc., n. 8;
provisions to secure validity of, 464;
annulment by senate, 107.
Legislation, regal, 177-82, 230;
centuriate, 230-9;
tribal, pre-decemviral, 269-74;
pre-Hortensian, 292-316;
from Hortensius to Gracchi, 330-362;
from Gracchi to Sulla, 303-411;
late republican, 412-61;
freed from obnuntiatio, 117, 445;
process of, 178 f., 465-70;
provided for by Twelve Tables, 233 f., 307, 368, 464, 474;
senatorial, 273;
transferred to tribes, 316;
to centuries, 406-8;
ballot in, 369;
fields of: administrative, 238, 306 f.;
 agrarian, 238, 265, n. 4, 272, 334, 363-7, 373 f., 385-7,
392, 395, 400, 403, 435 f., 438-41, 458;
colonial, 311, 350, 382 f., 393 ff., 457 f.;
financial, 310 f., 335-7, 351 f., 392, 403, 422, 438;
frumentarian, 372 f., 395, 401, 423, n. 8, 444;
judiciary, 358, 374-6, 402 f., 419, 424, 427 f., 442, 448, 455
f., 458 f.;
religious, 238, 295, n. 6, 308 f., 340, 358 f., 391 f., 435;
sumptuary, 337 f., 356, 388, n. 9, 423, 428, 448, 455 f.
Legum dictio, 110, 179, n. 7.
Lengle, on lex Cornelia Pompeia (88), 407, n. 2.
Lentus, L., consul (156), trial of, 255, n. 1.
Lex, meaning of word, 179;
data and rogata, 180.
Lex alearia, (before 204), 337.
⸺ auspical, 110.
⸺ centuriata de potestate, 185.
⸺ Coloniae Genetivae, 453, n. 4.
⸺ curiata de imperio, 31, 32, 112, 180, n. 7;
formula of, 183, 188;
sanctioning, 184;
Messala on, 185 f.;
dispensations from, 186, 190, 195, 199;
subject to veto, 187;
confirms imperium, 188;
functions performed without, 191;
 lack of in 49 b.c., 192, 194 f.;
one annually, 195;
becomes formality, 196 f.;
revived by optimates, 198;
strengthened by Sulla, 199;
de potestate, 190.
⸺ lenonia, 338, n. 5.
Leges regiae, 181.
Lex sacrata, so-called Icilian, 233, 272 f.;
on tribunes, 264;
meaning of, 264 f.;
mitigation of, 266;
renewed by Valerius and Horatius, 274;
list of leges s., 265, n. 1;
on centuriate trials, 268, n. 6.
⸺ satura, 396, 399.
⸺ de bello indicendo, 231.
⸺ de imperio, for triumphs, 334 f.;
Vespasiani, 464, n. 5;
see Lex curiata.
⸺ on driving nail, 238.
⸺ found at Ateste, 454, n. 3.
⸺ granting citizenship to priestesses of Ceres, 353.
⸺ creating dictatorship (501), 233.
⸺ instituting tribuni militum consulari potestate (445), 234,
294.
⸺ creating censors (443?), 234.
⸺ appointing prefect of market (440), 295, 305, n. 5.
⸺ on presenting crown to Jupiter (437), 295, n. 6.
⸺ on garments of candidates (432), 295.
⸺ increasing quaestors (421), 234.
⸺ creating special murder court (414), 253, 295.
⸺ as to residence on Capitoline hill (384), 295.
⸺ creating praetorship (367), 234.
⸺ creating curule aedileship (367), 234.
⸺ for election of 6 military tribunes (362), 234.
⸺ prohibiting comitia away from city (357), 297.
⸺ preparing for war (356), 297, n. 5.
⸺ granting triumph (356), 297, n. 5.
⸺ on interest and debts (347), 298.
⸺ granting citizenship to Privernates (329), 304 f.
⸺ creating promagistracy (t. 327), 305.
⸺ sending prefects to Capua (318), 306.
⸺ on dedication of temples, etc. (304), 309.
⸺ dispensing Q. Fabius from law (t. 298), 308.
⸺ creating triumviri coloniis deducendis (296), 311.
⸺ prolonging imperium (t. 295), 305.
⸺ granting Etruria to Fabius (295), 305 f.
⸺ on imperium of consul Q. Fabius (292), 289, 306.
⸺ creating special court (270), 254.
⸺ doubling number of quaestors (267), 332.
⸺ forbidding reëlection to censorship (265), 332.
⸺ instituting second praetor (242), 332.
⸺ granting privilege of riding (241), 332.
⸺ instituting 2 praetors (227), 341 f.
⸺ granting triumph (t. 223), 334.
⸺ on intermarriage of kin (241-219), 339 f.
⸺ on Sacred Spring (t. 217), 340.
⸺ dispensing consulars from law (t. 217), 343.
⸺ granting citizenship to Campanian knights (215), 340.
⸺ for election of pontifex maximus (before 212), 341;
for election of chief curio (before 209), 341.
⸺ creating 3 administrative boards (t. 212), 337.
⸺ on Campanian vectigalia (210), 337.
⸺ granting citizenship (t. 210), 353, n. 7.
⸺ for election of 24 military tribunes (207), 342.
⸺ dispensing C. Servilius from law (t. 203), 343, n. 2.
⸺ permitting oath by proxy (t. 200), 343, n. 2.
⸺ on qualification of plebeian tribunes and aediles (Flaminian
era), 342 f.
⸺ increasing praetors to 6 (198), 346.
⸺ on triumphs (after 180), 350.
⸺ forbidding reëlection of consul (151), 348;
dispensation from, 360;
repealed by Sulla, 415.
Leges, dispensing Scipio Aemilianus from laws (t. 148, 135),
360.
Lex, assigning seats to equites at theatre (t. 146?), 357.
⸺ abrogating proconsular imperium (136), 360.
⸺ granting Asia as province (t. 131), 381, n. 5.
⸺ on qualifications of senators (t. about 129), 369 f.
⸺ permitting reëlection of tribune (t. before 123), 369, 371.
⸺ agraria, amending Sempronian law (t. not after 118), 385.
⸺ founding Narbo Martius, 386, n. 1.
Leges, repealing Sempronian law on military service (about
115), 388 f.
Lex, on dedication of Capitoline temple (78), 341, n. 1.
⸺ on vectigalia (75), 424.
⸺ appointing decemviri for regulating Asia (t. 67), 433.
⸺ dispensing Caesar from law (t. 52), 449.
⸺ granting citizenship to Gades (49), 454.
Leges, recalling certain exiles (p. and t. 49), 454.
Lex, granting Caesar triumph over Juba (48), 335, n. 2.
Leges, conferring powers on Caesar (48-45), 451 f.
Lex, for founding Colonia Genetiva (t. 44), 453, n. 4.
Lex (?), for building temple to Isis (43), 459.
Lex, honoring triumviri (43), 459.
⸺ on birthday of Caesar (42), 457, n. 7.
⸺ granting lictors to Vestals (42), 459, n. 5.
Leges, honoring Octavia, Octavianus, and Livia (t. 35), 459 f.
Leges whose authors are given:
Lex Acilia de intercalatione (c. 191), 358.
⸺ Acilia repetundarum (t. 122), 375-8.
⸺ Acilia Calpurnia de ambitu (c. 67), 431;
amended by Cicero, 436.
⸺ Acilia Minucia, on peace with Carthage (t. 201), 344, n. 7.
⸺ Acilia Rubria, on worship of Jupiter (t. 122), 384, n. 4.
⸺ Aebutia, on legis actio, 339, n. 5.
Leges Aebutia et Licinia, on qualifications of candidates (t. after
194), 347 f.
Lex Aelia, colonial (t. 194), 350.
Leges Aelia et Fufia (t. about 150), 116 f., 358 f.;
amended by lex Clodia, 116 f., 445;
and curiate law, 198;
relation to tribunician comitia, 280.
Lex Aemilia, on censorship (d. 443), 237.
⸺ Aemilia de libertinorum suffragiis (c. 115), 388;
Sumptuaria, n. 9.
⸺ Aemilia frumentaria (c. 78), 423, n. 1, 444, n. 6.
⸺ Aemilia, for naming Caesar dictator (p. 49), 450.
⸺ Antia sumptuaria (t. 70?), 428.
⸺ Antistia, on punishment of Satricans (t. 319), 310;
serves as precedent, 340.
⸺ Antonia de Termessibus (t. 71), 425.
⸺ Antonia, on children of proscribed (t. 49), 453 f.;
colonial (c. 44), 237, 453, n. 4, 457 f.;
iudiciaria, 458;
 establishing appeal from quaestiones, 458 f.;
abolishing dictatorship, 237, 459;
a l. sacrata, 265, n. 1;
leges honoring Caesar, 452 n. 4;
lex confirming acts of Caesar, 457.
⸺ Antonia, on elections (t. 45), 454 f.;
agraria (t. 44), 458.
⸺ Antonia Tullia de ambitu (c. 63), 436 f.
⸺ Appuleia agraria (t. 100), 395;
colonial (t. 103, 100), 393 ff.;
frumentaria (t. 100), 395, 444, n. 6;
de maiestate, 394, 400;
de sponsu (103, 100?), 298, n. 1, 394, n. 5;
interdicting Metellus (t. 100), 257, n. 5, 395 f.
⸺ Aquilia de damno (t. 287?), 332 f.
⸺ Aternia Tarpeia de multae dictione (c. 454), 233, 269.
⸺ Atia, on election of sacerdotes (t. 63), 416, n. 6, 435.
⸺ Atia Ampia, honoring Pompey (t. 63), 435, n. 2.
⸺ Atilia, appointing special court (t. 210), 254, 340.
⸺ Atilia, on appointing tutors (242-186), 340.
⸺ Atilia Furia, for surrendering Mancinus (c. 136), 350.
⸺ Atilia Marcia, for electing 16 military tribunes (t. 311), 306.
⸺ Atinia, on stolen property (214?), 339, n. 5.
⸺ Atinia, for founding colonies (t. 197), 350.
⸺ Atinia, on right of tribunes to senatorship (t. 122-102), 391.
⸺ Atinia Marcia, on treaty with Macedon (t. 196), 349.
⸺ Aufidia, on importing wild beasts (t. 170), 356.
⸺ Aurelia, amending Cornelian law on tribunate (c. 75), 423 f.;
de iudiciis privatis, 424.
⸺ Aurelia iudiciaria (p. 70), 427, 448.
⸺ Baebia, colonial (t. 194), 350.
⸺ Baebia, on praetors (c. 181), 346.
⸺ Bantina, Latin, 370, n. 3, 379 f.
⸺ Boria (?) agraria (t. 118), 385.
⸺ Caecilia, appointing special court (154), 255.
⸺ Caecilia, abolishing vectigalia (p. 60), 438;
repealed by Caesar, 457, n. 6.
⸺ Caecilia, repealing lex Clodia on censorial stigma (c. 52),
450, n. 2.
⸺ Caecilia Cornelia, recalling Cicero (c. 57), 114, n. 7, 143,
446;
on cura annonae, 446.
⸺ Caecilia Didia, on rogations (c. 98), 396 f.;
amended, 438, n. 2.
⸺ Caelia tabellaria (t. 107), 253, 390.
⸺ Calidia, recalling Metellus (t. 98), 396, n. 1.
⸺ Calpurnia, for recovery of property, 339, n. 5.
⸺ Calpurnia repetundarum (t. 149), 358.
⸺ Calpurnia, recalling Popillius (t. 120), 388.
⸺ Calpurnia, granting citizenship (t. 89), 57, n. 5, 58, 402.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookultra.com