Get the full ebook with Bonus Features for a Better Reading Experience on ebookgate.

com

The Finite Difference Time Domain Method For

Electromagnetics with MATLAB Simulations 2nd
Edition Atef Z. Elsherbeni

https://ebookgate.com/product/the-finite-difference-time-
domain-method-for-electromagnetics-with-matlab-
simulations-2nd-edition-atef-z-elsherbeni/

OR CLICK HERE

DOWLOAD NOW

Download more ebook instantly today at https://ebookgate.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

A pragmatic introduction to the finite element method for

thermal and stress analysis with the matlab toolkit SOFEA
Petr Krysl
https://ebookgate.com/product/a-pragmatic-introduction-to-the-finite-
element-method-for-thermal-and-stress-analysis-with-the-matlab-
toolkit-sofea-petr-krysl/
ebookgate.com

The Finite Element Method An Introduction with Partial

Differential Equations 2nd Edition A. J. Davies

https://ebookgate.com/product/the-finite-element-method-an-
introduction-with-partial-differential-equations-2nd-edition-a-j-
davies/
ebookgate.com

The Wiener Hopf Method in Electromagnetics Vito G. Daniele

https://ebookgate.com/product/the-wiener-hopf-method-in-
electromagnetics-vito-g-daniele/

ebookgate.com

Numerical Analysis in Electromagnetics The TLM Method 1st

Edition Pierre Saguet

https://ebookgate.com/product/numerical-analysis-in-electromagnetics-
the-tlm-method-1st-edition-pierre-saguet/

ebookgate.com
The Finite Element Method for Three Dimensional
Thermomechanical Applications 1st Edition Guido Dhondt

https://ebookgate.com/product/the-finite-element-method-for-three-
dimensional-thermomechanical-applications-1st-edition-guido-dhondt/

ebookgate.com

Optical Wireless Communications System and Channel

Modelling with MATLAB 1st Edition Z. Ghassemlooy

https://ebookgate.com/product/optical-wireless-communications-system-
and-channel-modelling-with-matlab-1st-edition-z-ghassemlooy/

ebookgate.com

Time Dependent Problems and Difference Methods 2nd Edition

Bertil Gustafsson

https://ebookgate.com/product/time-dependent-problems-and-difference-
methods-2nd-edition-bertil-gustafsson/

ebookgate.com

The Scaled Boundary Finite Element Method 1st Edition John

P. Wolf

https://ebookgate.com/product/the-scaled-boundary-finite-element-
method-1st-edition-john-p-wolf/

ebookgate.com

Electric Power Systems with Renewables Simulations Using

PSSE 2nd Edition Mohan

https://ebookgate.com/product/electric-power-systems-with-renewables-
simulations-using-psse-2nd-edition-mohan/

ebookgate.com
The Finite-Difference Time-Domain
Method for Electromagnetics
with MATLAB‡ Simulations
 The ACES Series on Computational Electromagnetics and
Engineering (CEME)
Andrew F. Peterson, PhD – Series Editor

The volumes in this series encompass the development and application of numerical tech-
niques to electrical systems, including the modeling of electromagnetic phenomena over all
frequency ranges and closely-related techniques for acoustic and optical analysis. The scope
includes the use of computation for engineering design and optimization, as well as the
application of commercial modeling tools to practical problems. The series will include titles
for undergraduate and graduate education, research monographs for reference, and practi-
tioner guides and handbooks.

Series Editorial Board

Andrew F. Peterson – Series Editor, Georgia Institute of Technology
Atef Z. Elsherbeni – Associate Series Editor, Colorado School of Mines
James C. Rautio, Associate Series Editor, Sonnet Sofware, Inc.
Branislav M. Notaros, Associate Series Editor, Colorado State University

Titles in the Series

Elsherbeni and Demir – The Finite-Difference Time-Domain Method for Electromagnetics
with MATLAB Simulations, 2nd Edition (2015)
Elsherbeni, Nayeri, and Reddy – Antenna Analysis and Design Using FEKO Electro-
magnetic Simulation Software (2014)
Yu, Yang, and Li – VALU, AVX and GPU Acceleration Techniques for Parallel FDTD
Methods (2013)
Warnick – Numerical Methods for Engineering: An Introduction Using MATLAB and
Computational Electromagnetics (2011)
The Finite-Difference Time-Domain
Method for Electromagnetics
with MATLAB‡ Simulations
ACES Series
2nd Edition

Atef Z. Elsherbeni
Colorado School of Mines

Veysel Demir
Northern Illinois University

scitechpub.com
Published by SciTech Publishing, an imprint of the IET
www.scitechpub.com
www.theiet.org

The Institution of Engineering and Technology is registered as a Charity in England & Wales (no. 211014) and
Scotland (no. SC038698).

Copyright  2009, 2016 SciTech Publishing, Edison, NJ. All rights reserved.

First published 2009 (978-1-89112-171-5)

Reprinted 2013
Second edition 2015

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any
means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as permitted under Sections
107 or 108 of the United States Copyright Act of 1976, without either the prior written permission of the Publisher,
or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, 222 Rosewood
Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 646-8600, or on the web at copyright.com. Requests to the
Publisher for permission should be addressed to The Institution of Engineering and Technology, Michael Faraday
House, Six Hills Way, Stevenage, Herts, SG1 2AY, United Kingdom.

While the authors and publisher believe that the information and guidance given in this work are correct, all parties
must rely upon their own skill and judgement when making use of them. Neither the authors nor publisher assumes
any liability to anyone for any loss or damage caused by any error or omission in the work, whether such an error or
omission is the result of negligence or any other cause. Any and all such liability is disclaimed.

ISBN 978-1-61353-175-4 (hardback)

Typeset in India by MPS Limited

Printed in the UK by CPI Antony Rowe Ltd
To my wife, Magda, my daughters, Dalia and Donia, and my son, Tamer.
To the memory of my parents.
Atef Z. Elsherbeni
To my parents, Abdurrahman and Aysan, my wife, Minmei, and
my daughter, Laureen, and my son, Ronnie.
Veysel Demir
This page intentionally left blank
Contents

List of figures xv
List of tables xxiv
Preface xxv
Acknowledgements xxviii

1 Introduction to FDTD 1
1.1 The finite-difference time-domain method basic equations 2
1.2 Approximation of derivatives by finite differences 4
1.3 FDTD updating equations for three-dimensional problems 13
1.4 FDTD updating equations for two-dimensional problems 23
1.5 FDTD updating equations for one-dimensional problems 27
1.6 Exercises 32
2 Numerical stability and dispersion 33
2.1 Numerical stability 33
2.1.1 Stability in time-domain algorithm 33
2.1.2 CFL condition for the FDTD method 35
2.2 Numerical dispersion 37
2.3 Exercises 41
3 Building objects in the Yee grid 43
3.1 Definition of objects 43
3.1.1 Defining the problem space parameters 45
3.1.2 Defining the objects in the problem space 48
3.2 Material approximations 50
3.3 Subcell averaging schemes for tangential and normal components 52

vii
viii Contents

3.4 Defining objects snapped to the Yee grid 55

3.4.1 Defining zero-thickness PEC objects 57
3.5 Creation of the material grid 58
3.6 Improved eight-subcell averaging 66
3.7 Exercises 66
4 Active and passive lumped elements 71
4.1 FDTD updating equations for lumped elements 71
4.1.1 Voltage source 72
4.1.2 Hard voltage source 74
4.1.3 Current source 75
4.1.4 Resistor 76
4.1.5 Capacitor 77
4.1.6 Inductor 78
4.1.7 Lumped elements distributed over a surface or within a volume 79
4.1.8 Diode 81
4.1.9 Summary 85
4.2 Definition, initialization, and simulation of lumped elements 86
4.2.1 Definition of lumped elements 86
4.2.2 Initialization of FDTD parameters and arrays 89
4.2.3 Initialization of lumped element components 90
4.2.4 Initialization of updating coefficients 97
4.2.5 Sampling electric and magnetic fields, voltages, and currents 108
4.2.6 Definition and initialization of output parameters 111
4.2.7 Running an FDTD simulation: The time-marching loop 119
4.2.8 Displaying FDTD simulation results 129
4.3 Simulation examples 132
4.3.1 A resistor excited by a sinusoidal voltage source 132
4.3.2 A diode excited by a sinusoidal voltage source 135
4.3.3 A capacitor excited by a unit-step voltage source 137
4.4 Exercises 141
5 Source waveforms and time to frequency
domain transformation 143
5.1 Common source waveforms for FDTD simulations 143
5.1.1 Sinusoidal waveform 144
5.1.2 Gaussian waveform 145
5.1.3 Normalized derivative of a Gaussian waveform 148
5.1.4 Cosine-modulated Gaussian waveform 151
 Contents ix

5.2 Definition and initialization of source waveforms for FDTD

simulations 151
5.3 Transformation from time domain to frequency domain 155
5.4 Simulation examples 158
5.4.1 Recovering a time waveform from its Fourier transform 160
5.4.2 An RLC circuit excited by a cosine-modulated
Gaussian waveform 162
5.5 Exercises 167

6 S-Parameters 169
6.1 Scattering parameters 169
6.2 S-Parameter calculations 170
6.3 Simulation examples 179
6.3.1 Quarter-wave transformer 179
6.4 Exercises 184

7 Perfectly matched layer absorbing boundary 185

7.1 Theory of PML 185
7.1.1 Theory of PML at the vacuum–PML interface 185
7.1.2 Theory of PML at the PML–PML interface 188
7.2 PML equations for three-dimensional problem space 191
7.3 PML loss functions 192

7.4 FDTD updating equations for PML and MATLAB
implementation 194
7.4.1 PML updating equations – two-dimensional TEz case 194
7.4.2 PML updating equations – two-dimensional TMz case 197
7.4.3 MATLAB implementation of the two-dimensional FDTD
method with PML 199
7.5 Simulation examples 215
7.5.1 Validation of PML performance 215
7.5.2 Electric field distribution 220
7.5.3 Electric field distribution using DFT 225
7.6 Exercises 227

8 Advanced PML formulations 229

8.1 Formulation of CPML 229
8.1.1 PML in stretched coordinates 229
8.1.2 Complex stretching variables in CFS-PML 230
x Contents

8.1.3 The matching conditions at the PML–PML interface 231

8.1.4 Equations in the time domain 231
8.1.5 Discrete convolution 231
8.1.6 The recursive convolution method 232
8.2 The CPML algorithm 234
8.2.1 Updating equations for CPML 235
8.2.2 Addition of auxiliary CPML terms at respective regions 236
8.3 CPML parameter distribution 237

8.4 MATLAB implementation of CPML in the three-dimensional
FDTD method 238
8.4.1 Definition of CPML 239
8.4.2 Initialization of CPML 240
8.4.3 Application of CPML in the FDTD time-marching loop 246
8.5 Simulation examples 249
8.5.1 Microstrip low-pass filter 249
8.5.2 Microstrip branch line coupler 250
8.5.3 Characteristic impedance of a microstrip line 258
8.6 CPML in the two-dimensional FDTD method 264

8.7 MATLAB implementation of CPML in the two-dimensional
FDTD method 267
8.7.1 Definition of CPML 268
8.7.2 Initialization of CPML 268
8.7.3 Application of CPML in the FDTD time-marching loop 269
8.7.4 Validation of CPML performance 271
8.8 Auxiliary differential equation PML 273
8.8.1 Derivation of the ADE-PML formulation 273
8.8.2 MATLAB implementation of the ADE-PML formulation 275
8.9 Exercises 275
9 Near-field to far-field transformation 279
9.1 Implementation of the surface equivalence theorem 281
9.1.1 Surface equivalence theorem 281
9.1.2 Equivalent surface currents in FDTD simulation 282
9.1.3 Antenna on infinite ground plane 285
9.2 Frequency domain near-field to far-field transformation 285
9.2.1 Time-domain to frequency-domain transformation 285
9.2.2 Vector potential approach 286
9.2.3 Polarization of radiation field 287
9.2.4 Radiation efficiency 289
 Contents xi

9.3 MATLAB implementation of near-field to far-field

transformation 289
9.3.1 Definition of NF–FF parameters 289
9.3.2 Initialization of NF–FF parameters 290
9.3.3 NF–FF DFT during time-marching loop 293
9.3.4 Postprocessing for far-field calculation 297
9.4 Simulation examples 309
9.4.1 Inverted-F antenna 309
9.4.2 Strip-fed rectangular dielectric resonator antenna 315
9.5 Exercises 320
10 Thin-wire modeling 323
10.1 Thin-wire formulation 323

10.2 MATLAB implementation of the thin-wire formulation 327
10.3 Simulation examples 330
10.3.1 Thin-wire dipole antenna 330
10.4 An improved thin-wire model 335
10.5 MATLAB implementation of the improved
thin-wire formulation 339
10.6 Simulation example 339
10.7 Exercises 341
11 Scattered field formulation 345
11.1 Scattered field basic equations 345
11.2 The scattered field updating equations 346
11.3 Expressions for the incident plane waves 350

11.4 MATLAB implementation of the scattered field formulation 354
11.4.1 Definition of the incident plane wave 354
11.4.2 Initialization of the incident fields 355
11.4.3 Initialization of the updating coefficients 358
11.4.4 Calculation of the scattered fields 359
11.4.5 Postprocessing and simulation results 361
11.5 Simulation examples 365
11.5.1 Scattering from a dielectric sphere 365
11.5.2 Scattering from a dielectric cube 370
11.5.3 Reflection and transmission coefficients of a
dielectric slab 376
11.6 Exercises 380
xii Contents

12 Total field/scattered field formulation 381

12.1 Introduction 381
12.2 MATLAB implementation of the TF/SF formulation 386
12.2.1 Definition and initialization of incident fields 386
12.2.2 Updating incident fields 389
12.2.3 Updating fields on both sides of the TF/SF boundary 390
12.3 Simulation examples 393
12.3.1 Fields in an empty problem space 394
12.3.2 Scattering from a dielectric sphere 395
12.4 Exercises 396
13 Dispersive material modeling 397
13.1 Modeling dispersive media using ADE technique 398
13.1.1 Modeling Debye medium using ADE technique 398
13.1.2 Modeling Lorentz medium using ADE technique 400
13.1.3 Modeling Drude medium using ADE technique 401
13.2 MATLAB implementation of ADE algorithm for
Lorentz medium 402
13.2.1 Definition of Lorentz material parameters 402
13.2.2 Material grid construction for Lorentz objects 403
13.2.3 Initialization of updating coefficients 406
13.2.4 Field updates in time-marching loop 408
13.3 Simulation examples 410
13.3.1 Scattering from a dispersive sphere 410
13.4 Exercises 412
14 Analysis of periodic structures 413
14.1 Periodic boundary conditions 413
14.2 Constant horizontal wavenumber method 417
14.3 Source excitation 422
14.4 Reflection and transmission coefficients 424
14.4.1 TE mode reflection and transmission coefficients 425
14.4.2 TM mode reflection and transmission coefficients 427
14.4.3 TEM mode reflection and transmission coefficients 428
14.5 MATLAB implementation of PBC FDTD algorithm 429
14.5.1 Definition of a PBC simulation 429
 Contents xiii

14.5.2 Initialization of PBC 431

14.5.3 PBC updates in time-marching loop 434
14.6 Simulation examples 442
14.6.1 Reflection and transmission coefficients of a dielectric slab 442
14.6.2 Reflection and transmission coefficients of a dipole FSS 443
14.6.3 Reflection and transmission coefficients of a
Jarusalem-cross FSS 444
15 Nonuniform grid 447
15.1 Introduction 447
15.2 Transition between fine and coarse grid subregions 447
15.3 FDTD updating equations for the nonuniform grids 452
15.4 Active and passive lumped elements 454
15.5 Defining objects snapped to the electric field grid 457

15.6 MATLAB implementation of nonuniform grids 458
15.6.1 Definition of subregions 459
15.6.2 Initialization of subregions 460
15.6.3 Initialization of updating coefficients 464
15.6.4 Initialization of time step duration 466
15.7 Simulation examples 466
15.7.1 Microstrip patch antenna 466
15.7.2 Three-pole microstrip low-pass filter 467
16 Graphics processing unit acceleration of finite-difference
time-domain method 471
16.1 GPU programming using CUDA 472
16.1.1 Host and device 472
16.1.2 Thread hierarchy 474
16.1.3 Memory hierarchy 476
16.1.4 Performance optimization in CUDA 477
16.1.5 Achieving parallelism 477
16.2 CUDA implementation of two-dimensional FDTD 477
16.2.1 Coalesced global memory access 479
16.2.2 Thread to cell mapping 481
16.2.3 Use of shared memory 486
16.2.4 Optimization of number of threads 487
16.3 Performance of two-dimensional FDTD on CUDA 487
xiv Contents

APPENDIX A One-dimensional FDTD code 491

APPENDIX B Convolutional perfectly-matched layer
regions and associated field updates for
a three-dimensional domain 495
APPENDIX C MATLAB‡ code for plotting far-field patterns 505
APPENDIX D MATLAB‡ GUI for project template 509
References 511
About the authors 519
Index 523
List of figures

1.1 (a) Approximation of the derivative of f ðxÞ at x by finite differences:

forward difference; (b) approximation of the derivative of f ðxÞ at x by finite
differences: backward difference; (c) approximation of the derivative of
f ðxÞ at x by finite differences: central difference 5
1.2 (a) f ðxÞ, f 0 ðxÞ, and differences between f 0 ðxÞ and finite difference
approximations of f 0 ðxÞ for Dx ¼ p/5 and Dx ¼ p/10: f ðxÞ ¼ sin(x)e0.3x;
(b) f ðxÞ, f 0 ðxÞ, and differences between f 0 ðxÞ and finite difference
approximations of f 0 ðxÞ for Dx ¼ p/5 and Dx ¼ p/10: f 0 ðxÞ ¼ cos(x)e0.3x –
0.3 sin(x)e0.3x and finite difference approximations of f 0 ðxÞ for Dx ¼ p/5;
(c) f ðxÞ, f 0 ðxÞ, and differences between f 0 ðxÞ and finite difference
approximations of f 0 ðxÞ for Dx ¼ p/5 and Dx ¼ p/10: errorðf 0 ðxÞ) for
Dx ¼ p/5; (d) f ðxÞ, f 0 ðxÞ, and differences between f 0 ðxÞ and finite difference
approximations of f 0 ðxÞ for Dx ¼ p/5 and Dx ¼ p/10: error( f 0 ðxÞ) for
Dx ¼ p/10 8
1.3 Sample points of f ðxÞ 13
1.4 A three-dimensional FDTD computational space composed of (Nx  Ny  Nz)
Yee cells 14
1.5 Arrangement of field components on a Yee cell indexed as (i, j, k) 14
1.6 Material parameters indexed on a Yee cell 16
1.7 Field components around Ex(i, j, k) 17
1.8 Field components around Hx(i, j, k) 18
1.9 Explicit FDTD procedure 22
1.10 Two-dimensional TEz FDTD field components 24
1.11 Two-dimensional TMz FDTD field components 26
1.12 One-dimensional FDTD – positions of field components Ey and Hz 27
1.13 One-dimensional FDTD – positions of field components Ez and Hy 28
1.14 Snapshots of a one-dimensional FDTD simulation: (a) fields observed
after 100 time steps; (b) fields observed after 300 time steps; (c) fields
observed after 615 time steps; and (d) fields observed after 700 time steps 30

xv
xvi List of figures

2.1 The time–space domain grid with error e propagating with l ¼ 1/2 34
2.2 The time–space domain grid with error e propagating with l ¼ 1 35
2.3 The time–space domain grid with error e propagating with l ¼ 2 35
2.4 Maximum magnitude of Ez in the one-dimensional problem space for
simulations with Dt ¼ 3.3356 ps and Dt ¼ 3.3357 ps 36
2.5 (a) Maximum magnitude of Ez in the one-dimensional problem space:
simulation with Dx ¼ 1 mm; (b) maximum magnitude of Ez in the
one-dimensional problem space: simulation with Dx ¼ 4 mm 40
3.1 Parameters defining a brick and a sphere in Cartesian coordinates:
(a) parameters defining a brick and (b) parameters defining a sphere 50
3.2 An FDTD problem space and the objects defined in it 50
3.3 A cell around material component ez(i, j, k) partially filled with two media 51
3.4 A cell around material component ez(i, j, k) divided into eight subcells 52
3.5 Material component ez(i, j, k) parallel to the boundary of two different media
partially filling a material cell 53
3.6 Material component ez(i, j, k) normal to the boundary of two different
media partially filling a material cell 53
3.7 Material component ez(i, j, k) located between four Yee cells filled with four
different material types 55
3.8 Material component mz(i, j, k) located between two Yee cells filled with two
different material types 56
3.9 A PEC plate with zero thickness surrounded by four s material components
e
57
3.10 An FDTD problem space and the objects approximated by snapping to cells 62
3.11 Positions of the Ex components on the Yee grid for a problem space
composed of (Nx  Ny  Nz) cells 63
3.12 Positions of the Hx components on the Yee grid for a problem space
composed of (Nx  Ny  Nz) cells 64
3.13 Material grid on three plane cuts: (a) relative permittivity components
and (b) relative permeability components 67
3.14 The FDTD problem space of Exercise 3.1: (a) geometry of the problem
and (b) relative permittivity distribution 68
3.15 The FDTD problem space of Exercise 3.3: (a) geometry of the problem
and (b) electric conductivity distribution 69
4.1 Field components around Ez(i, j, k) 72
4.2 Voltage sources placed between nodes (i, j, k) and (i, j, k þ 1): (a) voltage
source with internal resistance (soft source) and (b) voltage source without
internal resistance (hard source) 73
4.3 Lumped elements placed between nodes (i, j, k) and (i, j, k þ 1): (a) current
source with internal resistance and (b) resistor 75
4.4 Lumped elements placed between nodes (i, j, k) and (i, j, k þ 1): (a) capacitor
and (b) inductor 77
 List of figures xvii

4.5 Parallel PEC plates excited by a voltage source at one end and terminated
by a resistor at the other end 80
4.6 A voltage source distributed over a surface and a resistor distributed over a
volume 80
4.7 Diodes placed between nodes (i, j, k) and (i, j, k þ 1): (a) diode oriented in
the positive z direction and (b) diode oriented in the negative z direction 82
4.8 A function f ðxÞ and the points approaching to the root of the function
iteratively calculated using the Newton–Raphson method 84
4.9 A diode defined between the nodes (is, js, ks) and (ie, je, ke) 96
4.10 A z-directed voltage source defined between the nodes (is, js, ks) and
(ie, je, ke) 99
4.11 A volume between PEC plates where a z-directed sampled voltage is required 110
4.12 A surface enclosed by magnetic field components and z-directed current
flowing through it 111
4.13 The y components of the magnetic and electric fields around the node
(is, js, ks): (a) magnetic field components and (b) electric field components 127
4.14 A voltage source terminated by a resistor 133
4.15 Sinusoidal voltage source waveform with 500 MHz frequency and
sampled voltage and current: (a) source waveform and sampled voltage
and (b) sampled current 136
4.16 Sinusoidal voltage source waveform with 500 MHz frequency and
sampled voltage: (a) a voltage source terminated by a diode and (b) source
waveform and sampled voltage 137
4.17 Unit step voltage source waveform and sampled voltage: (a) a voltage source
terminated by a capacitor and (b) sampled voltage and analytical solution 139
4.18 Dimensions of 50 W stripline 141
4.19 A 50 W stripline terminated by a resistor 142
4.20 A circuit composed of a voltage source, a capacitor, and a diode 142
5.1 A sinusoidal waveform excited for 4 s: (a) x(t) and (b) X(w) magnitude 144
5.2 A sinusoidal waveform excited for 8 s: (a) x(t) and (b) X(w) magnitude 146
5.3 A Gaussian waveform and its Fourier transform: (a) g(t) and (b) G(w)
magnitude 147
5.4 Gaussian waveform shifted by t0 ¼ 4.5t in time 149
5.5 Normalized derivative of a Gaussian waveform and its Fourier transform:
(a) g(t) and (b) G(w) magnitude 150
5.6 Cosine-modulated Gaussian waveform and its Fourier transform: (a) g(t)
and (b) G(w) magnitude 152
5.7 A Gaussian waveform and its Fourier transform: (a) g(t); (b) G(w); and
(c) g(t) recovered from G(w) 161
5.8 An RLC circuit: (a) circuit simulated in FDTD and (b) lumped element
equivalent circuit 162
xviii List of figures

5.9 Time-domain response, Vo, compared with source waveform, Vs 165

5.10 Frequency-domain response of source waveform Vs and output voltage
Vo: (a) Fourier transform of Vs and (b) Fourier transform of Vo 166
Vo ðwÞ
5.11 Transfer function TðwÞ ¼ Vs ðwÞ 167
6.1 An N-port network 170
6.2 A microstrip low-pass filter terminated by a voltage source and a resistor
on two ends: (a) three-dimensional view and (b) dimensions 171
6.3 S-parameters of the microstrip low-pass filter: (a) S11 and (b) S21 176
6.4 Sampled voltage at the second port of the microstrip low-pass filter 177
6.5 Sampled voltage at the second port of the microstrip low-pass filter with
se ¼ 0.2 177
6.6 S-parameters of the microstrip low-pass filter with se ¼ 0.2: (a) S11 and
(b) S21 178
6.7 The geometry and dimensions of a microstrip quarter-wave transformer
matching a 100 W line to 50 W line 179
6.8 The S11 and S21 of the microstrip quarter-wave transformer circuit 183
6.9 The problem space for the low-pass filter with absorbers on the xn, xp,
yn, yp, and zp sides 184
7.1 The field decomposition of a TEz polarized plane wave 186
7.2 The plane wave transition at the interface between two PML media 189
7.3 The loss distributions in two-dimensional PML regions 190
7.4 Nonzero regions of PML conductivities for a three-dimensional FDTD
simulation domain: (a) nonzero spex and spmx; (b) nonzero spey and spmy;
(c) nonzero spez and spmz; and (d) overlapping PML regions 193
7.5 Nonzero TEz regions of PML conductivities: (a) nonzero spex; (b) nonzero
spey; (c) nonzero spmx; and (d) nonzero spmy 196
7.6 Nonzero TMz regions of PML conductivities: (a) nonzero spmx; (b) nonzero
spmy; (c) nonzero spex; and (d) nonzero spey 198
7.7 A two-dimensional FDTD problem space with PML boundaries 203
7.8 TEz field components in the PML regions 206
7.9 Field components updated by PML equations 207
7.10 TMz field components in the PML regions 209
7.11 Field components updated by PML equations 210
7.12 A two-dimensional TEz FDTD problem terminated by PML boundaries and
its simulation results: (a) an empty two-dimensional problem space and
(b) sampled Hz in time 216
7.13 A two-dimensional TEz FDTD problem used as open boundary reference and
its simulation results: (a) an empty two-dimensional problem space and
(b) sampled Hz in time 218
7.14 Error in time and frequency domains 219
 List of figures xix

7.15 A two-dimensional problem space including a cylinder and a line source 220
7.16 Sampled electric field at a point between the cylinder and the line source 224
7.17 Magnitude of electric field distribution calculated by FDTD 224
7.18 Magnitude of electric field distribution calculated by BVS 225
7.19 Magnitude of electric field distribution calculated by FDTD using DFT at
1 GHz 227
7.20 Magnitude of electric field distribution calculated by FDTD using DFT at
2 GHz 227
8.1 The FDTD flowchart including the steps of CPML algorithm 234
8.2 Regions where CPML parameters are defined: (a) spex, spmx, aex, amx, kex,
and kmx; (b) spey, spmy, aey, amy, key, and kmy; (c) spez, spmz, aez, amz, kez,
and kmz; and (d) overlapping CPML regions 237
8.3 A problem space with 20  20  4 cells brick 240
8.4 Positions of the electric field components updated by CPML in the xp
region 246
8.5 Positions of the magnetic field components updated by CPML in the xp
region 246
8.6 Source voltage and sampled voltages observed at the ports of the low-pass
filter 250
8.7 Sampled currents observed at the ports of the low-pass filter 251
8.8 S11 of the low-pass filter 251
8.9 S21 of the low-pass filter 252
8.10 An FDTD problem space including a microstrip branch line coupler 252
8.11 S-parameters of the branch line coupler 259
8.12 An FDTD problem space including a microstrip line 259
8.13 Source voltage and sampled voltage of the microstrip line 263
8.14 Current on the microstrip line 263
8.15 S11 of the microstrip line 264
8.16 Characteristic impedance of the microstrip line 264
8.17 Updated fields and CPML parameters in TEz case: (a) xn and xp CPML
regions and (b) yn and yp CPML regions 266
8.18 Updated fields and CPML parameters in TMz case: (a) xn and xp CPML
regions and (b) yn and yp CPML regions 267
8.19 CPML and PML errors of sampled Hz in time and frequency domains:
(a) errors in time domain and (b) errors in frequency domain 272
8.20 The problem space for the quarter-wave transformer with CPML boundaries
on the xn, xp, yn, yp, and zp sides 276
8.21 The problem space for the quarter-wave transformer with the substrate
penetrating into the CPML boundaries on the xn and xp sides 276
8.22 The microstrip line reference case 277
xx List of figures

9.1 Near-field to far-field transformation technique: equivalent currents on an

imaginary surface 280
9.2 Two paths of the near-field to far-field transformation technique are
implemented to achieve different computation objectives 281
9.3 Surface equivalence theorem: (a) original problem and (b) equivalent
problem for region outside S 282
9.4 An imaginary surface is selected to enclose the antennas or scatterers 283
9.5 Equivalent surface currents on the imaginary closed surface 283
9.6 An imaginary closed surface on an infinite PEC ground plane using the
image theory 285
9.7 The equivalent surface current source and far field 286
9.8 Ex field components on the yn face of the NF–FF imaginary surface and the
magnetic currents generated by them 296
9.9 Hx field components around the yn face of the NF–FF imaginary surface
and the electric currents generated by them 297
9.10 Position vectors for sources on the zn and zp faces 302
9.11 Position vectors for sources on the xn and xp faces 303
9.12 Position vectors for sources on the yn and yp faces 304
9.13 An inverted-F antenna 309
9.14 Input reflection coefficient of the inverted-F antenna 314
9.15 Radiation patterns in the xy plane cut: (a) 2.4 GHz and (b) 5.8 GHz 314
9.16 Radiation patterns in the xz plane cut: (a) 2.4 GHz and (b) 5.8 GHz 315
9.17 Radiation patterns in the yz plane cut: (a) 2.4 GHz and (b) 5.8 GHz 315
9.18 A strip-fed rectangular DRA 316
9.19 Input reflection coefficient of the strip-fed rectangular DRA 319
9.20 Radiation patterns in the xy plane cut: (a) 3.5 GHz and (b) 4.3 GHz 319
9.21 Radiation patterns in the xz plane cut: (a) 3.5 GHz and (b) 4.3 GHz 320
9.22 Radiation patterns in the yz plane cut: (a) 3.5 GHz and (b) 4.3 GHz 320
9.23 A dipole antenna 321
9.24 A microstrip patch antenna 322
9.25 A microstrip patch antenna with a microstrip line feeding 322
10.1 A thin wire with its axis coinciding with Ez(i, j, k) and field components
surrounding Hy(i, j, k) 324
10.2 Magnetic field components surrounding the thin wire 324
10.3 A thin-wire dipole antenna 330
10.4 S11 of the thin-wire dipole antenna 333
10.5 Input impedance of the thin-wire dipole antenna 333
10.6 Radiation pattern in the xy plane cut 334
10.7 Radiation pattern in the xz plane cut 334
 List of figures xxi

10.8 Radiation pattern in the yz plane cut 335

10.9 Constant magnetic field on a loop circulating around a thin wire 336
10.10 A magnetic field component on the edge of a rectangular loop 336
10.11 Projection of magnetic field from a circular loop to a rectangular edge:
(a) large view and (b) details of the right edge 337
10.12 Electric field component Ex at the center of a cell face 338
10.13 Power reflection coefficient of the thin-wire dipole antenna 341
10.14 Input impedance of the thin-wire dipole antenna 342
10.15 Power reflection coefficient of the thin-wire dipole antenna: FDTD
implementation includes the end-caps model [41] and the stabilized RVS
model [42] 342
10.16 An antenna array composed of two thin-wire dipole antennas 343
10.17 A thin-wire loop 344
11.1 An incident plane wave 351
11.2 An incident plane wave delayed in time and shifted in space 352
11.3 Two incident plane waves: one traveling toward the origin and the
other traveling away from the origin 353
11.4 An FDTD problem space including a dielectric sphere 365
11.5 Scattered electric field captured at the origin and the incident field waveform 368
11.6 Bistatic RCS at 1 GHz in the xy plane 368
11.7 Bistatic RCS at 1 GHz in the xz plane 369
11.8 Bistatic RCS at 1 GHz in the yz plane 369
11.9 Calculated RCSq at 1 GHz in the xz plane compared with the analytical
solution 370
11.10 Calculated RCSf at 1 GHz in the yz plane compared with the analytical
solution 370
11.11 Scattered field due to a sphere captured on the xz plane cut: (a) at time
step 120; (b) at time step 140; (c) at time step 160; and (d) at time step 180 371
11.12 An FDTD problem space including a cube, and the surface mesh of the
cube used by the MoM solver 372
11.13 Bistatic RCS at 1 GHz in the xy plane 374
11.14 Bistatic RCS at 1 GHz in the xz plane 374
11.15 Bistatic RCS at 1 GHz in the yz plane 375
11.16 Calculated RCSq at 1 GHz in the xz plane compared with the MoM solution 375
11.17 Calculated RCSf at 1 GHz in the xz plane compared with the MoM solution 376
11.18 An FDTD problem space including a dielectric slab 376
11.19 Reflection and transmission coefficients of the dielectric slab 380
xxii List of figures

12.1 Total and scattered field regions for a three-dimensional FDTD

computational domain 382
12.2 Total and scattered field components and regions for a one-dimensional
FDTD computational domain 382
12.3 Ex and Hz field components that needs incident field compensation on the
yn face of the imaginary surface 384
12.4 Ez and Hx field components that needs incident field compensation on the
yn face of the imaginary surface 384
12.5 Electric field distribution on a yz plane-cut at three progressive time instants 394
12.6 Electric field distribution on the xz plane: (a) at time step 120; (b) at time
step 140; (c) at time step 160; and (d) at time step 180 395
13.1 Update sequence of fields in the presented Debye modeling algorithm 400
13.2 Update sequence of fields in the presented Lorentz and Drude modeling
algorithms 401
13.3 Complex relative permittivity versus frequency 410
13.4 Co-polarized RCS of a dispersive sphere at three frequencies 411
13.5 Cross-polarized RCS of a dispersive sphere at three frequencies 411
14.1 A periodic structure and a unit cell of two-dimension periodicity 414
14.2 Magnitude of reflection coefficient in the angle-frequency plane 415
14.3 Propagation vector of an incident plane wave and its horizontal component 415
14.4 Magnitude of reflection coefficient in the horizontal wavenumber-frequency
plane 416
14.5 Propagation vectors with the same horizontal wavenumber 417
14.6 An incident plane wave and a unit cell 418
14.7 Field components on an xy plane-cut of a unit cell in a three-dimensional
domain: (a) Ex , Ey , and Hz components and (b) Hx , Hy , and Ez components 419
14.8 Flow chart of PBC FDTD algorithm 420
14.9 Illustration of incident fields in the TE case 425
14.10 Illustration of incident fields in the TM case 427
14.11 Illustration of incident fields in the TEM case 428
14.12 Reflection and transmission coefficients of a dielectric slab 442
14.13 TE mode reflection coefficients of a DFSS: FDTD vs Ansoft Designer 443
14.14 TE mode transmission coefficients of a DFSS: FDTD vs Ansoft Designer 444
14.15 TM mode reflection coefficients of a DFSS: FDTD vs Ansoft Designer 444
14.16 TM mode transmission coefficients of a DFSS: FDTD vs Ansoft Designer 445
14.17 A Jarusalem-cross frequency selective circuit 445
14.18 Reflection and transmission coefficients of a Jarusalem-cross FSS: FDTD
vs Ansoft Designer 446
 List of figures xxiii

15.1 (a) An FDTD computational domain with coarse uniform grid; (b) an
FDTD computational domain with fine uniform grid; and (c) an FDTD
computational domain with nonuniform grid 448
15.2 Uniform and nonuniform subregions 450
15.3 Uniform electric and magnetic field grids in a one-dimensional space 451
15.4 Electric and magnetic field positions in a nonuniform grid 451
15.5 A fine grid subregion embedded in a coarse base grid 451
15.6 Electric field grid around Ex ði; j; k Þ, and magnetic grid cell sizes used to
update Ex ði; j; k Þ 453
15.7 Material component ez(i, j, k) located between four Yee cells filled with
four different material types 457
15.8 Material component mz(i, j, k) located between two Yee cells filled with
two different material types 458
15.9 Cell view of the microstrip rectangular patch antenna 467
15.10 Comparison of power reflection coefficient of fine grid, coarse grid, and
nonuniform grid simulations of a microstrip rectangular patch antenna 468
15.11 Geometry and dimensions of a three-pole low-pass filter 468
15.12 Comparison of scattering parameters of fine grid, coarse grid, and
nonuniform grid simulations of a three-pole low-pass microstrip filter 469
16.1 Parallel threads executing the same code on different set of data identified
by the threadIdx.x value 474
16.2 Grid of thread blocks 475
16.3 An extended two-dimensional computational domain 480
16.4 A grid of thread blocks that spans a 48  32 cells two-dimensional
problem space 483
16.5 Threads in a thread block 485
16.6 Throughput of CUDA FDTD calculations on a Tesla C1060 card 488
16.7 A snapshot of electric field distribution scattered from a dielectric cylinder
due to a line source with sinusoidal excitation 489
B.1 CPML regions where Ex is updated: (a) Ex is updated in yn and yp using
yexy and (b) Ex is updated in zn and zp using yexz 496
B.2 CPML regions where Ey is updated: (a) Ey is updated in zn and zp using
yeyz and (b) Ey is updated in xn and xp using yeyx 497
B.3 CPML regions where Ez is updated: (a) Ez is updated in xn and xp using
yezx and (b) Ez is updated in yn and yp using yezy 499
B.4 CPML regions where Hx is updated: (a) Hx is updated in yn and yp using
yhxy and (b) Hx is updated in zn and zp using yhxz 500
B.5 CPML regions where Hy is updated: (a) Hy is updated in zn and zp using
yhyz and (b) Hy is updated in xn and xp using yhyx 501
B.6 CPML regions where Hz is updated: (a) Hz is updated in xn and xp using
yhzx and (b) Hz is updated in yn and yp using yhzy 502
D.1 The GUI to create template problem definition files for a project 509
List of tables

1.1 Finite difference formulas for the first- and second-order derivatives
where the function f with the subscript i is an abbreviation of f ðxÞ. Similar
notation can be implemented for f (x þ Dx), f (x þ 2Dx), etc., as shown in
Figure 1.3. FD, forward difference; BD, backward difference; CD, central
difference 13
15.1 Comparison of fine grid, coarse grid, and nonuniform grid simulations of a
microstrip rectangular patch antenna 467
15.2 Comparison of fine grid, coarse grid, and nonuniform grid simulations of a
three-pole low-pass microstrip filter 469

xxiv
Preface

The contents of this book have evolved gradually and carefully from many years of teaching
graduate level courses in computational electromagnetics generally, and the Finite-Differ-
ence Time-Domain (FDTD) method specifically. The authors have further refined and
developed the materials by teaching numerous short courses on the FDTD method at various
educational institutions and at a growing number of international conferences. The theore-
tical, numerical, and programming experience of the second author, Veysel Demir, has been
a crucial factor in bringing the book to successful completion.

Objective
The objective of the book is to introduce the powerful FDTD method to students and
interested researchers and readers. An effective introduction is accomplished using a step-
by-step process that builds competence and confidence in developing complete working
codes for the design and analysis of various antennas and microwave devices. This book will
serve graduate students, researchers, and those in industry and government who are using
other electromagnetics tools and methods for the sake of performing independent numerical
confirmation. No previous experience with finite-difference methods is assumed of readers
to use this book.

Important topics
The main topics in the book include the following:
● finite-difference approximation of the differential form of Maxwell’s equations
● geometry construction in discrete space, including the treatment of the normal and
tangential electric and magnetic field components at the boundaries between different media
● outer-boundary conditions treatment
● appropriate selection of time and spatial increments
● selection of the proper source waveform for the intended application
● correct parameters for the time-to-frequency-domain transformation
● simulation of thin wires

xxv
xxvi Preface

● representation of lumped passive and some active elements

● total field/scattered field formulation
● definition and formulation of fields from near and far zone
● modeling of dispersive material
● analysis of periodic structures
● treatment of nonuniform grid
● use of graphical processing units for accelerating the simulations.

MATLAB‡ code organization

In the first edition of this book the chapters are presented in such a way that, by adding/
developing a new part of the code, chapter by chapter, at the end a well-developed FDTD
simulation package can be constructed. In the second edition, the first 11 chapters are either
the same as the ones in the first edition or extended with additional topics, therefore one can
progressively develop a package by studying the first 11 chapters.
Chapters 12–16 in the second edition are advanced topics in FDTD, and it is not feasible
to follow the progressive code development approach as implementation of these advanced
concepts, one over another, makes the code development cumbersome and deviate reader’s
attention from learning a concept to getting lost in a pile of complex code. As an alternative,
the functional code that is developed until Chapter 11 is reorganized as a base code and it is
being used as the start point for each of the advanced topics chapters.

Use of MATLAB‡
The development of the working code is based on the MATLAB programming language due
to its relative ease of use, widespread availability, familiarity to most electrical engineers,
and its powerful capabilities for providing graphical outputs and visualization. The book
illustrates how the key FDTD equations are derived, provides the final expressions to be
programmed, and also includes sample MATLAB codes developed for these equations.
None of the specialized MATLAB tool boxes is required to run the presented codes in this
book.
The MATLAB M-files for all programming examples shown in the text are available
from the publisher on request by emailing [email protected].

Key strengths
The strengths of the book can be summarized in four points:
● First, the derivations of the FDTD equations are presented in an understandable,
detailed, and complete manner, which makes it easy for beginners to grasp the FDTD
concepts. Further to that point, regardless of the different treatments required for var-
ious objects, such as dielectrics, conductors, lumped elements, active devices, or thin
 Preface xxvii

wires, the FDTD updating equations are provided with a consistent and unified
notation.
● Second, many three-dimensional figures are presented to accompany the derived
equations. This helps readers visualize, follow, and link the components of the equa-
tions with the discrete FDTD spatial domain.
● Third, it is well known that readers usually face difficulties while developing numerical
tools for electromagnetics applications even though they understand the theory.
Therefore, we introduce in this book a top-down software design approach that links
the theoretical concepts with program development and helps in constructing an FDTD
simulation tool as the chapters proceed.
● Fourth, fully worked out practical examples showing how the MATLAB codes for each
example is developed to promote both the understanding and the visualization of the
example configuration, its FDTD parameters and setup procedure, and finally the fre-
quency and/or time domain corresponding solutions.
At the end of each chapter, readers will find a set of exercises that will emphasize the key
features presented in that chapter. Since most of these exercises require code developments
and the orientation of geometries and sources can be chosen arbitrarily, a suggested con-
figuration of each in almost all of these exercises is provided in a three-dimensional figure.
The authors hope that with this coverage of the FDTD method, along with the supplied
MATLAB codes in a single book, readers will be able to learn the method, to develop their
own FDTD simulation tool, and to start enjoying, with confidence, the simulation of a
variety of electromagnetic problems.

Instructor resources
Upon adopting this book as the required text, instructors are entitled to obtain the solutions of
the exercises located at the end of each chapter. Most of these solutions are in the form of
MATLAB files, and can be obtained by contacting the publisher at [email protected].

Errors and suggestions

The authors welcome any reader feedback related to suspected errors and to suggestions for
improving the presentation of the topics provided in this book.
Acknowledgements

The authors would like to first thank God for giving us the endurance to initiate and complete
this book. Many thanks go to our family members for their support, patience, and sacrifice
during the preparation and completion of this book. The authors greatly acknowledge the co-
authorship of Chapter 14 with Drs. Khaled Elmahgoub and Fan Yang in the second edition
of the book, and the contribution of Dr. Matthew J. Inman to Chapter 12 in the first edition of
the book. They would also like to thank Dr. Branko Kolundzija for providing a copy of
WIPL-D software package to use for verifications for some of the presented examples.
The first author acknowledges the support and advice over many years of the late Charles
E. Smith. Without his continuous encouragement and support this book would have not been
finalized. He also thanks many of his colleagues, postdoctoral fellows and visiting scholars
who worked with him in topics related to FDTD, the graduate students who participated in
translating ideas into reality, and the undergraduate students who participated in research
projects with this author, in particular, Allen Glisson, Fan Yang, Abdel-Fattah A. Elsohly, Joe
LoVetri, Veysel Demir, Matthew J. Inman, Chun-Wen Paul Huang, Clayborne D. Taylor,
Mohamed Al Sharkawy, Vicente Rodriguez-Pereyra, Jianbing (James) Chen, Adel M. Abdin,
Bradford N. Baker, Asem Mokaddem, Liang Xu, Nithya L. Iyer, Shixiong Dai, Xuexun Hu,
Cuthbert Martindale Allen, Khaled Elmahgoub, and Terry Gerald. Chapter 14 is co-authored
by Drs. Khaled Elmahgoub and Fan Yang.
The authors thank Dudley Kay and Joanna Hughes of SciTech/IET Publishing for their
encouragement, patience, and useful suggestions during the course of the development of the
manuscript and production of this book.

xxviii
 CHAPTER 1

Introduction to FDTD

Computational electromagnetics (CEM) has evolved rapidly during the past decade to a
point where now extremely accurate predictions can be given for a variety of electro-
magnetic problems, including the scattering cross-section of radar targets and the precise
design of antennas and microwave devices. In general, commonly used CEM methods today
can be classified into two categories. The first is based on differential equation (DE) meth-
ods, whereas the second is based on integral equation (IE) methods. Both IE and DE solution
methods are based on the applications of Maxwell’s equations and the appropriate boundary
conditions associated with the problem to be solved. The IE methods in general provide
approximations for IEs in terms of finite sums, whereas the DE methods provide approx-
imations for DEs as finite differences.
In previous years, most numerical electromagnetic analysis has taken place in the
frequency domain where time-harmonic behavior is assumed. Frequency domain was
favored over time domain because a frequency-domain approach is more suitable for
obtaining analytical solutions for canonical problems, which are used to verify the numerical
results obtained as a first step before depending on a newly developed numerical method for
generating data for real-world applications. Furthermore, the experimental hardware avail-
able for making measurements in past years was largely confined to the frequency-domain
approach.
The recent development of faster and more powerful computational resources allowed for
more advanced time-domain CEM models. More focus is directed toward DE time-domain
approaches as they are easier to formulate and to adapt in computer simulation models
without complex mathematics. They also provide more physical insight to the characteristics
of the problems.
Therefore, an in-depth analysis and implementation of the commonly used time-domain
DE approach, namely, the finite-difference time-domain (FDTD) method for CEM applica-
tions, is covered in this book, along with applications related to antenna designs, microwave
filter designs, and radar cross-section analysis of three-dimensional targets.
The FDTD method has gained tremendous popularity in the past decade as a tool for
solving Maxwell’s equations. It is based on simple formulations that do not require complex
asymptotic or Green’s functions. Although it solves the problem in time, it can provide
frequency-domain responses over a wide band using the Fourier transform. It can easily
handle composite geometries consisting of different types of materials including dielectric,

1
2 CHAPTER 1 ● Introduction to FDTD

magnetic, frequency-dependent, nonlinear, and anisotropic materials. The FDTD technique

is easy to implement using parallel computation algorithms. These features of the FDTD
method have made it the most attractive technique of CEM for many microwave devices and
antenna applications.
FDTD has been used to solve numerous types of problems arising while studying many
applications, including the following:
● scattering, radar cross-section
● microwave circuits, waveguides, fiber optics
● antennas (radiation, impedance)
● propagation
● medical applications
● shielding, coupling, electromagnetic compatibility (EMC), electromagnetic pulse (EMP)
protection
● nonlinear and other special materials
● geological applications
● inverse scattering
● plasma

1.1 The finite-difference time-domain method

basic equations
The starting point for the construction of an FDTD algorithm is Maxwell’s time-domain
equations. The differential time-domain Maxwell’s equations needed to specify the field
behavior over time are
!
@D !
!
rH ¼ þ J; ð1:1aÞ
@t
!
! @B !
rE ¼  M; ð1:1bÞ
@t
!
r  D ¼ re ; ð1:1cÞ
!
r  B ¼ rm ; ð1:1dÞ
! !
where E is the electric field strength vector
!
in volts per meter, D is the electric displacement
vector in! coulombs per square meter, H is the magnetic field strength vector!in amperes per
meter, B is the magnetic flux density vector in webers !
per square meter, J is the electric
current density vector in amperes per square meter, M is the magnetic current density vector
in volts per square meter, re is the electric charge density in coulombs per cubic meter, and
rm is the magnetic charge density in webers per cubic meter.
Constitutive relations are necessary to supplement Maxwell’s equations and characterize
the material media. Constitutive relations for linear, isotropic, and nondispersive materials
can be written as
! !
D ¼ eE; ð1:2aÞ
! !
B ¼ mH ; ð1:2bÞ
 1.1 ● The finite-difference time-domain method basic equations 3

where e is the permittivity, and m is the permeability of the material. In free space

e ¼ e0  8:854  1012 farad=meter;

m ¼ m0 ¼ 4p  107 henry=meter:
We only need to consider curl equations (1.1a) and (1.1b) while deriving FDTD
equations because the divergence equations can be !
satisfied by the developed FDTD
updating!equations
!
[1]. The electric current density J is
!
the !sum !of the! conduction current
density J c ¼ se E and the !impressed
!
current
!
density
!
J i as J! ¼ J c þ J i : Similarly, for the
magnetic current density, M ¼ M c þ M i ; where M c ¼ sm H: Here se is the electric con-
ductivity in siemens per meter, and sm is the magnetic conductivity in ohms per meter. Upon
decomposing the current densities in (1.1) to conduction and impressed components and by
using the constitutive relations (1.2) we can rewrite Maxwell’s curl equations as
!
! @E ! !
rH ¼e þ se E þ J i ; ð1:3aÞ
@t
!
! @H ! !
r  E ¼ m  sm H  M i : ð1:3bÞ
@t
! ! ! !
This formulation treats only the electromagnetic fields E and H and not the fluxes D and B:
All four constitutive parameters e, m, se, and sm are present so that any linear isotropic material
can be specified. Treatment of electric and magnetic sources is included through the impressed
currents. Although only the curl equations are used and the divergence equations are not part of
the FDTD formalism, the divergence!
equations
! !
can! be used as a test on !the predicted
!
field
response, !so that!after forming D ¼ eE and B ¼ mH from the predicted E and H fields, the
resulting D and B must satisfy the divergence equations.
Equation (1.3) is composed of two vector equations, and each vector equation can be
decomposed into three scalar equations for three-dimensional space. Therefore, Maxwell’s
curl equations can be represented with the following six scalar equations in a Cartesian
coordinate system (x, y, z):
 
@Ex 1 @Hz @Hy
¼   sex Ex  J ix ; ð1:4aÞ
@t ex @y @z
 
@Ey 1 @Hx @Hz
¼   sy Ey  J iy ;
e
ð1:4bÞ
@t ey @z @x
 
@Ez 1 @Hy @Hx
¼   sz Ez  J iz ;
e
ð1:4cÞ
@t ez @x @y
 
@Hx 1 @Ey @Ez
¼   sx Hx  Mix ;
m
ð1:4dÞ
@t mx @z @y
 
@Hy 1 @Ez @Ex
¼   sy Hy  Miy ;
m
ð1:4eÞ
@t my @x @z
 
@Hz 1 @Ex @Ey
¼   sm
z zH  M iz : ð1:4f Þ
@t mz @y @x
4 CHAPTER 1 ● Introduction to FDTD

The material parameters ex, ey, and ez are associated with electric field components Ex,
Ey, and Ez through constitutive relations Dx ¼ exEx, Dy ¼ eyEy, and Dz ¼ ezEz, respectively.
Similarly, the material parameters mx, my, and mz are associated with magnetic field compo-
nents Hx, Hy, and Hz through constitutive relations Bx ¼ mxHx, By ¼ myHy, and Bz ¼ mzHz,
respectively. Similar decompositions for other orthogonal coordinate systems are possible,
but they are less attractive from the applications point of view.
The FDTD algorithm divides the problem geometry into a spatial grid where electric and
magnetic field components are placed at certain discrete positions in space, and it solves
Maxwell’s equations in time at discrete time instances. This can be implemented by first
approximating the time and space derivatives appearing in Maxwell’s equations by finite
differences and next by constructing a set of equations that calculate the values of fields at a
future time instant from the values of fields at a past time instant, therefore constructing a
time marching algorithm that simulates the progression of the fields in time [2].

1.2 Approximation of derivatives by finite differences

An arbitrary continuous function can be sampled at discrete points, and the discrete function
becomes a good approximation of the continuous function if the sampling rate is sufficient
relative to the function’s variation. Sampling rate determines the accuracy of operations
performed on the discrete function that approximates the operations on the continuous
functions as well. However, another factor that determines the accuracy of an operation on a
discrete function is the choice of the discrete operator. Most of the time it is possible to use
more than one way of performing an operation on a discrete function. Here we will consider
the derivative operation.
Consider the continuous function given in Figure 1.1(a–c), sampled at discrete points.
The expression for the derivative of f ðxÞ at point x can be written as
f ðx þ DxÞ  f ðxÞ
f 0ðxÞ ¼ lim : ð1:5Þ
Dx!0 Dx
However, since Dx is a nonzero fixed number, the derivative of f ðxÞ can be approxi-
mately taken as
f ðx þ DxÞ  f ðxÞ
f 0ðxÞ  : ð1:6Þ
Dx
The derivative of f ðxÞ is the slope of the dashed line as illustrated in Figure 1.1(a).
Equation (1.6) is called the forward difference formula since one forward point f (x þ Dx) is
used to evaluate f 0ðxÞ together with f ðxÞ.
It is evident that another formula for an approximate f 0ðxÞ can be obtained by using a
backward point f ðx  DxÞ rather than the forward point f ðx þ DxÞ as illustrated in Figure 1.1(b),
which can be written as
f ðxÞ  f ðx  DxÞ
f 0ðxÞ  : ð1:7Þ
Dx
This equation is called the backward difference formula due to the use of the backward
point fðx  DxÞ.
 1.2 ● Approximation of derivatives by finite differences 5

f (x + Δ x)

f (x – Δ x)

f (x)

(a) x – Δx x x + Δx

f (x + Δ x)

f (x – Δ x)
f (x)

(b) x – Δx x x + Δx

f (x +Δ x)

f (x – Δ x)

f (x)

(c) x – Δx x x + Δx

Figure 1.1 (a) Approximation of the derivative of f ðxÞ at x by finite differences: forward
difference; (b) approximation of the derivative of f ðxÞ at x by finite differences:
backward difference; (c) approximation of the derivative of f ðxÞ at x by finite
differences: central difference.
6 CHAPTER 1 ● Introduction to FDTD

The third way of obtaining a formula for an approximate f 0 ðxÞ is by averaging the for-
ward difference and backward difference formulas, such that

f ðx þ DxÞ  f ðx  DxÞ
f 0 ðxÞ  : ð1:8Þ
Dx

Equation (1.8) is called the central difference formula since both the forward and back-
ward points around the center are used. The line representing the derivative of f ðxÞ calcu-
lated using the central difference formula is illustrated in Figure 1.1(c). It should be noted
that the value of the function f ðxÞ at x is not used in central difference formula.
Examination of Figure 1.1 immediately reveals that the three different schemes yield
different values for f 0 ðxÞ, with an associated amount of error. The amount of error introduced
by these difference formulas can be evaluated analytically by using the Taylor series
approach. For instance, the Taylor series expansion of f ðx þ DxÞ can be written as

ðDxÞ2 00 ðDxÞ3 000 ðDxÞ4 0000

f ðx þ DxÞ ¼ f ðxÞ þ Dxf 0 ðxÞ þ f ðxÞ þ f ðxÞ þ f ðxÞ þ ⋯: ð1:9Þ
2 6 24

This equation gives an exact expression for f ðx þ DxÞ as a series in terms of Dx and
derivatives of f ðxÞ, if f ðxÞ satisfies certain conditions and infinite number of terms, theore-
tically, are being used. Equation (1.9) can be rearranged to express f 0 ðxÞ as

f ðx þ DxÞ  f ðxÞ Dx 00 ðDxÞ2 000 ðDxÞ3 0000

f 0 ðxÞ ¼  f ðxÞ  f ðxÞ  f ðxÞ  ⋯: ð1:10Þ
Dx 2 6 24

Here it can be seen that the first term on the right-hand side of (1.10) is the same as the
forward difference formula given by (1.6). The sum of the rest of the terms is the difference
between the approximate derivative given by the forward difference formula and the exact
derivative f 0 ðxÞ, and hence is the amount of error introduced by the forward difference
formula. Equation (1.10) can be rewritten as

f ðx þ DxÞ  f ðxÞ
f 0 ðxÞ ¼ þ OðDxÞ; ð1:11Þ
Dx

where O(Dx) represents the error term. The most significant term in O(Dx) is Dx/2, and the
order of Dx in this most significant term is one. Therefore, the forward difference formula is
first-order accurate. The interpretation of first-order accuracy is that the most significant
term in the error introduced by a first-order accurate formula is proportional to the
sampling period. For instance, if the sampling period is decreased by half, the error reduces
by half.
A similar analysis can be performed for evaluation of the error of the backward formula
starting with the Taylor series expansion of f ðx  DxÞ:

ðDxÞ2 00 ðDxÞ3 000 ðDxÞ4 0000

f ðx  DxÞ ¼ f ðxÞ  Dxf 0 ðxÞ þ f ðxÞ  f ðxÞ þ f ðxÞ þ ⋯: ð1:12Þ
2 6 24
 1.2 ● Approximation of derivatives by finite differences 7

This equation can be rearranged to express f 0 ðxÞ as

f ðxÞ  f ðx  DxÞ Dx 00 ðDxÞ2 000 ðDxÞ3 0000

f 0 ðxÞ ¼ þ f ðxÞ  f ðxÞ þ f ðxÞ  ⋯: ð1:13Þ
Dx 2 6 24

The first term on the right-hand side of (1.13) is the same as the backward difference
formula and the sum of the rest of the terms represents the error introduced by the backward
difference formula to the exact derivative of f ðxÞ, such that

f ðxÞ  f ðx  DxÞ
f 0ðxÞ ¼ þ OðDxÞ: ð1:14Þ
Dx

The order of Dx in the most significant term of O(Dx) is one; hence the backward dif-
ference formula is first-order accurate.
The difference between the Taylor series expansions of f ðx þ DxÞ and f ðx  DxÞ can be
expressed using (1.9) and (1.12) as

2ðDxÞ3 000
f ðx þ DxÞ  f ðx  DxÞ ¼ 2Dxf 0 ðxÞ þ f ðxÞ þ ⋯: ð1:15Þ
6

This equation can be rearranged to express f 0 ðxÞ as

f ðx þ DxÞ  f ðx  DxÞ ðDxÞ2 000 f ðx þ DxÞ  f ðx  DxÞ  

f 0 ðxÞ ¼  f ðxÞ þ ⋯ ¼ þ O ðDxÞ2 ;
2Dx 6 2Dx
ð1:16Þ

where the first term on right-hand side is the same as the central difference formula given in
(1.8) and the order of Dx in the most significant term of the error O((Dx)2) is two; hence the
central difference formula is second-order accurate. The interpretation of second-order
accuracy is that the most significant term in the error introduced by a second-order accurate
formula is proportional to the square of sampling period. For instance, if the sampling period
is decreased by half, the error is reduced by a factor of four. Hence, a second-order accurate
formula such as the central difference formula is more accurate than a first-order accurate
formula.
For an example, consider a function f ðxÞ ¼ sin(x)e0.3x as displayed in Figure 1.2(a).
The exact first-order derivative of this function is

f 0ðxÞ ¼ cosðxÞe0:3x  0:3 sinðxÞe0:3x :

This function f ðxÞ is sampled with a sampling period Dx ¼ p/5, and approximate deri-
vatives are calculated for f ðxÞ using the forward difference, backward difference, and central
difference formulas. The derivative of the function f 0 ðxÞ and its finite difference approx-
imations are plotted in Figure 1.2(b). The errors introduced by the difference formulas,
which are the differences between f 0 ðxÞ and its finite difference approximations, are plotted
in Figure 1.2(c) for the sampling interval Dx ¼ p/5. It is evident that the error introduced by
the central difference formula is smaller than the errors introduced by the forward difference
8 CHAPTER 1 ● Introduction to FDTD

0.8

0.6

0.4

0.2
f (x)

–0.2

–0.4

–0.6

–0.8

–1
0 5 10 15
(a) x

1
exact
0.8 forward difference
backward difference
0.6 Δ x = p /5
central difference

0.4

0.2
f ′ (x)

–0.2

–0.4

–0.6

–0.8

–1
0 5 10 15
(b) x

Figure 1.2 (a) f ðxÞ, f 0 ðxÞ, and differences between f 0 ðxÞ and finite difference approximations of
f 0 ðxÞ for Dx ¼ p/5 and Dx ¼ p/10: f ðxÞ ¼ sin(x)e0.3x; (b) f ðxÞ, f 0 ðxÞ, and differences between
f 0 ðxÞ and finite difference approximations of f 0 ðxÞ for Dx ¼ p/5 and Dx ¼ p/10: f 0 ðxÞ ¼
cos(x)e0.3x – 0.3 sin(x)e0.3x and finite difference approximations of f 0 ðxÞ for Dx ¼ p/5;
(c) f ðxÞ, f 0 ðxÞ, and differences between f 0 ðxÞ and finite difference approximations of f 0 ðxÞ
for Dx ¼ p/5 and Dx ¼ p/10: error(f 0 ðxÞ) for Dx ¼ p/5; (d) f ðxÞ, f 0 ðxÞ, and differences between
f 0 ðxÞ and finite difference approximations of f 0 ðxÞ for Dx ¼ p/5 and Dx ¼ p/10: error(f 0 ðxÞ)
for Dx ¼ p/10.
 1.2 ● Approximation of derivatives by finite differences 9

0.2
forward difference
0.15 Δ x = p /5 backward difference
central difference
0.1

0.05
Error [ f ' (x)]

–0.05

–0.1

–0.15

–0.2
0 5 10 15
(c) x
0.2
forward difference
0.15 Δ x = p /10 backward difference
central difference
0.1

0.05
Error [ f ' (x)]

–0.05

–0.1

–0.15

–0.2
0 5 10 15
(d) x

Figure 1.2 (Continued )

and backward difference formulas. Furthermore, the errors introduced by the difference
formulas for the sampling period Dx ¼ p/10 are plotted in Figure 1.2(d). It can be realized
that as the sampling period is halved, the errors of the forward difference and backward
difference formulas are halved as well, and the error of the central difference formula is
reduced by a factor of four.
The MATLAB code calculating f ðxÞ and its finite difference derivatives, and generat-
ing the plots in Figure 1.2 is shown in Listing 1.1.
10 CHAPTER 1 ● Introduction to FDTD

Listing 1.1 MATLAB code generating Figure 1.2(a–d)

1.2 ● Approximation of derivatives by finite differences 11
12 CHAPTER 1 ● Introduction to FDTD

Values of the function f ðxÞ at two neighboring points around x have been used to obtain a
second-order accurate central difference formula to approximate f 0(x). It is possible to obtain
formulas with higher orders of accuracy by including a larger number of neighboring
points in the derivation of a formula for f 0 ðxÞ. However, although there are FDTD
formulations developed based on higher-order accurate formulas, the conventional FDTD is
based on the second-order accurate central difference formula, which is found to be suffi-
ciently accurate for most electromagnetics applications and simple in implementation and
understanding.
It is possible to obtain finite-difference formulas for approximating higher-order
derivatives as well. For instance, if we take the sum of the Taylor series expansions of
f (x þ Dx) and f (x – Dx) using (1.9) and (1.12), we obtain

ðDxÞ4 0000
f ðx þ DxÞ þ f ðx  DxÞ ¼ 2f ðxÞ þ ðDxÞ2 f 00 ðxÞ þ f ðxÞ þ ⋯: ð1:17Þ
12

After rearranging the equation to have f 00(x) on the left-hand side, we get

f ðx þ DxÞ  2f ðxÞ þ f ðx  DxÞ ðDxÞ2 0000

f 00 ðxÞ ¼ 2
 f ðxÞ þ ⋯
ðDxÞ 12

f ðx þ DxÞ  2f ðxÞ þ f ðx  DxÞ

¼ 2
þ O ðDxÞ2 : ð1:18Þ
ðDxÞ

Using (1.18) we can obtain a central difference formula for the second-order derivative
f 00 (x) as

f ðx þ DxÞ  2f ðxÞ þ f ðx  DxÞ

f 00 ðxÞ  ; ð1:19Þ
ðDxÞ2

which is second-order accurate due to O(Dx)2.

Similarly, some other finite difference formulas can be obtained for the first- and
second-order derivatives with different orders of accuracy based on different sampling
points. A list of finite difference formulas is given for the first- and second-order derivatives
in Table 1.1 as a reference (Figure 1.3).
 1.3 ● FDTD updating equations for three-dimensional problems 13

Table 1.1 Finite difference formulas for the first- and second-order derivatives where the function
f with the subscript i is an abbreviation of f ðxÞ. Similar notation can be implemented
for f (x þ Dx), f (x þ 2Dx), etc., as shown in Figure 1.3. FD, forward difference; BD,
backward difference; CD, central difference.
2
Derivative ∂f=∂x Derivative ∂ f=∂x2

Difference scheme Type error Difference scheme Type error

fiþ1  fi FD OðD xÞ fiþ2  2fiþ1 þ fi FD OðD xÞ
Dx ðDxÞ2
fi  fi1 BD OðD xÞ fiþ2  2fi1 þ fi2 BD OðD xÞ
Dx ðDxÞ2
fiþ1  fi1 CD OððD xÞ2 Þ fiþ1  2fi þ fi1 CD OððD xÞ2 Þ
2D x ðDxÞ2
fiþ2 þ 4fiþ1  3fi FD OððD xÞ2 Þ CD OððD xÞ4 Þ
fiþ2 þ 16fiþ1  30fi þ 16fi1  fi2
2D x
12ðDxÞ2
3fiþ1  4fi1 þ fi2 BD OððD xÞ2 Þ
2D x
CD OððD xÞ4 Þ
fiþ2 þ 8fiþ1  8fi1 þ fi2
12D x

fi−2 fi−1 fi fi+1 fi+2

x − 2Δx x − Δx x x + Δx x + 2Δx

Figure 1.3 Sample points of f ðxÞ.

1.3 FDTD updating equations for three-dimensional

problems
In 1966, Yee originated a set of finite-difference equations for the time-dependent
Maxwell’s curl equations system [2]. These equations can be represented in discrete form,
both in space and time, employing the second-order accurate central difference formula. As
mentioned before, the electric and magnetic field components are sampled at discrete posi-
tions both in time and space. The FDTD technique divides the three-dimensional problem
geometry into cells to form a grid. Figure 1.4 illustrates an FDTD grid composed of
(Nx  Ny  Nz) cells. A unit cell of this grid is called a Yee cell. Using rectangular Yee
cells, a stepped or ‘‘staircase’’ approximation of the surface and internal geometry of the
structure of interest is made with a space resolution set by the size of the unit cell.
The discrete spatial positions of the field components have a specific arrangement in the
Yee cell, as demonstrated in Figure 1.5. The electric field vector components are placed at
14 CHAPTER 1 ● Introduction to FDTD

(Nx, Ny, Nz)

(i, j, k)

y (1, 1, 1)
x

Figure 1.4 A three-dimensional FDTD computational space composed of (Nx  Ny  Nz)

Yee cells.

Δx
node(i + 1, j + 1, k + 1)

Δz

Ez(i, j, k) Hz(i, j, k) Hy(i, j, k)

Hz(i, j, k)
Ey(i, j, k)
node(i, j, k) Δy
z Ex(i, j, k)
y
x

Figure 1.5 Arrangement of field components on a Yee cell indexed as (i, j, k).

the centers of the edges of the Yee cells and oriented parallel to the respective edges, and the
magnetic field vector components are placed at the centers of the faces of the Yee cells and
are oriented normal to the respective faces. This provides a simple picture of three-
dimensional space being filled by an interlinked array of Faraday’s law and Ampere’s law
contours. It can be easily noticed in Figure 1.5 that each magnetic field vector is surrounded
 1.3 ● FDTD updating equations for three-dimensional problems 15

by four electric field vectors that are curling around the magnetic field vector, thus simu-
lating Faraday’s law. Similarly, if the neighboring cells are also added to the picture, it
would be apparent that each electric field vector is surrounded by four magnetic field vectors
that are curling around the electric field vector, thus simulating Ampere’s law.
Figure 1.5 shows the indices of the field components, which are indexed as (i, j, k),
associated with a cell indexed as (i, j, k). For a computational domain composed of uniform
Yee cells having dimension Dx in the x direction, Dy in the y direction, and Dz in the z
direction, the actual positions of the field components with respect to an origin coinciding
with the position of the node (1, 1, 1) can easily be calculated as

Ex ði; j; kÞ ) ðði  0:5ÞDx; ðj  1ÞDy; ðk  1ÞDzÞ;

Ey ði; j; kÞ ) ðði  1ÞDx; ðj  0:5ÞDy; ðk  1ÞDzÞ;
Ez ði; j; kÞ ) ðði  1ÞDx; ðj  1ÞDy; ðk  0:5ÞDzÞ;
Hx ði; j; kÞ ) ðði  1ÞDx; ðj  0:5ÞDy; ðk  0:5ÞDzÞ;
Hy ði; j; kÞ ) ðði  0:5ÞDx; ðj  1ÞDy; ðk  0:5ÞDzÞ;
Hz ði; j; kÞ ) ðði  0:5ÞDx; ðj  0:5ÞDy; ðk  1ÞDzÞ:

The FDTD algorithm samples and calculates the fields at discrete time instants; how-
ever, the electric and magnetic field components are not sampled at the same time instants.
For a time-sampling period Dt, the electric field components are sampled at time instants
0, Dt, 2Dt, . . . , nDt, . . . ; however, the magnetic field components are sampled at time
instants 12 Dt; ð1 þ 12ÞDt; . . . ; ðn þ 12ÞDt; . . . . Therefore, the electric field components are
calculated at integer time steps, and magnetic field components are calculated at half-integer
time steps, and they are offset from each other by Dt/2. The field components need to be
referred not only by their spatial indices which indicate their positions in space, but also by
their temporal indices, which indicate their time instants. Therefore, a superscript notation is
adopted to indicate the time instant. For instance, the z component of an electric field vector
positioned at ((i – 1)Dx, ( j – 1)Dy, (k – 0.5)Dz) and sampled at time instant nDt is referred to
as Ezn ði; j; kÞ: Similarly, the y component of a magnetic field vector positioned at ((i – 0.5)Dx,
( j – 1)Dy, (k – 0.5)Dz) and sampled at time instant ðn þ 12ÞDt is referred to as Hynþ1=2 ði; j; kÞ:
The material parameters (permittivity, permeability, electric, and magnetic con-
ductivities) are distributed over the FDTD grid and are associated with field components;
therefore, they are indexed the same as their respective field components. For instance,
Figure 1.6 illustrates the indices for the permittivity and permeability parameters.
The electric conductivity is distributed and indexed the same as the permittivity, and the
magnetic conductivity is distributed and indexed the same as the permeability.
Having adopted an indexing scheme for the discrete samples of field components in both
time and space, Maxwell’s curl equations (1.4) that are given in scalar form can be expressed
in terms of finite differences. For instance, consider again (1.4a):
 
@Ex 1 @Hz @Hy
¼   sx Ex  J ix :
e
@t ex @y @z
16 CHAPTER 1 ● Introduction to FDTD

ex(i, j + 1, k + 1)

ey(i, j, k + 1)
mz(i, j, k + 1)
ey(i + 1, j, k + 1)
ex(i, j, k + 1)
uy(i, j + 1, k)
ez(i, j + 1, k) e (i + 1, j + 1, k)
mx(i + 1, j, k) z
ez(i, j, k) mz(i, j, k)
my(i, j, k)
ez(i + 1, j, k)
ex(i, j + 1, k)

ey(i, j, k) mz(i, j, k)
ey(i + 1, j, k)
node(i, j, k)
z ex(i, j, k)

y
x

Figure 1.6 Material parameters indexed on a Yee cell.

The derivatives in this equation can be approximated by using the central difference
formula with the position of Ex(i, j, k) being the center point for the central difference
formula in space and time instant ðn þ 12ÞDt as being the center point in time. Considering the
field component positions given in Figure 1.7, we can write
nþ12 nþ12
Exnþ1 ði; j; kÞ  Exn ði; j; kÞ 1 Hz ði; j; kÞ  Hz ði; j  1; kÞ
¼
Dt ex ði; j; kÞ Dy
nþ1 nþ1
1 Hy 2 ði; j; kÞ  Hy 2 ði; j; k  1Þ

ex ði; j; kÞ Dz
sex ði; j; kÞ nþ12 1 nþ1
 Ex ði; j; kÞ  J ix 2 ði; j; kÞ: ð1:20Þ
ex ði; j; kÞ ex ði; j; kÞ

It has already been mentioned that the electric field components are defined at integer
time steps; however, the right-hand side of (1.20) includes an electric field term at time
instant ðn þ 12ÞDt, that is, Exnþ1=2 ði; j; kÞ. This term can be written as the average of the terms
at time instants (n þ 1)Dt and nDt, such that

nþ1 Exnþ1 ði; j; kÞ þ Exn ði; j; kÞ

Ex 2 ði; j; kÞ ¼ : ð1:21Þ
2

Using (1.21) in (1.20) and arranging the terms such that the future term Exnþ1 ði; j; kÞ is
kept on the left side of the equation and the rest of the terms are moved to the right-hand side
of the equation, we can write
 1.3 ● FDTD updating equations for three-dimensional problems 17

z
Δx
y
x node(i + 1, j + 1, k + 1)
Δy

Hy(i, j, k)
Δz

node(i, j, k) Hz(i, j, k)
node(i, j − 1, k) Hz(i, j − 1, k)
Ex(i, j, k)

Hy(i, j, k − 1)

node(i, j, k − 1)

Figure 1.7 Field components around Ex(i, j, k).

2ex ði; j; kÞ þ Dtsex ði; j; kÞ nþ1 2ex ði; j; kÞ  Dtsex ði; j; kÞ n

Ex ði; j; kÞ ¼ Ex ði; j; kÞ
2ex ði; j; kÞ 2ex ði; j; kÞ

Dt  
nþ1 nþ1
þ Hz 2 ði; j; kÞ  Hz 2 ði; j  1; kÞ
ex ði; j; kÞDy

Dt  
nþ1 nþ1
 Hy 2 ði; j; kÞ  Hy 2 ði; j; k  1Þ
ex ði; j; kÞDz

Dt nþ1
 J ix 2 ði; j; kÞ
ex ði; j; kÞ
ð1:22Þ

After some manipulations, we get

2ex ði; j; kÞ  Dtsex ði; j; kÞ n

Exnþ1 ði; j; kÞ ¼ E ði; j; kÞ
2ex ði; j; kÞ þ Dtsex ði; j; kÞ x

2Dt  
nþ12 n þ12
þ
Hz ði; j; kÞ  Hz ði; j  1; kÞ
2ex ði; j; kÞ þ Dtsex ði; j; kÞ Dy

2Dt  
nþ1 nþ1

Hy 2 ði; j; kÞ  Hy 2 ði; j; k  1Þ
2ex ði; j; kÞ þ Dtsx ði; j; kÞ Dz
e

2Dt nþ1 ð1:23Þ

 J ix 2 ði; j; kÞ:
2ex ði; j; kÞ þ Dtsx ði; j; kÞ
e
18 CHAPTER 1 ● Introduction to FDTD

y
x
Ey(i, j, k + 1)
Δx
node(i, j, k + 1)
Δy

Ez(i, j + 1, k)
Ez(i, j, k)
Hx(i, j, k)
Δz

node(i, j + 1, k)

node(i, j, k) Ey(i, j, k)

Figure 1.8 Field components around Hx(i, j, k).

The form of (1.23) demonstrates how the future value of an electric field component can
be calculated by using the past values of the electric field component, the magnetic field
components, and the source components. This form of an equation is called an FDTD
updating equation. Updating equations can easily be obtained for calculating Eynþ1 ði; j; kÞ
starting from (1.4b) and Eznþ1 ði; j; kÞ starting from (1.4c) following the same methodology
that has been used to obtain (1.23).
Similarly, updating equations can be obtained for magnetic field components following
the same methodology. However, while applying the central difference formula to the time
derivative of the magnetic field components, the central point in time shall be taken as nDt.
For instance, (1.4c), which is
 
@Hx 1 @Ey @Ez
¼   sx Hx  Mix ;
m
@t mx @z @y

can be approximated using finite differences based on the field positions (as shown in
Figure 1.8) as

nþ1 n1
Hx 2 ði; j; kÞ  Hx 2 ði; j; kÞ 1 Eyn ði; j; k þ 1Þ  Eyn ði; j; kÞ
¼
Dt mx ði; j; kÞ Dz

1 Ezn ði; j þ 1; kÞ  Ezn ði; j; kÞ


mx ði; j; kÞ Dy

sm
x ði; j; kÞ n 1
 H ði; j; kÞ  M n ði; j; kÞ: ð1:24Þ
mx ði; j; kÞ x mx ði; j; kÞ ix
 1.3 ● FDTD updating equations for three-dimensional problems 19

nþ1
After some manipulations, the future term Hx 2 ði; j; kÞ in (1.24) can be moved to the left-
hand side and the other terms can be moved to the right-hand side such that

nþ1 2mx ði; j; kÞ  Dtsm x ði; j; kÞ n2

1
Hx 2 ði; j; kÞ ¼ Hx ði; j; kÞ
2mx ði; j; kÞ þ Dtsx ði; j; kÞ
m

2Dt  
þ
E n
ði; j; k þ 1Þ  E n
ði; j; kÞ
2mx ði; j; kÞ þ Dtsm x ði; j; kÞ Dz
y y

2Dt


Ezn ði; j þ 1; kÞ  Ezn ði; j; kÞ
2mx ði; j; kÞ þ Dtsx ði; j; kÞ Dy
m

2Dt
 Mixn ði; j; kÞ: ð1:25Þ
2mx ði; j; kÞ þ Dtsm
x ði; j; kÞ
nþ1
This equation is the updating equation for Hx 2 ði; j; kÞ. Similarly, updating equations can
nþ1 nþ1
easily be obtained for Hy 2 ði; j; kÞ starting from (1.4e) and Hz 2 ði; j; kÞ starting from (1.4f)
following the same methodology used to obtain (1.25).
Finally, (1.4a)–(1.4f) can be expressed using finite differences and can be arranged to
construct the following six FDTD updating equations for the six components of electro-
magnetic fields by introduction of respective coefficient terms:

Exnþ1 ði; j; kÞ ¼ Cexe ði; j; kÞ  Exn ði; j; kÞ

 
nþ1 nþ1
þ Cexhz ði; j; kÞ  Hz 2 ði; j; kÞ  Hz 2 ði; j  1; kÞ
 
nþ1 nþ1
þ Cexhy ði; j; kÞ  Hy 2 ði; j; kÞ  Hy 2 ði; j; k  1Þ
nþ1
þ Cexj ði; j; kÞ  J ix 2 ði; j; kÞ; ð1:26Þ
where
2ex ði; j; kÞ  Dtsex ði; j; kÞ
Cexe ði; j; kÞ ¼ ;
2ex ði; j; kÞ þ Dtsex ði; j; kÞ
2Dt
Cexhz ði; j; kÞ ¼
;
2ex ði; j; kÞ þ Dtsex ði; j; kÞ Dy
ð1:27Þ
2Dt
Cexhy ði; j; kÞ ¼ 
;
2ex ði; j; kÞ þ Dtsex ði; j; kÞ Dz
2Dt
Cexj ði; j; kÞ ¼  :
2ex ði; j; kÞ þ Dtsex ði; j; kÞ

Eynþ1 ði; j; kÞ ¼ Ceye ði; j; kÞ  Eyn ði; j; kÞ

 
nþ1 nþ1
þ Ceyhx ði; j; kÞ  Hx 2 ði; j; kÞ  Hx 2 ði; j; k  1Þ
 
nþ1 nþ1
þ Ceyhz ði; j; kÞ  Hz 2 ði; j; kÞ  Hz 2 ði  1; j; kÞ
nþ1
þ Ceyj ði; j; kÞ  J iy 2 ði; j; kÞ;
20 CHAPTER 1 ● Introduction to FDTD

where

2ey ði; j; kÞ  Dtsey ði; j; kÞ

Ceye ði; j; kÞ ¼ ;
2ey ði; j; kÞ þ Dtsey ði; j; kÞ

2Dt
Ceyhx ði; j; kÞ ¼   ;
2ey ði; j; kÞ þ Dtsey ði; j; kÞ Dz

2Dt
Ceyhz ði; j; kÞ ¼    ;
2ey ði; j; kÞ þ Dtsey ði; j; kÞ Dx

2Dt ð1:28Þ
Ceyj ði; j; kÞ ¼  :
2ey ði; j; kÞ þ Dtsey ði; j; kÞ

Eznþ1 ði; j; kÞ ¼ Ceze ði; j; kÞ  Ezn ði; j; kÞ

 
nþ1 nþ1
þ Cezhy ði; j; kÞ  Hy 2 ði; j; kÞ  Hy 2 ði  1; j; kÞ
 
nþ1 nþ1
þ Cezhx ði; j; kÞ  Hx 2 ði; j; kÞ  Hx 2 ði; j  1; kÞ

nþ1
þ Cezj ði; j; kÞ  J iz 2 ði; j; kÞ;

where

2ez ði; j; kÞ  Dtsez ði; j; kÞ

Ceze ði; j; kÞ ¼ ;
2ez ði; j; kÞ þ Dtsez ði; j; kÞ

2Dt
Cezhy ði; j; kÞ ¼
;
2ez ði; j; kÞ þ Dtsez ði; j; kÞ Dx

2Dt
Cezhx ði; j; kÞ ¼ 
;
2ez ði; j; kÞ þ Dtsez ði; j; kÞ Dy

2Dt ð1:29Þ
Cezj ði; j; kÞ ¼  :
2ez ði; j; kÞ þ Dtsez ði; j; kÞ
n þ12 n1
Hx ði; j; kÞ ¼ Chxh ði; j; kÞ  Hx 2 ði; j; kÞ
 
þ Chxey ði; j; kÞ  Eyn ði; j; k þ 1Þ  Eyn ði; j; kÞ


þ Chxez ði; j; kÞ  Ezn ði; j þ 1; kÞ  Ezn ði; j; kÞ

þ Chxm ði; j; kÞ  Mixn ði; j; kÞ;

 1.3 ● FDTD updating equations for three-dimensional problems 21

where
2mx ði; j; kÞ  Dtsmx ði; j; kÞ
Chxh ði; j; kÞ ¼ ;
2mx ði; j; kÞ þ Dtsmx ði; j; kÞ
2Dt
Chxey ði; j; kÞ ¼
;
2mx ði; j; kÞ þ Dtsm x ði; j; kÞ Dz
2Dt
Chxez ði; j; kÞ ¼ 
;
2mx ði; j; kÞ þ Dtsm
x ði; j; kÞ Dy ð1:30Þ
2Dt
Chxm ði; j; kÞ ¼  :
2mx ði; j; kÞ þ Dtsm
x ði; j; kÞ
n þ12 n1
Hy ði; j; kÞ ¼ Chyh ði; j; kÞ  Hy 2 ði; j; kÞ þ Chyez ði; j; kÞ


 Ezn ði þ 1; j; kÞ  Ezn ði; j; kÞ þ Chyex ði; j; kÞ


 Exn ði; j; k þ 1Þ  Exn ði; j; kÞ þ Chym ði; j; kÞ  Miyn ði; j; kÞ
where
2my ði; j; kÞ  Dtsm
y ði; j; kÞ
Chyh ði; j; kÞ ¼ ;
2my ði; j; kÞ þ Dtsm
y ði; j; kÞ
2Dt
Chyez ði; j; kÞ ¼   ;
2my ði; j; kÞ þ Dtsm
y ði; j; kÞ Dx
2Dt
Cbyex ði; j; kÞ ¼    ;
2my ði; j; kÞ þ Dtsm
y ði; j; kÞ Dz
ð1:31Þ
2Dt
Chym ði; j; kÞ ¼  :
2my ði; j; kÞ þ Dtsm
y ði; j; kÞ
nþ12 n12
Hz ði; j; kÞ ¼ Chzh ði; j; kÞ  Hz ði; j; kÞ


þ Chzex ði; j; kÞ  Exn ði; j þ 1; kÞ  Exn ði; j; kÞ
 
þ Chzey ði; j; kÞ  Eyn ði þ 1; j; kÞ  Eyn ði; j; kÞ
þ Chzm ði; j; kÞ  Mizn ði; j; kÞ;
where
2my ði; j; kÞ  Dtsmz ði; j; kÞ
Chzh ði; j; kÞ ¼ ;
2mz ði; j; kÞ þ Dtsz ði; j; kÞ
m

2Dt
Chzex ði; j; kÞ ¼
;
2mz ði; j; kÞ þ Dtsmz ði; j; kÞ Dy
2Dt
Chzey ði; j; kÞ ¼ 
;
2mz ði; j; kÞ þ Dtsm
z ði; j; kÞ Dx
2Dt
Chzm ði; j; kÞ ¼  :
2mz ði; j; kÞ þ Dtsm
z ði; j; kÞ
Exploring the Variety of Random
Documents with Different Content
Purslow Hundred, 696
Queen Anne’s Bounty, 25
Quatford, 652
Quatt, 654
— Jarvis, 654
— Malvern, 654
Quinta, The, 161
Ragdon, 539
Railways, 21
Ratlinghope, 704
Redcastle Hill, 288
Rednal, 204
Reilth, 714
Rhiston, 691
Rhos Goch, 694
Rhuddleford, 495
Richards Castle, 543
Ridge Higher, 239
— Lower, 239
Rindleford, 495
Ritton, 706
Rivers, 19
Roads, 21
Rock, 549
Rodington, 417
Roden, 390
Rodney’s Pillar, 672
Roman Invasion, 9
Romsley, 611
Roowood, 336
Rorrington, 689
Rossal, 674
Roughton, 495
Round Acton, 610
Rowley, 495
Rowton, 390
Rowton, 672
— (Stokesay), 553
Royal Oak, 463
Ruckley, 499
Rudge, 656
Rugantine, 712
Rushbury, 545
Roman Stations, 449, 545, 518, 654, 671, 694
Rushton, 451
Rushmore, 446
Ruthall, 590
Ruyton-of-the-Eleven-Towns, 196
Ryton, 472, 503
Sambrook, 261
Sandford, 204
— (Prees), 307
Sascott, 682
Saxon Gods, 14
Scrimage, 531
Selattyn, 199
Selley, 713
Severn, The, 19
Shadwell, 710
Shavington, 258
Shawbury, 309
Sheet, 540
Sheinton, 657
Shelbrook, 234
Shelderton, 700
Shelton & Oxon, 686
Shelve, 691
Shelvock, 198
Sheriff Hales, 397
Sherlowe, 385
Shiffnal, 473
Shineton, 657
Shipley, 469
Shipton, 547
Shotton, 144
Shooters Hill, 141
Shotatton, 199
Shrawardine, 254
— Little, 671
Shrewsbury, from 33 to 132
— Abbey, 49 and 72
— Abbots of, 74
— Almshouses, 85
— Annals, 79
— Anct. Mansions, 78
— Antiquarian and Nat. His. Society, 64
— Aquatic Excur., 72
— Assembly Rooms, 72
— Asylum, 66
— Barons of, 36
— Battle of, 36 & 133
— Baths Royal, 66
— Billiard Rooms, 72
— Bridges, 68
— Canal, 65
— Cattle Market, 68
— Chapels Ancient, 55
— Chapels Dissent, 55
— Charities, 82 to 93
— Charters, 44
— Coleham, 93
— Corporation, 42
— Council House, 78
— County Constab., 43
— County and Town Gaol, 63
— County Hall, 62
— Drapers Hall, 70
— Directory, 95
— Dispensary, 62
— Early Gov. of, 41
— Earls of, 35, 73
— Eye & Ear Dispensary, 62
— Frankwell, 93
— Fairs, 68
— Floods, 81
— Gaol, 63
— Gas Works, 67
— Gates & Posterns, 77
— Glass Staining, 71
— Great Parlia., 36
— Holy Cross and St. Giles, 93
— Hospital, St. Giles, 51
— Hill’s Mansion, 79
— House of Indus., 65
— House of Correc., 65
— Infirmary, 61
— Ireland’s Mansion, 79
— Jones’s Mansion, 79
— Kingsland, 72
— Library Subscription, 65
— Lord Hill’s Column, 67
— Markets, 68
— Market Hall, 63
— Market House, 63
— Mechanics’ Institute, 65
— Meole Brace, 93
— Mercer’s Hall, 71
— Monastic Foundation, 72
— Monks of, 73
— Municipal Act, 42
— Music Hall, 64
— Newspapers, 64
— News Room, 65
— Parishes of, 92
— Parliament at, 36
— Population, 34
— Public Buildings, 61
— Quarry The, 71
— Races, 72
— Railway Station, 67
— Savings’ Bank, 66
— Schools, 57 to 61
— Severn River, 33
— Show, 71
— — Cakes, 71
— — Brawn, 71
— Simnell Cake, 71
— Streets, 94
— Subscrip. Library, 65
— Tailors’ Hall, 71
— Theatre, 64
— Town Hall, 62
— Town Walls, 77
— Trade, 69
— Trade Directy., 115
— Water Works, 66
— St. Alkmund’s Parish, 92
— St. Chad’s Parish, 93
— St. Julian’s Parish, 93
— St. Mary’s Parish, 93
Shropshire Giant, 240
Siberscott, 682
Sibdon Carwood, 705
— Castle, 705
Sidbury, 657
Siefton, 532
Silvington, 670
Skeletons, 157
Skyborry, 714
Sleap, (Ercall), 388
Sleap, (Wem), 335
Smethcott, 144
Smethcott, 513
Snailbeach Mine, 678
Snedshill, 477
— Ironworks and Collieries, 477
Snitton, 665
Sodylt Hall, 234
Soil and Produce, 22
Sowdley Great, 262
Soulton, 335
South Bradford Hundred, 364
Spoad, 710
Spoonhill, 239
Spoonley, 258
Spray Hill, 384
Stableford, 495
St. Almund’s, 92
St. Chad’s, 93
St. Paul’s, Knowbury, 635
Stanford, 672
Stanmore, 495
Stanton-upon-Hine Heath, 314
— Lacy, 548
— Long, 550
— Shiffnal, 476
Stanwardine-in-the-Fields, 217
— in-the Woods, 217
Stanway, 547
Stapleton, 515
Steele, 307
Stiperstone Hill, 507
Stirchley, 418
— Hall, 418
— Ironworks, 418
St. Julian’s, 93
St. Martin’s, 158
St. Bryngwyla School, 159
St. Mary’s, 93
St. Winefred’s Well, 206
Stitt and Gatten, 705
Stocks and Coptiviney, 239
Stockett, 236
Stockton, 484
— Park, 485
Stockton-by-Newport, 400
Stockton-by-Chirbury, 689
Stoke-by-Burford, 667
Stoke-upon-Terne, 312
Stoke, St. Milborough, 550
Stoke, Say, 552
Stone Acton, 547
Stottesden, 657
— Hundred, 610
Stowe, 705
Strefford, 707
Stretton, 685
— All, 530
— Church, 524
— Little, 530
Styche & Woodlands, 297
Sundorne Castle, 138
Sugdon, 407
Sutherland, 1st Duke of, 394
Sutton, (Claverley) 469
— (Drayton), 278
Sutton-by-Chelmarsh, 636
Sutton-by-Shrewsbury, 515
— Spa, 516
Sutton-by-West Felton, 204
— Maddock, 486
— Little, 534
— Great, 535
— Court, 534
Swancote, 495
Swerney, 193
— Hall, 194
Sychtyn, 201
Sylattin, 199
Tan-coed-y-gaer, 193
Talbot John, 338 & 357
Tasley, 662
Tedsmere, 205
Tern, 392
— House, 392
Tetchill, 239
Thanes, 521
Thoughlands, 542
Ticklerton, 537
Tibberton, 384
Tilley, 336
— Green, 336
Tilsop, 667
Tilstock, 362
Timberth, 689
Tir-y-coed, 152
Tobacco Pipes Manufactory, 556
Tonge, 486
— Castle, 487
Totterton, 703
Trebert, 714
Trebrodier, 712
Trefarclawdd, 194
Treflach, 194
Trefnant, 672
Trefonnen, 195
Trelystan, 694
Trench, 240
Trench-by-Wem, 336
— Lane, 448
Treprenal, 157
Treverward, 710
Triptych, 666
Tugford, 553
Twyford, 205
Tylsoer Dr., 343
Tyn-y-rhos, 162
Uckington, 367
Uffington, 145
Uppington, 418
Uppington, 672
Upton Cresset, 662
— Magna, 419
— Parva, or Waters Upton, 421
Vennington, 685
Wackley Lodge, 232
Walcot-by-Chirbury, 689
Walcot-by-Wellington, 439
Walcot Hall, 702
Walford, 218
Walker’s Lowe, 661
Wallop, 685
Wall-under-Haywood, 547
Walton-by-Ercall, 392
Walton-by-Onibury, 543
Walton-by-Wenlock, 588
Walton-by-Worthen, 695
Walton-by-Stottesden, 661
Wappenshall, 439
Wars, 9
Waters Upton, 421
Watling Street, 426
Watts Dyke, 210
Wattlesborough, 672
Wellington, 421
— Fairs, 422
— Gas Works, 424
— History, 422 to 425
— Horticultural Society, 425
— Market Hall, 423
— News Room, 424
— Old Hall, 425
— Schools, 423
— Streets, 427
Welsh Frankton, 211
— Hampton, 255
Wem, 317
Wenlock Much, 579
— Edge, 589
— Franchise, 554
— Little, 565
Wentnor, 705
Westbury, 684
West Felton, 202
— Foreign Libty., 641
— Hamlet, 549
Westhope, 535
Westley, 503
Westley, 685
Weston-by-Clun, 709
Weston-by-Burford, 667
Weston Cotton, 195
Weston-by-Hopton, 579
Weston Lullingfield, 218
—Rhyn, 161
— Coalworks, 161
— Under Red Castle, 290
— Stowe, 705
Wettleton, 553
Whattall, 236
Wheathill, 663
Wheathall, 503
Wheel Green, 496
Whetmore, 667
Whitchurch, 337
Whitcott & Hardwick, 704
Whitcott Evan, 710
Whitcott Keysett, 711
White Ladies, 463
Whitley, 676
Whittington, 207
Whitton-by-Westbury, 685
Whitton-by-Burford, 667
Wicherley Hall, 218
Whixall, 307
Whigmore, 685
Whigwig, 589
Wikey, 199
Wilcott, 244
Wilderhope, 547
Wilderley, 508
Willaston, 308
Willey, 591
Wilmington, 689
Willstone, 524
Winnington, 672
Winsbury, 689
Winscote, 496
Winsley, 685
Wirswall, 364
Wistanstow, 706
Wistanswick, 373
Withington, 440
Wittingslow, 707
Wixhall, 291
Wollascott, 142
Wollaston, 672
Wollerton, 291
Wolf’s Head, 242
Wolverley, 336
Wombridge, 440
— Priory, 441
Woodbatch, 698
Woodcote, 442
Woodcote-by-St. Chad’s, 676
Woodhall, 681
Woodhouse, 477
Woodhouse, 204
Woodhouses New, 363
Woodhouses Old, 363
Woodlands, 297
Woodseaves, 278
Woodside, 477
Woofferton, 545
Woolstaston, 516
Woolston, 206
Woolston, 707
Woore, 298
Wooton, 196
Wootton, 549
Worfield, 491
Worthen, 692
Wotherton, 689
Woundale, 469
Wrentnall, 508
Wrickton, 661
Wrockwardine, 443
— Wood, 447
Wroxeter, 448
Wycherley The Poet, 141
Wyke, 476
Wyke-by-Wenlock, 558
Wyken, 496
Wykey, 199
Wytheford Magna, 312
Wytheford Parva, 312
Yeaton, 219
Yockleton, 685
Yorton, 134
 GENERAL HISTORY AND
DESCRIPTION OF SHROPSHIRE.

SHROPSHIRE is an inland county on the borders of Wales, bounded

on the north by Denbighshire, Cheshire, and a detached part of
Flintshire: on the east by Staffordshire: on the south by
Worcestershire, Herefordshire, and Radnorshire: and on the west by
Montgomery and Denbighshire. In length, from north to south, it is
about forty-five miles, and its extreme breadth thirty-five. Its
circumference is computed at 200 miles; and it comprises an area of
1,343 square statute miles, and, consequently, 859,520 acres. The
county, in Saxon annals, is called Scrobbesbyrig and Scrobbescire,
and by Latin authors, Comitates Salopiensis. It is one of the shires,
which, in the time of the Romans, was inhabited by the Cornavii,
whose province comprehended the counties of Cheshire, Salop,
Stafford, Warwickshire, and Worcestershire. At the census of 1801,
the county embraced a population of 167,639 souls: 1831, 222,800:
1841, 239,048, of whom 119,355 were males, and 119,693
females. At the same period, there were 47,208 inhabited houses,
2,086 uninhabited, and 293 houses building. The number of
persons born in the county in these returns was 203,689: in other
counties, 3,240: in Scotland, 391: in Ireland, 1,199: in the British
colonies, 14: foreigners in the county, 161: not specified where born,
1,144. Of the total population, 55,645 males, and 54,624 females,
were under 20 years of age: 12,189 were between sixty and seventy
years of age: 6,006 between seventy and eighty: 1,905 between
eighty and ninety: 139 between ninety and one hundred: and the
age of 5 persons exceeded one hundred years. The total population
of the fifteen unions, into which the county of Shropshire is divided,
at the census of 1851, are returned as containing 245,019
inhabitants, of whom 122,122 were males, and 122,997 females.
Shropshire is divided into the hundreds of Albrighton, Bradford,
Brimstree, Chirbury, Clun, Condover, Ford, Munslow, Oswestry,
Overs, Pimhill, Purslow, Stottesden, and Wenlock franchise, and
contains 224 parishes, and 5 extra-parochial places. By the recent
Reform and Division of Counties’ Acts, this county is divided into the
northern and southern divisions, each of which returns two members
to Parliament. The boroughs of Shrewsbury, Bridgnorth, Ludlow,
and Wenlock also return two members each. The expenditure of the
county for the year ending December, 1850, was £12,156. 17s.
4¼d., of which £3,587. 10s. 2d. was expended on the Gaol and
House of Correction; £2,257. 10s. 7d. in prosecutions; £605. 17s. 5d.
on bridges and roads; £562. 13s. 4d. on the Lunatic Asylum;
coroners, £501. 1s. 2d., and Clerk of the Peace, £436. 4s. 9d. Judge
Blackstone says:—England was first divided into counties, hundreds,
and tithings by Alfred the Great, for the protection of property and
the execution of justice. Tithings were so called because ten
freeholders formed one. Ten of these tithings were supposed to
form a hundred or wapentake, from an ancient ceremony, in which
the governor of a hundred met all the aldermen of his district, and
holding up his spear, they all touched it with theirs, in token of
subjection and union to one common interest. An indifferent
number of these wapentakes, or hundreds, form a county or shire,
for the civil government of which a shire-reeve or sheriff is elected
annually. The magistrate above the hundredry was called the
trithingman or lathgrieve, presided over three, four, or more,
hundreds, formed into what was called a trithing, in some places a
lathe, and in others a rape; hence the lathes of Kent, the rapes of
Sussex, the parts of Lincoln, and trithings or ridings of Yorkshire.
The kingdom was divided into parishes soon after the introduction of
Christianity, by Honorius, Archbishop of Canterbury, in the year 636,
and the boundaries of them, as marked in Doomsday book, agree
very nearly with the present division. The custom, which still
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookgate.com