Professional Documents
Culture Documents
An Introductory Course
Waves
Peter van den Berg (1943) was a member of the scientific staff of the
Laboratory of Electromagnetic Research, Delft University of
Technology, Delft, Netherlands. He graduated in 1971 from the Delft
University of Technology, and during the academic year 1973-1974 he
was Visiting Lecturer in the Department of Mathematics, University of
Dundee, Scotland. He was appointed as a full professor at the Laboratory
of Electromagnetic Research in 1981. Since 2003 he is a research
professor in the Faculty of Applied Sciences of the Delft University of
Technology. During his career, he has written an impressive amount of
papers on the numerical analysis of forward and inverse electromagnetic
wavefield problems, and he has been teaching various classes on
electromagnetic, acoustic and elastic waves.
Hans Blok (1935) was a member of the scientific staff of the Laboratory
of Electromagnetic Research, Delft University of Technology, Delft,
Netherlands. He graduated in 1970 from the Delft University of
Technology, and he was appointed full professor at the Laboratory of
Electromagnetic Research in 1980. He was dean of the Faculty of
Electrical Engineering, Delft University of Technology in the period
1980-1982. During the academic year 1983-1984 he was a visiting
scientist at Schlumberger-Doll Research, Ridgefield, Connecticut,
U.S.A. During his career, he has written a number of papers on
resonators and optical waveguides, and he has been teaching various
classes on electromagnetic waves and signal theory. He is
emeritus professor since 2000.
Electromagnetic Waves
— An Introductory Course
M.D. Verweij
P.M. van den Berg
H. Blok
VSSD
© VSSD
First edition 1999
Second edition 2001-2006
Published by:
VSSD
Leeghwaterstraat 42, 2628 CA Delft, The Netherlands
tel. +31 15 278 2124, telefax +31 15 278 7585, e-mail: hlf@vssd.nl
internet: http://www.vssd.nl/hlf
URL about this book: http://www.vssd.nl/hlf/e016.htm (a.o. to download the auxiliary
Repetitive Guide)
A collection of digital pictures and an elctronic version can be made available for
lecturers who adopt this book. Please send a request by e-mail to hlf@vssd.nl
All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system, or transmitted, in any form or by any means, electronic, mechanical, photo-
copying, recording, or otherwise, without the prior written permission of the publisher.
ISBN 90-407-1836-9 EAN 9789040718366
NUR 924
Keywords: electromagnetic waves
Preface
Preface v
1 Introduction 1
1.1. Cartesian vectors and their properties . . . . . . . . . . . . . 4
1.1.1. Addition, subtraction and multiplication of vectors . . 4
1.1.2. Differentiation with respect to a parameter . . . . . . 6
1.1.3. Differentiation with respect to the spatial coordinates 6
1.2. Exercises and problems . . . . . . . . . . . . . . . . . . . . . 17
Bibliography 241
Index 243
Chapter 1
Introduction
i1
......
.......
....
...
...
...
...
...
...
....
.
.
...... ............
.
.
.
O
.
....
. . .......
...
.. .......
..
...
..... .......
. .......
....... .........
..
..............
.
. ...........
.
i2 i3
Figure 1.1. Standard measuring rod and Cartesian reference frame {O, i1 , i2 , i3 }
in three-dimensional space.
....
...
..
...
...
...
..
................
....
... .. .......
.. .. .......
.... ...
....... ...
.......
....... ... .......
...... ... .......
.......... ... .......
. ....... ... .......
. ....... .
.
.... ..
. .......
....... ..
. ..........
... ....... ... ....... .
. .......... ...... ....
. .... ....... ....... .
.... i1 .....
.
........
..
x
.......
....... .......
...... .
....
... .
. ... .......................... ...
.
. .... ...... .
. .
. ...
.... ............... .... ..
.
....
.... ..........
.... . ..
.
O
... ..
.
...
i2 ......
.
..........
...........
..
. .........
............. ....
. . ...
...
.
... ........
.. ....... i3 ... .........
....... ....... .
....
...... . .......
................
.
.....
.. ..
..
.... ... .
. ..
......
x1 ....
.......
.......
....
.........
...
.
....... ....... .... ....... .
..........
.
....... . .................
....... .... .......
.......
.......
....... ...... .......
x3 ....... ....
. x2 ......
....
....... .......
.......... ......
.
ϕ v = ϕ v1 i1 + ϕ v2 i2 + ϕ v3 i3 . (1.3)
v · w = v1 · w1 + v2 · w2 + v3 · w3 = w · v , (1.4)
cartesian vectors and their properties 5
v × w = (v2 w3−v3 w2 )i1 + (v3 w1−v1 w3 )i2 + (v1 w2−v2 w1 )i3 = −w × v , (1.6)
or in matrix notation
i
1 i2 i3
v × w = v1 v2 v3 . (1.7)
w1 w2 w3
or in matrix notation
u
1 u2 u3
u · (v × w) = v1 v2 v3 . (1.9)
w1 w2 w3
u · (v × w) = v · (w × u) = w · (u × v) (1.10)
= −w · (v × u) = −v · (u × w) = −u · (w × v) .
∂t (v × w) = (∂t v) × w + v × ∂t w . (1.14)
Similarly, we have
div v = ∇ · v = ∂1 v1 + ∂2 v2 + ∂3 v3 , (1.20)
curl v = ∇ × v = (∂2 v3 −∂3 v2 )i1 + (∂3 v1 −∂1 v3 )i2 + (∂1 v2 −∂2 v1 )i3 . (1.21)
• vector rules;
We now summarize the rules for the differentiation with respect to the
spatial coordinates of the scalar functions ϕ = ϕ(x) and ψ = ψ(x), and of
the vector functions v = v(x) and w = w(x).
∇(ϕ + ψ) = ∇ϕ + ∇ψ , (1.22)
∇ · (v + w) = ∇ · v + ∇ · w , (1.23)
∇ × (v + w) = ∇ × v + ∇ × w , (1.24)
∇(ϕ ψ) = (∇ϕ) ψ + ϕ ∇ψ , (1.25)
∇ · (ϕ v) = (∇ϕ) · v + ϕ ∇ · v , (1.26)
∇ × (ϕ v) = (∇ϕ) × v + ϕ ∇ × v , (1.27)
∇ · (v × w) = (∇ × v) · w − v · (∇ × w) , (1.28)
∇ × (v × w) = (w · ∇)v − w ∇ · v − (v · ∇)w + v ∇ · w , (1.29)
∇(v · w) = w × (∇ × v) + (w · ∇)v + v × (∇ × w) + (v · ∇)w .
(1.30)
We note that we have assumed that the functions ϕ, ψ, v and w are differen-
tiable functions of the spatial coordinates. When we assume that ϕ = ϕ(x)
8 introduction
∇·x = 3, (1.38)
∇×x = 0, (1.39)
(∇ · ∇)|x| n
= n(n + 1)|x| n−2
, (1.40)
∇(a · x) = a , (1.41)
(a · ∇)x = a , (1.42)
(a × ∇) × x = −2a . (1.43)
Interpretation of grad ϕ
We consider a continuously differentiable scalar function ϕ = ϕ(x) and
we take the dot product of its gradient ∇ϕ (del ϕ) and an infinitesimal
increment of length
Thus we obtain
∇ϕ .
....
.........
..............
.
... .
..
...
...
..
... .... ... .................. ....
.
.... ....... .
.......... ...... ...
........ ...... ..
....... ...
.
..........
.......
......
......
.
. ..
....... ..
.... ϕ(x) = C + ∆C
....... ...... ....
....... ...... ... ....
...................
....... ....... .............
........ . .. ............
.... ...
......... .... .............. ......
...
............
........ ......... .
............................................. ...
.. ........... ...........
... ..............................................................
..
.....
.
...
................................... ..
........... ...
..
. ......... ......
.. .....
.. ....... .
....... ...... ...
.
ϕ(x) = C
...... ......
....... ...... ...
..
....... ......
...... ..... ..............................
.......
..........
.
.......
......
...
s .
.. .
.
.........
dx
.................... ............................... ..............
............
........................
...............
...... . .................. ........
....... ..
..... .. ...........
............. P s
................... ...
..... . ............................ .............
..... .........................................
.....
.
......
.....
.
.....
..
Q
xP .......
.....
.....
.
...
.....
.....
.....
i1 ..
......... .
...
.....
.....
... ....
.....
... ..........
... .......
......
.
.... ....
....... ............ ..
i2 ..........
........... O ...........
...
i3
dϕ = (∇ϕ) · dx = 0 , (1.46)
dϕ = dC = (∇ϕ) · dx . (1.47)
10 introduction
For a given dϕ, |dx| is a minimum when it is chosen parallel to ∇ϕ, or, for
a given |dx|, the change in the scalar function ϕ is maximized by choosing
dx parallel to ∇ϕ. This identifies ∇ϕ as a vector having the direction of
the maximum space rate of change of ϕ.
Very often the notion of directional derivative occurs. When τ is a unit
vector, the quantity τ · ∇ϕ is called the directional derivative of ϕ in the
direction of τ , and equals the rate of change of ϕ in the direction of τ , viz.,
τ · ∇ϕ = ∂τ ϕ = τ1 ∂1 ϕ + τ2 ∂2 ϕ + τ3 ∂3 ϕ . (1.48)
τ · ∇ϕ = ∂τ ϕ = 0 , (1.49)
Interpretation of div v
We consider a continuously differentiable vector function v = v(x). The
divergence operator ∇ · v (del dot v) results in a scalar quantity indicating
the outflow of a vector field. It can be obtained from the limiting behavior
of the net outflow integral for a vanishing small enclosed volume. To show
this we first compute the net outflow of a vector field v over the elementary
domain with volume dV = dx1 dx2 dx3 at the center of the elementary domain
(see Fig. 1.4). This latter point is given by xP = { 12 dx1 , 12 dx2 , 12 dx3 }. By
Taylor’s theorem, the field component v1 is
v1 (x) = v1 (xP ) + (∂1 v1 )(x1 − 12 dx1 ) + (∂2 v1 )(x2 − 12 dx2 ) + (∂3 v1 )(x3 − 12 dx3 )
The surface integral of the normal component of v (in the direction of the
outward normal) over the top surface {x1 = dx1 , 0 < x2 < dx2 , 0 < x3 <
dx3 } of the volume element, shown in Fig. 1.4, is
dx2 dx3
v1 (dx1 , x2 , x3 ) dA =[v1 (xP ) + 12 (∂1 v1 ) dx1 ]dx2 dx3
x2 =0 x3 =0
i1
......
.........
... .... ..
...
....
..
...
...
...
...
...
dx1
. .
...
.........
.
.... .... ...................................................
..
.
.
.
.....
.
..
.... ...
.
.
..........................
..........................
............ ∂V .
.
.....
.
. . .... ......
..... .... ..... ... ......
..
. . ..... .......
..
....
. . .
.
.
. ... ...... ..............
..... .... .... ....
..... .....
............... .. ..... ...
.... ........................................................ ..... ...
... ... ........................... .
..
...... ...
........................... .....
... ... ......... ...
... ... .
. ...
... ... .... ...
... ... .
. ...
... ... ..
. ...
... .. .. ...
s
. .
... ... .... ....
... .
. .
. ..
...
...
...
xP ..
.
....
.
.
..
.
....
.
.
r ........................................
ν
..
... ....
... .... .... ...
... ... ... ...
... ............................
.
. .
.
. ...
... .
. . . ...
O ..... ......................... .
.
... .. ...... . ...
...
... ..
....
.
.
..
...... ................................ .
.
.
.
.
.
. .
. .. .
..... ...
..
.
dx3
............................
.........................
... .
.... .
. ..... ...................................
...
... .........
....
.....
.
.
.
.
....
. .
........
.....
.
. ........ .... i3
.... ....... ..
. ..
.....
.. .... ....
................................................... .....
..... ........................... ... .........
..
.
.
.....
... .....
................. dx2 ...........................
..................................
i2
The surface integral of the normal component of v (in the direction of the
outward normal) over the bottom surface {x1 = 0, 0 < x2 < dx2 , 0 < x3 <
dx3 } of the volume element, shown in Fig. 1.4, is
dx2 dx3
− v1 (0, x2 , x3 ) dA = − [v1 (xP ) − 12 (∂1 v1 ) dx1 ]dx2 dx3
x2 =0 x3 =0
the negative sign in front of the integral coming in because the outward
pointing component of v is integrated, and the outward pointing component
of v for the bottom surface is −v1 . The sum of the surface integrals over these
two faces is therefore simply ∂1 v1 dx1 dx2 dx3 , to the order of approximation
considered here. The contributions to the other faces depend on v2 and v3
and can be computed in a similar way. The net outflow integral from the
volume
element is therefore
⊂⊃ ν · v dA = (∂1 v1 + ∂2 v2 + ∂3 v3 )dx1 dx2 dx3 = ∇ · v dV , (1.53)
x∈∂V
12 introduction
in which
V = dV (1.55)
x∈V
is the volume of the domain V. The integral x∈∂V ν · v dA is called the
flux of the vector field v through the closed surface ∂V.
∂D
ν
⊂⊃ ν · v dA = ∇ · v dV = div v dV , (1.56)
x∈∂D x∈D x∈D
Interpretation of curl v
We consider a continuously differentiable vector function v = v(x). The
curl operator ∇ × v (del cross v) results in a vectorial quantity indicating
the rate of circulation of a vector field. It can be obtained from the limiting
behavior of the net circulation integral around a vanishing small surface
area. To show this we first consider the case that the elementary area is
perpendicular to the i1 -axis. The center point of the area is given by xP =
{ 12 dx1 , 12 dx2 , 12 dx3 }. The circulation integral for the path shown in Fig. 1.6
is
dx2 dx3
τ · vdl = v2 ( 12 dx1 , l, 0) dl + v3 ( 12 dx1 , dx2 , l) dl
x∈∂A x2 =0 x3 =0
dx2 dx3
− v2 ( 12 dx1 , l, dx3 ) dl − v3 ( 12 dx1 , 0, l) dl ,
x2 =0 x3 =0
(1.59)
14 introduction
i3
.....
........
..........
...
....
..
...
...
...
...
...
... τ ..
∂A
.....
............................
dx3 ....
.. .. .....
.....
.....
... ....
.
...
...
...
...
...
...
...
... .... ..
r xP
........
....... ..
.. ......
... .......
..... .. ... ..
..
...
...
...
...
...
...
...
... .. ..
.............................................................................................................................................................................................................................
i2
O
.. ..
dx2
in which τ is the unit vector along the path and l is the arclength along the
path. Substituting the Taylor’s series expansion (see Eq. (1.50)) for v2 and
v3 we find
τ · vdl = v2 (xP ) dx2 − 12 (∂3 v2 )dx2 dx3 + v3 (xP ) dx3 + 12 (∂2 v3 )dx2 dx3
x∈∂A
−v2 (xP ) dx2 − 12 (∂3 v2 )dx2 dx3 − v3 (xP ) dx3 + 12 (∂2 v3 )dx2 dx3
plus higher order terms. Subsequently, the net circulation integral for the
elementary surface area perpendicular to the i2 -axis is (∂3 v1 − ∂1 v3 )dx3 dx1
and the net circulation integral for the elementary surface area perpendicular
to the i3 -axis is (∂1 v2 − ∂2 v1 )dx1 dx2 . Hence, the analogue of the curl is the
net circulation integral per unit of surface area at the point x and is written
as
τ · v dl
ν · curl v = lim x∈∂A = ν · (∇ × v) , (1.61)
A→0 A
cartesian vectors and their properties 15
in which
A= dA (1.62)
x∈A
is the area of the surface A and ∂A is the closed boundary of the surface
area A. In Eq. (1.61), ν is the unit vector normal to the surface area A and
oriented toward the side of advance of a right-hand screw as it is turned in
the direction of τ around ∂A.
τ · v dl = (ν × ∇) · v dA = ν · curl v dA , (1.63)
x∈∂S x∈S x∈S
∂S
in which ν is the unit vector normal to the surface area S and is oriented to-
ward the side of advance of a right-hand screw as it is turned in the direction
of τ around ∂S.
Two modifications of Stokes’ integral theorem are obtained as follows.
With v = a ϕ, in which a is an arbitrary constant vector and ϕ is a contin-
uous differentiable scalar function, we obtain
τ ϕ dl = (ν × ∇)ϕ dA = ν × grad ϕ dA , (1.64)
x∈∂S x∈S x∈S
With v = a × w, in which a is an arbitrary constant vector and w is a
continuous differentiable vector function, we obtain
τ × w dl = (ν × ∇) × w dA . (1.65)
x∈∂S x∈S
introduction 17
Exercise 1.1
Calculate the following expressions and eliminate as many vector products
as possible
(a) a × a
(b) (a × b) · c − a · (b × c)
( c ) (a × b) × c
(d) (a × b) · (c × d)
( e ) (a × b) × (c × d).
Exercise 1.2
What is the direction of the vector c = a × b with respect to the vectors a
and b? Calculate (a × b) · a and (a × b) · b.
Exercise 1.3
A two-dimensional rectangle in the plane x3 = 0, with dimensions d1 , d2 ,
can said to be spanned by two vectors along the two base vectors of the
Cartesian reference frame, viz., d1 = d1 i1 , d2 = d2 i2 . Let the unit normal
ν point in the positive x3 -direction. Give a geometrical interpretation of
A = ν · (d1 × d2 ).
Exercise 1.4
A three-dimensional brick in space, with dimensions d1 , d2 , d3 , can said to be
spanned by three vectors along the Cartesian reference frame base vectors,
viz., d1 = d1 i1 , d2 = d2 i2 , d3 = d3 i3 . What is the geometrical meaning of
the scalar triple product V = d1 · (d2 × d3 )?
Exercise 1.5
Sketch the following vector fields by selecting several points and by drawing
at each of these points an arrow with a direction that corresponds to the
direction of the vector field and with a length that is proportional to the
magnitude of the vector field
(a) v(x) = −x1 i1
(b) v(x) = x2 i1
18 introduction
( c ) v(x) = −x1 i1 − x3 i3
(d) v(x) = x1 i3 − x3 i1 .
Exercise 1.6
Determine the divergence and the rotation of the vector fields in Exercise
1.5.
Exercise 1.7
The operator ∇ · ∇ in Eq. (1.31) is a scalar operator that occurs often
in physics. It has therefore been given a name; it is called the Laplacian
(operator) and is often denoted as ∇2 . Given the at least twice differentiable
vector function A(x), write ∇2 A out in components.
Exercise 1.8
Show that if an arbitrary vector function E(x) can be written as E = −∇V ,
where V (x) is a differentiable scalar function, the vector field E is curl free,
i.e. irrotational, and hence ∇ × E = 0.
Exercise 1.9
Show that if an arbitrary vector function D(x) can be written as D =
∇ × C, where C(x) is a differentiable vector function, the vector field D(x)
is divergence free, i.e. solenoidal, and hence ∇ · D = 0.
Exercise 1.10
The distance between two points with position vectors x and x is given
by the length of the vector from x to x , d(x, x ) = |x − x|. When ∇
and ∇ denote partial differentiation with respect to x and x , respectively,
determine
(a) ∇d , ∇ d
( c ) ∇ · ∇d , ∇ · ∇ d
Exercise 1.11
Given the smooth surface S with unit normal ν, and a vector S = E × H.
Show that the normal component ν · S of S can also be written as ν · S =
ν · [(ν × E) × (ν × H)].
exercises and problems 19
Exercise 1.12
Given the smooth surface S with unit normal ν. Show that an arbitrary
vector H can be decomposed into H = (ν · H)ν + (ν × H) × ν. Give a
geometrical interpretation of the two terms.
Problem 1.1
Calculate the volume of a three-dimensional parallepiped spanned by the
arbitrarily oriented vectors a, b and c.
Problem 1.2
Given the smooth surface S with unit normal ν. Decompose the operator
∇ in two parts, one part tangential and the other normal to the surface S.
Problem 1.3
Use Gauss’ theorem to calculate x∈∂D ν dA, where ν is the outward unit
normal to the closed boundary ∂D of the domain D (see Fig. 1.5).
Problem 1.4
Use Stokes’ theorem to calculate x∈∂S τ dl, where τ denotes the unit tan-
gent along the closed boundary contour ∂S of the surface S. The integration
runs in the direction of circulation that forms a right-handed system with
the unit normal ν to S (see Fig. 1.7).
Chapter 2
In accordance with this, and with the conventions of the International Sys-
tem of Units, we postulate the expression (see Fig. 2.1)
F = qE + qµ0 v × H . (2.1)
In this expression
F = force (N),
q = electric charge (C),
v = velocity of point charge with respect to the observer (m/s),
µ0 = permeability in vacuum (H/m),
E = electric field strength (V/m),
H = magnetic field strength (A/m).
........
........
......
F
......
.
..
........
.
......
......
......
......
.. .........
.
......
......
......
......
......
i1 x u ......
......
.
......
.
......
....
......... ..
...
.
.....
.......... ........................................................... v
.......
...
..
... .......
....... q
......
... ......
... ......
... . .
...... .
... ......
.......
... .......
... .......
... .. ......
.
... .....
......
.. .......
.........
.
...
........ .............
...... .......
.......
.......
.......
O .......
.......
.......
...
.........
. .......
.......
. . .
...
...... .........
.
........... ............
i2 i3
Figure 2.1. Force F on an electric point charge of strength q and moving with
velocity v with respect to an observer at rest in the reference frame {0, i1 , i2 , i3 }.
F = qE (2.2)
F = qµ0 v × H (2.3)
Using the electric point charge as a measuring device one can investigate
the properties of an electromagnetic field in a vacuum domain. In particular,
it proves to be fruitful to investigate the relationships between the changes of
the field in space and the changes of the field in time. Quantitatively, this is
most simply done by mutually comparing the first-order partial derivatives
with respect to the spatial coordinates with the first-order partial derivatives
with respect to time. Carrying out the analysis, it is found that
−∇ × H + ε0 ∂t E = 0 , (2.4)
∇ × E + µ0 ∂t H = 0 . (2.5)
In these equations
Equations (2.4) and (2.5) are known as Maxwell’s equations (in vacuum).
Further, it is found that
∇·E = 0, (2.7)
∇·H = 0. (2.8)
The latter equations are not independent of Maxwell’s equations and are a
kind of compatibility relations. By noting that ∇ · (∇ × H) = 0, it follows
from Eq. (2.4) that ∂t ∇ · E = 0. Hence, if at some instant ∇ · E = 0, this
would be so at any instant. In particular, this applies when we study the
causal fields that are generated by sources that are switched on at a certain
instant, starting from an initially vanishing field. The same reasoning applies
to Eq. (2.5), from which, since ∇ · (∇ × E) = 0, it follows that ∂t ∇ · H = 0.
−∇ × H + ε0 ∂t E = −J mat , (2.9)
∇ × E + µ0 ∂t H = −K mat
. (2.10)
Here,
As we have said,
.................................
.............. .......
.........
........ .....
...
......... .....
...
...
...... ...
........ ...
...
..... ...
...
....
J mat
..
.. ..
...
.... ...
.....
. .
3 ..
.
...
..
. .
.
...
.. .
..
... ..
s
. .
... ...
... ...
...
...
.
...
... @ ...
...
...
vacuum
...
... @ ...
...
...
...
...
R K mat
@ ...
....
... ...
...
... ....
..... ..
..
....
matter ...
.....
.....
.
... .....
... ...
......
... ......
... .....
... ......
......
...
..... ....
........
...... .....
........ ........
..................................................
Figure 2.2. Piece of matter embedded in vacuum, with volume density of electric
current J mat and volume density of magnetic current K mat .
where
The latter quantities are particularly useful for describing the electromag-
netic properties of matter in case only fields varying slowly in time are
present. Substitution of Eqs. (2.12) - (2.15) in Eqs. (2.9) and (2.10) leads to
−∇ × H + ε0 ∂t E + J + ∂t P = −J ext , (2.16)
∇ × E + µ0 ∂t H + µ0 ∂t M = −K ext
. (2.17)
D = ε0 E + P , (2.18)
B = µ0 (H + M ) . (2.19)
where
With the aid of Eqs. (2.18) and (2.19), Eqs. (2.16) and (2.17) can be rewrit-
ten as
−∇ × H + J + ∂t D = −J ext , (2.20)
∇ × E + ∂t B = −K ext
. (2.21)
Equations (2.20) and (2.21) are known as Maxwell’s equations (in matter).
These equations constitute an incomplete system of equations since the num-
ber of equations is less than the number of unknown quantities, assuming
that the right-hand sides, which are representative of the action of exter-
nal sources, are known. The supplementing equations are furnished by the
28 the electromagnetic field equations
constitutive relations that describe the reaction of passive matter to the pen-
etration of an electromagnetic field.
As far as the compatibility relations are concerned, we observe that ap-
plication of the divergence operator to Eqs. (2.20) and (2.21) yields
∇ · J + ∂t ∇ · D = −∇ · J ext , (2.22)
∂t ∇ · B = −∇ · K ext
. (2.23)
instants, the medium is said to show relaxation; the property that only
the past is involved in relaxation phenomena, is known as the principle of
causality.
When the values of {J , D, B} at some position are related to the values of
{E, H} at the same position only, the medium is denoted as locally reacting.
Almost all media are locally reacting in their electromagnetic behavior.
If at a point in space the constitutive operators are orientation invariant,
the medium is denoted as isotropic at that point. If this property does not
apply, the medium is denoted as anisotropic.
In a domain in space where the constitutive operators are shift invariant,
the medium is denoted as homogeneous; in a domain in space where the
shift invariance does not apply, the medium is denoted as inhomogeneous or
heterogeneous.
For a wide class of materials, the quantities J and D only depend on E
(and not on H), while the quantity B only depends on H (and not on E).
For a medium that is, in addition, linear, time invariant, instantaneously
reacting, locally reacting and isotropic in its electromagnetic behavior, we
then have
where
σ = conductivity (S/m),
ε = permittivity (F/m),
µ = permeability (H/m).
In the tables of physical constants (see Table 2.1) one customarily spec-
ifies the dielectric properties of a material through
εr = ε/ε0 = relative permittivity
and its magnetic properties through
µr = µ/µ0 = relative permeability;
εr and µr are dimensionless.
εr = ε/ε0 σ (S/m)
air 1.0006
glass 6 1.0 × 10−12
quartz 4.6 1.0 × 10−12
bakelite 4.8 1.0 × 10−9
dry earth 5 10−4
wet earth 10 10−2
fresh water 81 10−3
seawater 81 4
salt water (20%) 81 20
aluminium 1 3.5 × 107
copper 1 5.8 × 107
silver 1 6.17 × 107
the system of field equations 31
Let the subscripts {1, 2, 3} denote the components of a vector along the
{i1 , i2 , i3 }-axes of the chosen reference frame, respectively. Then the elec-
tromagnetic field is governed by the six scalar equations for the three compo-
nents of the electric field strength and the three components of the magnetic
field strength. This system of electromagnetic field equations is
−(∂2 H3 − ∂3 H2 ) + σE1 + ε∂t E1 = −J1ext , (2.30)
−(∂3 H1 − ∂1 H3 ) + σE2 + ε∂t E2 = −J2ext , (2.31)
−(∂1 H2 − ∂2 H1 ) + σE3 + ε∂t E3 = −J3ext , (2.32)
...
.........
∂D
........
.......
.......
D(1) .
......
......
..
...
.
......
......
ν .
.....
.
......
.
D(2)
.....
.....
@
I
@ ..
..
.
.....
.....
.....
.....
.....
.....
.
.....
....
....
...
.....
...
...
...
...
....
.
...
...
...
...
...
...
...
....
.
..
...
...
..
Figure 2.3. Interface between two media with different electromagnetic properties.
34 the electromagnetic field equations
and
Equations (2.43) and (2.44) are the boundary conditions at a sourcefree in-
terface between two different media. In these equations only the components
of the electric and magnetic field strengths tangential to ∂D occur.
In view of the continuity of the tangential components of E and H across
the interface ∂D, the normal component of the Poynting vector S, ν · S =
ν · (E × H) = ν · [(ν × E) × (ν × H)], is continuous across this interface,
as it should be on physical grounds.
where ν is the unit vector along the normal to ∂D pointing away from D.
It is not allowed to prescribe boundary conditions for the tangential part of
the magnetic field strength in this case. In fact, the tangential part of the
magnetic field strength will, in general, exhibit a discontinuity upon crossing
∂D. Electrically impenetrable materials arise as limiting cases of materials
whose conductivity and/or permittivity go to infinity.
In view of the vanishing tangential components of E at the boundary
∂D, the normal component of the Poynting vector S, ν · S = H · (ν × E),
vanishes at this boundary as well.
the boundary conditions 35
ν ∂D..................................
.............................................
......
.....
@
I .. .
......
...
...
...
@ ...
..... ...
*
..
.
.... ...
x + δν .... ..
....
@ .
...... ..
1 ...
. ..
.....
. ..
D
. ..
x ..
.
....
.
.....
....
.
.
. .
. .
... ...
... ...
...
... ..
..
...
....
...
.
..
... ...
...
O ...
..
.
...
impenetrable object ...
......
.
... ...
... ...
..... ..
.....
....
.... ....
... .....
.....
...
... ........
.
... ......
......
... ......
... ......
....
..... .
...
........
....... ......
.......... .........
.......................................
where ν is the unit vector along the normal to ∂D pointing away from D.
It is not allowed to prescribe boundary conditions for the tangential part
of the electric field strength in this case. In fact, the tangential part of the
electric field strength will, in general, exhibit a discontinuity upon crossing
∂D. Magnetic impenetrable materials arise as limiting cases of materials
whose permeability goes to infinity.
In view of the vanishing tangential components of H at the boundary
∂D, the normal component of the Poynting vector S, ν · S = −E · (ν × H),
vanishes at this boundary as well.
36 the electromagnetic field equations
that automatically yields the value zero for t < 0 and for which numeri-
cal (Fast Fourier Transform) routines are available. In Eq. (2.48), j is the
imaginary unit (j 2 = −1).
frequency-domain representations 37
Equations (2.49) and (2.50) express the Fourier transformation for causal
fields.
Note that the length of a complex vector v̂ with components v̂1 , v̂2 and
v̂3 is defined as
which is in accordance with the definition of Eq. (1.5) for real vectors. The
asterisk denotes complex conjugate.
ext
−∇ × Ĥ + (σ+sε)Ê = −Ĵ , (2.52)
ext
∇ × Ê + sµĤ = −K̂ , (2.53)
f (x, t) = Re fˆ(x, jω)exp(jωt) . (2.60)
E(x, t) = Ê (x, jω) cos(ωt) + Ê (x, jω) sin(ωt) , (2.69)
where
Ê (x, jω) = Re Ê(x, jω) , (2.70)
Ê (x, jω) = −Im Ê(x, jω) , (2.71)
we observe that the field vector E(x, t) for any instant is located in the plane
through Ê (x, jω) and Ê (x, jω) while its endpoint describes an ellipse in
that plane. The electric field strength (and also) the electromagnetic field is
then said to be elliptically polarized at that point (Fig. 2.5).
When Ê (x, jω) and Ê (x, jω) have the same direction, i.e.,
∗
Ê(x, jω) × Ê (x, jω) = 0 , (2.72)
the ellipse degenerates into a straight line. The electric field strength is then
said to be linearly polarized at that point (Fig. 2.5). Finally, when Ê (x, jω)
and Ê (x, jω) have equal magnitudes and are perpendicular to each other,
i.e.,
the ellipse changes into a circle. The electric field strength is then said to be
circularly polarized at that point (Fig. 2.6).
40 the electromagnetic field equations
ωt = 12 π ..
ωt = 12 π........ ...
.
.....................................................
.......... ....... ......
........ ............ .....
....... ........
t ..........
..........
. ..
..
.....
....
... ...................
..
.. .
...
.
..
..
......
.
.
.
.
.
..
...
...
...
Ê ........... ..
... .
. ... .........
.....
.
.
..
.
.
.
.
..
Ê ...
...
.
............... ωt = 0
........
.......
.
..
.
.. ... .. .
....
...
.
.
.
.
..
.
.
. ...
.
..
Ê...........
.. .
. ..
. .
...
. .
.. ..
. ...
. . . ...
... ... ... ...
... ... ... .....
.. .. .. ...
...
s..
. Ê .......... ....
s...
...
ωt = π ...
.
..
.
..... ....... ....... ....... ....... ....... ...............................................................................................
... .. ..
..
ωt = 0 .
...
... .. ... ...
.... .. ...
.... . .... .
...
.
.. ... ... .
.. ...
.... .. ...
... ... ...
. .
... .... ...
... .. ... .
...
...
... .
.
..
..
...
.
....
.....
.....
ωt = π...............
.
... .....
... .. ..... .
... ... ...... ...
..... ......
.....
...... .
. .
............ .........
....... .. ....... .
.........
...........
.....................................
.
.
... ωt = 32 π
3 ...
ωt = 2π
ωt = 12 π
...........................................................
......... .......
....... ......... ......
t . .........
. .. ... .. ......
.......... ....
.....
...
.
....... .
.
.....
....
......
..
Ê .
.
.
..
.
.
.
...
...
...
...
. .
.... ..
. ...
...
.... .... ...
... ...
... ... ...
...
...
s ...
.......... ....... ....... ....... ....... .......................................................................................... Ê .
ωt = π ...
...
... .......... ....
.
.
ωt = 0
...
... .... ...
... . ...
... .... ...
... ...
... .
... ...
... .... ...
.... . ...
..... ..
.....
. ..
.....
...... ... .....
.....
......
....... ......
......... .. .......
....................................................
...
ωt = 32 π
The term
P out
= ∇ · (E × H)dV
x∈D
= ν · (E × H)dA = ν · SdA , (2.77)
x∈∂D x∈∂D
where Gauss’ integral theorem has been used, is written as an integral over
the surface ∂D through which D is in contact with its surroundings. There-
fore, it seems natural to interpret this term as the instantaneous power that
is, across ∂D, transferred from D toward its surroundings.
42 the electromagnetic field equations
.............................................
∂D................................... ......
.....
...
ν ....
..
.
.
.....
... ...
...
...
..... ...
@
I
@ .. .
........
. ...
..... ..
..
...
.
....
..
. ..
..
.
.
... ...
.
...... ...
...
. ..
.
D
.. ...
...
... ...
... ...
...
. ..
.
... ...
... ...
... ...
..
. .....
... ..
.. ...
.. ...
... ...
... σ, ε, µ .
.....
.
.
.... .....
....
... .....
... .....
... .........
.
... ......
... .....
... ......
......
....
..... .
...
.........
...... ...
......... .........
...........................................
ẇh = σE · E , (2.82)
we = 12 εE · E , (2.83)
wm = 12 µH · H . (2.84)
Steady-state analysis
In the steady-state analysis we define the time average Poynting’s vector
ST as the average of the time domain quantity S(x, t) = E(x, t)×H(x, t).
Let T = 2π/ω be the period in time of the fields, then we define the time
average
1 t +T
ST = E(x, t) × H(x, t)dt . (2.86)
T t=t
Using the complex representations of the type of Eq. (2.60), it follows that
∗
ST = 1
2 Re Ê × Ĥ , (2.87)
44 the electromagnetic field equations
in which Ê = Ê(x, jω) and Ĥ = Ĥ(x, jω) are the complex electric and
∗
magnetic field vectors. The vector Ê × Ĥ is known as the complex Poynting
vector.
To calculate the time-average power, we start with Eqs. (2.61) and (2.62),
and subsequently we take the dot product of the complex conjugate of
∗
Eq. (2.61) with Ê and the dot product of Eq. (2.62) with Ĥ , and add
the results. We then find
∗ ∗ ∗ ext∗ ∗ ext
∇·(Ê × Ĥ )+(σ −jωε)Ê · Ê +jωµĤ · Ĥ = −Ê · Ĵ − Ĥ · K̂ . (2.88)
∗ ∗ ext∗ ∗ ext
1
2 Re ∇ · (Ê × Ĥ ) + 1
2 σ Ê · Ê = − 12 Re Ê · Ĵ + Ĥ · K̂ .
(2.89)
The term
∗
ẇh T = 12 σ Ê · Ê (2.90)
represents the time average of the volume density of heat dissipated by elec-
tromagnetic field, while the right-hand side of Eq. (2.89) can be interpreted
as the time average of the volume density of electromagnetic power that is
generated by the specified source in D whose action has been accounted by
the volume densities of exterior electric current and magnetic current.
the electromagnetic field equations 45
Exercise 2.1
How many unknown quantities occur in Maxwell’s equations in matter?
Exercise 2.2
Substitution of Eq. (2.24) in Eq. (2.22), with J = 0, yields the local form
of the conservation law of electric charge, ∇ · J ext = −∂t ρ. If we integrate
both sides of this equation over the source domain Dext we obtain the global
x∈Dext ∇ · J dV = −∂t
ext
form, x∈Dext ρ dV . Use Gauss’ law to give
a physical interpretation of the global form of the law of conservation of
charge.
Exercise 2.3
Let the plane x3 = 0 with unit normal ν = i3 be the source-free interface
between two media with different electromagnetic properties. Apply the
reasoning, as it is used to derive the boundary conditions given in Section
2.6, to Eqs. (2.30) - (2.35) and derive the boundary conditions for E1 , E2 ,
H1 and H2 .
Exercise 2.4
Apply the reasoning, as it is used to derive the boundary conditions given in
Section 2.6, to Eqs. (2.22) and (2.23), and derive the boundary conditions
for J , D and B at a source-free interface.
Exercise 2.5
Find the polarization state (linear, circular or elliptical) of the following
fields given at a fixed position in space
(a) Ê = ji1 + i2
( c ) Ê = (1 + j)i2 − ji3
( e ) Ê = ji1 + j2i2 .
46 the electromagnetic field equations
Exercise 2.6
Give the dimensions of the following quantities in SI-units
(a) D · E , B·H
(b) J · E , E×H
(c) S , ST .
Exercise 2.7
1
Given the two fields Ê = (i1 + ji2 ) and Ĥ = ( µε00 ) 2 (−ji1 + i2 ) at a fixed
point in space, calculate ST . Then calculate E(t), H(t) and S(t).
Problem 2.1
Substitution of Eq. (2.24) in Eq. (2.22) for a conducting medium, with J ext =
0, yields ∇·J +∂t ρ = 0. In an isotropic homogeneous domain we can express
∇ · J in terms of ρ as ∇ · J = σε ρ. Substitution of this relation leads to a
first order ordinary differential equation for the electric charge density. The
factor τ = σε is called the charge relaxation time of the medium and is a
measure for the time it takes the medium to restore its equilibrium state
once it has been disturbed by an electromagnetic wave.
(a) Given a charge distribution ρ0 at t = 0, solve the ordinary differential
equation for the electric charge density in the time domain.
(b) Use Table 2.1 to find the relaxation times for copper, sea water, fresh
water and glass.
In the frequency domain we can write the equation for the charge density
as (jω + σε )ρ̂ = ρ0 , where the right-hand side comes from the initial charge
distribution considered as a source input. If σ ωε the medium can be
considered a conductor for that frequency; this is equivalent to ωε σ
1,
σ
where the factor ωε is known as the loss tangent also denoted as tan(δ).
( c ) Find the frequency range for which ‘wet earth’ can be regarded a con-
ductor. Use Table 2.1 and the condition tan(δ) > 10.
Problem 2.2
Let the plane x3 = 0 be the interface S between two media with different
exercises and problems 47
electromagnetic properties, D(1) with x3 < 0 and D(2) with x3 > 0, re-
spectively. In D(1) we have the constitutive constants ε(1) and µ(1) (while
σ (1) = 0) and in D(2) we have the constitutive constants ε(2) , µ(2) and σ (2) .
Let further in D(1) , in the limit of x3 ↑ 0, the electric flux density and the
(1) (1)
electric current density be given by D̂1 = 1 and Jˆ1 = 0. Give the values
(2) (2)
of D̂1 and Jˆ1 in the limit of x3 ↓ 0.
Problem 2.3
Show that an elliptically polarized field can be decomposed in two circularly
polarized fields, one left-handed and the other right-handed.
Hint: let Ê = (ai1 + bi2 ) with two arbitrary complex numbers a and
b, then let the superposition of two circularly polarized waves be Ê =
(a i1 + ja i2 ) + (b i1 − jb i2 ) and solve for a , b in terms of a, b.
Problem 2.4
Let the time-harmonic varying electric and magnetic field strengths at a
1
fixed point in space be given by Ê = i1 and Ĥ = ( µε00 ) 2 i2 . Determine
(a) E(t)
(b) H(t)
( c ) S(t).
Use the time-domain expressions for the electric and magnetic fields and the
Poynting vector to calculate
(d) ST
Problem 2.5
Assume that the electromagnetic sources are switched on during a finite
interval (0 < t < T ). Integrate both sides of Eq. (2.76) over all positive
times and give a physical interpretation of the resulting terms in analogy of
Eqs. (2.77) - (2.81).
Chapter 3
One-dimensional
Electromagnetic Waves
A reference frame is introduced such that the sheet coincides with the
plane x3 = 0 (Fig. 3.1). Let the impressed electric current be uniformly in
the x2 -direction and flowing in the direction of decreasing x1 , then
However, since Ê2 and Ĥ1 are not related to the source, we can take
...
...
σ, ε, µ ...
...
σ, ε, µ
...
E. ...
... E
.
... ... ...
.......... ...
... ..........
... ...
..
... ......
...
i1 ..
...
... ....... ...
.... ... ....
S ..........................................×f O r r............................................ S
f
............................
...
...
.... i3
...
H ...
...
...
H
...
...
...
...
...
...
...
...
...
...
....... ..
x3 < 0 ........
......
x3 > 0
J ext .
1
γ = [(σ + sε)sµ] 2 with Re(γ) ≥ 0. (3.8)
In the domain outside the source distribution, i.e., for x3
= 0, the modified
Helmholtz equation (3.7) has a zero right-hand side and admits exponential
functions of the type exp(±γx3 ) as their solutions. From Eq. (3.6) it follows
that the magnetic field also admits these exponential functions as their so-
lutions. The quantity of γ is the (s-domain) propagation coefficient. With
52 one-dimensional electromagnetic waves
the chosen value of the square root in the right-hand side, the propagation
factor exp(−γx3 ) is bounded as x3 → ∞ and unbounded as x3 → −∞, while
exp(γx3 ) is bounded as x3 → −∞ and unbounded as x3 → ∞. Since, due
to causality, the electromagnetic field in the half-space x3 < 0 must remain
bounded as x3 → −∞ and the electromagnetic field in the half-space x3 > 0
must remain bounded as x3 → ∞, we write
and
1
σ + sε 2
Y = . (3.15)
sµ
1
sµ 2
(3.18)
Z= .
σ + sε
and
ê+ − ê− = 0 . (3.23)
Combining Eqs. (3.22) and (3.23), with Eqs. (3.16) and (3.17), we arrive at
54 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
The inference is intended to be weak, not strong. But even so, exactly how it is
supposed to go and whether it is reasonable even as a weak inference is not at
all clear.
In what follows I will construct an idealized version of this argument,
which, although I cannot claim it to be Newton's, may nevertheless reflect
some features of his thought. (See Essay 3.) The argument will have two
essential components: objections to the wave theory, and an appeal to the
explanatory power of the particle theory. Moreover, it will be an argument
whose conclusion is drawn with probability. This probability will be reasona-
bly high but not high enough to achieve the certainty, or virtual certainty, of a
"deduction from phenomena." In constructing the argument I will assume
that the usual axioms of the probability calculus are satisfied. Probability can
be construed here as representing rational degrees of belief.
Let us assume to begin with that light is either a particle phenomenon or a
wave phenomenon. Newton himself offers no explicit argument for such an
assumption, although here is one that he was in a position to offer (and that is
very similar to one in fact offered by his wave-theoretic opponents in the nine-
teenth century; see Essay 3):
51. Ibid.
NEWTON'S CORPUSCULAR QUERY AND EXPERIMENTAL PHILOSOPHY 55
Let us assume that the premises are true, that they report "phenomena," and
that this is a "deduction from phenomena," so that the conclusion has the
highest certainty possible in experimental philosophy. Such an inference, a
type of causal simplification, would be permitted by Newton's Rule 2 requir-
ing that like causes be assigned to like effects, as far as possible. If in other
cases motion is caused by particles or waves, then, unless there is evidence to
the contrary, we should infer like causes in the case of the motion of light.
Let T1 be the hypothesis that light consists of particles, T2 the hypothesis
that light consists of waves, O the observed fact that light travels in straight
lines with uniform speed, and b the accepted background information, which
includes the information in the second premise above. We might express the
results of this argument probabilistically, as follows:
That is, the probability that either T1 or T2 is true, given O and b, is equal to
1. (With minor alterations an argument similar to that which follows can be
made if equation (1) is changed to say that the probability in question is close
to 1; see Essay 3.)
Let us suppose that by appeal to certain other observed facts about light —
call them O'—we can show that the probability of the wave theory is low, say
less than 1/2. Which facts these are, and how this is to be shown, will be taken
up in a moment. For the present let us simply write
Let O1, . . . ,On be various observed facts about light (e.g., reflection, refrac-
tion, variety of colors, fits of easy reflection and transmission, etc.) other
than those in O and O'. We would like to know how probable T1 is when these
facts are considered in addition to O&O'&b. Is p(T 1 /O 1 , . . . ,On&O'&b) also
high? Suppose that explanations of O1, . . . ,On by theory Tl can be con-
56 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
structed in such a way that O1, On follow deductively (in the ordi-
nary sense) from T1 together possibly with the background information b.
But if the particle theory Tl (together possibly with b) deductively entails
O1, On, then it follows from the probability calculus that p(T1:/
O1, . . . ,On&O&O'&b) p(T1/O&O'&b). So from (3), given the existence
of such explanations, we derive
which can be construed as the conclusion of the argument. It says that the
particle theory is probable (more probable than not) given not just a few
chosen optical phenomena but a range of them, including ones explained by
that theory.
The explanations of O1, . . . ,On provided by the particle theory do not
create the high probability for that theory, but they do sustain it. (In Essay 4 I
argue that, in general, derivational explanations, even those subject to certain
further conditions, will not by themselves suffice to guarantee high probabili-
ty, although they can increase it.) These explanations permit an inference
from (3) to (4). In an attempt to establish high probability for the particle
theory on the basis of a range of optical phenomena, this is an essential role
played by such explanations. To create high probability in the first place, a
type of eliminative argument is used in which the wave and particle theories
exhaust the probability but the probability of the wave theory is low. How is
the latter to be established?
Newton offers arguments of two types, one direct, the other indirect. An
argument appealing to diffraction is an example of the former. In the case of
other wave motions such as sound waves and water waves, diffraction into the
shadow of an obstacle is observed, but no such diffraction into the shadow
had been observed by Newton or others in the case of light (though Newton
had observed diffraction away from the shadow).52 So if we include in the
background information the fact of observed diffraction with other wave
phenomena and the absence of such observations in the case of light, then the
probability of the wave theory, given b, is low.
Second, Newton offers a more indirect type of argument. He points out
that to explain certain observed optical phenomena, the wave theory intro-
duces auxiliary assumptions that are either refuted by, or made very improba-
ble by, observations. For example, to explain differences in refrangibility of
rays emerging from a prism, wave theories introduce the auxiliary hypothesis
that the prism modifies rather than separates the rays. Newton argues that
this modification assumption is refuted or at least made extremely unlikely by
further refraction experiments. (See Experiment 5, Book I, pp. 34ff.) Can we
52. See Roger H. Stuewer, "A Critical Analysis of Newton's Work on Diffraction," Isis 61
(1970), pp. 188-205.
NEWTON'S CORPUSCULAR QUERY AND EXPERIMENTAL PHILOSOPHY 57
infer from this that the wave theory is improbable? We can if we suppose that
the probability of the modification assumption, given the wave theory and
the observations, is close to 1. Letting M be the modification assumption, T2
the wave theory, and O' the results of various of Newton's refraction experi-
ments, if p(M/O&O'&b) is close to zero and p(M/T2&O&O'&b) is close to 1,
then p(T2/O&O'&b) is close to zero. Even more generally, we get the same
result if we suppose simply that p(M/T2&O&O'&b) is much, much larger
than p(M/O&O'&b), without needing to suppose that the former probability
is close to 1. (See Essay 3.)
Accordingly, there are two sorts of arguments to show that the probability
of the wave theory is low, that is, (2). Once this is shown, we can infer the high
probability of the particle theory (3), and its continued high probability in
light of the various observed facts that it explains, that is, (4).
How "Newtonian" is the previous argument? In one respect, quite un-
Newtonian, since it explicitly invokes numerical probabilities and the proba-
bility calculus, neither of which, of course, Newton does. However, in certain
other respects it reflects what Newton seems to be doing in Queries 28 and 29.
Assuming, as I have been, that Newton intends to provide some reasons for
believing the particle theory, albeit "weak" ones, it gives a basis for inferring
that theory with probability rather than certainty. Moreover, in doing so it
takes into account Newton's criticisms of the wave theory and the explanatory
virtues of the particle theory (each of which Newton himself emphasizes),
showing how both contribute to the probability of the particle theory.
One objection that might be offered is that this argument is a type of
eliminative one, whereas Newton at one point rejects (a certain form of)
eliminative reasoning. In a letter to Oldenburg of July 1672 he writes:
I cannot think it effectual for determining truth to examine the several ways by
which phenomena may be explained, unless there can be a perfect enumeration of
those ways. You know, the proper method for inquiring after the properties of
things is to deduce them from experiments. And I told you that the theory, which I
propounded [the theory of the heterogeneity of light rays] was evinced to me, not by
inferring 'tis thus because not otherwise, that is, not by deducing it only from a
confutation of contrary suppositions, but by deriving it from experiments conclud-
ing positively and directly.53
E: Each of the hypotheses h1, . . . ,hn, if true, will correctly explain phe-
nomenon p.
But hypotheses h2,. . . ,hn are false.
Therefore, hypothesis h1 is true.
53. I. B. Cohen, ed., Newton's Papers and Letters (Cambridge, Mass., 1978), p. 93.
58 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
Such arguments, which infer the truth of an hypothesis from the falsity of
competitors, are fallacious, unless a complete enumeration can be made of all
the competitors. Newton appears to be thinking of deductive interpretations
of E in which if the premises are true the conclusion must also be true. And he
is correct in saying that arguments of form E, thus construed, are fallacious
unless a complete enumeration of hypotheses is given. They are also falla-
cious if construed nondeductively, that is, as being such that the premises
make the conclusion probable without entailing it. Unless some suitable as-
sumption is made about the probability of the disjunction of hypotheses that
are mentioned in the first premise, the conclusion that the probability of hl is
high cannot be drawn.
However, the particular eliminative argument I have constructed is not of
form E. The first step in the argument, which leads to (1) above, is not an
explanatory step, but one involving causal simplification. The claim is not
that the particle and wave theories will both explain the finite rectilinear
motion of light, but that in other observed cases when something travels with
uniform speed in a straight line this motion is caused by a series of bodies or a
series of wave pulses in a medium. Also, the claim in the first step is indeed
exhaustive, since it assigns a probability of 1 to the disjunction of hypotheses.
But even if it were not exhaustive in this sense, even if (1) were changed to read
"p(T 1 or T2/O&b) is close to but not equal to 1," a fallacy would not emerge
(although other changes would need to be made in the argument).
I conclude that reconstructing what Newton does in Queries 28 and 29 in
the form of a probabilistic argument that takes us from (1) to (4) above is in
conformity with certain important aspects of Newton's methodology. It com-
bines his explanatory reasoning in Query 29 with his criticisms of the wave
theory in Query 28 to provide some reason, though not the highest possible in
experimental philosophy, to believe the corpuscular hypothesis. Although it is
an eliminative argument it is not one of the type Newton rejects. Because the
hypothesis in question is inferred with a probability not sufficiently high to be
a virtual certainty, Newton could not construe the argument as a "deduction
from phenomena." While we should search for phenomena that will sanction
such a "deduction," we should acknowledge that, assuming its premises are
true, the present argument does provide a legitimate "weak" reason for believ-
ing the corpuscular hypothesis.
A "strong" inference furnishes the proposition inferred with the highest evi-
dence possible in experimental philosophy; a "weak" inference furnishes some
evidence but not the highest possible. I shall suppose that this difference can
NEWTON'S CORPUSCULAR QUERY AND EXPERIMENTAL PHILOSOPHY 59
(a) tells us that if the prior probability of h is not zero, then as the number n
of observed consequences of h and b gets larger and larger, the probability
that the n + 1 observational consequence of h and b will be true gets higher
and higher, approaching 1 as a limit. (b) tells us that if the prior probability of
h is not zero, then as the numbers m and n get larger, the probability that the
54. See John Earman, "Concepts of Projectibility and Problems of Induction," Nous 19
(1985), pp. 521-535.
60 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
(c) gives a set of sufficient conditions for the probability of (x)(Fx Gx),
given observed instances of the form Fa, Gai, to approach 1 as a limit.
Now let us apply these three probability results to Newtonian inductions.
In all three cases let us consider h's of the form (x)(Fx Gx), and O's instances
of the form Fai Gai. (a) tells us that if the prior probability of (x)(Fx Gx) is
not zero, then as the number of observed instances of (x)(Fx Gx) increases,
the probability that the next instance will obtain gets higher and higher,
approaching 1 as a limit. A similar claim can be made for (b). (a) and (b) —so
construed — correspond to Newton's inductions from some observed members
of a class to some other member(s) of that class. (c) corresponds to Newton's
inductions from some observed members of a class to all members. In all
three cases the probability in question approaches 1 as a limit, under certain
very weak assumptions. Intuitively, the probability that the next instance will
satisfy (x)(Fx Gx), that the next m instances will, and that all instances will,
gets higher and higher as more and more instances are observed. We get more
and more certainty in these cases with more and more observed instances of
(x)(Fx Gx).
However, it is not the case that for every number n of observed instances,
the probability that the next instance will satisfy (x)(Fx Gx), that the next m
instances will (for any m), and that all instances will, is close to 1. Consider
just the latter, and suppose that the prior probability of (x)(Fx Gx) is low,
and the O's are such that, with sufficiently small n, the prior probability of
the conjunction of O's is high. If (x)(Fx Gx) and b entails the O's, then by
Bayes' theorem,
Now if p((x)(Fx Gx)/b) is low and the O's are such that, with sufficiently
small n, p(O 1 , . . . ,On/b) is high, then the probability on the left will be
small, despite the fact that all the observed O's satisfy (x)(Fx Gx). One case
of this sort involves Goodmanesque properties such as "grue," where the prior
probability of the proposition "All emeralds are grue" is very low but where,
given appropriate background information, the probability that observed
items are grue if they are emeralds is very high. Strange Goodmanesque
properties or classes can prevent the probability on the left from being high
for a given n. But as n increases without bound, the probability on the left will
approach 1 as a limit, strange properties notwithstanding.
However, Goodmanesque properties are not the only things that can pre-
vent the probability on the left from being high for a given n. Recall the proof
of Proposition 1 of the Opticks. Here an induction is made from observations
of differences in refrangibility of blue and red rays in an experiment with the
sorts of prisms used by Newton to differences in refrangibility of any differ-
ently colored rays in any sort of refraction, whether or not the latter is
produced by a prism. A critic of Newton might argue as follows: (i) The
number n of observed instances of Proposition 1 (that lights that differ in
color differ in degrees of refrangibility) is quite low. (If we count as instances
here the results of types of experiments, rather than specific trials, then the
critic has some justification, since Newton cites only two experiments.) (ii)
The critic might agree that the probability of getting the results Newton
obtains with these types of experiments with prisms is high, while supposing
that obtaining analogous refraction results with other sorts of prisms or
without prisms is improbable.56 (iii) The critic might argue, on the basis of
background information b, that the prior probability of Newton's Proposi-
tion 1 is very low. At least, the critic might complain, Newton does nothing to
dispel doubts expressed in points (i) through (iii). But unless such doubts are
removed, the probability of Newton's Proposition 1, given the results of the
experiments Newton mentions, cannot be assumed to be high. The most we
can say is that this probability will increase toward 1 as the number of ob-
served instances of the proposition increases.
Confining our attention to ordinary deductions and inductions, we can
now answer the question "Are Newton's 'deductions from phenomena' guar-
anteed to be strong inferences?" in the following way. If they are deductions
(in the ordinary sense), they are so guaranteed. Any deductive inference from
O1, ,On and b to h—no matter what number n is — guarantees that the
probability of h given O1, . . . ,On and b is maximal. By contrast, it is not the
case that every particular inductive inference is guaranteed to be strong, no
matter how many instances are involved and no matter what the character of
56. See Simon Schaffer, "Glass Works: Newton's Prisms and the Uses of Experiment," in
David Gooding, Trevor Pinch, and Simon Schaffer, eds., The Uses of Experiment (Cambridge,
England, 1989), pp. 67-104.
62 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
probability of this hypothesis is not zero. This, of course, does not imply that
the probability of the "hypothesis" itself approaches certainty, but only that
the probability of its deductive consequences does.
Newton does not appear to be thinking of cases in which we make infer-
ences to deductive consequences of "hypotheses." But such inferences can be
strong ones, or at least they can get stronger and stronger as more and more
consequences are observed to hold. To be sure, Newton could claim that he is
classifying as "inductive" an inference from some observed consequences of h
to other not yet observed consequences. But his inductions appear to be
simply inductive generalizations from observed F's that are G's to other or all
F's being G's.
Let us turn, then, to result (c) involving the probability of h itself. And let
us consider the more general case in which h is any proposition that, together
with b, deductively implies O1, O2, Here we cannot obtain the result
that lim p(h/O 1 , . . . ,On&b) = 1 because we cannot in general assume that
n
lim
n—
p(O 1 , . . . ,On/b) = p(h/b). Indeed, the following are provable:
(d) Let h (together with b) entail O1, O2, . . . . If h has at least one
incompatible competitor h' that together with b also entails O1, O2,
. . . , and whose probability on b is greater than zero, then lim
p(h/O 1 , . . . ,On&b) 1.
(e) Let h together with b entail O1, O2, . . . . If h has at least one incom-
patible competitor h' that together with b also entails O1, O2, . . . ,
and is such that p(h'/b) p(h/b), then for any n, no matter how
large, p(hlO 1 , . . . ,On&b) .5.57
(f) If h, h1, . . . , hk form a partition on b, then for any O, and for each
hi ( h) in the partition, and for any number r greater than or equal to
0 and less than 1, p(h/O&b) > r if and only if p(h i /O&b) <
1 - r. i=1
Now suppose that we have some observed phenomena O, O1, ,On and
background information b, and we want to make a strong inference to h by
showing that the probability of h given the observed phenomena and back-
ground information is greater than some threshold value r for "very high"
probability. Using theorem (f), the following strategy is possible:
Strategy for showing that h has a very high probability (greater than some threshold
value r for very high probability), given observed phenomena O, O1, . . . , On and
background information b:
3. Show that O1, . . . ,On are derivable from h (together with b).
too modest, even for Newton, since the actual facts cited, Newton thought,
cast much more doubt on T2 than this. The two mentioned were diffraction
and refraction. If light is a wave phenomenon, then, like water waves and
sound waves, it ought to be diffracted into the shadow; but Newton observed
no such diffraction. Second, Newton (as well as defenders of the wave theory)
believed that, given the wave theory and the observations of differences in
degrees of refraction, the probability that light is modified by the refracting
prism is very high, say close to 1. But on the basis of his own refraction
experiments, Newton pretty clearly thought he had refuted the modification
assumption, that is, the probability of this assumption, given his experimen-
tal results, is close to zero. So, where M = the modification assumption, and
O' includes the results of Newton's refraction experiments, we have the result
ihat p(M/T2&O'&O&b) is close to 1, whereas p(M/O'&O&b) is close to zero.
It follows that p(T2 /O'&O&b) is close to zero. Letting O' also contain the
observed absence of diffraction into the shadow and b also contain observed
diffraction in the case of sound and water, we write
indicated what type of argument for this claim Newton could have given that
would be compatible with his general methodology. If the strategy is launched
by this assumption in step (1), then Newton's own arguments against the wave
theory can be used to justify the steps leading to the final (5).
Let us suppose that (1) is justified by inference from observed phenomena.
And let us assume that the remaining steps are also valid, so that the argu-
ment does establish the very high probability of a proposition given certain
observations and background information. If so it provides the basis for a
"strong" inference to that proposition from those observations and back-
ground information. Is the argument a "deduction from phenomena"?
In certain ways it seems quite different from the sorts of arguments New-
ton has in mind when he uses this expression. First, unlike the "deductions"
that Newton gives, it contains an inference to a disjunction of propositions in
the first step. Second, the argument is eliminative, whereas the "deductions"
Newton offers are not. Indeed, he rejects (certain types of) eliminative argu-
ments. Third, and most important, it makes use of the probability calculus,
which Newton never does. The inferences to (3), (4), and (5) are justified by
principles of probability. Whether Newton would have been willing to classify
such inferences as "deductive" is unclear.
Yet reasons might be offered for classifying the argument as a "deduction
from phenomena." First, the previous characterization of causal simplifica-
tion (as well as that of induction) does not preclude an inference to a disjunc-
tion of propositions. Second, although it is eliminative, it is not an elimina-
tive argument of the type that Newton rejects. Indeed, if the previous point is
accepted, it is an eliminative argument that uses causal simplification to
establish a disjunction of propositions and then to argue against one of the
disjuncts. Third, the probability principles generating steps (3), (4), and (5)
might be thought of as, or as akin to, mathematical principles, which for
Newton can serve as a basis for "deductions."
Accordingly, assuming the argument in question is valid, the following
possibilities emerge:
1. In a broad sense the argument is a "deduction from phenomena." If so
it does not refute the Newtonian claim that only "deductions from phenome-
na" guarantee strong inferences. However, if we construe it as a "deduction
from phenomena," then with this argument we must deny that the Newtonian
corpuscular hypothesis is an hypothesis. With this argument we will have
"deduced" the corpuscular hypothesis from the phenomena and thus rendered
it no longer hypothetical.
2. In a narrower sense (one that excludes probability arguments) the argu-
ment is not a "deduction from phenomena." Yet it provides the basis for a
"strong" inference to the corpuscular hypothesis. So if this narrower sense is
Newton's, then we need to reject his idea that only "deductions from phenom-
ena" can provide the highest certainty in experimental philosophy.
NEWTON'S CORPUSCULAR QUERY AND EXPERIMENTAL PHILOSOPHY 67
7. CONCLUSIONS
*For very helpful suggestions I am indebted to Robert Rynasiewicz, Doren Recker, Robert
Kargon, and Alan Shapiro.
This page intentionally left blank
ESSAY 3
Light Hypotheses
1. INTRODUCTION
At the beginning of the nineteenth century Thomas Young published papers
that defended the wave theory of light against the Newtonian particle theory.
Following this there occurred a lengthy and sometimes heated dispute be-
tween particle theorists and wave theorists which, it is alleged, stemmed from
deep divisions over scientific methodology. Particle theorists, it is said, partic-
ularly British ones, used the method of induction whereas wave theorists
employed the antithetical method of hypothesis. Thus Geoffrey Cantor
writes:
Although in the eighteenth century almost every British natural philosopher ac-
cepted without question the corpuscular interpretation of Newton's writings on
optics, by the 1830s most British natural philosophers had rejected Newton's cor-
puscular theory in favor of the wave theory of light. Intimately bound up with this
scientific "revolution" in optical theory was a change in scientific methodology: the
replacement of the method of induction by the method of hypothesis.1
1. Geoffrey Cantor, "The Reception of the Wave Theory of Light in Britain: A Case Study
Illustrating the Role of Methodology in Scientific Debate," Historical Studies in the Physical
Sciences 6 (1975), p. 109. Emphasis mine. See also his book Optics after Newton (Manchester,
1983), pp. 177-186.
69
70 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
The primary reason for opposition to ether theories was the widespread acceptance
among Scottish philosophers and scientists of a trenchant inductivism and empiri-
cism, according to which speculative hypotheses and imperceptible entities were
inconsistent with the search for reliable science.2
As this paper contains nothing which deserves the names either of experiment or
discovery, and as it is in fact destitute of every species of merit, we should have
allowed it to pass among the multitude of those articles which must always find
admittance into the collections of a Society which is pledged to publish two or three
volumes every year.3
2. Larry Laudan, "The Medium and its Message," in G. N. Cantor and M. J. Hodge, eds.,
Conceptions of the Ether (Cambridge, England, 1981), p. 170. Unlike Cantor, Laudan's principal
aim in this paper is to compare the methodology of those who defended the wave theory in the
nineteenth century with eighteenth- (rather than nineteenth-) century inductivist critiques of that
theory (and of ether theories generally). Since my main interest in what follows is (like that of
Cantor) in the debate between wave and particle theorists in the nineteenth century, Laudan's
claim that is of special concern to me is that nineteenth-century wave theorists utilized a method
of hypothesis.
3. Edinburgh Review 1 (1803), p. 450.
LIGHT HYPOTHESES 71
Young's defense of the wave theory on the grounds that it employs an unac-
ceptable method of hypothesis.
In what follows I propose the following:
1. To give an account of the method of hypothesis (or of various such
methods).
2. To argue, contrary to Cantor and Laudan, that in their actual practice,
as well as in their reflections on this practice, nineteenth-century wave theo-
rists such as Young, Fresnel, Lloyd, and Herschel typically employed a meth-
od that is significantly different from the method of hypothesis.
3. To argue that this method contains not only an explanatory component
present in the method of hypothesis, but an "independent warrant" compo-
nent that is not. For wave theorists the strategy for supplying independent
warrant is an eliminative one that can be justified by appeal to probabilistic
and inductive considerations. Of particular interest in this justification will be
probability considerations introduced in section 5 that pertain to the intro-
duction of auxiliary hypotheses.
4. To give an account of what nineteenth-century British particle theorists
such as Brougham meant by "induction" and how they utilized this method in
developing the particle theory. In doing this some attention will need to be
given to Newton's ideas, which exerted considerable influence on later particle
theorists.
5. To argue that there are strong similarities, if not identities, between the
inductivism of British particle theorists and the methodology of wave theo-
rists, and that the important debate is over particles versus waves, not meth-
odologies. Accordingly, the present case will not support a form of relativism
that states that fundamentally different theories employ fundamentally dif-
ferent methodologies.
Let us suppose the existence of the aether, with these its properties, to be destitute
of all direct evidence, still, if it serves to explain a great variety of phenomena, it
will have an indirect evidence in its favour by this means.4
4. David Hartley, Observations on Man, His Frame, His Duty, and His Expectations, 2nd
ed., 2 vols. (London, 1791), vol. 1, p. 15. Quoted by Laudan, op. cit., p. 161.
72 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
Basic method of hypothesis: The fact that hypothesis h if true would correctly
explain observed phenomena O1, . . . ,On constitutes at least some reason to think
that h is true.
. . . ,On (where these are significantly different) provides some reason to think that
h is true.
In point of fact, Whewell's version of the method seems even more com-
plex than this. Whewell notes that formulations like those above presuppose
that
the hypothesis with which we compare our fact [is] framed all at once, each of its
parts being included in the original scheme. In reality, however, it often happens
that the various suppositions which our system contains are added upon occasion
of different researches.7
all the additional suppositions tend to simplicity and harmony; the new supposi-
tions resolve themselves into the old ones, or at least require only some easy
modification of the hypothesis first assumed: the system becomes more coherent as
it is further extended.8
7. William Whewell, The Philosophy of the Inductive Sciences (New York, 1967), vol. 2, p.
68. Emphasis in original.
8. Ibid., p. 68. Emphasis in original.
74 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
believing such hypotheses other than the explanatory ones hypothesists men-
tion.
According to the wave theory, light consists of a wave motion or pulse trans-
mitted through some medium; the medium itself may be composed of parti-
cles that vibrate rather than exhibit translational motion. According to the
particle theory, light consists of discrete particles emanating from luminous
bodies; these particles are subject to forces obeying Newton's laws of motion;
if no such forces are acting the particles move in straight lines with constant
finite velocity.
The assumption that light is either a wave or particle phenomenon is made
on the grounds that these are the two main theories that have been proposed
by the physics community, or on the grounds of some empirical consideration
regarding motion, or both. Taking the former line Young writes:
It is allowed on all sides, that light either consists in the emission of very minute
particles from luminous substances, which are actually projected, and continue to
move with the velocity commonly attributed to light, or in the excitation of an
undulatory motion, analogous to that which constitutes sound, in a highly light
and elastic medium pervading the universe; but the judgments of philosophers of
all ages have been much divided with respect to the preference of one or the other
of these opinions.9
Humphrey Lloyd, after a few preliminaries, begins the body of his 1834
report on the present state of physical optics as follows:
The first property of light which claims our notice is its progressive movement.
Light we know, travels from one point of space to another in time, with a velocity
9. Thomas Young, A Course of Lectures on Natural Philosophy and the Mechanical Arts
(London, 1845), p. 359. Also taking the former line, Fresnel, like Young, begins by noting that
the particle and wave theories represent "the two systems which have up till now divided scientists
with respect to the nature of light." A. Fresnel, "Memoir on the Diffraction of Light" (1816),
reprinted (in part) in Henry Crew, ed., The Wave Theory of Light (New York, 1900). For parts of
this material not in Crew, I have used a translation provided by Laurence Selim.
LIGHT HYPOTHESES 75
of about 195,000 miles a second. The inquiry concerning the mode of this propaga-
tion involves that respecting the nature of light itself.
There are two distinct and intelligible ways of conceiving such a motion. Either
it is the self-same body which is found at different times in distant points of space;
or there are a multitude of moving bodies, occupying the entire interval, each of
which vibrates continually with certain limits, while the vibratory motion is com-
municated from one to another, and so advances uniformly. Nature affords numer-
ous examples of each of these modes of propagated movement; and in adopting
one or the other to account for the phenomena of light, we fall upon one or other
of the two rival systems, — the theories of Newton [particle theory] and of Huygens
[wave theory].10
Among the theories which philosophers have imagined to account for the phenom-
ena of light, two principally have commanded attention; the one conceived by
Newton ... in which light is conceived to consist of excessively minute molecules
of matter projected from luminous bodies. . . . The other hypothesis is that of
Huygens . . . , which supposes light to consist, like sound, in undulations or
pulses, propagated through an elastic medium.11
2. Show how each theory explains various observed optical phenomena (e.g.,
rectilinear propagation, reflection, refraction, diffraction, dispersion, etc.).
In this case it is maintained by Newton that the margin of the aperture possesses an
attractive force, which is capable of inflecting the rays. . . . In the Huygensian
10. Humphrey Lloyd, "Report on the Progress and Present State of Physical Optics," Reports
of the British Association for the Advancement of Science (1834), pp. 297-298.
11. J. F. W. Herschel, "Light," Encyclopedia Metropolitana (1845), vol. 4, p. 439.
76 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
3. Show that the particle theory, in explaining one or more of the observed
optical phenomena, introduces improbable hypotheses, while the wave
theory does not.
But there is some improbability in supposing that bodies of different forms and of
various refractive powers should possess an equal force of inflection, as they ap-
pear to do in the production of these effects; and there is reason to conclude from
experiments, that such a force, if it existed, must extend to a very considerable
distance from the surfaces concerned, at least a quarter of an inch, and perhaps
more, which is a condition not easily reconciled with other phenomena.13
The phenomena of diffraction do not at all depend upon the nature, the mass, or
the shape of the body which intercepts the light, but only upon the size of the
intercepting body or upon the size of the aperture through which it passes. We
must, therefore, reject any hypothesis which assigns these phenomena to attractive
and repulsive forces whose action extends to a distance from the body as great as
that at which rays are inflected.14
12. Young, op. cit., pp. 359-360. Young's wave explanation of diffraction, which is rather
sketchy, is perhaps this: waves (e.g., those of sound and water) diverge when admitted into a wide
space through an aperture. Light is a wave motion. That is why it diverges similarly. It remained
for Fresnel to give a quantitative wave-theoretic explanation of this and other diffraction phe-
nomena. In one such phenomenon light bends around obstacles into the shadow as well as away
from it. Newton (to whom Young refers in the quotation above) observed the external fringes but
not the internal ones. This was one of his reasons for rejecting the wave theory (see section 7).
13. Ibid. Young proceeds to show how in explaining numbers of other phenomena the
particle theory, by contrast to the wave theory, employs dubious hypotheses. This strategy is also
evident in one of his earliest papers ("Outlines of Experiments and Inquiries Respecting Sound
and Light," Philosophical Transactions of the Royal Society, 1800), in which he also argues in
favor of the wave theory, though with somewhat less assurance (see pp. 613-616).
14. Fresnel, op. cit., p. 99.
LIGHT HYPOTHESES 77
Not only is the hypothesis of fits improbable because of its complexity, and diffi-
cult to reconcile with the facts in its consequences, but it does not even suffice in
explaining the phenomenon of the colored rings, for which it was imagined.15
whatever be the luminous origin: the light of the sun, the fixed stars, the planets
and their satellites, being all propagated with the same swiftness. This conclusion
must be allowed to present a formidable difficulty in the theory of emission.16
The difficulty is that if light consists of particles (with mass), then a massive
object such as a fixed star should exert a force on them that "would be
sufficient to destroy the whole momentum of the emitted molecules, and the
star would be invisible at great distances" (p. 300). Lloyd asserts that the only
way to explain the fact that the velocity we observe is the same for all lumi-
nous bodies is to adopt an hypothesis of Arago "that the molecules of light
are originally projected with very different velocities: but that among these
velocities is but one which is adapted to our organs of vision, and which
produces the sensation of light" (pp. 300-301). Lloyd notes that such a suppo-
sition has some support from discoveries of invisible rays of the spectrum.
But he concludes that the supposition is not "easily reconciled with hy-
potheses which we are able to frame respecting the nature of vision" (p. 301;
for additional discussion of this point, see my "Light Problems: Reply to
Chen," Studies in History and Philosophy of Science 21 (1990)). By contrast,
Lloyd asserts, the fact that the observed velocity of light is the same for all
luminous bodies is explained without difficulty by the wave theory:
4. Conclude that the wave theory is (very probably) true, while the particle
theory is (very probably) false.
At the beginning of his paper Lloyd makes explicit the strategy he will follow:
To take, in the first instance, a rapid survey of the several leading classes of optical
phenomena which the labours of experimental philosophers have wrought out in
such rich profusion, and afterward to examine how far they are reducible to one or
other of the two rival theories which have alone advanced any claim to our consid-
eration. This is, in fact, the only way in which the truth of a physical theory can be
established; and the argument in its favor is essentially cumulative.17
Having seen how the two theories explain (or fail to explain) various optical
phenomena, and having argued that the particle theorists introduce hy-
potheses in these explanations that are improbable, whereas wave theorists do
not—or do not to such a great extent—it is concluded that the wave theory is
probably true.
1. Assume that either theory T1 or theory T2 is correct, and give grounds for
such an assumption.
2. Show how T1 and T2 explain various observed phenomena.
3. Show that T2 in explaining one or more of these phenomena introduces
improbable hypotheses, whereas T1, does not.
4. Conclude that T1 is probably true.
Let us examine the three steps leading to the conclusion, beginning with the
second.
Step 2 introduces the idea of explanation. To conclude that T1 is probably
true, a theorist using this strategy must show that T1 is capable of explaining
certain phenomena (by producing the explanations). Following the method of
hypothesis, we might say that such a theorist seeks a theory T that will satis-
fy a
18. These additional features of an explanation are emphasized by Lloyd, op. cit., pp. 349-
350.
80 THEORIES OF LIGHT: PARTICLES VERSUS WAVES
what follows. Rather, arguing in the manner of the previous section, wave
theorists are committed to the idea that it is likely that one theory or the other
is true, and reasons are given for thinking that this is so. What sorts of
reasons are these?
Lloyd, we recall, argues "there are two distinct and intelligible ways of
conceiving" the observed motion of light, and that "in adopting one or the
other [modes of propagated motion] to account for the phenomena of light"
we obtain the particle or the wave theory. Accordingly, a part of Lloyd's
reasoning is explanatory: both theories will explain the fact that light exhibits
motion. But to this claim Lloyd adds importantly that "nature affords numer-
ous examples of each of these modes of propagated movement." Had others
been observed, presumably further ways of conceiving the propagation of
light would have been noted. Accordingly, if, as seems plausible, Lloyd means
that these are the only modes of communication known (or perhaps even just
the most prevalent ones), then, in addition to the explanatory reason, there is
an inductive one: light is observed to be communicated from one point to
another in a finite time; the modes of communication observed in nature
consist in the motion of bodies from one point to another and in the vibratory
motion of a medium; so it is reasonable to assume that light is either a
particle or a wave phenomenon.
Sometimes (as with Young) the claim that light is either a particle or a wave
phenomenon is defended by saying that these are the two leading theories
proposed by physicists. If, as I am assuming, this is to be understood as
providing some reason not simply to examine these theories but to think that
one or the other is correct, then, even in this case, the reasoning involves some
inductive steps: The fact that these are the leading theories proposed by
physicists, together perhaps with an implicit appeal to the reputations of
those supporting each theory (e.g., Newton vs. Huygens) and to their other
successes, provides some reason to think that one of these theories is true.
(Certainly a particle theorist such as Brougham had no qualms about defend-
ing his theory, at least in part, by appeal to the success of his authority
Newton. And, as noted in Essay 1, Young in an 1802 paper also appealed to
the authority of Newton in defense of some of the assumptions of the wave
theory.)
Since the claim that light is either a particle or a wave phenomenon is not
simply assumed, but grounds are given for it, and since it is recognized that
the resulting accounts are not the only possible ones,19 just the most likely, it
19. An alternative theory, which had some defenders until the early years of the nineteenth
century, was that light is produced by the rectilinear motion, rather than the vibrations, of a
fluid. According to most of these theorists this fluid consists of particles. However, unlike the
particle theorists, fluid theorists refused to offer mechanical explanations of the interactions of
light and matter, i.e., explanations in terms of attractive and repulsive forces obeying Newtonian
Chapter 4
Two-dimensional
Electromagnetic Waves
We further assume that the sources of the electromagnetic wave field are
also independent of x2 , hence
and
In the first set of equations (4.4) - (4.6) the field components Ê1 , Ê3 and Ĥ2
occur, while in the second set of equations (4.7) - (4.9) the field components
Ĥ1 , Ĥ3 and Ê2 occur.
Parallel polarization
When K̂1ext , K̂3ext and Jˆ2ext are equal to zero in all space, the only non-
zero field components are Ê1 , Ê3 and Ĥ2 . Since the electric field vector
is parallel to the (x1 , x3 )-plane of observation, this electromagnetic field is
called to be parallelly polarized. Since the magnetic field vector is parallel
to the x2 -direction, this electromagnetic field is also denoted as the case
of H-polarization. The system of equations (4.4) - (4.6) are the governing
electromagnetic field equations. In a source-free domain, we observe that
electric field components Ê1 , Ê3 follow from the magnetic field component
Ĥ2 as
Perpendicular polarization
When Jˆ1ext , Jˆ3ext and K̂2ext are equal to zero in all space, the only non-
zero field components are Ĥ1 , Ĥ3 and Ê2 . Since the electric field vector is
perpendicular to the (x1 , x3 )-plane of observation, this electromagnetic field
is called to be perpendicularly polarized. Since the electric field vector is par-
allel to the x2 -direction, this electromagnetic field is also denoted as the case
of E-polarization. The system of equations (4.7) - (4.9) are the governing
electromagnetic field equations. In a source-free domain, we observe that
magnetic field components Ĥ1 , Ĥ3 follow from the electric field component
Ê2 as
where
. . . . . . . . . .
.... ... .... ... .... ... .... ... .... ...
... . ... .. ... . ... . ... .
.. ... ... .. ... ... ... ... ... ...
... . .
. .
. .
. .
.
... ... .... .... .... .... .... .... .... ....
.. ... .... .... .... ....
........................................................................................... .....
...
.....
..
.....
..
.....
..
.....
..
....... .. . . . .
....... .... ...
... .
.......................... β ..
... ....
. ..
.
. ..
... ....
.
. ..
... ....
.
. ..
... ....
.
. ..
... ....
.. ... .. ... . .. . ..
α ..
... .....
.. ..
... ...
... ..
... ...
... ..
... ....
... ..
... ....
...
... ... ... ... ... ... ... ... ... ...
Figure 4.1. Planes of equal phase (−−−−−) and planes of equal amplitude (− − −).
86 two-dimensional electromagnetic waves
or
α · α − β · β = −ω 2 εµ , (4.22)
2α · β = ωσµ . (4.23)
Since ωσµ ≥ 0 we observe that the angle between the vectors α and β is less
than or equal to 12 π. In a lossless medium we have σ = 0 and then either
the case of α = 0 and β · β = ω 2 εµ occurs, or the case of α
= 0 and β
= 0
with α · β = 0 occurs. In the latter case, the planes of equal amplitudes are
perpendicular to the planes of equal phase. A plane wave is called uniform
when the vectors α and β have the same direction. A plane wave is called
non-uniform when the vectors α and β have different directions (see Fig. 4.1)
The time average Poynting’s vector of a plane wave in the frequency
domain is given by
∗
ST = 12 Re Ê × Ĥ = S 0 exp[−(γ1 +γ1∗ )x1 − (γ3 +γ3∗ )x3 ] , (4.24)
in which
∗
S 0 = 12 Re ê × ĥ . (4.25)
plane waves in a homogeneous medium 87
When the plane wave is uniform, the vectors α and β have the same
direction. Let s be the unit vector in the direction of α and β. Then, we
may write
α = αs, β = βs, γ = γs , (4.26)
in which (cf. Eq. (4.19))
γ = α + jβ (4.27)
is the propagation coefficient, α is the attenuation coefficient and β is the
phase coefficient. Substitution of γ = γs into Eq. (4.18) yields, with s·s = 1,
γ 2 = (σ + jωε)jωµ , (4.28)
which is identical to Eq. (3.8). The uniform plane wave is a one-dimensional
wave in the direction of s; the vector s = {s1 , 0, s3 } can have an arbitrary
orientation with respect to the chosen coordinate system (in Chapter 2 we
have either s = i3 or s = −i3 ).
The phase factor exp(−jβ1 x1 − jβ3 x3 ) is periodic in the direction of
s = s1 i1 + s3 i3 with period
2π
λ= . (4.29)
β
ê......
............
.......
.....
..... ....
s
.....
..... ..............
..... .......
..... .....
..... ..
......
..... ....
..... .....
..... ĥ .....
r
f
.....
....
..... .. ...........
i.1 .......
..... .
. .
. .....
.....
........ ......
..... .
....... x
.... . .
...... ..
.....
... .......
.... .......
.......
.. ......
... ..... .
.... ..
... ... ...
... ....... ....
.......
r ... .......
.................................................................
i3
O
When the plane wave is uniform, we may use γ = γs = γs1 i1 + γs3 i3 . Then,
Eqs. (4.30) - (4.32) can be written as
and
s1 ê1 + s3 ê3 = s · ê = 0 , (4.35)
where we have used the expression for the wave impedance
1
γ sµ 2
Z(s) = = . (4.36)
σ + sε σ + sε
For a parallelly polarized, uniform, plane wave, we observe that the electric
field strength ê = {ê1 , 0, ê3 }, the magnetic field strength ĥ = {0, ĥ2 , 0} and
the direction of propagation s = {s1 , 0, s3 } form a mutually perpendicular,
right-handed triad (see Fig. 4.2).
Further, in the frequency domain, the Poynting vector S 0 defined in
Eq. (4.25) becomes
∗
S0 = 1
2 Re ê × ĥ = 12 Re −ê3 ĥ∗2 i1 + ê1 ĥ∗2 i3
plane waves in a homogeneous medium 89
1 ∗
= 2 Re [Z(jω)] ĥ2 ĥ2 s . (4.37)
Hence, the power flow of the parallelly polarized, uniform, plane wave is in
the direction of propagation.
When the plane wave is uniform, we may use γ = γs = γs1 i1 + γs3 i3 . Then,
Eqs. (4.38) - (4.40) can be written as
and
s1 ĥ1 + s3 ĥ3 = s · ĥ = 0 , (4.43)
where we have used the expression for the wave admittance
1
γ σ + sε 2
Y (s) = = . (4.44)
sµ sµ
For a perpendicularly polarized, uniform, plane wave, we see that the electric
field strength ê = {0, ê2 , 0}, the magnetic field strength ĥ = {ĥ1 , 0, ĥ3 } and
the direction of propagation s = {s1 , 0, s3 } form a mutually perpendicular,
right-handed triad (see Fig. 4.3).
90 two-dimensional electromagnetic waves
...
s
...................
..
.....
.
..
......
.
.....
.....
....
x rf .....
.....ê
.......
....... ....
i.1 ......
......
....... .....
.....
.....
..... . .....
.... .... .. .....
....... ... ...
.... .
...
... .... .....
.....
..... .....
... .... .. .....
.... ..
.. ... ... .....
.... .. .... ..... .
.. .
...
... .... ................
...
.... .......
......
....... ..
ĥ
.......
r .. ......
.................................................................
i3
O
Hence, the power flow of the perpendicularly polarized, uniform, plane wave
is in the direction of propagation.
In the domain where interference occurs, the total field strengths are given
by
(1) (2)
Ê = Ê + Ê , (4.51)
(1) (2)
Ĥ = Ĥ + Ĥ . (4.52)
We now assume that both waves are uniform plane waves. Then
and
(2) (2) (2) (2)
γ1 = γs1 , γ3 = γs3 , (4.54)
where
1
γ = [(σ + sε)sµ] 2 . (4.55)
92 two-dimensional electromagnetic waves
i.1
...
..........
...
...
...
s(2) .....
...
... s(1) ....
...............
.......
... ...................
..... ... ..... .
.
..... ... ..
...
..... .... .....
..... .....
... .. ......
... ....... .. ....... ....
θ r θ
..... ..................................................................................................
.... i3
O
Figure 4.4. The propagation directions of the two uniform plane waves.
Without loss of generality we always can choose such a reference frame that
the two waves both propagate in the positive x1 -direction, while they have
opposite propagation directions in the x3 -direction (see Fig. 4.4). Let θ be
the angle between s(1) and i3 and π − θ be the angle between s(2) and i3 .
Then,
(1) (1)
s(1) = s1 i1 + s3 i3 = sin(θ)i1 + cos(θ)i3 , (4.56)
and
(2) (2)
s(2) = s1 i1 + s3 i3 = sin(θ)i1 − cos(θ)i3 . (4.57)
Substituting Eqs. (4.46) - (4.49) in Eqs. (4.51) - (4.52), we obtain
Ê = exp[−γ sin(θ)x1 ] ê(1) exp[−γ cos(θ)x3 ] + ê(2) exp[γ cos(θ)x3 ] , (4.58)
(1) (2)
Ĥ = exp[−γ sin(θ)x1 ] ĥ exp[−γ cos(θ)x3 ] + ĥ exp[γ cos(θ)x3 ] . (4.59)
in which we have to substitute the expressions for the electric and magnetic
field strengths of Eqs. (4.58) and (4.59). We arrive at
(1)∗ (2)∗
ST = 1
2 Re ê (1)
× ĥ + ê (2)
× ĥ
(2)∗
+ 12 Re ê(1) × ĥ exp[−2jω cos(θ)x3 /c]
(1)∗
+ ê (2)
× ĥ exp[2jω cos(θ)x3 /c] . (4.62)
The first two terms of the right-hand side of Eq. (4.62) represent the power
flows of the individual plane waves. In view of the obvious presence of the
third and fourth term we conclude that the superposition principle only
applies to the electromagnetic field strengths, but not to the power flow.
In order to quantify the electromagnetic power flow of two interfering
waves we consider the cases of parallel and perpendicular polarization sepa-
rately.
(1) (1)∗ (1) (2) (2)∗ (2)
ST = 1
2Y ê2 ê2 s + ê2 ê2 s
(1) (2)∗
+Y Re ê2 ê2 sin(θ) exp[−2jω cos(θ)x3 /c] i1 . (4.66)
From the results of these cases of polarization, it follows that the power
flow in the i3 -direction is position invariant, but the power flow in the i1 -
direction is periodic in x3 with period
λ
λinterference = , (4.67)
2 cos(θ)
where λ is the wavelength of the uniform plane waves, see Eq. (4.29).
S1 T
....
........
..
...
...
...
...
..
.... ..... .... .... .... .... .... .... .... ..... ..
... ... .. ... .... ... .... ... .... ... .... ... .... ... .... ... ... ... .. ...
.. .. .. ... .. ... .. ... .. ... .. ... .. ... .. .. .. .. ... ..
... .. ... .. ... .. ... .. ... .. .. .. .. .. .. .. ... .. .. ..
... ... ....
.
. .. ....
. .. ....
. .. ....
. .. ....
. .. ....
. .. ....
. ... ....
. ... .....
. ...
.
.
.. .. .
. .. .
. .. .
. .. .
. .
.. .
. .
.. .
. .. .
. .. . .. .
... ... ... .. ... .. .. .. ... .. ... .. ... .. ... .. ... ... ... ..
... ... ... ..
..
... ... ...
... ... ... ... ... ... ... ... ... ... ... ... ... ...
... .. ... ... .. .. ... .. ... .. ... .. ... .. ... .. ... ..
... ..
.
. ... ...
.
. ... ..
. ... ..
. .
. ..
. .
. ..
. .
. ..
. ... ..
. ... ..
.
. .
. ..
.
.
... ... . ... . ... . ... . ... . ... . ... ...
.. .. ... .. ... ... ... ... ... .. ...
... .. ... .. ... .. ... ... ... ... ...
...
... .. ...
...
... .... ... ... ... .... ... ... ... ... ... ... ... ... ... ... .... ...
... .. ... ... ... ... ... .... ... .... ... .... ... .... ... .... ... .. ... ....
... .. ... ... ... .. . .
.
... .. .. .
.
... .. .. .
... .. .. .
... .. .. .
... .. .. .
... . .. .
. .
.
.. ... .. .. ... ..
... .. ... .. ... .. .. .. .. .. ... .. ... .. ... .. ... ... ... ...
.... .... ...... ....... ....... ....... ...... ...... ...... ....
............................................ x3
...................... .....................
λ
2 cos(θ)
i
Ê = êi exp(−γ1i x1 − γ3i x3 ) , (4.68)
i i
Ĥ = ĥ exp(−γ1i x1 − γ3i x3 ) , (4.69)
r
Ê = êr exp(−γ1r x1 − γ3r x3 ) , (4.70)
r r
Ĥ = ĥ exp(−γ1r x1 − γ3r x3 ) , (4.71)
in which
(γ1i )2 + (γ3i )2 = (γ1r )2 + (γ3r )2 = (σ + sε)sµ . (4.72)
From this equation, it is observed that for given values of γ1i two solutions
of γ3i exist. In view of causality with respect to the incident wave we take
1
γ3i = (σ + sε)sµ − (γ1i )2 2
, (4.73)
σ, ε, µ electrically
impenetrable
medium
.................
............
.... ........
.......
........
.......
.......
reflected wave ........
.
............................
i3
.........................
incident wave .. .
....... ...
........
.
...
.........
.
...
.......
......
x3 < 0 x3 = 0 x3 > 0
1
γ3r = − (σ + sε)sµ − (γ1r )2 2
, (4.74)
where Re[γ3r ] ≤ 0. The total field in the half-space −∞ < x3 < 0 is obtained
as
i r i r
Ê = Ê + Ê , Ĥ = Ĥ + Ĥ . (4.75)
or
Parallel polarization
In the case of parallel polarization, Ê1 , Ê3 and Ĥ2 are the non-zero elec-
tromagnetic field components. We only have to meet the boundary condition
of Eq. (4.77) for Ê1 . Using Eq. (4.10) and (4.75), the boundary condition at
x3 = 0 becomes
lim ∂3 (Ĥ2i + Ĥ2r ) = 0 . (4.79)
x3 ↑0
But, in that case, the boundary condition of Eq. (4.80) also requires
The two other non-zero vectors follow with the help of Eqs. (4.10) - (4.11)
as
Perpendicular polarization
In the case of perpendicular polarization, Ê2 , Ĥ1 and Ĥ3 are the non-
zero electromagnetic field components. We only have to meet the boundary
condition of Eq. (4.78) for Ê2 . Substituting the plane-wave representations
of Eqs. (4.68) and (4.70) into Eq. (4.78), we obtain
The two other non-zero vectors follow with the help of Eqs. (4.13) - (4.14)
as
σ, ε, µ electrically
impenetrable
medium
.......
....... sr si
..................
............. .........................
... ....... ... .
........ ....... ...
............... ..........
θr ....
... ....... .............
. ... .........................................................................
.. .........
... ...... .
θi .
...... .... .. .
. .
i3
.......
.......
.......
θr = θi . (4.94)
100 two-dimensional electromagnetic waves
In the case of steady-state (s = jω), the total wave field, being the su-
perposition of the incident wave and the reflected wave, has the character of
a travelling wave in the i1 -direction and a standing wave in the i3 -direction.
This phenomenon is clearly observable when the medium in the half-space
−∞ < x3 < 0 is lossless. Then
where c = (εµ)− 2 . Combining all the previous results, we conclude that the
1
Ĥ1 = 0 , (4.100)
Ĥ2 = 2ĥi2 exp(−jωsi1 x1 /c) cos(ωsi3 x3 /c) , (4.101)
Ĥ3 = 0 , (4.102)
Ê1 = 0 , (4.103)
Ê2 = −2jêi2 exp(−jωsi1 x1 /c) sin(ωsi3 x3 /c) , (4.104)
Ê3 = 0 , (4.105)
When a wave is incident upon the interface between two different media,
the wave is partly reflected by this interface and partly transmitted through
this interface. In the case that the incident wave is a plane wave and the
interface is a plane boundary between two semi-infinite media, the reflected
and transmitted waves are plane waves as well. Then, the pertaining reflec-
tion problem can be solved with the theory of plane waves.
We take the Cartesian coordinate system in such a way that the plane
x3 = 0 coincides with the interface between the two media. The half-space
−∞ < x3 < 0 consists of a homogeneous medium with constitutive constants
σ (1) , ε(1) and µ(1) and the half-space 0 < x3 < ∞ consists of a homogeneous
medium with constitutive constants σ (2) , ε(2) and µ(2) (see Fig. 4.8).
The incident (plane) wave is given by (cf. Eqs. (4.16) - (4.17))
i
Ê = êi exp(−γ1i x1 − γ3i x3 ) , (4.109)
i i
Ĥ = ĥ exp(−γ1i x1 − γ3i x3 ) , (4.110)
r
Ê = êr exp(−γ1r x1 − γ3r x3 ) , (4.111)
r r
Ĥ = ĥ exp(−γ1r x1 − γ3r x3 ) , (4.112)
in which
1
γ3i = (σ (1) + sε(1) )sµ(1) − (γ1i )2
2
, (4.114)
102 two-dimensional electromagnetic waves
.................
............
... ........
transmitted wave
....... ...............
........ ..................
....... ............. ....
....... .............
.............
reflected wave ........
. ....
...
...
...
..
............................
i3
.........................
incident wave .. .
....... ...
.......
.
...
..........
..
...
......
.
........
.......
.......
.......
.
...
..........
.
x3 < 0 x3 = 0 x3 > 0
t
Ê = êt exp(−γ1t x1 − γ3t x3 ) , (4.117)
t t
Ĥ = ĥ exp(−γ1t x1 − γ3t x3 ) , (4.118)
reflection and transmission of a plane wave 103
in which
(γ1t )2 + (γ3t )2 = (σ (2) + sε(2) )sµ(2) . (4.119)
In view of causality with respect to the transmitted wave we take
1
γ3t = (σ (2) + sε(2) )sµ(2) − (γ1t )2 2
, (4.120)
where Re[γ3t ] ≥ 0.
At x3 = 0 the electromagnetic field has to satisfy the boundary conditions
that the tangential components of the electric and magnetic field strengths
have to be continuous, i.e.,
i r t
lim i3 × (Ê + Ê ) = lim i3 × Ê , (4.121)
x3 ↑0 x3 ↓0
i r t
lim i3 × (Ĥ + Ĥ ) = lim i3 × Ĥ , (4.122)
x3 ↑0 x3 ↓0
or
Parallel polarization
In the case of parallel polarization, Ê1 , Ê3 and Ĥ2 are the non-zero
electromagnetic field components. We only have to meet the boundary con-
ditions of Eq. (4.123) for Ê1 and Eq. (4.126) for Ĥ2 . Using Eqs. (4.10) and
(4.116) the boundary conditions at x3 = 0 become
1 1
lim ∂3 (Ĥ2i + Ĥ2r ) = lim ∂3 Ĥ2t , (4.127)
x3 ↑0 σ (1) + sε(1) x3 ↓0 σ (2) + sε(2)
lim (Ĥ2i + Ĥ2r ) = lim Ĥ2t . (4.128)
x3 ↑0 x3 ↓0
104 two-dimensional electromagnetic waves
γ3i γ3r
ĥi exp(−γ1i x1 ) + ĥr exp(−γ1r x1 )
σ (1) + sε(1) 2 σ (1) + sε(1) 2
γ3t
= ĥt exp(−γ1t x1 ) , (4.129)
σ (2) + sε(2) 2
ĥi2 exp(−γ1i x1 ) + ĥr2 exp(−γ1r x1 ) = ĥt2 exp(−γ1t x1 ) . (4.130)
1
γ3t = (σ (2) + sε(2) )sµ(2) − (γ1i )2 2
. (4.133)
But, in that case, the boundary conditions of Eqs. (4.129) - (4.130) also
require
Let us now introduce the reflection coefficient R = R (s) and the transmis-
sion coefficient T = T (s) via the relations
Substituting the latter two expressions into Eqs. (4.134) - (4.135), divid-
ing the resulting two equations by the non-zero quantity ĥi2 , and using
Eq. (4.132), we arrive at the system of two algebraic equations,
γ3i γ3t
(1) (1)
−
R = σ +i sε σ (2) + sε(2) , (4.140)
γ3 γ3t
+
σ (1) + sε(1) σ (2) + sε(2)
γ3i
2
T = σ (1) + sε(1) . (4.141)
γ3i γ3t
+
σ (1) + sε(1) σ (2) + sε(2)
At this moment, the parallelly polarized reflected and transmitted wave are
determined completely.
In the special case of normal incidence, we have γ1i = 0. Then, we have
arrived at the reflection and transmission coefficients of the one-dimensional
waves (cf. Eqs. (3.95) derived from a magnetic field analysis), viz.,
Z (1) − Z (2)
R = , (4.142)
Z (1) + Z (2)
2Z (1)
T = , (4.143)
Z (1) + Z (2)
in which
1
(1) γ (1) sµ(1) 2
Z = (1) = (4.144)
σ + sε(1) σ (1) + sε(1)
106 two-dimensional electromagnetic waves
and 1
(2) γ (2) sµ(2) 2
Z = (2) = (4.145)
σ + sε(2) σ (2) + sε(2)
are the wave impedances in medium (1) and (2), respectively.
Perpendicular polarization
In the case of perpendicular polarization, Ê2 , Ĥ1 and Ĥ3 are the non-
zero electromagnetic field components. We only have to meet the boundary
conditions of Eq. (4.124) for Ê2 and Eq. (4.125) for Ĥ1 . Using Eqs. (4.13)
and (4.116) the boundary conditions at x3 = 0 become
1
γ3t = (σ (2) + sε(2) )sµ(2) − (γ1i )2 2
. (4.152)
reflection and transmission of a plane wave 107
But, in that case, the boundary conditions of Eqs. (4.148) - (4.149) also
require
Let us now introduce the reflection coefficient R⊥ = R⊥ (s) and the trans-
mission coefficient T⊥ = T⊥ (s) via the relations
Substituting the latter two expressions into Eqs. (4.153) - (4.154), divid-
ing the resulting two equations by the non-zero quantity êi2 , and using
Eq. (4.151) we arrive at the system of two algebraic equations,
1 + R⊥ = T⊥ , (4.157)
γ3i γ3i γ3t
− (1) R⊥ = T⊥ , (4.158)
sµ(1) sµ sµ(2)
from which R⊥ and T⊥ are obtained as
γ3i γ3t
−
µ(1) µ(2)
R⊥ = i , (4.159)
γ3 γ3t
+
µ(1) µ(2)
γ3i
2
µ(1)
T⊥ = . (4.160)
γ3i γ3t
+
µ(1) µ(2)
Y (1) − Y (2)
R⊥ = , (4.161)
Y (1) + Y (2)
2Y (1)
T⊥ = , (4.162)
Y (1) + Y (2)
in which 1
σ (1) + sε(1) σ (1) + sε(1) 2
Y (1) = = (4.163)
γ (1) sµ(1)
and 1
(2) σ (2) + sε(2) σ (2) + sε(2) 2
Y = = (4.164)
γ (2) sµ(2)
are the wave admittances in medium (1) and (2), respectively.
θr = θi . (4.167)
reflection and transmission of a plane wave 109
This relation is identical to Eq. (4.94) of Section 4.3 dealing with the reflec-
tion by an electrically impenetrable half-space. This relation is called Snell’s
law of reflection.
Subsequently, we investigate the conditions under which the transmitted
wave is a uniform plane wave. Let us assume that
1
γ t = γ1t i1 + γ3t i3 = [(σ (2) + jωε(2) )jωµ(2) ] 2 st , (4.168)
In the general case, real values of si1 and st1 do not meet this equation, and
in general the transmitted wave is a non-uniform plane wave. However, for
lossless media, Eq. (4.169) can be satisfied for real values of si1 and st1 .
Lossless media
In the case of lossless media, we introduce the (real-valued) index of
refraction n(1) of medium (1) as
1
n(1) = c0 [ε(1) µ(1) ] 2 , (4.170)
The angle θt between the vectors st and i3 is called the angle of transmission
(see Fig. 4.9). From Eq. (4.172) we obtain Snell’s law of transmission (also
called Snell’s law of refraction),
⎧ ⎫1
1 ⎨ n(1)
2⎬ 2
γ3t = jω ε(2) µ(2) 1 − (2) sin(θ )
i
2
, (4.174)
⎩ n ⎭
n(2)
0 ≤ sin(θi ) ≤ for n(1) > n(2) , (4.175)
n(1)
1 1
µ(1) µ(2)
cos(θi ) −
2 2
ε(1) ε(2)
cos(θt )
R = 1 1 , (4.176)
µ(1) 2 µ(2) 2
ε(1)
cos(θi ) + ε(2)
cos(θt )
1
µ(1) 2
2 ε(1)
cos(θi )
T = 1 1 , (4.177)
µ(1) 2 µ(2) 2
ε(1)
cos(θi ) + ε(2)
cos(θt )
1 1
ε(1) ε(2)
µ(1)
2
cos(θi ) − µ(2)
2
cos(θt )
R⊥ = 1 1 , (4.178)
ε(1) 2 ε(2) 2
µ(1)
cos(θi ) + µ(2)
cos(θt )
1
ε(1) 2
2 µ(1)
cos(θi )
T⊥ = 1 1 , (4.179)
ε(1) 2 ε(2) 2
µ(1)
cos(θi ) + µ(2)
cos(θt )
1.0 ..
...
1.0 .....
.......
... . ......
.....
|R | ...
..... |R⊥ |
.
. .. .
. . ...
. .... ...
.
...... . ...
. ..
..
...... . . ..
........ .
. ... ...
..... . ... ...
..... . . ..
....... .
. ... ..
.
ε(2) ....... . ε (2) . ..... ...
.
ε(1)
=8 ... .
ε(1) . . . .
=8
. .
..
.... ...
0.5 ..... 0.5 ..... ..
..
. . . . . .... . . . . . . .
. .
. . ..... . .
... ..
. . .
. . ..... ...
. . ..... . .... ...
... .. . .. .. ...
.
. ... .
.. .. . ...... ..... ...
. 4 ..
....
.
. . .. ...... ....
. ... .. . ...... ...... ...... .... ....
...... ...... ...... .... . ....
.. ......
......
......
.
.
4 ... . .
... ... . .
......
.
....
...... .. .. . .....
. . ......
0.25 ......
...... .
.
... .. .
... ..
.
0.25 .........
.......
...... 2
..... ... ... . ... .
......
.............................. ..
..... 2 . .
. .. . .
.
. .
. ..............................
...........
........
....... .... .... ... .
...... .. . . .
...... ... .... . ... .
......
..... ..... ... .
..
.... . . .. .
..... ...... ... .. . .
....... .... .
0 0
0 30 60 90 0 30 60 90
- θi - θi
1.0 .
.
...
.. .
.
.... 1.0 .
.
.
...
.
...
...
(1)
.
. ..
.
. (1) . .
.
.. ..
.
|R | ε ε
.
... ...
ε(2)
.
.
= 8 4 ....
....
...
....
2 |R⊥ | ε(2)
.
.
= 8 4 .
..
.
...
...
...
2
. . .
. . ..
.
... ... . .
... ...
6 .
. .
...
...
...
6 .
.
.
...
...
....
0.75 . ... .... 0.75 . .
.
. . .. . ... ...
.... ... . . ...
.
... . ... ..
. ... .... ..
. ..
.
.
. .
. . .. . . ...
. ... ... ... ...
. .. ... . . ...
. ..
. . ... ..
.
. .
.. ... . .
... ...
. ... . ..
0.5 .... .. 0.5 . ... ...
. .
.
. ..
.
.
. .
. ..
.
.
. .
. ... ... ... ...
. . . ... .
... ...
. .. .. ..
.
. .
.
..
.
. . ...... ...
.
. .. ... ..... ...
...... ...... . . .. ... ...... .....
...... . . .. ...
...... . ... ..
...
.. ... .... ..
.
....
0.25 .... .
. ....
.
...
.
..
...
..
0.25 ......
.....
.....
. . .. ... ..
.
. .
.. .. .
..... .......
..................... . . .. . ............
........ . .. ...
.............. . ...
.. ............. ..
. .. .
...
.. . ... ...... .
...
.. .... .... ...
. ..... ... ...
...
0 .
0
0 30 60 90 0 30 60 90
- θi - θi
Brewster angle
It is possible that the Fresnel reflection coefficient vanishes for a partic-
ular value of the angle of incidence. The pertaining angle of incidence is
i
called the Brewster angle. In the case of parallel polarization, this angle θB
follows from
1 1
µ(1) 2
µ(2) 2
cos(θ ) − i
cos(θt ) = 0 (4.180)
ε(1) ε(2)
⎛ ⎞12
µ(2) ε(2)
(1) − ε(1)
i
tan(θB )= ⎝µ ⎠ , for parallel polarization . (4.181)
(1) µ(2)
ε
ε(2)
− µ(1)
For dielectric media, where µ(1) = µ(2) = µ0 , this Brewster angle follows
from
1
i ε(2) 2
(4.182)
tan(θB ) = , for parallel polarization .
ε(1)
1 1
ε(1) 2
ε(2) 2
cos(θ ) − i
cos(θt ) = 0 (4.183)
µ(1) µ(2)
⎛ µ(2)
⎞12
ε(2)
(1) − µ(1)
i
tan(θB )= ⎝ε ⎠ , for perpendicular polarization . (4.184)
µ(1) ε(2)
µ(2)
− ε(1)
For dielectric media, where µ(1) = µ(2) = µ0 , this Brewster angle does not
exist; the right-hand side of Eq. (4.184) is not real valued.
114 two-dimensional electromagnetic waves
Total reflection
If n(1) > n(2) and the angle of incidence is larger than the critical angle,
i n(2)
θ > θci = arcsin , (4.185)
n(1)
⎧ 2 ⎫1
1 ⎨ n(1) ⎬2
γ3t = ω ε(2) µ(2) 2 sin(θ i
) − 1 . (4.186)
⎩ n(2) ⎭
In this case γ3t is real valued, while on account of Eq. (4.131), or (4.150), γ1t is
imaginary valued; hence, the transmitted wave is a non-uniform plane wave
(Fig. 4.12). The planes of equal phase are perpendicular to the interface
between the two different media, while the planes of equal amplitude are
parallel to the interface. Since we are dealing with lossless media, the planes
of equal amplitude are perpendicular to the planes of equal phase. Moreover,
since σ (1) = σ (2) = 0, γ3i is imaginary valued and γ3t is real valued, we
observe from Eq. (4.140) that |R | = 1, while from Eq. (4.159) we observe
that |R⊥ | = 1. The reflected wave has an amplitude identical to the one of
the incident wave; only the phase is different. This phenomenon is called
total reflection (see Fig. 4.11). There exists an electromagnetic field in the
half-space 0 < x3 < ∞, but there is no time average power transported in
the i3 -direction. This is easily observed from (cf. Eq. (4.25))
t t∗ t∗
Re[(Ê × Ĥ ) · i3 ] = Re[(êt × ĥ ) · i3 ] exp(−2γ3t x3 ) (4.187)
....
.......
...... ....... ....... .......
......
.
s........r. ....... ...... ....... ....... ....... .......... si
s.r.................... n(2) t
(1) s .................. .......
.....
...
.si
................. ....
..
.....
............
.....
... ....... .. ..... ..
..... ..
..
........... .......
...
. .... n . .
.
....... . ...
... ...... ..
... . ..... ..... .
.......... . . .
..... ... ... ..
.
. .
.
. .. ... ... ... ... ...
.. .....
....
.
.... . . .... ... ... ... ..
.. .....
..... ..... .... ... .... ..
..... .. ..... ... ... .. ...
... ..... .... ......... ... .. ... ... .... .
. ..
. ..... .. .....
..... .. ..... .. . ... ..... ...
. . ... ... .
... ............................................................................
. . .. ..
. ... . ... . ..
.. ..
. .. ..
.. .. ...............................................................................
.... .
...
..
..
.
.....
..
.. i3 ...
. ...
.. .
.... .
.
..
i3 .
.
..
... ..... . ..
.. ... ... . ..
..
.. . ... .. ...
..... . . .... . .
... ....
.. ... (2)
.
..... ...
. ...
..
.
(2)
.
.....
....
... n ....... ....
... ..
..... .....
.
... n ....... ....
... ..
.....
..... ...... (1) . .. .. (1) .
.....
. ......
. . n ..
...... ........ n ..
......
...... ...
.. . ....... ...
..
.. ...
..... .... ....... ....... .... ....... .......
1 ..
1
...
n(2) n(2)
(a) si1 = n(1)
(b) si1 > n(1)
Figure 4.11. Total reflection of a uniform plane wave, when n(1) > n(2) .
γ1t .....
............
γ1t
... ..... ...
...... ... ......
........ ... ........
....... ... ... ...
............ ...
. ..... ... ... ...
.....
..... .... ...
... ....
..... .. ... ..
.....
..... ... ... ...
..... .... ... ....
..... .. ... ..
..... ...
..... ... ... ...
..... .... ... ....
..... .. ... ..
..... ...
.....
..... ... ... ...
..... .... ... ....
..... .. ... ..............................................................................
..... ...
..... ... .....
.....
.....
.....
.....
...
...
...
...
γ3t
... ............................. ...... .............................
............... ..............
......... ..
θi = θci ..........
. ..
i3 θi > θci
..
...
...
.
i3
..... ...
..... ...
..... .
......
. ...
...... ...
..... ...
..... ...
..... ....
.....
.
.
..
....... ...
..... ...
..... ...
..... ...
......
. ...
.....
... ...
..... ...
...
....
.
..
...
...
...
Exercise 4.1
A parallelly polarized plane wave in free space has an electric field strength
E = (3i1 − 4i3 ) sin[3π × 109 t − 2π(4x1 + 3x3 )] .
(a) Find the frequency and the angle of propagation of this wave.
Exercise 4.2
Given the electric fields of two parallelly polarized uniform plane waves,
the total field of which is given by Eq. (4.58) with γ = jω(ε0 µ0 )1/2 and
|ê(1) | = |ê(2) |. Show that this total electric field corresponds to a wave
standing in the x3 -direction and propagating in the x1 -direction. Hint: use
the fact that s · ê = 0.
Exercise 4.3
A uniform perpendicularly polarized, steady state wave is incident from a
lossless halfspace upon the plane boundary surface x3 = 0 of a perfect con-
ductor. The incoming wave is uniform, and its electric field is expressed as
i
Ê = E0 exp[−(γ1i x1 + γ3i x3 )]i2 , for x3 < 0. Draw the amplitude of the total
i r
electric field, Ê = Ê + Ê , as a function of x3 .
Exercise 4.4
A uniform plane wave in free space, having an electric field strength
√ √
E i = ( 12 i1 − 12 3i3 ) E0 cos[6π × 109 t − 10π( 3x1 + x3 )] ,
is incident on the interface x3 = 0 between free space and a lossless dielectric
medium, σ (2) = 0, ε(2) = 1.5ε0 and µ = µ0 .
(a) Does this electric field correspond to a parallelly or a perpendicularly
polarized wave, why?
(b) Find the angle of incidence θi , the angle of transmission θt , and the
vectors si and st along the directions of propagation of the incident
and the transmitted wave, respectively.
exercises and problems 117
( c ) Find the expressions for the reflected and transmitted electric field
strength.
(d) Draw the electric field vector and the direction of propagation of the
incident, reflected and transmitted electric field in the configuration.
Exercise 4.5
A perpendicularly polarized, plane electromagnetic wave is incident from
medium (1) with ε(1) = 4ε0 , σ (1) = 0 and µ(1) = µ0 , upon the plane boundary
of medium (2) with ε(2) = ε0 , σ (2) = 0 and µ(2) = µ0 , at an angle of incidence
θi = π3 with the i3 -axis, i3 being the normal of the plane interface.
(a) What is the critical angle?
1
(b) Find γ1i and γ3i in terms of γ0 = jω(ε0 µ0 ) 2 .
(d) Find in medium (2) the distance d at which the electric field strength
decays a factor of exp(−1).
Exercise 4.6
If you are under water looking towards the water surface, only a small circle
in the water surface is light while the rest of the surface is dark. Explain
why there is only a cone of light connected with the circle. Calculate the
angle θ of the cone (see Fig. 4.13). For optical frequencies the constitutive
parameters of water are ε = 1.77ε0 , σ = 0 and µ = µ0 .
air
..... ...
..... .....
..... .....
.....
..... ..
......
. water
..... ...
.....
.....
θ
..... .................
..... .....
..........
.
Exercise 4.7
Calculate the critical angle θci and the Brewster angle θB i of an air/glass
Problem 4.1
A uniform plane wave is incident on a magnetically impenetrable wall at an
angle of incidence of θi = π6 . Give for both the parallelly and perpendicularly
polarized waves
(a) the boundary conditions at the interface
(b) the expressions for the total electric and magnetic field strengths.
Problem 4.2
A parallel polarized plane wave is incident from medium (1) on the plane
boundary between medium (1) and medium (2). Both media are perfect
dielectrics. You know that for an angle of incidence larger than the critical
angle no time averaged power is transmitted into medium (2); also the re-
flected power in medium (1) is zero for an angle of incidence equal to the
Brewster angle. Now imagine a situation where the Brewster angle is larger
than the critical angle. Then the incident wave will not be reflected, because
of the Brewster angle of incidence, and will not be transmitted, because of
post-critical incidence. Solve this paradox.
Problem 4.3
A sinusoidally varying uniform plane wave is incident on a plane interface
between two lossless dielectric media at an angle larger than the critical
angle, θi > θci .
(a) Find the Fresnel reflection and transmission coefficients.
(b) Find the phase shift of the reflected wave at the interface.
Problem 4.4
Indicate in the four graphs of Fig. 4.10 the Brewster angles and the critical
angles. Give the expressions for the Brewster angles, both for parallel and
perpendicular polarization, in case ε(1) = ε(2) while µ(1)
= µ(2) .
Problem 4.5
A parallelly polarized plane uniform wave in free space (n(1) = 1) is incident √
at an angle θi on a slab of lossless dielectric with index of refraction n(2) = 3
and thickness d = 3 m, while on the other side there is another lossless
dielectric medium that extends to infinity with index of refraction n(3) = 3
(see Fig. 4.14). Determine the angle of incidence in case the wave is totally
transmitted through both interfaces.
d -
Problem 4.6
A lossless dielectric slab, with thickness d = λ4 m and ε = 2.25ε0 , is on one
side coated with an electrically perfectly conducting material. A uniform
plane wave in free space (n(1) = 1) impinges on the dielectric slab at normal
incidence (see Fig. 4.15).
n(1) n(2)
d -
( c ) Find the electric and magnetic field strengths in the two regions by sub-
stituting the general expressions of the field strengths into the bound-
ary conditions at both interfaces.
Problem 4.7
A parallelly polarized plane uniform wave in a lossless dielectric medium (1)
with ε(1) = 12ε0 is incident at an angle θi = π6 on a lossless dielectric slab
with index of refraction n(2) = 2 and thickness d = 1 m, while on the other
side there is another lossless dielectric medium that extends to infinity with
index of refraction n(3) (see Fig. 4.14). Determine the maximum index of
refraction n(3) , for which the wave is totally reflected back into medium (1).
Chapter 5
Electromagnetic Rays in a
Two-dimensional Medium
are independent of the x2 -coordinate (cf. Eq. (4.1)), and that the sources
of the electromagnetic wave field are also independent of x2 (cf. Eq. (4.2)).
Then, the generated electromagnetic field will be independent of x2 (cf.
Eq. (4.3)). From Chapter 4 we know that in this case the resulting total
electromagnetic field may be decomposed in two constituents, viz. the paral-
lelly and perpendicularly polarized fields. Therefore, we separately consider
the two cases of parallel and perpendicular polarization.
L(x1 , x3 ) L(x1 , x3 )
E = e[t − ] H[t − ], (5.6)
c0 c0
L(x1 , x3 ) L(x1 , x3 )
H = h[t − ] H[t − ]. (5.7)
c0 c0
homogeneous, lossless medium 123
For a given time instant t, the electric and magnetic field strengths in all
points of the planes given by
L(x1 , x3 ) (s1 x1 + s3 x3 )
= = constant (5.8)
c0 c
have an equal value. The plane
L(x1 , x3 ) (s1 x1 + s3 x3 )
= =t (5.9)
c0 c
is in a certain respect special: here the values of the electric and the magnetic
field strengths are given by E = e(0) and H = h(0), respectively. This
implies that at this plane the disturbance of the source has just arrived. For
this reason the plane given in Eq. (5.9) is called the wavefront. In Fig. 5.1
the wavefront is depicted for some time instants. Obviously, plane waves
have plane wavefronts. These move away from the source.
Ray
trajectory
Wavefronts
Note that the results are equal to the results obtained above. However, from
the latter two equations it may be seen that the time domain wavefield at
the wavefront follows from the behavior of its complex frequency domain
counterpart for s → ∞.
In analogy with Eqs. (5.4) and (5.5), we now investigate a solution of the
form
s
Ê1 = ê1 exp[− L(x1 , x3 )] , (5.17)
c0
s
Ê3 = ê3 exp[− L(x1 , x3 )] , (5.18)
c0
s
Ĥ2 = ĥ2 exp[− L(x1 , x3 )] . (5.19)
c0
s
∂3 ĥ2 − (∂3 L)ĥ2 + (σ+sε)ê1 = 0 , (5.20)
c0
s
−∂1 ĥ2 + (∂1 L)ĥ2 + (σ+sε)ê3 = 0 , (5.21)
c0
s s
∂3 ê1 − (∂3 L)ê1 − ∂1 ê3 + (∂1 L)ê3 + sµĥ2 = 0 . (5.22)
c0 c0
without the factor s and retain the terms with the factor s. In the resulting
equations we divide by s/c0 . Then we obtain
Away from the wavefront the approximation remains valid as long as ê1 , ê3 ,
ĥ2 , and L, are slowly varying functions in space. This requires that ε, σ, and
µ, are slowly varying functions of position. The resulting approximation is
called the ray approximation, and the corresponding approximate solutions
of the form of Eqs. (5.17) - (5.19) are called electromagnetic rays. In each
point of a domain where the ray approximation is valid, Eqs. (5.23) - (5.25)
are linear algebraic equations for ê1 , ê3 , and ĥ2 , in which ∂1 L and ∂3 L have to
be determined in such a way that non-zero solutions exist. In order to derive
the condition to be satisfied by the partial derivatives of L, we eliminate ê1
and ê3 . Substituting Eqs. (5.23) and (5.24) into (5.25) yields
L(x1 , x3 )
= t(3)
L(x1 , x3 ) (2) c 0
=t
c0
L(x1 , x3 )
= t(1)
c0
s(1) (2)
s-
1
s(3)
q
Ray
trajectory
Wavefronts
so that
s1 ê1 + s3 ê3 = s · ê = 0 , (5.33)
while from Eq. (5.27) it follows that
1
n = c0 (εµ) 2 . (5.34)
In analogy with Eqs. (5.4) and (5.5), we now investigate a solution of the
form
perpendicular polarization 129
s
Ê2 = ê2 exp[− L(x1 , x3 )] , (5.38)
c0
s
Ĥ1 = ĥ1 exp[− L(x1 , x3 )] , (5.39)
c0
s
Ĥ3 = ĥ3 exp[− L(x1 , x3 )] . (5.40)
c0
s
−∂3 ê2 +(∂3 L)ê2 + sµĥ1 = 0 , (5.41)
c0
s
∂1 ê2 − (∂1 L)ê2 + sµĥ3 = 0 , (5.42)
c0
s s
−∂3 ĥ1 + (∂3 L)ĥ1 + ∂1 ĥ3 − (∂1 L)ĥ3 + (σ+sε)ê2 = 0 . (5.43)
c0 c0
For most inhomogeneous media it is impossible to solve these equations ex-
actly, and only approximate solutions may be obtained. Since the wavefront
is the most interesting part of a propagating field, it makes sense to derive
an approximation that focusses on the bahaviour at the wavefront. It fol-
lows from Tauber’s theorem that the behavior at the wavefront is found by
letting s → ∞. In that case we may neglect in Eqs. (5.41) - (5.43) the terms
without the factor s and retain the terms with the factor s. In the resulting
equations we divide by s/c0 . Then we obtain
Away from the wavefront the approximation remains valid as long as ê2 , ĥ1 ,
ĥ3 , and L, are slowly varying functions in space. This requires that ε, σ, and
µ, are slowly varying functions of position. The resulting approximation is
called the ray approximation, and the corresponding approximate solutions
of the form of Eqs. (5.38) - (5.40) are called electromagnetic rays. In each
130 electromagnetic rays in a two-dimensional medium
point of a domain where the ray approximation is valid, Eqs. (5.44) - (5.46)
are linear algebraic equations for ê2 , ĥ1 , and ĥ3 , in which ∂1 L and ∂3 L
have to be determined in such a way that non-zero solutions exist. In order
to derive the condition to be satisfied by the partial derivatives of L, we
eliminate ĥ1 and ĥ3 . Substituting Eqs. (5.44) and (5.45) into (5.46) yields
L(x1 , x3 )
=t (5.49)
c0
so that
s1 ĥ1 + s3 ĥ3 = s · ĥ = 0 , (5.54)
1
n = c0 (εµ) 2 . (5.55)
τ = ∂l x = ∂l x1 i1 + ∂l x3 i3 . (5.57)
Since Eq. (5.56) represents the ray trajectory, we simply have s = τ , hence,
s1 = ∂l x1 , s3 = ∂l x3 . (5.58)
n ∂l x1 (l) = ∂1 L , (5.59)
n ∂l x3 (l) = ∂3 L . (5.60)
τ =s
........
t ......................
................................................................
................. ..........
...
..
. .
............ .......... ........
.
........ .. .......
..
..
........
. ..
.
.
......
......
...
......
. ..
. .....
.....
....
.
.
....
.
..
.
.
.....
.
.
.
.
.
.
.
.
. .....
....
....
L = L(xQ )
... . ....
....
. .
. .... ....
.
.... . .... .....
.
...
.
.. ..
. .... ......
.
.. .
. ... ......
.
... .. ... ..........
Q t
..
.. .
. ... .....
..... ...
....
....
....
...
..
x ......
.......
...... .....
...
.
.....
. ..
.
. .
......... ...
...
.
.... .. ..
.... .. ...... ...
..
. .
. ...... ...
...
..
... .
.. ...
... .
L = L(xP ) ..
.....
.
.
.
.
.
.
.
.
..
...
...
...
...
...... . .
...... ....
. .
.
. ...
...... ... .. ...
...... . . ...
tP
...... .... ..
. ...
....... .
. ..
. .
.. .......... ..
.
. ...... .
.
.... ...... ..
.
..
. ...... .
.
.
.. ....
... ...
... ...
...
... ... t
...
...
..
.
.
O
..
....
.
.
which follows from either Eqs. (5.29) - (5.30) or Eqs. (5.50) - (5.51), yields
n s1 = constant (5.72)
along the ray trajectory. Let us now introduce θ as the angle between i3 and
s, and θ0 as the angle between i3 and s0 (see Fig. 5.4), then s1 = sin(θ),
and we obtain
i.3
....
.......
i.3 ....
...
....
........ .......
... ......
θ= π 1
2
u
..
...
.... s ..
...
...
...
...
...
...
...
...
...
.
. .
shor
...................................................................
i3 .......... θ . .
.............
... .................................
.
...
...
...
...
..............
. .............
..........
u
........
....... ...
.
.
........ .
...
............
.
... ...
......
.... s0 .....
...
.. .
........
θ0
... .................
........... ................
..
......
u ... ..........
.
.
...
.
......
......
......
...
......
......
......
.....
......
..
......
.
.....
.....
.....
x3;hor
..... x3
x3;0
i.3
....
.......
.....
...
..
...
r ..........................................
.... i1
O
in which
∂l [n cos(θ)] = ∂3 n , (5.75)
x..3 x..3
.. .
......... ..........
.... ....
.. .. ◦ ◦
4 ...
...
...
..
... 4 θ = 15 ... 0
...
...
θ = 30
...
.. ......
.0
.... θ = 45 .......
.......
.
0
◦
... . ......
... ... ... ... .... .......
... ... ... . .. ...... ............
.
... ..... ..........
3 ...
...
...
..
.. 3 ...
... ................. .......................... ...................... 0
θ = 60 ◦
...
...
...
..
..
..
..
...
...
...
s ...........................
..........
.........
.......
......
..... .
.........
.......
..........
x..3 x..3
... ...
........ ........
... ...
.... .... θ = 15 ◦
0 . ◦
4 ..
...
...
..
... 4 ..
...
...
θ = 30 ...
..
..
..... 0
.....
... ..
. ....
... ... ... . ...
.
.... ... .... ... ....
.....
.. ... .. ... .......
3 ...
...
..
.. 3 ...
...
... .....
.. ....... .................
..............................
........ .
...........
....... θ = 45 ◦
.. . .
... .. ... ... .... ...... ......
...... ...... 0
...
..
... ................................................................ ..... ...............................................
.. ..........
s
. ........
... .. ... ..................... ....... ....
. .
...... ..
.. ........
.
2 ...
..
2 ... ... ....... ..... ....... .....
.......
.......
... ..
.. ... ......... ......... ..............
. .
........ .
.........................................
θ = 60 ◦
.
... .. ... ...... .....
.
.. ....... ...... 0
... . ... ......... .......
... ..
.. ... .......................................
... ... ...
1 ... ... 1 ...
... .
.
. ...
... ... ...
... .. ...
..................................................................................... .......................................................................................................................................................................................................................................................................................................... x1
.. n ..
0 1 2 0 2 4 6 8 10
x..3 x..3
.... ....
........ ........
.... .... ◦ ◦
..
...
..
...
θ = 15 0 θ = 30 0
4 ...
..
... 4 ...
..
.
...
....
..
.... θ = 45 .
.......
....... 0
◦
... ... ... . .
. ...
...
... ... ... ... .... .......
... ... ... ..
. .
....
. ...
........
.
... ... ... ... .....
. .....
.......
3 ...
...
..
..
..
3 ...
...
... ..... ......... ................................
... ... ..... ....... ......... θ = 60 ◦
0
.. ... ........ ............ ....... ....
... ..
.. ... . ...... .......
... .. ... .. ... ..... ...... ..... ......
... ..
.. ... .... ... ..... ........ .....
. ..
.......
.. .. ... .... ..... ....
2 ...
... ..
. 2 ...
... .. .... ....
...................
.....
..... .....
.....
...
... ...
..
.
..
s
...
...
...........
......
.......
........
................ .................. .............
......
... .. ... ......
... .. ...
1 ...
...
... 1 ...
... .
.
. ...
... ... ...
... .. ...
....................................................................................... x1
n ...........................................................................................................................................................................................................................................................................................................
0 1 2 0 2 4 6 8 10
Figure 5.5. Ray trajectories for a piecewise linear profile of the refraction index;
n = 1 + 12 (x3 − 1) when 1 ≤ x3 < 2, n = 32 − 12 (x3 − 2) when 2 ≤ x3 ≤ 3.
ray trajectories 137
x..3 x..3
.. .
......... ..........
.... ....
.. .. θ = 15 ◦
... ... 0
4 ...
..
... 4 ... .
...
θ = 60
...
..... 0 ◦
... ... ... ... .....
... ... ... .. ....
... ... ... ..
. ...
......
. ..
... .. ... .. .....
3 ...
..
..
. 3 ... ..
.. ....
.....
...
...
... ..
...
. ...
...
...
s .. ......
........
... .. ...
..
...
2 ...
...
..
..
..
2 ...
...
... .. ...
... ..
.. ...
... .. ...
... .. ...
1 ...
...
.
... 1 ...
...
...
... ... ...
... .. ...
....................................................................................
... n .........................................................................................................................................................................................................................................................................................................
... x1
0 1 2 0 2 4 6 8 10
x..3 x..3
... ...
........ ........
... ...
.... .... θ = 15 ◦
0
4 ..
...
...
..
... 4 ..
...
...
.
..
θ = 60
..
..
... 0 ◦
... . ...
... ... ... .. ...
.... ... .... .. ...
... ...
.. .. .. .. ...
3 ...
... ..
..
. 3 ...
...
..
.. ...
...
....
. ...
... .
.
.. ... ... .........
... .. ... .. ...
2
...
... ...
..
.. 2
...
... s .. .......
......
... ..
.. ...
... .. ...
... ..
.. ...
... .. ...
... .. ...
1 ... .
... 1 ...
... ... ...
... ... ...
... .. ...
.................................................................................... .......................................................................................................................................................................................................................................................................................................... x1
.. n ..
0 1 2 0 2 4 6 8 10
x..3 x..3
.... ....
........ ........
.... ....
.. .. ◦ ◦
4 ...
...
..
... 4 θ = 15 ... 0
...
θ = 30 .
...
0.
θ = 45..
.. ...
...
0
◦
... ... ... ... . .
..
... ... ... .
.. ... ...
...
... ... ... .. ..
. ..
... .. ... .. .. ..
...
3 .. 3 .. ....
... . ... .. ....
... .. ... .. ... ....
... ..
.
... ... ...
. ...
.....
.
. ..
... .. ... ... .... .......
... .. ... ... .... .........
...
2 ...
...
.. 2 ...
...
... ... ......
.............................................................
s
.. . . .......
..
... .. ... ............. ......
... ..
.. ... ......
.....
... .. ... .....
... .. ...
1 ... .
... 1 ...
.....
.....
.....
... ... ... .....
... ... ... .....
..... ◦
... .. ... θ = 60 .
...................................................................................... ....................................................................................................................................................0 x1
n ...........................................................................................................................................................
0 1 2 0 2 4 6 8 10
Figure 5.6. Ray trajectories for a piecewise linear profile of the refraction index;
n = 32 − 12 (x3 − 1) when 1 ≤ x3 < 2, n = 1 + 12 (x3 − 2) when 2 ≤ x3 ≤ 3.
138 electromagnetic rays in a two-dimensional medium
x..3 x..3
.. .
......... ..........
.... ....
.. ..
θ = 49 ◦
... ... .. 0
4 ...
... ..
..
.. 4 ...
... .
....
....
.
... .. ... ....
... .....
... .. ... .....
... ..
.. ... .....
3 ... .. 3 ... ........
.
... .. ... .....
... ..
... ......
.. .......
... .. ... ........
...
...
..
.. ... ..... ........................
......................................................................................................................................................................... 0
θ = 49.45 ◦
2 ...
... ..
..
2 ...
... .............................................................................................
.................... ...........
... .. ... ............. ........
... ..
.. ... .
....
............ .......
..... ......
s
... .. ...
.. ..... .....
... ... ..... .....
1 ...
...
..
..
..
..
1 ...
...
.....
.....
.....
... .. ... .....
...
..
.. ... θ = 50 .....
.. ◦
................................................................................... .........................................................................................................................................................................................................................0
... n ........................................................................................ x1
0 1 2 0 2 4 6 8 10
Figure 5.7. Ray trajectories for a Gaussian profile of the refraction index;
n = 32 − 12 exp(−|x3 − 2|2 ).
Multiplying Eq. (5.81) with x3 and Eq. (5.82) with −x1 , and adding the
results, we arrive at
or equivalently
n(x3 ∂l x1 − x1 ∂l x3 ) = constant (5.85)
along the ray trajectory. Since ∂l x1 = s1 and ∂l x3 = s3 , we may write
along the ray trajectory. Let us now introduce φray as the angle between i3
and s, and φ as the angle between i3 and x, then
and
x3 s1 − x1 s3 = r sin(φray − φ) = r sin(θ) , (5.89)
where θ as the angle between x and s. Let us further introduce θ0 as the
angle between x0 and s0 (see Fig. 5.8), we obtain
in which
C0 = r0 n(r0 ) sin(θ0 ) , (5.91)
denotes the trajectory constant of the ray. Equation (5.90) is known as
Snell’s law for a radially layered medium. A second result follows from a
multiplication of Eq. (5.81) with x1 and Eq. (5.82) with x3 , and adding the
results, we arrive at
x
r.....
.....
..
...
....
........
.. ....
θ s
........
xang
.. ...
...............
t .................
....................................................................
........... ....... rang
x0 ......
.
.
.. ..
.......
. .
. .
..
..
..
.
.
.
.
........
.......
...... ......
.....
.........
.
...
. ..
....... ...
....... .. ..... .....
...... ..... ..............
r0 ..... s0 ......
......
... .......
. t θ= 1π ..
......... ..
....... .
... ...... ......... ....
2
... .... . ......
................ ..
...
...
... θ0
............................
.
...... .
..
.
..
.....
.....
.....
....
.... ....... .
... ...... . ..
........... sang
t ..
... ..... . ..
.. .......
...
.......
... .. r .......
. ...
..... .... .
. .
....
.
....
.
. .
... .
... ..
.
.
.... rang
.... .. ....
....
..
....
.
.... ...
.. r0...
..
..
..
....
..
.... .. . .....
... ...
... ... .....
.
... ..
.
... .
.
. .
..
.. ... . ....
... ..
.. ..
... ... ....
.. .
....
.
... ..... .. .
.. .....
........
..... t
O
For a chosen value of the starting position and the starting direction, the
trajectory can have a point where θ = 12 π; in such a point the trajectory
has a tangent in the angular direction. Let r = rang the value of the radial
coordinate, where the tangent is in the angular direction (see Fig. 5.8), then
substitution of θ = 12 π and r = rang into Eq. (5.90), with use of Eq. (5.91),
yields
rang n(rang ) = r0 n(r0 ) sin(θ0 ) = C0 . (5.96)
Since 0 ≤ sin(θ0 ) ≤ 1, Eq. (5.96) is only valid if rang n(rang ) ≤ r0 n(r0 ).
This means that in the starting position the trajectory has to start in the
direction of decreasing r n. After passing the point with horizontal tangent,
the ray bends in the direction of increasing r n. According to Eq. (5.90) the
angle θ can remain at the value of 12 π and r can remain at the value rang ,
but Eq. (5.95) shows that it is only possible if the value r = rang coincides
with the surface where ∂r (r n) = 0. Such a surface can only be reached
asymptotically, unless the trajectory starts in a horizontal direction at this
surface.
n0
1
r2 2
n(r) = n0 2 − 2
.....
........
a
.....
.................................................... .......
........... ........ ..............
......................................................................................................................................................................................................................................................................................
.
........ ............
.....
.........
......
q
.................
.......
..... ........
..... .....
.............................................................................................................................................................................................................................................................. .......
......
.....
.....
. ............... ......
...... ...........
.............
...... ....
...
... ...... ...
. ............. .
.
..................................................................................................................................................................................................................................................... ............
.....
...
...
...
.. .................. . .....
. ................... .......... ....... ...
... ............ ... ....... ....... ...
. . . . . .
. ........... .. . ..
.. .
................................................................................. ................................................................................................................................................................... ............ .. .
.. ......................... ......... ........... ...... .....
... .................... ......... ...... ... ..
................. ......... ...... .. ..
.. ....... .... .
............. ....... ..... .. ...
. .. .. ..
....................................................................................................................................................................................................................................................................................................
....................................................................................................
Or
.
.... .....................................
................................
... .....................................................................................................
.................................................................................................................................................................................................................................................................................................... ... .. ........ .................. .... ...
.
... . .... .
.
... .
. .... .......... .......... ..... ....... ....... ... ...
............. . ...... .... ... .
.. ........................ .......... ........... ...... ....
............................................................................................................................................................................................................................................................ .......... . ....
...
...................... ... ......... ..... ...
.
.. .......
... .............. ........ ....
.... ...
. . .
...
. ... ..
. .... ..
...
...
... . . ................ ........ ..... ...
................................................................................................................................................................................................................................. ...... .....
. ..
... ......... ..... ...
... .......... ...... ...
........... ...... ...
.... .............. ...... ....
.............................................................................................................................................................................................................................................................. ...... ....
..... ....... ....
...... ........ .....
...... ......... ......
... ............ .......
...................................................................................................................................................................................................................................................................................... ............
........... ..
...................................................
n0
1
r2 2
n(r) = n0 2 − 2
......
.......
a
.....
..................................................... .......
........... ........ ..............
.....
..................
........
..........................................................
............
.........
... q
..................
......
......... ........ .....
........... ....... .........
........ ..... ................................................. ....... ....
. .. . .
.. ............................................. .............. ...... ....
................ .... ........... ...... ...
............. ... ....................................................................
.......... ..... ...
.................. ......... ..... ..
................................................................................... .................
.............
........
........ ........
....
............ ........................................... .
.. ..... ............. ........... .........
................................................. .......................................................................
.....
. . ..
.................. . . . ......
... .
......................................................... ..... . . . . . . . ..
.. ......................... .............. ......... .........
.......... . .... .. ........ ... ..... ... .. ...
................ ..................... ...
................. .
.............. .............. ...................
................................................................................................................................ ............ ......... ..... ....
................................................. ...........................................................................................................
...
.. ............................................................................................................. ............................... ........... ................
..................... .......... ..
. .
...
.. . . . . . . . . . . . . ........
.. ......
...... .. . ........................ ......................................
Or s
... ........ ... ............... ...................
.................................................................................................................................................................................................................................................................................. ..........................................................................................................................................................
. . . . ........ . .......... ..........
.
.. .
.................................................................................
.... . . .
....
... .. . . . .. . . ........ .. ... .......................................................................... source
................................................. ......................................................................................................... .................... ..
........... ...... .... ....
. .
................................................................................................................................. ............................................................................................... ................ ..................................
....................... . .... ... ... .. .... .
..................... ........................................................................... .................. ........ ....... .....
................................ ....... ................................................... ..................................................................... ....... .......... ........... ..........
........................................ ... .. ...........
............. .................... ............ ....... .........
............. ............................ ...
............... ...... .
............. ....... .
..................... .......................................................................................................... ....... .......
........ ........
.............
..................... ..... ..... ........ ..... ...
...................... ................. ... .
.. ..
..... ....
.
........ .............................. ................................... .. ... .
........ .... .......... ...... ....
........... ....... .........
.......... ........ .....
................. .......... ......
...... ............................................................. .....
....... ..
......... .......
............ .........
................................................
Exercise 5.1
Given the profile for the index of refraction of a piecewise linear medium,
n = 2 − 12 (x3 − 1) when 1 ≤ x3 ≤ 3. Sketch in Fig. 5.11, where the ‘bullets’
denote the starting positions, the ray trajectories for the starting angles
θ0 = 45◦ , 60◦ , 90◦ and 120◦ . Whenever relevant, calculate the value of x3;hor
where the ray has a horizontal tangent and the angle θexit with which a ray
enters one of the homogeneous halfspaces.
Exercise 5.2 √
Given the starting position x3;0 = 2 2 − 1 in a medium with the profile
indicated in Fig. 5.5, what is the range of starting angles for which the
ray trajectories describe undulating paths that are confined to the region
1 < x3 < 3?
Exercise 5.3
Given a medium with the profile indicated in Fig. 5.5. Does each starting
position within the region 1 < x3 < 3 possess an associated range of starting
angles such that the ray trajectories are confined to the range 1 < x3 < 3? If
so, give the range of starting angles for all starting positions, if not, explain
your answer.
Exercise 5.4
For a medium with a profile as given in Fig. 5.6, find the starting positions
within the region 1 < x3 < 3 for which there exist rays with a horizontal
tangent at some value of x3 = x3;hor .
Exercise 5.5
For a medium with the profile indicated in Fig. 5.6, find the starting angle
belonging to each starting position within the region 1 < x3;0 < 3 for which
there exists a ray that reaches a horizontal tangent asymptotically at x3 = 2.
Problem 5.1
Given the profile for the index of refraction of a piecewise radially layered
medium, n(r) = 3 − r when 1 ≤ r ≤ 2. Consider a ray that comes from the
region r < 1 and enters the region 1 ≤ r ≤ 2 with a starting angle θ0 . Give
the range of starting angles for which the ray remains trapped.
144 electromagnetic rays in a two-dimensional medium
x..3 x..3
.. .
......... ..........
.... ....
.. ..
... ...
4 ...
...
..
... 4 ...
...
...
... ... ...
... ... ...
... .... ...
3 ... ... 3 ...
... ...
s
...
... ... ...
...
... ... ...
... ... ...
2 ...
...
...
...
...
2 ...
...
... ... ...
... ... ...
... ...
... ...
... ...
1 ...
...
.... 1 ...
...
...
... ... ...
... .. ...
.....................................................................................
... n .........................................................................................................................................................................................................................................................................................................
... x1
0 1 2 0 2 4 6 8 10
x..3 x..3
... ...
........ ........
... ...
.... ....
4 ..
...
...
..
... 4 ..
...
...
...
... ... ...
.... ... ....
.. .... ..
3 ...
...
...
...
3 ...
...
... ... ...
...
... ...
s
...
... ... ...
2 ...
...
...
...
...
2 ...
...
... ... ...
... ... ...
... ...
... ...
... ...
1 ... .... 1 ...
... ... ...
... ... ...
... .. ...
.................................................................................... .......................................................................................................................................................................................................................................................................................................... x1
.. n ..
0 1 2 0 2 4 6 8 10
x..3 x..3
.... ....
........ ........
.... ....
.. ..
... ...
4 ...
..
... 4 ...
... ... ...
... ... ...
... ... ...
... .... ...
3 ...
...
...
...
3 ...
...
... ... ...
...
... ... ...
... ... ...
2 ...
...
...
... 2 ...
...
...
...
...
...
...
s
...
...
... ...
... ...
1 ... 1 ...
... .... ...
... ... ...
... ... ...
... .. ...
......................................................................................
n ........................................................................................................................................................................................................................................................................................................... x1
0 1 2 0 2 4 6 8 10
Figure 5.11. Ray trajectories for a piecewise linear profile of the refraction index;
n = 2 − 12 (x3 − 1) when 1 ≤ x3 ≤ 3.
Chapter 6
Transmission Lines
.....
.....
.....
.....
. ........
....
.....
.....
.....
... ......
.....
.....
.....
.....
.
.. .....
.....
.....
.....
.....
...
.....
....
.....
.....
.....
..
.......
..... ...........................................
..... ........... .....
............. ..... ................
........ .....
.....
C0
.......
.....
......
.
.
..
.... ..
......
......
......
...... i3 ..
.
....
.
. .... ...... ................. .....
... ...
i3 .... .....
C2
..
...
.
.... .
....
. . ..... .................. . ..
......
.
... .... ........ .... ..... ....
...
...
...
.....
............................ ...........
..... ....
...... .....
.....
..... τ .......
........
.... ........
.....
.....
.....
..... ..... ...... ....
...
...
....
... τ
.....
. .
.............
... .... .........
.... ...
.....
.... . .. .
...
...
.
....
.....
.
..
.. ... .. ...... ..... ... .....
... ... . . ..... ... .....
...
...
....
..... .. ........... ......... ...
..
......
.
......
......... ...........
... ..... ... ....
C1 i 3
... ..... ... ....
... .......... ..... ... .....
... ..... ...............
. . .....
... ............................
. . . .
... ....... ..
......
.
... .. ...... ..... .. .. ...
... ... ........ ..... .... .....
.. ..... .... .....
...
... .... τ .........
............. .........
..... ...
.. .....
.....
D
... ... . .. .
..
.
... ... . .
... . ...... . . .
... ... .....
...
..........
..
... .....
... ....
..... .....
.... ...... ...... ... .....
....
.... .......... .......... .. .........
.
..... ....... .. ...
..... .. ......
..... ........
.....
...... ...........
.
...... ..
..
...... ....
....... .....
........ .....
......... ......
........... .... ..........
..............................................
6.1. TEM-waves
ˆ 3 , s) ,
Ĥ1 = h1 (x1 , x2 ) I(x (6.11)
ˆ 3 , s) ,
Ĥ2 = h2 (x1 , x2 ) I(x (6.12)
Ĥ3 = 0 , (6.13)
The TEM-wave of the type represented by Eqs. (6.8) - (6.14) satisfies the
partial differential equations (6.1) - (6.6) and the boundary conditions of
Eq. (6.7). To investigate this, we substitute the representations of Eqs. (6.8)
- (6.13) into Eqs. (6.1) - (6.7). This yields
h2 ∂3 Iˆ + (σ+sε)e1 V̂ = 0, (6.15)
−h1 ∂3 Iˆ + (σ+sε)e2 V̂ = 0, (6.16)
ˆ
−(∂1 h2 − ∂2 h1 ) I = 0 , (6.17)
∂3 Iˆ + γY0 V̂ = 0 , (6.22)
∂3 V̂ + γZ0 Iˆ = 0 , (6.23)
with
1
Z0 = , (6.24)
Y0
τ1 e1 + τ2 e2 = 0 on Cn . (6.31)
Note that γY0 and γZ0 play the role of the constants of separation. In order
that e1 and h2 satisfy Eqs. (6.25) and (6.29), and that e2 and h1 satisfy
Eqs. (6.26) and (6.28), we require that
1
γ = [(σ + sε)sµ] 2 with Re(γ) ≥ 0 . (6.32)
1
σ + sε 2
h1 = − Z0 e2 , (6.33)
sµ
1
σ + sε 2
h2 = Z0 e1 . (6.34)
sµ
∂ 1 e1 + ∂ 2 e2 = 0 . (6.35)
...........................................................................
............. ..........
.......... ........
.............. .......
....... .......
.
. ......... ......
......
.... .... .....
.....
....... .....
... ....
D
.
. ....
...
. ...
.... ...
... ...
. ...
... ...
... ........... ..................... .................. ....
.......... ...
.... ..... ..... ...
C2 C1 C0
..... ..
..... ..
... ........
... .....
. ... ...... ... ...
..
.... ..... ...
.. .
.
.
. ...
.. ...
... ... .. .... .. .
... ... . . . ...
... ... .
. .
..... ...
. .. .
....
.
......
... ..... .. .
. .. .... .. .
.. ... . .....
.
........ ... . .. ..
........ ... . .. .. ....
... ................................. ....... ................................. ....... ........
...
...
...
τ . .
τ .
...
.. . τ
..
... ...
... ...
... ...
.....
..... ......
.
..... .....
..... .....
...... ......
...... ......
.......
........ ....
........
......... .....
........... .........
............... ...........
.................................................................
∂2 ϕ = 0 . (6.39)
lim ϕ = 0 , (6.41)
x1 ↑a
lim ϕ = 1 . (6.42)
x1 ↓0
.............................................................................................. w ................................................................................................
x1 = a
E
.....
........
...
..
...
....
..
×f............................................ H
.
D
S
... i1
..........
...
r .... i2
............................. x1 = 0
O
Substituting Eqs. (6.44) and (6.45) in the normalization integral of Eq. (6.14)
and integrating over the domain {(x1 , x2 ) ∈ D; 0 < x1 < a, 0 < x2 < w},
where w is a characteristic length in the x2 -direction (see Fig. 6.3), we obtain
the characteristic impedance as
1
sµ 2 a (6.46)
Z0 = ,
σ + sε w
..............................................................
........... .........
......... ........
............. .......
.
...... ......
.......... ......
.....
..
..... .....
.....
....... ....
... ....
D
.
. ...
...
. ...
.
.. ....
....... ....... ....... . ..
.
.........
.
.
E ...
...
.... . . ...
. .. ...
...
...... .......
×f
... .. . . .. ...
... .... ...
.... ...
.
.......... .. . . . .
...
....... ...
...
...
...
.. ...... ........ ......... ...
...........
.. S .
..
........
. . ...
...
... ...
r i
..........
... . .... 1.....
.
.
.
..... . ....
.....
. . .
... .....
H ...
...
rO
.. . ... ... ..
.. .. ..... ... ..
.
......... .
.. .
. ...
... . . .. .. . ..
... ... ...
...
............................... ...
. .
. ..
.. .. ... ...... ... . . ..
a i
... . .... . .. . .
.
... ........... .. 2 ..
. ..
... . . . . .
.. .......... .... .
..... ... ...
... ...... ................... ...... ...
... ....... .........
... ........ ..................... ...
.. ..
... ...
.......... .. .
...
... ...... ...... .
....... . ..... ...
...
...
... ...............
....... b .......
....... ....... ...... ...
...
..
...
.......... .
....
... .
... ...
... ...
.... ...
....
..... ......
.
..... ....
..... .....
...... ......
...... ......
.......
........ . ...........
......... .......
............ .........
.........................................................
where a is the radius of the inner conductor and b is the radius of the outer
conductor of the coaxial line. Further, A and B are arbitrary constants.
To prove that ϕ satisfies the two-dimensional Laplace equation (6.37), we
compute
x1 x2
∂1 ϕ = B 2 , ∂2 ϕ = B 2 , (6.49)
r r
and
1 x2 1 x2
∂1 ∂1 ϕ = B( 2 − 2 41 ) , ∂2 ∂2 ϕ = B( 2 − 2 42 ) , (6.50)
r r r r
hence, the two-dimensional Laplace equation, ∂1 ∂1 ϕ+∂2 ∂2 ϕ = 0, is satisfied.
The constants A and B follow from the boundary conditions (cf. Eqs. (6.41)
- (6.42)
lim ϕ = 0 , (6.51)
r↑b
lim ϕ = 1 , (6.52)
r↓a
154 transmission lines
as
ln(1/b) 1
A= , B= . (6.53)
ln(a/b) ln(a/b)
Hence, the solution of the pertaining two-dimensional potential problem is
obtained as
ln(r/b)
ϕ= . (6.54)
ln(a/b)
Hence, from Eq. (6.36) it follows that
1 x1 1 x2
e1 = , e2 = , (6.55)
ln(b/a) r2 ln(b/a) r2
Substituting Eqs. (6.55) and (6.56) in the normalization integral of Eq. (6.14)
1
and integrating over the domain {(x1 , x2 ) ∈ D; a < (x21 + x22 ) 2 < b}, (see
Fig. 6.3), we obtain the relation
1
σ + sε 2 Z0 1
(e1 h2 − e2 h1 ) dA = dA = 1 .
(x1 ,x2 )∈D sµ [ln(b/a)]2 (x1 ,x2 )∈D r2
(6.57)
Since
b
1 1
dA = 2π dr = 2π ln(b/a) , (6.58)
(x1 ,x2 )∈D r2 r=a r
we arrive at the characteristic impedance as
1
sµ 2 ln(b/a) (6.59)
Z0 = .
σ + sε 2π
∂3 Iˆ + γY0 V̂ = 0 , (6.60)
∂3 V̂ + γZ0 Iˆ = 0 , (6.61)
∂3 ∂3 V̂ − γ 2 V̂ = 0 , (6.62)
for the electric potential V̂ . An analysis similar to Sections 3.1 and 3.6 yields
the general solution
ˆ 3 , s)
I(x ˆ 3 +dx3 , s)
I(x
e ...................................... r ............................... e
... ..
...... ......
......... .........
....
γZ0 dx3 ....
... ...
... ...
... ...
. .
Figure 6.5. The circuit equivalent of a transmission line of infinitesimal length dx3 .
156 transmission lines
where the s-dependent constants v̂ + , v̂ − , î+ and î− are determined by the
electromagnetic source strength at the beginning of the transmission line
and the terminating load at the end of the transmission line. On account of
Eq. (6.60) these constants have to satisfy the constraining relations
The functions v̂ + (s) exp(−γx3 ) and î+ (s) exp(−γx3 ) represent a wave trav-
eling in the positive x3 -direction, while v̂ − (s) exp(γx3 ) and î− (s) exp(γx3 )
represent a wave traveling in the negative x3 -direction.
ˆ s)
I(0, ˆ s)
I(L,
.... ...
...................................... ..... ... .. .......................................
.... ...
....... ........
......... .........
...
...
.. incident wave ...
.... ......................................................... ....
.. ..
... ...
... ...
... ...
transmission line
V̂ (0, s) V̂ (L, s) ZL
.. ...
...
Z0 , L .. ...
...
... ...
... ...
... .......................................................... ...
... ...
.. ..
..........
.......
reflected wave ..........
......
.
.... ....
..... .... ..
... ....
x3 = 0 x3 = L
ZL − Z0
Γ= exp(−2γL) , (6.72)
ZL + Z0
158 transmission lines
ZL + Z0 tanh(γL)
Zin = Z0 . (6.73)
Z0 + ZL tanh(γL)
We now determine the reflection coefficient and the input impedance for
various load impedances.
ZL − Z0
Γ= exp(−2jkL) (6.74)
ZL + Z0
and
ZL + jZ0 tan(kL)
Zin = Z0 , (6.75)
Z0 + jZL tan(kL)
where
1 ω
k = ω(εµ) 2 = . (6.76)
c
In the lossless case the characteristic impedance becomes real. For the
1
parallel-plate waveguide we then have Z0 = ( µε ) 2 wa and for the coaxial line we
1
have Z0 = ( µε ) 2 ln(b/a)
2π . We subsequently determine the reflection coefficient
and the input impedance for various load impedances.
propagation properties 159
where we have used Eqs. (6.8) - (6.13) and the normalization integral of
Eq. (6.14). Substitution of the expressions for V̂ = V̂ (x3 , jω) and Iˆ =
ˆ 3 , jω), i.e.,
I(x
It shows that in the lossless case the power flow in the transmission line is
equal to the difference of the power flow of the incident wave and the power
flow of the reflected wave.
160 transmission lines
I(0, t) .. ...
I(L, t)
....................................... ......... .. .......................................
...... ...
... ....
.......... ..........
. .. . ..
... ...
.. ..
.... ...
ZS ..
...
...
...
... ...
.... ....
transmission line
V (0, t) V (L, t) ZL
. ...
... Z0 , L . ...
...
VS ...
...
...
...
...
...
...
...
...
...
.......... ..........
........ ........
.
... ..
...
..... .... .
... ...
x3 = 0 x3 = L
where
ZL − Z0
ΓL = (6.84)
ZL + Z0
is the reflection coefficient of the reflected wave of a transmission line of zero
length. Applying the shift rule of the Laplace transformation, the corre-
sponding time domain result is
x3 2L−x3
V (x3 , t) = v + (t − ) + ΓL v + (t − ), (6.85)
c c
x3 2L−x3
I(x3 , t) = Y0 v + (t − ) − ΓL v + (t − ) . (6.86)
c c
Equations (6.85) and (6.86) show that on the transmission line a wave prop-
agates in the positive x3 -direction with wave speed c. At x3 = L, this wave
is reflected with reflection coefficient ΓL and propagates with wave speed c.
Once v + (t) is known the present problem is completely determined. This
quantity follows from the transient time behavior of the source.
When Zin = Zin (s) is the input impedance of the transmission line and
ZS is the internal resistance of the source, then from Fig. 6.7 we observe
that
Zin (s)
V̂ (0, s) = V̂S (s) , (6.87)
Zin (s) + ZS
where V̂S (s) is the Laplace transform of the time-dependent voltage VS (t) of
the source. Equation (6.87) cannot be transformed back to the time domain
by simple inspection. Therefore, we use Eqs. (6.66), (6.69) and (6.87) to
obtain
V̂ (0, s) Z0
v̂ + (s) = = V̂S (s) . (6.88)
1+Γ Z0 + ZS + (Z0 − ZS )Γ
Now in the first factor of the right-hand side of Eq. (6.88) only Γ =
ΓL exp(−2sL/c) contains an s-dependence. We therefore expand the first
factor of the right-hand side of Eq. (6.88) in terms of powers of Γ. We write
1 Z0
v̂ + (s) = V̂S (s) , (6.89)
1 − ΓS ΓL exp(−2sL/c) Z0 + ZS
in which
ZS − Z0 ZL − Z0
ΓS = , ΓL = . (6.90)
ZS + Z0 ZL + Z0
162 transmission lines
Together with Eqs. (6.85) and (6.86) the transient time behavior of the
propagation of waves on the transmission line is described completely. In
order to interpret the different terms of Eq. (6.85) and (6.92) we rearrange
the different terms as
Z0 x3
V (x3 , t) = [ + VS (t − )
Z0 + ZS c
2L−x3
+ΓL VS (t − )
c
x3 +2L
+ΓS ΓL VS (t − )
c
4L−x3 (6.93)
+ΓL ΓS ΓL VS (t − )
c
x3 +4L
+ΓS ΓL ΓS ΓL VS (t − )
c
6L−x3
+ΓL ΓS ΓL ΓS ΓL VS (t − )
c
+ ··· ].
Z0 VS (t)/(Z0 + ZS ). This pulse travels along the transmission line with wave
speed c and at t = L/c it arrives at the load. A reflection occurs with a re-
flection coefficient ΓL . The reflected pulse travels with wave speed c toward
the source and at t = 2L/c it arrives at the source, where a second reflec-
tion takes place with a reflection coefficient ΓS . The reflected pulse travels
again to the load with a voltage ΓS ΓL Z0 VS (t)/(Z0 + ZS ). This process is
continued. In Fig. 6.8 we illustrate these phenomena by plotting the voltage
V (x3 , t) at the middle, x3 = 12 L, of the transmission line as a function of
time.
VS (t) Z0 = 50 Ω L = 0.6 m
..
.......
........
1 ...
....... ZS = 150 Ω ZL = 10 Ω
.. ... ...
... .... ....
..
...
...
.. ..
... ...
... ....
c = 3 × 108 m/s
.. .... ....
.. ..
... ....
... ...
.... ....
.. ..
... ....
... .... ......................................................................
0 .. ..
.. ...
... .
... 2
................................................ t in (ns)
Z0 +ZS 1
Z0 ..V ( 2 L, t)
......
.........
...
....
..
...
...
1 .
......
... ....
... ...
..... ....
. ..
... ....
... ...
.... ....
.. ...
... ...
... ...
.... .... .... ..
.....
.. .. .. .. .. ..
.. .. .. ... ...... .. .. .......
.................................................... .... ......................................................................... .... ......................................................................... .... ............................................................................. ........................................................................................ .........................................
0 .. ..
.. .. ... ..
... ...
.
.. ...
.. .
... ..
2 .. ... ..
... ..
. 4 .
. .
.
.....
.. 6 8 10
... ...
... ....
... .. ................................................................................... t (in ns)
... ...
.. ..
... ...
......
Figure 6.8. The source voltage VS (t) and the voltage V ( 21 L, t).
164 transmission lines
Exercise 6.1
(a) Calculate the inner radius of the outer conductor b of a lossless coaxial
transmission line with Z0 = 50 Ω, ε = 1.5ε0 and an inner conductor
with outer radius a = 4.11 mm.
(b) Find the time-averaged power flow in the coaxial transmission line in
case of a matched load.
Exercise 6.2
For the circuit shown in Fig. 6.7, Z0 = 50Ω, ZS = 30Ω, ZL = 20Ω, L = 2 m,
c = 1 × 108 m/s and VS = 1 V for 0 ≤ t ≤ 1 ns, otherwise VS = 0 V.
(a) Calculate v + (t).
Exercise 6.3
For a two-wire transmission line as depicted in Fig. 6.7, the following data
are available: VS (t) = V0 H(t) V, H(t) denotes the unit step function starting
at t = 0; Z0 = 100 Ω. The voltages at the beginning and at the end of the
line are measured for the first 5 µs, they are
! !
100 V, 0 < t < 4 µs, 0 V, 0 < t < 2 µs,
V (0, t) = V (L, t) =
90 V, 4 < t < 5 µs, 75 V, 2 < t < 5 µs.
Exercise 6.4
In a steady-state situation, for what value(s) of L will the transmission line
terminated by a short circuit load behave like a capacitor? Conversely, for
what value(s) of L will the open circuit load behave as an inductor?
Exercise 6.5
The Voltage Standing Wave Ratio (VSWR) is a measure for the effectiveness
of the matching load at the end of a transmission line. In the steady-state
situation it is found as VSWR = |V |max /|V |min = (1 + |Γ|)/(1 − |Γ|). A loss-
less, two-wire line is terminated with a television set A giving a VSWR=3.
When A is replaced by the set B, it is meaured that VSWR=1.5. Which
of the two sets is better matched to the transmission line? Explain your
answer. What is the amplitude of the reflection coefficient of the set B?
Problem 6.1
Given a parallel-plate transmission line, made up of perfect conductors, of
width a = 1 mm and characteristic length w = 10 cm, with a perfect di-
electric medium between the plates, µ = µ0 . The voltage along the line,
for a uniform plane wave propagating along the line, is given by V (x3 , t) =
10 cos(3π × 108 t − 3πx3 ). Find
(a) the electric field strength E1 (x3 , t) of the wave
Problem 6.2
Consider the two-wire transmission line of Fig. 6.7 but now the source is
connected with the transmission line via a switch at x3 = 0. VS = 100 V,
ZS = 40 Ω, Z0 = 60 Ω and T = L/c = 1 µs. If the switch is closed at t = 0 s,
find the value of ZL for the following situations
(a) V ( 12 L, 1.6µs) = 75 V
(b) V ( 13 L, 2.5µs) = 44 V
( c ) V ( 34 L, t → ∞) = 80 V.
166 transmission lines
Problem 6.3
A lossless transmission line is short circuited at the far end. A variable
frequency voltage generator is connected as its input and the current drawn
is measured. It is found that the current reaches a maximum value for
f = 500 MHz while it reaches a minimum for f = 505 MHz. Determine
whether the current drawn would be a maximum, minimum or neither for
the following frequencies
(a) f = 450 Mhz
( c ) f = 335 Mhz.
Chapter 7
Electromagnetic Waveguides
axial
direction
exterior
domain
electromagnetically
impenetrable wall
interior
domain
. transverse
plane
metallic wall
metallic wall
(a) (b)
metallic outer
conductor
metallic inner
conductor
(c)
axial
direction
inhomogeneously
distributed
. penetrable
material
. transverse
plane
axis axis
mantle
fiber core
. .
.
(a) (b)
layer . layers .
.
. .
substrate substrate
(a) (b)
• Parallel polarization
∂1 ∂1 Ĥ2 + ∂3 ∂3 Ĥ2 + ω 2 εµĤ2 = 0 , (7.2)
x1 = a
ε, µ
propagation direction
....
......................................................................................
...
.....
i1
........
r
....
........................... i3
.... x1 = 0
O
TE-waves
We will now first discuss in detail the case of E-polarized waves or TE-
waves. Since we are interested in the propagation of an electromagnetic wave
in the positive x3 -direction we will assume a solution of Eq. (7.5) of the type
in which
k12 = ω 2 εµ − k32 . (7.10)
2j  sin(k1 a) = 0 . (7.16)
For a non-zero value of Â, the latter equation can only be satisfied if
k1 a = mπ , for m = 0, 1, 2, · · · . (7.17)
Substitution of this result into Eq. (7.10) leads to the dispersion relation
mπ 2
k32 = ω 2 εµ − ( ) , m = 0, 1, 2, · · · . (7.18)
a
Introducing the wavenumber
1
k = ω(εµ) 2 , (7.19)
⎧ 1
⎪
⎪ mπ 2 2 mπ
⎪
⎪ k −(
2
) when k ≥ ,
⎪
⎨ a a
k3 ≡ k3;m = (7.20)
⎪
⎪ 1
⎪
⎪ mπ 2 mπ
⎪
⎩ −j ( ) −k 2 2
when k < .
a a
parallel-plate waveguide 175
These modes decay in the positive x3 -direction. From Eqs. (7.21) - (7.22)
we find for these modes the time-averaged power flow density,
∗
1
2 Re (Ê × Ĥ ) · i3 = 1
2 Re −Ê2 Ĥ1∗
m m
∗
k3;m mπ
= 2 Re |Âm |2 [sin( x1 )]2 exp(−2jk3;m x3 )
ωµ a
= 0, (7.25)
TM-waves
Secondly, we will discuss in detail the case of H-polarized or TM-waves.
Since we are interested in the propagation of an electromagnetic wave in the
176 electromagnetic waveguides
where Eq. (7.4) has been used. From Eqs. (7.30) - (7.31) we see that
2j  sin(k1 a) = 0 . (7.34)
For a non-zero value of Â, the latter equation can only be satisfied if
k1 a = mπ , for m = 0, 1, 2, · · · . (7.35)
parallel-plate waveguide 177
Substitution of this result into Eq. (7.28) leads to the dispersion relation
mπ 2
k32 = ω 2 εµ − ( ) , m = 0, 1, 2, · · · . (7.36)
a
Introducing the wavenumber
1
k = ω(εµ) 2 , (7.37)
we find from Eq. (7.36) the propagation constants
⎧ 1
⎪
⎪ mπ 2 2 mπ
⎪
⎪ k −(
2
) when k ≥ ,
⎪
⎨ a a
k3 ≡ k3;m = (7.38)
⎪
⎪ 1
⎪
⎪ mπ 2 mπ
⎪
⎩ −j ( ) −k 2 2
when k < .
a a
1
the latter mode being a TEM-mode with propagation constant k = ω(εµ) 2 ,
see Section 3.6.
when 0 ≤ ω m < 1. These results lead to the dispersion curve of the m-th
mode in a lossless parallel-plate waveguide shown in Fig. 7.7.
. ..
αm .
.
. .....
...
....
.
. ..
. ....
βm .
.
. ...
....
. ..
. ....
6 .
.
.
....
...
. ..
2.0 .
.
. ....
...
....
.
. ..
. ....
...
.
. ...
. .
. ....
. ...
. ....
.
.
.
....
αm
. ...
. ....
.
.
. ....
1.0 ........................
..........
........
....... .
.
. ...
....
...... .
...... ...
..... . .
.....
......
...
...
βm
. .....
. ...
. ... ..
. ... ..
. ... ...
. ...
. ...
. ... ..
. ... ...
. ... ..
...
. ....
. ...
. ..
..
0 .
0 1.0 2.0 - ωm
" #$ % ........
evanescent propagation region
region
Plane-wave decomposition
Each mode can be written as a superposition of two plane waves. For a
TEm -mode the electric field strength may be written as (cf. Eq. (7.21))
and for a TMm -mode the magnetic field strength may be written as (cf.
Eq. (7.39)
where
mπ 1
k1;m = , k3;m = [k 2 − k1;m
2
]2 . (7.57)
a
For propagating modes k3;m is real-valued and each plane-wave constituent
of Eqs. (7.55) - (7.56) is a uniform one. The direction of propagation of such
a uniform plane wave is determined by the vector {k1 , k3 } ≡ {k1;m , k3;m } for
a wave propagating in the positive x1 -direction and the positive x3 -direction
and by the vector {k1 , k3 } ≡ {−k1;m , k3;m } for a wave propagating in the
negative x1 -direction and the positive x3 -direction. These directions of prop-
agation can be constructed in the (k1 , k3 )-plane (see Fig. 7.8). Each prop-
agating mode becomes non-propagating or evanescent in the x3 -direction,
when the operating frequency becomes below the cut-off frequency for that
mode. In Fig. 7.8, this phenomenon is also illustrated in the (k1 , k3 )-plane.
Finally, in Fig. 7.9, we present the amplitude distributions of three guided
TEm -modes, m = 1, 2, 3.
propagation of modes in a parallel-plate waveguide 181
k1
.....
.........
..........
...
....
..
...
...
evanescent ...
...
modes ... 4π ......
..
...
...
... a ....
...
......... ...
.... ...
.
.... ...... ..
. .. .
.. ..............................................................
........
k3 = {k1;3 , k3;3 }
...........
....... . ...... ..... .
..
3π .... .........
........
... . .
. ... ... ......
... ....... ... ........... ............
... ..
........ a .
. . ......
... ....
.... . .
.
. ...
. ......
..... .
. .. .....
... . .
..... .
. ... .....
.....
... ... .
. .. .....
... ..
.............. .
.
. . .
. .....
... .... ......... .
. .. ....
k2 = {k1;2 , k3;2 }
.
... ..... .
. .. ....
... ..
. .
...... 2π .. .
. .
. ...
.
... ....... . ... ............
..
. . a .
. .. ....... .....
.
propagating ..
.
.
.. .
.....
..
.....
.
.
.
.
....
...
..
.. ........ . ...
...
... ..... ... ...... ...
modes .. ..... ... ... .....
.. ..... . ........ ...
...
...
...
.
..
.
.
.
.
k ......
.......
.....
.
.
.
.
.
..
. ..
.
.
....
.
.....
. ...
........ ...
...
...
. π .... .
k1 = {k1;1 , k3;1 }
.. ..... .
. .
. . . ...
... . ..... . . .. ....................
... .... .
. ..
. ...... .
......
... .
.
.....
..... a ... .
.
.
.
. .
......
. ..
.... .
. ... .
.......... ...
... ..
.
.....
. .
.
. ..
. ..
......
.... ............ ...
..... . .. ...
... .. ..... .... .... ...... ...........
... .
... ..... .
. .. ..
......
. .... . . .
............
. ...
...
... .
.
.
.....
. . ... .......
.
. . ................. ..
........ . ..... .. .. ..... .... ...... ..
... . . ..
..... .... ..... .......... . . ..
. ...
... ..
.. ...................... ..
........................................................................................................................................................................................................................................................................................................................................................................................................................
.
..
O k3
Figure 7.8. The (k1 , k3 )-plane for the lossless parallel-plate waveguide.
x1 x1 x1
6 6 6
TE1 -mode TE2 -mode TE3 -mode
........... .............. ...........................
.......... ................ .........
........ ............ ..............
......
.... ................
.................... .......
.............................................
.. .................. ..................
.............................
...... ........... .........................
........... .
..... .....
.... ..
. . .............. .........
........... ................... ...........................
- - -
Ê(x1 , jω) Ê(x1 , jω) Ê(x1 , jω)
In Section 7.1 we have seen that two types of waves can exist independently:
E-polarized waves or TE-waves (perpendicular polarization) with field com-
ponents Ĥ1 , Ĥ3 and Ê2 that satisfy Eqs. (7.5) - (7.7), and H-polarized waves
or TM-waves (parallel polarization) with the field components Ê1 , Ê3 and
Ĥ2 that satisfy Eqs. (7.2) - (7.4).
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x1 = a
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . .
.
.
.
.
.
.
.
.
.
. . propagation direction . . . . . . . . . . . .
slab .
.
.
.
.
.
.
.
.
ε(2) , µ 0
.
.
.
.
.
.
....
....................................................................................
.. . .
.
.
.
.
.
.
.
.
.
.
.
. D(2)
. . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.......
.
. ....
i1 .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
r i3
. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . .
...
. . . . . . . . . . . . . . . . ........................... . . . . . . . . . . . . . .
.... x1 = 0
O
TE-waves
We will now first discuss in detail the case of E-polarized waves or TE-
waves. We are looking for a solution of the equations, cf. Eqs. (7.5) - (7.7),
together with the boundary conditions for the tangential field components
Ê2 and Ĥ3 at x1 = a and x1 = 0. We will further assume that the wave
we are looking for will propagate in the positive x3 -direction and will have
a nonuniform, transverse intensity distribution that is possibly restricted to
the slab region. That means we will require that the wavefield vanishes for
184 electromagnetic waveguides
(i) (i)
Ê = Â(i) exp(jk1 x1 ) + B̂ (i) exp(−jk1 x1 ) . (7.65)
(i) 1
In order to satisfy the condition at |x1 | → ∞, k1 = (ω 2 ε(i) µ0 −k32 ) 2 is chosen
(i)
such that Im(k1 ) ≤ 0, and in addition we take B̂ (1) = 0 and Â(3) = 0. The
other frequency-dependent constants Â(i) and B̂ (i) follow from the boundary
equations at x1 = a and x1 = 0, where
lim Ê2 = lim Ê2 =⇒ lim Ê = lim Ê , (7.66)
x1 ↑a x1 ↓a x1 ↑a x1 ↓a
and
lim Ê2 = lim Ê2 =⇒ lim Ê = lim Ê , (7.68)
x1 ↓0 x1 ↑0 x1 ↓0 x1 ↑0
These four equations containing the four unknowns Â(2) ,B̂ (2) , B̂ (3) and Â(1)
have a nonzero solution only if the determinant of the matrix of their coef-
ficients is zero. In this manner we find an equation which is known as the
eigenvalue equation or characteristic equation, because it leads to specific
values or eigenvalues for the still unknown propagation coefficients k3 . In
the lossless guiding structure, we expect those solutions of Eq. (7.70) for
(1) (3) (2)
which k1 and k1 are both negative imaginary and k1 is positive real.
Therefore, it is advantageous to introduce the quantities
(1) 1
κ(1) = jk1 = [k32 − ω 2 ε(1) µ0 ] 2 > 0 ,
(2) 1
κ(2) = k1 = [ω 2 ε(2) µ0 − k32 ] 2 > 0 ,
(3) 1
κ(3) = jk1 = [k32 − ω 2 ε(3) µ0 ] 2 > 0 . (7.71)
Then, the characteristic equation which must be satisfied if a solution of
Eq. (7.70) is to be obtained, is found as
exp(jκ(2) a) exp(−jκ(2) a) − exp(−κ(3) a) 0
(2)
κ exp(jκ(2) a) −κ exp(−jκ a) −jκ(3) exp(−κ(3) a)
(2) (2) 0
=0.
1 1 0 −1
κ(2) −κ (2) 0 jκ(1)
(7.72)
By straightforward algebra, Eq. (7.72) reduces to the form
κ(1) κ(3)
(2)
+ (2)
tan(κ(2) a) = κ (1)κ (3) . (7.73)
κ κ
1 − (2) (2)
κ κ
To obtain the relevant values of k3 , it is advantageous to use the addition
formula
tan(u) + tan(v)
tan(u + v) =
1 − tan(u) tan(v)
and to rewrite Eq. (7.73) as
κ(1) κ(3)
κ(2) a = arctan + arctan + mπ, m = 0, 1, 2, · · · , ∞ .
κ(2) κ(2)
(7.74)
186 electromagnetic waveguides
This equation, in which the quantities κ(i) are related to the propagation
coefficient k3 through Eq. (7.71), is called the dispersion equation. Its solu-
tions
k3 ≡ k3;m (ω) , m = 0, 1, 2, · · · , ∞ ,
form an innumerable set of real and complex numbers. The real values k3;m
that satisfy the inequalities in Eq. (7.71) form a finite set with
where
1
k (i) = ω(ε(i) µ0 ) 2 , i = 1, 2, 3 , (7.76)
⎧ (2) (3)
⎪
⎪
⎪
cos(κm a−ψm ) exp[−κm (x1 −a)] exp(−jk3;m x3 )
⎪
⎪
⎪
⎪ when a < x1 < ∞ ,
(1) ⎪
⎨ (2)
Âm cos(κm x1 −ψm ) exp(−jk3;m x3 )
Ê2;m =
cos(ψm ) ⎪
⎪
⎪ when 0 < x1 < a ,
⎪
⎪
⎪
⎪
(1)
cos(ψm ) exp(κm x1 ) exp(−jk3;m x3 )
⎪
⎩
when − ∞ < x1 < 0 ,
(7.77)
in which
(1)
κm
ψm = arctan (2)
. (7.78)
κm
dielectric slab waveguides 187
The magnetic field strength then follows from Eqs. (7.60) - (7.61) as
⎧
⎪
⎪ −k3;m (2) (3)
⎪
⎪ cos(κm a−ψm ) exp[−κm (x1 −a)] exp(−jk3;m x3 )
⎪
⎪ ωµ
⎪
⎪
0
⎪
⎪
⎪
when a < x1 < ∞ ,
⎪
⎪ −k
(1)
Âm ⎨ 3;m (2)
cos(κm x1 −ψm ) exp(−jk3;m x3 )
Ĥ1;m = ωµ0
⎪
cos(ψm ) ⎪
⎪
⎪ when 0 < x1 < a ,
⎪
⎪
⎪
⎪ −k 3;m (1)
⎪
⎪ cos(ψm ) exp(κm x1 ) exp(−jk3;m x3 )
⎪
⎪ ωµ0
⎪
⎩
when − ∞ < x1 < 0 ,
(7.79)
and
⎧
⎪
⎪
(3)
κm
⎪
⎪
(2) (3)
cos(κm a−ψm ) exp[−κm (x1 −a)] exp(−jk3;m x3 )
⎪
⎪
⎪
⎪ jωµ 0
⎪
⎪ when a < x1 < ∞ ,
⎪
⎪
⎪
⎪
(1) ⎪
⎨ κm
(2)
(2)
Ĥ3;m =
Âm sin(κm x1 −ψm ) exp(−jk3;m x3 )
cos(ψm ) ⎪ jωµ0
⎪
⎪
⎪
⎪ when 0 < x1 < a ,
⎪
⎪
⎪ (1)
⎪ −κm cos(ψ ) exp(κ(1) x ) exp(−jk x )
⎪
⎪
⎪
⎪ m m 1 3;m 3
⎪
⎪ jωµ0
⎩
when − ∞ < x1 < 0 .
(7.80)
From Eqs. (7.77) and (7.80) it follows that the time-averaged power flow
density in the x1 -direction,
∗ ∗
1
2 Re[(Ê m × Ĥ m ) · i1 ] = 12 Re[Ê2;m Ĥ3;m ]=0, (7.81)
in each of the three domains. The time-averaged power flow density in the
x3 -direction follows from Eqs. (7.77) and (7.79) as
188 electromagnetic waveguides
∗ ∗ ]
1
2 Re[(Ê m × Ĥ m ) · i3 ] = 12 Re[−Ê2;m Ĥ1;m
⎧
⎪
⎪ k3;m (2) (3)
⎪
⎪ [cos(κm a−ψm )]2 exp[−2κm (x1 −a)] , a < x1 < ∞ ,
⎪
⎪ ωµ0
⎪
⎪
⎪
⎪
Â(1) ⎪ 2
⎨
m k3;m (2)
= [cos(κm x1 −ψm )]2 , 0 < x1 < a ,
cos(ψm ) ⎪ ⎪ ωµ0
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ k3;m
⎪
⎩
(1)
[cos(ψm )]2 exp(2κm x1 ) , −∞ < x1 < 0 .
ωµ0
(7.82)
From Eqs. (7.81) and (7.82) we observe that the time-averaged power flow
of a guided mode is directed in the longitudinal direction and is confined to
(1)
the slab region. The frequency-dependent constant Âm is determined by
the electromagnetic source strength.
TM-waves
We will now first discuss in detail the case of H-polarized waves or TM-
waves. We are looking for a solution of the equations, cf. Eqs. (7.2) - (7.4),
∂1 ∂1 Ĥ2 + ∂3 ∂3 Ĥ2 + ω 2 ε(i) µĤ2 = 0 , in D(i) , i = 1, 2, 3 , (7.83)
in which
(i)
(k1 )2 = ω 2 ε(i) µ0 − k32 , i = 1, 2, 3 . (7.88)
Equation (7.87) constitutes a second-order differential equation. In each
subdomain D(i) , with i = 1, 2, 3, the general solution can be written as
(i) (i)
Ĥ = Â(i) exp(jk1 x1 ) + B̂ (i) exp(−jk1 x1 ) . (7.89)
(i) 1
In order to satisfy the condition at |x1 | → ∞, k1 = (ω 2 ε(i) µ0 −k32 ) 2 is chosen
(i)
such that Im(k1 ) ≤ 0, and in addition we take B̂ (1) = 0 and Â(3) = 0. The
other frequency-dependent constants Â(i) and B̂ (i) follow from the boundary
conditions at x1 = a and x3 = 0, where
and
These four equations containing the four unknowns Â(2) , B̂ (2) , B̂ (3) and
Â(1) have a nonzero solution only if the determinant of the matrix of their
coefficients is zero. In this manner we find an equation which is known as
the eigenvalue equation or characteristic equation, because it leads to specific
190 electromagnetic waveguides
ε(2) κ(1) ε(2) κ(3)
κ(2) a = arctan (1) (2) + arctan (3) (2) + mπ, m = 0, 1, 2, · · · , ∞ .
ε κ ε κ
(7.98)
dielectric slab waveguides 191
This equation, in which the quantities κ(i) are related to the propagation
coefficient k3 through Eq. (7.95), is called the dispersion equation. Its solu-
tions
k3 ≡ k3;m (ω) , m = 0, 1, 2, · · · , ∞ ,
form an innumerable set of real and complex numbers. The real values k3;m
that satisfy the inequalities in Eq. (7.95) form a finite set with
max[k (1) , k (3) ] < k3;m < k (2) , (7.99)
where
1
k (i) = ω(ε(i) µ0 ) 2 , i = 1, 2, 3 , (7.100)
with m = 0, 1, 2, · · · , M−1. Each value k3;m in Eq. (7.99) can be considered as
the propagation constant of the m-th guided mode of the thin-film waveguide.
The electromagnetic field distribution corresponding to each k3;m is denoted
as the TMm -mode (m = 0, 1, 2, · · · , M − 1) of the slab waveguide. The
(i)
quantities κ(i) corresponding to k3;m are denoted as κm . From Eq. (7.89) it
then follows that
⎧ (2) (3)
⎪
⎪
⎪
cos(κm a−ψm ) exp[−κm (x1 −a)] exp(−jk3;m x3 )
⎪
⎪
⎪
⎪ when a < x1 < ∞ ,
(1) ⎪
⎨ (2)
Âm cos(κm x1 −ψm ) exp(−jk3;m x3 )
Ĥ2;m =
cos(ψm ) ⎪
⎪
⎪ when 0 < x1 < a ,
⎪
⎪
⎪
⎪
(1)
cos(ψm ) exp(κm x1 ) exp(−jk3;m x3 )
⎪
⎩
when − ∞ < x1 < 0 ,
(7.101)
in which
(1)
ε(2) κm
ψm = arctan (1) (2) . (7.102)
ε κm
The electric field strength then follows from Eqs. (7.84) - (7.85) as
⎧
⎪
⎪
k3;m (2) (3)
⎪
⎪ cos(κm a−ψm ) exp[−κm (x1 −a)] exp(−jk3;m x3 )
⎪
⎪ ωε(3)
⎪
⎪ when a < x1 < ∞ ,
⎪
⎪
⎪
⎪
(1) ⎨ 3;m
k (2)
Âm cos(κm x1 −ψm ) exp(−jk3;m x3 )
Ê1;m = ωε(2)
⎪
cos(ψm ) ⎪
⎪
⎪ when 0 < x1 < a ,
⎪
⎪
⎪
⎪ k3;m (1)
⎪
⎪ cos(ψm ) exp(κm x1 ) exp(−jk3;m x3 )
⎪
⎪ (1)
⎩ ωε
when − ∞ < x1 < 0 ,
(7.103)
192 electromagnetic waveguides
and
⎧ (3)
⎪
⎪ −κm
⎪
⎪
(2) (3)
cos(κm a−ψm ) exp[−κm (x1 −a)] exp(−jk3;m x3 )
⎪
⎪ (3)
⎪
⎪ jωε
⎪
⎪ when a < x1 < ∞ ,
⎪
⎪
⎪
⎪
(1) ⎪
⎨ −κm
(2)
(2)
Ê3;m =
Âm sin(κm x1 −ψm ) exp(−jk3;m x3 )
cos(ψm ) ⎪ jωε(2)
⎪
⎪
⎪
⎪ when 0 < x1 < a ,
⎪
⎪
⎪
⎪
(1)
κm
⎪
⎪ (1)
⎪
⎪ cos(ψm ) exp(κm x1 ) exp(−jk3;m x3 )
⎪
⎪ jωε (1)
⎩
when − ∞ < x1 < 0 .
(7.104)
From Eqs. (7.101) and (7.104) it follows that the time-averaged power flow
density in the x1 -direction,
∗ ∗
1
2 Re[(Ê m × Ĥ m ) · i1 ] = 12 Re[−Ê3;m Ĥ2;m ]=0, (7.105)
in each of the three domains. The time-averaged power flow density in the
x3 -direction follows from Eqs. (7.101) and (7.103) as
∗ ∗ ]
1
2 Re[(Ê m × Ĥ m ) · i3 ] = 12 Re[Ê1;m Ĥ2;m
⎧
⎪
⎪ k3;m (2) (3)
⎪
⎪ [cos(κm a−ψm )]2 exp[−2κm (x1 −a)] , a < x1 < ∞ ,
⎪
⎪ ωε(3)
⎪
⎪
2⎪
 (1) ⎪
⎨
m k3;m (2)
=
cos(ψm ) ⎪
[cos(κm x1 −ψm )]2 , 0 < x1 < a ,
⎪
⎪ ωε(2)
⎪
⎪
⎪
⎪
⎪
⎪ k3;m
⎪
⎩ (1)
[cos(ψm )]2 exp(2κm x1 ) , −∞ < x1 < 0 .
ωε(1)
(7.106)
From Eqs. (7.105) and (7.106) we observe that the time-averaged power flow
of a guided mode is directed in the the longitudinal direction and is confined
(1)
to the slab region. The frequency-dependent constant Âm is determined by
the electromagnetic source strength.
propagation properties of a dielectric slab waveguide 193
k3;m
neff ,m = (7.107)
k0
3.5 ..............................................................
............................
....................................................
...............................................
neff ,m n(3) = 1.0 .....
. ..
...
.
......................
..
...
.
.
. .
.
..
..
..
................................................
..............................
....................... ...........................................
n(2) = 3.5 .....
....................
..................... ......... .
.
......................................
............................................
6
. ..
... .... ...
....... ...... ........................
...... ...... .......................
n(1) = 3.2 ...... ......
...... ......
.......... .........
......... ........
........................ .........
..... ......... . .....................
..
..
.... .... .
.. .. . . . . .....................
..... ....... ....... ........ .......... .........
... . . .. .. ................... ....... . ...........................
.. .. ...... ...... .. .
... ... ......... .........
3.4 ... ...
... ...
...... ......
...... .....
...... ......
........ .......
....... .......
....... .......
TE 0....... ....... ....... ......... ....................
.. .. TE . .
1............ ............
..
.. ..
...... .......
...... .....
........
........ ....
... ...
... ...
. .
.... .....
TE 2............................. ....... .......
....... .......
.... ... ..
.... .......
... ...
...... ......
..... .....
TE 3 .
.....
...... .....
..................
... ... TM ... .... ..... ..... ..... ......
... ... 0 ... .... ..... ...... ..... .......
... ...
.. ...
... ...
... ..... 1 TM ..... ....
..... ........
...... .....
..... ..........
.
... ...
..
. .
... ....
.
. . ..
..... .....
.
TM 2 . ..
..... .....
..... ......
. . .
... ...
... ...
... ...
... ...
.... ....
.... ...
..... .....
..... .....
TM 3
3.3 .... ....
.. ..
.
.
.. ....
.
... ...
.
... ...
.. ... ..... ...
...... .........
..... .....
.... .... ... .... . . ... ..... ..
...... .........
.
. . ... ... .. .. .... ....
.. .. ... ... .... ....
.. .. ... ... ... ... ... ....
.. .
.. .. ... ... ... ... ... ....
... .... ... ... ..... ..... ...... ......
.. .. ... ..
. .. .. .. ..
.. .. .. ... ... ... ... ...
.. .. ... .. ... .. ... ...
... .. .. ... ... ....
... .... ... .... ..... ..... ..... ......
.. .. .. ..
.. ..
.. ... ... .....
.. .. ... ... .... ....
.. ... ... ... ... .... .... ..
... .... .... .... ..... ..... .... ...
3.2
0 0.5 1.0 1.5 2.0
- a/λ0
Figure 7.11. Dispersion curves for the guided modes of a GaAs film located on a
AlGaAs substrate (a = film thickness, λ0 = 2π/k0 = wavelength in free space).
! 1 &
c0 [(n(1) )2 − (n(3) )2 ] 2
ωc,m = 1 arctan 1 + mπ .
a[(n(2) )2 − (n(1) )2 ] 2 [(n(2) )2 − (n(1) )2 ] 2
(7.113)
Substitution of k32 = ω 2 ε(1) µ0 = k02 (n(1) )2 in Eqs. (7.98) and (7.95) yields an
equation from which the cut-off frequencies of the TM-mode can be derived
as
! 1 &
c0 [(n(1) )2 − (n(3) )2 ] 2 ε(2)
ωc,m = 1 arctan 1 + mπ .
a[(n(2) )2 − (n(1) )2 ] 2 [(n(2) )2 − (n(1) )2 ] 2 ε(3)
(7.114)
The mode with the lowest cut-off frequency (the TE0 -mode) is often referred
as the fundamental mode.
Like for the parallel-plate waveguide in Section 6.2 we show in Fig. 7.12,
(2)
the (k1 , k3 )-plane for three different TE-modes in a lossless, asymmetric,
dielectric slab waveguide. Further it is interesting to investigate the influence
of the operating frequency on the amplitude distribution of the guided mode.
propagation properties of a dielectric slab waveguide 195
(2)
k1
..
.......
.......
.. .... ..
...
....
...
...
...
...
...
...
...
...
...
...
...
...
.
... ..
. ..... ..
... ..
.........................................................
...... .. .
.. ..
.......... ....
...........
.........
...
... ....... ..... ... ........
.......
..
....... ... .......
.. ...... (2)
{k1;2 , k3;2 }
. ......
... ......
.. .
.
. ......
... .......
. .
. ......
... ...
. .
.
. ............. ..........
... .
...... .
.
. .
.. .....
... ..
............. .
.
. .... .....
.....
... ..... ........... .... ...
.. ....
.... ...... .....
..... .
. ... ....
.
... .. .
.
. .... ...
...
..
. ..... .
. .
. ...
. . . (2)
propagating {k1;1 , k3;1 }
.. .
..... .
. ... ...
.. ....... .
. ..
... ..... .. ... ...........
. . ... ......... ....
modes ..
. .....
.....
.
.
. ... .
.... ...
.. ...
k (2)
..
. . .
.
. ... . .
... . ...
... ..... . ....
... ... ..... .... ... ....... ...
... ..... ... ... ....... ...
...
... ..
.
.....
.. .
. .
.... ............ ...
... ....
.....
..... ..
. .... .......... ...
...
... ... ..... .
.
. .
... ........
. ...
..... .. .... (2)
{k1;0 , k3;0 }
... . .
. . .
... .. .
... .
.
..
......
..
.
.
.
. ...
..
...
.........
.
... ... . .
. .
.........................
.
... ...
.
.....
.. .
.
. .
...
.
...
...... ....
... ... ... ............ ..
...
.....
..... .... ..... .............
. .......
..........
.. .
.. ................ ..
...
..... .. ... ....... ................ ..
.... ..... .. .......... .............................. ..
... ......................... ..
.........................................................................................................................................................................................................................................................................................................................................................................................................................
..
O k3
(2)
Figure 7.12. The (k1 , k3 )-plane for the lossless dielectric slab waveguide.
In Fig. 7.13, we have shown the dispersion curve of a TE0 -mode and its
amplitude distribution for three different values of the operating frequency.
Close to cut-off, a mode stretches far out in substrate and superstrate. Its
effective index of refraction neff approaches the value n(1) . Far from cut-off,
the mode is entirely concentrated in the thin film. As can be expected, now
its effective index neff approaches the value n(2) .
In Fig. 7.14, we have shown the dispersion curves of the TEm -modes,
m = 0, 1, 2, and their amplitude distributions for a particular frequency of
operation.
It is finally noted that from Eqs. (7.77) and (7.101), it can be seen that
in the substrate and the superstrate a guided mode locally behaves as a
non-uniform plane wave. The planes of equal phase and equal amplitude are
perpendicular to each other.
196 electromagnetic waveguides
n(2) ..............................
............................................................
.......................................
.................................
neff ,m n(3) = 1.0 .
...
...
...
..
.
..................
.......................
..
...........
n(2) = 3.5 .........
........
........
6 n(1) = 3.2 .
......
..
.........
......
.....
.....
.
.....
...
...
...
TE ...
0......
.
...
...
....
...
...
...
..
.
...
...
..
..
...
..
...
...
....
.
..
..
...
....
.
...
..
..
..
.
..
...
n(1) 0
ω1 ω2 ω3 ω4 4πc0 /a
- ω
ωc,0
x1 x1 x1 x1
...
...
6 ..... 6 .... 6 .... 6
... ..
.. ... ...
... ... ... ...
...
...
.....
......
...
...
...
.
.
.
.
. .. .
...
...
.
n(3)
....... ....... ... ...
........... ....... .......
..... ..... .........
. . . . . . . . ... . . . . . ...... . . . . ......... . . . . ................. . . . . . . . . . . .
. . . . . . . . ..... . . . . . . .... . . . . .
....... . .............
.... . . . . . . . . . . .
. .... . . .... . . ..... . ..... .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. ..... .
.
.
.
.
.
.
.
. ..... . .
.
.
.
.
.
...... .
.
.
.
.
.
.
.
. ........ .
.
.
.
.
.
.
.
.
.
.
.
. n(2)
. . . . . . . . . .... . . . . . ........ . . . . .............. . . . . . .......... . . . . . . . . . .
. ... .... .........
... ...... ........ .........
..
. ....... ....... .....
.
... ...
. ..
. ..
.
... ... ... ....
...
..
...
..
...
....
...
...
n(1)
..
..
.
...
- .
.
.. -
.
.
.. -
.... -
Ê(x1 , jω1 ) Ê(x1 , jω2 ) Ê(x1 , jω3 ) Ê(x1 , jω4 )
n(2) .......................................
............................................................
..............................
.................................
neff ,m n(3) = 1.0 ..
....
.....
..................
....................... .......................
.......... .....................
n(2) = 3.5 .......
.........
..........
..............
................
..................
6 .
...
.. ....
......
.
.. ....
n(1) = 3.2 ......
.....
......
.........
..........
...........
............
...
..... ........ ...........
......
. .. ......... ..........
.
.
.
.... ............
...
..............
..... ..
... ...... .........
... ...... ........
... ...... .......
... .......
TE 0...... .
. ....... .
...........
.
...
TE 1............
... .....
......
......
... ....
... TE 2..............
.... .
.. ...
..
.. .....
... .... .....
... ... .....
... ... .....
..
. ....
. .......
... .. ....
... .....
... ... ....
.. ... ...
..
... .... .....
. ..
.. ... ...
... ... ...
... ... ...
.... ... .....
. ... ..
.. ...
.. ... ...
... ... ...
.... ... ....
. .. ..
... .. ...
.. .. ...
.. .. ..
..
. ... ....
.. .. ...
... ... .....
n(1) 0
ωc,0 ωc,1 ωc,2 ω 4πc0 /a
- ω
x1 x1 x1
.... 6 ... 6 6 ....
... .... ...
.. .. ..
... ...
....
0 TE -mode ..
...
..
1 TE -mode ....
TE -mode 2
n(3)
..
...... ...
...
. ..........
.......... ................
. ........................
.
... ...
. . . . . . . . ..
. ........... . . . ........ .. ... .. . . . . . . . . . . . . ................. . . . . . . . .
. . . . . . . . .
. .......... . . .
. ................. . . . . . . . . . ......................... . ... . . . . . .
. ... ................. . ........................ .
. . . . . . . . . . . . ................ . . . . . ................. . . . . . . . . (2)
. . . . . . . . . ..
.
...... . . . . . . .
.....
.................. . . . ... ...
... ......................... . . . . . . . . n
. . . . . . . . . ............ . . . . . . . . . ......... . . . . . . ...................................... . . . . . . . .
.. .. .
......... ....
............. ........
..... ....... .........
... .... .....
.
.....
.... .... ...
.
.. .. ...
....
... ...
...
...
...
n(1)
..
. - ....
. - ....
. -
Ê(x1 , jω ) Ê(x1 , jω ) Ê(x1 , jω )
Exercise 7.1
A parallel-plate waveguide is used to guide TE-waves. The configuration is
depicted in Fig. 7.6 with ε = 4ε0 , µ = µ0 and a = 10 cm while the frequency
of operation is f = 1 GHz.
(a) Find the propagating and the evanescent modes.
(b) Give the expressions for the electric and magnetic field strengths of
the propagating mode(s).
Exercise 7.2
Find the maximum power that can be propagated in an air-filled parallel-
plate waveguide, with width a = 1 cm and characteristic length w = 5 cm,
without causing breakdown. In air breakdown occurs at approximately E =
2 × 106 V/m. Use a safety factor of 10.
Exercise 7.3
For frequencies above the cutoff frequencies, the propagation constants of
a parallel-plate waveguide are associated with the so-called ‘guided wave-
lengths’, λg,m , as k3;m = λ2π
g,m
for ω > ωc,m . Find the guided wavelengths of
the TEm -modes, m = 0, 1, 2, · · ·, with f = 300 MHz, c = 3 × 108 m/s and
a = 0.75 m.
Exercise 7.4
EHF electromagnetic waves, see Table 3.1, (also called millimeter waves) may
be guided by dielectric slab waveguides. Consider a slab, with ε = 10ε0 , in
free space. By selecting a proper thickness a it may be achieved that for
any frequency of the entire EHF frequency band only the TE0 -mode can
propagate in the slab. What is the range of possible values of a?
Exercise 7.5
Consider a symmetric dielectric slab waveguide in the following configura-
tion:
Superstrate : a < x1 < ∞, ε(1) ,
Slab : −a < x1 < a, ε(2) ,
Substrate : −∞ < x1 < −a, ε(1) .
exercises and problems 199
(b) Solve the equations for Â(2) and B̂ (2) in terms of Â(1) and also in terms
of B̂ (3) .
( c ) Use the results obtained in (b) to eliminate Â(2) and B̂ (2) and find
the two solutions with for each solution the corresponding relation
(2)
for κm a. Hint: One solution is found for m = 0, 2, 4, · · · (the even
solution), while the other is found for m = 1, 3, 5, · · · (the odd solution).
(d) Use the results in (b) and (c) to show that the electric field strength
is given by ⎧ (2) (1)
⎪
⎪ cos(κm a) exp[−κm (x1 − a)] exp(−jk3;m x3 ),
⎪
⎪
⎪
⎪ a < x1 < ∞,
⎪
⎨
(even) cos(κ(2) x ) exp(−jk −a < x1 < a,
Ê2;m = Ĉm m 1 3;m x3 ),
⎪
⎪
⎪
⎪ (2) (1)
⎪
⎪ cos(κm a) exp[κm (x1 + a)] exp(−jk3;m x3 ),
⎪
⎩
−∞ < x1 < −a,
with m = 0, 2, 4, · · ·, and
⎧ (2) (1)
⎪
⎪ sin(κm a) exp[−κm (x1 − a)] exp(−jk3;m x3 ),
⎪
⎪
⎪
⎪ a < x1 < ∞,
⎪
⎨ (2)
(odd)
Ê2;m = Ĉm sin(κm x1 ) exp(−jk3;m x3 ), −a < x1 < a,
⎪
⎪
⎪
⎪ (2) (1)
⎪
⎪−sin(κm a) exp[κm (x1 + a)] exp(−jk3;m x3 ),
⎪
⎩
−∞ < x1 < −a,
(even) (odd)
with m = 1, 3, 5, · · ·, and express the coefficients Ĉm and Ĉm in
(1)
terms of Âm .
Problem 7.1
For a parallel-plate waveguide find the relations for the phase velocity and
the group velocity, for propagating modes, and express them in terms of the
actual wave speed, c, the angular radial frequency of operation, ω, and the
modal cutoff angular frequency, ωc,m .
200 electromagnetic waveguides
Problem 7.2
A dielectric slab is at x1 = 0 coated with a perfectly conducting foil, see
Fig. 7.15. Find the solution for the electric field strength in the configuration
in case of guided TE-waves in the slab. Hint: you may use your results in
Exercise 7.5 to your advantage.
ε(1) , µ0
x1 = a
ε(2) , µ0 propagation
.......
direction
................................................................
....
... i1
..........
p ... ... i3
.................... x1 = 0
Problem 7.3
In this chapter we have considered one open waveguide structure, viz., the
slab waveguide, and have found solutions for guided electromagnetic waves.
One can also investigate whether guided wave solutions exist in a two-media
configuration (see Fig. 7.16). In that case the guided wave is a surface wave
that propagates along the interface between two different dielectric media
and is evanescent in both directions away from the interface. We simply
assume a solution for TM-waves of the type
Ĥ2 (x1 , x3 , jω) = Ĥ(x1 , jω) exp(−jk3 x3 ), −∞ < x1 < ∞ ,
and
Ĥ = B̂ (1) exp(−κ(1) x1 ), 0 < x1 < ∞ ,
Ĥ = Â (2)
exp(κ (2)
x1 ), −∞ < x1 < 0 ,
ε(1) , µ0 .. i1
..........
p .... .. i3
..................... x1 = 0
ε(2) , µ 0 propagation
.......
direction
................................................................
....
(1) 1 (2) 1
with κ(1) = jk1 = [k32 − ω 2 ε(1) µ0 ] 2 > 0 and κ(2) = jk1 = [k32 − ω 2 ε(2) µ0 ] 2 > 0,
where the condition Ĥ2 → 0 if |x1 | → ∞ is satisfied.
(a) Under what condition can there be a solution of this form?
(b) For an electron plasma the permittivity has a negative sign, when the
frequency is below the plasma frequency. When one of the two media
is such a plasma, give the solution for Ĥ2 and draw the amplitude of
the magnetic field as a function of x1 .
Excitation of
Two-dimensional
Electromagnetic Waves
A reference frame is introduced such that the sheet coincides with the
plane x3 = 0 (Fig. 8.1). Let the impressed electric current be uniform as a
function of x2 and flow from x1 = 12 a to x1 = − 12 a, then
!
−Iˆ∆ (x1 , s)δ(x3 ) , |x1 | < 12 a ,
Jˆ1ext = Jˆ2ext = 0 , Jˆ3ext = 0 , (8.1)
0, |x1 | > 12 a ,
where δ(x3 ) denotes the one-dimensional unit impulse (Dirac distribution),
and Iˆ∆ (in A/m) is the electric current per unit length (of the x2 -direction).
Since the electric-current sheet carries no magnetic current, we have
K̂1ext = 0 , K̂2ext = 0 , K̂3ext = 0 . (8.2)
The medium is assumed to be homogeneous and lossless with constitutive
constants ε and µ. Since the configuration is independent of x2 , the elec-
tromagnetic field is two-dimensional and in Chapter 4 we have seen that
the frequency-domain field equations (2.54) - (2.59) separate into two in-
dependent set of equations, cf. Eqs. (4.4) - (4.6) and Eqs. (4.7) - (4.9).
From Eqs. (8.1) - (8.2) we conclude that K̂1ext , K̂3ext and Jˆ2ext are equal to
zero, hence only a parallelly polarized electromagnetic field occurs. The non-
zero field components Ê1 , Ê3 and Ĥ2 satisfy the following set of equations
(s = jω, σ = 0, cf. Eqs. (4.4) - (4.6))
!
Iˆ∆ (x1 , jω)δ(x3 ) , |x1 | < 12 a ,
∂3 Ĥ2 + jωεÊ1 = (8.3)
0, |x1 | > 12 a ,
−∂1 Ĥ2 + jωεÊ3 = 0 , (8.4)
∂3 Ê1 − ∂1 Ê3 + jωµĤ2 = 0 . (8.5)
In Chapter 4 we have seen that in a domain outside the source distribution,
solutions of Eqs. (8.3) - (8.5) in the form of plane waves exist. We therefore
assume that the general solution consists of an infinite superposition of plane-
wave constituents
⎧
⎪
⎪ 1 ∞
⎪
⎨ 2π 1 (k1 , jω) exp(−jk1 x1 − jk3 x3 )dk1 , x3 > 0 ,
ê+
k1 =−∞
Ê1 (x1 , x3 , jω) =
⎪
⎪ 1 ∞
⎪
⎩ ê− (k1 , jω) exp(−jk1 x1 + jk3 x3 )dk1 , x3 < 0 ,
2π k1 =−∞ 1
(8.6)
sheet emitter with a parallel electric current 205
ε, µ ...........
.......... ..............
........
ε, µ
' (...................
.......
.......
......
... x = 12 a 1
H .....
...
. H
........ i1
....
O r ............................
...
...
...
... i3
...
..........
.......
' J ext .
x1 = − 12 a (
..... .......
....... .......
....... .......
............. .........
..
.
..... H H ..........
x3 < 0 x3 > 0
⎧
⎪
⎪ 1 ∞
⎪
⎨ ê+ (k1 , jω) exp(−jk1 x1 − jk3 x3 )dk1 , x3 > 0 ,
2π k1 =−∞ 3
Ê3 (x1 , x3 , jω) =
⎪
⎪ 1 ∞
⎪
⎩ ê− (k1 , jω) exp(−jk1 x1 + jk3 x3 )dk1 , x3 < 0 ,
2π k1 =−∞ 3
(8.7)
and
⎧
⎪
⎪ 1 ∞
⎪
⎨ ĥ+ (k1 , jω) exp(−jk1 x1 − jk3 x3 )dk1 , x3 > 0,
2π k1 =−∞ 2
Ĥ2 (x1 , x3 , jω) =
⎪
⎪ 1 ∞
⎪
⎩ ĥ− (k1 , jω) exp(−jk1 x1 + jk3 x3 )dk1 , x3 < 0,
2π k1 =−∞ 2
(8.8)
where ⎧
⎨
1 1
(ω 2 εµ − k12 ) 2 , |k1 | ≤ ω(εµ) 2 ,
k3 = (8.9)
⎩ −j(k 2 − ω 2 εµ) 21 , |k1 | > ω(εµ) 2 .
1
1
Note that the definition of the square root of k3 takes care of the bounded-
ness of the integrals as |x3 | → ∞. This guarantees the boundedness of the
electromagnetic field components as |x3 | → ∞. From Eqs. (8.3) and (8.4) it
206 excitation of two-dimensional electromagnetic waves
− − −
follows that the quantities ê+ + +
1 , ê3 , ê1 and ê3 are related to ĥ2 and ĥ2 as
k3 + −k1 +
ê+
1 = ĥ , ê+
3 = ĥ , (8.10)
ωε 2 ωε 2
−k3 − −k1 −
ê−
1 = ĥ , ê−
3 = ĥ . (8.11)
ωε 2 ωε 2
1
When x3 > 0 and |k1 | < ω(εµ) 2 we observe that the electromagnetic
field consists of a superposition of uniform plane waves propagating in the
1
{k1 , 0, k3 }-direction, while when |k1 | > ω(εµ) 2 the constituents are (non-
uniform) evanescent waves decaying exponentially in the x3 -direction.
The presence of the source distribution is accounted for by the application
of the excitation conditions
!
Iˆ∆ (x1 , jω) , |x1 | < 12 a ,
lim Ĥ2 (x1 , x3 , jω) − lim Ĥ2 (x1 , x3 , jω) =
x3 ↓0 x3 ↑0 0, |x1 | > 12 a ,
(8.12)
which is a consequence of Eq. (8.3), and
lim Ê1 (x1 , x3 , jω) − lim Ê1 (x1 , x3 , jω) = 0 , −∞ < x1 < ∞ , (8.13)
x3 ↓0 x3 ↑0
which is a consequence of Eq. (8.5). From Eqs. (8.13) and (8.6) it directly
follows that
ê− +
1 = ê1 , (8.14)
and from Eqs. (8.10) and (8.11) it follows that
ĥ−
2 = −ĥ2 .
+
(8.15)
We observe that the field components Ĥ2 and Ê3 are odd-symmetric func-
tions of x3 , while Ê1 is an even-symmetric function of x3 . Application of
Eqs. (8.12) and (8.15) to Eq. (8.8) leads to
∞ ! 1
ˆ |x1 | < 12 a ,
1 2 I∆ (x1 , jω) ,
ĥ+
2 (k1 , jω) exp(−jk1 x1 )dk1 =
2π k1 =−∞ |x1 | > 12 a .
0,
(8.16)
The left-hand side of this equation is a Fourier integral with respect to the
spatial variable x1 and with the transform parameter k1 . Fourier’s inversion
sheet emitter with a parallel electric current 207
theorem yields
1
a
2
1ˆ
ĥ+
2 (k1 , jω) = 2 I∆ (x1 , jω) exp(jk1 x1 ) dx1 . (8.17)
x1 =− 21 a
With Eqs. (8.17), (8.15), (8.10) - (8.11) and (8.6) - (8.8) the electromagnetic
field is determined completely. Once Iˆ∆ is prescribed, we can calculate the
integral of Eq. (8.17) either analytically or numerically, and once ĥ+ 2 is de-
termined we can calculate the Fourier integral of Eq. (8.8) either analytically
or numerically. The other field components can be computed in a similar
way.
When the point of observation is far enough away from the emitter, i.e.,
1
r = (x21 + x23 ) 2 (8.18)
when x3 > 0. In the case that r is large, the exponential function in the
integrand oscillates violently along the integration path, except in a point
where k1 x1 + k3 x3 is stationary. In such a stationary point the exponential
function does not vary rapidly in a small interval along the integration path.
In each part outside the stationary points the contribution to the integral
is negligibly small. Around the stationary point the contribution to the
integral is more significant. This fact is employed in Kelvin’s principle of
stationary phase. The stationary point k1 = k1s is the root of the equation
1
∂k1 [k1 x1 + k3 x3 ] = ∂k1 k1 x1 + (k 2 − k12 ) 2 x3 = 0 , (8.20)
208 excitation of two-dimensional electromagnetic waves
where
1
k = ω(εµ) 2 (8.21)
has been used. Hence, this point k1 = k1s follows from
k1
x1 − 1 x3 = 0 (8.22)
(k 2 − k12 ) 2
as
x1 x1
k1s = k 1 =k . (8.23)
(x21 + x23 ) 2 r
Note that
x3 1
k3s = [k 2 − (k1s )2 ] 2 = k
. (8.24)
r
Only around this stationary point the integrand of Eq. (8.19) will contribute
significantly. Assuming that the function ĥ+ 2 varies much slower than the
exponential function, we obtain the approximation
∞
1 + x1
Ĥ2 ≈ ĥ (k , jω) exp(−jk1 x1 − jk3 x3 )dk1 . (8.25)
2π 2 r k1 =−∞
1
x3 + x1 k 2
Ĥ2 ≈ ĥ2 (k , jω) exp(−jkr + j 14 π)
r r 2πr
1
when r = (x21 +x23 ) 2 → ∞ and x3 > 0 . (8.30)
holds, hence the electric field vector Ê = {Ê1 , 0, Ê3 } is perpendicular to the
direction of observation x = {x1 , 0, x3 }, see Fig. 8.2.
All the results pertaining to the electromagnetic field quantities in the
domain x3 < 0 can be derived in a similar way.
The time-averaged power flow density in the far field is given by
∗ ∗
1
2 Re[Ê × Ĥ ] = 1
2 Re[−Ê3 Ĥ2 i1 + Ê1 Ĥ2∗ i3 ]
2
1
x3 µ 2 k x1 x1 x3
≈ |ĥ+
2 (k , jω)|2 [ i1 + i3 ] ,
r ε 4πr r r r
2
1
x3 µ 2 k x1 x
≈ |ĥ+ (k , jω)|2 ,
r ε 4πr 2 r r
1
when r = (x21 +x23 ) 2 → ∞ , (8.34)
E
. ......
........
...
...
...
...
.............
........
S
(...................
...
.
.........
..........
.......
.......
H
.
......
..
... x
.......
.......
.......
.......
....... ......
. .θ ...... ....
.
.
.......
.......
..
...
................. .......
.. .. ............
..............
i1
.... .......
.......
.......... .......
.......
... ............
O r
............. .....
i3
......................
Figure 8.2. The orientation of the electromagnetic field vectors in the far field of
an emitter with parallel electric current.
sheet emitter with a parallel electric current 211
where x/r = x/|x| is the unit vector in the direction x of observation. The
electromagnetic power flow is in this direction and decays with 1/r as a
function of r. In conclusion we observe that the electric and magnetic field
strengths have, in the far-field region, the structure of a cylindrical wave that
expands radially from the origin of the coordinate system. In the angular
direction (see Fig. 8.2), the field amplitudes depend on θ via
x1 x3
cos(θ) = , sin(θ) = . (8.35)
r r
The electric and magnetic field strengths in the far-field region are transverse
to the local direction of propagation and behave locally as if the wave were a
uniform plane wave, with exp(−jkr) = exp(−jk1 x1 −jk3 x3 ), traveling in the
direction {k1 = k cos(θ), 0, k3 = k sin(θ)} away from the source. The time-
averaged power flow vanishes in the i1 -direction (θ = 0). All field amplitudes
are proportional to ĥ+ 2 (k cos(θ), jω). Consequently, in the far-field region
only the spatial Fourier transform of the electric current through the emitter
(see Eq. (8.17)) at the subset of the transform parameter k1 = k cos(θ) is
”seen”.
To study the angular dependence of the far-field characteristic, the di-
rective gain is introduced. This directive gain D(θ) is defined as the power
flow in the observation direction, normalized to the angular-averaged power
flow in the far field, viz.,
∗ x
1
2 Re (Ê × Ĥ ) ·
r
D(θ) = 2π . (8.36)
1 ∗ x
1
2 Re (Ê × Ĥ ) · dθ
2π θ=0 r
Using Eqs. (8.34) and (8.35) we find
| sin(θ) ĥ+
2 (k cos(θ), jω)|
2
D(θ) = .
(8.37)
1 2π
| sin(θ) ĥ+ 2
2 (k cos(θ), jω)| dθ
2π θ=0
In Fig. 8.3, the directive gain is presented for the case that Iˆ∆ is a constant
function, i.e.,
Iˆ∆ (x1 , jω) = î∆ (jω) , |x1 | < 12 a , (8.38)
212 excitation of two-dimensional electromagnetic waves
D(θ) D(θ)
...... ......
......... .........
a/λ = 0.1 ... .... ..
...
a/λ = 0.3 ... .... ..
...
.... ....
. . . .10 ..
.... . . . . . . . . 10 ..
.... . . . .
. . ... . . . . ... . .
. . ... . . . . ... . .
. . .. . . ..... . . .. . . .....
. .
. . ... . . . ...
. . .
.
.
. . ........... . .
.
.
.
. . . ...........
. . 1 .
. . . . . . 1 . .
.
. .
.. . . . . ... . . . . .
. .θ . .
. .
. . . . ... . . . . .
.
. .
.
θ
. . ..... ..
. .. . . .
.. ..... ..
. . . .
. . . .. . . . . .. .
. . . ......................... ... ............................ ..................... ... .....................
.......... . ....... . ..
...... . ...... . . . . . .
.
.......... . .... . ..........
. .....
. . . .
. . ...... .. .... .. .... . . . .... . . . . .
....... . ..... .. .... . . . .... . .
. . .. .. .
.
.
. .
0.1
.... .. ... .
.
.
. .
.
. .
.
.
.
. ....
.
.
. .
.... .. .
.
.
.
.
.
.
.... .. .... .
.
0.1 .
. .
.
.
.
.
. .
..
. ....
.
.
.
. . . .. . . . ..
. ... . . . . . .
. . .... .... .... . .
. . .
. ... . . .
... . .
. . .... .... .... . .
. . .... .
. .. .
.
. .
. . . . . . .
. .. .. . . .. .
.
. .
. . . . . . .
. .. . .
.. . . .. .......
.
..
. .. . . .. ......
. .
.. .
. . . . . ...... .. . . . . ... . . .. . . . ...... . . . . . . ... .
.
. . .
. .
. . .. . . . . .
. . .
... . .
. ........... .
. . . ..... . . . .
... . .
. ............ .
. . . ..... .
. .... . . . .
. . . .
. ...
.
. . . .... . . . .
. . . .
. ...
... .
. ... . . . ....... . . . . ... . . . ........ . . .
. ..
.. . . . . .... . . .
.
. . ...
. . . ... . . . .... . . .
.
. . .
.
.
. .
. . . . .. ... . . . . .. .. . . . . .. ... . . . . ..
. .. . . . . ... . .
. ... ... . . .
. .. . . .. .. . .. . . ...
. .. . . .
. .. . . . . . ... .... . . . . . .... . . . . . .. ... . . . . . . .
.
...
.... . . . .. . .. .. .. . . . .... . . ... . . . .. . .. .. .. . . . ...... .
.
.. . ....... . .. ..... .... . ... .. . . ....
. ..... ..
..... . . ..
. .... ..
. .... . .
. . ..... .
..... . . ...... . . ....... . .......... .
.
. . .
. . ..
...
. ..
. . .. . . . .. . ..
... ..... . .. .
. ...... . . ....... . .
.. . ........................... ..... ..
........................... . .. .. . ..................... ....... ..
.................... . ..
. . . .
. . . . . . . . . .
. .. .
. . .. . . .. .
. . .. .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . .
. . . . . . . .
. . . .
. . . . . . . .
.. . . .. .. . . ..
. . . . . .
. . . . . . . . . .
. . . . . . . . . . . .
..
D(θ) D(θ) ..
...... ......
....... .......
a/λ = 1 .. .... .. a/λ = 3 .. .... ..
.... ....
.... ....
10
. . . ...... . . . .
10
. . . ...... . . . .
. .
. . . ....
. .
. . . ....
. .
. . . . . .
. .... . . ..... . .... . . .....
. . .. . .... . . .. . ....
.
. . ... . . .......... .
. . ... . . ..........
.
.
. .
1
. . ... . . .
.
.
.
. .
. θ .
.
. .
1
. . ... . . .
.
.
.
. .
. θ
.
.
.
.. . .... . .. .
.
.
.
.. . .... . .. .
. . . ... . .. .
. . . ... . . . .
. . ... . .. . . ... . ..
. . . . . .
. .
. . . . . . . . .
. .
. . .
. . .
. . . . . .
. . .
. ................ 0.1 . . . . .. 0.1 . . .
. .... ........ . . . ... . . . .................................. .
.
. . . ... . . .
. . ..
. ..
......... . . . ....... .
. . .. . . .......
. . . ......
.
.
.
..
. . . . . .. .
.
.... . . ...... .
.. ..
. ........
. . . ..... . . ............... . . ..
. . . ............
.
. . . .
. . . . . . ..... .. ..... . . . ... . . ..... . ...... ..... .. ...... ....... . ....
... .. . .. . ... . . . . .. . . .
. ............................................................................................ .... ......................... ................................................................ .
.
. .... . . . . ...... ... ...... . . . . ... . ................... .............. .. .. ........... ....................
. .... . . . . ... ... ... . .
......... . . .
.
.... . ......... . . .......................................................... . .
.....
....
. ... . . . . . ... . . ... . ..
.
..
..
. .. .
.
..
.
..
.
..
...
.. ..
. ........ . . . . ..................... . .. .
. .
.
... . . .. ... . . . .. . ...... . . ................................... .................................. . ..
. . .
. ..
. ...
... .
. .
. . . ..... ...... . . .
. .
.. . . ............ . ... . ....... ................................................... .................................................................. .
.
....
.......
. ... . . . .
..... ....
. . . . . .
.... . . .
..
.
............................... ...... .. ...... .. ....
.. . . . . . . . .
.. . ... ..
.
....... ....... . ....... ................................. .
. . .. ... . ...
. ...
..... . .
. . ....
.... .
..
...........
.
. . . ... . . . ..................... . .. . . .................. . .
....... . . . ...... .. .. .... . ........ . . . . .. . . . .. .
. .... ....... . . . . . . . ..................................... .
. . . . . .
.
. . .................... . . . . .
. . . . . . . . .
. . . . . . . . . . . . . . . . . .
.. . . . . . . .. . . . . . .
.
.
. . . . .. .
.
. . . . ..
. .. . . .. . . .. . . .. .
. . . . . . . . . . . . . . .
. . . . .
. . . . . .
. . . . . .
. . . . . . . .
. . . .
. . . .
. . . . .
.
.. . . .. . .
. . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . .
Figure 8.3. The directive gain of the sheet emitter with parallel electric current.
sheet emitter with a perpendicular electric current 213
sin( 12 k1 a)
ĥ+
2 = î∆ . (8.39)
k1
The directive gain is presented for various values of a/λ, where λ = 2π/k
is the wavelength. For values of a/λ → 0, we observe that ĥ+ 2 ≈ 2 a î∆
1
and D(θ) ≈ 2 [sin(θ)] . From Fig. 8.3, we observe that this approximation
2
A reference frame is introduced such that the sheet coincides with the
plane x3 = 0 (Fig. 8.4). Let the impressed electric current flow uniformly in
the x2 -direction. It is present at x3 = 0 from x1 = − 12 a to x1 = 12 a. Hence,
!
Iˆ∆ (x1 , s)δ(x3 ) , |x1 | < 12 a ,
Jˆ1ext = 0 , Jˆ2ext = Jˆ3ext = 0 , (8.40)
0, |x1 | > 12 a ,
The non-zero field components Ĥ1 , Ĥ3 and Ê2 satisfy the following set of
equations (s = jω, σ = 0, cf. Eqs. (4.7) - (4.9))
⎧
⎪
⎪ 1 ∞
⎪
⎨ ĥ+ (k1 , jω) exp(−jk1 x1 − jk3 x3 )dk1 , x3 > 0,
2π k1 =−∞ 1
Ĥ1 (x1 , x3 , jω) =
⎪
⎪ 1 ∞
⎪
⎩ ĥ− (k1 , jω) exp(−jk1 x1 + jk3 x3 )dk1 , x3 < 0,
2π k1 =−∞ 1
(8.45)
ε, µ ..........
......... ..............
........
ε, µ
' '...................
.......
.......
.......
(x = 1 a
E ( 1 2 E
( .
........ i1
....
( ....
...r...........................
O( i3
(
(
' J ext x1 = − 12 a '
.......
....... .......
....... .......
....... ..........
..
.............
.
. E E ..........
.
x3 < 0 x3 > 0
⎧
⎪
⎪ 1 ∞
⎪
⎨ ĥ+ (k1 , jω) exp(−jk1 x1 − jk3 x3 )dk1 , x3 > 0 ,
2π k1 =−∞ 3
Ĥ3 (x1 , x3 , jω) = ∞
⎪
⎪ 1
⎪
⎩ ĥ− (k1 , jω) exp(−jk1 x1 + jk3 x3 )dk1 , x3 < 0 ,
2π k1 =−∞ 3
(8.46)
and
⎧
⎪
⎪ 1 ∞
⎪
⎨ ê+ (k1 , jω) exp(−jk1 x1 − jk3 x3 )dk1 , x3 > 0 ,
2π k1 =−∞ 2
Ê2 (x1 , x3 , jω) = ∞
⎪
⎪ 1
⎪
⎩ ê− (k1 , jω) exp(−jk1 x1 + jk3 x3 )dk1 , x3 < 0 ,
2π k1 =−∞ 2
(8.47)
where ⎧
⎨
1 1
(ω 2 εµ − k12 ) 2 , |k1 | ≤ ω(εµ) 2 ,
k3 = (8.48)
⎩ −j(k 2 − ω 2 εµ) 21 , |k | > ω(εµ) 21 .
1 1
Note that the definition of the square root of k3 takes care of the bounded-
ness of the integrals as |x3 | → ∞. This guarantees the boundedness of the
electromagnetic field components as |x3 | → ∞. From Eqs. (8.42) and (8.43)
− − −
it follows that the quantities ĥ+ + +
1 , ĥ3 , ĥ1 and ĥ3 are related to ê2 and ê2 as
−k3 + k1 +
ĥ+
1 = ê , ĥ+
3 = ê , (8.49)
ωµ 2 ωµ 2
k3 − k1 −
ĥ−
1 = ê , ĥ−
3 = ê . (8.50)
ωµ 2 ωµ 2
1
When x3 > 0 and |k1 | < ω(εµ) 2 we observe that the electromagnetic
field consists of a superposition of uniform plane waves propagating in the
1
{k1 , 0, k3 }-direction, while when |k1 | > ω(εµ) 2 the constituents are (non-
uniform) evanescent waves decaying exponentially in the x3 -direction.
The presence of the source distribution is accounted for by the application
of the excitation conditions
!
Iˆ∆ (x1 , jω) , |x1 | < 12 a ,
lim Ĥ1 (x1 , x3 , jω) − lim Ĥ1 (x1 , x3 , jω) =
x3 ↓0 x3 ↑0 0, |x1 | > 12 a ,
(8.51)
which is a consequence of Eq. (8.44), and
lim Ê2 (x1 , x3 , jω) − lim Ê2 (x1 , x3 , jω) = 0 , −∞ < x1 < ∞ , (8.52)
x3 ↓0 x3 ↑0
216 excitation of two-dimensional electromagnetic waves
which is a consequence of Eq. (8.42). From Eqs. (8.52) and (8.47) it directly
follows that
ê− +
2 = ê2 , (8.53)
and from Eqs. (8.49) and (8.50) it follows that
ĥ−
1 = −ĥ1 .
+
(8.54)
We observe that the field components Ĥ3 and Ê2 are even-symmetric func-
tions of x3 , while Ĥ1 is an odd-symmetric function of x3 . Application of
Eqs. (8.51) and (8.54) to Eq. (8.45) leads to
∞ ! 1
1 ˆ
2 I∆ (x1 , jω) , |x1 | < 12 a ,
ĥ+
1 (k1 , jω) exp(−jk1 x1 )dk1 =
2π k1 =−∞ |x1 | > 12 a .
0,
(8.55)
The left-hand side of this equation is a Fourier integral with respect to the
spatial variable x1 and with the transform parameter k1 . Fourier’s inversion
theorem yields
1
a
2
1ˆ
ĥ+
1 (k1 , jω) = 2 I∆ (x1 , jω) exp(jk1 x1 ) dx1 . (8.56)
x1 =− 21 a
With Eqs. (8.56), (8.54), (8.49) - (8.50) and (8.45) - (8.47) the electromag-
netic field is determined completely. Once Iˆ∆ is prescribed, we can calculate
the integral of Eq. (8.56) either analytically or numerically, and once ĥ+
1 is
determined we can calculate the Fourier integral of Eq. (8.45) either ana-
lytically or numerically. The other field components can be computed in a
similar way.
When the point of observation is far enough away from the emitter, i.e.,
1
r = (x21 + x23 ) 2 (8.57)
the sheet emitter with a perpendicular electric current 217
1
x3 + x1 k 2
Ĥ1 ≈ ĥ1 (k , jω) exp(−jkr + j 14 π)
r r 2πr
1
when r = (x21 +x23 ) 2 → ∞ and x3 > 0 . (8.58)
and
1
x3 + x1 k 2
Ê2 ≈ ê2 (k , jω) exp(−jkr + j 14 π)
r r 2πr
1
1
µ 2 x1 k 2
≈ − ĥ+
1 (k , jω) exp(−jkr + j 14 π)
ε r 2πr
1
when r = (x21 +x23 ) 2 → ∞ and x3 > 0 . (8.60)
H
. ......
........
...
...
...
...
...............
........
S
'...................
...
.
...............
........
.......
E
.
......
..
.........
. x
.......
.......
.......
.......
....... ......
θ . .
...... ....
.
.
....
.......
..
...
................. .......
.. .. ............
..............
i1
.... .......
....
.......... .......
.......
......
.... ............
O r ..................................
.. i3
Figure 8.5. The orientation of the electromagnetic field vectors in the far field of
an emitter with perpendicular electric current.
|ĥ+
1 (k cos(θ), jω)|
2
D(θ) = .
(8.65)
1 2π
|ĥ+ 2
1 (k cos(θ), jω)| dθ
2π θ=0
In Fig. 8.6, the directive gain is presented for the case that Iˆ∆ is a constant
function, i.e.,
Iˆ∆ (x1 , jω) = î∆ (jω) , |x1 | < 12 a , (8.66)
so that Eq. (8.56) yields
sin( 12 k1 a)
ĥ+
1 = î∆ . (8.67)
k1
The directive gain is presented for various values of a/λ, where λ = 2π/k
is the wavelength. For values of a/λ → 0, we observe that ĥ+ 1 ≈ 2 a î∆ and
1
for large values of a/λ the emitter radiates the electromagnetic wavefield
mainly in forward and backward directions.
...
D(θ) ...
D(θ)
....... .......
....... .......
a/λ = 0.1 .. ... ..
.. a/λ = 0.3 .. ... ..
..
.... ....
.. ..
. . . . ..... . . . .
10
.
. . . . ..... . . . .
10 .
. . . ....
. .
. . . ....
. .
. . . . . .
. ... . . ..... . ... . . .....
. .. ... . ... . .. ... . ...
. ... . . ........... . ... . . ...........
. . 1 ..
. . . . . 1 .
. . .
.
. .
. ........
...... .
........... . .
....
...
....
...
...
...
..........
........
. .
.
θ .
. . .
. . ..............
. .. ...
... .
.
.. .. ...............
......... .
.
. ..
..........
. .
.
θ
. ...... . ........ . ........ ..
...........
..... . ... . ..... . ... .
. ..... . ... . .....
..
. ..... . ... . ..... ..
. . . ..
..... . .
. .
....
.
. . . . . . ..
..... . .
. .
..... . .
. . . .... . . .. . .... .
. . . ...... . 0.1 .
.
. . ... . . . . . ....... . 0.1 .
.
. . ..... . .
.
.
.
.. . . . . . .... . . . . ..... . .
.
.
... . . . . . .... . . . . ...... .
. .. . . .
. .
. . . ... . . .
.. . . .
. .
. . . .... .
.
. . . . .
. . .. .
. . . . .
. .
. .. . .. . .. . .. . .. . . ... . .. . .. . .. . ..... .
. .. . . . ... . . . .
. .. . .. . . . ... . . . .
.
....
... .. . .... ..
.
. ... . . . ... . . . . . . .... . . . ... . . . . .. .
. ...... . .. . ...... .
. ... . . .. . . . . .... . . .. . .... .
. . . . . . .
. . . .. . . . . . . .
.. . . . . . . .. ... . . . . . . . ..
. . . . . . ... . .
. ..
.. . . . . ..
..
.
. ..
...
. . . . . . . . ..... .
. .. .. . . . . .
. . .. . . . . . . .
. ..
... . . . . . .
. . .. . . .... . . . . . .
. . ... .
... . . .. . . . . .... . .... . . .. . . . . ..... .
. . . . . . ...
. . . . . . . ...
. . ....... . . .
... . . . . . ...... . . .
.... . . .
. . .... . . .. . . . . ..... . . .. .
. .... .... . . ..... .... .
.. . ..... . . .... . .. .. . ..... . . ..... . ..
. ..... .
....... . ......... . . ...... . . ......... .
. ....... ...
........ . .
.........
........... ...
.............. .
. . ............
................................
. . . . . . ...................................... . . .
. . . . . . . . .
. . . . . . . .
. . . .
. . . . . . . .
.. . .. .. . ..
. . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . .
..
D(θ) D(θ)
..
...... ......
....... .......
a/λ = 1 ... .... ..
.. a/λ = 3 ... .... ..
..
.... ....
.. ..
10 . 10 .
. . . . ..... . . . . .
.
. . . . ..... . . . . . .
. . . .... . .
. . . .... . .
. ... . .
. ..... . ... .
. . .....
. .. ... . .... . .. ... . ....
. .
. . . .......... . .
. . . ..........
. . 1 ..
. . . . . 1 .. . .
.
. .
.. .
. . . ... . . .
.... . ..
. . θ
. .
. .
.. .
. . . ... . . .
.
.... . ..
. .
.
θ
. . .. . .. . . . .. . . . .
. . . . . .
. . . . . ... . . .. . . . . . ... . . ..
. . .
.
. . . . . . . .
. . . . .
. . .
.................................... . 0.1 .
.
. . ..
................................ . . . . . .
. . 0.1 ..... . . . . .
. ......... ...... . . ... . . ............ . .. ....... . . .. .
................ . .... . ............... . .. .
. ...... . . . .... ... .
. .. . .. . . .. .
...... . . . . .... .....
..
.. .
. . .....
. .. ...... .
. .
. . ...
.
. .... . ............ .... . ..... .. .... . .... ............ . .... . .
. .. . .. .... .. ... . . . ... . . . ..... . ........ .... . .. .. .. . .. ........ . ..... . . .
. . . . .... .
. ..
.. ..
. .... .
. . .. . . . . .
. .
.
. ............... . ................................... ................ ... .. .. . ............. . ................................. . ............... .... .......
......
.
...
. . . . . ..... ... .... . . . . ... . ................... .......... ..... ... .......... ....................
. .... . . . . ........... . . . . ... . ........ . . ............................ ....... .......... . .
.....
...
.............. . ..... ........................................ ..
. .... . . .... . . .
.
.
. . . . . .
................... ... . .. ..
. ...
... . . . .
. .... ..... . . ..
.
. . ..... . . ............................................................................. . ...
...
. ... . . . . .... .... . . .
.
... . . ............ . .......................................... ........................ .... ..................... ............................................ . ..
...... .
. ... . . . .... . . ...
. . . ... .
........................ . .
..... ................. . ... .... . ............... .......... ...................... . . . ..
...
. .
... . . . ... ..
.. .
.... . . . . ... .. ......
.. . . ... . .... ....... . .... .
. ... .
.... . . . .... .... . . ..
. . . .... ............ ...... ....... .. .. .............. . .
. ...... ..... . . . ....... . . . . .... ............ . . . ................ .
..... . . ..... . . . . . ............. . .
..
. ......................................... . ................................. . . . . .. . . . . .
. . . . . . . . . . . . . . . . .
. . . .
.. . .
. . . . .. .. . .
. . . . ..
. . . . . . . . . . . .
. .. .
. . .. . . .. .
. . .. .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . .
. . . .
. . . . . . . .
.. . .. .. . ..
. . . . . . . . . .
. . . . . . . .
. . . . . . . . . . . .
Figure 8.6. The directive gain of the sheet emitter with perpendicular electric
current.
excitation of two-dimensional electromagnetic waves 221
Exercise 8.1
In the stationary phase method the argument of the exponential function is
expanded in a Taylor series around the stationary point k1 = k1s . Why does
the term linear in (k1 − k1s ) vanish?
Exercise 8.2
In the case of a constant impressed current Iˆ∆ (x1 , jω) = ı̂∆ (jω), |x1 | ≤ 12 a,
for what angles θ is the directive gain function D(θ) equal to zero?
Exercise 8.3
Assume that the impressed current flows uniformly in the x2 -direction and is
ext
present at x1 = 0, x3 = 0. Hence, Ĵ = ı̂∆ (jω)δ(x1 , x3 )i2 , where δ(x1 , x3 )
is the two-dimensional unit impulse operative at x1 = 0, x3 = 0. This implies
that the electric current sheet is degenerated to a line source.
(a) Give the expression for ĥ+1.
(b) Give in the far-field approximation the expressions for the electric and
magnetic field strengths in the frequency domain.
( c ) Draw the far-field radiation characteristic for the electromagnetic field
6 , n ∈ IN . Indicate
conform Fig. 8.5, for the angles of observation θ = nπ
the magnitude of the field strengths by the lengths of the vectors for
Ĥ and the radii of the circles for Ê.
(d) Calculate and draw the directive gain as a function of the angle of
observation θ.
Problem 8.1
Let for a current sheet the impressed perpendicular electric current be given
by Iˆ∆ (jω) = a−1 ı̂∆ (jω), for |x1 | ≤ 12 a. Show that in case we let a → 0
the same value for ĥ+1 is obtained as in Exercise 8.3(a). This is most easily
sin(k1 a/2)
obtained by taking lim ĥ+ +
1 with ĥ1 = k1 a ı̂(jω).
a→0
Problem 8.2
Assume the impressed current flows uniformly in the x1 -direction; it is
present at x3 = 0 for |x1 | ≤ 12 a and is given by Ĵ = −(1− 2|xa1 | )ı̂∆ (jω)δ(x3 )i1 .
ext
222 excitation of two-dimensional electromagnetic waves
(b) Give in the far-field approximation the expressions for the electric and
magnetic field strengths.
Exercise 1.1
(a) 0
(b) 0
( c ) (a · c)b − (b · c)a
(d) Replace one of the vector products by a new vector e Rewrite the
expression Substitute the original vector product for e Again
rewrite the expression to finally obtain (a · c)(b · d) − (a · d)(b · c)
Exercise 1.2
Since c · a = (a × b) · a = 0 and c · b = (a × b) · b = 0, the vector c is normal
to both vectors a and b.
Exercise 1.3
A = d1 d2 is the area of the rectangle.
Exercise 1.4
V = d1 d2 d3 is the volume of the brick.
224 answers to exercises
Exercise 1.5
(a) x2 (b) x2
x1 x1
x3 x3
(c) x1 (d) x1
x3 x3
x2 x2
The vector fields are invariant in the direction of the coordinate axis that
points forward.
Exercise 1.6
(a) ∇ · v = −1 , ∇×v =0
(b) ∇ · v = 0 , ∇ × v = −i3
( c ) ∇ · v = −2 , ∇×v =0
(d) ∇ · v = 0 , ∇ × v = −2i2 .
Exercise 1.7
∇2 A = (∂12 + ∂22 + ∂32 )A1 i1 + (∂12 + ∂22 + ∂32 )A2 i2 + (∂12 + ∂22 + ∂32 )A3 i3 .
Exercise 1.8
∇ × E = −∇ × ∇V = 0, for all V (x) (see Eq. (1.32)).
Exercise 1.9
∇ · D = ∇ · (∇ × C) = 0, for all C(x) (see Eq. (1.33)).
answers to exercises 225
Exercise 1.10
Application of the chain rule of differentiation gives
x − x
(a) ∇d = − , ∇d = −∇d
|x − x|
x − x
(b) ∇d−1 = , ∇ d−1 = −∇d−1
|x − x|3
2
( c ) ∇ · ∇d = , ∇ · ∇ d = ∇ · ∇d
|x − x|
Exercise 1.11
Use Eq. (1.11) to obtain w × (ν × H) = (w · H)ν − (w · ν)H Substitute
w = ν × E (ν × E) × (ν × H) = [(ν × E) · H] · ν, since (ν × E) · ν = 0
Take the scalar product of the result with ν ν · [(ν × E) × ν × H)]
= (ν × E) · H = ν · (E × H)
Exercise 1.12
(ν · H)ν − ν × (ν × H) = (ν · H)ν − (ν · H)ν + (ν · ν)H = H
The term (ν · H)ν is the part of H that is normal to S; the term
−ν × (ν × H) = (ν × H) × ν is the part of H that is tangential to S.
Exercise 2.1
There are five unknown vectorial quantities in Maxwell’s equations in matter:
E, H, J , D and B, and hence fifteen unknown scalar quantities.
Exercise 2.2
Application of Gauss’ law yields x∈∂ D ext ν · J dA = −∂t x∈Dext ρ dV .
This equation states that the amount of electric current that is flowing out
of the domain Dext through its boundary ∂Dext equals the decrease per unit
time of the electric charge present in Dext .
Exercise 2.3
Following the reasoning of Section 2.6 the continuity of those components
that are differentiated in the i3 -direction must be enforced. Thus we must
enforce the continuity of H1 , H2 , E1 and E2 .
226 answers to exercises
Exercise 2.4
Following the reasoning of Section 2.6 the continuity of those components
that are differentiated in the direction normal to the interface must be en-
forced. Thus ν · J + ∂t ν · D is continuous across S and ν · B is continuous
across S.
Exercise 2.5
(a) circular polarization
( c ) elliptical polarization
( e ) linear polarization.
Exercise 2.6
(a) J/m3 , J/m3
( c ) W/m2 , W/m2 .
An easy way to arrive at these results is to employ Poynting’s theorem and
to realize that in physics all terms of an equation have equal dimensions.
Exercise 2.7
∗ 1
ST = 12 Re[Ê × Ĥ ] = 1
2 Re[Ê1 Ĥ2∗ − Ê2 Ĥ1∗ ]i3 = ( µε00 ) 2 i3 .
1
E(t) = cos(ωt)i1 − sin(ωt)i2 , H(t) = ( µε00 ) 2 [sin(ωt)i1 + cos(ωt)i2 ],
1 1
S = E(t) × H(t) = ( µε00 ) 2 [cos2 (ωt) + sin2 (ωt)]i3 = ( µε00 ) 2 i3 .
Exercise 3.1
(a) The polarization of the wave is independent of the value of x3
Take x3 = 0 to simplify the problem Ê = i2 and Ê = −i1
the wave is circularly polarized (cf. Exercise 2.2 (a)).
After inversion the expression for the phase shows that wave propa-
gates in the positive x3 -direction.
Since Ê , Ê and the direction of propagation form a right-handed
triad, the polarization is right-handed (also called counter clockwise).
answers to exercises 227
Exercise 3.2
f = 1 MHz, µ = µ0 and γ = α + jβ = 0.04 + j0.1.
(a) The wave is attenuated by a factor of exp{−Re[γx3 ]} An attenua-
tion of exp(−π) occurs over a distance x3 = 25π m.
Exercise 3.3
The velocity of electromagnetic waves in vacuum is c0 3 × 108 m/s
f = cλ0 The range is 3.75 × 1014 Hz (red light) - 1015 Hz (violet light), or
375 THz (terahertz) - 1 PHz (petahertz).
Exercise 3.4
1 1
2
δ = ( ωµσ ) 2 δCu = 9.3 mm and δAl = δCu ( 5.8
3.5 ) = 12 mm. Only in the
2
first skin depth there is a significant penetration of the electric field that
runs along the surface of the wire Only the first skin depth carries a
significant part of the electric current Using a radius larger than the skin
depth (δAl = 12 mm) would be a waste of (expensive) material since the
inner part will hardly contribute to the power transport.
Exercise 3.5
) 2 = (πf µσ)− 2 The skin depth is proportional to the inverse
1 1
2
δ = ( ωµσ
of the square root of frequency The lower frequency (f = 100 kHz)
determines the minimal thickness of the shield dAl = 5δAl = 1.3 mm and
dCu = 5δCu = 1.0 mm.
228 answers to exercises
Exercise 3.6
I∆ (t) = Re[Iˆ∆ (jω) exp(jωt)] = sin(ωt) Iˆ∆ (jω) = −j A/m E1 (x3 , t) =
1 1
2 Z sin{ω[t − (εµ) x3 ]} and H2 (x3 , t) = 2 sin{ω[t − (εµ) x3 ]}.
1 2
1 2
Exercise 3.7
H(0, t) = 1.5 cos(3π × 108 t)i2 = Re[Ĥ(0, jω) exp(jωt)] Ĥ(0, jω) =
H0 i2 with H0 = 1.5 A/m and ω = 3π × 108 rad/s.
Assume that the field propagates in the positive x3 -direction
γ
Ĥ(x3 , jω) = H0 exp(−γx3 )i2 Ê(x3 , jω) = H0 σ+jω ε exp(−γx3 )i1
α+jβ β
E(x3 , t) = H0 exp(−αx3 ) Re{ σ+jω ε exp[jω(t − ω x3 )]}.
Substitution of the actual parameter values gives
1
(a) γ = jβ = jω(εµ) 2 E(x3 , t) = 188 cos[3π × 108 (t − 10−8 x3 )]i1 V/m
1
(b) ωε = 7.5 × 10−2 σ γ = α + jβ = 0 + jω(εµ) 2 The answer is
the same as in case (a)
1
( c ) ωε = 7.5 × 10−2 σ γ = α + jβ = ( ωµσ 2 ) (1 + j)
2 E(x3 , t) =
8 77 8 77
11.5 exp(−77x3 ){cos[3π×10 (t− 3π×108 x3 )]−sin[3π×10 (t− 3π×10 8 x3 )]}×
i1 V/m.
Exercise 3.8
Assume a unit amplitude of the incident electric wave, and apply the
electric field analysis
i i 1
Ê = exp(−γx3 )i1 Ĥ = ( µε00 ) 2 exp(−γx3 )i2 ,
r r 1
Ê = R exp[γ(x3 − 2L)]i1 Ĥ = −( µε00 ) 2 R exp[γ(x3 − 2L)]i2 .
Apply the boundary condition for Ê1 at x3 = L R = −1.
In free space and for steady state it follows that γ = jβ = j cω0
i r
Ê = Ê + Ê = {exp[−j cω0 x3 ] − exp[j cω0 (x3 − 2L)]}i1
= exp(−j cω0 L) {exp[−j cω0 (x3 − L)] − exp[j cω0 (x3 − L)]}i1
= −2j exp(−j cω0 L) sin[ cω0 (x3 − L)]i1 .
Apply inversion E = 2 sin[ω(t − cL0 )] sin[ cω0 (x3 − L)]i1 Since the
spatial behavior and the temporal behavior of E are decoupled, this is a
standing wave.
1
In the same way, the total magnetic field follows as H = 2( µε00 ) 2 ×
cos[ω(t − cL0 )] cos[ cω0 (x3 − L)]i2 the phase difference with the electric
field is π2 rad.
answers to exercises 229
Exercise 3.9
i
(a) Apply Eqs. (2.54) - (2.56) with ∂1 = ∂2 = 0 and Ĵ = 0 Ê =
1
( µε00 ) 2 Ĥ0 exp(−γ (1) x3 )i1
r r 1
(b) Ĥ = R Ĥ0 exp(γ (1) x3 )i2 Ê = −( µε00 ) 2 R Ĥ0 exp(γ (1) x3 )i1 ,
t t 1
Ĥ = T Ĥ0 exp(−γ (2) x3 )i2 Ê = 2( εµ00 ) 2 T Ĥ0 exp(−γ (2) x3 )i1
( c ) From the modified Helmholtz equation and the principle of causality
1 1
it follows that γ (1) = s(ε0 µ0 ) 2 and γ (2) = 2s(ε0 µ0 ) 2 , i.e., in both cases
the root of γ 2 has to be chosen in such a way that Re(γ) ≥ 0
Exercise 3.10
t
Employ the electric field analysis Ê = T⊥ exp(−γ (sea) x3 )i1 and
t
Ĥ = Y (sea) T⊥ exp(−γ (sea) x3 )i2 .
The medium properties of seawater are ε = 81ε0 , µ = µ0 and σ = 4 S/m
for f = 100 Hz it is found that in the sea ωσε = 8.9 × 106 1
1
γ (sea) = α + jβ = ( σµ20 ω ) 2 (1 + j) = 4 × 10−2 (1 + j) m−1 Y (sea) = γ (sea)
σ
=
50(1 − j) S.
Y (air) = Y0 = ( µε00 ) 2 = 2.6 mS.
1
−5 (1 + j).
(sea) = 5.3 × 10
2Y0
T⊥ = Y +Y
0
For the transmitted field it follows that
t ∗t
ST = 2 Re[Ê × Ĥ ]
1
∗ ∗(sea)
= 1
2 Re[T⊥ T⊥ Y exp(−γ (sea) x3 ) exp(−γ ∗(sea) x3 )]i3
∗(sea)
2 |T⊥ | Re[Y
2
= 1
] exp{−2 Re[γ (sea) x3 ]}i3
−7
= 1.4 × 10 exp(−8 × 10−2 x3 )i3 W/m2
x3 = 0 m gives ST = 1.4 × 10−7 i3 W/m2 ; x3 = 100 m gives ST =
4.7× 10−11 i 3 W/m2 .
t
Ĵ = σ Ê = 4T⊥ exp(−γ (sea) x3 )i1 x3 = 0 m gives Ĵ = 2.1 × 10−4 ×
(1 + j)i1 A/m2 .
230 answers to exercises
Exercise 3.11
It could be included as a loss term in the inductance L The transmission
line equations, Eqs. (3.112) and (3.113), would then be
These are also known as the telegraph equations. The circuit parameter
R is of course the series resistance, while the parameter G is the shunt
conductance.
Exercise 4.1
(a) ω = 3π × 109 rad/s f = 1.5 GHz.
tan(θ) = ss13 = 43 θ = 0.927 rad (53.1◦ ).
Exercise 4.2
Parallel polarization s(1) · ê(1) = 0 and s(2) · ê(2) = 0 ê(1) = Ê (1) ×
[cos(θ)i1 − sin(θ)i3 ] and ê(2) = Ê (2) [cos(θ)i1 + sin(θ)i3 ].
|ê(1) | = |ê(2) | Ê (1) = Ê (2) = Ê0 (assuming their phases to be equal).
Subsitute in Eq. (4.58) and apply Euler’s formula
ω
Ê = 2Ê0 exp[−j sin(θ)x1 ]
c0
ω ω
× cos(θ) cos[ cos(θ)x3 ]i1 + j sin(θ) sin[ cos(θ)x3 ]i3
c0 c0
Propagating wave in the x1 -direction, standing wave in the x3 -direction.
answers to exercises 231
Exercise 4.3
i
Assume for x3 < 0 the presence of the incident field Ê = E0 ×
r
exp[−(γ1i x1+γ3i x3 )]i2 and the reflected field Ê =E0 R⊥ exp[−(γ1r x1+γ3r x3 )]i2 .
The boundary condition Ê2 = 0 for x3 ↑ 0 must hold for all x1 γ1r = γ1i ,
γ3r = −γ3i and R⊥ = −1.
In case of a lossless medium and steady state it is allowed to write γ1i = jβ1i
and γ3i = jβ3i .
For x3 < 0 the total field is
Ê = E0 [exp(−jβ1i x1 − jβ3i x3 ) − exp(−jβ1i x1 + jβ3i x3 )]i2
= E0 exp(−jβ1i x1 ) [exp(−jβ3i x3 ) − exp(jβ3i x3 )]i2
= −E0 exp(−jβ1i x1 ) 2j sin(β3i x3 )i2 .
Exercise 4.4
(a) Parallel polarization since the electric field is entirely in the plane of
propagation.
√ √
(b) si = 12 3i1 + 12 i3 sin(θi ) = si1 =√12 3 θi = π3 .
√ to Snell’s law, sin(θ ) = 1.5 sin(θ ) θ = 4 s =
i t t π t
According
√
1 1
2 2i1 + 2 2i3 .
(d) ...
...
...
er et
......... .
...... ............ ..... .....
....
.
.....
.......
... ........... ............
............
.. ...
π ........... .............. π
.
3 ...... ................ ..... 4
. ..
π ...............
3 ......
ei ..........
.......... ...........
...........
...
....
.
...
x3 = 0
Exercise 4.5
(2) (2) 1
(a) θci = arcsin( nn(1) ) = arcsin[( εε(1) ) 2 ] = arcsin( 12 ) = π
6 rad.
√
(b) γ1i = γ i sin(θi ) = γ0 3 and γ3i = γ i cos(θi ) = γ0 .
1 1
[−ω 2 ε(2) µ(2) − (γ i )2 ] 2 with Re[γ3t ] ≥ 0
( c ) γ3t = √ γ3t = (γ02 − 3γ02 ) 2 =
−jγ0 2.
Exercise 4.6
All the light that comes in from the air enters the water at angles smaller
than or equal to the critical angle; this is the reciprocal use of total reflection
at angles of incidence larger than the critical angle for plane waves from water
to air. The critical angle is θc = 0.85 rad (48.7◦ ).
Exercise 4.7
nglass
nair
= 32 The critical angle exists for a wave incident on the glass/air
interface and does not exist for a wave incident on the air/glass interface;
this is independent of the type of polarization.
glass
εglass
εair
= 2.25 and µµair = 1 The Brewster angle does not exist in case
of perpendicular polarization, whereas it exists for both the air/glass and
the glass/air interfaces in case of parallel polarization.
answers to exercises 233
Glass/air ⊥ arcsin( 23 ) —
arcsin( 23 ) arctan( 23 )
Air/glass ⊥ — —
— arctan( 32 )
Exercise 4.8
µ(2)
Take Eq. (4.181) and substitute µ(1)
=1
⎛ ⎞1 1 1
ε(2)
1−
2
ε(2) (ε(1) − ε(2) ) 2
ε(2) 2
i
tan(θB ) =⎝ ε(1) ⎠ = =
ε(1)
−1 ε(1) (ε(1) − ε(2) ) ε(1)
ε(2)
Exercise 5.1
2.50 45◦ Ray enters the halfspace x3 > 3 with θexit = 62◦
Ray turns at x3;hor = 2.84 and enters the half-
60◦
space x3 < 1 with θexit = 147◦
Ray starts at x3;hor = 2.50 and enters the half-
90◦
space x3 < 1 with θexit = 141◦
120◦ Ray enters the halfspace x3 < 1 with θexit = 147◦
Ray turns at x3;hor = 2.88 and enters the half-
2.00 45◦
space x3 < 1 with θexit = 148◦
Ray turns at x3;hor = 2.40 and enters the half-
60◦
space x3 < 1 with θexit = 139◦
Ray starts at x3;hor = 2.00 and enters the half-
90◦
space x3 < 1 with θexit = 131◦
120◦ Ray enters the halfspace x3 < 1 with θexit = 139◦
234 answers to exercises
(Continued)
Exercise 5.2
The extreme positions where a ray may have a horizontal tangent are
x3;hor = 1 and x3;hor = 3 For a horizontal tangent in between these
extreme positions it is found that n(x3;hor ) > 1 For√an undulating ray,
C0 = n(x3;hor ) sin( π2 ) > 1 C0 = n(x3;0 ) sin(θ0 ) = 2 sin(θ0 ) > 1
π 3π
4 < θ0 < 4 .
Exercise 5.3
(1) (2)
Yes: For any ray with 1 < x3;0 < 3 there exist a x3;hor and a x3;hor with
(1) (2) (1) (2)
1 < x3;hor < x3;0 < x3;hor < 3 and n(x3;0 ) > n(x3;hor ) = n(x3;hor ) > 1
For any 1 < x3;0 < 3 a range of θ0 may be found for which n(x3;0 ) sin(θ0 ) =
(1,2)
n(x3;hor ) sin( π2 ) > 1 arcsin( n(x13;0 ) ) < θ0 < π − arcsin( n(x13;0 ) ).
Exercise 5.4
For any 1 < x3;0 < 2 there exists a x3;hor in the range x3;0 < x3;hor < 2
such that n(x3;0 ) sin(θ0 ) = n(x3;hor ) sin( π2 ) may be solved for θ0 Any
starting position 1 < x3;0 < 2 has a ray with a horizontal tangent.
For any 2 < x3;0 < 3 there exists a x3;hor in the range 2 < x3;hor < x3;0
such that n(x3;0 ) sin(θ0 ) = n(x3;hor ) sin( π2 ) may be solved for θ0 Any
starting position 2 < x3;0 < 3 has a ray with a horizontal tangent.
Since ∂3 n
= 0 there is no ray with a horizontal tangent for x3;0 = 2.
Exercise 5.5
For rays with a horizontal asymptote it must be possible to solve
n(x3;0 ) sin(θ0 ) = n(2) sin( π2 ) for θ0 For all 1 < x3;0 < 2 a solution
θ0 = arcsin[ n(x13;0 ) ] may be found, and for all 2 < x3;0 < 3 a solution
answers to exercises 235
Exercise 6.1
(a) b = 11.4 mm.
( c ) Use Eq. (6.55) for φ = 100 |E| = (e21 + e22 )1/2 = r ln(b/a)
100
|E| has a maximum when r = a |E|max = a ln(b/a) = 2.38 × 104
100
V/m.
Exercise 6.2
)∞ )∞ −8
(a) v + (t) = Z0Z+Z
0
S n=0 (ΓS ΓL )
n V (t−2nL )
S c = 5
8
3 n
n=0 ( 28 ) VS (t−4×10 n).
(b) V ( 12 L, t) = v + (t − 2c L
) + ΓL v + (t − 2L−L/2
c ) = 58 [VS (t − 10−8 )−
−8 −8 −8
7 VS (t − 3 × 10 ) + 28 VS (t − 5 × 10 ) − 196 VS (t − 7 × 10 ) + · · ·] and
3 3 9
I( 12 L, t) = Z10 [v + (t − 2c L
) − ΓL v + (t − 2L−L/2
c
1
) = 80 [VS (t − 10−8 )+
−8 −8 −8
7 VS (t − 3 × 10 ) + 28 VS (t − 5 × 10 ) + 196 VS (t − 7 × 10 ) + · · ·].
3 3 9
∞ 1 10−9
( c ) WS = 0 VS (t)I(0, t) dt = Z0 +ZS 0 VS2 (t) dt = 12.5 pJ.
10−9 Z0 10−9
(d) W + = 0 V (0, t)I(0, t) dt = (Z0 +ZS )2 0 VS2 (t) dt = 7.8 pJ.
(e) WL = 0∞ V (L, t)I(L, t) dt = 0∞ (1 + ΓL )v + (t − L 1
c ) Z0 (1 − ΓL )×
v + (t − Lc ) dt = Z10 (1 − Γ2L ) 0∞ [v + (t − Lc )]2 dt.
)∞ (2n+1)L 2
[v + (t − Lc )]2 = [ Z0Z+Z
n=0 (ΓS ΓL ) VS (t −
0 n
c )] = ( Z0Z+Z
0
)2 ×
)∞ (2n+1)L
S S
2n 2
n=0 (ΓS ΓL ) VS (t− c ) since VS (t− (2n+1)L
c )VS (t− (2m+1)L
c )=0
for all n
= m.
∞ + L 2 Z0
2 )∞ (Γ Γ )2n ∞ V 2 (t− (2n+1)L ) dt =
0 [v (t− c )] dt = ( Z0 +ZS ) n=0 S L 0 S c
1−Γ2L
( Z0Z+Z
0
S
)2 1
1−(ΓS ΓL )2
× 10−9 WL = Z0
(Z0 +ZS )2 1−(ΓS ΓL )2
× 10−9 =
6.5 pJ .
Exercise 6.3
The 100 V signal travels in 2 µs from the beginning to the end T = 2 µs.
236 answers to exercises
At the end of the line the sum of the incoming 100 V signal and the once
reflected signal gives a 75 V signal over the load A once reflected signal
L −Z0
of −25 V signal originates at the end ΓL = Z ZL +Z0 = − 4
1
ZL = 60 Ω.
At the beginning of the line the reflected −25 V signal and the twice
reflected signal gives a −10 V signal over the source A twice reflected
S −Z0
signal of 15 V originates at the source ΓS = ZZS +Z0 = − 5 ZS = 25 Ω.
3
ω0 (2m+1)π .
m ∈ IN , provided that ∆ω 2
Exercise 6.5
−1
For a lossless transmission line, |Γ| = |ΓL | |ΓL | = VSWR
VSWR+1 |ΓL | = 2
1
for set A and |ΓL | = 5 for set B Set B is better matched to the trans-
1
mission line.
Exercise 7.1
1
(a) k = ω(εµ) 2 = 41.9 m−1 and πa = 31.4 m−1 k ≥ mπ a (propagating
modes) for m = 1, k < mπ
a (evanescent modes) for m = 2, 3, 4, · · · .
(b) The field strengths of the propagating TE1 -mode follows from Eqs.
(7.21) - (7.23) as
Exercise 7.2
∗
P = S · ν dA = 12 w 0a Re[(Ê × Ĥ ) · i3 ]mode dx1 For a given |E0 | the
1
power is largest for the TM0 - or TEM-mode and equals P = 12 wa( µε ) 2 |E0 |2 .
For safety take |E0 | = 2 × 105 V/m P = 26.5 kW.
Exercise 7.3
k = 2πf 2π
c = λ λg,m = k2π =√ 1
λg,0 = 1 m and λg,1 = 1.34 m
g,m 1−4m2 /9
while for m = 2, 3, 4, · · · the modes are evanescent and we can no longer
speak of a guided wavelength.
Exercise 7.4
The EHF band ranges from 30 GHz to 300 GHz The cutoff frequency
must be chosen in such a way that fc,0 ≤ 30 × 109 Hz < f < 300 × 109 Hz ≤
fc,1 .
c0
Equation (7.113) with n3 = n1 and m = 0 gives fc,0 = 0 Hz and fc,1 = 6a
fc,0 ≤ 30×109 Hz is satisfied for any value of a while fc,1 ≤ 300×109 Hz
requires that a < 0.167 mm.
Exercise 7.5
In each of the regions the general solution is
(i) (i)
Ê = Â(i) exp(jk1 x1 ) + B̂ (i) exp(−jk1 x1 ).
Exercise 8.1
In the Taylor series expansion around the stationary point k1s the linear term
is proportional to the first derivative at k1s , which by definition is equal to
zero.
Exercise 8.2
Parallel electric current:
D(θ) = 0 if sin(θ) = 0 θ = 0 and θ = π.
sin[ka cos(θ)/2]
D(θ) = 0 if ĥ+ 2 [k cos(θ), jω] = k cos(θ) ı̂(jω) = 0 ka cos(θ)
2 = nπ
with n ∈ ZZ \{0} θ = arccos( nλ
a ) with |nλ| ≤ |a|.
Perpendicular electric current:
sin[ka cos(θ)/2]
D(θ) = 0 if ĥ+
1 [k cos(θ), jω] = k cos(θ) ı̂(jω) = 0 ka cos(θ)
2 = nπ
with n ∈ ZZ \{0} θ= arccos( nλ
a ) with |nλ| ≤ |a|.
Exercise 8.3
(a) Use Eq. (8.56) with a delta function in the integrand ĥ+1 (k1 , jω) =
1
2 ı̂(jω).
(c) θ=0
⊗
.........................
..
...........
.........
.
.
..
........ . ⊗ ..
.
⊗ ....
..
..
....... ...
.
.
... .... H .....
... .. . ........
.. ... ...
.. ..
⊗ ⊗
... .
.. .. ...
. . . . ......
....
.
........
. ....
.......
...
..
.
.
.
. ....
.
..... ..
.
E
... .......
.......
...
.. .
i1 ..
..
.......
.......
...
....... ... ........... .. .... .........
..
....... .. .... .. ....... ....
O p
. . .
θ = 32 π ⊗ ⊗ θ = 12 π
. . .
...................................... .
...
...
. ....... ....... ....... ....... .......
...... . .
.................. ....
.......
i3 ....... ....... ....... ...
.
...
.........
.... ....... ... .
.......
....... . ... .......
...... ... .. .......
...
..... . .. ... ... ....... ...........
....... ... . .. ....... .....
E⊗ ... . ... .
.
..
...
.. ⊗ .
... ...
........... .
... ... .... ..
.
. ...
... .......
H ⊗ ... .. ..........
.......
....... .
..........
.
..
⊗
.......
⊗ .........................
θ=π
240 answers to exercises
∗ * +1
(d) The time-averaged power flow is 12 Re[Ê × Ĥ ] = |ı̂∆ (jω)|2 µε 2 16πr
k
×
[cos(θ)i1 + sin(θ)i3 ] D(θ) = 1 The directive gain is the unit
circle.
Bibliography
TEm , 175, 180, 186, 193 Plane wave, 49, 83, 84, 90, 95, 101,
TEM, 178 121, 126, 128, 130, 131,
TMm , 177, 180, 191, 193 167, 180, 204, 214
Mode index, 193 evanescent, 206, 215
non-uniform, 86, 109, 114, 195,
Normal incidence, 105, 108 203, 206, 215
parallelly polarized, 87, 93
One-dimensional wave, 49, 56, 58, perpendicularly polarized, 89,
60, 61, 66, 87, 105, 108, 93
167, 203 uniform, 86–89, 91, 99, 108,
Operator 109, 122, 126, 130, 203,
curl, 7, 13 206, 211, 215, 219
del, 6 Plane-wave decomposition, 180
div, 7, 10 Plasma, 201
grad, 6, 8 Plasma frequency, 201
Laplacian, 18 Point charge, 21
nabla, 6 Polarization
of Hamilton, 6 circular, 39
Optical communication system, 169 elliptical, 39
Optical fiber, 169 left-handed, 41
Orientation linear, 39
of electric field, 54 parallel, 82, 87, 93, 97, 100,
of magnetic field, 54 103, 111, 113, 114, 122,
Origin, 2 124, 172, 182, 204
perpendicular, 83, 89, 93, 98,
Parallel wires, 145 100, 106, 111, 113, 114,
Penetration depht, 80 122, 128, 172, 182, 213
Permeability, 22, 24, 29, 35, 81, right-handed, 41
121 Polarization state, 39
relative, 30 Position, 2
Permittivity, 24, 29, 34, 81, 121, Power, 114, 159
201 dissipated into heat, 42
relative, 30 generated by sources, 43
Phase, 55 instantaneous, 41
Phase coefficient, 57, 87, 178 volume density of, 43, 44
Phase factor, 58, 86, 87 Power flow, 32, 89, 90, 92–94, 159
Phase vector, 84 Power flow density, 175, 187, 192,
Phase velocity, 193 210, 218
Physical laws, 4 Power relation, 41
Physical quantity, 2, 4 Poynting theorem, 41
Plane complex, 44
of equal amplitude, 58, 86, 114, instantaneous, 41
195 Poynting vector, 31, 43, 58, 61, 86,
of equal phase, 58, 86, 114, 195 88, 90, 92
of observation, 82, 83 complex, 44
index 247
Transmission coefficient, 65, 69, 104, Voltage standing wave ratio, 165
105, 107, 108, 111 VSWR, 165
Transmission line, 49, 145, 167
lossless, 158, 160 Wave
multiconductor, 145 evanescent, 175, 177
two-conductor, 156 non-propagating, 175, 177
Transmission line equations, 73, 79, not plane, 121
150 plane, see Plane wave
Transmission line equivalent, 73 propagating, 175, 177
Transmitted field, 63, 68 standing, 100
Transmitted wave, 102, 109, 110, travelling, 100
114, 121 Wave admittance, 52, 55, 62, 63,
Transverse electric wave, 172 68, 89, 108
Transverse electromagnetic wave, Wave impedance, 53, 55, 64–66,
72, 145, 167 69, 88, 106
Transverse magnetic wave, 172 Wave speed, 56, 60, 61, 161, 163
Transverse plane, 168 Wavefront, 121, 123, 125, 127, 129,
Triple product 130
scalar, 5 Waveguide, 167
vectorial, 5 asymmetric dielectric slab, 194
Two-dimensional medium, 121 closed, 168
Two-dimensional wave, 81, 121, 167, cylindrical, 167, 168
203 dielectric slab, 171, 182, 193
graded-index slab, 182
Uniform line, 145 open, 168, 169
Unit, 4 parallel-plate, 72, 145, 151, 156,
Unit impulse 158, 171, 172, 178
one-dimensional, 50, 204, 213 planar, 171
two-dimensional, 221 planar graded-index, 171
planar step-index, 171
Vacuum, 30 step-index slab, 182
Vector, 3 symmetrical slab, 182
addition of, 4 thin-film, 182
components of, 4 uniformly cylindrical, 169
length of, 5, 37 Wavelength, 56, 87, 94
multiplication of, 4 Wavenumber, 174, 177, 193
properties of, 4
subtraction of, 4
Vector calculus, 4
Vector function, 6
derivative of, 6
Vectorial position, 3
Velocity
of observer, 55
of point charge, 22
Textbooks on science and technology from VSSD
Macroeconomics
James K. Galbraith and William Darity, Jr.
2005 / xxii+505 p. / ISBN 90-71301-57-5
http://www.vssd.nl/hlf/b006.htm
Introduction to bed, bank and shore protection Series on Mathematical Geodesy and Positioning
Engineering the interface of soil and water http://www.vssd.nl/hlf/a030.htm
Gerrit J. Schiereck Adjustment Theory
2004 / xii+ 400 p. / ISBN 90-407-1683-8 P.J.G. Teunissen
http://www.vssd.nl/hlf/f007.htm 2003 / vi + 201 p. / ISBN 90-407-1974-8
Principles of river engineering Hydrography
The non-tidal alluvial river C.D. de Jong, G. Lachapelle, S. Skone, I.A.Elema,
P.Ph. Jansen (ed.) 2003 / x+351 pp. / ISBN 90-407-2359-1
1994 / XVI+509 p. / A4 / ISBN 90-407-1280-8
http://www.vssd.nl/hlf/f006.htm Dynamic Data Processing
P.J.G. Teunissen
The vertical motion of foundations and pontoons 2001 / x + 241 p. / ISBN 90-407-1976-4
Godfried Kruijtzer
2002 / 49 p. / ISBN 90-71301-64-8 Testing Theory
http://www.vssd.nl/hlf/f016.htm P.J.G. Teunissen
2000 / viii + 147 p. / ISBN 90-407-1975-6
Chemical engineering
Mathematics
Mass Transfer in Multicomponent Mixtures
J.A. Wesselingh and R. Krishna Mathematical Systems Theory
2006 / 329 p. / ISBN 90-71301-58-3 G.J. Olsder and J.W. van der Woude
http://www.vssd.nl/hlf/d004.htm 2008 / x+208 p. / ISBN 90-71301-40-0
http://www.vssd.nl/hlf/a003.htm
Transport phenomena data companion
L.P.B.M. Janssen and Numerical Methods in Scientific Computing
M.M.C.G. Warmoeskerken J. van Kan, A. Segal and F. Vermolen
2006 / 168 p. / ISBN 90-71301-59-1 2005 / xii+277 pp. / ISBN 90-71301-50-8
http://www.vssd.nl/hlf/c017.htm http://www.vssd.nl/hlf/a002.htm
Order in Space
JA.K. van der Vegt
2006 / 93 pp. / ISBN 90-71301-61-3
http://www.vssd.nl/hlf/a017.htm
Electromagnetic Waves – An Introductory Course
Electromagnetic waves appear in many forms and
their applications are extremely widespread. Without
exaggeration it may be said that our ability to employ
and manipulate electromagnetic waves forms one of
the reasons that communication plays such an
important role in society.
The macroscopic theory of electromagnetic waves
has been formulated by Maxwell in 1864. But the
mathematical-physical nature of the subject makes it
difficult for students to master even today. The
continuous stream of new college textbooks shows
that many teachers encounter this problem and
attempt to resolve it by presenting the theory in some
suitable form.
In the Electrical Engineering curriculum of the Delft
University of Technology, the teaching of
electromagnetic waves has been divided into three
stages: 1) a basic course on Electricity and Magnetism,
2) an introductory course on Electromagnetic Waves,
and 3) advanced courses on the application and
computation of electromagnetic waves. The current
book is written to facilitate the introductory course
on Electromagnetic Waves. It is assumed that students
are already acquainted with the basic phenomena
and notions of the electric and the magnetic field, and
that they know in which way Maxwell's equations
describe the electromagnetic field. Starting from
Maxwell's equations, this book deals with the
derivation of plane wave propagation, plane wave
reflection and transmission, electromagnetic rays,
waves in two-wire transmission lines, waves in planar
waveguides, and the excitation of electromagnetic
waves. As such, the aim of the book is to provide a
solid understanding of how the basic ingredients that
make up the more sophisticated applications follow
from Maxwell's equations.
The aim of the introductory course on Electromagnetic
Waves is to teach students to manipulate the
fundamental formulas in order to solve a problem at
hand. To focus on this skill and to overcome the
problem of having to learn many formulas by heart,
an outline of this book is available from the website
as a printable Repetitive Guide.
Published by VSSD.
E016