3.

Plane Wave Solutions

3.1 A General Solution

of the Homogeneous Wave Equation

Let us consider the homogeneous wave equation l

(3.1)

and the direct and inverse Fourier transforms in space

Fx u = u(k) = (21f~d/2 ~d u(x)e-ik,xdx, (3.2a)

'1:"-1
J x u
A

=u x =
( ) 1
(21f)d/2
r
J][I{d u
A (k) ik'Xdk
e , (3.2b)

where i2 = -1, x E JRd and k E JRd. By applying the Fourier transform in

space to (3.1), we obtain the following ODE:

d 2A
~2 + c2 1kl 2 u = 0 (3.3)
dt '

whose solution is of the form

u(k, t) = A(k)eiclklt + B(k)e-iclklt. (3.4)

On the other hand, by applying to (3.3) the Fourier transform in time F t :

L2(JR) --+ L2(JR) where

Ftu = u(w) =
V
1
r.>=
21f
1 .
][I{
u(t)e-1wtdt, (3.5)

1 With a right-hand side, the correct equation would actually be

but the formulation given below is equivalent and easier to manipulate when
f = O.

G. C. Cohen, Higher-Order Numerical Methods for Transient Wave Equations

© Springer-Verlag Berlin Heidelberg 2002
26 3. Plane Wave Solutions

we obtain the dispersion relation of (3.1):

(3.6)

Now, by assuming w > 0, the solution can be written as

u(k, t) = A(k)e iwt + B(k)e- iwt . (3.7)

The inverse Fourier transform in space applied to u provides the following

form of the solution of (3.3)

u(x) = 1 r
(271" )d/2 JJRd
[A(k)ei(wt+k.a:l + B(k)ei(-wt+k.a:l] dk. (3.8)

Equation (3.8) shows that the solution of the homogeneous wave equation
can be expressed as a continuous superposition of the plane waves

(3.9)

whose amplitudes are A(k) and B(k). Hence the study of properties of the
solutions of the wave equation can be carried out by considering the plane
wave solution defined in (3.9).

Remarks
1. w is the pulsation and k the wave vector which indicates the direction
of propagation of the plane wave. Obviously, w/lkl is the velocity of the
propagated wave.
2. This result given in the case of the scalar wave equation, can be extended
to the other equations.

3.2 Application to the Maxwell Equations

3.2.1 The 3D Case

Now, let us look for a plane wave solution of the homogeneous anisotropic
Maxwell equations

aD
--VxH=O (3.10a)
at '
oB
7it+V x E=O, (3.10b)
D = EoE, (3.10c)
B=f..Lo H , (3.lOd)

of the form
 3.2 Application to the Maxwell Equations 27

E = Eoei(wHk."') , (3.11a)
H = H oei(wt+k."'), (3.11b)
D = Doei(wt+k."') , (3.11c)
B = Boei(wHk.",). (3.11d)

By inserting (3.11a)-(3.11d) into (3.lOa)-(3.lOd), we obtain the following

relations

wDo - k x H 0 = 0, (3.12a)
wBo +k x Eo = 0, (3.12b)
Do = coEo, (3.12c)
Bo = /-LoHo· (3.12d)

°
If w i- 0, (3.12a)-(3.12b) show that D is orthogonal to Hand k and B is
orthogonal to E and k. The case w = will be discussed below.

On the other hand, since co and /-Lo are symmetric positive definite, we
can write co = €O~o and /-Lo = iioiio. By multiplying (3.12c) and (3.12d) by
H 0 and E--; respectively, we obtain, thanks to these decompositions

~oEo· ~oHo = 0, (3.13a)

iioEo . iioH 0 = 0. (3.13b)

A similar process provides

~oEo . ~ok = 0, (3.14a)

- -
iioH 0 . iiok = 0, (3.14b)

which shows that the vector ~oE is orthogonal to ~oH and ~ok and the vector
iioH is orthogonal to iiok and iioE. -

Now, by eliminating Eo, Do and Bo in (3.12a)-(3.12d), we obtain the

following dispersion relation:

(3.15)

In other words, there are three velocities which are the square roots of the
eigenvalues of matrix Mo defined by

(3.16)

°°
Obviously, is an eigenvalue of Mo. The corresponding eigenvector is colinear
to k. So w = provides a stationary mode for which the field H is parallel to
28 3. Plane Wave Solutions

k. This stationary solution can be written as the gradient of a scalar potential

'ljJ. The other two modes lead to two dispersion relations which correspond
to two waves with different polarizations and different velocities.

In the isotropic case, one can easily check that the electric and magnetic
fields are orthogonal and are both orthogonal to k. The stationary mode still
exists but Mo has a double eigenvalue. Actually, when w -=I 0, the problem
can be written as
1
W Ho = -Ikl Ho·
2 2
(3.17)
coJ-Lo
Equation (3.17) leads to the following dispersion relation

cOJ-LOW2 = Ik1 2 , (3.18)

which shows that the velocity c of the waves is equal to 1/ VcoJ-Lo.

Remark

A similar development can be carried out for the electric field. In the isotropic
case the same dispersion relation is obtained.

3.2.2 The 2D Case

In the 2D case, we shall present the TM polarization defined in (1.16a) and

(1.16b) but an equivalent study can be carried out for the TE polarization.

So, let us consider the 2D Maxwell system in a homogeneous anisotropic

medium:
8E
co -8 - curlH = 0, (3.19a)
= t
8H
J-L0 m + curlE = 0, (3.19b)

for which we are looking for a plane wave solution derived from (3.11a)-
(3.11d).

Let us set:
co -1 -_
=
(a a
fJ
(3) .
'Y
(3.20)

After inserting this solution into (3.19a) and (3.19b) and eliminating Eo, we
obtain the following dispersion relation

(3.21 )

where k1 and k2 are the components of the vector k.

 3.3 Application to the Elastics System 29

Now, if we set kl = Ikl 2cosO and k2 = Ikl 2sinO, we obtain2

c= I~I = J:0 2 2
(asin 0+,cos 0-2jJsinOcosO), (3.22)

which provides an equation in polar coordinates of the velocity c. This equa-

tion indicates the anisotropy of c. For instance, when JLo = 1, a = 32/31,
jJ = -4/31 and, = 16/31 (which provides a matrix whose inverse is the
positive definite matrix A defined in Sect. 13.2.8), we obtain the curve drawn
in Fig. 3.1.

0.8

-0.8

Fig. 3.1. An example of an anisotropy curve of the velocity for the 2D Maxwell
system

3.3 Application to the Elastics System

We are looking for a plane wave solution of (1.36) when Lame's coefficients
are constant. The plane wave solution is

v = voei(wt+k.",). (3.23)
By inserting (1.6d) into (1.36), we obtain the dispersion relation

(JLlkl2 - s) Va + (A + JL)(k· vo)k = 0, (3.24)

where s = pw2.

2 The positivity of the expression under the square root comes from the positive
character of the matrix eo -1.
30 3. Plane Wave Solutions

Equation (3.24) has two classes of solutions:

These two classes provide two kinds of waves with two different dispersion
relations:

1. Waves parallel to the direction of propagation whose dispersion relation

is
(3.25)

2. Waves perpendicular to the direction of propagation whose dispersion

relation is
w~ = ~lkI2. (3.26)
P
So, by setting Vp = yip.. + 2/1)/ p and Vs = v;rp, we obtain the P-wave
and S-wave defined in (1.40) and (1.41).