EECS 117

Lecture 23: Oblique Incidence and Reection

Prof. Niknejad
University of California, Berkeley
University of California, Berkeley EECS 117 Lecture 23 p. 1/21
Review of TEM Waves
We found that E(z) = xE
i0
e
jz
is a solution to
Maxwells eq. But clearly this wave should propagate in
any direction and the physics should not change. We
need a more general formulation.
Consdier the following plane wave
E(x, y, z) = E
0
e
j
x
xj
y
yj
z
z
This function also satises Maxwells wave eq. In the
time-harmonic case, this is the Helmholtz eq.

2
E + k
2
E = 0
where k =

=

c
University of California, Berkeley EECS 117 Lecture 23 p. 2/21
Conditions Imposed by Helmholtz
Each component of the vector must satisfy the scalar
Helmholtz eq.

2
E
x
+ k
2
E
x
= 0
_

2
x
2
+

2
y
2
+

2
z
2
_
E
x
+ k
2
E
x
= 0
Carrying out the simple derivatives

2
x

2
y

2
z
+ k
2
= 0

2
x
+
2
y
+
2
z
= k
2
Dene k = x
x
+ y
y
+z
z
as the propagation vector
University of California, Berkeley EECS 117 Lecture 23 p. 3/21
Propagation Vector
The propagation vector can be written as a scalar times
a unit vector
k = ka
n
The magnitude k is given by k =

As well show, the vector direction a

n
denes the
direction of propagation for the plane wave
Using the dened relations, we now have
E(r) = E
0
e
jkr

x
= k x = ka
n
x

y
= k y = ka
n
y

z
= k z = ka
n
z
University of California, Berkeley EECS 117 Lecture 23 p. 4/21
Wavefront
w
a
v
e
f
r
o
n
t

p
l
a
n
e
r
k
Recall that a wavefront is a surface of constant phase
for the wave
Then a
n
R = constant denes the surface of constant
phase. But this surface does indeed dene a plane
surface. Thus we have a plane wave. Is it TEM?
University of California, Berkeley EECS 117 Lecture 23 p. 5/21
E is a Normal Wave
Since our wave propagations in a source free region,
E = 0. Or
E
0

_
e
jka
n
r
_
= 0

_
e
jka
n
r
_
=
_
x

x
+ y

y
+z

z
_
e
j(
x
x+
y
y+
z
z)
= j(
x
x +
y
y +
z
z)e
j(
x
x+
y
y+
z
z)
So we have
jk(E
0
a
n
)e
jka
n
r
= 0
This implies that a
n
E
0
= 0, or that the wave is
polarized transverse to the direction of propagation
University of California, Berkeley EECS 117 Lecture 23 p. 6/21
H is also a Normal Wave
Since H(r) =
1
j
E, we can calculate the direction
of the H eld
Recall that (fF) = fF +f F
H(r) =
1
j

_
e
jkr
_
E
0
H(r) =
1
j
E
0

_
jke
jkr
_
H(r) =
k

a
n
E(r) =
1

a
n
E(r)
=

k
=

=
_
/
University of California, Berkeley EECS 117 Lecture 23 p. 7/21
TEM Waves
So we have done it. We proved that the equations
E(r) = E
0
e
jkr
H(r) =
1

a
n
E(r)
describe plane waves where E is perpendicular to the
direction of propagation and the vector H is
perpendicular to both the direction of propagation and
the vector E
These are the simplest general wave solutions to
Maxwells equations.
University of California, Berkeley EECS 117 Lecture 23 p. 8/21
Wave Polarization
Now we can be more explicit when we say that a wave
is linearly polarized. We simply mean that the vector E
lies along a line. But what if we take the superposition
of two linearly polarized waves with a 90

time lag
E(z) = xE
1
(z) + yE
2
(z)
The rst wave is x-polarized and the second wave is
y-polarized. The wave propagates in the z direction
In the time-harmonic domain, a phase lag corresponds
to multiplication by j
E(z) = xE
10
e
jz
j yE
20
e
jz
University of California, Berkeley EECS 117 Lecture 23 p. 9/21
Elliptical Polarization
In time domain, the waveform is described by the
following equation
E(z, t) =
_
E(z)e
jt
_
E(z, t) = xE
10
cos(t z) + yE
20
sin(t z)
At a paricular point in space, say z = 0, we have
E(0, t) = xE
10
cos(t) + yE
20
sin(t)
Thus the wave rotates along an elliptical path in the
phase front!
We can thus create waves that rotate in one direction or
the other by simply adding two linearly polarized waves
with the right phase
University of California, Berkeley EECS 117 Lecture 23 p. 10/21
Oblique Inc. on a Cond. Boundary
Let the x- y plane dene
the plane of incidence.
Consider the polarization
of a wave impinging
obliquely on the boundary.
We can identify two
polarizations,
perpendicular to the plane
and parallel to the plane of
incidence. Lets solve
these problems separately.
Any other polarized wave
can always be decom-
posed into these two cases
E
i
H
i
E
r
H
r

r
k
i
k
r
=
A perpendicularly polarized
wave.
University of California, Berkeley EECS 117 Lecture 23 p. 11/21
Perpendicular Polarization
Let the angle of incidence and relfection we given by
i
and
r
. Let the boundary consists of a perfect conductor
E
i
= yE
i0
e
jk
1
r
where k
1
= k
1
a
ni
and a
ni
= xsin
i
+z cos
i
E
i
= yE
i0
e
jk
1
(xsin
i
+z cos
i
)
H
i
=
1

i
a
n
E
i
For the reected wave, similarly, we have
a
nr
= xsin
r
z cos
r
so that
E
r
= yE
r0
e
jk
1
(xsin
r
z cos
r
)
University of California, Berkeley EECS 117 Lecture 23 p. 12/21
Conductive Boundary Condition
The conductor enforces the zero tangential eld
boundary condition. Since all of E is tangential in this
case, at z = 0 we have
E
1
(x, 0) = E
i
(x, 0) +E
r
(x, 0) = 0
Substituting the relations we have
y
_
E
i0
e
jk
1
xsin
i
+ E
r0
e
jk
1
xsin
r
_
= 0
For this equation to hold for any value of x and , the
following conditions must hold
E
r0
= E
i0

i
=
r
University of California, Berkeley EECS 117 Lecture 23 p. 13/21
Snells Law
We have found that the angle of incidence is equal to
the angle of reection (Snells law)
The total eld, therefore, takes on an interesting form.
The reected wave is simply
E
r
= yE
i0
e
jk
1
(xsin
i
z cos
i
)
H
r
=
1

1
a
nr
E
r
H
r
=
E
i0

1
( xcos
i
z sin
i
)e
jk
1
(xsin
i
z cos
i
)
The total eld is thus
E
1
= E
i
+E
r
= yE
i0
_
e
jk
1
z cos
i
e
jk
1
z cos
i
_
e
jk
1
xsin
i
University of California, Berkeley EECS 117 Lecture 23 p. 14/21
The Total Field
Simplifying the expression for the total eld
E
1
= y2jE
i0
sin(k
1
z cos
i
)
. .
standing wave
e
jk
1
xsin
i
. .
prop. wave
H
1
=
2E
i0

1
( x cos
i
cos(k
1
z cos
i
)e
jk
1
xsin
i
+
z j sin
i
sin(k
1
z cos
i
)e
jk
1
xsin
i
)
University of California, Berkeley EECS 117 Lecture 23 p. 15/21
Important Observations
In the z-direction, E
1y
and H
1x
maintain standing wave
patterns (no average power propagates since E and H
are 90

out of phase. This matches our previous

calculation for normal incidence
Waves propagate in the x-direction with velocity
v
x
= /(k
1
sin
i
)
Wave propagation in the x-direction is a non-uniform
plane wave since its amplitude varies with z
University of California, Berkeley EECS 117 Lecture 23 p. 16/21
TE Waves
Notice that on plane surfaces where E = 0, we are free
to place a conducting plane at that location without
changing the elds outside of the region
In particular, notice that E = 0 when
sin(k
1
z cos
i
) = 0
k
1
z cos
i
=
2

1
z cos
i
= m
This holds for m = 1, 2, . . .
So if we place a plane conductor at z =
m
1
2 cos
i
, there
will be a guided wave traveling between the two
planes in the x direction
Since E
1x
= 0, this wave is a TE wave as H
1x
= 0
University of California, Berkeley EECS 117 Lecture 23 p. 17/21
Parallel Polarization (I)
E
i
H
i
E
r
H
r

r
k
i
k
r
=
Now the wave is
polarized in the plane
of incidence.
The approach is sim-
ilar to before but the
tangential component
of the electric eld de-
pends on the angle of
incidence
University of California, Berkeley EECS 117 Lecture 23 p. 18/21
Parallel Polarization (II)
Now consider an incident electric eld that is in the
plane of polarization
E
i
= E
i0
( xcos
i
z sin
i
)e
jkr
H
i
= y
E
i0

1
e
jkr
Likewise, the reected wave is expressed as
E
r
= E
r0
( xcos
i
+z sin
i
)e
jkr
H
r
= y
E
r0

1
e
jkr
Note that k
i,r
r = xsin
i,r
z cos
i,r
University of California, Berkeley EECS 117 Lecture 23 p. 19/21
Tangential Boundary Conditions
Since the sum of the reected and incident wave must
have zero tangential component at the interface
E
i0
cos
i
e
jk
1
xsin
i
+ E
r0
cos
r
e
jk
1
xsin
r
= 0
These equations must hold for all . Thus E
r0
= E
i0
as
before
Thus we see that these equations can hold for all
values of x if and only if
i
=
r
University of California, Berkeley EECS 117 Lecture 23 p. 20/21
The Total Field (Again)
Using similar arguments as before, when we sum the
elds to obtain the total eld, we can observe a
standing wave in the z direction and wave propagation
in the x direction.
Note that the magnetic eld H = H y is always
perpendicular to the direction of propagation but the
electric eld has a component in the x direction.
This type of wave is known as a TM wave, or
transverse magnetic wave
University of California, Berkeley EECS 117 Lecture 23 p. 21/21