Part 4: Waves On Boundaries: PHYS370 - Advanced Electromagnetism

PHYS370 Advanced Electromagnetism
Part 4: Waves on Boundaries

Reection and Transmission of EM Waves
When plane waves are incident on a boundary between dierent
media, some energy crosses the boundary, and some is
reected.
We dene transmission and reection coecients to quantify
the transmission and reection of wave energy. These
coecients are properties of the two media.
The transmission and reection coecients are determined by
matching the electric and magnetic elds in the waves at the
boundary between the two media.
Advanced Electromagnetism 1 Part 4: Waves on Boundaries
Reection and Transmission of EM Waves
In this part of the course, we shall consider:
Boundary conditions on electric and magnetic elds.
Boundary conditions on elds at the surfaces of conductors.
Monochromatic plane wave on a boundary:
directions of reected and transmitted waves (laws of reection and
refraction);
amplitudes of reected and transmitted waves (Fresnels equations);
the special case of a boundary between two dielectrics;
the special case of the surface of a conductor.
Monochromatic plane wave on a boundary between two dielectrics:
polarisation by reection;
total internal reection.
Reection coecient for a conducting surface.
Boundary Conditions 1: Normal Component of

B
We can use Maxwells equations to derive the boundary
conditions on the magnetic eld across a surface. Consider a
pillbox across the surface.

B
Take Maxwells equation:

B = 0 (1)
integrate over the volume of the pillbox, and apply Gauss
theorem:
_
V

BdV =
_
S
B

dS = 0 (2)
where V is the volume of the pillbox, and S is its surface. We
can break the integral over the surface into three parts: over
the at ends (S
1
and S
2
) and over the curved wall (S
3
):
_
S
1
B

dS +
_
S
2
B

dS +
_
S
3
B

dS = 0 (3)

B
In the limit that the length of the pillbox approaches zero, the
integral over the curved surface also approaches zero. If each
end has a small area A, then equation (3) becomes:
B
1n
A+B
2n
A = 0 (4)
or:
B
1n
= B
2n
(5)
In other words, the normal component of the magnetic eld

B
must be continuous across the surface.
Boundary Conditions 2: Tangential Component of

E
Consider a loop spanning the surface.

E

E =
B (6)
Integrate over the surface bounded by the loop and apply
Stokes theorem to get:
_
S

E

dS =
_
C
E

dl =
t
_
S
B

dS (7)
Now take the limit in which the width of the loop becomes
zero. The contributions to the integral around the loop C from
the narrow ends become zero, as does the integral of the
magnetic eld across the area bounded by the loop. We are
left with:
E
1t
l E
2t
l = 0 (8)
which means that:
E
1t
= E
2t
(9)
Therefore, the tangential component of the electric eld is
continuous across the boundary.

D
Consider a pillbox crossing the boundary.

D

D = (10)
Integrate over the volume of the pillbox, and apply Gauss
theorem:
_
V

DdV =
_
S
D

dS =
_
V
dV (11)
Now we take the limit in which the height of the pillbox
becomes zero. We assume that there is a surface charge density
s
. If the at ends of the pillbox have (small) area A, then:
D
1n
A+D
2n
A =
s
A (12)
Dividing by the area A, we arrive at:
D
2n
D
1n
=
s
(13)
Note that if the surface charge density is zero, the normal
component of

D is continuous across the surface. However,
this is not true for the normal component of

E, unless the two
materials have identical permittivities.

H
Consider a loop across the boundary.

H

H =

J +

D (14)
Integrate over the surface bounded by the loop, and apply
Stokes theorem to obtain:
_
S

H

dS =
_
C
H

dl =
_
S
J

dS +

t
_
S
D

dS (15)
As before, take the limit where the lengths of the narrow edges
of the loop become zero. Then we nd that:
H
1t
l H
2t
l = J
s
l (16)
or:
H
1t
H
2t
= J
s
(17)
where J
s
represents a surface current density perpendicular to
the direction of the tangential component of

H that is being
matched.

H
The concept of surface current density is analogous to that of
surface charge density: it represents a nite current in an
innitesimal layer of the material.
If the material has nite conductivity, then an innitesimal layer
of the material has innite resistance, and no current can ow
(if the electric eld is nite).
Therefore, for a material with nite conductivity, we have:
H
1t
= H
2t
(18)
That is, the tangential component of

H is continuous across
the boundary.
Summary of Boundary Conditions
The general conditions on electric and magnetic elds at the
boundary between two materials can be summarised as follows:
Boundary condition: Derived from... ...applied to:
B
2n
= B
1n

B = 0 pillbox
E
2t
= E
1t

E =
B loop
D
2n
D
1n
=
s

D = pillbox
H
2t
H
1t
= J
s

H =

J +

D loop
Boundary Conditions on Surfaces of Conductors
Static electric elds cannot persist inside a conductor. This is
simply because the free charges within the conductor will
re-arrange themselves to cancel any electric eld; this can
result in a surface charge density,
s
.
We have seen that electromagnetic waves can pass into a
conductor, but the eld amplitudes fall exponentially with
decay length given by the skin depth, :

(19)
As the conductivity increases, the skin depth gets smaller.
Since both static and oscillating electric elds vanish within a
good conductor, we can write the boundary conditions at the
surface of such a conductor:
E
1t
0 E
2t
0
D
1n

s
D
2n
0
Lenzs law states that a changing magnetic eld will induce
currents in a conductor that will act to oppose the change.
In other words, currents are induced that will tend to cancel
the magnetic eld in the conductor.
This means that a good conductor will tend to exclude
magnetic elds.
Thus the boundary conditions on oscillating magnetic elds at
the surface of a good conductor can be written:
B
1n
0 B
2n
0
H
1t
J
s
H
2t
0
We can consider an ideal conductor as having innite
conductivity.
In that case, we would expect the boundary conditions to
become:
B
1n
= 0 B
2n
= 0
E
1t
= 0 E
2t
= 0
D
1n
=
s
D
2n
= 0
H
1t
= J
s
H
2t
= 0
Strictly speaking, the boundary conditions on the magnetic
eld apply only to oscillating elds, and not to static elds.
But it turns out that for superconductors, static magnetic
elds are excluded as well as oscillating magnetic elds. This is
not expected for classical ideal conductors.
Superconductors and the Meissner Eect
Although superconductors have innite conductivity, they
cannot be understood in terms of classical theories in the limit
.
Superconductivity is a quantum phenomenon: one aspect of
this is the Meissner eect, which refers to the expulsion of all
magnetic elds (static as well as oscillating) from within a
superconductor.
In fact, even in a superconductor, the magnetic eld is not
completely excluded from the material but penetrates a small
distance (the London penetration depth, typically around
100 nm) into the material.
As long as the applied magnetic eld is not too large, a sample
of material cooled below its critical temperature will expel any
magnetic eld as it undergoes the phase transition to
superconductivity: when this happens, a magnet placed on top
of the sample will start to levitate.
The Meissner eect allows us to classify superconductors into
two distinct classes:
Type I superconductors: above a certain critical eld H
c
(which depends on the temperature), superconductivity is
abruptly destroyed.
Type II superconductors: above one critical eld value
H
c1
, the magnetic eld starts to penetrate, but the
electrical resistance remains zero. Above a second, higher
critical eld value H
c2
, superconductivity is abruptly
destroyed.
Critical Fields in Niobium (Type II Superconductor)
R.A. French, Intrinsic Type-2 Superconductivity in Pure Niobium,
Cryogenics, 8, 301 (1968). Note: t = T/T
c
. The critical temperature for
niobium is T
c
= 9.2 K.
Waves on Boundaries
We now apply the boundary conditions to an electromagnetic
wave incident on a boundary between two dierent materials.
We shall use the boundary conditions to derive the properties of
the reected and transmitted waves, for a given incident wave.
Consider a monochromatic wave incident at some angle on a
boundary. We must consider three waves: the incident wave
itself; the reected wave, and the transmitted wave on the far
side of the boundary.
The electric eld components for these waves can be written
(respectively):
E
I
(r, t) =

E
0I
e
j(
I
t
k
I
r)
(20)
E
R
(r, t) =

E
0R
e
j(
R
t
k
R
r)
(21)
E
T
(r, t) =

E
0T
e
j(
T
t
k
T
r)
(22)
Waves on Boundaries
Waves on Boundaries
Let us rst consider the time dependence of the waves. The
boundary conditions must apply at all times: for example, the
tangential component of the electric eld, E
t
, must be
continuous across the boundary at all points on the boundary
at all times.
This means that all waves must have the same time
dependence, and therefore:
I
=
R
=
T
= (23)
Reection at a boundary cannot change the frequency of an
incident monochromatic wave. Some surfaces reect some
wavelengths better than others, which is why they can appear
coloured under white light; but the frequency of the light does
not change.
Laws of Reection and Refraction
Now let us consider the relationships between the directions in
which the waves are moving.
We shall nd that these relationships are just the laws of
reection and refraction that we are familiar with from basic
optics.
However, our goal is now to derive these laws from Maxwells
equations, by applying the boundary conditions on elds in
waves across boundaries.
We start from the fact that the boundary conditions must be
satised at all points on the boundary.
This means that the waves must all change phase in the same
way as we move from one point to another on the boundary.
Since the phase of each of the waves at a position r is given by
k r, where

k is the appropriate wave vector, we must have:
k
I
p =
k
R
p =
k
T
p (24)
where p is any point on the boundary.
For simplicity, let us choose our coordinates so that the
boundary lies in the plane z = 0. Then any point p on the
boundary can be written:
p = (x, y, 0) (25)
Now we can (without loss of generality) further specify the
coordinate system so that

k
I
lies in the x z plane, i.e. the y
component of

k
I
is zero:
k
I
= (k
I
sin
I
, 0, k
I
cos
I
) (26)
where
I
is the angle between the direction of travel of the
incident wave and the boundary.
Now let us apply equation (24):
k
I
p =
k
R
p =
k
T
p
to points on the boundary with x = 0, i.e. p = (0, y, 0). We
nd:
k
Iy
= k
Ry
= k
Ty
= 0 (27)
Therefore, the directions of the incident, reected and
transmitted waves all lie in the plane y = 0.
Now let us consider points on the boundary with y = 0, i.e.
p = (x, 0, 0). This time, using equation (24) gives:
k
Ix
= k
Rx
= k
Tx
= k
I
sin
I
(28)
which (since the vertical components of the wave vectors are
all zero) can be written:
k
I
sin
I
= k
R
sin
R
= k
T
sin
T
(29)
But since the incident and reected waves are travelling in the
same material with the same frequency, the magnitudes of the
wave vectors must be the same:
k
I
= k
R
(30)
Combining equations (29) and (30) we nd:
I
=
R
the law of reection (31)
sin
I
sin
T
=
k
T
k
I
the law of refraction (Snells law) (32)
Reection and Refraction at a Boundary Between Dielectrics
As an example, consider a monochromatic wave incident on a
boundary between two dielectrics (e.g. air and glass).
Since the conductivity is zero on both sides of the boundary,
the wave vectors of all waves must be real.
Also, we have:
k
I
= v
1
,

k
T
= v
2
(33)
where v
1
is the phase velocity in medium 1, and v
2
is the phase
velocity in medium 2.
Then equation (32) gives us:
sin
I
sin
T
=
v
1
v
2
(34)
We dene the refractive index n of a material as the ratio of the
speed of light in a vacuum to the speed of light in the material:
n =
c
v
(35)
Then equation (34) can be written:
sin
I
sin
T
=
n
2
n
1
(36)
This is the familiar form of Snells law.
Reection and Refraction at the Surface of a Conductor
For a wave incident on a conductor,

k
T
will be complex:
k
T
= j
(37)
For a good conductor (i.e.
2
):

_
2
2
(38)
so:
k
T
=
_

2
+

2
=
2
(39)
Applying the law of refraction (32):
sin
I
sin
T
=
k
T
k
I
1
1 (40)
where we have assumed that
2

1
.
Since the largest value of sin
I
is 1, equation (40) tells us that
sin
T
0, so the direction of the transmitted wave in a good
conductor must be (close to the) normal to the surface.
Intensity of Reected and Refracted Waves
Having derived the relationships between the directions of the
incident, reected and refracted waves, we turn now to the
amplitudes of the waves.
We can again apply the boundary conditions on the
electromagnetic elds at a boundary to derive relationships
between the wave amplitudes.
It turns out that the relative amplitudes and phases of the
waves depend on the electromagnetic impedances of the
materials on either side of the boundary.
The results we nd are summarised in a set of equations known
as Fresnels equations.
Consider an electromagnetic wave incident on a boundary with
the electric eld in the wave normal to the plane of incidence.
We call this N polarisation.
Since the tangential component of the electric eld is
continuous across the boundary (at any time and any point on
the boundary), we have:
E
0I
+

E
0R
=

E
0T
(41)
The tangential component of the magnetic eld must also be
continuous across the boundary:
H
0I
cos
I
H
0R
cos
I
= H
0T
cos
T
(42)
The ratio of the electric eld to the magnetic intensity in a
wave is given by the impedance of the medium:
E
0
H
0
= Z (43)
Substituting from (43) into (42) we get:
(E
0I
E
0R
)
Z
1
cos
I
=
E
0T
Z
2
cos
T
(44)
Now we solve equations (41) and (44) for the amplitudes of
the reected and transmitted waves, as fractions of the
amplitude of the incident wave:
_
E
0R
E
0I
_
N
=
Z
2
cos
I
Z
1
cos
T
Z
2
cos
I
+Z
1
cos
T
(45)
_
E
0T
E
0I
_
N
=
2Z
2
cos
I
Z
2
cos
I
+Z
1
cos
T
(46)
It is important to remember that equations (45) and (46) apply
only to the case that the incident wave is polarised with the
electric eld normal to the plane of incidence, i.e. for N
polarisation.
Consider equations (45) and (46) in the special case of a
boundary between two dielectrics.
Since the refractive index n of a dielectric is given by:
n =
c
v
=
0
(47)
we can write:
Z =
_
0
Z
0
n
(48)
For non-magnetic dielectrics, we have
1
=
2
=
0
, so that:
Z
1
=
Z
0
n
1
, Z
2
=
Z
0
n
2
(49)
Then, equations (45) and (46):
_
E
0R
E
0I
_
N
=
Z
2
cos
I
Z
1
cos
T
Z
2
cos
I
+Z
1
cos
T
_
E
0T
E
0I
_
N
=
2Z
2
cos
I
Z
2
cos
I
+Z
1
cos
T
become:
_
E
0R
E
0I
_
N
=
n
1
cos
I
n
2
cos
T
n
1
cos
I
+n
2
cos
T
(50)
_
E
0T
E
0I
_
N
=
2n
1
cos
I
n
1
cos
I
+n
2
cos
T
(51)
The Fresnel equations above are frequently written in terms of
the refractive index rather than the impedance; but then the
equations are valid only for non-magnetic dielectrics.
Now consider the special case of a plane wave incident on the
surface of a good conductor, again polarised so that the
electric eld is parallel to the surface.
Recall that the impedance of a conductor is given in general by:
Z = (1 +j)
_
2
(52)
If the conductor is non-magnetic, so that
0
, and if the
permittivity is also close to the permittivity of free space, then
for a good conductor with :
|Z| Z
0
_
Z
0
(53)
If we assume that the wave is incident on a good conductor
from a dielectric with impedance of the same order of
magnitude as the impedance of free space, then we have:
Z
1
Z
0
, |Z
2
| Z
0
|Z
2
| Z
1
(54)
Then, equations (45) and (46) become:
_
E
0R
E
0I
_
N
=
Z
2
cos
I
Z
1
cos
T
Z
2
cos
I
+Z
1
cos
T
1 (55)
_
E
0T
E
0I
_
N
=
2Z
2
cos
I
Z
2
cos
I
+Z
1
cos
T
0 (56)
Therefore, there is nearly 100% reection from a metal surface,
with a phase change of 180
. The phase change means that

there is a cancellation between the tangential components of
the electric eld in the incident and reected waves.
Now consider the case of an monomchromatic wave incident on
a boundary, with the electric eld parallel to the plane of
incidence.
We call this P polarisation.
We apply the boundary conditions as before.
First, since the component of the electric eld tangential to the
boundary must be continuous across the boundary, we have:
E
0I
cos
I
+E
0R
cos
I
= E
0T
cos
I
(57)
Next, since the tangential component of the magnetic eld
must also be continuous across the boundary:
H
0I
H
0R
= H
0T
(58)
We can again express the magnetic intensity in terms of the
electric eld and the impedance:
E
0
H
0
= Z (59)
Proceeding as before, we nd for the ratios of the wave
amplitudes:
_
E
0R
E
0I
_
P
=
Z
2
cos
T
Z
1
cos
I
Z
2
cos
T
+Z
1
cos
I
(60)
_
E
0T
E
0I
_
P
=
2Z
2
cos
I
Z
2
cos
T
+Z
1
cos
I
(61)
Equations (60) and (61) are valid for the electric eld in the
incident wave parallel (P) to the plane of incidence.
Compare equations (60) and (61) with (45) and (46):
_
E
0R
E
0I
_
N
=
Z
2
cos
I
Z
1
cos
T
Z
2
cos
I
+Z
1
cos
T
_
E
0T
E
0I
_
N
=
2Z
2
cos
I
Z
2
cos
I
+Z
1
cos
T
When the materials on both sides of the boundary are
dielectrics, we can write equations (60) and (61) in terms of
the refractive indices n
1
and n
2
.
_
E
0R
E
0I
_
P
=
n
1
cos
T
n
2
cos
I
n
1
cos
T
+n
2
cos
I
(62)
_
E
0T
E
0I
_
P
=
2n
1
cos
I
n
1
cos
T
+n
2
cos
I
(63)
When the wave is incident from a dielectric onto the surface of
a good conductor ( ), if we assume non-magnetic
materials with permittivity close to that of vacuum, we have:
Z
1
Z
0
, |Z
2
| Z
0
|Z
2
| Z
1
(64)
Then:
_
E
0R
E
0I
_
P
=
Z
2
cos
T
Z
1
cos
I
Z
2
cos
T
+Z
1
cos
I
1 (65)
_
E
0T
E
0I
_
P
=
2Z
2
cos
I
Z
2
cos
T
+Z
1
cos
I
0 (66)
There is nearly 100% reection from a metal surface, with a
phase change of 180
: this is the case for both N and P

polarisation.
Fresnels Equations
The equations (45), (46), (60) and (61) are known as Fresnels
equations:
_
E
0R
E
0I
_
N
=
Z
2
cos
I
Z
1
cos
T
Z
2
cos
I
+Z
1
cos
T
_
E
0T
E
0I
_
N
=
2Z
2
cos
I
Z
2
cos
I
+Z
1
cos
T
_
E
0R
E
0I
_
P
=
Z
2
cos
T
Z
1
cos
I
Z
2
cos
T
+Z
1
cos
I
_
E
0T
E
0I
_
P
=
2Z
2
cos
I
Z
2
cos
T
+Z
1
cos
I
Fresnels Equations for Normal Incidence
For a wave incident normal to a boundary,
I
=
T
= 0, and
Fresnels equations (for electric elds both normal to, and
parallel to the plane of incidence) become:
E
0R
E
0I
=
Z
2
Z
1
Z
2
+Z
1
and
E
0T
E
0I
=
2Z
2
Z
2
+Z
1
(67)
Note that since the energy in a wave is proportional to the
square of its amplitude divided by the impedance of the
medium, we can easily show that energy is conserved in this
case:
E
2
0R
E
2
0I
+
Z
1
Z
2
E
2
0T
E
2
0I
=
(Z
2
Z
1
)
2
+4Z
1
Z
2
(Z
2
+Z
1
)
2
= 1 (68)
and hence:
E
2
0R
Z
1
+
E
2
0T
Z
2
=
E
2
0I
Z
1
(69)
Of course, energy is also conserved for general angles of
incidence, but the algebra is slightly more complicated.
Wave Incident on a Boundary Between Dielectrics
There are two important phenomena associated with
electromagnetic waves incident on a boundary between two
dielectrics. These are:
Polarisation by reection.
Total internal reection.
Both of these eects can be understood from the relationships
between the angles and intensities of the incident, reected and
transmitted waves that we have derived. We shall discuss each
in turn.
Expressing Fresnels equations in terms of the refractive indices
n
1
and n
2
for the two dielectrics:
_
E
0R
E
0I
_
N
=
n
1
cos
I
n
2
cos
T
n
1
cos
I
+n
2
cos
T
_
E
0T
E
0I
_
N
=
2n
1
cos
I
n
1
cos
I
+n
2
cos
T
_
E
0R
E
0I
_
P
=
n
1
cos
T
n
2
cos
I
n
1
cos
T
+n
2
cos
I
_
E
0T
E
0I
_
P
=
2n
1
cos
I
n
1
cos
T
+n
2
cos
I
The angles of incidence and transmission are related by the law
of refraction (32):
sin
I
sin
T
=
n
2
n
1
Since there are no imaginary terms in the ratios of the eld
amplitudes, E
0R
and E
0T
are either in phase with E
0I
(positive ratio) or out of phase (negative ratio).
For all n
1
and n
2
, E
0T
is always in phase with E
0I
.
If n
1
< n
2
(incident wave in low-density material, and
transmitted wave in high-density material):
_
E
0R
E
0I
_
N
< 0 (70)
so the reected and incident waves are always out of
phase.
Polarisation by Reection
Depending on the values of
I
and n
2
/n
1
, the ratio
_
E
0R
E
0I
_
P
(71)
can be positive or negative. It is zero when:
n
2
cos
I
= n
1
cos
T
(72)
Using the law of refraction (32):
n
1
n
2
=
sin
T
sin
I
(73)
we nd that the condition (72) can be written:
sin
I
cos
I
sin
T
cos
T
= 0 (74)
Using the trigonometric identity:
sin2A = 2sinAcos A (75)
equation (74) becomes:
sin2
I
sin2
T
= 0 (76)
Using a further trigonometric identity:
sinAsinB = 2cos
_
A+B
2
_
sin
_
AB
2
_
(77)
equation (76) becomes:
cos(
I
+
T
) sin(
I
T
) = 0 (78)
Therefore, the condition for zero reection (for the electric
eld parallel to the plane of incidence) becomes:
I
+
T
=

2
(79)
Substituting from equation (79) into the law of refraction (32),
we nd:
sin
T
sin
I
=
sin(
I
)
sin
I
=
cos
I
sin
I
=
n
1
n
2
(80)
The angle of incidence at which condition (80) is satised is
called the Brewster angle,
B
:
tan
B
=
n
2
n
1
(81)
When the angle of incidence is equal to the Brewster angle, a
wave with the electric eld parallel to the plane of incidence
has zero reected amplitude.
A wave with electric eld normal to the plane of incidence is
still reected. As a result, the wave becomes polarised.
Polarisation by reection enables us to use polarising lters to
reduce the amount of glare from sunlight reecting o the
surface of a lake, or o a wet road...
Total Internal Reection
Consider again the law of refraction (32):
sin
T
=
n
1
n
2
sin
I
(82)
If n
1
> n
2
, then for suciently large
I
, we apparently have:
sin
T
> 1 (83)
We interpret this as meaning that under these conditions, there
is no refracted wave: all the energy in the incident wave is
reected from the boundary.
The angle of incidence for which sin
T
= 1 is called the critical
angle,
c
:
sin
c
=
n
2
n
1
(84)
Total Internal Reection
Reection Coecient for a Good Conductor
Finally (for this part of the lecture course), we shall consider in
a little more detail reection from the surface of a good
conductor.
Recall that we found previously that, for polarisation both
normal and parallel to the plane of incidence, the amplitudes of
the reected and transmitted waves were given approximately
by:
_
E
0R
E
0I
_
P
=
Z
2
cos
T
Z
1
cos
I
Z
2
cos
T
+Z
1
cos
I
1
_
E
0T
E
0I
_
P
=
2Z
2
cos
I
Z
2
cos
T
+Z
1
cos
I
0
We shall try to take our analysis a stage further, so as to arrive
at a more accurate estimate for the amount of reected light.
For simplicity, we shall only consider normal incidence.
Consider a wave incident on the surface of a conductor at
normal incidence (
I
= 0). From Fresnels equations, the
amplitude of the reected wave is given by:
_
E
0R
E
0I
_
=
Z
2
Z
1
Z
2
+Z
1
(85)
The impedance for the wave in the conductor is given, in
general, by:
Z
2
= (1 +j)
2
2
2
(86)
Assuming non-magnetic materials, and that the conductor
satises the good conductor condition
2
, we shall have:
|Z
2
| Z
1
(87)
Thus, we can write equation (85) for the reected wave
amplitude in terms of a small parameter x:
_
E
0R
E
0I
_
=
1 x
1 +x
, x =
Z
2
Z
1
, |x| 1 (88)
Since x is small, we can make a Taylor series expansion:
(1 +x)
1
= 1 x +O
_
x
2
_
(89)
Using this in our expression for E
0R
/E
0I
from equation (88),
we nd:
E
0R
E
0I
= (1 x)(1 +x)
1
= (1 x)(1 x +O(x
2
))
= (1 2x +O(x
2
)) (90)
We nd:
E
0R
E
0I
1 +2
Z
2
Z
1
= 1 +(1 +j)
2
2
2
(91)
By a similar calculation, we nd:
E
0T
E
0I
2
Z
2
Z
1
= (1 +j)
2
2
2
(92)
Let us dene the real parameter , given by:
=
2
2
2
1 (93)
In terms of , the reected and transmitted wave amplitudes
are:
E
0R
E
0I
1 +(1 +j) (94)
E
0T
E
0I
(1 +j) (95)
These expressions are convenient for drawing the phases and
amplitudes of the reected and transmitted waves, relative to
the incident wave...
E
0R
( 1 +j) E
0I
(96)
E
0T
( +j) E
0I
(97)
In the case of a plane wave incident on a good conductor, we
nd that:
The phase of E
0T
leads the phase of E
0I
by 45
.
The phase of E
0R
leads the phase of E
0I
by:
tan
1
_

1
_
(98)
As
2
, E
0T
0 and E
0R
E
0I
as expected.
As
2
0, we do not get the expected dielectric formulae:
this situation violates many of the assumptions we made.
We dene the reection coecient R to be the fraction of
energy reected from a surface. R is the ratio of the time
averaged Poynting vector,

S =

E

H, for the incident and
reected waves.
For a plane wave with normal incidence on a good conductor:
R =

S
R
S
I
E
R

H
R
E
I

H
I
E
0R
E
0I
2
(99)
Since, from equation (94) the amplitude of the reected wave
is given by:
E
0R
E
0I
= 1 +(1 +j) (100)
we nd that:
E
0R
E
0I
2
= (1 +(1 +j)) (1 +(1 j)) 1 2 (101)
where we have used the fact that || 1 to drop a term in
2
.
Therefore:
R 1 2 = 1 2
2
2
2
(102)
Consider the example of 3 cm microwave radiation at normal
incidence on copper:
frequency of radiation = 2 10
10
s
1
conductivity of copper
2
= 5.6 10
7
(m)
1
permeability
1
=
2
=
0
permittivity
1
=
0
Using equation (102), we nd:
R 1 2
2
2
2
0.99972 (103)
Therefore, 99.97% of the incident power is reected.
From equation (102),
R 1 2
2
2
2
(104)
we observe that:
R gets closer to 1 at low frequencies.
Only at very high frequencies does the deviation from 1
become signicant for good conductors such as copper,
aluminium or silver.
We also note that since the skin depth is small, thin sheets of a
good conductor provide excellent shielding for long-wavelength
electromagnetic radiation.
Summary of Part 4: Waves on Boundaries
You should be able to:
Derive (from Maxwells equations) the boundary conditions on electric
and magnetic elds at the interface between two media.
Apply the boundary conditions on electric and magnetic elds to derive
the laws of reection and refraction.
Apply the boundary conditions to derive the relative amplitudes of the
reected and refracted waves for dierent polarisations (Fresnels
equations).
Apply Fresnels equations to boundaries between two dielectrics, and to
the surfaces of good conductors, to explain the behaviour of waves at
such boundaries.
Plot the variation in reected and refracted wave amplitudes as
functions of angle of incidence, for the boundary between two
dielectrics.
Explain the signicance of the Brewster angle and the critical angle,
and derive expressions for these angles, in terms of the refractive
indices of the media on either side of the boundary.

Part 4: Waves On Boundaries: PHYS370 - Advanced Electromagnetism

Diunggah oleh

Informasi Dokumen

Deskripsi Asli:

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Part 4: Waves On Boundaries: PHYS370 - Advanced Electromagnetism

Diunggah oleh

Hak Cipta:

Format Tersedia

PHYS370 Advanced Electromagnetism

Part 4: Waves on Boundaries

. The phase change means that

: this is the case for both N and P

Anda mungkin juga menyukai