Addis Ababa University

(CNCS)

Department of Physics

Electromagnetic Theory(Phys 602)

Assignment 2 (On Chapter 2)

Solutions

By:

Shinie Shewangizaw

ID.No. GSR/3647/12

Submitted to Dr. Teshome S.

May 07, 2020

 1

1. (7.2).
Solution

a) For Simplicity, we consider only the case where the incident is normal to the interface. The
Poynting flux
2
~ n = c Re[|~n.~k||E00 2 |]
S.~
8πω

Figure 1

Since ~k, ~n and S

~ are parallel, therefore it can be written as

c2 0
~ =
|S| Re[|~k||E02 |]
8πω
and
k2 n2
=
k1 n1
where n1 and n2 are refractive indices for the respective media.
 2
Sr nr  Er 
=
ki ni  Ei 

We have to calculate the ratio of magnitude of the square of electric field. Assuming that
the incident wave comes from the left and a refracted wave traveling to the right and
in the medium number 2 of thickness d the wave is the result of superposition of two
waves traveling in opposite directions, and in the third medium, only a transmitted wave
traveling to the left. This assumption does not lose generality. In fact it is the observed
consequences in the laboratory. Further more it is given that these media are non permeable
i.e., µ1 = µ2 = µ3 = 1.
• Medium 1

E~ i = E~0 eik1 z−iωt (1)

~ ~
B~ i = √1 k1 × Ei (2)
k1
E~r = E~1 e 1 z−iωt
−ik
(3)
~ ~
B~r = −√1 k1 × Er (4)
k1
 2

• Medium 2

B~m = (E~2+ eik2 z + E~2− e−ik2 z )e−iωt (5)

r ~
2 k2
~
Bm = × (E~2+ eik2 z − E~2− e−ik2 z )e−iωt (6)
µ2 k2

• Medium 3
~ t = E~3 eik3 z−iωt
E (7)
r ~ ~t
~t = 3 k3 × E
B (8)
µ3 k3

The boundary conditions:-

~ ⊥ Continuous
D
~ i + E~r ) + 2 E~m ].~n = 0
[1 (E (9)

~ ⊥ Continuous
B

[k~1 × E
~ i + k~1 × E~r − k~2 × E~m ].~n = 0 (10)

~ k Continuous
E

~ i + E~r − E~m ] × ~n = 0
[E (11)

~ k Continuous
H

1 ~ i + k~1 × E~r ) − 1 k~2 × E~m ] × ~n = 0

[ (k~1 × E (12)
µ µ2
Solving these equations, we obtained,
n3
(1 + n2
) ik d−ik d
E2+ = e
3 2
E3 (13)
2
(1 − nn23 )
E2− = eik3 d+ik2 d E3 (14)
2
(1 − nn12 )(1 + n3
n2
) n2
+ (1 +
n1
)(1 − nn23 )e2ik2 d
E1 = n2 n3 E0 (15)
(1 + n1
)(1 + n2
)+ (1 − nn21 )(1 − nn23 )e2ik0 d
4e−ik3 d E0
E3 = n2 n3 −ik2 d (16)
(1 − n1
)(1 − n2
)e + (1 + nn12 )(1 + nn32 )e−ik2 d
n n
Let rij be the ratio (1 − nji )/(1 + nji ).
The coefficient of reflection for medium 1,
 2
 E1 
R =   (17)
E0
2 2
r12 + r23 + 2r12 r23 cos 2k2 d
= 2 2
(18)
1 + 2r12 r23 cos 2k2 d + r12 r23
(r12 + r23 ) − 4r12 r23 sin2 k2 d
2
= (19)
(1 + 2r12 r23 )2 − 4r12 r23 sin2 k2 d
 3

The coefficient of reflection for medium 3,

 2
 E3 
R =   (20)
E0
 2
4  1 
= 2 2
  (21)
1 + 2r12 r23 cos 2k2 d + r12 r23 (1 + n2 /n1 )(1 + n3 /n2 ) 
 2
4  1 
= 2 2
  (22)
(1 + r12 r23 ) − 4r12 r23 sin k2 d (1 + n2 /n1 )(1 + n3 /n2 ) 


The transmission:-
4 (n1 n2 )2 n3
T ={ 2 2 2 2
} (23)
(1 + r12 r23 ) − 4r12 r23 sin k2 d (n1 + n2 ) (n2 + n3 ) n1
Notice that there are two forms of the results(i.e, one with positive and the other with cosine
in the expression, but they are are identical).
Case 1: n1 = 1, n2 = 2, n3 = 3, where k2 = 2π λ2
(or ki = nci ω

17 + 15 cos(4dω/c)
R =
113 + 15 cos(4dω/c)
16 − 15 sin2 (2πd/λ2 )
=
64 − sin2 (2πd/λ2 )
96
T =
113 + 15 cos(4dω/c)
48
=
64 − 15 sin2 (2πd/λ2 )
Case 2: n1 = 3, n2 = 2, n3 = 1. The results are the same as case 1.
Case 3: n1 = 2, n2 = 4, n3 = 1.
53 − 45 cos(8dω/c)
R =
117 − 45 cos(8dω/c
16 + 180 sin2 (2πd/λ2 )
=
144 + 180 sin(2πd/λ2 )
32
T =
9(4 + 5 sin2 (2πd/λ2 ))
53 − 45 cos(8πdω/c)
=
117 − 45 cos(8πdω/c)

b) For no reflected wave with frequency

(r12 + r23 )2 − 4r12 r23 sin2 (2πd/λ2 )
R = (24)
(1 + 2r12 r23 )2 − 4r12 r23 sin2 (kπd/λ2 )
therefore, the minimum occurs when the second term is maximum i.e., when (2πd/λ2 ) = π/2
(the lowest term) or d = λ43 .

(r12 − r23 )2
R= = 0; if r12 = r23
(1 − r12 r23 )2

√
∴ n2 = n1
 4

2. (7.5). A plane polarized electromagnetic wave E ~ = E ~ i ei~k.~x−iωt is incident on a flat uniform

sheet of an excellent conductor(σ >> ω0 ) having thickness D. Assuming that in space and in
the conducting sheet µµ0 = 00 = 1, discuss the reflection and transmission of the incident wave.

a) Show that the amplitudes of the reflected and transmitted waves, correct to the first order
in (0 ω/σ)1 /2, are

Er −(1 − e−2λ )
=
Ei (1 − e−2λ ) + γ(1 + e−2λ )

Et 2γe−λ
=
Ei (1 − e−2λ ) + γ(1 + e−2λ )

where
r
20 ω ωδ
γ= (1 − i) = (1 − i) (25)
σ c

(1 − i)D/σ (26)
p
and δ = 2/ωµσ is the penetration depth.
Solution
So long as we treat the conductor as a medium with complex dielectric constant
σ
/0 = 1 + i (27)
ω0

Electric field vectors of the form E ~ i ei~k.~x and E

~ r e−i~k.~x on the incident side, E
~ + eik~1 .~x and
~ − eik~1 .~x in the conductor, and E
E ~ t ei~k.~x−D~ on the transmitted side. We use matching for E ~
perpendicular to the plane of incidence (which corresponds to a sign convention of having
all electric fields pointing in the same direction). In this case, the matching becomes
First Interface

E k : Ei + Er = E+ + E−
H k : (Ei − Er ) = n(E+ − E− )

Second Interface

E+ eiφ + E− e−iφ = Et
n(E+ eiφ − E− e−iφ ) = Et

where n is the complex index of refraction

r r
 σ
n= = 1+i (28)
0 ω0

and φ is the phase change for going through the dielectric

r
ωn ωD σ
φ = k1 D = D= 1+i (29)
c c ω0
 5

~ t and E
Solving for E ~ r in terms of E
~ i , we obtain

Et 4/n 4/neiφ
= =
Ei (1 + 1/n)2 e−iφ − (1 − 1/n)2 eiφ (1 + 1/n2 )(1 − e−2iφ ) + 2/n(1 + e2iφ )

Er (1 − 1/n2 )(eiφ − e−iφ ) −(1 − 1/n2 )(1 − e2iφ )

= =
Ei (1 + 1/n)2 e−iφ − (1 − 1/n)2 eiφ (1 + 1/n2 )(1 − e−2iφ ) + 2/n(1 + e2iφ )

In fact, this problem can easily be generalized for incidence at an arbitrary angle i by taking
1/n → cos i/n cos r.We now take the limit where this is an excellent conductor,σ/ω0 >> 1.
In this case, the index of refraction (28) and phase change (29) may be approximated by
r r
σ σ 2
n≈ i = (1 + i) =
ω0 20 ω γ

r r
ωD ωD σ µ0 σω
φ= n ≈ (1 + i) n = (1 + i)D = iλ
c c 20 ω 2

For |γ| << 1 (equivalent to |n| >> 1) we drop terms of O(1/n2 ) compared to 1 to arrive at

Et 2γe−λ
= (30)
Ei (1 − e−2λ ) + γ(1 + e−2λ )
Er −(1 − e−2λ )
= (31)
Ei (1 − e−2λ ) + γ(1 + e−2λ )

where we have kept the O(γ) term in the denominator which becomes important in the
limit λ → 0.
b) Verify that for zero thickness and infinite thickness you obtain the proper limiting results.
Solution
The zero thickness limit corresponds to λ → 0. In this case, it is easy to see from the (30)
and (31) that

Et Er
λ→0: → 1, →0
Ei Ei
In the infinite thickness limit, we find instead
Et Er −1
λ→∞: → 0, →
Ei Ei 1+γ
Note that the reflection coefficient does not go to unity, as some of the power is dissipated
in the conductor. A perfect conductor (σ = ∞) has γ = 0, so all the power is reflected in
the perfect conductor limit.
c) Show that, except for sheets of very small thickness, the transmission coefficient is

8(Reγ)2 e−2D/δ
T =
1 − 2e−2D/δ cos(2D/δ) + e−4D/δ

Solution
To compute the transmission coefficient from (30) and(31) , we kept in mind that both γ
 6

and λ are complex. As long as we are not in the “very small thickness” limit, the O(γ) term
in the denominator can be ignored. in this case
Et 2γe−λ
≈
Ei (1 − e−2λ )
so that
 2
 Et  4|γ|2 e−2Reλ
T =  =
 
Ei 1 − 2Re(e−2λ ) + e−4Reλ
−2D
Taking |γ|2 = 2(Reγ)2 e δ then gives

8|Reγ|2 e−2D/δ
T =
1 − 2(e−2D/δ cos(2D/δ) + e−4D/δ
The very small thickness limit corresponds to when the O(γ) term becomes important. This
occurs when

|1 − e−2λ | ∼ |γ(1 + e−2λ )|

Expanding for small λ yields

D ωδ
|2λ| ∼ |2γ| ⇒ ∼
δ c
Hence small thickness corresponds to
ωδ 2
D<
c
3. (7.6). A plane wave of frequency is incident normally from vacuum on a semi-infinite slab of
material with a complex index of refraction n(ω)[n2 (ω) = (ω)/0 ].

a) Show that the ratio of reflected power to incident power is

 1 − n(ω) 2
 
 
 1 + n(ω) 

while the ratio of power transmitted into the medium to the incident power is
4Ren(ω)
T =
|1 + n(ω)|
Solution
While this problem involves a complex dielectric constant (ω), we note that the matching
conditions for incident and reflected waves at an interface hold for arbitrary (including
complex) values of µ and . For normal incidence, the expressions are simply
Er 1 − n(ω) Er 2
= , =
Ei 1 + n(ω) Ei 1 + n(ω)
where we have furthermore assumed that the material is non-permeable so that µ = µ0 . For
harmonic waves, the power is obtained from the real part of the Poynting vector
r ∗
~= E1 ~ ×H ~ =
∗ 1  ~ 2
S |E| n̂ (32)
2 2 µ∗
 7

The reflection coefficient is then straightforward

 2  2
Re(n̂.S~r )  Er   1 − n(ω) 
R= =  
Re(n̂.S~i ) Ei 1 + n(ω) 
For the transmission coefficient, we also have to account for the different material
s  2
~
Re(n̂.St ) (ω)∗  Et  4Re[n(ω)]
T = = Re =
~
Re(n̂.Si ) 0 Ei  |1 + n(ω)|2

~ D
b) Evaluate Re[iω(E. ~ ∗ − B.
~ H
~ ∗ )/2] as a function of (x, y, z). Show that this rate of change of
energy per unit volume accounts for the relative transmitted power T .
Solution
We write the electric and magnetic fields inside the material as
E~ =E ~ = √0 µ0 n(ω)n̂ × E
~ t eikn̂.~x , B ~ t eikn̂.~x

where the complex wavenumber k is given by

ωn(ω)
k(ω) =
c
In this case, the power per unit volume expression becomes
    
iω ~ ~ ∗ ~ ~ ∗ iω ∗ ~ 2 1 ~ 2
Re (E.D − B.H ) = Re (iω) |E| − |B| (33)
2 2 µ0
 
iω ∗ 2 ~ 2 −2Im[k(ω)]n̂.~
x
= Re ((ω) − 0 |n(ω)| )|E| e ) (34)
2
 
i0 ω 2∗ 2 ~ 2 −2Im[k(ω)]n̂.~x
= Re (n(ω) − |n(ω)| )|E| e ) (35)
2
0 ωn(ω)2 ~ 2 −2Im[k(ω)]n̂.~x
= |E| e (36)
2
= 0 ωRen(ω)k(ω)|E| ~ 2 e−2Im[k(ω)]n̂.~x (37)
r
0 ~ 2 e−2Im[k(ω)]n̂.~x
= Re[n(ω)]k(ω)|E| (38)
µ0
The power per area transmitted into the material may then be calculated as
Z ∞  
Pt iω ~ ~ ∗ ~ ~ ∗
= Re (E.D + B.H ) dz
A 0 2
r Z ∞
0 ~
= Re[n(ω)]Im[k(ω)]|E|2
e−2[k(ω)]n̂.~x )dz
µ0 0
r
1 0 ~ 2
= Re[n(ω)]|E|
2 µ0
On the other hand, the incident power per area may be obtained from (32)
r
Pi 1 0 ~ 2
= |E| (39)
A 2 µ0
This gives
 2
Pt  Et  4
= Re[n(ω)]   = Re[n(ω)] ×
Pi Ei |1 + n(ω)|2
 8

which agrees with the above calculation of the transmission coefficient.

Note that the complex Poynting vector inside the material is
r
~= 1 0 ~ 2 e−2Im[k(ω)]n̂.~x n̂
S n(ω)∗ |E|
2 µ0
Hence
r
Re(∇. ~ =1
~ S) 0 ~ 2 Im[k(ω)]e−2[k(ω)]n̂.~x
Re[n(ω)]|E| (40)
2 µ0
Comparing this with Poynting equations derived above demonstrates that the real part of
the complex Poynting’s theorem holds
~ S
∇. ~ = iω (E.
~ D~ ∗ − B.
~ H~ ∗ ) + 1 J~∗ E
~ =0
2 2
so long as we take J~ = 0 (i.e no free currents).
c) Solution For a conductor,
4πσ
n2 = 1 + i
ω
and the penetration depth
c
n2 = √
2πωσ
If ω << σ (assumed real), then n2 ≈ i 4πσ
ω

√ √
1 + i 4πσ 2πωσ c
∴n≈( √ ) √ = (1 + i) = (1 + i)( )
2 ω ω ωσ
Then
 c 2 c 2 c
|1 + n|2 = 1 + (1 + i)  = (1 + ) + ( )2

ωσ ωσ ωσ
c 2 c
|1 − n|2 = (1 − ) + ( )2
ωσ ωσ
c 2 c 2
(1 − ωσ
) + ( ωσ ) 4 ωσ
c
R= c 2 c 2 = 1 −
(1 + ωσ
) + ( ωσ ) 1 + (1 + 4ωσ
c
)2

4c
ωσ
4 ωσ
c
T = c 2 c 2 =
(1 + ωσ
) + ( ωσ ) 1 + (1 + 4ωσ
c
)2
ωσ 2ωσ
If c
<< 1, R ≈ 1 − c
T = 2 ωσ
c
−2z
2 −2z
1 ~ = Re σ|E|2 eImnz ωc = σ(E0 4e ) = ω E02 e σ
Re[(J~∗ .E)]
2 2 2|1 + n|2 π [1 + (1 + ωσc
)2 ]
We wish to compare this with
iω ~ ~ ∗ ~ ~ ∗ ω Re[n]Im[n] −2Im[n] ω z ωE02 e−2z/σ
Re[ (E.D − B.H )] = E02 e c =
8π π |1 + n|2 π[1 + ( ωσ
c
)2 ]
These two terms are equal and thus do not enter the complex form of Poynting’s theorem.
The complex dielectric constant can be taken to contain the conduction current contribution
by means of free electron contribution with zero binding frequency.
 9

4. (7.7). A ribbon beam of plane-polarized radiation of wavelength λ is totally reflected internally

at a plane boundary between two nonpermeable media with indices of refraction n and n0 (n0 <
n).

(a) Show that for a “monochromatic” ribbon beam of radiation in the z direction with an
electric field amplitude, E(x)e(ikz−iωt) , where E(x) is the smooth and finite in transverse
extent (but many wavelengths broad) the lower order approximation in terms of plane wave
is
Z
~
E(x, y, t) = ~ dκA(κ)eiκx+ikx−iωt

where ~ is a polarization vector, and A(κ) is the Fourier transform of E(x), with support
in κ around κ = 0 small compared to k. The finite beam consists of planes waves with a
small range of angles of incidence, centered around the geometrical optics value.
Solution
~κ = (∆κ, κ0 ), exact ~κ, ~κ0 = (0, κ0 ), central ~κ.
Z
d∆κ
E(x) = A(∆κ)ei∆κx
2π

ω
q
(∆κ)2 + κ20 =
c
The exact field in the Fourier space:
~ κ, t) = ~(κ)E0 ei∆κx+iκ0 z−iωt A(∆κ)
E(~

where

~κ.~(κ) = 0

Using
q
κ = (∆κ)2 + κ20 = κ0 and ~κ0 .~0 = 0, ~(κ) ≈ ~0

We have:
Exact
Z
~ d∆κ
E(x, z, t) = ~(κ)A(∆κ)ei∆κx+iκ0 z−iωt
2π
Z
~ d∆κ
E(x, z, t) ≈ ~0 A(∆κ)ei∆κx+iκz−iωt .
2π

(b) Consider the reflected beam and show that for i > i0 the electric field can be expressed
approximately as
~ 00 − δx)exp[ik~00 .~x − iωt + iφ(i, i0 )]
E~00 (x, y, z) = ~00 E(x

where ~00 is a polarization vector, x00 is the x coordinate perpendicular to k~00 , the reflected
wave vector, and δx = −(1/k)[dφ(i, i0 )/di].
Solution
Now consider the reflection of the wave packet. We derived that

Er (κ) = eiφ(θi ) EI (κ); κr = (κix , −κiz )

 10

So
Z
~ d∆κ
E(x, z, t) = ~r (κ~r )A(∆κ)eiφ(θi ) ei∆~κ.~x .
2π
Here φ(θi ) a function of κ~i . From this point of view, we see that changing κ0 → κ + ∆κ
changes the angle: δγ = δκ κ
. Upon reflection ~κr is related to κi

~κi = (~κ0 + ∆~κ)i = (κ sin(θi + δθi , κ cos(θi + δθi )

~κr = (κ sin(θi + δθi , −κ cos(θi + δθi )

= (κ sin(θi , − cos θi ) + κδθ(cos θi , sin thetai )
= (~κ0r − ∆κ~0r

∆κ is conjugate to a shift in the −x, direction.

The reflected wave is
Z
~ d∆κ
E(x, 0
z, t) = ~r A(∆κ)e−i∆κx eiκz−iωt eiφ(θi ) .
2π
So expanding
dφ
φ(θi ) = φ0 + δθi
dθi
dφ ∆k
= φ0 − .
dθi k
So Fourier transforming
Z
~r = d∆κ −i dφ ∆k
E ~0r A(∆κ)e−i∆κx+iκr z e−iφ0 e dθi k
2π
= ~r E(−(x − δx))eiκr z e−iφ0
0

So
dφ 1
δx = − (41)
dθi κ

(c) With the Frensnel expressions for the phases φ(i, i0 ) for the two states of plane polarization,
show that the lateral displacements of reflected beams with respect to the geometric optics
position are
λ sin i
D⊥ =
π (sin i − sin2 i0 )1/2
2

and
sin2 i0
Dk = D⊥ . .
[sin2 i − cos2 i. sin2 i0 ]
The displacement is known as the Goos-Hanchen effect.
Solution

nCT
φ0 (i) = −2 tan−1
Ci
 11
p
φ 1 sin2 i − sin2 i0 y
− tan (i) = ≡
sin2 i0
p
2 2
1 − sin i x

p !
2 2
φ 1 dφ d 1 sin i − sin i0
− sec2 = 2
p
2 2 di di sin i0 1 − sin2 i

dφ d y
= −2 cos2 (φ/2) ( ).
di di x

1 dφ 2 sin i sin2 i0
− =
κ di κ (sin2 i − sin2 i0 )1/2 (sin2 i − cos2 ı sin2 i0 )
λ sin i sin2 i0
= .
π (sin2 i − sin2 i0 )1/2 (sin2 i − cos2 ı sin2 i0 )

Therefore,
λ sin i
D⊥ =
π (sin i − sin2 i0 )1/2
2

and
sin2 i0
Dk = D⊥ . .
[sin2 i − cos2 i. sin2 i0 ]