Solutions to Problems in Electromagnetics, Vol.

1
Version 1.1

Steven W. Ellingson
ellingson.1@vt.edu
Virginia Tech

August 10, 2018

This manual accompanies Electromagnetics Vol. 1, an open textbook freely available at
https://doi.org/10.21061/electromagnetics-vol-1

c 2018 Steven W. Ellingson CC BY SA 4.0.

https://creativecommons.org/licenses/by-sa/4.0/

1
Contents

2 Electric and Magnetic Fields 3

3 Transmission Lines 6

5 Electrostatics 34

6 Steady Current and Conductivity 69

7 Magnetostatics 76

8 Time-Varying Fields 91

9 Plane Wave Propagation in Lossless Media 103

2
Chapter 2

Electric and Magnetic Fields

3
[m0002] [1]

2.2-1

From the problem statement, V = 1.5 V and d = 30 µm, so

V
|E| ≈ = 50 kV/m
d

4
[m0011] [1]

2.4-1

From the problem statement, V = 12 V, ǫr = 6, and d = 90 µm, so the electric field

intensity is
V ∼
|E| ≈ = 133 kV/m
d
Subsequently, the electric flux density is

|D| = ǫr ǫ0 |E| = 7.08 µC/m2

5
Chapter 3

Transmission Lines

6
[m0027] [1]

3.6-1

(a) The expression for the voltage Ve (z) traveling in the +z direction contains the factor
e−γz . The propagation constant γ = α + jβ, where α and β are real-valued constants.
Therefore, the ratio of the voltage at a distance l from some other point on the transmission
line is:
Ṽ (z + l) e−γ(z+l)
= −γz = e−γl = e−αl e−jβl
Ṽ (z) e
The magnitude of this difference is just the first factor; i.e., e−αl . We also know that
p
γ = (R′ + jωL′ ) (G′ + jωC ′ )

At 100 MHz, we find γ = 0.00850 + j3.14468 m−1 . Therefore, α = 0.00850 m−1 , and the
voltage after 1 m is


(1 V) exp − 0.00850 m−1 (1 m) = 0.9915 V

(b) From part (a) we know the phase of this difference is just the phase of the factor e−jβl .
Since β = 3.14468 rad/m, the phase of e−jβl is 179.8◦ for l = 1 m.

(c) For a radio wave in free space, there should be essentially no attenuation over 1 m, as
long as this 1 m span is located far from the transmitter. This is because free space propa-
gation contains no loss mechanisms analogous to R′ or G′ in the transmission line model. At
f = 100 MHz, the wavelength of the radio wave is λ = c/f ∼ = 3 m. That means the phase
rotates by 360◦ in 3 m, which is 120◦ in 1 m. Note that the wavelength of the radio wave
is significantly longer than the wavelength of the signal in the transmission line.

7
[m0027] [2]

3.6-2

The question is whether

Ve (z) = V0+ e−γz + V0− e+γz
is a solution to the TEM transmission line wave equation

∂2 e
2
V (z) − γ 2 Ve (z) = 0
∂z
where V0+ , V0− , and γ are complex-valued constants. To determine this, we substitute the
candidate solution into the equation and determine if equality holds. Taking the first deriva-
tive of the candidate solution:
∂ e
V (z) = −γV0+ e−γz + γV0− e+γz
∂z
Repeating to get the second derivative:

∂2 e
2
V (z) = +γ 2 V0+ e−γz + γ 2 V0− e+γz
∂z
Now making the substitutions into the left side of the wave equation:
 2 + −γz   
+γ V0 e + γ 2 V0− e+γz − γ 2 V0+ e−γz + V0− e+γz

= +γ 2 V0+ e−γz + γ 2 V0− e+γz − γ 2 V0+ e−γz − γ 2 V0− e+γz

=0
which is the the right hand side of the wave equation, as expected.

8
[m0052] [1]

3.7-1

It is true that the real part of the characteristic impedance must be positive.

Here’s a mathematical argument: Recall:

s
R′ + jωL′
Z0 =
G′ + jωC ′

Also note that R′ , L′ , G′ and C ′ must all be positive or zero. Therefore, the numerator
and denominator of the above expression, before taking the square root, must both have
phase in the range 0 to +π/2. This means the numerator divided by the denominator, again
before taking the square root, must have have phase in the range −π/2 to +π/2. Taking
the principal square root reduces the phase by a factor of two, the phase of Z0 must be in
the range −π/4 to +π/4. Subsequently, the real part of Z0 must be positive.

Here’s a physical argument: A positive-valued real-valued component of an impedance rep-

resents the dissipation of power (e.g., resistance) or transfer of power out of a system (e.g.,
to a load). Conversely, a negative-valued real-valued component of an impedance represents
the creation of power or the introduction of power into a system; in other words, an active
device. Since the concept of characteristic impedance applies to transmission lines, and since
transmission lines are passive devices, the real component of the characteristic impedance
must be positive.

9
[m0080] [1]

3.8-1

The physical current:

i(z, t) = (2 A) sin((3 rad/s)t + (4 rad/m)z + 5 rad)

= (2 A) cos((3 rad/s)t + (4 rad/m)z + 5 rad − π/2)

so h i
j ((5−π/2) rad)
˜
I(z) = (2 A) e ej(4 rad/m)z

10
[m0080] [2]

3.8-2

Converting to time-domain representation:

n o 
e
v(x, t) = Re V (x)e jωt
= Re V0 e+jβx ejωt

The problems statement implies V0 is complex-valued. To accomodate this, we define the

magnitude and phase of V0 as follows:

V0 , |V0 | ejπ/3

Then:  
v(x, t) = Re |V0 | ejπ/3 e+jβx ejωt = |V0 | Re ej(ωt+βx+π/3)
Finally, using the identity ejθ = cos θ + j sin θ, we obtain

v(x, t) = |V0 | cos (ωt + βx + π/3)

This wave is traveling in the −x direction.

11
[m0080] [3]

3.8-3

The form given in the problem statement is a phasor, describing a wave traving in the
+φ direction. To obtain the time domain form:

v(φ, t) = Re V0 e−jβφ ejωt = |V0 | cos (ωt − βφ + ψ)

where ψ, the phase of V0 , is not given. Note

2π ∼
β= = 62.832 rad/m
λ
Furthermore, we are told that v(φ = λ/4, t = 0) is a maximum, so:
   π 
λ
|V0 | cos −β + ψ = |V0 | cos − + ψ
4 2

is a maximum, which means ψ = +π/2. Therefore:

 π
v(φ, t) = |V0 | cos ωt − [62.832 rad/m] φ +
2

The problem statement does not provide sufficient information to determine |V0 | or ω.

12
[m0083] [1]

3.9-1

From the problem statement, Z0 = 72 Ω, L′ = 0.5 µH/m, f = p 80 MHz, and the low-loss
approximations apply. Using the low-loss approximation Z0 ≈ L′ /C ′ :

L′ ∼
′
C ≈ 2 = 96.4 pF/m
Z0

Subsequently, the phase velocity is

1 ∼
vp ≈ √
′ ′
= 1.44 × 108 m/s
LC
and the phase propagation constant is
√ √
β ≈ ω L′ C ′ = 2πf L′ C ′ ∼
= 3.49 rad/m

13
[m0143] [1]

3.10-1

The characteristic impedance Z0 of coaxial cable, assuming the low-loss assumptions ap-
ply, is
60 Ω b
Z0 ≈ √ ln
ǫr a
where ǫr is the relative permittivity of the spacer material, and a and b are the radii of the
inner and outer conductors, respectively. Air has ǫr ≈ 1 and is lossless to a very good ap-
proximation. We are also told the resistance of the inner and outer conductors is negligible.
Therefore, the low-loss assumptions apply, and we are justified in using the above expression.

One way to reduce Z0 from 90 Ω to 62 Ω is to replace the air spacer with a material spacer
having  2
90 Ω ∼
ǫr = = 2.11
62 Ω
Thus, one solution is to replace air with a low-loss material having ǫr ∼
= 2.11. Another way
is to reduce b/a. For the 90 Ω cable, we determine that
 
b 90 Ω ∼
≈ exp √ = 4.48
a (60 Ω) / 1

To reduce Z0 to 62 Ω, we require
 
b 62 Ω ∼
≈ exp √ = 2.81
a (60 Ω) 1

Thus, a second solution is to keep air as the spacer material but reduce b/a to 2.81.

14
[m0084] [1]

3.12-1

The voltage reflection coefficient is

ZL − Z0 500 Ω − 75 Ω ∼
Γ= = = +0.739
ZL + Z0 500 Ω + 75 Ω
Therefore, the peak voltage of the reflected wave at the antenna input is

(0.739) (30 V) ∼
= 22.2 V

The line is lossless, so there is no attenuation of the reflected wave along the return trip from
antenna to transmitter. Therefore, the peak voltage of the reflected wave at the output of
the transmitter is 22.2 V.

15
[m0084] [2]

3.12-2

From the problem statement, Ve0+ has magnitude 7 mV and phase 180◦ , so Ve0+ = −7 mV.
Also from the problem statement Z0 = 60 Ω and ZL = 20 Ω. Therefore,
ZL − Z0
Γ= = −0.5
ZL + Z0

Subsequently, Ve0− = ΓVe0+ = +3.5 mV. Thus, the magnitude of the reflected wave is 3.5 mV,
and the phase is 0◦ .

16
[m0084] [3]

3.12-3

The voltage reflection coefficient is

ZL − Z0 33 Ω − 140 Ω ∼
Γ= = = −0.6185
ZL + Z0 33 Ω + 140 Ω
Therefore, the magnitude of the reflected voltage wave is

|Γ (3 V)| ∼
= 1.86 V

and the phase of the reflected voltage wave is

170◦ + 180◦ → −10◦

17
[m0086] [1]

3.13-1

(a) On short-circuited line, Γ = 1. Therefore, the current at a voltage maximum is zero.

(b) The voltage at a short circuit is zero. The distance between voltage extrema is λ/4, so
λ/4 = 8 cm. The distance between voltage maxima is λ/2 = 16 cm. Therefore, the distance
between the short circuit and the second voltage maximum is 8 + 16 = 24 cm.

18
[m0081] [1]

3.14-1

First note
SWR − 1
|Γ| =
SWR + 1
So in this case
1.2 − 1 ∼
|Γ| ≤ = 0.091
1.2 + 1
Also note:
ZL − Z0
Γ=
ZL + Z0
where in this case Z0 = 50 Ω and ZL is the input impedance of the amplifier. Solving for ZL
we find:
1+Γ
ZL = Z0
1−Γ
Since the imaginary component of Z0 is zero, and since the imaginary component of ZL is
negligible, Γ must be real-valued. Therefore, −0.091 ≤ Γ ≤ +0.091 and

41.7 Ω ≤ ZL ≤ 60.0 Ω

19
[m0081] [2]

3.14-2

From the problem statement, Z0 = 72 Ω and ZL = 60 Ω. Therefore, the voltage reflec-

tion coefficient is
ZL − Z0 ∼
Γ= = −0.091
ZL + Z0
and the standing wave ratio is

1 + |Γ|
SWR = = 1.2
1 − |Γ|

20
[m0087] [1]

3.15-1

The input impedance of a lossless line is periodic in length, with period λ/2. Therefore,
the line is exactly 3 periods long, which means the input impedance is equal to the load
impedance 72 + j42 Ω.

21
[m0088] [1]

3.16-1

In this case, the input impedance is

Zstub = −jZ0 cot βl

where Z0 = 75 Ω, l = 13 cm, and

ω 2πf
β= =
vp 0.55c
where f = 900 MHz. Therefore, β ∼
= 34.3 rad/m, βl ∼
= 4.45 rad, and Zstub ∼
= −j19.7 Ω .

22
[m0088] [2]

3.16-2

From the problem statement: Z0 = 75 Ω, f = 1.5 GHz, Zin = +j300 Ω is desired, and
vp = 0.6c. Note that for a short circuit, in this case:

Zin = +jZ0 tan βl = +j300 Ω

so
βl ∼
= 1.3258 rad
Note
ω 2πf ∼
β= = = 52.36 rad/m
vp 0.6c
so l ∼
= 2.53 cm

23
[m0145] [1]

3.17-1

(a) For a bandpass response centered at 200 MHz, you want the input impedance into
the stub, which is attached in parallel to the line, to be an open circuit at 200 MHz. This is
accomplished using a short-circuited stub which is one quarter wavelength long at 200 MHz.
A wavelength in the transmission line is

0.67 (3 × 108 m/s)

λ= = 1.005 m (3.1)
200 MHz

so the stub length is 25.12 cm .

(b) See below:

24
[m0145] [2]

3.17-2

The smallest length for which the imaginary component of the impedance of an open-
circuited stub is positive is slightly greater than λ/4. The imaginary part of the impedance
remains positive until the length is slightly less than λ/2. In this transmission line,

ω 2πf
β= = = 89.8 rad/m
vp 0.7c

where f = 3 GHz. Therefore,

2π
λ= = 7 cm
β
and so the smallest contiguous range of transmission line length l is

1.75 cm < l < 3.5 cm

25
[m0145] [3]

3.17-3

For zero response centered at fc = 1.3 GHz you want the input impedance into the stub,
which is attached in parallel to the line, to be a short circuit at f = fc . This is accomplished
using an open-circuited stub that is one quarter wavelength long at f = fc . A wavelength
in the transmission line is
0.6 (3 × 108 m/s)
λ= = 13.84 cm (3.2)
1.3 GHz

so the stub length is 3.46 cm . The characteristic impedance is irrelevant.

26
[m0091] [1]

3.19-1

From the problem statement, we see that the design will consist of a quarter-wave matching
section followed by a line having a characteristic impedance Z02 = 300 Ω – i.e., equal to the
load impedance – and the total length will
√ be l = 5 cm. The characteristic impedance of
the quarter-wave section must be Z01 = ZS Z02 , where ZS is the source output impedance;
thus, we have Z01 = 122.5 Ω. The length l1 of the quarter-wave section is λ/4, where λ is
the wavelength in the transmission line. For FR4, we have
λ0 c/f (3 × 108 m/s) / (1.5 GHz)
λ= √ =q = q = 12.06 cm
ǫr,ef f 1
(ǫr + 1) 1
(4.5 + 1)
2 2

so l1 = 3.01 cm and l2 = l − l1 = 1.99 cm.

What’s left to figure out is the width w of the microstrip lines, which determines the charac-
teristic impedance since h = 1.6 mm and ǫr = 4.5 are already set. We know that h/w = 1/2
gives a characteristic impedance of 50Ω for FR4, so the width of a 50Ω line is 2h = 3.2 mm.
To get the higher characteristic impedance Z01 = 122.5Ω, w1 will have be smaller than
3.2 mm. An approximate but reasonable solution is simply to assume the characteristic
impedance scales with w in the same way (i.e., linearly) as it does in the “wide” (h/w ≪ 1)
case, so
50 Ω
w1 ≈ (3.2 mm) = 1.3 mm
122.5 Ω
and
50 Ω
w2 ≈ (3.2 mm) = 0.5 mm
300 Ω
You could also use the Wheeler (1977) formula or some other equation or reference; how-
ever, the increased accuracy is typically irrelevant in practice due to issues such as the large
variation in ǫr due to manufacturing issues. So, while it’s not wrong to take that approach,
it’s usually not worth the effort if you are able to instead “scale” from a known design as we
have done above.

So, your sketch should show the source, followed by 3.01 cm of line which is 1.3 mm wide,
followed by 1.09 cm of line which is 0.5 mm wide, followed by the load, as shown in the figure.

27
[m0091] [2]

3.19-2

From the problem statement: ZL = 200 Ω, l = λ/4, and Z0 = 100 Ω. Since this is a
quarter-wave line,
Z2
Zin = 0 = 50 Ω
ZL
.

28
[m0091] [3]

3.19-3

The problem statement implies that each of the stubs is short-circuited at the end opposite
the main line. For this to be a bandpass filter, the magnitude of the input impedance looking
into each stub must be very high – nominally infinite – since then the filter structure would
be in effect the main line by itself, with no stubs, and would therefore be well-matched at the
filter input and output. At any higher or lower frequency the magnitude of the stubs’ input
impedance can only be less; therefore, the input impedance of the filter would be increasingly
mismatched. This results in bandpass response.

The shortest length for which the magnitude of the input impedance of a short-circuited
transmission line is infinite is λ/4. Therefore, λ/4 = 3.38 mm and subsequently λ =
13.52 mm in the stub. Therefore, the center frequency is
vp 0.6c
f= = = 13.3 GHz
λ λ

29
[m0090] [1]

3.20-1

+ +
Summarizing the problem statement: Pav = 5 W and PL = 4.6 W. Therefore, PL /Pav = 0.92.
From this, we may find the magnitude of the reflection coefficient, |Γ|, using

PL
+
= 1 − |Γ|2
Pav

We find |Γ| ∼
= 0.283 and
1 + |Γ| ∼
SWR = = 1.79
1 − |Γ|

30
[m0090] [2]

3.20-2

+
From the problem statement, Γ = 0.3 + j0.4 and Pav = 3 W. Therefore,

+
PL = 1 − |Γ|2 Pav = 2.25 W

31
[m0094] [1]

3.23-1

From the problem statement, we have f = 220 MHz, antenna impedance ZA = 73 + j42 Ω,
and characteristic impedance Z0 = 50 Ω for both the transmission line and the stub. The
input impedance looking into a length d1 of transmission line terminated in impedance ZA
is
1 + Γe−j2βd1
Z1 (βd1 ) = Z0 (3.3)
1 − Γe−j2βd1
where
ZA − Z0
Γ= = 0.2719 + j0.2486 (3.4)
ZA + Z0
The first task is to find the smallest βd1 such that the real part of Y1 (βd1 ) = Z1−1 (βd1 ) equals
Y0 = Z0−1 = 0.02 Ω−1 . After a few minutes of trial and error one finds:

Y1 (βd1 = 1.345 rad) = 0.0200 + j0.0159 Ω−1 (3.5)

(You could also do this with a Smith chart if you are so inclined.) The match is accomplished
by attaching a stub having input admittance Y2 = −j0.0159 Ω−1 in parallel with Y1 , since
then the combined admittance will be Y1 + Y2 = Y0 = Z0−1 . For a short-circuited stub of
length d2 we would want:

Y2 = −jY0 cot βd2 = −j0.0159 Ω−1 ⇒ βd2 = 0.900 rad (3.6)

For an open-circuited stub of length d2 we would want:

Y2 = +jY0 tan βd2 = −j0.0159 Ω−1 ⇒ βd2 = 2.471 rad (3.7)

The short-circuited stub is shorter, so that’s the preferred solution. All that remains is to
figure out the physical lengths from the electrical lengths. For this, we need to know β. The
phase velocity is vp = 0.67c, so

2π 2π
β= = = 6.8771 rad/m (3.8)
λ 0.67c/f

Finally we have the solution:

βd1 1.345 rad
d1 = = = 19.6 cm distance from antenna terminals to stub (3.9)
β 6.877 rad/m

βd2 0.900 rad

d2 = = = 13.1 cm stub length (3.10)
β 6.877 rad/m
and the stub is short-circuited .

32
[m0094] [2]

3.23-2

In terms of the variables used in the book, the problems statement is indicating that
Y1 = 0.0128 − j0.0040 Ω−1 and that Zin is real-valued. Therefore, Yin = 1/Zin is real-valued,
and must be equal to the real part of Y1 ; i.e., Yin = 0.0128 Ω−1 . Therefore, Zin ∼
= 78.1 Ω ,
which is the answer to part (a).

The stub is being used to cancel the imaginary part of Y1 , so Ystub = +j0.0040 Ω−1 and
subsequently Zstub = −j250 Ω , which is the answer to part (b).

33
Chapter 5

Electrostatics

34
[m0102] [1]

5.1-1

The electric field due to a point charge q is

q
E(R) = R̂
4πǫR2
where R is the position-free vector pointing from the charge to the field point. From the
problem statement, q = −24 nC, ǫr = 2, and

R = x̂ + ŷ2 + ẑ3 m

Thus √
R , |R| = 12 + 22 + 32 ∼
= 3.74 m
R∼
R̂ , = x̂0.267 + ŷ0.534 + ẑ0.802
R
ǫ = ǫr ǫ0 = 2 · 8.854 × 10−12 F/m
Thus
E(R) ∼
= −x̂2.06 − ŷ4.12 − ẑ6.18 V/m

35
[m0103] [1]

5.2-1

From the problem statement,

q1 = +3 nC at r1 = −ẑd and
q2 = +3 nC at r2 = +ẑd where d = 0.5 m;
the field point of interest is r = +x̂x where x = +1.5 m, and
ǫ = ǫ0 ∼
= 8.854 × 10−12 F/m.

The electric field intensity due to two point charges is

 
1 r − r1 r − r2
E(r) = q1 + q2
4πǫ |r − r1 |3 |r − r2 |3

In this problem:
r − r1 = +x̂x − (−ẑd) = +x̂x + ẑd
√
|r − r1 | = x2 + d2
r − r2 = +x̂x − (+ẑd) = +x̂x − ẑd
√
|r − r2 | = x2 + d2
Substituting: " #
1 +x̂x + ẑd +x̂x − ẑd
E(+x̂x) = q+ q
4πǫ0 (x2 + d2 )3/2 (x2 + d2 )3/2
where we have made the definition q , q1 = q2 . Note that the ẑ-directed components
cancel, as expected from the symmetry of the problem. Eliminating these components and
simplifying:
q x
E(+x̂x) = x̂
2πǫ0 (x2 + d2 )3/2
Now take a moment to confirm that the solution is dimensionally-correct and makes physical
sense. Finally, substituting values, we obtain:

E(+x̂1.5 m) ∼
= +x̂ (20.5 V/m)

For a single charge q0 at the origin to create this field, we require

q0
x̂ = +x̂ (20.5 V/m)
4πǫ0 x2

which yields q0 ∼
= +5.12 nC .

36
[m0100] [1]

5.3-1

From the problem statement, ρv = Kr−2 where K = 2 C/m. From dimensional analysis,
it is clear that this is a volume charge density. We seek the total charge Q in a volumetric
region V bounded by the constant-coordinate surfaces r = a and r = b where a = 1 m and
b = 2 m. In general, Z
Q= ρv dv
V
In this case, using spherical coordinates:
Z b Z π Z 2π  
K 2


Q= r sin θ dr dθ dφ
r=a θ=0 φ=0 r2

Factoring into separate integrals:

Z b  Z π  Z 2π 
Q=K dr sin θ dθ dφ
r=a θ=0 φ=0

Evaluating the integrals:

Q = K [b − a] [2] [2π] = 4πK (b − a)

This a good point at which to check for dimensional consistency (i.e., correct units).

Using the given values of K, a, and b we obtain Q ∼

= 25.1 C .

37
[m0104] [1]

5.4-1

Interpreting the problem statement:

ρs,1 , +4 nC/m2 for sheet in the x = 0 plane. Let the field from this sheet be E1 .
ρs,2 , +16 nC/m2 for sheet in the y = 0 plane. Let the field from this sheet be E2 .
ρs,3 , +64 nC/m2 for sheet in the z = 0 plane. Let the field from this sheet be E3 .
Also, ǫ = 2ǫ0 .

The electric field intensity due to a single sheet of charge having charge density ρs in the
z = 0 plane is worked out in the book. It is:
ρs
+ẑ sgnz
2ǫ
This corresponds to the third sheet of charge above. Since the region of interest is z > 0:
ρs,3
E3 = +ẑ
4ǫ0

Similarly,
ρs,1
E1 = +x̂
4ǫ0
ρs,2
E2 = +ŷ
4ǫ0
The total field is the sum of these three fields. Thus:
ρs,1 ρs,2 ρs,3
E = x̂ + ŷ + ẑ
4ǫ0 4ǫ0 4ǫ0
Substituting values, we obtain:

E∼
= x̂ (113 V/m) + ŷ (452 V/m) + ẑ (1807 V/m)

38
[m0014] [1]

5.5-1

Note F12 = Q2 E1 where E1 is the electric field intensity associated with Q1 . Thus:
F12 Q1
E1 = = R̂12 2
Q2 4πǫR12

We can write this in terms of the electric flux density, assuming a isotropic and homogenous
medium:
Q1
D1 = ǫE1 = R̂12 2
4πR12
Now let’s put Q1 at the origin, and let S be a sphere of radius a centered at the origin.
Then, the left hand side of Gauss’ Law is:
I Z π Z 2π   Z Z 2π
Q1  2  Q1 π
D · ds = r̂ · r̂a sin θ dθ dφ = sin θ dθ dφ = Q1
S θ=0 φ=0 4πa2 4π θ=0 φ=0

Because we put Q1 at the origin and defined S to surround it, Qencl = Q1 , which is what we
expect from Gauss’ Law. Therefore, Coulomb’s Law is a solution to – a special case really
– of Gauss’ Law.

You might be inclined to object on the grounds that Gauss’ Law doesn’t say anything about
force or electric field intensity. This is true! However, electric field intensity is defined by
force; i.e., E1 = F12 /Q2 is a definition for E1 , and not derived from something else. Similarly,
D1 = ǫE1 is a definition for D1 , and not derived from something else. So, Gauss’ Law is as
fundamental as it gets.

39
[m0014] [4]

5.5-2

By symmetry, there can be no variation in the z or φ dimensions. Therefore, our answer

can depend only on ρ. The three regions to consider are inside the inner surface of the shell
(ρ < 1 m), inside the shell itself (1 ≤ ρ < 3 m), and outside the outer surface of the shell
(ρ > 3 m).

The integral form of Gauss’ Law is:

Z
D · ds = Qencl (5.1)
S

where S is any closed surface. Since we are asked for electric field intensity, we can use
D = ǫE to obtain: Z
1
E · ds = Qencl . (5.2)
S ǫ
Note that we use ǫ as opposed to ǫ0 , since the latter infers free space conditions, and we
haven’t been told that.

For ρ < 1 m, Qencl = 0; i.e., there is no surface that we can define that encloses charge.
Combined with the symmetry argument, we have that E = 0 in this region.

For 1 ≤ ρ < 3 m, Qencl depends on ρ. Combined with the symmetry argument, we have
E(ρ) = ρ̂E(ρ). Thus, a good choice for S is a cylinder centered on the z axis. This gives us:
Z 2π Z +L/2 Z ρ Z 2π Z +L/2
1
ρ̂E(ρ) · ρ̂ ρ dφ dz + 0 = ρv dρ ρdφ dz (5.3)
φ=0 z=−L/2 ǫ ρ=1 φ=0 z=−L/2

where “+0” on the left hand side is the contribution from the constant-z surfaces (the “end
caps”) of the cylinder – zero because the normal to those surfaces (ẑ) is perpendicular to E.
Now evaluating:
1
E(ρ) 2πρL = ρv πL(ρ2 − 1 m2 ) . (5.4)
ǫ
Finally:
 
ρv 1 m2
E(ρ) = ρ̂ ρ− in this region. (5.5)
2ǫ ρ
Assuming ρv is in C/m3 and ǫ is in F/m, E(ρ) will be in V/m. However, if you say simply
“1” as opposed to “1 m2 ” in the above expression, then you must indicate the units of ρ
(being meters) as well. In electromagnetics, a powerful (but unappreciated) technique for
checking your work is to make sure your solution has the right units. This is called dimen-
sional analysis. You should be able to substitute units for each of the quantities in the above
solution and find that the result has units of V/m – can you do this?

40
For ρ > 3 m, Qencl is constant at the maximum value (since all the charge has been enclosed),
but E(ρ) is still a variable function of ρ. Thus, the left hand side of Equation 5.4 remains
the same, but the right hand side is evaluated at ρ = 3 m. This yields:
1
E(ρ) 2πρL = ρv πL · 8 m2 . (5.6)
ǫ
Thus:
ρv 4 m 2
E(ρ) = ρ̂ in this region. (5.7)
ǫ ρ
Suggestion: Try dimensional analysis on this solution. Can you see why it is important to
say “4 m2 ” as opposed to just “4”?

41
[m0014] [2]

5.5-3

By symmetry, there can be no variation in the θ or φ dimensions. Therefore, our answer

can depend only on r. The three regions to consider are inside the inner surface of the shell
(r < 2 m), inside the shell itself (2 ≤ r < 4 m), and outside the outer surface of the shell
(r > 4 m).

The integral form of Gauss’ Law is:

Z
D · ds = Qenclosed (5.8)
S

where S is any closed surface. Since we are asked for electric field intensity, we can use
D = ǫE to obtain: Z
1
E · ds = Qenclosed . (5.9)
S ǫ
Note that we use ǫ as opposed to ǫ0 , since the latter infers free space conditions, and we
haven’t been told that.

For r < 2 m, Qenclosed = 0; i.e., there is no surface that we can define that encloses charge.
Therefore, E = 0 in this region.

For 2 ≤ ρ < 4 m, Qenclosed depends on r. Combined with the symmetry argument, we have
E(r) = r̂E(r). Thus, a good choice for S is a sphere centered at the origin. This gives us:
Z π Z 2π Z Z π Z 2π
2 1 r
r̂E(r) · r̂ r sin θ dθ dφ = ρv r2 sin θ dr dθ dφ (5.10)
θ=0 φ=0 ǫ r=2 θ=0 φ=0

Now evaluating:
ρv 4π 3  

4πr2 E(r) = r − 8 m3 . (5.11)
ǫ 3
Finally:
 
ρv 8 m3
E(r) = r̂ r− 2 in this region. (5.12)
3ǫ r
Assuming ρv is in units of C/m3 and ǫ is in F/m, E(ρ) will be in V/m. However, if you
say simply “8” as opposed to “8 m3 ” in the above expression, then you must indicate the
units of r (being meters) as well! In electromagnetics, a powerful (but sadly, unappreci-
ated) technique for checking your work is to make sure your solution has the right units.
This is called dimensional analysis. You should be able to substitute units for each of the
quantities in the above solution and find that the result has units of V/m – can you do this?

Also note that you have a second way to check your solution – it must be equal to the
solution for the first region for r = 2 m. Note that it is.

42
For r > 4 m, Qenclosed is constant at the maximum value (since all the charge has been
enclosed), but E(r) is still a variable function of r. Thus, the left hand side of Equation 5.11
remains the same, but the right hand side is evaluated at r = 4 m. This yields:
ρv 4π

4πr2 E(r) = 56 m3 . (5.13)
ǫ 3
Thus:
ρv 56 m3
E(r) = r̂ in this region. (5.14)
3ǫ r2
Suggestion: Try dimensional analysis on this solution. Can you see why it is important
to say “56 m3 ” as opposed to just “56”? Also, confirm that your answer agrees with the
Region 2 answer for r = 4 m.

43
[m0014] [3]

5.5-4

(a) A, B, and C have units of V/m4 , V/m3 , and V/m2 , respectively.

(b) According to the integral form of Gauss’ Law:

I I
Qencl = D · ds = ǫ0 E · ds
S S

Here, S is the surface of the box-shaped region, and ds is the normal to each of the six sides.
This integral is easiest to handle as the sum of integrals over each side, since then ds will be
constant over each of these integrals. Here we go:
Z Z y=+1 Z z=0
1
E · (−x̂ dy dz) = − A (−1) z 2 dy dz = + A
−x side y=0 z=−1 3
Z Z y=+1 Z z=0
1
E · (+x̂ dy dz) = A (+1) z 2 dy dz = + A
+x side y=0 z=−1 3
Z Z x=+1 Z z=0
E · (−ŷ dx dz) = − (−B (0) z) dx dz = 0
−y side x=−1 z=−1
Z Z x=+1 Z z=0
E · (+ŷ dx dz) = (−B (+1) z) dx dz = +B
+y side x=−1 z=−1
Z Z x=+1 Z y=+1
E · (−ẑ dx dy) = − Cx dx dy = 0
−z side x=−1 y=0
Z Z x=+1 Z y=+1
E · (+ẑ dx dy) = Cx dx dy = 0
+z side x=−1 y=0

So we find:
   
1 1 2
Qencl = ǫ0 A + A + 0 + B + 0 + 0 = ǫ0 A+B
3 3 3

having units of Coulombs if ǫ0 is in F/m and the dimensions are all in meters. The chances for
units-related confusion is reduced if consider what has actually happened in the integration
and say specifically:
  
2 5 4
Qencl = ǫ0 m A + (1 m ) B
3

44
[m0149] [1]

5.6-1

This is essentially the same problem shown as an example in the book, for which the electric
field intensity was found to be
ρl
E = ρ̂
2πǫρ
where here ρl = −2.1 mC/m and ρ is the distance from the z-axis. The electric flux density
is D = ǫE, so the permittivity doesn’t matter. The result is:

ρl ∼ 334 µC/m
D = ρ̂ = −ρ̂
2πρ ρ

45
[m0045] [1]

5.7-1

From Gauss’ Law, ρv = ∇ · D = ∇ · (ǫ0 ǫr E) = ǫ0 ǫr ∇ · E. Calculating the divergence:

∂
∂
∂

∇·E= (6 V/m2 )x + (2 V/m3 )yz + (1 V/m3 )xy
∂x ∂y ∂z

= (6 V/m2 )(1) + (2 V/m3 )(z) + (1 V/m3 )(0)

= (6 V/m2 ) + (2 V/m3 )z
ǫ0 ∼
= 8.854 × 10−12 F/m and ǫr = 4.5, so

ρv = 239.1 pC/m3 + (79.7 pC/m4 )z

46
[m0045] [2]

5.7-2

(a) ρv = ∇ · D = ǫ0 ∇ · E. Here it’s easiest to use Cartesian coordinates, for which

∂ ∂ ∂
∇ = x̂ + ŷ + ẑ
∂x ∂y ∂z
so we have
 
∂ ∂ ∂
ρv = ǫ 0 x̂ + ŷ + ẑ · [x̂ (2 V/m) sin x cos y − ŷ (2 V/m) cos x sin y]
∂x ∂y ∂z



= ǫ0 2 V/m2 cos x cos y − 2 V/m2 cos x cos y = 0
This is an example of a divergence-free field. It seems that there can be an electric field even
when there is no charge. This means simply that the source charge must lie entirely outside
the region begin considered.

(b) In this case we have

 
∂ ∂ ∂
ρv = ǫ0 x̂ + ŷ + ẑ · [x̂ (3 V/m) cos xy + ŷ (3 V/m) sin xy]
∂x ∂y ∂z


= 3 V/m3 ǫ0 [−y sin xy + x cos xy]
which has the expected units of C/m3 .

47
[m0045] [3]

5.7-3

(a) A, B, and C have units of V/m4 , V/m3 , and V/m2 , respectively.

(c) According to the differential form of Gauss’ Law, we have for this problem:
 
∂ 2 ∂ ∂  
ρv = ǫ 0 ∇ · E = ǫ 0 (Axz ) + (−Byz) + (Cx) = ǫ0 Az 2 − Bz + 0 .
∂x ∂y ∂z

This is charge density as a function of position. The enclosed charge is obtained by integrat-
ing over the region of interest:
Z Z x=+1 Z y=+1 Z z=0


Qencl = ρv dv = ǫ0 Az 2 − Bz dx dy dz
V x=−1 y=0 z=−1

The integrations over x and y factor out and are equal to 2 and 1 respectively. What’s left
is:
Z z=+1   
2

2 5
Qencl = (2)(1)ǫ0 Az − Bz dz = Qencl = ǫ0 m A + (1 m4 ) B
z=−1 3

Note that this result agrees with the result obtained using the more direct approach of us-
ing the integral form of Gauss’ Law. You should note that the reason the results are the
same is not really related to electromagnetics, but rather due to the Divergence Theorem
(of mathematics), which relates the behavior of a vector field in a volume to the behavior of
that same vector field over the enclosing surface.

48
[m0061] [1]

5.8-1

The change in the energy of the system resulting from moving the particle a small distance
l̂∆l is:
∆W ≈ −qE · l̂∆l
Power P is energy per time, so the power required to do this is:

∆W l̂∆l
P ≈ ≈ −qE ·
∆t ∆t
where ∆t is the time required for the particle to traverse the distance. Note that in the
limit as ∆t → 0, l̂∆l/∆t is the velocity v of the particle. Taking the limit and making the
substitution,
P = −qE · v
This is the “instantaneous power” required at time t and, through t, the position r(t).

Interpreting the problem statement: q = −4 mC;

E = E0 ẑ, where E0 = 3 V/m; and
r(t) = x̂a cos ωt + ŷb sin ωt + ẑct, where a = b = 2 m, ω = π rad/s, and c = 4 m/s.
Note
d
v , r(t) = −x̂aω sin ωt + ŷbω cos ωt + ẑc
dt
Therefore,

P = −qE · v = −qE0 c = − (−4 mC) (3 V/m) (4 m/s) = 48 mW

49
[m0064] [2]

5.12-1

Let us arbitrarily assume the charge is aligned along the z axis. Then the electric field
intensity is given by
ρl
E(ρ) = ρ̂ . (5.15)
2πǫρ
The potential difference is:
Z point 2 Z ρ2
ρl ρl ρ1
V21 = − E · dl̂ = − ρ̂ · ρ̂dρ = ln . (5.16)
point 1 ρ1 2πǫρ 2πǫ ρ2

50
[m0064] [1]

5.12-2

The electric field intensity resulting from a single line of uniform charge density is given
by
ρl
E(R) = R̂ . (5.17)
2πǫR
where R is the distance between the point of interest and closest point on the line, R̂ points
from that point on the line to the point of interest, and R = R̂R.

If we have two such lines of charge, then by superposition we could write:

ρl ρl
E(R) = R̂1 + R̂2 . (5.18)
2πǫR1 2πǫR2
where the subscripts “1” and “2” refer to the geometry relative to the first and second lines
of charge, respectively. In this problem, the lines of charge and the point of interest all lie
in the x − y (z = 0) plane. Since this is the case we may write simply:
ρl ρl
E(x, y) = x̂ + ŷ for z = 0. (5.19)
2πǫx 2πǫy
The potential difference is:
Z point 2
V21 = − E · dl . (5.20)
point 1
Remember that the answer should be the same for any path between the points, so you
might as well choose one that makes the problem simple. Here’s the result using one of two
equally-easy paths:
Z x2   Z y2  
ρl ρl ρl ρl
V21 = − x̂ + ŷ · x̂dx − x̂ + ŷ · ŷdy (5.21)
x=x1 2πǫx 2πǫy2 y=y1 2πǫx1 2πǫy

That is, first move from (2, 4) m to (1, 4) m along the y = y2 = 4 m line, and then move
from (1, 4) m to (1, 1) m along the x = x1 = 1 m line. Evaluating:
         
ρl x2 y2 ρl x1 y1 ρl
V21 = − ln + ln =+ ln + ln = 0.331 . (5.22)
2πǫ x1 y1 2πǫ x2 y2 ǫ

Here, you can check your results using dimensional analysis (C/m divided by F/m gives C/F
= V). You can also check that the sign is correct: Point 2 is closer to both lines of charge
than point 1, so when the charge is positive, work is being done and the potential difference
is positive. Said differently, the potential at point 2 is higher than the potential at point 1.

51
[m0063] [1]

5.14-1

To begin, it will be convenient to first convert r0 from Cartesian to spherical coordinates.

Here we go:
q
r0 = x20 + y02 + z02 ∼= 5.38 cm
θ0 = arccos (z0 /r0 ) ∼
= 42.0◦
φ0 = arctan (y0 /x0 ) ∼
= 56.3◦
(a)
V (r0 ) = V0 r02 cos θ0 ∼
= 10.8 mV
(b)  
∂ 1 ∂
E = −∇V = − r̂ + θ̂ +0 V
∂r r ∂θ
The third term in the gradient is zero because V in this problem does not vary with φ.
Continuing:
∂V 1 ∂V
E = −r̂ − θ̂ = −r̂2V0 r cos θ + θ̂V0 r sin θ
∂r r ∂θ
So:
E(r0 ) ∼
= −r̂400 + θ̂180 mV/m
(c)  
1 ∂ 2
1 ∂
ρv = ∇ · D = ǫ 0 ∇ · E = ǫ 0 2
r Er + (Eθ sin θ) + 0
r ∂r r sin θ ∂θ
where Er and Eθ are the r̂- and θ̂-directed components of E. The third term in the divergence
is zero because E in this problem does not vary with φ. Note:
1 ∂ 2
1 ∂ 2
1 ∂ 3

1 2


r E r = r (−2V 0 r cos θ) = −2V 0 r cos θ = −6V 0 r cos θ = −6V0 cos θ
r2 ∂r r2 ∂r r2 ∂r r2
and
1 ∂ 1 ∂ 1 ∂
1
(Eθ sin θ) = ((V0 r sin θ) sin θ) = V0 r sin2 θ = (2V0 r sin θ cos θ)
r sin θ ∂θ r sin θ ∂θ r sin θ ∂θ r sin θ
= 2V0 cos θ
Continuing:
ρv = ǫ0 [−6V0 cos θ + 2V0 cos θ] = −4V0 ǫ0 cos θ
(Good time for a units check...) At the point of interest:

ρv (r0 ) ∼
= −131 pC/m3

52
[m0063] [2]

5.14-2

From the problem statement,

V (r) = V0 r−1/2
where V0 , 4 V · m1/2 . So:
E(r) = −∇V (r)
∂ ∂ ∂
= −r̂ V0 r−1/2 + terms that go to zero because ∂θ
= ∂φ
=0
∂r
 
1 −3/2
= −r̂V0 − r
2
V0 −3/2
= +r̂ r
2


= +r̂ 2 V · m1/2 r−3/2

Note that the answer is dimensionally correct (and unambiguously so).

53
[m0067] [1]

5.15-1

Poisson’s Equation is
ρv
∇2 V = − . (5.23)
ǫ
The geometry of the problem suggests cartesian coordinates, and symmetry such that ∂V /∂y =
∂V /∂z = 0 is implied. Thus, the above equation becomes:
d2 ρv (x)
2
V (x) = − . (5.24)
dx ǫ
Integrating both sides with respect to x we have
Z
d 1 x
V (x) = − ρv (x) dx . (5.25)
dx ǫ −∞
Integrating both sides again with respect to x we have
Z Z x 
1 x
V (x) = − ρv (x) dx dx . (5.26)
ǫ −∞ −∞
The first chore is to take care of that sequence of integrations. We begin with mathematical
restatement of the given volume charge density:


 0 , − ∞ ≤ x < −b

−a , − b ≤ x < 0
ρv (x) = (5.27)

 +a , 0 ≤ x ≤ +b

0 , + b < x ≤ −∞
Integrating once: 
Z x 
 0 , − ∞ ≤ x < −b

−a(x + b) , − b ≤ x ≤ 0
ρv (x)dx = (5.28)
−∞ 
 +ax − ab , 0 < x ≤ +b

0 , + b < x ≤ −∞
If you have a hard time seeing this, consider sketching ρv (x) and then doing the integration
graphically. Integrating the second time:

Z x Z x  
 0 , − ∞ ≤ x < −b
 2 2
−(a/2)x − abx − ab /2 , − b ≤ x ≤ 0
ρv (x)dx dx = 2 (5.29)
−∞ −∞ 
 +(a/2)x − abx − ab2 /2 , 0 < x ≤ +b

−ab2 , + b < x ≤ −∞
Substituting this into Equation 5.26 we obtain:

 0

 2 2
, − ∞ ≤ x < −b
1 +(a/2)x + abx + ab /2 , − b ≤ x ≤ 0
V (x) = (5.30)
ǫ −(a/2)x2 + abx + ab2 /2 , 0 < x ≤ +b
 2
ab , + b < x ≤ −∞

54
Now check your answer. First, as always, check that is is dimensionally correct. Second,
note that V (x) should be a continuous function of x, since integration over any function
(specifically excluding the impulse or “delta” function) results in a continuous function.

To find the volume charge density a in terms of V, we simply evaluate:

ab2 ab2
Vd = V (x = +b) − V (x = −b) = −0 = . (5.31)
ǫ ǫ
and solve for a:
ǫVd
a= . (5.32)
b2

Given the relative permittivity of silicon ǫr ≈ 12, b = 100 µm, and Vd = 0.4 V, we find
a = 4.25 mC/m3 .

55
[m0067] [2]

5.15-2

The symmetry of this problem suggests a solution in spherical coordinates. Laplace’s Equa-
tion in spherical coordinates is:
   
2 1 ∂ 2 ∂V 1 ∂ ∂V 1 ∂ 2V
∇V = 2 r + 2 sin θ + 2 2 =0 (5.33)
r ∂r ∂r r sin θ ∂θ ∂θ r sin θ ∂φ2

The symmetry of the problem also requires that the potential V not vary with respect to θ
or φ; in other words:
∂V ∂V
= =0. (5.34)
∂θ ∂φ
Thus, V is a function of r only, and Laplace’s Equation simplifies to:
 
1 ∂ 2 ∂
r V (r) = 0 (5.35)
r2 ∂r ∂r

Multiplying through by r2 and then integrating with respect to r, we obtain:

∂
r2 V (r) = C , (5.36)
∂r
where C is an arbitrary constant. Now dividing through by r2 and integrating with respect
to r again, we obtain:
C1
V (r) = − + C2 , (5.37)
r
where C1 and C2 are constants that can be determined by boundary conditions. Applying
the boundary conditions, we obtain:
C1
V (r = 1 m) = − + C2 = 100 V , and (5.38)
1m
C1
V (r = 3 m) = − + C2 = 20 V . (5.39)
3m
A simple way to solve for C1 and C2 here is simply to subtract the second equation from
the first equation, which eliminates C2 , then solve for C1 and use that result to solve for C2 .
One finds C1 = −120 V·m and C2 = −20 V. Thus:

120 V · m
V (r) = + − 20 V , 1 m ≤ r ≤ 3 m. (5.40)
r

Note that an answer like “V (r) = 120/r − 20” is dangerously ambiguous, unless you specify
as part of the answer that r must be in meters and V will be in volts.

56
[m0068] [1]

5.16-1

The symmetry of this problem suggests a solution in cylindrical coordinates. Laplace’s

Equation in cylindrical coordinates is:
 
2 1 ∂ ∂V 1 ∂ 2V ∂ 2V
∇V = ρ + 2 2 + =0 (5.41)
ρ ∂ρ ∂ρ ρ ∂φ ∂z 2
The symmetry of the problem also requires that the potential V not vary with respect to φ
or z; in other words:
∂V ∂V
= =0. (5.42)
∂φ ∂z
Thus, V is a function of ρ only, and Laplace’s Equation simplifies to:
 
∂ ∂V
ρ =0 (5.43)
∂ρ ∂ρ
Integrating both sides with respect to ρ, we obtain:
∂
ρ V (ρ) = C1 , (5.44)
∂ρ
where C is an arbitrary constant. Now dividing through by ρ and integrating with respect
to ρ again, we obtain:
V (ρ) = C1 ln ρ + C2 , (5.45)
where C1 and C2 are constants that can be determined by boundary conditions. Applying
the boundary conditions, we obtain:

V (ρ = 0.001 m) = C1 ln (0.001 m) + C2 = 50 mV , and (5.46)

V (ρ = 0.002 m) = C1 ln (0.002 m) + C2 = 20 mV . (5.47)

A simple way to solve for C1 and C2 here is simply to subtract the second equation from
the first equation, which eliminates C2 ; then solve for C1 and use that result to solve for C2 .
One finds C1 = −43.3 mV and C2 = −249.0 mV. Thus:
 ρ 
V (ρ) = − (43.3 mV) ln − 249.0 mV , 1 mm ≤ ρ ≤ 2 mm. (5.48)
1m
Note that an answer that does not include “1 m” in the denominator of the argument of the
“ln” function is dangerously ambiguous unless you also specify that ρ must be in meters.
Here is another correct solution, this time with the argument in units of millimeters:
 ρ 
V (ρ) = − (43.3 mV) ln + 50.0 mV , 1 mm ≤ ρ ≤ 2 mm. (5.49)
1 mm
Note that C2 depends on the units of ρ in the argument of the logarithm function.

57
[m0021] [1]

5.18-1

A good way to get the charge density is to first find the electric field, and then to apply
the boundary condition that relates electric field to surface charge density on a conducting
surface. The electric field intensity is, in general:
∂V 1 ∂V ∂V
E = −∇V = −ρ̂ − φ̂ − ẑ . (5.50)
∂ρ ρ ∂φ ∂z
The last two terms are zero because the the answer cannot vary with respect to φ or z. So
we have:
∂V ∂ h  ρ  i 43.3 mV
E = −ρ̂ = −ρ̂ − (43.3 mV) ln − 249.0 mV = ρ̂ . (5.51)
∂ρ ∂ρ 1m ρ
The relevant boundary condition on the inner conductor is that the normal component of the
electric flux density D equals the surface charge density. The normal to the inner conductor
is +ρ̂, so we have:

ρs = (+ρ̂) · D|ρ=1 mm = ρ̂ · ǫE|ρ=1 mm = ǫr ǫ0 (43.3 V/m) . (5.52)

Since ǫ0 = 8.854 × 10−12 F/m and ǫr = 2.1, we have that the surface charge density on the
inner conductor is +804 pC/m2 .

58
[m0021] [2]

5.18-2

(a) Summarizing the problem statement, we have a sphere of radius a = 2 m containing

uniformly-distributed charge with volume density ρv = 3 pC/m3 , and the media is a dielec-
tric with ǫr = 4.5 everywhere. Poisson’s Equation is
ρv
∇2 V = − (5.53)
ǫ
Note that you could also do this problem by integrating over the charge distribution, and
that’s a great check. However, the problem statement requires you to use Poisson’s Equation.
The symmetry of the problem suggests the use of spherical coordinates. Noting that ∂V /∂θ
and ∂V /∂φ should be zero due to symmetry, we find
 
1 ∂ 2 ∂V ρv
r = − (5.54)
r2 ∂r ∂r ǫ
It’s straightforward to solve for V in this case. Here we go:
 
∂ 2 ∂V ρv
r = − r2 (5.55)
∂r ∂r ǫ
∂V ρv
r2 = − r 3 + C1 (5.56)
∂r 3ǫ
where C1 is an arbitrary constant. Continuing:
∂V ρv C1
=− r+ 2 (5.57)
∂r 3ǫ r
ρv 2 C 2
V (r) = − r + + C3 (5.58)
6ǫ r
where C2 and C3 are arbitrary constants. At this point you should confirm this result by
making sure it’s a solution to the original equation, and also by checking units.
Outside the sphere, ρv = 0. Thus:
Bo
V (r) = + Ao r>a (5.59)
r
Here we have replaced the constants C2 and C3 with Bo and Ao respectively. This is to remind
us that the constants may be different should we consider the region inside the sphere (as
we shall soon do). We can determine the value of the constant Ao by noting that V (r) → 0
as r → ∞, since the total charge is finite and contained within a finite region. Therefore,
Ao must be zero, leading to
Bo
V (r) = r>a (5.60)
r
To determine the value of the constant Bo we’re going to have to make some kind of con-
nection with V (r) inside the sphere. Inside the sphere:
ρv 2 B i
V (r) = − r + + Ai r≤a (5.61)
6ǫ r

59
Here we have replaced the constants C2 and C3 with Bi and Ai respectively. We can deter-
mine the value of the constant Bi by noting that V (r) must be finite as r → 0, since the
charge density is finite at r = 0. Therefore, Bi must be zero, leading to
ρv 2
V (r) = − r + Ai r≤a (5.62)
6ǫ
Now we apply the boundary condition at the surface of the sphere. Note that there is no
requirement for potential to be continuous (and it wouldn’t do us any good even if there
were, since we’d be stuck with one equation and two unknowns). The relevant boundary
condition at r = a is that the normal component of the electric field should be continuous:
Specifically,
[Do (r = a) − Di (r = a)] · r̂ = ρs (5.63)
where Do and Di are the electric flux densities outside and inside the sphere respectively,
and ρs is the surface charge density. The surface charge density ρs is zero, since all the charge
is taken into account as the volume charge density ρv . Also, D = ǫE; therefore, we have

[Eo (r = a) − Ei (r = a)] · r̂ = 0 (5.64)

Next we note E everywhere should be oriented in the r̂ direction due to symmetry. Thus,
we find:
Eo (r = a) = Ei (r = a) (5.65)
We can find the electric flux density by taking the gradient of the potential:
∂ h ρv 2 i ρv
E = −∇V = −r̂ − r + Ai = r̂ r r ≤ a (5.66)
∂r 6ǫ 3ǫ
 
∂ Bo Bo
E = −∇V = −r̂ = r̂ 2 r>a (5.67)
∂r r r
Now applying the boundary condition (Equation 5.65):

ρv Bo
a= 2 (5.68)
3ǫ a
Solving for Bo and substituting the result back into Equation 5.60, we obtain:

ρv a3 ρv a3
V (r) = = r>a (5.69)
3ǫr 3ǫr ǫ0 r
(Good time for a units check!) Finally, the answer:

0.201 V·m
V (r) = r>a (5.70)
r
(b) From the previous equation, V (3 m) = 66.9 mV.

60
[m0021] [3]

5.18-3

A good way to get the charge density is to first find the electric field, and then to apply
the boundary condition that relates electric field to surface charge density on a conducting
surface. The electric field intensity is:
∂V 1 ∂V 1 ∂V
E = −∇V = −r̂ − θ̂ − φ̂ . (5.71)
∂r r ∂θ r sin θ ∂φ
The last two terms are zero because the the answer cannot vary with respect to θ or φ. So
we have:  
∂V ∂ 120 V · m 120 V · m
E = −r̂ = −r̂ − 20 V = +r̂ . (5.72)
∂r ∂r r r2
The relevant boundary condition on the inner conductor is that the normal component of the
electric flux density D equals the surface charge density. The normal to the inner conductor
is +r̂, so we have:

ρs = (+r̂) · D|r=1 m = r̂ · ǫ0 E|r=1 m = ǫ0 (120 V/m) . (5.73)

Since ǫ0 = 8.854 × 10−12 F/m in free space, we have that the surface charge density on the
inner conductor is 1.06 nC/m2 .

61
[m0112] [1]

5.22-1

The net charge in the capacitor is zero because the charges on the two plates is equal
and opposite. The charge on the positively-charged plate is

Q+ = CV = (20 pF) (3 V) = +60 pC

62
[m0070] [1]

5.23-1

(a) The equivalent circuit is just a resistor R in parallel with a capacitor to which we’ll
assign the variable Cx . The impedance of the capacitor is −j/ωCx . Thus, the magnitude
of the capacitor’s impedance decreases with increasing frequency. The total impedance is
the parallel combination of R = 200 Ω and Cx . The effective resistance will decrease with
increasing frequency.

(b) This structure looks a lot like a parallel plate capacitor. Neglecting fringing fields,
capacitance is estimated as

ǫHW ǫ0 ǫr HW (8.854 × 10−12 F/m) · 37 · (0.3 mm) (0.3 mm)

Cx = = = = 49.1 fF
L L 0.6 mm
(c) The impedance is R k (−j/ωCx ). At f = 10 GHz, we have 144.8 − j89.4 Ω, so the
effective resistance is 144.8 Ω .

63
[m0070] [2]

5.23-2

From the problem statement: C < 3 pF, d = 2 mm, and ǫr = 3.0. Note:
ǫA ǫr ǫ0 A
C≈ =
d d
where A is the area in common. So:
(3 pF) d
A<
ǫr ǫ0

Therefore, the common area must be < 2.26 × 10−4 m2 .

64
[m0113] [1]

5.24-1

From the problem statement, C ′ = 30 pF/m and ǫr = 2.25 for polyethylene. The capacitance
of the original coaxial cable is
2πǫs
C′ =
ln (b/a)
where ǫs = ǫr ǫ0 for the spacer material and b/a is the ratio of the radius of the outer
conductor to that of the inner conductor. The capacitance of a coaxial cable that is identical
except polyethylene is replaced with air (ǫs = ǫ0 ) is:

′ 2πǫ0
Cnew =
ln (b/a)

Comparing the two equations, we observe:

′ C′ ∼
Cnew = = 13.3 pF/m
ǫr

65
[m0113] [2]

5.24-2

From the problem statement, a = 1 mm, b = 3 mm, ǫs = ǫ0 , and V = +1.5 kV mea-

sured at the outer conductor relative to the inner conductor. The capacitance of this cable
is
2πǫ0 ∼
C′ = = 50.6 pF/m
ln (b/a)
The outer conductor is positively-charged, and the line charge density on this conductor is

ρl = C ′ V ∼
= +76.0 nC/m

The circumference of the outer conductor is 2πb, so the surface charge density is
ρl
ρs = = +4.03 µC/m2
2πb

66
[m0114] [1]

5.25-1

From the problem statement, C = 4.7 mF and V = 16 V. The energy stored is

1
We = CV 2 ∼
= 602 mJ
2

67
[m0114] [2]

5.25-2

From the problem statement, C = 3.5 pF, d = 0.1 mm, ǫr = 10, and V = 3 V. For an
ideal parallel plate capacitor,
ǫA
C=
d
where ǫ = ǫr ǫ0 is the spacer permittivity and A is the plate area. In the present problem,
we find:
Cd ∼
A= = 3.95 × 10−6 m2
ǫ0 ǫr
The volume of the capacitor is

Ad ∼
= 3.95 × 10−10 m3

The energy in the capacitor is

1
We = CV 2 = 15.75 pJ
2
Therefore, the energy density is
We ∼ 3
= 39.8 mJ/m
Ad

68
Chapter 6

Steady Current and Conductivity

69
[m0071] [1]

6.4-1

(a) Resistance Rsteel per unit length l of the steel-only wire:

′ Rsteel 1 1
Rsteel = = = = 31.8 Ω/m
l σsteel · πa 2
(1.00 × 10 S/m) · π (0.1 mm)2
6

(b) Resistance per unit length of gold clad having outer radius b:

′ 1
Rgold =
σgold · π (b2 − a2 )
′
The total resistance per unit length Rtotal = 10 Ω/m is the parallel combination:
1 1 1 1

′
= ′
+ ′
= ′
+ σgold · π b2 − a2
Rtotal Rsteel Rgold Rsteel

Solving for b:
s 
1 1 1
b= ′
− ′
+ σgold · πa2 = 0.10263 mm
Rtotal Rsteel σgold · π

So the required thickness of gold is b − a, which is 2.63 µm .

70
[m0071] [2]

6.4-2

In the transmission line equivalent circuit (R′ ,G′ ,C ′ ,L′ ) model, R′ is a series resistance. Also,
any current applied to either conductor must return on the other conductor. Therefore, we
have
R′ = Ric ′
+ Roc ′
(6.1)
′ ′
where Ric is the resistance per length of the inner conductor and Roc is the resistance per
length of the outer conductor. Note

′ 1
Ric = (6.2)
σic Aic
where σic is the inner conductor conductivity and Aic is the cross-sectional area of the inner
′
conductor. Thus, Ric = 0.164 Ω/m. Also

′ 1
Roc = (6.3)
σoc Aoc
where σoc = σic (from the problem statement) and Aoc is the cross-sectional area of the outer
conductor, through which the current flows. Note:

Aoc = πb22 − πb21 (6.4)

where b1 and b2 are the radii of the inner and outer surfaces, respectively, of the outer
conductor. From the problem statement we have
 
0.05
b1 = 1 − b∼
= 0.1809 cm (6.5)
2
 
0.05
b2 = 1 + b∼
= 0.1901 cm (6.6)
2
(Check: the mean of b1 and b2 is (b1 + b2 ) /2 = b, as expected.) So Aoc ∼
= 1.081 × 10−6 m2 ,
′ ∼ ′ ∼
and subsequently, Roc =0.0406 Ω/m. Finally, we obtain R′ = Ric ′
+ Roc = 0.205 Ω/m.

71
[m0071] [3]

6.4-3

If the voltage drop is to be reduced by a factor of 2, then the resistance must be decreased
by a factor of 2. The DC resistance of a wire is l/σA where l is length, σ is conductivity,
and A is cross-sectional area. The use of the term “diameter” implies the wire has circular
cross section, so the original wire has A = π (D0 /2)2 and

l
R=
σπ (D0 /2)2
√
For this to be reduced by a factor of two, the new diameter must be 2 · D0 .

72
[m0071] [4]

6.4-4

The DC resistance of such a resistor is R = l/σA where A is cross-sectional area. In this

case, we may write
l l 4l
R= = 2 =
σA σπ (D/2) σπD2
√
where D is diameter. Note that D should increase by a factor of 2 in order to reduce R
by a factor of two. Thus, D becomes ∼= 1.41 mm.

73
[m0105] [1]

6.5-1

The conductance per unit length is

2πσs
G′ =
ln (b/a)

where σs is the spacer conductivity, and a and b are the radii of the inner and outer
conductors, respectively. From the book, RG-59 has σs ∼ = 5.9 × 10−5 S/m and exhibits
G′ ∼
= 200 µS/m normally. From the appendix “Conductivity of Some Common Materials,”
σs ≈ 5 S/m. The worst case is that spacer assumes the much higher conductivity of seawater,
in which case:
5 S/m ∼
G′ → (200 µS/m) = 17.0 S/m
5.9 × 10−5 S/m

74
[m0106] [1]

6.6-1

From the problem statement:

Length l = 1.2 cm,
radius a = 1.6 mm,
J is uniform (constant) in the resistor,
√
E = ẑE0 / ρ where E0 , 3 V · m−1/2 , and
P = 5 W.
Let σ be the conductivity of the material comprising the resistor. Then
Z
P = σ |E|2 dv
V

where V is the volume representing the resistor.

We cannot assume the material comprising the resistor is homogeneous. So, what do we
√
know about σ? Recall Ohm’s Law, J = σE. Since E is proportional to 1/ ρ and J is inde-
√
pendent of ρ, σ must have the form σ = σ0 ρ where σ0 is a constant having units of S·m−3/2 .

Continuing,  2
Z Z
√ E0
P = (σ0 ρ) √ dv = σ0 E02 ρ−1/2 dv
V ρ V

Let us assume the ends of the resistor are at z = 0 and z = l. Then:

Z a Z 2π Z l
2
P = σ0 E 0 ρ−1/2 [dρ (ρdφ) dz]
ρ=0 φ=0 z=0

Z a  Z 2π  Z l 
= σ0 E02 ρ +1/2
dρ dφ dz
ρ=0 φ=0 z=0
 
2 3/2
= σ0 E02 a (2π) (l)
3
Solving for σ0 :
3P ∼
σ0 = = 173 kS · m−3/2
4πa3/2 lE02
and subsequently,

√
σ∼
= 173 kS · m−3/2 ρ

75
Chapter 7

Magnetostatics

76
[m0115] [1]

7.1-1

Divergence of the electric field:

∇ · D = ρv , so
∇ · ǫE = ρv
where ǫ has units of F/m and ρv has units of C/m3 .
Curl of the electric field:
∇×E=0
Divergence of the magnetic field:
∇ · B = 0 , so
∇·H=0
Curl of the magnetic field:
∇×H=J
Noting J = σE:
∇ × H = σE
where σ has units of S/m.

77
[m0047] [1]

7.3-1

The differential form of Gauss’ Law for magnetism requires

∇·B=0

According the reported measurement:

 
∂ ∂ ∂
∇ · B = x̂ + ŷ + ẑ · x̂B0 x2 = 2B0 x
∂x ∂y ∂z

Therefore, the measurement is plausible only if B0 is zero.

78
[m0119] [1]

7.5-1

The magnetic flux density (not the same as magnetic flux!) in this case is given by

µ0 I
B(ρ) = φ̂ . (7.1)
2πρ

The magnetic flux is simply B integrated over the area of the loop (i.e., flux divided by area,
times area, is flux):
Z Z 23 cm Z 30 cm
µ0 I
Φ= B · ds = φ̂ · φ̂ dρ dz (7.2)
S ρ=3 cm z=0 2πρ

Where S is the area enclosed by the loop, and the absolute values of z don’t matter due to
symmetry, so you can pick any z’s you want as long as you cover 30 cm in the z direction.
Evaluating:
Z 23 cm  Z 30 cm   
µ0 I dρ µ0 I 23
Φ= dz = ln (30 cm) = 3 µT·m2 (7.3)
2π ρ=3 cm ρ z=0 2π 3

Solving for I we have

  −1

2π 23 T·m2
I = 3 µT·m 2
ln (0.3 m)−1 = 24.5 (7.4)
4π × 10−7 H/m 3 H

To get to units of A, the traditional units of current, recall that inductance (H) is defined
as magnetic flux (T·m2 ) divided by current (A), so we’re already there! Thus, we have
I= 24.5 A .

79
[m0119] [2]

7.5-2

The magnetic flux density (not the same as magnetic flux!) is given by

µ0 I
B(ρ) = φ̂ . (7.5)
2πρ

The magnetic flux is simply B integrated over the area of the loop (i.e., flux divided by area,
times area, is flux):
Z Z 0.02 m Z 0.10 m
µ0 I
Φ= B · ds = φ̂ · φ̂ dρ dz (7.6)
S ρ=0.01 m z=0 2πρ

Evaluating: Z  Z 
0.02 m 0.10 m
µ0 I dρ
Φ= dz (7.7)
2π ρ=0.01 m ρ z=0
 
(4π × 10−7 H/m)(3 A) 0.02
= ln (0.10 m) = 41.6 nT·m2 (7.8)
2π 0.01
Note that this may also be written as 41.6 nWb , since 1 Wb = 1 T/m2 .

80
[m0119] [3]

7.5-3

Ampere’s Law is Z
H · dl = Iencl (7.9)
C
where C is any path which encloses the current. A convenient path is just a constant-z circle
with radius a: Z 2π
H · φ̂ a dφ = I (7.10)
φ=0

Also, H = B/µ0 , so Z 2π
φ̂J0 a · φ̂ a dφ = I (7.11)
φ=0

Thus:
I = 2πa2 J0 = 20.0 mA . (7.12)

81
[m0119] [4]

7.5-4

Apply the right hand rule. When the thumb of the right hand points in the −ŷ direc-
tion, the curled fingers of the right hand point in the +ẑ direction at (+1, +1, 0) m.

82
[m0119] [5]

7.5-5

The measurements are explained by a wire aligned along the y-axis, with current flowing
in the −ŷ direction. This can be confirmed using the right-hand rule – point the thumb of
your right hand in the direction of the current, and the curled fingers of your right hand
point in the direction of the magnetic field.

83
[m0120] [1]

7.6-1

From the problem statement, we have that Coil 1 has N1 = 100 and I1 = 2 A. Coil 2
has N1 = 300 and I2 = 4 A, and is wound in the opposite direction. Both coils have
l = 10 cm and µ = µ0 . For Coil 1 we have
N1 I 1
B1 = b̂1 µ0
l

where b̂1 points in the direction of B1 inside the coil. For Coil 2 we have
N2 I 2
B2 = −b̂1 µ0
l
The total field B = B1 + B2 , so:
   N I 
N2 I2  µ0
   1 1
b̂1 · (B1 + B2 ) = µ0 − µ0 = |N1 I1 − N2 I2 | ∼
= 12.6 mT
l l  l

84
[m0049] [1]

7.7-1

The magnetic field in either a straight coil or a toroidal coil is proportional to current.
Current is proportional to the conductivity of the wire forming the coil. Therefore, doubling
the conductivity will double the magnetic field strength.

85
[m0123] [1]

7.12-1

The inductance of a linear inductor depends only on geometry and materials; therefore,
the inductance remains 1 H .

86
[m0124] [1]

7.13-1

If the loops are close together, then presumably the magnetic flux Φ through each wind-
ing is equal. Thus,
NΦ
L,
I
where N is number of linkages, which is this case is the number of windings. Thus,

LI LI
Φ= =
N 2

87
[m0124] [2]

7.13-2

From the problem statement:

length l = 5 cm,
radius a = 5/2 = 2.5 mm,
number of windings N = 300, and
relative permeability µr = 200.

Since l ≫ a and the winding density N/l is large, we may us the “long straight coil”
expression
µN 2 A
L≈
l
In the present problem:
(µr µ0 ) N 2 (πa2 ) ∼
L≈ = 8.88 mH
l
where µ0 = 4π × 10−7 H/m.

88
[m0127] [1]

7.15-1

Since
1
We = LI 2
2
We have r r
2We 2 · 2 mJ ∼
I= = = 292 mA
L 47 mH

89
[m0127] [2]

7.15-2

The energy initially stored in the inductor is

1
Wm = LI 2 = 6 nJ
2
The energy stored in the capacitor after the transfer is
1
We = CV 2 = 6 nJ
2

so V = 1.73 V .

90
Chapter 8

Time-Varying Fields

91
[m0055] [1]

8.3-1

Faraday’s Law says that the potential (or “emf”) induced in the coil is

d
Vemf = −N Φ(t)
dt
where Z
Φ(t) = B · ds
S

and where B(t) is the magnetic flux density and S is the surface defined by the cross-section
of the coil. When the magnetic flux is not varying with time, the potential is zero , so this
is the answer to parts (a) and (c).

While the magnetic field is being reduced, a non-zero potential is possible. Since the magnetic
field is spatially-uniform and parallel to the axis of the coil, the above integral simplies to

Φ(t) = B(t) · A

where B(t) is the scalar magnetic flux density and A is the cross-sectional area of the coil.
We do not know precisely how B varies with time (i.e., linearly with time? exponential
decay? etc.), so we cannot take a formal derivative. We can however estimate the derivative:

d ∆Φ B(t0 + 200 ms) · A − B(t0 ) · A

Φ(t) ≈ =
dt ∆t 200 ms
where t0 is the time at which the magnitude of the magnetic field begins to decrease. We
also note



B(t0 ) = µr µ0 H(t0 ) = 2 × 105 · 4π × 10−7 H/m · (20.0 mA/m) = 5.03 mT

and B(t0 + 200 ms) = 1.01 mT. Since A = 200 cm2 = 0.020 m2 , we may now calculate
 
 ∆Φ 

|Vemf | ≈ N   = 20.1 mV
∆t 

This is the answer to part (b). In some sense, this is the average emf generated in the coil
over the 200 ms period of interest; however, more precisely, this is merely the best estimate
of the instantaneous emf generated during that time, given the limited information about
the time dependence of the magnetic field over that time.

92
[m0055] [2]

8.3-2

Faraday’s Law says that the emf induced in a this loop is

d
Vemf = − Φ(t)
dt
(since a loop has N = 1 turn) where
Z
Φ= B · ds
S

Let us define Vemf to be across the resistor, with the “+” terminal on the right side and the
“−” terminal on the left side.1 Then:
Z +L/2 Z y0 +w
Φ= [ẑ B0 eay ] · [+ẑ dx dy]
x=−L/2 y=y0

where y0 is the location of the left side of the loop. Then:2

"Z # Z 
+L/2 y0 +w
ay B0 L ay0 aw
Φ = B0 dx e dy = e [e − 1]
x=−L/2 y=y0 a

Next we’re going to want to take the time derivative of Φ. However, to do that properly
we need to make sure we identify everything in the above expression for Φ that has a time
dependence. Only y0 depends on time. To make this clear, let us write y0 = ut + b, which
places the left side of the loop at y = b at time t = 0. Now we may write:
B0 L aut ab aw
Φ(t) = e e [e − 1]
a
so:3
d
Φ(t) = B0 Lueaut eab [1 − eaw ]
Vemf (t) = −
dt
The problem statement asks us to assess the situation when the left side of the loop is at
y = 0.5 m, so we choose t = 0 and b = 0.5 m (since we said earlier y0 = ut + b). Also from
the problem statement, u = −250 m/s. Thus:
d
Vemf (t = 0) = − Φ(t = 0) = B0 Lueab [1 − eaw ] = −7.60 A
dt
Finally, the current is simply this divided by R, which is −3.04 A. Because we chose the “+”
terminal to be on the right, the reference direction for current must be counter-clockwise
(i.e., this is the necessary direction for positive current to dissipate positive power in the
1
Not the only way to do it! This choice is arbitrary. Choosing the opposite reference polarity should give
you the exact same answer as long as you follow through correctly.
2
Once you complete this integration, it’s a good time for a units check!
3
Once you complete this differentiation, it’s a good time for a units check!

93
resistor, or alternatively you may think of this as being the necessary reference direction for
the loop to behave as a power source). We have found that the current is negative with
respect to this reference direction; therefore, the induced current is 3.04 A, clockwise .

You can check to make sure you got the correct current direction by using Lenz’s Law. If
the current is flowing clockwise, then the induced magnetic field in the loop is in the −ẑ
direction. The impressed magnetic flux is increasing, since the loop area is constant and
the magnetic field in the loop increases in the +ẑ direction as the loop slides to the left.
Therefore, the induced current is acting to oppose the change in the impressed magnetic
flux, as is required by Lenz’s Law.

94
[m0055] [3]

8.3-3

(a) Let’s stick with the SI system of units. Since “1” in the expression for B appears
to be unitless, B0 must have units of T (or Wb/m2 , if you prefer) . Since t is in s and “1”
appears to be unitless, k must have units of 1/s . Since a has units of m and equals vt, v
must have units of m/s .

(b) Here’s Faraday’s Law:

d
Vg (t) = −N Φ (8.1)
dt
where N = 1 since it is a loop (not a coil), and the magnetic flux is
Z Z Z
Φ= B · ds = [ẑB0 (1 + kt)] · [−ẑds] = −B0 (1 + kt) ds (8.2)
S S S

Note that ds is in the −ẑ direction. This is from the right-hand rule (of calculus), in which
your thump is along the loop (not the gap) pointing along the direction from the “−” terminal
to the “+” terminal. Since S represents the surface defined by the loop (actually any surface
defined by the loop, but we’ll keep it simple...), we have


Φ = −B0 (1 + kt) πa2 (8.3)

Substituting a = vt and expanding into two terms we get

Φ = −B0 πv 2 t2 − B0 kπv 2 t3 (8.4)

So Faraday’s Law says:

Vg (t) = 2B0 πv 2 t + 3B0 kπv 2 t2 (8.5)
which is more compactly written as:


Vg (t) = B0 πv 2 2t + 3kt2 (8.6)

Good time for a units check: Can you confirm that the result is dimensionally correct?

(c) The first problem is to determine the motional and transformer emf, so we should be
clear on what we mean by these terms. Motional emf is the contribution to the total emf
which is associated with changes in the size, shape, or orientation of the surface through
which the magnetic field lines are linked. Transformer emf is the contribution to the total
emf which is associated with changes in the magnetic field.
With that in mind, let’s consider an incorrect solution: You can’t set v = 0 and call the
result (in this case, zero) the transformer emf. This is wrong because if v were equal to zero
for some radius a > 0, then the calculated emf would be potentially non-zero because the
magnetic field is still time-varying.
However, it is true that setting k = 0 yields the motional emf (you can verify this for
yourself after reading through this solution), although this is pretty hard to justify, especially

95
since we just saw that arbitrarily setting constants to zero is dangerous. So, if you go that
approach you must be able to explain why this is reasonable.
With all this in mind, here are three reasonable ways to get a solution:
1. You could calculate the motional emf and transformer emf using Faraday’s Law, but in
two separate steps; that is, work out the emf for a static loop (to get the transformer
emf) and then for a static field (to get the motional emf).
2. You could calculate motional emf from magnetostatics (i.e., assume a static magnetic
field), then subtract this result from your answer for part (b) to get the transformer
emf.
3. The approach followed below, which is probably best since it deals directly with the
concepts of time-varying loop size vs. time-varying magnetic field.
Here we go: Z Z
d d d
Vg (t) = − Φ = − B · ds = − [ẑB(t)] · [−ẑds] (8.7)
dt dt S dt S
where B(t) ≡ B0 (1 + kt); i.e., the scalar component of the magnetic field. Continuing:
 Z 
d d
Vg (t) = + B(t) ds = + [B(t)A(t)] (8.8)
dt S dt
where A(t) ≡ πa2 (t); i.e., the area of the loop. The value in setting the problem up this
way is that we now have the magnetic field and the loop area set up as distinct, identifiable
factors in the solution. Now we differentiate using the chain rule:
   
d d
Vg (t) = B(t) A(t) + B(t) A(t) (8.9)
dt dt
Now we see clearly that the first term is the transformer emf and the second term is the
motional emf. Let’s label these Vgtr (t) and Vgm respectively. Now:
 
tr d
Vg (t) = B(t) A(t) = B0 k · πa2 (8.10)
dt
 
m d
Vg (t) = B(t) A(t) = B(t) · 2πa · v (8.11)
dt
Note that now we get the “expected” result when we set k = 0 and then v = 0. Also note
that transformer emf depends on loop area, and motional emf depends on loop perimeter –
you might have suspected this based on other problems you have encountered.
OK, now we’re ready to wrap up. To find out when the contributions of the transformer
emf and motional emf are equal, we set the above expressions equal and solve for time t = teq .
Here we go:
B0 k · πa2 = B(teq ) · 2πa · v (8.12)
Noting that a = vt and B(t) = B0 (1 + kt):

B0 k · πv 2 t2eq = B0 (1 + kteq ) · 2πv 2 teq (8.13)

96
Solving the above expression we find that teq = −2/k (good time for a units check, by the
way). Negative times don’t really make sense in the problem (how can the loop have zero
radius and before that negative radius?), so the transformer emf and motional emf are never
equal for k > 0 (i.e., magnetic field magnitude increasing). However, teq = +2/ |k| when
k < 0. So, the answer to the problem is:
at t = −2/k, and only if the magnitude of the magnetic field is decreasing.

97
[m0056] [1]

8.4-1

The measured voltage is the “transformer emf” induced by the magnetic flux through the
loop. According to Faraday’s Law, we have in general that:
Z
d d
Vemf = −N Φ = −N B · ds
dt dt S
Here, the number of turns, N , is 1; ds is perpendicular to the loop in the direction determined
by the Stoke’s Law convention (i.e., according to the reference polarity chosen for Vemf ),
and B can be written as
B = b̂B(t) = b̂B0 sin(2πf t + α)
where b̂ is simply a unit vector indicating the direction of B. We know b̂ is a constant with
respect to position because B was specified to be a uniform magnetic field. Thus, we have
for any particular orientation of the loop ds:
  Z 
d
Vemf = −B0 sin(2πf t + α) b̂ · ds .
dt S
 
V 
emf is maximized when b and ds point either in the same direction, or in exactly opposite
directions. In this case, the magnitude of the quantity in the rightmost square brackets is
simply the area of the loop, A, which here is 0.0314 m2 according to the problem statement.
For this orientation, we have:
 
V 
emf = B0 [2πf cos(2πf t + α)] A .
The above quantity is maximized when cos(2πf t + α) = 1, which corresponds to the peak
magnitude of Vemf , which is one-half of the peak-to-peak magnitude. Thus:

B0 2πf A = 0.5 × 20 mVpp , thus:

0.5 × 0.02 Vpp

B0 = = 507 nT .
2π(100 × 103 Hz)(0.0314 m2 )

98
[m0031] [1]

8.5-1

In the original scenario:

V2 N2
=p
V1 N1
where p = ±1 depending on the relative orientation of the windings (p = −1 for the example
shown in the book). Let V2′ be the new potential on the secondary coil. From Faraday’s
Law:
∂
V2′ = N2 Φ′2
∂t
′
where Φ2 is the magnetic flux through the secondary coil after the modification. Note

Φ2
Φ′2 =
2
since the secondary coil now intersects only half the flux it did previously. Subsequently,
Φ1
Φ′2 = p
2
Now:  
1 ∂ 1 N2 ∂ 1 N2
V2′ = p N2 Φ1 = p N1 Φ1 =p V1
2 ∂t 2 N1 ∂t 2 N1
From the problem statement, N1 = 200 and N2 = 300. Also p = −1 for the example shown
in the book. Therefore,
3
V2′ = − V1
4
with the sign depending on the relative orientations of the coil windings.

99
[m0031] [2]

8.5-2

The transformer is an application of Faraday’s law, which is intrinsic to the Maxwell-Faraday

Equation:
∂
∇×E=− B
∂t

100
[m0030] [1]

8.7-1

From the problem statement:

B0 = 2 T,
A = π (2 m2 ) = 4π m2 ,
the peak value of VT is 5 V, and
VT (t = 0) = 0 and increasing.

Since the loop is rotating in a static uniform magnetic field, VT must be sinusoidally-varying.
A general form for this variation is

VT (t) = AB0 ω cos (ωt + ψ)

where ω = 2πf is the angular frequency of rotation and ψ is an as-yet unknown phase offset.
However, it is known that
VT (t = 0) = AB0 ω cos(ψ) = 0
so ψ must be either π/2 or 3π/2. Since VT is increasing at t = 0, ψ must be 3π/2.

Furthermore, we know the peak value of VT . In the context of the general form, we find:

VT = AB0 ω = 5 V at maximum

and therefore, ω ∼
= 0.199 rad/s.

Putting this all together:

 
∼ 3π
VT (t) = (5 V) cos [0.199 rad/s] t +
2

101
[m0053] [1]

8.9-1

From the problem statement, the electric field intensity is

E = ŷ 3 V m−1 s−2 t2

(The fact that this is electric field intensity can be confirmed using dimensional analysis.)
The displacement current density in free space is
 
∂D ∂E

pA
= ǫ0 −1 −2
= ŷǫ0 6 V m s t∼ −2
= ŷ 53.1 pC m s −2
t = ŷ 53.1 2 t
∂t ∂t m s

102
Chapter 9

Plane Wave Propagation in Lossless

Media

103
[m0042] [1]

9.1-1

The general, time-domain, differential form of Ampere’s Law is:

∂D
∇×H=J+
∂t
The relationship between these quantities and the phasor representation of the same quan-
tities is: n o
H = Re He e jωt ,
n o
J = Re Je e jωt , and
n o
D = Re De e jωt
;
Now substituting these quantities into Ampere’s Law we have:
h n oi n o h n oi
∇ × Re He e jωt + ∂ Re De
e jωt = Re Je e jωt
∂t
The order of the “Re” operator and any linear real-valued operator can be exchanged (see
the textbook section on phasors for a proof of this). Taking advantage of this in the first
and last terms, we obtain:
n h io n o  h 
e jωt e jωt ∂ e jωt i
Re ∇ × He = Re Je + Re De
∂t

Note that the curl (“∇×”) operator operates only on position, and not on time. Thus, we
may rewrite the first term as shown below:
nh i o n o  h i
e e jωt e jωt ∂ e jωt
Re ∇ × H = Re Je + Re De
∂t

e being a
Note also that partial derivative in the last term operates only on time, whereas D,
phasor, is independent of time. Therefore, the partial derivative operates only on the factor
ejωt , and we have:
nh i o n o n o
Re ∇ × H ee jωt
= Re Jee jωt e
+ Re D · jωe jωt
.

Comparing terms above, we find that the phasor expression of Ampere’s Law that we seek
is:
∇×H e =J e + jω D
e .

104
[m0042] [2]

9.1-2

The expression worked out in the book is

e = −jω B
∇×E e

B is a flux density, so we use B = µH to obtain:

e = −jωµH
∇×E e

105
[m0036] [1]

9.2-1

e and H
Here are Maxwell’s Equations for source-free regions in terms of E e (only) in dif-
ferential form:
∇·E e =0 (9.1)
e = −jωµH
∇×E e (9.2)
e =0
∇·H (9.3)
e = +jωǫE
∇×H e (9.4)
e which satisfy at least the last three of the
The equation we seek must yield solutions for H
above four equations. We begin by taking the curl of Equation 9.4:
     
e e
∇ × ∇ × H = ∇ × +jωǫE = +jωǫ ∇ × E e (9.5)

e using Equation 9.2:

On the right, we can substitute for ∇ × E
   
+ jωǫ ∇ × E e = +jωǫ −jωµH e = +ω 2 µǫH
e (9.6)

On the left, we invoke the vector identity

∇ × ∇ × A = ∇ (∇ · A) − ∇2 A (9.7)

to obtain  
e =∇ ∇·H
∇×∇×H e − ∇2 H
e = −∇2 H
e (9.8)
e term. Substituting back into
where we have used Equation 9.3 to eliminate the ∇ · H
Equation 9.5 and rearranging terms we have
e + ω 2 µǫH
∇2 H e =0 (9.9)
√
Now substituting β = ω µǫ:
e + β 2H
∇2 H e =0 (9.10)
e
This is the homogeneous wave equation for H.

106
[m0036] [2]

9.2-2

e is
The wave equation for E
e + β 2E
∇2 E e =0

Also, we know that

90◦ π/2 rad π
β= = = rad/m
1m 1m 2
−1
which may also be expressed simply as π/2 m . So


e + 2.467 m−2 E
∇2 E e∼
=0

107
[m0038] [1]

9.4-1

(a) The wave equation for E e is ∇2 E

e + β 2E e = 0. In cylindrical coordinates, E
e = ρ̂E
eρ +
eφ + ẑE
φ̂E ez . The Laplacian operator in cylindrical coordinates is:
 
2 1 ∂ ∂ 1 ∂2 ∂2
∇ = ρ + 2 2+ 2
ρ ∂ρ ∂ρ ρ ∂φ ∂z

e
Thus, we have for the three components of E:
 
1 ∂ ∂ e 1 ∂2 e ∂2 e eρ = 0
ρ Eρ + 2 2 E ρ + Eρ + β 2 E
ρ ∂ρ ∂ρ ρ ∂φ ∂z 2
 
1 ∂ ∂ e 1 ∂2 e ∂2 e eφ = 0
ρ Eφ + 2 2 E φ+ Eφ + β 2 E
ρ ∂ρ ∂ρ ρ ∂φ ∂z 2
 
1 ∂ ∂ e 1 ∂2 e ∂2 e ez = 0
ρ Ez + 2 2 E z + Ez + β 2 E
ρ ∂ρ ∂ρ ρ ∂φ ∂z 2

e has no component in the ρ or φ direction, then E

(b) If E eρ = E eφ = 0 and we are down to
one equation:  
1 ∂ ∂ e 1 ∂2 e ∂2 e ez = 0
ρ Ez + 2 2 E z + Ez + β 2 E
ρ ∂ρ ∂ρ ρ ∂φ ∂z 2
ez /∂φ = ∂ E
If E is uniform in φ and z, then ∂ E ez /∂z = 0, so the second and third terms in
the above equation are zero. This leaves us with:
 
1 ∂ ∂ e ez = 0
ρ Ez + β 2 E
ρ ∂ρ ∂ρ

108
[m0039] [1]

9.5-1

(a) There are several ways to figure this out. One way is to start in phasor represenation,
in which the field is:
êE0 ej(ax+by+cz)
where a = 1 rad/m, b = 2 rad/m, and c = 3 rad/m. Note:

êE0 ej(ax+by+cz) = êE0 e+jax e+jby e+jcz

To see what’s going on here, consider a simpler version of the above equation, where the last
two terms are omitted:
êE0 e+jax
In this case, the answer we are looking for would be k̂ = −x̂, and in fact we also see that
the wavenumber β = a; i.e., 1 rad/m. Similarly, if we had just

êE0 e+jby

then the answer would be k̂ = −ŷ, with β = b = 2 rad/m. From this we can infer that a
vector (not necessarily a unit vector) that points in the direction of propagation in this case
is:
k = −ax̂ − bŷ − cẑ
Thus, the corresponding unit vector is
k −ax̂ − bŷ − cẑ
k̂ = =p
|k| (−a)2 + (−b)2 + (−c)2

Giving:
k̂ = −0.267x̂ − 0.534ŷ − 0.802ẑ

(b) In the above analysis, we find |k| = 3.74 rad/m. Following the reasoning above, this is
simply the wavenumber β. Thus, the wavelength is:
2π
λ= = 1.68 m .
β

(c) Since this is free space, and since we know the phase velocity in free space is c =
3.0 × 108 m/s, we also know the frequency, which is c/λ= 179 MHz .

109
[m0039] [2]

9.5-2

From the problem statement, the direction of propagation k̂ = −x̂ and H points in the
+ŷ direction. From the plane wave relationships:

E = −η k̂ × H

Therefore, E points in the − (−x̂) × ŷ = +ẑ direction.

110
[m0039] [3]

9.5-3

From the problem statement, |E| = 3 V/m and ǫr = 2. Since plastics are non-magnetic:
r
µ0 η0
η= =√
ǫr ǫ0 ǫr

where η0 ∼
= 376.7 Ω. Therefore,
|E| |E| √ ∼
|H| = = ǫr = 11.3 mA/m
η η0

111
[m0041] [1]

9.7-1

The spatial power density is

3W 3W
Save = = = 3 MW/m2
1 mm2 10−6 m2
Since Save = |E|2 /2η, and since η = η0 ∼
= 376.7 Ω in free space:
p
|E| = 2η0 Save ∼= 47.5 kV/m

112