Dielectrics, Capacitance, and Electric Energy Section 2.3: Solutions To Problems

P2 SOLUTIONS TO PROBLEMS
DIELECTRICS, CAPACITANCE, AND

ELECTRIC ENERGY
Section 2.3 Bound Volume and Surface Charge
Densities
PROBLEM 2.1 Nonuniformly polarized dielectric parallelepiped. This
problem is similar to Example 2.1.
(a) Combining Eqs.(2.19) and (1.167), the bound (polarization) volume charge den-
sity inside the parallelepiped is
p
= P =
_
P
x
x
+
P
y
y
+
P
z
z
_
= P
0
_
1
a
+
1
b
+
1
c
_
, (P2.1)
so it turns out to be constant throughout the volume (v) of the body.
With Eq.(2.23) in mind, we realize that there is no bound surface charge on
the parallelepiped sides belonging to planes x = 0, y = 0, and z = 0, respectively,
since P(0
+
, y, z) = P(x, 0
+
, z) = P(x, y, 0
+
) = 0. On the sides belonging to planes
x = a, y = b, and z = c, the bound surface charge density is given by
ps1
= n
d
P = x P(a
, y, z) = P
0
,
ps2
= y P(x, b
, z) = P
0
,
and
ps3
= z P(x, y, c
) = P
0
, (P2.2)
respectively; hence,
ps1
=
ps2
=
ps3
=
ps
= const, as well, on the entire surface
(S) of the parallelepiped.
(b) By means of Eqs.(1.30), (P2.1), and (P2.2),
Q
p
=
p
v +
ps
S =
p
abc +
ps
(ab +bc +ac) = 0 , (P2.3)
which conrms that the total bound charge of the dielectric parallelepiped is zero.
Section 2.4 Evaluation of the Electric Field and
Potential Due to Polarized Dielectrics
PROBLEM 2.2 Uniformly polarized disk on a conducting plane. The
distribution of bound charges of the disk is determined in Example 2.2; the volume
bound charge density (
p
) is zero, the surface densities on the upper and lower disk
bases,
ps1
and
ps2
, are those in Eqs.(2.28), while
ps3
= 0 on the side disk surface.
The electric eld in both the disk and the air equals the eld due to two circular
sheets of charge with densities
ps1
and
ps2
in free space, as well as their negative
images in the conducting plane in Fig.2.36. Namely, by image theory (Section 1.21),
49
50 Branislav M. Notaros: Electromagnetics (Pearson Prentice Hall)
the charges (of density
ps2
) right on the plane, at the coordinate z = 0
+
, and their
images (of density
ps2
) right below the plane, at the coordinate z = 0
, cancel
each other, so we are left with the two circular sheets of charge (in free space) shown
in Fig.P2.1.
z
O
a
e
0
e
0
r
ps1
-r
ps1
e
0
e
0
d
d
Figure P2.1 Two circular sheets of charge
ps1
= P and
ps1
= P in free space
equivalent, by virtue of image theory, to the uniformly polarized dielectric disk on
a conducting plane in Fig.2.36.
Like in Eq.(2.29), we then invoke the superposition principle and the expression
for the electric eld due to a thin charged disk, Eq.(1.63), and obtain the total
electric eld vector at a point dened by the coordinate z (0 < z < ) along the
z-axis in Fig.P2.1 (or Fig.2.36) as follows:
E = E
1
+E
2
=
P
2
0
_
z d
|z d|

z d
_
a
2
+ (z d)
2
1 +
z +d
_
a
2
+ (z +d)
2
_
z (z > 0) .
(P2.4)
In the conductor in the original structure, Fig.2.36, there is no electric eld, so
E = 0 for z < 0.
PROBLEM 2.3 Uniformly polarized hollow dielectric cylinder. This is,
essentially, very similar to the analysis performed in Example 2.2; the only important
dierence is that the two circular sheets of surface charge with densities
ps1
=
P and
ps2
= P, which, for the electric eld computation, can be considered
to be in free space, now have a circular hole. The eld vectors E
1
and E
2
in
Eq.(2.29) at the center of the hollow dielectric cylinder (point O), where E
1
= E
2
,
are computed, therefore, as an integral of elds due to elementary rings (as in
Fig.1.14), by merely changing the integration limits in Eq.(1.63) to a (starting) and
b (ending). Alternatively, we can represent the hollow disk (with inner and outer
radii a and b, respectively, and charge density
s
) as a superposition of a solid
(continuous) disk (as in Fig.1.14) with radius b and charge density
s
and another
one with radius a and charge density
s
, use Eq.(1.63) for the elds due to each of
them, and add together the results. In either way, we obtain the following expression
for the total eld at the point O, whose distance from the centers of each hollow
disk is h (Fig.P2.2):
E = E
1
+E
2
= 2E
1
=
Ph
0
_
1
a
2
+h
2
b
2
+h
2
_
z . (P2.5)
P2. Solutions to Problems: Dielectrics, Capacitance, Electric Energy 51
O
z
a
b
h
h
E
1
E
2
E
r
ps1
r
ps2
Figure P2.2 Evaluation of the electric eld intensity vector at the center of a
uniformly polarized hollow dielectric cylinder in Fig.2.37.
PROBLEM 2.4 Nonuniformly polarized thin dielectric disk. (a) By
means of Eq.(2.19) and the formula for the divergence in cylindrical coordinates,
Eq.(1.170), the bound volume charge density of the nonuniformly polarized dielec-
tric disk in Fig.2.38 amounts to
p
= P =
1
r
r
(rP
r
) =
2P
0
a
. (P2.6)
From Eq.(2.23), the bound surface charge density on the lateral (cylindrical) surface
of the disk is
ps
= r P(a
) = P
0
, (P2.7)
while
ps
= 0 on disk bases.
(b) Since the disk is very thin (d a), its bound volume charge, also being uniformly
distributed (
p
= const) throughout the disk volume, can be considered as an
equivalent circular sheet of charge, with density
s
=
p
d = 2P
0
d/a; namely,
the total bound volume charge of the disk,
p
a
2
d, must be equal to the total
surface charge of the equivalent sheet, so
s
a
2
, which yields this expression for
s
. By the same token, the belt of bound surface charge over the disk lateral
surface can be approximated by an equivalent circular line charge (ring) with density
Q
=
ps
2ad/(2a) =
ps
d = P
0
d. Consequently, Eqs.(1.63) and (1.43) can be used
for the electric eld at the axis of an innitely thin disk and an innitely thin ring
of charge, respectively, and, by the superposition of the two results, E at the z-axis
in Fig.2.38 comes out to be
E =

s
2
0
_
z
|z|

z
a
2
+z
2
_
z +
Q
az
2
0
(z
2
+a
2
)
3/2
z
=
P
0
d
0
a
_
_
z
|z|

z
a
2
+z
2
_
a
2
z
2 (z
2
+a
2
)
3/2
_
z (< z < ) . (P2.8)
PROBLEM 2.5 Uniformly polarized dielectric hemisphere. (a) The
bound surface charge density at the bottom (at) surface of the polarized hemi-
sphere is that in Eqs.(2.28) the second expression, and
ps
at the upper (spherical)
surface in Fig.2.39 is given by Eq.(2.30), so
ps
= P cos , where now 0 /2.
(b) We remove the conducting plane in Fig.2.39 using the image theory. The cir-
cular sheet of bound surface charge right against the plane and its negative image
cancel each other. The hemispherical sheet, considered to be in a vacuum, is sup-
plemented with another hemispherical sheet below the plane of symmetry, which is
also described by the function in Eq.(2.30), but with /2 (note that the
function cos is antisymmetrical with respect to = /2, which corresponds to a
negative image of the charge, exactly as required by the image theory). We thus
obtain the full spherical sheet of bound charge in Eq.(2.30), and conclude that the
polarized hemisphere on the conducting plane is equivalent to the polarized sphere
of Fig.2.7, in free space. Hence, the electric eld intensity vector at the point O in
Fig.2.39 is given by E = P/(3
0
), Eq.(2.32).
PROBLEM 2.6 Nonuniformly polarized large dielectric slab. (a) Refer-
ring to Fig.P2.3 and employing Eqs.(2.19), (1.167), and (2.23), the bound volume
and surface charge densities of the dielectric slab are given by
p
=
P
x
x
=
2P
0
x
a
2
,
ps1
= x P(a
) = P
0
,
ps2
= x P(a
+
) = P
0
.
(P2.9)
r
p
r
ps2
0
x a -a
r
ps1
x
E
P
Figure P2.3 Evaluation of the bound charges and electric eld of the innitely
large nonuniformly polarized dielectric slab in Fig.2.40.
(b) We note that the volume charge distribution described by
p
(x) in Eqs.(P2.9) is
the same as that in Eq.(1.153), the dierence being only the dierent multiplicative
constants. The electric eld due to this charge, considered to be in free space, is
therefore given by Eqs.(1.155) and (1.154), with the adjustment of constants. On
the other side, the electric eld due to two parallel oppositely charged sheets of
densities
ps1
and
ps2
in Eqs.(P2.9) and Fig.P2.3, obtained using Eq.(1.64) or the
result of Problem 1.60, amounts to (
ps1
/
0
) x. By the superposition of these
elds, the resultant eld in Fig.P2.3 comes out to be
E(x) =
P
0
0
a
2
_
x
2
a
2
_
x +
P
0
0
x =
P
0
x
2
0
a
2
x (|x| < a) ; E(x) = 0 (|x| > a) .
(P2.10)
(c) The voltage across the slab is
V =
_
a
x=a
E(x) dx =
P
0
0
a
2
_
a
a
x
2
dx =
2P
0
a
3
0
. (P2.11)
Section 2.5 Generalized Gauss Law
PROBLEM 2.7 Electric ux density vector. (a) Combining Eqs.(2.41) and
(2.32), the electric ux density vector at the center (point O) of the polarized
dielectric sphere in Fig.2.7 turns out to be
D =
0
E+P =
P
3
+P =
2P
3
. (P2.12)
(b) To nd the vector D along the axis (z-axis) of the polarized dielectric disk lying
on a conducting plane in Fig.2.36, we apply Eq.(2.41) for points inside the dielectric,
for 0 < z < d, where |z d| = (z d) and (z d)/|z d| = 1 in the electric eld
expression from Problem 2.2, and obtain
D =
0
E+P =
P
2
_
z d
_
a
2
+ (z d)
2
+
z +d
_
a
2
+ (z +d)
2
_
. (P2.13)
For d < z < along the z-axis in Fig.2.36, P = 0 and D =
0
E (air), as well
as |z d| = z d in the electric eld expression (Problem 2.2), which gives the same
expression for D as in Eq.(P2.13), so it holds true for 0 < z < .
Finally, D = 0 for z < 0 (below the conducting plane).
PROBLEM 2.8 Total (free plus bound) volume charge density. (a) Using
Eqs.(2.45), (2.19), and (2.41), the total (free plus bound) volume charge density in
the dielectric can, indeed, be expressed (in terms of the electric eld intensity vector,
E, only) as
tot
= +
p
= D+ ( P) =
0
E + P P =
0
E . (P2.14)
(b) By means of the above expression and the formula for the divergence in Cartesian
coordinates, Eq.(1.167),
tot
for the given function E(x, y, z) comes out to be
tot
=
0
_
E
x
x
+
E
y
y
+
E
z
z
_
= 8.85
_
4yz 3y
2
+ 3z
2
_
pC/m
3
(y, z in m) .
(P2.15)
PROBLEM 2.9 Uniform eld in a dielectric. From the relationship +
p
=
0
E (previous problem) and the fact that the divergence of the electric eld
vector in this dielectric region (E) is zero, since the eld is uniform, the bound
volume charge density (
p
) in the region equals the negative of the free volume
charge density () at the same point, namely,
p
=
0
E = (E = const) . (P2.16)
PROBLEM 2.10 Closed surface in a uniform eld. Applying Gauss law,
Eq.(1.133), to the closed surface S in Fig.2.41, over which the electric eld vector
is constant (uniform electric eld in the region), and can thus be taken out of the
ux integral, and inside which there is no charge (charge-free region), we obtain
_
S
E dS =
Q
S
0
E
_
S
dS = 0 (E = const , Q
S
= 0) . (P2.17)
We can rotate the parallel-plate capacitor in Fig.2.41, together with its (uniform)
eld, in an arbitrary way around the surface S, which is xed, so that the vector
(result of the integration)
_
S
dS does not change. Hence, the resulting relationship
in Eq.(P2.17) is satised (the dot product is zero) for an arbitrary direction of the
vector E and a xed vector
_
S
dS. This is possible only if
_
S
dS = 0 , (P2.18)
which concludes our proof of this vector identity.
PROBLEM 2.11 Flux of the electric eld intensity vector. With the use
of Eqs.(2.41) and (2.44), the ux of E through a closed surface S situated entirely
inside the polarized dielectric body can be expressed as
E
=
_
S
E dS =
1
0
__
S
D dS
_
S
P dS
_
=
1
0
__
v
dv
_
S
P dS
_
,
(P2.19)
where v denotes the volume enclosed by S.
Section 2.8 Electrostatic Field in Linear, Isotropic,
and Homogeneous Media
PROBLEM 2.12 Total enclosed bound and free charge. Combining
Eqs.(2.41), (2.47), (2.16), and (2.43), we obtain, similarly to Eqs.(2.60) and (2.61),
the following relationship between the total bound charge Q
pS
and free charge Q
S
enclosed by an imaginary closed surface S, via the permittivity of the homogeneous
dielectric ( = const):
P =

0
D
_
S
P dS =

0
_
S
D dS Q
pS
=

0
Q
S
.
(P2.20)
PROBLEM 2.13 Charge-free homogeneous medium. That there is no
bound volume charge in a homogeneous linear medium with no free volume charge
is obvious from Eq.(2.60):
= 0
p
=
r
1
r
= 0 . (P2.21)
PROBLEM 2.14 Dielectric cylinder with free volume charge. This is
similar to Example 2.5 (dielectric sphere with free nonuniform volume charge).
(a) Applying the generalized Gauss law, Eq.(2.44), to the same cylindrical Gaussian
surface as in Problem 1.58 and Example 1.19, we obtain the following expression
for the electric ux density in the dielectric cylinder with = const:
D(r) 2rh = r
2
h
. .
v
D(r) =
r
2
(0 r a) . (P2.22)
Having then in mind Eq.(2.47), the voltage between the axis and the surface of the
cylinder amounts to
V =
_
a
r=0
E(r) dr =
1
0
_
a
0
D(r) dr =
a
2
4
r
0
. (P2.23)
(b) As in Eqs.(2.67) and (2.68), the bound volume charge density inside the cylinder
and surface charge density on its surface [note that the polarization vector is zero
outside the cylinder, Eq.(2.21)] are
p
(r) =
r
1
r
(0 r < a) and
ps
= P(a
) =

r
1
r
D(a
)
=
(
r
1)a
2
r
, (P2.24)
respectively.
PROBLEM 2.15 Linear-exponential volume charge distribution. (a) For
the eld (observation) points in the p-type region of the semiconductor in Fig.2.42,
Eq.(2.74) now becomes
E
x
(x) =

0
a
_
x
x
=
x
e
x
/a
dx
0
a
_
1
x
a
_
e
x/a
(< x 0) , (P2.25)
where the integral is solved by integration by parts [
_
xe
x
dx = (1 +x) e
x
+C].
Similarly, in the n-type region [see Eq.(2.75)] we have
E
x
(x) =

0
a
__
0
e
x
/a
dx
+
_
x
0
x
e
x
/a
dx
_
=
0
a
_
1 +
x
a
_
e
x/a
(0 < x < ) . (P2.26)
(b) Like in Eqs.(2.76) and (2.77), the potential at points in the p-type region in
Fig.2.42, for the reference point at the center of the junction (x = 0), is found as
V (x) =
_
0
x
=x
E
x
(x
) dx
0
a
_
0
x
_
1
x
a
_
e
x
/a
dx
=

0
a
_
2a
_
e
x/a
1
_
xe
x/a
_
(< x 0) , (P2.27)
and in the n-type region,
V (x) =
_
x
0
E
x
(x
) dx
=

0
a
_
x
0
_
1 +
x
a
_
e
x
/a
dx
=

0
a
_
2a
_
1 e
x/a
_
xe
x/a
_
(0 < x < ) . (P2.28)
(c) The voltage between the n-type and p-type ends (built-in voltage) of the pn
junction (diode) is
V (x ) V (x ) =
4
0
a
2
. (P2.29)
Section 2.9 Dielectric-Dielectric Boundary
Conditions
PROBLEM 2.16 Dielectric-dielectric boundary conditions. (a) From the
boundary condition in Eq.(2.79), the tangential components of E
1
and E
2
, namely,
their x- and y-components, must be the same on the two sides of the boundary
(plane z = 0), so
E
2x
= E
1x
= 4 V/m and E
2y
= E
1y
= 2 V/m . (P2.30)
As there is no free surface charge on the boundary, Eq.(2.83) gives the following for
the normal component (z-component) of E
2
for z = 0
:
E
2z
= E
2n
=

r1
r2
E
1n
=

r1
r2
E
1z
= 10 V/m , (P2.31)
and hence
E
2
= (4 x 2 y + 10z) V/m (
s
= 0) . (P2.32)
(b) Now
s
in the plane z = 0 is nonzero, and we use the boundary condition in
Eq.(2.85) in place of Eq.(P2.31),
1
E
1n
2
E
2n
=
s
E
2z
=
1
r2
0
(
r1
0
E
1z
s
) = 7 V/m , (P2.33)
which, combined with Eqs.(P2.30), results in
E
2
= (4 x 2 y + 7z) V/m (
s
= 0) . (P2.34)
PROBLEM 2.17 Conductor-dielectric boundary conditions. Conductor-
dielectric and conductor-free space boundary conditions in Eqs.(1.186), (2.58), and
(1.190) are obtained from dielectric-dielectric boundary conditions in Eqs.(2.84) and
(2.85) by specifying that E
2
= 0 and D
2
= 0 in the second medium (conductor) [see
Eq.(1.181)] and D = E or D =
0
E in the rst medium (dielectric or free space).
PROBLEM 2.18 Water-air boundary. Marking water as medium 2 and air as
medium 1 in Fig.2.11, the law of refraction of the electric eld lines at the water-air
boundary in Eq.(2.87) gives
2

1

1
2
0
tan
1
tan
2
0
1
0 (
2
= 45
) ,
(P2.35)
meaning that the eld lines in air are approximately normal to the water surface,
which is sketched in Fig.P2.4. Note that the exact angle
1
is
1
= arctan
_
r1
r2
tan
2
_
= arctan
1
80
= 0.7
. (P2.36)
2
1
e
r2
e
r1
Figure P2.4 Refraction of electrostatic eld lines at a water-air interface.
Section 2.10 Poissons and Laplaces Equations
PROBLEM 2.19 Poissons equation for inhomogeneous media. (a) Com-
bining Eqs.(2.45), (2.47), and (1.101), we obtain the following second-order dier-
ential equation:
(V ) = . (P2.37)
By means of the identity (fa) = f a+(f) a (derived in Section 3.6), it can
be written also in the form
(V ) +V =
2
V +
1
V =
, (P2.38)
which we refer to as Poissons equation for an inhomogeneous medium. On the
other hand, if the dielectric medium is homogeneous, = const, so that = 0,
and this equation becomes Eq.(2.93), namely, the standard Poissons equation.
(b) The version of Eq.(P2.38) for a charge-free region ( = 0), that is, Laplaces
equation for an inhomogeneous medium, reads hence
2
V +
1
V = 0 , (P2.39)
which, in turn, reduces to the standard Laplaces equation, Eq.(2.95), in the case
of a homogeneous medium.
PROBLEM 2.20 Vacuum diode. (a) We apply the one-dimensional Poissons
equation in the x-coordinate, Eq.(2.98), from which the volume charge density in
the vacuum ( =
0
) diode in Fig.2.43 amounts to
(x) =
0
d
2
V (x)
dx
2
=
4
0
V
0
x
2/3
9d
4/3
(0 < x < d) . (P2.40)
(b)-(c) As in Eq.(2.101), the electric eld intensity in the diode is given by
E(x) = V =
dV (x)
dx
x =
4V
0
x
1/3
3d
4/3
x (0 < x < d) . (P2.41)
Eq.(1.190) then tells us that the surface charge densities on the cathode and anode
(Fig.2.43) are
s1
=
0
x E(0
+
) = 0 and
s2
=
0
( x) E(d
) =
4
0
V
0
3d
, (P2.42)
respectively.
(d) Performing a similar integration as in Eq.(1.149) and using the second expression
in Eqs.(1.30), the total charge of the diode turns out to be
Q =
_
v
(x) S dx
..
dv
+
s1
S +
s2
S =
4
0
V
0
S
9d
4/3
_
d
x=0
x
2/3
dx +
4
0
V
0
S
3d
= 0 , (P2.43)
where S stands for the (inner) surface area of the cathode or anode in Fig.2.43.
PROBLEM 2.21 Application of Poissons equation in spherical coordi-
nates. This is similar to the application of Poissons equation in Cartesian coordi-
nates in Example 2.7.
(a) Because of the spherical symmetry of the problem, the electric potential depends
only on the radial spherical coordinate r, so that the formula for the Laplacian in
spherical coordinates, Eq.(2.97), retains only the rst term. Combining this formula
with Poissons equation, Eq.(2.93), we have
2
V =
1
r
2
d
dr
_
r
2
dV
dr
_
=
(r)
0
=
0
r
0
a
(a < r < b) , (P2.44)
and the resulting second-order dierential equation in r is solved by two integrations,
like in Eq.(2.99), as follows:
r
2
dV
dr
=
_ _
0
r
3
0
a
_
dr =
0
r
4
4
0
a
+C
1
V (r) =
_ _
0
r
2
4
0
a
+
C
1
r
2
_
dr
=

0
r
3
12
0
a

C
1
r
+C
2
. (P2.45)
The integration constants, C
1
and C
2
, are computed from the boundary conditions
at the surfaces of metallic electrodes,
V (a) =
0
a
2
12
0
C
1
a
+C
2
= V
0
and V (b) =

0
b
3
12
0
a

C
1
b
+C
2
= 0 . (P2.46)
Substituting the numerical data (note that
0
= 3 C/m
3
), the two equations (in
two unknowns, C
1
and C
2
) become
100C
1
+C
2
= 12.82 and 20C
1
+C
2
= 352.9 (C
1
in V m; C
2
in V) ,
(P2.47)
and their solution is C
1
= 4.251 V m and C
2
= 438 V. Hence, the potential at
an arbitrary point between the electrodes, for a < r < b, given with respect to the
outer electrode, which is at zero potential, comes out to be
V (r) =
_
2.823 10
6
r
3
4.251
r
+ 438
_
V (r in m) . (P2.48)
(b) In the same way as in Eq.(P1.81), we obtain the electric eld vector between
the electrodes (a < r < b) from the result for V in Eq.(P2.48):
E(r) = V =
dV
dr
r =
_
8.47 10
6
r
2
4.251
r
2
_
r V/m (r in m) . (P2.49)
PROBLEM 2.22 Application of Laplaces equation in spherical coordi-
nates. Since = 0 between the electrodes, we are now solving the one-dimensional
Laplaces equation in the radial spherical coordinate r,
2
V =
1
r
2
d
dr
_
r
2
dV
dr
_
= 0 (a < r < b) , (P2.50)
and the two integrations result in
r
2
dV
dr
= C
1
V (r) =
_
C
1
r
2
dr =
C
1
r
+C
2
. (P2.51)
From the boundary conditions,
V (a) =
C
1
a
+C
2
= V
0
and V (b) =
C
1
b
+ C
2
= 0 , (P2.52)
and hence C
1
= 0.125 V m and C
2
= 2.5 V, with which the solution for the
potential in Eq.(P2.51) becomes
V (r) =
_
0.125
r
2.5
_
V (r in m) . (P2.53)
(b) The electric eld for a < r < b is given by
E(r) =
dV
dr
r =
0.125
r
2
r V/m (r in m) . (P2.54)
Section 2.11 Finite-Dierence Method for
Numerical Solution of Laplaces Equation
PROBLEM 2.23 FD computer program iterative solution. Computer
program for the nite-dierence analysis of the coaxial cable of square cross section
(Fig.2.13) based on Eq.(2.107) is given in the associated MATLAB exercise.
(a) Simulation results for the distribution of the potential and the electric eld
intensity in the space between the conductors and for the charge distribution of the
conductors of the square coaxial cable, taking the grid spacing to be d = a/10 and
the tolerance of the potential
V
= 10
8
V, are shown in Fig.P2.5.
(b) The computed total charges per unit length of the inner and outer conduc-
tors of the cable [Eq.(2.111)], taking d = a/N and N = 2, 3, 5, 7, 9, 10, 12, and25,
respectively, are tabulated in Table P2.1.
Table P2.1 Total charges p.u.l. of the
inner and outer conductors of the square
coaxial cable in Fig.2.13.
N Q
inner
[ C/m] Q
outer
[ C/m]
2 0.8854 10
10
1.0625 10
10
3 0.9391 10
10
1.0983 10
10
5 0.9847 10
10
1.1084 10
10
7 1.0066 10
10
1.1083 10
10
9 1.0201 10
10
1.1074 10
10
10 1.0252 10
10
1.1069 10
10
12 1.0332 10
10
1.1060 10
10
25 1.0579 10
10
1.1032 10
10
0
0.01
0.02
0.03
0
0.01
0.02
0.03
1
0.5
0
0.5
1

x [m]
Potential distribution for N = 10, iterative FD method
y [m]
V

[
V
]
1
0.8
0.6
0.4
0.2
0
0.2
0.4
0.6
0.8
1
0 0.005 0.01 0.015 0.02 0.025 0.03
0
0.005
0.01
0.015
0.02
0.025
0.03
x [m]
y

[
m
]
Electric field intensity vector at each node for N = 10, iterative FD method
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.5
1
1.5
2
2.5
3
3.5
x [cm]
r
s
[nC/m]
2
innerconductor
x [cm]
0 0.5 1 1.5 2 2.5 3
-2
-1.8
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
r
s
[nC/m]
2
outerconductor
Figure P2.5 Electric potential and eld intensity in the space between the con-
ductors, and the surface charge density on the surfaces of conductors, of the coaxial
cable of square cross section in Fig.2.13 results by the iterative nite-dierence
technique [Eqs.(2.107)-(2.110)] (FD computer program is provided in the associated
MATLAB exercise).
PROBLEM 2.24 FD computer program direct solution. Computer pro-
gram for the nite-dierence analysis of a square coaxial cable by directly solving
the system of linear algebraic equations with the potentials at interior grid nodes
in Fig.2.13(b) as unknowns is provided in the associated MATLAB exercise.
(a) The computed potential, eld, and charge distributions, using the direct FD
technique, are shown in Fig.P2.6. An excellent agreement with the results by the
iterative FD technique (previous problem) is observed [also in part (b)].
(b) Table P2.2 contains the results, obtained by the direct FD technique, for the
0
0.01
0.02
0.03
0
0.01
0.02
0.03
1
0.5
0
0.5
1

x [m]
Potential distribution for N = 10, direct FD method
y [m]
V

[
V
]
1
0.8
0.6
0.4
0.2
0
0.2
0.4
0.6
0.8
1
0 0.005 0.01 0.015 0.02 0.025 0.03
0
0.005
0.01
0.015
0.02
0.025
0.03
x [m]
y

[
m
]
Electric field intensity vector at each node for N = 10, direct FD method
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.5
1
1.5
2
2.5
3
3.5
x [cm]
r
s
[nC/m]
2
innerconductor
x [cm]
0 0.5 1 1.5 2 2.5 3
-2
-1.8
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
r
s
[nC/m]
2
outerconductor
Figure P2.6 Electric potential, eld, and charge distributions of the square coax-
ial cable in Fig.2.13 results by the direct nite-dierence technique (computer
program is given in the associated MATLAB exercise).
total charges per unit length of the inner and outer conductors of the cable.
Table P2.2 Total p.u.l. charges of ca-
ble conductors obtained by the direct
FD technique.
N Q
inner
[ C/m] Q
outer
[ C/m]
2 0.8854 10
10
1.0625 10
10
3 0.9391 10
10
1.0983 10
10
5 0.9847 10
10
1.1084 10
10
7 1.0066 10
10
1.1083 10
10
9 1.0201 10
10
1.1074 10
10
10 1.0252 10
10
1.1069 10
10
12 1.0332 10
10
1.1060 10
10
25 1.0580 10
10
1.1032 10
10
Section 2.13 Analysis of Capacitors with
Homogeneous Dielectrics
PROBLEM 2.25 Capacitance of the earth. Using Eq.(2.121), the capaci-
tance of the earth turns out to be
C
earth
= 4
0
R = 709.6 F (R = 6378 km) . (P2.55)
PROBLEM 2.26 Capacitance of a person. Taking the diameters of the
inscribed and overscribed spheres to be 2R
min
= 30 cm and 2R
max
= 170 cm,
respectively (note that these adopted dimensions are completely arbitrary, and are
just meant to be illustrative of an estimated value of the capacitance), that is,
R
min
= 15 cm and R
max
= 85 cm, we have 4
0
R
min
= 16.7 pF and 4
0
R
max
=
94.6 pF. Therefore, the capacitance of an average human body can be estimated to
be within the range 16.7 pF < C < 94.6 pF.
PROBLEM 2.27 Capacitance of a metallic cube, computed by the
MoM. The capacitance of the metallic cube numerically analyzed by the method
of moments in Problem 1.85 amounts to C = Q/V
0
= 73.27 pF. The capacitances
of the metallic sphere [Eq.(2.121)] inscribed in the cube (sphere radius r
a
= a/2),
the sphere overscribed about the cube (r
b
= a
3/2), the one with r

c
= (r
a
+r
b
)/2,
the sphere having the same surface as the cube [r
d
= a
_
3/(2)], and that with
the same volume as the cube {r
e
= a[3/(4)]
1/3
} come out to be C
a
= 55.63 pF,
C
b
= 96.36 pF, C
c
= 76 pF, C
d
= 76.88 pF, and C
e
= 69.02 pF, respectively. We
see that the capacitance of the sphere whose radius equals the arithmetic mean of
the radii of spheres inscribed in and overscribed about the cube (C
c
= 76 pF) is
quite close in value to the capacitance of the cube.
PROBLEM 2.28 RG-55/U coaxial cable. Using Eq.(2.123), the capacitance
per unit length of the RG-55/U coaxial cable amounts to C
= 2
r
0
/ ln(b/a) =
70 pF/m.
PROBLEM 2.29 Capacitance p.u.l. of a square coaxial cable, FD anal-
ysis. The capacitance per unit length of the coaxial cable of square cross section
in Fig.2.13 numerically analyzed by the nite-dierence technique presented in Sec-
tion 2.11 is obtained to be C
iterative
= 51.26006 pF/m by the iterative FD technique
(Problem 2.23) and C
direct
= 51.26008 pF/m by the direct FD technique (Prob-
lem 2.24), so practically the same result, with the tolerance of the potential of
V
= 10
8
V in the iterative solution and the grid spacing of d = a/10 in both solu-
tions. As a reference, the per-unit-length capacitance of a (standard) coaxial cable
(of circular cross section) having the same ratio of conductor radii (b/a = 3) and
dielectric (air) as the square cable is C
standard coax
= 50.61 pF/m [from Eq.(2.123)].
PROBLEM 2.30 Parallel-plate capacitor model of a thundercloud.
Based on Eqs.(2.127) and (2.126), the capacitance of the parallel-plate capacitor ap-
proximating a thundercloud, the voltage between the top and bottom of the cloud,
and the electric eld intensity in the cloud come out to be C =
0
S/d = 132.8 nF,
V = Q/C = 2.26 GV, and E = V/d = 2.26 MV/m, respectively.
PROBLEM 2.31 MoM numerical analysis of a parallel-plate capacitor.
Computer program based on the method of moments for the analysis of the parallel-
plate capacitor in Fig.2.19 is given in the associated MATLAB exercise. Using the
MoM program, the capacitance (C) of this capacitor for d/a ratios of 0.1, 0.5, 1, 2,
and 10 is found to be 117 pF, 38.3 pF, 28.7 pF, 24 pF, and 20.6 pF, respectively.
The corresponding C values obtained from Eq.(2.127), which neglects the fringing
eects, turn out to be 88.5 pF, 17.7 pF, 8.85 pF, 4.43 pF, and 0.885 pF.
PROBLEM 2.32 Nonsymmetrical thin two-wire line. The only dierence
with respect to the analysis of a symmetrical thin two-wire transmission line in air,
Fig.2.22, is the upper integration limit in computing the voltage between the line
conductors. Incorporating this change in Eqs.(2.140) and (2.141), we obtain the
following expression for the capacitance per unit length of a nonsymmetrical thin
two-wire line:
V =
_
db
x=a
E dx =
Q
2
0
_
_
db
a
dx
x

_
db
a
d(d x)
d x
_
=
Q
2
0
_
ln x|
db
a
ln(d x)|
db
a
_
=
Q
2
0
_
ln
d b
a
ln
b
d a
_
=
Q
2
0
ln
(d a)(d b)
ab

Q
2
0
ln
d
2
ab
C
=
Q
V

0
ln(d/
ab)
(P2.56)
(since d a d and d b d).
Similarly, we can pursue an alternative way of obtaining C
based on Eq.(1.197),
as in Eqs.(2.142) and (2.143), which now become
V
M1
=
Q
2
0
ln
d b
a
+
Q
2
0
ln
b
d a
C
=
Q
V
M1

0
ln(d/
ab)
. (P2.57)
Of course, for a = b, the result for C
in Eq.(P2.56) or (P2.57) reduces to that

in Eq.(2.141).
PROBLEM 2.33 Thick symmetrical two-wire line. The capacitance per
unit length of a (thick or thin) symmetrical two-wire line with the wire distance to
radius ratio d/a in air computed using the two expressions,
C
1
=

0
ln{d/(2a) +
_
[d/(2a)]
2
1}
and C
2
=

0
ln(d/a)
, (P2.58)
for dierent values of d/a are given in Table P2.3. We see that the results obtained
with the thin-wire approximation of the line (C
2
) are quite accurate for d/a 10,
and are acceptable even for d/a 5.
Table P2.3 Capacitance p.u.l. of a two-wire
line computed by expressions in Eqs.(P2.58).
d/a C
1
(pF/m) C
2
(pF/m) Error (%)
3 28.9 25.32 14.15
5 17.754 17.283 2.72
10 12.134 12.08 0.443
20 9.2931 9.2853 0.084
100 6.04036 6.04023 0.0022
PROBLEM 2.34 Two small metallic spheres in air. The capacitor consist-
ing of two small metallic spheres in air can be analyzed as the thin two-wire line
in Fig.2.22, by just changing the eld dependence Q
/(2
0
r) to Q/(4
0
r
2
). With
this, Eq.(2.139) becomes
E = E
1
+E
2
=
Q
4
0
_
1
x
2
+
1
(d x)
2
_
, (P2.59)
and the voltage of the capacitor is, in place of Eq.(2.140), found as
V =
_
da
x=a
E dx =
Q
4
0
_
_
da
a
dx
x
2

_
da
a
d(d x)
(d x)
2
_
=
Q
4
0
_
1
a

1
d a

_
1
d a

1
a
__
=
Q
2
0
_
1
a

1
d a
_
Q
2
0
a
(d a) .
(P2.60)
Its capacitance is
C =
Q
V
= 2
0
a . (P2.61)
Note that, since d a, the voltage between the two spheres can alternatively
be obtained using the expression for the potential of an isolated metallic sphere,
Eq.2.120, twice. Namely, V can be computed as the voltage from the rst sphere
(with charge Q) to the reference point for potential (at innity) plus the voltage
from the reference point to the second sphere (the one with charge Q), where the
two voltages equal the potential (with respect to innity) of the rst sphere and the
negative of the potential of the second sphere, respectively, so that
V = V
sphere with Q
+ (V
sphere with Q
) =
Q
4
0
a
+
_
Q
4
0
a
_
=
Q
2
0
a
. (P2.62)
PROBLEM 2.35 Four parallel wires in air. Let the rst conductor of the
transmission line (the upper pair of wires) in Fig.2.44 be charged by Q
and the
second conductor (the lower pair of wires) by Q
per unit of their length. Because

of symmetry, both charges, Q
and Q
, are distributed equally between the con-

nected wires, as shown in Fig.P2.7. Assuming that the second conductor of the line
(consisting of wires 3 and 4) is at a zero potential, the potential of the rst conduc-
tor, namely, the potential at the point M
1
on the surface of wire 1 (Fig.P2.7), with
respect to the reference point M
2
taken on the surface of wire 3 (V
M2
= 0) equals
the sum of the corresponding potentials (at M
1
, with respect to M
2
) due to charged
wires 1-4. Each of these potentials is computed using Eq.(1.197), as follows [also
see Eq.(2.144)]:
V
M1
=
Q
/2
2
0
ln
d
a
+
Q
/2
2
0
ln
d
2
d
+
Q
/2
2
0
ln
a
d
+
Q
/2
2
0
ln
d
d
2
=
Q
2
0
ln
d
2
a
,
(P2.63)
so that the capacitance per unit length of the line amounts to
C
=
Q
V
M1
=
2
0
ln(d
2/a)
= 9.86 pF/m . (P2.64)
Q' -
_
1
2
M
2
d
d
a a
a
a
1 2
3 4
M
1
Q'
_
1
2
Q'
_
1
2
Q' -
_
1
2
Figure P2.7 Evaluation of the capacitance p.u.l. of the transmission line in Fig.2.44
(cross section of the structure).
PROBLEM 2.36 Two wires at the same potential and a foil. (a) We
assume that the rst conductor (consisting of the two connected wires) of the trans-
mission line in Fig.2.45 (whose other conductor is the metallic foil) is charged by
Q
, i.e., that each of the wires is charged by Q
/2, and replace the foil by negative

images of the wires, as shown in Fig.P2.8. The voltage between wire 1 (or wire 2)
and the foil in the original system (Fig.2.45) equals a half of the voltage between
wires 1 and 3 in the equivalent system, so that, as in Eq.(2.146), the capacitance
per unit length of the original transmission line turns out to be twice that of the
equivalent line with wires 1 and 2 as one conductor and wires 3 and 4 as the other
conductor, in Fig.P2.8. This latter capacitance (C
e
) is evaluated using Eq.(1.197),
as in the previous problem,
V
M1
=
Q
/2
2
0
ln
2h
a
+
Q
/2
2
0
ln
D
d
+
Q
/2
2
0
ln
a
2h
+
Q
/2
2
0
ln
d
D
=
Q
2
0
ln
2hD
ad
C
e
=
Q
V
M1
=
2
0
ln[2hD/(ad)]
= 13.25 pF/m , D =
_
d
2
+ (2h)
2
, (P2.65)
Q' -
_
1
2
M
2
M
1
Q'
_
1
2
Q'
_
1
2
Q' -
_
1
2
h
d/2 d/2
a a
O
1 2
h
3
4
r
s
a a
D
E
1
a
Figure P2.8 Equivalent (in the upper half-space) transmission line to that in
Fig.2.45, by virtue of image theory.
and thus the capacitance p.u.l. of the line in Fig.2.45 (between the short-circuited
two-wire line and the metallic foil) is C
= 2C
e
= 26.5 pF/m.
(b) The induced surface charge density (
s
) at the point O on the foil (in Fig.2.45)
is obtained in a similar way to that carried out in Problem 1.88 or in Eqs.(1.220)
and (1.221). With reference to Fig.P2.8, we have
s
=
0
E
tot
=
0
(4E
1
cos ) , E
1
=
Q
/2
2
0
(D/2)
, Q
= C
V , cos =
2h
D
(P2.66)
(V = 20 V), and hence
s
= 5.4 nC/m
2
.
PROBLEM 2.37 Capacitance per unit length of a wire-corner line. To
nd the capacitance per unit length of the transmission line in Fig.1.57, whose
one conductor is the thin metallic wire and the other conductor is the 90
corner
metallic screen, we use the expression for the voltage V between the wire and the
screen obtained in Problem 1.89, and write
C
=
Q
V
=
2
0
ln(h
2/a)
= 13.06 pF/m . (P2.67)
PROBLEM 2.38 Equivalent circuit with two spherical capacitors. (a)-
(b) The system in Fig.1.41 can be replaced by two air-lled spherical capacitors,
one with radii of electrodes a and b, as in Fig.2.16, and the other with radii c
and d (in place of a and b), which gives rise to the equivalent circuit in
Fig.2.46. The capacitances of the capacitors in the schematic diagram are thus,
from Eq.(2.119),
C
ab
=
4
0
ab
b a
and C
cd
= 4
0
c , (P2.68)
where C
cd
can also be obtained from Eq.(2.121), as the capacitance of an isolated
metallic sphere of radius c in air.
(c) The charge of the rst (left) electrode of the rst capacitor (of capacitance C
ab
)
in Fig.2.46 equals the charge of the metallic sphere in Fig.1.41, so Q
a
= Q. The
charge of the second electrode of this capacitor is opposite, hence Q
b
= Q
a
=
Q. Since the metallic shell in Fig.1.41 is uncharged, and it is represented by the
second electrode of the rst capacitor and rst electrode of the second capacitor (of
capacitance C
cd
) in the equivalent circuit, we have Q
b
+Q
c
= 0 and Q
c
= Q
b
= Q
in Fig.2.46, and then Q
d
= Q
c
= Q for the second capacitor.
From Fig.2.46, the potential at the point O in Fig.1.41, that is, the potential of
the metallic sphere, can be obtained, using Eqs.(2.113) and (P2.68), as
V = V
ab
+V
cd
=
Q
a
C
ab
+
Q
c
C
cd
=
Q(bc ac +ab)
4
0
abc
, (P2.69)
which, of course, is the same result as in Eq.(1.202).
PROBLEM 2.39 Equivalent circuit with three spherical capacitors. (a)
With a similar reasoning as in the previous problem, we replace the system of three
metallic shells from Problem 1.77 by an equivalent circuit shown in Fig.P2.9, where
the capacitances of the three spherical capacitors amount to
C
ab
=
4
0
ab
b a
= 8.34 pF , C
cd
=
4
0
cd
d c
= 20 pF , C
ef
= 4
0
e = 11.13 pF .
(P2.70)
For the charges of individual electrodes of capacitors, we now have
Q
a
= Q = 10 nC , Q
b
= Q
a
= Q , Q
b
+Q
c
= Q
m
(unknown) ,
Q
d
= Q
c
, Q
d
+Q
e
= 0 (outer shell uncharged) . (P2.71)
V
io
V
+
+
O
Q
a
Q
b
Q
c
Q
d
C
ab
C
cd
Q
e
Q
f
C
ef
inner
shell
Q
m
middle
shell
outer
shell
Figure P2.9 Equivalent circuit (with three spherical capacitors) for the system of
three spherical metallic shells from Problem 1.77.
From the schematic diagram in Fig.P2.9 and Eqs.(P2.71), the given potential of
the middle shell with respect to the reference point at innity can be expressed as
V = V
cd
+V
ef
=
Q
c
C
cd
+
Q
e
C
ef
=
Q
d
C
cd
+
Q
d
C
ef
= Q
d
C
cd
+C
ef
C
cd
C
ef
(V = 1 kV) ,
(P2.72)
and hence, using Eqs.(P2.70) as well, the charge of the middle shell (Q
m
) turns out
to be
Q
d
=
C
cd
C
ef
V
C
cd
+C
ef
= 7.15 nC Q
m
= Q
b
+Q
c
= QQ
d
= 2.85 nC .
(P2.73)
The voltage between the inner and outer shells equals
V
io
= V
ab
+V
cd
=
Q
a
C
ab
+
Q
c
C
cd
=
Q
C
ab
+
Q
d
C
cd
= 1.55 kV , (P2.74)
the same as in Eq.(P1.154).
(b) For changed conditions in the circuit in Fig.P2.9 (and a slightly dierent nota-
tion), set in Problem 1.78, the charges of capacitor electrodes are given by
Q
a
= Q
1
= 2 nC , Q
b
= Q
a
= Q
1
, Q
b
+Q
c
= Q
2
= Q
m
(unknown) ,
Q
d
= Q
c
, Q
d
+Q
e
= Q
3
= 2 nC (outer shell is now charged) , (P2.75)
from which the charges of the second and third capacitors are Q
c
= Q
1
+ Q
2
and
Q
e
= Q
1
+ Q
2
+ Q
3
, respectively. Since the voltage between the inner and outer
shells is zero, we obtain the charge of the middle shell (Q
2
) as follows:
0 = V
ab
+V
cd
=
Q
a
C
ab
+
Q
c
C
cd
=
Q
1
C
ab
+
Q
1
+Q
2
C
cd
Q
2
= 6.8 nC . (P2.76)
Finally, the potential of the inner and outer shells (V
1
= V
3
) and that of the middle
shell (V
2
), with respect to the reference point at innity, are
V
1
= V
3
= V
ef
=
Q
e
C
ef
=
Q
1
+Q
2
+Q
3
C
ef
= 611 V ,
V
2
= V
cd
+V
3
=
Q
c
C
cd
+V
3
=
(Q
1
+Q
2
)
C
cd
+V
3
= 851 V , (P2.77)
which are the same results as in Problem 1.78.
PROBLEM 2.40 Equivalent circuit with parallel-plate capacitors. The
capacitors in the equivalent circuit in Fig.2.47 are parallel-plate capacitors, whose
capacitances are found from Fig.1.42 neglecting the fringing eects and using
Eq.(2.127):
C
1
= C
2
= C
3
=
0
S
d
= 442.7 pF . (P2.78)
We express the given potential V of the second electrode with respect to the ground
as the voltage, in Fig.2.47, between that electrode and the rst one (which is
grounded) and as the voltage to the fourth electrode (also grounded), respectively.
We also express the given charge Q of the third electrode as the sum of charges of
the corresponding capacitor plates (in Fig.2.47). What we obtain are the following
three equations:
V = V
2
= V
21
= V
12
=
Q
1
C
1
, V = V
23
+V
34
=
Q
2
C
2
+
Q
3
C
3
, Q
2
+Q
3
= Q ,
(P2.79)
whose solution is Q
1
= 885.4 nC, Q
2
= 557.3 nC, and Q
3
= 1.443 C. By means
of Eqs.(2.126) and (2.113), the electric eld intensities in capacitors in Fig.2.47, i.e.,
between the electrodes in Fig.1.42(b), amount to
E
1
=
V
12
d
=
Q
1
C
1
d
= 100 kV/m , E
2
=
V
23
d
=
Q
2
C
2
d
= 63 kV/m ,
E
3
=
V
34
d
=
Q
3
C
3
d
= 163 kV/m . (P2.80)
Of course, E
4
= 0 in Fig.1.42(b), as the eld in the (fourth) capacitor with short-
circuited plates, which is not shown in Fig.2.47.
PROBLEM 2.41 Equivalent circuit with cylindrical capacitors. Fig.P2.10
shows an equivalent circuit with three cylindrical capacitors, whose capacitances per
unit length of the system in Fig.1.55 are computed using the expression for C
of a
coaxial cable, Eq.(2.123),
C
12
=
2
0
ln[3d/(2d)]
= 137.2 pF/m , C
23
=
2
0
ln[5d/(4d)]
= 249.3 pF/m ,
C
34
=
2
0
ln[7d/(6d)]
= 360.9 pF/m . (P2.81)
Q
a
Q
b
Q
c
Q
d
Q
e
Q
f
1 2 3 4
C
12
C
23
C
34
+ V
3
Q
2
=0
Figure P2.10 Equivalent circuit (with three cylindrical capacitors) for the system
of four cylindrical conductors (shells) in Fig.1.55.
From the given potential of the third conductor (cylindrical shell) in Fig.1.55
with respect to the ground, expressed, in the circuit in Fig.P2.10, via the voltage
between that conductor and the fourth one, we get
V
3
=
Q
e
C
34
(V
3
= 1 kV) Q
e
= C
34
V
3
= 360.9 nC/m . (P2.82)
Since the second conductor is uncharged, we see (in Fig.P2.10) that
Q
b
+ Q
c
= Q
2
= 0 Q
c
= Q
b
= Q
a
. (P2.83)
Expressing then V
3
via the voltage on the other side, to the rst conductor, gives
V
3
=
Q
a
C
12
c
C
23
= Q
a
C
12
+C
23
C
12
C
23
Q
a
=
C
12
C
23
V
3
C
12
+C
23
= 88.5 nC/m .
(P2.84)
Hence, the total charges p.u.l. of the rst and third conductors (in Fig.1.55) come
out [from Fig.P2.10 and Eqs.(P2.82)-(P2.84)] to be
Q
1
= Q
a
= 88.5 nC/m , Q
3
= Q
d
+Q
e
= Q
c
+Q
e
= Q
a
+Q
e
= 450 nC/m ,
(P2.85)
and the results are the same as those obtained in Problem 1.79.
Section 2.14 Analysis of Capacitors with
Inhomogeneous Dielectrics
PROBLEM 2.42 Spherical capacitor with a solid and liquid dielectric.
In the rst (old) electrostatic state, the charge of the capacitor is, from Eq.(2.164),
Q = 4
0
_
b a
r1
ab
+
c b
r2
bc
_
1
V = 1.152 nC (V = V
source
= 100 V) . (P2.86)
Upon the voltage source is disconnected, and the capacitor is left to itself, this
charge remains the same in the new electrostatic state, in which the capacitance
of the capacitor is changed (oil is drained from the capacitor, so
r2
= 1), and the
voltage across open-circuited terminals of the capacitor turns out to be:
V
new
=
Q
4
0
_
b a
r1
ab
+
c b
bc
_
= 136.5 V . (P2.87)
PROBLEM 2.43 Oil drain without disconnecting the source. The capac-
itor charge in the old electrostatic state is Q = 1.152 nC (previous problem), and
that in the new state (with
r2
= 1) amounts to
Q
new
= 4
0
_
b a
r1
ab
+
c b
bc
_
1
V
source
= 843.3 pC . (P2.88)
The dierence equals the charge ow through the source circuit (in the direction
from the positive end of the source toward the positive electrode of the capacitor,
that with Q):
Q
ow
= Q = Q
new
Q = 308.3 pC . (P2.89)
PROBLEM 2.44 Metallic sphere with dielectric coating. (a) This is a
D-system, and the electric ux density vector, D, in both the dielectric coating and
air is given by Eqs.(2.161) and (2.162), where Q is the charge of the metallic sphere.
The electric eld intensity vector is E
1
= D/(
r
0
) in the coating and E
2
= D/
0
in
air. From this eld, we can express, by integration, the potential of the sphere (with
respect to the reference point at innity), V , in terms of Q, and nd its capacitance
(C) employing Eq.(2.116). Alternatively, we can consider the dielectrically coated
metallic sphere in Fig.2.48 as a special case of the spherical capacitor with two
dielectric layers in Fig.2.27, and specialize Eq.(2.164) specifying
r1
=
r
,
r2
= 1,
and c , to obtain
C = 4
0
_
b a
r
ab
+
1
b
_
1
=
4
r
0
ab
b a +
r
a
= 2.224 pF . (P2.90)
(b) Using Eq.(2.58) or (1.191), the density of free surface charge on the metallic
sphere in Fig.2.48 amounts to
s
= D(a
+
) =
Q
4a
2
=
CV
4a
2
= 1.77 C/m
2
. (P2.91)
(c) Eq.(2.60) tells us that there is no bound volume charge in a homogeneous linear
medium with no free volume charge; so, in both the dielectric coating and air in
Fig.2.48,
p
= 0.
(d) By means of Eqs.(2.61) and (P2.91), the bound surface charge density on the
surface of the dielectric coating next to the metallic sphere equals
psa
=
r
1
s
= 1.327 C/m
2
, (P2.92)
while, from Eqs.(2.23), (2.59), and (2.47), the bound surface charge density on the
other surface of the coating is
psb
= P(b
) =

r
1
r
D(b
) =

r
1
r
CV
4b
2
= 147.6 nC/m
2
. (P2.93)
Q
e
e
0
E
a
r
b
r
psa
r
psb
r
s1
r
s2
r
s3
r
s4
Figure P2.11 Evaluation of the free and bound surface charges in the spherical
capacitor half lled with a liquid dielectric in Fig.2.29.
PROBLEM 2.45 Charge densities in a half-lled spherical capacitor.
(a) With the electric eld intensity E(r) in the capacitor given in Eq.(2.175) and
notation in Fig.P2.11, Eq.(2.58) results in the following free surface charge densities
on metallic surfaces of the capacitor:
s1
=
r
0
E(a
+
) =

r
Q
2(
r
+ 1)a
2
= 2.98 C/m
2
,
s2
=
0
E(a
+
) =
Q
2(
r
+ 1)a
2
= 995 nC/m
2
,
s3
=
r
0
E(b
) =

r
Q
2(
r
+ 1)b
2
= 119 nC/m
2
,
s4
=
0
E(b
) =
Q
2(
r
+ 1)b
2
= 39.8 nC/m
2
. (P2.94)
(b) From Eq.(2.61) and (P2.94), bound surface charge densities on dielectric surfaces
next to capacitor electrodes in Fig.P2.11 are
psa
=
r
1
s1
= 1.99 C/m
2
,
psb
=
r
1
s3
= 79.6 nC/m
2
,
(P2.95)
while Eq.(2.23) tells us that there is no bound charge on the surfaces dielectric-air
in the capacitor, as vectors E and P are tangential on these surfaces.
PROBLEM 2.46 Empty and half-lled spherical capacitor. Using
Eq.(2.119), the charge of the empty (air-lled) spherical capacitor in the rst elec-
trostatic state equals
Q = CV =
4
0
abV
b a
(V = V
source
= 15 kV) . (P2.96)
Eq.(2.176) then gives the following for the voltage across the open-circuited termi-
nals of the capacitor, half lled with a liquid dielectric, in the new state (Q is the
same):
V
new
=
Q
C
new
=
(b a)Q
2(
r
+ 1)
0
ab
=
2V
r
+ 1
= 8.571 kV . (P2.97)
PROBLEM 2.47 Metallic sphere half embedded in a dielectric. (a) This
is an E-system, and the electric eld intensity vector (E) is entirely radial. Moreover,
it can be evaluated in the same way as in Fig.2.29 and Eq.(2.174), and is given in
Eq.(2.175). By integrating E(r) from the metallic sphere surface to the reference
point at innity, we can obtain the potential of the sphere and its capacitance.
Alternatively, we can use Eq.(2.176) with b . In either way, the capacitance of
the metallic sphere in Fig.2.49 comes out to be
C = 2(
r
+ 1)
0
a . (P2.98)
(b) The two terms in the application of the generalized Gauss law in Eq.(2.174)
actually show how the charge Q of the metallic sphere in Fig.2.49 is distributed
between its two halves; the ratio of the charge on the lower half, Q
1
,to the charge
of the upper half, Q
2
, is thus
Q
1
Q
2
=
D
1
D
2
=

r
0
E
0
E
=
r
, (P2.99)
and hence the portion of Q that is located on the upper half of the sphere amounts
to
Q
2
Q
=
1
r
+ 1
. (P2.100)
PROBLEM 2.48 Coaxial cable with two coaxial dielectric layers. This
is a cylindrical version of the structure in Fig.2.27, so Eq.(2.162) becomes
D(r) =
Q
2r
, a < r < c , (P2.101)
and the voltage between the conductors of the coaxial cable in Fig.2.50 is
V =
_
b
r=a
D(r)
1
dr +
_
c
r=b
D(r)
2
dr =
Q
2
_
1
1
ln
b
a
+
1
2
ln
c
b
_
. (P2.102)
Hence, the capacitance per unit length of the cable amounts to
C
=
Q
V
= 2
0
_
1
r1
ln
b
a
+
1
r2
ln
c
b
_
1
= 114.7 pF/m . (P2.103)
Alternatively, we can nd C
of the cable in Fig.2.50 as the equivalent total

p.u.l. capacitance of a series connection, Eq.(2.156), of two coaxial cables with
homogeneous dielectrics, whose p.u.l. capacitances, from Eq.(2.123), are given by
C
1
=
2
r1
0
ln(b/a)
and C
2
=
2
r2
0
ln(c/b)
. (P2.104)
PROBLEM 2.49 Coaxial cable with a radial variation of permittivity.
(a) This, again, is a cylindrical version of the structure analyzed in Example 2.19.
The electric ux density vector in the dielectric is radial with respect to the cable
axis and its magnitude is given by D(r) = Q
/(2r). Similarly to the integration in

Eq.(2.169), the voltage of the cable can be computed as
V =
_
b
a
E(r) dr =
_
b
a
D(r)
r
(r)
0
dr =
Q
2
0
_
b
a
dr
r
r
(r)
, (P2.105)
and its per-unit-length capacitance as
C
=
Q
V
= 2
0
_
_
b
a
dr
r
r
(r)
_
1
=
2
0
b
b a
= 69.54 pF/m . (P2.106)
(b) The volume and surface bound charge densities in the dielectric are found as
in Eqs.(2.170)-(2.172) [of course, here we use the formula for the divergence in
cylindrical coordinates, Eq.(1.170), to compute
p
]:
P(r) =

r
(r) 1
r
(r)
D(r) =

r
(r) 1
r
(r)
C
V
2r
=

0
bV
b a
_
1
r

a
r
2
_

p
= P =
1
r
d
dr
[rP(r)] =

0
abV
(b a)r
3
=
5
0
aV
4r
3
,
psa
= P(a
+
) = 0 ,
psb
= P(b
) =

0
V
5a
. (P2.107)
PROBLEM 2.50 Coaxial cable with four dielectric sectors. (a) This is
the same type of structure as the one in Fig.2.30, and we can modify (specialize),
according to the permittivity variation in Fig.2.51, either Eqs.(2.177) or just the
nal result the expression for the capacitance per unit length of the cable in
Eqs.(2.178) with the integral in , which now reduces to
C
=
_
2
0
() d
ln(b/a)
=

r1
0
/2 +
r2
0
/2 +
r3
0
/2 +
r4
0
/2
ln(b/a)
=
(
r1
+
r2
+
r3
+
r4
)
0
2 ln(b/a)
= 211 pF/m . (P2.108)
(b) As explained in Example 2.21, the electric eld intensity, E(r), in the dielectric
of the cable in Fig.2.51 is given in Eq.(2.124), so that its value near the inner
conductor of the cable amounts to
E(a
+
) =
V
a ln(b/a)
= 9.978 kV/m , for 0 2 . (P2.109)
Similarly to Eqs.(2.179) and (2.180), the free surface charge densities on the inner
conductor of the cable for the notation in Fig.P2.12 are:
s1
=
r1
0
E(a
+
) = 530 nC/m
2
,
s2
=
r2
0
E(a
+
) = 176.7 nC/m
2
,
s3
=
r3
0
E(a
+
) = 88.35 nC/m
2
,
s4
=
r4
0
E(a
+
) = 883.5 nC/m
2
. (P2.110)
e
r3
e
r2
e
r4
e
r1
90
o
90
o
90
o
a
b
coaxial
cable
r
s1
r
s2
r
s3
r
s4
Figure P2.12 Free surface charge densities on the inner conductor of the coaxial
cable in Fig.2.51 (cross section of the structure).
PROBLEM 2.51 Charge distribution for two coated wires. Since d a
in Fig.2.31, we neglect the electric eld due to the second conductor (which is far
away) while computing the charge distributions near the rst conductor and its
dielectric coating, and vice versa.
(a) Using Eqs.(2.58), (2.181), and (2.183), the densities of free surface charges on
the wire conductors, for the notation in Fig.P2.13, amount to
s1
= D
1
|
near wire 1
= D
1
(a
+
) =
Q
2a
=
C
V
2a
= 16.4 nC/m
2
,
s2
= D
2
|
near wire 2
=
s1
= 16.4 nC/m
2
(C
= 10.3 pF/m) . (P2.111)

e e
d a a a a
1 2
x
D
1
D
D
2
x
M
Q'
-Q'
e
0
r
s1
r
s2
r
ps1
r
ps2
r
ps3
r
ps4
Figure P2.13 Free and bound surface charge densities in the thin two-wire line
with dielectric coatings over conductors in Fig.2.31 (cross section of the structure).
(b) From Eq.(2.60),
p
= 0 (because = 0) in the dielectric coatings. By means
of Eqs.(2.61), (2.23), (2.59), and (2.47), the bound surface charge densities on indi-
vidual surfaces of coatings (Fig.P2.13) are
ps1
=
r
1
s1
= 12.3 nC/m
2
,
ps2
=
ps1
= 12.3 nC/m
2
,
ps3
= P(2a
) =

r
1
r
D(2a
) =

r
1
r
Q
2(2a)
=

r
1
r
C
V
4a
= 6.15 nC/m
2
,
ps4
=
ps3
= 6.15 nC/m
2
. (P2.112)
PROBLEM 2.52 Two metallic spheres with dielectric coating. The volt-
age between the two dielectrically coated spheres can, since d a, be computed as
the voltage from the positively charged sphere to the reference point for potential (at
innity) plus the voltage from the reference point to the negatively charged sphere.
By the same token, the capacitance between the two spheres (C of the associated
capacitor) can be obtained as the equivalent capacitance of a series connection of
two equal capacitors, using Eq.(2.156), so as
C =
C
one sphere
C
one sphere
C
one sphere
+C
one sphere
=
C
one sphere
2
= 1.112 pF , (P2.113)
where C
one sphere
is the capacitance of one (isolated) metallic sphere with the di-
electric coating (Fig.2.48), found in Problem 2.44. Of course, the same result (for
voltage and capacitance) can be obtained by integrating the electric eld vector, be-
tween the surfaces of the two metallic spheres, along the line connecting the centers
of the two spheres, as in Fig.2.31.
PROBLEM 2.53 Two metallic spheres half embedded in a dielectric. (a)
As in the previous problem, the capacitance of the capacitor in Fig.2.52 amounts to
C =
C
one sphere
2
= (
r
+ 1)
0
a = 0.695 pF , (P2.114)
with C
one sphere
now being the capacitance of one (isolated) metallic sphere (half
embedded in a dielectric half-space) in Fig.2.49, computed in Problem 2.47.
(b)-(c) Because d a, we evaluate the charge distributions on and near each of
the metallic spheres in Fig.2.52 neglecting the electric eld due to the other charged
electrode of the capacitor. Therefore, the electric eld intensity E(r) around each
of the spheres can in this computation be considered to be that in Eq.(2.175), so
E(r) =
Q
2(
r
+ 1)
0
r
2
=
CV
2(
r
+ 1)
0
r
2
=
aV
2r
2
. (P2.115)
Using Eqs.(2.58), (2.60), (2.61), and (2.23), the free and bound charge densities in
the system, for the notation in Fig.P2.14, are found to be
s1
=
r
0
E(a
+
) =

r
0
V
2a
= 708 nC/m
2
,
s2
=
0
E(a
+
) =

0
V
2a
= 177 nC/m
2
,
p
= 0 ,
ps1
=
r
1
s1
= 531 nC/m
2
,
ps2
=
ps3
= 0 . (P2.116)
e
0
e
d
a a
Q -Q
r
s1
r
s2
r
ps1
r
ps2
r
ps3
-r
s1
-r
s2
-r
ps1
-r
ps2
r
p
Figure P2.14 Free and bound charge densities in the system of two charged metal-
lic spheres pressed into a dielectric half-space in Fig.2.52.
PROBLEM 2.54 Permittivity gradient normal to capacitor plates. Let
us subdivide the dielectric into thin slices of thicknesses dx as shown in Fig.P2.15.
As each thin layer (slice) can be considered as being homogeneous, of permittivity
(x), it is obvious that this capacitor represents a D-system and, moreover, a gen-
eralization of the parallel-plate capacitor in Fig.2.25(a), which has only two such
layers. Therefore, the electric ux density vector in all thin layers is the same, given
by Eq.(2.147), with the fringing neglected, that is,
D = D x =
Q
ab
x (0 x d) , (P2.117)
which can be conrmed, for instance, applying the generalized Gauss law, Eq.(2.43),
to a rectangular closed surface enclosing the rst plate, with the right-hand side
positioned in either one of the slices in Fig.P2.15. The electric eld intensity vector
in the dielectric is
E(x) =
D
(x)
x =
Q
(x)ab
x . (P2.118)
The voltage of the capacitor amounts to
V =
_
d
x=0
E(x) dx =
Q
ab
_
d
0
dx
(x)
=
Q
2
0
ab
_
d
0
dx
1 + 3x/d
=
Qd ln4
6
0
ab
, (P2.119)
and its capacitance to
C =
Q
V
=
6
0
ab
d ln 4
. (P2.120)
Alternatively, this capacitance can be obtained by modeling each thin layer
in Fig.P2.15 (boundary surfaces between layers can be metalized) by a parallel-
plate capacitor with a homogeneous dielectric, plate separation dx, and capacitance
[Eq.(2.127)]
C
thin layer
= (x)
ab
dx
(0 x d) . (P2.121)
x
b
d
0
Q
-Q
e(x)
dx
E( ) x
D
Figure P2.15 Analysis of a parallel-plate capacitor with dielectric permittivity
gradient normal to plates [ = (x) in Fig.2.53].
Of course, all these capacitors are connected in series, and, in the equivalent circuit
representing a generalization of the circuit in Fig.2.26(a), the equivalent resultant
capacitance of such a connection equals, having Eqs.(2.155) and (2.156) in mind,
the inverse of the sum (integral) of the inverses of individual capacitances,
1
C
=
_
d
x=0
1
C
thin layer
=
_
d
0
dx
(x)ab
=
1
2
0
ab
_
d
0
dx
1 + 3x/d
=
d ln 4
6
0
ab
. (P2.122)
PROBLEM 2.55 Permittivity gradient parallel to capacitor plates. We
now subdivide the dielectric into thin layers of thicknesses dy as depicted in
Fig.P2.16, where each such layer can be considered as being homogeneous, of per-
mittivity (y), which implies that this is an E-system representing a generalization
of the parallel-plate capacitor with two homogeneous layers in Fig.2.25(b). The
electric eld intensity vector in all layers, in both capacitors, equals [Eq.(2.151)]
E = E x =
V
d
x (0 y b) , (P2.123)
and the electric ux density vector is
D(y) = (y)E x =
(y)V
d
x . (P2.124)
The generalized Gauss law applied to a rectangular box surface positioned about
the rst plate (Fig.P2.16) yields
_
b
y=0
D(y) a dy
..
dS
= Q
aV
d
_
b
0
(y) dy = Q , (P2.125)
where dS is the area of an elemental strip of length a and width dy, and Q is the
charge of the capacitor, and hence the capacitance of the capacitor
C =
Q
V
=
a
d
_
b
0
(y) dy =
2
0
a
d
_
b
0
_
1 + 3 sin
_
b
y
__
dy =
2( + 6)
0
ab
d
.
(P2.126)
d
Q -Q
b
D( ) y
E
y
0
dy e(y)
Figure P2.16 Analysis of a parallel-plate capacitor with dielectric permittivity
gradient parallel to plates [ = (y) in Fig.2.53].
An alternative approach based on an equivalent circuit as in Fig.2.26(b) is also
possible, where every thin layer in Fig.P2.16 is modeled by an elementary parallel-
plate capacitor with a homogeneous dielectric, plate area a dy, and plate separation
d, whose capacitance, from Eq.(2.127), is
dC = (y)
a dy
d
(0 y b) . (P2.127)
According to Eq.(2.159), the total capacitance of the parallel connection of all the
elementary capacitors amounts to
C =
_
b
y=0
dC =
_
b
0
(y)
a dy
d
=
2( + 6)
0
ab
d
. (P2.128)
PROBLEM 2.56 Capacitor with a nonlinear dielectric layer. (a)-(b)
This is similar to Example 2.23. Referring to Fig.P2.17, the condition V = 0 across
the capacitor (it is short-circuited) gives
Ed +E
0
d = 0 E = E
0
. (P2.129)
From the boundary condition for the normal components of the vector D in
Eq.(2.81), applied to the interface between the air-lled region and the ferroelectric
layer in Fig.P2.17, midway between the capacitor plates, we have
D
0
= D . (P2.130)
Air is a linear medium and the relationship between D and E in Eq.(2.51) holds
true. The ferroelectric is nonlinear and the relationship in Eq.(2.47) cannot be used,
but we can invoke the denition of D in Eq.(2.41), so that Eq.(P2.130) becomes
0
E
0
=
0
E +P . (P2.131)
Solving Eqs.(P2.129) and (P2.131), we obtain E
0
= P/(2
0
) (in air) and E =
P/(2
0
) (in the dielectric).
e
0 d
d
P
E
1
2
D
0
E
0
D
Figure P2.17 Electric eld intensities and ux densities in the nonlinear dielectric
layer and air region of the short-circuited parallel-plate capacitor in Fig.2.54.
Section 2.16 Electric Energy Density
PROBLEM 2.57 Energy of a spherical capacitor with two layers. (a)
Based on Eqs.(2.202), (2.199), and (2.162), the electric energy stored in the inner
dielectric layer in Fig.2.27 is obtained by integrating the electric energy density, w
e1
,
over the volume (v
1
) of this layer as follows:
W
e1
=
_
v1
w
e1
dv =
_
b
r=a
D(r)
2
2
1
4r
2
dr
. .
dv
=
_
b
a
Q
2
2
1
(4r
2
)
2
4r
2
dr
=
Q
2
8
1
_
1
a

1
b
_
=
(b a)Q
2
8
r1
0
ab
= 194.5 pJ , (P2.132)
where dv is adopted in the form of a thin spherical shell of radius r and thickness dr,
Eq.(1.33), and Q is the charge of the capacitor, computed in Eq.(2.164). Similarly,
the electric energy of the outer dielectric layer comes out to be
W
e2
=
_
v2
w
e2
dv =
_
c
b
D(r)
2
2
2
4r
2
dr =
(c b)Q
2
8
r2
0
bc
= 74.1 pJ . (P2.133)
(b) Alternatively, the energies in each of the dielectric layers can be found by repre-
senting the capacitor as a series connection [Fig.2.26(a)] of two spherical capacitors
with homogeneous dielectrics, whose capacitances, C
1
and C
2
, corresponding to the
individual layers, are given in Eqs.(2.165). Thus, using Eq.(2.192), we have
W
e1
=
Q
2
2C
1
=
(b a)Q
2
8
r1
0
ab
= 194.5 pJ and W
e2
=
Q
2
2C
2
=
(c b)Q
2
8
r2
0
bc
= 74.1 pJ .
(P2.134)
PROBLEM 2.58 Change in energy of a spherical capacitor. With the
use of Eq.(2.192), the change in electric energy of the capacitor between the two
electrostatic states turns out to be
W
e
= (W
e
)
new
(W
e
)
old
=
1
2
QV
new
1
2
QV
source
=
1
2
Q(V
new
V
source
) = 21 nJ ,
(P2.135)
where Q = 1.152 nC, V
new
= 136.5 V, and V
source
= 100 V (Problem 2.42).
PROBLEM 2.59 Energy of a coated metallic sphere. We can metalize the
equipotential surface r = b in the structure in Fig.2.48 and thus obtain two spherical
capacitors, with electrode radii of the second (air-lled) one being b and c .
Based on Eq.(2.192), the electric energies stored in the dielectric coating and in air
in Fig.2.48 can then, respectively, be expressed as
W
e1
=
Q
2
2C
1
and W
e2
=
Q
2
2C
2
, (P2.136)
where the charges of the two capacitors in the equivalent circuit in Fig.2.26(a)
are the same, Q, and their capacitances are obtained from Eqs.(2.165) specifying
r1
=
r
,
r2
= 1, and c :
C
1
=
4
r
0
ab
b a
and C
2
= 4
0
b . (P2.137)
The required equality of two energies in Eqs.(P2.136), i.e., the requirement that 1/2
of the total energy of the system is stored in the coating, implies the equality of the
capacitances in Eqs(P2.137), which further leads to
W
e1
= W
e2
C
1
= C
2

r
a
b a
= 1 b = (
r
+ 1) a = 5 cm .
(P2.138)
PROBLEM 2.60 Energy of a coaxial cable with two coaxial layers. This
is a cylindrical version of the energy computation in Problem 2.57.
(a) Having in mind Eqs.(2.206) and (2.207) and the expression for the electric ux
density in the cable dielectric, D(r) = Q
/(2r), we compute the electric energy

stored in the inner dielectric layer in Fig.2.50 per unit length of the coaxial cable by
integrating the electric energy density over the cross-sectional areas of the individual
layers, S
1
and S
2
, which results in
W
e1
=
_
S1
w
e1
dS =
_
b
r=a
D(r)
2
2
1
2r dr
. .
dS
=
_
b
a
Q
2
2
1
(2r)
2
2r dr =
Q
2
4
r1
0
ln
b
a
= 164 nJ/m , W
e2
=
_
S2
w
e2
dS =
_
c
b
D(r)
2
2
2
2r dr =
Q
2
4
r2
0
ln
c
b
= 409 nJ/m,
(P2.139)
where dS is the area of an elementary ring (adopted for integration) of radius r and
width dr, Eq.(1.60), Q
= C
V , and C
= 114.7 pF/m is the capacitance per unit

length of the cable, found in Problem 2.48.
(b) In the alternative approach, we represent the cable by the equivalent circuit
in Fig.2.26(a) and use the p.u.l. capacitances C
1
= 2
r1
0
/ ln(b/a) and C
2
=
2
r2
0
/ ln(c/b) (Problem 2.48), to obtain [see also Eq.(2.208)]
W
e1
=
Q
2
2C
1
=
Q
2
4
r1
0
ln
b
a
= 164 nJ/m , W
e2
=
Q
2
2C
2
=
Q
2
4
r2
0
ln
c
b
= 409 nJ/m.
(P2.140)
PROBLEM 2.61 Energy of a half-lled spherical capacitor. This is a
spherical version of the energy computation in Example 2.26. With the electric
eld intensity E(r) in the capacitor in Fig.2.29 given in Eq.(2.175), the electric
energy contained in the liquid, i.e., in the lower half of the space between electrodes
(volume v
1
), is computed as
W
e1
=
_
v1
w
e1
dv =
_
b
r=a
1
2

r
0
E(r)
2
. .
we1
2r
2
dr
. .
dv
=
_
b
a
1
2

r
0
Q
2
[2(
r
+ 1)
0
r
2
]
2
2r
2
dr
=

r
(b a)Q
2
4(
r
+ 1)
2
0
ab
, (P2.141)
where dv is the volume of a thin hemispherical shell of radius r and thickness dr
[1/2 of dv in Eq.(1.33)].
PROBLEM 2.62 Energy of a coaxial cable with four sectors. Like the
coaxial cable in Fig.2.33, this is an E-system, and the electric eld (E) is the same
as in the air-lled coaxial cable, Eq.(2.124). Similarly to the energy computation in
Eq.(2.210), the per-unit-length electric energy contained in the 90
dielectric sector
of relative permittivity
r1
(
r1
= 6) in Fig.2.51 is found to be
E(r) =
V
r ln(b/a)
W
e1
=
_
b
r=a
1
2

1
E(r)
2
. .
we1
2
r dr
. .
dS1
=

r1
0
V
2
4 ln
2
(b/a)
_
b
a
dr
r
=

r1
0
V
2
4 ln(b/a)
= 20.82 nJ/m , (P2.142)
with dS
1
standing for the surface area of a quarter (determined by the angle /2)
of a thin ring of radius r and width dr [ dS
1
is 1/4 of dS in Eq.(1.60)]. In the same
way, the p.u.l. energies in the remaining three 90
dielectric sectors in Fig.2.51 are

W
e2
=

r2
0
V
2
4 ln(b/a)
= 6.94 nJ/m , W
e3
=

r3
0
V
2
4 ln(b/a)
= 3.47 nJ/m ,
W
e4
=

r4
0
V
2
4 ln(b/a)
= 34.7 nJ/m . (P2.143)
PROBLEM 2.63 Energy of a capacitor with a variable permittivity.
Based on Eqs.(2.202) and (2.199) and the fact that D = const in the capacitor
dielectric (Problem 2.54), the total electric energy of the capacitor (W
e
) can be
computed as
W
e
=
_
v
w
e
dv =
_
d
x=0
D
2
2(x)
ab dx
. .
dv
=
abD
2
2
_
d
0
dx
(x)
=
abD
2
4
0
_
d
0
dx
1 + 3x/d
,
(P2.144)
where dv is adopted in the form of a thin at layer (slice) of thickness dx. By
the same token, the energy contained in the rst half (from x = 0 to x = d/2) of
the dielectric, W
e1
, can be obtained by replacing the upper limit (d) in the above
integrals by d/2. Hence, the percentage of W
e
contained in the rst half of the
dielectric turns out to be
W
e1
W
e
=
_
d/2
0
dx
1+3x/d
_
d
0
dx
1+3x/d
=
ln(5/2)
ln 4
= 66.1% . (P2.145)
PROBLEM 2.64 Energy of a capacitor with an inhomogeneous dielec-
tric. Given that now E = const in the dielectric of the capacitor (Problem 2.55),
we have
W
e
=
_
v
w
e
dv =
_
b
y=0
1
2
(y)E
2
ad dy
. .
dv
=
adE
2
2
_
b
0
(y) dy
=
0
adE
2
_
b
0
_
1 + 3 sin
_
b
y
__
dy , (P2.146)
so that the percentage of this energy stored in the lower half of the dielectric (from
y = 0 to y = b/2) amounts to
W
e1
W
e
=
_
b/2
0
[1 + 3 sin(y/b)] dy
_
b
0
[1 + 3 sin(y/b)] dy
=
1/2 + 3/
1 + 6/
= 50% . (P2.147)
PROBLEM 2.65 Energy of a system of spherical conductors. Using
Eq.(2.195), the energy of the system of three spherical conductors in Fig.1.56, in
which the third (outer) conductor is grounded (its potential is zero) and for which
the charges of the inner and middle conductors are, respectively, Q
1
= 1.85 pC and
Q
2
= 24.85 pC (Problem 1.80), is given by
W
e
=
1
2
(Q
1
V
1
+Q
2
V
2
+Q
3
V
3
) =
1
2
(Q
1
V
1
+Q
2
V
2
) = 138 pJ (V
3
= 0) . (P2.148)
PROBLEM 2.66 Energy of a system of at electrodes. (a) Using the
electric eld intensities between electrodes in the electrostatic system in Fig.1.42,
found in Example 1.28, the electric energy stored in the system is
W
e
=
1
2

0
E
2
1
. .
we1
Sd
..
v1
+
1
2

0
E
2
2
Sd +
1
2

0
E
2
3
Sd +
1
2

0
E
2
4
Sd
=

0
Sd
2
_
E
2
1
+E
2
2
+E
2
3
_
= 3.587 mJ (E
4
= 0) . (P2.149)
(b) Using the equivalent circuit in Fig.2.47, for which the capacitances and charges
are computed in Problem 2.40, the energy of the system equals
W
e
=
Q
2
1
2C
1
+
Q
2
2
2C
2
+
Q
2
3
2C
3
=
1
2C
1
_
Q
2
1
+Q
2
2
+Q
2
3
_
= 3.587 mJ (Q
4
= 0) ,
(P2.150)
which, of course, is the same result as in Eq.(P2.149).
PROBLEM 2.67 Energy of a system with free volume charge. The elec-
tric ux density in the dielectric cylinder is given by D(r) = r/2 (Problem 2.14),
and, similarly to the computation in Eqs.(2.213), (2.206), and (2.207), the electric
energy stored in the cylinder per unit of its length is found to be
W
e1
=
_
a
r=0
w
e
(r) 2r dr
. .
dS
=
_
a
0
D(r)
2
2
r
0
2r dr =

2
4
r
0
_
a
0
r
3
dr =

2
a
4
16
r
0
,
(P2.151)
with dS being the area of an elementary ring of radius r and width dr, Eq.(1.60),
in the cross section of the cylinder. The electric eld and ux density in air (outside
the cylinder), from Gauss law, are proportional to 1/r, and an integration like that
in Eq.(P2.151) results in W
e2
for the total per-unit-length energy contained in
the eld outside the cylinder. This means that such a charged cylinder cannot exist
alone in space; in fact, the sum of all charges p.u.l. of very long metallic/dielectric
systems in reality must be zero, as expressed by Eq.(11.27) in the eld analysis of
transmission lines.
PROBLEM 2.68 Energy of a pn junction. This is a planar version of the
energy calculation in Example 2.27.
The rst way to calculate the electric energy contained in the pn junction is to
use Eq.(2.202) and the electric eld expressions in Eqs.(2.74) and (2.75), as follows:
W
e
=
_
v
w
e
dv =
_

x=
1
2
E
x
(x)
2
S dx
..
dv
=

2
0
a
2
S
2
__
0
e
2x/a
dx +
_

0
e
2x/a
dx
_
=

2
0
a
3
S
2
, (P2.152)
where dv is an elementary volume for integration adopted in the form of a slice
of thickness dx across the junction [S is the area of the junction cross section
perpendicular to the x-axis in Fig.2.9(a)].
The second way is to employ Eq.(2.196) in a combination with the charge density
and potential expressions in Eqs.(2.69) and (2.76)-(2.77), respectively,
W
e
=
1
2
_
v
V dv =
1
2
_

x=
(x)V (x) S dx
..
dv
=

2
0
a
2
S
2
__
0
_
e
x/a
__
e
x/a
1
_
dx +
_

0
e
x/a
_
1 e
x/a
_
dx
_
=

2
0
a
3
S
2
.
(P2.153)
Of course, the results are the same.
Section 2.17 Dielectric Breakdown in Electrostatic
Systems
PROBLEM 2.69 Breakdown charge and energy of the earth. Based on
Eq.(1.193), dielectric breakdown of air near the earths surface occurs when
E(a
+
) =
Q
4
0
a
2
= E
cr0
= 3 MV/m , a = R = 6378 km (P2.154)
[E
cr0
is the dielectric strength of air, Eq.(2.53)], from which, and Eq.(2.192) and
the fact that C
earth
= 709.6 F (Problem 2.25), the maximum possible charge and
electric energy that could be stored on the earth turn out to be
Q = 4
0
R
2
E
cr0
= 1.36 10
10
C and W
e
=
Q
2
2C
= 1.3 10
23
J , (P2.155)
respectively.
PROBLEM 2.70 Maximum breakdown voltage of a spherical capacitor.
This is a spherical version of Example 2.29.
(a) Having in mind Eq.(2.117), the critical charge of the capacitor, at breakdown,
is given by
E(a
+
) =
Q
4
r
0
a
2
= E
cr
Q
cr
= 4
r
0
a
2
E
cr
. (P2.156)
Employing also Eq.(2.119), the breakdown voltage of the capacitor, for a given
radius a, is
V
cr
(a) =
Q
C
=
E
cr
b
a
2
+E
cr
a . (P2.157)
To optimize a such that V
cr
is maximum, assuming that b is xed, we equate to zero
the derivative of V
cr
with respect to a, which yields
dV
cr
da
= 0 E
cr
_
2
b
a + 1
_
= 0 a
opt
=
b
2
= 5 cm . (P2.158)
As the second derivative of V
cr
(a) for a = a
opt
is negative (equals 2E
cr
/b), this is
indeed a maximum (and not a minimum) of the function.
(b) The maximum breakdown voltage of the capacitor amounts to
(V
cr
)
max
= V
cr
(a
opt
) =
E
cr
b
4
= 625 kV . (P2.159)
(c) The corresponding value of the capacitor energy is
W
e
=
1
2
C (V
cr
)
2
max
=
1
2
4
r
0
a
opt
b
b a
opt
E
2
cr
b
2
16
=

r
0
b
3
E
2
cr
8
= 10.87 J . (P2.160)
PROBLEM 2.71 Breakdown in a wire-plane transmission line. (a) The
electric eld intensity of the wire-plane transmission line, in Fig.2.24(a), is the
largest very close to the surface of the wire conductor, where it is given by Eq.(2.224).
Namely, from the analysis in Problem 1.88, on the other side, the maximum eld
intensity near the ground plane, that straight below the wire (for x = 0), is
(E
near plane
)
max
=
Q
0
h

Q
2
0
a
(a h) . (P2.161)
The capacitance per unit length of the line (C
) is computed in Eq.(2.146), so that

the breakdown voltage of the line equals
E
max
=
Q
cr
2
0
a
= E
cr0
= 3 MV/m V
cr
=
Q
cr
C
=
Q
cr
2
0
ln
2h
a
= aE
cr0
ln
2h
a
= 159 kV . (P2.162)
(b) The maximum energy per unit length of the line is
W
e
=
1
2
C
V
2
cr
=

0
V
2
cr
ln(2h/a)
= 132.6 mJ/m . (P2.163)
(c) Using Eqs.(1.224), the largest possible electric force on the wire conductor in
Fig.2.24(a) per unit of its length comes out to be
F
e
=
Q
2
cr
4
0
h
=
(2
0
aE
cr0
)
2
4
0
h
=

0
a
2
E
2
cr0
h
= 25 mN/m . (P2.164)
PROBLEM 2.72 Grounded metallic sphere as a lightning arrester. (a)
With Q denoting the charge of the grounded small (a h) metallic sphere, as shown
in Fig.P2.18, the potential of the sphere due to Q and its image in the conducting
plane with respect to the plane, that is, the voltage between the surface of the
sphere and the plane (point 3 in Fig.P2.18), equals a half of the voltage between the
sphere and its image, so 1/2 of the voltage between two small metallic spheres in
air computed in Problem 2.34. It thus equals Q/(4
0
a), so that the total potential
V
1
at point 1 in Fig.P2.18 is given, in place of Eq.(2.228), by
V
1
=
Q
4
0
a
E
0
h . (P2.165)
The condition that the sphere is grounded then yields
V
1
= 0 Q = 4
0
E
0
ah . (P2.166)
(b) Since h a, the electric eld intensity on the surface of the sphere can be
evaluated using Eq.(1.193), neglecting the eld due to the image of Q,
E
1
=
Q
4
0
a
2
=
E
0
h
a
= 1200E
0
. (P2.167)
a
E
0
V=0
e
0
V
1
Q
1
2
3
h
H
Figure P2.18 Grounded small metallic sphere in a uniform atmospheric electric
eld above the surface of the earth (lightning arrester).
(c) Based on the electric eld expression from Problem 2.34, the eld intensities at
points 2 (H = 2 m) and 3 in Fig.P2.18 amount, for the reference direction of E
2
upward, to
E
2
=
Q
4
0
_
1
r
2
+
1
(2h r)
2
_
+E
0
= 0.9977E
0
(r = h H)
and E
3
=
Q
2
0
h
2
+E
0
= 0.9983E
0
(r = h) , (P2.168)
respectively.
Paralleling the computation in Eqs.(2.233), the eld intensities E
2
and E
3
cor-
responding to an eventual breakdown (in a thunderstorm) near the metallic sphere
in Fig.P2.18 [Eq.(2.232)] turn out to be
E
2
= 0.9977E
0
=
0.9977
1200
E
1
=
E
cr0
1202.8
, E
3
= 0.9983E
0
=
0.9833
1200
E
1
=
E
cr0
1202
,
(P2.169)
and they conrm the protective property of the grounded metallic sphere, serving
principally as a lightning arrester, in the same way as the grounded wire conductor
in Fig.2.34.
PROBLEM 2.73 Parallel-plate capacitor with two dielectric layers.
From Eq.(2.147), we obtain the following relationship between the electric eld
intensities in the two dielectric layers in Fig.2.25(a):
r1
E
1
=
r2
E
2
. (P2.170)
Let us rst assume that, for a high enough voltage V of the capacitor, a breakdown
occurs in the rst layer, which implies that E
1
= E
cr1
= 20 MV/m. At the same
time, Eq.(2.236) gives E
2
=
r1
E
1
/
r2
= 12 MV/m > E
cr2
= 11 MV/m, which
is impossible. We conclude that the second dielectric layer of the capacitor must
break down rst, in which case
E
2
= E
cr2
= 11 MV/m and E
1
=

r2
r1
E
2
= 18.33 MV/m . (P2.171)
The breakdown voltage of the capacitor is thus
V
cr
= E
1
d
1
+E
2
d
2
= 80.67 kV . (P2.172)
PROBLEM 2.74 Parallel-plate capacitor with two dielectric sectors.
The electric eld intensities in the two dielectric sectors in Fig.2.25(b) are the same,
Eq.(2.151), so it is obvious that, after a voltage of critical value is applied across
the capacitor plates, a breakdown will occur in the second sector (with a smaller
E
cr
). The corresponding eld values are
E
2
= E
cr2
= 11 MV/m and E
1
= E
2
= 11 MV/m < E
cr1
= 20 MV/m ,
(P2.173)
and the capacitor breakdown voltage comes out to be
V
cr
= E
1
d = 66 kV . (P2.174)
PROBLEM 2.75 Breakdown in a spherical capacitor with two layers.
(a) If the inner dielectric layer of the spherical capacitor in Fig.2.27 is made from
mica and the outer layer is oil, Eqs.(2.241) give
Q
(1)
cr
= 4
r1
0
a
2
E
cr1
= 48.07 C and Q
(2)
cr
= 4
r2
0
b
2
E
cr2
= 24.57 C .
(P2.175)
This means [see Eq.(2.242)] that the breakdown (for suciently high applied voltage
across the capacitor) occurs in the outer dielectric layer, and the breakdown value
of the capacitor charge amounts to
Q
cr
= Q
(2)
cr
= 24.57 C . (P2.176)
The breakdown voltage is
V
cr
=
Q
cr
C
old
= 2.134 MV , (P2.177)
where C
old
= 11.52 pF is the capacitance of the capacitor in the rst (old) electro-
static state (Problem 2.42).
(b) After the oil is drained from the capacitor, so that the outer dielectric layer is
now air, the second charge in Eqs.(P2.175) becomes
Q
(2)
cr
= 4
0
b
2
E
cr0
= 2.14 C (
r2
= 1 , E
cr0
= 3 MV/m) , (P2.178)
indicating that the outer layer breaks down rst, again. However, the breakdown
voltage changes, to
Q
cr
= Q
(2)
cr
= 2.14 C V
cr
=
Q
cr
C
new
= 253.7 kV , (P2.179)
with C
new
= 8.44 pF being the capacitance of the capacitor in the new state (Prob-
lem 2.42).
PROBLEM 2.76 Breakdown potential of a coated metallic sphere. (a)-
(b) For breakdown considerations, the value of the radius c in Fig.2.27 is actually
irrelevant, so it can even be c ; hence, we can reuse the expressions Q
(1)
cr
=
4
r1
0
a
2
E
cr1
and Q
(2)
cr
= 4
0
b
2
E
cr0
from the previous problem (for a spherical
capacitor with air as the second dielectric layer), which result in Q
(1)
cr
= 1.33 C
and Q
(2)
cr
= 0.3 C. We conclude that the breakdown would occur in air, near the
interface with the dielectric coating of the metallic sphere, and Q
cr
= Q
(2)
cr
= 0.3 C
[Eq.(2.242)]. Using the capacitance of the sphere computed in Problem 2.44, C =
2.224 pF, the breakdown potential of the sphere and the maximum electric energy
of this structure are
V
cr
=
Q
cr
C
= 135 kV and W
e
=
Q
2
cr
2C
= 20.2 mJ , (P2.180)
respectively.
PROBLEM 2.77 Breakdown in a coaxial cable with two layers. (a) Let
us pursue the rst way (based on the direct evaluation of the electric eld intensities
at characteristic locations in the structure) of the breakdown analysis presented for
a spherical capacitor with a two-layer dielectric in Example 2.32. For the coaxial
cable in Fig.2.50, Eqs.(2.234) and (2.235) take the following form:
E
1
(r) =
Q
2
r1
0
r
= E
1
(a
+
)
a
r
(a < r < b) , (P2.181)
E
2
(r) =
Q
2
r2
0
r
= E
2
(b
+
)
b
r
(b < r < c) , (P2.182)
and hence the relationship
r1
aE
1
(a
+
) =
r2
bE
2
(b
+
) . (P2.183)
An assumption that a breakdown occurs in the inner layer leads to
E
1
(a
+
) = E
cr1
= 40 MV/m E
2
(b
+
) = E
cr1
r1
a
r2
b
= 50 MV/m > E
cr2
,
(P2.184)
which is impossible. Hence, an eventual breakdown must be in the outer layer,
E
2
(b
+
) = E
cr2
= 20 MV/m E
1
(a
+
) = E
cr2
r2
b
r1
a
= 16 MV/m < E
cr1
.
(P2.185)
Similarly to Eq.(2.240), the breakdown voltage of the cable is computed, using
the eld-intensity values in Eqs.(P2.185), as
V
cr
= E
1
(a
+
) a
_
b
a
dr
r
+E
2
(b
+
) b
_
c
b
dr
r
=
r2
bE
cr2
_
1
r1
ln
b
a
+
1
r2
ln
c
b
_
= 38.8 kV.
(P2.186)
(b) The maximum energy that the cable can store per unit of its length amounts to
W
e
=
1
2
C
V
2
cr
= 86.4 mJ/m , (P2.187)
where C
= 114.7 pF/m is the capacitance p.u.l. of the cable, found in Problem

2.48.
PROBLEM 2.78 Simultaneous breakdown in two spherical layers. For
a suciently large applied voltage, dielectric breakdown will occur in both dielec-
tric layers in the capacitor in Fig.2.27 simultaneously if the two values of charges
Q
(1)
cr
and Q
(2)
cr
in Eqs.(2.241) are the same, which condition results in the following
relationship between the parameters of the capacitor:
Q
(1)
cr
= Q
(2)
cr

r1
a
2
E
cr1
=
r2
b
2
E
cr2
. (P2.188)
PROBLEM 2.79 Simultaneous breakdown in two coaxial layers. Let us
consider the critical value, for breakdown, of the charge Q
per unit length of the

coaxial cable in Fig.2.50, and denote by Q
cr1
and Q
cr2
the charge Q
in the case
of an eventual breakdown in the inner and outer dielectric layers, respectively, of
the cable. Based on the breakdown analysis in Problem 2.77, we then obtain the
following expressions for these charges, analogous to those in Eqs.(2.241) for the
corresponding spherical capacitor:
Q
cr1
= 2
r1
0
aE
cr1
and Q
cr2
= 2
r2
0
bE
cr2
. (P2.189)
Hence, we conclude that the breakdown will occur in both dielectric layers in
Fig.2.50 simultaneously if
Q
cr1
= Q
cr2

r1
aE
cr1
=
r2
bE
cr2
. (P2.190)
PROBLEM 2.80 Metallic sphere half immersed in a liquid dielectric.
Denoting by V
cr
the maximum potential of the metallic sphere in Fig.2.49 such that
dielectric breakdown will not occur after it is removed from the liquid and raised
in air high above the interface, the corresponding critical value of the charge of the
sphere, using the expression for its capacitance from Problem 2.47, turns out to be
Q
cr
= CV
cr
= 2(
r
+ 1)
0
aV
cr
. (P2.191)
This charge remains the same upon the elevation of the sphere above the liquid.
In the new electrostatic state, assuming that the sphere is isolated in air (as it
is high above the interface), the condition for a breakdown near the sphere surface,
expressed as in Problem 2.69, actually determines the value of Q
cr
, as follows:
Q
cr
= 4
0
a
2
E
cr0
(E
cr0
= 3 MV/m) . (P2.192)
Substituted back in Eq.(P2.191), for the rst electrostatic state, it leads to
V
cr
=
2aE
cr0
r
+ 1
= 30 kV . (P2.193)
Finally, we need yet to check whether the system in Fig.2.49 is safe from break-
down in the rst state. The electric eld intensity near the sphere surface in both
the liquid and air is, from Eqs.(2.175) and (P2.192),
E(a
+
) =
Q
cr
2(
r
+ 1)
0
a
2
=
2E
cr0
r
+ 1
=
E
cr0
2
= 1.5 MV/m . (P2.194)
Since E(a
+
) < E
cr0
(for air) and E(a
+
) < E
cr
= 20 MV/m (for the liquid dielec-
tric), we conclude that setting the sphere to the potential V
cr
or charging it with Q
cr
while half in the liquid (in the rst state) will not cause breakdown of the system,
which will occur only after the sphere is raised in air.
PROBLEM 2.81 Breakdown in a coaxial cable with a dielectric spacer.
(a) The electric eld intensity near the inner conductor of the cable in Fig.2.33 is
given by the expression in Eq.(2.218). This, for r = a
+
, is the largest electric eld
intensity in the cable. Since it is the same in both the dielectric spacer and the
air-lled part of the cable interior, an eventual dielectric breakdown will occur in
air (E
cr0
< E
cr
), if
E(a
+
) =
V
cr
a ln(b/a)
= E
cr0
= 3 MV/m , (P2.195)
and hence the breakdown voltage of the cable
V
cr
= E
cr0
a ln
b
a
= 6.59 kV . (P2.196)
(b) Having in mind Eqs.(2.178), as well as the capacitance found in Problem 2.50,
the capacitance per unit length of the cable (Fig.2.33) is
C
=

r
0
+
0
(2 )
ln(b/a)
=
(
r
/3 + 5/3)
0
ln(b/a)
= 84.4 pF/m , (P2.197)
so that the electric energy per unit length of the cable at breakdown amounts to
W
e
=
1
2
C
V
2
cr
= 1.83 mJ/m . (P2.198)

Dielectrics, Capacitance, and Electric Energy Section 2.3: Solutions To Problems

Uploaded by

Copyright:

Available Formats

Dielectrics, Capacitance, and Electric Energy Section 2.3: Solutions To Problems

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dielectrics, Capacitance, and Electric Energy Section 2.3: Solutions To Problems

Uploaded by

Copyright:

Available Formats

P2 SOLUTIONS TO PROBLEMS

DIELECTRICS, CAPACITANCE, AND

3/2), the one with r

in Eq.(P2.56) or (P2.57) reduces to that

per unit of their length. Because

, are distributed equally between the con-

, i.e., that each of the wires is charged by Q

/2, and replace the foil by negative

of the cable in Fig.2.50 as the equivalent total

/(2r). Similarly to the integration in

= 10.3 pF/m) . (P2.111)

/(2r), we compute the electric energy

= 114.7 pF/m is the capacitance per unit

dielectric sectors in Fig.2.51 are

) is computed in Eq.(2.146), so that

= 114.7 pF/m is the capacitance p.u.l. of the cable, found in Problem

per unit length of the

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.