Jackson 1.

15
From (1.11),
W =

0
2
_
V

2
dV.
Integrating by parts, we get
W =

0
2
__
A
dA
_
V
||
2
dV.
_
where the surface integral is over the bounding surfaces. Use E = and
Gauss Law:
W =

0
2
__
A
E dA +
_
V
||
2
dV.
_
=

0
2
_
Q
e
nc

0
+
_
V
||
2
dV
_
=

0
2
_
V
||
2
dV
(1)
since there is no enclosed charge. To extremize, we must set W = 0 when

= +. In this case, |

|
2
=

= +2()+O(
2
):
W =

0
2
_
V
2 () dV
=
0

0
_
V

2
dV
=
0
_
V

2
dV = 0.
In the last step, we use the assumption of equipotential which leads to = 0
and the absence of charge. Thus, the energy is extremized.
Jackson 2.1
Using the method of images, we assume an imaginary opposite charge on the
other side of the conducting plane. The potential of this two charge system
is given by
(x) =
1
4
0
_
q
|x

d|
+
q
|x +

d|
_
For the remainder of this problem, we will work in cylindrical coordinates
and ignore angular components.
1
2.1(a)
The surface charge is given by
() =
0
E
z
(, z = 0) =
0

z=0
Note that |x

d|
2
=
2
+ (r d)
2
. Plugging and chugging,

z
=
q
4
0

z
_
1
_
(z d)
2
+
2

1
_
(z + d)
2
+
2
_
=
q
4
0
_
z d
((z d)
2
+
2
)
3/2

z + d
((z + d)
2
+
2
)
3/2
_
Evaulated at z = 0, we arrive at the answer forthwith:
(r) =
0
E
z
(r, z = 0) =
dq
2(d
2
+
2
)
3/2
2.1(b)
Using Coulombs law on the image charge:

F =
1
4
0
qq
(2d)
2
z =
q
2
16
0
d
2
z
2.1.(c)
In this part we are to calculate the force on the plate by integrating
2
/2
0
over the plate. To wit:

F =
_

0
()
2
2
0
z 2d
= z
d
2
q
2
4
0
_

0

(
2
+ d
2
)
3
d
= z
d
2
q
2
4
0
1
4(d
2
+
2
)
2

0
=
q
2
16
0
d
2
z
2.1(d)
Note rst the generalization of the solution to 2.2(c):
F
z
(z) =
q
2
16
0
z
2
z
2
From there,
W =
_

d
F
z
(z) dz =
q
2
16
0
1
z

d
=
q
2
16
0
d
2.1(e)
In this case, we calculate the potential energy by multiplying the potential
due to the image charge by the real charge, q:
W =
1
8
0
qq
2d
=
q
2
8
0
d
2
.
In this case, the potential energy is twice that of the answer above. In
general, energy calculations with image charges do not work because moving
the original charge moves the image. Thus, above, moving the charge to
innity also moved the image to innity. The actual dierential along the
path was therefore 2dr, not dr as in the integral, since both charges moved.
2.1(f)
Plugging in the numbers, we get 3.6 eV, comparable to the work function for
most metals.
Jackson 2.3
There are three image line charges in this system: two of charge per unit
length to the left and below the original charge, and one of charge per
unit length diagonally from the original charge.
2.3(a)
The potential for the charge conguration above is just the sum of the indi-
vidual potentials:

T
(x) =

4
0
_
ln
R
2
|x x
0
|
2
ln
R
2
|x x
1
|
2
ln
R
2
|x x
2
|
2
+ ln
R
2
|x x
3
|
2
_
where x
i
are the locations of the charges (and r
i
= |x x
i
|. At the boundaries,
we notice immediately see that the = 0 due to symmetry: at x = 0,
|x x
0
| = |x x
3
| and |x x
1
| = |x x
2
|. Similar symmetry holds for y = 0.
Along the y = 0 plane, the tangential eld E
x
= /x. We note again
that the symmetry of the image charges: r
0
/x = r
1
/x and r
2
/x =
r
3
/x, such that all terms cancel.
3
2.3(b)
Before beginning this, we note that r
0
/y = r
1
/y and r
2
/y =
r
3
/y on the half-plane y = 0.
(x) =
0
E
y
(x, 0) =
0
(x, y)
y

y=0
=

_
y
0
(x x
0
)
2
+ y
2
0

y
0
(x + x
0
)
2
+ y
2
0
_
/ =
1

_
y
0
(x x
0
)
2
+ y
2
0

y
0
(x + x
0
)
2
+ y
2
0
_
2.3(b)
From a table of integrals,
_

0
1
(x x
0
)
2
+ y
0
dx =
1
2
2 arctan(x
0
/y
0
)
y
0
.
Thus,
Q
x
=
_

0
(x)dx =
2

tan
1
_
x
0
y
0
_
,
as required. Since the charge per unit length z is independent of z, integrating
over an innite plane yields innite charge.
Jackson 2.9
2.9(a)
The induced charge density is given by (2.15):
= 3
0
E cos .
We know that the pressure acting radially outward on the sphere from last
homework is F/A =
2
/2
0
. Thus the total force acting on the right hemi-
sphere is
F
z
=
_

2
2
0
r da
=
1
2
0
_
/2
/2
_

0
3
2
0
E
2
cos
2
Rcos dd
=
9
4

0
R
2
E
2
0
4
2.9(b)
The induced charge density in this case is given by:
= 3
0
E cos +
Q
4R
2
.
Thus,
F/A =

2
2
0
=
1
2
_
(3
0
E cos )
2
+
_
Q
4R
2
_
2
+
6Q
0
E cos
4R
2
_
.
We proceed as before:
F =
_

2
2
0
r da
=
1
2
0
_
/2
/2
_

0
(3
0
E cos )
2
+
_
Q
4R
2
_
2
+
6Q
0
E cos
4R
2
Rdd
=
9
4

0
R
2
E
2
0

QE
0
2
+
Q
2
32
0
R
2
Note the above corresponds to the left and right hemispheres. The QE
0
/2
terms in the expressions above point in the same direction, but since this
problem was solved in the direction away from the cut, the two hemispheres
have diering signs for that term. Additionally, since that term is in the same
direction for both halves, it cancels such that the foce on one hemisphere is
F =
9
4

0
R
2
E
2
0
+
Q
2
32
0
R
2
.
Jackson 2.15
2.15(a)
Obviously, the given form of the Greens function satises the boundary
conditions on x.

2
G =
_

2
x
2
+

2
y
2
_
G
= 2

n=1
_
n
2

2
g
n
(y, y

) sin(nx) sin(nx

) +

2
g
n
y
2
sin(nx) sin(nx

)
_
= 2

n=1
_

2
y
2
n
2

2
_
g
n
sin(nx) sin(nx

)
= 4(x x

)(y y

)
5
Using the completeness relation,
4(x x

)(y y

) = 4(y y

)
_
2

n=1
sin(nx) sin(nx

)
_
Thus we equate the two series and the individual terms in the series, giving
the relation in the book.
2.15(b)
_

2
y
2
n
2

2
_
g
n
= 4(y y

)
When y = y

,
g
n
(y, y

) = A
n
sinh(ny
<
) sinh(n(1 y
>
)).
Applying the jump condition in rst derivative,
g

n
(y, y

>y
g

n
(y, y

y>y

= 4,
yields
A
n
=
4
nsinh(n)
.
Problem II.A
Analyticity implies that there is a single-valued complex derivative of the
complex function f(z). In order for the derivative to exist, it must be the
same regardless of the direction of dierentiation.
df
dz

dz=dx
=
f
R
x
+ i
f
I
y
df
dz

dz=idy
= i
f
R
y
+
f
I
y
Equating the real and imaginary parts of the derivatives above yields the
Cauchy-Riemann equations:
f
R
x
=
f
I
y
f
I
y
=
f
R
y
6
Dierentiating f
R
, we nd that it solves Laplaces equation:

2
f
R
=

2
f
R
x
2
+

2
f
R
y
2
=

2
f
I
xy


2
f
I
yx
= 0.
Similarly,
2
f
I
= 0.
When f(z) = arcsin(z) = w,
exp iw = sin(w) + i cos(w) =

1 z
2
+ iz
f(z) = u + iv = i ln(
_
(1 z
2
) + iz) = i(| ln(
_
(1 z
2
)| + i arg(
_
(1 z
2
) + iz)
Thus the real part is u = arg(
_
(1z
2
) +iz), which corresponds to a dipole.
7