Jackson 3.

4 Homework Problem Solution

Dr. Christopher S. Baird, Fall 2012
University of Massachusetts Lowell

PROBLEM:
The surface of a hollow conducting sphere of inner radius a is divided into an even number of equal
segments by a set of planes; their common line of intersection is the z axis and they are distributed
uniformly in the angle  . (The segments are like the skin on wedges of an apple, or the earth's surface
between successive meridians of longitude.) The segments are kept at fixed potentials ±V, alternately.

(a) Set up a series representation for the potential inside the sphere for the general case of 2n segments,
and carry the calculation of the coefficients in the series far enough to determine exactly which
coefficients are different from zero. For the non-vanishing terms, exhibit the coefficients as an integral
over cos θ.

(b) For the special case of n = 1 (two hemispheres) determine explicitly the potential up to and
including all terms with l = 3. By a coordinate transformation verify that this reduces to result (3.36) of
Section 3.3.

SOLUTION:
There is no charge present, so we seek to solve Laplace's equation. In spherical coordinates this
becomes:

1 ∂2 ∂2 
∇ 2 =0 
r ∂r 2
r 
1
r 2 sin  ∂ 
∂ sin  ∂  1
∂ 
r 2 sin 2  ∂ 2
=0

Using the method of separation of variables, and when the full azimuthal range needs a valid solution,
the general solution is expressed in terms of the spherical harmonics Ylm:

∞ l
 r , ,=∑ ∑  Al , m r l B l ,m r −l−1 Y lm  , 
l =0 m=−l

In this problem, we require a valid solution at the origin, so that we must have Bl = 0 to keep those
terms from blowing up. The solution now becomes:

∞ l
 r , ,=∑ ∑ Al ,m r l Y lm  , 
l =0 m=−l

The boundary condition on the surface of the sphere is described mathematically as:
 { }
2i  2 i1
 V if
n n
 r=a=V  where V = where i is any of 0,1,...(n-1)
 2i1  2i 2
−V if 
n n

We apply this boundary condition:

∞ l
V =∑ ∑ Al , m al Y lm  , 
l =0 m=−l

Multiply both sides by Y *l ' m'  ,  and integrate over the surface of the sphere:

2  ∞ l 2 

∫ ∫ V Y *
l'm'  ,  sin d  d =∑ ∑ Al , m a l
∫ ∫ Y *l ' m '  , Y lm  , sin  d  d 
0 0 l =0 m=−l 0 0

Use the orthogonality of the spherical harmonics to pick one term from the double series:
2  ∞ l

∫ ∫ V Y *
l'm'  ,  sin d  d =∑ ∑ Al , m a l  l 'l  m ' m
0 0 l =0 m=−l

2 

∫ ∫ V Y *l m  ,sin  d  d = Al ,m a l
0 0

2 
Al , m=a−l ∫ ∫ V Y *l m  , sin  d  d 
0 0

Expand the definition of the spherical harmonics:

2 
Al , m=a−l
  2 l 1 l −m!
4  l m! 0
∫ V  e−i m  d ∫ P ml cos  sin  d 
0

Break the integral over the azimuthal angle into a sum over 2n integral pieces and plug in the explicit
value of the potential on the boundary:

[∑ ∫ ]
 2 n−1  j1  /n
Al , m=a−l
  2 l 1 l −m!
4  l m! ∫ 0
P ml  cos sin  d 
j =0 j/n
V  e−i m  d 

[∑  ∫ ]
 n−1  2 j1  /n  2 j2  / n

Al , m=a−l
  2 l 1 l −m!
∫ P m  cos sin  d 
4  l m! 0 l j=0 2 j/n
V e −i m  d  ∫
2 j 1  / n
−V  e−i m  d 

[∑  ]
 n−1  2 j1  /n  2 j2  /n

Al , m=V a −l

  2 l1 l−m!
∫ P m  cos sin  d 
4  lm! 0 l j=0
∫
2 j/n
−i m 
e d − ∫
 2 j1 / n
e−i m  d 
Let us call the part in brackets Bm and work it out separately:

 
n−1  2 j 1 / n  2 j2  / n

B m= ∑ ∫ e−i m  d − ∫ e−i m  d 
j=0 2 j  /n 2 j 1  / n

For m = 0 this reduces to:

 
n−1 2 j 1  / n  2 j2  / n

B0 =∑ ∫ d − ∫ d
j=0 2 j / n 2 j 1  / n

B0 =0 and thus Al , 0=0

For m ≠ 0 we have:
n−1
i
B m= ∑  e −i m  2 j1  /n−e−i m 2 j  /n −e−i m 2 j 2 / ne −i m  2 j1  /n 
m j=0

n−1
i
B m= ∑ e−i m 2 j  / n  2 e −i m  / n−1−e−i m 2  /n 
m j=0

n−1
−i 2
B m= ∑ e−i m  2 j  /n  e−i m/ n−1 
m j=0

m
Upon close inspection, if =k where k is some integer k = ...,-2, -1, 0, 1, 2,... then
2n
n−1
−i 2
B m= ∑ e−i m  2 j  /n  e−i 2 k −1 
m j=0

n−1
i 2
B m= ∑
m j=0
 1−1 

m
B m=0 if =k
2n

The only terms that do not vanish are the ones where m/2n does not equate to an integer. Thus the
terms that vanish are m=0,±2 n ,±4 n , ...

The final solution becomes:

l l


∞
r
 r , ,=∑ ∑ Al ,m Y lm  , 
l =0 m=−l a
 

 
n−1
2 l1 l−m ! i −i m / n 2
where Al , m=−V e −1  ∑ e −i m  2 j  /n ∫ P ml cos sin  d 
4 lm ! m j=0 0

and Al , m=0 for m=0,±2 n ,±4 n...

(b) For the special case of n = 1 (two hemispheres) determine explicitly the potential up to and
including all terms with l = 3. By a coordinate transformation verify that this reduces to result (3.36) of
Section 3.3.
l l


∞
r
 r , ,=∑ ∑ Al ,m Y lm  , 
l =0 m=−l ,odd a

where Al , m=−V
  2 l1 l−m ! i
4 lm ! m −1
4 ∫ P ml  x dx

Expand this into terms up to l = 3:

1 2 2 3 3
 r , ,= ∑
m=−1,odd
A1, m 
r
a
Y 1 m  ,  ∑ A2,m
m=−2,odd
r
a  Y 2 m  ,  ∑
m =−3, odd
A3, m 
r
a
Y 3 m  ,  ...

where the l = 0 term automatically drops out due to the fact that only coefficients for odd m are non-
vanishing.
2
 r , ,=
3
r
a[ A1,−1 Y 1,−1A1,1 Y 1, 1] 
r
a 
[ A2,−1 Y 2−1A2,1 Y 2 1 ]

 
r
a
[ A3,−3 Y 3,−3A3,−1 Y 3,−1A3,1 Y 3,1 A3,3 Y 3, 3 ] ...

We must calculate the coefficients explicitly by expanding the definition of the Legendre polynomials
and doing the integrals:

A1,−1=iV
 3
2
, A1,1 =i V
 3
2

A2,−1=0 , A2,1 =0

A3,−3=iV
 35 
256
, A3,3 =i V
 35 
256

A3,−1 =iV
 21 
256
, A3,1 =iV
 21 
256
We plug these back in:

 r , ,=   [
r
a
iV
3
3
2
Y 1,−1  ,Y 1,1  ,]


r
a  [
iV
35 
256
Y 3,−3Y 3, 3  ,
21
256  ]
Y 3,−1 Y 3,1   ...

3
3
 r , ,= V
2
r
a 
sin  sin 
r
a
V
64  [
35 3 21 2
sin  sin3  sin  5 cos −1 sin ...
64 ]
For n = 1 only, this problem is actually azimuthally symmetric if we make a coordinate transformation:

cos  '=sin  sin 

3
3
 r , ,= V
2
r
a 
cos  '−
7 r
8 a
V
5
2  [
3 3
cos  '− cos  ' ...
2 ]
3
3
 r , ,= V
2
r

P cos  '−
a 1
7 r
8 a  V P 3 cos  ' ...

This matches the solution found in the book using an azimuthal symmetry approach.