Jackson 3.

1 Homework Problem Solution

Dr. Christopher S. Baird
University of Massachusetts Lowell
PROBLEM:
Two concentric spheres have radii a, b (b > a) and each is divided into two hemispheres by the same
horizontal plane. The upper hemisphere of the inner sphere and the lower hemisphere of the outer
sphere are maintained at potential V. The other hemispheres are at zero potential. Determine the
potential in the region a r b as a series in Legendre polynomials. Include terms at least up to l = 4.
Check your solution against known results in the limiting case b , and a 0.
SOLUTION:
The geometry of the problem can be sketched as a cross-section of the spheres:
Because of the spherical geometry of the problem and the absence of charge, we choose to solve the
Laplace equation in spherical coordinates:

2
4=0
1
r

2
r
2
( r 4)+
1
r
2
sin0

0
(
sin 0
4
0
)
+
1
r
2
sin
2
0

2
4

2
=0
Using the method of separation of variables leads to the general solution:
4(r ,0 ,)=

l
( A
l
r
l
+B
l
r
l 1
)( A
m=0
+B
m=0
) P
l
m=0
(cos0)
+

m0, l
( A
l
r
l
+B
l
r
l 1
)( A
m
e
i m
+B
m
e
i m
) P
l
m
(cos0)
The boundaries in this problem are all azimuthally symmetric, so that the solution for the electric
potential will not be a function of . The only way to make the general solution independent of is to
set m = 0 and B
m=0
= 0.
4(r ,0 ,)=

l
( A
l
r
l
+B
l
r
l 1
) P
l
m=0
(cos0)
V
V
0
0
x
z
a
b
In other problems where we must have a valid solution at the origin, B
l
must be zero to keep the
solution from blowing up at the origin. But in this problem, we do not seek a valid solution at the
origin, so we cannot use this restriction.
Apply the boundary condition:
4(r=a)=V
1
where
V
1
=V if 0/ 2 and V
1
=0 if 0>/ 2
V
1
=

l
( A
l
a
l
+B
l
a
l 1
) P
l
(cos0)
Multiply both sides by
P
l '
(cos 0)sin0
and integrate

V
1
P
l '
(cos 0)sin0 d 0=

l
( A
l
a
l
+B
l
a
l 1
)

P
l '
(cos0) P
l
(cos0) sin 0d 0
Use the orthogonality condition of Legendre polynomials:
0

P
l '
(cos0) P
l
(cos 0)sin0 d 0=
2
2l +1
6
l 'l

V
1
P
l '
(cos 0)sin0 d 0=

l
( A
l
a
l
+B
l
a
l 1
)
2
2l +1
6
l ' l
The Kronecker delta makes every term in the sum collapse to zero except when l = l'

V
1
P
l
(cos0) sin0d 0=( A
l
a
l
+B
l
a
l 1
)
2
2l +1
Flip the equation and plug in the explicit form of V
1
A
l
a
l
+B
l
a
l1
=
2l +1
2
V

0
/ 2
P
l
(cos0) sin0d 0
Make a change of variables: x=cos 0 , dx=sin 0 d 0
A
l
a
l
+B
l
a
l1
=
2l +1
2
V

0
1
P
l
( x) dx
Apply the boundary condition:
4(r=b)=V
2
where
V
2
=0 if 0/ 2 and V
2
=V if 0>/ 2
V
2
=

l
( A
l
b
l
+B
l
b
l 1
) P
l
(cos0)
Multiply both sides by
P
l '
(cos 0)sin0
and integrate

V
2
P
l '
(cos0) sin0d 0=

l
( A
l
b
l
+B
l
b
l 1
)

P
l
(cos 0) P
l '
(cos0) sin 0d 0
Use the orthogonality condition of Legendre polynomials:
0

P
l '
(cos0) P
l
(cos 0)sin0 d 0=
2
2l +1
6
l 'l
A
l
b
l
+B
l
b
l1
=
2l +1
2

0

V
2
P
l
(cos0)sin0 d 0
Plug in the explicit form of V
2
A
l
b
l
+B
l
b
l1
=
2l +1
2
V

/ 2

P
l
( cos0)sin 0d 0
Make a change of variables: x=cos 0 , dx=sin 0 d 0
A
l
b
l
+B
l
b
l1
=
2l +1
2
V

1
0
P
l
( x) dx
Make a change of variables x -x and use the identity: P
l
(x)=(1)
l
P
l
( x)
A
l
b
l
+B
l
b
l1
=
2l +1
2
V (1)
l

0
1
P
l
( x) dx
The two equations in boxes form a system of equations that we can solve for A
l
and B
l
.
A
l
=
(a
l +1
+(1)
l
b
l+1
)
(a
2l+1
+b
2l +1
)
(
2l +1
2
)
V

0
1
P
l
( x)dx
B
l
=
a
l +1
b
2l +1
(1)
l
a
2l+1
b
l +1
a
2l +1
+b
2l+1
(
2l +1
2
)
V

0
1
P
l
( x)dx
The final solution is then:
4(r ,0 ,)=
V
2

l
( 2l +1)

0
1
P
l
( x) dx
(a
l +1
(1)
l
b
l+1
) r
l
a b(a
l
b
2l
(1)
l
a
2l
b
l
) r
l 1
|
a
2l+1
b
2l +1
P
l
(cos0)
We can do the integral, but we must be careful to do the l = 0 and the l>0 cases separately:
l = 0:

0
1
P
0
( x) dx=

0
1
dx=1
l > 0:

0
1
P
l
( x)dx=
1
2l+1

0
1
d
dx
P
l +1
( x)P
l 1
( x)| dx

0
1
P
l
( x)dx=
1
2l+1
P
l +1
( x)P
l 1
( x)|
0
1

0
1
P
l
( x)dx=
1
2l+1
P
l +1
(1)P
l 1
(1)P
l +1
(0)+P
l 1
(0)|

0
1
P
l
( x)dx=
1
2l+1
P
l 1
(0)P
l+1
(0)|
When the l = 0 term is taken out of the sum and the integral solutions plugged in, the final solution
becomes:
4(r ,0 ,)=
V
2
+
V
2

l =1

P
l 1
( 0)P
l+1
(0)|
(a
l +1
(1)
l
b
l +1
) r
l
a b( a
l
b
2l
(1)
l
a
2l
b
l
)r
l1
|
a
2l +1
b
2l +1
P
l
(cos0)
It should be noted that
P
l 1
(0)P
l+1
(0)|=0
when l is even, so that all the even terms drop out.
Because l is only odd, then (1)
l
=1
4(r ,0 ,)=
V
2
+
V
2

l=1,odd

P
l1
(0)P
l +1
(0)|
(a
l+1
+b
l +1
) r
l
a
l +1
b
l +1
(b
l
+a
l
)r
l1
|
a
2l +1
b
2l+1
P
l
(cos0)
This can be expressed explicitly as an expanded sum (up to term l = 4 is shown):
4(r ,0 ,)=
V
2
1+
(
3
2
)
(a
2
+b
2
) ra
2
b
2
(b+a)r
2
|
a
3
b
3
cos 0
+
(
7
16
)
(a
4
+b
4
)r
3
a
4
b
4
(b
3
+a
3
) r
4
|
a
7
b
7
(5cos
3
03cos 0)+... |
When we make a statement about the behavior of the solution as b , we are really making a
statement about b being very large compared to something. So we first get the solution in a form where
b is always the denominator of a ratio:
4(r ,0 ,)=
V
2
+
V
2

l=1, odd

P
l1
(0)P
l +1
(0)|
(( a/ b)
l +1
+1)( r/ b)
l
(1+(a/ b)
l
)(a/ r)
l +1
|
( a/b)
2l +1
1
P
l
(cos0)
Now we can make the statement that as b , (a/b) 0, and (r/b) 0 so that:
4(r ,0 ,)=
V
2
+
V
2

l=1, odd

P
l1
(0)P
l +1
(0)|
(
a
r
)
l +1
P
l
(cos 0)
This can be expressed explicitly as an expanded sum (up to term l = 4 is shown):
4(r ,0 ,)=
V
2

1+
(
3
2
)(
a
r
)
2
cos0+
(
7
16
)(
a
r
)
4
(5cos
3
03cos0)+...
|
We can check this by solving the problem outright. With azimuthal symmetry, the full angular range
included, and the condition that the solution does not blow up at infinity, the general solution becomes:
4(r ,0 ,)=A
0
+

l
B
l
r
l1
P
l
(cos0)
Apply the boundary condition (derived by physical arguments): 4(r -)=V / 2
V / 2=A
0
Apply the boundary condition:
4(r=a)=V
1
where
V
1
=V if 0/ 2 and V
1
=0 if 0>/ 2
V
1
=
V
2
+

l
B
l
a
l 1
P
l
( cos0)
Multiply both sides by
P
l '
(cos 0)sin0 d 0
where l' > 0 and integrate.

V
1
P
l
(cos0) sin0d 0=B
l
a
l1 2
2l +1
B
l
=
2l +1
2
a
l+1

V
1
P
l
(cos0) sin0d 0
B
l
=
2l +1
2
V a
l+1

0
1
P
l
( x)dx
B
l
=
V
2
a
l+1
P
l1
(0)P
l +1
(0)|
where l is odd
4(r ,0 ,)=
V
2
+
V
2

l=1,odd
P
l1
(0)P
l +1
(0)|
(
a
r
)
l+1
P
l
(cos 0)
This matches the solution found above.
The behavior of the original solution as a 0 is:
4(r ,0 ,)=
V
2
+
V
2

l=1,odd

P
l1
(0)P
l +1
(0)|
(a
l+1
+b
l +1
) r
l
a
l +1
b
l +1
(b
l
+a
l
)r
l1
|
a
2l +1
b
2l+1
P
l
(cos0)
4(r ,0 ,)=
V
2

V
2

l=1,odd

P
l1
(0)P
l +1
(0)|
(
r
b
)
l
P
l
(cos0)
This can be expressed explicitly as an expanded sum (up to term l = 4 is shown):
4(r ,0 ,)=
V
2

1
(
3
2
)
r
b
cos0
(
7
16
)(
r
b
)
3
(5cos
3
03cos0)+...
|

We can check this by solving the problem outright. With azimuthal symmetry, the full angular range
included, and the condition that the solution does not blow up at the origin, the general solution
becomes:
4(r ,0 ,)=

l
A
l
r
l
P
l
(cos0)
Apply the boundary condition:
4(r=b)=V
2
where
V
2
=0 if 0/ 2 and V
2
=V if 0>/ 2
V
2
=

l
A
l
b
l
P
l
( cos0)
Multiply both sides by
P
l '
(cos 0)sin0 d 0
and integrate.

V
2
P
l '
(cos0) sin0d 0=

l
A
l
b
l

P
l
(cos0) P
l '
(cos 0)sin0 d 0

V
2
P
l
(cos0)sin0 d 0=A
l
b
l 2
2l +1
A
l
=
2l +1
2
b
l

V
2
P
l
( cos0)sin 0d 0
A
l
=
2l +1
2
V b
l

1
0
P
l
( x) dx
A
l
=
2l +1
2
V b
l
(1)
l

0
1
P
l
( x) dx
A
l
=
V
2
b
l
P
l1
(0)P
l +1
(0)|
where l is odd, and
A
0
=
V
2
4(r ,0 ,)=
V
2

V
2

l=1,odd
P
l1
(0)P
l +1
(0)|
(
r
b
)
l
P
l
(cos0)
This matches the solution found above.