Jackson 3.

12 Homework Problem Solution

Dr. Christopher S. Baird
University of Massachusetts Lowell

PROBLEM:
An infinite, thin, plane sheet of conducting material has a circular hole of radius a cut in it. A thin, flat
disc of the same material and slightly smaller radius lies in the plane, filling the hole, but separated
from the sheet by a very narrow insulating ring. The disc is maintained at a fixed potential V, while the
infinite sheet is kept at zero potential.

(a) Using appropriate cylindrical coordinates, find an integral expression involving Bessel functions for
the potential at any point above the plane.

(b) Show that the potential a perpendicular distance z above the center of the disc is

Φ 0 (z )=V 1−
( z
√ a +z 2
2 )
(c) Show that the potential a perpendicular distance z above the edge of the disc is

Φ a (z )=
V
2(1−
kz
πa
K (k )
)
where k = 2a/(z2 + 4a2)1/2, and K(k) is the complete elliptic integral of the first kind.

SOLUTION:
(a) The general solution to the Laplace equation in cylindrical coordinates is:

Φ(ρ , ϕ , z )=( A0,0+B 0,0 ln ρ)(C 0,0+ D0,0 ϕ)( F 0, 0+G0, 0 z )

iν ϕ
+∑ ( Aν , 0 ρ + Bν ,0 ρ )(C ν ,0 e +D ν ,0 e )(F ν ,0+G ν ,0 z )
ν −ν −i ν ϕ

ν≠0

+∑ ( A0, k J 0 (k ρ)+B0, k N 0 (k ρ))(C 0,k +D0, k ϕ)( F 0,k e k z +G 0,k e−k z )

k≠0
+∑ ∑ ( Aν , k J ν (k ρ)+B ν ,k N ν ( k ρ))( C ν ,k e i ν ϕ+ Dν , k e−i ν ϕ )( F ν ,k e k z +G ν , k e−k z )
ν≠0 k≠0

The problem has azimuthal symmetry, immediately dictating that ν = 0 and the function be single-
valued in the azimuthal angle, leading to:

Φ(ρ , ϕ , z )=( A0+B 0 ln ρ)( F 0+G0 z )+∑ ( Ak J 0 (k ρ)+B k N 0 (k ρ))( F k e k z +G k e−k z )

k ≠0

The solution must be finite as z approaches infinity, telling us that Fk = 0 and G0 =0.
Also, the solution must be finite along the z axis (ρ = 0), telling us that Bk = 0 and B0 = 0. Our solution
is now:
 Φ(ρ , ϕ , z )=∑ A k J 0 (k ρ)e−k z
k

Typically at this point, we would apply a boundary condition at a finite radius, but there is not one for
this problem. With no constraint on the radius, k becomes a continuous spectrum and not a discrete set.
∞
Φ(ρ , ϕ , z )=∫ A( k ) J 0 (k ρ)e−k z dk
0

Apply the last boundary condition:

Φ( z=0)=V (ρ)
∞
V (ρ)=∫ A( k ) J 0 (k ρ)dk
0

We can take advantage of the orthogonality statement to solve this:

∞
1
∫ x J 0 (k x) J 0 ( k ' x)d x= k δ( k '−k ) (Jackson 3.108)
0

∞
A( k )=k ∫ V (ρ)ρ J 0 ( k ρ)d ρ
0

Insert the actual boundary condition in this problem:

a
A( k )=V k ∫ J 0 (k ρ)ρ d ρ
0

A( k )=V a J 1 (ka)

The final solution is:

∞
Φ(ρ , ϕ , z )=V a ∫ J 1 ( ka) J 0 (k ρ)e −k z dk
0

(b) The potential a perpendicular distance z above the center of the disc is:
∞
Φ(ρ=0)=V a ∫ J 1 (ka) J 0 (0)e−k z dk
0

∞
Φ(ρ=0)=V a ∫ J 1 (ka) e−k z dk
0

∞
Φ(ρ=0)=V ∫ J 1( x) e− x z /a dx
0
 2π
1
Use J 1( x)= ∫
2πi 0
e i (x cos θ+θ) d θ

2π ∞
1
Φ(ρ=0)=
2πi
V ∫ e i θ∫ e (i cosθ −z / a) x dx d θ
0 0

2π
1 1
Φ(ρ=0)=
2π
V ∫ e i θ ( cos θ+i z /a)
dθ
0

2π 2π
1 cos 2 θ+sin θ z /a −cos θ z /a+sin θ cos θ
Φ(ρ=0)= V [∫ 2 2 2
d θ+i ∫ d θ]
2π 0 (cos θ+z / a ) 0 (cos 2 θ+ z 2 / a 2 )

Several pieces go away due to symmetry.

2π
1 cos 2 θ
Φ(ρ=0)=
2π
V ∫ 2 2
(cos θ+z /a )
2
dθ
0

[ ]
π/2
2 z2 1
2 ∫
Φ(ρ=0)=V 1− dθ
π a 0 cos θ+z 2 /a 2
2

Substitute: u=
1
tan θ , cos 2
θ=
1
, d θ= √1+a 2 / z 2 du
√1+a 2 / z 2 2 2
1+(1+a / z )u
2
1+(1+a 2 / z 2 ) u2

[ ]
∞
2 z
Φ(ρ=0)=V 1− ∫ 1 2 du
π √ z +a 0 1+u
2 2

Φ(ρ=0)=V 1−
[ 2
π
z
√ z +a
2 2
−1 −1
[tan (∞)−tan (0)]
]
Φ(ρ=0)=V 1−
[ z
√ a +z 2 2 ]
(c) We can now find the potential a perpendicular distance z above the edge of the disc.
∞
Φ(ρ=a)=V a ∫ J 1 (ka ) J 0 ( k a)e−k z dk
0

∞
Φ(ρ=a)=V ∫ J 1 (x ) J 0 ( x) e−x z /a dx
0

2π 2π
1 1
Use J 1( x)= ∫
2πi 0
e i (x cos θ+θ) d θ and J 0 (x )= ∫
2π 0
ei x cos θ ' d θ '
 2π 2π ∞
1 1
Φ(ρ=a)= V ∫ ∫ e i θ∫ e (i cosθ +i cosθ '− z /a ) x dx d θ d θ '
2π 2πi 0 0 0

2π 2π
1 1 1
Φ(ρ=a)= V ∫ ei θ ∫ d θ'd θ
2π 2π 0 0 cos θ '+cos θ+i z /a

Many terms go away due to symmetry.

π π
11 cos θ cos θ '+cos 2 θ
Φ(ρ=a)= V ∫∫ d θ'd θ
π π 0 0 (cos θ '+cos θ)2+ z 2 /a 2

[ ]
π π
V 1 1 cos 2 θ '−cos 2 θ+z 2 /a 2
Φ(ρ=a)= 1− [∫ ∫ d θ d θ ']
2 π 0 π 0 ( cosθ '+cos θ)2+z 2 /a 2

( )
π
V 1 dθ
Φ= 1− ∫
2 π 0 √1+4 (a / z )2 cos2 θ

2
Substitute: k =
√( z /a)2 +4

( )
π /2
V kz dθ
Φ=
2
1−
πa
∫
0 √1−k 2 sin2 θ
π/ 2
Φ=
V
2
1−(kz
πa
K (k ) ) where K (k )= ∫
0
dθ
√ 1−k 2 sin 2 θ
Here K(k) is the complete elliptic integral of the first kind. Note that there is some ambiguity in the
literature, where some sources cite the “complete elliptic integral of the first kind” as:
π/ 2
dθ
K (k )= ∫
0 √ 1−k 2 sin 2 θ
and some cite it as:
π/ 2
dθ
K (k )= ∫
0 √ 1−k sin 2 θ
The first definition is used here because it gives the solution that exactly matches Jackson. In the end, it
should not matter which one you use as long as you are careful, consistent, and make clear which one
you are using. But using the one definition when the other one is meant leads to non-physical results.