California Institute of Technology SP 2025

EE 151 Electromagnetic Engineering Issued: 04/15/2025

Problem Set 5 Due: 04/24/2025

Problem Number 1: Potential in the Region Between Two Concentric Spherical Surfaces
Consider two concentric spherical surfaces of radii a and b (a < b). The
inner sphere (radius a) is held at zero potential (V = 0), and the outer
sphere (radius b) is held at a potential given by

V (b, θ, ϕ) = V0 cos θ.

Find the potential V (r, θ, ϕ) for a ≤ r ≤ b.

Hint: Use the general solution of Laplace’s equation in spherical coordi-

nates with symmetry and impose the boundary conditions on r = a and
r = b.

Solution

We have the region a ≤ r ≤ b with two concentric spherical surfaces:


V (a, θ, ϕ) = 0,
V (b, θ, ϕ) = V cos θ.
0

There is no ϕ dependence in the boundary conditions, so we expect no ϕ dependence in the solution. We

aim to solve Laplace’s equation,
∇2 V = 0,
in spherical coordinates with the stated boundary conditions.
Because of spherical symmetry (apart from the cos θ angular dependence), the most general form of the
solution (without ϕ terms) in the region a ≤ r ≤ b is
∞ h
X i
V (r, θ) = Al rl + Bl r− (l+1) Pl (cos θ),
l=0

where Pl are the Legendre polynomials.

Notice that the boundary condition on the outer sphere, V (b, θ) = V0 cos θ, involves only the l = 1 term,
because cos θ = P1 (cos θ). Hence, all terms with l ̸= 1 must vanish. Therefore, the solution simplifies to
V (r, θ) = A1 r + B1 r−2 P1 (cos θ).
 

Recalling P1 (cos θ) = cos θ, we write

V (r, θ) = A1 r + B1 r−2 cos θ.
 

(i) At r = a:
A1 a + B1 a−2 cos θ = 0
 
V (a, θ) = 0 =⇒ for all θ.
This implies
A1 a + B1 a−2 = 0.
(ii) At r = b:
A1 b + B1 b−2 cos θ = V0 cos θ.
 
V (b, θ) = V0 cos θ =⇒
This implies
A1 b + B1 b−2 = V0 .

1
Hence, we have two linear equations: 
A1 a + B1 a−2 = 0,
A b + B b−2 = V .
1 1 0

From the first equation:

B1 = − A1 a3 .
Substitute this into the second equation:

A1 b − A1 a3 b−2 = V0 A1 b − a3 b−2 = V0 .
 
=⇒

Factor out b:
V0 V0 V0 V0 b2
A1 b 1 − (a3 /b3 ) = V0 ,
 
=⇒ A1 =  3 =   = 3 3 = .
b 1 − ab3 b b b−a b3 − a3
 3
 
b 1 − ab3 3

So
V0 b2
A1 = .
b3 − a3
Then
V0 b2 a3
B1 = − A1 a3 = − .
b3 − a3
Substitute these back:
V0 b2 V0 b2 a3 −2
 
V (r, θ) = A1 r + B1 r−2 cos θ =
 
r − 3 r cos θ.
b3 − a3 b − a3

Factor out V0 b2 /(b3 − a3 ) if desired:

V0 b2 h 3 −2
i
V (r, θ) = r − a r cos θ.
b3 − a3

V 0 b2  a3 
V (r, θ) = r − 2 cos θ, a ≤ r ≤ b.
b3 −a 3 r

Problem Number 2: Potential of a Cylindrical Shell

A long cylindrical shell of radius R carries a uniform surface charge σ0
on the upper half and an opposite charge −σ0 on the lower half. Find
the electric potential inside and outside the cylinder.

Solution

Using separation of variables in cylindrical coordinates:

∞
X
s (ak cos kϕ + bk sin kϕ) + s−k (ck cos kϕ + dk sin kϕ)
 k 
V (s, ϕ) = a0 + b0 ln(s) +
k=1

To find the potential inside the cylinder, we can eliminate coefficients with ln(s) or s−k terms, as there terms
would cause our solution to blow up at the origin. We are thus left with a general solution of:

2
 ∞
X
Vin (s, ϕ) = a0 + sk (ak cos kϕ + bk sin kϕ)
k=1

To find the potential outside the cylinder, we can eliminate coefficients with ln(s) and sk terms, as they
would cause our solution to blow up at s → ∞. This gives us:
∞
X
Vout (s, ϕ) = s−k (ck cos kϕ + dk sin kϕ)
k=1

Using continuity of potential at s = R, we have

∞
X ∞
X
Rk (ak cos kϕ + bk sin kϕ) = R−k (ck cos kϕ + dk sin kϕ),
k=1 k=1

so ck = R2k ak , dk = R2k bk . Relate potential to surface charge, we have the following boundary condition:
 
∂Vin ∂Vout ρ
− = ,
∂s ∂s s=R ϵ0
Therefore,
∞
(
X
k−1 σ0 /ϵ0 , for 0 < ϕ < π;
2kR (ak cos kϕ + bk sin kϕ) =
k=1
−σ0 /ϵ0 , for π < ϕ < 2π
Using Fourier’s trick, multiply by (cos lϕ)dϕ and integrate from 0 to 2π, we have
π 2π o
σ0 π
Z Z 2π
σ0 n sin lϕ sin lϕ
2lRl−1 πal = [ cos lϕdϕ − cos lϕdϕ] = − = 0; al = 0.
ϵ0 0 π ϵ0 l 0 l π
Multiply by (sin lϕ)dϕ and integrate from 0 to 2π, we have
(
π 2π
σ0 π
Z 2π
0, if l is even
Z
σ0 n cos lϕ cos lϕ o σ0
2lRl−1 πbl = [ sin lϕdϕ− sin lϕdϕ] = − + = (2−2 cos lπ) =
ϵ0 0 π ϵ 0 l 0 l π lϵ0 4σ0 /lϵ0 , if l is odd
Therefore, (
0, if l is even
bl = 2 l−1
2σ0 /πϵ0 l R , if l is odd
Finally, the potential is

(
2σ0 R X 1 (s/R)k , (s < R)
V (s, ϕ) = sin kϕ
πϵ0 k2 (R/s)k , (s > R).
k=1,3,5,...

Problem Number 3: Potential in a Grounded Box

Consider a cubical box (edge length a) that consists of 6 metal plates
that are electrically isolated from each other. Determine the potential
inside the box for the following scenarios:

a) Only the top surface/wall is biased with voltage V0 and the others
are grounded. (Note: what should the potential be at the center of
the box. Is your result consistent with this value?)
b) The top and bottom walls are biased with voltages V1 , V2 respectively.
The others are grounded.
c) All of the walls are biased with voltage V0 .

3
Solution
To solve this problem, we first need to find the general solution to Laplace’s equation in Cartesian coordi-
nates, which is:
∂2V ∂2V ∂2V
+ + =0
∂x2 ∂y 2 ∂z 2
We wish to find solutions of the form: V (x, y, z) = X(x)Y (y)Z(z). This yields:
1 ∂2X 1 ∂2Y 1 ∂2Z
2
+ 2
+ =0
X ∂x Y ∂y Z ∂z 2
Therefore,
1 ∂2X 1 ∂2Y 1 ∂2Z
2
= C1 , 2
= C2 , = C3 , with C1 + C2 + C3 = 0.
X ∂x Y ∂y Z ∂z 2
a) For the following case, we have the following boundary conditions:
(i) V =0 when x = 0,
(ii) V =0 when x = a,
(iii) V =0 when y = 0,
(iv) V =0 when y = a,
(v) V =0 when z = 0,
(vi) V = V0 when z=a
Thus, we want the solution to be sinusoidal in x and y, and exponential in z. To do this, we set:
C1 = −k 2 , C2 = −l2 , C3 = l2 + k 2 . This yields:
X(x) = Asin(kx) + Bcos(kx)
Y (y) = Csin(ly) + Dcos(ly)
√ √
k2 +l2 z k2 +l2 z
Z(z) = Ee + Ge−

Applying the boundary conditions:

(i) =⇒ B = 0;
nπ
(ii) =⇒ k = ;
a
(iii) =⇒ D = 0;
mπ
(iv) =⇒ l = ;
a
(v) =⇒ E + G = 0
p z
thus Z(z) = 2E sinh (π (n2 + m2 ) )
a
This yields:
∞ X
X ∞ p
V (x, y, z) = Cn,m sin(nπx/a) sin(mπy/a) sinh(π n2 + m2 z/a)
n=1 m=1

To find the coefficients Cn,m , we can use boundary condition (vi):

X∞ X ∞ h p i
V0 = Cn,m sinh(π n2 + m2 ) sin(nπx/a) sin(mπy/a)
n=1 m=1
Z aZ a
p 2
Cn,m sinh(π n2 + m2 ) = ( )2 Vo sin(nπx/a) sin(mπy/a)dxdy
a 0 0
16V0
= 2 if n, m are odd
π nm

4
 Therefore, the potential in the box is:
√
16V0 X X 1 sinh(π n2 + m2 z/a)
V (x, y, z) = 2 sin(nπx/a) sin(mπy/a) √
π n=1,3,5... m=1,3,5...
nm sinh(π n2 + m2 )

Since there cannot be any maxima/minima, other than on the boundaries, the potential at the center of
the box is the average of the potential at the boundaries. That is, V (x = a2 , y = a2 , z = a2 ) = V60 .
b) We can use superposition to find the solution. Consider the two cases:

(i) Only the top wall is biased with voltage V1 . All other walls are grounded.
(ii) Only the bottom wall is biased with voltage V2 . All other walls are grounded.
Let Vi , Vii be the potential in the box corresponding to case (i), (ii) above. Then the total potential in
the box is

V (x, y, z) = Vi + Vii
" √ √ #
16 X X 1 sinh(π n2 + m2 az ) sinh(π n2 + m2 a−z
a )
= 2 sin(nπx/a) sin(mπy/a) V1 √ + V2 √
π n=1,3,5... m=1,3,5...
nm sinh(π n2 + m2 ) sinh(π n2 + m2 )

c) Since any solution to Laplace’s equation cannot have any maxima/minima, except on the boundaries, the
potential in the box must satisfy V (x, y, z) = V0 .
However, using superposition, an analytical equation can be determined to be:
" √ √ #
16V0 X 1 sinh(π n2 + m2 az ) sinh(π n2 + m2 a−z a )
V (x, y, z) = 2 sin(nπx/a) sin(mπy/a) √ + √
π nm sinh(π n 2 + m2 ) sinh(π n2 + m2 )
n,m odd
" √ √ #
1 sinh(π n2 + m2 ay ) sinh(π n2 + m2 a−y a )
+ sin(nπx/a) sin(mπz/a) √ + √
nm sinh(π n2 + m2 ) sinh(π n2 + m2 )
" √ √ #
1 sinh(π n2 + m2 xa ) sinh(π n2 + m2 a−x a )
+ sin(nπy/a) sin(mπz/a) √ + √
nm sinh(π n2 + m2 ) sinh(π n2 + m2 )

This solution converges numerically to V (x, y, z) = V0 , as expected.