Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
Homework Assignment #4 — Solutions
Textbook problems: Ch. 3: 3.4, 3.6, 3.9, 3.10
3.4 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 φ. (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
nonvanishing terms, exhibit the coeﬃcients as an integral over cos θ.
The general spherical harmonic expansion for the potential inside a sphere of
radius a is
Φ(r, θ, φ) =
¸
l,m
α
lm
r
a
l
Y
lm
(θ, φ)
where
α
lm
=
V (θ, φ)Y
∗
lm
(θ, φ)dΩ
In this problem, V (θ, φ) = ±V is independent of θ, but depends on the azimuthal
angle φ. It can in fact be thought of as a square wave in φ
2 π
V
V −
ϕ
n =4
This has a familiar Fourier expansion
V (φ) =
4V
π
∞
¸
k=0
1
2k + 1
sin[(2k + 1)nφ]
This is already enough to demonstrate that the m values in the spherical harmonic
expansion can only take on the values ±(2k+1)n. In terms of associated Legendre
polynomials, the expansion coeﬃcients are
α
lm
=
2l + 1
4π
(l −m)!
(l + m)!
2±
0
V (φ)e
−imφ
dφ
1
−1
P
m
l
(x) dx
=
4V
π
2l + 1
4π
(l −m)!
(l + m)!
∞
¸
k=0
1
2k + 1
2π
0
sin[(2k + 1)nφ]e
−imφ
dφ
×
1
−1
P
m
l
(x) dx
= −4iV
2l + 1
4π
(l −m)!
(l + m)!
∞
¸
k=0
δ
m,(2k+1)n
−δ
m,−(2k+1)n
2k + 1
1
−1
P
m
l
(x) dx
Using P
−m
l
(x) = (−)
m
[(l −m)!/(l +m)!]P
m
l
(x), we may write the nonvanishing
coeﬃcients as
α
l,−(2k+1)n
= (−)
n+1
α
l,(2k+1)n
= −
4iV
2k + 1
2l + 1
4π
(l −(2k + 1)n)!
(l + (2k + 1)n)!
1
−1
P
(2k+1)n
l
(x) dx
(1)
for k = 0, 1, 2, . . .. Since l ≥ (2k + 1)n, we see that the ﬁrst nonvanishing term
enters at order l = n. Making note of the parity of associated Legendre polyno
mials, P − l
m
(−x) = (−)
l+m
P
m
l
(x), we see that the nonvanishing coeﬃcients
are given by the sequence
α
n,n
, α
n+2,n
, α
n+4,n
, α
n+6,n
, . . .
α
3n,3n
, α
3n+2,3n
, α
3n+4,3n
, . . .
α
5n,5n
, α
5n+2,5n
, α
5n+4,5n
, . . .
.
.
.
b) For the special case of n = 1 (two hemispheres) determine explicitly the potential
up to an including all terms with l = 3. By a coordinate transformation verify
that this reduces to result (3.36) of Section 3.3.
For n = 1, explicit computation shows that
1
−1
P
1
1
(x) dx = −
π
2
,
1
−1
P
1
3
(x) dx = −
3π
16
,
1
−1
P
3
3
(x) dx = −
45π
8
Inserting this in to (1) yields
α
1,−1
= α
1,1
= iV
3π
2
α
3,−1
= α
3,1
= iV
21π
256
, α
3,−3
= α
3,3
= iV
35π
256
Hence
Φ = iV
¸
r
a
3π
2
(Y
1,1
+ Y
1,−1
)
+
r
a
3
21π
256
(Y
3,1
+ Y
3,−1
) +
35π
256
(Y
3,3
+ Y
3,−3
)
+· · ·
= −2V
¸
r
a
3π
2
Y
1,1
+
r
a
3
21π
256
Y
3,1
+
35π
256
Y
3,3
+· · ·
= 2V
¸
r
a
3
4
sin θ e
iφ
+
r
a
3
21
128
sin θ(5 cos
2
θ −1)e
iφ
+
35
128
sin
3
θ e
3iφ
+· · ·
= V
¸
r
a
3
2
sin θ sin φ
r
a
3
7
128
3 sin θ(5 cos
2
θ −1) sin φ + 5 sin
3
θ sin
3
φ
+· · ·
(2)
To relate this to the previous result, we note that the way we have set up the
wedges corresponds to taking the ‘top’ of the +V hemisphere to point along the
ˆ y axis. This may be rotated to the ˆ z
axis by a 90
◦
rotation along the ˆ x axis.
Explicitly, we take
ˆ y = ˆ z
, ˆ z = −ˆ y
, ˆ x = ˆ x
or
sin θ sin φ = cos θ
, cos θ = −sin θ
sin φ
, sin θ cos φ = sin θ
cos φ
Noting that sin 3φ = −sin
3
φ + 3 sin φcos
2
φ, the last line of (2) transforms into
Φ = V
¸
r
a
3
2
cos θ
+
r
a
3
7
128
3 cos θ
(5 sin
2
θ
sin
2
φ
−1)
+ 5(−cos
3
θ
+ 3 cos θ
sin
2
θ
sin
2
φ
)
+· · ·
= V
¸
3
2
r
a
cos θ
−
7
8
r
a
3
1
2
(5 cos
3
θ
−3 cos θ
) +· · ·
= V
¸
3
2
r
a
P
1
(cos θ
) −
7
8
r
a
3
P
3
(cos θ
) +· · ·
which reproduces the result (3.36).
3.6 Two point charges q and −q are located on the z axis at z = +a and z = −a,
respectively.
a) Find the electrostatic potential as an expansion in spherical harmonics and powers
of r for both r > a and r < a.
The potential is clearly
Φ =
q
4π
0
1
x −a 
−
1
x +a 
where a = aˆ z points from the origin to the positive charge. Using the spherical
harmonic expansion
1
x −x

= 4π
¸
l,m
1
2l + 1
r
l
<
r
l+1
>
Y
∗
lm
(ˆ x
)Y
lm
(ˆ x)
as well as a = aˆ z, we obtain
Φ =
q
0
¸
l,m
1
2l + 1
r
l
<
r
l+1
>
[Y
∗
lm
(0, φ
) −Y
∗
lm
(π, φ
)]Y
lm
(θ, φ) (3)
Noting that Y
lm
(0, φ) ∼ P
m
l
(1) and that P
m
l
(1) = δ
m,0
we see that only terms
with m = 0 contribute. This is also obvious from symmetry. Since
Y
l0
(0, φ) = (−)
l
Y
l0
(π, φ) =
2l + 1
4π
the potential (3) becomes
Φ =
q
4π
0
∞
¸
l=0
[1 −(−)
l
]
r
l
<
r
l+1
>
4π
2l + 1
Y
l0
(θ, φ)
=
q
2π
0
¸
l odd
r
l
<
r
l+1
>
P
l
(cos θ)
(4)
b) Keeping the product qa = p/2 constant, take the limit of a → 0 and ﬁnd the
potential for r = 0. This is by deﬁnition a dipole along the z axis and its
potential.
Since we will take a →0, we have r
<
= a and r
>
= r. This yields an expansion
of (4)
Φ =
qa
2π
0
r
2
∞
¸
k=0
a
r
2k
P
2k+1
(cos θ)
Setting qa = p/2 and taking a → 0, only the k = 0 term survives in the sum.
The result is
Φ =
p
4π
0
1
r
2
P
1
(cos θ) =
p
4π
0
cos θ
r
2
(5)
which is the potential due to a dipole.
c) Suppose now that the dipole of part b) is surrounded by a grounded spherical shell
of radius b concentric with the origin. By linear superposition ﬁnd the potential
everywhere inside the shell.
To account for the spherical shell, we add to (5) a solution to the (homogeneous)
Laplace’s equation. For an inside solution, we have
Φ =
p
4π
0
¸
1
r
2
P
1
(cos θ) +
∞
¸
l=0
A
l
r
l
P
l
(cos θ)
¸
The boundary condition Φ(r = b) = 0 corresponds to having
∞
¸
l=0
A
l
b
l+2
P
l
(cos θ) = −P
1
(cos θ)
Since the Legendre polynomials form an orthonormal set, the only term that can
show up on the left hand side is the l = 1 term. We then take A
1
= −1/b
l+2
, and
the resulting solution is
Φ =
p
4π
0
1
r
2
−
r
b
3
cos θ
3.9 A hollow right circular cylinder of radius b has its axis coincident with the z axis and its
ends at z = 0 and z = L. The potential on the end faces is zero, while the potential
on the cylindrical surface is given as V (φ, z). Using the appropriate separation of
variables in cylindrical coordinates, ﬁnd a series solution for the potential anywhere
inside the cylinder.
The general solution obtained by separation of variables has the form
Φ(ρ, φ, z) =
¸
J
m
(kρ) or N
m
(kρ)
¸
e
±imφ
¸
e
±kz
¸
However, since the potential vanishes on the endcaps, it is natural to take k →ik
so that the z function obeying boundary conditions is sin(nπz/L). The result is
to use the modiﬁed Bessel functions I
m
(kρ) and K
m
(kρ) instead. However, for
the solution to be regular at ρ = 0 we discard the K
ν
(kρ) functions, which blow
up at vanishing argument. The resulting series expression for the potential is
Φ(ρ, φ, z) =
∞
¸
m=0
∞
¸
n=1
I
m
nπ
L
ρ
a
mn
sin mφ + b
mn
cos mφ
sin
nπ
L
z
(6)
In order to satisfy the boundary conditions on the cylindrical surface, we need to
have
V (φ, z) =
¸
m,n
I
m
nπb
L
a
mn
sin mφ + b
mn
cos mφ
sin
nπ
L
z
This is a double Fourier series in φ and z. As a result, the Fourier coeﬃcients are
a
mn
b
mn
I
m
nπb
L
=
1
π
2π
0
dφ
2
L
L
0
dz V (φ, z)
sin mφ
cos mφ
sin
nπ
L
z
with the caveat that b
0,n
must be divided by two. This can be rewritten as
a
mn
b
mn
=
2
πLI
m
(nπb/L)
2π
0
dφ
L
0
dz V (φ, z)
sin mφ
cos mφ
sin
nπ
L
z
(7)
(where b
0,n
has to be divided by two).
3.10 For the cylinder in Problem 3.9 the cylindrical surface is made of two equal half
cylinders, one at potential V and the other at potential −V , so that
V (φ, z) =
V for −π/2 < φ < π/2
−V for π/2 < φ < 3π/2
a) Find the potential inside the cylinder.
To obtain the potential, we want to ﬁnd the coeﬃcients a
mn
and b
mn
of the
expansion (6) and (7). Noting ﬁrst that V (φ, z) = V (φ) = ±V is an even function
of φ, we see that all the a
mn
coeﬃcients vanish. We are left with
b
mn
=
2V
πLI
m
(nπb/L)
¸
π/2
−π/2
−
3π/2
π/2
¸
dφ cos mφ
L
0
dz sin
nπz
L
=
2V
π
2
I
m
(nπb/L)
4 sin(mπ/2)
m
1 −(−)
n
n
(m = 0)
This is nonvanishing only when both m and n are odd. Introducing m = 2k +1
and n = 2l + 1, we have
b
2k+1,2l+1
=
16V
π
2
I
2k+1
((2l + 1)πb/L)
(−)
k
(2k + 1)(2l + 1)
Inserting this into (6) yields
Φ =
16V
π
2
∞
¸
k=0
∞
¸
l=0
(−)
k
(2k + 1)(2l + 1)
I
2k+1
(
(2l+1)πρ
L
)
I
2k+1
(
(2l+1)πb
L
)
cos(2k+1)φsin
(2l + 1)πz
L
(8)
b) Assuming L b, consider the potential at z = L/2 as a function of ρ and φ and
compare it with twodimensional Problem 2.13.
For L b both ρ/L and b/L are much less than one. This allows us to use a
small argument expansion of the modiﬁed Bessel function
I
ν
(x) ≈
1
Γ(ν + 1)
x
2
ν
In addition, for z = L/2 we have
sin
(2l + 1)πz
L
= sin(l +
1
2
)π = (−)
l
Hence in this limit (8) becomes
Φ =
16V
π
2
¸
k,l
(−)
k
2k + 1
(−)
l
2l + 1
ρ
b
2k+1
cos(2k + 1)φ
=
16V
π
2
¸
¸
l
(−)
l
2l + 1
¸
¸
¸
k
(−)
k
2k + 1
ρ
b
e
iφ
2k+1
¸
Noting the Taylor series expansion for arctan
tan
−1
z =
¸
n
(−)
n
2n + 1
z
2n+1
we arrive at
Φ =
16V
π
2
tan
−1
(1)tan
−1
ρ
b
e
iφ
=
4V
π
tan
−1
ρ
b
e
iφ
To calculate tan
−1
z we recall that
tan
−1
a + tan
−1
b = tan
−1
a + b
1 −ab
Hence
tan
−1
z =
1
2
(tan
−1
z + tan
−1
z
∗
) =
1
2
tan
−1
z + z
∗
1 −z
2
For z = (ρ/b)e
iφ
we ﬁnd
Φ =
2V
π
tan
−1
2(ρ/b) cos φ
1 −(ρ/b)
2
=
2V
π
tan
−1
2bρ cos φ
b
2
−ρ
2
which reproduces the answer to Problem 2.13 (where V
1
= −V
2
= V ).
αn+2. 2 −1 Inserting this in to (1) yields α1. .n .1 = iV α3. αn+6. .1 = iV 3π 2 21π . α5n+2. the expansion coeﬃcients are αlm = = 2l + 1 (l − m)! 4π (l + m)! 4V π 2± 1 V (φ)e−imφ dφ 0 ∞ −1 Plm (x) dx sin[(2k + 1)nφ]e−imφ dφ 2l + 1 (l − m)! 4π (l + m)! k=0 1 2k + 1 2π 0 1 × −1 Plm (x) dx 1 = −4iV 2l + 1 (l − m)! 4π (l + m)! ∞ k=0 δm. Making note of the parity of associated Legendre polynomials. αn+4.3n . α5n.(2k+1)n − δm.polynomials.(2k+1)n =− 4iV 2k + 1 2l + 1 (l − (2k + 1)n)! 4π (l + (2k + 1)n)! 1 −1 Pl (2k+1)n (1) (x) dx for k = 0. we may write the nonvanishing coeﬃcients as αl.n . α3n+2.3n . . we see that the ﬁrst nonvanishing term enters at order l = n.. . we see that the nonvanishing coeﬃcients are given by the sequence αn. P − lm (−x) = (−)l+m Plm (x). 2.3. . . . . .n .−1 = α3.5n . 256 1 1 1 P3 (x) dx = − −1 1 3 P3 (x) dx = − −1 3π .−3 = α3.36) of Section 3.5n . . b) For the special case of n = 1 (two hemispheres) determine explicitly the potential up to an including all terms with l = 3. . 1. α3n+4.5n .−1 = α1.n . . Since l ≥ (2k + 1)n. For n = 1. 16 45π 8 α3. .−(2k+1)n 2k + 1 Plm (x) dx −1 Using Pl−m (x) = (−)m [(l − m)!/(l + m)!]Plm (x).−(2k+1)n = (−)n+1 αl. By a coordinate transformation verify that this reduces to result (3. . α3n.3n . α5n+4. explicit computation shows that π 1 P1 (x) dx = − .3 = iV 35π 256 . .
1 + 2 a 21π Y3. This may be rotated to the z axis by a 90◦ rotation along the x axis.3 + Y3.3 256 + ··· 3π r Y1. we note that the way we have set up the wedges corresponds to taking the ‘top’ of the +V hemisphere to point along the y axis. the last line of (2) transforms into Φ=V r a 3 r cos θ + 2 a 3 7 3 cos θ (5 sin2 θ sin2 φ − 1) 128 + 5(− cos3 θ + 3 cos θ sin2 θ sin2 φ ) + · · · =V 3 r 7 r 31 cos θ − (5 cos3 θ − 3 cos θ ) + · · · 2 a 8 a 2 3 r 7 r 3 =V P1 (cos θ ) − P3 (cos θ ) + · · · 2 a 8 a which reproduces the result (3.−1 ) 2 + = −2V = 2V r a r a r a 3 21π (Y3.1 + Y3. sin θ cos φ = sin θ cos φ Noting that sin 3φ = − sin3 φ + 3 sin φ cos2 φ. ˆ ˆ ˆ Explicitly. cos θ = − sin θ sin φ . ˆ ˆ z = −ˆ .1 + 256 3 sin θ eiφ 4 r 3 21 35 + sin θ(5 cos2 θ − 1)eiφ + sin3 θ e3iφ a 128 128 + ··· =V r a 3 sin θ sin φ 2 r 3 7 3 sin θ(5 cos2 θ − 1) sin φ + 5 sin3 θ sin3 φ + · · · a 128 (2) To relate this to the previous result.−1 ) + 256 3 35π (Y3. we take y=z.Hence Φ = iV r a 3π (Y1. ˆ y x=x ˆ ˆ or sin θ sin φ = cos θ .1 + Y1. .−3 ) + · · · 256 35π Y3.36).
0 we see that only terms with m = 0 contribute. This is by deﬁnition a dipole along the z axis and its potential. φ) 2l + 1 (4) 0 l l r< Pl (cos θ) rl+1 odd > b) Keeping the product qa = p/2 constant. Since we will take a → 0. φ )]Ylm (θ. respectively. φ) = (−)l Yl0 (π. φ) l+1 lm 2l + 1 r> l r< ∗ 1 Y (ˆ )Ylm (ˆ) x x l+1 2l + 1 r> lm l. φ) = the potential (3) becomes q Φ= 4π q = 2π ∞ 2l + 1 4π [1 − (−)l ] 0 l=0 l r< l+1 r> 4π Yl0 (θ. This is also obvious from symmetry. a) Find the electrostatic potential as an expansion in spherical harmonics and powers of r for both r > a and r < a. φ ) − Ylm (π. This yields an expansion of (4) ∞ a 2k qa P2k+1 (cos θ) Φ= 2π 0 r2 r k=0 . φ) ∼ Plm (1) and that Plm (1) = δm. The potential is clearly Φ= q 4π 1 1 − x − a  x + a  0 where a = aˆ points from the origin to the positive charge. Using the spherical z harmonic expansion 1 = 4π x − x  as well as a = aˆ. take the limit of a → 0 and ﬁnd the potential for r = 0.m (3) Noting that Ylm (0.m l r< 1 ∗ [Y ∗ (0. Since Yl0 (0.3.6 Two point charges q and −q are located on the z axis at z = +a and z = −a. we obtain z Φ= q 0 l. we have r< = a and r> = r.
To account for the spherical shell. φ. By linear superposition ﬁnd the potential everywhere inside the shell. z).Setting qa = p/2 and taking a → 0. z) = Jm (kρ) or Nm (kρ) e±imφ e±kz However. The general solution obtained by separation of variables has the form Φ(ρ. φ. while the potential on the cylindrical surface is given as V (φ. The potential on the end faces is zero. The resulting series expression for the potential is ∞ ∞ Φ(ρ. c) Suppose now that the dipole of part b) is surrounded by a grounded spherical shell of radius b concentric with the origin. and the resulting solution is Φ= p 4π 0 1 r − 3 2 r b cos θ 3. z) = m=0 n=1 Im nπ ρ L amn sin mφ + bmn cos mφ sin nπ z L (6) . the only term that can show up on the left hand side is the l = 1 term. We then take A1 = −1/bl+2 . The result is to use the modiﬁed Bessel functions Im (kρ) and Km (kρ) instead. For an inside solution. it is natural to take k → ik so that the z function obeying boundary conditions is sin(nπz/L). The result is p cos θ p 1 Φ= P1 (cos θ) = (5) 4π 0 r2 4π 0 r2 which is the potential due to a dipole. we add to (5) a solution to the (homogeneous) Laplace’s equation. for the solution to be regular at ρ = 0 we discard the Kν (kρ) functions. Using the appropriate separation of variables in cylindrical coordinates. only the k = 0 term survives in the sum. we have p Φ= 4π 0 1 P1 (cos θ) + r2 ∞ Al rl Pl (cos θ) l=0 The boundary condition Φ(r = b) = 0 corresponds to having ∞ Al bl+2 Pl (cos θ) = −P1 (cos θ) l=0 Since the Legendre polynomials form an orthonormal set. However.9 A hollow right circular cylinder of radius b has its axis coincident with the z axis and its ends at z = 0 and z = L. ﬁnd a series solution for the potential anywhere inside the cylinder. since the potential vanishes on the endcaps. which blow up at vanishing argument.
z) = V (φ) = ±V is an even function of φ.2l+1 = (−)k 16V π 2 I2k+1 ((2l + 1)πb/L) (2k + 1)(2l + 1) Inserting this into (6) yields 16V Φ= 2 π ∞ ∞ k=0 l=0 I2k+1 ( (2l+1)πρ ) (−)k (2l + 1)πz L cos(2k+1)φ sin (8) (2l+1)πb (2k + 1)(2l + 1) I2k+1 ( L ) L . we need to have V (φ. so that V (φ. As a result. z) = V −V for −π/2 < φ < π/2 for π/2 < φ < 3π/2 a) Find the potential inside the cylinder.n must be divided by two. This can be rewritten as amn bmn = 2 πLIm (nπb/L) 2π L dφ 0 0 dz V (φ. we want to ﬁnd the coeﬃcients amn and bmn of the expansion (6) and (7).9 the cylindrical surface is made of two equal halfcylinders.In order to satisfy the boundary conditions on the cylindrical surface. z) sin mφ cos mφ sin nπ z L (7) (where b0. z) 0 sin mφ cos mφ sin nπ z L with the caveat that b0. z) = m. To obtain the potential. Introducing m = 2k + 1 and n = 2l + 1. we have b2k+1. one at potential V and the other at potential −V . 3.n has to be divided by two). Noting ﬁrst that V (φ. We are left with bmn = = 2V πLIm (nπb/L) 2V π 2 Im (nπb/L) π/2 3π/2 L − −π/2 π/2 dφ cos mφ 0 n dz sin nπz L 4 sin(mπ/2) 1 − (−) m n (m = 0) This is nonvanishing only when both m and n are odd.10 For the cylinder in Problem 3.n Im nπb L amn sin mφ + bmn cos mφ sin nπ z L This is a double Fourier series in φ and z. the Fourier coeﬃcients are amn bmn Im nπb L 1 = π 2π 0 2 dφ L L dz V (φ. we see that all the amn coeﬃcients vanish.
For L b both ρ/L and b/L are much less than one.13 (where V1 = −V2 = V ). consider the potential at z = L/2 as a function of ρ and φ and compare it with twodimensional Problem 2.b) Assuming L b.13. This allows us to use a small argument expansion of the modiﬁed Bessel function Iν (x) ≈ In addition. . for z = L/2 we have sin (2l + 1)πz 1 = sin(l + 2 )π = (−)l L 1 Γ(ν + 1) x 2 ν Hence in this limit (8) becomes Φ= 16V π2 (−)k (−)l 2k + 1 2l + 1 (−)l 2l + 1 ρ b 2k+1 cos(2k + 1)φ (−)k 2k + 1 ρ iφ e b 2k+1 k.l 16V = 2 π l k Noting the Taylor series expansion for arctan tan−1 z = n (−)n 2n+1 z 2n + 1 we arrive at Φ= To calculate ρ iφ 4V 16V tan−1 (1) tan−1 e = 2 π b π tan−1 ρ iφ e b tan−1 z we recall that tan−1 a + tan−1 b = tan−1 a+b 1 − ab 1 2 Hence 1 tan−1 z = 2 (tan−1 z + tan−1 z ∗ ) = tan−1 z + z∗ 1 − z2 For z = (ρ/b)eiφ we ﬁnd Φ= 2V 2(ρ/b) cos φ 2V 2bρ cos φ tan−1 = tan−1 2 2 π 1 − (ρ/b) π b − ρ2 which reproduces the answer to Problem 2.
This action might not be possible to undo. Are you sure you want to continue?