Chapter 3 Potentials (Special Techniques) EM

3.1 Laplace’s Equation: 3.1.1 Introduction Tsun-Hsu Chang

1
Poisson’s equation: V =−
2
 (r)
0
Very often, we are interested in finding the potential in a
region where  = 0.
There may be plenty of charge elsewhere, but we’re
confining our attention to places where there is no charge.

Laplace’s equation:  V =0
2

V V V
2 2 2
In Cartesian coordinates, + + = 0
x 2
y 2
z 2

Introduction to Laplace and Poisson Equations:

[Link] 1
 3.1.2. Laplace’s Equation in 1D
Suppose V depends on only one variable, x.
2
d V
=0  V (x) = mx + b
dx 2

Two features of this solution:

1. Laplace’s equation is a kind of averaging instruction.

1
V (x) = (V (x − a) +V (x + a)) for any a
2
2. Laplace’s equation tolerates no local maxima or minima,
since the second derivative must be zero.

2
 3.1.3. Laplace’s Equation in 2D

Suppose V depends on two variables, x and y.

V
2 V
2 a partial differential equation (PDE);
+ =0 
x 2 y 2
not an ordinary differential equation (ODE).
Harmonic functions in two dimensions have the same
properties that we noted in one dimension:

3
 Features of Harmonic Function in 2D

1. The value of V at a point (x, y) is the average of those

around the point.
1
V (x, y) = 
2 R circle
Vdℓ

2. V has no local maxima or minima. All extrema occur at

the boundaries.

4
 3.1.4. Laplace’s Equation in 3D
V V V
2 2 2
+ + = 0 (partial differential equation (PDE))
x 2
y 2
z 2

In three dimensions we can neither provide you with an

explicit solution nor offer a suggestive physical example to
guide your intuition.
Nevertheless, the same two properties remain true.

1. The value of V at a point r is the average value of V

over a spherical surface of radius R centered at r:
1
V (r) =
4 R 2  Vda
sphere

5
 No Local Maxima or Minima in 3D
2. V has no local maxima or minima; the extreme values
must occur at the boundaries.

Ex. For a single point charge q located outside the sphere

of radius R as shown in the figure, find the potential at the
origin.
1 q 1 q
Sol:V = =
40 r 40 (z 2 + R2 − 2zR cos )1/2
1 q R2 sindd
so Vave (r = 0) =
4 R 2 
4 0 (z 2 + R 2 − 2zR cos )1/2
1 q −d cos
= 
2 40 (z 2 + R2 − 2zR cos )1/2
1 q 
= (z + R − 2zR cos )
2 2 1/2
2zR 40 0
1 q q
= ((z + R) − (z − R)) =
2zR 40 40 z 6
3.1.5. Boundary Conditions and Uniqueness Theorems

Laplace’s equation does not by itself determine V; a

suitable set of boundary conditions must be supplied.

What are appropriate boundary conditions, sufficient to

determine the answer and yet not so strong as to generate
inconsistencies? It is not easy to see.

For a given set of boundary conditions, is V uniquely

determined? Yes, it is. ➔ uniqueness theorem

7
Boundary Conditions and Uniqueness Theorems
First uniqueness theorem: the solution to Laplace’s
equation in some volume is uniquely determined if V is
specified on the boundary surface.
Proof:
Suppose there were two solutions to
Laplace's equation: 2V1 = 0 and 2V2 = 0
Their difference is:V3  V1 −V2.
This obays Laplace's equation,  V
2
3=0
Since V3 is zero on all boundaries and
Laplace's equation suggests that all extrema
occur on the boundary, so V3 = 0  V1 = V2.

8
 Uniqueness Theorems with Charges Inside
 V1
2 =−
 
and  V2 = − . Let V3  V1 −V2  2V3 = 0
2
0 0
Since V3 is zero on all boundaries and Laplace's equation suggests
that all extrema occur on the boundaries, so V3 = 0.  V1 = V2

Corollary: The potential in a volume is uniquely determined

if (a) the charge density throughout the region, and (b) the
value of V on all boundaries, are specified.

The uniqueness theorem frees your imagination. It doesn’t

matter how you come by your solution; if (a) it satisfies
Laplace’s equation and (b) it has the correct value on the
boundaries, then it is right.
9
optional 3.1.6. Conductors and the Second Uniqueness
Theorem
The simplest way to set the boundary conditions for an
electrostatic problem is to specify the value of V on all
surfaces surrounding the region of interest.
However, there are other circumstances in which we don’t
know the potential at the boundaries rather the charges on
various conducting surfaces. Is the electric field still
uniquely determined?
➔ Second uniqueness theorem.

10
optional Second Uniqueness Theorem
In a volume surrounded by conductors and containing a
specified charge density, the electric field is uniquely
determined if the total charge on each conductor is given.
Proof:
Suppose there are two solutions:

 E1 =
 and   E2 =

0 0
Both obey Gauss's law in integral form,
1
E1  da = Qi and
 
1
E2  da = Qi
0 0
ith conducting ith conducting
surface surface
Likewise, for the outer boundary
1
E1  da = Qtot and
 
1
E2  da = Qtot
0 0
outer outer
11
boundary boundary
optional

As before, we examine the difference E3  E1 − E2

which obeys  E3 = 0 in the region between the conductors, and

E3  da = 0 over each boundary surface.

Although we don't know how the charge distributes itself over
the conducting surface, we do know that each conductor is an
equipotential, and hence V3 = 0.
Invoking product rule, we find that
  (V3E3 ) = V3 (  E3 ) + E3 V3 = −(E3 )2

=0
 (  (V3E 3 ))d  = 3 3
(V E )  da =  3
−(E ) 2 d

v S v
E3 = 0 everywhere. Consequently, E1 = E2.
12
 3.2 The Method of Images: EM
Tsun-Hsu Chang
3.2.1 The Infinite Grounded Conducting Plane
Suppose a point charge q is held a distance d above an
infinite grounded conducting plane. What is the potential in
the region above the plane?

The boundary conditions of this case are:

1. V = 0 when z = 0 (since the conducting plane is grounded).
2. V → 0 far from the charge.

13
 The Image Charge
We can easily find a solution which
satisfies the boundary conditions as in the
figure.
The uniqueness theory guarantees that
this case is got to be the right answer.

The potential can then be written down as

1   q q 
V (x, y, z) = −  for z  0
40  x2 + y2 + (z − d )2 x + y + (z + d )
2 2 2 
 

Can we use this potential to find out the electric field,

surface charge distribution, and the force? Yes.

14
 3.2.2 Induced Surface Charge
It is straightforward to compute the surface charge 
induced on the conductor.
(E ⊥ − E⊥ 
b elo w )=
above
0
=0 since z0

V
−V  n̂  − for z  0
n
V V
 = − 0 = − 0
n z z=0

−1 −1  2(z − d )q 2(z + d )q 
=  − 2 3/ 2 
4 2  (x + y + (z − d ) )
2 2 2 3/ 2
(x + y + (z + d ) )  z=0
2 2

−1 −1 − 4qd −1 qd
= =
4 2 (x + y + d )
2 2 2 3/ 2
2 (x + y + d )
2 2 2 3/ 2
15
 Total Induced Charge

The total induced charge is (use the polar coordinate)

−1 qd −1 qd
= =
2 (x + y + d )
2 2 2 3/ 2
2 (r + d )
2 2 3/ 2

 2 −1 qd
Q =  da = 0 0 rdrd
2 (r + d )
2 2 3/ 2


 − qd qd
= dr 2
= = −q
0 2(r + d )
2 2 3/ 2
(r + d )
2 2 1/ 2
0

16
 3.2.3 Force and Energy

The charge q is attracted toward the plane, because of the

negative induced charge.
The force and the energy of this system can be analogous
to the case of two point charges.

q 2 q 2
F=− 1 zˆ ; W =− 1
4 0 4d 2 40 2d

Unlike the two point charges system, there is no field in the

conductor. Handle must be care.

17
 Work and Energy
Consider the work required to bring q in from infinity.
2 2
d d 1q 1 q
W =  Fdz =  dz = −
  4 4z 2
40 4d
0

which is half of that of the two point charge system.

This is because the conducting plane is grounded.

If the plane is not grounded, what would happen?

18
 3.2.4 The Grounded Spherical Conducting Shell

Any stationary charge distribution near a grounded conducting

plane can be treated in the same way, by introducing its mirror
image---method of images.

The image charges have opposite sign; this is what

guarantees that the plane will be at potential zero.

Can this method be applied to a curved surface? Yes.

Here is an example. A point charge is situated in front of a

grounded conducting sphere.

19
Example 3.2 A point charge is situated a distance a from the
center of a grounded conducting sphere of radius R. Find the
potential outside the sphere.

20
Sol: Assume the image charge q is placed at a distance b from
the center of the sphere. The potential is

 r = r 2 + a2 − 2ra cos

V (r) = 1  q q 
 +   for r  R
4 0  r r  
r = r + b − 2rb cos
2 2

It is equipotential on the surface of a grounded sphere.

Using two boundary conditions at P1 and P2 .
21
 1 q q 
At P1: ( + )=0
40 R − b a − R 
 two equations and two unknowns (q  and b)
1 q q
At P2 : ( + ) = 0
40 R + b a + R 
R2 R
b= , q = − q
a a
The force of attraction between charge and the sphere is

F= 1 
qq = −1 2
q Ra
40 (a − b) 2
40 (a − R )
2 2 2

If the sphere is connected to a fixed potential, can this method

still be applied? Yes.

Just imagine another image charge situated at the center of

the sphere, which provides a constant potential at the surface.

22
optional Ex. Two equal conducting spheres with radius R, each
carries a total charge Q1 and Q2 at a distance d from each
other. Find the electric field outside the conducting spheres.
Sol:

Assume the charges are located at the respective centers.

Using the image charge method, calculate the first level
induced charges. Then, calculated the second level induced
charges, and so on. The series should converges rather fast.
23
 EM
3.3 Separation of Variables Tsun-Hsu Chang

We shall attack Laplace’s equation directly, using the

method of separation of variables, which is the physicist’s
favorite tool for solving partial differential equations.
Applicability: The method is applicable in the circumstances
where the potential (V) or the charge density () is
specified on the boundaries of some region, and we are
asked to find the potential in the region where  = 0.

Laplace’s equation:  V =0
2

Basic strategy: Look for solutions that are products of

functions, each of which depends on only one of the
coordinates.
V (x, y, z) = X (x)Y ( y)Z (z)
24
 3.3.1 Cartesian Coordinates

Example 3.3 Two infinite grounded metal plates lie parallel

to the xz plane, one at y = 0, and the other at y = a. The left
end, at x = 0, is closed off with an infinite strip insulated
from the two plates and maintained at a specific potential
V0(y). Find the potential inside this “slot”.

25
 Boundary Conditions

The configuration is independent of z, so Laplace’s equation

reduces to two dimensions.
V V
2 2
+ =0
x 2
y 2

The potential inside is subject to the boundary conditions.

(i) V = 0 when y = 0,
(ii) V = 0 when y = a,
(iii) V = V0 ( y) when x = 0,
(iv) V → 0 as x → .

26
 Separation of Variables

The first step is to look for solutions in the form of products:

V (x, y) = X (x)Y ( y)
Substituting into Laplace’s equation, we obtain
2
d X 2
dY 1 1 2
d X 1 2
d Y
(Y +X = 0)   + =0
dx2 dy2 XY X dx 2 Y dy 2

The first term depends only on x and the second only on y.

The sum of these two functions is zero, which implies these
two functions must both be constant.

1 2
d X 1 2
d Y
= C0 and 2 = −C0
X dx2 Y dy
27
 A Simple Solution
Let C0 equal k2, for reasons that will appear in a moment.
=0 (iv)
1 d2X
=k 2
 X (x) = Ae kx + Be−kx
X dx2
2 =0 (i)
1d Y
= −k 2
 Y ( y) = C sin ky + D cosky
Y dy2
V (x, y) = ( Aekx −kx
+ Be )(C sin ky + D cosky)
The boundary condition (iv) requires that A equal zero,
and condition (i) demands that D equal zero.
Meanwhile (ii) yields sin ka = 0, from which it follows that
n
k= , n = 1, 2, 3 , … Why not n = 0?
a

V (x, y) = çBC e −k n x
sin kn y   Cne −n x a
sin(n y a)
Cn n=1 28
 A Complete Solution in Fourier Series
Now we have an infinite set of solutions.

−n x a
V (x, y) =  Cne sin(n y a)
n=1
Can we use the remaining boundary condition (iii) to
determine the coefficients Cn? Yes.

V (0, y) =  C n sin (ny a) = V0 ( y)
n=1

This is a Fourier sine series. Virtually any function V0(y)---

can be expanded in such a series. 這麼神奇!
We can use the so-called “Fourier’s trick” to find out the
coefficients Cn.

29
 The Fourier Trick

 a a
 Cn 0 sin (n y a) sin (n y a)dy =  V0 ( y) sin (n y a)dy
0
n=1
The integral on the left is
a
0 sin (n y a) sin (n y a)dy
 0, if n  n
1 a
=  (cos((n − n)
 y
) − cos((n + n)
 y 
)dy =  a
2 0 a a , if n = n
2
a
2
Cn  =  V0 ( y) sin (n y a)dy
a 0

30
 A Concrete Example
For a constant potential V0
 0, if n is even
2V0 a 2V0 
Cn = 
a 0
sin (n y a)dy =
n
(1− cosn ) =  4V0
, if n is odd
n



4V0 1 −nx a
So V (x, y) = e sin (ny a)
 n=1,3,5,... n

31
 EM
Completeness and Orthogonality Tsun-Hsu Chang

The success of this method hinges on two extraordinary

properties, i.e., completeness and orthogonality.

Completeness: If any other function f(y) can be expressed

as a linear combination of a complete function set fn(y):

f ( y) =  C n fn ( y)
n=1

Orthogonality: If the integral of the product of any two

different members of the set is zero:
a
0 f n ( y) f n ( y)dy = 0 for n  n

This allows us to kill off all terms but one ( n = n ) in the

infinite series and thereby solve for the coefficient Cn.
32
 Rectangular Metal Pipe

Example 3.5 An infinitely long rectangular metal pipe (sides

a and b) is grounded, but one end, at x = 0, is maintained at
a specified potential V0(y, z), as shown in the figure. Find the
potential inside the pipe.

33
 Boundary Condition
This is a genuinely three-dimensional problem,

V V V
2 2 2
+ + =0
x 2
y 2
z 2

The potential inside is subject to the boundary conditions.

(i) V = 0 when y = 0,
(ii) V = 0 when y = a,
(iii) V = 0 when z = 0,
(iv) V = 0 when z = b,
(v) V = V0 (y, z) when x = 0,
(vi) V → 0 as x → .

34
 Separation of Variables

The first step is to look for solutions in the form of products:

V (x, y, z) = X (x)Y ( y)Z (z)
Substituting into Laplace’s equation, we obtain
1 2
d X
1 1 2
d Y 2
d Z
+ + =0
X dx2 Y dy2 Z dz2

It follows that
1 2
d X 1 2
d Y 1 2
d Z
= (k + l ),
2 2
= −k 2
, = −l 2
X 2 Y 2 Z 2
dx dy dz
How do we know? Any other possibility?

35
 A Simple Solution
=0 (vi)
1  X
2 2 +l 2 x − k 2 +l 2 x
= (k + l )  X (x) = Ae
2 2 k + Be
X x2
=0 (i)
1 2Y
= −k 2  Y ( y) = C sin ky + D cos ky
Y y2
=0 (iii)
1  Z
2
= −l2  Z (z) = E sin lz + F cos lz
Z z2
Meanwhile (ii) and (iv) yields sin ka = 0 and sin ℓb = 0,
from which it follows that
n m
k= , n = 1,2,3,… ℓ= , m = 1,2,3,…
a b

36
 A Complete Solution in Fourier Series
The solution is
− ( an ) 2 +( mb ) 2 x
V (x, y, z) = BCEe sin (ny a) sin (mz b),
where n and m are unspecified integers.
Completeness: The solution can be written as
 

V (x, y, z) =  C n,m e
− ( an ) 2 +( mb ) 2 x
sin (ny a) sin (mz b)
n=1 m=1

Use the boundary condition (v) and the orthogonality to find

out the coefficients Cn,m.
 
V (0, y, z) =   Cn,m sin(n y a)sin(m z b) = V0 ( y, z)
n=1m=1
37
 The Fourier Trick & Constant Voltage Solution
  a b
  Cn,m 0 sin (n y a)sin (n y a)dy  sin (m z b)sin (m z b)dz
0
n=1m=1
a b
=  V0 ( y, z)sin (n y a)sin (m z b)dydz
0 0
4 a b
Cn,m =
ab 0 0
V0 ( y, z)sin(n y a)sin(m z b)dydz

If the end of the tube is a conductor at constant potentialV0

4V0 2a 2b 16V0
Cn,m = = if n and m are odd.
ab n m nm 2

=0 if n or m are even.

1 −

16V0 ( an ) 2 +( mb ) 2 x
V (x, y, z) = e sin (n y a) sin (m z b)
 2
n,m=1,3,5,… nm
38
 Homework of Chap. 3 (part I)
Problem 3.11 Two semi-infinite grounded conducting planes meet at right angles.
In the region between them, there is a point charge q, situated as shown in Fig. 3.15.
Set up the image configuration, and calculate the potential in this region. What
charges do you need, and where should they be located? What is the force on q?
How much work did it take to bring q in from infinity? Suppose the planes met
at some angle other than 90; would you still be able to solve the problem by the
method of images? If not, for what particular angles does the method work?

Problem 3.13 Find the potential in the infinite slot of Ex. 3.3 if the boundary at
x = 0 consists of two metal strips: one, from y = 0 to y = a/2, is held at a constant
potential V0 , and the other, from y = a/2 to y = a, is at potential −V0.

Problem 3.16 A cubical box (sides of length a) consists of five metal plates, which
are welded together and grounded (Fig. 3.23). The top is made of a separate sheet
of metal, insulated from the others, and held at a constant potential V0. Find the
potential inside the box. [What should the potential at the center (a/2, a/2, a/2)
be? Check numerically that your formula is consistent with this value.]11

39
 Homework of Chap. 3 (part I)
Problem 3.54 For the infinite rectangular pipe in Ex. 3.4, suppose the potential on
the bottom (y = 0) and the two sides (x = b) is zero, but the potential on the top
(y = a) is a nonzero constant V0. Find the potential inside the pipe. [Note: This is a
rotated version of Prob. 3.15(b), but set it up as in Ex. 3.4, using sinusoidal functions
in y and hyperbolics in x. It is an unusual case in which k = 0 must be included.
Begin by finding the general solution to Eq. 3.26 when k = 0.]26
  y 2  (−1)n cosh(n x / a) 
 Answer: V0 +  n=1 sin(n y / a)  . Alternatively, using sinusoidal
 a  n cosh(n b / a)
 
2V0  (−1)n sinh(n y)
functions of x and hyperbolics in y, −
b
 n=1 n sinh(na)
cos(n x), where

n  (2n −1) / 2b

40
 EM
3.3.2 Spherical Coordinates Tsun-Hsu Chang

For round objects spherical coordinates are more suitable.

In the spherical system, Laplace’s equation reads
1  2 V 1  V 1 V 2
(r )+ 2 (sin )+ 2 2 = 0
r r
2
r r sin   r sin   2

We will first treat the problem with azimuthal symmetry,

so that the potential is independent of .

 2 V 1  V
(r )+ (sin )=0
r r sin  

41
 Separation of Variables
The first step is to look for solutions in the form of products:

V (r, ) = R(r)( )
Substituting into spherical Laplace’s equation, we obtain
1 d 2 dR 1 d d
(r )+ (sin )=0
R dr dr sin d d
The first term depends only on r and the second only on .
The sum of these two functions is zero, which implies these
two functions must both be constant.
1 d dR 1
(r 2 ) = l ( l +1), d
(sin
d
) = −l(l +1)
R dr dr sin d d
Again, how do we know? Any other possibility?
42
 Simplest Case: A Metal Sphere
Example: A metal sphere of radius R, maintains a specified
potential V0. Find the potential outside the sphere.
Sol: The potential is independent of  and .
1 d dR
The Laplace’s equation is: (r 2
)=0
R dr dr
d 2 dR 2 dR
(r )=0  r = −A
dr dr dr
dR A A
=− 2  R= +B
dr r r
 R(r = R ) = A + B = V
 R0

0
R0
0
 R(r) = V0
R(r = ) = B = 0 r

43
 A Simple Solution & Legendre Polynomials
The general solutions for R and  are
d 2 dR l 1
(r ) = l ( l +1)R  R = Ar + B l+1
dr dr r
1 d d
(sin ) = −l(l +1) The solutions are not simple.
sin d d
The solutions are Legendre polynomials in the variable cos.
( ) = Pl (cos )
See lecture notes
of Jackson Chap.3

The polynomial is most conveniently defined by the

Rodrigues formula (generating function):
1 d ℓ 2
Pℓ (x) = ℓ ( ) (x −1) ℓ
2 ℓ! dx
44
 Rodrigues Formula
1 d ℓ 2
Prove: Pℓ (x) = ℓ ( ) (x −1) , x = cos
ℓ
2 ℓ! dx
1  Pℓ (cos )
where (sin ) = −ℓ(ℓ +1)Pℓ (cos )
sin  
Sol: [Link]
Let v = (x −1)
2 ℓ

v = 2ℓx(x −1)
2 ℓ−1
(x −1)
2

 (1− x )v + 2ℓxv = 0

2

(1− x )v − 2xv + 2ℓxv + 2ℓv = 0

2

(1− x )v + 2(ℓ −1)xv +1(2ℓ − 0)v = 0

2

(1− x )v + 2(ℓ − 2)xv  + 2(2ℓ −1)v = 0

2

2
(1− x )v (k +2)
+ 2(ℓ − k −1)xv (k +1)
+ (k +1)(2ℓ − k)v = 0
(k )

45
 d (x −1)
ℓ 2 ℓ
Let k = ℓ and u = v = (ℓ)
ℓ
= Pℓ (cos )(2 ℓ
ℓ!)
dx
(1− x )u − 2xu + ℓ(ℓ +1)u = 0
2

d Pℓ (cos )
2
dPℓ (cos )
 (1− cos  )
2
− 2x + ℓ(ℓ +1)P ℓ (cos ) = 0
dx2 dx

dPℓ (cos ) dPℓ (cos ) d 1 dPℓ (cos )

= =−
dx d dx sin d
d Pℓ (cos ) d
2
1 dPℓ (cos ) 1
= ( − )( − )
dx 2
d sin d sin
1 d Pℓ (cos ) cos dPℓ (cos )
2
= −
sin 
2
d 2
sin 
3
d

46
 1 d 2
Pℓ (cos ) cos dPℓ (cos )
(1− cos  )[ 2
2
− ]
sin  d 2 sin3  d
1 dPℓ (cos )
− 2 cos[− ] + ℓ(ℓ +1)Pℓ (cos )
sin d
d Pℓ (cos ) cos dPℓ (cos )
2
= + + ℓ(ℓ +1)P ℓ (cos )
d 2
sin d
1 d dPℓ (cos )
= (sin ) + ℓ(ℓ +1)Pℓ (cos ) = 0
sin d d

1 d (cos  −1)
ℓ 2 ℓ
 Pℓ (cos ) = ℓ
2 ℓ! d (cos ) ℓ

47
 Properties of Legendre Polynomials EM
The first few Legendre polynomials are listed Tsun-Hsu Chang
P0 (x) = 1
P1 (x) = x
P2 (x) = (3x2 −1) / 2
P3 (x) = (5x3 − 3x) / 2
P4 (x) = (35x4 − 30x2 + 3) / 8
P5 (x) = (63x 5 − 70x 3 +15x) / 8
Pℓ (x): an ℓth-order polynomial in x
Completeness: The Legendre polynomials constitute a
complete set of functions, on the interval –1 x 1.
Orthogonality: The polynomials are orthogonal functions:
1 
 Pl (x)P
−1 l  (x)dx =  Pl (cos 
0
)Pl  (cos  )sin  d 
 0 if l   l

= 2
2l +1 , if l  = l 48
 A Complete Solution in Legendre Polynomials
The Rodrigues formula generates only one solution. What
and where are other solutions?
These ”other solutions” blow up at  = 0 and/or  = , are
therefore unacceptable on physical grounds.

1
V (r, ) = (Ar + B
ℓ
ℓ+1
)Pℓ (cos ) See lecture notes
of Jackson Chap.3
r
The general solution is the linear combination of separable
solutions.

1
V (r, ) =  ( Aℓ r + Bℓ
ℓ
ℓ+1
)Pℓ (cos )
ℓ=0 r

49
Example 3.6 The potential V(R,) = V0sin2(/2) is specified on
the surface of a hollow sphere, of radius R. Find the potential
inside the sphere V(r,).

Sol: V (r, ) =  (Aℓ r + Bℓ
ℓ 1
ℓ+1
)Pℓ (cos )
ℓ=0 r
In this case Bℓ = 0 for all ℓ --- otherwise the potential would

blow up at the origin. Thus, V (r, ) =  ℓ
Aℓ r Pℓ (cos )
 ℓ=0
 Aℓ R ℓ
Pℓ (cos ) = V (R, )
ℓ=0 Ref. p.48
2ℓ +1 1 
Aℓ =
2 R ℓ 0
V (R, )Pℓ (cos ) sind
2ℓ +1 1  
= 0   
2
ℓ
V0 sin ( )Pℓ (cos )sin d
2 R 2
2ℓ +1 1 V
=
2 R ℓ 0 2 0
(1− cos  )P ℓ (cos  )sin  d 
2ℓ +1 1 V
=
2 R ℓ 0 2 0
(P 0 (cos  ) − P1 (cos ))P ℓ (cos ) sin  d 
50
 2ℓ +1 1  V0
Aℓ =
2 R ℓ 
0 2
(P0 (cos ) − P1 (cos ))Pℓ (cos ) sind

1  0 if ℓ   ℓ
−1 ℓ ℓ
P (x)P  (x)dx =  2
2ℓ+1
, if ℓ  = ℓ

V0
A0 = V0  r 
2  V (r, ) = − cos
1 
V0 2  R 
A1 = −
2R

51
Example 3.8 An uncharged metal
sphere of radius R is placed in an
otherwise uniform electric field E = E0 ẑ
Find the potential in the region outside
the sphere.

Sol: The sphere is an equipotential---we may as well set it

to zero.
The potential is azimuthally symmetric and by symmetry the
entire xy plane is at potential zero.
In addition, the potential is not zero at large z.
Boundary conditions are:
(i) V = 0 when r = R,
(ii) V → −E0r cos for r R.
52
 
V (r, ) =  (Aℓ r + Bℓ r
ℓ −(ℓ+1)
)Pℓ (cos )
ℓ=0

B.C. (i): V (R, ) =  ( Aℓ R + Bℓ R ℓ −(ℓ+1)
)Pℓ (cos ) = 0
ℓ=0

 Bℓ = − Aℓ R 2ℓ+1


B.C. (ii): V (r, ) =  ( Aℓ r )Pℓ (cos ) = −E 0r cos
ℓ

ℓ=0
 A1 = −E0 , all other Aℓ are zero.
3
R
V (r, ) = −E0 (r − 2 ) cos Why the electric field
r is enhanced?
3
R
E r= R = −V = E0 (1+ 2 3 ) cos rˆ = 3E0 cos  rˆ
R
 ( ) = 0 (3E0 cos r̂)  r̂ = 30 E0 cos

(E ⊥ − E⊥ ) =

b elo w
a bo ve
0
=−V nˆ =0 since z0 53
 EM
3.4 Multipole Expansion Tsun-Hsu Chang
3.4.1 Approximate Potential at Large Distance
If you are very far from a localized charge distribution, it
“looks” like a point charge, and the potential is---to good
approximation—(1/40)Q/r, where Q is the total charge.
But what if Q is zero?
Develop a systematic expansion for the potential of an
arbitrary localized charge distribution, in powers of 1/r.
1 1
V (r) = 
40 r − r
(r  )d  

Using the law of cosines,

1 1 Note, for simplicity,
=
r − r (r + (r) − 2rr cos )
2 2 r = rẑ

54
 Large Distance Approximation

1 1 1
= = (1+  ) −1/2
r − r (r + r − 2rrcos ) r
2
2

r r
where  = ( − 2 cos ) Taylor’s expansion
r r
1 1 1 3 2 5 3
(1+  ) = (1−  +  −  +… ) , if   1
−1/2
r r 2 8 16
1 1 1 r  r 3 r r 
So = (1− ( − 2 cos ) + ( ( − 2 cos )) 2

r − r r 2r r 8 r r
5 r r 
− ( ( − 2 cos )) +… ) 3

16 r r
1 r  r 2 (3cos   −1)
2
= (1+ ( ) cos  + ( ) ( ) +… )
r r r 2

55
 Legendre Polynomials & Multipole Expansion
1 1 (1+ ( r ) cos + ( r )2 ( (3cos   −1)
  2
=  ) +… )
r − r r r r 2

1 r ℓ
=  ( ) Pℓ (cos )
r ℓ=0 r

r ℓ
V (r) = 
1
 (
40 r ℓ=0 r
) P ℓ (cos  ) (r )d   This is the desired result.


=
1

40 r ℓ=0 r
1
ℓ 
(r ) ℓ
Pℓ (cos  ) (r )d  

1 1 
 2 
  (r  )d   + rcos   (r)d  
or more explicitly, V (r) = 1 r r

4 0 + 13  (r  ) ( 
2 3 1
cos   − ) (r)d  + … 
2

 r 2 2 

The multipole expansion of V in powers of 1/r.

56
 Legendre Polynomials & Multipole Expansion

V (r)
1 1 1 1 2 3 1
= (
4 0 r  (r)d  + 2  rcos (r)d  + 3  (r ) ( cos   − )(r)d  + … )
r r

2
2
2

57
 Dipoles
What is dipole? The arrangement of a pair
of equal and opposite charges separated by
some distance is called an electric dipole.

Permanent dipole: such as molecules of HCl, CO, and H2O.

Induced dipole: An electric field may also induce a charge

separation in an atom or a nonpolar molecule.

58
Example 3.10 An electric dipole consists of
two equal and opposite charges separated
by a distance d. Find the approximate
potential V at points far from the dipole.

Sol:
q 1 1 q −1/2 −1/2
V (r) = ( − )= ((1+  ) − (1−  ) )
40 r − d ẑ r + d2 ẑ 40r
2
r r r d
d
where  = ( − 2 cos )  − cos (since 1 and r = )
r r r r 2
q −1/2 −1/2
V (r) = ((1+  ) − (1−  ) )
40r
q d 1 qd cos
= ( cos ) =
40 r r 40 r2

59
 The Electric Field of a Dipole
1 qd cos 1 r p
ˆ p cos Why?
V (r) = = =
4 0 r 2
40 r 2
4 0 r 2
Just a convention.
where p = qd  pointing from the negative charge to the positive charge.
p 2 cos sin ˆ
E = −V (r) = ( r̂ + θ̂ + 0 )
4 0 r 3
r 3

p
 +  ˆ T 1 T 1 T ˆ
= (2 cos rˆ sin θ ) T = r̂ + ˆ + 
40 r 3
r r  r sin 

60
 Some Important Properties of Dipole
Potential and field due to a dipole:
p = Qd (− → + )
1 1
2 
V (r) = rcos (r)d 
4 0 r
1 1 1 1
= r̂   r (r)d  = r̂ p
4 0 r 2
– – 40 r 2

p: dipole moment
E = −V (r)

Torque in a uniform field:

τ = pE
Potential energy:

U = −p  E
61
 Homework of Chap. 3 (part II)
Problem 3.20 Suppose the potential V0 ( ) at the surface of a sphere is specified,
and there is no charge inside or outside the sphere. Show that the charge density on
the sphere is given by

0
 ( ) = 
2R l=0
(2l +1) 2
Cl Pl (cos ), (3.88)

where

0
Cl = V0 ( )Pl (cos ) sind . (3.89)

Problem 3.27 A sphere of radius R, centered at the origin, carries charge density
R
 (r, ) = k 2 (R − 2r) sin ,
r
where k is a constant, and r, are the usual spherical coordinates. Find the approximate
potential for points on the z axis, far from the sphere.

62
 Homework of Chap. 3 (part II)
Problem 3.43 A conducting sphere of radius a, at potential V0 , is surrounded by a
thin concentric spherical shell of radius b, over which someone has glued a surface
charge
 ( ) = k cos  ,
where k is a constant and  is the usual spherical coordinate.
(a) Find the potential in each region: (i) r > b, and (ii) a < r < b.
(b) Find the induced surface charge i ( ) on the conductor.
(c) What is the total charge of this system? Check that your answer is consistent
with the behavior of V at large r.
 aV0 / r + (b3 − a3 )k cos / 3r 20 , r  b
 Answer: V (r,  ) =  
  0
aV / r + (r 3
− a 3
)k cos  / 3r 2
0 , r  b

Problem 3.56 An ideal electric dipole is situated at the origin,

and points in the z direction, as in Fig. 3.36. An electric charge
is released from rest at a point in the xy plane. Show that it swings
back and forth in a semi-circular arc, as though it were a pendulum
supported at the origin.28

63