4-EM2011-Image_method_in_vacuum 3/21/2011 12:52:00 PM - 1 -

1
Electromagnetics
2011 #4 (3/21)

Solution of Electrostatic Problems; II
Electrostatic Image Charges (conductor cases)

The solution of a Laplaces equation is unique.
Therefore, if one obtains a solution of the Laplaces equation, by any means whatever,
and the solution satisfies all the boundary conditions of the system,
then, it is the complete and only solution of the problem.

The effect of the surface charges, induced by a known or given charge distribution, can
be replaced by one or group of image charges in some particular cases. This is called the
image charge method. It holds for some (not all) highly symmetric cases.


Example 1; A point charge near a conducting plane

A point charge located near a conducting plane induces some surface charges on the
plane itself. The effect of the induced surface charges can be replaced with a single
image charge q' .

When the charge q is located at x d = and the conducting plane is at 0 x = . The
induces image charge q' is assumed located at x d' = .

Then, out of the conductor, the electric potential is simply given by the sum of the
potentials produced from the original charge q and the image charge q' as

0
1
4
out
q q
r r
m
tc
'
| |
= +
|
'
\ .


The system has only one boundary; the conducting plane located at 0 x = .
The potential on the conducting plane is zero (grounded case) or constant (floating case).


For the grounded case, the boundary condition is
4-EM2011-Image_method_in_vacuum 3/21/2011 12:52:00 PM - 2 -
2
at 0 x = , 0
s
m =

It means
0
q q
r r
'
+ =
'


This boundary condition is simply satisfied by setting, in the diagram
q q ' = and
d d ' = .




It gives

0
1 1
4
out
q
r r
m
tc
| |
=
|
'
\ .


Where r is the length of the vector heading to the point P( , , ) x y z and originated from
the point ( ,0,0) d , and r' is the one from the point ( ,0,0) d .

The induced charge q' is the so-called image charge.

The real charge is the surface charges induced by the original charge q.

The surface charge density is obtained by taking the normal E field and using the
equation of
' r
O
q'
( ) q''
d'
d
r
q
P
x
4-EM2011-Image_method_in_vacuum 3/21/2011 12:52:00 PM - 3 -
3

0
n
E
o
c
=

The total induced charge is the same as q (for the infinite plane boundary case). It is
confirmed by taking the Gauss integration on the entire half space including the surface
of the conductor. (At a large r, the E field intensity decreases faster than 1/r. Thus the
surface integration becomes vanished. The image charge and the real original charge
form a kind of an electric dipole.)

The Coulomb force exerted on the original charge q by the surface charge on the
conductor plane is
q m' = V F

Where m' is the electric potential that is resulted from the image charge(s) only.

Since the image charge is the result of the induced surface charges, at the other side of
the conductor, the image charge is located at the mirror image position. In this problem,
the mirror-imaged image charge is located at the same position of the original charge.
Therefore, the net charge vanishes, which explains why there is no E-field inside of the
conductor.

What will happen when the conducting plane is not grounded but charged with a
constant potential? What is the effect of the grounding?


Example 2; Point charge near a conducting sphere

A point charge q is located in the vicinity of a conducting sphere.

When the sphere was initially neutral and not grounded, by using the image charge
method, we can get the location(s) and the magnitude(s) of the image charge(s).

When the sphere becomes grounded, the location(s) and the magnitude(s) of the image
charge(s) are not altered, but a new image charge(s) can be generated or cancelled out!
Therefore, we need to discuss the role of the ground carefully.

4-EM2011-Image_method_in_vacuum 3/21/2011 12:52:00 PM - 4 -
4


At first, lets think of a grounded sphere having a radius a. At the boundary, or at the
sphere surface, the electric potential vanishes (It is the definition of ground!);
( ) , , 0 r a m u m = =

When there are no external charges out of the sphere, the sphere would not have any
surface charge also.

When a point charge q is placed out of the sphere, at a distance d from the center of the
sphere, the sphere becomes to have induced surface charges due to the E field generated
by the point charge q.

The E field generated by the surface charges might be expressed by a set of discrete
point charges, so-called the image charges, located inside the sphere. At least, we can
try to assume the image charges and check the boundary conditions of the system. When
the image charges happen to satisfy all the boundary conditions of the system, we can
get the complete solution with the help of the image charge(s).

Lets assume an image charge located at a distance b from the center of the sphere but
still inside the sphere having a radius a.

b a d < <

2
R
r
a
O
q q'
q''
b
u
P
d
2
r
1
r
4-EM2011-Image_method_in_vacuum 3/21/2011 12:52:00 PM - 5 -
5
Due to the azimuthal symmetry, the image charge should be located on the line
connecting the point charge q and the center of the sphere O (the origin of the
coordinate);
( )
0 1 2
1
, ,
4
q q
r
r r
m u m
tc
'
= +
`
)


With the distances from the sources (real and image charges) to the measuring point P;

2 2 2
1
2 cos r r d rd u = +

2 2 2
2
2 cos r r b rb u = +

Each distance is given as a function of the polar angle u .

Note)
( ) ( )
2 2 2 2 2
1
2 2 cos r r d r d rd u = = + = + r d r d r d

Boundary condition;
At r a = ,
( )
2 2 2 2
2 1
0 1 2 1 2
1
, , 0 0 0
4
q q q q
a r q r q
r r r r
m u m
tc
' '
' = = + + = =
`
)


Substituting the distances into this gives

( ) ( )
2 2 2 2 2 2
2 cos 2 cos 0 q a b ab q a d ad u u ' + + =

The boundary condition should hold for any angle u . Thus, we have two equations;

One is from the constant terms

( ) ( )
2 2
2 2 2 2 2 2 2 2
2 2
0
a b
q a b q a d q q
a d
+
' ' + + = =
+


And the other equation is from the u dependent terms,
4-EM2011-Image_method_in_vacuum 3/21/2011 12:52:00 PM - 6 -
6

2 2 2 2
2 2 0
b
q ab q ad q q
d
' ' = =

There are two unknowns, q' and b, and we have two equations. Thus, we can solve them
at least in principle.

By equating them, we have the location of the image charge as
b d = and
2
a
b
d
=

However, the solution b d = is out of the sphere, so needs to be discarded. Thus, we
have the location of the image charge as
2
a
b
d
=

From this the amount of the image charge is obtained as

a
q q
d
' =

It fulfills the boundary condition; potential vanishes on the surface of the grounded
conducting sphere. Therefore, it is the solution of the problem.

The image charge q' is an induced charge. It comes from the ground to make the
potential vanished at the conducting sphere.

When the sphere is not grounded, there is no path of supplying the charges of the
image charge q' . Thus, we need one more image charge that is

located at the center of the sphere, thus, gives uniform potential
and has the equal amount of charge as q' , q q '' ' = .

In this case, the sphere has a nonzero potential!!

0 0 0
1 1 1
4 4 4
s
q q q a
R R d R
m
tc tc tc
'' '
= = =

The induced second image charge q'' generates a uniform potential on the surface of
4-EM2011-Image_method_in_vacuum 3/21/2011 12:52:00 PM - 7 -
7
the conducting sphere.

Example 3; A point charge near a conducting corner

We can use as many as image charges, provided they can satisfy the given boundary
conditions.

Example 4; A point charge near a conducting wedge

Even with infinitely many image charges, there is a situation where we can not satisfy
the given boundary conditions.

Example 5; Line charges near a conducting plane

The image charge can be in a form of a line charge, not in a localized point charge!


When the conducting boundary is replaced with a dielectric one, such as a dielectric
plane, is it also possible to use the image charge method?

If yes, is it possible to get the E field within the dielectric medium?