Electrodynamics I Midterm Exam - Part A - Closed Book KSU 2015/09/25

Name Electrodynamic Score = 120 / 120 points

Instructions: Use SI units. Where appropriate, define all variables or symbols you use, in words. Try
to tell about the physics involved, more than the mathematics, if possible.

1. (10) Write Gauss Law in integral form. State its physical significance.

I Z
1 1
da E n = d3 x (x) or E = Qenc .
S 0 V 0
E=electric field, n=outward normal unit vector, =volume charge density, 0 =permittivity of vac-
uum, E =electric flux out of the volume, Qenc =net charge enclosed in volume. It exemplifies how
electric flux (LHS integral) is produced by electric charge (RHS integral) within a volume.

2. (10) A fixed electric charge density (x) exists in space with no boundaries. Write an
expression for the electric field that it produces at a field point x.

(x x0 )
Z
1
E(x) = d3 x0 (x0 ) ,
40 |x x0 |3
where x0 is a source point and x is the field point where E is being calculated. The integral is over
all of space.

3. (10) A fixed electric charge density (x) exists in space with no boundaries. Write an
expression for the electrostatic potential that it produces at a field point x.

(x0 )
Z
1
(x) = d3 x0 ,
40 |x x0 |
where x0 is a source point and x is the field point where the potential is being calculated. The
integral is over all of space.

4. (10) Write an expression that gives the electrostatic field energy density in vacuum.

1
0 |E|2 ,
uE =
2
where uE is the energy density in joules/meter3 . This should be integrated over space to get the
total electrostatic energy in the field.

5. (10) From a solution for electric potential (x), how can you obtain the surface charge
density on a conductor boundary? Let n be the unit vector pointing perpendicularly
outward from the conductor.

You can use Gausss Law in integral form on small pillbox at the surface, and get
1
En A = A
0
where En = E n is the component of E outward from the conducting surface. This can be solved
and also related to the electric potential,


= 0 En = 0 .
n S

The derivative is along the axis outward from the surface, at the surface itself.

1
6. (10) From the previous two answers, how do you write the force per unit area caused by the
electric field on a conductor surface? Express the result as a vector.

The electrostatic energy in a small volume of thickness x and area A just outside a conducting
surface can be used to get the force:
UE
UE = uE V = uE Ax = F = = uE A.
x
1 2
Then the magnitude of force per unit area is the same as F/A = uE = 2 0 |E| . The direction is
parallel to q E, and from Question 5 with = 0 En we can write
1
F/A = E.
2
Curiously this is always outward from the surface, regardless of the sign of .

7. (10) Write Poissons equation for electrostatics. Explain what it is good for.

2 (x) = (x)/0 .
which is derived from the differential form of Gauss Law and the electrostatic assumption that
E(x) = .~ For a given volume charge density (x), the solution will give the electrostatic
potential, which then ultimately describes the electric field distribution.

8. (10) Write a differential equation that a Green function G(x, x0 ) for Poissons equation must
satisfy, for Dirichlet boundary conditions. Include a statement of the boundary conditions.

2 G(x, x0 ) = 4(x x0 ),
where x0 is the source point, x is the observation point or field point where a response (the potential
field) is measured. The boundary condition is that the Greens function vanishes on the boundary:

G(x, x0 ) = 0 x S, x0 S.

9. (10) A problem has Dirichlet boundary conditions. How do you write the general solution
to the Poisson equation for electrostatic potential (x) using a Greens function?

Z Z
1 1 G
(x) = d3 x0 G(x, x0 )(x0 ) da0 (x0 ),
40 V 4 S n0
where the second integral is over the surface S bounding the system, and n0 is a coordinate pointing
out of the system, at the boundary.

Compare, for the next problem with Neumann boundary conditions, where the potential solution is
Z Z
1 3 0 0 0 1
(x) = d x G(x, x )(x ) + da0 G(x, x0 ) 0 + hiS .
40 V 4 S n

The last term there is the average of the potential on the bounding surface, which is sometimes an
irrelevant constant.

2
10. (10) In an electrostatics problem with Neumann boundary conditions, what is the simplest
allowable boundary condition on the the Greens function G(x, x0 )? Hint: The result must
be consistent with the differential equation that G satisfies.

The Greens function solves 2 G(x, x0 ) = 4(x x0 ), and the simplest boundary condition is
G
n0 = A = constant. The constant cannot be zero, due to the divergence theorem applied to the
differential equation:
Z Z I
3 2 0
d x G(x, x ) = 4 3 0
d x (x x ) = ~ n da = 4.
G
V V S
Then the average outward gradient of the Greens function must be
  I
G 1 ~ n da = 4 = A = 4 ,
= G
n S S S S
where S is the total surface area of the system boundary.

11. Use delta-functions to express the charge density (x) for the following charge distributions,
in the indicated coordinate systems:
a) (10) A point charge q on the z-axis at z = c. Use cylindrical coordinates (, , z).

The charge density must have (z c) to get the correct location on the z-axis. Then it will
also need a normalization constant A, and some dependence on radial coordinate . Try
Z
e (x) = Af ()(z c) and require d3 x e (x) = q.

Use volume element d3 x = d d dz, then

Z Z Z 2 Z Z
3
d x e (x) = A f () d d dz (z c) = 2A f () d = q.
0 0 0

The only way for this to work is if A = q/(2) and f () = ()/. So we need
q
e (x) = ()(z c).
2
It has correct dimensions, charge length3 .

b) (10) A charge Q distributed uniformly over an infinitesimally thin circular disk of radius
a centered on the z-axis and lying in the plane z = 0. Use spherical coordinates (r, , ).

The charge is spread out over an area a2 , giving a uniform surface charge density = Q/(a2 ).
We need a delta at = /2 to place the charge in the xy-plane, a constant, and some function
f (r) to get the uniform surface charge density. A step function keeps the charges at r < a. Try
 
(x) = Af (r) H(a r).
2
For r < a, find the amount of charge in a circular ring of width dr, which should be
dq = da = 2r dr H(a r)
also Z Z 2
dq = r2 dr sin d d (x) = 2A r2 f (r)H(a r)
0 0
These require:
Q 1
A== and f (r) = .
a2 r
Then the final charge density is
Q  
(x) = H(a r).
a2 r 2
It is seen to have the correct dimensions.

3
Electrodynamics I Midterm Exam 1 - Part B - Open Book KSU 2015/09/25
Name Electrodynamic Score = 130 / 130 points
Instructions: Use SI units. Please show the details of your derivations here. Explain your reasoning for
full credit. Open-book only, no notes.

1. A point charge q is located a distance a above an infinite plane conductor held at zero
potential. Use the method of images to find

a) (20) The surface charge density on the plane, as a function of a radial coordinate .
With the charge q at x0 = (0, 0, a), by symmetry an image q 0 = q at x00 = (0, 0, a) will cause
the total potential to be zero on the entire xy-plane at z = 0. As a result, the total potential
at some field point x = (x, y, z) is
 
1 q q
(x) = .
40 |x x0 | |x x00 |

With x = (, , z) in cylindrical coordinates, where 2 = x2 + y 2 , this is

" #
q 1 1
(, , z) = p p .
40 2 + (z a)2 2 + (z + a)2

From Gauss Law one knows the relation, = 0 En = 0 Ez , which gives the surface charge
density in terms of the component of electric field at the conducting surface, z = 0. This is
  
 q za z+a
= 0 Ez = 0 = 2
z z=0 4 [2 + (z a)2 ]3/2 [ + (z + a)2 ]3/2 z=0
 
q 2a qa
= 2 2 3/2
= .
4 [ + a ] 2[ + a2 ]3/2
2

This diminishes away from the charge, also gets smaller if the charge is farther from the con-
ductor, and it has the opposite sign as q.

b) (20) The force per unit area on the plane, F/A, as a function of .
The forces acts on the surface charge density. In a previous question, the force per area has
been found to be given from F/A = 12 E, using the value of the electric field at the conducting
surface, which points in the z direction. Because is also negative, the force points in the +z
direction. The magnitude of force per area is also the same as the energy density, measured at
the surface,
2  2
(qa)2
 
Fz 1 1 2 1  1 q 2a
= Ez = 0 Ez = 0 = 0 = .
A 2 2 2 z  2 2 2
40 [ + a ] 3/2 8 0 (2 + a2 )3
2

z=0

This can be checked to have the correct dimensions, because q 2 /(0 2 ) is Coulomb force. The
greatest concentration of force density lies directly under the charge, as expected.

c) (20) The total electric force on the plane, by integrating the force per area. Is it the
result you expect?
Using a symmetrical area element, dA = 2 d, do the integration,

(qa)2 2 d (qa)2 d(2 + a2 ) (qa)2 1 q2

Z Z Z
Fz
Fz = dA = 2 2 2 3
= 2 3
= 4
= .
A 8 0 0 ( + a ) 80 0 ( + a2) 80 2a 160 a2

This is the same as Fz = q 2 /(40 d2 ) where d = 2a is the separation of the charge an its image,
so the result is the one expected.

4
2. A 2D region, a, 0 is bounded by conducting
surfaces at = 0, = 0 and = held at zero potential
(curved 2D corner, see Fig.). The potential is determined by
some charges far from the region.

a) (20) Write down a solution for the potential (, ) that satisfies the boundary conditions
for finite . It may have undetermined constants.
The basic solutions to 2D electrostatics as found from the Laplace equation must contain the
dependencies,
(a + b )(A ei + B ei )
with some undetermined constants. There is also a logarithmic term possible for = 0 az-
imuthally symmetric part of the solution, a0 + b0 ln . The total general solution looks
like

X
= a0 + b0 ln + (a + b )(A ei + B ei ).
=1

One must choose constants to force = 0 at = a and also at = 0 and = . The angular
terms can be made to obey the boundary conditions using sine functions, with m/.
 
i i m
(A e + B e ) Am sin , m = 1, 2, 3, ...

Also, the radial coefficients must be chosen so each term is zeroed at a,
 
am m/ + bm m/ = 0 = bm = am a2m/ .
=a

It is impossible to satisfy the boundary conditions at = 0, , if the logarithmic terms are

present. Hence, one needs to chose a0 = b0 = 0. So a solution that can satisfy the BCS is
 
X 
m/ 2m/ m/
 m
= Am a sin .
m=1

b) (20) In the case that = , the problem is that of a half cylinder on an infinite plane.
To leading order, determine how the surface charge density depends on along the
boundary at = 0, > a.
The surface charge density is again obtained from = 0 En , in terms of the component of the
electric field along the outward normal from a conducting surface. For the surface at = 0,
> a, the outward normal is in the direction. Therefore we do

1  1 X   m
= 0 E = 0 =  0 Am m/ a2m/ m/ cos(0).
=0
 m=1

The leading dependence is the first term, m = 1. With / = 1, this just gives

a2 a2
   
1
0 A1 = 0 A1 1 2 .

This goes to zero at = a, and nearly a constant for large . Far from the half-cylinder, there
will be a nearly uniform electric field perpendicular to that surface, while charge is sucked out
of the corner at = a. See Jackson problem 2.26 for more details.

5
3. For a point x in some volume V bounded by a surface S, Greens theorem can be arranged
into an integral statement about electrostatic potential,

1 (x0 ) 3 0 1 1 1
Z I   
(x) = d x + 0
0 da0 (1)
40 V R 4 S R n n R

where R = |x x0 | and the other symbols have their usual meanings.

Consider proving the mean-value theorem: For charge-free space the electrostatic potential
at a point is its average over the surface of any sphere centered on that point.

a) (5) To prove this theorem, first explain why the first term in Eq. 1 will be zero for the
specified situation.
There is no charge in the considered volume, (x0 ) = 0. That makes the first integral identically
zero.

b) (10) Next, give a convincing mathematical argument showing why the second term in
Eq. 1 will be zero for the specified situation.
Note that field point x is the center of a sphere, of radius R, whose surface is the surface of
integration for the last two terms. The vector R = x x0 points from that spherical surface,
to the field point x at the center of the sphere.

In the first surface integral term, there is the gradient of the potential, /n0 . But n0 is an
axis pointing out of the volume V , perpendicular to its surface. Then this gives the normal
component of the electric field at the surface, En0 = E n0 = /n0 . Because R is a constant
on the sphere, it comes out of the integrand. Then by Gauss Law we get
1 1 qenc
I I
1 0
0
da = E n0 da0 = = 0.
4R S n 4R S 4R 0

There is no enclosed charge in the spheres volume, by assumption, so the integral is zero.

c) (15) Finally, give a convincing mathematical argument showing what the value of the
last term in Eq. (1) is, and explaining how this proves the theorem.
The gradient of 1/R is needed for the last term, with respect to a coordinate n0 pointing outward
out of the sphere volume. But that radial coordinate is the spheres radius R. So we have
   
1 1 1
= = 2.
n0 R R R R

Again, this is a constant on the sphere, so now there results

   I
1 1
I I
1 1 1
(x) = (x0 ) 0 da0 = (x 0
) da 0
= (x0 ) da0
4 S n R 4 R2 S 4R 2
S

Of course, the integral is over the surface of the sphere, then divided by the spheres area. That
is the usual definition of the potential averaged over the surface of the sphere. So this proves
the mean-value theorem. Keep in mind that it only works in a charge-free region.