Electrodynamics FS 2015

Exam Solutions Prof. C. Anastasiou

Name: Student number:

Exercise Max. points Points Visum 1 Visum 2

1 15
2 15
3 15
4 15
Total 60

• The exam lasts 180 minutes.

• Start every new exercise on a new sheet.

• Write your name on every sheet you hand in.

• Do not use red color or pencil.

• You are only allowed to have a single two-sided handwritten formular.

1
2
Exercise 1. Spherical conductor inside a uniform electric field [15 points]

The general solution of the Laplace equation in spherical coordinates is

∞ X
l  
X
l Blm
Φ(r, θ, φ) = Alm r + l+1 Ylm (θ, φ), (1)
r
l=0 m=−l

where the spherical harmonics are given by

s
2l + 1 (l − m)! m
Ylm (θ, φ) = P (cos θ) eimφ . (2)
4π (l + m)! l

a) (1 point) How does Eq. (1) simplify for potentials with azimuthal symmetry?

A spherical conductor of radius R is placed in an (initially) uniform electric field E = E0 ẑ. We

place the origin of the coordinate system at the center of the sphere and we denote by θ the polar
angle with respect to the ẑ-axis. The potential of the conductor is set to zero Φ(r = R) = 0.

b) (3 points) What is the potential Φ(r) very far away from the conductor (r  R)?

c) (3 points) Show that the potential outside the sphere is given by

R3
   
R
Φ(r, θ) = −E0 r − 2 cos θ + C 1 − , (3)
r r

where C is a constant.

Recall: The Legendre polynomials Pl (x) are given by

1 dl
l
Pl (x) = l l
x2 − 1 . (4)
2 l! dx

d) (3 points) Find the induced charge density σ(θ) on the surface of the sphere, if the net
charge of the conductor is Q. What is the constant C of Eq (3) in that case?

e) (2 points) Calculate from first principles (e.g. Gauss’ law) the pressure exerted on a surface
element of the sphere for Q = 0.

f) (3 points) What is the force acting on the north hemisphere of the sphere and what is the
total force acting on the sphere for Q = 0?

Solution.

a) In the case of azimuthal symmetry the potential does not depend on φ, thus we only have the term m = 0
in the above sum. Moreover, since the functions Yl0 and Pl only differ by a normalization constant, we
introduce new parameters Al and Bl instead of Al0 and Bl0 . The resulting expression is
∞  
X Bl
Φ(r, θ) = Al rl + l+1 Pl (cos θ). (S.1)
r
l=0

b) At large distances from the sphere the potential does not go to zero, but we must recover the unperturbed
uniform field E = E0 ẑ. The potential (recall: E = −∇Φ) in this case is thus given by Φ(r  R) =
−E0 z + C = −E0 r cos θ + C, where θ is the angle with respect to the ẑ-axis.

3
c) The first condition we use is the surface potential Φ(r = R) = 0 and Eq. (S.1). They implies
Bl
Al Rl + =0 ⇒ Bl = −Al R2l+1 . (S.2)
Rl+1
Eq. (S.1) can thus be rewritten as
∞
R2l+1
X  
l
Φ(r, θ) = Al r − l+1 Pl (cos θ). (S.3)
r
l=0

Using now as second condition the one obtained in part (b), and noting that for r  R the second term in
parentheses in Eq. (S.3) is negligible, we have
∞
X
− E0 r cos θ + C = Al rl Pl (cos θ). (S.4)
l=0

Knowing that P0 (cos θ) = 1 and P1 (cos θ) = cos θ we conclude that A0 = C, A1 = −E0 and Ak = 0 ∀ k ≥ 2.
Finally we get
R3
   
R
Φ(r, θ) = −E0 r − 2 cos θ + C 1 − . (S.5)
r r
d) Here we use the discontinuity of the electric field across a surface charge σ, which is described by the
equation Eabove − Ebelow = σ/0 n̂, where above and below refers to the regions of space separated by the
surface and close to it. The vector n̂ is a unit vector pointing “upwards”. [It is simply derived taking a
small enough surface element in which σ is constant and applying Gauss’ and Stokes’ theorems with a thin
pillbox resp. thin rectangular loop.] Using the fact that E = −∇Φ and that the field inside (here Ebelow )
is zero allows us to find the induced charge density σ by looking at the derivative of the potential in the
direction normal to the surface. Using Eq. (S.5) we obtain
∂Φ  C

σ(θ) = −0 = 30 E0 cos θ − 0 . (S.6)
∂r r=R R


Now we use the fact that the net charge of the sphere is Q, which gives
Z Z Z
C
Q= σ ds = 30 E0 cos θ ds − 0 ds. (S.7)
S S R S

The first integral gives 0, while the second one simply gives 4πR2 . We thus see that C = − 4πQ0 R and the
potential in Eq. (3) reads

R3
   
Q 1 1
Φ(r, θ) = −E0 r − 2 cos θ + − . (S.8)
r 4π0 r R

e) We start by considering the force acting on an infinitesimal surface element due to the presence of everything
else, except the element itself (since it does not produce a force on itself). The force per unit surface
(electrostatic pressure) is given by f = σEav , where Eav = (Ein + Eout )/2 is the average field at the
position of the surface element. Here, since we have a conducting sphere, we have that Ein = 0 and the
field just above the conducting surface is given by Eout = σ/0 n̂ (again from the discontinuity of the electric
field across a surface charge: Eout − Ein = σ/0 n̂). Therefore the total pressure on a surface element is
given by the expression
σ2
P = , (S.9)
20
f) From symmetry, the components of the final force in the xy-plane do compensate and the resulting force
only has a z-component given by

σ2 9
fz = cos θ ds = 0 E02 cos3 θ ds, (S.10)
20 2
where θ is the angle between n̂ and ẑ and in the last step we used the surface charge density σ(θ) found
in part (d) when Q, i.e. C, is 0. We now integrate Eq. (S.10) over the surface of the northern hemisphere
(NH) in order to get the total force
Z Z 2π Z π/2
9
F = fz = 0 E02 R2 dφ cos3 θ sin θ dθ. (S.11)
NH 2 0 0

4
The integral over θ is easily solved using the substitution u = cos θ (du = − sin θ dθ) giving
Z 1
9
F = 9π0 E02 R2 u3 du = π0 E02 R2 . (S.12)
0 4
The total force acting on the sphere is found in the same way, but the integral over θ runs between 0 and
π instead of 0 and π/2, therefore giving F = 0.

5
Exercise 2. Magnetic field of a charged rotating spherical shell [15 points]

Figure 1: Charged rotating spherical shell. The external magnetic field refers only to part 5.

Consider a spherical shell of radius R with a charge Q uniformly distributed on its surface, which
rotates around its diameter with constant angular velocity ω ~ (see Fig. 1).

a) (2 points) Calculate the current density ~j.

b) (4 points) Show that the vector potential A ~ inside and outside the sphere is
 Q
 12π0 c2 R ω
 ~ × ~r r < R
~
A=
 QR2 ω

~ × r~r3 r > R
12π0 c2

The following integral may be used:

√ √
Z 1
ξdξ 2  
= (2 − α) 1 + α − (2 + α) 1 − α .
−1 (1 − αξ)
1/2 3α2

~ in the interior and the exterior of the spherical

c) (4 points) Determine the magnetic field B
shell.

d) (2.5 points) Determine the magnetic dipole moment m

~ for the rotating spherical shell.

e) (2.5 points) Assume now that the sphere is immersed in an external magnetic field

~ ext. = k ~el
B
r
where k is a constant and the angle between ~el and ω
~ is θ0 .Calculate the force acting on
the surface of the sphere due to the external magnetic field and determine the associated
potential energy.

Solution.

a) The current density is given by ~j = ρ~v , ~v = ω

~ × ~r, and the charge density is ρ = σδ(r − R) with
σ = Q/(4πR2 ). Then we get
~j(r~0 ) = Q δ(r0 − R)(~ω × r~0 ), (S.13)
4πR2

6
b) For the vector potential we have
1
Z ~j(r~0 )
~=
A dV 0 . (S.14)
4π0 c2 |~r − r~0 |
and so we obtain
~= Q ~ (~r)
A ~ ×F
ω (S.15)
c2 16π 2 0 R2
with
δ(r0 − R) ~0 0
Z
~ (~r) =
F r dV . (S.16)
|~r − r~0 |
Because of rotation symmetry it follows that F ~ = f · ~r/r. We choose the ~r-direction to be the z 0 -axis, then
θ is the angle between ~r and r~0 . By using dV 0 in spherical coordinates and integrating over φ0 we get
0

Z ∞
δ(r0 − R)r0 cos θ0 2πr02 sin θ0
Z π
F = dθ0 dr0 (S.17)
0 0 (r2 + r02 − 2rr0 cos θ0 )1/2
2πR3 cos θ0 sin θ0
Z π
= dθ0 2 . (S.18)
0 (r + R2 − 2rR cos θ0 )1/2
We make the substitution cos θ0 = ξ and we get
Z −1
ξdξ
F = −2πR3 . (S.19)
+1 (r2 + R2 − 2rRξ)1/2
This is an integral of the form
Z Z
ξdξ 1 ξdξ
I= = √ , (S.20)
(A − Bξ)1/2 A (1 − αξ)1/2
with A = r2 + R2 , B = 2r · R , α = B/A. We can then use the hint and finally obtain
2πR  2 2 2 2

F = (r + R − rR)(r + R) − (r + R + rR)|r − R| . (S.21)
3r2
Inside the sphere is r < R and so |r − R| = R − r and
4π
F = Rr. (S.22)
3
Outside we find
4π R4
F = . (S.23)
3 r2
The solution for the vector potential is
 Q

 12π0 c2 R
ω
~ × ~r r<R
~=
A
QR2 ~r


12π0 c2
ω
~ × r3
r>R

c) We can then get the magnetic field B ~ by taking B ~ =∇ ~ × A,

~

~ = Q 
~ · ~r) − (~ ~ r

B 2
~ (∇
ω ω · ∇)~ r < R. (S.24)
12π0 c R
Which gives, after explicit computation,
~ = Q
B ω
~ r < R. (S.25)
6π0 c2 R
For r > R we obtain,  
QR 2 3 ω~ · ~rr ~rr − ω
~ ~ · ~r)~r − r2 m
3(m ~
~ =
B ≡ . (S.26)
12π0 c 2 r 3 r5
~
d) We see that the B-field inside the sphere is homogeneous and proportional to ω. Outside the sphere we
find the behaviour of a dipole with
QR2 ω~
m~ = . (S.27)
12π0 c2
e) The potential is the one acting on the dipole moment and is given by

~ =− QR2 2
~ ext. = − QR ωBext. cos(θ0 ) = − QRk ω cos(θ0 ).
U = −m
~ ·B 2
~ ·B
ω (S.28)
12π0 c 12π0 c2 12π0 c2
The corresponding force is
~ = −∇U = − kQ ω cos(θ0 )~er
F (S.29)
12π0 c2

7
Exercise 3. Special Relativity [15 points]

A point-like charge q of mass m is initially at rest and it gets accelerated by a constant force F~ .

a) (2 points) What is the velocity of the charge as a function of time?

b) (2 points) What is the energy of the charge as a function of time?

c) (2 points) What is the distance travelled by the charge as a function of its energy?

d) (2 points) Calculate the derivative

dτ
, (5)
dE
where τ is the proper time of the particle and E the energy of the charge, as a function of
the energy.

e) (4 points) Calculate the instantaneous force felt by the charge in its own rest frame during
the linear acceleration.

f) (3 points) Assume now that the constant force F~ originates from a constant electric field.
What is the electromagnetic field strength tensor F µν in the instantaneous rest frame of
the charge? Is your answer consistent with the result of the previous question?
Hint. Recall that
−E 1 −E 2 −E 3
 
0
E 1 0 −B 3 B2 
F µν =
E 2
.
B3 0 −B 1 
E3 −B 2 B1 0

Solution.

mv
a) Since F = dp/dt, we have that p(t) = F t. Now on the other hand p(t) = γ(t)mv(t) = √ , and so
1−v 2 /c2
√ Ft
solving for v yields v(t) = .
m 1+F 2 t2 /(m2 c2 )

b) The energy is given by

√ r
mc2 mc2 mc2 m2 c2 + F 2 t2 F 2 t2
E(t) = γmc2 = p = q = = mc2 1 + 2 2 .
1 − v 2 /c2 2 t2
1 − m2 cF2 +F mc m c
2 t2

c) To find the distance, integrate v(t0 ) up to time t. We get

Z t " r #t
0 0 mc2 F 2 t2
x(t) = v(t )dt = 1+ 2 2 .
0 F m c
0

In terms of the energy, this yields

E − mc2 Ekin
x(E) = = .
F F
R
This result could have also been obtained through E = F dx together with the knowledge that F is
constant.
dt dτ 1 dt
d) To get the proper time as a function of the energy, note that dτ = γ
and so dE
= γ dE
. Then, we use

F 2t Fp
dE/dt = q = q
F 2 t2 F 2 t2
m 1 + m2 c2 m 1+ m 2 c2

to get q
F 2 t2
dτ m 1+ m2 c 2
= ,
dE γ Fp

8
 yielding
dτ E E 1 m
= 2 = p = p = p
dE c γF p F c2 γ γ 2 − 1mc F c E 2 /(mc2 ) − 1 F (E 2 − m2 c4 )
p
as p = γmv = γ 2 − 1mc.
e) We calculate this using the four-force, but we can do this both starting from the rest frame of the particle
or from the lab frame. In the rest frame, the four-force takes the simple form f 0µ = (0, F ~ 0 ). Assuming the
particle moves in the x direction, we can transform this with a Lorentz transformation of speed v to the
0ν
lab frame and obtain f µ = Lµ νf = (γvFx0 , γFx0 , Fy0 , Fz0 ). Now we know that in the lab frame f i = γFi for
i ∈ {1, 2, 3}, and so Fx = Fx , Fy = γFy = 0 and Fz0 = γFz = 0.
0 0

To calculate it starting from the lab frame is slightly more complicated: the four-force in the laboratory
frame is given by f µ = (mc dγ ~ ). For convenience, we shall denote by ~γ the vector (γ, 0, 0), and we shall
, γF
dτ
look at the force on the particle when it is travelling in x-direction (without loss of generality). Now, as
we change frame into the particle rest frame, we use a Lorentz transformation of speed v in x-direction,
yielding
dγ ~ , −mcγβ dγ + γ 2 Fx , γFy , γFz ).
f 0µ = Lµ ν f ν = (mcγ − γβ~γ · F
dτ dτ
~ ~
In the case of linear acceleration, F = (F, 0, 0), ~γ · F = γF and

dγ dγ F 2 tγ 2 2 F 2t
=γ = (m c + F 2 t2 )−1/2 = 2 2 ,
dτ dt mc m c
yielding an overall force four vector of

F 2t F 2t
f 0µ = (mcγ 2 2
− γ 2 βF, −mcγβ 2 2 + γ 2 F, 0, 0).
m c m c
The force felt by the particle is hence

F 2t F 3 t2
F 0 = −mcγβ 2 2
+ γ2F = γ2F − 2 2 = F
m c m c
in x-direction.
f) The electric field in this case transforms as Ek0 = Ek for the linear motion, so the results indeed coincide.
To calculate the result with the more general transformation rule, use F 0µν = Lµ ρ Lν σ F ρσ :
For our case of linear acceleration in an E field, the B field vanishes and E 1 = E, so the tensor F takes
the simple form  
0 −E 0 0
E 0 0 0
F µν =  .
 
0 0 0 0
0 0 0 0
With  
γ −p/m 0 0
−p/m γ 0 0
Lµ ν =
 
 0 0 1 0

0 0 0 1

9
 we get
   
γ −p/m 0 0 0 −E 0 0 γ −p/m 0 0
−p/m γ 0 0 E 0 0 0 −p/m γ 0 0
 
F 0µν = Lµ ρ Lν σ F ρσ =
 
 0 0 1 0  0 0 0 0  0 0 1 0
  
0 0 0 1 0 0 0 0 0 0 0 1
  
γ −p/m 0 0 Ep/m −Eγ 0 0
−p/m γ 0 0 −Ep/m 0 0
  Eγ

=
 
 0 0 1 0  0 0 0 0
 
0 0 0 1 0 0 0 0
 2 2 2

0 −Eγ + Ep /m 0 0
Eγ 2 − Ep2 /m2 0 0 0
=
 
0 0 0 0


0 0 0 0
 
0 −E 0 0
E 0 0 0
=
 
0 0 0 0

0 0 0 0
which yields Ek0 = Ek .
We could have alternatively employed tensor notation to solve this question. Then, we would only have to
check the relevant components though (and not need to do the whole transformation).

Exercise 4. Radiation emitted by an accelerated charge [15 points]

~ field generated by a moving charge q is given by

a) (2 points) The electric E
" #
q ~
n × {(~
n − ~
v ) × ˙}
~
v
~ x, t) =
E(~ + O(1/R2 ) (6)
4π (1 − ~v · ~n)3 R
ret
where v is the velocity of the charge,
γ = (1 − v 2 )−1/2 (7)
~ = ~n R, with R = |~x − ~r(tret )|,
R (8)
and ~r is the position vector of the charge. All the quantities on the right hand side of
Eq. 6 are computed at a “retarded time” tret while the electric field on the left hand side
is evaluated at an observation time t.
• What is the equation relating the retarded time and the observation time?
• Prove that
dt
= 1 − ~v · ~n.
dtret
b) (2 points) Show that the Poynting vector for a moving charge can be written as
 2  
~ = E
S ~  ~n − E~ · ~n E.
~ (9)

c) (2 points) Starting from the continuity equation satisfied by the Poynting vector show that
the energy radiated per unit of retarded time and unit of solid angle centered around the
moving charge at the retarded time is given by
dPret dE h
~ · ~nR2
i dt
≡ = S . (10)
dΩ dtret dΩ ret dtret

10
 Figure 2: Angles between ~v˙ , ~v , ~n

d) (4 points) Assume that the velocity of the charge is perpendicular to the acceleration, as
in Fig. 2. Show that the power radiated by the charge as a function of the angles θ, ϕ,
shown in Fig. 2, is

q2 |~v˙ |2 sin2 θ cos2 ϕ

 
dPret
= 1− 2 . (11)
dΩ 16π 2 (1 − v cos θ)3 γ (1 − v cos θ)2

Tip: It may be helpful to prove first the identity

 2
~  = ~c 2 (~n · d)
~n × (~c × d)
 ~ 2 + d~ 2 (~n · ~c)2 − 2(~c · d)(~
~ n · ~c)(~n · d)
~ (12)

e) (2 points) Explain qualitatively in which direction the power radiated is stronger for a
charge moving with a velocity close to the speed of light.

f) (3 points) What is the total power radiated in all directions?

Solution.

a) • The equation relating the retarded time and the observation time is

t − tret = R(tret ) (S.30)

with R(tret ) defined as in eq. (8).

• We have

t = tret + R(tret ) (S.31)

dt d
=1+ R(tret )
dt0 dtret
d p
=1+ x − ~r(tret ))2
(~
dtret
x − ~r(tret )
~
=1+ · (−~v )
|~
x − ~r(tret )|
= 1 − ~n · ~v (S.32)

b) The Poynting’s vector is given by

~=E
S ~ ×B
~ (S.33)
~ is related to E
where the magnetic field B ~ via

~ x, t) = [~n × E]
B(~ ~ ret (S.34)

We then have
~=E
S ~ × (~n × E)
~ (S.35)

11
 and in components
~ k
Si = ijk Ej (~n × E)
= ijk Ej lmk nl Em
= (δil δjm − δim δjl ) nl Ej Em
that is
~ = (~n |E|
S ~ 2−E
~ (~n · E))
~ (S.36)

c) From the continuity equation for the Poynting vector

dω ~ ·S
~=0
+∇ (S.37)
dt
where ω is the energy density, we get
Z Z
dω dE ~ · ~n R(tret (t))2 dΩ
= = S
V dt dt ∂V
dE dP ~ · ~n R(tret (t))2
= =S
dΩ dt dΩ
dE dP dt dPret ~ · ~n R(tret )2 dt
= = =S (S.38)
dΩ dtret dΩ dtret dΩ dtret
d) From eqs. (6) and (S.36) the Poynting vector takes here the form
~ = (~n |E|
S ~ 2−E ~ (~n · E))
~
"  2 #
q2 1  ~n × {(~n − ~v ) × ~v˙ } 
 
= ~
n + O(1/R3 ) (S.39)
16π 2 R2  (1 − ~v · ~n)3
 

ret

Then the power radiated becomes

dPret ~ · ~n) R2 dt
= (S
dΩ dtret
~ 2
= (S · ~n) R (1 − ~v · ~n)
  2
2

n × (~
n − ~
v ) × ˙ 
~
v
q ~
= (S.40)
16π 2 (1 − ~v · ~n)5
(S.41)

First we have to compute

  2
~n × (~n − ~v ) × ~v˙  = (ijk nj (klm (n − v)l v̇m )) · (iab na (bcd (n − v)c v̇d ))
 

= (δil δjm − δim δjl ) (δic δad − δid δac ) (n − v)l v̇m (n − v)c v̇d nj na
= (δcl δjm δad − δdl δjm δac − δcm δjl δad + δdm δjl δac ) (n − v)l v̇m (n − v)c v̇d nj na
 
= (~n − ~v )2 (~n · ~v˙ )2 − 2(~n · ~v˙ ) ~v˙ · (~n − ~v ) (~n · (~n − ~v )) + (~v˙ )2 (~n · (~n − ~v ))2 (S.42)

Now using the fact that ~v˙ ⊥ ~v and |~n|2 = 1, the previous expression reduces to
  2
~n × (~n − ~v ) × ~v˙  = (~n − ~v ) (~n · ~v˙ ) − 2(~n · ~v˙ ) (1 − ~n · ~v ) + (~v˙ ) (1 − ~n · ~v )
2 2 2 2 2
(S.43)
 

Using the polar angles as in the picture, we have

~n · ~v = v cos θ (S.44)
~n · ~v˙ = v̇ cos ϕ cos(π/2 − θ) = v̇ cos ϕ sin θ (S.45)
2 2
(~n − ~v ) = 1 + v − 2 v cos θ (S.46)

Inserting these quantities in (S.43), we find

  2
2 −2
~n × (~n − ~v ) × ~v˙  = (~v˙ ) (1 − v cos θ) − (~v˙ ) γ
2 2
sin2 θ cos2 ϕ (S.47)
 

with γ = (1 − v 2 )−1/2 . Hence the power radiated becomes

q2 |~v˙ |2 sin2 θ cos2 ϕ
 
dPret
= 1 − . (S.48)
dΩ 16π 2 (1 − v cos θ)3 γ 2 (1 − v cos θ)2

12
e) In the relativistic limit (v → c) the maximal power radiated is obtained in the forward direction (θ → 0),
as in the case of linear acceleration.
f) The integral we have to compute is
!
2π π
q2 |~v˙ |2 sin2 θ cos2 ϕ
Z Z 
dϕ dθ sin θ 1− 2 (S.49)
0 0 16π 2 (1 − v cos θ)3 γ (1 − v cos θ)2

The first term in the brackets will give

Z 2π Z π  π
sin θ 1 −2 

dϕ dθ = 2π − (1 − v cos θ)
0 0 (1 − v cos θ) 3 2v 
0
  
1 1 1
= 2π − −
2v (1 + v)2 (1 − v)2
 
1 −4v
= 2π −
2v (1 − v 2 )2
4π
= (S.50)
(1 − v 2 )2

The second term gives

2π π
sin2 θ
Z Z
dϕ cos2 ϕ dθ sin θ =
0 0 (1 − v cos θ)5
2π π
sin2 θ
Z Z
1
= dϕ (1 + cos 2ϕ)
dθ sin θ
0 2 0 (1 − v cos θ)5
1 2 + 6v 2
 
1 1
=π − 2
6v 2 (1 − v 2 )3 3v (1 − v 2 )2
4 1
=π (S.51)
3 (1 − v 2 )3

Combining the two results:

q2 ˙ 2
 
4 4
Pret = |~v | 4πγ 4 − π γ
16π 2 3
q2 2 ˙ 2 4
= |~v | γ (S.52)
4π 3
that is exactly Larmor’s formula for a charge in circular motion.

13