Problem 1.

Lienard-Wiechert for constant velocity

(a) For a particle moving with constant velocity v along the x axis show using Lorentz
transformation that gauge potential from a point particle is
e
Ax (t, x, x? = b) = p (1)
4⇡ b2 + 2 (x vt)2

at the point (t, r) = (t, x, y, z) = (t, x, b). So at the point (t, 0, b, 0) the gauge potential
Ax is is
e
Ax (t, x, y = b) = p (2)
4⇡ b2 + ( vt)2

(b) Start by noting the definitions

R R
T ⌘t R = |r r⇤ (T )| R ⌘ Rn ⌘ r r⇤ (T ) n⌘ (3)
c R
and drawing a picture for yourself. Then, after setting c = 1 and v = to simplify
algebra, show that the Lienard Wiechert result,

e v/c
A(t, r) = . (4)
4⇡ R(1 n · ) ret

gives the same result as Eq. (2).

(c) Show that the Lienard-Wiechert potential, Eq. (4), and analogous equation for ' can
be written covariantly 
µ e Uµ
A (X) = , (5)
4⇡ U · X ret
where X µ is the di↵erence in the space-time coordinate four vectors of the emission
and observation points, and U µ is the four velocity of the particle. What is X · X ⌘
X µ Xµ ? Can []ret be expressed covariantly?

1
Problem 2. Radiation during perpendicular acceleration
Consider an ultrarelativistic particle of velocity experiencing an acceleration a? perpen-
dicular to the direction of motion. Here a? points along the x-axis and points along the
z-axis.

(a) Show that the energy radiated per retarded time is approximately

dW e2 a2? sin2 ✓ cos2
= 1 (6)
dT d⌦ 16⇡ 2 c3 (1 cos ✓)3 2 (1 cos ✓)2

e2 a2? 4( ✓)2 cos2
' 2 3 1 (7)
2⇡ c (1 + ( ✓)2 )3 (1 + ( ✓)2 )

In the first equality, I give the full answer without approximation, but I will only grade
the second approximate result.
Hint, in working out this radiation pattern you might (as a start) show without ap-
proximation that

|n ⇥ (n ) ⇥ a|2 = (1 n · ) 2 a2 (n · a)2 (1 2
) (8)

by using the ”b(ac)-(ab)c” rule. Then select a coordinate system were

=(0, 0, ) (9)
a =(a? , 0, 0) (10)
n =(sin ✓ cos , sin ✓ sin , cos ✓) (11)

(b) Work in a ultra-relativistic approximation, and compute the total power by integrat-
ing over the solid angle (as done in class) to show that you obtain the appropriate
relativistic Larmour result
dW
= come on . . . you know it . . . right? (12)
dT

2
Problem 3. An oscillator radiating
(a) Determine the time averaged power radiated per unit sold angle for a non-relativistic
charge moving along the z-axis with instantaneous position, z(T ) = H cos(!o T ).

(b) Now consider relativistic charge executing simple harmonic motion. Show that the
instantaneous power radiated per unit solid angle is

dP (T ) dW e2 c 4 sin2 ✓ cos2 (!o T )

= = (13)
d⌦ dT d⌦ 16⇡ 2 H 2 (1 + cos ⇥ sin !o T )5
p
Here = !o H/c and = 1/ 1 2

(c) In the relativistic limit the power radiated is dominated by the energy radiated during
a short time interval around !o T = ⇡/2, 3⇡/2, 5⇡/2, . . .. Explain why. Where does
the outgoing radiation point at these times.

(d) Let T denote the time deviation from one of this discrete times, e.g. T = 3⇡/(2!o ) +
T . Show that close to one of these time moments:
dP ( T ) dW 2e2 c 4 6 ( !o T )2 ( ✓)2
= ' 2 2 (14)
d⌦ d T d⌦ ⇡ H (1 + ( ✓)2 + ( !o T )2 )5

(e) By integrating the results of the previous part over the T of a single pulse, show that
the time averaged power is

dP (T ) e2 c 4 5 5( ✓)2
= (15)
d⌦ 128⇡ 2 H 2 (1 + ( ✓)2 )7/2

(f) Make rough sketches of the angular distribution for non-relativistic and relativistic
motion.

3
Problem 4. Radiation during a collision
A classical non-relativistic charged particle of charge q and mass m is incident upon a repul-
sive mechanical potential U (r)
A
U (r) = 2 ,
r
so that the force on the particle is F = rU (r). The particle moves along the x-axis and
strikes the central potential head on as shown below. The incident kinetic energy (i.e. the
kinetic energy of the particle far from the origin) is K.

origin

(a) (2 points) Determine the particle’s classical trajectory x(t). Adjust the integration
constants so that the particle reaches its distance of closest approach at t = 0. Check
that for late times x(t) approaches vo t with the phyically correct value of vo . Check
that for small times x(t) behaves as x(t) ' xo + 12 ao t2 with the physically correct value
of xo .

(b) (4 points) Use dimensional reasoning and the Larmour formula to estimate the total
energy lost to electromagnetic radiation during the collision. How does the energy lost
scale with the incident velocity?

(c) (2 points) Calculate the energy lost to radiation during the collision processes. Some
relevant integrals are given at the end of this problem.

Now consider a detector placed along the y-axis far from the origin as shown below. The
front face of the detector has an area of ⇡R2 , and the detector is placed at a distance L from
the origin with L R.

R2
detector

origin

(d) (2 points) What is the direction of polarization of the observed light in the detector?
Explain.

(e) (2 points) What is the typical frequency of the photons that are emitted at 90o ?
Explain.

4
 (f) (5 points) For the detector described above, determine the average number of photons
received by the detector per unit frequency:
dN
. (16)
d!
Some relevant integrals are given at the end of the problem.
(g) (3 points) We have determined the photon radiation spectrum using classical electro-
dynamics. For what values of the parameters A and K is this approximation justified?

Useful integrals and formulas:

(a) For positive integer n, we note the integrals
Z 1
1
du = ⇡ cn (17)
1 (1 + u2 )n
where
1 3 5
c1 , c2 , c3 , c4 , . . . = 1, , , , . . . (18)
2 8 16
(b) For positive integers n, we note the integrals
Z 1
cos(xu) n
du 1 = cn x Kn (x) (19)
0 (u2 + 1)n+ 2
where
1 1 1
c1 , c2 , c3 , c4 , . . . = 1, , , ,... (20)
3 15 105
and Kn (x) are the modified Bessel functions, and the RHS of Eq. (19) is illustrated
below
1
n=1
0.9 n=2
0.8 n=3
n=4
0.7
cn x Kn(x)

0.6
0.5
n

0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7
x

5
Solution

(a) It is convenient to use dimensionless variables. The dimensional constants for the clas-
sical problem are
K, A, m , (6)
from which we can select a unit for meters, seconds, and kilograms. The unit of velocity can
be taken as r
1 2 2K
K = mvo =) vo ⌘ , (7)
2 m
which phyiscally is the velocity as r ! 1. The unit of meters is
r
A A
K = 2 =) xo ⌘ , (8)
xo K

which (by energy conservation) is the distance of closest approach2 . The unit of seconds is
therefore p
xo (mA)/2
to ⌘ ⌘ . (9)
vo K
We need to solve for the trajectory x(t). The velocity is given by the first integral (i.e.
enegy conservation)
1 2 A
mv (t) + =K. (10)
2 x(t)2
Switching to dimensionless variables,

v x(t)
v̄ = , x̄ = , (11)
vo xo
the dimensionless form of energy conservation reads
1
v̄ 2 + = 1. (12)
x̄2

Solving Eq. (12) for v̄ we have r

1
v̄ = . 1 (13)
x̄2
Finally, we write v̄ = dx̄/dt̄ and integrate Eq. (13) to find
p
x̄2 1 = t̄ + constant . (14)
2
Note that this is the distance of closest approach in the absence of energy loss due to radiation. In
the limit of classical electrodynamics one first determines the trajectories of charged particles (ignoring the
radiation), and then soves for the subsequent radiation. This is in e↵ect ignoring radiations back reaction.

17
We choose the integration constant to be zero, so that at t = 0 the trajectory is at the
turning point x̄ = 1 and find
p p
x̄2 1 = t̄ or x̄(t̄) = 1 + t̄2 (15)

Restoring units, the trajectory is

r
2Kt2 A
x(t) = + . (16)
m K
It is easy to check that this trajectory satisfies the appropriate limits.

(b) The energy lost to radiation is

Z 1
q 2 2a2
Eloss = dt (17)
1 4⇡ 3c3

We need to use dimensional reasoning to estimate a and the time interval over which the
acceleration is active.
Using the dimensional analysis of the previous section, the integral is of order
Z
2 vo2
dt a ⇠ . (18)
to
and thus
q 2 vo2 q2K 2
Eloss ⇠ ⇠ p . (19)
4⇡to c3 m Amc3
The energy lost scales as the velocity to the fourth power, K 2 / vo4 .

(c) We next evaluate the integral in Eq. (17) precisely. For reference we record the acceler-
ation:
vo d2 x̄ vo 1
a(t) = = . (20)
to dt̄2 to (1 + t̄2 )3/2
The relevant integral is
Z 1 Z 1 ✓ ◆2 Z 1 ✓ ◆
v2 d2 x̄ v2 dt̄ vo2 3⇡
dt a = o
2
dt̄ = o = . (21)
1 to 1 dt̄2 to 1 (1 + t̄2 )3 to 8

The energy lost is therefore

Z 1
q2 2
Eloss = dt a2 , (22)
4⇡ 3c3 1
q 2 vo2 ⇣ ⇡ ⌘
= . (23)
4⇡to c3 4

18
(d) The radiation electric field is
q
Erad (t, r) = n ⇥ n ⇥ a(te ) , (24)
4⇡rc2
where the emission time is
r
te = t . (25)
c
For the problem at hand a = a(t)x̂ and thus

n ⇥ n ⇥ x̂ = x̂ . (26)

So the radiation field is polarized in the x̂ direction.

(e) The typical frequency is given by dimension reasoning

1
!o ⇠ (27)
to

(f) To determine the yield of photons, we Fourier transform the radiation field and square
this Fourier transform. The Fourier transform of the electric field (in the x̂ direction)
reaching the detector Z 1
q
Erad (!, r) = dte i!t a(te ) . (28)
4⇡rc2 1
After switching to variables to integrate over the emission time,
!
ei!t = ei!(te +r/c) = eikr ei!te , k⌘ , (29)
c
the integral reads Z 1
q eikr i!te
Erad (!, r) = dte e a(te ) . (30)
4⇡rc2 1

Thus, after switching to dimensionless variables

te
!
¯ = !to t̄e = , (31)
to
Thus we find
Z 1
q eikr 1
Erad (!, r) = 2
vo dt̄e e i¯!t̄e , (32)
4⇡rc 1 (1 + t̄2e )3/2
q eikr
= vo [2¯
! K1 (¯! )] . (33)
4⇡rc2
Squaring the radiation field, we find the yield of photons
dN c
~! = |rErad (!, r)|2 . (34)
d!d⌦ ⇡

19
Assembling the ingredients, and expressing the result in terms of the fine structure constant
↵ = q 2 /(4⇡~c) = 1/137, we find

dN ↵ ⇣ v o ⌘2 1
= 2 [2¯ ! )]2 .
! K1 (¯ (35)
d!d⌦ 4⇡ c !
The solid angle is ⌦ = ⇡R2 /L2 , and thus we find

dN ⇡R2 ↵ ⇣ vo ⌘2 1
= 2 [2¯ ! )]2 .
! K1 (¯ (36)
d! L 4⇡ 2 c !

In the low frequency limit the term in brackets approaches

[2¯ ! )]2 ! 22 ,
! K1 (¯ (37)

and thus in low frequency limit we find

✓ ◆2
dN ↵ 2vo 1
= 2 . (38)
d!d⌦ 4⇡ c !

Notice that this expression is independent of A, and is in fact indentical to the radiation for
impulsive scattering where v(t) changes instantaneously:
(
vo x̂ t < 0
vimpulse (t) = . (39)
vo x̂ t>0

Indeed, in the low frequency limit the radiated waves do not have the temporal resolution
to resolve events of order the collision time to . Thus, as far as the radiation of these low
frequency waves is concerned, the collision happens instantaneously.

(g) To determine the validity of the classical approximation, we note that the typcal fre-
quency is 1/to . The energy of the emitted photon has to be small compared to the kinetic
energy of the particle for the classical approximation to be valid

~! ⌧ K . (40)
1 K
With to
=p , we find
(mA)/2
2~2
⌧ 1. (41)
mA

20
Problem 5. Periodic pulses
Consider a periodic motion that repeats itself with period To . Show that the continuous
frequency spectrum becomes a discrete spectrum containing frequencies that are integral
multiples of the fundamental, !o = 2⇡/To .
Let the electric field from a single pulse (or period) be E1 (t), i.e. where E1 (t) is non-
zero between 0 and To and vanishes elsewhere, t < 0 and t > To . Let E1 (!) be its fourier
transform.
(a) Suppose that the wave form repeats once so that two pulses are received. E2 (t) consists
of the first pulse E1 (t), plus a second pulse, E2 (t) = E1 (t) + E1 (t To ). Show that the
Fourier transform and the power spectrum is
E2 (!) = E1 (!) (1 + ei!To ) |E2 (!)|2 = |E1 (!)|2 (2 + 2 cos(!To )) (21)

(b) Now suppose that we have n (with n odd) arranged almost symmetrically around t = 0,
i.e.
En (t) = E1 (t+(n 1)To /2)+. . .+E1 (t+To )+E1 (t)+E1 (t To )+. . . E1 (t (n 1)To /2) ,
(22)
so that for n = 3
E3 (t) = E1 (t + To ) + E1 (t) + E1 (t To ) . (23)
Show that
sin(n!To /2)
En (!) = E1 (!) (24)
sin(!To /2)
and ✓ ◆2
2 2 sin(n!To /2)
|En (!)| = |E1 (!)| (25)
sin(!To /2)

(c) By taking limits of your expressions in the previous part show that after n pulses, with
n ! 1, we find
X 2⇡
En (!) = E1 (!m ) (! !m ) (26)
m
T o

and X 2⇡
|En (!)|2 = nTo ⇥ |E1 (!m )|2 (! !m ) (27)
|{z} To2
m
total time
where !m = 2⇡m/To .
Remark We have in e↵ect shown that if we define
X 1
(t) ⌘ (t nTo ) . (28)
n= 1

Then the Fourier transform of (t) is

X X 2⇡
ˆ (!) = e i!nTo = (! !m ) . (29)
n m
T o

6
 n=2 n=5
4 25
3 20
15
2
10
1 5
0 0
0 5 10 15 20 25 0 5 10 15 20 25
100
80 ⇣ ⌘2
sin(n!To /2)
60 !To /2
40
20 n = 10
0
0 5 10 15 20 25

7
(d) Show that a general expression for the time averaged power radiated per unit solid
angle into each multipole !m ⌘ m!o is:

dPm |rE(!m )|2

= (30)
d⌦ To2

Or Z 
To 2
dP̂m e2 !o4 m2 n · r⇤ (T )
= v(T ) ⇥ n exp i!m (T ) dT , (31)
d⌦ 32⇡ 4 c3 0 c
Here dP̂m /d⌦ is defined so that over along time period T , the energy per solid angle
is 1
dW X dP̂m
= T (32)
d⌦ m=1
d⌦
Also note that we are summing only over the positive values of m which is di↵erent
from how we had it in class:

dP̂m dPm dP m
⌘ + (33)
d⌦ d⌦ d⌦

8
Problem 6. Radiation spectrum of a SHO
(a) Show that for the simple harmonic motion of a charge discussed in Problem: An Oscil-
lator Radiating, the average power radiated per unit solid angle in the m-th harmonic
is
dP̂m e2 c 2
= 2 2 m2 tan2 ✓ [Jm (m cos ✓)]2 (34)
d⌦ 8⇡ H
(b) Show that in the non-relativistic limit the total power radiated is all in the fundamental
and has the value
e2 2 4 2
P = ! H (35)
4⇡ 3 o
where H 2 is the mean squared amplitude of the oscillation.

9
Problem 7. (Optional) Energy during a burst of deceleration
A particle of charge e moves at constant velocity, c, for t < 0. During the short time
interval, 0 < t < t its velocity remains in the same direction but its speed decreases
linearly in time to zero. For t > t, the particle remains at rest.

(a) Show that the radiant energy emitted per unit solid angle is

dW e2 2 (2 cos ✓) [1 + (1 cos ✓)2 ] sin2 ✓

= (36)
d⌦ 64⇡ 2 c t (1 cos ✓)4

(b) In the limit 1, show that the angular distribution can be expressed as

dW e2 2 4
⇠
' (37)
d⇠ 4⇡ c t (1 + ⇠)4

where ⇠ = ( ✓)2 .

(c) Show for 1 that the total energy radiated is in agreement with the relativistic
generalization of the Larmour formula.

10