Phys 602 Assignment III SolutionsProblem 9.

A radiating quadrupole consists of a square of side a with charges q at alternate corners. The square rotates with
angular velocity ω about an axis normal to the plane of the square and through its center. Calculate the quadrupole
moments, the radiation fields, the angular distribution of radiation, and the total radiated power, all in the long-
wavelength approximation. What is the frequency of the radiation?
Solution
we may represent the rotating quadrupole as shown in Fig. 1 In cartesian coordinate systems, the traceles quadrupole

q

q q


q

FIG. 1: Quadrupole arrangment for problem 9.2

tensor Qij is defined as

Z
Qij = (3x0i x0j − r02 δij )ρ(x0 )d3 x0 (1)

And the charge distribution is given in terms of the delta function as

 
a a a a
ρ(x0 , t) = qδ(z) δ(x − √ cos ωt)δ(y − √ sin ωt) − δ(x + √ cos ωt)δ(y − √ sin ωt)
2 2 2 2

 
a a a a
+qδ(z) δ(x + √ cos ωt)δ(y + √ sin ωt) − δ(x − √ cos ωt)δ(y + √ sin ωt) (2)
2 2 2 2
This integral is not straight forward. It is a unit step function. Let first perform the integrals separately.
Z Z 
a a a a
x2 ρ(x0 , t)d3 x0 = x2 qδ(z) δ(x − √ cos ωt)δ(y − √ sin ωt) − δ(x + √ cos ωt)δ(y − √ sin ωt)+
2 2 2 2


a a a a
δ(x + √ cos ωt)δ(y + √ sin ωt) − δ(x − √ cos ωt)δ(y + √ sin ωt) dxdydz
2 2 2 2

1 2 √ √ √ √
qa cos2 ωt U [2x − 2a cos ωt] − U [2x + 2a cos ωt] − 1 + U [2y − 2a sin ωt] − U [2y + 2a sin ωt]



=
2

= qa2 (2 cos2 −1) = qa2 cos 2ωt, (3)

1
 2

where U stands for Unit Step Function.

Z Z 
2 0 3 0 2 a a a a
y ρ(x , t)d x = x qδ(z) δ(x − √ cos ωt)δ(y − √ sin ωt) − δ(x + √ cos ωt)δ(y − √ sin ωt)+
2 2 2 2

a a a a
δ(x + √ cos ωt)δ(y + √ sin ωt) − δ(x − √ cos ωt)δ(y + √ sin ωt) dxdydz
2 2 2 2

1 2 2 √ √
√ √

= qa sin ωt − 1 + U [2x − 2a cos ωt] − U [2x + 2a cos ωt] U [2y − 2a sin ωt] − U [2y + 2a sin ωt]
2

= qa2 (2 sin2 −1) = −qa2 cos 2ωt, (4)

Z Z 
0 3 0 2 a a a a
xyρ(x , t)d x = x qδ(z) δ(x − √ cos ωt)δ(y − √ sin ωt) − δ(x + √ cos ωt)δ(y − √ sin ωt)+
2 2 2 2

a a a a
δ(x + √ cos ωt)δ(y + √ sin ωt) − δ(x − √ cos ωt)δ(y + √ sin ωt) dxdydz
2 2 2 2

1 2 √ √
√ √

= qa sin(2ωt) U [2x − 2a cos ωt] + U [2x + 2a cos ωt] U [2y − 2a sin ωt] + U [2y + 2a sin ωt]
4

= qa2 sin 2ωt, (5)

Since z = 0 all integrals related to z are zero. That is

Z Z
xzρ(x , t)d x = yzρ(x0 , t)d3 x0 = 0
0 3 0
(6)

Note that the indices 1 = x, 2 = y and 3 = z. Hence

Z Z  
2 2 2 2 0 3 0 2 2 0 3 0 2 2 −2iωt
Q11 = (3x − (x + y + z ))ρ(x )d x = (2x − y )ρ(x )d x = 3qa cos(2ωt) = 3qa < e . (7)

Z Z  
2 2 2 2 0 3 0 2 2 0 3 0 2 2 −2iωt
Q22 = (3y − (x + y + z ))ρ(x )d x = (2x − y )ρ(x )d x = −3qa cos(2ωt) = −3qa < e . (8)

Z  
Q12 = Q21 = xyρ(x0 )d3 x0 = −3qa2 sin(2ωt) = 3qa2 < ie−2iωt . (9)

Note that Q33 = Q13 = Q31 = Q23 = Q32 = 0. The traceless Quadrupole tensor can be written as
1 i 0
2 −2iωt
Qij = 3qa e i −1 0 (10)
0 0 0

From Eqn.(10) we see that the frequency of radiation is 2ω and hence k = 2ω/c. For the configuration shown in Fig.
1 the electric dipole moment vanish as it consists of two equal dipoles that are antiparallel to each other. Similarly,
the magnetic dipole moment for the present configuration is zero as the net charge is zero and no net current. As
both dipoles are zero what we expect is the electric quadrople term. Hence, we are interested to calculate the this
quadrupole radiation in long-wavelength approximation, we can employe equation (9.44) from Jackson

ick 3 eikr
H= n × Q(n), (11)
24π r
 3

where n = r̂ = x̂ sin θ cos φ + ŷ sin θ sin φ + ẑ cos θ and

3
X
Q(n) = (x̂Q1i ni + ŷQ2i ni + ẑQ3i ni ) = x̂(Q11 n1 + Q12 n2 ) + ŷ(Q21 n1 + Q22 n2 ), (12)
i=1

where other terms Q33 = Q13 = Q31 = Q23 = Q32 = 0. Using the result of equation (10) and noting that n1 =
cos φ sin θ, n2 = sin θ sin φ, equation equation (12) gives

Q(n) = 3qa2 e−i2ωt sin θ(x̂(cos φ + i sin φ) + ŷ(i cos φ − sin φ)) = 3qa2 e−i2ωt eφ (x̂ + iŷ). (13)

The vector product of equation (11) gives

x̂ ŷ ẑ
n × Q = 3qa2 sin θe−i2ωt eiφ sin θ cos φ sin θ sin φ cos θ . (14)
1 i 0

Dropping the time dependent part e−i2ωt and inserting equation (14) in to equation (11) yields

ick 3 eikr iφ
H=− e sin θ(−ix̂ cos θ + ŷ cos θ + ẑi sin θeiφ ). (15)
8π r
The corresponding electric field in the long-wave approximation is
iZ0
E= ∇ × H = Z0 H × n (16)
k
Then angular distribution of the of the power radiation is given by

dP c2 Z0 6 2 c2 Z0 6 ∗
= k |(n × Q(n)) × n| = k (Q (n) · Q(n) − |Q(n) · n|2 )
dΩ 1152π 2 1152π 2

c2 Z0 6
k 2(3qa2 sin θ)2 − (3qa2 sin θ)2 (sin2 θ cos2 φ + sin2 θ sin2 φ)


=
1152π 2

c2 Z0 6 c2 Z0 6 2 4 Z0 ω 6 2 4
k (3qa2 )2 sin2 θ(2 − sin2 θ) = k q a (1 − cos2 θ)(1 + cos2 θ = (q q )(1 − cos4 θ)


= 2 2
(17)
1152π 128π 2π 2 c4
where we used k = 2ω/c. The total power radiated by a rotating quadrupole

Z0 ω 6 2 4 8Z0 ω 6 2 4
Z Z
dP 4
P = Ω = (q q ) ω(1 − cos θ = )q a . (18)
dΩ 2π 2 c4 5πc4

Problem 9.5
(a) Show that for harmonic time variation at frequency ω the electric dipole scalar and vector potentials in the Lorenz
gauge and the long-wavelength limit are

eikr
Φ(x) = n · p(1 − ikr)
4π0 r2

µ0 ω eikr
A(x) = −i p [this is (9.16)]
4π r
where k = ω/c, n is a unit vector in the radial direction, p is the dipole moment (9.17), and the time dependence eiωt
is understood.
(b) Calculate the electric and magnetic fields from the potentials and show that they are given by (9.18).
 4

Solution
(a) In the Lorentz gauage the scalar potential is given by

ρ(x0 , t0 ) 0 |x − x0 |
Z Z
1 3 0
Φ(x, t) = d x dt0 δ(t + − t), (19)
4π0 |x − x0 | c
0 0
with ρ(x, t0 ) = ρ(x0 )e−iωt and t0 = t − |x−x
c
|
. Using these expressions of ρ(x, t0 ) and t equation (19) can be written as

ρ(x0 ) −iωt0 0 |x − x0 |
Z Z
1
Φ(x, t) = d3 x0 dt0 e δ(t + − t)
4π0 |x − x0 | c

ρ(x0 ) |x − x0 |
Z Z
1 ω 0
= d3 x0 ei c |x−x | dt0 e−iωt δ(t0 + − t). (20)
4π0 |x − x0 | c

From equation (20) the position dependent scalar potential can be deduced as

ρ(x0 ) ik|x−x0 | 3 0
Z
1
Φ(x) = e d x. (21)
4π0 |x − x0 |

In the far zone (long wave approximation. kr >> 1), the exponent in (21) oscillates rapidly and determine the behavior
of the potential. For far zone we can approximate

|x − x0 | ' r − n · x0

Hence
1 eikr
Z
0
lim Φ(x) = ρ(x0 )e−ikn·x d3 x0 . (22)
kr→∞ 4π0 r
0
For small size of source that is for source dimension, d << λ, the exponent e−in·x can be expanded in Taylor series as

0 (−ikn · x0 )2 X (−ik)n
e−in·x = 1 + (−ikn · x0 ) + + .... = (n · x0 )n
2 n
n!

Hence
1 eikr X (−ik)n
Z
lim Φ(x) = ρ(x0 )(n · x0 )n d3 x0 (23)
kr→∞ 4π0 r n n!

For dipole term n = 1 and including the 1/r2 term to make it valid every where outside the source in the long wave
approximation , we have

1 eikr 1 1 eikr 1
 Z  
0 0 3 0
Φ(x) = − ik ρ(x )(n · x )d x = − ik n · p, (24)
4π0 r r 4π0 r r

where p = ρ(x0 )x0 d3 x0 . We can proceed in the same way to determine the magnetic potential A
R

J(x0 , t0 ) 0 |x − x0 |
Z Z
µ0 3 0
A(x, t) = d x dt0 δ(t + − t), (25)
4π |x − x0 | c

where J(x0 , t) = J(x0 )e−iωt .

J(x0 ) |x − x0 |
Z Z
µ0 3 0 0
A(x, t) = d x dt0 e−iωt δ(t0 + − t)
4π |x − x0 | c

J(x0 ) ik|x−x0 | |x − x0 |
Z Z
µ0
= d3 x0 e dt0 e−iωt δ(t0 + − t) (26)
4π |x − x0 | c
 5

With |x − x0 | ' r − n · x0 , the position dependent vector potential in (26)

µ0 eikr
Z
0
lim A(x) = J(x0 )eikn·x d3 x0 . (27)
kr→∞ 4πr

with the Taylor expansion given above equation (23)the vector potential reads

µ0 eikr X (−ik)n
Z
lim A(x) = J(x0 )(n · x0 )n d3 x0 . (28)
kr→∞ 4πr n n!

For magnetic dipole term n = 0

µ0 eikr
Z
A(x) = J(x0 )d3 x0 (29)
4π r
1
(x × J)0 d3 x0 . Integrating by part
R
where m = 2
Z Z Z Z Z
∇ (x · J)d x = x (∇ · J)d x + J(∇ · x )d x = x (∇ · J)d + Jd3 x0
0 0 3 0 0 0 3 0 0 0 3 0 0 0 3

Z Z Z I Z Z
3 0 0 0 3 0 0 0 0 0 0
=⇒ Jd x = ∇ (x · J)d x − x (∇ · J)d = 3
(x · J)dS − 3
x (∇ · J)d = − x0 (∇0 · J)d3
S
0 0
The surface integral vanish as r → ∞. Recall that ∇ · J(x0 , t) = − ∂ρ(x
∂t
,t )
= iωρ(x0 , t),
Z Z
Jd3 x0 = −iω x0 ρ(x0 )d3 x0 = −iωp0 (30)

Therefore,

iωµ0 eikr
A(x0 ) = − p (31)
4π r

ikr

~ = 1∇ ~ = − iω (~n × p~) ∂ e
~ ×A ck 2 eikr  1 
H = ~n × p~ 1− (32)
µ0 4π ∂r r 4π r ikr

Choosing p~ = pẑ, ~n = r̂, and r̂ × p~ = −p sin θφ̂

2
~ = − ck p sin θ eikr 1 − 1 φ̂
 
H (33)
4π r ikr
The electric filed can be obtained from

~ = i ∇
E ~ ×H
~
ω0

eikr p sin θ ∂  eikr r

 
ik 1
1 ∂ 1

= −p 1− (sin2 θ)r̂ + 1− θ̂
4π0 r ikr r sin θ ∂θ r ∂r r ikr

eikr
   
~ = 1 2 1 ik ikr
E k (~n × p~) × ~n p · ~n) − p~) 3 − 2 e
+ (3~n(~ (34)
4π0 r r r

Problem 9.7
 6

(a) By means of Fourier superposition of different frequencies or equivalent means, show for a real electric dipole p(t)
that the instantaneous radiated power per unit solid angle at a distance r from the dipole in a direction n is
2
d2 p(t0 )
 
P (t) Z0
= n× ×n
dΩ 16π 2 c2 dt02

where t0 = t − r/c is the retarded time. For a magnetic dipole m(r), substitute (1/c)m̈ × n for (n × ¨p) × n.
(b) Show similarly for a real quadrupole tensor Qαβ (t) given by (9.41) with a real charge density ρ(x, t) that the
instantaneous radiated power per unit solid angle is
2
d3 Q(n, t0 )
 
P (t) Z0
= n× ×n
dΩ 576π 2 c4 dt03

where Q(n, t) is defined by (9.43).

Solution
(a) This problem is about to determine the power at instant of time rather than averaged power over a certain period.
It requires writing the power and the corresponding fields that give the power in terms of instantaneous quantities.
 
dP (t) 2
= r r̂ E(t) × H(t) (35)
dΩ

Note that the actual fields E(t) and H(t) are real and need the Fourier transform of E(x, t) and H(x, t) provided that
E(x, ω) = E(x, −ω)∗ and H(x, ω) = H(x, −ω)∗ .
Z +∞
E(x, t) = E(x, ω)e−iωt dω (36)
−∞

Z +∞
H(x, t) = H(x, ω)e−iωt dω (37)
−∞

Inserting
Z +∞  Z +∞ 
dP (t) 0
= r2 r̂ E(x, ω)e−iωt dω × H(x, ω 0 )e−iω t dω 0 (38)
dΩ ∞ ∞

From equations (31), (32) and (34) above the Fourier transform of the vector potential A, the electric field E and the
magnetic filed H, to the leading order in 1/r are given by

iµ0 ω eiωr/c
A(x, ω) = − p, (39)
4π r

ω 2 eiωr/c
H(x, ω) = n × p, (40)
4π r

µ0 ω 2 eiωr/c
E(x, ω) = n × (p × n). (41)
4π r
Now we can plug equation (40) and (41) into equation (38)
∞ ∞ 0
ω 02 e−i(ω+ω )r/c
2
Z Z  
dP 0 −i(ω+ω 0 )t µ0 ω 0
(t) = r2 dω dω e n · [n × (p(ω) × n)] × (n × p(ω ))
dΩ −∞ −∞ 4π 4cπ r2

Z ∞ Z ∞  
µ0 0
= dω dω 0 e−i(ω+ω )(t−r/c) ω 2 ω 02 p(ω) · p(ω 0 ) − n · p(ω)n · p(ω 0 )
16cπ 2 −∞ ∞
 7
Z ∞  Z ∞ 
dP µ0 0 0
(t) = dωe−iωt ω 2 (p(ω))j dω 0 e−iω t ω 02 (p(ω 0 ))k (δjk − nj nk ), (42)
dΩ 16cπ 2 −∞ −∞

where we used the definition of the retarded time t0 = t − r/c.

We can write the electric dipole moment p(x, ω) in time domain in its Fourier transform as
Z ∞
1 00
p(x, ω) = dt00 p(t00 )eiωt . (43)
2π −∞

Hence
∞ ∞
d2 τ d2
Z Z
1 1
ω 2 p(ω) = dτ p(τ )ω 2 eiωτ = − dτ 2
p(τ )eiωτ = − 2 p(τ )
2π −∞ 2π −∞ dτ dτ

Therefore, the integrand in the equation (42) gives

Z ∞
0 d2
dωe−iωt ω 2 p(ω) = − 2 p(τ ) (44)
∞ dτ τ =t0

Moreover, we now that the norm of the vector product

2
d2 d2 d2
(n × p(t0 )) × n =( p(t 0 2
)) − (n · p(t0 ))2 (45)
dt02 dt0 dt02

Inserting equation (44) and (45) in to (42) yields the required result,
2 2
dP µ0 d 0

Z0 d2 0


(t) = n × p(t ) × n = n × p(t ) × n , (46)
dΩ 16cπ 2 dt0 16c2 π 2 dt02

where µ0 = c210 = µ0 /0 1c = Zc0 , and Z0 = µ0 /0 .

p p

For magnetic dipole radiation, the electric dipole p is replaced by 1c n × m and equivalently (n × p̈) × n can be replaced
by (1/c)m̈ × n. This leads to the instantaneous power radiated per solid angle due to magnetic dipole moment
2
dP Z0 d2
(t) = n × m(t0 ) , (47)
dΩ 16c4 π 2 dt02
iω
(b) For electric quadrupole radiation, the electric dipole p replaced by 6c n · Q. Hence
∞
d3 iωτ 1 d3
Z
1 3 1
ω 2 p(ω) → ω n · Q(ω) = − dτ n · Q(τ ) 3
e = −r · Q
6c 12c −∞ dτ 6c dτ 3

Therefore,
∞
1 d3 1 d3
Z
i 0
dωe−iωt omega3 n · Q(ω)e−iωt = − Q(τ ) =− Q(x, t0 ). (48)
−∞ 6c 6c dτ 3 τ =t0 6c dτ 3

Hence the use of (48) in (42)gives

2
d3
 
dP Z0
(t) = n × 03 Q(x,t’) × n (49)
dΩ 576π 2 c6 dt

Problem 14.4 Using the Lienard-Wiechert fields, discuss the time-averaged power radiated per unit solid angle in
nonrelativistic motion of a particle with charge e, moving
(a) along the z axis with instantaneous position z(t) = a cos ω0 t,
(b) in a circle of radius R in the x-y plane with constant angular frequency ω0 . Sketch the angular distribution of the
radiation and determine the total power radiated in each case.
 8

Solution
(a)The instantaneous power radiated per unit solid angle for non-relativistic case (β << 1) is given by

dP e2 ~˙ 2
= |n × (n × β)| (50)
dω 4πc

2
r(t) = a cos(ω0 t)ẑ, ~˙ = −a ω0 cos(ω0 t)ẑ.
β (51)
c
Inserting (51) into (50)

dP e2 a2 ω04 e2 a2 ω04
= cos 2
(ω 0 t)|n × (n × ẑ)|2
= cos2 (ω0 t) sin2 θ (52)
dω 4πc3 4πc3
Not that the time average of cos2 (ω0 t) = sin2 (ωt) = 1/2 oever a period of ω0 /2π. Hence the time average paower of
dP/dΩ is

e2 a2 ω04 e2 a2 ω04 1
   
dP 2
= sin θ = (1 − cos(2θ)) (53)
dΩ 8πc3 8πc3 2

The total power radiated P can be obtained by radiating dipole can be obtained by integrating equation(52) over a
solid angl dΩ

e2 a2 ω04 π 3
Z 2π
e2 a2 ω04
Z
P = sin θdθ dφ = (54)
8πc3 0 0 3c3
(b) For a circle of constant radius R in x-y plane the radius vector is given by
1.0

0.5

0.4 0.2 0.2 0.4

0.5

1.0

 
3
FIG. 2: The average power radiated , dP
dΩ
,due to electric diploe moment in the units of (e2 a2 ω04 )/(8π c ) verses θ, θ from 0
to 2π.

2
r(t) = R(cos(ω0 t)x̂ + sin(ω0 t)ŷ), ~˙ = − ω0 R (cos(ω0 t)x̂ + sin(ω0 t)ŷ).
β (55)
c
Because this system has azimuthal symmetry (when averaged over a full period, we can rotate the coordinate system
such that the observer lies in the x-z plane. Thus, choosing n = cos θẑ + sin θx̂ allows us to evalute the vector products
in equation (50)as
2
~˙ = |n(n · β)
|n × (n × β)| ~˙ · n)| = ω0 R | − cos2 θ cos(ω0 t)x̂ − sin(ω0 t)ŷ + sin θ cos θ cos(ω0 t)ẑ|
~˙ − β(n (56)
c
 9

Hence
2
 
~˙ 2 = ω0 R
|n × (n × β)| 2 2 2
cos θ cos (ω0 t) + sin (ω0 t) (57)
c
The time average of both sin2 (ω0 t) = cos2 (ω0 t) = 1/2. Substituting equation (57) in to (50)we obtain,

e2 ω04 R2
 
dP
= cos2 θ cos2 (ω0 t) + sin2 (ω0 t)
dΩ 8πc3
The time average of both sin2 (ω0 t) = cos2 (ω0 t) = 1/2 which leads to the time averaged power as
e2 ω04 R2 1
 
dP
= ( (3 + cos 2θ)) (58)
dΩ 8πc3 2
The total power radiated due to electric quadrupole is

1.0

0.5

2 1 1 2

0.5

1.0

 
3
FIG. 3: The average power radiated, dP
dΩ
,due to electric quadrupole radiation in the units of (e2 a2 ω04 )/(8π c ) verses θ, θ from
0 to 2π.

π 2π
e2 ω04 R2 2e2 R2 ω04
Z Z
1
P = ( (3 + cos 2θ)) sin θdθ dφ = . (59)
8πc3 0 2 0 3c3

Problem 14.11
A particle of charge ze and mass m moves in external electric and magnetic fields E and B.
(a) Show that the classical relativistic result for the instantaneous energy radiated per unit time can be written
2 z 4 e4 6  ~ × B)2 − (β
~ · E)2

P = 2 3
γ (E + β
3m c
where E and B are evaluated at the position of the particle and γ is the particle’s instantaneous Lorentz factor.
(b) Show that the expression in part a can be put into the manifestly Lorentz- invariant form,
2z 4 r02 µν
P = F pν pλ Fλµ
3m2 c
where r0 = e2 /mc2 is the classical charged particle radius.
Solution
(a) Lets start by writing the radiated power due to realtivistic accelerated charge, that is
2 (ze)2 6 ~˙ 2 ~˙ 2 ].
~ × β)
P = γ [(β) − (β (60)
3 c
 10

~˙ 2 = β 2 (β)
~ × β)
Using the vector identity (β ~˙ 2 − (β ~˙ 2 , equation (60) becomes
~ · β)

2 (ze)2 6 ~˙ 2 ~˙ 2 − (β
~ · β)
2
~˙ 2 )] = 2 (ze) γ 6 [(β)
~˙ 2 (1 − β
~ 2 ) + (β
~ · β)
2
~˙ 2 ] = 2 (ze) γ 4 [(β)
~˙ 2 + γ 2 (β ~˙ 2 ]. (61)
~ · β)
P = γ [(β) − (β 2 (β)
3 c 3 c 3 c
~ can be determined from Newton’s second law as follow:
The relativistic acceleration dotβ

d~
p d d β~
F~ = = (γm~v ) = mc p ~˙ + β(
= γ 3 (1 − β 2 )β ~ β ~˙
~ · β)
dt dt d 1 − β2

~˙ + β(
~ β ~˙ =
~ · β) 1 ~
=⇒ (1 − β 2 )β F (62)
mcγ 3

~ to get
Now take the dot product of both sides of the last equation with β,

˙ =
~ ·~β 1 ~~
β βF
mcγ 3

Substituting this into (62)

~˙ = 1 ~ ~ ~ ~
β [F − β(β · F )] (63)
mcγ

Now insert equation (63) into (61).

2 (ze)2 γ 2 ~ ~ ~ ~ 2 2 2
~ · F~ )2 ] = 2 (ze) γ [F 2 − (β
~ · F~ )2 ]
P = 2 3
[(F − β(β · F )) + γ −2 (β (64)
3 m c 3 m c3
2

The Lorentz force for a particle of charge ze is given by

F~ = ze(E ~ × B)
~ +β ~ (65)

~ · F~ ) as follow
Using this we can evaluate F 2 and (β

F 2 = (ze)2 (E ~ × B)
~ +β ~ 2, ~ · F~ )2 = (ze)2 (β
(β ~ · E)
~ 2 (66)

Plugging equation (66) into (64) gives the desired result

2 z 4 e4 6  ~ × B)2 − (β
~ · E)2

P = γ (E + β (67)
3 m2 c3
(b) We can use employ the relativistic generalization of Larmor’s formula

2 (ze)2 dpµ dpµ

 
P =− (68)
3 m2 c3 dτ dτ

And
dpµ (ze) µν ze µν dpµ ze λ
= F Uν = F pν , =− p Fλµ (69)
dτ c mc dτ mc
where Uν = pnu /m. Substituting equation (69) into equation (68) gives

2 (ze)4 µν λ 2z 4 r02 µν
P = F p ν p F λµ = F pν pλ Fλµ (70)
3 m4 c5 3mc