Module III: Relativistic ED: applications

Lecture 3: EM potentials from a moving charge

(Lienard-Wiechert)

Amol Dighe
TIFR, Mumbai
Outline

Lienard-Wichert potentials: without relativity

Lienard Wichert potentials: using relativity

Outline

Lienard-Wichert potentials: without relativity

Lienard Wichert potentials: using relativity

Retarded potentials
I We have seen in the second module (Electromagnetic waves)
that a charge distribution ρ(~x) and a current distribution ~J(~x) give
rise to the retarded potentials
~ ~0
 
1
Z ρ ~x0 , t − |x−cx |
φ(~x, t) = d 3x 0 , (1)
4π0 |~x − ~x0 |
0
 
Z ~J ~x0 , t − |~x−~x |
~ ~x, t) = µ0 c
A( d 3 x 00 . (2)
4π |~x − ~x0 |
I For future convenience, we write them in a form that looks
symmetric in (~x, t) and (~x0 , t 0 ):
|~x − ~x0 | ρ(~x0 , t 0 )
Z  
1
φ(~x, t) = d 3 x 0 dt 0 δ t 0 − t + , (3)
4π0 c |~x − ~x0 |
|~x − ~x0 | ~J(~x0 , t 0 )
Z  
~ µ0 3 0 0 0
~
A(x, t) = d x dt δ t − t + , (4)
4π c |~x − ~x0 |
Note that we always use (~x, t) for the coordinates where we
measure the fields, whereas (~x0 , t 0 ) is the source position.
Potentials due to a moving charge
I Let a charge q be moving along a trajectory ~x00 (t 0 ) . Then

ρ(~x0 , t 0 ) = q δ 3 [~x0 − ~x00 (t 0 )] (5)

~J(~x0 , t 0 ) = 0
q ~v(t ) δ [x3 ~0 − ~x00 (t 0 )] . (6)

I Then for the scalar potential, after integrating over ~x0 using the
δ 3 function, we get

|~x − ~x00 (t 0 )|
Z  
q 0 0 1
φ(~x, t) = dt δ t − t + . (7)
4π0 c |x − x00 (t 0 )|
~ ~

I Similarly, for the vector potential,

|~x − ~x00 (t 0 )| ~v(t 0 )

Z  
~J(~x, t) = qµ0 0 0
dt δ t − t + . (8)
4π c |~x − ~x00 (t 0 )|
Integration over t 0
I Now we need to integrate over t 0 using the δ-function. Note that
Z
1
dy δ[f (y )] = (9)
|∂f (y )/∂y |y0

where y0 is the point such that f (y0 ) = 0.

r (t 0 )
I Let us denote ~r(t 0 ) ≡ ~x − ~x00 (t 0 ) . Then f (t 0 ) = t 0 − t + c ,
leading to
∂f ~v(t 0 ) · r̂(t 0 )
= 1 − (10)
∂t 0 c
where r̂(t 0 ) is a unit vector along the direction of ~r(t 0 ). Note
that the dependence of ~r and r on ~x and t is implicit.
I The quantity ∂f /∂t 0 needs to be calculated at t 0 that satisfies
f (t 0 ) = 0, i.e. at the time t 0 = tr , where tr satisfies the implicit
equation tr = t − r (tr )/c .
Problem
Obtain the above expression for ∂f /∂t 0
Lienard-Wichert potential
I Thus, we finally have

1 1 q
φ(~x, t) = h i (11)
4π0 r (tr ) 1 − (tr )·r̂(tr )
~
v
c

1 q
φ(~x, t) = h i . (12)
4π0 r − ~v·~r
c
tr

This is the Lienard-Wichert scalar potential.

I The Lienard-Wichert vector potential is, similarly,

~
~ ~x, t) = µ0 h q v(tr )i .
A( (13)
4π r − ~v·~r
c
tr

Problem
Show that, given ~x, t and ~x00 (t 0 ), the implicit equation for tr can
have at most one solution. What happens when there is no
solution ?
Lienard-Wichert potentials: simple form

Scalar potential
1 q
φ(~x, t) = (14)
4π0 s(tr )

Vector potential
~ ~x, t) = µ0 q~v(tr )
A( (15)
4π s(tr )

~v(tr )·~r(tr )
where s(tr ) ≡ r (tr ) − c

r (tr )
Note that tr is the solution of the implicit equation tr = t − c . We
shall denote quantities to be calculated at tr in magenta.
Comments on the Lienard-Wichert potential

I The scalar potential due to a moving charge is not simply equal

to 4π0qr (tr ) as one may naively guess. This is a purely geometric
effect, that arises from the Jacobian between dt 0 and dtr .
I Many different derivations of Lienard-Wichert potential may be
found in literature. The one in Griffiths gives a nice intuitive
understanding. Note that though that derivation describes the
situation in terms of a Jacobian between two volume integrals, it
is valid even for point charges.
I It is extremely crucial to keep track of times at which various
quantities are calculated: t, t 0 or tr . We shall try to be explicit
about it throughout the rest of the course. (We are already using
magenta for the quantities to be calculated at the retarded time
tr .)
I Note that ~r(~x, t, ~x0 (t 0 )), and hence r is also, in principle, a
function of ~x, t and ~x0 (t 0 ). The ~x and t dependence will be kept
implicit unless specifically needed.
Outline

Lienard-Wichert potentials: without relativity

Lienard Wichert potentials: using relativity

Lorentz invariants and Lorentz transformations

I Relativity provides us with a powerful tool: if we know the

4-potential in a particular frame, we can use Lorentz
transformations to convert it to any other frame.
I And of course we know the 4-potential for a point charge q in its
rest frame. Since φ = q/(4π0 |~r|) the 4-potential is simply
 
q
Ak = , ~0 (16)
4π0 |~r|c

I Applying Lorentz transformations directly to the above Ak is not

practical, since |~r| itself should change under these
transformations. To avoid this problem, we should write a
Lorentz-invariant quantity that equals |~r| in the rest frame of the
particle.
I We shall use brown to denote the 4-vectors and 4-tensors.
Lienard-Wichert 4-potential
I Such a quantity is r · u/c = rk uk /c, where rk ≡ (r , −~r) is the
position 4-vector from the source point (ctr , ~x0 ) to the observation
point (ct, ~x), and uk = (γc, γ~v) is the 4-velocity of the charge.
(0)
I In the rest frame of the particle, rk = (r , −~r) and uk (0) = (c, ~0),
(0)
so that rk uk (0) /c = r . Thus, r · u/c equals r in the rest frame of
the charge.
I The right form of the 4-potential in the rest frame of the charge is
then  
q
Ak (0) = , ~0 (17)
4π0 (r · u)

I In the frame where the charge is moving with a velocity ~v,

Lorentz transformations give

qγ~v quk
 
k qγ
A = , = (18)
4π0 (r · u) 4π0 c(r · u) 4π0 c(r · u)
This is the Lienard-Wichert 4-potential.
Components of the Lienard-Wichert 4-potential

I We have
~r · ~v
 
r · u = r γc − γ~v · ~r = γc r − =γcs. (19)
c

I The components of the 4-potential Ak are then

qγc q
φ = cA0 = = (20)
4π0 γcs 4π0 s
~µ qγ~v µ0 q~v
A = Aµ = 2
= (21)
4π0 γc s 4πs

the same as the potentials we had obtained without using

relativity explicitly.
I Note that the form of s appeared automatically from the Lorentz
invariant quantity.
Take-home message from this lecture

I The scalar and vector potential at a point due to a moving charge

is due to the location and velocity of the charge at the retarded
time tr , with an additional geometric factor of 1 − v (tr )·c r̂(tr ) .
I The geometric factor naturally emerges if we use covariant
notation.