Module I: Electromagnetic waves

Lecture 8: EM wave equation with sources

Amol Dighe
TIFR, Mumbai

Outline

~ with sources
Wave equation for and A

Solving the wave equation with sources

Coming up...

~ with sources
Wave equation for and A

Solving the wave equation with sources

~ and ?
Why the potentials A
~ and B
~ in the
We already have the wave equation satisfied by E
absence of any charge or current sources:
~
2 E

~
~
2E
E
 2 = 0
t
t

(1)

~
and similarly for B.
~
However when sources ( and ~J) are introduced, they affect E
~ in rather complicated ways. Therefore (with hindsight) we
and B
~ and , the vector and scalar
formulate our problem in terms of A
potentials, respectively.
When we come to relativity and covariance of equations, well
~ and even more.
appreciate the importance of A

A and : definitions
~ = 0 alows us to write B
~ as the curl of a
Maxwells equation B
~
vector, we define this vector as A:
~ =A
~
B

(2)

~ completely, since A
~ has not
Note that this does not define A
yet been defined, so the uniqueness theorem 1 is not satisfied.
~ = B/t
~
Maxwells equation E
then implies
~ =
E

~
A
t

(3)

This allows us to write

~
~ = A
E
t
where is a scalar. This is the definition of .
Note that is also not uniquely defined.

(4)

~ and
Gauge freedom for A
As obserbed earlier, A and are not uniquely defined.
Indeed, we can carry out simultaneous gauge trasformations
~0 = A
~ , 0 = +
A
t

(5)

~ 0 and 0
with any arbitrary scalar , and these new potentials A
~
~
will still give us the same E and B.
~ and B
~ are the physically measurable quntities, the
Since E
~
~ 0 , 0 ) are equivalent
potentials (A, ) and (A
This freedom of choosing any corresponds to the gauge
symmetry. We can choose to do the calculations in any
convenient gauge, the final measurable quantities will turn out to
be identical / gauge invariant.

~
Wave equation for A
~
We have already used two Maxwells equations while defining A
and : they will be satisfied automatically.
~ = ~J + ( E/t),
~
~ we get
Using B
where ~J = ~Jext + E,
!
!
2~
~

~ = ~Jext +
( A)
+  2
t
t
t
(6)
2~
~
~
Using ( A) = A + ( A), this leads to


~
~
A
2A
~
~ + () + 
2 A
 2 = ~Jext +( A)
t
t
t
(7)
If we now use our gauge freedom to make
~ + + (/t) = 0, (called as the Lorentz gauge), the
A
~
we get the wave equation for A:
~
2 A

~
~
A
2A
 2 = ~Jext
t
t

(8)

Wave equation for

We now see where the remaining Maxwells equation,
~ = / leads us to
E
!
~
A

~ 2 =

= ( A)
t

t


(9)

Now we use the same Lorentz condition as before to replace

~ by (/t), which leads to
A
2

 2 =
t
t


(10)

~
Note that the form of the equation for is the same as that for A,
with the charge a the source, instead of the current.

Coming up...

~ with sources
Wave equation for and A

Solving the wave equation with sources

The wave equations in vacuum

In vacuum, the wave equations take the form
2
~
2 A

1 2
c 2 t 2
~
1 2A
c 2 t 2

0

= 0~J

(11)
(12)

We drop the suffix on ~J for the sake of brevity.

We already know that for the static situation, i.e. when the
( 2 /t 2 ) terms are absent:
Z
(~x)
1
(~x0 ) 3 0
d x (13)
2 (~x) =
(~x) =
0
40
|~x ~x0 |
Z ~ 0
0
J(~x ) 3 0
2~ ~
~
~
~
~
A(x) = 0 J(x) A(x) =
d x
(14)
4
|~x ~x0 |
We expect (hope) that the solution to the time-dependent wave
equation may be similar.

Fourier analysis
Let us try solving a general equation
1 2
(~x, t) = g(~x, t)
(15)
c 2 t 2
by using the method of Fourier transform and Greens function.
Write the solution (~x, t) and the source g(~x, t) in terms of their
Fourier transforms and g :
Z
Z
it
(~x, t) =
(~x)e
d , g(~x, t) =
g (~x)eit d ,
2 (~x, t)

(16)
where the Fourier transforms are defined as
Z
Z
1
1
(~x, t))eit d , g (~x) =
g(~x, t))eit d ,
(~x) =
2
2
(17)
In terms of the Fourier transforms, the wave equation becomes
2
(~x) = g (~x)
(18)
c2
which well now try solving using the method of Greens function.
2 (~x) +

The method of Greens function

The method of Greens functions implies that:
If G(~x, ~x0 ) is a solution to the Greens equation
2
G(~x, ~x0 ) = (~x ~x0 )
c2
then the solution to (~x) is obtained as
Z
(~x) = g (~x0 )G(~x, ~x0 )d 3 x 0
2 G(~x, ~x0 ) +

(19)

(20)

This may be checked by explicit substitution.

The Greens equation is spherically symmetric, so we expect a
spherically symmetric solution, i.e. G(~x ~x0 ) is simply G(r ).
The Greens equation is then
1
[rG(r )] + k 2 G(r ) = (r )
r r

(21)

This has a solution (that may be checked by substitution):

G(r ) =

0
1 ikr
1
e
G(~x, ~x0 ) =
eik |~x~x |
4r
4|~x ~x0 |

(22)

Solution for (~x) and (~x, t)

The Greens function method has now given us the solution for
(~x):
Z
g (~x0 ) ik |~x~x0 | 3 0
1
(~x) =
e
d x
(23)
4
|~x ~x0 |
Inverse Fourier transform gives us the solution for (~x, t):
Z Z
1
g (~x0 ) i(tk |~x~x0 |) 3 0
(~x, t) =
e
d x
4 ~x0 |~x ~x0 |
In terms of t t k |~x ~x0 |/c, this becomes
Z
1
g (~x0 , t ) 3 0
(~x, t) =
d x
4 ~x0 |~x ~x0 |
This is the solution to our wave equation

(24)

(25)

Properties of the solution to wave equation

The solution formally looks the same as the solution for the static
case, except the time dependence, which appears through t+
and t : the advanced and retarded times respectively.
This implies that the potentials at any point depend on the
source distribution at some other times: in particular, at times
t = t |~x ~x0 |/c. This is akin to a signal taking time |~x ~x0 |/c
to travel from the source at ~x0 to affect the potential at x.
Thus, the disturbance caused by the sources travels with the
speed c. That is, the speed of light is c.
When we are dealing with the effect of time-varying sources on
the potentials, advanced solutions are not physical since they
would violate causality. They will need to be considered when,
later in the course, well be dealing with the back-reaction of the
changes in potential on the sources.

The retarded potentials

The retarded potentials, caused by time-varying sources, are:



Z ~x0 , t |~x~x0 |
Z
c
1
1
[(~x0 )] 3 0
3 0
~
(x, t) =
d
x
=
d x
40
40
|~x ~x0 |
|~x ~x0 |
(26)


Z ~J ~x0 , t |~x~x0 |
Z ~ 0
c
[J(~x )] 3 0
0
~ ~x, t) = 0
A(
d 3x 0 =
d x
4
4
|~x ~x0 |
|~x ~x0 |
(27)


~ ~0
where [f (~x)] is a convention used to write f ~x, t |xcx |

Recap of topics covered in this lecture

~ and , gauge freedom

Definitions of the potentials A
~ and in the presence
Lorentz gauge and wave equations for A
of sources (charges and currents)
Solution to the wave equation in vacuum, using Fourier
transforms and Greens function
~ and
Advanced and retarded solutions for the potentials A