Vibrations and Waves

Chapter 6 - Electromagnetic Waves - Maxwell’s equations

Andrew Lehmann
Last updated: October 9, 2023

1
Maxwell’s equations
Recap

After two semesters of electromagnetism we have Maxwell’s equations in integral

form
‹ ‹
~ · d ~S = Qenc (Gauss’ law)
E ~ · d ~S = 0 (No name)
B
ε0
˛ S
˛ S

~ ~ dΦB ~ · d~l = µ0 Ienc (Ampère’s law)

E · dl = − (Faraday’s law) B
L dt L

‚
The integrals S are integrals over a closed surface, and the d ~S infinitesimal
¸
vectors are normal to that surface, pointing towards the outside. The integrals L
are integrals over an oriented closed line, and the d~l infinitesimal vectors are along
that line, where the orientation creates a right hand rule convention that
determines the direction of the electromotive force in Faraday’s law or the sign of
the enclosed current in Ampère’s law.

2
First Goal

The first goal in this lesson will be to take the integral form of Maxwell’s equations
and derive their differential form:

~ = ρ
∇·E Gauss’ law
ε0
~
∇×E~ = − ∂B Faraday’s law
∂t
~
∇·B =0 No name
~
~ = µ0~J + µ0 ε0 ∂ E
∇×B Ampère’s law
∂t

3
Gauss’/Ostrogradsky’s theorem

For any vector field A ~ the integral of its contribution at any closed surface is equal
to the integral of its divergence in the volume enclosed by that same surface

~ d ~S
A S
‹ ˚
~ · d ~S =
A ~ dV
∇·A
S V(S)
V d ~S
d ~S ~
A

Note: the d ~S vectors must point towards the outside of the volume.
4
Gauss’ law

If we choose any volume in space, then the charge enclosed by that volume is an
integral of the charge density, ρ(x, y , z), at all points in the volume
˚
Qenc = ρ dV
V

Using Gauss’ theorem with the electric field we get

‹ ˚ ˚ ˚
~ ~ ~ ~ ρ
E · dS = ∇ · E dV =⇒ ∇ · E dV = dV
ε0
S V V V

Since this result is true for any arbitrary volume, the integrands must be indentical
at all points in space, and we get the differential form of Gauss’ law

~ = ρ
∇·E
ε0
5
Divergence of magnetic field

Starting from the integral form of the no-name law

‹
~ · d ~S = 0
B
S

we apply Gauss’ theorem and get

˚ ˚
~
∇ · B dV = 0 = 0 dV
V V

Since this result is true for any arbitrary volume, the integrands must be indentical
at all points in space, and we get the differential form of the law with no name

~ =0
∇·B

6
Stoke’s theorem

For any vector field A ~ the integral of its contribution along any closed line is equal
to the integral of its curl over the surface enclosed by that line

~
A d~
S

˛ ‹   L
~ · d~l =
A ~ · d ~S
∇×A
L d~l d~
S
S(L)

S
d~l ~
A

Note: the d~l vectors are related to the d ~S vectors. The orientation of the loop
determines, by a right hand rule, which direction from the surface the d ~S vectors
must point. 7
Faraday’s law

˛
For Faraday’s law ~ · d~l = − dΦB we must recall the definition of the magnetic
E
L dt
flux
‹
ΦB = ~ · d ~S
B
S

and so its time derivative is given by

‹ ‹ ~
dΦB d ~ · d ~S = ∂B
= B · d ~S
dt dt ∂t
S S

where the second equality is valid only if the imaginary surface S doesn’t change
with time.

8
Faraday’s law (cont.)

Using Stoke’s theorem on the left side of Faraday’s law we therefore have
˛ ‹  ‹  ‹ ~
!
~ · d~l =

~ · d ~S

~ · d ~S = ∂B
E ∇×E =⇒ ∇×E − · d ~S
L ∂t
S S S

Since this integral equality holds for any arbitrary surface S we must have that the
integrands are equal, giving the differential form of Faraday’s law

~
~ = − ∂B
∇×E
∂t

9
Ampère’s law
¸
For Ampère’s law L B ~ · d~l = µ0 Ienc we must recall how a current is calculated from
a current density. For the enclosed current we define an oriented loop L and add
up the current passing through that loop. This loop defines a surface at which the
current density, ~J, contributes a flow of current given by its component
perpendicular to the surface, the direction given by the normal vector d ~S

~J
d~
S
~J ~J ~J
‹
Ienc = ~J · d ~S
S(L) L S d~
S

~J ~J
~J

10
Ampère’s law

Using Stoke’s theorem on the left side of Ampère’s law we therefore have
˛ ‹   ‹   ‹  
~ ~
B · dl = ~
∇ × B · dS ~ =⇒ ~ ~
∇ × B · dS = µ0~J · d ~S
L S(L)
S S

Since this integral equality holds for any arbitrary surface S (or Ampèrian loop) we
must have that the integrands are equal, giving the differential form of Ampère’s
law before Maxwell

~ = µ0~J.
∇×B

This law has a problem. If we take its divergence we have

~ = µ0 ∇ · ~J
∇ · (∇ × B)

and mathematics tells us the divergence of any curl is always zero, whereas
physics tells us there is no problem with the divergence of a current density. The
right hand side may not be zero. 11
Conservation of charge

If current density, ~J, is diverging or con- ~J

~J
verging somewhere, then the charge ~J ~J
density, ρ, must change in time. This is gov-
erned by the continuity equation ~J

∂ρ The charge density at the circled

= −∇ · ~J
∂t location must increase.

12
Ampère’s law after Maxwell

The charge density is related to the electric field via Gauss’ law, and so the
divergence of the current density is
! !
∂   ~
∂E ~
∂E
∇ · ~J = − ~ =∇·
ε0 ∇ · E −ε0 =⇒ ∇ · ~J + ε0 =0
∂t ∂t ∂t

So Maxwell hypothesised that we should add another kind of current called the
∂E~
displacement current ~Jd = ε0 to the right side of Ampère’s law, to form a total
∂t
current density vector, to force the equation to be mathematically consistent. We
thus get the full differential form of Ampère’s law

~
~ = µ0~J + µ0 ε0 ∂ E
∇×B
∂t

13
Summary - Maxwell’s equations

We have completed our goal. With the help of the Gauss and Stokes theorems we
have taken the integral forms of Maxwell’s equations and derived their differential
forms:

~ = ρ
∇·E (Gauss’ law) ~ = 0 (No name)
∇·B
ε0
~ ~
~ = − ∂B
∇×E (Faraday’s law) ~ = µ0~J + µ0 ε0 ∂ E
∇×B (Ampère’s law)
∂t ∂t

14
Electromagnetic waves
Maxwell’s equations in a vacuum

In order to study electromagnetic waves, we consider a space with no charged

objects. That is, we take Maxwell’s equations in a vacuum, where there is no
charge or current density: ρ = 0 and ~J = 0. The equations reduce to

~ =0
∇·E ~ =0
∇·B
~ ~
~ = − ∂B
∇×E ~ = µ0 ε0 ∂ E
∇×B
∂t ∂t

15
Electromagnetic wave equation

Let’s take the curl of Faraday’s law

Ampère’s law
! z }| !{
~
∂B ∂  ~
~ =−∂ ∂E
  
~ =∇×
∇× ∇×E − =− ∇×B µ0 ε0 .
∂t ∂t ∂t ∂t
   
We need a vector identity ∇ × ∇ × A ~ =∇ ∇·A
~ − ∇2 A
~ for any vector field,
~ is the vector Laplacian:
where ∇2 A

~ = (∇2 Ax )~ex + (∇2 Ay )~ey + (∇2 Az )~ez

∇2 A for scalar Laplacian ∇2 Ak = ∇ · (∇Ax ).

Thus we find
!
~
~ =−∂ ∂E
 
∇ 
∇ ~ −∇2 E

·E µ0 ε0
| {z } ∂t ∂t
Gauss’ law 16
Electromagnetic wave equation (cont.)

We finally have

~
∂2E 1 ∂ 2 Ex 1
= ~
∇2 E → = ∇ · (∇Ex ) (for each component)
∂t 2 µ0 ε0 ∂t 2 µ0 ε0

where we recognise the wave equation (second time derivative proportional to

√
second spatial derivative) with phase speed c = 1/ µ0 ε0 . If we calculate this
speed using the measured values of the vacuum permeability and vacuum
permittivity:
µ0 = 1.25663706212 × 10−6 N A−2 and ε0 = 8.8541878128 × 10−12 F m−1
we find c ∼ 299 792 458 m/s. In 1865 Maxwell noticed this speed was so close to the
measured speed of light that he proposed that light (known to possess wave
properties like interference) was in fact an electromagnetic wave. This was
essentially accepted once Heinrich Hertz demonstrated the transmission of radio
wave signals in the 1880s.
This is considered as a great unification of two theories: the wave theory of light
and electromagnetism.
17
Electromagnetic wave equation (cont.)

If we take the curl of Ampère’s law

Faraday’s law
! z }| !{
~
∂E ∂  ~
~ = ∂ ∂B
  
~ =∇×
∇× ∇×B µ0 ε0 = µ0 ε0 ∇ × E −µ0 ε0 .
∂t ∂t ∂t ∂t

As before we use the curl of curl vector identity to find

!
  ∂ ∂ ~
B ~
∂2B 1
∇ ∇·B~ − ∇2 B~ =− µ0 ε0 =⇒ = ~
∇2 B
∂t ∂t ∂t 2 µ0 ε0

and so the magnetic field satisfies the same wave equation as the electric field.

18
Summary - Electromagnetic waves in a vacuum

We have shown that despite a vacuum containing no material, the equations of

electromagnetism

~ =0
∇·E ~ =0
∇·B
~
∂B ~
∂E
~ =−
∇×E ~ = µ0 ε0
∇×B
∂t ∂t
imply that electromagnetic waves can exist

~
∂2E 1 ~
∂2B 1
= ~
∇2 E and = ~
∇2 B
∂t 2 µ0 ε0 ∂t 2 µ0 ε0

1
and that they travel with phase speed c = √ .
µ0 ε0

19