Electrodynamics
J.Pearson
July 10, 2008
Abstract
These are a set of notes I have made, based on lectures given by R.Jones at the University of
Manchester JanJune ’08. Please email me with any comments/corrections: jap@watering.co.uk.
ii
CONTENTS iii
Contents
1 Linear Algebra & Introduction 1
1.1 Basis Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Rotations & Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2.1 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Brief Note on Important Vector Equations . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3.1 Important Electrodynamics Equations . . . . . . . . . . . . . . . . . . . . . . 4
2 Electromagnetic Field Equations 7
2.1 Maxwell’s Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.1 Electric Field Due to Point and Continuous Charges . . . . . . . . . . . . . . 7
2.1.2 Gauss’ Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1.3 Magnetic Vector Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.4 BiotSavart Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.5 Using the BiotSavart Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.6 Magnetic Force Between two Parallel Wires . . . . . . . . . . . . . . . . . . . 15
2.1.7 Amperes Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1.8 Faradays Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.1.9 Summary of Maxwell’s Equations in Vacuum . . . . . . . . . . . . . . . . . . 17
2.2 Maxwell’s Equations in Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.1 Dielectrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.2 Diamagnetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2.3 Summary of Maxwell’s Equations in Materials . . . . . . . . . . . . . . . . . 18
2.2.4 Fields Across Boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3 Potentials & Gauge Invariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 The Diracδ Function & Green Functions . . . . . . . . . . . . . . . . . . . . . . . . 22
2.4.1 Green Functions: Electrostatics . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.5 Poynting’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.6 Laplace Equation & Its Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.6.1 Solution to the Laplace Equation: Cartesian . . . . . . . . . . . . . . . . . . 29
iv CONTENTS
2.6.2 Solution to the Laplace Equation: Cylindrical Polars . . . . . . . . . . . . . . 32
2.6.3 Solution to the Laplace Equation: Spherical Polars . . . . . . . . . . . . . . . 33
2.6.4 Example: Dielectric Sphere in Uniform Eﬁeld . . . . . . . . . . . . . . . . . 35
2.6.5 Example: Charge Inside Spherical Cavity . . . . . . . . . . . . . . . . . . . . 37
2.7 Multipoles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.7.1 Electric Dipole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.7.2 General Multipole Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.7.3 Spherical Harmonic Expansion Multipoles . . . . . . . . . . . . . . . . . . . . 44
2.7.4 Relations Between Multipoles in Cartesian & Spherical Polars . . . . . . . . . 46
2.7.5 Properties of Multipoles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.8 Multipole Expansion of the Vector Potential . . . . . . . . . . . . . . . . . . . . . . . 48
2.9 Multipole Expansions Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.10 Method: Potentials and Surface Charges . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.11 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3 Retarded Potentials & Radiation 53
3.1 Introduction to Radiation from Accelerated Charges . . . . . . . . . . . . . . . . . . 53
3.1.1 Example: Larmor’s Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.1.2 Retarded Potentials & the Wave Equation . . . . . . . . . . . . . . . . . . . . 57
3.2 LienardWiechert Potentials: Point Charges . . . . . . . . . . . . . . . . . . . . . . . 58
3.2.1 Features of LienardWiechert Potentials . . . . . . . . . . . . . . . . . . . . . 61
3.2.2 Example: Particle Moving With Constant Velocity . . . . . . . . . . . . . . . 62
3.3 Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3.1 General Theory of Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3.2 Radiation: Acceleration & Velocity Parallel . . . . . . . . . . . . . . . . . . . 68
3.3.3 Radiation: Acceleration & Velocity Perpendicular . . . . . . . . . . . . . . . 69
3.3.4 Radiation: Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.3.5 Example: Minimum & Maximum Radiation . . . . . . . . . . . . . . . . . . . 71
3.3.6 Example: Charged Particle in Circular Orbit . . . . . . . . . . . . . . . . . . 73
3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
CONTENTS v
4 Relativistic Electrodynamics 75
4.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.2 Lorentz Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.2.1 Diﬀerentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.2.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.2.3 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.2.4 Inverse Lorentz Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.3 Lorentz 4Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.3.1 Proper Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.3.2 4velocity & 4momentum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.3.3 Electrodynamic 4vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.4 Electromagnetic Field Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.4.1 Maxwell’s Equations from F
µν
. . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.5 Lorentz Transformations of the Fields . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.6 LienardWiechert Fields from Lorentz Transformation . . . . . . . . . . . . . . . . . 99
4.7 Summary of 4Vectors & Transformations . . . . . . . . . . . . . . . . . . . . . . . . 102
4.8 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
A Vector Identities & Tricks 104
A.1 Vector Identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
A.1.1 Curl in Spherical & Cylindrical Polars . . . . . . . . . . . . . . . . . . . . . . 104
A.2 Useful Tricks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
B Worked Examples 107
B.1 Long Beam of Charge  Fields & Force . . . . . . . . . . . . . . . . . . . . . . . . . . 107
B.2 Spherical Shell Charge Distribution: σ(θ) = σ
0
cos θ . . . . . . . . . . . . . . . . . . 109
B.2.1 Spherical Shell Charge Distribution: σ(θ) = σ
0
cos 2θ . . . . . . . . . . . . . . 115
B.3 Show That T
µνλ
Is a Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
B.4 Lorentz Force in Covariant Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
C Legendre Equation & Spherical Harmonics 123
vi CONTENTS
C.0.1 Radial Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
C.1 Power Series Solution to the Ordinary Legendre Equation . . . . . . . . . . . . . . . 124
C.1.1 Orthogonality of Legendre Polynomials . . . . . . . . . . . . . . . . . . . . . 126
C.1.2 Using Legendre Polynomials as a Basis . . . . . . . . . . . . . . . . . . . . . . 127
C.2 Associated Legendre Polynomials & Spherical Harmonics . . . . . . . . . . . . . . . 127
C.3 The Addition Theorem for Spherical Harmonics . . . . . . . . . . . . . . . . . . . . . 128
C.4 Some Spherical Harmonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
C.5 Generating Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
C.5.1 Application: Expand
1
r
1
−r
2

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
D Books 135
1
1 Linear Algebra & Introduction
This section just lays some of the groundwork & very brief mathematical framework for some of
the things we will be using.
1.1 Basis Vectors
Suppose we deﬁne a set of vectors, such that:
e
i
e
j
= δ
ij
That is, they are of length one (normal), and are orthogonal (at ‘right angles’) to each other. Hence,
we have a set of orthonormal basis vectors. The cartesian basis vectors are:
e
1
= (1, 0, 0) e
2
= (0, 1, 0) e
3
= (0, 0, 1)
Using this basis, we are able to write vectors in terms of the basis. We use the equivalent notation:
x = x
1
e
1
+x
2
e
2
+x
3
e
3
= (x
1
, x
2
, x
3
) = x
i
e
i
Where the last line has used the Einstein summation convention of implied summations for repeated
indices. We will come back to this later, when discussing tensors in relativistic electrodynamics.
1.2 Rotations & Matrices
In the subsequent discussion, I will be being sloppy with notation, and write the vector x as x.
Let us write down the operation of some matrix M on a vector x, giving us some new vector x
:
_
_
M
11
M
12
M
13
M
21
M
22
M
23
M
31
M
32
M
33
_
_
_
_
x
1
x
2
x
3
_
_
=
_
_
x
1
x
2
x
3
_
_
M
ij
x
j
= x
i
The ﬁnal line M
ij
x
j
= x
i
is the index notation for matrix multiplication. The indices refer to
individual components, irrespective of the basis used.
Now, the transpose of the matrix M is:
M
T
ij
= M
ji
Notice then, that if:
x =
_
_
x
1
x
2
x
3
_
_
Then:
x
T
x = x
2
= (x
1
, x
2
, x
3
)
_
_
x
1
x
2
x
3
_
_
=
_
_
x
2
1
x
1
x
2
x
1
x
3
x
2
x
1
x
2
2
x
2
x
3
x
3
x
1
x
3
x
2
x
2
3
_
_
2 1 LINEAR ALGEBRA & INTRODUCTION
If a set of basis vectors e
i
are rotated to some new orientation e
i
, then the components of the
vectors in the primed system (new ones) are related to those in the unprimed system (old ones) by
x
= Rx, or:
x
i
= R
ij
x
j
Where we have that R
ij
= e
i
e
j
, the angle between bases in the primed and unprimed frame. The
cosines between the two frames. For a rotation about the zaxis, this looks like:
R =
_
_
cos θ sin θ 0
−sin θ cos θ 0
0 0 1
_
_
The inverse of R obviously has elements e
i
e
j
; which is just the transpose of R. Thus, R
−1
= R
T
.
This sort of matrix is said to be orthogonal. Hence, orthogonal transformations preserve lengths
and orientations of vectors within. This is equivalent to x y = x
y
. This will become very
important in our discussion on relativistic electrodynamics; in the invariance of quantities.
It is important that the distinction between the rotation matrix, R, and the operator M is under
stood.
Suppose we had that x is operated upon by some operator M, to give y:
y = Mx
Now, suppose we want to ﬁnd the equivalent in the primed frame. So, let us rotate y, via some
rotation matrix:
y
= Ry
Putting our expression for y in:
y
= R(Mx)
Now, a rotation followed by its inverse does nothing; its like saying, rotate by 90
◦
and back again.
Thus, R
T
R = 1 (essentially). Hence, if we put this factor in, nothing is changed:
y
= R(MR
T
Rx)
Let us collect the terms:
y
= RMR
T
(Rx)
Now, the eﬀect of x by rotation operator R is Rx = x
. Hence:
y
= RMR
T
x
Now, we also know that in the primed frame, transformations can still happen:
y
= M
x
So, upon comparison:
M
= RMR
T
Hence, we have an expression for the operator M in the primed frame. In index notation this is:
M
ij
= R
ik
M
kl
R
T
lj
= R
ik
R
jl
M
kl
Where in the ﬁrst equality we wrote down the operation, in a way conducive to matrix multiplication;
then used the transposed matrix element, and rearranged. The rearranged expression does not
represent matrix multiplication. Infact, it is a prototype for how a cartesian tensor, of second rank,
transforms.
1.3 Brief Note on Important Vector Equations 3
1.2.1 Tensors
This is an impressively brief introduction to these objects!
A cartesian tensor is a some T
ijkl
which transforms via:
T
abcd
= R
ai
R
bj
R
ck
R
dl
T
ijkl
This is for a rank4 tensor. Notice where the indices are in the rotation matrices R
ij
. The tensor
T
can be thought of as the ‘image’ of T, after the set of transformations given by the rotation
matrices R.
As another example, consider a rank1 tensor: a vector. It will transform via:
T
a
= R
ai
T
i
A matrix, which is a rank2 tensor will then transform via:
T
ab
= R
ai
R
bj
T
ij
Which is what we saw in the previous section: how an operator M transformed.
It is easy to see how this generalises to a rankn tensor.
We shall have a more complete discussion on tensors when we start to use them & their transfor
mation properties. In the present discussion, we have not made a distinction between tensors in
diﬀerent spaces (as we shall do later).
The salient points to take away from this is that a tensor (which is a set of quantities T) transforms
under a given set of rules (where the transformation has been denoted R here). That is, for a
general rankn tensor, it must transform like:
T
a
1
a
2
...an
= R
a
1
b
1
R
a
2
b
2
. . . R
anbn
T
b
1
b
2
...bn
This will become clearer when we use them.
1.3 Brief Note on Important Vector Equations
The ‘curl’ of a vector is the measure of how much it ‘rotates’. That is, something that has linear
ﬁeld lines, which just stretch oﬀ to inﬁnity has no curl.
The ‘div’ of a vector is a measure of the divergence of a ﬁeld from a source. The divergence of the
magnetic ﬁeld is zero  no magnetic monopoles.
The div grad = ∇ ∇φ = ∇
2
φ.
The divergence of a curl of a vector ﬁeld is zero. This can be visualised, as well as mathematically
proved, as the curl gives you a ﬁeld which is ‘going round in circles’, which will have no ‘radial’
component; hence its divergence is zero:
∇ (∇v) = 0
Stokes theorem: _
S
∇V dS =
_
V d
4 1 LINEAR ALGEBRA & INTRODUCTION
That is, the measure of how many ﬁeld lines pass through a closed path, enclosing a surface.
Divergence theorem:
_
V
∇ V dV =
_
S
V dS
That is, the measure of ﬂux of ﬁeld lines, through a surface enclosing some volume within which
the ﬁeld exists.
We will be making extensive use of the following:
ˆ r =
r
[r[
We shall try to be consistent with the notation that [r[ is the magnitude of the vector r, and that
ˆ r is a unit vector in the direction of r.
It is easy to derive various relations, but one that we use a lot is:
∇
1
r
= −
ˆ r
[r[
2
= −
r
[r[
3
Where r = xi +yj +zk, [r[ = r =
_
x
2
+y
2
+z
2
.
It is very useful to note that dotproducts commute. That is:
a b = b a
And also that crossproducts anticommute:
a b = −b a
1.3.1 Important Electrodynamics Equations
Most of these will be derived, or used at some point; but this is just a recap.
Maxwell’s equations:
∇ E =
ρ
ε
0
∇ B = 0 (1.1)
∇E = −
∂B
∂t
∇B = µ
0
J +µ
0
ε
0
∂E
∂t
(1.2)
In matter, these take on the following form:
∇ D = ρ
f
∇ B = 0 (1.3)
∇E = −
∂B
∂t
∇H = J
f
+
∂D
∂t
(1.4)
Where we have used the following deﬁnitions:
D = ε
0
E +P (1.5)
H =
1
µ
0
B −M (1.6)
1.3 Brief Note on Important Vector Equations 5
However, in linear media, these simplify to D = εE and H =
1
µ
B; Where ε ≡ ε
0
ε
r
, and similar for
the permeability.
The Lorentz force law:
F = q(E +v B) (1.7)
The Poynting vector:
P =
1
µ
0
(E B) (1.8)
Not to be confused with the polarisation vector above!
6 1 LINEAR ALGEBRA & INTRODUCTION
7
2 Electromagnetic Field Equations
2.1 Maxwell’s Equations
Let us start with Coulomb’s law: the force between two point charges.
Let us have two charges q
1
, q
2
, a distance d apart, then, the force between the two is given by:
[F
12
[ =
1
4πε
0
q
1
q
2
d
2
The factor of
1
4πε
0
is purely due to the SI units system used, and will give the ﬁeld in units Vm
−1
.
We can calculate the electric ﬁeld at the site of some test charge q
test
, due to some charge q. This
can be done for both a point charge q, and some continuous body of charge. We consider both
below.
2.1.1 Electric Field Due to Point and Continuous Charges
Figure 1: Setup for both point charge (a) and some continuous distribution of charge (b). Note, in
both cases, the observation point is at r, and the charge of interest is at r
.
Consider the setup in Fig (1)(a), for a point charge. The electric ﬁeld at some position r (that is,
at the position of the test charge q
test
), due to a point charge q, at r
, is given by:
E(r) =
F
q
test
=
1
4πε
0
q
[r −r
[
e
r−r
Where e
r−r
is the unit vector between r and r
. This is obviously just e
r−r
=
r−r
r−r

, their
separation, divided by the magnitude of their separation. Hence,
E(r) =
1
4πε
0
q(r −r
)
[r −r
[
3
8 2 ELECTROMAGNETIC FIELD EQUATIONS
Now, suppose we have a continuous distribution of charge, as in Fig (1)(b). Let us have a continuous
distribution ρ(r
). So, our ‘total charge’ q, which we used before, must now be substituted for an
integral over the volume of the charge distribution. Hence, the total charge is
_
ρ(r
) d
3
r
. Thus,
the electric ﬁeld at the position of the test charge (i.e. at the observation point P) is:
E(r) =
1
4πε
0
_
ρ(r
)
(r −r
)
[r −r
[
3
d
3
r
(2.1)
The integral will sweep over the distribution, picking up the little contributions from each ‘bit of
charge’ (which sounds like a contradiction for a continuous distribution, but its a useful way to
think about it).
Now, from vector calculus, we can derive the relation:
∇
1
[r −r
[
= −
r −r
[r −r
[
3
Hence, we can compare this with (2.1), and substitute in:
E(r) = −
1
4πε
0
∇
_
ρ(r
)
1
[r −r
[
d
3
r
= −∇φ(r)
Where we are now in a position to deﬁne the scalar static potential :
φ(r) =
1
4πε
0
_
ρ(r
)
1
[r −r
[
d
3
r
(2.2)
Thus:
E(r) = −∇φ(r) (2.3)
Hence, by Gauss’ law (which we will come to shortly), we have that ∇ E =
ρ
ε
0
. Thus, using the
above:
∇ E = −∇ ∇φ =
ρ
ε
0
Which is just:
∇
2
φ = −
ρ
ε
0
(2.4)
That is: Poissons equation. It is important to note (for subsequent discussions) that the expression
E = −∇φ only holds for electrostatic ﬁelds. That is, ones for which the charge is at rest. Thus, the
above poission equation also only holds for static ﬁelds. We shall come to timevarying ﬁelds later.
2.1.2 Gauss’ Law
Suppose we compute the integral over the (closed) surface containing some volume of charge density:
_
S
E dS =
1
ε
0
_
V
ρ(r) d
3
r (2.5)
2.1 Maxwell’s Equations 9
This is known as the integral version of Gauss’ law. By the divergence theorem, we are able to write
the LHS surface integral as a volume integral of the divergence of E. We can then make the two
(arbitrary) volumes on either side the same, leaving us with Maxwell’s 1
st
equation, by equating
the integrands:
∇ E =
ρ
ε
0
(2.6)
This holds true for time varying and electrostatic ﬁelds.
Example Suppose we have an inﬁnite rod, which carries a linecharge density λ (the rod is also
inﬁnitely thin). Let us compute the electric ﬁeld associated with such a system.
Let us choose our surface (i.e we are using the integral version of Gauss’ law ) to be a cylinder of
length , radius r enclosing the rod. Hence, the LHS of (2.5) is:
_
S
E dS = E
r
2πr
Where we have noted that all other possible components of the electric ﬁeld will be parallel to the
surface, and will hence not contribute to the surface integral. The RHS of (2.5) is just:
1
ε
0
_
V
ρ(r) d
3
r =
1
ε
0
eλ
Where we have used that the charge per unit length λ will carry a charge of e, hence the total
charge along the rod is its charge density multiplied by its length . Note, we have assigned a length
to an inﬁnite rod, but this isnt too much of a problem, as they cancel. Hence, equating the two
sides:
E
r
2πr =
1
ε
0
eλ
⇒E
r
=
eλ
2πε
0
r
We have hence computed the (radial) electric ﬁeld at a distance r away from an inﬁnite rod, with
charge per unit length λ.
Let us now consider a rod of radius a. We now have:
E
r
2πr =
1
ε
0
_
Aρ
V
d
Now, the terms in the integral on the RHS: ρ
V
is the volume charge density; which is just equal to
the line charge density divided by the area in which the charge resides: ρ
V
=
λe
πa
2
. A is the area
over which we are looking: that we have constructed our cylinder: A = πr
2
. Hence:
E
r
2πr =
1
ε
0
_
πr
2
λe
πa
2
d
Which is just:
E
r
2πr =
1
ε
0
r
2
a
2
λe
10 2 ELECTROMAGNETIC FIELD EQUATIONS
That is, the ﬁeld (which is radial) due to a rod of inﬁnite length, but ﬁnite radius a is given by:
E
r
(r) =
1
ε
0
rλe
2πa
2
r ≤ a
We shall now discuss the magnetic vector potential, and use it to derive (by analogy) the BiotSavart
law.
2.1.3 Magnetic Vector Potential
Now, from the Maxwell equation which states ∇ B = 0 (i.e. no magnetic monopoles), we are able
(via standard vector analysis) to deﬁne some vector A that B is the curl of, as the divergence of
the curl is zero. That is:
B = ∇A (2.7)
So that:
∇ B = ∇ ∇A = 0
Which, as was previously stated, is satisﬁed by standard vector analysis; and can be veriﬁed easily.
Now, another Maxwell equation (only for static ﬁelds) reads:
∇B = µ
0
J
Where J is the current density, which is the charge crossing per unit area per unit time. So,
inserting our expression for the magnetic vector potential:
∇∇A = µ
0
J
Which, again by a standard vector identity, can be written:
∇(∇ A) −∇
2
A = µ
0
J
Now, we use what we will call the Coulomb gauge, which is the statement that ∇ A = 0. Hence,
under the Coulomb gauge, the above becomes:
∇
2
A = −µ
0
J (2.8)
This is the vector form of Poissons equation: i.e. analogous to ∇
2
φ = −
ρ
ε
0
.
Magnetic ﬂux is just the surface integral of the magnetic ﬁeld:
Φ =
_
S
B dS
Which, by using the magnetic vector potential, is:
Φ =
_
S
∇A dS
Now, we can apply Stokes theorem to give us:
Φ =
_
A d (2.9)
So, we have the interpretation that magnetic ﬂux through an area is the same as the line integral
of the mangnetic ﬁeld vector over the line enclosing the surface.
2.1 Maxwell’s Equations 11
2.1.4 BiotSavart Law
We shall derive the BiotSavart law by analogy. Recall that we could ﬁnd the electric ﬁeld vector
from the scalar potential; so, let us suppose that we could ﬁnd the magnetic ﬁeld from the magnetic
vector potential. Also, by analogy, let us suppose that the vector potential may be found in the
same form as the scalar potential:
A =
µ
0
4π
_
J
dV
[r
12
[
Where r
12
≡ r −r
is the vector from the current loop carrying J to the observation point P. The
diﬀerent factor outfront is due to convention in wanting a speciﬁc set of units. Now, suppose that
Figure 2: How things are deﬁned in the BiotSavart law. The integral will sweep over the current
loop , picking up all its contributions.
we consider the loop of current to be a thin wire element, where the wire has crosssection S (which
doesn’t vary). So, we see that dV = Sd. Hence:
JdV = JSd = i
c
d
Where i
c
is the charge ﬂow per unit time. d is the lineelement. So now, our integral will sweep
over all the conributions from each line element to a ﬁeld at some observation point. So, we have
that:
A =
µ
0
i
c
4π
_
d
[r
12
[
Now, we know that we may ﬁnd B via B = ∇A. Thus:
B =
µ
0
i
c
4π
_
∇
d
[r
12
[
Again, we call upon a vector identity:
∇(av) = a∇v + (∇a) v
Where we have identiﬁed [r
12
[ as the scalar. Hence:
B =
µ
0
i
c
4π
__
1
[r
12
[
∇d +∇
_
1
[r
12
[
_
d
_
12 2 ELECTROMAGNETIC FIELD EQUATIONS
It is pretty trivial to show that the ﬁrst term is zero: if one does the cross product, one ﬁnds it to
be zero. We have also shown that:
∇
_
1
[r
12
[
_
= −
ˆ r
12
[r
12
[
2
Where ˆ r
12
is the unit vector from the current loop to the observation point. We then swop the
order of the crossproducts, noting that a b = −b a; resulting in:
B =
µ
0
i
c
4π
_
d ˆ r
12
[r
12
[
2
=
µ
0
i
c
4π
_
d r
12
[r
12
[
3
Hence, we have the BiotSavart law:
B =
µ
0
i
c
4π
_
d r
12
[r
12
[
3
(2.10)
Which ﬁnds the magnetic ﬁeld at a point, from a vector going from the current loop to the obser
vation point. The integral sweeps around the loop, picking up all contributions as it does. We will
sometimes use the BiotSavart law in the following form:
dB =
µ
0
i
c
4π
d r
12
[r
12
[
3
To be a little more consistent, supose that the current loop is at r
and the observer at r (relative
to some origin O), then the BiotSavart law reads:
dB(r) =
µ
0
i
c
4π
d (r −r
)
[r −r
[
3
2.1.5 Using the BiotSavart Law
Magnetic Field from a Current Loop Suppose we have a current I going round a loop, of
radius r. Suppose that the current is going clockwise, and that we have a coordinate system centred
on r = 0, with the ˆ r direction pointing radially out from the centre, and the ˆ z direction directly
out of the page.
Let us compute the magnetic ﬁeld at the origin. That is, compute B(0); when the current loop is
at r
= rˆ r.
2.1 Maxwell’s Equations 13
So:
dB(r) =
µ
0
i
c
4π
d (r −r
)
[r −r
[
3
⇒dB(0) =
µ
0
I
4π
d (−r
)
[ −r
[
3
⇒d (−r
) = [d[[ −r
[ ˆ z
= dr ˆ z
= dr ˆ z
⇒dB(0) =
µ
0
I
4π
dr
r
3
ˆ z
⇒B(0) =
µ
0
I
4π
_
d
r
2
ˆ z
=
µ
0
I
4π
2πr
r
2
ˆ z
⇒B =
µ
0
I
2r
Thus, we have found the magnetic ﬁeld due to a single loop of wire, carrying current I, at the origin.
If the current loop has N (tight) turns, this modiﬁes to:
B =
Nµ
0
I
2r
14 2 ELECTROMAGNETIC FIELD EQUATIONS
Magnetic Field due to a Long Straight Wire See Fig (3) to see how we have deﬁned things.
So, we have line element d = dx.
Figure 3: The setup for the magnetic ﬁeld due to a long thin wire. The wire goes oﬀ in either
direction to ±∞. The observer is at P, a vertical distance a from the wire. The wire carries a
current I. The vector r goes from the wire to P, along the ˆ r direction.
dB(r) =
µ
0
I
4π
d ˆ r
[r[
2
Let us look at the magnitude of the magnetic ﬁeld.
So, we ﬁrst see that:
[d ˆ r[ = [d[[ ˆ r[ sin θ
= dxsin θ
We also see that [r[
2
= r
2
. Now, by basic trigonometry:
r =
a
sin θ
Hence, let us put all this back into the BiotSavart expression:
dB =
µ
0
I
4π
[d ˆ r[
[r[
2
=
µ
0
I
4π
dxsin θ sin
2
θ
a
2
=
µ
0
I
4π
sin
3
θ
a
2
dx
Now, to make this easier to integrate, let us put x in terms of θ. We see that:
x = −
a
tan θ
⇒dx =
a
sin
2
θ
dθ
So, putting this all in:
dB =
µ
0
I
4πa
sin θ dθ
2.1 Maxwell’s Equations 15
We must integrate from θ = 0 →π. Hence:
B =
µ
0
I
4πa
_
π
0
sin θ dθ =
µ
0
I
2πa
Thus, we have found the magnetic ﬁeld due to a long thin, straight wire, at a distance a from it.
2.1.6 Magnetic Force Between two Parallel Wires
Now, we have just computed that a wire carrying a current generates a magnetic ﬁeld. We also
know that a charge will feel a force in the presence of a magnetic ﬁeld. Thus, a wire carrying a
current will feel a force in a magnetic ﬁeld.
Thus, two wires carrying a current will exert forces on each other.
Suppose we have two wires, 1 & 2, carrying currents of I
1
, I
2
a distance a from each other. Now,
the magnetic ﬁeld a distance a from wire 2 is given by:
B
2
=
µ
0
I
2
2πa
The force felt on a wire, in a magnetic ﬁeld, if the wire carries a charge q at velocity v is given by:
F = q(v B)
Or, in terms of magnitudes F = qvB. Now, qv = I. Hence, the force on wire 1, due to the
magnetic ﬁeld generated by wire 2 is:
F
1
=
µI
1
I
2
2πa
1
Where
i
is the length of wire i. Similarly, the force felt on wire 2, due to the ﬁeld generated by 1
is:
F
2
=
µI
1
I
2
2πa
2
Which is the same. Hence, the force per unit length is:
F
=
µI
1
I
2
2πa
So, if the magnitude of the force per unit length between two parallel wires carrying identical
currents, and are separated by 1m, is 2 10
−7
N/m, then the current in each wire is 1A.
This is the deﬁnition of the Ampere.
The deﬁnition of the Coulomb is similar: If a current of 1A is passing though a wire, then 1C of
charge passes a surface in 1s.
2.1.7 Amperes Law
Consider a closed loop enclosing some surface S; where there is a total current I through the
surface S. Then, Amperes law is:
_
B d = µ
0
_
S
J dS (2.11)
16 2 ELECTROMAGNETIC FIELD EQUATIONS
Using Stoke’s theorem, we get:
∇B = µ
0
J (2.12)
Note, this is only true for static ﬁelds. So, consider the following:
Charge conservation is the statement that the rate of ﬂow of current density plus the rate of change
of charge density is zero. That is, charge continuity:
∇ J +
∂ρ
∂t
= 0 (2.13)
Now, as it stands, Amperes law (2.12) does not satisfy the above. So, ﬁnding the divergence of
(2.12) gives:
∇ ∇B = µ
0
∇ J
The LHS is zero, by vector identities. Hence, we have that ∇ J = 0. Clearly, by (2.13), this
is only true if
∂ρ
∂t
= 0; which is not generally the case. That is, we have thus far only found a
staticequation. Let us consider the time varying equation, which we suppose to be:
∇B = µ
0
J +µ
0
ε
0
∂E
∂t
(2.14)
Taking the divergence of this:
∇ ∇B = µ
0
∇ J +µ
0
ε
0
∂∇ E
∂t
⇒0 = µ
0
∇ J +µ
0
ε
0
1
ε
0
∂ρ
∂t
= ∇ J +
∂ρ
∂t
Where we have used Gauss’ law. Hence, our modiﬁed timevarying version of Amperes law is
consistent with the continuity equation.
2.1.8 Faradays Law
This is the statement that the rate of change of a magnetic ﬁeld through some surface S generates
an electric ﬁeld (an EMF) in a loop enclosing the surface. That is:
−
∂
∂t
_
S
B dS =
_
E d (2.15)
Using Stokes’ theorem results in the diﬀerential form:
∇E = −
∂B
∂t
(2.16)
2.2 Maxwell’s Equations in Materials 17
2.1.9 Summary of Maxwell’s Equations in Vacuum
∇ E =
ρ
ε
0
(2.17)
∇ B = 0 (2.18)
∇E = −
∂B
∂t
(2.19)
∇B = µ
0
J +µ
0
ε
0
∂E
∂t
(2.20)
2.2 Maxwell’s Equations in Materials
Here we consider what happens to electric and magnetic ﬁelds if we put them in materials.
2.2.1 Dielectrics
This is for materials in electric ﬁelds.
We consider the eﬀect of charges induced in a material, when placed in an electric ﬁeld. Consider
that there is a free charge density ρ
free
and induced charge density ρ
ind
. Then, the total charge
density is just ρ = ρ
free
+ρ
ind
. Hence, Gauss’ law becomes:
∇ E =
1
ε
0
(ρ
free
+ρ
ind
)
Note, the induced charges only exist within the material!
If ρ
ind
= −∇ P, where P is the polarisation; so that the above easily becomes:
∇ (ε
0
E +P) = ρ
free
To tidy this up, let us deﬁne the electric displcement vector:
D = ε
0
E +P (2.21)
So, we have the form of Gauss’ law in materials:
∇ D = ρ
free
(2.22)
We also say that P = ε
0
χ
E
E; where χ
E
is the electrical susceptibility, which may be a tensor,
depending on the medium. In linear media only, we say that:
D = (1 +χ
E
)ε
0
E = ε
r
ε
0
E
Where ε
r
is the relative permitivity for linear media only.
18 2 ELECTROMAGNETIC FIELD EQUATIONS
2.2.2 Diamagnetics
This is for materials in a magentic ﬁeld, such as in a solenoid.
Let us initially consider the nontimevarying version of Amperes law; i.e. that
∂E
∂t
= 0. Also, as in
dielectrics, we have that the total current density is the sum of free and induced current densities.
Thus:
∇B = µ
0
(J
free
+J
ind
)
Also analogously, let J
ind
= ∇ M, where M is the magentisation. Hence, the above easily
becomes:
∇(B −µ
0
M) = µ
0
J
free
Again, let us clean up the above, by dividing through by µ
0
, and collecting terms into:
∇H = J
free
Where:
H =
1
µ
0
B −M
Which we call the magnetic intensity.
Again, we also say that M = χ
H
H, where χ
H
is the magnetic susceptibility. Therefore:
B = µ
0
(H +M) = µ
0
(1 +χ
H
)H = µ
0
µ
r
H
It is fairly easy to put the timevarying component back in, to see that:
∇H = J
free
+
∂D
∂t
(2.23)
2.2.3 Summary of Maxwell’s Equations in Materials
∇ D = ρ
free
(2.24)
∇ B = 0 (2.25)
∇E = −
∂B
∂t
(2.26)
∇H = J
free
+
∂D
∂t
(2.27)
2.2.4 Fields Across Boundaries
Consider the new Gauss’ law, which we can put into integral form easily, using the divergence
theorem: _
S
D dS =
_
V
ρ dV
Where we have dropped the subscript from the charge density.
Now, consider the surface S to be a box; so that it encloses a volume V , half of which is in medium
2.3 Potentials & Gauge Invariance 19
1, and the other part in medium 2. The normal unit vector ˆ n points into medium 1. So, the integral
on the LHS is just:
_
S
D dS = D
1
ˆ n −D
2
ˆ n
These then become only components perpendicular to the boundary (that is, parallel to the direction
of the normal), which we denote D
i⊥
. Thus, noting that the above is just equal to the surface charge
density:
D
1⊥
−D
2⊥
= ρ
s
That is, D is discontinuous at boundaries if ρ
s
,= 0; that is, a discontinuity due to free charges.
In a similar way, consider a small loop, crossing the boundary. The work done by the electric ﬁeld
in taking a small charge in the closed loop is zero. That is:
_
E d = 0
We ﬁnd that:
E
1//
−E
2//
= 0
That is, the electric ﬁeld is always continuous across boundaries.
2.3 Potentials & Gauge Invariance
Let us consider Faraday’s law:
∇E = −
∂
∂t
B
Now, we have already introduced the magnetic vector potential B = ∇ A, so that the above
becomes:
∇E = −
∂
∂t
(∇A)
Now, we know that curl grad is zero. That is, for some scalar χ, ∇∇χ = 0. So, by adding some
∇χ onto A, we would not ‘notice’ the change to B. So, let us write:
A −→A+∇χ
Now, taking a step backwards again, we have (from above) that Faradays law can be written:
∇
_
E +
∂A
∂t
_
= 0 (2.28)
So, let us change E so that the above is still valid. So, let us try:
E = −∇φ −
∂A
∂t
(2.29)
So, let us just check that if we choose (2.29), (2.28) is still valid:
∇
_
E +
∂A
∂t
_
= ∇
_
−∇φ −
∂A
∂t
+
∂A
∂t
_
= −∇∇φ
= 0
20 2 ELECTROMAGNETIC FIELD EQUATIONS
Thus, our choice in (2.29) has not altered (2.28). Let us see what (2.29) has done to Gauss law:
∇ E = ∇
_
−∇φ −
∂A
∂t
_
= −∇
2
φ −
∂∇ A
∂t
=
ρ
ε
0
⇒∇
2
φ +
∂∇ A
∂t
= −
ρ
ε
0
This is a coupled wave equation, in both ﬁelds φ and A, driven by charge density ρ.
Now, using Amperes law:
∇B = µ
0
J +
1
c
2
∂E
∂t
Where we have used that c
2
=
1
ε
0
µ
0
. Noting that B = ∇A
∇∇A = µ
0
J +
1
c
2
∂E
∂t
Using a vector identity:
∇∇A = ∇(∇ A) −∇
2
A
= µ
0
J +
1
c
2
∂E
∂t
Now, inserting our choice for E (2.29):
∇(∇ A) −∇
2
A = µ
0
J +
1
c
2
∂
∂t
_
−∇φ −
∂A
∂t
_
⇒∇
2
A−
1
c
2
∂
2
A
∂t
2
−∇
_
∇ A+
1
c
2
∂φ
∂t
_
= −µ
0
J
Now, a massive simpliﬁcation comes when we employ the Lorentz Gauge:
∇ A+
1
c
2
∂φ
∂t
= 0 (2.30)
Then, we are just left with:
⇒∇
2
A−
1
c
2
∂
2
A
∂t
2
= −µ
0
J
That is:
⇒
_
∇
2
−
1
c
2
∂
2
∂t
2
_
A = −µ
0
J (2.31)
Which is just a wave equation, which is generated by J. Notice that this is decoupled; in contrast
to the above driven by ρ.
2.3 Potentials & Gauge Invariance 21
Basically, as we have that B = ∇ A and E = −∇φ, we can change A and φ up to the point
that they dont change E, B. That is, we can make the transformations:
A
= A+∇χ (2.32)
φ
= φ −
∂χ
∂t
(2.33)
Where χ is some arbitrary scalar function. Let us now see that we can indeed do this. The criteria
we must satisfy is that by changing the potentials, we do not aﬀect the ﬁelds.
So, for the magentic ﬁeld:
B = ∇A
= ∇A+∇∇χ
= ∇A
Where the ﬁnal term is zero as the curl of the gradient of a ﬁeld is zero. Thus, we see that B is
invariant under the transformation.
Starting with Faradays law, i.e. ∇ E = −
∂
∂t
B, and inserting our transformed vector potential,
we end up with:
∇
_
E +
∂A
∂t
_
= 0
Which we have already shown. Now, we know that E = −∇φ−
∂A
∂t
. So, let us suppose that we can
transform our potentials somewhat:
E = −∇φ
−
∂A
∂t
= −∇φ
−
∂
∂t
(A+∇χ)
= −∇
_
φ
+
∂χ
∂t
_
−
∂A
∂t
Where, thus far, we have only used the relation between transformed & untransformed vector
potential. We shall now ﬁnd out how the scalar potential changes. Now, let:
φ = φ
+
∂χ
∂t
Hence, the transformed potential is:
φ
= φ −
∂χ
∂t
So everything is consistent, as we see that the above Eﬁeld is then unchanged.
We have introduced the Lorentz gauge, which is applicable to timevarying ﬁelds; whereas the
previous Coulomb gauge is only to be used on static ﬁelds.
Let us have a small mathematical diversion.
22 2 ELECTROMAGNETIC FIELD EQUATIONS
2.4 The Diracδ Function & Green Functions
These are very useful for dealing with point charges, or inﬁnitely thin sources.
The Diracδ function is deﬁned by:
_
∞
−∞
f(x)δ(x −x
) dx = f(x
) (2.34)
Notice, it can be used to pick up a single value of a function. One may think about this in the
following way: the integral sweeps over the xspace, but the δfunction will only return a nonzero
value when x = x
. Hence, it will return that value of any function within the integral.
And has the property that is has unit area:
_
∞
−∞
δ(x −x
) dx = 1
In 3D, we have:
δ
3
(r −r
) = δ(x −x
)δ(y −y
)δ(z −z
)
So, analogously:
_
f(r)δ
3
(r −r
) d
3
r = f(r
) (2.35)
Let us consider the delta function, having a function as an argument. Let us start with a simple
example; with just a constant inside the (delta) function. Now, let us state (then we shall prove it)
the following:
δ(a(x −x
)) =
1
[a[
δ(x −x
)
To show this, let us actually evaluate:
_
∞
−∞
f(x
)δ(a(x −x
))dx
Let us change variables:
y = ax y
= ax
So:
_
∞×a
−∞×a
f(
y
a
)δ(y −y
)dy
/a
Notice, if a < 0, then the integral is negative. To see this, replace a with −a everywhere above:
_
∞×(−a)
−∞×(−a)
f(
y
−a
)δ(y −y
)dy
/(−a) = −
1
a
_
−∞
∞
f(
y
−a
)δ(y −y
)dy
Now, notice, the eﬀect of having ‘minus’ an integral is to ﬂip the integration limits. So:
−
1
a
_
−∞
∞
f(
y
−a
)δ(y −y
)dy
=
1
a
_
∞
−∞
f(
y
−a
)δ(y −y
)dy
2.4 The Diracδ Function & Green Functions 23
So, let us take the modulus of a. Hence:
1
[a[
_
∞
−∞
f(
y
a
)δ(y −y
)dy
=
1
[a[
f(
y
a
) =
1
[a[
f(x)
Thus:
_
∞
−∞
f(x
)δ(a(x −x
))dx
=
1
[a[
f(x) (2.36)
Now, to consider and actual function:
δ(g(x))
So, we need to ﬁnd the zeros of g(x), as we know that the deltafunction is nonzero when its
argument is zero. Let them be at x
i
, so that we have that g(x
i
) = 0. Let us now do a Taylor
expansion about the zero:
g(x
i
+) = g(x
i
) + (x −x
i
)g
(x
i
) +. . .
Note, the ﬁrst term is zero, by deﬁnition. Note, here, a prime denotes derivative with respect to x,
wheras above, a prime is a way of distinguishing variables. Hence, near a zero:
g(x) ≈ (x −x
i
)g
(x
i
)
Near a zero. So, we have that:
δ(g(x)) = δ(g
(x
i
)(x −x
i
))
Near a zero. This is just like we had before, except we have more than one place the argument of
the deltafunction is zero. So, we must add the contributions up from all the zeros. Hence:
δ(g(x)) =
i
δ(x −x
i
)
[g
(x
i
)[
(2.37)
Where x
i
is a zero of the function g(x).
Now, suppose we had a set of charges, where charge q
i
resides at some position r
i
, then, we may
write the total charge density as:
ρ(r) =
i
q
i
δ(r −r
i
)
Which can be thought of as a set of ‘impluse charges’; which may however, be continuous. Or, as
another example, consider that some charges are distributed on an inﬁnitely thin shell, of radius a,
and that the distribution conforms to some σ(θ). Then, the charge distribution may be written:
ρ(r) = δ(r −a)σ(θ)
2.4.1 Green Functions: Electrostatics
If we wish to solve a diﬀerential equation of the form:
(L
x
u)(x) = f(x)
24 2 ELECTROMAGNETIC FIELD EQUATIONS
Plus boundary conditions; where L
x
is a linear hermitian operator. Then, to do so, we can always
write the solution to the DE, u(x) as:
u(x) =
_
G(x, x
)f(x
)dx
(2.38)
Where the Green function is deﬁned by:
L
x
G(x, x
) = δ(x −x
) (2.39)
We shall now supress the xsubscript on the operator: it is clear that it only operates on x, not x
.
To prove the above statement isn’t too hard:
(Lu)(x) = f(x)
⇒L
_
G(x, x
)f(x
)dx
=
_
LG(x, x
)f(x
)dx
=
_
δ(x −x
)f(x
)dx
= f(x)
Hence proven. In the ﬁrst line, we used the linearity of the operator, to be able to bring it inside the
integral (which is over x
anyway). Then we used the deﬁnitions of the Green function and delta
function.
Let us proceed by an example from our present electrostatic discussions.
Now, from Gauss’ law ∇ E = ρ/ε
0
and E = −∇Φ, we can easily derive Poissons equation:
∇
2
Φ(x) = −
ρ(x)
ε
0
Note, this is strictly for a static charge distribution, else we would have the A term as well. We
have the boundary condition that Φ(x) →0 as [x[ →∞.
Now, for a point charge q, at x
, we have a charge distribution which is just a delta function:
ρ(x) = qδ(x −x
)
Note, strictly, we should have written δ
3
, but that is understood, as its argument has 3 variables.
Now, we know that the resulting Poisson equation has the following solution:
∇
2
Φ(x) = −
q
ε
0
δ(x −x
) ⇒ Φ(x) =
q
4πε
0
1
[x −x
[
Now, upon comparison of the above formalism for Green function, Lu = f is just the Poisson
equation. That is, the operatior L is just the ∇
2
operator. Hence, from LG = δ(x − x
), we see
that the Green function we want, is that satisfying:
∇
2
G(x, x
) = δ(x −x
)
We have the form of G, from the above expression for the point charge. That is:
G(x, x
) = −
1
4π
1
[x −x
[
2.5 Poynting’s Theorem 25
Hence, we have that the solution to Poissons equation is given by:
Φ(x) =
_
G(x, x
)
_
−
ρ(x
)
ε
0
_
d
3
x
That is:
Φ(x) =
1
4πε
0
_
ρ(x
)
[x −x
[
d
3
x
Which is something we already knew, but we have derived it using Green functions; and in the
process, have identiﬁed a Green function.
Notice, we have also arrived at a useful relation:
∇
2
1
[x −x
[
= −4πδ(x −x
) (2.40)
Green function theory is extensive, and is used in solving diﬀerential equations (as was hinted at
previously); but we shall not go into that here.
2.5 Poynting’s Theorem
Let us start by stating the Lorentz force law:
F = q(E +v B)
So, the work done dW on a charge dq, when displaced a distance d is given by:
dW = F d
That is:
dW = dq(E +v B) d
Now, from v =
d
t
, a trivial relation, we can write the above as:
dW = dq(E +v B) vdt
That is:
dW
dt
=
_
dq(E +v B) v
However, note that:
dq = ρ
V
d
3
r
Hence:
dW
dt
=
_
V
ρ
V
(E +v B) v d
3
r
=
_
V
ρ
V
¦E v + (v B) v¦ d
3
r
=
_
V
E (ρ
V
v)d
3
r
26 2 ELECTROMAGNETIC FIELD EQUATIONS
Where we have used (the easily veriﬁable):
v (v B) = 0
We also note that J = ρ
V
v. Hence:
dW
dt
=
_
V
E J d
3
r (2.41)
That is, the total rate of doing work, by the ﬁelds, if there is a continuous distribution of charge
and current.
Now, consider Amperes’ law:
∇H = J +ε
0
∂E
∂t
That is:
J = ∇H −ε
0
∂E
∂t
So, putting this into (2.41):
dW
dt
=
_
V
E
_
∇H −ε
0
∂E
∂t
_
d
3
r
=
_
V
E (∇H) −ε
0
E
∂E
∂t
d
3
r
Now, we use a vector identity for the ﬁrst expression in the integral:
∇ (E H) = H (∇E) −E (∇H)
⇒E (∇H) = H (∇E) −∇ (E H)
So, after using this:
dW
dt
=
_
V
H (∇E) −∇ (E H) −ε
0
E
∂E
∂t
d
3
r
Now, we also know that from Faraday’s law:
∇E = −
∂B
∂t
= −µ
0
∂H
∂t
Thus, using this:
dW
dt
=
_
V
−µ
0
H
∂H
∂t
−∇ (E H) −ε
0
E
∂E
∂t
d
3
r
= −
_
V
∇ (E H) d
3
r −µ
0
_
V
H
∂H
∂t
d
3
r −ε
0
_
V
E
∂E
∂t
d
3
r
Now, we can use the divergence theorem on the far LHS integral. That is:
_
V
∇ (E H) d
3
r =
_
S
(E H) dS
2.5 Poynting’s Theorem 27
Putting this back in:
dW
dt
= −
_
S
(E H) dS −µ
0
_
V
H
∂H
∂t
d
3
r −ε
0
_
V
E
∂E
∂t
d
3
r
= −
_
S
(E H) dS −
∂
∂t
_
V
_
ε
0
[E[
2
+µ
0
[H[
2
_
d
3
r
Now, we start to recognise the Poynting vector, and energy densities of the electric and magnetic
ﬁelds:
P = E H (2.42)
U
E
=
_
V
ε
0
[E[
2
d
3
r (2.43)
U
M
=
_
V
µ
0
[H[
2
d
3
r (2.44)
Thus, we have:
dW
dt
= −
∂
∂t
(U
E
+U
M
) −
_
S
P dS (2.45)
So, we see that if the surface integral is zero; the rate at which particles gain energy is equal to the
rate at which the ﬁelds lose energy.
Also, if the ﬁelds are constant (i.e. their time derivatives are zero), and if work is still being done
on the particles, then this inﬂow of energy is provided by the Poynting vector. Thus, we see that
the Poynting vector represents the rate at which EM ﬁelds transport energy across a unit surface
(hence the integral).
As an example, let us consider the energy ﬂux for a plane harmonic wave. That is, ley uscompute
the Poynting vector.
So, a plane wave is given by:
E = E
0
e
i(k·r−ωt)
B = B
0
e
i(k·r−ωt)
And, we have that P = E H. We note that:
∇E = −
∂B
∂t
H =
1
µ
0
B
So:
∂B
∂t
= −iωB
And, by doing the crossproduct, and realising that the argument of the exponential is actually a
scalar, which is i(k
x
x + k
y
y + k
z
z − ωt); and that E
0
= (E
x
, E
y
, E
z
), where each component has
the same exponential factor, we can easily derive:
∇E = ik E
Thus, we have:
∇E = −
∂B
∂t
⇒ ik E = iωB
28 2 ELECTROMAGNETIC FIELD EQUATIONS
Hence:
k E = ωµ
0
H
Now, we also use the fact that:
ˆ
k =
k
[k[
Where
ˆ
k is a unit vector, in the direction of k. We shall write [k[ = k. We also have the standard
relation: ω = kc. So:
k E = ωµ
0
H
⇒k
ˆ
k E = ωµ
0
H
⇒H =
1
kωµ
0
ˆ
k E
=
1
µ
0
c
ˆ
k E
Hence, the Poynting vector is:
P = E H
=
1
µ
0
c
E
ˆ
k E
=
1
µ
0
c
[E[
2
This is, however, an instantaneous value. The time averaged value is just:
¸[E[
2
) =
1
2
[E[
2
2.6 Laplace Equation & Its Solutions 29
2.6 Laplace Equation & Its Solutions
Here, we shall denote the scalar potential as V , rather than φ, to avoid any possible confusion with
the coordinate.
Now, from Gauss’ law, using E = −∇V , we are able to easily derive the electrostatic Poisson
equation:
∇
2
V = −
ρ
ε
0
Where V is the scalar potential. We have seen that this has solution:
V (r) =
1
4πε
0
_
V
ρ(r
)
[r −r
[
d
3
r
If there is no charge density, we have the Laplace equation:
∇
2
V = 0 (2.46)
In solving the equation, we use the separation of variables technique; and its solution is unique  up
to additive constants.
We can use various boundary conditions:
• Dirichlet: V (r) is known on some surface;
• Neumann:
∂V
∂n
≡ ˆ n ∇V is known on S;
• Or, a mix of the above two.
We can solve using the inherent symmetry of the system:
• Rectangular  use Cartesian coordinates;
• Cylindrical  use cylindrical polars;
• Spherical  use spherical polars.
We shall now consider various types of solutions.
2.6.1 Solution to the Laplace Equation: Cartesian
This method will apply to all cartesian systems, with diﬀerent boundary conditions giving diﬀerent
ﬁnal results. The general method is moreorless unchanged.
Consider 2 plates which are inﬁnite in z. Let one edge of the plates be held at some potential V
0
.
So, consider the Dirichlet boundary conditions:
• V (x, 0) = V (x, a) = 0;
• V (0, y) = V
0
;
30 2 ELECTROMAGNETIC FIELD EQUATIONS
• lim
x→∞
V (x, y) = 0.
So, we have that the Laplace equation reduces to:
∂
2
V
∂x
2
+
∂
2
V
∂y
2
= 0
Where the potential V (x, y). So, using separation of variables, we have that:
V (x, y) = X(x)Y (y)
Putting this into the Laplace equation:
Y
d
2
X
dx
2
+X
d
2
Y
dy
2
= 0
Dividing through by V = XY :
1
X
d
2
X
dx
2
+
1
Y
d
2
Y
dy
2
= 0
So, we have that each of the above expressions must be a constant:
k
2
+ (−k
2
) = 0
That is:
1
X
d
2
X
dx
2
= k
2
1
Y
d
2
Y
dy
2
= −k
2
Each of these is easily solved:
X(x) = Ae
kx
+Be
−kx
(2.47)
Y (y) = C sin ky +Dcos ky (2.48)
So, we have that our general solution to the Laplace equation, in 2D (or 3D, with symmetry) is:
V (x, y) = ¦Ae
kx
+Be
−kx
¦¦C sin ky +Dcos ky¦ (2.49)
Where the constants must be determined by the boundary conditions, and will generally be linear
superpositions. So, let us continue, solving for our speciﬁc set of boundary conditions:
Let us use the boundary condition that V must decay as x →∞. That is, A = 0. Hence, we have:
V (x, y) = e
−kx
(C sin ky +Dcos ky)
Let us use another boundary condition: V (x, 0) = 0. Hence, we see that D = 0. Hence, we now
have:
V (x, y) = Ce
−kx
sin ky
2.6 Laplace Equation & Its Solutions 31
Let us use another boundary condition: V (x, a) = 0. So:
sin ka = 0
Thus:
k =
nπ
a
n = 1, 2, 3, . . .
Therefore:
V (x, y) = Ce
−
nπx
a
sin
_
nπy
a
_
Using the linear superposition of solutions, we have that:
V (x, y) =
∞
n=1
C
n
e
−
nπx
a
sin
_
nπy
a
_
(2.50)
Let us ﬁnd the constant C
n
. Let us use the ﬁnal boundary condition: V (0, y) = V
0
. Hence:
∞
n=1
C
n
sin
_
nπy
a
_
= V
0
(2.51)
To go further, we use the orthogonality of sine functions. That is, we use:
_
a
0
sin
_
nπy
a
_
sin
_
mπy
a
_
dy =
a
2
δ
nm
So, let us multiply both sides of (2.51) with “another”, and integrate. That is:
n
C
n
_
a
0
sin
_
nπy
a
_
sin
_
mπy
a
_
dy = V
0
_
a
0
sin
_
mπy
a
_
dy
Thus:
n
a
2
C
n
δ
nm
=
V
0
a
mπ
(1 −cos mπ)
That is:
a
2
C
m
=
V
0
a
mπ
(1 −(−1)
m
)
So, we see that if m is even, then C
m
is zero. And, if m is odd, then:
C
m
=
4V
0
mπ
∀m odd
Therefore, we have:
V (x, y) =
∞
m=1
4V
0
mπ
e
−
mπx
a
sin
_
mπy
a
_
m odd (2.52)
Thus, we have found the solution of the Laplace equation, in Cartesian coordinates, under our given
boundary conditions.
32 2 ELECTROMAGNETIC FIELD EQUATIONS
2.6.2 Solution to the Laplace Equation: Cylindrical Polars
Here, we use the coordinate system V (r, ϕ, z); so that the separation of variable happens thus:
V (r, ϕ, z) = R(r)Φ(ϕ)Z(z)
So, we have that the Laplacian, in cylindrical polars, takes on the form:
∇
2
V =
1
r
∂
∂r
r
∂V
∂r
+
1
r
2
∂
2
V
∂ϕ
2
+
∂
2
V
∂z
2
Thus, using our separated variables:
1
R
d
2
R
dr
2
+
1
Rr
dR
dr
+
1
r
2
Φ
d
2
Φ
dϕ
2
+
1
Z
d
2
Z
dz
2
= 0
Now, we notice a similar linear independence of the terms as for the Cartesian case. Thus, we let:
1
Z
d
2
Z
dz
2
= b
2
Hence, we have that the Laplace equation becomes:
r
2
R
d
2
R
dr
2
+
r
R
dR
dr
+
1
Φ
d
2
Φ
dϕ
2
+b
2
r
2
= 0
Now, we also let:
1
Φ
d
2
Φ
dϕ
2
= −α
2
Hence, the Laplace equation further reduces to:
r
2
R
d
2
R
dr
2
+
r
R
dR
dr
−α
2
+b
2
r
2
= 0
Cleaning up:
d
2
R
dr
2
+
1
r
dR
dr
+
_
b
2
−
α
2
r
2
_
R = 0
So, to summarise, we have that the solution to the Laplace equation, in 3D cylindrical polar coor
dinates, is the product of the solutions to the following equations:
d
2
Φ
dϕ
2
= −α
2
Φ (2.53)
d
2
Z
dz
2
= b
2
Z (2.54)
d
2
R
dr
2
+
1
r
dR
dr
+
_
b
2
−
α
2
r
2
_
R = 0 (2.55)
We have that the solutions to (2.53) and (2.54) are just:
Φ(ϕ) = C
α
cos αϕ +D
α
sin αϕ (2.56)
Z(z) = E
b
cosh bz +F
b
sinh bz (2.57)
2.6 Laplace Equation & Its Solutions 33
And that we note (2.55) is Bessel’s equation, which has solutions:
R(r) = A
ab
J
a
(br) +B
ab
N
a
(br) (2.58)
Where J
a
(br), N
a
(br) are Bessel functions of the ﬁrst and second kinds, respectively; and are gen
erally looked up.
So, we have that the general solution to the Laplace equation, in 3D cylindrical polars, is given by:
V (r, ϕ, z) =
a,b
¦A
ab
J
a
(br) +B
ab
N
a
(br)¦¦C
α
cos αϕ +D
α
sin αϕ¦¦E
b
cosh bz +F
b
sinh bz¦ (2.59)
Where the constants must be determined by initial conditions.
2.6.3 Solution to the Laplace Equation: Spherical Polars
We use the coordinate system V (r, θ, ϕ), where the Laplace equation takes on the form:
1
r
2
∂
∂r
r
2
∂V
∂r
+
1
r
2
_
1
sin θ
∂
∂θ
sin θ
∂V
∂θ
+
1
sin
2
θ
∂
2
V
∂
2
ϕ
2
_
= 0
Now, we separate V thus:
V (r, θ, ϕ) = R(r)Y
m
(θ, ϕ)
Where Y
m
are spherical harmonics.
We note that the above Laplacian has the
ˆ
L
2
operator. That is, the Laplace equation is:
1
r
2
∂
∂r
r
2
∂V
∂r
−
1
r
2
ˆ
L
2
= 0
Hence, we have that:
1
r
2
Y
m
d
dr
r
2
dR
dr
−
1
r
2
R
ˆ
L
2
Y
m
= 0 (2.60)
We know that Y
m
are eigenfunctions of
ˆ
L
2
thus:
ˆ
L
2
Y
m
(θ, ϕ) = ( + 1)Y
m
(θ, ϕ)
Hence, we have that the Laplace equation (2.60) reduces to:
Y
m
r
2
d
dr
r
2
dR
dr
−
R
r
2
( + 1)Y
m
= 0
Now, we try an ansatz: a power law, for R(r). So, let us try R(r) = r
α
. Hence, putting this into
the above yields:
Y
m
r
2
α(α + 1)r
α
=
r
α
r
2
( + 1)Y
m
Now, this is just:
α
2
+α − −
2
= 0
34 2 ELECTROMAGNETIC FIELD EQUATIONS
Which we can equivalently write as:
(α −)[α + ( + 1)] = 0
Hence, we see that the two solutions to the above equation, are just:
α = α = −( + 1)
And can be veriﬁed by substituting back in. So, we have that the general solution to the Laplace
equation, in spherical polars is:
V (r, θ, ϕ) =
,m
_
A
m
r
+
B
m
r
+1
_
Y
m
(θ, ϕ) (2.61)
Where we have used both solutions to the ansatz; each having diﬀerent coeﬃcients A
m
, B
m
; to be
determined by initial conditions.
Now, as Y
m
(θ, ϕ) = Θ
(θ)Φ
m
(ϕ) where Φ
m
(ϕ) = e
imϕ
, if the system has axial symmetry, then the
solution reduces to:
V (r, θ, ϕ) =
_
A
r
+
B
r
+1
_
P
(cos θ)
Where the P
(cos θ) are Legendre polynomials, and can be generated by Rodrigues formula:
P
(x) =
1
2
!
_
d
dx
_
(x
2
−1)
And the ﬁrst few are given by:
P
0
(x) = 1
P
1
(x) = x
P
2
(x) =
1
2
(3x
2
−1)
P
3
(x) =
1
2
(5x
3
−3x)
P
4
(x) =
1
8
(35x
4
−30x
2
+ 3)
If the system is symmetrical in both θ, ϕ; such as a point charge surrounded by dielectric; then the
Laplace equation looks like:
1
r
2
∂
∂r
r
2
∂V
∂r
= 0
And the potential is just V (r) = R(r). That is, we must solve:
1
r
2
d
dr
_
r
2
dR
dr
_
= 0
To see why there is no angular components, or diﬀerentials; consider that they are in there. That
is, the Laplace equation has the form we used previously. The diﬀerentials of the angular parts are
zero (they are constant: the symmetry of the system), and the radial diﬀerential will be multiplied
by the angular functions; however, these are then just divided out.
To solve this, note that we must have:
r
2
dR
dr
= D
2.6 Laplace Equation & Its Solutions 35
Where D is some constant. Then the Laplace equation is satisifed. We can ﬁnd R by the standard
method:
r
2
dR
dr
= D
⇒dR = D
dr
r
2
⇒R(r) =
_
D
dr
r
2
= −
D
r
+A
=
B
r
+A
Where A, B are both constants. Hence, we have that the potential in an angularsymmetric system
is given by:
V (r) = A+
B
r
For a more extensive treatise of spherical harmonics, see the relevant appendix.
2.6.4 Example: Dielectric Sphere in Uniform Eﬁeld
Consider a dielectric sphere, radius a, relative permittivity ε
r
, in a uniform electric ﬁeld, otherwise
vaccum. So, we immediately see that there is axialsymmetry; thus, we solve the Laplace equation
in spherical coordinates, with axial symmetry V (r, θ):
V (r, θ) =
∞
=0
_
A
r
+
B
r
+1
_
P
(cos θ)
The boundary conditions are implied, not necessarily given:
• The potential must go to zero at inﬁnity;
• The potential must remain ﬁnite (i.e. not diverge) as r →0;
• The ﬁeld at inﬁnity is just the ‘unperturbed’ ﬁeld, E
0
;
• The potential must be continuous at the boundary;
• The perpendicular D
⊥
must be continuous at boundaries; as no free charges.
So, as E = −∇V , we have that V = −E
0
z; just the integral. To see the motivation behind this,
consider how we choose the alignment of the coordinate system. The alignment is obviously arbi
trary, so we choose it to be aligned with the uniform applied ﬁeld. Remember, the transformation
between Cartesian and spherical polars, for z is z = r cos θ; hence the potential as r → ∞ is just
V = −E
0
r cos θ.
Now, note:
P
1
(x) = x ⇒ P
1
(cos θ) = cos θ
36 2 ELECTROMAGNETIC FIELD EQUATIONS
So, we have = 1 in the summation; by orthogonality considerations of the Legendre polynomials.
That is, consider that the potential (given by a sum over an inﬁnite number of multipoles) must go
to the dipole term (i.e. = 1) at inﬁnity. Thus, multiplying both sides by ‘another’ polynomial will
result in just the dipole term being ﬁltered out, on the potential side. This may be seen in more
detail in examples in the appendix. So, taking = 1, the ‘prototype’ for the potential is:
V (r, θ) =
_
Ar +
B
r
2
_
cos θ
Let us now continue by looking at the two regions (inside and outside) separately, with diﬀerent
constants in both cases. That is:
V
1
=
_
A
1
r +
B
1
r
2
_
cos θ r ≤ a
V
2
=
_
A
2
r +
B
2
r
2
_
cos θ r > a
Let us consider a boundary condition:
lim
r→∞
V = −E
0
r cos θ = V
2
Thus, we see that A
2
= −E
0
; but we can’t say anything about B
2
.
The consideration that V remain ﬁnite at the origin leads us to conclude that B
1
= 0. Hence, thus
far, we have:
V
1
= A
1
r cos θ r ≤ a
V
2
=
_
−E
0
r +
B
2
r
2
_
cos θ r > a
Let us now consider the ﬁrst ‘boundary’ condition; that the potential is continuous at the boundary.
That is:
V
1
(a, θ) = V
2
(a, θ)
So:
A
1
a cos θ =
_
−E
0
a +
B
2
a
2
_
cos θ
That is:
A
1
=
B
2
a
3
−E
0
(2.62)
Let us now apply the second ‘boundary’ condition; that D
⊥
is continuous at the boundary; this is
the case because there are no free charges on the boundary. If there were, there is a discontinuity
in D
⊥
, which we consider in later examples. So:
D
1⊥
(r = a) = D
2⊥
(r = a) ⇒ ε
r
E
1
(r = a) = E
2
(r = a)
2.6 Laplace Equation & Its Solutions 37
But, we have that E = −
∂V
∂r
; hence:
ε
r
∂V
1
∂r
¸
¸
¸
¸
r=a
=
∂V
2
∂r
¸
¸
¸
¸
r=a
⇒A
1
ε
r
cos θ =
_
−E
0
−
2B
2
a
3
_
cos θ
⇒A
1
ε
r
= −E
0
−
2B
2
a
3
So:
A
1
ε
r
= −E
0
−
2B
2
a
3
(2.63)
We can easily solve (2.62) and (2.63) for A
1
, B
2
; giving:
A
1
= −
3E
0
ε
r
+ 2
B
2
=
_
ε
r
−1
ε
r
+ 2
_
E
0
a
3
Hence, putting these back into our potentials for inside and outside:
V
1
(r, θ) = −
3E
0
ε
r
+ 2
r cos θ r ≤ a
V
2
(r, θ) = −E
0
r cos θ +E
0
cos θ
_
ε
r
−1
ε
r
+ 2
_
a
3
r
2
r > a
Hence solution found.
2.6.5 Example: Charge Inside Spherical Cavity
Consider a charge q at the centre of a spherical cavity; where outside the cavity is a dielectric,
permittivity ε
r
; and inside vacuum.
Again, the boundary conditions:
• Potential and D
⊥
continuous on boundary;
• Potential is zero at inﬁnity;
• Electric ﬁeld inside cavity, in limit r →0 is just the Coulomb ﬁeld.
Here, our system uses spherical polars, but with complete angular symmetry. Hence, the Laplace
equation is:
1
r
2
∂
∂r
r
2
∂V
∂r
= 0
We have already discussed the solution to this equation. It is just:
V (r) = A+
B
r
38 2 ELECTROMAGNETIC FIELD EQUATIONS
So, again, let us split the problem into inside/outside:
V
1
(r) = A
1
+
B
1
r
r ≤ a
V
2
(r) = A
2
+
B
2
r
r > a
To make the potential zero at r = ∞, we consider V
2
; and hence conclude that A
2
= 0. Let us
consider the continuity of the potential at the boundary:
A
1
+
B
1
a
=
B
2
a
Using A
2
= 0. That is:
B
2
= aA
1
+B
1
Let us consider the continuity of D
⊥
at the boundary. Hence, we have:
dV
1
dr
¸
¸
¸
¸
r=a
= ε
r
dV
2
dr
¸
¸
¸
¸
r=a
Noting that this time the vacuum is on the inside, and dielectic on the outside.
Thus:
−
B
1
a
2
= −ε
r
B
2
a
2
Thus:
B
1
= ε
r
B
2
Now, let us consider r →0. That is, inside:
E
1
→
q
4πε
0
r
2
Noting:
E
1
= −
dV
1
dr
= −
_
−
B
1
r
2
_
Hence:
B
1
= r
2
E
1
= r
2
q
4πε
0
r
2
=
q
4πε
0
We also have, via B
1
= ε
r
B
2
, that:
B
2
=
q
4πε
0
ε
r
And, via B
2
= aA
1
+B
1
, we have that:
A
1
= −
q
4πε
0
a
_
ε
r
−1
ε
r
_
Therefore, our solution is:
V
1
(r) = −
q
4πε
0
a
_
ε
r
−1
ε
r
_
+
q
4πε
0
r
r ≤ a
V
2
(r) =
q
4πε
0
ε
r
r
r > a
2.7 Multipoles 39
Let us consider what the surface charge induced is.
Recall that D = ε
0
E + P; where P is the polarisation. Also recall, that for a linear medium,
D = ε
r
ε
0
E. Hence, after a trivial rearangement:
P = (ε
r
−1)ε
0
E
Let us (as we are liberty to) consider the ﬁeld just outside the sphere. Then, we have that:
E
2
= −∇V
2
=
q
4πε
0
ε
r
r
2
ˆ r
Thus:
P =
q(ε
r
−1)
4πε
r
r
2
ˆ r
Finally, recall that ρ
ind
= −∇ P. Then, we have that:
ρ
ind
= −
q(ε
r
−1)
4πε
r
∇
ˆ r
r
2
= −
q(ε
r
−1)
ε
r
δ(r)
We have thus computed the surface charge density induced; where we have used:
∇
ˆ r
r
2
= 4πδ(r)
2.7 Multipoles
2.7.1 Electric Dipole
Figure 4: The electric dipole. Two charges, q, −q, separated by a distance d, observed at P; which
is at a position r, as measured from
d
2
Consider the potential due to two charges, q
1
and q
2
, at distances r
1
, r
2
from some observation
point P. Linear superposition allows us to write that the the ‘composite’ potential is just:
V (P) =
1
4πε
0
_
q
1
r
1
+
q
2
r
2
_
40 2 ELECTROMAGNETIC FIELD EQUATIONS
Now, consider the slightly more speciﬁc case, when q
1
= −q
2
= q. So:
V (P) =
q
4πε
0
_
1
r
1
−
1
r
2
_
Just a quick note, the cosine rule can be very easily derived. Consider:
(r −r
)
2
= r
2
+ (r
)
2
−2r r
= r
2
+ (r
)
2
−2rr
cos θ
So, with reference to the ﬁgure:
r
2
1
= r
2
+ (
d
2
)
2
−2r
d
2
cos θ
That is:
r
2
1
= r
2
+ (
d
2
)
2
−rd cos θ
Also, being careful that the angle is now cos α = cos(180 −θ) = −cos θ, we see that:
r
2
2
= r
2
+ (
d
2
)
2
+rd cos θ
So, if we combine these results:
r
2
1,2
= r
2
+
_
d
2
_
2
∓rd cos θ
That is, just:
r
2
1,2
= r
2
_
1 +
_
d
2r
_
2
∓
d
r
cos θ
_
Now, for r >> d, we can neglect such quadratic terms. Hence:
r
2
1,2
= r
2
_
1 ∓
d
r
cos θ
_
Thus, squarerooting:
r
1,2
= r
_
1 ∓
d
r
cos θ
_
1/2
Hence:
1
r
1,2
=
1
r
_
1 ∓
d
r
cos θ
_
−1/2
Now, to ﬁrst order binomial expansion; i.e. (1 +x)
n
≈ 1 +nx; we have:
1
r
1,2
=
1
r
_
1 ±
d
2r
cos θ
_
Hence:
1
r
1
−
1
r
2
=
1
r
_
1 +
d
2r
cos θ −1 +
d
2r
cos θ
_
=
d
r
2
cos θ
Therefore, we can write the potential, at r, due to a dipole:
V (r) =
q
4πε
0
_
1
r
1
−
1
r
2
_
=
q
4πε
0
d cos θ
r
2
(2.64)
Which we may write as:
V (r) =
1
4πε
0
p ˆ r
r
2
Where we have deﬁned the dipole moment p ≡ qd.
2.7 Multipoles 41
2.7.2 General Multipole Expansion
In the present multipole discussion, we assume that there is axial symmetry; which is why we use
Legendre polynomials, as will become apparent.
Here, let γ be the angle between r and r
, as shown in the ﬁgure, and R ≡ r −r
; so that R is the
Figure 5: The multipole. Consider a distribution of charges. Notice how things are deﬁned. We
have R ≡ r − r
, and that γ ≡ θ
− θ; where θ is the angle between the observation point P, and
the ˆ zaxis. Similarly, θ
is the angle between the axis and the charge distribution ‘bit’.
vector from the charge to the observation point, and γ the angle between charge and observation
point. So, under the cosine rule:
R
2
= r
2
+ (r
)
2
−2rr
cos γ
That is:
R
2
= r
2
_
1 +
_
r
r
_
2
−2
r
r
cos γ
_
Now, let:
δ ≡
r
r
_
r
r
−2 cos γ
_
(2.65)
Then:
R
2
= r
2
(1 +δ) ⇒ R = r
√
1 +δ
Hence:
1
R
=
1
r
(1 +δ)
−1/2
(2.66)
Now, the binomial expansion is such that:
(1 +x)
n
= 1 +nx +
n(n −1)
2!
x
2
+
n(n −1)(n −2)
3!
x
3
+. . .
Therefore:
(1 +δ)
−1/2
= 1 −
1
2
δ +
(−1/2)(−3/2)
2
δ
2
+
(−1/2)(−3/2)(−5/2)
6
δ
3
+. . .
= 1 −
1
2
δ +
3
8
δ
2
−
5
16
δ
3
+. . .
42 2 ELECTROMAGNETIC FIELD EQUATIONS
Hence, putting this expansion into (2.66), and taking ourselves out of using δ, via (2.65), we have:
1
R
=
1
r
_
1 −
1
2
_
r
r
__
r
r
−2 cos γ
_
+
3
8
_
r
r
_
2
_
r
r
−2 cos γ
_
2
+. . .
_
If we then gather appropriate terms:
1
R
=
1
r
_
1 +
r
r
cos γ +
_
r
r
_
2
3 cos
2
γ −1
2
+
_
r
r
_
3
5 cos
3
γ −3 cos γ
2
+. . .
_
Now, recall the Legendre polynomials:
P
0
(x) = 1 P
1
(x) = x P
2
(x) =
3x
2
−1
2
P
3
(x) =
5x
3
−3x
2
Thus, we notice those terms as present. Hence:
1
R
=
1
r
_
P
0
(cos γ) +
_
r
r
_
P
1
(cos γ) +
_
r
r
_
3
P
3
(cos γ) +. . .
_
We therefore write the whole thing as a sum:
1
R
=
1
r
∞
=0
_
r
r
_
P
(cos γ)
Which is just:
1
R
=
∞
=0
1
r
+1
(r
)
P
(cos γ)
Remembering that γ is the angle between r and r
. The above ‘motivation’ shows that the ex
pansion seems to work, but is by no means a proof! A proof of this may be found in the relevant
appendix, under the Legendre polynomial generating function, and the example relating to our
present discussion.
The order to which the sum is taken, is the order of the ‘pole’.
So, we have that the potential due to a small bit of charge is just:
dV =
1
4πε
0
dq
R
⇒ V (r) =
1
4πε
0
_
1
R
ρ(r)d
3
r
Hence, the total potential, at r, due to some charge density ρ(r
) is given by:
V (r) =
1
4πε
0
∞
=0
_
1
r
+1
P
(cos γ)(r
)
ρ(r
)d
3
r
To make this (possibly) a little more transparent, let us write out the ﬁrst few terms of the sum:
V (r) =
1
4πε
0
_
1
r
ρ(r
)d
3
r
+
r
r
2
cos γρ(r
)d
3
r
+
r
2
2r
3
(3 cos
2
γ −1)ρ(r
)d
3
r
=
1
4πε
0
_ _
1
r
+
r
r
2
cos γ +
r
2
2r
3
(3 cos
2
γ −1)
_
ρ(r
)d
3
r
2.7 Multipoles 43
We have that charges are arranged at some r
, relative to some orgin, somewhere roughly at the
centre of the distribution, and we are looking at the potential at a distance r away from the origin
of the system (i.e. at a position r). With computations on this, one must be very careful in noting
that:
cos γ = cos(θ
−θ) = cos θ cos θ
+ sin θ sin θ
And that the volume element is d
3
r
= r
2
sin θ
dr
dθ
dφ
.
So, we have a multipole expansion of V . We see that the ﬁrst term is the monopole = 0; then
dipole = 1, quadrupole = 2, and continuing up. Notice, at large distances, the multipole
expansion will be dominated by the monopole term; or the lowest pole term present.
For a point charge at the origin, all components other than the ﬁrst are zero; so the monopole is
the only contribution:
V (r) =
1
4πε
0
_
1
r
ρ(r
)d
3
r
=
q
4πε
0
r
The total charge is just q.
If the total charge vanishes, then the dipole term is dominant. Now, if we deﬁne the dipole moment
to be:
p =
_
r
ρ(r
)d
3
r
(2.67)
Then, as ˆ r =
r
r
, we have:
p ˆ r =
_
1
r
r r
ρ(r
)d
3
r
=
_
1
r
rr
cos γρ(r
)d
3
r
=
_
r
cos γρ(r
)d
3
r
Where we have noted that the angle between r and r
is γ, by deﬁnition, with reference to the
ﬁgure. Which is just the dipole term in the expansion:
V =
1
4πε
0
r
2
_
r
cos γρ(r
)d
3
r
=
1
4πε
0
r
2
p ˆ r
Recall the dipole moment we deﬁned for a pair of charges: p = qd. This is equivalent, as the
separation between charges is just r
, and total charge is given by the integral over charge density.
Thus, the dipole moment is better described in terms of the size, shape & density of the system, as
we have with the integral version.
Note, shifting the origin changes multipole moments; however, if the monopole term is zero, the
dipole moment is independant of the origin:
Suppose we have a dipole moment relative to some origin 0:
p =
_
r
ρ(r
)d
3
r
Then, suppose we shift the coordinates (the origin), so that r
= r
+ a. Hence, a new dipole
moment:
p
a
=
_
(r
+a)ρ(r
)d
3
r
=
_
r
ρ(r
)d
3
r
+
_
aρ(r
)d
3
r
= p +aq
44 2 ELECTROMAGNETIC FIELD EQUATIONS
But, as we said, q = 0; hence, p
a
= p.
Now, the electric ﬁeld due to a dipole:
V (r, θ) =
1
4πε
0
p cos θ
r
2
Hence, as E = −∇V , we have its components:
E
r
= −
∂V
∂r
=
1
4πε
0
2p cos θ
r
3
E
θ
= −
1
r
∂V
∂θ
=
1
4πε
0
p sin θ
r
3
E
ϕ
= 0
Hence:
E =
1
4πε
0
p
r
3
(2 cos θˆ r + sin θ
ˆ
θ)
It is common to notate this as:
E =
1
4πε
0
1
r
3
(3(p ˆ r)ˆ r −p)
So, to conclude, we have that we can write the scalar potential as a sum over multipole moments:
V (r, θ) =
1
4πε
0
∞
=0
p
r
+1
Where we have the pole moment:
p
=
_
r
P
(cos γ)ρ(r
)d
3
r
2.7.3 Spherical Harmonic Expansion Multipoles
Figure 6: The deﬁnition of angles in spherical polars. Notice that the angle between x and x
is γ.
Also, we shall be using r as x; and r
as x
; although this is a trivial assignment of symbols. Figure
from Jackson.
2.7 Multipoles 45
Recall the following expansion in terms of Legendre polynomials:
1
[r −r
[
=
1
r
+1
r
P
(cos γ)
Where γ is the angle between r and r
; and has the relation r r
= rr
cos γ. So, we had that the
potential may be written:
V (r) =
1
4πε
0
1
r
+1
_
r
P
(cos γ)ρ(r
)d
3
r
(2.68)
Now, we have the following relation:
P
(cos γ) =
4π
2 + 1
m=+
m=−
Y
m
(θ, φ)Y
∗
m
(θ
, φ
) (2.69)
We have that m = − → , in integer steps. Where, from the addition theorem for Spherical
harmonics:
cos γ = cos θ cos θ
+ sin θ sin θ
cos(φ −φ
)
Hence, inserting (2.69) into our potential expression (2.68) above:
V (r) =
1
4πε
0
4π
2 + 1
1
r
+1
m
_
r
Y
m
(θ, φ)Y
∗
m
(θ
, φ
)ρ(r
)d
3
r
=
1
4πε
0
4π
2 + 1
1
r
+1
m
Y
m
(θ, φ)
_
r
Y
∗
m
(θ
, φ
)ρ(r
)d
3
r
Now, to tidy this up. Let us write:
C
m
(θ, φ) ≡
_
4π
2 + 1
Y
m
(θ, φ) (2.70)
So that its conjugate is just (also changing its arguments):
C
∗
m
(θ
, φ
) =
_
4π
2 + 1
Y
∗
m
(θ
, φ
)
Then, we have that our potential expression cleans up to:
V (r) =
1
4πε
0
m
1
r
+1
C
m
(θ, φ)
_
r
C
∗
m
(θ
, φ
)ρ(r
)d
3
r
Further cleaning this up, let us deﬁne the multipole moment to be:
Q
m
≡
_
r
C
∗
m
(θ
, φ
)ρ(r
)d
3
r
(2.71)
Then:
V (r) =
1
4πε
0
m
1
r
+1
C
m
(θ, φ)Q
m
(2.72)
46 2 ELECTROMAGNETIC FIELD EQUATIONS
Remember, before, we had:
p
=
_
r
P
(cos γ)ρ(r
)d
3
r
Which is the axially symmetric version of the Q
m
above.
Now, in rectangular coordinates, we make the ‘connection’ that:
1
[r −r
[
= e
−r
·∇
1
r
So that the expansion is just:
e
−r
·∇
1
r
=
1
r
−(r
∇)
1
r
+
1
2
(r
∇)
2
1
r
+. . . +
(−1)
n
n!
(r
∇)
n
1
r
+. . .
Where we have used the standard expansion of the exponential:
e
x
=
∞
n=0
(−1)
n
n!
x
n
Now, we can evaluate some of the terms in the expansion:
1
[r −r
[
=
1
r
+
r
r
r
3
+
3(r
r)
2
−r
2
r
2
2r
5
+. . .
Then, using this as the expansion for a multipole:
V (r) =
1
4πε
0
_
_
q
r
+
p r
r
3
+
1
2
3
i,j=1
q
ij
x
i
x
j
r
5
+. . .
_
_
Where we have used the usual monopole and dipole expressions, as well as the ‘new’ quadrupole
tensor:
q
ij
=
_
(3x
i
x
j
−r
2
δ
ij
)ρ(r
)d
3
r
(2.73)
Which we state, and do not derive.
2.7.4 Relations Between Multipoles in Cartesian & Spherical Polars
Let us compute some components of the multipole:
Q
m
≡
_
r
C
∗
m
(θ
, φ
)ρ(r
)d
3
r
Where the coeﬃcients are given by (infact, its worth noting that they are the associated Legendre
polynomials, but we won’t go into that here):
C
∗
m
=
_
4π
2 + 1
Y
∗
m
(θ
, φ
)
2.7 Multipoles 47
Monopole Moment This is the component Q
00
. So, we look up the spherical harmonic:
Y
00
=
1
√
4π
⇒ Y
∗
00
= Y
00
Then, we have:
Q
00
=
_
√
4π
1
√
4π
ρ(r
)d
3
r
Which is just the total charge:
Q
00
=
_
ρ(r
)d
3
r
= q
Dipole Moment Let us look up the following spherical harmonics:
Y
10
=
_
3
4π
cos θ Y
11
= −
_
3
8π
sin θe
iφ
And the ﬁnal harmonic, i.e. Y
1−1
can be found from the relation:
Y
−m
= (−1)
m
Y
∗
m
Hence, let us start to compute things:
Q
10
=
_
r
_
4π
3
_
3
4π
cos θ
ρ(r
)d
3
r
=
_
r
ρ(r
) cos θ
d
3
r
= p
z
Where we have noted that r
cos θ
= z, as the standard conversion between spherical polars &
cartesian. Thus, we see that Q
10
is the zcomponent of the dipole moment.
Next:
Q
11
=
_
r
_
4π
3
_
−
_
3
8π
sin θ
e
−iφ
_
ρ(r
)d
3
r
= −
1
√
2
_
r
sin θ
e
−iφ
ρ(r
)d
3
r
= −
1
√
2
_
r
_
sin θ
(cos φ
−i sin φ
)
¸
ρ(r
)d
3
r
= −
1
√
2
__
r
sin θ
cos φ
ρ(r
)d
3
r
−i
_
r
sin θ
sin φ
ρ(r
)d
3
r
_
= −
1
√
2
(p
x
−ip
y
)
Again, where we have noted the use of the standard polarscartesian conversion. Finally:
Q
1−1
=
_
r
_
4π
3
_
_
3
8π
sin θ
e
iφ
_
ρ(r
)d
3
r
=
1
√
2
_
r
_
sin θ
(cos φ
+i sin φ
)
¸
ρ(r
)d
3
r
=
1
√
2
(p
x
+ip
y
)
48 2 ELECTROMAGNETIC FIELD EQUATIONS
Doing these, we must be careful that we conjugate the spherical harmonic before putting into the
integral.
Quadrupole Moment Here, we compute a component of the quadrupole moment. We look up
the following spherical harmonic:
Y
20
=
1
2
_
4
5π
(3 cos
2
θ −1)
Q
20
=
1
2
_
r
2
_
4π
5
_
4
5π
(3 cos
2
θ
−1)ρ(r
)d
3
r
=
1
2
_
[3(r
cos θ
)
2
−r
2
]ρ(r
)d
3
r
=
1
2
_
(3z
2
−r
2
)ρ(r
)d
3
r
=
1
2
q
33
That is, we ﬁnd that it is related to an element of the quadruple tensor.
2.7.5 Properties of Multipoles
We have see that the details of a charge distribution in a volume, are encoded into the multipoles
Q
m
. We have seen that:
• = 0 is the monopole moment: 2
0
= 1 charges;
• = 1 is the dipole moment: 2
1
= 2 charges;
• = 2 is the quadrupole moment: 2
2
= 4 charges;
• = 3 is the octupole moment: 2
3
= 8 charges.
We have also seen that Q
m
depend on the choice of origin; however, the ﬁrst nonzero moment is
frame independent.
The far ﬁeld potential is dominated by the ﬁrst nonzero moment.
2.8 Multipole Expansion of the Vector Potential
As we have previously seen, under the Coulomb gauge, we have that the magentic vector potential
satisﬁes the Poisson equation:
∇
2
A = −µ
0
J (2.74)
2.8 Multipole Expansion of the Vector Potential 49
Which, as we have seen, has solution:
A(r) =
µ
0
4π
_
J(r)
[r −r
[
d
3
r
(2.75)
Thus, exactly as we did for the scalar potential, we may expand this:
A(r) =
µ
0
4π
_
1
r
_
J(r
)d
3
r
+
r
r
3
_
r
J(r
)d
3
r
+. . .
_
This is a bit more transparent if we consider a single element of the potential:
A
i
(r) =
µ
0
4π
_
1
r
_
J
i
(r
)d
3
r
+
r
r
3
_
r
J
i
(r
)d
3
r
+. . .
_
(2.76)
Now, looking at the monopole term:
_
J
i
(r
)d
3
r
= 0
Which is zero (no monopoles). This can also be argued from vector calculus ground; which won’t
be done here. Looking at the dipole term (the second):
r
_
r
J
i
(r
)d
3
r
= x
j
_
x
i
J
i
d
3
r
= −
1
2
x
j
_
(x
i
J
j
−x
j
J
i
)d
3
r
= −
1
2
ijk
x
j
_
(r
J)
k
d
3
r
= −
1
2
_
r
_
(r
J)d
3
r
_
i
Hence, we have an expression for the magnetic dipole moment:
m =
1
2
_
r
J(r
)d
3
r
(2.77)
So that the dipole component of the magnetic vector potential is just (using a b = −b a):
A(r) =
µ
0
4π
mr
r
3
(2.78)
Thus, the magnetic ﬁeld induced by such a vector potential, from B = ∇A, is just:
B(r) =
µ
0
4π
3(m r)r −r
2
m
r
5
(2.79)
Where we have used the vector identity:
∇(r
n
a r) = r
n−2
[(n + 2)r
2
a −n(a r)r]
We shall stop with the magnetic analysis here.
50 2 ELECTROMAGNETIC FIELD EQUATIONS
2.9 Multipole Expansions Summary
If we have some axiallysymmetric charge distribution, at r
; with observation point at r, then,
we are able to expand the distribution in terms of multipoles, so that the scalar potential may be
written:
V (r, θ) =
1
4πε
0
∞
=0
1
r
+1
_
r
P
(cos γ)ρ(r
)d
3
r
(2.80)
Where the angle between the distribution and observation point is γ ≡ θ
−θ.
It can be useful to think about the pole term, which will be given by:
p
=
_
r
P
(cos γ)ρ(r
)d
3
r
(2.81)
So that the potential will be a sum over poles:
V (r, θ) =
1
4πε
0
∞
=0
1
r
+1
p
If, however, the charge distribution is not axiallysymmetric, then we must appeal to a sum over
spherical harmonics, as opposed to Legendre polynomials.
2.10 Method: Potentials and Surface Charges
Suppose we have the problem of ﬁnding the scalar potential in all space, in the presence of some
symmetric charge distribution.
The ﬁrst thing to do is to write the relations between D, E and φ at the boundary. That is:
D
out
−D
in
= ρ ε
out
ε
0
E
out
−ε
in
ε
0
E
in
= ρ ε
out
ε
0
∂φ
out
∂r
−ε
in
ε
0
∂φ
in
∂r
= −ρ φ
out
= φ
in
Where all are only valid at the boundary. The ﬁrst expression is that the perpendicular component
of the electric displacement ﬁeld is discontinuous in the presence of surface charge. The second and
third expressions just follow from the ﬁrst, using standard relations to link.
Usually, to ﬁnd φ, the solution in Spherical Polars will be used. This must then be divided into
two separate expressions: inside and outside. We must use ‘common sense’ boundary conditions
(it is usual that these are not stated in questions, but must be used). That is, at the origin and
inﬁnity, the potential does not diverge. Upon inspection of the potential expressions for inside &
outside, this immediately eliminates one coeﬃcient from each expression.
NOTE: this cannot be used if there is an applied electric ﬁeld. If there is an applied ﬁeld, then one
computes the potential (essentially) at inﬁnity, and uses this as a boundary condition.
Continuing, one is then able to express one coeﬃcient in terms of the other; by using the fact that
the potential is continuous at the boundary. To actually do this, one may either use linear inde
pendence of the coeﬃcients of Legendre polynomials, and read oﬀ the relation directly; or (possibly
the more complete method) use orthogonality: multiply the equation by ‘another’ polynomial.
2.11 Discussion 51
To then ﬁnd the remaining coeﬃcient, use the discontinuity of the derivative of the potential, at the
boundary. If possible, express the surface charge density in terms of Legendre polynomials, which
will make the next step a lot easier to do. Then, once the derivatives done, one must multiply
the whole equation by ‘another’ polynomial, and use orthogonality. Then, one will see that if the
surface charge is given by a ﬁnite number of polynomials, then the exact forms of the coeﬃcients
can be read oﬀ. Otherwise, only an approximate ‘inﬁnite sum’ will result as the coeﬃcients.
Hence, in this way, the scalar ﬁeld in all space may be found. This method works even if there is
a ﬁeld present, but care must be taken over the ‘inﬁnite’ boundary condition.
2.11 Discussion
So, let us review this section, and discuss the concepts introduced.
We started by introducing Maxwell’s four equations: Gauss’ law, no magnetic monopoles, Faradays
law & Amperes law. We discussed the electrostatic Coulomb gauge, which allowed us to derive a
wave equation for static ﬁelds only. We found that we must modify Amperes law, in light of the
continuity equation, to take account of time variation of ﬁelds.
We then looked a little at the eﬀect of matter & materials on the ﬁelds, with boundary conditions
for the presence of surface charge, and their eﬀect on the ﬁelds.
Next, we looked at time varying ﬁelds, employing the Lorentz gauge to derive wave equations which
are correct for timevarying ﬁelds. We also showed how much freedom we have in choosing the
ﬁelds, in terms of their associated potentials, in terms of invariance.
We then took a mathematical diversion, looking at the Diracδ function, and various properties &
uses. As well as a brief note on Green functions, with a speciﬁc case given.
We carried on with looking at Poynting’s theorem, which gives information on the energy of ﬁelds,
and the energy ﬂux of a ﬁeld, over a time period.
Then we discussed the solution to Laplaces equation, which arises in electrostatic systems, where
there are regions of no charge density. We discussed its solution in Cartesian, polar & spherical
polar coordinates, with solving some special cases with a speciﬁc set of boundary conditions.
Finally, we found a way of expressing a continuous distribution of charge, by approximating it to
a series of monopoles, dipoles etc. By writing the potentials as a sum like this, we are able to ﬁnd
some properties of the potentials, due to speciﬁc distribution of charges. We started with axially
symmetric systems (using Legendre polynomials), but continued with removing the axial symmetry
(using spherical harmonics: a fuller treatment of spherical harmonics & various related theorems
may be found in the appendix).
This concludes our discussion of electrostatic systems. We now move on to considering systems in
which charges move, and various consequences of the motion.
52 2 ELECTROMAGNETIC FIELD EQUATIONS
53
3 Retarded Potentials & Radiation
3.1 Introduction to Radiation from Accelerated Charges
Let us begin by considering the derivation of some wave equations. These will be waves ‘of potential
ﬁeld’, driven by ‘charge distributions’.
Consider the following Maxwell’s equations, for timevarying ﬁelds:
∇ E =
ρ
ε
0
∇B = µ
0
J +ε
0
µ
0
∂E
∂t
Consider also:
E = −∇V −
∂A
∂t
B = ∇A
So, inserting the ﬁrst into Gauss’ law:
∇
_
−∇V −
∂A
∂t
_
=
ρ
ε
0
Giving:
∇
2
V +
∂
∂t
∇ A = −
ρ
ε
0
(3.1)
And putting both into Amperes law:
∇∇A = µ
0
J +ε
0
µ
0
∂
∂t
_
−∇V −
∂A
∂t
_
Let us use the following vector identity, and relation:
∇∇A = ∇(∇ A) −∇
2
A c
2
=
1
ε
0
µ
0
Then, we have:
1
c
2
∂
2
A
∂t
2
−∇
2
A+∇
_
∇ A+
1
c
2
∂V
∂t
_
= µ
0
J (3.2)
Now, if we use the Lorentz gauge:
∇ A+
1
c
2
∂V
∂t
= 0 ⇒ ∇ A = −
1
c
2
∂V
∂t
(3.3)
Using this in both (3.1) and (3.2) gives:
1
c
2
∂
2
V
∂t
−∇
2
V =
ρ
ε
0
(3.4)
1
c
2
∂
2
A
∂t
2
−∇
2
A = µ
0
J (3.5)
54 3 RETARDED POTENTIALS & RADIATION
Hence, we have, using the Lorentz gauge, decoupled the equations into two inhomogeneous wave
equations. The solutions to these equations are given in terms of retarded scalar and vector poten
tials.
We must now come to the concept of retarded time. If we are at a point in space (i.e. the observer),
and we observe a charge distribution, which is moving, then, due to the ﬁnite speed of light, we
observe the distribution as it was when it emitted the light, not how it is now. That is, if we are at
a time t, then we see the charge distribution as it was at some time t
ret
, in the past.
So, we have solutions (which we shall later verify) to the scalar & vector potential:
V (r, t) =
1
4πε
0
_
ρ(r
, t
ret
)
[r −r
[
d
3
r
(3.6)
A(r, t) =
µ
0
4π
_
J(r
, t
ret
)
[r −r
[
d
3
r
(3.7)
Where, as usual, the observer is at r, and charge distribution of interest (i.e. the retarded distribu
tion) at r
. So, one supposes, it would be correct to say r
(t
ret
). That is, r
is the vector from the
retarded position to the origin.
Now, as the distance between observer and charge distribution is [r − r
[, the the time taken for
information to get from the charge distribution (as it was, at t
ret
), to the observer (at time t) is
just
1
c
[r −r
[. Hence:
t
ret
= t −
1
c
[r −r
[
_
t
adv
= t +
1
c
[r −r
[
_
Now, the solution to the right, with a +sign, is consistent; but would correspond to an advanced
time; which means that the eﬀect would precede the cause. Which is a violation of causality. That
is, a potential ﬁeld now would be due to the motion of charges in the future. As an aside, it has
actually been considered, by Feynman & Wheeler
1
that the potential now is actually a sum of the
advanced & retarded potential, and halved. They review a suggestion that radiation could be due
to interaction of the radiation, with an observer. The consequence of this is that (and this is an
example they give) if we look out of the window, at a star, which is (say) 4 light years away, the
only reason it emits radiation, is because the radiation is interacting with an absorber. That is,
when the light was emitted (4 years ago), it ‘knew’ that it would be interacting with material (i.e.
the observers eye) 4 year in its future. The cause is after the eﬀect. It is presumably like saying
that a tree does not make a sound if it falls in a forest, with noone around to listen. Anyway, we
digress:
Now, suppose we consider point charges:
ρ(r, t) = qδ(r −r
0
(t)) J(r, t) = qv(t)δ(r −r
0
(t))
Suppose that the charge distribution is travelling at velocity βc (at retarded time). Then, after a
fair amount of algebra, which we will do later; we end up with the LienardWiechart potentials:
V (r, t) =
1
4πε
0
_
q
κR
_
ret
(3.8)
A(r, t) =
µ
0
4π
_
cβ
κR
_
ret
(3.9)
1
Interaction with the Absorber as the Mechanism of Radiation: J.Wheeler & R.Feynman Rev.Mod.Phys Vol17,
No2, 1945
3.1 Introduction to Radiation from Accelerated Charges 55
Where, if R ≡ r −r
:
ˆ
R =
R
R
κ = 1 −
ˆ
R β
So, for velocities close to the speed of light (i.e. β ≈ 1), we see that there will be a strong angle
dependance, due to
ˆ
R β = βc cos θ. So, a strong dependance on the angle that the moving charge
is viewed, for speeds close to that of c.
So, how does this eﬀect the radiation of energy; i.e. the Poynting vector P = EH (which is the
energy density ﬂow, per unit area). If a charge is static, then A = 0. Hence, P = 0; hence, static
charges do not radiate.
Now, we know that if a particle is moving at constant velocity, then an inertial frame of reference
can be found in which the particle is stationary. Hence, for any charge which is moving at a
constant speed, an inertial frame of reference can be found in which it is stationary. Hence, it does
not radiate. Thus, any charge moving at a constant velocity does not radiate. Only accelerating
charges radiate.
Now, once the potentials are converted into ﬁelds, we will ﬁnd that there are two terms: one
dependant upon velocity: cβ, and one dependant upon acceleration: c
˙
β. Hence, due to our previous
arguement, only the acceleration terms will radiate.
Let us look at the radial dependancies of the energy ﬂux (i.e. of the radiation, P). Now, if
the maths is done, we ﬁnd a (electric) ‘velocity ﬁeld’, which is proportional to
1
R
2
, whereas the
‘acceleration ﬁeld’ is proportional to
1
R
. Now, as magnetic ﬁeld is proportional to electric ﬁeld, up
to B =
1
c
ˆ
k E, we see that for the velocity ﬁeld, P ∝
1
R
2
1
R
2
=
1
R
4
. Hence, the radiation ﬁeld, for
the velocity ﬁeld, is ∝
1
R
4
. This is the radiation per unit area, hence, the total radiation ﬂux is the
integral
_
Pd
2
r ∝
1
R
2
. Hence, the radiation ﬁeld for the velcity ﬁeld decays to zero.
Now, if we do this for the acceleration ﬁeld, we see that P ∝
1
R
2
⇒
_
Pd
2
r → 1. That is, the
radiation ﬁeld, for the acceleration ﬁeld, is a constant.
Hence, because the velocity term decays, but the acceleration term stays constant, we refer to the
acceleration term as the radiation ﬁeld.
The equations reﬀered to are:
E
rad
=
q
4πε
0
_
ˆ
R[(
ˆ
R−β)
˙
β]
c(1 −β
ˆ
R)
3
R
_
E
velocity
=
q
4πε
0
_
(
ˆ
R−β)(1 −β
2
)
c(1 −β
ˆ
R)
3
R
2
_
Both evaluated at the retarded time. Notice, the ﬁrst term is the acceleration ﬁeld, which (at the
end of our discussion) we call the radiation ﬁeld. These equations are merely stated here, but will
be discussed in detail later.
Consider the following magnitudes, for low β. That is, for nonrelativistic motion of charges:
E
vel
=
q
4πε
0
1
κ
3
R
2
E
rad
=
q
4πε
0
˙
β
cκ
3
R
Where κ ≡ 1 −β
ˆ
R. So, taking the ratio:
E
rad
E
vel
=
R
c
˙
β
56 3 RETARDED POTENTIALS & RADIATION
From which is see that if no acceleration (i.e.
˙
β = 0), then no radiation. Also, if c is inﬁnite, then
there would be no radiation.
Also, consider radiation with some characteristic frequency of oscillation, via c = νλ, then we get:
E
rad
E
vel
= R
β
λ
Hence, we see that the velocity ﬁeld dominates in the near zone, and radiation in the far zone.
3.1.1 Example: Larmor’s Formula
Consider the radiation from a nonrelativistic particle. So, we have that β << 1. Then, we have
that, in this limit:
E
rad
=
q
4πε
0
c
˙
β
R
sin θ
Where θ is the angle between the direction of acceleration and the observer. Now, we also have:
B
rad
=
1
c
ˆ
R
ret
E
rad
Hence, the Poynting vector:
P =
1
µ
0
E B
=
1
µ
0
c
E
rad
ˆ
R
ret
E
rad
=
1
µ
0
c
_
E
2
rad
ˆ
R
ret
−(E
rad
ˆ
R
ret
)E
rad
)
_
⇒[P[ = P =
1
µ
0
c
E
2
rad
Where we have used that E
rad
and
ˆ
R
ret
are perpendicular; as we will see later. So, inserting our
expressions in:
P =
1
µ
0
c
q
2
(4πε
0
c)
2
˙
β
2
R
2
sin
2
θ
=
q
2
16π
2
ε
0
c
˙
β
2
R
2
sin
2
θ
Hence, we have arrived at Larmor’s formula:
P =
q
2
16π
2
ε
0
c
˙
β
2
R
2
sin
2
θ (3.10)
Now, this is the power radiated per unit area, per unit time. So, the total power radiated, per unit
time, is just:
dW
dt
≡
¯
P =
_
PR
2
dΩ
3.1 Introduction to Radiation from Accelerated Charges 57
That is:
¯
P =
dW
dt
=
q
2
16π
2
ε
0
c
˙
β
2
_
sin
2
θ
R
2
R
2
sin θdθdφ
= Γ2π
_
π
θ=0
sin
3
θdθ
= Γ2π
_
1
−1
1 −x
2
dx
= Γ2π
4
3
=
q
2
6πε
0
c
˙
β
2
We shall denote this expression for the total power radiated, under Larmors assumptions:
¯
P
L
=
q
2
6πε
0
c
˙
β
2
(3.11)
Hence, just to check that this satisﬁes our intuition: as we cannot have negative power, the sign
of the charge must be even (i.e. the q
2
term). The power radiated must depend only on the
acceleration, and not velocity. Thus, our intuition is satisﬁed.
Again, even though we have assumed β << 1, if c →∞, then power radiated goes to zero. Hence,
that c is ﬁnite is important. Hence, radiation is a relativistic eﬀect.
Note, in Larmor’s formula, we see a sin
2
θ dependance. That is, no power is radiated along the
direction of acceleration.
3.1.2 Retarded Potentials & the Wave Equation
Now, in the previous section, we stated the retarded potentials:
V (r, t) =
1
4πε
0
_
ρ(r
, t
ret
)
[r −r
[
d
3
r
(3.12)
A(r, t) =
µ
0
4π
_
J(r
, t
ret
)
[r −r
[
d
3
r
(3.13)
Where t
ret
= t −
1
c
R, with R ≡ [r −r
[. That is, the above are:
V (r, t) =
1
4πε
0
_
ρ(r
, t −
1
c
R)
R
d
3
r
A(r, t) =
µ
0
4π
_
J(r
, t −
1
c
R)
R
d
3
r
Now, we wish to show that they do indeed satisfy the wave equations; and we do so for the scalar
potential, V :
_
∇
2
−
1
c
2
∂
2
∂t
2
_
V (r, t) = −
ρ(r, t)
ε
0
58 3 RETARDED POTENTIALS & RADIATION
Now, to proceed we compute the following Laplacian:
∇
2
ρ
R
= ρ∇
2
1
R
+
1
R
∇
2
ρ
Now, we also have the result:
∇
2
1
R
= −4πδ(R) = −4πδ(r −r
)
Now, as we have that ρ(r
, t −
1
c
R), we know it must be a solution to a wave equation (as that what
its solutions look like). Thus, its wave equation is:
∇
2
ρ =
1
c
2
∂
2
ρ
∂t
2
Hence:
∇
2
ρ
R
=
1
R
1
c
2
∂
2
ρ
∂t
2
−4πρδ(r −r
)
Now, the point of computing this was to be able to write:
∇
2
V =
1
4πε
0
_
∇
2
ρ(r
, t −
1
c
R)
R
d
3
r
=
1
4πε
0
_
1
R
1
c
2
∂
2
ρ
∂t
2
−4πρδ(r −r
)d
3
r
That is:
∇
2
V =
1
4πε
0
1
c
2
∂
2
∂t
2
_
ρ(r
, t −
1
c
R)
R
d
3
r
−
1
ε
0
_
δ(r −r
)ρ(r
, t −
1
c
R)d
3
r
Which is just:
∇
2
V =
1
c
2
∂
2
V
∂t
2
−
1
ε
0
ρ(r, t)
As we have noted that:
_
δ(r −r
)ρ(r
, t −
1
c
[r −r
[)d
3
r
= ρ(r, t)
Thus, we have shown that the proposed representation of the retarded scalar potential does indeed
satisfy the wave equation.
3.2 LienardWiechert Potentials: Point Charges
Here, we consider the scalar & vector potentials generated by a moving point charge q.
Consider a point charge, having some coordinates ρ(r
, t
). That is, we consider it at some (re
tarded) position & time: where it was, when it was. So, we may represent this as a deltafunction:
ρ(r
, t
) = qδ(r
−r
0
(t
))
That is, r
0
(t
) is some position within the charge. We have used that it is at the retarded time:
t
= t −
1
c
[r −r
[
3.2 LienardWiechert Potentials: Point Charges 59
Now, the potential is given by:
V (r, t) =
1
4πε
0
_
ρ(r
, t
)
[r −r
[
d
3
r
So, as one can imagine, integration here is complicated, as the retarded time is a function of the
integration variable:
V (r, t) =
q
4πε
0
_
1
[r −r
[
δ(r
−r
0
(t
))d
3
r
To go further, we look at the charge distribution again. We rewrite it using a delta function for
the retarded time:
ρ(r
, t
) = q
_
δ(τ −t
)δ(r
−r
0
(τ)) dτ t
≡ t −
1
c
[r −r
[
Notice then, this integral will only allow the value τ = t
in the position deltafunction. This integral
obviously evaluates to exactly what we had before; the reason for using this will (hopefully) become
clearer. Then, if we put this into the potential integral above:
V (r, t) =
q
4πε
0
_ _
1
[r −r
[
δ(τ −t
)δ(r
−r
0
(τ)) dτd
3
r
Notice, the d
3
r
integral will just have the eﬀect of sending r
→r
0
(τ). Thus:
V (r, t) =
q
4πε
0
_
1
[r −r
0
(τ)[
δ(τ −t
)dτ
To do this integral, we note that t
is a function of r:
V (r, t) =
q
4πε
0
_
1
[r −r
0
(τ)[
δ(τ −(t −
1
c
[r −r
0
(τ)[))dτ
Let us denote R(τ) ≡ r −r
0
(τ). Then, the above is:
V (r, t) =
q
4πε
0
_
1
R(τ)
δ(τ −(t −
1
c
R(τ)))dτ
Let us continue by noting some diracdelta theory:
δ(f(x)) =
i
δ(x −x
i
[f
(x
i
)[
Where the function f(x) has zeros at x
i
. Thus, in our case, we have f(τ) = τ −t +
1
c
R(τ). Then:
df
dτ
= 1 +
1
c
dR
dτ
And it has zeros at:
τ = t −
1
c
R(τ) ≡ t
60 3 RETARDED POTENTIALS & RADIATION
Infact, this t
is the same ‘class’ of coordinate as r
0
. So, we shall call it t
0
. Hence, using this:
δ(τ −(t −
1
c
R(τ))) = δ(τ −t
0
)
1
1 +
1
c
dR
dτ
¸
¸
(τ=t
0
)
Then, our integral for the potential becomes just:
V (r, t) =
q
4πε
0
1
1 +
1
c
dR
dτ
¸
¸
(τ=t
0
)
_
1
R(τ)
δ(τ −t
0
)dτ
Which is easily evaluated to be:
V (r, t) =
q
4πε
0
1
1 +
1
c
dR
dτ
¸
¸
τ=t
0
1
R(t
0
)
Now, we shall show that:
1
c
dR
dτ
¸
¸
¸
¸
τ=t
0
= −
β R
R
¸
¸
¸
¸
τ=t
0
So, let us use the chain rule:
dR
dτ
=
dR
dr
0
dr
0
dτ
So:
dR
dr
0
=
d
dr
0
[r −r
0
[ = −
r −r
0
[r −r
0
[
= −
R
R
And also, we see that the velocity of the beam is present:
dr
0
dτ
= v = βc
Hence:
1
c
dR
dτ
= −
R β
R
= −
ˆ
R β
Thus shown. Hence, our potential is:
V (r, t) =
q
4πε
0
1
1 −
R·β
R
1
R
Where everything is evaluated at τ = t
0
. Thus, we have the LienardWiechert scalar potential :
V (r, t) =
q
4πε
0
_
1
R(1 −β
ˆ
R)
_
ret
(3.14)
By very similar considerations, we can ﬁnd the vector potential:
A(r, t) =
µ
0
4π
_
J(r
, t
)
[r −r
[
d
3
r
Here, we use that:
J(r
, t
) = cβ(t
)ρ(r
, t
)
3.2 LienardWiechert Potentials: Point Charges 61
Then use the exact same argument, for deltafunctions. Giving the LienardWiechert vector poten
tial :
A(r, t) =
qµ
0
4π
_
cβ
R(1 −β
ˆ
R)
_
ret
(3.15)
Now, from these potentials, we are able to calculate the ﬁelds, from:
E = −∇V −
∂A
∂t
B = ∇A
After some (!!) algebra, one obtains the the LienardWiechert ﬁelds:
E =
q
4πε
0
_
(
ˆ
R−β)(1 −β
2
)
(1 −β
ˆ
R)
3
R
2
+
ˆ
R((
ˆ
R−β)
˙
β)
c(1 −β
ˆ
R)
3
R
_
(3.16)
B =
µ
0
qc
4π
_
(β
ˆ
R)(1 −β
2
)
(1 −β
ˆ
R)
3
R
2
+
˙
β
ˆ
R(β
ˆ
R)
c(1 −β
ˆ
R)
3
R
+
˙
β
ˆ
R
c(1 −β
ˆ
R)
2
R
_
(3.17)
Where:
˙
β ≡
d
dτ
β(τ) (3.18)
Notice, we have the accleration & velocity terms in both the electric and magnetic ﬁeld. The
acceleration ﬁelds are those with a
1
R
dependance, and velocity have
1
R
2
. We can also recover a
previously known relation:
B =
1
c
ˆ
R
ret
E (3.19)
3.2.1 Features of LienardWiechert Potentials
The potentials are relativistically correct  the are Lorentz covariant. That is, shifting reference
frames dosent change anything it shouldn’t!
For point charges, the magnetic ﬁeld is always perpendicular to the electric ﬁeld.
We have seen that we can decompose the ﬁelds into velocity and acceleration components:
E = E
v
+E
v
B = B
v
+B
a
Where:
E
v
, B
v
∝ β,
1
R
2
E
a
, B
a
∝
˙
β,
1
R
The energy radiated by a moving charge, per unit time is given by:
d
¯
P = P
ˆ
RdA = P
ˆ
RR
2
dΩ
This goes to:
d
¯
P
dΩ
= lim
R→∞
_
P
ˆ
RR
2
dt
dt
ret
_
(3.20)
62 3 RETARDED POTENTIALS & RADIATION
As the energy radiated, per unit time, per unit solid angle.
The components of the Poynting vector are:
P
vv
∝ [E
v
B
v
[ ∝
1
R
4
(3.21)
P
va
∝ [E
v
B
a
[ ∝
1
R
3
(3.22)
P
aa
∝ [E
a
B
a
[ ∝
1
R
2
(3.23)
And thus, the only nonvanishing component, as R →∞ is P
aa
.
One may think of the lackof radiation of the velocity components, as the energy being ‘convected’
along with the beam. The velocity term does have an associated energy  it is linked to the kinetic
energy of the beam.
Let us consider an example.
3.2.2 Example: Particle Moving With Constant Velocity
So, we have a particle moving with constant velocity cβ.
Figure 7: The setup for retarded motion.
If we have that the particle moves with constant velocity, then obviously
˙
β = 0. Thus, only the
velocity term of the LienardWiechert ﬁeld contributes:
E =
q
4πε
0
_
(
ˆ
R−β)(1 −β
2
)
(1 −β
ˆ
R)
3
R
2
_
ret
With reference to the ﬁgure, positions at the retarded time are denoted with a subscript r. Thus,
the above is just:
E =
q
4πε
0
(1 −β
2
)
(1 −β
ˆ
R
r
)
3
R
2
r
(
ˆ
R
r
−β) (3.24)
3.2 LienardWiechert Potentials: Point Charges 63
Now, the main goal is to get everything in terms of positions now, rather than at the retarded time.
We do this by careful vector algebra.
Consider the distance the particle travels between t
ret
and t. This is just its velocity multiplied by
the diﬀerence in times. That is, the distance the particle travels is just:
cβ(t −t
ret
)
Hence, with reference to the ﬁgure, we may then write:
R
r
= cβ(t −t
ret
) +R
p
(3.25)
Which is (obviously) completely the same as writing:
R
p
= R
r
−cβ(t −t
ret
)
Also, the distance a signal travel, at speed c, between t
ret
and t is just R
r
; by deﬁnition. Thus, we
also have:
R
r
= c(t −t
ret
)
Thus, combining the above two results:
R
p
= R
r
−R
r
β
That is, using the standard R = R
ˆ
R, just:
R
p
= R
r
(
ˆ
R
r
−β) ⇒ (
ˆ
R
r
−β) =
R
p
R
r
(3.26)
Hence, using this in (3.24) gives:
E =
q
4πε
0
(1 −β
2
)
(1 −β
ˆ
R
r
)
3
R
3
r
R
p
(3.27)
Thus, the only ‘vector’ left (and hence direction), is actually the actual position R
p
, rather than
the retarded position R
r
. Hence, the electric ﬁeld is in the direction of where the ﬁeld is now, but
due to where it was! So, let us continue, and get the denominator in terms of ‘now’ variables.
Now, let us square the lefthand expression in (3.26):
R
2
p
= R
2
r
(
ˆ
R
r
−β)
2
= R
2
r
(1 −2
ˆ
R
r
β +β
2
)
= R
2
r
−2R
2
r
ˆ
R
r
β +R
2
r
β
2
That is:
R
2
p
= R
2
r
−2R
2
r
ˆ
R
r
β +R
2
r
β
2
(3.28)
Now, with reference to the ﬁgure; we notice that the vertical in the two triangles is the same. So:
[R
r
β[
2
= [R
p
β[
2
(3.29)
64 3 RETARDED POTENTIALS & RADIATION
Let us expand the LHS, carefully:
[R
r
β[
2
= [R
r
β sin θ
r
]
2
= R
2
r
β
2
sin
2
θ
r
= R
2
r
β
2
(1 −cos
2
θ
r
)
= R
2
r
β
2
−(R
r
β)
2
Where θ
r
is the angle between β and R
r
. We have noted that the deﬁntion of the dotproduct
comes out, which we have then used. We can then easily see that (3.29) is just:
R
2
r
β
2
−(R
r
β)
2
= R
2
p
β
2
−(R
p
β)
2
(3.30)
Now, let us compute the following:
R
2
r
(1 −β
ˆ
R)
2
= (R
r
−β R
r
)
2
Which gives:
R
2
r
(1 −β
ˆ
R)
2
= R
2
r
−2R
r
β R
r
+ (β R
r
)
2
Now, substitute (3.28) into the above, for R
2
r
; giving:
R
2
r
(1 −β
ˆ
R)
2
= R
2
p
+ 2R
2
r
ˆ
R
r
β −R
2
r
β
2
−2R
r
β R
r
+ (β R
r
)
2
(3.31)
If we rearrange (3.30) slightly, we have:
(R
r
β)
2
−R
2
r
β
2
= (R
p
β)
2
−R
2
p
β
2
Substituting this into (3.31), for the third and last terms:
R
2
r
(1 −β
ˆ
R)
2
= R
2
p
+ 2R
2
r
ˆ
R
r
β −2R
r
β R
r
+ (R
p
β)
2
−R
2
p
β
2
Noting that two terms (note that 2R
2
r
ˆ
R
r
β = 2R
r
R
r
β) cancel gives:
R
2
r
(1 −β
ˆ
R)
2
= R
2
p
+ (R
p
β)
2
−R
2
p
β
2
Now, notice, the LHS of the above is ‘almost’ the same factor as we have in the denominator of the
electricﬁeld expression (3.27) (it is the same, if we take the 3/2 power of the above). And also, the
above is purely in terms of ‘now’ ! Let us expand this out now, getting some angular dependance.
R
2
r
(1 −β
ˆ
R)
2
= R
2
p
+ (R
p
β cos θ)
2
−R
2
p
β
2
= R
2
p
+R
2
p
β
2
(cos
2
θ −1)
= R
2
p
−R
2
p
β
2
sin
2
θ
= R
2
p
(1 −β
2
sin
2
θ)
Hence, we can use this in the denominator of (3.27):
E =
q
4πε
0
(1 −β
2
)
(1 −β
ˆ
R
r
)
3
R
3
r
R
p
=
q
4πε
0
(1 −β
2
)
(1 −β
2
sin
2
θ)
3/2
R
3
p
R
p
=
q
4πε
0
(1 −β
2
)
(1 −β
2
sin
2
θ)
3/2
R
2
p
ˆ
R
p
3.2 LienardWiechert Potentials: Point Charges 65
Thus, we have computed the electric ﬁeld in terms of the present position of the particle, and the
angle in which the particle is viewed:
E =
q
4πε
0
(1 −β
2
)
(1 −β
2
sin
2
θ)
3/2
R
2
p
ˆ
R
p
Also, we can ﬁnd the magnetic ﬁeld, from the following expression:
B =
1
c
ˆ
R
r
E
Now, from (3.25), we have that:
R
r
= cβ(t −t
ret
) +R
p
And hence:
ˆ
R
r
=
R
r
R
r
=
cβ(t −t
ret
)
R
r
+
R
p
R
r
We also had that R
r
= c(t −t
ret
). Hence:
ˆ
R
r
= β +
R
p
R
r
And therefore:
B =
1
c
_
β +
R
p
R
r
_
E
Now, R
p
and E are along the same line. Hence, that term is zero. Thus:
B =
1
c
β E
Hence, using our electric ﬁeld:
B =
q
4πε
0
c
(1 −β
2
)
(1 −β
2
sin
2
θ)
3/2
R
2
p
(β
ˆ
R
p
)
Also, note that:
1
4πε
0
c
=
µ
0
c
4π
Hence, to summarise what we have done: we have found the electric and magnetic ﬁelds due to a
moving particle. We have found that they do not depend upon retarded time, but only on the time
that the particle is in presently.
66 3 RETARDED POTENTIALS & RADIATION
3.3 Radiation
Let us consider the radiation emitted by a particle moving. We shall consider the two cases of velcity
and acceleration being parallel and perpendicular; but before then, we shall derive the general form
of radiation from a moving charge.
3.3.1 General Theory of Radiation
Now, we know that radiation is only emitted from the acceleration component of the electric ﬁeld:
E
e
=
q
4πε
0
_
ˆ
R(
ˆ
R−β)
˙
β
(1 −β
ˆ
R)R
_
ret
(3.32)
With the corresponding (acceleration components of) magnetic ﬁeld being:
B
a
=
1
c
[
ˆ
R]
ret
E
a
(3.33)
Now, from these expressions, we can see that both E
a
and B
a
are perpendicular to
ˆ
R. That is:
E
a
ˆ
R = 0 (3.34)
Now, the radiation per unit area is given by the Poynting vector, as we know:
P =
1
µ
0
E
a
B
a
Thus, using (3.33) in the above expression:
P =
1
µ
0
c
E
a
[
ˆ
R]
ret
E
a
Let us use the vector identity a b c = (a c)b −(a b)c; which gives:
P =
1
µ
0
c
_
E
2
a
[
ˆ
R]
ret
−(E
a
[
ˆ
R]
ret
)E
a
_
However, from (3.34), this is just:
P =
1
µ
0
c
E
2
a
[
ˆ
R]
ret
(3.35)
Now, the amount of energy dW, going through some bit of surface dA, per unit time is:
dW
dt
= P dA = [P
ˆ
RR
2
dΩ]
ret
Now, as the power
¯
P is d
¯
P = P
ˆ
RR
2
dΩ  the amount of ‘poynting vector’ going through an area,
we obviously have:
d
¯
P
dΩ
= [P
ˆ
RR
2
]
ret
3.3 Radiation 67
Hence, upon comparison of the two above expressions, we can write:
d
2
W
dtdΩ
= [P
ˆ
RR
2
]
ret
Now, if the charge is accelerated between times t
1
→t
2
, how much energy is lost? We can obviously
write:
dW
dΩ
=
_
t
2
t
1
[P
ˆ
RR
2
]
ret
dt
Hence, as we can convert between retarded times thus:
t
i
= t
i
+
R(t
i
)
c
⇒
dt
dt
= 1 +
1
c
dR
dt
Where we have previously seen that:
1
c
dR
dt
= −[
ˆ
R β]
ret
Hence, let us change integration variables:
dW
dΩ
=
_
t
2
t
1
[P
ˆ
RR
2
]
ret
dt
dt
dt
=
_
t
2
t
1
[P
ˆ
RR
2
]
ret
_
1 −[
ˆ
R β]
ret
_
dt
Therefore, divinding out the dt
:
d
2
W
dt
dΩ
=
_
(P
ˆ
RR
2
)(1 −
ˆ
R β)
_
ret
And this, upon comparison of the LHS and previous expressions, is:
d
¯
P
dΩ
=
_
(P
ˆ
R)(1 −
ˆ
R β)R
2
_
ret
Hence, using (3.35) and (3.32), this is just:
d
¯
P
dΩ
=
q
2
16π
2
ε
0
c
_
¸
_
_
ˆ
R(
ˆ
R−β)
˙
β
_
2
(1 −β
ˆ
R)
5
_
¸
_
ret
(3.36)
Now, we derived
¯
P in the comoving frame of the charge q; but the result is valid for any frame,
including the rest frame of the lab. Thus,
¯
P is invariant under Lorentz transformations.
Let us now consider colinear acceleration:
68 3 RETARDED POTENTIALS & RADIATION
3.3.2 Radiation: Acceleration & Velocity Parallel
This is sometimes denoted colinear motion.
Here, we have that the acceleration and velocity of the particle are in the same direction. Hence,
this immediately tells us that:
β
˙
β = 0
As is the condition for two things being parallel. Hence, using this:
d
¯
P
dΩ
=
q
2
16π
2
ε
0
c
_
¸
_
_
ˆ
R
ˆ
R
˙
β
_
2
(1 −β
ˆ
R)
5
_
¸
_
ret
Now, we use a vector identity
2
, to see that:
ˆ
R
ˆ
R
˙
β =
˙
β cos θ
ˆ
R−
˙
β
Thus, squaring it:
(
ˆ
R
ˆ
R
˙
β)
2
= (
˙
β cos θ
ˆ
R−
˙
β)
2
=
˙
β
2
cos
2
θ +
˙
β
2
−2
˙
β
2
cos
2
θ
=
˙
β
2
(1 −cos
2
θ)
=
˙
β
2
sin
2
θ
Hence, using this:
d
¯
P
dΩ
=
q
2
16π
2
ε
0
c
_
˙
β
2
sin
2
θ
(1 −β
ˆ
R)
5
_
ret
(3.37)
And, with β << 1, we obtain the previously derived Larmor formula:
d
¯
P
dΩ
=
q
2
16π
2
ε
0
c
˙
β
2
sin
2
θ (3.38)
Where θ is the angle between
˙
β and
ˆ
R.
Hence, we have derived the angular distribution of radiation, from a charge moving parallel to the
direction in which it is being accelerated. We have also derived the form for nonrelativistic motion.
Using (3.37), let us ﬁnd the angle at which most radiation is emmited.
To do this, let us isolate the θdependance, and diﬀerentiate:
d
dθ
sin
2
θ
(1 −β cos θ)
5
=
2 cos θ sin θ
(1 −β cos θ)
5
−
5β sin
3
θ
(1 −β cos θ)
6
= 0
And set it to zero (maximum). Solving:
[2 cos θ(1 −β cos θ) −5β sin
2
θ] sin θ = 0
2
a ×b ×c = (a · c)b −(a · b)c
3.3 Radiation 69
Solve this by the quadratic formula, with x ≡ cos θ, giving:
x =
−1 ±
√
1 + 15β
3β
And thus, the maximum (as opposed to the minimum) is, from x = cos θ:
θ
max
= cos
−1
_
−1 +
√
1 + 15β
3β
_
Let us ﬁnally ﬁgure out the total power radiated by a nonrelativistic particle. We hence integrate
(3.38):
¯
P
L
=
q
2
˙
β
2
16π
2
ε
0
c
_
dφ
_
dθ sin
2
θ sin θ
Which, with a substitution
3
, gives:
¯
P
L
=
q
2
˙
β
2
6πε
0
c
3.3.3 Radiation: Acceleration & Velocity Perpendicular
This corresponds to a charged particle being accelerated in a circular orbit. This will correspond
to
˙
β β = 0
Let us start with our general formula for radiation:
d
¯
P
dΩ
=
q
2
16π
2
ε
0
c
_
¸
_
_
ˆ
R(
ˆ
R−β)
˙
β
_
2
(1 −β
ˆ
R)
5
_
¸
_
ret
Let us, for convenience, write this in the following way:
d
¯
P
dΩ
=
q
2
16π
2
ε
0
c
_
Ξ
2
(1 −β
ˆ
R)
5
_
ret
Ξ ≡
ˆ
R(
ˆ
R−β)
˙
β
Let us expand out the top, using a vector identity as we go:
Ξ =
ˆ
R
ˆ
R
˙
β −
ˆ
Rβ
˙
β
= (
ˆ
R
˙
β)
ˆ
R−
˙
β −(
ˆ
R
˙
β)β + (
ˆ
R β)
˙
β
=
ˆ
R
˙
β(
ˆ
R−β) −
˙
β(1 −
ˆ
R β)
Hence, as the numerator in the expression for radiation is squared, let us square our expression (the
very last one):
Ξ
2
=
_
ˆ
R
˙
β(
ˆ
R−β) −
˙
β(1 −
ˆ
R β)
_
2
= (
ˆ
R
˙
β)
2
(
ˆ
R−β)
2
+
˙
β
2
(1 −
ˆ
R β)
2
−2ξ
3
See the appendix for a note on doing such integrals
70 3 RETARDED POTENTIALS & RADIATION
Where ξ is the ‘crossterm’ that always results from squaring a bracket:
ξ ≡ [(
ˆ
R
˙
β)(
ˆ
R−β)] [
˙
β(1 −
ˆ
R β)]
Let us compute this, carefully (remembering that terms that are already ‘dotted’ are a scalar, and
are just treated as ‘numbers’):
ξ ≡ [(
ˆ
R
˙
β)(
ˆ
R−β)] [
˙
β(1 −
ˆ
R β)]
= (
ˆ
R
˙
β)
ˆ
R
˙
β −(
ˆ
R
˙
β)(
ˆ
R
˙
β)(
ˆ
R β)
−(
ˆ
R
˙
β)β
˙
β + (
ˆ
R
˙
β)(β
˙
β)(
ˆ
R β)
Now, from the setup of the system, we have that velocity and acceleration are perpendicular. Hence,
as previously stated β
˙
β = 0. Thus, noting this results in the last two terms above being zero;
leaving us with the ﬁrst two (which we also clean up):
ξ = (
ˆ
R
˙
β)
2
−(
ˆ
R
˙
β)
2
(
ˆ
R β)
Taking out a common factor:
ξ = (
ˆ
R
˙
β)
2
(1 −
ˆ
R β)
Therefore, the squared expression for the numerator is:
Ξ
2
= (
ˆ
R
˙
β)
2
(
ˆ
R−β)
2
+
˙
β
2
(1 −
ˆ
R β)
2
−2ξ
= (
ˆ
R
˙
β)
2
(
ˆ
R−β)
2
+
˙
β
2
(1 −
ˆ
R β)
2
−2(
ˆ
R
˙
β)
2
(1 −
ˆ
R β)
Again, let us take out a common factor:
Ξ
2
= (
ˆ
R
˙
β)
2
_
(
ˆ
R−β)
2
−2 + 2
ˆ
R β
_
+
˙
β
2
(1 −
ˆ
R β)
2
= (
ˆ
R
˙
β)
2
_
1 −2
ˆ
R β +β
2
−2 + 2
ˆ
R β
_
+
˙
β
2
(1 −
ˆ
R β)
2
= −(
ˆ
R
˙
β)
2
(1 −β
2
) +
˙
β
2
(1 −
ˆ
R β)
2
Now, treating
ˆ
R
˙
β as the projection of
˙
β onto the xaxis, we have that
ˆ
R
˙
β =
˙
β sin θ cos φ, Also,
we have that
ˆ
R β = β cos θ. Hence, the above becomes:
Ξ
2
= −(
˙
β sin θ cos φ)(1 −β
2
) +
˙
β
2
(1 −β cos θ)
2
Also, we see that the denominator term in the radiation expression can be written:
(1 −β
ˆ
R)
5
= (1 −β cos θ)
5
Hence, using all this, our radiation expression, for acceleration being perpendicular to velocity, is:
d
¯
P
dΩ
=
q
2
16π
2
ε
0
c
Ξ
2
(1 −β
ˆ
R)
5
=
q
2
16π
2
ε
0
c
−(
˙
β sin θ cos φ)(1 −β
2
) +
˙
β
2
(1 −β cos θ)
2
(1 −β cos θ)
5
3.3 Radiation 71
Now, to expand out the top, and rearrange:
d
¯
P
dΩ
=
q
2
16π
2
ε
0
c
˙
β
2
(1 −β cos θ)
3
_
1 −
sin
2
θ cos
2
φ
γ
2
(1 −β cos θ)
2
_
Where we have noted that:
γ
2
=
1
1 −β
2
Now, we call this type of radiation synchrotron radiation, and we hence have the ﬁnal expression:
d
¯
P
dΩ
=
q
2
16π
2
ε
0
c
˙
β
2
(1 −β cos θ)
3
_
1 −
sin
2
θ cos
2
φ
γ
2
(1 −β cos θ)
2
_
(3.39)
3.3.4 Radiation: Summary
We have derived the radiated power, per unit solid angle, for the cases where the velocity and
acceleration are parallel and perpendicular.
For the case of parallel, we call the motion colinear acceleration, and usually denote the radiation
d
¯
P
//
. The expression we derived was:
d
¯
P
//
dΩ
=
q
2
16π
2
ε
0
c
_
˙
β
2
sin
2
θ
(1 −β cos θ)
5
_
(3.40)
For the case of perpendicular, we call the radiation synchrotron radiation, denote it dP
⊥
; and the
expression we derived was:
d
¯
P
⊥
dΩ
=
q
2
16π
2
ε
0
c
˙
β
2
(1 −β cos θ)
3
_
1 −
sin
2
θ cos
2
φ
γ
2
(1 −β cos θ)
2
_
(3.41)
3.3.5 Example: Minimum & Maximum Radiation
Let us consider two exercises:
• Let us show where no synchrotron radiation occurs, for φ = 0. Let us also show where
maximum radiation occurs.
• Let us ﬁnd the total radiated power, for both colinear and synchrotron radiation.
No synchrotron radiation Let us set φ = 0 in the expression for
d
¯
P
⊥
dΩ
. This gives:
d
¯
P
⊥
dΩ
=
q
2
16π
2
ε
0
c
˙
β
2
(1 −β cos θ)
3
_
1 −
sin
2
θ
γ
2
(1 −β cos θ)
2
_
If there is to be no radiation, this expression must be set to zero. That is:
q
2
16π
2
ε
0
c
˙
β
2
(1 −β cos θ)
3
_
1 −
sin
2
θ
γ
2
(1 −β cos θ)
2
_
= 0
72 3 RETARDED POTENTIALS & RADIATION
We must then solve this for θ. So, expanding out the (top) brackets, using sin
2
θ = 1 −cos
2
θ, and
the deﬁnition of γ, this becomes:
1
(1 −β cos θ)
5
_
(1 −β cos θ)
2
−(1 −cos
2
θ)(1 −β
2
)
¸
= 0
Thus, expanding out the top bracket again, and cleaning up, gives:
(β −cos θ)
2
(1 −β cos θ)
5
= 0
Hence, we see that we must have cos θ = β. That is, we have derived that there is no synchrotron
radiation, along φ = 0, if θ
min
= cos
−1
β.
To show where power is a maximum, we must diﬀerentiate the d
¯
P
⊥
, with respect to θ, then set to
zero; and solve for θ. We shall not do this here.
Total power radiated Let us ﬁnd the total power radiated, ﬁrst for the colinear case. So, we
must integrate the expression over solid angles. Thus:
_
d
¯
P
//
dΩ
dΩ =
q
2
16π
2
ε
0
c
_
_
˙
β
2
sin
2
θ
(1 −β cos θ)
5
_
dΩ
Remebering that dΩ = sin θdθdφ, this becomes:
¯
P
//
=
q
2
˙
β
2
16π
2
ε
0
c
_ _
sin
2
θ
(1 −β cos θ)
5
sin θdθdφ
That is:
¯
P
//
=
q
2
˙
β
2
16π
2
ε
0
c
2π
_
π
0
sin
3
θ
(1 −β cos θ)
5
dθ
Making the substitution of x = cos θ, the integral bit is:
_
1
−1
1 −x
2
(1 −βx)
5
dx =
4
3
1
(1 −β
2
)
3
Hence:
¯
P
//
=
q
2
˙
β
2
8πε
0
c
4
3
1
(1 −β
2
)
3
And, noting that 1/γ
2
= (1 −β
2
), this is:
¯
P
//
=
q
2
6πε
0
c
˙
β
2
γ
6
And, with reference to (3.11); the Larmor total power radiated, this is:
¯
P
//
=
¯
P
L
γ
6
¯
P
L
≡
q
2
6πε
0
c
˙
β
2
Now, if we do a similar integration, which is more complicated this time, for the synchrotron
radiation expression, we ﬁnd:
¯
P
⊥
=
¯
P
L
γ
4
3.4 Discussion 73
3.3.6 Example: Charged Particle in Circular Orbit
Suppose we have a charged particle, in a circular orbit (radius R). We have an applied magnetic ﬁeld
B which keeps the particles in such an orbit. Then, we have that the acceleration is perpendicular
to the velocity. Hence, we can use the following expression for the total power radiated:
¯
P
⊥
=
q
2
6πε
0
c
˙
β
2
γ
4
The velocity is v = cβ, and is direction at a tangent to the circle. The acceleration is a = c
˙
β, and
is directed from the particle to the centre of the orbit. Thus, using this expression:
¯
P
⊥
=
q
2
a
2
6πε
0
c
3
γ
4
Now, we have the relation a =
v
2
R
, from standard Newtonian mechanics for motion in a circle. And
also, note that v = cβ. Hence:
¯
P
⊥
=
q
2
cβ
4
6πε
0
R
2
γ
4
Now, from Newtons force equalling the Lorentz force:
γma = qvB ⇒ γm
β
2
c
2
R
= qβcB
Noting that everything is at rightangles. This easily gives:
R =
γmcβ
qB
Hence, using this in
¯
P
⊥
:
¯
P
⊥
=
q
4
β
2
B
2
6πε
0
m
2
c
γ
2
Hence, we have derived a formula giving the total power radiated by a charged particle, mass m, in
a circular orbit (governed by the externally applied magnetic ﬁeld B), at speed cβ.
3.4 Discussion
So, in this section, we have considered the relativistic eﬀect of the ﬁnite speed of light. We have
seen that a consequence of light having a ﬁnite speed, is that moving charges radiate. Infact, what
we have also seen, is that only accelerated charges radiate. We saw that this was because stationary
charges do not radiate, and that we can always put a charge moving with constant velocity into an
inertial frame, in which it is at rest.
We then derived the LienardWiechert potentials for point charges in motion, and stated the
associated ﬁelds.
Finally, we derived a general theorem of radiation, and specialised it to the cases of linear and
perpendicular motion; whilst ﬁnding positions of maximum and minimum radiation.
This concludes our discussion on retarded potentials; and we proceed onto relativistic electrody
namics.
74 3 RETARDED POTENTIALS & RADIATION
75
4 Relativistic Electrodynamics
The aim of this section is to show that the theory of electrodynamics is consistent with the special
theory of relativity
Figure 8: The standard setup for the Lorentz boost along the xdirection. It is standard to state
that the origins of the two systems coincide at t = 0, and that the motion of the second frame Σ
is inertial relative to the ﬁrst Σ; where the primed frame is at a speed u = cβ.
Let us state the Galilean transformations:
_
¸
¸
_
¸
¸
_
t = t
x = x
+ut
y = y
z = z
(4.1)
And also the Lorentz transformations:
_
¸
¸
_
¸
¸
_
t = γ(t
+
β
c
x
)
x = γ(x
+βct
)
y = y
z = z
(4.2)
γ ≡
1
_
1 −β
2
β ≡
v
c
(4.3)
These two sets of transformations are for ‘boosts along the xaxis’; and are the set we call the
“inverse transformation”.
We can show that Maxwell’s equations are not invariant under Galilean transformations, but are
invariant under Lorentz transformations. By ‘invariance’, we mean that the equations have the
same form in both the primed & unprimed frame. We could also show that Newton’s equation
F = m¨ a is invariant under Galilean, but not under Lorentz transformation.
The postulates of special relativity:
76 4 RELATIVISTIC ELECTRODYNAMICS
• All laws of nature are independent of the translational motion of the system as a whole;
• Information cannot travel faster than the speed of light.
There are many physical consequences of these postulates, including length contraction, and time
dilation.
4.1 Notation
We may be able to see, that from the Lorentz transformation equations, that the following is
invariant:
c
2
t
2
−x
2
−y
2
−z
2
= c
2
t
2
−x
2
−y
2
−z
2
≡ d
2
That is, the modulus of a 4vector remains unchanged under Lorentz transformation. We shall
denote a (contravariant; we shall come to this later) 4vector by x, as opposed to the 3vectors by
v. So, the above modulus is:
x x = x
x
= d
2
Minkowski space ‘mixes’ the spatial components with a temporal component. Minkowski space is a
space in which 4vectors reside, having their modulus invariant under Lorentz transformation. We
shall demonstrate this below.
Let us write the following
4
(what we will call contravariant) 4vectors:
x = (x
0
, x
1
, x
2
, x
3
) y = (y
0
, y
1
, y
2
, y
3
)
As we shall also see later, these contravariant vectors occupy the vector space. That is, they are
what we usually call ‘vectors’. Also, by way of notation, we can refer to the vector by x, or by
the set of components ¦x
µ
¦. We will usually be sloppy with using this notation, so that writing x
µ
implies the whole vector. However, in calculations, we will usually use x
µ
to refer to that single
component. By writing x
µ
we are not appealing to any basis system, such as Cartesian or spherical
polar.
The scalar product between the two vectors is then written:
x ∗ y =
3
ν,µ=0
x
µ
g
µν
y
ν
(4.4)
Where we have introduced the metric tensor of Minkowski space:
[g
µν
] =
_
_
_
_
1 0 0 0
0 −1 0 0
0 0 −1 0
0 0 0 −1
_
_
_
_
(4.5)
It is important to note that the expression g
µν
refers to the element, not the tensor itself, which
is why we used [g
µν
] above to refer to the whole tensor. Again, we will be sloppy with notation.
4
I have made a fairly comprehensive treatise of index notation & the various types of vectors, as well as preliminary
tensor calculus notes. They may be found on http://myweb.tiscali.co.uk/jonathanp. The most useful document is
“Index Notation”, which does use a diﬀerent metric signature of (−, +, +, +), rather than the (+, −, −, −) used here.
4.1 Notation 77
This may also be written as g
µν
= diag(1, −1, −1, −1). Under the Einstein summation convention
of implied summation over the available coordinate space, the scalar product may then be written:
x ∗ y = x
µ
g
µν
y
ν
Or, the modulus of a 4vector:
d
2
≡ x ∗ x = x
µ
g
µν
x
ν
This way of writing a modulus (or inner product), in terms of a metric, is very useful, as it has
removed the need of how the basis vectors ‘interact’ with each other. Consider the following:
In Cartesian space, we have the metric δ
ij
, which is the identity matrix. We shall write the dot
product between two vectors in Cartesian space:
x y = (x
i
e
i
) (y
j
e
j
) = x
i
y
j
e
i
e
j
Now, we cant go any further without stating how the basis vectors ‘interact’. The Cartesian basis
vector have the following relation:
e
i
e
j
= δ
ij
Then, we may write a Cartesian dotproduct, in terms of this rule:
x y = x
i
y
j
δ
ij
= x
i
y
i
Which is the familiar rule for ﬁnding the dotproduct of a vector. Now, the ﬁrst thing to note is
that we have used lower indices everywhere, and that they are all latin letter (as opposed to greek).
This convention will be used throughout. Second to note, is that instead of writing the vectors
as a sum of components with basis (i.e. x = x
i
e
i
), we could ignore the basis. In this way, the
dotproduct is written simply as:
x y = x
i
y
j
δ
ij
With no mention as to the nature of the basis vectors. Infact, this is technically incorrect: the
fact that we have used the expression δ
ij
means that we have used a preconceived rule for the
basis vectors. Generally, the rule is called a ‘metric’, and is usually written g
ij
. So, the metric for
Cartesian space is g
ij
= δ
ij
. Then, the dotproduct is:
x y = g
ij
x
i
y
j
Then, in writing it like this, we have seemingly generalised to any space, without need for a basis,
or how the basis interact (information of which is actually ‘inside’ the metric). So, in Minkowski
space (as above), the dotproduct of two 4vectors is written in terms of just the components (i.e.
not their basis; we have not considered once what the basis is) and the metric:
x y = g
µν
x
µ
y
ν
Where we have gone back to worrying about upper and lower indices. Let us now consider what
these correspond to.
Now, let us deﬁne a contravariant position vector thus:
x = x
µ
= (ct, x, y, z) (4.6)
78 4 RELATIVISTIC ELECTRODYNAMICS
And let us deﬁne a covariant position vector:
x
d
= x
µ
= (ct, −x, −y, −z) (4.7)
These two objects may appear to be the same, but they are not! What we used to call just a
‘vector’, is actually a contravariant vector. These types of vectors transform in a very diﬀerent way.
Let us make the following identiﬁcation:
x
µ
= (x
0
, x
1
, x
2
, x
3
) = (ct, x, y, z) (4.8)
x
µ
= (x
0
, x
1
, x
2
, x
3
) = (ct, −x, −y, −z) = (x
0
, −x
1
, −x
2
, −x
3
) (4.9)
Now, its worth noting how the metric acts upon co and contravariant vectors. Let us multiply the
metric by a contravariant vector:
g
µν
x
ν
⇒
_
_
_
_
1 0 0 0
0 −1 0 0
0 0 −1 0
0 0 0 −1
_
_
_
_
_
_
_
_
x
0
x
1
x
2
x
3
_
_
_
_
=
_
_
_
_
x
0
−x
1
−x
2
−x
3
_
_
_
_
Now, notice that the expression on the far right is actually how we deﬁned a covariant vector in
(4.9). Hence:
x
µ
= g
µν
x
ν
(4.10)
Being careful with the order of the indices
5
. Note that the operation of the metric lowers the index,
then relabels it. Similarly, we have:
x
µ
= g
µν
x
ν
(4.11)
That is, raising the index, by acting upon a covariant vector with the inverse metric (which we will
come to shortly).
Now, let us refer back to the expression we had for the modulus of a 4vector:
x ∗ x = x
µ
g
µν
x
ν
Now, we have an expression for ‘what happens’ if we multiply a contravariant vector by the metric,
in (4.10). Hence, let us use that:
x
µ
g
µν
x
ν
= x
µ
x
µ
We denote this, in terms of the actual vectors (as we used the ‘∗’ sign for the scalar product of two
contravariant vectors) as:
x x
d
= x
µ
x
µ
And this is the equivalent of the ‘dotproduct we knew before’. This is what we will call “the
interval”. So, for later reference, for inﬁnitesimals:
ds
2
≡ g
µν
dx
µ
dx
ν
5
Again, for explanations of these indices, and ways of manipulating them, I refer you to the afore mentioned “Index
Notation”.
4.1 Notation 79
Consider a “normal” 3vector x. Its length is given by the dotproduct with itself x
2
= x x. Then,
given a vector of inﬁnitesimals in each direction:
dx = (dx, dy, dz)
Its length is obviously:
ds
2
≡ dx dx = dx
2
+dy
2
+dz
2
Hence we see the reasoning behind calling the quantity ds
2
= g
µν
dx
µ
dx
ν
the ‘interval’. We shall
come back to this later, upon discussion of proper time.
The inverse of the metric g
µν
is written g
µν
. The inverse metric tensor has the same elements as
the noninverse metric, but it is not the same tensor. So, we may write that g
µν
= g
µν
, but that
merely means that its elements are the same, not that the tensors are identical. This highlights the
diﬀerence between representing a tensor by a matrix, and the tensor itself: we represent a tensor,
but we do not write the tensor itself. Also notice that g
µν
= g
νµ
: the tensors are symmetric.
We can arrive at an interesting relation by multiplying (4.10) by something that will raise the index
on the LHS:
g
ρµ
x
µ
= g
ρµ
g
µν
x
ν
Now, we can evaluate the LHS, via (4.11), to give:
x
ρ
= g
ρµ
g
µν
x
ν
Now, we must make this consistently true, which leads us to the relation:
g
ρµ
g
µν
= δ
ρ
ν
(4.12)
Which, if we plug it back in, results in:
x
ρ
= δ
ρ
ν
x
ν
= x
ρ
Which is indeed true! Notice, we can think about this in a slightly diﬀerent, but more ‘handwavey’
way: a matrix multiplied by its inverse gives the identity matrix; which is indeed the statement of
(4.12).
We say that multiplication always occurs between a covariant and contravariant vector (or vice
versa). We form the following scalar product between two 4vectors:
x y
d
= x
µ
y
µ
= g
µρ
x
ρ
g
µν
y
ν
= g
µρ
g
µν
x
ρ
y
ν
= δ
ρ
ν
x
ρ
y
ν
= x
ν
y
ν
Where we have just showed that x
µ
y
µ
= x
µ
y
µ
. To do so, we raised/lowered the indices, depending
on their initial location. Then, notice the inversenon multiplication, resulting in the Kronecker
delta, which we then evaluated.
By way of getting ‘used’ to playing around with these objects, consider the following expression:
g
µρ
g
ρλ
g
λν
We notice that the middle and last term are inverse times noninverse, giving:
g
µρ
g
ρλ
g
λν
= g
µρ
δ
ρ
ν
80 4 RELATIVISTIC ELECTRODYNAMICS
And we can then use the Kroneckerdelta:
g
µρ
δ
ρ
ν
= g
µν
Hence, we have shown:
g
µρ
g
ρλ
g
λν
= g
µν
We have introduced the concept of two diﬀerent types of vectors: covariant and contravariant
vectors. These are infact two diﬀerent types of vectors, each occupying their own vector space.
When one has a standard vector space, one can always deﬁne a dual vector space. An example
we may be familiar with, is in quantum mechanics: if we have a state in ‘ketspace’, so that [k)
represents such a vector. We also know that taking the Hermitian conjugate of such a vector gives
the ‘brastate’ vector ¸k[. So here, the vector space is the ketspace, and the dual the braspace.
The act of going from the ketspace to the braspace, via taking the Hermitian conjugate, is akin
to multiplying a contravariant vector by the metric, to get to the covariant space.
So, we have contravariant vectors occupying the vector space, and covariant vectors occupying the
dual vector space. What always used to call vectors, are occupying the vector space, and are thus
contravariant vectors.
As we have seen, the operation of a metric upon a contravariant vector gives covariant vector. So,
we can think of the metric as being a way of transferring between the vector space & the dual space
(and obviously the other way round via the inverse).
And, by way of (brief) introduction to the next section, a Lorentz transformation keeps within
one vector space, and gives the vector referred to a diﬀerent coordinate system. That is, given a
covariant vector, a Lorentz transformation of the covariant vector retains the status of ‘covariant’,
but will give the components of the vector with respect to a diﬀerent frame of reference within the
dual space of covariant vectors.
4.2 Lorentz Transformation
Let us restate the Lorentz transformations, in a slightly rejigged way:
ct
= γ(ct −βx)
x
= γ(x −βct)
y
= y
z
= z
Now, consider relabelling the coordinates thus
6
:
x
0
≡ ct x
1
≡ x x
2
≡ y x
3
≡ z
6
It should be obvious that these are not exponents, and are just indices.
4.2 Lorentz Transformation 81
Then, we can see that we may write the Lorentz transformations, for a boost along the xdirection,
as:
x
0
= γ(x
0
−βx
1
)
x
1
= γ(−βx
0
+x
1
)
x
2
= x
2
x
3
= x
3
Then, we can see that we can group these equations together into a single set of matrix equations:
_
_
_
_
x
0
x
1
x
2
x
3
_
_
_
_
=
_
_
_
_
γ −γβ 0 0
−γβ γ 0 0
0 0 1 0
0 0 0 1
_
_
_
_
_
_
_
_
x
0
x
1
x
2
x
3
_
_
_
_
(4.13)
Multiplication of these matrices reveals that we may indeed represent our system in this way. We
write this as:
x
µ
= Λ
µ
ν
x
ν
(4.14)
Notice that the transformation is symmetric (that is, the transpose of the transformation matrix is
the same as the untransposed version):
Λ
µ
ν
= Λ
ν
µ
Now, the next thing we do is a little more subtle. Consider the x
0
equation:
x
0
= γ(x
0
−βx
1
)
Let us diﬀerentiate this with respect to x
0
, x
1
, x
2
, x
3
separately:
∂x
0
∂x
0
= γ = Λ
0
0
∂x
0
∂x
1
= −γβ = Λ
0
1
∂x
0
∂x
2
=
∂x
0
∂x
3
= 0 = Λ
0
2,3
We see that these are the elements of the ﬁrst row of the transformation matrix. Hence, with a
little thought, we see that the following is true:
Λ
µ
ν
=
∂x
µ
∂x
ν
(4.15)
Putting this into (4.14):
x
µ
=
∂x
µ
∂x
ν
x
ν
(4.16)
This is actually the deﬁnition of a contravariant tensor, of ﬁrst rank. That is, any quantity A
µ
that
transforms via:
A
µ
=
∂x
µ
∂x
ν
A
ν
(4.17)
Is a contravariant tensor, of ﬁrst rank.
82 4 RELATIVISTIC ELECTRODYNAMICS
Let us ﬁnd the inverse transformation. We shall do this by considering the following expression,
the lowering of indices, by the metric, in ‘transformed’ space:
x
µ
= g
µν
x
ν
Then, the far right term we can express in terms of ‘unprimed’ coordinates, via (4.14), giving:
x
µ
= g
µν
Λ
ν
κ
x
κ
We may then express the far right ‘superscripted’ component in terms of a ‘subscripted’ expression,
from (4.11):
x
µ
= g
µν
Λ
ν
κ
x
κ
= g
µν
Λ
ν
κ
g
κρ
x
ρ
Now, a property of transformation matrices is that (we shall justify this in a later section):
g
µν
Λ
ν
κ
g
κρ
= (Λ
−1
)
ρ
µ
(4.18)
Hence, using this:
x
µ
= (Λ
−1
)
ρ
µ
x
ρ
Where we have that:
(Λ
−1
)
ν
µ
=
∂x
ν
∂x
µ
(4.19)
Let us write the transformation and the inverse, next to each other, as we would with matrix
multiplication (note: we expect the identity matrix out):
Λ
µ
ν
(Λ
−1
)
ν
ρ
=
∂x
µ
∂x
ν
∂x
ν
∂x
ρ
=
∂x
µ
∂x
ρ
= δ
µ
ρ
Which indeed conforms to expectation!
In analogue with the contravariant tensor, we have that any covariant tensor of ﬁrst rank transforms
like:
B
µ
=
∂x
ν
∂x
µ
B
ν
(4.20)
4.2.1 Diﬀerentiation
Let us write the 4vector diﬀerential covariant diﬀerential operator:
∂
µ
≡
∂
∂x
µ
=
_
∂
∂x
0
,
∂
∂x
1
,
∂
∂x
2
,
∂
∂x
3
_
=
_
1
c
∂
∂t
,
∂
∂x
,
∂
∂y
,
∂
∂z
_
Also, the contravariant diﬀerential vector:
∂
µ
≡
∂
∂x
µ
=
_
1
c
∂
∂t
, −
∂
∂x
, −
∂
∂y
, −
∂
∂z
_
4.2 Lorentz Transformation 83
That is (confusingly), the operator ∂
µ
is covariant, and ∂
µ
is contravariant. The confusion arises
when we consider that the diﬀerentials themselves have the opposite sign.
It may help to think of the following: consider the gradoperator in Euclidean space:
∇ =
∂
∂x
i
e
i
It is the equivalent of saying that ∇ is a covariant vector, but diﬀerentiates in a contravariant way.
It transforms as a covariant vector.
Notice then, forming the inner product:
∂
µ
∂
µ
=
1
c
2
∂
2
∂t
2
−
∂
2
∂x
2
−
∂
2
∂y
2
−
∂
2
∂z
2
=
1
c
2
∂
2
∂t
2
−∇
2
That is, we can write the D’Alembertian:
2
= ∂
µ
∂
µ
=
1
c
2
∂
2
∂t
2
−∇
2
We shall frequently use the notation:
∇ v =
∂v
i
∂x
i
Where it is understood that i = 1, 2, 3.
We have:
∂
µ
=
_
1
c
∂
∂t
, ∇
_
∂
µ
=
_
1
c
∂
∂t
, −∇
_
(4.21)
4.2.2 Examples
Let us bring together some notation, and use it to show some useful things.
Show that x
µ
x
µ
= x
µ
x
µ
To show this, we will show that the RHS is the same as the LHS. To
do this, let us write down the transformations:
x
µ
= Λ
µ
ν
x
ν
x
µ
= (Λ
−1
)
ρ
µ
x
ρ
Therefore:
x
µ
x
µ
= Λ
µ
ν
x
ν
(Λ
−1
)
ρ
µ
x
ρ
= (Λ
−1
)
ρ
µ
Λ
µ
ν
x
ν
x
ρ
= δ
ρ
ν
x
ν
x
ρ
= x
ν
x
ν
Hence proven.
84 4 RELATIVISTIC ELECTRODYNAMICS
Show that g
µν
g
νρ
= δ
ρ
µ
To show this, we start with x
µ
= g
µν
x
ν
. We then multiply the whole
thing by g
ρµ
:
g
ρµ
x
µ
= g
ρµ
g
µν
x
ν
Then, raise the indices on the LHS:
x
ρ
= g
ρµ
g
µν
x
ν
Therefore:
g
ρµ
g
µν
= δ
ρ
ν
Hence proven.
Show that ∂
µ
is a covariant vector Now, this is a little more tricky. We ﬁrst must write the
deﬁnitions of ∂
µ
and of a covariant vector:
∂
µ
≡
∂
∂x
µ
B
ν
=
∂x
µ
∂x
ν
B
µ
Now, let B
µ
→∂
µ
:
∂x
µ
∂x
ν
∂
µ
=
∂x
µ
∂x
ν
∂
∂x
µ
=
∂
∂x
ν
= ∂
ν
Where we have noticed that the same factors cancel oﬀ. Thus, we have shown that:
∂
ν
=
∂x
µ
∂x
ν
∂
µ
Hence proven.
4.2.3 Tensors
Let us brieﬂy extend the deﬁnition of transformation of tensors.
Consider a contravariant vector. It transforms thus:
A
µ
=
∂x
µ
∂x
ν
A
ν
This is a rank1 contravariant tensor. We can motivate a second rank tensor by considering two
ﬁrst rank contravariant tensors:
C
µ
=
∂x
µ
∂x
ρ
C
ρ
D
ν
=
∂x
ν
∂x
λ
D
λ
Then, let us write them next to each other:
C
µ
D
ν
=
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
C
ρ
D
λ
Then, let us deﬁne A
µν
≡ C
µ
D
ν
; then the above easily becomes the deﬁnition of how rank2
contravariant tensor transforms:
A
µν
=
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
A
ρλ
(4.22)
4.2 Lorentz Transformation 85
To show that a set of quantities T
µν
form a contravariant tensor, it is necessary to show that the
above holds. It is fairly easy to see the generalisation of this transformation rule to a contravariant
tensor of rankn:
A
µ
1
µ
2
µ
3
...µn
=
∂x
µ
1
∂x
ν
1
∂x
µ
2
∂x
ν
2
∂x
µ
3
∂ν
ν
3
. . .
∂x
µn
∂x
νn
A
ν
1
ν
2
ν
3
...νn
Similarly, a covariant tensor of rankn transforms:
B
µ
1
µ
2
µ
3
...µn
=
∂x
ν
1
∂x
µ
1
∂x
ν
2
∂x
µ
2
∂ν
ν
3
∂x
µ
3
. . .
∂x
νn
∂x
µn
B
ν
1
ν
2
ν
3
...νn
We can also talk of ‘mixed rank’ tensors. So, for example, a rank2 contravariant rank1 covariant
tensor would transform like:
T
µν
ρ
=
∂x
µ
∂x
α
∂x
ν
∂x
β
∂x
γ
∂x
ρ
T
αβ
γ
Just as we had the metric acting upon a vector, let us show what happens if the metric acts upon
a contravariant tensor, of second rank:
g
µρ
A
νρ
= A
ν
µ
It drops the index, and relabels it. Notice that we have retained the order of the indices. For a
more complicated mixedrank tensor:
g
µν
A
ργ
νλ
= A
ργµ
λ
g
µν
A
ρνλ
κ
= A
ρ λ
µ κ
It is useful to notice:
∂x
µ
∂x
ν
= δ
µ
ν
∂x
µ
∂x
ν
= δ
µ
ν
And also, with a small amount of thought:
∂x
µ
∂x
ρ
∂x
λ
∂x
µ
= δ
λ
ρ
Although not a precise reason, the above is easily realised if we notice we can cancel out the ∂
µ
factors from the top left & bottom right, leaving something we just saw was a Kroneckerdelta.
So, a tensor is a set of quantities which transform under a given rule. That is, given one set of
quantities, if we know the transformation to some other frame of reference (such as our Lorentz
transformation; which is only one example of many), we know what the set of quantities look like,
relative to that new frame of reference.
4.2.4 Inverse Lorentz Transformation
Let us consider an alternative way to think about & derive the inverse Lorentz transformation.
We saw that the metric acting on a covariant vector gave a contravariant vector:
g
µν
x
ν
= x
µ
We may also operate the metric upon a rank2 covariant tensor:
g
µν
A
νλ
= A
µ
λ
86 4 RELATIVISTIC ELECTRODYNAMICS
Equivalently, we may act the metric upon a mixedrank tensor:
g
µν
A
λ
ν
= A
λµ
Let us then create a composition:
g
νλ
g
µρ
A
λ
ρ
= g
νλ
A
λµ
= A
µ
ν
So, we ﬁrst acted the inner metric (this order is arbitrary), which raised the ρ, relabelling it µ (notice
that the relative order & column is preserved). The outer metric then dropped the λ, relabelling to
ν. So, as one operation:
A
µ
ν
= g
νλ
g
µρ
A
λ
ρ
Now, let us consider that the tensor in question is the Lorentz transformation:
Λ
µ
ν
= g
νλ
g
µρ
Λ
λ
ρ
Now, this, as it stands, is not the form in which we have matrix multiplication. Consider g
µρ
=
(g
ρµ
)
T
(which is infact a fairly pointless operation, since the metric is symmetric. We do it for
completeness). Then, replacing the middle metric above with its transpose:
Λ
µ
ν
= g
νλ
(g
ρµ
)
T
Λ
λ
ρ
Then, we may reorder the expression (as we are at liberty to: these are just numbers!):
Λ
µ
ν
= g
νλ
Λ
λ
ρ
(g
ρµ
)
T
This is now in the form of matrix multiplication:
Λ
= gΛg
T
The prime is to denote that it is in someway diﬀerent; not that it is in the primed frame! Let us
carry out this matrix multiplication; after noting that the transpose of the metric is the same the
untransposed metric:
gΛg
T
=
_
_
_
_
1 0 0 0
0 −1 0 0
0 0 −1 0
0 0 0 −1
_
_
_
_
_
_
_
_
γ −γβ 0 0
−γβ γ 0 0
0 0 1 0
0 0 0 1
_
_
_
_
_
_
_
_
1 0 0 0
0 −1 0 0
0 0 −1 0
0 0 0 −1
_
_
_
_
=
_
_
_
_
γ γβ 0 0
γβ γ 0 0
0 0 1 0
0 0 0 1
_
_
_
_
= Λ
So, upon comparison of this ﬁnal matrix (that which we called Λ
), we see that it is exactly the matrix
containing the elements of the inverse Lorentz tranformation. Hence, we may denote the inverse
Lorentz transformation matrix as Λ
µ
ν
. We must be very careful with the positions of the indices,
as we see that Λ
µ
ν
is the Lorentz transform. Thus, we see the justiﬁcation of (4.18), as promised!
So, common notation, which are completely equivalent, for the inverse Lorentz transformation:
Λ
ν
µ
= (Λ
−1
)
ν
µ
4.3 Lorentz 4Vectors 87
4.3 Lorentz 4Vectors
Now, we have already seen a 4vector, in the contravariant vector x
µ
= (ct, x); where this notation
implies that x = (x, y, z), the standard 3vector. We saw that it had the folowing transformation
properties:
x
µ
x
µ
= x
µ
x
µ
x
µ
= Λ
µ
ν
x
ν
Infact, any general 4vector, with components A
µ
has the same properties:
A
µ
A
µ
= A
µ
A
µ
A
µ
= Λ
µ
ν
A
ν
So that any 4vector (or two 4vectors) are invariant under Lorentz transformation. This is infact
a condition that a set of quantities must fulﬁll, in order to be called 4vectors.
Now, let us deﬁne the contravariant inﬁnitesimal:
dx
µ
= (cdt, dx) = (cdt, dx, dy, dz)
And the corresponding covariant inﬁnitesimal:
dx
µ
= (cdt, −dx)
We then have the line element (which we have already discussed):
ds
2
= g
µν
dx
µ
dx
ν
= dx
µ
dx
µ
And which we have shown to be invariant (and is only valid in Minkowski space; due to the form of
the metric). Let us write it in terms of the components of the corresponding inﬁnitesimal co and
contravariant vectors:
ds
2
= dx
µ
dx
µ
= c
2
dt
2
+ (dx)(−dx) + (dy)(−dy) + (dz)(−dz)
= c
2
dt
2
−dx
2
And therefore:
ds =
_
c
2
dt
2
−dx
2
=
_
dx
µ
dx
µ
4.3.1 Proper Time
Now, let us deﬁne the proper time τ as the line element divided by the speed of light. That is, in
inﬁnitesimal form:
dτ =
ds
c
(4.23)
88 4 RELATIVISTIC ELECTRODYNAMICS
Figure 9: Consider two diﬀerent situations. Consider (a) ﬁrst: a particle in its rest frame carves
out a vertical line in a spacetime diagram: it has motion through time, but not in space. Notice
that we have marked on the interval length ds
1
as the length of the line in total, in a spacetime
diagram; also notice that dt
1
is the amount of time experienced by such a motion. Now consider
the second case (b). The particle is now moving through space as well as time. Notice that it has
a diﬀerent projection onto the t axis. A moving particle experiences less time than a stationary
particle!
That is:
dτ =
1
c
_
dx
µ
dx
µ
=
1
c
_
c
2
dt
2
−dx
2
=
_
dt
2
−
1
c
2
dx
2
= dt
¸
1 −
1
c
2
_
dx
dt
_
2
= dt
_
1 −β
2
Where we have noticed that:
_
dx
dt
_
2
= (v)
2
= v β =
v
c
Also, recall that:
γ ≡
1
_
1 −β
2
Hence, the proper time is:
dτ =
dt
γ
(4.24)
4.3 Lorentz 4Vectors 89
Let us regeneralise this, by taking ourselves all the way back to the metricexpression:
dτ =
1
c
_
g
µν
dx
ν
dx
µ
Now, we can get a handle on what proper time is, noting the expression γdτ = dt. An observer
which is at ‘rest’ experiences less proper time than an observer ‘moving’. That is, proper time not
only takes into account movement in space, but also in time. This is the “normal” timedilation
eﬀect, but derived using a metric and spacetime intervals.
So, if we consider the usual scenario of a twin remaining on the earth while another travels on a
rocket & back again (ignoring any possible acceleration eﬀects). The proper time experienced by
the twin moving through space (i.e. the one on the rocket) is more than the twin that did not
move through space (i.e. the one on the earth). To think about it another way, consider a ball
being thrown, and a ball being rolled along the ﬂoor; and that they both have the same starting
and ending positions. It is classically obvious that they take diﬀerent times to traverse the same
distance, as the thrown ball must travel up as well as down & across; but the rolled ball only across.
So they travel diﬀerent distances, hence diﬀerent travel times. It is this exact same thing, except
the ‘extra distance’ travelled is motion through (improper) time.
If you think about sitting stationary; you are carving a line through the ‘time axis’, but not the
spatial axes. The length of the line that joins two events, whether you have moved through space or
not, gives the proper time. We can also think about this in terms of possesing a restmass energy,
even though you are not ‘moving’.
The line element ds is the interval of spacetime traversed in proper time dτ.
With reference to Fig (9), we see that if the two lines have the same length (i.e. ds
1
and ds
2
have the same length for both stationary and moving frames), their projections onto the taxis are
diﬀerent. The stationary line has a longer projection than the moving line. That the lines are the
same length is the statement that the proper time for the two observers, is the same.
So, consider the usual analogy of an observer on a rocket ship & another observer stationary on the
earth. Consider that both observers age by 5 years. That is, the proper time for both observers
is 5 years. However, their projections onto the taxis is diﬀerent. The stationary observer has a
longer t than the moving observer. Perhaps one can think of it in diﬀerent terms. Consider that
the moving observer starts his journey in the year 2065. The two observers (stationary & moving
in the rocket) both agree that the rocket leaves in the year 2065. Now consider that the rocket is
able to travel at a very large portion of the speed of light. Then, after the observer in the rocket
has aged by (say) 10years, he arrives back to earth. The observer (that was moving) then thinks
that the year is 2075. However, the stationary obsever on the earth thinks that the year is 2080.
This is the oddity of time dilation!
4.3.2 4velocity & 4momentum
Let us consider the 4velocity: u
µ
. We shall deﬁne it thus:
u
µ
≡
dx
µ
dτ
(4.25)
90 4 RELATIVISTIC ELECTRODYNAMICS
Hence, using the fact that dx
µ
= (cdt, x):
u
µ
=
_
c
dt
dτ
,
dx
dτ
_
Now, we see that from (4.24), the ﬁrst component of the 4vector is just u
0
= γc. Also, consider:
dx
dτ
=
dx
dt
dt
dτ
= γ
dx
dt
Therefore:
u
µ
= (cγ, γu) u =
dx
dt
(4.26)
So, let us brieﬂy consider an implication of the 4velocity vector u
µ
= (cγ, γu). For an observer
who is at rest (spatially), he will posess a component which is progressing along the taxis at the
speed of light. That is, if u = 0, then u
µ
= (c, 0) (after noting that γ = 1 if u = 0). Hence, even a
“stationary” particle is moving through the time coordinate.
And, similarly, the 4momentum is:
p
µ
= mu
µ
= (mcγ, γmu)
Now, we also have the relations:
E = γmc
2
p = γmu E
2
= p
2
c
2
+m
2
c
4
(4.27)
Therefore, we write the 4momentum as:
p
µ
= (E/c, p)
Now, as p
µ
is a 4vector, we then have that p
µ
p
µ
= p
µ
p
µ
; an invariant. Now, it is clear that
p
µ
= g
µν
p
ν
, so that:
p
ν
= (E/c, −p)
Therefore:
p
µ
p
µ
=
E
2
c
2
−p
2
p
2
≡ p
2
x
+p
2
y
+p
2
z
But, as we said, this is invariant; and so we call it the invariant mass. Hence:
E
2
c
2
−p
2
=
1
c
2
(E
2
−c
2
p
2
) = m
2
c
2
And therefore, we recover:
E
2
−c
2
p
2
= m
2
c
4
So, let us again consider the implication of p
µ
= (E/c, p). If a particle is at rest (spatially), then it
still posses an inherent energy: its rest energy. The rest energy comes about because the particle
still has motion through the time coordinate.
4.3 Lorentz 4Vectors 91
4.3.3 Electrodynamic 4vectors
Consider the continuity equation:
∂ρ
∂t
+∇ J = 0
It must be valid in all reference frames. Now, let us write the divergence part under the summation
convention:
∂ρ
∂t
+
∂J
i
∂x
i
= 0
Now, if we consider that we had x
µ
= (x
0
, x
1
, x
2
, x
3
) = (ct, x, y, z) as our Cartesian position 4
vector, we may think that we can write the above equation in a more compact form. Recall the
diﬀerential operator:
∂
µ
=
_
∂
∂x
0
,
∂
∂x
1
,
∂
∂x
2
,
∂
∂x
3
_
=
_
1
c
∂
∂t
,
∂
∂x
,
∂
∂y
,
∂
∂z
_
Then, if we deﬁne the 4vector current density thus:
J
µ
= (cρ, J) (4.28)
Then, the continuity equation becomes:
∂
µ
J
µ
=
1
c
∂cρ
∂t
+
∂J
i
∂x
i
= 0
That is, the continuity equation reads:
∂
µ
J
µ
= 0 (4.29)
Let us just check that we recover the continuity equation if we write things like this. So:
∂
µ
J
µ
=
∂J
0
∂x
0
+
∂J
i
∂x
i
=
1
c
∂cρ
∂t
+
∂J
i
∂x
i
=
∂ρ
∂t
+∇ J
= 0
Which is indeed true.
Recall that the the Lorentz gauge was:
1
c
2
∂φ
∂t
+∇ A = 0
Then, by analogy with the continuity equation, we see that we can write the 4vector potential:
A
µ
= (φ/c, A) (4.30)
92 4 RELATIVISTIC ELECTRODYNAMICS
And that the Lorentz gauge can be written:
∂
µ
A
µ
=
∂A
0
∂x
0
+
∂A
i
∂x
i
=
1
c
∂φ/c
∂t
+
∂A
i
∂x
i
=
1
c
2
∂φ
∂t
+
∂A
i
∂x
i
= 0
Therefore, we can write the 4vector potential in a manifestly Lorentez invariant form:
∂
µ
A
µ
= 0 (4.31)
Notice, by writing J
µ
= (cρ, J) and A
µ
= (φ/c, A), we have implicitly used the notation that J
is the standard 3vector current density, having components J = (J
x
, J
y
, J
z
); and similiarly for the
3vector vectorpotential A. Also, notice that J
0
= cρ, so that even for a charge ‘at rest’, there
exists a charge density. That is, if you look at an electron that appears to be at rest, it is still
carving out a line through the ‘time dimension’, which corresponds to the standard charge density
ρ. When it is moving relative to an observer, it then carves out a line in both space and time, and
in the process bringing the current density components into play. There is obviously the exact same
situation with A
0
; a charge at rest still generates a potential ﬁeld (as we have seen in electrostatics);
which we now see as being due to it carving out a line in time. But when it moves, it also produces
a vector potential ﬁeld.
Under the Lorentz gauge, we were able to derive the following wave equations:
1
c
2
∂
2
φ
∂t
2
−∇
2
φ = µ
0
c
2
ρ
1
c
2
∂
2
A
∂t
2
−∇
2
A = µ
0
J
Hence, we see that we can write these as (which we then verify):
∂
ν
∂
ν
A
µ
= µ
0
J
µ
(4.32)
So, that is:
_
1
c
2
∂
2
∂t
2
−
∂
2
∂x
2
i
_
A
µ
= µ
0
J
µ
Let us pick the component µ = 0: A
0
= φ/c and J
0
= cρ. Then:
_
1
c
2
∂
2
∂t
2
−
∂
2
∂x
2
i
_
φ
c
= µ
0
cρ
⇒
_
1
c
2
∂
2
∂t
2
−
∂
2
∂x
2
i
_
φ = µ
0
c
2
ρ
Which is indeed the equation corresponding to a wave driven by ρ. Let us pick the component
µ = j (just to be clear, the j
th
component of the vector potential A). Then, A
j
= A
j
and J
j
= J
j
(this seems to be odd notation, but it is ok, if one follows the meaning). Then:
_
1
c
2
∂
2
∂t
2
−
∂
2
∂x
2
i
_
A
j
= µ
0
cJ
j
4.4 Electromagnetic Field Tensor 93
Which is immediately satisﬁed. Notice that the corresponding wave equation, under summation
convention:
1
c
2
∂
2
A
∂t
2
−∇
2
A = µ
0
J
1
c
2
∂
2
A
j
∂t
2
−
∂
2
∂x
2
i
A
j
= µ
0
J
j
Let us consider an example:
If ρ
0
is the rest charge density in its rest frame, with J = 0, let us ﬁnd the 4vector current density
in the stationary frame. That is, consider standing on the earth, with a chunk of charge moving,
with respect you to as a stationary observer. That is, we have the ‘primed’ components, and let us
ﬁnd the ‘unprimed’ components of the 4vector current.
So, let us say that in the primed frame we have J
µ
= (cρ
0
, 0). We know the transformation:
J
µ
= (Λ
−1
)
µ
ν
J
ν
Now, this will actually be easier to see in matrix notation, as we must compute all the components:
_
_
_
_
J
0
J
1
J
2
J
3
_
_
_
_
=
_
_
_
_
γ γβ 0 0
γβ γ 0 0
0 0 1 0
0 0 0 1
_
_
_
_
_
_
_
_
J
0
J
1
J
2
J
3
_
_
_
_
=
_
_
_
_
γ γβ 0 0
γβ γ 0 0
0 0 1 0
0 0 0 1
_
_
_
_
_
_
_
_
cρ
0
0
0
0
_
_
_
_
=
_
_
_
_
γcρ
0
γβcρ
0
0
0
_
_
_
_
That is:
J
0
= γcρ
0
J
1
= γcβρ
0
Recall that:
u
µ
= (cγ, γu)
Then:
J
µ
= (γcρ
0
, γcβρ
0
, 0, 0) = ρ
0
(γc, γu, 0, 0)
So, taking u = (u, 0, 0), we see that:
J
µ
= ρ
0
u
µ
4.4 Electromagnetic Field Tensor
Recall the following expressions of the electric & magnetic ﬁelds, in terms of the vectors & scalar
potential:
E = −∇φ −
∂A
∂t
B = ∇A
94 4 RELATIVISTIC ELECTRODYNAMICS
Now, let us write the full equation out for the electric ﬁeld:
(E
1
, E
2
, E
3
) =
_
−
∂φ
∂x
, −
∂φ
∂y
, −
∂φ
∂z
_
−
_
∂A
1
∂t
,
∂A
2
∂t
,
∂A
3
∂t
_
Picking out the ‘1’ component, say; and writing x = x
1
, y = x
2
, z = x
3
:
E
1
= −
∂φ
∂x
1
−
∂A
1
∂t
Now, for generality, suppose we chose component ‘i’, rather than ‘1’:
E
i
= −
∂φ
∂x
i
−
∂A
i
∂t
Now, recall our previous deﬁnitions:
A
µ
= (φ/c, A
1
, A
2
, A
3
) x
µ
= (ct, x
1
, x
2
, x
3
) x
µ
= (ct, −x
1
, −x
2
, −x
3
)
Then, instead of writing φ, let us write cA
0
. And, instead of having t, let us have cx
0
. Thus:
E
i
= −c
∂A
0
∂x
i
−c
∂A
i
∂x
0
And, remembering that x
i
= −x
i
, and also that x
0
= x
0
:
E
i
c
=
∂A
0
∂x
i
−
∂A
i
∂x
0
Multiplying by 1:
−
E
i
c
=
∂A
i
∂x
0
−
∂A
0
∂x
i
≡ F
0i
That is, we have deﬁned an element of the electromagnetic ﬁeld tensor. Let us continue with the
ﬁelds. Let us write out the magnetic ﬁeld:
B =
¸
¸
¸
¸
¸
¸
e
1
e
2
e
3
∂
∂x
∂
∂x
2
∂
∂x
3
A
1
A
2
A
3
¸
¸
¸
¸
¸
¸
So that, picking the component B
1
:
B
1
=
∂A
3
∂x
2
−
∂A
2
∂x
3
=
∂A
2
∂x
3
−
∂A
3
∂x
2
≡ F
32
Similarly:
B
2
=
∂A
3
∂x
1
−
∂A
1
∂x
3
≡ F
13
Inspection of the order of the indices will show how we have arrived at the labelling we have. Also
notice that it is all consistent. So that:
F
µν
=
∂A
ν
∂x
µ
−
∂A
µ
∂x
ν
= ∂
µ
A
ν
−∂
ν
A
µ
(4.33)
4.4 Electromagnetic Field Tensor 95
Also notice that we can immediately see that it is antisymmetric:
F
µν
= −F
νµ
So, grouping together the terms we have computed, and those we antipicate:
F
µν
=
_
_
_
_
0 −E
1
/c −E
2
/c −E
3
/c
E
1
/c 0 −B
3
B
2
E
2
/c B
3
0 −B
1
E
3
/c −B
2
B
1
0
_
_
_
_
(4.34)
Hence, we have the electromagnetic ﬁeld tensor.
As a little aside on antisymmetry & tensors: if the antisymmetry of a tensor is deﬁned as A
µν
=
−A
νµ
then the components must conform to this. That is, if we represent it as a matrix, reﬂecting
the elements along the diagonal should ﬂip the sign of all components. Now, suppose the tensor
had elements along the diagonal. Then, they would not be aﬀected by the ﬂipping, nor the sign
ﬂipping. That is, A
µµ
= A
µµ
, and then we have some components which do not conform to the
deﬁnition. Therefore, we see that an antisymmetric tensor must have only zero diagonal entries.
Which is something we see from the ﬁeld tensor.
Now, we have not yet proved that the ﬁeld tensor does indeed transforms as a tensor. Let us do
so. So, we need to prove that the following holds:
F
µν
=
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
F
ρλ
As it is supposed to be a contravariant tensor, of second rank. Now, let us prove this by trying to
show that the following is true:
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
_
∂
ρ
A
λ
−∂
λ
A
ρ
_
= ∂
µ
A
ν
−∂
ν
A
µ
So, let us consider:
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
∂
ρ
A
λ
Now, we know that each of the following holds (using the inverse transformation):
∂
ρ
=
∂x
ρ
∂x
κ
∂
κ
A
λ
=
∂x
λ
∂x
σ
A
σ
Thus, using these:
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
∂
ρ
A
λ
=
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
∂x
ρ
∂x
κ
∂
κ
∂x
λ
∂x
σ
A
σ
Tidying up a little:
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
∂x
ρ
∂x
κ
∂x
λ
∂x
σ
∂
κ
A
σ
Now, we proceed by trying to make things into Kroneckerdeltas. We can use the rule that
∂x
ρ
∂x
ρ
= 1,
which is pretty obvious; but we use it over other diﬀerentials. We then notice that the ﬁrst and
third; and second and fourth term can be written like this. Thus, giving:
∂x
µ
∂x
κ
∂x
ν
∂x
σ
∂
κ
A
σ
96 4 RELATIVISTIC ELECTRODYNAMICS
However, each of the diﬀerentials left over are just Kroneckerdeltas:
δ
µ
κ
δ
ν
σ
∂
κ
A
σ
= ∂
µ
A
ν
And therefore, what we have shown is that:
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
∂
ρ
A
λ
= ∂
µ
A
ν
Conﬁrming the status of ∂
µ
A
ν
as a contravariant tensor of second rank. And although not a proof,
it is not a stretch of the mind to say that ∂
ν
A
µ
also transforms as a second rank contravariant
tensor. And therefore, F
µν
is a contravariant tensor, of second rank:
F
µν
=
∂x
µ
∂x
ρ
∂x
ν
∂x
λ
F
ρλ
(4.35)
This is, by deﬁnition of our Lorentz transforms:
F
µν
= Λ
µ
ρ
Λ
ν
λ
F
ρλ
Infact, this is a semitrivial proof, if we consider how F
µν
is constructed, and the construction of
(4.22).
4.4.1 Maxwell’s Equations from F
µν
Now, let us start by stating the following two equations, and then we shall recover Maxwell’s
equations from them:
∂
µ
F
µν
= µ
0
J
ν
(4.36)
∂
µ
F
νλ
+∂
ν
F
λµ
+∂
λ
F
µν
= 0 (4.37)
Firstly, notice that in the last equation, all indices appear in the same order: µ, ν, λ. The subsequent
expressions are then just even permutations of this combination.
Now, let us consider the ﬁrst equation:
∂
µ
F
µν
= µ
0
J
ν
(4.38)
Now, let us consider the case ν = 1:
∂
µ
F
µ1
= µ
0
J
1
Then, the LHS expression unpacks slightly to:
3
µ=0
∂
µ
F
µ1
= ∂
0
F
01
+∂
1
F
11
+∂
2
F
21
+∂
3
F
31
Now, we know all of the elements of this:
∂
0
=
1
c
∂
∂t
∂
1
=
∂
∂x
∂
2
=
∂
∂y
∂
3
=
∂
∂z
4.4 Electromagnetic Field Tensor 97
And, looking at the ﬁeld tensor elements:
F
01
= −E
1
/c F
11
= 0 F
21
= B
3
F
31
= −B
2
And therefore, we have:
3
µ=0
∂
µ
F
µ1
= −
1
c
2
∂E
1
∂t
+
∂B
3
∂y
−
∂B
2
∂z
And considering the whole equation together:
−
1
c
2
∂E
1
∂t
+
∂B
3
∂y
−
∂B
2
∂z
= µ
0
J
1
Shuﬄing around a bit:
∂B
3
∂y
−
∂B
2
∂z
= µ
0
J
1
+
1
c
2
∂E
1
∂t
Now, consider the ﬁrst component of the following crossproduct:
(∇B)
x
=
¸
¸
¸
¸
¸
¸
e
1
e
2
e
3
∂
∂x
∂
∂y
∂
∂z
B
1
B
2
B
3
¸
¸
¸
¸
¸
¸
x
=
∂B
3
∂y
−
∂B
2
∂z
Also, let us write Maxwell’s fourth equation (notice that it is a vector equation, and is actually a
set of 3 equations):
∇B = µ
0
J +
1
c
2
∂E
∂t
We see that taking ν = 1 in ∂
µ
F
µν
= µ
0
J
ν
, we found the ﬁrst component of the Maxwell equation.
It is not too hard to see that ν = 1, 2, 3 will produce the whole set of equations that gives Maxwells
fourth equation.
Then, what about ν = 0:
3
µ=0
∂
µ
F
µ0
= µ
0
J
0
We see that we will need the elements of the ﬁeld tensor which are down the ﬁrst column:
F
00
= 0 F
10
= E
1
/c F
20
= E
2
/c F
30
= E
3
/c
Thus (noting that J
0
= cρ):
3
µ=0
∂
µ
F
µ0
=
1
c
_
∂E
1
∂x
+
∂E
2
∂y
+
∂E
3
∂z
_
= µ
0
cρ
Which is completely equivalent to:
∇ E =
ρ
ε
0
Therefore, in total, we have seen that taking ν = 0 in ∂
µ
F
µν
= µ
0
J
ν
we are able to recover Gauss’
law, and taking ν = 1, 2, 3, we recover Amperes law.
98 4 RELATIVISTIC ELECTRODYNAMICS
Let us consider the second equation which we said represented Maxwells equations:
∂
µ
F
νλ
+∂
ν
F
λµ
+∂
λ
F
µν
= 0
We are trying to extract the following set of four equations:
∇E = −
∂B
∂t
∇ B = 0
Now, it is slightly easier to see whats going on if we write the following:
T
µνλ
≡ ∂
µ
F
νλ
+∂
ν
F
λµ
+∂
λ
F
µν
= 0
It is fairly easy to ﬁnd that T
023
, T
013
and T
012
give the three components to Faradays law; and that
T
123
gives the lack of magnetic monopoles equation. Notice, we could have reduced this equation
to T
µνλ
= 0, but for it to make sense, we would have to deﬁne it anyway, so we leave it in the above
form. It is possible to show that T
µνλ
transforms as a contravariant tensor of third rank; and is
proven in the appendix.
4.5 Lorentz Transformations of the Fields
Now, we have that the electromagnetic ﬁeld tensor transforms thus:
F
µν
= Λ
µ
ρ
Λ
ν
λ
F
ρλ
If we actually wish to compute the elements of the ﬁeld tensor, in the primed frame, in terms of
components in the unprimed frame, we can either go through many many summations, evaluating
the above explicitly, or, compute it via matrices. Notice that the above transformation may be
written:
F
= ΛFΛ
T
We must write it like this to allow us to use matrix multiplication. We can immediately start to do
this:
F
=
_
_
_
_
γ −γβ 0 0
−γβ γ 0 0
0 0 1 0
0 0 0 1
_
_
_
_
_
_
_
_
0 −E
1
/c −E
2
/c −E
3
/c
E
1
/c 0 −B
3
B
2
E
2
/c B
3
0 −B
1
E
3
/c −B
2
B
1
0
_
_
_
_
_
_
_
_
γ −γβ 0 0
−γβ γ 0 0
0 0 1 0
0 0 0 1
_
_
_
_
Doing the matrix multiplication of the two far right matrices:
F
=
_
_
_
_
γ −γβ 0 0
−γβ γ 0 0
0 0 1 0
0 0 0 1
_
_
_
_
_
_
_
_
γβE
1
/c −γE
1
/c −E
2
/c −E
3
/c
γE
1
/c −γβE
1
/c −B
3
B
2
γE
2
/c −γβB
3
−γβE
2
/c +γB
3
0 −B
1
γE
3
/c +γβB
2
−γβE
3
/c −γB
2
B
1
0
_
_
_
_
And the ﬁnal multiplications:
F
=
_
_
_
_
0 −E
1
/c −γ(E
2
/c −βB
3
) −γ(E
3
/c +βB
2
)
E
1
/c 0 −γ(B
3
−βE
2
/c) γ(B
2
+βE
3
/c)
γ(E
2
/c −βB
3
) γ(B
3
−βE
2
/c) 0 −B
1
γ(E
3
/c +βB
2
) −γ(B
2
+βE
3
/c) B
1
0
_
_
_
_
4.6 LienardWiechert Fields from Lorentz Transformation 99
Now, the ﬁnal thing to note is that the ﬁeld tensor in the primed frame has the same components
as the unprimed frame, in its frame. That is, we have:
F
=
_
_
_
_
0 −E
1
/c −E
2
/c −E
3
/c
E
1
/c 0 −B
3
B
2
E
2
/c B
3
0 −B
1
E
3
/c −B
2
B
1
0
_
_
_
_
And therefore we can read oﬀ the new components, in terms of the old ones:
E
1
= E
1
B
1
= B
1
E
2
= γ(E
2
−βcB
3
) B
2
= γ(B
2
+βE
3
/c)
E
3
= γ(E
3
+βcB
2
) B
3
= γ(B
3
−βE
2
/c)
Notice: the transformation was for a boost along the xdirection, the ﬁeld having components E
1
, B
1
in that direction. These components are unchanged. It is only the components perpendicular to
the boost direction which are changed.
Now, let us get this into vectorform. This is done mainly by inspection. We use that E
=
(E
1
, E
2
, E
3
); E = (E
1
, E
2
, E
3
). By inspection:
E
= γE + (1 −γ)E
1
ˆ x +γβc(B
2
ˆ z −B
3
ˆ y)
Which is infact:
E
= γE +
1 −γ
v
2
(v E)v +γv B
And, doing the same thing for the magnetic ﬁeld:
B
= γB +
1 −γ
v
2
(v B)v −
γv E
c
2
4.6 LienardWiechert Fields from Lorentz Transformation
Let us consider two frames: Σ & Σ
. At t = t
= 0, the two origins coincide. Now, we consider a
charge q at rest within Σ
. Then, its ﬁelds in its rest frame is just the Coulomb ﬁeld:
E
=
1
4πε
0
q
r
3
r
=
1
4πε
0
q
r
3
(x
, y
, z
) B
= 0
Now, we wish to compute the ﬁelds, from the observation point of Σ. Let us just formulate this in
a slightly less mathematical manner:
Consider that an observer is sat inside a box, with a charge at rest inside the box, next to him. That
is, an observer is in the rest frame of the charge. Then, the electric ﬁeld the observer “observes”
is just the standard Coulomb ﬁeld, with no magnetic ﬁelds. Then consider that another observer
is standing on the surface of the earth, watching the box moving, with the stationary charge inside
the box. The observer who is moving relative to the charge (i.e. the one on the earth) “observes”
a diﬀerent ﬁeld (infact, as we shall see, both electric & magentic) to the observer who is stationary
relative to the charge (i.e. the one inside the box). The problem at hand is to compute the electric
ﬁeld that the stationary observer observes, due to that moving charge.
100 4 RELATIVISTIC ELECTRODYNAMICS
We have seen that we have the following transformations of the ﬁeld components:
E
1
= E
1
B
1
= B
1
E
2
= γ(E
2
−βcB
3
) B
2
= γ(B
2
+βE
3
/c)
E
3
= γ(E
3
+βcB
2
) B
3
= γ(B
3
−βE
2
/c)
So, noting that B
= (B
1
, B
2
, B
3
) = (0, 0, 0) = 0, we are able to solve a little:
B
1
= 0 B
2
= −
β
c
E
3
B
3
=
β
c
E
2
So, using these:
E
1
= E
1
E
2
= γE
2
E
3
= γE
3
And therefore:
E = (E
1
, E
2
, E
3
) = (E
1
, γE
2
, γE
3
)
And, reference to the above Coulomb ﬁeld, for the charge in its rest frame:
E =
1
4πε
0
q
r
3
(x
, γy
, γz
)
Now we wish to express the components of E, in terms of coordinates in Σ. The Lorentz transfor
mations, for the coordinates, at t = 0, are just:
x
= γ(x −vt) = γx y
= y z
= z
And therefore:
E =
1
4πε
0
qγ
r
3
(x, y, z)
Now, let us ﬁnally express the r
in terms of unprimed coordinates. It is evident that:
r
2
= x
2
+y
2
+z
2
And that is, by our Lorentz transformations:
r
2
= γ
2
x
2
+y
2
+z
2
If we use the standard cartesianspherical polars coordinates transformation, except that we measure
axial angles from the xdirection:
x = r cos θ y = r sin θ cos φ z = r sin θ sin φ
Hence:
r
2
= γ
2
r
2
cos
2
θ +r
2
sin
2
θ
Let us write this as:
r
2
= γ
2
r
2
cos
2
θ +r
2
sin
2
θ
= γ
2
r
2
_
cos
2
θ +
1
γ
2
sin
2
θ
_
= γ
2
r
2
_
cos
2
θ + (1 −β
2
) sin
2
θ
_
= γ
2
r
2
(1 −β
2
sin
2
θ)
4.6 LienardWiechert Fields from Lorentz Transformation 101
And therefore, using this:
E =
1
4πε
0
qγ
γ
3
r
3
(1 −β
2
sin
2
θ)
3/2
r
Which is:
E =
1
4πε
0
q(1 −β
2
)
r
3
(1 −β
2
sin
2
θ)
3/2
r
This is the same as the LienardWiechert formula we derived (with considerable vector algebra) for
the electric ﬁeld of a moving point charge, with no acceleration.
102 4 RELATIVISTIC ELECTRODYNAMICS
4.7 Summary of 4Vectors & Transformations
So, let us bring together in one place, with little explanation, most of the transformations & metrics
discussed.
• Contravariant vector  occupies the vector space:
x
µ
= (ct, x, y, z)
• Covariant vector  occupies the dual vector space:
x
µ
= (ct, −x, −y, −z)
• Metric actions  transfer between vector & dual space:
x
µ
= g
µν
x
ν
x
µ
= g
µν
x
ν
g
µν
g
νρ
= δ
µ
ρ
• Lorentz transformation & inverse:
Λ
µ
ν
=
∂x
µ
∂x
ν
(Λ
−1
)
ν
µ
=
∂x
ν
∂x
µ
Λ
ρ
µ
(Λ
−1
)
µ
ν
= δ
ρ
ν
• Contravariant Lorentzian boost  transform between inertial frames:
x
µ
= Λ
µ
ν
x
ν
x
µ
= (Λ
−1
)
µ
ν
x
ν
• Covariant Lorentzian boost:
x
µ
= (Λ
−1
)
ν
µ
x
ν
x
µ
= Λ
ν
µ
x
ν
• Contravariant vector deﬁntion:
A
µ
=
∂x
µ
∂x
ν
A
ν
• Covariant vector deﬁnition:
B
µ
=
∂x
ν
∂x
µ
B
ν
• Contravariant & covariant diﬀerential operators:
∂
µ
≡
_
1
c
∂
∂t
, −∇
_
∂
µ
≡
_
1
c
∂
∂t
, ∇
_
• 4current & 4potential:
J
µ
= (cρ, J) A
µ
= (φ/c, A)
• Lorentz gauge, continuity equation & wave equation:
∂
µ
A
µ
= 0 ∂
µ
J
µ
= 0 ∂
µ
∂
µ
A
ν
= µ
0
J
ν
• Electromagnetic ﬁeld tensor:
F
µν
= ∂
µ
A
ν
−∂
ν
A
µ
• Maxwell’s equations:
∂
µ
F
µν
= µ
0
J
ν
∂
µ
F
νλ
+∂
ν
F
λµ
+∂
λ
F
µν
= 0
4.8 Discussion 103
4.8 Discussion
In this section, we have formulated electrodynamics in a way which is invariant under Lorentz
transformation.
We have introduced index notation, and contra and covariant vectors. We have looked at how to
transform between the two vector spaces via the Minkowski metric, and how to perform a Lorentz
boost, within one space.
We then looked at various 4vectors, and conﬁrmed that they did transform in an expected way. We
then formulated various previously known equations (such as the Lorentz gauge & the continuity
equation) in a way which was Lorentz invariant (that is, had the same form in diﬀerent inertial
frames of reference), which is in concordance with one of the postulates of special relativity.
We were then able to deﬁne the electromagnetic ﬁeld tensor, showing that it is indeed a tensor;
and ﬁnd a way of expressing Maxwell’s four equations in a very elegant way.
Finally, we found the transformations of the ﬁeld components; ﬁnishing with an example which
recovered the LienardWiechert ﬁelds that we derived in the previous section; but with considerably
less algebra!
104 A VECTOR IDENTITIES & TRICKS
A Vector Identities & Tricks
A.1 Vector Identities
Below are some commonly used vector identities:
a (b c) = b (c a)
a b c = (a c)b −(a b)c
∇∇ψ = 0
∇ (∇a) = 0
∇∇a = ∇(∇ a) −∇
2
a
∇ (ψa) = a ∇ψ +ψ∇ a
∇ (a b) = b (∇a) −a (∇b)
∇a b = a(∇ b) −b(∇ a) + (b ∇)a −(a ∇)b
All are easily proven, but some (if not all) are pretty tedious to actually show!
We have also made use of:
∇
ˆ r
r
2
= 4πδ(r)
∇
2
1
r
= −4πδ(r)
A.1.1 Curl in Spherical & Cylindrical Polars
Suppose we have V (r, θ, φ), so that its components are V
r
, V
θ
, V
φ
, then, its curl, in spherical polars,
is given by:
∇V =
¸
¸
¸
¸
¸
¸
e
r
/r
2
sin θ e
θ
/r sin θ e
φ
/r
∂
∂r
∂
∂θ
∂
∂φ
V
r
rV
θ
r sin θV
φ
¸
¸
¸
¸
¸
¸
(A.1)
Similarly, in cylindrical polars:
∇V =
¸
¸
¸
¸
¸
¸
e
r
/r e
φ
e
z
∂
∂r
∂
∂φ
∂
∂z
V
r
rV
φ
V
z
¸
¸
¸
¸
¸
¸
(A.2)
A.2 Useful Tricks
Below are some useful ‘tricks’ to bear in mind when evaluating certain expressions.
Suppose we have the integral:
1 =
_
π
0
cos θ sin θ dθ
A.2 Useful Tricks 105
Then, we can evaluate it via the subsitution:
x ≡ cos θ
Then, the lower limit becomes x = cos 0 = 1, and the upper limit x = cos π = −1. We also trivially
see that dx = −sin θdθ. Hence, let us rewrite the integral:
1 = −
_
−1
1
xdx
But, swopping the limits:
1 =
_
1
−1
xdx
Which results in zero. For a more general case, consider:
1 =
_
π
0
cos
n
θ sin θ dθ
Then, using the same substitution:
1 =
_
1
−1
x
n
dx =
_
0 nodd
2
n+1
neven
So:
_
π
0
cos
n
θ sin θ dθ =
_
0 nodd
2
n+1
neven
Although I have not checked, this will probably only work for n > 0.
106 A VECTOR IDENTITIES & TRICKS
107
B Worked Examples
B.1 Long Beam of Charge  Fields & Force
Consider a long straight beam of electrons, of radius a, and (uniform) line charge density eλ,
travelling along the z axis with velocity. Let us compute the electric and magnetic ﬁelds inside &
outside the beam; and then compute the force the beam feels.
Let us start by considering the electric ﬁeld, for r ≥ a. So, we write Gauss’ law:
_
S
E dS =
1
ε
0
_
V
ρ
V
dV
Which will ﬁnd the electric ﬁeld through a surface S, which encloses charge in a volume V .
Let us initially compute the charge enclosed by the beam, if it has length l. So:
_
ρ
V
dV =
_
ρ
l
A
Adl
Where we have noted that the volume charge density is just the line charge density over the area.
Also, the volume is just the area times length. Thus, we have:
Q
enc
=
_
ρ
l
A
Adl = λel
Hence, using Gauss’ law, where the surface is a cylinder, radius r, where r > a, and length l. Hence,
Gauss’ law results in:
E
r
2πrl =
1
ε
0
λel
That is:
E
r
=
λe
2πε
0
r
r > a
For inside the beam, we must reconsider the charge enclosed by the Gaussian surface. We see that
it is the total charge, multiplied by the area of the surface, and divided by the total area of the
beam. That is:
Q
enc
=
λel
πa
2
πr
2
Thus, the electric ﬁeld inside is given by:
E
r
=
λer
2πε
0
a
2
r < a
Let us now compute the magnetic ﬁeld, using Amperes law:
_
B d = µ
0
_
S
j dS
So, let us compute the magnetic ﬁeld outside. Consider a circle, radius r, which will enclose a
current I. However note that j = vρ. Thus: I = vλeˆ z. Hence:
B
θ
2πr = µ
0
vλe
108 B WORKED EXAMPLES
Hence:
B
θ
=
µ
0
vλe
2πr
r > a
Let us compute the magnetic ﬁeld inside. Let us use the same argument as we did for the charge
enclosed:
I
enc
= µ
0
vλe
πr
2
πa
2
Hence, the magnetic ﬁeld inside is:
B
θ
=
µ
0
vλer
2πa
2
r < a
Now, let us compute the force on the beam. We write the Lorenz force law:
F = q(E +v B)
Let us compute the force inside the beam. So, the cross product:
v B =
¸
¸
¸
¸
¸
¸
ˆ r
ˆ
θ ˆ z
0 0 v
0 B
θ
0
¸
¸
¸
¸
¸
¸
= −ˆ rvB
θ
Hence:
F
r
= λe
_
λer
2πε
0
a
2
−
µ
0
v
2
λer
2πa
2
_
=
λ
2
e
2
r
2πa
2
_
1
ε
0
−µ
0
v
2
_
=
λ
2
e
2
r
2πa
2
ε
0
_
1 −ε
0
µ
0
v
2
_
Now, we also notice that c
2
=
1
ε
0
µ
0
. Hence:
F
r
=
λ
2
e
2
r
2πa
2
ε
0
_
1 −
v
2
c
2
_
Thus, the total force is just:
F
r
=
_
a
0
F
r
dr
Which just results in:
F
r
=
λ
2
e
2
4πε
0
_
1 −
v
2
c
2
_
Now, notice, if v << c, then the force is large, and the beam spreads out. If v ≈ c, then the force
approaches zero; hence the beam retains its size. High energy particle accelerators take advantage
of this when focussing beams.
B.2 Spherical Shell Charge Distribution: σ(θ) = σ
0
cos θ 109
B.2 Spherical Shell Charge Distribution: σ(θ) = σ
0
cos θ
A spherical cavity of radius a is located in a linear dielecric medium, of permittivity ε
r
. On the
surface of the cavity, charge is distributed with a surface density σ(θ) = σ
0
cos θ.
Let us write down the relations between scalar potential, electric & displacement ﬁelds, at the
surface.
Now, the ﬁrst thing to note, is that the ‘cavity’ is more a ‘bubble’; that is, the medium inside is the
same as that outside. Now, at the surface, the electric displacement ﬁeld has the relation:
(D
2
−D
1
) ˆ n = ρ
s
If 1 is inside, and 2 outside, and the unit normal points from the boundary to region ‘2’: outside.
Figure 10: The direction of the normal vector, for shell of distributed charge. Region (1) is inside,
and (2) outside.
So, that is:
D
2,⊥
−D
1,⊥
= σ
0
cos θ
At r = a. Also, as we have that D = ε
r
ε
0
E. Hence:
ε
r
ε
0
E
2,⊥
−ε
r
ε
0
E
1⊥
= σ
0
cos θ
Notice, for arguments sake, that if the two regions had diﬀerent permittivities, then this reads:
ε
2,r
ε
0
E
2,⊥
−ε
1,r
ε
0
E
1⊥
= σ
0
cos θ
Going back to our system:
E
2,⊥
−E
1⊥
=
σ
0
ε
r
ε
0
cos θ
Now, we also have the relation that E = −∇V , where V is the scalar potential. Hence, in the radial
direction (which is the direction in which E
⊥
is in), we thus have:
∂V
2
∂r
−
∂V
1
∂r
= −
σ
0
ε
r
ε
0
cos θ
Have in mind that all the above relations are for r →a.
Now, let us compute the monopole and dipole moments, due to the charge distribution. And let
us do so for the case that the medium is vacuum. That is, χ
E
→0, where ε
r
= 1 +χ
E
. So, we have
that ε
r
= 1.
110 B WORKED EXAMPLES
Now, the charge distribution is given by (the axially symmetric):
ρ(r
) = δ(r
−a)σ(θ) = δ(r
−a)σ
0
cos θ
The potential, due to a multipole expansion, is given by:
V (r) =
1
4πε
0
∞
=0
1
r
+1
_
r
P
(cos γ)ρ(r
)d
3
r
Where γ ≡ θ
−θ. That is, the monopole contribution is just the n = 0 term, and dipole n = 1.
So, if we denote this as:
V (r) =
1
4πε
0
∞
=0
1
r
+1
p
Where:
p
≡
_
r
P
(cos γ)ρ(r
)d
3
r
Where γ ≡ θ
−θ. Note, the volume element is given by:
d
3
r
= r
2
sin θ
dr
dθ
dφ
Then we see that we can compute each ‘pole’contribution separately.
So, total charge (i.e. the monopole moment) is given by:
p
0
≡ q =
_
ρ(r
)d
3
r
That is (noting that P
0
(cos γ) = 1):
p
0
=
_
ρ(r
)r
2
sin θ
dr
dθ
dφ
=
_
2π
φ=0
_
π
θ=0
_
∞
r=0
δ(r
−a)σ
0
cos θ
r
2
sin θ
dr
dθ
dφ
Now, the dr
integral picks out only the point r
= a; due to the δfunction. So:
p
0
= σ
0
a
2
_
2π
φ=0
_
π
θ=0
cos θ
sin θ
dθ
dφ
That is:
p
0
= σ
0
a
2
2π
_
π
θ=0
cos θ
sin θ
dθ
To do this integral, either we notice immediately that the two functions are orthogonal, and hence
the integral is zero; or, we do a substitution:
x = cos θ ⇒ dx = −sin θdθ
Hence, the integral (just the integral, ignoring the constants) is:
I = −
_
−1
1
xdx =
_
1
−1
xdx =
1
2
(1 −1) = 0
B.2 Spherical Shell Charge Distribution: σ(θ) = σ
0
cos θ 111
That is, just zero. Hence, the monopole moment is zero:
p
0
= 0
Now, let us consider the dipole moment. This is given by the integral:
p
1
=
_
r
P
1
(cos γ)ρ(r
)d
3
r
=
_ _ _
r
cos γδ(r
−a)σ
0
cos θ
r
2
sin θ
dr
dθ
dφ
Where we have noted that P
1
(cos γ) = cos γ. Now, we must be careful in continuing. We must
remember that:
cos γ = cos(θ
−θ) = cos θ cos θ
+ sin θ sin θ
Hence, the integral becomes:
p
1
=
_ _ _
r
(cos θ cos θ
+ sin θ sin θ
)δ(r
−a)σ
0
cos θ
r
2
sin θ
dr
dθ
dφ
Let us initially evaluate the r
, φ
integrals; which just give:
p
1
= σ
0
a
3
2π
_
π
0
(cos θ cos θ
+ sin θ sin θ
) cos θ
sin θ
dθ
Notice, we must be careful about what is primed, and unprimed. Of course, this step could have
been done without inserting the expression for cos γ ﬁrst. So, let us split this integral into its two
additive parts:
(i)
_
cos θ cos θ
cos θ
sin θ
dθ
(ii)
_
sin θ sin θ
cos θ
sin θ
dθ
So that p
1
= σ
0
a
3
2π[(i) + (ii)]. Looking at (i) ﬁrst:
_
π
0
cos θ cos θ
cos θ
sin θ
dθ
= cos θ
_
π
0
cos
2
θ
sin θ
dθ
Which we can do via substitution x = cos θ
. Which easily gives:
cos θ
_
−
_
−1
1
x
2
dx
_
= cos θ
__
1
−1
x
2
dx
_
Thus:
(i) →
2
3
cos θ
Now, (ii):
sin θ
_
π
0
sin
2
θ
cos θ
dθ
112 B WORKED EXAMPLES
We substitute x = sin θ
:
→sin θ
__
0
0
x
2
dx
_
= 0
Hence, (ii) →0.
Hence, the dipole moment is:
p
1
= σ
0
a
3
2π
2
3
cos θ
That is:
p
1
=
4
3
σ
0
πa
3
cos θ
Hence, the calculated contributions to the potential, due to the multipole expansion (i.e. that due
to mono and dipole) is:
V (r) =
1
4πε
0
_
1
r
1
p
0
+
1
r
2
p
1
_
And is hence just:
V (r) =
1
4πε
0
1
r
2
4
3
σ
0
a
3
π cos θ
That is:
V (r) =
σ
0
a
3
cos θ
3ε
0
r
2
It is useful just to point out that V (r) actually means that it is a function of all 3 positions; that
is: V (r) = V (r, θ, φ). Thus, we have computed the scalar potential, up to the dipole term only.
Now, let us compute the exact analytical form of the scalar potential V (r), in the case that χ
E
,= 0.
We do so for both the regions inside r ≤ a and outside r ≥ a the ‘cavity’.
Now, we have that the solution to Laplace’s equation, in spherical polars, with axial symmetry, is
of the form:
V (r, θ) =
∞
=0
_
A
r
+
B
r
+1
_
P
(cos θ)
We have the case ‘1’ inside, and ‘2’ outside. Inside, we have that the potential must not diverge;
hence, we reject its Bterm. Similarly, we reject the outsides A term, as that would make the
potentials diverge for r →∞. Hence, we have:
V
1
=
A
r
P
(cos θ) r ≤ a
V
2
=
B
r
+1
P
(cos θ) r ≥ a
Now, to ﬁnd the constants, let us initially apply the boundary condition that the potential is
continuous at the boundary r = a. Hence:
A
a
P
(cos θ) =
B
a
+1
P
(cos θ)
B.2 Spherical Shell Charge Distribution: σ(θ) = σ
0
cos θ 113
Now, to ﬁnd the coeﬃcients, we apply a Fouriertype procedure. Multiply both sides by a ‘diﬀerent’
polynomial, P
(cos θ), say, and integrate. So, the LHS:
_
π
0
A
a
P
(cos θ)P
(cos θ) sin θdθ
Now, due to orthogonality of Legendre polynomials, we know that:
_
1
−1
P
(x)P
(x)dx =
_
0 ,=
2
2+1
=
Hence:
_
π
0
P
(cos θ)P
(cos θ) sin θdθ =
_
0 ,=
2
2p+1
=
From this, we can see that we will just pick out, from the sum over , the term =
:
2
2
+ 1
A
a
=
2
2
+ 1
B
a
+1
That is, reverting back to using , as it is just a dummy index:
A
a
=
B
a
+1
Hence:
B
= A
a
2+1
Now, another boundary condition we have is that of continuity of D
⊥
at the boundary. This, we
showed, is just:
_
∂V
2
∂r
−
∂V
1
∂r
_
r=a
= −
1
ε
r
ε
0
σ
0
cos θ
That is:
∂
∂r
_
B
a
+1
P
(cos θ) −A
a
P
(cos θ)
_
= −
1
ε
r
ε
0
σ
0
cos θ
Giving:
_
+ 1
a
+2
B
P
(cos θ) +A
a
−1
P
(cos θ)
_
=
1
ε
r
ε
0
σ
0
cos θ
Using our derived relation B
= A
a
2+1
, this results in:
_
( + 1)a
2+1
a
+2
A
P
(cos θ) +A
a
−1
P
(cos θ)
_
=
1
ε
r
ε
0
σ
0
cos θ
Which easily gives:
∞
=0
(2 + 1)A
P
(cos θ)a
−1
=
1
ε
r
ε
0
σ
0
cos θ
114 B WORKED EXAMPLES
Now, to ﬁnd the A
, we employ the same technique as we did before; namely to multiply by a
diﬀerent polynomial, and integrate. We shall be a bit more formal in the integration however, in
that we specify the limits as the period of the polynomials. So, the LHS is:
_
π
0
(2 + 1)A
a
−1
P
(cos θ)P
p
(cos θ) sin θ dθ
That is equal to, via orthogonality:
(2p + 1)A
a
−1
δ
2
2 + 1
= 2A
a
−1
The RHS is equal to, when multiplied and integrated:
1
ε
r
ε
0
σ
0
cos θ →
1
ε
r
ε
0
σ
0
_
π
0
cos θP
(cos θ) sin θdθ
Which is equivalent to, putting x = cos θ:
σ
0
ε
r
ε
0
_
1
−1
xP
(x)dx
Hence, we have:
2A
a
−1
=
σ
0
ε
r
ε
0
_
1
−1
xP
(x)dx
Thus, the coeﬃcients are given by:
A
=
σ
0
2a
−1
ε
r
ε
0
_
1
−1
xP
(x)dx
So, considering a few Legendre polynomials, and the coeﬃcents they will generate:
P
0
(x) = 1 ⇒ A
0
= 0
P
1
(x) = x ⇒ A
1
=
σ
0
3ε
r
ε
0
Infact, all higher A
i
are zero (a fact we gloss over here, but will spend more time on in other
examples). Thus, the potential inside:
V
1
=
A
r
P
(cos θ) r ≤ a
Thus:
V
1
=
σ
0
3ε
r
ε
0
r cos θ
And, outside:
V
2
=
B
r
+1
P
(cos θ) r ≥ a
B.2 Spherical Shell Charge Distribution: σ(θ) = σ
0
cos θ 115
Using our derived relation B
= A
a
2+1
, this is:
V
2
=
A
a
2+1
r
+1
P
(cos θ) r ≥ a
Thus:
V
2
=
σ
0
3ε
r
ε
0
a
3
r
2
cos θ r ≥ a
Hence, in summary, we have that the potential in all space is given by:
V (r, θ) =
_
σ
0
3εrε
0
r cos θ r ≤ a
σ
0
3εrε
0
a
3
r
2
cos θ r ≥ a
Note, the electric ﬁeld, inside the cavity is actually constant:
E
r
= −
∂V
1
∂r
= −
σ
0
3ε
r
ε
0
cos θ
That is, the ﬁeld inside the cavity is uniform.
An alternative way to do this, which is actually a bit easier to see, is to express the charge
distribution as a Legendre polynomial. Notice that σ(θ) = σ
0
cos θ = σ
0
P
1
(cos θ). Then, at the
stage of ‘ﬁnding coeﬃcients’, one can immediately see that only the = 1 term will contribute, by
orthogonality of the P
’s. We demonstrate this method in the next example.
B.2.1 Spherical Shell Charge Distribution: σ(θ) = σ
0
cos 2θ
Suppose we have a sphericalshell charge distribution; so that we have a charge density σ(θ) =
σ
0
cos 2θ at a radius a, in a medium with χ
E
= 0. Let us compute the potential, in all space, for
such a system.
Let us use a trig identity, to rewrite the charge distribution:
cos 2θ = 2 cos
2
θ −1
Thus:
σ(θ) = σ
0
(2 cos
2
θ −1)
Now, let us examine this. We note that we can conceive that this is made up of two Legendre
polynomials:
P
0
(cos θ) = 1 P
2
(cos θ) =
1
2
(3 cos
2
θ −1)
So, to express our charge distribution in terms of the above polynomials, let us suppose:
σ(θ) = σ
0
[αP
0
+βP
2
]
Then, equating:
α +
β
2
(3 cos
2
θ −1) = 2 cos
2
θ −1
116 B WORKED EXAMPLES
And, upon examination of the coeﬃcients, we see that:
α = −
1
3
β =
4
3
Hence, we have an expression for our charge distribution, in terms of Legendre polynomials:
σ(θ) =
σ
0
3
[4P
2
(cos θ) −P
0
(cos θ)]
This is a useful way to represent the charge distribution, as it will allow us to evaluate the potential
expansion exactly. Now, let us write the expansion of the potential:
V (r, θ) =
_
A
r
+
B
r
+1
_
P
(cos θ)
If we want the ﬁeld to converge (or at least, no diverge) as r →0, ∞, we must discard the relevant
terms, when evaluating the potential inside and outside the sphere. Thus:
V
1
=
A
r
P
(cos θ) r ≤ a
V
2
=
B
r
+1
P
(cos θ) r ≥ a
Now, we have that the potential is continuous at r = a. Hence:
V
1
(r = a) = V
2
(r = a) ⇒ B
= a
2+1
A
Where we have used a similar argument as we used in the other problem. Now, we have that the
electric ﬁeld is discontinuous at the boundary, according to the surfacecharge distribution. Thus:
_
∂V
2
∂r
−
∂V
1
∂r
_
r=a
= −
1
ε
0
σ(θ)
That is:
(B
(− −1)a
−−2
−A
a
−1
)P
(cos θ) = −
1
ε
0
σ(θ)
Using our derived relation between the inner and outer coeﬃcients, we end up with, after a little
algebra:
A
P
(cos θ)(2 + 1) =
a
1−
ε
0
σ(θ)
Now, let us ﬁnd the coeﬃcients. To do this, we multiply both sides by P
m
(cos θ) sin θ (the reason
for the P
m
is that of orthonormality; and that of the sin θ will become clear when we integrate for
orthonormality), and integrate over the Legendre polynomials period 0 →π. Thus:
A
(2 + 1)
_
π
0
P
(cos θ)P
m
(cos θ) sin θdθ =
a
1−
ε
0
_
π
0
σ(θ)P
m
(cos θ) sin θdθ
Let us consider the LHS. Let x = cos θ ⇒dx = −sin θdθ. Then, also changing the limits, we have:
A
(2 + 1)
_
1
−1
P
(x)P
m
(x)dx
B.2 Spherical Shell Charge Distribution: σ(θ) = σ
0
cos θ 117
This integrates to, taking = m:
A
(2 + 1)
_
1
−1
P
(x)P
m
(x)dx = A
(2 + 1)
2
2 + 1
= 2A
Hence, we have:
A
=
a
1−
2ε
0
_
π
0
σ(θ)P
(cos θ) sin θdθ
Now, let us insert our expression for the charge distribution (and the reason for expressing it so will
become clear):
A
=
a
1−
σ
0
2ε
0
3
_
π
0
[4P
2
(cos θ) −P
0
(cos θ)]P
(cos θ) sin θdθ
Hence, we see that only = 0, 2 will produce nonzero coeﬃcients; due to the orthonormality of the
Legendre polynomials. So, evaluating the = 0 case:
A
0
= −
aσ
0
6ε
0
_
π
0
P
2
0
(cos θ) sin θdθ
= −
aσ
0
6ε
0
2
= −
aσ
0
3ε
0
Where we have used the standard relation:
_
π
0
P
2
(cos θ) sin θdθ =
2
2 + 1
Which is put into the usual form:
_
1
−1
P
2
(x)dx =
2
2 + 1
By using the substitution x = cos θ. Which is the reason for multiplying by sin θ previously. Let us
compute the = 2 case:
A
2
=
σ
0
6aε
0
4
_
π
0
P
2
2
(cos θ) sin θdθ
=
σ
0
6aε
0
4
2
5
=
4σ
0
15aε
0
Now, let us write down the potential inside the shell:
V
1
(r, θ) =
A
r
P
(cos θ)
= A
0
P
0
(cos θ) +A
2
r
2
P
2
(cos θ)
= A
0
+A
2
r
2
P
2
(cos θ)
= −
aσ
0
3ε
0
+
4σ
0
15ε
0
a
r
2
P
2
(cos θ)
= −
σ
0
a
3ε
0
_
1 −
4r
2
5a
2
P
2
(cos θ)
_
r ≤ a
118 B WORKED EXAMPLES
And outside, using our derived relation between coeﬃcients:
V
2
(r, θ) =
a
2+1
r
−−1
A
P
(cos θ)
= −
σ
0
a
3ε
0
_
a
r
−
4a
3
5r
3
P
2
(cos θ)
_
r ≥ a
And therefore, we have, as the exact scalar potential:
V
1
(r, θ) = −
σ
0
a
3ε
0
_
1 −
4r
2
5a
2
P
2
(cos θ)
_
r ≤ a
V
2
(r, θ) = −
σ
0
a
3ε
0
_
a
r
−
4a
3
5r
3
P
2
(cos θ)
_
r ≥ a
B.3 Show That T
µνλ
Is a Tensor
From the relativistic electrodynamics section, we have the following deﬁnitions:
T
µνλ
≡ ∂
µ
F
νλ
+∂
ν
F
λµ
+∂
λ
F
µν
F
µν
≡ ∂
µ
A
ν
−∂
ν
A
µ
And we have that the inverse transformation of a ﬁrst rank contravariant tensor:
B
µ
=
∂x
µ
∂x
ν
B
ν
Now, in the main text, we have shown that the ﬁeld tensor transforms as a second rank contravariant
tensor:
F
αβ
=
∂x
α
∂x
µ
∂x
β
∂x
ν
F
µν
And we have that the 4current density transforms as a ﬁrst rank tensor, as well as the contravariant
derivative operator:
A
µ
=
∂x
µ
∂x
ν
A
ν
∂
µ
=
∂x
µ
∂x
ν
∂
ν
So, let us show that T
µνλ
transforms as a third rank contravariant tensor. That is, we must show
that the following is satisﬁed:
T
µνλ
=
∂x
µ
∂x
σ
∂x
ν
∂x
ρ
∂x
λ
∂x
π
T
σρπ
That is, expanding out the deﬁnition of T
µνλ
:
_
∂
µ
F
νλ
+∂
ν
F
λµ
+∂
λ
F
µν
_
=
∂x
µ
∂x
σ
∂x
ν
∂x
ρ
∂x
λ
∂x
π
(∂
σ
F
ρπ
+∂
ρ
F
πσ
+∂
π
F
σρ
)
Now, let us pick oﬀ the ﬁrst term:
∂x
µ
∂x
σ
∂x
ν
∂x
ρ
∂x
λ
∂x
π
∂
σ
F
ρπ
B.4 Lorentz Force in Covariant Form 119
Let us write the far right expressions in terms of their counterparts in the primed frame:
∂x
µ
∂x
σ
∂x
ν
∂x
ρ
∂x
λ
∂x
π
∂x
σ
∂x
α
∂x
ρ
∂x
β
∂x
π
∂x
γ
∂
α
T
βγ
We notice that most of this collapses into Kronecker deltas:
δ
µ
α
δ
ν
β
δ
λ
γ
∂
α
T
βγ
And that leaves us with:
∂
µ
T
νλ
Which is clearly what we needed to prove; for this ﬁrst term. It is tedious to do the same for the
other two terms, and it is plain that they transform in the same way.
Hence proven.
B.4 Lorentz Force in Covariant Form
We wish to show that the Lorentz force and rate of change of energy of a particle, charge q; induced
by electric & magnetic ﬁelds; can be written as:
dp
µ
dτ
= qF
µν
u
ν
Let us start by discussing the rate of change of energy. Let us denote the energy of the particle by
. Now, the work W done is scalar product of the force and displacement:
W = F d
Then the rate of change of energy is the rate of change of work:
dW
dt
=
d
dt
= F v
Now, the force on a charged particle is given by the Lorentz force law:
F = q(E +v B)
And therefore:
d
dt
= F v
= qv (E +v B)
= qv E +qv (v B)
= qE v
Where we have used the standard vector identity that a (b a) = 0. We shall now start to take a
diﬀerent “tack” on the problem.
120 B WORKED EXAMPLES
Now, let us state the following 4vectors:
u
µ
= (cγ, γu) p
µ
= (/c, p) dτ =
dt
γ
And the ﬁeldtensor:
F
µν
=
_
_
_
_
0 −E
1
/c −E
2
/c −E
3
/c
E
1
/c 0 −B
3
B
2
E
2
/c B
3
0 −B
1
E
3
/c −B
2
B
1
0
_
_
_
_
Now, notice:
dp
µ
dτ
= γ
dp
µ
dt
u
µ
= (cγ, −γu)
Now, let us multilply the ﬁeld tensor by the velocity 4vector:
F
µν
u
ν
=
_
_
_
_
0 −E
1
/c −E
2
/c −E
3
/c
E
1
/c 0 −B
3
B
2
E
2
/c B
3
0 −B
1
E
3
/c −B
2
B
1
0
_
_
_
_
_
_
_
_
cγ
−γu
1
−γu
2
−γu
3
_
_
_
_
=
_
_
_
_
E
1
γu
1
/c +E
2
γu
2
/c +E
3
γu
3
/c
E
1
γ +B
3
γu
2
−B
2
γu
3
E
2
γ +B
2
γu
1
−B
1
γu
2
E
3
γ +B
2
γu
1
−B
1
γu
2
_
_
_
_
Now, let us consider again the original equation:
dp
µ
dτ
= qF
µν
u
ν
That is, the µ
th
component of the LHS is the same as the µ
th
component of the RHS. Let us write
this out:
_
_
_
_
_
γ
c
d
dt
γ
dp
1
dt
γ
dp
2
dt
γ
dp
3
dt
_
_
_
_
_
= q
_
_
_
_
E
1
γu
1
/c +E
2
γu
2
/c +E
3
γu
3
/c
E
1
γ +B
3
γu
2
−B
2
γu
3
E
2
γ +B
2
γu
1
−B
1
γu
2
E
3
γ +B
2
γu
1
−B
1
γu
2
_
_
_
_
That is, equating components:
γ
c
d
dt
= q(E
1
γu
1
/c +E
2
γu
2
/c +E
3
γu
3
/c)
γ
dp
1
dt
= q(E
1
γ +B
3
γu
2
−B
2
γu
3
)
γ
dp
2
dt
= q(E
2
γ +B
2
γu
1
−B
1
γu
2
)
γ
dp
3
dt
= q(E
3
γ +B
2
γu
1
−B
1
γu
2
)
B.4 Lorentz Force in Covariant Form 121
Cancelling oﬀ various factors:
d
dt
= q(E
1
u
1
+E
2
u
2
+E
3
u
3
)
dp
1
dt
= q(E
1
+B
3
u
2
−B
2
u
3
)
dp
2
dt
= q(E
2
+B
2
u
1
−B
1
u
2
)
dp
3
dt
= q(E
3
+B
2
u
1
−B
1
u
2
)
The latter 3 equations are components of a crossproduct; and the ﬁrst of a scalar product. Finally,
notice:
F =
dp
dt
= q(E +v B)
d
dt
= v E
Examination of what we have from unpacking our notation, and what we have above, will reveal
that the following equation has, embedded inside, both the Lorentz force law & the rate of change
of energy of a particle:
dp
µ
dτ
= qF
µν
u
ν
(B.1)
Thanks to J.Agnew for helping out with this derivation!
122 B WORKED EXAMPLES
123
C Legendre Equation & Spherical Harmonics
In spherical polars (i.e. the (r, θ, φ) coordinate system), have have that the Laplace equation
∇
2
Φ = 0 has the form:
1
r
∂
2
∂r
2
(rΦ) +
1
r
2
sin θ
∂
∂θ
sin θ
∂Φ
∂θ
+
1
r
2
sin
2
θ
∂
2
Φ
∂φ
2
= 0 (C.1)
Now, let us assume that we can solve the equation using the standard ‘separation of variables’
technique. Thus:
Φ(r, θ, φ) =
U(r)
r
P(θ)Q(φ)
Where the usual R(r) is obviously just R(r) =
U(r)
r
. Hence, we have that:
PQ
d
2
U
dr
2
+
1
r
2
sin θ
UQ
d
dθ
sin θ
dP
dθ
+
1
r
2
sin
2
θ
UP
d
2
Q
dφ
2
= 0
Let us collect some terms together, and multiply by
r
2
sin
2
θ
RPQ
; giving:
r
2
sin
2
θ
_
1
U
d
2
U
dr
2
+
1
Pr
2
sin θ
d
dθ
sin θ
dP
dθ
_
+
1
Q
d
2
Q
dφ
2
= 0
Now, we can assign constant terms to this equation, so that:
1
Q
d
2
Q
dφ
2
= −m
2
(C.2)
Which, semitrivally, has solutions:
Q
m
(φ) = e
±imφ
(C.3)
So, we have:
r
2
sin
2
θ
_
1
U
d
2
U
dr
2
+
1
Pr
2
sin θ
d
dθ
sin θ
dP
dθ
_
−m
2
= 0
This is just:
r
2
U
d
2
U
dr
2
+
1
P sin θ
d
dθ
sin θ
dP
dθ
−
m
2
sin
2
θ
= 0
Again, let us use another constant term:
r
2
U
d
2
U
dr
2
= ( + 1) (C.4)
Then we have:
( + 1) +
1
P sin θ
d
dθ
sin θ
dP
dθ
−
m
2
sin
2
θ
= 0
That is:
1
sin θ
d
dθ
sin θ
dP
dθ
+
_
( + 1) −
m
2
sin
2
θ
_
P = 0
Now, this is more conventionally written, with x ≡ cos θ, as:
d
dx
_
(1 −x
2
)
dP
dx
_
+
_
( + 1) −
m
2
1 −x
2
_
P = 0 (C.5)
And is known as the generalised Legendre equation, solutions of which are known as the associated
Legendre polynomials.
124 C LEGENDRE EQUATION & SPHERICAL HARMONICS
C.0.1 Radial Solution
Let us quickly derive the solution to the radial part of the Laplace equation. That is, (C.4). So, we
have:
r
2
U
d
2
U
dr
2
= ( + 1)
Let us suppose that the solution is a power law. Let us try U(r) = r
α
. Then:
r
2
r
α
d
2
dr
2
(r
α
) = r
2
r
−α
α(α −1)r
α−1
= α(α −1)
So, we have that:
α(α −1) = ( + 1) ⇒ α
2
−α −
2
− = 0
Now, its not immediately obvious how to solve this; but, if the following expression is written, we
see that it is equivalent:
(α +)[α − −1] = α
2
−α −
2
−
Hence equivalent. Also, by writing in this form, we are able to see its solutions:
(α +)[α − −1] = 0
That is, two solutions:
α = − α = + 1
Hence, we have that the radial solution is of the form (putting in arbitrary constants):
U(r) = Ar
+1
+Br
−
(C.6)
We have also used the form R(r); hence, it is:
R(r) =
U(r)
r
= Ar
+Br
−−1
C.1 Power Series Solution to the Ordinary Legendre Equation
Now, in (C.5), we had a dependance upon m; that is, something that was not azimuthally symmetric.
Let us consider the case where m = 0. Then, the resulting equation is known as the ordinary
Legendre diﬀerential equation:
d
dx
_
(1 −x
2
)
dP
dx
_
+( + 1)P = 0 (C.7)
Now, recall that x = cos θ; where we have that θ ∈ [0, π]; so that the function must be well behaved
on −1 ≤ x ≤ 1. Now, we will assume that the solution takes on a power series form:
P(x) = x
α
∞
j=0
a
j
x
j
= a
j
x
j+α
C.1 Power Series Solution to the Ordinary Legendre Equation 125
So, calculating the necessary expressions (suppressing the summation signs):
dP
dx
= x
α
a
j
jx
j−1
+αx
α−1
a
j
x
j
= a
j
jx
α+j−1
+αa
j
x
α+j−1
= (a
j
j +αa
j
)x
α+j−1
⇒(1 −x
2
)
dP
dx
= (1 −x
2
)(a
j
j +αa
j
)x
α+j−1
= (a
j
j +αa
j
)[x
α+j−1
−x
α+j+1
]
⇒
d
dx
_
(1 −x
2
)
dP
dx
_
= a
j
(j +α)[(α +j −1)x
α+j−2
−(α +j + 1)x
α+j
]
Thus, the total equation is:
a
j
(j +α)[(α +j −1)x
α+j−2
−(α +j + 1)x
α+j
] +( + 1)a
j
x
j+α
= 0
Collecting likepowers of x, and putting summations back in:
∞
j=0
a
j
(j +α)(α +j −1)x
α+j−2
−
∞
j=0
a
j
[(j +α)(α +j + 1) −( + 1)]x
j+α
= 0
Now, we apply a standard ‘trick’ to solving these things further. In the ﬁrst expression, change
summation index from j to m = j −2. Then, the ﬁrst expression becomes:
∞
m=2
a
m+2
(m+ 2 +α)(α +m+ 1)x
α+m
Then, this will give us a recursion relation:
a
j+2
=
(α +j)(α +j + 1) −( + 1)
(α +j + 1)(α +j + 2)
a
j
From this, and the requirement that the series remain ﬁnite, we see that there is a cutoﬀ. We only
have the cases for α = 0, 1 that terminate, and they produce even, odd polynomials. Essentially, it
will give us the Legendre polynomials, of order :
P
)
(x) = 1 P
1
(x) = x P
2
(x) =
1
2
(3x
2
−1) P
3
(x) =
1
2
(5x
3
−3x)
Notice that each polynomial possesses only even or odd power; not both. This is down to the
recursive relation; and that the series must converge for [x[ ≤ 1. We also have that ≥ 0, and
integer. By a manipulation of the power series, we are able to (but is not done here) ﬁnd a generation
function: the Rodrigues’ formula:
P
(x) =
1
2
!
d
dx
(x
2
−1)
(C.8)
126 C LEGENDRE EQUATION & SPHERICAL HARMONICS
C.1.1 Orthogonality of Legendre Polynomials
Now, let us write down the Legendre equation again:
d
dx
_
(1 −x
2
)
dP
(x)
dx
_
+( + 1)P
(x) = 0
Let us multiply the equation by P
(x), and integrate over the interval:
_
1
−1
P
(x)
_
d
dx
_
(1 −x
2
)
dP
(x)
dx
_
+( + 1)P
(x)
_
dx = 0
Multiplying through:
_
1
−1
P
(x)
d
dx
_
(1 −x
2
)
dP
(x)
dx
_
+( + 1)P
(x)P
(x)dx = 0
Now, integrating the ﬁrst expression, by parts:
_
1
−1
P
(x)
d
dx
_
(1 −x
2
)
dP
(x)
dx
_
dx = P
(1 −x
2
)
dP
dx
¸
¸
¸
¸
1
−1
−
_
1
−1
(1 −x
2
)
dP
dx
dP
dx
dx
=
_
1
−1
(x
2
−1)
dP
dx
dP
dx
dx
As we have that the series terminates at ±1. Putting this back in:
_
1
−1
(x
2
−1)
dP
dx
dP
dx
+( + 1)P
(x)P
(x)dx = 0
Now, write this down again, but with →
; then subtract. This is easily shown to be just:
[( + 1) −
(
+ 1)]
_
1
−1
P
(x)P
(x)dx = 0 (C.9)
Which says that if ,=
, then the integral must vanish. However, to calculate the integral for the
=
(i.e. its ‘square’) then we must actually compute the integral, using the explicit form of the
Legendre polynomials; that is, using Rodrigues’ formula. So, for =
, the integral is therefore:
N
≡
_
1
−1
P
2
(x)dx
=
1
2
2
(!)
2
_
1
−1
d
dx
(x
2
−1)
d
dx
(x
2
−1)
dx
After some nontrivial integration (which is not repeated here), we are able to get to the standard
orthogonality condition:
_
1
−1
P
(x)P
(x)dx =
2
2 + 1
δ
(C.10)
C.2 Associated Legendre Polynomials & Spherical Harmonics 127
C.1.2 Using Legendre Polynomials as a Basis
Now, as the Legendre polynomials form a complete set of orthogonal functions, any function on
[−1, 1] can be expressed in terms of them; thus:
f(x) =
∞
=0
a
P
(x)
To ﬁnd the coeﬃcients a
, we apply a standard ‘fouriertype’ technique:
Multiply both sides by another polynomial, and integrate:
_
1
−1
P
(x)f(x)dx =
∞
=0
a
_
P
(x)P
(x)dx
Now, the integral on the RHS is just the orthogonality relation. Thus:
_
1
−1
P
(x)f(x)dx =
a
2
2 + 1
δ
Where the summation will just pick out a single value,
, as being nozero. Hence:
_
1
−1
P
(x)f(x)dx = a
2
2
+ 1
We can, of course, relabel the indices, back to . Hence:
a
=
2 + 1
2
_
1
−1
P
(x)f(x)dx
C.2 Associated Legendre Polynomials & Spherical Harmonics
Here, we consider the solutions to (C.5), for m ,= 0; that is, for problems without azimuthal
symmetry. So, we must solve the equation for arbitrary , m. We basicaly have the generalisation of
the Legendre polynomial P
(cos θ); namely, the associated Legendre polynomial P
m
(x). The version
of Rodrigues formula here is now:
P
m
(x) =
(−1)
m
2
!
(1 −x
2
)
m/2
d
+m
dx
+m
(x
2
−1)
(C.11)
Now, it can be shown that:
P
−m
(x) = (−1)
m
( −m)!
( +m)!
P
m
(x) (C.12)
In a similar fashion as we used previously, we have an orthogonality relation:
_
1
−1
P
m
(x)P
m
(x)dx =
2
2 + 1
( +m)!
( −m)!
δ
(C.13)
128 C LEGENDRE EQUATION & SPHERICAL HARMONICS
Now, recall that we had Q
m
as the azimuthal solution, (C.3). Now, they form a complete set on
φ ∈ [0, 2π]. Thus, as we have that P
m
(cos θ) form a complete set on θ ∈ [0, π], we can imagine that
some combination will give us full solutions, on the unit sphere. We call these solutions spherical
harmonics, and are given by the normalised product:
Y
m
(θ, φ) =
¸
2 + 1
4π
( −m)!
( +m)!
P
m
(cos θ)e
imφ
(C.14)
So, from (C.12), we can see that:
Y
−m
(θ, φ) = (−1)
m
Y
∗
m
(θ, φ) (C.15)
The normalisation/orthogonality relation is:
_
2π
φ=0
_
π
θ=0
Y
∗
m
(θ, φ)Y
m
(θ, φ) sin θdθdφ = δ
δ
mm
(C.16)
Note, for m = 0, we just have:
Y
0
(θ, φ) =
_
2 + 1
4π
P
(cos θ)
An arbitrary function may be expanded in terms of spherical harmonics, thus:
f(θ, φ) =
m
a
m
Y
m
(θ, φ) (C.17)
Where the coeﬃcients are found, in a similar way to before; to be:
a
m
=
_
Y
∗
m
(θ, φ)f(θ, φ) sin θdθdφ
Just as way of convenient notation, we use the solid angle element, dΩ ≡ sin θdθdφ, so that:
a
m
=
_
Y
∗
m
(θ, φ)f(θ, φ)dΩ
C.3 The Addition Theorem for Spherical Harmonics
Consider two coordinate vectors, x, x
with coordinates (r, θ, φ), (r
, θ
, φ
). The addition theorem
states:
P
(cos γ) =
4π
2 + 1
m
Y
∗
m
(θ
, φ
)Y
m
(θ, φ)
Where cos γ = cos θ cos θ
+ sin θ sin θ
cos(φ −φ
).
To prove the theorem, consider that the above Legendre polynomial may be expanded in terms of
spherical harmonics, thus:
P
(cos γ) =
m
A
m
(θ
, φ
)Y
m
(θ, φ)
C.4 Some Spherical Harmonics 129
Figure 11: The addition theorem, for spherical polars. Figure from Jackson
By various symmetry arguements, we see that we have only the case where =
in the sum. Thus:
P
(cos γ) =
m
A
m
(θ
, φ
)Y
m
(θ, φ)
Where the coeﬃcients are given by (found in the usual way):
A
m
(θ
, φ
) =
_
Y
∗
m
(θ, φ)P
(cos γ)dΩ
C.4 Some Spherical Harmonics
Below are some spherical harmonics Y
m
(θ, φ):
= 0
Y
00
=
1
√
4π
(C.18)
= 1
Y
11
= −
_
3
8π
sin θe
iφ
(C.19)
Y
10
=
_
3
4π
cos θ (C.20)
= 2
Y
22
=
1
4
_
15
2π
sin
2
θe
2iφ
(C.21)
Y
21
= −
_
15
8π
sin θ cos θe
iφ
(C.22)
Y
20
=
_
5
4π
(
1
2
cos
2
θ −1) (C.23)
130 C LEGENDRE EQUATION & SPHERICAL HARMONICS
= 3
Y
33
= −
1
4
_
35
4π
sin
3
θe
3iφ
(C.24)
Y
32
=
1
4
_
105
2π
sin
2
θ cos θe
2iφ
(C.25)
Y
31
= −
1
4
_
21
4π
sin θ(5 cos
2
θ −1)e
iφ
(C.26)
Y
30
=
_
7
4π
(
5
2
cos
3
θ −
3
2
cos θ) (C.27)
Note, we have previously written:
Y
−m
= (−1)
m
Y
∗
m
(C.28)
Which gives a way of generating the unwritten negative m harmonics above. Also, remember, for
a given , there are 2 + 1 diﬀerent spherical harmonics:
m = , −1, . . . , 0, . . . , − + 1, −
Also note, that a spherical harmonic, with m = 0 is just the normalisation constant, multiplied by
the Legendre polynomial:
Y
0
(θ, φ) =
_
2 + 1
4π
P
(θ) (C.29)
C.5 Generating Function
For the Legendre polynomials, we suppose that a function will generate the polynomials, then we
will check that it does:
f(x, t) =
∞
n=0
t
n
P
n
(x) (C.30)
=
1
√
1 −2tx +t
2
(C.31)
Now, let us ﬁnd both the x and tderivatives of this expression:
d
dx
:
∞
n=0
t
n
P
n
(x) = t(1 −2tx +t
2
)
−3/2
(C.32)
d
dt
:
∞
n=0
nt
n−1
P
n
(x) = (x −t)(1 −2tx +t
2
)
−3/2
(C.33)
Now, let us express the ﬁrst in the following way:
t(1 −2tx +t
2
)
−3/2
= t(1 −2tx +t
2
)
−1
(1 −2tx +t
2
)
−1/2
= t(1 −2tx +t
2
)
−1
n
t
n
P
n
(x)
C.5 Generating Function 131
Hence:
n
t
n
P
n
(x) = t(1 −2tx +t
2
)
−1
n
t
n
P
n
(x)
That is:
t
n
t
n
P
n
(x) = (1 −2tx +t
2
)
n
t
n
P
n
(x)
Which is just:
n
t
n+1
P
n
(x) =
n
(t
n
−2xt
n+1
+t
n+2
)P
n
(x)
Solve solve this (by equating coeﬃcients of like powers of t), we note the following:
n
(t
n
−2xt
n+1
+t
n+2
)P
n
(x) =
n
t
n
P
n
(x) −2x
n
t
n+1
P
n
(x) +
n
t
n+2
P
n
(x)
Where, in each sum above, n starts from zero. Now, we may relabel each summation variable thus:
⇒ t
n+1
P
n+1
(x) −2xt
n+1
P
n
(x) +t
n+1
P
n−1
(x)
Hence, now we equate powers of t, giving:
P
n
(x) = P
n+1
(x) −2xP
n
(x) +P
n−1
(x) (C.34)
Now, let us go back a bit. If we mutiply (C.32) by (x − t), and (C.33) by t, then they should be
equal. Hence:
n
(x −t)t
n
P
n
(x) =
n
nt
n
P(x)
That is:
xt
n
P
n
−t
n+1
P
n
−nt
n
P
n
= 0
Again, shifting the summation variable, to be able to equate coeﬃcients:
xt
n
P
n
−t
n
P
n−1
−nt
n
P
n
= 0
Hence:
xP
n
−P
n−1
= nP
n
(C.35)
Now, let us write down (C.34) and (C.35), together:
P
n
= P
n+1
−2xP
n
+P
n−1
nP
n
= xP
n
−P
n−1
From these, eliminate P
n−1
, giving:
(n + 1)P
n
= P
n+1
−xP
n
Now, in the above, let n →n + 1:
nP
n−1
= P
n
−xP
n−1
132 C LEGENDRE EQUATION & SPHERICAL HARMONICS
Now, add (C.35) times x to the above; giving:
x
2
P
n
−xP
n−1
+nP
n−1
= nxP
n
+P
n
−xP
n−1
Tidying up:
(1 −x
2
)P
n
= n(P
n−1
−xP
n
)
Diﬀerentiate both sides, with respect to x:
[(1 −x
2
)P
n
]
= n(P
n−1
−P
n
−xP
n
)
Now, using (C.35) again: xP
n
−P
n−1
= nP
n
, on the RHS
[(1 −x
2
)P
n
]
= −n(n + 1)P
n
Hence, we have shown:
[(1 −x
2
)P
m
(x)]
= −m(m+ 1)P
m
(x)
Hence completing the proof.
Let us compute f(1, t):
f(1, t) =
∞
n=0
P
n
(1)t
n
=
1
(1 −2tx +t
2
)
1/2
=
1
1 −t
=
n
t
n
Hence, we see that P
n
(1) = 1.
C.5.1 Application: Expand
1
r
1
−r
2

We often need to expand:
1
[r
1
−r
2
[
Now, we can do so in terms of orthogonal functions, in the angle between the two vectors. Let us
assume r
1
> r
2
. So:
1
[r
1
−r
2
[
=
1
_
r
2
1
+r
2
2
−2r
1
r
2
cos θ
=
1
r
1
1
_
1 + (
r
2
r
1
)
2
−2
r
2
r
1
cos θ
Now, we identify the expression with the generating function of the Legendre polynomials, as:
t =
r
2
r
1
x = cos θ
C.5 Generating Function 133
Hence, we have:
1
[r
1
−r
2
[
=
1
r
1
∞
n=0
_
r
2
r
1
_
n
P
n
(cos θ) (C.36)
For r
2
> r
1
. We can obviously rearrange the indices if the reverse is true.
134 C LEGENDRE EQUATION & SPHERICAL HARMONICS
135
D Books
There are a number of very useful texts, regarding this topic of Electrodynamics. Throughout the
course of writing this, I have consulted many of the below books.
• Heald & Marion: Classical Electromagnetic Radiation;
• Jackson: Classical Electrodynamics;
• Landau & Lifshitz : The Classical Theory of Fields;
• Shutz : A First Course in General Relativity;
• Zwiebach: A First Course in String Theory;
• Woodhouse: Special Relativity.
Heald & Marion is a very good text for many of the electrostatic problems, but is very light on
the relativistic part of the subject. Jackson is a classic text, designed for graduate courses in the
subject, so goes way beyond the scope of the current treatise; regardless, it is very good for the
multipole expansion, and especially in spherical harmonic theory.
The ﬁnal 4 books were consulted mainly for tensor theory. Zwiebach is an excellent start point for
Lorentz transformations & simple tensor manipulations; but is quite light if one is looking for a more
formal introduction. Hence, consulting Schutz is heavily recommended; infact, most introductory
general relativity texts tend to have a large section devoted to tensor theory & very helpful examples
& hints for manipulations. Schutz introduces tensors in a very formal manner, but grounds it ﬁrmly
in physical ideas with plently of examples. Woodhouse is a very mathematical text (whereas the
others are ﬁrmly grounded in physics), introducing special relativity in a very abstract manner.
The, frankly, classical theoretical physics series by Landau & Lifshitz are highly recommended. The
book relevant for electrodynamics is their Vol 2. It has an excellent section on the nonrelativistic
part; going into a lot of detail. The latter half of the book is on general relativity, so is well worth
looking at, as it introduces tensors very well.
ii
CONTENTS
iii
Contents
1 Linear Algebra & Introduction 1.1 1.2 Basis Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rotations & Matrices 1.2.1 1.3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 1 3 3 4 7 7 7 8 10 11 12 15 15 16 17 17 17 18 18 18 19 22 23 25 29 29
Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Brief Note on Important Vector Equations . . . . . . . . . . . . . . . . . . . . . . . . 1.3.1 Important Electrodynamics Equations . . . . . . . . . . . . . . . . . . . . . .
2 Electromagnetic Field Equations 2.1 Maxwell’s Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.1.6 2.1.7 2.1.8 2.1.9 2.2 Electric Field Due to Point and Continuous Charges . . . . . . . . . . . . . . Gauss’ Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Magnetic Vector Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BiotSavart Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Using the BiotSavart Law . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Magnetic Force Between two Parallel Wires . . . . . . . . . . . . . . . . . . . Amperes Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Faradays Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary of Maxwell’s Equations in Vacuum . . . . . . . . . . . . . . . . . .
Maxwell’s Equations in Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 2.2.2 2.2.3 2.2.4 Dielectrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Diamagnetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary of Maxwell’s Equations in Materials . . . . . . . . . . . . . . . . . Fields Across Boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3 2.4
Potentials & Gauge Invariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Diracδ Function & Green Functions . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1 Green Functions: Electrostatics . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5 2.6
Poynting’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Laplace Equation & Its Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.1 Solution to the Laplace Equation: Cartesian . . . . . . . . . . . . . . . . . .
. . . . . .6. . . . . . . . . . . . . . . . .5 2. . . . . . . . . . . . . . . . . . . . .7. . . . . . . . . . . .3. . . . . . . . . . Multipole Expansions Summary . . . . . . . . . . . . . . . . . . . . . . . . . . Example: Charged Particle in Circular Orbit . . . . .
2. . . .11 Discussion . .3. . . . . .iv 2. . . . . .1 2. . . . . . . . . . . . . . . . . . . . .1. . . . Example: Dielectric Sphere in Uniform Eﬁeld .
3.5 Electric Dipole . Spherical Harmonic Expansion Multipoles . . . . . . . . . . . . . . . . . . . . .4 2. . . . . . . . . . . . . . . . . . . . .1 3. . . . . . . . . . . . . . . . . . . . . .1. . . . . . . .3
Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 2. . . . . . . . . .2 2. . . . . . 2. Retarded Potentials & the Wave Equation . . . . . . . . . . . . . . . . . . . . . 3. Example: Particle Moving With Constant Velocity . . 3 Retarded Potentials & Radiation 3. . . . . .6. . . . . . . . . . . . .1 Introduction to Radiation from Accelerated Charges . . . . . . . . . . . . 3. . . . . . . . . . . . . . .3 2. . . . . . . . . . . . . . . . . Radiation: Acceleration & Velocity Perpendicular . . . Solution to the Laplace Equation: Spherical Polars . . . . . .9
Multipole Expansion of the Vector Potential . . . . . . . . . 32 33 35 37 39 39 41 44 46 48 48 50 50 51 53 53 56 57 58 61 62 66 66 68 69 71 71 73 73
Multipoles . . . . . . . . . . . . . . . . . . . .8 2. . . .3 3. .3 2. . . . . . . . . . Example: Minimum & Maximum Radiation . . . . . . .7. . . . . . . . . . . .2. . . . . . . . . . . .3. . . . . . .7. . . .6 General Theory of Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3. . . . . . . . . . . .2 Example: Larmor’s Formula . . . . . . . . .6.10 Method: Potentials and Surface Charges . . . . . . . . . . . . . . . .3. . . . 3. . . . . . . . . . . . . . . . . Properties of Multipoles .2.5 3. . . . . . . . . . .2 3. . .7
CONTENTS Solution to the Laplace Equation: Cylindrical Polars . . . . . . . . .3. . . . . . . . . . .7. . . . . . . . . . .2 3. . . . . . . . . . . . Example: Charge Inside Spherical Cavity . . . . . . . . . . . . . . 2. . . . . . . . . . . . . . . . Relations Between Multipoles in Cartesian & Spherical Polars . . . . Radiation: Summary . . . .4 3. . . . . . . . . . . . . . . . . . . .7. . . . .
LienardWiechert Potentials: Point Charges .1 3. . . . . . . . . . .
. . .3. . . . . . . . . . .6. . . . General Multipole Expansion . . . . . . . . . . . . . . . . .4
Discussion . . .
2. . . . .1 3. . . . . . . . . . . . . . . . . .2 2.2 Features of LienardWiechert Potentials . Radiation: Acceleration & Velocity Parallel . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . .
v 75 76 80 82 83 84 85 87 87 89 91 93 96 98 99
Lorentz 4Vectors . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . .1 4. . . . . . . . . . .3 Proper Time . . . 4. . . . . . . . . .3 Show That T µνλ Is a Tensor . . .1 Maxwell’s Equations from F µν . . . . . . . . . .
Summary of 4Vectors & Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Spherical Shell Charge Distribution: σ(θ) = σ0 cos θ . . . . . . . . . . . . . . .3. . . . . . . . . . . .
4. . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . .1 Long Beam of Charge . . . . . . . . . . . . .CONTENTS 4 Relativistic Electrodynamics 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . . . . . . . . . . . .2 Notation . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . .6 4. . . . . 104 B Worked Examples 107
B. . . . . . . . . . 104 A. . . . . . . Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 B. . . . . . . . .3 4. . . . . . .2.2 4. . . . Inverse Lorentz Transformation . . . . . . . . . . . . .1 4. . . . . . . . . 102 Discussion . 118 B. . . . .4 4. . . . . . . .
4. . . . . . . . . . . . . . . . . . . . . . . . . . 115 B. LienardWiechert Fields from Lorentz Transformation . . . . . . . . . . . . . . . . . Lorentz Transformation . . .1.2. . . . . . . . . . . . . . . . . .1 Curl in Spherical & Cylindrical Polars . . . . . . 104 A. . . . . . . .8
Lorentz Transformations of the Fields .2 4. . . . . . . . . . . . . . . . . . Electrodynamic 4vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 C Legendre Equation & Spherical Harmonics 123
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Lorentz Force in Covariant Form .1 4. . . .3 Diﬀerentiation . . .2. . . . . . . . . . . .2 Useful Tricks . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Vector Identities .5 4. . . . . . . . . . . . Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . .4
Electromagnetic Field Tensor .7 4. . . . . . . . . . . . . 4. . . . . . . . . .Fields & Force . . . . . . .1 Spherical Shell Charge Distribution: σ(θ) = σ0 cos 2θ .2. . . . . . . . 4velocity & 4momentum . . . 103 104
A Vector Identities & Tricks
A. .
.1. . . . . . . . . . . . . . . . . . . . . . . . 128 C.0. . . . . . . . 126 C. . . . .3 The Addition Theorem for Spherical Harmonics . . . . . . . . . . . . . . . . . . . . . . 132 135
D Books
. . . . . . . . . . . . . . . . . . . . . . . .2 Associated Legendre Polynomials & Spherical Harmonics . . . . .1. . . . . . . . . . . . .5 Generating Function . . . . . . . . . . . . . . .4 Some Spherical Harmonics . . . . . . . 129 C.1 Orthogonality of Legendre Polynomials . . . . . . .1 Application: Expand
1 r1 −r2 
. . . . . . . . . . . . . . . . . . 130 C. . . . . . . 127 C. . . . . . . . .1 Power Series Solution to the Ordinary Legendre Equation . . . . . .1 Radial Solution . .5. . 124 C. . . . . . . . . . . . . .2 Using Legendre Polynomials as a Basis . . . . . . . 127 C. . . . .vi
CONTENTS C. . . . . . . . . . 124 C. . . . . . . . . . . . . . . . . . . . . . . . . . .
Hence. The indices refer to individual components. 0) e2 = (0. we are able to write vectors in terms of the basis. 1)
Using this basis. and write the vector x as x.2
Rotations & Matrices
In the subsequent discussion. I will be being sloppy with notation.1
1
Linear Algebra & Introduction
This section just lays some of the groundwork & very brief mathematical framework for some of the things we will be using. 1. they are of length one (normal). x3 ) = xi ei Where the last line has used the Einstein summation convention of implied summations for repeated indices. Now. and are orthogonal (at ‘right angles’) to each other. Let us write down the operation of some matrix M on a vector x. the transpose of the matrix M is:
T Mij = Mji
Notice then. x3 ) x2 = x2 x1 x2 x2 x3 2 x3 x3 x1 x3 x2 x2 3
. irrespective of the basis used. when discussing tensors in relativistic electrodynamics. we have a set of orthonormal basis vectors. We use the equivalent notation: x = x1 e1 + x2 e2 + x3 e3 = (x1 .
1. 0.1
Basis Vectors
Suppose we deﬁne a set of vectors. that if: x1 x = x2 x3 Then: x1 x2 x1 x2 x1 x3 1 xT x = x2 = (x1 . giving us some new vector x : M11 M12 M13 x1 x1 M21 M22 M23 x2 = x2 x3 M31 M32 M33 x3 Mij xj = xi
The ﬁnal line Mij xj = xi is the index notation for matrix multiplication. such that: ei · ej = δij That is. 0.
1. 0) e3 = (0. We will come back to this later. x2 . The cartesian basis vectors are: e1 = (1. x2 .
RT R = 1 (essentially). let us rotate y. R−1 = RT . we also know that in the primed frame. it is a prototype for how a cartesian tensor. In index notation this is:
T Mij = Rik Mkl Rlj = Rik Rjl Mkl
Where in the ﬁrst equality we wrote down the operation. and rearranged. R. this looks like: cos θ sin θ 0 R = − sin θ cos θ 0 0 0 1 The inverse of R obviously has elements ei · ej . orthogonal transformations preserve lengths and orientations of vectors within. transforms. then used the transposed matrix element. to give y: y = Mx Now. Hence. the angle between bases in the primed and unprimed frame. It is important that the distinction between the rotation matrix. if we put this factor in. Infact. transformations can still happen: y =M x So. This sort of matrix is said to be orthogonal.2
1
LINEAR ALGEBRA & INTRODUCTION
If a set of basis vectors ei are rotated to some new orientation ei . suppose we want to ﬁnd the equivalent in the primed frame. Thus. via some rotation matrix: y = Ry Putting our expression for y in: y = R(M x) Now. So. or: xi = Rij xj Where we have that Rij = ei · ej . This is equivalent to x · y = x · y . in a way conducive to matrix multiplication. upon comparison: M = RM RT Hence. nothing is changed: y = R(M RT Rx) Let us collect the terms: y = RM RT (Rx) Now. Hence. Hence: y = RM RT x Now. the eﬀect of x by rotation operator R is Rx = x . The rearranged expression does not represent matrix multiplication. in the invariance of quantities. we have an expression for the operator M in the primed frame. and the operator M is understood. The cosines between the two frames. of second rank. For a rotation about the zaxis. its like saying. Suppose we had that x is operated upon by some operator M . then the components of the vectors in the primed system (new ones) are related to those in the unprimed system (old ones) by x = Rx.
. Thus. This will become very important in our discussion on relativistic electrodynamics. rotate by 90◦ and back again. a rotation followed by its inverse does nothing. which is just the transpose of R.
It will transform via: Ta = Rai Ti A matrix. something that has linear ﬁeld lines.. The tensor T can be thought of as the ‘image’ of T . it must transform like: Ta1 a2 .3
Brief Note on Important Vector Equations Tensors
3
1. The ‘div’ of a vector is a measure of the divergence of a ﬁeld from a source. . hence its divergence is zero: ·( Stokes theorem: × V dS =
S
× v) = 0
V ·d
.. This can be visualised. That is.
1. Ran bn Tb1 b2 . as the curl gives you a ﬁeld which is ‘going round in circles’..an = Ra1 b1 Ra2 b2 . The salient points to take away from this is that a tensor (which is a set of quantities T ) transforms under a given set of rules (where the transformation has been denoted R here). which is a rank2 tensor will then transform via: Tab = Rai Rbj Tij Which is what we saw in the previous section: how an operator M transformed. The divergence of a curl of a vector ﬁeld is zero.1. The div grad = · φ = 2 φ. as well as mathematically proved. The divergence of the magnetic ﬁeld is zero . We shall have a more complete discussion on tensors when we start to use them & their transformation properties. consider a rank1 tensor: a vector. for a general rankn tensor. That is. after the set of transformations given by the rotation matrices R.no magnetic monopoles..bn This will become clearer when we use them. which just stretch oﬀ to inﬁnity has no curl. which will have no ‘radial’ component. In the present discussion. As another example.2. . we have not made a distinction between tensors in diﬀerent spaces (as we shall do later). Notice where the indices are in the rotation matrices Rij . It is easy to see how this generalises to a rankn tensor.3
Brief Note on Important Vector Equations
The ‘curl’ of a vector is the measure of how much it ‘rotates’.1
This is an impressively brief introduction to these objects! A cartesian tensor is a some Tijkl which transforms via: Tabcd = Rai Rbj Rck Rdl Tijkl This is for a rank4 tensor.
enclosing a surface. That is: a·b=b·a And also that crossproducts anticommute: a × b = −b × a 1.1) (1.4
1
LINEAR ALGEBRA & INTRODUCTION
That is. the measure of ﬂux of ﬁeld lines.1 Important Electrodynamics Equations
Most of these will be derived.4)
. and that ˆ r is a unit vector in the direction of r. We will be making extensive use of the following: ˆ r= r r
We shall try to be consistent with the notation that r is the magnitude of the vector r. r = r = x2 + y 2 + z 2 . the measure of how many ﬁeld lines pass through a closed path. but this is just a recap.3) (1.6) ·B =0 × H = Jf +
∂D ∂t
·B =0 × B = µ0 J + µ0 ε0 ∂E ∂t
(1. but one that we use a lot is: ˆ 1 r r =− 2 =− 3 r r r Where r = xi + yj + zk.3.5) (1. Maxwell’s equations: ρ ε0 ∂B ×E =− ∂t ·E = In matter. Divergence theorem: · V dV =
V S
V · dS
That is. or used at some point. It is easy to derive various relations.2)
(1.
It is very useful to note that dotproducts commute. these take on the following form: · D = ρf ∂B ×E =− ∂t Where we have used the following deﬁnitions: D = ε0 E + P 1 H = B−M µ0 (1. through a surface enclosing some volume within which the ﬁeld exists.
7)
Not to be confused with the polarisation vector above!
. and similar for the permeability. in linear media. these simplify to D = εE and H = µ B.1. Where ε ≡ ε0 εr .3
Brief Note on Important Vector Equations
5
1 However.8) (1.
The Lorentz force law: F = q(E + v × B) The Poynting vector: P = 1 (E × B) µ0 (1.
6
1
LINEAR ALGEBRA & INTRODUCTION
.
then. The electric ﬁeld at some position r (that is.
2. for a point charge. the force between the two is given by: F12  = 1 q1 q2 4πε0 d2
1 The factor of 4πε0 is purely due to the SI units system used. and some continuous body of charge. We can calculate the electric ﬁeld at the site of some test charge qtest . divided by the magnitude of their separation. a distance d apart. This is obviously just er−r = separation. This can be done for both a point charge q.7
2
2. due to a point charge q. and the charge of interest is at r .1
Electromagnetic Field Equations
Maxwell’s Equations
Let us start with Coulomb’s law: the force between two point charges. is given by: E(r) = = F qtest 1 q er−r 4πε0 r − r 
r−r r−r  . Hence.1. We consider both below. E(r) = 1 q(r − r ) 4πε0 r − r 3
their
. at the position of the test charge qtest ). Note. q2 . the observation point is at r.
Where er−r is the unit vector between r and r . in both cases. Let us have two charges q1 . Consider the setup in Fig (1)(a). due to some charge q. and will give the ﬁeld in units Vm−1 .1
Electric Field Due to Point and Continuous Charges
Figure 1: Setup for both point charge (a) and some continuous distribution of charge (b). at r .
as in Fig (1)(b).e. we can derive the relation: 1 r−r =− r − r  r − r 3 Hence.1). Thus. from vector calculus. That is. which we used before. So. at the observation point P ) is: E(r) = 1 4πε0 ρ(r ) (r − r ) 3 d r r − r 3 (2. and substitute in: E(r) = − 1 4πε0 = − φ(r) ρ(r ) 1 d3 r r − r 
Where we are now in a position to deﬁne the scalar static potential : φ(r) = Thus: E(r) = − φ(r) Hence. It is important to note (for subsequent discussions) that the expression E = − φ only holds for electrostatic ﬁelds. Hence. we can compare this with (2. We shall come to timevarying ﬁelds later.8
2
ELECTROMAGNETIC FIELD EQUATIONS
Now. Now. by Gauss’ law (which we will come to shortly). 2.1.4)
That is: Poissons equation. but its a useful way to think about it). the electric ﬁeld at the position of the test charge (i.1)
The integral will sweep over the distribution. Let us have a continuous distribution ρ(r ). Thus. must now be substituted for an integral over the volume of the charge distribution.2 Gauss’ Law
Suppose we compute the integral over the (closed) surface containing some volume of charge density: E · dS =
S
1 ε0
ρ(r) d3 r
V
(2.5)
.2)
(2. the total charge is ρ(r ) d3 r . ones for which the charge is at rest.3) ·E =
ρ ε0 .
Thus. we have that above: ρ ·E =− · φ= ε0 Which is just:
2
1 4πε0
ρ(r )
1 d3 r r − r 
(2. using the
φ=−
ρ ε0
(2. suppose we have a continuous distribution of charge. picking up the little contributions from each ‘bit of charge’ (which sounds like a contradiction for a continuous distribution. the above poission equation also only holds for static ﬁelds. our ‘total charge’ q.
Hence. Let us choose our surface (i. Let us compute the electric ﬁeld associated with such a system. with charge per unit length λ. Example Suppose we have an inﬁnite rod.5) is just: 1 ε0 ρ(r) d3 r =
V
1 eλ ε0
Where we have used that the charge per unit length λ will carry a charge of e. A is the area over which we are looking: that we have constructed our cylinder: A = πr2 .2. Hence. we are able to write the LHS surface integral as a volume integral of the divergence of E. The RHS of (2. hence the total charge along the rod is its charge density multiplied by its length . and will hence not contribute to the surface integral. Note. leaving us with Maxwell’s 1st equation. We can then make the two (arbitrary) volumes on either side the same.5) is: E · dS = Er 2πr
S
Where we have noted that all other possible components of the electric ﬁeld will be parallel to the surface.1
Maxwell’s Equations
9
This is known as the integral version of Gauss’ law. radius r enclosing the rod. as they cancel. Let us now consider a rod of radius a.e we are using the integral version of Gauss’ law ) to be a cylinder of length . We now have: Er 2πr = 1 ε0 AρV d
Now. which is just equal to λe the line charge density divided by the area in which the charge resides: ρV = πa2 . the LHS of (2. by equating the integrands: ·E = ρ ε0 (2. which carries a linecharge density λ (the rod is also inﬁnitely thin). equating the two sides: Er 2πr = 1 eλ ε0 eλ 2πε0 r
⇒ Er =
We have hence computed the (radial) electric ﬁeld at a distance r away from an inﬁnite rod. Hence: Er 2πr = Which is just: Er 2πr = 1 r2 λe ε0 a 2 1 ε0 πr2 λe d πa2
. but this isnt too much of a problem. By the divergence theorem.6)
This holds true for time varying and electrostatic ﬁelds. we have assigned a length to an inﬁnite rod. the terms in the integral on the RHS: ρV is the volume charge density.
inserting our expression for the magnetic vector potential: × × A = µ0 J
2
×A
(2. is: Φ=
S
× A · dS
Now.9)
So. Now. which is the statement that under the Coulomb gauge. and use it to derive (by analogy) the BiotSavart law.8)
2φ
Now. 2.1. by using the magnetic vector potential. as the divergence of the curl is zero. Hence. can be written: ( · A) − A = µ0 J · A = 0. which is the charge crossing per unit area per unit time. we can apply Stokes theorem to give us: Φ= A·d (2. (2. 0
This is the vector form of Poissons equation: i.3 Magnetic Vector Potential
Now. we have the interpretation that magnetic ﬂux through an area is the same as the line integral of the mangnetic ﬁeld vector over the line enclosing the surface. again by a standard vector identity. and can be veriﬁed easily. the above becomes:
2
A = −µ0 J = − ερ . another Maxwell equation (only for static ﬁelds) reads: × B = µ0 J Where J is the current density. from the Maxwell equation which states · B = 0 (i. is satisﬁed by standard vector analysis. as was previously stated. we use what we will call the Coulomb gauge.7)
Which.10
2
ELECTROMAGNETIC FIELD EQUATIONS
That is. analogous to Magnetic ﬂux is just the surface integral of the magnetic ﬁeld: Φ=
S
B · dS
Which. That is: B= So that: ·B = · ×A=0 Which.e.
. no magnetic monopoles). we are able (via standard vector analysis) to deﬁne some vector A that B is the curl of. the ﬁeld (which is radial) due to a rod of inﬁnite length. So.e. but ﬁnite radius a is given by: Er (r) = 1 rλe ε0 2πa2 r≤a
We shall now discuss the magnetic vector potential.
suppose that
Figure 2: How things are deﬁned in the BiotSavart law. So now.1. Also. Recall that we could ﬁnd the electric ﬁeld vector from the scalar potential.1
Maxwell’s Equations BiotSavart Law
11
2. by analogy. where the wire has crosssection S (which doesn’t vary). our integral will sweep over all the conributions from each line element to a ﬁeld at some observation point. picking up all its contributions. Thus: B= Again. Now. So. The integral will sweep over the current loop . Hence: J dV = J Sd = ic d Where ic is the charge ﬂow per unit time. so. let us suppose that the vector potential may be found in the same form as the scalar potential: A= µ0 4π J dV r12 
Where r12 ≡ r − r is the vector from the current loop carrying J to the observation point P .4
We shall derive the BiotSavart law by analogy. we have that: µ0 i c d A= 4π r12  Now. we call upon a vector identity: × (av) = a × v + ( a) × v µ0 ic 4π × d r12 
Where we have identiﬁed r12  as the scalar.2. we know that we may ﬁnd B via B = × A. So. let us suppose that we could ﬁnd the magnetic ﬁeld from the magnetic vector potential. we see that dV = Sd . The diﬀerent factor outfront is due to convention in wanting a speciﬁc set of units. d is the lineelement. Hence: B= µ0 ic 4π 1 r12  ×d + 1 r12  ×d
. we consider the loop of current to be a thin wire element.
and the z direction directly out of the page.10)
Which ﬁnds the magnetic ﬁeld at a point. with the r direction pointing radially out from the centre.1. compute B(0). when the current loop is ˆ at r = rr . we have the BiotSavart law: µ0 ic 4π d × r12 r12 3
B=
(2. of radius r. The integral sweeps around the loop. That is. one ﬁnds it to be zero. Let us compute the magnetic ﬁeld at the origin. supose that the current loop is at r and the observer at r (relative to some origin O). then the BiotSavart law reads: µ0 ic d × (r − r ) 4π r − r 3
dB(r) =
2. from a vector going from the current loop to the observation point. We have also shown that: 1 r12  ˆ r12 r12 2
=−
ˆ Where r12 is the unit vector from the current loop to the observation point. We will sometimes use the BiotSavart law in the following form: µ0 ic d × r12 4π r12 3
dB =
To be a little more consistent.12
2
ELECTROMAGNETIC FIELD EQUATIONS
It is pretty trivial to show that the ﬁrst term is zero: if one does the cross product. and that we have a coordinate system centred ˆ ˆ on r = 0. resulting in: µ0 ic 4π µ0 ic 4π ˆ d × r12 r12 2 d × r12 r12 3
B = =
Hence. picking up all contributions as it does. noting that a × b = −b × a.5
Using the BiotSavart Law
Magnetic Field from a Current Loop Suppose we have a current I going round a loop. Suppose that the current is going clockwise. We then swop the order of the crossproducts.
.
1 So:
Maxwell’s Equations
13
µ0 ic d × (r − r ) 4π r − r 3 µ0 I d × (−r ) ⇒ dB(0) = 4π  − r 3 ˆ ⇒ d × (−r ) = d  − r z dB(r) = ˆ = d rz ˆ = d rz µ0 I d r ˆ z ⇒ dB(0) = 4π r3 d µ0 I ˆ ⇒ B(0) = z 4π r2 µ0 I 2πr ˆ = z 4π r2 µ0 I ⇒B = 2r Thus.2. this modiﬁes to: B= N µ0 I 2r
. carrying current I. we have found the magnetic ﬁeld due to a single loop of wire. If the current loop has N (tight) turns. at the origin.
to make this easier to integrate. let us put x in terms of θ. let us put all this back into the BiotSavart expression: dB = = = ˆ µ0 I d × r  4π r2 µ0 I dx sin θ sin2 θ 4π a2 3 µ0 I sin θ dx 4π a2
Now. The observer is at P . along the r direction. a vertical distance a from the wire.
Figure 3: The setup for the magnetic ﬁeld due to a long thin wire. we have line element d = dx. we ﬁrst see that: ˆ d × r  = d ˆ sin θ r = dx sin θ We also see that r2 = r2 . ˆ µ0 I d × r 2 4π r
dB(r) =
Let us look at the magnitude of the magnetic ﬁeld.14
2
ELECTROMAGNETIC FIELD EQUATIONS
Magnetic Field due to a Long Straight Wire See Fig (3) to see how we have deﬁned things. The wire carries a ˆ current I. We see that: x=− So. The wire goes oﬀ in either direction to ±∞. putting this all in: dB = µ0 I sin θ dθ 4πa a tan θ ⇒ dx = a dθ sin2 θ
. Now. So. So. by basic trigonometry: r= a sin θ
Hence. The vector r goes from the wire to P .
11)
. at a distance a from it. carrying currents of I1 . straight wire. Thus. and are separated by 1m.1. the force on wire 1. Hence. Hence. Now. two wires carrying a current will exert forces on each other. 2. is 2 × 10−7 N/m. This is the deﬁnition of the Ampere. Similarly. Amperes law is: B · d = µ0
S
J · dS
(2.2. where there is a total current I through the surface S. The deﬁnition of the Coulomb is similar: If a current of 1A is passing though a wire. in a magnetic ﬁeld. a wire carrying a current will feel a force in a magnetic ﬁeld. if the wire carries a charge q at velocity v is given by: F = q(v × B) Or. in terms of magnitudes F = qvB.7 Amperes Law
Consider a closed loop enclosing some surface S. then the current in each wire is 1A. 1 & 2. Hence: B= µ0 I 4πa
π
sin θ dθ =
0
µ0 I 2πa
Thus.1
Maxwell’s Equations
15
We must integrate from θ = 0 → π. if the magnitude of the force per unit length between two parallel wires carrying identical currents. the force per unit length is: F = µI1 I2 2πa
So. we have just computed that a wire carrying a current generates a magnetic ﬁeld. Suppose we have two wires. Thus. qv = I . Now.6 Magnetic Force Between two Parallel Wires
Now. then 1C of charge passes a surface in 1s. due to the magnetic ﬁeld generated by wire 2 is: µI1 I2 F1 = 1 2πa Where i is the length of wire i.1. 2. I2 a distance a from each other. We also know that a charge will feel a force in the presence of a magnetic ﬁeld. we have found the magnetic ﬁeld due to a long thin. due to the ﬁeld generated by 1 is: µI1 I2 F2 = 2 2πa Which is the same. the force felt on wire 2. the magnetic ﬁeld a distance a from wire 2 is given by: B2 = µ0 I2 2πa
The force felt on a wire. Then.
So.12) does not satisfy the above. as it stands.16 Using Stoke’s theorem.8
Faradays Law
This is the statement that the rate of change of a magnetic ﬁeld through some surface S generates an electric ﬁeld (an EMF ) in a loop enclosing the surface. charge continuity: ·J + ∂ρ =0 ∂t (2. we get:
2
ELECTROMAGNETIC FIELD EQUATIONS
× B = µ0 J
(2. That is.14)
Where we have used Gauss’ law.15)
Using Stokes’ theorem results in the diﬀerential form: ×E =− ∂B ∂t (2. So. we have thus far only found a ∂t staticequation.13). Hence. That is. by vector identities. Hence. by (2.
2. consider the following: Charge conservation is the statement that the rate of ﬂow of current density plus the rate of change of charge density is zero. which is not generally the case. our modiﬁed timevarying version of Amperes law is consistent with the continuity equation.12)
Note. Amperes law (2. ﬁnding the divergence of (2. Let us consider the time varying equation. we have that · J = 0.16)
. this is only true for static ﬁelds. Clearly. this is only true if ∂ρ = 0. That is: − ∂ ∂t B · dS =
S
E·d
(2.13)
Now. which we suppose to be: × B = µ0 J + µ0 ε 0 Taking the divergence of this: · × B = µ0 ·E ∂t 1 ∂ρ ⇒ 0 = µ 0 · J + µ 0 ε0 ε0 ∂t ∂ρ = ·J + ∂t · J + µ 0 ε0 ∂ ∂E ∂t (2.1.12) gives: · × B = µ0 ·J
The LHS is zero.
where χE is the electrical susceptibility.18) (2. where P is the polarisation.19) (2.
.2
Maxwell’s Equations in Materials Summary of Maxwell’s Equations in Vacuum
17
2.
2.22) (2. Hence.2.2
Maxwell’s Equations in Materials
Here we consider what happens to electric and magnetic ﬁelds if we put them in materials.20)
2. the total charge density is just ρ = ρf ree + ρind . which may be a tensor.2. we have the form of Gauss’ law in materials: · D = ρf ree (2. when placed in an electric ﬁeld. the induced charges only exist within the material! If ρind = − · P . Gauss’ law becomes: ·E = 1 (ρf ree + ρind ) ε0
Note.17) (2. In linear media only. depending on the medium. We consider the eﬀect of charges induced in a material. so that the above easily becomes: · (ε0 E + P ) = ρf ree To tidy this up. Then.9
ρ ε0 ·B = 0 ∂B ×E = − ∂t ·E = × B = µ 0 J + µ 0 ε0 ∂E ∂t
(2. Consider that there is a free charge density ρf ree and induced charge density ρind . let us deﬁne the electric displcement vector : D = ε0 E + P So.21)
We also say that P = ε0 χE E. we say that: D = (1 + χE )ε0 E = εr ε0 E Where εr is the relative permitivity for linear media only.1
Dielectrics
This is for materials in electric ﬁelds.1.
to see that: × H = Jf ree + ∂D ∂t (2. Let us initially consider the nontimevarying version of Amperes law.2.3
Summary of Maxwell’s Equations in Materials
· D = ρf ree ·B = 0 ×E = − ∂B ∂t ∂D ∂t
(2. the above easily × (B − µ0 M ) = µ0 Jf ree Again. where χH is the magnetic susceptibility. that ∂E = 0. i. we have that the total current density is the sum of free and induced current densities. Now.e. Hence.2. where M is the magentisation. using the divergence theorem: D · dS = ρ dV
S V
Where we have dropped the subscript from the charge density. Also. Therefore: B = µ0 (H + M ) = µ0 (1 + χH )H = µ0 µr H It is fairly easy to put the timevarying component back in. half of which is in medium
.2.25) (2. such as in a solenoid. we also say that M = χH H. Thus: × B = µ0 (Jf ree + Jind ) Also analogously. and collecting terms into: × H = Jf ree Where: H= 1 B−M µ0
Which we call the magnetic intensity. let us clean up the above.4
Fields Across Boundaries
Consider the new Gauss’ law.23)
2.24) (2. by dividing through by µ0 . Again. as in ∂t dielectrics.27)
× H = Jf ree +
2.26) (2. so that it encloses a volume V . let Jind = becomes: × M .2 Diamagnetics
2
ELECTROMAGNETIC FIELD EQUATIONS
This is for materials in a magentic ﬁeld.18 2. consider the surface S to be a box. which we can put into integral form easily.
noting that the above is just equal to the surface charge density: D1⊥ − D2⊥ = ρs That is. by adding some χ onto A. let us try: E =− φ− ∂A ∂t (2. The normal unit vector n points into medium 1. The work done by the electric ﬁeld in taking a small charge in the closed loop is zero. So. and the other part in medium 2.2. let us change E so that the above is still valid.3
Potentials & Gauge Invariance
19
ˆ 1.28) is still valid: × E+ ∂A ∂t = = − = 0 × − φ− × φ ∂A ∂A + ∂t ∂t
. for some scalar χ. Thus. (2.3
Potentials & Gauge Invariance
Let us consider Faraday’s law: ∂ B ∂t Now. parallel to the direction of the normal). the integral on the LHS is just: ˆ ˆ D · dS = D1 · n − D2 · n
S
These then become only components perpendicular to the boundary (that is.29). So. In a similar way. we know that curl grad is zero.28)
So. we would not ‘notice’ the change to B. That is: E·d =0 We ﬁnd that: E1// − E2// = 0 That is. × χ = 0. consider a small loop.
2. we have already introduced the magnetic vector potential B = × A. So. we have (from above) that Faradays law can be written: × E+ ∂A ∂t =0 (2. crossing the boundary. that is. That is. let us just check that if we choose (2. So. the electric ﬁeld is always continuous across boundaries. taking a step backwards again. let us write: ×E =− A −→ A + χ
Now. D is discontinuous at boundaries if ρs = 0. which we denote Di⊥ . so that the above becomes: ∂ × E = − ( × A) ∂t Now.29)
So. a discontinuity due to free charges.
we are just left with: ⇒ That is: ⇒
2 2
1 ∂φ =0 c2 ∂t 1 ∂2A = −µ0 J c2 ∂t2
(2. Now. a massive simpliﬁcation comes when we employ the Lorentz Gauge: ·A+ Then. Notice that this is decoupled.29) has done to Gauss law: ·E = ∂A ∂t ∂ ·A = − 2φ − ∂t ρ = ε0 ρ = − ε0 · − φ−
⇒
2
φ+
∂
·A ∂t
This is a coupled wave equation.
. our choice in (2. Let us see what (2.29) has not altered (2. driven by charge density ρ.29): ( ⇒
2
· A) − ·A+
2
A = µ0 J + = −µ0 J
1 ∂ c2 ∂t
− φ−
∂A ∂t
A−
1 ∂2A − c2 ∂t2
1 ∂φ c2 ∂t
Now. in contrast to the above driven by ρ.20
2
ELECTROMAGNETIC FIELD EQUATIONS
Thus.31)
Which is just a wave equation. in both ﬁelds φ and A. which is generated by J . using Amperes law: × B = µ0 J + Where we have used that c2 =
1 ε0 µ0 .28). inserting our choice for E (2.30)
A−
−
1 ∂2 c2 ∂t2
A = −µ0 J
(2.
1 ∂E c2 ∂t ×A 1 ∂E c2 ∂t
Noting that B = × × A = µ0 J +
Using a vector identity: × ×A = · A) − 2 A 1 ∂E = µ0 J + 2 c ∂t (
Now.
B.2. let us suppose that we can ∂t transform our potentials somewhat: E = − φ − ∂A ∂t ∂ = − φ − (A + χ) ∂t ∂χ ∂A = − φ + − ∂t ∂t
Where. i. Now. and inserting our transformed vector potential. So.33)
Where χ is some arbitrary scalar function. we see that B is invariant under the transformation.3
Potentials & Gauge Invariance
21
Basically. Let us have a small mathematical diversion. we know that E = − φ − ∂A .e. whereas the previous Coulomb gauge is only to be used on static ﬁelds. thus far. we can change A and φ up to the point that they dont change E. the transformed potential is: φ =φ− ∂χ ∂t ∂χ ∂t
So everything is consistent.
. we can make the transformations: A =A+ χ ∂χ φ =φ− ∂t (2. We shall now ﬁnd out how the scalar potential changes. as we have that B = × A and E = − φ. Now. Starting with Faradays law. The criteria we must satisfy is that by changing the potentials.32) (2. we do not aﬀect the ﬁelds. we have only used the relation between transformed & untransformed vector potential. which is applicable to timevarying ﬁelds. we end up with:
∂ × E = − ∂t B. for the magentic ﬁeld: B = = = ×A ×A+ ×A × χ
Where the ﬁnal term is zero as the curl of the gradient of a ﬁeld is zero. That is. Let us now see that we can indeed do this. as we see that the above Eﬁeld is then unchanged. So.
× E+
∂A ∂t
=0
Which we have already shown. Thus. We have introduced the Lorentz gauge. let: φ=φ + Hence.
notice. then the integral is negative. And has the property that is has unit area:
∞
δ(x − x ) dx = 1
−∞
In 3D. with just a constant inside the (delta) function. So: − 1 a
−∞ ∞ y f ( −a )δ(y − y )dy =
1 a
∞ −∞
y f ( −a )δ(y − y )dy
. analogously: f (r)δ 3 (r − r ) d3 r = f (r ) (2. Now. let us state (then we shall prove it) the following: 1 δ(a(x − x )) = δ(x − x ) a To show this. it can be used to pick up a single value of a function. it will return that value of any function within the integral. replace a with −a everywhere above:
∞×(−a) −∞×(−a) y f ( −a )δ(y − y )dy /(−a) = −
1 a
−∞ ∞
y f ( −a )δ(y − y )dy
Now. or inﬁnitely thin sources. we have: δ 3 (r − r ) = δ(x − x )δ(y − y )δ(z − z ) So. The Diracδ function is deﬁned by:
∞
f (x)δ(x − x ) dx = f (x )
−∞
(2. One may think about this in the following way: the integral sweeps over the xspace. if a < 0.35)
Let us consider the delta function.22
2
ELECTROMAGNETIC FIELD EQUATIONS
2. let us actually evaluate:
∞
f (x )δ(a(x − x ))dx
−∞
Let us change variables: y = ax So:
∞×a −∞×a
y = ax
f ( y )δ(y − y )dy /a a
Notice. Hence. Let us start with a simple example. the eﬀect of having ‘minus’ an integral is to ﬂip the integration limits.4
The Diracδ Function & Green Functions
These are very useful for dealing with point charges. but the δfunction will only return a nonzero value when x = x . having a function as an argument.34)
Notice. To see this.
2.4
The Diracδ Function & Green Functions
23
So, let us take the modulus of a. Hence: 1 a Thus:
∞ ∞ −∞
f ( y )δ(y − y )dy = a
1 y 1 f( ) = f (x) a a a
f (x )δ(a(x − x ))dx =
−∞
1 f (x) a
(2.36)
Now, to consider and actual function: δ(g(x)) So, we need to ﬁnd the zeros of g(x), as we know that the deltafunction is nonzero when its argument is zero. Let them be at xi , so that we have that g(xi ) = 0. Let us now do a Taylor expansion about the zero: g(xi + ) = g(xi ) + (x − xi )g (xi ) + . . . Note, the ﬁrst term is zero, by deﬁnition. Note, here, a prime denotes derivative with respect to x, wheras above, a prime is a way of distinguishing variables. Hence, near a zero: g(x) ≈ (x − xi )g (xi ) Near a zero. So, we have that: δ(g(x)) = δ(g (xi )(x − xi )) Near a zero. This is just like we had before, except we have more than one place the argument of the deltafunction is zero. So, we must add the contributions up from all the zeros. Hence: δ(g(x)) =
i
δ(x − xi ) g (xi )
(2.37)
Where xi is a zero of the function g(x). Now, suppose we had a set of charges, where charge qi resides at some position ri , then, we may write the total charge density as: ρ(r) = qi δ(r − ri )
i
Which can be thought of as a set of ‘impluse charges’; which may however, be continuous. Or, as another example, consider that some charges are distributed on an inﬁnitely thin shell, of radius a, and that the distribution conforms to some σ(θ). Then, the charge distribution may be written: ρ(r) = δ(r − a)σ(θ) 2.4.1 Green Functions: Electrostatics
If we wish to solve a diﬀerential equation of the form: (Lx u)(x) = f (x)
24
2
ELECTROMAGNETIC FIELD EQUATIONS
Plus boundary conditions; where Lx is a linear hermitian operator. Then, to do so, we can always write the solution to the DE, u(x) as: u(x) = Where the Green function is deﬁned by: Lx G(x, x ) = δ(x − x ) (2.39) G(x, x )f (x )dx (2.38)
We shall now supress the xsubscript on the operator: it is clear that it only operates on x, not x . To prove the above statement isn’t too hard: (Lu)(x) = f (x) ⇒L G(x, x )f (x )dx = = LG(x, x )f (x )dx δ(x − x )f (x )dx
= f (x) Hence proven. In the ﬁrst line, we used the linearity of the operator, to be able to bring it inside the integral (which is over x anyway). Then we used the deﬁnitions of the Green function and delta function. Let us proceed by an example from our present electrostatic discussions. Now, from Gauss’ law · E = ρ/ε0 and E = − Φ, we can easily derive Poissons equation:
2
Φ(x) = −
ρ(x) ε0
Note, this is strictly for a static charge distribution, else we would have the A term as well. We have the boundary condition that Φ(x) → 0 as x → ∞. Now, for a point charge q, at x , we have a charge distribution which is just a delta function: ρ(x) = qδ(x − x ) Note, strictly, we should have written δ 3 , but that is understood, as its argument has 3 variables. Now, we know that the resulting Poisson equation has the following solution:
2
Φ(x) = −
q δ(x − x ) ε0
⇒
Φ(x) =
q 1 4πε0 x − x 
Now, upon comparison of the above formalism for Green function, Lu = f is just the Poisson equation. That is, the operatior L is just the 2 operator. Hence, from LG = δ(x − x ), we see that the Green function we want, is that satisfying:
2
G(x, x ) = δ(x − x )
We have the form of G, from the above expression for the point charge. That is: G(x, x ) = − 1 1 4π x − x 
2.5
Poynting’s Theorem
25
Hence, we have that the solution to Poissons equation is given by: Φ(x) = That is: Φ(x) = G(x, x ) − ρ(x ) ε0 d3 x
1 4πε0
ρ(x ) 3 d x x − x 
Which is something we already knew, but we have derived it using Green functions; and in the process, have identiﬁed a Green function. Notice, we have also arrived at a useful relation:
2
1 = −4πδ(x − x ) x − x 
(2.40)
Green function theory is extensive, and is used in solving diﬀerential equations (as was hinted at previously); but we shall not go into that here.
2.5
Poynting’s Theorem
Let us start by stating the Lorentz force law: F = q(E + v × B) So, the work done dW on a charge dq, when displaced a distance d is given by: dW = F · d That is: dW = dq(E + v × B) · d Now, from v = d , a trivial relation, we can write the above as: t dW = dq(E + v × B) · vdt That is: dW = dt
dq(E + v × B) · v dq = ρV d3 r
However, note that: Hence: dW dt =
V
ρV (E + v × B) · v d3 r ρV {E · v + (v × B) · v} d3 r
V
= =
V
E · (ρV v)d3 r
the total rate of doing work.41): dW dt =
V
∂E ∂t ∂E ∂t
× H − ε0
E· E·(
V
× H − ε0
∂E ∂t
d3 r ∂E 3 d r ∂t
=
× H) − ε0 E ·
Now.41)
That is. putting this into (2. That is: · (E × H) d3 r =
V S
(E × H) · dS
. by the ﬁelds. Hence: dW = dt E · J d3 r
V
(2. we also know that from Faraday’s law: ×E =− Thus.26 Where we have used (the easily veriﬁable):
2
ELECTROMAGNETIC FIELD EQUATIONS
v · (v × B) = 0 We also note that J = ρV v. if there is a continuous distribution of charge and current. we can use the divergence theorem on the far LHS integral. using this: dW dt =
V
∂B ∂H = −µ0 ∂t ∂t
−µ0 H ·
∂H − ∂t
· (E × H) − ε0 E · H·
V
∂E 3 d r ∂t E·
V
= −
V
· (E × H) d3 r − µ0
∂H 3 d r − ε0 ∂t
∂E 3 d r ∂t
Now. after using this: dW = dt H ·(
V
× E) − E · ( × E) −
× H)
× H) = H · (
· (E × H)
× E) −
· (E × H) − ε0 E ·
∂E 3 d r ∂t
Now. we use a vector identity for the ﬁrst expression in the integral: · (E × H) = H · ( ⇒E·( So. consider Amperes’ law: × H = J + ε0 That is: J= So. Now.
the rate at which particles gain energy is equal to the rate at which the ﬁelds lose energy. a plane wave is given by: E = E0 ei(k·r−ωt) And. where each component has the same exponential factor. Ey . if the ﬁelds are constant (i. we see that if the surface integral is zero. We note that: ×E =− So: ∂B ∂t H= 1 B µ0 B = B0 ei(k·r−ωt)
∂B = −iωB ∂t And. we start to recognise the Poynting vector. Ez ). we can easily derive: × E = ik × E Thus.45)
So. their time derivatives are zero). Thus.e.43) (2.2. we see that the Poynting vector represents the rate at which EM ﬁelds transport energy across a unit surface (hence the integral).44)
UE = UM Thus. by doing the crossproduct. As an example. That is. So. we have: =
V
µ0 H2 d3 r
dW ∂ = − (UE + UM ) − dt ∂t
P · dS
S
(2. we have: ×E =− ∂B ∂t ⇒ ik × E = iωB
. then this inﬂow of energy is provided by the Poynting vector.42) (2. which is i(kx x + ky y + kz z − ωt). ley uscompute the Poynting vector. and realising that the argument of the exponential is actually a scalar. and that E0 = (Ex . and energy densities of the electric and magnetic ﬁelds: P = E×H ε0 E2 d3 r
V
(2. and if work is still being done on the particles. we have that P = E × H. Also. let us consider the energy ﬂux for a plane harmonic wave.5
Poynting’s Theorem
27
Putting this back in: dW dt = −
S
(E × H) · dS − µ0
V
H·
∂H 3 d r − ε0 ∂t
E·
V
∂E 3 d r ∂t
∂ = − (E × H) · dS − ∂t S
ε0 E2 + µ0 H2 d3 r
V
Now.
We also have the standard relation: ω = kc. in the direction of k. The time averaged value is just: 1 E2 = E2 2
.28 Hence:
2
ELECTROMAGNETIC FIELD EQUATIONS
k × E = ωµ0 H Now. So: k × E = ωµ0 H ˆ ⇒ k k × E = ωµ0 H 1 ˆ ⇒H = k×E kωµ0 1 ˆ = k×E µ0 c Hence. an instantaneous value. however. the Poynting vector is: P = E×H 1 ˆ = E×k×E µ0 c 1 = E2 µ0 c
This is. we also use the fact that: k ˆ k= k ˆ Where k is a unit vector. We shall write k = k.
Let one edge of the plates be held at some potential V0 . rather than φ.6
Laplace Equation & Its Solutions
29
2. consider the Dirichlet boundary conditions: • V (x.use Cartesian coordinates. 2.
.6. we are able to easily derive the electrostatic Poisson equation: ρ 2 V =− ε0 Where V is the scalar potential. a) = 0. from Gauss’ law. to avoid any possible confusion with the coordinate. • Cylindrical .2. Now. using E = − V . We can solve using the inherent symmetry of the system: • Rectangular .use spherical polars. • Spherical . a mix of the above two. we shall denote the scalar potential as V . y) = V0 . we have the Laplace equation:
2
V =0
(2.up to additive constants.6
Laplace Equation & Its Solutions
Here. We shall now consider various types of solutions. and its solution is unique . The general method is moreorless unchanged. Consider 2 plates which are inﬁnite in z. 0) = V (x. • Neumann:
∂V ∂n
ˆ ≡n·
V is known on S.
• Or. we use the separation of variables technique. We can use various boundary conditions: • Dirichlet: V (r) is known on some surface. • V (0.1 Solution to the Laplace Equation: Cartesian
This method will apply to all cartesian systems. with diﬀerent boundary conditions giving diﬀerent ﬁnal results.use cylindrical polars. We have seen that this has solution: V (r) = 1 4πε0 ρ(r ) 3 d r r − r 
V
If there is no charge density. So.46)
In solving the equation.
48) = k2 = −k 2 d2 X d2 Y +X 2 =0 dx2 dy
So. 0) = 0. in 2D (or 3D. y). we have that our general solution to the Laplace equation. we see that D = 0. A = 0. and will generally be linear superpositions. y) = {Aekx + Be−kx }{C sin ky + D cos ky} (2. y) = e−kx (C sin ky + D cos ky) Let us use another boundary condition: V (x.47) (2. y) = Ce−kx sin ky
. with symmetry) is: V (x. Hence. So. we have: V (x. That is. using separation of variables. we have that: V (x. Hence. Hence. we have that each of the above expressions must be a constant: k 2 + (−k 2 ) = 0 That is: 1 d2 X X dx2 1 d2 Y Y dy 2 Each of these is easily solved: X(x) = Aekx + Be−kx Y (y) = C sin ky + D cos ky (2. let us continue.49)
Where the constants must be determined by the boundary conditions. we now have: V (x.30 • limx→∞ V (x.
2
ELECTROMAGNETIC FIELD EQUATIONS
So. So. y) = X(x)Y (y) Putting this into the Laplace equation: Y Dividing through by V = XY : 1 d2 X 1 d2 Y + =0 X dx2 Y dy 2 So. y) = 0. solving for our speciﬁc set of boundary conditions: Let us use the boundary condition that V must decay as x → ∞. we have that the Laplace equation reduces to: ∂2V ∂2V + =0 ∂x2 ∂y 2 Where the potential V (x.
That is:
a
Cn
n 0
sin
nπy mπy sin a a
a
dy = V0
0
sin
mπy a
dy
Thus:
n
a V0 a Cn δnm = (1 − cos mπ) 2 mπ
That is:
V0 a a Cm = (1 − (−1)m ) 2 mπ So. y) =
m=1
4V0 − mπx mπy e a sin mπ a
m odd
(2. That is. 2. y) = V0 . in Cartesian coordinates.51) with “another”. we use:
a
sin
0
nπy mπy sin a a
dy =
a δnm 2
So.52)
Thus. y) =
n=1
Cn e −
nπx a
sin
nπy a
(2. let us multiply both sides of (2.6
Laplace Equation & Its Solutions
31
Let us use another boundary condition: V (x. and integrate. if m is odd.2. we use the orthogonality of sine functions.
.50)
Let us ﬁnd the constant Cn . And. .51)
To go further. under our given boundary conditions. y) = Ce−
nπx a
nπ a
n = 1. nπy a
sin
Using the linear superposition of solutions. we see that if m is even. 3. So: sin ka = 0 Thus: k= Therefore: V (x. then Cm is zero. Let us use the ﬁnal boundary condition: V (0. we have found the solution of the Laplace equation. then: Cm = Therefore. . we have that:
∞
V (x. Hence:
∞
Cn sin
n=1
nπy a
= V0
(2. . we have:
∞
4V0 mπ
∀m odd
V (x. a) = 0.
we have that the Laplace equation becomes: r2 d2 R r dR 1 d2 Φ + + + b2 r2 = 0 R dr2 R dr Φ dϕ2 Now.57)
.54) (2. ϕ.53) (2. in 3D cylindrical polar coordinates.53) and (2. we have that the Laplacian. Thus.54) are just: Φ(ϕ) = Cα cos αϕ + Dα sin αϕ Z(z) = Eb cosh bz + Fb sinh bz (2.56) (2. so that the separation of variable happens thus: V (r. is the product of the solutions to the following equations: d2 Φ dϕ2 d2 Z dz 2 d2 R 1 dR α2 + + b2 − 2 dr2 r dr r = −α2 Φ = b2 Z (2. we let: 1 d2 Z = b2 Z dz 2 Hence. we also let: 1 d2 Φ = −α2 Φ dϕ2 Hence. z) = R(r)Φ(ϕ)Z(z) So. the Laplace equation further reduces to: r2 d2 R r dR + − α2 + b2 r2 = 0 R dr2 R dr Cleaning up: d2 R 1 dR α2 + + b2 − 2 dr2 r dr r R=0
So. we notice a similar linear independence of the terms as for the Cartesian case. to summarise. z). ϕ. in cylindrical polars.32 2. we use the coordinate system V (r. we have that the solution to the Laplace equation. using our separated variables: 1 d2 R 1 dR 1 d2 Φ 1 d2 Z + + 2 + =0 R dr2 Rr dr r Φ dϕ2 Z dz 2 Now.2
2
ELECTROMAGNETIC FIELD EQUATIONS
Solution to the Laplace Equation: Cylindrical Polars
Here.55)
R = 0
We have that the solutions to (2.6. takes on the form:
2
V =
∂2V 1 ∂2V 1 ∂ ∂V + r + 2 r ∂r ∂r r ∂ϕ2 ∂z 2
Thus.
ϕ). Na (br) are Bessel functions of the ﬁrst and second kinds. ϕ)
Where Y m are spherical harmonics. ϕ.6. let us try R(r) = rα .58)
Where Ja (br). this is just: α2 + α − − 2 = 0
. the Laplace equation is: 1 ∂ 2 ∂V 1 ˆ r − 2 L2 = 0 2 ∂r r ∂r r Hence. z) =
a. So.60)
ˆ are eigenfunctions of L2 thus: ˆ L2 Y
m (θ. θ.6
Laplace Equation & Its Solutions
33
And that we note (2. we have that: 1 Y r2 We know that Y
m m
d 2 dR 1 ˆ r − 2 RL2 Y dr dr r
m
=0
(2. for R(r). we have that the Laplace equation (2. ϕ)
= ( + 1)Y
m (θ.2. ϕ)
Hence.55) is Bessel’s equation. we try an ansatz: a power law. and are generally looked up. we separate V thus: V (r. we have that the general solution to the Laplace equation. in 3D cylindrical polars.3 Solution to the Laplace Equation: Spherical Polars
We use the coordinate system V (r. ϕ) = R(r)Y
m (θ.59)
Where the constants must be determined by initial conditions. putting this into the above yields: Ym rα α(α + 1)rα = 2 ( + 1)Y m r2 r Now. which has solutions: R(r) = Aab Ja (br) + Bab Na (br) (2. Hence. ˆ We note that the above Laplacian has the L2 operator. is given by: V (r. So. respectively. That is.60) reduces to: Y m d 2 dR R r − 2 ( + 1)Y r2 dr dr r
m
=0
Now.b
{Aab Ja (br) + Bab Na (br)}{Cα cos αϕ + Dα sin αϕ}{Eb cosh bz + Fb sinh bz} (2. where the Laplace equation takes on the form: 1 ∂ 2 ∂V 1 1 ∂ ∂V 1 ∂2V r + 2 sin θ + =0 r2 ∂r ∂r r sin θ ∂θ ∂θ sin2 θ ∂ 2 ϕ2 Now. θ. 2.
B m. the Laplace equation has the form we used previously. ϕ) = Θ (θ)Φm (ϕ) where Φm (ϕ) = eimϕ . ϕ. That is.
Where we have used both solutions to the ansatz.m
A
mr
+
Bm r +1
Y
m (θ. such as a point charge surrounded by dielectric. however. as Y m (θ. we have that the general solution to the Laplace equation. then the Laplace equation looks like: 1 ∂ 2 ∂V r =0 r2 ∂r ∂r And the potential is just V (r) = R(r).34 Which we can equivalently write as:
2
ELECTROMAGNETIC FIELD EQUATIONS
(α − )[α + ( + 1)] = 0 Hence. in spherical polars is: V (r. note that we must have: dR r2 =D dr
. then the solution reduces to: B V (r. consider that they are in there. ϕ) =
. are just: α= α = −( + 1)
And can be veriﬁed by substituting back in.
to be
Now. or diﬀerentials. That is. ϕ)
(2. and can be generated by Rodrigues formula: P (x) = And the ﬁrst few are given by: P0 (x) = 1 P1 (x) = x P2 (x) = P3 (x) = P4 (x) =
2 1 2 (3x − 1) 3 1 2 (5x − 3x) 4 2 1 8 (35x − 30x
1 2 !
d dx
(x2 − 1)
+ 3)
If the system is symmetrical in both θ.61)
m. To solve this. we see that the two solutions to the above equation. θ. θ. So. each having diﬀerent coeﬃcients A determined by initial conditions. we must solve: 1 d r2 dr r2 dR dr =0
To see why there is no angular components. if the system has axial symmetry. these are then just divided out. and the radial diﬀerential will be multiplied by the angular functions. The diﬀerentials of the angular parts are zero (they are constant: the symmetry of the system). ϕ) = A r + +1 P (cos θ) r Where the P (cos θ) are Legendre polynomials.
Then the Laplace equation is satisifed. the transformation between Cartesian and spherical polars. So. • The ﬁeld at inﬁnity is just the ‘unperturbed’ ﬁeld. we solve the Laplace equation in spherical coordinates. see the relevant appendix. E0 . we immediately see that there is axialsymmetry. We can ﬁnd R by the standard method: r2 dR dr = D dr r2 D dr r2
⇒ dR = D ⇒ R(r) = = − =
D +A r
B +A r
Where A. in a uniform electric ﬁeld. • The perpendicular D⊥ must be continuous at boundaries.4 Example: Dielectric Sphere in Uniform Eﬁeld
Consider a dielectric sphere. θ) =
=0
Ar +
B r +1
P (cos θ)
The boundary conditions are implied. Remember. not diverge) as r → 0.2.6
Laplace Equation & Its Solutions
35
Where D is some constant. radius a.e. with axial symmetry V (r. 2. so we choose it to be aligned with the uniform applied ﬁeld. The alignment is obviously arbitrary. consider how we choose the alignment of the coordinate system. just the integral. relative permittivity εr . • The potential must be continuous at the boundary. as E = − V . thus. To see the motivation behind this. note: P1 (x) = x ⇒ P1 (cos θ) = cos θ
. for z is z = r cos θ. as no free charges. θ):
∞
V (r. Now.6. B are both constants. hence the potential as r → ∞ is just V = −E0 r cos θ. Hence. So. not necessarily given: • The potential must go to zero at inﬁnity. • The potential must remain ﬁnite (i. we have that the potential in an angularsymmetric system is given by: B V (r) = A + r For a more extensive treatise of spherical harmonics. otherwise vaccum. we have that V = −E0 z.
we have: V1 = A1 r cos θ V2 = B2 −E0 r + 2 r r≤a cos θ r>a
Let us now consider the ﬁrst ‘boundary’ condition. That is. So: D1⊥ (r = a) = D2⊥ (r = a) ⇒ εr E1 (r = a) = E2 (r = a)
. = 1) at inﬁnity. taking = 1. thus far. θ) = Ar + B r2 cos θ
Let us now continue by looking at the two regions (inside and outside) separately. Thus. So. That is: V1 (a. θ) So: A1 a cos θ = That is: A1 = B2 − E0 a3 (2. by orthogonality considerations of the Legendre polynomials. with diﬀerent constants in both cases.36
2
ELECTROMAGNETIC FIELD EQUATIONS
So. multiplying both sides by ‘another’ polynomial will result in just the dipole term being ﬁltered out. This may be seen in more detail in examples in the appendix. θ) = V2 (a. The consideration that V remain ﬁnite at the origin leads us to conclude that B1 = 0. on the potential side. this is the case because there are no free charges on the boundary. but we can’t say anything about B2 .e. If there were. which we consider in later examples. consider that the potential (given by a sum over an inﬁnite number of multipoles) must go to the dipole term (i. that the potential is continuous at the boundary.62) −E0 a + B2 a2 cos θ
Let us now apply the second ‘boundary’ condition. That is: V1 = V2 = B1 r2 B2 A2 r + 2 r A1 r + cos θ cos θ r≤a r>a
Let us consider a boundary condition:
r→∞
lim V = −E0 r cos θ = V2
Thus. that D⊥ is continuous at the boundary. we see that A2 = −E0 . there is a discontinuity in D⊥ . the ‘prototype’ for the potential is: V (r. Hence. we have = 1 in the summation.
where outside the cavity is a dielectric. θ) = −E0 r cos θ + E0 cos θ εr + 2 r 2 V1 (r. • Potential is zero at inﬁnity.5 Example: Charge Inside Spherical Cavity r≤a r>a
Consider a charge q at the centre of a spherical cavity. our system uses spherical polars. • Electric ﬁeld inside cavity. and inside vacuum. B2 . permittivity εr . It is just: V (r) = A + B r
. the boundary conditions: • Potential and D⊥ continuous on boundary. 2. giving: A1 = − 3E0 εr + 2 B2 =
2B2 a3
(2. hence: ∂r εr ∂V1 ∂r =
r=a
∂V2 ∂r
r=a
⇒ A1 εr cos θ = ⇒ A 1 εr So:
2B2 a3 2B2 = −E0 − 3 a −E0 −
cos θ
A1 εr = −E0 − We can easily solve (2. putting these back into our potentials for inside and outside: 3E0 r cos θ εr + 2 εr − 1 a 3 V2 (r.63)
εr − 1 εr + 2
E0 a3
Hence. Again. in limit r → 0 is just the Coulomb ﬁeld. but with complete angular symmetry.63) for A1 . the Laplace equation is: 1 ∂ 2 ∂V r =0 r2 ∂r ∂r We have already discussed the solution to this equation. θ) = − Hence solution found. we have that E = − ∂V .6.62) and (2. Here.2. Hence.6
Laplace Equation & Its Solutions
37
But.
Hence. let us consider r → 0. we have: dV1 dr = εr
r=a
B1 B2 = a a
dV2 dr
r=a
Noting that this time the vacuum is on the inside. Thus: B1 B2 − 2 = −εr 2 a a Thus: B1 = εr B2 Now. That is: B2 = aA1 + B1 Let us consider the continuity of D⊥ at the boundary. again. let us split the problem into inside/outside: B1 r B2 V2 (r) = A2 + r V1 (r) = A1 + r≤a r>a
To make the potential zero at r = ∞. via B2 = aA1 + B1 . our solution is: V1 (r) = − q 4πε0 a εr − 1 εr q 4πε0 r q V2 (r) = 4πε0 εr r + r≤a r>a q 4πε0 a εr − 1 εr dV1 B1 =− − 2 dr r q q = 2 4πε0 r 4πε0 q 4πε0 r2
q 4πε0 εr
. Let us consider the continuity of the potential at the boundary: A1 + Using A2 = 0. and dielectic on the outside. via B1 = εr B2 . we have that: A1 = − Therefore. inside: E1 → Noting: E1 = − Hence: B1 = r2 E1 = r2 We also have. and hence conclude that A2 = 0. That is. that: B2 = And. we consider V2 .38
2
ELECTROMAGNETIC FIELD EQUATIONS
So.
Also recall. that for a linear medium.1
Multipoles
Electric Dipole
Figure 4: The electric dipole.7. Hence. at distances r1 . Two charges. Then. Linear superposition allows us to write that the the ‘composite’ potential is just: V (P ) = 1 4πε0 q1 q2 + r1 r2
. Recall that D = ε0 E + P . observed at P . q. −q. we have that: E2 = − V2 = Thus: q ˆ r 4πε0 εr r2
Finally. separated by a distance d.2. r2 from some observation point P . recall that ρind = −
q(εr − 1) ˆ r 4πεr r2 · P . as measured from d 2 Consider the potential due to two charges. we have that: P = ρind = − q(εr − 1) 4πεr · ˆ r q(εr − 1) =− δ(r) 2 r εr
We have thus computed the surface charge density induced. q1 and q2 . Then.7
2. after a trivial rearangement: P = (εr − 1)ε0 E Let us (as we are liberty to) consider the ﬁeld just outside the sphere.7
Multipoles
39
Let us consider what the surface charge induced is. where P is the polarisation. D = εr ε0 E. where we have used: · ˆ r = 4πδ(r) r2
2. which is at a position r.
squarerooting: r1.2 = r2 +
d 2 d 2r
2
rd cos θ
2
That is. due to a dipole: V (r) = Which we may write as: ˆ 1 p·r 2 4πε0 r Where we have deﬁned the dipole moment p ≡ qd. we see that:
2 r2 = r2 + ( d )2 + rd cos θ 2
So. with reference to the ﬁgure:
2 r1 = r2 + ( d )2 − 2r d cos θ 2 2
That is:
2 r1 = r2 + ( d )2 − rd cos θ 2
Also. (1 + x)n ≈ 1 + nx.64)
. for r >> d. i. being careful that the angle is now cos α = cos(180 − θ) = − cos θ.2 r r Now. V (r) = q 4πε0 1 1 − r1 r2 = q d cos θ 4πε0 r2 (2. if we combine these results:
2 r1. the cosine rule can be very easily derived. to ﬁrst order binomial expansion.e. we can neglect such quadratic terms.2 = r2 1 +
d cos θ r
Now. when q1 = −q2 = q. at r. So: V (P ) = q 4πε0 1 1 − r1 r2
Just a quick note.40
2
ELECTROMAGNETIC FIELD EQUATIONS
Now. consider the slightly more speciﬁc case. just:
2 r1.2 r 2r Hence: 1 1 d d d 1 − = 1+ cos θ − 1 + cos θ = 2 cos θ r1 r2 r 2r 2r r Therefore. we have:
1 d 1 = 1± cos θ r1. we can write the potential.2 = r2 1
d cos θ r d cos θ r
1/2
Thus. Hence:
2 r1.2 = r 1 Hence: 1
−1/2 d 1 1 cos θ = r1. Consider: (r − r )2 = r2 + (r )2 − 2r · r = r2 + (r )2 − 2rr cos θ So.
2
In the present multipole discussion. 2 6 3 2 5 δ − δ3 + . so that R is the
Figure 5: The multipole. and γ the angle between charge and observation point. ..2. ˆ vector from the charge to the observation point. let γ be the angle between r and r . Here. where θ is the angle between the observation point P . and that γ ≡ θ − θ. θ is the angle between the axis and the charge distribution ‘bit’. 8 16 n(n − 1) 2 n(n − 1)(n − 2) 3 x + x + .66) ⇒ r r r − 2 cos γ r √ R=r 1+δ (2. let: δ≡ Then: R2 = r2 (1 + δ) Hence: 1 1 = (1 + δ)−1/2 R r Now. Notice how things are deﬁned. we assume that there is axial symmetry. which is why we use Legendre polynomials.... So. under the cosine rule: R2 = r2 + (r )2 − 2rr cos γ That is: R2 = r 2 1 + Now. the binomial expansion is such that: (1 + x)n = 1 + nx + Therefore: 1 (1 + δ)−1/2 = 1 − δ + 2 1 = 1− δ+ 2 (−1/2)(−3/2) 2 (−1/2)(−3/2)(−5/2) 3 δ + δ + . as shown in the ﬁgure. as will become apparent. Similarly.7
Multipoles General Multipole Expansion
41
2. 2! 3! (2. and R ≡ r − r .65) r r
2
−2
r cos γ r
. .7. Consider a distribution of charges. We have R ≡ r − r . and the z axis.
2
Now. recall the Legendre polynomials: P0 (x) = 1 P1 (x) = x P2 (x) = 3x2 − 1 2 P3 (x) = 5x3 − 3x 2
Thus. the total potential. due to some charge density ρ(r ) is given by: 1 V (r) = 4πε0
∞ =0
1 r
+1
P (cos γ)(r ) ρ(r )d3 r
To make this (possibly) a little more transparent. we have: 1 1 1 1− = R r 2 r r r − 2 cos γ r + 3 8 r r
2
r − 2 cos γ r
2
+ . . and the example relating to our present discussion. we have that the potential due to a small bit of charge is just: dV = 1 dq 4πε0 R ⇒ V (r) = 1 4πε0 1 ρ(r)d3 r R
Hence. is the order of the ‘pole’. Hence: 1 1 = P0 (cos γ) + R r r r P1 (cos γ) + r r
3
P3 (cos γ) + . under the Legendre polynomial generating function. at r. via (2..66). So.
If we then gather appropriate terms: r 1 1 1 + cos γ + = R r r r r
2
3 cos2 γ − 1 + 2
r r
3
5 cos3 γ − 3 cos γ + . we notice those terms as present.. . let us write out the ﬁrst few terms of the sum: V (r) = = 1 4πε0 1 4πε0 1 r r2 ρ(r )d3 r + 2 cos γρ(r )d3 r + 3 (3 cos2 γ − 1)ρ(r )d3 r r r 2r 1 r r2 + cos γ + 3 (3 cos2 γ − 1) ρ(r )d3 r r r2 2r
.42
2
ELECTROMAGNETIC FIELD EQUATIONS
Hence. The above ‘motivation’ shows that the expansion seems to work. but is by no means a proof! A proof of this may be found in the relevant appendix.. and taking ourselves out of using δ. The order to which the sum is taken. putting this expansion into (2.65).
We therefore write the whole thing as a sum: 1 1 = R r Which is just: 1 = R
∞ =0 ∞ =0
r r
P (cos γ)
1 r
+1
(r ) P (cos γ)
Remembering that γ is the angle between r and r ..
Hence. all components other than the ﬁrst are zero.e. and we are looking at the potential at a distance r away from the origin of the system (i. So. as we have with the integral version. the multipole expansion will be dominated by the monopole term. so the monopole is the only contribution: 1 1 q V (r) = ρ(r )d3 r = 4πε0 r 4πε0 r The total charge is just q.67)
Where we have noted that the angle between r and r is γ. as the separation between charges is just r . and continuing up. the dipole moment is better described in terms of the size. Note. with reference to the ﬁgure. If the total charge vanishes. at large distances. we have: ˆ p·r = 1 r · r ρ(r )d3 r = r 1 rr cos γρ(r )d3 r = r r cos γρ(r )d3 r r ρ(r )d3 r (2. quadrupole = 2. relative to some orgin. by deﬁnition. We see that the ﬁrst term is the monopole = 0. if the monopole term is zero. and total charge is given by the integral over charge density.7
Multipoles
43
We have that charges are arranged at some r . one must be very careful in noting that: cos γ = cos(θ − θ) = cos θ cos θ + sin θ sin θ And that the volume element is d3 r = r 2 sin θ dr dθ dφ . somewhere roughly at the centre of the distribution. however. Now. we have a multipole expansion of V . then the dipole term is dominant. Thus. as r = r .2. This is equivalent. With computations on this. shape & density of the system. at a position r). if we deﬁne the dipole moment to be: p= ˆ r Then. suppose we shift the coordinates (the origin). or the lowest pole term present. so that r = r + a. a new dipole moment: pa = = (r + a)ρ(r )d3 r r ρ(r )d3 r + aρ(r )d3 r
= p + aq
. shifting the origin changes multipole moments. For a point charge at the origin. then dipole = 1. the dipole moment is independant of the origin: Suppose we have a dipole moment relative to some origin 0: p= r ρ(r )d3 r
Then. Which is just the dipole term in the expansion: V = 1 4πε0 r2 r cos γρ(r )d3 r = 1 ˆ p·r 4πε0 r2
Recall the dipole moment we deﬁned for a pair of charges: p = qd. Notice.
as we said. pa = p. the electric ﬁeld due to a dipole: V (r. Also. as E = − V . to conclude. Notice that the angle between x and x is γ. we shall be using r as x. q = 0.7.44 But.3
Spherical Harmonic Expansion Multipoles
Figure 6: The deﬁnition of angles in spherical polars. θ) = Hence. θ) = Where we have the pole moment: p = r P (cos γ)ρ(r )d3 r 1 4πε0 p r
+1
2. although this is a trivial assignment of symbols. hence. we have that we can write the scalar potential as a sum over multipole moments: V (r. Now. and r as x . we have its components: Er = − Eθ Eϕ Hence: E= It is common to notate this as: E=
2
ELECTROMAGNETIC FIELD EQUATIONS
1 p cos θ 4πε0 r2
∂V 1 2p cos θ = ∂r 4πε0 r3 1 ∂V 1 p sin θ = − = r ∂θ 4πε0 r3 = 0 1 p ˆ ˆ (2 cos θr + sin θθ) 4πε0 r3 1 1 ˆr (3(p · r )ˆ − p) 4πε0 r3
∞ =0
So. Figure from Jackson.
.
φ )ρ(r )d3 r
(2. φ)
r C ∗m (θ . φ)Q m
(2.71)
r
+1
C
m (θ. we have that our potential expression cleans up to: V (r) = 1 4πε0 1
m
r
+1
C
m (θ. inserting (2.69)
We have that m = − → .68) above: V (r) = = 1 4πε0 1 4πε0 4π 1 2 + 1 r +1 4π 1 2 + 1 r +1 r Y
m ∗ m (θ. φ)
(2. φ)
≡
4π Y 2 +1
m (θ. So. in integer steps. φ ) = 4π Y ∗ (θ .φ )
(2. φ)Y m (θ
.7
Multipoles
45
Recall the following expansion in terms of Legendre polynomials: 1 = r − r  1 r
+1
r P (cos γ)
Where γ is the angle between r and r . to tidy this up. φ )ρ(r )d3 r
Y
m
m (θ. Where. let us deﬁne the multipole moment to be: Q Then: V (r) = 1 4πε0 1
m m
≡
r C ∗m (θ . we have the following relation: 4π P (cos γ) = 2 +1
m=+
Y
m=−
∗ m (θ.69) into our potential expression (2. φ )ρ(r )d3 r m
Now. and has the relation r · r = rr cos γ. φ ) 2 +1 m
Then. φ)Y m (θ
. Let us write: C
m (θ. from the addition theorem for Spherical harmonics: cos γ = cos θ cos θ + sin θ sin θ cos(φ − φ ) Hence. φ )ρ(r )d3 r
Further cleaning this up.2.68)
Now.72)
.70)
So that its conjugate is just (also changing its arguments): C ∗m (θ . φ)
r Y ∗ (θ . we had that the potential may be written: V (r) = 1 4πε0 1 r
+1
r P (cos γ)ρ(r )d3 r
(2.
4 Relations Between Multipoles in Cartesian & Spherical Polars (3xi xj − r 2 δij )ρ(r )d3 r (2. .j=1
Where we have used the usual monopole and dipole expressions. . ..73)
Let us compute some components of the multipole: Q
m
≡
r C ∗m (θ . . we had: p =
2
ELECTROMAGNETIC FIELD EQUATIONS
r P (cos γ)ρ(r )d3 r
m
Which is the axially symmetric version of the Q
above. using this as the expansion for a multipole: 1 q p · r 1 V (r) = + 3 + 4πε0 r r 2
3
qij xi xj + . r
Where we have used the standard expansion of the exponential:
∞
ex =
n=0
(−1)n n x n!
Now.7. we can evaluate some of the terms in the expansion: 1 1 r · r 3(r r)2 − r 2 r2 = + 3 + + .46 Remember. φ ) 2 +1 m
. . φ )ρ(r )d3 r
Where the coeﬃcients are given by (infact. in rectangular coordinates. . + (r · r n! 1 )n + . and do not derive. we make the ‘connection’ that: 1 1 = e−r · r − r  r So that the expansion is just: e−r · 1 1 = − (r · r r 1 1 ) + 2 (r · r 1 (−1)n )2 + . before.
Now. 2.. as well as the ‘new’ quadrupole tensor : qij = Which we state. but we won’t go into that here): C ∗m = 4π Y ∗ (θ . its worth noting that they are the associated Legendre polynomials. r − r  r r 2r5 Then. r5
i.
2.7
Multipoles This is the component Q00 . So, we look up the spherical harmonic: 1 ∗ ⇒ Y00 = Y00 Y00 = √ 4π Q00 = √ 1 4π √ ρ(r )d3 r 4π ρ(r )d3 r = q
47
Monopole Moment
Then, we have:
Which is just the total charge: Q00 =
Dipole Moment
Let us look up the following spherical harmonics: Y10 = 3 3 cos θ Y11 = − sin θeiφ 4π 8π can be found from the relation: Y
−m
And the ﬁnal harmonic, i.e. Y1−1
= (−1)m Y ∗ m 4π 3 3 cos θ ρ(r )d3 r 4π
Hence, let us start to compute things: Q10 = = = pz Where we have noted that r cos θ = z, as the standard conversion between spherical polars & cartesian. Thus, we see that Q10 is the zcomponent of the dipole moment. Next: Q11 = r 4π 3 − 3 sin θ e−iφ 8π ρ(r )d3 r r
r ρ(r ) cos θ d3 r
1 r sin θ e−iφ ρ(r )d3 r = −√ 2 1 = −√ r sin θ (cos φ − i sin φ ) ρ(r )d3 r 2 1 = −√ r sin θ cos φ ρ(r )d3 r − i r sin θ sin φ ρ(r )d3 r 2 1 = − √ (px − ipy ) 2 Again, where we have noted the use of the standard polarscartesian conversion. Finally: Q1−1 = = = r 4π 3 3 sin θ eiφ 8π ρ(r )d3 r
1 √ r sin θ (cos φ + i sin φ ) ρ(r )d3 r 2 1 √ (px + ipy ) 2
48
2
ELECTROMAGNETIC FIELD EQUATIONS
Doing these, we must be careful that we conjugate the spherical harmonic before putting into the integral.
Quadrupole Moment Here, we compute a component of the quadrupole moment. We look up the following spherical harmonic: Y20 = 1 2 4 (3 cos2 θ − 1) 5π
Q20 = = = =
4π 4 1 r2 (3 cos2 θ − 1)ρ(r )d3 r 2 5 5π 1 [3(r cos θ )2 − r 2 ]ρ(r )d3 r 2 1 (3z 2 − r 2 )ρ(r )d3 r 2 1 q33 2
That is, we ﬁnd that it is related to an element of the quadruple tensor.
2.7.5
Properties of Multipoles
We have see that the details of a charge distribution in a volume, are encoded into the multipoles Q m . We have seen that: • • • • = 0 is the monopole moment: 20 = 1 charges; = 1 is the dipole moment: 21 = 2 charges; = 2 is the quadrupole moment: 22 = 4 charges; = 3 is the octupole moment: 23 = 8 charges.
We have also seen that Q m depend on the choice of origin; however, the ﬁrst nonzero moment is frame independent. The far ﬁeld potential is dominated by the ﬁrst nonzero moment.
2.8
Multipole Expansion of the Vector Potential
As we have previously seen, under the Coulomb gauge, we have that the magentic vector potential satisﬁes the Poisson equation:
2
A = −µ0 J
(2.74)
2.8
Multipole Expansion of the Vector Potential
49
Which, as we have seen, has solution: A(r) = µ0 4π J (r) 3 d r r − r  (2.75)
Thus, exactly as we did for the scalar potential, we may expand this: A(r) = µ0 1 4π r J (r )d3 r + r · r3 r J (r )d3 r + . . .
This is a bit more transparent if we consider a single element of the potential: Ai (r) = µ0 1 4π r Ji (r )d3 r + r · r3 r Ji (r )d3 r + . . . (2.76)
Now, looking at the monopole term: Ji (r )d3 r = 0 Which is zero (no monopoles). This can also be argued from vector calculus ground; which won’t be done here. Looking at the dipole term (the second): r· r Ji (r )d3 r = xj xi Ji d3 r
1 = − xj (xi Jj − xj Ji )d3 r 2 1 = − ijk xj (r × J )k d3 r 2 1 = − r × (r × J )d3 r 2 i Hence, we have an expression for the magnetic dipole moment: m= 1 2 r × J (r )d3 r (2.77)
So that the dipole component of the magnetic vector potential is just (using a × b = −b × a): A(r) = µ0 m × r 4π r3 × A, is just: (2.79) (2.78)
Thus, the magnetic ﬁeld induced by such a vector potential, from B = B(r) = Where we have used the vector identity: × (rn a × r) = rn−2 [(n + 2)r2 a − n(a · r)r] We shall stop with the magnetic analysis here. µ0 3(m · r)r − r2 m 4π r5
then we must appeal to a sum over spherical harmonics. We must use ‘common sense’ boundary conditions (it is usual that these are not stated in questions.80)
Where the angle between the distribution and observation point is γ ≡ θ − θ. one is then able to express one coeﬃcient in terms of the other. That is.9
Multipole Expansions Summary
If we have some axiallysymmetric charge distribution. this immediately eliminates one coeﬃcient from each expression. one may either use linear independence of the coeﬃcients of Legendre polynomials. the potential does not diverge. Upon inspection of the potential expressions for inside & outside. The ﬁrst expression is that the perpendicular component of the electric displacement ﬁeld is discontinuous in the presence of surface charge. E and φ at the boundary.50
2
ELECTROMAGNETIC FIELD EQUATIONS
2. and uses this as a boundary condition. Continuing. with observation point at r. then.81)
So that the potential will be a sum over poles: 1 V (r. the solution in Spherical Polars will be used. as opposed to Legendre polynomials. To actually do this. at r . which will be given by: p = r P (cos γ)ρ(r )d3 r (2.10
Method: Potentials and Surface Charges
Suppose we have the problem of ﬁnding the scalar potential in all space. in the presence of some symmetric charge distribution. This must then be divided into two separate expressions: inside and outside. The second and third expressions just follow from the ﬁrst. by using the fact that the potential is continuous at the boundary.
2. we are able to expand the distribution in terms of multipoles. or (possibly the more complete method) use orthogonality: multiply the equation by ‘another’ polynomial. but must be used). the charge distribution is not axiallysymmetric. It can be useful to think about the pole term. at the origin and inﬁnity. however. Usually. and read oﬀ the relation directly. to ﬁnd φ. θ) = 4πε0
∞ =0
1 r
+1
p
If. If there is an applied ﬁeld. so that the scalar potential may be written: 1 V (r. That is: Dout − Din = ρ εout ε0 Eout − εin ε0 Ein = ρ εout ε0 ∂φout ∂φin − εin ε0 = −ρ ∂r ∂r φout = φin
Where all are only valid at the boundary. using standard relations to link. then one computes the potential (essentially) at inﬁnity. The ﬁrst thing to do is to write the relations between D. NOTE: this cannot be used if there is an applied electric ﬁeld.
. θ) = 4πε0
∞ =0
1 r
+1
r P (cos γ)ρ(r )d3 r
(2.
.11
Discussion
51
To then ﬁnd the remaining coeﬃcient. We also showed how much freedom we have in choosing the ﬁelds. the scalar ﬁeld in all space may be found. we found a way of expressing a continuous distribution of charge. where there are regions of no charge density. We then looked a little at the eﬀect of matter & materials on the ﬁelds. express the surface charge density in terms of Legendre polynomials. Then we discussed the solution to Laplaces equation. then the exact forms of the coeﬃcients can be read oﬀ. but continued with removing the axial symmetry (using spherical harmonics: a fuller treatment of spherical harmonics & various related theorems may be found in the appendix). polar & spherical polar coordinates. We discussed its solution in Cartesian. Faradays law & Amperes law. no magnetic monopoles. which arises in electrostatic systems. we are able to ﬁnd some properties of the potentials.11
Discussion
So. We now move on to considering systems in which charges move.2. due to speciﬁc distribution of charges. in terms of their associated potentials. only an approximate ‘inﬁnite sum’ will result as the coeﬃcients. Otherwise. which will make the next step a lot easier to do. and various consequences of the motion. employing the Lorentz gauge to derive wave equations which are correct for timevarying ﬁelds. We then took a mathematical diversion. in light of the continuity equation. Then. and the energy ﬂux of a ﬁeld. We found that we must modify Amperes law. This concludes our discussion of electrostatic systems. Next. which allowed us to derive a wave equation for static ﬁelds only. We discussed the electrostatic Coulomb gauge. which gives information on the energy of ﬁelds. with solving some special cases with a speciﬁc set of boundary conditions. Hence. in this way. let us review this section. at the boundary. As well as a brief note on Green functions. one will see that if the surface charge is given by a ﬁnite number of polynomials. If possible. looking at the Diracδ function. in terms of invariance. We started by introducing Maxwell’s four equations: Gauss’ law. Finally.
2. We carried on with looking at Poynting’s theorem. with boundary conditions for the presence of surface charge. This method works even if there is a ﬁeld present. but care must be taken over the ‘inﬁnite’ boundary condition. we looked at time varying ﬁelds. with a speciﬁc case given. We started with axially symmetric systems (using Legendre polynomials). By writing the potentials as a sum like this. and use orthogonality. one must multiply the whole equation by ‘another’ polynomial. dipoles etc. and their eﬀect on the ﬁelds. to take account of time variation of ﬁelds. and various properties & uses. Then. and discuss the concepts introduced. by approximating it to a series of monopoles. once the derivatives done. over a time period. use the discontinuity of the derivative of the potential.
52
2
ELECTROMAGNETIC FIELD EQUATIONS
.
2)
Using this in both (3. we have: 1 ∂2A − c2 ∂t2 Now.5)
2
A = µ0 J
. Consider the following Maxwell’s equations.4) (3.3)
2
×A=
(
· A) −
2
A
c2 =
1 ε0 µ 0
A+
·A+
1 ∂V c2 ∂t
= µ0 J
(3.1)
And putting both into Amperes law: × × A = µ 0 J + ε0 µ 0 ∂ ∂t − V − ∂A ∂t
Let us use the following vector identity. for timevarying ﬁelds: ·E = Consider also: E=− V − So. and relation: × Then.1
Retarded Potentials & Radiation
Introduction to Radiation from Accelerated Charges
Let us begin by considering the derivation of some wave equations.53
3
3.2) gives: 1 ∂2V − c2 ∂t 1 ∂2A − c2 ∂t2
2
V
=
ρ ε0
(3.1) and (3. These will be waves ‘of potential ﬁeld’. inserting the ﬁrst into Gauss’ law: · − V − Giving:
2
ρ ε0
× B = µ 0 J + ε0 µ 0
∂E ∂t
∂A ∂t
B=
×A
∂A ∂t
=
ρ ε0
V +
∂ ∂t
·A=−
ρ ε0
(3. if we use the Lorentz gauge: ·A+ 1 ∂V =0 c2 ∂t ⇒ ·A=− 1 ∂V c2 ∂t (3. driven by ‘charge distributions’.
decoupled the equations into two inhomogeneous wave equations. is because the radiation is interacting with an absorber. t) = 1 4πε0 µ0 A(r. if we are at a time t.e. Hence: c 1 1 tadv = t + r − r  tret = t − r − r  c c Now. As an aside. it ‘knew’ that it would be interacting with material (i.7)
Where. the retarded distribution) at r . the the time taken for information to get from the charge distribution (as it was. after a fair amount of algebra. then we see the charge distribution as it was at some time tret . which means that the eﬀect would precede the cause. at a star. with noone around to listen. That is. We must now come to the concept of retarded time. we digress: Now. the observers eye) 4 year in its future. It is presumably like saying that a tree does not make a sound if it falls in a forest. So. t) = (3. t) = qδ(r − r0 (t)) J (r. Anyway. The solutions to these equations are given in terms of retarded scalar and vector potentials. That is. Then. in the past. r is the vector from the retarded position to the origin. t) = 4π ρ(r . to the observer (at time t) is just 1 r − r . That is.8) 4πε0 κR ret µ0 cβ A(r. not how it is now. Which is a violation of causality. due to the ﬁnite speed of light. the observer is at r. and charge distribution of interest (i.Phys Vol17. as the distance between observer and charge distribution is r − r . a potential ﬁeld now would be due to the motion of charges in the future. If we are at a point in space (i. by Feynman & Wheeler1 that the potential now is actually a sum of the advanced & retarded potential.54
3
RETARDED POTENTIALS & RADIATION
Hence.Feynman Rev. with an observer. 1945
1
. which is moving. it would be correct to say r (tret ). and we observe a charge distribution. which is (say) 4 light years away. So. we observe the distribution as it was when it emitted the light. The consequence of this is that (and this is an example they give) if we look out of the window. with a +sign. and halved. we end up with the LienardWiechart potentials: 1 q V (r. t) = (3. which we will do later.Mod. suppose we consider point charges: ρ(r. when the light was emitted (4 years ago). tret ) 3 d r r − r  J (r . The cause is after the eﬀect.6) (3. as usual. is consistent.e. we have solutions (which we shall later verify) to the scalar & vector potential: V (r. we have. the observer). it has actually been considered. That is. at tret ).Wheeler & R. t) = qv(t)δ(r − r0 (t)) Suppose that the charge distribution is travelling at velocity βc (at retarded time). using the Lorentz gauge. Now. the solution to the right.e. the only reason it emits radiation. then. No2. one supposes.9) 4π κR ret
Interaction with the Absorber as the Mechanism of Radiation: J. tret ) 3 d r r − r  (3. They review a suggestion that radiation could be due to interaction of the radiation. but would correspond to an advanced time.
then A = 0. the total radiation ﬂux is the 2 r ∝ 1 . for any charge which is moving at a constant speed. it does not radiate. for nonrelativistic motion of charges: Evel = Erad = ˆ Where κ ≡ 1 − β · R. P ). The equations reﬀered to are: Erad = q 4πε0 ˙ ˆ ˆ R × [(R − β) × β] ˆ c(1 − β · R)3 R Evelocity = q 4πε0 ˆ (R − β)(1 − β 2 ) ˆ c(1 − β · R)3 R2
Both evaluated at the retarded time. whereas the 1 ‘acceleration ﬁeld’ is proportional to R . the Poynting vector P = E × H (which is the energy density ﬂow. So. P = 0. and one dependant upon acceleration: cβ. because the velocity term decays. Now.e. an inertial frame of reference can be found in which it is stationary. So. Notice. the ﬁrst term is the acceleration ﬁeld. hence. if R ≡ r − r : R ˆ ˆ R= κ=1−R·β R So. per unit area). Now. This is the radiation per unit area. Hence. static charges do not radiate. for the acceleration ﬁeld. we see that there will be a strong angleˆ dependance. for velocities close to the speed of light (i. due to R · β = βc cos θ. we ﬁnd a (electric) ‘velocity ﬁeld’. Hence. β ≈ 1). if we do this for the acceleration ﬁeld. for low β. as magnetic ﬁeld is proportional to electric ﬁeld. for speeds close to that of c. Thus. only the acceleration terms will radiate. That is.e. That is. the radiation ﬁeld. we know that if a particle is moving at constant velocity. we see that P ∝ R2 ⇒ P d2 r → 1. which (at the end of our discussion) we call the radiation ﬁeld.3. i. Now. Let us look at the radial dependancies of the energy ﬂux (i. These equations are merely stated here. Hence. due to our previous arguement.1
Introduction to Radiation from Accelerated Charges
55
Where. the radiation ﬁeld. So. If a charge is static. Hence. how does this eﬀect the radiation of energy. then an inertial frame of reference can be found in which the particle is stationary. once the potentials are converted into ﬁelds. we refer to the acceleration term as the radiation ﬁeld. Now. which is proportional to R2 . hence. of the radiation. but the acceleration term stays constant. we see that for the velocity ﬁeld. a strong dependance on the angle that the moving charge is viewed. integral P d R2 1 Now. up 1 1 1 ˆ to B = 1 k × E. Hence. is ∝ R4 . the radiation ﬁeld for the velcity ﬁeld decays to zero. P ∝ R2 R2 = R4 . for c 1 the velocity ﬁeld. we will ﬁnd that there are two terms: one ˙ dependant upon velocity: cβ. Consider the following magnitudes. Hence.e. if 1 the maths is done. any charge moving at a constant velocity does not radiate. Hence. Only accelerating charges radiate. is a constant. taking the ratio: Erad R˙ = β Evel c 1 q 3 R2 4πε0 κ ˙ q β 4πε0 cκ3 R
. but will be discussed in detail later.
this is the power radiated per unit area.e. So. Also. then there would be no radiation. we have arrived at Larmor’s formula: P = ˙ q2 β2 sin2 θ 16π 2 ε0 c R2 (3. as we will see later. we have that. Then. 3. inserting our expressions in: P = = ˙ β2 1 q2 sin2 θ µ0 c (4πε0 c)2 R2 ˙ β2 q2 sin2 θ 16π 2 ε0 c R2
Hence. β = 0). consider radiation with some characteristic frequency of oscillation. per unit time. Now. then no radiation. So. if c is inﬁnite. the total power radiated. in this limit: ˙ q β sin θ Erad = 4πε0 c R Where θ is the angle between the direction of acceleration and the observer. and radiation in the far zone. we have that β << 1. per unit time. then we get: Erad β =R Evel λ Hence.1. Also. we also have: 1ˆ Brad = Rret × Erad c Hence.56
3
RETARDED POTENTIALS & RADIATION
˙ From which is see that if no acceleration (i. we see that the velocity ﬁeld dominates in the near zone. the Poynting vector: P = = = ⇒ P  = P = 1 E×B µ0 1 ˆ Erad × Rret × Erad µ0 c 1 2 ˆ ˆ Erad Rret − (Erad · Rret )Erad ) µ0 c 1 2 E µ0 c rad
ˆ Where we have used that Erad and Rret are perpendicular. via c = νλ. is just: dW ¯ ≡ P = P R2 dΩ dt
.1 Example: Larmor’s Formula
Consider the radiation from a nonrelativistic particle. So.10)
Now.
if c → ∞. and not velocity. and we do so for the scalar potential. Note.13)
Where tret = t − 1 R. under Larmors assumptions: ¯ PL = q2 ˙ 2 β 6πε0 c (3. t) = 1 4πε0 µ0 A(r. Hence. we see a sin2 θ dependance. V : 1 ∂2 ρ(r. t − 1 R) 3 c d r R
Now.2 Retarded Potentials & the Wave Equation
Now. we stated the retarded potentials: V (r. no power is radiated along the direction of acceleration. tret ) 3 d r r − r  (3. Hence. in the previous section. t) 2 − 2 2 V (r. even though we have assumed β << 1. tret ) 3 d r r − r  J (r . The power radiated must depend only on the acceleration. t) = 4π ρ(r . the above are: c V (r. Again. Thus. then power radiated goes to zero. with R ≡ r − r .12) (3.1
Introduction to Radiation from Accelerated Charges
57
That is: ¯ dW P = dt = q2 ˙ β2 16π 2 ε0 c
π θ=0 1
sin2 θ 2 R sin θdθdφ R2
= Γ2π = Γ2π
−1
sin3 θdθ 1 − x2 dx
4 = Γ2π 3 q2 ˙ 2 β = 6πε0 c We shall denote this expression for the total power radiated. in Larmor’s formula. our intuition is satisﬁed. we wish to show that they do indeed satisfy the wave equations. t − 1 R) 3 c d r R J (r . 3. that c is ﬁnite is important. the sign of the charge must be even (i. just to check that this satisﬁes our intuition: as we cannot have negative power. t) = 4π ρ(r .11)
Hence. That is. t) = − c ∂t ε0
.3.1. t) = 1 4πε0 µ0 A(r.e. That is. radiation is a relativistic eﬀect. the q 2 term).
t) c Thus. t − 1 r − r )d3 r = ρ(r. the point of computing this was to be able to write:
2 2
V =
1 4πε0
2 ρ(r
. we may represent this as a deltafunction: ρ(r . we also have the result:
2
1 = −4πδ(R) = −4πδ(r − r ) R
Now. to proceed we compute the following Laplacian:
2
ρ =ρ R
2
1 1 + R R
2
ρ
Now. t ) = qδ(r − r0 (t )) That is. t − 1 R). when it was. r0 (t ) is some position within the charge. So. we consider the scalar & vector potentials generated by a moving point charge q. its wave equation is:
2
ρ=
1 ∂2ρ c2 ∂t2
Hence:
1 1 ∂2ρ ρ = − 4πρδ(r − r ) R R c2 ∂t2 Now. t − 1 R) 3 1 c d r − R ε0
2
1 1 ∂2ρ − 4πρδ(r − r )d3 r R c2 ∂t2
That is:
2
V =
1 1 ∂2 4πε0 c2 ∂t2
δ(r − r )ρ(r . as we have that ρ(r . we know it must be a solution to a wave equation (as that what c its solutions look like). t ). Consider a point charge. That is.2
LienardWiechert Potentials: Point Charges
Here. t − 1 R)d3 r c
Which is just: V = 1 ∂2V 1 − ρ(r. We have used that it is at the retarded time: t = t − 1 r − r  c
. we have shown that the proposed representation of the retarded scalar potential does indeed satisfy the wave equation. having some coordinates ρ(r . we consider it at some (retarded) position & time: where it was.
3. t − 1 R) 3 1 c d r = R 4πε0 ρ(r . Thus. t) c2 ∂t2 ε0
As we have noted that: δ(r − r )ρ(r .58
3
RETARDED POTENTIALS & RADIATION
Now.
t) = q 4πε0 1 δ(τ − t )δ(r − r0 (τ )) dτ d3 r r − r 
Notice. we note that t is a function of r: V (r. t) = q 4πε0 1 δ(τ − (t − 1 R(τ )))dτ c R(τ )
Let us continue by noting some diracdelta theory: δ(f (x)) =
i
δ(x − xi f (xi )
Where the function f (x) has zeros at xi . the potential is given by: V (r. Then. the reason for using this will (hopefully) become clearer. the above is: V (r. this integral will only allow the value τ = t in the position deltafunction. Then. We rewrite it using a delta function for the retarded time: ρ(r . Then: c 1 dR df =1+ dτ c dτ And it has zeros at: 1 τ = t − R(τ ) ≡ t c
. This integral obviously evaluates to exactly what we had before. as one can imagine. t) = 1 4πε0 ρ(r .3. Thus. t) = q 4πε0 1 δ(τ − t )dτ r − r0 (τ )
To do this integral. the d3 r integral will just have the eﬀect of sending r → r0 (τ ). as the retarded time is a function of the integration variable: q 1 V (r. t ) = q δ(τ − t )δ(r − r0 (τ )) dτ t ≡ t − 1 r − r  c
Notice then. we have f (τ ) = τ − t + 1 R(τ ). integration here is complicated. we look at the charge distribution again.2
LienardWiechert Potentials: Point Charges
59
Now. t) = q 4πε0 1 δ(τ − (t − 1 r − r0 (τ )))dτ c r − r0 (τ )
Let us denote R(τ ) ≡ r − r0 (τ ). Thus: V (r. in our case. t) = δ(r − r0 (t ))d3 r 4πε0 r − r  To go further. if we put this into the potential integral above: V (r. t ) 3 d r r − r 
So.
we see that the velocity of the beam is present: dr0 = v = βc dτ Hence: 1 dR R·β ˆ =− = −R · β c dτ R Thus shown.60
3
RETARDED POTENTIALS & RADIATION
Infact. Hence. we use that: J (r . t) = Now. our potential is: V (r. t ) µ0 4π J (r . we shall call it t0 . we have the LienardWiechert scalar potential : V (r. let us use the chain rule: dR dr0 dR = · dτ dr0 dτ So: dR R d r − r0 =− = r − r0  = − dr0 dr0 r − r0  R =−
τ =t0
1 1+
1 dR c dτ (τ =t0 )
q 4πε0 1 +
1
1 dR c dτ (τ =t0 )
1 δ(τ − t0 )dτ R(τ )
q 4πε0 1 +
1
1 dR c dτ τ =t0
1 R(t0 )
β·R R
τ =t0
And also. t) = q 4πε0 1 ˆ R(1 − β · R)
ret
(3. t) = q 1 1 R·β R 4πε0 1 −
R
Where everything is evaluated at τ = t0 . our integral for the potential becomes just: V (r. we shall show that: 1 dR c dτ So. we can ﬁnd the vector potential: A(r. Hence. t) = Which is easily evaluated to be: V (r. t ) 3 d r r − r 
. So. t ) = cβ(t )ρ(r .14)
By very similar considerations. using this: δ(τ − (t − 1 R(τ ))) = δ(τ − t0 ) c Then. Thus. this t is the same ‘class’ of coordinate as r0 . t) = Here.
2
LienardWiechert Potentials: Point Charges
61
Then use the exact same argument.18) q 4πε0 ˙ ˆ ˆ ˆ (R − β)(1 − β 2 ) R × ((R − β) × β) + ˆ ˆ (1 − β · R)3 R2 c(1 − β · R)3 R (3.15)
Now. from these potentials. one obtains the the LienardWiechert ﬁelds: E = B = Where: d ˙ β(τ ) β≡ dτ (3. R2 R The energy radiated by a moving charge. We can also recover a previously known relation: 1ˆ B = Rret × E c 3. and velocity have R2 .17)
˙ ˆ ˙ ˆ ˆ ˆ µ0 qc (β × R)(1 − β 2 ) β · R(β × R) β×R + + ˆ ˆ ˆ 4π (1 − β · R)3 R2 c(1 − β · R)3 R c(1 − β · R)2 R
Notice. Giving the LienardWiechert vector potential : A(r. shifting reference frames dosent change anything it shouldn’t! For point charges. Bv ∝ β. We have seen that we can decompose the ﬁelds into velocity and acceleration components: E = Ev + Ev Where: Ev . That is. B = Bv + Ba
1 ˙ 1 Ea . we are able to calculate the ﬁelds.1 Features of LienardWiechert Potentials (3.3.the are Lorentz covariant.16) (3. t) = qµ0 4π cβ ˆ R(1 − β · R)
ret
(3.20)
. for deltafunctions. per unit time is given by: ˆ ˆ ¯ dP = P · RdA = P · RR2 dΩ This goes to: ¯ dt dP ˆ = lim P · RR2 dΩ R→∞ dtret (3. from: E=− V − ∂A ∂t B= ×A
After some (!!) algebra. the magnetic ﬁeld is always perpendicular to the electric ﬁeld. we have the accleration & velocity terms in both the electric and magnetic ﬁeld.2. The 1 1 acceleration ﬁelds are those with a R dependance. Ba ∝ β.19)
The potentials are relativistically correct .
the only nonvanishing component. the above is just: E= q (1 − β 2 ) ˆ (Rr − β) 2 ˆ 4πε0 (1 − β · Rr )3 Rr (3. Thus. The velocity term does have an associated energy .22) (3.2 Example: Particle Moving With Constant Velocity
So.24)
. 3. as R → ∞ is Paa . The components of the Poynting vector are: Pvv ∝ Ev × Bv  ∝ Pva Paa 1 R4 1 ∝ Ev × Ba  ∝ 3 R 1 ∝ Ea × Ba  ∝ 2 R (3.23)
And thus. positions at the retarded time are denoted with a subscript r.
˙ If we have that the particle moves with constant velocity. Thus. per unit solid angle.
Figure 7: The setup for retarded motion. Let us consider an example. One may think of the lackof radiation of the velocity components. only the velocity term of the LienardWiechert ﬁeld contributes: E= q 4πε0 ˆ (R − β)(1 − β 2 ) ˆ (1 − β · R)3 R2
ret
With reference to the ﬁgure.2. as the energy being ‘convected’ along with the beam. we have a particle moving with constant velocity cβ.21) (3. per unit time.62
3
RETARDED POTENTIALS & RADIATION
As the energy radiated. then obviously β = 0.it is linked to the kinetic energy of the beam.
let us continue. between tret and t is just Rr . rather than at the retarded time.24) gives: E= (1 − β 2 ) q Rp 3 ˆ 4πε0 (1 − β · Rr )3 Rr (3. we notice that the vertical in the two triangles is the same.25)
Thus. with reference to the ﬁgure. but due to where it was! So. with reference to the ﬁgure. We do this by careful vector algebra. combining the above two results: Rp = Rr − Rr β ˆ That is.2
LienardWiechert Potentials: Point Charges
63
Now.3. we also have: Rr = c(t − tret ) Thus. the only ‘vector’ left (and hence direction). the distance the particle travels is just: cβ(t − tret ) Hence. the distance a signal travel. Hence.28)
Now. let us square the lefthand expression in (3. Thus. is actually the actual position Rp . Now.26) (3. rather than the retarded position Rr . using the standard R = RR. at speed c. That is. by deﬁnition. we may then write: Rr = cβ(t − tret ) + Rp Which is (obviously) completely the same as writing: Rp = Rr − cβ(t − tret ) Also. So: Rr × β2 = Rp × β2 (3.29)
. just: ˆ Rp = Rr (Rr − β) Hence. and get the denominator in terms of ‘now’ variables. Consider the distance the particle travels between tret and t. using this in (3. the electric ﬁeld is in the direction of where the ﬁeld is now.27) ⇒ ˆ (Rr − β) = Rp Rr (3.26):
2 2 ˆ Rp = Rr (Rr − β)2 ˆ = R2 (1 − 2Rr · β + β 2 ) r 2 2 ˆ 2 = Rr − 2Rr Rr · β + Rr β 2
That is:
2 2 2 ˆ 2 Rp = Rr − 2Rr Rr · β + Rr β 2
(3. the main goal is to get everything in terms of positions now. This is just its velocity multiplied by the diﬀerence in times.
27): E = = = q (1 − β 2 ) Rp 3 ˆ 4πε0 (1 − β · Rr )3 Rr q (1 − β 2 ) R 3 p 4πε0 (1 − β 2 sin2 θ)3/2 Rp q (1 − β 2 ) ˆ R 2 p 4πε0 (1 − β 2 sin2 θ)3/2 Rp
.
2 2 2 ˆ Rr (1 − β · R)2 = Rp + (Rp β cos θ)2 − Rp β 2 2 2 = Rp + Rp β 2 (cos2 θ − 1) 2 2 = Rp − Rp β 2 sin2 θ 2 = Rp (1 − β 2 sin2 θ)
Hence. which we have then used.27) (it is the same. getting some angular dependance. for the third and last terms:
2 2 2 ˆ 2 ˆ Rr (1 − β · R)2 = Rp + 2Rr Rr · β − 2Rr β · Rr + (Rp · β)2 − Rp β 2 2 ˆ Noting that two terms (note that 2Rr Rr · β = 2Rr Rr · β) cancel gives: 2 2 2 ˆ Rr (1 − β · R)2 = Rp + (Rp · β)2 − Rp β 2
Now. we can use this in the denominator of (3. We have noted that the deﬁntion of the dotproduct comes out. we have:
2 2 (Rr · β)2 − Rr β 2 = (Rp · β)2 − Rp β 2
Substituting this into (3.28) into the above. We can then easily see that (3. if we take the 3/2 power of the above). for Rr .29) is just:
2 2 Rr β 2 − (Rr · β)2 = Rp β 2 − (Rp · β)2
(3.31)
If we rearrange (3. the above is purely in terms of ‘now’ ! Let us expand this out now.30) slightly. the LHS of the above is ‘almost’ the same factor as we have in the denominator of the electricﬁeld expression (3. let us compute the following:
2 ˆ Rr (1 − β · R)2 = (Rr − β · Rr )2
Which gives:
2 2 ˆ Rr (1 − β · R)2 = Rr − 2Rr β · Rr + (β · Rr )2 2 Now.31). And also. giving: 2 2 2 ˆ 2 ˆ Rr (1 − β · R)2 = Rp + 2Rr Rr · β − Rr β 2 − 2Rr β · Rr + (β · Rr )2
(3.64 Let us expand the LHS. substitute (3. carefully:
3
RETARDED POTENTIALS & RADIATION
Rr × β2 = [Rr β sin θr ]2
2 = Rr β 2 sin2 θr 2 = Rr β 2 (1 − cos2 θr ) 2 = Rr β 2 − (Rr · β)2
Where θr is the angle between β and Rr .30)
Now. notice.
we have that: Rr = cβ(t − tret ) + Rp And hence: Rr cβ(t − tret ) Rp ˆ Rr = = + Rr Rr Rr
We also had that Rr = c(t − tret ).2
LienardWiechert Potentials: Point Charges
65
Thus. and the angle in which the particle is viewed: E= q (1 − β 2 ) ˆ R 2 p 4πε0 (1 − β 2 sin2 θ)3/2 Rp
Also. Rp and E are along the same line. from (3. We have found that they do not depend upon retarded time. that term is zero. we have computed the electric ﬁeld in terms of the present position of the particle. Hence: Rp ˆ Rr = β + Rr And therefore: B= 1 c β+ Rp Rr ×E
Now. q (1 − β 2 ) ˆ (β × Rp ) 2 4πε0 c (1 − β 2 sin2 θ)3/2 Rp
. but only on the time that the particle is in presently.25). using our electric ﬁeld: B= Also. we can ﬁnd the magnetic ﬁeld. Hence. to summarise what we have done: we have found the electric and magnetic ﬁelds due to a moving particle. Thus: 1 B = β×E c Hence. from the following expression: 1ˆ B = Rr × E c Now.3. note that: 1 µ0 c = 4πε0 c 4π Hence.
but before then.34). this is just: P = 1 2 ˆ E [R]ret µ0 c a (3. which gives: P = However. as the power P is dP = P · RR2 dΩ . we can see that both Ea and Ba are perpendicular to R. from these expressions.3
Radiation
Let us consider the radiation emitted by a particle moving. as we know: P = Thus.32)
ret
With the corresponding (acceleration components of) magnetic ﬁeld being: 1 ˆ Ba = [R]ret × Ea c (3.66
3
RETARDED POTENTIALS & RADIATION
3. 3. going through some bit of surface dA. we shall derive the general form of radiation from a moving charge. using (3.1 General Theory of Radiation
Now.3.35) 1 2 ˆ ˆ Ea [R]ret − (Ea · [R]ret )Ea µ0 c
Now.34)
Let us use the vector identity a × b × c = (a · c)b − (a · b)c.33)
ˆ Now. per unit time is: dW ˆ = P · dA = [P · RR2 dΩ]ret dt ˆ ¯ ¯ Now. That is: ˆ Ea · R = 0 Now. the radiation per unit area is given by the Poynting vector. from (3. We shall consider the two cases of velcity and acceleration being parallel and perpendicular. the amount of energy dW . we know that radiation is only emitted from the acceleration component of the electric ﬁeld: Ee = q 4πε0 ˙ ˆ ˆ R × (R − β) × β ˆ (1 − β · R)R (3.33) in the above expression: P = 1 ˆ Ea × [R]ret × Ea µ0 c 1 Ea × Ba µ0 (3.the amount of ‘poynting vector’ going through an area. we obviously have: ¯ dP ˆ = [P · RR2 ]ret dΩ
.
as we can convert between retarded times thus: ti = ti + Where we have previously seen that: 1 dR ˆ = −[R · β]ret c dt Hence. how much energy is lost? We can obviously write: t2 dW ˆ = [P · RR2 ]ret dt dΩ t1 Hence. this is just: ˙ ˆ ˆ ¯ dP q2 R × (R − β) × β = ˆ dΩ 16π 2 ε0 c (1 − β · R)5
2
ret
ret
(3.35) and (3. Let us now consider colinear acceleration:
. we derived P in the comoving frame of the charge q. if the charge is accelerated between times t1 → t2 . upon comparison of the LHS and previous expressions.3. but the result is valid for any frame. upon comparison of the two above expressions. using (3.32).36)
¯ Now. let us change integration variables: dW dΩ
t2
R(ti ) c
⇒
1 dR dt =1+ dt c dt
=
t1 t2
dt ˆ [P · RR2 ]ret dt dt ˆ ˆ [P · RR2 ]ret 1 − [R · β]ret dt
=
t1
Therefore. Thus. we can write: d2 W ˆ = [P · RR2 ]ret dtdΩ Now. ¯ including the rest frame of the lab.3
Radiation
67
Hence. P is invariant under Lorentz transformations. divinding out the dt : d2 W ˆ ˆ = (P · RR2 )(1 − R · β) dt dΩ
ret
And this. is: ¯ dP ˆ ˆ = (P · R)(1 − R · β)R2 dΩ Hence.
Hence. Hence. using this: 2 ˙ ˆ ˆ R×R×β ˆ (1 − β · R)5
ret
And. this immediately tells us that: ˙ β×β =0 As is the condition for two things being parallel.2
3
RETARDED POTENTIALS & RADIATION
Radiation: Acceleration & Velocity Parallel
This is sometimes denoted colinear motion.3. we have derived the angular distribution of radiation. to see that: ˙ ˙ ˙ ˆ ˆ ˆ R × R × β = β cos θR − β Thus. using this: ¯ dP q2 β˙2 sin2 θ = ˆ dΩ 16π 2 ε0 c (1 − β · R)5 (3. let us ﬁnd the angle at which most radiation is emmited. and diﬀerentiate: d sin2 θ 2 cos θ sin θ 5β sin3 θ = − =0 dθ (1 − β cos θ)5 (1 − β cos θ)5 (1 − β cos θ)6 And set it to zero (maximum). let us isolate the θdependance. Solving: [2 cos θ(1 − β cos θ) − 5β sin2 θ] sin θ = 0
2
(3.37)
ret
Hence. We have also derived the form for nonrelativistic motion.38)
a × b × c = (a · c)b − (a · b)c
. we obtain the previously derived Larmor formula: ¯ dP q2 ˙ = β 2 sin2 θ dΩ 16π 2 ε0 c ˙ ˆ Where θ is the angle between β and R. Here.68 3. from a charge moving parallel to the direction in which it is being accelerated. 2 ¯ dP q = 2ε c dΩ 16π 0 Now. To do this.37). with β << 1. squaring it: ˙ ˙ ˙ ˆ ˆ ˆ (R × R × β)2 = (β cos θR − β)2 2 2 2 ˙ ˙ ˙ = β cos θ + β − 2β 2 cos2 θ ˙ = β 2 (1 − cos2 θ) = β˙2 sin2 θ Hence. we have that the acceleration and velocity of the particle are in the same direction. we use a vector identity2 . Using (3.
3. with x ≡ cos θ. from x = cos θ: √ −1 −1 + 1 + 15β θmax = cos 3β Let us ﬁnally ﬁgure out the total power radiated by a nonrelativistic particle.38): ˙ q2β 2 ¯ PL = dφ dθ sin2 θ sin θ 2ε c 16π 0 Which. the maximum (as opposed to the minimum) is.3 Radiation: Acceleration & Velocity Perpendicular
This corresponds to a charged particle being accelerated in a circular orbit. This will correspond ˙ to β · β = 0 Let us start with our general formula for radiation: ˙ ˆ ˆ 2 ¯ dP q R × (R − β) × β = ˆ dΩ 16π 2 ε0 c (1 − β · R)5 Let us.3. gives: ˙ q2β 2 ¯ PL = 6πε0 c 3.3
Radiation
69
Solve this by the quadratic formula. for convenience. with a substitution3 . as the numerator in the expression for radiation is squared. write this in the following way: ¯ dP q2 Ξ2 = ˆ dΩ 16π 2 ε0 c (1 − β · R)5 ˙ ˆ ˆ Ξ ≡ R × (R − β) × β
ret
2
ret
Let us expand out the top. giving: √ −1 ± 1 + 15β x= 3β And thus. using a vector identity as we go: ˙ ˙ ˆ ˆ ˆ Ξ = R×R×β−R×β×β ˙ ˙ ˆ ˙ ˆ ˆ ˙ ˆ = (R · β)R − β − (R · β)β + (R · β)β ˙ ˆ ˙ ˆ ˆ = R · β(R − β) − β(1 − R · β) Hence. We hence integrate (3. let us square our expression (the very last one): ˙ ˆ ˙ ˆ ˆ Ξ2 = R · β(R − β) − β(1 − R · β)
3
2
˙ ˆ ˙ ˆ ˆ = (R · β)2 (R − β)2 + β 2 (1 − R · β)2 − 2ξ
See the appendix for a note on doing such integrals
.
we have that velocity and acceleration are perpendicular. let us take out a common factor: ˙ ˆ ˙ ˆ ˆ ˆ Ξ2 = (R · β)2 (R − β)2 − 2 + 2R · β + β 2 (1 − R · β)2 ˙ ˆ ˆ ˆ ˆ ˙ = (R · β)2 1 − 2R · β + β 2 − 2 + 2R · β + β 2 (1 − R · β)2 ˙ ˆ ˙ ˆ = −(R · β)2 (1 − β 2 ) + β 2 (1 − R · β)2 ˙ ˙ ˆ ˙ ˆ ˙ Now. from the setup of the system. Thus.70
3
RETARDED POTENTIALS & RADIATION
Where ξ is the ‘crossterm’ that always results from squaring a bracket: ˙ ˆ ˙ ˆ ˆ ξ ≡ [(R · β)(R − β)] · [β(1 − R · β)] Let us compute this. using all this. the above becomes: we have that R ˙ ˙ Ξ2 = −(β sin θ cos φ)(1 − β 2 ) + β 2 (1 − β cos θ)2 Also. we see that the denominator term in the radiation expression can be written: ˆ (1 − β · R)5 = (1 − β cos θ)5 Hence. Hence. we have that R · β = β sin θ cos φ. ˙ as previously stated β · β = 0. the squared expression for the numerator is: ˙ ˆ ˙ ˆ ˆ Ξ2 = (R · β)2 (R − β)2 + β 2 (1 − R · β)2 − 2ξ 2 ˆ 2 2 ˙ ˆ ˙ ˆ ˆ ˙ ˆ = (R · β) (R − β) + β (1 − R · β)2 − 2(R · β)2 (1 − R · β) Again. noting this results in the last two terms above being zero. leaving us with the ﬁrst two (which we also clean up): ˆ ˙ ˆ ˙ ˆ ξ = (R · β)2 − (R · β)2 (R · β) Taking out a common factor: ˆ ˙ ˆ ξ = (R · β)2 (1 − R · β) Therefore. is: ¯ dP dΩ = = q2 Ξ2 ˆ 16π 2 ε0 c (1 − β · R)5 ˙ ˙ q 2 −(β sin θ cos φ)(1 − β 2 ) + β 2 (1 − β cos θ)2 16π 2 ε0 c (1 − β cos θ)5
. Also. carefully (remembering that terms that are already ‘dotted’ are a scalar. Hence. our radiation expression. treating R · β as the projection of β onto the xaxis. ˆ · β = β cos θ. for acceleration being perpendicular to velocity. and are just treated as ‘numbers’): ˙ ˆ ˙ ˆ ˆ ξ ≡ [(R · β)(R − β)] · [β(1 − R · β)] ˆ ˙ ˆ ˙ ˆ ˙ ˆ ˙ ˆ = (R · β)R · β − (R · β)(R · β)(R · β) ˙ ˙ ˆ ˆ ˙ ˆ ˙ −(R · β)β · β + (R · β)(β · β)(R · β) Now.
The expression we derived was: ¯ dP// q2 β˙2 sin2 θ = dΩ 16π 2 ε0 c (1 − β cos θ)5 (3. for both colinear and synchrotron radiation. denote it dP⊥ .3. we call this type of radiation synchrotron radiation. we call the radiation synchrotron radiation. for the cases where the velocity and acceleration are parallel and perpendicular. and we hence have the ﬁnal expression: ¯ dP dΩ = ˙ sin2 θ cos2 φ q2 β2 1− 2 16π 2 ε0 c (1 − β cos θ)3 γ (1 − β cos θ)2 (3.3.4
Radiation: Summary
We have derived the radiated power. and rearrange: ¯ dP dΩ Where we have noted that: = ˙ sin2 θ cos2 φ q2 β2 1− 2 16π 2 ε0 c (1 − β cos θ)3 γ (1 − β cos θ)2 γ2 =
1 1 − β2 Now. For the case of parallel. to expand out the top. and the expression we derived was: ¯ dP⊥ dΩ 3.40)
For the case of perpendicular. No synchrotron radiation Let us set φ = 0 in the expression for
¯ dP⊥ dΩ . Let us also show where maximum radiation occurs. this expression must be set to zero. we call the motion colinear acceleration. • Let us ﬁnd the total radiated power.
This gives:
˙ ¯ dP⊥ q2 β2 sin2 θ = 1− 2 dΩ 16π 2 ε0 c (1 − β cos θ)3 γ (1 − β cos θ)2 If there is to be no radiation. and usually denote the radiation ¯ dP// .39)
3. That is: ˙ q2 β2 sin2 θ 1− 2 =0 16π 2 ε0 c (1 − β cos θ)3 γ (1 − β cos θ)2
.3
Radiation
71
Now.41)
Example: Minimum & Maximum Radiation
Let us consider two exercises: • Let us show where no synchrotron radiation occurs.3.5 = ˙ q2 β2 sin2 θ cos2 φ 1− 2 16π 2 ε0 c (1 − β cos θ)3 γ (1 − β cos θ)2 (3. per unit solid angle. for φ = 0.
and solve for θ. along φ = 0. for the synchrotron radiation expression. we have derived that there is no synchrotron radiation. the Larmor total power radiated. Thus: ¯ dP// q2 dΩ = dΩ 16π 2 ε0 c Remebering that dΩ = sin θdθdφ. we must integrate the expression over solid angles. with respect to θ. So. Total power radiated Let us ﬁnd the total power radiated. then set to zero. We shall not do this here. the integral bit is:
¯ P// =
1 −1
1 − x2 4 1 dx = 5 (1 − βx) 3 (1 − β 2 )3
Hence:
˙ 1 q2β 2 4 ¯ P// = 8πε0 c 3 (1 − β 2 )3 q2 ˙ 2 6 β γ 6πε0 c q2 ˙ 2 β 6πε0 c
And. ﬁrst for the colinear case. this becomes: ¯ P// = That is: ˙ q2β 2 16π 2 ε0 c sin2 θ sin θdθdφ (1 − β cos θ)5 β˙2 sin2 θ dΩ (1 − β cos θ)5
π ˙ sin3 θ q2β 2 2π dθ 5 16π 2 ε0 c 0 (1 − β cos θ) Making the substitution of x = cos θ. and cleaning up. this is: ¯ P// =
And. which is more complicated this time. and the deﬁnition of γ. if θmin = cos−1 β. using sin2 θ = 1 − cos2 θ.11). with reference to (3. ¯ To show where power is a maximum. we ﬁnd: ¯ ¯ P⊥ = PL γ 4
. So. we must diﬀerentiate the dP⊥ . expanding out the (top) brackets. That is. this becomes: 1 (1 − β cos θ)2 − (1 − cos2 θ)(1 − β 2 ) = 0 (1 − β cos θ)5 Thus. gives: (β − cos θ)2 =0 (1 − β cos θ)5 Hence. noting that 1/γ 2 = (1 − β 2 ). we see that we must have cos θ = β. expanding out the top bracket again.72
3
RETARDED POTENTIALS & RADIATION
We must then solve this for θ. this is: ¯ ¯ P// = PL γ 6 ¯ PL ≡
Now. if we do a similar integration.
and is direction at a tangent to the circle. and specialised it to the cases of linear and perpendicular motion. at speed cβ. Hence. This concludes our discussion on retarded potentials. we have derived a formula giving the total power radiated by a charged particle. using this in P⊥ : γmcβ qB
q4β 2B 2 2 ¯ P⊥ = γ 6πε0 m2 c Hence. note that v = cβ. and that we can always put a charge moving with constant velocity into an inertial frame. we derived a general theorem of radiation. Infact. Then.
. we have that the acceleration is perpendicular to the velocity. what we have also seen.3. we can use the following expression for the total power radiated: ¯ P⊥ = q2 ˙ 2 4 β γ 6πε0 c
˙ The velocity is v = cβ. in a circular orbit (radius R). and we proceed onto relativistic electrodynamics. We saw that this was because stationary charges do not radiate.6
Suppose we have a charged particle. The acceleration is a = cβ. in which it is at rest. is that only accelerated charges radiate. we have considered the relativistic eﬀect of the ﬁnite speed of light. Thus. We have seen that a consequence of light having a ﬁnite speed. And
q 2 cβ 4 4 ¯ P⊥ = γ 6πε0 R2 Now.4
Discussion
So.
from standard Newtonian mechanics for motion in a circle. Hence:
v2 R. and is directed from the particle to the centre of the orbit.3. This easily gives: γma = qvB ⇒ γm R= ¯ Hence. is that moving charges radiate. Finally.
3. using this expression: q 2 a2 4 ¯ γ P⊥ = 6πε0 c3 Now. whilst ﬁnding positions of maximum and minimum radiation. mass m. in a circular orbit (governed by the externally applied magnetic ﬁeld B). We have an applied magnetic ﬁeld B which keeps the particles in such an orbit.4
Discussion Example: Charged Particle in Circular Orbit
73
3. from Newtons force equalling the Lorentz force: β 2 c2 = qβcB R Noting that everything is at rightangles. in this section. and stated the associated ﬁelds. We then derived the LienardWiechert potentials for point charges in motion. we have the relation a = also.
74
3
RETARDED POTENTIALS & RADIATION
.
we mean that the equations have the same form in both the primed & unprimed frame.3)
These two sets of transformations are for ‘boosts along the xaxis’.1)
+ βx ) c + βct ) y z β≡ v c
(4.
Let us state the Galilean transformations: t t= x = x + ut y y= z= z And also the Lorentz transformations: t = γ(t x = γ(x y= z= 1 γ≡ 1 − β2
(4.75
4
Relativistic Electrodynamics
The aim of this section is to show that the theory of electrodynamics is consistent with the special theory of relativity
Figure 8: The standard setup for the Lorentz boost along the xdirection. It is standard to state that the origins of the two systems coincide at t = 0. but are invariant under Lorentz transformations.2)
(4. and that the motion of the second frame Σ is inertial relative to the ﬁrst Σ. We could also show that Newton’s equation F = m¨ is invariant under Galilean. a The postulates of special relativity:
. where the primed frame is at a speed u = cβ. but not under Lorentz transformation. We can show that Maxwell’s equations are not invariant under Galilean transformations. and are the set we call the “inverse transformation”. By ‘invariance’.
4
. x2 . −) used here. +).µ=0
xµ gµν y ν
(4. which does use a diﬀerent metric signature of (−. We shall denote a (contravariant. the above modulus is: x · x = x · x = d2 Minkowski space ‘mixes’ the spatial components with a temporal component. as well as preliminary tensor calculus notes.76
4
RELATIVISTIC ELECTRODYNAMICS
• All laws of nature are independent of the translational motion of the system as a whole. the modulus of a 4vector remains unchanged under Lorentz transformation. by way of notation. we can refer to the vector by x. Again. That is. having their modulus invariant under Lorentz transformation.1
Notation
We may be able to see.5)
It is important to note that the expression gµν refers to the element. So. Let us write the following4 (what we will call contravariant) 4vectors: x = (x0 .uk/jonathanp. There are many physical consequences of these postulates. such as Cartesian or spherical polar.co.4)
Where we have introduced the metric tensor of Minkowski space: 1 0 0 0 0 −1 0 0 [gµν ] = 0 0 −1 0 0 0 0 −1
(4. Minkowski space is a space in which 4vectors reside. By writing xµ we are not appealing to any basis system. However. that from the Lorentz transformation equations. They may be found on http://myweb. and time dilation. • Information cannot travel faster than the speed of light. y 1 . We shall demonstrate this below. which is why we used [gµν ] above to refer to the whole tensor. not the tensor itself. +. +. that the following is invariant: c2 t2 − x2 − y 2 − z 2 = c2 t 2 − x 2 − y 2 − z 2 ≡ d2 That is. we will usually use xµ to refer to that single component. Also. −. in calculations. so that writing xµ implies the whole vector. as opposed to the 3vectors by v. they are what we usually call ‘vectors’. We will usually be sloppy with using this notation. y 3 )
As we shall also see later.tiscali.
4. x1 . we shall come to this later) 4vector by x. or by the set of components {xµ }. including length contraction. y 2 . The scalar product between the two vectors is then written:
3
x∗y =
ν. rather than the (+. these contravariant vectors occupy the vector space. The most useful document is “Index Notation”. x3 ) y = (y 0 . we will be sloppy with notation. −.
I have made a fairly comprehensive treatise of index notation & the various types of vectors.
Now. we may write a Cartesian dotproduct. the dotproduct of two 4vectors is written in terms of just the components (i.e. in writing it like this.4. This convention will be used throughout. as it has removed the need of how the basis vectors ‘interact’ with each other. Under the Einstein summation convention of implied summation over the available coordinate space. −1). the dotproduct is written simply as: x · y = xi yj δij With no mention as to the nature of the basis vectors. we have the metric δij . x. let us deﬁne a contravariant position vector thus: x = xµ = (ct. z) (4. In this way. the scalar product may then be written: x ∗ y = xµ gµν y ν Or. or how the basis interact (information of which is actually ‘inside’ the metric). the dotproduct is: x · y = gij xi yj Then.6)
. The Cartesian basis vector have the following relation: ei · ej = δij Then. we could ignore the basis. is that instead of writing the vectors as a sum of components with basis (i. in terms of this rule: x · y = xi yj δij = xi yi Which is the familiar rule for ﬁnding the dotproduct of a vector. the modulus of a 4vector: d2 ≡ x ∗ x = xµ gµν xν This way of writing a modulus (or inner product). Now. Generally. is very useful. and is usually written gij . We shall write the dot product between two vectors in Cartesian space: x · y = (xi ei ) · (yj ej ) = xi yj ei · ej Now. Let us now consider what these correspond to. in terms of a metric. the rule is called a ‘metric’. we cant go any further without stating how the basis vectors ‘interact’. we have seemingly generalised to any space. Infact.e. not their basis. this is technically incorrect: the fact that we have used the expression δij means that we have used a preconceived rule for the basis vectors. So. without need for a basis. So. which is the identity matrix. Second to note. Consider the following: In Cartesian space. x = xi ei ). we have not considered once what the basis is) and the metric: x · y = gµν xµ yν Where we have gone back to worrying about upper and lower indices. in Minkowski space (as above). y. −1. the ﬁrst thing to note is that we have used lower indices everywhere. and that they are all latin letter (as opposed to greek).1
Notation
77
This may also be written as gµν = diag(1. Then. the metric for Cartesian space is gij = δij . −1.
we have an expression for ‘what happens’ if we multiply a contravariant vector by the metric. z) xµ = (x0 . Note that the operation of the metric lowers the index.11)
That is. x2 . −y. −x. notice that the expression on the far right is actually how we deﬁned a covariant vector in (4. is actually a contravariant vector. x1 . x. then relabels it.10)
Being careful with the order of the indices5 .9). x3 ) = (ct.7)
These two objects may appear to be the same. So.78 And let us deﬁne a covariant position vector:
4
RELATIVISTIC ELECTRODYNAMICS
xd = xµ = (ct. x2 . I refer you to the afore mentioned “Index Notation”. x1 . Let us multiply the metric by a contravariant vector: 0 0 1 0 0 0 x x 0 −1 0 0 x1 −x1 gµν xν ⇒ 0 0 −1 0 x2 = −x2 0 0 0 −1 x3 −x3 Now.10). for explanations of these indices. Let us make the following identiﬁcation: xµ = (x0 . y. in terms of the actual vectors (as we used the ‘∗’ sign for the scalar product of two contravariant vectors) as: x · x d = xµ xµ And this is the equivalent of the ‘dotproduct we knew before’. for inﬁnitesimals: ds2 ≡ gµν dxµ dxν
Again. Now. x3 ) = (ct. Hence. −z) = (x .and contravariant vectors. These types of vectors transform in a very diﬀerent way. let us use that: xµ gµν xν = xµ xµ We denote this. but they are not! What we used to call just a ‘vector’. −z)
(4. and ways of manipulating them. for later reference. we have: xµ = g µν xν (4. This is what we will call “the interval”. Similarly. −y. −x . let us refer back to the expression we had for the modulus of a 4vector: x ∗ x = xµ gµν xν Now.9)
Now. its worth noting how the metric acts upon co.8) (4. by acting upon a covariant vector with the inverse metric (which we will come to shortly). raising the index. in (4. Hence: xµ = gµν xν (4.
5
. −x )
0 1 2 3
(4. −x. −x .
We can arrive at an interesting relation by multiplying (4. To do so. By way of getting ‘used’ to playing around with these objects. and the tensor itself: we represent a tensor. but we do not write the tensor itself. we can evaluate the LHS. This highlights the diﬀerence between representing a tensor by a matrix. but it is not the same tensor. which we then evaluated. not that the tensors are identical. which is indeed the statement of (4. So. resulting in the Kroneckerdelta. Also notice that gµν = gνµ : the tensors are symmetric. which leads us to the relation:
ρ g ρµ gµν = δν
(4. we raised/lowered the indices.10) by something that will raise the index on the LHS: g ρµ xµ = g ρµ gµν xν Now. Then.11). notice the inversenon multiplication. via (4. results in:
ρ xρ = δν xν = xρ
Which is indeed true! Notice. to give: xρ = g ρµ gµν xν Now.12)
Which. We shall come back to this later.1
Notation
79
Consider a “normal” 3vector x. but more ‘handwavey’ way: a matrix multiplied by its inverse gives the identity matrix. we can think about this in a slightly diﬀerent. giving:
ρ gµρ g ρλ gλν = gµρ δν
. Then. if we plug it back in. upon discussion of proper time.4.12). we must make this consistently true. given a vector of inﬁnitesimals in each direction: dx = (dx. depending on their initial location. we may write that gµν = g µν . consider the following expression: gµρ g ρλ gλν We notice that the middle and last term are inverse times noninverse. The inverse of the metric gµν is written g µν . but that merely means that its elements are the same. The inverse metric tensor has the same elements as the noninverse metric. dy. We say that multiplication always occurs between a covariant and contravariant vector (or viceversa). We form the following scalar product between two 4vectors:
ρ x · y d = xµ yµ = g µρ xρ gµν y ν = g µρ gµν xρ y ν = δν xρ y ν = xν y ν
Where we have just showed that xµ yµ = xµ y µ . Its length is given by the dotproduct with itself x2 = x · x. dz) Its length is obviously: ds2 ≡ dx · dx = dx2 + dy 2 + dz 2 Hence we see the reasoning behind calling the quantity ds2 = gµν dxµ dxν the ‘interval’.
So here. each occupying their own vector space. via taking the Hermitian conjugate. So. in a slightly rejigged way: ct x y z = γ(ct − βx) = γ(x − βct) = y = z
Now.
. The act of going from the ketspace to the braspace.80 And we can then use the Kroneckerdelta:
4
RELATIVISTIC ELECTRODYNAMICS
ρ gµρ δν = gµν
Hence. consider relabelling the coordinates thus6 : x0 ≡ ct
6
x1 ≡ x
x2 ≡ y
x3 ≡ z
It should be obvious that these are not exponents. we have contravariant vectors occupying the vector space. What always used to call vectors. and are thus contravariant vectors.
4. to get to the covariant space. by way of (brief) introduction to the next section. one can always deﬁne a dual vector space. These are infact two diﬀerent types of vectors. the vector space is the ketspace. but will give the components of the vector with respect to a diﬀerent frame of reference within the dual space of covariant vectors. is akin to multiplying a contravariant vector by the metric. and covariant vectors occupying the dual vector space. As we have seen. a Lorentz transformation keeps within one vector space. so that k represents such a vector. When one has a standard vector space. we have shown: gµρ g ρλ gλν = gµν We have introduced the concept of two diﬀerent types of vectors: covariant and contravariant vectors. given a covariant vector. And.2
Lorentz Transformation
Let us restate the Lorentz transformations. and the dual the braspace. We also know that taking the Hermitian conjugate of such a vector gives the ‘brastate’ vector k. and gives the vector referred to a diﬀerent coordinate system. So. That is. the operation of a metric upon a contravariant vector gives covariant vector. An example we may be familiar with. we can think of the metric as being a way of transferring between the vector space & the dual space (and obviously the other way round via the inverse). and are just indices. are occupying the vector space. a Lorentz transformation of the covariant vector retains the status of ‘covariant’. is in quantum mechanics: if we have a state in ‘ketspace’.
14)
Notice that the transformation is symmetric (that is.3 ∂x2 ∂x3
We see that these are the elements of the ﬁrst row of the transformation matrix. with a little thought. of ﬁrst rank. x1 . the transpose of the transformation matrix is the same as the untransposed version): Λµ ν = Λν µ Now.4.17)
. we see that the following is true: Λµ ν = Putting this into (4. x3 separately: ∂x 0 = γ = Λ00 ∂x0 ∂x 0 = −γβ = Λ01 ∂x1 ∂x 0 ∂x 0 = = 0 = Λ02.16) ∂x µ ∂xν (4.13) 1 0 x2 0 1 x3
Multiplication of these matrices reveals that we may indeed represent our system in this way. Hence. any quantity Aµ that transforms via: Aµ = Is a contravariant tensor. we can see that we can group these equations 0 x γ −γβ x 1 −γβ γ x2 = 0 0 0 0 x3 together into a single set of matrix equations: 0 x 0 0 x1 0 0 (4. Consider the x 0 equation: x 0 = γ(x0 − βx1 ) Let us diﬀerentiate this with respect to x0 . ∂x µ ν A ∂xν (4.14): xµ = ∂x µ ν x ∂xν (4. We write this as: x µ = Λµ ν xν (4. the next thing we do is a little more subtle.2
Lorentz Transformation
81
Then. of ﬁrst rank.15)
This is actually the deﬁnition of a contravariant tensor. That is. for a boost along the xdirection. as: x 0 = γ(x0 − βx1 ) x 1 = γ(−βx0 + x1 ) x 2 = x2 x 3 = x3 Then. x2 . we can see that we may write the Lorentz transformations.
giving: xµ = gµν Λν κ xκ We may then express the far right ‘superscripted’ component in terms of a ‘subscripted’ expression. 0 ∂x1 ∂x2 ∂x3 ∂x 1∂ ∂ ∂ ∂ . using this: xµ = (Λ−1 )µρ xρ Where we have that: (Λ−1 )µν = ∂xν ∂x µ (4.− . c ∂t ∂x ∂y ∂z
.11): xµ = gµν Λν κ xκ = gµν Λν κ g κρ xρ Now.2. as we would with matrix multiplication (note: we expect the identity matrix out): Λµ ν (Λ−1 )ρν = Which indeed conforms to expectation! In analogue with the contravariant tensor.19) (4. . by the metric. in ‘transformed’ space: xµ = gµν x ν Then.14).− c ∂t ∂x ∂y ∂z ∂ ∂ ∂ ∂ . the lowering of indices. we have that any covariant tensor of ﬁrst rank transforms like: Bµ = 4.18)
Let us write the transformation and the inverse. from (4. via (4.20) ∂x µ ∂xν ∂x µ µ = = δρ ∂xν ∂x ρ ∂x ρ
Let us write the 4vector diﬀerential covariant diﬀerential operator: ∂µ ≡ ∂ ∂xµ = = Also.82
4
RELATIVISTIC ELECTRODYNAMICS
Let us ﬁnd the inverse transformation.− . the contravariant diﬀerential vector: ∂µ ≡ ∂ = ∂xµ 1∂ ∂ ∂ ∂ . .1 Diﬀerentiation ∂xν Bν ∂x µ (4. a property of transformation matrices is that (we shall justify this in a later section): gµν Λν κ g κρ = (Λ−1 )µρ Hence. the far right term we can express in terms of ‘unprimed’ coordinates. . next to each other. . We shall do this by considering the following expression.
Show that xµ xµ = x µ xµ To show this.4. we will show that the RHS is the same as the LHS.21)
4.2.2
Lorentz Transformation
83
That is (confusingly).− c ∂t (4. but diﬀerentiates in a contravariant way. We have: ∂µ = 1∂ . 2. 3. we can write the D’Alembertian:
2
= ∂µ ∂ µ =
1 ∂2 − c2 ∂t2 ∂vi ∂xi
2
We shall frequently use the notation: ·v = Where it is understood that i = 1. c ∂t ∂µ = 1∂ . the operator ∂µ is covariant. and use it to show some useful things.2
Examples
Let us bring together some notation. forming the inner product: ∂µ ∂ µ = 1 ∂2 ∂2 ∂2 ∂2 1 ∂2 − 2− 2− 2 = 2 2− c2 ∂t2 ∂x ∂y ∂z c ∂t
2
That is. The confusion arises when we consider that the diﬀerentials themselves have the opposite sign. It may help to think of the following: consider the gradoperator in Euclidean space: = ∂ ei ∂xi
It is the equivalent of saying that is a covariant vector. and ∂ µ is contravariant. It transforms as a covariant vector. Notice then. To do this.
. let us write down the transformations: x µ = Λµ ν xν Therefore: x µ xµ = Λµ ν xν (Λ−1 )µρ xρ = (Λ−1 )µρ Λµ ν xν xρ
ρ = δν xν xρ
xµ = (Λ−1 )µρ xρ
= xν xν Hence proven.
22)
. We then multiply the whole g ρµ xµ = g ρµ gµν xν
Then.3 Tensors
Let us brieﬂy extend the deﬁnition of transformation of tensors. then the above easily becomes the deﬁnition of how rank2 contravariant tensor transforms: A µν = ∂x µ ∂x ν ρλ A ∂xρ ∂xλ (4. We can motivate a second rank tensor by considering two ﬁrst rank contravariant tensors: Cµ= ∂x µ ρ C ∂xρ Dν = ∂x ν λ D ∂xλ
Then. let Bµ → ∂µ : ∂ ∂xµ Bν = ∂xµ Bµ ∂x ν
∂xµ ∂xµ ∂ ∂ ∂µ = = = ∂ν ∂x ν ∂x ν ∂xµ ∂x ν Where we have noticed that the same factors cancel oﬀ. We ﬁrst must write the deﬁnitions of ∂µ and of a covariant vector: ∂µ ≡ Now. Consider a contravariant vector. Show that ∂µ is a covariant vector Now. It transforms thus: Aµ = ∂x µ ν A ∂xν
This is a rank1 contravariant tensor. we start with xµ = gµν xν . 4. we have shown that: ∂ν = ∂xµ ∂µ ∂x ν
Hence proven. let us write them next to each other: C µD ν = ∂x µ ∂x ν ρ λ C D ∂xρ ∂xλ
Then. raise the indices on the LHS: xρ = g ρµ gµν xν Therefore:
ρ g ρµ gµν = δν
Hence proven. Thus.2. this is a little more tricky. let us deﬁne Aµν ≡ C µ Dν .84
ρ Show that gµν g νρ = δµ thing by g ρµ :
4
RELATIVISTIC ELECTRODYNAMICS
To show this.
let us show what happens if the metric acts upon a contravariant tensor.µn = ∂xν1 ∂xν2 ∂ν ν3 ∂xνn . with a small amount of thought: gµν Aρνλκ = Aρµλκ ∂x µ µ = δν ∂x ν
∂x µ ∂xλ λ = δρ ∂xρ ∂x µ Although not a precise reason.. a covariant tensor of rankn transforms: Bµ1 µ2 µ3 .ν ∂x µ1 ∂x µ2 ∂x µ3 ∂x µn 1 2 3 n
We can also talk of ‘mixed rank’ tensors..4. A A µ1 µ2 µ3 . it is necessary to show that the above holds. a rank2 contravariant rank1 covariant tensor would transform like: ∂x µ ∂x ν ∂xγ αβ T µνρ = T γ ∂xα ∂xβ ∂x ρ Just as we had the metric acting upon a vector. if we know the transformation to some other frame of reference (such as our Lorentz transformation.... and relabels it.2. So. Notice that we have retained the order of the indices. For a more complicated mixedrank tensor: g µν Aργ νλ = Aργµλ It is useful to notice: ∂xµ µ = δν ∂xν And also. So. a tensor is a set of quantities which transform under a given rule..4 Inverse Lorentz Transformation
Let us consider an alternative way to think about & derive the inverse Lorentz transformation. 4. the above is easily realised if we notice we can cancel out the ∂ µ factors from the top left & bottom right. which is only one example of many).2
Lorentz Transformation
85
To show that a set of quantities T µν form a contravariant tensor. We saw that the metric acting on a covariant vector gave a contravariant vector: g µν xν = xµ We may also operate the metric upon a rank2 covariant tensor: g µν Aνλ = Aµ λ
. Bν ν ν . of second rank: gµρ Aνρ = Aνµ It drops the index.µn = ∂xν1 ∂xν2 ∂ν ν3 ∂xνn Similarly..νn ∂x µ1 ∂x µ2 ∂x µ3 . for example. relative to that new frame of reference... That is. leaving something we just saw was a Kroneckerdelta. we know what the set of quantities look like.... given one set of quantities. It is fairly easy to see the generalisation of this transformation rule to a contravariant tensor of rankn: ∂x µn ν1 ν2 ν3 .
as we see that Λµ ν is the Lorentz transform. which are completely equivalent. relabelling it µ (notice that the relative order & column is preserved). We do it for completeness). upon comparison of this ﬁnal matrix (that which we called Λ ). we ﬁrst acted the inner metric (this order is arbitrary). we may denote the inverse Lorentz transformation matrix as Λν µ .18). let us consider that the tensor in question is the Lorentz transformation: Λν µ = gνλ g µρ Λλ ρ Now. We must be very careful with the positions of the indices. we see that it is exactly the matrix containing the elements of the inverse Lorentz tranformation. replacing the middle metric above with its transpose: Λν µ = gνλ (g ρµ )T Λλ ρ Then. relabelling to ν. not that it is in the primed frame! Let us carry out this matrix multiplication. as one operation: Aν µ = gνλ g µρ Aλ ρ Now. we may act the metric upon a mixedrank tensor: g µν Aλ ν = Aλµ Let us then create a composition: gνλ g µρ Aλ ρ = gνλ Aλµ = Aν µ So. Then.86
4
RELATIVISTIC ELECTRODYNAMICS
Equivalently. this. which raised the ρ. as promised! So. after noting that the transpose of the metric is the same the untransposed metric: 1 0 0 0 γ −γβ 0 0 1 0 0 0 0 −1 0 0 −γβ γ 0 0 0 −1 0 0 gΛg T = 0 0 −1 0 0 0 0 −1 0 0 1 0 0 0 0 −1 0 0 0 1 0 0 0 −1 γ γβ 0 0 γβ γ 0 0 =Λ = 0 0 1 0 0 0 0 1 So. we may reorder the expression (as we are at liberty to: these are just numbers!): Λν µ = gνλ Λλ ρ (g ρµ )T This is now in the form of matrix multiplication: Λ = gΛg T The prime is to denote that it is in someway diﬀerent. Consider g µρ = (g ρµ )T (which is infact a fairly pointless operation. as it stands. since the metric is symmetric. Thus. The outer metric then dropped the λ. So. for the inverse Lorentz transformation: Λµ ν = (Λ−1 )µ ν
. we see the justiﬁcation of (4. is not the form in which we have matrix multiplication. common notation. Hence.
dx) = (cdt.3. in the contravariant vector xµ = (ct.3
Lorentz 4Vectors
87
4. dy. dx. due to the form of the metric). y. the standard 3vector. where this notation implies that x = (x. any general 4vector. Let us write it in terms of the components of the corresponding inﬁnitesimal co. That is. dz) And the corresponding covariant inﬁnitesimal: dxµ = (cdt. in order to be called 4vectors. z).3
Lorentz 4Vectors
Now.and contravariant vectors: ds2 = dxµ dxµ = c2 dt2 + (dx)(−dx) + (dy)(−dy) + (dz)(−dz) = c2 dt2 − dx2 And therefore: ds = c2 dt2 − dx2 = dxµ dxµ
4. We saw that it had the folowing transformation properties: xµ xµ = x µ xµ x µ = Λµ ν xν
Infact. with components Aµ has the same properties: Aµ Aµ = A µ Aµ A µ = Λµ ν Aν
So that any 4vector (or two 4vectors) are invariant under Lorentz transformation.1
Proper Time
Now. x). we have already seen a 4vector. let us deﬁne the contravariant inﬁnitesimal: dxµ = (cdt.23)
. Now. This is infact a condition that a set of quantities must fulﬁll.4. let us deﬁne the proper time τ as the line element divided by the speed of light. −dx) We then have the line element (which we have already discussed): ds2 = gµν dxµ dxν = dxµ dxµ And which we have shown to be invariant (and is only valid in Minkowski space. in inﬁnitesimal form: dτ = ds c (4.
the proper time is: dτ = dt γ (4. The particle is now moving through space as well as time. but not in space. Consider (a) ﬁrst: a particle in its rest frame carves out a vertical line in a spacetime diagram: it has motion through time. Now consider the second case (b). A moving particle experiences less time than a stationary particle!
That is: dτ = = = = dt = dt Where we have noticed that: dx dt Also. in a spacetime diagram.88
4
RELATIVISTIC ELECTRODYNAMICS
Figure 9: Consider two diﬀerent situations. Notice that it has a diﬀerent projection onto the t axis.24) 1 1 − β2
2
1 c 1 c
dxµ dxµ c2 dt2 − dx2 dt2 − 1 2 dx c2 dx dt
2
1 1− 2 c 1 − β2
= (v)2 = v
β=
v c
. Notice that we have marked on the interval length ds1 as the length of the line in total. recall that: γ≡ Hence. also notice that dt1 is the amount of time experienced by such a motion.
3
Lorentz 4Vectors
89
Let us regeneralise this. consider a ball being thrown. gives the proper time. So. The line element ds is the interval of spacetime traversed in proper time dτ . and a ball being rolled along the ﬂoor. Perhaps one can think of it in diﬀerent terms. If you think about sitting stationary. That the lines are the same length is the statement that the proper time for the two observers. The two observers (stationary & moving in the rocket) both agree that the rocket leaves in the year 2065. the stationary obsever on the earth thinks that the year is 2080. This is the “normal” timedilation eﬀect. Then. the one on the earth). The stationary observer has a longer t than the moving observer. Consider that both observers age by 5 years. even though you are not ‘moving’. We can also think about this in terms of possesing a restmass energy. after the observer in the rocket has aged by (say) 10years. except the ‘extra distance’ travelled is motion through (improper) time.4. noting the expression γdτ = dt. The stationary line has a longer projection than the moving line. you are carving a line through the ‘time axis’. To think about it another way. their projections onto the taxis is diﬀerent. The observer (that was moving) then thinks that the year is 2075. and that they both have the same starting and ending positions. This is the oddity of time dilation!
4. The proper time experienced by the twin moving through space (i. ds1 and ds2 have the same length for both stationary and moving frames). With reference to Fig (9). their projections onto the taxis are diﬀerent. That is. we can get a handle on what proper time is. So they travel diﬀerent distances. by taking ourselves all the way back to the metricexpression: dτ = 1 c gµν dxν dxµ
Now. he arrives back to earth. The length of the line that joins two events.3. proper time not only takes into account movement in space.25)
. So. is the same. we see that if the two lines have the same length (i. but derived using a metric and spacetime intervals.e.2
4velocity & 4momentum
Let us consider the 4velocity: uµ . hence diﬀerent travel times. We shall deﬁne it thus: uµ ≡ dxµ dτ (4. but also in time. if we consider the usual scenario of a twin remaining on the earth while another travels on a rocket & back again (ignoring any possible acceleration eﬀects).e. as the thrown ball must travel up as well as down & across. However. the one on the rocket) is more than the twin that did not move through space (i. It is classically obvious that they take diﬀerent times to traverse the same distance. but not the spatial axes. It is this exact same thing.e. However. An observer which is at ‘rest’ experiences less proper time than an observer ‘moving’. Now consider that the rocket is able to travel at a very large portion of the speed of light. whether you have moved through space or not. That is. the proper time for both observers is 5 years. Consider that the moving observer starts his journey in the year 2065. consider the usual analogy of an observer on a rocket ship & another observer stationary on the earth. but the rolled ball only across.
the 4momentum is: pµ = muµ = (mcγ. it is clear that pµ = gµν pν . Now. γu) u= dx dt (4.90 Hence. That is. an invariant. 0) (after noting that γ = 1 if u = 0). the ﬁrst component of the 4vector is just u0 = γc. consider: dx dx dt dx = =γ dτ dt dτ dt Therefore: uµ = (cγ. if u = 0.26)
So. then it still posses an inherent energy: its rest energy. dτ dτ
Now. he will posess a component which is progressing along the taxis at the speed of light. we write the 4momentum as: pµ = (E/c. let us brieﬂy consider an implication of the 4velocity vector uµ = (cγ. and so we call it the invariant mass. this is invariant. γu). Hence: E2 1 − p2 = 2 (E 2 − c2 p2 ) = m2 c2 c2 c And therefore. p). The rest energy comes about because the particle still has motion through the time coordinate. as we said. Hence.27)
Therefore. For an observer who is at rest (spatially). −p) Therefore: pµ pµ = E2 − p2 c2 p2 ≡ p2 + p2 + p2 x y z
But. p) Now.
. as pµ is a 4vector. we see that from (4. then uµ = (c. And. even a “stationary” particle is moving through the time coordinate. using the fact that dxµ = (cdt. Also. γmu) Now. x): uµ = c
4
RELATIVISTIC ELECTRODYNAMICS
dt dx . similarly. we also have the relations: E = γmc2 p = γmu E 2 = p2 c2 + m2 c4 (4. If a particle is at rest (spatially). we recover: E 2 − c2 p2 = m2 c4 So.24). let us again consider the implication of pµ = (E/c. so that: pν = (E/c. we then have that pµ pµ = p µ pµ .
3. we may think that we can write the above equation in a more compact form. . if we deﬁne the 4vector current density thus: J µ = (cρ. z) as our Cartesian position 4vector. . .3
Lorentz 4Vectors Electrodynamic 4vectors
91
4. x.4. Recall that the the Lorentz gauge was: 1 ∂φ + c2 ∂t ·A=0
Then. we see that we can write the 4vector potential: Aµ = (φ/c. A) (4. Now. y. the continuity equation reads: ∂µ J µ = 0 Let us just check that we recover the continuity equation if we write things like this. the continuity equation becomes: ∂µ J µ = That is. So: ∂µ J µ = ∂J 0 ∂Ji + ∂x0 ∂xi 1 ∂cρ ∂Ji = + c ∂t ∂xi ∂ρ + ·J = ∂t = 0 (4. J ) Then. x3 ) = (ct.3
Consider the continuity equation: ∂ρ + ·J =0 ∂t It must be valid in all reference frames. if we consider that we had xµ = (x0 .29) 1 ∂cρ ∂Ji + =0 c ∂t ∂xi (4.30)
. 0 ∂x1 ∂x2 ∂x3 ∂x 1∂ ∂ ∂ ∂ . let us write the divergence part under the summation convention: ∂ρ ∂Ji =0 + ∂t ∂xi Now. by analogy with the continuity equation. . x2 . Recall the diﬀerential operator: ∂µ = = ∂ ∂ ∂ ∂ . c ∂t ∂x ∂y ∂z
Then. x1 .28)
Which is indeed true.
and in the process bringing the current density components into play. Aj = Aj and J j = Jj (this seems to be odd notation.92 And that the Lorentz gauge can be written: ∂µ Aµ = = =
4
RELATIVISTIC ELECTRODYNAMICS
∂A0 ∂Ai + ∂x0 ∂xi 1 ∂φ/c ∂Ai + c ∂t ∂xi 1 ∂φ ∂Ai =0 + c2 ∂t ∂xi
Therefore. it then carves out a line in both space and time. Let us pick the component µ = j (just to be clear. Then: 1 ∂2 ∂2 φ − 2 = µ0 cρ c2 ∂t2 ∂xi c 1 ∂2 ∂2 ⇒ − 2 φ = µ0 c2 ρ c2 ∂t2 ∂xi Which is indeed the equation corresponding to a wave driven by ρ. That is. Jy . by writing J µ = (cρ.32)
Let us pick the component µ = 0: A0 = φ/c and J 0 = cρ. we were able to derive the following wave equations: 1 ∂2φ − c2 ∂t2 1 ∂2A − c2 ∂t2
2
φ = µ0 c2 ρ
2
A = µ0 J
Hence. having components J = (Jx . notice that J 0 = cρ. There is obviously the exact same situation with A0 . we have implicitly used the notation that J is the standard 3vector current density. the j th component of the vector potential A). Then. that is: ∂2 1 ∂2 − 2 c2 ∂t2 ∂xi Aµ = µ 0 J µ (4. we can write the 4vector potential in a manifestly Lorentez invariant form: ∂ µ Aµ = 0 (4. Jz ). it is still carving out a line through the ‘time dimension’. J ) and Aµ = (φ/c. we see that we can write these as (which we then verify): ∂ ν ∂ν Aµ = µ0 J µ So. Then: 1 ∂2 ∂2 − 2 c2 ∂t2 ∂xi Aj = µ0 cJ j
. Also. so that even for a charge ‘at rest’. there exists a charge density. a charge at rest still generates a potential ﬁeld (as we have seen in electrostatics). A). When it is moving relative to an observer. which we now see as being due to it carving out a line in time. Under the Lorentz gauge. but it is ok. if you look at an electron that appears to be at rest. But when it moves. if one follows the meaning).31)
Notice. which corresponds to the standard charge density ρ. and similiarly for the 3vector vectorpotential A. it also produces a vector potential ﬁeld.
as we must compute all the components: 0 0 J γ γβ 0 0 J γβ γ 0 0 J 1 J1 J2 = 0 0 1 0 J 2 J3 0 0 0 1 J3 cρ0 γ γβ 0 0 γβ γ 0 0 0 = 0 0 1 0 0 0 0 0 0 1 γcρ0 γβcρ0 = 0 0 That is: J 0 = γcρ0 Recall that: uµ = (cγ. consider standing on the earth. under summation convention: 1 ∂2A 1 ∂ 2 Aj ∂2 − 2 A = µ0 J − 2 Aj = µ0 Jj c2 ∂t2 c2 ∂t2 ∂xi Let us consider an example: If ρ0 is the rest charge density in its rest frame. γu) Then: J µ = (γcρ0 . γcβρ0 . We know the transformation:
ν
J µ = (Λ−1 )ν µ J
Now. we have the ‘primed’ components. with respect you to as a stationary observer. 0) = ρ0 (γc. let us say that in the primed frame we have J
µ
= (cρ0 . Notice that the corresponding wave equation. taking u = (u. 0. γu. 0) So. 0. this will actually be easier to see in matrix notation. That is.4
Electromagnetic Field Tensor
Recall the following expressions of the electric & magnetic ﬁelds. we see that: J µ = ρ0 uµ J 1 = γcβρ0
4. That is.4
Electromagnetic Field Tensor
93
Which is immediately satisﬁed. with a chunk of charge moving. let us ﬁnd the 4vector current density in the stationary frame. So.4. 0). in terms of the vectors & scalar potential: ∂A E =− φ− B = ×A ∂t
. 0). and let us ﬁnd the ‘unprimed’ components of the 4vector current. 0. with J = 0.
∂t ∂t ∂t
Picking out the ‘1’ component. remembering that xi = −xi . −x3 ) ∂φ ∂Ai − ∂xi ∂t
Then. rather than ‘1’: Ei = − Now. E 2 . So that: B2 = F µν = ∂Aν ∂Aµ − = ∂ µ Aν − ∂ ν Aµ ∂xµ ∂xν (4. suppose we chose component ‘i’. A2 . and writing x = x1 . −x2 . And. y = x2 . let us write the full equation out for the electric ﬁeld: (E 1 .94
4
RELATIVISTIC ELECTRODYNAMICS
Now. x3 ) xµ = (ct. let us have cx0 . Also notice that it is all consistent. A3 ) xµ = (ct. Thus: E i = −c ∂Ai ∂A0 −c 0 ∂xi ∂x
And. −x1 . x2 . A1 . x1 . Let us continue with the ﬁelds. Let us write out the magnetic ﬁeld: − e1 B= So that. instead of writing φ. . we have deﬁned an element of the electromagnetic ﬁeld tensor.− ∂x ∂y ∂z − ∂A1 ∂A2 ∂A3 . say. E 3 ) = − ∂φ ∂φ ∂φ .33) ∂A3 ∂A2 ∂A2 ∂A3 − ≡ F 32 − = ∂x2 ∂x3 ∂x3 ∂x2
∂ ∂x A1
e2
∂ ∂x2 A2
e3
∂ ∂x3 A3
. let us write cA0 . z = x3 : E1 = − ∂φ ∂A1 − ∂x1 ∂t
Now. for generality. picking the component B 1 : B1 = Similarly: ∂A3 ∂A1 − ≡ F 13 ∂x1 ∂x3 Inspection of the order of the indices will show how we have arrived at the labelling we have.− . recall our previous deﬁnitions: Aµ = (φ/c. instead of having t. and also that x0 = x0 : ∂A0 ∂Ai Ei = − c ∂xi ∂x0 Multiplying by 1: ∂Ai ∂A0 Ei = − ≡ F 0i c ∂x0 ∂xi That is.
As a little aside on antisymmetry & tensors: if the antisymmetry of a tensor is deﬁned as Aµν = −Aνµ then the components must conform to this. suppose the tensor had elements along the diagonal. That is. they would not be aﬀected by the ﬂipping. using these: ∂x µ ∂x ν ρ λ ∂x µ ∂x ν ∂xρ κ ∂xλ σ ∂ A = ∂ A ∂xρ ∂xλ ∂xρ ∂xλ ∂x κ ∂x σ Tidying up a little: ∂x µ ∂x ν ∂xρ ∂xλ κ σ ∂ A ∂xρ ∂xλ ∂x κ ∂x σ ρ Now. reﬂecting the elements along the diagonal should ﬂip the sign of all components. We can use the rule that ∂xρ = 1. giving: ∂x µ ∂x ν κ σ ∂ A ∂x κ ∂x σ ∂xρ κ ∂ ∂x κ Aλ = ∂xλ σ A ∂x σ So. and then we have some components which do not conform to the deﬁnition. Now. we know that each of the following holds (using the inverse transformation): ∂ρ = Thus. ∂x which is pretty obvious.34)
Hence. Then. Which is something we see from the ﬁeld tensor. and those we antipicate: 0 −E1 /c −E2 /c −E3 /c E /c 0 −B3 B2 F µν = 1 E2 /c B3 0 −B1 E3 /c −B2 B1 0
(4.4. if we represent it as a matrix. of second rank. Therefore. and second and fourth term can be written like this. let us prove this by trying to show that the following is true: ∂x µ ∂x ν ∂ ρ Aλ − ∂ λ Aρ = ∂ µ A ν − ∂ ν A µ ∂xρ ∂xλ ∂x µ ∂x ν ρ λ ∂ A ∂xρ ∂xλ Now. Now. So. we see that an antisymmetric tensor must have only zero diagonal entries. we need to prove that the following holds: F
µν
=
∂x µ ∂x ν ρλ F ∂xρ ∂xλ
As it is supposed to be a contravariant tensor. Thus. let us consider:
. Let us do so. we have not yet proved that the ﬁeld tensor does indeed transforms as a tensor. we proceed by trying to make things into Kroneckerdeltas. nor the sign ﬂipping. Now. grouping together the terms we have computed. we have the electromagnetic ﬁeld tensor. That is. but we use it over other diﬀerentials. Aµµ = Aµµ . We then notice that the ﬁrst and third.4
Electromagnetic Field Tensor
95
Also notice that we can immediately see that it is antisymmetric: F µν = −F νµ So.
this is a semitrivial proof. And therefore.35)
This is. The subsequent expressions are then just even permutations of this combination. we know all of the elements of this: ∂0 = 1∂ c ∂t ∂1 = ∂ ∂x ∂2 = ∂ ∂y ∂3 = ∂ ∂z
. and the construction of (4. it is not a stretch of the mind to say that ∂ ν Aµ also transforms as a second rank contravariant tensor. the LHS expression unpacks slightly to:
3
(4. let us start by stating the following two equations. all indices appear in the same order: µ. notice that in the last equation.22). And although not a proof. λ. if we consider how F µν is constructed. each of the diﬀerentials left over are just Kroneckerdeltas:
µ ν δκ δσ ∂ κ A σ = ∂ µ A ν
And therefore.38)
∂µ F µ1 = ∂0 F 01 + ∂1 F 11 + ∂2 F 21 + ∂3 F 31
µ=0
Now.96
4
RELATIVISTIC ELECTRODYNAMICS
However. let us consider the ﬁrst equation: ∂µ F µν = µ0 J ν Now.1 Maxwell’s Equations from F µν
Now. F µν is a contravariant tensor. what we have shown is that: ∂x µ ∂x ν ρ λ ∂ A = ∂ µA ν ∂xρ ∂xλ Conﬁrming the status of ∂ µ Aν as a contravariant tensor of second rank.37)
+∂ F
ν
λµ
+∂ F
λ
µν
Firstly. 4.36) (4. by deﬁnition of our Lorentz transforms: F
µν
= Λµ ρ Λν λ F ρλ
Infact. and then we shall recover Maxwell’s equations from them: ∂µ F µν ∂ F
µ νλ
= µ0 J ν = 0
(4. ν. let us consider the case ν = 1: ∂µ F µ1 = µ0 J 1 Then.4. Now. of second rank: F
µν
=
∂x µ ∂x ν ρλ F ∂xρ ∂xλ
(4.
let us write Maxwell’s fourth equation (notice that it is a vector equation. and taking ν = 1. consider the ﬁrst component of the following crossproduct: e1 ( × B)x =
∂ ∂x
1 ∂E1 ∂B3 ∂B2 + − = µ0 J 1 c2 ∂t ∂y ∂z
e2
∂ ∂y
e3
∂ ∂z x
=
B1 B2 B3
∂B3 ∂B2 − ∂y ∂z
Also. we found the ﬁrst component of the Maxwell equation. we recover Amperes law. 3. what about ν = 0:
3
∂µ F µ0 = µ0 J 0
µ=0
We see that we will need the elements of the ﬁeld tensor which are down the ﬁrst column: F 00 = 0 F 10 = E1 /c Thus (noting that J 0 = cρ):
3
F 20 = E2 /c
F 30 = E3 /c
∂µ F µ0 =
µ=0
1 c
∂E1 ∂E2 ∂E3 + + ∂x ∂y ∂z ρ ε0
= µ0 cρ
Which is completely equivalent to: ·E =
Therefore.
. 2.4. in total. 3 will produce the whole set of equations that gives Maxwells fourth equation. and is actually a set of 3 equations): 1 ∂E × B = µ0 J + 2 c ∂t µν = µ J ν . 2. We see that taking ν = 1 in ∂µ F 0 It is not too hard to see that ν = 1. looking at the ﬁeld tensor elements: F 01 = −E1 /c And therefore. Then. we have:
3
F 11 = 0 F 21 = B3
F 31 = −B2
∂µ F µ1 = −
µ=0
1 ∂E1 ∂B3 ∂B2 + − c2 ∂t ∂y ∂z
And considering the whole equation together: − Shuﬄing around a bit: ∂B3 ∂B2 1 ∂E1 − = µ0 J 1 + 2 ∂y ∂z c ∂t Now. we have seen that taking ν = 0 in ∂µ F µν = µ0 J ν we are able to recover Gauss’ law.4
Electromagnetic Field Tensor
97
And.
in the primed frame. It is possible to show that T µνλ transforms as a contravariant tensor of third rank. in terms of components in the unprimed frame. We can immediately start to do 0 −E1 /c −E2 /c −E3 /c 0 E1 /c 0 0 −B3 B2 0 E2 /c B3 0 −B1 1 E3 /c −B2 B1 0 γ −γβ 0 0 −γβ γ 0 0 0 0 1 0 0 0 0 1
Doing the matrix multiplication of the two far right matrices: γβE1 /c −γE1 /c −E2 /c −E3 /c γ −γβ 0 0 −γβ γ 0 0 γE1 /c −γβE1 /c −B3 B2 F = 0 γE2 /c − γβB3 −γβE2 /c + γB3 0 1 0 0 −B1 γE3 /c + γβB2 −γβE3 /c − γB2 B1 0 0 0 0 1 And the ﬁnal multiplications: 0 −E1 /c −γ(E2 /c − βB3 ) −γ(E3 /c + βB2 ) E1 /c 0 −γ(B3 − βE2 /c) γ(B2 + βE3 /c) F = γ(E2 /c − βB3 ) γ(B3 − βE2 /c) 0 −B1 γ(E3 /c + βB2 ) −γ(B2 + βE3 /c) B1 0
.5
Lorentz Transformations of the Fields
Now. so we leave it in the above form.98
4
RELATIVISTIC ELECTRODYNAMICS
Let us consider the second equation which we said represented Maxwells equations: ∂ µ F νλ + ∂ ν F λµ + ∂ λ F µν = 0 We are trying to extract the following set of four equations: ×E =− ∂B ∂t ·B =0
Now. evaluating the above explicitly. we would have to deﬁne it anyway. compute it via matrices. we have that the electromagnetic ﬁeld tensor transforms thus: F
µν
= Λµ ρ Λν λ F ρλ
If we actually wish to compute the elements of the ﬁeld tensor. we can either go through many many summations. it is slightly easier to see whats going on if we write the following: T µνλ ≡ ∂ µ F νλ + ∂ ν F λµ + ∂ λ F µν = 0 It is fairly easy to ﬁnd that T 023 . Notice that the above transformation may be written: F = ΛF ΛT We must write it like this this: γ −γβ 0 −γβ γ 0 F = 0 0 1 0 0 0 to allow us to use matrix multiplication. or.
4. but for it to make sense. we could have reduced this equation to T µνλ = 0. and is proven in the appendix. Notice. T 013 and T 012 give the three components to Faradays law. and that T 123 gives the lack of magnetic monopoles equation.
We use that E = (E1 . with the stationary charge inside the box. z ) 3 4πε0 r 4πε0 r 3 B =0
Now. the ﬁeld having components E1 . from the observation point of Σ.6
LienardWiechert Fields from Lorentz Transformation
Let us consider two frames: Σ & Σ . E3 ).e. The observer who is moving relative to the charge (i. next to him. with a charge at rest inside the box. due to that moving charge. E2 . the two origins coincide. That is. we consider a charge q at rest within Σ . in its frame. Then consider that another observer is standing on the surface of the earth. we have: 0 −E1 /c −E2 /c −E3 /c E /c 0 −B3 B2 F = 1 E2 /c B3 0 −B1 E3 /c −B2 B1 0 And therefore we can read oﬀ the new components. let us get this into vectorform. the ﬁnal thing to note is that the ﬁeld tensor in the primed frame has the same components as the unprimed frame. we wish to compute the ﬁelds. as we shall see.
. That is. These components are unchanged. E = (E1 . the one on the earth) “observes” a diﬀerent ﬁeld (infact. Now. It is only the components perpendicular to the boost direction which are changed. At t = t = 0. Then. watching the box moving.4. The problem at hand is to compute the electric ﬁeld that the stationary observer observes.6
LienardWiechert Fields from Lorentz Transformation
99
Now. By inspection: ˆ ˆ ˆ E = γE + (1 − γ)E1 x + γβc(B2 z − B3 y) Which is infact: 1−γ (v · E)v + γv × B v2 And. with no magnetic ﬁelds. Let us just formulate this in a slightly less mathematical manner: Consider that an observer is sat inside a box. y . an observer is in the rest frame of the charge. in terms of the old ones: E1 = E1 B1 = B1 E2 = γ(E2 − βcB3 ) B2 = γ(B2 + βE3 /c) E3 = γ(E3 + βcB2 ) B3 = γ(B3 − βE2 /c) Notice: the transformation was for a boost along the xdirection. E3 ). E2 . This is done mainly by inspection. B1 in that direction. the one inside the box). doing the same thing for the magnetic ﬁeld: E = γE + B = γB + γv × E 1−γ (v · B)v − 2 v c2
4. Then.e. both electric & magentic) to the observer who is stationary relative to the charge (i. Now. its ﬁelds in its rest frame is just the Coulomb ﬁeld: E = 1 q 1 q r = (x . the electric ﬁeld the observer “observes” is just the standard Coulomb ﬁeld.
except that we measure axial angles from the xdirection: x = r cos θ Hence: r 2 = γ 2 r2 cos2 θ + r2 sin2 θ Let us write this as: r 2 = γ 2 r2 cos2 θ + r2 sin2 θ 1 = γ 2 r2 cos2 θ + 2 sin2 θ γ = γ 2 r2 cos2 θ + (1 − β 2 ) sin2 θ = γ 2 r2 (1 − β 2 sin2 θ) y = r sin θ cos φ z = r sin θ sin φ
. in terms of coordinates in Σ.100
4
RELATIVISTIC ELECTRODYNAMICS
We have seen that we have the following transformations of the ﬁeld components: E1 = E1 B1 = B1 E2 = γ(E2 − βcB3 ) B2 = γ(B2 + βE3 /c) E3 = γ(E3 + βcB2 ) B3 = γ(B3 − βE2 /c) So. 0) = 0. using these: E1 = E1 And therefore: E = (E1 . The Lorentz transformations. 0. are just: x = γ(x − vt) = γx And therefore: E= y =y z =z
1 qγ (x. y. by our Lorentz transformations: r 2 = γ 2 x2 + y 2 + z 2 If we use the standard cartesianspherical polars coordinates transformation. noting that B = (B1 . E2 . for the coordinates. at t = 0. for the charge in its rest frame: E= 1 q (x . we are able to solve a little: B1 = 0 So. B3 ) = (0. γE3 ) And. E3 ) = (E1 . γy . reference to the above Coulomb ﬁeld. It is evident that: r2 = x2+y2+z2 And that is. z) 4πε0 r 3
Now. γE2 . let us ﬁnally express the r in terms of unprimed coordinates. B2 . γz ) 4πε0 r 3 E2 = γE2 E3 = γE3 β B2 = − E3 c B3 = β E2 c
Now we wish to express the components of E.
4. using this: E= Which is: E= qγ 1 r 3 r 3 (1 − β 2 sin2 θ)3/2 4πε0 γ q(1 − β 2 ) 1 r 4πε0 r3 (1 − β 2 sin2 θ)3/2
This is the same as the LienardWiechert formula we derived (with considerable vector algebra) for the electric ﬁeld of a moving point charge.
.6
LienardWiechert Fields from Lorentz Transformation
101
And therefore. with no acceleration.
−x. z) • Covariant vector . x.transfer between vector & dual space: xµ = g µν xν xµ = gµν xν
µ g µν gνρ = δρ
• Lorentz transformation & inverse: ∂xν ∂x µ (Λ−1 )µν = Λµ ν = ∂xν ∂x µ x µ = Λµ ν xν • Covariant Lorentzian boost: xµ = (Λ−1 )µν xν • Contravariant vector deﬁntion: Aµ = • Covariant vector deﬁnition: Bµ =
ρ Λρ µ (Λ−1 )νµ = δν
• Contravariant Lorentzian boost . with little explanation. • Contravariant vector .transform between inertial frames: xµ = (Λ−1 )νµ x ν
xµ = Λν µ xν ∂x µ ν A ∂xν ∂xν Bν ∂x µ 1∂ . most of the transformations & metrics discussed.− c ∂t ∂µ ≡
. y.occupies the dual vector space: xµ = (ct. A) • Lorentz gauge. −z) • Metric actions .7
Summary of 4Vectors & Transformations
So. continuity equation & wave equation: ∂µ Aµ = 0 • Electromagnetic ﬁeld tensor: F µν = ∂ µ Aν − ∂ ν Aµ • Maxwell’s equations: ∂µ F µν = µ0 J ν ∂ µ F νλ + ∂ ν F λµ + ∂ λ F µν = 0 ∂µ J µ = 0 ∂µ ∂ µ Aν = µ0 J ν 1∂ .occupies the vector space: xµ = (ct. J ) Aµ = (φ/c. let us bring together in one place. −y. c ∂t
• Contravariant & covariant diﬀerential operators: ∂µ ≡ • 4current & 4potential: J µ = (cρ.102
4
RELATIVISTIC ELECTRODYNAMICS
4.
and ﬁnd a way of expressing Maxwell’s four equations in a very elegant way.and covariant vectors. We then looked at various 4vectors. We were then able to deﬁne the electromagnetic ﬁeld tensor.8
Discussion
In this section. we found the transformations of the ﬁeld components. had the same form in diﬀerent inertial frames of reference). We have introduced index notation. Finally. We then formulated various previously known equations (such as the Lorentz gauge & the continuity equation) in a way which was Lorentz invariant (that is. and conﬁrmed that they did transform in an expected way. we have formulated electrodynamics in a way which is invariant under Lorentz transformation. but with considerably less algebra!
.4. and contra. which is in concordance with one of the postulates of special relativity. and how to perform a Lorentz boost. showing that it is indeed a tensor. ﬁnishing with an example which recovered the LienardWiechert ﬁelds that we derived in the previous section. within one space.8
Discussion
103
4. We have looked at how to transform between the two vector spaces via the Minkowski metric.
in cylindrical polars: er /r ×V =
∂ ∂r
rVθ
r sin θVφ
eφ
∂ ∂φ
ez
∂ ∂z
(A.1.104
A
VECTOR IDENTITIES & TRICKS
A
A. is given by: er /r2 sin θ eθ /r sin θ ×V =
∂ ∂r ∂ ∂θ
eφ /r
∂ ∂φ
(A.1
Vector Identities & Tricks
Vector Identities
Below are some commonly used vector identities: a · (b × c) = b · (c × a) a × b × c = (a · c)b − (a · b)c × ·( × ψ = 0 ×a = ( · a) − ψ+ψ · b) − b(
2
× a) = 0 a × b) )a − (a · )b · (ψa) = a · ·a · a) + (b ·
· (a × b) = b · ( × a × b = a(
× a) − a · (
All are easily proven. but some (if not all) are pretty tedious to actually show! We have also made use of: · ˆ r r2 21 r = 4πδ(r) = −4πδ(r)
A. θ. so that its components are Vr . Vφ . in spherical polars. Vθ .1
Curl in Spherical & Cylindrical Polars
Suppose we have V (r. Suppose we have the integral: I=
0 π
cos θ sin θ dθ
. its curl. then.2
Useful Tricks
Below are some useful ‘tricks’ to bear in mind when evaluating certain expressions.1)
Vr Similarly. φ).2)
Vr
rVφ Vz
A.
For a more general case. swopping the limits:
1
I=
−1
x dx
Which results in zero. We also trivially see that dx = − sin θdθ. the lower limit becomes x = cos 0 = 1. we can evaluate it via the subsitution: x ≡ cos θ Then. using the same substitution:
1
I=
−1
xn dx =
0
2 n+1
n odd n even n odd n even
So:
0
π
cosn θ sin θ dθ =
0
2 n+1
Although I have not checked. this will probably only work for n > 0. Hence.2
Useful Tricks
105
Then. and the upper limit x = cos π = −1. let us rewrite the integral:
−1
I=−
1
xdx
But.
.A. consider:
π
I=
0
cosn θ sin θ dθ
Then.
106
A
VECTOR IDENTITIES & TRICKS
.
where the surface is a cylinder. we must reconsider the charge enclosed by the Gaussian surface. radius r. which encloses charge in a volume V . and length l. where r > a. Also. So.1
Worked Examples
Long Beam of Charge . and divided by the total area of the beam. multiplied by the area of the surface. However note that j = vρ. we write Gauss’ law: E · dS =
S
1 ε0
ρV dV
V
Which will ﬁnd the electric ﬁeld through a surface S. of radius a. using Amperes law: B · d = µ0
S
j · dS
So. travelling along the z axis with velocity. Let us start by considering the electric ﬁeld. Hence. We see that it is the total charge. which will enclose a ˆ current I. Gauss’ law results in: 1 Er 2πrl = λel ε0 That is: λe Er = r>a 2πε0 r For inside the beam. and then compute the force the beam feels.107
B
B. we have: Qenc = ρl Adl = λel A
Hence. and (uniform) line charge density eλ. using Gauss’ law. Thus: I = vλez. if it has length l.Fields & Force
Consider a long straight beam of electrons. So: ρV dV = ρl Adl A
Where we have noted that the volume charge density is just the line charge density over the area. the volume is just the area times length. radius r. Consider a circle. for r ≥ a. let us compute the magnetic ﬁeld outside. Let us compute the electric and magnetic ﬁelds inside & outside the beam. the electric ﬁeld inside is given by: Er = λer 2πε0 a2 r<a
Let us now compute the magnetic ﬁeld. That is: λel 2 Qenc = πr πa2 Thus. Thus. Hence: Bθ 2πr = µ0 vλe
. Let us initially compute the charge enclosed by the beam.
then the force approaches zero.
λer µ0 v 2 λer − 2πε0 a2 2πa2 λ2 e2 r 1 − µ0 v 2 2πa2 ε0 λ2 e2 r 1 − ε0 µ 0 v 2 2πa2 ε0
Hence: Fr = λ2 e2 r 2πa2 ε0
a
1−
v2 c2
Thus. the total force is just: Fr =
0
Fr dr
Which just results in: Fr = λ2 e2 4πε0 1− v2 c2
Now. the cross product: ˆ ˆ ˆ r θ z 0 0 v 0 Bθ 0
v×B =
= −ˆvBθ r
Hence: Fr = λe = = Now. So. let us compute the force on the beam. if v << c. the magnetic ﬁeld inside is: Bθ = µ0 vλer 2πa2 r<a
Now. we also notice that c2 =
1 ε0 µ0 .
. High energy particle accelerators take advantage of this when focussing beams. If v ≈ c.108 Hence: Bθ = µ0 vλe 2πr r>a
B
WORKED EXAMPLES
Let us compute the magnetic ﬁeld inside. and the beam spreads out. hence the beam retains its size. notice. We write the Lorenz force law: F = q(E + v × B) Let us compute the force inside the beam. then the force is large. Let us use the same argument as we did for the charge enclosed: πr2 Ienc = µ0 vλe 2 πa Hence.
at the surface. And let us do so for the case that the medium is vacuum.2
Spherical Shell Charge Distribution: σ(θ) = σ0 cos θ
A spherical cavity of radius a is located in a linear dielecric medium.r ε0 E2. that is.⊥ − E1⊥ = ∂V2 ∂V1 σ0 − =− cos θ ∂r ∂r εr ε0 Have in mind that all the above relations are for r → a. we thus have: E2. at the surface. That is. Region (1) is inside. then this reads: ε2. Hence.⊥ − ε1. Hence: εr ε0 E2. Now. of permittivity εr . charge is distributed with a surface density σ(θ) = σ0 cos θ. and the unit normal points from the boundary to region ‘2’: outside. the ﬁrst thing to note. and (2) outside. Let us write down the relations between scalar potential. the medium inside is the same as that outside. the electric displacement ﬁeld has the relation: ˆ (D2 − D1 ) · n = ρs If 1 is inside.⊥ − D1. Also. Now. let us compute the monopole and dipole moments. that if the two regions had diﬀerent permittivities. On the surface of the cavity.r ε0 E1⊥ = σ0 cos θ Going back to our system: σ0 cos θ εr ε0 Now.⊥ = σ0 cos θ At r = a. is that the ‘cavity’ is more a ‘bubble’.⊥ − εr ε0 E1⊥ = σ0 cos θ Notice.B. where V is the scalar potential. where εr = 1 + χE . for shell of distributed charge.
. χE → 0. and 2 outside. we also have the relation that E = − V . Now.
Figure 10: The direction of the normal vector. that is: D2.2
Spherical Shell Charge Distribution: σ(θ) = σ0 cos θ
109
B. as we have that D = εr ε0 E. So. electric & displacement ﬁelds. for arguments sake. in the radial direction (which is the direction in which E⊥ is in).
So. we have that εr = 1. due to the charge distribution.
and hence the integral is zero. ignoring the constants) is:
−1 1
I=−
1
xdx =
−1
xdx = 1 (1 − 1) = 0 2
. Note. or. the volume element is given by: d3 r = r 2 sin θ dr dθ dφ Then we see that we can compute each ‘pole’contribution separately. we do a substitution: x = cos θ ⇒ dx = − sin θdθ
Hence. if we denote this as: ∞ 1 1 p V (r) = 4πε0 r +1
=0
Where: p ≡ r P (cos γ)ρ(r )d3 r
Where γ ≡ θ − θ. due to a multipole expansion. the dr integral picks out only the point r = a. So.110 Now. either we notice immediately that the two functions are orthogonal. the monopole moment) is given by: p0 ≡ q = That is (noting that P0 (cos γ) = 1): p0 = =
φ=0 θ=0 r=0
ρ(r )d3 r
ρ(r )r 2 sin θ dr dθ dφ
2π π ∞
δ(r − a)σ0 cos θ r 2 sin θ dr dθ dφ
Now. due to the δfunction. That is. and dipole n = 1. the charge distribution is given by (the axially symmetric): ρ(r ) = δ(r − a)σ(θ) = δ(r − a)σ0 cos θ The potential.e. the monopole contribution is just the n = 0 term. is given by: V (r) = 1 4πε0
∞ =0
B
WORKED EXAMPLES
1 r
+1
r P (cos γ)ρ(r )d3 r
Where γ ≡ θ − θ. So. So:
2π π
p 0 = σ0 a2
φ=0 θ=0 π
cos θ sin θ dθ dφ
That is: p0 = σ0 a2 2π
cos θ sin θ dθ
θ=0
To do this integral. total charge (i. the integral (just the integral.
This is given by the integral: p1 = = r P1 (cos γ)ρ(r )d3 r r cos γδ(r − a)σ0 cos θ r 2 sin θ dr dθ dφ
Where we have noted that P1 (cos γ) = cos γ. So. this step could have been done without inserting the expression for cos γ ﬁrst. Which easily gives:
−1 1
cos θ −
1
x2 dx = cos θ
−1
x2 dx
Thus: (i) → Now. We must remember that: cos γ = cos(θ − θ) = cos θ cos θ + sin θ sin θ Hence. φ integrals.B. the monopole moment is zero: p0 = 0 Now. just zero. which just give:
π
p1 = σ0 a3 2π
0
(cos θ cos θ + sin θ sin θ ) cos θ sin θ dθ
Notice. Hence. let us split this integral into its two additive parts: (i) (ii) cos θ cos θ cos θ sin θ dθ sin θ sin θ cos θ sin θ dθ
So that p1 = σ0 a3 2π[(i) + (ii)]. Now. we must be careful in continuing. we must be careful about what is primed. let us consider the dipole moment. and unprimed. (ii):
π
2 cos θ 3
sin θ
0
sin2 θ cos θ dθ
. Looking at (i) ﬁrst:
π π
cos θ cos θ cos θ sin θ dθ = cos θ
0 0
cos2 θ sin θ dθ
Which we can do via substitution x = cos θ . the integral becomes: p1 = r (cos θ cos θ + sin θ sin θ )δ(r − a)σ0 cos θ r 2 sin θ dr dθ dφ
Let us initially evaluate the r .2
Spherical Shell Charge Distribution: σ(θ) = σ0 cos θ
111
That is. Of course.
Inside. we have computed the scalar potential. we have: V1 = V2 = A r P (cos θ) B P (cos θ) r +1 r≤a r≥a
Now. with axial symmetry.112 We substitute x = sin θ : → sin θ
0
B
WORKED EXAMPLES
0
x2 dx = 0
Hence. in spherical polars. Now. (ii) → 0.and dipole) is: 1 1 1 p + 2 p1 V (r) = 1 0 4πε0 r r And is hence just: V (r) = That is: V (r) = 1 1 4 σ0 a3 π cos θ 4πε0 r2 3 σ0 a3 cos θ 3ε0 r2
It is useful just to point out that V (r) actually means that it is a function of all 3 positions. we have that the potential must not diverge. Hence. we reject its Bterm. we have that the solution to Laplace’s equation. to ﬁnd the constants. Now. due to the multipole expansion (i. up to the dipole term only. θ) = A r + +1 P (cos θ) r
=0
We have the case ‘1’ inside. we reject the outsides A term. φ).e. is of the form: ∞ B V (r. the dipole moment is: 2 p1 = σ0 a3 2π cos θ 3 That is: 4 p1 = σ0 πa3 cos θ 3
Hence. let us compute the exact analytical form of the scalar potential V (r). Thus. in the case that χE = 0. Hence: A a P (cos θ) = B P (cos θ) a +1
. We do so for both the regions inside r ≤ a and outside r ≥ a the ‘cavity’. Hence. and ‘2’ outside. as that would make the potentials diverge for r → ∞. that due to mono. the calculated contributions to the potential. hence. that is: V (r) = V (r. Similarly. let us initially apply the boundary condition that the potential is continuous at the boundary r = a. θ.
−1
B 1 P (cos θ) − A a P (cos θ) = − σ0 cos θ εr ε0 a +1 1 σ0 cos θ εr ε0
P (cos θ) =
this results in:
−1
A P (cos θ) + A a
P (cos θ) =
1 σ0 cos θ εr ε0
(2 + 1)A P (cos θ)a
=0
−1
=
1 σ0 cos θ εr ε0
. say. the LHS:
π
A a P (cos θ)P (cos θ) sin θdθ
0
Now. we can see that we will just pick out. we know that:
1
P (x)P (x)dx =
−1
0
2 2 +1
= =
Hence:
π
P (cos θ)P (cos θ) sin θdθ =
0
0
2 2p+1
= = = :
From this. another boundary condition we have is that of continuity of D⊥ at the boundary. and integrate. we apply a Fouriertype procedure. This. P (cos θ). is just: ∂V2 ∂V1 1 =− − σ0 cos θ ∂r ∂r r=a εr ε 0 That is: ∂ ∂r Giving: +1 B P (cos θ) + A a a +2 Using our derived relation B = A a2 ( + 1)a2 a +2 Which easily gives:
∞ +1 +1 . So. as it is just a dummy index: Aa = Hence: B = A a2
+1
B a +1
Now.2
Spherical Shell Charge Distribution: σ(θ) = σ0 cos θ
113
Now. Multiply both sides by a ‘diﬀerent’ polynomial.B. the term 2 B 2 A a = 2 +1 2 + 1 a +1 That is. we showed. to ﬁnd the coeﬃcients. reverting back to using . from the sum over . due to orthogonality of Legendre polynomials.
the coeﬃcients are given by: A = σ0 −1 ε ε 2a r 0
1 −1 1 π
cos θP (cos θ) sin θdθ
0
xP (x)dx
−1
=
σ0 εr ε0
1
xP (x)dx
−1
xP (x)dx
−1
So. considering a few Legendre polynomials. but will spend more time on in other examples). the potential inside: V1 = Thus: V1 = And. the LHS is:
π
(2 + 1)A a
0
−1
P (cos θ)Pp (cos θ) sin θ dθ
That is equal to. all higher Ai are zero (a fact we gloss over here. outside: V2 = B P (cos θ) r ≥ a r +1 A r P (cos θ) r≤a
σ0 r cos θ 3εr ε0
. we employ the same technique as we did before.114
B
WORKED EXAMPLES
Now. So. and integrate. namely to multiply by a diﬀerent polynomial. We shall be a bit more formal in the integration however. in that we specify the limits as the period of the polynomials. we have: 2A a Thus. putting x = cos θ: σ0 εr ε0 Hence. to ﬁnd the A . via orthogonality: (2p + 1)A a
−1
δ
2 = 2A a 2 +1
−1
The RHS is equal to. Thus. when multiplied and integrated: 1 1 σ0 cos θ → σ0 εr ε0 ε r ε0 Which is equivalent to. and the coeﬃcents they will generate: P0 (x) = 1 P1 (x) = x ⇒ ⇒ A0 = 0 A1 = σ0 3εr ε0
Infact.
let us examine this. one can immediately see that only the = 1 term will contribute. equating: α + β (3 cos2 θ − 1) = 2 cos2 θ − 1 2
.2
Spherical Shell Charge Distribution: σ(θ) = σ0 cos θ
+1 . for such a system. to rewrite the charge distribution: cos 2θ = 2 cos2 θ − 1 Thus: σ(θ) = σ0 (2 cos2 θ − 1) Now. We demonstrate this method in the next example. we have that the potential in all space is given by: V (r. the electric ﬁeld.
115
Using our derived relation B = A a2 V2 = Thus:
this is: A a2 +1 P (cos θ) r ≥ a r +1 σ0 a3 cos θ 3εr ε0 r2
V2 =
r≥a
Hence.2. in a medium with χE = 0. Then. let us suppose: σ(θ) = σ0 [αP0 + βP2 ] Then. We note that we can conceive that this is made up of two Legendre polynomials: P0 (cos θ) = 1 P2 (cos θ) = 1 (3 cos2 θ − 1) 2 So. θ) =
σ0 3εr ε0 r cos θ σ0 a3 3εr ε0 r2 cos θ
r≤a r≥a
Note. in summary. inside the cavity is actually constant: Er = − ∂V1 σ0 =− cos θ ∂r 3εr ε0
That is. Notice that σ(θ) = σ0 cos θ = σ0 P1 (cos θ). An alternative way to do this.
B. Let us compute the potential. the ﬁeld inside the cavity is uniform.1
Spherical Shell Charge Distribution: σ(θ) = σ0 cos 2θ
Suppose we have a sphericalshell charge distribution. by orthogonality of the P ’s.B. is to express the charge distribution as a Legendre polynomial. so that we have a charge density σ(θ) = σ0 cos 2θ at a radius a. at the stage of ‘ﬁnding coeﬃcients’. in all space. Let us use a trig identity. to express our charge distribution in terms of the above polynomials. which is actually a bit easier to see.
116 And, upon examination of the coeﬃcients, we see that: α = −1 3 β=
4 3
B
WORKED EXAMPLES
Hence, we have an expression for our charge distribution, in terms of Legendre polynomials: σ(θ) = σ0 [4P2 (cos θ) − P0 (cos θ)] 3
This is a useful way to represent the charge distribution, as it will allow us to evaluate the potential expansion exactly. Now, let us write the expansion of the potential: V (r, θ) = Ar + B P (cos θ) r +1
If we want the ﬁeld to converge (or at least, no diverge) as r → 0, ∞, we must discard the relevant terms, when evaluating the potential inside and outside the sphere. Thus: V1 = V2 = A r P (cos θ) B P (cos θ) r +1 r≤a r≥a
Now, we have that the potential is continuous at r = a. Hence: V1 (r = a) = V2 (r = a) ⇒ B = a2
+1
A
Where we have used a similar argument as we used in the other problem. Now, we have that the electric ﬁeld is discontinuous at the boundary, according to the surfacecharge distribution. Thus: ∂V2 ∂V1 − ∂r ∂r That is: (B (− − 1)a−
−2
=−
r=a
1 σ(θ) ε0
1 σ(θ) ε0 Using our derived relation between the inner and outer coeﬃcients, we end up with, after a little algebra: a1− σ(θ) A P (cos θ)(2 + 1) = ε0 −A a
−1
)P (cos θ) = −
Now, let us ﬁnd the coeﬃcients. To do this, we multiply both sides by Pm (cos θ) sin θ (the reason for the Pm is that of orthonormality; and that of the sin θ will become clear when we integrate for orthonormality), and integrate over the Legendre polynomials period 0 → π. Thus:
π
A (2 + 1)
0
P (cos θ)Pm (cos θ) sin θdθ =
a1− ε0
π
σ(θ)Pm (cos θ) sin θdθ
0
Let us consider the LHS. Let x = cos θ ⇒ dx = − sin θdθ. Then, also changing the limits, we have:
1
A (2 + 1)
−1
P (x)Pm (x)dx
B.2
Spherical Shell Charge Distribution: σ(θ) = σ0 cos θ = m:
1
117
This integrates to, taking
A (2 + 1)
−1
P (x)Pm (x)dx = A (2 + 1) a1− 2ε0
π
2 = 2A 2 +1
Hence, we have: A = σ(θ)P (cos θ) sin θdθ
0
Now, let us insert our expression for the charge distribution (and the reason for expressing it so will become clear): a1− σ0 π A = [4P2 (cos θ) − P0 (cos θ)]P (cos θ) sin θdθ 2ε0 3 0 Hence, we see that only = 0, 2 will produce nonzero coeﬃcients; due to the orthonormality of the Legendre polynomials. So, evaluating the = 0 case: A0 = − aσ0 π 2 P (cos θ) sin θdθ 6ε0 0 0 aσ0 = − 2 6ε0 aσ0 = − 3ε0
π
Where we have used the standard relation: P 2 (cos θ) sin θdθ =
0
2 2 +1
Which is put into the usual form:
1
P 2 (x)dx =
−1
2 2 +1
By using the substitution x = cos θ. Which is the reason for multiplying by sin θ previously. Let us compute the = 2 case: A2 = = =
π σ0 4 P 2 (cos θ) sin θdθ 6aε0 0 2 σ0 2 4 6aε0 5 4σ0 15aε0
Now, let us write down the potential inside the shell: V1 (r, θ) = A r P (cos θ)
= A0 P0 (cos θ) + A2 r2 P2 (cos θ) = A0 + A2 r2 P2 (cos θ) aσ0 4σ0 2 = − + r P2 (cos θ) 3ε0 15ε0 a σ0 a 4r2 = − 1 − 2 P2 (cos θ) 3ε0 5a
r≤a
118 And outside, using our derived relation between coeﬃcients: V2 (r, θ) = = − a2
+1 − −1
B
WORKED EXAMPLES
r
A P (cos θ) r≥a
σ0 a a 4a3 − 3 P2 (cos θ) 3ε0 r 5r
And therefore, we have, as the exact scalar potential: 4r2 σ0 a 1 − 2 P2 (cos θ) 3ε0 5a σ0 a a 4a3 V2 (r, θ) = − − 3 P2 (cos θ) 3ε0 r 5r V1 (r, θ) = − r≤a r≥a
B.3
Show That T µνλ Is a Tensor
From the relativistic electrodynamics section, we have the following deﬁnitions: T µνλ ≡ ∂ µ F νλ + ∂ ν F λµ + ∂ λ F µν F µν ≡ ∂ µ Aν − ∂ ν Aµ
And we have that the inverse transformation of a ﬁrst rank contravariant tensor: Bµ = ∂xµ ν B ∂x ν
Now, in the main text, we have shown that the ﬁeld tensor transforms as a second rank contravariant tensor: ∂x α ∂x β µν F αβ = F ∂xµ ∂xν And we have that the 4current density transforms as a ﬁrst rank tensor, as well as the contravariant derivative operator: ∂x µ ν ∂x µ ν Aµ = A ∂µ= ∂ ∂xν ∂xν So, let us show that T µνλ transforms as a third rank contravariant tensor. That is, we must show that the following is satisﬁed: ∂x µ ∂x ν ∂x λ σρπ T µνλ = T ∂xσ ∂xρ ∂xπ That is, expanding out the deﬁnition of T µνλ : ∂ µF
νλ
+ ∂ νF
λµ
+ ∂ λF
µν
=
∂x µ ∂x ν ∂x λ σ ρπ (∂ F + ∂ ρ F πσ + ∂ π F σρ ) ∂xσ ∂xρ ∂xπ
Now, let us pick oﬀ the ﬁrst term: ∂x µ ∂x ν ∂x λ σ ρπ ∂ F ∂xσ ∂xρ ∂xπ
can be written as: dpµ = qF µν uν dτ Let us start by discussing the rate of change of energy. We shall now start to take a diﬀerent “tack” on the problem.
B.B. Hence proven. for this ﬁrst term. charge q. the force on a charged particle is given by the Lorentz force law: F = q(E + v × B) And therefore: d dt = F ·v = qv · (E + v × B) = qv · E + qv · (v × B) = qE · v Where we have used the standard vector identity that a · (b × a) = 0. the work W done is scalar product of the force and displacement: W =F ·d Then the rate of change of energy is the rate of change of work: dW d = =F ·v dt dt Now. Now.4
Lorentz Force in Covariant Form
We wish to show that the Lorentz force and rate of change of energy of a particle. It is tedious to do the same for the other two terms.4
Lorentz Force in Covariant Form
119
Let us write the far right expressions in terms of their counterparts in the primed frame: ∂x µ ∂x ν ∂x λ ∂xσ ∂xρ ∂xπ α ∂ T ∂xσ ∂xρ ∂xπ ∂x α ∂x β ∂x γ We notice that most of this collapses into Kronecker deltas:
µ ν λ δα δβ δγ ∂ α T βγ βγ
And that leaves us with: ∂ µT
νλ
Which is clearly what we needed to prove. and it is plain that they transform in the same way. Let us denote the energy of the particle by . induced by electric & magnetic ﬁelds.
.
equating components: γd c dt dp1 γ dt dp2 γ dt dp3 γ dt = q(E1 γu1 /c + E2 γu2 /c + E3 γu3 /c) = q(E1 γ + B3 γu2 − B2 γu3 ) = q(E2 γ + B2 γu1 − B1 γu2 ) = q(E3 γ + B2 γu1 − B1 γu2 )
. the µth component of the LHS is the same as the µth component of the RHS. p) dτ =
B
WORKED EXAMPLES
dt γ
0 −E1 /c −E2 /c −E3 /c E /c 0 −B3 B2 = 1 E2 /c B3 0 −B1 E3 /c −B2 B1 0
Now. −γu)
Now. Let us write this out: γ d E1 γu1 /c + E2 γu2 /c + E3 γu3 /c c dt γ dp1 E1 γ + B3 γu2 − B2 γu3 dt dp2 = q E2 γ + B2 γu1 − B1 γu2 γ dt dp3 E3 γ + B2 γu1 − B1 γu2 γ dt That is. notice: dpµ dτ = γ dpµ dt uµ = (cγ. let us state the following 4vectors: uµ = (cγ.120 Now. let us consider again the original equation: dpµ = qF µν uν dτ That is. γu) And the ﬁeldtensor: F µν pµ = ( /c. let us multilply the ﬁeld tensor by the velocity 4vector: 0 −E1 /c −E2 /c −E3 /c E1 /c 0 −B3 B2 = E2 /c B3 0 −B1 E3 /c −B2 B1 0 E1 γu1 /c + E2 γu2 /c + E3 γu3 /c E1 γ + B3 γu2 − B2 γu3 = E2 γ + B2 γu1 − B1 γu2 E3 γ + B2 γu1 − B1 γu2 cγ −γu1 −γu2 −γu3
F µν uν
Now.
1)
. notice: dp d F = = q(E + v × B) =v·E dt dt Examination of what we have from unpacking our notation.4
Lorentz Force in Covariant Form
121
Cancelling oﬀ various factors: d dt dp1 dt dp2 dt dp3 dt = q(E1 u1 + E2 u2 + E3 u3 ) = q(E1 + B3 u2 − B2 u3 ) = q(E2 + B2 u1 − B1 u2 ) = q(E3 + B2 u1 − B1 u2 )
The latter 3 equations are components of a crossproduct. and the ﬁrst of a scalar product. both the Lorentz force law & the rate of change of energy of a particle: dpµ = qF µν uν dτ Thanks to J.B. will reveal that the following equation has. embedded inside. and what we have above. Finally.Agnew for helping out with this derivation! (B.
122
B
WORKED EXAMPLES
.
5)
And is known as the generalised Legendre equation. θ. have have that the Laplace equation 2 Φ = 0 has the form: 1 ∂2 ∂ 1 ∂Φ 1 ∂2Φ (rΦ) + 2 sin θ + 2 2 =0 (C. has solutions: Qm (φ) = e±imφ So. we can assign constant terms to this equation.
. φ) coordinate system). and multiply by r2 sin2 θ
giving:
1 d dP 1 d2 Q 1 d2 U + sin θ + =0 U dr2 P r2 sin θ dθ dθ Q dφ2 1 d2 Q = −m2 Q dφ2
Now. the (r. this is more conventionally written. so that: (C.e.
Let us collect some terms together. let us use another constant term: r2 d2 U = ( + 1) U dr2 Then we have: ( + 1) + That is: 1 d dP m2 sin θ − =0 P sin θ dθ dθ sin2 θ (C. Thus: U (r) Φ(r. as: dP d (1 − x2 ) + dx dx ( + 1) − m2 P =0 1 − x2 (C.1) 2 r ∂r r sin θ ∂θ ∂θ r sin θ ∂φ2 Now. Hence. with x ≡ cos θ. let us assume that we can solve the equation using the standard ‘separation of variables’ technique.3)
1 d dP m2 sin θ + ( + 1) − P =0 sin θ dθ dθ sin2 θ Now.123
C
Legendre Equation & Spherical Harmonics
In spherical polars (i.2)
Which. solutions of which are known as the associated Legendre polynomials. we have that: r PQ d2 U 1 d dP 1 d2 Q + 2 U Q sin θ + 2 2 UP 2 = 0 dr2 r sin θ dθ dθ dφ r sin θ
r2 sin2 θ RP Q .4) 1 d2 U 1 d dP + sin θ − m2 = 0 2 2 sin θ dθ U dr Pr dθ (C. φ) = P (θ)Q(φ) r Where the usual R(r) is obviously just R(r) = U (r) . θ. we have: r2 sin2 θ This is just: r2 d2 U 1 d dP m2 + sin θ − =0 2 U dr P sin θ dθ dθ sin2 θ Again. semitrivally.
4).6)
C. Now. we have: r2 d2 U = ( + 1) U dr2 Let us suppose that the solution is a power law. Then. Then: r2 d2 α (r ) = r2 r−α α(α − 1)rα−1 rα dr2 = α(α − 1) So.1
Power Series Solution to the Ordinary Legendre Equation
Now. Let us consider the case where m = 0. so that the function must be well behaved on −1 ≤ x ≤ 1. we have that the radial solution is of the form (putting in arbitrary constants): U (r) = Ar We have also used the form R(r). where we have that θ ∈ [0. we will assume that the solution takes on a power series form:
∞
P (x) = xα
j=0
aj xj = aj xj+α
. π]. its not immediately obvious how to solve this. Also. we see that it is equivalent: (α + )[α − − 1] = α2 − α − 2 − Hence equivalent. it is: R(r) = U (r) = Ar + Br− r
−1 +1
+ Br−
(C.124 C. two solutions: α=− α= +1 Hence.0. (C. by writing in this form. but. that is. in (C.5). we have that: α(α − 1) = ( + 1) ⇒ α2 − α −
2
− =0
Now. So.7)
Now.1 Radial Solution
C
LEGENDRE EQUATION & SPHERICAL HARMONICS
Let us quickly derive the solution to the radial part of the Laplace equation. the resulting equation is known as the ordinary Legendre diﬀerential equation: d dP (1 − x2 ) + ( + 1)P = 0 dx dx (C. something that was not azimuthally symmetric. That is. if the following expression is written. we are able to see its solutions: (α + )[α − − 1] = 0 That is. Let us try U (r) = rα . recall that x = cos θ. we had a dependance upon m. hence.
we see that there is a cutoﬀ.8)
. and the requirement that the series remain ﬁnite. the total equation is: aj (j + α)[(α + j − 1)xα+j−2 − (α + j + 1)xα+j ] + ( + 1)aj xj+α = 0 Collecting likepowers of x. We only have the cases for α = 0. and that the series must converge for x ≤ 1. 1 that terminate. not both. This is down to the recursive relation. and putting summations back in:
∞ ∞
aj (j + α)(α + j − 1)xα+j−2 −
j=0 j=0
aj [(j + α)(α + j + 1) − ( + 1)]xj+α = 0
Now. We also have that ≥ 0. and they produce even. change summation index from j to m = j − 2.1
Power Series Solution to the Ordinary Legendre Equation
125
So. Then. odd polynomials.C. Essentially. the ﬁrst expression becomes:
∞
am+2 (m + 2 + α)(α + m + 1)xα+m
m=2
Then. this will give us a recursion relation: aj+2 = (α + j)(α + j + 1) − ( + 1) aj (α + j + 1)(α + j + 2)
From this. we are able to (but is not done here) ﬁnd a generation function: the Rodrigues’ formula: P (x) = d (x2 − 1) 2 ! dx 1 (C. In the ﬁrst expression. it will give us the Legendre polynomials. and integer. we apply a standard ‘trick’ to solving these things further. By a manipulation of the power series. of order : P) (x) = 1 P1 (x) = x
1 P2 (x) = 2 (3x2 − 1) 1 P3 (x) = 2 (5x3 − 3x)
Notice that each polynomial possesses only even or odd power. calculating the necessary expressions (suppressing the summation signs): dP dx = xα aj jxj−1 + αxα−1 aj xj = aj jxα+j−1 + αaj xα+j−1 = (aj j + αaj )xα+j−1 ⇒ (1 − x2 ) dP dx = (1 − x2 )(aj j + αaj )xα+j−1 = (aj j + αaj )[xα+j−1 − xα+j+1 ] ⇒ d dP (1 − x2 ) dx dx = aj (j + α)[(α + j − 1)xα+j−2 − (α + j + 1)xα+j ]
Thus.
126 C. for = . let us write down the Legendre equation again: dP (x) d (1 − x2 ) + ( + 1)P (x) = 0 dx dx Let us multiply the equation by P (x). This is easily shown to be just:
1
Now. we are able to get to the standard orthogonality condition:
1
P (x)P (x)dx =
−1
2 δ 2 +1
(C. and integrate over the interval:
1
P (x)
−1
d dP (x) (1 − x2 ) + ( + 1)P (x) dx = 0 dx dx
Multiplying through:
1
P (x)
−1
d dP (x) (1 − x2 ) + ( + 1)P (x)P (x)dx = 0 dx dx
Now.e. by parts:
1
P (x)
−1
d dP (x) (1 − x2 ) dx = dx dx =
P (1 − x2 )
1
dP dx
1
1
−
−1 −1
(1 − x2 )
dP dP dx dx dx
(x2 − 1)
−1
dP dP dx dx dx
As we have that the series terminates at ±1. integrating the ﬁrst expression. its ‘square’) then we must actually compute the integral. However. write this down again. then subtract. to calculate the integral for the = (i. the integral is therefore:
1
N
≡
−1
P 2 (x)dx 1 2 ( !)2 2
1 −1
=
d d (x2 − 1) (x2 − 1) dx dx dx
After some nontrivial integration (which is not repeated here). So. using Rodrigues’ formula.9)
Which says that if = . Putting this back in:
1
(x2 − 1)
−1
dP dP + ( + 1)P (x)P (x)dx = 0 dx dx → . using the explicit form of the Legendre polynomials. but with
[ ( + 1) − ( + 1)]
−1
P (x)P (x)dx = 0
(C.10)
.1
C
LEGENDRE EQUATION & SPHERICAL HARMONICS
Orthogonality of Legendre Polynomials
Now.1. that is. then the integral must vanish.
1] can be expressed in terms of them. that is.
1
P (x)f (x)dx = a
−1
We can. the associated Legendre polynomial P m (x). namely.1.13)
. So.11)
In a similar fashion as we used previously. Hence: 2 2 +1
Where the summation will just pick out a single value. thus:
∞
f (x) =
=0
a P (x)
To ﬁnd the coeﬃcients a . for problems without azimuthal symmetry. Hence: a = 2 +1 2
1
P (x)f (x)dx
−1
C. We basicaly have the generalisation of the Legendre polynomial P (cos θ). we consider the solutions to (C. as the Legendre polynomials form a complete set of orthogonal functions. and integrate:
1 ∞
P (x)f (x)dx =
−1 =0
a
P (x)P (x)dx
Now. of course. as being nozero. any function on [−1. m. we have an orthogonality relation:
1
P m (x)P m (x)dx =
−1
2 ( + m)! δ 2 + 1 ( − m)!
(C. Thus:
1
P (x)f (x)dx =
−1
a
2 δ 2 +1 . for m = 0. The version of Rodrigues formula here is now: P m (x) = Now. the integral on the RHS is just the orthogonality relation.2
Associated Legendre Polynomials & Spherical Harmonics Using Legendre Polynomials as a Basis
127
C. we apply a standard ‘fouriertype’ technique: Multiply both sides by another polynomial. we must solve the equation for arbitrary . back to .5).2
Associated Legendre Polynomials & Spherical Harmonics
Here.12) d +m 2 (−1)m (x − 1) (1 − x2 )m/2 2 ! dx +m (C. relabel the indices.2
Now. it can be shown that: P −m (x) = (−1)m ( − m)! m P (x) ( + m)! (C.C.
φ)
= (−1)m Y ∗ (θ. θ . θ. thus: P (cos γ) =
m
A
m (θ
. φ)f (θ. so that: a
m
=
Y ∗ (θ. φ)
. (r . φ) sin θdθdφ m
Just as way of convenient notation. φ ). φ)
(C. in a similar way to before. and are given by the normalised product: Y
m (θ. dΩ ≡ sin θdθdφ. 2π]. φ) = 2 +1 P (cos θ) 4π
An arbitrary function may be expanded in terms of spherical harmonics. We call these solutions spherical harmonics. φ )Y m (θ. we can see that: Y
−m (θ.3
The Addition Theorem for Spherical Harmonics
Consider two coordinate vectors. φ )Y
m (θ. they form a complete set on φ ∈ [0. The addition theorem states: 4π Y ∗ (θ .16)
Note. φ)Y
m (θ. from (C. consider that the above Legendre polynomial may be expanded in terms of spherical harmonics. φ)f (θ. as we have that P m (cos θ) form a complete set on θ ∈ [0. φ)dΩ m
C.3). φ) =
m
a
m Y m (θ. for m = 0.17)
Where the coeﬃcients are found. we can imagine that some combination will give us full solutions. x. to be: a
m
=
Y ∗ (θ.14)
So. we use the solid angle element.12). we just have: Y 0 (θ. recall that we had Qm as the azimuthal solution.128
C
LEGENDRE EQUATION & SPHERICAL HARMONICS
Now. x with coordinates (r. To prove the theorem. Thus. φ) sin θdθdφ
= δ δmm
(C.15)
The normalisation/orthogonality relation is:
2π φ=0 π θ=0
Y ∗m (θ. φ). Now. π]. φ) m
(C. thus: f (θ. (C. φ) P (cos γ) = 2 +1 m m Where cos γ = cos θ cos θ + sin θ sin θ cos(φ − φ ). on the unit sphere. φ)
=
2 + 1 ( − m)! m P (cos θ)eimφ 4π ( + m)!
(C.
Figure from Jackson
By various symmetry arguements. for spherical polars. φ ) = Y ∗ (θ.20)
(C.19) (C.22) (C. φ )Y
m (θ. φ)
Where the coeﬃcients are given by (found in the usual way): Am (θ . φ):
Below are some spherical harmonics Y =0
1 Y00 = √ 4π =1 Y11 = − Y10 = =2 Y22 = Y21 Y20 15 sin2 θe2iφ 2π 15 = − sin θ cos θeiφ 8π 5 1 = ( cos2 θ − 1) 4π 2 1 4 3 sin θeiφ 8π 3 cos θ 4π
(C.21) (C. φ)P (cos γ)dΩ m
C. Thus:
Am (θ .18)
(C.4
Some Spherical Harmonics
m (θ.23)
. we see that we have only the case where = P (cos γ) =
m
in the sum.C.4
Some Spherical Harmonics
129
Figure 11: The addition theorem.
. Also.27)
−m
= (−1)m Y ∗ m
(C. t) =
n=0
tn Pn (x) √ 1 1 − 2tx + t2
(C.33)
ntn−1 Pn (x) = (x − t)(1 − 2tx + t2 )−3/2
n=0
Now. − 1. . there are 2 + 1 diﬀerent spherical harmonics: m = . .30) (C. . 0. for a given . multiplied by the Legendre polynomial: Y 0 (θ. let us ﬁnd both the x.25) (C.32) (C.31)
=
Now. . we suppose that a function will generate the polynomials.26) (C.5
Generating Function
For the Legendre polynomials.130 =3
C
LEGENDRE EQUATION & SPHERICAL HARMONICS
Y33 = − Y32 Y31 Y30 Note. φ) = 2 +1 P (θ) 4π (C. we have previously written: Y
35 sin3 θe3iφ 4π 1 105 sin2 θ cos θe2iφ = 4 2π 1 21 = − sin θ(5 cos2 θ − 1)eiφ 4 4π 7 5 = ( cos3 θ − 3 cos θ) 2 4π 2 1 4
(C. remember. − Also note. that a spherical harmonic. . then we will check that it does:
∞
f (x. . with m = 0 is just the normalisation constant. let us express the ﬁrst in the following way: t(1 − 2tx + t2 )−3/2 = t(1 − 2tx + t2 )−1 (1 − 2tx + t2 )−1/2 = t(1 − 2tx + t2 )−1
n
tn Pn (x)
.and tderivatives of this expression: d : dx d : dt
∞ ∞
tn Pn (x) = t(1 − 2tx + t2 )−3/2
n=0
(C.29)
C.28)
Which gives a way of generating the unwritten negative m harmonics above.24) (C. − + 1. .
35). Now.34) and (C. let us go back a bit. giving: Pn (x) = Pn+1 (x) − 2xPn (x) + Pn−1 (x) (C. giving: (n + 1)Pn = Pn+1 − xPn Now. in the above. together: Pn = Pn+1 − 2xPn + Pn−1 nPn = xPn − Pn−1 From these.35)
.C. now we equate powers of t. we may relabel each summation variable thus: ⇒ tn+1 Pn+1 (x) − 2xtn+1 Pn (x) + tn+1 Pn−1 (x)
Hence.34)
Now. If we mutiply (C. to be able to equate coeﬃcients: xtn Pn − tn Pn−1 − ntn Pn = 0 Hence: xPn − Pn−1 = nPn Now.33) by t. n starts from zero. in each sum above.5
Generating Function
131
Hence: tn Pn (x) = t(1 − 2tx + t2 )−1
n n
tn Pn (x)
That is: t
n
tn Pn (x) = (1 − 2tx + t2 )
n
tn Pn (x)
Which is just: tn+1 Pn (x) =
n n
(tn − 2xtn+1 + tn+2 )Pn (x)
Solve solve this (by equating coeﬃcients of like powers of t). then they should be equal. Hence: (x − t)tn Pn (x) = ntn P (x)
n n
That is: xtn Pn − tn+1 Pn − ntn Pn = 0 Again. let n → n + 1: nPn−1 = Pn − xPn−1 (C. we note the following: (tn − 2xtn+1 + tn+2 )Pn (x) =
n n
tn Pn (x) − 2x
n
tn+1 Pn (x) +
n
tn+2 Pn (x)
Where. shifting the summation variable.32) by (x − t). and (C. let us write down (C. eliminate Pn−1 .
t):
∞
f (1. we can do so in terms of orthogonal functions.132
C
LEGENDRE EQUATION & SPHERICAL HARMONICS
Now. on the RHS [(1 − x2 )Pn ] = −n(n + 1)Pn Hence. Let us assume r1 > r2 . t) =
n=0
Pn (1)tn 1 (1 − 2tx + t2 )1/2 1 1−t tn
n
= = = Hence.35) again: xPn − Pn−1 = nPn . using (C. we have shown: [(1 − x2 )Pm (x)] = −m(m + 1)Pm (x) Hence completing the proof. Let us compute f (1. giving: x2 Pn − xPn−1 + nPn−1 = nxPn + Pn − xPn−1 Tidying up: (1 − x2 )Pn = n(Pn−1 − xPn ) Diﬀerentiate both sides. So: 1 r1 − r2  = = 1 r1 1
2 r1
+
2 r2
− 2r1 r2 cos θ 1
1 + ( r2 )2 − 2 r2 cos θ r1 r1
Now. we see that Pn (1) = 1. as: t= r2 r1 x = cos θ
.5. with respect to x: [(1 − x2 )Pn ] = n(Pn−1 − Pn − xPn ) Now. in the angle between the two vectors.35) times x to the above. C. add (C.1 Application: Expand
1 r1 −r2 
We often need to expand: 1 r1 − r2  Now. we identify the expression with the generating function of the Legendre polynomials.
5
Generating Function
133
Hence.
.C.36)
For r2 > r1 . we have: 1 1 = r1 − r2  r1
∞ n=0
r2 r1
n
Pn (cos θ)
(C. We can obviously rearrange the indices if the reverse is true.
134
C
LEGENDRE EQUATION & SPHERICAL HARMONICS
.
infact.135
D
Books
There are a number of very useful texts. • Woodhouse: Special Relativity. but grounds it ﬁrmly in physical ideas with plently of examples. Throughout the course of writing this. • Shutz : A First Course in General Relativity. so is well worth looking at. but is quite light if one is looking for a more formal introduction. • Heald & Marion: Classical Electromagnetic Radiation. but is very light on the relativistic part of the subject. The book relevant for electrodynamics is their Vol 2. • Zwiebach: A First Course in String Theory. regardless. classical theoretical physics series by Landau & Lifshitz are highly recommended. so goes way beyond the scope of the current treatise. Heald & Marion is a very good text for many of the electrostatic problems. Zwiebach is an excellent start point for Lorentz transformations & simple tensor manipulations. and especially in spherical harmonic theory. regarding this topic of Electrodynamics. I have consulted many of the below books. consulting Schutz is heavily recommended. The latter half of the book is on general relativity. as it introduces tensors very well. It has an excellent section on the nonrelativistic part. designed for graduate courses in the subject. • Jackson: Classical Electrodynamics. The ﬁnal 4 books were consulted mainly for tensor theory. introducing special relativity in a very abstract manner. it is very good for the multipole expansion. Hence. frankly. going into a lot of detail.
. most introductory general relativity texts tend to have a large section devoted to tensor theory & very helpful examples & hints for manipulations. The. • Landau & Lifshitz : The Classical Theory of Fields. Jackson is a classic text. Woodhouse is a very mathematical text (whereas the others are ﬁrmly grounded in physics). Schutz introduces tensors in a very formal manner.