8.

ELECTROSTATICS AND
MAGNETOSTATICS

Matthew Baring — Lecture Notes for PHYS 532, Spring 2023

1 Coulomb’s Law

With the complete set of electromagnetic field equations assembled, the task L&L
now is to assemble more details of how charges generate and respond to Sec. 36
fields in configurations that approximate those commonly encountered. The
starting point is electrostatics, where the electric field is time-independent
and the magnetic field is zero. The two pertinent Maxwell’s equations are

∇ · E = 4πρ and ∇ × E = 0 . (1)

The latter can be expressed in terms of the potential φ(r) using the definition
E = −∇φ , and the former then takes the form of Poisson’s equation:

∇2 φ = −4πρ . (2)

This can be solved for simple charge configurations using the mathematical
technique of separation of variables. When there is no charge, the solutions
are necessarily linear functions of coordinates, that if bounded at infinity,
correspond to a trivial φ = constant solution.

For a single point charge q at the origin, the simplest way to determine the
field form is to use Gauss’ theorem from vector analysis to convert a volume
integration of Gauss’ law to a surface integral. Thus
Z Z Z
4πq = 4πρ dV = ∇ · E dV = E · dΣ = 4πE r2 (3)

1
since the potential must be spherically symmetric and the field therefore
radial as it threads Σ isotropically. The magnitude and direction of the
electric field are then described via Coulomb’s law:
qr̂ q
E = ⇒ φ(r) = . (4)
r2 r

The potential form is simply obtained using the gradient operator.

For a system of charges, this form can be superposed to obtain the contribu-
tion from the entire ensemble:
X q X
φ = n
⇒ ∇2 φ = −4π qn δ 3 (r − rn ) . (5)
|r − rn |
n n

We therefore can identify the Green’s function 1/|r − rn | for the Coulomb
problem for any stationary charge configuration.

Motion in the Coulomb Field: The pedagogy will only cover charge
motions in a static Coulomb potential through problems; this is the expedient L&L
path to learning. Sec. 39

• For unbound charges interacting with a stationary test charge, the in-
teraction leads to a scattering deflection of the mobile charge, whether the
interaction is attractive or repulsive.

• For bound interactions, which are necessarily attractive (i.e., between

charges of opposite sign), the relativistic nature of the interaction leads to
unclosed orbits, in distinct contrast to the closed elliptical trajectories for
the motion of non-relativistic charges in the Coulomb field.

2
1.1 Electrostatic Energies of Charges

The Coulomb field possesses an energy that is readily expressible via L&L
1
Z Sec. 37
Uem = E 2 dV . (6)
8π
Since the potential scales as 1/r , the electric field magnitude E ∼ 1/r2 and
the integration down to zero radii is divergent. This poses a problem that
underlines the limitations of a classical theory of electromagnetism on small
scales. We can re-express the Coulomb field energy using E = −∇φ :
Z Z Z
1 1
 1
Uem = − E · ∇φ dV = − ∇ Eφ dV +
 φ ∇ · E dV . (7)
8π 8π 8π
This is essentially an integration by parts. The first integral on the right is of
a pure derivative and is thus zero if the volume extends to infinity. The sec-
ond integral on the right can be re-written using Gauss’ law for electrostatics,
∇ · E = 4πρ . The result is the compact form
Z
1 1 X
Uem = ρ φ dV → qn φ(rn ) . (8)
2 2
n

The second form here is the discrete charge ensemble distillation of the energy
integral, with φ(rn ) being the potential resulting from the ensemble at the
point of charge qn . For a single charge, the potential diverges at its position,
and so one deduces an infinite self-energy for electrostatics.

• This divergence is untenable, and defines a limitation of classical elec-

trodynamics. Such a self-energy could, in principle, serve as an attribution
for mass. It certainly cannot exceed mc2 . Thus, for an electron, the validity
of classical electrodynamics must be restricted to radius domains such that
e2 e2 −13
. me c2 ⇒ r & r0 ≡ 2 ≈ 2.818 × 10 cm . (9)
r me c
Accordingly, we define the fundamental limiting scale of classical E/M, namely
the classical electron radius r0 ; at comparable lengths, quantum mechan-
ics must be introduced, arising naturally on scales ∼ ~/me c = r0 /αf .

• Specifying a self-energy is therefore an approximate practice, and even in

quantum mechanics, divergences appear and have to be eliminated through
renormalization techniques.

3
1.2 Fourier Transform of Electrostatic Fields

As a preparation for wave elements of our electrodynamics pedagogy, it is L&L

instructive to consider the spatial Fourier transform of the Coulomb field. Sec. 51
This is germane to structured static systems of charges, such as in a lattice.
For single point charge at the origin, Poisson’s equation assumes the form

∇2 φ = −4πq δ 3 (r) . (10)

Let us decompose the field in terms of its 3D space Fourier components φk

via the transform
Z ∞
d3 k
Z
φ = ik·r
e φk 3 ⇔ φk = e−ik·r φ(r) d3 x . (11)
−∞ (2π)

Here, k is the wavevector of the field transform, with λ = 2π/|k| being the
effectively wavelength or spatial scale of longitudinal potential variations.
Observe that this construction employs an asymmetric convention for the
Fourier transform/inverse transform pair. Poisson’s equation yields
Z ∞
d3 k 3
Z
2 2 ik·r 3 ik·r d k
∇φ = (ik) e φk = −4πq δ (r) = −4πq e .
−∞ (2π)3 (2π)3
(12)
The last identity stems from the Fourier representation of the Dirac delta
function. Equating integrands, it then follows that

4πq
φk = (13)
k2

is the Fourier transform of the 3D Coulomb potential. If one extends this

to a screened Coulomb field (say in a classical model of atom) with a
potential φ(r) = qe−µr /r , then the Fourier transform becomes

4πq
φk = . (14)
k 2 + µ2

Accordingly, 2π/µ defines the lengthscale for screening of the bare Coulomb
potential. This is pertinent to the Thomas-Fermi model of atoms, and Debye
screening in warm plasmas.

4
 • The electric field Fourier component Ek is simply ascertained by taking
the gradient. Thus,
Z ∞
d3 k
E ≡ eik·r Ek
−∞ (2π)3
Z ∞
d3 k
= −∇ eik·r φk (15)
−∞ (2π)3
Z ∞
d3 k
= −i k eik·r φk .
−∞ (2π)3

The Fourier transform of the electric field is thus

4πq k
Ek = −i k φk = −i . (16)
k2

2 Field of a Charge in Uniform Motion

To prepare for the radiation portions of the course, we need to characterize

fields of charges in uniform motion. Let the rest frame of a charge be K 0 , L&L
moving with a velocity v = vx̂ = βcx̂ with respect to our K frame. Sec. 38

Plot: K frame and frame K 0 in which charge is at rest.

The scalar potential is easily transformed between the two frames, since the
vector potential in the charge rest frame is A0 = 0 . Thus,

φ0 q
φ = p = 0p for (r0 )2 = (x0 )2 + (y 0 )2 + (z 0 )2 . (17)
1−β 2 r 1 − β2

These K 0 frame coordinates need to be expressed in terms of our K frame

coordinates (t, x, y, z) . This is effected using the Lorentz boost relations:

x0 = γ(x − βct) , y0 = y , z0 = z . (18)

It follows then that

(r0 )2 = γ 2 (x − βct)2 + y 2 + z 2 , (19)

5
Two Inertial Frames in Relative Motion
which can be substituted directly into the Coulomb form. The vector poten-
tial in the K frame can be similarly prescribed using A = γφ0 v/c = φv/c ,
since A0 = 0 . If one then sets

2 (r0 )2
r∗ ≡ = (x − βct)2 + 1 − β 2 (y 2 + z 2 ) ,


2 (20)
γ
then the components of the four-potential in the K frame are
q qv qβ
φ = , A = = ∗ x̂ . (21)
r∗ cr∗ r

The electric field can be obtained in either of two ways. One path is to use
the Coulomb field form in the charge’s rest frame and Lorentz transform it
using our standard boost relations. This gives
qx0 γqy 0 γqz 0
Ex = Ex0 = , Ey = γEy0 = , Ez = . (22)
(r0 )3 (r0 )3 (r0 )3
This is the simplest approach, yet the same equations will result by taking the
4-gradient of the 4-potential (φ, A) using the expression for r0 in Eq. (19).
The field can then be written in vector notation:
1 ∂A qr
E ≡ −∇φ − = 2 ∗ 3 for r = (x − βct, y, z) . (23)
c ∂t γ (r )
This is the first indication that there is a retardation or time delay in
establishing the field: causality applies in communicating electromagnetic
information from a moving charge.

If the radius vector r makes an angle θ to the boost direction, then clearly
y 2 + z 2 = r2 sin2 θ and (x − βct)2 = r2 cos2 θ . This leads to a re-expression
for the electric field vector form:
qr
−3/2
E = 2 3 1 − β 2 sin2 θ since (r∗ )2 = r2 1 − β 2 sin2 θ


. (24)
γ r
This highlights the angular dependence of the electric field relative to the
boost direction. The two extreme cases are parallel ( θ = 0 ) to and perpen-
dicular to ( θ = π/2 ) the boost:
1 q q
Ek ≡ E(θ = 0) = , E⊥ ≡ E(θ = π/2) = γ (25)
γ 2 r2 r2