Annales de la Fondation Louis de Broglie, Volume 30 no 1, 2005 87

Free falling electric charge in a static

homogeneous gravitational field

Luca Fabbri

Dipartimento di Fisica dell’Università di Bologna

46 Via Irnerio, 40126 Bologna, Italy

ABSTRACT. We reconsider the problem of a free falling electric charge

in a static homogeneous gravitational field, specifically in a space-time
domain in which the Riemann tensor vanishes and no electromagnetic
field is present. We choose to describe the radiation emitted by the
charge in terms of a general covariant quantity. We show that, under
these assumptions, the charge, however accelerated, does not radiate,
so that no contradiction arises with the Principle of Equivalence, which
remains valid also for charged matter.
RÉSUMÉ. On reconsidère le problème d’une particule chargée en chute
libre en un champ gravitationnel statique et homogène, spécifiquement
dans un domaine d’espace-temps où le tenseur de Riemann s’annule
et en absence de champs éléctromagnétiques. On choisit de décrire la
radiation émise par la charge en termes d’une quantité covariante en
sens général. On montre que, dans ces hypothèses, la charge, n’im-
porte comment accélérée, n’emit pas de radiation, de façon qu’on ne
relève aucune contradiction avec le principe d’equivalence, qui garde sa
validité en présence de matière chargée.

1 Introduction
Einstein’s Principle of Equivalence does not draw any explicit distinction
between charged and neutral matter, whereas charged and neutral par-
ticles behave quite differently according to the Lorentz-Dirac equation.
Since both the Principle of Equivalence and the Lorentz-Dirac equation
describe the motion of a particle, but they do it in such a different way,
a contradiction seems to arise.
More specifically, consider an electric charge in free fall in a static
homogeneous gravitational field: an observer at rest with respect to the
matter generating the gravitational field sees the charge in an accelerated
88 L. Fabbri

motion (with respect to a system of reference which may be supposed

to be inertial in the traditional sense) and so he would be tempted to
conclude that the charge radiates; from the point of view of a free falling
observer, the charge is at rest in a locally inertial reference frame, so that,
on the basis of the Principle of Equivalence, he would instead conclude
that the charge does not radiate.
The purpose of the present work is to give another contribution to
the solution of this (apparent) contradiction.
In Sec. 1 we briefly rediscuss the Principle of Equivalence and the
Lorentz-Dirac equation of motion. In Sec. 2 we recall the latest and,
in our view, more interesting attempts at solving the problem, trying to
explain why we consider them not yet fully satisfactory, while, at the
same time, outlining the framework and the main lines of our approach.
The calculations necessary to obtain the solution of our problem are
developed in Sec. 3, where we also re-examine the scope of the Principle
of Equivalence in the light of the conclusion reached about the question
at hand.

2 The Principle of Equivalence

Consider a torsion-less space-time, so that (via Weyl’s theorem) it is
always possible to associate a locally inertial system of reference to any
space-time point.

Principle of Equivalence 1 A free falling system of reference in a

gravitational field is a locally inertial system of reference, that is, the laws
of physics are valid in it in the form established by Special Relativity.

We shall need to analyse the problem in an extended space-time

region. We therefore need considering a formal set-up in which global and
local statements are equivalent. Since the connection is the mathematical
tool that describes accelerations, and thereby gives information about
inertiality, the above demand translates into the following mathematical
assumption: if a system of reference is found such that the connection
vanishes at a point, then it vanishes everywhere in that system.
There exists a theorem that translates the above assumption to an
equivalent one, but in a form for which is used a tensorial language: in
this formulation, one can say the Riemann tensor vanishes everywhere.
µ
Rαβρ =0 (1)
Free falling electric charge . . . 89

A gravitational field for which the Riemann tensor is null is called a

static homogeneous gravitational field [1].
With this assumption, we may give the statement of the Principle of
Equivalence in the following form:

Principle of Equivalence 2 A free falling system of reference in a

static homogeneous gravitational field is a (globally) inertial system of
reference, that is, the laws of physics are valid in it in the form established
by Special Relativity.

This is the form of the Principle of Equivalence that will be considered

hereafter.

3 Equation of motion
The most general equation of motion is the Lorentz-Dirac equation, that
we can write for a Riemannian space in which the metric has signature
(1, −1, −1, −1) as follows

δuµ 2 δ 2 uµ δuα δuα µ

m = eg µα Fαβ uβ + e2 ( 2 + uµ ) + e2 T (Rαβρ ) (2)
δs 3 δs δs δs
µ
where T (Rαβρ ), called tail, is related to the Riemann tensor and vanishes
if and only if the Riemann tensor vanishes (see [2] and [3]).
µ
Taking into account condition Rαβρ = 0, we shall deal with a Lorentz-
Dirac equation of the form

δuµ 2 δ 2 uµ δuα δuα

m = eg µα Fαβ uβ + e2 ( 2 + uµ )
δs 3 δs δs δs
As we can see, there are two terms depending on the charge: the first
describes the interaction between the charged particle and an external
elecromagnetic field; the second describes the selfinteraction of the charge
with its own electromagnetic field. As a consequence, the equation dis-
criminates between motions of differently charged particles in the pres-
ence of an external electromagnetic field through the first term, but,
due to the second one it may also possibly discriminate between them
intrinsically.
In case a charge dependent motion should arise when a non vanishing
external electromagnetic field is present, we would be unable to ascertain
if the differences of the two motions are due to the intrinsic character or
90 L. Fabbri

to the external action of the electromagnetic field. So we choose to work

in a space in which there is no any external electromagnetic field

Fµν ≡ 0 (3)

so that the Lorentz-Dirac equation reduce to

δuµ 2e2 δ 2 uµ δuα δuα

= ( 2 + uµ ) (4)
δs 3m δs δs δs
The is the form of the Lorentz-Dirac equation of motion which shall be
used hereafter.

4 The problem of free falling electric charge in a static

homogeneous gravitational field
Suppose one tried to tackle the problem of a free falling electric charge
in a static homogeneous gravitational field on the basis of the following
hypotheses:

a) Radiation emitted by a charge should be treated as a covariant

quantity, that is it should be described in tensorial terms

b) Principle of Equivalence is valid also for charged particles

c) In general (deliberatly vague), any accelerated charge emits

radiation

Analyse the case of free falling electric charge in a static homogeneous

gravitational field according to these hypotheses. Since, as we said, from
the point of view of a comoving observer, the charge does not radiate, the
tensor describing the emitted radiation should vanish in the free falling
system; since, on the other hand, from the point of view of an observer
at rest with respect to the matter generating the gravitational field ,
the charge should radiate, the tensor describing the emitted radiation
should not vanish in that system. The contradiction can be avoided
only dropping (at least) one of the three hypotheses, that is:

a) Consider the emitted radiation as a physical quantity depending

on the system of reference

b) Consider the Principle of Equivalence valid only for the mechanics

Free falling electric charge . . . 91

c) Consider that, in some situation, even an accelerated charge may

not radiate

In the past, many attempts at solving the problem have been made.
Some of them tried to solve the problem considering a theory in
which non tensorial quantities were used (see, for example, [1]). In these
attempts, in fact, one defined the emitted radiation as

2 2 duµ duν
I0 = e gµν (5)
3 ds ds
which is a tensor only underclosest linear transformations of coordinates
but not a general covariant tensor, so that it is possible to find a system
of reference in which it vanishes, even if it does not in another system of
reference.
This is exactly what happens, namely one has I0 6= 0 in the system
of reference at rest with respect to the matter, I0 = 0 in the system of
reference in free fall with the charge, and the two frames are connected
by a non linear coordinate transformation, under which I0 is not a tensor.
Even though in this way no contradiction arises, we do not find this
attempt quite convincing. And this for two reasons: on the one hand,
we deem unsatisfactory to deal with an emission of radiation in a not
generally covariant term, and this when trying to solve an apparent
contradiction arising in the framework of general relativity. For this
reason, we will consider in this work

2 2 δuα δuα
I= e (6)
3 δs δs
as the true expression for the emitted radiation. On the other hand, one
is faced by another apparent contradiction, namely that arising from the
circumstance that energy carried by electromagnetic radiation should be
revealed or not by an array of counters depending on its state of motion
with respect to the charge.
A different and more organic approach was considered first by
M.Born in 1909, then by D.L.Drukey and finally by M.Bondi and T.Gold
in the fifties and sixties; in this new approach the charge was not left
to itself, its free fall being checked by an external electromagnetic field.
The question whether the charge radiates was then given an affermative
answer ([4], [5] and [6]).
92 L. Fabbri

The consequent contradiction about the validity of the Principle of

Equivalence was removed in this manner: on one hand there is emission
of radiation but on the other hand the Riemann tensor does not really
vanish everywhere, but only locally. So, we can separate the space into
two domains: in an immediate neighborhood of the charge, the Riemann
tensor vanishes and one has no radiation; far from the charge, one has
radiation and the Riemann tensor exhibits a curvature of the space-time:
the Principle of Equivalence does not apply far from the charge, while
near to it the Principle of Equivalence does not lead to any contradiction.
We do not find fully satisfactory this approach either: the necessity
of considering an electromagnetic field appears to leave quite open the
question concerning a free falling charge, an object which, after all, seems
to be entitled to exist.
Our choice to consider a charge in actual free fall (Fµν = 0) in a static
µ
homogeneous gravitational field (Rαβρ = 0) is somewhat complementary
to the last described.

5 Solving the problem

As we have just said, we shall take into account the conditions

Fµν ≡0 (7)
µ
Rαβρ =0 (8)

and the situation of a free falling electric charge in a static homogeneous

gravitational field.
These assumptions leave us with a Lorentz-Dirac equation of motion
of the form
δuµ 2e2 δ 2 uµ δuα δuα
= ( 2 + uµ ) (9)
δs 3m δs δs δs
The last equation must be supplied with the two sets of initial conditions
concerning the four-position and the four-velocity; and, since one deals
with a third order differential equation, the two sets must be supple-
mented by a third set concerning the four-acceleration. The three sets
of initial conditions are

t(s = 0)=0
x(s = 0)=x0
y(s = 0)=0
z(s = 0)=0
Free falling electric charge . . . 93

and

ut (s = 0)=1
ux (s = 0)=0
uy (s = 0)=0
uz (s = 0)=0

and, finally

ω t (s = 0)=0
ω x (s = 0)=0
ω y (s = 0)=0
ω z (s = 0)=0

µ
Since, according with the condition Rαβρ = 0, the space is flat, there
exists an overall inertial system, in which one can write the Lorentz-
Dirac equation of motion in the form established by Special Relativity,
that is
2e2 dω µ
ωµ = ( + uµ ω α ωα ) (10)
3m ds
In a static homogeneous gravitational field the free motion is character-
ized by the condition
ω α ωα = −g 2 (11)
with g constant. Taking the first derivatives with respect to s one has

dω α
ωα = 0
ds
and taking the contraction between the Lorentz-Dirac equation and the
acceleration and simplifying one obtains

2e2 dω µ
−g 2 = ωµ = 0
3m ds
so that
ω α ωα = 0 (12)
This result implies
I=0 (13)
94 L. Fabbri

in the inertial system of reference, but, since it is a tensor, the result

holds in all the systems of reference. This means that no radiation is
emitted by the charge.
We next show that the motion is a geodesic motion. The motion of
the partcle is characterized by the condition
δuα δuα
=0 (14)
δs δs
and the Lorentz-Dirac equation reduces to
δuµ 2e2 δ 2 uµ
= (15)
δs 3m δs2
so that the equations for the single components have been separated.
Their solutions are
2e2 ω µ (0) 3m
uµ (s) = uµ (0) + [exp( 2 s) − 1]
3m 2e
which, taking into account the initial conditions, become

uµ (s) = uµ (0)

hence
ω µ (s) ≡ 0 (16)
in the inertial system of reference. In this system the connection is also
vanishing. Since, in general,
δuµ duµ
= + Γµαβ uα uβ
δs ds
one obtains
δuµ
≡0 (17)
δs
and, since this is a tensor, the previous relation holds in every system
of reference.
The solution of the Lorentz-Dirac equation in our situation is a space-
time geodesic. This result is the same one obtain considering a neutral
particle, so that there is no distinction between the two kinds of objects.

One further work on this problem is due to Rohrlich and Fulton

[7]; in their paper, they give a proof of the fact that the right side of
Free falling electric charge . . . 95

the Lorentz-Dirac equation vanishes, namely, writing the Lorentz-Dirac

equation of motion in a flat space as
δuµ
m = eg µα Fαβ uβ + Γµ (18)
δs
they show that
Γµ ≡ 0 (19)
But the term that describes the radiation is decomposable into two
terms; hence, even if the term describing the radiation vanishes, the
single terms in which it is decomposed may not. In our situation, we
have also required the absence of the external electromagnetic field. In
this case, if the term describing the radiation vanishes then the single
terms vanish too and the motion is once again a geodesic.

6 Conclusions
To conclude: we considered two observers: the observer at rest with
respect to the matter source of the gravitational field sees the charge
perform the hyperbolic motion, typical of relativistic free fall, hence not
radiating; the observer in free fall with the charge keeps seeing it at
rest, henceagain not radiating, as required by covariance and without
any contradiction.
If the Principle of Equivalence is not valid for all of electromag-
netism, then a proof for this fact cannot be obtained by this kind of
an experiment.

Acknowledgements
I would like to thank Prof. S.Bergia for discussions and foundamental
help in the drawing up of the final version of this paper.

References
[1] Rohrlich F., Ann. Phys., 22, 169, (1963)
[2] Dirac P.A.M., Proc. Roy. Soc., A167, 148, (1938)
[3] De Witt B.S. and Brehme R.W., Ann. Phys., 9, 220, (1960)
[4] Born M., Ann. der Phys., 30, 1, (1909)
[5] Drukey D.L., Phys. Rev., 76, 545, (1949)
[6] Bondi M. and Gold T., Proc. Roy. Soc., A229, 416, (1955)
[7] Rohrlich F. and Fulton T., Ann. Phys., 9, 499, (1960)

(Manuscrit reçu le 16 avril 2004)