Causality Electromagnetic

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

INSTITUTE OF PHYSICS PUBLISHING EUROPEAN JOURNAL OF PHYSICS

Eur. J. Phys. 25 (2004) 287296 PII: S0143-0807(04)69749-2


Presenting electromagnetic theory in
accordance with the principle of
causality
Oleg D Jemenko
Physics Department, West Virginia University, PO Box 6315, Morgantown,
WV 26506-6315, USA
Received 1 October 2003
Published 28 January 2004
Online at stacks.iop.org/EJP/25/287 (DOI: 10.1088/0143-0807/25/2/015)
Abstract
Amethod of presenting electromagnetic theory in accordance with the principle
of causality is described. Two causal equations expressing time-dependent
electric and magnetic elds in terms of their causative sources by means
of retarded integrals are used as the fundamental electromagnetic equations.
Maxwells equations are derived from these causal equations. Except for the
fact that Maxwells equations appear as derived equations, the presentation is
completely compatible with Maxwells electromagnetic theory. An important
consequence of this method of presentation is that it offers newinsights into the
cause-and-effect relations in electromagnetic phenomena and results in simpler
derivations of certain electromagnetic equations.
1. Introduction
One of the most important tasks of physics is to establish causal relations between physical
phenomena. No physical theory can be complete unless it provides a clear statement and
description of causal links involved in the phenomena encompassed by that theory. In
establishing and describing causal relations it is important not to confuse equations which
we call basic laws with causal equations. A basic law is an equation (or a system of
equations) from which we can derive most (hopefully all) possible correlations between the
various quantities involved in a particular group of phenomena subject to the basic law. A
causal equation, on the other hand, is an equation that unambiguously relates a quantity
representing an effect to one or more quantities representing the cause of this effect. Clearly,
a basic law need not constitute a causal relation, and an equation depicting a causal relation
may not necessarily be among the basic laws in the above sense.
Causal relations between phenomena are governed by the principle of causality.
According to this principle, all present phenomena are exclusively determined by past events.
Therefore equations depicting causal relations between physical phenomena must, in general,
be equations where a present-time quantity (the effect) relates to one or more quantities (causes)
that existed at some previous time. An exception to this rule are equations constituting causal
0143-0807/04/020287+10$30.00 2004 IOP Publishing Ltd Printed in the UK 287
288 O D Jemenko
relations by denition; for example, if force is dened as the cause of acceleration, then the
equation F = ma, where F is the force and a is the acceleration, is a causal equation by
denition.
In general, then, according to the principle of causality, an equation between two or
more quantities simultaneous in time but separated in space cannot represent a causal relation
between these quantities. In fact, even an equation between quantities simultaneous in time
and not separated in space cannot represent a causal relation between these quantities because,
according to this principle, the cause must precede its effect. Therefore the only kind of
equations representing causal relations between physical quantities, other than equations
representing cause and effect by denition, must be equations involving retarded (previous-
time) quantities.
Let us apply these considerations to the basic electromagnetic eld laws. Traditionally
these laws are represented by the four Maxwells equations, which, in their differential form,
are
D = , (1)
B = 0, (2)
E =
B
t
, (3)
and
H = J +
D
t
, (4)
where E is the electric eld vector, Dis the displacement vector, His the magnetic eld vector,
B is the magnetic ux density vector, J is the current density vector, and is the electric charge
density. For elds in a vacuum, Maxwells equations are supplemented by the two constitutive
equations,
D =
0
E (5)
and
B =
0
H, (6)
where
0
is the permittivity of space, and
0
is the permeability of space.
Since none of the four Maxwells equations is dened to be a causal relation, and since
each of these equations connects quantities simultaneous in time, none of these equations
represents a causal relation. That is, D is not a consequence of (and vice versa), E
is not a consequence of B/t (and vice versa), and H is not a consequence of J + D/t
(and vice versa). Thus, Maxwells equations, even though they are basic electromagnetic
equations (since most electromagnetic relations are derivable fromthem), do not depict cause-
and-effect relations between electromagnetic phenomena and leave the question of causality
in electromagnetic phenomena unanswered.
The purpose of this paper is to show how Maxwellian electromagnetic theory can be
reformulated and presented in a classroomin compliance with the principle of causality so that
the causal relations between fundamental electromagnetic phenomena are clearly revealed.
2. Causal equations for electric and magnetic elds
A reformulation and presentation of Maxwells electromagnetic theory in accordance with
the principle of causality must be based on causal electromagnetic equations that are at
least as general as Maxwells equations and are in complete accord with the latter. What
should be the form of such causal electromagnetic equations? Since an effect can be a
combined or cumulative result of several causes, it is plausible that in causal equations a
quantity representing an effect should be expressed in terms of integrals involving quantities
Presenting electromagnetic theory in accordance with the principle of causality 289
representing the various causes of that effect. And since, by the principle of causality, the
cause must precede its effect, the integrals in causal equations must be retarded, that is, the
integrands in these integrals must involve quantities as they existed at a time prior to the time
for which the quantity representing the effect is being computed.
The following equations for the electric elds E and the magnetic eld H in a vacuum
satisfy the above requirements for causal electromagnetic equations, and we shall use them as
basic laws for presenting Maxwells electromagnetic theory in compliance with the principle
of causality:
E =
1
4
0

[]
r
3
+
1
r
2
c
[]
t

r dV

1
4
0
c
2

1
r

J
t

dV

, (7)
and
H =
1
4

[J]
r
3
+
1
r
2
c

J
t

r dV

. (8)
The square brackets in these equations are the retardation symbol indicating that the quantities
between the brackets are to be evaluated for the retarded time t

= t r/c, where t is the


time for which E and H are evaluated, is the electric charge density, J is the current density,
r is the distance between the eld point x, y, z (point for which E and H are evaluated) and
the source point x

, y

, z

(volume element dV

), and c is the velocity of light. The integrals


are extended over all space.
According to equation (7), the electric eld has three causative sources: the retarded charge
density [], the retarded time derivative of the charge density []/t , and the retarded time
derivative of the current density [J/t ]. Likewise, according to equation (8), the magnetic
eld has two causative sources: the retarded current density [J] and the retarded time derivative
of the current density [J/t ].
As we shall presently see, in order to be equivalent to Maxwells electromagnetic theory,
the electromagnetic theory based on the causal equations (7) and (8) needs a third basic
equation: the familiar continuity equation representing the conservation of electric charge
J =

t
. (9)
Furthermore, equations (7)(9) need to be supplemented by the constitutive relations,
equations (5) and (6), unless only the elds E and H, without the elds D and B, are used.
In the mode of presentation of electromagnetic theory described here, the three laws,
equations (7)(9), are postulated, and their correctness is proved by demonstrating that
they are in complete agreement with Maxwellian electrodynamics, that is, by demonstrating
that Maxwells equations can be derived from them. Therefore it is not necessary to
discuss the original considerations that led to the formulation of equations (7)(9). It
may be noted, however, that equations (7) and (8) have been originally obtained from
inhomogeneous equations for electromagnetic waves [1] and can also be obtained from the
retarded electromagnetic potentials [2] as well as from the wave equation for the retarded
electromagnetic potentials [3].
3. Deriving Maxwells equations from the causal equations for E and H
To derive Maxwells equations from the causal equations (7) and (8), we rst transform these
equations to a somewhat different formwith the help of vector identities listed in the appendix.
Using vector identity (A.4), we replace the two terms in the integrand of the rst integral of
equation (7) by a single term, obtaining
E =
1
4
0


[]
r
dV


1
4
0
c
2

1
r

J
t

dV

. (10)
290 O D Jemenko
Transforming now the integrand in the rst integral of equation (10) by means of vector
identity (A.5), we obtain (note that the ordinary operator operates upon the eld-point
coordinates, whereas the primed operator

operates upon the source-point coordinates)


E =
1
4
0

]
r
dV

+
1
4
0

[]
r
dV

1
4
0
c
2

1
r

J
t

dV

. (11)
The second integral in the last expression can be transformed into a surface integral by means
of vector identity (A.2). But this surface integral vanishes, because is conned to a nite
region of space, while the surface of integration is at innity. We thus have
E =
1
4
0

]
r
dV

1
4
0
c
2

1
r

J
t

dV

. (12)
Similarly, applying vector identity (A.4) to equation (8), we obtain
H =
1
4


[J]
r
dV

. (13)
Transforming equation (13) by means of vector identity (A.5) and eliminating

{[J]/r} by
means of vector identity (A.3) (see the explanation belowequation (11); note that J is conned
to a nite region of space), we obtain for the magnetic eld
H =
1
4

J]
r
dV

. (14)
Maxwells equations can now be obtained from equations (12), (14) and (9) as follows:
The rst Maxwells equation. From equation (12) we have
E =
1
4
0

]
r
dV

1
4
0
c
2

1
r

J
t

dV

=
1
4
0

]
r
dV


1
4
0
c
2


1
r

J
t

dV

, (15)
(the operator can be placed under the integral sign because it operates on the eld-
point coordinates x, y, z, while the integration is over the source-point coordinates x

, y

, z

).
Applying vector identity (A.5) to equation (15) and eliminating

{[

]/r} by means of
vector identity (A.1) (see the explanation below equation (11)), we obtain
E =
1
4
0

]
r
dV


1
4
0
c
2

1
r

J)
t

dV

, (16)
and, using now equation (9) to replace

J by /t , we get
E =
1
4
0


1
c
2

t
2

r
dV

. (17)
According to vector identity (A.6), the right side of equation (17) is simply (1/
0
),
so that, replacing E in equation (17) by D with the help of equation (5), we obtain the rst
Maxwells equation
D = . (1)
The second Maxwells equation. From equation (14) we have
H =
1
4

J]
r
dV

=
1
4

J]
r
dV

(18)
(the operator can be placed under the integral sign because it operates on the eld-
point coordinates x, y, z, while the integration is over the source-point coordinates x

, y

, z

).
Applying vector identity (A.5) to equation (18) and eliminating

{[

J]/r} by means
Presenting electromagnetic theory in accordance with the principle of causality 291
of vector identity (A.1) (see the explanation below equation (11); note that J is conned to a
nite region of space), we obtain
H =
1
4

J)]
r
dV

, (19)
and since

0,
H = 0, (20)
which, by equation (6), yields the second Maxwells equation
B = 0. (2)
The third Maxwells equation. From equation (12) we have
E =
1
4
0

]
r
dV


1
4
0
c
2

1
r

J
t

dV

=
1
4
0

]
r
dV

1
4
0
c
2


1
r

J
t

dV

, (21)
where we have placed the operator under the integral sign (this can be done because
operates on unprimed coordinates, while the integration is over the primed coordinates).
Applying vector identity (A.5) to equation (21) and eliminating

{[

]/r} by means
of vector identity (A.3) (see the explanation below equation (11)), we obtain
E =
1
4
0

]
r
dV

1
4
0
c
2


1
r

J
t

dV

, (22)
and since

0,
E =
1
4
0
c
2


1
r

J
t

dV

. (23)
Let us now multiply equation (14) by
0
and differentiate it with respect to time. Since,
by equation (6), B =
0
H, we have
B
t
=
0
H
t
=

0
4

J]
r
dV

. (24)
Transforming the integrand in equation (24) by means of the vector identities (A.5) and
eliminating

{[

J]/r} by means of vector identity (A.3) (see the explanation below


equation (11); note that J is conned to a nite region of space), we obtain
B
t
=
0
H
t
=

0
4


[J]
r
dV

. (25)
Differentiating under the integral sign and taking into account that [J]/t = [J/t ] and
that
0
= 1/
0
c
2
, we obtain
B
t
=
1
4
0
c
2


1
r

J
t

dV

, (26)
which together with equation (22) yields the third Maxwells equation
E =
B
t
. (3)
The Fourth Maxwells Equation. From equation (14) we have
H =
1
4

J]
r
dV

=
1
4

J]
r
dV

, (27)
292 O D Jemenko
where we have placed the operator under the integral sign (this can be done because
operates on unprimed coordinates, while the integration is over the primed coordinates).
Applying vector identity (A.5) to equation (27) and eliminating

{[

J]/r} by means
of vector identity (A.3) (see the explanation below equation (11); note that J is conned to a
nite region of space), we obtain
H =
1
4

J)]
r
dV

. (28)
Let us now nd the time derivative of D by using equation (12). Since, by equation (5),
D =
0
E, we have
D
t
=
0
E
t
=
1
4


t
[

]
r
dV


1
4c
2


t
1
r

J
t

dV

=
1
4

r
dV


1
4c
2

1
r

2
J
t
2

dV

(29)
and, making use of the continuity equation, equation (3), we obtain
D
t
=
1
4

J)]
r
dV


1
4c
2

1
r

2
J
t
2

dV

. (30)
Next, let us subtract equation (30) from (28). Placing the derivative D/t on the right side
of the resulting equation and replacing the three integrals by a single integral, we obtain
H =
1
4

J)

J)
1
c
2

2
J
t
2

r
dV

+
D
t
. (31)
But, according to vector identity (A.7), the rst term on the right in equation (30) is simply
the current density J. Replacing this term by J, we obtain the fourth Maxwells equation
H = J +
D
t
. (4)
4. Discussion
Although the presentation of electromagnetic theory on the basis of the causal equations
for electric and magnetic elds, equations (7) and (8), is somewhat more complex than
the traditional presentation based directly on Maxwells equations, such a presentation, as
we shall presently see, simplies the derivation of some electromagnetic formulae, offers
important new insights into certain electromagnetic phenomena and dispels certain erroneous
views on electromagnetic cause-and-effect relations. In particular, as is explained below,
the presentation based on the causal electromagnetic equations shows that the traditional
explanation of the very important phenomenon of electromagnetic induction is incorrect and
reinforces the original explanation of this phenomenon provided by Faraday and Maxwell.
But let us rst demonstrate how the presentation of electromagnetic theory based
on the causal electromagnetic equations simplies the derivation of some representative
electromagnetic formulae.
To start with, let us note that in time-independent systems there is no retardation and the
time derivatives vanish. Therefore, for time-independent systems we immediately obtain from
equation (7) the Coulomb eld equation,
E =
1
4
0


r
3
r dV

, (32)
and from equation (8) we immediately obtain the BiotSavart law,
H =
1
4

J
r
3
r dV

. (33)
Presenting electromagnetic theory in accordance with the principle of causality 293
Next, let us quickly derive some formulae whose derivation requires considerable effort
in the conventional presentation of electromagnetic theory.
Factoring out the operator fromunder the rst integral of equation (10), we immediately
obtain the relation for the retarded electric scalar potential
E =
1
4
0
c
2

1
r

J
t

dV

, (34)
where is
=
1
4
0

[]
r
dV

, (35)
reducing to the ordinary scalar potential
=
1
4
0


r
dV

, (36)
for time-independent electric eld.
Likewise, factoring out the operator from under the integral of equation (13) and using
equation (6), we immediately obtain the equations for the retarded vector potential A
B = A, (37)
with
A =

0
4

[J]
r
dV

, (38)
which reduces to the ordinary vector potential
A =

0
4

J
r
dV

(39)
in the case of time-independent magnetic eld.
Next, taking into account that
0
= 1/
0
c
2
, and noting that the last termin equation (34) is
the partial time derivative of the retarded vector potential A given by equation (38), we obtain
the equation expressing time-dependent electric eld in terms of the retarded scalar electric
potential and the retarded magnetic vector potential:
E =
A
t
. (40)
To complete these representative derivations, we shall now quickly obtain the Lorenz
condition. Evaluating A by using equation (38), using vector identity (A.5), and then
eliminating

{[J]/r} by means of vector identity (A.1), we have


A =

0
4


[J]
r
dV

=

0
4

[

J]
r


[J]
r

dV

=

0
4

J]
r
dV

. (41)
Using the continuity equation, equation (9), we replace

J by /t , obtaining
A =

0
4

1
r

dV

0
4

[]
r
dV

. (42)
Replacing the integral in equation (42) with the help of (35), we obtain the Lorenz condition
(frequently erroneously referred to as the Lorentz condition)
A =
0

t
. (43)
And now let us discuss the very important consequence of the causal equation,
equations (7) and (8), pertaining to the phenomenon of electromagnetic induction. There is a
widespread belief that time-variable electric and magnetic elds can cause, induce, each other.
294 O D Jemenko
It is traditionally asserted that, according to Maxwells equation (3), a changing magnetic eld
produces an electric eld (Faraday induction) and that, according to Maxwells equation (4),
a changing electric eld produces a magnetic eld (Maxwell induction). The very useful
and successful method of calculating induced voltage (emf) in terms of changing magnetic
ux appears to support the reality of Faraday induction. And the existence of electromagnetic
waves appears to support the reality of both Faraday induction and Maxwell induction. Note,
however, that as explained in section 1, Maxwells equation (3), which is usually considered
as depicting Faraday induction, does not represent a cause-and-effect relation because in this
equation the electric and the magnetic eld is evaluated for the same moment of time. Note also
that in electromagnetic waves electric and magnetic elds are in phase, that is, simultaneous
in time, and hence, according to the principle of causality (which states that the cause always
precedes its effect), the two elds cannot cause each other (by the principle of causality, the
elds should be out of phase if they create each other).
Maxwells equations by themselves do not provide an answer to whether or not the
Faraday induction or Maxwell induction are real physical phenomena. In Maxwells
equations electric and magnetic elds are linked together in an intricate manner, and neither
eld is explicitly represented in terms of its sources. It is true, of course, that whenever
there exists a time-variable electric eld, there also exists a time-variable magnetic eld. This
follows fromour equations (7) and (8) as well as fromMaxwells equations (3) and (4). But, as
already mentioned, according to the causality principle, Maxwells equations do not reveal a
causal link between electric and magnetic elds. On the other hand, equations (7) and (8) show
that in time-variable systems electric and magnetic elds are always created simultaneously,
because these elds have a common causative source: the changing electric current [J/t ]
(the last term of equation (7) and the last term in the integral of equation (8)).
It is important to note that neither Faraday (who discovered the phenomenon of
electromagnetic induction) nor Maxwell (who gave it a mathematical formulation) explained
this phenomenon as the generation of an electric eld by a magnetic eld (or vice versa).
After discovering the electromagnetic induction, Faraday wrote in a letter of November
29, 1831, addressed to his friend Richard Phillips [4]:
When an electric current is passed through one of two parallel wires it causes at rst a
current in the same direction through the other, but this induced current does not last a moment
notwithstanding the inducing current (from the Voltaic battery) is continued. . ., but when the
current is stopped then a return current occurs in the wire under induction of about the same
intensity and momentary duration but in the opposite direction to that rst found. Electricity in
currents therefore exerts an inductive action like ordinary electricity (electrostatics, ODJ) but
subject to peculiar laws: the effects are a current in the same direction when the induction is
established, a reverse current when the induction ceases and a peculiar state in the interim. . ..
Quite clearly, Faraday speaks of an inducingcurrent, and not at all of an inducing magnetic
eld. (In the same letter Faraday referred to the induction by magnets as a very powerful proof
of the existence of Amperian currents responsible for magnetization.)
Similarly, Maxwell wrote in his Treatise [5]:
It is only since the denitions of electromotive force. . . and its measurement have been
made more precise, that we can enunciate completely the true lawof magneto-electric induction
in the following terms: the total electromotive force acting round a circuit at any instant is
measured by the rate of decrease of the number of lines of magnetic force which pass through
it. . .. Instead of speaking of the number of lines of magnetic force, we may speak of the
magnetic induction through the circuit, or the surface integral of magnetic induction extended
over any surface bounded by the circuit.
As we see, Maxwell, too, considered the electromagnetic induction as a phenomenon in
which a current (or electromotive force) is induced in a circuit, but not as a phenomenon in
which a changing magnetic eld causes an electric eld. He clearly says that the induced
electromotive force is measured by, not caused by, the changing magnetic eld. Just like
Faraday, he made no allusion to any causal link between magnetic and electric elds.
Presenting electromagnetic theory in accordance with the principle of causality 295
And there is one more fact that supports the conclusion that what we call electromagnetic
induction is not the creation of one of the two elds by the other. In the covariant formulation
of electrodynamics, electric and magnetic elds appear as components of one single entity
the electromagnetic eld tensor. Quite clearly, a component of a tensor cannot be a cause of
another component of the same tensor, just like a component of a vector cannot be a cause of
another component of the same vector.
We must conclude therefore that the true explanation of the phenomenon of
electromagnetic induction is provided by the causal electromagnetic equations, equations (7)
and (8). According to these equations, in time-variable systems electric and magnetic elds are
always created simultaneously, because they have a common causative source: the changing
electric current [J/t ]. Once created, the two elds coexist from then on without any effect
upon each other. Hence electromagnetic induction as a phenomenon in which one of the elds
creates the other is an illusion. The illusion of the mutual creation arises from the facts
that in time-dependent systems the two elds always appear prominently together, while their
causative sources (the time-variable current in particular) remain in the background
1
.
Thus, even though a presentation of electromagnetic theory on the basis of the causal
electromagnetic equations is somewhat more complicated than the traditional presentation on
the basis of Maxwell equations, such a presentation is well justied by the new possibilities
that it offers and by the important new results revealed by it.
Appendix
Vector identities
In the vector identities listed below, U is a scalar point function; V is a vector point function;
X is a scalar or vector point function of primed coordinates (source-point coordinates) and
incorporates an appropriate multiplication sign (dot or cross for vectors); the operator
operates upon unprimed coordinates (eld-point coordinates); the operator

operates upon
primed coordinates (source-point coordinates).
Identities for the calculation of surface and volume integrals

AdV

A dS

(Gausss theorem) (A.1)

U dV

U dS

(A.2)

AdV

A dS

. (A.3)
Identities for operations with retarded quantities
r[X]
r
3
+
r
r
2
c

X
t

=
[X]
r
(A.4)

[X]
r
=
[

X]
r

[X]
r
(A.5)
U =
1
4

U
1
c
2

2
U
t
2

r
dV

(A.6)
A =
1
4

A)

A)
1
c
2

2
A
t
2

r
dV

. (A.7)
1
The author has been unable to determine by whom, where and why it was rst suggested that changing electric and
magnetic elds create each other. One thing appears certain howeverthe idea did not originate with either Faraday
or Maxwell.
296 O D Jemenko
References
[1] Jemenko OD 1989 Electricity and Magnetism 2nd edn (Star City: Electret Scientic) pp 5146 (the same pages
in the rst edition of the book (Appleton-Century-Crofts, New York, 1966))
[2] Rosser W G V 1997 Interpretation of Classical Electromagnetism (Dordrecht: Kluwer) pp 824
[3] Jackson J D 1999 Classical Electrodynamics 3rd edn (New York: Wiley) pp 2437
[4] Thompson S P 1898 Michael Faraday, His Life and Work (New York: Macmillan) p 115, 116
[5] Maxwell J C 1891 A Treatise on Electricity and Magnetism 3rd edn, vol 2 (New York: Dover) pp 1767 (reprint)

You might also like