Professional Documents
Culture Documents
Analysis of The Lorenz Gauge
Analysis of The Lorenz Gauge
(1.1)
*
Department of Physics, Voronezh State University, Universitetskaya Sq.1, Voronezh, 394693,
Russia; E-mail: [email protected]
APEIRON Vol. 7 Nr. 1-2, January-April, 2000 Page 39
Initial conditions are zero.
. 0
t
U
; 0 U
0 t
0 t
=
=
=
=
First solution. Function U is equal to zero. U
1
=0. (1.2)
Second solution. We may write very much other solution of the task. For example, the
second solution is
2 2 2
2
( ) ( )
2
( )
0 ( )
( , ) [ ]
2 2
1
[ ( , ) ]
2
x ct
x ct x ct
x ct
x c t t
x t
x c t
A A
U x t e e e d
c
f d d Ae e
c
+
+
+
= + +
+
(1.3)
where
2
2
2
2
( , ) [2 (1 2 ) ] f Ae e
c
= + . , , A are constant ( >0, >0).
The second solution U
2
(x,t) is not equal to zero, and it has no singularities.
Thus, we have two different solutions of wave equation under the given initial condi-
tions. The obtained outcome is not exclusive or unique. Always it is possible to find simi-
lar examples, in which there is a second solution, for the inhomogeneous equations and
tasks with boundary conditions. In other words, we can generalise this example. The fur-
ther research of the problem is a task for mathematicians.
Choice of Second Solution
Now we shall show one of many ways for searching a second solution. Earlier we investi-
gated the problem in (Kuligin et al., 1989), (Kuligin et al., 1990). To simplify the analysis
we write a wave equation without boundary conditions for free space. This condition not
is limitation. We can use this way for the tasks with boundary conditions.
) t ; ( f
t
U
v
1
U
2
2
2
r =
(1.6)
where: U is some function; v is propagation speed of it; f is source of function U;
=
a > r when 0
; a r when t) ; F(
) t ; ( f
r
r
(1.7)
A solution of equation (1.6) should satisfy to the following initial condition.
) (
t
U
); ( ) 0 ; ( U
0 t
r r r =
=
=
(1.8)
It is a standard initial value problem for a hyperbolic equation without boundary condi-
tions.
When the solution exists (in general case it exist (Tikhonov, Samarsky, 1953)) then we
write it as ) t ; ( U r . The instantaneous functions in the solution are absent.
Now we shall begin to seek a solution of equation (1.6) as a sum of two functions.
) t ; ( V ) t ; ( u U r r + = (1.9)
After substitution of expression (1.9) in equation (1.6) we find
Page 40 APEIRON Vol. 7 Nr. 1-2, January-April, 2000
) t ; ( f
t
V
v
1
V
t
u
v
1
u
2
2
2 2
2
2
r =
(1.10)
As we have entered two new unknown functions we should add an appropriate condition.
There are many variants to give a side condition. For instance, we may use following
equations.
) t ; ( f
t
V
V ); t ; ( f V V ); t ; ( f V
3
3 2 2 1
r r r =
+ = + =
(1.11)
where and are some factors.
We assume that function u satisfies to the Poisson equation. We should specially no-
tice that the instantaneous functions do not break the principle of causality in physics
(Kuligin, 1987).
0 V lim ); t ; ( f V
1
r
1
= =
r
1.12)
The solution of Poisson equation (1.12) exists also (Tikhonov, Samarsky, 1953). Naturally,
function u
1
should satisfy to the new equation:
2
1
2
2 2
1
2
2
1
2
1
2
2
1
t
V
v
1
) t ; ( f
t
V
v
1
V
t
u
v
1
u
= +
+ =
r
(1.13)
Function u
1
has the new initial conditions.
.
t
V
) (
t
u
); 0 ; ( V ) ( ) 0 ; ( u
0 t
1
0 t
1
1 1
= =
= r r r r
(1.14)
Other conditions are conserved.
We shall consider that in general case a solution of equation (1.12) exists, and it is
known to us. In this case the right part of equation (1.13) and initial conditions (1.14) are
determined and are known. In general case a solution of equation (1.14) exists (Tikhonov,
Samarsky, 1953). We write it as ) t ; ( V
1
r .
Thus, we receive the new solution of equation (1.6), which differs from solution U(r;t).
It has the following form.
) t ; ( V ) t ; ( u U
1 1 1
r r + = (1.15)
Expression (1.15) is solution of equation (1.6) under the initial conditions. It can be con-
firmed by direct check.
In the same way we can find other solutions of equation (1.11):
3 3 3 2 2 2
V u U ; V u U + = + = .
Moreover, any solution, which is composed with U, U
1
, U
2
, U
3
, is also one of same
task.
1 a a a a ; U a U a U a U a U
3 2 1 0 3 3 2 2 1 1 0 new
= + + + + + + = (1.16)
where a
i
are free constant (i=0,1,2,3).
The method of choice of new solutions can also be used, for example, for Maxwell
equations and Shroedingers equation.
APEIRON Vol. 7 Nr. 1-2, January-April, 2000 Page 41
Mathematical Gauges
The procedure of a choice of the second solution has immediate connection with a choice
of kinds of functions having different space and time properties. It contains also the
choice of some equations to have these properties. We shall name this procedure as
mathematical gauge of primary equations. We mark the following singularities of mathe-
matical gauge procedure.
1. The solution of any fixed gauge is unique.
2. In general case the solutions of different gauges differ from each other.
The gauge invariance of mathematical gauges has no a place.
2. Mathematical Gauges in Electrodynamics
Maxwell equations
The problem of mathematical gauge is immediately connected to a problem of a physical
covariance of equations. The violation of uniqueness of a solution reduces to a series of
problems. We shall not consider the problem of physical covariance in this paper. We
shall consider, that the inertial frame of observer, in which the equations are fair, exists,
and we shall compare solutions of different gauges only in this frame of observer.
We write Maxwell equations.
= =
+
= =
E
H
E
H j
E
H
div ; 0
t
rot
0 div ;
t
rot
(2.1)
In set of equations (2.1) two vectors E and H are connected among themselves. We
can note independent equations for E and H.
= =
+
E
j E
E
H j
H
H
div ;
t
t c
1
rotrot
0 div ; rot
t c
1
rotrot
2
2
2
2
2
2
(2.2)
The equations (2.2) have wave solutions E
w1
and H
w1
. The index w means that they
are direct solutions of equations (2.2) without introduction of additional mathematical
gauges.
Now we shall search for a solution of equation (2.1) as the following.
ins w
E E E + = (2.3)
where: E
w
is solution of a wave equation; E
ins
is solution of Poisson equation.
The index ins means that we deal with a solution of Poisson equation.
We add the following equations to destroy indeterminacy.
0 rot ; 0 div
ins w
= = E E (2.4)
Now the complete set of t he equations is defined and one is
Page 42 APEIRON Vol. 7 Nr. 1-2, January-April, 2000
div ; 0 rot
0 div ;
t c
1
-
t
t c
1
rotrot
0 div ; rot
t c
1
rotrot
ins ins
w
2
ins
2
2 2
w
2
2
w
w
2
w
2
2
w
= =
=
+
= =
+
E E
E
E j E
E
H j
H
H
(2.5)
The equations (2.5) have electric field strength E
2
= E
ins
+ E
w2
and magnetic intensity
H
w2
. A solution of equation (2.2) differs from a solution of equation (2.5) even by that the
first solution does not contain instantaneous electric field strength E
ins
.
Here there is a parallel between equations (2.2), (2.5) and gauges of Maxwell equations.
If we replace in equation (2.5) electric field strength E
ins
by -grad, E
w
by
t
A
and mag-
netic intensity H
w
by rotA then we obtain Coulombs gauge of Maxwell equations.
0 = div ;
; grad
t
c
1
t c
1
2 2
2
2
A
j
A
A
+ =
(2.6)
If we replace E
w
by
t
A
and H
w
by rotA in equation (2.2) then we receive the Lorenz
gauge of Maxwell equations.
t c
1
2
2
2
j
A
A =
;
t c
1
2
2
2
;
0 =
t
c
1
div
2
+ A
(2.7)
Thus, physical gauge of Maxwell equations, which is determined by a condition for divA,
as a matter of fact coincides with mathematical gauge. It is a particular case of mathemati-
cal gauge. Therefore there is no sense to distinguish these gauges further.
Now we consider properties of solutions of equations (2.2) and (2.5). The second solu-
tion (2.5) contains instantaneous field strength E
ins
, which is absent in solution (2.2). In the
books (for example (Ginzburg, 1987)) they state that E
ins
is compensated by other field
strength E
w
. Due to this fact Coulombs gauge is equivalent to the Lorenz gauge. We
consider this conclusion as fallacy. It is shown in Appendix 1 that two gauges are not
equivalent.
We can tell that the casual concurrence cannot be generalised without the proofs, and
in the general case the uniqueness of a solution of Maxwell equations has no a place.
Solutions of the Lorenz gauge
Now we consider two solutions of the Lorenz gauge (2.7). The first solution is wave solu-
tion of equations (2.7) :
. rot ;
t
grad
1w w 1
A H
A
E =
=
(2.8)
The second solution may be obtained with the standard way. Let potentials A and be
a sum of other potentials.
w ins w ins
; + = + = A A A (2.9)
APEIRON Vol. 7 Nr. 1-2, January-April, 2000 Page 43
We give special condition for potentials
ins ins
and A . The potentials should be solu-
tions of Poisson equation.
0 grad
t
c
1
div ; ;
ins
2
ins ins ins
=
= = A j A
(2.10)
Hence, potentials
w w
and A are solutions of wave equations.
t c
1
t c
1
2
ins
2
2 2
w
2
2
w
A A
A
;
t c
1
t c
1
2
ins
2
2 2
w
2
2
w
;
0 =
t
c
1
div
w
2
w
+ A
(2.11)
The second solution is the following.
. rot rot
;
t
grad
t
ins w 2ins 2w 2
ins
ins
w
ins 2 w 2 2
A A H H H
A A
E E E
+ = + =
= + =
(2.12)
We see that the first solution (2.8) differs from the second solution (2.12). Thus, the
Lorenz gauge has many solutions too.
Limiting transfer
We should notice that the transfer to the limit c takes place
= =
c
w 1 ins 2
c
w 1 ins 2
lim ; lim H H E E
. (2.13)
Expression (2.13) is a casual concurrence, as the solution of equation (2.8) is not a so-
lution of equation (2.12). We should use the transfer to limit c very cautiously. Ot h-
erwise, we shall make illegal transfer from one gauge in another. The solutions of wave
equation (2.8) do not contain instantaneous fields. We can not force an instantaneous
field to propagate with light velocity.
Now we shall draw some brief conclusions.
1. The different gauges of Maxwell equations have different solutions. In gen-
eral case the gauge invariance is absent.
2. The electromagnetic potentials play the important role in classical electrody-
namics. Due to this circumstance we should have special gauge, which else
is necessary to find.
3. The transfer to limit c is not always lawful in an electrodynamics.
3. Law of Energy Conservation for Lorenzs Gauge
Before to investigate singularity law of energy conservation of the Lorenz gauge, we do
some explanations.
At first, the interaction in the classical theories is a basis. The interaction has two as-
pects. They are the force and power aspects. The electric field strength is the reflecting of
force aspect, and the electric potential is the reflecting of energy aspect. They speak that
Page 44 APEIRON Vol. 7 Nr. 1-2, January-April, 2000
the potential has no physical sense due to it is defined to a free constant. It is incorrect.
The potential energy also is defined with the same exactitude.
In second, in general case the solution of a wave equation can contain fields of differ-
ent kinds. Each field has the own equation. Due to this circumstance each kind of a field
has its law of energy conservation, energy density and energy flux density. Moreover,
each law has its special interpretation (Kuligin, Kuligina, 1986), (Kuligin et al., 1996).
Poyntings theorem is not universal, and one may not be used for all task of electrody-
namics.
The Proof
Let us consider equations (2.7). If we multiple the first equation by ( ) ( ) 1 t A and
the second equation by
t
then after simple transformations we receive
0
t
] )
t c
( ) div ( ) rot [(
t 2
1
] div
t
- rot
t
[ div
1
2 2 2
=
+ +
j
A A
A A A
A
A
A
(3.1)
0
t
] )
t c
( ) grad [(
t 2
+ ] grad
t
div[-
2 2
=
(3.2)
Equations (3.1) and (3.2) have the known form of a law of energy conservation:
0 p w
t
div = +
+ S
(3.3)
Expressions (3.1) and (3.2) are laws of energy conservation within the framework of the
Lorenz gauge.
Three energy flux densities
Now we must analyse the expressions and must show that three different fluxes exist in
the framework of the Lorenz gauge. We give an explanation mathematically.
Vector potential A and current density j can be submitted as a sum of two independent
components. They are solenoidal and irrotatational components of potential and current
density.
A = A
1
+ A
2
; j = j
1
+
j
2
(3.4)
where: div A
1
= 0; div j
1
= 0; rot A
2
= 0; rot j
2
= 0.
This is not a new gauge.
The convenience of similar division of potentials and current densities is dictated also
by following circumstance. Now we consider a point source that creates fields of vector
potentials A
1
and A
2
. The components of electric and magnetic fields created by potential
A
1
are always orthogonal to an electric field created by potential A
2
. By other words, the
following relationships take place:
0 )
t
(rot ; 0 )
t t
(
2
1
2 1
=
A
A
A A
(3.5)
Expressions (3.5) remain fair for any current densities and charge densities provided that
the fields and potentials are considered for very large distances from the source. We will
use these relationships further.
With expression (3.4) we can write three wave equations in the Lorenz gauge (2.7).
APEIRON Vol. 7 Nr. 1-2, January-April, 2000 Page 45
1
2
1
2
2
1
t c
1
j
A
A =
(3.6)
t c
1
2
2
2
2
2
2
j
A
A =
(3.7)
2
2
2
t c
1
(3.8)
0
t
c
1
div
2
2
=
+ A
(3.9)
Each wave equation describes own energy density, each wave equation has own vector
of energy flux density and law of energy conservation (Kuligin et al., 1990), (Kuligin et al.,
1997). These laws have an identical general form:
3) 2; 1; = (k 0 p
t
w
div
k
k
k
= +
+ S
(3.10)
where: S
k
is vector of flux density; w
k
is energy density; p
k
is power density of applied
electromotive forces; k = 1 corresponds to the law of conservation of vector potential A
1
;
k = 2 corresponds to potential A
2
; k = 3 corresponds to scalar potential .
Expressions of S
k
, p
k
and w
k
are given in Table 1. We see that within the framework of
the Lorenz gauge there are three various energy fluxes, which are described by three vari-
ous vectors of flux density. The first vector S
1
is classical vector that describes a trans-
versal electromagnetic wave. The second vector is vector S
2
that describes a longitudinal
wave of vector potential. The third vector S
3
is vector of flux density of longitudinal wave
of scalar potential .
Table 1.
Transversal Waves of Vector Potential
1
1
rot
t
1 A
S
=
] )
ct
( ) rot [(
2
1
w
2 1 2
1 1
=
A
A
t
p
1
1 1
=
A
j
Longitudinal Waves of Vector Potential
2
1
div
t
1
A
A
S
=
] )
ct
( ) div [(
2
1
w
2 2 2
2 2
=
A
A
t
p
2
2 2
=
A
j
Longitudinal Waves of Scalar Potential
= grad
t
3
S
] )
ct
( ) grad [(
2
w
2 2
3
=
t
p
3
=
Due to absence of uniqueness of the solution we cannot use expressions of energy
density and energy flux density, which were given by Poynting, within the framework of
the Lorenz gauge.
Page 46 APEIRON Vol. 7 Nr. 1-2, January-April, 2000
4. Energy Cancellation of Longitudinal Waves
From Maxwell equations follow that a longitudinal electric field must decreases not slower
than r
2
when r (Appendix 2). It is known, that any wave, which propagates from a
local source to infinity, in general case decreases as r
1
for r .
Thus, we deal with the contradiction. On the one hand, according to expression (B.4)
the longitudinal electric field decreases not slower than r
2
. On the other hand, the wave
equation predicts exis tence of longitudinal electrical fields of scalar and vector potentials,
which decrease proportionally r
1
. This contradiction should be investigated.
For the erasure of the contradiction we may assume that vectors S
2
and S
3
should de-
stroy each other even with r . We write two integrals for search of a necessary and
sufficient condition of longitudinal wave cancellation.
= d d sin r r S
(4.1)
= d d sin r w r W
2
(4.2)
where: is common flow of electromagnetic waves through the spherical surface; W is
energy inside thin layer r as shown in Figure 1;
= grad
t
] div
t
rot
t
[ A
A
A
A
S
(4.3)
] )
ct
( ) drad [(
2
+ ] )
ct
( ) div ( ) rot [(
2
w
2 2 2 2 2
+ +
=
A
A A
(4.4)
and are factors, them modules are equal to 1, and signs will be established later.
Now we present vector potential A as sum (3.4), and we suppose that all sources of
potentials locate inside some sphere with radius a (r >> a). Lets consider potentials on the
spherical surface r .We may use orthogonality of fields (3.5) when r .
Integrals (4.1) and (4.2) can be easily divided into two independent groups.
Figure 1.
APEIRON Vol. 7 Nr. 1-2, January-April, 2000 Page 47
d d sin ]r )
ct
( + ) [(rot
2
r lim W
d d sin r ] rot
t
[ lim
2 2 1 2
1
r
T
1
1
r
T
A
A
r A
A
(4.5)
0 d d sin ]}r )
ct
( + ) [(grad
2
+ ] )
ct
( + ) [(div
2
{ r lim W
0 d d sin r ] grad
t
div
t
[ lim
2 2 2 2 2 2
2
r
L
2
2
r
L
=
=
=
A
A
r A
A
(4.6)
where:
T
and
L
are flows of transversal and longitudinal waves; W
T
and W
L
are
energy values inside thin layer r for transversal and longit udinal waves accordingly.
The transversal waves carry away energy at infinity. Therefore integrals (4.5) are not
equal to zero.
As the longitudinal waves cannot transfer energy at infinity as two integrals (4.6) must
be equal to zero. The first equation (4.6) establishes that the common flow of longitudinal
waves does not carry away energy at infinity. The second equation (4.6) shows that field
energy inside the thin spherical layer quickly decreases because of the previous condition
Two equations (4.6) are satisfied simultaneously if expressions inside integrals are
equal to zero.
0 ] grad
t
div
t
[ r lim
2
2
r
=
A
A
r
(4.7)
0 = ]} )
ct
( ) grad [(
2
] )
t c
( ) div [(
2
{ r lim
2 2 2 2 2
2
2
r
+
A
A
(4.8)
In general case divA
2
, -grad,
t
2
A
and
t c
may decrease as r
-1
. The sums of
square terms inside braces of equation (4.8) are always positive. Hence, the expression
(4.8) may be equal to zero only provided that = -= 1 or = = +1.
Let = 1 and = +1. Taking into account this condition and equation (3.8) we trans-
form expressions (4.7) and (4.8).
] grad
t
][ grad
t
[
2
r lim ; 0 ] grad
t
[
t
r lim
2 2 2
r
2
r
+
A A A
r
(4.9)
From expression (4.9) follows that both equations are equal to zero simultaneously if and
only if
0 ] grad
t
[ lim lim
2
r
L
r
=
=
A
r E r
(4.10)
where E
L
is longitudinal component of an electric field strength.
Expression (4.10) is always right due to expression (B.4). In other words, the set of si-
multaneous equations (4.6) will be identically equal to zero if only the summarised longi-
tudinal electric field strength decreases faster than r
1
. It is possible only under the condi-
tion = .
So, we can formulate the final conclusion. The longitudinal waves of scalar and vector
potentials can compensate each other under c only provided that energy flux density
Page 48 APEIRON Vol. 7 Nr. 1-2, January-April, 2000
and energy density of scalar potential field should have an opposite sign with respect to
the appropriate densities of vector potential.
The return thesis also is fair. If the longitudinal waves of scalar and vector potentials,
which are described by the wave equations, compensate each other with r , then
potential energy and energy flow of scalar potential and, accordingly, potential energy
and energy flow of vector potential should have opposite signs.
5. Crash of Lorenzs Gauge
As is evident from the equations (3.7)(3.8) there are always longitudinal waves of scalar
and vector potentials within the framework of the Lorenz gauge. However, these waves
are compensated in the solutions of Maxwell equations, i.e. mutually cancellated each
other with r . Hence, the Lorenz gauge either deals with negative energy flux density
and negative energy density of charge fields of scalar potential or deals with negative flux
density of transversal electromagnetic waves because of the necessary and sufficient
condition of cancellation of longitudinal waves.
Firstly, = 1and = +1. We deal with violation of Coulombs law. If energy of scalar
potential is negative then the same charges should be attracted, and the unlike charges
should be repelled! But it contradicts reality!
Secondly, = + 1and = 1. Here a radiation of negative energy by aerials also is ab-
surd.
So, we see that the Lorenz gauge is in conflict with the physical phenomena. The
gauge cannot and should not be used for the description of electrodynamical phenomena.
This is a crash of the Lorenz gauge. Now it is essential to find a new transformation of the
equations. It is possible that light velocity is not the limit of all velocities. We consider
that not only the wave fields exist. Instantaneous fields also can exist in a nature.
We could add following argument. We have shown that the limiting transfer c is
illegal within the framework of the Lorenz gauge. Therefore, no the the Lorenz gauge may
explain quasistatical phenomena of electrodynamics.
Special Relativity theory, which use mathematical construction of Lorenzs theory and
the Lorenz gauge, now may be considered as questionable theory. Our conclusions are
not original. Similar conclusions are found in other works (Wallace, 1969), (Hayden, 1993),
where the authors compare experimental results to results following from the Special Rela-
tivity theory, in our research (Kuligin et al., 1990), (Kuligin et al., 1989), (Kuligin et al.,
1996), (Kuligin et al., 1994) and many papers of other authors.
Conclusions
We have proved that the wave equation has no a unique solution. Is shown that the
gauge procedure is a choice of different kinds of fields, which will give the contribution to
a solution of a wave equation. Generally, gauge invariance is not existing.
We have proved laws of energy conservation for wave fields in framework of the Lo-
renz gauge, which are expressed with wave potentials.
The problem of cancellation of longitudinal waves is investigated. Is shown with the
laws that the Lorenz gauge can not correctly to describe electrodynamics effects.
We consider that the experimental inspection of Maxwell equations, which will define
the exact gauge and equations for electromagnetic potentials, is necessary.
APEIRON Vol. 7 Nr. 1-2, January-April, 2000 Page 49
Acknowledgement
We sincerely thank editor R. Keys and Prof. C. Whitney for support of new ideas. Also we
are grateful to the referees and personally to Prof. A. Chubykalo for helpful remarks.
Appendix1. Compensation of Instantaneous Potential by Retarded
Potential
Till present day it was considered that the initial problem of wave equation has the unique solution.
By virtue of it the natural conclusion about independence of the solution from gauge selection.
Differently, the solutions within the framework of the Lorenz gauge and within frameworks of
Coulombs gauge were considered equivalent.
Now we should show that such conclusion is not right. Lets consider Coulombs gauge of
Maxwell equations (2.6).
0 = div ;
; grad
t
c
1
t c
1
w ins
ins
2 2
w
2
2
w
A
j
A
A
+ =
(?.1)
The solution of these equations should contain a field of instantaneous potential
ins
, i.e. -grad
ins
.
They assert that the field of a vector potential
t
w
A
compensates this instantaneous compo-
nent of field strength grad
ins
in all free space. To show an inaccuracy of this conclusion we shall
transform the right member of vector equation using an equation of continuity of scalar potential
ins
0
t
div
ins
ins
=
+ v
(A.2)
For simplicity we shall consider the charge velocity v is constant.
In outcome the right part of the vector equation of expression (A.1) is:
ins
2
ins ins
ins
2
ins
2
rotrot
c
1
graddiv
t
grad
c
1
- grad
t
c
1
=
=
+
v v v
v j
(A.3)
A compensation of field - grad
ins
by field
t
w
A
is impossible for following reasons. At first,
the right part of the equation of the vector potential is proportional v/c. Therefore, when v/c equal
or close to zero we can not have full compensation in free space. In second, the source of the vector
potential field (the right part of equation (A.3)) is solenoidal. It can create only solenoidal vector
potential A. The attempt to construct in free space a field of polar vector, using only sources of
solenoidal fields, is senseless.
Sometimes they speak, that compensation of instantaneous field by field of vector potential fol-
lows directly from gauge invariance linking Coulombs gauge to the Lorenz gauge. Lets consider
also this approach.
It is known that electric field E and the magnetic field H are saved with a following transforma-
tion.
t
f
= ' ; f grad '
+ = A A
(A.4)
Page 50 APEIRON Vol. 7 Nr. 1-2, January-April, 2000
where: A and old electromagnetic potentials; A and new electromagnetic potentials;
t
f
=
is
some function (gauge potential) satisfying to a homogenous wave equation:
0
t
f
c
1
f
2
2
2
=
(A.5)
Obviously from equation (A.5), the gauge potential f is not instantaneous.
Lets now record the Lorenz gauge of Maxwell equations.
t c
1
2
w
2
2
w
j
A
A =
;
t c
1
2
w
2
2
w
;
0 =
t
c
1
div
w
2
w
+ A
(A.6)
To receive from these equations of Coulombs gauge of Maxwell equation they enter the condition
t
f
w
=
. Using this condition and expression (A.4) we shall substitute A and , which are ex-
pressed with A and , into expression (A.6).
In outcome we find:
0 = ' div ; '
; ' grad
t
c
1
t
'
c
1
'
w w
ins
2 2
w
2
2
w
A
j
A
A
+ =
(?.7)
It would seem that now we really deal with Coulombs gauge. Actually we should remember that
potential f and, therefore, potentials
w
and
w
should be the solutions of homogeneous wave equa-
tion (see (A.5)).
0 )
t
'
c
1
' (
t
)
t c
1
(
t
t
f
c
1
f
2
w
2
2
w
2
w
2
2
w
2
w
2
2
w
=
(A.8)
It contradicts with Poisson equation of scalar potential (A.7)
=
w
'
(A.9)
The equations (A.8) and (A.9) are incompatible. Therefore, Coulombs gauge is not a corollary of
the Lorenz gauge because of gauge invariance. This conclusion has basic value for the quantum
electrodynamics.
Appendix 2. Longitudinal Electric Field at Infinity
Lets consider pattern of electric charges, which are taking place inside sphere limited in some radius
a. To estimate a rate of electric field change for r > > a we write integral
=
i
2 o
q d d sin r n E
(B.1)
where: E is electric intensity; n
o
is unit normal to the surface of sphere; q
i
is i-charge inside sphere
of radius a.
It is obvious that the transversal electrical field of vector potential A
1
does not give the contribu-
tion to integral. The electric field strength is orthogonal to vector n
o
. Therefore when r we can
estimate a rate of change of a longitudinal electrical field formed by potentials A
2
and .
= grad
t
2
L
A
E
(B.2)
Now with expressions (B.1) and (B.2), we estimate a rate of change of longitudinal electrical field
E
L
. For the aim we write expression (B.1) as an approximation.
i
2
L
q r 4 E
(B.3)
With expression (B.3) it is easily to write:
APEIRON Vol. 7 Nr. 1-2, January-April, 2000 Page 51
2
i
L
r 4
q
E
(B.4)
Thus, when the charge sum inside sphere is different from zero then the longitudinal electrical
field decreases as r
-2
. If the sum is equal to zero and the pattern of electric charges has a form of a
multipole source then the longitudinal field decreases even faster.
References
Ginzburg V.L., 1987. Teoreticheskaia Fizika i Astrofizika, Moscow, Nauka, GIFML (in Russian).
Hayden H.C., 1993. Stellar aberration. Galilean Electrodynamics, vol. 4, # 5.
Kuligin V.A., 1987. Causality and physical interactions. In Determinism in a modern science
Voronezh University Press. (In Russian).
Kuligin V.A., Kuligina G.A., 1986. The mechanics of quasineutral systems of charged particles and
laws of conservation of nonrelativistic electrodynamics, deposited with VINITI, Sep. 4, #
6451-V86. (in Russian).
Kuligin V.A., Kuligina G.A., Korneva M.V., 1989. Lorentzs transformation and theory of knowl-
edge, deposited with VINITI, Jan. 24, # 546-V89. (in Russian).
Kuligin V.A., Kuligina G.A., Korneva M.V., 1990. The paradoxes of relativistic mechanics and
electrodynamics, deposited with VINITI, July 24, # 4180-V90. (in Russian).
Kuligin V.A., Kuligina G.A., Korneva M.V., 1994. Epistemology and Special Relativity. Apeiron,
(20:21).
Kuligin V.A., Kuligina G.A., Korneva M.V., 1996. The electromagnetic mass of a charged particle.
Apeiron, vol.3, # 1
Kuligin V.A., Kuligina G.A., Korneva M.V., 1997. Energy of radiators and transfer lines with
dispersion, deposit ed with VINITI, Feb. 20, # 538-V97. (in Russian).
Tikhonov A.N., Samarsky A.A., 1953. The equations of mathematical physics. GIFML. (in Rus-
sian).
Wallace B.G., 1969. Radar testing of the relative velocity of light in space. Spectroscope Letters,
vol.2, # 12. Pp. 361-367.
Addendum
We write the action integral of an electromagnetic field
= d ] A j F
4
[
ic
1
S
k k
2
ik
(C.1)
where:
k
i
i
k
ik
x
A
x
A
F
=
is the tensor of the electromagnetic field; d is elementary 4-volume (dx;
dy; dz; icdt).
We shall vary only potentials A
k
to find equations of the electromagnetic field (Landau,Lifshitz,
1975) . Also we shall consider that j
k
does not depend on potentials A
k
.
0 d } A j ] A
x x
Ak
A
x x
A
A
x x
A
[
2
{
ic
1
S
k k i
k i
k
k i
k
k
i i
k
=
=
(?.2)
After integration by parts we obtain
0 dS A
x
Ai
ic
1
d A ] j A
x
A
[
ic
1
S
i
S
i k
k
k k k
2
i
k
2
=
=
(?.3)
where dS
i
is an element of a 4-surface.
Page 52 APEIRON Vol. 7 Nr. 1-2, January-April, 2000
In the second integral we should evaluate values for limits of integration. The integration limits of
spatial co-ordinates is infinity. Here field is equal to zero. Under the initial condition the variation
of potentials is equal to zero an initial point t
a
and final point t
b
when we integrate over time. Thus,
the second integral is equal to zero (Landau,Lifshitz, 1975). The variation dA
k
is arbitrary. There-
fore expression in brackets of the first integral is equal to zero.
0 j
x
A
k
2
i
k
2
=
(C.4)
We write down the expression in the classical form
=
]
t c
1
[ ; ]
t c
1
[
1
2
2
2 2
2
2
j
A
A
(C.5)
Lorenzs condition ?
i
/ x
i
=0 follows from equation j
i
/ x
i
=0 and expressions (C.5).
As the second integral of expression (C.3) do not give any components in equations (C.4) and
(C.5) we can give other expression of the action integral.
0 d ] A j [
ic
1
S
k k
= =
(C.6)
where
2
i
k
)
x
A
(
2
=
(C.7)
is LaGrange function density of the electromagnetic field;
=
b
a
t
t
2 2 2 2 2
dxdydzdt } ] )
ct
( ) grad [( ] )
ct
( ) div ( ) rot [(
2
1
{ S jA
A
A A
(C.8)
With expression (C.7) and using (Landau,Lifshitz, 1975) we can write an energy -momentum ten-
sor of the electromagnetic field.
]
x
A
x
A
)
x
A
(
2
1
[ T
l
k
i
k 2
i
l
il il
=
(C.9)
From expression (C. 9) follows that energy density of the electromagnetic field (T
44
) is
] )
ct
( ) grad [( ] )
ct
( ) div ( ) rot [(
2
1
w
2 2 2 2 2
+ +
=
A
A A
(C.10)
Now we writ e energy of the electromagnetic field in a volume V
+
+ +
= =
V
2 2 2 2 2
V
dv ]} )
ct
( ) grad [( ] )
ct
( ) div ( ) rot [(
2
1
{ wdv W
A
A A
(C.11)
We consider energy change in the volume.
dv ]}
t
t
c
1
grad
t
grad [
]
t
t
c
1
div
t
div rot
t
rot [
1
{
dt
dW
2
2
2
2
2
2
V
A A
A A A A
(C.12)
After integration of expression (C.12) by parts and using of equations (C. 5) we obtain:
+ +
V
2 2 2 2 2
dv ]} )
ct
( ) grad [( ] )
ct
( ) div ( ) rot [(
2
1
{
A
A A
S
o
ds ] grad
t
) div
t
rot
t
(
1
[ n A
A
A
A
(?.13)
dv )
t t
(
V
A
j
APEIRON Vol. 7 Nr. 1-2, January-April, 2000 Page 53
where n
o
is a unit normal vector on surface S.
The potentials A and are independent and we may write the integral form of the law of energy
conservation
0 dv )
t
( dS ) div
t
rot
t
(
1
dv ] )
ct
( ) div ( ) rot [(
2
1
V V S
o 2 2 2
=
+ +
A
j n A
A
A
A A
A A
0 dv )
t
( dS ) grad
t
( dv ] )
ct
( ) grad [(
2
V V S
o 2 2
=
n
(C.14)
From expression (C.14) are visible that Pointings law of energy conservation not is unique. The
form of the law and its contents depend on gauge selection. It not is randomness, as the solution of
Maxwell equations depends on gauge.
We also want to pay attention to expression (C.11). Within the framework of the Lorenz gauge
the field energy of the scalar potential is negative. This fact is known for the experts of quantum
electrodynamics. Often they admire with a mathematical formalism, but sometimes overlook about
an essence of physical phenomena. The purpose of our article is to fill in this vacuum.
Reference
Landau L.D., Lifshitz E.M., 1975. The Classical Theory of Fields. Pergamon Press. New York.