The Maxwell equations including magnetic monopoles

© W.D.Bauer 27.02.04 [email protected]

Abstract:

The derivation of the Maxwell equations is reproduced whereby

magnetic charges are included. This ansatz yields following

results:

1) Longitudinal Ampère forces in a differential magnetostatic

force law are improbable. Otherwise a electric current would

generate magnetic charges.

2) Simple magnetic and electric induced polarization phenomena

are completely analogous and are described by a Laplace equation.

3) Permanent magnetic fields can be understood to be caused by

magnetic charges. Consequently, a moving permanent magnet

represents a magnetic current which generates a electric field.

4) The electromagnetic tensors of energy and momentum have some

additional terms which are written down generally.

5) If the electric material parameters are influenced by non-

electric variables (for instance temperature or pressure), the

formalism of electrodynamics is not sufficient to describe the

system and has to be completed by further differential equations

from the other areas of physics.

6) Nonlinear electro-thermodynamic systems may violate the second

law of thermodynamics. This is illustrated by a electric cycle

with a data storing FET invented by Yusa & Sakaki.

1
1) Introduction

The Maxwell equations are about 150 Jahre old. They are the

mathematical compilation of the experiments and considerations

based on the original work of Cavendish, Coulomb, Poisson,

Ampère, Faraday and others [1]. Mathematically they are partial

differential equations. Different notations exist for them: most

popular is the vector notation (O. Heaviside), which replaced the

original notations in quaternions (J.C. Maxwell). More modern

is the tensor notation (H. Minkowski, A. Einstein), which is able

to describe situations which are discussed in the theory of

relativity [2]. All notations are equivalent in the non-

relativistic limit.

The Maxwell equations were and still are very successful. Until

today their range of applicability grows permanently.

Here a short derivation is given which especially takes account

for the newer developments of material descriptions. Furthermore,

monopoles are included because Ehrenhaft proved their existence

already 50 years ago [3-6]. It will be shown that the theory

needs also their existence for a full description of all

problems. This explains perhaps effects which are regarded

generally as dubious because they cannot be understood in a

conventional approach.

2
2) The equations of the electromagnetic field

a) The laws of Coulomb and the equation of Poisson

The so called Coulomb law describes the force between electric

charges. It was discovered by Priestley in 1767 [1, 7]. Cavendish

rediscovered it again and measured as well the dielectricity

constant. However, due to many contributions to the knowledge

about electricity it has the name of the third discoverer

Couloumb [1].

The Coulomb law in the notation of today is [8]

1
2
M q q |x(x xx|)
PP!(x)!(x´) (x|x x´x´)| dx´
igj
i j
i
i j

j
3
3
dx 3 (1)

with the definitions F:=force, q:=single charge, x´ space

coordinate and i,j are indices. It can also be written as

P
F
!(x´)E(x´) dx´ 3 (2)

by using the definition of the electric field E

(x x´)
P
E
!(x´)
|x x´ | 3
dx´ 3 (3)

The electric field E can be derived from a potential - E

by using

the definition

E
/- E (4)

Then, the E-field is defined by

3
 - E

!(x´) dx´ 3
P|x x´ |
(5)

- E
has a empirical meaning. The E-field can be measured

experimentally by difference of voltage - (r) - (r

E E ref)
between a

point in space at r and a reference point at rref which oftenly is

set to infinity where no field exists. Using the Poisson equation

the charges the field can be calculated from the potential

°-E
/E
4Œ! (6)

If matter is in the field the empirical potential - E

consists of

the induced charges !matter(x´) and the contributions from the

charged surface of the conductors !conductor(x´)

-
-
E conductor
 -P
(!conductor(x´)!matter(x´)) 1 dx´ 3
P |x x´ |
(7)

where - P
is the “mean field” of the material charges. Thus, per

definition only the charges on the conductor are detected in the

experiment. In order to obtain a expression with empirical

variables similar to (6) the equation (7) is rewritten

!conductor(x´)
- :
- -
-
D E P conductor

P |x x´ |
dx´ 3 (8)

Contrary to the empirical meaning of - ,-

E D
has only a formal

character. Using -D
in the Poisson equation the charges to be

measured in or on the conductors can be calculated. One defines

4
 -
!conductor(x´) x x´ dx´ 3
D :
0ik E:
E4ŒP:
/ D
P |x x´ |
(9)

where 0ik is the dielectric tensor of the material.

Using the mathematical relations / |x x´ | 1
/´ |x x´ | 1 and

/´ 2 |x x´ | 1
4Œ/(x x´) and the redefinition ! :
!conductor the Poisson
equation is

Œ-D(x)
/.[0ik(x)/-E(x)]
4Œ!(x) (10)

Using (9) and (10) follows

/.D(x)
4!(x) (11)

Important special cases:

surface charges

An electric potential can exist due to a surface density 1

1conductor(x´)
- :
-
D conductor

P |x x´ |
dx´ 2 (12)

Then, the electric field

-
1(x´) x x´ 3 da´
D:
/ D
P |x x´ |
(13)

constraints for the material properties

In the most cases it is possible to make simplifying constraints

for the material properties. In order to explain this it is

necessary to write down the potential - (x)

P of the multipole

5
expansion of the charges in the material [9]. A multipole

expansion of the potential calculates the distribution of charge

in space about a origin 0 as a serie of moments

- (x)
!r  p
P
dipol

r3
.x
1
2
M
i, j
Qij
xix j
r5
 .... (14)

cf. appendix 1. Here the definitions of the dipol moment pdipol and

the quadrupol moment Qij are

P
p dipol :
x !(x )dx  3
P
Qij :
(3x  i x  j r 2/ij)!(x )dx  3 (15)

This consideration is done for all points in space. Using Pdipol

as density of polarisation then follows


Pdipol(x ) .(x x ) (xi xi )(xj x j)
- (x ,x)

û P
 !(x )
|xi xi|


|xi xi|
 3
1
2
M Q (x )
i, j
ij


|xi xi|
 5
 .... dV (16)

The first term represents induced charges for instance if

recombination processes in semiconductors have to be accounted

for. However, for the most problems electric neutrality can be

assumed and the first term becomes zero. Furthermore oftenly

higher terms are neglected because they are quantitatively

irrelevant. Then, after integration over the whole space holds

(x x ) /Pdipol(x ) 

- (x)
PP
P dipol(x

). dx  3

P dx  3 (17)
|xi xi | |xi xi |
 3 

If (17) is inserted in (8) one can identify: P=Pdipol

6
 b) Ampère´s law

The discovery of electromagnetism by Oersted [10, 11] in 1820

inspired some researchers in France to find the quantitative laws

of these effects. Especially, Biot&Savart and Ampère tackled the

task to solve this problem by intelligent experiments [1, 11].

In order to fit their experiment by the theory they made

additional assumptions which filled up some lacking observations.

This led to different laws for the forces between differential

current elements of a circuit. For closed circuits, however, the

different laws coincided in one. The discussion of this problem

is running until today.

Biot and Savart [12-14] found out that “the total force which is

exerted by a file of infinite lenght under current on a element

of austral or boreal magnetism in the distance FA or FB, is

perpendicular on the shortest distance between the molecule and

fig.1a: the Biot-Savart - setup Fig.1b: the Biot-Savart - setup

A magnetic needle is under the influence of the measuring the time constant of the torsion
field of current CZ . A cover protects against pendulum it is concluded on the force of the
the movement of air. The magnet A’B’ compen- field on the needle, if the current flows.
sates the magnetism of earth where the needle Distance and angle of the file are varied in the
is located. experiments.

7
the file (see figs.1)”. This law is written today in a form which

goes back to Grassmann [11, 15]. It holds [8]

ds1×(ds2×r). ds .r
dF  i1.i2
i1.i2 (ds1.ds2) r3  13 ds2 (18)
|r|3 |r| |r|

with i1/2 := current, ds1/2 := length of file element, r:=

distance between file elements.

Ampère idolized Newton a little bit. So he overtook, 1) that

Newton’s 3.axiom (actio-reactio) of mechanics also holds for

electromagnetism, and 2) that the force between single elements

of current is a central force and lies on the distance line

between the elements.

Based on his own experiments on closed circuits Ampère included

the following observations [16, 17] in his theory, see fig.2a-d:

1) The force of a file under current reverses if the current

reverses, see fig.2a.

2) the forces of a current, which flows in a smooth circular

circuit, is the same, if the “circle” of the current is not

smooth but sinoidal, see fig.2b.

3) the force of a closed current on a single current element is

perpendicular to it, see fig.2c .

4) the force between two current elements does not change if all

spatial dimensions of the setup are enlarged by a constant

factor, see fig. 2d.

Applying these observations Ampère constructed his force law.

Based on Ampere´s assumption it holds for the force F  r . The

observations 1)+ 2) suggest for first order

8
fig.2a: Ampère´s first experiment fig.2b: Ampère´s second experiment
AB is a fixed conductor under current. The in a trench PQ flows a current straight on in a
circuits d’c’fe and cde’f’ are stiffly connected , conductor, in the trench SR in a sinoidal
are symmetrical over AB and can rotate about conductor. The circuits BCDE and FGHI
the axis x’y’. Their orientation of the current is mounted stiffly together, but can rotate around
opposite in these circuits; experimental result: the Axis AK . The same current flows through
no rotation due to complete balance of opposite them, however in opposite direction.
forces experimental result: only if the circuit is exaxtly
in the middle between the conductors all forces
compensate and no movement is observable.

fig.2c: Ampère´s third Experiment fig.2d: Ampère´s fourth Experiment

M and M’ are trenches filled with mercury, arm the outer circuits are fixed, the circuit in the
OC can be turned. The current flows over the middle can move. Only, if the diameters fullfil
troughs M back to the arm OC. The arm turns the relation dleft :dmiddle =dmiddle:dright, all forces
into the middle, where a equilibrium of torque compensate and the circuit in the middle NOM
exists and where all forces on OC apply does not move.
perpendiculary.

9
F  i1.i2[3(r).(ds1.ds2)  %(r)(ds1.r).(ds2.r)] , the combination of both

proportionalities result in F  i1.i2r[3(r).(ds1.ds2)  %(r)(ds1.r).(ds2.r)] .

Observation 4) implies 3(r)=A/r3 and %(r)=B/r5 with A and B as

constants to be determined. These can be calculated applying

observation 3) as shown in the proof below. So follows B = -3A/2.

Proof[1]:
Imagine two circuits located with an angle of 90° between. Due to observation 3)

holds for a closed circuit

F ,i .i r rA (ds .ds )  rB (ds .r).(ds .r) ds

0
1 2 3 1 2 5 1 2 2

This equation is rewritten as

A (ds1.ds2).(ds2.r) B(ds1.r).(ds2.r)2

r3 r5

Because the integral over the circuit is zero, a potential exist and consequently

also a total differential. If the circuit is chosen to be a round circuit one can

replace by ds1=-dr and write

A d(dr.ds2).(ds2.r) B(ds1.r).(ds2.r)2

r3 r5

Due to the potential property follows Qxy = Qyx and then

A
d
3

B5 (ds1.r)
2r r

With ds1=-dr this becomes

3A4 dr
B4 dr
2r r

and B = -3A/2 follows. q.e.d a

10
So Ampère´s law is written :

i1.i2 2 3
F
r[ .(ds1.ds2) (ds1.r).(ds2.r)] (19)
2 3
c r r5

Riemann [18] and Whittaker [1] checked this derivation and

realized, that Ampère’s workout is only one possible ansatz to

explain the observations. They doubted in Ampère’s assumption,

that the force between current elements is a central force,

because the forces could be as well angular moments [19]. They

found other possible formulas, which could explain all

observations. Whittaker enlarged Ampère’s formula, by adding

terms, which were in accordance with the observations on closed

current loops, because they were zero after integration over a

closed loop. So he made the general ansatz:

i.i´ 2 3
F
r[ .(ds.ds´) (ds.r).(ds´.r)]
2 3
c r r5
 $(r)(ds´.r).ds  $(r)(ds.r)ds´  $(r).(ds.ds´).r (20)
 1 $´(r).(ds.r).(ds´.r)
r

Whittaker dropped Ampère’s assumption, that the force should be

a central force and he applied only Newton’s law actio-reactio.

He made the most simple possible choices $(r)

i.i´/(c 2r 3)
, $´(r)
3i.i´/(c 2r 3) and obtained the force law

i.i´
F
[(ds.r)ds´  (ds´.r)ds r(ds.ds´) ] (21)
c 2r 3

11
Tabel 1: different versions of magnetostatic

force law between current elements ([20] and [1])

general form of the magnetostatic force law:

i.i´
F
k [r.(A.(ds.ds´)B.(r.ds)(r.ds´)/r 2)C.(r.ds´)dsD.(r.ds)ds´]
3
r

name year ref. A B C D comment

------------------------------------------------------------------------

Ampere 1823 [17] -2 3 0 0 central force

Grassmann 1845 [11, 15] -1 0 0 1 no monopoles

Riemann 1875 [18] -1 0 1 1 moment conserved

Whittaker 1912 [1, 15] see under Riemann

Brown 1955 ???? 1 -6 6 6 ?????

Aspden 1987 [21] -1 0 1 -1 cons.angular moment

Marinov 1993 [22] -1 0 0.5 0.5 experiment

Cavallieri 1998 [23] see under Grassmann experiment

Of course this force law was not convincing as well.

For the basic idea of Riemann and Whittaker was used by many

others who built their “own” force laws using other assumptions.

The discussion is running until today, see[23] and tab.1.

All different forms yielded the same result for the magnetic

field H , if they were integrated over a closed circuit.

I
c
l x |x× ds´|
3
or generally H

1
c
l j(x´) × |xx x´|x´
3
d 3x  (22)

12
general scheme of proof for every magnetostatic force law:
According to a general theorem of vector analysis, see appendix 2, every

vector field can be decomposed into a vortex field and a potential field. The

vortex field is caused by currents, the potential field by charges. If this

is compared with theorem 2 in appendix 2, then the Biot-Savart law generates

a vortex field. All other fields deviating from Biot-Savart, have to be

written as

field law = Biot-Savart-law + additional terms

These additional terms must be identified as a potential field. If the

current is integrated over a closed circle the potential terms cancel to

zero1 . a

The force of a closed circuit on a differential current element

is according to Biot-Savart, see (18) and (22),

i
dF
H×ds (23)
c

If one integrates over two interacting closed circuits the von

Neumann force law is obtained [1, 8, 24, 25]

I1 I2
c 2
ll |xx | ds ds
12
3 1 2 (24)
12

From (22) also follows, that the magnetic field can be calculated

The Biot-Savart law is probably the correct version for

physical currents. It does not generate “magnetic charges” and
coincides with the B-field of a moving charge according to
Lienard-Wiechert(in the special case of zero acceleration).
Experimentally the Biot-Savart - law is supported by the
measurements of Cavallieri [23].

13
from a vector potential H := /×A with

A(x)

1
c
l |xj(x´)
x´|
3
d x 
(25)

Then follows

div rot A
div H
0 (26)

So Ampère concluded: The cause of the magnetic field are not

magnetic charges but only currents.

Ampère’s theory includes as well para-, dia- oder ferromagnetic

“excited” materials. The total magnetic field B includes the

field from the measurable currents j and the field M of the

magnetism of the material, where the field M (according to

Ampère) is generated exclusively by currents in the material.

Then follows

4Œ
rot B
(j j ) (27)
c conductor material

with

B:
µH oder B:
H4ŒM (28)

Analogously like for charges a relation is sought between the

empirical variables. So the unknown current jmaterial is eliminated.

If compared with electrostatics, see eq. (14) to (17), it can be

derived for currents, that for magnetostatics holds

/.j!
0 and j()
0 , i.e.
Pj d x
P/.j d x
P! d x
0 . Here is applied
3 2

2

14
no currents exist at the boundary in the infinite. Thus no

charges can be built up there and only dipol terms and terms of

higher order can contribute to the result. So a definition (29)

analogous to (17) is used for the magnetization M of the material

/×M :
j material /c (29)

Then, using (27),(28) and (29) Ampère´s laws are derived

4Œ
rotH
j divB
0 (30)
c conductor

In order to derive the present version Ampère´s law is rewritten

as [8]:

/×H
rot rotA
grad divA /2A
(31)

/ j(x´) /. 1 d 3x´ j(x´) /2 1 d 3x´
P c |x x´| P c |x x´|

With the mathematical relations / |x x´| 1

/´|x x´| 1 and

/´ 2 |x x´| 1
4Œ/(x x´) this becomes

4Œ
/×H
/ j(x´) ./´ 1
d 3x´  j (32)
P c |x x´| c

because A also fulfills the Poisson equation /2A

4Πj/c .
If the integral in (32) is integrated partially using that j

vanishes at boundary in the infinite, then follows

/×H
4Œ j / /´ j(x´) d 3x´ (33)
c c|x x´|

15
Now, the observation is used that no charges build up during

magnetostatic experiments. Using the continuity equation this

fact can translated into mathematics by /´ j´  N

0 .
This yields Ampère´s law of magnetostatics:

/×H
4Πj oder lHds
4ΠjdA
c cP (34)
S

Comparing the coefficients of (31) and (33) follows grad div A=0.

Oftenly, it is assumed div A =0. This expression is known as

Coulomb-gauge. The vector potential A is not a unique function,

because A can be replaced by A*= A + /f(x) . The important point

for the choice of vector potential A is that grad div A=? has to

be chosen such, that a physically motivated constraint is

fulfilled - the continuity equation [26].

At the time of Biot&Savart and Ampère this was not known fully

and only the closed circuits could be tested out. So the result

(22) for the H-Feld was ok. . However later, after the discovery

of the electron by J.J. Thomson [27], discussions came up due to

the basic problem behind the approaches of Biot&Savart and

Ampère: Not every magnetic problem could be discussed by a closed

electric circuit. Freely moving charges (as differential current

elements) could exist and the question for their field had to be

solved. So observations were published that longitudinal forces

existed in railguns [28, 29] and in plasma tubes [30, 31] (See

also the review article [32]). These forces seemed to be

explained by Ampere’s differential force law, but not by Biot-

16
Savart´s version. Althought these problems seem to be solved

today not in favour for longitunal forces2 the problem will be

left open here for further considerations. So all mathematically

possible field configurations will be included in the discussion

by adding a magnetic potential to the magnetic vector field. So

any vector field F can be decomposed into two terms FC and FV,

derived from a potential (for FC) and a vector potential of

vortex field (for FV), see the proof in appendix 2 [26] and [35,

36]. Thus any H-Feld is described by

H
HV  HC

1
c
,j(x´) × |xx x´|x´ d x ,N (x´) |xx x´|x´ d x
3
3 
H 3
3 
(35)

Here NH is the magnetic charge distribution due to the deviation

from Biot-Savart´s differential law, see eq.(18).

If a concrete system is solved with a boundary problem, a Laplace

field HL has to be added which satisfies rot HL =0 and div HL =0.

H
H V  HC  HL

1
c
,j(x´) × |xx x´|x´ d x ,N (x´) |xx x´|x´ d x /3(x)
3
3 
H 3
3 
(36)

Here is HL :
/3(x) the Laplace field. This potential describes a

field, which is generated outside of the defined area of the

problem. The field HL helps to adapt the solution to the given

Both observations were explained later by Rambaut

&Vigier[33], see as well [34]. They pointed out, that these
observations do not answer the question, because a closed
moving circuit shows a “longitudinal” mechanical expansion due
to a “expansion” pressure of a loop due to the Lorenz force.

17
boundary condition of the problem. Then (36) changes to

/×A / /3(x)
/×
H
HV  H C  HL
/
1
c
, |xj(x´)
x´|
3 
/,
d x /
N (x´)
H
|x x´|

d x /3(x) (37)
3

with A:
c 1 ,j(x´)/|x x´|d 3x´ as magnetic vector potential and

 :
,NH(x´)/|x x´|d 3x´ as potential function of the magnetic charges.

If magnetic charges are included the magnetic field becomes a

general field and loses all symmetry properties with respect of

parity. So every field configuration can be described generally.

It will be shown here that this is useful for problems with

induced and permanent magnetization. Only a reinterpretation of

the conventional point of view leads to a Poisson equation for

magnetic charges.

Proof:
The conventional theory for problems with permanent magnetization [8]

(without exciting field from outside) assumes, that

/.B0
/.(H0  4ŒM)
0 (38)

Here is M the magnetization of the material and H0 the inner magnetic field

which generates the magnetization. If a field is applied additionally from

outside, this equation is enlarged

/.B0
/.(H0  H  4ŒM)
0 (39)

with H a the exciting H-field from outside, which is added. Because no

currents are obvious in matter as cause for the inner H0-field it holds

/×H0
0 . This means that H0 can be derived from a potential H 0
according

to H0
/ H 0
and the magnetostatic Poisson-equation follows [8]

18
 ŒH0
4Œ!M (40)

with /.H0 :
4 Œ!M defined as “effective magnetic charge density” in [8].

From (39) and (40) follows

/.B :
/.(H  4ŒM)
4Œ!M :
4Œ!H (41)

With these redefinitions the conventional equation (39) is written down with

magnetic charges in a form, which is completely analog to electrostatics.

Then, analogously to electrostatics, the empirical field is the H-field

contrary to the conventional interpretation taking the B-field. q.e.d. a

Thus, magnetostatic boundary problems can be worked out

analogously to electrostatics with changed boundary conditions.

Textbooks show [8], that the solution of magnetic boundary

problems are sometimes completely analogous to electrostatics.

Similarly, permanent magnets (like an analog to ferroelectrets)

fig.3: polarized bowl in a potential field

boundary condition between inner of bowl and outside:no charges and no currents, i.e.
Binnen=Baußen , similar like in electrostatics the field is given by the charge distribution at the outer
boundary: By=0 , /Blinks = -/Brechts . It holds the Laplace equation û3 =/B=0. For the equations of a
metal bowl in the electric field the magnetic variables have to be replaced by electric ones.

19
can be regarded to consist of magnetic charges.

For the simple phenomena of induced polarization, however , see

fig.3, the fields are derived by the assumptions, that B is the

solution of the equation /.B

0 . The boundary condition represent
either given current distributions exciting the material, either

they are the existing field in a distance far from the object

under consideration. This means, that the magnetic field H in the

neighborhood of a magnetic material fullfils locally always

û3
/.B
0 and /×H
0 . Or the Laplace equation holds for the
induced magnetism.

Only, if the experiment deviates from the theory, magnetic

charges are probable. This is the case for the ferromagnetic

hysteresis of iron. If compared with the conventional parity

tabel, see tab.2, the B-field has (-1) parity under time

inversion, i.e. if the current is inversed, the field has to be

inversed as well. If a hysteresis exists, this is not the case,

because the hysteresis line B(H) is not unique. For a change of

parity with fields of the strenght of the coercitivity, the

change in parity can easily be disproved. In this case

inhomogenities or gradients of magnetic permeability µ(x) can

induce magnetic charges. Then it holds

/.B(x)
/.(µ(x).H(x))
µ(x)/.H(x)  H(x)./µ(x)
4ŒNM g 0 (42)

20
fig.4: field lines of magnetic H-field of a cylindric permanent magnet
the magnet is modeled here as capacity of magnetic charges.The magnetic charges are distributed
on the surfaces of north and south pole. The iron has a permeability of µ=10000

Tab.2: symmetry properties of conventional electrodynamics

It holds generally: F(u) = P . F(-u)

variable u field F parity P sort of field

x --> -x E -1 potential

D -1 potential

H 1 vortex

B 1 vortex

cause

t --> -t E 1 charge

D 1 charge

H -1 current

B -1 current

21
Consequently, a magnet can be modeled as well using magnetic

charges, see fig.4. Any permanent magnetism destroys any parity

of a B-field similar like it is proved in the original

experiments of violation of parity in beta-decay.

Of course a general B-field with no parity cannot be explained

solely by a vector potential A,

1) because B = rot A has a defined parity;

2) because B:=rot A implies div rot A = div B = 0 follows, which

is contradicting to the physical result div Bg0.

So, for magnetic charges the magnetic potential  has to be

introduced. Similar like the magnetic vector potential A it has

a more formal character, because it is not known very much about

magnetic charges except of Ehrenhaft´s [5, 6] and Mikhailov´s

experiments [37-48]. Important questions about concentrating,

storing and conducting of magnetic charges are open.

The potentials of the magnetic field are

1 jB,H,M(x´) 3 NB,H,M(x´)
AB,H,M
d x´ B,H,M
d 3x´ (43)
c P |x x´| P |x x´|

Then, the magnetic fields can be derived

B V
rotAB , BC
/B

HV
rotAH , HC
/H (44)

M V
rotAM 4Π, MC
/M 4Œ

using the definitions B C :

HC  4ŒM C , B V :
HV  4ŒM V , AB :
AHA M and

B :
HM . The empirical magnetic field is

22
 H
HC  H V  HL (45)

For the B-field holds:

B :
B V  B C  B L (46)

Ampère´s law are written (using rot HC/L=0 and div BV/L=0):

/×H
/×HV
4Πj /.B
/.BC
4Œ!B (47)
c

Then the general force law of magnetism is:

1
F
j× B V dx´ 3 NH.(HCHL) dx´ 3 (48)
cP P

Later Ampère´s law /×Hv

4Πj/c was extended by Maxwell. Maxwell
realized [49], that it could not describe cases, where electric

charge appeared, which were stored in capacitances. Maxwell

solved the problem by a hypothesis, which turned out to be very

useful, especially with respect to the theory of electromagnetic

waves. He changed Ampère’s equation to

/×HV
4Πj  1 0D (49)
c c 0t

Introducing the dielectric displacement dD/dt Maxwell removed a

contradiction between physics and mathematics, because the

continuity equation as a constraint could always be fulfilled

4Π1 dD

4Πdiv j  E
0
dN
div rot HV
div j (50)
c c dt c dt

23
This form of Ampère´s law holds until today. It can describe as

well the cases where charges are generated, for instance

electron-positron pairs in high energy physics, electron-hole

pairs in semiconductors, or dissociations into ions in chemistry.

Maxwell´s improvement does not change as well the gauge relation,

because using (31) it can be calculated

grad divA
/
/´.j(x´) d 3x´
/ N E (x´) d 3x´
1 dD (51)
c|x x´| c|x x´| c dt

So the vector potential for Ampère´s law (34) can be retained.

c) Faraday’s law

The induction law has been found by Faraday. Using his

formulation it is written

d
U
(52)
dt

For Faraday the flux =,B dA were the number of field lines,
which go through a closed circuit. For a expanding or contracting

circuit this is written today [9]

l C
Eds

1 d
c dt
, S
B dA

1
c
, S
0B dA  1
0t c
, /×(B×v) dA  1c ,
S S
v(/.B) dA (53)

A simple derivation can be done using the formalism of special

relativity, see section e). This law can be formulated

alternatively using (53) , 4ŒNH

/.B and jH
!H v

24
 /×E
1 0B 1 /×(v×B)  4Œ jH (54)
c 0t c c

It will be shown in the next section, that this equation is

consistent with a gauge by a continuity equation for magnetic

monopoles.

d) the complete Maxwell equations

The Maxwell equation describe the coupling of fields with moving

charges in space. They can be generalized that they hold for

solids and for gases and liquids.

The notations for the indices here are C:=charge, V:=vortex,

E:=electric field and H:=magnetic field.

If magnetic charges are included the Maxwell equations are

(using the definitions v=velocity and j:=v!)

lEVds
d :
1 d BdA lHVds
d, :
1 d PDdA oder
dt c dt P dt c dt

/×EV
1 0B /×( v ×B)  4Œ !Hv /×H V
1 0D /×( v ×D)  4Œ !Ev
c 0t c c c 0t c c
div DC
4ŒNE div B C
4ŒNH
(55)
v v
N E /.jE
div /×( ×B)
0 N H /.jH
div /×( ×D)
0
c c
Œ3D
div D L
0 Œ3B
div BL
0
D
D V  DC  DL B
BV  B C  BL

For a mixed system of charged particles the individual equations

of each sort of particle have to be added together.

In the version above Ampere´s law is extended by the so called

25
 Y Y
%
' +
 

(
fig.5a: E-field due to the Lorenz force fig.5b: H-field due to the Rowlands force
at the expansion (or contraction) of a circuit in a at the roll out of a conducting foil over a
magnetic field polarized electret material

Rowlands term which is electric analog to the Lorenz force. This

term takes account for a H-field, which is generated, if a

capacitance grows in an electric field, see fig.5b.

Similarly the Laplace field is accounted for in (55).

To complete the theory an electric vector potential must also be

introduced. It is generated by magnetic currents.

All generating potentials are listed in (56)

ND,E,P(x´)
- D,E,P

P |x x´|
d 3x´ +D,E,P
1
cP
jB,H,M(x´)
|x x´|
d 3x´

(56)
1 jD,E,P(x´) 3 NB,H,M(x´)
AB,H,M
d x´ B,H,M
d 3x´
c P |x x´| P |x x´|

They are interconnected with the fields by

26
 B V
rotA B , BC
/B
HV
rotA H , HC
/H
M V
rotA M 4Π, M C
/M 4Œ
(57)
DV
rot+D , DC
/-D
EV
rot+E , EC
/-E
PV
rot+P 4Π, PC
/-P 4Œ

Summarizing it can be said about the Maxwell equations:

Electric and magnetic fields can be described mathematically as

general fields. Their causes are charges and currents of electric

and magnetic particles, which fullfil the continuity equation as

a constraint. Due to the mathematics the electric and magnetic

fields can be decomposed into a vortex, a potential field and

a Laplace field. The charges build up the potential fields, the

currents the vortex field and the Laplace field adapts to the

boundary conditions.

e) The Maxwell equations and the theory of relativity

In the theory of relativity the Maxwell equations are formulated

in the terminology of tensor calculus.

The theory of relativity relates the variables measured in a

reference system to the variables of another system which moves

relative to the first system. The transformation applies for a

movement in z-direction (using the definitions :=v/c,  :

1/ 1 2 )

27
  0 0 i
0x´i 0 1 0 0
aij

(58)
0xj 0 0 1 0
i 0 0 

Similarly vectors are transformed (using the Einstein convention)

A´ i
a ij . Aj (59)

Tensors T´ij are transformed by

T´ ij
a ik . ajl . Tkl (60)

The 4-vectors of the theory of relativity are, cf. appendix 3,

space coordinates : x
( x, y, z, ict)
momentum : p
( px, py, pz, imc)
i
wave number : k
( kx, ky, kz, &)
c
electric 4 current : j E
( jx , jy , jz , ic!E)
E E E
(61)
magnetic 4 current : j H
( jx , jy , jz , ic!H)
H H H

electric Lorenz vector : L E

( Ax , Ay , Az , ic-D)
E E E

magnetic Lorenz vector : L H

( +x , +y , +z , icB)
H H H

The 4-vectors are invariant, i.e. the length of a vector is

independent from the state of movement of the reference system.

From this property and from (61) follows the continuity equation

d i dN d i dN
j E
div jE  E
0 j H
div jH  H
0 (62)
dxi dt dxi dt

An analogous equation - the Lorenz gauge -holds as well for

Lorenz vectors, see appendix 3. The definitions for the

28
electromagnetic tensor field at no current (v=0) are

0 E3 E2 iB1 0 H3 H2 iD1
E3 0 E1 iB2 H3 0 H1 iD2
F :

ij
G :

ij
(63)
E2 E1 0 iB3 H2 H1 0 iD3
iB1 iB2 iB3 0 iD1 iD2 iD3 0

If there no current is flowing, (j=!v=0) the 4-currents are

jE
( 0, 0, 0, ic!E) jH
( 0, 0, 0, ic!H) (64)

Then the Maxwell equations can be written

d ij d
F
4ΠjH i G ij
4ΠjE i (65)
dxj dxj

The complete system (55) of Maxwell equations follows if the

charges move. This is described by the following coordinate

transformation

d d 0x´n 0x´k ij 0x´k i

F´ kn
F
4Πj
4Œ j´H k
dx´n dx´n 0xj 0xi 0xi H
(66)
d d 0x´n 0x´k ij 0x´
G´ kn
G
4Πk jE i
4Œ j´E k
dx´n dx´n 0xj 0xi 0xi

(66) represents the complete Maxwell equations in tensor

notation, cf. (55). One consequence should be emphasized:

if currents exist the complete Maxwell equations have to be

applied including the terms of Lorenz and Rowlands force.

29
 f) the electromagnetic tensors of momentum and energy

The electromagnetic conservation of energy

The power of a electrically and magnetically charged particle is

v v
(using FH
qH (H ×D) , FE
qE (E ×B) and F
FE  FH )
c c
dEmech

F.v :
qE .E.v  q H .H .v (67)
dt

This equation integrated over the whole space yields with j:= v!

dEmech

F.v :
( jE .E  j H .H ) dx 3 (68)
dt P

If the Maxwell equations are solved for the currents, (i.e.

1 0D
/×E 1 0B /×( v ×B) and j E
4Œ /×H c 0t /×( c ×D) ) and inserted
c c v
jH

4Πc 0t c
in (67), and using /.(a×b)
b.(/×a) a.(/×b) , it follows a modified

Poynting energy conservation equation for the energy density:

/.S 0U  c /×( v ×B) .H  c /×( v ×D) .E
demech
(69)
dt 0t 4Πc 4Πc

Here the following definitions have been used

c
S:
(E×H)
4Œ
(70)
dU 1 dD
:
E  H dB
dt 4Πdt dt

The last two terms in (69) are non-standard, because the energy

conservation is derived always without Rowlands and Lorenz terms.

30
 The electromagnetic conservation of momentum

The force on a charge distribution of electromagnetic charge is

jE jH
Fmech
!E ( E  ×B)  !H ( H  ×D) dx´ 3 (71)
P c c

Using again Maxwell’s equations solved for j this can be written

Fmech
[E /.D  H /.B  (/×H )×B(/×E )×D 1 0D ×B  1 0B ×D ]dx´ 3 (72)
P c 0t c 0t

Using the definitions E  :

Ev/c×B and H  :
H v/c×D and the

calculation in footnote3 follows

d  dE
Fmech
Tik D B dH d D×B
dxk dxk dxk dt 4Œc
(73)
v v
 (/×( ×B))×B  (/×( ×D))×D dx´ 3
c c

Here Tik is defined as Tik:

EiDjH iBj . The fourth term of the first

line of (73) is pFeld:

(D×B)/(4Œc) which is defined as the

3
The first three vector terms can be written in the
terminology of the tensor calculus:

0Es 0H
E./.D  H./B  (/×H)×B(/×E)×D
0ijk0jls Dk  0ijk0jls s Bk .
0xl 0xl
Using 0ijk0jls
/kl/is /ks/il the first term is transformed to

0Es 0H 0D 0E 0E 0B 0H 0H
0ijk0jls Dk  0ijk0jls s Bk
Ei j Di k Dk k  Hi j Bi k Bk k
0xl 0xl 0xj 0xk 0xi 0xj 0xk 0xi

In the 2nd and 5th term k can be exchanged with j without

changing the result.Then follow the first three terms of (73).

31
electromagnetic momentum pFeld of the field.

If the generality of (73) is restricted (i.e. if only materials

are used with purely linear constitutive relations like B=µH and

D=0E) then the first three terms of (73) represent the Maxwell

energy tensor:

dTik /
(EiDkHiB k ik (0E 2µH 2))
d
:
(74)
dxk dxk 2

This equation is found in the textbooks normally. The last two

terms of (73) are omitted always, because “shorted” Maxwell

equation are used which is wrong in the general case according

to the author´s opinion.

The equations of conservation of energy and momentum describe the

behaviour of a generalized capacitive-inductive- electronic

element. Special cases for the energy equation are the pure

capacitance (if H=0 and B=0) and the pure coil (if E=0 and D=0),

see (73). For these cases the equation says, that the energy

flowing into the electronic element can be identified with the

electric or magnetic field energy.

The definition of electromagnetic work can be done if (69) is

applied

dE mech
Wel
dt (75)
P dt

It should be said that the discussion about the “correct”

equations (69) and (73) is alive until today, cf. [50].

32
It is remarkable that the derivation with monopoles yields the

same result as without. The cause of this may be, that many

Maxwell equations are the solutions from the the theory of

general relativity, because one degree of freedom remains

undetermined during the derivation [8] [51, 52]. These conside-

rations were done for the shorted Maxwell equations (55), i.e.

/.D
4Œ!E /×H
1 dB  4ΠjH
c dt c
(76)
/.B
4Œ!H /×E
1 dD  4ΠjE
c dt c

It can be shown, that all these equation can be transformed by

E
E  cos  H  sin D
D  cos  B  sin
H
E  sin  H  cos B
D  sin  B  cos
  (77)
!E
!E cos  !H sin jE
jE cos  j H sin


 
!H
!E sin  !H cos j H
jE sin  j H cos


If the parameter  in (77) is chosen appropriately, the

conventional Maxwell equations without magnetic charges are the

result. It is shown that relativistic pressure tensor (shorted

calculation without Lorenz and Rowlands terms !) is invariant

under these transformations.

If it is believed, that every electric charge is in a constant

proportion with a magnetic charge, -so the argumentation and the

calculation of Harrison[53] and Katz [52]- the combined charge

is regarded as a new “elementary charge”, and built up a

transformed (shorted) system of Maxwell equations with div B=0

33
[8, 52, 53]. So it is understandable, that Mikhailov [38, 48, 54]

tried to determine the proportion between electric and magnetic

charge, especially because the first workout of his measurements

[38] spoke against the generally accepted theoretical value of

Dirac [55, 56]. Anyway, in the light of these opinions of

Harrison[53] and Katz [52], one can ask why Mikhailov sees any

effects at all. For author the discussion is not at the end here.

Perhaps, parity checks can solve this question.

g) boundary conditions

stationary discontinous boundary conditions by charges

In order to derive boundary condition equation (13) is applied

on a fictive “pillbox” at the boundary between two materials of

a potential field [8], see fig.6a .

So one obtains the relation (with 1:=surface charge density)

/.FC dV
FC.n da
(FC(1) FC(2)).n Œa
4Œ1Œa
P P (78)
S

Equation (78) shows a relation between vector components of the

field F1 in region 1 and F2 in region 2 which both are normal to

the surface. This yields for the vertical components of

dielectric displacement DC :

3
4Œ1E
(DC(1) DC(2)).n (79)

34
 F2 F2
n
t

F1 F1

fig.6a the pillbox - construction fig.6b the circuit - construction

for the determination of boundary conditions for the determination of boundary conditions
due to charges due to currents

i.e. at the boundary there is a discontinuity which is determined

by the surface charge density. An analog holds for the

vertical component of the magnetic field BC :

3
4Œ1H
(B C(1) B C(2)).n (80)

For a electric or magnetic conducting surface holds

-
constant 
constant (81)

stationary boundary conditions by currents

Equation (29) can be applied to derive a boundary condition if

a surface current k flows at the boundary between regions of

different materials, see fig.6b . So one obtains [8]

/.×FV dA
FV ds
(n×t).(FV(1) FV(2))ûl
4Œ k.t ûl
P P c (82)
S

Equation (82) is a relation between the vector components F1 and

F2 which flow tangentially on the surface of the boundary

between two regions 1 and 2 of different materials.

35
 discontinuities of the magnetic vortex field

for tangents to the surface

3×(HV(1) HV(2))

n
4Œ
K (83)
c E

discontinuity of the electric vortex field

for tangents to the surface

3×(EV(1) EV(2))

n
4Œ
K (84)
c H

For more general, nonstationary boundary conditions at moving

surfaces, see [8].

h) the constitutive equations of the material

The system of Maxwell equations can be solved after the

constitutive equation are known which describe the material

properties. They couple the electric variables (E,D) and the

magnetic variables (B,H) which can be represented generally by

D E
B

coupling q H (85)

In the most cases these couplings are simple, i.e.

as resistor: j
1.E

or capacitively: D
0E (86)

or inductively: B
µH

36
Initially the material constant were constants which described

the simple cases of material properties. Later more complicated

nonlinear functions were found which could generate phase

transitions, i.e.

1
1(E), 0
0(E), µ
µ(H) (87)

After the fundamental crystal structures were known, the material

properties could be correlated to the symmetry of the crystals.

Then, the constitutive equation were described by tensors

1
1 (E), 0
0 (E),
ik ik µ
µ ik(H) (88)

which were first linear, then non-linear.

Then, materials were discovered whose properties were magnetic

and electric, and where an electric field influenced the magnetic

properties and vice versa [57] [58].

The theory of relativity found out that dielectric or magnetic

polarized material behaved different if it was set in motion. The

following equations are from [59]

v
E´
E ×M
c
(89)
v
H 
H ×P
c

A further complication of the constitutive relations are space-

dependence of the material properties which are realized for

instance as electronic elements.

Furthermore all materials had their own dynamics in time in the

37
form of relaxation time.

If all material properties are accounted for then the general

constitutive equations can be abstracted as additional

differential equations which help to solve the complete system

of partial differential equations. This system can be written as


E f (E,D,H,B;T,! ,x,
 &,...)(x,t)

f 1(E,D,H,B;T,! i,x, (90)

H 2 i  &,....)(x,t)

or


D f (E,D,H,B;T,! ,x,
 &,...)(x,t)

f 1(E,D,H,B;T,! i,x, (91)

B 2 i  &,....)(x,t)

The variables after the semicolon show that the constitutive

equations may not depend only from electromagnetic parameters,

but can depend as well from mechanic or thermodynamic material

properties. This means that the electrodynamics cannot be

separated from the other areas of physics. If these the material

properties drift under the influence of electromagnetic fields

then a purely electrodynamic description is not sufficient and

further differential equations from other areas of physics have

to be added to a complete partial differential equation system

to be solved.

Examples:

1) Known examples are electric motors and generators. Here the

mechanic equations of motion of the motors are added. They

describe the motion by the angular coordinate of the rotor.

38
2) Other systems are magnetic materials, for which the Landau-

Lifshitz-Gilbert - equation [57] [60] hold


.M×H  ..M×(M×H )
M (92)
eff eff

It generates a system of partial differential equation (:=PDE)

if it is combined with the equation of the magnetostatic

potential (41) [61]. It allows to calculate magnetic domains in

ferromagnetic materials.

3) A homogeneous thermostatic system like a polymer solution is

described by a free energy density f. The system plus field is

described by the free energy density f 

f  !E-E . Then using the

definitions of the global chemical potential µi*:=df*/dxi and

xi:=volume ratio the PDE-system hold

°-D(xi(r), r)
4Œ!E
0µ i
(x (r), -E(r))
0
(93)
0r i

For a magnetic system (for instance a ferrofluid solution) the

electric variables (E,D) are replaced by magnetic ones (H,B). The

magnetic charge density ! is set to zero, because no magnetic

charges can be detected during the magnetization, see [62].

4) If the problem depends from time additionally, it is necessary

to replace the second equation of (93) by the thermodynamic

functions for non-equilibrium. Then one can write

39
 °-D(xi(r), r)
4Œ!E

0ni(r) Z en (r)E(r) (94)

ji
Di  i i i
0r RT

Here hold the definitions n:=concentration, r:=space coordinate

Z:=number of charges per ion, e:=elementary charge, E:= electric

field, R:=Avogadro-constant, T:=temperature, D:=diffusion

constant, :=mobility. The second equation of (94) is the Nernst-

Planck equation, which should coincide with the second equation

(93) for j=0. So electrochemical problems are discussed, cf.[63].

5) In semiconductors the charge densities depend from chemical

potential or quasi-Fermi level, which can be influenced by the

electric potential. A good example for such a system is a InAs-

quantum dot-doted FET invented by Yusa&Sakaki [64]. Its structure

is shown in fig.7. The FET can be used for storing data by

charging the gate capacitance.

The theoretical model of this FET stems from Rack et al.[65].

The PDE´s of the system is:

Poisson equation: 00 0z[0(z).0z-(z)]

!(z) 
with !(z)
e[N D (z) n 3d(z) n QD(z)]
1
current: 0t n(z)
0 j(z) f(nQD(z,t),n(z))
0 (95)
e z
recombinations: 0tn QD(z,t)
f(nQD(z,t),n(z))

Here are 00 := dielectric constant of vacuum, 0:=dielectric

constant of the material, !:=charge density, ND:=density of
donators, n3d:=charge density of electrons, nQD:=charge density of

40
fig.7: structure of a InAs-quantum dot-doted GaAs-FET
a two-dimensional electron gas (2DEG) is located in the boundary between AlGaAs and GaAs. It
represents the zero potential of the system. The electric potential is applied to the Al layer,cf. figs.9

electron trapped im quantum dots, n(z):=free electron density

function specified in the article, j:=current in the FET, and

f(nQD,n) is a specific function, which characterizes the

recombination process, see [65]. Figs.8 show the electron density

in the 2DEG versus voltage. Remarkable is the orientation of the

electric cycle which is opposite to the ferroelectric loss

hysteresis. This suggests a “gain hysteresis”.

It is known that electric work can be changed to mechanic work

fig.8a the experiment of Yusa-Sakaki- cf. [64] fig.8b the theoretical calculation of the
hysteresis of a InAs-quantum dot-doted FET Yusa-Sakaki-FET by Rack et al.
electron charge density of the two-dimensional electron charge density of the two-dimensional
electron gas (2DEG) vs. gate voltage electron gas (2DEG) vs. gate voltage

41
with efficiencies until 100% in the best electromotors. So

electric work should be equivalent to mechanical work in a

thermodynamical sense. An “isothermically” proceeded electric

cycle with an orientation like in fig.8a can fulfill the energy

balance only if heat flows in from outside. Thus, the FET is a

candidate for second law violation because only heat and

electricity can be exchanged. According to own recent work [62]

such cycles could be possible and further evidence can be found:

Cooling effects in semiconductors have been predicted by [66].

These considerations support the considerations for the FET

discussed above. According to [66] the FET is cooled down if it

is set under voltage. So the electrons are enforced into the

quantum dots below the quasi-Fermi niveau, where they stick due

to their binding energy. After the electric discharge of the FET-

capacitance the FET goes back to the equilibrium either if the

voltage is slightly inverted, cf.fig.8a, either if the wavelength

of the thermal radiation is suffiently high to overcome the

binding energy of 0.25eV, which holds the electrons in the

quantum dot potentials. So, the system can be regarded also as

a concretisation of Maxwell´s demon. The electric energy is

lended probably from the quantum dots to be paid back after some

time from the thermic influx of environment. Further evidence for

this idea can be found from the results of fig. 9a-c, which show

the conduction band edge (which is here equivalent to the

potential) in the FET at the beginning of the cycle, after

charging it with voltage, and after discharging the capacitance.

42
fig.9a the conduction band edge vs. position in the FET of Yusa&Sakaki
before the cycle: voltage U=0 V

fig.9b the conduction band edge vs. position in the FET of Yusa&Sakaki
in the cycle: voltage U=0.9 V

fig.9c the conduction band edge vs. position in the FET of Yusa&Sakaki
after the cycle: voltage U=0 V
the band edge is changed due to the storage of charges in the quantum dots, cf. fig.9a

43
From the slope in the diagrams one calculates the electric fields

in the FET. If one regards the FET as a capacitance and applies

(69) one can estimate the energy exchanged after a cycle. From

(69) follows for a pure capacitance

T
ûW
PU. I dt
1 PP E dD dV (96)
4Œ
0

If one reads off electric field values from the slopes in fig.

9a to fig. 9c one obtains the electric field energies in the FET:

before charging the gate capacitance

W1  E2*V  (1V/600nm)2 *600nm = 0.00166666

after discharging the gate capacitance

W2  E2i*Vi (.38V/200nm)2*200nm +(.62V/400nm)2*0.400nm = 0.001683

energy balance

ûW  -(W2-W1)  -0.00001633
The energy difference of 1% is negative meaning that electric
energy is released by the FET after the electric cycle is closed.

The Second Law is violated by the hysteresis of the equilibrium

state. The effect is due to the nonlinear behaviour of the FET.

Of course, all evidence of the experiment with the Yusa-Sakaki

FET is indirectly concluded here. More decisive would be a full

balance of all electrons in the calculation or the experiment.

Herewith, the constitutive equations are characterized from the

simple case to the most complicated systems. Generally, the

description of a system may be very sophisticated. However,

normally the description is made as simple as possible.

44
fig.10: strength of the --component of the electric vortex field +of a rotating magnetic ring
cross section view: ring radius 1m, ring width 5cm, ring heigth 12cm, center of rotation is to the
left, (not to be seen in picture). rainbow scale: blue is minus min., red is plus max., see appendix 4.

3. Conclusions

It has been shown that the existence of magnetic charges is

justified at least theoretically especially if fields of

permanent magnetism are described.

This result suggests the following consequences to be proved:

If magnetic charges can be separated in space - for instance in

the form of charges of polarisation in a permanent magnet - and

if this magnet moves in a circle, two opposite magnetic currents

are generated which itself should generate a electric field

according to Faraday´s law extended for magnetic charge currents.

Measurement of the electric field from moving permanent magnets

45
fig.11a: E-field strength around a rotating magnetic ring (cross section)
ring radius 1m, ring width 5cm, ring heigth 12cm, center of rotation is to the left, (not shown in the
picture). Arbitrary units. Picture is calculated from the data of fig.10,see appendix 4.

fig.11b position angle of the E-field around a rotating magnetic ring (cross section)
radius 1m,ring width 5cm,ring heigth 12cm,center of rotation is to the left,(not shown in the picture)

46
can answer the question whether the field stems from currents of

moving magnetic charges or from the dB/dt - term. According to

the theory above both possibilities can be calculated.

The setup of such experiments would be similar to constructions

which are known in the unofficial subscene of physics. J.

Searl[67-70], D. Hamel [71] and Godin&Roschin [72, 73] claim,

that they have observed strong electrostatic effects around fast

moving permanent magnets.

Fig.10 shows the calculated electrical vortex field + due to a

moving permanent magnet ring representing two currents of

opposite magnetic surface charge which are placed on top and

bottom of the ring, cf. fig.4. The electric field is calculated

from the electric vortex field by E = rot +, see fig.11a and fig.
11b: the pictures show the electric field strength and the

position angle of the field around the cross section of the right

half of the ring. In appendix 4 the method of the calculation is

shown. If one compares the order of magnitude of the calculation

with the data of Godin&Roschin [72, 73] then this suggests for

the electric field values an agreement between theory and

experiment.

47
Appendix 1: the derivation of the multipole expansion

First the term 1/|x-x´| is written as:

1 1


1 1
|x x´| x 2x´ 2 2x.x´ |x| x´ 2 2x.x´
1
|x|2

with the abbreviation . :

(x´ 2 2x.x´)/|x|2 « 1 .

This expression is expanded in a series

3 x´ 2 2x.x´
2

1 .  3 .2 ± .....
1 1 x´2  2 x.x´
2
1
 ± ....
1. 2 8 2 |x| 2 |x|2 8 |x|2

Using the definitions x0:=x/|x| and |x|:=r one obtains

1

1  12 (x .x0)  13 [ 3 (x .x0)2 1 x 2]  O( 14 )
  

|x x | r

r r 2 2 r

If this result is applied to the potential definition one gets

x0 x0
-(x)
1
P
!d 3x´  12 x0 !x´d 3x´  i 3 j ![3x´ix´j x´n x´n/ij]d 3x´  O( 14 )
r r P 2r P r

This can be written as well

Qij.x0 x0
-(x)
q
r

p.x0
2
 i

3
j
 O( 14 )
r 2r r

using the definitions

q:
!(x )d 3x  p :
x !(x )d 3x  Q ij :
(3x´ i x´ j x´n x´n/ij)!(x )d 3x 
P P P

48
Appendix 2: decomposition of a general vector field into a

potential field and a vortex field

Theorem 1:

The derivative of a vector field F can be decomposed in a

symmetric (index=C) and a antisymmetric part (index=V), i.e.

0Fi 0Fi V 0Fi C


0xj 0xj 0xj

FC is the symmetric part and is a gradient of a potential field

0Fi C 0Fj C 0U(xl)

or rot FC
0 with Fl
C

0xj 0xi 0xl

(with U(xl):=potential function)

FV is a antisymmetric vortex field

0Fi V 0Fj V


0xj 0xi

Proof:

The derivatives of the field F can be decomposed according to

0Fi 1 0Fi 0Fj 0Fi 0Fj

  1
0xj 2 0xj 0x i 2 0xj 0x i

for the symmetric part holds:

49
 0Fi 0Fj 1 0Fi 0Fi
C V
0Fj C 0Fj V 0Fi C

0U
2
1

  

2 0xj 0xi 2 0x j 0xj 0x i 0xi 0xj 0x i0xj

for the antisymmetric part holds:

1 0Fi 0Fj 1 1 0Fi 0Fi

C V
0Fj C 0Fj V 0Fi V

rotF


2 0xj 0xi 2 2 0xj 0x j 0xi 0x i 0xj

It can be checked, that

0Fi 0Fi V 0Fi C


0xj 0xj 0xj

q.e.d.

Theorem 2:

F is a field with a defined boundary condition 0F around the

space which is interesting for the problem. Divergence and

rotation are defined according to

/×F
j(x) /.F
!(x)

and the boundary condition 0F

0F: F.n
f(r)

Then it holds:

50
F can be calculated as sum of a gradient FC of a potential, plus

rotation of a vector potential FV , plus a Laplace field FL

according to

F :
F C  F V  F L
1 !(x )(x x ) 3
FC

4ΠP |x x |3
1
d x´
/
!(x ) 3
4ΠP |x x |
1
d x´
/
4Œ
-
1 j(x )×(x x ) 3 1 j(x ) 3 1
FV
d x´
/× d x´
/×A
4ΠP |x x |  3 4ΠP |x x | 
4Œ
FL
/3

It holds:

/×FC
0 /.FC
!(x)
/.FV
0 /×FV
j(x)
/.FL
û3
0

scheme of the proof [26]:

1) It is looked for the solution of

/×FC
0 /.FC
!(x)

This is the potential field

1 !(x )(x x ) 3 1 !(x ) 3

FC
d x´
/ d x´
4ΠP |x x |3 4ΠP |x x |

2) It is looked for the solution of

/.FV
0 /×FV
j(x)

This is the vortex field

51
 1 j(x )×(x x ) 3 1 j(x ) 3
FV
d x´
/× d x´
4ΠP |x x |3 4ΠP |x x |

3) It is looked for

/.FL
0 /×FL
0

using the boundary condition

FL.n
F.n FC.n FV.n

The solution is the Laplace field

/.FL
û3
0

4) The general solution is the sum of 1) - 3). This can be

checked using the vector relations divrot A=0 and rotgrad -=0 .
So one obtains

F
FC  FV  FL

q.e.d

The Laplace field is a “generalized constant of integration”. It

allows to adapt to the boundary conditions. It is needed, if

boundary conditions for F exist which are non-zero in the

infinite, see fig.3.

52
Appendix 3: Derivation of the Lorenz gauge

The continuity equation is

dNE
div j(x´)  (x´)
0
dt

It can be written as

div j(x´)  N E(x´)

dx´3
0
|x x´|

The divergence term is changed using partial integration. One

term can be canceled during partial integration, because j(x’)=0

holds for x’=  . So it is obtained

N (x´)
j(x´)/´ 1  E dx´3
0
|x x´| |x x´|

With / |x x´| 1
/´|x x´| 1 one yields

N (x´)
/ j(x´)  E dx´3
0
|x x´| |x x´|

This is the Lorenz gauge

0-E
/.AH 1
0
c 0t

53
Appendix 4:

Model calculation 1:

In order to estimate the field, which could be generated by a

magnetic current, we calculate here the non-real case of a coil

of one turn which is driven by a current of magnetic charge. The

coil is modelled as a rotating tube which is charged with

magnetic surface charges.

We use the formulas of magnetostatics applied for magnetic

currents (in SI-units) by exchanging the magnetic variables by

the analogous electric variables.

Datas of the setup:

1 magnetic tube charged with magnetic charges
diameter: d= 2m
height: h= 10cm
number of turns: n= 1
magnetic field strength at the surface: B0 = 1T = 1 Vs/m2
magnetic permeability: µ = 10001
speed of rotation: f = 10Hz.

Using this data the magnetic current IH can be calculated to

IH = surface charge * speed of rotation = (µ-1)*B0*d*Œ*h*f

Then, the electrical field of a magnetic current, cf. (80)

E = IH*n/h = (µ-1)*B0*d*Œ*f =2*Œ*105 V/m

This means: electrical fields generated by magnetic currents

54
should be sufficiently strong to be detected easily. It should

be possible to reach the breakdown voltage of air (30 kV/cm at

1 bar) if the parameters are chosen accordingly high.

Model calculation 2:

We estimate here the field of a magnetic cylinder ring which

turns around its central axis. The upper surface of the ring is

the north pole, the lower the south pole.

Data of the setup:

1 magnetic ring magnet
upper rim: north-, lower rim: south pole
diameter d= 2m
height h= 12cm
width b= 5cm
number of turns n= 1
magnetic field strength at the pole surfaces: B0= 1T
magnetic permeability µ= 10000
speed of rotation: f= 10Hz.

The origin of the coordinate ]

system is the centre of

symmetry on the middle of the

D

central axis. The distribution U


of the electric field lines of
1 \

the setup can be calculated by

,

using a known example and

adapting it for the present [

setup. For a simple ring

current,see fig. 12, Jackson[8] Fig.12: the coordinate system

calculates a vortex vector-field in chapter 5.5, equation 5.37.

The formula transferred to magnetic currents yields

55
 ]

U

1
U D
U Due to geometry it holds:

]
z coordinates: r1.cos1z0
r.cos
2 r2.cos2 z0
r.cos
1 \

] radius projection r1.sin1

r.sin
into x y plane: r2.sin2
r.sin

fig.13 the geometric situation of a field point due to a circulating magnetic dipol
the field is composed from two opposite circulating magnetic currents

4I H a (2 m) K(m) 2E(m) 4arsin()

+-(r,)
with m

m m a 2
r 22arsin()

Here are K(m) and E(m) elliptic integrals of first and second

order, which are calculated numerically by a program.

This formula is applied for two circuits which are shifted by z0

upward and downwards. In both circuits the magnetic current flows

in opposite directions. From the geometry of the setup the

appropriate radii and angles of each circuit can be determined,

see fig. 13 . The system of equations from fig. 13 are solved

r.cos z0
1
arcot  r1
r.sin /sin1
r.sin
r.cosz0
2
arcot  r2
r.sin /sin2
r.sin

56
Then it is possible to write down the electric vortex potential

+ which has only one component in --direction which is

perpendicular to the plane of the paper, see fig.11

+-(r,)
+-(r1,1)  +-(r2,2)

For the purpose of a simple calculation the equally distributed

magnetic charge on the surface was approximated by 11 charges at

equidistant positions on the surfaces each. The calculated

intensity of the --component of the electric vortex field + is

already shown in fig.10. Then, the E-field is calculated

according to E = rot +, fig. 11 a)+b). So the field has been

estimated having a maximum 80 kV/cm near the edges at the surface

of the moving magnet. This seems to coincide with the

observations of Godin&Roschin [72, 73]. They observed a

luminescence and therefore a ionisation of the air near the

surface of moving magnets. However, the prediction of

electrodynamics is contrary to the other reported claims like the

self-acceleration and the weight loss of the setup at higher

angular velocities. Electrodynamics predicts a braking down due

to the Lenz-rule. Any currents generated by the E-field leads to

braking and dissipation as well. Whether electric cycles can be

built up with a gain hysteresis or not has to be proved

separately.

57
Bibliography:

1. Whittaker, E., A History of the Theories of Aether and Electricity - The Classical
Theories. Vol. 1. 1973, New York: Humanities Press.
2. Whittaker, E., A History of the Theories of Aether and Electricity - The Modern
Theories 1900 -1926. Vol. 2. 1973, New York: Humanities Press.
3. Ehrenhaft, F., Über die Photophorese, die wahre magnetische Ladung und die
schraubenförmige Bewegung der Materie in Feldern Erster Teil. Act. phys. austr.,
1951. 4: p. 461.
4. Ehrenhaft, F., Über die Photophorese, die wahre magnetische Ladung und die
schraubenförmige Bewegung der Materie in Feldern Zweiter Teil. Act. phys.
austr., 1952. 5: p. 12.
5. Bauer, W.D., Magnetic Monopoles in Theory and Experiment. 2002. 2002.
URL: http://www.overunity-theory.de/Ehrenhaft/talk.zip
6. Bauer, W.D., Über elektrische und magnetische Aufladungen - Das Werk von
Felix Ehrenhaft, . 2003: Berlin.
URL: http://www.overunity-theory.de/Ehrenhaft/Ehrenhaft.zip
7. Priestley, J., The History and Present State of Electricity, with Original
Experiments. 1767, London.
8. Jackson, J.D., Classical Electrodynamics Second Edition. 2 ed. 1975, New York:
John Wiley & Sons.
9. Risken, H., Vorlesungsskript Elektrodynamik. 1976, Ulm: Uni Ulm Selbstverlag.
10. Oersted, H.C., Versuche über die Wirkung eines elektrischen Stromes auf die
Magnetnadel. Annals of Philosophy, 1820. 16: p. 273.
11. Tricker, R.A.R., Frühe Elektrodynamik. 1974, Braunschweig: Vieweg.
12. Biot, J.B. and F.Savart, Eine Bemerkung zum Magnetismus der Volta'schen
Säule. Ann.Chim.Phys., 1820. 15: p. 222-223.
13. Biot, J.B. and F. Savart, Journal de Physique, 1820. 91: p. 151.
14. Biot, J.B. and F. Savart, Magnetisierung von Metallen mittels sich bewegender
Elektrizität, in Précis Elémentaire de Physique. 1824: Paris. p. 707-723.
15. Grassmann, H., Neue Theorie der Elektrodynamik. Poggendorf's Annalen der
Physik und Chemie, 1845. 64(1): p. 1 - 18.
16. Ampère, A.M., Ann. Chim. Phys., 1820. 15: p. 59-76, 177-208.
17. Ampère, A.M., Mém. de l'Acad., 1825. 6: p. 175.
18. Riemann, B., Schwere, Elektrizität und Magnetismus, nach den Vorlesungen von
B.Riemann. 1875, Hannover.
19. Cavallieri, G., G. Spavieri, and G. Spinelli, The Ampère and Biot-Savart force
laws. Eur. J. Phys., 1996. 17: p. p.205-207.
20. Curé, J.C., Action and reaction in Electrodynamics. Deutsche Physik, 1995.
4(13): p. 5 - 10.
21. Aspden, H., Physics without Einstein. 1969, Southampton: Sabberton.
22. Marinov, S., Marinov's formula is the only viable formula in magnetism. Deutsche
Physik, 1994. 3(11): p. 18 - 34.
23. Cavallieri, G., et al., Experimental proof of standard electrodynamics by
measuring the self-force on a part of a current loop. Phys. Rev. E, 1998. 58(2):

58
 p. 2502 - 2517.
24. Neumann, F.v., Abhandlungen I. 1845, Berlin.
25. Neumann, F.v., Abhandlungen II. 1848, Berlin.
26. Bronstein, I.N. and K.A. Semendjajew, Taschenbuch der Mathematik. 20 ed.
1983, Thun: Harri Deutsch.
27. Thomson, J.J., Phil. Mag., 1899. 48(5): p. 547.
28. Graneau, P., Comment on "The motionally induced back EMF in railguns". Phys.
Lett. A, 1991. 160: p. 490 - 491.
29. Graneau, P. and N. Graneau, Newtonian Electrodynamics. 1995, New York:
World Scientific Publishing Co.
30. Tanberg, R., On the cathode of an arc drawn in vacuum. Phys. Rev, 1929. 35:
p. 1080.
31. Kobel, E., Pressure and high velocity vapour jets at cathodes of a mercury
vacuum arc. Phys. Rev., 1930. 36: p. 1636.
32. Johansson, L., Longitudinal electrodynamic forces - and their possible
technological application, in Department of Electromagnetic Theory. 1996, Lund
Institute of Technology P.O. Box 118 S-22100 Schweden: Lund.
URL: http://www.df.lth.se/~snorkelf/LongitudinalMSc.pdf
33. Rambaut, M. and J.P. Vigier, Ampère forces considered as collective non-
relativistic limit of the sum of all Lorentz interactions acting on individual current
elements: possible consequences for electromagnetic discharge stability and
tokamak behaviour.
Phys. Lett. A, 1990. 148(5): p. 229-238.
34. Beauregard, O.C.d., Statics of filaments and magnetostatics of currents:
Ampère tension and the vector potential. Phys. Lett. A, 1993. 183: p. 41 - 42.
35. Duschek, A. and A. Hochrainer, Grundzüge der Tensorrechnung in analytischer
Darstellung 2. Tensoranalysis. Vol. 2. 1950, Berlin: Springer-Verlag.
36. Duschek, A. and A. Hochrainer, Grundzüge der Tensorrechnung in analytischer
Darstellung 1. Tensoralgebra. 3. Auflage ed. Vol. 1. 1954, Berlin: Springer-
Verlag.
37. Mikhailov, V.F. and L.I. Mikhailova, Preprint HEPI-82-01, . 1982, Academy of
Sciences of Kazakh SSR: Alma Ata.
38. Mikhailov, V.F., The Magnetic Charge Phenomen on Ferromagnetic Aerosols.
Phys. Lett. B, 1983. 130: p. 331.
39. Mikhailov, V.F., J. Phys. A: Math. Gen., 1985. 18: p. L903.
40. Mikhailov, V.F., Ann. Fond. Louis de Broglie, 1987. 12: p. 491.
41. Mikhailov, V.F. and L.I. Mikhailova, On some regularities of aerosol particle
motion in electromagnetic fields, . 1987, Kazakh Akademy of Sciences, USSR:
Alma Ata.
42. Mikhailov, V.F., Preprint HEPI-88-17, . 1988, Academy of Sciences of Kazakh
SSR: Alma Ata.
43. Mikhailov, V.F., Preprint HEPI-88-05, . 1988, Academy of Sciences of Kazakh
SSR: Alma Ata.
44. Mikhailov, V.F. and L.I. Mikhailova, J. Phys. A: Math. Gen., 1989. 23: p. 53 - 63.
45. Mikhailov, V.F. and L.I.Mikhailova, Preprint HEPI-90-07, . 1990, Academy of
Sciences of Kazakh SSR: Alma Ata.
46. Mikhailov, V.F. and L.I.Mikhailova, Preprint HEPI-91-04, . 1991, Academy of

59
 Sciences of Kazakh SSR: Alma Ata.
47. Mikhailov, V.F., Observation of apparent magnetic charges carried by
ferromagnetic particles in water droplets. J. Phys. A: Math. Gen., 1991. 24: p. 53
- 57.
48. Mikhailov, V.F., Experimental detection of Dirac's magnetic charge ? J. Phys. D:
Appl. Phys., 1996. 29: p. 801 - 804.
49. Tricker, R.A.R., Faraday und Maxwell. 1974, Berlin: Akademie-Verlag.
50. Obukhov, Y.N. and F.W. Hehl, Electromagnetic energy-momentum and forces
in matter. Phys. Lett. A, 2003. 311: p. 277 - 284.
51. Rainich, G.Y., Electrodynamics in the General Relativity Theory. Trans. Am.
Math. Soc., 1925. 27(1): p. 106 - 136.
52. Katz, E., Concerning the Number of Independent Variables of the Classical
Electromagnetic Field. Am. J. Phys., 1965. 33: p. 306 - 312.
53. Harrison, H., et al., Possibility of Observing the Magnetic Charge of an Electron.
Am. J. Phys., 1963. 31: p. 249.
54. Akers, D., Mikhailov´s Experiments on Detection of Magnetic Charge. Int. J.
Theor. Phys., 1988. 27(8): p. 1019 - 1022.
55. Dirac, P.A.M., Quantizised Singularities in the Electromagnetic Field. Proc. R.
Soc. London A, 1931. 133: p. 60.
56. Dirac, P.A., The Theory of Magnetic Poles. Phys. Rev., 1948. 74(7): p. 830.
57. Landau, L.D. and E.M. Lifshitz, Lehrbuch der theoretischen Physik -
Elektrodynamik der Kontinua. 5 ed. Vol. 8. 1990, Berlin: Akademie-Verlag.
58. O´Dell, T.H., The Electrodynamics of Magneto-electric Media. Series of
monographs on selected topics in solid state physics. 1970, London: North-
Holland Amsterdam.
59. Penfield, P. and H. Haus, Electrodynamics of moving media. 1967, Cambridge
Mass.
60. Hubert, A. and R.Schäfer, Magnetic Domains. 2001, Berlin,Heidelberg: Springer-
Verlag.
61. Ternovsky, V., B. Luk'yanchuk, and J.P. Wang, Remanent States of Small
Ferromagnetic Cylinder. JETP Letters, 2001. 73(12): p. 661 - 665.
62. Bauer, W.D., Second Law versus Variation Principles, . 2001.
URL: http://xxx.lanl.gov/pdf/physics/0009016.pdf
63. Christoph, J., Musterbildung auf Elektrodenoberflächen, in Chemie, FU Berlin:
Fachbereich Chemie, Berlin., 2000
64. Yusa, G. and H. Sakaki, Trapping of photogenerated carriers by InAs quantum
dots and persistent photoconductivity in novel GaAs/n-AlGaAs field-effect
transistor structures. Appl. Phys. Lett., 1997. 70(3): p. 345 - 347.
65. Rack, A., et al., Dynamic bistability of quantum dot structures: Role of the Auger
process. Phys.Rev. B, 2002. 66: p. 165429.
see as well URL: http://www.hmi.de/people/rack/diplom/index.html
66. Rego, L.G.C. and G. Kirczenow, Electrostatic mechanism for semiconductor
heterostructures. Appl. Phys. Lett., 1999. 75(15): p. 2262 -2264.
67. Ehlers, H.J., raum&zeit-Interview mit John Roy Robert Searl, in Raum&Zeit
special, H.J. Ehlers, Editor. 1994, Ehlers-Verlag: Sauerlach. p. 148 - 149.
68. Sandberg, S.G., Der Searl-Effekt und der Searl-Generator,
in Raum&Zeit special 7 Wunschtraum der Menschheit: Freie Energie,

60
 H.J. Ehlers, Editor. 1994, Ehlers Verlag: Sauerlach. p. 149 - 157.
see as well URL: http://www.rexresearch.com
69. Schneider, H. and H. Watt, Dem Searl-Effekt auf der Spur (I), in Raum&Zeit
special, H.J. Ehlers, Editor. 1994, Ehlers Verlag: Sauerlach. p. 174 - 180.
70. Schneider, H. and H. Watt, Dem Searl-Effekt auf der Spur (II), in Raum&Zeit
special, H.J. Ehlers, Editor. 1994, Ehlers-Verlag: Sauerlach. p. 181 - 185.
71. Manning, J. and P. Sinclair, The Granite Man and the Butterfly. 1995, Fort
Langley, B.C. , Canada: Project Magnet Inc.
see as well URL: http://jnaudin.free.fr/html/hammnu.htm
72. Roschin, V.V. and S.M. Godin, An Experimental Investigation of the Physical
Effects in a Dynamic Magnetic System. Technical Physics Letters, 2000. 26(12):
p. 1105 - 1107.
see as well URL: http://www.rexresearch.com/roschin/roschin.htm
73. Roschin, V.V. and S.M. Godin,
Experimental Research of the Magnetic-Gravity Effects, . 2000.
see as well URL: http://alexfrolov.narod.ru/russearl.html

61