Spin-Curvature and The Unification of Fields in A Twisted Space

INDRANU SUHENDRO
SPIN-CURVATURE
AND THE UNIFICATION
OF FIELDS
IN A TWISTED SPACE S
V
E
N
S
K
A
F
Y
S
I
K
A
R
K
I
V
E
T
2
0
0
8
Indranu Suhendro
Spin-Curvature
and the Unication of Fields
in a Twisted Space
Spinn-krokning
och foreningen av falt
i ett tvistat rum
2008
Swedish physics archive
Svenska fysikarkivet
Svenska fysikarkivet, that means the Swedish physics archive, is a publisher regis-
tered with the National Library of Sweden (Kungliga biblioteket), Stockholm.
Please send all comments on this book or requests to reprint to the Author.
E-mails: wings.of.solitude@gmail.com; spherical symmetry@yahoo.com
Edited by Stephen J. Crothers and Dmitri Rabounski
Copyright c Indranu Suhendro, 2008
Copyright c Typesetting and design by Dmitri Rabounski, 2008
Copyright c Publication by Svenska fysikarkivet, 2008
Copyright Agreement: All rights reserved. The Author does hereby grant Sven-
ska fysikarkivet non-exclusive, worldwide, royalty-free license to publish and dis-
tribute this book in accordance with the Budapest Open Initiative: this means that
electronic copying, print copying and distribution of this book for non-commercial,
academic or individual use can be made by any user without permission or charge.
Any part of this book being cited or used howsoever in other publications must ac-
knowledge this publication. No part of this book may be reproduced in any form
whatsoever (including storage in any media) for commercial use without the prior
permission of the copyright holder. Requests for permission to reproduce any part of
this book for commercial use must be addressed to the Author. The Author retains
his rights to use this book as a whole or any part of it in any other publications and
in any way he sees t. This Copyright Agreement among the Author and Svenska
fysikarkivet shall remain valid even if the Author transfers copyright of the book
to another party. The Author hereby agrees to indemnify and hold harmless Sven-
ska fysikarkivet for any third party claims whatsoever and howsoever made against
Svenska fysikarkivet concerning authorship or publication of the book.
Cover photo: Spiral Galaxy M74. This image is a courtesy of the Hubble Space
Telescope Science Institute (STScI) and NASA. This image is a public domain prod-
uct; see http://hubblesite.org/copyright for details. We are thankful to STScI and
NASA for the image. Acknowledgement: R. Chandar (University of Toledo) and
J. Miller (University of Michigan).
This book was typeset using teT
E
X typesetting system and Kile, a T
E
X/L
A
T
E
X editor
for the KDE desktop. Powered by Ubuntu Linux.
ISBN: 978-91-85917-01-3
Printed in Sweden and the USA
Contents
Preface of the Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Foreword. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Chapter 1 A four-manifold possessing an internal spin
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2 Geometric construction of a mixed, metric-compatible
four-manifold R
4
possessing an internal spin space S
p
. . 11
Chapter 2 The unified field theory
2.1 Generalization of Kaluzas projective theory. . . . . . . . . . . . 20
2.2 Fundamental eld equations of our unied eld theory.
Geometrization of matter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Chapter 3 Spin-curvature
3.1 Dynamics in the microscopic limit. . . . . . . . . . . . . . . . . . . . . . 44
3.2 Spin-curvature tensor of S
p
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3 Wave equation describing the geometry of R
4
. . . . . . . . . . 48
Chapter 4 Additional considerations
4.1 Embedding of generalized Riemannian manifolds
(with twist) in N =n+p dimensions . . . . . . . . . . . . . . . . . . . 53
4.2 Formulation of our gravoelectrodynamics by means
of the theory of distributions. Massive quantum elec-
tromagnetic eld tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.3 On the conservation of currents . . . . . . . . . . . . . . . . . . . . . . . . 64
4.4 On the wave equations of our unied eld theory. . . . . . . 65
4.5 A more compact form of the generalized Gauss-
Codazzi equations in IR
5
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Appendix The fundamental geometric properties of a curved
manifold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Preface of the Editor
I am honoured to present this book by Indranu Suhendro, in which he
adduces theoretical solutions to the problems of spin-curvature and the
unication of elds in a twisted space. A twist of space is given herein
through the appropriate formalism, and is related to the anti-symmetric
metric tensor. Kaluzas theory is extended and given an appropriate
integrability condition. Both matter and the isotropic electromagnetic
eld are geometrized through common eld equations: trace-free eld
equations, giving the energy-momentum tensor for the electromagnetic
eld via only the generalized Ricci curvature tensor and scalar, are
obtained. In the absence of the electromagnetic eld the theory goes to
Einsteins 1928 theory of distant parallelism where only the matter eld
is geometrized (through the twist of space-time). Therefore the above
results, in common with respective wave equations, are joined into a
unied theory of semi-classical gravoelectrodynamics.
There have been few attempts to introduce spin-particles into the
theory of relativity (which is the theory of elds, in the sense pro-
pounded by Landau and Lifshitz). Frankly speaking, only two of the
attempts were complete. The rst attempt, by A. Papapetrou (1951),
was a frontal approach to this problem: he introduced a spin-particle
as a swiftly rotating gyro (Proc. Roy. Soc. A, 1951, v. 209, 248258 and
259268). This approach however doesnt match experimental data due
to that fact that, considering an electron as a solid ball, the linear
velocity of its rotation at its surface should be 70 times greater than
the velocity of light. The second attempt, by me and L. Borissova
(2001), introduced a spin-particle through the variational principle and
Lagranges function for such a system (see Chapter 4 of the book Fields,
Vacuum, and the Mirror Universe, Editorial URSS, Moscow, 2001). De-
spite some success related to its immediate application to the theory of
elementary particles, this method however could not be considered as
purely geometric: spin, the fundamental property of a particle, wasnt
expressed through the geometric properties of the basic space, but in-
stead still remained a non-geometrized fundamental characteristic of
matter.
Insofar as the geometrization of distributed matter is concerned,
just one solution was successful before this book. It was given for
an isotropic electromagnetic eld. Such elds are geometrized via the
Preface of the Editor 5
well-known Rainich condition and the Nortvedt-Pagels condition: the
energy-momentum tensor of such a eld expresses itself through the
components of the fundamental metric tensor, so Einsteins equations
contain only the geometric left-hand-side by moving all the right-
hand-side terms to the left-hand-side so the right-hand-side becomes
zero. Various solutions given by the other authors are particular to the
Rainich/Nortvedt-Pagels condition, or express an electromagnetic eld
under a very particular condition.
Both problems are successfully resolved in this book with the use
of only geometric concepts on a common basis (a twisted space), which
is the great advantage of this work. This fact places this book in the
same class of great ideas in the theory of elds, produced during the last
century, commencing with the time Albert Einstein rst formulated his
eld equations.
I am therefore very pleased to bring this book to the reader. I recom-
mend it to anyone who is seriously interested in the Theory of Relativity
and the geometric approach to physics in particular.
February, 2008 Dmitri Rabounski
Foreword
In the present research, I consider a unication of gravity and electro-
magnetism in which electromagnetic interaction is seen to produce a
gravitational eld. The eld equations of gravity and electromagnetism
are therefore completely determined by the fundamental electromag-
netic laws. Insight into this unication is that although gravity and
electromagnetism have dierent physical characteristics (e.g., they dif-
fer in strength), it can be shown through the algebraic properties of the
curvature and the electromagnetic eld tensors that they are just dif
ferent aspects of the geometry of space-time. Another hint comes from
the known speed of interaction of gravity and electromagnetism: elec-
tromagnetic and gravitational waves both travel at the speed of light.
This means that they must somehow obey the same wave equation.
This indeed is unity. Consequently, many dierent gravitational and
electromagnetic phenomena may be described by a single wave equation
reminiscent of the scalar Klein-Gordon equation in quantum mechan-
ics. Light is understood to be a gravoelectromagnetic wave generated
by a current-producing oscillating charge. The charge itself is gener-
ated by the torsion of space-time. This electric (or more generally,
electric-magnetic) charge in turn is responsible for the creation of mat-
ter, hence also the transformation of matter into energy and vice versa.
Externally, the gravitational eld manifests itself as the nal outcome
of the entire process. Hence gravity and electromagnetism obey the
same set of eld equations, i.e., they derive from a common origin. As
a result, the charge produces the so-called gravitational mass. Albeit
the geometric non-linearity of gravity, the linearity of electromagnetism
is undisturbed: an idea which is central also in quantum mechanics.
Therefore I preserve the most basic properties of matter such as energy,
momentum, mass, charge and spin through this linearity. It is a mod-
est attempt to once again achieve a comprehensive unication which
explains that gravity, electromagnetism, matter and light are only dif-
ferent aspects of a single theory.
Karlstad, Sweden, 2004 Indranu Suhendro
Acknowledgements
This work was originally intended to be my Ph.D. thesis in 2003. How-
ever, life has led me through its own existential twists and turns. In
the process of publishing this book, many individuals have made their
splendid contributions. It would be rather dicult to give credit with-
out overlooking some of these sincere, wonderful persons. I would like
especially to single out my dear colleagues, Dmitri Rabounski, Stephen
Crothers, and Augustina Budai, who most passionately and sincerely
helped prepare the entire manuscript in the present form. I would also
like to express my most sincere gratitude to Bengt and Colette Johnsson,
who gave me shelter and showered me with tenderness during my most
creative years in Sweden. I would also like to thank my dear friends,
Barbara

Smilgin, Dima Shaheen, Quinton Westrich, Liw Raskog, Vi-
taliy Kazymyrovych, and Sana Raq, with whom I share the subtle
passions that normally lie hidden in the solitary, silent depths of life.
Amidst lifes high tides and despite all human frailties, in our togeth-
erness, I believe, we have done enough to motivate each other to reach
the height of existence. I must also recognize the most beautiful love,
patience, and understanding of my father, mother, and the rest of my
family throughout my life of seasons. Finally, this little scientic cre-
ation is dedicated with innite love, yearning, gratitude, and humility
to A. S. and M. H. for always being there for me somewhere between
my slumber and wakefulness in this sojourn called life.
February, 2008 Indranu Suhendro
Chapter 1
A FOUR-MANIFOLD POSSESSING
AN INTERNAL SPIN
1.1 Introduction
Attempts at a consistent unied eld theory of the classical elds of
gravitation and electromagnetism and perhaps also chromodynamics
have been made by many great past authors since the eld concept
itself was introduced by the highly original physicist, M. Faraday in
the 19th century. These attempts temporarily ended in the 1950s:
in fact Einsteins denitive version of his unied eld theory as well
as other parallel constructions never comprehensively and compellingly
shed new light on the relation between gravity and electromagnetism.
Indeed, they were biased by various possible ways of constructing a uni-
ed eld theory via dierent geometric approaches and interpretations
of the basic geometric quantities to represent the eld tensors, e.g., the
electromagnetic eld tensor in addition to the gravitational eld ten-
sor (for further modern reference of such attempts, especially the last
version of Einsteins gravoelectrodynamics see, e.g., various works of
S. Antoci). This is put nicely in the words of Infeld: . . . the problem of
generalizing the theory of relativity cannot be solved along a purely for-
mal way. At rst, one does not see how a choice can be made among the
various non-Riemannian geometries providing us with the gravitational
and Maxwells equations. The proper world-geometry which should lead
to a unied theory of gravitation and electricity can only be found by
an investigation of its physical content. In my view, one way to justify
whether a unied eld theory of gravitation and electromagnetism is
really true (comprehensive) or really refers to physical reality is to
see if one can derive the equation of motion of a charged particle, i.e.,
the (generalized) Lorentz equation, if necessary, eortlessly or directly
from the basic assumptions of the theory. It is also important to be able
to show that while gravity is in general non-linear, electromagnetism is
linear. At last, it is always our modest aim to prove that gravity and
electromagnetism derive from a common source. In view of this one
must be able to show that the electromagnetic eld is the sole ingre-
dient responsible for the creation of matter which in turn generates a
gravitational eld. Hence the two elds are inseparable.
Chapter 1 A Four-Manifold Possessing an Internal Spin 9
Furthermore, I nd that most of the past theories were based on the
Lagrangian formulation which despite its versatility and exibility may
also cause some uneasiness due to the often excessive freedom of choos-
ing the eld Lagrangian. This strictly formal action-method looks like
a short cut which does not lead along the direct route of true physical
progress. In the present work we shall follow a more fundamental (nat-
ural) method and at the same time bring up again many useful classical
ideas such as the notion of a mixed geometry and Kaluzas cylinder con-
dition and ve-dimensional formulation. Concerning higher-dimensional
formulations of unied eld theories, we must remember that there is
always a stage in physics at which direct but narrow physical arguments
can hardly impinge upon many hidden properties of Nature. In fact the
use of projective geometries also has deep physical reason and displays a
certain degree of freedom of creativity: in this sense science is an art, a
creative art. But this should never exclude the elements of mathemati-
cal simplicity so as to provide us with the very conditions that Natures
manifest four-dimensional laws of physics seemingly take.
For instance, Kaluzas cylinder condition certainly meets such a re-
quirement and as far as we speak of the physical evidence (i.e., there
should possibly be no intrusion of a particular dependence upon the
higher dimension(s)), such a notion must be regarded as important
if not necessary. An arbitrary ane (n+1)-space can be represented
by a projective n-space. Such a pure higher-dimensional mathematical
space should not strictly be regarded as representing a real higher-
dimensional world space. Physically saying, in our case, the ve-
dimensional space only serves as a mathematical device to represent
the events of the ordinary four-dimensional space-time by a collection
of congruence curves. It in no way points to the factual, exact number
of dimensions of the Universe with respect to which the physical four-
dimensional world is only a sub-space. In this work we shall employ a
ve-dimensional background space simply for the sake of convenience
and simplicity.
On the microscopic scales, as we know, matter and space-time itself
appear to be discontinuous. Furthermore, matter arguably consists of
molecules, atoms and smaller elements. A physical theory based on a
continuous eld may well describe pieces of matter which are so large in
comparison with these elementary particles, but fail to describe their be-
havior. This means that the motion of individual atoms and molecules
remains unexplained by physical theories other than quantum theory in
which discrete representations and a full concept of the so-called mate-
rial wave are taken into account. I am convinced, indeed, that if we had
10 I. Suhendro Spin-Curvature and the Unification of Fields
a sucient knowledge of the behavior of matter in the microcosmos, it
would, and it should, be possible to calculate the way in which matter
behaves in the macrocosmos by utilizing certain appropriate statistical
techniques as in quantum mechanics. Unfortunately, such calculations
prove to be extremely dicult in practice and only the simplest systems
can be studied this way. Whats more, we still have to make a number of
approximations to obtain some real results. Our classical eld theories
alone can only deal with the behavior of elementary particles in some
average sense. Perhaps we must humbly admit that our understanding
and knowledge of the behavior of matter, as well as space-time which
occupies it, is still in a way almost entirely based on observations and
experimental tests of their behavior on the large scales. This is a matter
for experimental determination but a theoretical framework is always
worth constructing. As generally accepted, at this point one must aban-
don the concept of the continuous representation of physical elds which
ignores the discrete nature of both space-time and matter although it
doesnt always treat matter as uniformly distributed throughout the re-
gions of space. Current research has centered on quantum gravity since
the departure of the 1950s but we must also acknowledge the fact that
a logically consistent unication of classical elds is still important. In
fact we do not touch upon the formal, i.e., standard construction of a
quantum gravity theory here. We derive a wave equation carrying the
information of the quantum geometry of the curved four-dimensional
space-time in Chapter 4 by rst assuming the discreteness of the space-
time manifold on the microscopic scales in order to represent the pos-
sible inter-atomic spacings down to the order of Plancks characteristic
length.
Einstein-Riemann space(-time) R
4
(a mixed, four-dimensional one)
endowed with an internal spin space S
p
is rst considered. We stick
to the concept of metricity and do not depart considerably from ane-
metric geometry. Later on, a ve-dimensional general background space
IR
5
is introduced along with the ve-dimensional and (as a brief di-
gression) six-dimensional sub-spaces O
n
and V
6
as special coordinate
systems.
Conventions: Small Latin indices run from 1 to 4. Capital Latin in-
dices run from 1 to 5. Round and square brackets on particular tensor
indices indicate symmetric and skew-symmetric characters, respectively.
The covariant derivative is indicated either by a semi-colon or the sym-
bol . The ordinary partial derivative is indicated either by a comma
or the symbol . Einstein summation convention is, as usual, employed
throughout this work. Finally, by the word space we may also mean
space-time.
1.2 Geometric construction of a mixed, metric-compatible
four-manifold R
4
possessing an internal spin space S
p
Our four-dimensional manifold R
4
is endowed with a general asymmetric
connection
i
jk
and possesses a fundamental asymmetric tensor dened
herein by
ij
=
(ij)
+
[ij]

1
2
_
g
(ij)
+g
[ij]
_
, (1.1)
where g
(ij)
2
(ij)
will play the role of the usual geometric metric ten-
sor with which we raise and lower indices of tensors while g
[ij]
2
[ij]
will play the role of a fundamental spin tensor (or of a skew- or anti-
symmetric metric tensor). We shall also refer both to as the funda-
mental tensors. They satisfy the relations
g
(ij)
g
(kj)
=
k
i
, (1.2a)
g
[ij]
g
[kj]
=
k
i
, (1.2b)
g
(ij)
g
[jk]
= g
[ij]
g
(jk)
. (1.2c)
We may construct the fundamental spin tensor as a generalization
of the skew-symmetric symplectic metric tensor in any M-dimensional
space(-time), where M =2, 4, 6, . . . , M =2 +p, embedded in (M +n)-
dimensional enveloping space(-time), where n=0, 1, 2, . . . . In M
dimensions, we can construct p =M2 null (possibly complex) normal
vectors (null n-legs) z
1
, z
2
, . . . , z
p
with z
m
z
n
=0 (m, n, . . . =1, . . . , p
and , , . . . =1, . . . , M). If we dene the quantity
where
the skew-symmetric, self-dual null bivector
denes a null rotation,

then these null n-legs are normal to the (hyper)plane

M+n2
R
M
(contained in R
M
) dened in such a way that
=
...
z
1
z
2
. . . z
p
= (M p)!
_
z
1[
z
2| ]
+ +z
1[
z
p| ]
_
,
= 0 ,
...
=

g
...
,
...
=
1
g

...
, g = det g.
Here
...
is the Levi-Civita permutation symbol. Hence in four
dimensions we have
1
z
2
=
1
2

= z
1
z
2
z
1
z
2
,
1
z
2
z
1
z
2
,
tr
1
z
2
z
1
z
2
= 0 ,
where
is the generalized Kronecker delta and
. (The minus sign holds if the manifold is Lorentzian and

vice versa.) The particular equation tr =0 is of course valid in M
dimensions as well. In M dimensions, the fundamental spin tensor of
our theory is dened as a bivector satisfying
g
[]
g
[]
=
+
1
M 1
_
_
. (1.3a)
Hence the above relation leads to the identity
g
[]
g
[]
=
. (1.3b)
In the particular case of M =2 and n=1, the
vanishes and the

fundamental spin tensor is none other than the two-dimensional Levi-
Civita permutation tensor:
g
[AB]
=
AB
=

g
_
0 1
1 0
_
,
g
[AB]
g
[CD]
=
C
A
D
B

C
B
D
A
,
g
[AC]
g
[BC]
=
B
A
,
where A, B=1, 2. Lets now return to our four-dimensional manifold
R
4
. We now have
g
[ij]
g
[kl]
=
kl
ij
+
1
3
_
k
i

l
j

k
j
l
i
_
,
g
[ij]
g
[kj]
=
k
i
.
Hence the eigenvalue equation is arrived at:
g
[ij]
=
kl
ij
g
[kl]
. (1.3c)
We can now construct the symmetric traceless matrix Q
k
i
through
Q
k
i

kl
ij
u
l
u
j
=
ijpq
klrs
z
p
1
z
q
2
z
r1
z
s2
u
l
u
j
,
kl
ij
Q
k
i
u
j
u
l
,
Q
ik
= Q
ki
,
tr Q = 0 ,
Q
ik
u
k
= 0 ,
where u
i
is the unit velocity vector, u
i
u
i
=1. Lets introduce the unit
spin vector:
v
i
= g
[ik]
u
k
,
v
i
v
i
= 1, u
i
v
i
= 0 ,
g
[ik]
v
i
u
k
= 1 .
Multiplying both sides of (1.3c) by the unit spin tensor, we get
g
[ij]
u
j
= Q
k
i
g
[kr]
u
r
.
In other words,
v
i
= Q
i
k
v
k
. (1.3d)
Now we can also verify that
ij
kj
=
k
i
. (1.4)
Note that since the fundamental tensor is asymmetric it follows that
ij
jk
= g
(ij)
g
[jk]
,=
ij
kj
_
,=
k
i
_
. (1.5)
The line-element of R
4
can then be given through the asymmetric
fundamental tensor:
ds
2
=

2
ik
dx
i
dx
k
= g
(ik)
dx
i
dx
k
.
There exists in general no relation such as g
[kj]
g
(ji)
=
k
i
. However,
we have the relations
g
(rs)
g
[ir]
g
[js]
= g
(ij)
(a)
g
[rs]
g
(ir)
g
(js)
= g
[ij]
(b)
_
. (1.6)
We now introduce the basis g
l
which spans the metric space of R
4
and its associate
l
which spans the spin space S
p
R
4
(we identify
the manifold R
4
as having the Lorentzian signature 2, i.e., it is a
space-time). These bases satisfy the algebra
i
= g
[ik]
g
k
(a)
g
i
= g
[ki]
k
(b)
(g
i
g
j
) = (
i

j
) = g
(ij)
(c)
_
g
i
g
j
_
=
_
i

j
_
=
i
j
(d)
(
i
g
j
) = g
[ij]
(e)
_
_
. (1.7)
We can derive all of (1.2) and (1.6) by means of (1.7). In a pseudo-
ve-dimensional space
n
=R
4
n (a natural extension of R
4
which
includes a microscopic fth coordinate axis normal to all the coordinate
patches of R
4
), the algebra is extended as follows:
[g
i
, g
j
] = C
k
ij
g
k
+g
[ij]
n (a)
[n, g
i
] = g
[ij]
g
j
(b)
i
= [n, g
i
]
_
= g
[ij]
g
j
_
(c)
_
_
, (1.8)
where the square brackets [ ] are the commutation operator, C
k
ij
stands
for the commutation functions and n is the unit normal vector to the
manifold R
4
satisfying (n n) = 1. Here we shall always assume that
(n n) = +1 anyway. In summary, the symmetric and skew-symmetric
metric tensors can be written in
n
=R
4
n as
g
(ik)
= (g
i
g
k
) (d)
g
[ik]
= ([g
i
, g
k
] n) (e)
_
. (1.8)
The (intrinsic) curvature tensor of the space R
4
is given through the
relations
R
i
jkl
() =
i
jl,k

i
jk,l
+
m
jl
i
mk

m
jk
i
ml
,
a
i;j;k
a
i;k;j
= R
l
ijk
a
l
2
l
[jk]
a
i;l
,
for an arbitrary vector a
i
. The torsion tensor
i
[jk]
is introduced through
the relation
;i;k

;k;i
= 2
r
[ik]
,r
,
which holds for an arbitrary scalar eld . The connection of course can
be written as
i
jk
=
i
(jk)
+
i
[jk]
. The torsion tensor
i
[jk]
together with
the spin tensor g
[ij]
shall play the role associated with the internal spin
of an object moving in space-time. On the manifold R
4
, lets now turn
our attention to the spin space S
p
and evaluate the tangent component
of the derivative of the spin basis
l
with the help of (1.7):
(
j
i
)
T
=
_
g
[ik],j
g
[lk]
g
[ik]
g
[lm]
k
mj
_
l
(1.9)
since (
j
g
i
)
T
=
k
ij
g
k
. Now with the help of (1.3), we have
(
j

i
)
T
=
_
g
[ik],j
g
[lk]
1
3
_
l
i

l
ij
_
_
l
,
where we have put
j
=
i
ij
. On the other hand one can easily show that
by imposing metricity upon the two fundamental tensors (use (1.7) to
prove this), the following holds:
i
jk
=
_
j
g
i
g
k
_
=
_
i

k
_
. (1.10)
Thus, solving for a tetrad-independent connection, we have
i
jk
=
3
2
g
[ir]
g
[jr],k

1
2

i
j
k
. (1.11)
The torsion tensor is therefore
i
[jk]
=
1
4
_
i
k

k
i
j
_
+
3
4
g
[il]
_
g
[jl],k
g
[kl],j
_
=
=
1
4
_
i
k

k
i
j
_
+
3
4
g
[il]
g
[jk],l
,
(1.12)
where we have assumed that the fundamental spin tensor is a pure curl:
g
[ij]
=
i,j

j,i
, g
[ij],k
+g
[jk],i
+g
[ki],j
= 0 .
This expression and the symmetric part of the connection:
i
(jk)
=
1
2
g
(li)
_
g
(lj),k
g
(jk),l
+g
(kl),j
_
g
(li)
g
(jm)
m
[lk]
g
(li)
g
(km)
m
[lj]
,
therefore determine the connection uniquely in terms of the fundamental
tensors alone:
i
jk
=
1
2
g
(li)
_
g
(lj),k
g
(jk),l
+g
(kl),j
_
+
1
2
_
i
k
j
g
(ir)
g
(jk)

r
_
+
+
3
4
g
[ir]
_
g
[jr],k
g
[kr],j
_
3
4
g
(li)
g
(jr)
g
[rs]
_
g
[ls],k
g
[ks],l
_
3
4
g
(li)
g
(kr)
g
[rs]
_
g
[ls],j
g
[js],l
_
.
(1.13)
There is, however, an alternative way of expressing the torsion ten-
sor. The metric and the fundamental spin tensors are treated as equally
fundamental and satisfy the ansatz
g
(ij);k
= g
[ij];k
= 0 and
ij;k
= 0 .
Therefore, from g
[ij];k
=0, we have the following:
g
[ij],k
= g
[mj]
m
ik
+g
[im]
m
jk
.
Letting W
jik
=g
[jm]
m
ik
, we have
g
[ij],k
= W
ijk
W
jik
. (1.14)
Solving for W
i[jk]
by making cyclic permutations of i, j and k, we get
W
i[jk]
=
1
2
_
g
[ij],k
g
[jk],i
+g
[ki],j
_
+W
j(ik)
W
k(ij)
.
Therefore
W
ijk
= W
i(jk)
+W
i[jk]
=
=
1
2
_
g
[ij],k
g
[jk],i
+g
[ki],j
_
+W
i(jk)
+W
j(ik)
W
k(ij)
.
Now recall that W
ijk
=g
[im]
m
jk
. Multiplying through by g
[il]
, we get
i
jk
=
1
2
g
[li]
_
g
[lj],k
g
[jk],l
+g
[kl],j
_
+
+g
[li]
g
[jm]
m
(lk)
g
[li]
g
[km]
m
(lj)
+
i
(jk)
.
(1.15)
On the other hand,
i
jk
=
1
2
g
(li)
_
g
(lj),k
g
(jk),l
+g
(kl),j
_

g
(li)
g
(jm)
m
[lk]
g
(li)
g
(km)
m
[lj]
+
i
[jk]
.
(1.16)
From (1.15) the torsion tensor is readily read o as
i
[jk]
=
1
2
g
[li]
_
g
[lj],k
g
[jk],l
+g
[kl],j
_
+
+g
[li]
g
[jm]
m
(lk)
g
[li]
g
[km]
m
(lj)
.
(1.17)
We denote the familiar symmetric Levi-Civita connection by
_
i
jk
_
=
1
2
g
(li)
_
g
(lj),k
g
(jk),l
+g
(kl),j
_
.
If we combine (1.13) and (1.17) with the help of (1.2), (1.3) and (1.4),
after a rather lengthy but straightforward calculation we may obtain a
solution:
i
[jk]
=
1
2
g
[li]
_
g
[lj],k
g
[jk],l
+g
[kl],j
_
+
+g
[li]
g
[jm]
_
m
lk
_
g
[li]
g
[km]
_
m
lj
_
+
1
3
_
m
[mk]
i
j

m
[mj]
i
k
_
.
(1.18)
Now the spin vector
i
[ik]
is to be determined from (1.12). If we
contract (1.12) on the indices i and j, we have
i
[ik]
=
3
4
g
[ij]
_
g
[ij],k
g
[kj],i
_
3
4

k
. (1.19)
But from (1.14):
k
=
i
ik
=
1
2
g
(ij)
g
(ij,k
=
1
2
g
[ij]
g
[ij],k
. (1.20)
Hence (1.19) becomes
i
[ik]
S
k
=
3
8
g
[ij]
g
[ij],k

3
4
g
[ij]
g
[kj],i
. (1.21)
Then with the help of (1.21), (1.18) reads
i
[jk]
=
1
2
g
[li]
_
g
[lj],k
g
[jk],l
+g
[kl],j
_
+
+g
[li]
g
[jm]
_
m
lk
_
g
[li]
g
[km]
_
m
lj
_
+
+
1
8
g
[mn]
_
g
[mn],k
i
j
g
[mn],j
i
k
_
+
1
4
g
[mn]
_
g
[km],n
i
j
g
[jm],n
i
k
_
.
(1.22)
So far, we have been able to express the torsion tensor, which shall
generate physical elds in our theory, in terms of the components of the
fundamental tensor alone.
In a holonomic frame,
i
[jk]
=0 and, of course, we have from (1.16)
the usual Levi-Civita (or Christoel) connection:
i
jk
=
_
i
jk
_
=
1
2
g
(li)
_
g
(lj),k
g
(jk),l
+g
(kl),j
_
.
If therefore a theory of gravity adopts this connection, one may argue
that in a strict sense, it does not admit an integral concept of internal
spin in its description. Such is the classical theory of General Relativity.
In a rigid frame (constant metric) and in a pure electromagnetic
gauge condition, one may have
i
(jk)
=0 and in this special case we
have from (1.15)
i
jk
=
1
2
g
[li]
_
g
[lj],k
g
[jk],l
+g
[kl],j
_
,
which is exactly the same in structure as
_
i
jk
_
with the fundamental
spin tensor replacing the metric tensor. We shall call this connection
the pure spin connection, denoted by
L
i
jk
=
1
2
g
[li]
_
g
[lj],k
g
[jk],l
+g
[kl],j
_
. (1.23)
Lets give an additional note to (1.20). Lets nd the expression for
i
ki
, provided we know that
k
=
i
ik
=
1
2
g
(ij)
g
(ij),k
=
=
1
2
g
[ij]
g
[ij],k
=
=
_
ln
g
_
,k
=
_
i
ik
_
.
Meanwhile, we express the following relations:
S
k
=
i
[ik]
=
i
[ki]
=
=
3
8
g
[ij]
g
[ij],k

3
4
g
[ij]
g
[kj],i
=
=
3
4
_
k
g
[ij]
g
[kj],i
_
,
i
(ik)
=
_
i
ik
_
g
(li)
g
(im)
m
[lk]
g
(li)
g
(km)
m
[li]
=
k

i
[ik]
=
=
1
4

k
+
3
4
g
[ij]
g
[kj],i
.
Therefore
i
ki
=
i
(ki)
+
i
[ki]
=
i
(ki)

i
[ik]
=
=
3
2
g
[ij]
g
[kj],i

1
2

k
,
which can also be derived directly from (1.11). From (1.11) we also see
that
g
[ij],k
=
2
3
g
[rj]
r
ik
+
1
3
g
[ij]
k
, (1.24a)
g
[rj]
r
ik
= g
[ri]
r
jk
. (1.24b)
Also, for later purposes, we derive the condition for the conservation
of charges:
g
[ik]
,k
=
1
3
g
[ik]
k
. (1.24c)
Having developed the basic structural equations here, we shall see in
the following chapters that the gravitational and electromagnetic ten-
sors are formed by means of the fundamental tensors g
(ij)
2
(ij)
and g
[ij]
2
[ij]
alone (see Section 4.2). In other words, gravity and
electromagnetism together arise from this single tensor. We shall also in-
vestigate their fundamental relations and ultimately unveil their union.
Chapter 2
THE UNIFIED FIELD THEORY
2.1 Generalization of Kaluzas projective theory
We now assume that the space-time R
4
is embedded in a general
ve-dimensional Riemann space IR
5
. This is referred to as embedding
of class 1. We shall later dene the space
n
=R
4
n to be a special
coordinate system in IR
5
. The ve-dimensional metric tensor g
(AB)
of
IR
5
of course satises the usual projective relations
g
(AB)
= e
i
A
e
j
B
g
(ij)
+n
A
n
B
,
e
A
i
n
A
= 0 ,
where e
A
i
=
i
x
A
is the tetrad. If now g
l
denotes the basis of R
4
and
e
A
of IR
5
:
e
A
= e
i
A
g
i
+n
A
n,
g
i
= e
A
i
e
A
, g
(ij)
= e
A
i
e
B
j
g
(AB)
,
e
A
i
e
i
B
=
A
B
n
A
n
B
, e
A
i
e
j
A
=
j
i
.
The derivative of g
i
R
4
IR
5
is then
j
g
i
=
k
ij
g
k
+
ij
n.
We also have the following relations:
g
i;j
=
ij
n,
i
n =
j
i
g
j
, e
A;B
= 0 , (
B
e
A
=
C
AB
e
C
) ,
e
A
i;j
=
ij
n
A
, n
A
;i
=
j
i
e
A
j
, e
i
A;B
= n
A
i
j
e
j
B
, g
(ij);k
= g
(AB);C
= 0 .
In our work we shall, however, emphasize that the exterior curvature
tensor
ij
is in general asymmetric:
ij
,=
ji
just as the connection
i
jk
is. This is so since in general
j
e
A
i
,=
i
e
A
j
. Within a boundary , the
metric tensor g
(ij)
may possess discontinuities in its second derivatives.
Now the connection and exterior curvature tensor satisfy
k
ij
= e
k
A
j
e
A
i
+e
k
A
A
BC
e
B
i
e
c
j
(a)
ij
= n
A
j
e
A
i
+n
A
A
BC
e
B
i
e
c
j
(b)
_
, (2.1)
Chapter 2 The Unified Field Theory 21
where
j
e
A
i
= e
A
k

k
ij

A
BC
e
B
i
e
C
j
+
ij
n
A
. (2.2)
We can also solve for
A
BC
in (2.1) with the help of the projective
relation e
A
= e
i
A
g
i
+n
A
n. The result is, after a quite lengthy calculation,
A
BC
= e
A
i

C
e
i
B
+e
A
k

k
ij
e
i
B
e
j
C
+
ij
e
i
B
e
j
C
n
A
+
+ (
C
n
B
)n
A
i
j
e
A
i
e
j
C
n
B
.
(2.3)
If we now perform the calculation (
k
j

j
k
) g
i
with the help of
some of the above relations, we have in general
(
k

j

j
k
) g
i
= R
A
BCD
e
B
i
e
C
j
e
D
k
e
A
+
+
_
k
e
C
j

j
e
C
k
_
A
BC
e
B
i
e
A
+
+ (
k
j

j
k
) e
A
i
e
A
.
(2.4)
Here we have also used the fact that
(
C

B

B
C
) e
A
= R
D
ABC
e
D
.
On the other hand,
i
n =
j
i
g
j
,and
j
g
i
=
_
l
ij
g
l
+
ij
n
_
,
k
=
=
_
l
ij,k
+
m
ij
l
mk

ij
l
k
_
g
l
+
_
ij,k
+
l
ij
lk
_
n.
Therefore we obtain another expression for (
k

j

j
k
) g
i
:
(
k
j

j
k
) g
i
=
_
R
l
ijk
+
ik
l
j

ij
l
k
_
g
l
+
+
_
ij;k

ik;j
+ 2
l
[jk]
il
_
n,
(2.5a)
(
k

j

j
k
) e
A
i
+
_
k
e
C
j

j
e
C
k
_
A
BC
e
B
i
S
A
ijk
. (2.5b)
Combining (2.4) and (2.5), we get, after some algebraic mani-
pulations,
R
ijkl
=
ik
jl

il
jk
+R
ABCD
e
A
i
e
B
j
e
C
k
e
D
l
S
Ajkl
e
A
i
(a)
ij;k

ik;j
= R
ABCD
n
A
e
B
i
e
C
j
e
D
k
2
l
[jk]
il
+S
Aijk
n
A
(b)
_
_
_
. (2.6)
We have thus established the straightforward generalizations of the
equations of Gauss and Codazzi.
Now the electromagnetic content of (2.6) can be seen as follows:
rst we split the exterior curvature tensor
ij
into its symmetric and
skew-symmetric parts:
ij
=
(ij)
+
[ij]
(ij)
k
ij
,
[ij]
f
ij
_
. (2.7)
Here the symmetric exterior curvature tensor k
ij
has the explicit
expression
k
ij
=
1
2
n
A
_
j
e
A
i
+
i
e
A
j
_
+n
A
A
(BC)
e
B
i
e
C
j
=
=
1
2
e
A
i
e
B
j
(n
A;B
+n
B;A
) .
(2.8)
Furthermore, in our formalism, the skew-symmetric exterior curva-
ture tensor f
ij
is naturally equivalent to the electromagnetic eld tensor
F
ij
. It is convenient to set f
ij
=
1
2
F
ij
. Hence the electromagnetic eld
tensor can be written as
F
ij
= n
A
_
j
e
A
i

i
e
A
j
_
+ 2n
A
A
[BC]
e
B
i
e
C
j
=
= e
A
i
e
B
j
(n
A;B
n
B;A
)
(2.9)
The ve-dimensional electromagnetic eld tensor is therefore
F
AB
=
A
n
B

B
n
A
.
2.2 Fundamental eld equations of our unied eld theory.
Geometrization of matter
We are now in a position to simplify (2.6) by invoking two conditions.
The rst of these, following Kaluza, is the cylinder condition: the laws
of physics in their four-dimensional form shall not depend on the fth
coordinate x
5
n
. We also assume that x
5
y is a microscopic co-
ordinate in
n
. In short, the cylinder condition is written as (by rst
putting n
A
=e
A
5
)
g
(ij),5
= g
(ij),A
n
A
= e
B
i
e
C
j
(n
B;C
+n
C;B
) = 0 ,
where we have now assumed that in IR
5
the dierential expression e
i
A,B
e
i
B,A
vanishes. However, from (2.2), we have the relation
e
A
i,j
e
A
j,i
= 2e
A
k

k
[ij]
+F
ij
n
A
.
Furthermore, the cylinder condition implies that n
A;B
+ n
B;A
=0
and therefore we can nullify (2.8). This is often called the assumption
of weakness. The second condition is the condition of integrability
imposed on arbitrary vector elds, e.g., on
i
(say) in R
4
. The necessary
and sucient condition for a vector eld
,i

i
(a one-form) to be
integrable is
i,j
=
j,i
. If this is applied to (2.4), we will then have
R
A
BCD
e
B
i
e
C
j
e
D
k
= S
A
ijk
. Therefore (2.6) will now go into
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) , (2.10a)
F
ij;k
F
ik;j
= 2
l
[jk]
F
il
. (2.10b)
These are the sought unied eld equations of gravity and electro-
magnetism. They form the basic eld equations of our unied eld
theory.
Altogether they imply that
R
ik
=
1
4
F
ij
F
j
k
= R
(ik)
, (2.11a)
R
[ik]
= 0 , (2.11b)
R =
1
4
F
ik
F
ik
, (2.11c)
F
;j
ij
= 2
l
[ik]
F
k
l
J
i
. (2.11d)
These relations are necessary and sucient following the two condi-
tions we have dealt with. These eld equations seem to satisfy a denite
need. They tell us a beautiful and simple relation between gravity and
electromagnetism: (2.10a) tells us that both inside and outside charges,
a gravitational eld originates in a non-null electromagnetic eld (as
in Rainichs geometry), since according to (2.10b), the electromagnetic
current is produced by the torsion of space-time: the torsion produces
an electromagnetic source. The electromagnetic current is generated by
dynamic electric-magnetic charges. In a strict sense, the gravitational
eld cannot exist without the electromagnetic eld. Hence all matter
in the Universe may have an electromagnetic origin. Denoting by d a
three-dimensional innitesimal boundary enclosing several charges, we
have, from (2.11d)
e = 2
_
r
[ik]
u
i
F
k
r
d.
We may represent a negative charge by a negative spin produced
by a left-handed twist (torsion) and a positive one by a positive spin
produced by a right-handed twist. (For the conservation of charges
(currents) see Sections 4.3 and 4.4.) Now (2.11c) tells us that when the
spatial curvature, represented by the Ricci scalar, vanishes, we have a
null electromagnetic eld, also it is seen that the strength of the electro-
magnetic eld is equivalent to the spatial curvature. Therefore gravity
and electromagnetism are inseparable. The electromagnetic source, the
charge, looks like a microscopic spinning hole in the structure of the
space-time R
4
, however, the Schwarzschild singularity is non-existent in
general. Consequently, outside charges our eld equations read
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) , (2.12a)
F
ij
;j
= 0 , (2.12b)
which, again, give us a picture of how a gravitational eld emerges
(outside charges).
In this way, the standard action integral of our theory may take the
form
I =
_
_
R
_
R
ik
R
ik
_
1
2
_
g d
4
x = (2.13)
=
_
_
R
_
1
16
_
F
ik
F
ik
_
2
R
ijkl
R
ijkl
1
4
R
ijkl
F
il
F
jk
_1
2
_
g d
4
x.
Here

R denotes the Ricci scalar built from the symmetric Christoel
connection alone.
From the variation of which, we would arrive at the standard
Einstein-Maxwell equations. However, we do not wish to stress heavy
emphasis upon such an action-method (which seems like a forced short
cut) in order to arrive at the eld equations of our unied eld the-
ory. We must emphasize that the equations (2.10)-(2.13) tell us how
the electromagnetic eld is incorporated into the gravitational eld in a
very natural manner, in other words theres no need here to construct
any Lagrangian density of such. We have been led into thinking of how
to couple both elds using dierent procedures without realizing that
these elds already encapsulate each other in Nature. But here our
space-time is already a polarized continuum in the sense that there ex-
ists an electromagnetic eld at every point of it which in turn generates
a gravitational eld.
Remark 1
Without the integrability condition we have, in fairly general conditions,
the relation
(
k
j

j
k
) e
A
i
= R
l
ijk
e
A
l
+R
A
BCD
e
B
i
e
C
j
e
D
k
+
+
_
ij;k

ik;j
+ 2
l
[jk]
il
_
n
A
+
_
ik
l
j

ij
l
k
_
e
A
l

2
A
BC
_
l
[jk]
e
C
l
+
[jk]
n
C
_
e
B
i
,
(2.a)
S
A
ijk
(
k
j

j
k
) e
A
i
+ 2
A
BC
_
l
[jk]
e
C
l
+
[jk]
n
C
_
e
B
i
=
= (
k
j

j
k
) e
A
i
+
A
BC
_
k
e
C
j

j
e
C
k
_
e
B
i
.
Hence we have
R
ijkl
=
ik
jl

il
jk
+R
ABCD
e
A
i
e
B
j
e
C
k
e
D
l
S
Ajkl
e
A
i
, (2.b
1
)
ij;k

ik;j
= R
ABCD
n
A
e
B
i
e
C
j
e
D
k
+S
Aijk
n
A
2
l
[jk]
il
. (2.b
2
)
These are just the equations in (2.6). Upon employing a suitable
cylinder condition and putting
[ij]
=
1
2
F
ij
(within suitable units), we
have the complete set of eld equations of gravoelectrodynamics:
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) +R
ABCD
e
A
i
e
B
j
e
C
k
e
D
l
S
Ajkl
e
A
i
, (2.c
1
)
1
2
(F
ij;k
F
ik;j
) = R
ABCD
n
A
e
B
i
e
C
j
e
D
k
+S
Aijk
n
A
l
[jk]
F
il
. (2.c
2
)
End of Remark 1
Sub-remark
Lets consider the space S
5
=R
4
Y which describes a ve-dimensional
thin shell where Y is the microscopic coordinate representation
spanned by the unit normal vector to the four-manifold R
4
. The coor-
dinates of this space are characterized by y
=
_
x
i
, y
_
where the Greek
indices run from the 1 to 5 and where the extra coordinate y is taken
to be the Planck length:
y =
_
G
c
3
,
which gives the thickness of thin shell. Here G is the gravitational
constant of Newton, is the Planck constant divided by 2 and c is
the speed of light in vacuum. (From now on, since the Planck length is
extremely tiny, we may drop any higher-order terms in y.) Then the
basis
of the space S
5
can in general be split into
i
=
_
k
i
y
k
i
_
g
k
,
5
= n.
It is seen that the metric tensor of the space S
5
, i.e.,
, has the
following non-zero components:
ik
= g
(ik)
2y
(ik)
,
55
= 1 .
The simplest sub-space of the space S
5
is given by the basis
g
i
=
i
_
x
i
, 0
_
,
g
5
= n,
where g
l
is of course the tangent basis of the manifold R
4
. We shall
denote this pseudo-ve-dimensional space as the special coordinate sys-
tem
n
=R
4
n whose metric tensor g
can be arrayed as
g
=
_
g
(ik)
|
4x4
0
0 1
_
.
Now the tetrad of the space S
5
is then given by
A
=
_
e
A

_
, which
can be split into
A
i
= e
A
i
y
k
i
e
A
k
,
A
5
= n
A
.
Then we may nd the inverse to the tetrad
A
i
as follows:
i
A
= e
i
A
+y
i
k
e
k
A
,
5
A
= n
A
= y
,A
.
From the above relations, we have the following:
i
A
A
k
=
i
k
,
A
k
e
i
A
=
i
k
y
i
k
,
i
A
e
A
k
=
i
k
+y
i
k
.
The ve-dimensional index of the tetrad
A
i
is raised and lowered
using the metric tensor g
(AB)
. The inverse of the tetrad is achieved
with the help of the metric tensor
G
AB
= e
i
A
e
k
B
g
(ik)
+n
A
n
B
2y
(ik)
e
i
A
e
k
B
,
which reduces to g
(AB)
due to the cylinder condition. The
four-dimensional metric tensor g
(ik)
is used to raise and lower the four-
dimensional index. Again, we bring in the electromagnetic eld tensor
F
ij
via the cylinder condition, which yields
ik
=
1
2
F
ik
. The connection
of the space S
5
is then
_
x
i
, y
_
=

1
2
F
y
,

1
2
F
y
,

1
2
yF
,

1
2
yF
,
where
is the connection of the space

n
=R
4
n with the electro-
magnetic eld tensor derived from it: F
=2
5
, F
= 2
5
, F
5
=0.
Therefore its only non-zero components are F
ik
. At the base of the space
S
5
, the connection is
_
x
i
, 0
_
=

1
2
F
y
,

1
2
F
y
,
.
From the above expression, we see that
k
=
k

1
2
F
k
y
,
,
5
=
1
2
F
5
= F
.
As can be worked out, the ve-dimensional connection of the back-
ground space IR
5
is related to that of S
5
through
A
BC
_
x
i
, y
_
=
A
B,C
+
A
C

1
2
F
B
n
C
1
2
y
A
B
F
k
i
e
i
C
.
Now the ve-dimensional curvature tensor R
ABCD
=R
ABCD
_
x
i
, 0
_
is to be related once again to the four-dimensional curvature tensor
of R
4
, which can be directly derived from the curvature tensor of the
space S
5
as R
ijkl
=
S
5
R
ijkl
_
x
i
, 0
_
. With the help of the above geometric
objects, and after some laborious work-out, we arrive at the relation
R
ABCD
= e
A
e
B
e
C
e
D
R

1
2
F
AB
F
CD
+
ABCD
,
where we have a new geometric object constructed from the electromag-
netic eld tensor:
ABC
=
1
2
F
(e
B
n
C
e
C
n
B
) e
+
1
2
F
,
(n
B
e
C
n
C
e
B
) e
A
+
+
1
2
F
A,C
(x
, 0) n
B

A,B
(x
, 0) n
C
_
+
+
1
2
F
(n
B
e
C
n
C
e
B
) e
A
BCD

BCD
e
A
.
Dene another curvature tensor:
ijkl
(R
ABCD

ABCD
) e
A
i
e
B
j
e
C
k
e
D
l
.
Then we have the relation
R
ijkl
=
ijkl
+
1
2
F
ij
F
kl
.
By the way, the curvature tensor of the space
n
=R
4
n is here
given by
R
,
(x
, 0)
,
(x
, 0) +
.
Expanding the connections in the above relation, we obtain
R
,
+
+
+
1
2
_
F
,
F
,
_
y
,
+
1
2
F
y
,

y
,
_
+
+
1
2
_
F
_
y
,
+
1
2
_
F
_
y
,
+
+
1
4
F
_
F
y
,
F
y
,
_
y
,
+
1
4
F
_
F
y
,
F
y
,
_
y
,
+
+
1
4
_
F
_
y
,
y
,
.
We therefore see that the electromagnetic eld tensor is also present
in the curvature tensor of the space
n
=R
4
n. In other words, elec-
tromagnetic and gravitational interactions are described together on an
equal footing by this single curvature tensor.
Direct calculation shows that some of its four-dimensional and mixed
components are
R
i
jkl
=
i
jl,k

i
jk,l
+
r
jl
i
rk

r
jk
i
rl
,
R
5
ijk
= R
5ijk
=
1
2
_
F
ij;k
F
ik;j
+ 2
r
[jk]
F
ir
_
,
R
ij5k
= 0 .
Furthermore, we obtain the following equivalent expressions:
R
ABCD
= R
ijkl
e
i
A
e
j
B
e
k
C
e
l
D
+
1
2
_
F
ik;l
F
il;k
+ 2
m
[kl]
F
im
_
e
i
A
n
B
e
k
C
e
l
D
+
+
1
2
_
F
jl;k
F
jk;l
+2
m
[lk]
F
jm
_
n
A
e
j
B
e
k
C
e
l
D
1
2
F
AB
F
CD
+
ABCD
,
R
ABCD
=
1
2
(F
AC;D
F
AD;C
) n
B
+
1
2
(F
BD;C
F
BC;D
) n
A
1
2
F
AB
F
CD

ABCD
+R
ijkl
e
i
A
e
j
B
e
k
C
e
l
D
+
+ 2
m
[kl]
F
im
e
i
A
n
B
e
k
C
e
l
D
+ 2
m
[lk]
F
jm
n
A
e
j
B
e
k
C
e
l
D
,
where
ABCD
=
1
2
_
F
MA
_
F
M
C
n
D
F
M
D
n
C
_
n
B
F
MB
_
F
M
C
n
D
F
M
D
n
C
_
n
A
_
ABCD
.
When the torsion tensor of the space R
4
vanishes, we have the rela-
tion
R
ABCD
=
1
2
(F
AC;D
F
AD;C
) n
B
+
1
2
(F
BD;C
F
BC;D
) n
A
1
2
F
AB
F
CD

ABCD
+R
ijkl
e
i
A
e
j
B
e
k
C
e
l
D
which, again, relates the curvature tensors to the electromagnetic eld
tensor.
Finally, if we dene yet another ve-dimensional curvature tensor:
R
ABCD

R
ABCD
+
1
2

F
AB
F
CD

ABCD
,
where

R
ABCD
,

F
AB
and
ABCD
are the extensions of R
ABCD
, F
AB
and
ABCD
which are dependent on y, we may obtain the relation
R
ABCD
e
A
i
e
B
j
e
C
k
e
D
l
=
= R
ijkl
+
1
2
yF
r
i
R
rjkl
+
1
2
yF
r
j
R
irkl
+
1
2
yF
r
k
R
ijrl
+
1
2
yF
r
l
R
ijkr
.
End of Sub-remark
Lets now write the eld equations of our unied eld theory as
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) +
ijkl
, (2.d
1
)
F
ij;k
F
ik;j
= 2
l
[jk]
F
il
+
ijk
, (2.d
2
)
ijkl
= R
ABCD
e
A
i
e
B
j
e
C
k
e
D
l
S
Ajkl
e
A
i
, (2.e
1
)
ijk
= 2
_
R
ABCD
e
B
i
e
C
j
e
D
k
S
Aijk
_
n
A
, (2.e
2
)
Consider the invariance of the curvature tensor under the gauge
transformation
i
jk
=
i
jk
+
i
j
,k
(2.f)
for some function =(x). This is analogous to the gauge transforma-
tion of the electromagnetic potential, i.e.,

i
=
i
+
,i
, with a scaling
constant , which leaves the electromagnetic eld tensor invariant.. We
dene the electromagnetic potential vector
i
and pseudo-vector
i
via
k
ki
=
i
+
i
where is a constant.
Then we see that the electromagnetic eld tensor can be expressed as
F
ik
=
i,k

k,i
=
1
(
i,k

k,i
) . (2.g)
More specically, the two possible electromagnetic potentials
i
and
i
transform homogeneously and inhomogeneously, respectively, accord-
ing to
i
= e
A
i

A
,
i
= e
A
i

A
+e
k
A
e
A
k,i
n
A
A
BC
n
B
e
C
i
.
The two potentials become equivalent in a coordinate system where
g equals a constant. Following (2.g), we can express the curvature

tensor as
R
ijkl
=
1
(F
ik
F
jl
F
il
F
jk
) +
2
_
g
(ik)
g
(jl)
g
(il)
g
(jk)
_
, (2.h)
where
1
and
2
are invariants. (The term
0
F
ij
F
kl
would contribute
nothing.) Hence
R
ik
=
1
F
il
F
l
k
+ 3
2
g
(ik)
. (2.i)
Putting
1
=
1
4
in accordance with (2.d
1
) and contracting (2.i) on
the indices i and k we see that
2
=
1
12
R
1
48
F
ik
F
ik
. Consequently, we
have the important relations
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) +
1
12
_
g
(ik)
g
(jl)
g
(il)
g
(jk)
_
R
1
48
_
g
(ik)
g
(jl)
g
(il)
g
(jk)
_
F
rs
F
rs
,
(2.j)
R
ik
=
1
4
F
il
F
l
k
+
1
4
g
(ik)
R
1
16
g
(ik)
F
rs
F
rs
. (2.k)
Comparing (2.j) and (2.e
1
) we nd
ijkl
= R
ABCD
e
A
i
e
B
j
e
C
k
e
D
l
S
Ajkl
e
A
i
=
=
1
12
_
g
(ik)
g
(jl)
g
(il)
g
(jk)
_
R
1
48
_
g
(ik)
g
(jl)
g
(il)
g
(jk)
_
F
rs
F
rs
.
(2.l)
Hence also
ik
=
1
4
g
(ik)
R
1
16
g
(ik)
F
rs
F
rs
, (2.m
1
)
= R
1
4
F
rs
F
rs
. (2.m
2
)
Note that our above consideration produces the following traceless
eld equation:
R
ik

1
4
g
(ik)
R =
1
4
_
F
il
F
l
k

1
4
g
(ik)
F
rs
F
rs
_
. (2.n)
In a somewhat particular case, we may set
1
4
g
(ik)
R =
_
c
2
u
i
u
k
+t
ik
_
where is a coupling constant and t
ik
is the general-
ized stress-metric tensor, such that R=
e
, where now
e
=c
2
+ t
is the eective material density. We also have
R
ik
1
2
g
(ik)
R =
_
c
2
u
i
u
k
+t
ik
_
+
1
4
_
F
il
F
l
k

1
4
g
(ik)
F
rs
F
rs
_
. (2.o)
The above looks slightly dierent from the standard eld equation
of General Relativity:
R
ik
1
2
g
(ik)
R = k
_
c
2
u
i
u
k
+t
ik
_
F
il
F
l
k
1
4
g
(ik)
F
rs
F
rs
__
, (2.p)
which is usually obtained by summing altogether the matter and elec-
tromagnetic terms. Here

R
ik
and

R are the Ricci tensor and scalar
built out of the Christoel connection and k ,= is the usual coupling
constant of General Relativity. Lets denote by m
0
, and c, the point-
mass, material density and speed of light in vacuum. Then the vanishing
of the divergence of (2.p) leads to the equation of motion for a charged
particle:
Du
i
Ds
u
i
;k
u
k
=
e
m
0
c
2
F
i
k
u
k
.
However, this does not provide a real hint to the supposedly miss-
ing link between matter and electromagnetism. We hope that theres
no need to add an external matter term to the stress-energy tensor.
We may interpret (2.n), (2.o) and (2.p) as telling us that matter and
electromagnetism are already incorporated, in other words, the elec-
tromagnetic eld produces material density out of the electromagnetic
current J. In fact these are all acceptable eld equations. Now, for in-
stance, we have R=
e
=
_
c
2
J
i
u
i
+t
_
. From (2.m
2
), the classical
variation follows:
I =
_

g d
4
x =
=
_ _
R
1
4
F
ik
F
ik
_
g d
4
x = 0 ,
(2.q)
which yields the gravitational and electromagnetic equations of Ein-
stein and Maxwell endowed with source since the curvature scalar here
contains torsion as well.
Finally, lets investigate the explicit relation between the Weyl tensor
and the electromagnetic eld tensor in this theory. In four dimensions
the Weyl tensor is
C
ijkl
= R
ijkl

1
2
_
g
(ik)
R
jl
+g
(jl)
R
ik
g
(il)
R
jk
g
(jk)
R
il
_
+
+
1
6
_
g
(ik)
g
(jl)
g
(il)
g
(jk)
_
R.
Comparing the above equation(s) with (2.j) and (2.k), we have
C
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) +
1
24
_
g
(ik)
g
(jl)
g
(il)
g
(jk)
_
F
rs
F
rs
1
8
_
g
(ik)
F
jr
F
r
l
+g
(jl)
F
ir
F
r
k
g
(il)
F
jr
F
r
k
g
(jk)
F
ir
F
r
l
_
.
(2.r)
We see that the Weyl tensor is composed solely of the electromag-
netic eld tensor in addition to the metric tensor. Hence we come to
the conclusion that the space-time R
4
is conformally at if and only
if the electromagnetic eld tensor vanishes. This agrees with the fact
that, when treating gravitation and electromagnetism separately, it is
the Weyl tensor, rather than the Riemann tensor, which is compatible
with the electromagnetic eld tensor. From the structure of the Weyl
tensor as revealed by (2.r), it is understood that the Weyl tensor ac-
tually plays the role of an electromagnetic polarization tensor in the
space-time R
4
. In an empty region of the space-time R
4
with a vanish-
ing torsion tensor, when the Weyl tensor vanishes, that region possesses
a constant sectional curvature which conventionally corresponds to a
constant energy density.
Lets for a moment turn back to (2.6). We shall show how to get the
source-torsion relation, i.e., (2.11d) in a dierent way. For this purpose
we also set a constraint S
A
ijk
=0 and assume that the background ve-
dimensional space is an Einstein space:
R
AB
= g
(AB)
,
where is a cosmological constant. Whenever =0 we say that the
space is Ricci-at or energy-free, devoid of matter. Taking into ac-
count the cosmological constant, this consideration therefore takes on a
slightly dierent path than our previous one. We only wish to see what
sort of eld equations it will produce.
We rst write
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) +R
ABCD
e
A
i
e
B
j
e
C
k
e
D
l
,
F
ij;k
F
ik;j
= 2R
ABCD
n
A
e
B
i
e
C
j
e
D
k
2
l
[jk]
F
il
.
Dene a symmetric tensor:
B
ik
R
ABCD
e
A
i
n
B
e
C
k
n
D
= B
ki
. (2.14)
It is immediately seen that
R
AB
e
A
i
e
B
k
= g
(ik)
, (2.15a)
R
AB
n
A
n
B
= , (2.15b)
R
AB
e
A
i
n
B
= 0 . (2.15c)
Therefore
R
ik
=
1
4
F
il
F
l
k
+R
AB
e
A
i
e
B
k
R
ABCD
e
A
i
n
B
e
C
k
n
D
=
=
1
4
F
il
F
l
k
+ g
(ik)
+B
ik
.
(2.16)
From (2.14) we also have, with the help of (2.15b), the following:
B = g
(ik)
B
ik
= R
ABCD
n
B
n
D
_
g
(AC)
n
A
n
C
_
=
= R
AB
n
A
n
B
+R
ABCD
n
A
n
B
n
C
n
D
=
= .
Hence we have
R =
1
4
F
ik
F
ik
+ 3. (2.17)
The Einstein tensor (or rather, the generalized Einstein tensor en-
dowed with torsion)
G
ik
R
ik

1
2
g
(ik)
R
up to this point is therefore
G
ik
= B
ik

1
8
g
(ik)
F
rs
F
rs
1
2
g
(ik)

1
4
F
il
F
l
k
. (2.18)
From the relation
F
ij;k
F
ik;j
= 2R
ABCD
n
A
e
B
i
e
C
j
e
D
k
2
l
[jk]
F
il
we see that
F
kj
;k
= 2R
BC
AD
n
A
e
k
B
e
j
C
e
D
k
2
j
l[.k]
F
kl
=
= 2R
C
A
e
j
C
n
A
2R
BC
AD
n
A
n
B
e
j
C
n
D
2
j
l[k]
F
kl
=
= 2
j
l[k]
F
kl
= J
j
.
In other words,
J
i
= 2
l
[ik]
F
k
l
,
which is just (2.11d). We will leave this consideration here and commit
ourselves to the eld equations given by (2.10) and (2.11) for the rest
of our work.
Lets obtain the (generalized) Bianchi identity with the help of (2.10)
and (2.11). Recall once again that
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) ,
F
ij;k
F
ik;j
= 2
l
[jk]
F
il
.
Performing the covariant derivative on R
ijkl
, the result of the cyclic
summation over the indices k, l and m is
R
ijkl;m
+R
ijlm;k
+R
ijmk;l
=
=
1
4
(F
ik;m
F
im;k
) F
jl
+
1
4
(F
jl;m
F
jm;l
) F
ik
+
+
1
4
(F
il;k
F
ik;l
) F
jm
+
1
4
(F
jm;k
F
jk;m
) F
il
+
+
1
4
(F
im;l
F
il;m
) F
jk
+
1
4
(F
jk;l
F
jl;k
) F
im
.
(2.19)
Or equivalently,
R
ijkl;m
+R
ijlm;k
+R
ijmk;l
=
=
1
2

n
[mk]
F
in
F
jl
+
1
2

n
[ml]
F
jn
F
ik
+
1
2

n
[kl]
F
in
F
jm
+
+
1
2

n
[km]
F
jn
F
il
+
1
2

n
[lm]
F
in
F
jk
+
1
2
n
[lk]
F
jn
F
im
.
(2.20)
From (2.10) if we raise the index i and then perform a contraction
with respect to the indices iand k, we have
R
jl;m
R
i
jml;i
R
jm;l
=
=
1
4
J
m
F
jl

1
4
J
l
F
jm
+
1
4
F
i
l
F
jm;i

1
4
F
i
l
F
ji;m
+
+
1
4
F
ji
F
i
m;l

1
4
F
ji
F
i
l;m
+
1
4
F
i
m
F
ji;l

1
4
F
i
m
F
jl;i
.
If we raise the index j and then contract on the indices j and l, we
have the expression
R
;m
2R
i
m;i
=
_
i
m
R 2R
i
m
_
;i
=
=
1
2
J
i
F
im
+
1
2
F
i
l
_
F
l
m;i
F
l
i;m
_
.
Therefore we have
_
R
i
m

1
2

i
m
R
_
;i
=
1
4
J
i
F
im
+
1
4
F
i
l
_
F
l
i;m
F
l
m;i
_
=
=
1
4
J
i
F
im
+
m
,
(2.21)
where
m
=
1
4
F
i
l
_
F
l
i;m
F
l
m;i
_
=
1
2
F
i
l
k
[im]
F
l
k
,
We can also write
_
R
ik
1
2
g
(ik)
R
_
;k
=
1
4
J
k
F
ik
+
i
. (2.22)
On the other hand, repeating the same contraction steps on
(2.20) gives
R
;m
2R
i
m;i
=
_
i
m
R 2R
i
m
_
;i
g
m
=
=
1
2

n
[kl]
F
k
n
F
l
m
+
1
2
n
[km]
F
l
n
F
k
l
+
1
2

n
[lm]
F
k
n
F
l
k
+
+
1
2

n
[lk]
F
l
n
F
k
m
=
=
n
[kl]
F
k
n
F
l
m
+
n
[km]
F
l
n
F
k
l
=
=
n
[kl]
F
k
n
F
l
m
2
m
.
Therefore we have
_
R
ik
1
2
g
(ik)
R
_
;k
=
1
2
g
i
, (2.23a)
where g
i
, a non-linear quantity, can be seen as a complementary tor-
sional current:
g
i
=
1
2
J
k
F
ik
+
i
=
n
[kl]
F
k
n
F
li
2
i
. (2.23b)
In the most general case, by the way, the Ricci tensor is asymmetric.
If we proceed further, the generalized Bianchi identity and its contracted
form will be given by
R
ijkl;m
+R
ijlm;k
+R
ijmk;l
=
= 2
_
r
[kl]
R
ijrm
+
r
[lm]
R
ijrk
+
r
[mk]
R
ijrl
_
,
(2.24a)
_
R
ik
1
2
g
(ik)
R
_
;i
= 2g
(ik)
r
[ji]
R
j
r
+
r
[ij]
R
ijk
...r
. (2.24b)
Remark 2
Consider a uniform charge density. Again, our resulting eld equation
(2.21) reads
_
R
ik
1
2
g
(ik)
R
_
;k
=
1
4
F
i
k
J
k
+
1
4
g
(ik)
_
F
r
s
_
F
s
r;k
F
s
k;r
__
,
where we have set
i
=
1
4
g
(ik)
_
F
r
s
_
F
s
r;k
F
s
k;r
__
. Recall the Lorentz
equation of motion: m
0
c
2 Du
i
Ds
=eF
i
k
u
k
. Setting =
1
4
_
e
_
, we obtain
m
0
c
2
Du
i
Ds
=
1
__
R
ik
1
2
g
(ik)
R
_
;k
i
_
or
m
0
c
2
Du
i
Ds
=
1
__
R
ik
1
2
g
(ik)
R
_
;k
2g
(ik)
s
[rk]
R
r
s
_
.
In the absence of charge density (when the torsion tensor is zero),
i.e., in the limit , we get the usual geodesic equation of motion
of General Relativity:
d
2
x
i
ds
2
+
_
i
jk
_
dx
j
ds
dx
k
ds
= 0 .
The eld equation can always be brought into the form
R
ik
= k
_
T
ik

1
2
g
(ik)
T
_
.
Here the energy-momentum tensor need not always vanish outside
the world-tube and in general we can write T
ik
;k
=
1
2
g
i
. Now the Ein-
stein tensor is
G
ik
= R
ik

1
2
g
(ik)
R =
1
4
F
il
F
l
k

1
8
g
(ik)
F
rs
F
rs
= T
ik
.
The right-hand side stands more appropriately as the eld strength
rather than the classical conservative source term as T =
1
2
F
ik
F
ik
(again, is a coupling constant). The equation of gravoelectrodynamics
can immediately be written in the form
R
ik

1
4
g
(ik)
R =
1
4
_
F
il
F
l
k

1
4
g
(ik)
F
rs
F
rs
_
= T
ik
,
T
ik
= F
il
F
l
k

1
4
g
(ik)
F
rs
F
rs
as expected. In this eld equation, as can be seen, the material density
arises directly from electromagnetic interaction.
End of Remark 2
The (sub-)spaces
n
= R
4
n and V
6
=
n
m. The
n
-covariant
derivative
We now consider the space
n
= R
4
n IR
5
, a sub-space of IR
5
with
basis
A
satisfying
A
= (g
i
, n). Therefore this basis spans a special
coordinate system in IR
5
. We dene the
n
-covariant derivative to be
a projective derivative which acts upon an arbitrary vector eld of the
form =
_
i
,
_
or, more generally, upon an arbitrary tensor eld of
the form T =
_
T
ij...
kl...
, t
i...
kl...
(1)
, t
j...
kl...
(2)
, . . . , t
ij...
k...
(N1)
, t
ij...
l...
(N)
_
, where t
i...
kl...
(1)
= T
i5...
kl...
,
t
j...
kl...
(2)
= T
5j...
kl...
etc., and which is projected onto the four-dimensional
physical space-time R
4
. For instance, for =
i
g
i
+n then
j
=
i
;j
g
i
+
i
g
i;j
+
;j
n +n
;j
=
=
i
;j
g
i
+
1
2

i
F
ij
n +
;j
n
1
2
F
i
j
g
i
=
=
_
i
;j

1
2
F
i
j
_
g
i
+
_
;j
+
1
2
F
ij
i
_
n.
The projection of this onto R
4
manifests in
i
|j
=
i
;j

1
2
F
i
j
.
The stroke [ represents the
n
-covariant derivative which takes up
the notion of cylindricity. When this is applied, for instance, to a second
rank tensor eld T, we have T
ij
|k
=T
ij
;k

1
2
t
i
(1)
F
j
k
1
2
t
j
(2)
F
i
k
. For a tensor
of arbitrary rank, we therefore have
T
ij...
kl...|m
= T
ij...
kl...;m

1
2
t
i...
kl...
(1)
F
j
m

1
2
t
j...
kl...
(2)
F
i
m

1
2
t
ij...
k...
(N1)
F
lm

1
2
t
ij...
l...
(N)
F
km
. . . .
(2.25)
So the electromagnetic eld extends the covariant derivative. To pass
on from General Relativity to our unied eld theory we merely need to
replace ordinary (horizontal) covariant derivatives with
n
-covariant
derivatives. Recall that the equation of geodesic motion in General
Relativity is u
i
;j
u
j
=0 where u
i
=dx
i
/ds is of course the unit velocity
vector. Generalizing by letting u=(u
i
, ) and setting u
i
;j
u
j
u
i
|j
u
j
=0,
we have u
i
|j
u
j
=
_
u
i
;j

1
2
F
i
j
_
u
j
= 0.
Setting =2
_
e
m
0
c
2
_
where, again, e is the electric charge, m
0
is the
point-mass and c is the speed of light in vacuum, we have the Lorentz
equation of motion:
Du
i
Ds
=
e
m
0
c
2
F
i
j
u
j
,
Du
i
Ds
u
i
;j
u
j
.
The fth component of the momentum is consequently given by
p
5
= 2
_
e
c
2
_
. (2.26)
We will now show that the quantity is indeed constant along the
world-line. First we write u=u
i
g
i
+ n , then, as before, we have
j
u=
_
u
i
;j

1
2
F
i
j
_
g
i
+
_
;j
+
1
2
F
ij
u
i
_
n.
Applying the law of parallel transport in
n
=R
4
n, i.e., 4
u
u=0
(in the direction of
4
u) where
4
u represents the ordinary tangent four-
velocity eld, we get two equations of motion:
_
u
i
;j

1
2
F
i
j
_
u
j
= 0 , (2.27a)
_
,j
+
1
2
F
ij
u
i
_
u
j
= 0 . (2.27b)
The rst of these is just the usual Lorentz equation in a general co-
ordinate system, the one weve just obtained before using the straight-
forward notion of vertical-horizontal
n
-covariant derivative. Mean-
while, the second reads, due to the vanishing of its second term:
d
ds
=0,
which establishes the constancy of with respect to the world-line.
Therefore it is justied that forms a fundamental constant of Nature
in the sense of a correct parameterization. We also have
d e
ds
=
dm
0
ds
=0.
Again, theres a certain possibility for the electric charge and the mass,
to vary with time, perhaps slowly in reality. We shall now consider the
unit spin vector eld in the spin space S
p
:
4
v u
i
i
= g
[ik]
u
i
g
k
, (2.28)
which has been dened in Section 1.2. This spin (rotation) vector is
analogous to the ordinary velocity vector in the spin space representa-
tion. For the moment, let v =
_
v
i
,
_
; v
i
=g
[ik]
u
k
where v =v
i
i
+n.
Therefore we see that
j
v = u
i
;j
g
i
+u
i
g
i;j
+
;j
n +n
;j
=
= v
i
;j
i
+v
i
i;j
+
;j
n +n
;j
=
=
_
v
i
;j

1
2
g
[ik]
F
kj
_
i
+
_
,j
+
1
2
g
[kl]
F
l
j
v
k
_
n,
(2.29)
with the help of (1.7). If the law of parallel transport 4
u
u=0 applies
for the velocity eld u, it is intuitive that in the same manner it must
also apply to the spin eld v:
4
u
v = 0 . (2.30)
This states that spin is geometrically conserved. We then get
_
v
i
;j

1
2
g
[ik]
F
kj
_
u
j
= 0 , (2.31a)
_
,j
+
1
2
g
[kl]
F
l
j
v
k
_
u
j
= 0 , (2.31b)
which are completely equivalent to the equations of motion in (2.27a)
and (2.27b).
Lets also observe that
G
ik
|k
= G
ik
;k
, (2.32a)
F
ij|k
+F
jk|i
+F
ki|j
= F
ij;k
+F
jk;i
+F
ki;j
, (2.32b)
i.e., the vertical-horizontal
n
-covariant derivative operator when ap-
plied to the Einstein tensor and the electromagnetic eld tensor equals
the ordinary covariant derivative operator. We shall be able to prove
this statement. First G
ik
;j
G
ik
| j
=G
ik
;j

1
2
X
i
F
k
j

1
2
Y
k
F
i
j
where
X
i
=G
i5
=Y
i
(due to the symmetry of the tensor G
ik
, G
5i
=R
i5
), so
that G
ik
|k
=G
ik
;k

1
2
Y
k
F
i
k
. Now the ve-dimensional curvature tensors
(the Riemann and Ricci tensors) in
n
are
R
A
BCD
=
A
BD,C

A
BC,D
+
E
BD
A
EC

E
BC
A
ED
,
R
AB
=
C
AB,C

C
AC,B
+
E
AB
C
EC

E
AC
C
EB
.
In this special coordinate system we have
k
5i
= g
k
(
i
n) =
1
2
g
k
F
j
i
g
j
=
1
2
F
k
i
,
5
5i
= n (
i
n) =
1
2
n F
k
i
g
k
= 0 =
5
i5
=
k
55
,
k
5k
= g
k
(
k
n) =
1
2
g
k
F
j
k
g
j
=
1
2
F
j
j
= 0 =
k
k5
,
R
5i
=
k
5i,k

k
5k,i
+
k
5i
l
kl

l
5k
k
li
=
=
1
2
_
F
k
i,k
+
k
lk
F
l
i

k
li
F
l
k
_
=
=
1
2
_
F
k
i;k
+ 2
l
[ik]
F
k
l
_
=
=
1
2
(J
i
+J
i
) = 0
with the help of (2.11d). Hence G
ik
|k
=G
ik
;k
. (2.32b) can also be easily
proven this way.
As a brief digression, consider a six-dimensional manifold V
6
=
n
m
where m is the second normal coordinate with respect to R
4
. Let
i
g
[5i]
, a
5
, b
6
,
i
k

_
i
5k
_
,
i
k

_
i
6k
_
, =g
[56]
, and
i
g
[6i]
. Casting (1.12) and (1.22) into six dimensions, the electro-
magnetic eld tensor can be written in terms of the fundamental spin
tensor as the following equivalent expressions:
F
ik
=
5
4
_
r
_
g
[ir],k
g
[kr],i
_
(
i,k

k,i
)
_
=
=
5
4
_
r
g
[ik],r
(
i,k

k,i
)
_
,
(2.33a)
F
ik
=
l
_
g
[li],k
g
[ik],l
+g
[kl],i
_
a
_
i,k

k,i
_
b (
i,k

k,i
)
2
l
_
g
[im]
_
m
lk
_
g
[km]
_
m
li
__
2a
_
g
[im]
m
k
g
[km]
m
i
_
2b
_
g
[im]
m
k
g
[km]
m
i
_
.
(2.33b)
Since the basis in this space is given by g
=
_
g
i
, n, m
_
, the funda-
mental tensors are
g
()
=
_
_
_
_
g
(ik)
4x4
0 0
0 1 0
0 0 1
_
_
_, (2.34a)
g
[]
=
_
_
_
_
_
g
[ik]
4 x 4

i

i
0
i

i
0
_
_
_
_
. (2.34b)
Remark 3 (on the modied Maxwells equations)
Using (2.25) we can now generalize Maxwells eld equations through
the new extended electromagnetic tensor F
ik
=
i|k

k|i
where
F
ik
(old)
=
i;k

k;i
=
i,k

k,i
+ 2
l
[ik]
l
= F
ik
+ 2
l
[ik]
l
,
where
i|k
=
i;k

1
2
F
ik
. Now
5
= is taken to be an extra scalar
potential. Therefore F
ik
=
i,k

k,i
F
ik
or
F
ik
=
1
1 +
(
i,k

k,i
) (
i,k

k,i
) . (2.a)
For instance, the rst pair of Maxwells equations can therefore be
generalized into
E =
1
1 +
_
1
c

A
t

_
, (2.b
1
)
B =
1
1 +
x

A, (2.b
2
)
div

B =
1
(1 +)
2
_

x

A
_
+
1
1 +

x

A, (2.b
3
)
curl
_
1
1 +
E
_
=
1
c
t
1
1 +
B
1
c
B
t
1
1 +
, (2.b
4
)
where

A is the three-dimensional electromagnetic vector potential:
A= (A
a
), is the electromagnetic scalar potential,

E is the electric
eld,

B is the magnetic eld and

is the three-dimensional (curvilin-
ear) gradient operator and
2
=4. Here we dene the electric
and magnetic elds in such a way that
F
4a
= (1 +)
1
E
a
, F
12
= (1 +)
1
B
3
,
F
31
= (1 +)
1
B
2
, F
23
= (1 +)
1
B
1
.
We also note that in (2.b
3
) the divergence

x
A is in general
non-vanishing when torsion is present in the three-dimensional curved
sub-space. Direct calculation gives
div curl

A =
1
2

abc
_
R
d
cab
A
d
2
d
[ab]
A
c;d
_
,
(curl grad )
c
=
ab
.c

d
[ab]
,d
.
So the magnetic charge with density
m
in the innitesimal volume
d is given by
=
1
2
_
abc
_
R
d
cab
A
d
2
d
[ab]
A
c;d
_
d.
End of Remark 3
Chapter 3
SPIN-CURVATURE
3.1 Dynamics in the microscopic limit
We now investigate the microscopic dynamics of our theory. Lets in-
troduce an innitesimal coordinate transformation into
n
through the
dieomorphism
x
i
= x
i
+
i
with an external Killing-like vector =
_
i
,
_
:
i
=
i
(x), = (x)
(not to be confused with the internal Killing vector which describes
the internal symmetry of a particular conguration of space-time or
which maps a particular space-time onto itself). The function here
shall play the role of the amplitude of the quantum mechanical state
vector [). Recall that R
4
represents the four-dimensional physical
world and n is a microscopic dimension. In its most standard form
(x) Ce
(2i/h) (Etpr(x,y,z))
is the quantum mechanical scalar wave
function; h is the Planck constant, E is energy and p is the three-
momentum. Dene the extension of the space-time
n
=R
4
n by
ij
=
1
2
D
g
(ij)
. (3.1)
We would like to express the most general symmetry, rst, of the
structure of
n
and then nd out what sort of symmetry (expressed
in terms of the Killing-like vector) is required to describe the non-local
statics or non-deformability of the structure of the metric tensor
g (the lattice arrangement). Our exterior derivative is dened as the
variation of an arbitrary quantity with respect to the external eld .
Unlike the ordinary Killing vector which maps a space-time onto itself,
the external eld and hence also the derivative D
g
(ij)
map R
4
onto,
say, R
4
which possesses a deformed metrical structure of g R
4
, g
.
We calculate the change in the metric tensor with respect to the ex-
ternal eld, according to the scheme g
(ij)
= (g
i
g
j
) g
(ij)
=
_
g
i
g
j
_
,
as follows:
D
g
(ij)
= (g
i
D
g
j
) + (D
g
i
g
j
) ; D
g
i
= g
i
g
i
=
i
.
The Lie derivative D
denotes the exterior change with respect to the

innitesimal exterior eld, i.e., it dynamically measures the deformation
Chapter 3 Spin-Curvature 45
of the geometry of the space-time R
4
.
Thus
D
g
(ij)
= (g
i

j
) + (
i
g
j
) ,
where
i
=
_
k
;i

1
2
F
k
i
_
g
k
+
_
,i
+
1
2
F
ki
k
_
n.
By direct calculation, we thus obtain
D
g
(ij)
=
i;j
+
j;i
. (3.2)
3.2 Spin-curvature tensor of S
p
As an interesting feature, we point out that the change D
g
(ij)
in the
structure of the four-dimensional metric tensor does not involve the
wave function . The space-time R
4
will be called static if g does
not change with respect to . The four-dimensional (but not the ve-
dimensional) metric is therefore static whenever
i;j
+
j;i
=0.
Now let R
4
be an innitesimal copy of R
4
. To arrive at the lattice
picture, let also R
4
, R
4
, R
4
, . . . be n such copies of R
4
. Imagine the
space S
n
consisting of these copies. This space is therefore populated
by R
4
and its copies. If we assume that each of the copies of R
4
has the
same metric tensor as R
4
, then we may have
= (0, ) .
We shall call this particular fundamental symmetry normal sym-
metry or spherical world-symmetry. Then the n copies of R
4
exist
simultaneously and each history is independent of the four-dimensional
external eld
4
and is dependent on the wave function only. In other
words, the special lattice arrangement = (0, ) gives us a condition
for R
4
and its copies to co-exist simultaneously independently of how
the four-dimensional external eld deforms their interior metrical struc-
ture. Thus the many sub-manifolds nR
4
represent many simultaneous
realities which we call world-pictures. More specically, at one point in
R
4
, there may at least exist two world-pictures. In other words, a point
seen by an observer conned to lie in R
4
may actually be a line or curve
whose two end points represent the two solutions to the wave function
. The splitting of R
4
into its copies occurs and can only be perceived
on the microscopic scales with the wave function describing the en-
tire process. Conversely, on the macroscopic scales the inhabitants of
n
=R
4
n may perceive the collection of the nR
4
(sub-spaces of R
4
)
representing a continuous four manifold R
4
=R
4
R
4
R
4
. . . There
may be an innite number of nR
4
composing the space-time R
4
, pos-
sessing the same uctuating metric tensor g, i.e., that of R
4
(the total
space). On the microscopic scales, uctuations in the metric do occur.
In special cases, the cylinder condition may prevent the topology of the
spaces nR
4
to change which cannot be perceived directly by an external
observer in R
4
for it takes place along the microscopic coordinate y. The
uctuations induce the many dierent world-pictures. Our task now is
to nd the wave equation describing the entire process. Now, with the
help of (1.7) and (1.8) we can write
i
= (
i
g
j
) g
j
+ (
i
n) n =
=
1
2
((
i
g
j
) + (
j
g
i
)) g
j
+
+
1
2
((
i
g
j
) (
j
g
i
)) g
j
+ (
i
n) n =
=
1
2
D
g
(ij)
g
j
_
(
j
n)
j
, g
i
+
+g
[jk]
(
j
n) C
l
ki
g
l
+
3
2
g
[kl]
(
k
g
l
) g
[ij]
g
j
3
2

kl
ij
(
k
g
l
) g
j
=
=
1
2
D
g
(ij)
g
j
_
3
2
g
[kl]
(
k
g
l
) n (
j
n)
j
_
, g
i
+
+g
[jk]
(
j
n) C
l
ki
g
l

3
2

kl
ij
(
k
g
l
) g
j
.
(3.3)
Setting
S =
3
2
g
[kl]
(
k
g
l
) n (
j
n)
j
,
X
i
= g
[jk]
(
j
n) C
l
ki
g
l

3
2

kl
ij
(
k
g
l
) g
j
,
we have
i
=
1
2
D
g
(ij)
g
j
+ [S, g
i
] +X
i
. (3.4)
To see that the vector S represents a spin (rotation) vector is not
dicult as we know that (g
i
n) =0 and hence D(g
i
n) =0 where the
change in n can be represented by an internal rotation: Dn= [S, n].
With the help of (1.8), we now have
(g
i
Dn) = (g
i
[S, n]) = (S [n, g
i
]) = (S
i
) = S
i
.
Therefore
S
i
= (S
i
) = (
i
n) = (n
i
) .
Now
S = S
i
i
+
3
2
g
[ij]
(
i
g
j
) n =
= S
i
i
+n,
(3.5)
S
i
= (n
i
) =
,i

1
2
F
k
i
k
, (3.6a)
=
3
2
g
[ij]
(
i
g
j
) =
=
3
2
g
[ij]
_
j;i

1
2
F
ji
_
.
(3.6b)
We can also calculate the exterior variation of the electromagnetic
eld tensor F
ik
:
D
F
ik
= 2D
(
i
n g
k
) =
= 2 (
i
D
n g
k
) + 2 (
i
n D
g
k
) =
= 2 (
i
[S, n] g
k
) + 2 (
i
n D
g
k
) =
= 2 (
i
S
k
) 2
_
_
S,
1
2
F
j
i
g
j
g
k
_
_
F
j
i
g
j

k

_
.
(3.7a)
Hence we can write
D
F
ik
= 2 (
i
S
k
)
_
F
j
i
g
j
(
k
[S, g
k
])
_
=
= 2 (
i
S
k
) F
j
i
_
g
j

1
2
D
g
(kl)
g
l
_
F
j
i
(g
j
X
k
) =
=
_
2H
ik
+
1
2
F
j
i
D
g
(jk)
+F
j
i
X
jk
_
,
(3.7b)
where we have just dened the spin-curvature tensor H
ik
:
H
ik
= (
i
S
k
) , (3.8)
which measures the internal change of the spin in the direction of the
spin basis. We further posit that the spin-curvature tensor satises
the supplementary identities (which are deduced from the conditions
4
u
u=0 and 4
u
S =0)
H
ik
u
i
= 0 , (3.9a)
tr H = 0 . (3.9b)
The transverse condition (3.9a) reproduces the Lorentz equation of
motion while (3.9b) describes the internal properties of the structure of
physical elds which corresponds to the quantum limit on our manifold
(for details see Section 4.2).
The explicit expression of H
ik
can be found to be
H
ik
= (
i
S
k
) =
=
1
2
g
[kl]
F
l
i
S
k;i
(3.10)
from (3.6a) and (3.6b). Furthermore, still with the help of (3.6a) and
(3.6b), we obtain, after some simplications,
H
ik
=
;k;i
+
1
4
F
l
i
(
l;k

k;l
)
1
4
F
l
i
F
lk
+
1
2
_
F
l
k
l
_
;i
. (3.11)
However, we recall that
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) , R
ik
=
1
4
F
il
F
l
k
and therefore we obtain the spin-curvature relation in the form
H
ik
=
;k;i
R
ik
+
1
4
F
l
i
(
l;k

k;l
) +
1
2
_
F
l
k
l
_
;i
. (3.12)
3.3 Wave equation describing the geometry of R
4
If we contract (3.12) with respect to the indices i and k, we have
H = (2 R) +F
ik
i;k
+
1
2
J
i
i
, (3.13)
where 2 is the covariant four-dimensional Laplacian, again, R is the
curvature scalar and J
i
is the current density vector. However, using
(3.9a) and (3.9b) and associating with the space
n
=R
4
n the fun-
damental world-symmetry = (0, ), then we obtain, from (3.12), the
equation of motion:
;k;i
= H
ik
+R
ik
. (3.14)
From (3.9a), (3.9b) and (3.14), we obtain the wave equation:
(2 R) = 0 . (3.15a)
This resembles the scalar Klein-Gordon wave equation except that
we have the curvature scalar R in place of M
2
= (m
0
c/)
2
(we normally
expect this in generalizing the scalar Klein-Gordon equation). Note also
that the ordinary Klein-Gordon and Dirac equations do not explicitly
contain any electromagnetic terms. This means that the electromag-
netic eld must somehow already be incorporated into gravity in terms
of M. Since is just the amplitude of the state vector [), we can also
write
(2 R) [) = 0 . (3.15b)
If the curvature scalar vanishes, there is no source (or actually, no
electromagnetic eld strength) and we have 2 [) =0 which is the wave
equation of massless particles.
Remark 4
Recall (3.7):
D
F
ik
=
_
2H
ik
+
1
2
F
j
i
D
g
(jk)
+F
j
i
X
jk
_
, (3.a)
where
D
g
(ik)
=
i;k
+
k;i
= 2
ik
,
H
ik
= (
i
S
k
) .
Meanwhile, for an arbitrary tensor eld T, we have in general
D
T
ij...
kl...
=
m
T
ij...
kl...;m
+T
ij...
ml...
m
;k
+T
ij...
km...
m
;l
+
T
mj...
kl...

i
;m
T
im...
kl...

j
;m
. . .
Therefore
D
F
ik
=
l
F
ik;l
+F
lk

l
;i
+F
il
l
;k
. (3.b)
Comparing this with (3.a), we have, for the spin-curvature ten-
sor H
ik
,
H
ik
=
1
2
_
l
F
ik;l
+F
lk
l
;i
+F
il
l
;k
_
1
4
F
l
i
(
l;k
+
k;l
)
1
2
F
l
i
X
lk
(3.c)
in terms of the electromagnetic eld tensor.
End of Remark 4
Finally, lets dene the following tensor:
A
ij
S
i;j

1
2
F
k
j
k|i
, (3.16)
where, as before,
k |i
=
k;i

1
2
F
ki
is the
n
-covariant derivative of
k
, the notion of which we have devel-
oped in Section 2.1 of this work, and
S
i
=
,i

1
2
F
k
i
k
is the spin vector (3.6a). The meaning of the tensor (3.16) will become
clear soon. It has no classical analogue. We are now in a position
to decompose (3.16) into its symmetric and alternating parts. The
symmetric part of (3.16):
A
(ij)
=
1
2
_
S
i;j
+S
j;i

1
2
F
k
i
k|j

1
2
F
k
j
k|i
_
(3.17a)
may be interpreted as the tension of the spin eld.
Now its alternating part:
A
[ij]
=
1
2
_
S
i;j
S
j;i
+
1
2
F
k
i
k|j

1
2
F
k
j
k|i
_
(3.17b)
represents a non-linear spin eld. (However, this becomes linear when
we invoke the fundamental world-symmetry = (0, ).) If we employ
this fundamental world- symmetry, (3.17a) and (3.17b) become
A
(ij)
=
1
2
(S
i;j
+S
j;i
) +R
ij
, (3.18a)
A
[ij]
=
1
2
(S
i;j
S
j;i
) . (3.18b)
With the help of (3.6a) and the relation
;i;k

;k;i
= 2
r
[ik]
,r
,
(3.18b) can also be written
A
[ij]
=
k
[ij]
,k
. (3.19)
Now
A
ij
=
;i;j

1
2
_
F
k
i
k
_
;j

1
2
F
k
j

k;i
+
1
4
F
k
j
F
ki
=
=
_
;i;j
R
ij
+
1
2
_
F
k
i
k
_
;j
+
1
2
F
k
j

k;i
_
.
Therefore the Ricci tensor can be expressed as
R
ik
=
1
2

1
F
r
k

r|i

1
_
;i;k
+S
i;k
+
1
2
(F
r
i
r
)
;k

1
2
F
r
k

r;i
_
.
(3.20)
We can still obtain another form of the wave equation of our quan-
tum gravity theory. Taking the world-symmetry = (0, ), we have,
from r =r
n,
g
i
= h
i

,i
n +
1
2
F
r
i
g
r
, (3.21)
where h
i
r
,i
is the basis of the space-time R
4
. Now the metric tensor
of the space-time R
4
is
g
(ik)
= (g
i
.g
k
) =
= h
ik

,k
i
+
1
2
F
r
k
ir

,i
k
+
,i
,k
+
+
1
2
F
r
i
kr
+
1
4

2
g
(rs)
F
r
i
F
s
k
,
where h
ik
(h
i
h
k
) is the metric tensor of R
4
,
i
(h
i
n) and
ik
(h
i
g
k
). Direct calculation shows that
i
=
,i
,
ik
= g
(ik)
+
1
2
F
ik
.
Then we arrive at the relation
g
(ik)
= h
ik

,i
,k

1
4

2
g
(rs)
F
r
i
F
s
k
. (3.22)
Now from (3.2) we nd that this is subject to the condition
D
g
(ij)
= 0 . (3.23)
Hence we obtain the wave equation
,i
,k
=
1
4

2
g
(rs)
F
r
i
F
s
k
. (3.24a)
Expressed in terms of the Ricci tensor, the equivalent form of
(3.24a) is
,i
,k
=
2
R
ik
. (3.24b)
Expressed in terms of the Einstein tensor G
ik
=R
ik
1
2
g
(ik)
R,
(3.24b) becomes
_
r
i
s
k

1
2
g
(ik)
g
(rs)
_
,r
,s
=
2
G
ik
. (3.25)
If in particular the space-time R
4
has a constant sectional curvature,
then R
ijkl
=
1
12
_
g
(ik)
g
(jl)
g
(il)
g
(jk)
_
and R
ik
=g
(ik)
, where
1
4
R
is constant, so (3.24b) reduces to
,i
,k
=
2
g
(ik)
. (3.26)
Any axisymmetric solution of (3.26) would then yield equations that
could readily be integrated, giving the wave function in a relatively
simple form. Multiplying now (3.24b) by the contravariant metric tensor
g
(ik)
, we have the wave equation in terms of the curvature scalar as
follows:
g
(ik)
,i
,k
=
2
R. (3.27)
Finally, lets consider a special case. In the absence of the scalar
source, i.e., in void, the wave equation becomes
g
(ik)
,i
,k
= 0 . (3.28)
This wave equation therefore describes a massless, null electromag-
netic eld where
F
ik
F
ik
= 0 .
In this case the electromagnetic eld tensor is a null bivector. There-
fore, according to our theory, there are indeed seemingly void regions
in the Universe that are governed by null electromagnetic elds only.
Chapter 4
ADDITIONAL CONSIDERATIONS
4.1 Embedding of generalized Riemannian manifolds (with
twist) in N =n +p dimensions
In connection to Chapter 2 of this work where we considered an embed-
ding of class 1, we outline the most general formulation of embedding
theory of class p in N =n+p dimensions where n now is the number
of dimensions of the embedded Riemannian manifold. First, let the em-
bedding space R
N
be an N-dimensional Riemannian manifold spanned
by the basis e
A
. For the sake of generality we take R
N
to be an
N-dimensional space-time. Let also R
n
be an n-dimensional Rieman-
nian sub-manifold (possessing torsion) in R
N
spanned by the basis g
l
where now the capital Latin indices A, B, . . . run from 1 to N and

the ordinary ones i, j, . . . from 1 to n. If now g
AB
= (e
A
e
B
) and
g
ij
= (g
i
g
j
) denote the metric tensors of R
N
and R
n
, respectively, and
if we introduce the p-unit normal vectors (also called n-legs) n
()
(where
the Greek indices run from 1 to p and summation over any repeated
Greek indices is explicitly indicated otherwise there is no summation),
then
g
ij
= e
A
i
e
B
j
g
AB
,
g
AB
= e
i
A
e
j
B
g
ij
+
n
()
A
n
()
B
,
e
A
i
n
()
A
= 0 ,
(n
()
n
()
) =
()
,
()
= 1 ,
g
i,j
=
k
ij
g
k
+
()
()
ij
n
()
,
g
i;j
=
()
()
ij
n
()
,
e
A,B
=
C
AB
e
C
,
k
ij
= e
k
A
e
A
i,j
+e
k
A
A
BC
e
B
i
e
C
j
,
A
BC
= e
A
i
e
i
B,C
+e
A
k

k
ij
e
i
B
e
j
C
+
()
()
ij
e
i
B
e
j
C
n
()A
+
+
n
()A
n
()
B,C

()
()
kj
g
ki
e
A
i
e
j
C
n
()
B
,
()
ij
= n
()
A
e
A
i,j
+n
()
A

A
BC
e
B
i
e
C
j
,
e
A
i;j
=
()
()
ij
n
()A
,
e
A
i,j
= e
A
k

k
ij

A
BC
e
B
i
e
C
j
+
()
()
ij
n
()A
.
Now since e
A
=e
i
A
g
i
+
()
n
()
A
n
()
and
()
ij
= e
A
i
n
()
A;j
for the
asymmetric p-extrinsic curvatures, we see that
e
i
B
()
ij
=
_
A
B

()
n
()A
n
()
B
_
n
()
A;j
=
= n
()
B;j
+
()
n
()
A;j
n
()A
n
()
B
.
Lets dene the p-torsion vectors by
i
= n
()
A;i
n
()A
,
i
=
i
.
Hence
n
()
A;i
=
()
ki
e
k
A
+
()
i
n
()
A
or
n
()
;i
=
()
ki
g
k
+
()
i
n
()
.
Now
n
()
A;i;k
=
()
ik
e
i
A

g
rs
()
ri

()
sk
n
()
A
+
()
i;k
n
()
A

i

()
rk
e
r
A
+
()
k
n
()
A
.
Hence we obtain the expression
n
()
A;i;k
n
()A
= g
rs
()
ri

()
sk
+
i;k
+
()
k
.
Chapter 4 Additional Considerations 55
Meanwhile,
n
()
A;i
= n
()
A;B
e
B
i
,
n
()
A;i;k
=
_
n
()
A;B
e
B
i
_
;C
e
C
k
=
= n
()
;B;C
e
B
i
e
C
k
+n
()
A;B
e
A
i;k
=
= n
()
A;B;C
e
B
i
e
C
k
+
n
()
A;B

()
()
ik
n
()B
,
but
n
()
A;B
= n
()
A;i
e
i
B
=
()
ki
e
k
A
e
i
B
.
Hence
n
()
A;B
n
()B
= 0 ,
n
()
A;B
n
()A
= 0 .
We also see that
n
()
A;i;k
n
()A
= n
()
A;B;C
n
()A
e
B
i
e
C
k
.
Consequently, we have
_
n
()
A;i;k
n
()
A;k;i
_
n
()A
=
_
n
()
A;B;C
n
()
A;C;B
_
n
()A
e
B
i
e
C
k
=
= R
ABCD
n
()A
n
()B
e
C
i
e
D
k
.
On the other hand, we see that
_
n
()
A;i;k
n
()
A;k;i
_
n
()A
=
i;k

k;i
+g
rs
_
()
rk

()
si

()
ri

()
sk
_
+
+
()
(
i
) +
+R
ABCD
n
()A
n
()B
e
C
i
e
D
k
.
Combining the last two equations we get the Ricci equations:
i;k

k;i
= g
rs
_
()
ri

()
sk

()
rk

()
si
_
()
(
i
) .
Now from the relation
g
i,j
=
k
ij
g
k
+
()
()
ij
n
()
,
we obtain the expression
g
i,jk
=
_
r
ij,k
+
s
ij
r
sk

g
rs
()
()
ij

()
sk
_
g
r
+
+
()
()
rk

r
ij
+
()
()
ij,k
+
()
()
ij

k
_
n
()
.
Hence consequently,
g
i,jk
g
i,kj
=
_
R
r
ijk
+
g
rs
()
_
()
sj

()
ik

()
sk

()
ij
_
_
g
r
+
+
()
_
()
ij;k

()
ik;j
+ 2
r
[jk]
()
ir
_
n
()
+
+
()
_
()
ij

k

()
ik

j
_
n
()
.
On the other hand, we have
g
i,jk
=
_
A
BC,D
+
E
BC
A
ED
_
e
B
i
e
C
j
e
D
k
e
A
+
+
_
e
A
i,jk
+
A
BC
e
B
i,j
e
C
k
+
A
BC
e
B
i,k
e
C
j
+
A
BC
e
B
i
e
C
j,k
_
e
A
,
g
i,jk
g
i,kj
=
_
R
A
BCD
e
B
i
e
C
j
e
D
k
+S
A
ijk
_
e
A
,
where, just as in Chapter 2,
S
A
ijk
= e
A
i,jk
e
A
i,kj
+
_
e
C
j,k
e
C
k,j
_
A
BC
e
B
i
.
Combining the above, we generalize the Gauss-Codazzi equa-
tions into
R
ijkl
=
()
_
()
ik

()
jl

()
il

()
jk
_
+R
ABCD
e
A
i
e
B
j
e
C
k
e
D
l
S
Ajkl
e
A
i
,
()
ij;k

()
ik;j
= R
ABCD
n
()
e
B
i
e
C
j
e
D
k
+S
Aijk
n
()A
2
r
[jk]
()
ir
+
+
()
_
()
ij

k

()
ik

j
_
.
Finally, when R
n
is embedded isometrically in R
N
, i.e., when
the embedding manifold R
N
is an Euclidean or pseudo-Euclidean
N-dimensional space(-time) or if we impose a particular integrability
condition on the n-vectors in R
n
the way we derived (2.10) in Chap-
ter 2, we have the system of equations
R
ijkl
=
()
_
()
ik

()
jl

()
il

()
jk
_
,
()
ij;k

()
ik;j
= 2
r
[jk]
()
ir
+
()
_
()
ij

k

()
ik

j
_
,
i;k

k;i
= g
rs
_
()
ri

()
sk

()
rk

()
si
_
()
(
i
) .
Turning to physics, a unied eld eld theory of gravity and electro-
magnetism may be framed by the geometric quantities above either in
5 = 4+1 background dimensions or, if we wish to extend it, in N = 4+p
background dimensions. As an alternative, in N = 4 +p dimensions we
may also observe that the following assumptions can be made possible:
1. The rst alternating p-exterior curvature
(1)
[ik]
is equivalent to the
ordinary electromagnetic eld tensor F
ik
;
2. The remaining p-exterior curvatures
(2)
[ik]
, . . . ,
(p)
[ik]
,
(1)
(ik)
, . . . ,
(p)
(ik)
represent elds beyond the known electromagnetic and gravita-
tional elds;
3. As in Chapter 2, if desired, the four-dimensional metric tensor g
ij
may not depend on the extra p-coordinates
_
x
N4
, x
N3
, . . . , x
N
_
.
Thus the cylinder condition represented by the equations
n
()
A;B
+n
()
B;A
= 0 or
()
(ij)
= 0
is, again, arrived at.
4.2 Formulation of our gravoelectrodynamics by means of
the theory of distributions. Massive quantum electro-
magnetic eld tensor
In Chapter 2 we have assumed a type of parallel transport applied
to the pseudo-ve-dimensional velocity eld u
_
u
A
_
=u
i
g
i
+ n in
n
=R
4
n, i.e., 4
u
u=0 where
4
u=u
i
g
i
; u
i
=
dx
i
ds
and =
2e
m
0
c
2
=u
A
n
A
;
d
ds
=0. To describe non-diverging point-like objects which may experi-
ence no change in energy even when accelerated, like electrons, we now
introduce a special kind of autoparallelism through the relation
i
u=0.
This says that the pseudo-ve-dimensional velocity eld u=(u
i
, ) is an
autoparallel vector eld in the sense of the theory of distributions, whose
magnitude is independent of the four coordinates of R
4
. In other words,
u
i
;k
=
1
2
F
i
k
,
,i
=
1
2
F
k
i
u
k
.
From these we have
Du
i
Ds
=
e
m
0
c
2
F
i
k
u
k
,
u
i
;i
= 0 ,
d
ds
= 0 .
Were now in a position to derive the eld equations of gravoelec-
trodynamics with the help of these relations. Afterwards, we shall show
that these relations, indeed, lead to acceptable equations of motion in
both four and ve dimensions. We may also emphasize that the ve-
dimensional space IR
5
is a Riemann space. First we note that
u
i;k;l
=
1
2
F
ik;l
+
1
2

,l
F
ik
=
=
1
2
F
ik;l

1
4
F
r
l
F
ik
u
r
.
Now
u
i;k;l
u
i;l;k
= R
r
ikl
u
r
2u
i;r
r
[kl]
=
=
1
2
(F
ik;l
F
il;k
) +
1
4
(F
r
k
F
il
F
r
l
F
ik
) u
r
.
Therefore we have
R
i
jkl
u
i
=
1
4
_
F
i
k
F
jl
F
i
l
F
jk
_
u
i
+
1
2
_
F
jk;l
F
jl;k
+ 2
r
[kl]
F
jr
_
.
But u
i
=
A
i
u
A
and =u
A
n
A
, so by means of symmetry, we can
lop o the u
A
:
R
i
jkl
A
i
=
1
4
_
F
i
k
F
jl
F
i
l
F
jk
_
A
i
+
1
2
_
F
jk;l
F
jl;k
+ 2
r
[kl]
F
jr
_
n
A
.
From this relation, we derive the unied eld equations
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) ,
F
ij;k
F
ik;j
= 2
r
[jk ,]
F
ir
as expected. We have thus no need to assume that the curvature of
the background ve-dimensional space IR
5
vanishes. To strengthen our
proof, recall that g
i;j
=
ij
n. Hence
g
i;j;k
=
ij;k
n +
ij
n
;k
=
=
ij;k
n
ij
r
k
g
r
,
g
i;j;k
g
i;k;j
= (
ij;k

ik;j
) n +
_
ik
r
j

ij
r
k
_
g
r
.
However,
g
i;j;k
g
i;k;j
= R
r
ijk
g
r
2
r
[jk]
n
ir
.
Combining the above, we have
R
ijkl
=
ik
jl

il
jk
.
ij;k

ik;j
= 2
r
[jk]
ir
.
Invoking the cylinder condition, again we get
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) ,
F
ij;k
F
ik;j
= 2
r
[jk]
F
ir
.
We now assume that the ve-dimensional equation of motion in IR
5
is in general not a geodesic equation of motion. Instead, we expect an
equation of motion of the form
d
2
x
A
ds
2
+
A
BC
dx
B
ds
dx
C
ds
=
F
A
B
u
B
,
where u
A
=e
A
i
u
i
, and
ds
2
=
5
ds
2
= g
(AB)
dx
A
dx
B
=
_
e
i
A
e
k
B
g
(ik)
+n
A
n
B
_
dx
A
dx
B
=
= g
(ik)
dx
i
dx
k
=
4
ds
2
,
A
BC
=
1
2
g
(DA)
_
g
(DB),C
g
(BC),D
+g
(CD),B
_
,
F
AB
F
AB
+
AB
,
F
AB
= (n
A;B
n
B;A
) ,
F
ik
= e
A
i
e
B
k
F
AB
,
F
AB
n
B
= 0 ,
AB
e
A
i
e
B
k
= 0 .
Note also that the connection transforms as
A
BC
= e
A
i
e
i
B,C
+e
A
i

i
jk
e
j
B
e
k
C
+
+
1
2
F
ik
e
i
B
e
k
C
n
A
+n
A
n
B,C

1
2
F
i
k
e
A
i
e
k
C
n
B
.
Now with the help of the relation
e
A
i,k
= e
A
r

r
ik

A
BC
e
B
i
e
C
k
+
1
2
F
ik
n
A
,
we have
d
2
x
A
ds
2
=
d
ds
_
e
A
i
dx
i
ds
_
= e
A
i,k
dx
ds
i
dx
ds
k
+e
A
i
d
2
x
i
ds
2
=
= e
A
i
d
2
x
i
ds
2
+e
A
i

i
jk
dx
j
ds
dx
k
ds

A
BC
dx
B
ds
dx
C
ds
.
Hence
d
2
x
A
ds
2
+
A
BC
dx
B
ds
dx
C
ds
= e
A
i
_
d
2
x
i
ds
2
+
i
jk
dx
j
ds
dx
k
ds
_
=
F
A
B
u
B
.
Setting
AB
=0 and =
e
m
0
c
2
, since F
A
B
e
i
A
u
B
=F
A
B
e
i
A
e
B
k
u
k
=F
i
k
u
k
,
we then obtain the equation of motion:
Du
i
Ds
=
e
m
0
c
2
F
i
k
u
k
.
A straightforward way to obtain the ve-dimensional equation of
motion is as follows: from our assumptions we have
_
e
A
i
u
A
_
;k
=
1
2
F
ik
,
F
ik
u
A
n
A
+u
A;B
e
A
i
e
B
k
=
1
2
F
ik
.
Therefore recalling that u
A
n
A
=0 and multiplying through by u
k
,
we get
u
A
;B
e
i
A
u
B
=
1
2
F
i
k
u
k
.
Again, noting that u
A
;B
u
B
=u
A
;k
u
k
=
Du
A
Ds
and multiplying through
by e
C
i
, we get
Du
C
Ds
n
A
Du
A
Ds
n
C
=
e
m
0
c
2
F
C
A
u
A
.
But
u
A
Dn
A
Ds
= n
A
Du
A
Ds
= n
A;B
u
A
u
B
= n
A;k
u
A
u
k
=
=
1
2
u
A
F
ik
e
i
A
u
k
=
=
1
2
F
ik
u
i
u
k
= 0 .
Therefore
Du
A
Ds
=
e
m
0
c
2
F
A
B
u
B
.
Lets now recall the spin-curvature tensor (3.8) as well as (3.5)
and (3.6):
H
ik
= (
i
S
k
) =
=
1
2
g
[kl]
F
l
i
S
k;i
,
S = S
i
i
+
3
2
g
[ij]
(
i
g
j
) n =
= S
i
i
+n,
S
i
= (n
i
) =
,i

1
2
F
k
i
k
,
=
3
2
g
[ij]
(
i
g
j
) =
=
3
2
g
[ij]
_
j;i

1
2
F
ji
_
.
Recall (3.9a):
H
ik
u
i
= 0 .
Multiplying (3.10) by u
i
, we get the equation of motion:
DS
k
Ds
=
1
2
g
[kr]
F
r
i
u
i
,
Equivalently,
DS
i
Ds
=
1
2
g
[ik]
F
kr
u
r
.
We can now compare this with (2.31a). Now this leads us to consider
a case in our theory in which the spin vector S
i
is normalized (in the
quantum limit). Then it is given by
S
i
= v
i
= g
[ik]
u
k
.
Multiplying once again by g
[ij]
, we get
g
[ij]
DS
i
Ds
=
1
2
F
jr
u
r
Du
i
Ds
=
1
2
F
i
k
u
k
,
which is the Lorentz equation of motion. In this case then it automati-
cally follows that
= 2
_
e
m
0
c
2
_
=,
d
ds
= 0 .
Corresponding to our ongoing analysis, lets also recall (3.9b):
tr H = 0 .
This gives the divergence equation
S
i
;i
=
1
2
g
[ik]
F
ik
or
g
[ik]
_
1
2
F
ik
u
i;k
_
= 0 .
On the other hand, from (3.6a), and by employing our fundamental
symmetry, we have
S
i
;i
= 2 .
Therefore, with the help of (3.15a), we see that
2 =
1
2
g
[ik]
F
ik
= R .
1
2
g
[ik]
F
ik
=
1
4
F
ik
F
ik
.
The simplest solution for massive, electrically charged particles of
this would then be
F
ik
= 2g
[ik]
.
which expresses the proportionality of the already quantized electro-
magnetic eld tensor to the fundamental spin tensor of our unied eld
theory. In other words,
F
ik
= 4
_
e
m
0
c
2
_
1
g
[ik]
.
Finally, with the help of (3.27), we get the fundamental quantum
relations (here we mean R ,= 0, ,= 0):
F
ik
= 4
_
e
m
0
c
2
_
Rg
[ik]
,
R =
1
2
.
Now, following (1.1), we can write our asymmetric fundamental ten-
sor
ik
as
ik
=
1
2
_
g
(ik)
+
1
4
_
m
0
c
2
e
_
F
ik
_
,
which satises (1.4):
ij
kj
=
k
i
. On the other hand, now we see that
ir
kr
=
1
2
_
k
i
+
1
4
_
m
0
c
2
e
_
2
2
R
k
i
_
,
ir
rk
=
1
2
_
k
i
+
1
2
_
m
0
c
2
e
_
F
k
i

1
4
_
m
0
c
2
e
_
2
2
R
k
i
_
.
Hence we get the following expression for the Ricci tensor:
R
ik
= 4
_
e
m
0
c
2
_
2
2
g
(ik)
.
We now dene an inverse wave function:
= 2
_
e
m
0
c
2
_
1
.
We can then express the Ricci and the electromagnetic eld ten-
sors as
R
ik
=
2
g
(ik)
,
F
ik
= 2g
[ik]
.
Then it follows from (1.2b) and (1.2c) that
F
ik
= 2
1
g
[kr]
R
r
i
,
R =
1
2
g
[ik]
F
ik
= 4
2
.
We have now therefore fullled our promise in the beginning (for
instance, at the end of Section 1.2) to express gravity and electro-
magnetism in terms of the components of the fundamental tensor, i.e.,
g
(ik)

2
(ik)
and g
[ik]

2
[ik]
, alone.
4.3 On the conservation of currents
We need to recall the basic eld equations:
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) ,
F
ij;k
F
ik;j
= 2
r
[jk]
F
ir
,
R
ik
=
1
4
F
i
r
F
rk
,
R
ik
;k
=
1
4
F
i
r;k
F
rk
1
4
F
i
r
F
rk
;k
=
=
1
4
F
i
k
J
k
+
1
4
F
i
k
k
[rs]
F
rs
,
J
i
= 2 g
(ik)
r
[ks]
F
s
r
.
To guarantee conservation of currents, we now introduce the one-
form
i

n
:
i
= R
ik
g
k
+
i
n,
where
i
represents another current. Requiring, in the sense of the
theory of distributions, that the covariant derivative of
i
vanishes at
all points of R
4
also means that its covariant divergence also vanishes:
i
;i
= 0 .
Hence
R
ik
;i
g
k
+R
ik
g
i;k
+
i
;i
n +
i
n
;i
= 0 ,
R
ik
;i
g
k

1
2
R
ik
F
ik
n +
i
;i
n
1
2

i
F
k
i
g
k
= 0 .
Then we have
R
ik
;k
=
1
2
F
i
k

k
,
i
;i
= 0 .
Comparing the last two equations above with the fourth equation,
one must nd
i
=
1
2
_
J
i
i
[kl]
F
kl
_
.
Meanwhile, in the presence of torsion the Bianchi identity for the
electromagnetic eld tensor and the covariant divergence of the four-
current J are
F
ij;k
+F
jk;i
+F
ki;j
= 2
_
r
[ij]
F
kr
+
r
[jk]
F
ir
+
r
[ki]
F
jr
_
,
J
i
;i
=
_
i
[kl]
F
kl
_
;i
.
Now since
i
;i
=0, we see that
_
i
[kl]
F
kl
_
;i
= 0
and (in a general setting)
J
i
=
i
[kl]
F
kl
,= 2
i
for some constant , are necessary and sucient conditions for the cur-
rent density vector J to be conserved. Otherwise both
i
and J
i
are
directly equivalent to each other. Of course one may also dene another
conserved current through
j
i
= F
ik
||k
j
i
||i
= 0 ,
where the double stroke represents a covariant derivative with respect
to the symmetric Levi-Civita connection.
In general the current J will automatically be conserved if the or-
thogonality condition imposed on the twist vector, derived from the
torsion tensor, and the velocity vector:
i
u
i
= 0 ,
where
i
=
k
[ki]
holds. For more details of the conservation law for charges, see Sec-
tion 4.4 below.
4.4 On the wave equations of our unied eld theory
We start again with the basic eld equations of our unied eld theory:
R
ijkl
=
1
4
(F
ik
F
jl
F
il
F
jk
) ,
F
ij;k
F
ik;j
= 2
r
[jk]
F
ir
,
J
i
= 2 g
(ik)
s
[kr]
F
r
s
.
We remind ourselves that these eld equations give us a set of com-
plete relations between the curvature tensor, the torsion tensor, the elec-
tromagnetic eld tensor and the current density vector. From these eld
equations, we are then able to derive the following insightful algebraic
relations:
F
ij
R
ijkl
= R
r
k
F
rl
R
r
l
F
rk
,
F
ik
= R
1
(F
rs
R
risk
+F
ir
R
r
k
) ,
R
2
= R
ik
R
ik
+
1
4
R
ijkl
F
ik
F
jl
=
= R
ik
R
ik
+R
ijkl
R
ijkl
+
1
4
R
ijkl
F
il
F
jk
,
J
i
= 2R
1
g
(ik)
q
[kp]
_
F
rs
R
p
r.sq
+F
p
r
R
r
q
_
,
R = 2
2
J
i
q
[ik]
_
F
rs
R
k
r.sq
+F
k
r
R
r
q
_
,
where the density corresponds to possible electric-magnetic charge
distribution. As has been shown previously, we also have the traceless
eld equation:
R
ik

1
4
g
(ik)
R =
1
4
_
F
il
F
l
k

1
4
g
(ik)
F
rs
F
rs
_
.
Now recall that the tetrad and the unit normal vector satisfy
e
A
i;k
=
1
2
F
ik
n
A
,
n
A
;i
=
1
2
F
k
i
e
A
k
.
Then we see that
e
A
i;j;k
=
1
2
F
ij;k
n
A
1
4
F
ij
F
r
k
e
A
r
,
g
(jk)
e
A
i;j;k
=
1
2
J
i
n
A
1
4
F
ik
F
rk
e
A
r
.
With the help of the basic eld equations, we obtain the tetrad wave
equation of our unied eld theory:
2e
A
i
=
1
2
J
i
n
A
R
k
i
e
A
k
.
This expression gives a wave-type equation of the tetrad endowed
with two sources: the electromagnetic source, i.e., the electromagnetic
current density vector and the Ricci curvature tensor which represents
the gravitational source in standard General Relativity.
In addition, we can also obtain the following wave equation:
2n
A
=
1
2
J
i
e
A
i
Rn
A
.
In other words,
(2 +R) n
A
=
1
2
J
i
e
A
i
.
Furthermore, it follows that
e
k
A
2e
A
i
= R
k
i
,
n
A 2e
A
i
=
1
2
J
i
,
n
A2n
A
= R,
e
i
A
2n
A
=
1
2
J
i
.
We may now express the current density vector as
J
i
= 2n
A
2e
i
A
.
Again, we obtain the conservation law of electromagnetic currents
as follows:
J
i
;i
= 2n
A
;i
2e
i
A
+ 2n
A
2e
i
A;i
=
= F
k
i
e
A
k
2e
i
A
=
= F
k
i
R
i
k
=
= 0 .
(Here we have also used the relation 2e
i
A;i
=
1
2
2
_
F
i
i
n
A
_
=0.)
Meanwhile, since the equation 2g
(ik)
=0 must be satised uncon-
ditionally by the metric tensor g
(ik)
=e
Ai
e
A
k
and also since g
(AB);i
=
g
(AB);C
e
C
i
=0, we then get
e
Ak2e
A
i
= e
A
i
2e
Ak
2R
ik
.
Now the curvature tensor can be expressed in terms of the tetrad as
R
ijkl
= g
(AB)
_
k
e
A
i

l
e
B
j

l
e
A
i

k
e
B
j
_
.
Then the Ricci tensor is
R
ik
= g
(AB)
r
e
A
i

r
e
B
k
.
We can also express this as
R
ik
= g
(AB)
e
A
i
2e
B
k
,
which can be written equivalently as
R
ik
=
1
2
g
(AB)
_
e
A
i

r
r
e
B
k
+e
A
k
r
e
B
i
_
.
Consider now a source-free region in the space-time R
4
. As weve
seen, the absence of source is characterized by the vanishing of the
torsion tensor. In such a case, if the space-time has a constant sectional
curvature K, we obtain the tetrad wave equation for the empty region:
(2 +
1
4
R) e
A
i
= 0 .
In other words,
(2 +K) e
A
i
= 0 .
Combining our tetrad wave equation 2e
A
i
=
1
2
J
i
n
A
R
k
i
e
A
k
with the
equation for the Ricci tensor we have derived in the quantum limit (in
Section 4.2), which is
R
ik
= 4
_
e
m
0
c
2
_
2
2
g
(ik)
,
we obtain the following wave equation in the presence of the electro-
magnetic current density:
(2 +
1
4
R) e
A
i
=
1
2
J
i
n
A
,
R = C + 4
_
_
r
[ik]
R
ik
rs
2
i
[rs]
R
r
i
_
dx
s
,
where C =4K is constant.
Finally, lets have a look back at the wave equation given by (3.15b):
(2 R) [) = 0 .
If we fully assume that the space-time R
4
is embedded isometrically
in IR
5
spanned by the time coordinate =ct and the four space co-
ordinates u, v, w, y which together form the line-element ds
2
=d
2
du
2
dv
2
dw
2
dy
2
d
2
d
2
dy
2
, and if the wave function
represented by the state vector [) does not depend on the microscopic
fth coordinate y, we can write the wave equation in the simple form:
_

2
2

2
2
_
[) = R[) .
4.5 A more compact form of the generalized Gauss-Codazzi
equations in IR
5
We write the generalized Gauss-Codazzi equations (see (2.6) for in-
stance) in IR
5
once again:
R
ijkl
=
ik
jl

il
jk
+R
ABCD
e
A
i
e
B
j
e
C
k
e
D
l
S
Ajkl
e
A
i
,
ij;k

ik;j
= R
ABCD
n
A
e
B
i
e
C
j
e
D
k
2
r
[jk]
ir
+S
Aijk
n
A
in terms of the general asymmetric extrinsic curvature tensor. Here, as
usual,
S
A
ijk
= e
A
i,jk
e
A
i,kj
+
_
e
C
j,k
e
C
k,j
_
A
BC
e
B
i
.
From the fundamental relations
e
A
i;j
=
ij
n
A
,
n
A
;i
=
k
i
e
A
k
,
we have
e
A
i;j;k
=
ij;k
n
A
ij
r
k
e
A
r
.
Hence
e
A
i;j;k
e
A
i;k;j
= (
ij;k

ik;j
) n
A
+
_
r
j
ik

r
k
ij
_
e
A
r
.
With the help of the generalized Gauss-Codazzi equations above and
the identity
e
i
A
e
B
i
=
B
A
n
A
n
B
,
we see that that the generalized Gauss-Codazzi equations in IR
5
can be
written somewhat more compactly as a single equation:
e
A
i;j;k
e
A
i;k;j
= R
r
ijk
e
A
r
R
A
BCD
e
B
i
e
C
j
e
D
k
2
r
[jk]
ir
n
A
+S
A
ijk
.
Appendix: The Fundamental Geometric Properties
of a Curved Manifold
Let us present the fundamental geometric objects of an n-dimensional
curved manifold. Let
a
=
X
i
x
a
E
i
=
a
X
i
E
i
(the Einstein summation
convention is assumed throughout this work) be the covariant (frame)
basis spanning the ndimensional base manifold C
with local coordi-

nates x
a
=x
a
_
X
k
_
. The contravariant (coframe) basis
b
is then given
via the orthogonal projection

b
,
a
_
=
b
a
, where
b
a
are the compo-
nents of the Kronecker delta (whose value is unity if the indices coin-
cide or null otherwise). The set of linearly independent local directional
derivatives E
i
=

X
i
=
i
gives the coordinate basis of the locally at
tangent space T
x
(M) at a point x C
. Here M denotes the topo-

logical space of the so-called n-tuples h (x) =h
_
x
1
, . . . , x
n
_
such that
relative to a given chart (U, h(x)) on a neighborhood U of a local co-
ordinate point x, our C
-dierentiable manifold itself is a topological

space. The dual basis to E
i
spanning the locally at cotangent space
T
x
(M) will then be given by the dierential elements dX
k
via the rela-
tion

dX
k
,
i
_
=
k
i
. In fact and in general, the one-forms dX
k
indeed
act as a linear map T
x
(M) IR when applied to an arbitrary vector
eld F T
x
(M) of the explicit form F =F
i
X
i
=f
a
x
a
. Then it is
easy to see that F
i
=FX
i
and f
a
=Fx
a
, from which we obtain the
usual transformation laws for the contravariant components of a vector
eld, i.e., F
i
=
a
X
i
f
a
and f
i
=
i
x
a
F
i
, relating the localized compo-
nents of F to the general ones and vice versa. In addition, we also see
that

dX
k
, F
_
=F X
k
=F
k
.
The components of the symmetric metric tensor g =g
ab
b
of the
base manifold C
are readily given by

g
ab
=
a
,
b
)
satisfying
g
ac
g
bc
=
b
a
,
where g
ab
=

a
,
b
_
. It is to be understood that the covariant and
contravariant components of the metric tensor will be used to raise and
the (component) indices of vectors and tensors.
The components of the metric tensor g (x
N
) =
ik
dX
i
dX
k
de-
scribing the locally at tangent space T
x
(M) of rigid frames at a point
Appendix: The Fundamental Properties of a Curved Manifold 71
x
N
=x
N
(x
a
) are given by
ik
= E
i
, E
k
) = diag (1, 1, . . . , 1) .
In four dimensions, the above may be taken to be the components of
the Minkowski metric tensor, i.e.,
ik
= E
i
, E
k
) =diag (1, 1, 1, 1).
Then we have the expression
g
ab
=
ik
a
X
i
b
X
k
.
The line-element of C
is then given by
ds
2
= g = g
ab
_
i
x
a
k
x
b
_
dX
i
dX
k
,
where
a
=
i
x
a
dX
i
.
Given the existence of a local coordinate transformation via
x
i
=x
i
(x
) in C
, the components of an arbitrary tensor eld T C
of rank (p, q) transform according to

T
ab...g
cd...h
= T
...
...

x
a
x
b
. . .
x
g
c
x
d
x
. . .
h
x
.
Let
i
1
i
2
...i
p
j
1
j
2
...j
p
be the components of the generalized Kronecker delta.
They are given by
i
1
i
2
...i
p
j
1
j
2
...j
p
=
j
1
j
2
...j
p
i
1
...i
p
= det
_
_
_
_
_
_
i
1
j
1

i
2
j
1
. . .
i
p
j
1
i
1
j
2

i
2
j
2
. . .
i
p
j
2
. . . . . . . . . . . .
i
1
j
p
i
2
j
p
. . .
i
p
j
p
_
_
_
_
_
_
where
j
1
j
2
...j
p
=
_
det (g)
j
1
j
2
...j
p
and
i
1
i
2
...i
p
=
1
det(g)
i
1
i
2
...i
p
are
the covariant and contravariant components of the completely anti-
symmetric Levi-Civita permutation tensor, respectively, with the ordi-
nary permutation symbols being given as usual by
j
1
j
2
...j
q
and
i
1
i
2
...i
p
.
Again, if is an arbitrary tensor, then the object represented by
j
1
j
2
...j
p
=
1
p!

i
1
i
2
...i
p
j
1
j
2
...j
p
i
1
i
2
...i
p
is completely anti-symmetric.
Introducing a generally asymmetric connection via the covariant
derivative
a
=
c
ab
c
,
i.e.,
c
ab
=
c
,
b
a
) =
c
(ab)
+
c
[ab]
,
where the round index brackets indicate symmetrization and the square
ones indicate anti-symmetrization, we have, by means of the local coor-
dinate transformation given by x
a
=x
a
(x
) in C
b
e
a
=
c
ab
e
a
e
b
,
where the tetrads of the moving frames are given by e
a
=
a
x
and
e
a
x
a
. They satisfy e
a
b
=
a
b
and e
a
e
a
. In addition, it can
also be veried that
e
a
e
a

a
bc
e
b
e
c
b
e
a
= e
a
b

a
cb
e
c
.
We know that is a non-tensorial object, since its components trans-
form as
c
ab
= e
c
b
e
a
+e
c
a
e
b
.
However, it can be described as a kind of displacement eld since
it is what makes possible a comparison of vectors from point to point
in C
. In fact the relation

b
a
=
c
ab
c
denes the so-called metricity
condition, i.e., the change (during a displacement) in the basis can be
measured by the basis itself. This immediately translates into
c
g
ab
= 0 ,
where we have just applied the notion of a covariant derivative to an
arbitrary tensor eld T:
m
T
ab...g
cd...h
=
m
T
ab...g
cd...h
+
a
pm
T
pb...g
cd...h
+
b
pm
T
ap...g
cd...h
+ +
g
pm
T
ab...p
cd...h
p
cm
T
ab...g
pd...h

p
dm
T
ab...g
cp...h

p
hm
T
ab...g
cd...p
such that (
m
T)
ab...g
cd...h
=
m
T
ab...g
cd...h
.
The condition
c
g
ab
=0 can be solved to give
c
ab
=
1
2
g
cd
(
b
g
da

d
g
ab
+
a
g
bd
) +
c
[ab]
g
cd
_
g
ae
e
[db]
+g
be
e
[da]
_
from which it is customary to dene
c
ab
=
1
2
g
cd
(
b
g
da

d
g
ab
+
a
g
bd
)
as the Christoel symbols (symmetric in their two lower indices) and
K
c
ab
=
c
[ab]
g
cd
_
g
ae
e
[db]
+g
be
e
[da]
_
as the components of the so-called contorsion tensor (anti-symmetric in
the rst two mixed indices).
Note that the components of the torsion tensor are given by
a
[bc]
=
1
2
e
a
c
e
b

b
e
c
+e
c
e
b
_
,
where we have set
c
=
c
, such that for an arbitrary scalar eld
we have
(
a
b

b
a
) = 2
c
[ab]
c
.
The components of the curvature tensor R of C
are then given via

the relation
(
q
p

p
q
) T
ab...s
cd...r
= T
ab...s
wd...r
R
w
cpq
+T
ab...s
cw...r
R
w
dpq
+ +T
ab...s
cd...w
R
w
rpq
T
wb...s
cd...r
R
a
wpq
T
aw...s
cd...r
R
b
wpq
. . .T
ab...w
cd...r
R
s
wpq
2
w
[pq]
w
T
ab...s
cd...r
,
where
R
d
abc
=
b
d
ac

c
d
ab
+
e
ac
d
eb

e
ab
d
ec
=
= B
d
abc
() +

b
K
d
ac

c
K
d
ab
+K
e
ac
K
d
eb
K
e
ab
K
d
ec
,
where

denotes covariant dierentiation with respect to the Christoel
symbols alone, and where
B
d
abc
() =
b
d
ac

c
d
ab
+
e
ac
d
eb

e
ab
d
ec
are the components of the Riemann-Christoel curvature tensor of C
.
From the components of the curvature tensor, namely, R
d
abc
, we
have (using the metric tensor to raise and lower indices)
R
ab
R
c
acb
= B
ab
() +

c
K
c
ab
K
c
ad
K
d
cb
2
c
[ac]
+ 2K
c
ab
d
[cd]
,
R R
a
a
= B() 4g
ab
c
[bc]
2g
ac
b
[ab]
d
[cd]
K
abc
K
acb
,
where B
ab
() B
c
acb
() are the components of the symmetric Ricci
tensor and B() B
a
a
() is the Ricci scalar. Note that K
abc
g
ad
K
d
bc
and K
acb
g
cd
g
be
K
a
de
.
Now since
b
ba
=
b
ba
=
b
ab
=
a
_
ln
_
det (g)
_
,
b
ab
=
a
_
ln
_
det (g)
_
+ 2
b
[ab]
,
we see that for a continuous metric determinant, the so-called homoth-
etic curvature vanishes:
H
ab
R
c
cab
=
a
c
cb

b
c
ca
= 0 .
Introducing the traceless Weyl tensor W, we have the following de-
composition theorem:
R
d
abc
= W
d
abc
+
1
n 2
_
d
b
R
ac
+g
ac
R
d
b

d
c
R
ab
g
ab
R
d
c
_
+
+
1
(n 1) (n 2)
_
d
c
g
ab

d
b
g
ac
_
R,
which is valid for n > 2. For n=2, we have
R
d
abc
= K
G
_
d
b
g
ac

d
c
g
ab
_
,
where
K
G
=
1
2
R
is the Gaussian curvature of the surface. Note that (in this case) the
Weyl tensor vanishes.
Any n-dimensional manifold (for which n > 1) with constant sec-
tional curvature R and vanishing torsion is called an Einstein space. It
is described by the following simple relations:
R
d
abc
=
1
n(n 1)
_
d
b
g
ac

d
c
g
ab
_
R,
R
ab
=
1
n
g
ab
R.
In the above, we note especially that
R
d
abc
= B
d
abc
() ,
R
ab
= B
ab
() ,
R = B() .
Furthermore, after some lengthy algebra, we obtain, in general, the
following generalized Bianchi identities:
R
a
bcd
+R
a
cdb
+R
a
dbc
=
= 2
_
a
[bc]
+
b
a
[cd]
+
c
a
[db]
+
a
eb
e
[cd]
+
a
ec
e
[db]
+
a
ed
e
[bc]
_
,
e
R
a
bcd
+
c
R
a
bde
+
d
R
a
bec
= 2
_
f
[cd]
R
a
bfe
+
f
[de]
R
a
bfc
+
f
[ec]
R
a
bfd
_
,
a
_
R
ab
1
2
g
ab
R
_
= 2 g
ab
c
[da]
R
d
c
+
a
[cd]
R
cdb
...a
,
for any metric-compatible manifold endowed with both curvature and
torsion.
In the last of the above set of equations, we have introduced the
generalized Einstein tensor, i.e.,
G
ab
R
ab

1
2
g
ab
R.
In particular, we also have the following specialized identities, i.e.,
the regular Bianchi identities:
B
a
bcd
+B
a
cdb
+B
a
dbc
= 0 ,

e
B
a
bcd
+

c
B
a
bde
+

d
B
a
bec
= 0 ,
a
_
B
ab
1
2
g
ab
B
_
= 0 .
In general, these hold in the case of a symmetric, metric-compatible
connection. Non-metric dierential geometry is beyond the scope of our
present consideration.
We now dene the so-called Lie derivative which can be used to
dene a dieomorphism invariant in C
. For a vector eld U and a

tensor eld T, both arbitrary, the invariant derivative represented (in
component notation) by
L
U
T
ab...g
cd...h
=
m
T
ab...g
cd...h
U
m
+T
ab...g
md...h
c
U
m
+T
ab...g
cm...h
d
U
m
+ +
+ +T
ab...g
cd...m
h
U
m
T
mb...g
cd...h

m
U
a
T
am...g
cd....h

m
U
b
T
ab...m
cd...h

m
U
g
denes the Lie derivative of T with respect to U. With the help of the
torsion tensor and the relation
b
U
a
=
b
U
a
a
cb
U
c
=
b
U
a
a
bc
2
a
[bc]
_
U
c
,
we can write
L
U
T
ab...g
cd...h
=
m
T
ab...g
cd...h
U
m
+T
ab...g
md...h
c
U
m
+T
ab...g
cm...h
d
U
m
+ +
+T
ab...g
cd...m
h
U
m
T
mb...g
cd...h

m
U
a
T
am...g
cd...h

m
U
b

T
ab...m
cd...h

m
U
g
+ 2
a
[mp]
T
mb...g
cd...h
U
p
+ 2
b
[mp]
T
am...g
cd...h
U
p
+ +
+ 2
g
[mp]
T
ab...m
cd...h
U
p
2
m
[cp]
T
ab...g
md...h
U
p
+ 2
m
[dp]
T
ab...g
cm...h
U
p

2
m
[hp]
T
ab...g
cd...m
U
p
.
Hence, noting that the components of the torsion tensor, namely,
i
[kl]
, indeed transform as components of a tensor eld, it is seen that the
L
U
T
ij...s
kl...r
do transform as components of a tensor eld. Apparently, the
beautiful property of the Lie derivative (applied to an arbitrary tensor
eld) is that it is connection-independent even in a curved manifold.
Conclusion
We have shown that gravity and electromagnetism are intertwined in a
very natural manner, both ensuing from the melting of the same under-
lying space-time geometry. They obey the same set of eld equations.
However, there are actually no objectively existing elementary particles
in this theory. Based on the wave equation (73), we may suggest that
what we perceive as particles are only singularities which may be in-
terpreted as wave centers. In the microcosmos everything is essentially
a wave function that also contains particle properties. Individual wave
function is a fragment of the universal wave function represented by the
wave function of the Universe in (73). Therefore all objects are essen-
tially interconnected. We have seen that the electric(-magnetic) charge
is none other than the torsion of space-time. This charge can also be
described by the wave function alone. This doesnt seem to be contra-
dictory evidence if we realize that nothing exists in the quantum realm
save the quantum mechanical wave function (unfortunately, we have not
made it possible here to carry a detailed elaboration on this statement).
Although we have not approached and constructed a quantum theory of
gravity in the strictly formal way (through the canonical quantization
procedure), internal consistency of our theory awaits further justica-
tion. For a few more details of the underlying unifying features of our
theory, see Chapter Additional Considerations.
Bibliography
(Other Attempts at Unication by the Author)
1. Suhendro I. A four-dimensional continuum theory of space-time and the
classical physical elds. Progress in Physics, 2007, no. 3, 3446.
2. Suhendro I. A new semi-symmetric unied eld theory of the classical
elds of gravity and electromagnetism. Progress in Physics, 2007, no. 3,
4762.
3. Suhendro I. A new conformal theory of semi-classical quantum General
Relativity. Progress in Physics, 2007, no. 3, 96103.
4. Suhendro I. A unied eld theory of gravity, electromagnetism, and the
Yang-Mills gauge eld. Progress in Physics, 2008, no. 1, 3137.
5. Suhendro I. On a geometric theory of generalized chiral elasticity with
discontinuities. Progress in Physics, 2008, no. 1, 5874.
6. Suhendro I. On the geometry of background currents in General Relativ-
ity. Progress in Physics, 2008, no. 2, 121124.
Spin-Curvature and the Unication of Fields in a Twisted Space
by Indranu Suhendro
ISBN 978-91-85917-01-3. Svenska fysikarkivet, 2008, 78 pages
The book draws theoretical ndings for spin-curvature and the unication of
elds in a twisted space. A space twist, represented through the appropriate
formalism, is related to the anti-symmetric metric tensor. Kaluzas theory is
extended and given an appropriate integrability condition. Both matter and
the isotropic electromagnetic eld are geometrized through common eld
equations: trace-free eld equations giving the energy-momentum tensor
for such an electromagnetic eld solely via the (generalized) Ricci curvature
tensor and scalar are obtained. In the absence of electromagnetic elds
the theory goes to Einsteins 1928 theory of distant parallelism where only
matter eld is geometrized (through the twist of space-time). The above
results in common with respective wave equations are joined into a unied
eld theory of semi-classical gravoelectrodynamics.
Spinn-kr okning och f oreningen av f alt i ett tvistat rum
av Indranu Suhendro
ISBN 978-91-85917-01-3. Svenska fysikarkivet, 2008, 78 sidor
Boken skildrar teoretiska forskningsresultat f or spinn-kr okning och f orening
av f alt i ett tvistat rum. En rumsartad tvist, representerat genom l amplig
formalism, relateras till den antisymmetriska metriska tensorn. Kaluzas
teori ut okas och ges ett l ampligt villkor f or integrabilitet. B ade materief alt
och isotropiska elektromagnetiska f alt geometriseras genom gemensamma
f altekvationer som ar sp arfria f altekvationer som ger energi-impulstensorn
f or ett elektromagnetiskt f alt enbart via den (generaliserade) Ricci kr okning-
stensorn och skal aren. I avsaknad av elektromagnetiska f alt leder teorin
till Einsteins teori fr an 1928 om avl agsen parallellism d ar bara materief alt
geometriseras (genom en tvistning av rumstiden). Ovanst aende resultat till-
sammans med respektive v agekvationer f orenas till en f orenad f altteori av
semiklassisk gravitationselektrodynamik.

Spin-Curvature and The Unification of Fields in A Twisted Space

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Spin-Curvature and The Unification of Fields in A Twisted Space

Uploaded by

Copyright:

Available Formats

INDRANU SUHENDRO

denes a null rotation,

is the generalized Kronecker delta and

. (The minus sign holds if the manifold is Lorentzian and

vanishes and the

is the connection of the space

g equals a constant. Following (2.g), we can express the curvature

denotes the exterior change with respect to the

where now the capital Latin indices A, B, . . . run from 1 to N and

with local coordi-

. Here M denotes the topo-

-dierentiable manifold itself is a topological

are readily given by

, the components of an arbitrary tensor eld T C

of rank (p, q) transform according to

. In fact the relation

are then given via

. For a vector eld U and a

You might also like