Higher-Dimensional Physics: Old English Saying

1.
HIGHER-DIMENSIONAL PHYSICS
Theres more to this than meets the eye (Old English saying)
1.1
Introduction
Theoretical physics is in the happy situation of being able to
pluck good ideas from philosophy, work them through using the machinery of algebra, and produce something which is both stimulating
and precise. It goes beyond words and equations, because when
properly done it encapsulates what many people regard as reality.
We sometimes tend to forget what a stride was made when
Newton realized that the force which causes an apple to fall to the
ground is the same one which keeps the Moon in its orbit and which
is now known to influence the motions of even the most remote galaxies. Nowadays, gravity has to be considered in conjunction with
electromagnetism plus the weak and strong forces of particle physics.
Even so, it is still possible to give an account of modern physics in a
few hundred pages or so. On reflection, this is remarkable. It comes
about because of the enormous efficiency of mathematics as the natural language of physics, coupled with the tradition whereby physicists
introduce the least number of hypotheses necessary to explain the
natural world (Occams razor of old). At present, it is commonly believed that the best way to explain all of the forces of physics is via
the idea of higher dimensions.
In this regard, five-dimensional field theory is particularly
useful, as it is the basic extension of the four-dimensional spacetime
FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

World Scientific Publishing Co. Pte. Ltd.
http://www.worldscibooks.com/physics/6029.html
Five-Dimensional Physics
of Einstein gravity and is widely regarded as the low-energy limit of

higher-dimensional theories which more fully address the particle interactions. This slim volume is a concise account of recent developments in 5D theory and their implications for classical and quantum
physics.
1.2
Dimensions Then and Now

The idea of a dimension is primitive and at least partly in-
tuitive. Recent histories of the idea are given in the books by Wesson
(1999) and Halpern (2004). It was already established by the time of
Newton, who realized that mass was a more fundamental concept
than density, and that a proportionality between physical quantities
could be converted to an equation if the latter balanced its ingredients
of mass, length and time (i.e., was dimensionally homogeneous).
Hence the introduction of a parameter G, which we now call Newtons constant of gravity.
The coordinates of an object (x, y, z) in ordinary space and
that of local time (t) are, of course, the basic dimensions of geometry.
But the concept of force, at least the gravitational kind, obliges us to
introduce another dimension related to the mass of an object (m).
And modern physics recognizes other such, notably the one which
measures a bodys electric charge (q). The role of the so-called fundamental constants of physics is primarily to transpose quantities like
mass and charge into geometrical ones, principally lengths (Wesson
1999, pp. 2-11). This is illustrated most cogently by the conversion

Higher-Dimensional Physics
of the time to an extra coordinate x 4 ct via the use of the speed of

light, a ploy due to Minkowski and Einstein which forms the foundation of 4D spacetime.
The idea of a dimension is, to a certain extent, malleable. It is
also important to notice that modern field theories, like general relativity, are written in terms of tensor equations which are not restricted
in their dimensionality. One can speculate that had Einstein been
formulating his theory of gravity today, he might have established this
anonymity of dimension as a principle, on a par with the others with
which we are familiar, such as that of equivalence (see Chapter 3). It
is this freedom to choose the dimensionality which underlies the numerous extensions of general relativity. These include the original
5D Kaluza-Klein theory, its modern variants which are called induced-matter and membrane theory, plus the higher extensions such
as 10D supersymmetry, 11D supergravity and the higher-D versions
of string theory.
Kaluza initiated field theory with more than the 4 dimensions
of spacetime in 1921, when he published a paper which showed how
to unify gravity (as described by Einsteins equations) with electromagnetism (as described by Maxwells equations). It is well known
that Einstein kept Kaluzas paper for a couple of years before finally
as referee allowing it to go forward. However, Einstein was then and
remained in his later years an advocate of extra dimensions. For example, a letter to Kaluza from Einstein in 1919 stated The formal
unity of your theory is astonishing (Halpern, 2004, p.1). Indeed, the

natural way in which the 4 Maxwell equations fall out of the 15 field
equations of what is a kind of general relativity in 5D, has since come
to be called the Kaluza-Klein miracle. However, the mathematical
basis of the unification is simple: In 5D there are 15 independent
components of the metric tensor, of which one refers to a scalar field
which was not at the time considered significant and was so suppressed. For similar reasons, to do with the presumed unobservability
of effects to do with the extra dimension, all derivatives of the other
metric coefficients with respect to the extra coordinate were set to
zero (the cylinder condition).
which
could
depend
on
the
This left 14 metric coefficients,

4
coordinates
of
spacetime
( x , = 0,123 for t , xyz ). These 14 coefficients were determined

by
14 field equations. The latter turned out to be the 10 Einstein
equations and the 4 Maxwell equations. Voila: a unification of gravity and electromagnetism.
Klein pushed the 5D approach further in 1926, when he published a paper which showed how to incorporate quantum effects into
the theory. He did this by the simple device of assuming that the topology of the extra dimension was not flat and open, but curved into a
circle. In other words, while a local orbit in spacetime (x) would be
straight, an orbit in the extra dimension (x4) would merely go around
and around. This cyclic behaviour would lead to quantum effects,
provided the extra dimension were rolled up to a microscopic size
(compactification). The size of the extra dimension was presumed
to be related to the parameter typical of quantum phenomena, namely

Plancks constant h. Among other consequences of the closed topology of the extra dimension, it was shown that the cyclic momentum
could be related to the charge of the electron e, thus explaining its
quantization.
The brainwaves of Kaluza and Klein just summarized are the
kind which are neat and yet powerful. They continued to be held in
high regard for many years in theoretical physics, even though the
latter was redirected by the algebraically simple and effective ideas on
wave mechanics that were soon introduced by Schrodinger, Heisenberg and Dirac. Kaluza-Klein theory later underwent a revival, when
Einsteins theory was recognized as the best basis for cosmology. But
something has to be admitted: Kaluza-Klein theory in its original
form is almost certainly wrong.
By this, it is not meant that an experiment was performed
which in the standard but simplistic view of physics led to a disproval
of the 5D theory. Rather, it means that the original Kaluza-Klein theory is now acknowledged as being at odds with a large body of modern physical lore. For example, the compactification due to Klein
leads to the prediction that the world should be dominated by particles
with the Planck mass of order 10-5 g, which is clearly not the case.
(This mismatch is currently referred to as the hierarchy problem, to
which we will return.) Also, the suppression of the scalar field due to
Kaluza leaves little room to explain the dark energy currently believed to be a major component of the universe. (This is a generic
form of what is commonly referred to as the cosmological-constant

problem, to which attention will be given later.) Further, the cylinder

condition assumed by both fathers of 5D field theory effectively rules
out any way to explain matter as a geometrical effect, something
which Einstein espoused and is still the goal of many physicists.
It is instructive to recall at this juncture the adage which
warns us not to throw out the baby with the bath-water. In this instance, the baby is the concept of a 5 (or higher) D space; whereas the
water is the smothering algebraic restrictions which were applied to
the theory in its early days as a means of making progress, but which
are now no longer needed.
Hence modern Kaluza-Klein theory,
which is algebraically rich and exists in several versions.

1.3
Higher-Dimensional Theories
These may be listed in terms of their dimensionality and
physical motivation. However, all are based on Einsteins theory of

general relativity. The equations for this and its canonical extension
will be deferred to the next section.
Induced-matter theory is based on an unrestricted 5D manifold, where the extra dimension and derivatives with respect to the
extra coordinate are used to explain the origin of 4D matter in terms
of geometry. (For this reason, it is sometimes called space-timematter theory.) As mentioned above, this goal was espoused by Einstein, who wished to transpose the base-wood of the right-hand side
of his field equations into the marble of the left-hand side. That is,
he wished to find an algebraic expression for what is usually called

( )
the energy-momentum tensor T , which was on the same footing

as the purely geometrical object we nowadays refer to as the Einstein
tensor G . That this is possible in practice was proved using an

algebraic reduction of the 5D field equations by Wesson and Ponce de
Leon (1992). They were, however, unaware that the technique was
guaranteed in principle by a little-known theorem on local embeddings of Riemannian manifolds by Campbell (1926). We will return
to the field equations and their embeddings below. Here, we note that
the field equations of 5D relativity with a scalar field and dependence
on the extra coordinate in general lead to 15 second-order, non-linear
relations. When the field equations are set to zero to correspond to a
5D space which is apparently empty, a subset of them gives back the
10 Einstein field equations in 4D with sources. That is, there is an
effective or induced 4D energy-momentum tensor which has the
properties of what we normally call matter, but depends on the extra
metric coefficients and derivatives with respect to the extra coordinate. The other 5 field equations give back a set of 4 Maxwell-like or
conservation equations, plus 1 scalar relation which has the form of a
wave equation. Following the demonstration that matter could be
viewed as a consequence of geometry, there was a flurry of activity,
resulting in several theorems and numerous exact solutions (see Wesson 1999 for a catalog). The theory has a 1-body solution which satisfies all of the classical tests of relativity in astrophysics, as well as
other solutions which are relevant to particle physics.

Membrane theory is based on a 5D manifold in which there is

a singular hypersurface which we call 4D spacetime. It is motivated
by the wish to explain the apparently weak strength of gravity as
compared to the forces of particle physics. It does this by assuming
that gravity propagates freely (into the 5D bulk), whereas particle interactions are constrained to the hypersurface (the 4D brane). That
this is a practical approach to unification was realized by Randall and
Sundrum (1998, 1999) and by Arkani-Hamed, Dimopoulos and Dvali
(1998, 1999). The original theory helped to explain the apparently
small masses of elementary particles, which is also referred to as the
hierarchy problem. In addition, it helped to account for the existence
and size of the cosmological constant, since that parameter mediates
the exponential factor in the extra coordinate which is typical of distances measured away from the brane. As with induced-matter theory, the membrane approach has evolved somewhat since its inception. Thus there has been discussion of thick branes, the existence of
singular or thin branes in (4+d) dimensions or d-branes, and the possible collisions of branes as a means of explaining the big bang of traditional 4D cosmology. It should also be mentioned that the field
equations of induced-matter and membrane theory have recently been
shown to be equivalent by Ponce de Leon (2001; see below also).
This means that the implications of these approaches for physics owes
more to interpretation than algebra, and exact solutions for the former
theory can be carried over to the latter.

Theories in N > 5 dimensions have been around for a considerable time and owe their existence to specific physical circumstances. Thus 10D supersymmetry arose from the wish to pair every
integral-spin boson with a half-integral-spin fermion, and thereby
cancel the enormous vacuum or zero-point fields which would ensue
otherwise. The connection to ND classical field theory involves the
fact that it is possible to embed any curved solution (with energy) of a
4D theory in a flat solution (without energy) of a higher-dimensional
theory, provided the larger manifold has a dimension of N 10.
From the viewpoint of general relativity or a theory like it, which has
10 independent components of the metric tensor or potentials, this is
hardly surprising. The main puzzle is that while supersymmetry is a
property much to be desired from the perspective of theoretical particle physics, it must be very badly broken in a practical sense. The
reason for this apparent conflict between theory and practice may
have to do with our (perhaps unjustified) wish to reduce physics to
4D, and/or our (probably incomplete) knowledge of how to categorize
the properties of particles using internal symmetry groups. The latter
have, of course, to be taken into account when we attempt to estimate
the size of the space necessary to accommodate both gravity and
the particle interactions. Hence the possible unification in terms of
(4+7)D or 11D supergravity. However, a different approach is to
abandon completely the notion of a point with its implied singularity and instead model particles as strings (Szabo 2004, Gubser and
Lykken 2004). The logic of this sounds compelling, and string theory

10 Five-Dimensional Physics
offers a broad field for development. But line-like singularities are

not unknown, and some of the models proposed have an unmanageably high dimensionality (e.g., N =26). One lesson which can be
drawn, though, from N > 5D theory is that there is no holy value of N
which is to be searched for as if it were a shangri-la of physics.
Rather, the value of N is to be chosen on utilitarian grounds, in accordance with the physics to be studied.
1.4
Field Equations in N 4 Dimensions

Just as Maxwells equations provided the groundwork for
Einsteins equations, so should general relativity be the foundation for

field equations that use more than the 4 dimensions of spacetime.
Einsteins field equations are frequently presented as a match
between a geometrical object G and a physical object T , via a
coupling constant , in the form G = T ( , = 0,123) . Here the
Einstein tensor G R Rg / 2 depends on the Ricci tensor, the
Ricci scalar and the metric tensor, the last defining small intervals in
4D by a quadratic line element ds 2 = g dx dx .
The energy-
momentum tensor T depends conversely on common properties of

matter such as the density and pressure p, together with the
4-velocities u dx / ds . However, even Einstein realized that this
split between geometry and matter is subjective and artificial. One
example of this concerns the cosmological constant . This was

11
originally added to the left-hand side of the field equations as a geometrical term g , whence the curvature it causes in spacetime corresponds to a force per unit mass (or acceleration) rc 2 / 3 , where r
is the distance from a suitably chosen origin of coordinates. But
nowadays, it is commonly included in the right-hand side of the field
equations as an effective source for the vacuum, whose equation of
state is pv = v c 2 , where v = c 2 / 8 G corresponds to the density of a non-material medium. (Here we take the dimensions of
as length-2 and retain physical units for the speed of light c and gravitational constant G, so the coupling constant in the field equations is
= 8 G / c 4 .) The question about where to put is largely one of

semantics. It makes little difference to the real issue, which is to obtain the g or potentials from the field equations.
The latter in traditional form are
1
8 G
R Rg + g = 4 T
c
2
(1.1)
Taking the trace of this gives R = 4 8 G / c 4 T
where
T g T . Using this to eliminate R in (1.1) makes the latter read
R =
8 G
1
T T g + g
4
c
2

(1.2)
Here the cosmological constant is treated as a source term for the

vacuum, along with the energy-momentum tensor of ordinary matter. If there is none of the latter then the field equations are
R = g
(1.3)
These have 10 independent components (since g is symmetrical).

They make it clear that measures the mean radius of curvature of a
4D manifold that is empty of conventional sources, i.e. vacuum. If
there are no sources of any kind or if the ordinary matter and vacuum fields cancel as required by certain symmetries then the field
equations just read
R = 0
(1.4)
It is these equations which give rise to the Schwarzschild and other

solutions of general relativity and are verified by observations.
The field equations of 5D theory are taken by analogy with
(1.4) to be given by
R AB = 0 ( A, B = 0,123, 4 )
(1.5)
Here the underlying space has coordinates x A = (t , xyz, l ) where the

last is a length which is commonly taken to be orthogonal to spacetime. The associated line element is dS 2 = g AB dx A dx B , where the
5D metric tensor now has 15 independent components, as does (1.5).
However, the theory is covariant in its five coordinates, which may be

13
chosen for convenience. Thus a choice of coordinate frame, or gauge,

may be made which reduces the number of g AB to be determined
from 15 to 10. This simplifies both the line element and the field
equations.
The electromagnetic gauge was used extensively in earlier
work on 5D relativity, since it effectively separates gravity and electromagnetism. A more modern form of this expresses the 5D line
element as parts which depend on g (akin to the Einstein gravitational potentials with associated interval ds 2 = g dx dx ), (a
scalar field which may be related to the Higgs field by which particles
acquire their masses), and A (related to the Maxwell potentials of
classical electromagnetism). The 5D line element then has the form
dS 2 = ds 2 + 2 ( dx 4 + A dx )
(1.6)
Here = 1 determines whether the extra dimension g 44 = 2 is

spacelike or timelike: both are allowed by the mathematics, and we
will see elsewhere that = 1 is associated with particle-like behaviour while = +1 is associated with wave-like behaviour. Most work
has been done with the former choice, so we will often assume that
the 5D metric has signature (+ ) . Henceforth, we will also absorb the constants c and G by a suitable choice of units. Then the dynamics which follows from (1.6) may be investigated by minimizing

the 5D interval, via dS = 0 (Wesson 1999, pp. 129-153). In
general, the motion consists of the usual geodesic one found in Einstein theory, plus a Lorentz-force term of the kind found in Maxwell
theory, and other effects due to the extended nature of the geometry
including the scalar field. Further results on the dynamics, and the
effective 4D energy-momentum tensor associated with the offdiagonal terms in line elements like (1.6), have been worked out by
Ponce de Leon (2002). We eschew further discussion of metrics of
this form, however, to concentrate on a more illuminating case.
The gauge for neutral matter has a line element which can be
written
dS 2 = g ( x , l ) dx dx + 2 ( x , l )
(1.7)
In this we have set the electromagnetic potentials ( g 4 ) to zero, but

the remaining degree of coordinate freedom has been held in reserve.
(It could in principle be used to flatten the scalar potential via
g 44 = 1 , but while we will do this below it is instructive to see what

effects follow from this field.) The components of the 5D Ricci tensor for metric (1.7) have wide applicability. They are:
R = 4 R
, ;
,4 g ,4
2 2
+ g g ,4 g ,4
g ,44
g g ,4 g ,4

R4
15
g g ,4 g g ,4
g g , g g ,4
=
+
2
4
g g , g g ,4
R44 =
g ,4 g ,4
,4 g g ,4
2
g g ,44
2
g g g ,4 g ,4
4
(1.8)
Here a comma denotes the ordinary partial derivative, a semicolon

denotes the ordinary 4D covariant derivative,
1
g , ; and
2 . Superscripts are used here and below for the 5D tensors

2
and their purely 4D parts, whenever there is a risk of confusion.

When the components (1.8) are used with the 5D field equations
(1.5), it is clear that we obtain tensor, vector and scalar equations
which have distinct applications in physics.
The tensor components of (1.8), in conjunction with the 5D
field equations R AB = 0 (1.5), give the 10 field equations of Einsteins general relativity. The method by which this occurs is by now
well known (Wesson and Ponce de Leon 1992). In summary, we
form the conventional 4D Ricci tensor, and with it and the 4D Ricci
scalar construct the 4D Einstein tensor G 4 R 4 R g / 2 . The

remaining terms in 5 R of (1.8) are then used to construct an effective or induced 4D energy-momentum tensor via G = 8 T .
Several instructive results emerge during this process. For example,
the 4D scalar curvature just mentioned may be shown using all of
(1.8) to ge given by
4
R=
[g
,4
g , 4 + g g , 4
)]
2
(1.9)
This relation has been used implicitly in the literature, but explicitly
as here it shows that: (a) What we call the curvature of 4D spacetime
can be regarded as the result of embedding it in an x 4 -dependent 5D
manifold; (b) the sign of the 4D curvature depends on the signature of
the 5D metric; (c) the magnitude of the 4D curvature depends
strongly on the scalar field or the size of the extra dimension
(g
44
= 2 ) , so while it may be justifiable to neglect this in astro-
physics (where the 4D curvature is small) it can be crucial in cosmology and particle physics. Another instructive result concerns the form
of the 4D energy-momentum tensor. It is given by
8 T =
, ;
,4 g ,4
g g ,4 g ,4
2
g ,44 + g g ,4 g ,4
g
2
. (1.10)
g
g
g
g
+
(
)
,4
,4
,4

17
This relation has been used extensively in the literature, where it has
been shown to give back all of the properties of ordinary matter (such
as the density and pressure) for standard solutions. However, it has
further implications, and shows that: (a) What we call matter in a
curved 4D spacetime can be regarded as the result of the embedding
in an x 4 -dependent (possibly flat) 5D manifold; (b) the nature of the
4D matter depends on the signature of the 5D metric; (c) the 4D
source depends on the extrinsic curvature of the embedded 4D spacetime and the scalar field associated with the extra dimension, which
while they are in general mixed correspond loosely to ordinary matter
and the stress-energy of the vacuum. In conclusion for this paragraph,
we see that a 5D manifold which is apparently empty contains a
4D manifold with sources, where the tensor set of the 5D field equations corresponds to the 4D Einstein equations of general relativity.
The vector components of (1.8), in conjunction with (1.5), can
be couched as a set of conservation equations which resemble those
found in Maxwellian electromagnetism and other field theories. They
read
P; = 0
(1.11)
where the 4-vector concerned is defined via
1
g g , 4 g g , 4
2
(1.12)
These are usually easy to satisfy in the continuous fluid of inducedmatter theory, and are related to the stress in the surface (x4 = 0) of

membrane theory with the Z2 symmetry (see below). It should be

noted that these relations do not come from some external criterion
such as the minimization of the line element, but are derived from and
are an inherent part of the field equations.
The scalar or last component of (1.8), when set to zero in accordance with the field equations (1.5), yields a wave-type equation
for the potential associated with the fifth dimension g 44 = 2 in

the metric (1.7). It is
g ,4 g ,4
,4 g g ,4
+ g g ,44
. (1.13)
Here as before g , ; and some of the terms on the righthand side are present in the energy-momentum tensor of (1.10). In
fact, one can rewrite (1.13) for the static case as a Poisson-type equation with an effective source density for the -field. In general (1.13)
is a wave equation with a source induced by the fifth dimension.
Let us now leave the gauge for neutral matter (1.7) and focus
on a special case of it, called the canonical gauge. This was the
brainchild of Mashhoon, who realized that if one factorizes the 4D
part of a 5D metric in a way which mimics the use of cosmic time in
cosmology, significant simplification follows for both the field equations and especially the equations of motion (Mashhoon, Liu and
Wesson 1994). The efficacy of this gauge is related to the fact that a
quadratic factor in l on the 4D part of a 5D model has algebraic con-

19
sequences similar to those of a quadratic factor in t on the 3D part of a

4D cosmological model. The latter case, in the context of 4D Friedmann-Robertson-Walker (FRW) cosmologies, is known as the Milne
universe. This has several interesting properties (Rindler 1977). We
will come back later to the Milne universe as a lower-dimensional
example of highly-symmetric 5D manifolds. For now, we note the
form of the metric and summarize its properties.
The 5D canonical metric has a line element given by
dS 2 =
l2
g x , l ) dx dx dl 2
2 (
L
(1.14)
where x 4 = l is the extra coordinate and L is a constant length introduced for the consistency of physical dimensions. There is an extensive literature on (1.14), both with regard to solutions of the field
equations (1.5) and the equations of motion which follow from minimizing the interval S in (1.14). Some of the consequences of (1.14)
can be inferred from what we have already learned, while some will
become apparent from later study. But for convenience we here
summarize all of its main properties following Wesson (2002):
(a) Mathematically (1.14) is general, insofar as the five available coordinate degrees of freedom have been used to set g 4 = 0, g 44 = 1 .
Physically, this removes the potentials of electromagnetic type and
flattens the potential of scalar type. (b) The metric (1.14) has been
extensively used in the field equations, and many solutions are
known. These include solutions for the 1-body problem and cosmol-

ogy which have acceptable dynamics and solutions with the opposite
sign for g 44 which describe waves. (c) When g / l = 0 in (1.14),
the 15 field equations R AB = 0 of (1.5) give back the Einstein equations as described above, now in the form G = 3 g / L2 . These in
general identify the scale L as the characteristic size of the 4-space.
For the universe, the last-noted relations define an Einstein space with
= 3 / L2
(1.15)
which identifies the cosmological constant. (d) This kind of local

embedding of a 4D Riemann space in a 5D Ricci-flat space can be
applied to any N, and is guaranteed by Campbells theorem. We will
take this up in more detail below. (e) The factorization in (1.14) says
in effect that the 4D part of the 5D interval is (l / L ) ds , which defines
a momentum space rather than a coordinate space if l is related to m,
the rest mass of a particle. This has been discussed in the literature as
a way of bridging the gap between the concepts of acceleration as
used in general relativity, and force (or change of momentum) as used
in quantum theory. (f) Partial confirmation of this comes from a
study of the 5D geodesic and a comparison of the constants of the
motion in 5D and 4D. In the Minkowski limit, the energy of a parti-
cle moving with velocity v is E = l 1 v 2
1/ 2
in 5D, which agrees
with the expression in 4D if l = m. (g) The five components of the

geodesic equation for (1.14) split naturally into four spacetime com-

21
ponents and an extra component. For g / l 0 , the former contain terms parallel to the 4-velocity u , which do not exist in 4D general relativity. We will look into this situation later. But we note now
that for g / l = 0 , the motion is not only geodesic in 5D but geodesic in 4D, as usual. Indeed, for g / l = 0 , we recover the 4D
Weak Equivalence Principle as a kind of symmetry of the 5D metric.
The preceding list of consequences of the canonical metric
(1.14) shows that it implies departures from general relativity when
its 4D part depends on the extra coordinate, but inherits many of the
properties of Einsteins theory when it does not. In the latter case, the
4D cosmological constant is inherited from the 5D scaling, and has a
value = 3 / L2 depending on the signature of the extra dimension
( = 1) .
This is a neat result, and elucidates the use of de Sitter and
anti-de Sitter spaces in approaches to cosmology and particle production, which use quantum-mechanical approaches such as tunneling.
However, in general we might expect the potentials of spacetime to
depend on the extra coordinate. Both for this case as in (1.14), and
for the case where the scalar potential is significant as in (1.7), the
vacuum will have a more complicated structure than that implied by
the simple cosmological constant just noted. It was shown in (1.10)
that in general the effective 4D energy-momentum tensor for neutral
matter in 5D theory contains contributions from both ordinary matter
and the vacuum. Ordinary matter (meaning material particles and
electromagnetic fields) displays an enormous complexity of structure.

Vacuum matter (meaning the scalar field and virtual particles which
defy Heisenbergs uncertainty relation) may display a corresponding
complexity of structure. To use a clich, 5D induced-matter theory
implies that we may have only scratched the surface of matter.
Membrane theory uses an exponential rather than the quadratic of (1.14) to factorize the 4D part of a 5D metric. Thus a generalized form of the type of metric considered by Randall and Sundrum
(1998, 1999) is
dS 2 = e F ( l ) g dx dx dl 2
(1.16)
Here F ( l ) is called the warp factor, and is commonly taken to depend on the cosmological constant and the extra coordinate x 4 = l
in such a way as to weaken gravity away from the brane (l = 0 ) . Particle interactions, by comparison, are stronger by virtue of being confined to the brane, which is effectively the focus of spacetime. An
important aspect of (at least) the early versions of brane theory is the
assumption of Z2 symmetry, which means in essence that the physics
is symmetric about the hypersurface l = 0. This prescription is simple
and effective, hence the popularity of membrane theory. However, a
comparison of (1.16) and (1.7) shows that the former is merely a special case of the latter, modulo the imposition of the noted symmetry.
In fact, examination shows that membrane theory and induced-matter
theory are basically the same from a mathematical viewpoint, even if
they differ in physical motivation. The most notable difference is that
for membrane theory particles are confined to the spacetime hypersur-

23
face by the geometry, which is constructed with this in mind; whereas

for induced-matter theory particles are only constrained by solutions
of the 5D geodesic equation, and can wander away from spacetime at
a slow rate governed by the cosmological constant or oscillate around
it. That the field equations of membrane theory and space-timematter theory are equivalent was shown by Ponce de Leon (2001).
His work makes implicit use of embeddings, and we defer a discussion of these plus the connection between brane and STM worlds to
the next section.
Embeddings must, however, play an important role in the extension of 5D theories to those of even higher dimension. That this is
so becomes evident when we reflect on the preceding discussion. In
it, we have morphed from 4D general relativity with Einsteins equations in the forms (1.1)(1.4), to 5D relativity with the apparently
empty field equations (1.5). These lead us to consider the electromagnetic gauge (1.6) and the gauge for neutral matter (1.7). The latter has
associated with it the 5D Ricci components (1.8), which imply the 4D
Ricci scalar (1.9) and the effective 4D energy-momentum tensor
(1.10). The latter balances the Einstein equations, and leaves us with
the 4 vector terms which satisfy (1.11) by virtue of (1.12), plus the 1
scalar wave equation (1.13). When the 5D metric has a 4D part which
is factorized by a quadratic in the extra coordinate, we obtain the canonical metric (1.14), which leads us to view the cosmological constant (1.15) as a scale inherited from 5D. When alternatively the 5D
metric has a 4D part which is factorized by an exponential in the extra

coordinate, we obtain the warp metric (1.16), which leads us to view

spacetime as a singular surface in 5D. All of these results are entrained in the sense that they follow from the smooth embedding of
4D in 5D. Certain rules of differential geometry underly this embedding. The main one of these is a theorem of Campbell (1926), which
was revitalized by Tavakol and coworkers, who pointed out that it
also constrains the reduction from general relativity in 4D to models
of gravity in 3D and 2D which may be more readily quantized (see
Rippl, Romero and Tavakol 1995). It is not difficult to see how to
extend the formalism outlined above for N > 5, so yielding theories of
supersymmetry, supergravity, strings and beyond. But in so doing
there is a danger of sinking into an algebraic morass. An appreciation
of embedding theorems can help us avoid this and focus on the
physics.
1.5
A Primer on Campbells Theorem

Embedding theorems can be classified as local and global in
nature. We are primarily concerned with the former because our field
equations are local. (The distinction is relevant, because global theorems are more difficult to establish; and since they may involve
boundary conditions, harder to satisfy.) There are several local embedding theorems which are pertinent to ND field theory, of which the
main one is commonly attributed to Campbell (1926). He, however,
only outlined a proof of the theorem in a pedantic if correct treatise on
differential geometry. The theorem was studied and established by

25
Magaard (1963), resurrected as noted above by Rippl, Romero and

Tavakol (1995), and applied comprehensively to gravitational theory
by Seahra and Wesson (2003). The importance of Campbells theorem is that it provides an algebraic method to proceed up or down the
dimensionality ladder N of field theories like general relativity which
are based on Riemannian geometry. Nowadays, it is possible to prove
Campbells theorem in short order using the lapse-and-shift technique
of the ADM formalism. The latter also provides insight to the connection between different versions of 5D gravity, such as inducedmatter and membrane theory.
We will have reason to appeal to
Campbells theorem at different places in our studies of 5D field theory. In the present section, we wish to draw on results by Ponce de
Leon (2001) and Seahra and Wesson (2003), to give an ultra-brief
account of the subject.
Campbells theorem in succinct form says: Any analytic
Riemannian space Vn ( s, t ) can be locally embedded in a Ricci-flat
Riemannian space Vn +1 ( s + 1, t ) or Vn +1 ( s, t + 1) .
We are here using the convention that the small space has
dimensionality n with coordinates running 0 to n 1, while the large
space has dimensionality n +1 with coordinates running 0 to n. The
total dimensionality is N = 1+ n, and the main focus is on N = 5.
To establish the veracity of this theorem (in a heuristic fashion at least), and see its relevance (particularly to the theories considered in the preceding section), consider an arbitrary manifold n in a

Ricci-flat space Vn +1 . The embedding can be visualized by drawing a

line to represent n in a surface, the normal vector n A to it satisfying
n n n A n A = = 1 . If eA form an appropriate basis and the extrinsic curvature of N is K , the ADM constraints read
GAB n A n B =
1
R + K K K 2 ) = 0
(
2
GAB eA n B = K; K , = 0
(1.17)
These relations provide 1 + n equations for the 2 n ( n + 1) / 2 quantities g , K . Given an arbitrary geometry g for n , the constraints therefore form an under-determined system for K , so infinitely many embeddings are possible. This implies that the embedding of a system of 4D equations like (1.1)(1.4) in a system of 5D
equations like (1.5) is always possible.
This demonstration of Campbells theorem can easily be extended to the case where Vn +1 is a de Sitter space or anti-de Sitter
space with an explicit cosmological constant, as in brane theory. De-
pending on the application, the remaining n ( n + 1) ( n + 1) = n 2 1
degrees of freedom may be removed by imposing initial conditions on

the geometry, physical conditions on the matter, or conditions on a
boundary.

27
The last is relevant to membrane theory with the Z2 symmetry.

To see this, let us consider a fairly general line element with
dS 2 = g ( x , l ) dx dx + dl 2 where g = g (x ,+l ) for l 0
and g = g x ,l for l 0 in the bulk (Ponce de Leon 2001).

Non-gravitational fields are confined to the brane at l = 0, which is a
singular surface. Let the energy-momentum in the brane be represented by ( l ) S AB (where S AB n A = 0 ) and that in the bulk by TAB.
Then the field equations read GAB = ( l ) S AB + TAB where is a
5D coupling constant.
The extrinsic curvature discussed above
changes across the brane by an amount K ( l >0 ) K ( l <0 )

which is given by the Israel junction conditions. These imply
= S Sg
3
(1.18)
But the l =0 plane is symmetric, so
K ( l >0 ) = K ( l <0 ) = S Sg
2
3
(1.19)
This result can be used to evaluate the 4-tensor
P K Kg =

(1.20)
However,
(g
,4
is
actually
identical
to
the
4-tensor
g g g ,4 ) / 2 of induced-matter theory, which we noted
above in (1.12). It obeys the field equations P; = 0 of (1.11), which

are a subset of R AB = 0 . That is, the conserved tensor P of induced-matter theory is essentially the same as the total energymomentum tensor in Z2-symmetric brane theory. Other correspondences can be established in a similar fashion.
The preceding exercise confirms the inference that inducedmatter theory and membrane theory share the same algebra, and helps
us understand why matter in 4D can be understood as the consequence of geometry in 5D.
1.6
Conclusion
In this chapter, we have espoused the idea that extra dimen-
sions provide a way to better understand known physics and open a

path to new physics.
The template is Einsteins general relativity, which is based
on a fusion of the primitive dimensions of space and time into 4D
spacetime. The feasibility of extending this approach to 5D was
shown in the 1920s by Kaluza and Klein, and if we discard their restrictive conditions of cylindricity and compactification we obtain a
formalism which many researchers believe can in principle offer a
means of unifying gravity with the forces of particle physics (Section
1.2). 5D is not only the simplest extension of general relativity, but is

29
also commonly regarded as the low-energy limit of higher-N theories.

Most work has been done on two versions of 5D relativity which are
similar mathematically but different physically. Induced-matter (or
space-time-matter) theory is the older version. It views 4D mass and
energy as consequences of the extra dimension, so realizing the dream
of Einstein and others that matter is a manifestation of geometry.
Membrane theory is the newer version of 5D relativity. It views 4D
spacetime as a hypersurface or brane embedded in a 5D bulk, where
gravity effectively spreads out in all directions whereas the interactions of particles are confined to the brane and so stronger, as observed. These theories are popular because they allow of detailed calculations, something which is not always the case with wellmotivated but more complicated theories for N > 5 (Section 1.3). The
field equations of all theories in N 4 dimensions have basically the
same structure, and this is why we treated them together in Section
1.4. There we concentrated again on the case N = 5, paying particular
attention to the equations which allow us to obtain the 15 components
of the metric tensor. In the classical view, these are potentials, where
the 10-4-1 grouping is related to the conventional split into gravitational, electromagnetic and scalar fields. In the quantum view, the
corresponding particles are the spin-2 graviton, the spin-1 photon and
the spin-0 scalaron. The extension of the metric and the field equations to N > 5 is obvious, in which case other particles come in.
However, the extension of general relativity to N > 4 needs to be
guided by embedding theorems. The main one of these dates again

from the 1920s, when it was outlined by Campbell. The plausibility

of Campbells theorem can be shown in short order using modern
techniques, as can the mathematical equivalence of induced-matter
theory and membrane theory (Section 1.5). In summary, the contents
of this chapter provide a basis for writing down the equations for
2 < N < and deriving a wealth of physics.

In pursuing this goal, however, some fundamental questions
arise. In studying (say) 5D relativity, we introduce an extra coordinate (x4 = l ), and an extra metric coefficient or potential (g44). The
two are related, and by analogy with proper distance in the ordinary
3D space of a curved 4D manifold we can define
g ( x , l ) dl
44
1/ 2
as
the size of the extra dimension. Even at this stage, two issues arise
which need attention.
What is the nature of the fifth coordinate? Possible answers
are as follows: (1) It is an algebraic abstraction. This is a conservative but sterile opinion. It implies that l figures in our calculations,
but either does not appear in our final answer, or is incapable of
physical interpretation once we arrive there. (2) It is related to mass.
This is the view of induced-matter theory, where quantities like the
density and pressure of a fluid composed of particles of rest mass m
can be calculated as functions of l from the field equations. Closer
inspection shows that for the special choice of gauge known as the
pure-canonical metric, l and m are in fact the same thing. We will
return to this possibility in later chapters, but here note that in this interpretation the scalar field of classical 5D relativity is related to the

31
Higgs (or mass-fixing) field of quantum theory. (3) It is a length perpendicular to a singular hypersurface. This is the view of membrane
theory, where the hypersurface is spacetime. It is an acceptable opinion, and as we have remarked it automatically localizes the 4D world.
But since we are made of particles and so confined to the hypersurface, our probes of the orthogonal direction have to involve quantities
related to gravity, including masses.
The other issue which arises at the outset with 5D relativity
concerns the size of the extra dimension, defined as above to include
both the extra coordinate and its associated potential. This is a separate, if related, issue to what we discussed in the preceding paragraph.
We should recall that even in 4D relativity, drastic physical effects
can follow from the mathematical behaviour of the metric coefficients. (For example, near the horizon of an Einstein black hole in
standard Schwarzschild coordinates, the time part of the metric
shrinks to zero while the radial part diverges to infinity.) This issue is
often presented as the question: Why do we not see the fifth dimension? Klein tried to answer this, as we have seen, by arguing that the
extra dimension is compactified (or rolled up) to a microscopic size.
So observing it would be like looking at a garden hose, which appears
as a line from far away or as a tube from close up. Since distances are
related to energies in particle experiments, we would only expect the
finite size of the fifth dimension to be revealed in accelerators of
powers beyond anything currently available. This is disappointing.
But more cogently, and beside the fact that it leads to conflicts, many

researchers view compactification in its original form as a scientific

cop-out. The idea can be made more acceptable, if we assume that
the universe evolves in such a way that the fifth dimension collapses
as the spatial part expands. But even this is slightly suspect, and better alternatives exist.
Thus for membrane theory the problem is
avoided at the outset, by the construction of a 5D geometry in which

the world is localized on a hypersurface. For induced-matter theory,
particles are constrained with respect to the hypersurface we call
spacetime by the 5D equations of motion. In the latter theory, modest
excursions in the extra dimension are in fact all around us to see, in
the form of matter.
Let us assume, for the purpose of going from philosophy to
physics, that a fifth dimension may exist and that we wish to demonstrate it. We already know that 4D general relativity is an excellent
theory, in that it is soundly based in logic and in good agreement with
observation. We do not desire to tinker with the logic, but merely
extend the scope of the theory. Our purpose, therefore, is to look for
effects which might indicate that there is something bigger than
spacetime.
References
Arkani-Hamed, N., Dimopoulos, S., Dvali, G. 1998, Phys. Lett. B429,
263.
Arkani-Hamed, N., Dimopoulos, S., Dvali, G. 1999, Phys. Rev. D59,
086004.

33
Campbell, J.E. 1926, A Course of Differential Geometry (Clarendon,

Oxford).
Gubser, S.S., Lykken, J.D. 2004, Strings, Branes and Extra Dimensions (World Scientific, Singapore).
Halpern, P. 2004, The Great Beyond (Wiley, Hoboken).
Magaard, L. 1963, Ph.D. Thesis (Kiel).
Mashhoon, B., Liu, H., Wesson, P.S. 1994, Phys. Lett. B331, 305.
Ponce de Leon, J. 2001, Mod. Phys. Lett. A16, 2291.
Ponce de Leon, J., 2002, Int. J. Mod. Phys. 11, 1355.
Randall, L., Sundrum, R. 1998, Mod. Phys. Lett. A13, 2807.
Randall, L., Sundrum, R. 1999, Phys. Rev. Lett. 83, 4690.
Rindler, W. 1977, Essential Relativity (2nd. ed., Springer, Berlin).
Rippl, S., Romero, C., Tavakol, R. 1995, Class. Quant. Grav. 12,
2411.
Seahra, S.S., Wesson, P.S. 2003, Class. Quant. Grav. 20, 1321.
Szabo, R.J. 2004, An Introduction to String Theory and D-Brane Dynamics (World Scientific, Singapore).
Wesson, P.S., Ponce de Leon, J. 1992, J. Math. Phys. 33, 3883.
Wesson, P.S. 1999, Space-Time-Matter (World Scientific, Singapore).
Wesson, P.S. 2002, J. Math. Phys. 43, 2423.


Higher-Dimensional Physics: Old English Saying

Uploaded by

Copyright:

Available Formats

Higher-Dimensional Physics: Old English Saying

Uploaded by

Document Information

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

Higher-Dimensional Physics: Old English Saying

Uploaded by

Copyright:

Available Formats

1.

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

of Einstein gravity and is widely regarded as the low-energy limit of

Dimensions Then and Now

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

of the time to an extra coordinate x 4 ct via the use of the speed of

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

This left 14 metric coefficients,

( x , = 0,123 for t , xyz ). These 14 coefficients were determined

14 field equations. The latter turned out to be the 10 Einstein

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

problem, to which attention will be given later.) Further, the cylinder

Hence modern Kaluza-Klein theory,

which is algebraically rich and exists in several versions.

physical motivation. However, all are based on Einsteins theory of

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

the energy-momentum tensor T , which was on the same footing

tensor G . That this is possible in practice was proved using an

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

Membrane theory is based on a 5D manifold in which there is

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

offers a broad field for development. But line-like singularities are

Field Equations in N 4 Dimensions

Einsteins equations, so should general relativity be the foundation for

momentum tensor T depends conversely on common properties of

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

= 8 G / c 4 .) The question about where to put is largely one of

Taking the trace of this gives R = 4 8 G / c 4 T

T g T . Using this to eliminate R in (1.1) makes the latter read

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

Here the cosmological constant is treated as a source term for the

These have 10 independent components (since g is symmetrical).

It is these equations which give rise to the Schwarzschild and other

Here the underlying space has coordinates x A = (t , xyz, l ) where the

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

chosen for convenience. Thus a choice of coordinate frame, or gauge,

Here = 1 determines whether the extra dimension g 44 = 2 is

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

the 5D interval, via dS = 0 (Wesson 1999, pp. 129-153). In

In this we have set the electromagnetic potentials ( g 4 ) to zero, but

g 44 = 1 , but while we will do this below it is instructive to see what

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

Here a comma denotes the ordinary partial derivative, a semicolon

2 . Superscripts are used here and below for the 5D tensors

and their purely 4D parts, whenever there is a risk of confusion.

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

= 2 ) , so while it may be justifiable to neglect this in astro-

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

where the 4-vector concerned is defined via

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

membrane theory with the Z2 symmetry (see below). It should be

for the potential associated with the fifth dimension g 44 = 2 in

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

sequences similar to those of a quadratic factor in t on the 3D part of a

FIVE-DIMENSIONAL PHYSICS - Classical and Quantum Consequences of Kaluza-Klein Cosmology

which identifies the cosmological constant. (d) This kind of local

cle moving with velocity v is E = l 1 v 2