^{1}

^{1}

^{*}

We show how the metric of a five-dimensional hyperspace-time can be used to model the quantum nature of electromagnetic interactions. The space-time neighborhood of the point where such an interaction takes place bends according to the curl and the derivative of the local electromagnetic four-potential, both calculated in the direction of the latter. In this geometric setting, the presence of a non-gravitational field is needed to induce the discretization of any gravitational field. We also exploit two variants of the classical Kaluza-Klein five-dimensional theory to obtain coupled generalizations of Einstein’s and Maxwell’s equations. The first variant involves an unspecified scalar field that may be related to the inflaton. The equations of the second variant show a direct interdependency of gravitation and electromagnetism that would emerge or be activated through the production of electromagnetic waves.

In Kaluza’s [

With the above properties assumed to apply to the fifth dimension, Kaluza and Klein were able to obtain a formal unification of gravitation and electromagnetism. However, this unification is unsatisfactory because it does not show how gravitational and electromagnetic fields interact. A deeper unification of gravitation and electromagnetism requires a revision of the postulate applying to the geometric structure of the fifth dimension. As it is, the postulate concerns only local symmetries of the group U(1) used to characterize the continuous functions describing classical electromagnetism. A more complete geometric modeling of electromagnetism must also allow expressing the discontinuity of certain functions associated with an electrically charged particle, such as its energy when it is involved in an interaction. To model this discontinuous variation, we shall postulate that the geometric structure of the fifth dimension has a global symmetry making it multiconnected [3-5]. The fifth dimension then decomposes into equal length intervals, each topologically equivalent to the circle S^{1}. This multiconnectivity generates a structure with a discrete group of symmetry G over the fifth dimension. The group G establishes an order into the set of these intervals and renders homologous points of the fifth dimension. The multiconnectivity of the fifth dimension reduces its metric measure to that of only one of its finite intervals. Thus it is no longer necessary for the fifth dimension to have a strong curvature to be unobservable. The flexibility to choose one of these intervals as the geometric framework necessary to describe the continuous properties of electromagnetism is a degree of freedom similar to those of continuous gauges in standard field theories: here it corresponds to a discontinuous gauge.

This five-dimensional hyperspace-time can be seen as a fibre bundle whose base is space-time and fibres are copies of the fifth dimension. Above each point of the trajectory of a charged particle in space-time, a finite interval of the fifth dimension must be chosen to describe the properties of the continuous symmetries related to a possible interaction with an exterior electromagnetic field. Without such a field, the particle is free and the finite interval can be arbitrarily chosen above each point of its trajectory: this choice determines a horizontal section in the hyperspace-time fibre bundle. When the particle undergoes an electromagnetic interaction at a given point in space-time, the quantum effect on the particle generates a modification, over this point and with respect to the horizontal section, of the finite interval needed to describe the continuous properties related to the interaction. Since neighborhoods of homologous points belonging to different finite intervals have the same local geometry, the continuous properties of electromagnetism are invariant under changes of this interval. At the point of space-time where the interaction takes place, this change of interval implies a discontinuous variation of the hyperspace-time metric along the fifth dimension. Outside this point, the hyperspace-time metric stays independent of the fifth dimension, in accordance with the original Kaluza-Klein postulate. Our postulate thus differs from that of Kaluza and Klein only at the interaction point. Relying on this new postulate, the present paper first aims to show how the metric of a five-dimensional hyperspace-time can be used to unify gravitational and electromagnetic theories, and simultaneously explain qualitatively the quantum nature of electromagnetic interactions. In its second part, we shall reconsider the field interaction aspect of the original Kaluza-Klein theory and obtain coupled generalizations of Einstein’s and Maxwell’s equations.

In this paper, Latin indices are for space-time and run from 1 to 4. Greek letters are hyperspace-time indices and run from 1 to 5. The signatures of space-time and hyperspace-time metrics are respectively (−1,1,1,1) and (−1,1,1,1,1). A line over a space-time symbol will indicate that we consider its hyperspace-time version. The electromagnetic tensor is defined by F_{mn}=A_{n,m}-A_{m,n} where A_{m} designates the covariant components of the electromagnetic four-potential.

In a completely geometric unified field theory, the hyperspace-time metric should allow to describe all properties of fields. To express the fact that physics seems to depend only on space-time coordinates, Kaluza and Klein postulated that the components are independent of x^{5}. They also postulated that under the coordinate transformation the new space-time coordinates are independent of x^{5}. Therefore, the original Kaluza-Klein postulate concerning the fifth dimension is

Taking into account these equations, it is easily shown that the most general expression for such a metric is

where (g_{mn}) is the space-time metric, A_{m} is the electromagnetic four-potential and k is an arbitrary constant.

To express mathematically the discrete nature of an electromagnetic interaction occurring at a point x_{0} of hyperspace-time, we assume that the mathematical mechanism modeling the electromagnetic interaction is related only to the hyperspace-time metric components which are independent of the space-time metric. Inspired by the way how electrons change their orbit around a nucleus, we postulate that if varies with respect to x^{5}, then

where designates the Dirac distribution. This new postulate will be said of discretization due to the fifth dimension. To simplify the presentation, we shall set until Section 5. It is then straightforward to deduce the hyperspace-time scalar curvature:

where R is the space-time scalar curvature.

Let S denotes the action associated with Equation (3). We designate by g the determinant of the space-time metric and by the four-dimensional space of coordinates x^{2}, x^{3}, x^{4} and x^{5} between two given values of time x^{1}. Then

By successively varying the space-time metric and the electromagnetic four-potential, the first term on the right-hand side of Equation (4) yields the Einstein and Maxwell equations deduced by Klein. Its second term is an expression of the electromagnetic interaction at x_{0}. Bringing in the Lorenz gauge, Equation (4) becomes

The second term of the right-hand side of Equation (5) shows that the interaction can be described as a multiple of the sum of partial variations of the electromagnetic four-potential’s squared norm evaluated at x_{0} in the direction of its largest variation. This expression can also be written

According to [_{0} in the direction of this four-potential. Depending on the space-time curvature, and on the value of the electromagnetic four-potential in the neighborhood of x_{0}, the change of horizontal section will thus occur more or less rapidly, and with a rotation more or less pronounced. These effects should emerge physically as contributions to the 4-momentum and angular momentum of the particle subject to this interaction.

Observe that since

The space-time metric, i.e. the gravitational field, thus varies discontinuously at x = x_{0} and depends on the value of the electromagnetic 4-vector at this point. If this four-potential is identically zero, the postulate of discretization due to the fifth dimension then reduces to Klein’s original postulate and there is no discretization. The presence of a non-zero electromagnetic four-potential is thus necessary to induce the discretization of the gravitational field.

Let us now try to extend the mathematical mechanism modeling the electromagnetic interaction to all components of the hyperspace-time metric, i.e. to replace Equations (2) with

A straightforward calculation shows that the hyperspace-time scalar curvature then becomes:

This expression involves products of the Dirac distribution, which is impossible within the framework of the theory [

Therefore, the postulate based on Equations (7) does not allow deducing the equations of a discretized gravitational field.

Even if Klein was aware that can be any differentiable scalar function of space-time coordinates, he chose to replace it with the proportionality constant k between and in order to facilitate his calculations. However, this choice led him to the well known unsatisfactory unification of gravitation and electromagnetism. Since Kaluza’s and Klein’s works, several hundred papers were written trying to improve this situation (see e.g. [

The equations of motion in space-time are then obtained by varying the action built up with Equation (8) separately with respect to the three sets of independent dynamical variables g^{mn}, A_{m} and. Direct calculations give, respectively,

where

In this unified fields theory, the Equations (9), (10) and (11) correspond to three different laws of nature. If the function α is constant, then Equations (9) and (10) respectively reduce to the Einstein equations for a spacetime with T_{mn} as energy-momentum tensor, and the Maxwell equations for a free space-time. Since Equation (11) results from the variations of α, this equation does not exist if α is a constant.

Considering a situation where the three laws describe by Equations (9)-(11) apply, we can substitute Equations (10) into Equation (11), and the resulting expression into Equations (9). The space-time where this situation occurs is then described by

(12)

Observe that if k = 0, or if the electromagnetic fourpotential is identically zero, then Equations (12) reduce to

Such an equation may be used to model a universe whose expansion is regulated by the scalar field α.

In the preceding Section, the independence of the scalar field α with respect to the variables g^{mn} and A_{m} was assumed to simplify the calculations. But such a scalar field is hard to find. We shall now consider an easily reachable scalar field by observing that it was arbitrary at the starting point of Klein’s development. If we allow it to depend on the variables g^{mn} and A_{m}, then we can use the electromagnetic energy density of space-time given by. For this scalar field, the equations of motion in space-time are found by varying the action determined by Equation (8) separately with respect to g^{mn} and A_{m}. Straightforward calculations yield, respectively,

where and are the partial derivatives of with respect to and respectively.

One way to solve the system of Equations (13) and (14) is first to rewrite Equations (14) as four partial differential equations of order one for R. The common solution of these equations is an expression of R in terms of the electromagnetic four-potential and of an arbitrary scalar function of the space-time coordinates. By substituting this expression of R into Equations (13), we obtain equations for R_{mn} analogous to Equations (12), but with a more complex expression and an arbitrary function on its right-hand side. This arbitrary function here plays a role similar to that of α in Section 5.

In the first part of this paper, we have shown how the metric of a five-dimensional hyperspace-time can be used to model the quantum nature of electromagnetic interactions. In this framework, the neighborhood of the point where an interaction takes place bends according to the curl and the derivative of the local electromagnetic four-potential, both calculated in the direction of the latter. We have also seen that a non-gravitational field is required to induce the discretization of any gravitational field.

In its second part, we have reconsidered the field interaction aspect of the original Klein’s unifying scheme which led to two independent laws of nature. With the method of Section 5, we have obtained a system of three coupled groups of equations for g_{mn}, A_{m} and α. These groups of equations correspond to three different laws of nature. Due to the coupling between the groups of equations, these laws are not independent: each one will depend on g_{mn}, A_{m} and α. Observe that this unifying scheme calls in a scalar function acting at a middle scale and which must not be of gravitational nor of electromagnetic origin. It is hard to find such a scalar field; at very large scale it could be identified with the field describing the inflationary phase in the early universe.

Envisaging the possibility of controlling the unspecified scalar field of Section 5, we have replaced it in Section 6 with the space-time local density of electromagnetic energy. Other choices are possible. This has led us to a system of two groups of equations corresponding to two interdependent laws of nature. If we consider Equations (13) and (14) as generalizations of Einstein’s equations in the presence of an electromagnetic field and of Maxwell’s equations, respectively, these groups of equations show that it would be possible to influence the gravitational field in a region of space-time by locally changing the electromagnetic field. Reciprocally, a variation of the gravitational field would cause a variation of the electromagnetic field. Since any variation of the electromagnetic field generates electromagnetic waves, the two above predictions could be tested through phenomena involving a large and sudden emission of electromagnetic waves.