Dec 19, 2011 - Since the geometrization of gravitation by General Relativity (GR) where the .... theory are easily seem to be equivalent to Einstein's...

0 downloads 13 Views 248KB Size

Gravitation and Electromagnetism as Geometrical Objects of a Riemann-Cartan Spacetime Structure J. Fernando T. Giglio(1) and Waldyr A. Rodrigues Jr.(2) . (1)

FEA-CEUNSP. 13320-902 - Salto, SP, Brazil.

[email protected] (2) Institute of Mathematics, Statistics and Scientific Computation IMECC-UNICAMP 13083-859 Campinas, SP, Brazil [email protected] or [email protected]

May 29, 2018 Abstract In this paper we first show that any coupled system consisting of a gravitational plus a free electromagnetic field can be described geometrically in the sense that both Maxwell equations and Einstein equation having as source term the energy-momentum of the electromagnetic field can be derived from a geometrical Lagrangian proportional to the scalar curvature R of a particular kind of Riemann-Cartan spacetime structure. In our model the gravitational and electromagnetic fields are identified as geometrical objects of the structure. We show moreover that the contorsion tensor of the particular Riemann-Cartan spacetime structure of our theory encodes the same information as the one contained in Chern-Simons term A ∧ dA that is proportional to the spin density of the electromagnetic field. Next we show that by adding to the geometrical Lagrangian a term describing the interaction of a electromagnetic current with a general electromagnetic field plus the gravitational field, together with a term describing the matter carrier of the current we get Maxwell equations with source term and Einstein equation having as source term the sum of the energy-momentum tensors of the electromagnetic and matter terms. Finally modeling by dust charged matter the carrier of the electromagnetic current we get the Lorentz force equation. Moreover, we prove that our theory is gauge invariant. We also briefly discuss our reasons for the present enterprise.

1

1

Introduction

Since the geometrization of gravitation by General Relativity (GR) where the gravitational field, generated by an energy-momentum tensor1 T ∈ sec T02 M , ¯ τg , ↑i is represented by a particular Lorentzian spacetime structure2 hM, g, ∇, together with Einstein equation, several classical models have been proposed which try to geometrize the description of the electromagnetic field, with the obvious interest in describing both fields, i.e., gravitational and electromagnetic, through a unique principle: the geometric one. This has been tried by generalizing the Lorentzian spacetime structure, i.e., utilizing more general geometries incorporating additional degrees of freedom, which hopefully permits in principle a description of the electromagnetic field as some aspect of the underlie chosen geometry. In this sense, there has been many lines of investigation of this problem, some of the most well known are: (i) Weyl theory, where non metric compatible symmetric connections [51, 52] are used. The resulting geometry has non zero Riemann tensor and a null torsion tensor and is now known as Weyl geometries [58]. We mention also in this class Eddington unified theory which is described by a non metric compatible connection with non null Riemann and torsion tensors [15]. (ii) Theories based on spacetimes with more than four dimensions [33, 36], known as Kaluza-Klein theories. These theories have been studied in the last decades in connection with fiber bundle formulations of the four fundamental interactions [3, 12]. (iii) Introduction of metric compatible connections in four dimensions, other than that the Levi-Civita connection, like, e.g., in the so-called RiemannCartan geometries [56, 9], which in general have non zero Riemann and torsion tensors. (iv) The non symmetric metric theory of Einstein [16]. (v) Theories on Finslerian spaces [18, 6, 60, 61]. (vi) Einstein teleparallel theory3 . The fact is that all these theories are problematic. According to the majority view, Weyl theory received what is thought to be a knockdown by Einstein [1] but according to Eddington, non metricity can not be completely ruled out by experiment if due care is taken [15]. Concerning (ii) it remains always a problem to explain why the extra dimensions are not observable, or why they compactify4 [12]. Concerning (iii), several theories that includes Cartan’s torsion in GR have been proposed. One motivation was to obtain a unified geometrical description V this paper the notation sec Tsr M means section the the Tsr M bundle. Also, sec r T ∗ M means section of the bundle of r-form fields. 2 In the structure hM, g, ∇, ¯ τg , ↑i, the pair hM, gi is called a Lorentzian manifold, M being a 4-dimensional Hausdorff paracompact locally compact manifold and g ∈ sec T02 M a Lorentzian ¯ is the Levi-Civita connection of g, τg ∈ sec V4 T ∗ M metric of signature (1, −1−, 1 − 1). ∇ and (↑) define respectively a spacetime orientation and a time orientation for M . More details about time orientation may be found, e.g., in [54]. 3 See complete list of Einstein papers on teleparallelism in [53]. 4 The problem exists also in modern string theory, for no reason is given for the compatification of the extra dimensions. 1 In

2

of electromagnetism and gravitation [18, 25]. These theories, and that of (i) and (ii), at least to what refers to a classical unified description of the gravitational and electromagnetic fields, are not in general totally accepted because of their failure in obtaining simultaneously the electromagnetic field equations (Maxwell equations), the energy-momentum tensor of the electromagnetic field and the Lorentz force equation for the motion of charged matter as they are known in the physical situations involving gravitation and electromagnetism in a four dimensional universe. Also, some authors5 are of the opinion that RiemannCartan geometry is necessary to describe besides gravitation, also the spinning matter, which is supposed to be the source of the torsion field [9, 28, 25, 29, 30] and it seems also that a non vanishing torsion tensor appears as a necessary ingredient in the gauge formulation of GR when the Poincar´e invariance is taken locally [25, 59, 34, 55, 39]. Besides that in string theory, in the low energy limit, the effective Lagrangian has an antisymmetric field that is interpreted as torsion [19]. Torsion derived from scalar field, vector field or antisymmetric tensor field can be found in Hammond [26] and references cited therein. The non-symmetric theory of Einstein has recently be developed by Moffat [40] with a very different interpretation aiming to describe dark matter, but will not be commented here, nor will we discuss the status of theories that use Finslerian spacetimes. However we comment that Einstein’s teleparallel theory, that he originally interpreted as a unified theory of the gravitational plus the electromagnetic field is indeed a non sequitur. The case is that Einstein’s preferred version of his teleparallel theory has a Lagrangian that is equivalent to the Einstein-Hilbert Lagrangian of GR6 . This according to [53] has been informed by Lancoz to Einstein and put, so to say, an end to teleparallelism as an unified field theory Moreover, it is now known [50] that a theory of the gravitational field can be V1 formulated for the gravitational potentials ga ∈ sec T ∗ M (with at least one of the ga for a = 0, 1, 2, 3 non closed, i.e., F a = dga 6= 0) living on Minkowski spacetime hM, η, D, τη , ↑i and satisfying field equations derived from a postulated Lagrangian density, thus dispensing the geometrical interpretation of gravitation as a Lorentzian or a teleparallel spacetime. The field equations of the theory are easily seem to be equivalent to Einstein’s equations once we introduce ¯ and a field g = ηab ga ⊗gb ∈ sec T02 M together with its Levi-Civita connection ∇ ¯ interpret the structure hM, g, ∇, τg , ↑i as an effective Lorentzian spacetime7 . It is also the case that for the formulation of the electromagnetic field theory we do not even need a metric field defined on a manifold that serves as support 5 Well, this is indeed a polemical view, not endorsed, e.g., by Weinberg. See his exchange of letters with Hehl in Physics Today at: http://ptonline.aip.org/journals/doc/PHTOAD-ft/vol 60/iss 3/16 2.shtml?bypassSSO=1. Anyway there are some proposals in the literature to observe experimentally the existence of torsion, see, e.g., [20, 57]. 6 This is clear, e.g., in [2] where the torsion tensor is used to describe the gravitational field of Einstein’s GR in a particular Riemann-Cartan spacetime structure known nowadays as Weitzenb¨ ock (or teleparallel) spacetime hM, g, ∇, τg , ↑i where the curvature tensor of ∇ is null, and the torsion tensor of ∇ is non null. See also [50]. 7 There are other possibilities also involving more complicated geometrical structures, see, e.g., [44].

3

for that field [31]. Indeed, it is well know that Maxwell equations can be written in a star shape manifold [11] as the compatibility equations for a closed 2-form V2 V3 field F ∈ sec T ∗ M (dF = 0) and a closed current J ∈ sec T ∗ M (dJ = 0). These equations imply in a star shape manifold V1 V2 the existence of a 1-form field A ∈ sec T ∗ M and a 2-form field G ∈ sec T ∗ M such that F = dA and J = −dG in such a way that Maxwell equations read dF = 0, dG = −J

(1)

It is also well known that in general G is related to F through the socalled constitutive equations of the medium. In that sense the gravitational field in GR modeled by a Lorentzian spacetime serves as an effective medium for the propagation of the electromagnetic field and the constitutive equations are given simple by G = ⋆F, (2) g

and thus defining J = ⋆J ∈ sec g

V1

T ∗ M we can write Maxwell equations as

dF = 0, δF = −J ,

(3)

where ⋆ is the Hodge dual operator and δ is the Hodge coderivative operag

tor8 . The intrinsic Maxwell equations even when expressed in a Lorentzian or Riemann-Cartan spacetime structures do not need the use of the covariant derivative operator of those structures for their writing, although they can be formulated with such operators (see below9,10 ). Having said all that we can ask: is it possible, in the same sense that the gravitational field can be describe by the potentials ga living in Minkowski ¯ τg , ↑i or a teleparallel structure spacetime or by a Lorentzian structure hM, g, ∇, hM, g, ∇, τg , ↑i to describe gravitation and electromagnetism geometrically, i.e., through a particular Riemann-Cartan spacetime structure where those fields are represented by some of the geometrical objects associated with that structure? As we shall see, the answer is positive once we use a Riemann-Cartan structure equipped with a particular connection whose contortion tensor is given by Eq.(23) below11 . In our very simple model we are able to obtain a unified description of the gravitational and free electromagnetic fields as geometrical aspects of a particular Riemann-Cartan spacetime, with the Einstein and the free Maxwell 8 Details,

may be found, e.g., in [48]. also [49] for the formulation of Maxwell equations using the the covariant derivative operator of a general Riemann-Cartan spacetime structure. 10 Moreover it can be shown that at least in Minkowski spacetime those equations and energy-momentum conservation of field plus matter imply in a unique coupling between F and J, namely the Lorentz force law [14]. 11 We observe that in many of the papers dealing with the subject of our study torsion is generally taken in particular forms. Besides that let us also add that torsion as resulting from spinning matter has never been experimentally observed [27, 22], although there are proposals to this end [57, 20]. 9 See

4

equations being derived from a geometrical Lagrangian, i.e., a Lagrangian proportional to the scalar curvature R of a particular Riemann-Cartan connection. The main mathematical tools for doing that is presented in Section 2 where we also show that the information contained in the contortion tensor of the particular Riemann-Cartan spacetime structure of our theory is the same as the one contained in the Chern-Simons term A ∧ dA that as well known now [48, 21] is proportional to the spin density of the electromagnetic field. Moreover, we show in Section 3 that by adding to the geometrical Lagrangian an interaction term proportional to J ·A describing the source of the electromagnetic field and its interaction with that field and a term describing the matter carrier of the current we get Maxwell equations with source term and Einstein equations having as source term the sum of the energy-momentum tensors of the electromagnetic and matter terms. In Section 4, modeling by dust charged matter the carrier of the electromagnetic current we get from the nullity of Riemann-Cartan covariant derivative of the sum of energy-momentum tensor of matter plus the electromagnetic field the Lorentz force equation. In Section 5 we show that our theory is well defined by proving its gauge invariance. Finally, in Section 6 we present our conclusions.

2

Maxwell Equations

¯ and In what follows hM, gi as defined above is a Lorentzian manifold. Let ∇ ∇ be respectively the Levi-Civita connection and a particular metric compatible Riemann-Cartan [43] connection of g on M . Let U ⊂ M and hxµ i be a coordinates for U ⊂ M , and heµ = ∂/∂xµ i a basis of T U (µ = 0, 1, 2, 3) and hϑµ = dxµ i the corresponding dual basis i.e., a basis for T ∗ U . We also introduce the reciprocal basis heµ i of heµ i for T M and the reciprocal basis hϑµ i of hϑµ i for T ∗ M , such that g = gµν ϑµ ⊗ ϑν = g µν ϑµ ⊗ ϑν , eµ = g µν eν ,

g µα gαν = δνµ ,

ϑµ = gµν ϑν .

(4)

Moreover we introduce as metric for the cotangent bundle the object g ∈ sec T20 M, g = g µν eµ ⊗ eν = gµν eµ ⊗ eν and define the scalar product of arbitrary 1-form fields X and Y by X · Y = g(X, Y )

(5)

¯ ··λ ¯ Let moreover Γ··λ µν· and Γµν· be the connection coefficients of ∇ and ∇ in the ··λ ··λ ¯ ¯ coordinate basis just introduced, i.e., ∇∂µ ∂ν = Γµν· ∂λ and ∇∂µ ∂ν = Γµν· ∂λ . As ··λ ¯ ··λ it is well know (see, e.g., [37, 48]), Γ µν· and Γµν· are related by ··λ ¯ ··λ Γ··λ µν· = Γµν· + Kµν· ,

5

(6)

where the connection coefficients Γ··λ µν· of the Levi-Civita connection are given by: 1 λα Γ··λ (∂µ gνα + ∂ν gµα − ∂α gµν ) (7) µν· = g 2 ··β ··λ and the Kµν are the components of the contorsion tensor K = Kµν· eβ ⊗ dxµ ⊗ ν 2 12 dx ∈ T M ⊗ sec T1 M defined by :

1 λβ ··ρ ··ρ ··ρ (g gλρ Tµν· − g λβ gνρ Tµλ· − g λβ gµρ Tνλ· ) 2 1 ·· β ·β· β·· = (Tµν· − Tµ·ν + T·νµ ). 2

··β Kµν· :=

(8)

··· However, taking into account that we have the (bastard [23]) symmetry Kµνλ := ··β ··· gβλ Kµν· = −Kµλν we prefer in what follows to take the contortion as the object V V K ∈ 1 T ∗ M ⊗ 2 T ∗ M (which carries the same information as K) defined by

K=

1 ··· µ 1 ··λ µ Kµν· ϑ ⊗ ϑν ∧ ϑλ = Kµνλ ϑ ⊗ ϑν ∧ ϑλ . 2 2

Also, the ··λ ··λ Tµν· = Γ··λ µν· − Γνµ·

(9)

(10) V2

··λ are the components of torsion tensor Θ = 21 Tµν· eλ ⊗ϑµ ∧ϑν ∈ sec T M ⊗ T ∗ M of the Riemann-Cartan connection ∇, but hereVwe will prefer to use as torsion V2 1 ··λ ϑλ ⊗ ϑµ ∧ ϑν ∈ sec T M ⊗ T ∗ M which encodes tensor the object Θ = 21 Tµν· the same information than Θ.

We now proceed by introducing a particular Riemann-Cartan spacetime structure hM, g, ∇, τg , ↑i where the contorsion tensor is defined by K := −C B⊗ F ∈ sec with components13 ··λ ·λ Kµν· = −CBµ Fν· ,

V1

T ∗M ⊗

V2

T ∗M

··· ··· Kµνλ = −CBµ Fνλ = −Kµλν .

(11)

(12)

V and where the constant C and the Bµ , which are the components of B ∈ sec 1 T ∗ M , ·λ are to be determined and where the Fν· := g λα Fνα = −g αλ Fαν = −F·νλ· are the V2 ∗ components of F ∈ sec T M , i.e., F :=

1 1 ·ν µ Fµν ϑµ ∧ ϑν = Fµ· ϑ ∧ ϑν . 2 2

(13)

In what follows we propose to give a physical interpretation for those objects associated to the structure hM, g, ∇, τg , ↑i. 12 Note that this differs from the definition in [28] by a signal and a factor 1/2 due to conventions.used here with are the ones in [48]. 13 This form is analogous to such that is taken in [8] for torsion.

6

Before proceeding to build our theory we recall some formulas that will be used latter. We start calculating the covariant derivative ∇µ of F µν 14 . Using the connection given by Eq.(6) we obtain ··ν µδ ¯ µ F µν + K ··µ F δν + Kµδ ∇µ F µν = ∇ F . µδ

(14)

But the last two terms in Eq.(14) cancel out due to Eq.(12), and we have ¯ µ F µν = ∇µ F µν = ∇

1 (− det g)

1

1 2

∂µ [(− det g) 2 F µν ].

(15)

Next, we observe that from Eq.(14) it follows immediately that ¯ [µ Fνλ] ∇[µ Fνλ] = ∇µ Fνλ + ∇ν Fλµ + ∇λ Fµν = ∇ = ∂µ Fνλ + ∂ν Fλµ + ∂λ Fµν .

(16)

From Eqs.(15) and (16), a natural assumption is to define F = dA, where V1 A ∈ sec T ∗ M , and of course, ¯ µ Aν − ∇ ¯ ν Aµ . Fµν = ∂µ Aν − ∂ν Aµ = ∇

(17)

This suggests to interpret F as the electromagnetic field and A as its potential, and we are going to show that this is indeed the case. Also, Eq.(15) defines in general a conserved current 1-form field J = Jµ ϑµ = V ν J ϑν ∈ sec 1 T ∗ M such that (c being the velocity of light in vacuum) 4π ν ¯ µ F µν = (− det g)− 12 ∂µ [(− det g) 12 F µν ] J := ∇µ F µν = ∇ c

and of course,

1

∂ν [(− det g) 2 J ν ] = 0.

(18)

(19)

With our choise of F the second member of Eq.(16) is null. We then recognize Eq.(16) and Eq.(18) as Maxwell equations written on a Lorentzian spacetime [38, 49].

3

Action Principle, Maxwell and Einstein Equations and their Source Terms

In a Riemann-Cartan spacetime the curvature tensor can be written as (see, e.g., [37, 48]) ···χ ··ρ ··χ ··χ ¯ ···χ + ∇ ¯ µ K ··χ − ∇ ¯ ν K ··χ + K ··ρ Kµρ· Rµνλ· =R − Kµλ· Kνρ· , µνλ· νλ· µλ· νλ·

(20)

14 More precisely we write , e.g., for a tensor t = t·ν ∂ ⊗ dxα ∈ sec T M ⊗ T ∗ M , ∇ ∂µ t = α ν α where ∇ t·ν = ∂ t·ν + Γ··ν t·δ − Γ··δ t·ν . (∇µ t·ν )∂ ⊗ dx ν µ µ µα δ α α α µδ α

7

where the bars as already said above refers to quantities defined with the LeviCivita connection. The last two terms in Eq.(20) cancel out because of Eq.(12) and then from Eq.(20) we have for the scalar curvature: ···µ ·νµ ¯ + 2∇ ¯ µ Kν·· R = g νλ Rµνλ· =R .

(21)

·νµ ¯ µ Kν·· ¯ µ Bν − ∇ ¯ ν Bµ = Doing the evaluation of ∇ and taking into account that ∇ ∂µ Bν − ∂ν Bµ we get

¯ + C (∂µ Bν − ∂ν Bµ ) F µν + 8πC Bµ J µ . R=R c

(22)

··λ Eq.(22) suggests to us to identify the Bµ in Kµν· with the components of µ the electromagnetic potential A = Aµ dx , i.e., we take from now on ··λ ·λ Kµν· := −CAµ Fν· ,

(23)

since this identification will permit us to interpret (a factor apart) the fist and second terms on the r.h.s. of Eq.(22) as the gravitational field and the free electromagnetic field (Jµ = 0) Lagrangian in GR. Indeed, for that case we are in position of interpreting those fields as parts of a Riemann-Cartan spacetime structure hM, g,∇, τg, ↑i by taking as Lagrangian of the system L :=

−c3 R 16πG

(24)

and the action is S=

−c3 16πG

Z

1

¯ + CFµν F µν )(− det g) 2 d4 x. (R

(25)

We can now give a geometrical model for the interaction of the electromagnetic field F , its current J and the gravitational field g by interpreting those fields as parts of a Riemann-Cartan spacetime structure hM, g,∇, τg, ↑i. Variation ( δ ) of S in Eq.(25) with respect to the contravariant components of g g

gives Z Z 1 1 −c3 ¯ 2 d4 x + C Fµν F µν (− det g) 2 d4 x δ [ R(−g) 16πG g Z −c3 ¯ − 8πCTµν )δg µν (− det g) 12 d4 x, ¯ µν − 1 gµν R (R = 16πG 2

δS = g

(26) (27)

where

1 1 (−Fµβ Fνβ + gµν F αβ Fαβ ), (28) 4π 4 are the components of the energy-momentum tensor of electromagnetic electromagnetic field, showing moreover that we must take C as Tµν =

C= 8

G . c4

(29)

Remark 1 This complete the proof of our claim in the introduction that any coupled system consisting of a gravitational field and electromagnetic field can be fully geometrized by a special Riemann-Cartan spacetime structure. Remark 2 We observe that if the contorsion tensor just introduced and whose information is contained in K is theVsame as the one contained in the Chern3 Simons object [43, 13] A ∧ dA ∈ sec T ∗ M . Indeed, C = A ∧ dA = A ∧ F 1 = Aµ Fνλ ϑµ ∧ ϑν ∧ ϑλ 2 1 = (Aµ Fνλ + Aλ Fµν + Aν Fλµ )ϑµ ∧ ϑν ∧ ϑλ 3!

(30)

It is eventually opportune to observe that A ∧ dA has been called in [35] the topological torsion, although in [48] it has been argued that this was not a good nomenclature since this object is proportional to the spin density of the electromagnetic field. This results seems to be endorsed by the nice analysis in [21]. R 1 Remark 3 Before proceeding we note that [45] δ Aµ J µ (− det g) 2 d4 x = 0. g

Indeed, since J is a time like 1-form field there must be (at least ) one coordinate system where J = J 0 := ρq ϑ0 and thus Aµ J µ = A0 ρq . Consequently we have that Z Z 1 µ 4 2 Aµ J (− det g) d x = Q A0 dx0 , (31) where Q is the total charge in space and then δ g

R

1

1

Aµ J µ ((− det g) 2 ) 2 d4 x = 0.

Based on the last remark, we proceed with the building of our theory by postulating a Lagrangian for the gravitational and electromagnetic fields and their sources which must include an electromagnetic current J and a material medium carrying that current (which as in GR cannot be geometrized) as 8πC −c3 (R + Aµ J µ ) + Lm . 16πG c Thus the total action for our theory is L :=

St = S + Sm Z −c3 ¯ + CFµν F µν + 16πC Aµ J µ )(− det g) 12 d4 x = (R 16πG c Z 1 1 Lm (− det g) 2 d4 x, + c

(32)

(33)

where G is the gravitational constant and we require as usual in field theories that the equations of motion of the theory are giving by

9

δ(S + Sm ) = 0.

(34)

Now, the energy-momentum of the material charge distribution is defined by [38] Z 1 δ Sm := Teµν δg µν (− det g)1/2 d4 x, g c ∂[(− det g)1/2 Lm ] 1 ∂ ∂[(− det g)1/2 Lm ] 1e , Tµν (− det g) 2 = − ∂ 2 ∂g µν ∂xλ g µν ∂xλ

(35)

and performing the δ variation of (S + Sm ) we get: g

¯ µν = R ¯ µν − 1 gµν R ¯ = 8πG (Tµν + Teµν ) , G 2 c4

(36)

¯ µν are the components of the Einstein tensor associated with the Leviwhere G Civita connection. Eq.(36) (Einstein equation in components form) gives the well know relation between the Einstein tensor and the energy-momentum tensor of the electromagnetic plus the energy-momentum tensor of matter on a Lorentzian spacetime [38]. Varying Eq.(32) with respect to and Aµ gives, as well known [38], Eq. (18), the non homogeneous Maxwell equations15 , and this completes the proof of our claim that with a special Riemann-Cartan connection it is possible to present the electromagnetic and gravitational fields as parts of the Riemann-Cartan structure hM, g, ∇, τg , ↑i with a Lagrangian giving by Eq.(32)

4

Lorentz Force Equation

For simplicity, we will consider in what follows, a continuous distribution of non-interacting incoherent charged matter, or “dust” as the material support of the electromagnetic current J. Let T˜ = ρ0 c2 V ⊗ V ∈ sec T02 M be the energy momentum of the charged “dust” where V = V µ ϑµ is the 1-form velocity field of the dust (g(V , V ) = 1) and ρ0 is its proper charged mass density. We calculate now the components of the covariant Riemann-Cartan covariant derivative of (T + T˜), i.e., ¯ µ (T µν + T˜ µν ) + K µ (T δν + T˜ δν ) + Kµδν (T µδ + T˜ µδ ). (37) ∇µ (T µν + T˜ µν ) = ∇ µδ Since from Eq.(36) it is: ¯ µ (T µν + T˜ µν ) = 0, ¯ µν = 8πG ∇ ¯ µG ∇ c4 15 The

homogeneous Maxwell equations follows trivially form F = dA.

10

(38)

we get recalling Eq.(23) that ··µ δν ··ν Kµδ· T + Kµδ· T µδ = 0.

(39)

··µ ˜ δν ··ν ˜ µδ ∇µ (T µν + T˜ µν ) = Kµδ· T + Kµδ· T .

(40)

Then, we can write

Now, recall that from Maxwell equations it follows trivially that ∇µ T µν =

1 µν F Jµ . c

(41)

Using this result in Eq.(40) we have G 1 ρ0 c2 V µ ∇µ V ν + V ν ∇µ (ρ0 c2 V µ ) = − F µν Jµ + 4 ρ0 c2 Aµ V δ (Fδ··µ V ν + Fδ··ν V µ ). c c (42) Due to Vµ V µ = 1 and the skew symmetry of Fµν , we get contracting Eq. (42) with Vν that G ·µ ν V . (43) ∇µ (ρ0 c2 V µ ) = 4 ρ0 c2 Aµ Fν· c With Eq.(43), we have from Eq.(42), G 1 ρ0 c2 V µ ∇µ V ν = − F µν Jµ + 4 ρ0 c2 Aµ V µ Fδ··ν V δ . c c

(44)

But for each integral line σ (parametrized by proper time s) with tangent vector field σ∗s at σ(s) of the flow defined by the velocity field V we can write (with g(σ∗s , ) = V |σ ) V µ ∇µ V ν =

dV ν ¯ ··ν V µ V δ + K ··ν V µ V δ , +Γ µδ· µδ· ds

(45)

and then, substituting Eq.(45) in Eq.(44) and using Eq.(23), we obtain ρq ·ν µ dV ν µ δ ¯ ··ν F V = 0 ,. +Γ µδ· V V − ds ρ0 c2 µ·

(46)

where have used that J µ = cρq V µ with the rest charge density given by ρq . Eq.(46) is then identified as the Lorentz force law on a Lorentzian spacetime [38]. Note also that evaluating explicitly the covariant derivative Eq.(43), with the aid of Eq.(15), we get ¯ µ (ρ0 c2 V µ ) = ∂µ [(−g) 12 ρ0 c2 V µ ] = 0 , ∇

(47)

i.e., in this model we have matter conservation. Remark 4 We observe here that if we model the matter as a Dirac-field living in the Riemann-Cartan background we may obtain like in [46] that the torsion tensor is also a source of the spin density. This will be discussed elsewhere. 11

5

The Gauge Invariance

Recall that in our model the Lagrangian (excluding the electromagnetic coupling of the electromagnetic potential with the electromagnetic current) for the gravitational plus the electromagnetic field is geometrized, i.e., it is given by the scalar curvature of the Riemann-Cartan connection according to Eq.(22) ¯ + G Fµν F µν = R − 8πG Aµ J µ . R c4 c4

(48)

Now, we investigate what happens if we make a gauge transformation A 7→ A+ dϕ in the definition of the contorsion. That transformation changes K 7→ K ¸ where the components of K ¸ are ··λ K ¸ ··µν·λ = Kµν· +

G ·λ (∂µ ϕ)Fν· . c4

(49)

Then, we get an new Riemann-Cartan curvature tensor whose components are

G ¯ ·χ ¯ ν (F ·χ ∂µ ϕ)] [∇µ (Fλ· ∂ν ϕ) − ∇ λ·. c4 From Eq.(50), the new scalar curvature R ¸ is ···χ R ¸ ···χ µνλ· = Rµνλ· +

R ¸ =R+

8πG (−g)

1 2

c5

1

∂µ [(−g) 2 ϕJ µ ].

(50)

(51)

Then since (R ¸ −R) differs by an exact differential, if we take R ¸ as the new Lagrangian for the electromagnetic plus gravitational fields we get the same equations of motion as before. We conclude that the freedom in choosing an electromagnetic gauge in our physical equations means the freedom, within a gauge, to choose the RiemannCartan curvature tensor R ¸ ···χ ¸ ··λ µν· ). So, there is a class of gauge µνλ· (through K equivalent Riemann-Cartan structure describing the same gravitational plus electromagnetic field.

6

Conclusions.

We have obtained a set of equations, namely Einstein equations, Maxwell equations and the Lorentz force equations describing a continuous distribution of charged matter interacting with the electromagnetic and gravitational fields from a geometric point of view. We note, however, that those equations have the same form as if they were written on a Lorentzian spacetime, although we have postulated as Lagrangian for the free electromagnetic plus gravitation field the scalar curvature of a (particular) Riemann-Cartan spacetime. Then, if there is an electromagnetic field generated by a current distribution on a Lorentzian spacetime modeling a gravitational field we can think of these two fields as geometrical properties of a particular Riemann-Cartan spacetime structure, the one 12

whose contortion tensor is given by Eq.(23), although this is not apparent in the usual physical equations. Also the contortion tensor of our theory encodes the same information as the one that is contained in the Chern-Simons term A∧dA, which is proportional to the spin density of the electromagnetic field. Finally we observe that despite the fact that the contortion tensor and the RiemannCartan curvature tensor of our theory is not gauge invariant, the resulting field equations obtained through the complete Lagrangian involving the coupled interacting system consisted of gravitational field plus the electromagnetic fields and the charge current produce equations for those fields and equations of motion of the charged matter that are gauge invariant, as they should be. So, for each gauge we have a different, but equivalent Riemann-Cartan spacetime structure defined by another gauge Acknowledgments Authors dedicate this paper to the memory of Professor Jaime Keller. They also would like to thank J. Vaz, Jr., M. A. Faria Rosa and R. da Rocha for useful discussions.

References [1] Adler, R., Bazin, M., and Schiffer, M. Introduction to General Relativity, chapter 13, McGraw-Hill, New York, 1965. [2] Andrade,V.C., and Pereira, J. G., Torsion and the Electromagnetic Field., Int. J. Mod. Phys. D 8, 141-151(1999) [3] Appelquist, T., Chodos, A., and Freund, P.G.O., Modern Kaluza-Klein Theories, Addison-Wesley, New York, 1987. [4] Babak, S. V., and Grishckuk, L. P., The Energy Momentum Tensor for the Gravitational Field., Phys. Rev. D 61, 024038.1-18 (1999) [arXiv: gr-qc/9907027 v2] [5] Barut, A. O., Elecrodynamics and Classical Theory of Fields and Particles, Dover, New York, 1980. [6] Beil, R. G. Electrodynamics from a Metric, Int. J. Theor. Phys. 26, 189-197 (1986). [7] Borchsenius, K., An extension of the Nonsymmetric Unified Field Theory, Gen. Rel. Grav. 7, 527- 534 (1976). [8] Capozzielo, S., Lambiase, G., Stornailo, C. Geometric Classification of Torsion Tensor of Space-Time, Ann. Phys. 10,713-727(2001) [arXiv: gr-qc 0101038 v1] [9] Cartan, E., On Manifolds with an Affine Connection and the Theory of General Relativity, Bibliopolis, Napoli, 1986. 13

[10] Chyba, C.F., Kaluza-Klein Unified Field Theory And Apparent FourDimensional Space-Time, Am. J. Phys. 53, 863-872 (1985) [11] Choquet-Bruhat, Y., DeWiit-Morette, C, and Dillard-Bleick, M., Analysis, Manifolds and Physics (revised edition), North-Holland, Amsterdam, 1982. [12] Coquereaux, R., and Jadczyk, A. Riemannian Geometry, Fiber Bundles, Kaluza-Klein Theories And All That, Lectures Notes in Physics 16, World Scientific, Singapore, 1988. [13] Cottingham, W. N., and Greenwood, D. A., An Introduction to the Standard Model of Particle Physics (second edition), Cambridge Univ. Press, Cambridge, 2007. [14] da Rocha, R. and Rodrigues, W. A. Jr. Pair and Impar, Even and Odd Differential Forms and Electromagnetism, Ann. Phys. (Berlin) 19, 6-34 (2010). [15] Eddington, A. S., The Mathematical Theory of Relativity, 3th. ed., Chelsea, New York, 1975. [16] Einstein, A., The Meaning of Relativity, appendix II, Princeton University Press, Princeton, 1956. [17] Fern´ andez, V. V. and Rodrigues, W. A. Jr., Gravitation as a Plastic Distortion of the Lorentz Vacuum, Fundamental Theories of Physics 168, Springer, Heidleberg, 2010. [18] Fontaine, M., and Amiot, P., Geometrical Aspects of Electromagnetism, I., Ann. Phys. 147, 269-280 (1983). [19] Fradkin, E. S., and Tseytlin, A. A., Effective Field Theory from Quantized Strings, Phys. Lett. B 158, 316-322 (1985). [20] de Andrade, L. C. G., and Lopes, M, Detecting Torsion from Massive Electrodynamics, G. Rel. Grav. 25, 1101-1106. (1993). [21] de Vries, H., On the electromagnetic Chern Simons spin density as hidden variable and EPR correlations. [http://physics-quest.org/ChernSimonsSpinDensity.pdf] [22] Garecki, J., Is Torsion Needed in Theory of Gravity ?, Rel. Grav. Cosmol. 1, 43-59 (2004) [arXiv:

gr-qc/0103029 v1]

[23] G¨ ockeler, M, and Sch¨ ucker, T., Difeerentil Geoemtry, Gauge, Gauge Theories, and Gravity, Cambridge University Press, Cambridge, 1987. [24] Hammond, R. T., Gravitation, Torsion, and Electromagnetism, Gen. Rel. Grav. 20, 813-827 (1988). 14

[25] Hammond, R. T., Spin, Torsion, Forces, Gen. Rel. Grav. 26, 247-263 (1994). [26] Hammond, R. T., New Fields in General Relativity, Cont. Phys. 36, 103114 (1995). [27] Hammond, R. T., Upper Limit on the Torsion Coupling Constant, Phys. Rev. D 52, 6918-6921 (1995). [28] Hehl, F. W., von der Heyde, P., and Kerlick, G., General Relativity with Spin and Torsion: Foundations and Prospects, Rev. Mod. Phys. 48, 393-416 (1976). [29] Hehl, F. W., Spin and Torsion in General Relativity: I. Foundations, Gen. Rel. Grav., 4, 333-349 (1973). [30] Hehl, F. W., and Obukhov, Y. N. How Does the Electromagnetic Field Couple to Gravity, in Particular to Metric, Nonmetricity, Torsion and Curvature?, Lecture Notes in Phys. 562, 479-504 (2001)[arXiv: gr-qe/0001010 v2] [31] Hehl, F. W., and Obukhov, Y. N., Foundations of Classical Electrodynamics, Charge, Flux and Metric, Birkh¨auser, Boston, (2003). [32] Jacubiec, A., and Kijowski, J., On Interaction of The Unified MaxwellEinstein Field with Spinorial Matter, Lett. Math. Phys. 9, 1-11 (1985). [33] Kaluza, Th., Zum Unit¨atsproblem in der Physik, Sitzungsber Preuss. Akad. Wiss. Berlin, 966-972 (1921). [34] Kibble, T., Lorentz Invariance and the Gravitational Field, J. Math. Phys. 2, 212-221 (1961). [35] Kiehn, R. M., Non-Equilibrium and Irreversible Thermodynamics from a Topological Perspective. Adventures in Applied Topology vol.1, 2008. [http://www.lulu.com/spotlight/kiehn] [36] Klein, O., Quantentheorie und F¨ unfdimensionale Relativit¨ atstheorie, Z. Phys. 37, 895-906 (1926). [37] Kleinert, H., Gauge Fields in Condensed Matter, World Scientific, Singapore, 1989. [38] Landau, L. D., and Lifshitz, E. M., The Classical Theory of Fields, Pergamon, Oxford, 1971. [39] Maluf, J. W., Hamiltonian Formulation of the Teleparallel Description of General Relativity, J. Math. Phys. 35, 335-343 (1994). [40] Moffat, J. W., Space-time Structure in a Generalization of Gravitation Theory, Phys. Rev. D 15, 3520-3529 (1977). 15

[41] Moffat, J. W., Reinventing Gravity, Smithsonian Books, Harper Coolins Publ, New York, 2008. [42] McKellar, R. J., Asymmetric Connection Treatment of the EinsteinMaxwell Field Equations, Phys. Rev. D 20, 356-361 (1979). [43] Nakahara, M, Geometry, Topology and Physics, IOP Publ. Ltd, Bristol and Philadephia, 1990 [44] Note-Cuello, E. A., da Rocha, R. and Rodrigues, W. A. Jr., Some Thoughts on Geometries and on the Nature of the Gravitational Field , J. Phys. Math. 2, 20-40 (2010) [45] Pauli, W., Theory of Relativity, Dover Publ. Inc., New York, 1981. [46] Poplawski N. J., Torsion as Electromagnetism and Spin, Int. J. Theor. Phys. 49, 1481-1488 (2010). [47] Ringermacher, H., An Electrodynamic Connection, Class. Quant. Grav. 11, 2383-2394 (1997). [48] Rodrigues, W. A. Jr. and Capelas de Oliveira, E., The Many Faces of Maxwell, Dirac and Einstein Equations, Lecture Notes in Physics 722, Springer, Heidleberg, 2007. [49] Rodrigues, W. A. Jr., Differential Forms on Riemannian (Lorentzian) and Riemann-Cartan Structures and Some Applications to Physics, Ann. Fond. L. de Broglie, 32 (special issue dedicated to torsion), 425-478 (2007).[arXiv:0712.3067v6 [math-ph]] [50] Rodrigues, W. A. Jr., On the Nature of the Gravitational Field and its Legitimate Energy-Momentum Tensor. [arXiv:1109.5272[math-ph]] [51] Weyl, H., Gravitation und Elektrizit¨ at, Sitzungsber. Preuss. Akad. Wiss 26, 465-478 (1918) [52] Weyl, H., Space,Time, Matter, 4th. ed., Dover Publications, New York, 1950. [53] Sauer, T., Field Equations in Teleparallel Spacetime: Einstein’s Fernparallelismus Approach Towards Unified Field Theory., Historia Math. 33, 399-439 (2006). [arXiv:physics/0405142v1] [54] Sachs, R. K. and Wu, H., General Relativity for Mathematicians, SpringerVerlag, New York, 1997. [55] Sciama, D., The Physical Structure of General Relativity, Rev. Mod. Phys 36, 463–469 (1964). [56] Schr¨odinger, E., The General Unitary Theory of the Physical Fields, Proc. R. Irish Acad. A 49, 43-58 (1943). 16

[57] Sivaram, C., and de Andrade, L. C. G., Torsion Gravity Effects on CargedParticle and Neutron Interferometers. [arXiv: gr-qc/0111009 v1] [58] Souza, Q. A. G., and Rodrigues, W. A. Jr, The Dirac Operator and the Structure of Riemann-Cartan-Weyl Spaces,. in Letelier, P., and Rodrigues, W. A. Jr. (eds.), Gravitation: The Spacetime Structure, pp. 179-212, World Scientific, Singapore, (1994). [59] Utiyama, D., Invariant Theoretical Interpretation of Interaction, Phys. Rev. 101, 1597-1607 (1956). [60] Vargas, J. G., Geometrization of the Physics with Teleparallelism I. The Classical Interactions., Found. Phys. 22, 507-526 (1992) [61] Vargas, J. G., Torr, D. G., and Lecompte, A.,Geometrization of the Physics with Teleparallelism. II. Towards a Fully Geometric Dirac Equation, Found. Phys. 22, 527-547 (1992).

17