Lorentz-invariant equations for weak gravitational fields 1

Chapter 1

L ORENTZ - INVARIANT EQUATIONS FOR WEAK

GRAVITATIONAL FIELDS

D. A. Pérez-Carlos1, A. Gutiérrez-Rodrı́guez1, *A. Espinoza-Garrido2,

M. A. Hernández-Ruı́z3 and E. Cruz-Albaro1
1
Unidad Académica de Fı́sica,
Universidad Autónoma de Zacatecas, México.
2
Unidad Académica de Ciencia y Tecnologı́a de la Luz y la Materia,
Universidad Autónoma de Zacatecas, México.
3
Unidad Académica de Ciencias Quı́micas,
Universidad Autónoma de Zacatecas, México.

February 3, 2024

Abstract
It is well known that the linearized Einstein field equations give us a good ap-
proximation when the gravitational fields are weak. In this sense, they are valid, for
example, to measure phenomena within our Solar System. In this work, we will get
gravitational equations invariant under Lorentz transformations. They will allow us
to obtain an energy-momentum tensor for the gravitational field, although this is a
first approximation, it serves to describe the conservation of energy and momentum of
matter and fields, that is, it explains the transfer of energy and momentum of any dis-
tribution of masses and gravitational field in weak gravitational fields. We have given
a physical interpretation of its components. We also will study the flux and the energy
density of the gravitational field.

1. Introduction
When we work with the Einstein Field Equations in the linear approximation, we obtain a
system of equations that can be written in a manner analogous to the Maxwell equations
of Electromagnetic Theory. This idea was proposed for the first time by Oliver Heaviside
in 1893 [1]. At that time there were only two kinds of interaction in physics, as we know:
* E-mail address: [email protected]
2 D. A. Pérez-Carlos, et al.

gravitational and electromagnetic interactions. When Heaviside made this assumption, it

was necessary to postulate the existence of a second gravitational field. This field must be a
consequence of moving masses and must act only on moving masses. This second gravita-
tional field is now called Heaviside field, and is denoted by the vector field k, the ordinary
gravitational field, known as Newtonian gravitational field is denoted by g. We define the
gravitodynamical field as the set of gravitational and Heaviside fields. The formulation
of this system of equations resulting from this analogy between the two theories is called
Gravitodynamical Theory. The gravitational field g is defined as a perturbation of space-
time1 due to the distribution of mass located at some region and capable of interacting with
a test mass mt according to
Fg = mt g, (1)

where Fg is the force exerted by a mass distribution acting on the test mass.
The Heaviside field is a perturbation of space-time due to a moving mass distribution
at some region which interacts on a test mass mt moving with velocity v, according to the
formula
Fk = mt (v × k), (2)

The work begun by Heaviside was continued by Jefimenko almost a century after in his
books [2, 3], improving Heaviside’s ideas, developing a gravitational theory that explains
a wide variety of physical phenomena that are not considered by the Newtonian theory
of gravitation, and explaining phenomena considered exclusive to the general theory of
relativity, such as the bending of light in the presence of massive bodies, the prediction of
gravitational waves and the precession of perihelion of mercury.
The gravitodynamical equations were obtained by Jefimenko supposing an analogy be-
tween his retarded solutions for the electromagnetic field and expressions for the fields g
and k. This means he postulated the gravitodynamical solutions and he obtained a system
of equations analog to the Maxwell equations. This system of equations is obtained by
Arbab [4] using quaternions, and also is obtained by Segura [5] from the linearized field
equations of the general theory of relativity. As we know, the linearized field equations
are valid only for weak fields, and therefore, it may be reasonable to think why to study a
linearized theory of gravitation if there are non-linear theories that explain gravitation re-
lated to super-massive bodies (See, for example, [6, 7]). We must note that Einstein Field
Equations lead us to the linear gravitodynamical theory in the first-order of approximation,
but we call the zero-order approximation the Newtonian theory, also, we have developed
a better approximation to second order in [8] which gives a better approximation to time-
dependent gravitational systems. Also, our motivation is that there is a methodological and
historical importance in the understanding of a linear theory of gravitation, this theory can
take us to a deeper understanding of gravitational phenomena. At last, we think there are
details and properties of gravitational fields that are not understood because they were not
allowed to appear in a linear theory. We must consider, by the way, that linear gravitational
equations are enough to describe phenomena within our solar system because they are valid
for weak gravitational fields.
1
In the linear case, and the regime of low velocities we can adopt only one time, which means that there are
no effects of retardation, for practical purposes.
 Lorentz-invariant equations for weak gravitational fields 3

Another advantage of working with a linear theory of gravitation is that it allows us

to obtain an energy-momentum tensor of the gravitational field, although the quantities
obtained are only a first approximation, this tensor describes the conservation of the four-
momentum, that is, the conservation of energy and momentum of fields and matter. This
tensor obtained explains the transfer of energy and momentum of any distribution of masses
and gravitodynamical field. It is important to mention the differences between this approach
of gravitation and the laws of electromagnetism. It is well-known there are two types of
electric charge and the interaction between them is repulsive and attractive, there is only
a kind of mass that is measured as positive, but it is known that this interaction is always
attractive, but from this theory the gravitational field g have another contribution due to the
movement of masses in the same radial direction but with a repulsive interaction, but with
a very little contribution. Whereas the magnetic field in electromagnetic theory is dextro-
rotatory, and the Heaviside field is levorotatory, it can be seen from the minus sign in the
analog to the Ampere law of electromagnetism. The mass current density J is considered
the ”flow” of the mass as an all and not the flow of particles inside a conductor as in the
case of its electromagnetic counterpart. In gravitodynamical theory, there are five contri-
butions to the gravitodynamical field, they are due to the mass density ρ, its variation on
time ∂ρ/∂t, the mass density J, and its variation on time ∂J/∂t. This can be seen from the
solutions postulated by Jefimenko to get the set of equations in the International System of
Units, namely,
Z     
[ρ] 1 ∂ρ γ 1 ∂J
Z
′
g(r, t) = −γ + 2 rdV + 2 dV ′ , (3)
r3 r c ∂t c r ∂t
Z   
γ [J] 1 ∂J
k(r, t) = − 2 + 2 × rdV ′ , (4)
c r3 r c ∂t
where γ is the Newtonian constant of gravitation, the quantities between brackets, are re-
tarded quantities, which means that they are evaluated in the retarded time t′ = t − r/c and
r is the distance between the source point and the field point, the point where the field is
measured and, c is the speed of the propagation of the gravitational interaction. There is
an indirect measurement for this speed in[9], but the direct measurement is given together
with the detection of gravitational waves and reported by Abbot et al [10]. In both cases,
this speed is equal to the speed of light.

2. Equations of gravitation invariant under Lorentz transfor-

mations
It is possible to write the gravitodynamical equations in a Lorentz-invariant form. To sim-
plify our calculations, we will start our derivation by writing the Jefimenko equations in the
new system of units introduced by us in previous works [12, 13], namely, the Gravitational
Gaussian System (GGS). These equations are

∇ · G = −4π̺, (5)

∇ · K = 0, (6)
4 D. A. Pérez-Carlos, et al.

Table 1. Gravitational rationalized quantities in GGS.

Gravitational field Heaviside field

Formula G = γ −1 g K = γ −1 ck
Units [G] = M L−2 = Jef [K] = M L−2 = Jef

1 ∂K
∇×G = − , (7)
c ∂t
4π 1 ∂G
∇×K=− J+ , (8)
c c ∂t
where ̺ is the mass density, J is the mass current density and c is the velocity of propagation
of the plane wave solutions of the gravitodynamical field.
In this system the gravitodynamical quantities G and K are defined by means of the
table 1, where g is the ordinary gravitational field and k is the Heaviside field in the Inter-
national System of Units. Heaviside field is the gravitational analog to the magnetic field
in the electromagnetic theory, called in General Relativity as gravitomagnetism or Lense-
Thirring effect, considered as a distortion of space-time due to the rotation of masses. and
tested its existence by means of the Gravity Probe B [11].
We can define the quantities ̺ and J by means of the Dirac delta function as follows
X
̺(r, t) = mn δ3 (r − rn (t)) (9)
n

and X
J(r, t) = mn vn (t)δ3 (r − rn (t)), (10)
n
where we are considering a system of particles of mass mn , whose positions in the instant
t are given by the position vectors rn .
If the Heaviside field K satisfies Eq. (5), then we can write it as the curl of some vector
quantity Γ
K = ∇ × Γ, (11)
where Γ is called Heaviside vector potential. If we substitute Eq. (11) in Eq. (7) we obtain
 
1 ∂Γ
∇× G+ = 0.
c ∂t
Because the quantity within the parentheses can be written as the gradient of the gravita-
tional scalar potential τ , we can express the gravitational field G
1 ∂Γ
G = −∇τ − . (12)
c ∂t
Substituting Eq. (12) in Eq. (5) we obtain
1 ∂
∆τ + (∇ · Γ) = 4π̺, (13)
c ∂t
 Lorentz-invariant equations for weak gravitational fields 5

where ∆ = ∇2 is the Laplacian operator.

And substituting Eqs. (11) and (12) in Eq. (8), and using the fact that ∇ × ∇ × V =
∇(∇ · V) − ∆V is valid for some arbitrary vector function V we have

1 ∂2Γ
 
1 ∂τ 4π
∆Γ − 2 2 − ∇ ∇ · Γ + = J. (14)
c ∂t c ∂t c
We can impose an additional condition over Γ and τ in a similar way to the Lorentz gauge
in the electromagnetic field, that is,
1 ∂τ
∇·Γ=− . (15)
c ∂t
Using the condition given by Eq. (15) in Eqs. (13) and (14), we obtain inhomogeneous
wave equations for the potentials τ and Γ, i. e.
1 ∂2τ
∆τ − = 4π̺ (16)
c2 ∂t2
and
1 ∂2Γ 4π
∆Γ − = J. (17)
c2 ∂t2 c
We define the four-dimensional mass current density as

J µ = (c̺, J), (18)

where µ = 0,1,2, 3,4. Then the continuity equation

∂̺
∇·J+ =0 (19)
∂t
becomes
∂µ J µ = 0. (20)
We also define the four-dimensional gravitodynamical potential as follows

Γµ = (τ, Γ). (21)

Then Eqs. (16) and (17) can be written in a covariant form

4π µ
Γµ = J , (22)
c
and condition (15) can be expressed as

∂µ Γµ = 0. (23)

Considering the fields G and K in terms of the potentials τ and Γ (Eqs. (11) and (12)), can
be demonstrated that the components of both fields can be put together in an antisymmetric
second rank gravitodynamical tensor H µν 2 .

H µν = ∂ µ Γν − ∂ ν Γµ . (24)
2
We define the gravitodynamial tensor by H µν in honour of Heaviside.
6 D. A. Pérez-Carlos, et al.

From Eq. (24) we can identify the components of this tensor as the gravitational and Heav-
iside fields
H 0ν = −Gν (25)
and
H µν = −ǫµνξ Kξ , (26)
where ǫµνξ is the three dimensional Levi-Civita symbol defined by

+1 if (µ, ν, ξ) is (1, 2, 3), (2, 3, 1) or (3, 1, 2)

µνξ
ǫ = −1 if (µ, ν, ξ) is (3, 2, 1), (1, 3, 2) or (2, 1, 3) (27)

0 if µ = ν, or µ = ξ or ν = ξ.


Expressing tensor (24) in components we have

 
0 −Gx −Gy −Gz
Gx 0 −Kz Ky 
H µν =
Gy Kz
 (28)
0 −Kx 
Gz −Ky Kx 0

or  
0 Gx Gy Gz
−Gx 0 −Kz Ky 
Hµν =
−Gy Kz
. (29)
0 −Kx 
−Gz −Ky Kx 0
Therefore, Eqs. (5) and (8) can be written as

4π ν
∂µ H µν = − J , (30)
c
and Eqs. (6) and (7) become

∂µ Hνξ + ∂ν Hξµ + ∂ξ Hµν = 0. (31)

We define the gravitodynamical dual tensor as follows

1
Hµν = ǫµναβ Hαβ , (32)
2

where ǫµναβ is the four-dimensional Levi-Civita symbol, defined as


+1 if (µ, ν, α, β) is an even permutation of (0, 1, 2, 3),

µναβ
ǫ = −1 if (µ, ν, α, β) is an odd permutation of (0, 1, 2, 3), (33)

0 otherwise.


We can use the symbol given by (33) to express Eq. (31) (or Eqs. (6) and (7)) as the next
equation
ǫµναβ Hαβ,ν = 0 (34)
 Lorentz-invariant equations for weak gravitational fields 7

Now, we will show the dual gravitodynamical field tensor in components

 
0 −Kx −Ky −Kz
Kx 0 Gz −Gy 
Hµν =  Ky −Gz
 (35)
0 Gx 
Kz Gy −Gx 0
or  
0 Kx Ky Kz
−Kx 0 Gz −Gy 
Hµν =  (36)
−Ky −Gz 0 Gx 
−Kz Gy −Gx 0.
The importance of the gravitodynamical tensor Hµν lies in the fact that it will be utilised
for the construction of the part of the action that corresponds to the gravitodynamical field.

3. Flux and energy density of gravitodynamical field

We can multiply with dot product Eq. (7) by K and Eq. (8) by −G and adding them, and
taking into account the vectorial identity ∇ · (A × B) = B · (∇ × A) − A · (∇ × B) for
any vectors A and B, we have
4π 1 ∂
∇ · (K × G) = − J·G+ (G2 + K 2 ) (37)
c 2c ∂t
or
∂w
+ J · G = −∇ · S, (38)
∂t
where
G2 + K 2
w=− (39)
8π
and
c
S= K × G, (40)
4π
where w is the gravitodynamical energy density, which measures the gravitational field
energy contained in a region of the gravitodynamical field when we integrate over the region
under consideration. S is the gravitational Poynting vector field, it represents the direction
and rate of energy flow per unit area at a point of space under consideration. Eq. (38)
is the gravitational Úmov-Poynting theorem in differential form, and expresses the energy
conservation for a system formed by masses and gravitodynamical field. This point will be
demonstrated later. If we integrate over a three-dimensional volume and use Gauss theorem
in the second member, we get
d
Z Z I
wd3 x + J · Gd3 x = − S · ds. (41)
dt V V s

Using the expression of the mass current (10), we can calculate the second integral in the
first member, that is
dE
Z X
J · Gd3 x = mn vn · G(rn ) = , (42)
V n
dt
8 D. A. Pérez-Carlos, et al.

where E is the sum of the kinetic energies of the particles inside the volume V . Here, we
supposed there are not particles in the boundary of the volume. Considering Eq. (42),
Eq.(41) becomes Z 
d
I
3
wd x + E = − S · ds. (43)
dt V s

This is the gravitational Úmov-Poynting theorem in integral form. The physical meaning
of this result is interpreted as follows, the temporal variation of the total energy of the
particles and the gravitodynamical field inside the volume, is equal, to the flux, taken with
the opposite sign, of the Poynting vector through the surface that encloses the volume. If we
integrate over all the three-dimensional space, then the surface integral in Eq. (43) vanishes,
because the gravitodynamical field is zero in the infinity. This result is given by
Z 
d 3
wd x + E = 0. (44)
dt V

This result expresses the conservation energy law in a closed system constituted by masses
and fields. We will obtain the same result expressed in Eq. (43) from the energy-momentum
tensor.

4. Principle of least action

We can obtain the equations of motion, not only for particles but also for fields, from ex-
tremal conditions of the functional called action.
In our case, as we see in [14], the principle of least action can be generalized for fields
noting that there is a similarity of continuous media and fields, so, we will consider as a
basic magnitude for the description of the fields, a Lagrangian density L defined by
Z
L = Ld3 x, (45)

then the principle of least action tells us that the action given by
Z  µ µ
 Z  
µ ∂Γ 3 1 µ ∂Γ
S = L Γ , µ d xdt = L Γ , µ d4 x, (46)
∂x c ∂x

where d4 x = cdtdxdydz, S must be an extremal, and Γµ are quantities describing the

state of the system and they correspond to the components of the four-potential (21) of the
gravitodynamical field.
µ µ
Abbreviating ∂Γ∂xµ = Γ,µ , we will vary S following the principle of least action to obtain
the field equations, that is
Z  
1 ∂L µ ∂L µ
δS = δΓ + δΓ d4 x
c ∂Γµ ∂Γµ,µ ,µ
Z    
1 ∂L µ ∂ ∂L µ ∂ ∂L µ
= δΓ + µ δΓ − µ µ δΓ,µ d4 x = 0. (47)
c ∂Γµ ∂x ∂Γµ,µ ,µ ∂x ∂Γ,µ
 Lorentz-invariant equations for weak gravitational fields 9

If we integrate over all space and we use Gauss’s theorem, the second term of the integral
is null, and we obtain
∂L ∂ ∂L
− = 0, (48)
∂Γµ ∂xµ ∂Γµ,µ
is the Euler-Lagrange variational equation. We note that if we add to Eq. (48) an integral
of divergence, this is converted into a surface integral, whose variation is zero, indeed

∂ µ µ
L → L′ = L + r (Γ ). (49)
∂xµ

5. Noether’s theorem
We will enunciate Noether’s theorem to find Noether’s current, quantity that we will use to
derive in section 7 the gravitodynamical energy-momentum tensor.
Theorem 1 (Noether’s Theorem). Any continuous transformation of the coordinates and
the fields that leave the action invariant in a four-dimensional volume corresponds to a
current conserved J µ in the evolution, which satisfies ∂µ j µ = 0.

Proof. We suppose that four-vectors and field functions transform as a result of some con-
tinuous transformation belonging to a Lie’s group3 :

x → x′ and u(x) → u′ (x′ ). (50)

Now, we will consider the action functional

 
∂u
Z
S[u(x)] = L u, , x d4 x, (51)
Ω ∂x

where Ω is an arbitrary integration region.

We consider the transformation

u(x) → u′ (x) = u(x) + δ̃u(x), (52)

where δ̃u(x) is the local variation of u(x).

Being as
∂u µ
u′ (x′ ) = u′ (x) + δx , (53)
∂xµ
the total variation is given by δu(x) = u′ (x′ ) − u(x), so that

∂u µ
δu(x) = δ̃u(x) + δx . (54)
∂xµ
Now, we will assume the action is invariant with respect to transformations (50), this means

S[u′ (x′ )] = S[u(x)] = inv. (55)

3
Informally, a Lie group is a group of symmetries where the symmetries are continuous.
10 D. A. Pérez-Carlos, et al.

Then δS = 0. We want to find the explicit form of the variation of the action S.
Z
δS = [L(δd4 x) + d4 xδL], (56)
Ω

∂x′
where δd4 x = d4 x′ − d4 x. Here d4 x′ = ∂x d4 x. The Jacobian is

∂x′ ∂x′µ ∂δxµ

   
= det = det δνµ + . (57)
∂x ∂xν ∂δxν
We can write the Jacobian in another way if we use the matrix identity

ln(det A) = tr(ln A)

on Eq. (57),
∂x′
  ′    ′ 
∂x ∂x
= exp ln det = exp tr ln
∂x ∂x ∂x
µ
  ′

∂δx ∂δx
≈ exp tr ≈1+ , (58)
∂x ∂xµ
hence
∂δxµ 4
δd4 x = d x. (59)
∂xµ
Now, we will calculate the variation of the Lagrangian density δL
 ′   
′ ′ ∂u ∂u ∂L
′
δL = L u (x ), ′ , x − L u(x), , x = δ̃L + µ δxµ , (60)
∂x ∂x ∂x

where the local variation of the Lagrangian density δ̃L is

   
∂L ∂L ∂L ∂ ∂ ∂L ˜
δ̃L = δ̃u + δ̃u,ν = − ν δ̃u + ν δu .
∂u ∂u,ν ∂u ∂x ∂x ∂u,ν
If the field satisfies the Lagrange equations
δS ∂L ∂ ∂L
= − ν = 0,
δu ∂u ∂x ∂u,ν
then  
∂ ∂L
δ̃L = δ̃u . (61)
∂xν ∂u,ν
We substitute Eqs. (59) and (61) in Eq. (56) in order to obtain the variation of the action
 
∂ ∂L
Z
µ ν
δS = µ
Lδx + (δu − u,ν δx ) d4 x. (62)
Ω ∂x ∂u,µ
If the volume Ω is arbitrary and δS = 0, then the divergence is null,
  
∂ µ ∂L ν ∂L
Lδ ν − u ,ν δx + δu = 0. (63)
∂xµ ∂u,µ ∂u,µ
 Lorentz-invariant equations for weak gravitational fields 11

The transformations form a Lie’s group with parameters κl , where l = 1, 2, ..., p; therefore

∂xν l ∂u l
δxν = δκ , and δu = δκ ,
∂κl ∂κl

where
∂xν ∂u
= (Xl )νµ xµ , and = X l u,
∂κl ∂κl
and Xl are the group generators. We define the Noether’s current by means of

δxν
 
∂L ∂L δu
J µl = −Lδµν + u,ν l
− . (64)
∂u,µ δκ ∂u,µ δκl

Hence, by virtue of the Eq. (63), due to the independence of the parameters one with respect
to another, we have
∂J µl
= 0. (65)
∂xµ
At the first order, Noether’s theorem can be expressed as
Z Z
δS = ∂µ J µl d4 x = J µl dsµ = 0, (66)
Ω s

where we have applied the Gauss theorem for fields satisfying the Euler-Lagrange equa-
tions. Also, we can write this result in a differential form

∂µ J µl = 0. (67)

The current J µl is not uniquely determined, we can always add any trivial conservative
current C µ such as ∂µ C µ = 0.
It is possible to define the energy-momentum tensor from Noether’s theorem, and to
do this, we will consider transformation of the Poincaré’s group. For translations xµ →
xµ + aµ , in the Noether’s current given by Eq. (64), we get

δxν δu
κl → aµ , µ
= δνµ ; = 0, (68)
δa δaµ

where the last equation in (68) is conditioned by the homogeneity of space, that is, invari-
ance of the fields with respect to translations: u′ (x′ ) = u(x).
Therefore, the Noether’s current (64) takes the form

∂L A
T µν = −Lδµν + u,ν . (69)
∂uA
,µ

This is the definition of the energy-momentum tensor, it applies to any field.

12 D. A. Pérez-Carlos, et al.

6. Action for the gravitodynamical field

Following the analogy between electrodynamic theory and gravitation and based on the
results obtained by Landau and Lifshitz in [15], the action for a system constituted by the
gravitodynamical field and particles located there, has the form of the next sum

S = Sp + Sf + Spf , (70)

where the subindex p denotes ”particles”, f denotes ”field” and pf denotes ”particle-field”.
Sp is the part of the action that depends only on the particles, this is the action for free
particles. It is well known that the action for a free particle is given by
Z
S1 = −mc ds,

and the Lagrangian is

r
2 v2
L = −mc 1− .
c2
Therefore, the action for a system formed by several particles is given by the sum
X Z
Sp = − mc ds. (71)

Sf is part of the action that corresponds to a field in the absence of masses, this means,
depends only on properties of the gravitodynamical field.
To get this part of the action we must consider the next argument taking as a starting
point the superposition principle. Because any solution of the field equations represents a
field existing in nature, the sum of any number of such fields must be a field existing in
nature and must satisfy the field equations.
The field equations must be linear homogeneous differential equations because they
satisfy the superposition principle. We obtain the field equations varying the action, and
such variation reduces the degree of the integrand in a unity, for this, the action must be a
quadratic expression of the components of such field and must be a scalar invariant. The
quantity that satisfies these conditions is the product Hµν H µν , therefore,Sf must be
Z
Sf = a Hµν H µν d4 x, (72)

where the integral extends over all space and all time between two given instants, a is a
1
constant that in the Gaussian system is a = 16π , therefore

1
Z
Sf = Hµν H µν d4 x, d4 x = cdtdxdydz. (73)
16πc

Now, we will obtain the part of the action that corresponds to the interaction between the
field and the particles, that is, Spf . This part of the action must contain quantities char-
acterizing the particles and the field. On one side, the mass m of a particle determines its
 Lorentz-invariant equations for weak gravitational fields 13

properties concerning the interaction with the field. On the other hand, the properties of the
field are characterized by the four-potential Γµ . In this way the action of a particle is
m
Z
Spf = − Γµ dxµ ,
c
where we introduce the factor 1c for convenience and the Γµ are taken in points of the line
of the universe of the particle. Therefore, for a system of particles, we have
XmZ
Spf = − Γµ dxµ . (74)
c
In each term of the sum, Γµ is the potential of field in that point of space-time in which is
the particle.
So, we write the sum given by (70) in an explicit form considering Eqs. (71), (73) and
(74), therefore
1 XmZ
X Z Z
S=− mc ds + Hµν H µν d4 x − Γµ dxµ , (75)
16πc c
is the total action corresponding to the system formed by particles and gravitodynamical
field.

7. Equation of motion of a particle in a gravitodynamical field

Taking the variation of Eq. (75) and following the results obtained in [16] but in our case
for the gravitodynamical field, to obtain the equation of motion of a particle inside a grav-
itodynamical field. We can despise the part of the action corresponding to the free gravito-
dynamical field Sf .
Thus, the action becomes
XZ  mn 
S=− mn cdsn + Γµ dxµa . (76)
n
c
This integral is taken along the universe line of the n-th particle, between two fixed universe
points.
The variation δS of Eq. (76) is
XZ  mn mn 
δS = − mn cδdsn + Γµ δdxµn + δΓµ dxµn . (77)
n
c c
p
Since ds = dxµ dxµ ,
dxµ δdxµ
δds = = uµ δdxµ ,
ds
where uµ is the four velocity.
Now, we will commute the operators variation δ and differentiation d, and integrating
by parts the first two terms of (77), we obtain
X
n mn  µ XZ
δS = − mn cuµ + Γµ δxn + (mn dunµ δxµn + mn dΓµ δxµn − mn δΓν dxνn ).
n
c n
(78)
14 D. A. Pérez-Carlos, et al.

Due to that δxµn = 0 in the region of integration, the first sum in Eq. (78) is null.
Therefore Eq. (78) transforms as follows
X Z  dunµ 
δS = mn − mn Hµν uµn δxµn dsn , (79)
n
ds

where we have used the next relations

duµ
duµ = ds,
ds
dΓµ = uν ∂ν Γµ ds,
and
δΓν = ∂µ Γν δxµ .
By virtue of arbitrariness of δxµn and due that δS = 0, by the principle of least action, we
have
duµ 1
= 2 Hµν uν , (80)
ds c
where we depreciate the index n. Eq. (80) are the four invariant relativistic equations of
motion of a particle.
We can write Eq. (80) as the next equations

dp
= m(G + v × K) (81)
dt
and
dε
= mv · G, (82)
dt
where we have used the relativistic momentum and energy given by

mv mc2
p= q and ε = q
v2 2
1− c2
1 − vc2

and r
v2
ds = 1− dt.
c2
Eq. (81) is the equation of motion and is the force acting on a particle with mass m inside
a gravitodynamical field. This equation corresponds to the sum of Eqs. (1) and (2) given in
the introduction of this paper and is the analog of the Lorentz force given in electromagnetic
theory. On the other hand, Eq. (82) gives us the temporal variation of the kinetic energy of
the particle, and it is determined by the work realized by the gravitational field.
Eqs. (81) and (82) are not independent of each other, in fact, using

dp dε
v· =
dt dt
and multiplyig scalarly Eq. (81) by v, we can see that Eq.(82) is consequence of Eq. (81).
 Lorentz-invariant equations for weak gravitational fields 15

8. The gravitodynamical energy-momentum tensor

In this section, we will obtain the gravitodynamical energy-momentum tensor. To begin we
will consider that there is no mass presence, in order to simplify our calculations.
When we are dealing with Hamilton’s principle for fields, we must be regarding as
a basic magnitude for its description a Lagrangian density, because there is a similarity
between continuous media and fields, and in this case, there are no discrete particles, hence,
we decompose the continuous media in infinitesimal volumes. Because the kinetic and
potential energies are functions of the mass, we take into account their volume densities.
Then, the Lagrangian density L is given by
Z
L = L(Γµ , Γµ,µ )d3 x, (83)

The Lagrangian density corresponding to the free gravitodynamical field to the action (73)
is given by
1
Lf = Hµν H µν . (84)
16π
In section 5. we defined the energy-momentum tensor as the Noether’s current given by Eq.
(69), whose conservation is conditioned by the invariance of the action of the system with
respect to transformations of the group of space-time translations.
Here, we will consider that magnitudes q correspond to the gravitodynamical four-
potential Γµ , so that
∂Lf
T νµ = Γα,µ − δνµ Lf . (85)
∂(Γα,ν )
To calculate the derivative present in Eq. (85), we will calculate the variation of the La-
grangian density δLf

1 µν 1 µν
δLf = H δHµν = H δΓν,µ , (86)
8π 4π
where we have used the covariant form of the gravitodynamical field tensor in terms of the
four-potential Γµ
Hµν = ∂µ Γν − ∂ν Γµ .
and the property of antisymmetry of Hµν . Hence, from Eq. (86) it is deduced

∂Lf 1 να
= H . (87)
∂(Γα,ν ) 4π

At last, using Eq. (87) we obtain the gravitodynamical energy-momentum canonical tensor
1 1 ν
T νµ = Γα,µ H να − δ Hαβ H αβ . (88)
4π 16π µ
Written in contravariant components, we have
1 1 νµ
T νµ = Γα,µ H να − η Hαβ H αβ . (89)
4π 16π
16 D. A. Pérez-Carlos, et al.

Until this expression is not gauge invariant, we can always obtain an expression gauge
invariant, if we add to Eq. (89) a divergence given by
1 µ ν ,α 1
− (Γ H α ) = − Γµ,α H να , (90)
4π 4π
where we have to consider that in the absence of masses, the field equation is H να
,α =
0. Therefore the energy-momentum tensor of the gravitodynamical field is expressed in a
symmetric form as follows
1 µα ν 1 µν
T µν = H Hα − η H αβ H αβ . (91)
4π 16π
We will analysing the component T 00 . If we write this component in an explicit form, we
have
1
T 00 = − (K 2 + G2 ). (92)
8π
If we integrate T 00 over a volume V we obtain the energy of the gravitodynamical field E
and is the field energy contained in the gravitodynamical field
Z
T 00 d3 x = E. (93)
V

The components T 0i , where i = 1, 2, 3 match with the components S i of the energy flux
density given by Eq. (40) divided by c , that is

Si 1
T 0i = = ǫijk K j Gk , (94)
c 4π
where ǫijk is the Levi-Civita symbol. Also, T i0 = T 0i and S i is the Poynting vector.
We will define the Jefimenko stress tensor as the three dimensional tensor ̟jk = −T jk
whose components are given by
 
1 1 2 2
̟jk = Gj Gk + Kj Kk − δjk (G + K ) . (95)
4π 2
Once obtained the components of the gravitodynamical energy-momentum tensor, we can
write it in components as follows
 
Sx Sy Sx
w c c c
 Sx
 c ̟xx ̟xy ̟xz 

µν
T =  Sy . (96)
 c ̟yx ̟yy ̟yz 
Sz
c ̟zx ̟zy ̟zz

Now, we are going to give a physical interpretation of the components of the stress tensor
(95). For this task, we take the four-divergence of the gravitodynamical energy-momentum
tensor (91), we have
 
1 1
T µν
,ν = H ν µα
H
α ,ν + H µα ν
H α,ν − H αβ H αβ,µ
. (97)
4π 2
 Lorentz-invariant equations for weak gravitational fields 17

This tensor can be reduced to

µν
T ,ν = −H µν Jν , (98)
making use of Jefimenko equations (30) and (34).

We will demonstrate that Eq. (98) is the conservation law for the total four-momentum
of the system. Substituting the current given by Eq. (10) in Eq. (98) we obtain

dxν 3
T µ0 µi
X
,0 + T ,i = − mn H µν δ (r − rn (t)). (99)
n
dt

Making use of the equation of motion of a punctual mass (80) and integrating Eq. (99) in a
volume V , we have !
d X I
µ
P + pn = − T µk dsk .
µ
(100)
dt n s

P µ is the total four-momentum of the gravitodynamical field in the considered volume. The
sum is taken only with respect to the particles inside the volume.
For µ = 0 Eq. (100) coincides with the Umov-Poynting theorem given by Eq. (43).
For µ = 1, 2, 3 we obtain the law of variation of the total three-dimensional momentum of
the system in the given volume, that is
!
d X I
µ
P + pn = − T ik dsk .
i
(101)
dt n s

The quantity of momentum outgoing from the volume per unit time is given by the second
member of Eq. (101). Consequently, the magnitudes T ik are the components of the three-
dimensional momentum flux density tensor.
i
If the gravitodynamical field is time independent, then dP
dt = 0, and considering ̟ik =
−T ik , Eq. (101) becomes
d X i
I
p = ̟ik dsk . (102)
dt n n s

We can use Eq. (81) to transform the first member of Eq. (102), therefore

d X X Z
pn = mn [G(rn ) + v × K(rn )] = (̺G + J × K)d3 x. (103)
dt n n

We conclude that in a static field the stress tensor flux through a surface limiting the volume,
gives us the total force acting over the particles inside the volume.

9. Application of the results obtained

We have a spherical mass rotating with angular velocity ω, and it has a translational velocity
equal to v directed along the y-axis, its mass density is considered uniform for practical
purposes (see fig. 1). The field is acting on a test mass mt located in the x-axis at a distance
r greater than a the radius of the spherical mass density ρ. We present the solutions (3) and
18 D. A. Pérez-Carlos, et al.

(4) postulated by Jefimenko in the Gaussian System of Units using the table 1, considering
the regime of low velocities v << c, we can ignore retardation brackets symbols, the
gravitodynamical field in the point P at r >> a4 , where the mass mt is located at instant
t = 0, and the sphere pass through the origin, we have of solutions (3) and (4), that is,
Z  
ρ 1 ∂ρ 1 1 ∂(ρv) 3 ′
Z
3 ′
G=− 3
+ 2
rd x + 2
d x, (104)
r cr ∂t c r ∂t
Z  
1 ρv 1 ∂(ρv)
K=− 3
+ 2 × rd3 x′ . (105)
c r cr ∂t
Note that the first integral in Eq. (104) is directed in the radial direction between the centers
of the masses interacting, but the second one is directed in the direction of the mass current
density J = ρv.

Figure 1. The sphere of radius a is rotating with an angular velocity ω, and it is moving in
the direction of positive y-axis with a velocity v, the field point P , the point where the field
is measured is located at r << a.

The second integral of Eq. (104) has special properties analysed by Chubykalo et al
[17]. This integral is known as gravikinetic field, denoted by Gk and can be expressed as
4
At r >> a the sphere can be considered as a punctual mass.
 Lorentz-invariant equations for weak gravitational fields 19

follows,
1 ∇′ (v · J) 3 ′
Z
Gk = − d x − v × K. (106)
c r
Using vector identities
∇′ (X) (X) (X)
=∇ + ∇′ (107)
r r r
and Z I
∇φd3 x = φds, (108)

we have
I v · dl
I
Gk = − ∇ − v × K, (109)
c r
where I is the mass current. This integral is called apparent gravitational dipole, and is
denoted by Gd , then
Gk = −Gd − v × K. (110)
Considering that the dimensions of the apparent gravitational dipole are smaller than r , Gd
can be written as follows  
I r
Gd = ∇ ·v×s , (111)
c r3
where s is the surface area of the mass current loop forming the gravitational dipole. We
have used the identity I Z
φdl = ds × ∇φ. (112)

A gravitational dipole of mass m moving with speed v << c appears to have the additional
mass (See Appendix 4 in [2]),
mv 2
mapp = . (113)
2c2
The field G from (104) is then

m mv 2 1 1 ∂J 3 ′
Z
G = − 3 r − 2 3 r + Gd + d x. (114)
r 2c r c r ∂t

Considering Eq. (110) and Gk as the last integral of Eq- (104), we can see that

Gd + Gf = −v × K, (115)

where Gf denotes the first integral of Eq. (104). The Heaviside field is found in the
integration of the external field of a spherical shell rotating, this field is

mωa2
K= ẑ. (116)
5cr 3
This is the Heaviside field at the point field P in the instant of time when the sphere passes
through the origin. Thus, we have the total force experimented by the test mass mt , from
Eq. (81),
20 D. A. Pérez-Carlos, et al.

m′ m m′ mv 2 m′ mvωa2
F=− x̂ − x̂ − x̂. (117)
r2 2c2 r 2 5c2 r 3
The force exerted by the mass m on the test mass mt is different from the result obtained
by Newton’s law of gravitation. We can appreciate this expression differs in the last two
terms dependent on the velocity of translation v and the angular frequency ω of our sphere.
We identify the field G as
m mv 2
G = − 2 x̂ − 2 2 x̂, (118)
r 2c r
and K is given by (116).
In this way, we use Eq. (40) to calculate the gravitational Poynting vector field, which
measures the total flux of gravitodynamic energy,

m2 ωa2 v2
 
S=− 1 − 2 ŷ. (119)
20πr 5 2c

The minus sign in (119) appears as a consequence of the gravitational Umov-Poynting

theorem in its integral form Eq. (43). Using Gauss theorem
I Z
S · ds = ∇ · S d3 x′ = 0, , (120)
s V

where d3 x = r 2 sin θdrdθdφ if we integrate over all the space, this integral vanishes, which
means that the gravitodynamic field vanishes at infinity, according to Eq. (44). The energy
density is " #
2
1 m2 v2 ω 2 a4
w=− 1− 2 + . (121)
8π r 2 2c 25c2 r 4
v 4
We can expand the binomial and disregard the third term 4c 4 because it is a very small
quantity. The last term can be also neglected, due to the same reasons, and can be considered
when the angular velocity is very fast, and so we are left with

1 m2 v2
 
w=− 1− 2 . (122)
8π r 2 c

The energy density of the gravitodynamical field is negative in the sense that no energy
can be extracted from the gravitodynamical field by destroying the field. Energy must be
delivered to the field to destroy it.
We show in Fig. 2 how the energy density behaves for the next values corresponding to our
Solar System

• r = 1.5 × 1013 cm

• a = 1.39 × 1011 cm

• ω = 2.6 × 10−6 s−1

• c = 3 × 1010 cm/s
 Lorentz-invariant equations for weak gravitational fields 21

• m = 2 × 1033 g

• v = 6.175 × 107 cm/s

This data can be consulted by going to the next url; https://www.oarval.org/SSsp.htm. The
units were changed to the GGS of units.

Figure 2. We show here how the energy density behaves for the values corresponding to the
Sun and the average distance between the Sun and the Earth.

References
[1] O. Heaviside, The electrician 31, 5125-5134 (1893).

[2] O. Jefimenko, Causality, electromagnetic induction and gravitation: A different ap-

proach to the theory of electromagnetic and gravitational fields, 2nd ed. (Princeton
University Press, Princeton, NJ, 2000).

[3] O. Jefimenko, Gravitation and Cogravitation: Developing Newton’s Theory of Grav-

itation to its Physical and Mathematical Conclusion, 1st ed. (Electret Scientific Star
City, Waynesburg, PA, 2006).

[4] A. I. Arbab, Astrophys Space Sci 330, 61–68 (2010).

[5] W. S. González, Gravitoelectromagnetismo y principio de Mach (eWT Ediciones,

Cádiz, 2013).
22 D. A. Pérez-Carlos, et al.

[6] A. Einstein, The collected papers of Albert Einstein Vol. 6 The foundation of the gen-
eral theory of relativity, (Princeton University Press, Princeton, NJ, 1997).

[7] A. A. Logunov, M. A. Mestvirishvili, Found. Phys. 16 1 (1986).

[8] D. A. Pérez Carlos, A. Espinoza Garrido, A. E. Chubykalo. The European Physical

Journal C, 80 1-7 (2007).

[9] E. B. Fomalont and S. M. Kopeikin, Astrophys. J. 598 (1), 704-711 (2003).

[10] Abbot et al, Phys. Rev. Lett., 116, 241103, (2016).

[11] B. Vető 2010 Eur. J. Phys. 31 1123 (2010).

[12] D. Pérez-Carlos, A. Espinoza, A. Chubykalo, Open Access Journal of Physics 1(2),

1-12 (2017).

[13] D. A. Pérez Carlos,,A. Espinoza Garrido, A. B. Gutiérrez Rodrı́guez, and Z. Oziewicz.

Weak Gravitational Fields in the Geometric Algebra Approach. Scientific Legacy of
Professor Zbigniew Oziewicz: Selected Papers from the International Conference
“Applied Category Theory Graph-Operad-Logic” (2023).

[14] W. S. González, Teorı́a de campo relativista (eWT Ediciones, Cádiz,2014).

[15] L. D. Landau and E. M. Lifshitz, Classical field theory (Pergamon press LTD, 1971).

[16] D.V. Galtsov, Iu. V. Grats and V. Ch. Zhukovski, Campos clásicos: Enfoque moderno
(Editorial URSS,2005).

[17] A. Chubykalo, A Espinoza, D. Pérez-Carlos, J. Mod. Phys, 9 (8), 1522-1544, (2018).