Introduction
In his seminal paper entitled “On the electrodynamics of moving bodies” [1 ] and published in the “Annalen der Physik” in 1905 during the “annus mirabilis”, Albert Einstein laid the foundation of special relativity by rooting firmly his theory on the transformation laws for the electromagnetic fields derived independently in 1899 by Hendrik A. Lorentz [2 ] and in 1900 by Henri Poincaré. [3 ] Shortly thereafter, an important effort was made to establish relativistic theories for all branches of physics, notably by Albert Einstein himself [4 ] in 1907. Max Planck [5 ] derived in 1908 the first relativistic transformation law for the temperature,
T = γ − 1 T ~ (Einstein and Planck) (1) T = \gamma^{-1} \widetilde{T} \qquad\text{(Einstein and Planck)} \tag*{(1)} T = γ − 1 T (Einstein and Planck) (1)
where T ~ \widetilde{T} T is the temperature in the rest frame R ~ \widetilde{\mathcal{R}} R and T T T is the temperature in an inertial frame R \mathcal{R} R and γ > 0 \gamma> 0 γ > 0 is the Lorentz factor. This transformation law was at the heart of the reigning paradigm of relativistic thermodynamics for over 50 years adopted notably by Louis de Broglie [6 ] , Olivier C. de Beauregard [7 ] , Max von Laue [8 ] , Richard C. Tolman [9 ] and Wolfgang Pauli [10 ] . Heinrich Ott called into question this paradigm. [11 ] Taking into account the relativistic transformation law for the mass, i.e. M = γ M ~ M = \gamma\widetilde{M} M = γ M , where M ~ \widetilde{M} M is the mass in the rest frame R ~ \widetilde{\mathcal{R}} R and M M M is the mass in an inertial frame R \mathcal{R} R , Ott showed in 1963 that the relativistic transformation law for the temperature becomes, [11 ]
T = γ T ~ (Ott and Arzelies) (2) T = \gamma\widetilde{T} \qquad\text{(Ott and Arzelies)} \tag*{(2)} T = γ T (Ott and Arzelies) (2)
and Henri Arzelies reached independently the same conclusion in 1965. [12 ] The difference between the relativistic transformation laws (1) and (2) for the temperature is referred to unfortunately as the Einstein and Ott controversy. A private letter sent by Einstein to von Laue in 1952 shows that Einstein’s point of view clearly changed over time. [13 ] The transformation law for the temperature derived in this letter is identical to the transformation law (2) derived later by Ott. So the controversy should be renamed adequately the Planck and Einstein controversy. A third relativistic transformation law of the temperature was postulated by Peter T. Landsberg [14 ] and Nicolaas G. van Kampen [15 ] who argued independently in 1967 and 1968 that the temperature should be a Lorentz invariant quantity,
T = T ~ (Landsberg and van Kampen) (3) T = \widetilde{T} \qquad\text{(Landsberg and van Kampen)} \tag*{(3)} T = T (Landsberg and van Kampen) (3)
All these authors agree on the frame independence of the entropy that should be a Lorentz invariant quantity as advocated initially by Albert Einstein [4 ] and Max Planck, [5 ]
S = S ~ (Einstein, Planck et al.) (4) S = \widetilde{S} \qquad\text{(Einstein, Planck et al.)} \tag*{(4)} S = S (Einstein, Planck et al.) (4)
where S ~ \widetilde{S} S is the entropy in the rest frame R ~ \widetilde{\mathcal{R}} R and S S S is the entropy in an inertial frame R \mathcal{R} R . Since the entropy is Lorentz invariant and the temperature is the intensive conjugate variable of the entropy, the relativistic transformation law for the internal energy has to be the same as the relativistic transformation law for the temperature. The relativistic transformation law for the pressure is another essential result for a relativistic theory of thermodynamics. Albert Einstein [4 ] , Max Planck [5 ] and Arzelies [12 ] reached the conclusion that the pressure is Lorentz invariant,
p = p ~ (Einstein, Planck and Arzelies) (5) p = \widetilde{p} \qquad\text{(Einstein, Planck and Arzelies)} \tag*{(5)} p = p (Einstein, Planck and Arzelies) (5)
where p ~ \widetilde{p} p is the pressure in the rest frame R ~ \widetilde{\mathcal{R}} R and p p p is the pressure in an inertial frame R \mathcal{R} R . These authors used the mechanical definition of the pressure as the ratio of the normal force and the surface area. [16 ] Sutcliffe derived another transformation law using the thermodynamic definition of the pressure as the intensive conjugate variable of the volume in 1965, [17 ]
p = γ 2 p ~ (Sutcliffe) (6) p = \gamma^{2}\widetilde{p} \qquad\text{(Sutcliffe)} \tag*{(6)} p = γ 2 p (Sutcliffe) (6)
Our aim is to establish a rigorous theoretical framework for the relativistic thermodynamics of a perfect fluid based on the Lorentz invariance between an inertial frame R \mathcal{R} R and a local rest frame R ~ \widetilde{\mathcal{R}} R . Our main objective is to settle once for all the debate on the relativistic transformations laws of thermodynamic quantities by using a consistent and systematic approach. Another important purpose is to determine the structure of the relativistic continuity equations for extensive thermodynamic scalar and vector quantities. Such an approach is expected to lead to a clear identification of the relativistic corrections to the dynamics of a perfect fluid. Finally, our synthetic approach allows us to obtain elegant axiomatic statements of the relativistic first and second law of thermodynamics.
This publication is structured as follows : the structure of the relativistic continuity equations for scalar and vector quantities is established in § 2 by following the approach outlined by Stückelberg. [18 ] The relativistic statements of the second and first law are presented in § 3 and 4 together with the relativistic continuity equations for the entropy and energy-momentum. In § 5 , we establish the relativistic continuity equation for the orbital angular momentum, which implies the symmetry of the stress-energy-momentum tensor. The relativistic matter continuity equation is derived in § 6 . As a prelude to the study of the dynamics and thermodynamics of a perfect fluid, we first examine the dynamics of a relativistic dust in § 7 and derive the corresponding stress-energy-momentum tensor in § 8 . Taking into account the pressure and the internal energy density, we generalise our analysis to the study of the dynamics of a perfect fluid in § 9 and derive the corresponding stress-energy-momentum tensor in § 10 . This leads to the introduction of a relativistic inertia density that differs from the mass density. Using the thermodynamic definition of the temperature, the pressure and the chemical potential, we establish rigorously the relativistic transformation laws for these three intensive quantities in § 11 , 12 and 13 respectively. The relativistic force density is established in § 14 . The relativistic continuity equations for the energy and the momentum are derived from the relativistic continuity equation for the energy-momentum in § 15 and 16, and the relativistic corrections are highlighted. In § 17 , we derive a relativistic continuity equation for the inertia and show how and why it differs from the continuity equation for the mass that is usually assumed to hold. We derive the relativistic Euler equation from the relativistic continuity equations for the momentum and the inertia in § 18 and highlight the relativistic corrections. Finally, we conclude our analysis of the relativistic thermodynamics of a perfect fluid in § 19 . In Appendices Appendix A and Appendix B, we present the fundamental concepts of relativistic kinematics and dynamics of a particle, which are the cornerstone of a relativistic theory.
Relativistic continuity equations
We establish in this section the structure of the relativistic continuity equations for a frame-independent scalar quantity F F F and for the contravariant component F ν F^\nu F ν of a vector in Minkowski space. We follow the covariant approach based on the divergence theorem in Minkowski space outlined by Stückelberg. [18 ] In order to do so, we consider a thermodynamic system within a fixed enclosure. An arbitrary extensive physical property is described by an extensive scalar function F ( t ) F(t) F ( t ) of time t t t in an inertial frame R \mathcal{R} R . By extensivity, the scalar function F ( t ) F(t) F ( t ) is obtained by integrating the density f ( t , x ) f(t,\boldsymbol{x}) f ( t , x ) over the spatial volume V 3 ( t ) V_3(t) V 3 ( t ) of the system, [19 ]
F ( t ) = ∫ V 3 ( t ) d V 3 ( x ) f ( t , x ) (7) F(t)=\int_{V_3(t)} dV_3(\boldsymbol{x})\,f(t,\boldsymbol{x}) \tag*{(7)} F ( t ) = ∫ V 3 ( t ) d V 3 ( x ) f ( t , x ) (7)
where x \boldsymbol{x} x is the position spatial vector, t t t is the time, d V 3 ( x ) dV_3(\boldsymbol{x}) d V 3 ( x ) is the spatial infinitesimal volume in the inertial frame R \mathcal{R} R . To describe the time evolution of the extensive scalar function F ( t ) F(t) F ( t ) in a relativistic manner, we define a covariant vector representing an infinitesimal spatial volume that is written in components in the inertial frame R \mathcal{R} R as,
d V 3 μ ( x ) = d V 3 ( x ) δ 0 μ (8) dV_{3\mu}(\boldsymbol{x})=dV_3(\boldsymbol{x})\,\delta_{0\mu} \tag*{(8)} d V 3 μ ( x ) = d V 3 ( x ) δ 0 μ (8)
The covariant time component of the infinitesimal spatial volume vector (8) in the inertial frame R \mathcal{R} R reads,
d V 30 ( x ) = d V 3 ( x ) (9) dV_{30}(\boldsymbol{x})=dV_3(\boldsymbol{x}) \tag*{(9)} d V 30 ( x ) = d V 3 ( x ) (9)
We introduce a current density vector of the physical property F ( t ) F(t) F ( t ) of contravariant components j f μ ( t , x ) j_f^\mu(t,\boldsymbol{x}) j f μ ( t , x ) in the inertial frame R \mathcal{R} R . In the local rest frame R ~ \widetilde{\mathcal{R}} R , the contravariant component of the current density vector j ~ f μ ( t ~ , x ~ ) \widetilde{j}_f^\mu(\widetilde{t},\widetilde{\boldsymbol{x}}) j f μ ( t , x ) is defined as a purely temporal vector,
j ~ f μ ( t ~ , x ~ ) = c f ~ ( t ~ , x ~ ) δ μ 0 (10) \widetilde{j}_f^\mu(\widetilde{t},\widetilde{\boldsymbol{x}})=c\,\widetilde{f}(\widetilde{t},\widetilde{\boldsymbol{x}})\,\delta^{\mu0} \tag*{(10)} j f μ ( t , x ) = c f ( t , x ) δ μ 0 (10)
where f ~ ( t ~ , x ~ ) \widetilde{f}(\widetilde{t},\widetilde{\boldsymbol{x}}) f ( t , x ) is the density of the frame independent physical property F ~ ( t ~ ) \widetilde{F}(\widetilde{t}) F ( t ) in the local rest frame R ~ \widetilde{\mathcal{R}} R ,
F ( t ) = F ~ ( t ~ ) (11) F(t)=\widetilde{F}(\widetilde{t}) \tag*{(11)} F ( t ) = F ( t ) (11)
Thus, the contravariant time component of the current density vector j ~ f 0 ( t , x ) \tilde{j}^{0}_{f}(t,\mathbf{x}) j ~ f 0 ( t , x ) in the local rest frame R ~ \widetilde{\mathcal{R}} R is written as,
j ~ f 0 ( t ~ , x ~ ) = c f ~ ( t ~ , x ~ ) (12) \tilde{j}^{0}_{f}(\tilde{t},\tilde{\mathbf{x}})=c\,\tilde{f}(\tilde{t},\tilde{\mathbf{x}}) \tag*{(12)} j ~ f 0 ( t ~ , x ~ ) = c f ~ ( t ~ , x ~ ) (12)
In view of the contravariant component of the current density vector (10), the contravariant components of the current density vector j f μ ( { x ρ } ) j^{\mu}_{f}(\{x^{\rho}\}) j f μ ({ x ρ }) in an inertial frame R \mathcal{R} R are related to the contravariant components of the current density vector j ~ f ν ( { x ~ ρ } ) \tilde{j}^{\nu}_{f}(\{\tilde{x}^{\rho}\}) j ~ f ν ({ x ~ ρ }) in the local rest frame R ~ \widetilde{\mathcal{R}} R by an inverse Lorentz transformation,
j f μ ( { x ρ } ) = ( Λ − 1 ) μ ν j ~ f ν ( { x ~ ρ } ) = ( Λ − 1 ) μ 0 c f ~ ( { x ~ ρ } ) (13) j^{\mu}_{f}(\{x^{\rho}\})=(\Lambda^{-1})^{\mu}{}_{\nu}\,\tilde{j}^{\nu}_{f}(\{\tilde{x}^{\rho}\})=(\Lambda^{-1})^{\mu}{}_{0}\,c\,\tilde{f}(\{\tilde{x}^{\rho}\}) \tag*{(13)} j f μ ({ x ρ }) = ( Λ − 1 ) μ ν j ~ f ν ({ x ~ ρ }) = ( Λ − 1 ) μ 0 c f ~ ({ x ~ ρ }) (13)
In view of the component of the Lorentz transformation (A.40) and the contravariant component of the current density vector (13), the contravariant time component of the current density vector is given by,
j f 0 ( { x ρ } ) = ( Λ − 1 ) 0 0 c f ~ ( { x ~ ρ } ) = γ c f ~ ( { x ~ ρ } ) (14) j^{0}_{f}(\{x^{\rho}\})=(\Lambda^{-1})^{0}{}_{0}\,c\,\tilde{f}(\{\tilde{x}^{\rho}\})=\gamma c\,\tilde{f}(\{\tilde{x}^{\rho}\}) \tag*{(14)} j f 0 ({ x ρ }) = ( Λ − 1 ) 0 0 c f ~ ({ x ~ ρ }) = γ c f ~ ({ x ~ ρ }) (14)
and the contravariant spatial components of the current density vector are given by,
j f j ( { x ρ } ) = ( Λ − 1 ) j 0 c f ~ ( { x ~ ρ } ) = γ v i f ~ ( { x ~ ρ } ) δ i j (15) j^{j}_{f}(\{x^{\rho}\})=(\Lambda^{-1})^{j}{}_{0}\,c\,\tilde{f}(\{\tilde{x}^{\rho}\})=\gamma v^{i}\,\tilde{f}(\{\tilde{x}^{\rho}\})\,\delta^{j}_{i} \tag*{(15)} j f j ({ x ρ }) = ( Λ − 1 ) j 0 c f ~ ({ x ~ ρ }) = γ v i f ~ ({ x ~ ρ }) δ i j (15)
In view of the contravariant temporal and spatial components of the velocity vector (A.21) and (A.24), and the contravariant temporal and spatial components of the current density vector (14) and (15), the contravariant components of the current density in the inertial frame R \mathcal{R} R are given by,
j f μ ( { x ρ } ) = f ~ ( { x ~ ρ } ) u μ ( { x ρ } ) (16) j^{\mu}_{f}(\{x^{\rho}\})=\tilde{f}(\{\tilde{x}^{\rho}\})\,u^{\mu}(\{x^{\rho}\}) \tag*{(16)} j f μ ({ x ρ }) = f ~ ({ x ~ ρ }) u μ ({ x ρ }) (16)
In view of the contravariant time component of the current density vector (12) in the local rest frame R ~ \widetilde{\mathcal{R}} R , the contravariant time component of the current density vector in the inertial frame R \mathcal{R} R is written as,
j f 0 ( { x ρ } ) = c f ( { x ρ } ) (17) j^{0}_{f}(\{x^{\rho}\})=c\,f(\{x^{\rho}\}) \tag*{(17)} j f 0 ({ x ρ }) = c f ({ x ρ }) (17)
Identifying the contravariant time components of the current density vector (14) and (17), we deduce that the density f ( t , x ) f(t,\mathbf{x}) f ( t , x ) in the inertial frame R \mathcal{R} R is related to the density f ~ ( t ~ , x ~ ) \tilde{f}(\tilde{t},\tilde{\mathbf{x}}) f ~ ( t ~ , x ~ ) in the local rest frame R ~ \tilde{\mathcal{R}} R ~ by,
f ( t , x ) = γ f ~ ( t ~ , x ~ ) (18) f(t,\mathbf{x}) = \gamma\tilde{f}(\tilde{t},\tilde{\mathbf{x}}) \tag*{(18)} f ( t , x ) = γ f ~ ( t ~ , x ~ ) (18)
According to the covariant time component of the infinitesimal spatial volume (9) and the contravariant time component of the current density vector (17), the integrand of the integral relation (7) for the scalar quantity F F F can be recast in the inertial frame R \mathcal{R} R as,
d V 3 ( x ) f ( t , x ) = 1 c d V 30 ( x ) j f 0 ( t , x ) (19) dV_3(\mathbf{x}) f(t,\mathbf{x}) = \frac{1}{c}dV_{30}(\mathbf{x}) j_f^0(t,\mathbf{x}) \tag*{(19)} d V 3 ( x ) f ( t , x ) = c 1 d V 30 ( x ) j f 0 ( t , x ) (19)
According to the covariant components (8) and the covariant time component (9) of the infinitesimal spatial volume vector,
d V 3 μ ( { x ρ } ) j f μ ( { x ρ } ) = d V 3 ( x ) δ 0 μ j f μ ( t , x ) = d V 30 ( x ) j f 0 ( t , x ) (20) dV_{3\mu}(\{x^\rho\})j_f^\mu(\{x^\rho\}) = dV_3(\mathbf{x})\delta_{0\mu}j_f^\mu(t,\mathbf{x}) = dV_{30}(\mathbf{x})j_f^0(t,\mathbf{x}) \tag*{(20)} d V 3 μ ({ x ρ }) j f μ ({ x ρ }) = d V 3 ( x ) δ 0 μ j f μ ( t , x ) = d V 30 ( x ) j f 0 ( t , x ) (20)
In view of the contracted identity (20), the integrand (19) reduces to,
d V 3 ( x ) f ( t , x ) = 1 c d V 3 μ ( { x ρ } ) j f μ ( { x ρ } ) (21) dV_3(\mathbf{x})f(t,\mathbf{x}) = \frac{1}{c}dV_{3\mu}(\{x^\rho\})j_f^\mu(\{x^\rho\}) \tag*{(21)} d V 3 ( x ) f ( t , x ) = c 1 d V 3 μ ({ x ρ }) j f μ ({ x ρ }) (21)
which is a frame independent scalar since it is the contraction of two vectors. Time t t t in the inertial frame R \mathcal{R} R defines a three dimensional spatial hypersurface Σ 3 ( t ) \Sigma_3(t) Σ 3 ( t ) orthogonal to the worldline Σ 1 \Sigma_1 Σ 1 . The volume of the system is a subset of this hypersurface, i.e. V 3 ( t ) ⊂ Σ 3 ( t ) V_3(t) \subset\Sigma_3(t) V 3 ( t ) ⊂ Σ 3 ( t ) . In view of the integrand (21), the extensive scalar function (7) can be recast as,
F ( t ) = 1 c ∫ V 3 ( t ) d V 3 μ ( { x ρ } ) j f μ ( { x ρ } ) (22) F(t) = \frac{1}{c}\int_{V_3(t)} dV_{3\mu}(\{x^\rho\})j_f^\mu(\{x^\rho\}) \tag*{(22)} F ( t ) = c 1 ∫ V 3 ( t ) d V 3 μ ({ x ρ }) j f μ ({ x ρ }) (22)
In order to determine the continuity equation of the scalar function F ( t ) F(t) F ( t ) , we write the variation of this function between during a time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] ,
Δ F i → f ≡ F ( t f ) − F ( t i ) = 1 c ∫ V 3 ( t f ) d V 3 μ ( { x ρ } ) j f μ ( { x ρ } ) − 1 c ∫ V 3 ( t i ) d V 3 μ ( { x ρ } ) j f μ ( { x ρ } ) (23) \Delta F_{i\to f} \equiv F(t_f)-F(t_i) = \frac{1}{c}\int_{V_3(t_f)} dV_{3\mu}(\{x^\rho\})j_f^\mu(\{x^\rho\})
-\frac{1}{c}\int_{V_3(t_i)} dV_{3\mu}(\{x^\rho\})j_f^\mu(\{x^\rho\}) \tag*{(23)} Δ F i → f ≡ F ( t f ) − F ( t i ) = c 1 ∫ V 3 ( t f ) d V 3 μ ({ x ρ }) j f μ ({ x ρ }) − c 1 ∫ V 3 ( t i ) d V 3 μ ({ x ρ }) j f μ ({ x ρ }) (23)
The spatial volumes V 3 ( t i ) V_3(t_i) V 3 ( t i ) and V 3 ( t f ) V_3(t_f) V 3 ( t f ) of the system are represented in space-time by two sections of the parallel spatial hypersurfaces Σ 3 ( t i ) \Sigma_3(t_i) Σ 3 ( t i ) and Σ 3 ( t f ) \Sigma_3(t_f) Σ 3 ( t f ) (Fig. 1). The worldsheet of the fixed enclosure of the system defines a hypersurface of volume V 3 e x t ( t i , t f ) V_3^{\mathrm{ext}}(t_i,t_f) V 3 ext ( t i , t f ) connecting the spatial volumes V 3 ( t i ) V_3(t_i) V 3 ( t i ) and V 3 ( t f ) V_3(t_f) V 3 ( t f ) . We define an infinitesimal volume vector of components d V 3 μ e x t ( { x ρ } ) dV_{3\mu}^{\mathrm{ext}}(\{x^\rho\}) d V 3 μ ext ({ x ρ }) locally orthogonal to the hypersurface of volume V 3 e x t ( t i , t f ) V_3^{\mathrm{ext}}(t_i,t_f) V 3 ext ( t i , t f ) containing the velocity vector u μ u^\mu u μ associated to the worldline of the local rest frame. The orthogonality condition between the volume vector and the worldline is written in the inertial frame R \mathcal{R} R as,
Figure 1. Worldsheet of the system of hypervolume V 4 V_4 V 4 during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] in space-time represented by a cylinder. At the initial time t i t_i t i , the system is represented by a circular section of spatial volume V 3 ( t i ) V_3(t_i) V 3 ( t i ) . At the final time t f t_f t f , the system is represented by a circular section of spatial volume V 3 ( t f ) V_3(t_f) V 3 ( t f ) .
d V 3 μ e x t ( { x ρ } ) u μ = 0 (24) dV_{3\mu}^{\mathrm{ext}}(\{x^\rho\})u^\mu= 0 \tag*{(24)} d V 3 μ ext ({ x ρ }) u μ = 0 (24)
According to contravariant component of the current density vector (16) and the null identity (24), we obtain the orthogonality condition,
d V 3 μ e x t ( { x ρ } ) j f μ ( { x ρ } ) = d V 3 μ e x t ( { x ρ } ) f ~ ( { x ~ ρ } ) u μ ( { x ρ } ) = 0 (25) dV_{3\mu}^{\mathrm{ext}}(\{x^\rho\})j_f^\mu(\{x^\rho\}) = dV_{3\mu}^{\mathrm{ext}}(\{x^\rho\})\widetilde{f}(\{\widetilde{x}^\rho\})u^\mu(\{x^\rho\}) = 0 \tag*{(25)} d V 3 μ ext ({ x ρ }) j f μ ({ x ρ }) = d V 3 μ ext ({ x ρ }) f ({ x ρ }) u μ ({ x ρ }) = 0 (25)
Integrating the orthogonality condition (25), we obtain the null identity,
1 c ∫ V 3 e x t ( t i , t f ) d V 3 μ e x t ( { x ρ } ) j f μ ( { x ρ } ) = 0 (26) \frac{1}{c}\int_{V_3^{\mathrm{ext}}(t_i,t_f)} dV^{\mathrm{ext}}_{3\mu}\left(\{x^\rho\}\right)j_f^\mu\left(\{x^\rho\}\right)=0 \tag*{(26)} c 1 ∫ V 3 ext ( t i , t f ) d V 3 μ ext ( { x ρ } ) j f μ ( { x ρ } ) = 0 (26)
In view of relation (26), the variation of the scalar function F F F (22) during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] is recast as,
Δ F i → f = 1 c ∫ V 3 ( t f ) d V 3 μ ( { x ρ } ) j f μ ( { x ρ } ) + 1 c ∫ V 3 ( t i ) ( − d V 3 μ ( { x ρ } ) ) j f μ ( { x ρ } ) + 1 c ∫ V 3 e x t ( t i , t f ) d V 3 μ e x t ( { x ρ } ) j f μ ( { x ρ } ) (27) \Delta F_{i\to f}=\frac{1}{c}\int_{V_3(t_f)}dV_{3\mu}\left(\{x^\rho\}\right)j_f^\mu\left(\{x^\rho\}\right)+\frac{1}{c}\int_{V_3(t_i)}\left(-dV_{3\mu}\left(\{x^\rho\}\right)\right)j_f^\mu\left(\{x^\rho\}\right)+\frac{1}{c}\int_{V_3^{\mathrm{ext}}(t_i,t_f)}dV^{\mathrm{ext}}_{3\mu}\left(\{x^\rho\}\right)j_f^\mu\left(\{x^\rho\}\right) \tag*{(27)} Δ F i → f = c 1 ∫ V 3 ( t f ) d V 3 μ ( { x ρ } ) j f μ ( { x ρ } ) + c 1 ∫ V 3 ( t i ) ( − d V 3 μ ( { x ρ } ) ) j f μ ( { x ρ } ) + c 1 ∫ V 3 ext ( t i , t f ) d V 3 μ ext ( { x ρ } ) j f μ ( { x ρ } ) (27)
The four-dimensional worldsheet of the system during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] has a hypervolume V 4 V_4 V 4 . [18 ] The boundary of the worldsheet of the system of hypervolume V 4 V_4 V 4 consists of the three-dimensional volumes V 3 ( t i ) V_3(t_i) V 3 ( t i ) , V 3 ( t f ) V_3(t_f) V 3 ( t f ) and V 3 e x t ( t i , t f ) V^{\mathrm{ext}}_3(t_i,t_f) V 3 ext ( t i , t f ) . The infinitesimal volume vectors d V 3 μ dV_{3\mu} d V 3 μ , − d V 3 μ -dV_{3\mu} − d V 3 μ and d V 3 μ e x t dV^{\mathrm{ext}}_{3\mu} d V 3 μ ext are orthogonal to the boundaries and oriented outside the four-dimensional worldsheet of the system. Thus, the variation (27) of the scalar function F F F during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] is recast as a closed integral over the boundary of the four-dimensional worldsheet of the system during this time interval,
Δ F i → f = 1 c ∮ ∂ V 4 ( t i , t f ) d V 3 μ ( { x ρ } ) j f μ ( { x ρ } ) (28) \Delta F_{i\to f}=\frac{1}{c}\oint_{\partial V_4(t_i,t_f)}dV_{3\mu}\left(\{x^\rho\}\right)j_f^\mu\left(\{x^\rho\}\right) \tag*{(28)} Δ F i → f = c 1 ∮ ∂ V 4 ( t i , t f ) d V 3 μ ( { x ρ } ) j f μ ( { x ρ } ) (28)
According to the divergence theorem, [18 ] the integral on the three dimensional closed boundary of worldsheet of the system during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] is equal to the integral of the divergence of the integrand on the four dimensional worldsheet of the system during this time interval,
Δ F i → f = 1 c ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ j f μ ( { x ρ } ) (29) \Delta F_{i\to f}=\frac{1}{c}\int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)\partial_\mu j_f^\mu\left(\{x^\rho\}\right) \tag*{(29)} Δ F i → f = c 1 ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ j f μ ( { x ρ } ) (29)
If the scalar function F F F represents a conserved physical property, there is no variation (29) during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] , i.e. Δ F i → f = 0 \Delta F_{i\to f}=0 Δ F i → f = 0 . Otherwise, there is scalar function Σ F i → f \Sigma_{F\,i\to f} Σ F i → f representing the physical source of F F F in the system,
Δ F i → f = Σ F i → f (30) \Delta F_{i\to f}=\Sigma_{F\,i\to f} \tag*{(30)} Δ F i → f = Σ F i → f (30)
which is the relativistic balance equation of F F F . The physical source of F F F in the system is the integral of the scalar source density σ f \sigma_f σ f of the scalar function density f f f over the worldsheet of the system during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] divided by c c c ,
Σ F i → f = 1 c ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) σ f ( { x ρ } ) (31) \Sigma_{F\,i\to f}=\frac{1}{c}\int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)\sigma_f\left(\{x^\rho\}\right) \tag*{(31)} Σ F i → f = c 1 ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) σ f ( { x ρ } ) (31)
In view of the integral relations (29) and (31) for the variation of the scalar function F F F , the relativistic balance equation (30) multiplied by c c c yields,
∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ j f μ ( { x ρ } ) = ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) σ f ( { x ρ } ) (32) \int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)\partial_\mu j_f^\mu\left(\{x^\rho\}\right)=\int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)\sigma_f\left(\{x^\rho\}\right) \tag*{(32)} ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ j f μ ( { x ρ } ) = ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) σ f ( { x ρ } ) (32)
The identification of the integrands in the relativistic balance equation (32) yields the relativistic continuity equation of F F F ,
∂ μ j f μ ( { x ρ } ) = σ f ( { x ρ } ) (33) \partial_\mu j_f^\mu\left(\{x^\rho\}\right)=\sigma_f\left(\{x^\rho\}\right) \tag*{(33)} ∂ μ j f μ ( { x ρ } ) = σ f ( { x ρ } ) (33)
In view of the contravariant components of the current density (16) , the relativistic continuity equation is recast as, [18 ]
∂ μ ( f ~ ( { x ~ ρ } ) u μ ( { x ~ ρ } ) ) = σ f ( { x ρ } ) (inertial frame) (34) \partial_\mu\left(\tilde{f}\left(\{\tilde{x}^\rho\}\right)u^\mu\left(\{\tilde{x}^\rho\}\right)\right)=\sigma_f\left(\{x^\rho\}\right)\qquad\text{(inertial frame)} \tag*{(34)} ∂ μ ( f ~ ( { x ~ ρ } ) u μ ( { x ~ ρ } ) ) = σ f ( { x ρ } ) (inertial frame) (34)
The covariant time component of the partial derivative vectorial operator is written is written as,
∂ 0 = 1 c ∂ t (35) \partial_0=\frac{1}{c}\partial_t \tag*{(35)} ∂ 0 = c 1 ∂ t (35)
The gradient ∇ \nabla ∇ is expressed in covariant components in the orthonormal dual basis frame as,
∇ = e j ∂ j (36) \nabla=e^j\partial_j \tag*{(36)} ∇ = e j ∂ j (36)
Using the contravariant components of the velocity vector (A.24), the contravariant components of the current density vector (16) and the relativistic transformation law for the density (18) , the contravariant spatial components of the current density vector are given by,
j f j ( { x ρ } ) = γ f ~ ( t ~ , x ~ ) v j ( t , x ) = f ( t , x ) v j ( t , x ) (37) j_f^j\left(\{x^\rho\}\right)=\gamma\tilde{f}(\tilde{t},\tilde{x})v^j(t,x)=f(t,x)v^j(t,x) \tag*{(37)} j f j ( { x ρ } ) = γ f ~ ( t ~ , x ~ ) v j ( t , x ) = f ( t , x ) v j ( t , x ) (37)
In view of the frame duality (A.8), the spatial velocity vector (A.23), the contravariant temporal and spatial components of the current density vector (17) and (37), the partial time derivative (35) and the gradient (36), the divergence of the current density vector is written as,
∂ μ j f μ ( { x ρ } ) = ∂ 0 j f 0 ( { x ρ } ) + ∂ j j f j ( { x ρ } ) = ∂ t f ( t , x ) + ( e k ∂ k ) ⋅ ( f ( t , x ) v j ( t , x ) e j ) = ∂ t f ( t , x ) + ∇ ⋅ ( f ( t , x ) v ( t , x ) ) (38) \begin{aligned}
\partial_\mu j_f^\mu(\{x^\rho\}) &= \partial_0 j_f^0(\{x^\rho\})+\partial_j j_f^j(\{x^\rho\}) \\
&= \partial_t f(t,\mathbf{x})+(\mathbf{e}^k\partial_k)\cdot\left(f(t,\mathbf{x})\mathbf{v}^j(t,\mathbf{x})\mathbf{e}_j\right) \\
&= \partial_t f(t,\mathbf{x})+\nabla\cdot\left(f(t,\mathbf{x})\mathbf{v}(t,\mathbf{x})\right) \tag*{(38)}
\end{aligned} ∂ μ j f μ ({ x ρ }) = ∂ 0 j f 0 ({ x ρ }) + ∂ j j f j ({ x ρ }) = ∂ t f ( t , x ) + ( e k ∂ k ) ⋅ ( f ( t , x ) v j ( t , x ) e j ) = ∂ t f ( t , x ) + ∇ ⋅ ( f ( t , x ) v ( t , x ) ) (38)
In view of the divergence of the current density vector (38), the relativistic continuity equation (33) is written as,
∂ t f ( t , x ) + ∇ ⋅ ( f ( t , x ) v ( t , x ) ) = σ f ( t , x ) (inertial frame) (39) \partial_t f(t,\mathbf{x})+\nabla\cdot\left(f(t,\mathbf{x})\mathbf{v}(t,\mathbf{x})\right)=\sigma_f(t,\mathbf{x})\qquad\text{(inertial frame)} \tag*{(39)} ∂ t f ( t , x ) + ∇ ⋅ ( f ( t , x ) v ( t , x ) ) = σ f ( t , x ) (inertial frame) (39)
and in view of the contravariant components (37), the current density spatial vector in the inertial frame R \mathcal{R} R is given by,
j f ( t , x ) = f ( t , x ) v ( t , x ) (40) \mathbf{j}_f(t,\mathbf{x})=f(t,\mathbf{x})\mathbf{v}(t,\mathbf{x}) \tag*{(40)} j f ( t , x ) = f ( t , x ) v ( t , x ) (40)
According to the relativistic continuity equation (33), the source density of f f f is a frame independent scalar,
σ f ( { x ρ } ) = σ ~ f ( { x ~ ρ } ) (41) \sigma_f(\{x^\rho\})=\widetilde{\sigma}_f(\{\widetilde{x}^\rho\}) \tag*{(41)} σ f ({ x ρ }) = σ f ({ x ρ }) (41)
In the local rest frame R ~ \widetilde{\mathcal{R}} R , in view of relations (10) and (41), the relativistic continuity equation (34) becomes,
∂ ~ μ ( f ~ ( { x ~ ρ } ) u ~ μ ( { x ~ ρ } ) ) = σ ~ f ( { x ~ ρ } ) (local rest frame) (42) \widetilde{\partial}_\mu\left(\widetilde{f}(\{\widetilde{x}^\rho\})\widetilde{u}^\mu(\{\widetilde{x}^\rho\})\right)=\widetilde{\sigma}_f(\{\widetilde{x}^\rho\})\qquad\text{(local rest frame)} \tag*{(42)} ∂ μ ( f ({ x ρ }) u μ ({ x ρ }) ) = σ f ({ x ρ }) (local rest frame) (42)
Using the identities,
u ~ μ = c δ 0 μ and ∂ ~ μ = c − 1 ∂ t ~ (43) \widetilde{u}^\mu=c\,\delta^{0\mu}
\qquad\text{and}\qquad
\widetilde{\partial}_\mu=c^{-1}\partial_{\widetilde{t}} \tag*{(43)} u μ = c δ 0 μ and ∂ μ = c − 1 ∂ t (43)
the relativistic continuity equation (42) in the local rest frame R ~ \widetilde{\mathcal{R}} R becomes,
∂ t ~ f ~ ( t ~ , x ~ ) = σ ~ f ( t ~ , x ~ ) (local rest frame) (44) \partial_{\widetilde{t}}\widetilde{f}(\widetilde{t},\widetilde{\mathbf{x}})=\widetilde{\sigma}_f(\widetilde{t},\widetilde{\mathbf{x}})\qquad\text{(local rest frame)} \tag*{(44)} ∂ t f ( t , x ) = σ f ( t , x ) (local rest frame) (44)
The extension of the current approach to determine the relativistic continuity equation of a vector function is straightforward. By analogy with the scalar function (22) of time F ( t ) F(t) F ( t ) , the contravariant scalar components F ν ( t ) F^\nu(t) F ν ( t ) of the vector function, where ν ∈ { 0 , 1 , 2 , 3 } \nu\in\{0,1,2,3\} ν ∈ { 0 , 1 , 2 , 3 } , are written as, [18 ]
F ν ( t ) = 1 c ∫ V 3 ( t ) d V 3 ( { x ρ } ) j f μ ν ( { x ρ } ) (45) F^\nu(t)=\frac{1}{c}\int_{V_3(t)} dV_3\left(\{x^\rho\}\right)j_f^{\mu\nu}\left(\{x^\rho\}\right) \tag*{(45)} F ν ( t ) = c 1 ∫ V 3 ( t ) d V 3 ( { x ρ } ) j f μν ( { x ρ } ) (45)
where j f μ ν ( { x ρ } ) j_f^{\mu\nu}\left(\{x^\rho\}\right) j f μν ( { x ρ } ) are the components of the current density of the physical property F ν F^\nu F ν . The current density of a scalar (rank-0 tensor) property F F F is a vector (rank-1 tensor) of components j μ j^\mu j μ and the current density of a vector (rank-1 tensor) property of components F ν F^\nu F ν is a rank-2 tensor of components j μ ν j^{\mu\nu} j μν . By analogy with the scalar function (22) of time F ( t ) F(t) F ( t ) , the variation of the scalar components F ν ( t ) F^\nu(t) F ν ( t ) of a vector function is written as,
Δ F i → f ν = 1 c ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ j f μ ν ( { x ρ } ) (46) \Delta F^\nu_{i\to f}=\frac{1}{c}\int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)\partial_\mu j_f^{\mu\nu}\left(\{x^\rho\}\right) \tag*{(46)} Δ F i → f ν = c 1 ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ j f μν ( { x ρ } ) (46)
If the vector function (46) of scalar components F ν F^\nu F ν represents a conserved physical property, there is no variation during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] , i.e. Δ F i → f ν = 0 \Delta F^\nu_{i\to f}=0 Δ F i → f ν = 0 . Otherwise, there is scalar function Σ F i → f ν \Sigma^\nu_{F\,i\to f} Σ F i → f ν representing the physical source of F ν F^\nu F ν in the system,
Δ F i → f ν = Σ F i → f ν (47) \Delta F^\nu_{i\to f}=\Sigma^\nu_{F\,i\to f} \tag*{(47)} Δ F i → f ν = Σ F i → f ν (47)
which is the component of relativistic balance equation of the vector function. The physical source of F ν F^\nu F ν in the system is the integral of the component σ f ν \sigma_f^\nu σ f ν of the vector source density of the component f ν f^\nu f ν of the vector function density over the worldsheet of the system during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] divided by c c c ,
Σ F i → f ν = 1 c ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) σ f ν ( { x ρ } ) (48) \Sigma^\nu_{F\,i\to f}=\frac{1}{c}\int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)\sigma_f^\nu\left(\{x^\rho\}\right) \tag*{(48)} Σ F i → f ν = c 1 ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) σ f ν ( { x ρ } ) (48)
In view of the integral relations (46) and (48), the relativistic balance equation (47) multiplied by c c c yields,
∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ j f μ ν ( { x ρ } ) = ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) σ f ν ( { x ρ } ) (49) \int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)\partial_\mu j_f^{\mu\nu}\left(\{x^\rho\}\right)=\int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)\sigma_f^\nu\left(\{x^\rho\}\right) \tag*{(49)} ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ j f μν ( { x ρ } ) = ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) σ f ν ( { x ρ } ) (49)
The identification of the integrands in the relativistic balance equation (49) yields the relativistic continuity equation of F ν F^\nu F ν , [18 ]
∂ μ j f μ ν ( { x ρ } ) = σ f ν ( { x ρ } ) (50) \boxed{\partial_\mu j_f^{\mu\nu}\left(\{x^\rho\}\right)=\sigma_f^\nu\left(\{x^\rho\}\right)} \tag*{(50)} ∂ μ j f μν ( { x ρ } ) = σ f ν ( { x ρ } ) (50)
Relativistic second law of thermodynamics
We state in this section the relativistic first law of thermodynamics and establish the relativistic continuity equation of the frame-independent scalar function entropy S S S based on the results of § 2 . In particular, we derive the transformation for the entropy density and show that the entropy source density is a frame independent quantity. On a microscopic scale, the elementary constituents of the fluid can be considered as hard spheres in local rest frame R ~ \widetilde{\mathcal{R}} R . Thus, the elementary constituents appear as ellipsoids flattened in the direction of motion in the inertial frame R \mathcal{R} R . The volume contraction in the direction of motion (A.43) then ensures that the number Ω \Omega Ω of microscopic states corresponding to a macroscopic state remains unchanged under a Lorentz transformation. Thus, in agreement with Einstein [4 ] and Planck [5 ] , the entropy of the system is a frame independent quantity,
S ( t ) = S ~ ( t ~ ) (51) S(t)=\widetilde{S}(\widetilde{t}) \tag*{(51)} S ( t ) = S ( t ) (51)
The relativistic second law of thermodynamics reads :
For every thermodynamic system there exists a scalar extensive state function entropy S ( t ) S(t) S ( t ) . During the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] , where t i < t f t_i<t_f t i < t f , the entropy source Σ S i → f \Sigma_{S\,i\to f} Σ S i → f is a monotonically increasing function of time t t t ,
Δ S i → f = Σ S i → f ≥ 0 (52) \Delta S_{i\to f}=\Sigma_{S\,i\to f}\geq0 \tag*{(52)} Δ S i → f = Σ S i → f ≥ 0 (52)
In the distant future, the entropy S ( t ) S(t) S ( t ) of an isolated system tends to a finite maximum S max S_{\max} S m a x that is compatible with the constraints imposed on the system, such as its internal walls or characteristics of the enclosure,
lim t → ∞ S ( t ) = S max (53) \lim_{t\to\infty}S(t)=S_{\max} \tag*{(53)} t → ∞ lim S ( t ) = S m a x (53)
The contravariant components of the entropy current density vector (16) in the inertial frame R \mathcal{R} R are written as, [18 ]
j s μ ( { x ρ } ) = s ~ ( { x ~ ρ } ) u μ ( { x ρ } ) (54) j_s^\mu(\{x^\rho\}) = \tilde{s}(\{\tilde{x}^\rho\})u^\mu(\{x^\rho\}) \tag*{(54)} j s μ ({ x ρ }) = s ~ ({ x ~ ρ }) u μ ({ x ρ }) (54)
where s ~ ( { x ~ ρ } ) \tilde{s}(\{\tilde{x}^\rho\}) s ~ ({ x ~ ρ }) is the entropy density in the local rest frame R ~ \widetilde{\mathcal{R}} R . According to the relativistic transformation law for a density (18), the entropy density s ( t , x ) s(t,\mathbf{x}) s ( t , x ) in the inertial frame R \mathcal{R} R is related to the entropy density s ~ ( t ~ , x ~ ) \tilde{s}(\tilde{t},\tilde{\mathbf{x}}) s ~ ( t ~ , x ~ ) in the local rest frame R ~ \widetilde{\mathcal{R}} R by,
s ( t , x ) = γ s ~ ( t ~ , x ~ ) (55) s(t,\mathbf{x}) = \gamma\tilde{s}(\tilde{t},\tilde{\mathbf{x}}) \tag*{(55)} s ( t , x ) = γ s ~ ( t ~ , x ~ ) (55)
The entropy (22) reads,
S ( t ) = 1 c ∫ V 3 ( t ) d V 3 μ j s μ ( { x ρ } ) (56) S(t) = \frac{1}{c}\int_{V_3(t)} dV_3\,\mu\,j_s^\mu(\{x^\rho\}) \tag*{(56)} S ( t ) = c 1 ∫ V 3 ( t ) d V 3 μ j s μ ({ x ρ }) (56)
Using Gauss’ theorem, the entropy variation (29) during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] is written as,
Δ S i → f = 1 c ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ j s μ ( { x ρ } ) (57) \Delta S_{i\to f} = \frac{1}{c}\int_{V_4(t_i,t_f)} dV_4(\{x^\rho\})\,\partial_\mu j_s^\mu(\{x^\rho\}) \tag*{(57)} Δ S i → f = c 1 ∫ V 4 ( t i , t f ) d V 4 ({ x ρ }) ∂ μ j s μ ({ x ρ }) (57)
The entropy source (31) is written as,
Σ S i → f = 1 c ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) σ s ( { x ρ } ) (58) \Sigma_{S\,i\to f} = \frac{1}{c}\int_{V_4(t_i,t_f)} dV_4(\{x^\rho\})\,\sigma_s(\{x^\rho\}) \tag*{(58)} Σ S i → f = c 1 ∫ V 4 ( t i , t f ) d V 4 ({ x ρ }) σ s ({ x ρ }) (58)
In view of the second law of thermodynamics (52), the identification of the integrands of the entropy variation (57) and the entropy source (58) yields the relativistic entropy continuity equation,
∂ μ j s μ ( { x ρ } ) = σ s ( { x ρ } ) ≥ 0 (59) \partial_\mu j_s^\mu(\{x^\rho\}) = \sigma_s(\{x^\rho\}) \ge0 \tag*{(59)} ∂ μ j s μ ({ x ρ }) = σ s ({ x ρ }) ≥ 0 (59)
In view of the contravariant components of the entropy current density vector (54), the relativistic entropy continuity equation (59) is recast as, [18 ]
∂ μ ( s ~ ( { x ~ ρ } ) u μ ( { x ρ } ) ) = σ s ( { x ρ } ) ≥ 0 (inertial frame) (60) \partial_\mu\left(\tilde{s}(\{\tilde{x}^\rho\})u^\mu(\{x^\rho\})\right) = \sigma_s(\{x^\rho\}) \ge0 \qquad\text{(inertial frame)} \tag*{(60)} ∂ μ ( s ~ ({ x ~ ρ }) u μ ({ x ρ }) ) = σ s ({ x ρ }) ≥ 0 (inertial frame) (60)
The relativistic continuity equation for the entropy (39) in the inertial frame R \mathcal{R} R is written as,
∂ t s ( t , x ) + ∇ ⋅ ( s ( t , x ) v ( t , x ) ) = σ s ( t , x ) ≥ 0 (inertial frame) (61) \partial_t s(t,\mathbf{x}) + \nabla\cdot\left(s(t,\mathbf{x})\,\mathbf{v}(t,\mathbf{x})\right) = \sigma_s(t,\mathbf{x}) \ge0 \qquad\text{(inertial frame)} \tag*{(61)} ∂ t s ( t , x ) + ∇ ⋅ ( s ( t , x ) v ( t , x ) ) = σ s ( t , x ) ≥ 0 (inertial frame) (61)
and the entropy current density spatial vector (40) in the inertial frame R \mathcal{R} R is given by,
j s ( t , x ) = s ( t , x ) v ( t , x ) (62) j_s(t,x) = s(t,x)\,v(t,x) \tag*{(62)} j s ( t , x ) = s ( t , x ) v ( t , x ) (62)
According to the relativistic continuity equation (59) , the entropy source density is a frame independent scalar,
σ s ( { x ρ } ) = σ ~ s ( { x ~ ρ } ) (63) \sigma_s(\{x^\rho\}) = \tilde{\sigma}_s(\{\tilde{x}^\rho\}) \tag*{(63)} σ s ({ x ρ }) = σ ~ s ({ x ~ ρ }) (63)
In the local rest frame R ~ \widetilde{\mathcal{R}} R , in view of the entropy source density (63) and the contravariant components of the entropy current density vector,
j ~ s μ ( { x ~ ρ } ) = s ~ ( { x ~ ρ } ) u ~ μ ( { x ~ ρ } ) (64) \tilde{j}_s^\mu(\{\tilde{x}^\rho\}) = \tilde{s}(\{\tilde{x}^\rho\})\tilde{u}^\mu(\{\tilde{x}^\rho\}) \tag*{(64)} j ~ s μ ({ x ~ ρ }) = s ~ ({ x ~ ρ }) u ~ μ ({ x ~ ρ }) (64)
the relativistic continuity equation (59) becomes,
∂ ~ μ ( s ~ ( { x ~ ρ } ) u ~ μ ( { x ~ ρ } ) ) = σ ~ s ( { x ~ ρ } ) ≥ 0 (local rest frame) (65) \tilde{\partial}_\mu\left(\tilde{s}(\{\tilde{x}^\rho\})\tilde{u}^\mu(\{\tilde{x}^\rho\})\right) = \tilde{\sigma}_s(\{\tilde{x}^\rho\}) \ge0 \qquad\text{(local rest frame)} \tag*{(65)} ∂ ~ μ ( s ~ ({ x ~ ρ }) u ~ μ ({ x ~ ρ }) ) = σ ~ s ({ x ~ ρ }) ≥ 0 (local rest frame) (65)
Using the identities (43) , the relativistic entropy continuity equation (65) in the local rest frame R ~ \widetilde{\mathcal{R}} R becomes,
∂ t ~ s ~ ( t ~ , x ~ ) = σ ~ s ( t ~ , x ~ ) ≥ 0 (local rest frame) (66) \partial_{\tilde{t}}\tilde{s}(\tilde{t},\tilde{x}) = \tilde{\sigma}_s(\tilde{t},\tilde{x}) \ge0 \qquad\text{(local rest frame)} \tag*{(66)} ∂ t ~ s ~ ( t ~ , x ~ ) = σ ~ s ( t ~ , x ~ ) ≥ 0 (local rest frame) (66)
Relativistic first law of thermodynamics
We state in this section the relativistic first law of thermodynamics and establish the relativistic continuity equation of the contravariant component P ν P^\nu P ν of the momentum vector in Minkowski space based on the results of § 2 . In order to do so, we use the fact that the momentum source is proportional to the contravariant component F ν F^\nu F ν of the net force exerted on the system. The relativistic first law of thermodynamics reads:
For every thermodynamic system there exists a vector extensive state function momentum of contravariant components P ν P^\nu P ν . During the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] , where t i < t f t_i < t_f t i < t f , the momentum source Σ P i → f ν \Sigma^\nu_{P\,i\to f} Σ P i → f ν is the net force exerted on the system multiplied by the time interval F ν Δ t i → f F^\nu\Delta t_{i\to f} F ν Δ t i → f ,
Δ P i → f ν = F ν Δ t i → f (67) \Delta P^\nu_{i\to f} = F^\nu\Delta t_{i\to f} \tag*{(67)} Δ P i → f ν = F ν Δ t i → f (67)
For an isolated system, there is not net force acting on the system, i.e. F i → f ν F^\nu_{i\to f} F i → f ν . Thus the momentum variation during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] vanishes, i.e. Δ P i → f ν = 0 \Delta P^\nu_{i\to f}=0 Δ P i → f ν = 0 , which implies that the components of the momentum vector are conserved,
P ν ( t ) = c o n s t ( isolated system ) (68) P^\nu(t) = \mathrm{const} \qquad(\text{isolated system}) \tag*{(68)} P ν ( t ) = const ( isolated system ) (68)
The scalar components of the momentum vector (45) read,
P ν ( t ) = 1 c ∫ V 3 ( t ) d V 3 μ ( { x ρ } ) T μ ν ( { x ρ } ) (69) P^\nu(t) = \frac{1}{c}\int_{V_3(t)} dV_{3\mu}\left(\{x^\rho\}\right)T^{\mu\nu}\left(\{x^\rho\}\right) \tag*{(69)} P ν ( t ) = c 1 ∫ V 3 ( t ) d V 3 μ ( { x ρ } ) T μν ( { x ρ } ) (69)
where T μ ν T^{\mu\nu} T μν are the contravariant components of the rank-2 stress-energy-momentum tensor. The momentum variation (46) is written as,
Δ P i → f ν = 1 c ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ T μ ν ( { x ρ } ) (70) \Delta P^\nu_{i\to f} = \frac{1}{c}\int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)\partial_\mu T^{\mu\nu}\left(\{x^\rho\}\right) \tag*{(70)} Δ P i → f ν = c 1 ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) ∂ μ T μν ( { x ρ } ) (70)
The net force (48) exerted on the system multiplied by the time interval F ν Δ t i → f F^\nu\Delta t_{i\to f} F ν Δ t i → f is written as,
F ν Δ t i → f = 1 c ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) f ν ( { x ρ } ) (71) F^\nu\Delta t_{i\to f} = \frac{1}{c}\int_{V_4(t_i,t_f)}dV_4\left(\{x^\rho\}\right)f^\nu\left(\{x^\rho\}\right) \tag*{(71)} F ν Δ t i → f = c 1 ∫ V 4 ( t i , t f ) d V 4 ( { x ρ } ) f ν ( { x ρ } ) (71)
where f ν f^\nu f ν is the component of the force density vector. In view of the first law of thermodynamics (67) , the identification of the integrands in the momentum variation (70) and the momentum source (71) yields the relativistic energy-momentum continuity equation,
∂ μ T μ ν ( { x ρ } ) = f ν ( { x ρ } ) (72) \partial_\mu T^{\mu\nu}\left(\{x^\rho\}\right)=f^\nu\left(\{x^\rho\}\right) \tag*{(72)} ∂ μ T μν ( { x ρ } ) = f ν ( { x ρ } ) (72)
Relativistic orbital angular momentum continuity equation
We show in this section that in the absence of intrinsic angular momentum, the relativistic continuity equation for the contravariant components of the orbital angular momentum tensor L μ ν L^{\mu\nu} L μν and the relativistic energy-momentum continuity equation (72) , established in § 4 , imply that the contravariant components T μ ν T^{\mu\nu} T μν of the stress energy momentum tensor are symmetric. The contravariant components of the orbital angular momentum tensor are defined as the antisymmetric product of the contravariant components of the position vector of the centre of mass and the momentum vector,
L μ ν = X μ P ν − X ν P μ (73) L^{\mu\nu}=X^\mu P^\nu-X^\nu P^\mu \tag*{(73)} L μν = X μ P ν − X ν P μ (73)
In view of the contravariant components of the momentum vector (69) the contravariant components of the orbital angular momentum tensor (73) are written as, [20 , 21 ]
L μ ν = 1 c ∫ V 3 ( t ) d V 3 ρ ( x μ T ρ ν − x ν T ρ μ ) (74) L^{\mu\nu}=\frac{1}{c}\int_{V_3(t)}dV_3\,\rho\left(x^\mu T^{\rho\nu}-x^\nu T^{\rho\mu}\right) \tag*{(74)} L μν = c 1 ∫ V 3 ( t ) d V 3 ρ ( x μ T ρ ν − x ν T ρ μ ) (74)
Using Gauss’ theorem, the variation of the contravariant components of the orbital angular momentum tensor during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] are written as,
Δ L i → f μ ν = 1 c ∫ V 4 ( t i , t f ) d V 4 ∂ ρ ( x μ T ρ ν − x ν T ρ μ ) (75) \Delta L^{\mu\nu}_{i\to f}=\frac{1}{c}\int_{V_4(t_i,t_f)}dV_4\,\partial_\rho\left(x^\mu T^{\rho\nu}-x^\nu T^{\rho\mu}\right) \tag*{(75)} Δ L i → f μν = c 1 ∫ V 4 ( t i , t f ) d V 4 ∂ ρ ( x μ T ρ ν − x ν T ρ μ ) (75)
The orbital angular momentum source is the torque,
Σ L μ ν i → f = 1 c ∫ V 4 ( t i , t f ) d V 4 t μ ν (76) \Sigma_{L^{\mu\nu}\,i\to f}=\frac{1}{c}\int_{V_4(t_i,t_f)}dV_4\,t^{\mu\nu} \tag*{(76)} Σ L μν i → f = c 1 ∫ V 4 ( t i , t f ) d V 4 t μν (76)
where the contravariant components of the torque density tensor are defined as,
t μ ν = x μ f ν − x ν f μ (77) t^{\mu\nu}=x^\mu f^\nu-x^\nu f^\mu \tag*{(77)} t μν = x μ f ν − x ν f μ (77)
The relativistic orbital angular momentum theorem is written as,
Δ L i → f μ ν = Σ L μ ν i → f (78) \Delta L^{\mu\nu}_{i\to f}=\Sigma_{L^{\mu\nu}\,i\to f} \tag*{(78)} Δ L i → f μν = Σ L μν i → f (78)
In view of the angular momentum theorem (78) , the identification of the integrands in the angular momentum variation (75) and the angular momentum source (76) yields the relativistic orbital angular momentum continuity equation,
∂ ρ ( x μ T ρ ν − x ν T ρ μ ) = t μ ν (79) \partial_{\rho}\left(x^{\mu}T^{\rho\nu}-x^{\nu}T^{\rho\mu}\right)=t^{\mu\nu} \tag*{(79)} ∂ ρ ( x μ T ρ ν − x ν T ρ μ ) = t μν (79)
According to relativistic energy-momentum continuity equation (72) and the identity ∂ ρ x μ = δ ρ μ \partial_{\rho}x^{\mu}=\delta_{\rho}^{\mu} ∂ ρ x μ = δ ρ μ , the relativistic angular momentum continuity equation (79) yields the identity,
T μ ν − T ν μ + x μ f ν − x ν f μ = t μ ν (80) T^{\mu\nu}-T^{\nu\mu}+x^{\mu}f^{\nu}-x^{\nu}f^{\mu}=t^{\mu\nu} \tag*{(80)} T μν − T ν μ + x μ f ν − x ν f μ = t μν (80)
In view of the relativistic torque density tensor (77) , the identity (80) implies that the stress-energy-momentum tensor is symmetric, [21 ]
T μ ν = T ν μ (81) T^{\mu\nu}=T^{\nu\mu} \tag*{(81)} T μν = T ν μ (81)
provided there is no intrinsic angular momentum, i.e. S μ ν = 0 S^{\mu\nu}=0 S μν = 0 .
Relativistic matter continuity equation
We establish in this section the relativistic continuity equation of the frame-independent scalar function number of moles of substance N N N based on the results of § 2. We also derive the relativistic transformation law for the number density.
In order to do so, we consider a relativistic fluid consisting of a single substance.
In special relativity, the number of moles of substance in a system is a frame independent quantity,
N = N ~ (82) N=\widetilde{N} \tag*{(82)} N = N (82)
The contravariant components of the current density vector of substance (16) is written as, [18 ]
j n μ = n ~ u μ (83) j_{n}^{\mu}=\widetilde{n}u^{\mu} \tag*{(83)} j n μ = n u μ (83)
where n ~ \widetilde{n} n is the number density in the local rest frame R ~ \widetilde{\mathcal{R}} R . According to the relativistic transformation law for the number density (18) , the number density n n n in the inertial frame R \mathcal{R} R is related to the number density n ~ \widetilde{n} n in the local rest frame R ~ \widetilde{\mathcal{R}} R by,
n = γ n ~ (84) n=\gamma\widetilde{n} \tag*{(84)} n = γ n (84)
The number of moles of substance (22) reads,
N = 1 c ∫ V 3 ( t ) d V 3 μ j n μ (85) N = \frac{1}{c} \int_{V_3(t)} dV_{3\mu} j_n^\mu \tag*{(85)} N = c 1 ∫ V 3 ( t ) d V 3 μ j n μ (85)
Using Gauss’ theorem, the variation of the number of moles of substance (29) during the time interval [ t i , t f ] [t_i,t_f] [ t i , t f ] is written as,
Δ N i → f = 1 c ∫ V 4 ( t i , t f ) d V 4 ∂ μ j n μ (86) \Delta N_{i\to f} = \frac{1}{c} \int_{V_4(t_i,t_f)} dV_4 \partial_\mu j_n^\mu \tag*{(86)} Δ N i → f = c 1 ∫ V 4 ( t i , t f ) d V 4 ∂ μ j n μ (86)
There is no matter source (31) for a relativistic perfect fluid consisting of a single substance,
Σ N i → f = 1 c ∫ V 4 ( t i , t f ) d V 4 σ n = 0 (87) \Sigma N_{i\to f} = \frac{1}{c} \int_{V_4(t_i,t_f)} dV_4 \sigma_n = 0 \tag*{(87)} Σ N i → f = c 1 ∫ V 4 ( t i , t f ) d V 4 σ n = 0 (87)
The matter conservation equation for a relativistic perfect fluid is written as,
Δ N i → f = Σ N i → f = 0 (88) \Delta N_{i\to f} = \Sigma N_{i\to f} = 0 \tag*{(88)} Δ N i → f = Σ N i → f = 0 (88)
In view of the matter conservation equation (88), the identification of the integrands in the variation of the number of moles (86) and the matter source (87) yields the relativistic matter continuity equation, [18 ]
∂ μ j n μ = 0 (89) \partial_\mu j_n^\mu= 0 \tag*{(89)} ∂ μ j n μ = 0 (89)
In view of the contravariant components of the current density vector of substance (83), the relativistic matter continuity equation (89) is recast as,
∂ μ ( n ~ u μ ) = 0 (inertial frame) (90) \partial_\mu\left(\widetilde{n} u^\mu\right) = 0 \qquad\text{(inertial frame)} \tag*{(90)} ∂ μ ( n u μ ) = 0 (inertial frame) (90)
The relativistic matter continuity equation (39) in the inertial frame R \mathcal{R} R is written as,
∂ t n + ∇ ⋅ ( n v ) = 0 (inertial frame) (91) \partial_t n + \nabla\cdot(n v) = 0 \qquad\text{(inertial frame)} \tag*{(91)} ∂ t n + ∇ ⋅ ( n v ) = 0 (inertial frame) (91)
and the spatial vector matter current density (40) in the inertial frame R \mathcal{R} R is given by,
j n = n v (92) j_n = n v \tag*{(92)} j n = n v (92)
In the local rest frame R ~ \widetilde{\mathcal{R}} R , in view of the contravariant components of the current density vector of substance,
j ~ n μ = n ~ u ~ μ (93) \widetilde{j}_n^\mu= \widetilde{n}\widetilde{u}^\mu \tag*{(93)} j n μ = n u μ (93)
the relativistic continuity equation (89) becomes,
∂ ~ μ ( n ~ u ~ μ ) = 0 (local rest frame) (94) \boxed{\widetilde{\partial}_{\mu}\left(\widetilde{n}\,\widetilde{u}^{\mu}\right)=0\qquad\text{(local rest frame)}} \tag*{(94)} ∂ μ ( n u μ ) = 0 (local rest frame) (94)
Using the identities (43) , the relativistic matter continuity equation (94) in the local rest frame R ~ \widetilde{\mathcal{R}} R becomes,
∂ t ~ n ~ = 0 (local rest frame) (95) \boxed{\partial_{\widetilde{t}}\widetilde{n}=0\qquad\text{(local rest frame)}} \tag*{(95)} ∂ t n = 0 (local rest frame) (95)
Dynamics of a relativistic dust
In order to generalise the relativistic dynamics of a particle described in § Appendix B, we consider in this section a relativistic dust, which is a continuous medium consisting of relativistic non interacting particles. In particular, we derive the relativistic transformation laws for the mass density and the energy density. In the local rest frame R ~ \widetilde{\mathcal{R}} R , the mass density of a relativistic dust is defined as, [20 ]
ρ ~ = n ~ M ~ (96) \widetilde{\rho}=\widetilde{n}\,\widetilde{M} \tag*{(96)} ρ = n M (96)
In view of the mass density (96) in the local rest frame R ~ \widetilde{\mathcal{R}} R , the mass density of a relativistic dust is defined in the inertial frame R \mathcal{R} R as,
ρ = n M (97) \rho=nM \tag*{(97)} ρ = n M (97)
According to the relativistic transformation laws for the mass (B.10) and the number density (84) , the mass density ρ \rho ρ in the inertial frame R \mathcal{R} R is expressed in terms of the mass density ρ ~ \widetilde{\rho} ρ in the rest frame R ~ \widetilde{\mathcal{R}} R as, [22 ]
ρ = γ 2 ρ ~ (98) \boxed{\rho=\gamma^{2}\widetilde{\rho}} \tag*{(98)} ρ = γ 2 ρ (98)
which shows that under a Lorentz transformation the mass density transforms as a component of a rank-2 tensor and not as a component of a vector. In view of the momentum vector (B.3), the relativistic transformation law of the number density (84) and the mass density (96) , the contravariant components of the momentum density vector of a relativistic dust in the inertial frame R \mathcal{R} R are given by,
p μ = n P μ = n M ~ u μ = γ n ~ M ~ u μ = γ ρ ~ u μ (99) p^\mu= n P^\mu= n \widetilde{M} u^\mu= \gamma\widetilde{n}\widetilde{M} u^\mu= \gamma\widetilde{\rho} u^\mu \tag*{(99)} p μ = n P μ = n M u μ = γ n M u μ = γ ρ u μ (99)
In view of the covariant components of the velocity vector (A.24), and the momentum vector (B.13), the relativistic transformation law of the number density (84) and the mass density (96) , the covariant components of the momentum density vector of a relativistic dust in the inertial frame R \mathcal{R} R are written as,
p μ = n P μ = n M ~ u μ = γ n ~ M ~ u μ = γ ρ ~ u μ (100) p_\mu= n P_\mu= n \widetilde{M} u_\mu= \gamma\widetilde{n}\widetilde{M} u_\mu= \gamma\widetilde{\rho} u_\mu \tag*{(100)} p μ = n P μ = n M u μ = γ n M u μ = γ ρ u μ (100)
In contrast to the momentum, the momentum density is not a vector since the contravariant (99) or covariant (100) components of the momentum density vector are the product of the frame dependent factor γ ρ ~ \gamma\widetilde{\rho} γ ρ and the contravariant u μ u^\mu u μ or covariant u μ u_\mu u μ components of the velocity vector. Using the energy (B.7) and the mass (96) , the energy density e ~ \widetilde{e} e in the local rest frame R ~ \widetilde{\mathcal{R}} R is written as,
e ~ = n ~ E ~ = n ~ M c 2 = ρ ~ c 2 (101) \widetilde{e} = \widetilde{n}\widetilde{E} = \widetilde{n} M c^2 = \widetilde{\rho}c^2 \tag*{(101)} e = n E = n M c 2 = ρ c 2 (101)
Using the relativistic transformation laws for the energy (B.8), the number density (84) and the mass density (98) , and the rest energy density (101) , the energy density e ~ \widetilde{e} e in the inertial frame R \mathcal{R} R is written as,
e = n E = γ 2 n ~ E ~ = γ 2 n ~ M c 2 = γ 2 ρ ~ c 2 = ρ c 2 (102) e = n E = \gamma^2 \widetilde{n}\widetilde{E} = \gamma^2 \widetilde{n} M c^2 = \gamma^2 \widetilde{\rho}c^2 = \rho c^2 \tag*{(102)} e = n E = γ 2 n E = γ 2 n M c 2 = γ 2 ρ c 2 = ρ c 2 (102)
The energy density (102) in the inertial frame R \mathcal{R} R is related to the energy density (101) in the local rest frame R ~ \widetilde{\mathcal{R}} R by,
e = γ 2 e ~ (103) e = \gamma^2 \widetilde{e} \tag*{(103)} e = γ 2 e (103)
In view of the relativistic transformation law for the mass density (98) , the contravariant components of the momentum density vector (99) , and the energy density (102) , the contravariant time component of the momentum density in the inertial frame R \mathcal{R} R is given by,
p 0 = e c = ρ c = γ 2 ρ ~ c (104) p^0 = \frac{e}{c} = \rho c = \gamma^2 \widetilde{\rho}c \tag*{(104)} p 0 = c e = ρ c = γ 2 ρ c (104)
In view of the relativistic transformation law for the mass density (98) , the covariant component of the momentum density vector (100) , and the energy density (102) , the covariant time component of the momentum density in the inertial frame R \mathcal{R} R is written as,
p 0 = − e c = − ρ c = − γ 2 ρ ~ c (105) p_0=-\frac{e}{c}=-\rho c=-\gamma^2\tilde{\rho}c \tag*{(105)} p 0 = − c e = − ρ c = − γ 2 ρ ~ c (105)
According to the spatial velocity vector (A.23), the relativistic transformation law for the mass density (98) and the contravariant components of the momentum density vector (99) , the spatial momentum density vector in the inertial frame R \mathcal{R} R is written as,
p = p j e j = γ 2 ρ ~ v j e j = γ 2 ρ ~ v = ρ v (106) \boldsymbol{p}=p^j\boldsymbol{e}_j=\gamma^2\tilde{\rho}v^j\boldsymbol{e}_j=\gamma^2\tilde{\rho}\boldsymbol{v}=\rho\boldsymbol{v} \tag*{(106)} p = p j e j = γ 2 ρ ~ v j e j = γ 2 ρ ~ v = ρ v (106)
where p j e j = p j e j p^j\boldsymbol{e}_j=p_j\boldsymbol{e}^j p j e j = p j e j .
Stress-energy-momentum tensor of a relativistic dust
We determine the contravariant components T μ ν T^{\mu\nu} T μν of a stress energy-momentum tensor for a relativistic dust. Since a relativistic dust is made of non interacting particles, the only non zero contravariant component of the stress-energy-momentum tensor of a relativistic dust in the local rest frame R ~ \widetilde{\mathcal{R}} R is the contravariant time component defined as the rest energy density, [20 ]
T ~ 00 = e ~ = ρ ~ c 2 (107) \widetilde{T}^{00}=\tilde{e}=\tilde{\rho}c^2 \tag*{(107)} T 00 = e ~ = ρ ~ c 2 (107)
Thus, the contravariant components of the stress-energy-momentum tensor of a relativistic dust are formally written as,
T ~ μ ν = ρ ~ c 2 δ μ 0 δ ν 0 (108) \widetilde{T}^{\mu\nu}=\tilde{\rho}c^2\delta^{\mu0}\delta^{\nu0} \tag*{(108)} T μν = ρ ~ c 2 δ μ 0 δ ν 0 (108)
The contravariant components of the velocity vector are purely time-like in the local rest frame R ~ \widetilde{\mathcal{R}} R ,
u ~ μ = c δ μ 0 (109) \tilde{u}^{\mu}=c\delta^{\mu0} \tag*{(109)} u ~ μ = c δ μ 0 (109)
In view of the contravariant components of the velocity vector (109) , the contravariant components of the stress-energy-momentum tensor (108) become,
T ~ μ ν = ρ ~ u ~ μ u ~ ν (110) \widetilde{T}^{\mu\nu}=\widetilde{\rho}\,\widetilde{u}^{\mu}\,\widetilde{u}^{\nu} \tag*{(110)} T μν = ρ u μ u ν (110)
In view of the rest energy density (107) , the contravariant components of the stress-energy-momentum tensor (110) are recast as,
T ~ μ ν = e ~ c 2 u ~ μ u ~ ν (111) \widetilde{T}^{\mu\nu}=\frac{\widetilde{e}}{c^{2}}\,\widetilde{u}^{\mu}\,\widetilde{u}^{\nu} \tag*{(111)} T μν = c 2 e u μ u ν (111)
The contravariant components of the stress-energy-momentum tensor T μ ν T^{\mu\nu} T μν of a relativistic dust in the inertial frame R \mathcal{R} R are obtained by performing inverse Lorentz transformations on the contravariant components of the stress-energy-momentum tensor T ~ μ ν \widetilde{T}^{\mu\nu} T μν of a relativistic dust in in the local rest frame R ~ \widetilde{\mathcal{R}} R ,
T μ ν = ( Λ − 1 ) ρ μ ( Λ − 1 ) λ ν T ~ ρ λ (112) T^{\mu\nu}=(\Lambda^{-1})^{\mu}_{\ \rho}(\Lambda^{-1})^{\nu}_{\ \lambda}\,\widetilde{T}^{\rho\lambda} \tag*{(112)} T μν = ( Λ − 1 ) ρ μ ( Λ − 1 ) λ ν T ρ λ (112)
Using the components of the stress-energy-momentum tensor (108) in the local rest frame R ~ \widetilde{\mathcal{R}} R , the components of the stress-energy-momentum tensor (112) in the inertial frame R \mathcal{R} R are recast as,
T μ ν = ( Λ − 1 ) 0 μ ( Λ − 1 ) 0 ν ρ ~ c 2 (113) T^{\mu\nu}=(\Lambda^{-1})^{\mu}_{\ 0}(\Lambda^{-1})^{\nu}_{\ 0}\,\widetilde{\rho}c^{2} \tag*{(113)} T μν = ( Λ − 1 ) 0 μ ( Λ − 1 ) 0 ν ρ c 2 (113)
In view of the components of the inverse Lorentz transformation (A.40) and the energy density (102) , the contravariant time component of the stress-energy-momentum tensor (113) of a relativistic dust is the energy density,
T 00 = ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 0 ρ ~ c 2 = γ 2 ρ ~ c 2 = e (114) T^{00}=(\Lambda^{-1})^{0}_{\ 0}(\Lambda^{-1})^{0}_{\ 0}\,\widetilde{\rho}c^{2}=\gamma^{2}\widetilde{\rho}c^{2}=e \tag*{(114)} T 00 = ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 0 ρ c 2 = γ 2 ρ c 2 = e (114)
According to the contravariant time component of the velocity vector (A.21), the contravariant time component of the stress-energy-momentum tensor (114) is recast as, [22 , 20 ]
T 00 = ρ ~ ( γ c ) ( γ c ) = ρ ~ u 0 u 0 (115) T^{00}=\widetilde{\rho}(\gamma c)(\gamma c)=\widetilde{\rho}\,u^{0}u^{0} \tag*{(115)} T 00 = ρ ( γ c ) ( γ c ) = ρ u 0 u 0 (115)
In view of the components of the inverse Lorentz transformation (A.40) and the energy density (102) , the contravariant temporal and spatial components of the stress-energy-momentum tensor (113) of a relativistic dust are the components of the energy current density vector up to a factor c − 1 c^{-1} c − 1 , [22 , 20 ]
T j 0 = ( Λ − 1 ) 0 j ( Λ − 1 ) 0 0 ρ ~ c 2 = γ 2 ρ ~ c v j = e c v j (116) T^{j0} = \left(\Lambda^{-1}\right)^{j}_{0}\left(\Lambda^{-1}\right)^{0}_{0}\tilde{\rho}c^{2} = \gamma^{2}\tilde{\rho}cv^{j} = \frac{e}{c}v^{j} \tag*{(116)} T j 0 = ( Λ − 1 ) 0 j ( Λ − 1 ) 0 0 ρ ~ c 2 = γ 2 ρ ~ c v j = c e v j (116)
According to the contravariant temporal and spatial components of the velocity vector (A.21) and (A.24), the contravariant temporal and spatial components of the stress-energy-momentum tensor (116) are recast as,
T j 0 = ρ ~ ( γ c ) ( γ v j ) = ρ ~ u 0 u j (117) T^{j0} = \tilde{\rho}(\gamma c)(\gamma v^{j}) = \tilde{\rho}u^{0}u^{j} \tag*{(117)} T j 0 = ρ ~ ( γ c ) ( γ v j ) = ρ ~ u 0 u j (117)
In view of the components of the inverse Lorentz transformation (A.40) and the contravariant components of the spatial momentum density vector (106) , the contravariant space and time components of the stress-energy-momentum tensor (113) of a relativistic dust are the components of the momentum density up to a factor c c c , [22 , 20 ]
T 0 j = ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 j ρ ~ c 2 = γ 2 ρ ~ c v j = c p j (118) T^{0j} = \left(\Lambda^{-1}\right)^{0}_{0}\left(\Lambda^{-1}\right)^{j}_{0}\tilde{\rho}c^{2} = \gamma^{2}\tilde{\rho}cv^{j} = cp^{j} \tag*{(118)} T 0 j = ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 j ρ ~ c 2 = γ 2 ρ ~ c v j = c p j (118)
According to the contravariant temporal and spatial components of the velocity vector (A.21) and (A.24) and the contravariant of the momentum density vector (99) , the contravariant temporal and spatial components of the stress-energy-momentum tensor (118) are recast as,
T 0 j = ρ ~ ( γ v j ) ( γ c ) = ρ ~ u j u 0 (119) T^{0j} = \tilde{\rho}(\gamma v^{j})(\gamma c) = \tilde{\rho}u^{j}u^{0} \tag*{(119)} T 0 j = ρ ~ ( γ v j ) ( γ c ) = ρ ~ u j u 0 (119)
In view of relation (A.40) and (106) , the contravariant space and time components of the stress-energy-momentum tensor (113) of a relativistic dust are the components are the contravariant components of the momentum current density up to a factor γ \gamma γ , [22 , 20 ]
T j k = ( Λ − 1 ) 0 j ( Λ − 1 ) 0 k ρ ~ c 2 = γ 2 ρ ~ v j v k (120) T^{jk} = \left(\Lambda^{-1}\right)^{j}_{0}\left(\Lambda^{-1}\right)^{k}_{0}\tilde{\rho}c^{2} = \gamma^{2}\tilde{\rho}v^{j}v^{k} \tag*{(120)} T j k = ( Λ − 1 ) 0 j ( Λ − 1 ) 0 k ρ ~ c 2 = γ 2 ρ ~ v j v k (120)
According to the contravariant spatial components of the velocity vector (A.24), the contravariant spatial components of the stress-energy-momentum tensor (120) are recast as,
T j k = ρ ~ ( γ v j ) ( γ v k ) = ρ ~ u j u k (121) T^{jk} = \tilde{\rho}(\gamma v^{j})(\gamma v^{k}) = \tilde{\rho}u^{j}u^{k} \tag*{(121)} T j k = ρ ~ ( γ v j ) ( γ v k ) = ρ ~ u j u k (121)
In view of the contravariant temporal and spatial components of the stress-energy-momentum tensor (115), (117), (119) and (121), the contravariant components of the stress-energy-momentum tensor are written as, [22 , 20 ]
T μ ν = ρ ~ u μ u ν (122) T^{\mu\nu}=\tilde{\rho}u^\mu u^\nu \tag*{(122)} T μν = ρ ~ u μ u ν (122)
In view of the rest energy density (101), the contravariant components (122) are recast as,
T μ ν = e ~ c 2 u μ u ν (123) T^{\mu\nu}=\frac{\tilde{e}}{c^2}u^\mu u^\nu \tag*{(123)} T μν = c 2 e ~ u μ u ν (123)
Dynamics of a relativistic perfect fluid
In order to generalise the relativistic dynamics of a dust described in § 7, we consider in this section a relativistic perfect fluid, which is a continuous medium consisting of relativistic particles undergoing elastic collisions. In particular, we derive the relativistic transformation law for the internal energy density. For a relativistic dust, the internal energy density vanishes. However, for a relativistic perfect fluid in the local rest frame R ~ \widetilde{\mathcal{R}} R , the difference between the energy density e ~ \tilde{e} e ~ and the internal energy density u ~ \tilde{u} u ~ is the rest energy density ρ ~ c 2 \tilde{\rho}c^2 ρ ~ c 2 ,
e ~ − u ~ = ρ ~ c 2 (124) \tilde{e}-\tilde{u}=\tilde{\rho}c^2 \tag*{(124)} e ~ − u ~ = ρ ~ c 2 (124)
For a relativistic perfect fluid in the inertial frame R \mathcal{R} R , the difference between the energy density e e e and the internal energy density u u u is the sum of the rest energy density ρ ~ c 2 \tilde{\rho}c^2 ρ ~ c 2 and the kinetic energy density ( γ 2 − 1 ) ρ ~ c 2 (\gamma^2-1)\tilde{\rho}c^2 ( γ 2 − 1 ) ρ ~ c 2 ,
e − u = ρ ~ c 2 + ( γ 2 − 1 ) ρ ~ c 2 = γ 2 ρ ~ c 2 (125) e-u=\tilde{\rho}c^2+(\gamma^2-1)\tilde{\rho}c^2=\gamma^2\tilde{\rho}c^2 \tag*{(125)} e − u = ρ ~ c 2 + ( γ 2 − 1 ) ρ ~ c 2 = γ 2 ρ ~ c 2 (125)
In view of the relativistic transformation law for the mass density (98), the energy density difference (125) becomes,
e − u = ρ c 2 (126) e-u=\rho c^2 \tag*{(126)} e − u = ρ c 2 (126)
The energy density difference (125) in the inertial frame R \mathcal{R} R is related to the energy density difference (124) in the local rest frame R ~ \widetilde{\mathcal{R}} R by,
e − u = γ 2 ( e ~ − u ~ ) (127) e-u=\gamma^2(\tilde{e}-\tilde{u}) \tag*{(127)} e − u = γ 2 ( e ~ − u ~ ) (127)
In view of the relativistic transformation laws for the energy density (103) and the energy density difference (127), the internal energy density u u u in the inertial frame R \mathcal{R} R is related to the internal energy density u ~ \tilde{u} u ~ in the local rest frame R ~ \widetilde{\mathcal{R}} R by,
u = γ 2 u ~ (128) u = \gamma^{2}\tilde{u} \tag*{(128)} u = γ 2 u ~ (128)
Stress-energy-momentum tensor of a relativistic perfect fluid
We determine the contravariant components T μ ν T^{\mu\nu} T μν of a stress energy-momentum tensor for a relativistic perfect fluid. Such a fluid is characterised by the random motion of the microscopic constituents that generates an isotropic pressure on a macroscopic scale. Thus, in the local rest frame R ~ \widetilde{\mathcal{R}} R , the spatial part of the stress-energy-momentum tensor of a relativistic perfect fluid is the identity tensor 1 3 \mathbf{1}_{3} 1 3 multiplied by the scalar pressure p ~ \tilde{p} p ~ in the local rest frame. It is written in contravariant components as,
T ~ j k = p ~ δ j k (129) \widetilde{T}^{jk} = \tilde{p}\,\delta^{jk} \tag*{(129)} T j k = p ~ δ j k (129)
Since a relativistic perfect fluid is a generalisation of a relativistic dust, the internal energy density u ~ \tilde{u} u ~ needs to be added to the rest energy density ρ ~ c 2 \tilde{\rho}c^{2} ρ ~ c 2 . In order to generalise the contravariant time component of the stress-energy-momentum tensor of a dust (107), the contravariant time component of the stress-energy-momentum tensor of a relativistic perfect fluid in the local rest frame R ~ \widetilde{\mathcal{R}} R is written as,
T ~ 00 = ρ ~ c 2 + u ~ = e ~ (130) \widetilde{T}^{00} = \tilde{\rho}c^{2} + \tilde{u} = \tilde{e} \tag*{(130)} T 00 = ρ ~ c 2 + u ~ = e ~ (130)
In view of the temporal and spatial contravariant components of the stress-energy-momentum tensor (129) and (130), the contravariant components of the stress-energy-momentum tensor in the local rest frame R ~ \widetilde{\mathcal{R}} R read,
T ~ μ ν = ( ρ ~ c 2 + u ~ ) δ μ 0 δ ν 0 + p ~ ( η μ ν + δ μ 0 δ ν 0 ) (131) \widetilde{T}^{\mu\nu} = \left(\tilde{\rho}c^{2}+\tilde{u}\right)\delta^{\mu0}\delta^{\nu0}+\tilde{p}\left(\eta^{\mu\nu}+\delta^{\mu0}\delta^{\nu0}\right) \tag*{(131)} T μν = ( ρ ~ c 2 + u ~ ) δ μ 0 δ ν 0 + p ~ ( η μν + δ μ 0 δ ν 0 ) (131)
The contravariant components of the stress-energy-momentum tensor for a perfect fluid (131) in the local rest frame R ~ \widetilde{\mathcal{R}} R reduce to the contravariant components of the stress-energy-momentum tensor for relativistic dust (108) in the absence of pressure and internal energy density, i.e. p ~ = u ~ = 0 \widetilde{p} = \widetilde{u} = 0 p = u = 0 . The contravariant components of the stress-energy-momentum tensor (131) in the local rest frame R ~ \widetilde{\mathcal{R}} R can be recast as,
T ~ μ ν = ( ρ ~ + 1 c 2 ( u ~ + p ~ ) ) c 2 δ μ 0 δ ν 0 + p ~ η μ ν (132) \widetilde{T}^{\mu\nu} = \left(\widetilde{\rho} + \frac{1}{c^2}(\widetilde{u}+\widetilde{p})\right)c^2\delta^{\mu0}\delta^{\nu0}+\widetilde{p}\eta^{\mu\nu} \tag*{(132)} T μν = ( ρ + c 2 1 ( u + p ) ) c 2 δ μ 0 δ ν 0 + p η μν (132)
In relativistic thermodynamics, the inertia density m ~ \widetilde{m} m in the local rest frame R ~ \widetilde{\mathcal{R}} R is different from the mass density ρ ~ \widetilde{\rho} ρ since the pressure and the internal energy carry inertia. It is defined as,
m ~ = ρ ~ + 1 c 2 ( u ~ + p ~ ) = 1 c 2 ( e ~ + p ~ ) (133) \boxed{\widetilde{m}=\widetilde{\rho}+\frac{1}{c^2}(\widetilde{u}+\widetilde{p})=\frac{1}{c^2}(\widetilde{e}+\widetilde{p})} \tag*{(133)} m = ρ + c 2 1 ( u + p ) = c 2 1 ( e + p ) (133)
In view of the inertia density (133) , the contravariant components of the stress-energy-momentum tensor (132) in the local rest frame R ~ \widetilde{\mathcal{R}} R reduce to,
T ~ μ ν = m ~ c 2 δ μ 0 δ ν 0 + p ~ η μ ν (134) \widetilde{T}^{\mu\nu}=\widetilde{m}c^2\delta^{\mu0}\delta^{\nu0}+\widetilde{p}\eta^{\mu\nu} \tag*{(134)} T μν = m c 2 δ μ 0 δ ν 0 + p η μν (134)
Using the contravariant components of the velocity vector (109) in the local rest frame R ~ \widetilde{\mathcal{R}} R , the contravariant components of the stress-energy-momentum tensor (134) in the local rest frame R ~ \widetilde{\mathcal{R}} R reduce to,
T ~ μ ν = m ~ u ~ μ u ~ ν + p ~ η μ ν (135) \boxed{\widetilde{T}^{\mu\nu}=\widetilde{m}\,\widetilde{u}^{\mu}\widetilde{u}^{\nu}+\widetilde{p}\eta^{\mu\nu}} \tag*{(135)} T μν = m u μ u ν + p η μν (135)
In view of the inertia density (133) , the contravariant components of the stress-energy-momentum tensor (135) in the local rest frame R ~ \widetilde{\mathcal{R}} R are recast as,
T ~ μ ν = ( e ~ + p ~ c 2 ) u ~ μ u ~ ν + p ~ η μ ν (136) \boxed{\widetilde{T}^{\mu\nu}=\left(\frac{\widetilde{e}+\widetilde{p}}{c^2}\right)\widetilde{u}^{\mu}\widetilde{u}^{\nu}+\widetilde{p}\eta^{\mu\nu}} \tag*{(136)} T μν = ( c 2 e + p ) u μ u ν + p η μν (136)
The contravariant components of the stress-energy-momentum tensor T μ ν T^{\mu\nu} T μν of relativistic perfect fluid in the inertial frame R \mathcal{R} R are obtained by performing inverse Lorentz transformations on the contravariant components of the stress-energy-momentum tensor T ~ μ ν \widetilde{T}^{\mu\nu} T μν of a relativistic perfect fluid in in the local rest frame R ~ \widetilde{\mathcal{R}} R ,
T μ ν = ( Λ − 1 ) μ ρ ( Λ − 1 ) ν λ T ~ ρ λ (137) T^{\mu\nu}=(\Lambda^{-1})^{\mu}{}_{\rho}(\Lambda^{-1})^{\nu}{}_{\lambda}\widetilde{T}^{\rho\lambda} \tag*{(137)} T μν = ( Λ − 1 ) μ ρ ( Λ − 1 ) ν λ T ρ λ (137)
Using the components of the stress-energy-momentum tensor (134) in the local rest frame R ~ \widetilde{\mathcal{R}} R , the components of the stress-energy-momentum tensor (137) in the inertial frame R \mathcal{R} R are recast as,
T μ ν = ( Λ − 1 ) 0 μ ( Λ − 1 ) 0 ν m ~ c 2 + ( ( Λ − 1 ) 0 μ ( Λ − 1 ) 0 ν η 00 + ( Λ − 1 ) j μ ( Λ − 1 ) k ν η j k ) p ~ (138) T^{\mu\nu}=(\Lambda^{-1})^\mu_{0}(\Lambda^{-1})^\nu_{0}\tilde{m}c^2+\left((\Lambda^{-1})^\mu_{0}(\Lambda^{-1})^\nu_{0}\eta^{00}+(\Lambda^{-1})^\mu_{j}(\Lambda^{-1})^\nu_{k}\eta^{jk}\right)\tilde{p} \tag*{(138)} T μν = ( Λ − 1 ) 0 μ ( Λ − 1 ) 0 ν m ~ c 2 + ( ( Λ − 1 ) 0 μ ( Λ − 1 ) 0 ν η 00 + ( Λ − 1 ) j μ ( Λ − 1 ) k ν η j k ) p ~ (138)
The contravariant time component of the stress-energy-momentum tensor (138) of a relativistic fluid in the inertial frame R \mathcal{R} R reads,
T 00 = ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 0 m ~ c 2 + ( ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 0 η 00 + ( Λ − 1 ) j 0 ( Λ − 1 ) k 0 η j k ) p ~ (139) T^{00}=(\Lambda^{-1})^0_{0}(\Lambda^{-1})^0_{0}\tilde{m}c^2+\left((\Lambda^{-1})^0_{0}(\Lambda^{-1})^0_{0}\eta^{00}+(\Lambda^{-1})^0_{j}(\Lambda^{-1})^0_{k}\eta^{jk}\right)\tilde{p} \tag*{(139)} T 00 = ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 0 m ~ c 2 + ( ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 0 η 00 + ( Λ − 1 ) j 0 ( Λ − 1 ) k 0 η j k ) p ~ (139)
Using the Minkowski metric (A.6), the Lorentz factor (A.26) and the components of the inverse Lorentz transformation (A.40), we obtain the relation,
v j v k = ( v i δ i j ) ( v i δ i k ) = v i 2 δ j k = v 2 δ j k (140) v^jv^k=\left(v^i\delta^j_i\right)\left(v^i\delta^k_i\right)=v^{i2}\delta^{jk}=v^2\delta^{jk} \tag*{(140)} v j v k = ( v i δ i j ) ( v i δ i k ) = v i 2 δ j k = v 2 δ j k (140)
which yields the identity,
( Λ − 1 ) 0 0 ( Λ − 1 ) 0 0 η 00 + ( Λ − 1 ) j 0 ( Λ − 1 ) k 0 η j k = γ 2 ( v 2 c 2 − 1 ) = − 1 (141) (\Lambda^{-1})^0_{0}(\Lambda^{-1})^0_{0}\eta^{00}+(\Lambda^{-1})^0_{j}(\Lambda^{-1})^0_{k}\eta^{jk}=\gamma^2\left(\frac{v^2}{c^2}-1\right)=-1 \tag*{(141)} ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 0 η 00 + ( Λ − 1 ) j 0 ( Λ − 1 ) k 0 η j k = γ 2 ( c 2 v 2 − 1 ) = − 1 (141)
Taking into account the metric (A.6), the contravariant time component of the velocity vector (A.21) and the identity (141) , the contravariant time component of the stress-energy-momentum tensor of a relativistic perfect fluid (139) reduces to,
T 00 = m ~ ( γ c ) ( γ c ) − p ~ = m ~ u 0 u 0 + p ~ η 00 (142) T^{00}=\tilde{m}(\gamma c)(\gamma c)-\tilde{p}=\tilde{m}u^0u^0+\tilde{p}\eta^{00} \tag*{(142)} T 00 = m ~ ( γ c ) ( γ c ) − p ~ = m ~ u 0 u 0 + p ~ η 00 (142)
The contravariant space and time components of the stress-energy-momentum tensor (138) of a relativistic fluid in the inertial frame R \mathcal{R} R read,
T j 0 = ( Λ − 1 ) 0 j ( Λ − 1 ) 0 0 m ~ c 2 + ( ( Λ − 1 ) 0 j ( Λ − 1 ) 0 0 η 00 + ( Λ − 1 ) k j ( Λ − 1 ) ℓ 0 η k ℓ ) p ~ (143) T^{j0}=(\Lambda^{-1})^j_{0}(\Lambda^{-1})^0_{0}\tilde{m}c^2+\left((\Lambda^{-1})^j_{0}(\Lambda^{-1})^0_{0}\eta^{00}+(\Lambda^{-1})^j_{k}(\Lambda^{-1})^0_{\ell}\eta^{k\ell}\right)\tilde{p} \tag*{(143)} T j 0 = ( Λ − 1 ) 0 j ( Λ − 1 ) 0 0 m ~ c 2 + ( ( Λ − 1 ) 0 j ( Λ − 1 ) 0 0 η 00 + ( Λ − 1 ) k j ( Λ − 1 ) ℓ 0 η k ℓ ) p ~ (143)
Using the Minkowski metric (A.6) and the components of the inverse Lorentz transformation (A.40), we obtain the identity,
( Λ − 1 ) 0 j ( Λ − 1 ) 0 0 η 00 + ( Λ − 1 ) k j ( Λ − 1 ) ℓ 0 η k ℓ = − γ 2 c ( v j − δ k j v k ) = 0 (144) (\Lambda^{-1})^j_{0}(\Lambda^{-1})^0_{0}\eta^{00}+(\Lambda^{-1})^j_{k}(\Lambda^{-1})^0_{\ell}\eta^{k\ell}=-\frac{\gamma^2}{c}\left(v^j-\delta^j_kv^k\right)=0 \tag*{(144)} ( Λ − 1 ) 0 j ( Λ − 1 ) 0 0 η 00 + ( Λ − 1 ) k j ( Λ − 1 ) ℓ 0 η k ℓ = − c γ 2 ( v j − δ k j v k ) = 0 (144)
Taking into account the metric (A.6), the contravariant temporal and spatial components of the velocity vector (A.21) and (A.24), and the identity (144) , the contravariant temporal and spatial components of the stress-energy-momentum tensor of a relativistic perfect fluid (143) reduce to,
T j 0 = m ~ ( γ v j ) ( γ c ) = m ~ u j u 0 + p ~ η j 0 (145) T^{j0}=\tilde{m}\left(\gamma v^{j}\right)(\gamma c)=\tilde{m}u^{j}u^{0}+\tilde{p}\eta^{j0} \tag*{(145)} T j 0 = m ~ ( γ v j ) ( γ c ) = m ~ u j u 0 + p ~ η j 0 (145)
The contravariant space and time components of the stress-energy-momentum tensor (138) of a relativistic fluid in the inertial frame R \mathcal{R} R read,
T 0 j = ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 j m ~ c 2 + ( ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 j η 00 + ( Λ − 1 ) k 0 ( Λ − 1 ) ℓ j η k ℓ ) p ~ (146) \begin{aligned}
T^{0j}={}&\left(\Lambda^{-1}\right)^{0}_{0}\left(\Lambda^{-1}\right)^{j}_{0}\tilde{m}c^{2}\\
&+\left(\left(\Lambda^{-1}\right)^{0}_{0}\left(\Lambda^{-1}\right)^{j}_{0}\eta^{00}+\left(\Lambda^{-1}\right)^{0}_{k}\left(\Lambda^{-1}\right)^{j}_{\ell}\eta^{k\ell}\right)\tilde{p} \tag*{(146)}
\end{aligned} T 0 j = ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 j m ~ c 2 + ( ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 j η 00 + ( Λ − 1 ) k 0 ( Λ − 1 ) ℓ j η k ℓ ) p ~ (146)
Using the Minkowski metric (A.6) and the components of the inverse Lorentz transformation (A.40), we obtain the identity,
( Λ − 1 ) 0 0 ( Λ − 1 ) 0 j η 00 + ( Λ − 1 ) k 0 ( Λ − 1 ) ℓ j η k ℓ = − γ 2 c ( v j − δ k j v k ) = 0 (147) \left(\Lambda^{-1}\right)^{0}_{0}\left(\Lambda^{-1}\right)^{j}_{0}\eta^{00}+\left(\Lambda^{-1}\right)^{0}_{k}\left(\Lambda^{-1}\right)^{j}_{\ell}\eta^{k\ell}=-\frac{\gamma^{2}}{c}\left(v^{j}-\delta^{j}_{k}v^{k}\right)=0 \tag*{(147)} ( Λ − 1 ) 0 0 ( Λ − 1 ) 0 j η 00 + ( Λ − 1 ) k 0 ( Λ − 1 ) ℓ j η k ℓ = − c γ 2 ( v j − δ k j v k ) = 0 (147)
Taking into account the metric (A.6), the contravariant temporal and spatial components of the velocity vector (A.21) and (A.24), and the identity (147) , the contravariant temporal and spatial components of the stress-energy-momentum tensor of a relativistic perfect fluid (146) reduce to,
T 0 j = m ~ ( γ c ) ( γ v j ) = m ~ u j u 0 + p ~ η 0 j (148) T^{0j}=\tilde{m}(\gamma c)(\gamma v^{j})=\tilde{m}u^{j}u^{0}+\tilde{p}\eta^{0j} \tag*{(148)} T 0 j = m ~ ( γ c ) ( γ v j ) = m ~ u j u 0 + p ~ η 0 j (148)
The contravariant spatial components of the stress-energy-momentum tensor (138) of a relativistic fluid in the inertial frame R \mathcal{R} R read,
T j k = ( Λ − 1 ) 0 j ( Λ − 1 ) 0 k m ~ c 2 + ( ( Λ − 1 ) 0 j ( Λ − 1 ) 0 k η 00 + ( Λ − 1 ) ℓ j ( Λ − 1 ) m k η ℓ m ) p ~ (149) \begin{aligned}
T^{jk}={}&\left(\Lambda^{-1}\right)^{j}_{0}\left(\Lambda^{-1}\right)^{k}_{0}\tilde{m}c^{2}\\
&+\left(\left(\Lambda^{-1}\right)^{j}_{0}\left(\Lambda^{-1}\right)^{k}_{0}\eta^{00}+\left(\Lambda^{-1}\right)^{j}_{\ell}\left(\Lambda^{-1}\right)^{k}_{m}\eta^{\ell m}\right)\tilde{p} \tag*{(149)}
\end{aligned} T j k = ( Λ − 1 ) 0 j ( Λ − 1 ) 0 k m ~ c 2 + ( ( Λ − 1 ) 0 j ( Λ − 1 ) 0 k η 00 + ( Λ − 1 ) ℓ j ( Λ − 1 ) m k η ℓ m ) p ~ (149)
Using the Minkowski metric (A.6), the components of the inverse Lorentz transformation (A.40) and the relation (140) , we obtain the identity,
( Λ − 1 ) 0 j ( Λ − 1 ) 0 k η 00 + ( Λ − 1 ) ℓ j ( Λ − 1 ) m k η ℓ m = − γ 2 ( v 2 c 2 − 1 ) δ j k = δ j k (150) \left(\Lambda^{-1}\right)^{j}_{0}\left(\Lambda^{-1}\right)^{k}_{0}\eta^{00}+\left(\Lambda^{-1}\right)^{j}_{\ell}\left(\Lambda^{-1}\right)^{k}_{m}\eta^{\ell m}=-\gamma^{2}\left(\frac{v^{2}}{c^{2}}-1\right)\delta^{jk}=\delta^{jk} \tag*{(150)} ( Λ − 1 ) 0 j ( Λ − 1 ) 0 k η 00 + ( Λ − 1 ) ℓ j ( Λ − 1 ) m k η ℓ m = − γ 2 ( c 2 v 2 − 1 ) δ j k = δ j k (150)
Taking into account the metric (A.6), the contravariant spatial components of the velocity vector (A.24) and the identity (150) , the contravariant spatial components of the stress-energy-momentum tensor of a relativistic perfect fluid (149) reduce to,
T j k = m ~ ( γ v j ) ( γ v k ) + p ~ δ j k = m ~ u j u k + p ~ η j k (151) T^{jk}=\tilde{m}\left(\gamma v^j\right)\left(\gamma v^k\right)+\tilde{p}\delta^{jk}=\tilde{m}u^j u^k+\tilde{p}\eta^{jk} \tag*{(151)} T j k = m ~ ( γ v j ) ( γ v k ) + p ~ δ j k = m ~ u j u k + p ~ η j k (151)
In view of the contravariant temporal and spatial components of the stress-energy-momentum tensor of a relativistic perfect fluid (142) , (145) , (148) and (151) , the contravariant components of the stress-energy-momentum tensor of a relativistic perfect fluid in the inertial frame R \mathcal{R} R are written as,
T μ ν = m ~ u μ u ν + p ~ η μ ν (152) T^{\mu\nu}=\tilde{m}u^\mu u^\nu+\tilde{p}\eta^{\mu\nu} \tag*{(152)} T μν = m ~ u μ u ν + p ~ η μν (152)
In view of the inertia density (133) , the contravariant components of the stress-energy-momentum tensor of a relativistic perfect fluid (152) are recast as, [22 , 20 ]
T μ ν = ( e ~ + p ~ c 2 ) u μ u ν + p ~ η μ ν (153) T^{\mu\nu}=\left(\frac{\tilde{e}+\tilde{p}}{c^2}\right)u^\mu u^\nu+\tilde{p}\eta^{\mu\nu} \tag*{(153)} T μν = ( c 2 e ~ + p ~ ) u μ u ν + p ~ η μν (153)
Relativistic temperature
We derive the relativistic transformation law for the temperature in this section, which sheds light on the Einstein, Planck and Ott controversy. [4 , 5 , 11 ] In the local rest frame R ~ \widetilde{\mathcal{R}} R , the temperature T ~ \widetilde{T} T is defined as,
T ~ = ∂ u ~ ∂ s ~ (154) \widetilde{T}=\frac{\partial\widetilde{u}}{\partial\widetilde{s}} \tag*{(154)} T = ∂ s ∂ u (154)
In the inertial frame R \mathcal{R} R , the temperature T T T is defined as,
T = ∂ u ∂ s (155) T=\frac{\partial u}{\partial s} \tag*{(155)} T = ∂ s ∂ u (155)
In view of the transformation laws for the internal energy density (128) and the entropy density (55) , the temperature in the inertial frame R \mathcal{R} R is recast as,
T = ∂ ( γ 2 u ~ ) ∂ ( γ s ~ ) = ∂ ( γ 2 u ~ ) ∂ s ~ ( ∂ ( γ s ~ ) ∂ s ~ ) − 1 (156) T=\frac{\partial(\gamma^2\widetilde{u})}{\partial(\gamma\widetilde{s})}
=\frac{\partial(\gamma^2\widetilde{u})}{\partial\widetilde{s}}
\left(\frac{\partial(\gamma\widetilde{s})}{\partial\widetilde{s}}\right)^{-1} \tag*{(156)} T = ∂ ( γ s ) ∂ ( γ 2 u ) = ∂ s ∂ ( γ 2 u ) ( ∂ s ∂ ( γ s ) ) − 1 (156)
Since the internal energy density u ~ \tilde{u} u ~ and the entropy density s ~ \tilde{s} s ~ in the local rest frame R ~ \tilde{\mathcal{R}} R ~ are independent of the Lorentz factor γ \gamma γ , we obtain the identities,
∂ ( γ 2 u ~ ) ∂ s ~ = γ 2 ∂ u ~ ∂ s ~ and ∂ ( γ s ~ ) ∂ s ~ = γ (157) \frac{\partial(\gamma^{2} \tilde{u})}{\partial\tilde{s}} = \gamma^{2} \frac{\partial\tilde{u}}{\partial\tilde{s}}
\quad\text{and} \quad
\frac{\partial(\gamma\tilde{s})}{\partial\tilde{s}} = \gamma \tag*{(157)} ∂ s ~ ∂ ( γ 2 u ~ ) = γ 2 ∂ s ~ ∂ u ~ and ∂ s ~ ∂ ( γ s ~ ) = γ (157)
According to the identities (157) , the temperature (156) in the inertial frame R \mathcal{R} R reduces to,
T = γ ∂ u ~ ∂ s ~ (158) T = \gamma\frac{\partial\tilde{u}}{\partial\tilde{s}} \tag*{(158)} T = γ ∂ s ~ ∂ u ~ (158)
The temperature (158) in the inertial frame R \mathcal{R} R is related to the temperature (154) in the local rest frame R ~ \tilde{\mathcal{R}} R ~ by,
T = γ T ~ (159) T = \gamma\tilde{T} \tag*{(159)} T = γ T ~ (159)
which implies that a relativistic fluid appears hotter when it is in relativistic motion. This is the conclusion reached by Einstein in a private letter to von Laue [13 ] and by Ott [11 , 23 ] in the famous controversy that opposed their results to the previous work performed by Einstein and Planck. [4 , 5 ] In fact, the initial conclusion of Einstein and Planck can be reached by defining the temperature T ′ T' T ′ in the inertial frame R \mathcal{R} R as,
T ′ = ∂ u ~ ∂ s ~ = ∂ u ~ ∂ ( γ s ~ ) = ∂ u ~ ∂ s ~ ( ∂ ( γ s ~ ) ∂ s ~ ) − 1 (160) T' = \frac{\partial\tilde{u}}{\partial\tilde{s}} = \frac{\partial\tilde{u}}{\partial(\gamma\tilde{s})} = \frac{\partial\tilde{u}}{\partial\tilde{s}} \left(\frac{\partial(\gamma\tilde{s})}{\partial\tilde{s}}\right)^{-1} \tag*{(160)} T ′ = ∂ s ~ ∂ u ~ = ∂ ( γ s ~ ) ∂ u ~ = ∂ s ~ ∂ u ~ ( ∂ s ~ ∂ ( γ s ~ ) ) − 1 (160)
According to identities (157) , the temperature (160) reduces to,
T ′ = 1 γ ∂ u ~ ∂ s ~ (161) T' = \frac{1}{\gamma} \frac{\partial\tilde{u}}{\partial\tilde{s}} \tag*{(161)} T ′ = γ 1 ∂ s ~ ∂ u ~ (161)
The temperature (161) in the inertial frame R \mathcal{R} R is related to the temperature (154) in the local rest frame R ~ \tilde{\mathcal{R}} R ~ by,
T ′ = 1 γ T ~ (162) T' = \frac{1}{\gamma} \tilde{T} \tag*{(162)} T ′ = γ 1 T ~ (162)
The relativistic transformation (162) for the temperature is flawed for the following reason : the internal energy density is not frame independent in a relativistic framework according relation (128) . Therefore, the temperature in the inertial frame R \mathcal{R} R has to be defined as (158) rather than (161) . The factor γ 2 \gamma^{2} γ 2 associated with the Lorentz transformation of the internal energy density is essential in deriving this result. In order to test the relativistic transformation law for the temperature, we can do the following “Gedankenexperiment”: a black body is moving at relativistic velocity with respect to the inertial frame R \mathcal{R} R of the observer. A photon of frequency ν ~ \widetilde{\nu} ν in the local rest frame R ~ \widetilde{\mathcal{R}} R of the black body is emitted orthogonally to the direction of motion of the black body. This photon is detected in the inertial frame R \mathcal{R} R of the observer with a frequency ν \nu ν . Since the photon is emitted orthogonally to the direction of motion of motion, the detected frequency ν \nu ν is related to the emitted frequency ν ~ \widetilde{\nu} ν through a relativistic transverse Doppler effect,
ν = γ ν ~ (163) \nu= \gamma\widetilde{\nu} \tag*{(163)} ν = γ ν (163)
In the local rest frame R ~ \widetilde{\mathcal{R}} R of the black body, the emitted spectral radiance of the black body is a function of the emitted frequency ν ~ \widetilde{\nu} ν of the photon and the temperature T ~ \widetilde{T} T given by Planck’s law,
B ~ ( ν ~ , T ~ ) = 2 h ν ~ 3 c 2 1 e h ν ~ k B T ~ − 1 (164) \widetilde{B}(\widetilde{\nu},\widetilde{T}) = \frac{2h\widetilde{\nu}^{3}}{c^{2}} \frac{1}{e^{\frac{h\widetilde{\nu}}{k_{B}\widetilde{T}}}-1} \tag*{(164)} B ( ν , T ) = c 2 2 h ν 3 e k B T h ν − 1 1 (164)
In the inertial frame R \mathcal{R} R of the observer, the detected spectral radiance of the black body is a function of the detected frequency ν \nu ν of the photon and the measured temperature T T T given by Planck’s law,
B ( ν , T ) = 2 h ν 3 c 2 1 e h ν k B T − 1 (165) B(\nu,T) = \frac{2h\nu^{3}}{c^{2}} \frac{1}{e^{\frac{h\nu}{k_{B}T}}-1} \tag*{(165)} B ( ν , T ) = c 2 2 h ν 3 e k B T h ν − 1 1 (165)
As shown by Johnson and Teller [24 ] , the ratio of the spectral radiance and the frequency cubed is frame invariant,
B ~ ( ν ~ , T ~ ) ν ~ 3 = B ( ν , T ) ν 3 (166) \frac{\widetilde{B}(\widetilde{\nu},\widetilde{T})}{\widetilde{\nu}^{3}} = \frac{B(\nu,T)}{\nu^{3}} \tag*{(166)} ν 3 B ( ν , T ) = ν 3 B ( ν , T ) (166)
which implies that,
h ν ~ k B T ~ = h ν k B T (167) \frac{h\widetilde{\nu}}{k_{B}\widetilde{T}} = \frac{h\nu}{k_{B}T} \tag*{(167)} k B T h ν = k B T h ν (167)
In view of the frame invariance (167) and the transverse Doppler effect (163), the temperature T T T of the black body measured in the inertial frame R \mathcal{R} R of the observer is related to the temperature of the black body T ~ \widetilde{T} T in the local rest frame R ~ \widetilde{\mathcal{R}} R of the black body by,
T T ~ = v v ~ = γ thus T = γ T ~ (168) \frac{T}{\widetilde{T}}=\frac{v}{\widetilde{v}}=\gamma\qquad\text{thus} \qquad T=\gamma\widetilde{T} \tag*{(168)} T T = v v = γ thus T = γ T (168)
in agreement with the prediction (159) as it should.
Relativistic chemical potential
We derive the relativistic transformation law for the chemical potential in this section. In the local rest frame R ~ \widetilde{\mathcal{R}} R , the chemical potential μ ~ \widetilde{\mu} μ is defined as,
μ ~ = ∂ u ~ ∂ n ~ (169) \widetilde{\mu}=\frac{\partial\widetilde{u}}{\partial\widetilde{n}} \tag*{(169)} μ = ∂ n ∂ u (169)
In the inertial frame R \mathcal{R} R , the chemical potential μ \mu μ is defined as,
μ = ∂ u ∂ n (170) \mu=\frac{\partial u}{\partial n} \tag*{(170)} μ = ∂ n ∂ u (170)
In view of the transformation laws for the internal energy density (128) and the number density density (84), the chemical potential in the inertial frame R \mathcal{R} R is recast as,
μ = ∂ ( γ 2 u ~ ) ∂ ( γ n ~ ) = ∂ ( γ 2 u ~ ) ∂ n ~ ( ∂ ( γ n ~ ) ∂ n ~ ) − 1 (171) \mu=\frac{\partial\left(\gamma^{2}\widetilde{u}\right)}{\partial\left(\gamma\widetilde{n}\right)}
=\frac{\partial\left(\gamma^{2}\widetilde{u}\right)}{\partial\widetilde{n}}
\left(\frac{\partial\left(\gamma\widetilde{n}\right)}{\partial\widetilde{n}}\right)^{-1} \tag*{(171)} μ = ∂ ( γ n ) ∂ ( γ 2 u ) = ∂ n ∂ ( γ 2 u ) ( ∂ n ∂ ( γ n ) ) − 1 (171)
Since the internal energy density u ~ \widetilde{u} u and the number density n ~ \widetilde{n} n in the local rest frame R ~ \widetilde{\mathcal{R}} R are independent of the Lorentz factor γ \gamma γ , we obtain the identities,
∂ ( γ 2 u ~ ) ∂ n ~ = γ 2 ∂ u ~ ∂ n ~ and ∂ ( γ n ~ ) ∂ n ~ = γ (172) \frac{\partial\left(\gamma^{2}\widetilde{u}\right)}{\partial\widetilde{n}}=\gamma^{2}\frac{\partial\widetilde{u}}{\partial\widetilde{n}}
\qquad\text{and} \qquad
\frac{\partial\left(\gamma\widetilde{n}\right)}{\partial\widetilde{n}}=\gamma \tag*{(172)} ∂ n ∂ ( γ 2 u ) = γ 2 ∂ n ∂ u and ∂ n ∂ ( γ n ) = γ (172)
According to the identities (172), the chemical potential (171) in the inertial frame R \mathcal{R} R reduces to,
μ = γ ∂ u ~ ∂ n ~ (173) \mu=\gamma\frac{\partial\widetilde{u}}{\partial\widetilde{n}} \tag*{(173)} μ = γ ∂ n ∂ u (173)
The chemical potential (173) in the inertial frame R \mathcal{R} R is related to the chemical potential (169) in the local rest frame R ~ \widetilde{\mathcal{R}} R by, [25 ]
μ = γ μ ~ (174) \mu=\gamma\widetilde{\mu} \tag*{(174)} μ = γ μ (174)
which implies that a relativistic fluid has a higher chemical potential when it is in relativistic motion.
Relativistic pressure
We use the relativistic transformation laws for the internal energy, the entropy, the number density, the temperature and the chemical potential in order to derive the relativistic transformation law for the pressure in this section. In order to do so, we begin with the Euler relation per unit volume in the local rest frame R ~ \widetilde{\mathcal{R}} R ,
u ~ = T ~ s ~ − p ~ − μ ~ n ~ (175) \widetilde{u}=\widetilde{T}\widetilde{s}-\widetilde{p}-\widetilde{\mu}\widetilde{n} \tag*{(175)} u = T s − p − μ n (175)
which implies that the pressure in the local rest frame R ~ \widetilde{\mathcal{R}} R is given by,
p ~ = T ~ s ~ − μ ~ n ~ − u ~ (176) \widetilde{p}=\widetilde{T}\widetilde{s}-\widetilde{\mu}\widetilde{n}-\widetilde{u} \tag*{(176)} p = T s − μ n − u (176)
The Euler relation per unit volume in the inertial frame R \mathcal{R} R reads,
u = T s − p − μ n (177) u=Ts-p-\mu n \tag*{(177)} u = T s − p − μ n (177)
which implies that the pressure in the inertial frame R \mathcal{R} R is given by,
p = T s − μ n − u (178) p=Ts-\mu n-u \tag*{(178)} p = T s − μ n − u (178)
In view of the transformation laws for the entropy density (55) , the number density (84) , the internal energy density (128) , the temperature (159) and the chemical potential (174) , the pressure (178) in the inertial frame R \mathcal{R} R is recast as,
p = ( γ T ~ ) ( γ s ~ ) − ( γ μ ~ ) ( γ n ~ ) − ( γ 2 u ~ ) (179) p=(\gamma\widetilde{T})(\gamma\widetilde{s})-(\gamma\widetilde{\mu})(\gamma\widetilde{n})-(\gamma^2\widetilde{u}) \tag*{(179)} p = ( γ T ) ( γ s ) − ( γ μ ) ( γ n ) − ( γ 2 u ) (179)
Thus, in agreement with Sutcliffe [17 ] , the pressure (179) in the inertial frame R \mathcal{R} R is related to the pressure (176) in the local rest frame R ~ \widetilde{\mathcal{R}} R by,
p = γ 2 p ~ (180) p=\gamma^2\widetilde{p} \tag*{(180)} p = γ 2 p (180)
which implies that a relativistic fluid has a larger pressure when it is in relativistic motion. The thermodynamic pressure that satisfies the transformation law (180) has to be distinguished from the mechanical pressure that is invariant under a relativistic frame transformation, as noted by Einstein and Planck [4 , 5 ] . As mentioned by Sutcliffe : “The two pressures are the same only in the local rest frame R ~ \widetilde{\mathcal{R}} R . This distinction is necessary in order for relativistic thermodynamics to be consistent with relativistic mechanics.” [17 ] In the non-relativistic limit, the thermodynamic and mechanical pressures coincide. As an illustration, the equation of state of an ideal gas in the local rest frame R ~ \widetilde{\mathcal{R}} R reads, [26 ]
p ~ = n ~ R T ~ (ideal gas) (181) \widetilde{p}=\widetilde{n}R\widetilde{T}\qquad\text{(ideal gas)} \tag*{(181)} p = n R T (ideal gas) (181)
where R R R is the ideal gas constant. Multiplying the equation of state (181) in the local rest frame R ~ \widetilde{\mathcal{R}} R by γ 2 \gamma^2 γ 2 yields,
( γ 2 p ~ ) = ( γ n ~ ) R ( γ T ~ ) (182) \left(\gamma^2\widetilde{p}\right)=\left(\gamma\widetilde{n}\right)R\left(\gamma\widetilde{T}\right) \tag*{(182)} ( γ 2 p ) = ( γ n ) R ( γ T ) (182)
Using the transformation laws for the number density (84) , the temperature (159) and the pressure (180) , equation (182) is identified as the equation of state of an ideal gas in the inertial frame R \mathcal{R} R , [27 ]
p = n R T (ideal gas) (183) p=nRT\qquad\text{(ideal gas)} \tag*{(183)} p = n R T (ideal gas) (183)
as expected. In order to test the relativistic transformation law for the pressure, we can do the following “Gedankenexperiment”: a black body is moving at relativistic velocity with respect to the inertial frame R \mathcal{R} R of the observer. A monochromatic filter selects the photons of frequency ν ~ \widetilde{\nu} ν and density n ~ \widetilde{n} n in the local rest frame R ~ \widetilde{\mathcal{R}} R of the black body that are emitted orthogonally to the direction of motion of the black body. The photons are detected in the inertial frame R \mathcal{R} R of the observer with a frequency ν \nu ν and a density n n n . Since the photons are emitted orthogonally to the direction of motion of motion, the detected frequency ν \nu ν is related to the emitted frequency ν ~ \widetilde{\nu} ν through the relativistic transverse Doppler effect (163) . The internal energy density u ~ \widetilde{u} u of the emitted monochromatic photons in the local rest frame R ~ \widetilde{\mathcal{R}} R of the black body is given by the Planck-Einstein relation,
u ~ = n ~ ℏ ν ~ (184) \widetilde{u}=\widetilde{n}\hbar\widetilde{\nu} \tag*{(184)} u = n ℏ ν (184)
The internal energy density u u u of the detected monochromatic photons in the local rest frame R \mathcal{R} R of the observer is given by the Planck-Einstein relation,
u = n ℏ ν (185) u=n\hbar\nu \tag*{(185)} u = n ℏ ν (185)
The radiation pressure p ~ \tilde{p} p ~ of the emitted photons in the local rest frame R ~ \widetilde{\mathcal R} R of the black body reads,
p ~ = 1 3 u ~ = 1 3 n ~ ℏ ν ~ (186) \tilde{p} = \frac{1}{3}\tilde{u} = \frac{1}{3}\tilde{n}\hbar\tilde{\nu} \tag*{(186)} p ~ = 3 1 u ~ = 3 1 n ~ ℏ ν ~ (186)
The radiation pressure p p p of the detected photons in the inertial frame R \mathcal R R of the observer reads,
p = 1 3 u = 1 3 n ℏ ν (187) p = \frac{1}{3}u = \frac{1}{3}n\hbar\nu \tag*{(187)} p = 3 1 u = 3 1 n ℏ ν (187)
In view of the transformation law (84) for the number density and the relativistic transverse Doppler effect (163) , the radiation pressure p p p of the detected photons (187) in the inertial frame R \mathcal R R of the observer is related to the radiation pressure p ~ \tilde{p} p ~ of the emitted photons (186) in the local rest frame R ~ \widetilde{\mathcal R} R of the black body by the transformation law,
p = 1 3 n ℏ ν = 1 3 ( γ n ~ ) ℏ ( γ ν ~ ) = γ 2 p ~ (188) p = \frac{1}{3}n\hbar\nu= \frac{1}{3}(\gamma\tilde{n})\hbar(\gamma\tilde{\nu}) = \gamma^2\tilde{p} \tag*{(188)} p = 3 1 n ℏ ν = 3 1 ( γ n ~ ) ℏ ( γ ν ~ ) = γ 2 p ~ (188)
in agreement with the prediction (180) as it should.
Relativistic force density
We determine the properties of the contravariant components of the force density vector in this section. According to the relativistic equation of motion (B.17) of a particle, the contravariant components of the force F μ F^\mu F μ exerted on an infinitesimal fluid element are defined as the derivative of the contravariant components of the momentum P μ P^\mu P μ with respect to the proper time τ ~ \tilde{\tau} τ ~ in the local rest frame R ~ \widetilde{\mathcal R} R . Thus, the contravariant components of the force vector F μ F^\mu F μ exerted on an infinitesimal fluid element are obtained by multiplying the contravariant components of the force density vector f μ f^\mu f μ by the infinitesimal volume d V ~ 3 d\widetilde{V}_3 d V 3 in the local rest frame R ~ \widetilde{\mathcal R} R , [28 ]
F μ = f μ d V ~ 3 (189) F^\mu= f^\mu d\widetilde{V}_3 \tag*{(189)} F μ = f μ d V 3 (189)
In view of the contravariant components of the force vector (189) , the contravariant time component of the force vector F 0 F^0 F 0 is related to the contravariant time component of the force density vector f 0 f^0 f 0 by,
F 0 = f 0 d V ~ 3 (190) F^0 = f^0 d\widetilde{V}_3 \tag*{(190)} F 0 = f 0 d V 3 (190)
and the contravariant spatial component of the force vector F j \mathbf{F}^{j} F j is related to the contravariant spatial component of the force density vector f j \mathbf{f}^{j} f j by,
F j = f j d V ~ 3 (191) \mathbf{F}^{j}=\mathbf{f}^{j}\,\mathrm{d}\widetilde{V}_{3} \tag*{(191)} F j = f j d V 3 (191)
The spatial force density vector f \mathbf{f} f in the inertial frame R \mathcal{R} R is defined as,
f = f j e j (192) \mathbf{f}=\mathbf{f}^{j}\mathbf{e}_{j} \tag*{(192)} f = f j e j (192)
In view of the volume contraction (A.43), the spatial force vector (B.21), the spatial contravariant component of the force vector (191) and the spatial force density vector (192) , the force vector is expressed as,
F = 1 γ F j e j = 1 γ f j e j d V ~ 3 = f j e j d V 3 = f d V 3 (193) \mathbf{F}=\frac{1}{\gamma}\mathbf{F}^{j}\mathbf{e}_{j}=\frac{1}{\gamma}\mathbf{f}^{j}\mathbf{e}_{j}\,\mathrm{d}\widetilde{V}_{3}=\mathbf{f}^{j}\mathbf{e}_{j}\,\mathrm{d}V_{3}=\mathbf{f}\,\mathrm{d}V_{3} \tag*{(193)} F = γ 1 F j e j = γ 1 f j e j d V 3 = f j e j d V 3 = f d V 3 (193)
According to the contravariant time component of the force vector (190) and the spatial force vector (193) , the contravariant time component of the force vector (B.25) is recast as,
f 0 d V ~ 3 = γ c f ⋅ v d V 3 (194) f^{0}\,\mathrm{d}\widetilde{V}_{3}=\frac{\gamma}{c}\mathbf{f}\cdot\mathbf{v}\,\mathrm{d}V_{3} \tag*{(194)} f 0 d V 3 = c γ f ⋅ v d V 3 (194)
In view of the volume contraction (A.43) and the contravariant time component of the force vector (194) , the power density in the inertial frame R \mathcal{R} R is given by, [28 ]
f 0 = 1 c f ⋅ v (195) f^{0}=\frac{1}{c}\mathbf{f}\cdot\mathbf{v} \tag*{(195)} f 0 = c 1 f ⋅ v (195)
Relativistic energy continuity equation
We determine the relativistic energy continuity equation by taking the contravariant time components of the energy-momentum continuity equation (72) in this section. In view of the stress-energy-momentum tensor of a relativistic perfect fluid (153) , the relativistic energy-momentum continuity equation (72) of a relativistic perfect fluid becomes, [29 ]
∂ μ ( ( e ~ + p ~ c 2 ) u μ u ν + p ~ η μ ν ) = f ν (196) \partial_{\mu}\left(\left(\frac{\widetilde{e}+\widetilde{p}}{c^{2}}\right)u^{\mu}u^{\nu}+\widetilde{p}\,\eta^{\mu\nu}\right)=f^{\nu} \tag*{(196)} ∂ μ ( ( c 2 e + p ) u μ u ν + p η μν ) = f ν (196)
In view of the Minkowski metric (A.6), the contravariant components of the velocity vector (A.21) and the contravariant time component of the force density (195), the component ν = 0 \nu=0 ν = 0 of the relativistic energy-momentum continuity equation (196) yields the relativistic energy continuity equation,
∂ μ ( γ e ~ + p ~ c u μ − p ~ δ μ 0 ) = 1 c f ⋅ v (197) \partial_{\mu}\left(\gamma\frac{\tilde{e}+\tilde{p}}{c}u^{\mu}-\tilde{p}\,\delta^{\mu0}\right)=\frac{1}{c}\boldsymbol{f}\cdot\boldsymbol{v} \tag*{(197)} ∂ μ ( γ c e ~ + p ~ u μ − p ~ δ μ 0 ) = c 1 f ⋅ v (197)
Using the frame duality (A.8), the contravariant temporal and spatial components of the velocity vector (A.21) and (A.24), as well as the partial time derivative (35), the relativistic energy continuity equation (197) is recast as,
1 c ∂ t ( γ 2 e ~ + p ~ ( γ 2 − 1 ) ) + 1 c ( e i ∂ i ) ⋅ ( γ 2 ( e ~ + p ~ ) v i e i ) = 1 c f ⋅ v (198) \frac{1}{c}\partial_t\left(\gamma^2\tilde{e}+\tilde{p}\left(\gamma^2-1\right)\right)+\frac{1}{c}\left(\boldsymbol{e}^{i}\partial_i\right)\cdot\left(\gamma^2(\tilde{e}+\tilde{p})v^{i}\boldsymbol{e}_i\right)=\frac{1}{c}\boldsymbol{f}\cdot\boldsymbol{v} \tag*{(198)} c 1 ∂ t ( γ 2 e ~ + p ~ ( γ 2 − 1 ) ) + c 1 ( e i ∂ i ) ⋅ ( γ 2 ( e ~ + p ~ ) v i e i ) = c 1 f ⋅ v (198)
In view of the spatial velocity vector (A.23), the gradient (36), the transformation laws for the energy density (103) and the pressure (180), the relativistic energy continuity equation (198) multiplied by c c c becomes,
∂ t ( e + ( 1 − 1 γ 2 ) p ) + ∇ ⋅ ( ( e + p ) v ) = f ⋅ v (199) \partial_t\left(e+\left(1-\frac{1}{\gamma^2}\right)p\right)+\boldsymbol{\nabla}\cdot\left((e+p)\boldsymbol{v}\right)=\boldsymbol{f}\cdot\boldsymbol{v} \tag*{(199)} ∂ t ( e + ( 1 − γ 2 1 ) p ) + ∇ ⋅ ( ( e + p ) v ) = f ⋅ v (199)
According to the Lorentz factor (A.26), the relativistic energy continuity equation (199) in the inertial frame R \mathcal{R} R is recast as,
∂ t ( e + v 2 c 2 p ) + ∇ ⋅ ( e v ) = ( f − ∇ p ) ⋅ v (relativistic) (200) \partial_t\left(e+\frac{v^2}{c^2}p\right)+\boldsymbol{\nabla}\cdot(e\boldsymbol{v})=(\boldsymbol{f}-\boldsymbol{\nabla}p)\cdot\boldsymbol{v}\quad\text{(relativistic)} \tag*{(200)} ∂ t ( e + c 2 v 2 p ) + ∇ ⋅ ( e v ) = ( f − ∇ p ) ⋅ v (relativistic) (200)
where the term,
v 2 c 2 p = ( 1 − 1 γ 2 ) p = p − p ~ (201) \frac{v^2}{c^2}p=\left(1-\frac{1}{\gamma^2}\right)p=p-\tilde{p} \tag*{(201)} c 2 v 2 p = ( 1 − γ 2 1 ) p = p − p ~ (201)
is a relativistic correction of the energy density due to the pressure difference between the inertial frame R \mathcal{R} R and the local rest frame R ~ \widetilde{\mathcal{R}} R which leads to a relativistic coupling between heat and work. In the non-relativistic limit, the energy continuity equation (200) reduces to, [19 ]
∂ t e + ∇ ⋅ ( e v ) = ( f − ∇ p ) ⋅ v (non-relativistic) (202) \partial_t e+\boldsymbol{\nabla}\cdot(e\boldsymbol{v})=(\boldsymbol{f}-\boldsymbol{\nabla}p)\cdot\boldsymbol{v}\quad\text{(non-relativistic)} \tag*{(202)} ∂ t e + ∇ ⋅ ( e v ) = ( f − ∇ p ) ⋅ v (non-relativistic) (202)
as expected.
Relativistic momentum continuity equation
We determine the relativistic momentum continuity equation by multiplying the contravariant spatial components of the energy-momentum continuity equation (72) with the corresponding spatial orthonormal basis vectors in this section. In view of the contravariant spatial components of the velocity vector (A.24), the component ν = i \nu=i ν = i of the relativistic energy-momentum continuity equation (196) yields the relativistic momentum continuity equation, [29 ]
∂ μ ( γ ( e ~ + p ~ c 2 ) u μ v i − p ~ δ μ i ) = f i (203) \partial_{\mu}\left(\gamma\left(\frac{\tilde{e}+\tilde{p}}{c^{2}}\right)u^{\mu}v^{i}-\tilde{p}\,\delta^{\mu i}\right)=f^{i} \tag*{(203)} ∂ μ ( γ ( c 2 e ~ + p ~ ) u μ v i − p ~ δ μ i ) = f i (203)
Using the Minkowski metric (A.6), the contravariant temporal and spatial components of the velocity vector (A.21) and (A.24), and the partial time derivative (35), the relativistic momentum continuity equation (203) is recast as,
∂ t ( γ 2 ( e ~ + p ~ c 2 ) v i ) + ∂ j ( γ 2 ( e ~ + p ~ c 2 ) v i v j ) + ∂ i p ~ = f i (204) \partial_{t}\left(\gamma^{2}\left(\frac{\tilde{e}+\tilde{p}}{c^{2}}\right)v^{i}\right)+\partial_{j}\left(\gamma^{2}\left(\frac{\tilde{e}+\tilde{p}}{c^{2}}\right)v^{i}v^{j}\right)+\partial^{i}\tilde{p}=f^{i} \tag*{(204)} ∂ t ( γ 2 ( c 2 e ~ + p ~ ) v i ) + ∂ j ( γ 2 ( c 2 e ~ + p ~ ) v i v j ) + ∂ i p ~ = f i (204)
According to the frame duality (A.8), the contraction of the relativistic momentum continuity equation (204) with the spatial orthonormal basis vectors e i e_i e i is written as,
∂ t ( γ 2 ( e ~ + p ~ c 2 ) v i e i ) + ( e k ∂ k ) ⋅ ( γ 2 ( e ~ + p ~ c 2 ) v i v j e i ⊗ e j ) + ( e i ∂ i ) p ~ = f i e i (205) \begin{aligned}
\partial_{t}\left(\gamma^{2}\left(\frac{\tilde{e}+\tilde{p}}{c^{2}}\right)v^{i}e_i\right)+\left(e^{k}\partial_{k}\right)\cdot\left(\gamma^{2}\left(\frac{\tilde{e}+\tilde{p}}{c^{2}}\right)v^{i}v^{j}e_i\otimes e_j\right) \\
&+\left(e^{i}\partial_{i}\right)\tilde{p}=f^{i}e_i \tag*{(205)}
\end{aligned} ∂ t ( γ 2 ( c 2 e ~ + p ~ ) v i e i ) + ( e k ∂ k ) ⋅ ( γ 2 ( c 2 e ~ + p ~ ) v i v j e i ⊗ e j ) + ( e i ∂ i ) p ~ = f i e i (205)
since e i ∂ i = e i ∂ i e^{i}\partial_{i}=e_i\partial^{i} e i ∂ i = e i ∂ i . In view of the frame duality (A.8), the spatial velocity vector (A.23), the gradient (36), the transformation laws for the energy density (103) and the pressure (180), and the force density spatial vector (193), the relativistic momentum continuity equation (205) is recast as,
∂ t ( ( e + p c 2 ) v ) + ∇ ⋅ ( ( e + p c 2 ) v ⊗ v ) + ∇ p ~ = f (206) \partial_{t}\left(\left(\frac{e+p}{c^{2}}\right)\mathbf{v}\right)+\nabla\cdot\left(\left(\frac{e+p}{c^{2}}\right)\mathbf{v}\otimes\mathbf{v}\right)+\nabla\tilde{p}=\mathbf{f} \tag*{(206)} ∂ t ( ( c 2 e + p ) v ) + ∇ ⋅ ( ( c 2 e + p ) v ⊗ v ) + ∇ p ~ = f (206)
Using the decomposition (126) of the energy density e e e , and the enthalpy density h h h in the inertial frame R \mathcal{R} R ,
h = u + p (207) h=u+p \tag*{(207)} h = u + p (207)
the relativistic momentum continuity equation (206) becomes,
∂ t ( ( ρ + h c 2 ) v ) + ∇ ⋅ ( ( ρ + h c 2 ) v ⊗ v ) + ∇ p ~ = f (208) \partial_t\left(\left(\rho+\frac{h}{c^2}\right)\mathbf{v}\right)+\nabla\cdot\left(\left(\rho+\frac{h}{c^2}\right)\mathbf{v}\otimes\mathbf{v}\right)+\nabla\tilde{p}=\mathbf{f} \tag*{(208)} ∂ t ( ( ρ + c 2 h ) v ) + ∇ ⋅ ( ( ρ + c 2 h ) v ⊗ v ) + ∇ p ~ = f (208)
The symmetric spatial stress tensor in the inertial frame R \mathcal{R} R is defined as,
τ = − ( ρ + h c 2 ) ( v ⊗ v ) − p ~ 1 3 (209) \boldsymbol{\tau}=-\left(\rho+\frac{h}{c^2}\right)(\mathbf{v}\otimes\mathbf{v})-\tilde{p}\mathbf{1}_3 \tag*{(209)} τ = − ( ρ + c 2 h ) ( v ⊗ v ) − p ~ 1 3 (209)
where 1 3 \mathbf{1}_3 1 3 is the spatial identity rank-2 tensor. In view of the momentum density (106) , the spatial stress tensor (209) in the inertial frame R \mathcal{R} R becomes,
τ = − p ⊗ v − p ~ 1 3 − h ( v c ⊗ v c ) (relativistic) (210) \boldsymbol{\tau}=-\mathbf{p}\otimes\mathbf{v}-\tilde{p}\mathbf{1}_3-h\left(\frac{\mathbf{v}}{c}\otimes\frac{\mathbf{v}}{c}\right)\qquad\text{(relativistic)} \tag*{(210)} τ = − p ⊗ v − p ~ 1 3 − h ( c v ⊗ c v ) (relativistic) (210)
where the last term is a relativistic correction of the spatial stress tensor which leads to a relativistic coupling between heat and work. In view of the momentum density vector (106) and the spatial stress tensor (210) , the relativistic momentum continuity equation (208) in the inertial frame R \mathcal{R} R is recast as,
∂ t ( p + h c 2 v ) − ∇ ⋅ τ = f (relativistic) (211) \partial_t\left(\mathbf{p}+\frac{h}{c^2}\mathbf{v}\right)-\nabla\cdot\boldsymbol{\tau}=\mathbf{f}\qquad\text{(relativistic)} \tag*{(211)} ∂ t ( p + c 2 h v ) − ∇ ⋅ τ = f (relativistic) (211)
where the relativistic correction to the momentum density p \mathbf{p} p also leads to a relativistic coupling between heat and work. In the non-relativistic limit, the momentum continuity equation (211) reduces to, [19 ]
∂ t p − ∇ ⋅ τ = f (non-relativistic) (212) \partial_t\mathbf{p}-\nabla\cdot\boldsymbol{\tau}=\mathbf{f}\qquad\text{(non-relativistic)} \tag*{(212)} ∂ t p − ∇ ⋅ τ = f (non-relativistic) (212)
and the spatial stress tensor (210) reduces to,
τ = − p ⊗ v − p 1 3 (non-relativistic) (213) \boldsymbol{\tau}=-\mathbf{p}\otimes\mathbf{v}-p\mathbf{1}_3\qquad\text{(non-relativistic)} \tag*{(213)} τ = − p ⊗ v − p 1 3 (non-relativistic) (213)
as expected.
Relativistic inertia continuity equation
The entropy (51) and the number of moles of substance (82) are frame independent scalar quantities. By contrast, the mass (B.10) is a frame dependent quantity. Thus, the mass continuity equation cannot be established by following the same theoretical approach as for the entropy in § 3 and for the quantity of substance in § 6. In fact, since the enthalpy density h h h of the fluid carries inertia, the matter continuity equation is established in this section for the relativistic inertia density m = ρ + h / c 2 m = \rho+ h/c^2 m = ρ + h / c 2 . This is achieved by projecting the relativistic energy-momentum continuity equation (72) along the worldline of the local element of fluid defined by the velocity in the inertial frame R \mathcal{R} R . The contraction of the relativistic energy-momentum continuity equation (72) with the velocity vector reads,
( ∂ μ T μ ν ) u ν = f ν u ν (214) (\partial_\mu T^{\mu\nu})u_\nu= f^\nu u_\nu \tag*{(214)} ( ∂ μ T μν ) u ν = f ν u ν (214)
In view of the orthogonality condition (B.19) between the force and velocity vectors and the contravariant components of the force vector (189), we obtain the identity,
F ν u ν = ( f ν u ν ) d V ~ 3 = 0 (215) F^\nu u_\nu= (f^\nu u_\nu)\,d\widetilde{V}_3 = 0 \tag*{(215)} F ν u ν = ( f ν u ν ) d V 3 = 0 (215)
According to the identity (215), the contracted relativistic energy-momentum continuity equation (214) reduces to,
u ν ( ∂ μ T μ ν ) = 0 (216) u_\nu(\partial_\mu T^{\mu\nu}) = 0 \tag*{(216)} u ν ( ∂ μ T μν ) = 0 (216)
In view of the contravariant components of the stress energy momentum tensor (152), the contracted relativistic energy-momentum continuity equation (216) becomes,
u ν ∂ μ ( m ~ u μ u ν ) + u ν ∂ μ ( η μ ν p ~ ) = 0 (217) u_\nu\partial_\mu(\widetilde{m}\,u^\mu u^\nu) + u_\nu\partial_\mu(\eta^{\mu\nu}\,\widetilde{p}) = 0 \tag*{(217)} u ν ∂ μ ( m u μ u ν ) + u ν ∂ μ ( η μν p ) = 0 (217)
In view of the contraction of the velocities (A.17), the relativistic contracted energy-momentum continuity equation (217) is recast as,
− c 2 ∂ μ ( m ~ u μ ) + m ~ u μ ( u ν ∂ μ u ν ) + u μ ∂ μ p ~ = 0 (218) -c^2\partial_\mu(\widetilde{m}\,u^\mu) + \widetilde{m}\,u^\mu(u_\nu\partial_\mu u^\nu) + u^\mu\partial_\mu\widetilde{p} = 0 \tag*{(218)} − c 2 ∂ μ ( m u μ ) + m u μ ( u ν ∂ μ u ν ) + u μ ∂ μ p = 0 (218)
According to the contraction of the velocities (A.17), we obtain the identity,
∂ μ ( u ν u ν ) = − ∂ μ ( c 2 ) = 2 u ν ∂ μ u ν = 0 (219) \partial_\mu(u_\nu u^\nu) = -\partial_\mu(c^2) = 2u_\nu\partial_\mu u^\nu= 0 \tag*{(219)} ∂ μ ( u ν u ν ) = − ∂ μ ( c 2 ) = 2 u ν ∂ μ u ν = 0 (219)
In view of the identity (219), the contracted relativistic energy-momentum continuity equation (218) yields the relativistic inertia continuity equation,
∂ μ ( m ~ u μ ) = u μ ∂ μ ( p ~ c 2 ) (220) \partial_\mu\left(\tilde{m}u^\mu\right)=u^\mu\partial_\mu\left(\frac{\tilde{p}}{c^2}\right) \tag*{(220)} ∂ μ ( m ~ u μ ) = u μ ∂ μ ( c 2 p ~ ) (220)
According to the inertia density (133) and the enthalpy density in the local rest frame R ~ \tilde{\mathcal{R}} R ~ ,
h ~ = u ~ + p ~ (221) \tilde{h}=\tilde{u}+\tilde{p} \tag*{(221)} h ~ = u ~ + p ~ (221)
the relativistic inertia continuity equation (220) is recast as,
∂ μ ( ρ ~ u μ ) = − ∂ μ ( h ~ c 2 u μ ) + u μ ∂ μ ( p ~ c 2 ) (relativistic) (222) \partial_\mu\left(\tilde{\rho}u^\mu\right)=-\partial_\mu\left(\frac{\tilde{h}}{c^2}u^\mu\right)+u^\mu\partial_\mu\left(\frac{\tilde{p}}{c^2}\right)\qquad\text{(relativistic)} \tag*{(222)} ∂ μ ( ρ ~ u μ ) = − ∂ μ ( c 2 h ~ u μ ) + u μ ∂ μ ( c 2 p ~ ) (relativistic) (222)
The relativistic inertia continuity equation (222) differs from the relativistic mass continuity equation, i.e. ∂ μ ( ρ ~ u μ ) = 0 \partial_\mu\left(\tilde{\rho}u^\mu\right)=0 ∂ μ ( ρ ~ u μ ) = 0 , that is usually assumed to hold in textbooks on relativity. [22 , 20 ] As shown here, such an assumption is not warranted since the right hand side of the relativistic inertia continuity equation (222) clearly does not vanish. Since the Lorentz factor γ \gamma γ is a first-order relativistic correction factor, the derivative of the Lorentz factor ∂ μ γ \partial_\mu\gamma ∂ μ γ is a second-order relativistic correction to the dynamics that can be neglected to first-order. Thus, in view of the relativistic transformation law for the pressure (180), the relativistic inertia continuity equation (222) multiplied by γ \gamma γ can be recast as,
∂ μ ( γ m ~ u μ ) = γ u μ ∂ μ ( p ~ c 2 ) = 1 γ u μ ∂ μ ( p c 2 ) (223) \partial_\mu\left(\gamma\tilde{m}u^\mu\right)=\gamma u^\mu\partial_\mu\left(\frac{\tilde{p}}{c^2}\right)=\frac{1}{\gamma}u^\mu\partial_\mu\left(\frac{p}{c^2}\right) \tag*{(223)} ∂ μ ( γ m ~ u μ ) = γ u μ ∂ μ ( c 2 p ~ ) = γ 1 u μ ∂ μ ( c 2 p ) (223)
In view of the frame duality (A.8), the contravariant time component of the velocity vector (A.21), the spatial velocity vector (A.23), the contravariant spatial components of the velocity vector (A.24), the partial time derivative (35) and the gradient (36), we obtain the identities,
∂ μ ( γ m ~ u μ ) = ∂ t ( γ 2 m ~ ) + ( ∂ j e j ) ⋅ ( γ 2 m ~ v k e k ) = ∂ t m + ∇ ⋅ ( m v ) (224) \partial_\mu\left(\gamma\tilde{m}u^\mu\right)=\partial_t\left(\gamma^2\tilde{m}\right)+\left(\partial_j e^j\right)\cdot\left(\gamma^2\tilde{m}v^k e_k\right)=\partial_t m+\nabla\cdot(m\boldsymbol{v}) \tag*{(224)} ∂ μ ( γ m ~ u μ ) = ∂ t ( γ 2 m ~ ) + ( ∂ j e j ) ⋅ ( γ 2 m ~ v k e k ) = ∂ t m + ∇ ⋅ ( m v ) (224)
u μ ∂ μ ( p c 2 ) = γ ( ∂ t + v ⋅ ∇ ) p c 2 (225) u^\mu\partial_\mu\left(\frac{p}{c^2}\right)=\gamma\left(\partial_t+\boldsymbol{v}\cdot\nabla\right)\frac{p}{c^2} \tag*{(225)} u μ ∂ μ ( c 2 p ) = γ ( ∂ t + v ⋅ ∇ ) c 2 p (225)
The total time derivative of the pressure in the inertial frame R \mathcal{R} R is defined as,
p ˙ = ( ∂ t + v ⋅ ∇ ) p (226) \dot{p}=\left(\partial_t+\boldsymbol{v}\cdot\nabla\right)p \tag*{(226)} p ˙ = ( ∂ t + v ⋅ ∇ ) p (226)
In view of the identities (224) and (225) , and the total time derivative of the pressure (226) , the relativistic inertia continuity equation (223) in the inertial frame R ~ \widetilde{\mathcal{R}} R is recast as,
∂ t m + ∇ ⋅ ( m v ) = p ˙ c 2 (relativistic) (227) \partial_t m + \nabla\cdot(m\mathbf{v}) = \frac{\dot{p}}{c^2}\qquad\text{(relativistic)} \tag*{(227)} ∂ t m + ∇ ⋅ ( m v ) = c 2 p ˙ (relativistic) (227)
In the non-relativistic limit, the inertia continuity equation (227) reduces to the mass continuity equation, [19 ]
∂ t ρ + ∇ ⋅ ( ρ v ) = 0 (non-relativistic) (228) \partial_t \rho+ \nabla\cdot(\rho\mathbf{v}) = 0\qquad\text{(non-relativistic)} \tag*{(228)} ∂ t ρ + ∇ ⋅ ( ρ v ) = 0 (non-relativistic) (228)
as expected.
Relativistic Euler equation
We deduce the relativistic Euler equation from the relativistic momentum continuity equation using the inertia density continuity equation in this section. In view of the inertia density (133) in the local rest frame R ~ \widetilde{\mathcal{R}} R , the inertia density in the inertial frame R \mathcal{R} R is given by,
m = e + p c 2 (229) m = \frac{e+p}{c^2} \tag*{(229)} m = c 2 e + p (229)
Using the transformation laws for the energy density (103) and the pressure (180) , the inertia density (229) is recast as,
m = γ 2 ( e ~ + p ~ c 2 ) (230) m = \gamma^2\left(\frac{\widetilde{e}+\widetilde{p}}{c^2}\right) \tag*{(230)} m = γ 2 ( c 2 e + p ) (230)
The inertia density (230) in the inertial frame R \mathcal{R} R is related to the inertia density (133) in the local rest frame R ~ \widetilde{\mathcal{R}} R by,
m = γ 2 m ~ (231) m = \gamma^2\widetilde{m} \tag*{(231)} m = γ 2 m (231)
In view of the inertia density (229) , the relativistic momentum continuity equation (206) reduces to,
∂ t ( m v ) + ∇ ⋅ ( m v ⊗ v ) + ∇ p ~ = f (232) \partial_t(m\mathbf{v})+\nabla\cdot(m\mathbf{v}\otimes\mathbf{v})+\nabla\widetilde{p}=\mathbf{f} \tag*{(232)} ∂ t ( m v ) + ∇ ⋅ ( m v ⊗ v ) + ∇ p = f (232)
The relativistic momentum continuity equation (232) is expanded as,
m ( ∂ t v + v ⋅ ∇ v ) + ( ∂ t m + ∇ ⋅ ( m v ) ) v = f − ∇ p ~ (233) m\left(\partial_t \boldsymbol{v}+\boldsymbol{v}\cdot\nabla\boldsymbol{v}\right)+\left(\partial_t m+\nabla\cdot\left(m\boldsymbol{v}\right)\right)\boldsymbol{v}=\boldsymbol{f}-\nabla\tilde{p} \tag*{(233)} m ( ∂ t v + v ⋅ ∇ v ) + ( ∂ t m + ∇ ⋅ ( m v ) ) v = f − ∇ p ~ (233)
The spatial acceleration vector of a fluid is written as,
a = v ˙ = ( ∂ t + v ⋅ ∇ ) v (234) \boldsymbol{a}=\dot{\boldsymbol{v}}=\left(\partial_t+\boldsymbol{v}\cdot\nabla\right)\boldsymbol{v} \tag*{(234)} a = v ˙ = ( ∂ t + v ⋅ ∇ ) v (234)
According to the spatial acceleration vector (234), the relativistic momentum continuity equation (233) yields the relativistic Euler equation in the inertial frame R \mathcal{R} R ,
m a + ( ∂ t m + ∇ ⋅ ( m v ) ) v = f − ∇ p ~ (235) m\boldsymbol{a}+\left(\partial_t m+\nabla\cdot\left(m\boldsymbol{v}\right)\right)\boldsymbol{v}=\boldsymbol{f}-\nabla\tilde{p} \tag*{(235)} m a + ( ∂ t m + ∇ ⋅ ( m v ) ) v = f − ∇ p ~ (235)
In view of the inertia density (133), the enthalpy density (207) and the inertia continuity equation (227), the relativistic Euler equation (235) in the inertial frame R \mathcal{R} R is recast as,
( ρ + h c 2 ) a = f − ∇ p ~ − p ˙ c 2 v (relativistic) (236) \left(\rho+\frac{h}{c^2}\right)\boldsymbol{a}=\boldsymbol{f}-\nabla\tilde{p}-\frac{\dot{p}}{c^2}\boldsymbol{v}\qquad\text{(relativistic)} \tag*{(236)} ( ρ + c 2 h ) a = f − ∇ p ~ − c 2 p ˙ v (relativistic) (236)
where the relativistic corrections are the terms that are functions of the enthalpy density h h h , which leads to a relativistic coupling between heat and work, or of the time derivative of the pressure p ˙ \dot{p} p ˙ . In the non-relativistic limit, the Euler equation (236) reduces to, [19 ]
ρ a = f − ∇ p (non-relativistic) (237) \rho\boldsymbol{a}=\boldsymbol{f}-\nabla p\qquad\text{(non-relativistic)} \tag*{(237)} ρ a = f − ∇ p (non-relativistic) (237)
as expected.
Conclusion
Using the covariant approach based on the divergence theorem in Minkowski space outlined by Stü̈ckelberg [18 ] , we derived the relativistic continuity equations for Lorentz invariant scalar functions such as the entropy S S S and the number of moles of substance N N N , and for the vector components P ν P^\nu P ν of the energy momentum vector. This approach lead naturally to a relativistic formulation of the first law (67) and second law (52) of thermodynamics,
Δ P i → f ν = F ν Δ t i → f and Δ S i → f = Σ S i → f ≥ 0 (238) \Delta P^\nu_{i\to f}=F^\nu\Delta t_{i\to f}\qquad\text{and}\qquad\Delta S_{i\to f}=\Sigma_{S_{i\to f}}\geq0 \tag*{(238)} Δ P i → f ν = F ν Δ t i → f and Δ S i → f = Σ S i → f ≥ 0 (238)
where the variation of the state functions Δ P i → f ν \Delta P^\nu_{i\to f} Δ P i → f ν and Δ S i → f \Delta S_{i\to f} Δ S i → f are due to the source terms F ν Δ t i → f F^\nu\Delta t_{i\to f} F ν Δ t i → f and Σ S i → f \Sigma_{S\,i\to f} Σ S i → f in an arbitrary inertial frame R \mathcal{R} R . On one hand, the entropy S S S and the number of moles of substance N N N are frame independent quantities. It implies that the relativistic continuity equations for these quantities (61) and (91) are Lorentz-invariant. On the other hand, the energy-momentum vector is not frame independent. Thus, for a perfect fluid the relativistic continuity equations for these quantities (200) and (211) are not Lorentz-invariant,
∂ t ( e + v 2 c 2 p ) + ∇ ⋅ ( e v ) = ( f − ∇ p ) ⋅ v (energy) ∂ t ( p + h c 2 v ) − ∇ ⋅ τ = f (momentum) (239) \begin{aligned}
\partial_t\left(e+\frac{v^2}{c^2}p\right)+\nabla\cdot(e\boldsymbol{v})&=(\boldsymbol{f}-\nabla p)\cdot\boldsymbol{v} &&\text{(energy)}\\
\partial_t\left(\boldsymbol{p}+\frac{h}{c^2}\boldsymbol{v}\right)-\nabla\cdot\boldsymbol{\tau}&=\boldsymbol{f} &&\text{(momentum)} \tag*{(239)}
\end{aligned} ∂ t ( e + c 2 v 2 p ) + ∇ ⋅ ( e v ) ∂ t ( p + c 2 h v ) − ∇ ⋅ τ = ( f − ∇ p ) ⋅ v = f (energy) (momentum) (239)
which leads to relativistic corrections due to the relativistic motion at velocity v \boldsymbol{v} v of the local rest frame R ~ \widetilde{\mathcal{R}} R with respect to the inertial frame R \mathcal{R} R . These relativistic corrections couple the relativistic work and heat. The first relativistic correction is an additional power density term ∂ t ( v 2 c 2 p ) \partial_t\left(\frac{v^2}{c^2}p\right) ∂ t ( c 2 v 2 p ) due to the time derivative of the pressure p p p in the relativistic continuity equation for the energy. The second relativistic correction is an additional spatial force density term ∂ t ( h c 2 v ) \partial_t\left(\frac{h}{c^2}\boldsymbol{v}\right) ∂ t ( c 2 h v ) due to the time derivative of the enthalpy density h h h in the relativistic continuity equation for the momentum. These relativistic continuity equations are derived from the stress energy-momentum tensor (152) that is written in contravariant components in the inertial frame R \mathcal{R} R as,
T μ ν = m ~ u μ u ν + p ~ η μ ν = ( ρ ~ + 1 c 2 ( u ~ + p ~ ) ) u μ u ν + p ~ η μ ν (240) T^{\mu\nu}=\widetilde{m}\,u^\mu u^\nu+\widetilde{p}\,\eta^{\mu\nu}=\left(\widetilde{\rho}+\frac{1}{c^2}(\widetilde{u}+\widetilde{p})\right)u^\mu u^\nu+\widetilde{p}\,\eta^{\mu\nu} \tag*{(240)} T μν = m u μ u ν + p η μν = ( ρ + c 2 1 ( u + p ) ) u μ u ν + p η μν (240)
where the relativistic inertia density m ~ \widetilde{m} m is due mainly to the mass density ρ ~ \widetilde{\rho} ρ and also to the internal energy density u ~ \widetilde{u} u and to the pressure p ~ \widetilde{p} p that carry inertia for a relativistic motion. The difference between the inertia density m ~ \widetilde{m} m and the mass density ρ ~ \widetilde{\rho} ρ is a relativistic correction. Since the mass density ρ \rho ρ is part of a component of a rank-two tensor, the relativistic continuity equation for the mass was deduced from the relativistic continuity equation for the energy momentum by projection along the worldline defined by the velocity vector. In fact, it is the relativistic continuity equation (227) for the inertia that was derived in this manner,
∂ t m + ∇ ⋅ ( m v ) = p ˙ c 2 (inertia) (241) \partial_t m + \nabla\cdot(m\mathbf{v}) = \frac{\dot{p}}{c^2} \qquad\text{(inertia)} \tag*{(241)} ∂ t m + ∇ ⋅ ( m v ) = c 2 p ˙ (inertia) (241)
where the relativistic corrections are the difference between the inertia and mass densities, i.e. m − ρ = h c 2 m-\rho=\frac{h}{c^2} m − ρ = c 2 h , and the mass flow rate p ˙ c 2 \frac{\dot{p}}{c^2} c 2 p ˙ due to the proper time derivative of the pressure p p p . In view of this relativistic continuity equation, we determined the relativistic Euler equation (236) ,
( ρ + h c 2 ) a = f − ∇ p ~ − p ˙ c 2 v (Euler equation) (242) \left(\rho+\frac{h}{c^2}\right)\mathbf{a}=\mathbf{f}-\nabla\tilde{p}-\frac{\dot{p}}{c^2}\mathbf{v} \qquad\text{(Euler equation)} \tag*{(242)} ( ρ + c 2 h ) a = f − ∇ p ~ − c 2 p ˙ v (Euler equation) (242)
The first relativistic correction is an additional mass density term h c 2 \frac{h}{c^2} c 2 h due to the enthalpy density of the perfect fluid that carries inertia. The second relativistic correction is an additional spatial force density term − p ˙ c 2 v -\frac{\dot{p}}{c^2}\mathbf{v} − c 2 p ˙ v due to the proper time derivative of the pressure p p p .
In our theoretical derivation of relativistic thermodynamics, we reached the conclusion that the relativistic transformation laws for the intensive quantities, namely the temperature (159), the pressure (180) and the chemical potential (174) are given by,
T = γ T ~ and p = γ 2 p ~ and μ = γ μ ~ (243) T=\gamma\tilde{T} \qquad\text{and} \qquad p=\gamma^2\tilde{p} \qquad\text{and} \qquad\mu=\gamma\tilde{\mu} \tag*{(243)} T = γ T ~ and p = γ 2 p ~ and μ = γ μ ~ (243)
The relativistic transformation law for the temperature is in agreement with Ott [11 ] , Einstein [13 ] in his latter years and Arzelies [12 ] . The relativistic transformation law for the pressure agrees with the result obtained by Sutcliffe [17 ] . These laws for the temperature and the pressure are in agreement with the “Gedankenexperiment” where a black body moving at relativistic velocity emits photons orthogonally to the direction of motion. The relativitic transverse Doppler effect accounts for the frequency shift affecting the temperature and the radiation pressure. Moreover, in view of the relativistic transformation law for the volume (A.43),
V = γ − 1 V ~ (244) V=\gamma^{-1}\tilde{V} \tag*{(244)} V = γ − 1 V ~ (244)
the relativistic transformation laws for the densities of extensive quantities, namely the entropy density (55), the number density (84),
s = γ s ~ and n = γ n ~ (245) s = \gamma\tilde{s} \qquad\text{and} \qquad n = \gamma\tilde{n} \tag*{(245)} s = γ s ~ and n = γ n ~ (245)
in agreement with Einstein [4 ] and Planck [5 ] . The relativistic transformation laws for the other densities of extensive quantities, namely the mass density (98), the energy density (103) and the internal energy density (128) are given by,
ρ = γ 2 ρ ~ and e = γ 2 e ~ and u = γ 2 u ~ (246) \rho= \gamma^{2} \tilde{\rho} \qquad\text{and} \qquad e = \gamma^{2} \tilde{e} \qquad\text{and} \qquad u = \gamma^{2} \tilde{u} \tag*{(246)} ρ = γ 2 ρ ~ and e = γ 2 e ~ and u = γ 2 u ~ (246)
The difference in the behaviour of the relativistic transformation laws (245) and (246) , namely the power of the Lorentz factor γ \gamma γ , is due to the fact that the entropy density s s s and the number density n n n are components of vectors in Minkowski space whereas the mass density ρ \rho ρ , the energy density e e e and the internal energy density u u u are parts of a component of a rank-2 tensor in Minkowski space. Another important aspect of our derivation is the Lorentz invariance of the entropy source density (63), which is a consequence of the Lorentz invariance of the relativistic continuity equation for the entropy,
ρ s = ρ ~ s ~ (247) \rho_s = \tilde{\rho}\tilde{s} \tag*{(247)} ρ s = ρ ~ s ~ (247)
It implies that the relativistic irreversible thermodynamics of a perfect fluid can be described in a frame independent manner. This opens interesting perspectives for the derivation of the relativistic transformation laws for physical laws such as Fourier’s law [30 ] and Ohm’s law [31 ] , and physical effects like the Seebeck effect and the Nernst effect.
In view of the relativistic transformation laws (243) , (245) and (246) , the relativistic thermodynamics of a perfect fluid appears like a γ \gamma γ game. Are you interested in playing the game?
Acknowledgements
The authors would like to thank Aude Maier for initiating this work as well as Jean-Philippe Ansermet and François A. Reuse for insightful discussions.
Appendix A Relativistic kinematics
We briefly recall in this appendix the basic notions of relativistic kinematics that are essential to build a consistent theory of relativistic thermodynamics. According to the fundamental postulate of special relativity, the speed of light c c c is the same with respect to any inertial frame. To describe the underlying structure of space and time, we consider a spherical light wave emitted at time t = t ′ = 0 t = t' = 0 t = t ′ = 0 in an inertial frame R \mathcal{R} R and in another inertial frame R ′ \mathcal{R}' R ′ respectively. In Cartesian coordinates, the infinitesimal distances squared travelled by the light wave in the two inertial frames satisfies the condition, [22 , 20 ]
d x 2 + d y 2 + d z 2 − c 2 d t 2 = d x ′ 2 + d y ′ 2 + d z ′ 2 − c 2 d t ′ 2 = 0 (248) dx^2 + dy^2 + dz^2 - c^2dt^2 = dx'^{2} + dy'^{2} + dz'^{2} - c^2dt'^{2} = 0 \tag*{(248)} d x 2 + d y 2 + d z 2 − c 2 d t 2 = d x ′ 2 + d y ′ 2 + d z ′ 2 − c 2 d t ′ 2 = 0 (248)
In view of the condition (A.1), the infinitesimal frame-independent space-time interval squared d s 2 ds^2 d s 2 is defined as,
d s 2 = − c 2 d t 2 + d x 2 + d y 2 + d z 2 (249) ds^2 = -c^2dt^2 + dx^2 + dy^2 + dz^2 \tag*{(249)} d s 2 = − c 2 d t 2 + d x 2 + d y 2 + d z 2 (249)
and called the infinitesimal line element. It vanishes for a light wave, i.e. d s 2 = 0 ds^2 = 0 d s 2 = 0 , and it is negative for a point mass moving slower than light, i.e. d s 2 < 0 ds^2 < 0 d s 2 < 0 . With the formal notation in contravariant space-time coordinates,
d x 0 = c d t , d x 1 = d x , d x 2 = d y , and d x 3 = d z (250) dx^0 = c\,dt,\qquad dx^1 = dx,\qquad dx^2 = dy,\qquad\text{and}\qquad dx^3 = dz \tag*{(250)} d x 0 = c d t , d x 1 = d x , d x 2 = d y , and d x 3 = d z (250)
the infinitesimal line element (A.2) is recast as,
d s 2 = − ( d x 0 ) 2 + ( d x 1 ) 2 + ( d x 2 ) 2 + ( d x 3 ) 2 (251) ds^2 = -\left(dx^0\right)^2 + \left(dx^1\right)^2 + \left(dx^2\right)^2 + \left(dx^3\right)^2 \tag*{(251)} d s 2 = − ( d x 0 ) 2 + ( d x 1 ) 2 + ( d x 2 ) 2 + ( d x 3 ) 2 (251)
Using the Einstein summation convention, the infinitesimal line element (A.4) is written in contravariant coordinates { x ρ } \{x^\rho\} { x ρ } in R 1 , 3 \mathbb{R}^{1,3} R 1 , 3 as, [21 ]
d s 2 = η μ ν d x μ d x ν (252) ds^2 = \eta_{\mu\nu}\,dx^\mu dx^\nu \tag*{(252)} d s 2 = η μν d x μ d x ν (252)
where μ , ν ∈ { 0 , 1 , 2 , 3 } \mu,\nu\in\{0,1,2,3\} μ , ν ∈ { 0 , 1 , 2 , 3 } and the components of the Minkowski flat metric read,
η μ ν = δ μ ν − 2 δ μ 0 δ ν 0 (253) \eta_{\mu\nu} = \delta_{\mu\nu} - 2\,\delta_{\mu0}\,\delta_{\nu0} \tag*{(253)} η μν = δ μν − 2 δ μ 0 δ ν 0 (253)
which defines the space-time signature ( − , + , + , + ) (-,+,+,+) ( − , + , + , + ) . The set of orthonormal basis vectors in flat space-time R 1 , 3 \mathbb{R}^{1,3} R 1 , 3 is { e μ } \{e_\mu\} { e μ } where μ ∈ { 0 , 1 , 2 , 3 } \mu\in\{0,1,2,3\} μ ∈ { 0 , 1 , 2 , 3 } and the set of orthonormal dual basis vectors is { e ν } \{e^\nu\} { e ν } where ν ∈ { 0 , 1 , 2 , 3 } \nu\in\{0,1,2,3\} ν ∈ { 0 , 1 , 2 , 3 } . This duality is written as,
e ν ⋅ e μ = δ μ ν (254) e^\nu\cdot e_\mu= \delta^\nu_\mu \tag*{(254)} e ν ⋅ e μ = δ μ ν (254)
The set of orthonormal basis vectors in space R 3 \mathbb{R}^3 R 3 is { e j } \{e_j\} { e j } where j ∈ { 1 , 2 , 3 } j \in\{1,2,3\} j ∈ { 1 , 2 , 3 } and the set of orthonormal dual basis vectors is { e k } \{e^k\} { e k } where k ∈ { 1 , 2 , 3 } k \in\{1,2,3\} k ∈ { 1 , 2 , 3 } . This duality is written as,
e k ⋅ e j = δ j k (255) e^k \cdot e_j = \delta^k_j \tag*{(255)} e k ⋅ e j = δ j k (255)
The metric is given by the scalar product of basis vectors in space-time and in Minkowski space, [32 ]
e μ ⋅ e ν = η μ ν and e j ⋅ e k = δ j k (256) e_\mu\cdot e_\nu= \eta_{\mu\nu} \qquad\text{and} \qquad e_j \cdot e_k = \delta_{jk} \tag*{(256)} e μ ⋅ e ν = η μν and e j ⋅ e k = δ j k (256)
where δ j k = η j k \delta_{jk} = \eta_{jk} δ j k = η j k . The infinitesimal spatial displacement vector d x dx d x is expressed in contravariant components in the orthonormal vector basis frame or in covariant components in the orthonormal dual vector basis frame as,
d x = d x j e j = d x k e k (257) dx = dx^j e_j = dx_k e^k \tag*{(257)} d x = d x j e j = d x k e k (257)
The infinitesimal temporal displacement is expressed in contravariant and covariant coordinates as,
d x 0 = c d t and d x 0 = η 0 μ d x μ = η 00 d x 0 = − c d t (258) dx^0 = c\,dt \qquad\text{and} \qquad dx_0 = \eta_{0\mu}\,dx^\mu= \eta_{00}\,dx^0 = -c\,dt \tag*{(258)} d x 0 = c d t and d x 0 = η 0 μ d x μ = η 00 d x 0 = − c d t (258)
In view of the infinitesimal spatial displacement (A.10) and the infinitesimal temporal displacement (A.11), the infinitesimal line element (A.5) is recast as,
d s 2 = d x 0 d x 0 + d x j d x j = − ( c d t ) 2 + ( d x ) 2 (259) ds^2 = dx_0\,dx^0 + dx_j\,dx^j = -(c\,dt)^2 + (dx)^2 \tag*{(259)} d s 2 = d x 0 d x 0 + d x j d x j = − ( c d t ) 2 + ( d x ) 2 (259)
where d x 2 = d x ⋅ d x dx^2 = dx \cdot dx d x 2 = d x ⋅ d x . In the rest frame R ~ \widetilde{R} R , time is the proper time t ~ \widetilde{t} t and there is no spatial displacement, i.e. ( d x ~ ) 2 = 0 (d\widetilde{x})^2 = 0 ( d x ) 2 = 0 . Thus, the infinitesimal line element (A.12) reduces to, [21 ]
d s 2 = − c 2 d t ~ 2 (260) ds^2 = -c^2\,d\widetilde{t}^{\,2} \tag*{(260)} d s 2 = − c 2 d t 2 (260)
In view of the infinitesimal line elements (A.5) and (A.13), the time derivative of the line element is given by,
d s 2 d t ~ 2 = η μ ν d x μ d t ~ d x ν d t ~ = − c 2 (261) \frac{ds^2}{d\tilde{t}^2} = \eta_{\mu\nu} \frac{dx^\mu}{d\tilde{t}} \frac{dx^\nu}{d\tilde{t}} = -c^2 \tag*{(261)} d t ~ 2 d s 2 = η μν d t ~ d x μ d t ~ d x ν = − c 2 (261)
The contravariant components of the velocity vector of a particle are defined as, [33 ]
u μ = d x μ d t (262) u^\mu= \frac{dx^\mu}{dt} \tag*{(262)} u μ = d t d x μ (262)
The covariant components of the velocity vector are given by,
u μ = η μ ν d x ν d t ~ (263) u_\mu= \eta_{\mu\nu} \frac{dx^\nu}{d\tilde{t}} \tag*{(263)} u μ = η μν d t ~ d x ν (263)
In view of the time derivative of the infinitesimal line element (A.14), the contravariant (A.15) and covariant (A.16) components of the velocity vector, the contraction of these components is frame independent,
u μ u μ = − c 2 (264) u_\mu u^\mu= -c^2 \tag*{(264)} u μ u μ = − c 2 (264)
The Lorentz factor γ \gamma γ is defined as, [22 , 20 ]
γ = d t d t ~ thus d t = γ d t ~ (265) \gamma= \frac{dt}{d\tilde{t}} \qquad\text{thus} \qquad dt = \gamma d\tilde{t} \tag*{(265)} γ = d t ~ d t thus d t = γ d t ~ (265)
which describes time dilatation. The contravariant components of the velocity vector (A.15) can be recast in terms of the Lorentz factor (A.18) as,
u μ = d x μ d t d t d t ~ = γ d x μ d t (266) u^\mu= \frac{dx^\mu}{dt} \frac{dt}{d\tilde{t}} = \gamma\frac{dx^\mu}{dt} \tag*{(266)} u μ = d t d x μ d t ~ d t = γ d t d x μ (266)
The covariant components of the velocity vector (A.16) can also be recast in terms of the Lorentz factor (A.18) as,
u μ = η μ ν d x ν d t d t d t ~ = γ d x μ d t (267) u_\mu= \eta_{\mu\nu} \frac{dx^\nu}{dt} \frac{dt}{d\tilde{t}} = \gamma\frac{dx_\mu}{dt} \tag*{(267)} u μ = η μν d t d x ν d t ~ d t = γ d t d x μ (267)
In view of relations of the contravariant components of the velocity vector (A.15) and the Lorentz factor (A.18), the contravariant time component of the velocity vector is written as,
u 0 = d x 0 d t = c d t d t ~ = γ c (268) u^0 = \frac{dx^0}{dt} = c \frac{dt}{d\tilde{t}} = \gamma c \tag*{(268)} u 0 = d t d x 0 = c d t ~ d t = γ c (268)
and the covariant time component of the velocity vector is given by,
u 0 = η 0 μ u μ = η 00 u 0 = − γ c (269) u_{0}=\eta_{0\mu}u^{\mu}=\eta_{00}u^{0}=-\gamma c \tag*{(269)} u 0 = η 0 μ u μ = η 00 u 0 = − γ c (269)
The spatial velocity vector v \mathbf{v} v is expressed in contravariant components in the orthonormal basis frame as,
v = v j e j = d x j d t e j = d x j d t ~ d t ~ d t e j = 1 γ u j e j (270) \mathbf{v}=v^{j}\mathbf{e}_{j}=\frac{dx^{j}}{dt}\mathbf{e}_{j}=\frac{dx^{j}}{d\tilde{t}}\frac{d\tilde{t}}{dt}\mathbf{e}_{j}=\frac{1}{\gamma}u^{j}\mathbf{e}_{j} \tag*{(270)} v = v j e j = d t d x j e j = d t ~ d x j d t d t ~ e j = γ 1 u j e j (270)
or in covariant components in the orthonormal dual basis frame as v j = v j e j v^{j}=v_{j}\mathbf{e}^{j} v j = v j e j . According to the contravariant components of the velocity vector (A.19) and the spatial velocity vector (A.23), the contravariant and covariant component of the velocity vector are written as,
u j = γ v j and u j = γ v j (271) u^{j}=\gamma v^{j}\qquad\text{and}\qquad u_{j}=\gamma v_{j} \tag*{(271)} u j = γ v j and u j = γ v j (271)
In view of the frame duality (A.8), the contravariant (A.21) and covariant (A.22) time components of the velocity and the spatial components (A.24) of the velocity, the contraction of the velocities (A.17) is recast as,
u μ u μ = u 0 u 0 + u i u i = γ 2 ( − c 2 + ( v j e j ) ⋅ ( v k e k ) ) = γ 2 ( − c 2 + v 2 ) (272) u_{\mu}u^{\mu}=u_{0}u^{0}+u_{i}u^{i}=\gamma^{2}\left(-c^{2}+\left(v_{j}\mathbf{e}^{j}\right)\cdot\left(v^{k}\mathbf{e}_{k}\right)\right)=\gamma^{2}\left(-c^{2}+v^{2}\right) \tag*{(272)} u μ u μ = u 0 u 0 + u i u i = γ 2 ( − c 2 + ( v j e j ) ⋅ ( v k e k ) ) = γ 2 ( − c 2 + v 2 ) (272)
where v 2 = v ⋅ v v^{2}=\mathbf{v}\cdot\mathbf{v} v 2 = v ⋅ v . The identification of the velocity contractions (A.17) and (A.25) yields a kinematic expression [1 ] for the Lorentz factor (A.18),
γ ≡ γ ( v ) = ( 1 − v 2 c 2 ) − 1 / 2 (273) \gamma\equiv\gamma(v)=\left(1-\frac{v^{2}}{c^{2}}\right)^{-1/2} \tag*{(273)} γ ≡ γ ( v ) = ( 1 − c 2 v 2 ) − 1/2 (273)
The contravariant components of the acceleration vector are defined as, [20 ]
a μ = d u μ d t ~ = d x ν d t ~ d u μ d x ν = u ν ∂ ν u μ (274) a^{\mu}=\frac{du^{\mu}}{d\tilde{t}}=\frac{dx^{\nu}}{d\tilde{t}}\frac{du^{\mu}}{dx^{\nu}}=u^{\nu}\partial_{\nu}u^{\mu} \tag*{(274)} a μ = d t ~ d u μ = d t ~ d x ν d x ν d u μ = u ν ∂ ν u μ (274)
According to the velocity contraction (A.17) and the contravariant components of the acceleration vector (A.27), the velocity and acceleration vectors are orthogonal, which is written in components as,
u μ a μ = u μ u ν ∂ ν u μ = 1 2 u ν ∂ ν ( u μ u μ ) = − 1 2 u ν ∂ ν c 2 = 0 (275) u_{\mu}a^{\mu}=u_{\mu}u^{\nu}\partial_{\nu}u^{\mu}=\frac{1}{2}u^{\nu}\partial_{\nu}\left(u_{\mu}u^{\mu}\right)=-\frac{1}{2}u^{\nu}\partial_{\nu}c^{2}=0 \tag*{(275)} u μ a μ = u μ u ν ∂ ν u μ = 2 1 u ν ∂ ν ( u μ u μ ) = − 2 1 u ν ∂ ν c 2 = 0 (275)
According to the Lorentz factor (A.18), the contravariant components of the acceleration vector are recast as,
a μ = d u μ d t ~ = d t d t ~ d u μ d t = γ d u μ d t (276) a^\mu= \frac{du^\mu}{d\tilde{t}} = \frac{dt}{d\tilde{t}}\frac{du^\mu}{dt} = \gamma\frac{du^\mu}{dt} \tag*{(276)} a μ = d t ~ d u μ = d t ~ d t d t d u μ = γ d t d u μ (276)
In an inertial frame R \mathcal{R} R , the infinitesimal hypervolume in space-time is expressed as,
d V 4 = d x 0 d x 1 d x 2 d x 3 = c d t d V 3 (277) dV_4 = dx^0 dx^1 dx^2 dx^3 = c\,dt\,dV_3 \tag*{(277)} d V 4 = d x 0 d x 1 d x 2 d x 3 = c d t d V 3 (277)
where d V 3 dV_3 d V 3 is the infinitesimal spatial volume in this frame. In the rest frame R ~ \widetilde{\mathcal{R}} R , the infinitesimal hypervolume in space-time is expressed as,
d V ~ 4 = d x ~ 0 d x ~ 1 d x ~ 2 d x ~ 3 = c d t ~ d V ~ 3 (278) d\widetilde{V}_4 = d\widetilde{x}^0 d\widetilde{x}^1 d\widetilde{x}^2 d\widetilde{x}^3 = c\,d\widetilde{t}\,d\widetilde{V}_3 \tag*{(278)} d V 4 = d x 0 d x 1 d x 2 d x 3 = c d t d V 3 (278)
where we denote physical quantities evaluated in the rest frame R ~ \widetilde{\mathcal{R}} R with a tilde. The infinitesimal hypervolume in the rest frame is related to the infinitesimal hypervolume in the inertial frame R \mathcal{R} R by, [33 ]
d x ~ 0 d x ~ 1 d x ~ 2 d x ~ 3 = ∂ ( x ~ 0 , x ~ 1 , x ~ 2 , x ~ 3 ) ∂ ( x 0 , x 1 , x 2 , x 3 ) d x 0 d x 1 d x 2 d x 3 (279) d\widetilde{x}^0 d\widetilde{x}^1 d\widetilde{x}^2 d\widetilde{x}^3 = \frac{\partial\left(\widetilde{x}^0,\widetilde{x}^1,\widetilde{x}^2,\widetilde{x}^3\right)}{\partial\left(x^0,x^1,x^2,x^3\right)}\,dx^0 dx^1 dx^2 dx^3 \tag*{(279)} d x 0 d x 1 d x 2 d x 3 = ∂ ( x 0 , x 1 , x 2 , x 3 ) ∂ ( x 0 , x 1 , x 2 , x 3 ) d x 0 d x 1 d x 2 d x 3 (279)
where the Jacobian is the determinant of the Lorentz transformation Λ \Lambda Λ ,
∂ ( x ~ 0 , x ~ 1 , x ~ 2 , x ~ 3 ) ∂ ( x 0 , x 1 , x 2 , x 3 ) = det ( Λ ) (280) \frac{\partial\left(\widetilde{x}^0,\widetilde{x}^1,\widetilde{x}^2,\widetilde{x}^3\right)}{\partial\left(x^0,x^1,x^2,x^3\right)} = \det\left(\Lambda\right) \tag*{(280)} ∂ ( x 0 , x 1 , x 2 , x 3 ) ∂ ( x 0 , x 1 , x 2 , x 3 ) = det ( Λ ) (280)
and the components of the Lorentz transformation are given by,
Λ ν μ = ∂ x ~ μ ∂ x ν (281) \Lambda^\mu_{\ \nu} = \frac{\partial\widetilde{x}^\mu}{\partial x^\nu} \tag*{(281)} Λ ν μ = ∂ x ν ∂ x μ (281)
For an orthochronous Lorentz transformation where the rest frame R ~ \widetilde{\mathcal{R}} R is moving with respect to the inertial frame R \mathcal{R} R at scalar velocity v i v^i v i along the i i i -axis, the components are given by, [21 ]
Λ 0 0 = γ and Λ j 0 = Λ 0 j = − 1 c u j = − γ c v j = − γ c v i δ i j Λ k j = ( γ − 1 ) δ i j δ k i + δ k j (282) \begin{aligned}
\Lambda^0_{\ 0} &= \gamma\quad\text{and} \quad\Lambda^0_{\ j} = \Lambda^j_{\ 0} = -\frac{1}{c}u^j = -\frac{\gamma}{c}v^j = -\frac{\gamma}{c}v^i\delta_i^j \\
\Lambda^j_{\ k} &= (\gamma-1)\delta_i^j\delta^i_k+\delta^j_k \tag*{(282)}
\end{aligned} Λ 0 0 Λ k j = γ and Λ j 0 = Λ 0 j = − c 1 u j = − c γ v j = − c γ v i δ i j = ( γ − 1 ) δ i j δ k i + δ k j (282)
In view of the Lorentz factor (A.26) and the components of the Lorentz transformation (A.35), the determinant of the Lorentz transformation yields,
det ( Λ ) = Λ 0 0 Λ i i − Λ i 0 Λ 0 i = γ 2 ( 1 − v 2 c 2 ) = 1 (283) \det(\Lambda) = \Lambda^0_{\ 0}\Lambda^i_{\ i}-\Lambda^0_{\ i}\Lambda^i_{\ 0} = \gamma^2\left(1-\frac{v^2}{c^2}\right)=1 \tag*{(283)} det ( Λ ) = Λ 0 0 Λ i i − Λ i 0 Λ 0 i = γ 2 ( 1 − c 2 v 2 ) = 1 (283)
where v 2 = v j v j v^2 = v_j v^j v 2 = v j v j . This means that a Lorentz transformation is an isometry in Minkowski space. The infinitesimal hypervolume in the inertial frame R \mathcal{R} R is related to the infinitesimal hypervolume in the rest frame R ~ \widetilde{\mathcal{R}} R by, [33 ]
d x 0 d x 1 d x 2 d x 3 = ∂ ( x 0 , x 1 , x 2 , x 3 ) ∂ ( x ~ 0 , x ~ 1 , x ~ 2 , x ~ 3 ) d x ~ 0 d x ~ 1 d x ~ 2 d x ~ 3 (284) dx^0 dx^1 dx^2 dx^3 = \frac{\partial(x^0,x^1,x^2,x^3)}{\partial(\widetilde{x}^0,\widetilde{x}^1,\widetilde{x}^2,\widetilde{x}^3)}d\widetilde{x}^0d\widetilde{x}^1d\widetilde{x}^2d\widetilde{x}^3 \tag*{(284)} d x 0 d x 1 d x 2 d x 3 = ∂ ( x 0 , x 1 , x 2 , x 3 ) ∂ ( x 0 , x 1 , x 2 , x 3 ) d x 0 d x 1 d x 2 d x 3 (284)
where the Jacobian is the determinant of the inverse orthochronous Lorentz transformation Λ − 1 \Lambda^{-1} Λ − 1 ,
∂ ( x 0 , x 1 , x 2 , x 3 ) ∂ ( x ~ 0 , x ~ 1 , x ~ 2 , x ~ 3 ) = ( ∂ ( x ~ 0 , x ~ 1 , x ~ 2 , x ~ 3 ) ∂ ( x 0 , x 1 , x 2 , x 3 ) ) − 1 = det ( Λ ) − 1 = det ( Λ − 1 ) (285) \frac{\partial(x^0,x^1,x^2,x^3)}{\partial(\widetilde{x}^0,\widetilde{x}^1,\widetilde{x}^2,\widetilde{x}^3)}
=
\left(\frac{\partial(\widetilde{x}^0,\widetilde{x}^1,\widetilde{x}^2,\widetilde{x}^3)}{\partial(x^0,x^1,x^2,x^3)}\right)^{-1}
= \det(\Lambda)^{-1} = \det(\Lambda^{-1}) \tag*{(285)} ∂ ( x 0 , x 1 , x 2 , x 3 ) ∂ ( x 0 , x 1 , x 2 , x 3 ) = ( ∂ ( x 0 , x 1 , x 2 , x 3 ) ∂ ( x 0 , x 1 , x 2 , x 3 ) ) − 1 = det ( Λ ) − 1 = det ( Λ − 1 ) (285)
The inverse orthochronous Lorentz transformation is also an isometry in Minkowski space, i.e. det ( Λ − 1 ) = 1 \det(\Lambda^{-1}) = 1 det ( Λ − 1 ) = 1 , in view of the determinant (A.36) since,
det ( Λ ) det ( Λ − 1 ) = det ( Λ Λ − 1 ) = det ( 1 4 ) = 1 (286) \det(\Lambda)\det(\Lambda^{-1}) = \det(\Lambda\Lambda^{-1}) = \det(\mathbb{1}_4) = 1 \tag*{(286)} det ( Λ ) det ( Λ − 1 ) = det ( Λ Λ − 1 ) = det ( 1 4 ) = 1 (286)
For the orthochronous inverse Lorentz transformation where the inertial frame R \mathcal{R} R is moving with respect to the rest frame R ~ \widetilde{\mathcal{R}} R at velocity − v i -v^i − v i along the x i x^i x i -axis, the components are given by, [21 ]
( Λ − 1 ) 0 0 = γ and ( Λ − 1 ) j 0 = ( Λ − 1 ) 0 j = 1 c u j = γ c v j = γ c v i δ i j (\Lambda^{-1})^0_0 = \gamma\qquad\text{and} \qquad(\Lambda^{-1})^0_j = (\Lambda^{-1})^j_0 = \frac{1}{c}u^j = \frac{\gamma}{c}v^j = \frac{\gamma}{c}v^i\delta^j_i ( Λ − 1 ) 0 0 = γ and ( Λ − 1 ) j 0 = ( Λ − 1 ) 0 j = c 1 u j = c γ v j = c γ v i δ i j
( Λ − 1 ) k j = ( γ − 1 ) δ i j δ k i + δ k j (287) (\Lambda^{-1})^j_k = (\gamma- 1)\delta^j_i\delta^i_k + \delta^j_k \tag*{(287)} ( Λ − 1 ) k j = ( γ − 1 ) δ i j δ k i + δ k j (287)
According to the infinitesimal volumes in the inertial frame (A.30) and the rest frame (A.31), the volume transformation law (A.32), the Jacobian (A.33) and the determinant of the Lorentz transformation (A.36), the infinitesimal hypervolume in space-time is invariant,
d V 4 = d V ~ 4 (288) dV_4 = d\widetilde{V}_4 \tag*{(288)} d V 4 = d V 4 (288)
In view of the infinitesimal volumes in the inertial frame (A.30) and the rest frame (A.31), and the infinitesimal hypervolume in space-time (A.41), we obtain the identity,
d t d V 3 = d t ~ d V ~ 3 (289) dt\,dV_3 = d\widetilde{t}\,d\widetilde{V}_3 \tag*{(289)} d t d V 3 = d t d V 3 (289)
According to the time dilatation (A.18) and the hypervolume identity (A.42), we obtain a volume contraction due to a length contraction in the direction of motion along the i i i -axis, [22 ]
d V 3 = 1 γ d V ~ 3 (290) dV_3 = \frac{1}{\gamma}d\widetilde{V}_3 \tag*{(290)} d V 3 = γ 1 d V 3 (290)
Appendix B Relativistic dynamics
We briefly review in this appendix the foundations of the relativistic dynamics of a particle within a special relativistic framework. In particular, we derive the relativistic transformation laws for the mass and energy of the particle between the rest frame R ~ \widetilde{\mathcal{R}} R and an inertial frame R \mathcal{R} R . The contravariant time component of the momentum vector of a particle in an inertial frame R \mathcal{R} R is the energy up to a factor c − 1 c^{-1} c − 1 , [22 ]
P 0 = E c (291) P^0 = \frac{E}{c} \tag*{(291)} P 0 = c E (291)
and the covariant time component of the momentum vector in the inertial frame R \mathcal{R} R is given by,
P 0 = η 0 μ P μ = η 00 P 0 = − E c (292) P_0 = \eta_{0\mu}P^\mu= \eta_{00}P^0 = -\frac{E}{c} \tag*{(292)} P 0 = η 0 μ P μ = η 00 P 0 = − c E (292)
The contravariant component of the momentum vector P μ P^\mu P μ is the product of the mass M ~ \widetilde{M} M of the particle in the rest frame R \mathcal{R} R and the contravariant component of the velocity vector u μ u^\mu u μ , [34 ]
P μ = M ~ u μ (293) P^\mu= \widetilde{M}u^\mu \tag*{(293)} P μ = M u μ (293)
In view of the contravariant time component of the velocity vector (A.21) and the contravariant component of the momentum vector (B.3), the energy for a relativistic particle (B.2) is written as, [22 ]
E = c P 0 = c M ~ u 0 = γ M ~ c 2 (294) E = cP^0 = c\widetilde{M}u^0 = \gamma\widetilde{M}c^2 \tag*{(294)} E = c P 0 = c M u 0 = γ M c 2 (294)
In view of the Lorentz factor (A.26), the energy (B.4) is recast explicitly as,
E = ( 1 − v 2 c 2 ) − 1 / 2 M ~ c 2 (295) E = \left(1-\frac{v^2}{c^2}\right)^{-1/2}\widetilde{M}c^2 \tag*{(295)} E = ( 1 − c 2 v 2 ) − 1/2 M c 2 (295)
In the non-relativistic limit, to first-order in v 2 / c 2 v^2/c^2 v 2 / c 2 , the energy (B.5) reduces to,
E ≃ ( 1 + 1 2 v 2 c 2 ) M c 2 = M ~ c 2 + 1 2 M ~ v 2 (non-relativistic) (296) E \simeq\left(1+\frac{1}{2}\frac{v^2}{c^2}\right)Mc^2=\widetilde{M}c^2+\frac{1}{2}\widetilde{M}v^2 \qquad\text{(non-relativistic)} \tag*{(296)} E ≃ ( 1 + 2 1 c 2 v 2 ) M c 2 = M c 2 + 2 1 M v 2 (non-relativistic) (296)
as expected. The energy E ~ \widetilde{E} E of the particle in the rest frame R ~ \widetilde{\mathcal{R}} R is the rest energy,
E ~ = M ~ c 2 (297) \widetilde{E}=\widetilde{M}c^2 \tag*{(297)} E = M c 2 (297)
The energy E E E of the particle (B.4) in the inertial frame R \mathcal{R} R is expressed in terms of the energy E ~ \widetilde{E} E of the particle (B.7) in the rest frame R ~ \widetilde{\mathcal{R}} R as, [22 , 20 ]
E = γ E ~ (298) E=\gamma\widetilde{E} \tag*{(298)} E = γ E (298)
In view of the energy E ~ \widetilde{E} E of the particle (B.7) in the rest frame R ~ \widetilde{\mathcal{R}} R , the energy E E E of the particle in the inertial frame R \mathcal{R} R is recast in terms of the mass M M M in the inertial frame R \mathcal{R} R as,
E = M c 2 (299) E=Mc^2 \tag*{(299)} E = M c 2 (299)
According to the energy of the particle (B.4) and (B.9), the mass M M M of the particle in the inertial frame R \mathcal{R} R is expressed in terms of the mass M ~ \widetilde{M} M of the particle in the rest frame R ~ \widetilde{\mathcal{R}} R as, [22 , 20 ]
M = γ M ~ (300) M=\gamma\widetilde{M} \tag*{(300)} M = γ M (300)
The spatial momentum vector P \mathbf{P} P is expressed in contravariant components in the orthonormal basis frame or in covariant components in the orthonormal dual basis frame as,
P = P j e j = P k e k (301) \mathbf{P}=P^j e_j=P_k e^k \tag*{(301)} P = P j e j = P k e k (301)
In view of the spatial velocity vector (A.23) and the contravariant components of the momentum vector (B.3), the spatial momentum vector (B.11) is recast as,
P = P j e j = M ~ u j e j = γ M ~ v j e j = γ M ~ v (302) \mathbf{P}=P^j e_j=\widetilde{M}u^j e_j=\gamma\widetilde{M}v^j e_j=\gamma\widetilde{M}\mathbf{v} \tag*{(302)} P = P j e j = M u j e j = γ M v j e j = γ M v (302)
The covariant component of the momentum vector P μ P_\mu P μ is the product of the mass M ~ \widetilde{M} M in the rest frame R \mathcal{R} R and the covariant component of the velocity vector
P μ = M ~ u μ (303) P_{\mu}=\widetilde{M}u_{\mu} \tag*{(303)} P μ = M u μ (303)
According to the contraction of the velocities (A.17), the contravariant and covariant components of the momentum vector (B.3) and (B.13), the contraction of the momenta is expressed as,
P μ P μ = M ~ 2 u μ u μ = − M ~ 2 c 2 (304) P_{\mu}P^{\mu}=\widetilde{M}^{2}u_{\mu}u^{\mu}=-\widetilde{M}^{2}c^{2} \tag*{(304)} P μ P μ = M 2 u μ u μ = − M 2 c 2 (304)
which is a Casimir invariant of the Poincaré algebra. In view of the frame duality (A.8), the contravariant and covariant temporal and spatial components of the momentum vector (B.1), (B.2), (B.3) and (B.13), the Casimir invariant is recast as,
P μ P μ = P 0 P 0 + P j P j = − E 2 c 2 + ( P j e j ) ⋅ ( P k e k ) = − E 2 c 2 + P 2 (305) P_{\mu}P^{\mu}=P_{0}P^{0}+P_{j}P^{j}=-\frac{E^{2}}{c^{2}}+(P_{j}e^{j})\cdot(P^{k}e_{k})=-\frac{E^{2}}{c^{2}}+P^{2} \tag*{(305)} P μ P μ = P 0 P 0 + P j P j = − c 2 E 2 + ( P j e j ) ⋅ ( P k e k ) = − c 2 E 2 + P 2 (305)
where P 2 = P ⋅ P P^{2}=P\cdot P P 2 = P ⋅ P . The identification of the Casimir invariants (B.14) and (B.15) yields the energy-momentum invariant for the relativistic motion of a particle, [22 ]
E 2 = P 2 c 2 + M ~ 2 c 4 (306) E^{2}=P^{2}c^{2}+\widetilde{M}^{2}c^{4} \tag*{(306)} E 2 = P 2 c 2 + M 2 c 4 (306)
where P 2 c 2 P^{2}c^{2} P 2 c 2 is the square of the relativistic kinetic energy and M ~ 2 c 4 \widetilde{M}^{2}c^{4} M 2 c 4 is the square of the rest energy in the rest frame R ~ \widetilde{R} R . The relativistic law of motion states that the force vector is the time derivative of the momentum vector. In view of the definition (A.18) of the Lorentz factor, the contravariant components of the force vector are given by, [22 , 21 ]
F μ = d P μ d t ~ = d P μ d t d t d t ~ = γ d P μ d t (307) F^{\mu}=\frac{dP^{\mu}}{d\widetilde{t}}=\frac{dP^{\mu}}{dt}\frac{dt}{d\widetilde{t}}=\gamma\frac{dP^{\mu}}{dt} \tag*{(307)} F μ = d t d P μ = d t d P μ d t d t = γ d t d P μ (307)
Since the mass M ~ \widetilde{M} M in the rest frame R ~ \widetilde{R} R is constant, the contravariant components of the acceleration vector (A.29) and momentum vector (B.3) imply that the contravariant components of the force vector can be recast as,
F μ = d d t ( M ~ u μ ) = M ~ d u μ d t = M ~ a μ (308) F^{\mu}=\frac{d}{dt}\left(\widetilde{M}u^{\mu}\right)=\widetilde{M}\frac{du^{\mu}}{dt}=\widetilde{M}a^{\mu} \tag*{(308)} F μ = d t d ( M u μ ) = M d t d u μ = M a μ (308)
In view of the orthogonality between the velocity and the acceleration vectors (A.28) and the relativistic law of motion (B.17), the velocity and force vectors are orthogonal,
u μ F μ = M ~ u μ a μ = 0 (309) u_{\mu}F^{\mu}=\widetilde{M}u_{\mu}a^{\mu}=0 \tag*{(309)} u μ F μ = M u μ a μ = 0 (309)
According to the non-relativistic law of motion, the spatial force vector F \mathbf{F} F in the inertial frame R \mathcal{R} R is expressed in contravariant components in the orthonormal basis frame as the time derivative of the spatial momentum vector (B.11),
F = d P d t = d P j d t e j (310) \mathbf{F}=\frac{d\mathbf{P}}{dt}=\frac{dP^{j}}{dt}\mathbf{e}_{j} \tag*{(310)} F = d t d P = d t d P j e j (310)
In view of the contravariant components of the relativistic law of motion (B.17), the spatial force vector in the inertial frame R \mathcal{R} R can be recast as,
F = 1 γ F j e j (311) \mathbf{F}=\frac{1}{\gamma}F^{j}\mathbf{e}_{j} \tag*{(311)} F = γ 1 F j e j (311)
In view of the contravariant time component of the momentum vector (B.1) and the relativistic law of motion (B.17), the contravariant time component of the force vector is given by,
F 0 = γ d P 0 d t = γ c d E d t (312) F^{0}=\gamma\frac{dP^{0}}{dt}=\frac{\gamma}{c}\frac{dE}{dt} \tag*{(312)} F 0 = γ d t d P 0 = c γ d t d E (312)
According to the covariant time component of the velocity vector (A.22), the spatial components of the velocity vector (A.24) and the time component of the force vector (B.22), the orthogonality condition between the velocity and force vectors (B.19) is written as,
u 0 F 0 + u j F j = − γ 2 d E d t + γ v j F j = 0 (313) u_{0}F^{0}+u_{j}F^{j}=-\gamma^{2}\frac{dE}{dt}+\gamma v_{j}F^{j}=0 \tag*{(313)} u 0 F 0 + u j F j = − γ 2 d t d E + γ v j F j = 0 (313)
In view of the frame duality (A.8), the spatial velocity vector (A.23) and the force vector (B.21), the orthogonality condition between the velocity and force vectors (B.23) yields the time derivative of the energy,
d E d t = 1 γ F j v j = ( 1 γ F j e j ) ⋅ ( v k e k ) = F ⋅ v (314) \frac{dE}{dt}=\frac{1}{\gamma}F^{j}v_{j}=\left(\frac{1}{\gamma}F^{j}\mathbf{e}_{j}\right)\cdot\left(v_{k}\mathbf{e}^{k}\right)=\mathbf{F}\cdot\mathbf{v} \tag*{(314)} d t d E = γ 1 F j v j = ( γ 1 F j e j ) ⋅ ( v k e k ) = F ⋅ v (314)
According to the time derivative of the energy (B.24), the contravariant time component of the force vector (B.22), which represents the relativistic power exerted by the force, is recast as, [22 , 21 ]
F 0 = γ c F ⋅ v (315) F^{0}=\frac{\gamma}{c}\mathbf{F}\cdot\mathbf{v} \tag*{(315)} F 0 = c γ F ⋅ v (315)