Generalized thermodynamic relations for perfect spin hydrodynamics

Abstract

Generalized thermodynamic relations are introduced into the framework of a relativistic perfect spin hydrodynamics. They allow for consistent treatment of spin degrees of freedom, including the use of spin tensors whose structure follows from microscopic calculations. The obtained results are important for establishing consistency between different formulations of spin hydrodynamics and form the basis for introducing dissipative corrections.

Introduction — Recent measurements of non-zero spin polarization of hyperons [1, 2, 3] and vector mesons [4] produced in relativistic heavy-ion collisions have triggered broad interest in the spin polarization phenomena in strongly interacting matter, for a recent review see Ref. [5]. On the theory side, there exist several approaches to incorporate spin degrees of freedom into the framework of relativistic hydrodynamics  [6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29]. The latter has become the main theoretical tool used to describe the spacetime evolution of strongly interacting matter produced in heavy-ion collisions [30, 31, 32], hence, the inclusion of spin dynamics in the hydrodynamics formalism seems to be an inevitable necessity.

A certain difficulty in develo** the formalism of spin hydrodynamics is the fact that there are different formulations of this approach using different assumptions. The differences appear already at the basic level of relativistic thermodynamic relations and definitions of the fundamental macroscopic quantities such as the spin tensor.

In this work we critically reexamine thermodynamic relations used in perfect spin hydrodynamics of particles with spin ${\nicefrac{{1}}{{2}}}$ and propose to introduce their generalized forms that can be used for large values of the spin polarization tensor, $\omega_{\mu\nu}$, and with kinetic-theory motivated forms of the spin tensor $S^{\lambda,\mu\nu}$. In this way, we remove a gap between the works that use kinetic-theory concepts as the starting point [33, 34] and the works that use phenomenological expressions for the spin tensor and construct dissipative corrections using the positivity of the entropy production as the main physical ansatz [10, 16].

Interestingly, the new tensor forms of the thermodynamic relations proposed in this work include terms whose mathematical structure is typical for dissipative corrections. In our case their presence is not related to the entropy production but results from a richer description of the system that requires introduction of the spin polarization tensor (reminding us of the structure of relativistic magnetohydrodynamics). An important consequence of the fact that such terms appear at the perfect-fluid level is that they should be taken into account in theoretical constructions aiming at the development of dissipative spin hydrodynamics.

Notation and conventions — For the Levi-Civita tensor $\epsilon^{\mu\nu\alpha\beta}$ we follow the convention $\epsilon^{0123}=-\epsilon_{0123}=+1$. The metric tensor is of the form $g_{\mu\nu}=\textrm{diag}(+1,-1,-1,-1)$. Throughout the text we make use of natural units, $\hbar=c=k_{B}=1$. The scalar product of two four-vectors $a$ and $b$ reads $a\cdot b=a^{0}b^{0}-{\bm{a}}\cdot{\bm{b}}$, where the three-vectors are denoted by bold font.

Scalar and tensor forms of thermodynamic relations — The fundamental thermodynamic relations used in standard relativistic hydrodynamics consist of the identity

$$\varepsilon+P=T\sigma+\mu n$$ (1)

and the first law of thermodynamics

$$d\varepsilon=Td\sigma+\mu dn.$$ (2)

Here $\varepsilon$, $P$, $T$, $\sigma$, $\mu$ and $n$ are the local energy density, pressure, temperature, entropy density, baryon chemical potential, and baryon number density, respectively. The identity (1) is a direct consequence of the extensivity of energy, entropy and baryon number (they are all proportional to the system’s volume). Equations (1) and (2) imply the Gibbs-Duhem relation

$$dP=\sigma dT+nd\mu.$$ (3)

In order to take into account dissipation effects, one usually rewrites Eqs. (1)–(2) in a tensor (four-vector) form. This is achieved by multiplication of Eqs. (1)–(3) by the local four-velocity of the fluid $u^{\mu}$, which leads to the following expressions

$$S^{\mu}=\sigma u^{\mu}=P\beta^{\mu}-\xi N^{\mu}+\beta_{\lambda}T^{\lambda\mu},$$ (4)

$$dS^{\mu}=-\xi dN^{\mu}+\beta_{\lambda}dT^{\lambda\mu},$$ (5)

$$d(P\beta^{\mu})=N^{\mu}d\xi-T^{\lambda\mu}d\beta_{\lambda}.$$ (6)

Here we have introduced common notation: $\beta^{\mu}=u^{\mu}/T$, $\beta=\sqrt{\beta^{\lambda}\beta_{\lambda}}=1/T$, and $\xi=\mu/T$. The tensors $N^{\mu}$ and $T^{\lambda\mu}$ describe the baryon current and energy-momentum tensor for a perfect fluid, namely, $N^{\mu}=nu^{\mu}$ and $T^{\lambda\mu}=(\varepsilon+P)u^{\lambda}u^{\mu}-Pg^{\lambda\mu}=\varepsilon u^{\lambda}u^{\mu}-P\Delta^{\lambda\mu}$, where the tensor $\Delta^{\lambda\mu}=g^{\lambda\mu}-u^{\lambda}u^{\mu}$ projects on the space orthogonal to flow.

Many formulations of relativistic spin hydrodynamics as their starting points choose an extension of Eqs. (1)–(3) that includes the spin polarization tensor $\omega_{\alpha\beta}$, the tensor spin chemical potential $\Omega_{\alpha\beta}=T\,\omega_{\alpha\beta}$, and the spin density tensor $S^{\alpha\beta}$. They read

$$\varepsilon+P=T\sigma+\mu n+{\scriptstyle{\frac{1}{2}}}\Omega_{\alpha\beta}S^{\alpha\beta},$$ (7)

$$d\varepsilon=Td\sigma+\mu dn+{\scriptstyle{\frac{1}{2}}}\Omega_{\alpha\beta}dS^{\alpha\beta},$$ (8)

$$dP=\sigma dT+nd\mu+{\scriptstyle{\frac{1}{2}}}S^{\alpha\beta}d\Omega_{\alpha\beta}.$$ (9)

We note that $\omega_{\alpha\beta}$, $\Omega_{\alpha\beta}$, and $S_{\alpha\beta}$ are all rank-2 antisymmetric tensors. Below we will use the following parametrization of the spin polarization tensor [9]

$$\omega_{\alpha\beta}=k_{\alpha}u_{\beta}-k_{\beta}u_{\alpha}+t_{\alpha\beta},$$ (10)

where $t_{\alpha\beta}=\epsilon_{\alpha\beta\gamma\delta}u^{\gamma}\omega^{\delta}$, and the four-vectors $k$ and $\omega$ are orthogonal to the flow vector $u$, namely, $k\cdot u=0$ and $\omega\cdot u=0$. By multiplying Eqs. (7)–(9) by $u^{\mu}$, we obtain

$$S^{\mu}=P\beta^{\mu}-\xi N^{\mu}+\beta_{\lambda}T^{\lambda\mu}-{\scriptstyle{\frac{1}{2}}}\omega_{\alpha\beta}S^{\mu,\alpha\beta},$$ (11)

$$dS^{\mu}=-\xi dN^{\mu}+\beta_{\lambda}dT^{\lambda\mu}-{\scriptstyle{\frac{1}{2}}}\omega_{\alpha\beta}dS^{\mu,\alpha\beta},$$ (12)

$$d(P\beta^{\mu})=N^{\mu}d\xi-T^{\lambda\mu}d\beta_{\lambda}+{\scriptstyle{\frac{1}{2}}}S^{\mu,\alpha\beta}d\omega_{\alpha\beta}.$$ (13)

Here, we have introduced the spin tensor $S^{\mu,\alpha\beta}$ defined by the expression

$$S^{\mu,\alpha\beta}=u^{\mu}S^{\alpha\beta}$$ (14)

which is an analog of the perfect-fluid forms of $N^{\mu}$ and $T^{\lambda\mu}$ given below Eq. (6).

A direct consequence of Eq. (12) is that it implies the entropy conservation for a system that conserves baryon number, energy, linear momentum and spin, namely, the conservation laws $\partial_{\mu}N^{\mu}=0$, $\partial_{\mu}T^{\mu\lambda}=0$, and $\partial_{\mu}S^{\mu,\alpha\beta}~{}=~{}0$ imply $\partial_{\mu}S^{\mu}=0$. We note that the spin conservation is a direct consequence of using a symmetric energy-momentum tensor in the considered formalism. In general, only the total angular momentum is conserved, $\partial_{\mu}J^{\mu,\alpha\beta}=0$ with $J^{\mu,\alpha\beta}=x^{\alpha}T^{\mu\beta}-x^{\beta}T^{\mu\alpha}+S^{\mu,\alpha\beta}$, which implies $\partial_{\mu}S^{\mu,\alpha\beta}=T^{\beta\alpha}-T^{\alpha\beta}$. Thus, the divergence of the spin tensor is determined by an antisymmetric part of the energy-momentum tensor that vanishes in our case.

The use of the expression (14) can be traced back to the very first model of a spinning fluid by Weyssenhoff and Raabe [35]. It was also used in Ref. [9], where the first formulation of relativistic hydrodynamics for particles with spin ${\nicefrac{{1}}{{2}}}$ was proposed. The form (14) has been subsequently used in many works that followed the methods of Israel and Stewart (positivity of the entropy current) to construct the framework of dissipative spin hydrodynamics [10, 16, 36, 37, 38, 39].

Although Eq. (14) has been used in numerous works, its form disagrees with expressions for the spin tensor obtained from the microscopic calculations [40] (and used in the spin hydrodynamics formulations that directly refer to kinetic theory [33, 41]). The latter usually lead to a more complex structures. As an extension of Eq. (14) we may consider the form

$$S^{\lambda,\mu\nu} = u^{\lambda}\left[A\left(k^{\mu}u^{\nu}-k^{\nu}u^{\mu}\right)+A_{1}t^{\mu\nu}\right]$$

$$+\frac{A_{3}}{2}\left(t^{\lambda\mu}u^{\nu}-t^{\lambda\nu}u^{\mu}+\Delta^{\lambda\mu}k^{\nu}-\Delta^{\lambda\nu}k^{\mu}\right),$$

where $A,A_{1}$ and $A_{3}$ are some scalar functions. In the case $A_{3}=0$, we reproduce Eq. (14). Moreover, for $A=A_{1}$ (with $A_{3}=0$) we find that the spin density tensor is proportional to the spin polarization tensor, namely $S^{\mu\nu}=A\,\omega^{\mu\nu}$ .

The above discussion indicates an important problem encountered in the formulations of spin hydrodynamics — a transition from Eqs. (7)–(9) to Eqs. (11)–(13) obtained by the multiplication of Eqs. (7)–(9) by the flow vector $u^{\mu}$ is inconsistent with the use of a microscopically derived spin tensor as the latter contains parts orthogonal to $u^{\mu}$. Consequently, the formulations of spin hydrodynamics that start from Eqs. (7)–(9) and use the spin tensor of the form (14) seem to be inconsistent with the formulations based on the kinetic-theory arguments. In this work we argue that the solution to the above problem lies in revising the thermodynamic relations (7)–(9). Our reasoning is supported by an analysis of a kinetic model presented below.

Insights from kinetic theory — Let us turn now to the discussion of a simple kinetic model that treats spin classically. It has been shown that for small polarization tensor the results obtained with such a model are consistent with the results obtained from the calculations using a semiclassical expansion of the Wigner function [41].

In the classical treatment of spin [43, 44], one introduces the internal angular momentum tensor $s^{\alpha\beta}$ defined in terms of the particle’s four-momentum $p$ (with $p^{\mu}p_{\mu}=m^{2}$ being the particle mass squared) and spin four-vector $s$

$$s^{\alpha\beta}=\frac{1}{m}\epsilon^{\alpha\beta\gamma\delta}p_{\gamma}s_{\delta}.$$ (16)

Equation (16) implies that $s^{\alpha\beta}=-s^{\beta\alpha}$ and $s^{\alpha\beta}p_{\beta}=0$. The spin four-vector is orthogonal to four-momentum $s\cdot p=0$, hence we can write $s^{\alpha}=1/(2m)\,\epsilon^{\alpha\beta\gamma\delta}p_{\beta}s_{\gamma\delta}$. In the particle’s rest frame (PRF), where $p^{\mu}=(m,0,0,0)$, the four-vector $s^{\alpha}$ has only space components, $s^{\alpha}=(0,{\bm{s}}_{*})$, with the normalization $|{\bm{s}}_{*}|=\,\textgoth{s:}$. For particles with spin ${\nicefrac{{1}}{{2}}}$ we use the value of the Casimir operator $\,\textgoth{s:}^{2}={\nicefrac{{1}}{{2}}}\left(1+{\nicefrac{{1}}{{2}}}\right)={\nicefrac{{3}}{{4}}}$.

The basic object used in the kinetic theory is the phase-space distribution function $f(x,{\bm{p}})$. For particles with spin, $f(x,{\bm{p}})$ is generalized to a spin dependent distribution $f(x,{\bm{p}},s)$. In local equilibrium, the spin dependent distribution functions for particles ($+$) and antiparticles ($-$) have the form 

$$f^{\pm}(x,p,s)=\exp\left(-p_{\mu}\beta^{\mu}\pm\xi+{\scriptstyle{\frac{1}{2}}}\omega_{\alpha\beta}s^{\alpha\beta}\right).$$ (17)

where $\beta^{\mu},\xi$ and $\omega_{\alpha\beta}$ are functions of space and time coordinates $x$ and play the same role as $\beta^{\mu},\xi$ and $\omega_{\alpha\beta}$ defined above. By integrating the equilibrium distribution functions over momentum and spin degrees of freedom, one obtains the macroscopic currents and tensors

$$N^{\mu}\!=\!\int dP\,dS\,p^{\mu}\,\left[f^{+}(x,p,s)-f^{-}(x,p,s)\right],$$ (18)

$$T^{\mu\nu}\!=\!\int dP\,dS\,p^{\mu}p^{\nu}\,\left[f^{+}(x,p,s)+f^{-}(x,p,s)\right],$$ (19)

$$S^{\lambda,\mu\nu}\! = \!\!\int\!dP\,dS\,\,p^{\lambda}\,s^{\mu\nu}\left[f^{+}(x,p,s)+f^{-}(x,p,s)\right].$$ (20)

Here we have introduced the integration measures in momentum, $dP=d^{3}p/((2\pi)^{3}E_{p})$, and spin space [41]

$$dS=\frac{m}{\pi\,\textgoth{s:}}\,d^{4}s\,\delta(s\cdot s+\,\textgoth{s:}^{2})\,\delta(p\cdot s),$$ (21)

with the normalization $\int dS=2$ that reflects two possible orientations of the spin ${\nicefrac{{1}}{{2}}}$.

In addition to $N^{\mu}$, $T^{\mu\nu}$, and $S^{\lambda,\mu\nu}$, we introduce the entropy current using the standard Boltzmann definition [45]

$$S^{\mu}=\!-\!\int dP\,dS\,p^{\mu}\left[f^{+}\left(\ln f^{+}\!\!-\!1\right)\!+\!f^{-}\left(\ln f^{-}\!\!-1\!\right)\right],$$ (22)

which directly leads to the formula [46]

$$S^{\mu}=T^{\mu\alpha}\beta_{\alpha}-\frac{1}{2}\omega_{\alpha\beta}S^{\mu,\alpha\beta}-\xi N^{\mu}+{\cal N}^{\mu},$$ (23)

where we have defined the particle four-current

$${\cal N}^{\mu}=\coth\xi\,\,N^{\mu}.$$ (24)

Generalized thermodynamics — We reach now the key moment of our discussion. It is important to realize that Eq. (23) has exactly the same structure as Eq. (11) except that the term $P\beta^{\mu}$ in Eq. (11) is replaced by ${\cal N}^{\mu}$. Obviously, these two tensors agree in the spinless case. Moreover, they also agree if only linear corrections in the spin polarization tensor are included, as the spin effects in both ${N}^{\mu}$ and ${\cal N}^{\mu}$ start with the quadratic terms in $k$ and $\omega$. However, in general we have $P\beta^{\mu}\neq{\cal N}^{\mu}$. Thus, Eq. (23) represents a generalization of the standard thermodynamic relation (11) to the case including the spin degrees of freedom.

Starting from the definition (24), one can derive two additional relations

$$dS^{\mu}=-\xi dN^{\mu}+\beta_{\lambda}dT^{\lambda\mu}-{\scriptstyle{\frac{1}{2}}}\omega_{\alpha\beta}dS^{\mu,\alpha\beta},$$ (25)

$$d{\cal N}^{\mu}=N^{\mu}d\xi-T^{\lambda\mu}d\beta_{\lambda}+{\scriptstyle{\frac{1}{2}}}S^{\mu,\alpha\beta}d\omega_{\alpha\beta}.$$ (26)

Equation (25) has the same form as Eq. (12), however, with the spin tensor (14) replaced by the formula (Generalized thermodynamic relations for perfect spin hydrodynamics). Equation (26) agrees with Eq. (13) only if we can again set $P\beta^{\mu}={\cal N}^{\mu}$. Equation (25) also shows that the entropy conservation is a direct consequence of three other conservation laws: for baryon number, energy, linear momentum, and spin.

A set of Eqs. (23), (25), and (26) represents our first important result. For spin polarized media, it should replace the set of Eqs. (11)–(13). Strictly speaking, Eq. (23) was derived for the first time in Ref. [46], however, in the subsequent papers only the terms linear in $\omega_{\mu\nu}$ were included that resulted in neglecting all the products (contractions) of the tensors $\omega_{\mu\nu}$ and $S^{\lambda,\mu\nu}$. One should emphasize that Eq. (23) holds for any values of the spin polarization tensor. Equation (25) was used earlier in the works that derived the form of the dissipative corrections in spin hydrodynamics. However, in this series of investigations, to maintain the products of $\omega_{\mu\nu}$ and $S^{\lambda,\mu\nu}$ in the formalism, it was assumed that $S^{\lambda,\alpha\beta}$ was of the form (14) with $S^{\alpha\beta}$ being of the zeroth order in $\omega^{\alpha\beta}$. These assumptions contradict the microscopic results which suggest the form (Generalized thermodynamic relations for perfect spin hydrodynamics) with $A_{3}\neq 0$. We conclude this part of our discussion with the statement that Eqs. (23), (25), and (26) should be used with at least second order corrections in $\omega$ to include the spin degrees of freedom in a non-trivial and consistent way .

Generalized scalar thermodynamic relations — The arguments presented above indicate that the generalized tensor thermodynamic relations include currents and tensors that contain parts orthogonal to the flow vector $u^{\mu}$. Such terms typically appear in dissipative hydrodynamics, however, in the spin hydrodynamics that may appear at the perfect-fluid (entropy conserving) level.

It becomes clear now, that such orthogonal corrections cannot appear if one starts from Eqs. (7)–(9) and multiplies them by $u^{\mu}$. Hence, a natural question arises, if there exists an analog of such scalar thermodynamic relations that is valid in the case of arbitrary $\omega^{\mu\nu}$ and for microscopically motivated spin tensor of the form (Generalized thermodynamic relations for perfect spin hydrodynamics).

In this case the baryon current $N^{\mu}$ has the structure 

$$N^{\mu}={\bar{n}}u^{\mu}+n_{t}t^{\mu},$$ (27)

where

$$t^{\mu}=t^{\mu\nu}k_{\nu}=\epsilon^{\mu\nu\alpha\beta}k_{\nu}u_{\alpha}\omega_{\beta}.$$ (28)

The four-vector $t$ is orthogonal to the vectors $u,k$ and $\omega$. In the local rest frame (LRF), where $u^{\mu}=(1,0,0,0)$, one finds that ${\bm{t}}={\bm{k}}\times{\bm{\omega}}$. The current ${\cal N}^{\mu}$ can be obtained from Eq. (24) that always holds for the Boltzmann statistics. In analogy to Eq. (27) we find

$$S^{\mu}={\bar{\sigma}}u^{\mu}+\sigma_{t}t^{\mu}.$$ (29)

The scalar functions ${\bar{n}}$ and ${\bar{\sigma}}$ depend on $\xi,T,k^{2}$ and $\omega^{2}$. In the spinless case, they reduce to standard densities depending only on $T$ and $\mu$. The “transverse” components (those with the subscript $t$) do not appear in the spinless case as they are multiplied by $t^{\mu}$ that vanishes in the limit $k,\omega\to 0$.

The energy-momentum decomposition reads

$$T^{\mu\nu} = {\bar{\varepsilon}}u^{\mu}u^{\nu}-{\bar{P}}\Delta^{\mu\nu}+P_{k}\,k^{\mu}k^{\nu}$$ (30)

$$+P_{\omega}\,\omega^{\mu}\omega^{\nu}+P_{t}\,(t^{\mu}u^{\nu}+t^{\nu}u^{\mu}),$$

where ${\bar{\varepsilon}}$ and ${\bar{P}}$ depend also on $\xi,T,k^{2}$ and $\omega^{2}$. Finally, the form of the spin tensor is given by (Generalized thermodynamic relations for perfect spin hydrodynamics), which gives

$$\frac{1}{2}\omega_{\alpha\beta}S^{\mu,\alpha\beta} = u^{\mu}(Ak^{2}-A_{1}\omega^{2})+A_{3}t^{\mu}$$ (31)

$$\equiv {\bar{s}}u^{\mu}+s_{t}t^{\mu}.$$

The use of the above decompositions in (23) leads (after comparing the terms multiplying $u^{\mu}$ and $t^{\mu}$) to two equations

$${\bar{\varepsilon}}+\coth\xi\,{\bar{n}}T = T{\bar{\sigma}}+\mu{\bar{n}}+{\bar{s}}T,$$ (32)

$$P_{t}+\coth\xi\,n_{t}T = T\sigma_{t}+\mu n_{t}+s_{t}T.$$ (33)

In the spinless case, all terms in Eq. (33) vanish, while Eq. (32) reduces to Eq. (1) – the term $\coth\xi\,{\bar{n}}T$ becomes equal to the equilibrium pressure of spinless particles.

In this way we arrive at our second main point. Our analysis shows that Eq. (7) is not an appropriate starting point for introducing thermodynamics of spin polarized media. We need at least two scalar equations, Eqs. (32) and (33), to introduce mutual relations between functions describing densities of various physical quantities.

Summary and conclusions — In this work we have introduced generalized thermodynamic relations into the framework of a relativistic perfect spin hydrodynamics. They allow for a consistent treatment of spin degrees of freedom, including the use of spin tensors whose structure follows from microscopic calculations. The obtained results are important for establishing consistency between different formulations of spin hydrodynamics. They also form the correct starting point for introducing dissipative corrections. In the future investigations, the explicit results obtained herein may serve as a reference point for comparisons among other approaches aiming at the construction of the entropy current in spin hydrodynamics, in particular, those that directly refer to the underlying quantum field theory [47].

Acknowledgements — This work was supported in part by the Polish National Science Centre Grant No. 2022/47/B/ST2/01372.

SUPLEMENTAL MATERIALS

I Tensor decomposition of the equilibrium currents

I.1 1. Baryon current

To derive Eq. (27) and find explicit forms of the coefficients ${\bar{n}}$ and $n_{t}$, we start with Eq. (18) and use the expression (17) for the spin dependent equilibrium distribution functions. Expanding the exponential function up to the second order terms in the spin polarization tensor $\omega_{\alpha\beta}$ we find

$$N^{\mu} = 2\sinh{\xi}\int dP\,p^{\mu}e^{-p\cdot\beta}\int dS\left[1+\frac{1}{2}\omega_{\alpha\beta}s^{\alpha\beta}+\frac{1}{8}(\omega_{\alpha\beta}s^{\alpha\beta})(\omega_{\sigma\gamma}s^{\sigma\gamma})+\cdots\right]$$ (34)

$$= 2\sinh{\xi}\int dP\,p^{\mu}e^{-p\cdot\beta}\left[2+\frac{1}{8}\int dS(\omega:s)(\omega:s)\right]$$

$$= 4\sinh{\xi}\int dP\,p^{\mu}e^{-p\cdot\beta}\left[\left(1+\frac{\,\textgoth{s:}^{2}}{12}\omega^{\alpha\beta}\omega_{\alpha\beta}\right)+\frac{\,\textgoth{s:}^{2}}{6m^{2}}p^{\alpha}p^{\beta}\omega^{\gamma}_{\phantom{\gamma}\alpha}\omega_{\beta\gamma}\right]$$

$$= 4\sinh{\xi}\left(1+\frac{\,\textgoth{s:}^{2}}{12}\,\omega:\omega\right)\underset{Z^{\mu}}{\underbrace{\int dP\,p^{\mu}e^{-\beta\cdot p}}}+\frac{2\textgoth{s:}^{2}\sinh{\xi}}{3m^{2}}\underset{Z^{\mu\alpha\beta}}{\underbrace{\int dP\,p^{\mu}p^{\alpha}p^{\beta}e^{-\beta\cdot p}}}\omega^{\gamma}_{\phantom{\gamma}\alpha}\omega_{\beta\gamma}.$$

Here we use the notation $\omega:s=\omega_{\alpha\beta}s^{\alpha\beta}$ and $\omega:\omega=\omega^{\alpha\beta}\omega_{\alpha\beta}=2(k^{2}-\omega^{2})$. The integral over the spin degrees of freedom is done according to the rules given in Ref. [41]. We have underlined the integrals that define the tensors $Z^{\mu}$ and $Z^{\mu\alpha\beta}$, whose explicit forms are taken from Ref. [48] and divided by $(2\pi)^{3}$, since our integration measure $dP$ includes this extra factor in the denominator. The intermediate steps of the calculations include the contractions

$$u^{\alpha}u^{\beta}u^{\nu}\omega^{\gamma}_{\phantom{\gamma}{\alpha}}\omega_{\beta\gamma}=-k^{2}u^{\nu}$$ (35)

and

$$(g^{\mu\alpha}u^{\beta}+g^{\mu\beta}u^{\alpha}+g^{\beta\alpha}u^{\mu})\omega^{\gamma}_{\phantom{\gamma}{\alpha}}\omega_{\beta\gamma}=2\omega^{\mu}_{\phantom{\mu}\gamma}k^{\gamma}-2k^{2}u^{\mu}+2\omega^{2}u^{\mu}=(2\omega^{2}-4k^{2})u^{\mu}+2t^{\mu},$$ (36)

which leads to the decomposition

$$N^{\mu}=(n_{0}+n_{2}^{k}+n_{2}^{\omega})u^{\mu}+n_{t}t^{\mu}.$$ (37)

Here the coefficients $n_{0},n_{2}^{k},n_{2}^{\omega}$, and $n_{t}$ have form:

$$n_{0}=\frac{2\sinh{\xi}}{\pi^{2}}z^{2}T^{3}K_{2}(z),\hskip 11.38092pt{n}_{2}^{k}=-\frac{2\,\textgoth{s:}^{2}\sinh{\xi}}{3\pi^{2}}zT^{3}K_{3}(z)k^{2},\hskip 11.38092pt{n}_{2}^{\omega}=-\frac{\,\textgoth{s:}^{2}\sinh{\xi}}{3\pi^{2}}zT^{3}\left[zK_{2}(z)+2K_{3}(z)\right]\omega^{2},$$ (38)

$$n_{t}=-\frac{2\,\textgoth{s:}^{2}\sinh{\xi}}{3\pi^{2}}zT^{3}K_{3}(z),$$ (39)

where $z=m/T$ and $K_{n}(z)$’s are the modified Bessel functions of the second type. The coefficients $n_{0}$ describes the baryon density of a relativistic spinless gas.

I.2 2. Particle current ${\cal N}^{\mu}$

The baryon current is the difference of the particle and antiparticle currents, $N^{\mu}=N^{\mu}_{+}-N^{\mu}_{-}$. The particle current ${\cal N}^{\mu}$ is the sum ${\cal N}^{\mu}=N^{\mu}_{+}+N^{\mu}_{-}$. Since we consider Boltzmann statistics in this work, $N^{\mu}$ and ${\cal N}^{\mu}$ are related as in Eq. (24). We note that throughout this work we consider $\mu\neq 0$ ($\xi\neq 0)$.

I.3 3. Energy-momentum tensor

When considering the energy-momentum tensor, we start with the definition (19) and use expressions for the tensors $Z^{\mu\nu}$ and $Z^{\mu\nu\alpha\beta}$ given in Ref. [48]

$$T^{\mu\nu}=4\cosh{\xi}\left(1+\frac{\,\textgoth{s:}^{2}}{12}\omega:\omega\right)\underset{Z^{\mu\nu}}{\underbrace{\int dP\,p^{\mu}p^{\nu}e^{-\beta\cdot p}}}+\frac{2\,\textgoth{s:}^{2}\cosh{\xi}}{3m^{2}}\omega^{\gamma}_{\phantom{\gamma}\alpha}\omega_{\beta\gamma}\underset{Z^{\mu\nu\alpha\beta}}{\underbrace{\int dP\,p^{\mu}p^{\nu}p^{\alpha}p^{\beta}e^{-\beta\cdot p}}}.$$ (40)

The main part of algebra involves the contraction $Z^{\mu\nu\alpha\beta}\omega^{\gamma}_{\phantom{\gamma}\alpha}\omega_{\beta\gamma}$, which contains three terms:

$$(g^{\mu\nu}g^{\alpha\beta}+g^{\mu\alpha}g^{\nu\beta}+g^{\alpha\nu}g^{\mu\beta})\omega^{\gamma}_{\phantom{\gamma}{\alpha}}\omega_{\beta\gamma}=-g^{\mu\nu}(\omega:\omega)+2\omega_{\gamma}^{\phantom{\gamma}(\mu}\omega^{\nu)\gamma}$$

$$=(4\omega^{2}-2k^{2})g^{\mu\nu}-(2k^{2}+2\omega^{2})u^{\mu}u^{\nu}-2(k^{\mu}k^{\nu}+\omega^{\mu}\omega^{\nu})+4u^{(\mu}t^{\nu)},$$

$$(g^{\mu\nu}u^{\alpha}u^{\beta}+g^{\mu\alpha}u^{\nu}u^{\beta}+g^{\alpha\nu}u^{\mu}u^{\beta}+g^{\mu\beta}u^{\alpha}u^{\nu}+g^{\beta\alpha}u^{\nu}u^{\mu}+g^{\beta\nu}u^{\mu}u^{\alpha})\omega^{\gamma}_{\phantom{\gamma}{\alpha}}\omega_{\beta\gamma}$$

$$=-k^{2}g^{\mu\nu}+(2\omega^{2}-6k^{2})u^{\mu}u^{\nu}+4u^{(\mu}t^{\nu)},$$

and

$$u^{\mu}u^{\nu}u^{\alpha}u^{\beta}\omega^{\gamma}_{\phantom{\gamma}{\alpha}}\omega_{\beta\gamma}=-k^{2}u^{\mu}u^{\nu}.$$ (41)

Here the small round brackets denote symmetrization of the indices. In this way we arrive at Eq. (30),

$$T^{\mu\nu}=(\varepsilon_{0}+{\varepsilon}_{2}^{k}+{\varepsilon}_{2}^{\omega})u^{\mu}u^{\nu}-({P}_{0}+{P}^{k}_{2}+{P}^{\omega}_{2})\Delta^{\mu\nu}+P_{k}k^{\mu}k^{\nu}+P_{\omega}\omega^{\mu}\omega^{\nu}+P_{t}(t^{\mu}u^{\nu}+t^{\nu}u^{\mu}),$$ (42)

where:

$${\varepsilon}_{0}=\frac{2\cosh{\xi}}{\pi^{2}}z^{2}T^{4}\left[zK_{3}(z)-K_{2}(z)\right],$$ (43)

$${\varepsilon}_{2}^{k}=-\frac{2\,\textgoth{s:}^{2}\cosh{(\xi)}}{3\pi^{2}}zT^{4}\left[zK_{2}(z)+5K_{3}(z)\right]k^{2},\hskip 11.38092pt{\varepsilon}_{2}^{\omega}=-\frac{\,\textgoth{s:}^{2}\cosh{(\xi)}}{3\pi^{2}}zT^{4}\left[zK_{2}(z)+(z^{2}+10)K_{3}(z)\right]\omega^{2},$$ (44)

$${P_{0}}=\frac{2\cosh{\xi}}{\pi^{2}}z^{2}T^{4}K_{2}(z),$$ (45)

$${P}^{k}_{2}=-\frac{4\,\textgoth{s:}^{2}\cosh{\xi}}{3\pi^{2}}zT^{4}K_{3}(z)k^{2},\hskip 11.38092pt{P}^{\omega}_{2}=-\frac{\,\textgoth{s:}^{2}\cosh{\xi}}{3\pi^{2}}zT^{4}\left[zK_{2}(z)+4K_{3}(z)\right]\omega^{2},$$ (46)

$$P_{t}=\frac{2\,\textgoth{s:}^{2}\cosh{\xi}}{3\pi^{2}}zT^{4}\left[K_{3}(z)-zK_{4}(z)\right],\hskip 11.38092ptP_{k}=P_{\omega}=-\frac{2\,\textgoth{s:}^{2}\cosh{\xi}}{3\pi^{2}}z\,T^{4}K_{3}(z).$$ (47)

Obviously, the quantities $\varepsilon_{0}$ and $P_{0}$ correspond to the energy density and pressure of spinless particles, respectively. We also have $P_{0}=\coth\xi\,n_{0}T$, which is the relativistic version of the Clapeyron equation.

I.4 4. Spin tensor

Finally, we consider the spin tensor $S^{\lambda,\mu\nu}$ and again express it in terms of the tensors $Z$:

$$S^{\lambda,\mu\nu} = 2\cosh{\xi}\int dPp^{\lambda}e^{-p\cdot\beta}\int dSs^{\mu\nu}\left[1+\frac{1}{2}\omega_{\alpha\beta}s^{\alpha\beta}\right]$$ (48)

$$= 2\cosh{\xi}\int dPp^{\lambda}e^{-p\cdot\beta}\,\frac{1}{2}\omega_{\alpha\beta}\int dSs^{\mu\nu}s^{\alpha\beta}$$

$$= \frac{4\,\textgoth{s:}^{2}}{3m^{2}}\cosh{\xi}\left[z^{2}T^{2}\omega^{\mu\nu}\underset{Z^{\lambda}}{\underbrace{\int dP\,p^{\lambda}e^{-\beta\cdot p}}}+\omega^{\nu}_{\phantom{\nu}\alpha}\underset{Z^{\lambda\alpha\mu}}{\underbrace{\int dP\,p^{\lambda}p^{\alpha}p^{\mu}e^{-\beta\cdot p}}}-\omega^{\mu}_{\phantom{\mu}\alpha}\underset{Z^{\lambda\alpha\nu}}{\underbrace{\int dP\,p^{\lambda}p^{\alpha}p^{\nu}e^{-\beta\cdot p}}}\right].$$

To obtain the final form, we need the explicit expression for $Z^{\lambda\alpha\mu}\omega^{\nu}_{\phantom{\nu}\alpha}-Z^{\lambda\alpha\nu}\omega^{\mu}_{\phantom{\mu}\alpha}$. In this case, it is useful to introduce the tensor

$$t^{\lambda\mu\nu}=\omega^{\nu\lambda}u^{\mu}-\omega^{\mu\lambda}u^{\nu}+g^{\lambda\mu}k^{\nu}-g^{\lambda\nu}k^{\mu}.$$ (49)

Then, the spin tensor $S^{\lambda,\mu\nu}$ can be written as

$$S^{\lambda,\mu\nu}=A_{1}u^{\lambda}\omega^{\mu\nu}+A_{2}u^{\lambda}u^{[\mu}k^{\nu]}+\frac{1}{2}A_{3}t^{\lambda\mu\nu}$$ (50)

where

$$A_{1}=\frac{2\,\textgoth{s:}^{2}\cosh{\xi}}{3\pi^{2}}zT^{3}\left[zK_{2}(z)+2K_{3}(z)\right],\hskip 19.91684ptA_{2}=\frac{4\,\textgoth{s:}^{2}\cosh{\xi}}{3\pi^{2}}z^{2}T^{3}K_{4}(z),\hskip 19.91684ptA_{3}=-\frac{4\,\textgoth{s:}^{2}\cosh{\xi}}{3\pi^{2}}zT^{3}K_{3}(z).$$ (51)

Which is consistent with the decomposition used in Ref. [46]. Moreover, with $A=A_{1}-A_{2}/2-A_{3}$ we reproduce Eq. (Generalized thermodynamic relations for perfect spin hydrodynamics).

I.5 5. Entropy current

Starting from the formula

$$S^{\mu}=T^{\mu\alpha}\beta_{\alpha}-\frac{1}{2}\omega_{\alpha\beta}S^{\mu,\alpha\beta}-\xi N^{\mu}+{\cal N}^{\mu}$$ (52)

and using the orthogonality conditions $t\cdot u=k\cdot u=\omega\cdot u=0$ as well as the contractions

$$\frac{1}{T}T^{\mu\alpha}u_{\alpha}=\frac{\bar{\varepsilon}}{T}u^{\mu}+\frac{P_{t}}{T}t^{\mu}$$ (53)

and

$$-\xi N^{\mu}+\mathcal{N}^{\mu}=(\coth(\xi)-\xi)N^{\mu}=(\coth(\xi)-\xi)\bar{n}u^{\mu}+(\coth(\xi)-\xi)n_{t}t^{\mu}$$ (54)

we find

$$\frac{1}{2}S^{\mu,\alpha\beta}\omega_{\alpha\beta}=\bar{s}u^{\mu}+s_{t}t^{\mu}.$$ (55)

The forms of $\bar{s}$ and $\bar{s}_{t}$ are given in (31). Then, we have the next form of entropy current

$$S^{\mu}={\bar{\sigma}}u^{\mu}+\sigma_{t}t^{\mu}.$$ (56)

Where coefficients hold:

$$\bar{\sigma}=\frac{\bar{\varepsilon}}{T}+(\coth\xi-\xi)\bar{n}-\bar{s},\hskip 19.91684pt\sigma_{t}=\frac{P_{t}}{T}+(\coth{\xi}-\xi)n_{t}-s_{t},$$ (57)

which agrees with Eqs. (32) and (33).