betaFeedback

WorksPhysics and Astronomy

Spin alignment of K∗K^\ast induced by strange-baryon density inhomogeneity

Abstract

The difference between the spin alignments of K∗K^\ast and those of ϕϕ at the low collision energies is a puzzle raised by the recent experiments. Unlike ϕϕ meson, K∗K^\ast, carrying a unit strange charge, should react to strange chemical potential μSμ_S. In this paper, we shall first convince you that μSμ_S is not small in a brayon-rich medium for keeping strange neutrality, and then derive the spin alignment induced by the gradient of μSμ_S, and hence of baryon chemical potential μBμ_B, using linear response theory, with the transport coefficients expressed, without any approximation, in terms of the K∗K^\ast's in-medium spectral properties by employing Ward-Takahashi identity. It turns out that such an effect applies mainly to the particles whose longitudinal and transverse modes diverge, and induces only the local spin alignment in a static medium. The magnitudes of these coefficients will be further estimated under the quasi-particle approximation.

Introduction

The strong magnetic fields and relativistic flows generated in high energy heavy-ion collisions (HICs) induce probably the most abundant spin polarization phenomena on this planet, including chiral magnetic effect [1, 2], chiral vortical effect [3, 4, 5, 6, 7, 8], shear induced polarization [9, 10], and spin Hall effect [11, 12], which open a new window probing the properties of quark-gluon plasma (QGP). Among these polarization phenomena, the tensor polarization, or spin-alignment, of the vector mesons are under active investigations and discussions.

Spin alignment, defined as δρ00≡ρ00−1/3\delta\rho_{00} \equiv\rho_{00} - 1/3 with ρ00\rho_{00} being the probability to find a vector meson in ∣0⟩|0\rangle, characterizes the tendency for the vector mesons to be in ∣±1⟩|\pm1\rangle or ∣0⟩|0\rangle. Proposed in 2005 based on quark coalescence model, the spin alignment was considered to have originated from the polarization of the constituent quarks [13, 14], and was further estimated with thermal model as δρ00∼O(ω2/T2)∼O(10−4)\delta\rho_{00} \sim O(\omega^2/T^2) \sim O(10^{-4}) in HICs [15] with ω\omega and TT representing the vorticity and temperature of the rotating fireball, respectively, under the assumption that the polarization is purely induced due to vortical effect. This estimated value is, however, much smaller than the later on measurements carried out for various vector mesons, including K∗K^*, ϕ\phi and J/ψJ/\psi, in Relativistic Heavy-Ion Collider (RHIC) and Large Hadron Collider (LHC) [16, 17, 18, 19].

Substantial efforts have been made to fix the mismatch between the theoretical estimations and experimental measurements, including improving the quark coalescence model [20], introducing an exotic fluctuating strong field [21], and connecting the spin alignment with the turbulent color fields [22] or the fluctuating glasma field in the HICs [23]. Among these theoretical efforts, the discovery of shear induced tensor polarization (SITP) was made recently via both the linear response theory [24, 25] and the quantum kinetic theory [26], and is promising to describe the data obtained in the Au-Au collisions at 200 GeV [24]. Furthermore, it is illustrated in Ref. [24] that the SITP mechanism bridges two key problems in high energy nuclear physics, i.e., the spin alignment and the in-medium spectral properties of the vector mesons, making the former a potential probe of the latter.

Besides the magnitude, the dependencies of the spin alignments on the transverse momenta (pTp_T), centralities, collisional energies and particle species are puzzling as well. The pTp_T and centrality dependencies might be promisingly described in the SITP scenario with sincere efforts [24]. In this paper, I will focus on the specie dependence of the spin alignments, i.e., the difference between the spin alignments of K∗K^* and those of ϕ\phi, which is enlarged in the low energy collisions.

Unlike ϕ\phi meson, K∗K^* carries a unit strange charge, making it reacting to the strange chemical potential μS\mu_S. Therefore, the gradient ∂(βμS)\partial(\beta\mu_S) could contribute to the spin alignments of K∗K^*, but not to those of ϕ\phi. It might make the spin alignments of both the particle species diverge. Since ∂(βμS)\partial(\beta\mu_S) is usually related to the diffusion of strange charge, I would like to call this effect diffusion induced tensor polarization (DITP). I would not conclude that the DITP effect is the key machinery leading to the difference between K∗K^* and ϕ\phi’s spin alignments. In fact, it will be pointed out in Section IV that the DITP effect only contributes to the “local” spin alignments in a static medium, and hence most probably is not the reason for the observed difference on the “global” spin alignment between the particle species. But still, it is an interesting effect. As explained later, DITP could be affected drastically by the spinodal decomposition [27, 28], which makes it a potential signal of first order phase transition that has been sought for a decade in the beam energy scan (BES) experiments [29].

Following the avenue paved in Ref. [24], I will show the derivation of DITP using linear response theory, with the corresponding transport coefficients evaluated beyond random phase approximation. By employing Ward-Takahashi identity [30], these coefficients can be expressed non-perturbatively in terms of meson’s spectral properties, and further estimated under the quasi-particle approximation. But before the derivation, I would first like to clarify shortly in the next section that μS\mu_S is NOT negligible in the low energy HICs.

Strange-Baryon Correspondence

The significance of μS\mu_S in the low energy HICs can be intuitively illustrated by the simulation on the Ar+KCl collisions at 1.76 GeV [31], where the produced kaons amount to 50 times the number of anti-kaons, which is reasonable, since in such a baryon rich medium, ss quarks are more likely to be bounded with di-quarks into hyperons, rather than paired with anti-quarks into anti-kaons. From the perspective of thermal model, a large ratio NK/NKˉ∼exp⁡(2μS/T)N_K/N_{\bar K}\sim\exp(2\mu_S/T) suggests a large μS\mu_S in these low energy HICs.

In fact, μS\mu_S is positively dependent on μB\mu_B for any strange neutral systems. In partonic phase, given that a ss quark carries 1/31/3 baryon charge and −1-1 strange charge, μS\mu_S should be equal to μB/3\mu_B/3 so that there are equal numbers of ss and sˉ\bar{s} quarks in the system. The contribution from electrical chemical potential is negligible at all collisional energies [32], and is hence neglected in the above analysis. For the strange neutral hadron gas composed of all the particle species listed in the PDG notebook [33], the dependencies of μS\mu_S on μB\mu_B under various temperatures are plotted by the solid lines in Figure 1. As shown, in all the thermal equilibrium cases, μS\mu_S increases with μB\mu_B, in consistent with the results shown in Ref. [32], and approaches the partonic limit, i.e., μB/3\mu_B/3, under high temperatures.

Figure 1. The dependence of strange chemical potential μS\mu_S on baryon chemical potential μB\mu_B for a strange neutral system either in partonic phase or in hadronic phase under various temperatures.

The strange-baryon correspondence demonstrated above suggests not only the significance of the driving force of the DITP effect, i.e., ∂(βμS)\partial(\beta\mu_S), in the low energy HICs, but also the potential connection between the DITP and the first order phase transition, since, as illustrated in Figure 1, ∂(βμS)\partial(\beta\mu_S) is positively dependent on ∂(βμB)\partial(\beta\mu_B), while the latter will be enlarged with density inhomogeneity generated during a first order phase decomposition [27, 28], which is observed and quantified by the density moments in the simulations on the HICs at the BES energies [34, 35]. It makes the DITP effect a potential novel probe of QCD first order phase transition. After all these commercials, let us look into the derivation of DITP in detail.

Derivation of DITP

The derivation of DITP is based on the linear response theory, where ∂(βμS)\partial(\beta\mu_S) plays the role of the driving force and δρ00\delta\rho_{00} is the response of the system proportional to the driving force. Before I go further, let me first express δρ00\delta\rho_{00} in terms of the field operators.

Observable

Following Ref. [24], δρ00\delta\rho_{00} is again expressed in terms of the Wigner function of the vector mesons. In detail, the single particle density matrix element ρss′(p)\rho_{ss'}(\mathbf{p}) is defined as N−1as†(p)as′(p)\mathcal{N}^{-1}a_s^\dagger(\mathbf{p})a_{s'}(\mathbf{p}) where as†a_s^\dagger and asa_s are the operators generating and annihilating a vector meson in a state ∣s⟩\lvert s\rangle, and N\mathcal{N} is the normalization factor keeping ∑sρss=1\sum_s\rho_{ss}=1. The annihilating and generating operators can be further expressed in terms of the field operators (see, e.g., Chpt. 5 in Ref. [36]) using

ϕ~μ(p)=(2π)4−322Epδ(p0−Ep)∑sϵsμ(p)as(p),(1)\tilde{\phi}^{\mu}(p)=\frac{(2\pi)^{4-\frac{3}{2}}}{\sqrt{2E_p}}\delta(p^0-E_p)\sum_s\epsilon_s^\mu(p)a_s(\mathbf{p}), \tag*{(1)}

where Ep=p2+m2E_p=\sqrt{\mathbf{p}^2+m^2} is the on-shell energy, ϕ\phi represents the particle annihilation field and ϵ\epsilon represents the on-shell covariant polarizer satisfying ϵs∗(p)⋅ϵs′(p)=−δss′\epsilon_s^*(p)\cdot\epsilon_{s'}(p)=-\delta_{ss'} and ϵs(p)⋅p~=0\epsilon_s(p)\cdot\tilde{p}=0 with p~≡(Ep,p)\tilde{p}\equiv(E_p,\mathbf{p}) representing the on-shell 4-momentum. It finally leads to

ρss′(p)=N−1ϵsμ(p)∫d3XWμν(X,p)ϵs′ν∗(p)(2)\rho_{ss'}(\mathbf{p})=\mathcal{N}^{-1}\epsilon_s^\mu(p)\int d^3XW_{\mu\nu}(X,\mathbf{p})\epsilon_{s'}^{\nu*}(p) \tag*{(2)}
Wμν(X,p)≡Ep∫dp0πd4yeip⋅yϕμ†(X−y2)ϕν(X+y2),W_{\mu\nu}(X,\mathbf{p})\equiv E_p\int\frac{dp^0}{\pi}d^4y e^{ip\cdot y}\phi_\mu^\dagger\left(X-\frac{y}{2}\right)\phi_\nu\left(X+\frac{y}{2}\right),

where WW is the Wigner function characterizing the phase space distribution of the vector mesons, with the integration on p0p^0 covering only the positive and time-like frequencies to exclude the contributions from the anti-particle sectors. The normalization factor can thus be determined, using ∑sϵsμϵsν∗≡P~μν\sum_s\epsilon_s^\mu\epsilon_s^{\nu*}\equiv\tilde{P}^{\mu\nu}, as

N=P~μν∫d3XWμν(X,p),(3)\mathcal{N}=\tilde{P}^{\mu\nu}\int d^3XW_{\mu\nu}(X,\mathbf{p}), \tag*{(3)}

where Pμν(p)=−gμν+pμpν/p2P^{\mu\nu}(p)=-g^{\mu\nu}+p^\mu p^\nu/p^2 is the projector orthogonal to momentum, satisfying P⋅p=0P\cdot p=0, Pμμ=−3P^\mu{}_\mu=-3 and P⋅P=−PP\cdot P=-P, and P~\widetilde{P} is a shorthand of P(p~)P(\widetilde{p}) satisfying P~⋅ϵs=−ϵs\widetilde{P}\cdot\epsilon_s=-\epsilon_s. Given that ϵs\epsilon_s is real for s=0s=0, the spin alignment can thus be written as

δρ00(p)=N−1ϵ0⟨μ(p)ϵ0ν⟩(p)∫d3XW(μν)(X,p).(4)\delta\rho_{00}(p)=\mathcal{N}^{-1}\epsilon_0^{\langle\mu}(p)\epsilon_0^{\nu\rangle}(p)\int d^3X W_{(\mu\nu)}(X,p). \tag*{(4)}

where the round brackets mean to make the tensor symmetric about the indices in the brackets, i.e., A(αβ)≡Aαβ/2+Aβα/2A_{(\alpha\beta)}\equiv A_{\alpha\beta}/2+A_{\beta\alpha}/2, and the angle brackets mean to make the tensor further “trace”-less, i.e., A⟨αβ⟩≡A(αβ)−P~αβP~μνAμν/3A_{\langle\alpha\beta\rangle}\equiv A_{(\alpha\beta)}-\widetilde{P}_{\alpha\beta}\widetilde{P}_{\mu\nu}A^{\mu\nu}/3, resulting in ϵ0μϵ0ν=ϵ0⟨μϵ0ν⟩−P~μν/3\epsilon_0^{\mu}\epsilon_0^{\nu}=\epsilon_0^{\langle\mu}\epsilon_0^{\nu\rangle}-\widetilde{P}^{\mu\nu}/3. (4) exhibits a general dependence of spin alignment on the Wigner function, whose expectation value can be decomposed as W=W(0)+W(1)W=W^{(0)}+W^{(1)} with W(0)W^{(0)} and W(1)W^{(1)} representing the thermal expectation value of WW and its correction proportional to the “driving force” respectively. After keeping the leading order of W(1)W^{(1)}, one obtain

δρ‾00=α0+N0−1[ϵ0⟨μϵ0ν⟩−α0P~μν]∫d3XW(μν)(1).(5)\overline{\delta\rho}_{00}=\alpha_0+\mathcal{N}_0^{-1}\left[\epsilon_0^{\langle\mu}\epsilon_0^{\nu\rangle}-\alpha_0\widetilde{P}^{\mu\nu}\right]\int d^3X W^{(1)}_{(\mu\nu)}. \tag*{(5)}

where

α0=N0−1ϵ0⟨μϵ0ν⟩∫d3XW(μν)(0),(6)\alpha_0=\mathcal{N}_0^{-1}\epsilon_0^{\langle\mu}\epsilon_0^{\nu\rangle}\int d^3X W^{(0)}_{(\mu\nu)}, \tag*{(6)}
N0=P~μν∫d3XW(μν)(0)\mathcal{N}_0=\widetilde{P}^{\mu\nu}\int d^3X W^{(0)}_{(\mu\nu)}

characterizes the zeroth order tensor polarization, originating from the in-medium divergence between the longitudinal and the transverse modes of the vector mesons, which has been proposed and studied in Ref. [24, 25], and is proven negligible at least for ϕ\phi meson [25]. In the remaining part of this paper, I shall focus on the second term in Eq. (5) and evaluate W(1)W^{(1)} originating from the gradient ∂(βμS)\partial(\beta\mu_S) using the linear response theory. But before exploring W(1)W^{(1)} with linear response theory, let me

end this subsection by mentioning the last point, i.e., in a real HIC, the spatial integration ∫d3X\int d^3X in Eqs. (4) and (5) should be replaced with the integration over the hyper-surface at freeze-out.

Linear Response

Following the approach developed by Zubarev [37, 38], W(1)W^{(1)} under an in-homogeneous strange charge distribution can be obtained as

W(1)⟨αβ⟩=−lim⁡k0→0ρWJαβ;μ(k0,k=0)k0∂μ(βμS)β.(7)W_{(1)}^{\langle\alpha\beta\rangle}=-\lim_{k^0\to0}\frac{\rho_{WJ}^{\alpha\beta;\mu}(k^0,\mathbf{k}=0)}{k^0}\frac{\partial_\mu(\beta\mu_S)}{\beta}. \tag*{(7)}
ρWJαβ;μ(k)≡∫d4xeik⋅x⟨[W(αβ)(x′+x2,p),jμ(x′−x2)]⟩.\rho_{WJ}^{\alpha\beta;\mu}(k)\equiv\int d^4x e^{ik\cdot x}\left\langle\left[W^{(\alpha\beta)}\left(x'+\frac{x}{2},\mathbf{p}\right),j^\mu\left(x'-\frac{x}{2}\right)\right]\right\rangle.

After substituting WαβW^{\alpha\beta} with the field operators using (2), and transforming the commutators in ϱ\varrho into

a Matsubara propagator via an analytical continuation (see, e.g., Chpt. 2.4 in Ref. [39]), ρ\rho can be further expressed as the three-point Green’s function illustrated in Fig. Figure 2 as

Figure 2. The Feynman diagram for calculating ρWJαβ;μ\rho_{WJ}^{\alpha\beta;\mu} in Eq. (8). The fields ϕα\phi^\alpha and ϕβ\phi^\beta are at the same time, resulting in the summation on the Matsubara frequency iωli\omega_l. The momentum flows to jμj^\mu vanishes.

ρWJαβ;μ(k)=4Ep∫d3ye−ip⋅y∫0βd4xeiωnx0−ik⋅xIm⁡⟨Tτϕ†α(x′+x2−y2)ϕβ(x′+x2+y2)jμ(x′−x2)⟩∣iωn→k0+i0+=4EpT∑lIm⁡[S~βρ(iωn+iωl,p+k2)Γ~ρσμ(iωn+iωl,p+k2;iωl,p−k2)S~σα(iωl,p−k2)]=2Ep∫C++C−dp02πiIm⁡[S~βρ(iωn+p0,p+k2)Γ~ρσμ(iωn+p0,p+k2;p0,p−k2)S~σα(p0,p−k2)]coth⁡p02T.(8)\begin{aligned} \rho_{WJ}^{\alpha\beta;\mu}(k)={}&4E_p\int d^3y e^{-i\mathbf{p}\cdot\mathbf{y}}\int_0^\beta d^4x e^{i\omega_n x^0-i\mathbf{k}\cdot\mathbf{x}}\operatorname{Im}\left\langle T_\tau\phi^{\dagger\alpha}\left(x'+\frac{x}{2}-\frac{y}{2}\right)\phi^\beta\left(x'+\frac{x}{2}+\frac{y}{2}\right)j^\mu\left(x'-\frac{x}{2}\right)\right\rangle\bigg|_{i\omega_n\to k^0+i0^+}\\ ={}&4E_pT\sum_l\operatorname{Im}\left[\widetilde{S}^{\beta\rho}\left(i\omega_n+i\omega_l,\mathbf{p}+\frac{\mathbf{k}}{2}\right)\widetilde{\Gamma}^{\mu}_{\rho\sigma}\left(i\omega_n+i\omega_l,\mathbf{p}+\frac{\mathbf{k}}{2};i\omega_l,\mathbf{p}-\frac{\mathbf{k}}{2}\right)\widetilde{S}^{\sigma\alpha}\left(i\omega_l,\mathbf{p}-\frac{\mathbf{k}}{2}\right)\right]\\ ={}&2E_p\int_{C_+ + C_-}\frac{dp^0}{2\pi i}\operatorname{Im}\left[\widetilde{S}^{\beta\rho}\left(i\omega_n+p^0,\mathbf{p}+\frac{\mathbf{k}}{2}\right)\widetilde{\Gamma}^{\mu}_{\rho\sigma}\left(i\omega_n+p^0,\mathbf{p}+\frac{\mathbf{k}}{2};p^0,\mathbf{p}-\frac{\mathbf{k}}{2}\right)\widetilde{S}^{\sigma\alpha}\left(p^0,\mathbf{p}-\frac{\mathbf{k}}{2}\right)\right]\coth\frac{p^0}{2T}. \tag*{(8)} \end{aligned}

where S~\widetilde{S} and Γ~\widetilde{\Gamma} represent the non-perturbative meson propagator and vertex for conserved current respectively, with the Matsubara frequencies ωn\omega_n and ωl\omega_l taking the discrete values, i.e., 2πnT2\pi nT and 2πlT2\pi lT. The transition from the summation over ωl\omega_l to the integration over p0p^0 at the last equality, with C+C_+ and C−C_- representing the upward and downward integration contours slightly left and right to the imaginary axis respectively, is based on the Residue’s theorem as explained in Ref. [39].

Eq. (8) is EXACT, without any approximation. The non-perturbative propagator S~\widetilde{S} can be further written in the spectral representation (see, e.g., Chpt. 2.4 of Ref. [39]) as

S~αβ(iωn,p)=∫dp02π∑a=L,TPaαβ(p)Aa(p)iωn−p0(9)\widetilde{S}^{\alpha\beta}(i\omega_n,\mathbf{p})=\int\frac{\mathrm{d}p^0}{2\pi}\sum_{a=L,T}\frac{P_a^{\alpha\beta}(\mathbf{p})\mathcal{A}_a(\mathbf{p})}{i\omega_n-p^0} \tag*{(9)}

where AL\mathcal{A}_L and AT\mathcal{A}_T are the spectral functions of the longitudinal and transverse modes of the vector mesons, and PLP_L and PTP_T defined in Eq. (27) are the longitudinal and transverse projectors respectively. In vacuum, one should expect AL=AT∝δ(p0−Ep)\mathcal{A}_L=\mathcal{A}_T\propto\delta(p^0-E_p) for stable particles. However, in a thermal medium, the spectral functions shall be broadened with their peak locations being shifted. Furthermore, AL\mathcal{A}_L and AT\mathcal{A}_T might diverge as well, with the difference vanishing for the zero-momentum particles in the medium rest frame (MRF). As demonstrated in Appendix § 1.2.1, “longitudinal” and “transverse” mean being parallel and perpendicular to the particle’s 3-momentum respectively in the MRF, so both the modes should degenerate for the particles with vanishing 3-momenta.

Under the assumption that there is no pole of p0p^0 in the vertex Γ~ρσμ(iωn+p0,p+k2;p0,p−k2)\widetilde{\Gamma}^{\mu}_{\rho\sigma}(i\omega_n+p^0,\mathbf{p}+\frac{\mathbf{k}}{2};p^0,\mathbf{p}-\frac{\mathbf{k}}{2}), the integration over p0p^0 in Eq. (8) can be conducted by collecting only the residues of the poles in S~\widetilde{S} indicated in the denominator of Eq. (9), which results in a simple relation as follows,

W(1)(αβ)=(Θαβ;μ+Υαβ;μ)∂μ(βμS)T,W_{(1)}^{(\alpha\beta)}=\left(\Theta^{\alpha\beta;\mu}+\Upsilon^{\alpha\beta;\mu}\right)\frac{\partial_\mu(\beta\mu_S)}{T},
Θαβ;μ=T2Ep∫dp0π∑a,b=LTAa(p)Ab(p)Pa(αγ)(p)Pbβ)ρ(p)[Re⁡Γ~ρσμ(p;p)]∂n∂p0,\Theta^{\alpha\beta;\mu}=T^2E_p\int\frac{\mathrm{d}p^0}{\pi}\sum_{a,b=L}^{T}\mathcal{A}_a(p)\mathcal{A}_b(p)P_a^{(\alpha\gamma)}(p)P_b^{\beta)\rho}(p)\left[\operatorname{Re}\widetilde{\Gamma}^{\mu}_{\rho\sigma}(p;p)\right]\frac{\partial n}{\partial p^0},
Υαβ;μ=T2EpP∫dωˉ dωˉ′π2∑a,b=LTPb(αγ)(ωˉ′,p)Paβ)ρ(ωˉ,p)Aa(ωˉ,p)Ab(ωˉ′,p)ωˉ−ωˉ′\Upsilon^{\alpha\beta;\mu}=T^2E_p\mathcal{P}\int\frac{\mathrm{d}\bar{\omega}\,\mathrm{d}\bar{\omega}'}{\pi^2}\sum_{a,b=L}^{T}P_b^{(\alpha\gamma)}(\bar{\omega}',p)P_a^{\beta)\rho}(\bar{\omega},p)\frac{\mathcal{A}_a(\bar{\omega},p)\mathcal{A}_b(\bar{\omega}',p)}{\bar{\omega}-\bar{\omega}'}
×lim⁡k0→012k0Im⁡[Γ~ρσμ(ωˉ,p;ωˉ−k0−i0+,p)coth⁡ωˉ2T−Γ~ρρμ(ωˉ′+k0+i0+,p;ωˉ′,p)coth⁡ωˉ′2T],(10){}\times\lim_{k^0\to0}\frac{1}{2k^0}\operatorname{Im}\left[\widetilde{\Gamma}^{\mu}_{\rho\sigma}(\bar{\omega},p;\bar{\omega}-k^0-i0^+,p)\coth\frac{\bar{\omega}}{2T}-\widetilde{\Gamma}^{\mu}_{\rho\rho}(\bar{\omega}'+k^0+i0^+,p;\bar{\omega}',p)\coth\frac{\bar{\omega}'}{2T}\right], \tag*{(10)}

where nn is the Bose distribution function and P\mathcal{P} represents evaluating the principle value of the integration. Notice that both Θ\Theta and Υ\Upsilon are dimensionless.

The contribution from Υαβ;μ\Upsilon^{\alpha\beta;\mu} in Eq. (10) vanishes if the vertex Γ~\widetilde{\Gamma} is TIME REVERSAL SYMMETRIC and ∂q0Γ~σρμ(p+q;p)∣q→0\partial_{q^0}\widetilde{\Gamma}^{\mu}_{\sigma\rho}(p+q;p)|_{q\to0} is REAL or if A(ω,p)\mathcal{A}(\omega,p) is ANALYTIC in the closed upper half-plane of ω\omega and REAL on the real axis, where the latter is similar to the assumption employed in the QCD sum rules [40]. To see the validity of the above statement, one can exchange ωˉ\bar{\omega} and ωˉ′\bar{\omega}' in the second term of Υαβ;μ\Upsilon^{\alpha\beta;\mu}, and obtains

Υαβ;μ=T2Ep∑a,b=LTP∫dωˉπPa(αγ)(ωˉ,p)Aa(ωˉ,p)P∫dωˉ′πPbβ)σ(ωˉ′,p)Ab(ωˉ′,p)ωˉ−ωˉ′\Upsilon^{\alpha\beta;\mu}=T^2E_p\sum_{a,b=L}^{T}\mathcal{P}\int\frac{\mathrm{d}\bar{\omega}}{\pi}P_a^{(\alpha\gamma)}(\bar{\omega},p)\mathcal{A}_a(\bar{\omega},p)\mathcal{P}\int\frac{\mathrm{d}\bar{\omega}'}{\pi}P_b^{\beta)\sigma}(\bar{\omega}',p)\frac{\mathcal{A}_b(\bar{\omega}',p)}{\bar{\omega}-\bar{\omega}'}
×lim⁡k0→012k0Im⁡[Γ~ρσμ(ωˉ,p;ωˉ−k0−i0+,p)+Γ~σρμ(ωˉ+k0+i0+,p;ωˉ,p)]coth⁡ωˉ2T=−T2Ep∑a,b=LTP∫dωˉπPa(αγ)(ωˉ,p)Aa(ωˉ,p)P∫dωˉ′πPbβ)σ(ωˉ′,p)Ab(ωˉ′,p)ωˉ−ωˉ′\begin{aligned} {}\times\lim_{k^0\to0}\frac{1}{2k^0}\operatorname{Im}\left[\widetilde{\Gamma}^{\mu}_{\rho\sigma}(\bar{\omega},p;\bar{\omega}-k^0-i0^+,p)+\widetilde{\Gamma}^{\mu}_{\sigma\rho}(\bar{\omega}+k^0+i0^+,p;\bar{\omega},p)\right]\coth\frac{\bar{\omega}}{2T} \\ &=-T^2E_p\sum_{a,b=L}^{T}\mathcal{P}\int\frac{\mathrm{d}\bar{\omega}}{\pi}P_a^{(\alpha\gamma)}(\bar{\omega},p)\mathcal{A}_a(\bar{\omega},p)\mathcal{P}\int\frac{\mathrm{d}\bar{\omega}'}{\pi}P_b^{\beta)\sigma}(\bar{\omega}',p)\frac{\mathcal{A}_b(\bar{\omega}',p)}{\bar{\omega}-\bar{\omega}'} \end{aligned}
×∂k0Im⁡Γ~σρμ(ωˉ+k0,p;ωˉ,p)∣k0→0coth⁡ωˉ2T.(11){}\times\left.\partial_{k^0}\operatorname{Im}\widetilde{\Gamma}^{\mu}_{\sigma\rho}(\bar{\omega}+k^0,p;\bar{\omega},p)\right|_{k^0\to0}\coth\frac{\bar{\omega}}{2T}. \tag*{(11)}

The last equality in Eq. (11) holds since Γ~ρσμ(ωˉ,p;ωˉ−k0−i0+,p)=Γ~σρμ(ωˉ−k0−i0+,p;ωˉ,p)\widetilde{\Gamma}^{\mu}_{\rho\sigma}(\bar{\omega},p;\bar{\omega}-k^0-i0^+,p)=\widetilde{\Gamma}^{\mu}_{\sigma\rho}(\bar{\omega}-k^0-i0^+,p;\bar{\omega},p) due to time reversal symmetry, and is further approximately equal to Γ~ρσμ(ωˉ+k0−i0+,p;ωˉ,p)−2k0∂q0Γ~σρμ(p+k;p)∣k→0\widetilde{\Gamma}^{\mu}_{\rho\sigma}(\bar{\omega}+k^0-i0^+,p;\bar{\omega},p)-2k^0\partial_{q^0}\widetilde{\Gamma}^{\mu}_{\sigma\rho}(p+k;p)|_{k\to0} where Γ~ρσμ(ωˉ+k0−i0+,p;ωˉ,p)\widetilde{\Gamma}^{\mu}_{\rho\sigma}(\bar{\omega}+k^0-i0^+,p;\bar{\omega},p) happens to be the complex conjugate of Γ~ρσμ(ωˉ+k0+i0+,p;ωˉ,p)\widetilde{\Gamma}^{\mu}_{\rho\sigma}(\bar{\omega}+k^0+i0^+,p;\bar{\omega},p). Eq. (11) hence vanishes if ∂q0Γ~σρμ(p+q;p)∣q→0\partial_{q^0}\widetilde{\Gamma}^{\mu}_{\sigma\rho}(p+q;p)|_{q\to0} is real, which holds at least in vacuum with pp and p+qp+q being both on-shell [41, 42].

On the other hand, the integration over ωˉ′\bar{\omega}' in Eq. (11) is condensed in the following short expression

P∫dωˉ′πPbβσ(ωˉ′,p)Ab(ωˉ′,p)ωˉ−ωˉ′,\mathcal{P}\int\frac{\mathrm{d}\bar{\omega}'}{\pi}P_b^{\beta\sigma}(\bar{\omega}',p)\frac{\mathcal{A}_b(\bar{\omega}',p)}{\bar{\omega}-\bar{\omega}'},

which, according to Kramers–Kronig relation or dispersion relation (see, e.g., Chpt. 10.8 in Ref. [36]), can be evaluated as

P∫dω′πPbβσ(ωˉ′,p)Re⁡Ab(ωˉ′,p)ωˉ−ωˉ′=−Pbβσ(ωˉ,p)Im⁡Ab(ωˉ,p),\mathcal{P}\int\frac{d\omega'}{\pi}P_b^{\beta\sigma}(\bar{\omega}',\mathbf{p})\frac{\operatorname{Re}A_b(\bar{\omega}',\mathbf{p})}{\bar{\omega}-\bar{\omega}'}=-P_b^{\beta\sigma}(\bar{\omega},\mathbf{p})\operatorname{Im}A_b(\bar{\omega},\mathbf{p}),
P∫dω′πPbβσ(ωˉ′,p)Im⁡Ab(ωˉ′,p)ωˉ−ωˉ′=Pbβσ(ωˉ,p)Re⁡Ab(ωˉ,p),\mathcal{P}\int\frac{d\omega'}{\pi}P_b^{\beta\sigma}(\bar{\omega}',\mathbf{p})\frac{\operatorname{Im}A_b(\bar{\omega}',\mathbf{p})}{\bar{\omega}-\bar{\omega}'}=P_b^{\beta\sigma}(\bar{\omega},\mathbf{p})\operatorname{Re}A_b(\bar{\omega},\mathbf{p}),

if A(ω,p)A(\omega,\mathbf{p}) is analytic in the closed upper half-plane of ω\omega, and vanishes, leading to a vanishing Υαβ;μ\Upsilon^{\alpha\beta;\mu} as well, if I further require Im⁡A(ωˉ,p)=0\operatorname{Im}A(\bar{\omega},\mathbf{p})=0, which sounds reasonable. So, in the remaining part of the paper, I will get rid of the contribution from Υ\Upsilon, and focus on the contribution

W(1)(αβ)=Θαβ;μ∂μ(βμS)T.(12)W_{(1)}^{(\alpha\beta)}=\Theta^{\alpha\beta;\mu}\frac{\partial_\mu(\beta\mu_S)}{T}. \tag*{(12)}

Non-perturbative Vertex

To further evaluate Θαβ;μ\Theta^{\alpha\beta;\mu} in Eq. (10), one need to know the detailed expression of the vertex Γ~\widetilde{\Gamma} dressed by the thermal or quantum loops. Fortunately enough, one do not need to list and calculate the Feynman diagrams for the dressed vertex order by order, rather, the vertex Γ~\widetilde{\Gamma} can be determined using Ward-Takahashi identity [30, 43].

Based on the conservation law and the canonical commutation relation, the vertex of jμj^\mu is linked with the propagator of the charged particle S~\widetilde{S} as

Γ~ρσμ(p;p)=q∂pμS~ρσ−1(p),(13)\widetilde{\Gamma}^{\mu}_{\rho\sigma}(p;p)=q\partial_{p_\mu}\widetilde{S}^{-1}_{\rho\sigma}(p), \tag*{(13)}

with qq representing the particle charge. For vector mesons,

S~μν−1(p)=−p2Pμν−m2gμν−ΠLPμνL−ΠTPμνT,(14)\widetilde{S}^{-1}_{\mu\nu}(p)=-p^2P_{\mu\nu}-m^2g_{\mu\nu}-\Pi_LP^L_{\mu\nu}-\Pi_TP^T_{\mu\nu}, \tag*{(14)}

where ΠL\Pi_L and ΠT\Pi_T represent the longitudinal and transverse self-energies respectively, which are more fundamental than the spectral functions. In fact, both ALA_L and ATA_T can be expressed in terms of the self-energies as

AL/T(p)=2Im⁡ΠL/T(p2−m2+Re⁡ΠL/T)2+Im⁡ΠL/T2.(15)A_{L/T}(p)=\frac{2\operatorname{Im}\Pi_{L/T}}{(p^2-m^2+\operatorname{Re}\Pi_{L/T})^2+\operatorname{Im}\Pi_{L/T}^2}. \tag*{(15)}

Since both ΠL\Pi_L and ΠT\Pi_T are Lorentz scalars, they should be expressed as the functions of two independent scalars, i.e., ΠL/T=ΠL/T(ϵ,ς)\Pi_{L/T}=\Pi_{L/T}(\epsilon,\varsigma) with ϵ≡u⋅p\epsilon\equiv u\cdot p and ς≡p2\varsigma\equiv p^2, for uu representing the flow velocity. In the MRF, ϵ=p0\epsilon=p^0 and κ2≡ϵ2−ς=p2\kappa^2\equiv\epsilon^2-\varsigma=\mathbf{p}^2 represent the energy and square of the 3-momentum, respectively. Hence, the momentum gradient, presenting in Eq. (13), on the self-energies are

∂pμΠL/T=∂ϵΠL/Tuμ+2∂ςΠL/Tpμ.(16)\partial_{p_\mu}\Pi_{L/T}=\partial_\epsilon\Pi_{L/T}u^\mu+2\partial_\varsigma\Pi_{L/T}p^\mu. \tag*{(16)}

General Form of DITP

After substituting Γ~ρσμ\widetilde{\Gamma}^{\mu}_{\rho\sigma} in Eq. (10) with Eq. (13), Eq. (14) and Eq. (16), and employing the last identity in Eq. (28) and then Eq. (35), one obtains

Θαβ;μ=−qT2Ep∫dp0π∑a,b=LTAaAb{[2pμ(1+∂ςRe⁡Πa)+uμ∂ϵRe⁡Πa]Paαβδab+Re⁡(ΠT−ΠL)PaασPbβρ ∂pμPTρσ}∂n∂p0=−qT2Ep∫dp0π{AL2[2pμ(1+∂ςRe⁡ΠL)+uμ∂ϵRe⁡ΠL]PLαβ+AT2[2pμ(1+∂ςRe⁡ΠT)+uμ∂ϵRe⁡ΠT]PTαβ+2ϵκ−2ALATRe⁡(ΠT−ΠL)u⊥(βP~Tα)μ}∂n∂p0.(17)\begin{aligned} \Theta^{\alpha\beta;\mu}={}&-qT^2E_p\int\frac{dp^0}{\pi}\sum_{a,b=L}^{T}A_aA_b\left\{\left[2p^\mu(1+\partial_\varsigma\operatorname{Re}\Pi_a)+u^\mu\partial_\epsilon\operatorname{Re}\Pi_a\right]P_a^{\alpha\beta}\delta_{ab}\right.\\ &\left.\qquad+\operatorname{Re}(\Pi_T-\Pi_L)P_a^{\alpha\sigma}P_b^\beta{}_\rho\,\partial_{p_\mu}P_T^{\rho}{}_\sigma\right\}\frac{\partial n}{\partial p^0}\\ ={}&-qT^2E_p\int\frac{dp^0}{\pi}\left\{A_L^2\left[2p^\mu(1+\partial_\varsigma\operatorname{Re}\Pi_L)+u^\mu\partial_\epsilon\operatorname{Re}\Pi_L\right]P_L^{\alpha\beta}\right.\\ &\left.\qquad+A_T^2\left[2p^\mu(1+\partial_\varsigma\operatorname{Re}\Pi_T)+u^\mu\partial_\epsilon\operatorname{Re}\Pi_T\right]P_T^{\alpha\beta}\right.\\ &\left.\qquad+2\epsilon\kappa^{-2}A_LA_T\operatorname{Re}(\Pi_T-\Pi_L)u_\perp^{(\beta}\widetilde{P}_T^{\alpha)\mu}\right\}\frac{\partial n}{\partial p^0}. \tag*{(17)} \end{aligned}

Let me further liberate all the tensors with indices in Eq. (17) out of the integration by transforming the off-shell tensors and vectors, including pp, u⊥u_\perp and PL/TP_{L/T}, into the on-shell ones using Eq. (30), Eq. (31) and Eq. (32). Given that the α\alpha, β\beta indices in Eq. (17) will eventually be contracted with those of the on-shell polarizers and projectors as demonstrated in Eq. (5), Θαβ;μ\Theta^{\alpha\beta;\mu} can be equivalently decomposed as

Θαβ;μ=(vLpp~μm+vLuuμ)P~Lαβ+(vTpp~μm+vTuuμ)P~Tαβ+vΔu(αP~Tβ)μ,(18)\Theta^{\alpha\beta;\mu}=\left(v_L^p\frac{\widetilde{p}^\mu}{m}+v_L^u u^\mu\right)\widetilde{P}_L^{\alpha\beta}+\left(v_T^p\frac{\widetilde{p}^\mu}{m}+v_T^u u^\mu\right)\widetilde{P}_T^{\alpha\beta}+v_{\Delta}u^{(\alpha}\widetilde{P}_T^{\beta)\mu}, \tag*{(18)}

where the dimensionless scalar transport coefficients, whose values can be evaluated in the MRF, are

ϑLp=−2qmT2Ep∫dεπAL2(1+∂ςRe⁡ΠL)∂n∂ε(1+Ω2κ2m2ς),\vartheta^{p}_{L}=-2qmT^{2}E_{p}\int\frac{\mathrm{d}\varepsilon}{\pi}\mathcal{A}_{L}^{2}(1+\partial_{\varsigma}\operatorname{Re}\Pi_{L})\frac{\partial n}{\partial\varepsilon}\left(1+\frac{\Omega^{2}\kappa^{2}}{m^{2}\varsigma}\right),
ϑTp=−2qmT2Ep∫dεπAT2(1+∂ςRe⁡ΠT)∂n∂ε,\vartheta^{p}_{T}=-2qmT^{2}E_{p}\int\frac{\mathrm{d}\varepsilon}{\pi}\mathcal{A}_{T}^{2}(1+\partial_{\varsigma}\operatorname{Re}\Pi_{T})\frac{\partial n}{\partial\varepsilon},
ϑLu=−qT2Ep∫dεπAL2(2Ω+2Ω∂ςRe⁡ΠL+∂εRe⁡ΠL)∂n∂ε(1+Ω2κ2m2ς),\vartheta^{u}_{L}=-qT^{2}E_{p}\int\frac{\mathrm{d}\varepsilon}{\pi}\mathcal{A}_{L}^{2}(2\Omega+2\Omega\partial_{\varsigma}\operatorname{Re}\Pi_{L}+\partial_{\varepsilon}\operatorname{Re}\Pi_{L})\frac{\partial n}{\partial\varepsilon}\left(1+\frac{\Omega^{2}\kappa^{2}}{m^{2}\varsigma}\right),
ϑTu=−qT2Ep∫dεπAT2(2Ω+2Ω∂ςRe⁡ΠT+∂εRe⁡ΠT)∂n∂ε,\vartheta^{u}_{T}=-qT^{2}E_{p}\int\frac{\mathrm{d}\varepsilon}{\pi}\mathcal{A}_{T}^{2}(2\Omega+2\Omega\partial_{\varsigma}\operatorname{Re}\Pi_{T}+\partial_{\varepsilon}\operatorname{Re}\Pi_{T})\frac{\partial n}{\partial\varepsilon},
ϑΔ=2qT2Ep∫dεπεκ2ALATRe⁡(ΠT−ΠL)∂n∂ε(1−εΩς),(19)\vartheta_{\Delta}=2qT^{2}E_{p}\int\frac{\mathrm{d}\varepsilon}{\pi}\frac{\varepsilon}{\kappa^{2}}\mathcal{A}_{L}\mathcal{A}_{T}\operatorname{Re}(\Pi_{T}-\Pi_{L})\frac{\partial n}{\partial\varepsilon}\left(1-\frac{\varepsilon\Omega}{\varsigma}\right), \tag*{(19)}

with Ω≡u⋅(p−p~)\Omega\equiv u\cdot(p-\widetilde{p}) characterizing the deviation of the particle’s energy p0p^{0} from its on-shell kinetic energy EpE_{p}. After substituting W(1)W^{(1)} in Eq. (5) with Eq. (12) and (18), and employing the identities in Eq. (28) and (A7), one should finally obtain the spin alignment as

δρ‾00=α0+N0−1∫d3X[(ϑsp.pp~μm+ϑsp.quμ)ϵ⟨α0ϵβ⟩0P~Lαβ+ϑΔϵα0ϵβ0u(αP~Tβ)μ−α0(ϑtt.pp~μm+ϑtt.quμ)]∂μ(βμS)T,(20)\overline{\delta\rho}_{00}=\alpha_{0}+N_{0}^{-1}\int\mathrm{d}^{3}X\left[\left(\vartheta^{p}_{\mathrm{sp.}}\frac{\widetilde{p}^{\mu}}{m}+\vartheta^{q}_{\mathrm{sp.}}u^{\mu}\right)\epsilon^{0}_{\langle\alpha}\epsilon^{0}_{\beta\rangle}\widetilde{P}_{L}^{\alpha\beta}+\vartheta_{\Delta}\epsilon^{0}_{\alpha}\epsilon^{0}_{\beta}u^{(\alpha}\widetilde{P}_{T}^{\beta)\mu}-\alpha_{0}\left(\vartheta^{p}_{\mathrm{tt.}}\frac{\widetilde{p}^{\mu}}{m}+\vartheta^{q}_{\mathrm{tt.}}u^{\mu}\right)\right]\frac{\partial_{\mu}(\beta\mu_{S})}{T}, \tag*{(20)}
ϑsp.p≡ϑLp−ϑTp,ϑtt.p≡ϑLp+2ϑTp,ϑsp.q≡ϑLq−ϑTq,ϑtt.q≡ϑLq+2ϑTq.\vartheta^{p}_{\mathrm{sp.}}\equiv\vartheta^{p}_{L}-\vartheta^{p}_{T},\qquad\vartheta^{p}_{\mathrm{tt.}}\equiv\vartheta^{p}_{L}+2\vartheta^{p}_{T},\qquad\vartheta^{q}_{\mathrm{sp.}}\equiv\vartheta^{q}_{L}-\vartheta^{q}_{T},\qquad\vartheta^{q}_{\mathrm{tt.}}\equiv\vartheta^{q}_{L}+2\vartheta^{q}_{T}.

Again, the spatial integration ∫d3X\int\mathrm{d}^{3}X should be replaced with the integration over the freeze-out hyper-surface. Notice that all the non-vanishing transport coefficients in Eq. (20), including α0\alpha_{0}, ϑsp.p/q\vartheta^{p/q}_{\mathrm{sp.}} and ϑΔ\vartheta_{\Delta}, are mainly originating from the splitting of the spectral properties between the longitudinal and the transverse modes. Hence, the DITP effect applies mainly to the vector mesons whose longitudinal and transverse modes are different.

Eq. (20) and Eq. (19) demonstrate the key discovery of this work. In the next two sections, I shall look into the tensor structures in Eq. (20) in detail in the MRF, and further estimate the transport coefficients in Eq. (19) under quasi-particle approximation.

Contribution to the Global and Local Spin Alignment

I now evaluate the contributions from each term in Eq. (20) to the “global” spin alignment, i.e., the spin alignment obtained after integrating out the particles’ 3-momentum, or at least integrating out the orientation of the 3-momentum. One should expect that some terms may vanish after the momentum (or orientation) integration due to symmetry. Let us figure out who they are.

In the MRF, under the assumption that all the transport coefficients are functions of p2\mathbf{p}^{2} near thermal equilibrium, there are three tensor structures in Eq. (20) relevant to the momentum orientation, i.e., ϵ⟨α0ϵβ⟩0P~Lαβp~⋅∂(βμS)\epsilon^{0}_{\langle\alpha}\epsilon^{0}_{\beta\rangle}\widetilde{P}_{L}^{\alpha\beta}\widetilde{p}\cdot\partial(\beta\mu_{S}), ϵ⟨α0ϵβ⟩0P~Lαβu⋅∂(βμS)\epsilon^{0}_{\langle\alpha}\epsilon^{0}_{\beta\rangle}\widetilde{P}_{L}^{\alpha\beta}u\cdot\partial(\beta\mu_{S}) and ϵα0ϵβ0u(αP~Tβ)μ∂μ(βμS)\epsilon^{0}_{\alpha}\epsilon^{0}_{\beta}u^{(\alpha}\widetilde{P}_{T}^{\beta)\mu}\partial_{\mu}(\beta\mu_{S}). The terms proportional to α0\alpha_{0} do not contribute additional tensor structures, since α0∝ϵ⟨α0ϵβ⟩0P~Lαβ\alpha_{0}\propto\epsilon^{0}_{\langle\alpha}\epsilon^{0}_{\beta\rangle}\widetilde{P}_{L}^{\alpha\beta} [24].

The polarizer ϵ0\epsilon_{0} can be expressed explicitly in the MRF as

ϵ0=1m(Epp)mΠ+(Ep−m)p^p^(0z^),(21)\epsilon_{0}=\frac{1}{m}\left(\begin{array}{c} E_{p} \\ \mathbf{p} \end{array}\right)m\Pi+\left(E_{p}-m\right)\hat{p}\hat{p}\left(\begin{array}{c} 0 \\ \hat{z} \end{array}\right), \tag*{(21)}

where p^≡p/∣p∣\hat{p}\equiv\mathbf{p}/|\mathbf{p}|, and z^\hat{z} is the spin quantization axis in the particle rest frame (PRF). Hence, in the MRF,

ϵα0ϵβ0P~Lαβp~⋅∂=[(p⋅z^)2p2−13](Ep∂t+p⋅∇),\epsilon^{0}_{\alpha}\epsilon^{0}_{\beta}\widetilde{P}_{L}^{\alpha\beta}\widetilde{p}\cdot\partial=\left[\frac{(\mathbf{p}\cdot\hat{z})^{2}}{\mathbf{p}^{2}}-\frac{1}{3}\right](E_{p}\partial_{t}+\mathbf{p}\cdot\nabla),
ϵα0ϵβ0P~Lαβu⋅∂=[(p⋅z^)2p2−13]∂t,(22)\epsilon^{0}_{\alpha}\epsilon^{0}_{\beta}\widetilde{P}_{L}^{\alpha\beta}u\cdot\partial=\left[\frac{(\mathbf{p}\cdot\hat{z})^{2}}{\mathbf{p}^{2}}-\frac{1}{3}\right]\partial_{t}, \tag*{(22)}
ϵα0ϵβ0u(αP~Tβ)μ∂μ=−(p⋅z^)2m[p^×(z^×p^)]⋅∇.\epsilon^{0}_{\alpha}\epsilon^{0}_{\beta}u^{(\alpha}\widetilde{P}_{T}^{\beta)\mu}\partial_{\mu}=-\frac{(\mathbf{p}\cdot\hat{z})^{2}}{m}\left[\hat{p}\times(\hat{z}\times\hat{p})\right]\cdot\nabla.

All the above terms vanish after the momentum or orientation integration, indicating that the DITP effect induces only the “local” spin alignment in a static medium, which is illustrated intuitively in Figure 3, where the orientation dependencies of the spin alignments in a static medium contributed by the tensor structures listed in Eq. (22) are plotted for the cases with ∇(βμS)\nabla(\beta\mu_{S}) being perpendicular (left column) and parallel (right column) to z^\hat{z} respectively. A more general orientation dependencies of the spin alignment induced by DITP in a static medium is the superposition of the patterns illustrated in Figure 3. These orientation dependencies, apparently different from those generated in the SITP [24] effect, may be regarded as the signal of the DITP effect and, as demonstrated in section II, a potential probe of the first order phase transition.

Figure 3. The momentum orientation dependence of the spin alignment induced by the DITP effect contributed by the tensor structure ϵα⟨β0P~Lαβp~0∂μ\epsilon_{\alpha\langle\beta}^{0}\widetilde{P}_{L}^{\alpha\beta}\widetilde{p}^{0}\partial_{\mu} (upper panels) and ϵαϵ0p~0P~Tαβ∂μ\epsilon_{\alpha\epsilon}^{0}\widetilde{p}^{0}\widetilde{P}_{T}^{\alpha\beta}\partial_{\mu} (lower panels) in a static and near-equilibrium medium, with the chemical potential gradient ∇(βμS)\nabla(\beta\mu_{S}) being perpendicular (left column) and parallel (right column) to the spin quantization axis z^\hat{z}, respectively, where θ\theta represents the angle between p\mathbf{p} and z^\hat{z}, and ϕ\phi represents the azimuth relative to z^\hat{z}. A more general orientation dependence of the spin alignment induced by DITP is the superposition of these patterns.

It should be noted that the above discussions apply only to cases where the spin alignments are counted in a synchronous bulk. If they are counted on a realistic freeze-out hyper-surface, which means the spatial integration in Eq. (20) is replaced by ∫dΣαpα\int d\Sigma_{\alpha}p^{\alpha} [11, 24], the DITP effect might contribute to the “global” spin alignment due to the extra orientation dependence in the integration measure.

DITP under quasi-particle approximation

Although the transport coefficients has already been given in Eq. (19), and can thus be evaluated by employing the self-energies and spectral functions of the vector mesons obtained systematically via a non-perturbative formalism such as functional renormalization group (FRG) method [44], it is still meaningful to express these coefficients in terms of the quantities with intuitive physical meanings, such as spectral width and mass-shift, under the quasi-particle approximation.

Under the quasi-particle approximation, the self-energies are modeled as

Im⁡ΠL/T∼EpΓL/T,Re⁡ΠL/T∼−2Ep2EL/T+2μSE+μS2,(23)\operatorname{Im}\Pi_{L/T} \sim E_{p}\Gamma_{L/T}, \qquad \operatorname{Re}\Pi_{L/T} \sim-2E_{p}^{2}\mathcal{E}_{L/T}+2\mu_{S}\mathcal{E}+\mu_{S}^{2}, \tag*{(23)}

with Γ\Gamma and E\mathcal{E}, depending only on p2p^{2} in the MRF, representing the width and mass-shift respectively. The terms with μS\mu_{S} originate from transforming p2→(p0+μ)2−p2p^{2}\to(p^{0}+\mu)^{2}-\mathbf{p}^{2} in the denominator of Eq. (15) in the presence of finite chemical potential. The momentum derivatives presenting in Eq. (16) reduce to

1+∂ςRe⁡Π=1−F,1+\partial_{\varsigma}\operatorname{Re}\Pi=1-\mathcal{F},
2Ω+2Ω∂ςRe⁡Π+∂ϵRe⁡Π=2(Ω+μS+EpF),2\Omega+2\Omega\partial_{\varsigma}\operatorname{Re}\Pi+\partial_{\epsilon}\operatorname{Re}\Pi=2(\Omega+\mu_{S}+E_{p}\mathcal{F}),

accordingly, with F≡2∂p2(EpE)\mathcal{F}\equiv2\partial_{p^{2}}(E_{p}\mathcal{E}). The spectral function can thus be written approximately in the Breit-Wigner form as

AL/T≈2EpΓL/T4Ep2(Ω+μS−EL/T)2+Ep2ΓL/T2,(24)\mathcal{A}_{L/T}\approx\frac{2E_{p}\Gamma_{L/T}}{4E_{p}^{2}(\Omega+\mu_{S}-\mathcal{E}_{L/T})^{2}+E_{p}^{2}\Gamma_{L/T}^{2}}, \tag*{(24)}

which satisfies

Ep∫dϵπAL/T2=1EpΓL/T,E_{p}\int\frac{d\epsilon}{\pi}\mathcal{A}_{L/T}^{2}=\frac{1}{E_{p}\Gamma_{L/T}},
Ep∫dϵπΩAL/T2=EL/T′EpΓL/T,E_{p}\int\frac{d\epsilon}{\pi}\Omega\mathcal{A}_{L/T}^{2}=\frac{\mathcal{E}'_{L/T}}{E_{p}\Gamma_{L/T}},
Ep∫dϵπΩ2AL/T2=14ΓL/T2+EL/T′2EpΓL/T,E_{p}\int\frac{d\epsilon}{\pi}\Omega^{2}\mathcal{A}_{L/T}^{2}=\frac{\frac{1}{4}\Gamma_{L/T}^{2}+\mathcal{E}_{L/T}^{\prime2}}{E_{p}\Gamma_{L/T}},
Ep∫dϵπALAT=ΓˉEp(Γˉ2+EΔ2),E_{p}\int\frac{d\epsilon}{\pi}\mathcal{A}_{L}\mathcal{A}_{T}=\frac{\bar{\Gamma}}{E_{p}(\bar{\Gamma}^{2}+\mathcal{E}_{\Delta}^{2})},
Ep∫dϵπΩALAT=Eˉ′Γˉ+14EΔΓΔEp(Γˉ2+EΔ2),E_{p}\int\frac{d\epsilon}{\pi}\Omega\mathcal{A}_{L}\mathcal{A}_{T}=\frac{\bar{\mathcal{E}}'\bar{\Gamma}+\frac{1}{4}\mathcal{E}_{\Delta}\Gamma_{\Delta}}{E_{p}(\bar{\Gamma}^{2}+\mathcal{E}_{\Delta}^{2})},
Ep∫dϵπΩ2ALAT=ΓLΓTΓˉ+2ET2ΓL+2EL2ΓT4Ep(Γˉ2+EΔ2),(25)E_{p}\int\frac{d\epsilon}{\pi}\Omega^{2}\mathcal{A}_{L}\mathcal{A}_{T}=\frac{\Gamma_{L}\Gamma_{T}\bar{\Gamma}+2\mathcal{E}_{T}^{2}\Gamma_{L}+2\mathcal{E}_{L}^{2}\Gamma_{T}}{4E_{p}(\bar{\Gamma}^{2}+\mathcal{E}_{\Delta}^{2})}, \tag*{(25)}

where Γˉ≡(ΓL+ΓT)/2\bar{\Gamma}\equiv(\Gamma_{L}+\Gamma_{T})/2 and Eˉ≡(EL+ET)/2\bar{\mathcal{E}}\equiv(\mathcal{E}_{L}+\mathcal{E}_{T})/2 represent the average spectral width and mass-shift, ΓΔ≡ΓL−ΓT\Gamma_{\Delta}\equiv\Gamma_{L}-\Gamma_{T} and EΔ≡EL−ET\mathcal{E}_{\Delta}\equiv\mathcal{E}_{L}-\mathcal{E}_{T} represent the differences on the width and mass-shift between the longitudinal and transverse modes, and E′≡E−μS\mathcal{E}'\equiv\mathcal{E}-\mu_{S}.

The following discussion is under the conjectured sequence of magnitudes that EΔ∼ΓΔ≪E≪Γˉ∼T≪m<Ep\mathcal{E}_{\Delta}\sim\Gamma_{\Delta}\ll\mathcal{E}\ll\bar{\Gamma}\sim T\ll m<E_{p}, which is justified, e.g., in the one-loop calculation on the ϕ\phi meson spectral properties [25], where the EΔ\mathcal{E}_{\Delta} and ΓΔ\Gamma_{\Delta} turn out to be negligible compared to E\mathcal{E} and Γˉ\bar{\Gamma}, and Γˉ\bar{\Gamma} is approximately half of the temperature. T≪mT\ll m is apparent given the huge K∗K^{*} mass, and leads to eEp/T≫1→n≪1e^{E_{p}/T}\gg1\to n\ll1 with nn being the Bose distribution. The spectral function given by Eq. (24) is peaked around Ω∼−μS\Omega\sim-\mu_{S}. I therefore expand the Bose distribution n(Ep+Ω)n(E_{p}+\Omega) around Ω∼−μS\Omega\sim-\mu_{S} as ∂n/∂ϵ∼[−1+(μS+Ω)/T]n∗/T\partial n/\partial\epsilon\sim[-1+(\mu_{S}+\Omega)/T]n_{*}/T with n∗≡n(Ep−μS)n_{*}\equiv n(E_{p}-\mu_{S}). Furthermore, according to Ref. [25], the mass-shift varies slowly with ∣p∣|\mathbf{p}|, which makes F≡2∂p2(EpE)∼O(E/m)≪1\mathcal{F}\equiv2\partial_{p^{2}}(E_{p}\mathcal{E})\sim\mathcal{O}(\mathcal{E}/m)\ll1.

The transport coefficients in Eq. (19) can be estimated by first expanding the integrands till the second order of Ω\Omega, conducting the integrals using Eq. (25), further expanding the integrated results in terms of O(Γˉ/Eˉ)\mathcal{O}(\bar{\Gamma}/\bar{\mathcal{E}}), O(T/m)\mathcal{O}(T/m) or the other ratios at the similar order of magnitude according to the above conjectured sequence, and keeping only the leading contributions for each transport coefficient. The results are listed below as

ϑspp≈−2qmn∗Ep[EΔΓˉ+(T−Eˉ)ΓΔΓˉ2+Γˉ2+4(Eˉ−μS)24Γˉm(μS−T)p2m3]\vartheta_{sp}^{p} \approx-\frac{2q m n_*}{E_p}\left[\frac{\mathcal{E}_{\Delta}}{\bar{\Gamma}}+\frac{(T-\bar{\mathcal{E}})\Gamma_{\Delta}}{\bar{\Gamma}^{2}}+\frac{\bar{\Gamma}^{2}+4(\bar{\mathcal{E}}-\mu_S)^{2}}{4\bar{\Gamma}m}\frac{(\mu_S-T)\mathbf{p}^{2}}{m^{3}}\right]
ϑttp≈6qn∗mEpT−EˉΓˉ\vartheta_{tt}^{p} \approx6q n_*\frac{m}{E_p}\frac{T-\bar{\mathcal{E}}}{\bar{\Gamma}}
ϑspu≈−2qn∗[Γˉ2+4(Eˉ−μS)24Γˉm(μS−T)p2m3Bˉ+μEp+T−EˉEp(Bˉ+EˉΓΔ/Γˉ−BΔ+EΔΓˉ)+Bˉ+EˉΓˉEΔEp+ΓΔ4Ep]\vartheta_{sp}^{u} \approx-2q n_*\left[\frac{\bar{\Gamma}^{2}+4(\bar{\mathcal{E}}-\mu_S)^{2}}{4\bar{\Gamma}m}\frac{(\mu_S-T)\mathbf{p}^{2}}{m^{3}}\frac{\bar{B}+\mu}{E_p}+\frac{T-\bar{\mathcal{E}}}{E_p}\left(\frac{\bar{B}+\bar{\mathcal{E}}\Gamma_{\Delta}/\bar{\Gamma}}{}-\frac{B_{\Delta}+\mathcal{E}_{\Delta}}{\bar{\Gamma}}\right)+\frac{\bar{B}+\bar{\mathcal{E}}}{\bar{\Gamma}}\frac{\mathcal{E}_{\Delta}}{E_p}+\frac{\Gamma_{\Delta}}{4E_p}\right]
ϑttu≈6qn∗[T−EˉEp(Bˉ+EˉΓˉ)−Γˉ4Ep]\vartheta_{tt}^{u} \approx6q n_*\left[\frac{T-\bar{\mathcal{E}}}{E_p}\left(\frac{\bar{B}+\bar{\mathcal{E}}}{\bar{\Gamma}}\right)-\frac{\bar{\Gamma}}{4E_p}\right]
ϑΔ≈−4n∗Ep(T−Eˉ)p2EΔΓˉ(26)\vartheta_{\Delta}\approx-\frac{4n_*E_p(T-\bar{\mathcal{E}})}{\mathbf{p}^{2}}\frac{\mathcal{E}_{\Delta}}{\bar{\Gamma}} \tag*{(26)}

where Bˉ≡2Ep∂p2(EpEˉ)\bar{B}\equiv2E_p\partial_{\mathbf{p}^{2}}(E_p\bar{\mathcal{E}}) and BΔ≡2Ep∂p2(EpEΔ)B_{\Delta}\equiv2E_p\partial_{\mathbf{p}^{2}}(E_p\mathcal{E}_{\Delta}).

One can see from Eq. (26) that O(ϑspu/ϑspp)∼O(ϑttu/ϑttp)∼O(T/m)≪1\mathcal{O}(\vartheta_{sp}^{u}/\vartheta_{sp}^{p})\sim\mathcal{O}(\vartheta_{tt}^{u}/\vartheta_{tt}^{p})\sim\mathcal{O}(T/m)\ll1, and O(ϑspu)∼O(α0ϑttp)\mathcal{O}(\vartheta_{sp}^{u})\sim\mathcal{O}(\alpha_{0}\vartheta_{tt}^{p}), O(ϑspu)∼O(α0ϑttu)\mathcal{O}(\vartheta_{sp}^{u})\sim\mathcal{O}(\alpha_{0}\vartheta_{tt}^{u}) since O(α0)∼O(EΔ/T)\mathcal{O}(\alpha_{0})\sim\mathcal{O}(\mathcal{E}_{\Delta}/T) [24]. ϑΔ\vartheta_{\Delta} is sensitive to p2\mathbf{p}^{2} and seems divergent at p2=0\mathbf{p}^{2}=0. Such a divergence is however an illusion. As demonstrated before, the longitudinal and transverse modes should degenerate for the zero momentum particles in the MRF, resulting a vanishing EΔ\mathcal{E}_{\Delta} at p2=0\mathbf{p}^{2}=0 as well. On the other hand, according to equipartition theorem for the non-relativistic particles, pˉ2∼3Tm\bar{\mathbf{p}}^{2}\sim3Tm, which makes O(ϑΔ)∼O(ϑspp)\mathcal{O}(\vartheta_{\Delta})\sim\mathcal{O}(\vartheta_{sp}^{p}) in the average sense. To conclude, under the quasi-particle approximation, the terms with the transport coefficient ϑspp\vartheta_{sp}^{p}, α0ϑttp\alpha_{0}\vartheta_{tt}^{p}, and ϑΔ\vartheta_{\Delta} are the leading contributions to DITP. All these coefficients are, however, way smaller than the leading ones in the SITP effect [24].

Meanwhile, similar to the transport coefficients in the SITP effects [24], all the coefficients in Eq. (26) are Todd, i.e., odd in the spectral width, whose physical meaning will be demonstrated in the next section.

Summary

In summary, I discover a new machinery, called diffusion induced tensor polarization (DITP), contributing to the spin alignment of the strange (or charged) vector mesons with a magnitude proportional to the gradient of the strange chemical potentials, which is proven not small in the low energy HICs. The DITP effect is derived using the linear response theory, with the transport coefficients evaluated non-perturbatively, under certain assumptions, by employing the Ward-Takahashi identity. It turns out that these non-vanishing coefficients are mostly originating from the splitting between the longitudinal and transverse spectral properties, and are estimated to be much smaller than those of the SITP effect, under the quasi-particle approximation. The tensor structure of the DITP effect indicates that it contributes only to the “local” spin alignment in a static medium, and is hence most probably elusive in the current experiments where only the “global” spin alignments are measured. However, this should not undermine the significance of the DITP effect. Since the strange chemical potential is proven positively dependent on the baryon chemical potential for maintaining strange neutrality, DITP should also increase with the gradient of the baryon chemical potential, while the latter shall be drastically enhanced during the first order phase transition. Once the “local” spin alignments are measured, the DITP effect might be a novel probe of the QCD first order phase transition.

Future efforts will be made from two aspects. Theoretically, the transport coefficients defined in Eq. (19) will be calculated with the spectral properties being evaluated using the FRG formalism. Phenomenologically, I shall generate the temperature and chemical potential distribution on the freeze-out hyper-surface from realistic simulations based on either the hydrodynamic or the transport models.

This study bridges two key and interesting problems in the high energy nuclear physics, i.e., the spin alignment of the vector mesons and the properties of the QCD phase transition, and hence opens new perspectives for improving our knowledge on the both sides.

Properties of Projectors

The longitudinal and transverse projectors defined as

PLμν≡u⊥μu⊥νu⊥2,PTμν≡Pμν−PLμν(27)P_{L}^{\mu\nu}\equiv\frac{u_{\perp}^{\mu}u_{\perp}^{\nu}}{u_{\perp}^{2}},\qquad P_{T}^{\mu\nu}\equiv P^{\mu\nu}-P_{L}^{\mu\nu} \tag*{(27)}

with u⊥≡P⋅uu_{\perp}\equiv P\cdot u for uu representing the flow velocity, take the components not only perpendicular to the particle momentum, but also parallel and perpendicular to the flow velocity, respectively, and satisfy the following identities

P⋅PL=PL⋅PL=−PL,P⋅PT=PT⋅PT=−PT,P \cdot P_L=P_L \cdot P_L=-P_L,\quad P \cdot P_T=P_T \cdot P_T=-P_T,
PL⋅PT=0,PLμμ=−1,PTμμ=−2,Pμμ=−3,P_L \cdot P_T=0,\quad P_L^{\mu}{}_{\mu}=-1,\quad P_T^{\mu}{}_{\mu}=-2,\quad P^{\mu}{}_{\mu}=-3,
p⋅P(p)=p⋅PL(p)=p⋅PT(p)=0,p\cdot P(p)=p\cdot P_L(p)=p\cdot P_T(p)=0,
P⋅u=PL⋅u=u⊥,u⋅PT=0,(28)P\cdot u=P_L\cdot u=u_{\perp},\quad u\cdot P_T=0, \tag*{(28)}
PLαβPLβγ∂pμPρσ=PLαγuLγ=⋯ ,P_L^{\alpha\beta}P_L^{\beta\gamma}\partial_{p^\mu}P_{\rho\sigma}=P_L^{\alpha\gamma}u_L^\gamma=\cdots,
PaαρPbβσ∂pμPρσ=0for a,b=L or T,P_a^{\alpha\rho}P_b^{\beta\sigma}\partial_{p^\mu}P_{\rho\sigma}=0\quad\text{for }a,b=L\text{ or }T,
PaαρPbβσ∂pμPρσL=−PaαρPbβσ∂pμPρσTfor a,b=L or T.P_a^{\alpha\rho}P_b^{\beta\sigma}\partial_{p^\mu}P^L_{\rho\sigma} =-P_a^{\alpha\rho}P_b^{\beta\sigma}\partial_{p^\mu}P^T_{\rho\sigma}\quad\text{for }a,b=L\text{ or }T.

The physical meanings of PTP_T is more transparent in the medium rest frame (MRF), where u=(1,0)u=(1,0) and

PLMRF=p02p2(v2vp^vp^p^p^),PTMRF=(0001−p^p^),(29)P_L^{\mathrm{MRF}}=\frac{p_0^2}{p^2}\begin{pmatrix} \mathbf{v}^2 & \mathbf{v}\hat{\mathbf{p}}\\ \mathbf{v}\hat{\mathbf{p}} & \hat{\mathbf{p}}\hat{\mathbf{p}}\end{pmatrix},\quad P_T^{\mathrm{MRF}}=\begin{pmatrix}0&0\\0&\mathbf{1}-\hat{\mathbf{p}}\hat{\mathbf{p}}\end{pmatrix}, \tag*{(29)}

with p^≡p/∣p∣\hat{\mathbf{p}}\equiv\mathbf{p}/|\mathbf{p}| being the unit 3-vector in the direction of p\mathbf{p}, and v≡p/p0\mathbf{v}\equiv\mathbf{p}/p^0 being the off-shell particle 3-velocity. In the MRF, PTMRFP_T^{\mathrm{MRF}} takes the spatial components perpendicular to p\mathbf{p}, this is the reason why PTP_T is regarded “transverse”.

(29) exhibits a key property of PTP_T, i.e., PTMRFP_T^{\mathrm{MRF}} does not depend on p0p^0, neither depend on the particle mass. Furthermore, since a general PTP_T is linked with PTMRFP_T^{\mathrm{MRF}} by a Lorentz boost with the frame velocity equal to uu, i.e., PT=Λ(u)⋅PTMRF⋅ΛT(u)P_T=\Lambda(u)\cdot P_T^{\mathrm{MRF}}\cdot\Lambda^T(u), it does NOT depend on the particle mass in an ARBITRARY frame either. Hence,

PT(p)=PT(p~)≡P~T,(30)P_T(p)=P_T(\tilde p)\equiv\tilde P_T, \tag*{(30)}

with p~=(Ep,p)\tilde p=(E_p,\mathbf{p}) is the on-shell 4-momentum. The connection between PL(p)P_L(p) and P~L\tilde P_L is not as simple as (30). However, if sandwiched by the on-shell polarizers or projectors, PL(p)P_L(p) and P~L\tilde P_L obey the simple relations as follows:

ϵ0αP~αβL(p)ϵ0β=ϵ0αP~αβLϵ0β(1+Ω2κ2m2p2),\epsilon_0^{\alpha}\tilde P^L_{\alpha\beta}(p)\epsilon_0^{\beta} =\epsilon_0^{\alpha}\tilde P^L_{\alpha\beta}\epsilon_0^{\beta}\left(1+\frac{\Omega^2\kappa^2}{m^2p^2}\right),
P~αβPαβL(p)=P~αβP~αβL(1+Ω2κ2m2p2),(31)\tilde P^{\alpha\beta}P^L_{\alpha\beta}(p) =\tilde P^{\alpha\beta}\tilde P^L_{\alpha\beta}\left(1+\frac{\Omega^2\kappa^2}{m^2p^2}\right), \tag*{(31)}

with Ω≡u⋅(p−p~)\Omega\equiv u\cdot(p-\tilde p), κ2≡(u⋅p~)2−m2=(u⋅p)2−p2\kappa^2\equiv(u\cdot\tilde p)^2-m^2=(u\cdot p)^2-p^2. Similarly,

ϵ0⋅u⊥(p)=−ϵ0⋅u(1−p⋅uΩp2),\epsilon_0\cdot u_{\perp}(p)=-\epsilon_0\cdot u\left(1-\frac{p\cdot u\Omega}{p^2}\right),
P~⋅u⊥(p)=−P~⋅u(1−p⋅uΩp2).(32)\tilde P\cdot u_{\perp}(p)=-\tilde P\cdot u\left(1-\frac{p\cdot u\Omega}{p^2}\right). \tag*{(32)}

Furthermore, the on-shell projectors obey the identity:

ϵ0αP~αβLϵ0β=−ϵ0αP~αβTϵ0β.(33)\epsilon_0^{\alpha}\tilde P^L_{\alpha\beta}\epsilon_0^{\beta} =-\epsilon_0^{\alpha}\tilde P^T_{\alpha\beta}\epsilon_0^{\beta}. \tag*{(33)}

In the end, let us look into the momentum derivative on PTP_T presenting in (17). Given (29), in the MRF,

PaαμPbβν∂pμPρσT=−PaαiPbβj(δiμpjp2+δjμpip2−2pipjpμp4)δμ≠0=p0(PaαμPbβ0p2+Paα0Pbβμp2−2Paα0Pbβ0p0pμp4)δμ≠0.(34)\begin{aligned} P_a^{\alpha\mu}P_b^{\beta\nu}\partial_{p^\mu}P^T_{\rho\sigma} =-P_a^{\alpha i}P_b^{\beta j}\left(\frac{\delta^\mu_i p_j}{\mathbf{p}^2}+\frac{\delta^\mu_j p_i}{\mathbf{p}^2}-2\frac{p_i p_j p^\mu}{\mathbf{p}^4}\right)\delta_{\mu\ne0} \\ &=p^0\left(\frac{P_a^{\alpha\mu}P_b^{\beta0}}{\mathbf{p}^2} +\frac{P_a^{\alpha0}P_b^{\beta\mu}}{\mathbf{p}^2} -2\frac{P_a^{\alpha0}P_b^{\beta0}p^0p^\mu}{\mathbf{p}^4}\right)\delta_{\mu\ne0}. \tag*{(34)} \end{aligned}

(34) can be extended to an arbitrary frame by conducting the following transformation: p0→u⋅pp^0\to u\cdot p, p2→κ2\mathbf{p}^2\to\kappa^2, Paμ0→Paμνuν=0P_a^{\mu0}\to P_a^{\mu\nu}u_\nu=0 (for a=Ta=T) or u⊥μu_\perp^\mu (for a=La=L), and Vμδμ≠0→ΔμνVνV^\mu\delta_{\mu\ne0}\to\Delta^{\mu\nu}V_\nu with VV representing an arbitrary 4-vector and Δμν≡gμν−uμuν\Delta^{\mu\nu}\equiv g^{\mu\nu}-u^\mu u^\nu being the flow projector that takes the components perpendicular to uu. Under these transformations, (34) are thus extended to

PTασPTβρ∂pμPρσT=0,P_T^{\alpha\sigma}P_T^{\beta\rho}\partial_{p^\mu}P^T_{\rho\sigma}=0,
PLασPTβρ∂pμPρσT=u⋅p(u⊥αPTβνΔμνκ2)=u⋅p(u⊥αP~Tβμκ2),P_L^{\alpha\sigma}P_T^{\beta\rho}\partial_{p^\mu}P^T_{\rho\sigma} =u\cdot p\left(\frac{u_\perp^\alpha P_T^{\beta\nu}\Delta^\mu{}_\nu}{\kappa^2}\right) =u\cdot p\left(\frac{u_\perp^\alpha\tilde P_T^{\beta\mu}}{\kappa^2}\right),
PTασPLβρ∂pμPρσT=u⋅p(u⊥βP~Tαμκ2),P_T^{\alpha\sigma}P_L^{\beta\rho}\partial_{p^\mu}P^T_{\rho\sigma} =u\cdot p\left(\frac{u_\perp^\beta\tilde P_T^{\alpha\mu}}{\kappa^2}\right),
PLασPLβρ∂pμPρσT=2u⋅p(uνκ2+u2 u⋅p pνκ4)PLαβΔμν=0.(35)P_L^{\alpha\sigma}P_L^{\beta\rho}\partial_{p^\mu}P^T_{\rho\sigma} =2u\cdot p\left(\frac{u^\nu}{\kappa^2}+\frac{u^2\,u\cdot p\,p^\nu}{\kappa^4}\right)P_L^{\alpha\beta}\Delta^\mu{}_\nu=0. \tag*{(35)}

Acknowledgments

The author thanks Shuai Y.F. Liu and Min He for inspiring and helpful discussions, especially for those on the analyticity of the spectral functions with Shuai Y.F. Liu. This research is supported by NSFC No. 12105129 and NSFC No. 12205090.

References

  1. [3]A. Vilenkin, Phys. Rev. D 20, 1807 (1979).
  2. [4]D. T. Son and P. Surowka, Phys. Rev. Lett. 103, 191601 (2009), arXiv:0906.5044 [hep-th].DOI
  3. [5]D. E. Kharzeev and D. T. Son, Phys. Rev. Lett. 106, 062301 (2011), arXiv:1010.0038 [hep-ph].DOI
  4. [6]J.-H. Gao, Z.-T. Liang, S. Pu, Q. Wang, and X.-N. Wang, Phys. Rev. Lett. 109, 232301 (2012), arXiv:1203.0725 [hep-ph].DOI
  5. [7]Y. Jiang, X.-G. Huang, and J. Liao, Phys. Rev. D 92, 071501 (2015), arXiv:1504.03201 [hep-ph].DOI
  6. [8]D. E. Kharzeev, J. Liao, S. A. Voloshin, and G. Wang, Prog. Part. Nucl. Phys. 88, 1 (2016), arXiv:1511.04050 [hep-ph].DOI
  7. [9]S. Y. F. Liu and Y. Yin, JHEP 07, 188 (2021), arXiv:2103.09200 [hep-ph].DOI
  8. [10]B. Fu, S. Y. F. Liu, L. Pang, H. Song, and Y. Yin, Phys. Rev. Lett. 127, 142301 (2021), arXiv:2103.10403 [hep-ph].DOI
  9. [11]S. Y. F. Liu and Y. Yin, Phys. Rev. D 104, 054043 (2021), arXiv:2006.12421 [nucl-th].DOI
  10. [12]B. Fu, L. Pang, H. Song, and Y. Yin, (2022), arXiv:2208.00430 [hep-ph].DOI
  11. [13]Z.-T. Liang and X.-N. Wang, Phys. Lett. B 629, 20 (2005), arXiv:nucl-th/0411101.DOI
  12. [14]Z.-T. Liang and X.-N. Wang, Phys. Rev. Lett. 94, 102301 (2005), [Erratum: Phys.Rev.Lett. 96, 039901 (2006)], arXiv:nucl-th/0410079.DOI
  13. [15]F. Becattini, F. Piccinini, and J. Rizzo, Phys. Rev. C 77, 024906 (2008), arXiv:0711.1253 [nucl-th].DOI
  14. [16]B. I. Abelev et al. (STAR), Phys. Rev. C 77, 061902 (2008), arXiv:0801.1729 [nucl-ex].
  15. [17]S. Acharya et al. (ALICE), Phys. Rev. Lett. 125, 012301 (2020), arXiv:1910.14408 [nucl-ex].
  16. [18]S. Singha (STAR), Phys. As. A 1005, 121733 (2021), arXiv:2002.07427 [nucl-ex].DOI
  17. [19]M. S. Abdallah et al. (STAR), Nature 614, 244 (2023), arXiv:2204.02302 [hep-ph].DOI
  18. [20]X.-L. Sheng, Q. Wang, and X.-N. Wang, Phys. Rev. D 102, 056013 (2020), arXiv:2007.05106 [nucl-th].DOI
  19. [21]X.-L. Sheng, L. Oliva, Z.-T. Liang, Q. Wang, and X.-N. Wang, Phys. Rev. Lett. 131, 042304 (2023), arXiv:2205.15689 [nucl-th].DOI
  20. [22]A. Kumar, B. Müller, and D.-L. Yang, Phys. Rev. D 107, 076025 (2023), arXiv:2212.13354 [nucl-th].DOI
  21. [23]A. Kumar, B. Müller, and D.-L. Yang, Phys. Rev. D 108, 016020 (2023), arXiv:2304.04181 [nucl-th].DOI
  22. [24]F. Li and S. Y. F. Liu, (2022), arXiv:2206.11890 [nucl-th].arxiv.org/abs/2206.11890
  23. [25]W.-B. Dong, Y.-L. Yin, X.-L. Sheng, S.-Z. Yang, and Q. Wang, (2023), arXiv:2311.18400 [hep-ph].arxiv.org/abs/2311.18400
  24. [26]D. Wagner, N. Weickgennant, and E. Speranza, Phys. Rev. Res. 5, 03187 (2023), arXiv:2207.01111 [nucl-th].arxiv.org/abs/2207.01111
  25. [27]F. Li and C. M. Ko, Phys. Rev. C 93, 035205 (2016), arXiv:1601.00026 [nucl-th].DOI
  26. [28]F. Li and C. M. Ko, Phys. Rev. C 95, 055203 (2017), arXiv:1606.05012 [nucl-th].DOI
  27. [29]J. Adam and et al., (2014), https://drupal.star.bnl.gov/STAR/starnotes/public/sn0598.
  28. [30]J. C. Ward, Phys. Rev. 78, 182 (1950).
  29. [31]F. Li, L.-W. Chen, C. M. Ko, and S. H. Lee, Phys. Rev. C 85, 064902 (2012), arXiv:1204.1327 [nucl-th].DOI
  30. [32]S. Bhattacharyya, D. Biswas, S. K. Ghosh, R. Ray, and P. Singha, Phys. Rev. D 101, 054002 (2020), arXiv:1911.04828 [hep-ph].DOI
  31. [33]R. L. Workman et al. (Particle Data Group), PTEP 2022, 083C01 (2022).
  32. [34]K.-J. Sun, C. M. Ko, F. Li, J. Xu, and L.-W. Chen, Eur. Phys. J. A 57, 313 (2021), arXiv:2006.08929 [nucl-th].DOI
  33. [35]K.-J. Sun, W.-H. Zhou, L.-W. Chen, C. M. Ko, F. Li, R. Wang, and J. Xu, (2022), arXiv:2205.11010 [nucl-th].arxiv.org/abs/2205.11010
  34. [36]S. Weinberg, The Quantum theory of fields. Vol. 1: Foundations (Cambridge University Press, 2005).
  35. [37]D. N. Zubarev, Nonequilibrium statistical thermodynamics (Consultants Bureau, 1974).
  36. [38]A. Hosoya, M.-a. Sakagami, and M. Takao, Annals Phys. 154, 229 (1984).
  37. [39]M. L. Bellac, Thermal Field Theory, Cambridge Monographs on Mathematical Physics (Cambridge University Press, 2011).
  38. [40]P. Colangelo and A. Khodjamirian, , 1495 (2000), arXiv:hep-ph/0010175.DOI
  39. [41]M. S. Bhagwat and P. Maris, Phys. Rev. C 77, 025203 (2008), arXiv:nucl-th/0612069.DOI
  40. [42]Y.-Z. Xu, D. Binosi, Z.-F. Cui, B.-L. Li, C. D. Roberts, S.-S. Xu, and H. S. Zong, Phys. Rev. D 100, 114038 (2019), arXiv:1911.05199 [nucl-th].DOI
  41. [43]Y. Takahashi, II Nuovo Cimento (1955-1965) 6, 371 (1957).
  42. [44]C. Jung, F. Rennecke, R.-A. Tripolt, L. von Smekal, and J. Wambach, Phys. Rev. D 95, 036020 (2017), arXiv:1610.08754 [hep-ph].DOI

Paper details

vlnc:W_e56dasat
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 6

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents