betaFeedback

WorksPhysics and Astronomy

Local Spin Polarization in Anisotropic Gubser Flow: Suppression Mechanism and Formulation Dependence

Abstract

We analytically study the longitudinal spin polarization in relativistic heavy-ion collisions using a perturbed Gubser flow solution. In the large-system-size limit, we derive analytical expression of the local spin polarization along the beam direction. In our treatment, the contributions from thermal vorticity and thermal shear are of comparable magnitudes. The thermal vorticity yields the polarization with a sign opposite to that observed experimentally, while the thermal shear counteracts this effect, helping recover the desired sign. We find that the choice of the reference unit vector aligned with the fluid velocity gives the experimentally observed sign only at low transverse momenta, whereas another formulation with the unit vector fixed along the laboratory time direction yields the desired sign for a wide range of transverse momenta. Notably, a recent formulation with the unit vector normal to the freeze-out hypersurface exhibits an exact cancellation between contributions from thermal vorticity and thermal shear at leading order in the large-system-size limit. We identify a general cancellation pattern with acceleration dominance, which is manifest particularly in the latter two formulations. Thus, the total polarization originates from non-acceleration effects, which need not be substantial even when the elliptic flow is finite, as clearly demonstrated in our analytical results. For comparison, we also discuss the spin polarization in the Hubble flow with rotation.

Introduction

In noncentral relativistic heavy-ion collisions, the quark-gluon plasma (QGP) contains a large orbital angular momentum oriented perpendicular to the reaction plane. Part of the orbital angular momentum is subsequently transferred to the spin of produced particles, e.g., Λ\Lambda hyperons, causing an average polarization along the direction of the orbital angular momentum [1, 2, 3]. This phenomenon is known as global spin polarization. Experimental measurements from STAR Collaboration [4], ALICE Collaboration [5], and HADES Collaboration [6] have already confirmed the existence of the global spin polarization. Theoretically, the experimental data can be well described by various phenomenological models [7, 8, 9, 10, 11, 12, 13, 14, 15] using the modified Cooper-Frye formula with the thermal vorticity [16, 17], confirming the mutual conversion between the orbital angular momentum of QGP and the spin of produced particles.

However, the successful theoretical description encounters challenges when extended to local spin polarization, which is a function of the produced particle’s momentum. Along the beam direction, the local spin polarization induced by elliptic flow and triangular flow has already been measured [18, 19, 20, 21]. Compared to the experimental results, theoretical calculations based on thermal vorticity alone give an opposite sign in the azimuthal-angle dependence [22, 23, 24, 25, 26, 27]. This discrepancy, called the sign puzzle, implies that the original method based on thermal vorticity alone is insufficient to describe the local observable. The puzzle has stimulated substantial theoretical activity, accelerating the development of spin physics in heavy-ion collisions. For example, quantum kinetic theory [28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42] and relativistic spin hydrodynamics [43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59] have witnessed rapid development in recent years.

To address the sign puzzle, theoretical frameworks have been extended by incorporating the thermal shear tensor into the modified Cooper-Frye formula [31, 60, 61]. The subsequent numerical simulations [62, 63, 64] have revealed that the polarization induced by thermal shear has a sign opposite to the polarization from thermal vorticity, implying a cancellation when adding them together. Under certain conditions, the thermal shear contribution exceeds that of the thermal vorticity, thereby recovering the correct sign of the local spin polarization [14, 15, 62, 63, 64, 65, 66, 67, 68, 69]. For a comprehensive overview, we refer readers to the review articles [38, 70, 71].

Currently, several distinct formulations have been proposed to quantify the thermal shear effect on local spin polarization. The main difference among them stems from the choice of the unit vector in the thermal shear term. Specifically, the Fu-Liu-Pang-Song-Yin (FLPSY) formulation [60, 62] aligns this vector with the fluid velocity, whereas the Becattini-Buzzegoli-Palermo-Inghirami-Karpenko (BBPIK) formulation [61, 63] sets it to be the time direction in the laboratory frame. The Sheng-Becattini-Roselli (SBR) formulation [72] defines the vector as the normal to the freeze-out hypersurface. While

angle dependence [22, 23, 24, 25, 26, 27]. This discrepancy, called the sign puzzle, implies that the original method based on thermal vorticity alone is insufficient to describe the local observable. The puzzle has stimulated substantial theoretical activity, accelerating the development of spin physics in heavy-ion collisions. For example, quantum kinetic theory [28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42] and relativistic spin hydrodynamics [43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59] have witnessed rapid development in recent years.

To address the sign puzzle, theoretical frameworks have been extended by incorporating the thermal shear tensor into the modified Cooper-Frye formula [31, 60, 61]. The subsequent numerical simulations [62, 63, 64] have revealed that the polarization induced by thermal shear has a sign opposite to the polarization from thermal vorticity, implying a cancellation when adding them together. Under certain conditions, the thermal shear contribution exceeds that of the thermal vorticity, thereby recovering the correct sign of the local spin polarization [14, 15, 62, 63, 64, 65, 66, 67, 68, 69]. For a comprehensive overview, we refer readers to the review articles [38, 70, 71].

Currently, several distinct formulations have been proposed to quantify the thermal shear effect on local spin polarization. The main difference among them stems from the choice of the unit vector in the thermal shear term. Specifically, the Fu-Liu-Pang-Song-Yin (FLPSY) formulation [60, 62] aligns this vector with the fluid velocity, whereas the Becattini-Buzzegoli-Palermo-Inghirami-Karpenko (BBPIK) formulation [61, 63] sets it to be the time direction in the laboratory frame. The Sheng-Becattini-Roselli (SBR) formulation [72] defines the vector as the normal to the freeze-out hypersurface. While these formulations share the common goal, they lead to quantitatively different predictions. A systematic comparison of these formulations is therefore warranted. To date, numerical comparisons of the first two formulations have been conducted in some cases [69, 73]. However, it is still challenging to uncover underlying mechanisms from the numerical results.

In contrast, analytical studies can explicitly show how the local spin polarization scales with initial conditions and transverse momentum, as well as the cancellation between thermal shear and thermal vorticity effects. At present, the existing analytical studies addressing local spin polarization are based on the modified blast-wave model [74, 75], in which the dynamical variables are decoupled from the hydrodynamic equations. To the best of our knowledge, there are currently no analytic studies of spin polarization derived directly from hydrodynamic solutions.

In this work, we fill this gap by studying longitudinal spin polarization using a perturbed analytical solution of relativistic hydrodynamics. Specifically, we will use an anisotropic extension of the Gubser flow that incorporates both elliptic and triangular geometric deformations [76, 77, 78, 79]. We will derive explicit analytical expressions for the spin polarization for three formulations. This analytical approach enables us to systematically examine the parameter dependence and perform a direct comparison among these formulations. We aim to clarify the conditions under which the desired sign of local spin polarization is recovered, and to identify potential cancellation between thermal shear and thermal vorticity effects.

The paper is organized as follows. In Sec. II, we present several analytical hydrodynamic solutions. In Sec. III, we review the three formulations for spin polarization. In Sec. IV, we calculate the spin polarization in the rotating Hubble flow. In Sec. V, we analyze the longitudinal spin polarization in the perturbed Gubser flow. Finally, we summarize our findings and discuss their implications in Sec. VI. Throughout this work, we adopt the metric gμν=diag⁡{+,−,−,−}g_{\mu\nu} = \operatorname{diag}\{+,-,-,-\} for flat space-time.

Analytical Fluid Solutions

We focus on ideal relativistic hydrodynamics without charge currents for simplicity. We can generalize our approach to viscous hydrodynamics if needed. In the ideal case, the energy-momentum tensor is decomposed as

Tμν=euμuν−pΔμν,(1)T^{\mu\nu} = e u^\mu u^\nu- p\Delta^{\mu\nu}, \tag*{(1)}

where uμu^\mu, ee, and pp represent the fluid velocity, the energy density, and the thermodynamic pressure, respectively. In our convention the projection tensor is defined as Δμν=gμν−uμuν\Delta^{\mu\nu} = g^{\mu\nu} - u^\mu u^\nu. The energy-momentum tensor satisfies the conservation law:

∇μTμν=0,(2)\nabla_\mu T^{\mu\nu} = 0, \tag*{(2)}

and the equation of state is required to close the system of equations. Here, we adopt the conformal equation of state, i.e.,

e=3p=λT4,(3)e = 3p = \lambda T^4, \tag*{(3)}

where λ\lambda is a dimensionless constant and TT is the fluid temperature.

We are constructing the analytical solutions of relativistic hydrodynamics to approximate the QGP in heavy-ion collisions. The rotating Hubble flow and the perturbed Gubser flow are two representative examples, as schematically illustrated in Figure 1. The former is a simple solution with spherical expansion and rigid rotation at finite angular velocity, ω⃗\vec{\omega}, qualitatively corresponding to global polarization, while the latter is a solution of our present interest with longitudinal and transverse expansion, including a nonzero elliptic flow, v2v_2. In the subsequent sections, we will explain these two solutions in order.

Figure 1. Schematic illustration of the rotating Hubble flow (left) and the perturbed Gubser flow (right). The rotating Hubble flow comprises spherical expansion (red arrows) with rigid rotation (blue arrows) about the yy-axis. The perturbed Gubser flow builds upon the original Gubser flow (red arrows) by introducing additional anisotropy (blue arrows) in the xyxy-plane.

Rotating Hubble Flow

For spherically expanding systems, the Hubble flow is a simple solution of the relativistic hydrodynamic equations [80]. This solution can be generalized to include rotation as [81]

(ut,u⃗)=γω(t,x⃗+ω⃗×x⃗),T=T0γω,(4)(u^t,\vec{u}) = \gamma_\omega(t,\vec{x}+\vec{\omega}\times\vec{x}), \qquad T = T_0\gamma_\omega, \tag*{(4)}

where ω⃗\vec{\omega} is the angular velocity vector, T0T_0 is a dimensionless constant determined by the initial condition, and γω=[t2−(x⃗+ω⃗×x⃗)2]−1/2\gamma_\omega= [t^2-(\vec{x}+\vec{\omega}\times\vec{x})^2]^{-1/2}. The velocity profile captures a spherically symmetric expansion coupled with rigid rotation along the axis parallel to ω⃗\vec{\omega}. In the ω⃗→0\vec{\omega}\to0 limit, Eq. (4) reduces to the standard Hubble flow. This exact solution provides an analytical toy model for relativistic fluids with rigid rotation and can be applied, for example, in the cosmological context [82]. Also in the heavy-ion collision, the Hubble flow is a qualitatively reasonable description as long as the collision energy is low, while higher energy collisions would produce a fireball expanding predominantly along the beam direction with approximate boost invariance.

Bjorken Flow

For boost-invariant and cylindrically symmetric systems, the Bjorken flow [83] is the best-known. To describe the Bjorken flow, it is natural to work in Milne coordinates (τ,r,ϕ,η)(\tau, r, \phi, \eta), where τ=t2−z2\tau= \sqrt{t^2-z^2} is the longitudinal proper time, r=x2+y2r = \sqrt{x^2+y^2} is the transverse radius, ϕ=tan⁡−1(y/x)\phi= \tan^{-1}(y/x) is the azimuthal angle, and η=tanh⁡−1(z/t)\eta= \tanh^{-1}(z/t) is the rapidity. The corresponding metric reads:

ds2=dτ2−dr2−r2dϕ2−τ2dη2.(5)ds^2 = d\tau^2 - dr^2 - r^2 d\phi^2 - \tau^2 d\eta^2 . \tag*{(5)}

In Milne coordinates, the velocity profile of the Bjorken flow takes a simple form [83]:

uμ=(1,0),T=T0τ−1.(6)u^\mu= (1,0), \qquad T = T_0\tau^{-1} . \tag*{(6)}

Although the Bjorken flow is useful to approximate bulk properties of the hydrodynamic evolutions at high enough energy, it is too simple to discuss spin polarization, as the (1+1)(1+1)-dimensional solution inevitably yields zero spin polarization. To obtain a nontrivial result, we should consider more realistic hydrodynamic solutions. Since transverse expansion is crucial for local spin polarization phenomena, we need to include the transverse dynamics, which has been implemented in the Gubser flow [76, 77]. For our purpose, moreover, it is also indispensable to introduce transverse anisotropy, which has been discussed perturbatively [78]. We will explain constructions of these extensions in order.

Gubser Flow

The Gubser flow is a conformal solution with longitudinal and transverse expansion. It is convenient to formulate it in terms of x^μ\hat{x}^\mu on dS3×RdS_3 \times\mathbb{R}. The Minkowski spacetime, R3,1\mathbb{R}^{3,1}, is transformed to dS3×RdS_3 \times\mathbb{R} by a Weyl rescaling, so that the line element in Eq. (5) is deformed as

ds^2=τ−2ds2=dρ2−cosh⁡2ρ(dθ2+sin⁡2θdϕ2)−dη2.(7)d\hat{s}^2 = \tau^{-2}ds^2 = d\rho^2 - \cosh^2\rho(d\theta^2+\sin^2\theta d\phi^2) - d\eta^2 . \tag*{(7)}

where ρ\rho and θ\theta are defined via

sinh⁡ρ=−L2−τ2+r22Lτ,tan⁡θ=2LrL2+τ2−r2.(8)\sinh\rho= -\frac{L^2-\tau^2+r^2}{2L\tau}, \qquad\tan\theta= \frac{2Lr}{L^2+\tau^2-r^2} . \tag*{(8)}

We will explain the meaning of the parameter LL later. With this Weyl rescaling, the temperature and fluid velocity transform as

T^=τT,u^μ=1τ∂xν∂x^μuν,(9)\hat{T} = \tau T, \qquad\hat{u}_\mu= \frac{1}{\tau}\frac{\partial x^\nu}{\partial\hat{x}^\mu}u_\nu, \tag*{(9)}

where x^μ=(ρ,θ,ϕ,η)\hat{x}^\mu=(\rho,\theta,\phi,\eta) and xμ=(τ,r,ϕ,η)x^\mu=(\tau,r,\phi,\eta). The conservation equation Eq. (2) with the equation of state Eq. (3) leads to

T^u^μ∇^μu^α−Δ^αμ∇^μT^=0,(10)\hat{T}\hat{u}^\mu\hat{\nabla}_\mu\hat{u}^\alpha-\hat{\Delta}^{\alpha\mu}\hat{\nabla}_\mu\hat{T}=0 , \tag*{(10)}
u^μ∇^μT^+13T^∇^μu^μ=0.(11)\hat{u}^\mu\hat{\nabla}_\mu\hat{T}+\frac{1}{3}\hat{T}\hat{\nabla}_\mu\hat{u}^\mu=0 . \tag*{(11)}

The projection tensor acting on x^μ\hat{x}^\mu is Δ^μν=g^μν−u^μu^ν\hat{\Delta}^{\mu\nu}=\hat{g}^{\mu\nu}-\hat{u}^\mu\hat{u}^\nu with the metric g^μν\hat{g}^{\mu\nu} deduced from Eq. (7). We note that ∇^μ\hat{\nabla}_\mu is the covariant derivative defined in terms of g^μν\hat{g}^{\mu\nu}.

The fluid velocity profile in the Gubser flow is given in the x^μ\hat{x}^\mu coordinates by

(u^ρ,u^θ,u^ϕ,u^η)=(1,0,0,0).(12)(\hat{u}_\rho,\hat{u}_\theta,\hat{u}_\phi,\hat{u}_\eta)=(1,0,0,0) . \tag*{(12)}

This solution trivially satisfies Eq. (10a), while Eq. (10b) results in

T^=T^0(cosh⁡ρ)−2/3,(13)\hat{T}=\hat{T}_0(\cosh\rho)^{-2/3} , \tag*{(13)}

where T^0\hat{T}_0 is determined by the initial condition.

We can grasp an intuitive picture by transforming the solution back to Minkowski spacetime using Eq. (9). The profile in the Milne coordinates becomes

(uτ,ur,uϕ,uη)=(L2+τ2+r22Lτcosh⁡ρ,−rLcosh⁡ρ,0,0),(14)(u_\tau,u_r,u_\phi,u_\eta)=\left(\frac{L^2+\tau^2+r^2}{2L\tau\cosh\rho},\frac{-r}{L\cosh\rho},0,0\right) , \tag*{(14)}

from which we see that the fluid velocity has both longitudinal and transverse radial expansion. It is easy to confirm that the large-LL limit reduces the Gubser flow to the Bjorken flow as cosh⁡ρ→L/(2τ)\cosh\rho\to L/(2\tau) for L→∞L\to\infty. It should be emphasized that the expansion of the Gubser flow is still isotropic on the transverse plane.

The above expression of uru_r implies that LL is a typical scale along the rr direction. Indeed, the rescaled energy density behaves as ϵ^∝(cosh⁡ρ)−8/3\hat{\epsilon}\propto(\cosh\rho)^{-8/3}. In the early time regime τ≈0\tau\approx0, the energy density ϵ^\hat{\epsilon} decays along the rr direction as

ϵ^(r=L)ϵ^(r=0)≈0.16.(15)\frac{\hat{\epsilon}(r=L)}{\hat{\epsilon}(r=0)}\approx0.16 . \tag*{(15)}

Thus, the parameter LL characterizes the initial transverse size of the QGP, inside of which the energy density is localized.

Perturbed Gubser Flow

We need to break transverse isotropy to describe local spin polarization phenomena. For this purpose, we construct an approximate solution with perturbative anisotropy on top of the Gubser flow. Following Refs. [77, 78], we consider the perturbed solution parametrized as

(u^ρ,u^θ,u^ϕ,u^η)=(1,vθ(ρ,θ,ϕ),vϕ(ρ,θ,ϕ),0)(16)(\hat{u}_{\rho},\hat{u}_{\theta},\hat{u}_{\phi},\hat{u}_{\eta})=(1,v_{\theta}(\rho,\theta,\phi),v_{\phi}(\rho,\theta,\phi),0) \tag*{(16)}

with ∣vθ,ϕ∣≪1|v_{\theta,\phi}|\ll1 for the fluid velocity, and the local temperature is modified accordingly as

T^(ρ,θ,ϕ)=T^0(cosh⁡ρ)−2/3[1+T(ρ,θ,ϕ)](17)\hat{T}(\rho,\theta,\phi)=\hat{T}_{0}(\cosh\rho)^{-2/3}[1+\mathcal{T}(\rho,\theta,\phi)] \tag*{(17)}

with ∣T∣≪1|\mathcal{T}|\ll1. Substituting them into the conservation equations (10), we obtain:

3∂ρvθ−2vθtanh⁡ρ−3∂θT=0(18)3\partial_{\rho}v_{\theta}-2v_{\theta}\tanh\rho-3\partial_{\theta}\mathcal{T}=0 \tag*{(18)}
3∂ρvϕ−2vϕtanh⁡ρ−3∂ϕT=0(19)3\partial_{\rho}v_{\phi}-2v_{\phi}\tanh\rho-3\partial_{\phi}\mathcal{T}=0 \tag*{(19)}
3∂ρT−sech⁡2ρ(∂θvθ+vθcot⁡θ+∂ϕvϕcsc⁡2θ)=0(20)3\partial_{\rho}\mathcal{T}-\operatorname{sech}^{2}\rho(\partial_{\theta}v_{\theta}+v_{\theta}\cot\theta+\partial_{\phi}v_{\phi}\csc^{2}\theta)=0 \tag*{(20)}

up to the linear order in the perturbative expansion. We can analytically solve the linear differential equations (17) using the method of separation of variables. The general solution takes the following form:

vθ=∑l=0∞χ(l)(ρ)∂θX(l)(θ,ϕ)(21)v_{\theta}=\sum_{l=0}^{\infty}\chi^{(l)}(\rho)\partial_{\theta}X^{(l)}(\theta,\phi) \tag*{(21)}
vϕ=∑l=0∞χ(l)(ρ)∂ϕX(l)(θ,ϕ)(22)v_{\phi}=\sum_{l=0}^{\infty}\chi^{(l)}(\rho)\partial_{\phi}X^{(l)}(\theta,\phi) \tag*{(22)}
T=∑l=0∞σ(l)(ρ)X(l)(θ,ϕ)(23)\mathcal{T}=\sum_{l=0}^{\infty}\sigma^{(l)}(\rho)X^{(l)}(\theta,\phi) \tag*{(23)}

where

X(l)(θ,ϕ)=∑m=0lPlm(cos⁡θ)[clmcos⁡(mϕ)+dlmsin⁡(mϕ)](24)X^{(l)}(\theta,\phi)=\sum_{m=0}^{l}P_{l}^{m}(\cos\theta)[c_{lm}\cos(m\phi)+d_{lm}\sin(m\phi)] \tag*{(24)}

with constant coefficients clmc_{lm} and dlmd_{lm}. As usual, Plm(x)P_{l}^{m}(x) represents the associated Legendre polynomial. The functions χ(l)\chi^{(l)} and σ(l)\sigma^{(l)} satisfy, respectively:

3ddρχ(l)−2χ(l)tanh⁡ρ−3σ(l)=0(25)3\frac{d}{d\rho}\chi^{(l)}-2\chi^{(l)}\tanh\rho-3\sigma^{(l)}=0 \tag*{(25)}
3ddρσ(l)+l(l+1)χ(l)sech⁡2ρ=0(26)3\frac{d}{d\rho}\sigma^{(l)}+l(l+1)\chi^{(l)}\operatorname{sech}^{2}\rho=0 \tag*{(26)}

It would be technically difficult to cope with the analytical solutions of Eqs. (25) and (26). Instead, to simplify them, let us focus on the large-size case [78], i.e.,

τ/L≪1(27)\tau/L\ll1 \tag*{(27)}

which corresponds to ρ→−∞\rho\to-\infty. In this case, the leading-order asymptotic solution of Eqs. (25) and (26) is

χ(l)≃32,σ(l)(ρ)≃1,(28)\chi^{(l)}\simeq\frac{3}{2},\qquad\sigma^{(l)}(\rho)\simeq1, \tag*{(28)}

where an overall multiplicative factor has been absorbed into the coefficients clmc_{lm} and dlmd_{lm}.

The choice of clmc_{lm} and dlmd_{lm} depends on the physical setup of our interest. For our present purpose, we will focus on a special choice with m=lm=l, leading to

X(l)(θ,ϕ)=cllPll(cos⁡θ)cos⁡(lϕ)(29)X^{(l)}(\theta,\phi)=c_{ll}P_{l}^{l}(\cos\theta)\cos(l\phi) \tag*{(29)}

for l=2,3l=2,3 only. This choice is motivated by the heavy-ion collision geometry. For non-central collisions, the nuclear overlapping region in the transverse plane is approximately elliptical [84]. Without loss of generality, we can take the shorter side along the xx axis (i.e., ϕ=0,π\phi=0,\pi) and the longer side along the yy axis (i.e., ϕ=π/2,3π/2\phi=\pi/2,3\pi/2). To construct a solution with elliptic deformation, the ϕ\phi dependence is dominated by terms with m=2m=2 in Eq. (24), among which sin⁡(2ϕ)\sin(2\phi) should be removed because they violate the reflection symmetry along the xx and yy axes; thus, dl2=0d_{l2}=0 is fixed.

We conclude that l=2l=2 should give the dominant contribution as follows. Physically, the isothermal geometry is deformed elliptically, and for a fixed value of r>0r>0, the temperature on the xx axis should be lower than that on the yy axis, namely,

T∣ϕ=0<T∣ϕ=π/2(30)\left.\mathcal{T}\right|_{\phi=0}<\left.\mathcal{T}\right|_{\phi=\pi/2} \tag*{(30)}

for r>0r>0. We note that Pl2(cos⁡θ)P_{l}^{2}(\cos\theta) has (l−2)(l-2) distinct zeros for cos⁡θ∈(−1,1)\cos\theta\in(-1,1) [85]. Since cos⁡θ≃(L2−r2)/(L2+r2)\cos\theta\simeq(L^{2}-r^{2})/(L^{2}+r^{2}) under the condition in Eq. (27), Pl2(cos⁡θ)P_{l}^{2}(\cos\theta) exhibits (l−2)(l-2) distinct zeros for r∈(0,∞)r\in(0,\infty). If modes with l>2l>2 dominated in Eq. (24), then the physical constraint in Eq. (30) would be violated as Pl>22(cos⁡θ)P_{l>2}^{2}(\cos\theta) could change the sign at finite rr. If the dominant component arises from l=m=2l=m=2, the condition (30) is guaranteed.

As long as the l=2l=2 mode is sufficiently large, we can still consider higher order perturbations originating from initial fluctuations. Indeed, recent experimental measurements have observed spin polarization corresponding to the triangular flow, implying fluctuation-driven geometric deformations in non-central collisions [20]. We shall effectively incorporate the fluctuation effects by retaining the subdominant term with m=3m=3 in Eq. (24). To ensure the monotonic thermal profile along the radial direction, we will focus on the mode with l=m=3l=m=3 only. Unlike the m=2m=2 mode, symmetry properties cannot constrain the combination of cos⁡(3ϕ)\cos(3\phi) and sin⁡(3ϕ)\sin(3\phi). We note that cl3cos⁡(3ϕ)+dl3sin⁡(3ϕ)c_{l3}\cos(3\phi)+d_{l3}\sin(3\phi) is proportional to cos⁡[3(ϕ+ϕ0)]\cos[3(\phi+\phi_{0})] with a phase factor ϕ0\phi_{0}. This ϕ0\phi_{0} does not change the local spin polarization up to a phase factor, and the conclusions of this work do not depend on ϕ0\phi_{0}. Thus, we can safely set ϕ0=0\phi_{0}=0, omitting all sin⁡(3ϕ)\sin(3\phi) terms in Eq. (24).

Under the condition in Eq. (27), the associated Legendre polynomial simplifies as follows:

Pll(cos⁡θ)≈(−1)l(2l−1)!!(2Lr)l(L2+r2)l(31)P_{l}^{l}(\cos\theta)\approx(-1)^{l}(2l-1)!!\frac{(2Lr)^{l}}{(L^{2}+r^{2})^{l}} \tag*{(31)}

Substituting Eqs. (28), (29), (31) into Eqs. (18) and transforming back to the Minkowski spacetime, we obtain the perturbed solution as [78]

uτ≃1,uη=0,(32)u_{\tau} \simeq1,\qquad u_{\eta}=0, \tag*{(32)}
ur≃−τLf(r)−3ε2τf(r)f′(r)cos⁡(2ϕ)−92ε3τf2(r)f′(r)cos⁡(3ϕ),(33)\begin{aligned} u_{r} &\simeq-\frac{\tau}{L}f(r)-3\varepsilon_{2}\tau f(r)f'(r)\cos(2\phi)\\ &\quad-\frac{9}{2}\varepsilon_{3}\tau f^{2}(r)f'(r)\cos(3\phi), \tag*{(33)} \end{aligned}
uϕ≃3ε2τf2(r)sin⁡(2ϕ)+92ε3τf3(r)sin⁡(3ϕ),(34)u_{\phi}\simeq3\varepsilon_{2}\tau f^{2}(r)\sin(2\phi)+\frac{9}{2}\varepsilon_{3}\tau f^{3}(r)\sin(3\phi), \tag*{(34)}

with the corresponding temperature distribution given by

T≃T^0(2L)2/3τ1/3(L2+r2)2/3[1−ε2f2(r)cos⁡(2ϕ)−ε3f3(r)cos⁡(3ϕ)].(35)T\simeq\frac{\hat{T}_{0}(2L)^{2/3}}{\tau^{1/3}(L^{2}+r^{2})^{2/3}}\left[1-\varepsilon_{2}f^{2}(r)\cos(2\phi)-\varepsilon_{3}f^{3}(r)\cos(3\phi)\right]. \tag*{(35)}

Here, f′(r)=df(r)/drf'(r)=df(r)/dr and

f(r)=2LrL2+r2.(36)f(r)=\frac{2Lr}{L^{2}+r^{2}}. \tag*{(36)}

The new coefficients ε2,3\varepsilon_{2,3} are defined as ε2=−3c22\varepsilon_{2}=-3c_{22} and ε3=15c33\varepsilon_{3}=15c_{33}. To ensure the perturbative hierarchy between elliptic and triangular deformations, we shall impose the following condition:

0<ε3≪ε2≪1.(37)0<\varepsilon_{3}\ll\varepsilon_{2}\ll1. \tag*{(37)}

Compared to the original Gubser flow, the perturbed solution introduces the ϕ\phi-dependence in both temperature and velocity, describing the anisotropy in the transverse plane. For example, using (t,x,y,z)(t,x,y,z) coordinates, the anisotropic part related to ε2\varepsilon_{2} of the perturbed velocity is expressed as

δux=12ε2L2τx(L2−x2+3y2)(L2+x2+y2)3,(38)\delta u^{x}=\frac{12\varepsilon_{2}L^{2}\tau x(L^{2}-x^{2}+3y^{2})}{(L^{2}+x^{2}+y^{2})^{3}}, \tag*{(38)}
δuy=−12ε2L2τy(L2+3x2−y2)(L2+x2+y2)3,(39)\delta u^{y}=-\frac{12\varepsilon_{2}L^{2}\tau y(L^{2}+3x^{2}-y^{2})}{(L^{2}+x^{2}+y^{2})^{3}}, \tag*{(39)}

which are independent of the first term −τf(r)/L-\tau f(r)/L in Eq. (27b). The distribution of the anisotropic part of the perturbed velocity field is illustrated in Figure 2. The arrows represent the direction of the velocity field, and the color indicates the magnitude of the velocity. We note that this perturbed velocity field exhibits a clear elliptic deformation, representing a finite elliptic-flow component. The elliptic deformation by ε2\varepsilon_{2} is consistent with the geometry of non-central heavy-ion collisions, while the triangular deformation by ε3\varepsilon_{3} is induced by initial fluctuations.

Figure 2. Transverse distribution of (δux,δuy)(\delta u^x,\delta u^y) in Eqs. (38) and (39) with the mode coefficients, ε2=0.1\varepsilon_2=0.1 and ε3=0\varepsilon_3=0, and the physical setting, L=10L=10 fm and τ=1\tau=1 fm. The arrows represent the direction of the velocity field, and the color indicates the magnitude of the velocity. The velocity field exhibits a clear elliptic deformation.

We emphasize that the perturbed solution in Eqs. (27) and (35) relies on the large size condition in Eq. (27), which we assume holds even on the freeze-out hypersurface. Denoting the freeze-out temperature as TfT_f, we determine the freeze-out hypersurface by

T(τf,r,ϕ)=Tf,(40)T(\tau_f,r,\phi)=T_f, \tag*{(40)}

Solving this equation yields the freeze-out proper time:

τf≃T^03L2Tf3(L2+r2)2[1−3ε2f2(r)cos⁡(2ϕ)−3ε3f3(r)cos⁡(3ϕ)].(41)\tau_f\simeq\frac{\hat{T}_0^3L^2}{T_f^3(L^2+r^2)^2}\left[1-3\varepsilon_2f^2(r)\cos(2\phi)-3\varepsilon_3f^3(r)\cos(3\phi)\right]. \tag*{(41)}

Consistency with the condition in Eq. (27) then requires

T^03(TfL)3≪1,(42)\frac{\hat{T}_0^3}{(T_fL)^3}\ll1, \tag*{(42)}

which we assume throughout the subsequent analyses.

Three Different Formulations

The sign puzzle in local spin polarization suggests that the description based on thermal vorticity alone is incomplete, which has motivated the inclusion of shear-induced polarization. Although the polarization induced by thermal vorticity is well-established, at least three distinct formulations for shear-induced polarization currently coexist in the literature. We briefly review these formulations below.

The spin polarization of the Λ\Lambda hyperon in the Belinfante pseudo-gauge [86, 87, 88], as a function of its momentum, pμp^\mu, has two contributions, i.e., Sμ(p)=Sϖμ(p)+Sξμ(p)S^\mu(p)=S^\mu_{\varpi}(p)+S^\mu_{\xi}(p), the former from the thermal vorticity tensor ϖμν\varpi_{\mu\nu} and the latter from the thermal shear tensor ξμν\xi_{\mu\nu}. These contributions are calculated by the modified Cooper-Frye formula on the freeze-out hypersurface Σf\Sigma_f as [60, 61]

Sϖμ(p)=−18mN(p)∫Σf[dΣ]ϵμνρσpνϖρσ(43)S^\mu_{\varpi}(p)=-\frac{1}{8m\mathcal{N}(p)}\int_{\Sigma_f}[d\Sigma]\epsilon^{\mu\nu\rho\sigma}p_\nu\varpi_{\rho\sigma} \tag*{(43)}
Sξμ(p)=−14mN(p)∫Σf[dΣ]ϵμνρσpνnρp⋅nξσλ(44)S^\mu_{\xi}(p)=-\frac{1}{4m\mathcal{N}(p)}\int_{\Sigma_f}[d\Sigma]\epsilon^{\mu\nu\rho\sigma}\frac{p_\nu n_\rho}{p\cdot n}\xi_{\sigma\lambda} \tag*{(44)}

Before explaining the notation, we note that p⋅Sϖ,ξ=0p\cdot S_{\varpi,\xi}=0 holds, and in the following, we work in the lab frame; we therefore focus on the spatial part, S⃗\vec{S}, only.

In the above expressions, we introduced a shorthand notation [dΣ][d\Sigma] defined as [dΣ]=dΣ⋅p nF(1−nF)[d\Sigma]=d\Sigma\cdot p\,n_F(1-n_F), where dΣμd\Sigma^\mu is the future-directed integration measure and nF=(1+eβ⋅p)−1n_F=(1+e^{\beta\cdot p})^{-1} is the Fermi-Dirac distribution function. Since eβ⋅p≫1e^{\beta\cdot p}\gg1 in practical applications, we can approximate the weight factor as nF(1−nF)≃nF≃e−β⋅pn_F(1-n_F)\simeq n_F\simeq e^{-\beta\cdot p}.

Here, the thermal vorticity tensor ϖμν\varpi_{\mu\nu} and thermal shear tensor ξμν\xi_{\mu\nu} are given by

ϖμν=−12(∂μβν−∂νβμ)(45)\varpi_{\mu\nu}=-\frac{1}{2}(\partial_\mu\beta_\nu-\partial_\nu\beta_\mu) \tag*{(45)}
ξμν=12(∂μβν+∂νβμ)(46)\xi_{\mu\nu}=\frac{1}{2}(\partial_\mu\beta_\nu+\partial_\nu\beta_\mu) \tag*{(46)}

with βμ=uμ/T\beta_\mu=u_\mu/T. The normalization factor N(p)\mathcal{N}(p) is calculated from

N(p)=∫ΣfdΣ⋅p nF.(47)\mathcal{N}(p)=\int_{\Sigma_f}d\Sigma\cdot p\,n_F. \tag*{(47)}

The choice of the unit vector nμn^\mu in Eq. (44) may vary across different theoretical formulations. In this work, we examine the following three distinct choices:

  1. FLPSY formulation [60, 62]: nμn^\mu is identified with the fluid velocity uμu^\mu.

  1. BBPIK formulation [61]: nμn^\mu is set to be nμ=δ0μn^\mu=\delta^\mu_0 in (t,x,y,z)(t,x,y,z) coordinates. In addition, as proposed in [63], explicit TT-gradients are removed by the replacements of ϖμν→ϖμνK\varpi_{\mu\nu}\to\varpi^K_{\mu\nu} and ξμν→ξμνK\xi_{\mu\nu}\to\xi^K_{\mu\nu}, where

ϖμνK=−12β(∂μuν−∂νuμ)(48)\varpi^K_{\mu\nu}=-\frac{1}{2}\beta(\partial_\mu u_\nu-\partial_\nu u_\mu) \tag*{(48)}
ξμνK=12β(∂μuν+∂νuμ).(49)\xi^K_{\mu\nu}=\frac{1}{2}\beta(\partial_\mu u_\nu+\partial_\nu u_\mu). \tag*{(49)}
  1. SBR formulation [72]: nμn^\mu is the normal vector to Σf\Sigma_f. If Σf\Sigma_f is isothermal, i.e., with constant temperature, nμ=−∂μT/∂νT∂νTn^\mu=-\partial^\mu T/\sqrt{\partial^\nu T\partial_\nu T}. Then, TT-gradient terms cancel out automatically in Sμ(p)S^\mu(p), though each Sϖ,ξμ(p)S^\mu_{\varpi,\xi}(p) receives TT-gradient contributions.

It should be noted that the SBR formulation is applicable to an arbitrary Σf\Sigma_f, and its complete form in [72] is more general than Eqs. (43) and (44). In the present work, we will calculate the polarization vector on an isothermal Σf\Sigma_f determined by Eq. (40). In this way, we will omit unphysical contributions mentioned in [72] and use formulas in Eqs. (43) and (44) to evaluate the spin polarization in the SBR formulation.

For the mass mm appearing in Eqs. (43) and (44), the Λ\Lambda hyperon mass is the common choice, but as argued in Ref. [62], mm can be the s-quark mass if the Λ\Lambda spin polarization originates from the constituent s quark. In the subsequent analyses, therefore, we will compare both scenarios:

m={1116 MeV(Λ hyperon),300 MeV(s quark).(50)m=\begin{cases} 1116\ \mathrm{MeV} & (\Lambda\ \text{hyperon}),\\ 300\ \mathrm{MeV} & (\text{s quark}). \end{cases} \tag*{(50)}

Spin Polarization with the Rotating Hubble Flow

We now calculate the spin polarization by using the analytical solutions presented in Sec. II. Let us first consider the spin polarization in the rotating Hubble flow. For simplicity, we assume ∣ω⃗∣≪1|\vec{\omega}|\ll1 and ∣ω⃗×x⃗/t∣≪1|\vec{\omega}\times\vec{x}/t|\ll1, and expand the relevant integrals in powers of ω⃗\vec{\omega}. Then, a nonzero polarization first appears at linear order in ω⃗\vec{\omega}. In order to distinguish the results clearly, we will use superscripts FLPSY, BBPIK, and SBR to label three formulations shown in Sec. III.

In the FLPSY and the SBR formulations, the polarization induced by thermal vorticity is parallel to ω⃗\vec{\omega}:

S⃗ϖFLPSY(p)=S⃗ϖSBR(p)=Ep4mT0ω⃗,(51)\vec{S}^{\mathrm{FLPSY}}_{\varpi}(p)=\vec{S}^{\mathrm{SBR}}_{\varpi}(p)=\frac{E_p}{4mT_0}\vec{\omega}, \tag*{(51)}

where Ep=p0E_p=p^0. We note that S⃗ξFLPSY(p)=S⃗ξSBR(p)=0\vec{S}^{\mathrm{FLPSY}}_{\xi}(p)=\vec{S}^{\mathrm{SBR}}_{\xi}(p)=0 to all orders. This is because the thermal shear tensor is proportional to the metric, i.e., ξμν=T0−1gμν\xi_{\mu\nu}=T_0^{-1}g_{\mu\nu} in this case, making the integration in Eq. (44) identically zero. Therefore, the spin polarization receives only a vorticity-induced part, and no cancellation between S⃗ϖ\vec{S}_{\varpi} and S⃗ξ\vec{S}_{\xi} can occur for the rotating Hubble flow. This observation is consistent with the experimental data of global Λ\Lambda polarization, which is well described by the thermal vorticity alone [4].

In the BBPIK formulation, we obtain

S⃗ϖBBPIK(p)=Ep4mT0[1+TfmK3(m/Tf)K2(m/Tf)]ω⃗,(52)\vec{S}^{\mathrm{BBPIK}}_{\varpi}(p)=\frac{E_p}{4mT_0}\left[1+\frac{T_f}{m}\frac{K_3(m/T_f)}{K_2(m/T_f)}\right]\vec{\omega}, \tag*{(52)}
S⃗ξBBPIK(p)=K4(m/Tf)8mT0EpK2(m/Tf)[(ω⃗⋅p⃗)p⃗−p⃗ 2ω⃗],(53)\vec{S}^{\mathrm{BBPIK}}_{\xi}(p)=\frac{K_4(m/T_f)}{8mT_0E_pK_2(m/T_f)}\left[(\vec{\omega}\cdot\vec{p})\vec{p}-\vec{p}^{\,2}\vec{\omega}\right], \tag*{(53)}

where Kν(x)K_\nu(x) is the modified Bessel function of the second kind. In a similar way to the FLPSY and the SBR formulations, S⃗ϖBBPIK(p)\vec{S}^{\mathrm{BBPIK}}_{\varpi}(p) remains parallel to ω⃗\vec{\omega}. There is a difference in the vorticity-induced part between the FLPSY/SBR results in Eq. (51) and the BBPIK results in Eq. (52). This difference is a function of m/Tfm/T_f, and it decreases with increasing m/Tfm/T_f. The shear-induced polarization S⃗ξBBPIK\vec{S}^{\mathrm{BBPIK}}_{\xi} is always perpendicular to p⃗\vec{p}. For global polarization, we average Eqs. (52) and (53) over the momentum p⃗\vec{p}. Upon averaging, the direction of S⃗ϖBBPIK\vec{S}^{\mathrm{BBPIK}}_{\varpi} remains unchanged, while S⃗ξBBPIK\vec{S}^{\mathrm{BBPIK}}_{\xi} becomes strictly anti-parallel to S⃗ϖBBPIK\vec{S}^{\mathrm{BBPIK}}_{\varpi}. The two contributions therefore partially cancel, yet the net polarization remains aligned with ω⃗\vec{\omega}. We emphasize that, for global polarization, the shear-induced polarization contributions in the BBPIK formula remains nonzero.

Spin Polarization with the Perturbed Gubser Flow

In this section, we consider the spin polarization using the perturbed Gubser flow given in Eqs. (27) and (35).

Analytical Expressions of the Spin Polarization

We parametrize the momentum pμ=(Ep,px,py,pz)p^\mu=(E_p,p^x,p^y,p^z) as follows:

Ep=mTcosh⁡ηp,pz=mTsinh⁡ηp,E_p=m_T\cosh\eta_p,\qquad p^z=m_T\sinh\eta_p,
px=pTcos⁡ϕp,py=pTsin⁡ϕp,(54)p^x=p_T\cos\phi_p,\qquad p^y=p_T\sin\phi_p, \tag*{(54)}

where, conversely, we have ηp=tanh⁡−1(pz/Ep)\eta_p=\tanh^{-1}(p^z/E_p), ϕp=tan⁡−1(py/px)\phi_p=\tan^{-1}(p^y/p^x), pT=(px2+py2)1/2p_T=(p_x^2+p_y^2)^{1/2}, and mT=(pT2+m2)1/2m_T=(p_T^2+m^2)^{1/2}. With this parametrization, the integration measure dΣ⋅pd\Sigma\cdot p on the freeze-out hypersurface Σf\Sigma_f can be expressed as

dΣ⋅p=dr dϕ dη τf[rmTcosh⁡(ηp−η)−rpTcos⁡(ϕp−ϕ)∂rτf−pTsin⁡(ϕp−ϕ)∂ϕτf](55)d\Sigma\cdot p=dr\,d\phi\,d\eta\,\tau_f\left[rm_T\cosh(\eta_p-\eta)-rp_T\cos(\phi_p-\phi)\partial_r\tau_f-p_T\sin(\phi_p-\phi)\partial_\phi\tau_f\right] \tag*{(55)}

with integration domains r∈[0,+∞)r\in[0,+\infty), ϕ∈[0,2π)\phi\in[0,2\pi), and η∈(−∞,+∞)\eta\in(-\infty,+\infty).

After evaluating the integrals in Eqs. (43) and (44), we will average the results over a symmetric interval of ηp\eta_p centered at the origin, following the experimental procedure. Then, Sx=Sy=0S^x=S^y=0, because the integrands are odd under (η,ηp)→(−η,−ηp)(\eta,\eta_p)\to(-\eta,-\eta_p), reflecting the underlying symmetry of the collision system. Thus, only the longitudinal component, Sz(p)S^z(p), survives in the averaging process, which significantly simplifies our subsequent analyses.

To derive analytical expressions of the longitudinal spin polarization, we expand the integrals for small ϵ2,3\epsilon_{2,3} and large LL. For the normalization factor in Eq. (47), the leading order term is given by

N(p)=8πT^03ϑTf2K1(ϑ),ϑ=m2+pT2Tf.(56)\mathcal{N}(p)=\frac{8\pi\hat{T}_0^3\vartheta}{T_f^2}K_1(\vartheta),\qquad\vartheta=\frac{\sqrt{m^2+p_T^2}}{T_f}. \tag*{(56)}

For the longitudinal spin polarization SzS^z, we first take an expansion in Eqs. (43) and (44) in powers of ϵ2,3\epsilon_{2,3} up to the linear order:

Sϖz(p)=∑i=23ϵiS(i),ϖ(p)sin⁡(iϕp),(57)S^z_{\varpi}(p)=\sum_{i=2}^{3}\epsilon_iS_{(i),\varpi}(p)\sin(i\phi_p), \tag*{(57)}
Sξz(p)=∑i=23ϵiS(i),ξ(p)sin⁡(iϕp).(58)S^z_{\xi}(p)=\sum_{i=2}^{3}\epsilon_iS_{(i),\xi}(p)\sin(i\phi_p). \tag*{(58)}

Then, we expand the Fourier coefficients, S(2)(p)S_{(2)}(p) and S(3)(p)S_{(3)}(p), for large LL such that the condition in Eq. (42) is satisfied. We keep the leading order terms correspondingly, by changing the integration variable as r→s=r/Lr\to s=r/L in Eqs. (43) and (44), and expanding the integrands in powers of 1/(TfL)1/(T_fL). We then perform the integrations over ϕ∈[0,2π)\phi\in[0,2\pi), s∈[0,+∞)s\in[0,+\infty), and η∈(−∞,+∞)\eta\in(-\infty,+\infty).

The analytical results are summarized as follows. For the polarization induced by thermal vorticity, we find the relation:

S(i),ϖFLPSY(p)=2S(i),ϖBBPIK(p)=S(i),ϖSBR(p)(59)S^{\mathrm{FLPSY}}_{(i),\varpi}(p)=2S^{\mathrm{BBPIK}}_{(i),\varpi}(p)=S^{\mathrm{SBR}}_{(i),\varpi}(p) \tag*{(59)}

for i=2,3i=2,3, and

S(2),ϖSBR(p)=−C(p)[ϑK0(ϑ)+3K1(ϑ)],(60)S^{\mathrm{SBR}}_{(2),\varpi}(p)=-C(p)\left[\vartheta K_0(\vartheta)+3K_1(\vartheta)\right], \tag*{(60)}
S(3),ϖSBR(p)=−D(p)[ϑK0(ϑ)+5K1(ϑ)],(61)S^{\mathrm{SBR}}_{(3),\varpi}(p)=-D(p)\left[\vartheta K_0(\vartheta)+5K_1(\vartheta)\right], \tag*{(61)}

where the coefficients, C(p)C(p) and D(p)D(p), read:

C(p)=96πT^06pT2cosh⁡ηp35mTf7L4N(p),(62)C(p)=\frac{96\pi\hat{T}_0^6p_T^2\cosh\eta_p}{35mT_f^7L^4\mathcal{N}(p)}, \tag*{(62)}
D(p)=144πT^09pT3cosh⁡ηp55mTf11L7N(p).(63)D(p)=\frac{144\pi\hat{T}_0^9p_T^3\cosh\eta_p}{55mT_f^{11}L^7\mathcal{N}(p)}. \tag*{(63)}

For large LL, we find C(p)≫D(p)C(p)\gg D(p), which is consistent with the assumption that the elliptic deformation is dominant.

We note that in Eq. (59) S(i),ϖBBPIK(p)S^{\mathrm{BBPIK}}_{(i),\varpi}(p) is half of S(i),ϖFLPSYS^{\mathrm{FLPSY}}_{(i),\varpi} and S(i),ϖSBRS^{\mathrm{SBR}}_{(i),\varpi}. The perturbed solution in Eqs. (27) and (35) satisfies

ϖμν=−β2(∂ρuσ−∂σuρ)−12(uσ∂ρβ−uρ∂σβ)=−β(∂ρuσ−∂σuρ)+O(ϵ2,32),(64)\begin{aligned} \varpi_{\mu\nu}=-\frac{\beta}{2}(\partial_\rho u_\sigma-\partial_\sigma u_\rho)-\frac{1}{2}(u_\sigma\partial_\rho\beta-u_\rho\partial_\sigma\beta) \\ &=-\beta(\partial_\rho u_\sigma-\partial_\sigma u_\rho)+\mathcal{O}(\epsilon_{2,3}^2), \tag*{(64)} \end{aligned}

while ϖμν\varpi_{\mu\nu} is replaced with −12β(∂ρuσ−∂σuρ)-\frac{1}{2}\beta(\partial_\rho u_\sigma-\partial_\sigma u_\rho) in the BBPIK formula as indicated in Eq. (48). This explains the factor of 1/21/2 in Eq. (59) [89].

For the polarization induced by thermal shear tensor, we obtain:

S(2),ξFLPSY(p)=3C(p)[3K1(ϑ)−∫ϑ∞dt K0(t)],(65)S^{\mathrm{FLPSY}}_{(2),\xi}(p)=3C(p)\left[3K_1(\vartheta)-\int_{\vartheta}^{\infty}dt\,K_0(t)\right], \tag*{(65)}
S(3),ξFLPSY(p)=3D(p)[10ϑ−1K0(ϑ)+3K1(ϑ)](66)S^{\mathrm{FLPSY}}_{(3),\xi}(p)=3D(p)\left[10\vartheta^{-1}K_0(\vartheta)+3K_1(\vartheta)\right] \tag*{(66)}

for the FLPSY formulation, and

S(2),ξBBPIK(p)=C(p)2cosh⁡2ηp[ϑK0(ϑ)+11K1(ϑ)],(67)S^{\mathrm{BBPIK}}_{(2),\xi}(p)=\frac{C(p)}{2\cosh^2\eta_p}\left[\vartheta K_0(\vartheta)+11K_1(\vartheta)\right], \tag*{(67)}
S(3),ξBBPIK(p)=D(p)2cosh⁡2ηp[17K1(ϑ)+(ϑ+24ϑ−1)K0(ϑ)](68)S^{\mathrm{BBPIK}}_{(3),\xi}(p)=\frac{D(p)}{2\cosh^2\eta_p}\left[17K_1(\vartheta)+(\vartheta+24\vartheta^{-1})K_0(\vartheta)\right] \tag*{(68)}

for the BBPIK formulation, and

S(2,3),ξSBR(p)=−S(2,3),ϖSBR(p)(69)S^{\mathrm{SBR}}_{(2,3),\xi}(p)=-S^{\mathrm{SBR}}_{(2,3),\varpi}(p) \tag*{(69)}

for the SBR formulation. For large LL, in all the cases, S(i),ξ(p)S_{(i),\xi}(p) turns out to be of the same order as S(i),ϖ(p)S_{(i),\varpi}(p). In particular for the SBR formulation, they cancel exactly and the total polarization vanishes.

Dependence on the Azimuthal Angle and the Transverse Momenta

Let us first examine the dependence on the azimuthal angle ϕp\phi_p. For illustration, SϖzS_{\varpi}^{z}, SξzS_{\xi}^{z}, and their sum are plotted as functions of ϕp\phi_p in Figure 3 and Figure 4. We set ϵ2=0.1\epsilon_2=0.1 and ϵ3=0\epsilon_3=0 to isolate the sin⁡(2ϕp)\sin(2\phi_p) structure in Figure 3, and ϵ2=0\epsilon_2=0 and ϵ3=0.01\epsilon_3=0.01 to extract the sin⁡(3ϕp)\sin(3\phi_p) structure in Figure 4. Other parameters are chosen as T^0=3\hat{T}_0=3, L=10 fmL=10\ \mathrm{fm}, and Tf=150 MeVT_f=150\ \mathrm{MeV}. The results are averaged over the rapidity range, ηp∈[−1,1]\eta_p\in[-1,1] and the transverse momentum, pT∈[0.5,3] GeVp_T\in[0.5,3]\ \mathrm{GeV}.

Figure 3. Longitudinal spin polarization with ϵ2=0.1\epsilon_2=0.1 and ϵ3=0\epsilon_3=0 as a function of the azimuthal angle ϕp\phi_p. The green dashed line represents SϖzS_{\varpi}^{z}, the blue dashed line represents SξzS_{\xi}^{z}, and the red solid line represents the total polarization. The panels display results for different formulations: (a) and (b) for FLPSY, (c) for BBPIK, and (d) for SBR. The mass mm is set to 300 MeV in panel (b) and 1116 MeV in others. The remaining parameters are fixed as T^0=3\hat{T}_0=3, L=10 fmL=10\ \mathrm{fm}, and Tf=150 MeVT_f=150\ \mathrm{MeV}. The results are averaged over the rapidity range ηp∈[−1,1]\eta_p\in[-1,1] and the transverse momentum range pT∈[0.5,3] GeVp_T\in[0.5,3]\ \mathrm{GeV}.

Figure 4. Longitudinal spin polarization with ϵ2=0\epsilon_2=0 and ϵ3=0.01\epsilon_3=0.01 as a function of the azimuthal angle ϕp\phi_p. The green dashed line represents SzϖS^{\varpi}_z, the blue dashed line represents SzξS^{\xi}_z, and the red solid line represents the total polarization. The panels display results for different formulations: (a) and (b) for FLPSY, (c) for BBPIK, and (d) for SBR. The mass mm is set to 300 MeV in panel (b) and 1116 MeV in others. The remaining parameters are fixed as T^0=3\hat{T}_0=3, L=10L=10 fm, and Tf=150T_f=150 MeV. The results are averaged over the rapidity range ηp∈[−1,1]\eta_p \in[-1,1] and the transverse momentum range pT∈[0.5,3]p_T \in[0.5,3] GeV.

From Figure 3 and Figure 4, we see that SξzS_{\xi}^{z} has the desired sign consistent with the experimental data [18, 20], while SϖzS_{\varpi}^{z} always has the opposite sign. This can be understood from our formulas with C(p),D(p)>0C(p),D(p)>0 in Eqs. (62) and (63). To reproduce the desired sign of total longitudinal spin polarization, the magnitude of the thermal-shear contribution should exceed that of thermal vorticity. For panels (a) and (b) of Figure 3 in the FLPSY formulation, the total spin polarization shown by the solid line has the undesired sign regardless of the mass choice. Interestingly, we find that the total longitudinal spin polarization in the FLPSY formalism becomes positive in panel (b) of Figure 4. In the SBR formulation in panel (d) of Figure 3 and Figure 4, the total polarization is identically zero as a result of exact cancellation at leading order. The BBPIK formulation in panel (c) of Figure 3 and Figure 4 produces the desired sign, arising after a significant cancellation between SϖzS_{\varpi}^{z} and SξzS_{\xi}^{z}. For the BBPIK and SBR formulations, here, we present only the results for m=1116 MeVm=1116\ \mathrm{MeV}; we have numerically checked that the sign in the BBPIK formulation remains unchanged for m=300 MeVm=300\ \mathrm{MeV}.

As another interesting observation, we find that the longitudinal spin polarization vanishes when ϵ2=ϵ3=0\epsilon_2=\epsilon_3=0. This indicates that finite initial eccentricities generate anisotropic flow, which in turn leads to longitudinal spin polarization. Therefore, initial-state fluctuations can influence longitudinal spin polarization.

Next, we shall discuss the dependence on the transverse momentum pTp_T. Since the relative magnitudes of SϖzS_{\varpi}^{z} and SξzS_{\xi}^{z} determine the sign of their sum, we plot the ratio, S(i),ϖ/S(i),ξS_{(i),\varpi}/S_{(i),\xi}, averaged over ηp∈[−1,1]\eta_p\in[-1,1] as a function of pTp_T in Figure 5. Interestingly, this ratio is independent of T^0\hat{T}_0 and LL, implying that both SϖzS_{\varpi}^{z} and SξzS_{\xi}^{z} identically respond to the initial conditions. Obtaining the experimentally desired sign for the total longitudinal spin polarization requires S(i),ϖ/S(i),ξ>−1S_{(i),\varpi}/S_{(i),\xi}>-1.

Figure 5. S(i),ϖ/S(i),ξS_{(i),\varpi}/S_{(i),\xi} as a function of the transverse momentum pTp_T. Panel (a) shows S(2),ϖ/S(2),ξS_{(2),\varpi}/S_{(2),\xi} (i=2i=2), while panel (b) shows S(3),ϖ/S(3),ξS_{(3),\varpi}/S_{(3),\xi} (i=3i=3). The freeze-out temperature is taken as Tf=150T_f=150 MeV. S(i),ϖS_{(i),\varpi} and S(i),ξS_{(i),\xi} are separately averaged over ηp∈[−1,1]\eta_p\in[-1,1].

Comparing the results for m=1116 MeVm=1116\ \mathrm{MeV} and m=300 MeVm=300\ \mathrm{MeV}, we find that a smaller mass yields a larger ratio, S(i),ϖ/S(i),ξS_{(i),\varpi}/S_{(i),\xi}, which can be understood from the (pνnρ)/(p⋅n)(p_\nu n_\rho)/(p\cdot n) factor in SξμS_{\xi}^{\mu} [89]. In particular, for the FLPSY formulation, Figure 5 shows S(i),ϖFLPSY/S(i),ξFLPSY>−1S_{(i),\varpi}^{\mathrm{FLPSY}}/S_{(i),\xi}^{\mathrm{FLPSY}}>-1 for small pTp_T in the case of m=300 MeVm=300\ \mathrm{MeV}. Although the ratio lies above −1-1 in the small-pTp_T region giving the correct sign there, the ratio decays rapidly as pTp_T increases and becomes much less than −1-1 at large pTp_T. This behavior is consistent with the results in Figure 3 and Figure 4, where the spin polarization averaged over pT∈[0.5,3]p_T \in[0.5,3] GeV has an undesired sign.

The BBPIK formulation works well when pT<3p_T<3 GeV, as shown in Figure 5. As pTp_T increases, the ratio decreases slowly but eventually crosses −1-1. At large pTp_T, we find S(i),ϖBBPIK/S(i),ξBBPIK<−1S^{\mathrm{BBPIK}}_{(i),\varpi}/S^{\mathrm{BBPIK}}_{(i),\xi}<-1, which suggests that the BBPIK formulation also predicts the opposite sign for the longitudinal spin polarization at large transverse momenta.

Acceleration-driven Near Cancellation

From Figure 3–Figure 5, we find a significant cancellation between SzϖS^{\varpi}_z and SzξS^{\xi}_z in the BBPIK formulation. This cancellation can partially be understood by comparing the integrands of Szϖ(p)S^{\varpi}_z(p) in Eq. (43) and Szξ(p)S^{\xi}_z(p) in Eq. (44). Recalling that the replacements, ϖμν→ϖμνK\varpi_{\mu\nu}\to\varpi^K_{\mu\nu} and ξμν→ξμνK\xi_{\mu\nu}\to\xi^K_{\mu\nu}, are adopted in the BBPIK formulation, we then find

12ϵμνρσpνϖρσK=pxϖtyK−pyϖtxK+⋯(70)\frac{1}{2}\epsilon^{\mu\nu\rho\sigma}p_\nu\varpi^K_{\rho\sigma}=p_x\varpi^K_{ty}-p_y\varpi^K_{tx}+\cdots \tag*{(70)}
ϵμνρσpνnρp⋅npλξσλK=pxξtyK−pyξtxK+⋯ .(71)\epsilon^{\mu\nu\rho\sigma}\frac{p_\nu n_\rho}{p\cdot n}p^\lambda\xi^K_{\sigma\lambda}=p_x\xi^K_{ty}-p_y\xi^K_{tx}+\cdots. \tag*{(71)}

Here, the ellipses represent the contributions from other components of ϖμνK\varpi^K_{\mu\nu} and ξμνK\xi^K_{\mu\nu}. We include a factor of 1/21/2 in Eq. (70) to account for the factor-two difference between the denominators in Eqs. (43) and (44). According to the definitions of ϖμνK\varpi^K_{\mu\nu} and ξμνK\xi^K_{\mu\nu}, the acceleration terms ∂tux\partial_tu_x and ∂tuy\partial_tu_y cancel exactly in the sum of two contributions in Eqs. (70) and (71) [89]. Therefore, the acceleration effect does not cause the longitudinal spin polarization in the BBPIK formulation. We emphasize that this argument does not rely on any specific velocity profile.

In the perturbed Gubser flow given by Eq. (27), the terms represented by the ellipses in Eq. (70) vanish, and ϖtxK=−ξtxK\varpi^K_{tx}=-\xi^K_{tx} and ϖtyK=−ξtyK\varpi^K_{ty}=-\xi^K_{ty} immediately follow. This implies that Szϖ(p)S^{\varpi}_z(p) is completely eliminated by acceleration part of Szξ(p)S^{\xi}_z(p) in the BBPIK formulation. In this specific case, the total longitudinal spin polarization depends only on the non-acceleration components of ξμν\xi_{\mu\nu}, but not ϖμν\varpi_{\mu\nu}.

The cancellation of the acceleration terms also occurs in the SBR formulation. In the QGP generated in relativistic heavy-ion collisions, the longitudinal expansion dominates over the transverse expansion. If ηp\eta_p dependence is negligible, the anisotropic expansion leads to the following condition for the temperature gradients:

∣∂xT∣, ∣∂yT∣≪∣∂zT∣.(72)|\partial_xT|,\ |\partial_yT|\ll|\partial_zT|. \tag*{(72)}

The unit vector, nμ=−∂μT/∣∂νT∂νT∣n^\mu=-\partial^\mu T/\sqrt{|\partial^\nu T\partial_\nu T|}, in the SBR formulation can thus be approximated as

(nt,nx,ny,nz)≃(−∂tT,0,0,∂zT)(∂tT)2−(∂zT)2.(73)(n^t,n^x,n^y,n^z)\simeq\frac{(-\partial_tT,0,0,\partial_zT)}{\sqrt{(\partial_tT)^2-(\partial_zT)^2}}. \tag*{(73)}

Since ∂tT=−∂zTcoth⁡η\partial_tT=-\partial_zT\coth\eta due to ∂ηT=0\partial_\eta T=0, we can write the above expression as

nμ=δ0μ+δnμ,(74)n^\mu=\delta^\mu_0+\delta n^\mu, \tag*{(74)}

where δnt=cosh⁡η−1\delta n^t=\cosh\eta-1, δnz=sinh⁡η\delta n^z=\sinh\eta, δnx≃0\delta n^x\simeq0, δny≃0\delta n^y\simeq0 under the condition in Eq. (72). In the mid rapidity region ηp≈0\eta_p\approx0, the distribution function, nFn_F, provides an exponential suppression factor e−ϑcosh⁡ηe^{-\vartheta\cosh\eta}. Hence, the dominant contribution to the integral in Eq. (44) comes from the region where η≈0\eta\approx0, and thus δnt≈δnz≈0\delta n^t\approx\delta n^z\approx0. The vector nμn^\mu can now be approximated as δ0μ\delta^\mu_0. Thanks to the exponential suppression, this approximation becomes increasingly accurate in the SBR formulation as ϑ\vartheta increases, i.e., pTp_T increases. Recall that, in the SBR formulation, the replacement of ϖμν→ϖKμν\varpi^{\mu\nu}\to\varpi_K^{\mu\nu} and ξμν→ξKμν\xi^{\mu\nu}\to\xi_K^{\mu\nu} does not change the total polarization vector Sμ(p)S^\mu(\boldsymbol{p}) [72]. Therefore, Eqs. (70) and (71) are approximately applicable to the SBR formulation, so that the acceleration terms in SϖzS^z_\varpi and SξzS^z_\xi cancel each other, as in the BBPIK formulation.

Finally, we make a remark about the total longitudinal spin polarization in the SBR formulation, which is

exactly vanishing as shown in Figure 5, which still lacks a theoretical explanation, and is probably an accidental cancellation. As discussed above, SϖzS^z_\varpi is completely eliminated by the acceleration part of SξzS^z_\xi in the perturbed Gubser flow. The exact cancellation requires that the non-acceleration part of SξzS^z_\xi should also vanish. We have numerically verified that the integrand of the complicated non-acceleration part of SξzS^z_\xi is nonzero, while the integration over the full hypersurface makes the final result vanish. We have also examined the effect of introducing finite cutoffs for η\eta and rr in the integration. Although this renders total longitudinal spin polarization finite, its magnitude remains much smaller than SϖzS^z_\varpi and SξzS^z_\xi.

Conclusion

In this work, we have analytically studied the longitudinal spin polarization of Λ\Lambda hyperons in relativistic heavy-ion collisions using analytical hydrodynamic solutions. Based on the rotating Hubble flow and the perturbed Gubser flow including elliptic and triangular fluctuations, we have derived analytical expressions for the spin polarization vector in three distinct ways: the FLPSY, BBPIK, and SBR formulations. Our analytical approach provides a useful insight to investigate the thermal-vorticity and thermal-shear contributions to the longitudinal spin polarization.

In the rotating Hubble flow, which could qualitatively be regarded as an approximation of the global polarization setup, the polarization induced by thermal vorticity is found to be parallel to the angular velocity vector ω⃗\vec{\omega} in all formulations under consideration. The contribution from the thermal shear tensor vanishes identically in the FLPSY and SBR formulations, while in the BBPIK formulation it becomes nonzero and contains a component antiparallel to ω⃗\vec{\omega}, leading to the tendency of partial cancellation. For global polarization, the shear-induced polarization in the BBPIK formulation becomes strictly antiparallel to ω⃗\vec{\omega} after an average over spatial momentum, yet the net polarization remains aligned with ω⃗\vec{\omega}.

In perturbed Gubser flow, in contrast, our results have revealed that, across all evaluated formulations, the thermal vorticity alone always yields a sign opposite to the experimental observations. This motivates the inclusion of thermal shear tensor, which is expected to recover the desired overall sign. However, the final result is highly sensitive to the chosen formulation. In the FLPSY formulation, we have found that the experimentally favored sign of the longitudinal spin polarization is recovered only when the ss quark mass (ms=300m_s=300 MeV) is used, and only within the small transverse momentum regime. In the BBPIK formulation, the predicted total longitudinal polarization has a correct sign. In the SBR formulation, it vanishes at leading order after integration over the full isothermal freeze-out hypersurface.

We have found that the acceleration terms cancel in both the BBPIK and SBR formulations. We have demonstrated that the acceleration terms arising from the thermal vorticity are canceled by the acceleration terms in the thermal shear tensor. Consequently, the longitudinal spin polarization is determined primarily by the non-acceleration effects. This cancellation partly explains why the net polarization is significantly smaller than the individual thermal-vorticity and thermal-shear contributions. We emphasize that such cancellation does not rely on velocity profiles. In the present work, our aim was to deepen our understanding about the roles of the acceleration and non-acceleration terms in the thermal-vorticity and thermal-shear contributions at the analytical level. Thus, we did not try to adjust the parameters, T0T_0, LL, ϵ2,3\epsilon_{2,3}, to fit the experimental data. However, the perturbed Gubser flow could be a reasonable approximation at sufficiently high collision energies as measured at LHC [19]. Then, in the future, it is worth extending our analytical approach for phenomenological study. For this purpose, it would be crucial to relax the large-size condition in Eq. (27). Future studies could also generalize our consideration to more realistic solutions of relativistic hydrodynamics including viscous terms to study the spin polarization, uncovering or confirming approximate cancellation mechanisms.

Acknowledgments

The authors are grateful to Zheng Cao for useful discussions particularly in the early stage of this work. The authors thank Xu-Guang Huang and Yi Yin for helpful discussions. KF is supported in part by Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Nos. 22H05118, 25K24464, 26K00698. SP is supported in part by the National Key Research and Development Program of China under Contract No. 2022YFA1605500, by the Chinese Academy of Sciences (CAS) under Grant No. YSBR-088, and by the National Natural Science Foundation of China (NSFC) under Grant No. 12135011. DLW is supported in part by the NSFC under Grant No. 125B2110.

References

References

  1. [1]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
  2. [2]J.-H. Gao, S.-W. Chen, W.-t. Deng, Z.-T. Liang, Q. Wang, and X.-N. Wang, Phys. Rev. C 77, 044902 (2008), arXiv:0710.2943 [nucl-th].DOI
  3. [3]S. A. Voloshin, arXiv:nucl-th/0410089 (2004).DOI
  4. [4]L. Adamczyk et al. (STAR), Nature 548, 62 (2017), arXiv:1701.06657 [nucl-ex].
  5. [5]S. Acharya et al. (ALICE), Phys. Rev. C 101, 044611 (2020), [Erratum: Phys.Rev.C 105, 029902 (2022)], arXiv:1909.01281 [nucl-ex].
  6. [6]R. Abu Yassin et al. (HADES), Phys. Lett. B 835, 137506 (2022), arXiv:2207.05160 [nucl-ex].
  7. [7]I. Karpenko and F. Becattini, Eur. Phys. J. C 77, 213 (2017), arXiv:1610.04717 [nucl-th].DOI
  8. [8]H. Li, L.-G. Pang, Q. Wang, and X.-L. Xia, Phys. Rev. C 96, 054908 (2017), arXiv:1704.01507 [nucl-th].DOI
  9. [9]Y. Sun and C. M. Ko, Phys. Rev. C 96, 024906 (2017), arXiv:1706.09467 [nucl-th].DOI
  10. [10]D.-X. Wei, W.-T. Deng, and X.-G. Huang, Phys. Rev. C 99, 014905 (2019), arXiv:1810.00151 [nucl-th].DOI
  11. [11]O. Vitiuk, L. V. Bravina, and E. E. Zabrodin, Phys. Lett. B 803, 135298 (2020), arXiv:1910.06292 [hep-ph].DOI
  12. [12]B. Fu, K. Xu, X.-G. Huang, and H. Song, Phys. Rev. C 103, 024903 (2021), arXiv:2011.03740 [nucl-th].DOI
  13. [13]A. Lei, D. Wang, D.-M. Zhou, B.-H. Sa, and L. P. Csernai, Phys. Rev. C 104, 054903 (2021), arXiv:2110.13485 [nucl-th].DOI
  14. [14]S. Ryu, V. Jupic, and C. Shen, Phys. Rev. C 104, 054908 (2021), arXiv:2106.08125 [nucl-th].DOI
  15. [15]X.-Y. Wu, C. Yi, G.-Y. Qin, and S. Pu, Phys. Rev. C 105, 064909 (2022), arXiv:2204.02218 [hep-ph].DOI
  16. [16]F. Becattini, V. Chandra, L. Del Zanna, and E. Grossi, Annals Phys. 338, 32 (2013), arXiv:1303.3431 [nucl-th].DOI
  17. [17]R.-h. Fang, L.-g. Pang, Q. Wang, and X.-n. Wang, Phys. Rev. C 94, 024904 (2016), arXiv:1604.04036 [nucl-th].DOI
  18. [18]J. Adam et al. (STAR), Phys. Rev. Lett. 123, 132301 (2019), arXiv:1905.11917 [nucl-ex].DOI
  19. [19]S. Acharya et al. (ALICE), Phys. Rev. Lett. 128, 172005 (2022), arXiv:2107.11183 [nucl-ex].
  20. [20]M. Abdulhamid et al. (STAR), Phys. Rev. Lett. 131, 202301 (2023), arXiv:2303.09074 [nucl-ex].DOI
  21. [21]A. Hayrapetyan et al. (CMS), Phys. Rev. Lett. 135, 132301 (2025), arXiv:2502.07898 [nucl-ex].
  22. [22]F. Becattini and I. Karpenko, Phys. Rev. Lett. 120, 012302 (2018), arXiv:1707.07984 [nucl-th].DOI
  23. [23]X.-L. Xia, H. Li, Z.-B. Tang, and Q. Wang, Phys. Rev. C 98, 024905 (2018), arXiv:1803.00867 [nucl-th].DOI
  24. [24]W. Florkowski, A. Kumar, R. Ryblewski, and A. Mazeliauskas, Phys. Rev. C 100, 054907 (2019), arXiv:1904.00002 [nucl-th].DOI
  25. [25]H.-Z. Wu, L.-G. Pang, X.-G. Huang, and Q. Wang, Phys. Rev. Research 1, 033058 (2019), arXiv:1906.09385 [nucl-th].DOI
  26. [26]X.-L. Xia, H. Li, X.-G. Huang, and H. Z. Huang, Phys. Rev. C 100, 014913 (2019), arXiv:1905.03120 [nucl-th].DOI
  27. [27]F. Becattini, G. Cao, and E. Speranza, Eur. Phys. J. C 79, 741 (2019), arXiv:1905.03123 [nucl-th].DOI
  28. [28]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
  29. [29]J.-W. Chen, S. Pu, Q. Wang, and X.-N. Wang, Phys. Rev. Lett. 110, 262301 (2013), arXiv:1210.8312 [hep-th].DOI
  30. [30]Y. Hidaka, S. Pu, and D.-L. Yang, Phys. Rev. D 95, 091901 (2017), arXiv:1612.04630 [hep-th].DOI
  31. [31]Y. Hidaka, S. Pu, and D.-L. Yang, Phys. Rev. D 97, 016004 (2018), arXiv:1710.00278 [hep-th].DOI
  32. [32]J.-H. Gao and Z.-T. Liang, Phys. Rev. D 100, 056021 (2019), arXiv:1902.06510 [hep-ph].DOI
  33. [33]N. Weickgenannt, X.-L. Sheng, E. Speranza, Q. Wang, and D. H. Rischke, Phys. Rev. D 100, 056018 (2019), arXiv:1902.06513 [hep-ph].DOI
  34. [34]Y.-C. Liu, K. Mameda, and X.-G. Huang, Chin. Phys. C 44, 094101 (2020), [Erratum: Chin.Phys.C 45, 089001 (2021)], arXiv:2002.03753 [hep-ph].DOI
  35. [35]N. Weickgenannt, E. Speranza, X.-L. Sheng, Q. Wang, and D. H. Rischke, Phys. Rev. Lett. 127, 052301 (2021), arXiv:2005.01506 [hep-ph].DOI
  36. [36]N. Weickgenannt, E. Speranza, X.-L. Sheng, Q. Wang, and D. H. Rischke, Phys. Rev. D 104, 016022 (2021), arXiv:2103.04896 [nucl-th].DOI
  37. [37]X.-L. Sheng, N. Weickgenannt, E. Speranza, D. H. Rischke, and Q. Wang, Phys. Rev. D 104, 016029 (2021), arXiv:2103.10636 [nucl-th].DOI
  38. [38]Y. Hidaka, S. Pu, Q. Wang, and D.-L. Yang, Prog. Part. Nucl. Phys. 127, 103989 (2022), arXiv:2201.07644 [hep-ph].DOI
  39. [39]S. Fang, S. Pu, and D.-L. Yang, Phys. Rev. D 106, 016002 (2022), arXiv:2204.11519 [hep-ph].DOI
  40. [40]W.-B. Dong, Y.-L. Yin, and Q. Wang, Phys. Rev. C 106, 054909 (2022), arXiv:2209.12402 [nucl-th].DOI
  41. [41]S. Fang, S. Pu, and D.-L. Yang, Phys. Rev. D 109, 034034 (2024), arXiv:2311.15197 [hep-ph].arxiv.org/abs/2311.15197
  42. [42]S. Fang and S. Pu, Phys. Rev. D 111, 034015 (2025), arXiv:2408.09877 [hep-ph].DOI
  43. [43]D. Montenegro, L. Tinti, and G. Torrieri, Phys. Rev. D 96, 076016 (2017), arXiv:1703.03079 [hep-th].DOI
  44. [44]W. Florkowski, B. Friman, A. Jaaiswal, and E. Speranza, Phys. Rev. C 97, 041901 (2018), arXiv:1705.00587 [nucl-th].DOI
  45. [45]F. Becattini, W. Florkowski, and E. Speranza, Phys. Lett. B 789, 419 (2019), arXiv:1807.10994 [hep-th].DOI
  46. [46]W. Florkowski, A. Kumar, and R. Rybleswki, Prog. Part. Nucl. Phys. 108, 103709 (2019), arXiv:1811.04409 [nucl-th].DOI
  47. [47]K. Hattori, M. Hongo, X.-G. Huang, M. Matsuo, and H. Taya, Phys. Lett. B 795, 100 (2019), arXiv:1901.06615 [hep-th].DOI
  48. [48]K. Fukushima and S. Pu, Phys. Lett. B 817, 136346 (2021), arXiv:2010.01608 [hep-th].DOI
  49. [49]S. Li, M. A. Stephanov, and H.-U. Yee, Phys. Rev. Lett. 127, 082302 (2021), arXiv:2011.12318 [hep-th].DOI
  50. [50]M. Hongo, X.-G. Huang, M. Kaminski, M. Stephanov, and H.-U. Yee, JHEP 11, 150 (2021), arXiv:2107.14231 [hep-th].DOI
  51. [51]D. She, A. Huang, D. Hou, and J. Liao, Sci. Bull. 67, 2265 (2022), arXiv:2105.04060 [nucl-th].DOI
  52. [52]D.-L. Wang, S. Fang, and S. Pu, Phys. Rev. D 104, 114043 (2021), arXiv:2107.11726 [nucl-th].DOI
  53. [53]N. Weickgenannt, D. Wagner, E. Speranza, and D. H. Rischke, Phys. Rev. D 106, L0091901 (2022), arXiv:2208.01955 [nucl-th].DOI
  54. [54]N. Weickgenannt, D. Wagner, E. Speranza, and D. H. Rischke, Phys. Rev. D 106, 096014 (2022), arXiv:2203.04766 [nucl-th].DOI
  55. [55]Z. Cao, K. Hattori, M. Hongo, X.-G. Huang, and H. Taya, PTEP 2022, 071D01 (2022), arXiv:2205.08051 [hep-th].DOI
  56. [56]S. Bhadury, W. Florkowski, A. J. Kumar, and R. Rybleswki, Phys. Rev. Lett. 129, 192301 (2022), arXiv:2204.01357 [nucl-th].DOI
  57. [57]W. Florkowski and M. Hontarenko, Phys. Rev. Lett. 134, 082302 (2025), arXiv:2405.03263 [hep-ph].DOI
  58. [58]X.-G. Huang, Nucl. Sci. Tech. 36, 208 (2025), arXiv:2411.11753 [nucl-th].DOI
  59. [59]S. Fang, K. Fukushima, S. Pu, and D.-L. Wang, arXiv:2506.20698 [nucl-th] (2025).arxiv.org/abs/2506.20698
  60. [60]S. Y. F. Liu and Y. Yin, JHEP 07, 188 (2021), arXiv:2103.09200 [hep-ph].DOI
  61. [61]F. Becattini, M. Buzzegoli, and A. Palermo, Phys. Lett. B 820, 136519 (2021), arXiv:2103.10917 [nucl-th].DOI
  62. [62]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
  63. [63]F. Becattini, M. Buzzegoli, G. Inghirami, I. Karpenko, and A. Palermo, Phys. Rev. Lett. 127, 272302 (2021), arXiv:2103.14621 [nucl-th].DOI
  64. [64]C. Yi, S. Pu, and D.-L. Yang, Phys. Rev. C 104, 064901 (2021), arXiv:2106.00238 [hep-ph].DOI
  65. [65]W. Florkowski, A. Kumar, A. Mazeliauskas, and R. Ryblewski, Phys. Rev. C 105, 064901 (2022), arXiv:2112.02799 [hep-ph].DOI
  66. [66]M. Buzzegoli, F. Becattini, G. Inghirami, I. Karpenko, and A. Palermo, Acta Phys. Polon. Supp. 16, 1 (2023), arXiv:2208.04449 [nucl-th].arxiv.org/abs/2208.04449
  67. [67]A. Palermo, F. Becattini, M. Buzzegoli, G. Inghirami, and I. Karpenko, EPJ Web Conf. 276, 01026 (2023), arXiv:2208.09874 [nucl-th].DOI
  68. [68]A. Palermo, E. Grossi, I. Karpenko, and F. Becattini, Eur. Phys. J. C 84, 920 (2024), arXiv:2404.14295 [nucl-th].arxiv.org/abs/2404.14295
  69. [69]C. Yi, X.-Y. Wu, J. Zhu, S. Pu, and G.-Y. Qin, Phys. Rev. C 111, 044901 (2025), arXiv:2408.04296 [hep-ph].arxiv.org/abs/2408.04296
  70. [70]F. Becattini, M. Buzzegoli, T. Niida, S. Pu, A.-H. Tang, and Q. Wang, Int. J. Mod. Phys. E 33, 2430006 (2024), arXiv:2402.04540 [nucl-th].arxiv.org/abs/2402.04540
  71. [71]T. Niida and S. A. Voloshin, Int. J. Mod. Phys. E 33, 2430010 (2024), arXiv:2404.11042 [nucl-ex].DOI
  72. [72]X.-L. Sheng, F. Becattini, and D. Roselli, arXiv:2509.14301 [nucl-th] (2025).DOI
  73. [73]S. Alzhrani, S. Ryu, and C. Shen, Phys. Rev. C 106, 014905 (2022), arXiv:2203.15718 [nucl-th].DOI
  74. [74]A. Arslan, W.-B. Dong, G.-L. Ma, S. Pu, and Q. Wang, Phys. Rev. C 111, 044911 (2025), arXiv:2411.17285 [nucl-th].arxiv.org/abs/2411.17285
  75. [75]A. Arslan, W.-B. Dong, C. Gale, S. Jeon, Q. Wang, and X.-Y. Wu, arXiv:2509.00796 [nucl-th] (2025).arxiv.org/abs/2509.00796
  76. [76]S. S. Gubser, Phys. Rev. D 82, 085027 (2010), arXiv:1006.0006 [hep-th].DOI
  77. [77]S. S. Gubser and A. Yarom, Nucl. Phys. B 846, 469 (2011), arXiv:1012.1314 [hep-th].DOI
  78. [78]Y. Hatta, J. Noronha, G. Torrieri, and B.-W. Xiao, Phys. Rev. D 90, 074026 (2014), arXiv:1407.5952 [hep-ph].DOI
  79. [79]X. Ren, J.-Y. Hu, H.-j. Xu, and S. Pu, arXiv:2604.11058 [nucl-th] (2026).arxiv.org/abs/2604.11058
  80. [80]T. Csörgő, F. Grassi, Y. Hama, and T. Kodama, Phys. Lett. B 565, 107 (2003), arXiv:nucl-th/0305059.DOI
  81. [81]Y. Hatta, J. Noronha, and B.-W. Xiao, Phys. Rev. D 89, 114011 (2014), arXiv:1403.7693 [hep-th].DOI
  82. [82]S. Weinberg, Gravitation and Cosmology: Principles and Applications of the General Theory of Relativity (John Wiley and Sons, New York, 1972).
  83. [83]J. D. Bjorken, Phys. Rev. D 27, 140 (1983).
  84. [84]K. Yagi, T. Hatsuda, and Y. Miake, Quark-gluon plasma: From big bang to little bang, Cambridge Monographs on Particle Physics, Nuclear Physics and Cosmology, Vol. 23 (Cambridge University Press, Cambridge, 2005).
  85. [85]V. Henner, T. Belozerova, and K. Forinash, Mathemat- ical Methods in Physics: Partial Differential Equations, Fourier Series, and Special Functions (A K Peters/CRC Press, New York, 2009).
  86. [86]F. J. Belinfante, Physica 6, 887 (1939).
  87. [87]F. J. Belinfante, Physica 7, 449 (1940).
  88. [88]L. Rosenfeld, Mém. Acad. Roy. Belg. 18, 1 (1940).
  89. [89]Z. Cao, Private communication.

Paper details

vlnc:W_xnzh9ttz
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

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