betaFeedback

WorksPhysics and Astronomy

Exact Polarization of Particles of Any Spin at Global Equilibrium

Abstract

The polarization of the ΛΛ particle offers the unique opportunity to study the hydrodynamic gradients in the Quark-Gluon Plasma formed in heavy-ion collisions. However, the theoretical formula commonly used to calculate polarization is only a linear order expansion in thermal vorticity and neglects higher-order corrections. Here, I present an exact calculation to all orders in (constant) thermal vorticity at global equilibrium, obtaining the analytic form of the spin density matrix and the polarization vector for massive particles of any spin. Finally, I extend these results to local equilibrium and assess their phenomenological impact by numerically calculating the polarization vector in a 3+1 hydrodynamic simulation.

Introduction

The experimental measurement of the polarization vector of Λ\Lambda particles in heavy ion collisions paved a new way in the hydrodynamic studies of the QGP [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14]. In fact, with spin polarization one can access the thermodynamic gradients in the plasma at freeze-out. According to the hydrodynamic model and based on a statistical calculation, it was found early on that the polarization vector of massive spin-SS fields is [15, 16, 17]:

Sμ(p)=−S(S+1)312mϵμνρσpσ∫dΣ⋅p ϖνρn(p)[1+(−1)2Sn(p)]∫dΣ⋅p n(p)(1)S^\mu(p)=-\frac{S(S+1)}{3}\frac{1}{2m}\epsilon^{\mu\nu\rho\sigma}p_\sigma\frac{\int\mathrm{d}\Sigma\cdot p\,\varpi_{\nu\rho}n(p)[1+(-1)^{2S}n(p)]}{\int\mathrm{d}\Sigma\cdot p\,n(p)} \tag*{(1)}

where pp is the four-momentum of the particle at hand, the integral is calculated on the decoupling hypersurface, n(p)=[eβ⋅p+(−1)2S+1]−1n(p)=[e^{\beta\cdot p}+(-1)^{2S+1}]^{-1} is the Fermi-Dirac or the Bose-Einstein distribution function and ϖ\varpi is the thermal vorticity:

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

βμ\beta^\mu being the four-temperature vector.

Additional contributions to spin polarization have been found later on, including the thermal shear [18, 19, 20, 21] and the spin hall effect [22, 23, 24], which I will not discuss further here. It is important to stress that all the terms mentioned so far are the leading order contributions to the spin vector in linear response theory. Even in the case of global equilibrium, where the only contribution to the spin vector comes from the constant thermal vorticity, there are no results beyond the first order of approximation in ϖ\varpi. Here, I show how to compute the spin vector to all orders in ϖ\varpi for fields of any spin at global equilibrium [25]. I will also provide a finite and analytic expression of the spin density matrix, which can be used to compute any spin observable, including the vector meson spin alignment.

Spin density matrix and polarization vector

To compute the polarization vector or any other spin-related quantity, one uses the spin density matrix Θ\Theta. For massive fields of spin SS, the spin density matrix is a (2S+1)×(2S+1)(2S+1)\times(2S+1) hermitian matrix, characterized by S2−1S^2-1 real numbers. The definition of the spin density matrix adopted in the context of relativistic heavy-ion collisions is [26]:

Θsr(p)=Tr⁡(ρ^a^r†(p)a^s(p))∑lTr⁡(ρ^a^l†(p)a^l(p))=⟨a^r†(p)a^s(p)⟩∑l⟨a^l†(p)a^l(p)⟩.(3)\Theta_{sr}(p)=\frac{\operatorname{Tr}\left(\hat{\rho}\hat{a}^{\dagger}_{r}(p)\hat{a}_{s}(p)\right)}{\sum_l\operatorname{Tr}\left(\hat{\rho}\hat{a}^{\dagger}_{l}(p)\hat{a}_{l}(p)\right)}=\frac{\left\langle\hat{a}^{\dagger}_{r}(p)\hat{a}_{s}(p)\right\rangle}{\sum_l\left\langle\hat{a}^{\dagger}_{l}(p)\hat{a}_{l}(p)\right\rangle}. \tag*{(3)}

In the above equation, ρ^\hat{\rho} is the statistical operator, and a^\hat{a} and a^†\hat{a}^{\dagger} are the annihilation and creation operators of states with momentum pp and spin rr, ss.

The spin density matrix is related to the spin vector SμS^\mu through [26]:

Sμ(p)=∑i=13[p]μitr⁡(Θ(p)DS(Ji)),(4)S^\mu(p)=\sum_{i=1}^{3}[p]^\mu{}_{i}\operatorname{tr}\left(\Theta(p)D^{S}(J^{i})\right), \tag*{(4)}

where DS(Ji)D^{S}(J^{i}) is the ii-th generator of SO(3) in the SS-representation. The transformation denoted as [p][p] is the so-called standard boost: a Lorentz transformation that maps the rest frame momentum p=(m,0)p=(m,\mathbf{0}) to the momentum p=(ϵ,p)p=(\epsilon,\mathbf{p}), i.e. pμ=[p]μνpνp^\mu=[p]^\mu{}_{\nu}p^\nu. Other formulae for the polarization vector, such as those involving the Wigner function, are always derived from Eqs. (3) and (4).

Exact spin physics at global equilibrium

From equation (4), the polarization vector can be computed once the spin density matrix is known. Here, I will compute exactly the spin density matrix for fields of any spin at global equilibrium.

I will consider the following density operator:

ρ^=1Zexp⁡[−b⋅P^+ϖ:J^2+ζQ^],(5)\hat{\rho}=\frac{1}{Z}\exp\left[-b\cdot\hat{P}+\frac{\varpi:\hat{J}}{2}+\zeta\hat{Q}\right], \tag*{(5)}

where Q^\hat{Q} is the charge operator, and P^\hat{P} and J^\hat{J} are the generators of the Poincaré group: the four-momentum and the angular momentum-boost operator, respectively. The vector bμb^\mu and antisymmetric tensor ϖμν\varpi^{\mu\nu}, that is the thermal vorticity, are constants, and they define the four-temperature vector as the Killing vector:

βμ(x)=bμ+ϖμνxν.\beta^\mu(x)=b^\mu+\varpi^{\mu\nu}x_\nu.

In addition, ζ=μ/T\zeta=\mu/T is also constant, μ\mu being the chemical potential. The proper temperature is T=(β⋅β)−1T=(\sqrt{\beta\cdot\beta})^{-1}.

In Refs. [27, 28], it was realized that factorizing the operator (5) and performing the analytic continuation ϖ↦−iϕ\varpi\mapsto-i\phi provides a powerful technique to compute exact expectation values. This method allows the calculation of ⟨a^r†(p)a^s(p)⟩\langle\hat{a}^{\dagger}_{r}(p)\hat{a}_{s}(p)\rangle:

⟨a^r†(p)a^t(p′)⟩=2ε′∑n=1∞(−1)2S(n+1)δ3(Λnp−p′)D(W(Λn,p))tse−bˉ⋅∑k=1nΛkpenζ,(6)\left\langle\hat{a}^{\dagger}_{r}(p)\hat{a}_{t}(p')\right\rangle= 2\varepsilon' \sum_{n=1}^{\infty}(-1)^{2S(n+1)}\delta^{3}\left(\Lambda^{n}p-p'\right)D\left(W\left(\Lambda^{n},p\right)\right)_{ts}e^{-\bar{b}\cdot\sum_{k=1}^{n}\Lambda^{k}p}e^{n\zeta}, \tag*{(6)}

where Λ=exp⁡[−iϕ⋅J/2]\Lambda=\exp[-i\phi\cdot J/2] is the Lorentz transformation with parameter ϕ\phi equal to the imaginary vorticity and W(Λ,p)=[Λp]−1Λ[p]W(\Lambda,p)=[\Lambda p]^{-1}\Lambda[p] is the so-called Wigner rotation. The vector b~(ϕ)\tilde{b}(\phi) is:

b~μ(ϕ)=∑k=0∞1(k+1)!ϕ ν1μϕ ν2ν1⋯ϕ νkνk−1bνk,(7)\tilde{b}^{\mu}(\phi)=\sum_{k=0}^{\infty}\frac{1}{(k+1)!}\phi^{\mu}_{\ \nu_{1}}\phi^{\nu_{1}}_{\ \nu_{2}}\cdots\phi^{\nu_{k-1}}_{\ \nu_{k}}b^{\nu_{k}}, \tag*{(7)}

and its form is a consequence of the fact that the Poincaré group is non-Abelian.

Plugging eq. (6) into eq. (3), yields:

Θrs(p)=∑n=1∞(−1)2S(n+1)e−bˉ⋅∑k=1nΛkpenζDrs(W(Λn,p))δ3(Λnp−p)∑n=1∞(−1)2S(n+1)e−bˉ⋅∑k=1nΛkpenζtr⁡[D(W(Λn,p))]δ3(Λnp−p).(8)\Theta_{rs}(p)=\frac{\sum_{n=1}^{\infty}(-1)^{2S(n+1)}e^{-\bar{b}\cdot\sum_{k=1}^{n}\Lambda^{k}p}e^{n\zeta}D_{rs}\left(W\left(\Lambda^{n},p\right)\right)\delta^{3}\left(\Lambda^{n}p-p\right)}{\sum_{n=1}^{\infty}(-1)^{2S(n+1)}e^{-\bar{b}\cdot\sum_{k=1}^{n}\Lambda^{k}p}e^{n\zeta}\operatorname{tr}\left[D\left(W\left(\Lambda^{n},p\right)\right)\right]\delta^{3}\left(\Lambda^{n}p-p\right)}. \tag*{(8)}

To proceed further in the calculation, and as long as the thermal vorticity is imaginary, the constraint Λp=p\Lambda p=p must be enforced. Indeed, from eq. (6), one sees that if this is not the case ⟨a^s†(p)a^t(p)⟩=0\langle\hat{a}^{\dagger}_{s}(p)\hat{a}_{t}(p)\rangle=0, and the spin density matrix would become trivial.

Imposing the constraint Λp=p\Lambda p=p, it follows that:

ϕμν=ϵμνρσξρpσm,ξρ=−12mϵρμνσϕμνpσ.(9)\phi^{\mu\nu}=\epsilon^{\mu\nu\rho\sigma}\xi_{\rho}\frac{p_{\sigma}}{m},\qquad\xi^{\rho}=-\frac{1}{2m}\epsilon^{\rho\mu\nu\sigma}\phi_{\mu\nu}p_{\sigma}. \tag*{(9)}

Thanks to this simplification, and after some algebra (see Ref. [25] for the full derivation), the spin density matrix becomes:

Θ(p)=∑n=1∞(−1)2S(n+1)e−nb⋅penζe−inξ0⋅DS(J)∑n=1∞(−1)2S(n+1)e−nb⋅penζtr⁡(e−inξ0⋅DS(J)),\Theta(p)=\frac{\sum_{n=1}^{\infty}(-1)^{2S(n+1)}e^{-nb\cdot p}e^{n\zeta}e^{-in\xi_{0}\cdot D^{S}(J)}}{\sum_{n=1}^{\infty}(-1)^{2S(n+1)}e^{-nb\cdot p}e^{n\zeta}\operatorname{tr}\left(e^{-in\xi_{0}\cdot D^{S}(J)}\right)},

where ξ0μ=[p]  ν−1μξν\xi_{0}^{\mu}=[p]^{-1\mu}_{\ \ \nu}\xi^{\nu}, ξ0\xi_{0} being its spatial part, and DS(J)D^{S}(J) is the three-vector of the generators of the rotation group in the SS-representation. As long as the absolute value of the eigenvalues of the matrix exp⁡(−b⋅p+ζ)exp⁡[−iξ0⋅DS(J)]\exp(-b\cdot p+\zeta)\exp[-i\xi_{0}\cdot D^{S}(J)] is less than one (that is b⋅p>ζb\cdot p>\zeta) the series can be summed as a geometric series of matrices [29]. Defining:

θμ=−12mϵμνρσϖνρpσ,(10)\theta^{\mu}=-\frac{1}{2m}\epsilon^{\mu\nu\rho\sigma}\varpi_{\nu\rho}p_{\sigma}, \tag*{(10)}

the spin density matrix after the analytic continuation ϕ↦iϖ\phi\mapsto i\varpi reads:

Θ(p)=[(−1)2SI−e−b⋅p+θ0⋅DS(J)]−1−(−1)2SI∑k=−SS(eb⋅p−k−θ2−(−1)2S)−1.(11)\Theta(p)=\frac{\left[(-1)^{2S}\mathbb{I}-e^{-b\cdot p+\theta_{0}\cdot D^{S}(J)}\right]^{-1}-(-1)^{2S}\mathbb{I}}{\sum_{k=-S}^{S}\left(e^{b\cdot p-k\sqrt{-\theta^{2}}}-(-1)^{2S}\right)^{-1}}. \tag*{(11)}

From eq. (11), one can compute all spin-related quantities, including spin polarization and spin alignment. In particular, the mean spin vector reads, using eq. (4):

Sμ(p)=θμ−θ2∑k=−SSk[eb⋅p−ζ−k−θ2−(−1)2S]−1∑k=−SS[eb⋅p−ζ−k−θ2−(−1)2S]−1.(12)S^{\mu}(p)=\frac{\theta^{\mu}}{\sqrt{-\theta^{2}}}\frac{\sum_{k=-S}^{S}k\left[e^{b\cdot p-\zeta-k\sqrt{-\theta^{2}}}-(-1)^{2S}\right]^{-1}}{\sum_{k=-S}^{S}\left[e^{b\cdot p-\zeta-k\sqrt{-\theta^{2}}}-(-1)^{2S}\right]^{-1}}. \tag*{(12)}

This formula reproduces the previous literature, including the small vorticity approximation and the case of Boltzmann statistics. In the case of very large thermal vorticity, it predicts a polarization equal to one [25].

Polarization in heavy-ion collisions

One can now evaluate the effect of the formula (12) in phenomenological applications. Although the derivation holds in global equilibrium, eq. (12) can be generalized to local equilibrium by promoting the thermal vorticity to be a local variable, i.e. ϖ↦ϖ(x)\varpi\mapsto\varpi(x), and integrating eq. (12) over the decoupling hypersurface.

SLEμ(p)=∫dΣ⋅p n(p)Sμ(p)∫dΣ⋅p n(p).(13)S^\mu_{\mathrm{LE}}(p)=\frac{\int\mathrm{d}\Sigma\cdot p\,n(p)S^\mu(p)}{\int\mathrm{d}\Sigma\cdot p\,n(p)}. \tag*{(13)}

Here, in Figure 1, I report the result of Eq. (13) compared to that of Eq. (1) in Au-Au collisions at sNN=30\sqrt{s_{NN}}=30 GeV in the centrality class 10-60%, simulated using the code vHLLE [30] for the hydrodynamic evolution. The initial state is an averaged entropy density profile generated by GLISSANDO v.2.702 [31]. For the figure, I consider the case of Dirac fermions at midrapidity. I remind the reader that the polarization is related to the spin vector by Pμ=Sμ/SP^\mu=S^\mu/S, where SS is the spin. One can see that the two formulae give the same result, and it follows that the linear approximation is a good one, at least for sNN>30\sqrt{s_{NN}}>30 GeV.

The components of the polarization vector along the total angular momentum and the beam direction as functions of the azimuthal angle

Figure 1. The components of the polarization vector along the total angular momentum (left panel) and the beam direction (right panel) as a function of the azimuthal angle. The blue dashed line is the linear approximation (1), whereas the solid orange line corresponds to Eq. (13).

Conclusions

In this work, I have shown how to calculate the spin density matrix and the polarization vector for massive spin-SS fields to all orders in thermal vorticity and obtained an analytic finite expression for both quantities. Using the spin density matrix, all spin observables can be computed, including alignment. The polarization vector is a finite sum, representing the average of the spin state weighted by Bose or Fermi distribution functions.

I have shown a numerical calculation assessing the importance of high-order corrections in vorticity to the phenomenology of heavy ion collisions. As it turns out, the exact solution and the linear approximation precisely coincide for AuAu collisions at 30 GeV. Yet, a small discrepancy may arise in lower-energy collisions, where the thermal vorticity is larger. This work shows that the linear approximation of the polarization vector is an excellent one in most practical applications to heavy-ion collisions.

References

  1. [1]L. Adamczyk et al. (STAR), Nature 548, 62 (2017), 1701.06657
  2. [2]J. Adam et al. (STAR), Phys. Rev. Lett. 123, 132301 (2019), 1905.11917
  3. [3]J. Adam et al. (STAR), Phys. Rev. Lett. 126, 162301 (2021), 2012.13601
  4. [4]S. Acharya et al. (ALICE), Phys. Rev. Lett. 128, 172005 (2022), 2107.11183
  5. [5]S. Acharya et al. (ALICE), Phys. Rev. C 101, 044611 (2020), [Erratum: Phys.Rev.C 105, 029902 (2022)], 1909.01281
  6. [6]W. Florkowski, B. Friman, A. Jaiswal, E. Speranza, Phys. Rev. C 97, 041901 (2018), 1705.00587
  7. [7]W. Florkowski, A. Kumar, R. Ryblewski, Prog. Part. Nucl. Phys. 108, 103709 (2019), 1811.04409
  8. [8]N. Weickgenannt, X.L. Sheng, E. Speranza, Q. Wang, D.H. Rischke, Phys. Rev. D 100, 056018 (2019), 1902.06513
  9. [9]N. Weickgenannt, E. Speranza, X.I. Sheng, Q. Wang, D.H. Rischke, Phys. Rev. Lett. 127, 052301 (2021), 2005.01506
  10. [10]J.H. Gao, Z.T. Liang, Q. Wang, X.N. Wang, Lect. Notes Phys. 987, 195 (2021), 2009.04803
  11. [11]F. Becattini, M.A. Lisa, Ann. Rev. Nucl. Part. Sci. 70, 395 (2020), 2003.03640
  12. [12]F. Becattini, Rept. Prog. Phys. 85, 122301 (2022), 2204.01144
  13. [13]J.H. Gao, G.L. Ma, S. Pu, Q. Wang, Nucl. Sci. Tech. 31, 90 (2020), 2005.10432
  14. [14]M. Hongo, X.G. Huang, M. Kaminski, M. Stephanov, H.U. Yee, JHEP 11, 150 (2021), 2107.14231
  15. [15]F. Becattini, V. Chandra, L. Del Zanna, E. Grossi, Annals Phys. 338, 32 (2013), 1303.3431
  16. [16]R.H. Fang, L.g. Pang, Q. Wang, X.n. Wang, Phys. Rev. C 94, 024904 (2016), 1604.04036
  17. [17]F. Becattini, G. Cao, E. Speranza, Eur. Phys. J. C 79, 741 (2019), 1905.03123
  18. [18]F. Becattini, M. Buzzegoli, A. Palermo, Phys. Lett. B 820, 136519 (2021), 2103.10917
  19. [19]F. Becattini, M. Buzzegoli, G. Inghirami, I. Karpenko, A. Palermo, Phys. Rev. Lett. 127, 272302 (2021), 2103.14621
  20. [20]B. Fu, S.Y.F. Liu, L. Pang, H. Song, Y. Yin, Phys. Rev. Lett. 127, 142301 (2021), 2103.10403
  21. [21]S.Y.F. Liu, Y. Yin, JHEP 07, 188 (2021), 2103.09200
  22. [22]B. Fu, L. Pang, H. Song, Y. Yin (2022), 2201.12970
  23. [23]Y.B. Ivanov, Phys. Rev. C 103, L031903 (2021), 2012.07597
  24. [24]M. Buzzegoli (2022), 2211.04549
  25. [25]A. Palermo, F. Becattini, Eur. Phys. J. Plus 138, 547 (2023), 2304.02276
  26. [26]F. Becattini, Lect. Notes Phys. 987, 15 (2021), 2004.04050
  27. [27]F. Becattini, M. Buzzegoli, A. Palermo, JHEP 02, 101 (2021), 2007.08249
  28. [28]A. Palermo, M. Buzzegoli, F. Becattini, JHEP 10, 077 (2021), 2106.08340
  29. [29]J. Hubbard, B. Hubbard, Vector Calculus, Linear Algebra, and Differential Forms: A Unified Approach (Prentice Hall, 2002), ISBN 9780130414083, https://books.google.com/books?id=bW92QgAACAAJ
  30. [30]I. Karpenko, P. Huovinen, M. Bleicher, Comput. Phys. Commun. 185, 3016 (2014), 1312.4160
  31. [31]M. Rybczynski, G. Stefanek, W. Bronowski, P. Bozek, Comput. Phys. Commun. 185, 1759 (2014), 1310.5475

Paper details

vlnc:W_f2ktx8vk
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