betaFeedback

WorksPhysics and Astronomy

TT-matrix Analysis of Static Wilson Line Correlators from Lattice QCD at Finite Temperature

Abstract

We utilize a previously constructed thermodynamic TT-matrix approach to the quark-gluon plasma (QGP) to calculate Wilson line correlators (WLCs) of a static quark-antiquark pair and apply them to the results from 2+1-flavor lattice-QCD (lQCD) computations with realistic pion mass. The self-consistent TT-matrix results, which include constraints from the lQCD equation of state in the light-parton sector, can describe the lQCD data for WLCs fairly well once refinements of the input parameters are implemented. In particular, the input potential requires less screening than used in previous TT-matrix analyses. Pertinent predictions for the spectral and transport properties of the QGP are discussed, including the spatial diffusion coefficient for heavy quarks which turns out to have a rather weak temperature dependence, in approximate agreement with recent lQCD results.

Introduction

The study of the quark-gluon plasma (QGP) provides unique opportunities to understand how emergent (many-body) phenomena arise from the fundamental interactions between the partons in the Quantum Chromodynamics (QCD). In particular, heavy-flavor (HF) particles are widely regarded as excellent probes of the transport and hadronization properties of the QCD medium in ultrarelativistic heavy-ion collisions (URHICs) [1, 2, 3], for several reasons. Heavy-quark and -antiquark pairs are produced in initial hard processes and approximately conserved throughout the evolution of fireball. Their subsequent propagation through the medium can be characterized by well-defined transport coefficients, most notably the spatial diffusion coefficient. The interactions between heavy quarks and the medium occur with small energy transfer and thus enable the use of potential approximations. The spectra of HF particles can preserve a memory of their interaction history as their thermalization time, which is parametrically enhanced by their mass-over-temperature ratio, is comparable to or (for bottom) even longer than the fireball lifetime.

Bound states of heavy quark and antiquark, i.e., quarkonia, provide further insights into the properties of the QGP in URHICs [4, 5, 6], as their in-medium properties are closely related to the in-medium QCD force. These are, in turn, critical inputs to their transport properties which govern the abundance and transverse-momentum spectra of quarkonia [7]. For example, quarkonium kinetics critically depends on their inelastic reaction rates, which, in turn, depend on the in-medium binding energies and are closely related to individual heavy-quark (HQ) interactions with the medium. Lattice-QCD (lQCD) computations have provided ample information on the in-medium properties of heavy quarkonia through HQ free energies and Euclidean correlators [8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25] which constrain the spectral functions that can be computed in effective models and that subsequently serve as an input to phenomenological applications [26, 27, 28].

In the present paper we focus on for Wilson line correlators (WLCs) of a static quark antiquark pair at finite temperature. Pertinent lQCD results have recently been obtained using realistic 2+1-flavor lQCD calculations [29]. When compared to predictions from hard-thermal loop (HTL) perturbation theory, marked disagreement was found. On the other hand, when using schematic spectral functions based on parametrizations using different ansätze, it was inferred that the underlying potential exhibits a relatively weak screening, even at rather large distances, while the widths of the spectral peaks turned out to be substantial, albeit quantitatively with a rather large spread. These results were refined and essentially confirmed in a more recent analysis [30], and call for microscopic analysis in a nonperturbative approach. Toward this end, we will employ quantum many-body theory, i.e., a thermodynamic TT-matrix approach, that has been previously constructed to describe the interactions between partons in a strongly coupled QGP [31, 32]. Utilizing a Hamiltonian with a non-perturbative input potential, the 1- and 2-body correlation functions evaluated selfconsistently [33, 34, 35]. In the vacuum, the kernel is the standard Cornell potential, while its finite-temperature corrections have been constrained by lQCD data for the HQ free energy and equation of state (EoS). The WLCs from lQCD provide a novel opportunity to improve the constraints on the in-medium potential, especially since their dependence on euclidean time, τ\tau, for different values of spatial separation, rr, provides a much extended dynamical reach compared to the HQ free energies which are evaluated at τ=1/T\tau= 1/T.

The structure of this article is as follows. In § 2, we briefly recollect the key components of the thermodynamic TT-matrix approach as needed in the present context. In Sec. III introduces the formalism of static WLCs and how to compute them within the TT-matrix framework. The constraints imposed on the in-medium corrections to the potential from lQCD data for EoS and static WLCs are detailed in Sec. IV. In § 5 we compute the transport coefficients of charm and static quarks predicted by the newly inferred potential and compare the diffusion coefficient to recent lQCD results. Our summary and conclusions are contained in Sec. VI.

T-matrix Approach

The thermodynamic TT-matrix is a quantum many-body approach to evaluate 1- and 2-body correlation functions selfconsistently by resumming an infinite series of ladder diagrams; thus, it is suitable to study both bound and scattering states for strongly interacting systems. Originally devised for the study of HF particles within the QGP context [32, 33], this approach was subsequently extended to encompass the light-parton sector [36]. By reducing the 4-dimensional (4D) Bethe-Salpeter equation into a more manageable 3D Lippmann-Schwinger equation [37], followed by a partial-wave expansion, one obtains a 1D scattering equation [35, 36],

TijL,a(z,p,p′)=VijL,a(p,p′)+2π∫−∞∞k2dkVijL,a(p,k)Gij0(z,k)TijL,a(z,k,p′)(1)T_{ij}^{L,a}(z,p,p') = V_{ij}^{L,a}(p,p') + \frac{2}{\pi}\int_{-\infty}^{\infty} k^{2}dk V_{ij}^{L,a}(p,k)G_{ij}^{0}(z,k)T_{ij}^{L,a}(z,k,p') \tag*{(1)}

that features the intermediate 2-parton propagator,

Gij0(z,k)=∫−∞∞dω1dω2[1±ni(ω1)±nj(ω2)]z−ω1−ω2×ρi(ω1,k)ρj(ω2,k)(2)G_{ij}^{0}(z,k) = \int_{-\infty}^{\infty} d\omega_{1}d\omega_{2}\frac{[1\pm n_{i}(\omega_{1})\pm n_{j}(\omega_{2})]}{z-\omega_{1}-\omega_{2}}\times\rho_{i}(\omega_{1},k)\rho_{j}(\omega_{2},k) \tag*{(2)}

which is a convolution of two single-parton spectral functions,

ρi(ω,k)=−1πIm⁡Gi(ω+iϵ,k)(3)\rho_{i}(\omega,k)=-\frac{1}{\pi}\operatorname{Im}G_{i}(\omega+i\epsilon,k) \tag*{(3)}

which are given in terms of the (positive-energy projected) propagators,

Gi(ω,k)=1[ω−ϵi(k)−Σi(k)](4)G_{i}(\omega,k)=\frac{1}{[\omega-\epsilon_{i}(k)-\Sigma_{i}(k)]} \tag*{(4)}

In Eq. (1), VijL,aV_{ij}^{L,a} denotes the potential between particle ii and jj in color channel aa with angular-momentum LL, the nin_i are Bose (++) or Fermi (−-) distribution functions, ϵi=Mi2+k2\epsilon_i=\sqrt{M_i^2+k^2} on-shell energies with particle mass Mi=−12∫d3p(2π)3Viia=1(p)+Mi0M_i=-\frac{1}{2}\int\frac{d^3p}{(2\pi)^3}V_{ii}^{a=1}(\mathbf{p})+M_i^0 composed of the selfenergy from the color-singlet (a=1a=1) potential and the bare mass, Mi0M_i^0; the p,p′p,p' are the moduli of incoming and outgoing momenta in the center-of-mass (CM) frame. The 1-particle selfenergies, Σi(k)\Sigma_i(k), in Eq. (4) are calculated by closing the TT-matrix with an in-medium 1-parton propagator [35, 36].

The static in-medium potential in color-singlet state,

V~(r,T)=−43αs[e−mdrr+md]−σms[e−msr−(cbmsr)2−1](5)\widetilde{V}(r,T)=-\frac{4}{3}\alpha_s\left[\frac{e^{-m_d r}}{r}+m_d\right]-\frac{\sigma}{m_s}\left[e^{-m_s r-(c_bm_s r)^2}-1\right] \tag*{(5)}

reduces to the well-known Cornell potential, V~(r)=−43αsr+σr\widetilde{V}(r)=-\frac{4}{3}\frac{\alpha_s}{r}+\sigma r, in vacuum, where the coupling constant αs=0.27\alpha_s=0.27, and string tension, σ=0.225 GeV2\sigma=0.225\ {\rm GeV}^2, are fitted to the vacuum free energy from lQCD data [21, 27, 38, 39, 40, 41], with a string breaking distance of rSB=1r_{\rm SB}=1 fm. For the in-medium potential, mdm_d (msm_s) is the Debye screening mass for color-Coulomb (confining) interaction; the parameter cbc_b in the quadratic term of the exponential for the confining term controls the saturation of the confining potential, essentially mimicking in-medium string breaking. For the interactions between particles with finite masses, the static potential in momentum space acquires relativistic corrections induced by the underlying Lorentz structure [32]

Vij(p,p′)=RijvecVvec(p−p′)+RijscaVsca(p−p′)(6)V_{ij}(\mathbf{p},\mathbf{p}')=R_{ij}^{\rm vec}V^{\rm vec}(\mathbf{p}-\mathbf{p}')+R_{ij}^{\rm sca}V^{\rm sca}(\mathbf{p}-\mathbf{p}') \tag*{(6)}

where VvecV^{\rm vec} (VscaV^{\rm sca}) denotes the static vector (scalar) potential. One has

Rijvec=1+p2ϵi(p)ϵj(p)1+p′2ϵi(p′)ϵj(p′)(7)R_{ij}^{\rm vec}=\sqrt{1+\frac{p^2}{\epsilon_i(p)\epsilon_j(p)}}\sqrt{1+\frac{p'^2}{\epsilon_i(p')\epsilon_j(p')}} \tag*{(7)}
Rijsca=MiMjϵi(p)ϵj(p)MiMjϵi(p′)ϵj(p′)(8)R_{ij}^{\rm sca}=\sqrt{\frac{M_iM_j}{\epsilon_i(p)\epsilon_j(p)}}\sqrt{\frac{M_iM_j}{\epsilon_i(p')\epsilon_j(p')}} \tag*{(8)}

The color-Coulomb potential is characterized by an entirely vector Lorentz structure, while the confining potential is commonly assumed to be scalar [42, 43], in which case Vvec=VCV^{\rm vec}=V_C and Vsca=VSV^{\rm sca}=V_S. However, Refs. [44, 45, 46, 47, 48] have suggested that the confining potential exhibits a combination of vector and scalar Lorentz structures, expressed as Vvec=VC+(1−χ)VSV^{\rm vec}=V_C+(1-\chi)V_S and Vsca=χVSV^{\rm sca}=\chi V_S, rather than being exclusively scalar. In this context, the key parameter is the mixing coefficient, χ\chi, wherein χ=1\chi=1 represents a purely scalar confining potential, and values below one admix a vector component. In a previous study of 1/mQ1/m_Q corrections (mQm_Q: HQ mass) to the HQ interaction [35] we have found that a mixing coefficient of χ=0.6\chi=0.6 leads to a marked improvement of the spin-orbit and spin-spin splittings in vacuum charmonium and bottomonium spectroscopy [49] over the χ=1\chi=1 case. In addition, the resulting charm-quark diffusion coefficient showed better agreement with 2+1-flavor lQCD data [50]. Therefore, we also introduce the mixing effect in this study. The potential is extended to different color channels by multiplication with pertinent Casimir coefficients [35, 36].

Static Wilson Line Correlators from the T-matrix

The static Wilson line correlator in Euclidean time, which is amenable to lQCD computations, is connected to the static QQ‾Q\overline{Q} spectral function ρQQ‾\rho_{Q\overline{Q}}, through a Laplace transform,

W(r,τ,T)=∫−∞∞dEe−EτρQQ‾(E,r,T),(9)W(r,\tau,T)=\int_{-\infty}^{\infty}dE e^{-E\tau}\rho_{Q\overline{Q}}(E,r,T), \tag*{(9)}

where rr is the distance between QQ and Q‾\overline{Q}, and EE their total energy which is measured relative to the mass threshold of the bare HQ mass, 2MQ02M_Q^0 (numerically taken as 2×1042 \times10^4 GeV). The constituent static HQ mass is the sum of the bare mass and the mass shift originating from the selfenergy, i.e., MQ=MQ0+V~(r→∞)/2M_Q=M_Q^0+\widetilde{V}(r\to\infty)/2 [36]. The inversion of Eq. (9) is a challenging (if not ill-posed) problem, making it difficult to reconstruct spectral functions from lQCD results for WLCs. On the other hand, if one can calculate the WLCs with spectral functions obtained from an effective model using Eq. (9), the comparison of the model-derived WLCs with lattice data can test the model capabilities and provide microscopic insights.

In the TT-matrix formalism, the QQ‾Q\overline{Q} spectral function takes the same form as given in Ref. [36],

ρQQ‾(E,r,T)=−1πIm⁡[1E−V~(r,T)−Φ(r,T)ΣQQ‾(E,T)](10)\rho_{Q\overline{Q}}(E,r,T)=-\frac{1}{\pi}\operatorname{Im}\left[\frac{1}{E-\widetilde{V}(r,T)-\Phi(r,T)\Sigma_{Q\overline{Q}}(E,T)}\right] \tag*{(10)}

where V~(r,T)\widetilde{V}(r,T) is the static in-medium potential introduced in § 2. The two-body selfenergy, ΣQQ‾\Sigma_{Q\overline{Q}}, is related to the two-body propagator by

[GQQ‾0(E)]−1=E−V~(r→∞)−ΣQQ‾(E).(11)\left[G^0_{Q\overline{Q}}(E)\right]^{-1}=E-\widetilde{V}(r\to\infty)-\Sigma_{Q\overline{Q}}(E). \tag*{(11)}

The two-body selfenergy is reduced by interference effects that depend on the relative distance between the quarks and are sometimes referred to the “imaginary part of potential” [51]. In the TT-matrix formalism, these correspond to 3-body diagrams, which are difficult to calculate explicitly [36]. However, one can approximately implement them through an rr-dependent suppression factor (or interference function), ϕ(r)\phi(r) [36], i.e., ΣQQ‾(E,r)=ΣQQ‾(E)ϕ(r)\Sigma_{Q\overline{Q}}(E,r)=\Sigma_{Q\overline{Q}}(E)\phi(r). In perturbation theory, ϕ(r)\phi(r) has been calculated, essentially corresponding to an atanh-function (which vanishes for r→0r\to0 in the color-singlet channel and goes to 1 for r→∞r\to\infty) [51]. In a nonperturbative setting, it has been supplemented with an extra stretch-factor in the argument, which has been determined based on lQCD constraints from the static free energies [36]. The interference effect is expected to be significant for deeply bound heavy quarkonia and has been found to improve the description of Euclidean quarkonium correlators within the TT-matrix approach [36].

To facilitate the understanding of the physical meaning of the lattice results for the WLCs and to what extent these can constrain the QQ‾Q\overline{Q} spectral function, Ref. [29] has defined its nnth-order cumulants as

m1(r,τ,T)=−∂τln⁡W(r,τ,T),(12)m_1(r,\tau,T)=-\partial_\tau\ln W(r,\tau,T), \tag*{(12)}
mn=∂τmn−1(r,τ,T),n>1.(13)m_n=\partial_\tau m_{n-1}(r,\tau,T),\quad n>1. \tag*{(13)}

The first cumulant, m1m_1, essentially corresponds to an effective mass that is commonly used in lattice QCD. In general, the WLC contains many states at T=0T=0. At large τ\tau (i.e., small energies) it will be dominated by the ground state that defines the potential at zero temperature. At higher energies, excited states contributing to the WLC are related to hybrid potentials (such as those between DD-mesons) which will cause m1m_1 to have a rather complex behavior at small τ\tau. At non-zero temperature, where only data at relatively small τ\tau are available, this fact renders the analysis quite involved. We therefore employ the lQCD results for m1m_1 where the excited-state contributions to the WLC have been subtracted, following the procedure outlined in Ref. [29], which renders a much simpler behavior at small τ\tau.

Returning to the TT-matrix expression, Eq. (10), one can show by expanding m1m_1 for τ→0\tau\to0 and considering the analytic properties of the two-body selfenergy, that m1(r,τ=0,T)m_1(r,\tau=0,T) reduces to the potential V~(r,T)\widetilde{V}(r,T), see Appendix § 4.1 for more details. The second cumulant m2m_2, or equivalently the slope of m1m_1, is a measure of the width of ρQQ‾\rho_{Q\overline{Q}}, characterizing the interacting strength between QQ‾Q\overline{Q} and medium.

In-Medium Potentials Constrained by lQCD

In this section, we lay out the procedure of inferring the modifications to the in-medium potential by means of HQ WLCs and QGP EoS within a selfconsistent quantum many-body approach [35, 36]. It consists of 2 self-consistency loops, as follows. Initially, the parametrization of the in-medium potential is refined by calculating the cumulants of static WLCs (m1m_1) using the TT-matrix approach and fitting them to the lQCD data. The fit parameters in this step are the screening masses, mdm_d (color-Coulomb) and msm_s (confining), as well as cbc_b which controls the “string breaking”. Subsequently, these refined in-medium potentials are used to compute the EoS of QGP, and the main parameters in this step include the in-medium light-quark and gluon masses. Since the calculation of EoS requires the one-body spectral functions and two-body scattering amplitudes (which depend on each other through the parton selfenergies, cf. Eqs. (1)–(4)), one has to solve a selfconsistency problem which is done by numerical iteration. The selfenergies of the heavy quarks, calculated from heavy-light TT-matrices closed off with thermal parton spectral functions, are then reinserted into the calculation of the WLCs via the in-medium HQ spectral functions. A more refined in-medium two-body potential is created by refitting the screening masses to the lQCD WLC data, and this potential is then processed to provide a new fit to the EoS, and the process is iterated until convergence has been achieved.

In the remainder of this section, we discuss our numerical fits of the EoS (following Ref. [36]) in Sec. IV A and the results for the WLCs in Sec. IV B.

Equation of State

The equation of state is encoded in the pressure, P(T,μ)P(T,\mu), of a many-body system as a function of temperature and chemical potential(s). In the present context, lQCD data for the EoS are used to adjust the bare light-parton masses which are due to effects that are not explicitly part of the many-body calculations, e.g., due to quark or gluon condensates (similar to what is done in quasiparticle models). In the grand canonical ensemble the pressure is related to the grand potential per unit volume via Ω=−P\Omega=-P. For an interacting system, it can be calculated diagrammatically within the Luttinger-Ward-Baym (LWB) formalism where all closed-loop “skeleton” diagrams are computed with fully dressed propagators [52, 53, 54], also referred to as a 2-particle irreducible approach. This constitutes a conserving (i.e., thermodynamically consistent) approximation scheme, which is compatible with the ladder resummation in the selfconsistent TT-matrix and selfenergies. For a Hamiltonian approach to the QCD, this has been carried out in Refs. [36, 55]. In particular, the Luttinger-Ward functional, Φ\Phi, could be resummed utilizing a matrix-logarithm resummation technique which accounts for the possibility of dynamically formed bound states contributing to the EoS. In compact form the result can be written as

Ω=∑j∓dj∫dp~{ln⁡(−Gj(p~)−1)+[Σj(p~)−12log⁡Σj(p~)]Gj(p~)}.(14)\Omega=\sum_j \mp d_j\int d\tilde{p}\left\{\ln\left(-G_j(\tilde{p})^{-1}\right)+\left[\Sigma_j(\tilde{p})-\frac{1}{2}\log\Sigma_j(\tilde{p})\right]G_j(\tilde{p})\right\}. \tag*{(14)}

using the notation ∫dp~≡−β−1∑n∫d3p/(2π)3\int d\tilde{p}\equiv-\beta^{-1}\sum_n\int d^3p/(2\pi)^3 with p~≡(iωn,p)\tilde{p}\equiv(i\omega_n,\mathbf{p}) and β=1/T\beta=1/T. The summation in Eq. (14) includes all light-parton channels with the spin-color degeneracy djd_j, and ∓\mp denotes bosons (upper) or fermions (lower). The three pieces in Eq. (14), ln⁡(−G−1)\ln(-G^{-1}), ΣG\Sigma G, and log⁡ΣG\log\Sigma G, correspond to the contributions from quasiparticles, selfenergies and the Luttinger-Ward functional (LWF) characterizing two-body interactions, respectively.

We present the fits to the pressure as obtained from lQCD [41] together with LWF contributions, as well as the corresponding light-parton masses in Figure 1. In accordance with the findings in Refs. [35, 36], the two-body interaction (LWF) plays an increasingly important role as temperature decreases, becoming the leading contribution at T=0.195T=0.195 GeV, indicating a transition in the degrees of freedom (from QGP to mesons and diquarks) in the system.

Figure 1. Left panel: the pressure scaled by T4T^4 (blue line) compared to the pertinent 2+1-flavor lQCD data [41], as well as the LWF contribution (dashed line). Right panel: the total in-medium masses (i.e., bare plus selfenergies) for light quarks (blue) and gluons (orange) as a function of temperature.

Static Wilson Line Correlators

Next we turn to the fits to the cumulants of the WLCs, which are, in fact, part of the combined fit procedure as outlined at the beginning of this section, cf. Figure 2. A fair overall agreement with the lQCD results can be achieved for both the intercepts at T=0T=0 and the slopes of m1m_1, although the description somewhat worsens at the highest temperature. We will return to the reason for that below.

Figure 2. The first cumulants of WLCs (lines) as a function of imaginary time at different temperatures and distances in comparison with the corresponding 2+1-flavor lQCD data [29].

For the interference functions, ϕ(r,T)\phi(r,T), we simply took the results from previous Ref. [36] (interpolated to the set of temperatures used here) which were based on fits to the lQCD data for HQ free energies, see Figure 3, but allowing for ±10%\pm10\% variations to illustrate underlying uncertainties. They show the qualitatively expected features of pushing the interference effects deeper into the bound state as temperature increases.

Figure 3. The interference function, ϕ(r)\phi(r), as a function of distance at different temperatures as interpolated from Ref. [36].

The main difference relative to the previous TT-matrix results lies in the screening parameters of the potential parameters as a function of temperature, shown in Figure 4. Most notably, it turns out the fits to the WLCs can be carried out with a constant Debye mass for the string interaction: at ms=0.2m_s=0.2 GeV is comparable to the low-TT values in previous work but does not increase with TT. Instead, it turns out that an increase in the cbc_b parameter is in order (which was assumed to be constant at cb=1.3c_b=1.3–1.55 in the previous works), implying that the screening of the string interaction is relegated to larger distances. On the other hand, the screening mass for the color-Coulomb interaction did not change much in either magnitude or temperature dependence.

Figure 4. The in-medium potential parameters as a function of temperature.

The pertinent in-medium potentials are displayed in Figure 5. Following the discussion of the screening parameters, the main difference to previous TT-matrix extractions largely based on HQ free energies lies at the higher temperatures and larger distances where less screening is present based on the WLCs fits. We also plot lQCD data points extracted from the small-rr limit of m1m_1. They essentially overlap with the extracted potential, except at the largest temperature where the interplay of the medium and large-distance points is not quite captured, and which is at the origin of the discrepancy in the WLC fits in lower right panel in Figure 2. This might be in part due to the fact that to observe string breaking on the lattice, it is not sufficient to consider WLCs or Wilson loops alone, but operators corresponding to static-light mesons should also be included in the analysis [56] (recall our remark following Eq. (12)). We also note that our fits utilize a mixing coefficient for the vector component of the confining potential of χ=0.8\chi=0.8, compared to 0.6 in Ref. [35] and 1 in Ref. [36]. While this choice still takes advantage of the improvement in the hyperfine splittings in the vacuum spectroscopy, it mitigates somewhat the aforementioned problem in the potential at the highest temperature which would be exacerbated for χ=0.6\chi=0.6.

Figure 5. The in-medium potentials (lines) as a function of distance at different temperatures in comparison to the 2+1-flavor lQCD results (dots) for the first cumulants of WLCs at τ=0\tau=0 [29].

The bridge between the underlying potential and the WLCs is provided by the static QQˉQ\bar{Q} spectral functions which are obtained from the selfconsistent solutions of the former and subsequently injected into Eq. (9) and used to compute the pertinent moment; they are depicted in Figure 6. The effective mass of QQˉQ\bar{Q} state, characterizing the pole position of ρQQˉ\rho_{Q\bar{Q}} at distance rr, is largely determined by the potential, V~(r)\widetilde{V}(r) (and identical to the latter in the limit m1(τ=0)m_1(\tau= 0)). However, they are not exactly the same since the non-zero real part of QQˉQ\bar{Q} self-energy, ΣQQˉ\Sigma_{Q\bar{Q}} in Eq. (10), slightly shifts the pole position away from V~(r)\widetilde{V}(r). Nevertheless, m1(τ=0)m_1(\tau= 0) remains an indicator of the QQˉQ\bar{Q} effective mass, which turns out to be rather temperature-independent as a consequence of the approximately temperature-independent potentials at small and intermediate distances (as shown in Figure 5). The increase of effective mass with increasing rr results from the smaller attraction (less binding) between QQ and Qˉ\bar{Q} at larger rr.

Figure 6. The spectral functions (with EE shifted by twice static-quark mass) of static quark antiquark pair as a function of energy at different temperatures and distances.

The width of the static spectral function indicates the interaction between QQˉQ\bar{Q} and the medium partons and is quantified by the imaginary part of the QQˉQ\bar{Q} self-energy, Φ(r,T)Im⁡ΣQQˉ(E,T)\Phi(r,T)\operatorname{Im}\Sigma_{Q\bar{Q}}(E,T), in Eq. (10); its dependence on temperature and distance are consistent with the slope of m1m_1 in Figure 2, exhibiting a strong broadening with increasing temperature and distance. The dependence on temperature is a consequence of several competing effects: a larger QGP density and less interference tend to increase width, while a smaller interaction strength with increasing temperatures decreases it. The distance dependence has two basic components, i.e., a long-range force enables a parton to interact with a larger number of thermal partons in the heat bath, proportional to the volume of the spherical shell which grows as r2r^2 [57], and the ceasing of interference effects with ϕ(r)\phi(r) approaching 1, enabling both QQ and Qˉ\bar{Q} to fully interact with the QGP.

Finally, we display in Figure 7 the spectral functions of light quarks, gluons and charm quarks as following from the selfconsistent TT-matrix solution based on the WLC fit. Not surprisingly, the most significant changes are at the two higher temperatures (T=0.293T=0.293, 352 GeV) where

Figure 7. Single-parton spectral functions for light quarks (first row), gluons (second row) and charm quarks (third row) as a function of energy for various 3-momenta in each panel. From left to right, the four columns correspond to temperatures of T=0.195T=0.195, 0.2510.251, 0.2930.293 and 0.352 GeV0.352\,\mathrm{GeV}, respectively.

the interaction strength is larger than before [35]. In the light-parton sector this leads to broader “quasiparticle” peaks and generates stronger collective modes on the low-energy shoulder of the “quasiparticle peaks” at zero parton momentum; note that the effective quark mass is around 0.5 GeV while the peaks of the collective mode are located near ω≃0.1\omega\simeq0.1 GeV. For the zero-momentum gluon spectral functions at the two higher temperatures, the collective modes emerge near ω=0.2\omega=0.2 GeV, which is, in fact, not far from gluon condensation. It is this feature that, especially at T=0.352T=0.352 GeV, prevents us from improving the fit of the WLC, as a stronger input potential at large distances results in an unstable fitting procedure due to the emergence of gluon condensation (signalled by a disappearance of the low-energy peak and a real part of the propagator flipping from negative to positive). Whether this “feature” can be turned into a framework where the bare parton masses are selfconsistently obtained from a condensed ground state is beyond the scope of this work (it would also have to be implemented at the lower temperatures). Even in the charm-quark sector the low-momentum spectral functions at the higher temperatures are significantly broader than before, again accompanied with a collective low-energy peak which implies notable deviations from a simple “quasiparticle” spectral shape.

Charm-Quark Transport Coefficients

We are now in position to compute the charm-quark transport properties in the QGP with the refined potential. Two main ingredients to the charm-quark transport coefficients are parton spectral functions, discussed in the previous section, and the heavy-light scattering amplitudes. The latter are displayed in Figure 8 for SS-wave cqˉc\bar{q} scattering in the color-singlet channel (which provides the largest contribution, and together with color-anti-triplet diquark amplitudes makes up the dominant contribution to the transport coefficient). In our recent work [35], we have found that the inclusion of a vector component in the string force (with mixing coefficient χ=0.6\chi=0.6) leads to significantly harder momentum dependence of the scattering amplitudes than the previous results for χ=0\chi=0 [36]. This trend persists here, although a little less pronounced due to the choice of χ=0.8\chi=0.8, leading to slightly reduced amplitudes for T=0.195 GeVT=0.195\,\mathrm{GeV} (compared to χ=0.6\chi=0.6) while they are quite comparable for T=0.251 GeVT=0.251\,\mathrm{GeV} at finite momentum. For the two higher temperatures, the effect of the stronger potential becomes important again.

Figure 8. The imaginary part of the SS-wave charm-light TT-matrices as a function of CM energy in the color-singlet channel at different temperatures for various CM momenta in each panel.

As the resonance peaks in the heavy-light amplitudes are largely below the “nominal” 2-parton threshold (e.g., Ecm≃2 GeVE_{\mathrm{cm}}\simeq2\,\mathrm{GeV} for the SS-wave “DD-meson” resonances in Figure 8 vs. Ethr=mq+mc≃2.3 GeVE_{\mathrm{thr}}=m_q+m_c\simeq2.3\,\mathrm{GeV} at T=0.195 GeVT=0.195\,\mathrm{GeV}), it is mandatory to account for the nontrivial spectral functions in the evaluation of the HQ transport coefficient. Following the previous study [57] in this context, the off-shell effects in calculating the HQ friction coefficient (relaxation rate) can be implemented by utilizing Kadanoff-Baym equations to yield

A(p)=∑i12ϵc(p)∫dω′d3p′(2π)32ϵc(p′)dνd3q(2π)32ϵi(q)dν′d3q′(2π)32ϵi(q′)δ(4)(2π)41dc∑a,l,s∣M∣2ρc(ω′,p′)ρi(ν,q)ρi(ν′,q′)[1−nc(ω′)]ni(ν)[1±ni(ν′)](1−p⋅p′p2).(15)A(p)=\sum_i \frac{1}{2\epsilon_c(p)}\int\frac{d\omega' d^3p'}{(2\pi)^3 2\epsilon_c(p')} \frac{d\nu d^3q}{(2\pi)^3 2\epsilon_i(q)} \frac{d\nu' d^3q'}{(2\pi)^3 2\epsilon_i(q')} \delta^{(4)}(2\pi)^4\frac{1}{d_c}\sum_{a,l,s}|M|^2\rho_c(\omega',p')\rho_i(\nu,q)\rho_i(\nu',q')\left[1-n_c(\omega')\right]n_i(\nu)\left[1\pm n_i(\nu')\right]\left(1-\frac{\mathbf{p}\cdot\mathbf{p}'}{p^2}\right). \tag*{(15)}

Here δ(4)\delta^{(4)} is a short-hand notation for the energy-momentum conserving δ\delta-function in the 2→22\to2 scattering process, dc=6d_c=6 the spin-color degeneracy of charm quarks, and the summation, ∑i\sum_i, is over all light-flavor quarks and gluons making up the thermal medium, i=u,uˉ,d,dˉ,s,sˉ,gi=u,\bar{u},d,\bar{d},s,\bar{s},g (the masses for light and strange quarks are assumed to be the same). Only the incoming charm quark is assumed to be a quasiparticle of definite momentum and corresponding on-shell energy, while the off-shell effects are implemented by energy convolutions over the spectral functions of the light parton and the outgoing charm quark. The heavy-light scattering matrix elements, ∣M∣2|M|^2 in Eq. (15), are directly related to the TT-matrix in the CM frame which incorporates the summation over all possible two-body color and partial-wave channels [57].

In Figure 9 we plot our results for the charm-quark friction coefficient, A(p;T)A(p;T). At the lower two temperatures the WLC-based result is a little lower than the one from [35], mostly due to the smaller value of χ\chi (0.8 vs. 0.6), but at temperatures T≳0.3T\gtrsim0.3 GeV, it becomes larger at low momentum (due to the stronger potential) and comparable at momenta p≳8p\gtrsim8 GeV (where the color-Coulomb interaction starts to dominate). The spatial diffusion coefficient, Ds=T/(McA(p=D_s=T/(M_cA(p= 0)), which is proportional to the relaxation time, τc=1/A(p=0)\tau_c = 1/A(p = 0), is commonly scaled by the inverse thermal wavelength, 2πT2\pi T, to render a dimensionless quantity. We display the diffusion coefficient in Figure 10 as a function of temperature. In the static limit our result is in fair agreement with recent lQCD data [50], and about a factor of 2-3 larger than the result from the AdS/CFT correspondence which is believed to provide a quantum lower bound for this quantity [58]. The temperature dependence is a bit weaker than our previous results, again a consequence of the stronger potential at the higher temperatures.

Figure 9. The charm-quark friction coefficient as a function of charm-quark 3-momentum at different temperatures.

Spatial diffusion coefficient for static, bottom, and charm quarks

Figure 10. Our result for the spatial diffusion coefficient for static (blue solid line), bottom (blue dashed line) and charm (blue dotted line) quarks as a function of temperature (scaled by Tc=0.155T_c=0.155 GeV) compared to the 2+1-flavor lQCD data [50] (red dots with error bars) and the AdS/CFT estimate [58] (black line). The bare quark masses for charm and bottom are 1.359 and 4.681 GeV, respectively, and we use 10 GeV to approximate the static limit.

The phenomenological consequences of our updated transport coefficients for HF observables in URHICs, which will also require the inclusion of radiative contributions [59], remain to be worked out.

Conclusions

We have applied the thermodynamic TT-matrix approach to compute static Wilson line correlators and utilized them to analyze pertinent lattice data. By varying the screening properties of the input potential and carrying out selfconsistent calculations of the 1- and 2-body correlation functions that encompass a description of lQCD results for the QGP equation of state, solutions were found that result in a fair agreement with the lQCD data for the first cumulants as a function of euclidean time. While the input potential at low temperatures is quite similar to previous solutions that were based on fits to heavy-quark free energies, more significant adjustments were required at temperatures above ∼300\sim300 MeV,

amounting to a significantly less screened (i.e., stronger) potential. This reinforces earlier findings that remnants of the confining force play a critical role in the properties of the strongly coupled QGP (well) above the critical temperature, with parton collision widths in excess of 0.5 GeV.

An immediate consequence of the stronger potential is an enhancement of low-energy collective modes in the light-quark and gluon spectral functions that develop well below the nominal values of their masses, implying strong deviations from the quasiparticle picture. Even for charm-quark spectral functions this distortion is still significant. We have also computed the pertinent HQ transport coefficients. Their temperature dependence turns out to be more gradual than before, and the predicted spatial diffusion coefficient in the static limit is still in approximate agreement with recent 2+1-flavor lQCD results. It will be interesting to see how these calculations fare when implemented into phenomenological applications to HF data in heavy-ion collisions. Work in this direction is in progress.

Acknowledgments

This work has been supported by the U.S. National Science Foundation under grant nos. PHY-1913286 and PHY-2209335, and by the The U.S. Department of Energy, Office of Science, Office of Nuclear Physics through contract No. DE-SC0012704 and the Topical Collaboration in Nuclear Theory on Heavy-Flavor Theory (HEFTY) for QCD Matter under award no. DE-SC0023547.

Relation between First Cumulant of WLC and Potential

In this Appendix we prove the identity that the first cumulant of the WLCs in the limit of vanishing euclidean time recovers the static potential, m1(r,τ=0)=V~(r)m_1(r,\tau= 0) = \widetilde{V}(r) (for simplicity we suppress the temperature dependence), as mentioned at the end of Sec. III. Expanding W(r,τ)W(r,\tau) in the vicinity of τ=0\tau= 0, one obtains

W(r,τ)=∫−∞∞dEe−EτρQQˉ(E,r)≈∫−∞∞dEρQQˉ(E,r)−τ∫−∞∞dEEρQQˉ(E,r)=1−τV~(r).(16)W(r,\tau) = \int_{-\infty}^{\infty} dE e^{-E\tau}\rho_{Q\bar{Q}}(E,r) \approx\int_{-\infty}^{\infty} dE \rho_{Q\bar{Q}}(E,r) - \tau\int_{-\infty}^{\infty} dE E\rho_{Q\bar{Q}}(E,r) = 1-\tau\widetilde{V}(r). \tag*{(16)}

The last identity in Eq. (16) has been proved in Ref. [36] using a contour integral and the fact that the two-body selfenergy, ΣQQˉ\Sigma_{Q\bar{Q}}, in the spectral function of Eq. (10), is analytic. In addition, the identity ∫−∞∞dEρQQˉ(E,r)=1\int_{-\infty}^{\infty} dE\rho_{Q\bar{Q}}(E,r)=1 is nothing but the normalization condition of spectral function. The first cumulant of the WLC, m1(r,τ)m_1(r,\tau), then becomes

m1(r,τ)=−∂τln⁡W(r,τ)=V~(r)(17)m_1(\mathbf{r},\tau)=-\partial_\tau\ln W(\mathbf{r},\tau)=\widetilde{V}(\mathbf{r}) \tag*{(17)}

and all the higher-order terms of τ\tau vanish at τ=0\tau=0.

References

References

  1. [1]F. Prino and R. Rapp, J. Phys. G 43, 093002 (2016), 1603.00529.
  2. [2]A. Beraudo et al., Nucl. Phys. A 979, 21 (2018), 1803.03824.
  3. [3]X. Dong and V. Greco, Prog. Part. Nucl. Phys. 104, 97 (2019).
  4. [4]R. Rapp, D. Blaschke, and P. Crochet, Prog. Part. Nucl. Phys. 65, 209 (2010), 0807.2470.
  5. [5]L. Kluberg and H. Satz, Color Deconfinement and Charmonium Production in Nuclear Collisions (2010), 0901.3831.
  6. [6]P. Braun-Munzinger and J. Stachel, Landolt-Bornstein 23, 424 (2010), 0901.2500.
  7. [7]X. Du, S. Y. F. Liu, and R. Rapp, Phys. Lett. B 796, 20 (2019), 1904.00113.
  8. [8]F. Karsch, S. Datta, E. Laermann, P. Petreczky, S. Stickan, and I. Wetzorke, Nucl. Phys. A 715, 701 (2003), hep-ph/0209028.
  9. [9]S. Datta, F. Karsch, P. Petreczky, and I. Wetzorke, Phys. Rev. D 69, 094507 (2004), hep-lat/0312037.
  10. [10]P. Petreczky and K. Petrov, Phys. Rev. D 70, 054503 (2004), hep-lat/0405009.
  11. [11]A. Mocsy and P. Petreczky, Eur. Phys. J. C 43, 77 (2005), hep-ph/0411262.
  12. [12]O. Kaczmarek and F. Zantow, Phys. Rev. D 71, 114510 (2005), hep-lat/0503017.
  13. [13]A. Jakovac, P. Petreczky, K. Petrov, and A. Velytsky, Phys. Rev. D 75, 014506 (2007), hep-lat/0611017.
  14. [14]G. Aarts, C. Allton, M. B. Oktay, M. Peardon, and J.-I. Skullerud, Phys. Rev. D 76, 045219 (2007), 0705.2198.
  15. [15]P. Petreczky, Eur. Phys. J. C 62, 85 (2009), 0810.0258.
  16. [16]G. Aarts, S. Kim, M. P. Lombardo, M. B. Oktay, S. M. Ryan, D. K. Sinclair, and J. I. Skullerud, Phys. Rev. Lett. 106, 061602 (2011), 1010.3725.
  17. [17]G. Aarts, C. Allton, S. Kim, M. P. Lombardo, M. B. Oktay, S. M. Ryan, D. K. Sinclair, and J. I. Skullerud, JHEP 11, 103 (2011), 1109.4496.
  18. [18]H. T. Ding, A. Francis, O. Kaczmarek, F. Karsch, H. Satz, and W. Soeldner, Phys. Rev. D 86, 014509 (2012), 1204.4945.
  19. [19]S. Kim, P. Petreczky, and A. Rothkopf, Phys. Rev. D 91, 054511 (2015), 1409.3630.
  20. [20]A. Bazavov, F. Karsch, Y. Maezawa, S. Mukherjee, and P. Petreczky, Phys. Rev. D 91, 054503 (2015), 1411.3018.
  21. [21]A. Bazavov, N. Brambilla, P. Petreczky, A. Vairo, and J. H. Weber (TUMQCD), Phys. Rev. D 98, 054511 (2018), 1804.10600.
  22. [22]R. Larsen, S. Meinel, S. Mukherjee, and P. Petreczky, Phys. Rev. D 100, 074506 (2019), 1908.08437.
  23. [23]R. Larsen, S. Meinel, S. Mukherjee, and P. Petreczky, Phys. Lett. B 800, 135119 (2020), 1910.07374.
  24. [24]R. Larsen, S. Meinel, S. Mukherjee, and P. Petreczky, Phys. Rev. D 102, 114508 (2020), 2008.00100.
  25. [25]P. Petreczky, S. Sharma, and J. H. Weber, Phys. Rev. D 104, 054511 (2021), 2107.11368.
  26. [26]R. Rapp and H. van Hees (2010), pp. 111–206, 0903.1096.
  27. [27]A. Mocsy, P. Petreczky, and M. Strickland, Int. J. Mod. Phys. A 28, 1340012 (2013), 1302.2180.
  28. [28]M. He, H. van Hees, and R. Rapp, Prog. Part. Nucl. Phys. 130, 104020 (2023), 2204.09299.
  29. [29]D. Bala, O. Kaczmarek, R. Larsen, S. Mukherjee, G. Parkar, P. Petreczky, A. Rothkopf, and J. H. Weber (HotQCD), Phys. Rev. D 105, 054513 (2022), 2110.11659.
  30. [30]A. Bazavov, D. Hoying, O. Kaczmarek, R. N. Larsen, S. Mukherjee, P. Petreczky, A. Rothkopf, and J. H. Weber (2023), 2308.16587.
  31. [31]D. Cabrera and R. Rapp, Phys. Rev. D 76, 114506 (2007), hep-ph/0611134.
  32. [32]F. Riek and R. Rapp, Phys. Rev. C 82, 035201 (2010), 1005.0769.
  33. [33]M. Mannarelli and R. Rapp, Phys. Rev. C 72, 064905 (2005), hep-ph/0505080.
  34. [34]F. Riek and R. Rapp, New J. Phys. 13, 045007 (2011), 1012.0019.
  35. [35]Z. Tang and R. Rapp (2023), 2304.02060.
  36. [36]S. Y. F. Liu and R. Rapp, Phys. Rev. C 97, 034918 (2018), 1711.03282.
  37. [37]R. Brockmann and R. Machleidt, Int. Rev. Nucl. Phys. 8, 121 (1999), nucl-th/9612004.
  38. [38]M. Cheng et al., Phys. Rev. D 77, 014511 (2008), 0710.0354.
  39. [39]P. Petreczky, J. Phys. G 37, 094009 (2010), 1001.5284.
  40. [40]A. Bazavov, Y. Burnier, and P. Petreczky, Nucl. Phys. A 932, 117 (2014), 1404.4267.
  41. [41]A. Bazavov et al. (HotQCD), Phys. Rev. D 90, 094503 (2014), 1407.6387.
  42. [42]V. D. Mur, V. S. Popov, Y. A. Simonov, and V. P. Yurov, J. Exp. Theor. Phys. 78, 1 (1994), hep-ph/9401203.
  43. [43]W. Lucha, F. F. Schoberl, and D. Gromes, Phys. Rept. 200, 127 (1991).
  44. [44]N. Brambilla and A. Vairo, Phys. Rev. D 55, 3974 (1997), hep-ph/9606344.
  45. [45]N. Brambilla and A. Vairo, Phys. Lett. B 407, 167 (1997), hep-ph/9703377.
  46. [46]A. P. Szczepaniak and E. S. Swanson, Phys. Rev. D 55, 3987 (1997), hep-ph/9611310.
  47. [47]D. Ebert, V. O. Galkin, and R. N. Faustov, Phys. Rev. D 57, 5663 (1998), (Erratum: Phys.Rev.D 59, 019902 (1999)), hep-ph/9712318.
  48. [48]D. Ebert, R. N. Faustov, and V. O. Galkin, Phys. Rev. D 67, 014027 (2003), hep-ph/0210381.
  49. [49]M. Tanabashi et al. (Particle Data Group), Phys. Rev. D 98, 030001 (2018).
  50. [50]L. Altenkort, O. Kaczmarek, R. Larsen, S. Mukherjee, P. Petreczky, H.-T. Shu, and S. Stendebach (HotQCD), Phys. Rev. Lett. 130, 231902 (2023), 2302.08501.
  51. [51]M. Laine, O. Philipsen, P. Romatschke, and M. Tassler, JHEP 03, 054 (2007), hep-ph/0611300.
  52. [52]J. M. Luttinger and J. C. Ward, Phys. Rev. 118, 1417 (1960).
  53. [53]G. Baym and L. P. Kadanoff, Phys. Rev. 124, 287 (1961).
  54. [54]G. Baym, Phys. Rev. 127, 1391 (1962).
  55. [55]S. Y. F. Liu and R. Rapp, Eur. Phys. J. A 56, 44 (2020), 1612.09138.
  56. [56]G. S. Bali, H. Neff, T. Duessel, T. Lippert, and K. Schilling (SESAM), Phys. Rev. D 71, 114513 (2005), hep-lat/0505012.
  57. [57]S. Y. F. Liu, M. He, and R. Rapp, Phys. Rev. C 99, 055201 (2019), 1806.05669.
  58. [58]J. Casalderrey-Solana and D. Teaney, Phys. Rev. D 74, 085012 (2006), hep-ph/0605199.
  59. [59]S. Y. F. Liu and R. Rapp, JHEP 08, 168 (2020), 2003.12536.

Paper details

vlnc:W_gannhcv8
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

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