betaFeedback

WorksPhysics and Astronomy

Thermal quarks and Polyakov loops in two-color dense QCD

Abstract

We study confinement and deconfinement in dense two color QCD by analyzing the dynamics of thermal quarks and gluons. The Polyakov loop is used as a probe of the relevant thermal excitations, distinguishing quark and hadron dominated regimes in dense matter. To describe the Polyakov loop, we adopt both lattice informed phenomenological models and the massive Yang Mills framework. After calibrating these models at zero density, we investigate in medium modifications of the Polyakov loops and gluon propagators at finite temperature and density. Diquark gaps control the screening at zero temperature, whereas the screening due to thermal quarks is sensitive to the Polyakov loop. Inclusion of the Polyakov loop helps to reproduce lattice data at low temperature, suggesting that thermal excitations are predominantly hadronic rather than uncorrelated quarks.

Article

Introduction

Recently, QCD-like theories such as two-color QCD (QC2D\mathrm{QC}_{2}\mathrm{D}) and QCD at finite isospin chemical potential (QCDI\mathrm{QCD}_{I}) have attracted much attention as laboratories to test methodologies and concepts concerning dense matter in quantum chromodynamics (QCD) [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]. For these theories, lattice Monte Carlo simulations can be carried out without encountering the sign problem [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25], and the results have been used to assess the reliability of effective models as well as weak-coupling approaches, with or without pairing effects [26] [27, 28]. In recent years, much effort has been devoted to studies of the equation of state (EOS) of dense matter. Notably, several concepts relevant to neutron star EOS—such as the appearance of a sound-speed peak, the stiffening of matter due to pairing effects [29–32], and the significance of quark effects prior to hadron overlap [33–35]—have been confirmed or supported by lattice results in QCD-like theories.

In QC2D\mathrm{QC}_{2}\mathrm{D} and QCDI\mathrm{QCD}_{I} at zero temperature, simple models that incorporate both diquark (meson) and quark degrees of freedom—such as quark–diquark (meson) models [9] [36, 37]—provide good agreement with lattice results [34,38–43]. Although these models do not explicitly realize confinement, they are able to capture the overall features of hadronic matter at low density and quark matter at high density. A notable advantage is that, within a single framework, both causality and thermodynamic stability are automatically satisfied from low to high densities, especially in the domain where hadronic matter transforms into quark matter. By construction, such models implement the essential features of the previous crossover approaches in which hadronic and quark EOS were interpolated phenomenologically [29,30,44–53]. For more explicit realization of the crossover, see, e.g., recent developments based on the quarkyonic matter picture [35,54–58].

As mentioned, simple quark–hadron models have not directly addressed the nature of confinement–deconfinement transitions; rather this most subtle and intriguing question has been detoured. In this paper, taking dense two-color QCD as our platform, we investigate confinement/deconfinement in dense matter by studying the thermal properties of dense matter. We delineate the lattice data for the Polyakov loop and gluon propagators and compare them with analytic results. Our central question is whether thermal excitations are dominated by quarks or hadrons; answering to this question helps us to understand confinement/deconfinement in dense matter.

In QC2_2D, quarks acquire the color-singlet gap and thus generate neither Debye nor Meissner masses [59–63]. As a consequence, infrared gluons are protected from screening and remain confining. With increasing density, the protected phase space gradually shrinks, but it disappears only asymptotically—a scenario known as asymptotic confinement [64,65]. Quark matter with confining gluons provides a concrete realization of the quarkyonic matter proposed by McLerran and Pisarski [66]. Although qualitative arguments for asymptotic confinement are plausible, a detailed characterization remains lacking. Given the significant progress in lattice simulations since the original proposal of asymptotic confinement, it is timely to revisit this concept in a more quantitative manner. In this paper, we pursue this goal through a set of analyses based on effective models.

In thermal models of QCD, the Polyakov loop plays a central role [67]; see Ref. [68] for a review. The expectation value of the Polyakov loop is related to the free energy of an infinitely heavy quark, ⟨L⟩∼e−FQ\langle L\rangle\sim e^{-F_Q}. In pure Yang–Mills (pYM) theory, the flux string from an isolated color charge extends to infinity, implying FQ→∞F_Q \to\infty and hence ⟨L⟩→0\langle L\rangle\to0. In contrast, ⟨L⟩\langle L\rangle becomes finite when the medium is populated with color fields, allowing the flux string of a heavy quark to dissolve. In QCD with light quarks, the string can also be terminated by dynamical quark–antiquark creation, so ⟨L⟩\langle L\rangle may already be finite even before the medium is saturated with color fields. A key question is how rapidly ⟨L⟩\langle L\rangle grows with temperature, which directly reflects whether the dominant thermal excitations are hadrons or quarks.

In this paper, we examine two models for the Polyakov loop. The first is the so-called inverse Weiss potential, derived from nonperturbative gluon and ghost propagators [69–73]. In pYM theory, this model successfully reproduces the critical temperature of the confinement–deconfinement transition. Here, we investigate how its predictions are modified once quarks are coupled. The second model is a phenomenological construction calibrated to lattice simulation data, and it has been widely used in Polyakov–Nambu–Jona–Lasinio (PNJL) and Polyakov–Quark–Meson (PQM) frameworks. Although both models describe pYM well, the inclusion of quarks clearly differentiates them: the inverse Weiss potential becomes excessively modified and deviates from lattice results, whereas the phenomenological model continues to reproduce the data. This distinction becomes even more pronounced in the regime of dense quark matter.

We further investigate gluon propagators at various values of μq\mu_q and TT, incorporating Polyakov loop effects that suppress thermal quarks. Lattice studies of gluon propagators in two-color dense QCD have been reported in Ref. [74], while model analyses with massive gluon propagators have been carried out both at zero temperature [60] and at finite temperature with quark one-loop corrections [61]. These analytic computations represent finite-μq\mu_q applications of the framework successfully developed for pYM [75–78] and later extended to QCD with dynamical quarks [79,80]; see Ref. [81] for a review. In the present work, we extend these earlier analyses for QC2_2D by explicitly including the Polyakov-loop suppression of thermal quarks, and we argue that this effect is helpful to explain the lattice data at low temperature.

The remainder of this paper is organized as follows. In § 2, we discuss Polyakov loop potentials in pYM theory within both phenomenological and massive-gluon models. In § 3, we introduce quarks into these Polyakov loop models. In § 4, the model is calibrated at μq=0\mu_q=0 against lattice results. In § 5, we analyze the behavior of Polyakov loops at finite density. In § 6, we compute the in-medium screening of gluon propagators in a Polyakov loop background.

In § 7, we examine the density and temperature dependence of the screening masses. In § 8, we study gluon propagators at finite momenta. Finally, § 9 is devoted to discussions and a summary.

Polyakov Loop Potential in Pure Yang–Mills

We begin with the gluon sector in the presence of background fields A0=A0aλa2A_0 = A_0^a \frac{\lambda_a}{2}, where λa\lambda_a are the Pauli matrices for color. The background field A0A_0 is assumed to be constant in space-time, and we consider amplitudes along the λ3\lambda_3 direction. The mean-field Polyakov loop is then given as follows:

Φ=12tr⁡[PeigβA03λ3/2]=cos⁡βR2,R≡gA03.(1)\Phi= \frac{1}{2}\operatorname{tr}\left[\mathcal{P}e^{ig\beta A_0^3\lambda_3/2}\right] = \cos\frac{\beta\mathcal{R}}{2}, \qquad\mathcal{R} \equiv gA_0^3. \tag*{(1)}

In the high temperature limit (β→0\beta\to0), Φ\Phi approaches unity, corresponding to the deconfined phase. At low temperatures, A0A_0 is expected to scale as ∼T\sim T. If βR→π\beta\mathcal{R} \to\pi is in this limit, the system is in the confined phase (Φ=0\Phi= 0). Thus, as the temperature increases, the system undergoes a transition from the confined to the deconfined phase.

Below, we examine two typical models for the Polyakov loop potential: a phenomenological model commonly employed in PNJL and PQM frameworks, and the inverse Weiss potential, which can be computed from nonperturbative gluon and ghost propagators.

Phenomenological Potential

The phenomenological potentials used in PNJL or PQM models [68] are constructed to reproduce lattice results both with and without quarks. For example, the two-color potential used by Brauner, Fukushima, and Hidaka (BFH) [8] is

ΩBFH=−bT[24Φ2e−βa+ln⁡(1−Φ2)],(2)\Omega_{\mathrm{BFH}}=-bT\left[24\Phi^2e^{-\beta a}+\ln(1-\Phi^2)\right], \tag*{(2)}

where b=(0.2105 GeV)3b=(0.2105\,\mathrm{GeV})^3 and a=0.8581 GeVa=0.8581\,\mathrm{GeV}. The potential vanishes for the confined solution Φ→0\Phi\to0. The first term, proportional to −Te−βa-Te^{-\beta a}, can be interpreted as an entropic term favoring the deconfined solution. The second term, ln⁡(1−Φ2)\ln(1-\Phi^2), favors the confined solution; it diverges as Φ→1\Phi\to1, preventing the system from reaching complete deconfinement (Φ=1\Phi= 1). The logarithmic form originates from strong-coupling expansions, where the Haar measure yields a logarithmic contribution, and it is related to the presence of ghosts. The three-color version of this model is known to successfully reproduce lattice data at Nc=3N_c=3, and this situation is also true at Nc=2N_c=2, as we see in the following.

Inverse Weiss Potential

To construct the inverse Weiss potential, we compute the effective potential in the presence of the background field A03A_0^3. For the potential derived from the original perturbative propagators, see Refs. [82,83]. The Lagrangian for the pYM part is

LpYM=−14GμνAGAμν,GμνA=∂μAνA−∂νAμA+gfABCAμBAνC,(3)\mathcal{L}_{\mathrm{pYM}}=-\frac{1}{4}G_{\mu\nu}^A G^{A\mu\nu}, \qquad G_{\mu\nu}^A=\partial_\mu A_\nu^A-\partial_\nu A_\mu^A+gf_{ABC}A_\mu^B A_\nu^C, \tag*{(3)}

where fABCf_{ABC} is the structure constant and gg is the coupling constant. To construct the one-particle-irreducible (1PI) effective potential as a function of A0=A03λ32A_0=A_0^3\frac{\lambda_3}{2}, we first separate the gauge field as Aμ=A0δμ0+aμA_\mu=A_0\delta_{\mu0}+a_\mu, with aμa_\mu being quantum fields. The GμνAG_{\mu\nu}^A can be rewritten as GμνA=[Dμ,aν]A−[Dν,aμ]A+gfABCaμBaνCG_{\mu\nu}^A=[\mathcal{D}_\mu,a_\nu]^A-[\mathcal{D}_\nu,a_\mu]^A+gf_{ABC}a_\mu^B a_\nu^C, with the background covariant derivative defined by Dμaν=∂μaν−iRδμ0[λ32,λC2]aνC\mathcal{D}_\mu a_\nu=\partial_\mu a_\nu-i\mathcal{R}\delta_{\mu0}\left[\frac{\lambda_3}{2},\frac{\lambda_C}{2}\right]a_\nu^C. Using the Cartan bases λ3/2\lambda_3/2 and λ±/2=(λ1±iλ2)/2\lambda_\pm/2=(\lambda_1\pm i\lambda_2)/2, one can write

Dμaν=∂μaν3λ32+(∂μ−iRδμ0)aν−λ+2+(∂μ+iRδμ0)aν+λ−2.(4)\mathcal{D}_\mu a_\nu=\partial_\mu a_\nu^3\frac{\lambda_3}{2}+(\partial_\mu-i\mathcal{R}\delta_{\mu0})a_\nu^-\frac{\lambda_+}{2}+(\partial_\mu+i\mathcal{R}\delta_{\mu0})a_\nu^+\frac{\lambda_-}{2}. \tag*{(4)}

We impose the background gauge fixing condition, Dμaμ=0D^\mu a_\mu=0, which breaks the gauge invariance of aμa_\mu but maintains the invariance for AμA_\mu.

To model the non-perturbative propagators in the IR, we add a mass term for gluons to capture the qualitative trend of the Landau gauge propagators in a vacuum; see Ref. [81] for a recent review and Refs. [84,85], which give the physical motivation based on the Gribov–Zwanziger framework. We can either add a term ∝mg2aμ2\propto m_g^2a_\mu^2, which is local but not gauge invariant, or a term which is non-local but gauge invariant,

Lmg=−mg22aμ(gμν−∂μ∂ν∂2)aν.(5)\mathcal{L}_{m_g}=-\frac{m_g^2}{2}a_\mu\left(g^{\mu\nu}-\frac{\partial^\mu\partial^\nu}{\partial^2}\right)a_\nu. \tag*{(5)}

Including all these elements and using the Faddeev–Popov trick, the massive YM model is

LmYM=LpYM+(Dμaμ)22ξ+Lmg+cˉDμDμc+Lint,(6)\mathcal{L}_{\mathrm{mYM}}=\mathcal{L}_{\mathrm{pYM}}+\frac{(D^\mu a_\mu)^2}{2\xi}+\mathcal{L}_{m_g}+\bar{c}D^\mu D_\mu c+\mathcal{L}_{\mathrm{int}} , \tag*{(6)}

where cc and cˉ\bar{c} are ghost and anti-ghost fields. Taking ξ→0\xi\to0, the gluon and ghost propagators are (the gluon propagator DμνAB∝⟨(aμA)∗(−k)aνB(k)⟩D_{\mu\nu}^{AB}\propto\langle(a_\mu^A)^*(-k)a_\nu^B(k)\rangle)

[DABμν(k)]tree=−δABkA2−mg2Pkμν,[DcAB(k)]tree=−δABkA2,(7)\left[D_{AB}^{\mu\nu}(k)\right]_{\mathrm{tree}}=-\frac{\delta_{AB}}{k_A^2-m_g^2}P_k^{\mu\nu},\qquad\left[D_c^{AB}(k)\right]_{\mathrm{tree}}=-\frac{\delta^{AB}}{k_A^2}, \tag*{(7)}

where δAB=1\delta_{AB}=1 for (A,B)=(3,3),(+,+),(−,−)(A,B)=(3,3),(+,+),(-,-), with the other components being zero. The projection operator for the transverse component is Pkμν=gμν−kμkν/k2P_k^{\mu\nu}=g^{\mu\nu}-k^\mu k^\nu/k^2. The momentum for (3,−,+)(3,-,+)-components is given by

(k3)μ=kμ,(k±)μ=kμ±Rδμ0.(8)(k_3)_\mu=k_\mu,\qquad(k_\pm)_\mu=k_\mu\pm\mathcal{R}\delta_{\mu0}. \tag*{(8)}

With these propagators, one can construct the one-loop effective potential. We define

F(r;m)=T∑n∫k⃗ln⁡[(ωn+r)2+k⃗ 2+m2].(9)F(r;m)=T\sum_n\int_{\vec{k}}\ln\left[(\omega_n+r)^2+\vec{k}^{\,2}+m^2\right]. \tag*{(9)}

where ωn=2nπT\omega_n=2n\pi T (n∈Zn\in\mathbb{Z}) is the Matsubara frequency for bosons. The free energy densities for gluons and ghosts can be written as

ΩgmYM(T;R)=12∑r=0,±R[3F(r;mg)+F(r;0)],ΩghmYM(T;R)=−∑r=0,±RF(r;0),(10)\Omega_g^{\mathrm{mYM}}(T;\mathcal{R})=\frac{1}{2}\sum_{r=0,\pm\mathcal{R}}\left[3F(r;m_g)+F(r;0)\right],\qquad\Omega_{gh}^{\mathrm{mYM}}(T;\mathcal{R})=-\sum_{r=0,\pm\mathcal{R}}F(r;0), \tag*{(10)}

where TT-independent constants are neglected. The factor 3 in the first term accounts for the trace over the projectors PμνP^{\mu\nu}, while the last term arises from the longitudinal component, which decouples from radiative corrections and hence remains massless. Ghosts contribute with the opposite sign to gluons.

Summing the gluon and ghost contributions, the one-loop potential reads

Ωg+ghmYM(T;R)=12∑r=0,±R[3F(r;mg)−F(r;0)],(11)\Omega_{g+gh}^{\mathrm{mYM}}(T;\mathcal{R})=\frac{1}{2}\sum_{r=0,\pm\mathcal{R}}\left[3F(r;m_g)-F(r;0)\right], \tag*{(11)}

where the sum over rr can be explicitly evaluated as

∑r=0,±RF(r;m)=T∫k⃗(ln⁡[1−e−βEm]+ln⁡[1−2e−βEm(2Φ2−1)+e−2βEm]).(12)\sum_{r=0,\pm\mathcal{R}}F(r;m)=T\int_{\vec{k}}\left(\ln\left[1-e^{-\beta E_m}\right]+\ln\left[1-2e^{-\beta E_m}(2\Phi^2-1)+e^{-2\beta E_m}\right]\right). \tag*{(12)}

In standard perturbation theory, ghosts cancel two gluon modes, leaving only the two transverse modes. In the present massive theory, the massless longitudinal mode is canceled, but the cancellation of one mode is incomplete due to the gluon mass. At low temperatures, massive gluons are suppressed, and the ghost contribution, which favors the confining solution, dominates the potential.

The gluon and ghost potentials are shown in Figure 1 for low temperature (T=20T = 20 MeV, confined) and high temperature (T=260T = 260 MeV, deconfined). It should be emphasized that the Polyakov loop in the fundamental representation does not suppress thermal excitations in the adjoint representation. Physically, dynamical gluons screen adjoint charges and form color-singlet states. Indeed, the logarithmic terms with Boltzmann factors can be rewritten as Tradj.⁡[1−Ladje−βE]T\operatorname{radj.}[1-L_{\mathrm{adj}}e^{-\beta E}]. Here, LadjL_{\mathrm{adj}} denotes the Polyakov loop operator in the adjoint representation, which can be expressed in terms of the fundamental representation LfundL_{\mathrm{fund}} through a group-theoretic relation, (Nc2−1)Ladj=Nc2LfundLfund−1(N_c^2-1)L_{\mathrm{adj}}=N_c^2L_{\mathrm{fund}}L_{\mathrm{fund}}-1. At the mean-field level, a vanishing Polyakov loop for quarks (Φ=0\Phi= 0) necessarily implies a non-vanishing Polyakov loop for gluons and ghosts. As a result, the thermal pressure (=−ΩmYM=-\Omega^{\mathrm{mYM}}) is unphysically large at low temperature.

Image not extracted

Figure 1. The Polyakov loop potential before (left) and just after (right) the deconfinement transition. The results at T=20T = 20 and 260 MeV are shown for illustrative purposes. The potential is normalized by T4T^4. The normalized potential for ghosts does not depend on TT. The absolute value of the gluonic part grows with TT, shifting the minimum from Φ=0\Phi= 0 to 1.

In addition, we examine the temperature and Φ\Phi-dependence of the mass mgm_g. Such dependences have been studied in Ref. [86]; see also Ref. [87]. We adopt the following parametrization:

mgT=(mgvac)2+cgΦngT2,(13)m_g^T = \sqrt{\left(m_g^{\mathrm{vac}}\right)^2+c_g\Phi^{n_g}T^2}, \tag*{(13)}

which interpolates between the computations at Φ=0\Phi= 0 and 1 within the Gribov–Zwanziger framework. We choose ng=2n_g = 2 and cg=0c_g = 0 and 4. At Φ=1\Phi= 1, the mass at large TT eventually scales as mg∼gsTm_g\sim g_sT, while for Φ=0\Phi= 0, the mass is saturated, mg∼const⁡m_g\sim\operatorname{const} for Φ=0\Phi= 0. For the vacuum mass, we choose mgvac=800m_g^{\mathrm{vac}} = 800 MeV, which is consistent with the lattice studies for the Landau gauge propagators in pYM [72,73].

Comparison of the Polyakov Loop Potential

Here, we compare the phenomenological BFH potential with the mYM potential. The left panel of Figure 2 shows both potentials as functions of Φ\Phi. In both cases, the critical temperature lies in the range 240–280 MeV. A major difference is that the mYM potential is significantly flatter than the BFH one. Consequently, the Polyakov loop grows much more rapidly with temperature in the mYM case, as illustrated in the right panel of Figure 2. This rapid growth can be moderated by introducing a TT- and Φ\Phi-dependent gluon mass, which suppresses the gluonic entropic contribution. With cg≃2.0c_g \simeq2.0–4.0 and ng=2n_g = 2, the thermal evolution of the Polyakov loop in the mYM model can be brought into qualitative agreement with the BFH result, although the potential structure at Φ≳0.5\Phi\gtrsim0.5 still shows substantial differences. As we will see in the next section, the behavior of the potential for Φ≳0.5\Phi\gtrsim0.5 plays a crucial role once quarks are included.

Image not extracted

Figure 2. (Left) Comparisons of the Polyakov loop potential for the mYM for mgvac=800m_g^{\mathrm{vac}} = 800 MeV with and without the Φ\Phi-dependence of mgm_g, and the pYM part used in the PNJL model (BFH). For Φ\Phi-dependent mgTm_g^T, we chose cg=4c_g = 4 and ng=2n_g = 2 in Equation (13). All these models lead to the transition temperature being somewhere between T=240T = 240 and 280 MeV. The potential used in the PNJL model has an energy cost around Φ∼1\Phi\sim1, much bigger than in the mYM. (Right) The temperature dependence of the Polyakov loop Φ\Phi for mYM models with cg=0c_g = 0, 2, and 4, and for the BFH model.

Including Quarks: PNJL Model

Quasi-Quark Excitations

We now include quarks within a quasi-particle description. To be concrete, we use the PNJL model. As a first step, we examine the μq=0\mu_q = 0 case to calibrate our models against lattice data. In particular, the lattice data were often obtained for a nonzero diquark source with which quarks form color-singlet diquark condensates and acquire a gap even at μq=0\mu_q = 0. Without the diquark source term, the diquark condensate is formed only for μB≥mπ\mu_B \ge m_\pi (equivalently μq≥mπ/2\mu_q \ge m_\pi/2), as predicted by the chiral perturbation theory [88,89].

For later convenience, from the very beginning, we derive the general expression of quark spectra at nonzero μq\mu_q and Δ\Delta. In the presence of a temporal background field, the kinetic term reads

\mathcal{L}_{\mathrm{kin},q}=\bar{q}\left[i\slashed{\partial}+g\slashed{A}+\mathcal{R}T_3\gamma_0+i\mu\gamma_0-M_q\right]q-q^T\Delta q, \tag*{(14)}

with T3=λ3/2T_3=\lambda_3/2 and Δ=γ5τ2λ2Δ\Delta=\gamma_5\tau_2\lambda_2\Delta, where τ2\tau_2 is the Pauli matrix in flavor space. Employing the Nambu–Gor’kov formalism, we define ψ=q\psi=q and its charge-conjugated field as ψC(p)=λ2τ2CqˉT(−p)\psi_C(p)=\lambda_2\tau_2C\bar{q}^T(-p). Then,

\bar{\psi}(p)\slashed{p}\psi(p)=\bar{\psi}_C(p)\slashed{p}\psi_C(p),\qquad\bar{\psi}(p)\gamma_0A_0T_3\psi(p)=\bar{\psi}_C(p)\gamma_0A_0T_3\psi_C(p), \tag*{(15)}

where we used λ2(−T3T)λ2=T3\lambda_2(-T_3^T)\lambda_2=T_3. The momentum space representation of the Lagrangian now reads

\mathcal{L}_{\mathrm{kin},q}=\bar{\Psi}(p) \begin{bmatrix} \slashed{p}+(\mathcal{R}T_3+i\mu)\gamma_0-M_q & \Delta\gamma_5\\ -\Delta^*\gamma_5 & \slashed{p}+(\mathcal{R}T_3-i\mu)\gamma_0-M_q \end{bmatrix} \Psi(p), \tag*{(16)}

where Ψ(p)=(ψ(p),ψC(p))T/2\Psi(p)=(\psi(p),\psi_C(p))^T/\sqrt{2}.

We note that p0p_0 and RT3\mathcal{R}T_3 always appear in the combination p0+RT3p_0+\mathcal{R}T_3, while the chemical potential μ\mu enters the quasiparticle poles as ξp,a(p⃗)=(Eq(p⃗)∓μ)2+Δ2\xi_{p,a}(\vec{p})=\sqrt{(E_q(\vec{p})\mp\mu)^2+\Delta^2}. In the thermodynamic potential, the shift p0→p0+RT3p_0\to p_0+\mathcal{R}T_3 modifies the standard entropic logarithm according to

tr⁡Cln⁡[1+e−β(ξki+iRT3)]=ln⁡[1+e−β(ξki+iR/2)]+ln⁡[1+e−β(ξki−iR/2)],(17)\operatorname{tr}_C\ln\left[1+e^{-\beta(\xi_k^i+i\mathcal{R}T_3)}\right]=\ln\left[1+e^{-\beta(\xi_k^i+i\mathcal{R}/2)}\right]+\ln\left[1+e^{-\beta(\xi_k^i-i\mathcal{R}/2)}\right], \tag*{(17)}

where the trace is taken over color indices. Combining these logarithms, the thermal quark contribution to the free energy (or equivalently, minus the pressure) becomes as follows:

Ωkin,qT=−Pkin,qT=−4T∑i=p,a∫kln⁡(1+2Φe−βξki+e−2βξki).(18)\Omega^{T}_{\mathrm{kin},q}=-\mathcal{P}^{T}_{\mathrm{kin},q}=-4T\sum_{i=p,a}\int_k \ln\left(1+2\Phi e^{-\beta\xi_k^i}+e^{-2\beta\xi_k^i}\right). \tag*{(18)}

The pressure at Φ=0\Phi=0 (confined) or Φ=1\Phi=1 (deconfined) are

Pkin,qT→{4T∑i=p,a∫kln⁡(1+e−βNcξki)(Φ→0)4NcT∑i=p,a∫kln⁡(1+e−βξki)(Φ→1)\mathcal{P}^{T}_{\mathrm{kin},q}\to \begin{cases} 4T\sum_{i=p,a}\int_k\ln\left(1+e^{-\beta N_{\mathrm{c}}\xi_k^i}\right) & (\Phi\to0)\\ 4N_{\mathrm{c}}T\sum_{i=p,a}\int_k\ln\left(1+e^{-\beta\xi_k^i}\right) & (\Phi\to1) \end{cases}

where we write Nc=2N_{\mathrm{c}}=2. The thermal quark pressure is larger at least by a factor NcN_{\mathrm{c}} in the deconfined phase than in the confined phase. The difference would be even greater in the high temperature limit; for Φ=0\Phi=0 with ξk∼∣k⃗∣\xi_k\sim|\vec{k}|, we rescale the variable as Nc∣k⃗∣=∣k⃗′∣N_{\mathrm{c}}|\vec{k}|=|\vec{k}'| to get

Pkin,qT→∞(Φ=0)∼Nc−4Pkin,qT→∞(Φ=1).(19)\mathcal{P}^{T\to\infty}_{\mathrm{kin},q}(\Phi=0)\sim N_{\mathrm{c}}^{-4}\mathcal{P}^{T\to\infty}_{\mathrm{kin},q}(\Phi=1). \tag*{(19)}

In order to examine the impact of the Polyakov loop, in Figure 3, we fix Mq=400M_q=400 MeV (later, we consider the dynamical evolution of MqM_q) and show the thermal quark pressure at μq=0\mu_q=0 with Φ∗\Phi_* minimizing the potentials for mYM + quarks and BFH. We also show the Φ=0\Phi=0 and 1 cases for comparison. It is clear that the suppression effects due to the Polyakov loop are substantial; Pkin,qT\mathcal{P}^{T}_{\mathrm{kin},q} at Φ=0\Phi=0 is much smaller than the Φ=1\Phi=1 case, as anticipated from Equation (19) for the quark sector. For dynamical computations allowing the variance of Φ\Phi, the growth of Φ\Phi is driven by the thermal quark (hadronic) excitations. Thermal hadronic contributions rise rapidly around T∼150T\sim150–200200 MeV as in the Hagedorn gas, as also does the Polyakov loop. In this regime, hadrons strongly overlap, and there are color fields to which quarks couple become available to form a quark–gluon–plasma; here, hadrons dissociate into a quark–gluon plasma, and quark descriptions should be valid.

Image not extracted

Figure 3. The Polyakov loop dependence of the quark pressure in Equation (18) at μq=0\mu_q=0. Here, we fix Mq=400M_q=400 (500) MeV for solid (dotted) curves. The Polyakov loop values Φ∗\Phi_* are dynamically determined for the mYM and BFH with quarks. The results for fixed Φ=0\Phi=0 and Φ=1\Phi=1 (common for the mYM and BFH models) are also shown.

Evolution of Condensates

To determine the evolution of the in-medium gaps, we include both the zero-point energy and the potential terms associated with the gaps and the Polyakov loop. For this purpose, we adopt the pNJL-type thermodynamic potential

Ωq=−4∑i=p,a∫k⃗[ξki+Tln⁡(1+2Φe−βξki+e−2βξki)]+(Mq−mq)2+(Δ−Δext)22G,(20)\Omega_q=-4\sum_{i=p,a}\int_{\vec{k}}\left[\xi_k^i+T\ln\left(1+2\Phi e^{-\beta\xi_k^i}+e^{-2\beta\xi_k^i}\right)\right]+\frac{(M_q-m_q)^2+(\Delta-\Delta_{\mathrm{ext}})^2}{2G}, \tag*{(20)}

where GG denotes the NJL coupling constant, mqm_q the current quark mass, and Δext\Delta_{\mathrm{ext}} an external diquark source term. The current quark mass mqm_q is adjusted to reproduce the pion mass employed in the lattice simulations. In the lattice setup, the diquark source term Δext\Delta_{\mathrm{ext}} is introduced to stabilize the numerical calculations; we take its value accordingly to ensure consistency with the lattice data.

Including the pYM term, the gap equations for Φ\Phi, MqM_q, and Δ\Delta are (ΩQC2D=ΩpYM+Ωq)(\Omega_{\mathrm{QC2D}}=\Omega_{\mathrm{pYM}}+\Omega_q)

∂ΩQC2D∂Φ=∂ΩpYM∂Φ−4T∑i=p,a∫k⃗[Θ(ΛNJL−∣k⃗∣)cosh⁡(βξki)+Φ+Θ(∣k⃗∣−ΛNJL)cosh⁡(βξki∣η=ηext)+Φ],\frac{\partial\Omega_{\mathrm{QC2D}}}{\partial\Phi}=\frac{\partial\Omega_{\mathrm{pYM}}}{\partial\Phi}-4T\sum_{i=p,a}\int_{\vec{k}}\left[\frac{\Theta(\Lambda_{\mathrm{NJL}}-\lvert\vec{k}\rvert)}{\cosh(\beta\xi_k^i)+\Phi}+\frac{\Theta(\lvert\vec{k}\rvert-\Lambda_{\mathrm{NJL}})}{\cosh(\beta\xi_k^i\vert_{\eta=\eta_{\mathrm{ext}}})+\Phi}\right],
∂ΩQC2D∂η=−4∑i=p,a∫k⃗Θ(ΛNJL−∣k⃗∣)∂ξki∂ηsinh⁡(βξki)cosh⁡(βξki)+Φ+η−ηext2G,(21)\frac{\partial\Omega_{\mathrm{QC2D}}}{\partial\eta}=-4\sum_{i=p,a}\int_{\vec{k}}\Theta(\Lambda_{\mathrm{NJL}}-\lvert\vec{k}\rvert)\frac{\partial\xi_k^i}{\partial\eta}\frac{\sinh(\beta\xi_k^i)}{\cosh(\beta\xi_k^i)+\Phi}+\frac{\eta-\eta_{\mathrm{ext}}}{2G}, \tag*{(21)}

where η=Mq\eta=M_q or Δ\Delta. In the low temperature limit (β→+∞\beta\to+\infty), the thermal factors in the integrand depend on Φ\Phi very weakly; the thermal factor behaves as 1/(cosh⁡(βξki)+Φ)≃e−βξki1/(\cosh(\beta\xi_k^i)+\Phi)\simeq e^{-\beta\xi_k^i} for the Φ\Phi part and sinh⁡(βξki)/(cosh⁡(βξki)+Φ)≃1\sinh(\beta\xi_k^i)/(\cosh(\beta\xi_k^i)+\Phi)\simeq1 for the MqM_q and Δ\Delta parts.

Here, we have assumed that the gaps damp at high energy and parametrize such damping scale as ΛNJL\Lambda_{\mathrm{NJL}}. This scale appears automatically if we use momentum-dependent interactions, e.g., the ∼1/p2\sim1/p^2 type. In the NJL model, the contact interactions yield only momentum-independent gaps. Thus, we have to impose such a scale ΛNJL\Lambda_{\mathrm{NJL}} by hand so that ∂ξki/∂η\partial\xi_k^i/\partial\eta is nonzero only for ∣k⃗∣≤ΛNJL\lvert\vec{k}\rvert\leq\Lambda_{\mathrm{NJL}}. This results in the standard NJL gap equations. Meanwhile, the integral in the equation for Φ\Phi should include both thermal quarks with and without the effective masses, so we should keep the integral for ∣k⃗∣≥ΛNJL\lvert\vec{k}\rvert\geq\Lambda_{\mathrm{NJL}}.

The Pion Mass

To calibrate our model parameters to the lattice setup, we first compute the pion mass within the model. In the framework of the RPA, the pion mass at μq=T=0\mu_q=T=0 is determined by

1−4GNcδab∫p⃗Θ(ΛNJL−∣k⃗∣)Ep−mπ2/4Ep=0.(22)1-4GN_c\delta_{ab}\int_{\vec{p}}\frac{\Theta(\Lambda_{\mathrm{NJL}}-\lvert\vec{k}\rvert)}{E_p-m_\pi^2/4E_p}=0. \tag*{(22)}

In the chiral limit, this equation reduces to the gap equation for MqM_q at mπ=0m_\pi=0, in accordance with the Nambu–Goldstone theorem.

The Polyakov Loop in QC2D: Calibration at q=0q=0

We first briefly summarize the qualitative differences between the mYM and phenomenological BFH potentials after including quarks, then we proceed to quantitative comparisons with lattice results.

We begin by examining the structure of the Polyakov loop potentials at μq=0\mu_q=0. Figure 4 shows the Polyakov loop potential, ΩQC2DT=ΩpYM+ΩqT\Omega_{\mathrm{QC2D}}^T=\Omega_{\mathrm{pYM}}+\Omega_q^T, with quarks included (left), and the corresponding temperature dependence of the Polyakov loop (right). Here, ΩpYM\Omega_{\mathrm{pYM}} is replaced by either Ωgl+gh\Omega_{\mathrm{gl+gh}} for the mYM or ΩBFH\Omega_{\mathrm{BFH}} for the phenomenological model. For the mYM potential, we consider cg=0.0c_g=0.0 and 4.04.0 with ng=2n_g=2 and mgvac=800 MeVm_g^{\mathrm{vac}}=800\ \mathrm{MeV}.

Image not extracted

Figure 4. (Left) The Polyakov loop potential at μq=0\mu_q = 0 in the same setup as in Figure 2 except that thermal quark contributions in Equation (18) are added and the temperatures are now T=100T = 100 and 150150 MeV. The effective quark mass is fixed to Mq=400M_q = 400 MeV. (Right) The temperature dependence of the Polyakov loop Φ\Phi. The solid (dotted) lines are results for Mq=400M_q = 400 (500) MeV. The lattice results (renormalization scheme A in Ref. [15]) are also plotted for comparison.

Adding quarks to the pYM potential significantly alters its qualitative structure, leading to distinct behaviors between the mYM and BFH models. While both models exhibit similar transition temperatures in the pYM case, the mYM potential is much flatter with respect to Φ\Phi, so the deconfined solution is strongly favored once quarks are included. In contrast, the BFH potential incurs a large energy cost near Φ≃1\Phi\simeq1 (diverging as Φ→1\Phi\to1) due to the Haar measure term ∼ln⁡(1−Φ2)\sim\ln(1-\Phi^2), which delays the growth of Φ\Phi. These differences are directly reflected in the temperature dependence of the Polyakov loop.

We now turn to more quantitative discussions. To compare our model calculations with lattice simulation results, we must use model parameters adjusted to the setup used in lattice computations. First, we note that there is no unique way to set the scale in QC2D [14] since there are no real-world counterparts. In this work, we consider the lattice results in which the string tension is fixed as σs≃440\sqrt{\sigma_s} \simeq440 MeV. For the current quark masses, the “fine ensemble” lattice simulations [15] use a relatively heavy quark mass that yields a pion-to-rho mass ratio of mπ/mρ∼0.8m_\pi/m_\rho\sim0.8, which is substantially larger than the corresponding ratio in the real world, mπ/mρ≃0.18m_\pi/m_\rho\simeq0.18. The heavy pion mass of mπ≃637m_\pi\simeq637 MeV was used to reduce the computational cost for extensions to finite μq\mu_q. We take into account this heavy current quark mass in our model predictions. We test two parameter sets, set I and set II, in Table 1.

Namemπm_\pi [MeV]ΛNJL\Lambda_{\mathrm{NJL}} [MeV]GG [GeV−2^{-2}]mqm_q [MeV]MqM_q [MeV]
set I6376575.42121378
set II7.23110512

Table 1. Two parameter sets are used for the quark model. We tune the current quark mass mqm_q for a given coupling GG so that the RPA yields a pion pole at mπ=637m_\pi= 637 MeV. The cutoff ΛNJL\Lambda_{\mathrm{NJL}} is chosen to be the same for both parameter sets.

Another important step is to set the renormalization factor of the Polyakov loop. We are not aware of how to rigorously match the Polyakov loop in our model and lattice data. For this reason, we simply choose the renormalization factor for the lattice results to make model results and lattice data consistent at μq=0\mu_q = 0 (Figure 5), then we make model predictions for the Polyakov loop at μq≠0\mu_q \ne0. Later, these model predictions are compared with the lattice results.

Image not extracted

Figure 5. The temperature dependence of the Polyakov loop and the effective quark mass MqM_q for the set I (solid) and II (dotted) at μq=0\mu_q = 0 with Δext=0\Delta_{\rm ext} = 0. For the lattice data with mπ≃637m_\pi\simeq637 MeV, we use two renormalization schemes with ΔER=ER−EB=220\Delta E_R = E_R - E_B = 220 and 270 MeV to convert the lattice data in scheme B; see Equation (27).

The renormalized and bare Polyakov loops are related by (in Ref. [15], the notation “LL” is used in place of Φ\Phi)

ΦR(T)=ZΦNτΦ0((Nτa)−1)(23)\Phi_R(T) = Z_\Phi^{N_\tau}\Phi_0\left((N_\tau a)^{-1}\right) \tag*{(23)}

where T=(Nτa)−1T = (N_\tau a)^{-1} and ZΦZ_\Phi is the renormalization factor. Ref. [15] discussed two renormalization schemes,

ΦRscheme A(T=1/4a)=1.0,ΦRscheme B(T=1/4a)=0.5,(24)\Phi_R^{\mathrm{scheme\ A}}(T = 1/4a) = 1.0,\qquad\Phi_R^{\mathrm{scheme\ B}}(T = 1/4a) = 0.5, \tag*{(24)}

where 1/4a≃356.8751/4a \simeq356.875 MeV. In Ref. [15], most results are presented in scheme B. In the continuum expression, the renormalization factors correspond to the choice of the zero-point energy,

ZΦNτ=exp⁡[βln⁡ZΦa]≡eβ(ER−E0).(25)Z_\Phi^{N_\tau} = \exp\left[\beta\frac{\ln Z_\Phi}{a}\right] \equiv e^{\beta(E_R-E_0)}. \tag*{(25)}

The Polyakov loops in schemes A and B can be written as follows:

ΦRscheme A/B(T;ER)=eβ(ERA/B−E0)Φ0(T).(26)\Phi_R^{\mathrm{scheme\ A/B}}(T;E_R) = e^{\beta(E_R^{A/B}-E_0)}\Phi_0(T). \tag*{(26)}

In Ref. [15], most of the data are given in scheme B. Hence, it is convenient to write the Polyakov loops at the energy scale ERE_R as follows:

ΦR(T;ER)=eβ(ER−EB)ΦRscheme B(T).(27)\Phi_R(T;E_R) = e^{\beta(E_R-E_B)}\Phi_R^{\mathrm{scheme\ B}}(T). \tag*{(27)}

For instance, to convert the data from scheme B to A, we substitute ERAE_R^A such that ERA−ERB≃247E_R^A - E_R^B \simeq247 MeV. For fitting to our model at μq=0\mu_q = 0, we use ER≃270E_R \simeq270 MeV unless otherwise stated.

It is important to note that these renormalization factors affect not only the magnitude of the Polyakov loop but also its temperature derivatives. For instance, the location of the susceptibility peak depends on the renormalization scheme. This scheme dependence can be eliminated by considering the ratio ΦR(T;μq)/ΦR(T;μq=0)\Phi_R(T;\mu_q)/\Phi_R(T;\mu_q=0), allowing a physically meaningful discussion of μq\mu_q effects. At low temperatures, this ratio is inconvenient since both numerators and denominators become vanishing. For this reason, we do not show the ratio but plot Φ\Phi itself after calibrating ER−EBE_R-E_B so that our Φ\Phi at μq=0\mu_q=0 matches the lattice results. We then examine the μq\mu_q-dependence of the Polyakov loop.

The Polyakov Loop in Baryonic and Quark Matter

Outline

We now explore the finite μq\mu_q domain, beginning with the overall behavior of various condensates. It is well established that for Δext=0\Delta_{\mathrm{ext}}=0, a diquark condensate forms when μq≥mπ/2\mu_q \ge m_\pi/2. Initially, the diquarks are of the Bose–Einstein condensation (BEC) type, forming molecule-like pairs, but with increasing density, they evolve into Bardeen–Cooper–Schrieffer (BCS)-type pairs near the quark Fermi surface. For finite Δext≠0\Delta_{\mathrm{ext}}\ne0, diquark condensates can form even before μq=mπ/2\mu_q=m_\pi/2. Figure 6 shows MqM_q, Δ\Delta, and Φ\Phi at a fixed temperature T=10T=10 MeV for various μq\mu_q using parameter set I. The solid (dotted) lines correspond to Δext=45\Delta_{\mathrm{ext}}=45 MeV, which matches ja=0.03ja=0.03 in the lattice results of Ref. [15]. As mentioned, nonzero Δext\Delta_{\mathrm{ext}} allows the diquark condensate to form even for μq<mπ/2\mu_q<m_\pi/2. Meanwhile, the reduction of Δ\Delta at very large μq\mu_q is simply an artifact of the UV cutoff; beyond the Fermi surface, with increasing μq\mu_q, the available phase space for pairing diminishes because of the UV cutoff; hence, diquarks cannot form at very large densities. The gap Δ\Delta vanishes around μq≃800\mu_q\simeq800 MeV, leaving quarks gapless on the Fermi surface. In this regime, even a small temperature of T∼10T\sim10 MeV is sufficient to induce substantial Polyakov loop expectation values.

Image not extracted

Figure 6. The μq\mu_q-dependence of the Polyakov loop, the effective quark mass MqM_q, and the diquark gap Δ\Delta for the set I at T=10T=10 MeV with Δext=45\Delta_{\mathrm{ext}}=45 MeV (solid) and 0 MeV (dotted). The reduction in Δ\Delta at large μq\mu_q is the UV cutoff artifact, so we compare our model results with the lattice data at μq\mu_q less than 714 MeV (shown as the gray line). For Δext=45\Delta_{\mathrm{ext}}=45 MeV, the Δ\Delta is nonzero even at μq=0\mu_q=0, and transitions are crossover everywhere, as the chiral and U(1)BU(1)_B symmetries are explicitly broken. The nonzero Polyakov loop can be found only for μq≳800\mu_q\gtrsim800 MeV in the Δext=0\Delta_{\mathrm{ext}}=0 case where quarks become gapless because of the UV cutoff artifact.

Figure 7 shows the diquark gaps Δ\Delta for μq=286\mu_q=286, 428, 571, and 714 MeV as a function of temperature. For Δext=0\Delta_{\mathrm{ext}}=0 MeV, parameter set I predicts a second-order transition at Ts∼110T_s\sim110–140 MeV. In Ref. [15], the results at finite jj are extrapolated to j→0j\to0, giving an estimate Ts≃100T_s\simeq100–120 MeV, which is roughly consistent with set I. In contrast, set II, used with the BFH potential, predicts Ts≃170T_s\simeq170–200 MeV, indicating a different definition of the theory from that in Ref. [15]. This explains why the NJL coupling GG in set I is approximately 0.75 times smaller than in the BFH setup. For comparisons, in figure 8 we also show the results for fixed values of Polyakov loop at Φ=0\Phi=0 and 1.

Image not extracted

Figure 7. The temperature dependence of the diquark gap Δ\Delta for set I with Δext=45\Delta_{\mathrm{ext}} = 45 MeV (solid) and 0 MeV (dotted).

Image not extracted

(a)

Image not extracted

(b)

Figure 8. The same plot as Figure 7, but Φ\Phi is fixed to 0 (left panel) or 1 (right panel).

Having established reasonable consistency in TsT_s and Δ\Delta between model set I and the lattice results of Ref. [15], we now extend the comparison to finite μq\mu_q. Figure 9 shows the Polyakov loop Φ\Phi for set I with Δext=45\Delta_{\mathrm{ext}} = 45 MeV, together with the lattice data at ja=0.03j_a = 0.03. At low temperatures, corresponding to Φ≲0.8\Phi\lesssim0.8, the model reproduces the lattice results well. At higher temperatures (Φ≳0.8\Phi\gtrsim0.8), the model Φ\Phi begins to saturate, while the lattice Φ\Phi continues to grow, as the renormalization factor allows Φ\Phi to exceed unity. This high-temperature deviation occurs for all μq\mu_q studied, and its origin remains unclear. Nevertheless, our model is expected to be more reliable at low energies, and the agreement in this regime is noteworthy.

Image not extracted

Figure 9. The temperature dependence of the Polyakov loop for set I at various μq\mu_q with Δext=45\Delta_{\mathrm{ext}}=45 MeV. The lattice data is from the ja=0.03j_a=0.03 in Ref. [15]. For the density interval from μq≃300\mu_q \simeq300 MeV to ≃700\simeq700 MeV, the size of Δ\Delta is ∼250\sim250–300300 MeV; see Figure 6.

At low temperatures, it is crucial to maintain a sizable Δ\Delta. Figure 10 illustrates that if Δ\Delta is artificially set to zero, quarks become gapless. These gapless excitations induce a rapid rise of the Polyakov loop, resulting in substantial disagreement between our model and the lattice data.

Image not extracted

Figure 10. Same as Figure 9, except: (Left panel) For the external diquark source, Δex=0\Delta_{\mathrm{ex}}=0 is chosen. Here, the phase transition from the diquark condensed phase to the normal phase is the second-order phase transition; (Right panel) Δ=0\Delta=0 is fixed by hand for all TT. The gapless quarks trigger a rapid increase in the Polyakov loop already at temperatures less than ∼50\sim50 MeV.

We emphasize that, in all these analyses, the parameters for the gluonic part of the Polyakov loop potential, ΩpYM\Omega_{\mathrm{pYM}}, are kept unchanged. In principle, in-medium gluon screening could modify the gluon propagator and weaken confinement. Nevertheless, our minimal setup—without major modifications to ΩpYM\Omega_{\mathrm{pYM}}—yields reasonable agreement with lattice results, suggesting that drastic changes in ΩpYM\Omega_{\mathrm{pYM}} are disfavored for the density range considered here. In the next section, we reinforce this conclusion by examining the effect of the diquark gap on the gluon propagators.

Gluon Propagators at One-Loop with Polyakov Loops

We focus on medium corrections to the static gluon propagators, D(k)D(k), at zero frequency (k0=0k_0=0). Processes with k0≪μqk_0 \ll\mu_q keep quarks near the Fermi surface and hence are most relevant for nonperturbative effects. At the same time, the momentum transfer ∣k⃗∣\lvert\vec{k}\rvert can be large if quarks remain close to the Fermi surface before and after scattering. Therefore, we concentrate on small k0k_0 while considering a wide range of ∣k⃗∣|\vec{k}| to ∼2μq\sim2\mu_q.

To reproduce the lattice results of Ref. [74], it is crucial to suppress quark excitations in the infrared region. Gapless quarks would otherwise over-screen the electric gluons [62]. Including a diquark gap of Δ∼200\Delta\sim200 MeV reduces the electric screening, improving agreement with lattice data [60,61]. In contrast, in the magnetic sector, static screening remains absent for any value of Δ\Delta, as in the normal phase. The magnetic screening is usually regarded as non-perturbative effects.

At finite temperatures, thermally excited particles—either quarks or hadrons—respond to external momenta as gapless modes [61,90]. If the thermal excitations are quark-dominated, they induce static electric screening as in the normal phase, while the static magnetic modes remain unscreened. In contrast, if the thermal excitations are confined, the electric screening is expected to be weaker, since color-singlet objects of typical size RsizeR_{\mathrm{size}} cannot respond effectively to external fields at wavelengths much longer than their size (k≪Rsize−1k \ll R_{\mathrm{size}}^{-1}).

We account for the suppression of thermal quarks by including Polyakov loops. The gluon self-energy is described by the polarization function

−iΠμνAB(kA)=(igs)2∫x,ye−ik(x−y)⟨(qˉγμλA2q)(x)(qˉγνλB2q)(y)⟩.(28)-\mathrm{i}\Pi_{\mu\nu}^{AB}(k_A) = (\mathrm{i}g_s)^2 \int_{x,y} e^{-\mathrm{i}k(x-y)} \left\langle\left(\bar{q}\gamma_\mu\frac{\lambda_A}{2}q\right)(x)\left(\bar{q}\gamma_\nu\frac{\lambda_B}{2}q\right)(y)\right\rangle. \tag*{(28)}

where kAk_A was defined in Equation (8). Explicit evaluation shows that ΠAB\Pi^{AB} depends on kAk_A and is diagonal in color space, with nonzero components only for (A,B)=(3,3),(+,+),(−,−)(A,B)=(3,3),(+,+),(-,-). All other combinations vanish.

As usual, we decompose the gluon self-energy into electric and magnetic components. In the presence of the background field, the momenta entering the projection tensors are kAk_A, which ensures the Ward–Takahashi identity, kAμΠμνAB=0k_A^\mu\Pi_{\mu\nu}^{AB}=0, is satisfied. The decomposition then reads

ΠμνAB(kA)=δABPμνE(kA)ΠE(kA)+δABPμνM(kA)ΠM(kA)(29)\Pi_{\mu\nu}^{AB}(k_A)=\delta^{AB}P_{\mu\nu}^{E}(k_A)\Pi_E(k_A)+\delta^{AB}P_{\mu\nu}^{M}(k_A)\Pi_M(k_A) \tag*{(29)}

where kAk_A was defined in Equation (8).

The Vacuum Subtraction

The polarization function is formally UV divergent. We extract the UV divergent vacuum contribution and combine it with the counter terms for the renormalization,

[ΠμνAB(kA)]R=[ΠμνAB(kA)+δZgkA2PμνkA]=δABPμνkA[Πvac(kA)+δZgkA2]+PμνE(kA)δΠEAB(kA)+PμνM(kA)δΠMAB(kA),(30)\begin{aligned} \left[\Pi_{\mu\nu}^{AB}(k_A)\right]^R=\left[\Pi_{\mu\nu}^{AB}(k_A)+\delta Z_g k_A^2 P_{\mu\nu}^{k_A}\right] \\ &=\delta^{AB}P_{\mu\nu}^{k_A}\left[\Pi_{\mathrm{vac}}(k_A)+\delta Z_g k_A^2\right]+P_{\mu\nu}^{E}(k_A)\delta\Pi_E^{AB}(k_A)+P_{\mu\nu}^{M}(k_A)\delta\Pi_M^{AB}(k_A), \tag*{(30)} \end{aligned}

where the renormalized vacuum part is (K2≡−(kA)2K^2\equiv-(k_A)^2, αs=gs2/4π\alpha_s=g_s^2/4\pi)

ΠvacR(kA)=Πvac(kA)+δZgkA2=−K22αsπ∫01dx x(1−x)ln⁡((Mqvac)2+x(1−x)K2(Mqvac)2+x(1−x)μR2)(31)\Pi_{\mathrm{vac}}^R(k_A)=\Pi_{\mathrm{vac}}(k_A)+\delta Z_g k_A^2=-K^2\frac{2\alpha_s}{\pi}\int_0^1 dx\,x(1-x)\ln\left(\frac{(M_q^{\mathrm{vac}})^2+x(1-x)K^2}{(M_q^{\mathrm{vac}})^2+x(1-x)\mu_R^2}\right) \tag*{(31)}

with μR\mu_R being the renormalization scale. The medium part is defined by subtracting the vacuum part,

δΠE,M≡ΠE,M−Πvac,(32)\delta\Pi_{E,M}\equiv\Pi_{E,M}-\Pi_{\mathrm{vac}}, \tag*{(32)}

where the subscript ABAB is omitted for the moment.

This subtracted function is UV convergent, but its finite part depends on regularization methods. In particular, if we integrate q0q_0 first to pick up poles and then integrate q⃗\vec{q}, there emerge UV finite terms which violate the conservation law kAμΠμνAB(kA)=0k_{A}^{\mu}\Pi_{\mu\nu}^{AB}(k_{A})=0 when the quark propagators in ΠE,M\Pi_{E,M} and Πvac\Pi_{\mathrm{vac}} contain, e.g., different MqM_{q} or Δ\Delta [59].

One way to restore the conservation law is to introduce counterterms that cancel regularization artifacts [59]. Another approach is to split the vacuum subtraction into two steps [61]:

δΠE,M=δSΠE,M+δμqΠE,M=[ΠvacMq→M~q−Πvac]dim reg+[ΠE,M−ΠvacMq→M~q]3d reg.(33)\delta\Pi_{E,M}=\delta_{S}\Pi_{E,M}+\delta_{\mu q}\Pi_{E,M}=\left[\Pi_{\mathrm{vac}}^{M_{q}\to\widetilde{M}_{q}}-\Pi_{\mathrm{vac}}\right]_{\mathrm{dim\ reg}}+\left[\Pi_{E,M}-\Pi_{\mathrm{vac}}^{M_{q}\to\widetilde{M}_{q}}\right]_{\mathrm{3d\ reg}}. \tag*{(33)}

Here, δμqΠE,M\delta_{\mu q}\Pi_{E,M}—which is the difference between ΠE,M\Pi_{E,M} and Πvac\Pi_{\mathrm{vac}} evaluated with the same gap parameters but different μq\mu_{q}—has no regularization artifacts; hence, one can evaluate it using the standard residue method with a three-dimensional cutoff. Meanwhile, each term in δSΠE,M\delta_{S}\Pi_{E,M} can be evaluated via dimensional regularization; hence, δSΠE,M\delta_{S}\Pi_{E,M} can be made artifact-free. The parameter M~q(μq,T)≡Mq2+Δ2\widetilde{M}_{q}(\mu_{q},T)\equiv\sqrt{M_{q}^{2}+\Delta^{2}}, determined at a given μq\mu_{q}, is uniquely fixed to ensure that the conservation law is maintained. Explicitly, δSΠE,M\delta_{S}\Pi_{E,M} reads as follows:

δSΠE,M=−K22αsqπ∫01dx x(1−x)ln⁡M~q2+x(1−x)K2(Mqvac)2+x(1−x)K2.(34)\delta_{S}\Pi_{E,M}=-K^{2}\frac{2\alpha_{s}^{q}}{\pi}\int_{0}^{1}dx\,x(1-x)\ln\frac{\widetilde{M}_{q}^{2}+x(1-x)K^{2}}{(M_{q}^{\mathrm{vac}})^{2}+x(1-x)K^{2}}. \tag*{(34)}

If the gap parameters in vacuum and medium remain the same, i.e., M~q=Mqvac\widetilde{M}_{q}=M_{q}^{\mathrm{vac}}, then δSΠE,M\delta_{S}\Pi_{E,M} vanishes by construction.

The Medium Corrections

Physical medium effects are largely reflected by δμqΠE,M\delta_{\mu q}\Pi_{E,M}. Decomposing the normal and anomalous quark propagators into the particle and antiparticle components, then calculating the residue integral for q0q_{0}, we find (before the vacuum subtraction)

−iΠE,MAB(kA)=gs22∑s,s′=p,a∫dq02πKE,Mss′[CE,Mss′GABss′+C~E,Mss′G~ABss′](q⃗+,q⃗−;k0),(35)-\mathrm{i}\Pi_{E,M}^{AB}(k_{A})=\frac{g_{s}^{2}}{2}\sum_{s,s'=p,a}\int\frac{dq_{0}}{2\pi}\mathcal{K}_{E,M}^{ss'}\left[\mathcal{C}_{E,M}^{ss'}\mathcal{G}_{AB}^{ss'}+\widetilde{\mathcal{C}}_{E,M}^{ss'}\widetilde{\mathcal{G}}_{AB}^{ss'}\right](\vec{q}_{+},\vec{q}_{-};k_{0}), \tag*{(35)}

where q⃗±=q⃗±k⃗/2\vec{q}_{\pm}=\vec{q}\pm\vec{k}/2. We call KE,Mss′\mathcal{K}_{E,M}^{ss'}, CE,Mss′\mathcal{C}_{E,M}^{ss'}, and GABss′,G~ABss′\mathcal{G}_{AB}^{ss'},\widetilde{\mathcal{G}}_{AB}^{ss'} the kinematic, coherence, and propagator factors. The propagator G~\widetilde{\mathcal{G}} is non-vanishing only at nonzero temperatures.

Kinematic Factors

The explicit forms of the kinematic factors are

KEpp=KEaa=−k⃗2k⃗2(1+q⃗ 2−k⃗2/4+Mq2Eq+Eq−),\mathcal{K}_{E}^{pp}=\mathcal{K}_{E}^{aa}=-\frac{\vec{k}^{2}}{\vec{k}^{2}}\left(1+\frac{\vec{q}^{\,2}-\vec{k}^{2}/4+M_{q}^{2}}{E_{q+}E_{q-}}\right),
KEpa=KEap=−k⃗2k⃗2(1−q⃗ 2−k⃗2/4+Mq2Eq+Eq−),\mathcal{K}_{E}^{pa}=\mathcal{K}_{E}^{ap}=-\frac{\vec{k}^{2}}{\vec{k}^{2}}\left(1-\frac{\vec{q}^{\,2}-\vec{k}^{2}/4+M_{q}^{2}}{E_{q+}E_{q-}}\right),
KMpp=KMaa=−1+(∣q⃗∣cos⁡θ)2−k⃗2/4+Mq2Eq+Eq−,\mathcal{K}_{M}^{pp}=\mathcal{K}_{M}^{aa}=-1+\frac{(|\vec{q}|\cos\theta)^{2}-\vec{k}^{2}/4+M_{q}^{2}}{E_{q+}E_{q-}},
KMpa=KMap=−1−(∣q⃗∣cos⁡θ)2−k⃗2/4+Mq2Eq+Eq−,(36)\mathcal{K}_{M}^{pa}=\mathcal{K}_{M}^{ap}=-1-\frac{(|\vec{q}|\cos\theta)^{2}-\vec{k}^{2}/4+M_{q}^{2}}{E_{q+}E_{q-}}, \tag*{(36)}

where θ\theta in cos⁡θ\cos\theta is the angle between q⃗\vec{q} and k⃗\vec{k}. These kinematic factors depend on MqM_{q} and the spatial momenta, and they are common across different states of matter. It is useful to note that, at low momenta, the electric components behave as KEpp→2+O(k⃗2)\mathcal{K}_{E}^{pp}\to2+O(\vec{k}^{2}) and KEpa→O(k⃗2)\mathcal{K}_{E}^{pa}\to O(\vec{k}^{2}). This indicates that in-medium corrections to the electric part are dominated by particle–hole excitations.

Coherence Factors

The coherence factors, which characterize the diquark condensed phase, are

CE,Mpp=12(1−(Eq+−μq)(Eq−−μq)±∣Δ∣2ξ~p+ξ~p−),CE,Maa=12(1−(Eq++μq)(Eq−+μq)±∣Δ∣2ξ~a+ξ~a−),CE,Mpa=12(1+(Eq+−μq)(Eq−+μq)∓∣Δ∣2ξ~p+ξ~a−),(37)\begin{aligned} C^{\mathrm{pp}}_{E,M} &= \frac{1}{2}\left(1-\frac{(E_{q+}-\mu_q)(E_{q-}-\mu_q)\pm\lvert\Delta\rvert^2}{\tilde{\xi}_p^+\tilde{\xi}_p^-}\right),\\ C^{\mathrm{aa}}_{E,M} &= \frac{1}{2}\left(1-\frac{(E_{q+}+\mu_q)(E_{q-}+\mu_q)\pm\lvert\Delta\rvert^2}{\tilde{\xi}_a^+\tilde{\xi}_a^-}\right),\\ C^{\mathrm{pa}}_{E,M} &= \frac{1}{2}\left(1+\frac{(E_{q+}-\mu_q)(E_{q-}+\mu_q)\mp\lvert\Delta\rvert^2}{\tilde{\xi}_p^+\tilde{\xi}_a^-}\right), \tag*{(37)} \end{aligned}

and the factor relevant only at finite temperature (since G~\tilde{\mathcal{G}} is nonzero only at finite temperature, see below),

C~E,Mpp=12(1+(Eq+−μq)(Eq−−μq)±∣Δ∣2ξ~p+ξ~p−),C~E,Maa=12(1+(Eq++μq)(Eq−+μq)±∣Δ∣2ξ~a+ξ~a−),C~E,Mpa=12(1−(Eq+−μq)(Eq−+μq)∓∣Δ∣2ξ~p+ξ~a−),(38)\begin{aligned} \widetilde{C}^{\mathrm{pp}}_{E,M} &= \frac{1}{2}\left(1+\frac{(E_{q+}-\mu_q)(E_{q-}-\mu_q)\pm\lvert\Delta\rvert^2}{\tilde{\xi}_p^+\tilde{\xi}_p^-}\right),\\ \widetilde{C}^{\mathrm{aa}}_{E,M} &= \frac{1}{2}\left(1+\frac{(E_{q+}+\mu_q)(E_{q-}+\mu_q)\pm\lvert\Delta\rvert^2}{\tilde{\xi}_a^+\tilde{\xi}_a^-}\right),\\ \widetilde{C}^{\mathrm{pa}}_{E,M} &= \frac{1}{2}\left(1-\frac{(E_{q+}-\mu_q)(E_{q-}+\mu_q)\mp\lvert\Delta\rvert^2}{\tilde{\xi}_p^+\tilde{\xi}_a^-}\right), \tag*{(38)} \end{aligned}

where CE,Mpa(q+,q−)=CE,Map(q−,q+)C^{\mathrm{pa}}_{E,M}(q_+,q_-)=C^{\mathrm{ap}}_{E,M}(q_-,q_+) and C~E,Mpa(q+,q−)=C~E,Map(q−,q+)\widetilde{C}^{\mathrm{pa}}_{E,M}(q_+,q_-)=\widetilde{C}^{\mathrm{ap}}_{E,M}(q_-,q_+). It is important to note that, at zero temperature, the electric component is vanishing in the small momentum limit, CEpp∼O(q⃗⋅k⃗)C^{\mathrm{pp}}_E\sim O(\vec{q}\cdot\vec{k}). Later, we see that the propagator factor GEpp\mathcal{G}^{\mathrm{pp}}_E remains finite for k⃗→0\vec{k}\to0 so that the product CEppGEppC^{\mathrm{pp}}_E\mathcal{G}^{\mathrm{pp}}_E vanishes, meaning that the electric screening is absent in the diquark condensed phase. This situation differs from the normal phase where the propagator exhibits a singularity k⃗\vec{k} as ∼1/q⃗⋅k⃗\sim1/\vec{q}\cdot\vec{k} for small ∣k⃗∣\lvert\vec{k}\rvert and hence makes the product CEppGEppC^{\mathrm{pp}}_E\mathcal{G}^{\mathrm{pp}}_E finite. In this case, the Debye mass emerges, cutting off gluons in the IR. In contrast, the particle–hole contribution to the magnetic component is finite, CMpp∼Δ2/ξ~p2C^{\mathrm{pp}}_M\sim\Delta^2/\tilde{\xi}_p^2.

Propagator Factors

Finally, we examine the propagator factors where the effects of the Polyakov loop enter. First, we generalize the Fermi–Dirac distribution for quarks as

nF(x±iQ/2)=1+Φeβxe2βx+2Φeβx+1±iΦ~eβxe2βx+2Φeβx+1≡LFΦ(x)±iL~FΦ(x),(39)n_F(x\pm iQ/2)=\frac{1+\Phi e^{\beta x}}{e^{2\beta x}+2\Phi e^{\beta x}+1}\pm i\frac{\widetilde{\Phi}e^{\beta x}}{e^{2\beta x}+2\Phi e^{\beta x}+1}\equiv L_F^\Phi(x)\pm i\widetilde{L}_F^\Phi(x), \tag*{(39)}

where Φ=Φ+iΦ~=cos⁡(βR/2)+isin⁡(βR/2)\Phi=\Phi+i\widetilde{\Phi}=\cos(\beta R/2)+i\sin(\beta R/2). For deconfined quarks at (Φ,Φ~)=(1,0)(\Phi,\widetilde{\Phi})=(1,0), this function is reduced to the usual Fermi–Dirac distribution, LFΦ=1(x)→nF(x)L_F^{\Phi=1}(x)\to n_F(x) and L~FΦ=1(x)→0\widetilde{L}_F^{\Phi=1}(x)\to0. The low temperature behavior of the real part is LFΦ=1(ξs)∼e−βξsL_F^{\Phi=1}(\xi_s)\sim e^{-\beta\xi_s}.

For confined quarks at (Φ,Φ~)=(0,1)(\Phi,\widetilde{\Phi})=(0,1), we find LFΦ=i(x)→nF(2x)L_F^{\Phi=i}(x)\to n_F(2x) and L~FΦ=i(x)→eβxnF(2x)\widetilde{L}_F^{\Phi=i}(x)\to e^{\beta x}n_F(2x). At low temperatures, their behaviors are LFΦ=i(ξs)∼e−2βξsL_F^{\Phi=i}(\xi_s)\sim e^{-2\beta\xi_s} and L~FΦ=i(ξs)∼e−βξs\widetilde{L}_F^{\Phi=i}(\xi_s)\sim e^{-\beta\xi_s}, respectively. For low energy excitations, the real part exhibits a stronger suppression, ∼e−2βΔ\sim e^{-2\beta\Delta}, which can be interpreted as the Boltzmann factor for hadronic excitations. Meanwhile, the imaginary part is suppressed as ∼e−βΔ\sim e^{-\beta\Delta}, reflecting quark excitations decaying into the colored background.

Using these generalized distribution functions, the A=B=3A=B=3 components of the propagator factors are

Gss′A=B=3(q+,q−;Φ)=[1−LFΦ(ζs−)−LFΦ(ζs′+)](1k0+ζs−+ζs′++1−k0+ζs−+ζs′+),\mathcal{G}^{A=B=3}_{ss'}(q_+,q_-;\Phi)=\left[1-L_F^\Phi(\zeta_s^-)-L_F^\Phi(\zeta_{s'}^+)\right]\left(\frac{1}{k_0+\zeta_s^-+\zeta_{s'}^+}+\frac{1}{-k_0+\zeta_s^-+\zeta_{s'}^+}\right),
G~ss′A=B=3(q+,q−;Φ)=[LFΦ(ζs−)−LFΦ(ζs′+)](1k0−ζs−+ζs′++1−k0−ζs−+ζs′+),(40)\widetilde{\mathcal{G}}^{A=B=3}_{ss'}(q_+,q_-;\Phi)=\left[L_F^\Phi(\zeta_s^-)-L_F^\Phi(\zeta_{s'}^+)\right]\left(\frac{1}{k_0-\zeta_s^-+\zeta_{s'}^+}+\frac{1}{-k_0-\zeta_s^-+\zeta_{s'}^+}\right), \tag*{(40)}

which contain only the real part of the thermal distribution functions. On the other hand, the (A,B)=(∓,∓)(A,B)=(\mp,\mp) components are

Gss′∓∓=[1−LFΦ(ζs−)−LFΦ(ζs′+)](1k0∓iR+ζs−+ζs′++1−(k0∓iR)+ζs−+ζs′+)\mathcal{G}^{\mp\mp}_{ss'}=\left[1-L_F^\Phi(\zeta_s^-)-L_F^\Phi(\zeta_{s'}^+)\right]\left(\frac{1}{k_0\mp iR+\zeta_s^-+\zeta_{s'}^+}+\frac{1}{-(k_0\mp iR)+\zeta_s^-+\zeta_{s'}^+}\right)
±i[L~FΦ(ζs−)+L~FΦ(ζs′+)](1k0∓iR−ζs−+ζs′+−1−(k0∓iR)−ζs−+ζs′+),\qquad{}\pm i\left[\widetilde{L}_F^\Phi(\zeta_s^-)+\widetilde{L}_F^\Phi(\zeta_{s'}^+)\right]\left(\frac{1}{k_0\mp iR-\zeta_s^-+\zeta_{s'}^+}-\frac{1}{-(k_0\mp iR)-\zeta_s^-+\zeta_{s'}^+}\right),
G~ss′∓∓=[LFΦ(ζs−)−LFΦ(ζs′+)](1k0∓iR−ζs−+ζs′++1−(k0∓iR)−ζs−+ζs′+)\widetilde{\mathcal{G}}^{\mp\mp}_{ss'}=\left[L_F^\Phi(\zeta_s^-)-L_F^\Phi(\zeta_{s'}^+)\right]\left(\frac{1}{k_0\mp iR-\zeta_s^-+\zeta_{s'}^+}+\frac{1}{-(k_0\mp iR)-\zeta_s^-+\zeta_{s'}^+}\right)
±i[L~FΦ(ζs−)+L~FΦ(ζs′+)](1k0∓iR−ζs−+ζs′+−1−(k0∓iR)−ζs−+ζs′+),(41)\qquad{}\pm i\left[\widetilde{L}_F^\Phi(\zeta_s^-)+\widetilde{L}_F^\Phi(\zeta_{s'}^+)\right]\left(\frac{1}{k_0\mp iR-\zeta_s^-+\zeta_{s'}^+}-\frac{1}{-(k_0\mp iR)-\zeta_s^-+\zeta_{s'}^+}\right), \tag*{(41)}

which contain the real and imaginary parts of the thermal distribution functions. It is important to note that GAB\mathcal{G}^{AB} and G~AB\widetilde{\mathcal{G}}^{AB} are functions of kAk_A; this verifies Equation (28), where we wrote ΠAB\Pi^{AB} as functions of kAk_A. We also note that the imaginary parts of G\mathcal{G} and G~\widetilde{\mathcal{G}} are common, Im⁡Gss′∓∓=Im⁡G~ss′∓∓\operatorname{Im}\mathcal{G}^{\mp\mp}_{ss'}=\operatorname{Im}\widetilde{\mathcal{G}}^{\mp\mp}_{ss'}. These functions vanish at (kA)0=0(k_A)_0=0, and we will therefore not discuss them further.

The Full Expression After Adding the pYM Type Self-Energy

Below, we examine the electric and magnetic propagators at K2=−kA2K^2=-k_A^2. It is known that the vacuum propagators in the pYM theories are well described by the massive gluon propagator, including the gluon one-loop polarization function, ΠvacG(k;mg,αsG,μR)\Pi_{\mathrm{vac}}^G(k;m_g,\alpha_s^G,\mu_R), whose explicit expression is given by Equation (17) in Ref. [75]. Here, μR\mu_R is the renormalization scale for which we take μR=1 GeV\mu_R=1\ \mathrm{GeV}, and αsG\alpha_s^G is the coupling in the gluon sector, which generally differs from the quark–gluon coupling αsq\alpha_s^q. In the following, we begin with the vacuum propagators fitted to lattice data and then examine the impact of in-medium quark contributions. The propagators are expressed as

DμνAB(kA)=DEABPμνE+DMABPμνM,DE,MAB(kA)=δABK2+mg2+ΠvacG+ΠvacR+δΠE,M,(42)D_{\mu\nu}^{AB}(k_A)=D_E^{AB}P_{\mu\nu}^E+D_M^{AB}P_{\mu\nu}^M,\qquad D_{E,M}^{AB}(k_A)=\frac{\delta^{AB}}{K^2+m_g^2+\Pi_{\mathrm{vac}}^G+\Pi_{\mathrm{vac}}^R+\delta\Pi_{E,M}}, \tag*{(42)}

where ΠvacR\Pi_{\mathrm{vac}}^R and δΠE,M\delta\Pi_{E,M} are defined in Equations (31) and (33).

One Loop vs. Lattice Results: Screening Masses

From now on, we compare our computations with the lattice results of Ref. [74]. For concreteness, we focus on the data at β=2.1\beta=2.1, with the inverse lattice spacing a−1=1.41 GeVa^{-1}=1.41\ \mathrm{GeV}, the pion mass mπ≃637 MeVm_\pi\simeq637\ \mathrm{MeV}, spatial size Ls≃2.21 fmL_s\simeq2.21\ \mathrm{fm}, and temporal size LtL_t corresponding to temperatures ranging from T≃44 MeVT\simeq44\ \mathrm{MeV} to 353 MeV353\ \mathrm{MeV}. This setup is very similar to the “fine ensemble” lattice data in Ref. [15] for the Polyakov loop in the previous section, where β≃2.1\beta\simeq2.1, a−1=1.43 GeVa^{-1}=1.43\ \mathrm{GeV}, and mπ≃637 MeVm_\pi\simeq637\ \mathrm{MeV}. In Ref. [74], the results at T≃44 MeVT\simeq44\ \mathrm{MeV} are regarded as effectively zero temperature data. Here, however, we explicitly treat them as T=44 MeVT=44\ \mathrm{MeV} results, since analytic calculations at T=0T=0 and T=44 MeVT=44\ \mathrm{MeV} show considerable deviations at high densities.

Unfortunately, the lattice data at small momenta are sensitive to system size, and the statistical errors are non-negligible. Moreover, the lattice propagators were not renormalized, so the overall normalization of the field amplitude is arbitrary. To minimize both statistical and systematic uncertainties, we normalize each propagator by its value at K2=0K^2 = 0 at a chosen reference point of (μq,T)(\mu_q, T).

For the analytic results, we generally use (Φ,Mq,Δ)(\Phi, M_q, \Delta) obtained from our gap Equation (21) with Δext=45\Delta_{\rm ext} = 45 MeV, except for the μq=0\mu_q = 0 case, where the lattice computations set the external field to Δext=0\Delta_{\rm ext} = 0. Unless otherwise stated, we take αsG=1.0\alpha_s^G = 1.0; variations in αsG\alpha_s^G can largely be compensated by adjusting mgm_g, for which we test several values.

Density Evolution of the Screening Masses: Qualitative Guides from Fixed Φ\Phi and Δ\Delta

To establish theoretical baselines, we first examine the screening masses with Polyakov loops set to Φ=0\Phi= 0 (confined) and Φ=1\Phi= 1 (deconfined) for various diquark gaps, Δ=1\Delta= 1, 100, and 200 MeV. The quark mass MqM_q is determined by solving the gap equations for each (Φ,Δ)(\Phi, \Delta) configuration. These results serve as guidelines for comparison with the lattice data.

We define the screening masses as follows:

mE/M2(μq,T)≡1/DE/M(k=0;μq,T).(43)m^2_{E/M}(\mu_q,T) \equiv1/D_{E/M}(k=0;\mu_q,T). \tag*{(43)}

However, these expressions are not directly usable until a normalization is specified. To remove this arbitrariness and reduce statistical fluctuations in the lattice data, we normalize the screening masses by their values at zero density and "zero temperature" for each μq\mu_q:

m‾E/M(μq,T)≡mE/M(μq,T)mE/M(μq∗,T∗),(μq∗,T∗)=(0,44) MeV.(44)\overline{m}_{E/M}(\mu_q,T) \equiv\frac{m_{E/M}(\mu_q,T)}{m_{E/M}(\mu_q^*,T^*)}, \qquad(\mu_q^*,T^*)=(0,44)\ {\rm MeV}. \tag*{(44)}

Shown in Figure 11 are the normalized electric (left panels) and magnetic (right panels) screening masses at low temperature, T≃44T \simeq44 MeV. For the analytic results, we display two cases: Φ=1\Phi= 1 (upper panels, deconfined) and Φ=0\Phi= 0 (lower panels, confined). The error bands on the theoretical curves correspond to variations of αsG\alpha_s^G from 0.5 to 2, with the central solid curve representing αsG=1\alpha_s^G = 1.

Figure 11. Theoretical guidelines vs lattice data for the normalized electric (left panels) and magnetic (right panels) screening masses, mˉE,M\bar{m}_{E,M}, for the Φ=1\Phi= 1 (upper panels, deconfined) and Φ=0\Phi= 0 (lower panels, confined) cases. The diquark gaps Δ=1\Delta= 1, 100, and 200 MeV and mg=500m_g = 500 MeV are chosen for theoretical curves. The error bands in theoretical curves originate from the variation in αsq\alpha_s^q, ranging from 0.5 to 2, with the central solid curve being αsq=1\alpha_s^q = 1. To fit the lattice data in the electric sector, it is essential to quench thermal quarks by the diquark gap and/or Polyakov loop effects; even at temperatures as low as T≃44T \simeq44 MeV, the abundance of thermal quarks is not negligible at large densities. In contrast, the theory predicts that the magnetic sector is free from the quark screening effects.

For the electric sector, it is essential to consider a relatively large gap, Δ∼200\Delta\sim200 MeV, and/or to include the Polyakov loop suppression of thermal quarks. The theoretical curves for the electric screening mass suggest that, for Δ≲100\Delta\lesssim100 MeV, the screening masses increase too rapidly once the BCS regime sets in. This estimate of Δ∼200\Delta\sim200 MeV is consistent with the lattice results for the superfluid critical temperature, Ts≃100T_s \simeq100 MeV, in the BCS region; applying the BCS estimate, ΔBCS≃Ts/0.57∼180\Delta^{\rm BCS} \simeq T_s/0.57 \sim180 MeV [91], supports this observation.

Including the Polyakov loop suppresses thermal quark contributions, tempering the evolution of the electric screening toward high densities. Near the Fermi surface, the abundance of thermal quarks and baryons (diquarks) scales as ∼Φe−βΔ\sim\Phi e^{-\beta\Delta}, ∼Φe−βΔ\sim\Phi e^{-\beta\Delta} and ∼e−2βΔ\sim e^{-2\beta\Delta}, respectively, controlling the strength of quark screening effects. The impact of the Polyakov loop is small for Δ≃0\Delta\simeq0 as e−βΔ≃e−2βΔ≃1e^{-\beta\Delta} \simeq e^{-2\beta\Delta} \simeq1, while for a substantial Δ\Delta, roughly twice the temperature is required for the confined case (Φ=0\Phi= 0) to develop a screening mass as large as in the deconfined case (Φ=1\Phi= 1). Equivalently, for a fixed TT, the results for Δ\Delta with Φ=0\Phi= 0 resemble those for 2Δ2\Delta with Φ=1\Phi= 1. This correspondence can be verified by comparing the curve for Δ=100\Delta= 100 MeV with Φ=0\Phi= 0 to that for Δ=200\Delta= 200 MeV with Φ=1\Phi= 1, i.e., e−2β×100 MeV=e−β×200 MeVe^{-2\beta\times100\,{\rm MeV}} = e^{-\beta\times200\,{\rm MeV}}, as shown in Figure 11.

The situation is very different in the magnetic sector. The theoretical curves show essentially no in-medium modifications of the magnetic screening mass with increasing μq\mu_q (and also with increasing TT, see Figure 12). The lattice results, however, suggest that the magnetic mass starts to increase from μq∼700\mu_q \sim700 MeV, a trend that cannot be captured by the present one-loop calculations.

Thermal evolution of normalized electric and magnetic screening masses

Figure 12. The thermal evolution of the normalized electric (left panels) and magnetic (right panels) screening masses mE,Mm_{E,M} for several μq\mu_q. The cases for mg=400m_g = 400 MeV (upper panels) and 500 MeV (lower panels) are displayed. The error bands are set in the same way as Figure 11.

The values of (Φ,Δ)(\Phi,\Delta) dynamically determined from the gap equations remain within the range Φ∼0\Phi\sim0 and Δ∼200\Delta\sim200–300300 MeV. Consequently, the screening masses closely follow the curves for Δ=200\Delta= 200 MeV and Φ=0\Phi= 0, where both the electric and magnetic masses remain nearly constant over the range of μq\mu_q explored in this work. For this reason, we do not separately display the results of dynamically determined gaps here.

Thermal Evolution of the Screening Masses

Next, we examine the temperature evolution of the static screening masses for μq=0\mu_q = 0, 282, 423, 564, and 705 MeV. Shown in Figure 12 are the normalized electric and magnetic screening masses. For the analytic results, we present two cases with mg=400m_g = 400 MeV (upper panels) and 500 MeV (lower panels). The error bands for the theoretical curves reflect the variation in αsq\alpha_s^q from 0.5 to 2, with the central solid curve corresponding to αsq=1\alpha_s^q = 1.

For the electric sector, the screening mass increases with TT, as expected from the conventional picture of the thermal Debye mass. The electric screening is primarily governed by the gap Δ\Delta; in the absence of Δ\Delta, the electric mass of order ∼gsμq\sim g_s\mu_q already emerges at T=0T = 0. The Polyakov loop suppresses thermal quark fluctuations and thereby moderates the growth of thermal contributions to the screening. The results with mg=400m_g = 400 MeV appear to reproduce the lattice data well, whereas those with mg=500m_g = 500 MeV tend to slightly underestimate them. At present, however, it is difficult to pin down the optimal value of mgm_g from these computations, since its effect can be compensated by the behavior of Δ\Delta, which is not yet conclusively determined. In fact, our determination of Δ\Delta is not entirely free from cutoff artifacts, as already mentioned. We suspect that the value of Δ\Delta in our PNJL model—estimated as ∼\sim220–300 MeV for Δex=45\Delta_{\mathrm{ex}} = 45 MeV—and the resulting superfluid critical temperature Tc∼150T_c \sim150 MeV may be somewhat too large. If this expectation is correct, the quark contributions to mgm_g should be more pronounced than in the present computations, in which case a larger mgm_g or αsq\alpha_s^q value may be acceptable.

For the magnetic sector, the quark contributions to the screening mass vanish for all μq\mu_q and TT. Consequently, the magnetic screening mass remains insensitive to variations in μq\mu_q and TT. This trend is consistent with the lattice data up to T≃250T \simeq250 MeV. At zero momentum (∣k⃗∣=0|\vec{k}| = 0), we find no in-medium modifications for the analytically computed magnetic propagators, whereas at finite spatial momenta such modifications do appear, as will be discussed in the next subsection.

One Loop vs. Lattice Results: Gluon Propagators at Finite Momenta

Next, we examine the momentum dependence of the gluon propagators. We focus on the static case with k0=0k_0 = 0 but cover a wide range of spatial momenta, since such modes play the dominant role in low-energy processes. We select specific (μq,T)(\mu_q, T) points where lattice data are available [74]. Our analysis begins with the thermal evolution at μq=0\mu_q = 0 and μq=705\mu_q = 705 MeV, followed by the density evolution at T=44T = 44 MeV and T=118T = 118 MeV. The results at T=118T = 118 MeV, however, turn out to be very sensitive to the location of the diquark critical temperature and therefore require some additional analysis.

Thermal Evolution at q=0 and q=705 MeV

First, we examine the thermal evolution of the propagators at μq=0\mu_q = 0 and μq=705\mu_q = 705 MeV. Figure 13 compares our analytic results with the lattice data. The propagators are normalized as follows:

Figure 13. The thermal evolution of the normalized electric and magnetic propagators D‾E,M(k0=0,k⃗)\overline{D}_{E,M}(k_0 = 0,\vec{k}) for μq=0\mu_q = 0 (upper panels) and μq=705\mu_q = 705 MeV (lower panels). The value of mgm_g is fixed to 500 MeV, and the error bands are set in the same way as Figure 11. The magnetic propagators show the paramagnetic enhancement at finite momenta.

D‾E/M(μq,T)≡DE/M(μq,T)DE/M(μq,T∗),T∗=44 MeV.(45)\overline{D}_{E/M}(\mu_q,T) \equiv\frac{D_{E/M}(\mu_q,T)}{D_{E/M}(\mu_q,T^*)}, \qquad T^* = 44\ \mathrm{MeV}. \tag*{(45)}

Here, we show the results for mg=500m_g = 500 MeV, with the range of αsq\alpha_s^q varied between 0.5 and 2.0, as in the previous analyses.

For μq=0\mu_q = 0, the electric propagators in the lattice data show a slightly stronger suppression than in our analytic results. This discrepancy becomes less pronounced if we reduce the value of mgm_g, so we consider it to lie within the range of theoretical uncertainties. The magnetic propagators, on the other hand, are well reproduced across the entire temperature range studied.

Next, we turn to the μq=705\mu_q = 705 MeV case. The electric propagators show good consistency between the analytic and lattice results. In contrast, the magnetic propagators exhibit a sizable discrepancy at ∣k⃗∣=0|\vec{k}| = 0, for which we currently have no clear explanation. This suggests that computations beyond one-loop order or genuinely non-perturbative effects may be required. Meanwhile, the analytic and lattice results are in good agreement for ∣k⃗∣≳500|\vec{k}| \gtrsim500 MeV.

Density Evolution at T=44T = 44 MeV and T=118T = 118 MeV

Here, we examine the density evolution of the propagators at fixed temperatures, T=44T = 44 MeV and T=118T = 118 MeV. Shown in Figure 14 are comparisons between the analytic results and lattice data. Here, the propagators are normalized as follows:

Figure 14. The density evolution of the normalized electric and magnetic propagators D‾E,M(k0=0,k⃗)\overline{D}_{E,M}(k_0 = 0,\vec{k}) for T=44T = 44 MeV (upper panels) and T=118T = 118 MeV (lower panels). The value of mgm_g is fixed to 500 MeV, and the error bands are set in the same way as Figure 11. The magnetic propagators show the paramagnetic enhancement at finite momenta. At T=118T = 118 MeV, the TT dependence of the electric propagators is considerably weaker than the lattice data, likely because ΔT=118 MeV\Delta_{T=118\ \mathrm{MeV}} is overestimated.

D‾E/M(μq,T)≡DE/M(μq,T)DE/M(μq∗,T),μq∗=0.(46)\overline{D}_{E/M}(\mu_q,T) \equiv\frac{D_{E/M}(\mu_q,T)}{D_{E/M}(\mu_q^*,T)}, \qquad\mu_q^* = 0 . \tag*{(46)}

First, we consider the T=44T = 44 MeV results. In the lattice data, the electric propagators at μq=423\mu_q = 423 and 564564 MeV show an enhancement at low momenta, but we suspect that this feature is simply due to statistical fluctuations (see also Figure 11). For μq=282\mu_q = 282 and 705705 MeV, the analytic results are consistent with the lattice data at ∣k⃗∣∼0|\vec{k}| \sim0 but remain slightly smaller than the lattice values for ∣k⃗∣≳500|\vec{k}| \gtrsim500 MeV. This tension can be alleviated by increasing mgm_g. Taking these observations together, we conclude that the agreement between the analytic and lattice results at T=44T = 44 MeV is acceptable.

As for the magnetic propagators at finite momenta, we observe a significant quark contribution through paramagnetic enhancement, which grows with increasing μq\mu_q. It is important to note that choosing too large a value of αsq\alpha_s^q leads to a serious disagreement between the analytic and lattice results. Furthermore, extrapolating the one-loop results with fixed αsq\alpha_s^q to very high density eventually causes chromo-magnetic instability [92,93], which has been extensively discussed in the context of gapless color superconductivity [94,95]. A possible way to avoid such an instability is to invoke a “gluon saturation” mechanism, as in the Gribov–Zwanziger model [84,85], which would drive mgm_g to increase with μq\mu_q before the paramagnetic enhancement causes the instability. This conceptual issue, however, lies beyond the scope of the present paper.

Now, we turn to the case of T=118T = 118 MeV. Our analytic results show substantially weaker electric screening compared to the lattice results. This discrepancy may be understood as follows. As mentioned earlier, the Δ\Delta and the superfluid critical temperature TsT_s in our PNJL model are about 1.2–1.5 times larger than those inferred from the lattice. Consequently, while T=118T = 118 MeV lies slightly above the lattice TsT_s, it is still smaller than our model’s TsT_s by about 20–50 MeV. To examine how this discrepancy affects the results, we compute the analytic propagators with fixed values of Δ\Delta of Δ=10\Delta= 10 and 100 MeV while determining (Φ,Mq)(\Phi, M_q) self-consistently from the gap equation. The corresponding results are shown in Figure 15. Using smaller Δ\Delta than predicted in our PNJL model enhances electric screening and hence suppresses the propagators, thereby improving agreement with the lattice data.

Figure 15. The normalized electric and magnetic propagators DˉE,M(k0=0,k⃗)\bar{D}_{E,M}(k_0 = 0,\vec{k}) at T=118T = 118 MeV. The diquark gaps are held fixed to ΔT=118=10\Delta_{T=118} = 10 MeV (upper panel) and 100 MeV (lower panel).

For the magnetic propagators at T=118T = 118 MeV, the agreement with the lattice results is generally good, as shown in Figure 14 and Figure 15. We also notice that reducing Δ\Delta slightly weakens the paramagnetic enhancement at finite ∣k⃗∣|\vec{k}|.

Discussion and Summary

In this paper, we delineate the role of thermal quarks in dense QC2D and their impact on gluon dynamics, focusing specifically on the Polyakov loop and gluon propagators.

Our initial approach was to start from the massive gluon propagators constrained by lattice data and then use them to compute the Polyakov loop potential. However, the resulting potential for the pYM\mathrm{pYM} turns out to be too flat; the inclusion of quark contributions drastically deforms the structure of the Polyakov loop potential, leading to an overly rapid deconfinement that disagrees with lattice QCD results. We therefore speculate that the massive gluons appearing in the propagators and those responsible for confinement (the Polyakov loop potential) are somewhat distinct: the former corresponds to excitations above the confining vacuum, while the gluons forming the Polyakov loop potential are not like quasi-particles and appear to be more robust against quark-induced perturbations.

To investigate the nature of thermal excitations in dense matter, we employ phenomenological Polyakov loop potentials used in the literature and examine their correlation with thermal quarks in a manner similar to the PNJL and PQM models. We tune our setup to be consistent with lattice results; in particular, we adjust our model parameters to reproduce a pion mass of mπ∼640m_\pi\sim640 MeV and introduce a finite diquark source term, Δex=45\Delta_{\mathrm{ex}} = 45 MeV, which is commonly used to stabilize lattice simulations. The renormalization of the Polyakov loop is also taken into account.

After calibrating our model to reproduce lattice results at μq=0\mu_q = 0, we study the Polyakov loop at finite μq\mu_q and find reasonable agreement with the lattice data. Achieving this agreement requires a diquark gap of Δ∼200\Delta\sim200–300300 MeV and a finite Δex\Delta_{\mathrm{ex}}. If Δ\Delta is artificially set to zero, the Polyakov loop reaches the deconfined solution too rapidly due to the large abundance of gapless quarks.

We next delineate in-medium gluon propagators in the Landau–DeWitt gauge. At the one-loop level, it is known that at zero temperature, the color-singlet diquark gap protects gluons from screening by quarks. Thermal quarks overcoming the gap energy contribute to the screening as in the normal phase; hence, the thermal modifications of the gluon mass can be used to study the nature of thermal excitations and how confinement affects them.

The Polyakov loop effects quenches the colored quarks and suppresses their contributions to the screening. These confining effects are important at the lowest temperature measured on the lattice, T∼44T \sim44 MeV, and should be taken into account when confronting theoretical computations with the lattice data. The thermal evolution of the screening masses at various μq\mu_q appears reasonably consistent with the lattice results for the range of our parameters, αsq=0.5\alpha_s^q = 0.5–2.02.0 and mg=400m_g = 400–500500 MeV.

Overall, our analyses with the current lattice data suggest that thermal colored excitations in dense two-color QCD matter are strongly suppressed. In the Φ=1\Phi= 1 case, the electric screening grows too rapidly compared to lattice results, unless Δ\Delta is considerably larger than ∼200\sim200 MeV (this is unlikely however), allowing thermal quarks to be effectively neglected even without Polyakov loop suppression. Including the Polyakov loop increases the minimum excitation from ∼Δ\sim\Delta to ∼2Δ\sim2\Delta, corresponding to the energy scale of baryonic (diquark) excitations.

To draw more definitive conclusions about confinement and deconfinement in dense QC2_2D matter, however, lattice data with smaller statistical fluctuations and better-controlled finite-volume and lattice-spacing effects are called for. On the theoretical side, the analytic computations should be improved by including higher-order loop corrections and genuinely non-perturbative effects.

While this work focuses on the proof of principle about confinement/deconfinement in dense quark matter, ultimately, the framework should be applied to three-color QCD. The quark back reaction to the gluon sector highly depends on the phase structure. For instance, in the case of the color-flavor-locked phase, there are both Debye and Meissner masses with which both electric and magnetic gluons are screened. Another probability, as suggested in the quarkyonic matter scenario, is that the Fermi surface is dominated by color-singlet composites which do not strongly screen gluons, as in the color-singlet diquarks in QC2_2D. These physics, dependent on the phase structure, are left for future studies.

Author Contributions: Conceptualization, Y.K., T.K., and D.S.; formal analysis, Y.K. and T.K.; writing—original draft preparation, Y.K. and T.K.; writing—review and editing, T.K. and D.S. All authors have read and agreed to the published version of the manuscript.

Funding: T.K. was supported by JSPS KAKENHI Grant No. 23K03377; D.S. by JSPS KAKENHI Grant No. 23H05439, No. 23H05439, and No. 25K17386.

Data Availability Statement: The data will be provided upon request.

Acknowledgments: We thank A. Maas and J. Skullerud for kindly providing us with their lattice data in Refs. [15] [74], and Itou for discussions.

Conflicts of Interest: The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:

QC2DQC_{2}D two-color quantum chromodynamics

QCDIQCD_{I} isospin QCD

EOS equations of state

pYM pure Yang–Mills

mYM massive Yang–Mills

PNJL Polyakov–Nambu–Lasinio

BFH Brauner–Fukushima–Hidaka

BCS Bardeen–Cooper–Schrieffer

BEC Bose–Einstein condensation

References

Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

References

  1. [1]S.; Kenny, P.; Kim, S.; Skullerud, J.I. Lattice Study of Dense Matter with Two Colors and Four Flavors. Eur. Phys. J. A 2011, 47, 60. doi.org/10.1140/epja/i2011-11060-1.DOI
  2. [2]antsev, N.; Braguta, V.; Ilgenfritz, E.; Kotov, A.; Nikolaev, A. Lattice study of thermodynamic properties of dense QC2D. Phys. Rev. D 2020, 102, 074507. https://doi.org/10.1103/PhysRevD.102.074507.
  3. [3]ov, V.G.; Braguta, V.V.; Nikolaev, A.A.; Rogalyov, R.N. Effects of Dense Quark Matter on Gluon Propagators in Lattice QC2D. Phys. Rev. D 2020, 102, 114511. https://doi.org/10.1103/PhysRevD.102.114511.
  4. [4]S.; Nakamura, A.; Nonaka, C. Behavior of hadrons at finite density: Lattice study of color SU(2) QCD. Phys. Lett. B 2003, 551, 305–310. doi.org/10.1016/S0370-2693(02)03065-4.
  5. [5]R.; Detmold, W.; Romero-López, F.; Davoudi, Z.; Illa, M.; Parreño, A.; Perry, R.J.; Shanahan, P.E.; Wagman, M.L. Lattice quantum dynamics at large isospin density. Phys. Rev. D 2023, 108, 114506. https://doi.org/10.1103/PhysRevD.108.114506.
  6. [6]R.; Detmold, W.; Illa, M.; Parreño, A.; Perry, R.J.; Romero-López, F.; Shanahan, P.E.; Wagman, M.L. QCD Constraints on Isospin-Dense Matter and the Nuclear Equation of State. Phys. Rev. Lett. 2025, 134, 011903. https://doi.org/10.1103/PhysRevLett.134.011903.
  7. [7]B.; Cuteri, F.; Endrodi, G. Equation of state and speed of sound of isospin-asymmetric QCD on the lattice. J. High Energy Phys. 2023, https://doi.org/10.1007/JHEP07(2023)055.
  8. [8]B.B.; Endrodi, G.; Schmalzbauer, S. QCD phase diagram for nonzero isospin-asymmetry. Phys. Rev. D 2018, 97, 054514. https://doi.org/10.1103/PhysRevD.97.054514.
  9. [9]o, Y. Interplay between the weak-coupling results and the lattice data in dense QCD. arXiv 2024. https://doi.org/10.48550/arXiv.2408
  10. [10]o, Y. Enhanced contribution of the pairing gap to the QCD equation of state at large isospin chemical potential. Phys. Rev. D 2024, https://doi.org/10.1103/PhysRevD.109.054035.
  11. [11]ma, K.; Minato, S. Speed of sound and trace anomaly in a unified treatment of the two-color diquark superfluid, the pion-condensed spin matter, and the 2SC quark matter. Phys. Rev. D 2025, 111, 094006. https://doi.org/10.1103/PhysRevD.111.094006.
  12. [12]Powell, P.D.; Song, Y.; Baym, G. Phenomenological QCD equation of state for massive neutron stars. Phys. Rev. D 2015, 91, 045003. doi.org/10.1103/PhysRevD.91.045003.DOI
  13. [13]Baym, G.; Hatsuda, T. Implications of NICER for Neutron Star Matter: The QHC21 Equation of State. Astrophys. J. 2022, 934, 46. doi.org/10.3847/1538-4357/ac8786.DOI
  14. [14]A.; Gorda, T.; Braun, J. Color superconductivity under neutron-star conditions at next-to-leading order. arXiv 2025. https://doi.org/arXiv.2504.03834.
  15. [15]A.; Gorda, T.; Braun, J. Pressure and speed of sound in two-flavor color-superconducting quark matter at next-to-leading order. Phys. 2024, 110, 014034. https://doi.org/10.1103/PhysRevD.110.014034.
  16. [16]Suenaga, D. Peaks of sound velocity in two color dense QCD: Quark saturation effects and semishort range correlations. Phys. Rev. D 2025, 076001. https://doi.org/10.1103/PhysRevD.105.076001.
  17. [17].; Kojo, T. Sound velocity peak and conformality in isospin QCD. Phys. Rev. D 2024, 109, 076006. https://doi.org/10.1103/PhysRevD.
  18. [18]Stiffening of matter in quark–hadron continuity: A mini-review. J. Subat. Part. Cosmol. 2025, 4, 100088. https://doi.org/10.1016/j.5.100088.
  19. [19]ekal, L. Universal Aspects of QCD-like Theories. Nucl. Phys. B Proc. Suppl. 2012, 228, 179–220. https://doi.org/10.1016/j.nuclphysbps.006.
  20. [20]do, K.; Strothoff, N.; von Smekal, L.; Wambach, J. Fluctuations in the quark–meson model for QCD with isospin chemical potential. Phys. Rev. D 2013, 718, 1044–1053. https://doi.org/10.1016/j.physletb.2012.11.055.
  21. [21]en, J.O.; Adhikari, P.; Kneschke, P. Pion condensation and QCD phase diagram at finite isospin density. In Proceedings of the XIII Confinement and the Hadron Spectrum—Confinement 2018, Maynooth, Ireland, 31 July–6 August 2018; 197. https://doi.org/10.22323
  22. [22]R.; Kojo, T.; Suenaga, D. Thermal effects on sound velocity peak and conformality in isospin QCD. Phys. Rev. D 2024, 110, 054037. doi.org/10.1103/PhysRevD.110.054037.DOI
  23. [23]B.B.; Chelnokov, V.; Endrodi, G.; Marko, G.; Scheid, D.; von Smekal, L. Renormalization group invariant mean-field model for QCD at finite isospin density. Phys. Rev. D 2025, 112, 054038. https://doi.org/10.1103/fryz-f3vw.
  24. [24]gno, J.P.; Gomez Dumm, D.; Scoccola, N.N. Cold isospin asymmetric baryonic rich matter in nonlocal NJL-like models. Phys. Rev. D 2025, 109, 094041. https://doi.org/10.1103/PhysRevD.109.094041.
  25. [25]S.; Duarte, D.C.; Farias, R.L.S.; Ramos, R.O. Speed of sound peak in isospin QCD: A natural prediction of the Medium Separation. arXiv 2025. https://doi.org/10.48550/arXiv.2507.14343.
  26. [26].; Lopes, B.S.; Farias, R.L.S.; Parra, L.C. On the origin of the peak of the sound velocity for isospin imbalanced strongly interacting. Phys. Lett. B 2025, 864, 139396. https://doi.org/10.1016/j.physletb.2025.139396. , K.; Hatsuda, T.; Takatsuka, T. Hadron-Quark Crossover and Massive Hybrid Stars with Strangeness. Astrophys. J. 2013, 764, 12. doi.org/10.1088/0004-637X/764/1/12.

Paper details

vlnc:W_2hvczg96
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