Does the QCD plasma contain gluons?

Abstract

Comparison of two appropriately chosen screening masses of colour singlet operators in the pure glue QCD plasma indicates that at sufficiently high temperature it contains a weakly-interacting massive quasi-particle with the quantum numbers of the electric gluon. Still in the deconfined phase, but closer to Tc, the same mass ratio is similar to that at zero temperature, indicating that the propagating modes are more glueball-like, albeit with a lower scale for the masses. We observe a continuity between these two regimes. With the RHIC fully operational and busy taking data, certain questions about the treatment of the QCD plasma have become urgent. One of the most basic is about the modes of excitation in a plasma: does it contain weakly interacting quark and gluon quasi-particles, or some more complicated collective excitations? We show here that lattice computations yield detailed answers to this question.The Debye screening mass, m D , in pure gauge QCD, at temperature T > T c has an expansion in the strong coupling g of the form-where the leading term g is well known, the second term has been extracted in perturbation theory [1] and the non-perturbative coefficients b = 2.46 ± 0.15 and c = −0.49 ± 0.15 have been computed in a lattice simulation of dimensionally reduced QCD [2]. The QCD coupling g has to be evaluated at the scale 6.742T . Since T c /Λ M S = 1.15 ± 0.05 [3], g = O(1) for T /T c ≃ 3 or less, and the error due to the neglect of the g 4 term is about 35d%, where d is the coefficient of this term. As a result, it becomes difficult to validate perturbation theory by comparing this screening mass to lattice data [4].In this work we address a prior question: at temperatures of interest to current and near-future experiments what mediates the longest correlations in the plasma? We examine this by comparing screening masses, m, obtained from correlations of two gauge invariant operators with different symmetry properties. They are chosen in such a way that one would be obtained by the exchange of two electric gluons while the other would need three, if indeed such gluons are the lightest excitations in the plasma. As a result, the two screening masses would be roughly in the ratio 3/2.Since screening masses involve the transfer matrix in a spatial direction, they are classified by the symmetry group of the lattice sliced perpendicular to a spatial direction [5,6,7]. For the thermodynamics of a 3+1 dimensional field theory realised on a hypercubic Euclidean lattice, since the Euclidean time direction is distinguished from the spatial directions, this is the group, where D 4 is the tetragonal group, Z 2 (C) is the charge-conjugation symmetry of the fields, and the Z 2 (T ) factor arises from the symmetry t ↔ −t [5,6]. The transfer matrix can be block diagonalised in irreps of this group.Extensive thermodynamic quantities depend only on the lowest eigenvalue of the transfer matrix, and the phase structure is determined by the degeneracies and symmetries of the corresponding eigenvectors. In high temperat...

Screening masses in the SU(3) pure gauge theory and universality

“…Such a behaviour was observed also by the authors of Ref. [37], whose results, when the comparison is possible, agree with ours. …”

Debye screening mass near deconfinement from holography

“…[9] where m D was defined as the largest inverse screening length among all the possible Euclidean correlation functions involving pairs of CT -odd operators in the thermal gauge field theory. Previous studies concerning thermal screening lengths in non-Abelian plasmas include lattice calculations [10][11][12][13][14][15][16], non-perturbative analyses of the gluon propagator at finite temperature [17][18][19][20], other analytical approaches [21,22], and holographic calculations [23][24][25].…”

Charmonium systems after the deconfinement transition

“…The ground state peaks and their strengths reproduce reasonably well the mass and amplitude obtained from a fit. The peaks at higher ω scale roughly with the lattice spacing and are probably dominated by lattice artefacts (see the discussion in reference [6]). …”

“…Here we report an update of our results in reference [3]. Further details, including discussion of systematics, can be found in [6].…”