Dileptons at Colliders as Probes of the Quark-Gluon Plasma

TL;DR

The paper reviews dileptons as penetrating electromagnetic probes of the QGP in ultra-relativistic heavy-ion collisions, highlighting how the invariant mass $m_{ m ll}$ differentiates partonic from hadronic emission and enables access to the medium’s early temperature, equilibration, and possible chiral-symmetry restoration via vector-meson spectral functions. It contrasts two main theoretical frameworks—effective many-body theory (via in-medium spectral functions) and microscopic transport models (e.g., PHSD)—and discusses pre-equilibrium contributions modeled with QCD kinetic theory, emphasizing their complementary roles in predicting dilepton yields across the hadronic and partonic phases. The review summarizes experimental results from RHIC and the LHC, including low-mass excesses consistent with thermal radiation and high-mass signals compatible with early QGP emission, while detailing the significant backgrounds from hadronic decays and heavy-flavor processes and the methods used to mitigate them (e.g., vertexing and DCA-based techniques). It also outlines the ongoing and planned upgrades (ALICE upgrades, ALICE 3, LHCb-UII) that will enable more precise, differential, and time-resolved dilepton measurements, with the goal of constraining transport coefficients like $rac{ ilde{ u}}{s}$, the QCD equation of state near $T_c$, and the dynamics of early thermalization.

Abstract

Ultra-relativistic heavy-ion collisions are used to create a deconfined state of quarks and gluons, the quark-gluon plasma (QGP), similar to the matter in the early universe. Dileptons are a unique probe of the QGP. Being emitted during all stages of the collision without interacting strongly with the surrounding matter, they carry undistorted information about the medium evolution. The mass of the lepton-antilepton pair gives a unique mean to separate partonic from hadronic radiation. Thus, dileptons can be used to study the QGP equilibration time, its average temperature but also effects related to the restoration of chiral symmetry in the hot medium via vector meson decays. This information is not accessible with hadrons. The price to pay is a large background from ordinary hadron decays. We summarize the potential of dilepton measurements, the results obtained so far at colliders, and the ongoing efforts for future experiments with further increased sensitivity.

Figures (6)

• Figure 1: Measured dielectron yields a a function of invariant mass in the STAR (left panel from Reference STAR:2015tnn) and PHENIX (right panel from Reference PHENIX:2015vek) acceptances in minimum bias Au--Au collisions at $\sqrt{s_{\mathrm{NN}}}=0.2$ TeemV, compared to the contributions of known hadronic decays (cocktail). Statistical and systematic uncertainties on the data points are shown separately by vertical bars and boxes, respectively. The ratios data over cocktail simulations are shown in the bottom panels, where the band around unity indicates the uncertainties on the cocktail calculations.
• Figure 2: Left panel: Invariant mass distribution of the dielectron excess (data - cocktail) in the STAR acceptance in minimum bias Au--Au collisions at $\sqrt{s_{\mathrm{NN}}}=0.2$ TeemV compared to models. Panel from Reference STAR:2015tnn. Right panel: Dielectron yields scaled by $N_{\rm part}$ for the $\rho$-like (A) region with the cocktail subtracted and for the $\omega$-like (B) and the $\phi$-like (C) regions without cocktail subtraction as a function of $N_{\rm part}$, together with model calculations (see text) and a fit of the data in the $\rho$-like region. Statistical and systematic uncertainties from the data points are shown separately by vertical bars and grey boxes, respectively. Green brackets depict the total systematic uncertainties including those from cocktails (to be ignored for B and C). Panel from Reference STAR:2015tnn.
• Figure 3: Left panel: Acceptance-corrected dilepton excess mass spectra normalised by the charged-particle multiplicity in different colliding systems and at different $\sqrt{s_{\mathrm{NN}}}$ from STAR STAR:2024bpcHP2024starprelbis and NA60 NA60:2008ctj, fitted to extract temperature parameters. Vertical bars and boxes around data points represent the statistical and systematic uncertainties, respectively. Panel from Reference HP2024starprelbis. Right panel: Collision-energy dependence of the integrated dilepton excess yields in the mass range $0.4 < m_{\rm ll} < 0.75$ GeemV$/c^2$ measured by STAR STAR:2015zalSTAR:2023wtaHan:2024nzrsQM2024starprelHP2024starprelSTAR:2024bpc, NA60 NA60:2008ctj and HADES HADES:2019auv, normalised by the neutral-pion multiplicity and compared to calculations Rapp:2013nxa. Panel from Reference HP2024starprel.
• Figure 4: Left panel: Dielectron $m_{\rm ee}$-differential yield in the 10% most central Pb--Pb collisions at $\sqrt{s_{\mathrm{NN}}}=5.02$ TeemV, compared with the expected $\rm e^{+}e^{-}$ contributions from known hadronic decays (cocktails), including two different estimations for dielectrons from correlated heavy-flavour hadron decays. The error bars and boxes represent the statistical and systematic uncertainties of the data, whereas the bands show the uncertainties of the hadronic cocktails. Panel from Reference ALICE:2023jef. Right panel: Corresponding excess yield of dielectrons with respect to the cocktails, compared with predictions from the model of Rapp Rapp:2013nxa. The error bars and boxes represent the total statistical and systematic uncertainties including those from the cocktails ALICE:2023jef. Panel from Reference ALICE:2023jef.
• Figure 5: Left panel: Sketch of the dielectron spatial topology with the definition of the distance-of-closest approach (DCA) to the reconstructed collision vertex of the single $\rm e^{\pm}$ and $\rm e^{+}e^{-}$ pairs. Panel provided by Sebastian Scheid. Right panel: Fit of the inclusive $\rm e^{+}e^{-}$ yield in the 10% most central Pb--Pb collisions at $\sqrt{s_{\mathrm{NN}}}=5.02$ TeemV as a function of DCA$_{\rm ee}$ in the mass range $1.2 < m_{\rm ee} < 2.6$ GeemV$/c^2$. The error bars and boxes represent the statistical and systematic uncertainties of the data. Panel from Reference ALICE:2023jef.