Quark–gluon plasma (QGP) is a state of matter in which quarks and gluons — the particles that make up protons and neutrons — are no longer bound inside individual hadrons but move freely through a volume far larger than a nucleus. It is produced in the laboratory by colliding heavy atomic nuclei at close to the speed of light, and it is what the universe was made of in the first microseconds after the Big Bang. In the two decades since it was first created, it has turned out to be less like a gas of free particles than like a nearly frictionless liquid.[1][2]
Why ordinary matter cannot show it
Quantum chromodynamics, the theory of the strong interaction, has a peculiar property: its force carrier, the gluon, interacts with itself. At ordinary energies that self-interaction makes the force grow with distance, so quarks can never be pulled apart — they are permanently confined inside hadrons. But the same theory predicts that at sufficiently high temperature or density the coupling weakens and confinement dissolves, leaving quarks and gluons free to move over nuclear distances. Lattice calculations of QCD put the change at a pseudo-critical temperature of about 155 MeV, around two trillion kelvin, and describe it not as a sharp transition but as a rapid crossover. Chiral symmetry, broken in the vacuum, is partly restored at the same time.[3][4]
Nothing in the laboratory is that hot. The way round the problem is to heat a small volume extremely quickly: two heavy nuclei are accelerated to within a whisker of light speed and smashed together, and the energy deposited in the overlap region creates a fireball a few femtometres across. It lives for about 10 fm/c — some 10-23 seconds — expands and cools, and by the time it reaches the detectors it has condensed back into ordinary hadrons. Everything known about the plasma is inferred from the debris.[5]
The signatures
Four observations, none decisive alone, together make the case. The first is jet quenching: when a violent parton–parton collision produces two back-to-back jets of particles, one jet in a heavy-ion collision is markedly suppressed relative to what proton–proton collisions predict, because the parton has had to plow through dense coloured matter and radiate energy away. The suppression is quantified by the nuclear modification factor RAA, and it is the closest thing the field has to a direct measurement of the medium’s opacity.[6]
The second is collective flow. Collisions are never perfectly head-on, so the overlap region is almond-shaped; pressure gradients push the matter outwards more strongly along the short axis, and the resulting anisotropic distribution of particles — elliptic flow — is a hydrodynamic response. Reproducing it requires a fluid with extraordinarily low shear viscosity to entropy density ratio, close to a conjectured quantum lower bound. That is why the plasma is described as a near-perfect liquid rather than a gas.[7][8]
The third is strangeness enhancement: strange quarks are heavy enough that ordinary hadron collisions produce them rarely, but in a plasma with a high gluon density they are made copiously, so multistrange baryons appear far more often in heavy-ion collisions than in proton collisions. The fourth is quarkonia melting: bound states of a heavy quark and antiquark, such as the J/ψ, are pulled apart by Debye screening once the density of colour charges is high enough, so their yield drops — a thermometer calibrated by the binding energy of each state. Direct photons, which escape the fireball without interacting, provide a fourth, complementary channel.[9][10]
The surprise was not that deconfinement happens but what it looks like: the freest quarks in nature behave collectively, like a liquid with almost no viscosity, which is the opposite of the weakly interacting gas everyone had drawn on the blackboard.
How small can a plasma be?
For years the answer seemed to be: not very. QGP was a phenomenon of lead–lead or gold–gold collisions, and proton–proton collisions were treated as a trivial baseline. That assumption eroded through the 2010s, when the same flow patterns turned up in high-multiplicity proton–proton and proton–lead events. Whether those small systems genuinely form a plasma, or produce similar correlations by other means, has been the field’s most contested question.[11]
The LHC’s 2025 oxygen run was designed to settle it. Oxygen nuclei contain just 16 protons and neutrons, against about 200 in lead, and since the first oxygen–oxygen collisions all four large LHC experiments saw flow patterns reminiscent of the heavy-ion case. In July 2026 ALICE reported the sharper test: evidence for parton energy loss in oxygen–oxygen collisions, obtained by comparing neutral pions in oxygen–oxygen with those in proton–oxygen, where conventional nuclear effects should be similar but no medium large enough to stop a parton should form. The residual suppression is 4.9σ from the null hypothesis — roughly a one in two million chance of being an accident — extending the clearest QGP signature to the smallest nuclear system yet studied. CMS had already hinted at the same effect.[12]
Where the transition lies
The phase diagram of strongly interacting matter is plotted against temperature and baryon chemical potential, and the region of interest is where the two meet. Lattice QCD says the crossover at near-zero chemical potential is smooth; models suggest that at high baryon density the transition becomes first order, which would imply a critical point where the two regimes meet. Finding it is the goal of energy scans — RHIC running at lower collision energies, and fixed-target programmes at the SPS — which look for non-monotonic behaviour in fluctuations of conserved quantities as the beam energy is changed. Two decades of searching have not yet located it, and the field is candid that it may not be experimentally accessible.[13][14]
India’s contribution
Indian groups have been part of ALICE since its inception, funded by the Department of Atomic Energy and the Department of Science and Technology. The hardware contribution is concrete: the Photon Multiplicity Detector was conceived, designed and commissioned entirely in India, and Indian teams built components of the muon spectrometer, silicon pad detectors and the MANAS chip. Indian physicists have also been part of the analysis that established the plasma’s existence and characterised its properties, and the community runs its own detector R&D and simulation effort through institutions including VECC in Kolkata and centres participating in the ALICE–STAR–India programme.[15][16]
What the next decade holds
ALICE finished Run 3 on 14 June 2026 with about 53 billion lead–lead collisions recorded in continuous readout, alongside more than 150 pb-1 of proton–proton data, and entered Long Shutdown 3 at the end of that month. The analysis of that sample — precision measurements of the plasma’s transport properties, of charm and beauty quarks moving through it, and of the flow of light nuclei and hypernuclei — will occupy the field for years. Hardware upgrades already in train include a forward calorimeter and a new inner tracking system, with the much larger ALICE 3 detector planned for the 2030s.[17][18]
The open questions are as large as the data set. Nobody has a first-principles calculation of the plasma’s transport coefficients from QCD alone; the answers come from comparing hydrodynamic models to data. The fate of the critical point is unresolved. And the boundary between a plasma and a merely busy proton collision is still being mapped. What is settled is that the state exists, that it filled the early universe, and that it behaves like nothing the textbooks of the 1980s predicted.[19]
Select a year and month above to explore cold archived discussions.
Please select an archived topic above to view the discussion logs.