The first 'heavy-ion run' at CERN's Large Hadron Collider (LHC) took place in November last year. In place of the usual proton beams, the accelerator was primed with bunches of lead ions, brought to collision at a centre-of-mass energy of 2.76 TeV per nucleon pair. The run was primarily for the benefit of ALICE, one of the four massive detectors in the LHC ring, designed to capture the debris of these heavy-ion collisions and investigate the formation of quark–gluon plasma.

But that didn't stop two of the other detectors — ATLAS and CMS — getting in on the act. Both are general-purpose detectors that are at the forefront of LHC searches for evidence in proton–proton collisions of the Higgs boson or supersymmetry. Both succeeded in recording data during the heavy-ion run, coping with the huge particle multiplicities that are characteristic of such collisions (and illustrated here in this event from CMS).

Credit: CERN

Now, in Physical Review Letters, the CMS collaboration report their analysis of the production of upsilon mesons in their heavy-ion data and in a complementary set of proton–proton data taken earlier this year. The comparison reveals a suppression of these mesons in the lead–lead collisions, which could indicate the formation of quark–gluon plasma.

At sufficiently high energy density, quarks and gluons are no longer confined but exist in a plasma, evidence of which has already emerged in earlier CERN experiments and particularly at the Relativistic Heavy Ion Collider at Brookhaven. In such an environment, bound states such as the upsilon — a bottom and antibottom quark pair — melt away. The melting temperature depends on binding energy, with excited states melting at lower temperatures than the ground state.

Upsilons decay into pairs of muons, which is a distinctive signature (the red tracks seen emerging above and below, here) that can even be picked out of heavy-ion data. Using their sample of such events, CMS have measured the ratio, for lead collisions compared with proton collisions, of another ratio — that of the production of the first two excited upsilon states to the ground state. The result is 0.31+0.19−0.15± 0.03 (statistical and systematic errors, respectively), suggesting that upsilons are indeed melting away in the heavy-ion collisions.