Has the deconfinement phase transition been seen?

In the mid-1990s a study of results from experiments at CERN (with a collision energy in the centre of mass of the nucleon pair of √s_NN = < 20 GeV) and the Alternating Gradient Synchrotron (AGS) at Brookhaven (√s_NN = 5.5 GeV) indicated some intriguing changes in the energy dependence of hadron production between top AGS and SPS energies (M Gazdzicki and D Röhrich 1996). Within a statistical model of the early stage of the collision process, these changes could be attributed to the onset of the deconfinement phase transition, where quarks and gluons are no longer confined within hadrons (M Gazdzicki and M Gorenstein 1999). The model predicted a sharp maximum in the multiplicity ratio of strange hadrons (hadrons that contain strange and anti-strange quarks) to pions (the lightest hadron) at the beginning of the transition region, at about √s_NN ~ 7.5 GeV. This prediction triggered a new experimental programme at the SPS – the energy scan programme – in which the NA49 experiment recorded head-on (central) collisions of two lead nuclei (Pb+Pb) at several energies, √s_NN = 6.3, 7.6, 8.7 and 12.3 GeV. Other heavy-ion experiments at the SPS (NA45, NA50, NA57 and NA60) participated in selected runs of the programme.

Recently published results from the energy scan, obtained mainly by the NA49 collaboration, have confirmed expectations. They indicate that rapid changes of hadron production properties occur within a narrow energy range of √s_NN = 7-12 GeV (V Friese et al. 2003). The “Collision energy dependence” figure shows these latest results, together with earlier data from the SPS and the AGS, and data from the Relativistic Heavy-Ion Collider (RHIC). Data from proton_proton collisions are also included for comparison. The top panel of the figure shows that the number of pions produced per nucleon participating in the collision increases with energy as expected in both proton-proton and nucleus-nucleus reactions. However, the rate of increase in nucleus-nucleus collisions becomes larger within the SPS energy range and then stays constant up to the RHIC domain.

The most dramatic effect, shown in the middle panel of the figure, is seen in the energy dependence of the ratio <K⁺>/<π⁺> of the mean multiplicities of K⁺ and π⁺ produced in central Pb+Pb collisions. Following a fast threshold rise, the ratio passes through a sharp maximum in the SPS range and then seems to settle to a lower plateau value at higher energies. Kaons are the lightest strange hadrons and <K⁺> count for about half of all the anti-strange quarks produced in the collisions. Thus, the relative strangeness content of the produced matter passes through a sharp maximum at the SPS in nucleus-nucleus collisions. This feature is not observed for proton-proton reactions.

A third important result is the constant value of the apparent temperature of K⁺ mesons in central Pb+Pb collisions at SPS energies, as shown in the bottom panel of the figure. The plateau at SPS energies is preceeded by a steep rise of the apparent temperature measured at the AGS and followed by a further increase indicated by the RHIC data. Very different behaviour is measured in proton-proton interactions.

So far, only the statistical model of the early stage reproduces the sharp maximum and the following plateau in the energy dependence of the <K⁺>/<π⁺> ratio. In this model, the spike reflects the decrease in the number ratio of strange to non-strange degrees of freedom and changes in their masses when deconfinement sets in. Moreover, the observed steepening of the increase in pion production is consistent with the expected excitation of the quark and gluon degrees of freedom.

Finally, in the fireball of particles created in the collision, the apparent temperature is related to the thermal motion of the particles and their collective expansion velocity. Collective expansion effects are expected to be important only in heavy-ion collisions, as they result from the pressure generated in the dense interacting matter. The stationary value of the apparent temperature of K⁺ mesons may thus indicate an approximate constancy of the early stage temperature and pressure in the SPS energy range due to the coexistence of hadronic and deconfined phases.

These results suggest the deconfinement phase transition exists in nature (and thus the quark-gluon plasma) and that in Pb+Pb collisions it begins to occur in the SPS energy range. From the composition of hadrons resulting from the decay of the fireball, the temperature at which the transition takes place can be estimated to be T ≅ 2 x 10¹² K (170 MeV), coinciding with the limiting temperature of hadrons suggested at CERN many years ago by Rolf Hagedorn.

The observation of anomalies in the energy dependence of hadron production in Pb+Pb collisions in the SPS energy range requires further study. Analysis of data taken last year continues in search of further phenomena caused by the deconfinement phase transition, such as anomalies in the event-by-event fluctuations expected in the vicinity of the second-order critical end-point (M Stephanov, K Rajagopal and E Shuryak 1999). In future, it would be interesting to extend measurements of the energy dependence to central collisions of light nuclei as well as to proton-proton and proton-nucleus interactions. Such measurements should significantly constrain models of the collision process and, in particular, help us to understand the role played by the volume of the droplet of strongly interacting matter in determining the onset of the deconfinement phase transition.