Mysterious long-range correlations seen in pPb collisions

The CMS collaboration has published its first result on proton–lead (pPb) collisions (CMS collaboration 2012), related to the observation of a phenomenon that was seen first in nucleus–nucleus collisions but also detected by CMS in 2010 in the first LHC proton–proton (pp) collisions at a centre-of-mass energy of 7 TeV (V Khachatryan et al. CMS collaboration 2010). The effect is a correlation between pairs of particles formed in high-multiplicity collisions – that is, collisions producing a high number of particles – which manifests as a ridge-like structure.

About once in every 100,000 pp collisions with the highest produced particle multiplicity, CMS observed an enhancement of particle pairs with small relative azimuthal angle Δφ (figure 1a). Such correlations had not been observed before in pp collisions but they were reminiscent of effects seen in nucleus–nucleus collisions first at Brookhaven’s Relativistic Heavy-Ion Collider (RHIC) and later in collisions of lead–lead nuclei (PbPb) at the LHC (figure 1b shows peripheral PbPb collisions from CMS).

Nucleus–nucleus collisions produce a hot, dense medium similar to the quark–gluon plasma (QGP) thought to have existed in the first microseconds after the Big Bang. The long-range correlations in PbPb collisions are interpreted as a result of a hydrodynamic expansion of this medium and are used to determine its fluid properties. Remarkably, this matter is found to have low frictional resistance (shear viscosity/entropy density ratio), behaving as a (nearly) perfect liquid. Because a QGP medium was not expected in the small pp system, the CMS results led to a large variety of theoretical models, which attempted to explain the origin of these ridge-like correlations (Wei Li 2012).

In September 2012, the LHC provided a short pilot run of pPb collisions at a centre-of-mass energy of 5 TeV per nucleus, for just a few hours. CMS collected two million pPb collisions (figure 2) – and now the first correlation analysis of these data has revealed strong long-range correlations, most easily visible as the ridge-like structure highlighted in figure 1c. As was the case for the pp data, the most common simulations of pPb collisions do not show ridge-like correlations, thus indicating a new, still unexplained phenomenon. Surprisingly, the effect in pPb collisions is much stronger than in pp collisions. In fact, it is similar to that seen in PbPb collisions.

The 2013 pPb run should yield at least a 30,000-fold increase in the pPb data sample at the same collision energy. Combined with the surprisingly large magnitude of the observed correlations, this will enable detailed studies and open a new testing ground for basic questions in the physics of strongly interacting systems and the nature of the initial state of nuclear collisions.