We have investigated the role of non-flow correlations in the elliptic flow observable. Imagine an off-central collision of two heavy ions at very high energies. In the plane transverse to the beam axis the nuclear overlap has a shape of an almond. If a quark-gluon plasma is formed in this overlap region then, due to the azimuthal asymmetry of the almond, the pressure gradients would be different in different transverse directions creating an azimuthally asymmetric distribution of produced particles. This effect, known as elliptic flow, would be a signature of plasma formation. However, to measure the effect experimentally one needs to either find the reaction plane direction or study two-particle azimuthal correlations. (The two methods are identical for a large number of particles.) We [57] have demonstrated that usual two-particle correlations generated by the particle production mechanism in the early stages of the collision heavily pollute the observable preventing separation of the "real" elliptic flow from these correlation effects. Our model of two-particle correlations successfully describes the elliptic flow data while the hydrodynamical models based on QGP scenario fail to do so.
In our second paper on the subject [58] we have shown how these non-flow correlations pollute the observables in all three major ways of measuring elliptic flow. Our model can be tested by studying proton-proton collisions (pp). There creation of QGP is unlikely and the observed correlations are entirely non-flow. The prediction of our model is that in proton-proton collisions non-flow correlations are even larger than in AA giving elliptic flow observable v2 about 6 times larger than in AA. The prediction has been confirmed by the analysis of the preliminary pp data at RHIC by PHENIX and STAR experimental groups.
I presented the above works at Quark Matter 2002 conference in Nantes, France and contributed to conference proceedings. The above mentioned papers together with the conference presentation generated a lot of interest and many discussions. They have also inspired experimental groups to re-analyze their data on elliptic flow in light of our results.