Correcting correlation function measurements

Abstract

Correlation functions measured as a function of $\ensuremath{\Delta}\ensuremath{\eta},\ensuremath{\Delta}\ensuremath{\phi}$ have emerged as a powerful tool to study the dynamics of particle production in nuclear collisions at high energy. They are however subject, like any other observables, to instrumental effects which must be properly accounted for to extract meaningful physics results. We compare the merits of several techniques used towards measurement of these correlation functions in nuclear collisions. We discuss and distinguish the effects of finite acceptance, and detection efficiency that may vary with collision parameters such as the position of the event in the detector and the instantaneous luminosity of the beam. We focus in particular on instrumental effects which break the factorization of the particle pair detection efficiency, and describe a technique to recover the robustness of correlation observables. We finally introduce a multidimensional weight method to correct for efficiencies that vary simultaneously with particle pseudo rapidity, azimuthal angle, transverse momentum, and the collision vertex position. The method can be generalized to account for any number of ``event variables'' that may break the factorability of the pair efficiency.