Background estimator for jet studies in p+p and A+A collisions

Abstract

Experimentally, jet physics studies face an unavoidable task: Distinguishing, at the detector level, the particles produced in the hard partonic scattering from the ones created in unrelated soft processes such as pileup interactions in high-luminosity proton-proton scattering or the underlying event in heavy-ion collisions. The fluctuating nature of the background constitutes the main source of uncertainty for any subtraction algorithm. Aiming at mitigating the effect of such fluctuations, we present a new method to estimate the background contribution to the transverse momentum on a jet-by-jet basis. Our approach is based on estimating the median background momentum density stored above a pT-cut applied at the constituent level and an experimentally accessible correction term related to the signal contribution below the cut. This allows to trade part of the uncertainty due to background contamination for that of the signal below the cut, similarly to the SoftKiller method. We propose to reduce the fluctuations of the latter by exploiting intrinsic correlations among the soft and hard sectors generated in the branching process of QCD jets. Our data-driven approach is tested against pythia8 and jewel dijet events embedded in minimum bias events and thermal background, respectively, and compared to the area-median and SoftKiller methods. The main result of this study is a ∼5%-35% improvement on the resolution of the reconstructed jet pT compared to previous methods in both high-luminosity proton-proton and heavy-ion collisions.