Study of sPlot Background Subtraction for Particle Identification Efficiency Determination at Belle II

Sumitted to PubDB: 2024-09-19

Category: Bachelor Thesis, Visibility: Public

Tags: -

Authors Lars von der Werth, Stefan Wallner
Date Jan. 1, 2024
Belle II Number BELLE2-UTHESIS-2024-003
Abstract Particle identification is a central experimental challenge in the precise measurement of Standard Model parameters and in the search for new physics beyond the standard model at the Belle II experiment. This also includes an accurate knowledge of the imperfections in particle identification efficiencies. This efficiency is typically determined from Monte Carlo simulations, but crosschecking and validating these simulations, requires to determine efficiencies directly from the real data. In this thesis, we present the determination of muon identification efficiencies using J/Ψ → μ+μ− decays. To remove background events, that remain after event selection, we do a background subtraction using the sPlot method. In the existing implementation of the efficiency calculation, we observe in simulated J/Ψ → μ+μ− decays, that the efficiencies computed using background subtraction disagree with the Monte Carlo truth efficiencies in the momentum range below 1.5 GeV/c. The sPlot method can be impaired by correlations between the kinematic variables of the sample. Therefore, we investigate the effect of correlations to understand and mitigate a possible bias introduced by them to the determination of particle identification efficiencies. We find out, that it is possible to improve the efficiency estimates by taking into account correlations between kinematic variables in the background subtraction. However, these improvements alone are not sufficient to reach agreement with the true efficiencies in the full momentum range. A possible explanation for the remaining discrepancies are observed imperfections of the parametrization of the probability density function used to model the invariant mass distribution of the two particle system, which is needed for the sPlot method.
Conference Garching

Files