Thomas Kretzschmar ; Allen Caldwell

2022
TUM Munich

Abstract: The Belle II experiment at SuperKEKB – an asymmetric e+e− collider – aims at a total integrated luminosity of 50 ab−1 to pursue a rich program of Standard Model and Beyond the Standard Model physics. Belle II collected about 427.79 fb−1 at the Υ (4S) resonance until the middle of 2021. This thesis presents a search for the hypothetical lepton-flavor violating process τ → ` + α(invisible), where τ is a tau particle that decays, ` is a lepton, and α is a Goldstone boson. Several models of beyond-theStandard-Model physics predict this kind of topology, including Z 0 and axion-like particle models. With an examined dataset of 63 fb−1 , this analysis significantly improves the limit compared to the previous best limit, obtained by ARGUS with an integrated luminosity of 475 pb−1 . Critical elements of the analysis are the event selection, reconstruction, and the interpretation of the overall event kinematics used to search for the two-body decay signature of the signal with an irreducible background of τ → `ντν` decays. The event selection is a one-dimensional cut-based strategy. Advanced machine learning algorithms, MLA, were evaluated but did not show substantial gains with the same input parameters; MLAs make controlling systematic uncertainties harder. A precise knowledge of the tau restframe is required to exploit the two-body kinematics of the signal decay. The analysis uses different strategies to cope with the missing ντ kinematic information, including a novel method developed explicitly for this search. This technique, referred to as generalised known kinematics, GKK, fully propagates the probability density function, pdf, of the ντ to obtain a pdf for the τ kinematics per event. Beyond the τ → `α decay, this thesis evaluates the potential of GKK to improve the τ mass measurement. Upper limits for the branching ratio of τ → `α for several masshypothesis of α, mα, are obtained with several frequentist approaches. The 95% upper limit on the ratio of branchin ratios, Br(τ → `α)/Br(τ → `ντν` ), is in the range of 10−2 to 10−3 , depending on mα ranging from 0 to 1.6GeV/c2 . A discussion of an electron identification performance study – crucial for the Belle II physics program – complements the physics analysis and its results.

Note: Presented on 03 11 2022
Note: PhD