Search for the lepton flavour violating decays $\tau \to e\ell\ell'$ and tracking studies at the Belle II experiment

Sumitted to PubDB: 2025-05-12

Category: Phd Thesis

Tags:

Principal Authors Justine Serrano, Arthur Thaller
Date 2025-05-12
Belle II Number BELLE2-PTHESIS-2025-015
Abstract This thesis reports the search for the Lepton Flavour Violating (LFV) decays $\tau \to e \ell \ell^{'} $ ($\ell^{(')} = e,\mu$) using the Run 1 dataset of the Belle II experiment, corresponding to an integrated luminosity $\mathcal{L} = 424 $ fb$^{-1}$. The analysis is based on an untagged reconstruction of the signal candidates in $e^{+}e^{-} \to \tau^{+}\tau^{-}$ events, where only the signal $\tau$ is reconstructed while the other is left unconstrained. The background suppression is performed on data sidebands due to the important contribution of QED processes ($e^{+}e^{-} \to \ell^+\ell^-\ell^{+'}\ell^{-'}$, $\ell^{(')} = e,\mu$) that are mis-modeled in the simulations. The analysis relies on a combination of selection criteria and a BDT classifier to achieve an almost background-free scenario, while maintaining a signal reconstruction efficiency in the range [15-24]%. The branching fraction is extracted with an unbinned maximum likelihood fit to the $M_\_{e\ell\ell'}$ distribution, and the upper limits are computed using the $CL_S$ method. The systematic uncertainties affecting the signal efficiency and the background modelling are evaluated. No signal events are observed across the five modes, and the observed upper limits in four of the modes ($e^+e^-e^+$, $e^+e^-\mu^+$, $\mu^+\mu^- e^+$ and $\mu^+ e^-\mu^+$ ) are more stringent than the previous Belle's results obtained with 762 fb$^{-1}$ while the fifth mode ($e^+\mu^-e^+$) limit is comparable to Belle's measurement. In addition, three studies of the fake tracks rate are included. The first one implements a track time variable based on measurement from the silicon vertex detector (SVD), the second one uses a classifier to reject detector hits based on the time information in degraded data acquisition conditions and the third one studies the fake tracks rate in the proposed upgrade of the vertex detector (VTX).
Institute CPPM

Files