Towards High Precision Measurements of Higgs Boson Properties in the Di-Tau Decay with the ATLAS Detector

Abstract

The Standard Model of particle physics (SM) describes the known fundamental particles and their interactions. The model is found to be incomplete with a need for extensions in order to cover all current experimental observations. The Higgs boson is associated to the Higgs field, which is responsible for the acquisition of the elementary particle masses. With its central role in the SM, the Higgs and precision measurements of its properties provide anchor points for possible model extensions. <br /> The Higgs decay into two &tau; leptons is well-suited to explore the boson's properties. The distinct decay signature and high branching ratio of the decay channel offer strong sensitivity in the measured observables and provide prominent access to the Yukawa coupling of leptons to the Higgs boson. The differential exploration of the available phase-space promises hints for unknown physics phenomena. The selection of phase-space regions is based on the Simplified Template Cross-Section (STXS) approach, which is optimized to reduce theoretical uncertainties while maximizing sensitivity for physics beyond the SM. <br /> The H&rarr;&tau;&tau; measurement employs a binned-profile likelihood fit and is based on the ATLAS Run 2 dataset recorded at the Large Hadron Collider (LHC). The most precise cross-section measurement in the Vector-Boson-Fusion (VBF) production mode to date yields a value of 0.93^+0.17_-0.15 relative to the SM prediction. A selection of 18 phase-space regions are measured in the STXS analysis approach. Eight VBF-related regions comprise the first measurements of the high p^H_T>200 GeV and high m_jj>1.5 TeV phase-space region (1.29^+0.39_-0.34), as well as the most precise exploration of the low p^H_T<200 GeV region in combination with a high m_jj>1.5 TeV selection (0.12^+0.34_-0.33). <br /> Precision in SM measurements necessitates the validation and optimization of advanced analysis strategies within a collaborative effort. Towards an improved reconstruction of the second largest background in the analysis, the misidentified &tau; leptons (fakes), a decay-mode dependent fake estimation is investigated. Apart from a more detailed description of the background, the new approach enables the definition of phase-space regions with increased signal over background ratio. To improve sensitivity by reduced uncertainties, a combination of fake-templates across phase-space regions is implemented. The combination of templates reduces statistical uncertainties on the shape and adds to the stability of the fit setup. <br /> In the upcoming high-luminosity era of the LHC, reliable and computationally efficient reconstruction is a core-requirement. The optimization of a Recurrent Neural Network for electron particle identification bridges the current reconstruction algorithms to the future of object reconstruction in high-energy physics.