What to expect when you are expecting new physics: searches for new phenomena in multilepton final states with the ATLAS detector

This dissertation describes my work on three diverse but related aspects of the ATLAS experiment at CERN. The central idea of this thesis is structured around searches for new phenomena in multilepton final states using data from proton-proton collisions at a center-of-mass energy of 13 TeV recorded by the ATLAS detector at the Large Hadron Collider during Run 2. Three distinct models, covering final states with multiple light leptons, tau leptons, jets and missing transverse energy are considered. The first analysis presented is the first-ever search for flavour violating heavy Higgs bosons decaying to multilepton plus multi b-jets final states, motivated primarily by the ttW -like and tttt-like tensions in data reported by the ATLAS and CMS experiments. The targeted signals lead to final states with either a same-sign top-quark pair, three top-quarks, or four top-quarks. A broad excess is observed over the considered parameter space of mass and couplings, with the largest local significance being 2.8 standard deviations. The second analysis is, yet again, the first-ever direct LHC search for a doublet or singlet vector-like lepton (VLL) coupling to first and second generation Standard Model leptons, motivated by the anomalous muon magnetic moment and flavor anomalies. No significant excess is observed over the SM expectation value for any of the models under study, and to-date, the strictest limits are placed on the production of VLLs from these models. Finally, the third analysis covered as part of this dissertation presents preliminary studies for the search for VLLs preferentially coupled to third generation SM fermions, in final states with multiple light leptons, tau-leptons, b-jets and missing transverse energy. The preliminary sensitivity to the model is estimated from the these channels and is competitive with that from tau-only channels at low VLL masses, which exclude VLLs up to 1 TeV. While no significant excess was seen in either of the searches, they serve as an attempt to add another piece to the jigsaw of finding a suitable extension of the SM, and a playground to build and test cutting-edge data analysis techniques for the next generation of analyses. My work on searching for BSM physics is complemented by the other two topics covered in this dissertation, on studying the TileCal light yield degradation with the data obtained by the integrator read-out of the ATLAS TileCal during Run 2 and deriving TileCal-related uncertainties for the ATLAS luminosity measurement in Run 3. The former study not only provides valuable feedback to the TileCal Phase-II upgrade plans, but is also essential for re-processing physics data and indirectly affects the determination of jet and missing transverse energy related uncertainties. The study estimated the effect of damage and ageing of the scintillators and wavelength shifting fibres in TileCal cells to be between 10% to 30% depending on the location of the cell from the interaction point. The latter measurement affects counting of the measured number of events of a certain theoretical process, is used by every ATLAS measurement using the dataset and is a limiting factor in precision analyses measuring physics cross-sections. The overall uncertainty in the luminosity determination in 2022 was estimated to be 2.19% (1.91% from TileCal) and in 2023 was estimated to be 2.04% (1.2% from TileCal).