Studies of photon isolation in a large-pileup environment and search for axion-like particles decaying into two photons with the ATLAS detector at the LHC

This thesis presents my contributions to the ATLAS experiment, consisting mainly on two topics: studies of photon isolation, and a search for new resonances in the diphoton decay channel. The work described is based on proton-proton collision data from the Large Hadron Collider at a center-of-mass energy of 13 TeV recorded with the ATLAS detector from 2015 to 2018, corresponding to an integrated luminosity of 139 $^{-1}$. Prompt photons at hadron colliders are an important probe of the theory of Quantum Chromodynamics and a powerful signature in searches for new physics. Both objectives require a deep understanding of photon production processes and possible background sources. Photon candidates are reconstructed and identified using information from ATLAS calorimetric and tracking systems. For a precise prompt photon identification, the isolation of photons from nearby energy deposits is required. Measurements of photon isolation efficiency are presented for photons with transverse energies ($E^{\gamma}_{T}$) in the range from 20 GeV up to 1.5 TeV in different pseudorapidity ($\eta^{\gamma}$) regions. The methodology presented requires the subtraction of background events, primarily from neutral mesons in hadronic jets, estimated using data-driven techniques. Photon isolation performances are heavily affected by the increase of pileup noise from additional proton-proton collisions, since energy from uncorrelated processes may be found around the photon candidate. A prospective study on how to improve photon isolation against increasing pileup using additional information from energy deposits in the calorimeters is presented. The implementation of this study will allow, in future data-taking periods, to improve prompt photon identification at various steps in photon reconstruction. The second part of this thesis presents a search for new physics with two photons in the final state. Light and weakly interacting particles are predicted in a variety of theoretical models which would appear as resonances in the low mass part of the spectrum. Phenomenologically, such resonances gather interest as possible Dark Matter mediators. Previous to this work, the ATLAS and CMS experiments have searched for diphoton resonances at masses below the Higgs boson mass down to 65 GeV with no observed evidence for new physics. The analysis described in this thesis uses a novel approach, exploiting the particular topology of events with collimated pairs of photons reconstructed in the detector. This topology allows to reach invariant masses below 65 GeV, a region poorly explored in hadron colliders due to the experimental challenge it poses for recording low energy photons. The expected limits provide the strongest bound on couplings of axion-like-particles to gluons and photons for masses below 65 GeV. My contributions to this analysis span over almost all the steps of the study, being responsible for the generation and validation of signal and background simulation samples, analysis strategy, background estimation, evaluation of systematic uncertainties and computation of the expected limit. I also contributed to the validation of the next generation of diphoton triggers. Future data-taking periods will benefit from novel energy thresholds allowing to further extend the search towards lower masses. A prospective study is presented with the expected improvement in analyses targetting low invariant diphoton masses.