A new algorithm for the identification of boosted Z → e+e- decays for heavy resonance searches with the ATLAS detector at the LHC

The identification of W, Z, and Higgs bosons with large transverse momenta is crucial in many searches for new heavy resonances. Thus far, the development of algorithms for the tagging of boosted bosons focuses on the reconstruction and identification of hadronic boson decays, while no dedicated algorithm to identify boosted Z → e+e- decays exists. The performance of the standard electron reconstruction and identification algorithms degrades with increasing transverse momentum of the e+e- pairs. This thesis describes the development and performance of a new approach for reconstructing and identifying highly boosted Z → e+e- decays with the ATLAS detector. A Z → e+e- candidate decay is reconstructed via a single jet, clustered via the anti-kt algorithm using a radius parameter of 0.4. For the identification of the Z → e+e- decay, a deep neural network, trained on the jet properties including inner detector and calorimeter information, is used. Signal efficiencies and background rejection rates are evaluated as a function of the jet kinematics. Furthermore, the mass and transverse momentum response scales and resolutions are studied. Finally, the Z → e+e- identification and reconstruction approach is tested in the search for a Z’ boson based on Monte Carlo simulations of the data taken with the ATLAS detector during the LHC Run 2. Expected exclusion limits on the production cross section times branching ratio at 95% confidence level are determined.