So far, the Large Hadron Collider (LHC) has revealed no confirmed departure from the Standard Model (SM) of particle physics, yet the SM at its current status is known to be incomplete, as it lacks a candidate for dark matter, it does not account for gravitational interactions and offers no explanations for the very large energy gap between the electroweak and Planck scales.
Four-lepton proton-proton final state events are rare in SM processes and the fact that the leptons carries the full kinematic information of the event allows several invariant masses to be reconstructed from the same collision: the four-lepton invariant mass, as well as the dilepton invariant mass of each same-flavour, opposite-sign couple of leptons (electrons or muons). A new particle decaying into leptons - directly or indirectly - would result in a localized excess of data compared to the SM predictions - a resonance - in one or multiple mass distributions, or in other kinematic quantities.
The key point of the project is: how can four-leptons events recorded by ATLAS be used to search for new particles beyond the SM, and how much can Machine Learning-powered analysis techniques improve the sensitivity of such a search? This can include a resonance decaying into a pair of Z bosons, or kinematic variables that separate the signal from the SM ZZ, Z+jets and top-quark backgrounds. The Z-boson-pair resonance has already been studied for Randall-Sundrum graviton and radion excitations and for the heavy scalars of two-Higgs-doublet models, with simulated signal samples already available. As an extension, the project can also build on the group's active dilepton+MET (Z'+MET) analysis, exploring whether requiring missing transverse momentum adds sensitivity to scenarios with an invisible particle.
The analysis will use ROOT- and Python-based ATLAS tools and the ATLAS exotics group’s existing analysis framework, combined with proton-proton collision data recorded by ATLAS over the course of the last two data collection periods: Run 2 (2015-2018) and Run 3 (2022-2026).
Milestones:
M1: Acquire the theoretical background related to searches in four-lepton final states (SM limitations, new physics signal benchmark models such as heavy ZZ resonances) and get trained on the relevant analysis tools and languages: ROOT, Python, ATLAS software tools, possibly use of neural networks in particle physics
M2: Reproduce a documented four-lepton event selection using the available signal and SM simulated samples
M3: Study the four-lepton and dilepton invariant-mass distributions and other kinematic variables, in particular the decay angle of a new resonance; are we in the presence of a spin-2 excited graviton? Or is it a new spin-0 (Higgs boson-like)?
M4: Propose and implement improvements to the reference selection, evaluate the improvement’s impact on signal efficiency, background rejection and expected sensitivity to discover or exclude new physics
M5: Finalize the analysis of the full Run3 (and possibly Run 2) proton proton data at 13.6 TeV, evaluate systematics, complete the thesis
Courses relevant from the bachelor's degree:
Recommended courses for theoretical curriculum in the master's programme:
(Mandatory for the project)
(Other relevant courses)