Searching for signs of hidden sectors at the LHC: long-lived particle methods and models

Mysteries beyond the Standard Model (BSM) inspire theories, experiments and analysis tools that cover gaps in parameter space for new particles. Open questions ask the origins of neutrino masses, the mechanisms underlying flavour mass hierarchies, the matter-antimatter asymmetry and signs of dark matter. These questions are solved in theories of new long-lived particles (LLPs) in hidden sectors, but in practice the current search program may miss key data, especially for MeV-GeV range masses of transverse-boosted LLPs. The high-luminosity Large Hadron Collider (HL-LHC) will produce more data than any previous LHC run with a renewed focus on rare LLP signatures. This thesis develops three techniques to extend reach during the HL-LHC era: one, building a Z’ model that predicts testable signatures; two, producing sensitivity limits of the proposed dedicated LLP detector ANUBIS to Heavy Neutral Leptons (HNLs), and three, creating a graph-based strategy for detecting anomalous signals. The U(1) extension with Z’ addresses the flavour puzzle and anomalies by interpolating between lepton flavour coupling scenarios, but also predicts HNL signatures. Similar signatures are simulated for acceptance calculations with ANUBIS, revealing the mass and coupling ranges where the experiment will reach for electron- and muon-coupled scenarios. The analysis tool is a proof-of-principle using LHC dataset-wide Graph Neural Networks (GNNs) to enhance signal-background separation in leptoquark simulations. The thesis identifies cross-talk between theory and experiment through these three approaches to BSM LLP discovery.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2025

Author dc:creator

• Mullin, Anna

Advisor dc:contributor.advisor

• Brandt, Oleg

• Beyond Standard Model physics
• Long-lived particle
• Hidden sector
• Graph neural network
• Heavy neutral lepton
• Effective field theory

Licence

• https://www.repository.cam.ac.uk/bitstreams/2d05fe3c-fb61-477a-9b88-ee540751e920/download
• http://purl.org/NET/rdflicense/allrightsreserved

Language dc:language

eng

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.130903

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/404209