Back to search

University of British Columbia

Search for long-lived particles decaying into displaced hadronic jets in the ATLAS Calorimeter using novel machine learning algorithms

Abstract

dc:description

The Standard Model (SM) has been hugely successful at explaining our natural world at the smallest scales, including the fundamental particles and forces. However, experimental evidence, such as dark matter and baryon asymmetry point to the SM as being an incomplete model. One such candidate to extend the SM are hidden sectors, in which long-lived particles could be the missing link between the SM and a group of hidden particles. Firstly, work on performance expectations of the upgrade to the ATLAS Inner Detector to enable data taking for the next decade will be presented, showing increases in reconstruction efficiency at high momentum and particle density compared to the current detector. Then, a search for hidden sector particles in 2016 ATLAS data, totalling 33.0 fb⁻¹ from LHC proton-proton collisions at center-of-mass energy of 13 TeV, will be discussed. This search focuses on long-lived particles decaying back to SM particles in the ATLAS calorimeters. No significant excess was found, and limits were set on cross section times branching fraction as a function of proper decay length. For the 125 GeV mediator, a few cm to a few m are excluded, assuming a branching fraction of 10%. For higher mass mediators, up to 1 TeV, cross section times branching fraction of 0.1 pb are typically excluded between a few cm to a few m. Finally, a search for these same long-lived hidden sector particles was also performed for the full Run 2 ATLAS dataset, totalling 139.0 fb⁻¹. Novel machine learning techniques were used, such as an adversarial neural network which greatly reduced the impact of simulation mis-modelling. No significant excess was found, and for a 125 GeV Higgs Boson mediator, assuming a 10% branching ratio to long-lived particles, proper decay lengths between about 1 cm and a few tens of meters are excluded, improving 2016 results by about an order of magnitude. For higher mass mediators, cross section times branching fraction of 0.1 pb can be excluded for a proper decay length between 1 cm to a few tens of meters, improving 2016 limits by a factor of 2-3 depending on the model.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy - PhD
Level thesis:degree_level
doctoral
Discipline thesis:degree_discipline
Physics
Grantor dc:publisher
University of British Columbia
Year dc:date
2022

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Cormier, Félix
Contributors dc:contributor
  • Lister, Alison

Rights

dc:rights
Statement dc:rights
  • Attribution-NonCommercial-NoDerivatives 4.0 International
Language dc:language
eng

Identifiers

dc:identifier.*
Handle dc:identifier
http://hdl.handle.net/2429/81501

Chain of custody

source
Harvested from
University of British Columbia
Base URL
circle.library.ubc.ca/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Cormier, Félix. Search for long-lived particles decaying into displaced hadronic jets in the ATLAS Calorimeter using novel machine learning algorithms. doctoral thesis, University of British Columbia, 2022. http://hdl.handle.net/2429/81501