Search for new physics in the scalar sector using advanced τ lepton reconstruction and identification techniques in the ATLAS experiment
Loading...
DOI
Date
2025
Authors
Juzek, Monika
Journal Title
Journal ISSN
Volume Title
Thesis supervisors
Publisher
The Henryk Niewodniczański Institute of Nuclear Physics Polish Academy of Sciences
License Type
Creative commons license
Abstract
Model Standardowy (SM) fizyki cząstek elementarnych, pomimo licznych potwierdzeń doświadczalnych, a w szczególności odkrycia bozonu Higgsa w 2012 roku, pozostaje teorii niepełną. Nie wyjaśnia on takich zjawisk jak masy neutrin, ciemna materia czy asymetria między materią a antymaterią we Wszechświecie. Liczne proponowane rozszerzenia SM, takie jak np. Two-Higgs-Doublet Model (2HDM) oraz Minimalny Supersymetryczny Model Standardowy (MSSM), przewidują istnienie dodatkowych cząstek, w tym naładowanych bozonów Higgsa.
Celem niniejszej rozprawy jest poszukiwanie sygnałów nowej fizyki w sektorze skalarnym z wykorzystaniem leptonów τ w danych zderzeń proton-proton zebranych przez detektor ATLAS podczas Run-2 akceleratora LHC przy energii w środku masy √s= 13 TeV, odpowiadających łącznej scałkowanej świetlności 140 fb−1. Analiza koncentruje się na poszukiwaniach naładowanych bozonów Higgsa (H^±) w kanale rozpadu H^± → τ^± ν_τ , badając dwie sygnatury stanów końcowych: τ+dżety oraz τ+lepton (elektron lub mion), w zakresie mas 80-3000 GeV. W ramach pracy opracowano i zoptymalizowano metodę szacowania tła pochodzącego od dżetów błędnie identyfikowanych jako hadronowe rozpady τ (τ_(had-vis)), opartą na współczynnikach Fake-Factor (FF). Niepewność wyznaczenia tego tła stanowi jedno z głównych źródeł niepewności systematycznej poszukiwania H^± w całym zakresie masy. Zastosowano metodę dopasowania do kształtu (template-fit method) do uwzględnienia składu dżetów gluonowych i kwarkowych w obszarach sygnałowych. Otrzymane współczynniki FF zostały zweryfikowane w obszarach kontrolnych, uzyskując dobrą zgodność z danymi. Analiza niepewności systematycznych wykazała, że główne źródło niepewności stanowi¡ ciężkie dżety kwarkowe błędnie klasyfikowane jako τ_(had-vis). Nie stwierdzono istotnych odchyleń od przewidywań Modelu Standardowego. Ustalono górne ograniczenia na poziomie 95% przedziału ufności dla iloczynu przekroju czynnego i częstości rozpadu B〖(H〗^± → τ^± ν_τ), w zakresie od 4.51 pb do 0.4 fb, w zależności od masy H^±. Dla niskich mas (80-130 GeV) odpowiada to ograniczeniom na B(t → bH^± ) ×B〖(H〗^± → τ^± ν_τ) w zakresie 0.27%-0.04%.
Przedstawiono również wkład autora w rozwój Systemu Kontroli Detektora (DCS) dla Półprzewodnikowego Detektora śladów (SCT) oraz w prace modernizacyjne nad nowym Detektorem śladów ITk przygotowywanym na fazę wysokiej świetlności LHC (HL-LHC). Dodatkowo przeprowadzono badania obejmujące walidację polaryzacji i korelacji spinowych leptonów τ w próbkach Monte Carlo oraz opracowanie narzędzi Rivet do analizy próbek Monte Carlo z leptonami τ w stanie końcowym, wykorzystywanych w eksperymencie ATLAS. Wyniki pracy wpisuj¡ się w szeroki program badawczy eksperymentu ATLAS, obejmujący zarówno rozwój infrastruktury detektora, jak i zaawansowane analizy danych w poszukiwaniu fizyki wykraczającej poza Model Standardowy.
While the Standard Model (SM) of particle physics has been validated in many experiments, most notably with the discovery of the Higgs boson in 2012, it remains an incomplete theory. It does not account for phenomena such as neutrino masses, dark matter, and the cosmic matter–antimatter asymmetry. Various extensions of the SM have been proposed, introducing new physics processes beyond the SM (BSM), most notably the Two-Higgs-Doublet Model (2HDM) and the Minimal Supersymmetric Standard Model (MSSM), both predicting an extended scalar sector containing particles such as a charged Higgs boson. This thesis presents a search for new physics in the scalar sector using τ leptons in proton–proton collision data recorded by the ATLAS detector at the LHC. Due to their distinct decay signatures, τ leptons are powerful tools for discriminating potential BSM signals from SM backgrounds. The main goal of the thesis is the search for charged Higgs bosons (H^±) decaying via H^± → τ^± ν_τ, using proton–proton collision data at √s= 13 TeV collected with the ATLAS detector during Run-2 of the LHC, corresponding to an integrated luminosity of 140 fb−1. The analysis targets two final state signatures: τ+jets and τ+lepton (electron or muon), covering a wide mass range of 80—3000 GeV. The author's primary task was the estimation of backgrounds from jets misidentified as hadronically decaying τ leptons (τ_(had-vis)). This background estimation remains among the leading sources of systematic uncertainty across the full H^± mass range. A data-driven Fake-Factor (FF) method was developed and optimized for the new τ identification algorithm based on recurrent neural networks employed by the ATLAS experiment. A template-fit method was applied to derive combined FFs reflecting the gluon/quark jet composition in the signal regions. These FFs were validated in control regions enriched with fake τ_(had-vis) backgrounds, demonstrating good modeling. Systematic uncertainty studies identified heavy-flavour jets misidentified as τ_(had-vis) as the dominant contribution. No significant excess above the SM background was observed. Upper limits at the 95% confidence level were set on the production cross-section times branching ratio B〖(H〗^± → τ^± ν_τ), with values ranging from 4.51 pb to 0.4 fb across the H^± mass range. For low masses (80—130 GeV), this corresponds to limits on B(t → bH^± ) ×B〖(H〗^± → τ^± ν_τ) of 0.27%–0.04%. The author also contributed to the Detector Control System for the current ATLAS Semiconductor Tracker (SCT) and to the upgrade of the Inner Tracker (ITk) for the High-Luminosity LHC (HL-LHC) phase. In addition, supporting studies were conducted, including validation of τ lepton polarization and spin correlations in Monte Carlo samples, as well as development of Rivet tools for validation of MC simulations with τ leptons in final states, now used by the ATLAS Collaboration. Each component of the presented work contributes to the broader ATLAS physics programme, spanning detector operations and advanced data analysis in the ongoing quest for new physics.
While the Standard Model (SM) of particle physics has been validated in many experiments, most notably with the discovery of the Higgs boson in 2012, it remains an incomplete theory. It does not account for phenomena such as neutrino masses, dark matter, and the cosmic matter–antimatter asymmetry. Various extensions of the SM have been proposed, introducing new physics processes beyond the SM (BSM), most notably the Two-Higgs-Doublet Model (2HDM) and the Minimal Supersymmetric Standard Model (MSSM), both predicting an extended scalar sector containing particles such as a charged Higgs boson. This thesis presents a search for new physics in the scalar sector using τ leptons in proton–proton collision data recorded by the ATLAS detector at the LHC. Due to their distinct decay signatures, τ leptons are powerful tools for discriminating potential BSM signals from SM backgrounds. The main goal of the thesis is the search for charged Higgs bosons (H^±) decaying via H^± → τ^± ν_τ, using proton–proton collision data at √s= 13 TeV collected with the ATLAS detector during Run-2 of the LHC, corresponding to an integrated luminosity of 140 fb−1. The analysis targets two final state signatures: τ+jets and τ+lepton (electron or muon), covering a wide mass range of 80—3000 GeV. The author's primary task was the estimation of backgrounds from jets misidentified as hadronically decaying τ leptons (τ_(had-vis)). This background estimation remains among the leading sources of systematic uncertainty across the full H^± mass range. A data-driven Fake-Factor (FF) method was developed and optimized for the new τ identification algorithm based on recurrent neural networks employed by the ATLAS experiment. A template-fit method was applied to derive combined FFs reflecting the gluon/quark jet composition in the signal regions. These FFs were validated in control regions enriched with fake τ_(had-vis) backgrounds, demonstrating good modeling. Systematic uncertainty studies identified heavy-flavour jets misidentified as τ_(had-vis) as the dominant contribution. No significant excess above the SM background was observed. Upper limits at the 95% confidence level were set on the production cross-section times branching ratio B〖(H〗^± → τ^± ν_τ), with values ranging from 4.51 pb to 0.4 fb across the H^± mass range. For low masses (80—130 GeV), this corresponds to limits on B(t → bH^± ) ×B〖(H〗^± → τ^± ν_τ) of 0.27%–0.04%. The author also contributed to the Detector Control System for the current ATLAS Semiconductor Tracker (SCT) and to the upgrade of the Inner Tracker (ITk) for the High-Luminosity LHC (HL-LHC) phase. In addition, supporting studies were conducted, including validation of τ lepton polarization and spin correlations in Monte Carlo samples, as well as development of Rivet tools for validation of MC simulations with τ leptons in final states, now used by the ATLAS Collaboration. Each component of the presented work contributes to the broader ATLAS physics programme, spanning detector operations and advanced data analysis in the ongoing quest for new physics.
Description
Keywords
Citation
Sponsorship:
Narodowe Centrum Nauki; NAWA - Program STER
Grantnumber:
2022/47/B/ST2/03059; BPI/STE/2023/1/00027/U/00001
License Type
Attribution 4.0 International