Searches for Higgs boson pair production in the bb-bb- final state with the ATLAS detector

This thesis presents searches for Higgs boson pair production in the $b\bar{b}b\bar{b}$ final state using data collected by the ATLAS detector from proton-proton collisions at $\sqrt{s}=13$ TeV during the entirety of LHC Run 2 and at $\sqrt{s}=13.6$ TeV over a portion of Run 3. Three distinct analyses are detailed. The first analysis presents a search for scalar resonances in the asymmetric production of two scalar bosons using the full ATLAS Run 2 dataset. This Beyond-the-Standard-Model (BSM) scenario posits the production of a heavy Higgs-like resonance $X$ via gluon-gluon fusion, which then decays into a Standard Model Higgs boson ($H$) and a lighter scalar ($S$). This study, representing the first ATLAS search in this decay topology, sets upper limits on the cross section for resonant scenarios with $m_X$ between 200~GeV and 3~TeV and $m_S$ from 70~GeV to 2.5~TeV. No significant excesses were observed. The resulting limits are competitive with a corresponding CMS search, which reported its highest excess with a local significance of 4.1$\sigma$ (global 2.8$\sigma$) at $(m_{X},m_{S})=(700,400)$~GeV. No such excess was found in this ATLAS search at a similar mass hypothesis of $(m_{X},m_{S})=(750,400)$~GeV. The second analysis targets non-resonant Higgs boson pair production, probing both Standard Model $HH$ production and potential deviations in the Higgs trilinear self-coupling using the Run 2 data. No significant evidence for Standard Model $HH$ production is observed, but stringent upper limits on the cross section for Higgs boson pair production via gluon-gluon fusion and vector-boson fusion are established at 5.4 (8.1) times the Standard Model prediction for observed (expected) yields. These results mark a substantial improvement in sensitivity relative to earlier Run 2 searches, positioning the $b\bar{b}b\bar{b}$ final state as the third most sensitive channel in the ATLAS Run 2 $HH$ combination. Furthermore, cross section limits are set across a range of Higgs self-coupling values, parameterized by $\kappa_{\lambda} = \lambda_{HHH}/\lambda^{SM}_{HHH}$, with the observed interval $-3.9 \leq \kappa_{\lambda} \leq 11.1$ ($-4.6 \leq \kappa_{\lambda} \leq 10.8$ expected). The third analysis addresses non-resonant $HH$ production through gluon-gluon fusion in the resolved topology, incorporating both advanced reconstruction and analytical techniques for the Run 2 dataset and extending to include partial Run 3 data from 2022 and 2023. While this study is ongoing, with further optimization and systematic studies underway, the core analysis workflow is established. For the most sensitive discriminant, the expected statistics-only upper limit on the $HH$ production cross section is 2.80 times the SM prediction, reflecting a notable gain in sensitivity compared to the Run 2 stat-only result. This corresponds to an 18\% improvement beyond what would be anticipated from the increase in luminosity alone. Cross section limits are again set as a function of $\kappa_{\lambda}$, which is constrained to $-2.86 \leq \kappa_{\lambda} \leq 10.02$ in the expected, stat-only scenario. Finally, this thesis also describes new methods and results for deriving dedicated jet calibrations and systematic uncertainties in the context of ATLAS fast calorimeter simulation.