A la poursuite du Fond Diffus de Neutrinos de Supernovae dans les expériences Super-Kamiokande et Hyper-Kamiokande

Core-collapse supernovae (CCSNe) are cataclysmic explosions at the end of massive stars’ lives that occur roughly once per second in the observable universe, emitting over 99% of their energy as neutrinos. Such relic neutrinos from all past CCSNe form the Diffuse Supernova Neutrino Background (DSNB), a faint whisper of the cosmic past – one that has yet to be observed experimentally. The Super-Kamiokande (Super-K) water Cherenkov detector in Japan currently leads the world in sensitivity to a DSNB detection with inverse beta decay (IBD) as its target detection channel. This has been enhanced since the 2020s by the addition of gadolinium (Gd) to better identify neutrons.This thesis takes hold of key tasks in the DSNB search, including the reduction of dominant backgrounds coming from atmospheric neutrino interactions, the development of a new technique to distinguish neutron captures from muon and pion decays, and the validation of a Machine Learning tool for neutron identification. To search for the DSNB signal, two parallel statistical approaches—a “binned” and “spectral” analysis—are employed. Particular care was taken with the data unblinding procedure, background modeling, event selection optimization, and the evaluation of systematic uncertainties affecting both the DSNB signal and its backgrounds. This led to results in this thesis constituting the world’s tightest limits on the DSNB to date with the binned analysis. Additionally, the first direct comparisons of results from both statistical approaches in Super-K are presented and the first model-independent DSNB sensitivity study for the next-generation Hyper-Kamiokande (Hyper-K) experiment is introduced. Finally, these projections for Hyper-K are compared to those of a possible extension of Super-K beyond 2029 and to those of the Jiangmen Underground Neutrino Observatory (JUNO) experiment, contextualizing the immediate future of the DSNB search across the globe.