The role of multimessenger astronomy in constraining supernova explosion mechanisms

Supernova explosions are undoubtedly one of the most energetic phenomena in the Universe and one of the long-standing riddles of stellar astrophysics. Their astonishing brightness has captivated human attention for thousands of years. Although supernovae are rare events, their detection provides invaluable insights into high-energy processes in the cosmos. The most recent groundbreaking detection was SN1987A. It was the last supernova that was so near and bright to be visible by naked eye. The observation of a neutrino flux from SN1987A confirmed the theoretical predictions and, most importantly, marked the first time neutrinos were used for astronomical observations. This peculiar observation is commonly considered the dawn of the so-called multimessenger era. Multimessenger astronomy represents a fundamental paradigm shift. The traditional approach of observing the sky solely through visible light has dramatically evolved in the last decades. Today, astronomers leverage information across the entire electromagnetic spectrum and benefit from the detection of other cosmic messengers, namely cosmic rays, neutrinos, and gravitational waves. This multidisciplinary approach has enriched our understanding of the Universe and emphasised the need for efficient collaboration among various experiments. The current era of multimessenger astronomy has established a vast, well-organised network of telescopes capable of communicating and collaborating. The next decade promises many upgrades on current facilities and the construction of new generation detectors that will allow us to observe the Universe with an unprecedented precision. In this context, supernovae serve as crucial laboratories for studying nuclear matter under extreme conditions. Neutrinos and gravitational waves provide unique insights into the mechanisms driving these explosive events, helping to answer longstanding questions about the processes involved in shock revival and explosion initiation. After a detailed introduction about the state-of-art of supernova models and potentiality of multimessenger astronomy, the thesis describes my contribution within the KM3NeT experiment, a next generation neutrino telescope. I present my contribution on the development of a machine learning tool, trained to distinguish between neutrinos and atmospheric muons, with the aim of improving the background rejection capability. Thanks to the performance achieved, which are shown in this thesis, this tool is already used in the online event processing and it is planned to be applied to all current analyses. Indeed, most part of my work inside the collaboration was devoted to the development of the so-called Real-Time Analysis framework, a platform that takes care of the online processing of incoming events, thus providing all the valuable physical information about potential neutrino events in a very short amount of time. The rapidity of the online event processing indeed is a crucial aspect for a real-time platform, especially in the context of supernovae, where the detection of a neutrino can be used to trigger dedicated analyses that searches for a counterpart emissions, $e.g.$ gravitational wave counterpart, inside the interested time window. In line with this scenario, I developed a pipeline called MUSE (or Multi-messenger Understanding of Supernova Explosions). MUSE is based on machine learning techniques and it is trained to perform targeted searches for supernova signals within gravitational wave data. The model was trained on a set of phenomenological waveforms designed to mimic features observed in modern 3D simulations and then tested on simulation outcomes in the context of Einstein Telescope, a third generation gravitational wave detector. Indeed, thanks to the impressive growth of the available computing power, current numerical simulations are at the stage of accurately predicting the gravitational wave emission from a stellar explosion. However, despite not being conclusive, numerical simulations have enhanced our knowledge about the insight mechanisms driving such phenomena. Specifically, multi-dimensional simulations, $i.e.$ 2D and 3D, have shown that, in many cases, the explosion is supported by non-radial instabilities developing between the shock front and the proto-neutron star surface. One of the most notable effect is the Standing Accretion Shock Instability (or SASI), a large-scale non-radial instability. Interestingly, this effect induces a clear modulation in both neutrino and gravitational wave emissions, thus establishing a direct connection between these two cosmic messengers. Indeed, the study of this correlation will lead be a breakthrough in astrophysics, that will allow us to unveil the inner mechanisms triggering one of the most powerful phenomena in the Universe. In my research, I have investigated the possibility of adopting the Hilbert-Huang transform to identify and extract SASI signatures from simulated gravitational wave signals.