Thèse de doctorat
Résumé : The nature of dark matter (DM) is one of the most important open questions in modern physics. Despite decades of experimental and theoretical investigations, its fundamental composition is still unknown. Among the many proposed candidates, primordial black holes (PBHs) constitute a particularly intriguing possibility. While these hypothetical black holes from the early Universe have been excluded from constituting the entirety of the DM over most mass ranges, there remains a window in the asteroid-mass regime, spanning approximately ~10^17 g to ~10^23 g, in which they could still account for all of it. In this thesis, we develop a new approach to constrain primordial black holes in this window, based on their capture by main-sequence stars and the subsequent destruction of those stars, with the aim of definitively excluding (or, with some luck, discovering) this dark matter candidate.The first chapter is dedicated to the study of PBH capture by stars. We consider two mechanisms: capture via adiabatic contraction during star formation, and direct capture, in which initially unbound PBHs lose energy as they pass through a star and become bound. We then investigate how captured PBHs continue to lose energy and eventually settle inside stellar objects, primarily due to dynamical friction. The role of external gravitational perturbers in this process is also discussed. Applying these results to main-sequence stars, we find that the probability of PBH capture can approach unity in DM–dominated environments with low velocity dispersions, such as ultra-faint dwarf galaxies.In the second chapter of this thesis, we study the growth of black holes trapped inside main-sequence stars. Starting from the Bondi accretion framework, we examine the microphysics of the plasma, including nuclear energy generation and energy transport within the accretion flow. We also discuss the role of stellar rotation. Importantly, we consider scenarios in which the black holes orbit within the star rather than remaining fixed at its center. We find that the Bondi solution provides an accurate description for black holes with masses between ~10^21 g and ~10^29 g. We conclude that black holes with masses > 10^21 g will either accrete their host stars on timescales much shorter than the age of the Universe or produce sufficient luminosity to leave observable signatures. The case of lighter black holes requires incorporating energy diffusion and convection within the accretion flow, and is left for future work.In the third and final chapter, we assume that all asteroid-mass PBHs destroy their host stars in a finite time. Based on the observation that more massive stars are more likely to capture PBHs, and are therefore more likely to be destroyed, we investigate how this process modifies the stellar mass function of ultra-faint dwarf galaxies, making it more bottom-heavy. Specifically, we use Hubble Space Telescope data from three such galaxies and compare their color–magnitude diagrams with model stellar populations that account for stellar destruction by PBHs. From this comparative analysis, and given the absence of the predicted signature, we exclude PBHs with masses around ~10^19 g from constituting more than 78% of the DM at the 3 sigma level. Future dedicated work on PBH capture in binary systems will allow these constraints to be extended to higher masses.This thesis lays the groundwork for a new approach to constraining primordial black holes. Future theoretical and observational developments will further refine this method and ultimately help close the asteroid-mass window.