Primordial Nucleosynthesis in the Presence of MeV-scale Dark Sectors

In this thesis, we perform a comprehensive study of Big Bang nucleosynthesis constraints on different dark-sector models with $\mathrm{MeV}$-scale particles which are neither fully relativistic nor fully non-relativistic during all relevant temperatures. To this end, we derive a generic set of equations that can be used to determine the light-element abundances for many different dark-sector scenarios. In particular, we take into account all relevant effects that might alter the creation of light elements in the early universe, including modifications to the Hubble rate and time-temperature relation, an adjusted best-fit value for the baryon-to-photon ratio due to an altered effective number of neutrinos, a modified neutrino-decoupling temperature as well as late-time modifications of the nuclear abundances due to photodisintegration. We then solve these equations for the case of dark sectors with $\mathrm{MeV}$-scale particles that are created in the early universe and later decay into either dark or electromagnetic radiation. In the former case, the dark and the visible sectors are completely decoupled, and we show that even such scenarios can be severely constrained if the initial temperature ratio of both sectors is sufficiently large. In the latter case, the particle decay can instead lead to a severe entropy injection into the SM heat bath, and we show that the final constraints can be very different from the naive order-of-magnitude estimates that are usually employed for such scenarios. We then turn to the case of $\mathrm{MeV}$-scale dark matter that can annihilate into all kinematically available SM states, including electrons/positrons, photons, and neutrinos. In this context, we first update the lower bound on the mass of thermal dark matter using improved determinations of the nuclear abundances. Afterwards, we calculate the corresponding constraints on the annihilation cross-section of dark matter and show that, for p-wave suppressed annihilations, the bounds from nucleosynthesis are much stronger than the ones from the CMB and even competitive with the strongest bounds from other indirect searches. Finally, we apply our results to models with self-interacting dark matter as well as to models with axion-like particles. In the former case, we show that for scalar mediators, most parts of parameter space leading to sizable self-interactions are already excluded by a combination of direct-detection experiments and constraints from nucleosynthesis. In the latter case, we further evaluate the robustness of our constraints by allowing various additional effects that may weaken the bounds of the standard scenario. We find that, while the bounds can indeed be weakened, very relevant robust constraints remain