Dark matter indirect detection and phenomenology of flavored dark matter models

Dark matter remains one of the most fascinating puzzles in modern physics, with its existence being well established through a variety of astrophysical observations. However, despite extensive experimental and theoretical efforts, the properties of dark matter remain largely unknown. In this thesis, we contribute to the search for dark matter by studying two different strategies: First, we study indirect detection with cosmic ray antideuterons which can provide insights into the interactions of dark matter with Standard Model particles. Second, we develop dark matter models to help explain the observational evidence and study specifically the phenomenology of flavored dark matter models. In the first part, we investigate the indirect detection of dark matter with cosmic ray antideuterons which are expected to have a much lower background compared to other antiparticles, making them a promising target for dark matter searches. We compute their injection spectrum using Monte Carlo simulations, taking into account the uncertainties that arise from the modeling of the production of antideuterons. We simulate the propagation of antideuterons through the Galaxy using two different propagation models and include the uncertainties of the propagation parameters. The computed antideuteron fluxes are used to develop and train a neural network emulator to predict the antideuteron flux with a significant speed-up compared to simulations. Using the emulator, we investigate the sensitivity of current and future experiments to antideuterons produced in dark matter annihilations. We find that our tool is accurate and efficient and can be used for the interpretation of future antideuteron data in the context of dark matter. In the second part of the thesis we focus on the phenomenology of flavored dark matter models. These models offer a rich phenomenology that can evade current constraints and allow for a more dynamic freeze-out phenomenology. First, we study a flavored Majorana dark matter model with couplings to up-type quarks and focus on more general freeze-out scenarios. We perform detailed scans over the parameter space to obtain viable points that are not excluded by current constraints and find that including coannihilation and conversion processes significantly increases the viable parameter space of the model. The parameter space is then confronted with constraints from a variety of LHC searches, and we discuss possible signatures to distinguish the Majorana dark matter model from the Dirac case. We find that the charge asymmetry in the production of a top quark and a jet is a promising signature. Lastly, we expand the study to include a variety of flavored dark matter models and develop a framework to perform an automated study of the phenomenology of these models. This framework includes all relevant constraints from the relic abundance, direct and indirect detection experiments as well as flavor and collider constraints and allows for a systematic study of the viable parameter space of flavored dark matter models. To demonstrate the capabilities of the framework, we perform a study of a model with couplings to right-handed leptons. We conclude that the phenomenology of the model can be studied very efficiently with this framework and find that it is mostly constrained by the relic abundance and flavor observables.