Exploring the Galactic potential: Insights from cosmological simulations of the Milky Way

Recent advances in cosmological simulations enable a detailed investigation of the structure of Milky Way--like galaxies, which form and evolve in a self-consistent manner within these simulations. Using a publicly available sample of such models from Illustris TNG50, we conducted a comprehensive analysis of their dynamical structure, focusing on the contributions of individual components (e.g., bulge, disk, dark halo) to the rotation curve. This allowed us to further refine the sample by selecting only those models that best match the rotation curve data of our Galaxy, resulting in a subset of 12 galaxy models. For each model, we calculated a triplet of actions (${J}_{\mathrm{R}}$, ${J}_{z}$, ${L}_{z}$) and the adiabatic invariant ${J}_{v}={J}_{\mathrm{R}}+{J}_{z}+|{L}_{z}|/2$ for each stellar particle comprising such systems. We found that the two-dimensional $({L}_{z},{J}_{v})$ maps for all barred models consistently exhibit a distinct thin or thick stripe associated with the bar. In some cases, a counterrotating component characterized by ${L}_{z}<0$ and/or a roughly axisymmetric vertically extended component appearing above the bar stripe in action space are observed. By comparing the actual contributions of individual components to those derived from standard rotation curve decomposition techniques, we found that the dark matter halo in the considered cosmological models is more concentrated than typical Navarro-Frenk-White profiles and instead follows a pseudoisothermal profile, while the stellar component is overall less massive. Separating the bar from the disk via actions, we obtained that the bulge, which effectively represents the bar while decomposing the rotation curve, is found to be approximately one and a half times as massive as the actual bar component.