Presentation #100.19 in the session AGN.
Contrary to the standard lore, there is mounting observational evidence for active galactic nuclei (AGN) feedback at the low-mass end of the galaxy population. However, most modern cosmological simulations fail to reproduce the observed dwarf AGN population due to several implicit modeling choices that have been calibrated towards AGN in massive galaxies. We have investigated the impact of stellar and AGN feedback models on dwarf galaxies with a suite of high-resolution cosmological zoom-in simulations. Importantly, we find that there are sufficient amounts of gas to power brief Eddington-limited accretion episodes in dwarfs, which generate strong gas outflows and significantly suppress star formation. However, with the standard Bondi model, AGN accretion in dwarfs is always inefficient especially for light black hole seeds. This demonstrates the need for a more sophisticated black hole accretion model which allows for AGN activity across the whole galaxy stellar mass range. We have developed a novel accretion-disk based model, which changes from a thin alpha disk to an advection dominated accretion flow as the Eddington ratio decreases, covering all radiative regimes and incorporating black hole spin evolution due to Lense-Thirring precession. I will present some first results from a diverse suite of idealized massive black hole binary simulations, using this novel model to track the black holes’ evolution and inject feedback in accordance with the accretion disk state. I will also discuss the crucial impact of the disk state on electromagnetic counterpart predictions with a view towards future gravitational wave observatories such as LISA or IPTA.