Title:Sowing black hole seeds: Direct collapse black hole formation with realistic Lyman-Werner radiation in cosmological simulations

Abstract:We study the birth of supermassive black holes from the direct collapse process and characterize the sites where these black hole seeds form. In the pre-reionization epoch, molecular hydrogen (H$_2$) is an efficient coolant, causing gas to fragment and form Population III stars, but Lyman-Werner radiation can suppress H$_2$ formation and allow gas to collapse directly into a massive black hole. The critical flux required to inhibit H$_2$ formation, $J_{\rm crit}$, is hotly debated, largely due to the uncertainties in the source radiation spectrum, H$_2$ self-shielding, and collisional dissociation rates. Here, we test the power of the direct collapse model in a self-consistent, time-dependant, non-uniform Lyman-Werner radiation field -- the first time such has been done in a cosmological volume -- using an updated version of the SPH+N-body tree code Gasoline with H$_2$ non-equilibrium abundance tracking, H$_2$ cooling, and a modern SPH implementation. We vary $J_{\rm crit} $ from $30$ to $10^3$ in units of $J_{21}$ to study how this parameter impacts the number of seed black holes and the type of galaxies which host them. We focus on black hole formation as a function of environment, halo mass, metallicity, and proximity of the Lyman-Werner source. Massive black hole seeds form more abundantly with lower $J_{\rm crit}$ thresholds, but regardless of $J_{\rm crit}$, these seeds typically form in halos that have recently begun star formation. Our results do not confirm the proposed atomic cooling halo pair scenario; rather black hole seeds predominantly form in low-metallicity pockets of halos which already host star formation.