Supermassive black hole seeds from sub-keV dark matter

Abstract

Quasars observed at redshifts $z\sim 6-7.5$ are powered by supermassive black holes which are too large to have grown from early stellar remnants without efficient super-Eddington accretion. A proposal for alleviating this tension is for dust and metal-free gas clouds to have undergone a process of direct collapse, producing black hole seeds of mass $M_\textrm{seed}\sim10^5 M_\odot$ around redshift $z \sim 17$. For direct collapse to occur, a large flux of UV photons must exist to photodissociate molecular hydrogen, allowing the gas to cool slowly and avoid fragmentation. We investigate the possibility of sub-keV mass dark matter decaying or annihilating to produce the UV flux needed to cause direct collapse. We find that annihilating dark matter with a mass in the range of $13.6 \textrm{ eV} \le m_{dm} \le 20 \textrm{ eV}$ can produce the required flux while avoiding existing constraints. A non-thermally produced dark matter particle which comprises the entire dark matter abundance requires a thermally averaged cross section of $\langleσv \rangle \sim 10^{-35}$ cm$^3/$s. Alternatively, the flux could originate from a thermal relic which comprises only a fraction $\sim10^{-9}$ of the total dark matter density. Decaying dark matter models which are unconstrained by independent astrophysical observations are unable to sufficiently suppress molecular hydrogen, except in gas clouds embedded in dark matter halos which are larger, cuspier, or more concentrated than current simulations predict. Lastly, we explore how our results could change with the inclusion of full three-dimensional effects. Notably, we demonstrate that if the $\mathrm{H}_2$ self-shielding is less than the conservative estimate used in this work, the range of both annihilating and decaying dark matter models which can cause direct collapse is significantly increased.