Abstract

The co-evolution of massive black holes and their host galaxies is well established within ΛCDM cosmology. The repeated mergers, accretion, and feedback that conspire to regulate this process can be studied in large-scale cosmological simulations, such as Illustris, FABLE , MillenniumTNG and Flamingo. These simulations resolve key galaxy formation processes at ~kpc scales, but are plagued with numerical inaccuracies at the smaller scales of black holes. This scale discrepancy presents significant challenges for investigating black hole properties and generating testable predictions, e.g. for future JWST , Gaia, LISA and IPTA observations of isolated and binary black holes. In this talk I will discuss the black hole population in FABLE . Our results show that the numerical treatment of black holes in cosmological simulations leads to a misleading picture, even at the well-resolved large scale of galaxies. In particular, a large fraction of black holes coalesce well before their host galaxies merge and thus require extra delays on the order of a few Gyrs. These delays, governed by the dynamical timescale of the merging host galaxies, occur before and in addition to any delays arising from unresolved ‘sub-grid’ physics describing BH hardening mechanisms on parsec scales. This effect has profound implications for the black hole merger rates predicted from these large-scale cosmological simulations as well as for the multi-messenger predictions, once black hole growth during these dynamical galaxy merger delays is accounted for.