Simulations of supermassive black hole growth in high-redshift disk galaxies

TL;DR

This study uses high-resolution simulations to explore how supermassive black holes grow in high-redshift disk galaxies, revealing that dense cloud accretion can significantly contribute to black hole growth and AGN activity.

Contribution

The paper presents detailed simulations showing that dense molecular cloud accretion in isolated z=2 disk galaxies can dominate black hole growth, a novel insight into galaxy evolution.

Findings

Black holes in gas-rich disks undergo episodic Eddington-limited accretion.

High-fgas disks host weak to moderate AGNs 25-10% of the time.

Black hole growth in high-fgas galaxies is up to 1000 times higher than in low-fgas galaxies.

Abstract

Observations suggest that a large fraction of black hole growth occurs in normal star-forming disk galaxies. Here we describe simulations of black hole accretion in isolated disk galaxies with sufficient resolution (~5 pc) to track the formation of giant molecular clouds that feed the black hole. Black holes in z=2 gas-rich disks (fgas=50%) occasionally undergo ~10 Myr episodes of Eddington-limited accretion driven by stochastic collisions with massive, dense clouds. We predict that these gas-rich disks host weak AGNs 1/4 of the time, and moderate/strong AGNs 10% of the time. Averaged over 100 Myr timescales and the full distribution of accretion rates, the black holes grow at a few per cent of the Eddington limit -- sufficient to match observations and keep the galaxies on the MBH-Mbulge relation. This suggests that dense cloud accretion in isolated z=2 disks could dominate cosmic black hole growth. In z=0 disks with fgas=10%, Eddington-limited growth is extremely rare because typical gas clouds are smaller and more susceptible to disruption by AGN feedback. This results in an average black hole growth rate in high-fgas galaxies that is up to 1000 times higher than that in low-fgas galaxies. In all our simulations, accretion shows variability by factors of 10^4 on a variety of time scales, with variability at 1 Myr scales driven by the structure of the interstellar medium.