Modelling the Growth of Supermassive Black Holes in Cosmological Simulations

TL;DR

This paper compares two models of supermassive black hole growth in cosmological simulations, evaluating their ability to reproduce observed black hole properties and highlighting challenges in modeling their mass functions.

Contribution

It introduces and compares the modified Bondi-Hoyle and ADP black hole growth models within large-scale cosmological simulations, analyzing their effectiveness and limitations.

Findings

Both models reproduce local scaling relations.

ADP model is offset at high black hole masses.

Simulated black hole mass functions are too steep.

Abstract

There is strong evidence that supermassive black holes reside in all galaxies that contain a stellar spheroid and their mass is tightly correlated with properties such as stellar bulge mass and velocity dispersion. There are also strong theoretical arguments that feedback from supermassive black holes plays an important role in shaping the high mass end of the galaxy mass function, hence to accurately model galaxies we also need to model the black holes. We present a comparison of two black hole growth models implemented within a large-scale, cosmological SPH simulation including star formation and feedback. One model is a modified Bondi-Hoyle prescription that grows black holes based on the smooth density of local gas, while the other is the recently proposed Accretion Disc Particle (ADP) method. This model swallows baryonic particles that pass within an accretion radius of the black hole and adds them to a subgrid accretion disc. Black holes are then grown by material from this disc. We find that both models can reproduce local scaling relations, although the ADP model is offset from the observed relations at high black hole masses. The total black hole mass density agrees between models to within a factor of three, but both struggle to reproduce the black hole mass function. The simulated mass functions are too steep and underestimate the number of intermediate and high mass black holes. In addition, the ADP model swallows excessive amounts of material at the resolution of large-scale, cosmological simulations producing unrealistically large accretion discs. Future work needs to be performed to improve the black hole mass function within simulations. This should be done through the mass growth and feedback as they are strongly coupled and should not be treated as separate entities.