# How important is non-ideal physics in simulations of sub-Eddington accretion onto spinning black holes?

## TL;DR

This study presents the first 3D general relativistic simulations of sub-Eddington accretion onto spinning black holes including non-ideal plasma effects, revealing that pressure anisotropy is significant but does not alter the overall flow structure.

## Contribution

It introduces the first global simulations incorporating pressure anisotropy and heat flux in low-density accretion flows onto Kerr black holes, challenging ideal MHD assumptions.

## Key findings

- Pressure anisotropy is comparable to magnetic pressure.
- Heat flux remains dynamically unimportant.
- Flow structure similar to ideal plasma simulations.

## Abstract

Black holes with accretion rates well below the Eddington rate are expected to be surrounded by low-density, hot, geometrically thick accretion disks. This includes the two black holes being imaged at sub-horizon resolution by the Event Horizon Telescope. In these disks, the mean free path for Coulomb interactions between charged particles is large, and the accreting matter is a nearly collisionless plasma. Despite this, numerical simulations have so far modeled these accretion flows using ideal magnetohydrodynamics. Here, we present the first global, general relativistic, 3D simulations of accretion flows onto a Kerr black hole including the non-ideal effects most likely to affect the dynamics of the disk: the anisotropy between the pressure parallel and perpendicular to the magnetic field, and the heat flux along magnetic field lines. We show that for both standard and magnetically arrested disks, the pressure anisotropy is comparable to the magnetic pressure, while the heat flux remains dynamically unimportant. Despite this large pressure anisotropy, however, the time-averaged structure of the accretion flow is strikingly similar to that found in simulations treating the plasma as an ideal fluid. We argue that these similarities are largely due to the interchangeability of the viscous and magnetic shear stresses as long as the magnetic pressure is small compared to the gas pressure, and to the sub-dominant role of pressure/viscous effects in magnetically arrested disks. We conclude by highlighting outstanding questions in modeling the dynamics of low collisionality accretion flows.