The impact of spurious collisional heating on the morphological evolution of simulated galactic discs

TL;DR

This study investigates how artificial collisional heating in N-body simulations affects the morphology and kinematics of galactic discs, highlighting the importance of particle resolution to avoid biased results.

Contribution

It demonstrates the impact of spurious dark matter particle interactions on galaxy evolution in simulations and provides guidelines for particle resolution to mitigate these effects.

Findings

Spurious heating increases stellar velocity dispersion and decreases mean azimuthal velocity.

Galaxies with fewer than 10^6 dark matter particles are highly affected by artificial heating.

Heating causes discs to become rounder, lose rotational support, and alters scaling relations.

Abstract

We use a suite of idealised N-body simulations to study the impact of spurious heating of star particles by dark matter particles on the kinematics and morphology of simulated galactic discs. We find that spurious collisional heating leads to a systematic increase of the azimuthal velocity dispersion ($\sigma_\phi$) of stellar particles and a corresponding decrease in their mean azimuthal velocities ($\overline{v}_\phi$). The rate of heating is dictated primarily by the number of dark matter halo particles (or equivalently, by the dark matter particle mass at fixed halo mass) and by radial gradients in the local dark matter density along the disc; it is largely insensitive to the stellar particle mass. Galaxies within haloes resolved with fewer than $\approx 10^6$ dark matter particles are particularly susceptible to spurious morphological evolution, irrespective of the total halo mass (with even more particles required to prevent heating of the galactic centre). Collisional heating transforms galactic discs from flattened structures into rounder spheroidal systems, causing them to lose rotational support in the process. It also affects the locations of galaxies in standard scaling relations that link their various properties: at fixed stellar mass, it increases the sizes of galaxies, and reduces their mean stellar rotation velocities and specific angular momenta. Our results urge caution when extrapolating simulated galaxy scaling relations to low masses where spurious collisional effects can bias their normalisation, slope and scatter.