Cosmological simulations of dwarfs: the need for ISM physics beyond SN feedback alone

TL;DR

This study uses high-resolution cosmological simulations to show that supernova feedback alone is insufficient to regulate dwarf galaxy properties, highlighting the need for additional ISM physics such as turbulent support and stellar feedback mechanisms.

Contribution

The paper demonstrates that supernova feedback alone cannot account for dwarf galaxy regulation in cosmological simulations, emphasizing the importance of other physical processes.

Findings

SN feedback often insufficient in dwarf galaxies, leading to excessive stellar mass and compactness.

Dense gas buildup prior to SN events reduces feedback effectiveness.

Results are robust against resolution and background variations.

Abstract

The dominant feedback mechanism in low mass haloes is usually assumed to take the form of massive stars exploding as supernovae (SNe). We perform very high resolution cosmological zoom-in simulations of five dwarf galaxies to z = 4 with our mechanical SN feedback model. This delivers the correct amount of momentum corresponding to the stage of the SN remnant evolution resolved, and has been shown to lead to realistic dwarf properties in isolated simulations. We find that in 4 out of our 5 simulated cosmological dwarfs, SN feedback has insufficient impact resulting in excessive stellar masses, extremely compact sizes and central super-solar stellar metallicities. The failure of the SN feedback in our dwarfs is physical in nature within our model and is the result of the build up of very dense gas in the early universe due to mergers and cosmic inflows prior to the first SN occurring. We demonstrate that our results are insensitive to resolution (provided that it is high enough), details of the (spatially uniform) UV background and reasonable alterations within our star formation prescription. We therefore conclude that the ability of SNe to regulate dwarf galaxy properties is dependent on other physical processes, such as turbulent pressure support, clustering and runaway of SN progenitors and other sources of stellar feedback.