How well is angular momentum accretion modelled in semi-analytic galaxy formation models?

TL;DR

This study evaluates how accurately semi-analytic galaxy formation models simulate gas cooling and angular momentum transfer compared to hydrodynamical simulations, revealing varying degrees of overestimation and underestimation among models.

Contribution

It provides a detailed comparison of multiple semi-analytic models against hydrodynamical simulations, highlighting their strengths and limitations in modeling angular momentum during gas accretion.

Findings

Gas retains 70-80% of angular momentum during accretion.

MORGANA overestimates angular momentum of cooled gas.

Older GALFORM models underestimate angular momentum.

Abstract

Gas cooling and accretion in haloes delivers mass and angular momentum onto galaxies. In this work, we investigate the accuracy of the modelling of this important process in several different semi-analytic (SA) galaxy formation models (GALFORM, L-GALAXIES and MORGANA) through comparisons with a hydrodynamical simulation performed with the moving-mesh code AREPO. Both SA models and the simulation were run without any feedback or metal enrichment, in order to focus on the cooling and accretion process. All of the SA models considered here assume that gas cools from a spherical halo. We found that the assumption that the gas conserves its angular momentum when moving from the virial radius, $r_{\rm vir}$, to the central region of the halo, $r\sim 0.1 r_{\rm vir}$, is approximately consistent with the results from our simulation, in which gas typically retains $70-80\%$ of its angular momentum during this process. We also found that, compared to the simulation, the MORGANA model tends to overestimate the mean specific angular momentum of cooled-down gas, the L-GALAXIES model also tends to overestimate this in low-redshift massive haloes, while the two older GALFORM models tend to underestimate the angular momentum. In general, the predictions of the new GALFORM cooling model developed by Hou et al. agree the best with the simulation.