In situ versus ex situ drivers of galaxy quenching: critical black hole mass and main sequence universality in the FLAMINGO simulation
Published version
Peer-reviewed
Repository URI
Repository DOI
Type
Change log
Abstract
ABSTRACT Exploiting a large sample of 5.3 million galaxies with $M_\ast , {=}, 10^{10-11}, {\rm M}\odot$ from the highest-resolution FLAMINGO simulation, we carry out a statistical analysis of quiescent and star-forming galaxies to explore quenching mechanisms. From redshift $z, {\simeq }, 7$ to 0, we find that the median star formation rate of main-sequence galaxies is independent of the environment and of whether a galaxy is a central or satellite, whereas the fraction of quiescent galaxies is highly sensitive to both. By employing Random Forest (RF) classifiers, we demonstrate that black hole (BH) feedback is the most responsible quenching mechanism for both centrals and satellites, while halo mass is the second most significant. For satellites, a notable importance given by RF to stellar mass implies in situ pre-quenching rather than ex situ pre-processing prior to infall to the current host halo. In the cosmic afternoon of $z, {=},$0 –1, we identify two distinct regimes of evolution: at $M{\rm BH}, {\gtrsim }, 10^7, {\rm M}\odot$, essentially all galaxies are quenched regardless of their environment; at $M{\rm BH}, {\lesssim }, 10^7, {\rm M}\odot$, quenching is determined mainly by halo mass. Galaxies undergo a sharp transition from the main sequence to quiescence once their BH mass reaches $M{\rm BH}, {\simeq }, 10^7, {\rm M}\odot$ (typically when $M\ast , {\simeq }, 10^{10.5}, {\rm M}\odot$ and $M{\rm h}, {\simeq }, 10^{12}, {\rm M}_\odot$) with a short quenching time-scale of ${<}$ 1 Gyr. This transition is driven by a sudden change in the gas mass in the inner circumgalactic medium. Our results indicate that galaxy quenching arises from a combination of in situ and ex situ physical processes.
Description
Acknowledgements: We thank the referee for constructive comments that improved the paper. SL (Seunghwan Lim) and RM (Roberto Maiolino) acknowledge support by the Science and Technology Facilities Council (STFC) and by the UKRI (UK Research and Innovation) Frontier Research grant RISEandFALL. RM also acknowledges funding from a research professorship from the Royal Society. This work used the DiRAC(Distributed Research using Advanced Computing)@Durham facility managed by the Institute for Computational Cosmology on behalf of the STFC (Science and Technology Facilities Council) DiRAC HPC (High Performance Computing) Facility (www.dirac.ac.uk). The equipment was funded by BEIS (Business, Energy and Industrial Strategy) capital funding via STFC capital grants ST/K00042X/1, ST/P002293/1, ST/R002371/1 and ST/S002502/1, Durham University and STFC operations grant ST/R000832/1. DiRAC is part of the National e-Infrastructure. This work is partly funded by research programme Athena 184.034.002 from the Dutch Research Council (NWO).
Funder: NWO; doi: https://doi.org/10.13039/501100003246
Journal Title
Conference Name
Journal ISSN
1365-2966

