Quenching time-scales in the IllustrisTNG simulation

The time-scales for galaxy quenching offer clues to its underlying physical drivers. We investigate central galaxy quenching time-scales in the IllustrisTNG 100-1 simulation, their evolution over time, and the pre-quenching properties of galaxies that predict their quenching time-scales. Defining quenching duration τ_q as the time between crossing specific star formation rate (sSFR) thresholds, we find that ${\sim} 40{{\ \rm per\ cent}}$ of galaxies quench rapidly with τ_q < 1 Gyr, but a substantial tail of galaxies can take up to 10 Gyr to quench. Furthermore, 29 per cent of galaxies that left the star-forming main sequence (SFMS) more than 2 Gyr ago never fully quench by z = 0. While the median τ_q is fairly constant with epoch, the rate of galaxies leaving the SFMS increases steadily over cosmic time, with the rate of slow quenchers being dominant around z ~ 2-0.7. Compared to fast quenchers (τ_q < 1 Gyr), slow-quenching galaxies (τ_q > 1 Gyr) were more massive, had more massive black holes, had larger stellar radii, and accreted gas with higher specific angular momentum (AM) prior to quenching. These properties evolve little by z = 0, except for the accreting gas AM for fast quenchers, which reaches the same high AM as the gas in slow quenchers. By z = 0, slow quenchers also have residual star formation in extended gas rings. Using the expected relationship between stellar age gradient and τ_q for inside-out quenching we find agreement with Mapping Nearby Galaxies at Apache Point Observatory (MaNGA) integral field unit (IFU) observations. Our results suggest the accreting gas AM and potential well depth determine the quenching time-scale.