Quenching of star formation in galaxies

The puzzle

Galaxies fall into two broad populations: those still forming stars (blue, gas-rich, with active Hα and UV emission) and those that have stopped (red, gas-poor, with old stellar populations). The split is sharp; the bimodality is strong enough to motivate a single label, quenching, for the transition.

The quiescent fraction also depends strongly on environment: it rises sharply in groups and clusters, signalling that something about the dense environment shuts down star formation. The questions driving this work are which physical processes are responsible (ram-pressure stripping, starvation, harassment, tidal interactions, feedback), when in a galaxy's orbital history they take effect, and how fast they complete the transition from blue to red.

What my group has found

Quenching is fast, efficient, and tied to first pericentre

In Oman & Hudson 2016 we applied the orbit-library method to an SDSS sample of cluster satellites (10⁹–10^11.5 M_⊙) and found that quenching by massive (> 10¹³ M_⊙) clusters is essentially 100% efficient: all satellites quench on their first infall, at or within a Gyr of first pericentric passage. Higher-mass satellites quench earlier, with little dependence on host-cluster mass.

Catching the transition with IFU spectroscopy

In Owers, Hudson, Oman et al. 2019 we used SAMI integral-field spectroscopy to identify a population of Hδ-strong cluster galaxies caught mid-transition. Their location (inside 0.6 R₂₀₀) and elevated velocity dispersion identify them as a first-infall population — consistent with the orbit-library prediction — and their spatially-resolved star formation is consistent with outside-in quenching by ram-pressure stripping on the first passage.

Separating gas stripping from star-formation quenching

In Oman, Bahé, Healy, Hess, Hudson & Verheijen 2021 we extended the framework with HI fluxes from ALFALFA, fitting separate stripping and quenching delay times. Stripping in massive clusters happens at or just before first pericentre; in lower-mass (10^13.5 M_⊙) groups it is delayed by ~3 Gyr. Balmer emission then fades a further ~3.5 Gyr (5.5 Gyr in groups) after pericentre, with (g−r) reddening following a few hundred Myr later. The delay-time scales are remarkably independent of satellite mass.

Quenching timescales from forward-modelled stellar ages

In Reeves, Hudson & Oman 2023 we forward-modelled both mass-weighted stellar ages and quiescent fractions in projected phase space, breaking the long-standing degeneracy between quenching onset and duration. We find total quenching times of t_Q ≈ 3.7–4.0 Gyr after first pericentre across 10⁹–10^10.5 M_⊙, with the onset close to or just after pericentre and a substantially longer SFR-decline timescale. This is in slight tension with the rapid-stripping picture and suggests that ram-pressure stripping is not complete on first passage.

How many stars form in galaxy mergers?

In Reeves & Hudson 2024 we forward-modelled the difference in stellar age between post-coalescence mergers and matched controls (same stellar mass, environment, and redshift) for 445 visually confirmed mergers from Bickley et al. Post-mergers are systematically younger by up to ~1.5 Gyr at lower stellar mass. Modelling the inspiral phase plus a coalescence burst gives a best-fit stellar-mass burst fraction f_burst = ΔM_☆/M_☆,merger = 0.18 ± 0.02 over 10 < log(M_☆/M_⊙) < 11, with no mass trend. Burst durations span ~120–250 Myr and the remaining cold-gas fraction predicted by the model matches observations. The inferred burst fractions are significantly larger than hydrodynamical-simulation predictions.