Why do galactic satellites shut down early?

Satellite galaxies around a virtual Milky Way show an interesting pattern: the less massive ones tend to shut down star formation more easily than their larger counterparts. This emerges from a comparison of three cosmological simulation suites – FIREbox, FIRE-2 zoom‑in and IllustrisTNG50 – with real‑world observations from SAGA, ELVES and the Milky Way+M31 system. To make a fair comparison, researchers selected only hosts with halo masses between 10^11.9 and 10^12.2 solar masses and satellites with stellar masses from 10^7 to 10^10 solar masses, applying the same projected aperture and a common quenching definition. All simulations reproduce the observed trend: the quenched fraction rises as stellar mass declines, matching the data from SAGA, ELVES and the MW+M31 system. This shows that the mass dependence is a robust outcome of current galaxy‑formation models. Radial trends, however, differ. SAGA and ELVES display a gentle decline in quenched fraction with projected distance, signalling strong environmental quenching near the host. TNG50 follows this behaviour closely, while FIREbox shows an almost flat curve within uncertainties and the FIRE-2 zoom‑ins exhibit a suppressed central quenched fraction, largely because their Milky Way‑M31‑like hosts lack massive satellites and show a clear radial split between star‑forming and quenched systems. These results imply that the host halo’s assembly history and environment can shape satellite fate, accounting for part of the diversity seen across the different simulations.

The simulations employ different physical prescriptions for stellar feedback and black‑hole feedback, as well as distinct treatments of interstellar gas dynamics. In FIREbox and FIRE-2, energetic feedback from supernovae and stellar winds is modelled with an impulsive approach that injects momentum and thermal energy directly into gas particles near young stars, driving galactic winds that can strip gas from low‑mass satellites even at modest halo masses. This wind‑driven stripping is especially efficient in systems with shallow gravitational potentials, which explains why the least massive satellites exhibit higher quenched fractions even far from the host. IllustrisTNG50, by contrast, uses a two‑stage feedback model: stellar winds are followed by thermal heating from supermassive black‑hole (AGN) activity in the more massive hosts. This creates an external pressure that compresses the circumgalactic medium and boosts the efficiency of ram‑pressure stripping when a satellite passes through the host halo. Spatial resolution differences also affect the ability to resolve sub‑kiloparsec features such as stripping tails and turbulence in the intergalactic medium. FIREbox’s relatively high uniform resolution can follow filamentary gas structures around hosts, whereas the FIRE-2 zoom‑ins, despite excellent resolution near the main galaxies, have a limited representation of the more distant halo because of their smaller simulated volume, leading to a lower central quenched fraction. Finally, the host halo’s accretion history – quantified by its mass‑accretion rate and concentration – alters the hot‑gas distribution and the frequency of close encounters, leaving an environmental imprint that shows up in the observed radial trends. These points highlight that feedback physics and halo assembly history are key factors driving the diversity of quenching predictions across the different simulation suites.