Double Trouble: Two spectroscopically confirmed low-mass quiescent galaxies at z>5 in overdensities

We present the discovery of two low-mass, high-redshift, quiescent galaxies, GS-z5-Q1 and COS-z5-Q1, using JWST NIRSpec spectroscopy alongside NIRCam and MIRI photometry. Observed at a redshift of z=5.39 and z=5.11 respectively, and with stellar masses of $\rm 10^{9.6}M_\odot$ and $\rm 10^{9.5}M_\odot$, GS-z5-Q1 and COS-z5-Q1 are two of the most distant quiescent galaxies spectroscopically confirmed to-date, and are by far the least massive ($\sim10\times$ lower mass). Full spectrophotometric modelling reveals that COS-z5-Q1 appears to have quenched more than 300Myr prior to observation ($z\sim 7$) and has a formation redshift of around z$\sim$11, whilst GS-z5-Q1 formed and quenched in a single burst around 150Myr prior to observation ($z\sim6$). GS-z5-Q1 is found to lie near the centre of a known high-z overdensity in GOODS-S, as would be expected by galaxy formation models, while COS-z5-Q1 lies towards the outskirts of an overdense region. This highlights the role that environment could play in accelerating galaxy evolutionary processes and could possibly be linked to the galaxies’ quiescent nature. By modelling their stellar populations, we show that these types of low-mass quiescent galaxies could potentially be descendants of the higher-z “mini-quenched” galaxies. The discovery of these two low-mass $z>5$ quiescent galaxies illuminates a previously undiscovered galaxy population and motivates dedicated follow-up surveys to investigate the overall population.

💡 Research Summary

This paper reports the spectroscopic confirmation of two low‑mass, high‑redshift quiescent galaxies—GS‑z5‑Q1 at z = 5.39 and COS‑z5‑Q1 at z = 5.11—using JWST NIRSpec PRISM/CLEAR low‑resolution spectra combined with deep NIRCam and MIRI photometry. Both objects have stellar masses of ≈10⁹·⁵ M⊙, making them roughly an order of magnitude less massive than any previously spectroscopically confirmed quiescent galaxy at comparable redshift, and placing them among the most distant quiescent systems known.

Candidate selection began with the extended UVJ colour criteria and a specific‑star‑formation‑rate (sSFR) cut (sSFR < 0.2/t_age(100 Myr)) applied to a photometric catalogue drawn from the five CANDELS fields (Baker et al. 2025d). After cross‑matching with spectra available in the Dawn JWST Archive, GS‑z5‑Q1 and COS‑z5‑Q1 emerged as the only objects satisfying both the photometric and spectroscopic requirements. Their UVJ colours lie outside the classic Schreiber et al. (2015) quiescent box but within the more inclusive Belli et al. (2019) and Baker et al. (2025b) selections, illustrating the necessity of broadened colour cuts for low‑mass high‑z quiescent searches.

The authors reduced the imaging data (PRIMER‑COSMOS v7.4 and GOODS‑S v7.2) with the Grizli pipeline, extracting 0.5″ circular aperture photometry and applying aperture‑to‑total corrections. Spectra were extracted with msaexp from the v4.3 DJA release, yielding 2451 s of integration for GS‑z5‑Q1 (observed as part of the Barrufet et al. 2025 campaign) and 17069 s for COS‑z5‑Q1 (from the CAPERS survey).

Joint spectro‑photometric fitting was performed with Prospector, employing the FSPS stellar population synthesis models, a flexible non‑parametric continuity star‑formation history (12 look‑back time bins), a Chabrier IMF, Calzetti dust law (A_V = 0–3 mag), and a metallicity prior spanning log Z/Z⊙ = −2.0 to 0.19. Redshifts were allowed to vary around the NIRSpec values with a Gaussian σ = 0.05, and a 10 % error floor was added to the photometry. Parameter inference used Dynamic Nested Sampling via dynesty.

Both spectra display a pronounced Balmer break and lack strong nebular emission lines, confirming a quiescent state with negligible star formation over the past ≈100 Myr. GS‑z5‑Q1’s best‑fit SFH shows a single, brief burst that formed the bulk of its stellar mass ≈150 Myr before observation (≈z 6), followed by rapid quenching. The inferred SFR during the burst is ≈40 M⊙ yr⁻¹, comparable to star‑forming galaxies observed at similar epochs. COS‑z5‑Q1 exhibits a more extended SFH: formation began ≈300 Myr before observation (≈z 7), with quenching occurring between z 6 and z 9. The uncertainties are larger, reflecting the flexibility of the continuity prior, but the overall picture is of an early‑forming system that ceased star formation well before the epoch of observation. Both galaxies have modest dust attenuation (A_V < 0.5 mag) and intermediate metallicities (log Z/Z⊙ ≈ −0.5).

Environmental analysis reveals that GS‑z5‑Q1 resides near the centre of a known overdensity in GOODS‑S, while COS‑z5‑Q1 lies toward the outskirts of the same structure. This spatial segregation aligns with theoretical expectations that low‑mass galaxies in protocluster cores experience accelerated evolution, possibly due to enhanced merger rates, ram‑pressure stripping, or early AGN feedback. The authors argue that the environment likely played a pivotal role in the rapid quenching of GS‑z5‑Q1, whereas COS‑z5‑Q1’s more gradual shutdown may reflect a less extreme local density.

The paper also situates these objects within the emerging “mini‑quenched” population (M★ < 10⁹ M⊙, quenching timescales of 20–50 Myr). While the two galaxies are 0.5–0.7 dex more massive, their quiescent nature and early formation suggest they could be the descendants of mini‑quenched systems that have continued to grow before finally shutting down.

In summary, the authors have expanded the frontier of spectroscopically confirmed quiescent galaxies to lower masses and higher redshifts, providing the first direct evidence that low‑mass quiescent galaxies can exist at z > 5 and that their quenching can be strongly modulated by environment. The work underscores the need for deeper NIRSpec observations and systematic surveys of high‑z overdensities to quantify the number density, formation pathways, and evolutionary fate of this previously hidden galaxy population.