A systematic search for dormant galaxies at z~5-7 from the JWST NIRSpec archive

TL;DR

The study addresses the existence and mass distribution of dormant galaxies at z ~ 5–7 and their bursty star-formation histories. It introduces a systematic JWST NIRSpec prism search that leverages weak Hα emission (EW < 50 Å) and strong Balmer breaks to identify galaxies that ceased star formation at least 10 Myr before observation. The authors identify 14 dormant galaxies with $M_* \,\sim\ 10^{7.6}-10^{10.5}\,M_\odot$ that lie below the main sequence and halted SF 10–25 Myr prior, revealing an intermediate-mass dormant population across a wide mass range and demonstrating a viable spectroscopic selection method in prism data. This work implies that dormant phases are common enough to be detectable in large JWST surveys and informs burstiness models, while highlighting the need for complete samples and extension to higher redshifts via additional emission lines such as [OIII].

Abstract

JWST has revealed a population of ``dormant'' galaxies at $z>5$ that have recently halted their star formation and are characterized by weak emission lines and significant Balmer breaks. Until now, only four such galaxies have been reported at $z>5$, three with low stellar masses, $M_*<10^9M_\odot$ (so-called mini-quenched galaxies), and one massive quiescent galaxy with $M_*=10^{10.2}M_\odot$; no such galaxy had been reported at intermediate masses. Here, we present a systematic search for dormant galaxies at $5<z<7.4$ that halted star formation at least 10 Myr before the time of observation. To do this, we made use of all the publicly available NIRSpec prism data in the DAWN JWST Archive (DJA) and selected galaxies with low H$α$ equivalent widths ($EW_{0}<50$Å) and strong Balmer breaks ($F_{ν,4200}/F_{ν,3500}>1.4$). We find 14 dormant galaxies with stellar masses ranging from $10^{7.6}-10^{10.5}$, revealing an intermediate-mass population. By construction, these 14 sources are located about 1 dex below the star-forming main sequence. Their star formation histories suggest that they halted star formation between 10 and 25 Myr before the time of observation which, according to models, is comparable with the timescales of internally regulated bursts driving a ``breathing'' mode of star formation. Our results show that $\sim1\%$ of the galaxies in the DJA are in a dormant phase of their star formation histories, and they span a wide stellar mass range. These galaxies can be empirically selected using only their spectral features in NIRSpec prism data.

Figures (7)

• Figure 1: Balmer beaks and rest-frame H$\alpha$ EWs of model galaxies that halted star formation at a grid of times before observation. The Bagpipes models assume a constant SFH of 100 Myr before star formation is stopped. They were computed with three different metallicities and two dust scenarios: $A_V=0$ and $A_V=0.8$. The gray region denotes the parameter space from which galaxies were selected for our sample. We find that the evolution of the Balmer breaks is metallicity- and dust-dependent, while H$\alpha$ EWs experience minimal variations between different metallicities, with no variations with different dust attenuation since we applied the same dust screen to the nebular lines as to the continuum.
• Figure 2: Observed $f_\nu$ spectra (blue) and posterior spectra from Bagpipes (orange) for four of the dormant galaxies selected in this work (the full set of spectra is shown in the appendix). The filled circles are the photometry values, which we fit simultaneously with the spectra. We observe strong Balmer breaks and weak emission lines. Insets: are the red-green-blue (RGB) cutouts of the galaxies, with the slits used for these observations overplotted in pink, showing the large size range of these sources.
• Figure 3: Distribution of Balmer break ratios and rest-frame H$\alpha$ EWs of our 14 dormant galaxies, highlighted against the distribution of the total galaxy sample. We find a strong correlation between the two quantities in the entire galaxy population at $5<z<7.4$, which is a key result for our understanding of burstiness. The dormant galaxies selected in this work lie on the edge of this distribution (shadowed in gray), with the strongest Balmer breaks and the weakest EWs.
• Figure 4: Past SFHs of the dormant galaxies in our sample. Left: sSFRs vs lookback time for all 14 individual sources. Shaded areas correspond to the 16-84th percentiles of the bagpipes posterior SFHs. Right: Mean sSFH of the 14 dormant galaxies against that of the entire sample of $1,598$ objects. The time bins for the continuity SFH were set to be at 0, 3, 10, 25, 50, 100, 250, 500, and 700 Myr. We observe that star formation halted in these galaxies between 10 and 25 Myr before observation. Previous to that, all galaxies experienced a continuous rise in star formation before it halted, except the galaxy with ID 4188, which experienced this burst 250 Myr before observation and whose star formation has been gradually decreasing ever since.
• Figure 5: Distribution of mass and H$\alpha$-derived SFR of our sample of 1,598 galaxies (purple contours). Also shown are the main sequence fits from Ciesla24 for $6<z<7$ galaxies, from McClymont25 for $z\sim6$, and from Cole25 for $5<z<6$. The shaded green area represents the $0.5$ dex scatter in the main sequence reported by Ciesla24. The starred points are the 14 dormant galaxies selected in this work and the filled circles are those reported to date at $z>5$: at $z=5.2$Strait+23, at $z=7.3$Looser+24, at $z=7.3$Weibel+24, and at $z=8.5$Baker+25; H$\alpha$ cannot be observed for the sources at $z>7.4$, and therefore we plot SFR$_{10}$ instead of SFR$_{\textrm{H}\alpha}$ for those literature sources. We observe that all of our dormant galaxies lie below the main sequence and span a stellar mass range of $M_*\sim10^{7.6}-10^{10.5}M_\odot$, without an empty gap between the low- and high-mass regime.