A census of quiescent galaxies across $0.5 < z < 8$ with JWST/MIRI: Mass-dependent number density evolution of quiescent galaxies in the early Universe

Abstract

JWST observations reveal numerous quiescent galaxies (QGs) at high redshift ($z \sim 4-8$), challenging models of early galaxy formation and quenching. Accurate number density estimates are crucial for comparison with theory but remain uncertain. We systematically study QGs at $0.5 < z < 8$ using a mass-complete sample from the JWST/PRIMER survey with deep NIRCam and MIRI imaging. The MIRI data, probing rest-frame near-infrared at $z \sim 3-8$, are vital for robust stellar mass measurement and QG identification. We find that nearly all photometrically selected, point-like QG candidates located in the UVJ QG region are actually "Little Red Dots", for which the UVJ colors were wrongly estimated due to inaccurate photometric redshift estimation. MIRI reduces significantly contamination to high-mass QGs from star-forming galaxies, yielding lower number densities than previous studies. The evolution of QG number density is strongly mass-dependent. The density of high-mass QGs ($\log (M_{\star}/M_{\odot}) > 10.3$) decreases rapidly from $n = 1\times10^{-5}~\mathrm{Mpc^{-3}}$ at $z=3-4$ to $n=2\times10^{-6}~\mathrm{Mpc^{-3}}$ at $z = 4-5$, becoming negligible ($n \lesssim 10^{-6}~\mathrm{Mpc^{-3}}$ ) at $z > 5$. Conversely, low-mass QGs ($9<\log (M_{\star}/M_{\odot})<10.3$) maintain a nearly constant number density ($n\sim3\times10^{-6}~\mathrm{Mpc^{-3}}$) across $z = 4-8$. This suggests low-mass QGs at $z > 4$ are likely temporarily quenched, akin to mini-quenched galaxies. Comparison with major hydrodynamical and semi-analytical models shows most underestimate high-mass QG densities at $z>4$ and fail to reproduce the constant low-mass QG density at $z>5$.

Figures (8)

• Figure 1: UVJ diagram of all galaxies with publicly available spectra at $3<z_{spec}<8$, color-coded by sSFR derived with BAGPIPES. There are two black horizontal dividing lines: the upper one at UV=1.3 was proposed by williams_detection_2009. This work shifts this line downward to 0.85 to include high-redshift PSBs. We also define an extension area where the properties of galaxies mostly fall between those of SFGs and QGs. Because we aim at assess the results of the photometric sample selection, the UVJ colors here are derived from SED fitting using only photometric data. However, when fitting the SFH, we include spectral with the emission lines ($H_{\alpha}$ and [O III]) masked out. A total of 12 galaxies fall within the PG+PSB region (4 PGs, 8 PSBs); their images, spectra, and fitted SFHs are shown in the figure, numbered in descending order of UV color. The SFHs of different galaxies are aligned together, with x-axis label represents the time before the galaxies are observed.
• Figure 2: The UVJ color-color diagram with $M_{\star}>10^{9}M_{\odot}$ in photometric sample, color-coded by sSFR. The comoving volume of all redshift slices is nearly the same (except for the highest-redshift bin), allowing direct interpretation of the evolution in QG and SFG number densities.
• Figure 3: Impact of MIRI photometry on UVJ color classifications. We show the UVJ colors of galaxies in the spectroscopic sample that exhibit inconsistent classifications between the MIRI and no‑MIRI runs. The left panel displays the rotated UVJ diagram. Black and yellow open circles, connected by arrows, denote the same galaxies in the MIRI and no‑MIRI groups, respectively; blue and red circles indicate galaxies of different masses. The two panels on the right present representative SED fitting results for two conditions (MIRI-exclude QG and MIRI-include QG), respectively.
• Figure 4: The mass-dependent number density evolution of QGs. The error bars represent Poisson errors only. The relative error for the low-redshift data is approximately 10%, and is not visible in the figure. Results from valentino_atlas_2023 and carnall_surprising_2023 are also shown, but the mass range is slightly different from ours. Empty circles mark the density calculated in the full sample. The single detection limit, defined as the inverse of the volume, marks the upper limit of the number density if only one source is detected in this redshift range. The result is compared with the number density in EAGLE, Illustris-TNG, GAEA, L-Galaxies, and Shark. We use the sSFR threshold $\mathrm{sSFR}<0.2/t_{Hubble}$) to select QGs in models.
• Figure 5: The evolution of QG fractions with redshfits for at different mass bins. Results from EAGLE, Illustris‑TNG, GAEA, L‑Galaxies, and Shark, together with observational data from muzzin_evolution_2013, are also plotted. Results for the full sample are shown as open circles and are plotted only when they exceed the Poisson error. The QG selection criterion for models is $\mathrm{sSFR}<0.2/t_{\mathrm{Hubble}}$. Data points from EAGLE and TNG are plotted only up to $z\sim4$, indicating an absence of QGs at higher redshifts. Although not perfect, the best galaxy formation models reproduce the evolution of the QG fraction up to $z=8$.