RUBIES Reveals a Massive Quiescent Galaxy at z=7.3

TL;DR

The paper reports the spectroscopic confirmation of RUBIES-UDS-QG-z7, a massive quiescent galaxy at $z=7.29$, formed most of its stellar mass by $z\sim8-9$ in a burst and quenched within ~50–100 Myr. Its compact size ($R_e\approx209$ pc) and high central stellar density ($\log(\Sigma_{*,e})\approx10.85$) indicate that galactic cores comparable to those in local ellipticals were already in place less than 700 Myr after the Big Bang. Non-parametric SFH modeling with JWST NIRSpec/PRISM and NIRCam/MIRI photometry yields $\log(M_*/M_\odot)=10.23\pm0.04$ and a sustained low sSFR, challenging galaxy formation models which underpredict such systems at $z>7$ (e.g., $\log(n/\mathrm{Mpc^{-3}})\approx-5.8$). The results imply that some galaxy cores formed extremely early, prompting revisions to feedback and star-formation prescriptions and motivating wider, deeper JWST surveys to constrain the abundance and evolution of early quiescent galaxies.

Abstract

We report the spectroscopic discovery of a massive quiescent galaxy at $z_{\rm spec}=7.29\pm0.01$, just $\sim700\,$Myr after the Big Bang. RUBIES-UDS-QG-z7 was selected from public JWST/NIRCam and MIRI imaging from the PRIMER survey and observed with JWST/NIRSpec as part of RUBIES. The NIRSpec/PRISM spectrum reveals one of the strongest Balmer breaks observed thus far at $z>6$, no emission lines, but tentative Balmer and Ca absorption features, as well as a Lyman break. Simultaneous modeling of the NIRSpec/PRISM spectrum and NIRCam and MIRI photometry (spanning $0.9-18\,μm$) shows that the galaxy formed a stellar mass of log$(M_*/M_\odot)=10.23^{+0.04}_{-0.04}$ before $z\sim8$, and ceased forming stars $50-100\,$Myr prior to the time of observation, resulting in $\log(\rm{sSFR/Gyr}^{-1})<-1$. We measure a small physical size of $209_{-24}^{+33}\,{\rm pc}$, which implies a high stellar mass surface density within the effective radius of $\log(Σ_{*,\rm e}/M_\odot\,kpc^{-2})=10.85_{-0.12}^{+0.11}$ comparable to the highest densities measured in quiescent galaxies at $z\sim2-5$. The 3D stellar mass density profile of RUBIES-UDS-QG-z7 is remarkably similar to the central densities of local massive ellipticals, suggesting that at least some of their cores may have already been in place at $z>7$. The discovery of RUBIES-UDS-QG-z7 has strong implications for galaxy formation models: the estimated number density of quiescent galaxies at $z\sim7$ is $>100\times$ larger than predicted from any model to date, indicating that quiescent galaxies have formed earlier than previously expected.

Figures (6)

• Figure 1: NIRSpec/PRISM Spectrum of RUBIES-UDS-QG-z7. Top: 2D SNR spectrum. Bottom: 1D spectrum of RUBIES-UDS-QG-z7 in red, with 1$\sigma$ uncertainties in gray. The NIRCam photometry is shown as orange dots and the best-fitting SED from Prospector in blue (see Section \ref{['sec:prospector']}). The 1D PRISM spectrum has been scaled by the calibration vector inferred by Prospector to account for slit-loss. A zoom-in to the region around $\lambda_{\rm rest}\sim0.4\,\mu$m is shown in the inset panel, where we highlight the position of various absorption features. Note also that the best-fitting SED latches on the Balmer absorption lines.
• Figure 2: Top: NIRCam and MIRI imaging cutouts of RUBIES-UDS-QG-z7. Middle: The posterior median SED from Prospector along with the NIRCam + MIRI photometry, which constrains the fit over the rest-frame optical to near-IR. We further plot the MSA slit position on an inset showing a $2\hbox{$.\!\!^{\prime\prime}$}04\times2\hbox{$.\!\!^{\prime\prime}$}04$ RGB-cutout around the source. Bottom: Fitted non-parametric SFHs for the low-metallicity fit ($Z=0.11\,{\rm Z_\odot}$) in red and the high-metallicity ($Z=1.19\,{\rm Z_\odot}$) fit in khaki. This shows a rising SFH for RUBIES-UDS-QG-z7 which peaks at $z\sim8-9$, followed by a rapid decline in the SFR at $z\sim8$ and little to no star formation in the past $\gtrsim50\,$Myr. For comparison, we overplot bright sources at $z\gtrsim8$ with their SFR-values as published in Akins2023 (COS-z8M1), Bunker2023b (GN-z11), Castellano2024 (GHZ2/GLASS-z12), and Carniani2024 (JADES-GS-z14-0), as well as three samples of UV-bright objects at $z>7.7$ plotted with their published values of $M_{\rm UV}$, converted to SFR with the relation from Kennicutt1998, assuming no dust. Grey: photometrically selected targets for follow-up with ALMA through REBELS from Bouwens2022b, light blue: a compilation of NIRSpec observations by Roberts-Borsani2024, and light green: objects from the JWST BoRG-survey Roberts-Borsani2024b. This illustrates that plausible progenitors of RUBIES-UDS-QG-z7 at $z\sim8-9$, and perhaps at even higher redshifts are either dust-obscured or have yet to be discovered or at least spectroscopically confirmed.
• Figure 3: Left: Projected stellar mass surface density within $R_{\rm e}$ versus redshift for RUBIES-UDS-QG-z7 compared to the massive quiescent galaxy at $z=4.9$ from deGraaff2024, a massive quiescent galaxy at $z=4.66$ from Carnall2023b, the core component of a massive quiescent galaxy at $z=3.97$ from Setton2024, the densest star cluster found in the Sunrise arc by Vanzella2023, and a sample of 225 massive quiescent galaxies at $z\sim2-3$ selected from the 3DHST survey. The surface density of RUBIES-UDS-QG-z7 is consistent with the densities of massive quiescent galaxies at $z\sim4-5$ and the densest systems at $z\sim2-3$, and only a factor $\sim4$ below that of the densest known star clusters at $z\sim6$. Right: Mass density profiles of some of the objects shown on the left, as well as of the triply imaged LRD from Furtak2024. For the quiescent galaxies at $z\sim2-3$, we show the median stellar mass profile and shade the region between the 16th and 84th percentiles. Similarly, we compile a sample of 514 massive quiescent galaxies at $z\sim0$ from the GAMA survey, and show the respective median and percentile profiles. The stellar mass density of RUBIES-UDS-QG-z7 at $R\sim300\,$pc is consistent with the latter, indicating that the cores of some local ellipticals may be in place at $z\sim7$.
• Figure 4: Comparison of RUBIES-UDS-QG-z7 to various spectra from the literature: the low-mass, recently quenched galaxies (also referred to as (mini-)quenched) at $z\sim7.3$ from Looser2024 and at $z\sim5.2$ from Strait2023, the triply imaged LRD from Furtak2024, and the massive quiescent galaxy at $z=4.9$ from deGraaff2024. While the former two show much stronger UV-emission than RUBIES-UDS-QG-z7, the LRD shows a much redder rest-optical continuum and a broad H$\beta$ emission line. For its redshift, the spectrum of RUBIES-UDS-QG-z7 is unique. However, it shows remarkable agreement with the spectral shape of the $z=4.9$ massive quiescent galaxy.
• Figure 5: Balmer Break strength vs. redshift for the 5 sources shown in Figure \ref{['fig:spectrum_comparison']}, three objects with strong breaks from Wang2024b, YD4 as published in Witten2024, the median relation from stacked NIRSpec spectra from Roberts-Borsani2024, and a sample of objects with NIRSpec spectroscopic redshifts and breaks measured from photometry Kuruvanthodi2024. RUBIES-UDS-QG-z7 shows one of the strongest Balmer breaks measured at $z\gtrsim7$ to date, comparable to the break strength measured in the massive quiescent galaxy at $z=4.9$deGraaff2024.