A population of red candidate massive galaxies ~600 Myr after the Big Bang

TL;DR

JWST enables exploration of massive galaxies at $z>7$ by providing 1–5 μm coverage that reveals the Balmer break in the rest-frame optical. The authors apply an empirical double-break SED selection to CEERS NIRCam data and perform multi-method photometric redshift and stellar mass inferences with EAZY (emission-line templates), Prospector-α, and Bagpipes across varied priors to assess systematics. They identify six galaxies with fiducial $M_* > 10^{10} M_{ m sun}$ in the range $7.4 \le z \le 9.1$, with one candidate near $M_* \sim 10^{11} M_{ m sun}$ at $z \approx 7.5$; if confirmed, the measured mass densities in redshift bins $7<z<8.5$ and $8.5<z<10$ would exceed previous UV-selected predictions by factors up to $\sim 20$ at $z\sim 8$ and $\sim 1000$ at $z\sim 9$. The study also shows that, depending on attenuation laws and SFH assumptions, derived masses can vary by large factors, underscoring the need for spectroscopic confirmation and direct continuum or dynamical mass measurements. Overall, the results imply a more substantial early population of massive galaxies than previously thought, with JWST revealing a substantial extension to the high-mass end of the galaxy stellar mass function at $z>7$.

Abstract

Galaxies with stellar masses as high as $\sim 10^{11}$ solar masses have been identified out to redshifts $z \sim 6$, approximately one billion years after the Big Bang. It has been difficult to find massive galaxies at even earlier times, as the Balmer break region, which is needed for accurate mass estimates, is redshifted to wavelengths beyond $2.5\mum$. Here we make use of the $1-5\mum$ coverage of the JWST early release observations to search for intrinsically red galaxies in the first ~750 million years of cosmic history. In the survey area, we find six candidate massive galaxies (stellar mass $>10^{10}$ solar masses) at $7.4 < z < 9.1$, 500 - 700 Myr after the Big Bang, including one galaxy with a possible stellar mass of $\sim 10^{11}$ solar masses. If verified with spectroscopy, the stellar mass density in massive galaxies would be much higher than anticipated from previous studies based on rest-frame ultraviolet-selected samples.

Figures (7)

• Figure 1: Redshifts and tentative stellar masses of double-break selected galaxies. Shown in gray circles are EAZY-determined redshifts and stellar masses using emission-line enhanced templates (Salpeter IMF) for objects with$\mathrm{S} / \mathrm{N}>8$ in the F444W band. Fiducial redshifts and masses of the bright galaxies (F444W $<27 \mathrm{AB}$ ) that satisfy our double-break selection are shown by the large red symbols. Uncertainties are the $16^{\text{th }}-84^{\text{th }}$ percentile of the posterior probability distribution. All galaxies have photometric redshifts $6.5<\mathrm{z}<9$. 1 . Six galaxies are candidate massive galaxies with fiducial $\mathrm{M}^{*}>10^{10} \mathrm{M}_{\odot}$.
• Figure 2: Images of the six galaxies with the highest apparent masses as a function of wavelength. The fiducial stellar masses of the galaxies are ($\left.\log \left(\mathrm{M} * / \mathrm{M}_{\odot}\right)>10\right)$. Each cutout has a size of 2.4" × 2.4". The filters range from the $0.6 \mu \mathrm{~m}$ F606W filter of HST/ACS to the $4.4 \mu \mathrm{~m}$ F444W JWST/NIRCam filter. The galaxies are undetected in the optical filters; blue in the short-wavelength NIRCam filters; and red in the long-wavelength NIRCam filters. The color stamps show F150W in blue, F277W in green, and F444W in red.
• Figure 3: Spectral energy distributions and stellar population model fits. Top rows: photometry (black squares), best-fitting EAZY models (red lines) and redshift probability distribution$\mathrm{P}(\mathrm{z})$ (gray filled histograms) of six galaxies with apparent fiducial masses $\log \left(\mathrm{M} * / \mathrm{M}_{\odot}\right)>10$. The flux density units are $\mathrm{f}_{\mathrm{v}}$. Uncertainties and upper limits (triangles) are $1 \sigma$. Fiducial best-fit stellar masses and redshifts are noted. The seds are characterized by a double break: a Lyman break and an upturn at $>3 \mu \mathrm{~m}$. Emission lines are visible in the longest wavelength bands in several cases. Bottom panel: average rest-frame SED of the 6 candidate massive galaxies (red dots) and the $16^{\text{th }}-84^{\text{th }}$ percentile of the running median (shaded area). The red line is the best-fit median EAZY model. Green squares and the green line show average rest-frame UV-selected galaxies at $\mathrm{z}=8,10$ from HST+Spitzer ${ }^{15,34}$. Gray triangles show two spectroscopically confirmed galaxies at $\mathrm{z} \sim 9^{23,36,44}$. The double break selected galaxies are significantly redder than previously identified objects at similar redshifts. This may be due to high M/L ratios or effects that are not included in our modeling, such as AGN or exotic lines.
• Figure 4: Cumulative stellar mass density, if the fiducial masses of the JWST-selected red galaxies are confirmed. The solid symbols show the total mass density in two redshift bins,$7<\mathrm{z}<8.5$ and $8.5<\mathrm{z}<10$, based on the three most massive galaxies in each bin. Uncertainties reflect Poisson statistics and cosmic variance. The dashed lines are derived from Schechter fits to UV-selected samples. ${ }^{3}$ The JWST-selected galaxies would greatly exceed the mass densities of massive galaxies that were expected at these redshifts based on previous studies. This indicates that these studies were highly incomplete or that the fiducial masses are overestimated by a large factor.
• Figure 5: Extended Data Figure 3. Spectral energy distributions of all $\mathbf{1 3}$ galaxies that satisfy the color-color selection. The layout of the figure is identical to Fig. 3 in the main text. In addition, an alternative model fit (model E, see Methods) is shown that produces low stellar masses (blue), but generally requires extremely young ages ( $<5 \mathrm{Myr}$ ) at specific narrow redshift intervals. The panel at the lower right shows the averaged rest-frame SED of the seven galaxies with fiducial $\log \left(\mathrm{M} * / \mathrm{M}_{\odot}\right)<10$, compared to previously-found galaxies at similar redshifts (see Fig. 3).