Excavating the Ruins: An Ancient z = 2.675 Galaxy Which Formed in the First 500 Myr

We present the analysis of an ancient galaxy at z = 2.675, which we dub “Eridu.” Simultaneously modeling the JWST/NIRSpec G140M and G235M spectra from the SMILES program and 0.4–25 μm Hubble Space Telescope, JWST/NIRCam, and JWST/MIRI photometry from the JADES+SMILES photometric catalogs shows that Eridu is massive and quiescent, with stellar mass log(M*/M⊙)=10.96−0.01+0.01 and average star formation rate <1M⊙ yr−1 over the last 100 Myr. Star formation histories (SFHs) inferred from various models produce disconcertingly early and fast formation within ∼300 Myr of the Big Bang and quenching 2 Gyr prior to observation (z ∼ 10). This stellar mass assembly implies a M* ≈ 1011M⊙ progenitor at z > 10, nearly 2 orders of magnitude more than the most massive current high-redshift observations. From Eridu’s spectrum, we infer [Mg/Fe]=+0.65−0.19+0.20 , indicating its stellar population is extremely α-enhanced, consistent with the rapid formation timescale inferred from its SFH. We show that the inferred metallicity varies ∼0.4 dex between solar-scaled and α-enhanced models. As α-enhancement is expected in high-z quiescent galaxies, we speculate the common practice of spectral energy distribution modeling with solar-scaled abundance patterns could systematically affect the inferred ages of these objects. Eridu inhabits a massive protostructure that offers additional explanations for rapid mass assembly and quenching via environmental mechanisms, e.g., major mergers. Although its inferred formation is at odds with observations of the brightest z > 10 galaxies, future high-redshift galaxy formation studies and updated α-enhanced stellar models will unearth how Eridu and the first quiescent galaxies formed in the extremely early Universe.