The Identification of Two JWST/NIRCam-Dark Starburst Galaxies at $z=6.6$ with ALMA

Fengwu Sun, Jinyi Yang, Feige Wang, Daniel J. Eisenstein, Roberto Decarli, Xiaohui Fan, George H. Rieke, Eduardo Bañados, Sarah E. I. Bosman, Zheng Cai, Jaclyn B. Champagne, Luis Colina, Francesco D'Eugenio, Yoshinobu Fudamoto, Mingyu Li, Xiaojing Lin, Weizhe Liu, Jianwei Lyu, Chiara Mazzucchelli, Xiangyu Jin, Hyunsung D. Jun, Yunjing Wu, Huanian Zhang

First listed 2025-06-06 | Last updated 2025-06-06

Abstract

We analyze two dusty star-forming galaxies at $z=6.6$. These galaxies are selected from the ASPIRE survey, a JWST Cycle-1 medium and ALMA Cycle-9 large program targeting 25 quasars and their environments at $z\simeq6.5 - 6.8$. These galaxies are identified as companions to UV-luminous quasars and robustly detected in ALMA continuum and [C II] emission, yet they are extraordinarily faint at the NIRCam wavelengths (down to $>28.0$ AB mag in the F356W band). They are more obscured than galaxies like Arp220, and thus we refer to them as "NIRCam-dark" starburst galaxies (star formation rate $\simeq 80 - 250\,\mathrm{M}_{\odot}\,\mathrm{yr}^{-1}$). Such galaxies are typically missed by (sub)-millimeter blank-field surveys. From the star-formation history (SFH), we show that the NIRCam-dark galaxies are viable progenitors of massive quiescent galaxies at $z\gtrsim4$ and descendants of UV-luminous galaxies at $z>10$. Although it is hard to constrain their number density from a quasar survey, we conclude that NIRCam-dark galaxies can be as abundant as $n\sim10^{-5.5}$ Mpc$^{-3}$ assuming a light halo occupation model. If true, this would equal to $\sim$30% of the number densities of both the quiescent galaxies at $z\gtrsim4$ and UV-luminous galaxies at $z>10$. We further predict that analogs at $z\sim8$ should exist according to the SFH of early massive quiescent galaxies. However, they may fall below the current detection limits of wide JWST and ALMA surveys, thus remaining "JWST-dark". To fully trace the evolution of massive galaxies and dust-obscured cosmic star formation at $z\gtrsim8$, wide-field JWST/NIRCam imaging and slitless spectroscopic surveys of early protoclusters are essential.

Short digest

Two extreme, dust-enshrouded starbursts at z=6.6 are identified in ASPIRE quasar fields: they are bright in ALMA 1.2 mm continuum and [C II] 158 µm, yet essentially disappear in JWST/NIRCam (F356W >28 AB). SEDs anchored to ALMA fluxes imply heavier obscuration than Arp 220 and star formation rates of ≃80–250 M⊙ yr⁻¹. Their inferred star-formation histories make them plausible progenitors of massive quiescent galaxies at z≳4 and descendants of UV-luminous z>10 systems, with an abundance as high as n≈10⁻⁵⋅⁵ Mpc⁻³ under a light halo-occupation model (~30% of those comparison populations). A key caveat is that number densities come from quasar-companion fields; analogous z∼8 systems may remain undetected (“JWST-dark”) at current survey depths.

Key figures to inspect

• Figure 1: Inspect the NIRCam cutouts versus ALMA 1.2 mm and [C II] spectra to see the stark NIRCam non-detections (F356W >28 AB) despite robust millimeter/line detections; note the quasar-subtracted panels near J1526–2050 and the ALMA contour overlap with the NIRCam-dark positions.
• Figure 2: Compare the best-fit cigale SEDs to Haro 11, Arp 220, and ALESS optically faint DSFG templates (all scaled to 1.2 mm) to verify the extreme attenuation—“more obscured than Arp 220”—and read off the implied LIR and SFR ≃80–250 M⊙ yr⁻¹.
• Figure 3: Place the two sources on the M⋆–z and SFR–z context alongside JWST-confirmed massive quiescients (z≳4), z>10 “blue monsters,” and ALMA starbursts (REBELS-25, MACS0416_Y1) to assess whether the NIRCam-dark points plausibly bridge blue monsters to quiescients via bursty SFHs.
• Figure 4: Examine how the light vs heavy host-halo models drive very different number-density inferences; check the quoted n≈10⁻⁵⋅⁵ Mpc⁻³ against quiescient and blue-monster densities and note the uncertainty from using quasar-environment companions.