A SPectroscopic Survey of Biased Halos in the Reionization Era (ASPIRE): Spectroscopically Complete Census of Obscured Cosmic Star Formation Rate Density at z = 4–6

Abstract

We present a stringent measurement of the dust-obscured star formation rate density (SFRD) at z = 4–6 from the ASPIRE JWST Cycle-1 medium and ALMA Cycle-9 large program. We obtained JWST/NIRCam grism spectroscopy and ALMA 1.2 mm continuum map along 25 independent quasar sightlines, covering a total survey area of ∼35 arcmin ^2 where we search for dusty star-forming galaxies (DSFGs) at z = 0–7. We identify eight DSFGs in seven fields at z = 4–6 through the detection of H α or [O iii ] λ 5008 lines, including fainter lines such as H β , [O iii ] λ 4960, [N ii ] λ 6585, and [S ii ] λλ 6718,6733 for six sources. With this spectroscopically complete DSFG sample at z = 4–6 and negligible impact from cosmic variance (shot noise), we measure the infrared luminosity function (IRLF) down to L _IR ∼ 2 × 10 ^11 L _⊙ . We find flattening of IRLF at z = 4–6 towards the faint end (power-law slope $\alpha =0.5{9}_{-0.45}^{+0.39}$ ). We determine the dust-obscured cosmic SFRD at this epoch to be $\mathrm{log}[{\rho }_{\mathrm{SFR},\mathrm{IR}}/({M}_{\odot }\,{\mathrm{yr}}^{-1}\,{\mathrm{Mpc}}^{-3})]=-1.5{2}_{-0.13}^{+0.14}$ . This is significantly higher than previous determinations using ALMA data in the Hubble Ultra Deep Field, which is void of DSFGs at z = 4–6 because of strong cosmic variance (shot noise). We conclude that the majority (66% ± 7%) of cosmic star formation at z ∼ 5 is still obscured by dust. We also discuss the uncertainty of SFRD propagated from far-IR spectral energy distribution and IRLF at the bright end, which will need to be resolved with future ALMA and JWST observations.

Keywords

• James Webb Space Telescope
• Starburst galaxies
• High-redshift galaxies
• Luminous infrared galaxies
• Galaxy evolution