CEERS: Possibly forging the first dust grains in the universe

Aims. This work aims to investigate the coevolution of metals and dust for 173 galaxies at 4.0 < z ≤ 11.4 spectroscopically observed by the NIRSpec instrument onboard the James Webb Space Telescope (JWST) in the Cosmic Evolution Early Release Science Survey (CEERS) project. More specifically, we want to study and analyse the properties of a sample of galaxies that show an extremely low dust attenuation and try to understand the possible physical processes at play in these galaxies. Methods. We developed a new version of the CIGALE code that accepts spectroscopic and photometric data. From a statistical comparison of the observations with the modelled spectra, we derived a set of physical parameters that allowed us to constrain the above physical processes. Results. Our analysis reveals a population of 49 extremely low-dust-attenuation galaxies (GELDAs) consistent with AFUV = 0.0 within 2σA_FUV and Mstar<109 M⊙. After stacking the spectra of the 49 GELDAs to increase the signal-to-noise ratio, we measured a very blue UV slope of βFUV=−2.451±0.066 and a Balmer decrement of Hα/Hβ = 2.932±0.660 without underlying absorption and consistent with no dust attenuation; Case B assumes an underlying absorption of 2.5%. Furthermore, the proportion of GELDAs is much higher at z > 8.8 (83.3% of the total sample) than at z < 8.8 (26.3% of the total sample). This suggests that GELDAs became dominant in the early Universe. Assuming a prior far-infrared dust spectrum from the ALPINE sample, we performed an analysis of the properties of this galaxy population. The trends observed in the Mdust versus Mstar diagram feature an upper and a lower sequence linked by objects that can be transitional. A comparison with models suggests that we might observe a critical transition at Mstar≈108.5 M⊙, corresponding to a critical metallicity of Zcrit = 12+log10(O/H) ≈ 7.60 (i.e. Z/Z⊙≈0.1). At this point, galaxies transition from being dominated by stellar-dust production (mainly from supernovae) to grain growth through gas–dust accretion in the ISM. The observational critical metallicity Zcrit derived in this paper is in good agreement with predictions from theoretical models for the onset of efficient grain growth. Furthermore, the mean gas-mass fraction of our entire sample at 4.0 < z < 11.4 is very high: fgas≳0.9. All of our galaxies, including GELDAs at all redshifts, contain a large amount of gas that was not expelled from the galaxies. Finally, the small size of the galaxies combined with the mass of gas lead to very high surface-gas densities – which put our sample below high-redshift sub-millimeter galaxies – at relatively low star formation efficiency. The population of high-redshift GELDAs would provide us with a natural and inherent explanation for the origin of the apparent tension between observations and theoretical models in the number density of bright galaxies at z ≳ 9.