The dust enrichment of early galaxies in the JWST and ALMA era

Abstract: Recent observations with the James Webb Space Telescope are yielding tantalizing hints of an early population of massive, bright galaxies at $z > 10$, with Atacama Large Millimeter Array (ALMA) observations indicating significant dust masses as early as $z\sim 7$. To understand the implications of these observations, we use the DELPHI semi-analytic model that jointly tracks the assembly of dark matter halos and their baryons, including the key processes of dust enrichment. Our model employs only two redshift- and mass-independent free parameters (the maximum star-formation efficiency and the fraction of supernova energy that couples to gas) that are tuned against all available galaxy data at $z \sim 5-9$ before it is used to make predictions up to $z \sim 20$. Our key results are: (i) the model under-predicts the observed ultraviolet luminosity function (UV LF) at $z > 12$; observations at $z>16$ lie close to, or even above, a "maximal" model where all available gas is turned into stars; (ii) UV selection would miss 34\% of the star formation rate density at $z \sim 5$, decreasing to 17\% by $z \sim 10$ for bright galaxies with $\rm{M_{UV}} < -19$; (iii) the dust mass ($M_d$) evolves with the stellar mass ($M_$) and redshift as $\log(M_d) = 1.194\log(M_) + 0.0975z - 5.433$; (iv) the dust temperature increases with stellar mass, ranging between $30-33$ K for $M_* \sim 10^{9-11}M_\odot$ galaxies at $z \sim 7$. Finally, we predict the far infrared LF at $z \sim 5-20$, testable with ALMA observations, and caution that spectroscopic redshifts and dust masses must be pinned down before invoking unphysical extrema in galaxy formation models.