After the Big Bang, the baryonic Universe was entirely made up of primordial elements (hydrogen, helium and lithium). Heavier elements were produced only later on through nuclear fusion reactions in stellar cores as the first generation of stars in galaxies formed from this primordial gas. Through the recycling of interstellar gas over various generations of stars, the interstellar material gradually became enriched with heavy elements and some of these elements condensed to form solid particles - also knows as dust grains - with sizes ranging from nanometres to microns. These dust grains were long thought to grow in stellar environments during late stellar evolutionary phases, after which they were expelled into the interstellar medium. Recently, stellar dust production was demonstrated to be insufficient to account for present-day dust masses in galaxies based on the accurate dust mass determinations with Herschel and ALMA. An alternative mechanism to grow dust through accretion of elements in the interstellar medium was invoked to solve this conundrum and to produce most of the galactic dust in the local Universe.

As the very first galaxies in the early Universe still need to build up their reservoir of heavy elements, these pristine galaxies are thought to have relatively little dust grains present in their interstellar medium. The detections of massive dust reservoirs in some of the very first galaxies that formed when the Universe was only a few 100Myr old, seem to contradict with this hypothesis that little dust should have formed in these galaxies, and imposes the need for efficient dust production either from stellar sources or through grain growth.

We propose - in the first place - to derive accurate dust masses for a statistical sample of galaxies with dust detections in the Early Universe (e.g., the large ALMA program REBELS detecting galaxies between redshifts of 6.5 and 9 of which the PI is a member). Due to the availability of only one or a few wavebands with detected dust emission, it is a tedious job to constrain the temperature and mass of the dusty reservoirs in high-redshift galaxies. We will exploit a Bayesian modelling technique (Lamperti, Saintonge, De Looze+ 2020, Lamperti+ in prep.) to accurately model the dust emission for a sample of dusty high-redshift galaxies. Secondly, we will model the inferred dust masses with our Dust and Element evolUtion modelS (De Looze+2020). These models - and their relative contributions from dust production (grain growth, stellar dust) and destruction processes (e.g., astration, supernova shocks) - have been calibrated for the local Universe, and will have to be updated to model the dust formation and destruction in the Early Universe. By modelling a sample of extremely dusty high-redshift galaxies, we will push our models to the extreme to finally understand whether grains are mostly condensed in stellar environments or grown through accretion in the interstellar medium.

This project is ideally suited for a 2-year post-doctoral position.