Dwarf satellite galaxies at high-redshift: physical understanding and JWST predictions

According to the cosmological ΛCDM (Lambda cold dark matter) model, dwarf galaxies are the most abundant type of galaxies not only in our Local Universe, where many are observed as satellites of our Milky Way (M* ~ 10^(11) Msun), but also at high-redshift. They constitute the building blocks that, assembling through merger events, formed the massive galaxies we see today. They also were the first galaxies to form in the Universe and that, already during the first billion years after the Big Bang (z > 6), drastically impacted all subsequent structure formation through feedback processes: supernovae (SN) explosions, for instance, are capable of both polluting the pristine gas with heavy elements, both injecting energy, heating and evacuating the gas, and possibly suppressing of star formation. Due to their shallow potential wells, dwarf galaxies are very sensitive to feedback processes, as well as possible gravitational interactions, making them the ideal laboratories to study these effects during early galaxy formation. Observing low-mass galaxies at high-z represents a great challenge due to their low luminosities. However, a new frontier can now be reached in their investigation thanks to the James Webb Space Telescope (JWST), that in its first few months of observations has already started to reveal the existence of faint galaxies at cosmic dawn (e.g. Castellano et al. 2022, Finkelstein et al. 2022, Matthee et al. 2022). It is therefore essential to tackle the many open issues related to early cosmic evolution and to provide predictions and tools for interpreting the upcoming data. With this aim in mind, in my Ph.D. work I focused on the study of dwarf galaxies in the early Universe and I addressed the following key questions: which physical mechanisms drive the evolution of high-z dwarf systems? How do they assemble to form massive galaxies? What is the role of feedback processes and can they quench star formation at high-z? Which among the galaxies populating the faint-end slope of the UV luminosity function can we detect with JWST? I have addressed these questions by taking advantage of state-of-the-art numerical simulations (Pallottini et al. 2017, Pallottini et al. (including Gelli) 2022) that follow the evolution of typical massive Lyman break galaxies (LBG, M* ~ 10^(10) Msun) up to z = 6. I developed a satellite finder that enabled the search of dwarf satellites in the densest regions surrounding massive galaxies in cosmological simulations and I found that, even in such extremely dense environments, their evolutionary and chemical properties are mainly regulated by their mass, ranging from M* ~ 10^6 Msun to ~ 10^9 Msun (Gelli et al. 2020). Indeed, low-mass dwarf galaxies (M* < 5x10^8 Msun) experience short and intense bursts of star formation ( < 50 Myr), whereas high-mass satellites (M* > 5x10^8 Msun) have much longer star formation histories, fuelled by the numerous merger events characterising the dense environment surrounding massive LBGs. Despite all satellites showing a remarkably rapid metal enrichment leading their stellar populations to solar values Z*~Zsun, low-mass systems contain a larger amount ( > 50 % ) of metal-poor stars (logZ*/Zsun<−0.5), similar to the behaviour of Local Group dwarf galaxies (e.g. Salvadori et al. 2015). In order to understand whether JWST will be able to shed light and uncover these faint systems, I modelled their spectral energy distributions (SEDs) and made accurate predictions for their observability (Gelli et al. 2021): producing synthetic images of the systems, I demonstrated that the emission of the satellites can be disentangled from the one of the central LBG if they are located at a distance of at least > 0.25′′ from its centre, i.e. in the vast majority of cases. By deriving color-magnitude diagrams, I found that the color F200W − F356W < − 0.25 is a powerful tool to identify low-mass star-bursting metal-poor dwarf galaxies. Since JWST is expected to detect tens of LBGs when performing deep surveys (e.g. JADES), we will have the unprecedented opportunity to detect “for free” also their dwarf satellite population. In view of the significant amount of detections of faint dwarf galaxies expected in upcoming JWST observations, it is important to provide meaningful statistical predictions to interpret them. For this reason, I performed an in-depth analysis of the overall population of dwarf satellites at 6 < z < 8 within the large suite of simulations SERRA (Gelli et al. to be submitted). Interestingly, I found that 31% of them are passive galaxies with no ongoing star formation and that they dominate the low-mass regime (M* < 10^8 Msun), once again similar to what is witnessed in present-day local dwarf galaxies. The principal process leading to their quenching at high-z is internal SN feedback heating up and evacuating the remaining gas. Most low-mass satellites later experience starvation due to the inefficiency in pulling additional gas onto their weak potential wells. By reconstructing their SEDs I demonstrated that, even though such passive satellites are expected to be very faint (mAB(F444W) > 29), their detection is feasible in deep surveys with the JWST/NIRCam instrument. I identified the color-color diagram F356W-F444W vs F200W-F356W as a unique and effective tool that can allow us to easily identify passive satellites, providing the opportunity to uncover this population of early quenched systems for the first time. Finally, while I was writing my PhD Thesis, the JWST detection of the first passive low-mass galaxy (M* = 10^(8.7) Msun) at z ~ 7.3 was reported by Looser et al. 2023, providing the unique opportunity to study the imprint of feedback process at early cosmic times. Comparing the SED of this observed passive system, JADES-GS-z7-01-QU, with that of simulated galaxies with similar stellar mass and fraction of time spent in activity (f_duty = 0.7), I found that the star formation rate decline is a powerful diagnostic to distinguish different feedback mechanisms and that the observed galaxy should have experienced an abrupt quenching caused by a process faster SN feedback alone (Gelli et al. 2023).