Analysis of the early stages in the formation of massive galaxies with JWST and HST data

The most massive galaxies in the nearby Universe are generally quiescent and present relatively old stellar populations. Understanding the origins of these massive galaxies, such as the Milky Way and even more massive systems, and how they formed their stars is a primary objective in astrophysics. The reason for this is that these galaxies hold very valuable information to unravel the mysteries of cosmic evolution and the processes that govern how galaxies form and evolve.

Constraining the epoch in which these galaxies emerged and analyzing the early stages of the stellar mass assembly in their likely progenitors at higher redshifts would represent an important step forward in our comprehension of the complex process of galaxy formation and evolution. This thesis presents a comprehensive study which combines cutting-edge observational capabilities with a sophisticated  analysis  technique  to  investigate  the  first  stages in the formation and stellar mass assembly of massive progenitors at 1 < z < 4 of these galaxies. This work not only addresses the question of when they began to form their stellar populations but also compares the results with predictions of current cosmological simulations to explore its potential limitations and revisions required in their cosmological and/or galaxy formation models.

This research has been conducted by combining of the exceptional optical capabilities of the Hubble Space Telescope (HST) and the unprecedented infrared capabilities of the James Webb Space Telescope (JWST). Our dataset includes optical-to-NIR broad-band observations from the Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS) with HST and the Cosmic Evolution Early Release Science (CEERS) Survey with JWST. These cosmological surveys enable us to access the spatially-resolved emission of a representative sample of massive galaxies at 1 < z < 4. Our approach is to combine the information provided by these stellar populations in two dimensions (2D), derived from stellar population synthesis in 2D (2D SPS), and develop a methodology that allows us to infer the first stages of their integrated star formation histories (SFHs) in a more robust way than when only using integrated emission. This thesis can be divided into a first part on the development and validation of our 2D-SPS methodology with simulated imaging data, and a second part on the application of this 2D-SPS methodology to real JWST + HST observations.

The development of this 2D-SPS methodology has been performed using the Illustris numerical simulation. Illustris is a large-scale hydrodynamical simulation which reproduces the general relationships observed for galaxies at different redshifts. In addition, Illustris provides synthetic images that imitate those of real cosmological surveys from JWST and HST. We use the latter synthetic images from Illustris to develop and optimize our 2D-SPS methodology for inferring the early stages of the SFH of a galaxy, built by combining the information of their spatially-resolved stellar populations. The advantage of using Illustris is that we can access the individual simulated particles that comprise each galaxy in the simulation, which provides us with the ground-truth information regarding its formation and stellar mass assembly. This is crucial to test and evaluate our 2D-SPS-derived SFHs, especially their first stages, with the ground-truth values provided by the simulated particles belonging to the galaxies.

One of the benefits of using Illustris is that, using its merger trees and tracking galaxies in time down to z=0, we can select only massive 1 < z < 4 galaxies which are bona-fide progenitors of the most massive (M_★ > 10^11 M_⊙) local galaxies. We use this sample of massive 1 < z < 4 progenitors to test our 2D-SPS method and to study its effectiveness in recovering the first episodes of stellar mass assembly from the SFH of these galaxies. For our analysis, we quantify the first stages of stellar mass formation by calculating the formation times at which the galaxy formed 5%, 10%, and 25% of its stellar mass, computed directly from the corresponding SFH. We evaluate the goodness of our estimations in terms of the accuracy, defined as the median relative difference between the formation times measured and those obtained for the ground-truth extracted from Illustris. Our method proves to be successful in recovering the formation times with a median accuracy below 5%. In addition, the comparison of the formation times inferred for our sample of Illustris massive progenitors at 1 < z < 4 with those of their descendants at z=0, together with those inferred of the whole population of M_★ > 10^11 M_⊙ galaxies at z=0, gives us information about the limitations and biases we may also encounter in real (naturally magnitude-limited) observations.

With our 2D-SPS method already validated, we are ready to apply it to real massive galaxies observed in the first epoch of JWST/CEERS observations executed in June 2022. These observations consist of six NIRCam pointings which overlap the majority of the CANDELS/EGS field, for which HST data are already available. Our overall findings reveal that massive 1 < z < 4 galaxies in CEERS began its stellar mass assembly at very early ages of the Universe, challenging our previous assumptions regarding the formation of the first galaxies at high redshift, in line with other recent JWST works. We compare our results from the predictions for massive galaxies in Illustris and another state-of-art and improved simulation, the IllustrisTNG (The Next Generation). Both simulations exhibit significant discrepancies regarding the cosmic times at which galaxies began to form their stars when compared to the formation epochs of CEERS galaxies. Our results highlight potential shortcomings in the current galaxy formation models of these simulations at early epochs.

The research of this thesis not only advances in our comprehension of the early formation and assembly of the stellar mass in massive galaxies, but also provides observational constraints for future cosmological and galaxy formation models, adding another small piece to the puzzle of understanding of our cosmic origins.