Searching for high-redshift progenitors of massive galaxies

This Thesis focuses on the process of star formation and its relationship with the formation and evolution of the most massive galaxies in the Universe. The main hypothesis behind this work is that the picture of the high-z Universe based on deep optical/NIR surveys is potentially incomplete, since it misses most of the contribution from dusty sources that can be unveiled only with multi-wavelength data. In this Thesis, I follow two complementary approaches to support this hypothesis. In the initial chapters, I focus on the radio-selected (RS-) NIR-dark galaxies: a population of likely dusty star-forming galaxies selected at radio frequencies. I characterize these sources by taking advantage of the almost complete photometric coverage offered by the COSMOS field and involve in the analysis state-of-the-art facilities such as ALMA and JWST. The main result of this first part of the Thesis is the determination of the physical properties of the RS-NIRdark galaxies, picturing them as a population of highly dust-obscured (Av~4 mag), massive (M~10^(10.5-11) Msun) and star-forming (SFR~300-500 Msun/yr) galaxies mainly located at z~3 and beyond. These properties are then employed to estimate the expected contribution of these sources to the cosmic star formation rate density. The second approach, followed in the fifth chapter of the Thesis, focuses on a class of extremely red sources unveiled in the first data collected with JWST and missed by old-generation facilities. The analysis presented in this work pictures them as extremely massive (M~10^11 Msun) sources at z~5-7. The existence of such massive galaxies in the first Gyr of cosmic history suggests a much more efficient star formation in the early Universe than what is commonly derived by observing low-z galaxies. These two approaches, combined, shed new light on the evolutionary paths that bring to the formation of massive galaxies in the low-z Universe.