Uncovering the hidden universe: luminous type-2 QSOs at cosmic noon and beyond

Abstract

This thesis investigates the population, physical properties, and environments of luminous obscured quasars at cosmic noon and beyond. A mid-infrared selection method combining SDSS, DESI, and WISE photometry was developed to identify optically faint, extremely red sources. Follow-up spectroscopy with Gemini and Keck achieved an 83% success rate, confirming 53 new Type-2 quasars, 75% of them at redshift z > 2. Their spectral energy distributions reveal extreme bolometric luminosities, massive black holes, and similarities to JWST little red dots. The sample shows substantial spectral diversity, including broad H-alpha emission in about one-third of optically classified Type-2 quasars, indicating that orientation alone cannot explain the Type-1/Type-2 distinction. Keck/KCWI and ALMA observations further reveal rich gaseous environments: 80% of the targeted quasars host giant Lyman-alpha nebulae, and several occupy massive halos with companion galaxies. The thesis also explores artificial intelligence in astronomy through a convolutional neural network for detecting damped Lyman-alpha systems in DESI spectra and a benchmark for evaluating physical reasoning in large multimodal models.