As society pursues sustainable alternatives to conventional semiconductors, halide perovskites have emerged as exceptionally promising materials for next-generation optoelectronics. Their remarkable rise from laboratory curiosities to record-breaking solar cells and efficient light-emitting devices reflects unique optoelectronic properties – long carrier diffusion lengths, defect tolerance, and compositional tunability – that distinguish them from traditional semiconductors. However, realizing their full technological potential requires overcoming critical challenges: the toxicity of lead-based formulations, fundamental limits imposed by many-body recombination processes, and incomplete understanding of interface-controlled carrier dynamics. These complex, multiscale phenomena demand theoretical approaches capable of bridging atomic-scale mechanisms with device-level performance. Modern first-principles methods, from density functional theory to many-body perturbation theory, provide the predictive framework necessary to unravel these challenges and guide rational materials design. In this thesis, I present comprehensive ab initio investigations of halide perovskites and double perovskites that resolve device puzzles, quantify fundamental performance limits, and unlock new multifunctional capabilities.

For conventional halide perovskites, I addressed two critical device-performance-limiting phenomena through complementary theoretical frameworks. First, I explained the apparent contradiction between large hole injection barriers observed in surface spectroscopy and high-efficiency perovskite LED performance by demonstrating that surfactant additives create interface dipoles inducing near-surface hole accumulation that locally rebalances carrier populations despite nominal energy barriers. This interfacial mechanism establishes design principles for optimizing carrier balance through additive chemistry. Second, I developed a comprehensive many-body treatment of Auger recombination incorporating dynamic dielectric screening for the first time in halide perovskites. Through frequency-dependent screening implemented via low-scaling GW calculations, I revealed that conventional static approximations overestimate Auger coefficients by 50-60%, shifting the radiative-Auger crossover to nearly twice the carrier density and substantially extending the high-brightness operating window. Together, these advances provide both interface-level control and quantitative many-body limits essential for accurate device modeling.

Leveraging the expanded chemical space of halide double perovskites enabled exploration of multifunctional properties beyond conventional optoelectronics. Systematic high-throughput screening of transition-metal-substituted compositions mapped magnetic phase spaces and identified rare ferromagnetic outliers with half-metallic character – Cs₂AgNiCl₆, Cs₂AuNiCl₆, and Cs₂HgCrCl₆ – despite general antiferromagnetic tendencies in halide perovskite-based frameworks. The analysis revealed how extended superexchange pathways and ionic bonding suppress magnetic interactions compared to oxide analogues, while specific orbital configurations enable robust spin polarization. Complementary investigations of sub-bandgap engineering through palladium doping created optically active intragap states extending photoresponse to 1380 nm with experimental validation by collaborators, establishing transferable design principles for spectral extension without bandgap modification. Finally, through careful implementations ensuring phase-consistent exciton-phonon coupling calculations, I applied this framework to distinguish dynamic phonon dressing from static self-trapping in wide-gap systems, demonstrating that Cs₂NaYCl₆ exhibits self-trapped exciton character with strong underlying exciton-phonon coupling driving the system toward localization. These studies collectively demonstrate how double perovskite structural and chemical flexibility enables targeted design of magnetic, optical, and excitonic functionalities within environmentally benign frameworks.