Revealing the Role of Confined Molecular H$_2$ in the Passivation of Defective Silicon Using First-Principles Simulations

The passivation of silicon dangling bonds by hydrogen is a crucial requirement for silicon-based optoelectronic technology, especially for solar cells. Recent experiments on intense light soaking of silicon heterojunction solar cells unveiled interesting dynamical aspects of hydrogen passivation that are linked to Si-H bond breaking and repassivation. These processes take place predominantly in porous regions near the amorphous/crystalline interface, where hydrogen can exist in molecular form. This work addresses the question of whether molecular H$_2$ directly participates in Si-H depassivation and repassivation. Using density functional theory, we calculate and compare formation energies of point defects, multivacancy cavities and the Si(100) surface to identify relevant passivated and depassivated states. Furthermore, we employ nudged elastic band calculations to determine the activation barriers of the corresponding pathways. We find that sufficient local free volume enables a direct double-H pathway for depassivation through the formation of confined molecular H$_2$. Despite involving the breaking of two Si-H bonds, the double-H process can be energetically and kinetically competitive with the single-H process and can exhibit a reverse repassivation barrier as low as $0.15\,$eV under p-type conditions. These findings provide a plausible atomistic explanation for passivation recovery during light soaking and illuminated annealing in porous regions near amorphous/crystalline silicon interfaces.