Control of defect processes in photovoltaic materials is essential for realising high-efficiency solar cells and related optoelectronic devices. The concentrations of native defects and extrinsic dopants tune the Fermi level and enable semiconducting p-n junctions; however, fundamental limits to doping exist in many compounds. Optical transitions involving defect states can enhance photocurrent generation through sub-bandgap absorption; however, such states are often responsible for carrier trapping and non-radiative recombination events that limit open-circuit voltage. Many classes of materialsincluding metal oxides, chalcogenides, and halidesare being examined for next-generation solar energy applications, and each technology faces distinct challenges that could benefit from point defect engineering. We review the evolution in point defect behaviour from Si-based photovoltaics to thin-film CdTe and Cu(In,Ga)Se2 technologies, through to the latest generation halide perovskite (CH3NH3PbI3) and kesterite (Cu2ZnSnS4) devices. We focus on the chemical bonding that underpins the defect chemistry, and the atomistic processes associated with the photophysics of charge carrier generation, trapping, and recombination in solar cells. Finally, we outline general principles to enable defect control in complex semiconducting materials.

The semiconducting behaviour and optoelectronic response of gallium nitride is governed by point defect processes, which, despite many years of research, remain poorly understood. The key difficulty in the description of the dominant charged defects is determining a consistent position of the corresponding defect levels, which is difficult to derive using standard supercell calculations. In a complementary approach, we take advantage of the embedded cluster methodology that provides direct access to a common zero of the electrostatic potential for all point defects in all charge states. Charged defects polarise a host dielectric material with long-range forces that strongly affect the outcome of defect simulations; to account for the polarisation we couple embedding with the hybrid quantum mechanical/molecular mechanical (QM/MM) approach and investigate the structure, formation and ionisation energies, and equilibrium concentrations of native point defects in wurtzite GaN at a chemically accurate hybrid-density-functional-theory level. N vacancies are the most thermodynamically favourable native defects in GaN, which contribute to the n-type character of as-grown GaN but are not the main source, a result that is consistent with experiment. Our calculations show no native point defects can form thermodynamically stable acceptor states. GaN can be easily doped n-type, but, in equilibrium conditions at moderate temperatures acceptor dopants will be compensated by N vacancies and no significant hole concentrations will be observed, indicating non-equilibrium processes must dominate in p-type GaN. We identify spectroscopic signatures of native defects in the infrared, visible and ultraviolet luminescence ranges and complementary spectroscopies. Crucially, we calculate the effective-mass-like-state levels associated with electrons and holes bound in diffuse orbitals. These levels may be accessible in competition with more stronglylocalised states in luminescence processes and allow the attribution of the observed 3.46 and 3.27 eV UV peaks in a broad range of GaN samples to the presence of N vacancies.