Hidden complexity in defect-assisted nonradiative capture during energy conversion

Abstract

Defect-assisted nonradiative recombination can drastically reduce the efficiency of electronic and optoelectronic devices. Conventionally, a common approach to judge the likelihood of a defect acting as a strong nonradiative recombination center is to examine the relative position of the charge-state transition level of the defect in the band gap. By performing rigorous first-principles calculations for the nonradiative hole capture coefficient of the carbon-on-nitrogen (C_N) defect in the Al_0.5Ga_0.5N alloy, we show that this widely used approach is oversimplified and may lead to misleading conclusions. Unexpectedly, the capture coefficient shows almost no correlation with the transition energy (i.e., the energy difference between the defect transition level and the band edges). Instead, we find that the semiclassical energy barrier for the capture process is actually the decisive factor for the capture coefficient; it is influenced not only by the transition energy, but also by other competing factors such as the vibrational frequencies and the lattice relaxations associated with the nonradiative capture process. Our results demonstrate that assessing the nonradiative capture coefficient solely based on transition levels is unreliable, highlighting the necessity to explicitly compute the defect-assisted nonradiative capture coefficient from first principles.