Defect calculations in Ga-based semiconductors using optimal hybrid functionals

点击翻译按钮获取中文摘要

The functionality of semiconductors is closely connected to their point defects, which control the electronic and optical behavior of the material. The emphasis of defect physics has shifted in recent years from micro- to optoelectronics and photovoltaics, i.e., from the traditional elemental and binary semiconductors, with relatively narrow gap, to wide band gap materials of more complicated structures. Electronic structure calculations in these materials have revealed that local and semi-local approximations to density functional theory (DFT) cannot be trusted: partly because the bigger error in the width of the band gap but also because of the artificial delocalization of defect states (which, make the defect level shallower and preclude the reproduction of small polaron states). While first-principles total energy calculations beyond (semi)local DFT cannot yet be carried out for supercells, we have shown that the HSE06 screened hybrid functional yields very accurate results for defects in C, Si, SiC, Ge and TiO2. We have suggested that the success is connected to the fact that HSE06 in these materials provides the total energy as an appropriate piece-wise linear function of the occupation numbers, because of error compensation between semi-local and Hartree-Fock exchange.Gallium based semiconductors have immense technological importance: besides the GaN LEDs, Ga2O3 has also great potential for application as UV-transparent electrode and se-miconductor for power MOSFETs, and Ga-based chalcopyrites (CuGaS2, CuGaSe2) are considered in photovoltaics. HSE06 underestimates the band gap of these materials and we have recently shown that simple tuning of the mixing parameter to correct that (as usual in the literature), does not lead to correct defect levels. Tuning the mixing and the screening parameters, to reproduce the gap and to fulfill the generalized Koopman theorem, has allowed us to achieve the correct piece-wise linear behavior of the total energy as a function of the occupation numbers and, consequently, the accurate description of defect states. Of course, this procedure leads to material specific parameters. The aim of this project is twofold: on the one hand we would like to investigate the trends in the optimal parameters across the Ga-based semiconductors mentioned above (but also for CuInS2 and CuInSe2), in order to establish general rules for finding a screened hybrid for a class of materials and, on the other, perform accurate calculations to identify important defect centers in these materials.