Origin of the discrepancy between the fundamental and optical gaps and native defects in two dimensional ultra-wide bandgap semiconductor: Gallium thiophosphate

Abstract

Ultra-wide bandgap (UWBG) semiconductors have great potential for high-power electronics, radio frequency electronics, deep ultraviolet optoelectronic devices, and quantum information technology. Recently, the two-dimensional UWBG GaPS4 was first applied to the solar-blind photodetector in experiments, which was found to have remarkable performance, such as high responsivity, high quantum efficiency, etc., and promising applications in optoelectronic devices. However, the knowledge of monolayer (ML) GaPS4 for us is quite limited, which hinders its design and application in optoelectronic devices. Here, we focus on the properties of electronic structure and intrinsic defects in ML GaPS4 by first-principles calculations. We confirmed that the fundamental gap of ML GaPS4 is 3.87 [Formula: see text], while the optical gap is 4.22 [Formula: see text]. This discrepancy can be attributed to the inversion symmetry of its structure, which limits the dipole transitions from valence band edges to conduction band edges. Furthermore, we found that intrinsic defects are neither efficient p-type nor n-type dopants in ML GaPS4, which is consistent with experimental observations. Our results also show that if one expects to achieve p-type ML GaPS4 by selecting the appropriate dopant, P-rich conditions should be avoided for the growth process, while for achieving n-type doping, S-rich growth conditions are inappropriate. This is because due to the low strain energy, [Formula: see text] has very low formation energy, which leads to the Fermi levels ([Formula: see text]) pinning at 0.35 [Formula: see text] above the valence band maximum and is not beneficial to achieve p-type ML GaPS4 under the P-rich conditions; the large lattice relaxation largely lowers the formation energy of [Formula: see text], which causes the [Formula: see text] pinning at 0.72 [Formula: see text] below the conduction band minimum and severely prevents ML GaPS4 from being n-type doping under the S-rich conditions. Our studies of these fundamental physical properties will be useful for future applications of ML GaPS4 in optoelectronic devices.