Joint first principles and experimental investigation on the improved performance of the Ta2O5 interface layer for an MoS2/WS2 channel

Transition metal dichalcogenides are attractive channel materials for an aggressive dimension scaling of double-gated field-effect transistors. Some challenges, however, remain nowadays, including finding a suitable top high-κ dielectric and a way to deposit it. Recently, an amorphous Ta2O5 interface layer grown on top of MoS2 has been shown to improve the device's electrical performance compared to a deposited amorphous Al2O3 film. In this work, we investigate the origin of such an improvement, combining experimental and first-principles approaches. The former suggests a comparable n-type doping of the 2D material for the Al2O3 and Ta2O5 interface layer based on x-ray photoelectron spectroscopy. Within the first-principles approach, we construct several amorphous Ta2O5 atomistic models and their interfaces with a MoS2 or WS2 monolayer. We demonstrate that a clean interface can in principle be achieved between the 2D material and Ta2O5 with a proper surface passivation, with a defect-free electronic bandgap a potential fluctuation within the 2D material under control. Little differences are found between MoS2 and WS2 in the corresponding interface at the theoretical level. The electronic properties are found to strongly depend on the coordination of the metal atoms at the oxide surface. Combining these analyses, we conclude that a higher interface quality for the Ta2O5 layer with MoS2 than for the Al2O3 one is a plausible explanation for the improved performances.