Effects of intermixing and oxygen vacancies on a two-dimensional electron gas at the polar (TbScO3/KTaO3) (001) interface

3d-5d perovskite oxides, ABO(3) (where A and B are 3d or 5d elements), form polar surfaces in the (001)stacked thin films. As a result, the polar-polar (001) interface between two ABO(3) insulators could create polar discontinuity potentially producing a two-dimensional electron gas of higher density and stronger spatial localization compared to the widely studied polar-nonpolar oxide interfaces, such as (001) LaAlO3/SrTiO3. Here, as a model system, we explore the interface between polar (001) TbScO3 and polar (001) KTaO3 using first-principles density functional theory. We find that the intermixed interface Ta0.75Sc0.25O2/Tb0.75K0.25O maintaining the bulk perovskite charge stacking (e.g., ... + 1/ - 1/ + 1 ...) is insulating and has a lower energy than the metallic interface TbO/TaO2 breaking such stacking. This intermixed interface is, however, prone to the formation of oxygen vacancies which make it conducting. We emphasize that the driving force for the formation of the two-dimensional electron gas (2DEG) here is not a built-in electric field stemming from the polar discontinuity but the interface stoichiometry. We find that the ratio of oxygen vacancy concentration is a factor of 30 times larger at the interface than in bulk KTO at room temperature. The oxygen vacancy-induced 2DEG resides on the Ta-5d electronic orbitals with d(xy) and d(xz)/(yz) occupation dominating overall charge density near and far away from the interface.

Automated summary

This article investigates the polar interfaces in 3d-5d perovskite oxides and their potential to create a two-dimensional electron gas. The study shows that at certain interfaces, oxygen vacancies can lead to the formation of a highly localized and dense 2DEG. The driving force for this phenomenon is the interface stoichiometry, rather than the polar discontinuity.