Seoul National University
2DEG state at BaSnO3/LaInO3 interface created by interface polarization
After describing all the experimental data on the conductance enhancement at BaSnO3 (BSO)/LaInO3 (LIO) interface, we will present analysis of 2DEG state at BSO/LIO interface by 1-dimensional Poisson-Schrödinger equation. Experimental data to analyze are the n2D of BSO/LIO interface as a function of LIO thickness; the n2D peaks at 4 unit cell thickness of LIO and then decreases slowly by about 25 % to a constant value. In our analyses, the required materials parameters are the effective masses, the dielectric constants, and the conduction band offset between BSO and LIO. In addition, as a way to accommodate the polar discontinuity, we took into account a polarization in LIO.
We will show first that any uniform polarization in LIO cannot explain the observed n2D data of BSO/LIO. When the uniform polarization is 50 μC/cm2, the n2D approaches 3x1014 /cm2, the value predicted by the “charge discontinuity” model, in the thick LIO limit. The only way to generate the experimental n2D is by assuming a highly non-uniform polarization, namely the “interface polarization” model. In such a model, only a couple of unit cells of LIO at the interface possess high polarization of about 50 μC/cm2 and then the polarization decays to zero in the next two unit cells. Using such an “interface polarization” model we could reproduce the experimental n2D. We will also discuss the effect of deep level states (donors or acceptors) in wideband gap oxide heterostructures. By combining the “interface polarization” with the deep level states, we are able to explain the entire LIO thick dependence, La-doping dependence, and the substrate dependence.
The origin for the “interface polarization” will be further discussed in detail. Although the primary cause of the “interface polarization” is the epitaxial strain, the details of how the epitaxial strain at the interface causes the polarization seems complicated. We will present the recent experimental and theoretical evidences for the “interface polarization”.
Using such BSO/LIO 2DEG states, all perovskite transparent FETs with excellent properties were made. We believe that the stable oxygen stoichiometry and the ability to control the local doping level in LaInO3/BaSnO3 heterostructures will lead to more complete understanding of its 2DEG behavior, important in creating new science and technology.