Layer-resolved band bending at the n-SrTiO3(001)/p-Ge(001) interface

TL;DR

This study investigates the layer-specific band bending at the n-SrTiO3(001)/p-Ge(001) heterojunction using advanced x-ray photoelectron spectroscopy and theoretical modeling, revealing significant band alignment effects and implications for water splitting.

Contribution

It introduces a novel layer-resolved analysis method for band bending at the heterojunction interface and combines experimental and theoretical approaches for comprehensive understanding.

Findings

Layer-resolved mapping of band bending in Ge

Significant reduction in valence band offset due to Ge band bending

Good agreement between experimental data and DFT-based interface models

Abstract

The electronic properties of epitaxial heterojunctions consisting of the prototypical perovskite oxide semiconductor,n-SrTiO3 and the high-mobility Group IV semiconductor p-Ge have been investigated. Hard x-ray photoelectron spectroscopy with a new method of analysis has been used to determine band alignment while at the same time quantifying a large built-in potential found to be present within the Ge. Accordingly, the built-in potential within the Ge has been mapped in a layer-resolved fashion. Electron transfer from donors in the n-SrTiO3 to the p-Ge creates a space-charge region in the Ge resulting in downward band bending which spans most of the Ge gap. This strong downward band bending facilitates visible-light, photo-generated electron transfer from Ge to STO, favorable to drive the hydrogen evolution reaction associated with water splitting. Ti 2p and Sr 3d core-level line shapes reveal that the STO bands are flat despite the space-charge layer therein. Inclusion of the effect of Ge band bending on band alignment is significant, amounting to a ~0.4 eV reduction in valence band offset compared to the value resulting from using spectra averaged over all layers. Density functional theory allows candidate interface structural models deduced from scanning transmission electron microscopy images to be simulated and structurally optimized. These structures are used to generate multi-slice simulations that reproduce the experimental images quite well. The calculated band offsets for these structures are in good agreement with experiment.