Table of Contents
Fetching ...

Decoupling Composition and Band Gap in $κ$-Ga$_2$O$_3$ Heterostructures via STEM-EELS

Annett Thøgersen, Georg Muntingh, Lasse Vines, Øystein Prytz, Max Kneiß, Marius Grundmann, Holger von Wenckstern, Ingvild J. T. Jensen

TL;DR

This work tackles whether interfacial band-gap variations in oxide heterostructures track compositional changes or are dominated by strain and measurement artifacts. It combines monochromated, probe-corrected STEM-EELS with an automated quantitative analysis framework to map band-gap excitations at nanometer scales and to quantify inelastic-delocalization effects. Applied to $κ$-Ga$_2$O$_3$ heterostructures grown on ITO and ZnO templates, it reveals a strain-driven gradient from $5.08$ eV to $4.28$ eV over ~ $10$ nm in the defect-free ITO case, while ZnO templates facilitate strain relaxation and gaps that align with composition. Overall, the study demonstrates decoupling of composition and local electronic structure at oxide interfaces, establishing STEM-EELS with automated analysis as a robust tool for nanoscale electronic characterization of wide-band-gap materials.

Abstract

High-resolution mapping of electronic properties at oxide heterointerfaces remains challenging due to probe delocalization and overlapping signals. In this work, we employ monochromated, probe-corrected scanning transmission electron microscopy combined with electron energy-loss spectroscopy (STEM-EELS) to resolve band gap variations across $κ$-Ga$_2$O$_3$-based multilayers with nanometer-scale precision. A custom automated quantitative-based EELS analysis framework enabled automated band gap fitting and visualization, ensuring reproducibility and high spatial resolution. By optimizing acquisition parameters and quantifying inelastic delocalization, we demonstrate reliable extraction of band gap excitations from layers only a few nanometers thick. For heterostructures grown on ITO templates, strain at defect-free interfaces induces a gradual band gap transition from $5.08~\mathrm{eV}$ to $4.28~\mathrm{eV}$ over $\sim 10~\mathrm{nm}$, despite an abrupt compositional change. In contrast, ZnO-based templates introduce structural defects that relieve strain, yielding band gaps consistent with composition. These results establish STEM-EELS as a powerful tool for nanoscale electronic characterization and highlight the dominant role of interfacial strain over composition in governing local band structure.

Decoupling Composition and Band Gap in $κ$-Ga$_2$O$_3$ Heterostructures via STEM-EELS

TL;DR

This work tackles whether interfacial band-gap variations in oxide heterostructures track compositional changes or are dominated by strain and measurement artifacts. It combines monochromated, probe-corrected STEM-EELS with an automated quantitative analysis framework to map band-gap excitations at nanometer scales and to quantify inelastic-delocalization effects. Applied to -GaO heterostructures grown on ITO and ZnO templates, it reveals a strain-driven gradient from eV to eV over ~ nm in the defect-free ITO case, while ZnO templates facilitate strain relaxation and gaps that align with composition. Overall, the study demonstrates decoupling of composition and local electronic structure at oxide interfaces, establishing STEM-EELS with automated analysis as a robust tool for nanoscale electronic characterization of wide-band-gap materials.

Abstract

High-resolution mapping of electronic properties at oxide heterointerfaces remains challenging due to probe delocalization and overlapping signals. In this work, we employ monochromated, probe-corrected scanning transmission electron microscopy combined with electron energy-loss spectroscopy (STEM-EELS) to resolve band gap variations across -GaO-based multilayers with nanometer-scale precision. A custom automated quantitative-based EELS analysis framework enabled automated band gap fitting and visualization, ensuring reproducibility and high spatial resolution. By optimizing acquisition parameters and quantifying inelastic delocalization, we demonstrate reliable extraction of band gap excitations from layers only a few nanometers thick. For heterostructures grown on ITO templates, strain at defect-free interfaces induces a gradual band gap transition from to over , despite an abrupt compositional change. In contrast, ZnO-based templates introduce structural defects that relieve strain, yielding band gaps consistent with composition. These results establish STEM-EELS as a powerful tool for nanoscale electronic characterization and highlight the dominant role of interfacial strain over composition in governing local band structure.

Paper Structure

This paper contains 9 sections, 2 equations, 9 figures, 1 table.

Figures (9)

  • Figure 1: A) DF-STEM image of the 3-layered Ga$_2$O$_3$ on ITO substrate, B) DF-STEM image of doped, multi-layered Ga$_2$O$_3$ on ZnO, C) HR-STEM image of the first interface in the 3-layered Ga$_2$O$_3$ on ITO sample, D) HR-STEM image of an interface in the multi-layered Ga$_2$O$_3$ on ZnO, E) GPA strain map ($\varepsilon_{xx}$) of C), F) GPA strain map ($\varepsilon_{xx}$) of D), G) Strain profiles extracted from GPA for the two samples, showing variations across the heterointerfaces. Blue: 3-layered Ga$_2$O$_3$ on ITO; Green: multi-layered Ga$_2$O$_3$ on ZnO.
  • Figure 2: A) BF-STEM image of the 3-layered Ga$_2$O$_3$ on ITO substrate, B) band gap map of the same region, C) fitted mean band gap (black) and corresponding $\pm 1$ standard deviation (gray) calculated row-wise from the EELS spectrum image, and D) fitted mean band gap across the two first layers from the line scan EELS data.
  • Figure 3: Fitted band gap of the A) (Al$_{0.27}$Ga$_{0.73})_2$O$_3$ layer, B) (In$_{0.18}$Ga$_{0.82})_2$O$_3$, C) Ga$_2$O$_3$, and D) ITO layer.
  • Figure 4: A) The signal from the line scan of the 3-layered Ga$_2$O$_3$ on ITO substrate, within the energy window of 4.2 eV to 6.8 eV (black curve) plotted with the intensity derived from the ADF image (red curve) near the protective carbon layer, with the ADF image as an inset. B) The EELS low loss signal from the red and blue region in (A).
  • Figure 5: A) BF STEM image showing the location of the two spectra. B) Low loss EELS signal with the fitted band gap at 19 nm into the (Al$_{0.27}$Ga$_{0.73})_2$O$_3$ layer, at the point where the band gap of both layers are visible (band gap of (In$_{0.18}$Ga$_{0.82})_2$O$_3$ shown with black arrow). C) EELS spectrum at 24.1 nm, inside the (In$_{0.18}$Ga$_{0.82})_2$O$_3$ layer. This is the point where only the band gap of this layer is visible.
  • ...and 4 more figures