Mapping correlations between structural and plasmonic properties in SrSnO₃ through Nanoscale Imaging and Spectroscopy with Machine learning Assistance (NISMA)

Strontium stannate (SrSnO₃) is an ultra-wide bandgap oxide semiconductor suitable for applications in high-power electronics. Thin films of SrSnO₃ were perfected through epitaxial strain to promote the high mobility structural phase. Even in epitaxially strained films, competing structural phases impair mobility [1]. Near-field microscopy enables nano-imaging and spectroscopy at the spatial scales of phase coexistence. Using polarization-resolved laser-based infrared microscopy with nanoscale microscopy and spectroscopy applied to films of La-doped SrSnO₃, we identify how structural phases in SrSnO₃ self-organize and resolve their vibrational and electronic properties. Further, we develop and deploy Nanoscale Imaging and Spectroscopy with Machine learning Assistance (NISMA), a novel method to spatially map correlations between structural and plasmonic properties in dense nanometer-resolved datasets. Our results challenge basic expectations that structural phases uniquely determine local electronic properties in these films at the nanoscale. Rather, NISMA implies that latent properties including arrangement of structural domain walls and accommodation strain are mutually influential factors impacting nanoscale electronic properties of these films. Our study lays the groundwork for promoting desired macroscopic behaviors in wide-bandgap semiconductor films by exposing their rich characteristics at the nanoscale.



[1] Truttmann, T.K., et al. Commun Phys 4, 241 (2021).