Methods for Rapid and Smart Characterization of Vanadium Oxide Thin Films and Thermally Aged Samples Using Raman Spectroscopy and Multivariate Analysis

Vanadium oxides (VOxs) are technologically important materials because of their versatile properties, including a reversible metal-insulator transition. However, the complex vanadium-oxygen phase diagram includes numerous stable stoichiometries and metastable polymorphs, making material identification challenging. This work presents a data-driven, nondestructive characterization workflow that combines Raman spectroscopy and principal component analysis (PCA) to provide rapid and intelligent feedback during VOx thin-film fabrication. Although Raman spectroscopy provides ultrasensitive chemical information, conventional spectral interpretation requires specialized expertise and extensive reference data. An in-house fabrication protocol based on spin-casting chemical precursors and thermal annealing was developed to produce thin-film standards of VO2, V2O5, V3O7, and V6O13. Raman fingerprints from these standards were used to construct two- and three-dimensional PCA models that distinguish stoichiometric variation and provide interpretable feedback during fabrication-process development. The standards also serve as references for evaluating thermally corroded samples. These results demonstrate that PCA-enabled Raman analysis can serve as an adaptive sensing approach for rapid assessment of VOx stoichiometry, simplified spectral interpretation, and fabrication optimization.