Tokyo University of Science researchers have developed an exciting new AI-based approach for conducting detail X-ray absorption spectroscopy (XAS) analysis, paving the way for streamlined analysis of semiconductors, Internet-of-Things devices, and next generation energy storage. Although XAS provides an incredibly detailed look into the electronic state of these materials, current analysis techniques require extensive expertise and manual labor.
Published in the journal Scientific Reports, the AI based method provides a streamlined approach for XAS analysis, an analysis technique often hindered by a lack of reference data and subjectivity of interpretations during analysis.
"AI-based data-driven methods, such as machine learning, can be powerful tools for efficiently analyzing and interpreting measurement data, providing objective insights," explained Professor Masato Kotsugi.
In their study, the team tested four different machine learning techniques with the Uniform Manifold Approximation and Projection (UMAP) technique standing out by performing exceptionally well in classifying complex structural data. Not only could the technique identify global trends, it was also able to detect subtle differences in phases and defect types.
"Our findings show that UMAP can be a valuable tool for rapid, scalable, automated, and importantly, objective material identification using complex experimental spectral data," added Prof. Kotsugi.
The method will soon be implemented as software at the Nano-Terasu synchrotron radiation center, where the AI-based approach will be used to accelerate the development of novel materials in key fields such as semiconductors, catalysis, and energy storage.
"Our method demonstrates the potential of autonomous structural identification, opening up new possibilities for data-driven material design and development of novel materials.," concluded Prof. Kotsugi.