Supplementary material for: A Transformer network for high-throughput materials characterization with X-ray photoelectron spectroscopy

This study presents a Transformer-based neural network designed to predict chemical concentrations from X-ray photoelectron spectroscopy (XPS) survey spectra of homogeneous bulk inorganic and organic compounds. We propose a workflow for generating a representative synthetic XPS dataset, complemented by experimentally informed data augmentation strategies. The resulting dataset is employed for training, validation, and testing of the Transformer architecture, with the ultimate objective of developing an expert system to support quantitative XPS analysis.