Rapid and Accurate Thin Film Thickness Extraction via UV-Vis and Machine Learning

Thin-film processes are ubiquitous in photovoltaics research and are increasingly incorporated into high-throughput experimentation (HTE) equipment. However, HTE is limited by the slowest steps, and accurate thickness measurements have emerged as a bottleneck. This study demonstrates rapid yet accurate thin-film thickness extraction by leveraging machine learning (ML) in combination with non-destructive optical measurements (UV-Vis). We achieve 86.9% accuracy of thickness prediction within 10-percentage-error bounds on simulated data.