Beyond manual processing: Robotic-arm automation coupled with machine learning for precision control and correlation insights in perovskite solar cells

Spin coating, widely used in perovskite solar cell research, is inherently limited by manual processing, leading to uncontrollable process parameters. A major challenge lies in the anti-solvent dripping process, where precise control of dripping speed by manual methods is impractical. Consequently, quantitative analysis of thin films and device performance based on dripping speed has remained challenging, despite its significant impact on film quality. To address this limitation, we developed a fully customized robotic arm-based system that precisely mimics human motion while enabling quantitative control of previously unmanageable parameters. This automated system systematically explores the effects of various process variables, including anti-solvent chemical properties, through machine learning-driven data analysis.