In-situ Raman investigation of Laser-Induced Graphene using Machine Learning

High-quality graphene was laser-induced from precursors graphene oxide and commercial polyimide. Laser annealing is a promising method to create graphene-based devices including sensors, biomedical equipment and thin-film transistors. A Bayesian optimization-based machine-learning strategy was used to predict optimal process parameters. We custom-built our system to allow simultaneous laser patterning processing and in-situ Raman spectroscopy characterizations. The Raman G/D ratio is a good indication of the quality of the laser-induced graphene. Rapid and significant improvements in the G/D ratios were seen with the machine-learning predicted process parameters. This experimental setup has enormous potential in Autonomous Research Systems for new materials discovery and can be scaled up for advanced-manufacturing patterned graphene-based electronics.