Model Based Predictive Control of Thin Film Synthesis with Pulsed Laser Deposition

The last decade has seen substantial investment into methods that incorporate machine learning and high-throughput workflows in materials science. Years of success on the computational front in using artificial intelligence (AI) to quickly identify new interesting material systems is tempered by the same old issue: predicting new materials is straightforward compared to the experimental validation through some synthesis modality. Moreover, many predicted materials are only metastable, which requires that synthesis trajectories be carefully controlled during growth to drive the system towards the desired metastable state. Only sporadic effort has been made in the past 30 years towards intelligent real-time control of synthesis in the science domain due to the lack of alignment between experiments and theory. Here, we demonstrate an approach to enable the real-time control of thin film synthesis through a combination of in situ optical diagnostics and a recently developed Bayesian state estimation method. We developed a physical model for thin film growth in pulsed laser deposition (PLD) and applied the direct filter method for real-time estimation of nucleation and growth rates during PLD of transition metal dichalcogenides. We validated the approach on simulated and previously acquired reflectivity data for WSe₂ growth and ultimately deployed the algorithm on an autonomous PLD system during growth of 1T’-MoTe₂ under various synthesis conditions. We found that the film growth parameters can be robustly determined in real-time at very early stages of growth, down to 15% monolayer area coverage. This approach opens new opportunities for adaptive film growth control based on a fusion of in situ diagnostics, modern data assimilation methods, and physical models which promises to enable control of synthesis trajectories towards desired material states.

This work was supported by the Center for Nanophase Materials Sciences (CNMS), which is a US Department of Energy, Office of Science User Facility at Oak Ridge National Laboratory. Materials synthesis was supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Materials Sciences and Engineering Division.