Towards Autonomous Physical Vapor Deposition of Inorganic Thin Film Materials

Growth and characterization of thin film materials require long experimental campaigns to tune the processing conditions until desired material properties and chemical composition are achieved. If devices are under study, the cycle of synthesis, characterization and optimization is stretched even further, since now phenomena occurring at the materials interfaces play a significant role.<br/>In this context, autonomous experimentation is transforming the landscape of synthesis and characterization of materials and devices. The incorporation of automation of repetitive tasks and artificial intelligence techniques to plan experiments based on the available data accelerate the materials discovery chain while minimizing human intervention, freeing researchers to focus on more strategically relevant tasks.<br/>In this presentation, I will present the progress on autonomous experimentation at NREL for synthesis of inorganic thin films by physical vapor deposition, namely sputtering¹ and molecular beam epitaxy², all enabled by automated instruments driven by artificial intelligence algorithms. Equipped with sophisticated capabilities such as multidimensional Bayesian analysis and computer vision techniques, these systems intelligently navigate through large parameter spaces, such as substrate and effusion cell temperature, power, pressure, gas flow and shutter times.<br/>I will also discuss the many challenges and solutions encountered during the adaptation of existing laboratory infrastructure to accommodate autonomous workflows, such as instrument automation, workflow orchestration, data management, software development, and establishment of sub networks for communication between a remote server and command-and-control computers.<br/><br/>1. Fébba, D. M. <i>et al.</i> Autonomous Sputter Synthesis of Thin Film Nitrides with Composition Controlled by Bayesian Optimization of Optical Plasma Emission. <i>APL Materials</i> vol. 11 (2023).<br/>2. Schaefer, S. <i>et al.</i> Rapid screening of molecular beam epitaxy conditions for monoclinic (InxGa1−x)2O3 alloys. <i>J. Mater. Chem. A</i> (2024) doi:10.1039/D3TA07220G.