<ul><li>Self-driving labs (SDLs) utilizing automation and machine learning have potential in accelerating experimental procedures for material discovery.</li><li>This work introduces an SDL based on magnetron co-sputtering, enabling accurate composition mapping on multi-element thin films in-situ, without the need for time-consuming ex-situ analysis.</li><li>The method employs machine learning techniques, including active learning using Gaussian processes, to predict composition distribution in combinatorial thin films based on in-situ measurements from sensors in the sputter chamber.</li><li>The framework enhances efficiency by eliminating the requirement for extensive characterization or calibration, showcasing the ability of ML-guided SDLs to expedite materials exploration.</li></ul>

Machine learning for in-situ composition mapping in a self-driving magnetron sputtering system

Discover more