Machine Learning

AI-Driven A-Lab: Synthesizing Inorganic Materials With Robotics And ML

Machine Learning

AI-Driven A-Lab: Synthesizing Inorganic Materials With Robotics And ML

An autonomous laboratory for the accelerated synthesis of novel materials UC Berkeley and Berkeley Lab established A-Lab, which merges robotics, databases, machine learning, and literature data for autonomous optimization of inorganic powder synthesis. The A-Lab planned and interpreted the outcomes of experiments performed by robots. The Materials Project identified novel, air-stable targets. Machine learning natural-language models trained on scientific literature proposed synthesis recipes. A fully-automated robotic laboratory tested the recipes by performing 1) powder dosing, 2) sample heating, and 3) product characterization via XRD. The sample purity was assessed via XRD and analyzed by ML models. When samples were impure (< 50% target yield), the A-Lab proposed novel synthesis recipes based on a thermodynamics-based active-learning algorithm. A-Lab's success highlights AI's potential in material discovery, bridging computational and experimental approaches.

The A-Lab, an autonomous laboratory, synthesized 41 novel inorganic compounds out of 58 targets over only 17 days.

Academia | P. 232 Source: N.J. Szymanski, B. Rendy, Y. Fei, R.E. Kumar, T. He, D. Milsted, M.J. McDermott, M. Gallant, E. Dogus Cubuk, A. Merchant, H. Kim, A. Jain, C.J. Bartel, K. Persson, Y. Zeng & G. Ceder

Extracted from the PDF; charts and images may read imperfectly.