Integrating high-throughput and machine learning methods to accelerate discovery of molten salt corrosion-resistant alloys

Prof. Jason Hattrick- Simpers and his collaborators in University of Wisconsin and Argonne National Lab have demonstrated the first approach that combines high-throughput experiments, computational modeling, and machine learning to systematically study the corrosion resistance of high entropy alloys in molten salt environments.

“By employing explainable machine learning models we were able to not only model the corrosion resistance but to also identify a novel sacrificial mechanism which could be used to protect structurally important elements from being dissolved by the salts,” says Prof. Hattrick-Simpers.

These findings were recently published in the journal of Advanced Science.

Schematic of the developed HTP and automated methods for corrosion-resistant alloy development.