PNNL

Abstract

Multi-principal element alloys (MPEAs) continue to gain research prominence due to their promising high-temperature microstructural and mechanical properties. Recently, machine learning (ML) and materials informatics have been used extensively for screening MPEAs, however, most of these efforts were focused on constructing models for predicting phase stability and mechanical properties of known compositions. These approaches may accelerate screening but optimizing new compositions with desirable properties within a practical time frame from an infinitely large design space of MPEA systems remains a grand challenge. To tackle this MPEAdiscovery challenge, we utilize a generative adversarial network coupled with a neural-network ML model to enable inverse design of MPEAs by filtering compositions that have excellent hardness. We successfully found two compositions with higher hardness, one of which far exceeds (> 940 HV) that is 50% higher than hardness of any alloy provided in the training set. ML-predicted hardnesses are validated by Vickers microhardness measurements. We also used density functional theory to provide thermodynamic and electronic insights to higher hardness of the new MPEA found in this work. Our finding suggests that generative ML can greatly accelerate materials discovery by identifying novel compositions, which can serve as a data-informed tool to guide experiments.