A new class of “supermetals” has taken flight in the aerospace industry. High-entropy alloys, or HEAs, are made up of roughly equal proportions of five or more metals and feature unique superpowers—they can be incredibly strong, can withstand extreme temperatures, and can be fashioned into a variety of shapes and sizes.
“The seemingly limitless potential of this novel alloy design paradigm is exciting,” said Zhidong Leong, a Research Scientist from A*STAR’s Institute of High Performance Computing (IHPC). “Next-generation HEAs could produce unprecedented combinations of material properties, spurring significant developments in advanced manufacturing and the aerospace industry.”
Leong did point out, however, that building and testing different HEA “recipes” is no small task given the limitless ways the metal components can be combined. This has proved to be a long-standing barrier in the field; conventional computer modelling and experimental methods take too long to complete.
To that end, Leong and his colleagues tackled the problem from a different perspective. “Our approach stems from the vision that physics-driven machine learning is the key to tackling many of the unsolved problems in materials design,” said Leong.
The team built a computational analysis platform for Mo-V-Nb-Ti-Zr, a HEA known for its myriad aerospace uses. Their platform was powered by machine learning and cluster expansion, a popular computational technique used for modelling the structural properties of alloys. The programme generated simulations of how various Mo-V-Nb-Ti-Zr “recipes” react as they gradually cool down from high temperatures. This data could then be used to predict the structural and mechanical properties of new alloy formulations.
A side-by-side look at data from existing experimental databases showed that the researchers’ computational approach was on the right track: they could accurately predict how HEAs would behave under a wide range of experimental conditions and even use the platform to help design novel HEAs with superior mechanical properties.
“Our results provide the highly desired insights into the HEA’s microstructures over the vast compositional space,” said Leong. “We are now extending our computational capabilities to study complex materials in the sustainability domain.”
The A*STAR-affiliated researchers contributing to this research are from the Institute of Materials Research and Engineering (IMRE) and the Institute of High Performance Computing (IHPC).
