The Short-Range Order (SRO), the arrangement of atoms over small distances, plays a crucial role in materials’ properties, yet it has been understudied in metallic alloys. However, recent attention has been drawn to this concept as it is a contributing step towards developing high-performing alloys known as high-entropy alloys. Understanding how atoms self-arrange can pose significant challenges and require vigorous lab experiments or computer simulations.
At MIT’s Department of Materials Science and Engineering, graduate students Killian Sheriff and Yifan Cao, are utilizing machine learning to explore the complex chemical arrangements that form SRO. Their results were recently published in the Proceedings of the National Academy of Sciences.
High-entropy alloys, due to their intricate compositions, exhibit superior properties. These alloys consist of multiple elements, leading to an extensive design space. Understanding and utilizing SRO in tailoring the properties of high-entropy alloys can be beneficial in multiple industries, such as aerospace, biomedicine, and electronics.
SRO refers to the propensity of atoms to form specific chemical arrangements. This arrangement isn’t random but follows a certain pattern. A thorough understanding of SRO and its distribution is critical to understand the full potential of high-entropy alloys.
The traditional approaches for understanding SRO provide an incomplete picture and have limitations. The team at MIT turned to machine learning to overcome these limitations. After simulating chemical bonds in high-entropy alloys, they used 3D neural networks to identify billions of atom arrangements, also known as “motifs”. These motifs, even when rotated, mirrored, or inverted, were identified by the networks regardless of symmetry.
This research will continue further with the help of the U.S. Department of Energy’s INCITE program, which will allow the team to utilize the world’s fastest supercomputer to examine how SRO alters during standard metal processing conditions such as casting and cold-rolling. This can lead to the intentional design of new classes of materials, revolutionizing the materials science field.
The research was funded by the MathWorks Ignition Fund, MathWorks Engineering Fellowship Fund, and the Portuguese Foundation for International Cooperation in Science, Technology and Higher Education in the MIT–Portugal Program.