Researchers at the Indian Institute of Technology Guwahati, collaborating with various universities, have developed a machine learning-based approach for creating advanced metal alloys that exclude Critical Raw Materials (CRMs). This method is aimed at identifying sustainable, high-performance alloys less reliant on fragile supply chains. High-Entropy Alloys (HEAs), which balance multiple metals in their composition, have emerged as promising alternatives to traditional alloys, often used in critical industries like aerospace and nuclear power. However, many HEAs utilize CRMs, which present challenges due to their limited availability and environmental impact.
To tackle this issue, the team formulated a machine learning-assisted framework that analyzes Multi-Principal Element Alloys (MPEAs) while avoiding CRMs. They categorized CRMs based on supply risk and economic significance, compiling a database of 3,608 alloy compositions primarily from non-critical elements. Utilizing the Extra Trees Regressor model, they successfully predicted the hardness of numerous alloys and identified a CRM-free alloy, “Ti₀.₀₁₁₁NiFe₀.₄Cu₀.₄,” which surpassed the hardness of established CRM-containing alloys.
The researchers synthesized this promising alloy at IIT Kanpur, confirming its properties aligned with predictions, demonstrating the efficacy of their AI-driven method. Prof. Shrikrishna N. Joshi highlighted the significance of this framework, which focuses on compositional data and can predict various mechanical and functional characteristics, extending its applicability beyond traditional parameters. The research, emphasizing both performance and sustainability, is documented in Scientific Reports, co-authored by several team members from IIT Guwahati and partner universities. Potential applications for the newly developed alloy include wear-resistant components and parts for machinery.