IIT Guwahati and UK Universities Build ML Framework for Sustainable Metal Alloys
Researchers from IIT Guwahati, working with London South Bank University, the University of Manchester, and the University of Leeds, have developed a machine learning framework for designing high-performance metal alloys that avoid Critical Raw Materials (CRMs) like tantalum, niobium, tungsten and hafnium — materials that are scarce, costly, and create supply-chain risk. Using a database of 3,608 alloy compositions and an Extra Trees Regressor model, the team identified a CRM-free alloy, "Ti₀.₀₁₁₁NiFe₀.₄Cu₀.₄," predicted to exceed the hardness of a well-known CRM-containing benchmark alloy (~480 HV). They then produced it at lab scale at IIT Kanpur, with measured hardness closely matching the prediction, validating the approach. Researchers say the framework is the first validated computational method of its kind and could extend to optimising other properties like strength, ductility and corrosion resistance, with plans to test the alloys with industry partners next. The work, led by Prof Shrikrishna N Joshi, was published in Scientific Reports.
