IIT Guwahati and UK Universities Develop AI-Based Method for Sustainable Metal Alloys

In a remarkable advancement towards sustainable materials and reduced reliance on scarce resources, researchers from the Indian Institute of Technology Guwahati (IIT-G) have collaborated with leading universities in the United Kingdom to develop a machine learning-based framework. This innovative approach aims to design advanced metal alloys without the use of Critical Raw Materials (CRMs).

Collaboration and Research Background

The research was conducted in collaboration with London South Bank University, the University of Manchester, and the University of Leeds. This partnership provides a practical pathway to developing high-performance alloys while addressing supply chain vulnerabilities and environmental concerns associated with the mining of rare elements.

For centuries, alloying has been a fundamental process in improving metals by mixing a base metal with small amounts of other elements. Recently, scientists have focused on High-Entropy Alloys (HEAs), also known as Multi-Principal Element Alloys (MPEAs). These alloys contain several elements in nearly equal proportions and are recognized for their exceptional strength and stability, particularly at high temperatures, making them ideal for demanding applications.

The Challenge of Critical Raw Materials

Despite their advantages, many high-performance HEAs utilized in sectors such as aerospace, gas turbines, and nuclear power depend on critical raw materials like tantalum, niobium, tungsten, and hafnium. These elements pose significant challenges due to their high costs, extraction difficulties, and limited availability. This situation increases import dependence and places considerable pressure on global supply chains and the environment.

Machine Learning-Assisted Alloy Design Framework

To address these challenges, the IIT Guwahati-led research team developed a machine learning-assisted alloy design framework that prioritizes materials free from the most critical raw elements. The researchers categorized CRMs into three distinct groups based on supply risk, economic importance, and global availability. They also created a comprehensive database of 3,608 alloy compositions, focusing on systems constructed from relatively abundant elements.

Among the several models tested, the Extra Trees Regressor emerged as the most effective in predicting Vickers hardness. This model was integrated with optimization techniques inspired by natural processes to identify alloy compositions that deliver high hardness without relying on CRMs.

Development of a CRM-Free Alloy

Using this innovative approach, the team identified a CRM-free alloy, Ti₀.₀₁₁₁NiFe₀.₄Cu₀.₄, which was predicted to have a hardness exceeding that of a widely used CRM-containing alloy with a hardness of approximately 480 HV. The alloy was subsequently developed at laboratory scale at IIT Kanpur, where experimental testing confirmed that its measured hardness closely matched the predicted value.

Significance of the Study

Prof. Shrikrishna N. Joshi of the Department of Mechanical Engineering at IIT Guwahati emphasized the significance of the newly developed alloy, stating that it is well-suited for applications where high hardness is critical. “The developed CRM-free alloy offers strong performance while avoiding the use of critical raw materials, making it suitable for both performance-driven and sustainability-focused applications,” he noted.

Broader Implications of the Framework

Prof. Joshi also highlighted the broader implications of the framework. He stated that it represents the first validated computational approach for designing CRM-free MPEAs using a database based solely on unary and binary alloy compositions, without relying on microstructural or processing parameters. This framework, built entirely on compositional data and machine learning models, can be easily adapted to other material systems where experimental data is limited. Furthermore, it can be extended to predict multiple properties such as strength, ductility, fracture toughness, corrosion resistance, thermal conductivity, and wear resistance.

Potential Applications

The newly designed alloys have a wide range of potential applications, including:

• Wear-resistant components
• Tooling and surface-contact parts
• Automotive components
• Industrial machinery components

Future Directions

The findings of this research have been published in Scientific Reports, a journal of the Nature Publishing Group. The paper is co-authored by Prof. Joshi, Dr. Swati Singh of IIT Guwahati, Prof. Saurav Goel of London South Bank University, Dr. Mingwen Bai of the University of Leeds, and Prof. Allan Matthews of the University of Manchester.

As a next step, the research team plans to collaborate with industry partners and research laboratories to test the alloys under real operating conditions and move towards large-scale deployment.

Note: The development of sustainable metal alloys represents a significant step towards reducing environmental impact and enhancing material performance in various industries.