Research Projects

High-entropy materials (HEMs), composed of multiple principal elements stabilized by high configurational entropy, represent a promising class of materials with exceptional mechanical, thermal, and chemical properties. Unlike conventional alloys or ceramics, HEMs exhibit tunable functionalities and are increasingly explored for applications in catalysis, energy storage, aerospace, defence, and environmental technologies. Their versatility enables performance in high-temperature structural alloys, catalysts for CO₂ reduction and water splitting, solid-state electrolytes in batteries, and corrosion- or radiation-resistant coatings. However, rational design of HEMs remains a major challenge due to their vast compositional space and complex structure–property relationships. This challenge is particularly acute when focusing on environmentally sustainable systems using earth-abundant elements, as most current HEMs rely on critical or rare metals that are costly and have supply risks. Thus, there is a pressing need to design next-generation HEMs using abundant, low-cost transition metals such as Fe, Mn, Co, Ni, and Cu, which also offer diverse electronic and chemical characteristics. Machine learning (ML) offers a powerful route to accelerate the discovery of such materials by enabling efficient prediction of properties, phase stability, and synthesis feasibility. Although significant global progress has been made in applying machine learning to the design of high-entropy materials, efforts within the Indian research scenario—particularly toward sustainable, earth-abundant systems—are still in the early stages of development. This project aims to address this gap by developing interpretable, domain-specific ML frameworks for the inverse design of high-entropy alloys and ceramics based on earth-abundant elements. The methodology includes a combination of classical algorithms, graph neural networks (GNNs), transfer learning, and large language models, trained on curated datasets drawn from experimental literature and high-throughput computations. The central hypothesis is that such models can reveal composition–structure–property relationships and guide the discovery of materials with targeted functionalities. By integrating explainable Artificial Intelligence (AI) techniques, the project will not only predict material performance but also uncover key physicochemical factors governing HEM behavior. Expected outcomes include validated ML models, new sustainable material candidates, and scientific insights that advance AI-driven materials discovery and support national goals for clean energy and technological innovation.