AI-Assisted Synthesis Optimization of Low-Dimensional Materials Using Large Language Models
Faculty: Prof. Ananth Govind Rajan (CE) and Prof. Chiranjib Bhattacharyya (CSA)
This project will develop large language model (LLM)-based frameworks for optimizing the synthesis of low-dimensional materials (2D materials and nanoparticles). Building on recent advances in generative AI and LLM-driven scientific reasoning, the work will explore how text-based models can extract, structure, and reason over synthesis knowledge scattered across the materials science literature to propose, refine, and optimize synthesis protocols. The project will combine LLM-based natural language understanding of experimental procedures with structured representations of low dimensional material properties (composition, layer number, defect density, morphology) to build predictive and generative tools that recommend synthesis conditions (precursors, temperature, pressure, growth time, substrate choice) likely to yield materials with targeted structural and electronic properties. A key focus will be closing the loop between literature-mined synthesis knowledge, automated reasoning/planning via LLM agents, and (where feasible) active-learning-guided experimental validation, to accelerate the discovery of reproducible synthesis routes for technologically relevant low-dimensional materials.