Title: Artificial Intelligence Across Disciplines: Research, Innovation, and Intelligent Solutions
Editors: Dr. Subita Bhagat, Dr. A. Balamurugan, Dr. P. Krishna Kumar, and Mrs. S. Nandhini Devi
ISBN: 978-81-69857-83-3
Chapter: 19
DOI: https://doi.org/10.59646/815/19
Authors: T. Keerthana, and Dr. A. Balamurugan
Abstract
The convergence of artificial intelligence, high-throughput computational screening, and automated experimental synthesis has established a paradigm shift in materials chemistry and nanomaterial development. Conventional trial-and-error methodologies face substantial limitations due to the combinatorial complexity of chemical composition spaces, non-linear nanoscale surface phenomena, and lengthy synthesis-characterization cycles. This chapter details a unified machine learning framework combining crystal graph convolutional neural networks, equivariant message-passing interatomic potentials, and Bayesian active learning for the targeted discovery and optimization of functional nanomaterials. Focusing on lead-free halide double perovskite nanocrystals () for optoelectronic applications, the methodology navigates a compositional space exceeding
candidates. The integrated framework accelerates structural relaxation, accurately predicts thermodynamic stability distances from the convex hull (
), and tailors target electronic bandgaps. This deployment reduces required density functional theory evaluations by over an order of magnitude and guides closed-loop robotic precursor formulation to maximize photoluminescence quantum yields.
Keywords: Materials Informatics, Graph Neural Networks, Active Learning, Nanocrystals, Density Functional Theory, High-Entropy Nanomaterials