COMPUTATIONAL MATERIALS SECTION

Overview

The Computational Materials Section at CSIR-CGCRI develops and applies advanced modelling and simulation methods to understand, predict, and design new materials. Developing materials with targeted properties by trial and error is slow and costly, particularly when the space of possible compositions and structures is vast. The section addresses this challenge by combining physics-based modelling with data-driven methods, working closely with the institute’s experimental groups and with academic researchers across the country.

Material properties such as mechanical strength, optical response, and thermal and chemical stability ultimately come from structure and bonding at the atomic scale. The section’s work connects these length scales. Molecular dynamics (MD) simulations, using classical and machine-learned interatomic potentials, show how atomic arrangements, network topology, and dynamics control the behaviour of disordered and crystalline materials, from structural relaxation and phase stability to plasticity and fracture.

Building on this, the section develops AI/ML models, including graph neural networks that learn directly from atomic structure, to capture composition–structure–property relationships. These models enable inverse design, in which generative models and Bayesian optimisation start from target properties and propose candidate materials for experimental validation. Density functional theory (DFT) complements this with insight into electronic structure, chemical bonding, and defect energetics, and provides high-fidelity data for training machine-learned potentials.

At the component scale, finite element analysis (FEA) uses material properties from simulation and experiment to predict stress distribution, thermal and residual stresses, deformation, and failure under realistic operating conditions. This lets designs be tested and optimised virtually, reducing costly prototyping.

The work is supported by the institute’s high-performance computing facility and carried out in collaboration with experimental groups, academic researchers across the country, and industry.

Mission

The mission of the Computational Materials Section is to speed up the discovery and design of advanced materials through predictive modelling grounded in statistical mechanics and quantum mechanics and strengthened by artificial intelligence. By linking atoms to applications, from electronic and atomic-scale mechanisms to the performance of engineered components, the section aims to turn fundamental scientific insight into practical materials solutions for national priorities in energy, photonics, healthcare, and strategic sectors.