Materials Discovery


Prescience provides computational prediction of material properties with high accuracy even before they are synthesized. We capitalize on the advances in machine learning (ML) for computational methodologies and simulations to predict bulk macroscopic and microscopic properties. We provide solutions based on computational methodologies to the soft materials like surfactants (ionic/non-ionic), proteins and lipids (charged/neutral), polymers, ionic liquids and hard materials like semiconductors, metals, alloys etc. Our approach is to use atomistic level information e.g., electronic structures, atomic level interactions and chemical structures to design and predict new materials.


Smart materials are responsive materials reacting to external stimulations and have one or more properties. We provide solutions to the industrial, technological and domestic applications designing smart materials viz., rubber, polymers, and its composites for smart applications like sensors, screens, thermo-responsive materials etc. We use ab-initio quantum chemical methods and classical molecular simulations to design and screen materials targeted for specific applications.


Deep learning has been receiving increasing attention and has achieved great improvements in both time efficiency and prediction accuracy. It is well known that computational simulation and experimental measurement are two conventional methods that are widely adopted in the field of materials science. We are using deep learning and machine learning along with experiment data and computer simulation to design new materials and material compositions. These methodologies are getting useful and saving cost for development of rubber, surfactant, paint, catalyst, lubricant compositions.


Our service also includes prediction of failure of materials and its mechanism. Prediction of failure of materials at different external and internal conditions are important aspect of materials design. Durability and safety are key elements in tyre, paint, semiconductor, metal, alloy, composite industries. We routinely use quantum chemical, Monte Carlo, molecular dynamics methods to predict the weak points (chemical bonds), crack propagation pathways and kinetics of it. Currently we are also using machine learning on the data available for certain materials and trying to connect that to the materials compositions.