Case Study: Universite de Lorraine achieves high-accuracy adsorption enthalpy predictions with Mat3ra

Universit'e de Lorraine - Customer Case Study

Université de Lorraine needed a way to predict adsorption enthalpies in zeolites with high accuracy, but the required random phase approximation (RPA) calculations were extremely computationally expensive. To make the work feasible, the research team used Mat3ra’s platform, combined with machine learning and high-memory cloud compute, to support advanced quantum materials modeling for large zeolite systems.

Mat3ra enabled the team to train ML models on a modest set of RPA calculations and use them to predict energies for many additional structures, dramatically reducing the number of costly calculations needed. The project studied more than 500 zeolite configurations in parallel, with up to 12,800 cores running at peak, and produced adsorption enthalpies that were generally within chemical accuracy of experiment.

Universit'e de Lorraine