Low-carbon hydrogen can be a powerful asset in ensuring energy security and reaching the “Net Zero by 2045” goal fixed by the Scottish Government. This ambitious objective requires the development of a supply chain and a new generation of “hydrogen-ready” materials, such as selective membranes for hydrogen purification and barriers for the safe storage and delivery of such gas.
Physical models are an efficient strategy for the design of new materials with desired performance and can replace long and expensive experimental campaigns. Their application to novel materials, however, is limited by the lack of input parameters.
We aim at overcoming this issue by using Machine Learning (ML) algorithms to estimate the material parameters required by physical models, based on their molecular structure. This ML-aided physical modeling platform will facilitate the screening and design of Hydrogen-ready materials and enable the fast deployment of a Hydrogen infrastructure in Scotland and the UK.

Collaborators: Dr. Eleonora Ricci, Marie Curie PostDoctoral fellow, Demokritos Institute, Greece, coordinator of the project ML-MULTIMEM