Discovery of materials has benefited tremendously from advanced supercomputers, including nanoscale materials deposited as thin films on surfaces for harnessing their properties. However, existing models and methodologies are incomplete and prevent us from obtaining fundamental atomic-scale insights into thin-film processing, as they cannot describe movement and interaction of atoms accurately over realistic time scales. The aim of this project—to be conducted at the University of Helsinki—is to overcome this challenge by leveraging strengths of existing methods in the framework of machine learning. This will allow us to study the way by which metal and modern materials (e.g., graphene) interact, a key step for fabricating high-performance devices. On a broader scale our work will provide a universal platform for designing new nanostructured materials.

2026

2028