Predicting structure and properties of mol2Dmat heterostructures by machine learning (A06)

• Subject Area: Theoretical Chemistry: Molecules, Materials, Surfaces
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 555467911

A06 develops a machine-learning framework to predict the structure and emergent electronic and vibrational states in mol2Dmat heterostructures, resulting in ab-initio accuracy and prediction power for systems with 105 atoms. The framework will address simulating donor and acceptor molecules on graphene, hBN, transition metal dichalcogenides, and nanotubes by coupling random structure-search algorithms to machine-learning engines and electronic structure codes. A06 will obtain electron-phonon couplings and 2D material/molecule interactions.

• DFG Programme: Collaborative Research Centres
• Subproject of: SFB 1772: Heterostructures of Molecules and 2D- Materials
• Applicant Institution: Freie Universität Berlin
• Project Head: Dr. Mariana Rossi