CO₂ sequestration in Mg- and Ca-rich minerals is a promising route for long-term carbon storage, but the atomistic mechanisms controlling sequestration at mineral–water interfaces remain poorly understood. This M2 project will use AI-accelerated molecular simulation to investigate the interfacial physics governing CO₂ sequestration and identify surface properties that enhance its efficiency. PROPOSITION_stageM2_DIADEM Pezz… Starting from an existing machine-learning model of the forsterite–H₂O–CO₂ interface, the student will combine MACE foundation potentials, targeted ab initio calculations, active learning, reactive molecular dynamics and free-energy calculations. The project will explore how surface structure, hydroxylation and acidity modify the free-energy landscape and key microscopic processes involved in CO₂ sequestration, including CO₂ speciation, surface protonation, cation dissolution and ion-pair formation. The objective is to extract physically meaningful descriptors linking interfacial structure and dynamics to sequestration efficiency, and ultimately provide design rules for optimized mineral surfaces. The project is funded by the DIADEM Academy.