Carbon sequestration is a critical component of global efforts to mitigate the effects of anthropogenic climate change. The accumulation of atmospheric carbon dioxide (CO₂), largely due to industrial activities, is the primary driver of global warming. To limit the impacts of climate change, it is essential to not only remove CO₂ from the atmosphere but also store long-term. One promising approach to this is the natural process of carbonate formation, where CO₂ reacts with silicates and other minerals to form stable carbonate compounds [1-4]. In this study, we applied both Density Functional Theory (DFT) and a recently developed method called Machine Learning Molecular Dynamics (MLMD) to enable a much longer time scale to investigate the reaction mechanism but at reduced costs within the accuracy of Ab Initio Molecular Dynamics (AIMD) [5]. Molecular modeling focused on the diopside surface to determine optimal parameters—such as temperature and the addition of NaCl or NaF—to facilitate cation dissociation for use in carbonation reactions. For DFT analysis four minerals (diopside, forsterite, wollastonite, and lizardite) were selected to compare their efficacy in carbonation. This molecular-level investigation aims to deepen our understanding of the carbonation process and identify strategies to enhance the reaction, thereby contributing to more effective carbon sequestration methods.