Embedded correlated wavefunction driven multi-level quantum mechanics/molecular dynamics simulations of aqueous reactions

Accurate simulation of aqueous chemistry is challenging for density functional theory (DFT) based molecular dynamics (MD) simulations. This is due to DFT functionals’ approximate electron exchange and correlation description, which often yield sizeable errors when computing aqueous reaction free energies. Such errors are of particularly large magnitude in the case of ionic species, where commonly utilized pure density functionals display large charge delocalization errors. A remedy may be provided by correlated wavefunction (CW) based methods; however, the computational scaling of such methods prohibits their use in MD simulations, making them unable to capture structural dynamic and explicit solvation effects, also resulting in errors when computing aqueous reactivity. Here, we implement multi-level quantum mechanics/MD simulations, which combine embedded CW theory with DFT-MD to capture accurately effects arising from both structural dynamics and electron exchange-correlation. We apply this framework to elucidate the atomic-level processes involved in carbonate formation, including CO₂ speciation, Mg²⁺ and Ca²⁺ dehydration, and Mg²⁺/Ca²⁺-CO₃^2- ion-pairing, which are relevant to elementary processes in carbon capture and storage technologies and remain not fully understood.