Surface Confinement of Finite-Size Water Droplets for SO3 Hydrolysis Reaction Revealed by Molecular Dynamics Simulations Based on a Machine Learning Force Field

As an important source for sulfuric acid in the atmosphere, hydrolysis of sulfur trioxide (SO3) takes place with water clusters of sizes from several molecules to several nanometers, resulting in various final products, including neutral (H2SO4)-(H2O) clusters and ionic (HSO4)--(H3O)+ clusters. The diverse products may be due to the ability of proton transfer and the formation of hydrated ions for water cluster of finite sizes, especially the sub-micrometer ones. However, the detailed molecular-level mechanism is still unclear due to the lack of available characterization and simulations tools. Here, we developed a quantum chemistry-level machine learning (ML) model to simulate the hydrolysis of SO3 with water clusters of sizes up to nanometers. The simulation results demonstrate diverse reaction paths taking place between SO3 and water clusters of different sizes. Generally, neutral (H2SO4)-(H2O) clusters are preferred by water clusters of ultra-small size, and a loop structure-mediated mechanism with SO3(H2O)n≤4 structures and a non-loop structure-mediated mechanism with structure relaxation are observed. As the water cluster size increases to (H2O)8, a (HSO4)--(H3O)+ ion-pair product emerges; and the Eigen-Zundel ion conversion-like proton transfer mechanism takes place and stabilizes the ion pairs. As the water cluster sizes further increase beyond several nanometers ((H2O)n≥32), the (SO4)2-[(H3O)+]2 ion-pair product appears. The reason could be that the surface of these water clusters is large enough to screen Coulomb repulsion between two tri-coordinated ion-pair complexes. These findings would provide new perspectives for understanding SO3 hydrolysis in the real atmosphere and sulfuric acid chemistry in atmospheric aerosols.