Our progress in interfacial chemistry has traditionally relied on understanding how surfaces interact directly with reactive species. However, a growing number of studies highlight an equally important, indirect role of surfaces: reshaping the local solvation environment. This effect is challenging to rationalize, arising from a subtle interplay between the water network and surface properties, which limits our ability to predict interfacial reactivity. Here, we address weather interfacial solvation, as measured by cavitation free energies, can be quantitatively inferred from the structural perturbation that a surface induces on the interfacial water network. By using machine learning, we show that it can be quantitatively predicted with a single descriptor measuring the packing of interfacial water molecules within the adlayer physisorbed on the surface. Leveraging this descriptor, we construct a minimal phenomenological model to predict the effect of the surface on hydrophobic solvation free energy at solid/liquid interfaces, validated across diverse surfaces. The model provides intuitive insights for engineering solvation effects in interfacial chemistry.
Jassar et al. (Mon,) studied this question.