Molecular determinants of solvent nanoseparation by nanoporous carbon materials

The manifold tunable properties of nanoporous carbon materials, including high surface area, mechanical and chemical stability, electrical and thermal conductivity, rich surface chemistry, and biocompatibility render them a perfect platform for energy storage and conversion, catalysis, nanoseparation, water purification, and drug delivery. Here, we construct molecular models of nanoporous carbon materials and unveil the molecular determinants of solvent nanoseparation and self-diffusion by coarse-grained molecular dynamics simulations. Best nanoseparation is achieved at pore diameters just above size exclusion and surface oxidation is the major selective modifier of self-diffusion of polar molecules in solvent mixtures. The shape of the solvents, and to a lesser extent, the geometry of the pore network also influence nanoseparation. To quantify the impact of pore-network geometry, we derived a Markov state model, which estimates the probability of a molecule following a certain path. Our simulation framework for material construction, simulation, and analysis provides a robust foundation for future investigations on how solvent nanoseparation in amorphous nanoporous materials is governed by the interplay of surface chemistry, pore geometry, and molecular properties.