Reactivity or yield for solution-phase catalysis in nanoporous materials with inhibited transport can exhibit a strong dependence on effective pore diameter or width, W. For example, the yield for PNB conversion to an aldol product in mesoporous silica nanoparticles is strongly enhanced upon increasing W from around 1 to 2 nm. To provide a high-fidelity description of such behavior, a multiscale modeling framework is developed incorporating three coupled levels of analysis: (i) Molecular Dynamics (MD) simulation with Effective Fragment Potentials (EFP) on the time scale of tens of ps to assess diffusion coefficients for the reactant and product species, where simulation explicitly incorporates the solvent utilized in experiment; (ii) strongly damped Langevin simulations describing the Brownian dynamics of a reactant and product pair inside the pore in an implicit solvent to assess the propensity, P, for the passing on longer time scales of these species as a function of W; EFP-MD diffusion coefficients provide input to the Langevin simulations; we emphasize that P is shown to be a key parameter controlling overall reactivity; (iii) spatially discrete coarse-grained (CG) stochastic modeling which captures the overall catalytic reaction-diffusion process on the appropriate time- and length-scales. P provides key input to these CG simulations. Kinetic Monte Carlo (KMC) simulation of this CG stochastic model allows assessment of, e.g., reactivity versus pore diameter, thereby connecting with experiment.
Kim et al. (Wed,) studied this question.