Abstract

Water is ubiquitous in zeolite chemistry, yet its effect is poorly understood and standard computational simulation techniques often lack the accuracy to make quantitative predictions. We study the interaction of Brønsted acidic active sites (Al–O(H)–Si, BAS ) with one to three water molecules in the zeolite H–MFI. Considering that BAS can be located in different topological regions of the MFI framework (site heterogeneity), we investigate a representative set of six BAS positions and predict chemically accurate (±4 kJ mol−1) heats of water adsorption. To remedy the shortcomings of density functional theory (DFT) at general gradient approximation (GGA) level, we couple second–order Møller–Plesset perturbation theory (MP2) with dispersion–corrected density functional theory (PBE+D2) and optimise structures with hybrid MP2 :PBE+D2. To reach chemical accuracy, we add coupled–cluster corrections to the energies (MP2:(PBE+D2)+ΔCC).[1] Our predictions align excellently with experimental data for loading 1 H2O /BAS.[2,3] For 2 H2O /BAS, however, our predictions show significantly weaker adsorption than indicated by the experimental values. This discrepancy between predictions and experiment indicates an effect that goes beyond our model of ideal, isolated BAS , leading to stronger adsorption of the second water molecule.

[1] J. Sauer, Acc. Chem. Res. 2019, 52, 3502. [2] J. Lercher and co-workers, Angew. Chem., Int. Ed. 2019, 58, 3450-3455. [3] D. Olson and co-workers, Micropor. Mesopor. Mat. 2000, 35-36, 435-446.