Quantification and Statistical Analysis of Errors Related to the Approximate Description of Active Site Models in Metal-Exchanged Zeolites

摘要

Metal-exchanged zeolites are complex catalytic systems with great potential for applications in hydrocarbon conversion processes. The numerous combinations of zeotypes and metal-based Lewis acid motifs create a vast material space that can only be effectively navigated with robust structure-function relationships. Exhaustive modeling of extraframework metal sites in all crystallographically distinct framework positions is not practical; thus, we quantified the uncertainties arising from the use of simplified active site representations using the C-H bond activation in methane as a probe reaction. Density functional theory calculations and statistical analysis of associated error ensembles suggest that the Lewis acid identity primarily determines the reactivity of the zeolite towards methane activation, with other factors such as zeolite model characteristics deemed secondary. The error ensemble distributions were used to predict the relative probability of creating active C-H bond scission fragments in metal-exchanged zeolites, and thus can be used to efficiently screen various metal-exchange candidates in their efficacy towards hydrocarbon conversion.