Dynamic Evolution and Stability of Ketenes in MAPO-18 (M = Si or Mg): Molecular Insights into the Reaction Mechanism for CO2-to-Hydrocarbons

AEI topology molecular sieves (MAPO-18) have shown promising properties as components of OX-ZEO type tandem catalysts, where the Brønsted acid sites (BAS) introduced by different framework substitutions lead to distinct catalytic mechanisms on the CO2-to-hydrocarbons conversion. Ketenes are key intermediates in the conversion of methanol to hydrocarbons. This study provides molecular-level insights into the reactions of three ketenes (ketene, methyl ketene, and dimethyl ketene) with the BAS in MAPO-18 (M = Si, Mg) molecular sieves at operando conditions through first-principles molecular dynamics (FPMD) simulations combined with enhanced sampling techniques. Free energy surfaces constructed from FPMD simulations revealed distinct kinetic and thermodynamic preferences, linking them to different reaction routes for the production of olefins. Prior studies suggested that ketenes and their protonated analogues are key intermediates in two different pathways to olefins formation, and the three ketenes exhibited higher kinetic stability than their protonated forms in H-SAPO-18 compared to H-MgAPO-18, suggesting a high tendency for olefin production via the (cyclo)addition-decarboxylation route in H-SAPO-18. In contrast, the increased stability of the cationic intermediates and low protonation barrier for methyl and dimethyl ketenes in MgAPO-18 favor their direct decarbonylation to olefins. Surface-bound species displayed decreasing stability from surface acetate to surface propionate to surface isobutyrate , aligning with established trends for surface alkoxides. Moreover, a comparison with static calculations demonstrates their limited ability to capture the entropic contributions and dynamic effects that dominate the behavior of active intermediates under realistic reaction conditions, highlighting the necessity of molecular dynamics (MD) approaches for accurate mechanistic modeling of catalytic reactions. Benchmarking against high-level random phase approximation (RPA) calculations further revealed that the density functional theory (DFT) method in combination with the revPBE-D3 functional used in the FPMD simulations may overestimate the stability of cationic intermediates and underestimate the protonation barriers of ketene. Together, these findings underscore the importance of combining dynamic simulations with accurate potential energy descriptions to reliably model active intermediates in confined microporous environments. Overall, this study provides key steps of ketene reactivity in zeolite frameworks, bridging computational and experimental insights into CO2-to-hydrocarbon conversion pathways. These results emphasize how subtle variations in the framework composition and substituents dictate the reaction mechanisms, offering guidance for the rational design of molecular sieves tailored for selective catalytic transformations.