Combined DFT and Kinetic Monte Carlo Study of UiO-66 Catalysts for γ-Valerolactone Production

Abstract

Zr-based metal–organic frameworks (MOFs) are excellent heterogeneous porous catalysts due to their thermal stability. Their tunability via node and linker modifications makes them amenable for theoretical studies on catalyst design. However, detailed benchmarks on MOF-based reaction mechanisms combined with kinetics analysis are still scarce. Thus, we here evaluate different computational models and density functional theory (DFT) methods followed by kinetic Monte Carlo studies for a case reaction relevant in biomass upgrading, i.e., the conversion of methyl levulinate to γ-valerolactone catalyzed by UiO-66. We show the impact of cluster versus periodic models, the importance of the DF of choice, and the direct comparison to experimental data via simulated kinetics data. Overall, we found that Perdew–Burke–Ernzerhof (PBE), a widely employed method in plane-wave periodic calculations, greatly overestimates reaction rates, while M06 with cluster models better fits the available experimental data and is recommended whenever possible.