Catalytic MOFs under Operation: From Static Descriptors to Dynamic Mechanistic Control

Metal-organic frameworks (MOFs) have traditionally been viewed as ideal heterogeneous catalysts because their crystallographic precision enables atomic-level structure-property correlations. Yet our work repeatedly showed that catalytic turnover rarely occurs on these idealized structures. Instead, active sites continuously evolve through substrate adsorption, transient coordination changes, charge redistribution, and surface reconstruction. The central challenge is therefore not simply designing stable frameworks but designing frameworks whose structural evolution remains mechanistically predictable. Therefore, how should structurally well-defined MOFs be interpreted as catalysts when their actual function depends not only on the as-synthesized lattice but also on this inevitable structural evolution during operation? We argue that this question extends beyond MOFs. Many modern heterogeneous catalysts─including oxides, single-atom catalysts, and coordination polymers─operate through dynamic active sites whose structure differs from the as-synthesized material. MOFs provide an unusually well-defined platform for understanding these transformations because both the starting structure and the subsequent evolution can be experimentally resolved. In this Account, we outline a necessary shift in MOF catalysis from treating static descriptors as universally predictive to controlling dynamic reaction mechanisms based on their elementary steps. To bridge this gap between idealized models and operational realities, we developed the framework-retaining selective ligand removal (SeLiRe) strategy. By exploiting the differential lability of mixed ligands, SeLiRe generates deterministic open metal sites and hierarchical pores across both hard-hard and soft-soft coordination networks without fully sacrificing the parent topology. We further investigated charge generation and extraction dynamics, demonstrating how modifications intended to improve optical absorption can inadvertently alter charge localization and introduce kinetic traps, while the dimensionality of inorganic secondary building units fundamentally dictates charge carrier lifetimes. Transitioning to harsher electrocatalytic regimes, we employ operando spectroscopy to capture the potential-driven evolution of these static open metal nodes. Specifically, we tracked the in situ transformation of low-valence sites into highly reactive, hydroxylated transient intermediate states alongside the controlled reconstruction of bulk frameworks into highly conductive surfaces. We propose that the principal advantage of MOFs is not that they remain structurally invariant during catalysis, but that they provide an experimentally accessible platform in which structural evolution can be designed, monitored, and mechanistically interpreted. This perspective shifts catalyst design from optimizing static descriptors toward engineering controlled dynamic active sites.