Elucidating and Tuning Catalytic Sites on Zirconium- and Aluminum-Containing Nodes of Stable Metal–Organic Frameworks

Abstract

ConspectusMetal-organic frameworks (MOFs) are a huge, rapidly growing class of crystalline, porous materials that consist of inorganic nodes linked by organic struts. Offering the advantages of thermal stability combined with high densities of accessible reactive sites, some MOFs are good candidate materials for applications in catalysis and separations. Such MOFs include those with nodes that are metal oxide clusters (e.g., Zr6O8, Hf6O8, and Zr12O22) and long rods (e.g., [Al(OH)]n). These nanostructured metal oxides are often compared with bulk metal oxides, but they are in essence different because their structures are not the same and because the MOFs have a high degree of uniformity, offering the prospect of a deep understanding of reactivity that is barely attainable for most bulk metal oxides because of their surface heterogeneity. This prospect is being realized as it has become evident that adventitious components on MOF node surfaces, besides the linkers, are crucial. These ligands arise from modulators, solvents, or products of solvent decomposition in MOF synthesis solutions, and because they are minor components that are often irregularly placed on defects, they may not show up in X-ray diffraction (XRD) crystal structures. Hydroxyl groups on the nodes (like those on bulk metal oxides) are regarded as native functional groups arising from solvent water, but they may barely be present initially, with common ligands instead being formate and acetate formed from modulators formic acid and acetic acid. (Formate also arises from the decomposition of dimethylformamide (DMF) solvent.) Replacement and control of the node ligands is facilitated by postsynthesis reactions (e.g., with alcohols or aqueous HCl/H2SO4 solutions) or as a result of high-temperature decomposition. In catalysis, adventitious node ligands can be (a) reaction inhibitors that block active sites on the nodes (e.g., formate blocking Zr, Hf, or Al Lewis acid sites); (b) reaction intermediates (e.g., ethoxy in ethanol dehydration); or (c) active sites themselves (e.g., terminal OH groups in tert-butyl alcohol (TBA) dehydration). Surprisingly, in view of the catalytic importance of such ligands on bulk metal oxides, their subtle chemistry on MOF nodes is only recently being determined. We describe (1) methods for identifying and quantifying node ligands (especially by IR spectroscopy and by 1H NMR spectroscopy of MOFs digested in NaOH/D2O solutions); (2) node ligand surface chemistry expressed as reaction networks; (3) catalysis, with mechanisms and energetics determined by density functional theory (DFT) and spectroscopy; and (4) MOF unzipping by reactions of linker carboxylate ligands with reactants such as alcohols that break node-linker bonds, a cause of catalyst deactivation and also an indicator of node-linker bond strength and MOF stability.