Rates and selectivities for alkene epoxidations depend sensitively on the identity of the active metal center for both heterogeneous and homogeneous catalysts. While group 6 metals (Mo, W) have greater electronegativities and the corresponding molecular complexes have greater rates for epoxidations than group 4 or 5 metals and molecular complexes, these relationships are not established for zeolite catalysts. Here, we combine complementary experimental methods to determine the effects of metal identity on the catalytic epoxidation of 1hexene with H 2 O 2 for active sites within the BEA framework. Postsynthetic methods were used to incorporate groups 4−6 transition-metal atoms (Ti, Nb, Mo, W) into the framework of zeolite BEA. In situ Raman and UV−vis spectroscopies show that H 2 O 2 activates to form peroxides (M-(η 2 -O 2 )) and hydroperoxides (M-OOH) on all M-BEA but also metal oxos (MO) on W-and Mo-BEAs, the latter of which leaches rapidly. Changes in turnover rates for epoxidation as functions of reactant concentrations and the conformation of cis-stilbene epoxidation products indicate that epoxide products form by kinetically relevant O-atom transfer from M-OOH or M-(η 2 -O 2 ) intermediates to the CC bond and show two distinct kinetic regimes where H 2 O 2 -derived intermediates or adsorbed epoxide molecules prevail on active sites. Ti-BEA catalyzes epoxidations with turnover rates 60 and 250 times greater than Nb-BEA and W-BEA, which reflect apparent activation enthalpies (ΔH ‡ ) for both epoxidation and H 2 O 2 decomposition that are lower for Ti-BEA than for Nb-and W-BEAs. Values of ΔH ‡ for epoxidation differ much more between metals than barriers for H 2 O 2 decomposition and give rise to large differences in 1hexene epoxidation selectivities that range from 93% on Ti-BEA to 20% on W-BEA. Values of ΔH ‡ for both pathways scale linearly with measured enthalpies for adsorption of 1,2-epoxyhexane from the solvent to active sites measured by isothermal titration calorimetry. These correlations confirm that linear free-energy relationships hold for these systems, despite differences in the coordination of active metal atoms to the BEA framework, the identity and number of pendant oxygen species, and the complicating presence of solvent molecules.

A pedagogical review that deconvolutes the excess free energy effects of several solvent phenomena and connects findings across a variety of catalytic reactions and materials.

Ti atoms incorporated into the framework of zeolite *BEA (Ti-BEA) or grafted onto SBA-15 (Ti-SBA-15) catalyze alkene epoxidation with hydrogen peroxide (H 2 O 2 ), t-butyl hydrogen peroxide (TBHP), or cumene hydroperoxide (CHP). The rates of epoxidation, however, differ by orders of magnitude depending on the combination of an oxidant, an alkene, and a support used. Within Ti-BEA, the rates of 1-octene epoxidation with H 2 O 2 are 30 and 170 times greater than reactions with TBHP or CHP, respectively. In contrast, 1-octene epoxidation rates in Ti-SBA-15 with H 2 O 2 are 7-and 40-fold higher than in reactions with TBHP or CHP, respectively. Moreover, comparisons of 1-alkene (C 6 −C 10 ) epoxidations within Ti-BEA and Ti-SBA-15 show that the turnover rates depend differently on the length of alkene reactants depending on the oxidant identity, which is due to complex interactions among the alkene, the activated oxidant, and the solvent-filled pore of the Ti-silicate. Thermochemical analyses of apparent activation free energies within a Born−Haber cycle reveal three distinct contributions that affect catalysis, which include charge transfer between Ti-OOR and CC functions within the transition state; the adsorption of the alkene into the silicate pores; and the structural rearrangement of the reactive Ti-OOR intermediates and solvent to accommodate the alkene. First, epoxidations with H 2 O 2 give the highest rates among these oxidants because Ti-OOH intermediates are more electrophilic than Ti-OOtBu or Ti-OOcumyl species as a consequence of electron donation from alkyl groups that increase intrinsic barriers for O-atom transfer. Second, differences in the epoxidation rates between Ti-BEA and Ti-SBA-15 largely reflect the changes in the stability of the alkenes adsorbed within the pores of each silicate. Third, the distinct sensitivities of epoxidation rates on oxidant identity within Ti-BEA and Ti-SBA-15 are caused by differences between the inner-sphere interactions among Ti-OOR intermediates and adsorbed alkenes that depend on the surrounding environment. We present a thermodynamic model that quantitatively describes how inner-sphere interactions among epoxidation transition states depend on the steric bulk of the reacting species and how these interactions are conferred by the topology of the surrounding pore. The mesopores of Ti-SBA-15 allow transition states to access conformations that lower the free energy of the complex relative to analogous transition states in Ti-BEA, which explains why epoxidation rates in mesoporous solids are less sensitive to the identity of the oxidant than within microporous silicates.

Solvent molecules within zeolite pores provide interactions that influence the stability of reactive intermediates and impact rates and selectivities for catalytic reactions. We show the kinetic and thermodynamic consequences of these interactions and reveal their origins using alkene epoxidations in titaniumsubstituted *BEA (Ti-BEA) zeolites. Epoxidation turnover rates vary widely among primary n-alkenes (C 6 −C 18 ) in hydrophilic (Ti-BEA-OH) and hydrophobic (Ti-BEA-F) catalysts in aqueous acetonitrile (CH 3 CN). Apparent activation enthalpies (ΔH app ‡ ) and entropies (ΔS app ‡) increase with alkene carbon number in both catalysts; however, the span of ΔH app ‡ values in Ti-BEA-OH (68 kJ mol −1 ) greatly exceeds that in Ti-BEA-F (18 kJ mol −1 ). These trends, and commensurate gains in ΔS app ‡ , reflect the displacement and reorganization of solvent molecules that scale with the size of transition states and the numbers of solvent molecules stabilized by silanol defects near active sites. Experimental and computational assessments of intrapore solvent composition from 1 H NMR, infrared spectroscopy, and grand canonical molecular dynamics (GCMD) simulations show that Ti-BEA-OH uptakes larger quantities of both CH 3 CN and H 2 O than Ti-BEA-F. The Born−Haber decomposition of simulated enthalpies of adsorption (ΔH ads,epox ) for C 6 −C 18 epoxides attributes ΔH ads,epox that become more endothermic for larger adsorbates to the displacement of greater numbers of solvent molecules bound to silanol defects into the bulk solvent. A strong correlation between ΔH app ‡ and ΔH ads,epox (from GCMD and isothermal titration calorimetry) gives evidence that the disruption of solvent structures provides excess thermodynamic contributions (e.g., G ε ) that depend on the solvent composition in the pores, the excluded volume of reactive species, and the density of silanol groups near active sites. Altering G ε values offers opportunities to control selectivities and rates of reactions through the design of extended active site environments.

The type and density of structural defects within zeolites and zeotypes affect the stabilities of adsorbed species, which, in turn, impact the performance of these materials as catalysts and adsorbents. Despite the recognized importance of silanol groups (SiOH) on the properties of a zeolite catalyst or adsorbent, the densities of SiOH have not been quantitatively linked to the concentration of hydroxide (OH − ) and fluoride (F − ) ions within the synthesis gel. Here, we present a method for the synthesis of siliceous or heteroatom-substituted MFI zeolites (M-MFI; M = Si, Ti, Nb, or Ta) with tunable densities of SiOH, which depend simply on the ratio of hydrofluoric acid (HF) to structuredirecting agent (SDA; tetrapropylammonium hydroxide) used within the synthesis gel. The equilibrated ion exchange between OH − and F − ions forms tetrapropylammonium fluoride in situ, which does not lead to the formation of SiOH defects within M-MFI. Comparisons of infrared spectra from 15 distinct M-MFI materials show that the densities of SiOH groups within M-MFI decrease linearly with the ratio of HF/SDA, independent of the identity of the heteroatom within the framework. Materials synthesized within purely OH − media possess SiOH densities 3 and 100 times greater than analogous materials synthesized with HF/SDA ratios of 1 and 1.5, respectively. The use of HF forms metal fluoride complexes, detected by Raman spectroscopy, which are not readily incorporated into the zeolite framework during synthesis and lead to a decrease in the efficiency of transition-metal incorporation with increasing amounts of HF. The quantity of the heteroatom incorporated into the framework increases linearly with the concentration of metal precursor in the synthesis gel, which provides a method to mitigate the lower yields introduced by the use of HF. Comparisons between H 2 O vapor adsorption isotherms show that M-MFI materials synthesized with an HF/SDA ratio of 1.5 adsorb 4−10-fold less H 2 O than M-MFI synthesized with equal amounts of HF and SDA and 100 times less H 2 O than M-MFI synthesized in OH − media. Comparisons of water uptake within hydrophobic M-MFI materials show that framework Ti and Nb sites stabilize 5 and 7−8 H 2 O molecules, respectively, near saturation vapor pressures. These findings provide a flexible strategy to control the densities of silanol and hydroxyl groups (e.g., Nb-OH) within MFI and will likely extend to the synthesis of other zeolite frameworks.