(465b) Group IV and V Periodic Trends in Olefin Epoxidation: Effects of Local Environment and Electronic Structure

Hydrogen peroxide (H₂O₂) is a green oxidant that can epoxidize a variety of olefins over site-isolated group IV and V metal oxides (e.g., grafted onto mesoporous silica¹ or incorporated into zeolites^2,3). These group IV and V metals incorporated into the *BEA framework (M-BEA) have been shown to activate H₂O₂ to form epoxidation-active hydroperoxide (M-OOH) or superoxide (M-(O)₂^-) (collectively referred to as M-(O₂)*) species, respectively. The reactivity and selectivity towards epoxidation is proportional to the functional Lewis acid strength of the catalyst (i.e., adsorption enthalpies of Lewis-acid bound deuterated acetonitrile) or the ligand-to-metal charge transfer (LMCT) energy of the reactive intermediate.³ In this work, we explore the electronic properties of these activated metal complexes and extend our previous findings to include an analogous series of catalysts grafted onto mesoporous silica (M-SiO₂) via calixarene complexes. The differences in the reactivity towards olefin epoxidation over M-BEA and M-SiO₂ catalysts are attributed to two main effects. First, the microporous structure of the *BEA framework solvates, and thus stabilizes, the interaction between the olefin and M-(O₂)* intermediate in the transition state, which decreases the apparent activation enthalpies relative to SiO₂. Second, the coordination environment of the M-BEA or M-SiO₂ affects the electronic structure of the active intermediate (as detected via UV-vis spectroscopy), which results in a systematic decrease in the LMCT band energies for M-SiO₂ catalysts.

Turnover rates for styrene (C₈H₈O) oxide formation through primary reaction pathways as a function of styrene (C₈H₈) and H₂O₂ concentrations over M-BEA and M-SiO₂ (M = Ti, Nb, or Ta) reveal two distinct regimes, which correspond to a change in the most abundant surface intermediate (MASI). The observed dependencies are consistent with a rate expression that describes the irreversible activation of H₂O₂ to form a pool of M-(O₂)* intermediates, which then react with C₈H₈ or H₂O₂ via an Eley-Rideal mechanism to form C₈H₈O or H₂O₂ decomposition products, respectively. Time-resolved UV-vis measurements acquired in situ during reaction of cyclohexene with H₂O₂-activated M-BEA and M-SiO₂, in conjunction with molecular probe reactions with cis-stilbene, support the trend that the reactive intermediate on Ti-based catalysts is Ti-OOH (i.e., hydroperoxide), while Nb- and Ta-materials react through a M-(O₂)^- (i.e., superoxide) intermediate.

Understanding of the fundamental properties of the catalysts and reactive intermediates that are responsible for the differences in apparent activation barriers for epoxidation require comparisons of these catalysts made under similar conditions (i.e., MASI). Activation enthalpies for C₈H₈O formation (ΔH^‡_E) are measured on similar M-(O₂)* saturated surfaces on all catalysts (shown by rates that are proportional to [C₈H₈] and independent of [H₂O₂]), and correlate linearly with both the measured adsorption enthalpies of pyridine (ΔH_Py, determined via van’t Hoff analysis of FTIR spectra of pyridine coordinated to the Lewis acidic Ti, Nb, and Ta atoms within the BEA framework or grafted onto SiO₂) and also to the LMCT band energy of the reactive intermediates on these materials. These comparisons reveal two critical differences among these materials. First, ΔH^‡_E and ΔH_Py values for M-SiO₂ materials are systematically greater by 19 ± 2 and 11 ± 2 kJ mol^-1 than M-BEA materials, respectively, which is consistent with the difference in bulk-averaged adsorption enthalpies for C₈H₈ and pyridine in the *BEA framework and in mesoporous SiO₂ (measured in the absence of metal atoms). This difference is attributed to the microporous environment of *BEA solvating the transition state (relative to the initial state) to a larger extent than SiO₂, which results in the systematic decrease of apparent activation and adsorption enthalpies between *BEA and SiO₂. Second, the linear relationship between ΔH^‡_E and the LMCT band energy of the active intermediate possesses the same systematic offset in ΔH^‡_E between *BEA and SiO₂ that matches the differences in adsorption enthalpies of C₈H₈ (vide supra). Interestingly, the M-SiO₂ catalysts possess systematically lower energy LMCT bands, which may be attributed to one of two phenomena, or both. First, the hydroperoxidolysis of a M-OSi bond in the zeolite framework leaves a vicinal SiOH group in a tetrahedral position that may contribute a small amount of electron density towards the metal atom, which decreases the Lewis acid strength of the metal atom. Second, the neighboring SiOH group in *BEA likely distorts the geometry of the M-(O₂)* intermediates, which will lower the extent of orbital overlap, lengthening the M-O bonds, and thus increase the LMCT band energy. The overarching trend relating ΔH^‡_E and the LMCT band energy suggests that metal centers that are highly electron withdrawing (i.e., possess more highly polarized M-(O₂)* bonds) result in more electrophilic and reactive intermediates. To develop an intuition for the electrophilicity of the reactive intermediates, turnover rates for the epoxidation of para-substituted styrene (X-C₈H₈, where X = OMe, Me, H, Br, and NO₂) were measured under identical reaction conditions (3 mM X-C₈H₈, 0.01 M H₂O₂, 313 K). The corresponding Hammett plot reveals a reaction constant (ρ) of -0.9 ± 0.05 for all M-β and M-SiO₂, which definitively shows that the reactive intermediates implicated in olefin epoxidation have a strong preference to react with electron-rich olefins. Collectively, these data show that the rational design of next-generation epoxidation catalysts should seek to increase the Lewis acid strength of the metal center to increase the electrophilicity and reactivity of the active intermediates and leverage the decrease in apparent activation barriers that are attributed to the solvation of the transition states in microporous environments.

References:

1) Thornburg, N.E.; Thompson, A.B.; Notestein, J.M.; ACS Catal., 2015, 5, 5077-5088.

2) Bregante, D.T.; Priyadarshini, P.; Flaherty, D.W.; J. Catal., 2017, 348, 75-89.

3) Bregante, D.T.; Flaherty, D.W.; 2017, In Revision