(544gr) Catalytic Thiophene Oxidation By Groups 4 and 5 Zeolite BEA with H2O2: Mechanistic and Spectroscopic Evidence for the Effects of Metal Lewis Acidity and Solvent Lewis Basicity

Bregante, D. T., University of Illinois, Urbana-Champaign
Patel, A., University of Illinois, Urbana-Champaign
Johnson, A., University of Illinois, Urbana-Champaign
Flaherty, D., University of Illinois, Urbana-Champaign
Sulfoxidation of thiophenes with hydrogen peroxide (H2O2) is a promising alternative to hydrodesulfurization (HDS) for the desulfurization of crude-oil streams to produce low- and ultra-low-sulfur content fuels and gasolines, as HDS requires long residence times and harsh reaction conditions to remove highly-substituted polyaromatic thiophenes. Highly disperse group 4 and 5 transition metals grafted onto mesoporous silica and incorporated into zeolites selectively activate H2O2 for sulfide oxidation, which produces sulfoxide moieties that enable facile methods for removing sulfur from product streams (e.g., liquid-liquid extraction).

Here, we present a combination of kinetic data and in situ UV-visible spectroscopy to identify and implicate specific surface species for sulfoxidation and the mechanism by which 2,5-dimethylthiophene (C6H8S; i.e., a model substituted thiophene) oxidation occurs [1]. Product selectivities as a function of C6H8S conversion show that C6H8O is resistant to further oxidation, as this is process breaks aromaticity. Notably, these product selectivities differ from non-aromatic sulfide oxidation (e.g., thioanisole) [2], which are readily oxidized to form the corresponding sulfone (i.e., the product resulting from two sequential oxidation reactions). Sulfoxidation rates vary by five thousand among this series of materials, while H2O2 selectivities do not vary significantly. Mechanistic interpretation of kinetics measured in the absence of mass-transfer restrictions show that group 4 and 5 metal atoms (Ti, Zr, Nb, and Ta) incorporated into zeolite *BEA (M-BEA) reversibly adsorb then irreversibly activate H2O2 to form pools of hydroperoxide and peroxide intermediates (collectively referred to as M-(O2)). These reactive intermediates then combine with C6H8S to form the corresponding sulfoxide (C6H8SO). UV-vis spectra collected in situ corroborate kinetic analysis that implies surfaces are saturated with specific surface species (e.g., M-(O2) and sulfoxide saturated surfaces at low and high [C6H8S]:[H2O2], respectively) under different well-defined kinetic regimes. This mechanistic insight allowed us to measure activation enthalpies (ΔH‡) for C6H8S oxidation (ΔH‡Ox) and H2O2 decomposition (ΔH‡Dec). ΔH‡Ox and ΔH‡Dec are similar in value and decrease with a similar functional dependence on the adsorption enthalpy for CD3CN bound to Lewis acid sites, likely because both reactions involve the interaction of an electron lone pair (on S and O atoms of C6H8S and H2O2, respectively) with the reactive M-(O2) intermediates. These differences in ΔH‡Ox between the four M-BEA and lack of differences between ΔH‡Ox and ΔH‡Dec for a given M-BEA explain why there are such large differences in the rates of sulfoxidation without a corresponding change in H2O2 selectivities.

Measured turnover rates for C6H8S sulfoxidation depend strongly on the solvent for that catalytic reaction and change by more than ten thousand. For example, the rates of reaction are immeasurable (<10-7 (mol C6H8SO)(mol Ti • s)-1; 0.01 M C6H8S, 0.01 M H2O2) when dimethylsulfoxide is used as a solvent, whereas, rates are several orders of magnitude higher in acetonitrile (~0.007 (mol C6H8SO)(mol Ti • s)-1). UV-vis spectra collected in situ on Ti-BEA show that differences in turnover rates are attributed to competition between these Lewis basic solvents and H2O2 for adsorption onto the Lewis acidic active sites, which result in decreased rates (by decreasing the relative abundance of M-(O2) intermediates). More generally, turnover rates depend exponentially on the solvent nucleophilicity (N1), which is shown for a variety of academic- and industrially-relevant solvents (acetonitrile, methanol, ethanol, acetone, and p-dioxane) as well as solvent mixtures (methanol and acetonitrile).

Together, the data and interpretations presented in this work provide a self-consistent mechanism for the sulfoxidation of C6H8S with H2O2 over M-BEA, and demonstrates that the rates of reaction depend highly on the electron affinity of the active site, while H2O2 selectivities are unaffected. Moreover, this work identifies the dependence of turnover rates on solvent nucleophilicity (a proxy for “Lewis base strength”) and shows that these Lewis basic solvents competitively adsorb to the active sites to form Lewis acid-base adducts, which decrease turnover rates. Consequently, this work shows that turnover rates for sulfoxidation are highest when highly electrophilic active sites (i.e., stronger Lewis acids) are paired with non-coordinating solvents (i.e., weaker Lewis bases) and will guide the design of increasingly productive catalytic systems for sulfide oxidation.


[1] Bregante, D.T.; Patel, A.Y.; Johnson, A.M.; Flaherty, D.W.; In Review

[2] Thornburg, N.E.; Notestein, J.M.; ChemCatChem 2017, 9, 3714-3724.