(688g) Dioxygen Activation Pathways in Selective Oxidations Catalyzed By Metal Oxides

The oxidative dehydrogenation (ODH) of hydrocarbons and oxygenates is an essential route to convert undervalued feedstocks into marketable chemicals [1,2]. ODH reactions on reducible metal oxides occur via Mars-van Krevelen redox cycles at lattice O-atoms (O*) with H-abstractions from the organic reactants as the kinetically-relevant step, followed by rapid re-oxidation of the reduced centers by O₂ [3]. Compared with the relatively well established reduction part of the cycle, the re-oxidation steps have received much less attention in spite of their consequences for selectivity in ODH reactions [4]. The activated dioxygen intermediates [4,5] formed via O₂ activations can mediate O-insertion steps for species that act as intermediates in undesired combustion reactions [4,6]. Here, we investigate plausible O-insertion routes in ODH reactions by examining propene epoxidation during methanol-O₂ reactions on Mo-based Keggin polyoxometalate (POM) clusters.

Propene epoxidation (PO) products were detected from O₂-C₃H₆ reactants only in the presence of CH₃OH on Mo-based Keggin clusters, consistent with the involvement of O₂-derived species generated on reduced centers that form as CH₃OH ODH intermediates. Two possible activated dioxygen intermediates, peroxo (OO*) and hydroperoxo (HOO*) species, may form via O₂ activation at reduced centers consisting of O-vacancies (*) and surface hydroxyls (OH*), respectively. DFT indicates that these two pathways differ in their kinetic response to O₂ concentrations, because OO* formation is limited by the formation of the O-vacancy formation step, which does not depend on O₂ pressure; in contrast, HOO* formation is mediated by H-transfer from hydroxyls to O₂ as the kinetically-relevant step, making HOO* concentrations proportional to O₂ pressure. Measured PO formation rates from CH₃OH-O₂-C₃H₆ reactants (453 k, 0.1-4 kPa CH₃OH, 1-60 kPa O₂, 1-35 kPa C₃H₆) were proportional to O₂ pressures at low pressures (10 < kPa), suggesting that HOO* act as the main O₂-derived epoxidation intermediates. DFT calculations also indicate the low activation barrier (â??E^â?¡ = 32 kJ/mol) for C₃H₆ reactions with HOO* sites. These HOO* intermediates can also react with OH* to form OO* and H₂O, but such steps occur with much higher barriers (â??E^â?¡ = 81 kJ/mol) than for OOH* reactions with C₃H₆. In contrast, OO* species, formed via either O₂ reactions with vacancies or from OOH*, preferably activate C-H bonds in CH₃OH (â??E^â?¡ = 31 kJ/mol) over O-insertion reactions with C₃H₆ (â??E^â?¡ = 39 kJ/mol). The DFT-derived barriers for CH₃OH C-H activation on OO* (â??E^â?¡ = 31 kJ/mol) are much lower than for the respective reaction with O* species (â??E^â?¡ = 69 kJ/mol) for a give O-atom in a POM cluster, suggesting that such steps occur rapidly after O₂ activation to consume the OO* species and complete a catalytic ODH turnover. Alternate steps involving O-atom migration from OO* to reduced centers require multiple steps, each with significant larger barriers (â??E^â?¡ = 104 kJ/mol) than for direct C-H activation in CH₃OH by OO* .

The first-order dependence of PO formation rates from CH₃OH-O₂-C₃H₆ mixtures on C₃H₆ and O₂ pressures at low pressures, approached a zero-order kinetic regime on both C₃H₆ and O₂; rates also increased with increased CH₃OH pressures along with the concomitant increase in ODH rates. These trends are consistent with the proposed reaction steps for PO occurring on the HOO* sites that are generated from hydroxyl centers by O₂, where the formation rate is in steady-state with consumption rates of OOH* via PO or decomposition to OO* and H₂O. These data, and their mechanistic interpretation, indicate that the ratio of O-insertion to C-H activation rates for a given C₃H₆/CH₃OH reactant ratio would increase as the HOO*/OO* increases, because of their relative reactivities in these two reactions. Indeed, the presence of H₂O increased measured PO/ODH ratios, because O-vacancies (*) react with H₂O to form OH*, thus causing a concomitant increase in HOO*/OO* ratios during O₂ activation on oxide surfaces. These mechanistic inferences, derived from the scavenging of O₂-derived species by propene, taken together with theoretical treatments, illustrate their complexity and selectivity consequences, while providing guidance for catalysts and materials that can enhance or inhibit the relative contributions from pathways mediated by HOO* and OO* intermediates. They also provide fundamental insights into how isolated vacancies, formed as intermediates in ODH reactions, react with the O₂ co-reactants to complete ODH catalytic turnovers.

The authors acknowledge the financial support of the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences (DE-AC05-76RL0-1830) and computational resources from EMSL at Pacific Northwest National Laboratory (PNNL) and from XSEDE supported by the National Science Foundation.

References

[1] T. Punniyamurthy, S. Velusamy, J. Iqbal, Chem. Rev. 105 (2005) 2329.

[2] R.K. Grasselli, Top. Catal. 21 (2002) 79.

[3] P. Deshlahra, R.T. Carr, S.H. Chai, E. Iglesia, ACS Catal. 5 (2015) 666.

[4] X. Rozanska, E. Kondratenko, J. Sauer, J. Catal. 256 (2008) 84.

[5] a. Goodrow, A.T. Bell, J. Phys. Chem. C 111 (2007) 14753.

[6] Y. V Geletii, C.L. Hill, R.H. Atalla, I.A. Weinstock, J. Am. Chem. Soc. 128 (2006) 17033.