(502d) Mechanistic Consequences of Composition and Descriptors of Reactivity for Alkanol Odh Reactions on Polyoxometalate Clusters

Prashant Deshlahra and Enrique Iglesia*

 Department of Chemical Engineering, University of California, Berkeley, CA 94720

*iglesia@berkeley.edu

Polyoxometalate (POM) clusters are oxides with Keggin type structure, which consist of central tetrahedral units enclosed in a metal oxide shell and charge balancing protons [H_8-n (Xⁿ⁺O₄)(M₁₂O₃₆); X= P, Si, Al, Co, M = W, Mo, V etc.]. Protons in these clusters act as catalytic Brønsted acid sites while MO_x addenda species can undergo redox cycles to catalyze oxidative dehydrogenation (ODH) reaction. Such material exhibit well-defined atomic connectivity over diverse compositional range, making them well-suited for combining rigorous kinetic and isotopic and spectroscopic studies with density functional theory (DFT) estimates of their redox and acid properties and the energetics of the relevant elementary steps. We report here mechanistic interpretations of oxidative dehydrogenation (ODH) rates of CH₃OH-O₂ mixtures on Mo- and W-based POM clusters, with different central atoms (H₃PMo₁₂O₄₀, H₄SiMo₁₂O₄₀) or addenda atoms (H_3+nPV_n=1-3Mo_12-nO₄₀, H₄PV₁W₁₁O₄₀).

ODH reactions involve kinetic coupling between steps that reduce and oxidize the oxide clusters [1]; the dependence of rates on O₂ and CH₃OH pressures show that reduction steps, which hydrogenate the clusters using CH₃OH, typically limit rates for most O₂/CH₃OH reactant ratios. ODH rates showed a Langmuir-type behavior with changing CH₃OH pressures, suggesting saturation of oxide surface with adsorbed CH₃OH with increasing pressure. Initial linear CH₃OH dependence provides a measure of first-order rate constant (k_1st,ODH), which reflects stability of kinetically relevant transition state (TS) with respect to gaseous CH₃OH reactant. Measured kinetic isotope effects (r_CH3OH/r_CD3OD = 2.5, r_CH3OH/r_CH3OD = 1.0) suggest that this TS involves H-abstraction from C-H bond of CH₃OH, but cannot discern between H-abstraction a CH₃OH molecule and that from a dissociated methoxide species; the latter are typically presumed to be prevalent despite lacking spectroscopic evidence of reactive methoxides on several oxide surfaces during ODH conditions [2]. DFT calculations can distinguish between the molecular and the dissociative routes, and we show that H-abstraction from molecularly adsorbed CH₃OH are significantly more favorable than dissociative route and its calculated pathways are consistent with experimental kinetic and isotopic measurements (calculated KIE, r_CH3OH/r_CD3OD = 2.2). The Measured k_1st,ODH values reflect ability of oxide clusters to accept electrons during H-abstraction, in a step mediated by a late transition state (TS). The TS stabilities correlate strongly with H-addition energies (HAE) to catalysts’ O-atoms (slope of E_TS versus HAE ~ 0.8), which probe energetics of reduction analogously to the H-abstraction step. HAE values change with location on clusters and need exponentially weighted reaction-averaging (HAE_avg) over all locations; these HAE_avg values provide accurate descriptors of the reduction part of redox cycle. HAE_avg values decrease with changing POM clusters’ addenda-atoms from W to Mo and their V-substitution suggesting improved ability to accept electrons in reduction step, consistent with higher measured ODH rate constants. ODH rates were independent of O₂ pressure and adding H₂O to CH₃OH-O₂ reactant mixtures decreased rates without changing O₂ pressure dependence, suggesting fast re-oxidation steps resulting in low concentration of reduced centers and inhibition of rates by H₂O blocking the reactive oxidized centers.

Re-oxidation of clusters is kinetically silent but their details are relevant to elementary steps that complete the redox turnovers and are accessible through accurate detection of reduced centers in UV-vis spectra. Spectra from experiments and time dependent DFT calculations indicate small extents of reduction; the concentration of reduced centers decreases with increasing O₂ and H₂O pressures, consistent with increased re-oxidation rates and with blocking of oxidized centers by H₂O, respectively. Mechanistic interpretations of the steady-state concentration of reduced centers and of transient response [3] of change in the concentration after changing reactant pressure are consistent with fast re-oxidation rates; the proportionality of these rates to O₂ pressure increase with the extents of reduction, suggesting non-Langmuirian behavior of the elementary steps involved in re-oxidation. DFT calculation show that re-oxidation steps can proceed via a pathway involving O-vacancy formation in reduced cluster followed by O₂ activation at the vacancy, and another pathway involving formation hydroperoxide radicals by direct reaction of O₂ with the H-atom in the reduced cluster without vacancy formation; prevalence of the two routes is determined by H₂O concentration in the gas-phase. Re-oxidation elementary steps lead to formation of peroxide species on the cluster by addition O₂ to reduced centers, involvement of such species were probed using propene oxidation concurrent with CH₃OH-O₂reactions because propene molecules form epoxide products by reacting with such peroxide species.

These results describe detailed mechanistic interpretations using kinetic, isotopic and spectroscopic methods along with theoretical calculations to establish the rigorous relations between catalyst properties and stability of relevant transition states for reduction and re-oxidation steps ubiquitous in catalysis by oxide materials.

References

[1] J.M. Tatibouet, Appl. Catal. A.: General. 148 (1997) 213; P. Mars, D.W. van Krevelen, Chem. Eng. Sci. 3 (1954) (Special Suppl.) 41.

[2] Seman, M., Kondo, J.N., Domen, K., Radhakrishnan, R., Oyama, S.T., J. Phys. Chem. B, 106, 12965 (2002); Jehng J.M., Hu H., Gao X., Wachs I.E. Catal. Today 28, 335 (1996).

[3] Argyle, M.D., Chen, K., Iglesia, E., Bell, A.T., J. Phys. Chem. B, 109, 2414 (2005).