(664e) Mechanistic Interpretations and Consequences of Hydrogen Spillover in Toluene Hydrogenation Catalysis

Spillover effects are ubiquitous in catalysis and involve the transport of active species formed on one function to another one that does not itself form the active species at the conditions of the experiment. Hydrogen spillover, in particular, has attracted significant attention because of its practical and mechanistic consequences for alkene¹ and arene² hydrogenations, alkane hydroisomerization,³ and methanol synthesis.⁴ In many cases, spillover has been proposed to occur via an activated H-atom (I_H) that is formed on a metal function, denoted as the “emitter,” and transported to an “acceptor” site present on the support where it is scavenged via subsequent reactions. Here, we explore hydrogen spillover phenomena and their mechanistic consequences for toluene hydrogenation, for which metal oxide supports have been implicated as the acceptor functions in such H-spillover intermediates, but without a definitive mechanistic description of such reaction pathways or the definitive identification of the intermediates that mediate such cascade reactions.

The role of hydrogen spillover on rates of toluene hydrogenation was measured on SiO₂-supported Pt catalysts (Pt/SiO₂) as physical intrapellet mixtures with SiO₂, Al₂O₃ or ZrO₂ or as loose mixtures of Pt/SiO₂ and γ-Al₂O₃ aggregates similar in size. These mixtures were treated in-situ in 10% O₂/He at 573 K in order to remove any adventitious carbonaceous contaminants derived from ambient exposure and potentially able to block Pt sites; they were then treated in H₂ at 573 K to reduce Pt nanoparticles. Hydrogenation turnover rates on Pt/SiO₂ (393 K, 90 kPa H₂, 0.35 kPa toluene), reported based on the number of exposed Pt atoms, were independent of Pt loading (0.2 to 0.5% wt.) or dispersion (1.5 to 5 nm) and unaffected by intrapellet dilution of Pt/SiO₂ by SiO₂. Turnover rate enhancements as large as five-fold were observed on Pt/SiO₂ catalysts when prepared instead as physical mixtures with γ-Al₂O₃ or ZrO₂, two oxides that did not form toluene hydrogenation products at these conditions. Similar promotional effects were observed for ethene hydrogenation (303 K, 20 kPa H₂, 2 kPa ethene) and 1,3,5-tri-isopropylbenzene hydrogenation (443K, 90 kPa H₂, 1 kPa 1,3,5-triisopropylbenzene). CO oxidation turnover rates (473 K, 10 kPa O₂, 0.5 kPa CO) on these same mixtures, in contrast, were insensitive to the amount or identity of the diluent used, suggesting that the dispersion of the Pt nanoparticles was identical in all mixtures.

Hydrogenation turnover rates increased monotonically for a given Pt/SiO₂ catalyst with increasing extent of dilution by γ-Al₂O₃, before reaching asymptotic rates that were unaffected by additional diluent. Moreover, similar asymptotic rates were observed when a given Pt/SiO₂ catalyst was mixed physically with either γ-Al₂O₃ or ZrO₂. These data show that asymptotic rates reflect the presence of an acceptor function at concentrations sufficient to fully scavenge all the I_H intermediates formed on the Pt function. Consequently, spillover-mediated rates (defined as the difference between the turnover rate on a diluted mixture and on Pt/SiO₂) depend only on the rate of I_H formation on the emitter function, but not on the amount, site density, and site reactivity of the acceptor. No intermediate species I_H were detected in the fluid phase at any conditions, suggesting that these intermediates are present at concentrations below their detection limits and set by their rate of formation and scavenging at steady-state. Such species were also undetected on Pt/SiO₂, which lacks an acceptor function, indicating that even their equilibrium concentrations are too low to detect. As a result, close proximity between emitter and acceptor sites is required in order to circumvent these equilibrium limitations and mediate detectable spillover hydrogenation events.

Turnover rate enhancements were also observed on separate Pt/SiO₂ with γ-Al₂O₃ aggregates of similar size present as loose mixtures (20 to 100 µm radius). These mixtures gave asymptotic rates that decreased as the size of the Pt/SiO₂ aggregates increased. Reaction-transport models of the formation of I_H within Pt/SiO₂ aggregates and its consumption on acceptor domains based on the Thiele formalisms show that concentration gradients of I_H develop as equilibrium pressures of I_H are reached within the inner regions of these aggregates and spillover-mediated rates become limited by diffusion of I_H within the emitter function. These results illustrate the effect of severe concentration gradients of these unstable intermediates on spillover-mediated hydrogenation events in the absence of similar gradients in toluene or hydrogen reactant concentrations, consistent with the requirement for proximity.

H-atoms in their neutral or charged forms have been assumed to mediate spillover as the active and mobile intermediate I_H that allows the emitter and acceptor functions to communicate, but only H-atoms (or H⁺-e^- pairs) can migrate over larger than atomic distances and effect hydrogenation events on oxide surfaces. Gaseous H radicals, even at equilibrium with H₂, would collide with acceptor functions at rates much lower than measured rates of spillover-mediated hydrogenation reactions. The prevalent alternate hypothesis, summarized by Prins⁵, is that H-atoms migrate on surfaces, leading to spillover-mediated hydrogenation events. Such processes are unlikely to occur on insulating oxides (and even on semiconductors) at typical hydrogenation conditions⁵ and would, in any case, be contradicted by the measured kinetic dependences of monofunctional and spillover-mediated rates on the pressures of H₂ and the unsaturated co-reactant. We propose instead that H-atoms are transported via gaseous molecular shuttles that form via partial hydrogenation of toluene and donate H-atoms at acceptor sites separated beyond atomic distances. The participation of such intermediates is supported by the observed dependences of I_H formation rates on reactant concentrations on the emitter function and in agreement with dependences of mixture geometry on the transport of intermediates across emitter and acceptor functions.

These results on toluene hydrogenation establish trends in spillover-mediated rate enhancements that identify regimes of kinetic control and mass-transfer control which enable the interpretation of kinetic data and informs the future design of bifunctional catalysts that maximize spillover effects.

The authors acknowledge financial support from Chevron Energy Technology Company and the NSF Graduate Research Fellowship Program.

References

(1) Sinfelt, J. H.; Lucchesi, P. J. J. Am. Chem. Soc. 1963, 85 (21), 3365.

(2) Lin, S.; Vannice, A. J. Catal. 1993, 143 (2), 539.

(3) Zhang, A.; Nakamura, I.; Aimoto, K.; Fujimoto, K. Ind. Eng. Chem. Res. 1995, 34 (4), 1074.

(4) Jung, K. D.; Bell, A. T. J. Catal. 2000, 193 (2), 207.

(5) Prins, R. Chem. Rev. 2012, 112 (5), 2714.