The Kinetic Consequences of Oxygen Chemical Potential in Methane Oxidation Reactions on Group VIII Metal Clusters

Ya-Huei (Cathy) Chin 1, Corneliu Buda2, Mnica Garca-Diguez1, Matthew Neurock2**, and Enrique Iglesia1* 1 Department of Chemical Engineering, University of California, Berkeley, CA 94720, USA. 2 Department of Chemical Engineering, University of Virginia, Charlottesville, VA 22904, USA. * iglesia@berkeley.edu, ** mn4n@virginia.edu Introduction Reactions of CH4 with H2O, CO2, and O2 oxidants on Group VIII metal clusters are important for synthesis gas production and the latter (CH4-O2 reactions) for heat generation. Despite the significant efforts spent on elucidating the mechanistic details for these reactions, the kinetically-relevant steps, identity of surface sites, and the relation between them that govern the relative rates and selectivities have remained controversial. Herein, we apply kinetic and isotopic labeling studies together with density functional theory (DFT) calculations to probe the catalytic requirements for CH4-O2 reactions on Group VIII metal clusters and then compare with CH4-CO2/H2O reactions. All of these reactions proceed via an identical sequence of elementary steps, but the identities of the kinetically-relevant step and the most abundance surface intermediates differ depending on oxygen virtual pressure (O2,vir), a fictitious O2 pressure that is equilibrated with chemisorbed oxygen (O*) and therefore a rigorous surrogate of oxygen chemical potential on the surfaces of metal clusters. The O2,vir is derived from kinetic coupling of relevant elementary steps and is thus a strict function of rate and thermodynamic constants. We show that O2,vir prescribes the oxygen coverage and reactivity and also the thermodynamically-stable phase of metal clusters during chemical turnovers. A change in O2,vir leads to a commensurate shift in the identities of the kinetically-relevant step and molecular site structures, in turn resulting in different kinetic regimes with seemingly different rates, apparent kinetic parameters, oxygen selectivity, and structure sensitivity trends. Results and Discussion CO/CO2 ratios in CH4-H2O/CO2 reactions are given by the water-gas-shift thermodynamics on Pt, Ru, Rh, Ni, and Ir clusters under all relevant conditions, as reported elsewhere [1]. In CH4-O2 reactions, however, the oxygen selectivities toward CO are determined by the relative reactive collision probabilities of CO and CH4 with O* intermediates, which are equal to the rate constant ratios of CO to CH4 oxidation. The oxygen selectivities were determined from competitive 13CH4-12CO oxidation reactions; their values were much larger than unity (> 400, 873 K, 33 nm Pt cluster), indicating that CO, if formed and desorbed, would undergo sequential oxidation at much faster rates than the kineticallyrelevant C-H bond activation steps for CH4 conversion. These results are consistent with the exclusive formation of CO2, measured by varying the residence time and O2/CH4 ratio independently, and the ensemble-averaged selectivities (> 150) predicted from DFT barriers for CO and CH4 reactions on a cubo-octahedral Pt cluster of 201 atoms. Thus, CO is formed from a sequential combustion-reforming route. CH4-O2 reactions exhibit complex reaction dynamics. The pseudo first-order rate constants (rCH4/CH4) are a single-valued function that is either constant, inversely proportional to, or proportional to O2/CH4 ratio (labeled as KR1-3, respectively, Fig. 1). These distinct kinetic responses are consequences of the dynamic transitions in O* coverage from saturation to uncovered as the O2,vir decreases. In response to this decrease, the identities of the most abundant surface intermediates shift from O* to Pt (*) and the kinetically-relevant steps from C-H bond activation on oxygen atom site-pairs (O*-O*;

KR1) to oxygen-vacancy site-pairs (O*-*; KR2), then to O2 dissociation on metal atom sitepairs (*-*; KR3). Upon O2 depletion, CH4-H2O/CO2 reactions occur via C-H bond activation on *-* (KR4) site-pairs. C-H bond activation steps in KR 1, 2, and 4 are kinetically-relevant, as confirmed from normal CH4/CD4 kinetic isotope effects (KIE), but are irrelevant in KR 3 with a KIE of near unity. The energetics of the reactant, transition, and product state of these steps are probed using DFT studies on a model cubo-octahedral Pt cluster (1.8 nm diameter). C-H bond activation on O*-O* site-pairs proceeds via hydrogen abstraction coupled with the formation of a methyl radical-like transition state with a barrier of 149 kJ mol-1, consistent with KR 4 KR 3 KR 2 KR 1 KR 5 the measured barrier of 155 kJ mol-1. In contrast, the calculated barriers were lower for similar steps on O*-* (93-120 kJ mol-1 depending on O* coverages) and *-* (79 kJ KR 3 mol-1) site-pairs because the oxidative addition O /CH 200 of the * site into the C-H bond stabilizes the CH3 at the transition state. Rates for O*KR 5 150 assisted C-H activation (on O*-* and O*-O* X 10 Pd sites) increase with increasing Pt cluster size, KR 2 100 because weakly bound O* atoms prevalent on larger clusters are more effective for H 50 KR 2 KR 1 abstraction. The opposite trend was observed Pt KR 1 on *-* site-pairs because smaller clusters with KR 4 X 10 Pd 0 higher Pt coordination are less effective in the 0 0.2 0.4 0.6 0.8 1 O /CH oxidative addition step. Rates do not vary with Figure 1. Pseudo first-order rate constants (873 K) as a single-valued cluster size when they are limited by O2 function of O /CH ratio on Pt and Pd clusters (0.2 % wt. on Al O ) and the transition state structures determined from DFT calculations for the various dissociation. Similar kinetic responses on O2,vir kinetic regimes. are found on Pd and Rh clusters. The relative reactivities of Pt, Rh, and Pd are correlated to the O* binding strengths when O* are involved in the kinetically-relevant steps (KR1-3) but to C* binding strengths when * are involved (KR4). On Pd, however, the clusters undergo phase transition to PdO that is governed solely by the O2,vir, leading to an additional kinetic regime (KR5) where the rates are limited by C-H bond activation on Pd2+-O2- site-pairs. In the oxidation step, the O2,vir is equal to O2 pressure because O2 activation steps are equilibrated, as confirmed from oxygen exchange rates measured in CH4-16O2-18O2 identical to those at chemical equilibrium. In contrast, O2,vir that governs the dynamics of PdO decomposition and the associated activity changes are derived from kinetic coupling of O* scavenging by CH4 and O2 dissociation. As a result, it is determined by the O2/CH4 ratio and the rate constant ratio of C-H bond and O* recombination steps because the higher rates on PdO than Pd prevent O* equilibration on such surfaces. CH4 reactions with O2, H2O, and CO2, are described by an identical sequence of elementary steps but they exhibit complex kinetics, O* selectivity, structure sensitivity and periodic trends that are governed by the chemical potential of oxygen. Significance We provide a rigorous interpretation of the complex kinetics in CH4-O2 reactions, the structure sensitivity effects, periodic trends and their correlation to the chemical potential of oxygen using kinetic, isotopic, and DFT methods. Reference 1. J. Wei and E. Iglesia, J. Phys. Chem. 108, 4094 (2004).
0 20 40 60 80
2 4

rC-H /(CH4) [mol (g-atom Metalsurface-s-kPa)-1]