(618c) Descriptors for Reactivity and Selectivity of Dioxygen Activation Routes on Metal Oxides
The re-oxidation tendencies of reduced centers and their consequences for O2 activation rates and pathways can be assessed from theory by rigorously examining the activation barriers for inner and outer sphere routes on oxide catalysts that differ in composition and in electronic properties. Such assessments, however, require realistic models of relevant oxide catalysts and rigorous treatments of triplet-singlet spin crossings and radical-like species along the O2 activation reaction coordinate. Keggin-type polyoxometalate (POM) clusters represent appropriate model oxides because of their known and stable atomic connectivity, modest cluster size, and compositional diversity, which provide diverse reduction properties without significant concomitant changes in atomic structure. These materials have been recently exploited in the identification of accurate reactivity descriptors for acid5 and oxidation6 reactions on metal oxides through systematic benchmarking of theory and experiments. Here, we use density functional theory (DFT) to assess energies for the intermediates and transition states involved in inner and outer sphere O2 activation routes and to develop more accurate descriptors for O2 activation pathways. Two-electron reduced centers formed at various locations in acidic form of Mo (H-Mo) and W based (H-W) POM clusters and at a specific location in Mo-POM clusters with NH4+ charge-balancing cations (nNH4-Mo) were examined, each of which exhibited different stabilities and thus reactivities for O2 activations.
Inner sphere routes involve the cleavage and re-formation of metal-oxygen (M-O) bonds via H2O evolution from two-electron reduced centers consisting of OH pairs (H/OH*), formed in reduction steps, to form O-vacancies (*), which then react with O2 to form OO* species. H2O desorption energies to form * from H/OH* can be expressed as the sum of energies that are required to remove two protons from H/OH* (deprotonation energy; DPE) and to remove an O2- anion from a divalent POM anion (O2--removal energy; ORE) and that are released by forming gaseous H2O from two protons and one O2- removed from the oxide cluster, the value that is independent of the oxide properties. DPE values reflect the stability of the O-H bonds, which can be dissected into ionic (DPEion) and covalent (DPEcov) components. DPEion values reflect the ionic interaction energies between protons and the POM anions and DPEcov reflects the energy required to reorganize structures and charges within the POM cluster upon dissociation of the O-H bonds.5 Analogously, ORE values reflect the stability of M-O bonds, which also have the ionic (OREion) and covalent (OREcov) components. The ionic components of the DPE and ORD values depend strongly on the charge balancing cations (H+ or NH4+) and on the location of the reduced center in a given POM cluster. The sites exhibiting higher local electron density present higher DPEion values but lower OREion; these effects tend to cancel out, leading to H2O desorption energies that depend very weakly on charge-balancing cations or location of the reduced centers for POM clusters with a given addenda atom. The covalent components of DPE and ORE depend weakly on charge-balancing cations and on reduced centers location but are quite sensitive to the identity of the addenda atoms (e.g., Mo or W). W-POM clusters exhibited lower DPEcov values but higher OREcov than that of Mo-POM; these values, however, do not completely off set each other, leading to higher H2O desorption energy (~ 30 kJ mol-1) for W-POM than for Mo-POM clusters. Once formed, O-vacancies (*) can react with O2 to form OO* species via a sequential transfer of two electrons. Transition states for the first electron transfer present the highest barrier along the reaction coordinate. Their energies and structures, however, are similar to that of the reactants (noted as âearlyâ transition states), leading to the activation barriers that are insensitive to the reaction energies and thus to the stabilities of the vacancy site.
Outer sphere routes involve a sequential transfer of proton-coupled-electrons from vicinal surface OH groups at reduced centers (H/OH*) to each O-atom in O2 to form H2O2; such routes circumvent any involvements of M-O bonds during the O2 activation process.1 The kinetically-relevant step involves the transfer of the first H-atom and is mediated by a âlateâ transition state, whose energy and structure resemble that of the product state. Consequently, activation barriers are correlated to the reaction energy and depend on the energy required to dissociate one of the H-atoms from H/OH* species (H-atom dissociation energies, HDE) and on the interaction energy between the âOOH radicals and the surface at the transition state (EintTS).
The different dependence of the two O2 activation barriers on the stability of the reduced centers (* or H/OH*) leads to ratios of their respective rates that depend strongly on H2O desorption energies, which determine the contribution of inner sphere routes, and on HDE and EintTS values, which determine the contribution of outer sphere routes. More stable H/OH* species with larger H2O desorption energies and with smaller HDE and larger EintTS values favor outer sphere routes involving H2O2 as molecular shuttles on one O-atom between distant reduced centers. Such trends are consistent with experimental data from solvated POM clusters in liquid media, which show the preferential outer sphere O2 activation route occurring on W-POM compared to the parallel inner and outer sphere pathways involved in Mo-based POM clusters.1,8 Theoretical methods shown in this work demonstrate the conceptual framework to develop catalytic descriptors for the rates and selectivities of O2 activation routes on metal oxides. The results of this work, in turn, provide the guidance to design oxidation catalysts with high selectivity for either inner or outer sphere O2 activation routes.
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 Extreme Science and Engineering Discovery Environment (XSEDE) supported by the National Science Foundation.
(1) Duncan, D. C.; Hill, C. L. J. Am. Chem. Soc. 1997, 119, 243.
(2) Sazinsky, M. H.; Lippard, S. J. Acc. Chem. Res. 2006, 39, 558.
(3) Rozanska, X.; Kondratenko, E.; Sauer, J. J. Catal. 2008, 256, 84.
(4) Kopp, D. A.; Lippard, S. J. Curr. Opin. Chem. Biol. 2002, 6, 568.
(5) Deshlahra, P.; Carr, R. T.; Iglesia, E. J. Am. Chem. Soc. 2014, 136, 15229.
(6) Deshlahra, P.; Iglesia, E. J. Phys. Chem. C 2016, 120, 16741.
(7) Deshlahra, P.; Carr, R. T.; Chai, S.-H.; Iglesia, E. ACS Catal. 2015, 5, 666.
(8) Duncan, D. C.; Chambers, R. C.; Hecht, E.; Hill, C. L. J. Am. Chem. Soc. 1995, 117, 681.