Reactivity and Selectivity Descriptors for the Activation of C-H Bonds in Hydrocarbons and Oxygenates on Metal Oxides

Reactivity and Selectivity Descriptors for the Activation of C-H Bonds in Hydrocarbons and Oxygenates on Metal Oxides

Prashant Deshlahra and Enrique Iglesia*

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

*iglesia@berkeley.edu

C-H bond activations at lattice O-atoms on oxides mediate some of the most important chemical transformations of small organic molecules [1-2]. The relations between molecular and catalyst properties and C-H activation energies are discerned in this study for the diverse C-H bonds prevalent in C₁-C₄ hydrocarbons and oxygenates using lattice O-atoms of polyoxometalate (POM) clusters with a broad range of H-atom abstraction properties. These activation energies determine, in turn, attainable selectivities and yields of desired oxidation products, which differ from reactants in their C-H bond strength. Brønsted-Evans-Polanyi (BEP) linear scaling relations [3-4] predict that C-H activation energies depend solely and linearly on the C-H bond dissociation energies (BDE) in molecules and on the H-atom addition energies (HAE) of the lattice oxygen abstractors. These relations omit critical interactions between organic radicals and surface OH groups that form at transition states that mediate the H-atom transfer, which depend on both molecular and catalyst properties (Fig. 1a); they also neglect deviations from linear relations caused by the lateness of transition states. Thus, HAE and BDE values, properties that are specific to a catalyst and a molecule in isolation, represent incomplete descriptors of reactivity and selectivity in oxidation catalysis (Fig. 1b). These effects are included here through crossing potential formalisms that account for the lateness of transition states in estimates of activation energies from HAE and BDE and by estimates of molecule-dependent, but catalyst-independent, parameters that account for diradical interactions that differ markedly for allylic and non-allylic C-H bonds (Fig. 1c). The systematic ensemble-averaging of activation energies for all C-H bonds in a given molecule show how strong abstractors and high temperatures decrease an otherwise ubiquitous preference for activating the weakest C-H bonds in molecules, thus allowing higher yields of products with C-H bonds weaker than in reactants than predicted from linear scaling relations based on molecule and abstractor properties [5].  Such conclusions contradict the prevailing guidance to improve such yields by softer oxidants and lower temperatures, a self-contradictory strategy, given the lower reactivity of such weaker H-abstractors. The diradical-type interactions, not previously considered as essential reactivity descriptors in catalytic oxidations, may expand the narrow yield limits imposed by linear free energy relations by guiding the design of solids with surfaces that preferentially destabilize allylic radicals relative to those formed from saturated reactants at C-H activation transition states.

Figure 1. (a) A thermochemical cycle description of the C-H activation transition-state energy relative to a gaseous organic reactant and a surface metal oxide site (MO*) as a sum of, (i) the energy required to separate H atom from the C atom in the reactant to form a radical species (C-H BDE), (ii) the energy of adding the H atom to the O atom of MO* (HAE), (iii) and interaction energy between the radical (R•) and hydroxylated metal oxide (•MOH*) at the transition state ( ). (b) DFT-derived C-H bond activation energies in alkanes (closed symbols), alkanols (open symbols), alkenes (half-filled symbols) and alkanals (crossed symbols) at a specific O-atom location in H₃PMo₁₂O₄₀ POM cluster as a function of their C-H bond dissociation energy (BDE). (c) DFT-derived C-H bond activation energies as a function of the sum of the C-H BDE, of reactants, the HAE of different abstractors and the product state interaction energy ( ). Dashed curves represent regressed values of the activation energies to the functional form described by C-H O-H crossing-potential models for C-H activation.

Acknowledgments

This work was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under contract number DE-AC05-76RL0-1830. Computational facilities were provided by the Environmental Molecular Science Laboratory (EMSL) at Pacific Northwest National Laboratory (PNNL), a DOE Office of Science User Facility, under proposal number 48772.

References

[1] Mamedov, E. A.; Corberán, V. C. Appl. Catal. A: General 1995, 127, 1-40.

[2] Labinger, J. A.; Bercaw J. E. Nature 2002, 417, 507-514.

[3] Bronsted, J. N. Chem. Rev. 1928, 5, 231-338.

[4] Evans, M. G.; Polanyi, M. Trans. Faraday Soc. 1938, 34, 11-24.

[5] Zboray, M.; Bell, A. T.; Iglesia, E. J. Phys. Chem. C 2009, 113 12380-12386.