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

C-H bond activation at lattice O atoms on oxides mediates some of the most important chemical transformations of small organic molecules. 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 C-1-C-4 hydrocarbons and oxygenates using lattice O atoms 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. Bronsted-Evans-Polanyi (BEP) linear scaling relations 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; 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. These effects are included here through crossing potential formalisms that account for the lateness in 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 nonallylic C-H bonds. 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. 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.