The reaction mechanisms for alkane hydroxylation catalyzed by non-heme FeVO complexes presented in the literature vary from rebound stepwise to concerted highly asynchronous processes. The origin of these important differences is still not completely understood. Herein, in order to clarify this apparent inconsistency, the hydroxylation of a series of alkanes (methane and substrates bearing primary, secondary, and tertiary C
H bonds) through a FeVO species, [FeV(O)(OH)(PyTACN)]2+ (PyTACN = 1-(2′-pyridylmethyl)-4,7-dimethyl-1,4,7-triazacyclononane), has been computationally examined at the gas phase and in acetonitrile solution. The initial breaking of the C
H bond can occur via hydrogen atom transfer (HAT), leading to an intermediate where there is an interaction between the radical substrate and [FeIV(OH)2(PyTACN)]2+, or through hydride transfer to form a cationic substrate interacting with the [FeIII(OH)2(PyTACN)]+ species. Our calculations show the following: (i) except for methane in the rest of the alkanes studied, the intermediate formed by R+ and [FeIII(OH)2(PyTACN)]+ is more stable than that involving the alkyl radical and the [FeIV(OH)2(PyTACN)]2+ complex; (ii) in spite of (i), the first step of the reaction mechanism for all substrates is a HAT instead of hydride abstraction; (iii) the HAT is the rate-determining step for all analyzed cases; and (iv) the barrier for the HAT decreases along methane → primary → secondary → tertiary carbon. The second part of the reaction mechanism corresponds to the rebound process. Therefore, the stereospecific hydroxylation of alkane C
H bonds by non-heme FeV(O) species occurs through a rebound stepwise mechanism that resembles that taking place at heme analogues. Finally, our study also shows that, to properly describe alkane hydroxylation processes mediated by FeVO species, it is essential to consider the solvent effects during geometry optimizations. The use of gas-phase geometries explains the variety of mechanisms for the hydroxylation of alkanes reported in the literature

This work has been supported by Ministerio de Economiá y Competitividad of Spain (Projects CTQ2014-54306-P, CTQ2014-52525-P, and CTQ2012-37420-C02-01, Ramón y Cajal contract to A.C., and grant No. BES-2012-052801 to V.P.), Generalitat de Catalunya (project numbers 2014SGR931, 2014SGR862, Xarxa de Refereǹ cia en Quiḿ ica Teor̀ ica i Computacional, and ICREA Academia prizes for M.S. and M.C.), the European Commission (ERC-2009-StG-239910 to M.C. and FP7-PEOPLE-2011-CIG-303522 to A.C.), and European Fund for Regional Development (FEDER grant UNGI10-4E-801)