The methane-to-methanol reactions that we aimed to evaluate consist of four critical steps starting from the initial reactants (CH4, NH4+, oxo): C-H activation followed by a hydrogen atom abstraction (HAA) from CH4 to form some combination of the hydrogenated complex, ammonium or ammonia, and a methyl radical; a radical rebound (RR) to form a methanol adduct and ammonium; methanol dissociation from the metal; and catalysis regeneration via oxygen atom transfer. We modeled these steps as stationary points throughout the reaction coordinate, each point consisting of a sum of the reactants, intermediates, or products of each step. Our research mainly focuses on the energy barrier of the C-H

*…show more content…*

We performed all calculations with the Gaussian09 package [17] at the B97D/6-31+G(d) level of theory [18], with reactions simulated in an acetonitrile solvent, a polar solvent, by the SMD continuum solvation model [19] to stabilize anion and cation behavior. Our jobs specifically involved density functional theory (DFT) to approximate solutions to the Schrödinger equation of our systems [31]. These methods assume certain parameters for the system and work off these parameters to arrive at a solution, which communicates information about a system’s electronic behavior, structure, and thermochemistry. By extracting key data points concerning electronic and thermodynamic properties of interest, we can then draw conclusions on the favorability of our various metal

*…show more content…*

Compared to the straight optimization of a non-transition state system to a ground state, transition states are more numerically sensitive and take more time to calculate. To achieve this level of accuracy, we continuously modified second derivative (force constants and frequencies) calculation criteria, which was either calculated initially by Gaussian and then approximated in further steps (Opt=CalcFC), or re-calculated at each step (Opt=CalcAll). Approximating the force constants was usually enough, though we turned to a detailed re-calculation of force constants at each step for particularly difficult optimizations, albeit at the expense of computational resources and time. Using the ModRedundant function, we also isolated the minima and maxima optimizations by freezing the active bond(s) of interest and forcing a minimization of the surrounding system, then unfreezing the active bond and optimizing the structure again to a local maximum. Monitoring the average root mean squared (RMS) of the forces in our system at every step of the reaction cycle was also helpful in resolving calculation errors. RMS forces lower as the calculation approaches convergence; recording this property allowed us to reevaluate molecules using