Molecular electrocatalysis of hydrogen evolution: catalytic design studied by density functional theory

A theoretical framework has been formulated to gain mechanistic insight into molecular hydrogen evolving electrocatalysts. Reduction potentials and pKa’s were calculated for a variety of complexes with density functional theory in conjunction with appropriate experimental references in order to characterize the thermodynamics of the free energy of reaction for proton reduction catalysis. Cobalt diglyoxime catalysts, such as Co(dmgBF2)2 (dmg = dimethylglyoxime), were shown to produce molecular hydrogen from an acidic acetonitrile solution at moderate overpotentials through either a monometallic or bimetallic mechanism. For monometallic hydrogen production, a single Co(II)-hydride was proposed to react with acid in the hydrogen production elementary step. For bimetallic hydrogen production, two Co(III)-hydride intermediates were proposed to react in the hydrogen production elementary step. Changes in experimental conditions, such as acid strength and overpotential, were shown to affect the probabilities of evolving hydrogen through each pathway. Two current peaks in cyclic voltammetry experiments were assigned by the calculations: a reassignment of a peak at ca. –1 V vs SCE of Co(dmgBF2)2 in acetonitrile from the Co(III/II)-hydride couple to the Co(II/I)-hydride couple; and the assignment of a previously unassigned peak just negative of the Co(II/I) couple of Co(dpgBF2)2 in acetonitrile (dpg = diphenylglyoxime) as the Co(III/II)-hydride couple. Theoretical calculations on a series of complexes Co(dRgBF2)2 in acetonitrile provided linear correlations of thermodynamic properties, including reduction potentials and hydride pKa’s, with respect to the Hammett constant of the substituents R, which range from strongly electron-donating to strongly electron-withdrawing. Because two of the reduction potential lines intersect at a Hammett constant consistent with a slightly electron-donating substituent, we predicted that the Co(II/I) couple and Co(III/II)-hydride couple would overlap for complexes with electron-donating substituents and separate for complexes with electron-withdrawing substituents. This prediction was found to be consistent with the peaks in experimental cyclic voltammograms of Co(dmgBF2)2 that we reassigned and Co(dpgBF2)2 that we assigned. The linear correlations enabled the calculation of all thermodynamic properties of a catalyst from the Hammett constant of its substituents, which is a powerful predictive tool for catalyst design. Additional theoretical calculations on a comprehensive set of cobalt, nickel, and iron diglyoxime and diimine-dioxime complexes revealed trends in thermodynamic properties with respect to metal center and oxime bridge. For cobalt and nickel complexes, the anodic shift due to ligand protonation at proton bridges was found to be greater than the electron-withdrawing effect of replacing the proton with BF2 in both acetonitrile and water. The reverse trend was found for iron, which required much greater overpotential to catalyze proton reduction. Theoretical investigations on synthetic avenues for catalyst design of metal oxime electrocatalysts have demonstrated how catalyst modification can affect the mechanism and energetics of hydrogen production as functions of experimental conditions. The concept of ligand noninnocence, in particular ligand protonation along the hydrogen evolution reaction pathway, was applied to a series of cobalt dithiolene complexes. The complex that initially reduces at the most anodic potential electrochemically, Co(mnt)2 (mnt = maleonitrile-2,3-dithiolate), required the most overpotential for hydrogen evolution catalysis, despite having the most strongly electron-withdrawing substituents in the series and having produced hydrogen with the highest turnover frequency photochemically. The theoretical calculations were used to assign protonation states of the ligand, which provided an explanation for the anomalous behavior of Co(mnt)2. While the other complexes in the series became protonated at both dithiolene ligands along the hydrogen evolution reaction pathway, the more strongly electron-withdrawing substituents of Co(mnt)2 resulted in only one ligand protonation. The key cobalt hydride intermediate was proposed to form via intramolecular proton transfer, which was calculated to be thermodynamically favorable after reduction. Understanding the impact of ligand protonation on electrocatalytic activity, along with insight gained from studying the effects of other synthetic modifications, is important for designing more effective electrocatalysts for solar devices