How reliable is DFT for predicting the redox potentials in polypyridine-based hydrogen evolution catalysts?

Earth-abundant 3d metal-based polypyridine complexes have recently emerged as highly promising candidates for hydrogen evolution reaction (HER) catalysts in aqueous or proton-rich solutions. However, the unpredictable variation of the energies and structures of metal complexes with exchange-correlation functionals poses a significant barrier for computational discovery of improved catalytic modules. In the present work, we have performed a comprehensive and systematic benchmark study for accurate and reliable calculation of the redox potential of the electron transfer process, which is considered to be a crucial event in the pathway of hydrogen generation. We have considered a benchmark set of 15 experimentally well-characterized polypyridine-based catalysts with sufficiently diverse chemical functional groups. The catalytic complexes contain Co^II, Fe^III, Ni^II, and Cu^II metal ions. A set of 10 GGA functionals, each of them with one all-electron and one ECP-corrected basis set, has been systematically applied to the catalyst series. The functional space contains pure, hybrid, dispersion-corrected, and long-range corrected functionals. Our proposed methodologies work impressively well in the case of Fe-containing catalysts, which is particularly encouraging considering the high abundance, low cost, and low toxicity of iron metal. The B3LYP-Def2TZVP, B3LYP-LANL2DZ(f)/6-311++G**, and M06-LANL2DZ(f)/6-311++G** are the three top-performing methods, which have mean average errors (MAEs) of 0.24 V, 0.26 V, and 0.25 V, respectively, when the entire benchmark set of all the catalytic complexes is considered. These methodologies can be preferably used for screening and discovery of novel polypyridine-based catalyst modules for the generation of hydrogen gas, which can be a source of clean and renewable energy.