(415f) More Accurate Depiction of Adsorption Energy on Transition Metals Using Work Function As One Additional Descriptor

Transition metals (TMs), benefiting from their active interactions with chemical species owing to partial occupancy of the d bands, represent one most important category of catalysts and have been widely used in many reactions. The catalytic properties of TMs in a chemical reaction are primarily determined by adsorption energy (E_ads) of the involved species. In this regard, it is of high research significance to get accurate E_ads values, however, which is often challenging be measured experimentally. Active research interests have been devoted to determine Eads by establishing an exact relationship with measurable catalyst parameters like the d-band (εd) theory, which was based on the Newns-Anderson model and suggested for use by Hammer and Nørskov. It describes a simple E_ads–ε_d linear correlation and has been widely adopted for application, which turns out to work well for certain adsorption systems, for instance H₂ and CO adsorption, but becomes less effective for many other systems, such as O, OH and OOH adsorption. These controversial results suggest the current E_ads–ε_dlinear correlation could be oversimplified to describe these adsorption systems and εd might not be the only influencing parameter, thus improvements are needed.

Considering the fact that band hybridization and electron transfer could occur simultaneously when a molecule adsorbs to a TM surface, we propound that E_ads contains a mixture of covalent and ionic contributions (i.e., E_ads=E_covalent + E_ionic). The E_covalent term would be describable using the ε_d parameter as it accounts for the band hybridization, whereas the E_ionic term is determined by the charge transfer between TM and adsorbing molecules and thus would be related to the work function (W) parameter. Herein by taking both the ionic and covalent contributions into account, we suggest a new description model for E_ads by dividing it into E_covalent and E_ionic components and adding W as one additional descriptor. We study the adsorption of O, OH, and OOH groups on TM surfaces using the new E_ads–(ε_d, W) model with the explicit formula as E_ads = f(ε_d, W) = z₀+ a × ε_d + k ×(W - W₀)² and achieve significantly better goodness of fit compared to the current E_ads–ε_d model. We further demonstrate the new E_ads–(ε_d, W) model can better predict activity property of TMs in oxygen reduction reaction (ORR) benefiting from a more accurate description of the oxygen binding energy (E_O).