The reduction degree of TiO₂ support is critical to the performances of metal catalysts. In many previous theoretical calculations, only the bridge oxygen vacancy (O_v) was considered as the electron-donating defect on reduced rutile TiO₂ (r-TiO_2−x) supports. However, titanium adatoms (Ti_ad.), oxidized titanium islands (Ti_ad.O_n), and acid hydroxyls (O_brH) also exist at the metal/support interface. By conducting density functional theory (DFT) calculations and ab initio molecular dynamics (AIMD) simulations, we compared r-TiO_2−x surfaces with O_v, Ti_ad., Ti_ad.O_n, and O_brH sites loaded with Au nanoparticles (NPs). The results showed the Au NPs were oxygen-phobic but titanium-philic, resulting in wetting of O_v and Ti_ad. but short contact with Ti_ad.O_n and O_brH. The Bader charges of Au NPs (Q_M) showed a good linear relationship with the ideal number of donating electrons (N_e) from the defective sites (Q_M = −K_eN_e + Q_M,S), demonstrating the intrinsic electron allocation at the interface. The O_v, Ti_ad., and Ti_ad.O_n exhibited similar slopes (K_e), relatively steeper than that of O_brH. That means in the scope of Au NP charge state, the Ti_ad. and Ti_ad.O_n have a close electron-donating ability with O_v, but the O_brH donates relatively fewer electrons. This linear relationship can be extended approximately to other metals. The higher the metal work function, the steeper the K_e for easier electron donation from defective sites. The stronger the metal oxygen affinity, the more positive the intercept (Q_M,S). That explains the easy generation of metallic or negative Pt and Au NPs on r-TiO_2−x, but hard for Cu and Zn in experiment. That provides theoretical guidance for regulating the charge of metal NPs over TiO_2−x supports.

Electrochemical CO₂-reduction reaction (CO₂RR) is a promising way to alleviate energy crisis and excessive carbon emission. The Cu-based electrocatalysts have been considered for CO₂RR to generate hydrocarbons and alcohols. However, the application of Cu electrocatalysts has been restricted by a high onset potential for CO₂RR and low selectivity. In this study, we have designed a series of Cu-based single-atom alloy catalysts (SAAs), denoted as TM₁/Cu (111), by doping isolated 3d-transition metal (TM) atom onto the Cu (111) surface. We theoretically evaluated their stability and investigated the activity and selectivity toward CO₂RR. Compared to the pure Cu catalyst, the majority TM₁/Cu (111) catalysts are more favorable for hydrogenating CO₂ and can efficiently avoid the hydrogen-evolution reaction due to the strong binding of carbonaceous intermediates. Based on the density functional theory calculations, instead of the HCOOH or CO products, the initial hydrogenation of CO₂ on SAAs would form the *CO intermediate, which could be further hydrogenated to produce methane. In addition, we have identified the bond angle of adsorbed *CO₂ can describe the CO₂ activation ability of TM₁/Cu (111) and the binding energy of *OH can describe the CO₂RR activity of TM₁/Cu (111). We speculated that the V/Cu (111) can show the best activity and selectivity for CO₂RR among all the 3d-TM-doped TM₁/Cu (111). This work could provide a rational guide to the design of new type of single-atom catalysts for efficient CO₂RR.