A Combined Density Functional Theory and Microkinetics Simulation Study of Electrochemical CO₂ Reduction on Ceria-Supported Bismuth

Direct electrochemical CO2 reduction (ECR) into carbon-based fuels and chemicals is a promising way to upgrade waste CO2 with renewable energy, contributing to closing carbon cycles and mitigating climate change. Here, we investigate the ECR of Bi–CeO2 catalysts. Using a combination of density functional theory (DFT), artificial neural networks (ANN), genetic algorithms (GA), and microkinetics simulations, we conducted a comprehensive exploration of active sites, the CO2-to-formic acid (HCOOH) mechanism and the electrochemical behavior of Bix/CeO2 catalysts. Three representative models were investigated: (i) a Bi atom adsorption on CeO2 (Bi1/CeO2), (ii) a single Bi atom doped in the CeO2 surface (Bi1–CeO2), and (iii) a small cluster of eight Bi atoms adsorbed on CeO2 (Bi8/CeO2). ANN-GA was employed to identify the optimal structure of the Bi8 clusters on the CeO2 surface. Our investigation shows various reaction pathways for converting CO2 to HCOOH and CO. For the structural models featuring Bi on the surface, the HCOO pathway toward HCOOH is the predominant one. Bi doping in CeO2 predominantly favors the COOH pathway, resulting in CO as the main product. The former models leading to HCOOH exhibit higher current densities than the doped model, which mainly produces CO. Electronic structure analysis shows that stronger electron donation in the Bi1/CeO2 model enhances HCOOH current densities by weakening the O–H bond and stabilizing the transition state. We discuss the kinetic differences in current density and selectivity as a function of the electrochemical potential. These findings not only elucidate various CO2 conversion pathways, which can explain the formation of desirable HCOOH and unwanted CO, but also offer theoretical guidance for the design of electrocatalysts.