Electron Transfer Limitation in Carbon Dioxide Reduction Revealed by Data-Driven Tafel Analysis

Carbon dioxide (CO₂) recycling holds promise to mitigate anthropogenic emissions and to increase the sustainability of many chemical and fuel production processes. Despite marked advances in catalyst activity and selectivity at laboratory scale, fundamental understanding of the electrocatalytic reduction of CO₂ remains limited, resulting in great uncertainty when extrapolating data to industrially relevant reaction rates. Importantly, the predominant models apply linear Tafel extrapolation, which drastically overpredicts the current density at large overpotentials. Researchers have posited several models to explain the curvature in Tafel behavior for CO₂ reduction catalysis. Here we compare the ability of select models using Bayesian inference to explain curvature in Tafel behavior within the context of CO₂ reduction to CO catalyzed by gold surfaces. By harvesting Tafel data on gold surfaces from multiple literature sources in a variety of reactor configurations, we identify three important features common to the aggregate data on Au-mediated CO₂ reduction: (1) curvature in the Tafel plot at high overpotentials is only partly caused by mass transfer limitations; (2) the Marcus-Hush-Chidsey model for rate-limiting single-electron transfer kinetics provides the best fit to the data of the models tested; and finally, (3) the highly varied data collapse onto a single curve governed by the maximum predicted current in the electron-transfer-limited model. This analysis sets a foundation for determining more accurate activity-driving force relationships for CO₂ reduction on electrocatalytic surfaces, both improving the quality of system-level analyses and motivating further research into the underlying mechanisms of CO₂ reduction catalysis.