CO2 Reduction to CO on Au: Facet Dependent Activity and Selectivity

Abstract

Due to the drastic drop in renewable electricity prices, electrochemical CO2 reduction has received renewed interest. While CO2 reduction can lead to a wide variety of products, almost all go through a CO intermediate, thus understanding the 2 electron transfer reaction of CO2 to CO is important not just for CO production, but also for products such as ethylene, ethanol, methane and others. Au has been shown to be the most active electrocatalyst for CO production, however major fundamental knowledge on this reaction has yet to be discovered. While most works focus on highly active nanostructured Au electrodes, there has yet to be a comprehensive study on the intrinsic electrochemical activity of individual Au facets.This talk will discuss our work on the electrochemical conversion of CO2 to CO on single crystal Au <100>, <111>, <110>, and <211> facets as well as polycrystalline Au. Our results show a 25 fold difference between the least and most active crystal facets. All facets follow approximately the same Tafel slope between the tested potentials of -0.6 V to -0.8 V vs. RHE. Pb underpotential deposition (UPD) was used to deposit a monolayer of Pb on the Au gold surface to determine surface area. However, the potential at which Pb UPD deposits is facet dependent. By setting a correct potential we could selectively deposit lead on a given facet. While PB UPD favored depositing on the most active sites, we still could gain fundamental catalysis knowledge. On polycrystalline Au, 20% of a monolayer of Pb yielded a 50% decrease in CO production, but only a 30% decrease in H2 production. 50% of a Pb monolayer on Au yielded a 95% decrease in CO production, but only a 50% decrease on H2 production. The decrease in CO production as a function of Pb coverage indicated that not only could we passivate the most active sites, but also that the facet dependent activity was quite different for CO production versus the competing H2 evolution side reaction. To further utilize Pb UPD to isolate sites we reinvestigated one of our least active facets, Au <111>. As expected, STM images show that our single crystal is not completely flat and there is a small percentage of step sites. We used selective Pb UPD to deposit 4% of a monolayer in the attempts to passivate these sites. This led to a 50% decrease in CO activity, but the same H2 evolution activity. Covering the Au with 15% of a monolayer led to a 90% drop in CO activity, but the same H2 evolution activity. These results suggest that our intrinsic activity of our least active facets is probably much lower than measured and is dominated by a small number of highly active defect sites. While these results focus on Au, no single crystal is defect free. Thus these results may lead us to also question whether there are other situations where the non-optimal facets for a given reaction are actually much less active than originally thought.