Disclosing the Oxidation State of Copper Oxides upon Electrochemical CO2-Reduction

Abstract

In recent years, an increasing amount of work has been devoted to study the electrochemical reduction of carbon dioxide, which is regarded as a means to valorize CO2 while possibly closing the carbon cycle. Despite this interest and the large variety of CO2-reduction products that have been reported in the literature, only the reduction towards formate, CO or certain alcohols is expected to be cost competitive if the latter can be produced in sufficiently large volumes.1,2 Specifically, CO2-electroreduction to methanol or ethanol has preponderantly been carried out using catalysts based on copper oxides (CuOx),3 for which the (surface) oxidation state in the course of the reaction and its concomitant effect on the product selectivity remain controversial. This assignment is further complicated by the large surface roughness customarily displayed by these copper oxides (that are often prepared through electrodeposition and thermal treatment steps),4 which upon reduction would lead to an abundance of low-coordinated surface sites that reportedly reduce CO2 to ethanol.5 To untangle these effects, in this study we used DC magnetron sputtering to prepare low roughness Cu2O thin films (TFs) with a controlled oxide thickness of 100 nm, and used these as electrocatalysts for the reduction of CO2. Interestingly, these model electrodes did not display the enhanced selectivity towards ethanol and ethylene usually reported for CuOx, but instead led to a product distribution similar to the one attained with polycrystalline copper. Consistently, grazing angle X-ray diffraction measurements of the TFs prior to and after chronoamperometric (CA) CO2-reduction at - 1.0 V vs. the reversible hydrogen electrode (RHE) suggested the co-reduction of the copper oxide in the course of the reaction. To verify this and additionally study the TFs’ surface oxidation state, complementary X-ray photoelectron spectroscopy (XPS) and Ar-sputtering measurements were performed; chiefly, sample re-oxidation upon disassembly of the electrochemical cell and/or transfer to the XPS were avoided by performing the electrochemical measurements in a N2-filled glovebox and transferring the sample in an air-tight chamber, respectively. As illustrated in Figure 1, this resulted in clear changes in the Cu-2p and -LMM spectra – specifically, whereas in the former no significant differences were observed when the post-CA samples were transferred in air or N2 (likely due to the identical Cu-2p spectra exhibited by Cu0 and Cu2O),6 the Auger spectra were much more sensitive to the transfer means, and the excellent agreement6 between the spectrum of the N2-transferred sample and reports for Cu0 confirmed the complete reduction of the TF upon CO2-electroreduction. Ultimately, these results indirectly point at the surface roughness as the factor determining the enhanced alcohol selectivity of CO2-reduction electrocatalysts derived from (reduced) copper oxides, while highlighting the importance of air-free sample transfer for the post-mortem assignment of surface oxidation states.