A Combined Spectroscopic and Electrochemical Approach to Investigate the Selectivity Determining Factors for CO2 Electroreduction on Pd-Nanoparticles

Abstract

The electrochemical conversion of carbon dioxide into added-value products keeps on receiving increased interest as a means to valorize CO2 while diminishing its anthropogenic emissions. A broad spectrum of CO2-electroreduction products (e.g., CH4, C2H4) has been reported, but only a few are expected to be economically competitive with regards to industrially established chemical production pathways [1]. Interestingly, Pd-based nanoparticle catalysts are particularly well suited for the production of two of such products, since they feature the unique ability to selectively produce formate and carbon monoxide (CO) at low vs. high overpotentials, respectively [2]. This switching of the selectivity with the applied potential has been linked to a difference in the nature of the Pd catalyst and corresponding reaction mechanism. Specifically, at potentials below − 0.4 VRHE at which CO is the main reduction product [2,3], the reaction is believed to proceed on a fully CO-covered Pd-surface [4]. At higher potentials between − 0.4 and 0 VRHE, though, formate is predominantly produced [2,3] through a hypothesized CO2-hydrogenation mechanism on palladium hydride (PdHx) [5].Chiefly, the design of better-performing catalysts requires a better understanding of the relation between the precise stoichiometry of this hydride phase (i.e., the value of x in PdHx) and the applied potential, reaction duration and corresponding coverage of surface adsorbates (e.g., CO).To determine the relation between the above-described parameters, in this study we used a commercial carbon-supported palladium catalyst (40 % Pd/C) that displayed the above-described switchable selectivity between formate and CO in online gas chromatography measurements in a newly designed parallel plate cell (Figure 1). Specifically, very high selectivities of up to 90 % for formate were found at very low overpotentials (− 0.1 to − 0.3 VRHE), while at high overpotentials (− 0.7 to − 0.9 VRHE) CO is the main product with faradaic efficiencies of over 60 %, and only low amounts of formate are found. The interplay between surface adsorbates and palladium hydride formation was then investigated by quantifying the catalyst’s time and potential-dependent hydride content using chronoamperometry complemented with operando X-ray absorption spectroscopy (XAS) in our group's spectroelectrochemical flow cell [6]. Here, we investigated the influence of CO2 presence in the electrolyte acquiring XAS over 10 minutes of potential hold in both N2- and CO2-saturated potassium bicarbonate.As an example of our results, Figure 2 shows the X-ray absorption near edge structure (XANES) spectra after ten minutes of holding potential at − 0.1 VRHE. While we observed full hydride formation in N2-saturated KHCO3, the palladium in the CO2 saturated electrolyte remained in full metallic state, suggesting that in the presence of carbon dioxide surface adsorbates prevent the formation of PdHx. Additionally, we were able to monitor changes of the palladium hydride stoichiometry in the seconds- to minutes-timescales, thereby enabling us to investigate the hydride formation speeds with an unprecedented time resolution. Finally, the strongly adsorbed surface species formed during CO2 electroreduction were identified and their surface coverage quantified using a novel combined rotating disk electrode (RDE) cyclic voltammetry and in situ external reflection Fourier-transform infrared spectroscopy (FTIR) setup. This allowed us to determine that in the CO2 saturated bicarbonate electrolyte the Pd-surface is covered by a full monolayer of surface-adsorbed CO that prevents hydrogen adsorption and subsequently palladium hydride formation, as unveiled by our XAS results.In conclusion, we have combined electrochemical with spectroscopic methods providing us with the means to improve our understanding of the palladium hydride stoichiometry and adsorbed surface species under CO2 electroreduction conditions. Thus, this contribution will provide novel insight into the relation among the time-dependent bulk and surface composition of Pd-nanoparticles and their link to their CO2-reduction selectivity.Literature:[1] J. Durst et al., Chimia, 2015, 69, 769. [2] M. Rahaman et al., ChemSusChem, 2017, 10, 1733. [3] D. Gao et al., NanoResearch, 2017, 10, 2181. [4] R. Kortlever et al., Catalysis Today, 2015, 244, 58. [5] X. Min et al., J. Am. Chem. Soc., 2015, 137, 4701. [6] T. Binninger et al., J. Elec. Soc., 2016, 163, H906. Figure 1