What can Raman spectroscopy really say about the adsorbed CO on roughened Cu electrodes in CO₂ electroreduction...

Electrochemical CO₂ reduction (CO₂RR) offers a promising strategy to recycle carbon by converting CO₂ into valuable fuels and chemicals. So far, Cu-based catalysts remain the most effective for producing high-value multi-carbon products from CO₂. Under the operating conditions of the CO₂RR, Cu is known to undergo significant surface reconstruction. Furthermore, this reconstruction is crucial for creating the active sites for the CO₂RR, whereas flat surfaces catalyze only the hydrogen evolution reaction (HER). This poses challenges for accurately elucidating the structure-activity relationship. To address this, recent research has employed CO as a probe molecule, and combined in situ surface-enhanced Raman spectroscopy (SERS) and density functional theory (DFT) to reveal how surface reconstruction affects not only the macroscopic morphology but also the local active sites, adsorption of key intermediates, and the overall catalytic performance. A specific metric tracked in SERS studies is the intensity ratio of the Cu-CO stretching band to the CO rotational band, as it is believed to be related to the CO coverage. Here, we challenge this assumption. We combine grand canonical genetic algorithm (GCGA) global optimization with grand canonical density functional theory (GCDFT) calculations to find optimal accessible CO coverages under CO₂RR conditions. Then, we use these structures to investigate how different surface roughness, facets, and CO adsorption sites influence the Raman spectra, and the intensity ratio of interest. We show that the changes in the Raman intensity ratio are not solely determined by CO coverage. The adsorption site has a significant influence on the spectral signature, while the realistic reconstruction of the Cu surface impacts the available sites in a potential-dependent manner. The local environment of the adsorbed CO, including the site, coverage, and secondary coordination sphere, affect the stretching and rotation mode frequencies and intensities. We conclude that the SERS signature, while it can be exceptionally useful, alone is insufficient to establish a rigorous, quantitative correlation between the surface coverage or density of species and the production rates of target products required for a clear structure-activity relationship.