High-Accuracy Quantum Monte Carlo Study of CO Reduction Intermediates on Cu(111)

Computational understanding of the electrochemical reduction of carbon monoxide (COR) on copper surfaces is crucial for developing efficient catalysts for sustainable fuel and chemical production. Density Functional Theory (DFT) often fails to accurately predict CO* adsorption energies and sites on copper, leading to uncertainties in reaction mechanisms. We employed fixed-node diffusion Monte Carlo (FNDMC) to investigate the adsorption energies of CO, hydrogen, and key intermediates COH and CHO on Cu(111), achieving high-accuracy many-body electronic structure calculations.



Our FNDMC results match experimental CO adsorption data, effectively resolving the "CO adsorption puzzle" by accurately predicting the top site as the most stable at low coverages. In contrast, DFT favors other sites due to its inability to capture electron correlation effects. Both CO* and H* adsorption energies from FNDMC align with experimental benchmarks confirming our method's accuracy. Next, using FNDMC we find that CHO binds stronger than COH on Cu(111) in the vacuum, but water solvation significantly reduces such energy difference. Finally, we discuss the possibility of fully explicit solvation in QMC. Overall, our findings enhance our understanding and precision of adsorption processes on copper surfaces, demonstrating FNDMC's superior accuracy over DFT, and aiding the development of new catalysts.