CO2 reduction on copper nanoclusters: a DFT analysis using GGA and hybrid functionals

The problem of global warming related to CO2 still demands extensive work to shed light on the mechanisms of its conversion into value-added products in different catalysts. In this sense, theoretical studies generally employ the DFT approach with GGA functionals and periodic surfaces, which has largely amplified our understanding of this process. Nevertheless, achieving a deeper understanding may require improved precision. Here, we analyse the impact of a hybrid functional (ωB97X-D3) on the calculated values of several properties, to shed light on where the most widely used approach can provide accurate results, and to identify possible deficiencies. In this context, reaction free energies, barriers, and solvation effects were evaluated using PBE+D3BJ and ωB97X-D3 functionals on nanocluster surfaces of different sizes (Cu13 and Cu55). For the reaction energies, although we have found significant discrepancies between the two functionals in specific reaction configurations, the result after selecting only the lowest energy mechanism in each functional tends to hide these specific deficiencies. The most relevant caution point is in the reaction barriers, where the two functionals show larger differences, and may cause large deviations in the predicted kinetics, given their sensitivity to errors in the barrier height.