Identifying systematic DFT errors in catalytic reactions

Using CO2 reduction reactions as examples, we present a widely applicable method for identifying the main source of errors in density functional theory (DFT) calculations. The method has broad applications for error correction in DFT calculations in general, as it relies on the dependence of the applied exchange-correlation functional on the reaction energies rather than on errors versus the experimental data. As a result, improved energy corrections can now be determined for both gas phase and adsorbed reaction species, particularly interesting within heterogeneous catalysis. We show that for the CO2 reduction reactions, the main source of error is associated with the C = O bonds and not the typically energy corrected OCO backbone.