Correcting DFT-based energies to enhance computational catalysis predictions

Kimika Teorikoa Seminar

Density functional theory (DFT) is widely employed in computational catalysis to model reaction mechanisms. Unfortunately, DFT-based predictions of gas-phase free energies entail large errors that prevent an accurate estimation of equilibrium potentials, reaction energies and adsorption energies, which are vital for computational models to rationally screen and design improved electrocatalysts. In this talk, I will show how to calculate such errors for several compounds using different exchange correlation functionals. Subsequently, I will show that such errors can be connected to chemical properties and, therefore, be systematically corrected. Finally, I will detail how DFT errors affect widespread tools such as free-energy diagrams, adsorption-energy scaling relations and volcano plots. Our main conclusion is that even at high levels of theory, gas-phase errors are so large that overlooking them limits the predictive capabilities of computational models of massive use in electrocatalysis.