Solvation effects on DFT predictions of ORR activity on metal surfaces

We present a comparison study between the implicit and explicit solvation approach for density functional theory (DFT) predictions of the oxygen reduction reaction (ORR) activity on Pt (111) and other metal surfaces under acidic conditions. DFT calculations with a self-consistent polarizable continuum model implement in VASPsol results in more accurate predictions of onset potentials for Pt(111) than vacuum DFT calculations due to the extra stabilization of the surface intermediates (OOH*, O*, OH*). Implicit solvation also preserves the scaling relationship among ORR intermediates and volcano-shape activity relationships that correlate ORR onset potential or activity to the free energy absorption of O* or OH*. Moreover, VASPsol predicts variation in the solvation energies across different surfaces, suggesting the use of universal solvation corrections may not be valid. However, VASPsol exhibits significant weaker OH* solvation energies (by ∼ 0.4 eV) on Pt (111) compared to literature values using an explicit water bilayer and therefore underestimates the onset potential on Pt (111). We attribute this lack of OH* solvation energies by VASPsol to its inability to address hydrogen bonds and the absence of intermediate stabilization of the bilayer structure. Strategies to mitigate this problem, including tuning VASPsol parameters and a hybrid approach incorporating one to two water molecules, are also examined but not found to alleviate the underestimation of the solvation of OH*.