“…Currently, continuum solvation models such as Vaspsol (Mathew et al, 2014(Mathew et al, , 2019, COSMO (Klamt and Schüürmann, 1993), or Jaguar (Bochevarov et al, 2013) are commonly used to describe solvation at the electrode/electrolyte interface (Abidi et al, 2020;Basdogan et al, 2020b;Groß, 2020). However at least in the case of the intermediates relevant to the ORR and CO 2 reduction reactions on Cu, Au, and Pt, the widely used continuum solvation models did not improve the accuracy of binding energies compared to the respective calculations in vacuum (Heenen et al, 2020). Due to the complexity and computational costs of AIMD, solvation models by the inclusion of machine-learning techniques have been developed (Basdogan et al, 2020b), which are a promising alternative to continuum solvation models.…”