Improved hydrogen evolution activity descriptors from first-principles electrochemical kinetics

Catalyst activity in the Hydrogen Evolution Reaction (HER) is traditionally estimated based on Sabatier’s principle, with hydrogen’s adsorption strength as its sole variable. This elegant descriptor qualitatively captures HER activity trends but lacks information on electrolyte composition and exhibits distinct outliers like Pt’s exceptionally high and Cu’s low activity. Numerous studies have added mechanistic insights from first-principles; however, they have generally been limited to analyzing kinetics in acid. Here, we use constant-potential DFT coupled with Gaussian process-based transition state search to compile microkinetic models for alkaline HER. Accounting for both non-Nernstian effects and adsorbate-adsorbate interactions, we reproduce experimental activity trends, as well as outliers in the conventional volcano. Relating activities to descriptors, we find that both refining the traditional descriptor with catalyst surface charge information and applying the H binding strength in the (generally) unfavorable top-site exhibit largely improved predictive power for exchange current densities. This stems from their ability to capture trends in the kinetics of electrochemical hydrogen adsorption, the rate-limiting step on all investigated surfaces, as a consequence of adsorbate-adsorbate interactions. However, these descriptors are not applicable in predicting trends in the kinetics of non-electrochemical H₂-desorption, which instead follow the difference in binding strengths between fcc- and top-sites and an effective electronic coupling term. Combining all identified descriptors, we define novel activity volcanoes, which can aid screening for abundant and active HER catalysts. This shows how mechanistic analysis of even the simplest electrocatalytic reactions can still improve our understanding of reactive processes at the electrified solid–liquid interface.