Converging Grand Canonical Ensemble Models to the Computational Hydrogen Electrode—Appropriate Microsolvation Models

The computational hydrogen electrode (CHE), a nearly 20 year old model for computing reaction energetics of coupled proton-electron transfer steps, has been a transformative tool for understanding Sabatier limits in catalytic processes relevant to energy storage and retrieval. However, a critical shortcoming of the CHE is its intrinsic inability to capture <i>kinetics</i>, a key quantity in predictive models for catalytic activity. Towards addressing this gap, grand canonical ensemble (GCE) models have emerged in the past 10 years, enabling a more realistic charging response and calculation of energetics at constant potential. In this work, we illustrate that nearly all reports involving GCE calculations, particularly those that involve production of hydroxide, significantly underestimate ionic solvation energies due to an insufficient microsolvation model that is slower to converge than may be intuitively expected. The consequences of this underestimation are discussed, with significant implications for our understanding of pH effects in electrocatalysis for fundamental reactions. Recommended practices, including the sufficiency of semi-local functionals, are discussed, with our results suggesting that GGA functionals may fail to capture ionic solvation energetics.