The impact of coverage dependence, arising from lateral adsorbate interactions, on microkinetic modeling predictions is well known. The influence of coverage dependence on binding/activation energies of surface reactions can be elucidated using Density Functional Theory (DFT) calculations, which treat the surface as a periodically repeated unit cell. However, the translation of these results to appropriately parameterize microkinetic models that invoke commonly-employed mean-field assumptions remains an intellectual challenge. In this work, we develop a strategy to parameterize a mean-field approach for a prototypical two-step catalytic reaction network with a generic adsorbate interaction model, using results from previously employed and accurate lattice-based kinetic Monte Carlo (kMC) simulations as the benchmark. We compare statistically averaged activation barriers from kMC simulations across a broad range of reaction conditions, against activation barriers predicted from configurationally-distinct coverage representations on a 4x4 periodic unit cell and deduce the appropriate active site and spectator arrangements from these comparisons. We posit a single descriptor, capturing the neighboring information around the active site and show that it correlates one-to-one with the steady state coverage from the kMC simulations. These results point to a lucid protocol that utilizes a handful of carefully chosen DFT calculations to construct and evaluate coverage-cognizant mean-field microkinetic models.
