In this study, we focused on the dissociation of CO* on Co(0001) surfaces, exploring how varying coverages of non-reactive CO* (spectators) influence the transition state (TS) geometries and reaction energy barriers. At low coverage, we apply the nudged elastic band (NEB) method to locate the first-order saddle point, but finding first-order saddle points at high coverage is not trivial. The interactions between adsorbates destabilize the initial and final structures obtained from low-coverage conditions. Moreover, increasing the number of adsorbates increases the number of possible configurations combinatorially, complicating the preparation of NEB calculations.
To address this, our strategy is to use the transition state obtained at low coverage as an initial guess for high coverage. We fixed the slab atoms and the TS guess, using the constrained minima hopping method to place CO molecules on the surface and identify the global minimum configurations at different coverages. Then, we released the constraint for TS and further optimized the system until the accurate first-order saddle point was found.
In this talk, we will discuss both the computational workflow for determining the coverage dependence of activation energies as well as possible strategies for generalizing these results for more high-throughput calculations. The computed coverage-dependent energy barriers will serve as valuable input for the Reaction Mechanism Generator (RMG), enabling more accurate reaction mechanism predictions for complex surfaces. Additionally, we hope this work can contribute to the development of coverage-dependent BEP relationships.