Konstantijn T. Rommens, Thobani G. Gambu, Mark Saeys
First-principles microkinetic models are instrumental to elucidate reaction mechanisms and guide catalyst design, yet they often rely on low-coverage DFT calculations and therefore fail to describe the crowded surfaces present during low-temperature and high-pressure catalytic reactions. By employing realistic surface coverages in DFT calculations for intermediates and transition states, a dual-site microkinetic model for the methanation of CO 2 over nickel catalysts was constructed. Several pathways were considered, including H- and OH-assisted CO x activation, which are crucial to lowering the activation energies and increasing reaction rates. Simulations for a packed bed reactor show that the reaction occurs via intermediate formation of CO, quickly reaching a pseudo-steady-state CO partial pressure where the net rate of CO formation from CO 2 is equal to the rate of CO conversion to CH 4 . Simulations based on realistic surface coverage show significant quantitative and qualitative improvement over microkinetic models constructed with low-coverage DFT calculations. Turnover frequencies vary around 10 –3 s –1, close to experimentally observed values, with high CH 4 selectivities at CO 2 conversions above 1%. Simulated total coverages are close to the coverages used in the reference DFT calculations, but the ratio between CO* and H* is strongly temperature- and pressure-dependent. In the reaction mechanism, CO 2 activation primarily occurs on the terraces and is the main rate-controlling step for CO 2 activation, while CO is activated on B 5 sites via a COH # intermediate, one of the rate-controlling steps in CH 4 formation. Without B 5 sites, the CO concentration reaches a reverse water–gas shift equilibrium before slow methane formation starts. The dominant reaction mechanism identified by microkinetic models based on DFT calculations for realistic surface coverages qualitatively differs from the mechanism identified based on low-coverage DFT calculations, illustrating the crucial role of modeling realistic coverages for reactions on crowded surfaces.