Predicting CO Interaction and Activation on Inhomogeneous Ru Nanoparticles Using Density Functional Theory Calculations and Machine Learning Models

Accurately predicting catalyst performance based on theoretical models is crucial for the rapid and cost-effective design of materials with specific catalytic functions and enhanced durability. In this study, we investigate the interaction and activation of CO on isolated Ru nanoparticles of varying sizes, approximated by adsorption energies and vibrational frequencies, respectively. By employing a linear combination of the nanoparticles’ geometric and electronic properties prior to the CO interaction, we describe the CO interaction and activation. Through rigorous validation, we demonstrate the reliability of the models and their ability to predict the influence of the nanoparticle size on the CO adsorption energies and vibrational frequencies. To estimate the effect of Al₂O₃ as a support material on CO adsorption and activation, we utilize the validated models, effectively bypassing the high computational cost of density functional theory calculations. Our findings reveal that the presence of the support material leads to increased instability in CO adsorption, particularly near the Ru/Al₂O₃ interface. Furthermore, this study uncovers the localized effect of the support material, shedding light on its pivotal role in facilitating CO activation and lowering the activation energy. These insights have significant implications for catalytic reactions, particularly those encountered in Fischer-Tropsch reactors used for the production of synthetic liquid fuel.