Computationally modeling the effectiveness of Pd/Cu as a diesel oxidation catalyst

Diesel engine emissions are a major cause of air pollution producing carbon monoxide (CO), hydrocarbons, nitrogen oxide molecules (NOx), and particulate matter (PM). Low temperature combustion engines offer a promising solution for reducing NOx and PM emissions, but in turn, the lower temperature interferes with the diesel oxidation catalyst (DOC) causing an increase in CO and hydrocarbon emissions. To combat this, the development of new catalysts is critical. Recent studies show that a Pd/Cu alloy could lead to inhibition-free low temperature oxidation reactions, but the mechanistic origin of the improvement over Pd/Pt alloys remains unknown. To elucidate the mechanism, we first used SurfaceEP, a machine-learning package to rapidly estimate binding energies and identified certain ensembles with promising oxygen binding properties. For isolated Pd atoms in the surface of Cu, we obtained density functional theory (DFT) data for all CO oxidation steps. We are currently incorporating this information into a kinetic Monte Carlo (kMC) model, which will allow us to study the mechanism and activity of well-defined site ensembles. This will provide the necessary fundamental insight that is required to further improve the composition and surface architecture of Pd/Cu diesel oxidation catalyst.