Unraveling the Role of CeO2 Facets in Regulating the Dispersion, Stability, and Activity of Cu Sub-Nanometer Clusters for Water−Gas Shift Reaction

Abstract The exposed facets of a support critically govern the performance of supported metal catalysts, yet their atomic-scale regulatory mechanisms remain unclear. This study employs multiscale simulations to elucidate how three low-index CeO2 facets—(111), (110), and (100)—regulate the structure, stability, and catalysis of supported sub‑nanometer Cu clusters (Cun, n = 2−10) for the water−gas shift reaction (WGSR). A multistep structure screening approach, combining machine learning molecular dynamics (MLMD) and DFT calculations, reveals that the metal−support interaction (MSI) strength follows the order (111) < (110) < (100), inversely correlating with the intrinsic stability of the bare facets. Stronger MSI on less stable facets promotes higher Cu dispersion, with the dominant species evolving from bilayer clusters (i.e., Cu10) on (111) to linear clusters (i.e., Cu2) on (110) and nearly atomically dispersed states (i.e., Cu2) on (100). To evaluate facet-dependent activity, Cu7—a model with established high activity and notable population for the (111) surface—and the dominant Cu2 species were thus selected as representative models for the (111) and (110)/(100) facets, respectively. The (110) facet achieves optimal stability against Ostwald ripening due to its moderate MSI and highly anisotropic Cu diffusion barriers. Under realistic conditions, the surfaces are precovered by H/OH species, which nonlinearly modulate adsorption and activity through electronic and steric effects. Microkinetic modeling (kinetic Monte Carlo simulation) predicts the general activity trend Cu7/CeO2(111) > Cu2/CeO2(110) > Cu2/CeO2(100). Carboxyl formation is identified as the rate‑determining step on (111) and (110), while a high barrier for H2 formation limits the activity on the (100) facet. This work provides a detailed atomic‑scale picture of the “structure → MSI → dispersion → performance” relationship, revealing that overall catalytic performance is governed by a trade-off between thermodynamic stability (governing the abundance of active species) and kinetic factors (intrinsic activity and sintering resistance). CeO2(110) optimally balances this trade-off, offering moderate cohesion energy to stabilize small clusters without triggering detrimental over-dispersion or thermodynamically driven sintering. These insights offer key guidelines for rational design of supported catalysts via facet control.