Ultrasmall Au clusters supported on pristine and defected CeO₂: Structure and stability

Area of Science:

• Materials Science
• Computational Chemistry
• Catalysis

Background:

• Atomistic simulations of supported metal catalysts are complex due to dual components.
• Efficient theoretical tools are needed to map potential energy surfaces (PES) for metal/support interfaces.

Purpose of the Study:

• To apply the stochastic surface walking-neural network (SSW-NN) method to explore the PES of the Au/CeO2 system.
• To identify stable structures and understand the behavior of gold clusters on ceria surfaces.

Main Methods:

• Developed a ternary Au-Ce-O global neural network (G-NN) potential.
• Utilized SSW global optimization and enhanced molecular dynamics.
• Explored segregation and diffusion pathways for gold clusters on CeO2(111).

Main Results:

• Identified ultrasmall cationic gold clusters (e.g., Au4O2) attached to surface defects as the only stable pattern.
• Observed a strong thermodynamic preference for other clusters to form bulk gold.
• Found a high kinetic barrier (>1.4 eV) for segregation and diffusion, despite thermodynamic sintering tendencies.

Conclusions:

• The high thermodynamic stability of ultrasmall cationic gold clusters and kinetic stability on ceria drive the high activity of Au/CeO2 catalysts.
• Global PES exploration using SSW-NN is critical for understanding metal cluster morphology and kinetics on oxide supports.