Structural transformations in single-crystalline AgPd nanoalloys from multiscale deep potential molecular dynamics

Area of Science:

• Materials Science
• Computational Chemistry
• Nanotechnology

Background:

• Silver-palladium (AgPd) nanoalloys exhibit structural changes during catalysis, but simulation mechanisms are limited by simplified potentials.
• Understanding nanoalloy restructuring is crucial for designing efficient catalysts.

Purpose of the Study:

• To develop an accurate deep-learning potential for AgPd nanoalloys.
• To investigate the shape reconstruction mechanism of AgPd nanoalloys from cuboctahedron (Oh) to icosahedron (Ih) geometries.
• To explore the influence of vacancies and diffusion on nanoalloy transformation.

Main Methods:

• Developed a deep-learning interatomic potential trained on a multiscale dataset.
• Simulated AgPd nanoalloys using the developed potential to study structural evolution.
• Analyzed surface restructuring, phase changes, and atomic diffusion during shape reconstruction.

Main Results:

• The deep-learning potential accurately predicts mechanical properties and formation energies, outperforming Gupta potentials for surface energies.
• AgPd nanoalloys spontaneously transform from Oh to Ih geometries, with restructuring times dependent on composition and size.
• Observed concurrent surface and internal structural changes, influenced by vacancies and promoting Ag outward diffusion in Ih structures.

Conclusions:

• The developed deep-learning potential provides a reliable tool for simulating AgPd nanoalloys.
• The Oh to Ih shape transformation is thermodynamically favorable and involves displacive mechanisms.
• Understanding these restructuring dynamics is key for controlling nanoalloy behavior in catalytic applications.