Signatures of paracrystallinity in amorphous silicon from machine-learning-driven molecular dynamics.

Area of Science:

• Materials Science
• Condensed Matter Physics
• Computational Materials Science

Background:

• Decades of research on amorphous silicon structure have yielded two main theories: continuous random network and paracrystalline models.
• The paracrystalline model suggests localized structural order within an amorphous network, but its extent and experimental verification remain debated.
• Conflicting experimental interpretations have hindered a unified understanding of amorphous silicon's structure.

Purpose of the Study:

• To investigate the compatibility of paracrystalline models with experimental observations of amorphous silicon.
• To elucidate the boundary between amorphization and crystallization in silicon using advanced simulation techniques.
• To provide a unified explanation for existing, seemingly contradictory theories on amorphous silicon structure.

Main Methods:

• Utilized quantum-mechanically accurate, machine-learning-driven simulations to explore silicon's configurational space.
• Systematically sampled configurations to define the amorphization-crystallization boundary.
• Analyzed simulation data using structural and local-energy descriptors.

Main Results:

• Demonstrated that signatures of paracrystallinity are consistent with experimental data for amorphous silicon.
• Simulation results support the validity of paracrystalline models in explaining experimental observations.
• Established a unified framework reconciling continuous random network and paracrystalline theories.

Conclusions:

• Paracrystalline models offer a consistent explanation for experimental findings in amorphous silicon.
• Advanced simulations provide crucial insights into the structural nuances of amorphous materials.
• This work unifies long-standing debates regarding the structure of amorphous silicon.