Neural Network Quantum Molecular Dynamics, Intermediate Range Order in GeSe₂, and Neutron Scattering Experiments.

Area of Science:

• Materials Science
• Condensed Matter Physics
• Computational Chemistry

Background:

• Covalent glasses and melts exhibit intermediate range order.
• This order is observed as the first sharp diffraction peak (FSDP) in neutron scattering.
• Previous atomistic simulations had limitations in system size or accuracy.

Purpose of the Study:

• Investigate the first sharp diffraction peak (FSDP) in Germanium Diselenide (GeSe2) glass and melt.
• Utilize a novel simulation method for accurate, large-scale atomistic modeling.
• Enable quantitative comparison between simulation and experimental neutron scattering data.

Main Methods:

• Employed neural-network quantum molecular dynamics (NNQMD) for large-scale simulations.
• Validated NNQMD against quantum molecular dynamics (QMD) and experimental data.
• Analyzed system-size dependence of FSDP height.

Main Results:

• NNQMD achieved validated quantum mechanical accuracy for large systems.
• Calculated FSDP heights quantitatively matched experimental neutron scattering data for GeSe2.
• Presented detailed structural analyses including pair distribution and structure factors.

Conclusions:

• NNQMD is a powerful tool for studying intermediate range order in glasses and melts.
• The method overcomes limitations of previous simulation techniques.
• Accurate simulation of FSDP provides insights into glass and melt structures.