Accessing negative Poisson's ratio of graphene by machine learning interatomic potentials

Area of Science:

• Materials Science
• Computational Physics
• Nanotechnology

Background:

• Negative Poisson's ratio (NPR) materials offer unique mechanical properties for advanced applications.
• Understanding the origin of NPR is crucial for designing novel functional materials.
• Conflicting explanations exist for NPR mechanisms, particularly in graphene.

Purpose of the Study:

• To elucidate the fundamental mechanism behind the negative Poisson's ratio in graphene.
• To develop accurate computational models for simulating NPR phenomena.
• To reconcile discrepancies in existing research findings regarding NPR origins.

Main Methods:

• Constructed a moment tensor potential for molecular dynamics (MD) simulations of graphene.
• Analyzed the evolution of key geometric parameters during simulations.
• Validated findings against first-principles calculations.

Main Results:

• The increase in bond angle, not bond length, is identified as the primary cause of NPR in graphene.
• Machine learning-based potentials improve the accuracy of MD simulations for NPR.
• Results align with and refine existing first-principles findings.

Conclusions:

• The study clarifies the origin of NPR in graphene, attributing it to bond angle changes.
• Machine learning interatomic potentials enhance the predictive power of MD simulations.
• This work advances the application of ML in multiscale simulations for functional materials.