ACCELERATED DISCOVERIES OF MECHANICAL PROPERTIES OF BORON NITRIDE NANOSHEETS USING MACHINE LEARNING

This research investigates the mechanical properties of Boron Nitride Nanosheets (BNNS) by integrating molecular dynamics (MD) simulation with machine learning (ML) techniques. The work highlights the impact of various factors, including dimension, temperature, strain rate and chirality orientation and their influence on mechanical properties like fracture stress, fracture strain, ultimate tensile strength, and Young's modulus within the context of nanoelectromechanical systems (NEMS). The novel combination of MD simulations and ML models - specifically decision trees, random forest, support vector machines, and polynomial regressions demonstrates that temperature significantly impacts the Young's modulus, and strain rate estimates true agreement with theoretical expectations. These results highlight machine learning potential to advance material property predictions providing a new methodology for interpreting mechanical properties at the nanoscale and paving the way for innovative applications in NEMS.