Microstructure evolution under thermo-mechanical operating of rocksalt-structure TiN via neural network potential

Area of Science:

• Materials Science
• Computational Materials Science
• Nanotechnology

Background:

• Protective coatings on cutting tools degrade at high temperatures, impairing mechanical properties.
• The atomic-scale mechanisms of coating failure under combined temperature and stress are not well understood.
• Simulating these complex interactions requires methods capable of large system sizes and high accuracy.

Purpose of the Study:

• To develop a quantum-accurate, large-scale atomic simulation method for titanium nitride (TiN) systems.
• To investigate the fracture behavior of TiN coatings under extreme thermodynamic conditions.
• To elucidate the atomic mechanisms behind coating failure in cutting tools.

Main Methods:

• Developed a deep neural network (DNN) potential for Ti-N binary systems using first-principles datasets.
• Performed large-scale atomic simulations of TiN under varying temperature and stress.
• Compared DNN-potential predictions with traditional empirical interatomic potentials (e.g., embedded-atom-method).

Main Results:

• The DNN-potential accurately predicts lattice constants, phonon, and mechanical properties of TiN.
• Successfully simulated the atomic evolution of fracture in large-scale TiN systems.
• Identified interatomic brittle fracture as the failure mode when TiN exceeds its tensile yield point.

Conclusions:

• DNN potentials offer a powerful tool for quantum-accurate, large-scale atomic simulations.
• The study provides new insights into the fracture mechanisms of TiN coatings under operational stress and temperature.
• Findings can inform the design and understanding of advanced coating tools for extreme environments.