Code and Data Availability: Thermal Decomposition Mechanism of CF3SO2F
  via Deep Learning Molecular Dynamics

This repository contains the custom code and relevant data files
associated with the manuscript *"Unraveling the Thermal Decomposition
Mechanism of a Promising SF6 Alternative by Deep Learning Molecular
Dynamics"*.


    🗂️ Repository Contents

  *

    |src/|: json for training and inferring with the deep learning potential (DLP)

  *

    |data/|: Input structures, initial training datasets

  *

    |models/|: Deep neural network potential model

  *

    |scripts/|: Example scripts for running molecular dynamics (MD) simulations using the trained DLP

  *

    |results/|: Sample output files including trajectories and decomposition products analysis


    ⚙️ Requirements

To reproduce the results, install the required dependencies. A suggested
environment includes:

  *

    Python ≥ 3.8

  *

    DeepMD-kit 

  *

    LAMMPS 
  *

    PyTorch or TensorFlow