I-ReaxFF: stands for Intelligent-Reactive Force Field

• I-ReaxFF is a differentiable ReaxFF framework based on TensorFlow, with which we can get the first and higher-order derivatives of energies, and also can optimize ReaxFF and ReaxFF-nn (Reactive Force Field with Neural Networks) parameters with integrated optimizers in TensorFlow.

---

• ffield.json: the parameter file from machine learning
• reaxff_nn.lib the parameter file converted from ffield.json for usage with GULP

Installation

The following package needs to be installed

1. TensorFlow, pip install tensorflow --user or conda install tensorflow
2. Numpy, pip install numpy --user
3. matplotlib, pip install matplotlib --user

Install this package after downloading this package and run the command in the shell in the I-ReaxFF root directory pip install . --user. or using a command with editable mode:

pip install . -e

Alternatively, this package can be installed without downloading the package through pip pip install --user irff.

Usage

1. Generating a dataset by DFT calculations
2. Prepare the parameter file 'ffield.json'
3. Train the model

Citation

1. Feng Guo et al., Intelligent-ReaxFF: Evaluating the reactive force field parameters with machine learning, Computational Materials Science 172, 109393, 2020.

2. Feng Guo et al., ReaxFF-MPNN machine learning potential: a combination of reactive force field and message passing neural networks, Physical Chemistry Chemical Physics, 23, 19457-19464, 2021.

3. Feng Guo et al., ReaxFF-nn: A Reactive Machine Learning Potential in GULP/LAMMPS and the Applications in the Thermal Conductivity Calculations of Carbon Nanostructures, Physical Chemistry Chemical Physics, 27, 10571-10579, 2025.

Use ReaxFF-nn with LAMMPS:

https://gitee.com/fenggo/ReaxFF-nn_for_lammps

https://github.com/fenggo/ReaxFF-nn_for_lammps