nnrf_nitramines: Training Data, NNRF Potentials, and Iterative Code for HE materials

Neural network reactive force field for C, H, N, O systems

npj Comput Mater 7, 9 (2021). https://doi.org/10.1038/s41524-020-00484-3

Abstract:
Reactive force fields have enabled an atomic level description of a wide range of phenomena, from chemistry at extreme conditions to the operation of electrochemical devices and catalysis. While significant insight and semi-quantitative understanding have been drawn from such work, the accuracy of reactive force fields limits quantitative predictions. We developed a neural network reactive force field (NNRF) for CHNO systems to describe the decomposition and reaction of the high energy nitramine 1,3,5-Trinitroperhydro-1,3,5-triazine (RDX). NNRF was trained using energies and forces of a total of 3100 molecules (11941 geometries) and 15 condensed matter systems (32973 geometries) obtained from density functional theory calculations with semi-empirical corrections to dispersion interactions. The training set is generated via a semi-automated iterative procedure that enables refinement of the NNRF until a desired accuracy is attained. The RMS error of NNRF on a testing set of configurations describing the reaction of RDX is one order of magnitude lower than current state of the art potentials.

The NNRF Potentials

The NNRF potentials for Gen1.x, Gen2.3 and Gen3.7 are provided and can be used for Molecule Dynamic simulations for Nitramines.

Gen1.X NNRF for thermal decomposition for RDX crystal

Gen1.1 ~ Gen1.9 are for the thermal decomposition of RDX crytal.

Gen2.X NNRF for thermal decomposition for RDX crystal

Gen2.8 is the NNRF for the thermal decomposition of RDX, HMX, NM, CL20, TNT, TATB, PETN crytals.

Gen3.X NNRF for Shock dynamics for PETN crystal and Liquid NM

Gen3.7 is the NNRF for the shock dynamics of PETN crystal and Liquid NM.

The Iterative code

This is python module to run a iterative process for NNRF training.

The Training data

This is the collection of data for every iteration of the training process for Gen1.X ~ Gen2.X Individual file is different geometry in the format of n2p2 data file.
Please check the input data format of n2p2 (https://compphysvienna.github.io/n2p2/Topics/cfg_file.html).
All individual files were concatenated to train the NNRF. The training data for shock response (Gen3.X) will be available soon.

The energy and forces of data files are corresponding to E(DFT) - E(Reference Potential) and F(DFT) - F(Reference Potential).

In the training of the NNRF, NNRF was designed to learn and predict this difference.
In the evaluation of the NNRF with trained parameters, NNRF + ReaxFF VC will be corresponding to PBE-D2 (Ground truth).

Installing the NNRF python module

Add instruction