EEM and ACKS2 ReaxFF validation

Disclaimer

This is work in progress! Do not trust the results yet. Things to be done:

• Figure out why ACKS2 results are slightly different from ReaxFF:
  • Compare xmortr and elcvec arrays.
  • Also compare EEM results
• Figure out why constraints on ACKS2 do not give the same result as disabling charge transfer between fragments. (This could be the right behavior.)

Description

This is a completely self-contained reference implementation of the EEM and ACKS2 models as they are implemented in the ReaxFF model. (In a broader context, EEM and ACKS2 may take different forms.) Many structures are provided to generate reference data that can be used to validate the charges obtained with full-blown ReaxFF implementations.

Requirements

• Python >= 3.0
• Numpy >= 1.0
• Pytest (for testing only)

Installation

Minimal setup

git clone git@github.com:tovrstra/validation_eem_acks2_reaxff.git
cd validation_eem_acks2_reaxff
pip install .

Development setup

git clone git@github.com:tovrstra/validation_eem_acks2_reaxff.git
cd validation_eem_acks2_reaxff
python -m venv venv
echo "source venv/bin/activate" > .envrc
direnv allow
pip install -e .
pip install pytest
hash -r
pytest -v

How to run an example

Just type the following on the command line:

qcalc {eem,acks2} reaxff_parameter_file xyz_file

Energy (due to charges) and charges are just printed on screen. XYZ files for molecules and periodic structures are included in the repository. Two dummy ReaxFF parameter files are provided: ffield_eem and ffield_acks2.

References

Molecular data sets taken from

• S66: http://begdb.com
• Crystals from COD: http://www.crystallography.net/cod/

The mathematical form of the EEM and ACKS2 models in ReaxFF was recently summarized in the following paper:

• 1. Islam, M. M., Kolesov, G., Verstraelen, T., Kaxiras, E. & van Duin, A. C. T. eReaxFF: A Pseudoclassical Treatment of Explicit Electrons within Reactive Force Field Simulations. J. Chem. Theory Comput. 12, 3463–3472 (2016). http://dx.doi.org/10.1021/acs.jctc.6b00432

The ACKS2 model (in its more general form) is described in the following two papers:

• Verstraelen, T., Ayers, P. W., Van Speybroeck, V. & Waroquier, M. ACKS2: atom-condensed Kohn-Sham DFT approximated to second order. J. Chem. Phys. 138, 074108 (2013). http://dx.doi.org/10.1063/1.4791569

• Verstraelen, T., Vandenbrande, S. & Ayers, P. Direct computation of parameters for accurate polarizable force fields. J. Chem. Phys. 141, 194114 (2014). http://dx.doi.org/10.1063/1.4901513