Hi Sriram, thanks a lot for the suggestion. i will check this out!

Hi Marcos,

As you mentioned this is a very unusual geometry (sp3-like but  planar carbonaceous framework with stretched C-C) , chances are that existing ReaxFF potentials may have problems describing it.

Although ReaxFF parameters are transferable, they have been parametrized for a particular set of geometries, forces, etc. So  you may need to think about reparameterizing to incorporate this particular peculiar configuration into your reference set (if it is indeed a stable geometry at the DFT level).

Best wishes,

Fedor

Maybe ReaxFF parameters fitted for Ni-C-H system developed by Prof. Adri van Duin is better than that in LAMMPS. In my test on that set parameters, the energy curves versus C-C bond has three minimums corresponding to C-C single, double and triple bond respectively. 

Hi Goumans and Sriram,

Unfortunately it seems that I might have to really try and parametrize a new reaxff potential for my weird geometry, which doesn't seem to be a simple task. Sriram's potential might not help me much, as it doesn't include the C-H interactions.

Zongyang, thanks for the suggestion. Would you have the reference for the potential you mentioned? Looks like it has not been included in the December 2014 release of LAMMPS.

Best regards,

Marcos

ReaxFF parameters I mentioned is in supporting information of this reference (J. Phys. Chem. C, 2010, 114 (11), pp 4939–4949). If it is not good enough, you can replace parameters of carbon with that in Sriram's potential. One tips: the format of reax.ffield is strict in LAMMPS due to being read by Fortran code. 

Hi Marcos,

Indeed you can just grab the Ni/C/H force field parameters from Mueller et al. from the SI, they should be in the right format. 

You are also welcome to give the ReaxFF implementation in the ADF modeling suite a try (http://www.scm.com/trial). The currently included force fields are listed here: http://www.scm.com/ReaxFF/forcefields-fix2014.html

We also include parameter optimization using the Metropolis Monte Carlo global optimization technique (by Eldhose Iype et al., http://dx.doi.org/10.1002/jcc.23246, see web presentation by him: http://www.scm.com/Videos/Webinars/OptimizingReaxFFParameters.webm) and we are working on further integration Bernd Hartke's genetic algorithms for ReaxFF parametrization.

With our molecular DFT code ADF and the periodic DFT module BAND you can generate the reference data for (re)parametrization of your ReaxFF force field, but you can use other reference data as well.

Best regards,

Fedor