ReaxFF parameter file format error

LAMMPS LAMMPS General Discussion
,
1

I want to, in advance, say I’m not sure if this is something I should be asking here but I would appreciate it very much if you’d share your experience.
I am very familiar with ReaxFF and have developed several parameter sets myself, but for the life of me, I can’t figure out what is wrong with this file that I got from this paper:
https://pubs.acs.org/doi/10.1021/acs.jpcb.9b04298?src=getftr&utm_source=sciencedirect_contenthosting&getft_integrator=sciencedirect_contenthosting

NH3_x25.lmpdat (8.1 KB)
substrate_1slab2.lmpdat (161.0 KB)
in_1slab.input (2.3 KB)
COHN.reax (8.4 KB)

I have uploaded some files for you to reproduce. I get this error regarding the Atoms section of the file.
ERROR on proc 0: Not a valid floating-point number: ‘H’ (…/reaxff_ffield.cpp:584)
Last command: pair_coeff * * COHN.reax C O N C C C H H C N H
ORTE_ERROR_LOG: Out of resource in file …/…/…/orte/util/show_help.c at line 501

Which says probably it can’t parse the columns right and it think a floating number should be where ‘H’ atom type is defined.

Thank you very much and I understand if you respond with: this is not a LAMMPS problem.

2

Please see ERROR on proc 0: Not a valid floating-point number: 'O' (../reaxff_ffield.cpp:584) - #6 by Iakovos_Delasoudas. This seems to be the same issue with a similar but different parameter set.
I can read/parse both files with the latest stable update version and the latest develop version.

3

The issue in my case was that the parameter set could not support the low gradient vdW correction term. So if you put the lgvdw no maybe will work.

4

I compiled the lammps-patch_2Apr2025 and I got "not compatible with lgvdW’ error.
and yes

Yes, I removed the lgvdW from the pair_style and it is working now.

This from the documentation helped:
“Note: Force field files are different for the original or lg corrected pair styles, using the wrong ffield file generates an error.”