ReaxFF error: inconsistent force field file

LAMMPS LAMMPS General Discussion
1

I would like to use the ReaxFF parameterisation published here. Happily the authors have supplied the parameters in the supporting information. However, when using these parameters with LAMMPS (29 Oct 2020) I am getting the following error:

ERROR: Inconsistent force field file (…/reaxc_ffield.cpp:341)
Last command: pair_coeff * * Ostadhossein_2017_H_S_Mo.reaxff H S Mo

I’ve dug in to reaxc_ffield.cpp and line 341 is in the section which reads the two-body interactions. It seems to check that the atom numbers are positive integers and that the line contains enough parameters.

The relevant section of my force field file is

6 ! Nr of bonds; at1;at2;De(sigma);De(pi);De(pipi);p(be1);p(b p(be2);p(bo3);p(bo4);n.u.;p(bo1);p(bo2)
  1  1 156.0973   0.0000   0.0000  -0.1377   0.0000   1.0000   6.0000   0.8240
         2.9907   1.0000   0.0000   1.0000  -0.0593   4.8358   0.0000   0.0000
  1  2 183.1582   0.0000   0.0000  -0.7544   0.0000   1.0000   6.0000   0.3725
        11.7366   1.0000   0.0000   1.0000  -0.0595   4.6177   1.0000   0.0000
  2  2  86.8868  69.1367   0.0000  -0.9993  -0.4781   1.0000  17.8574   0.0999
         0.2799  -0.1677   8.2557   1.0000  -0.1131   6.1440   1.0000   0.0000
  1  3  77.9675   0.0000   0.0000  -0.5019  -0.3000   0.0000  36.0000   0.1374
         0.0697  -0.3027  15.0243   1.0000  -0.0500   5.9618   1.0000   0.0000
  2  3  64.9870  33.5938 137.6546   1.0000  -0.2304   1.0000  19.1688   0.2000
         1.3265  -0.1497   7.0000   1.0000  -0.1447   6.6437   1.0000   0.0000
  3  3  51.8235   0.0000   0.0000   0.8271  -0.3000   0.0000  16.0000   0.2670
         0.2248  -0.3000  16.0000   1.0000  -0.1908   7.3978   0.0000   0.0000

It looks to me like none of the conditions to throw the error would be met so I’m wondering what I’m missing here! Hopefully it’s a simple fix. Any help or guidance here would be much appreciated. Let me know if you need any more files or further information about the simulation.

2

First off, I would suggest to try with the latest stable release of LAMMPS. This has a largely refactored ReaxFF force field reader with (we hope) better checking and better error messages.

But in order to check this, I would need the entire force field file and your input file as well (data file would be optional since to check force field reading, no atoms would be needed).

3

Thanks for the quick response! I’ve uploaded the input files as a ZIP file which you can download here.

I’ll look in to compiling the latest stable version locally as well as getting the supercomputer IT team to make it available.

4

Thanks for the files. I had a closer look and below is the location where the problem is.

This should be two lines instead of one:

6 ! Nr of bonds; at1;at2;De(sigma);De(pi);De(pipi);
     p(be1);p(b p(be2);p(bo3);p(bo4);n.u.;p(bo1);p(bo2)

Because of this the parser will skip the first line of the bond parameter block and parse the second line. That has 8 numbers instead of the expected 10 and thus you get an error about an incorrect format.

5

Thanks, that’s great! That must have been an issue with how I transcribed the force field from the supplementary material PDF.