Stability of a fullerenes with reax/c potential

Hello,

I'm using the LAMMPS version from 4 Apr 2014 and I have a problem with fullerenes simulated with the reax/c potential. When my timestep equals 0.1 fs, a kinetic energy of a single fullerene particle quickly increase to approx. 300 eV, which I think is way too much - the kinetic energy of the same particle simulated with the AIREBO potential (with the same timestep) never exceeded 0.01 eV. Could someone explain to me, where that additional energy came from? I had tried ffield.reax.cho, ffield.reax.rdx and ffield.reax.mattsson potentials from "potentials" directory.

I've attached C60 datafile and there is my input script below.

Thanks in advance,
Michal Kanski

  Script:

units metal
atom_style charge
boundary f f f
read_data fullerene.dat

#pair_style reax/c NULL
#pair_coeff * * ffield.reax.cho H C C
#fix reaxc all qeq/reax 1 0.0 10.0 1e-6 reax/c

pair_style airebo 4.0
pair_coeff * * CH.airebo H C C

thermo 100
thermo_modify lost warn flush yes

fix 1 all nve
timestep 0.0001

thermo_style custom elapsed time dt atoms ke pe temp
thermo_modify lost warn norm no flush yes

dump 6 all custom 100 fullerene.lammpstrj id element x y z
dump_modify 6 element H C C sort id

run 5000

fullerene.dat (10.7 KB)

Hello,

I'm using the LAMMPS version from 4 Apr 2014 and I have a problem with fullerenes simulated with the reax/c potential.

Which uses “units real"

When my timestep equals 0.1 fs, a kinetic energy of a single fullerene particle quickly increase to approx. 300 eV, which I think is way too much - the kinetic energy of the same particle simulated with the AIREBO potential (with the same timestep) never exceeded 0.01 eV. Could someone explain to me, where that additional energy came from? I had tried ffield.reax.cho,

From bad unit style.

ffield.reax.rdx and ffield.reax.mattsson potentials from "potentials" directory.

You should not try different force fields blindly, but look through the ReaxFF literature to see if there is one specifically designed (trained) for fullerenes.

Ray

I am grateful for your response. After changing units everything works fine with a proper potential.

Yours sincerely,
Michal Kanski