[lammps-users] Carbon Nanotubes bend with tersoff potential

Dear lammps users,

     I run a simulation that a 200 nanometer (8,8) carbon nanotube
equilibrates with NVE or NVT ensemble at 300 K for a while. I applied
either AIREBO or Tersoff potential in my simulations. Surprisingly,
the VMD visualization shows that the CNT bends when using Tersoff
potential (SiC.tersoff). While, the CNT keeps straight when using
AIREBO potential.

     Initially, I thought the problem is that the default bond-bond
distance of 1.42 Angstrom fits AIREBO potential well but not fits for
Tersoff potential. Thus, I relaxed the unit cell (32 Carbon atoms)
with Tersoff potential. It turns out that the equilibrium bond-bond
distance to be 1.46 Angstrom. However, the CNT still bends when I run
the 200 nm tube simulation again with this relaxed structure.

     I have checked the pressure of system in tersoff potential. The
pressure vibrates between -200 ~ 200 bar, which I assume is normal. I
have attached the log file. Can anyone point out why the CNT bends in
Tersoff potential? Did anyone has the same problem before?

Haibin Chen Ph.D.
Mechanical Engineering Dept
Carnegie Mellon University
Pittsburgh, Pa, 15213

log.txt (13.1 KB)

Hi Haibin,

Can it be because the Tersoff potential has interactions up to the 3rd
NN while the interactions in AIREBO are longer range (4th NN and
farther if you include torsion and LJ)?


Dear Zhun-Yong, thanks for your reply. I did not include torsion and LJ in
AIREBO. I expect the impact of 4th NN in AIREBO is very small especially
for a pristine CNT.