graphene tension test, results depend on box size in perpendicular direction?

Dear All,

Here is a weird problem in my simulations. I am using AIREBO potential with cut off for LJ as 3.0, which means the interaction radius is around 10.2 angstrom.

Graphene sheet is in xy plane. Periodic boundary condition in all three dimension are used. Therefore, if the height of simulation box is bigger than 10.2 angstrom, no interaction should be present.

However, different box heights generate quite different results. How could I understand that?

By the way, the simulated max strength is only half of the value reported. What could be the problem?

The input and data files are attached below for your reference. Thanks a lot!

dimension 3
boundary p p p
units metal
atom_style atomic
read_data data.vmd50
pair_style airebo 3.0 1 1
pair_coeff * * /opt/lammps-2May09/potentials/CH.airebo C

Dear Yanbiao,
Are you using the current LAMMPS version? I notice that your lammps
directory for your CH.airebo potential file lists the version as 2May09.
Older versions of AIREBO had some bugs, which were fixed for recent