Stone–Wales Defect in Two-Dimensional Heterostructure: Problem in Equilibration

LAMMPS LAMMPS General Discussion
1

I am trying to simulate a Silicene-hBN bilayer heterostructure with a Stone-Wales defect. After running a fix nvt, during fix nve, I found that the system temperature suddenly jumped to a higher value. I have tried running the NVT for a longer period (5 ns), considering that the defect will be equilibrated by then, but the problem persists.

LAMMPS Version 29 Aug 2024 - Update 3

read_data 30x10_sw_1%.lmp

-- Potentials
pair_style      hybrid sw tersoff lj/cut 10
pair_coeff * * tersoff BNC.tersoff B N NULL
pair_coeff * * sw silicene.sw NULL NULL Si
pair_coeff 1 3  lj/cut 0.011654 3.732   # B-Si
pair_coeff 2 3  lj/cut 0.007215 3.544   # N-Si

-- Settings
timestep        0.0001          # 0.1 fs
variable T      equal 300       # target temperature (K)
variable tdamp  equal 0.01


fix  1            all  momentum 10000 linear 0 0 1
fix  tempControl  all  nvt  temp ${T} ${T} ${tdamp}  drag 1
run  50000000
unfix tempControl
unfix 1


fix  EnergyControl  all  nve
run  10000000

image
image1232×865 225 KB

30x10_sw_1%.lmp (1.3 MB)

What can be done here to resolve this issue?

2

As you are running nve, I think the most likely reason would be abrupt shift of atom coordinates in a short time (probably a diffuse event of a defect?). Maybe you can try decreasing timestep or using a smooth variation of LJ, but sometimes the problem lies in the functional form of the force field, and cannot be easily solved.

Dumping out the trajectory may be helpful in understanding what is happening.

3

Since you are already using a timestep of 0.1fs, which is unusually small for the kinds of atoms you have (I would expect that you should be able to safely use 1fs, possibly even larger), I support the suspicion of @initialize that there could be something wrong with your choice of force field.