hBN Thermal Conductivity

LAMMPS LAMMPS Beginners
1

Dear Users,
I am trying to find thermal conductivity of hBN zigzag nano ribbon (10 nm X 3 nm) at 300K. First I minimized system energy with steepest descent method, then applied fix nvt and finally fix nve command. I used BNC.tersoff potential from “Examples” directory of LAMMPS.
However, in the nvt stage, pressure became negative. As i checked the dump file, most of the bonds of the hBNNR were broken. Would appreciate any suggestion regarding the possible source of error.

My data file and input script are attached below.
BN_data2.txt (29.0 KB)
input.lammps (4.3 KB)

2
  1. your system is not that of a ribbon but that of a sheet due to periodic boundaries in x and y. to make it a ribbon, you would need to set also the y dimension to fixed or shrinkwrap
  2. your box dimension in z is excessive, which may interfere with efficient MPI parallelization +/- 5 should do.
  3. a negative pressure indicates that your system “wants” to shrink. you can check this by adding fix box/relax and relax your system in x direction during minimization. if the box changes a lot, you have to re-think and re-check your initial geometry. perhaps there was a typo or some miscalculations somewhere. if not, then perhaps the potential is not suitable for this specific system.
  4. why use steepest descend and not conjugate gradient?
  5. if you start MD at finite temperature of a very strained system, all hell can break loose.
1 Like
3

  1. Initially, it was from -50 to 50 and the system still became chaotic. As pressure was very high, I tweaked it to check if that was the source of problem.

  2. I used steepest descent method as it was mentioned in the paper I was trying to regenerate. However, I also applied CG algorithm and the result was same.`

I will try fix box/relax and re-check my initial geometry.
Thank you very much for your suggestions.