[lammps-users] Airebo

Hi Axel and Yoel,

There are indeed a few bugs in the Lammps version of the AIREBO potential. Various members of the Harrison / Stuart groups are working on finding and fixing them. Heating/melting as you discovered causes problems. Using the original serial code, you should be able to do what you are trying to do, with no problems. We routinely use our version of AIREBO with the LJ and torsion off to create a-C films. Within Lammps, I tried all the suggestions Axel mentioned, nve/limit, etc. with no better results than you observe.

As Axel mentions, you probably will not end up with a good structure if you use NPT, unless you apply a very large pressure. Systems prepared this way will have very low density. My typical recipe is as follows:

  1. Set the desired final density by scaling the atomic and box coordinates to the appropriate values.
  2. Thermalize at 300K and Heat to 8000K in NVE. I like to actually melt the diamond.
  3. Equilibrate in NVE @ 8000K.
  4. Quench to 300K. Quench times vary but I usually use only a few picoseconds.
  5. Equilibrate at 300K in NPT to relieve any stresses and/or minimize energy, depending on your final objective.

A few other cravats:
2nd generation REBO generates films with ~90 sp2 content every time no matter what the starting conditions. If you want to make DLC with high sp3 content, you can use the first generation REBO potential (not in Lammps) or Tersoff (included in lammps but not parameterized for H).

I have a few papers out there discussing this method. Do a search for Schall, Harrison, and amorphous carbon.

Good luck, and look for a post in the future saying the airebo has been fixed.