LAMMPS users, I am able to reproduce the density as reported in 2003 ReaxFF paper for quartz, coesite, and cristobalite. Yet, for stishovite, I am not able to do so. Initially, I use NPT to equilibrate at 1 atm and 300 K, like I did with other polymorphs.
But with stishovite, the temperature shoots up to ~2,000 K before going back to 300 K. However, the density drops to about ~3.2 g/cc, much lower than what Dr. van Duin reported of ~4.29 g/cc.
I know my structure is stable because when I use other potentials (Tersoff, BKS, COMB10), the structure is stable, and by that, I mean the temperature does not shoot up, and the density calculated is reasonable compared to exp.
I am using a timestep of 0.2 fs upon the suggesting of Dr. van Duin, and even when changing this to 0.1 fs, or even 0.05 fs, I get the same results.
I am using real units.
My npt ensemble looks like:
fix 1 all npt temp 300 300 20 x 1 1 200 y 1 1 200 z 1 1 200
Does anyone know what could be the culprit behind the hike in temp from 300 K to 2000 K, before it drops slowly back down to 300 K?
Note I am also modeling ~768 atoms, but no change when modeling ~3,000 atoms.
Thanks,
Ben