Dear LAMMPS developer & users,
I’m conducting a GCMC simulation involving water molecules (SPC/E model with all its bonds being constrained by SHAKE method). However, I find the SHAKE fails as it’s used with the fix_gcmc command.
The water molecule file (molecule.H2O), input script (in.gcmc_water_vapor) and the log file (log.lammps) are attached above. I simulated a cubic box filled with water vapor. Initially, a single water molecule is placed in the box. The run consists of two stages: NVE and GCMC. In the NVE stage, the water molecule looks normal and the SHAKE works well (indicated by the tiny E_mol energy). However, in the GCMC stage, I hope to fill the cubic box with water molecules at 400 K and 1 bar. The command used is “fix 2 Vapor gcmc 10 10 20 0 9248 400.0 0 1.0 mol H2O shake 1 pressure 1.0 tfac_insert 2 full_energy”. However, the SHAKE fails occasionally. In the thermal output, E_mol may suddenly increases (such as timestep 11500) in the log file. I dump the molecular position and find a broken molecule at timestep 11500 as below. The H-O bonds and the angle bond are broken. I’m not sure whether it’s a bug or a mistake in my input script. (PS: I’m using the latest version of LAMMPS-Mar2017).
Any suggestion or comment is appreciated.
ITEM: TIMESTEP
11500
ITEM: ATOMS id mol type x y z vx vy vz fx fy fz
1783 595 1 11.9581 92.6583 99.3903 9.47262 -0.634907 -2.31436 46.4906 -53.2122 -172.189
1784 595 2 12.0246 91.426 98.7354 9.47262 -0.634907 -2.31436 -23.2863 190.711 144.204
1785 595 2 11.8447 92.6759 99.7708 9.47262 -0.634907 -2.31436 -23.2043 -137.499 27.9852
Regards,
----Tengfei LIANG
molecule.H2O (449 Bytes)
in.gcmc_water_vapor (1.9 KB)
log.lammps (71.3 KB)