Dear LAMMPS users,
I’m trying to simulate halogen bonding in organic molecules using a pseudo atom (PA) on halogen atoms.
The simulation takes place in the npt ensemble at 100K and I keep getting the following error after a few timesteps:
Bond atoms %d %d missing on proc %d at step %ld.
I guess the problem comes from the PA because everything is fine when I take it off.
After inspection of the log file, the problem seems to arise from the thermostat: the temperature is clearly more than 100K when the PA is present.
I switched from npt to nve with a Berendsen thermostat, but this did not solve the problem, neither did the tuning of the PA bonding or non-bonding parameters.
Finally, I took a look at the trajectories using VMD and the problem seems to be that, despite the quite stiff bonding parameters of the PA particle, some of them move in a non-physical way, eventually leading to the explosion of a nearby bond on the same molecule or on a neighbor molecule; even at temperature lower than 100K.
Any idea on what could cause this?
These errors occur when the simulation is run on 1 or several processors.
The LAMMPS version is the 28 Jul 2015 version.
Thanks a lot,
Clement.
inputFile (957 Bytes)
log_npt (12 KB)
log_without_PA (28.8 KB)