This is another case of “The stuff you have to figure out that articles don’t explicitly tell you about.”[1]
There are issues to have with LigParGen but as far as I am concerned, most of my problems were located between my keyboard and my chair when using it. Since it does not provide forcefield settings for LAMMPS, you have to figure them out yourself which is error prone.
Please have a look at this discussion. This should help you understand what is missing in your input file. I’ve no problem running a condensed phase with your molecule data file and the simulation doesn’t crash with proper settings of the forcefield. Let me advise you to double check your previous simulations with other molecules that supposedly “run fine”.
As a personal note I also tend to fix H covalent bonds with fix shake with timestep of 1 fs and higher.
As a side note, I am very skeptical with articles stating “We used forcefield XXX” without providing parameters of forcefield XXX in SI but a single ref in text because, sometimes, they actually didn’t use parameters of forcefield XXX or changed some stuff and only tell you if asked. ↩︎