Water box with a sheet of graphene

GROMACS version: 2020.1-Ubuntu-2020.1-1
GROMACS modification: No

Hello everybody.
I am trying to create a box with a sheet of graphene solvated with H2O, with a size of 17.09 x 25 x 27.13 (angstrom). The system preparation parameters are:

rlis = rcoulomb = rvdw = 0.6

System optimization normally occurs with grompp and mdrun, but the problem is when I use the NPT ensemble. NPT.mdp details:

dt = 0.002
nsteps = 10000
rlis = rcoulomb = rvdw = 0.6
Tcoupl = nose-hoover
Pcoupl = Parrinello-Rahman
pcoupltype = anisotropic
tau_p = 5.0
compressibility = 0 4.5e-5 0 0 0 0 which would be at 1 atm
ref_p = 0 1 0 0 0 0
periodic_molecules = yes

When actually starting the simulation, the error that is reported is the following:

starting mdrun ‘Grephene in Water’
10000 steps, 20.0 feet.
*** Process received signal ***
Signal: Segmentation fault (11)
Signal code: Address not mapped (1)
Failing at address: 0x55b4ce7797e0
[ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x14420)[0x7f6511d54420]

*** End of error message ***
Segmentation fault (core image recorded)

If I lower the dt to dt = 0.0001:
Fatal error:
One of the box vectors has become shorter than twice the cut-off length or box_yy-|box_zy| or box_zz has become smaller than the cut-off.

Would anyone have an idea how to resolve this issue?
Should I change the barostat to Berendsen?

HI,

By ‘optimization’ you mean energy-minimization? Did you get any warning there?
You could try to first run NVT (so constant volume) with dt=0.001 for a few hundred steps, and then switch to NPT (possibly already with dt=0.002). This usually worked out for me.
You could also try to switch thermostat to Berendsen (which I understand to be more stable in case of systems that are initialized far from equilibrium), or C-rescale (was it already implemented in Gromacs 2020? Maybe you want to upgrade to a more recent release.).

The problem was on my computer that couldn’t generate the files and had a tracking error. I tried on another computer and it worked. But thanks for the help.