[lammps-users] Why did my liquid water become solid

Hi everyone,

I am doing a simple simulation with only 400 CG beads, one bead represent 4 water atoms, as mentioned inMarrink’s work at J. Phys. Chem. B, Vol. 108, No. 2, 2004, by using LAMMPS.

But after several thousands time steps, beads formed kind of lattice shape as in solid material. Dose anyone can help me with this?

The attachments are my data file and output - xyz file, which can be visible through VMD.
The following is my input file:

units real
atom_style full
boundary p p p

pair_style lj/cut 12.0

read_data CG_real_Water_5-0.dat

print “read finished”

pair_coeff * * 1.195 4.7

neighbor 0.1 bin

neigh_modify delay 5 page 10000000

thermo_style custom step temp press vol
thermo 1000

velocity all create 300.0 4928459 rot yes dist gaussian

fix 2 all npt 300.0 300.0 1000.0 xyz 1.0 1.0 1000.0

timestep 1
run 5000

compute mo_temp water temp

fix tem water ave/time 1000 2 5000 c_mo_temp file temp_real_Water_5-0_npt.profile

dump a all xyz 50000 CG_real_Water_5-0_npt.xyz
print “start running”

timestep 5

run 5000000

CG_real_Water_5-0_npt.xyz (117 KB)

CG_real_Water_5-0.dat (19.3 KB)

Hi everyone,

I am doing a simple simulation with only 400 CG beads, one bead
represent 4 water atoms, as mentioned inMarrink's work at J. Phys.
Chem. B, Vol. 108, No. 2, 2004, by using LAMMPS.

But after several thousands time steps, beads formed kind of lattice
shape as in solid material. Dose anyone can help me with this?

you are most likely a victim of the "flying ice cube syndrom",
which can usually be dealt with be (infrequently) removing the
center of mass momentum.

but if a recall correctly, also the marrink water potential has
a tendency to freeze if you use a comparatively short time step
like you do. the proponents of that model claim that the length
of the time step is part of the parametrization and claim that
you have to use larger time steps, but i personally find this
a bit strange.

i have to say at this point that our group is also working on
coarse grain potentials (including water) and that thus i am
of course biased towards our model. feel free to check out the
CG-CMM user package in LAMMPS and the related publications.

cheers,
    axel.

If Axel is correct (flying ice cube) then try fix nph
with a fix langevin for thermostatting. The Langevin
thermostat often does not suffer from the flying ice cube effect.

Steve