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

Hello,

I am facing a similar problem, but happens in 5-10ps of the simulation. The difference is its not CG, water is on top of a TiO2 crystal. (Details below). Will nph prevent such layering in this case and is the same kind of “flying ice cube syndrome” ?
Thanks.
Ritwik.

Init

units real
dimension 3
newton on

processors px,py,pz (let lammps choose)

boundary p p p
atom_style full
kspace_style ewald 1.0e-6
kspace_modify mesh 32 32 32 order 8
read_data data.386-reconst
group wat type 3 4
angle_style harmonic
angle_coeff 1 37.5 113.47
bond_style harmonic
bond_coeff 1 529.583 1.012
pair_style hybrid buck/coul/long 8.0 lj/cut/coul/long 8.0
pair_coeff 1 1 buck/coul/long 717647.4 0.154 121.0676
pair_coeff 2 2 buck/coul/long 271716.3 0.234 696.8883
pair_coeff 1 2 buck/coul/long 391049.1 0.194 290.3317
pair_coeff 1 3 buck/coul/long 28593.02 0.265 148.000
pair_coeff 2 3 buck/coul/long 271716.3 0.234 696.8883
pair_coeff 2 4 buck/coul/long 7088.0 0.2563 51.975
pair_coeff 3 3 lj/cut/coul/long 0.1553 3.166
pair_coeff 1 4 lj/cut/coul/long 0.00144 3.7475
pair_coeff 4 4 lj/cut/coul/long 0.00 0.0
dump 1 all atom 1 minimize.dump
minimize 1.0e-5 1.0e-7 55555 10000
undump 1
timestep 1.0
dump 1 all atom 500 atom.dump
fix 1 wat nvt 300.0 300.0 100.0
fix 2 all msd 500 msd.dat
fix 3 all rdf 500 rdf.dat 100 1 1 1 2 1 3 1 4 2 2 2 3 2 4 3 3 4 4
thermo_style multi
thermo 100
run 500000

Hello,

hello ritwik,

I am facing a similar problem, but happens in 5-10ps of the

no, you don't. you only integrate the EOM for the water molecules,
thus it is impossible to have to whole system do a collective
translational motion.

simulation. The difference is its not CG, water is on top of a TiO2
crystal. (Details below). Will nph prevent such layering in this case

please have a closer look at steve's answer. it is not the nph
integrator, but rather the langevin thermostat that makes the
difference.

and is the same kind of "flying ice cube syndrome" ?

in any case, it _cannot_ be the same. it is difficult to
make any statements beyond this without seeing more details,
but there are a few things in your input that don't make
much sense, to me:

how can you run a stable MD on water with a 1fs time step,
if you don't shake the water? why do you dump the tio2 crystal
coordinates, if they don't move (waste of disk)? why do you
compute the MSD for the whole system, if only a part of the
atoms move?

...and the final question: why do you think it is wrong
if the water molecules attach strongly to your surface?
with a simulation you generally get what out what you put
in as parameters. if you get a frozen out water layer,
so it could just be a property of the potentials yuo are
using.

cheers,
    axel.