Deform Box and perform NVT

LAMMPS LAMMPS General Discussion
1

Hi, after obtaining the polymer model from moltemplate, I used fix deform to reduce the size of the box from (300, 180, 180) to (180 180 180). Following that, I hope to do NVT.
The protocol is to create a well-mixed polymer system (aka like a melt) for further simulation.
The problem is that my rdf, after 5,000,000 steps, looks like the one I uploaded.

rdf-error
rdf-error1181×787 116 KB

units           real
atom_style      full
bond_style      harmonic

pair_style      lj/cut 14
read_data ${to_read_data}

bond_coeff     *    0.6   7
pair_coeff      * * 0.148 7 14

timestep        2.0

thermo_style    custom step temp pe etotal press vol density lx ly lz
thermo          5000 

fix             2 all deform 1 x final 0 ${lx}

run             50000
unfix           2

write_data  system_after_deform_lx-${lx}.data

##########
# NVT ensemble - 1

thermo_style    custom step temp pe etotal press vol density lx ly lz
thermo          5000  # time interval for printing out "thermo" data

dump            1 all custom 10000 nvt-1.dump id mol type x y z ix iy iz

fix             1 all nvt temp 300.0 300.0 $(100.0*dt)
restart         5000 nvt-1.restart

run             5000000
unfix           1

write_data      nvt-1.data

Any suggestions on how to proceed?

2

The problem, that you don’t get the results you expect is not automatically an error and specifically not a LAMMPS problem and thus more a topic for discussion with people familiar with your specific research than a public forum focused on the LAMMPS software. Thus you should discuss with your adviser, tutor(s), colleague(s), and collaborator(s).

From remote and without knowledge about (and vested interest in) your research, there is very little that can be said. You don’t provide any details on that, not even what shape of g(r) you would expect. How should anybody from remove know more than you about that?

The only thing of note seems to be that you are simulating a bead-spring polymer with only bonds (no angles and dihedrals) and non-bonded interactions. For that I would have expected that you would change the special_bonds settings, which by default excludes the non-bonded interactions between 1-2, 1-3, and 1-4 neighbors, which would with your force field settings allow 1-3 and 1-4 atoms in the same molecule come very close to each other or even overlap. A more common setting for bead-spring polymers is to only exclude 1-2 non-bonded interactions (or sometimes not even thouse). This is achieved with:

special_bonds lj/coul 0.0 1.0 1.0