Distributing atoms

Dear lammps user and Merry Christmas

I have a problem with creating a binary mixture of fluid droplet. I have made a droplet by packmol to creat spherical form of the droplet. But when I apply fix nvt to the system the droplet collapses. All atoms are distributed to the vacant space in simulation box. I changed all the parameters which I think might be effective such as cut-off radius (20A), different initial atoms coordinates, long coulomb interaction(PPPM method) and truncated coulomb, increasing temperature NVT ensemble, the number of molecules from 2000 to 5000. I don’t know exactly what is the problem. In the following you can see my script as well:

units real
atom_style full
boundary p p p
special_bonds lj/coul 0.0 0.0 0.0
pair_style lj/cut/coul/long 15
bond_style harmonic
angle_style harmonic
kspace_style pppm 0.000001
read_data merging.data

pair_coeff 1 1 0.1553 3.166
pair_coeff 1 2 0.0 1.583
pair_coeff 1 3 0.16248 3.118
pair_coeff 1 4 0.0 1.583
pair_coeff 1 5 0.1793 3.4705

pair_coeff 2 2 0.0 0.0
pair_coeff 2 3 0.0 1.535
pair_coeff 2 4 0.0 0.0
pair_coeff 2 5 0.0 1.8875

pair_coeff 3 3 0.17 3.07
pair_coeff 3 4 0.0 1.535
pair_coeff 3 5 0.18759 3.4225

pair_coeff 4 4 0.0 0.0
pair_coeff 4 5 0.0 1.8875

pair_coeff 5 5 0.207 3.775

bond_coeff 1 1000 1.0
bond_coeff 2 386.0 1.425
bond_coeff 3 553.0 0.945

angle_coeff 1 100 109.47
angle_coeff 2 55.0 108.5

group spce type 1 2
group methanol type 3 4 5

group fluidbox union spce methanol

neighbor 3.0 bin
neigh_modify every 2 delay 0 every 1 check yes

fix fixnvtf all nvt temp 280 298.13 100.0
minimize 1.0e-4 1.0e-6 100000 400000
fix fixshake fluidbox shake 0.00001 20 10 b 1 2 3 a 1 2

timestep 1
restart 50000 R.restart
run 250000

thank you in advance for your time and your responses.

Sincerely,

Reza

have you tested your parameters for each component of your mixture
separately, preferably as bulk and as droplet system separately?
can you reproduce published properties of these materials?

only if this is successful, you should look into mixtures. your input
hints at simulating a water/methanol mixture. this is a well
understood and well studied system, so you should check the published
literature for results and studies that you can reproduce.

i would not recommend to even start your system of interest until you
have validated all settings and parameters. what you are describing
you did seems to be doing the procedure backwards. the problem here
is, that it is near impossible to identify problems, because there are
so many opportunities, where things can go wrong.

axel.