make hydrogen move

dear all lammps user.
I want to simulate adsorption of hydrogen storage in CNT.
It already ran, but I think there’s some mistakes that I find yet.
I want to make the hydrogen move, so it can be catched by the CNT. but in simulation with VMD, hydrogen atoms have attraction each other.
I have already assign the initial velocity of hydrogen.

Can somebody help me??

best regards,

mardi

here’s the input script :

3-d LJ hydrogen storage

dimension 3
boundary p p p

atom_style full
neighbor 0.3 bin
neigh_modify delay 5

read_data lmp_rev.dat

#kspace_style ewald 1.0e-2
bond_style harmonic
bond_coeff 1 1201.1 0.254

LJ potentials

pair_style lj/cut/coul/cut 3.0
pair_coeff 1 1 1.0 1.0
pair_coeff 2 2 1.737 1.166
pair_coeff 1 2 1.858 1.083

Define groups

group cnt type 2
group hyd type 1

Initial velocities

compute mobile hyd temp
velocity hyd create 1.0 482748 temp mobile
fix 1 all nve
fix 2 hyd temp/rescale 100 0.26316 0.26316 0.02 1.0
fix_modify 2 temp mobile

Static nanotube

velocity cnt set 0.0 0.0 0.0 units box
fix 3 cnt setforce 0.0 0.0 0.0
fix 4 cnt aveforce 0.0 0.0 0.0

Run

timestep 0.0005
thermo 10000
thermo_style custom step atoms temp pe ke evdwl
thermo_modify temp mobile

dump mydump all atom 10000 dump.asem_kecut
run 100000

You're using LJ units with an incredibly large
bond strength and a very small timestep.
What makes you think LJ coeffs of sigma,eps = 1.0
is a good model for hydrogen. With a cutoff of 3 sigma,
they will attract each other. Is this meant
to be a physical model for real materials
or an idealized system?

Steve

dear Steve,
its for real material.
what is a good LJ coeffs and cutoff in this simulation?

Dear lammps users

I’m using a Stillinger Weber potential in a simulation with various types of atoms, I wanted to know if the “neigh_modify exclude” command will correctly turn off the interactions between two types of atoms, because in the manual it says that it turn of pairwise interactions, but SW potential has olso the three body interaction term

Thanks in advance

Matias

dear Steve,
its for real material.
what is a good LJ coeffs and cutoff in this simulation?

do i understand you right?

you are setting up a simulation and you have
not done any research about what parameters
represent your system well?

what did motivate your choice of parameters then?

axel.

Dear lammps users

I'm using a Stillinger Weber potential in a simulation with various types of
atoms, I wanted to know if the "neigh_modify exclude" command will correctly
turn off the interactions between two types of atoms, because in the manual
it says that it turn of pairwise interactions, but SW potential has olso the
three body interaction term

it should do this, because the three body term
is computed by looping over the neighbors of the
second atom. so if the first pair interaction or
the second pair is excluded, nothing is computed.

i would still recommend to set up a few simplistic
test systems and verify this. this is very easy
to do and you can be a hundred percent certain
after doing this kind of test.

cheers,
    axel.

As always, I suggest you find a paper where
someone has modeled the system you are interested
in, learn how they did it, and start from there
before changing the model.

Steve

The definition of an "interaction" between 2 atoms
is fuzzy for many-body potentials.

Axel is correct that SW (and all the many-body potentials)
use pairwise neighbor lists to find interactions.
And the exclude settings will drop those pairwise
interactions from the list. However the 3-body term
in the SW case is actually a double loop over the
neighbors of atom I (not a loop over neighbors of J).

You'll have to decide if the
missing pairwise neighbors leave you with what you
desire in that case. In some many-body potentials
you could drop the A/C interaction from the pairwise
list but the 2 atoms would still "interact" thru B because
A/B and B/C interactions would still exist.

Steve