Hi, Lammps user/developer
I am trying to perform MD simulation for a TeO2 glass and for the same
I am using core-shell model. I have some queries related to my
simulation.
1. While optimising the interatomic potential (Buckingham) using GULP
for Te and O core-shell, I made the shells massless. Now, it seems
that I cannot put masses to be zero in lammps structure file. So I
have given a mass of 0.001 a.u.. How, strongly is that going to affect
my dynamics.
2. A common way to perform MD simulation for glass is to create a
random initial configuration. But doing so results in poor dynamics
and lammps keeps on giving "certain atom or bonds are missing error".
I have tried reducing the timestep and check every bin option. But
still couldn't figure it out.
3. In the output file, I am getting very large negative potential
energy in the second step. What could be the possible reason?
Attached file:
Input file = in.TeO2
log file = log.lammps
coordinate file = structure.dat
potential file = pot.mod
in.TeO2 (2.09 KB)
structure.dat (31.7 KB)
log.lammps (614 Bytes)
pot.mod (554 Bytes)
Hi, Lammps user/developer
I am trying to perform MD simulation for a TeO2 glass and for the same
I am using core-shell model. I have some queries related to my
simulation.
- While optimising the interatomic potential (Buckingham) using GULP
for Te and O core-shell, I made the shells massless. Now, it seems
that I cannot put masses to be zero in lammps structure file. So I
have given a mass of 0.001 a.u… How, strongly is that going to affect
my dynamics.
why don’t you just make a test with different settings. What are acceptable choices should be discussed in the relevant publications describing the derivation of the core-shell method or its applications.
- A common way to perform MD simulation for glass is to create a
random initial configuration. But doing so results in poor dynamics
and lammps keeps on giving “certain atom or bonds are missing error”.
I have tried reducing the timestep and check every bin option. But
still couldn’t figure it out.
how to “unoverlap” a system has been discussed many times on this mailing list.
the best way to start would be using the “soft” pair style, which is finite at r=0 and its potential height can be adjusted (even during a run). see the LAMMPS manual for examples (and the “example/micelle” folder, if I remember correctly).
- In the output file, I am getting very large negative potential
energy in the second step. What could be the possible reason?
atoms of different charge getting too close. doing dynamics with very light particles requires a very short timestep, as those will otherwise be hurled around like with a sling.
axel.
how to “unoverlap” a system has been discussed many times on this mailing list.
the best way to start would be using the “soft” pair style, which is finite at r=0 and its potential height can be adjusted (even during a run). see the LAMMPS manual for examples (and the “example/micelle” folder, if I remember correctly).
Comment : As far as I can see in VMD there are no overlapping atomic coordinates other than the shells and cores. Also, in an attempt to resolve this issue I did a ABINTIO MD run for 5ps at 300K followed by 5ps st 600K using CP2K code. So, I am sure that there isn’t any overlap.
Raghvender