Dear lammps users
I want to simulate an amorphous SiO nanoparticle. For this purpose I need to switch off the periodic boundary conditions. When I use non periodic boundary conditions f f f I run into lost atoms problems since the pressure in the system is rather high beta-cristobalite for instance exists at temperature 1600 K and pressure 5000 bar. May I use s s s? It works but what exactly happens using this boundary condition? I saw in my simulations that the box size is not kept exactly constant using sss. It is not really clear to me from the manual.
Thank you in advance.
Regards
Sabine
Dear lammps users
dear sabine,
I want to simulate an amorphous SiO nanoparticle. For this purpose I need to
switch off the periodic boundary conditions. When I use non periodic
boundary conditions f f f I run into lost atoms problems since the pressure
in the system is rather high beta-cristobalite for instance exists at
temperature 1600 K and pressure 5000 bar. May I use s s s? It works but what
exactly happens using this boundary condition? I saw in my simulations that
with shrinkwrap boundary conditions, if atoms move out of
the current box, the box gets extended.
the box size is not kept exactly constant using sss. It is not really clear
to me from the manual.
that is what is supposed to happen.
obviously your system is evaporating atoms.
you have now two choices.
1) keep using periodic boundaries, but have a large enough
vacuum area between your nanoparticle and its periodic images.
2) use fixed boundaries, but add reflecting or soft (lj) walls.
which one would be the better choice depends on a number
of details, e.g. the potential that you are using, the systems size
and what properties you want to study.
cheers,
axel.
Thinking about the physics of this problem, it sounds like you may need to use a pressurizing medium, like in a diamond anvil cell, since a beta-cristobalite nanoparticle at zero pressure and 1600 K will either evaporate, or rearrange its structure, or both. Consider adding in a whole load of LJ particles using pair_style hybrid, periodic boundaries, and fix npt T=1600K P=5000 bar. There is a nice paper on CdSe nanoparticles where this was done (Grunwald, Rabani, Dellago, PRL 96 255701 (2006).