I want to perform a MD simulation of a molecular crystal that have phase ch=
ange from triclinic cell to monoclinic cell. While there is ongoing structu=
ral phase transition, there is volume change, however the no. of atom in un=
it cell also change. Like example from 42 atoms to 34 atoms. But I think I =
cant use npt ensemble because it fix mole hence no. of atom. So is there an=
yway for me to make the atom missing? |
I want to perform a MD simulation of a molecular crystal that have phase ch=
ange from triclinic cell to monoclinic cell. While there is ongoing structu=
this is quite a tricky thing. are you sure that the potentials that
you are planning to use, would actually be capable to correctly
or even approximately represent that?
ral phase transition, there is volume change, however the no. of atom in un=
it cell also change. Like example from 42 atoms to 34 atoms. But I think I =
cant use npt ensemble because it fix mole hence no. of atom. So is there an=
yway for me to make the atom missing?
this is not at all a problem when using a supercell. the minimum you'd need
would contain 714 atoms. you probably have to choose a (much) larger
one to overcome finite size effects.
your question seems to indicate, that you may not have very much
experience with MD. in that case, i would strongly suggest to start
with an easier project. in any case, a thorough search of the literature
on modeling phase transition is highly recommended.
cheers,
axel.
ral phase transition, there is volume change, however the no. of atom in un=
it cell also change. Like example from 42 atoms to 34 atoms. But I think I =
cant use npt ensemble because it fix mole hence no. of atom. So is there an=
yway for me to make the atom missing?this is not at all a problem when using a supercell. the minimum you'd need
would contain 714 atoms. you probably have to choose a (much) larger
one to overcome finite size effects.
Just to amplify Axel's point - once you have built a large system, LAMMPS
has no concept of a unit cell. It simply models the entire collection
of N atoms and they can form whatever sub-structures or sub-periodicty
they wish. However, if you are doing this in an overall periodic
box, then the large box does impose constraints on the final structure
due to periodic boundary conditions. Also LAMMPS is not setup to
do grand canonical MD (adding/deleting atoms on the fly) for solid (dense)
systems.
Steve