[lammps-users] Allowing for material expansion?

My system is as follows: My unit cell is a slice of a
nanowire embedded in a surrounding matrix, and the
natural lattice constant of the wire is greater than
that of the matrix. The initial state of the system
would force the lattice constant of the nanowire to be
that of the surrounding matrix, and then as the system
equilibrates, the wire would expand and induce
stresses on the matrix. The catch is that I need to
allow for some expansion on the part of the matrix as
well, so an NVE condition isn't going to be
sufficient. Note that all the materials here are
supposed to be solid semiconductors.

I was thinking of using an NPT or NPH ensemble here,
but I'm not sure if it would be appropriate to assume
atmospheric pressure or zero pressure, or even if
there might be something I can do in LAMMPS that
didn't rely on assuming a pressure. What would be a
sensible way to go here?

The real issue is whether your system is periodic
or not, i.e. an array of wires embedded in a
matrix, or a single isolated wire. If its
periodic, NPT or NPH will work fine and expand
the box as necessary. If not periodic, then
use non-periodic BC and NVE will work fine
and the system will expand on the edges.


The system is periodic.

What I ended up doing is increasing the size of the
simulation box a little bit, relaxing the atoms,
increasing the size of the simulation box a little bit
again, relaxing the atoms again, and so on, until the
potential energy starts to increase. The size of the
box where the potential energy is minimum is the size
that I choose. That seems to work decently, and I
don't have to worry about assuming a particular value
for the pressure.