Dear lammps users,
I’m simulating a liquid-liquid phase transition of Si, and I want the system to be fully equilibrated at a given point during the phase transition. However, this is not easy for me, since the system is quite unstable, and easily crystallizes during and after the liquid-liquid phase transition. Do you have any good ideas to avoid the crystallization and fluctuation, and prolong the equilibration time for my liquid-liquid phase transition ?
Bests,
Bo
Just a quck reply.
There may be better sampling methods to use, and I don't know how
well this approach works near a phase transition, but you can try
using replica-exchange.
http://en.wikipedia.org/wiki/Parallel_tempering
LAMMPS implements this using the "temper" command.
http://lammps.sandia.gov/doc/temper.html
You will need a much higher number of temperatures (small temperature
spacing) near the phase transition. And (as always with
replica-exchange), you will need to check that the temperature swaps
are able cross over the phase transition barrier. (IE make sure that
the various replicas are able to mix their temperatures. This means
systems at low temperatures are eventually able to reach high
temperatures, and visa-versa.)
I hope this gets you started.
Andrew