Hello,
I am using ReaxFF to model iron-water interaction. I put 4 angstrom on the iron side as vacuum and I am using 8 cores for my simulation. Attached are input and geometry files. The problem is that the whole system moves to the vacuum side and then after 413800 steps I have “lost atom” error. Can I have your idea what I am doing wrong.
Thank you in advance,
Hossein,
in.Ferun (1.5 KB)
Geometry.dat (120 KB)
Hello,
I am using ReaxFF to model iron-water interaction. I put 4 angstrom on the
iron side as vacuum and I am using 8 cores for my simulation. Attached are
input and geometry files. The problem is that the whole system moves to the
vacuum side and then after 413800 steps I have "lost atom" error. Can I have
your idea what I am doing wrong.
you must not do time integration (via fix nvt) on all atoms.
set up a group, that does not include the bottom 3 or 4 layers of iron atoms
and do time integration only on those.
that will a) avoid reconstruction on the part representing the interface to bulk
b) restrict the motion of the full system.
there are discussions on more elaborate thermalization/time
integration schemes for such slab systems (e.g. using combinations of
fix nve and fix langevin) in the mailing list archives.
axel.