Dear all,
I am using NVE ensemble to do a simple relaxation calculation. As we know that in this ensemble, we should keep the Number of atoms, and Volume, and Energy constant.
But if some atoms fly out of the box, while we do not use the pbc, which means the ‘N’ is changed. So, in this case whether NVE still applies?
If so, it seems contrary to use NVE …
Thanks for your clarification.
Nothing breaks in NVE integration if atoms disappear. E.g.
it could be used for a system where atoms evaporate from
a surface. But the remaining atoms might slowly cool,
but that's physical.
Steve
I am using NVE ensemble to do a simple relaxation calculation. As we
know that in this ensemble, we should keep the Number of atoms, and
Volume, and Energy constant.
I think it's the other way around: if you choose dynamical rules that
conserve N, V, and E, and that sample the resulting available phase space
ergodically and reversibly, then you generate an ensemble of
configurations that is known as the NVE ensemble (in the limit of infinite
trajectory length).
But if some atoms fly out of the box, while we do not use the pbc, which
means the 'N' is changed. So, in this case whether NVE still applies?
If so, it seems contrary to use NVE ...
You don't really choose an ensemble, you choose dynamical rules. If your
rules are such that particles can disappear, then you don't generate the
NVE ensemble. Well-defined ensembles in which the number of particles can
fluctuate are based on a constant value of the chemical potential, which
AFAIK lammps does not support.
In your specific case, it depends on what your boundary conditions are, and
where the particles actually go when they "fly out of the box".
Hope this helps,
Lutz