[gmx-users] Freezing groups and energy conservation

[Francesco Mambretti]

Dear all, I read a discussion about that but there are no news since then,
so I would like to know if there have been recent improvements.

http://article.gmane.org/gmane.science.biology.gromacs.user/63051/match=freeze+groups+energy+conservation

The problem is: in Gromacs (also in my 5.0.2 version) freezing groups and
obtaining energy conservation (even with a run explicitly in NVE,
cutoff-scheme =verlet and verlet-buffer-tolerance = 10^(-8) ) seem to be
incompatible. Is that true?

I simply want to freeze an FCC-like core, made of 13 Argon atoms interacting
via LJ, in a liquid supercooled (liquid arrangement of the 256 atoms of my
system but velocities rescaled to a temperature typical of a solid
configuration) system. The idea is that this nucleus (that should therefore
mantain its shape) should guide the system to crystallize.


Apart from using position restrains, other alternatives to do that?
In fact also position restraints cause to me a negative energy drift, in
this case.

Thanks

Francesco Mambretti, LCP, University of Milan