[gmx-users] Re: Anomaly in Temperature Behavior

Andrey Frolov andrey.i.frolov at mail.ru
Thu Jan 24 08:27:26 CET 2013

S. Alireza B wrote
> My concern is that at the end of my 0.5 ns NVT run the T is ~300K;
> however,
> at the first step of my NVE, when no forces has exerted on unfrozen
> molecules yet, the T appears to be ~185K!

Your explanation of this issue given in the first post sounds very
meaningful. From general sense it is reasonable to consider the frozen
particles just as sources of external potential field acting on the rest of
the particles. So, one should not take into account frozen particles to
calculate temperature of the system, what, i think, gromacs does.

If you still have concerns, there is a way to figure it out of a simple
test. E.g. generate a simulation system with two point particles. Freeze one
particle, assign velocity to another one, evaluate temperature by mdrun,
then unfreeze the first particle, check if it has a zero velocity, evaluate
temperature by mdrun. If the temperature displayed by mdrun drops down in
the second evaluation, then you were correct.

S. Alireza B wrote
> The output configuration of my system shows that all of my solid phase has
> melted. So, no crystallization occurred.

Sorry, the output after 0.5 ps of simulation time in NVE shows complete

If it is true, that this illustrates that your crystal phase is really far
from equilibrium with respect to internal degrees of freedom (has high
potential energy). When you unfreeze this in your NVE simulation the
potential energy tends to decrease, and the only way for it to decrease is
to transform into kinetic energy (which increases), leading to increase of

View this message in context: http://gromacs.5086.n6.nabble.com/Anomaly-in-Temperature-Behavior-tp5004808p5004843.html
Sent from the GROMACS Users Forum mailing list archive at Nabble.com.

More information about the gromacs.org_gmx-users mailing list