[gmx-users] Temperature of the system when using stochastic dynamics in vacuum

[Berk Hess]

________________________________
> Date: Tue, 12 Feb 2008 23:31:14 +0800
> From: goushenzhiyuan at gmail.com
> To: gmx-users at gromacs.org
> Subject: [gmx-users] Temperature of the system when using stochastic dynamics in vacuum
> 
> 
> Hi all,
> 
> Recently, I run a carbon nanotube in vacuum using stochastic dynamics (Gromacs 3.3.1, integrator=sd) at 300K (ref_t=300.0) for 800 ps. The number of atoms of the CNT is about 5000, and the tube is left free to vibrate except for a few atoms are position restrained. When I set the tau_t=100, 0.2, 0.05, the actual simulation temperature fluctuates around an average value ~72K, 283K, 299K respectively.
> 
> I also run a peptide with about 110 atoms in vacuum using stochastic dynamics at 300K for 1 nanosecond. When I set the tau_t=100, 0.1, the actual simulation temperature fluctuates around an average value ~147K, 302K respectively.
> 
> So I want to know: Why different tau_t (with the same ref_t=300.0) can result in different temperature when the system is in vacuum? Can I perform the stochastic dynamic simulation in vacuum? If I can, can I used the tau_t=0.05, 0.1 for the CNT and peptide mentioned above respectively to couple the temperature of the system to 300K?
> 
>  Thanks,
> 
>  Xiu Peng.

>From the number you have posted I guess that you have very short cut-off's
which introduce a lot of integration error. Therefore energy will be removed
from your system. Stochastic dynamics will put energy into your system again,
but the rate at which it does this depends on tau_t.
Normally one should be able to run with tau_t on the order of 1 ps.
I don't know what your system setup is, but I guess you should increase
your cut-off distance(s) and/or improve your electrostatics metod,
if you have electrostatics.

Berk.

_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today it's FREE!
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/