[ase-users] Langevin thermostat friction parameter

[Riccardo Petraglia]

Hello everybody,
this is the first time I am writing to this mailing-list.
I am using ASE to run dynamics with an "home-made dispersion corrected"
version of DFTB+ and Gaussian. To compare the values I obtain with other,
more reliable, methods, I have to use the Langevin thermostat (I am used to
the Nose-Hoover one). Now... I would like to compare my trajectories with
some trajectories at DFT level obtained with therachem. The guy who did
this trajectory, in order to have a fast thermalization, used a friction
value of 50fs (time not time^(-1)). I did some calculation and I obtained
that the same value in a.u. (to be used in ASE) is 0.0005 (time^(-1)). From
my point of view, the manual is not precise about this point: have the
friction to be specified as a time^(-1) or as a time? The value has some
relation with the ts or they are completely independent (from the
implementation point of view)?

I am asking these questions because with the friction value I found, the
temperature std-deviation is much higher then the one obtained with
therachem and I am wondering if this is related to a wrong friction value
or to the fact that the methods used are different.

Thank you for your time,
Riccardo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20140520/adcc968c/attachment.html>