PIGLET thermostat with DFTB+

[Nicolas Künzel]

Hello,

I read on the webpage for the GLE thermostats that you already included the possibility to work with the PIGLET thermostat in LAMMPS, CP2K, CPMD and FHI-AIMS.
How about DFTB+?
Since it is a package one usually uses instead of full DFT for calculations which have quite a strong computational effort. Therefore it would be considerably useful to also use the PIGLET thermostat since that reduces the computational costs a lot and would lead to the possibility of quantum calculations on normally not really accessible systems due to too much effort.

Cheers,

Nicolas

[Michele Ceriotti]

I guess you mean to ask if GLE methods are available in DFTB+ natively (as of course they are through i-PI). 

AFAIK they are not, but this is a question that is best asked to the DFTB+ developers. If I were them, I would not spend time implementing that, as one has access to all of these methods through i-PI, and the communication overhead is negligible given the cost of a TB force calculation.

Cheers

[Nicolas Künzel]

No I guess I just misunderstood the information on the website. Didn't realise it was meant that the thermostat was fully included in the MD codes. Thought it was about i-PI and the MD codes. But nice that I can use it with i-PI.

I have another question though:
Is the PIGLET thermostat just working when considering structural properties or could one also use it for calculations of e.g. free energy differences for proton transfers?

Also: the \omega_{max} and \omega_{min} on the webpage when generating parameters, are they physical frequencies or frequencies of the normal mode representation?
For the hw/kT term just the physical frequencies have to be considered, right?