[Wien] RMT choice and cell size

[Gerhard H Fecher]

I often try to start the optimization with the lowest lattice parameter 
(volume)  that I like to have for the E(V) dependence. Then its easily 
checked in a first initialization if the rmts are too big.
However, sometimes one has to take care that small rmts may lead to a "close 
to ghostband" situation if you increase the lattice parameter, as the 
interstitial takes more and more volume one may need more and more plane 
waves. Just to say it in flapsy words.

For complex ferromagnetic materials you may also check whether the energy 
differences are caused by changes between ferri and ferro magnetic order of 
different atoms. It appears that the convergence is sometimes running into a 
"wrong" minimum. The "correct" dependence may then depend critically on your 
settings in the .inst file. Most oftenly I had to play arround with the 
initial settings to find satisfactory results.
I usually name it experimental theory.

Ciao
Gerhard

Am Mittwoch, 22. März 2006 13:55 schrieb L. D. Marks:
> On Wed, 22 Mar 2006, Chandrika wrote:
> 
> > Just another question : Is there a 'rule of thumb' about how much smaller 
> > than covalent/atomic radius the RMT should generally be.
> 
>  	I roughly follow what is in the UG, section 4.3, where it talks 
> about the RMT's (search for "atomic sphere radius"). It would be nice if 
> there was a good method to determine the best RMT's for a case beyond what 
> setrmt will do, but at present I do not believe that there is one.
> 
> > How do you say that 1.8 would be too big RMT for SiC?
> 
>  	It's a guess. You should check the logs, e.g. case.outputM during 
> the minimization to see if it's got touching RMTs.
> 
> -----------------------------------------------
> Laurence Marks
> Department of Materials Science and Engineering
> MSE Rm 2036 Cook Hall
> 2220 N Campus Drive
> Northwestern University
> Evanston, IL 60201, USA
> Tel: (847) 491-3996 Fax: (847) 491-7820
> email: L-marks at northwestern dot edu
> http://www.numis.northwestern.edu
> -----------------------------------------------
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>