[Wien] Ionic versus covalent systems

[Rocquefelte]

Dear Subin Adhikari,

Just a general idea: more the system is ionic more the dispersion of the 
band structure will be flat.
In some cases the flat character of the band structure may lead to 
mixing problems due to charge oscillations from one cycle to another.
Usually one solution when you are facing such problem of convergence is 
to modify the mixing (smaller) or to modify the integration scheme in 
lapw2 (from tetra to GAUSS with a broadening that will reduce the 
oscillations).

Regards

Xavier


jadhikari at clarku.edu a écrit :
> Dear Wien2k users,
>
> I did calculations of compounds containing "Calcium ions" and it was
> extremely tough to get the converged results. The calculations of similar
> compounds, replacing Calcium with "Lead ions" was extremely easy and of
> course superbly fast. All that was done was "Ca" replacex with "Pb".
>
> So the questions ( basic) are-
> Is it the former being more ionic we are having problems and that the
> latter being more covalent all of the troubles are gone?
>
> Is is because of localized versus delocalized systems?
>
> Following compounds- LiZnN, LiMgN, LiMn2O4, NaNbO3, MgAl2O4, CaCu3Ti4O12
> possess ionic characteristics and NaNbO3 is an example of a tougher system
> for convergence. 9 IBZ with RKmax of 6 or greater is extremly time
> consuming. The published article is silent about the RKmax and the test
> for higher IBZ points.
>
> Is it because of "Na" ion in the system or it does not matter?
>
> Looking forward for just a general idea.
>
> Thanking you
>
> Subin Adhikari
>
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
>