[Wien] Charge convergence

[Laurence Marks]

It is possible for :ENE (Energy, -ec) to seem to converge, but has not
-- so it is always wise IMO to also use the charge converge (:DIS,
-cc). A value of .0004130 is probably OK, I suggest using the "Check"
command which gives more information. (Peter added it to version 16,
presumably 17 as well as a command Check-mixing.)

>From the variational principle, the density converges more slowly than
the energy. Depending upon what you are trying to do your charge
convergence may be fine. For something like phonons it probably is
not. I will note that very good charge convergence often requires good
k-mesh, RKMAX etc, more than is needed for a routine calculation.

N.B., with more atoms charge convergence can be harder to achieve --
and less important.

On Tue, Jul 11, 2017 at 9:40 AM, Luis Ogando <lcodacal at gmail.com> wrote:
> Dear Prof. Marks and Wien2k community,
>
>    Good afternoon !
>    Just a curiosity about the charge convergence criteria. I am optimizing a
> supercell and the charge convergence limit I am using is 0.001. The same
> numerical value for the energy convergence.
>    The energy is finally converged, but the charge is not despite the fact
> that its change is below the limit in the last 10 iterations.
>
> CarbazoleLDAopt.dayfile::CHARGE convergence:  1 0.001 -.0002028
> CarbazoleLDAopt.dayfile::CHARGE convergence:  0 0.001 .0006007
> CarbazoleLDAopt.dayfile::CHARGE convergence:  1 0.001 -.0002590
> CarbazoleLDAopt.dayfile::CHARGE convergence:  0 0.001 .0007240
> CarbazoleLDAopt.dayfile::CHARGE convergence:  0 0.001 .0008640
> CarbazoleLDAopt.dayfile::CHARGE convergence:  0 0.001 .0005344
> CarbazoleLDAopt.dayfile::CHARGE convergence:  0 0.001 .0005217
> CarbazoleLDAopt.dayfile::CHARGE convergence:  0 0.001 .0002525
> CarbazoleLDAopt.dayfile::CHARGE convergence:  0 0.001 .0004560
> CarbazoleLDAopt.dayfile::CHARGE convergence:  0 0.001 .0004130
>
>    If I am not wrong, the convergence is achieved after 3 consecutive
> iterations with changes below the limit, but in my case this is happening
> for more than 10 iterations  without convergence.
>    Could you please kindly explain what may be happening ?
>    Thank you !
>                     Luis
>



-- 
Professor Laurence Marks
"Research is to see what everybody else has seen, and to think what
nobody else has thought", Albert Szent-Gyorgi
www.numis.northwestern.edu ; Corrosion in 4D: MURI4D.numis.northwestern.edu
Partner of the CFW 100% program for gender equity, www.cfw.org/100-percent
Co-Editor, Acta Cryst A