[Wien] A question about RMT and AIM

[Laurence Marks]

Aim is a hard calculation to get right because there is not currently
any really good algorithm for finding the BCP and surface, and it is
done by a brute force search method. What can certainly create
problems are:

1) The discontinuities at the RMT's. You need to use very high clm's
(in case.in2) to reduce this; for a small calculation like you have
you may want to use an lapw basis set (rather than APW+lo) since this
has no discontinuity. (I have a small utility to calculate the
discontinuities.)

2) Insufficient sampling. At the end of case.outputaim there are some
error guesstimates which are approximately correct (within a factor of
2). If these are large you need to increase the sampling in
case.inaim.

3) Using too large an initial R and/or too large steps.

In general, calculations with smaller RMT's should be more reliable
albeit slower.

On 11/27/06, G.C.ZHOU <gc_zhou1982 at 163.com> wrote:
>
>
>
>  Dear Professor and users:
>  I am doing the AIM calculation of ZnO, it has three
> structure:wurtzite,zinc-blende and rocksalt.
>  when I do the calculation of BCP,the other setting are all same,I find RMT
> affect the result very greatly.
>
> the structure is wurtzite,
>    and set R(Zn)=1.75, R(O)=1.85, the Laplace of BCP is 5.003;
>    when set R(Zn)=1.85, R(O)=1.90, the Laplace of BCP is 0.580,and the
> position of BCP is not correct.
>
> the structure is zinc-blende,
>   and set R(Zn)=1.75, R(O)=1.85, the Laplace of BCP is 4.993;
>   when set R(Zn)=1.85, R(O)=1.90, I can't find the BCP.
>
> the structure is rocksalt,
>   and set R(Zn)=1.75, R(O)=1.85, the Laplace of BCP is 4.511;
>   when set R(Zn)=1.94, R(O)=2.10, the Laplace of BCP is 3.256;
>
> I am puzzled with this results, but the two results of evety structure are
> too different with only the RMT is changing. And the Laplace of rocksalt's
> BCP tells that, in the rocksalt structure, the covanlent of Zn-O bond turns
> more then wurtzite, it's contrary with our earlier cognition, which is the
> bonding became more ionic in this structure.
> even worse, when I reduce the RMT, the laplace will turns smaller yet.
>
> I set RKmax=9.0, LM=10, and set Ecut=-6.5Ry, there is no electron leak out.
> and use x lapw -c to do the program.
>
> Can tell me why the BCP change so much in different RMT, and which
> result should I take as the right one.
>
>  Any advices is welcome!
>
> Thanks.
>
>  ________________________________
>
> G.C.ZHOU
> 2006-11-27
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
>
>


-- 
Laurence Marks
Department of Materials Science and Engineering
MSE Rm 2036 Cook Hall
2220 N Campus Drive
Northwestern University
Evanston, IL 60208, USA
Tel: (847) 491-3996 Fax: (847) 491-7820
email: L-marks at northwestern dot edu
http://www.numis.northwestern.edu