[Wien] new version & LDA+U for triclinic system

[Hua Wu]

Dear WIEN2k users,

   I have tested the LDA+U fix as described below for the triclinic
system I mentioned two weeks ago. The fix works for that problem !!
Thanks a lot to the WIEN2k authors.  -- H. Wu

On Thu, 4 Nov 2004, Peter Blaha wrote:

> Dear WIEN users!
>
> A new version is on the WIEN2k update page.
>
> The reason for this extra email is, that we think we found a possible severe
> bug in the LDA+U part of the program. The situation is not yet completely
> clear but our tests seem to indicate that these fixes are necessary.
>
> Problems with LDA+U might have been in 2 situations:
>
> a) Real case, without spin-orbit coupling (with inversion symmetry),
>   but the density matrix (of some atoms) has imaginary parts:
>
>   In this case we believe now that both, the LDA+U eigenvalues (due to a
>   bug in SRC_lapw1/horb.F) AND the LDA+U total-energy (due to a problem in
>   SRC_orb/vldau.f) are wrong.
>
> b) Complex case (without inversion symmetry): The LDA+U eigenvalues are ok
>   (also the scf-solution is ok), but the total-energy is wrong.
>
> Please check all your LDA+U calculations wether or not the density matrix
> has an imaginary part (just check the scf file!).
>
> A purely "real" dmat gave probably correct results.
>
> With the fixes in the new versions several tests with real/complex code and
> supercells with reduced symmetry seem to give consistent results.
> Unfortunately we do NOT yet really understand the fix for E-tot.
>
> Regards
>
>                                      P.Blaha
> --------------------------------------------------------------------------
> Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
> Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
> Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
> --------------------------------------------------------------------------
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>