[Wien] problem with supercell

[Stefaan Cottenier]

> I never did this myself, but I remember that it has been discussed on the
> mailing list about a year ago (you can check the digest).  People asked
> about using dummy atoms (i.e., empty spheres in your structure).  I remember
> that P. Blaha was not really fond of this idea, but to those who persisted
> in trying, he advised not to use a dummy atom with Z=0 during
> initialization, but one with a small charge (eg. 1).  Then you would have to
> correct this in case.struct (and also the number of electrons in case.in2, I
> suppose) before starting the SCF.
> I'm not sure how that might relate to your problems, but I thought I'd
> mention it anyway.

Kevin, I don't think this dummy atom is appropriate here. Vacancies can be 
calculated by just removing an atom, and such a calculation should proceed 
without any special problems (just slower: less symmetry and more vacuum). 

Angelica: did you initialize this supercell-with-vacancy from scratch? If you 
first did the cell with the original atom, then removed the atom and 
continued, you will have problems.

Stefaan