[Wien] Diamond lattice

[Peter Blaha]

> > In the case of Diamond you need t put the atom at 1/4 1/4 1/4. Only one
> > atom.
> > Below is a struct file which works.
>
> Thank you, that works. Do you happen to know why this has to be this way? When
> using a drawing program, I can construct the lattice using the spacegroup and
> one atom at (0,0,0), a 8a-Wyckoff position. The lattice composed by WIEN is
> the same, although the cell looks rather awkwardly due to the 1/4-shift.

This spacegroup #227 has two different settings of origin. One with
(0,0,0) at 16c, pointgroup -3m, i.e. it has inversion symmetry at the origin
of the cell. In this setting the two C atoms are at position 8a (1/8,1/8,1/8)

The "more common" setting of this spacegroup has the origin (0,0,0) at 8a,
which has pointgroup -43m, i.e. it DOES NOT have inversion. This is the setting
in which the diamond structure is usually given (since a C atom is at
the origin of the cube) and people like this drawing.

However, in this setting all matrix elements in a calculation would be
complex numbers and the calculation would take 4x as long and need 2x memory.

Thus WIEN2k REQUIRES to use a setting where inversion is at the origin of
the cell.

What do you learn from this: The name of the spacegroup + some coordinates
of an atom are sometimes NOT sufficient to describe a structure. You must
either also specify which "setting" of the spacegroup was chosen or
specify which Wyckoff position was given (and from this you know which setting
was used).


                                      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/
--------------------------------------------------------------------------