[Wien] Re : Supercell problem of Al2O3

[Michael Fischer]

Hello,

I think I can identify the problem, without knowing how to solve it (in 
an efficient way): The bottom of the struct-file must contain all 
symmetry operations of the hexagonal setting of space group 167, i.e. 
the 12 ones of the rhombohedral setting times 3 (translations in 
hexagonal cell) -> 36 symmetry operations.
As we constructed your struct-file by modifications of the one for the 
rhombohedral setting, it contains only 12 symmetry operations. This is 
where the error in the outputs-file comes from, where you read 
something like this:

   The (multiplicity of this atom)*(number of pointgroup-operations) is 
NOT
   = (number of spacegroup-operations)
  MULT:          18  ISYM:           2  NSYM          12 				<--- if 
NSYM was 36, everything would be fine
  Check your struct file with      sgroup -wien case.struct

I suspect the ROTDEF-error is connected to this, too.

As I said, I do not have a clue how to fix this efficiently. It could 
be possible to add the necessary symmetry operations at the bottom of 
the file by hand, but I think/hope that more experienced users will 
have a better solution.
Alternatively, you could specify all 30 atoms on formally inequivalent 
positions, each with MULT = 1.

Regards
Michael