[Wien] x symmetry problem

[Mr C.A. Bridges]

On Fri, 21 Nov 2003, Peter Blaha wrote:

>> I’m trying to run SCF cycle for a monoclinic TiO. In the data I have
>> (program ICSD) for the space group is written A 1 1 2/m, #12. I saw that
>> #12 in the StructGentm is B2/m. When I started xgroup I got that:
>>
>> sgroup found: 12 (C 2/m) [unique axis c] cell choice 2
>>
>> In the next step: x symmetry, I got:
>> FORTRAN STOP -3 missing
>
>Centered monoclinic cells are only partially supported by WIEN.
>The monoclinic angle MUST be gamma, so you must transform your spacegroup
>representation into B2/m (basically interchanging some axis and positions)
>and use this representation.
>
>Furthermore, please note that there is a yet unresolved bug for monoclinic
>centered cells, which I can only summarize as: SOMETIMES the nonspherical
>potential terms are wrong for this symmetry.
>As SOMETIMES is not very specific, you MUST check your results using a
>doubled P cell (or a triclinic P cell of same size, but no symmetry)


Sorry, I would just like to clarify this for myself. When you refer to a
'doubled P cell', do you mean that as the asymmetric unit of the
primitive cell is twice that of the higher symmetry centered cell, you
have to double the asymmetric unit size (rather than the entire cell)?
And when you refer to a 'triclinic P cell of the same size', this is
referring to a cell the same size as the monclinic centered cell?

Basically, I am wondering if it is sufficient to lower the symmetry from
C112/m (for example) to the subgroup P112/m? This would have the same
size of unit cell (twice the size of asymmetric unit), and unique axis
c.

Please note for users of this list that there is a freely available
windows program called "Powder Cell for MS-Windows" (clearly a windows
program) on the site "http://www.ccp14.ac.uk" that will readily
transform crystal structures between various space group settings.

Craig