Supercell - CIF

als

This might be a little off topic for this forum, but I thought I give it a try.

I need to do a Rietveld refinement with a supercell. I have not done this before, typically I just use a standard CIF for my Rietveld refinements in TOPAS.

The structure I have is trigonal, R-3 with about 10 atoms in the asymmetric unit (these are what I would typically refine). However I need a 1x1x2 supercell. I have created this in VESTA, however when loaded into TOPAS the list of atoms to be refined is a few hundred! This is simply the asymmetric unit * multiplicity of the cell * supercell.

Is there a way to generate supercell either outside of TOPAS and import, or within TOPAS, such that I would only have about 20 atoms in the asymmetric unit and the rest generated by symmetry elements?

johnsoevans

I think vesta lets you do this via the "transform" feature and then outputting a cif.
I tend to use the isodistort (https://stokes.byu.edu/iso/isodistort.php) or isocif tools. You can write supercells out either in cif format or str format. The str format is set up to let you use a distortion mode basis for refinement, but the cif format can be read into topas/jEdit and handled in the usual way.
The instructions at the start of the tutorial at https://topas.webspace.durham.ac.uk/tutorial_isoriet_wo3_simple/ describes creating a supercell. Here it's 2x2x2 in P1 but you can change settings as you need.

rowlesmr

I know that it doesn't help very much, but when I did this (maybe twice? it was a while ago), I just did it in P1 and manually linked all of the atom positions with the required symmetry operations.

als

After working through some possibilities, I'm leaning towards going down the route of making a 1x1x2 supercell in P1 and manually linking the atom positions with the required symmetry operations.

Would you be able to provide an example of linking atoms with symmetry with required symmetry operations?

johnsoevans

You'd just need a set of equations for all the fractional coordinates like:

site A1_1 x xA1 0.1 y yA1 0.2 zA1 0.3 .......
site A1_2 x = xA1+0.5; y = yA1+0.25; z = zA1; ..........

You'd need to figure out the equations using e.g. international tables.

I'm not sure what you're trying to do, but if you do set all this up the diffraction pattern calculated in the p1 supercell will be identical to a conventional description but will be slower to calculate and it will be easy to introduce errors. The ISODISTORT approach is designed to make some group-subgroup work easier and less error prone.