Partial occupancy groups get fixed by themselves

[orex]

Partial occupancy groups get fixed by themselves

[orex]

In the example you wrote, there is only one site for the disordered atom. So, it is occupied either by 1 Fe or 1 Si atom. Please increase the cell size to get the result. For example,

supercell -d -i <input>.cif -m -s 2x2x2

[gagarofalo]

But in the Ca2Al2SiO7 example the disordered site is distributed even in the 1x1x1 run.
I would like to avoid increasing the cell size because, in the end, I'd like to do just one 1x1x1 big cluster (100s of atoms big) randomly assigning the site species. Is that possible?

[orex]

Again. You have ONE site with multiplicity 1 in you cell. When you have only physical 1 place where you can put your atom, it is not possible to distribute several atoms around one site (see output).

...
Group 1 (1 atomic position in supercell):
...

For Ca2Al2SiO7 example:

....
 Group 2 (4 atomic positions in supercell):
  * Site #1: AlT2 (occ. 0.5) -> distributed over 2 positions out of 4 (actual occ.: 0.500).
  * Site #2: SiT2 (occ. 0.5) -> distributed over 2 positions out of 4 (actual occ.: 0.500).
  Number of combinations for the group is 6
...

[gagarofalo]

Ok thanks I see.