# How to analyze the space group of a relaxed structure that has fractional site occupancies

I am using VASP package to model a material that has fractional site occupancies. An appropriate supercell (cif) was generated using supercell package and converted to POSCAR using VESTA. Then the structural relaxation was performed. Now I have the relaxed structure (CONTCAR) and I want to analyze its space group.

I tried pymatgen, but it doesn't seem to be working for fractional site occupant materials, because the original POSCAR structure is Pm-3m, however is identified as P4/mmm by pymatgen.

Is there any other tool I can use to analyze such structures?

• I was desperately looking for materials modelling Q&A site and finally here we are. Hope it will help a lot! – Achintha Ihalage Jun 11 '20 at 19:52
• I'm very happy that you found us. I've been trying to recruit people since 9 months ago and it would have been nice if I managed to reach you back then. I also tried to make a site like this (though different) in 2018, but StackExchange rejected the proposal. How did you find us? Which Hot-Network-Question? We are only 1 month old so this is the best time to get involved! – Nike Dattani Jun 11 '20 at 19:56
• I was actually on StackExchange Chemistry and saw a hot-network-question (I don't exactly remember). Always reluctant to ask materials questions on chemistry site. I hope this site will attract many experts too. – Achintha Ihalage Jun 11 '20 at 20:12
• Was it this one, or something else? – Nike Dattani Jun 11 '20 at 20:16
• I definitely recall this question. Probably it was the one I saw. – Achintha Ihalage Jun 11 '20 at 20:21

### ASE

The ASE python package can find the space group of a structure using the get_spacegroup function as follows.

import ase.io as io
from ase.spacegroup import get_spacegroup

sp = get_spacegroup(structure,symprec=1e-5)
print(sp)


I am not sure as to whether it can deal with fractional occupancies, but it seems it can from the source of the Spacegroup class.

• Thanks. This module can deal with fractional site occupancies (handling cif, POSCAR etc,, whereas pymatgen just gives an error). However, I had to do a small workaround to get the space group. – Achintha Ihalage Jun 12 '20 at 15:30
• @AchinthaIhalage It would be helpful for future visitors if you could edit the workaround into the answer and I can accept the edit. Or ofcourse you can write another answer :). – Rashid Rafeek Jun 13 '20 at 1:38
• It was nothing major. I was focused on one fractional site occupant material. It's space group was correctly identified by ASE using the cif file as it is, I guess it will work for other materials too. But when I created a supercell POSCAR, the identified space group was wrong, so I had to convert it back to cif to get the correct space group. And I'm not sure if this dirty workaround works for other materials. – Achintha Ihalage Jun 14 '20 at 9:16

### spg-lib

Another light-weight option is the spg-lib by Atsushi Togo and Isao Tanaka described here with multiple interfaces that can be found here.

Especially, the tolerance parameters for the detection of symmetries are quite handy.

### Bilbao Crystallographic Server

You can take a look at the Bilbao Crystallographic Server.

It has several tools for symmetry and structure analysis in the Solid State Theory Applications section.

### FINDSYM

Another useful tool to determine crystal symmetry is FINDSYM. It works with the .cif format exclusively, so you may have to convert your structural files to .cif before using it.