Dear Abinit Community.
I am attempting to investigate the formation energy and lattice structure of substitutional rutile lattices. I have run a 48 atom super cell of TiO2 with a single Titanium atom replaced with Vanadium the total ground state energy is as expected different but when identifying the space group instead of Cm m m (# 65); Bravais oC (1-face-center ortho.) instead of the expected space group of P4_2/m n m (#136) Bravais tP (primitive tetrag.). When working with a pure TiO2 or VO2 crystal the space group is correctly identified, but when impurity atoms are added the space group changes even though both Pure TiO2 and VO2 share the same space group of #136. The pseudopotentials used are paw type (JTH v1.1) xc (PBEsol) accuracy (standard) format (xml). The files attached contain the input, output and log files of a simulation of a single substitution of V in place of Ti. Are there any methods to force the symmetry operations into an expected space group?
TiO2_48atombulk15Ti1V.abi (3.9 KB)
TiO2_48atombulk15Ti1V.log (2.1 MB)
TiO2_48atombulk15Ti1V.abo (793.8 KB)
Dear cmlyons6,
If you are certain of the space group symmetry of your doped system, then you can simply use symrel. In this keyword you can assign all the matrices corresponding to the symmetry operations related to the space group that you expect to have.
You can find an example of its use here.
Another (easier) possibility is to increase the symmetry tolerance above the default one using this keyword tolsym. Does the space group change when you go to 1E-4 or 1E-3? You can also perform this check directly on the website GUI (Structure Analyzer) where you can tune two tolerance parameters. Depending on the strictness of your relaxation convergence, this will reflect on how many digits you can rely on in your obtained structure, so the symmetry tolerance is a good indicator in your situation.
Let us know!
Bogdan
Hi, I may be misunderstanding your question, but it reads to me as though you are expecting that the doped supercell should have the same symmetry as the original undoped TiO2 and VO2 supercells? However, if you replace a Ti by V at one site in the TiO2 supercell, say, then the symmetry of the new system is certainly going to be lower than in the undoped case. I suspect that abinit is correctly identifying this lower symmetry (providing I’m understanding your question correctly!)