I want to generate equivalent vector SQS with good correlation functions for phonon calculations. However, the mcsqs code normally doesn’t provide me equivalent vector SQS structures.
I know there is a way, which chooses the equivalent vector in sqscell.out and run "mcsqs -rc" to find the equivalent vector supercells, but this method cannot give me good correlation functions, which is not random enough.
So my question is that is there any way to convert the existed SQS structures to equivalent vector SQS structures.
Sorry, there is really no other approach than to provide your own preferred cell in sqscell.out and use the -rc option of mcsqs. If you can’t find a good match, then you need to try a bigger cell.
I have to tell you, however, Ithat do not see why you would want to have a cubic cell. The sqs formalism does not require this at all. If you are concerned that properties such as elastic constant would not have cubic symmetry, then using a cubic cell won’t ensure this either. The right way to do this is symmetrize the tensors you get from the low-symmetry sqs as described in https://dx.doi.org/10.1103/PhysRevB.72.144109.
Hi Professor Axel
Thank you for answer my questions. I just want to do phonon calculations. If the my SQS structures have the same lattice parameter. It will make my phonon calculation much easier. I just want to try what this article did. They chose the SQS structure with almost same lattice parameters.
Wang, Y., Zacherl, C.L., Shang, S., Chen, L.Q. and Liu, Z.K., 2011. Phonon dispersions in random alloys: a method based on special quasi-random structure force constants. Journal of Physics: Condensed Matter, 23(48), p.485403.
Ok I see. Really, you have to specify your own sqscell.out - that’s the only way. Maybe making the cell twice as big would allow a better sqs and still make the phonon calculations easily comparable.