SQS Different Bravais Lattice

Alloy Theoretic Automated Toolkit (ATAT)
1

As a follow-up to the question in a previous thread below, if the generated SQS is of a different Bravais lattice from the original supercell I defined in my RNDSTR.IN, what should I look out for when trying to link subsequent SCF or VC-RELAX calculations to my original unit cell?

For example, if I am attempting to derive cubic lattice constants for a disordered rocksalt structure, but the SQS returned is body-centered orthorhombic. What assumptions can I make to infer cubic lattice constants from VC-RELAX calculations on the orthorhombic SQS?

Many thanks in advance!

2

The right way to do this would be to symmetrize your lattice constants (in a way analogous to what was done in https://dx.doi.org/10.1103/PhysRevB.72.144109 for elastic constants).
In other words, take your relaxed SQS and apply all the point group operation of the original unit cell (here bcc). Then average these transformed lattice constant to get the symmetrize value that will satisfy the bcc symmetry by construction.