lattice basis and selective dynamics

[rpnabar]

I was trying to find an equlibrium crystal structure by a grid of calculations varying lattice parameters a, b and c(orthorhombic cell) each calculation being an ISIF=0 caclulation.(fixed cell shape and volume)

The lattice basis consists of some ions at general positions(x,y,z) and others at the special positions eg. (1/2 1/2 1/2) etc.

Is it necessary to relax each ionic position or would it be acceptible to keep the special positions fixed by selective dynamics.(that would save a lot of time too)

[admin]

that depends: if you give the atomic positions in direct coordinates and if there is strong experimental evidence that some of the atoms are in high-symmerty positions, you can leave these atoms fixed if you use ISIF=0.
In any case, I would do one test run for the converged cell to see whether these atoms change their positions if they are not restricted anymore.

[Nice]

thanks admin for you answer, but with ISIF=0 the ions will be relaxed, I do not want to relax them, I want to see the band structure for the experimental results. for this I choose to use a selective dynamics with (FFF) for all atoms, together with ISIF=2 and IBRION=2 also. What do you think about this? or have I to set IBRION=5 or 6 only?