how to interpret spin-orbit calculation results

[SangLee]

Dear members:

I would like to ask questions about spin-orbit calculations.

1. Which coordinate system is used to report the magnetic moments printed at the end of the OUTCAR obtained from a spin-orbit calculation ?
Is it the cartesian coordinate system used in POSCAR (or m^axis ) or the cartesian coordinate system set by SAXIS in INCAR when these two system are different each other ?

2. I have calculated a molecular cation containing a heavy atom. Let me call one chemical bond in the cation as a reference bond. After the reference bond was aligned along x direction, I performed spin-orbit calculation. I repeated the calculations with the reference bond aligned y and z axis. The resulting three E0 energies do not agree each other. It is confusing me because I believe that these three energies must be the same. Would somebody give me an explanation about it ?

Thank you in advance.

Sang

[admin]

1) the written information in OUTCAR refers to the global carthesian x,y,z directions, not to SAXIS
2) I am not quite sure what you did exactly: in principle you should obtain the same total energies only if you rotate eg the complete complex such that the principal axis eg rotates from x to z and if you also rotate the spin quantization axes accordingly. if you transform just one of them , the energies need not be the same (pelase read the manual: LSORBIT is meant to lift the degeneracy of the magnetic moments with rotation)