My Community

Dear Admin,

Has the "spin contamination" problem been looked in the unrestricted HF in VASP? The problem is related to a mixture of spin states with higher spin multiplicity, so the wave function is no longer an eigenfunction of the total spin. Since I am looking at an open shell system, I worry that this error may affect the total potential energy surface or transition states.

Some codes seem to allow printing out the expectation value of the total spin to check the spin contamination. Can this be done in VASP? If you know any references on this issue using VASP, please let me know.

Thanks,

Na

the "spin contamination" problem is not addressed at all in vasp,
the partial occupancies of the Bloch states are calculated within the
minimization of the total energy.