My Community

Dear Vasp Users and developers.

I'm now working impurity problem in magnetic materials.

I set the non-magnetic impurity in the antiferromagnetic
and ferromagnetic materials, but these the self-consistent calculation does not converged. The magnetic material we
treat is metal.

How should we generally treat the non-convergence problem by Vasp.

We set

AMIX=0.1
BMIX=0.001
AMIX_MAG=0.1
BMIX_MAG=0.001

but the calculation did not converge, and change AMIX AMIX_MAG more small but the results is same.

I also change the mixing scheme like
IMIX = 1 ( Kerker mixing)
AMIX = 0.1
BMIX=10.0
AMIX_MAG=0.1
BMIX_MAG = 10.0

(We set rather large BMIX for Kerker mixing, is it right for
difficult convergence problem ? )


I want to know the best mixing scheme for difficult convergence problem, and how should we treat AMIX, BMIX
for each mixing scheme.

Sincerely.

usually a linear mixing scheme is best for the starting iterations of cells which are hard to converge. Have you checked OUTCAR for possibly different reasons for your bad convergence?