convergence problem with large magnetic system with impurities

[okuno]

Dear Vasp Users.

I 'm now attempting to calculate large magnetic system ( ferri-magnetic system) with impurities.

The calculation does not converged and oscillate.

I had change the values of MAGMOM from 2 to 4 but the results does not amended.
(atom's bare spin value is 5 )

So I want to set approximate values of AMIX BMIX, AMIX_MAG BMIX_MAG.

The VASP's manual says that in the magnetic system
AMIX=0.2 , BMIX = 0.0001 AMIX_MAG=0.8 BMIX_MAG=0.0001
is appropriate values, so I changed thiese parameter but also
in this case, the results was not improved.

Are there any remedy to get converged result in such a complex
magnetic system ?

I want to know how to set AMIX BMIX AMIX_MAG, BMIX_MAG to
improve the bad convergence, or how to get good initial value
for complex ferri magnetic system.

Sincerely.

My Input file is as follows.


#ISTART=1
ENCUT=650
PREC=High
ISMEAR=0
SIGMA = 0.01
ISPIN=2
#ICHARG=2
VOSKOWN=1
EDIFF=1e-5
MAGMOM = 0.0 0.0 3.0 3.0 3.0 3.0 \
-3.0 -3.0 -3.0 -3.0 -3.0 -3.0 \
3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 \
0 0 0 0 0 0 0 0 0 \
0 0 0 0 0 0 0 0 0 \
0 0 0 0 0 0 0 0 0 \
0 0 0 0 0 0 0 0 0 0 0 0

AMIX=0.1
BMIX=1.0
#BMIX=0.00001
AMIX_MAG=0.1
BMIX_MAG=1.0
#BMIX_MAG=0.00001

NELMIN=7

NELM=70

LWAVE=.FALSE.
LREAL=Auto

LDAU=.TRUE.
LDAUL=-1 2 -1 -1
LDAUU=0.0 6.0 0.0 0.0
LDAUJ=0.0 0.0 0.0 0.0
LDAUTYPE=2
LDAUPRINT=2

LORBIT=11

[admin]

without knowing further details, I would recommend to try the following:
please
keep the low mixing parameters
check if the k-mesh is converged
try if a different BZ-integration (ISMEAR=1) and slightly larger smearing (SIGMA) helps
set LMAXMIX=6 if your system contains d-elements

[okuno]

Dear administrator.

Thank you for reply.

About the mixing parameters, should be all the values, AMIX AMIX_MAG, BMIX, BMIX_MAG keep small ?

[admin]

yes, I would keep them all small.