Regarding broken-symmetry calculations in vasp

[rajaraman]

Dear Vasp users,
I'm doing a single-point broken-symmetry DFT calculation of a [Co3(pytag)(py)6Cl3]+ cationic molecule. Here each Co center is in +II, d7, high-spin state. However the broken symmetry state (Co1; S = 3/2, Co2; S = 3/2, Co3; S= -3/2) is the ground state. I'm using the following INCAR file
SYSTEM = Co3_BS1
ISTART = 0
ICHARG = 2
NCORE = 16
NSW = 0
EDIFF = 1E-05
LREAL = AUTO
ISMEAR = 0
PREC = Medium
SIGMA = 0.01
NELM = 300
ALGO = FAST
ISPIN = 2
NELECT = 363
MAGMOM = 1*3.90 1*3.90 1*-3.90 3*0 15*0 49*0 45*0
LORBIT = 11
For the broken symmetry state, we expect spin densities on Co1, Co2, and Co3 around 3.0, 3.0, and -3.0 respectively. However, in the OUTCAR file, the spin densities are coming around 0.989, 1.105, and 1.018. Can anyone please let me know what are the corrections needed to get the correct spin densities?

[merzuk.kaltak]

Dear Rajaraman,

Your system might require a DFT+U treatment (probably a U>0 on Co) to reach a specific solution.
However, we can tell you more, if you provide a proper error report.

[alex]

Hello Rajaraman,

two things you might consider:
a) spin densities are computed with the help of alpha and beta charge (!) densities (alpha - beta).
Since charges are not localized, you will probably never see the desired spin in spin densities. ~0.5e for one electron is in my experience a good indication. But be careful with that.
b) check OSZICAR for the magnetic moment, here you'll see if you reach the proper value for your system.

Best regards,

alex