Groundstate magnetic calculations on Fe

Message

Jiaye Li · #1 Post by **Jiaye Li** » Tue Apr 29, 2008 1:30 am

Hi

I am performing calculations on Fe surface, aiming to find the groundstate configuration and energy.

Should I only set ISPIN=2 in the INCAR for the magnetic groundstate calculation? Some others used generalized gradient spin approximation (GGSA) for the calculation with the same system, what is GGSA? Is it GGA + ISPIN=2?

Thanks

Jiaye Li · #2 Post by **Jiaye Li** » Tue Apr 29, 2008 1:51 am

I read from the vasp mannul "In some rare cases, the real LDA groundstate might differ from the configuration for which the pseudopotential was generated (most transition metals, see Sec. 11). For the atomic all electron reference calculation the occupancies have been set by hand (for Pd a was chosen to be the reference configuration, which is not the LDA groundstate of the atom). In this case it is also necessary to set the occupancies for VASP also by hand, to get the same energy as that one found in the POTCAR file. This can be done including the following lines in the INCAR file:

LDIAG = .FALSE. ! keep ordering of eigenstates fixed
ISMEAR = -2 ! keep occupancies fixed
FERWE = 5*0.9 0.5 ! set the occupancies manually

(5*0.9 is interpreted as 0.9 0.9 0.9 0.9 0.9). "

My question is how to set the FERWE and FERDO flags? Are they neccessary when MAGMOM is set in the INCAR?

Thanks.

Jiaye Li · #3 Post by **Jiaye Li** » Tue Apr 29, 2008 2:18 am

I did some calculations with Fe bulk, after one iteration it said "Your highest band is occupied at some k-points! Unless you are |
| performing a calculation for an insulator or semiconductor, without |
| unoccupied bands, you have included TOO FEW BANDS!! Please increase |
| the parameter NBANDS in file 'INCAR' to ensure that the highest band |
| is unoccupied at all k-points. It is always recommended that one |
| include a few unoccupied bands to accellerate the convergence of |
| molecular dynamics runs (even for insulators or semiconductors) |
| Because the presence of unoccupied bands improves wavefunction |
| prediction, and helps to suppress 'band-crossings.' "

Should I add NBANDS in INCAR, and how much I should assign?

forsdan · #4 Post by **forsdan** » Tue Apr 29, 2008 8:40 am

1. Yes. The ground state for Fe is a ferromagnetic bcc lattice. So if you want to calculate the bulk and surface properties for this system you have to set ISPIN = 2, which allows for spinpolarized calculations.

2. You're correct. GGSA = GGA + ISPIN=2.

3. The FERWE/FERDO tags are for the atomic references calculations, if you want to look at a specific configuration. You don't need to assign them for bulk calculations. The occupancy will be determined automatically.

4. Instructions for NBANDS can be found in the manual:
http://cms.mpi.univie.ac.at/vasp/vasp/node215.html

Best regards,
/Dan Fors

Jiaye Li · #5 Post by **Jiaye Li** » Sat May 03, 2008 5:08 am

Thank you!

I test the FCC bulk Fe by setting MAGMOM=4 (or some other values), the calculated magnetic value is 2.28, which is close to the experimental value of 2.12.

If the MAGMOM flag didn't exist in INCAR, it will meet with "Your highest band is occupied at some k-points! Unless you are |
| performing a calculation for an insulator or semiconductor, without |
| unoccupied bands, you have included TOO FEW BANDS!! Please increase |
| the parameter NBANDS in file 'INCAR' to ensure that the highest band |
| is unoccupied at all k-points. It is always recommended that one |
| include a few unoccupied bands to accellerate the convergence of |
| molecular dynamics runs (even for insulators or semiconductors) |
| Because the presence of unoccupied bands improves wavefunction |
| prediction, and helps to suppress 'band-crossings.' |"

by setting NBANDS in INCAR will eliminate the above warnings, but will slow down the convergence. It is worthy of being noted that the calculated magnetic value by setting NBANDS is consistant with that by setting MAGMOM. So I think NBANDS should be avoided in INCAR to speed up convergence if one doesn't need to calculate the band structure(I didn't check the difference). Say, by only setting MAGMOM it can give rise to a ground state energy for Fe calculation. What I speculated is reasonable?

Jiaye Li · #6 Post by **Jiaye Li** » Sun May 04, 2008 1:38 pm

For paramagnetism calculation, should I set ISPIN=1 or 2?

#7 Post by **admin** » Mon May 05, 2008 1:35 pm

1) your calculation must have gone wrong, there are always several unccupied bands included by default, and the default choice usually is more than sufficient to keep a few bands un-occupied.
2) set ISPIN=2

Jiaye Li · #8 Post by **Jiaye Li** » Tue May 06, 2008 12:13 pm

This is my INCAR:
SYSTEM = Fe
Startparameter for this Run:
NWRITE = 2 verbosity write-flag
LPETIM= F write-flag & timer #
ISTART = 0 0-new 1-cont 2-samecut job
ISPIN = 2 1-no, 2-yes
INIWAV = 1 0-jellium WF; 1-random numbers (recomm.)

Electronic Relaxation
! ENCUT = 400.00 #Ecut (eV)
PREC = Medium * #precission: #Low/Medium/High
NELM = 150 max. # of electronic SC steps
NELMDL = 5 # of non-consistent steps at the beggining:>0 or <0
EDIFF = 1.0e-4 stopping-criterion for electronic upd.
VOSKOWN = 1
ISMEAR = 2
SIGMA = 0.1
AMIX = 0.2
BMIX = 0.0001

Ionic relaxation
EDIFFG = 1.0e-3 stopping-conditions for ionic relax.loop (EDIFF*10)
NSW = 250 * # of steps for ionic update (def:0)
IBRION = 2 * ionic relax: -1:no move; 0-MD;1-Newton;2-CG;3-damped;4-?
ISIF = 2 * calc.stress:0-no;1-tr;2-7 yes; 3-change vol.4-change shape
! ISYM = 2 symmetry:0-OFF, 1-ON,2-ON+memory conserving for CD
NBLOCK = 1
POTIM = 0.3 if IBRION=1,2,3:scaling constant forces;IBRION=0,time st.
! LPLANE = .TRUE.
! NPAR = 8
IALGO = 48 * algorithm: use only 8 (CG) or 48 (RMM-DIIS)
LDIAG = .TRUE. sub-space diagonalisation
GGA = 91 * xc-type: PB, PW (86), LM, 91 (Perdew-Wang 91)

LWAVE = .T. * write WAVECAR
LCHARG = .T. * write CHGCAR and CHG
LVTOT = .T. * write the local potential LOCPOT
LELF = .F. create ELFCAR file
LORBIT = .F. create PROOUT
LREAL = Auto

POSCAR

Fe
1.00000000000000
5.7328 0 0
0 5.7328 0
0 0 5.7328
16
Selective dynamics
Direct
-0.00000 -0.00000 -0.00000 T T T ! Fe
-0.00000 -0.00000 0.50000 T T T ! Fe
-0.00000 0.50000 -0.00000 T T T ! Fe
-0.00000 0.50000 0.50000 T T T ! Fe
0.50000 -0.00000 -0.00000 T T T ! Fe
0.50000 -0.00000 0.50000 T T T ! Fe
0.50000 0.50000 -0.00000 T T T ! Fe
0.50000 0.50000 0.50000 T T T ! Fe
0.25000 0.25000 0.25000 T T T ! Fe
0.25000 0.25000 0.75000 T T T ! Fe
0.25000 0.75000 0.25000 T T T ! Fe
0.25000 0.75000 0.75000 T T T ! Fe
0.75000 0.25000 0.25000 T T T ! Fe
0.75000 0.25000 0.75000 T T T ! Fe
0.75000 0.75000 0.25000 T T T ! Fe
0.75000 0.75000 0.75000 T T T ! Fe
KPOINTS

Fe !Comment
0 !automatic generation of k-points
Monkhorsh-Pack !use Monkhorst Pack
4 4 4 !grid 4x4x4
0 0 0 !shift (usually 0 0 0)

Would you please figure out where the problem is?
Thanks.

#9 Post by **admin** » Wed May 07, 2008 9:08 am

the problem can be solved by setting the moments of each atom (MAGMOM) explicitely to (at least, in your case) 2 or 3:
if MAGMOM is not set, a default value of 1 per atom is assumed and the number of bands (NBANDS) is estimated from the expected number of majority spin bands (+ a small surplus) for a local magnetic moment of 1muBohr for each atom. If the magnetic moments at each site are substantially higher (as in your case), this may not be sufficient to have enough spin up bands to be filled.

Jiaye Li · #10 Post by **Jiaye Li** » Sat May 10, 2008 6:45 am

I set MAGMOM in my INCAR and performed the calculation again, at this time no warning concerning the insufficiency in spinup bands was observed. Thank you very much for your attention!