# What is the source of error in this mBJ-LDA calculation?

I am trying to do MBJLDA calculation for bandstructure for WS2 monolayer unit cell using VASP and following the method provided here : Band Structure mBJ-LDA using Vasp, but I am getting an error message. The INPUTS are below:

INCAR:

ENCUT=600
PREC = Accurate
EDIFF=1E-6
NELM=300
ISMEAR=-5

###MBJ:
METAGGA=MBJ
LASPH=.TRUE.
LMAXMIX=4
GGA=CA


LDA POTCAR:

  PAW W_pv 06Nov2007
TITEL  = PAW W_pv 06Nov2007
PAW S 27Mar2009
TITEL  = PAW S 27Mar2009


POSCAR:

 W2S4
1.00000000000000
3.1912863255000001    0.0000000000000000    0.0000000000000000
-1.5956425563000001    2.7637348281000000    0.0000000000000000
0.0000000000000000    0.0000000000000000   20.1269199999999984
W    S
1     2
Direct
0.6666670084824631  0.3333329922369330  0.5000000000000000
0.3333330949386806  0.6666670699615054  0.4218851941843354
0.3333330949386806  0.6666670699615054  0.5781148058156645

0.00000000E+00  0.00000000E+00  0.00000000E+00
0.00000000E+00  0.00000000E+00  0.00000000E+00
0.00000000E+00  0.00000000E+00  0.00000000E+00


The calculations was based on SCF-PBE WAVECAR.

• Is that the end of the error message? Feb 12, 2021 at 16:33
• I suggest you first repeat the result of the linked tutorial with a 3D semiconductortor.
– Jack
Feb 14, 2021 at 23:07
• Following the linked tutorial, I can't repeat your error.
– Jack
Feb 14, 2021 at 23:08
• It should be noted that the MBJ-LDA method for 2D materials is tricky to obtain converged results.
– Jack
Feb 15, 2021 at 2:59
• +1. Related: mattermodeling.stackexchange.com/q/391/5 and mattermodeling.stackexchange.com/a/1470/5. A little more on mBJ: mattermodeling.stackexchange.com/a/1048/5. Feb 15, 2021 at 8:18

I assume that:

For monolayer, it is very hard/tricky to converge at MBJ self-consistent calculation. Usually, one possible way is you can obtain the CMBJ=XXX from its bulk counterpart.

I have tried this but get unsatisfactory result:

• Bulk WS2 (SOC is ignored!):

• PBE band:

• MBJ band:

grep CMBJ OUTCAR ====> CMBJ=1.2327

• Monolayer WS2:

• PBE result:

• MBJ result:

### INCAR

 ### Read PBE wavefunction
ISTART=1
ICHARG=0
ENCUT=400
EDIFF=1E-6
NELM=300
ISMEAR=-5

###MBJ:
METAGGA=MBJ
LASPH=.TRUE.
LMAXMIX=4
GGA=CA
CMBJ=1.2327


This unsatisfactory result may be related to the badly converged MBJ self-consistent calculation. Unfortunately, I haven't found any solutions for this (still trying...).

By the way, the system contains only three atoms, why not using the HSE06 method?

Hope it helps.

• I am also getting the same error for a bulk material, my system is quite big. I am not sure how did you fix the question asked by Chi kou. Like I know that setting a value for CBMJ helps with convergence but I do not have any value for my specific case and I have tried a few values none of which helo this issue ? Furthermore, I do not get this error in PBE so this should be something related to mBJ only.
– Chan
Sep 7 at 9:20
• So actually my issue went away if I simply did not put ISMEAR = -5 in my INCAR. But very confused as to why this problem was happening in the first place.
– Chan
Sep 7 at 10:23