GW band structure

[wwlphys]

Hi, I am trying to plot GW band structure. The figure (attached "band structure.png") seems not right at the boundary of Brillouin zone.

The calculation consists of three steps:
First step: GW

INCAR
ISMEAR = 0
SIGMA = 0.05
ALGO = EVGW0
LOPTICS = .TRUE.
NELMGW = 4
MAXMEM = 7848
LWAVE = .TRUE.

Second step:

INCAR
ISMEAR = 0
SIGMA = 0.05
ALGO = NONE
NELM = 1
MAXMEM = 7848
NBANDS = 7040
LWAVE = .FALSE.
LCHARG = .FALSE.
LWANNIER90=.TRUE.
WANNIER90_WIN = "
exclude_bands 301-7040
"

Third step:

INCAR
ISMEAR = 0
SIGMA = 0.05
ALGO = NONE
NELM = 1
MAXMEM = 7848
NBANDS = 7040
LWAVE = .FALSE.
LCHARG = .FALSE.
LWANNIER90_RUN = .TRUE.

wannier90.win
bands_plot = true

begin kpoint_path
R 0.50000000 0.50000000 0.50000000 G 0.00000000 0.00000000 0.00000000
G 0.00000000 0.00000000 0.00000000 X 0.50000000 0.00000000 0.00000000
X 0.50000000 0.00000000 0.00000000 M 0.50000000 0.50000000 0.00000000
M 0.50000000 0.50000000 0.00000000 G 0.00000000 0.00000000 0.00000000
end kpoint_path

exclude_bands 301-7040

# This part was generated automatically by VASP
num_bands = 300
num_wann = 120
#use_bloch_phases = .true.
begin unit_cell_cart
8.2575924 -0.0000000 0.0000000
-0.0000000 8.5244477 0.0000000
0.0000000 0.0000000 11.8865529
end unit_cell_cart
begin atoms_cart
Cs 4.2677600 3.7684559 2.9716382
Cs 0.1389638 0.4937680 8.9149147
Cs 8.1186286 8.0306798 2.9716382
Cs 3.9898324 4.7559918 8.9149147
Pb 4.1287962 -0.0000000 5.9432765
Pb -0.0000000 4.2622239 0.0000000
Pb -0.0000000 4.2622239 5.9432765
Pb 4.1287962 -0.0000000 0.0000000
Br 4.7527079 8.3979065 8.9149147
Br 0.6239116 4.3887651 2.9716382
Br 6.5702288 1.7636265 0.4591257
Br 1.6873634 6.7608214 11.4274273
Br 2.4414328 2.4985972 11.4274273
Br 5.8161598 6.0258503 0.4591257
Br 2.4414328 2.4985972 6.4024020
Br 5.8161598 6.0258503 5.4841509
Br 1.6873634 6.7608214 6.4024020
Br 6.5702288 1.7636265 5.4841509
Br 7.6336807 4.1356826 8.9149147
Br 3.5048845 0.1265413 2.9716382
end atoms_cart
mp_grid = 2 2 2
begin kpoints
0.000000000000 0.000000000000 0.000000000000
0.500000000000 0.000000000000 0.000000000000
0.000000000000 0.500000000000 0.000000000000
0.500000000000 0.500000000000 0.000000000000
0.000000000000 0.000000000000 0.500000000000
0.500000000000 0.000000000000 0.500000000000
0.000000000000 0.500000000000 0.500000000000
0.500000000000 0.500000000000 0.500000000000
end kpoints

Four step: gnuplot -persist wannier90_band.gnu

OUTCARs and wannier.wout are attached.

[ferenc_karsai]

This calculation is unfortunately to big for us to redo. Your input tags look ok, so could not spot there anything at first glance.

We should narrow down the problem. So first please try to plot the bandstructure with wannier but using PBE instead of GW. This means replace only step 1 by a fast PBE calculation. If the error persists there is most likely something wrong in the wannierization.

[wwlphys]

Hi Ferenc,
I replace the first INCAR with
ISMEAR = 0
SIGMA = 0.05
#ALGO = EVGW0
LOPTICS = .TRUE.
#NELMGW = 4
MAXMEM = 7848
LWAVE = .TRUE.

The result is similiar.
Can you redo this?

[henrique_miranda]

To get a good Wannierization of the band structure can be tricky and it often requires a lot of trial and error in getting good initial projections, energy windows, k-points number of bands etc...
I would recommend that you first try to get a good Wannierization at the DFT level and only once that is done then you try to do it after the GW calculation.

That being said, I succeeded in getting a good Wannierization for this system with the following input file:

INCAR:

Code: Select all

ISMEAR = 0
SIGMA = 0.05

NBANDS = 150

LWANNIER90_RUN=.TRUE.
LWANNIER90_AUTO_WINDOW=.TRUE.
LWANNIER90_WIN="
# disentanglement
dis_num_iter = 500
dis_conv_tol = 1e-8

#iterations
num_iter          = 500
num_print_cycles  = 50
conv_tol          = 1e-4
"

LSCDM=.TRUE.
LWRITE_MMN_AMN=.FALSE.
LWAVE=.FALSE.
LCHARG=.FALSE.

KPOINTS:

Code: Select all

pymatgen v2020.8.13 with grid density = 1277 / number of atoms
0
Gamma
4 4 4

This uses the SCDM method:
wiki/index.php/Category:Wannier_Functions
Try it yourself (first at the DFT level only) and let me know how it works for you.
You might have to remove the previous wannier90.win file.

[wwlphys]

Hi Henrique,
It works. Thank you!