Band Structure Calculation with HSE06 Functional

[KieranSong]

Dear All,

I am using HSE06 hybrid functional to calculate the band structure. I copied IBZKPT file as the KPOINTS file and pasted the k point path for the band structure after the IBZKPT file. The k point path for the band structure was obtained by another band structure calculation with PBE functional for the same system before. The new KPOINTS files is shown below.
Automatically generated mesh
61
Reciprocal lattice
0.00000000000000 0.00000000000000 0.00000000000000 1
0.09090909090909 0.00000000000000 0.00000000000000 2
...
-0.09090909090909 0.45454545454546 0.00000000000000 2
0.00000000 0.00000000 0.00000000 0.000
...
0.50000000 0.50000000 0.00000000 0.000

The INCAR file is shown below.
Syetem = C_based
ISTART = 1
LWAVE = .FALSE.
LCHARGE = .TRUE.
ICHARG = 11
IVDW = 12
ISPIN = 2
MAGMOM = 0.156 -0.155 -0.096 0.096 -0.156 0.155 -0.156 0.155 0.096\
-0.096 0.156 -0.155 -0.156 0.155 0.096 -0.096 0.156 -0.155\
0.156 -0.155 -0.096 0.096 -0.156 0.155
LORBIT = 11
ENCUT = 500
ISMEAR = 0
SIGMA = 0.03
PREC = Accurate
LHFCALC = .TRUE.
AEXX = 0.25
HFSCREEN = 0.2
ALGO = Damped
TIME = 0.4
PRECFOCK = Fast
ISYM = 3
LREAL = Auto
NELMIN = 5
NELM = 700
NELMDL = -10
EDIFF = 1e-5
NPAR = 32
KPAR= 16

After the NSCF calculation done, the band structure is completely wrong and the output file shows the warning information below.

-----------------------------------------------------------------------------
| |
| W W AA RRRRR N N II N N GGGG !!! |
| W W A A R R NN N II NN N G G !!! |
| W W A A R R N N N II N N N G !!! |
| W WW W AAAAAA RRRRR N N N II N N N G GGG ! |
| WW WW A A R R N NN II N NN G G |
| W W A A R R N N II N N GGGG !!! |
| |
| Your generating k-point grid is not commensurate to the symmetry |
| of the lattice. This can cause slow convergence with respect |
| to k-points for HF type calculations |
| suggested SOLUTIONS: |
| ) if not already the case, use automatic k-point generation |
| ) shift your grid to Gamma (G) (e.g. required for hex or fcc lattice) |
| |
-----------------------------------------------------------------------------

Would anyone please give me some suggestions on the solution?

Thank you very much.

Kieran

[KieranSong]

Dear All,

I upload the band structure as the attachment, which I obtained through using p4v software to read the vasprun.xml file in the non-self consistent calculation step with HSE06 hybrid functional.



Would anyone please kindly give me some suggestions to solve it?

Thank you very much in advance.

Kind regards,

Kieran

[andreas.singraber]

Dear Kieran,

can you please post a zipped archive with all the required input files (see forum guidlines)?

Thank you!

Best,
Andreas Singraber

[KieranSong]

Dear Andreas,

I compressed all input files as zip file and uploaded it as the attachment below.

input_files.zip

I copied 'single_point_step_IBZKPT' file as the KPOINTS in the NSCF calculation step; then, I pasted 'k-points in reciprocal lattice and weights: k-points along high symmetry lines' part in the 'previous_NSCF_step_with_PBE_OUTCAR' file to the end of this KPOINTS file. This 'previous_NSCF_step_with_PBE_OUTCAR' file was obtained through carrying out another previous band structure calculation with PBE functional.

Would you please have a look at them and give me some suggestions on the solutions? Thank you very much.

Kind regards,

Kieran

[marie-therese.huebsch]

Dear Kieran,

Sorry for the delay. I would like to help you figure out how to calculate the band structure for your system.

Could you please tell me along which high-symmetry points you want to plot the band structure? Unfortunately, the explicit k points you have provided are not high symmetry points. That is also why you have received the error message from VASP. A very convenient tool to figure out the k path is SeeK-path by the Materials Cloud.

Try to compute the PBE band structure for your system along the chosen k path and share the result here. You can use fewer k points, less strict convergence criteria, smaller ENCUT, etc., to make your calculation computationally cheap. This way, you can check if your workflow is correct before wasting resources on an incorrect calculation. Mind that you will need to increase these parameters again to produce publishable results.

Best regards,
Marie-Therese

[KieranSong]

Dear Marie-Therese,

Thank you for your kind suggestion.

I did the band structure calculation with the PBE functional before I carried out the same calculation with the HSE06 hybrid functional. Here is the KPOINTS file in the band structure calculation with the PBE functional for the same system.

k-points along high symmetry lines
100
Line-mode
rec
0.00 0.00 0.00 !Gamma
0.00 0.50 0.00 !Y
0.00 0.50 0.00 !Y
0.50 0.50 0.00 !S
0.50 0.50 0.00 !S
0.50 0.00 0.00 !X
0.50 0.00 0.00 !X
0.00 0.00 0.00 !Gamma
0.00 0.00 0.00 !Gamma
0.50 0.50 0.00 !S

The band structure under the calculation with the PBE functional is pasted below.

total_band.png

After this band structure calculation with the PBE functional done, I copied the 'k-points in reciprocal lattice and weights: k-points along high symmetry lines' part in the OUTCAR file and pasted it after the 'single_point_step_IBZKPT' file, which was obtained in the single point calculation with HSE06 hybrid functional, to generate the new KPOINTS file for the non-self consistent calculation with the HSE06 hybrid functional. This 'single_point_step_IBZKPT' file was uploaded as the attachment in my previous reply.

I also copied the WAVECAR file from the single point calculation with the HSE06 hybrid functional but the non-self consistent calculation with the HSE06 functional went quite slowly. The energy difference (dE) value oscillated around 0.001 eV after 72-hour running. Please, refer to my another post yesterday. (forum/viewtopic.php?f=4&t=18261)

I would really appreciate that you could give me some further suggestions on the the solution.

Kind regards,

Kieran

[marie-therese.huebsch]

Dear Kieran,

Your PBE band structure looks correct. I could reproduce the same result.

For the HSE06-band-structure calculation, try to do the following steps:

1. Set up INCAR, KPOINTS, POTCAR and POSCAR in a new directory to perform a selfconsistent HSE06 calculation. Copy WAVECAR and CHGCAR of the PBE calculation to that directory. In KPOINTS, specify a uniform k mesh, which can be auto-generated in the same way as for the PBE calculation. Run vasp_std until selfconsistency is reached.
2. Create a new directory for the HSE06-band-structure calculation and copy INCAR, POTCAR, POSCAR, WAVECAR and CHGCAR of the HSE06 calculation to that directory.
3. Add ICHARG = 11 to the INCAR file to run a non-selfconsistent calculation.
4. Create a KPOINTS file. It should be in the same format as the one attached to this post, but probably you will use more k points. The KPOINTS file can be constructed by the following steps:
   1. Copy IBZKPT of the HSE06 calculation to KPOINTS. This is a uniform k mesh needed to compute the Fock exchange.
   2. In the OUTCAR of the PBE-band-structure calculation search for the keyword KPOINTS. At the first occurrence, you will find a list of k points in reciprocal coordinates. Append the list of k points to KPOINTS and set all weights of the appended list of k points to 0! If the weights of the k points for the band structure are not zero, the Fock exchange cannot be computed correctly.
   3. Finally, update the number of k points in the second line of the KPOINTS file. For instance, if the list of k points from point b) contains 50 points, add this to the number of points given in the IBZKPT file
5. Run vasp_std.

Also, note that 100 in line two of KPOINTS in your PBE-band-structure calculation is a lot. In line mode, the second line sets the number of k points in each line segment. So in your case, that results in 500 k points.

As a first trial, you could limit the number of electronic steps in (1.) using NELM, i.e., run the HS06 calculation without reaching full convergence. Additionally, you may compute the HSE06 band structure only from Gamma to Y and compare it to the PBE band structure. And perhaps 10 k points suffice for the line segment. This will speed up your calculation tremendously.

If the suggested calculation fails, please gather all input files and OUTCAR in separate directories for each calculation (PBE, PBE band, HSE06, HSE06) and zip them all. Then, upload the zip and describe how the calculation failed.

Best regards,
Marie-Therese

[KieranSong]

Dear Marie-Therese,

Thank you very much for your help and suggestions with details.

During last few days, I have been doing the band structure with the HSE06 hybrid functional several times but the band structure is still incorrect. I zip the input files (single_point_PBE; NSCF_PBE; single_point_HSE06; NSCF_HSE06) and upload them below.

inputfile.zip

Here is the band structure calculated with the HSE06 hybrid functional.



Would you please give me more suggestions on my calculations?

Thank you again and I look forward to your reply.

Kind regards,

Kieran