About the band structure calculation by GW in vasp 5.2

[lxbyf]

I want do some band sturcture calculation by GW, but there is few information about this method. So, anyone can give me some suggestion? Thanks very much......

[panda]

I wrote a blog post about this that might help you:

http://www.quantumquibbles.com/blog/

It describes how to perform a GWo band structure calculation from start to finish for Si.
<span class='smallblacktext'>[ Edited Wed May 19 2010, 04:50PM ]</span>

[lxbyf]

Thanks for your reply very much.

I followed your steps, but there is something wrong in KPOINT file in the last step.

I always use a fortran program "gk.f" to obtain the Line-mode KPOINT file in band structure calculation. But when I run the last calculation, the vasp showed warning as follows:

error in IBZKPT_HF: two k-points are equivalent 1 16
this will cause problems in the HF routine

and

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)

My calculation system is wurtzite structure, can you give me some suggestion?

[Mahlaga]

Dear Panda

I'm trying to reproduce the calculation for Si as following the steps you described on http://www.quantumquibbles.com/blog/
but I get the same error:

error in IBZKPT_HF: two k-points are equivalent 1 20
this will cause problems in the HF routine

How can I solve this preoblem?

[panda]

It is very picky. I sometimes forget to change the KPOINTS file from 4x4x4 to points of high symmetry, so I have to go back and run again. In fact, I remember doing that on this calculation. Could you try running with ICHARG = 11 with the 4x4x4 mesh first, then re running it with ICHARG = 11 for the first Brillouin zone? Maybe that will correct the problem. I don't see why it would matter but have observed the same error and have followed the above advice and it corrected the problem. I think it has something to do with the symmetry not being a full star? If anyone knows for sure why this error occurs I would love to hear more! Keep me updated as to how it goes, if not you can start the calculation over and I will help with the steps if you email me.

[Mahlaga]

Thank you. I'll try it again and keep you informed about progress.

[VASP001]

Does panda's suggestion work ?

[kuh]

Dear Panda,

I followed your steps but encountered the same problem.
I then changed the KPOINTS file to just:

k-points along high symmetry lines
20 ! 20 intersections
Line-mode
rec
0 0 0 ! gamma
0.5 0.0 0 ! X

and reran the last step and got this:

Code: Select all

distr:  one band on    1 nodes,    8 groups
 vasp.5.2.2 15Apr09 complex
 POSCAR found :  1 types and    2 ions
 LDA part: xc-table for Pade appr. of Perdew

 -----------------------------------------------------------------------------
|                                                                             |
|           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) |
|                                                                             |
 -----------------------------------------------------------------------------

 found WAVECAR, reading the header
  number of k-points has changed, file:    36 present:    20
  trying to continue reading WAVECAR, but it might fail
 WARNING: stress and forces are not correct
 POSCAR, INCAR and KPOINTS ok, starting setup
 WARNING: small aliasing (wrap around) errors must be expected
 FFT: planning ...(           1 )
 reading WAVECAR
 the WAVECAR file was read sucessfully
 charge-density read from file: unknown
 calculate exact exchange contribution

Now there is no
"error in IBZKPT_HF: two k-points are equivalent" warning.
but still I can't get the band structure.

What am I doing wrong?





<span class='smallblacktext'>[ Edited Mon Jun 14 2010, 03:48PM ]</span>

[kuh]

<span class='smallblacktext'>[ Edited Mon Jun 14 2010, 03:54PM ]</span>

[panda]

You need to follow the instructions from the error in your output - try using a gamma centered grid. You get this error when you are trying to perform the band structure part of the calculation because the k-points you generated in the previous step of the calculation don't match the symmetry of the line you are specifying. What value are you using for ICHARG and LORBIT? Can you provide full INCAR for each step and details of exactly how you preformed the calculations? Because when I go through the steps I outlined in the blog post above I get the result. I have had this error before though, and it stems from how your grid was generated. Are you switching symmetry off?

[kuh]

Dear Panda,

I tried to do exactly what you described in your blog.
Here are the INCAR and KPOINTS files I used.

----- RUN 1 (DFT calc) -----
[INCAR]

Code: Select all

ISTART = 0
ISYM = 0
ICHARG = 2
ISMEAR = 0
SIGMA = 0.1
LORBIT = 11
NSW = 200
IBRION = 2
ISIF = 3
EDIFF = 0.0001
EDIFFG = -0.01

[KPOINTS]
 

Code: Select all

Monkhorst Pack
0
Gamma
4 4 4
0 0 0

----- RUN 2 (HF-type calc) -----
[INCAR]

Code: Select all

ISTART = 1
ICHARG = 2
EDIFF = 0.0001
EDIFFG = -0.01
ENCUT = 400
ENAUG = 800
LREAL = .FALSE.
LWAVE = .TRUE.
LCHARG = .TRUE.
NELM = 200
NSW = 0
IBRION = -1
LMAXMIX = 4
ISMEAR = 0
SIGMA = 0.1
NSIM = 4
IALGO = 48
ISYM = 0
LHFCALC = .TRUE.
HFSCREEN = 0.2
ALGO = Damped
TIME = 0.4
ENCUTFOCK = 0
NKRED = 2
AEXX = 0.25 

[KPOINTS] Same as above.

----- RUN 3 (calculate dielectric matrix) -----
[INCAR] (same as above except the addition of "LOPTICS = .TRUE." at the end)

Code: Select all

ISTART = 1
ICHARG = 2
EDIFF = 0.0001
EDIFFG = -0.01
ENCUT = 400
ENAUG = 800
LREAL = .FALSE.
LWAVE = .TRUE.
LCHARG = .TRUE.
NELM = 200
NSW = 0
IBRION = -1
LMAXMIX = 4
ISMEAR = 0
SIGMA = 0.1
NSIM = 4
IALGO = 48
ISYM = 0
LHFCALC = .TRUE.
HFSCREEN = 0.2
ALGO = Damped
TIME = 0.4
ENCUTFOCK = 0
NKRED = 2
AEXX = 0.25 
LOPTICS = .TRUE.

[KPOINTS] Same as above.

----- RUN 4 (G0W0 on regular mesh) -----
[INCAR]

Code: Select all

ISTART = 1
ICHARG = 11
EDIFF = 0.0001
EDIFFG = -0.01
ENCUT = 400
ENAUG = 800
LREAL = .FALSE.
LWAVE = .TRUE.
LCHARG = .TRUE.
NELM = 200
NSW = 0
IBRION = -1
LMAXMIX = 4
ISMEAR = 0
SIGMA = 0.1
NSIM = 4
IALGO = 48
ISYM = 0
LHFCALC = .TRUE.
HFSCREEN = 0.2
ALGO = GW0
TIME = 0.4
ENCUTFOCK = 0
NKRED = 2
AEXX = 0.25 
LOPTICS = .TRUE.
NOMEGA = 50
ENCUTGW = 100
LORBIT = 11

[KPOINTS] Same as above.

----- RUN 5 (G0W0 along high-symmetry line) -----
[INCAR] Same as above.
[KPOINTS]

Code: Select all

k-points along high symmetry lines
20 ! 20 intersections
Line-mode
rec
0 0 0 ! gamma
0.5 0.0 0 ! X

The KPOINTS file in the last step is somewhat
simpler than that in your blog. But using your
KPOINTS didn't solve the problem. (The same error
occurred as Mahlaga and lxbyf reported.)

Thank you in advance for your help.



<span class='smallblacktext'>[ Edited Sun Jun 20 2010, 10:22AM ]</span>

[panda]

I don't know, everything looks fine. It worked for me when I did it, don't know what to tell you. Let me look at it more and I will get back to you, but just glancing over your input everything looks in order.

[kuh]

Thank you for your reply.
Seems like I'm at a dead end.

Let me know if there's any new finding.

[monazam]

I followed the procedure mentioned here for Si but at the end it comes out with this error message:

internal ERROR in RE_READ_KPOINTS_RD: the new k-point set for the reduced symmetry case does not contain all original k-points. Try to switch off symmetry

considering that ISYM=0, what does it mean to switch symmetry off.
I would appreciate if somebody could help.

[gloomy2004]

why not 2 steps? one is DFT, the second is GW. why HSE06 stay here? for what?
<span class='smallblacktext'>[ Edited Mon Jan 30 2012, 07:36PM ]</span>