Page 1 of 1

Interpretation of reduced KPOINT Set

Posted: Thu Sep 30, 2021 1:49 pm
by Vasp1a
Dear Vasp Users,

I have a question regarding the reduced KPOINT Set generated by VASP.
I want to calculate the effective mass tensor using the "mstar" script, which works fine, following the tutorial. It generates the effective mass tensor at each k point specified in IBZKPT. However, I now want to know the effective mass not only at the Gamma point, but at all High Symmetry Points (HSPs). For a sufficiently dense k-point grid, each HSP point should be approximated close enough. However, I am not quite sure how to match them in the reduced form.

E.G. my KPOINTS Input (for silicon) is:
Automatic
0
Gamma
7 7 7
0 0 0

My resulting IBZKPOINT is:

Automatically generated mesh
20
Reciprocal lattice
0.00000000000000 0.00000000000000 0.00000000000000 1
0.14285714285714 0.00000000000000 0.00000000000000 8
0.28571428571429 0.00000000000000 0.00000000000000 8
0.42857142857143 0.00000000000000 0.00000000000000 8
0.14285714285714 0.14285714285714 0.00000000000000 6
0.28571428571429 0.14285714285714 0.00000000000000 24
0.42857142857143 0.14285714285714 0.00000000000000 24
-0.42857142857143 0.14285714285714 0.00000000000000 24
-0.28571428571429 0.14285714285714 0.00000000000000 24
-0.14285714285714 0.14285714285714 0.00000000000000 12
0.28571428571429 0.28571428571429 0.00000000000000 6
0.42857142857143 0.28571428571429 0.00000000000000 24
-0.42857142857143 0.28571428571429 0.00000000000000 24
-0.28571428571429 0.28571428571429 0.00000000000000 12
0.42857142857143 0.42857142857143 0.00000000000000 6
-0.42857142857143 0.42857142857143 0.00000000000000 12
0.42857142857143 0.28571428571429 0.14285714285714 24
-0.42857142857143 0.28571428571429 0.14285714285714 48
-0.42857142857143 0.42857142857143 0.14285714285714 24
-0.28571428571429 0.42857142857143 0.14285714285714 24


How do I, e.g., find out which point is closest to my L point, which I know is at 0.5, 0.5, 0.5?

Or is there maybe a smarter way of doing this, i.e. adding the high symmetry points explicitly somehow to the calculation, so that I do not even need to find the best approximation anymore?


Thanks a lot and best regards

Re: Interpretation of reduced KPOINT Set

Posted: Tue Oct 05, 2021 4:00 pm
by marie-therese.huebsch
Hi,

Thank you for bringing up this topic.

First of all, yes, there is a better way. After finding a selfconsistent solution, you can perform a non-selfconsistent calculation and compute the Kohn-Sham orbitals at any list of k points you chose explicitly. Check out the ICHARG tag to see how to perform a non-selfconsistent calculation.

Would you mind linking the tutorial you refer to in your question? Note that I am not sure if you require a uniform k mesh for your analysis.

Lastly, let me comment on the reduced k mesh. You can obtain all k points by applying the spacegroup operations of your system. This might be cumbersome, so instead, you could switch off the exploitation of symmetry in VASP using the ISYM tag and autogenerate the k mesh. Then, the IBZKPT file will contain a list of all k points (instead of only the reduced k mesh), some of which are equivalent according to the symmetry of the system. Be careful not to unintentionally run a selfconsistent calculation without the use of symmetry, as that can be computationally expensive. A single step suffices to get the IBZKPT file.

Best regards,
Marie-Therese

Re: Interpretation of reduced KPOINT Set

Posted: Fri Oct 08, 2021 11:03 am
by Vasp1a
Dear Marie-Therese,

thank you for your reply!
What I am using (or trying to at least) is this tool:

https://github.com/rubel75/mstar
https://github.com/rubel75/mstar/wiki

I am not sure wether the kp-method/Waveder file needs a whole grid in order to function properly, hence I was reluctant to use an nscf calculation with only the kpoints of interest.
What I have tried so far is to generate the kpoints from IBZKPT (as you suggested), and then restarted the calculation with the extracted kpoints and adding the points of interest at the end with a weight of Zero.
....
-0.25000000000000 0.50000000000000 0.20833333333333 24
-0.25000000000000 0.50000000000000 0.25000000000000 6
0.0 0.0 0.0 0
0.5 -0.0 0.5 0
0.5 0.25 0.75 0


This technically seems to work fine, however I am unsure wether the results are correct. The wiki only discusses silicon briefly (I could recreate the results, but it is not explaining how they know that kpoint 15 is the conduction band minimum). I therefore tried with GaP, which worked from the technical side, but the results do not really match the experimental values I could find online: 0.6m_e for the holes (should be at Gamma), 0.8 for the electrons (should be at L) (https://www.tf.uni-kiel.de/matwis/amat/ ... 2_3_1.html):

The output contains inverse effective masses (m0/m_ij*)

# band 1=xx; 2=yy; 3=zz; 4=yz; 5=xz; 6=xy :---: DOS 1/m*, Cond 1/m* -----> Energy, Occupation


# KP: 736 NBCDER: 16 NEMAX: 450 (0.0000,0.0000,0.0000) --> \Gamma-Point
1 7.816E-01 7.816E-01 7.816E-01 -5.768E-06 -7.973E-06 -1.179E-05:---: 7.816E-01, 7.816E-01 -----> -9.1784, 1.00000
2 7.816E-01 7.816E-01 7.816E-01 -2.805E-06 -3.266E-07 -3.849E-06:---: 7.816E-01, 7.816E-01 -----> -9.1784, 1.00000
3 -4.757E+00 -4.758E+00 -4.752E+00 -1.400E-03 -1.979E-03 1.728E-03:---:-4.756E+00, -4.756E+00 -----> 2.9975, 1.00000
4 -4.757E+00 -4.758E+00 -4.752E+00 -1.399E-03 -1.978E-03 1.730E-03:---:-4.756E+00, -4.756E+00 -----> 2.9975, 1.00000
5 -5.970E+00 -6.991E+00 -2.099E+00 -1.145E+00 -1.619E+00 1.367E+00:---:-3.744E+00, -5.020E+00 -----> 3.0842, 1.00000
6 -5.970E+00 -6.991E+00 -2.099E+00 -1.145E+00 -1.619E+00 1.367E+00:---:-3.744E+00, -5.020E+00 -----> 3.0842, 1.00000
7 -4.069E+00 -3.047E+00 -7.945E+00 1.146E+00 1.621E+00 -1.369E+00:---:-4.224E+00, -5.020E+00 -----> 3.0842, 1.00000
8 -4.069E+00 -3.047E+00 -7.945E+00 1.146E+00 1.621E+00 -1.369E+00:---:-4.224E+00, -5.020E+00 -----> 3.0842, 1.00000
9 9.768E+00 9.768E+00 9.768E+00 -1.379E-04 -1.963E-04 -1.077E-04:---: 9.768E+00, 9.768E+00 -----> 4.6847, 0.00000
10 9.768E+00 9.768E+00 9.768E+00 -1.353E-04 -1.892E-04 -1.017E-04:---: 9.768E+00, 9.768E+00 -----> 4.6847, 0.00000
11 1.507E+00 1.507E+00 1.507E+00 -1.887E-05 -2.664E-05 -1.409E-05:---: 1.507E+00, 1.507E+00 -----> 6.8973, 0.00000
12 1.507E+00 1.507E+00 1.507E+00 -1.860E-05 -2.633E-05 -1.329E-05:---: 1.507E+00, 1.507E+00 -----> 6.8973, 0.00000
13 1.432E+00 1.509E+00 1.127E+00 -3.022E-02 -4.261E-02 -1.085E-01:---: 1.342E+00, 1.356E+00 -----> 7.0618, 0.00000
14 1.432E+00 1.509E+00 1.127E+00 -2.984E-02 -4.234E-02 -1.084E-01:---: 1.343E+00, 1.356E+00 -----> 7.0618, 0.00000
15 1.280E+00 1.203E+00 1.585E+00 -1.379E-01 -1.950E-01 -4.952E-01:---: 1.251E+00, 1.356E+00 -----> 7.0618, 0.00000
16 3.714E+00 4.067E+00 2.312E+00 2.998E-02 4.245E-02 1.085E-01:---: 3.268E+00, 3.364E+00 -----> 7.0618, 0.00000


# KP: 741 NBCDER: 16 NEMAX: 450 (0.5000,0.5000,0.5000) --> L-Point
1 9.351E-03 3.351E-01 -1.295E+00 -7.989E-01 -1.130E+00 -4.613E-01:---:-9.982E-01, -3.169E-01 -----> -7.3374, 1.00000
2 9.478E-03 3.357E-01 -1.295E+00 -7.987E-01 -1.129E+00 -4.611E-01:---:-9.983E-01, -3.166E-01 -----> -7.3374, 1.00000
3 1.208E+00 8.300E-01 2.721E+00 9.264E-01 1.310E+00 5.348E-01:---: 9.234E-01, 1.586E+00 -----> -3.3632, 1.00000
4 1.208E+00 8.301E-01 2.721E+00 9.264E-01 1.310E+00 5.349E-01:---: 9.234E-01, 1.586E+00 -----> -3.3632, 1.00000
5 -3.252E+00 -3.795E+00 -1.088E+00 1.324E+00 1.873E+00 7.683E-01:---: 2.157E+00, -2.712E+00 -----> 1.9575, 1.00000
6 -3.252E+00 -3.795E+00 -1.088E+00 1.324E+00 1.873E+00 7.683E-01:---: 2.157E+00, -2.712E+00 -----> 1.9575, 1.00000
7 -3.361E+00 -3.915E+00 -1.138E+00 1.362E+00 1.927E+00 7.827E-01:---: 2.197E+00, -2.805E+00 -----> 2.0275, 1.00000
8 -3.361E+00 -3.915E+00 -1.138E+00 1.362E+00 1.927E+00 7.827E-01:---: 2.197E+00, -2.804E+00 -----> 2.0275, 1.00000
9 5.932E+00 6.698E+00 2.868E+00 -1.876E+00 -2.653E+00 -1.083E+00:---: 3.167E+00, 5.166E+00 -----> 4.6834, 0.00000
10 5.932E+00 6.698E+00 2.868E+00 -1.876E+00 -2.653E+00 -1.083E+00:---: 3.167E+00, 5.166E+00 -----> 4.6834, 0.00000
11 3.252E-01 3.224E-01 3.426E-01 1.349E-02 1.910E-02 4.057E-03:---: 3.294E-01, 3.301E-01 -----> 7.7973, 0.00000
12 3.253E-01 3.224E-01 3.427E-01 1.351E-02 1.913E-02 4.077E-03:---: 3.294E-01, 3.301E-01 -----> 7.7973, 0.00000
13 2.693E-01 2.497E-01 3.414E-01 4.127E-02 5.834E-02 2.764E-02:---: 2.782E-01, 2.868E-01 -----> 7.8609, 0.00000
14 2.693E-01 2.497E-01 3.414E-01 4.128E-02 5.836E-02 2.766E-02:---: 2.782E-01, 2.868E-01 -----> 7.8609, 0.00000
15 1.990E+00 2.339E+00 5.911E-01 -9.742E-01 -1.378E+00 -5.625E-01:---:-1.741E+00, 1.640E+00 -----> 11.2887, 0.00000
16 4.149E+00 4.547E+00 2.558E+00 -8.566E-01 -1.211E+00 -4.946E-01:---: 3.329E+00, 3.751E+00 -----> 11.2887, 0.00000

I would have at least expected the Gamma point to be correct, but maybe my reference for the experimental values are wrong. I e.g. found for the DOS effective mass for the gamma valley for the electrons an effective mass of 0.09m_e (http://www.ioffe.ru/SVA/NSM/Semicond/GaP/bandstr.html), which complies with my result (about 0.10 m_e).

So if someone has a reliable (possibly ab-initio) source for effective masses of a compound at the Gamma and another High-Symmetry Point, that would be very much appreciated as well.

Thanks and all the best,

Carl-Friedrich Schön

Re: Interpretation of reduced KPOINT Set

Posted: Mon Oct 11, 2021 8:23 am
by marie-therese.huebsch
Dear Carl-Friedrich,

Maybe some user has experience with this tool and can help!

Regarding the results: Of course, ultimately you want to check if the results agree with the experiment, but for confirming whether you are using the tool correctly it is better to check against other numerical DFT results. Otherwise, it is not clear whether the theory, implementation, or the usage of the tool is the source of discrepancy.

Did you check the original paper? (https://arxiv.org/abs/2007.03816) They seem to show at least 3 examples.
It also seems that mstar does not require a uniform k mesh, but do not take my word for it. VASP cannot support or give proper advice on tools provided by others. Please confirm this with the developers of mstar.

Please, also consider sharing your experience here afterward!

Best regards,
Marie-Therese

Re: Interpretation of reduced KPOINT Set

Posted: Mon Oct 11, 2021 10:24 am
by rubel
It is possible to use mstar with an arbitrary set of k-points (even a single k-point of interest).

Probably the most 'economical' way is to make (1) SCF and store the charge density, then (2) run non-SCF calculation for the k-point of interest and many bands.

Checking the results: Please also read a note about limitations (https://github.com/rubel75/mstar/wiki/G ... le-in-VASP). s-p chemical bonds is not a problem. But if you have d-electrons, I would recommend checking agains the band curvature. Also XX_GW potentials perform better for high-energy states, which is useful here too.

Best regards
Oleg

Re: Interpretation of reduced KPOINT Set

Posted: Tue Oct 12, 2021 11:34 am
by Vasp1a
Dear Marie-Therese, dear Rubel,


thank you for your suggestions.

I have looked again at Silicon example given in the original publication by Rubel. I suppose the Delta Point given in Table I is a reference point that is not the Gamma Point that could be used to check the consistency of my calculations.
The Delta Point is not given explicitely, but from a previous band structure calculation I extracted that it should be located at (0.0000,0.2406,0.1701). Correct me if I am wrong.

I have not yet implemented the smarter way you suggested yet, i.e. I have just attached the Delta point at the end of the KPOINT file after the reduced set for the scf cycle. I obtain the following results:

# band 1=xx; 2=yy; 3=zz; 4=yz; 5=xz; 6=xy :---: DOS 1/m*, Cond 1/m* -----> Energy, Occupation
# KP: 426 NBCDER: 16 NEMAX: 200 (0.0000,0.2406,0.1701) --> Delta-Point
1 8.230E-01 7.971E-01 8.422E-01 1.699E-02 -3.488E-06 -6.008E-05:---: 8.205E-01, 8.208E-01 -----> -5.4427, 1.00000
...
8 1.303E+02 1.993E+00 1.470E+00 1.173E+00 -2.400E-03 -1.818E-03:---: 5.873E+00, 4.460E+01 -----> 4.1413, 1.00000
9 5.276E+00 1.054E+00 1.645E+00 -9.191E-01 -2.925E-06 -7.080E-05:---: 1.674E+00, 2.658E+00 -----> 7.2948, 0.00000
10 5.276E+00 1.054E+00 1.645E+00 -9.191E-01 9.772E-06 1.351E-05:---: 1.674E+00, 2.658E+00 -----> 7.2948, 0.00000
...
15 7.337E-01 -1.102E+00 4.907E-02 1.999E-01 -3.269E-05 -2.201E-04:---:-4.102E-01, -1.066E-01 -----> 10.4714, 0.00000

Band 8(9) should be the one relevant for electrons.
I get for the cardinal directions effective masses of about 0.19, 0.95 and 0.6. The values of 0.19 and 0.95 seem reasonable (and are also stated in the paper in Table 1), the 0.6 however are not correct I think.

What are your opinions?

Best,

Carl-Friedrich

Re: Interpretation of reduced KPOINT Set

Posted: Wed Oct 13, 2021 11:51 am
by rubel
Hi Carl-Friedrich,

you are right, your results look strange. Here is my output (Si, CBM) and required input files. It is a simple single-run calculation. Please try to reproduce.

I hope it will help
Oleg
~~
minv_ij.dat

Code: Select all

# KP: 15 NBCDER: 16 NEMAX: 180
...
 9  5.181E+00  1.062E+00  5.181E+00 -1.453E-06 -1.321E-04 -1.445E-06
10  5.181E+00  1.062E+00  5.181E+00 -1.443E-06 -1.318E-04 -1.437E-06
...
KPOINTS file:

Code: Select all

Automatic
 0
Gamma
 7 7 7
 0 0 0
POSCAR

Code: Select all

Si experimental a0
   5.431
     0.5    0.5    0.0000000000000000
     0.0000000000000000    0.5    0.5
     0.5   0.0000000000000000    0.5
   Si
     2
Direct
  0.0000000000000000  0.0000000000000000  0.0000000000000000
  0.2500000000000000  0.2500000000000000  0.2500000000000000
  
INCAR

Code: Select all

SYSTEM=Si
ISTART=0
ICHARG=2

LOPTICS = T    # compute dipole matrix elements and write them in WAVEDER file
LPEAD = F      # do not use finite difference in k space to get derivative of the cell-periodic part of the orbitals
NBANDS = 180   # more empty bands

ALGO = Normal
EDIFF=1.0E-09

LCHARG = F     # store charge density?
LWAVE = F      # store FWs?

ISMEAR=0; SIGMA=0.01

# SOC
LNONCOLLINEAR = .TRUE.
NUPDOWN=0
LSORBIT = .TRUE.
SAXIS =   0 0 1 # (quantisation axis for spin)
GGA_COMPAT = .FALSE.
POTCAR: PAW_PBE Si 05Jan2001

Re: Interpretation of reduced KPOINT Set

Posted: Mon Oct 18, 2021 8:43 am
by Vasp1a
Dear Oleg,

thank you for the reply!
I could reproduce the result by looking at KPOINT 15 as well as inserting the coordinates of this kpoint separately to the KPOINTS file, meaning the general procedure seems to be working. I think the conduction band minimum coordinates I initally used were not correct.
This leaves me with the question if vasp is actually okay with inserting non-reduced kpoints to the reduced kpoint list generated in IBZKPT.

Would it be possible for you to check your results for the L-Point (0.5, 0.5, 0.5)? If these match with mine, I should be good.

Thanks and all the best,

Carl-Friedrich

Re: Interpretation of reduced KPOINT Set

Posted: Tue Oct 19, 2021 9:36 am
by rubel
Dear Carl-Friedrich,

I am glad it worked eventually.
This leaves me with the question if vasp is actually okay with inserting non-reduced kpoints to the reduced kpoint list generated in IBZKPT.
I do not think you can mix "Cartesian" with "reciprocal" in VASP KPOINTS file. The IBZKPT file should be in the "reciprocal" mode. If you look at the crystallographic servers (https://www.cryst.ehu.es/cgi-bin/cryst/ ... ph-kv-list) there are a "Conventional basis" (most common in literature) and also a "Primitive basis" (used in the IBZKPT file, provided you set up correctly the primitive lattice vectors).

I would recommend reading wiki/index.php/KPOINTS and also verify the Conventional <-> Primitive conversion.

Best regards
Oleg

Re: Interpretation of reduced KPOINT Set

Posted: Thu Oct 21, 2021 12:50 pm
by Vasp1a
Dear Oleg,

thanks a lot! Yes, it seems that it was a problem with mixing the kpoint conventions indeed.
I have switched to a separate nscf calculation with a kpoint file only using the reciprocal High Symmetry Points.
In case someone digs out this thread, for silicon the KPOINTS file looks like this now:

Code: Select all

Automatically generated mesh
15
Reciprocal
0.0 0.0 0.0 1
0.5 0.0 0.5 1
0.5 0.25 0.75 1
0.375 0.375 0.75 1
0.0 0.0 0.0 1
0.5 0.5 0.5 1
0.625 0.25 0.625 1
0.5 0.25 0.75 1
0.5 0.5 0.5 1
0.375 0.375 0.75 1
0.625 0.25 0.625 1
0.5 -0.0 0.5 1
0.0 0.24055 0.1701 1
0.42857142857143 0.42857142857143 0.0000000000000 1
0.413 0.413 0.0000000000000 1
the last two points are supposed to be the conduction band minimum. As thie minimum is not a HSP, I added two test points. The output for Gamma is (restructured from the raw output):

Code: Select all

# band 1=xx;     2=yy;      3=zz;     4=yz;       5=xz;      6=xy   :---: DOS 1/m*,   Cond 1/m* -----> Energy, Occupation

# KP: 1 NBCDER: 16 NEMAX: 200  (0.0000,0.0000,0.0000) --> \Gamma-Point 
 1  8.713E-01  8.713E-01  8.713E-01 -6.321E-07 -9.443E-08 -1.266E-07:---: 8.713E-01,  8.713E-01 -----> -6.0323, 1.00000
 2  8.713E-01  8.713E-01  8.713E-01  7.249E-09  4.235E-07  3.518E-07:---: 8.713E-01,  8.713E-01 -----> -6.0323, 1.00000
 3 -4.402E+00 -4.402E+00 -4.402E+00 -3.451E-04 -3.441E-04 -3.443E-04:---:-4.402E+00, -4.402E+00 ----->  5.9280, 1.00000
 4 -4.402E+00 -4.402E+00 -4.402E+00 -3.432E-04 -3.418E-04 -3.432E-04:---:-4.402E+00, -4.402E+00 ----->  5.9280, 1.00000
 5 -5.212E+00 -5.212E+00 -5.212E+00 -2.666E+00 -2.666E+00 -2.666E+00:---:-4.088E+00, -5.212E+00 ----->  5.9756, 1.00000
 6 -5.212E+00 -5.212E+00 -5.212E+00 -2.666E+00 -2.666E+00 -2.666E+00:---:-4.088E+00, -5.212E+00 ----->  5.9756, 1.00000
 7 -3.840E+00 -3.840E+00 -3.840E+00  2.667E+00  2.667E+00  2.667E+00:---: 3.984E+00, -3.840E+00 ----->  5.9756, 1.00000
 8 -3.840E+00 -3.840E+00 -3.840E+00  2.667E+00  2.667E+00  2.667E+00:---: 3.984E+00, -3.840E+00 ----->  5.9756, 1.00000
 9  2.715E+00  2.715E+00  2.715E+00  6.856E-05  6.789E-05  6.826E-05:---: 2.715E+00,  2.715E+00 ----->  8.4561, 0.00000
10  2.715E+00  2.715E+00  2.715E+00  8.218E-05  8.514E-05  8.186E-05:---: 2.715E+00,  2.715E+00 ----->  8.4561, 0.00000
11  2.882E-01  2.882E-01  2.882E-01 -2.628E-01 -2.628E-01 -2.628E-01:---:-4.162E-01,  2.882E-01 ----->  8.4907, 0.00000
12  2.882E-01  2.882E-01  2.882E-01 -2.628E-01 -2.628E-01 -2.628E-01:---:-4.161E-01,  2.882E-01 ----->  8.4907, 0.00000
13  4.960E+00  4.960E+00  4.960E+00  2.627E-01  2.627E-01  2.627E-01:---: 4.946E+00,  4.960E+00 ----->  8.4908, 0.00000
14  4.960E+00  4.960E+00  4.960E+00  2.627E-01  2.627E-01  2.627E-01:---: 4.946E+00,  4.960E+00 ----->  8.4908, 0.00000
15  5.945E+00  5.945E+00  5.945E+00 -1.074E-05 -1.311E-05 -1.137E-05:---: 5.945E+00,  5.945E+00 ----->  9.3315, 0.00000
16  6.179E+00  6.179E+00  6.179E+00 -9.833E-06 -1.061E-05 -1.048E-05:---: 6.179E+00,  6.179E+00 ----->  9.3315, 0.00000
and for the conduction band minimum:

Code: Select all

# KP: 15 NBCDER: 16 NEMAX: 200  (0.4130,0.4130,0.0000) --> CBM-Point 
 1  5.954E-01  6.864E-01  5.954E-01 -2.465E-08 -2.542E-05 -8.494E-07:---: 6.243E-01,  6.257E-01 -----> -3.1408, 1.00000
 2  5.955E-01  6.864E-01  5.955E-01  1.180E-06  5.607E-05  1.740E-07:---: 6.244E-01,  6.258E-01 -----> -3.1408, 1.00000
 3 -2.178E-01  3.856E-01 -2.181E-01 -1.794E-05 -1.715E-04 -3.165E-05:---: 2.636E-01, -1.675E-02 -----> -0.4552, 1.00000
 4 -2.172E-01  3.857E-01 -2.172E-01  6.225E-05  3.793E-04  1.085E-04:---: 2.630E-01, -1.625E-02 -----> -0.4552, 1.00000
 5 -2.464E+00  8.611E-01 -2.464E+00 -1.579E-04 -6.053E-03 -7.382E-05:---: 1.736E+00, -1.356E+00 ----->  3.1932, 1.00000
 6 -2.464E+00  8.616E-01 -2.464E+00  6.548E-05 -5.835E-03  1.540E-04:---: 1.736E+00, -1.355E+00 ----->  3.1932, 1.00000
 7 -2.320E+00  8.625E-01 -2.320E+00 -4.338E-05  5.897E-03 -8.038E-05:---: 1.668E+00, -1.259E+00 ----->  3.1957, 1.00000
 8 -2.320E+00  8.627E-01 -2.320E+00  9.025E-05  6.018E-03  2.718E-05:---: 1.668E+00, -1.259E+00 ----->  3.1957, 1.00000
 9  5.280E+00  1.036E+00  5.280E+00 -1.735E-06 -1.896E-04 -3.045E-06:---: 3.068E+00,  3.865E+00 ----->  6.4911, 0.00000
10  5.280E+00  1.036E+00  5.280E+00 -1.480E-07  3.500E-05 -1.103E-06:---: 3.069E+00,  3.866E+00 ----->  6.4911, 0.00000
11  4.294E+00  1.242E+00  4.293E+00 -8.093E-06  6.323E-05 -7.153E-06:---: 2.839E+00,  3.276E+00 ----->  7.1205, 0.00000
12  4.294E+00  1.242E+00  4.294E+00  1.090E-05  2.140E-04  2.138E-05:---: 2.839E+00,  3.276E+00 ----->  7.1205, 0.00000
13 -7.729E+00 -2.024E-02 -7.729E+00 -5.373E-05  1.197E-04 -3.901E-05:---:-1.065E+00, -5.159E+00 -----> 14.7218, 0.00000
14 -7.728E+00 -2.023E-02 -7.728E+00 -1.404E-05  4.324E-04  1.630E-06:---:-1.065E+00, -5.159E+00 -----> 14.7218, 0.00000
15 -7.704E+00 -3.210E-02 -7.704E+00 -1.543E-05 -3.526E-04 -3.676E-05:---:-1.240E+00, -5.147E+00 -----> 14.7306, 0.00000
The results are in good accordance with your results. I did not relax the structure/might have used slightly different parameters etc., hence the results are a bit different.

Thanks and all the best,

Carl-Friedrich

Re: Interpretation of reduced KPOINT Set

Posted: Mon Oct 25, 2021 2:13 pm
by rubel
Dear Carl-Friedrich,

thanks for confirming. Now your results look good, but you did not list the INCAR file used with your KPOINTS
Automatically generated mesh
15
Reciprocal
0.0 0.0 0.0 1
0.5 0.0 0.5 1
0.5 0.25 0.75 1
0.375 0.375 0.75 1
0.0 0.0 0.0 1
0.5 0.5 0.5 1
0.625 0.25 0.625 1
0.5 0.25 0.75 1
0.5 0.5 0.5 1
0.375 0.375 0.75 1
0.625 0.25 0.625 1
0.5 -0.0 0.5 1
0.0 0.24055 0.1701 1
0.42857142857143 0.42857142857143 0.0000000000000 1
0.413 0.413 0.0000000000000 1
In case you do it in a single shot (1 SCF run, e.g., with HSE06), then I would use this KPOINTS file

Code: Select all

Automatically generated mesh
      20
Reciprocal lattice
    0.00000000000000    0.00000000000000    0.00000000000000             1
    0.14285714285714    0.00000000000000    0.00000000000000             8
    0.28571428571429    0.00000000000000    0.00000000000000             8
    0.42857142857143    0.00000000000000    0.00000000000000             8
    0.14285714285714    0.14285714285714    0.00000000000000             6
    0.28571428571429    0.14285714285714    0.00000000000000            24
    0.42857142857143    0.14285714285714    0.00000000000000            24
   -0.42857142857143    0.14285714285714    0.00000000000000            24
   -0.28571428571429    0.14285714285714    0.00000000000000            24
   -0.14285714285714    0.14285714285714    0.00000000000000            12
    0.28571428571429    0.28571428571429    0.00000000000000             6
    0.42857142857143    0.28571428571429    0.00000000000000            24
   -0.42857142857143    0.28571428571429    0.00000000000000            24
   -0.28571428571429    0.28571428571429    0.00000000000000            12
    0.42857142857143    0.42857142857143    0.00000000000000             6
   -0.42857142857143    0.42857142857143    0.00000000000000            12
    0.42857142857143    0.28571428571429    0.14285714285714            24
   -0.42857142857143    0.28571428571429    0.14285714285714            48
   -0.42857142857143    0.42857142857143    0.14285714285714            24
   -0.28571428571429    0.42857142857143    0.14285714285714            24
   0.413 0.413 0.0000000000000 0
Please note that points on the mesh are properly weighted, while extra points (here it is the last one) have '0' weight.

Also, we experimented more with other materials and found that GW potentials perform better (especially for materials with significant contribution of d states at VBM/CBM, e.g., TMDs) than standard potentials meant for ground state properties. It is because the high energy states are better described in the GW set and those states are also important for mstar.

Best regards
Oleg