GW band interpolation

[anhhv]

Dear all,

I want to interpolate G0W0 band structure along symmetry lines using yambo 4.4.0. I did the fist step to remove time-reversal symmetry. This is the input file

Code: Select all

fixsyms                        # [R] Reduce Symmetries
#% Efield1
# 0.00     | 0.00     | 0.00     |        # First external Electric Field
#%
#RmAllSymm                     # Remove all symmetries
RmTimeRev                      # Remove Time Reversal

When I tried to run ypp, it created FixSymm folder which contains SAVE folder inside. However, SAVE is totally empty. The log file showed that "Symmetries did not change. Program terminaned."

Code: Select all

 <---> [01] CPU structure, Files & I/O Directories
 <---> [02] Y(ambo) P(ost)/(re) P(rocessor)
 <---> [03] Core DB
 <---> :: Electrons             : 100.0000
 <---> :: Temperature       [ev]: 0.000000
 <---> :: Lattice factors [a.u.]: 10.07497  10.07497  10.07497
 <---> :: K points              :  47
 <---> :: Bands                 :  672
 <---> :: Symmetries            :  48
 <---> :: RL vectors            : 276249
 <---> [04] K-point grid
 <---> :: Q-points   (IBZ):  47
 <---> :: X K-points (IBZ):  47
 <---> [05] CORE Variables Setup
 <---> [05.01] Unit cells
 <---> Axis vectors are left handed
 <---> [05.02] Symmetries
 <---> [05.03] RL shells
 <---> [05.04] K-grid lattice
 <---> [05.05] Energies [ev] & Occupations
 <---> Symmetries-respecting SAVE written to./FixSymm
 <---> BZ -> IBZ reduction |                                        | [000%] --(E) --(X)
 <---> BZ -> IBZ reduction |########################################| [100%] --(E) --(X)
 <---> Symmetries did not change. Program terminaned.
 <---> [06] Game Over & Game summary

Can you give me an advice? Many thanks.

Viet-Anh Ha

[Daniele Varsano]

Dear Viet-Anh Ha,
please include your affiliation in our signature, this is a rule of the forum. You can do once for all by editing your user profile.
In order to understand what is going on it would be useful if you could include your set report file (in order to upload you need to rename it e.g. with a suffix .txt).
However, if you are using the 4.4 release it should not be needed to remove the TR symmetry anymore.

Best,
Daniele

[claudio]

Try to do it, without removing time-reversal symmetry

best
Claudio

[anhhv]

Hi Daniele & Claudio,

Thank you for your suggestion. It works without removal of TR symmetry. I tried with DFT band structure interpolation first. This is the input file of mine

Code: Select all

electrons                      # [R] Electrons (and holes)
bnds                           # [R] Bands
INTERP_mode= "BOLTZ"           # Interpolation mode (NN=nearest point, BOLTZ=boltztrap aproach)
cooIn= "rlu"                   # Points coordinates (in) cc/rlu/iku/alat
cooOut= "rlu"                  # Points coordinates (out) cc/rlu/iku/alat
% BANDS_bands
  37 |  52 |                   # Number of bands
%
% INTERP_Grid
-1 |-1 |-1 |                             # Interpolation BZ Grid
%
INTERP_Shell_Fac= 40.00000     # The bigger it is a higher number of shells is used
CIRCUIT_E_DB_path= "none"      # SAVE obtained from the QE `bands` run (alternative to %BANDS_kpts)
BANDS_path= "G X W K G L U W L K" # BANDS path points labels (G,M,K,L...)
BANDS_steps= 30                # Number of divisions
#BANDS_built_in       # Print the bands of the generating points of the circuit using the nearest internal point
%BANDS_kpts # K points of the bands circuit
%
%BANDS_kpts                    # K points of the bands circuit
 0.000000| 0.000000| 0.000000|
 0.500000| 0.000000| 0.500000|
 0.500000| 0.250000| 0.750000|
 0.375000| 0.375000| 0.750000|
 0.000000| 0.000000| 0.000000|
 0.500000| 0.500000| 0.500000|
 0.625000| 0.250000| 0.625000|
 0.500000| 0.250000| 0.750000|
 0.500000| 0.500000| 0.500000|
 0.375000| 0.375000| 0.750000|
%

However, in the output, it seems that the coordinates of k-points given in the input file changed, therefore, I got the weird symmetry-line band structure (in comparison with DFT one). I could not uploaded the files (I do not know why) but the coordinates in the o.bands_interpolated now are

Code: Select all

0.0000        0.0000        0.0000  [G]
0.0000        1.0000        1.0000  [X]
0.5000        1.5000        1.0000  [W]
1.5000        0.7500        0.7500  [K]
0.0000        0.0000        0.0000  [G]
1.0000        1.0000        1.0000  [L]
1.5000        0.7500        0.7500  [U]
0.5000        1.5000        1.0000  [W]
1.0000        1.0000        1.0000  [L]
1.5000        0.7500        0.7500  [K]

Many thanks.[/i]

[Daniele Varsano]

Dear Viet-Anh Ha,
you can try to assign the path explcitly as you did in %BANDS_kpts namelist avoding to assign the BANDS_path variable.
Careful that in your input file %BANDS_kpts is repeated twice.

Daniele


PS: In order to upload file you need to rename them with allowed suffix (e.g. .txt)

[anhhv]

Hi Daniele,

Thank you so much. It works perfectly when I comment #BANDS_path variable. But the code does not recognize the point labels in the name list %BANDS_kpts

Code: Select all

%BANDS_kpts                    # K points of the bands circuit
 0.000000 | 0.000000 | 0.000000 | 'G'
 0.500000 | 0.000000 | 0.500000 | 'X'
 0.500000 | 0.250000 | 0.750000 | 'W'
 0.375000 | 0.375000 | 0.750000 | 'K'
 0.000000 | 0.000000 | 0.000000 | 'G'
 0.500000 | 0.500000 | 0.500000 | 'L'
 0.625000 | 0.250000 | 0.625000 | 'U'
 0.500000 | 0.250000 | 0.750000 | 'W'
 0.500000 | 0.500000 | 0.500000 | 'L'
 0.375000 | 0.375000 | 0.750000 | 'K'
%

By the way, it's not a big deal. Many thanks.

[pyadav]

Dear yambo team,

I'm trying to get a gw/dft band of a material of which the input and output files are attached. My DFT band calculated from QE is in agreement with literature but when I'm trying to interpolate the DFT bands from ypp ( created using yambo -s b (or yambo -s b -V qp) ) are not giving the same bands as the QE results. My system is a Dirac semimetal (with inversion as well as TRS). The noticeable thing is that the Dirac points are not preserved in the ypp interpolated bands. The desired k-path can be found at the bottom of the yambo band data file.

I'll really appreciate any help regarding the same.
Thanks
Pushpendra

[Daniele Varsano]

Dear Pushpendra,

the QE bands are calculated, while it seems that the Yambo interpolator is not able to catch the crossing at the Dirac point.
My suggestion is the following:
--For the KS band structure you can stay with the QE results
--If you need to look at QP bands, the interpolator will fail but you can try to interpolate QP corrections instead of QP energies as indicated in this post:
--or alternatively use a Wannier interpolator as indicated in this other post.


Best,
Daniele

[claudio]

Dear Pushpendra Yadav

yambo uses a smooth Fourier interpolation, that fails in all crossing points, including the direct points,
because derivatives are not analytic there.
The simplest solution is to use a k-point sampling that include your Dirac point in the QE calculation,
and then interpolate bands. In this case the Dirac point will be included in the interpolation

best
Claudio

[pyadav]

Dear Daniele and Claudio,

Thanks to both of you for the quick reply. I'll see how the WANNIER90 interface can be used for this purpose. I'll appreciate it if any post/webpage can be suggested for technical help to link the Yambo with the WANNIER90 package.

Wishes,
Pushpendra