GW bandstrcutre and GW corrected band gaps

[ljzhou86]

I have finished the calculation based GW approximation, then I use nypp -s b -V qp to generate ypp.in and to get the bandstructure, however, some err occur:

[ERROR] STOP signal received while in :[04.05] Energies [ev] & Occupations
[ERROR]Set BANDS_steps > 0

what's the mean of the parameter "BANDS_steps "? And how to set it?

When I set BANDS_steps = 1, then ,another err come out as follows:

[Interp] Max error = 0.00000197 average error = 0.00000020forrtl: severe (66): output statement overflows record, unit -5, file Internal Formatted Write
Image PC Routine Line Source
ypp 00000000006B628A Unknown Unknown Unknown
ypp 00000000006B4D86 Unknown Unknown Unknown
ypp 0000000000672720 Unknown Unknown Unknown:

What's the reason and how to resolve it?


In addition, how to get the new GW corrected direct and indirect band gaps rapidly without ploting bandstructure after a GW calculation? We know the r-* file only contains the KS gaps instead of GW gaps.

[Daniele Varsano]

Dear ljzhou86,
let me ask you to fill your signature in your profile, with your affiliation. This is a rule of the forum.
About ypp, people that code it will answer you soon. Anyway, you can see your QP energies both at the end of the report file and in the output file o.qp. If you do not have the output file it means that something went wrong in your calculations.

Best,

Daniele

[ljzhou86]

Dear ljzhou86,
let me ask you to fill your signature in your profile, with your affiliation. This is a rule of the forum.
About ypp, people that code it will answer you soon. Anyway, you can see your QP energies both at the end of the report file and in the output file o.qp. If you do not have the output file it means that something went wrong in your calculations.

Best,

Daniele

I have finished such a calculation and obtained the reprot and o.qp file, however, these files donot contain the GW corrected band gaps directle and consequently I have to plot the bandstructure to get the GW band gap. It is inconvenient if I only want to get the GW corrected band gap.

[claudio]

Dear ljzhou86

we corrected the bug in the interpolation in the last version of yambo, please download it from subversion.

Regarding the indirect/direct gap in GW you needs to know which are the k-points responsable for both the gap,
and then calculate the GW correction in these points.

Notice that you can interpolate the GW band structure

http://www.yambo-code.org/input_file/ya ... _bands.php

or calculate the GW correction also in points that are note present in the initial k-points grid

http://www.yambo-code.org/input_file/ya ... ypp_kk.php

regards
Claudio Attaccalite

[ljzhou86]

Why the GW corrected band structure of 2D semimetallic silicene shows a big band gap(see in the attched figure)?

The ypp.in is as followed:
bnds # [R] Bands
electrons # [R] Electrons (and holes)
cooIn= "rlu" # Points coordinates (in) cc/rlu/iku/alat
% BKpts
0.00 | 0.00 | 0.00 | # Bands circuit
0.33333333333| 0.333333333 | 0.00 | # Bands circuit
0.00 | 0.50 | 0.00 | # Bands circuit
0.00 | 0.00 | 0.00 | # Bands circuit
%
BANDS_steps= 30 # Number of divisions
% BANDS_range
1 | 8 | # Bands Range
%
% INTERPGrid
-1 |-1 |-1 | # Interpolation Grid
%
GfnQPdb= "E < ./SAVE/ndb.QP"

The yambo.in by " yambo -p p -g n" is as followed:
ppa # [R Xp] Plasmon Pole Approximation
gw0 # [R GW] GoWo Quasiparticle energy levels
HF_and_locXC # [R XX] Hartree-Fock Self-energy and Vxc
em1d # [R Xd] Dynamical Inverse Dielectric Matrix
EXXRLvcs= 7271 RL # [XX] Exchange RL components
Chimod= "Hartree" # [X] IP/Hartree/ALDA/LRC/BSfxc
% QpntsRXp
1 | 73 | # [Xp] Transferred momenta
%
% BndsRnXp
1 | 20 | # [Xp] Polarization function bands
%
NGsBlkXp= 1 RL # [Xp] Response block size
% LongDrXp
1.000000 | 0.000000 | 0.000000 | # [Xp] [cc] Electric Field
%
PPAPntXp= 27.21138 eV # [Xp] PPA imaginary energy
% GbndRnge
1 | 20 | # [GW] G[W] bands range
%
GDamping= 0.10000 eV # [GW] G[W] damping
dScStep= 0.10000 eV # [GW] Energy step to evalute Z factors
DysSolver= "n" # [GW] Dyson Equation solver (`n`,`s`,`g`)
%QPkrange # [GW] QP generalized Kpoint/Band indices
1| 73| 1| 20|
%
%QPerange # [GW] QP generalized Kpoint/Energy indices
1| 73| 0.0|-1.0|
%

[Daniele Varsano]

Dear Zhou,
I do not know if this is the problem, anyway

NGsBlkXp= 1 RL # [Xp] Response block size

considering only one G vector in the screening you are practically neglecting the correlation part of the Self Energy in the GW correction and considering only the HF part. Let me suggest you to take your time and go through the GW tutorial before approaching production run.

Best,

Daniele

[ljzhou86]

I have tried to recalculate the 2D metallic silicene via changing the variable by "NGsBlkXp= 205 RL # [Xp] Response block size", however, the problem still exist. Why?

[Daniele Varsano]

Dear Zhou,
There could many reason, and we do not have the details of your calculations.
First of all check that your LDA band structure it is in agreement with the literature i.e. you have the same direct 0 gap at K and the same gaps at gamma etc?
Next I suggest you to look to more than two bands.
If it so, are you sure your results are converged (k-points, bands in \chi, bands in G, Gvecs, intersheet vacuum etc.?).
If it so, I suggest you to look to specific k-points and bands and look to the difference contribution to the GW corrections \Sigma_x, V_xc, Sigmac (all are in the
report ) in order to see if there is some specific contribution that is giving bad results.

Cheers,

Daniele

[ljzhou86]

I have obtained the right LDA band strcutre by using another type of pseudopotential of Si. However, the GW-corrected result seem to be not OK based on two points:

1) the Dirac points of silicene donot overlap together, and the one of GW shift up about 0.4eV
2) there is a variation for the slope of this GW valuence band marked in blue ellipse

What's the reason for the above two points? Whether the K-point in the LDA calculation (18*18*1) is not larger or the ecutwfc is not high (40Ry)? Note that the intersheet distance is 30 angstom.

[Daniele Varsano]

Dear Zhou,

1) the Dirac points of silicene donot overlap together, and the one of GW shift up about 0.4eV

Absolute energies here do not have real meaning, only energy differences, so the shift of the band structure it is not a problem.

2) there is a variation for the slope of this GW valuence band marked in blue ellipse

I do not know if this is physical or an artefact, this is not a question related on the code, so it is up you to check that the calculation is at convergence,
the reason of such behavior and if it is physical or not.

Best,

Daniele