the convergence problem between k-point and BLongDir (111)

Concerns issues with computing quasiparticle corrections to the DFT eigenvalues - i.e., the self-energy within the GW approximation (-g n), or considering the Hartree-Fock exchange only (-x)

Moderators: Davide Sangalli, andrea.ferretti, myrta gruning, andrea marini, Daniele Varsano

Post Reply
baran1386
Posts: 2
Joined: Wed Jul 05, 2017 6:39 am

the convergence problem between k-point and BLongDir (111)

Post by baran1386 » Tue Aug 01, 2017 9:11 am

Hi all,
I carry out the calculation of the absorption spectra for bulk TiO2 using BSE+GW. In the last step when I increase the k-points mesh from 8×8×8 to 9×9×9, the difference between these two k-points mesh at absorption spectra for BLongDir (111) is strange. While, for BLongDir (100), (010) and (001) reach to reasonable convergence. I do not understand why it happens just for BLongDir (111).
Best Regards,
Elham Goliaei
**********************
PhD student
Theoretical Physical Chemistry Group
Department of Chemistry
K. N. Toosi University of Technology
Tehran, Iran

User avatar
myrta gruning
Posts: 239
Joined: Tue Mar 17, 2009 11:38 am
Contact:

Re: the convergence problem between k-point and BLongDir (11

Post by myrta gruning » Thu Aug 03, 2017 10:56 am

Hallo Elham,

it is difficult to draw any conclusion from the information you provided. It can be everything.
What do you mean by strange?
It is true that being 1,1,1 an average of all the directions (001 010 and 100), one would expect that the convergence properties should follow the same trend.
The advice I can give is to inspect
-the input to check the consistency of variables with other calculations
-the log and the output to check if the calculation was completed, ended without errors and that the correct databases were read etc. Note that for example when you use the Haydock iterative solution of the BSE, the spectrum may be printed out, but not converged (for several reasons). You can check this by comparing column 2 with column 6 of the o.eps* file.
Further, what direction did you choose for the screening (LongDrXs)?

Best,
Myrta
Dr Myrta Grüning
School of Mathematics and Physics
Queen's University Belfast - Northern Ireland

http://www.researcherid.com/rid/B-1515-2009

baran1386
Posts: 2
Joined: Wed Jul 05, 2017 6:39 am

Re: the convergence problem between k-point and BLongDir (11

Post by baran1386 » Thu Aug 03, 2017 5:01 pm

Dear Myrta,
Thanks for your reply. My o.eps outputs have 3 columns E/ev[1], eps /Im[2] and eps /Re[3].
When I have used 111 for LongDrXs, then I put 111 for BLongDir, and when I have used 100 for LongDrXs, then I put 100 for BLongDir. I mean, it depends on what direction I choose for calculating spectra.
My inputs are shown here, this might help. It should be mentioned that my unitcell has 6 atoms.
Best Regards,
Elham
**********************
PhD student
Theoretical Physical Chemistry Group
Department of Chemistry
K. N. Toosi University of Technology
Tehran, Iran

01_gw_xyz.in
#
# ** ** ** **** **** ****** *******
# //** ** **** /**/** **/** /*////** **/////**
# //**** **//** /**//** ** /** /* /** ** //**
# //** ** //** /** //*** /** /****** /** /**
# /** **********/** //* /** /*//// **/** /**
# /** /**//////**/** / /** /* /**//** **
# /** /** /**/** /** /******* //*******
# // // // // // /////// ///////
#
# GPL Version 4.0.2 Revision 90
# MPI Build
# http://www.yambo-code.org
#
gw0 # [R GW] GoWo Quasiparticle energy levels
ppa # [R Xp] Plasmon Pole Approximation
HF_and_locXC # [R XX] Hartree-Fock Self-energy and Vxc
X_all_q_CPU= "6 6 1 1" # [PARALLEL] CPUs for each role
X_all_q_ROLEs= "q k c v" # [PARALLEL] CPUs roles (q,k,c,v)
X_all_q_CPU= "6 6 1 1" # [PARALLEL] CPUs for each role
X_all_q_ROLEs= "q k c v" # [PARALLEL] CPUs roles (q,k,c,v)
X_all_q_nCPU_invert=0 # [PARALLEL] CPUs for matrix inversion
SE_CPU= "6 6 1" # [PARALLEL] CPUs for each role
SE_ROLEs= "q qp b" # [PARALLEL] CPUs roles (q,qp,b)
em1d # [R Xd] Dynamical Inverse Dielectric Matrix
EXXRLvcs= 1500 RL # [XX] Exchange RL components
Chimod= "" # [X] IP/Hartree/ALDA/LRC/BSfxc
% BndsRnXp
1 | 32 | # [Xp] Polarization function bands
%
NGsBlkXp= 3 RL # [Xp] Response block size
% LongDrXp
1.000000 | 1.000000 | 1.000000 | # [Xp] [cc] Electric Field
%
PPAPntXp= 27.21138 eV # [Xp] PPA imaginary energy
% GbndRnge
1 | 32 | # [GW] G[W] bands range
%
GDamping= 0.10000 eV # [GW] G[W] damping
dScStep= 0.10000 eV # [GW] Energy step to evaluate Z factors
GTermKind= "none" # [GW] GW terminator ("none","BG" Bruneval-Gonze,"BRS" Berger-Reining-Sottile)
DysSolver= "n" # [GW] Dyson Equation solver ("n","s","g")
%QPkrange # [GW] QP generalized Kpoint/Band indices
1| 75| 15| 19|
%
02_screening.in
#
# ::: ::: ::: :::: :::: ::::::::: ::::::::
# :+: :+: :+: :+: +:+:+: :+:+:+ :+: :+: :+: :+:
# +:+ +:+ +:+ +:+ +:+ +:+:+ +:+ +:+ +:+ +:+ +:+
# +#++: +#++:++#++: +#+ +:+ +#+ +#++:++#+ +#+ +:+
# +#+ +#+ +#+ +#+ +#+ +#+ +#+ +#+ +#+
# #+# #+# #+# #+# #+# #+# #+# #+# #+#
# ### ### ### ### ### ######### ########
#
#
# GPL Version 4.0.2 Revision 90
# MPI Build
# http://www.yambo-code.org
#
em1s # [R Xs] Static Inverse Dielectric Matrix
X_all_q_CPU= "6 6 1 1" # [PARALLEL] CPUs for each role
X_all_q_ROLEs= "q k c v" # [PARALLEL] CPUs roles (q,k,c,v)
Chimod= "hartree" # [X] IP/Hartree/ALDA/LRC/BSfxc
ElecTemp= 0.272114 eV # Electronic Temperature
BSSmod= "d" # [BSS] (h)aydock/(d)iagonalization/(i)nversion/(t)ddft
BSENGexx= 9871 RL # [BSK] Exchange components
BSENGBlk= 100 RL # [BSK] Screened interaction block size
% BndsRnXs
1 | 32 | # [Xs] Polarization function bands
%
NGsBlkXs= 3 RL # [Xs] Response block size
% DmRngeXs
0.10000 | 0.10000 | eV # [Xs] Damping range
%
% LongDrXs
1.000000 | 1.000000 | 1.000000 | # [Xs] [cc] Electric Field
%

03_bse+gw.in

#
# /$$ /$$ /$$$$$$ /$$ /$$ /$$$$$$$ /$$$$$$
# | $$ /$$//$$__ $$| $$$ /$$$| $$__ $$ /$$__ $$
# \ $$ /$$/| $$ \ $$| $$$$ /$$$$| $$ \ $$| $$ \ $$
# \ $$$$/ | $$$$$$$$| $$ $$/$$ $$| $$$$$$$ | $$ | $$
# \ $$/ | $$__ $$| $$ $$$| $$| $$__ $$| $$ | $$
# | $$ | $$ | $$| $$\ $ | $$| $$ \ $$| $$ | $$
# | $$ | $$ | $$| $$ \/ | $$| $$$$$$$/| $$$$$$/
# |__/ |__/ |__/|__/ |__/|_______/ \______/
#
# GPL Version 4.0.2 Revision 90
# MPI Build
# http://www.yambo-code.org
#
optics # [R OPT] Optics
bse # [R BSE] Bethe Salpeter Equation.
X_all_q_CPU= "6 2 1 1" # [PARALLEL] CPUs for each role
X_all_q_ROLEs= "q k c v" # [PARALLEL] CPUs roles (q,k,c,v)
X_all_q_CPU= "6 2 1 1" # [PARALLEL] CPUs for each role
X_all_q_ROLEs= "q k c v" # [PARALLEL] CPUs roles (q,k,c,v)
X_all_q_nCPU_invert=0 # [PARALLEL] CPUs for matrix inversion
SE_CPU= "6 2 1" # [PARALLEL] CPUs for each role
SE_ROLEs= "q qp b" # [PARALLEL] CPUs roles (q,qp,b)
BSEmod= "causal" # [BSE] resonant/causal/x_coupling/xc_coupling
BSKmod= "IP" # [BSE] IP/Hartree/HF/ALDA/SEX
% BEnRange
2.00000 | 8.00000 | eV # [BSS] Energy range
%
% BDmRange
0.10000 | 0.10000 | eV # [BSS] Damping range
%
BEnSteps= 100 # [BSS] Energy steps
% BLongDir
1.000000 | 1.000000 | 1.0000000 | # [BSS] [cc] Electric Field
%
% BSEBands
15 | 19 | # [BSK] Bands range
%

User avatar
myrta gruning
Posts: 239
Joined: Tue Mar 17, 2009 11:38 am
Contact:

Re: the convergence problem between k-point and BLongDir (11

Post by myrta gruning » Thu Aug 10, 2017 9:25 am

Hallo,

if you are sure that everything is consistent, then the only advice is to continue with the convergence study on kpts.

Two things I noticed in your input:

BSKmod= "IP" means you are not performing Bethe-Salpeter, but independent particles. See http://www.yambo-code.org/input_file/va ... BSKmod.php

3 RL for the Response block size, NGsBlkXp, looks a bit low, and if you did not, you should check if it is converged. Of course, if you performed IP calculations any value looks converged, as the screening is not used in that case. So when you changed BSKmod= "SEX" you need to study convergence of the the screening and BSE parameters.

Best,
Myrta
Dr Myrta Grüning
School of Mathematics and Physics
Queen's University Belfast - Northern Ireland

http://www.researcherid.com/rid/B-1515-2009

Post Reply