Electronic Band Gap

Run-time issues concerning Yambo that are not covered in the above forums.

Moderators: andrea marini, Daniele Varsano, myrta gruning, Conor Hogan

martinspenke
Posts: 149
Joined: Tue Apr 08, 2014 6:05 am

Electronic Band Gap

Post by martinspenke » Tue Dec 30, 2014 3:05 am

Dear Daniele,

Is there any way to calculate the ELECTRONIC band gap from
the experimentally measured optical absorption spectrum ?

Bests
Martin
Martin Spenke, PhD Student
Theoretisch-Physikalisches Institut
Universität Hamburg, Germany

User avatar
Daniele Varsano
Posts: 2723
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Electronic Band Gap

Post by Daniele Varsano » Tue Dec 30, 2014 8:55 am

Dear Martin,
I would say no if excitonic effects are present. Usually electronic band gap is measured by direct and inverse photoemission experiments.
Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

martinspenke
Posts: 149
Joined: Tue Apr 08, 2014 6:05 am

Re: Electronic Band Gap

Post by martinspenke » Tue Dec 30, 2014 5:41 pm

Dear Daniele,

I simply wonder how these people made it in this paper :

L. R. Painter, R. N. Hamm, E. T. Arakawa and R. D. Birkhoff,
Phys. Rev. Lett., 1968, 21, 282–284.

to calculate the electronic band gap from spectrum
which is clearly affected by excitonic effects !!!
Unfortunately they do not say how they did it.

Best
Martin
Martin Spenke, PhD Student
Theoretisch-Physikalisches Institut
Universität Hamburg, Germany

User avatar
Daniele Varsano
Posts: 2723
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Electronic Band Gap

Post by Daniele Varsano » Tue Dec 30, 2014 6:01 pm

Dear Martin,
it is not clear to me neither how do they estimate the band gap.
Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Khaled
Posts: 21
Joined: Fri Apr 10, 2015 10:11 am
Location: South Africa

Re: Electronic Band Gap

Post by Khaled » Wed Apr 22, 2015 9:57 pm

Respected Dr. Daniele Varsano,

I am doing calculation of Silicon and GaAs Band gap in G0W0 within PPA and I am comparing the results with the results are came out from the following paper "http://arxiv.org/pdf/cond-mat/0604447.pdf " which are including G0W0 band gap and their experimental values for Si and GaAs (you will find it in table 4).

For Silicon I think is fine, I did it. but for GaAs I faced some problems related to the band gap, first of all, I made calculation for the QP correction from Band indices 9 to 10(because the max-level in valence band is 9) by using the parameter QPKrange. when I calculated the Band gap (Min- conduction Energy - max-Valence energy) from the output file o, always I found minimum conduction energy (at band indices 10 ) greater than the maximum energy (at the band indices 9).

when I searched about Ga element, I found it was a metal. I thought that problem because I didn't include smearing in DFT calculation, then I did it and I found the problem still the same. I enclosed with this message the calculation with smearing and without smearing.

I am waiting for your valuable respond. Thank you in advance.

Best regards.
You do not have the required permissions to view the files attached to this post.
Khaled Shehata,
PhD Student,
School of Physics,
Witwatersrand University

User avatar
Daniele Varsano
Posts: 2723
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Electronic Band Gap

Post by Daniele Varsano » Thu Apr 23, 2015 8:21 am

Dear Khaled,
I'm not really sure I've understood your problem.
Anyway looking at your files:
please note in the report:

Code: Select all

 States summary         : Full        Metallic    Empty
                           0001-0004               0005-0070
  Indirect Gaps      [ev]: 1.038438  3.881507
  Direct Gaps        [ev]: 1.084434  6.769703
You have 4 occupied bands !! So I do not know why are you looking at bands 9-10.
Beside that looking at your input in general:

Code: Select all

 NGsBlkXp= 1            RL    # [Xp] Response block size
You are including a 1G vector only in the screening, this is surely underconverged!!
Also the number of bands (10) could be not enough, I invite you to have a look to the tutorials concerning convergence issue.
About Ge, this is an (in)famous well know case where simple Kohn-Sham gives a metallic behavior instead of semiconductor, you can check the literature about that.
Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Khaled
Posts: 21
Joined: Fri Apr 10, 2015 10:11 am
Location: South Africa

Re: Electronic Band Gap

Post by Khaled » Fri Apr 24, 2015 11:25 am

Respected Dr. Daniele Varsano,

Thank you a lot for your help. First of all, I did convergence calculation for Si, but let me first show you, the way, I am using to calculate the band gap for that example.
from the beginning, I know the value of Maximum valence band level from pseudo potential file "from the value of z_valence=4 multiplied by number of atom "nat" , then the result divided by 2" for Si max.valence level is 4 and for GaAs is 9 , according to these information I specified the value of QPKrange from 4 | 5 in case of Si and 9 |10 for GaAs. then from the output file "o.file " I used the following command " awk '{if ($1!="#" && $2==4.0) print $3+$4 }' o.file name | sort " , to collect the max. value of valence band (band 4) which is corresponding to( E0 + E0+E ). and I collected the minimum value of conduction Band (Band 5). from that information I can get the value of Band gap. I hope you get me. Is that true ?, if not how can I calculate it from that side.

Second. Something make me confused a lot with respect to dealing with the following information from r-file, when I made comparison in Si bulk tuotrials files between files "GoWo_PPA_gap_kpt.dat" and "r-02Cohsex_em1s_HF_and_locXC_gw0_cohsex" which are prepared for drawing Direct and in direct band gap Vs Kpoints. I found different information of Direct and Indirect gaps wrote down in ". dat" file. my question, why I have two values of both indirect and direct gaps and how can I extract the actual value of both.

States summary : Full Metallic Empty
0001-0004 0005-0050
Indirect Gaps [ev]: 0.551130 4.540434
Direct Gaps [ev]: 2.576077 8.242179

I am very sorry if my questions are stupid questions. Thank you in advance Sir. I appreciate your help.

Best regards,

Khaled.
Khaled Shehata,
PhD Student,
School of Physics,
Witwatersrand University

User avatar
Daniele Varsano
Posts: 2723
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Electronic Band Gap

Post by Daniele Varsano » Fri Apr 24, 2015 1:47 pm

Dear Khaled,
I know the value of Maximum valence band level from pseudo potential file
The total number of electrons are also reported in the QE output and r_setup of Yambo.
for Si max.valence level is 4 and for GaAs is 9
OK, but in the file you posted you had 4 occupied bands in the report, and in the output you calculated bands 9 and 10.
from the output file "o.file " I used the following command " awk '{if ($1!="#" && $2==4.0) print $3+$4 }' o.file name | sort "
Ok, not following the awk script, anyway in the output and report you have all the needed information (E0 and Delta E), you can write script, plot the values, or inspect the output/report.

I found different information of Direct and Indirect gaps wrote down in ". dat" file.
States summary : Full Metallic Empty
0001-0004 0005-0050
Indirect Gaps [ev]: 0.551130 4.540434
Direct Gaps [ev]: 2.576077 8.242179
Careful, this information in the report are the values of the Kohn-Sham calculations, so from E0 without applying corrections. The E0,DeltaE number reported in the output and report are the same. In the report you have also the different contribution <Sigma_x>, <Sigma_c> etc..

why I have two values of both indirect and direct gaps
The first value is the gap (direct or indirect), the second value is the max energy different for the same bands. You can safely ignore the second value.

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

haseebphysics1
Posts: 148
Joined: Sat Aug 17, 2019 2:48 pm

Re: Electronic Band Gap

Post by haseebphysics1 » Fri Feb 19, 2021 6:38 pm

Dear all,

I want to take the suggestion:

1: If PBE is underestimating the electronic gap from "1.6" to "0" (the metallic state from PBE). How one can do the excited state calculations on that system using Yambo? Will usual GW with PPA or full GW make sense? or should I first try to open the gap using Hybrid functionals before using Yambo?

1: If we can run the Yambo on top of PBE in this case, would I need any extra care in the input files? I mean "Drude frequency" or something like this to be set?

Many thanks,
Haseeb.
Haseeb Ahmad
MS - Physics,
LUMS - Pakistan

User avatar
Daniele Varsano
Posts: 2723
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Electronic Band Gap

Post by Daniele Varsano » Fri Feb 19, 2021 6:45 pm

Dear Haseeb,

1) It depends on the origin of the bad description of the gap, i.e. of the problem you want to study.
Here, it does not seems an underestimation, but a qualitative different ground state. In this case I would no do a perturbation theory on top of a qualitatively wrong result. In this sense it would be safer to open the gap with another approximation (e.g. hybrids). If instead, you have at least a very small gap in the sense that occupations are correct, or a defect state placed in conduction instead that in the gap, maybe GW can put things in order, if you are lucky.

2) I would say no, you do not want include metallic screening for a system that should not have it. Anyway, Drude term can be applied only with a full frequency treatment.

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Post Reply