Gw gap

[halima]

dear All
i have tested the Gw correction in the GaN compound which has Eg (experimental)=3.4 Ev , i found the value :
Indirect Gaps [ev]: 3.212099 6.227104
Direct Gaps [ev]: 4.24827 10.46290
is this correct?and is there a solution to make the value close to the experimental one
thanks for your help
H.ZAARI
PhD Student in laboratory of magnetism and physics of high energy
Faculty of Sciences in Rabat - Morocco
Email: halimazaari@gmail.com

[Daniele Varsano]

Dear Halima,
1) I never did calculations on GaN, but I think it should have a direct gap.
2) From the format you report the gap it seems you copied it from the report, and this is the DFT gap, not the GW.
3) Contrary if this the gap you calculated from the o.qp file, I cannot say anything about the quality of your calculations, i.e. if you checked all the convergences (please look at the tutorial, in GW calculation there are many variable to bring to convergence), as you did not say anything on the procedure you followed and what are the final parameters you used.
4) GW is rather good approximation for GaN, so you should check first if your DFT band structure is ok (there are enormous amount of DFT calculation in the literature you can compare), and then you can compare with other GW calculation (there also many).

Best,
Daniele

[halima]

Dear all
it seam that i have probelm and i need some clarification
in the Quantum espresso i dit the scf calculation ; nscf and dos so i get the value Egap=2.49 eV from the dos value underestimated the experimental one is 3.49 eV
my objectif is to correct this value .
in the yambo code i follow these step:
p2y
yambo -i
yambo
yambo -g n -p p
i get the file o.qp
in the file r-setup, i get :
Fermi Level [ev]: 15.72918
Electronic Temp. [ev K]:0.1361E-3 1.579
Bosonic Temp. [ev K]:0.1361E-3 1.579
States summary : Full Metallic Empty
0001-0008 0009-0010
Indirect Gaps [ev]: 3.244010 6.511537
Direct Gaps [ev]: 4.24553 10.45276
X BZ K-points : 32
the value of 4.24 is that calculated in DFT
in the file o.qp
I searched the value that matches the the maximum and minimum of the valence and concuction band .I was doing correspondence for the value corcted (E-E0) .i get 4.24; 3.1 eV respectively
is that correct ? and why I get a value 4.24 from DFT
H.ZAARI
PhD Student in laboratory of magnetism and physics of high energy
Faculty of Sciences in Rabat - Morocco
Email: halimazaari@gmail.com

[Daniele Varsano]

Dear Halima,
the band structure of qe (nscf run) and r_setup have to be exactly the same, and you should find the same gap. If this is not the case please post the nscf output and the r_setup file.
Once this is clarified please note that your GW calculation I think is *totally* under convergence. You are summing up 10 bands (where you have 8 occupied), ie only 2 unoccupied bands and I do suspect this is not enough, and most importantly your are considering only 1 Gvector in your screening.

Code: Select all

| NGsBlkXp= 1 

which is surely not a good approximation. Please have a look to the tutorials.
Anyway before affording the GW step, you have to be sure you have a reasonable DFT ground state.
Best,
Daniele