[Pw_forum] Convergence or band-gap problem in orthorhombic CsPbBr3

[Federico Iori]

Hi. My two cents: 

do you have some experimental references to compare your band structure and band gap with? 

In my opinion, if you really want to use yambo, maybe better than a scissor operator (how big it should be?) try to calculate the GW correction to your eigenvalues. 

If so, I think you should read carefully the exact procedure to prepare a yambo run. 
Look at the tutorial. 
http://www.yambo-code.org/tutorials/index.php 

HTH 
Federico 

	
Federico IORI 
Marie Curie Fellow 
Laboratoire de Physique des Solides 
Bâtiment 510 - Rue André Rivière 
91400 Orsay 


----- Mensaje original -----

De: "Christoph Wolf(신소재공학과)" <chwolf at postech.ac.kr> 
Para: "pw forum" <pw_forum at pwscf.org> 
Enviados: Miércoles, 7 de Septiembre 2016 8:50:59 
Asunto: [Pw_forum] Convergence or band-gap problem in orthorhombic CsPbBr3 



Dear all! 



I am currently trying to simulate the electronic properties of CsPbBr3 in its low-temperature phase (orthorhombic; a=8.244, b=11.735, c=8.198 angstrom, according to Stoumpos; Cryst. Growth Des. 2013); My principle plan is to follow Radi et al. (arXiv:1405.1706v2), who have done this using Wien2k, muffin-tin potentials. Eventually I want to continue processing this with YAMBO (doesn’t take US PP), so I am using normcons PBE (from the standard repository). 



However, the band-gap obtained is really tiny (without SOC, SOC is known to further reduce the gap in this material); I have done the following: 



1) Used the CIF file provided by stoumpos et al to create my unit cell and double checked with xcrysden –pwi if it turns out properly 

2) Tested convergence wrt. Total energy leading to a good compromise with ecutwfc =60.0 (tested between 30 and 180 Ry) and a unshifted automatic 4x3x4 k-grid (tested up to 16x16x16). 

3) Run Relax (gamma only) – scf – bands along G-X-S-R-G 



The resulting band structure is in the pdf. The band gap at R is so tiny (0.x eV) that I am not sure if applying scissors in yambo is justified. I also tried smaller degauss and fixed occupations to not much success. If someone can spot an obvious error in my input or has any hint how I can improve my results I would be very grateful! Thank you in advance for your time and help! 



Yours, 

Chris 



PS: I have also tried a VC-relax which clearly favors a shrinking cell (the cubic phase of CsPbBr3 for example is a=5.605 A, so the Pnma cell is probably far away from 0K – how to deal with this problem when trying to do a 300K phase…?) 



My scf input: 



&control 

calculation = 'scf' 

restart_mode='from_scratch', 

prefix='cspbbr3', 

pseudo_dir = '. /pseudo/', 

outdir='./' 

wf_collect=.true. 

verbosity='high' 

/ 

&SYSTEM 

ibrav = 8 

celldm(1)= 15.5789 

celldm(2)= 1.4235 

celldm(3)= 0.9944 

nat = 5 

ntyp = 3 

ecutwfc =60.0 

occupations='smearing' 

smearing='mp' 

degauss=0.002 

force_symmorphic=.true. 

vdw_corr='grimme-d2' 

/ 

&electrons 

diago_full_acc=.true., 

conv_thr=1.0d-8 

diagonalization='david' 

electron_maxstep=300 

/ 

ATOMIC_SPECIES 

Pb 207.20000 Pb.pbe-d-hgh.UPF 

Cs 132.90500 Cs.pbe-sp-hgh.UPF 

Br 79.90400 Br.pbe-hgh.UPF 

ATOMIC_POSITIONS (crystal) 

Pb 0.500000000 0.495569542 0.500000000 

Cs 0.000000000 -0.000054173 0.000000000 

Br 0.000000000 0.496389920 0.500000000 

Br 0.500000000 0.002099521 0.500000000 

Br 0.500000000 0.496980388 0.000000000 

K_POINTS {automatic} 

4 3 4 0 0 0 





My bands input 

nbnd=200 ! include plenty of empty bands for yambo 



And 



K_POINTS {crystal_b} 

5 

0.0000000000 0.0000000000 0.0000000000 10 !G 

0.5000000000 0.0000000000 0.0000000000 10 !X 

0.5000000000 0.5000000000 0.0000000000 10 !S 

0.5000000000 0.5000000000 0.5000000000 10 !R 

0.0000000000 0.0000000000 0.0000000000 10 !G 








Christoph Wolf 




Department of Materials Science and Engineering, POSTECH 

RIST #3, 3152, San 31 Hyoja-dong, Nam-gu, Pohang, Kyungbuk, 790-784, Republic of Korea 












30POSTECH 

_______________________________________________ 
Pw_forum mailing list 
Pw_forum at pwscf.org 
http://pwscf.org/mailman/listinfo/pw_forum 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20160907/1bf46676/attachment.html>