the formation energy defect of BaZrS3 perovskite

[moh1100]

i want to calculate the defects formation energy Ef[Xq]=Etot[Xq]−Etot[bulk]−∑iniμi+(qEVBM+Δμe)+Ecorr of the compound BaZrS3 based on the article enclosed, first of all they calculated We first calculated the formation energies for charge-neutral defects as a function of chemical potential, and then, the formation energy versus the fermi level. I've read the tutorial http://docs.quantumwise.com/tutorials/charged_defects.html but i couldn't understand well. Can you please explain in details please the steps that i can follow to calculate each term, especially that the compound has 12 defects: three vacancies (V (Ba) , V (Zr) , V (S) ), three interstitials (Bai , Zri , Si ), two cation substitutions (Ba (Zr) , Zr (Ba) ), and four antisite substitutions (Ba (S) , Zr (S) , S (Ba) , and S (Zr) )

. thank you in advance, Regards,

[zh]

The tutorial has already given the detailed steps for the calculation of defect formation energy.
Have you go through the tutorial following the steps one by one to an exercise?

What you need to do is replace the GaAs compound to your case BaZrS3. The difference is that you want to consider 12 types of defect. At least, you can follow the tutorial to calculate the defect formation energy of vacancy defects such as V_Ba, V_Zr, and V_S for your case. The difficulty may be the choice of chemical potentials for the Ba, Zr, and S, which are limited in a range because of thermodynamic conditions.  You have to determine the range of chemical potential of Ba, Zr, and S.

[moh1100]

Dear Zh,

thank you for your reply,
i will follow the steps to calculate the  vacancies formation energie. But what about the other defects, do you have an idea about the steps i should follow? and how can we determine the range of chimical potentials.

Regards,

[zh]

The interstitial defects could be done by adding one additional atom into the supercell. But for the position of added atom, you need to try different vacant space in the lattice and then find out lowest energy position.

The substitution defect can be done by replace one of lattice atoms with other different type of atom.

The remaining steps are similar to the calculations of vacancy defects.

[bina]

Dear Sir,

I have also a problem in the formation energy of defected ZnO monolayer.
Here, I have computed the formation energy of oxygen vacancy in ZnO monolayer (ZnO_Ov) and zinc vacancy in ZnO monolayer (ZnO_Znv) using the https://docs.quantumwise.com/tutorials/formation_energies/formation_energies.html tutorial. and found that ZnO_Ov has more formation energy than ZnO_Znv that is absolutely wrong.
Computational details which i have used are
GGA-PBE with double zeta double polarized and 6x6x1 k points
force convergence criterial = 0.001

result i obtained
ZnO (Pristine) = 31662.38 eV
ZnO_Ov = 31226.73 eV
ZnO_Znv = 30113.38 eV
Zn = 1541.73 eV
O = 427.81 eV

Ef (ZnO_Ov) = 7.83 eV
Ef (ZnO_Znv) = 7.26 eV

[Jess Wellendorff]

1) Those formations energies are very close to each other in relative size. How big of a difference do you expect to find?
2) Consider using a better pseudopotential + basis set, i.e. SG15, perhaps that helps.

[bina]

Dear Sir,
Formation energy of ZnO_Ov should be smaller than ZnO_Znv but in my case i found the opposite result means formation energy of ZnO_Ov larger than ZnO_Znv that is absolutely wrong.
you can read this paper http://www.sciencedirect.com/science/article/pii/S1567173912004889
this paper did the whole calculation using Wien2K and found that the formation energy ZnO_Ov < ZnO_Znv.

[bina]

Dear ATK developers,
you have any explanation about my  previous question.

Thank you

[Jess Wellendorff]

Did you thoroughly check your computational setup, i.e. structures, k-point sampling, pseudopotentials, basis set, grid cutoff energies, etc. ?

[bina]

Yes I have checked all the steps and vary the basis sets and exchange correlation potentials but found the same trend. that is why i have doubt in this calculation

[Petr Khomyakov]

You may consider having a ghost atom at the vacancy site, see the following tutorial on the use of ghost atoms for work function calculations, https://docs.quantumwise.com/tutorials/work_function_ag_100/work_function_ag_100.html, as well as the last section at https://docs.quantumwise.com/tutorials/formation_energies/formation_energies.html. In other words, you have to replace the O atom with the corresponding ghost atom to model the vacancy more accurately.

[bina]

Dear Petr,

I am computing the result by considering the ghost atom but can you give me detailed explanation about this ghost atom? what are the limitation of ghost atom? and where it is used? and also send me a link of reference paper of ghost atom.

Thank you.

[bina]

Dear Petr,

I did the calculation by considering the vacancy of atom as a ghost atom but the problem could not be resolve. you have any other suggesting to find the accurate computational result. because your tool results challenge the experimental and other computational results.
 Thank you and waiting for your kind response.

[Petr Khomyakov]

Did you investigate how the formation energy depends on the unit cell size? It might be that you are using too small cell.

[bina]

I am using a 4*4*1 supercell, So i think it is not a too small supercell