[salmon-user:00369] Re: Negative band gap in spin-polarized calculations

[Lee, Inho]

Dear Salmon developers and users,

I found that I used the unit of system as [unit_system = 'a.u.'] instead of [unit_system = 'A_eV_fs']

I re-calculated ground state of CrI3 with [unit_system = 'A_eV_fs'], and the result was changed. However the band gap is still negative.

Thank you,

Inho Lee

*****************************************

Inho Lee, Ph.D

Theoretical Division, T-4

Los Alamos National Laboratory

Los Alamos, New Mexico 87545

Email: inholee.at.lanl.gov

---

From: Lee, Inho
Sent: Thursday, July 3, 2025 6:12 PM
To: salmo...@salmon-tddft.jp <salmo...@salmon-tddft.jp>
Subject: Negative band gap in spin-polarized calculations

 

Dear Salmon developers and users,

I am Inho Lee, postdoc in Los Alamos National Lab.

I performed the spin-polarized LDA calculation of CrI3 monolayer, however it shows unphysical results with negative band gap.

(Moreover, DOS is not correct.)

Therefore, I tried calculation with larger r-grid, however the band gap is still negative.

I used r-grid with ~0.19 Angstrom (a,b-axis) and ~0.205 Angstrom(c-axis).

Also I used 8*8*1 kgrid.

Is r-grid with 0.2 Ang not enough? or Is there inappropriate option in my input file?

(I checked the atomic coordinate, and crystal structure is correct)

I attached input/pseudopotential/output(w/o intermediate iteration data) files.

I welcome any comments and suggestions,

Thank you,

Sincerely,

Inho Lee

*****************************************

Inho Lee, Ph.D

Theoretical Division, T-4

Los Alamos National Laboratory

Los Alamos, New Mexico 87545

Email: inholee.at.lanl.gov

[Lee, Inho]

[Shunsuke Yamada]

Dear Dr. Inho Lee, 

Sorry for the delayed response due to mail system issues. 

Thank you for updating the status. Your input file seems to be no problem.

It might be due to the pseudopotentials and nonorthogonal cell. Some pseudopotentials containing core electrons make it worse. Maybe, more softer pseudopotentials (such as FHI) are better.

Also, the r-grid convergence is sometimes very severe for nonorthogonal coordinate. The use of orthogonal cell might improve it.

I have no experience with CrI3, so I apologize for not being able to offer useful advice.

Sincerely,

Shunsuke

山田俊介
Shunsuke Yamada
量子技術基盤研究部門 関西光量子科学研究所
Kansai Institute for Photon Science, National Institutes for Quantum Science and Technology

2025/07/07 6:57、Lee, Inho <inh...@lanl.gov>のメール:

<CrI3_gs.out.txt><input_LDA.dat>