| Re: Band structure for SnS2 including Hubbard ( No.1 )|
- Date: 2023/02/01 21:46
- Name: Aleksey Baglov
A few thoughts.
1. Basis set can be reduced almost without accuracy loss. Try standard or even quick choices.
S S7.0-s2p2d1 S_PBE19
Sn Sn7.0-s3p2d1 Sn_PBE19
S S7.0-s2p2d1f1 S_PBE19
Sn Sn7.0-s3p2d2 Sn_PBE19
2. If you have two or more PAOs on orbital use you need to be careful with Hubbard’s parameter. Use it for the first value of PAOs set. Example: basis s2p2d1, Hubbard 1s 0.0 2s 0.0 1p 5.0 2p 0.0 1d 0.0. In general, you need to know, how to influence each PAO from a basis set on the band structure. You can combine different values for different PAO: 1p 3.3 2p 5.1 3p 0.1, but you need to know, why you do it.
3. Hubbard’s parameter values from calculations within the plane wave basis set (VASP, Quantum ESPRESSO) are not equivalence for the localized basis set (OpenMX). From my experience, the OpenMX should be used a smaller value of Hubbard’s parameter for reproducing plane wave codes results.
| Re: Band structure for SnS2 including Hubbard ( No.2 )|
- Date: 2023/02/06 16:08
- Name: Ziba Torkashvand <email@example.com>
- Thank you Dr. Baglov for your explanation.
Actually, I have tried so many sets of values but unfortunately, for large numbers, it can not be converged and for lower ones, there is no difference in bandgap with respect to without Hubbard.