I'm a new user of Elk. In the elk manual, the part of "dft+u" lists several types of DFT+U. But I'm confused which to adopt.
Taking BiFeO3 as an example, in WIEN2K, VASP and pwscf, I usually empoly PBE (GGA) plus an effective Hubbard U = 2 eV to reproduce the electronic and magnetic structures. Then in our Elk, which type of double counting (FLL or AFM?) and dftu input parameters (Slater, Racah or Yukawa?) should I use and how to determine the values of U and J?
Besides, the manual lists several references in the part of dft+u, one of them is PRB, 52, R5467 (1995), does this paper corresbond to the type of Slater parameters?
Best regards
Fang
Last edit: Fang 2015-10-22
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Elk's implementation is somewhat more sophisticated than in many other codes. You can think of U_eff = U - J, so if you choose J = 0 you should get more or less the same results when compared to other codes. The standard way to set the parameters is to use U and J, inpdftu=1. The FLL is what you normally choose for localized electron systems.
Note however, that the +U is applied to a specific orbital, so the result will depend on the muffin-tin radii. When compared to other codes, they might use different MT or augmentation radii, so the same U value may give rise to different results for this reason.
Regards
Markus
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
If I want to use the type of LDAU in PRB, 52, R5467 (1995) for BiFeO3, how cloud I determin the input parameters? I read the example about NiO implemented in Elk, I haven't foud how to generate the resonable parameters for the calculations.
In the paper http://arxiv.org/pdf/0911.1342v1.pdf using Elk, one sentence is "In this approach the Hund’s rule exchange parameter J varies automatically between 0 and 1 eV". Does it mean that we can calculate the corresponding values of J for U in Elk?
Best regards
Fang
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear developers and users,
I'm a new user of Elk. In the elk manual, the part of "dft+u" lists several types of DFT+U. But I'm confused which to adopt.
Taking BiFeO3 as an example, in WIEN2K, VASP and pwscf, I usually empoly PBE (GGA) plus an effective Hubbard U = 2 eV to reproduce the electronic and magnetic structures. Then in our Elk, which type of double counting (FLL or AFM?) and dftu input parameters (Slater, Racah or Yukawa?) should I use and how to determine the values of U and J?
Besides, the manual lists several references in the part of dft+u, one of them is PRB, 52, R5467 (1995), does this paper corresbond to the type of Slater parameters?
Best regards
Fang
Last edit: Fang 2015-10-22
Dear Fang,
Elk's implementation is somewhat more sophisticated than in many other codes. You can think of U_eff = U - J, so if you choose J = 0 you should get more or less the same results when compared to other codes. The standard way to set the parameters is to use U and J, inpdftu=1. The FLL is what you normally choose for localized electron systems.
Note however, that the +U is applied to a specific orbital, so the result will depend on the muffin-tin radii. When compared to other codes, they might use different MT or augmentation radii, so the same U value may give rise to different results for this reason.
Regards
Markus
Dear Dr. Markus,
Thanks for your answer.
Best regards
Fang