spin orbit & LDA+U for f-electrons
Posted: Mon Jul 04, 2005 8:30 pm
Dear all,
I have problems to calculate spin orbit coupling with both the Dudarev and Liechtenste LSDA+U approaches for systems with f-electrons.
For example: Let's assume there two Ce3+ ion in the unit cell, which both have the electron configuration [Xe] 4f1.
When I apply both LSDA+U as well as the LDA+U approaches implemented in VASP 4.6 with a proper Hubbard U, the obtained atomic partial magnetic moments of the Ce3+ ions are around 1.0.
However, when I switch on the spin orbit coupling (LSORBIT = .TRUE.) the calculated sclar norms of the atomic partial magnetic moments are only 0.2
Provided VASP 4.6 has the capability to calculate spin orbit interactions for f-elements using DFT methods extended with Hubbard U: Which are the critical parameters one has to play with?
Thank you for your help.
Sincerely,
René Windiks
I have problems to calculate spin orbit coupling with both the Dudarev and Liechtenste LSDA+U approaches for systems with f-electrons.
For example: Let's assume there two Ce3+ ion in the unit cell, which both have the electron configuration [Xe] 4f1.
When I apply both LSDA+U as well as the LDA+U approaches implemented in VASP 4.6 with a proper Hubbard U, the obtained atomic partial magnetic moments of the Ce3+ ions are around 1.0.
However, when I switch on the spin orbit coupling (LSORBIT = .TRUE.) the calculated sclar norms of the atomic partial magnetic moments are only 0.2
Provided VASP 4.6 has the capability to calculate spin orbit interactions for f-elements using DFT methods extended with Hubbard U: Which are the critical parameters one has to play with?
Thank you for your help.
Sincerely,
René Windiks