[Wien] Ueff

[pieper]

Until one of the real experts (I am not among them) finds time to 
answer: My guess is that both is correct, Ueff=U-J as well as 
Ueff=U-J/2. That is, the meaningful thing to do is set J=0 (see UG 
section on ORB), and not estimate it from DFT.

The common ground between ORB or DFT and the magnetism of atoms as in 
Hunds rules is a multiband Hubbard model.

For the latter, the multiple correlation energy parameters of such a 
model are summarized into one intra-orbital U, one inter-orbital V, and 
a Hund's rule parameter J reducing the inter-orbital correlation energy 
for same spin direction. [1].

For the former, the parameter Ueff represents on-site contributions to 
the correlation energy which are underestimated by L(S)DA. Ueff is, 
therefore, an approximation that has to estimate the contribution 
already accounted for by L(S)DA  - which leads to options 1,2,3 for 
different self-interaction corrections.

I doubt that there is a general and safe way to assign specific parts of 
the correlation energy to U, V, and J. At least, the authors of the ORB 
code recommend to avoid the ambiguities [2], and to set J=0.

[1] see for example, the talk 'Localized Electrons with Wien2k' by Elias 
Assmann Wien, workshop 2013, ldau.pdf among the documentations on the 
Wien2k site
[2] G. K. H. Madsen and P. Novák, Europhys. Lett. 69, 777 (2005)

---
Dr. Martin Pieper
Karl-Franzens University
Institute of Physics
Universitätsplatz 5
A-8010 Graz
Austria
Tel.: +43-(0)316-380-8564


Am 28.11.2017 08:42, schrieb karima Physique:
> Dear prof. P. Blaha and wien2k users :
> 
> I am still waiting for your answers and I thank you in advance.
> 
> my question was :
> 
> Does Ueff=U-J or Ueff=U- 1/2 J? how to estimate J using constrained
> DFT.?
> 
> Thank you in advance
> 
> 2017-11-27 12:05 GMT+01:00 karima Physique
> <physique.karima at gmail.com>:
> 
>> Dear Wien2k users:
>> 
>> Does Ueff=U-J or Ueff=U- 1/2 J? how to estimate J using constrained
>> DFT.?
>> 
>> Thank you in advance
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html