[Pw_forum] DFT+U

Kucukbenli Emine emine.kucukbenli at epfl.ch
Sun Apr 28 16:19:31 CEST 2013


Dear Suza,

The Hubbard correction to the energy is determined as a function of the occupation of the Hubbard orbitals (such as d for Pb).
How do we find the occupation of these orbitals?
By projecting the Kohn Sham orbitals (KS) on the Hubbard orbitals (d) :
 |<d|KS>|^2 * (occ. of |KS>)

Now perhaps you can answer your own question:
If you have constructed your system with only  s and p electrons of Pb (5d in the core )
how would the projection of Kohn Sham solutions of this system on Pb 5d would look like?
Would this projection be what you want to apply U with ?

Then there is the technical part of where to read the Hubbard orbitals from, if they are not written in pseudo. But there are alternatives for that. I think the other question above is more significant.

About the two different hubbard U for two different orbitals on the same site question:
We have implemented this recently in the current version of the code, it is still very fresh and we are testing it.
The U+V formalism[1] allows >1 orbital per site to have U correction, and even off-site correction.
It is indeed physically meaningful, and easy to understand from the piecewise linearity picture
(or at least that is how i understand [2] :)

best
emine kucukbenli, postdoc at theos, epfl, switzerland

[1]V. L. Campo Jr. and M. Cococcioni, "Extended DFT+U+V method with on-site and inter-site electronic interactions", Journal of Physics: Condensed Matter 22, 055602 (2010)
[2]I think Cococcioni's thesis explains this link between piecewise linearity and U very well: www.sissa.it/cm/thesis/2002/cococcioni.pdf


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20130428/38e41448/attachment.html>


More information about the users mailing list