[Pw_forum] lda+U abinitio U

[Lorenzo Paulatto]

Thank you Matteo,
this should be doable, and thank you for providing the tutorial, even if it is 
a bit rough on the edges it was very useful.


On Monday, September 5, 2016 12:11:42 AM CEST Matteo Cococcioni wrote:
> Hi Lorenzo,
> 
> the two tutorial should be quite equivalent, although the one Federico
> found is more recent and probably I updated some details.
> 
> Regarding your questions: as far as I understand you are only interested in
> U on one specific atomic species of your materials. You could do that by
> doing two perturbations: one on the atom you are interested in, the second
> on all the other atoms (which you treat as a "macro-atom") at the same
> time. Of course the occupation of this second "extended atom" is the sum of
> the occupations of the atoms it is composed by. you then construct a 2x2
> response matrix that, once inverted, should give you the U you are looking
> for (throwing away the one for the "macro-atom").
> This only limits the number of perturbation but still requires a supercell
> of sufficiently large size.
> 
> We are working on a way to make the whole procedure more efficient, through
> automatizing the linear-response, but the code is not ready yet.
> 
> Best,
> 
> Matteo
> 
> On Fri, Sep 2, 2016 at 5:55 PM, Lorenzo Paulatto <
> 
> lorenzo.paulatto at impmc.upmc.fr> wrote:
> > Hello,
> > I have a question about how to compute the abinitio U, I'm following the
> > online Santa Barbara 2009 tutorial (and trying to use the script therein)
> > and
> > PRB 71, 35105 (2005).
> > 
> > I have a system that is not very large but quite heavy to compute, and
> > have
> > several atomic species (3 at first, then 4) but only one is problematic,
> > as it
> > is could be Cerium or other Lanthanoids.
> > 
> > It would be quite anoying and very cpu time consuming for me to properly*
> > compute U by finite response for all the species, but I'm not sure that
> > 1. it is possible with the available toolchain
> > 2. it makes sense
> > 
> > Also, if somebody (e.g. Matteo) has some more recent and/or more user
> > friendly
> > code to compute dalpha/dn I'm a taker.
> > 
> > thank you for your help!
> > 
> > *) by properly I mean take care that it is converged with the supercell
> > size
> > 
> > --
> > Dr. Lorenzo Paulatto
> > IdR @ IMPMC -- CNRS & Université Paris 6
> > +33 (0)1 44 275 084 / skype: paulatz
> > http://www.impmc.upmc.fr/~paulatto/
> > 23-24/4é16 Boîte courrier 115,
> > 4 place Jussieu 75252 Paris Cédex 05
> > 
> > _______________________________________________
> > Pw_forum mailing list
> > Pw_forum at pwscf.org
> > http://pwscf.org/mailman/listinfo/pw_forum


-- 
Dr. Lorenzo Paulatto 
IdR @ IMPMC -- CNRS & Université Paris 6
phone: +33 (0)1 44275 084 / skype: paulatz
www:   http://www-int.impmc.upmc.fr/~paulatto/
mail:  23-24/4é16 Boîte courrier 115, 4 place Jussieu 75252 Paris Cédex 05