[Pw_forum] Occupation matrix in LDA+U

[elbuesta at icqmail.com]

 Dear Dr. Cococcioni,

Thanks very much for your explanation and your tips. They were very valuable.

All best!

 



-----Original Message-----
From: Matteo Cococcioni <matteo at umn.edu>
To: PWSCF Forum <pw_forum at pwscf.org>
Sent: Thu, Mar 25, 2010 10:01 pm
Subject: Re: [Pw_forum] Occupation matrix in LDA+U

 
 
 
 
Dear Luca and Fabio 
 
here are my answers. 
 
1) the diagonal elements of the occupation matrix are the projection on  
the atomic orbitals in the order 
you wrote. Of course also non diagonal elements are associated with d  
orbitals. they are not projection strictly 
speaking but you can see them as generalized overlap terms between d  
atomic orbitals in the subspace of occupied Kohn-Sham states. 
 
2) yes. from Doc/INPUT_PW: 
 
   +-------------------------------------------------------------------- 
   Variable:       starting_ns_eigenvalue(m,ispin,I) 
 
   Type:           REAL 
   Default:        -1.d0 that means NOT SET 
   Description:    In the first iteration of an LDA+U run it overwrites 
                   the m-th eigenvalue of the ns occupation matrix for the 
                   ispin component of atomic species I. Leave unchanged 
                   eigenvalues that are not set. This is useful to suggest 
                   the desired orbital occupations when the default choice 
                   takes another path. 
   +-------------------------------------------------------------------- 
 
keep in mind that you are adjusting the eigenvalue of the matrix. So you  
better give a look at the occupation matrix 
and its spectrum after the first iteration of a non constraint run. 
 
hope this helps, 
 
Matteo 
 
 
 
elbuesta at icqmail.com wrote: 
> Dear all, 
> 
> We have a fews questions/doubts concerning the ns occupation matrix  
> for LDA+U calculations (starting_ns_eigenvalue(m,ispin,I)). Following  
> the idea presented in example 25 of quantum espresso, we are trying to  
> fix the initial occupation of the "d" orbital of Ni in a monolayer of  
> NiO in an anti-ferromagnetic state. The questions we have are: 
> 
> 1) Are each of the diagonal elements of the occupation matrix  
> associated with the d orbitals? Something like, 
> (1,1) -> dz2, (2,2) -> dzx, (3,3) -> dzy, (4,4) -> d(x2-y2) and  (5,5)  
> -> dxy   ? 
> 2) If there is this correspondence, is it possible to specify an  
> initial occupation of each of the orbitals after the initial guess is  
> evaluated in the first iteration? For example, in the attached input,  
> how could we specify that only the dz2, dxz and dyz orbitals are  
> initially completely occupied with 2 electrons? In other words, how  
> can we change the "spin down occupation matrix" of the Ni atom from 
> 0.6 0.0 0.0 0.0 0.0            1.0 0.0 0.0 0.0 0.0 
> 0.0 0.6 0.0 0.0 0.0            0.0 1.0 0.0 0.0 0.0 
> 0.0 0.0 0.6 0.0 0.0     ->    0.0 0.0 1.0 0.0 0.0 
> 0.0 0.0 0.0 0.6 0.0            0.0 0.0 0.0 0.0 0.0 
> 0.0 0.0 0.0 0.0 0.6            0.0 0.0 0.0 0.0 0.0 
> ?? 
> 
> Thanks very much for your time, 
> With best regards, 
> 
> Luca Sementa, Fábio Negreiros 
> Postdoc students, 
> IPCF - CNR - Pisa, Italy 
> 
> = 
> ------------------------------------------------------------------------ 
> 
> _______________________________________________ 
> Pw_forum mailing list 
> Pw_forum at pwscf.org 
> http://www.democritos.it/mailman/listinfo/pw_forum 
 
_______________________________________________ 
Pw_forum mailing list 
Pw_forum at pwscf.org 
http://www.democritos.it/mailman/listinfo/pw_forum 
 

 

=
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.democritos.it/pipermail/pw_forum/attachments/20100326/2b68e9e0/attachment.htm