<div style="font-family: arial; color: black; font-size: 10pt;"> Dear Dr. Cococcioni,<br>
<br>
Thanks very much for your explanation and your tips. They were very valuable.<br>
<br>
All best!<br>
</div>
<div> <br>
</div>
<div><br>
</div>
-----Original Message-----<br>
From: Matteo Cococcioni <matteo@umn.edu><br>
To: PWSCF Forum <pw_forum@pwscf.org><br>
Sent: Thu, Mar 25, 2010 10:01 pm<br>
Subject: Re: [Pw_forum] Occupation matrix in LDA+U<br>
<br>
<br>
<div id="AOLMsgPart_0_616a8555-52fb-47ab-8d95-41aff696dbc2" style="margin: 0px; font-family: Tahoma,Verdana,Arial,Sans-Serif; font-size: 12px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);"> <br>
<br>
<pre style="font-size: 9pt;"><tt> <br>
Dear Luca and Fabio <br>
<br>
here are my answers. <br>
<br>
1) the diagonal elements of the occupation matrix are the projection on <br>
the atomic orbitals in the order <br>
you wrote. Of course also non diagonal elements are associated with d <br>
orbitals. they are not projection strictly <br>
speaking but you can see them as generalized overlap terms between d <br>
atomic orbitals in the subspace of occupied Kohn-Sham states. <br>
<br>
2) yes. from Doc/INPUT_PW: <br>
<br>
+-------------------------------------------------------------------- <br>
Variable: starting_ns_eigenvalue(m,ispin,I) <br>
<br>
Type: REAL <br>
Default: -1.d0 that means NOT SET <br>
Description: In the first iteration of an LDA+U run it overwrites <br>
the m-th eigenvalue of the ns occupation matrix for the <br>
ispin component of atomic species I. Leave unchanged <br>
eigenvalues that are not set. This is useful to suggest <br>
the desired orbital occupations when the default choice <br>
takes another path. <br>
+-------------------------------------------------------------------- <br>
<br>
keep in mind that you are adjusting the eigenvalue of the matrix. So you <br>
better give a look at the occupation matrix <br>
and its spectrum after the first iteration of a non constraint run. <br>
<br>
hope this helps, <br>
<br>
Matteo <br>
<br>
<br>
<br>
<a href="mailto:elbuesta@icqmail.com">elbuesta@icqmail.com</a> wrote: <br>
> Dear all, <br>
> <br>
> We have a fews questions/doubts concerning the ns occupation matrix <br>
> for LDA+U calculations (starting_ns_eigenvalue(m,ispin,I)). Following <br>
> the idea presented in example 25 of quantum espresso, we are trying to <br>
> fix the initial occupation of the "d" orbital of Ni in a monolayer of <br>
> NiO in an anti-ferromagnetic state. The questions we have are: <br>
> <br>
> 1) Are each of the diagonal elements of the occupation matrix <br>
> associated with the d orbitals? Something like, <br>
> (1,1) -> dz2, (2,2) -> dzx, (3,3) -> dzy, (4,4) -> d(x2-y2) and (5,5) <br>
> -> dxy ? <br>
> 2) If there is this correspondence, is it possible to specify an <br>
> initial occupation of each of the orbitals after the initial guess is <br>
> evaluated in the first iteration? For example, in the attached input, <br>
> how could we specify that only the dz2, dxz and dyz orbitals are <br>
> initially completely occupied with 2 electrons? In other words, how <br>
> can we change the "spin down occupation matrix" of the Ni atom from <br>
> 0.6 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 <br>
> 0.0 0.6 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 <br>
> 0.0 0.0 0.6 0.0 0.0 -> 0.0 0.0 1.0 0.0 0.0 <br>
> 0.0 0.0 0.0 0.6 0.0 0.0 0.0 0.0 0.0 0.0 <br>
> 0.0 0.0 0.0 0.0 0.6 0.0 0.0 0.0 0.0 0.0 <br>
> ?? <br>
> <br>
> Thanks very much for your time, <br>
> With best regards, <br>
> <br>
> Luca Sementa, Fábio Negreiros <br>
> Postdoc students, <br>
> IPCF - CNR - Pisa, Italy <br>
> <br>
> = <br>
> ------------------------------------------------------------------------ <br>
> <br>
> _______________________________________________ <br>
> Pw_forum mailing list <br>
> <a href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a> <br>
> <a target="_blank" href="http://www.democritos.it/mailman/listinfo/pw_forum">http://www.democritos.it/mailman/listinfo/pw_forum</a> <br>
<br>
_______________________________________________ <br>
Pw_forum mailing list <br>
<a href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a> <br>
<a target="_blank" href="http://www.democritos.it/mailman/listinfo/pw_forum">http://www.democritos.it/mailman/listinfo/pw_forum</a> <br>
</tt></pre> <br>
</div>
<br>