Dear Dr. Cococcioni, Thanks very much for your explanation and your tips. They were very valuable.
All best! -----Original Message----- From: Matteo Cococcioni <[email protected]> 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, F?bio 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
