[Pw_forum] DFT+U and starting_ns_eigenvalue(m,ispin,I)

Wajood A Diery wdiery at mit.edu
Thu Mar 26 14:44:55 CET 2015 

Dear Matteo,
What exactly meant is if there is a way to control the occupation matrix like the one discuss in this article http://chemistry.tcd.ie/staff/people/gww/gw_new/research/methodology/, and here http://journals.aps.org/prb/abstract/10.1103/PhysRevB.79.235125, where you can set up or define the occupation matrix.
Best
Wajood
________________________________
From: pw_forum-bounces at pwscf.org [pw_forum-bounces at pwscf.org] on behalf of Matteo Cococcioni [matteo at umn.edu]
Sent: 25 March 2015 10:13
To: PWSCF Forum
Subject: Re: [Pw_forum] DFT+U and starting_ns_eigenvalue(m,ispin,I)


Dear Wajood,

to be precise: the starting_ns_eigenvalue flag allows you to change the eigenvalue keeping the eigenvector fixed at what it was after the first iteration (i.e., when the first non trivial occupation matrix is obtained).
But I still do not understand what you want to do. What is the difference between "filling of d levels" and "eigenvalues of the occupation matrix"? What do you mean exactly for "d levels"? If you mean Kohn-Sham states yes, you can change their occupations (there is a card to do that, after you choose occupations = 'from_input" if I'm not wrong) but they are not going to be purely d states (i.e. they are not going to correspond exactly to Bloch sums of atomic d states). Maybe some will but it all depends on the electronic structure of the system.

Matteo

On Wed, Mar 25, 2015 at 2:53 PM, Wajood A Diery <wdiery at mit.edu<mailto:wdiery at mit.edu>> wrote:
Dear Matteo,
I dont think this is what am doing. As far as I understand, starting_ns_eigenvalue(m,ispin,I) overwrites the mth eigenvalue not the occupation matrix it self. My question was if i can define the occupation matrix that corresponds to different filling of the d levels.
Best
Wajood

________________________________
From: pw_forum-bounces at pwscf.org<mailto:pw_forum-bounces at pwscf.org> [pw_forum-bounces at pwscf.org<mailto:pw_forum-bounces at pwscf.org>] on behalf of Matteo Cococcioni [matteo at umn.edu<mailto:matteo at umn.edu>]
Sent: 25 March 2015 09:24
To: PWSCF Forum
Subject: Re: [Pw_forum] DFT+U and starting_ns_eigenvalue(m,ispin,I)

Dear Wajood,

isn't this what you are already doing?

Matteo

On Wed, Mar 25, 2015 at 11:33 AM, Wajood A Diery <wdiery at mit.edu<mailto:wdiery at mit.edu>> wrote:
Thank you so much.  One more question: is there a way in quantum espresso to control the occupation matrix where the user can define the occupation of the orbital in the occupation matrix?
Best
Wajood
________________________________
From: pw_forum-bounces at pwscf.org<mailto:pw_forum-bounces at pwscf.org> [pw_forum-bounces at pwscf.org<mailto:pw_forum-bounces at pwscf.org>] on behalf of Matteo Cococcioni [matteo at umn.edu<mailto:matteo at umn.edu>]
Sent: 24 March 2015 13:56
To: PWSCF Forum
Subject: Re: [Pw_forum] DFT+U and starting_ns_eigenvalue(m,ispin,I)

Dear Wajood,

it is not a problem. the eigenvalues are printed in growing order. the important thing is that you get the xz state corresponding to the 0 eigenvalue. and it seems to be the case (although it is not a pure xz and has a component on z^2)

regards,

Matteo

On Tue, Mar 24, 2015 at 6:30 PM, Wajood A Diery <wdiery at mit.edu<mailto:wdiery at mit.edu>> wrote:
Dear Pwscf users,
I have few questions about starting_ns_eigenvalue(m,ispin,I) and the occupation matrix for DFT+U. Am trying to fix the occupation of the d orbitals in Mn in BiMnO3. I've used starting_ns_eigenvalue(m,ispin,I) to specify the following occupation: 4 electrons in dz2, dyz, dx2-y2 and dxy and zero occupation for dxz. I was expecting to get eigenvalues such as  1.000  0.000  1.000  1.000  1.000, which give the occupation of the d-orbital,  but I got  0.000  1.000  1.000  1.000  1.000?  i've tried different occupations of the d-orbital and I always  get the same occupation or eigenvalues : 0.000  1.000  1.000  1.000  1.000? I am not sure if I miss understand the statring_ns_eigenvalue tag or there is something else that am missing? any help or suggestion will  be appreciated.
Wajood
Mechanical Engineering Department
MIT
This is the part of the input file related to my Question
&system
    ibrav = 0,
     celldm(1)= 18.013436
    nat = 40,
    ntyp = 3,
    ecutwfc = 60,
    nspin = 2,
    starting_magnetization(2)=1,
   occupations='smearing', smearing='gauss',
    degauss=0.01
   lda_plus_u=.true.
   lda_plus_u_kind=1
     Hubbard_U(2)=8
    Hubbard_J(1,2)=1.06
starting_ns_eigenvalue(1,1,2)=1.0,
starting_ns_eigenvalue(2,1,2)=0.0,
starting_ns_eigenvalue(3,1,2)=1.0,
starting_ns_eigenvalue(4,1,2)=1.0,
starting_ns_eigenvalue(5,1,2)=1.0,


The output file is:
LDA+U parameters:
U(  2) =   8.0000   J(  2) =   1.0600   B(  2) =   0.1217
atom    9   Tr[ns(na)] (up, down, total) =   4.00000  0.19494  4.19494
   spin  1
    eigenvalues:
  0.000  1.000  1.000  1.000  1.000
    eigenvectors:
  0.110  0.701  0.037  0.000  0.153
  0.795  0.016  0.023  0.000  0.166
  0.000  0.004  0.885  0.000  0.111
  0.096  0.279  0.055  0.000  0.570
  0.000  0.000  0.000  1.000  0.000
    occupations:
  0.890 -0.295  0.000 -0.102  0.000
 -0.295  0.205  0.000 -0.276  0.000
  0.000  0.000  1.000  0.000  0.000
 -0.102 -0.276  0.000  0.904  0.000
  0.000  0.000  0.000  0.000  1.000
   spin  2
    eigenvalues:
  0.020  0.020  0.021  0.065  0.070
    eigenvectors:
  0.000  0.512  0.416  0.071  0.000
  0.000  0.004  0.193  0.803  0.000
  0.542  0.000  0.000  0.000  0.458
  0.000  0.484  0.390  0.126  0.000
  0.458  0.000  0.000  0.000  0.542
    occupations:
  0.023  0.010  0.000  0.005  0.000
  0.010  0.056  0.000  0.014  0.000
  0.000  0.000  0.043  0.000 -0.025
  0.005  0.014  0.000  0.026  0.000
  0.000  0.000 -0.025  0.000  0.047
atomic mag. moment =     3.805060

_______________________________________________
Pw_forum mailing list
Pw_forum at pwscf.org<mailto:Pw_forum at pwscf.org>
http://pwscf.org/mailman/listinfo/pw_forum


_______________________________________________
Pw_forum mailing list
Pw_forum at pwscf.org<mailto:Pw_forum at pwscf.org>
http://pwscf.org/mailman/listinfo/pw_forum


_______________________________________________
Pw_forum mailing list
Pw_forum at pwscf.org<mailto:Pw_forum at pwscf.org>
http://pwscf.org/mailman/listinfo/pw_forum

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20150326/c8c9d9fc/attachment.html>

More information about the users mailing list