[Wien] Problem in LDA+U calculation

Hena Das hd249 at cornell.edu
Thu Jun 28 17:55:24 CEST 2012


Okay...Thanks.

________________________________________
From: wien-bounces at zeus.theochem.tuwien.ac.at [wien-bounces at zeus.theochem.tuwien.ac.at] on behalf of tran at theochem.tuwien.ac.at [tran at theochem.tuwien.ac.at]
Sent: Thursday, June 28, 2012 11:36 AM
To: A Mailing list for WIEN2k users
Subject: Re: [Wien] Problem in LDA+U calculation

Ok. There was another problem with your case.struct: the RMT of
some atoms were too small, leading to core leakage (during init_lapw).
Also, with LDA+U it is better to use large RMT because U is applied only
inside the sphere.

On Thu, 28 Jun 2012, Hena Das wrote:

> Thank you for the solution, it is working now.
>
> ________________________________________
> From: wien-bounces at zeus.theochem.tuwien.ac.at [wien-bounces at zeus.theochem.tuwien.ac.at] on behalf of tran at theochem.tuwien.ac.at [tran at theochem.tuwien.ac.at]
> Sent: Wednesday, June 27, 2012 1:54 PM
> To: A Mailing list for WIEN2k users
> Subject: Re: [Wien] Problem in LDA+U calculation
>
> I have the same problem and the error message in case.outputorbup is
> "Conflict in atom orb. number: lorb           3 ne ll           2".
> The reason is that in case.dmatup(dn), the blocks for a given atom are
> ordered in increasing value of l (even if they are not ordered in case.indmc),
> while orb thinks that the order is like in case.inorb.
> I solved the problem by ordering the values of l in case.inorb (don't
> forget to also order the values of U and J accordingly):
>   1 2 2 3
>   2 2 2 3
>   3 1 2
> Then it runs.
>
> Another unrelated problem: during init_lapw, lstart complains about the
> values of R0 in case.struct. You have to regenerate case.struct, such that
> the values of R0 (they are set automatically) are ok. The values of R0
> should never be set manually.
>
>
> On Wed, 27 Jun 2012, Hena Das wrote:
>
> > @Robert: Yes I used Wien2k orb executables.
> > @Tran: I followed your suggestion and tried to do the same in Wien2k. The same error appeared even in Wien2k.
> > I tried both options :   1 2 3 2
> >                                 2 2 3 2
> >                                    &
> >                                 1 2 2 3
> >                                 2 2 2 3
> > Below are the case.indmc and case.inorb files that I have used:
> > case.indmc
> > -9.                      Emin cutoff energy
> >  3                       number of atoms for which density matrix is calculated
> >  1  2  3  2   index of 1st atom, number of L's, L1
> >  2  2  3  2   dtto for 2nd atom, repeat NATOM times
> >  3  1  2      dtto for 2nd atom, repeat NATOM times
> >  0 0           r-index, (l,s)index
> > case.inorb
> >   1  3  0                     nmod, natorb, ipr
> > PRATT  1.0                    BROYD/PRATT, mixing
> >   1 2 3 2
> >   2 2 3 2
> >   3 1 2                          iatom nlorb, lorb
> >   1                              nsic 0..AMF, 1..SIC, 2..HFM
> >    0.53 0.07
> >    0.43 0.07            U J (Ry)
> >    0.53 0.07
> >    0.43 0.07
> >    0.43 0.07
> >
> > I hope I am not using any wrong input. Can you please check.
> >
> > Hena
> >
> > ________________________________________
> > From: wien-bounces at zeus.theochem.tuwien.ac.at [wien-bounces at zeus.theochem.tuwien.ac.at] on behalf of tran at theochem.tuwien.ac.at [tran at theochem.tuwien.ac.at]
> > Sent: Wednesday, June 27, 2012 3:22 AM
> > To: wien at zeus.theochem.tuwien.ac.at
> > Subject: Re: [Wien] Problem in LDA+U calculation
> >
> > Dear Hena,
> >
> > Actually, it should work with WIEN2k (I don't know with Wienncm).
> > I have tried myself on NiO with
> > l=1 and 2 on nickel and l=1 on oxygen and it's running properly.
> > The first thing you should do is to try with WIEN2k and
> > see if it works or not. If it works with WIEN2k, but not with
> > Wienncm, it could be that also a c-shell script of Wienncm has to be
> > modified.
> >
> > But just in case, you can also try this:
> >
> > replace
> >
> >   1 2 3 2
> >   2 2 3 2
> >
> > by
> >
> >   1 2 2 3
> >   2 2 2 3
> >
> > in case.indmc and case.inorb
> >
> > F. Tran
> >
> > On 25.06.2012 17:45, Hena Das wrote:
> >
> > >I am doing a scf calculation for a particular non-collinear spin
> > >configuration. A am using LDA+U method. When I use the following
> > >case.inorb and case.indmc files:
> > >case.inorb
> > > 1  3  0                     nmod, natorb, ipr
> > >PRATT  1.0                    BROYD/PRATT, mixing
> > >  1 1 3
> > >  2 1 3
> > >  3 1 2
> > >  1                              nsic 0..AMF, 1..SIC, 2..HFM
> > >   0.65 0.07                        U J (Ry)
> > >   0.65 0.07
> > >   0.45 0.07
> > >case.indmc
> > >-9.                      Emin cutoff energy
> > > 3                       number of atoms for which density matrix is
> > >calculated
> > > 1  1  3      index of 1st atom, number of L's, L1
> > > 2  1  3      dtto for 2nd atom, repeat NATOM times
> > > 3  1  2      index of 1st atom, number of L's, L1
> > > 0 0           r-index, (l,s)index
> > >
> > >the program executes properly without any error. However when I use the
> > >other set of case.inorb and case.indmc files:
> > >case.inorb
> > >  1  3  0                     nmod, natorb, ipr
> > >PRATT  1.0                    BROYD/PRATT, mixing
> > >  1 2 3 2
> > >  2 2 3 2
> > >  3 1 2
> > >  1                              nsic 0..AMF, 1..SIC, 2..HFM
> > >   0.65 0.07
> > >   0.45 0.07                        U J (Ry)
> > >   0.65 0.07
> > >   0.45 0.07
> > >   0.45 0.07
> > >case.indmc
> > >-9.                      Emin cutoff energy
> > > 3                       number of atoms for which density matrix is
> > >calculated
> > > 1  2  3  2      index of 1st atom, number of L's, L1
> > > 2  2  3  2      dtto for 2nd atom, repeat NATOM times
> > > 3  1  2      index of 1st atom, number of L's, L1
> > > 0 0           r-index, (l,s)index
> > >the program stops by giving the error: error in Vorb. It is not writing
> > >case.vorbup/dn files. In the next set I just use U at the d states as
> > >well
> > >as at the f states for atom 1 and 2. Am I giving any wrong input? Help me
> > >to solve this problem.
> > >
> > >Waiting for suggestions.
> > >
> > >Best,
> > >Hena
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
_______________________________________________
Wien mailing list
Wien at zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien


More information about the Wien mailing list