[Wien] New exchange-correlation potential

[F. Tran]

Hello,

For volume cell optimization you have two possibilities:

(1) Do the volume cell optimization using LDA or GGA for
both the energy functional and the potential, and then do a
self-consistent MBJ calculation at the obtained volume.

(2) Do the optimization of the unit cell volume (but not the internal
parameters) with indxc=28 (LDA for energy and MBJ for potential).

I recommend to use (1), because the scf convergence with the MBJ potential
can be very slow and probably you don't want to deal with this slow
convergence at every geometry.

In general, comparing total energies by plugging into the energy
functional an electron density which was obtained with a potential
which is not the derivative of the functional is not very good.

Actually, when comparing the total energy of two magnetic states, the
result will depend also on the energy functional. We could have chosen,
e.g., PBE, for the energy functional associated with the MBJ potential.

So, MBJ is interesting for the band structure and magnetic moments, but
much less for the total energy and geometry optimization.


On Wed, 9 Jun 2010, Ricardo Faccio wrote:

> Dear F. Tran
> What about the use of total energy for: Volume cell optimization and/or for
> evaluation of a particular magnetic arrangement ? Can it be used for
> comparative purposes? or the procedure indicates that we need to optimize
> everything first (forces and volume) for a particular "variational"
> xc-potentinal (eg. PBE, etc) and then proceed with MBJ by doing a single
> point?
> Best regards
> Ricardo
> -------------------------------------------------------------------------
>  Dr. Ricardo Faccio
>  Prof. Adjunto de Física
>  Mail: Cryssmat-Lab., Cátedra de Física, DETEMA
>  Facultad de Química, Universidad de la República
>       Av. Gral. Flores 2124, C.C. 1157
>       C.P. 11800, Montevideo, Uruguay.
>  E-mail: rfaccio at fq.edu.uy
>  Phone: 598 2 924 98 59
>              598 2 929 06 48
>  Fax:    598 2 9241906
>  Web:  http://cryssmat.fq.edu.uy/ricardo/ricardo.htm
> ---------------------------------------------------------------------------------
> ----- Original Message ----- From: "F. Tran" <tran at theochem.tuwien.ac.at>
> To: "A Mailing list for WIEN2k users" <wien at zeus.theochem.tuwien.ac.at>
> Sent: Wednesday, June 09, 2010 6:54 PM
> Subject: Re: [Wien] New exchange-correlation potential
> 
> 
> > Hello,
> >
> > In order to use the modified Becke-Johnson exchange potential
> > you have to choose indxc=28 in case.in0 and indxc=50 in case.in0_grr.
> >
> > A few more things:
> >
> > In case.in0 you also need to replace NR2V by R2V such that the
> > exchange-correlation potential is written in case.r2v, which will
> > be used in the next iteration.
> >
> > You need case.inm_vresp such that case.vrespsum (which contains a
> > part of the kinetic-energy density) is created.
> >
> > There is no adjustable parameters.
> >
> > The LDA correlation potential is added to the MBJ exchange potential.
> >
> > The LDA exchange-correlation functional is used for the energy.
> >
> > The MBJ exchange potential is not the derivative of an energy functional,
> > which means that the method is not variational (i.e., the orbitals are
> > not the ones which minimize the total energy, which is LDA).
> > In particular, it means that you can not use forces to optimize internal
> > parameters.
> >
> >
> > On Wed, 9 Jun 2010, yhzhao wrote:
> >
> > > Dear Prof. Pblaha,
> > >
> > > I have tried to calculate something using the new exchange potential
> > > with indxc=50 as suggested by the usersguide. At first I want to repeat
> > > some data given by PRL 102, 226401 (2009), for example, the band gap of
> > > GaAs. I can get a band gap of 0.7 eV for GaAs, which is much better than
> > > 0.3 eV of lda, but not much better than the value given by you in PRL
> > > 102, 226401 (2009). In the calculation, I use the experimental lattice
> > > constant. The k-grid have been increased as dense as 20x20x20 and RKmax
> > > as large as 9.0.
> > >
> > > Is there any adjustable parameter for the MBJ calculation? Thanks.
> > >
> > > Regards,
> > >
> > > yonghong
> > >
> > > _______________________________________________
> > > 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
>