The ABINIT Users Mailing List ( CLOSED )

["Anglade Pierre-Matthieu" <anglade@gmail.com>]

> the calculation will never converge! Having one k-point for
> a large box with minimal dispersion should be fine.

That's nice if you don't fear the flaws this can introduce into your
work. Otherwise there is many otherway you can try:
1)Then instead of tsmear/occopt/diemac you might want to use iprcel.
It can do "miracles".
2) If you don't like it and since you seem to like large number of
step you can also try the simples iscf 2 with a small value for
diemix.
3) Otherwise you may also try to converge with tsmear then run a
second dataset based on the previous wave funtion but just decreasing
temperature. You'll get a converged result and you'll see what the
effect of tsmear is on your calculation.

About k-point on dielectric materials I've no experience. However, I
myself would never compare energies between two calculations to a
larger extent than the absolute convergence I have in term of
k-points.

regards

PMA


> I'll try to the use method 1 you mention and see how it
> goes. Many thanks for that.
>
> best regards,
> Aloysius
>
> Anglade Pierre-Matthieu wrote:
> > Hi,
> >
> > 1) I'm not sure your method for computing the binding energy is
> > perfect. There may be some flaws here. Althought I'm not an expert on
> > this matter, I would recommand you to try plotting the Energy vs
> > interatomic distance curve. I've also been told that van der Waals
> > interactions are (beacuse of error cancelations) best taken into
> > acount with LDA.
> >
> > 2) You calculation may not be perfectly set up.  Main points are:
> > a) you use occopt  3 : is you oxygen metallic?
> > b) you have really few k points in your cell. Is this really enough?
> >
> > regards
> >
> > PMA
> >
> > On 6/27/06, Aloysius Soon <aloysius@physics.usyd.edu.au> wrote:
> >> Dear all,
> >> I've been playing around for quite a while with various
> >> examples found in the Test_v* and Tut, but I can't seem to
> >> get the right answer for a simple (although not that simple)
> >> like O2 molecule in a box. Upon subtraction, the binding
> >> energy that I've got out of a typical GGA calculation is 4
> >> to 5 eV per O atom, which is way-off the usual that
> >> literature or other codes I've used (~3.1 to 3.3 eV). I've
> >> attached one of the many input files that I've used for the
> >> atom and molecule. Hope someone could help me out in this.
> >>
> >> Many thanks.
> >>
> >> == input file for oxygen atom ==
> >> occopt 3           # Could not converge with default
> >> tsmear 0.03 eV
> >> #band  16
> >> #iscf   17
> >> #npulayit 3
> >> nsym  1
> >> nsppol 2           # Needed for spin-polarized run
> >> kptopt 0
> >> kpt    0.25 0.25 0.25
> >>
> >> #Definition of the unit cell
> >> acell   9.0   10.0   11.0 angstrom
> >> #angdeg  90  90  90
> >> rprim
> >>                1.00000000000000    0.00000000000000
> >> 0.00000000000000
> >>                0.00000000000000    1.00000000000000
> >> 0.00000000000000
> >>                0.00000000000000    0.00000000000000
> >> 1.00000000000000
> >> chkprim 0
> >>
> >> #Definition of the atom types
> >> ntypat 1
> >> znucl  8
> >>
> >> #Definition of the atoms
> >> natom 1
> >> typat 1
> >> xred
> >>              0.00000000000000    0.00000000000000
> >> 0.00000000000000
> >>
> >> #Exchange-correlation functional
> >> ixc 11             # GGA-PBE
> >>
> >> #Definition of the planewave basis set
> >> ecut         70.0 Ry
> >>
> >>
> >> #Definition of the SCF procedure
> >> nstep 99
> >> toldfe 1.0d-12
> >> #tolvrs 1.0d-8
> >> diemac 1.5
> >> diemix 0.5
> >> ===EOF===
> >>
> >> == input file for oxygen molecule ==
> >> occ 1.0  1.0  1.0  1.0  1.0  1.0  1.0  0.0
> >>      1.0  1.0  1.0  1.0  1.0  0.0  0.0  0.0
> >> occopt 2
> >> nsppol 2
> >> nband  8 8
> >> nsym  1
> >> kptopt 0
> >> kpt    0.25 0.25 0.25
> >>
> >> #Definition of the unit cell
> >> acell     9.0   10.0   11.0 angstrom
> >>
> >> #angdeg  90  90  90
> >> rprim
> >>                1.00000000000000    0.00000000000000
> >> 0.00000000000000
> >>                0.00000000000000    1.00000000000000
> >> 0.00000000000000
> >>                0.00000000000000    0.00000000000000
> >> 1.00000000000000
> >> chkprim 0
> >>
> >> #Optimization of the lattice parameters
> >> optcell 0
> >> ionmov  3
> >> ntime   99
> >> dilatmx 1.1
> >> ecutsm  0.5
> >> prtgeo  2
> >>
> >> #Definition of the atom types
> >> ntypat 1
> >> znucl  8
> >>
> >> #Definition of the atoms
> >> natom 2
> >> typat 1 1
> >> xangst
> >>                0.00000000000000     0.00000000000000
> >> 0.60856624824230
> >>                0.00000000000000     0.00000000000000
> >> -0.60856624824230
> >>
> >> #Exchange-correlation functional
> >> ixc 11             # GGA-PBE
> >>
> >> #Definition of the planewave basis set
> >> ecut         70.0 Ry
> >>
> >> #Definition of the SCF procedure
> >> nstep 99
> >> toldfe 1.0d-8
> >> diemac   1.5
> >> diemix   0.5
> >> ===EOF===
> >>
> >>
> >>
> >> Best regards,
> >> Aloysius
> >>
> >> --
> >> Aloysius Soon
> >> Condensed Matter Theory Group
> >> School of Physics A28, Room 361
> >> The University of Sydney, NSW 2006 Australia
> >> Phone: +61 2 903 65389
> >> Fax: +61 2 935 17726
> >> Email: aloysius@physics.usyd.edu.au
> >> Web: http://www.physics.usyd.edu.au/~aloysius/
> >> ==============================================
> >> - CARPE DIEM -
> >> "Gather ye rosebuds while ye may,
> >> Old time is still a-flying,
> >> And this same flower that smiles today,
> >> To-morrow will be dying."
> >>                         ROBERT HERRICK
> >>                         1591-1674
> >> ==============================================
> >>
> >>
> >
> >
>
> --
> Aloysius Soon
> Condensed Matter Theory Group
> School of Physics A28, Room 361
> The University of Sydney, NSW 2006 Australia
> Phone: +61 2 903 65389
> Fax: +61 2 935 17726
> Email: aloysius@physics.usyd.edu.au
> Web: http://www.physics.usyd.edu.au/~aloysius/
> ==============================================
> - CARPE DIEM -
> "Gather ye rosebuds while ye may,
> Old time is still a-flying,
> And this same flower that smiles today,
> To-morrow will be dying."
>                         ROBERT HERRICK
>                         1591-1674
> ==============================================


--
Pierre-Matthieu Anglade