Hi all,

My three cents on the subject.

Graphene needs a very dense in-plane sampling for a correct description of
its physical properties. It is likely that bilayer graphene does need a high
sampling as well. Since it's an isolated bilayer, it would be fine to use
just 1 MP division in the direction perpendicular to the graphene planes.
However, only one point is likely to be insufficient for catching the
electronic dispersion close to K. My suggestion: try different MP samplings
for the supercell, to check that the forces and electronic structure are
indeed converged. Only then, using this sampling as the minimum one, try to
calculate the phonon dispersion. From the input sent to the list, it seems
to me that only gamma-point MP sampling is being used in the calculation,
and this surely makes a big difference.

Cheers,

Marcos

On Mon, Oct 10, 2011 at 6:03 PM, <[email protected]> wrote:

> > Dear Postnikov,
> >        Thanks for your help ! As your advice, I changed the supercell (3
> 3
> > 0) into (4 4 0), however, new problem occured. I tried many
> > parameters, but there was always the same error:
> >           "+  Segmentation fault "
> >         I don't konw what this error means and how to eliminate it .
>
> Dear Liu,
> this message most probably hints that your system is
> too big to fit into memory.
> Note that you don't need to use the same the same multiplication
> numbers along both directions. Increase just one in order
> to scan the dispersion along one direction.
>
> Did you check that your Gamma phonons are OK,
> in order to be sure what MeshCutoff you need,
> and not taking it excessively high.
>
> Best regards
>
> Andrei
>
> >
> > sincerely
> > Liu
> >
> > ____________________________________________________
> > NumberOfSpecies 1
> > %block chemicalspecieslabel
> > 1 6 C
> > %endblock chemicalspecieslabel
> > PAO.BasisSize DZP
> > MeshCutoff   500.0 Ry/ 300Ry / 200 Ry / 100Ry
> > DM.MixingWeight  0.001 / 0.01/ 0.0001
> > DM.NumberPulay    3
> > DM.Tolerance     1.0d-4
> > %include FC.fdf
> >  NumberOfAtoms        4
> > LatticeConstant      2.45951215  Ang
> > %block LatticeVectors
> > 0.8660254   0.5      0.0
> > 0.8660254  -0.5      0.0
> > 0.0         0.0      8.0
> > %endblock LatticeVectors
> > AtomicCoordinatesFormat   NotScaledCartesianAng
> > %block AtomicCoordinatesAndAtomicSpecies
> >   -0.000007    0.000017   -0.127708   1    12.01
> >    1.419987    0.000012   -0.127709   1    12.01
> >   -0.000021   -0.000010    3.725312   1    12.01
> >    1.419990   -0.000005    3.725319   1    12.01
> > %endblock AtomicCoordinatesAndAtomicSpecies
> > SuperCell_1          4
> > SuperCell_2          4
> > SuperCell_3          0
> > AtomicDispl          0.2  Bohr
> > BandLinesScale       pi/a
> > %block BandLines
> > 1  0.000      0.000        0.000  \Gamma
> > 173  1.1547005  0.000        0.000       M
> > 100  1.1547005  0.666666667  0.000       K
> > 200  0.000      0.000        0.000  \Gamma
> > %endblock BandLines
> >
> > _________________________________________________
> > My outfile also attached. Thanks for your help again!
> >
> >
> >
> >
> >
> >
> > At 2011-10-10 04:38:35,[email protected] wrote:
> >>> Dear siesta users,
> >>>        As we konw that there should be no negative value in phonon
> >>> spectrum, but when I tried to draw the phonon spectrum of
> >>> bilayer-graphene, there are always some negative intervals. I'am
> >>> very confused about this.
> >>
> >>Dear Liu,
> >>
> >>such kind of "noise" in the vicinity of Gamma appears
> >>when the force constants are not enough converged in the real space
> >>(with distance to neighbors).
> >>You use the supercell (3 3 0); will (4 4 0) be affordable?
> >>
> >>However! - the supercell size will affect just the vicinity of Gamma,
> >>but NOT Gamma itself - which ought to be good even in (0 0 0)
> >>supercell. I cannot judge from your plot whether you have
> >>an imaginary frequency also at Gamma, or it is just traced so
> >>through many points. If you have a problem at Gamma,
> >>then you should fix it first, staying with a (0 0 0) supercell.
> >>It must be due to insufficient MeshCutoff
> >>(170 Ry you use is not that high).
> >>
> >>Best regards
> >>
> >>Andrei Postnikov
> >>
> >>
> >>
> >>
> >>>        My program attached.
> >>>
> >>> Best regards
> >>> Liu
> >>>  _____________________________________________________________-
> >>>
> >>> NumberOfSpecies 1
> >>>
> >>>
> >>>
> >>> %block chemicalspecieslabel
> >>>
> >>> 1 6 C
> >>>
> >>> %endblock chemicalspecieslabel
> >>>
> >>>
> >>>
> >>> PAO.BasisSize DZP
> >>>
> >>>
> >>>
> >>> MeshCutoff 170.0 Ry
> >>>
> >>> DM.MixingWeight 0.001 /0.01/0.1
> >>>
> >>> DM.NumberPulay   3
> >>>
> >>> DM.Tolerance    1.0d-5
> >>>
> >>> %include FC.fdf
> >>>
> >>> NumberOfAtoms       4
> >>>
> >>> LatticeConstant     2.45951215 Ang
> >>>
> >>>
> >>>
> >>> %block LatticeVectors
> >>>
> >>> 0.8660254  0.5     0.0
> >>>
> >>> 0.8660254 -0.5     0.0
> >>>
> >>> 0.0        0.0     8.0
> >>>
> >>> %endblock LatticeVectors
> >>>
> >>>
> >>>
> >>> AtomicCoordinatesFormat  NotScaledCartesianAng
> >>>
> >>>
> >>>
> >>> %block AtomicCoordinatesAndAtomicSpecies
> >>>
> >>>  -0.000007   0.000017  -0.127708  1   12.01
> >>>
> >>>   1.419987   0.000012  -0.127709  1   12.01
> >>>
> >>>  -0.000021  -0.000010   3.725312  1   12.01
> >>>
> >>>   1.419990  -0.000005   3.725319  1   12.01
> >>>
> >>> %endblock AtomicCoordinatesAndAtomicSpecies
> >>>
> >>>
> >>>
> >>> SuperCell_1         3
> >>>
> >>> SuperCell_2         3
> >>>
> >>> SuperCell_3         0
> >>>
> >>>
> >>>
> >>> AtomicDispl         0.2 Bohr
> >>>
> >>>
> >>>
> >>> BandLinesScale      pi/a
> >>>
> >>> %block BandLines
> >>>
> >>> 1 0.000     0.000       0.000 \Gamma
> >>>
> >>> 173 1.1547005 0.000       0.000      M
> >>>
> >>> 100 1.1547005 0.666666667 0.000      K
> >>>
> >>> 200 0.000     0.000       0.000 \Gamma
> >>>
> >>> %endblock BandLines
> >>>
> >>>
> >>
> >
>
>

Responder a