[ase-users] CaF2 Bandstructure Plot

[Ask Hjorth Larsen]

Hi,

2018-08-01 5:54 GMT-05:00 Christoph Bernkopf <christoph.bernkopf at cern.ch>:
>
> Hi Ask,
>
> I have an update for you. After updating ASE (3.16.2) and GPAW (1.4.0) I ran the CaF2 files again. I’m getting the following special points:
>
> "G": [0, 0, 0],
> "K": [0.375, 0.375, 0.75],
> "L": [0.5, 0.5, 0.5],
> "U": [0.625, 0.25, 0.625],
> "W": [0.5, 0.25, 0.75],
> "X": [0.5, 0, 0.5],
>
> The bs plot looks worse than before but the line: special_points = kpoints.get_special_points(atoms.cell) works now. Do you know how I can fix the bandstructure plot?
>
> Attached is the plot and the code.

The code still uses calc.set(kpts=(20, 20, 7)).

So you still use the old kpoints that do not form any path.  Use
kpts=kpts, so you actually pass the newly defined kpoint path to the
calculator.

Best regards
Ask

>
> Cheers,
> Chris
>
>
>
>
>
> > On 26.07.2018, at 16:22, Ask Hjorth Larsen <asklarsen at gmail.com> wrote:
> >
> > 2018-07-26 16:17 GMT+02:00 Christoph Bernkopf <christoph.bernkopf at cern.ch>:
> >> Hi Ask,
> >>
> >> Thank you for your help. I’m looking to what you find out!
> >>
> >> Just as a quick note: Using fcc produces an error as well: (AttributeError:
> >> 'NoneType' object has no attribute ‘kd’)
> >
> > This probably works better with newer versions (both ASE and GPAW).
> > Also remember atoms.get_potential_energy() before getting the band
> > structure.
> >
> > Best regards
> > Ask
> >
> >>
> >> Cheers,
> >> Chris
> >>
> >>
> >> On 26.07.2018, at 14:29, Ask Hjorth Larsen <asklarsen at gmail.com> wrote:
> >>
> >> Hi,
> >>
> >> 2018-07-25 18:55 GMT+02:00 Christoph Bernkopf <christoph.bernkopf at cern.ch>:
> >>
> >> Hi Ask,
> >>
> >> Thank you for your reply!
> >>
> >> I looked up CaF2 on the materials project
> >> (https://materialsproject.org/materials/mp-2741/#). I converted the cif file
> >> provided by the materials project using the ase-gui (command: “ase-gui
> >> CaF2.cif -o CaF2.py”). I ran the code using the converted cif file but I’m
> >> still getting errors.
> >>
> >> When I’m omitting ‘cubic’ I get the following error:
> >> special_points = kpoints.get_special_points(atoms.cell)
> >> TypeError: get_special_points() takes at least 2 arguments (1 given)
> >>
> >> When I use 'cubic’ I get the following error:
> >>   assert abc.ptp() < eps and abs(angles - pi / 2).max() < eps
> >> AssertionError
> >> Traceback (most recent call last):
> >> File "bandstructure_caf2.py", line 58, in <module>
> >>   special_points = kpoints.get_special_points('cubic', atoms.cell)
> >>
> >>
> >> 'cubic' will only work on a cell which is literally cubic.  The
> >> reduced cell is not, which is the problem here.
> >>
> >> The reduced cell is considered an 'fcc' type in ASE but obviously the
> >> structure is not strictly fcc because only the Ca atoms fall on the
> >> actual fcc sites.
> >>
> >> The question now is how to solve this, in the sense of getting an
> >> appropriate BZ path.  Though one can always specify a path manually.
> >>
> >> I think the current determination of default BZ paths is probably
> >> deficient in that it assumes that it's working with simple structures.
> >> I will try to have a look at this.
> >>
> >> Best regards
> >> Ask
> >>
> >>
> >> My file is attached.
> >>
> >> Cheers,
> >> Chris
> >>
> >>
> >>
> >>
> >> On 24.07.2018, at 15:24, Ask Hjorth Larsen <asklarsen at gmail.com> wrote:
> >>
> >> Hi Christoph, (please reply to list)
> >>
> >> 2018-07-24 15:03 GMT+02:00 Christoph Bernkopf <christoph.bernkopf at cern.ch>:
> >>
> >> Hi Ask,
> >>
> >> Thank you for your quick answer! I tried out your suggestion:
> >>
> >> (CaF2 is a cubic crystal system, cF12)
> >>
> >> special_points = kpoints.get_special_points('cubic', atoms.cell)
> >> print("\nSpecial Points: ", special_points)  # decide some nice points
> >> bandpath = 'GWUXL'
> >> kpts, xcoords, labels = kpoints.bandpath(bandpath, atoms.cell, 100)
> >>
> >> I get the following error in the first line:
> >>  assert abc.ptp() < eps and abs(angles - pi / 2).max() < eps
> >> AssertionError
> >>
> >> Do you know what I’m doing wrong? The whole file is attached.
> >>
> >>
> >> Are you quite sure the structure is correct?  Try visualizing it and
> >> repeating it e.g. in the GUI.
> >>
> >> It should not be necessary to specify that you want special points for
> >> a cubic cell.  It should automatically figure out which cell it is.
> >>
> >> Best regards
> >> Ask
> >>
> >>
> >> Cheers,
> >> Christoph
> >>
> >>
> >>
> >>
> >> On 24.07.2018, at 14:14, Ask Hjorth Larsen <asklarsen at gmail.com> wrote:
> >>
> >> Hi,
> >>
> >> 2018-07-24 12:24 GMT+02:00 Christoph Bernkopf via ase-users
> >> <ase-users at listserv.fysik.dtu.dk>:
> >>
> >>
> >> Hi All,
> >>
> >> I’m currently using a CaF2 model to calculate the band structure (python
> >> code attached). My output is attached also (caf2_bandstructure.png). It
> >> looks wrong to me as the output usually has 4-5 entries on the X axis
> >> (attached example: bandstructure_si.png). Also the labels are missing in my
> >> graphic. Does anybody know what I’m doing wrong here?
> >>
> >>
> >> You should specify a good path for the band structure.  Right now it
> >> only has a Monkhorst-Pack sampling.
> >>
> >> import ase.dft.kpoints as kpoints
> >> special_points = kpoints.get_special_points(atoms.cell)
> >> print(special_points)  # decide some nice points
> >> banpath = 'GWUXL'
> >> kpts, xcoords, labels = kpoints.bandpath(path, atoms.cell, 100)
> >>
> >> calc.set(kpts=kpts, ...)
> >>
> >> Best regards
> >> Ask
> >>
> >> Cheers and thanks for the help,
> >> Chris
> >>
> >>
> >>
> >>
> >> _______________________________________________
> >> ase-users mailing list
> >> ase-users at listserv.fysik.dtu.dk
> >> https://listserv.fysik.dtu.dk/mailman/listinfo/ase-users
> >>
> >>
>