This thread is locked.Only browsing is available.
Top Page > Browsing
Disagreement between bandgaps in band dispersion and DOS
Date: 2022/08/12 04:51
Name: ELIE ALBERT   <elio-physics@live.com>

Dear OPENMX users,

I am trying to obtain the band structure (BS) and DOS of a nanotube I am working on.

There is a mismatch in the bandgap values. From BS, the bandgap is clearly around 0.18eV, whereas the DOS is ~ 0 around the Fermi energy, with no apparent bandgap.

What could have gone wrong?

OBS: Since the nanotube is 1D, the band dispersion path is obligatorily defined as:

Band.Nkpath 1
<Band.kpath
600 0.0 0.0 0.0 0.0 0.0 0.5 G Z
Band.kpath>


Regards

Elio

メンテ
Page: [1]

Re: Disagreement between bandgaps in band dispersion and DOS ( No.1 )
Date: 2022/08/12 12:04
Name: T. Ozaki

Hi,

Did you take long cell vectors perpendicular to the nanotube axis so that the interaction with copied images can be avoided?
If you share your input file with us, we may be able to suggest definitely.

Regards,

TO
メンテ
Re: Disagreement between bandgaps in band dispersion and DOS ( No.2 )
Date: 2022/08/12 12:56
Name: ELIE ALBERT  <elio-physics@live.com>

Dear Prof. Ozaki,

Thanks for your reply.

I did take large unit vectors along the x and y directions. Find below part of the input file:

Atoms.UnitVectors.Unit Ang
<Atoms.UnitVectors
31.000000000000000 0.000000000000000 0.000000000000000
0.000000000000000 31.000000000000000 0.000000000000000
0.000000000000000 0.000000000000000 23.002099999999995
Atoms.UnitVectors>
#
# SCF or Electronic System
#

scf.XcType GGA-PBE # LDA|LSDA-CA|GGA-PBE

# DFT+U part #
scf.Hubbard.U Off # On|Off, default=off
#scf.Hubbard.Occupation dual # onsite|full|dual, default=dual
#scf.DFTU.Type 2 # 1:Simplified(Dudarev)|2:General, default=1
#scf.dc.Type cFLL # sFLL|sAMF|cFLL|cAMF, default=sFLL
scf.Slater.Ratio 0.625 # default=0.625
scf.Yukawa off # default=off
##############

scf.SpinPolarization off # On|Off
scf.SpinOrbit.Coupling off
scf.partialCoreCorrection On
scf.ElectronicTemperature 350.0 # default=300 (K) we probably need to converge
scf.energycutoff 230.0 # default=150 (Ry) we need to converge
scf.maxIter 6000 # default=40
scf.EigenvalueSolver band # Recursion|Cluster|Band
scf.Kgrid 1 1 4 # means 4x4x4 we need to converge
scf.Mixing.Type Rmm-Diis # Simple|Rmm-Diis|Gr-Pulay
scf.stress.tensor on # off|on
scf.Init.Mixing.Weight 0.70 # default=0.30
scf.Min.Mixing.Weight 0.01 # default=0.001
scf.Max.Mixing.Weight 0.10 # default=0.40
scf.Kerker.factor 10.0 # default=1.00
scf.Mixing.History 40 # default=5
scf.Mixing.StartPulay 40 # default=6
scf.criterion 2.0e-5 # default=1.0e-6 (Hartree)

#
# MD or Geometry Optimization
#

MD.Type Nomd # Nomd|Opt|DIIS|NVE|NVT_VS|NVT_NH
MD.applied.pressure 0.004 # Can be values other than 0 to simulate systems under pressure.
MD.maxIter 10000 # default=1
MD.Opt.StartDIIS 35
MD.TimeStep 0.05 # default=0.5 (fs)
MD.Opt.criterion 5e-3 # default=1.0e-4 (Hartree/bohr)

#
# Band dispersion
#
#

Voronoi.Charge off


Band.dispersion on # on|off, default=off
# if <Band.KPath.UnitCell does not exist,
# the reciprical lattice vector is employed.
Band.Nkpath 1
<Band.kpath
600 0.0 0.0 0.0 0.0 0.0 0.5 G Z
Band.kpath>

#
# MO output
#

MO.fileout off # on|off
num.HOMOs 1 # default=1
num.LUMOs 1 # default=1
MO.Nkpoint 1 # default=1
<MO.kpoint
0.0 0.0 0.0
MO.kpoint>

#
# DOS and LDOS
#

Dos.fileout on # on|off , default=off
Dos.Erange -7.0 7.0 # default = -20 20
Dos.Kgrid 1 1 8 # default = Kgrid1 Kgrid2 Kgrid3

#
# output Hamiltonian and overlap
#

HS.fileout off # on|off, default=off


Regards




メンテ
Re: Disagreement between bandgaps in band dispersion and DOS ( No.3 )
Date: 2022/08/12 14:35
Name: T. Ozaki

Hi,

Did you try the gaussian broadening method with a small width instead of the tetrahedron method?
If you see the gap using the gaussian broadening method, it indicates that the tetrahedron method failed to interporate
the eigenenergies properly. This problem tends to happen especially for low dimensional systems.
You may change the number of 8 in "Dos.Kgrid 1 1 8" to a larger number, and see what happens with the tetrahedron method.

Also, you can check the standard output to make sure that the system is really treated as an one-dimensional system in your calculation.
In this case, you must see the following message in the standard output.
<Check_System> The system is chain.

Regards,

TO
メンテ
Re: Disagreement between bandgaps in band dispersion and DOS ( No.4 )
Date: 2022/08/12 21:21
Name: ELIE ALBERT  <elio-physics@live.com>

Dear prof. Ozaki

Once again, thanks for the feedback.

The message "The system is a chain" is appearing in the std output.

Regarding the DOS calculation, I am actually using the Gaussian smearing with a width of 0.035. Is that too large?

I will try to use the tetrahedrom method and see what I will get.

Regards
メンテ
Re: Disagreement between bandgaps in band dispersion and DOS ( No.5 )
Date: 2022/08/12 22:52
Name: T. Ozaki

Hi,

If you show us the full input file, we will be able to give a more definite point.

Regards,

TO
メンテ
Re: Disagreement between bandgaps in band dispersion and DOS ( No.6 )
Date: 2022/08/13 04:16
Name: ELIE ALBERT  <elio-physics@live.com>

Dear Prof. Ozaki,

I think the problem is solved.

I have computed the DOS using different Gaussian broadening values, and I got the following:

0.0018 0.171eV
0.002 0.169eV
0.005 0.15eV
0.01 0.114eV
0.015 0.078eV
0.02 0.044eV
...

Since the 0.0018 broadening gives the closest value of the bandgap to that obtained by the BS calculation, does it mean we can adopt the corresponding DOS?

I have increased the DOS grid and there were no changes in the spectrum

Regards
メンテ

Page: [1]