Are you calculating the bandgap at the Gamma Point?It is around 8.40 eV at
Gamma point (my test calculation). Sure, this will be different at the other
symmetry points in the BZ.Gul Dr. Gul Rahman
Assistant Professor,
Department of Physics,
Quaid-i-Azam University,
Islamabad, Pakistanhttp://www.qau.edu.pk/profile.php?id=818020
On Thursday, 6 August 2015, 16:36, 毛飞 <[email protected]> wrote:
Dear Zhen and Rahman Thank you for your useful suggestions. Please excuse
me to bother you again. I calculate the band structure of LiF again, and I
think I have already corrected the errors in the new input file, but the result
is still unreasonable, the band gap of LiF is 15.8 eV (it is too large). You
are right, the structure of LiF is the same with NaCl. I define the primitive
cell as the simulation cell (see block LatticeVectors) in the input file.
Regards, Fei #################### Input file #######################
SystemName bulk.LiF SystemLabel bulk_LiF NumberOfAtoms
2 NumberOfSpecies 2 %block ChemicalSpeciesLabel 1 3 Li 2 9
F %endblock ChemicalSpeciesLabel # basis PAO.BasisSize DZP
PAO.EnergyShift 200 meV LatticeConstant 4.08 Ang # Lattice
constant alat %block LatticeVectors # primitive cell 0.000
0.500 0.500 0.500 0.000 0.500 0.500 0.500 0.000 %endblock
LatticeVectors %block kgrid_Monkhorst_Pack 10 0 0 0.0 0 10 0 0.0 0 0
10 0.0 %endblock kgrid_Monkhorst_Pack BandLinesScale
ReciprocalLatticeVectors %block BandLines 1 0.0 0.0 0.0 #
Gamma-point 20 0.5 0.0 0.5 # X-point 30 0.5 0.25 0.75 #
W-point 30 0.375 0.375 0.75 # K-point 30 0.0 0.0 0.0 #
Gamma-point %endblock BandLines %block ProjectedDensityOfStates -20 10
0.1 300 eV %endblock ProjectedDensityOfStates MeshCutoff 300.0
Ry # SCF options MaxSCFIterations 100 # Maximum number of SCF
iter DM.NumberPulay 3 # One Pulay every 3 iter
DM.MixingWeight 0.35 # New DM amount for next SCF cycle
DM.Tolerance 5.d-4 # Tolerance in maximum difference
SolutionMethod diagon # OrderN or Diagon OccupationFunction FD
ElectronicTemperature 25 meV # Temp. for Fermi smearing #
exchange-correlation functional XC.functional LDA XC.authors CA #
structural infomation AtomCoorFormatOut Ang # output information
WriteCoorXmol T WriteMDXmol T WriteForces T WriteKpoints .true.
WriteEigenvalues .true. WriteKbands .true. WriteBands
.true. WriteMullikenPop 1 # options for MD MD.MaxForceTol
0.01 eV/Ang MD.TypeOfRun CG MD.NumCGSteps 400 MD.Broyden.History.Steps 6
MD.Broyden.Initial.Inverse.Jacobian 1.0 # options for continue #
MD.UseSaveCG T # neccessary! # MD.UseSaveXV T
# neccessary! # DM.UseSaveDM true # to use
continuae..ion files AtomicCoordinatesFormat ScaledCartesian # Format for
coordinates AtomicCoorFormatOut Ang %block
AtomicCoordinatesAndAtomicSpecies 0.00000 0.00000
0.00000 1 0.50000 0.50000 0.50000 2
%endblock AtomicCoordinatesAndAtomicSpecies
在2015-08-05，Gul Rahman <[email protected]> 写道:
-----原始邮件-----
发件人: Gul Rahman <[email protected]>
发送时间: 2015年8月5日 星期三
收件人: "[email protected]" <[email protected]>
主题: Re: [SIESTA-L] The energy band gap of LiF
Hello,As suggested by Dr. Zhen that your unit cell is wrong. In your input
file you are considering a BCC lattice using the conventional unit cell. I
guess you want to use the primitive cell of LiF, but you have used wrong
lattice vectors for the primitive cell of LiF. I dont know the exact structure
of LiF, but I guess it is NaCl-type structure. If it is NaCl-type then you can
easily use the primitive lattice vectors of FCC lattice, and use the fractional
coordinates (only two atoms). You can also use the conventional unit cell as
given by Zen, but it will take longer computational time, which we always want
to reduce:)Plz, always visualise your structure before calculating its
properties, e.g., band, optical, etc.
I hope it will help.
Good Luck for LiF.Gul
Dr. Gul Rahman
Assistant Professor,
Department of Physics,
Quaid-i-Azam University,
Islamabad, Pakistanhttp://www.qau.edu.pk/profile.php?id=818020
On Wednesday, 5 August 2015, 10:07, Zhen Zhu <[email protected]> wrote:
Hi Fei,
your unit cell is wrong.
Here is the correct conventional unit cell:Li4 F41.0 4.0834274292
0.0000000000 0.0000000000 0.0000000000 4.0834274292
0.0000000000 0.0000000000 0.0000000000 4.0834274292
Li F 4 4Cartesian 0.000000000 0.000000000
0.000000000 0.000000000 2.041713715 2.041713715
2.041713715 0.000000000 2.041713715 2.041713715
2.041713715 0.000000000 0.000000000 2.041713715
0.000000000 0.000000000 0.000000000 2.041713715
2.041713715 2.041713715 2.041713715 2.041713715
0.000000000 0.000000000
Good luck--Zhen
On Tue, Aug 4, 2015 at 9:59 PM, 毛飞 <[email protected]> wrote:
>Dear Khan and RahmanI calculate the band gap of LiF with the corrected input
>file, but the energy gap is still very large, is about 18 eV (see attachment).
>The input file is attached below. Any comment or suggestions?Regards, Fei
>################### Input file #####################SystemName
>bulk.LiFSystemLabel bulk_LiF NumberOfAtoms 2NumberOfSpecies
>2 %block ChemicalSpeciesLabel 1 3 Li 2 9 F%endblock
>ChemicalSpeciesLabel # basis PAO.BasisSize DZPPAO.EnergyShift 200
>meV LatticeConstant 4.02 Ang # Lattice constant alat%block
>LatticeParameters # conventional cell1.00 1.00 1.00 90.0 90.0 90.0
>%endblock LatticeParameters %block kgrid_Monkhorst_Pack 10 0 0 0.0 0 10 0
>0.0 0 0 10 0.0%endblock kgrid_Monkhorst_Pack BandLinesScale
>ReciprocalLatticeVectors%block BandLines1 0.0 0.0 0.0 #
>Gamma-point20 0.5 0.0 0.5 # X-point30 0.5 0.25 0.75 #
>W-point30 0.0 0.0 0.0 # Gamma-point%endblock BandLines %block
>ProjectedDensityOfStates -20 10 0.1 300 eV%endblock
>ProjectedDensityOfStates MeshCutoff 300.0 Ry # SCF
>optionsMaxSCFIterations 100 # Maximum number of SCF
>iterDM.NumberPulay 3 # One Pulay every 3
>iterDM.MixingWeight 0.35 # New DM amount for next SCF
>cycleDM.Tolerance 5.d-4 # Tolerance in maximum difference
>SolutionMethod diagon # OrderN or DiagonOccupationFunction
>FDElectronicTemperature 25 meV # Temp. for Fermi smearing #
>exchange-correlation functionalXC.functional LDAXC.authors CA # structural
>infomationAtomCoorFormatOut Ang # output informationWriteCoorXmol TWriteMDXmol
>TWriteForces T WriteKpoints .true.WriteEigenvalues
>.true.WriteKbands .true.WriteBands
>.true.WriteMullikenPop 1 # options for MDMD.MaxForceTol 0.01
>eV/AngMD.TypeOfRun CGMD.NumCGSteps 400MD.Broyden.History.Steps
>6MD.Broyden.Initial.Inverse.Jacobian 1.0 # options for continue# MD.UseSaveCG
>T # neccessary!# MD.UseSaveXV T
># neccessary!# DM.UseSaveDM true # to use continuae..ion
>files AtomicCoordinatesFormat Fractional %block
>AtomicCoordinatesAndAtomicSpecies # Number of irreducible atoms in the
>conventional cell is 2 0.0000000 0.0000000 0.0000000
>1 Li 0.5000000 0.5000000 0.5000000 2
>F%endblock AtomicCoordinatesAndAtomicSpecies
在2015-08-05，Younas Khan <[email protected]> 写道:
-----原始邮件-----
发件人: Younas Khan <[email protected]>
发送时间: 2015年8月5日 星期三
收件人: "[email protected]" <[email protected]>
主题: Re: [SIESTA-L] The energy band gap of LiF
I think you have incorrectly written the lattice vectors. Kindly check it, they
may not be in proper format.
Younas Khan
On Tuesday, August 4, 2015, 毛飞 <[email protected]> wrote:
Dear Rahman Thank you for your reply, I will try the calculation with your
suggestions. Fei
在2015-08-04，Gul Rahman <[email protected]> 写道:
-----原始邮件-----
发件人: Gul Rahman <[email protected]>
发送时间: 2015年8月4日 星期二
收件人: "[email protected]" <[email protected]>
主题: Re: [SIESTA-L] The energy band gap of LiF
Dear Fei Mao,Yes, your band structure is strange, it has flat band (like
molecules) and the band gap is very large, which is wrong.Plz, look carefully
into your input file.First, try to reproduce the band structure and lattice
parameters of bulk (primitive) LiF, then you can generate the SuperCell either
using your pencil in hands (:) or using the SIESTA or any other computer code.
And then put the Atomic coordinates of the Supercell explicitly in your fdf
file. You can see you have defined SuperCell (2x2x2) in your fdf file.Also, try
to use fractional coordinates if you are dealing with a bulk system. Whenever
you generate .XV file, you must visualise it to see what's going on. Also,
your k-points sampling is not good for the bulk LiF system. You must use dense
k-points. I will not use 1x1x1 for bulk. I have to converge it.I hope it may
help.Regards,Gul
Dr. Gul Rahman
Assistant Professor,
Department of Physics,
Quaid-i-Azam University,
Islamabad, Pakistanhttp://www.qau.edu.pk/profile.php?id=818020
On Tuesday, 4 August 2015, 14:21, 毛飞 <[email protected]> wrote:
Dear Rahman and Vaghela Thank you for kind replies.The Fermi energy level is at
-5.9 eV in the calculation, I modify the fig and mark the Fermi energy by a
black line (which is attached). Now, I am puzzled by two points in this fig. 1.
As shown in the fig, the conduction and valence band are both composed by
many straight lines, it seems to me very strange. Generally, these energy
levels evolves smoothly with changes in k, forming a smooth band of states, and
there are conduction band bottom and valence band top. However, these typical
characters are absence in the fig. 2. It can be seen from the fig that,
the energy gap of LiF is about 16 eV (the experimental one is 14 eV), but the
LDA calculations usually underestimate the gap. Any comment is appreciated,
and thank you in advance.Sincerely Fei MaoUniversity of South China, China
在2015-08-04，Mayuri Vaghela <[email protected]> 写道:
-----原始邮件-----
发件人: Mayuri Vaghela <[email protected]>
发送时间: 2015年8月4日 星期二
收件人: [email protected]
主题: Re: [SIESTA-L] The energy band gap of LiF
first check systemlabel.EIG file,
you will get fermi energy from that file.
now draw clear line using set axis command of gnuplot at appropriate energy
value
from that you can distinguish band gap between CB &VB...
On Tue, Aug 4, 2015 at 10:47 AM, Altaf Ur Rahman <[email protected]> wrote:
Dear when you calculate the band structure by using like gnuplot you must see
the fermi energy from your band data file/EIG file and then when you plot the
you can write in the gnuplot this sentence set zero axis 3gnuplot> plot
'./Band.data' u ($1)+3.23:2 w l here in above case your fermi energy is 3.23
and you can see the fermi line will appear that clearly distiguish the
conduction and valance band
On Mon, Aug 3, 2015 at 6:04 PM, 毛飞 <[email protected]> wrote:
Dear siesta developers and users I am a newbie in band structure calculation. I
calculate the band structure of LiF, and a 2*2*2 supercell of 64 atoms is
selected. In this calculation, the local density approximation (LDA) is adopted
for the exchange-correlation functional term, the band structure is obtained
along the high symmetric line Γ->X->W->Γ, only gamma point is used for
Brillouin zone sampling, more details can be found in the input file which is
attached below. I can not understand the results of the E(k) (two figs are also
attached), it seems there are two band gaps formed in the calculation (the
Fermi energy is shifted to 0 eV), I want to know which one (gap 1 or gap 2) is
the band gap in my calculation, and how does another gap come from? Is there
anything wrong in the input file? Please help me, any suggestion or comment is
appreciated. Sincerely Fei MaoUniversity of South China, China
############################ Input file for the LiF
############################SystemName bulk.LiFSystemLabel
bulk_LiF NumberOfAtoms 8NumberOfSpecies 2 %block ChemicalSpeciesLabel
1 3 Li 2 9 F%endblock ChemicalSpeciesLabel # basis PAO.BasisSize
DZPPAO.EnergyShift 200 meV LatticeConstant 4.04 Ang # Lattice
constant alat%block LatticeVectors # Lattice vectors, in units of
latticeconstant 2.0 0.0 0.0 0.0 2.0 0.0 0.0 0.0
2.0 %endblock LatticeVectors %block SuperCell 2 0 0 0.0 0 2 0 0.0
0 0 2 0.0%endblock SuperCell %block kgrid_Monkhorst_Pack 1 0 0 0.0 0 1 0
0.0 0 0 1 0.0%endblock kgrid_Monkhorst_Pack BandLinesScale
ReciprocalLatticeVectors%block BandLines1 0.0 0.0 0.0 #
Gamma-point20 0.5 0.0 0.5 # X-point30 0.5 0.25 0.75 #
W-point30 0.0 0.0 0.0 # Gamma-point%endblock BandLines %block
ProjectedDensityOfStates -20 10 0.1 300 eV%endblock
ProjectedDensityOfStates MeshCutoff 300.0 Ry # SCF
optionsMaxSCFIterations 100 # Maximum number of SCF
iterDM.NumberPulay 3 # One Pulay every 3 iterDM.MixingWeight
0.35 # New DM amount for next SCF cycleDM.Tolerance
5.d-4 # Tolerance in maximum difference SolutionMethod diagon
# OrderN or DiagonOccupationFunction FDElectronicTemperature 25 meV
# Temp. for Fermi smearing # exchange-correlation functionalXC.functional
LDAXC.authors CA # structural infomationAtomCoorFormatOut Ang # output
informationWriteCoorXmol TWriteMDXmol TWriteForces T WriteKpoints
.true.WriteEigenvalues .true.WriteKbands .true.WriteBands
.true.WriteMullikenPop 1 # options for MDMD.MaxForceTol
0.01 eV/AngMD.TypeOfRun CGMD.NumCGSteps 400MD.Broyden.History.Steps
6MD.Broyden.Initial.Inverse.Jacobian 1.0 # options for continue# MD.UseSaveCG T
# neccessary!# MD.UseSaveXV T #
neccessary!# DM.UseSaveDM true # to use continuae..ion
files AtomicCoordinatesFormat Ang%block AtomicCoordinatesAndAtomicSpecies
-5.963747 -2.022693 -2.026392 1 -6.051639
-5.949852 -2.013945 2 -2.147364 -2.056847
-2.009531 2 -2.039835 -5.953723 -2.024337
1 -6.085698 -1.925561 2.008577 2 -5.941415
-5.910826 2.052174 1 -2.040599 -2.032355
2.028040 1 -2.129653 -6.096888 2.008118
2%endblock AtomicCoordinatesAndAtomicSpecies
--
Zhen Zhu, Postdoctoral Researcher
Materials Department
University of CaliforniaSanta Barbara, CA 93106-5050 U.S.A.
Tel: +1-517-488-2298Email: [email protected]
[email protected]
