Hi Fei, are you sure your Fermi level is correct?
--Zhen On Thu, Aug 6, 2015 at 11:43 PM, Gul Rahman <[email protected]> wrote: > 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, Pakistan > http://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, Pakistan > http://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 F4 > 1.0 > 4.0834274292 0.0000000000 0.0000000000 > 0.0000000000 4.0834274292 0.0000000000 > 0.0000000000 0.0000000000 4.0834274292 > Li F > 4 4 > Cartesian > 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 Rahman > I 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.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 4.02 Ang # Lattice constant alat > %block LatticeParameters # conventional cell > 1.00 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 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 0.0 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 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] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);>> > 写道: > > -----原始邮件----- > *发件人:* Gul Rahman <[email protected] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);> > > > *发送时间:* 2015年8月4日 星期二 > *收件人:* "[email protected] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);>" > <[email protected] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);> > > > *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, Pakistan > http://www.qau.edu.pk/profile.php?id=818020 > > > > On Tuesday, 4 August 2015, 14:21, 毛飞 <[email protected] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);>> > 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 Mao > University of South China, China > > > > > > 在2015-08-04，Mayuri Vaghela <[email protected] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);>> > 写道: > > -----原始邮件----- > *发件人:* Mayuri Vaghela <[email protected] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);> > > > *发送时间:* 2015年8月4日 星期二 > *收件人:* [email protected] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);> > *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] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);> > > 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 3* > gnuplot> 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] > <http://stu.mail.bnu.edu.cn/jy3/editor/java_script:_e(%7B%7D,%27cvml%27,%[email protected]%27);> > > 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 Mao > University of South China, China > > ############################ Input file for the LiF > ############################ > SystemName bulk.LiF > SystemLabel bulk_LiF > > NumberOfAtoms 8 > NumberOfSpecies 2 > > %block ChemicalSpeciesLabel > 1 3 Li > 2 9 F > %endblock ChemicalSpeciesLabel > > # basis > PAO.BasisSize DZP > PAO.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 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 0.0 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 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 California > Santa Barbara, CA 93106-5050 U.S.A. > Tel: +1-517-488-2298 > Email: [email protected] > [email protected] > > > > > > > -- Zhen Zhu, Postdoctoral Researcher Materials Department University of California Santa Barbara, CA 93106-5050 U.S.A. Tel: +1-517-488-2298 Email: [email protected] [email protected]
