Mismatch in wannier band and SCF band

[leedairo]

Dear all.

Hello.

If wannier interpolation is correct, should wannier interpolated band and DFT band overlap exactly?

In my calculation,
(I think) Wannier spread is converged and show reasonalbe value, but there is difference between wannier band and DFT band nearby CBM and VBM.

How can I increase accuracy of wannier band structure?
I included my band structure, epw input, and wannier output.

Thank you!!

https://drive.google.com/file/d/0ByhlkG ... sp=sharing

Code: Select all

 Cycle:   2000
  WF centre and spread    1  (  0.300343,  0.520147,  7.495584 )     1.42081938
  WF centre and spread    2  (  0.300350, -0.520208,  7.495669 )     1.42083881
  WF centre and spread    3  ( -0.600613, -0.000022,  7.495548 )     1.42081257
  WF centre and spread    4  ( -0.000068,  0.000070,  8.306052 )     1.55350147
  WF centre and spread    5  (  0.300309,  0.520209, 12.504366 )     1.42077011
  WF centre and spread    6  (  0.300301, -0.520145, 12.504457 )     1.42075008
  WF centre and spread    7  (  0.000057, -0.000074, 11.693905 )     1.55345494
  WF centre and spread    8  ( -0.600680,  0.000023, 12.504350 )     1.42077233
  WF centre and spread    9  (  2.776741,  0.000027,  9.058750 )     1.73777614
  WF centre and spread   10  (  2.033985, -0.000046,  8.062220 )     1.92072116
  WF centre and spread   11  (  1.662502,  0.643363,  9.058605 )     1.73783671
  WF centre and spread   12  (  1.662485, -0.643267,  9.058721 )     1.73778923
  WF centre and spread   13  (  2.033904,  0.000048, 11.937748 )     1.92066638
  WF centre and spread   14  (  2.776794, -0.000027, 10.941312 )     1.73781320
  WF centre and spread   15  (  1.662563,  0.643244, 10.941167 )     1.73775231
  WF centre and spread   16  (  1.662580, -0.643342, 10.941285 )     1.73780139
  Sum of centres and spreads ( 16.271553,  0.000000,159.999738 )    25.89987621

   2000    -0.675E-13     0.0000018061       25.8998762185     146.46  <-- CONV
        O_D=      0.0925681 O_OD=      2.8481559 O_TOT=     25.8998762 <-- SPRD
 Delta: O_D=  0.9020562E-15 O_OD= -0.6616929E-13 O_TOT= -0.6750156E-13 <-- DLTA
 ------------------------------------------------------------------------------

 Writing checkpoint file P2Si2-aa.chk... done
 Final State
  WF centre and spread    1  (  0.300343,  0.520147,  7.495584 )     1.42081938
  WF centre and spread    2  (  0.300350, -0.520208,  7.495669 )     1.42083881
  WF centre and spread    3  ( -0.600613, -0.000022,  7.495548 )     1.42081257
  WF centre and spread    4  ( -0.000068,  0.000070,  8.306052 )     1.55350147
  WF centre and spread    5  (  0.300309,  0.520209, 12.504366 )     1.42077011
  WF centre and spread    6  (  0.300301, -0.520145, 12.504457 )     1.42075008
  WF centre and spread    7  (  0.000057, -0.000074, 11.693905 )     1.55345494
  WF centre and spread    8  ( -0.600680,  0.000023, 12.504350 )     1.42077233
  WF centre and spread    9  (  2.776741,  0.000027,  9.058750 )     1.73777614
  WF centre and spread   10  (  2.033985, -0.000046,  8.062220 )     1.92072116
  WF centre and spread   11  (  1.662502,  0.643363,  9.058605 )     1.73783671
  WF centre and spread   12  (  1.662485, -0.643267,  9.058721 )     1.73778923
  WF centre and spread   13  (  2.033904,  0.000048, 11.937748 )     1.92066638
  WF centre and spread   14  (  2.776794, -0.000027, 10.941312 )     1.73781320
  WF centre and spread   15  (  1.662563,  0.643244, 10.941167 )     1.73775231
  WF centre and spread   16  (  1.662580, -0.643342, 10.941285 )     1.73780139
  Sum of centres and spreads ( 16.271553,  0.000000,159.999738 )    25.89987621

         Spreads (Ang^2)       Omega I      =    22.959152194
        ================       Omega D      =     0.092568075
                               Omega OD     =     2.848155949
    Final Spread (Ang^2)       Omega Total  =    25.899876218

Code: Select all

--
&inputepw
   prefix      = 'P2Si2-aa'
   amass(1)    = 30.974
   amass(2)    = 28.085
   outdir      = './'

   system_2d = .true.

   iverbosity  = 1

   elph        = .true.
   epbwrite    = .true.
   epbread     = .false.

  epwwrite    = .true.
  epwread     = .false.

  etf_mem     = .true.

   nbndsub     =  16
   nbndskip    =  0

   wannierize  = .true.
   num_iter    = 2000
   iprint      = 2
   proj(1)     = 'P:sp3'
   proj(2)     = 'Si:sp3'

wdata(1) = 'bands_plot = .true.'
wdata(2) = 'begin kpoint_path'
wdata(3) = 'G 0.00 0.00 0.00 X 0.50 0.50 0.00'
wdata(4) = 'X 0.50 0.50 0.00 M 0.666666667 0.333333333 0.00'
wdata(5) = 'M 0.666666667 0.333333333 0.00 G 0.00 0.00 0.00'
wdata(8) = 'end kpoint_path'
wdata(9) = 'bands_plot_format = gnuplot'
wdata(10) = 'guiding_centres = .true.'
wdata(11) = 'dis_num_iter = 2000'
wdata(12) = 'num_print_cycles = 10'
wdata(13) = 'dis_mix_ratio = 1.0'

  elecselfen  = .true.
  phonselfen  = .false.
  a2f         = .false.

  parallel_k  = .true.
  parallel_q  = .false.

  fsthick     = 15 ! eV 
  eptemp      = 300 ! K 
  degaussw    = 0.01! eV
!  degaussq    = 0.05 !meV

  dvscf_dir   = './save'
  filukk      = './data.ukk'
  filkf       = './mesh.dat'

!  efermi_read = .true.
!  fermi_energy = -1.5283


!  nkf1        = 100
!  nkf2        = 100
!  nkf3        = 1

  nqf1        = 100
  nqf2        = 100
  nqf3        = 1
  
  nk1         = 10
  nk2         = 10
  nk3         = 1

  nq1         = 10
  nq2         = 10
  nq3         = 1
 /
      14 cartesian
   0.000000000000000E+00   0.000000000000000E+00   0.000000000000000E+00   0.0200000
   0.577350269209576E-01   0.999999999896335E-01   0.000000000000000E+00   0.1200000
   0.115470053841915E+00   0.199999999979267E+00   0.000000000000000E+00   0.1200000
   0.173205080762873E+00   0.299999999968900E+00   0.000000000000000E+00   0.1200000
   0.230940107683830E+00   0.399999999958534E+00   0.000000000000000E+00   0.1200000
  -0.288675134604788E+00  -0.499999999948167E+00   0.000000000000000E+00   0.0600000
   0.173205080762873E+00   0.999999999896335E-01   0.000000000000000E+00   0.1200000
   0.230940107683830E+00   0.199999999979267E+00   0.000000000000000E+00   0.2400000
   0.288675134604788E+00   0.299999999968900E+00   0.000000000000000E+00   0.2400000
  -0.230940107683830E+00  -0.599999999937801E+00   0.000000000000000E+00   0.2400000
   0.346410161525746E+00   0.199999999979267E+00   0.000000000000000E+00   0.1200000
  -0.173205080762873E+00  -0.699999999927434E+00   0.000000000000000E+00   0.2400000
  -0.115470053841915E+00  -0.599999999937801E+00   0.000000000000000E+00   0.1200000
  -0.577350269209576E-01  -0.699999999927434E+00   0.000000000000000E+00   0.1200000

[sponce]

Hello,

Indeed, small spread does not necessarily mean you are ok.
Large spread usually means you are not ok.

In this case, it is indeed not very good.

Some suggestions:
- exclude the last two (or even 4) bands
- you can increase the initial k-point grid. This will force Wannier to go through more points
- You can play with other variables from Wannier90 to improve it. See the Wannier90 website (user guide and tutorial).

Best,
Samuel

[leedairo]

Dear Samuel.

Thank you for kind reply.

I solved problem by slightly changing initial wannier projection
Then, I can get exact same band structure between wannier and scf.

Thank you!!

[vinpandey]

Dear Sir, I want to calculate electron self-energy of the PbI2 2D system,
I have been able to generate my wannier band structure which is matching with band structure calculation by PW basis.
but when I am doing this EPW calculation, where I am giving exactly same parameter as in the case.win file, It is not giving me correctly matched band structure.
Actually, I have given nband = 35 in my nscf file, when I do band structure calculation explicitly through W90 (by excluding bands 1-16,35), it runs fine and gives the band strucutre which matches with pw band strucutre exacly. But when I do epw calculation (for all 35 bands), it does n't give right band structure. when I use a flag "exclude_bands" in epw (with help of wdata) it is giving me error , just after it does mmn calculation, and the error is
"= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= PID 26134 RUNNING AT cn32.cms.local
= EXIT CODE: 6
= CLEANING UP REMAINING PROCESSES
= YOU CAN IGNORE THE BELOW CLEANUP MESSAGES"

but this error does not show up when I do wannier calculation explicitly
please guide me.
thanking you for your reply in advance
I am giving my epw input file here
&inputepw
prefix = 'pbi2-5' ,
amass(1) = 207.20000
amass(2) = 126.90447
outdir = './'
dvscf_dir = './save'
elph = .true.
kmaps = .false.
epbwrite = .true.
epbread = .false.
epwwrite = .true.
epwread = .false.
lpolar = .true.
wannierize = .true.
system_2d = .true.
nbndsub = 18
nbndskip = 0

efermi_read= .true.
fermi_energy= -1.03385

dis_win_max = 10.0d0
dis_froz_max = 0.0d0
dis_win_min = -5.0d0
iprint = 2



num_iter = 5000

proj(1)= 'Pb:p'
proj(2)= 'I:p'
wdata(1)= 'dis_num_iter = 10000'
wdata(2)= 'conv_window = 1000'
wdata(3)= 'num_print_cycles = 1000'
wdata(4) = 'dis_mix_ratio = 0.5d0'
wdata(5) = 'guiding_centres=T'
wdata(6) = 'exclude_bands = 1-16,35'
wdata(7) = 'bands_plot = .true.'
wdata(8) = 'begin kpoint_path'
wdata(9) = ' K 0.3333333333 0.3333333333 0.00000 G 0.00000 0.00000 0.00000 '
wdata(10) = ' G 0.00000 0.00000 0.00000 M 0.50000 0.00000 0.00000 '
wdata(11) = ' M 0.50000 0.00000 0.00000 K 0.3333333333 0.3333333333 0.00000 '
wdata(12) = 'end kpoint_path'
wdata(13) = 'bands_num_points = 100'

elecselfen = .true.
phonselfen = .false.


a2f = .false.
parallel_k = .true.
parallel_q = .false.
fsthick = 4 ! eV
eptemp = 300 ! K
degaussw = 0.01 ! eV
!prtgkk = .true.
filkf = 'path.dat'
!nkf1 = 10
!nkf2 = 10
!nkf3 = 1
nqf1 = 1
nqf2 = 1
nqf3 = 1
nk1 = 15
nk2 = 15
nk3 = 1
nq1 = 5
nq2 = 5
nq3 = 1
/
5 cartesian
0.000000000000000E+00 0.000000000000000E+00 0.000000000000000E+00
0.000000000000000E+00 0.230940107618367E+00 0.000000000000000E+00
0.000000000000000E+00 0.461880215236735E+00 0.000000000000000E+00
0.200000000000000E+00 0.346410161427551E+00 0.000000000000000E+00
0.200000000000000E+00 0.577350269045918E+00 0.000000000000000E+00
and my case.win file is
num_bands = 18
num_wann = 18
spinors='true'
exclude_bands = 1-16,35
dis_win_max = 10.0d0
dis_froz_max = 0.0d0
dis_win_min = -5.0d0
dis_num_iter = 10000
dis_mix_ratio = 0.5d0
guiding_centres=T
num_iter = 5000
num_print_cycles = 1000
conv_window = 1000

Begin Atoms_Frac
Pb 0.333333333 0.666666667 0.117489654
I 0.000000000 0.000000000 0.001974014
I 0.666666667 0.333333333 0.232984032
End Atoms_Frac

Begin Projections

Pb:p
I:p
End Projections
bands_num_points = 100

begin kpoint_path
K 0.3333333333 0.3333333333 0.00000 G 0.00000 0.00000 0.00000
G 0.00000 0.00000 0.00000 M 0.50000 0.00000 0.00000
M 0.50000 0.00000 0.00000 K 0.3333333333 0.3333333333 0.00000
end kpoint_path
restart = plot
bands_plot = true
!bands_plot_format=xmgr
Begin Unit_Cell_Cart
bohr
8.81679607 0.000000000 0.000000000
-4.40839803500000000 7.63556937850736228 0.000000000
0.000000000 0.000000000 30.23561583056647381
End Unit_Cell_Cart


mp_grid = 7 7 1
begin kpt
....
end kpts

[sponce]

Hello,

The flag "exclude band" cannot be used in wannier90 (not supported by EPW).

You need to use "nbndskip" within EPW.

Best wishes,
Samuel