The ABINIT Users Mailing List ( CLOSED )

[Alexey Kosobutsky <kosobutsky@list.ru>]

Dear Abinit users,

I am interested in the calculation of the correct position of Si conduction band levels.
For lower conduction bands the results of GW calculation (obtained with using of Abinit 5.8.4) seem to be reasonable, but qusiparticle corrections for levels starting from 11-12 are rather strange: many of them are negative with large absolute values.
For example, for k-point (-0.125, -0.250, 0.000) from a 4x4x4 mesh I have obtained the following result:

k =   -0.125  -0.250   0.000
Band     E0 <VxcLDA>   SigX SigC(E0)      Z dSigC/dE  Sig(E)    E-E0       E
   1  -5.418 -10.505 -17.059   6.668   0.597  -0.676 -10.437   0.068  -5.350
   2   2.652 -10.845 -13.432   2.000   0.744  -0.344 -11.281  -0.437   2.215
   3   4.563 -10.815 -12.460   1.081   0.762  -0.312 -11.246  -0.431   4.132
   4   4.641 -10.851 -12.328   0.940   0.763  -0.310 -11.261  -0.410   4.231
   5   7.998  -9.855  -5.904  -3.727   0.775  -0.290  -9.681   0.174   8.172
   6   9.579 -10.423  -5.761  -4.418   0.762  -0.312 -10.238   0.185   9.764
   7  10.380 -10.413  -5.576  -4.581   0.758  -0.320 -10.220   0.194  10.574
   8  10.627  -9.563  -4.653  -4.594   0.767  -0.303  -9.320   0.242  10.869
   9  13.367 -10.468  -4.758  -5.488   0.737  -0.358 -10.304   0.164  13.531
  10  14.235  -9.918  -4.040  -5.494   0.741  -0.350  -9.634   0.284  14.520
  11  15.745  -8.530  -2.655  -5.884   0.722  -0.385  -8.536  -0.007  15.739
  12  17.387  -8.595  -2.469  -6.227   0.695  -0.438  -8.666  -0.070  17.317
  13  19.090  -9.078  -2.448  -6.954   0.654  -0.528  -9.290  -0.212  18.878
  14  20.359  -9.098  -2.371  -7.261   0.623  -0.605  -9.431  -0.332  20.027
  15  23.609  -9.946  -2.666  -9.303   0.432  -1.316 -10.820  -0.874  22.735
  16  23.992  -9.955  -2.619  -9.848   0.464  -1.157 -11.120  -1.165  22.827
  17  25.101  -8.495  -1.678  -9.661   0.720  -0.389 -10.543  -2.048  23.053
  18  27.486 -10.036  -2.321  -8.687   2.669   0.625 -12.629  -2.592  24.894
  19  27.716 -10.426  -2.445  -9.965   8.090   0.876 -26.477 -16.051  11.665
  20  29.017 -10.024  -2.232  -8.679   1.023   0.023 -10.932  -0.908  28.109
  21  32.152 -10.437  -2.111  -8.320   1.730   0.422 -10.426   0.011  32.163
  22  33.120 -10.525  -2.214  -8.377   0.810  -0.234 -10.578  -0.053  33.067
  23  34.348 -10.485  -2.040  -7.689   1.144   0.126  -9.619   0.866  35.214
  24  35.686  -9.563  -1.642  -7.270   1.029   0.029  -8.892   0.671  36.357
  25  37.497  -9.407  -1.464  -6.995   1.005   0.005  -8.454   0.953  38.450
  26  38.027 -10.288  -1.976  -6.171   1.425   0.298  -7.238   3.050  41.077
  27  38.070  -9.253  -1.467  -6.888   1.528   0.345  -7.881   1.372  39.442
  28  39.153  -9.729  -1.659  -5.469   0.687  -0.455  -7.941   1.788  40.941
  29  39.671  -9.579  -1.471  -6.684   0.580  -0.725  -8.753   0.826  40.497
  30  40.535  -9.818  -1.569  -6.329   0.846  -0.182  -8.194   1.624  42.159

As one can see, the QP correction E-E0 can be as large as -16 eV (for 19th band) that looks quite unphysically. The content of input file used for calculation is shown below, it is modified tgw1_1.in file with sufficiently high (as I think) values of convergence controlling parameters. The corresponding log file is attached, the majority of warnings notify about "Values of Re Sig_c are not linear". The question is: how to calculate high levels of conduction band with sufficient accuracy?

Best wishes,
Alexey


Input file:
# Crystalline silicon
# Calculation of the GW corrections
# Dataset 1: ground state calculation and of the kss file for 10 k-points in IBZ
# Dataset 2: calculation of the screening (epsilon^-1 matrix for W)
# Dataset 3: calculation of the Self-Energy matrix elements (GW corrections)

ndtset      3

# Definition of parameters for the calculation of the KSS file
nbandkss1   -1         # Number of bands in KSS file (-1 means the maximum possible)
nband1       9         # Number of (occ and empty) bands to be computed
#istwfk1     10*1

# Calculation of the screening (epsilon^-1 matrix)
optdriver2  3        # Screening calculation
gwpara2     2
getkss2     -1       # Obtain KSS file from previous dataset
nband2      54       # Bands to be used in the screening calculation
ecutwfn2    4.0      # Cut-off energy of the planewave set to represent the wavefunctions
ecuteps2    8.0      # Cut-off energy of the planewave set to represent the dielectric matrix
ppmfrq2    16.7 eV  # Imaginary frequency where to calculate the screening

# Calculation of the Self-Energy matrix elements (GW corrections)
optdriver3  4        # Self-Energy calculation
gwpara3     1
getkss3     -2       # Obtain KSS file from dataset 1
getscr3     -1       # Obtain SCR file from previous dataset
nband3       150       # Bands to be used in the Self-Energy calculation
ecutwfn3     8.0     # Planewaves to be used to represent the wavefunctions
ecutsigx3    8.0     # Dimension of the G sum in Sigma_x
                    # (the dimension in Sigma_c is controlled by npweps)
nkptgw3      2               # number of k-point where to calculate the GW correction
kptgw3                       # k-points
  -0.125  -0.250   0.000
  -0.125   0.250   0.000
bdgw3    1 30  1 30           # calculate GW corrections for 2 k-points for bands from 1 to 30

# Data common to the three different datasets

# Definition of the unit cell: fcc
acell  3*10.217        # This is equivalent to   10.217 10.217 10.217
rprim  0.0  0.5  0.5   # FCC primitive vectors (to be scaled by acell)
       0.5  0.0  0.5
       0.5  0.5  0.0

# Definition of the atom types
ntypat  1         # There is only one type of atom
znucl 14          # The keyword "znucl" refers to the atomic number of the
                 # possible type(s) of atom. The pseudopotential(s)
                 # mentioned in the "files" file must correspond
                 # to the type(s) of atom. Here, the only type is Silicon.
                       # Definition of the atoms
natom 2           # There are two atoms
typat  1 1        # They both are of type 1, that is, Silicon.
xred              # Reduced coordinate of atoms
     0.0  0.0  0.0
     0.25 0.25 0.25


# Definition of the k-point grid
kptopt  1            # Option for the automatic generation of k points,
nkpt    10
ngkpt   4 4 4
nshiftk 4
shiftk  0.5 0.5 0.5  # These shifts will be the same for all grids
       0.5 0.0 0.0
       0.0 0.5 0.0
       0.0 0.0 0.5
istwfk  10*1         # This is mandatory in all the GW steps.

# Use only symmorphic operations
symmorphi 0

# Definition of the planewave basis set (at convergence 16 Rydberg 8 Hartree)
ecut 8.0          # Maximal kinetic energy cut-off, in Hartree

# Definition of the SCF procedure
nstep   10        # Maximal number of SCF cycles
toldfe  1.0d-6    # Will stop when this tolerance is achieved on total energy
diemac  12.0      # Although this is not mandatory, it is worth to
                 # precondition the SCF cycle. The model dielectric
                 # function used as the standard preconditioner
                 # is described in the "dielng" input variable section.
                 # Here, we follow the prescription for bulk silicon.


# This line added when defaults were changed (v5.3) to keep the previous, old behaviour
 iscf 5

Attachment: log.zip
Description: Binary data