The ABINIT Users Mailing List ( CLOSED )

[Andrey TITOV <and80t@yahoo.com>]

Dear colleagues,
Thank you for your previous answers. I try to calculate electronic states of a
Si10H16 charged cluster in a big box using GW approximation (charge=-1.0,
version abinit-5.4.4). I obtain these GW energies:
           1
  0.000000  0.000000  0.000000
  19
    22  -6.2353  -0.6129   0.0028
    23  -5.7809  -0.4317   0.0068
    24  -5.7771  -0.4279   0.0068
    25  -5.7707  -0.4215   0.0068
    26  -5.7573  -0.5252   0.0018
    27  -5.7583  -0.5262   0.0018
    28  -5.7588  -0.5267   0.0018
    29   0.1360   0.8574   0.0000
    30   1.9205   2.2868   0.0000
    31   1.9190   2.2853   0.0000
    32   1.9165   2.2828   0.0000
    33   1.6850   1.9313   0.0000
    34   1.6873   1.9337   0.0000
    35   1.2768   1.1995  -0.0010
    36   2.2875   2.1017  -0.0015
    37   2.2874   2.1017  -0.0015
    38   2.2874   2.1016  -0.0015
    39   2.5163   2.0420  -0.0007
    40   1.8490   1.2706  -0.0020

The Fermi level is in the band 29. We see that the order of bands above the
Fermi level is not monotonic: some bands (33, 34, 35) are placed in energy
below previous bands (30, 31, 32). However previous LDA calculation gives a
monotonic positions of bands in energy. I do not understand that: is it a
physical effect? The same result is obtained if I use plasmon pole model or
contour integration for screening calculation.

This effect is also observed in the neutral cluster, but at higher energy 
(band
42):
           1
  0.000000  0.000000  0.000000
  27
    24  -6.7802  -0.8995   0.0000
    25  -6.7751  -0.8943   0.0000
    26  -6.7580  -0.9965   0.0000
    27  -6.7585  -0.9971   0.0000
    28  -6.7583  -0.9969   0.0000
    29   1.3567   2.4416   0.0000
    30   1.7790   2.5854   0.0000
    31   1.7777   2.5841   0.0000
    32   1.7763   2.5827   0.0000
    33   1.8384   2.4054   0.0000
    34   1.8391   2.4062   0.0000
    35   2.3342   2.4982  -0.0014
    36   2.3345   2.4986  -0.0013
    37   2.3331   2.4972   0.0000
    38   2.4201   2.2714   0.0000
    39   2.9348   2.7214   0.0001
    40   2.9367   2.7233  -0.0009
    41   2.9489   2.7355  -0.0009
    42   2.0714   1.6368  -0.0008
    43   3.2147   2.7079  -0.0005
    44   3.2152   2.7084   0.0000
    45   3.2117   2.7049  -0.0001
    46   2.2551   1.2258  -0.0005
    47   2.2520   1.2228  -0.0002
    48   2.2506   1.2213   0.0000
    49   2.4290   1.2941   0.0001
    50   3.4170   2.1483   0.0015

I send you my input file for the charged cluster. May be someone observed this
effect in other materials? Is there an explanation of this result?
Thank you,
Andrey Titov.

Institut matériaux microélectronique et nanosciences de Provence 




# Si10h16 cluster

ndtset   4

kptopt   1            
ngkpt    1 1 1        

charge  -1.0

# Dataset1: usual self-consistent ground-state calculation
# Definition of the k-point grid
nkpt1    1
nshiftk1 1
shiftk1  0.0 0.0 0.0  # This grid is the most economical
         0.5 0.0 0.0
         0.0 0.5 0.0
         0.0 0.0 0.5
prtden1  1         # Print out density
prteig1  1
nband1   150
istwfk1  1*1



# Dataset2: calculation of kss file
# Definition of k-points
nkpt2     1           
nshiftk2  1
shiftk2   0.0 0.0 0.0  
          0.0 0.5 0.5
          0.5 0.0 0.5
          0.5 0.5 0.0
istwfk2   1*1                    
iscf2     -2            
getden2   -1            
nband2    150
nbandkss2 150           
kssform2  3

# Dataset3: Calculation of the screening (epsilon^-1 matrix)
optdriver1  3        
getkss1     2        
nband1      150      
ecutwfn1    3.0      
ecuteps1    3.0      
nfreqim1    4
nfreqre1    10
freqremax1  1.0
gwcalctyp1  2

# Dataset4: Calculation of the Self-Energy matrix elements (GW corrections)
optdriver2  4        
getkss2     2        
getscr2    -1        
nband2      150      
ecutwfn2    3.0      
ecutsigx2   3.0      
                     
nkptgw2     1        
kptgw2                     
  0.000    0.000    0.000    
bdgw2       22  40           
gwcalctyp2  2


# Definition of the unit cell
acell  3*25.0            
rprim  1.0  0.0  0.0     
       0.0  1.0  0.0   
       0.0  0.0  1.0

# Definition of the atom types
ntypat  2         
znucl 14 1        
                         

# Definition of the atoms
natom 26           
typat  1 1 1 1 1 1 1 1 1 1              
       2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
xcart               # Reduced coordinate of atoms
-2.56577463  -2.56577463  -2.56577463
 2.56577463  -2.56577463   2.56577463
-2.56577463   2.56577463   2.56577463
 2.56577463   2.56577463  -2.56577463
 0.0          0.0         -5.13155235
 0.0          0.0          5.13155235
 0.0         -5.13155235   0.0
 0.0          5.13155235   0.0
-5.13155235   0.0          0.0
 5.13155235   0.0          0.0
 6.75466170  -1.62310994  -1.62310994
-1.62310994   6.75466170  -1.62310994
-1.62310994  -1.62310994   6.75466170
 1.62310994  -1.62310994  -6.75466170
-6.75466170  -1.62310994   1.62310994
 1.62310994   6.75466170   1.62310994
-1.62310994   1.62310994  -6.75466170
-1.62310994  -6.75466170   1.62310994
 6.75466170   1.62310994   1.62310994
 1.62310994  -6.75466170  -1.62310994
-6.75466170   1.62310994  -1.62310994
 1.62310994   1.62310994   6.75466170
-4.18888535  -4.18888535  -4.18888535
 4.18888535  -4.18888535   4.18888535
-4.18888535   4.18888535   4.18888535
 4.18888535   4.18888535  -4.18888535


# Definition of the planewave basis set 
ecut 11.0          # Maximal kinetic energy cut-off, in Hartree

# Use only symmorphic operations
symmorphi 0

# Definition of the SCF procedure
nstep   200       # Maximal number of SCF cycles
diemac  3.0       # 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.
tolwfr   1.0d-22

iscf 3