[Wannier] Energy mismatch among symmetry equivalent states in silicon

[Vahid Askarpour]

Dear Arash,

The discrepancy does not exist if I wannierse the four valence bands. I used the setting in the example03 in the wannier90-2.0.1 which gets installed with QE package.

It uses the following .win file:

*********************************************
num_bands         =   12       

! For valence+conduction
num_wann          =   8     

! For valence only
! num_wann   =  4      

dis_win_max       = 17.0d0     ! remove for valence only calculation
dis_froz_max      =  6.4d0
dis_num_iter      =  120
dis_mix_ratio     = 1.d0

num_iter          = 50
num_print_cycles  = 10

Begin Atoms_Frac
Si  -0.25   0.75  -0.25
Si   0.00   0.00   0.00 
End Atoms_Frac

Begin Projections     
! For valence+conduction
Si : sp3 
! For valence only
! f=-0.125,-0.125, 0.375:s
! f= 0.375,-0.125,-0.125:s
! f=-0.125, 0.375,-0.125:s
! f=-0.125,-0.125,-0.125:s
End Projections

begin kpoint_path
L 0.50000  0.50000 0.5000 G 0.00000  0.00000 0.0000
G 0.00000  0.00000 0.0000 X 0.50000  0.00000 0.5000
X 0.50000 -0.50000 0.0000 K 0.37500 -0.37500 0.0000 
K 0.37500 -0.37500 0.0000 G 0.00000  0.00000 0.0000
end kpoint_path

Begin Unit_Cell_Cart
-2.6988 0.0000 2.6988
0.0000 2.6988 2.6988
-2.6988 2.6988 0.0000
End Unit_Cell_Cart
*****************************************

The final spreads for only the valence bands are:

Final State
 WF centre and spread    1  ( -0.129387,  0.129387, -0.129387 )     3.25457766
 WF centre and spread    2  ( -0.129387, -0.129387,  0.129387 )     3.25457766
 WF centre and spread    3  (  0.129387,  0.129387,  0.129387 )     3.25457766
 WF centre and spread    4  (  0.129387, -0.129387, -0.129387 )     3.25457766
 Sum of centres and spreads ( -0.000000,  0.000000, -0.000000 )    13.01831064

When I wannierise 4 valence+4 conduction, I get:

Final State
 WF centre and spread    1  (  0.447952,  0.447952,  0.447952 )     2.58541760
 WF centre and spread    2  (  0.447952, -0.447952, -0.447952 )     2.58541760
 WF centre and spread    3  ( -0.447952,  0.447952, -0.447952 )     2.58541760
 WF centre and spread    4  ( -0.447952, -0.447952,  0.447952 )     2.58541760
 WF centre and spread    5  ( -1.125561,  1.564973,  1.564973 )     2.96886521
 WF centre and spread    6  ( -1.125561,  1.125561,  1.125561 )     2.96886521
 WF centre and spread    7  ( -1.564973,  1.564973,  1.125561 )     2.96886521
 WF centre and spread    8  ( -1.564973,  1.125561,  1.564973 )     2.96886521
 Sum of centres and spreads ( -5.381069,  5.381069,  5.381069 )    22.21713123

So with 8 bands, I get smaller spreads but the energy mismatch appears. It seems that I get better WFs with 8 bands than 4 bands. Or perhaps I am 
missing something here. 

Thank you,
Vahid


> On Sep 1, 2016, at 10:09 AM, Mostofi, Arash <a.mostofi at imperial.ac.uk> wrote:
> 
> Dear Vahid
> Does the same discrepancy arise if you wannierise just the four valence bands, or does it only occur when you disentangle 8 WFs? If the latter, you may need to adjust you disentanglement settings to get a better set of WFs.
> Best wishes
> Arash 
> --
> Departments of Physics & Materials
> Imperial College London
> 
> Sent from my iPhone
> 
>> On 31 Aug 2016, at 23:05, Vahid Askarpour <vh261281 at dal.ca> wrote:
>> 
>> Dear Wannier Users,
>> 
>> I have calculated the interpolated band structure of silicon (8 wannier functions) using a k grid of 8x8x8. I considered two segments:
>> 
>> 1. From 0.05 0.4 -0.4 to 0.05 0.4 -0.5
>> 2. From 0.85 0.4 0.3 to 0.85 0.4 0.4
>> 
>> My purpose in choosing these two segments is to compare the electronic energies at two symmetry equivalent points: (0.05,0.4,-0.45) and (0.85,0.4,0.35).
>> 
>> Here are the eight band energies (in eV) for the two points:
>> 
>> -2.2651301     -2.262256
>> -0.7977982    -0.7918293
>> 2.5481431      2.5474092
>> 3.3210793      3.4353798
>> 7.3564264      7.340448
>> 8.7099476      8.6832805
>> 14.055428      14.064489
>> 14.828398      14.849574
>> 
>> In case of diamond, these energies are identical. But for silicon, for example, band 4 is different by 100meV! Has anyone encountered this energy mismatch
>> and is there a way to deal with it?
>> 
>> Thank you,
>> 
>> Vahid
>> 
>> 
>> Vahid Askarpour
>> Department of Physics and Atmospheric Science
>> Dalhousie University,
>> Halifax, NS, Canada
>> 
>> _______________________________________________
>> Wannier mailing list
>> Wannier at quantum-espresso.org
>> http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier
> _______________________________________________
> Wannier mailing list
> Wannier at quantum-espresso.org
> http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier