[Pw_forum] How to set nbnd value for berry phase calculations?

Gabriele Sclauzero sclauzer at sissa.it
Thu Jul 1 12:44:41 CEST 2010


Il giorno 30/giu/2010, alle ore 09.43, Lim Chiang Huay, Freda ha scritto:

> Hi all,
>  
> I’m doing Berry Phase calculations for the first time for BiFeO3. I met with some problems in setting nbnd values for the scf and nscf calculations. And I can’t find an answer in previous forum post.
>  
> I used the pseudopotentials available in the website http://www.pwscf.org/pseudo.php
> They are: 
> Bi.pbe-d-mt.UPF
> Fe.pbe-nd-rrkjus.UPF
> O.pbe-mt.UPF
> and from the pseudopotentials, there should be 41 electrons in a cell. From the manual nbnd = nelect/2.

In your cell you have 2 formula units, hence 82 electrons. Let's suppose that it is an insulator.

If you do spin-unpolarized calculations (nspin=1) you need at least 41 bands, hence nbnd in case (1) is too little.

>  
> I’ve tried different combinations of nbnd values for scf and nscf jobs and non was able to give me the polarization value.
> The following are the some selected cases and their error msg:
> Case(1)
> *.scf, nbnd=21; *.nscf, nbnd=21
> Both jobs crash
> Error for scf: too few bands


> Error for nscf: starting and expected charges differ

You cannot perform the nscf calculation if the scf one has not run successfully.

>  
> Case(2)
> *.scf, nbnd=43; *.nscf, nbnd=42
> Scf converged. Nscf crashed.
> Error for nscf: Polarization only for insulators and no empty bands

The code prevents one from doing Berry's phase calculations for ferromagnetic insulators (though I don't know if these systems actually exist...). There is a check that stops the calculation if nbnd /= nelec/2, which is always the case for these systems (see below). I don't know if there is something from the theoretical point of view that prevents one from doing such calculation. 

In any case, the number of bands in the nscf run is too small to include all occupied spin-up states.

>  
> Case(3)
> *.scf, nbnd=42; *.nscf, nbnd=42
> Scf converged. Nscf crashed.
> Error for scf: too few spin up bands

Since you calculation is spin-polarized there will be an unbalance between the number of occupied spin-up and spin-down bands, hence you need the code to compute more than 42 bands. This stems from the fact that in the pwscf code the number of total spin-up bands (i.e. occuppied+empty=nbnd) is equal to the total number of spin-down bands.

> Error for nscf: starting and expected charges differ

As above.


HTH

GS

>  
> My input files are as follows:
> SCF JOB:
> &control
>     calculation  = 'scf',
>     restart_mode = 'from_scratch'
>     pseudo_dir   = '/Home6/cc/ihpclimc/BP_CAL/PPOT'
>     outdir       = '/Scratch/ihpclimc/BERRY_PHASE_CALCULATIONS/Try_BP/OUTDIR',
>  /
>  &system
>     ibrav=5
>     celldm(1)=10.6523,
>     celldm(4)=0.50988,
>     nat=10
>     ntyp=3
>     nbnd=42
>     ecutwfc=50.0
>     occupations = 'fixed'
>     degauss=0.00
>     nspin=2
>     multiplicity=5
>     lda_plus_u=.true.
>     Hubbard_U(2)=6
>  /
>  &electrons
>     conv_thr = 1e-12,
>     mixing_beta=0.3,
>     diagonalization='cg',
> /
> ATOMIC_SPECIES
>   Bi   209.0      Bi.pbe-d-mt.UPF
>   Fe    55.85     Fe.pbe-nd-rrkjus.UPF
>   O     16.00     O.pbe-mt.UPF
>  
> ATOMIC_POSITIONS crystal
>   Bi    0.000    0.000    0.000         0       0       0
>   Bi    0.500    0.500    0.500         0       0       0
>   Fe    0.219    0.219    0.219
>   Fe    0.719    0.719    0.719
>   O     0.445    0.032    0.897
>   O     0.897    0.445    0.032
>   O     0.032    0.897    0.445
>   O     1.397    0.532    0.945
>   O     0.532    0.945    1.397
>   O     0.945    1.397    0.532
>  
> K_POINTS {automatic}
>   1 1 1 1 1 1
>  
> NSCF JOB:
> &control
>     calculation  = 'nscf',
>     pseudo_dir   = '/Home6/cc/ihpclimc/BP_CAL/PPOT'
>     outdir       = '/Scratch/ihpclimc/BERRY_PHASE_CALCULATIONS/Try_BP/OUTDIR',
>     lberry       = .true.
>     gdir         = 3
>     nppstr       = 7
>  /
>  &system
>     ibrav=5
>     celldm(1)=10.6523,
>     celldm(4)=0.50988,
>     nat=10
>     ntyp=3
>     nbnd=42
>     ecutwfc=50.0
>     occupations = 'fixed'
>     degauss=0.00
>     nspin=2
>     multiplicity=5
>     lda_plus_u=.true.
>     Hubbard_U(2)=6
>  /
>  &electrons
>     conv_thr = 1e-12,
>     mixing_beta=0.3,
>     diagonalization='cg',
> /
> ATOMIC_SPECIES
>   Bi   209.0      Bi.pbe-d-mt.UPF
>   Fe    55.85     Fe.pbe-nd-rrkjus.UPF
>   O     16.00     O.pbe-mt.UPF
>  
> ATOMIC_POSITIONS crystal
>   Bi    0.000    0.000    0.000         0       0       0
>   Bi    0.500    0.500    0.500         0       0       0
>   Fe    0.219    0.219    0.219
>   Fe    0.719    0.719    0.719
>   O     0.445    0.032    0.897
>   O     0.897    0.445    0.032
>   O     0.032    0.897    0.445
>   O     1.397    0.532    0.945
>   O     0.532    0.945    1.397
>   O     0.945    1.397    0.532
>  
> K_POINTS {automatic}
>   4 4 4 1 1 1
>  
>  
> Can anyone help?
>  
> Thanks,
> Rgds
>  
> Freda Lim
> Institute of High Performance Computing
> 1 Fusionopolis Way, #16-16 Connexis
> Singapore 138632
> Email: limch at ihpc.a-star.edu.sg
>  
>  
> 
> This email is confidential and may be privileged. If you are not the intended recipient, please delete it and notify us immediately. Please do not copy or use it for any purpose, or disclose its contents to any other person. Thank you.
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum


§ Gabriele Sclauzero, EPFL SB ITP CSEA
   PH H2 462, Station 3, CH-1015 Lausanne

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.democritos.it/pipermail/pw_forum/attachments/20100701/90c3411c/attachment-0001.htm 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1753 bytes
Desc: not available
Url : http://www.democritos.it/pipermail/pw_forum/attachments/20100701/90c3411c/attachment-0001.bin 


More information about the Pw_forum mailing list