The ABINIT Users Mailing List ( CLOSED )

[Xavier Gonze <gonze@pcpm.ucl.ac.be>]

Dear Lee Geunsik,

Pseudopotential might be the problem, as suggested by Matthieu.
Another possibility is the number of bands. You might
need more bands than in the ground state.

Xavier

Lee Geunsik wrote:
> Dear abinitors,
> 
> I am looking for unstable phonon modes of beta-tin (body-centered 
> tetragonal) Si.
> Beta-tin is the high pressure phase next to diamond.
> The problem is to test phonon freq. convergence with respect to input 
> parameters.
> 
> With the volume corresponding beta-tin phase, electronic calculation shows 
> that it is metalic.
> So I computed phonon freq. for q=(0,0,0) with occopt=3, tsmear=0.004, 
> nband=6.
> Following is relevant input file.
> 
> ndtset  6   udtset 3 2
> 
> ecut1?   16.0
> ngkpt1?   3*16
> 
> ecut2?   16.0
> ngkpt2?   3*18
> 
> ecut3?   18.0
> ngkpt3?   3*16
> 
> #Ground state
> kptopt?1   1
> tolvrs?1   1.0d-18
> prtden?1   1
> iscf?1     5
> 
>  getwfk?2  -1
> rfatpol?2  1 2
>   rfdir?2  1 1 1
>  rfphon?2  1
>    nqpt?2  1
>  kptopt?2  2
>  tolvrs?2  1.0d-8
>    iscf?2  5
>   prtwf?2  0
> 
> qpt?2  0.0 0.0 0.0
> 
> #Common input variables
> occopt 3   tsmear 0.004
> nband 6
> 
> fftalg 412
> acell  3*8.88
> rprim  0.0  0.5  0.5
>        0.5  0.0  0.5
>        0.5  0.5  0.0
> ntypat 1
> znucl 14
> natom 2
> typat 1 1
> xred  0.0  0.0  0.0
>       1/4  1/4  1/4
> ixc 2
> nshiftk 1
> shiftk 0.5 0.5 0.5
> nstep 40
> diemac 12.0
> 
> 
> Following is the results.
> 
> ecut=16 Hart.  ngkpt=3*16
>  Phonon frequencies in cm-1    :
> -  7.573314E+01  7.699138E+01  7.720520E+01  7.450543E+02  7.451827E+02
> -  7.452047E+02
> 
> ecut=16 Hart.  ngkpt=3*18
>  Phonon frequencies in cm-1    :
> -  7.572324E+01  7.699129E+01  7.720675E+01  7.448678E+02  7.452402E+02
> -  7.453041E+02
> 
> ecut=18 Hart.  ngkpt=3*16
>  Phonon frequencies in cm-1    :
> -  7.630890E+01  7.747308E+01  7.767107E+01  7.450455E+02  7.451757E+02
> -  7.451981E+02
> 
> All accoustic freq. are quite larger than zero.
> What is the possible reasons?
> Could you give some information?
> 
> Thanks a lot.
> Geunsik.
> 
> 
>