The ABINIT Users Mailing List ( CLOSED )

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

Dear Javier,

It might be that you do not include enough bands (input variable  
nband). This number is chosen automatically
by ABINIT, but might not be enough for your calculation (phonons in a  
metal).  Indeed, it has been observed that
phonon frequencies are sensitive to the explicit treatment of bands  
that are higher than the Fermi energy,
and only very weakly populated. Such bands usually do not play a  
major role in the total energy,
but this might not be the case for the phonons.

Xavier


On 08 Jun 2006, at 19:40, javier.torres@unito.it wrote:

> Dear Abinit Users
>
> I performed a response function calculation in order to get the  
> phonons at gamma point of AgB2... For this I used the optimized  
> gometry of my system, performed with ionmov 2 and optcell 2 as you  
> can see in the attached files, the obtained lattice parameters  
> correspond to a stress tensor with components smaller than 10^-8,  
> the atomic coordinates were not optimized because the atoms are  
> special positions...
>
> The result of the phonon calculation with this geometry is a set of  
> freqs with very negative values of about -46 cm^-1 ... How can I  
> get the adequate geometry of my system to get the true frequencies???
>
> Kind regards
>
> F. T.
>
> OPT...
> # Hexagonal Primitive (P6/mmm) AgB2 GS SHIFTK  0 0 0
> # Geometric block
> spgroup  191
> acell   5.4424112657  5.4424112657  6.6707332527  #Experimental  
> Values Lal
> angdeg  90 90 120
> ntypat  2
> znucl   47 5
> natom   3
> typat   1 2 2
> xred
>         0.0 0.0 0.0
>         1/3 2/3 1/2
>         2/3 1/3 1/2
> #
> # OPTCELL
> #
> ionmov   2
> optcell  2    #try opcell 2 too!!!
> tolmxf   1.0d-5
> ntime    50
> dilatmx  1.2
> strfact  10
> #
> # Methods and Parameters
> #
> intxc    0         #better for fft changed  disable for RF calc
> occopt   7         #added for metalic occupation Gaussian smearing
> tsmear   0.01      #convergence
> brav     4         #added for hexagonal lattice
> ixc      1
> ecut     30.0
> ecutsm   0.5
> ngkpt    10 10 10       #16x16x16 exceeds the limit of memory  
> allocation with kopt=2
> frzfermi 1
> nshiftk  1
> shiftk   0.0 0.0 0.0
> #
> # SCF defined in each dataset
> #
> nstep     1000
> #
> # DS_1 gs and density calculation
> #
> iscf     3
> kptopt   1
> kptgen   1
> toldfe   1.0d-12   #important for freqcalc
> #irdwfk  1
>
>
>
> Phon...
>
> # Hexagonal Primitive (P6/mmm) AgB2 GS SHIFTK  0 0 0
> ndtset 2
> # Geometric block
> spgroup  191
> acell   5.6504357844  5.6504357843  7.5478571140
> angdeg  90 90 120
> ntypat  2
> znucl   47 5
> natom   3
> typat   1 2 2
> xred
>         0.0 0.0 0.0
>         1/3 2/3 1/2
>         2/3 1/3 1/2
> #
> # Methods and Parameters
> #
> intxc    0         #better for fft changed  disable for RF calc
> occopt   7         #added for metalic occupation Gaussian smearing
> tsmear   0.01      #convergence
> brav     4         #added for hexagonal lattice
> ixc      1
> ecut     43.0
> ecutsm   0.5
> ngkpt    10 10 10       #16x16x16 exceeds the limit of memory  
> allocation with kopt=2
> #frzfermi 1
> nshiftk  1
> shiftk   0.0 0.0 0.0
> #
> # SCF defined in each dataset
> #
> nstep     1000
> #
> # DS_1 gs and density calculation
> #
> iscf1     3
> kptopt1   1
> kptgen1   1
> tolwfr1   1.0d-20   #important for freqcalc
> #irdwfk1   1
> #
> # DS_2 whole phonon
> #
>   rfphon2   1            # Will consider phonon-type perturbation
>  rfatpol2   1 3          # Only the first atom is displaced
>    rfdir2   1 1 1        # Along the first reduced coordinate axis
>     nqpt2   1            # One wavevector is to be considered
>      qpt2   0 0 0        # This wavevector is q=0 (Gamma)
>   kptopt2   2            # Automatic generation of k points, taking
>                         # into account the time-reversal symmetry only
>   tolvrs2   1.0d-10       # SCF stopping criterion
>     iscf2   3            # Self-consistent calculation, using  
> algorithm 3
>   getwfk2   -1            # Read the ground-state wavefunctions