[Pw_forum] pp.x ILDOS plots with ultrasoft pseudopotentials

[Henning Glawe]

On Thu, Feb 10, 2011 at 07:31:01AM +0100, Paolo Giannozzi wrote:
> On Feb 9, 2011, at 8:23 , Henning Glawe wrote:
> > Is this a known problem?
> 
> no, it isn't. Please provide a (small) test showing this behavior

You'll find one attached to this mail (both pw.x scf input, pp.x ildos and
pp.x density input).
I did some further investigations: for a set of espresso versions, I tried
different compiler versions in order to exclude a possible mis-compilation
problem (gfortran-debian-4.4.5-8, ifort-11.1, ifort-12.0.2, all on amd64
architecture systems).

Details of the problem: 
- use USPP
- calculate ildos for emin=-\infty, emax=E_{Fermi} and compare to \rho
Observations:
- in faulty versions, the range of values is [-41,+240], while in working
  versions, it is [0,0.76]

Affected quantum espresso versions: >=4.2

Espresso versions up to 4.1.3 are OK, the calculated ildos \approx \rho if
integration range is -\infty to E_{Fermi} (the smearing contibutes only in
the scf density, but not in the ildos, which causes small differences).
 I could observe these proboblem only when using ultrasoft pseudopotentials,
results with norm-conserving pseudos seem to be ok.

-- 
Mit freundlichen Grüßen
Henning Glawe

Max-Planck-Institut für Mikrostrukturphysik
Weinberg 2, 06120 Halle (Saale), Germany http://www.mpi-halle.de/~theory
Phone: +49-345-5582-613 Fax: +49-345-5511223  Email: glawe at mpi-halle.de
-------------- next part --------------
&control
   calculation   = 'scf'
   restart_mode  = 'from_scratch'
   prefix        = 'CaC6'
   pseudo_dir    = '../pseudo'
   outdir        = 'scratch/'
   verbosity     = 'high'
!   tstress       = .true.
!   tprnfor       = .true.
!   etot_conv_thr = 1.0d-5
!   forc_conv_thr = 1.0d-4
   wf_collect    = .true.
/
&system
   ibrav         = 5
   celldm(1)     = 9.6963160
   celldm(4)     = 0.64213864405089686974
   nat           = 7
   ntyp          = 2
   ecutwfc       = 30
   ecutrho       = 300
   occupations   = 'smearing'
   smearing      = 'methfessel-paxton'
   degauss=0.06
/
&electrons
   diagonalization  = 'david'
   mixing_beta      = 0.2
   conv_thr =  1.0d-10
   mixing_beta = 0.3
/
ATOMIC_SPECIES
   Ca 40.078  Ca.pw91-nsp-van.UPF
   C  12.011  C.pw91-van_ak.UPF 
ATOMIC_POSITIONS crystal
   Ca       0.000000000   0.000000000   0.000000000
   C        0.166666667   0.833333333   0.500000000
   C        0.500000000   0.166666667   0.833333333 
   C        0.833333333   0.500000000   0.166666667
   C        0.833333333   0.166666667   0.500000000
   C        0.500000000   0.833333333   0.166666667
   C        0.166666667   0.500000000   0.833333333
K_POINTS AUTOMATIC 
   8  8  8 0 0 0 
-------------- next part --------------
&inputPP
   outdir   = 'scratch/',
   prefix   = 'CaC6'
   plot_num = 10
   filplot  = 'CaC6.all.ildos'
   emin     = -50
   emax     = 9.8040
!   emin     = 9.7540
!   emax     = 9.8540
/
 
&plot
   iflag=3,
   output_format=6
   fileout='CaC6.all.ildos.cube'
/ 
-------------- next part --------------
&inputPP
   outdir   = 'scratch/',
   prefix   = 'CaC6'
   plot_num = 0
   filplot  = 'CaC6.density'
   emin     = -50
   emax     = 9.6261
/
 
&plot
   iflag=3,
   output_format=6
   fileout='CaC6.density.cube'
/