[Pw_forum] Calculation of specific irreps only

[Turan Birol]

   Dear members of the pwscf community,

   I am calculating the phonon frequencies of Sr4 Ti3 O10 at gamma
point. What I am interested in particular is a soft mode, which,
unfortunately, is the hardest one to converge. So, I was wondering if
it is possible to get the frequencies of only a group of phonons,
without doing the whole calculation.

   The topic have been apparently briefly discussed in the forum in
July 2006, but I believe the suggestion given there (use of modenum)
is not suitable for my purpose. Instead I specified the relevant
representations in the following way:

&inputph
 tr2_ph=1.0d-16,
 prefix='Sr2TiO4',
 lnscf=.true.,
 nrapp=2,
/
0.0 0.0 0.0
5 6

in the input of ph.x. However, the code stops, after doing the
calculation and achieving convergence for only the relevant modes as
wished, saying:

     Stopping because representation    1 is not done

   and timing analysis follows. But the frequencies of these modes are
not given. As I understand it, using the nrapp keyword as I have done
it should be possible to calculate the frequencies of phonons
belonging to specific representations, without having to calculate all
the phonons at the given wave vector. Is this correct? If not, is it
due to the way things are implemented, or is it not possible in DFPT,
in which case I have to resort to frozen phonons? (I do not see a
reason for the latter, as different irreps correspond to different
types of distortions which -in principle- are decoupled; however,
there are people in the list who know more about DFPT than me, for
sure. :-) )

   Thank you very much for your time.

   Regards,
   Turan Birol
   Phd student, Cornell University Department of Physics, USA