The ABINIT Users Mailing List ( CLOSED )

[Jorge Iniguez <jiniguez@nist.gov>]

Thank you all for your responses! OK, so the problem is very probably
related to the XC-FFT loss of translational symmetry, which may be worse
for alkali atoms, and things would improve if we could use intxc=1 in the
response calculations (and things *will* improve if I compute the gamma
phonons via ground state calculations with intxc=1; I will check).  I
guess I could have imagined... I was misled by the fact that the
insulating case worked so well, but Matthieu's proposed explanation for
this is probably right. (By the way, Matthieu, I am using TM pseudos from
the Khein-Allan distribution, so the calculations should be converged in
the cutoff.) OK then, I will keep playing with these things and let you
know if I find something noteworthy.

Cheers,
Jorge