[Pw_forum] GWW-band (P. Umari)

[umari at democritos.it]

Dear Ali,

First of all, note that the GWW is implemented only at the gamma point,
so it requires large supercells, (or better it is appropriate for
non-crystalline systems).
However the long-range parts of the symmetric dielectric matrix
and of the screened-interaction can be calculated using k-points.
With the code ph.x you are calculating the head and the wings of
the symmetric dielectric matrix.
E_HEAD are the head terms of the symmetric dielectric matrix
calculated on the imaginary frequency axis (e.g. "E_HEAD : 58"
is relative to the 58th frequency step). These terms together with
the wings are calculated using linear response.
Then, also the long range part  (element G=0,G=0)
of the screened interaction can also be calculated using a k-points grid.
The calculated elements are <\Psi_i|G(i \tau)v|\Psi_i>,
where \Psi_i are the KS states G(i\tau) is the KS Green function
in imaginary time and v is the bare Coulomb interaction.
The term "X energy 1" is the exchange energy relative to the 1st KS state.

> I am trying to use implemented approach GWW in espresso to obtain
> quasiparticle state for bulk TiO2 ,but I didn,t succeed to get true band
> gap by GW density of state.
> I used 10 10 12 kmesh for k sampling but I don't know which file and which
> rows contains dielectric constant. when I grep E_HEAD ,the last lines is
> E_HEAD : 58 (-0.288322417972075195E-19,-0.347622164532722579E-06)
> E_HEAD : 59 (0.136048521030191330E-19,-0.268035068560476896E-06)
..... .....
> what is these two columns? where do I look for dielectric constant?
> also when I grep the energy in exchange file I see
> X energy 1 -3.13056306094735648
> X energy 2 -3.13125888267243768

Best regards,

P. Umari (Democritos)

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.