Drude frequency in LRC

[Stephan]

Hi Daniele,
Have you already found out what went wrong in the LRC-calculations with the Drude contributuion?

Best regards
Stephan

[Daniele Varsano]

Dear Stephan,
Unfortunately I did not have found the time to check it. In the meanwhile if you have time, can you do a calculation
including local-filed effects (hartree) with NGblk > 1, avoiding to use the LRC functional, including the Drude term and
post here the results?

Best,
Daniele

[Daniele Varsano]

Dear Stephan,
we have been checking the Drude contribution which seems ok for RPA calculation, including local field effects.
The problem arise indeed when using the LRC functional, which seems to kill the zero frequency divergence.

Daniele

[Davide Sangalli]

Dear Stephan,
few more comments.

As you may know the macroscopic part of the classical induced field is responsible for the difference between the epsilon, which in a metal has a divergence at w=0 and the eels, which has a pole at w=w_p. The alpha/q^2 kernel is exactly equal to such term if alpha=4\pi. Thus switching on alpha the kernel moves the divergence away from w=0 to w=w_p as one may expect.

Best,
Davide