nlo of metallic systems

[ilukacevic]

Dear colleagues,

I'm trying to calculate the nlo rf of metallic systems. I optimized my structures using occopt 4. But when I start the rf nlo, abinit stops in the beginning saying that the value 4 is not allowed in nlo. Then I tried 3 and 6 and again it happened the same.
Abinit suggests me to use occopt 1. As I understand from the input variables help, occopt 1 gives a semicond. Since I know that the studied systems are metallic, how can I incorporate these with rf nlo calculations?

Thank you all for any help!

Yours,

Igor Lukacevic

[gonze]

Dear Igor,

By nlo, I guess you mean optdriver=5 .
Please note that, at present, optdriver=5 allows to compute
non-resonant Raman scattering - being non-resonant, metals are excluded -,
and non-linear dielectric susceptibility at long-wavelength - being at long wavelength,
metals are excluded - .

For metals, frequency-dependent nlo should possible, in the simple sum-over-state
approximation, using the "optics" utility.
See Tv4#57

Best,
X.

[ilukacevic]

Dear prof. Gonze,

thank You for the instructions. I remembered yesterday afternoon that there is no non-resonant Raman in metals.

Igor Lukacevic