Joint density of states

[ramkrishna]

Dear Sir,
In optical spectrum we can directly calculate the real and imaginary part of the dielectric. However, we know the imaginary part is determined by the joint density of states (joint DOS). Is it possible to calculate (or extract from the Im_epsilon) the joint DOS?

Thanks
Ramkrishna

[zh]

By using an approximated relationship between the JDOS  and the imaginary part of the dielectric (\epsilon_2):  $\epsilon \propto  \frac{JDOS} { \omega^2}$,
(see http://etsf.polytechnique.fr/sites/default/files/dft-II-silva.pdf)
you may obtain the JDOS.

[ams_nanolab]

This Eq. for the JDOS seems confusing. The absorption eps_2 contains info about dipole matrix elements (oscillator strength) while the JDOS does not, how can both be proportional?

[zh]

Please read the 22nd page of the Silvana Botti's presentation (http://etsf.polytechnique.fr/sites/default/files/dft-II-silva.pdf). There is an assumption that the optical matrix elements are assumed to constant.