Epsilon_1

[Yunfeng]

Dear All,

How does Yambo caculate the real part of the complex dielectric function: epsilon_1? I mean what's the equation, in (yambo -b -o b -y h) and (yambo -o c).

Thank you very much
Sincerely, Yunfeng

[andrea marini]

How does Yambo caculate the real part of the complex dielectric function: epsilon_1? I mean what's the equation, in (yambo -b -o b -y h) and (yambo -o c).

Dear Yunfeng,

Yambo always calculates the full, complex, response function. To understand how it does it look at the these pages of the dicumentation: LR, Response, BSE solvers.

If something will remain still unclear let us know and we will help you.

[Yunfeng]

Dear Andrea.

Thank you very much! I was actually just confused with one of previous post by Myrta elsewhere. I got a feeling that the epsilon_1 was calculated by Kramers-Kronig relation. see below:

The warning regards the eel, so you should worry only if you want to calculate it . Within the Tamm Dancoff approximation for the BSE, this spectra is obtained from the eps one, and a Kramers-Kronig has to be performed. In order for this transform to be accurate you should choose the energy range in such a way that the Im(eps) is (approx.) zero at the beginning and at the end of the range.

My understanding is that EEL relies only on epsilon_1 and epsilon_2 by
EELS= epsilon_2/(epsilon_1^2+epsilon_2^2).

If both are calculated in direct way, we need not use the Kramers-Kronig relation, though it is also good.

Thanks
Sincerely, Yunfeng

[andrea marini]

Thank you very much! I was actually just confused with one of previous post by Myrta elsewhere. I got a feeling that the epsilon_1 was calculated by Kramers-Kronig relation. see below:

The warning regards the eel, so you should worry only if you want to calculate it . Within the Tamm Dancoff approximation for the BSE, this spectra is obtained from the eps one, and a Kramers-Kronig has to be performed. In order for this transform to be accurate you should choose the energy range in such a way that the Im(eps) is (approx.) zero at the beginning and at the end of the range.

My understanding is that EEL relies only on epsilon_1 and epsilon_2 by
EELS= epsilon_2/(epsilon_1^2+epsilon_2^2).

If both are calculated in direct way, we need not use the Kramers-Kronig relation, though it is also good.

Dear Yungfeng,

Myrta is right in what concerns the EELS. In the case of the Bethe-Salpeter the use of the Tamm-Dancoff approximation imposes to use a Kramers-Kronig integration to build the negative frequency part of the response function to buil the EELS. In the reciprocal space solvers (-o c), instead, there is no need of using a KK transformation. So, I was right in the sense the Yambo always calculate the full complex response function. However in the case of the BSE you get only the resonant part of epsilon_1, that is not enough to build the EELS.