Page **1** of **1**

### Drude term

Posted: **Tue Sep 04, 2018 12:18 pm**

by **neets**

Hello,

I’m confused about how to get the data of DrudeWXd in a specific metallic system. Shoud I calculate the Drude frequency from the Eq(2.17) in the Andrea' thesis.

Best,

neets

### Re: Drude term

Posted: **Thu Sep 06, 2018 4:05 pm**

by **Daniele Varsano**

Dear Neets,

please sign your post with your full name and affiliation, this is a rule of the forum, you can do once for all by filling the signature in your profile.

In order to take into account the Drude tail, Yambo uses the Drude model, so the DrudeWXd frequency enters in the input as an empirical parameter.

Usually, the value of the Drude frequency and width is taken from the literature, in general from experiments or previous calculations.

The expression you found in Andrea's thesis is not implemented in Yambo anymore as very heavy and impractical.

Best,

Daniele

### Re: Drude term

Posted: **Fri Sep 07, 2018 6:23 am**

by **neets**

Dear Daniele,

Thank you for your reply. As you said, the Drude frequency and width is taken from experiments or previous calculations generally. What should we do when we are interested in the Plasmon Resonance of a new metallic system ?

### Re: Drude term

Posted: **Tue Sep 11, 2018 5:39 pm**

by **Daniele Varsano**

Dear Neets,

it depends on what you are interested in,

if you are interested in plasmons I suggest you:

1) to look at the eels spectrum, you can see it at RPA level and if you

are interested in a particular q vector and it converges with respect the k point sampling then intraband transitions

are included and you should not need the Drude term.

2) If you are interested instead to the q=0 response you can see at what frequency the plasmon is located. If it is far from w=0 then the Drude term even if

there should not affect that much the position of the plasmon.

3) If this is not the case, a possibility could be to look at finite q increasing k point sampling for smaller and smaller q point toward zero, fit the Drude parameters and

extrapolate to q=0. Not sure about the precision of this workaround.

Best,

Daniele