Decompose of dielectric function in metallic system

[andyamygto]

Hi all

I am trying to calculate the q-dependent dielectric function of metallic system using linear response method.
I know that we can use % BndsRnXd to choose the band in the Polarization function.
However, in the metallic system it seems like the code force to sum all the band crossing the Fermi level (namely metallic band)

Is there a way to do decompose the contribution of different metallic band to the dielectric function like the PRB 104, 125101 Fig.7?

Thanks in advance.

[Daniele Varsano]

Dear Chin,
unfortunately, there is not an easy way to make this decomposition from input.
You can do it by selecting your band of interest inside the code. At the moment, I do not have better ideas.

Best,
Daniele

[andyamygto]

Dear Daniele

Thanks for the information.

Actually I have tried to comment the line 59~64 in the src/pol_function/X_pre_setup.F

if (X%ib(1)>minval(Xen%nbf)+1) X%ib(1)=minval(Xen%nbf)+1
if (X%ib(2)<maxval(Xen%nbm) ) X%ib(2)=maxval(Xen%nbm)

if (X%ib(1)<0.or.X%ib(1)>maxval(Xen%nbm).or.X%ib(2)<X%ib(1)) X%ib(1)=1
if (X%ib(2)>Xen%nb.or.X%ib(2)<minval((/Xen%nbf(:n_sp_pol)/))) X%ib(2)=Xen%nb

It seems like the code will crush if I comment these lines the code will crush.
Can you hint that which part in the code is calculation the Screening matrix elements ?

Best,
Andy Hsu