Oddness of Imaginary Dielectric Function

Deals with issues related to computation of optical spectra, in RPA (-o c) or by solving the Bethe-Salpeter equation (-o b). Includes local field effects, excitons, etc.

Moderators: Davide Sangalli, Daniele Varsano, andrea.ferretti, andrea marini, Conor Hogan, myrta gruning

Post Reply
xueshanxihua
Posts: 18
Joined: Tue Dec 06, 2016 9:43 am

Oddness of Imaginary Dielectric Function

Post by xueshanxihua » Tue Sep 24, 2019 7:38 am

Dear developers,

As required by its oddness, we know that the imaginary part of dielectric function strictly vanishes at ω=0.

Indeed, that's what we get in an IP spectra calculation of Yambo.

However, I was confused that the usually used expression, as given in Eq.(1) of https://iopscience.iop.org/article/10.1 ... 48X/ab15d0, seems not to guarantee the oddness. I.e., the Eq.(1) can't ensure a strictly-zero-value at ω=0, since a finite broadening of spectrum will lead to an accumulation of intensity at ω=0.

Thus, I'd like to know if Yambo had use any technical approaches or additional constraints to ensure the imaginary dielectric function to be strictly zero at ω=0.



Thank you very much,
Zeyu Jiang
Postdoc
Department of Physics, Applied Physics, and Astronomy
Rensselaer Polytechnic Institute
Troy, New York

User avatar
Daniele Varsano
Posts: 2660
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Oddness of Imaginary Dielectric Function

Post by Daniele Varsano » Tue Sep 24, 2019 10:44 am

Dear Zeyu,
in the equation, you mentioned it is reported the Time-Ordered version of the response function which is not strictly zero for w=0. This should be the default used by Yambo. Anyway, the ordering of the response function can be changed (by using the verbosity -V resp), you will have the following variable:

Code: Select all

GrFnTpXd= "r"                  # [Xd] Green`s function (T)ordered,(R)etarded,(r)senant,(a)ntiresonant [T, R, r, Ta, Ra]
For the expression to each time-ordering you can have a look to:
src/common/Lorentzian_FT.F
The Retarted ordering would provide exactly zero for w=0.
The time ordering is reported in the header of the output files e.g.:

Code: Select all

# - The Green`s function is T-ordered - 
so you can check your output what kind ordering you have in your calculation, and also repeat the calculation with the desired ordering comparing the outputs,
In order to check the parity (e.g. considering also negative energies).

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Post Reply