[QE-developers] turboEELS related question

[Moldabekov, Zhandos (FWU-D)]

Dear turboEELS developers,

My name is Zhandos Moldabekov. I work in a theoretical group 
investigating warm dense matter at the Helmholtz-Zentrum Dresden-Rossendorf.

I used turboEELS to compute the dynamic structure factor of electrons in 
warm dense matter (dense plasmas) [ arXiv:2502.04921 ]. I conducted a 
few tests comparing the results of turboEELS with those from "standard" 
LR-TDDFT and quantum Monte Carlo methods. The performance and accuracy 
of turboEELS are exceptionally good. Currently, turboEELS is the primary 
code I use for analyzing X-ray Thomson scattering (XRTS) of hot dense 
matter.

I have a technical question regarding the variable "approximation." In 
the turboEELS paper published in Computer Physics Communications (2015), 
it states that for the GGA functional, the Time-Dependent Generalized 
Gradient Approximation is used.

Does this imply that adiabatic GGA  exchange-correlation kernel is 
computed as described in Phys. Rev. B 99, 035151 (2019)?

Or one does not need an XC kernel in the Liouville-Lanczos method; thus, 
turboEELS uses only the XC potential?

I would appreciate your feedback.

Best,

Zhandos