I think you should take care of all possible interactions to get close to the real picture. In period solids, there might be electron-hole interaction (solve BSE equation for it), el-phonon coupling, etc. Note that, QE epsilon.x is the lowest level of approximation for the solids (IPA) and it doesn't include any non-local part and local field effects. Moreover, you can incorporate the many-body effects in the solids using certain TDDFT kernels as well, like TDDFT-LRC (static and dynamic) and if you have solid with low electronic dielectric function and small lattice screening than the excitonic effects might play a big role even in the bulk systems, so you better solve BSE in those cases instead of RPA level of spectra. I recommend to take your system into account carefully and then choose the right theory to apply.

Best of luck,

Haseeb Ahmad.

I think you should take care of all possible interactions to get close to the real picture. In periodic solids, there might be electron-hole interaction (solve BSE equation for it), el-phonon coupling, etc. Note that, QE epsilon.x is the lowest level of approximation for the solids (IPA) and it doesn't include any non-local part and local field effects. Moreover, you can incorporate the many-body effects in the solids using certain TDDFT kernels as well, like TDDFT-LRC (static and dynamic) and if you have solid with low electronic dielectric function and small lattice screening than the excitonic effects might play a big role even in the bulk systems, so you better solve BSE in those cases instead of RPA level of spectra. I recommend to take your system into account carefully and then choose the right theory to apply.

Best of luck,

Haseeb Ahmad.