Only reflectivity not RAS

[burkzdemir]

I want to calculate the reflectivity of a slab structure. Can I use the real-space cutoff option for epsilon in yambo -o c / -o b calculations not in RAS calculations?

[Daniele Varsano]

Dear Burak,
the cutoff coulomb potential it is meant to avoid spurious interaction in non-periodic direction in non3D systems.
It is not related to any specific kind of calculation, where did you find that it is related to the RAS?
Anyway, the cutoff potential it is not in real space, but in reciprocal space (see Phys. Rev. B 73, 205119 (2006).).
NB: if you plan to use a box shape for your slab, have in mind that the side of the box have to be doubled, it is discussed in the forum.
Best,
Daniele

[burkzdemir]

I read about the cutoff used in RAS calculation mentioned in the yambo web-site (http://www.yambo-code.org/tutorials/Sur ... /index.php).

Cutoff # [RAS] Cutoff mode (Uncomment to use)
CutZero= 0.250000 # [RAS] Zero position of cutoff fn (Frac)
CutStep= 0.500000 # [RAS] Width of cutoff function (Frac)

[Daniele Varsano]

Dear Burak,
I understand now, you mean the "boxcar" function used to calculate the alpha_surf, introduced here:
C. Hogan, R. DelSole, G. Onida, Phys. Rev. B 68 (2003) 035405.
while I was talking about reciprocal space cutoff, sorry for the misunderstanding.
Unfortunately I cannot help on that, but people working with surface with yambo will give you the answer.
Best,
Daniele

[Conor Hogan]

Dear Burak,
The RAS module will also generate the files o.eps_cell-rpa which contain the cell dielectric tensors including the cutoff. From these, you can compute whatever you like (using standard relations for Fresnel reflectivity, or perhaps more advanced ones including the slab anisotropy/light polarization). However I would recommend doing everything first without cutoff and a symmetric slab, to make sure that you scale the components correctly.
The real space cutoff can only be used however for the basic level: RPA without local fields.
Conor

[burkzdemir]

Can it be used for RPA+Drude?

[Conor Hogan]

It should work, yes, although I've never tried it. The surface spectra module is (more or less) just a driver for the normal optics routines, and so the subroutine X_Drude is called in the normal manner.