full GW+BSE calculation

[Daniele Varsano]

Dear Khaled,

1) When you solve the BSE, the meaningful part of the exciton energy is the real part. If you are interested in exciton lifetime you need to do a model for it first.
2) If you want to extract the transition dipole matrix you can write them either looking into ./src/bse/K_diag_compute_epsilon.F, when the epsilon is built after the diagonalization of the dielectric matrix.
Otherwise with the postprocessing ypp, when excitons are sorted out (have a look in ./ypp/exciton_sort_and_report.F), the BS_R are read, normalized and printed out.

Best,

Daniele

[Khaled]

Dear Dr. Daniele Varsano,

Thanks a lot Dr. Daniele for your great help. I was able to extract the transition dipole matrix, but I am not sure a bout the units of the transition dipole matrix $ | \sum_{vck} \langle vq-k | e^{-iqr} | ck \rangle A_{\lambda} |^{2} $
because I extracted this term alone (even I couldn't figure out if you used the electric dipole approimation for this term or not). Please, Can you help me for getting this information. Thank you in advance.

Best regards,

Khaled.

[Daniele Varsano]

Dear Khaled,
the internal unit in the code is the atomic unit (bohr for length, Hartree for energy etc..).
The electric dipole approximation it is used for the q=0 term.

Best,

Daniele

[Flex]

What is exactly the unit of the matrix elements ?

A transition dipole has SI units of Coulomb.meters, so here, should it be electron charge.bohr ?

[Daniele Varsano]

Dear Thierry,
yes, it should be as you wrote.

Daniele

[haseebphysics1]

Dear Yambors,

I have a pretty simple question.

In the BSE input file, one of the important parameter is BSSmod:

BSSmod= "d" # [BSS] (h)aydock/(d)iagonalization/(i)nversion/(t)ddft`

I was curious about the last option (t), here t represetns TDDFT? since it is just the method to solve BSE equation, so what does TDDFT means here. I have done TDDFT using LRCs approx. already.


Thanks,

[Daniele Varsano]

Dear Haseeb,
this variable was meant for a tddft kernel derived from the BSE equation: see e.g. Phys. Rev. Lett. 91, 256402 (2003). This feature is not present anymore in the code.
Best,
Daniele