Yambo Community Forum

Hello,

Is it possible to get very fine splitting 1s, 2s exciton peaks i.e Rydberg series by performing BSE calculations using Yambo code as shown in this paper for GaAs (Fig2: https://doi.org/10.1103/PhysRevLett.81.2312). These peaks are lying in meV range. Basically I am interested only in the bound exciton region i.e. very narrow region, let say for GaAs in meV range.

In this paper "Phys. Rev. B, 86, 125203 (2012)", main focus is to get the optical absorption spectrum close to the experimental spectra by speeding up the solution of the Bethe-Salpeter equation by a double-grid method and Wannier interpolation. However, in my case, I just want to have a clear picture of the bound exciton as shown in the (Fig 2, https://doi.org/10.1103/PhysRevLett.81.2312). My question is, could it be possible to capture these very fine splitting of the exciton (in meV range) using Yambo code.

If you can share few papers focussing on exciton fine splitting using Yambo code would be really helpful.

Thanks,

Dear Pooja,
please sign your post with your name and affiliation, this is a rule of the forum, and you can do once for all filling the signature in your user profile.

You can capture the fine splitting, of course it is a matter to push the convergence parameters, so it depends on the availability of your computational resources.

A way to push with respect the k points is the double grid method you indicated in your post (Phys. Rev. B, 86, 125203 (2012)), anyway pictures like Fig.2 for GaAs can be probably obtained with standard methods.

Best,
Daniele

Dear developers,

Could you please help me to know, is it possible in yambo code to select the finite K points near the band extrema only, not the whole Brillouin zone, for the BSE calculations to obtain a converged representation of the 1s, 2s, and 2p excitons (bound excitons). If yes, please let me know, how to proceed or what are the steps to be taken into account?

Thanks,
Pooja

Dear Pooja,
I'm afraid this is not possible as the BSE is built considering transitions for the whole BZ, but maybe others can give you some hints.

Best,
Daniele

In this paper PRL, 81, 2312 (1998), they have mentioned that:

"Figure 1 does not exhibit bound exciton states because of the limited resolution 􏰈􏰀0.15 eV􏰁 arising from the finite set of 500 k points in the BZ used. To describe the discrete exciton states, we have to increase the k-point grid density to 10^8 k points in the Brillouin zone. Fortunately, the use of about 1000 of these k points near the position of the band extrema allows already for a converged representation of the 1s, 2s, and 2p excitons."


My question is, Is it possible using yambo code to impose the finite K-point sampling at the band extrema only, to observe the exciton fine splitting, like in the mentioned paper they have used 1000 of K points near band extrema? I hope my question is clear to you.

Thanks,
Pooja

Your question is clear, and I hoped my answer was too. Yambo implements bse on a regular grid spanning the Bz, so what you are asking is not possible. What you can do is to calculate relevant quantities as the screening and bands ona a regular grid, and then implement a bse on an effective mass model, interpolate the screning in a denser grid in your region of interest and then include the transition of interest in your model.
Best,
Daniele

Dear Daniel,

Thankyou for your quick response,

Could you suggest me tutorial to perform these steps in yambo, basically how to implement bse on an effective mass model?

Thankyou for your patience
Regards,
Pooja

Dear Pooja,
there is not yambo tutorial for that, I just suggested a way to overcome your problem. Finally it is very similar to what it is done in the Rohlfing and Luoie paper where the screening is calculated for a coarse grid and used for the finer grid. In any case, you should implement a BSE calculation on a small portion of the BZ.

Best,
Daniele

Dear Pooja,
maybe this thread can be useful for you:
viewtopic.php?p=9867#p9867

Best,
Daniele