Phonon assisted excitonic DOS

[sitangshu]

Dear Yambo guys,

Sorry, I am posting this here. I couldn't find a broad index for this topic.

1. I am practicing a phonon assisted PL spectra in hBN monolayer following the tutortial: https://www.yambo-code.eu/wiki/index.ph ... emperature I am using the version 5.1.0 Revision 21422 with SlepC. For now, I haven't used the nondiagonal supercell + finite displacement method to compute this PL spectra.

2. The doubt I am facing is about the term "DOSERange". What does this means?

3. I am familiar with the thesis by Paleari and can recover the temperature dependent (q -->0) optical spectra with double grids (both nscf & matdyn) and indirect exciton dispersion. I found out the hBN monolayer exciton dispersion which seems to be correct

4. However, while running the ypp_ph -e p after generating the (converged) gkkp elements (ndb.PH_Double_Grid), when I plot "o-output.excitons_ph_dos.50K" I find the spectra is shifted in energy axis? Is this plot reasonable? I am attaching the two files o- and l-output.


Regards,
Sitangshu

[claudio]

Dear Sitangshu

> 2. The doubt I am facing is about the term "DOSERange". What does this means?

is the range of energy where you want to calculate the phonon-assisted transitions

> 4. However, while running the ypp_ph -e p after generating the (converged) gkkp elements (ndb.PH_Double_Grid), when I plot "o-output.excitons_ph_dos.50K" I find the spectra is shifted in energy axis? Is this plot reasonable? I am attaching the two files o- and l-output.

if I open the your report file r-output_excitons_ph_ass_dos
I notice that you have an exciton at q=(1/3, 1/3,0) with energy E_min=5.18 eV

so your spectra is not so strange, because the peaks appear at en energy E_min - E_ph that should be compatible with what you found.

Also in your interpolated file, "bn.o-output.excitons_interpolated" you can see the lowest exciton around 5.18 eV.

The phonon-assisted spectra correspond to the luminescence spectra without considering the exciton-phonon matrix elements and dipoles.
This means that it just gives you a density of states of where you can expect peaks.
Then in the reality some of those peaks can be dark to the matrix-elements that are absent in the plot, and their relative high also depends from these matrix
elements that are missing.
Anyway this simple tool give you an idea of where could be located the luminescence peaks.

Let me know if you need more explanations

best
Claudio

[sitangshu]

Many thanks Professor Claudio for your comments. Certainly, the exciton-phonon matrix elements are not here so, I can't say a more detailed scattering mechanism.
1. But It seems that the higher end of the peaks follows an exponential decay. I don't see however see any phonon replica like in bulk hBN (need more check probably).
2. How can I assign the phonon modes to the peaks (like in Cassabois, Valvin and Gil, 2016)? Do I need the Eliashberg spectra to correlate?
3. What could the be experimental DOS broadening range when actual experiments are performed?

Regards
Sitangshu

[claudio]

Dear Sitangshu

in order to check which phonon mode is responsible of your spectra
you can modify the file

ypp/excitons/excitons_ph_ass_dos.F

just change the loop

do i_l=1,ph_modes

in somethinkg like

do i_l=1,1 ! to check the first mode

than

do i_l=2,2

and so on.

In this case the code will produce only the replica due to the first phonon, than the second an so on.
I will add this in the input in the new version, in this way you can also assign the different peaks at the different phonon modes.

best
Claudio