Query on "ELPhExcStates" and "ELPhExcSum" in Exciton-phonon_coupling_and_luminescence

[sitangshu]

Hello guys,

I have a doubt on exciton-phonon coupling and luminescence.
We are trying to analyse the exciton phonon couplings in a 2D system as per the guidelines in Fulvio's thesis here: https://orbilu.uni.lu/handle/10993/41058.
We have done the following steps:
1. A SAVE folder is generated after performing electron phonon calculations which now contains ndb.elph_gkkp* , ndb.PH_Double_Grid, etc. files. The total number of k and q points are 37 (18 18 1 sampling). The phonon q-double grid is on 60x60x1.
2. A finite momentum BSE is also performed. A total of 5 lowest excitons are used to compute the finite momentum BSE using SlepC. The exciton dispersion seems to be reasonable. All files ndb.BS_diago_Q1, ndb.BS_diago_Q2, ..., ndb.em1s*, ndb.dipoles are inside the folder named "output".
3. The filled electronic states are upto band number 26. A total of 200 bands were included in the calculation.
What we want to know is how to proceed not for the phonon assisted luminosity using exciton-phonon couplings. We are struggling with the last part of the exciton-phonon matrix elements described here: https://www.yambo-code.eu/wiki/index.ph ... minescence

Here are a few queries:
1. We are using the "yambo-excph" version: 5.2.0 Revision 22427 Hash 49ca835f5
2. What should we write in these: "ELPhExcStates" and "ELPhExcSum"? Should it be 1 | 5 in both cases?
3. There are two temperatures now: "BoseTemp" and "EXCTemp". Is the later one exciton thermalization temperature as shown in Fig. 1(c) of the hBN bulk Nature paper by Cassabois et. al. in vol. 10, pp. 262, (2016)?
4. Should this be computed using nompi and apply RIM and RIM_W?

We would be glad for your response.
Regards,

Sitangshu

[sitangshu]

Hi,
Waiting for your response...

Regards

[claudio]

Dear Sitangshu

here my answers

>Here are a few queries:
>1. We are using the "yambo-excph" version: 5.2.0 Revision 22427 Hash 49ca835f5

you should use this version of Yambo
https://www.yambo-code.eu/wiki/index.ph ... g_the_code

> 2. What should we write in these: "ELPhExcStates" and "ELPhExcSum"? Should it be 1 | 5 in both cases?

ELPhExcStates are the states that will emit/absorb light
ELPhExcSum are the states that enters in the exciton-phonon scattering

to be more precise, ELPhExcStates correspond to the \lambda sum in Eq. 17 of https://arxiv.org/pdf/2212.10407
while ELPhExcSum corresponds to the \beta sum of the same equation.

if you are studying the light emission you need very few ELPhExcStates because their occupation is suppressed by the Boltzman
factor and only the lowest will contribute.

ELPhExcSum is a quantity you should converge, increase this number until the result does not change, it converges very fast

You can try to put 2 in ELPhExcStates, and then 3,4,5 in ELPhExcSum to check if result is ok. Then you can also increase ELPhExcStates to 5
I expect results will be the same

>3. There are two temperatures now: "BoseTemp" and "EXCTemp". Is the later one exciton thermalization temperature as shown in Fig. 1(c) of the hBN bulk Nature paper by Cassabois et. al. in vol. 10, pp. >262, (2016)?

Yes there are two temperature because one is the excitonic Boltzman temperature T_exc in Eq. 17 of https://arxiv.org/pdf/2212.10407, while
BoseTemp is the temperature of the phonons, the Bose Factor n_{q,u} in Eq. 17 https://arxiv.org/pdf/2212.10407.
Usually BoseTemp has not a strong effect on the final results. If you look at Fig. S6 of https://arxiv.org/pdf/2212.10407 you can see that in hBN
EXCTemp is about 2 times the BoseTemp.

>4. Should this be computed using nompi and apply RIM and RIM_W?

I think you can calculate with or without mpi anyway calculations are very fast.
The RIM, RIM_W is used in the GW and BSE but once you calculated this two you do not need them anymore,
you just use the already calcualted matrix elements of the BSE

best
Claudio

[sitangshu]

Dear Professor Claudio,

Many thanks for your responses. We did some work on PL emission and indeed we now understood its concept and workflow. I am attaching a relevant plot done by us. Right now I did not find any satellites and replicas, but the structure (possessing a direct optical gap) guarantees that the phonon assisted PL should be small enough.
We are also able to plot (by updating the codes) the necessary excitonic gkkps and their maps.
However, we are stuck on how to obtain the RS lifetime along the excitonic BZ, such as shown in Fig. 2a here: https://journals.aps.org/prl/abstract/1 ... 125.107401
The code only show lifetimes at Q=0. Can you help us modifying the script?

Regards,
Sitangshu

[sitangshu]

Dear Professor Claudio,

I am able to generate the file "o-output.LifeTimes_of_q" which produces this:

Code: Select all

#     |q| (a.u.)         e1                 e2                 e3                 e4                 e5                q_x (rlu)          q_y (rlu)          q_z (rlu)
#
       0.00000000        0.518650625E-2     0.560177065E-2     0.725484749E-2     0.104749948E-1     0.306912690E-6      0.00000000         0.00000000         0.00000000   
      0.667863312E-1     0.292935051E-1     0.295540829E-1     0.206869579E-1     0.205167614E-1     0.213172212E-2      0.00000000        0.555555556E-1      0.00000000   
      0.667863312E-1     0.292934572E-1     0.295540605E-1     0.206012734E-1     0.206027546E-1     0.213149538E-2    -0.555555556E-1    -0.277777778E-1      0.00000000   
      0.667863312E-1     0.292935060E-1     0.295541221E-1     0.206037327E-1     0.206001752E-1     0.213152509E-2     0.555555556E-1    -0.277777778E-1      0.00000000   
      .........			.......				......				....				......			......				.....			......			.....

I suppose that the lifetimes appearing are in (meV)^-1.

If this is so, it also seems that these lifetimes are written for all q-points (expanded) in the BZ. I have 18 18 1, and therefore total q-points written are 324. But neither q-indices nor these lifetimes are written along the high-symmetry route, such as along the exction dispersion. Although I can recognize some q-points at |gamma, M, K, but not in-between. Can you please help me writing lifetimes along the BZ also?

Regards,
Sitangshu

[claudio]

Dear Sitangshu

unfortunately the interpolation in all the BZ is not implemented yet, we will do it soon
in a new release.
You can try to modify the subroutine that calculation interpolation of exciton band structure and force it to read also the life-time,
and then interpolate it.
Or try to use YamboPy.

best
Claudio