how to extract the transition from ypp output

[haseebphysics1]

Dear

I've followed the following tutorial (although not reproduced it!)
http://www.yambo-code.org/wiki/index.ph ... e_excitons
Here it is written that "Open the first file and look inside. The first exciton is at 4.83 eV"...

And in the weight file, there is the following output,

Code: Select all

#  Band_V     Band_C     K  ibz     Symm.      Weight     Energy
#
 4.000000   5.000000   7.000000   2.000000   0.922095   4.401093
 4.000000   5.000000   7.000000   1.000000   0.922086   4.401093
The first exciton is essentially done of only single particle transitions from VBM to CBM at K (last k-point of the grid).

So, if the last column is the KS ENERGY DIFFERENCE NOT THE QP GAP? Because it can not be smaller than the exciton position (4.8 eV)! Am I right here?

2: Moreover, how it was concluded from the above data,

"The first exciton is essentially done of only single-particle transitions from VBM to CBM at K (last k-point of the grid)."?
Because in the data there are two transactions from band 4 to 5 and was k-point # 7 was the last in IBZ? And is this a direct transition because no K-point of CB was given?

3: And lastly, in the same tutorial, the amplitude vs energy plot is given, there is a peak at 4.4 eV, again at the KS level, not at the 4.83 eV (exciton)? Were not we supposed to plot the exciton peak instead of KS-peak?

Many thanks,

[Daniele Varsano]

Dear Haseeb,

So, if the last column is the KS ENERGY DIFFERENCE NOT THE QP GAP? Because it can not be smaller than the exciton position (4.8 eV)! Am I right here?

It is the KS energy difference, reported for each transition participating, what is not clear here?
In general it can be also smaller if the exciton is not bound, note that it is KS and not QP. It is anyway an information added to the output, if it causes you a problem you can just ignore it, it is not needed to calculate the binding energy.

2. What is not clear here? two transitions participate with equal weights, they correspond to a particular k point in the IBZ and another k point related by. symmetry to the first one. It is a direct transition because in the GPL version only BSE for q=0 is implemented, now in the 4.5 you have an output prepared also for finite Q BSE, but actually that part of the code has been not yet released.

3. Because that plot it is not a spectrum and it provides different information, i.e. the KS amplitude of the transition participating to form that particular exciton.

Best,
Daniele

[haseebphysics1]

Dear Daniele, thank you, it is much clear now to me!

I was aware that exciton can always be seen before the bandgap but I was mixing KS with the QP gap.

And the second reason was asking this in the calculation of B.E! since, I've not done the GW QP corrections, but I will take the exp gap as a reference. So, to find the exciton B.E. what I will do is to add the scissor corrections in the energies of the o.exc_qpt1_E_sorted file and then subtract that from the experimental gap.
i.e;

Exciton B.E = Exp gap - Energy of o.exc_qpt1_E_sorted + sciccor correction

And similarly for other excitons...

Am I right in this procedure, dear Daniele?


Thanking you,

[Daniele Varsano]

Am I right in this procedure, dear Daniele?

yes, it makes sense!

Daniele

[haseebphysics1]

Dear Daniele,

What do you have in the 4th file?

For state 1-3, I have four files for weight and amplitude and are attached. Similarly, if I put 4687-4687 in states variable, I got a whole range of files! more than 200 files! I was supposed to get only one for state # 4687!

2:

in the o.exc* are reported the eigenvalues and intensities for the complete spectrum of your BSE independently to the energy interval the spectrum is plotted.

So, in the o.exc* ypp calculations how we can control the energy windows? How ypp decides the range itself if it does not take the range from the BSE input spectrum?

3: In all the optical calculations for a single q point (q > 0), we also have eels spectra! So, I want to know how significant is this because we are allowing to change moments and still getting eels data?

4: In ypp output can we get some info somewhere about the f-sum rule (Thomas-Reiche-Kuhn sum rule)?

5: Lastly, about the meaning of singlet and triplet in BSE! I have read if we switch off the exchange we will go to the triplet calculations! But these kinds of terminology, we use in Spin-Orbit Coupling! Right? So, is this BSE has something to do with SOC effects?

Thanks,
Take care and stay safe,

[Daniele Varsano]

Dear Haseeb,

1. This happens because you have degenerate excitons: from the report:

Code: Select all

 [WARNING] 1 excitonic states added to the USER selection to preserve degeneracy

note 3 and 4 are the same as they are degenerate and yambo considered a linear combination.
The energy threshold when 2 excitations are considered degenerate can be controlled by input (Degen_Step).

2. The number of excitations is equal to the dimension of the matrix that is diagonalized.

3. BSE is for q=0 and you have the eels for q=0. If you calculate spectra in linear response you can also calculate eels for any q > 0.

4. In order to fulfil the f-sum rule you need to integrate to all the energy, it hardly converges with respect the number of included bands anyway you should have all the info to calculate it

5. Singlet and triplet are meaningful with a collinear spin description. In the presence of SOC you deal with spinors and they are not well defined anymore.

Best,
Daniele

PS: It seems that your weight files have been modified, attaching the content of the amplitude.

[haseebphysics1]

Dear Daniele,

In the general ypp weight output file, at the beginning of the weight file there are a lot of lines for K-points (transitions?) as compare to their corresponding details of VB and CB info(below in the same ypp weight file)! For example in the attached file contains just 5 lines of VB and CB info while a lot for K-points! And their weights are not corresponding with each other? So, how we can be sure that for a particular k-value which v>c transition is happening?

Should we consider the max weight from k-list and also from bands info list and match those?

Code: Select all

#                 K-point [iku]             Weight
#     0.000000     0.000000     0.000000     1.000000
#     0.000000     0.000000     0.166667     0.466940
#     0.000000    -0.500000     0.000000     0.659483
#     0.000000    -0.500000     0.166667     0.367642
#     0.166667     0.000000     0.000000     0.726685
#     0.166667     0.000000     0.166667     0.187905
#     0.166667    -0.500000     0.000000     0.593406
#     0.166667    -0.500000     0.166667     0.169595
#      0.33333      0.00000      0.00000      0.06427
#      0.33333     -0.50000      0.00000      0.06216
#    -0.166667     0.000000     0.166667     0.188187
#    -0.166667     0.500000     0.166667     0.169848
#    
#    Band_V       Band_C       Kv-q ibz     Symm_kv      Kc ibz       Symm_kc      Weight       Energy
#    E    [eV]    Amplitude
 160.0000   162.0000       1.0000       1.0000       1.0000       1.0000       0.1871       2.3594
    160.0        161.0        5.000        1.000        5.000        1.000       0.7036E-1     2.413   
    160.0        162.0        9.000        1.000        9.000        1.000       0.6275E-1     2.443   
    160.0        162.0        9.000        2.000        9.000        2.000       0.6271E-1     2.443   
    159.0        162.0        5.000        1.000        5.000        1.000       0.5432E-1     2.442   

And dear Daniele, does the attached pic seems converged BSE in terms of K-points? And if we have the converged BSE spectrum then will the same # of k-points also be sufficient for the exciton plot?

Thanks,

[Daniele Varsano]

Dear Haseeb,
1. it seems to me that you are thinking things are more complicated than they are.
The first commented lines are transition ordered by k points and integrated over all cv transitions at that k point.
Below you have the transition detailed for each K point and c,v pair.
Note that first set (kpoint integrated over cv) weights are normalized to the larger value.
The integrated values are useful if you want to represent the contribution of the excitations in the BZ independently in the particular cv transition at that k point.

2. Yes, it is perfectly converged.
2b Exciton plot, try it and see what does it change, then you decide.


Daniele

[Reza_Reza]

Dear Daniele,

I am trying to understand the transitions from the ypp output. However, I have some questions and ambiguity. In an example file that I attached the optical absorption, I have computed the exciton amplitude. The energy range between the transition corresponding to the strongest exciton (index = 1) and the exciton energy is different. Here is an output file:

# Electron-Hole pairs that contribute to Excitonic State 1 for iq=1 more than 5.000000%
# K-point [iku] Weight
# : 0.469999999 0.00000000 0.00000000 0.128007874
# : 0.479999989 0.00000000 0.00000000 0.391013265
# : 0.490000010 0.00000000 0.00000000 1.00000000
# : -0.500000000 0.00000000 0.00000000 0.630462587
#
#
# Band_V Band_C Kv-q ibz Symm_kv Kc q ibz Symm_kc Weight Energy [eV]
#
66 67 51 1 51 1 0.283119 0.079455
66 67 50 5 50 5 0.224529 0.141083
66 67 50 1 50 1 0.224528 0.141083
66 67 49 5 49 5 0.087780 0.246223
66 67 49 1 49 1 0.087780 0.246223
#
# 01/02/2025 at 17:15 ypp @ mc-n020.mesu [start]
# 01/02/2025 at 17:15 [end]
#
# .-Input file exciton_amp.in
# | excitons # [R] Excitonic properties
# | amplitude # [R] Amplitude
# | States= "1 - 5" # Index of the BS state(s)
# | BSQindex= 1 # Q-Index of the BS state(s)
# | Degen_Step= 0.010000 eV # Maximum energy separation of two degenerate states
-------------------------------------------------------------------------------------------------------------------------
The energy is between 0.246223 and 0.079455, which is very different from
# E [ev] Strength Index
#
0.516867377E-1 1.00000000 1.00000000
in qpt1_E_sorted file that I attached here. What is the reason for that?

Best,

Reza

[Daniele Varsano]

Dear Reza,

these are two different quantities. The exciton energy is the eigenvalue of the excitonic BSE hamiltonian, while in the weight output are reported the Kohn-Sham energy differences of the transitions participating in that excitons together with their weights.

Best,
Daniele