how to extract the transition from ypp output

[zsjan]

Dear all:
I have calculated the optical spectra in transition space with yambo -o b -k hartree and now I want to extract the individual transition that contributes to the spectra. However, when I specify the states entry in the inputfile for ypp using ypp -e a, it seems that I still cannot get such information. For example, when I set States= "1 - 10", I am expecting it gives me the contribution from transition from state 1 to state 10, is it right? however, the obtained o.exc_amplitude_at_1 file always gives me the highest peak about exciton contribution. Is something wrong? The log file is attached.

[Daniele Varsano]

Dear Shijun,
no attached file is found.
Anyway if you specify "1-1)", ypp will give you the KS transition that mostly contribute to the first 10 excited states (one excited states in each output file).
So, o.exc_amplitude_at_1 will give you the KS contributions to the first excited states (indexed with 1) in o.exc_E_sort.
Hope it is clear,
Best,

Daniele

[zsjan]

Dear Shijun,
no attached file is found.
Anyway if you specify "1-1)", ypp will give you the KS transition that mostly contribute to the first 10 excited states (one excited states in each output file).
So, o.exc_amplitude_at_1 will give you the KS contributions to the first excited states (indexed with 1) in o.exc_E_sort.
Hope it is clear,
Best,

Daniele

Sorry, but I still cannot understand. Actually, when I change the states variable, the obtained o.exc_amplitude_at_1 is always the same. For example, 1-2, 1-3, 1-10... Is this reasonable? The log file of ypp is the following, in which the " Merged with states 1 -> 13 " always appears, why?


<---> [01] Y(ambo) P(ost)/(re) P(rocessor)
<---> [02] Core DB
<---> :: Electrons : 8.000000
<---> :: Temperature [ev]: 0.000000
<---> :: Lattice factors [a.u.]: 3.851738 3.851738 3.851738
<---> :: K points : 10
<---> :: Bands : 50
<---> :: Symmetries : 48
<---> :: RL vectors : 1831
<---> [03] K-point grid
<---> :: Q-points (IBZ): 19
<---> :: X K-points (IBZ): 10
<---> [04] CORE Variables Setup
<---> [04.01] Unit cells
<---> [04.02] Symmetries
<---> [04.03] RL shells
<---> [04.04] K-grid lattice
<---> [04.05] Energies [ev] & Occupations
<---> [05] Excitonic Properties
<---> [M 0.282 Gb] Alloc BS_mat ( 0.281)
<---> :: Sorting energies
<---> [05.01] Excitonic Amplitude
<---> Processing 1 states
<---> :: State 6 Merged with states 1 -> 13
<---> [05.02] Reporting amplitude and weights
<---> [M 0.001 Gb] Free BS_eh_table BS_mat A_weight ( 0.281)
<---> [06] Game Over

The logfile of the calculations are attached now!

[Daniele Varsano]

Dear Shijun,
the code merges states because they are very closed in energy. Check it in your excitation sorted file.
You can avoid to merge them by setting the variable:

Code: Select all

Degen_Step=   0.0100   eV   

to a resolution smaller than the energy difference between your excitations.
When you change the states variable form 1-2, 1-3, 1-10, you are calculating the exciton amplitudes for the
excitons wit indexes 1 and 2 (you will have two files)
excitons with indexes 1,2 and 3 (you will have 3 files0
excitons with indexes 1,2, ....up to 10. (you will have 20 files)

If there are degeneracies they can be merged, but you can avoid this as stated above.

o.exc_amplitude_at_1 is always related to the first exciton, you are just recalculating it many times.

Best,
Daniele

[zsjan]

Dear Shijun,
the code merges states because they are very closed in energy. Check it in your excitation sorted file.
You can avoid to merge them by setting the variable:

Code: Select all

Degen_Step=   0.0100   eV   

to a resolution smaller than the energy difference between your excitations.
When you change the states variable form 1-2, 1-3, 1-10, you are calculating the exciton amplitudes for the
excitons wit indexes 1 and 2 (you will have two files)
excitons with indexes 1,2 and 3 (you will have 3 files0
excitons with indexes 1,2, ....up to 10. (you will have 20 files)

If there are degeneracies they can be merged, but you can avoid this as stated above.

o.exc_amplitude_at_1 is always related to the first exciton, you are just recalculating it many times.

Best,
Daniele

Thank you very much!
However, I still have a question. The results from states 2-2, 3-3, 4-4... are the same for my calculations: the files o.exc_amplitude_at_2, o.exc_amplitude_at_3, o.exc_amplitude_at_4... have the same contents. why?
Anyway, I want to know the transition contributed from state M to state N. How to get this information?

[Daniele Varsano]

Dear Shijun,

The results from states 2-2, 3-3, 4-4... are the same for my calculations: the files o.exc_amplitude_at_2, o.exc_amplitude_at_3, o.exc_amplitude_at_4... have the same contents. why?

There can be degeneracies, anyway it is impossible to say anything without knowing what are you calculating. Try to understand what is happening, if it makes sense or not. If you want explanations, please post all the information that can be useful (report/outputs).

I want to know the transition contributed from state M to state N. How to get this information?

It is not clear to me what you mean for state M and N. Here you are calculating the transition contributing to the excitation from the ground state to the first excited states (1,2,3 etc.).

Best,
Daniele

[zsjan]

Dear Shijun,

The results from states 2-2, 3-3, 4-4... are the same for my calculations: the files o.exc_amplitude_at_2, o.exc_amplitude_at_3, o.exc_amplitude_at_4... have the same contents. why?

There can be degeneracies, anyway it is impossible to say anything without knowing what are you calculating. Try to understand what is happening, if it makes sense or not. If you want explanations, please post all the information that can be useful (report/outputs).

I want to know the transition contributed from state M to state N. How to get this information?

It is not clear to me what you mean for state M and N. Here you are calculating the transition contributing to the excitation from the ground state to the first excited states (1,2,3 etc.).

Best,
Daniele

Well, the system I tested is the LiF as provided in the tutorial. There is a peak around 9eV in the RPA level. However, my o.exc_amplitude_at_* always have a peak around 9eV. The output of RPA calculations have been uploaded in the above.
What I want to see is that is it possible to identify where the peak at 9eV come from? For example, from the Li(s) state to Li(p) state....

[zsjan]

Dear Shijun,

The results from states 2-2, 3-3, 4-4... are the same for my calculations: the files o.exc_amplitude_at_2, o.exc_amplitude_at_3, o.exc_amplitude_at_4... have the same contents. why?

There can be degeneracies, anyway it is impossible to say anything without knowing what are you calculating. Try to understand what is happening, if it makes sense or not. If you want explanations, please post all the information that can be useful (report/outputs).

I want to know the transition contributed from state M to state N. How to get this information?

It is not clear to me what you mean for state M and N. Here you are calculating the transition contributing to the excitation from the ground state to the first excited states (1,2,3 etc.).

Best,
Daniele

Well, the system I tested is the LiF as provided in the tutorial. There is a peak around 9eV in the RPA level. However, my o.exc_amplitude_at_* always have a peak around 9eV. The output of RPA calculations have been uploaded in the above.
What I want to see is that is it possible to identify where the peak at 9eV come from? For example, from the Li(s) state to Li(p) state....

[Daniele Varsano]

Dear Shijun,

what you need to do is to sort your excitation energies (ypp -e s). Inspect the output and you can recognize the peak looking at the excitation energies and oscillator strength. You can also plot them (energy wrt intensity) on top your spectrum. Once you have individuated the excitations, you look at the corresponding index and analyze it with ypp -a (setting in input the index).
The output will give you the KS orbital participating in that excitation. Then you can inspect the character of that orbital by plotting them.
It is possible that you have many excitations beneath the peaks at 9eV.
Best,
Daniele

[haseebphysics1]

Dear Daniele,

As I have already confirmed from you in another post that when we run yambo - o b -k hartree , then we will be doing RPA (including LFE) but in transition space. But the input file made contains parameters from BSE calculation. And you clarified that in this case BSEBands will acts like BndsRnXd as was used in RPA level optics. Now, the same calculations which was done in chi space, I am doing that in transition space to know the electronic transitions. But my calculations is not processing probably due to memory requirements!

I think I can't include the same number of BSEBands as I was doing in BndsRnXd, correct? input and log files are attached.


Thanks,