Bright vs dark exciton

claudio · Post by **claudio** » Sat Dec 13, 2025 3:52 pm

I move this discussion here from my website:
https://www.attaccalite.com/yambo-input ... /#comments

I hope you are doing well. I am using Yambo to study excitonic properties, and I am a bit confused about how to identify bright and dark excitons in my calculations.

Specifically, I would like to understand:

How to distinguish optically allowed (bright) and optically forbidden (dark) excitons in Yambo.
Which excitons follow optical selection rules and which ones are forbidden.
How this classification is affected by spin–orbit coupling (SOC), since my calculations include SOC.
How to analyze this at q = 0, where optical transitions are expected.
I could not find clear documentation or examples addressing the spin character of excitons and the bright/dark classification in SOC-based Yambo calculations, so I would really appreciate your guidance on this.

Abbas
Master Student
IIUI Isalamabad pakistan

claudio · Post by **claudio** » Sat Dec 13, 2025 4:00 pm

Dear Abbas

I reply you here in the Lumen forum:

you can plot exciton dipole with the command

Code: Select all

 ypp -e s

this will produce two file with excitons sorted by energy and dipole intensity.

In this way you can distinguish dark and bright excitons.
For more infor see here:
https://www.lumen-code.org/wiki/index.p ... e_excitons

Then you can also look at YamboPy for further analysis see here:
https://www.lumen-code.org/wiki/index.p ... in_YamboPy

Let me know if it is clear
best
Claudio

abbas05 · Post by **abbas05** » Tue Dec 23, 2025 1:51 pm

thanks sir
How does Yambo allow us to separate symmetry-forbidden (zero dipole) excitons from spin-forbidden excitons in BSE calculations with SOC.

abbas
master student
IIUI islamabad Pakistan

Davide · Post by **Davide** » Mon Dec 29, 2025 10:09 am

In a calculation with SOC, Sz is not a good quantum number. Accordingly, transitions are not exactly spin flip or spin conserving.

However, if SOC is small, the concept holds with a good degree of approximation.

You can compute <Sz>. Input file generation with

Code: Select all