calculating triplet exciton in BSE.

[jack]

Dear all,

I am new to Yambo and currently trying to calculate excitons using the BSE method. However, I am a bit confused about how triplet excitons are handled in BSE. I have read several discussions on the forum regarding triplet calculations and have summarized what I understand so far as follows:

1. For a non-spin-polarized calculation, setting BSENGexx = 0 eV yields the singlet–triplet transition energy.

2. For a spin-polarized calculation, Yambo computes both singlet and triplet excitons even without setting BSENGexx = 0 eV.

3. For magnetic systems, setting BSENGexx = 0 eV behaves the same way as described in point 2.

4. For magnetic systems, spin-flip transitions correspond to magnons when setting BSEprop = "magn".

5. Does the option BSKmod = "TRIPLET" work? I noticed that the forum often recommends using BSENGexx = 0 eV instead to include triplet excitons.

Could you please let me know whether my understanding is correct?

Thank you in advance for your guidance.

[Daniele Varsano]

Dear Jack,
please sign your posts with your full name and affiliation, this is a rule of the forum and you can do once for all by filling your signature in your user profile.
Yes, what you report is essentially correct.

1. Yes, when you set the exchange kernel to zero you compute triplet excitations and this gives access to s-t splitting.
2. In non-collinear case (spinors) you have access to all the excitations but note they cannot be strictly classified as singlets and triplets.
3. Magnetic systems can be treated in collinear and non-collinear framework, in non-collinear case you have all the excitations, also in collinear case you cannot distinguish anymore singlets and triplets (The ground state already breaks spin symmetry), what you can distinguish spin-conserving and spin-flip excitations.
4. Correct
5. BSKmod = "TRIPLET" is not recognized. Triplets can be calculated in non spin polarized calculations setting the exchange kernel to zero.

Best,

Daniele

[jack]

Dear Jack,
please sign your posts with your full name and affiliation, this is a rule of the forum and you can do once for all by filling your signature in your user profile.
Yes, what you report is essentially correct.

1. Yes, when you set the exchange kernel to zero you compute triplet excitations and this gives access to s-t splitting.
2. In non-collinear case (spinors) you have access to all the excitations but note they cannot be strictly classified as singlets and triplets.
3. Magnetic systems can be treated in collinear and non-collinear framework, in non-collinear case you have all the excitations, also in collinear case you cannot distinguish anymore singlets and triplets (The ground state already breaks spin symmetry), what you can distinguish spin-conserving and spin-flip excitations.
4. Correct
5. BSKmod = "TRIPLET" is not recognized. Triplets can be calculated in non spin polarized calculations setting the exchange kernel to zero.

Best,

Daniele

Dear Daniele,

Thank you for your response. I am now trying to calculate excitons in a magnetic system and have two further questions:
1. If I set BSENGexx = 0 eV, will I be able to obtain the triplet exciton within the spin-conserving channel?
2. I noticed that the optDipAverage tag can be used to account for material anisotropy. However, I am not sure in which calculation this tag is effective—does it apply to the dielectric constant calculation or the GW calculation?
Best,
Jiang