Overestimated binding energy

[Nhat Minh Pham]

Dear users and developers,
I'm using Yambo to get monolayer MoS2 optical absorption spectrum from solving BSE.
But the result has a large redshift with over 1 eV of binding energy, which causes the first peak to locate at 1.5 eV.
I used Yambo 5.1 to get a gap of about 2.55 eV and calculated BSE from that with 15x15 grid (non-SO) and
45x45 grid (with SO). They both show the first peak at 1.5 eV, far from PhysRevB.88.045412 result.
Please help me fix it. My BSE input and spectrum are attached below.

[Daniele Varsano]

Dear Pham Nhat Minh,

one inconsistency I can see in your input is that you ask for 20Ry in BSENGBlk, but only 8Ry are considered in NGsBlkXs.
Check if this is a source of error. Next, it would be useful to look at the report file in order to see if there are other problems.

Best,
Daniele

[Nhat Minh Pham]

Thank you, Daniele

My calculation for 45x45 grid was bad, and the fast convergence of binding energy from references made me panic.
BSE is slow to converge with Kpoint-sample. Is there any good way to make it fast as Yambo 5.1 does with GW?

Best,
Minh

[Daniele Varsano]

Dear Minh,

the Wrim technique in principle can be also used for BSE, but I would not suggest using it. Even if it can speed up the calculation, it can also brings you to non accuarte results depending on the nature (localized or delocalized) of the exciton.

Best,

Daniele