[salmon-user:00383] Questions about functional and spin calculations

[沈荣琦]

I'm currently calculating the ground state of a fused silica model containing an oxygen vacancy defect. I'm encountering a problem where spin calculations and the functional TBmBJ cannot be used simultaneously; using them together results in an error. If I don't calculate the spin and only use the functional TBmBJ, the program runs normally. I don't know what the problem is. I've attached the input files to the email; I hope you can take a look and look forward to your reply.

[Yamada Shunsuke]

Thank you for the question and I am so sorry for that inconvenience. Currently, TBmBJ with spin is not implemented yet. We will improve this point in a future version up.

Sincerely,

Shunsuke 

---

差出人: 沈荣琦 <srqla...@126.com>
送信日時: 金曜日, 11月 14, 2025 10:37 午前
宛先: salmo...@salmon-tddft.jp <salmo...@salmon-tddft.jp>
件名: [salmon-user:00383] Questions about functional and spin calculations

 

[沈荣琦]

Another question I'd like to ask is whether a fused silica model containing a single oxygen vacancy defect needs spin calculations? I noticed that single-crystal silicon models containing a single vacancy defect don't require spin calculations.

[Yamada Shunsuke]

Dear user,

Whether spin calculations are necessary depends on the purpose of your study. For example, if you only want to investigate properties of excitation energies, spin is often not required.

However, this issue is highly system-dependent, so in practice, you may need to explore it through trial and error. We do not have prior experience with this specific case, so it could be considered part of new research.

If the reason for including spin was simply because the calculation did not converge, adjusting some input parameters may help improve convergence without spin.

For example, try the following settings in the &scf section. Mixrate is the mixing parameter for the simple mixing method of the electron density in the SCF loop. If convergence is still problematic, decrease mixrate further:

  method_mixing = 'simple'

  mixrate = 0.1

  yn_preconditioning = 'y'

Sincerely,

Shunsuke

差出人: 沈荣琦 <srqla...@126.com>
日付: 土曜日, 2025年11月15日 11:54
宛先: salmo...@salmon-tddft.jp <salmo...@salmon-tddft.jp>
件名: [salmon-user:00385] Questions about functional and spin calculations