Question: can present numerical atomic orbital always reproduce band gap of PW calculation in theoretical aspect?

kirk0830 commented 8 months ago

Details

Recently I realize the released version of numerical atomic orbital for ABACUS lcao calculation only use occupied bands as reference, by either explicitly specifying number of bands to refer, or set to auto. However, I am concerned about whether this generation strategy is correct theoretically. If not always true, an alternative way to define the keyword auto for determing number of bands to refer might be 50-50, say set number of bands to refer to double of the number of occupied bands, I guess.

Have you read FAQ on the online manual http://abacus.deepmodeling.com/en/latest/community/faq.html

[X] Yes, I have read the FAQ part on online manual.

Task list for Issue attackers (only for developers)

[x] Understand the problem or question described by the user.
[x] Check if the issue is a known problem or has been addressed in the documentation.
[ ] Test the issue or problem on a similar system or environment, if possible.
[x] Identify the root cause or provide clarification on the user's question.
[x] Provide a step-by-step guide, including any necessary resources, to resolve the issue or answer the question.
[ ] If the issue is related to documentation, update the documentation to prevent future confusion (optional).
[ ] If the issue is related to code, consider implementing a fix or improvement (optional).
[ ] Review and incorporate any relevant feedback from users or developers.
[ ] Ensure the user's issue is resolved or their question is answered and close the ticket.

kirk0830 commented 8 months ago

no matter what answer will be here, I plan to add band structure reproduction test in ABACUS-Pseudopot-Nao-Square

kirk0830 commented 7 months ago

Seems this question has no certain answer, just empirically the band gap estimation seems not go wrong. However, it is better to generate orbitals including empty states.

hongriTianqi commented 6 months ago

There is some clue that including unoccupied bands in orbital generation might increase convergence and correctness of systems with excited magnetic moments. LiCoO test using orbitals without unoccupied bands: https://app.bohrium.dp.tech/abacustest/?request=GET%3A%2Fapplications%2Fabacustest%2Fjobs%2Fjob-abacustest-v0.3.113-7142e5

This result shows 1 Co4+ with large spin, 206 scf steps to converge.

LiCoO test using orbitals including 3 unoccupied bands in dimer and 11 unoccupied bands in trimer. https://app.bohrium.dp.tech/abacustest/?request=GET%3A%2Fapplications%2Fabacustest%2Fjobs%2Fjob-abacustest-v0.3.113-59ee47

This result shows 3 Co4+ with large spin, 137 scf steps to converge.

deepmodeling / abacus-develop