We do not currently have an active space selection workflow, and don't yet support active space calculations. Active space selection is currently a very difficult problem, and there have been a few suggestions:
If you want to do it manually, visualizing them is the best solution. π orbitals with extended conjugation and any orbitals that are stretched are typically the most important. One should select a balanced active space, with complementary sets of bonding and antibonding orbitals. For benzene, the frontier orbitals include π₂, π₃, π₄, and π₅, but π₁ is the HOMO–4 and π₆ is the LUMO+7 (https://labs.rowansci.com/s/benzene_orbitals), something that would be missed if merely selecting based on energies.
Hey guys - do you have workflow solutions for easily selecting active spaces?
We do not currently have an active space selection workflow, and don't yet support active space calculations. Active space selection is currently a very difficult problem, and there have been a few suggestions:
- UHF Natural Orbitals (https://doi.org/10.1063/1.4922352)
- MP2 Natural Orbitals (https://doi.org/10.1063/1.453884)
- DMRG (https://arxiv.org/pdf/1602.03835)
- ML (https://pubs.acs.org/doi/10.1021/acs.jctc.9b01297)
If you want to do it manually, visualizing them is the best solution. π orbitals with extended conjugation and any orbitals that are stretched are typically the most important. One should select a balanced active space, with complementary sets of bonding and antibonding orbitals. For benzene, the frontier orbitals include π₂, π₃, π₄, and π₅, but π₁ is the HOMO–4 and π₆ is the LUMO+7 (https://labs.rowansci.com/s/benzene_orbitals), something that would be missed if merely selecting based on energies.