Molcas Forum

huangtf

Member

Registered: 2017-04-23

Posts: 18

Bond order analysis —— some questions

Hello everyone,

I have four questions about bond order analysis to consult you, see below, please!

1. On page 227, the manual says that the LOCNatural keyword gives natural occupation numbers and that it has proven useful for Effective Bond Order (EBO) analysis. In other word, I should do a single-point &localization using the LOCNatural keyword to localise the last .RasOrb.1 file previously outputted in my caspt2 optimization, and then, extract the occupation numbers of active orbitals from the .LocOrb file, and thirdly, EBO is calculated as bonding minus antibonding occupancies divided by two. Do you think so?

2. Can .Pt2Orb file be used to do EBO analysis? (These orbitals are the latest ones in my previous caspt2 optimization)

3. EBO analysis seems difficult to apply to polyatomic systems, eg., acrolein and 1,3-butadiene, alternatively, Mulliken bond order analysis might be a feasible way. Thus, in the case that the .molden file will not be produced (basis functions with angular momentum larger than g), is there a possibility to do Mulliken bond order analysis using .RasOrb.1 or .Pt2Orb generated by Molcas 8.0?

4. Mulliken population analysis is performed automatically in both &RASSCF and &CASPT2 by Molcas 8.0, however, for Mulliken bond order analysis, is there such a keyword or module?

Any insight will be appreciated!

Best regards,
huangtf

Last edited by huangtf (2017-07-26 14:45:21)