| 5:55:15 AM PDT - Thu, Aug 23rd 2018 |
|
NWChem currently doesn't have any facilities for doing Ziegler-Rauk (ZR) or Kitaura-Morokuma (KM) energy decomposition analysis (EDA) .
[see e.g. 1. https://core.ac.uk/download/pdf/132550427.pdf and/or 2. https://onlinelibrary.wiley.com/doi/epdf/10.1002/wcms.1345)
KM EDA can be done to some extent manually in NWChem, allowing for easy the calculation of E_int, E_orb, and E_prep through the generation of movecs corresponding to the different fragments.
However, I don't see how to compute E_elstat or E_Pauli (one would be enough to allow for the computation of the other). In particular the latter seems like something which shouldn't be too hard to implement in NWChem ('antisymmetrize and renormalise the product wavefunction'), but I'm not the person to do so (lacking theoretical and programming skills).
I think implementing some sort of EDA in NWChem would be valuable to the community since:
- NBO can currently only do its version, NEDA, in conjunction with GAMESS US. Generating an NBO file and using e.g. NBOPro won't do it.
- The codes that do offer it -- Turbomole, ADF (EDA-NOCV), molpro (SAPT) -- are not easily accessible to most. Note that Gaussian doesn't.
- AFAIK the only open code that does offer EDA is GAMESS US (CMOEDA, LMOEDA), and I have never had much luck working with e.g. polyanionic species in that code. It also appears to be much more demanding in terms of memory than other codes I've used.
Are there any plans/possibilities of implementing at least a basic form of KM or ZR EDA in NWChem? Or is that non-trivial?
/Andy
|
|
|