Ensure dispersion correction is added to scfenergies for all packages

This is to help us keep track. It may not be relevant for all packages, but it should be investigated for each. Because this is relatively critical, the first pass should make sure scfenergies includes all relevant energies; for example, a 3-body energy may be optionally calculated, and might not be present in the final dispersion energy.

• ADF (???)
• DALTON (???)
• Firefly (???)
• GAMESS-US (???) ~~ * [ ] GAMESS-UK (???)~~
• Gaussian (???) ~~ * [ ] Jaguar (???)~~
• Molcas (???)
• Molpro (???)
• MOPAC (???)
• NWChem (???)
• ORCA (issue: https://github.com/cclib/cclib/issues/660, PR: https://github.com/cclib/cclib/pull/418)
• Psi3 (???)
• Psi4 (???)
• Q-Chem (???)
• Turbomole (???)

[berquist]

I was thinking that this was more about emergency checks for the dispersion energy being included in scfenergies, because is isn’t in ORCA for some reason.

[ghutchis]

No, I just tried a bunch of methods and they don’t give anything explicit. I realized that cclib-data doesn’t have many MOPAC files. If you want to create an issue to generate more examples, I’d be happy to do that (frequency, CI, etc.)

It might be useful for some people to separate out the calculated heat-of-formation from the actual SCF total energy. We’ve found the latter to be more useful for machine learning, for example.