The reason that we use the stability analysis (on singlets) is to make sure we are actually modeling the ground state of the real system in question, but this also seems like it could be a little bit counter-intuitive because the resulting wavefunction is then spin-contaminated. This is supposed to happen, right? If so, what is a good value / rule-of-thumb estimate to expect after a stability reoptimization of the wavefunction?
Also, this next question is not quite as important, but still worth asking. What does it mean when the spin-projected wavefunction has more spin-contamination than the original? I have only seen this happen once in a singlet calculation of ammonia-coordinated iron(II), but it definitely happened. This is approximately what happened: = ~0.8 -> 1.1 after annihilation, for reference.
Thank you for your input, good sir. I have read through the papers but I still am not 100% clear of what the spin contamination means in this calculation, especially if it is not a nice multiple of 1/2. Am I just over-thinking it?
I think I understand the concept of a singlet diradical, but not something like a =1.2 species (0.7 < S < 0.71). Is that because it is actually bogus, or might something like that exist? This is just a hypothetical example so it is surely possible that it doesn't exist.
To all of you that follow this question, I think I have some more insight based on empirical evidence.
It appears that, if there is spin contamination in the wavefunction, it prefers to adopt a different electronic state than the one you have chosen.
For example, a doublet with =10 probably prefers to be a sextet (ideal =8.75). From what I can tell, it is that easy and I was just thinking too much.
As another example, a stabilized singlet with =1.8 is likely to be more stable when the radical electrons are unpaired, thus a triplet optimization should give better results. This may not always be the case but it seems to be a good indication.
I still don't know why a spin-annihilated wavefunction could have more spin than the original, but I haven't done any real math so maybe I'll let you know if I ever end up doing that...
And thank you, Dr. Gadzheiv, for leading me closer to this conclusion. If you think my analysis is incorrect then please let me know. It's too bad that I haven't been publishing much recently...
Hello,
How to estimate spin contamination particularly triplet spin contamination from the TD files.
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory
- Control Systems
- Control Systems Engineering