Thanks for the replies!
@hokru, I aim to show the effect of mixing of the ground state wavefunction with the doubly excited wavefunction. At the HF level, both the bonding and doubly excited wavefunctions have equal probabilities for both electrons to be on each nucleus. But, at the wavefunction amplitude level, they have different signs. Adding the two wavefunctions then reduces the joint probability density that both electrons are on the same nucleus. So, at the HF level showing the wavefunction would be optimal.
At the CI level showing the joint probability density would still convey the essential insight. But this would not be the usual density! I am trying to show electron correlation, which depends on all the electron coordinates. I want to plot the amplitude (or probability) that electron two can be at position r2 given that electron one is at r1. That information is contained in Psi(r1,r2).
@jmisiewicz, thanks for the insights. As mentioned above, for the CI calculation yes the density might work – if I can get to the multi electron density | Psi(r1,r2) |^2 .
The output file says DETCI (I just went with the default). I don’t know if FNOCC is an option – I want to show that CI gets the convergence of the singlet and triplet energies at large rAB. There are some words in the documentation that say FNOCC is implemented for closed shell references, that would be a problem for the triplet state, is that right?
Thank you both for your time!