DFT fails to create diabatic states

[Oleksii]

Hi everyone, 

I am testing the NWChem electron transfer module, and have run into some difficulties generating diabatic states. To investigate, I have used the minimal example from the manual (He(0)-He(+) electron transfer), and tested it with both Hartree-Fock and DFT.

I ran calculations with several distances, ranging from 5 to 15 Å, but in all cases DFT produced unexpected results: the overlap between the two diabatic states was close to unity, leading to unrealistically large electron coupling. I further analysed the spin densities in both states and noticed that, although the individual guesses were correct, the spin density in one of the diabatic states was delocalised between the donor and acceptor (see the attached picture). Interestingly, these states were generated correctly in HF.

Could you please provide any hints on resolving this issue in DFT? It sounds unlikely that the tiny exchange-correlation term can affect the spin density distribution in such a simple system, especially at longer distances.

I have uploaded the input/output files for this minimal example here (distance 7 Å between the donor and acceptor):

https://drive.google.com/file/d/1Q-QxHdAD9QDyP8JDDuUR8InXUnlDAzIu/view?usp=sharing

In order to make the setups comparable, I used the "dft" engine for both, but using "xc hfexch" for the Hartree-Fock case.

I would very much appreciate your help.

Best regards,

Oleksii