RHF mos as UHF guess

[acapobia]

Hi all,

I wonder if it is possible (TM 6.3) to use converged RHF MOs as a guess for UHF. Suppose you have a set of converged MOs for a neutral closed shell molecule.  Supposing at  first that no symmetry changes are involved, is it possible to use the above mos as a guess for a DSCF (DFT) computation of the same molecule in its singly ionized doublet state obtained by removing or adding an electron?

Sorry if this is a duplicate post, but I wasn't able to find anything relevant.

[Arnim]

Hi!

Yes, that is indeed possible. Here is a short HOWTO.

First, create alpha and beta like this:
cp mos alpha
cp mos beta
In alpha change $scfmo to $uhfmo_alpha .
In beta $scfmo to $uhfmo_beta .

In control change
$scfmo   file=mos
$closed shells
 a       1-21                                   ( 2 )
to
$uhfmo_alpha   file=alpha
$uhfmo_beta   file=beta
$uhf
$alpha shells
 a       1-21                                   ( 1 )
$beta shells
 a       1-21                                   ( 1 )

Now, you have just created an UHF contrl from an RHF control.
Restarting dscf should give the same energy as before.
To charge the system just put electrons in the specific shell, e.g.:
$alpha shells
 a       1-22                                   ( 1 )

Cheers,

Arnim

[acapobia]

It worked fine, thanks.
Now a much more sophisticated related question. Suppose that something like Jahn Teller effects are involved in going from the neutral to the ionic species, so that a reduced symmetry is obtained by ionizing the system;  suppose, further, you know the correct occupation scheme for the new reduced symmetry. Is there a way to recover the original RHF guess, for the UHF computation ?

[Arnim]

I am not sure, if I understand what you want to do. But the 'use' option in define can transform mos from one symmetry to another.