scforbitalshift

[keyari79]

Dear all
I would like to use the $scforbitalshift keyword to increase the HOMO-LUMO gap in a scf calculation. I do not have any convergence problems, I just need a  large H-L gap for other purposes. But this keyword seems to be only usefull when it comes down to making the calculation converge!. What I need however is a way of manually increasing the gap for even harmless cases. Is there a way to do this in TURBOMOLE?

with regards
keyarash

[christof.haettig]

There is no automatic way to do this. All post-SCF (or DFT) programs in TM also need the correct orbital energies which belong to
the given structure and basis set. Why do you want to work with futched orbital energies?

Christof

[keyari79]

Hi
I want to perform structure optimization (QM/MM) for the excited CT state.  I am facing the well known problem of CT underestimation even if I employ the BHLYP functional (there are no point charge gradients for the CC2 modul available in TM, therefore I have to use TD-DFT for my QM/MM purposes!      ).
I am planning to perform single point CC2 on top of TD-DFT optimized structures. The problem of  sever underestimation of CT excitation energies during the geometry optimization however does not allow me go any further along the optimized TD-DFT path.  I thought, perhaps it is possible to widen the H-L gap (CT state is dominated by a HOMO-LUMO transition). It seems as if the only option (    ) is to use TD-HF or CIS for the geometry optimization.

thanks for your comment

[christof.haettig]

I see. I will fix the problem with the missing point charge gradients.
A work around for the version which don't have these gradients would be to incorporate the point charges as 'naked nuclei' without basis functions as individually assign charges in $coord and $atoms.

Best regards,
C.H.