PC GAMESS-related discussion club

Alex Granovsky
gran@classic.chem.msu.su

this message is usually related with triplet instability of the reference wavefunction. In this case, due to the nature of TDDFT equations, the transition energies become purely imaginary. However, in some rare cases this can be simply an artefact of numerical scheme used to solve TDDFT equations. As your system is small, try to increase the number of initial guess vectors or to use incore diagonalization - this should help in the latter case.

In the case of TDA/TDDFT, the equations are different and always result in real energies. In the worst case you'll simply get a negative excitation energy.

Best regards,
Alex


On Tue May 8 '07, Pedro Silva wrote
-----------------------------------
>Dear friends,

>When performing a TDDFT calculation on O2 (with 6 core orbitals) I get the following error:

> UNIT VECTOR GUESS AT TDDFT COEFICIENTS ...
> Error: non positive-defined subspace A-B matrix!

>The error does not occur if I select TDA=.true.

>Why does this happen?

>Pedro S.
>

[ This message was edited on Fri May 11 '07 by the author ]