CCL: Oscillator strength

[Ulf Ekström <uekstrom * gmail.com>]

 Sent to CCL by: =?ISO-8859-1?Q?Ulf_Ekström?= [uekstrom:gmail.com]
 On Wed, Mar 31, 2010 at 4:50 PM, Anton Nizovtsev nanto[A]mail.ru
 <owner-chemistry.%x%.ccl.net> wrote:
 >
 > Sent to CCL by: "Anton  Nizovtsev" [nanto-$-mail.ru]
 > Thank you for response!
 > I am interesting in singlet-singlet transition from A(S0)B(S0) to
 A*(T1)B*(T1). This case seems to be not trivial. The question is to calculate
 oscillator strength of such transition. Id like to know your opinion about
 applicability of linear response methods (namely, TDDFT, CCn etc.) and other
 possible methods to my problem.
 Since this is a double excitation I think you will have trouble with
 TDDFT. In principle it sounds like something that would be suitable
 for perturbation theory (i.e. compute the T1 states separately and
 couple them afterwards), but I don't immediately know of a program
 that can do this. Best bet seems to be a CI with a suitably defined
 state space, but your result may depend strongly on this space. I
 think Molcas can compute overlap between two configurations defined
 > from different sets of orbitals, which is what you get if you do a
 separate state optimization. Maybe linear response SOPPA (available in
 Dalton) can do it, but then you have to carefully identify the right
 state. I was trying to think of some approximation where you compute,
 say, A*(T1)B(S0) as an unrestricted ground state, and do tddft from
 there, but didn't come up with anything useful.
 Regards,
 Ulf Ekstrom, VU University Amsterdam