Author Topic: Excited state optimizations: General questions  (Read 7168 times)

tom

  • Newbie
  • *
  • Posts: 4
  • Karma: +0/-0
Excited state optimizations: General questions
« on: October 22, 2008, 12:04:44 PM »
Hello!
I am trying to optimize the geometry of the excited state of a molecule with Turbomole 5.7.1.
Now, some practical questions arose:

1.) I read in another thread that the ground state geometry has to be described accurately. Does this mean, I have to necessarily use the same functionals, basis sets etc. for ground state and excited state optimizations? Or does it simply mean that the ground state geometry has to be meaningful in terms of correlating with a real-life geometry?

2.) What about symmetry? My ground state is of c3v symmetry, but since I don't exactly know what the excited state will look like, should I rather not use symmetry?

3.) What exactly do the "number of roots" mean to the geometry of the excited state that will result? In the ground state, the HOMO and the LUMO are non-degenerate and belong to the same irrep, the excitation is symmetry forbidden. The (HOMO-1), however, is twofold degenerate and (HOMO-1) -> LUMO transitions have appreciable oscillator strengths. This is what I also observe in the UV absorption spectrum: a band of very little intensity at the wavelength calculated (TDDFT) for the HOMO -> LUMO transition, and a much stronger band located at the wavelength calculated for the (HOMO-1) -> LUMO transition.
Will the geometry, in theory, be any different if I select, say, three states instead of one?

4.) If I look at a plot of the frontier orbitals, can I qualitatively say something like: the HOMO consists mainly of bonding X-X orbitals, whereas the LUMO consists of antibonding X-Y orbitals, so the excited state geometry will probably contain elongated X-X bonds (missing electron density in the bonding orbital) and elongated X-Y bonds (populated antibonding orbitals)? If so, should I modify the starting geometry accordingly in this direction, or should I start from the ground state geometry?

5.) Are there some general guidelines when to use RPA, and when to use TDA? What advantages does each one have over the other in describing real-life problems and computational performance? 

6.) Will I ever actually see a populated (formerly) virtual orbital from the occupation numbers in the job.last file in an excited state geometry optimization using TDDFT or CIS? Because the occupation numbers seem to be always the same as they are in the ground state.

7.) Can I, with any excited state optimization method in Turbomole, select to optimize a certain occupation, say
          a   b 
(HOMO-1): 0   1
HOMO:     1   1
LUMO:     1   0

just to see what the geometry would look like?
« Last Edit: October 23, 2008, 03:36:37 PM by tom »

christof.haettig

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 291
  • Karma: +0/-0
    • Hattig's Group at the RUB
Re: Excited state optimizations: General questions
« Reply #1 on: January 11, 2009, 12:58:30 PM »
Well, to find the answers to most of your question you should consult some standard quantum chemistry textbooks. Without understanding the quantum chemical background, it will be very difficult to do meaningfull calculations for excited states.

Christof