# How to interpret Gaussian output for excited states?

I have calculated the excited energies for benzene in Gaussian. The output therefore was:

 Excited states from <AA,BB:AA,BB> singles matrix:
Excited State   1:      Singlet-A      5.5804 eV  222.18 nm  f=0.0000  S**2=0.000
20 -> 22         0.41871
20 -> 23         0.27308
21 -> 22         0.27309
21 -> 23        -0.41870
Excited State   2:      Singlet-A      6.4315 eV  192.78 nm  f=0.0000  S**2=0.000
20 -> 22        -0.27213
20 -> 23         0.41722
21 -> 22         0.41725
21 -> 23         0.27213
Excited State   3:      Singlet-A      7.9411 eV  156.13 nm  f=0.0000  S**2=0.000
18 -> 23         0.49564
19 -> 22         0.49673


Now I am wondering from where to where are these transitions. I know that the first one is from HOMO to LUMO. Thinking of QM I'd assume that the second transition is from HOMO-2 to HOMO+1 and the third one would be from HOMO-1 to HOMO+2. AM I thinking correct?

• Saying that electronic excitations are the result of transitions from one orbital to another is kind of the 0th order approximation and quite primitive. Sadly, things are not that simple. Could you perhaps explain how you got to your statements? What do you wish to conclude from them? – Raditz_35 Aug 14 '18 at 12:32
• @Raditz_35 Firstly, I optimized the molecules ground state structure using DFT/B3-LYP/6-31G with only respecting singlets. After that I used TD-SCF for DFT with B3-LYP and 6-31G and singlets only again to obtain my excitation energies. I want to calculate excitation energies for poly(p-phenylene) with n from 1 to 8 where n is the number of monomers. Then I want to compare the excitation energies obtained from Gaussian with the ones I calculated using particle in a box model. That's why I thought of transitions from one orbital to another. – p_punkt Aug 14 '18 at 12:56
• The output is fairly obvious: The first excited state has excitations from MO 20 to 22, 20 to 23, 21 to 22, and a de-excitation from 23 to 21. So the first one is certainly not from HOMO to LUMO.// Please edit your question to include more details, especially the input files you have used to generate this. Also specify the version of Gaussian you are using. – Martin - マーチン Aug 14 '18 at 15:41