Molcas Forum

turnsek

Member

Registered: 2023-12-14

Posts: 1

A possible bug in RASSI resulting in divergent osc. strengths?

Hi all!

I have been trying to calculate a simple PEC for the 3p excited state of helium. I can get converged orbitals with great energy accuracy using SA-RASSCF + RMS-CASPT2.
This is where the niceness ends though. When trying to calculate the oscillator strengths between the ground state and the excited states of sym. gr. 5 (in the D2h group) I get a ton of warnings of type:

 ###############################################################################
 ###############################################################################
 ###                                                                                                                                                  ###
 ###                                                                                                                                                  ###
 ###    WARNING: A possible bug was detected.                                                                            ###
 ###                                                                                                                                                  ###
 ###                                                                                                                                                  ###                                  
 ###############################################################################
 ###############################################################################
  WARNING: Non-zero matrix element computed
  which should be zero by spin symmetry!
               Spins S1, S2:   0.0000000000000000        0.0000000000000000     
  Spin projections SM1, SM2:   0.0000000000000000        0.0000000000000000     
     Operator has S=1.0, SM:   0.0000000000000000     
  Clebsch-Gordan:   0.0000000000000000     
  Size is TMATEL=   3.7325662560760975E-009

and the resulting osc. strengths are

++ Dipole transition strengths (spin-free states):
   -----------------------------------------------
     for osc. strength at least  1.00000000E-05

      From   To        Osc. strength     Einstein coefficients Ax, Ay, Az (sec-1)    Total A (sec-1)
     -----------------------------------------------------------------------------------------------
         1    2       5.20432021E+26  0.00000000E+00  0.00000000E+00  9.43286204E+36  9.43286204E+36
         1    3       1.67828202E+27  0.00000000E+00  0.00000000E+00  3.22359225E+37  3.22359225E+37
         1    4       1.35162735E+25  0.00000000E+00  0.00000000E+00  3.03314392E+35  3.03314392E+35
         1    5       8.65194149E+35  0.00000000E+00  0.00000000E+00  1.96774217E+46  1.96774217E+46
         1    6       1.04071829E+26  0.00000000E+00  0.00000000E+00  2.36695398E+36  2.36695398E+36
         1    7       2.41784030E+27  0.00000000E+00  0.00000000E+00  5.50554665E+37  5.50554665E+37
         1    8       2.81973328E+25  0.00000000E+00  0.00000000E+00  6.75361370E+35  6.75361370E+35
         1    9       1.39341170E+39  0.00000000E+00  0.00000000E+00  3.35527899E+49  3.35527899E+49
     -----------------------------------------------------------------------------------------------

(expected result would be right around 0.05.

The energies of the states and orbital population analyses are all as expected. I am using a large basis set [29,25,24,23](15,18,15,14) which I contracted just for this purpose and made sure to include the desired orbitals in the genano. The basis set exponents were taken from [10.1063/1.4772468].

Now for the worst part. The oscillator strengths calculated in this way for the states of sym. gr. 2 - the Pi states - are in quantitative agreement with the experiments.

I would very much appreciate it if anyone had any idea what might be the main culprit here. Thank you for your help!