Cubic response calculations crashed for excited states of small molecules
Hello, I am using the MCSCF cubic response algorithm to calculate the intensity-dependent refractive index (IDRI) in Dalton. In particular, I am interested in that cubic response for excited states of CO2.
To benchmark my calculation, I first calculated the third-harmonic generation (THG)of CO2 at the ground state. Using CAS(12,13)/d-aug-cc-pVQZ, the THG I got is within the error bar in the experiment. However, when I chose the reference state to be one of the excited state (e.g. 1Pi_g) for the IDRI, the numerics blowed up. Only when I reduce the basis size to be cc-pVTZ, the program output some large numbers. But the same method gave a pretty far off THG for the ground state. I attached the relevant files here.
Is there anyone who understands why?
I tried molecules like methanol, and I experienced similar problems when I used the aug-cc-pVTZ basis set. I also tried to reduce the CAS space, but the calculations still blowed up. I have other files to share upon request.
I truly appreciate your comments/help.
Isaac
idri_cas13_d2h_1pig_co2_d2h_davqz.out
idri_cas13_d2h_1pig_co2_d2h_avqz.out
idri_cas13_d2h_1pig_co2_d2h_cvtz.out