Question about spin-orbit coupling

[behnam nikoobakht]

Hello,

I have a question regarding the evaluation of spin-orbit coupling (SOC) matrix elements between the singlet and triplet states in a neutral system.  I have computed SOC matrix elements using MRCI wavefunction along distorted molecular geometry. When I plot SOC matrix elements as a function of bond length, I can see a sudden job at some of the geometries. My question is that is there any way to get rid of these sudden jumps for the evaluation of SOC matrix elements?

Thank you for your time,

Behnam

[iri...@verizon.net]

Hi Behnam,

It sounds like you are getting different states (before SOC step) at different geometries. Especially when dissociating, many states become close in energy and it can be tricky to maintain a consistent set across different geometries.  If you're using state-averaged MCSCF orbitals, maybe dynamical weighting would smooth some of the bumps. 

Sudden changes may be physically correct, if they result from a curve crossing (adiabatic state changing character from one diabatic state to another). Of course, that would affect how you interpret the matrix elements. 

Good luck,

Karl