You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
TS optimization is possible, and following an excited state is possible. The main difficulty is psi4 is lacking cheap excited state methods with analytic gradients. So no CIS, and TDDFT would use finite-difference gradients. If it's EOM-CC2 or EOM-CCSD (or a well-characterized by docc/socc SCF), Psi can help.
Hello
is it possible for excited state optimization in psi4 and search for a transition in the excited state?
The text was updated successfully, but these errors were encountered: