I want to calculate the XANES spectrum of benzene at the Carbon K-edge using the CVS-ADC(2)-x method. As the six Carbon atoms are equivalent I want to use effective core potentials for five of the carbon atoms and then calculate transitions starting from only one of the C(1s) orbitals but I don’t know how to properly set my input parameters. Do I set CC_REST_OCC to 1 as there is only 1 Carbon atom left? And how can I tell the program that I want ECPs for only 5 (and not all) Carbon atoms?
What you want is something like ECP = MIXED (analogous to BASIS = MIXED that allows different basis sets for different instances of the same element). Unfortunately it looks like Q-Chem does not support that at present. ECP = GEN supports different types of ECPs (i.e., mixing and matching named pseudopotentials) but ECP = MIXED is not recognized, and that’s consistent with my recollection that this feature is not implemented.
A workaround for what you want to do would be to localize the MOs (e.g., Boys MOs will usually put the 1s on a single atom, even for symmetry-equivalent atoms), then use some kind of CVS-type (frozen orbital) approximation to exclude the C(1s) MOs that you don’t want to be active.
We also encountered similar problems. We need to set different ECPs for the same element, ECP = MIXED(similar to BASIS = MIXED). Whether to consider adding this feature in the next release?
