Excited states

MOPAC can calculate electronic excitations using both its native thermochemistry (PMx) models and also Zerner’s INDO/S (ZINDO) model. Each option has a different set of advantages and limitations. The PMx models provide a baseline potential energy surface from which to study excited-state potential energy surfaces. However, only atomic excitation energies are in the PMx training data, so their accuracy for delocalized electronic excitations relies on model transferability. The INDO/S model is trained primary on electronic excitation data of molecules, but it can only predict vertical excitation energies and does not provide access to a complete excited-state potential energy surface.

All of MOPAC’s excited state functionality is based on the configuration interaction (CI) method, with a variety of options for specifying the set of active configurations in the CI calculation. See the online MOPAC manual for a complete description of these CI options.

For a PMx calculation, the simplest way to perform an excited-state calculation is to add the keywords MECI C.I.=x ROOT=y to a MOPAC input file. The MECI keyword prints information about the CI calculation in the .out output file, C.I.=x requests a CI active space constructed from the x/2 (rounded up) highest-energy occupied orbitals and the x/2 (rounded down) lowest-energy virtual orbitals, and ROOT=y uses the yth lowest-energy state of the CI calculation to define the excited-state potential energy surface.

The simplest way to perform an INDO/S calculation is to add the keywords INDO C.I.=x to a MOPAC input file. The INDO keyword prints information about the INDO/S calculation in the .out output file, and C.I.=x requests a set of active configurations for CI by exciting the lowest energy configuration with single-electron excitations from the x/2 (rounded down) highest-energy occupied orbitals to the x/2 (rounded up) lowest-energy virtual orbitals. Note that the C.I. keyword does not have consistent behavior between the INDO and MECI features.