Alexander A. Voityuk Journal of Chemical Theory and Computation 2014, ASAP
Contributed by +Jan Jensen
This paper presents a significantly more accurate re-parameterization of the popular INDO/S method. The re-parameterization is based on the TBE-2 data set: vertical excitation energies and oscillator strengths computed for the valence excited states of 28 medium-sized organic molecules using multistate-CASPT2 and aug-cc-pVTZ basis set.
The author sums it up succinctly: "The mean absolute deviations of the INDO/X excitation energies relative to the TBE-2 data is 0.26 eV for singlet states and 0.33 eV for triplet states. The corresponding values for INDO/S are 0.56 and 0.64 eV. ... Unlike INDO/S where two different Hamiltonians are employed for singlet and triplet excitations, both types of excited states are calculated within the same INDO/X scheme." The MADs are thus similar to the MAD for TD-B3LYP calculations (0.27 eV) and significantly better than TD-BP86 (0.53 eV) and TD-BHLYP (0.48 eV) calculations.
Like INDO/S, INDO/X uses the CIS approach and can therefore be applied to quite large molecules. Furthermore, the author notes that the results are quite insensitive to the chosen active space.
Thanks to @CompChemNews for alerting me to this paper.
This work is licensed under a Creative Commons Attribution 4.0 International License.
Contributed by +Jan Jensen
This paper presents a significantly more accurate re-parameterization of the popular INDO/S method. The re-parameterization is based on the TBE-2 data set: vertical excitation energies and oscillator strengths computed for the valence excited states of 28 medium-sized organic molecules using multistate-CASPT2 and aug-cc-pVTZ basis set.
The author sums it up succinctly: "The mean absolute deviations of the INDO/X excitation energies relative to the TBE-2 data is 0.26 eV for singlet states and 0.33 eV for triplet states. The corresponding values for INDO/S are 0.56 and 0.64 eV. ... Unlike INDO/S where two different Hamiltonians are employed for singlet and triplet excitations, both types of excited states are calculated within the same INDO/X scheme." The MADs are thus similar to the MAD for TD-B3LYP calculations (0.27 eV) and significantly better than TD-BP86 (0.53 eV) and TD-BHLYP (0.48 eV) calculations.
Like INDO/S, INDO/X uses the CIS approach and can therefore be applied to quite large molecules. Furthermore, the author notes that the results are quite insensitive to the chosen active space.
Thanks to @CompChemNews for alerting me to this paper.
This work is licensed under a Creative Commons Attribution 4.0 International License.
No comments:
Post a Comment