Concerning on your wondered that, “why I persist in using LSDA for calculating the optical properties of all molecules?” which had shown in previous report. I had chosen LSDA approximation because considering only the electronic results, when comparison is nearest with experimental result but incorrect for predicting the properties. From learn more about difference of exchange-correlation, it found that local spin density and generalized gradient approximations cannot describe bound excitonic states due to lack of an effective attractive Coulomb interaction between photo-excited electron and holes, which are not accurate enough for describing quasi-particles (such as excitons, phonons etc.) .In contrast, hybrid density functional, include long-rang non-local and non-adiabatic corrections in from of a fractions of exact exchange (Hartree-Fock), are able to reproduce the excitonic effects.
Therefore, authors have removed the optical absorption results of the TD-LSDA and instead by TD-B3LYP results in order to according with the ionization potential (IP) and electron affinity (EA) results have been obtained from the same method.