Why do TD-DFT excitation energies of BODIPY/aza-BODIPY families largely deviate from experiment? Answers from electron correlated and multireference methods

Mohammad R. Momeni, Alex Brown

Research output: Contribution to journalArticle

84 Scopus citations

Abstract

The vertical excitation energies of 17 boron-dipyrromethene (BODIPY) core structures with a variety of substituents and ring sizes are benchmarked using time-dependent density functional theory (TD-DFT) with nine different functionals combined with the cc-pVTZ basis set. When compared to experimental measurements, all functionals provide mean absolute errors (mean AEs) greater than 0.3 eV, larger than the 0.1-0.3 eV differences typically expected from TD-DFT. Due to the high linear correlation of TD-DFT results with experiment, most functionals can be used to predict excitation energies if corrected empirically. Using the CAM-B3LYP functional, 0-0 transition energies are determined, and while the absolute difference is improved (mean AE = 0.478 eV compared to 0.579 eV), the correlation diminishes substantially (R2 = 0.961 to 0.862). Two very recently introduced charge transfer (CT) indices, qCT and dCT, and electron density difference (EDD) plots demonstrate that CT does not play a significant role for most of the BODIPYs examined and, thus, cannot be the source of error in TD-DFT. To assess TD-DFT methods, vertical excitation energies are determined utilizing TD-HF, configuration interaction CIS and CIS(D), equation of motion EOM-CCSD, SAC-CI, and Laplace-transform based local coupled-cluster singles and approximate doubles LCC2∗ methods. Moreover, multireference CASSCF and CASPT2 vertical excitation energies were also obtained for all species (except CASPT2 was not feasible for the four largest systems). The SAC-CI/cc-pVDZ, LCC2∗/cc-pVDZ, and CASPT2/cc-pVDZ approaches are shown to have the smallest mean AEs of 0.154, 0.109, and 0.100 eV, respectively; the utility of the LCC2∗ approach is demonstrated for eight extended BODIPYs and aza-BODIPYs. We found that the problems with TD-DFT arise from difficulties in dealing with the differential electron correlation (as assessed by comparing CCS, CC2, LR-CCSD, CCSDR(T), and CCSDR(3) vertical excitation energies for five compounds) and from contributions of multireference character and double excitations (from analysis of the CASSCF wave functions).

Original languageEnglish (US)
Pages (from-to)2619-2632
Number of pages14
JournalJournal of Chemical Theory and Computation
Volume11
Issue number6
DOIs
StatePublished - Jun 9 2015

Fingerprint Dive into the research topics of 'Why do TD-DFT excitation energies of BODIPY/aza-BODIPY families largely deviate from experiment? Answers from electron correlated and multireference methods'. Together they form a unique fingerprint.

  • Cite this