Predicting accurate fluorescent spectra for high molecular weight polycyclic aromatic hydrocarbons using density functional theory

The ability of density functional theory (DFT) methods to predict accurate fluorescence spectra for poly cyclic aromatic hydrocarbons (PAHs) is explored. Two methods, PBEO and CAM-B3LYP, are evaluated both in the gas phase and in solution. Spectra for several of the most toxic PAHs are predicted and compared to experiment, including three isomers of C24H14 and a PAH containing heteroatoms. Unusually high resolution experimental spectra are obtained for comparison by analyzing each PAH at 4.2 K in an n-alkane matrix. All theoretical spectra visually conform to the profiles of the experimental data but are systematically offset by a small amount. Specifically, when solvent is included the PBEO functional overestimates peaks by 16.1 +/- 6.6 nm while CAM-B3LYP underestimates the same transitions by 14.5 +/- 7.6 nm. These calculated spectra can be empirically corrected to decrease the uncertainties to 6.5 +/- 5.1 and 5.7 +/- 5.1 nm for the PBEO and CAM-B3LYP methods, respectively. A comparison of computed spectra in the gas phase indicates that the inclusion of n-octane shifts peaks by +11 nm on average and this change is roughly equivalent for PBEO and CAM-B3LYP. An automated approach for comparing spectra is also described that minimizes residuals between a given theoretical spectrum and all available experimental spectra. This approach identifies the correct spectrum in all cases and excludes approximately 80% of the incorrect spectra, demonstrating that an automated search of theoretical libraries of spectra may eventually become feasible. (C) 2016 Elsevier Inc. All rights reserved.