Chemometric Analysis of Multidimensional Fluorescence Data Recorded from Benzo[a]pyrene Metabolites in Shpol\u27skii Like Matrixes

This dissertation extends the application of Shpol\u27skii spectroscopy to polar metabolites in polar organic solvents. For the first-time, we reported experimental evidence on the line-narrowing effect caused by primary alcohols on the spectral features of PAH metabolites at 77 K and 4.2 K. The effect of primary alcohols (RCH2OH) on the 77 K and 4.2 K fluorescence characteristics of PAH metabolites showed the best spectral narrowing when matching the length of the alkyl group (R) of the primary alcohol to the length of the n-alkane that best fitted the dimensions of the parent PAH. The spectral narrowing and the fluorescence enhancements observed for 1-hydroxypyrene, 2-hydroxyfluorene, 9-hydroxyphenanthrene, 3-hydroxybenzo[a]pyrene , 4-hydroxybenzo[a]pyrene , 5-hydroxybenzo[a]pyrene , B[a]P-trans-7,8-dihydrodiol (±), B(a)P-trans-9,10-dihydrodiol (±), B[a]P-r-7,t-8-dihydrodiol-c-9,10-epoxide(±) (syn-BPDE), and B[a]P-r-7,t-8-dihydrodiol-t-9,10-epoxide(±) (anti-BPDE) demonstrated the potential for the qualitative and quantitative analysis of PAH metabolites at trace concentration levels. The obtained enhancements for B[a]P metabolites provided limits of detection varying from 0.1 ng.mL-1 (anti-BPDE (±)) to 0.8ng.mL-1 . Since B[a]P is the most toxic PAH in EPA list, four B[a]P metabolites were selected for chemometric studies. These included 1-Hydroxybenzo[a]pyrene, Benzo[a]pyrene-cis-4,5-dihydrodiol, Benzo[a]pyrene-r-7, t-8, t-9, c-10-tetrahydrotetrol (+/-) and Benzo[a]pyrene-r-7, t-8, t-9, t-10-tetrahydrotetrol (+/-). In all cases, 1-octanol was used as the Shpol\u27skii matrix. MCR-ALS, UPLS/RBL, and PARAFAC were tested for multi-way calibration purposes. These algorithms carry with them the higher order advantage, which refers to their ability to perform a mathematical separation, identification, and quantitation of analytes in complex samples with chemically unknown composition. Three-way data formats consisted of 77 K EEMs recorded with the spectrofluorimeter. Four-way data formats were 4.2 K TREEMs recorded with the instrumental system built in-house. In both cases, the best prediction abilities were obtained with PARAFAC. This algorithm was able to provide accurate determinations at the parts-per-billion concentration levels