Chemometric Deconvolution of Continuous Electrokinetic Injection Micellar Electrokinetic Chromatography Data for the Quantitation of Trinitrotoluene in Mixtures of Other Nitroaromatic Compounds.

摘要

The chemometric deconvolution of spectral data from continuous electrokinetic injection of mixtures of nitroaromatics for the quantitation of trinitrotoluene (TNT) is presented. Partial least squares regression (PLS) models were constructed for TNT, and the ability of these models to correctly predict TNT concentration in the presence of other nitroaromatic interferents was explored. Spectral similarities between TNT and other nitroaromatic compounds are shown to result in incorrect identification and quantitation of those compounds as TNT when interferent data is not incorporated into the PLS models, highlighting an important limitation of the technique. Using properly calibrated PLS models, TNT concentration is correctly identified and quantified in a solution containing 2,4-DNT, 2,6-DNT, and 4-Amino-2,6-DNT when all three are treated as either interferents, or alternatively as analytes of interest (whether they are also identified and quantified).