Tributyltin determination in natural waters using second order multivariate calibration and fluorescence spectroscopy

BRAVO, M.; OLIVIERI, A. C.; ESCANDAR, G. M.; AGUILAR, F.; QUIROZ, W.

Buzios

Congreso; XXXVII Colloquium Spectroscopicum Internationale; 2011

CSI

Due to its widespread use as an antifouling agent in boat paints, tributyltin (TBT) is a common contaminant of marine and freshwater ecosystems. Exposure to water and sediments contaminated with TBT induces its accumulation on marine biota, and leads to biological effects on aquatic organism. Besides, potential harmful effects on human health may also result from the consumption of contaminated seafood. Recently, the European Union has classified to TBT as priority pollutant (Water Framework Directive 2000/60/EC, Pollutant Emission Register 2000/479/EC). Unfortunately, present and future restrictions will not immediately remove TBT from aquatic environments.For control of TBT levels in aquatic ecosystems, several analytical methodologies have been proposed. In general, they are based on a gas chromatographic separation, followed by selective detection. However, most of these analytical methodologies require a succession of additional discrete steps (extraction, derivatization, preconcentration and clean-up) which may all be prone to errors and demand high skilled analysts, time and money which severely hamper that these methodologies can be implemented for routine analysis.In this work, a rapid and reliable analytical method is proposed for quantitation of tributyltin in natural waters (seawater, tap and river water samples) using excitation-emission fluorescence matrices (EEM) and second-order multivariate calibration. The principle of detection of TBT is based on the reaction between tributyltin and 3,5,7,2?,4?-pentahydroxyflavone (morin) in a triton X-100 micellar medium, which yields a fluorescent derivative. Several multivariate algorithms were evaluated for processing matrix fluorescence data in order to achieve the second advantage in presence of potential interferents and real water matrix. Finally, analyte quantitation was possible in spiked water samples containing uncalibrated spectral interferences after a preconcentration step on a C-18 membrane, reaching a detection limit in ng L?1 level, demonstrating the analytical potential of the proposed methodology for analysis of contaminated water samples.