ANALYTICAL QUALITY PARAMETERS OF THE SOLI D-PHASE EXTRACTION METHOD FOR THE DETERMINATION OF PESTICIDE-RESIDUES IN WATER BY GAS-CHROMATOGRAPHY

A study was done to establish the linearity, analytical sensitivity, p recision, and detection limit of 29 pesticides intensively used in Chi le and of well-known persistence in the environment by the statistical model of linear regression. To this end, extractions were carried out in waters containing 12,5 to 300 ng/L for pesticides determined by GC -ECD, and 100 to 750 ng/L for pesticides determined by GC-NPD. Only tr iadimefon, heptachlor, and aldrin showed a lack of linearity in extrac tion due to low precision at the lower concentration levels and, possi bly, to interference caused by the extraction columns, undetected in b lank measurements. The slope values in the regression equation Y = a bX (where Y was the amount of recovered analyte, and X the amount add ed) ranged from 0,818 to 1,041 except for aldrin (0,561), dieldrin (0, 786), and triadimefon (1,254),which accounted, in the first two, for l osses produced in the extraction process. Systematic errors were detec ted for most of the compounds determined for GC-ECD, where the smalles t errors appeared in estimating the regression equation. In the case o f compounds determined for GC-NPD, systematic errors were detected onl y for dichlorvos, ethoprophos, simazine, napropamide, and triadimefon. The detection limits established through the model are higher than th ose determined for the instrumental method; in most of the cases they are lower than the maximum allowed limit for waters by the European Un ion (100 ng/L), ranging from 6 to 36 ng/L for pesticides determined by GC-ECD, and from 54 to 118 ng/L for those determined by GC-NPD.