COMPARISON OF GAS-CHROMATOGRAPHY AND IMMUNOASSAY METHODS FOR THE DETECTION OF ATRAZINE IN WATER AND SOIL

Leachate and soil samples collected from different tillage systems wer e analyzed for atrazine using gas chromatography (GC) and an enzyme-li nked immunosorbent assay (ELISA) based on magnetic particle technology . Solid-phase extraction (SPE) was used to concentrate atrazine residu es in leachate samples and soil extracts before GC analysis. Atrazine concentrations determined by GC ranged from 0.1 to 600 mu g L-1 for wa ter samples and from 1.0 to 700 mu g kg(-1) for soil samples. Atrazine concentrations in 92 leachate samples as determined by ELISA were wel l-correlated (R = 0.97) with GC levels over the entire concentration r ange. Soil samples (215) were prepared and analyzed by three combinati ons of extraction/detection methods: 1) conventional extraction for GC /detection by GC analysis; 2) conventional extraction for GC/detection by ELISA analysis; 3) extraction for ELISA using a commercially avail able field kit/detection by ELISA analysis. Methanol (MeOH) in water w as the common extractant. Although the initial comparison of soil extr acts between the two different systems (Method 1 versus Method 3) was favorable (R = 0.97), two-thirds of the samples contained levels below the lower threshold for atrazine detection by both methods and some e xtracts were perceived to provide unfavorable substrate conditions (> 10% MeOH). Elimination of these data points reduced the correlation va lue (R = 0.77). To determine possible sources of variability, the extr action and detection methods were examined separately. In a comparison of extraction methods (Method 2 versus Method 3), ELISA analysis of k it extracts underestimated (R = 0.71) atrazine levels compared to thos e conventionally extracted, suggesting that differences in extraction time between methods may have accounted for reduced kit efficiency. Wh ere detection methods (Method 1 versus Method 2) were compared on spec ific extracts (< 10% MeOH), good agreement (R = 0.99) was achieved bet ween ELISA and GC values, illustrating that control of extractant conc entration is critical in using this assay for atrazine detection in so il.