Simultaneous Filtration and Solid-Phase Extraction Combined with Large-Volume Injection in GC/MS for Ultra-Trace Analysis of Polar Pesticides in Surface Water
Résumé
A method combining simultaneous filtration and solid-phase extraction (SPE) with large-volume injection (LVI) in gas chromatography/mass spectrometry (GC/MS) was developed to determine 13 polar pesticides in surface water. The selected pesticides - 4 organophosphorus, 7 organonitrogens and 2 triazine degradation products - were extracted from 0.5-L samples of filtered and raw water using cartridges filled with a silica-bonded material (1 g of ISOLUTE triazine, C-18) and a depth filter. No obstruction was observed during the extraction of raw water drawn from the St. Lawrence River (concentration of suspended particulate matter (SPM) ranging from 2 to 58 mg L−1). Overall percent recoveries were satisfactory for all the target pesticides (>60%) except desisopropyl-atrazine (more polar), which varied from 29 to 46% according to sample pH. The coefficient of variation was below 10% for the majority of the target pesticides and detection limits ranged from 0.1 to 0.8 ng L−1. Applied to real samples drawn from the St. Lawrence River, this method allowed for the detection of atrazine, cyanazine, desethyl-atrazine (DEA), desisopropyl-atrazine (DIA), metolachlor and simazine, at concentrations of 6 to 91 ng L−1. Using atrazine and metolachlor as examples, the correlation between filtered and raw water samples was more significant for the former (r = 0.87) than for the latter (r = 0.67). Temporal variations in atrazine and metolachlor in filtered water drawn from the St. Lawrence River, for example, were similar whether using the established method, based on liquid-liquid large-volume extraction (LVE) combined with GC/NPD analysis, or the one proposed herein. The latter method, however, systematically found atrazine concentrations 62% higher than those obtained by the older one, applied to the same field samples. Thus, the switch to the new analytical method will require the application of a correction factor to the atrazine concentration time series acquired with the previously used method.