Determination of glyphosate, AMPA and glufosinate in water samples by an improved UHPLC-MS/MS method

REPETTI, MARÍA ROSA; DEMONTE, LUISINA D.; MICHLIG, NICOLÁS; GARCÍA, SILVIA; DE JESÚS, JUAN JOSÉ; BELDOMÉNICO, HORACIO

Dublin

Workshop; 10th European Pesticide Residue Workshop; 2014

Glyphosate (N-phosphonomethyl glycine) is the herbicide with more extended use in Argentina. Together with glufosinate are almost exclusively associated with the advanced agricultural technology of transgenic crops. Despite the extense use of glyphosate its residue occurrence in food and environment is comparatively poorly evaluated. This is mainly due to its difficult determination at trace levels, and the limited sensitivity and robustness of various existing methods.The goal of our study was to develop an improved analytical method for the quantification of glyphosate, its major metabolite aminomethylphosphonic acid (AMPA) and glufosinate at the low ppb level in surface and groundwater and to analyse different water samples from the central region of Argentina.A derivatization step followed by liquid-liquid partition clean-up was performed before the UHPLC-ESI-MS/MS determination. Quantitation of analytes was performed in positive ESI mode, using the multiple reaction monitoring (MRM) mode with three transitions for each analyte to enhance the specificity of the method and avoid false positives. The method involves a pretreatment of 2 hours reaction step, followed by L-L partition with dichloromethane without further concentration of the sample. Sample preparation development assays involved optimization of variables such as sample size, derivatization steps (reagents, reaction time, pH) and cleanup process. We focused also on alternatives to improve analytes identification, confirmation and sensitivity. UHPLC separation was performed with mobile phase combinations of water and MeOH or MeCN using formic acid and ammonium formiate as modifiers. Electrospray ionization in modes ESI(+) and ESI(-) were compared. The ionization conditions (capillary and extractor voltages, source and desolvation temperatures and cone and desolvation flows) were optimized using statistical approaches, while cone voltage and fragmentation conditions were optimized through infusion assays.The method was validated according with SANCO/12571/2013 guidelines. Calibration range was 0.04 ? 250 μg/L running solvent and matrix matched curves, that presented adequate linearity (residuals ≤ 20%, r2>0.99). Trueness and precision was evaluated through recovery study at two concentration levels (LOQ and 200 ppb) obtaining an acceptable range of 85-110 % for all analytes. LOD and LOQ values (μg/L) were 0.2 and 0.6 (glyphosate), 0.1 and 0.2 (AMPA), 0.02 and 0.1 (glufosinate) respectively. Comparatively with other existing approaches, the developed procedure has throughput advantages, such as simpler sample preparation and faster chromatographic analysis. The method was applied to analyze real water samples (n=97) from several regional areas of Santa Fe and Entre Rios with high food production concern (soy, cereals, milk). Surface water (river, streams, channels) and groundwater were satisfactorily analyzed contributing with relevant data not previously known in the region, very useful for monitoring and risk asessment purposes.