P.Y. Communal and A. Royer, Girpa, Angers, France
C. Schauber and T. Faye, Varian, Inc., France
Phenoxy acids are used as herbicides in agricultural and domestic areas. These compounds have to be detected at trace level to ensure a safe drinking water and food supply and verify appropriate remediation of these pollutants from contaminated sites.
Among the analytical methods developed, LC/UV exhibits poor sensitivity and inadequate selectivity for complex vegetable matrices. Negative ion ESI LC/MS has exceptionally high sensitivity for these herbicides and, with the added selectivity of MS/MS, provides accurate analyses even for the most complex sample matrices.
A highly sensitive and fast method is described below for high throughput identification and quantitation four common herbicides in vegetable extracts (Figure 1).
Varian ProStar 410 AutoSampler
Varian ProStar 210 Solvent Delivery Modules
Varian 1200L LC/MS with ESI source
Materials and Reagents
Standard solutions: pesticides were provided by the GIRPA group, France. ASE 300 (Dionex)
Florisil (Carlo Erba)
Hydromatrix bulk material (Varian Part No. 198003)
All other chemicals are reagent grade or HPLC grade.
Sample Preparation and Extraction
Liquid/solid extraction is performed using the Varian Sample Preparation Products Hydromatrix (25g) and matrix (25g onion). Two grams of Florisil is added to the homogeneous mixture followed by Accelerated Solvent Extraction (ASE) with ethyl acetate (100 mL, 100 bar, 50 0C). After 100 μL of ndodecane is added to the extract, the solution is evaporated to dryness and the residue is dissolved wi th 25 mL of acetonitrile. The final solution is then spiked with 5 ppb of herbicides.
Results and Discussion
The MS/MS conditions shown above are determined with the Automated MS/MS Breakdown software (also see Varian LC/MS Application Note 11 and 12).
The LC method uses a fast gradient program for a complete analysis in less than ten minutes. Due to the structural similarity of the acidic herbicides, incomplete resolution is observed under faster separation conditions.
The relative selectivity of LC/MS (SIM mode) and LC/MS/MS (MRM mode) are directly compared using spiked matrix samples. For pure standards, the sensitivity in the SIM mode is excellent, but even at the highest spiked concentration of 5 ppb, matrix interferences are strong. In Figure 2, SIM plots for all four herbicides shows extra peaks and baseline instability. For mecoprop, an interfering peak clearly overlaps with the analyte. As with a LC/UV detector, reliable quantitation is not possible with SIM, especially at lower concentrations.
Using the same spiked matrix extract, the increased selectivity of MS/MS is readily apparent. The MRM chromatograms in Figure 3 are interference free. The baseline is clear from any false positive peaks and no cross talk effect is observed when similar product ions are monitored in the same segment time (Table 1 and Figure 3). With the elimination of these matrix interferences, quantitation with high confidence is possible.
The MRM method has excellent linearity for all the herbicides from 0.1 to 5 ppb as shown for dichlorprop in Figure 4. The developed MRM method is also very sensitive with a LOD of 0.1 ppb for all the pesticides under the described instrument conditions.
This 1200L LC/MS/MS method is simple, fast, and hig hly sensitive. The method can quantitatively analyze pesticides at sub-ppb levels and maintain accuracy even with complex matrices. The 1200L LC/MS/MS system is recommended when sensitivity and selectivity are required at very low concentration.