LC/MS/MS analysis of glyphosate in food matrices without the use of derivatization

Poster Presentation

Prepared by D. Stevens
Waters Corporation, 34 Maple Street, Milford, MA, 01520, United States

Contact Information:; 508-369-8311


Glyphosate is a broad spectrum herbicide, which accounts for more than half of global herbicide sales. While discussions on the toxicological concerns of glyphosate and associated compounds continue, maximum residue limits (MRLs) are enforced globally. Analysis of glyphosate and its metabolites can be challenging due to lack of retention by reverse phase LC. Common alternatives include derivatization and ion chromatography. However, due to time-consuming sample preparation, MS incompatible solvents and the need for specialized equipment and reagents, an underivatized LC-MS/MS approach is preferred. Here, a method is presented for the direct analysis of glyphosate and its metabolite which gives sufficient retention and provides robust sensitivity to meet necessary MRLs.
Aminomethylphosphonic acid (AMPA), chlorate, ethephon, fosetyl aluminium, glufosinate, glyphosate, maleic hydrazide and phosphonic acid were targeted for a inclusion in a single LC/MS/MS method. Extracts of various foodstuffs, including juices and cereals were prepared in accordance with the Quick Polar Pesticides (QuPPe) extraction method. [1] Separation was achieved on a mixed mode/ hydrophilic interaction liquid chromatography (HILIC) column, using a mobile phase gradient at neutral pH. Targeted multi-reaction monitoring methods were used, with at least two transitions per compound to quantify and confirm analyte detection.
Preliminary data
Following evaluation of a selection of mixed mode LC columns, certain limitations were determined for the analysis. Chromatographic performance was evaluated in accordance with SANTE guideline document 11954/2015. Retention of all analytes was greater than two times the void volume, while the retention time tolerance for each analyte was within 0.1 minute of the corresponding matrix matched standard.
Overall method performance, in the absence of isotopically labeled internal standard, was evaluated by assessing linearity, accuracy and sensitivity. Satisfactory linearity was determined for each analyte, in the range from 0.001 to 0.25 mg/kg (R2> 0.995, residuals< 20%) across all matrices. Accuracy of the method was statically determined in terms of precision, repeatability and trueness where samples were spiked, pre-extraction, to three concentration levels (0.01, 0.05 and 0.1 mg/kg, n=9 for each). Recoveries were within 90 % to 107 %, while repeatability and trueness for all analytes were < 10% (%RSD) and within 10 % tolerance (%bias), respectively. Overall system robustness was determined throughout these analytical runs.
Furthermore, excellent sensitivity was achieved for all analytes in matrix below 0.001 mg/kg (except for ethephon in tomato juice at 0.0025 mg/kg). Residues of chlorate, maleic hydrazide and phosphonic acid were detected in samples. In the absence of isotopically labeled internal standards, these detections were confirmed and quantified applying standard addition calibration, as permitted by SANTE guideline document 11954/2015.
[1] EURL QuPPe method v9.2 available at (last accessed Jan. 24th 2017)