Ana Lozano

Overcoming matrix effects using the dilution approach in multiresidue methods for fruits and vegetables.

During recent years matrix effects in liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) have quickly become a major concern in food analysis. The phenomenon of ion suppression can lead to errors in the quantification of the analytes of interest, as well as can affect detection capability, precision, and accuracy of the method. Sample dilution is an easy and effective method to reduce interfering compounds, and so, to diminish matrix effects. In this work, matrix effects of 53 pesticides in three different matrices (orange, tomato and leek) were evaluated. Several dilutions of the matrix were tested in order to study the evolution of signal suppression. Dilution of the extracts led to a reduction of the signal suppression in most of the cases. A dilution factor of 15 demonstrated to be enough to eliminate most of the matrix effects, opening the possibility to perform quantification with solvent based standards in the majority of the cases. In those cases where signal suppression could not be reduced, a possible solution would be to use stable isotope-labelled internal standards for quantification of the problematic pesticides.

Determination of pesticide residues in high oil vegetal commodities by using various multi-residue methods and clean-ups followed by liquid chromatography tandem mass spectrometry.

Several extraction methods were evaluated in terms of recoveries and extraction precision for 113 pesticides in avocado: QuEChERS with various d-SPE clean-ups (Z-Sep, Z-Sep+, PSA+C18 and silica), miniLuke and ethyl acetate. Extracts were analysed using liquid chromatography coupled with triple quadrupole mass spectrometer working in multi-reaction monitoring mode. Z-Sep and Z-Sep+ are new types of material for high lipid matrices - these two sorbents contain ZrO2, which improves fat removal from the extracts. The QuEChERS protocol with Z-Sep provided the highest number of pesticides with recoveries in the 70-120% range along with the lowest amount of coextracted matrix compounds. Subsequently, this method was validated in two matrices - avocado and almonds. In the validation recoveries at two levels - 10 and 50μg/kg - limit of quantitation, linearity, matrix effects, as well as the inter- and intraday precision were studied. In the avocado samples, 107 analytes had LOQs equal to 10μg/kg (signal to noise of quantitative transition was equal 20 or more). In the almond samples, 92 pesticides had LOQs equal to 10μg/kg (S/N≥20) and 2 pesticides at 50μg/kg. The validated method was employed in the analysis of real avocado and almond samples.

Large-scale pesticide testing in olives by liquid chromatography-electrospray tandem mass spectrometry using two sample preparation methods based on matrix solid-phase dispersion and QuEChERS.

In this work we have evaluated the performance of two sample preparation methodologies for the large-scale multiresidue analysis of pesticides in olives using liquid chromatography-electrospray tandem mass spectrometry (LC-MS/MS). The tested sample treatment methodologies were: (1) liquid-liquid partitioning with acetonitrile followed by dispersive solid-phase extraction clean-up using GCB, PSA and C18 sorbents (QuEChERS method - modified for fatty vegetables) and (2) matrix solid-phase dispersion (MSPD) using aminopropyl as sorbent material and a final clean-up performed in the elution step using Florisil. An LC-MS/MS method covering 104 multiclass pesticides was developed to examine the performance of these two protocols. The separation of the compounds from the olive extracts was achieved using a short C18 column (50 mm x 4.6 mm i.d.) with 1.8 microm particle size. The identification and confirmation of the compounds was based on retention time matching along with the presence (and ratio) of two typical MRM transitions. Limits of detection obtained were lower than 10 microgkg(-1) for 89% analytes using both sample treatment protocols. Recoveries studies performed on olives samples spiked at two concentration levels (10 and 100 microgkg(-1)) yielded average recoveries in the range 70-120% for most analytes when QuEChERS procedure is employed. When MSPD was the choice for sample extraction, recoveries obtained were in the range 50-70% for most of target compounds. The proposed methods were successfully applied to the analysis of real olives samples, revealing the presence of some of the target species in the microgkg(-1) range. Besides the evaluation of the sample preparation approaches, we also discuss the use of advanced software features associated to MRM method development that overcome several limitations and drawbacks associated to MS/MS methods (time segments boundaries, tedious method development/manual scheduling and acquisition limitations). This software feature recently offered by different vendors is based on an algorithm that associates retention time data for each individual MS/MS transition, so that the number of simultaneously traced transitions throughout the entire chromatographic run (dwell times and sensitivity) is maximized.

Pesticide analysis in teas and chamomile by liquid chromatography and gas chromatography tandem mass spectrometry using a modified QuEChERS method: Validation and pilot survey in real samples

This paper presents the validation of a modified QuEChERS method in four matrices - green tea, red tea, black tea and chamomile. The experiments were carried out using blank samples spiked with a solution of 86 pesticides (insecticides, fungicides and herbicides) at four levels - 10, 25, 50 and 100 μg/kg. The samples were extracted according to the citrate QuEChERS protocol; however, to reduce the amount of coextracted matrix compounds, calcium chloride was employed instead of magnesium sulphate in the clean-up step. The samples were analysed by LC-MS/MS and GC-MS/MS. Included in the scope of validation were: recovery, linearity, matrix effects, limits of detection and quantitation as well as intra-day and inter-day precision. The validated method was used in a real sample survey carried out on 75 samples purchased in ten different countries. In all matrices, recoveries of the majority of compounds were in the 70-120% range and were characterised by precision lower than 20%. In 85% of pesticide/matrix combinations the analytes can be detected quantitatively by the proposed method at the European Union Maximum Residue Level. The analysis of the real samples revealed that large number of teas and chamomiles sold in the European Union contain pesticides whose usage is not approved and also pesticides in concentrations above the EU MRLs.

Pesticide residue analysis of fruit juices by LC-MS/MS direct injection. One year pilot survey.

For this work, thirteen types of fruit juices (orange, pineapple, peach, apple, multifruit, mango, strawberry, tomato, pear, mandarin, grape, banana and grapefruit) were selected to develop an analytical method for the analysis of 53 pesticides by direct injection in LC-MS/MS. The preparation of the samples was very simple: an aliquot of the juice was centrifuged and it was ten-times diluted prior to analysis, which allowed reducing considerably the time and cost of the analyses. Besides, dilution of the samples permits reducing the amount of matrix going into the system, and thus, decreasing the matrix effects, so common in this type of commodities, opening the possibility to perform quantification with solvent based standards. Validation of the method was carried out in accordance with EU guidelines. Calibration curves covering three orders of magnitude were performed, and they were linear over the concentration range studied for all the matrices (from 0.1 to 100 μg L(-1)). Practical limits of quantification were in the low μg L(-1) range, far below the maximum residue levels (MRLs) of the EU regulations, which do not set specific MRLs for juices, and in this cases of processed food, MRLs of the raw product are applied. Repeatability of the instrumental method was studied in all matrices, obtaining good intra- and inter-day relative standard deviations (RSDs). The proposed method was applied to 106 real fruit juice samples purchased in different local markets during a one-year survey in order to validate the suitability for routine analysis. 43% of the analysed samples gave positive results (higher than the practical limits of quantification).

Evaluation of zirconium dioxide-based sorbents to decrease the matrix effect in avocado and almond multiresidue pesticide analysis followed by gas chromatography tandem mass spectrometry

Two sorbents containing ZrO₂ (Z-Sep and Z-Sep+) were tested as a d-SPE clean-up in combination with the QuEChERS and ethyl acetate multiresidue method in the pesticide residues extraction in avocado. All extracts were analysed using gas chromatography coupled with a triple quadrupole mass spectrometer working in multi-reaction monitoring mode. GC QToF was used to compare the amount of matrix compounds present in the final extracts, prepared according to different protocols. The highest number of pesticides with acceptable recoveries and the lowest amount of coextracted matrix compounds were provided by QuEChERS with Z-Sep. Subsequently, this method was fully validated in avocado and almonds. Validation studies were carried out according to DG Sanco guidelines including: the evaluation of recoveries at two levels (10 and 50 μg/kg), limit of quantitation, linearity, matrix effects, as well as interday and intraday precision. In avocado, 166 pesticides were fully validated compared to 119 in almonds. The method was operated satisfactorily in routine analysis and was applied to real samples.

Benefits and pitfalls of the application of screening methods for the analysis of pesticide residues in fruits and vegetables.

The goal of this study was to expand knowledge on the performance of screening methods based on accurate mass measurements using a liquid chromatography electrospray quadrupole time-of-flight mass spectrometry (LC-ESI-QTOF-MS) system operating in full scan mode and with automatic identification based on the use accurate-mass databases. The study involved the analysis of 97 pesticides, in five matrices (tomato, pepper, zucchini, orange and leek) and at three concentration levels (20, 50 and 100 μg kg(-1)). Aspects concerning optimization of the search parameters, sensitivity, matrix effects, efficiency of the algorithm search, usefulness of fragment ions, etc., are evaluated in deep. Sensitivity requirements have been identified as the main obstacle affecting the automatic identification of pesticides, especially in complex matrices, where the ionization suppression reduces the detectability of analytes. In addition, we have detected some failures in the software used for automatic data processing in terms of analysis of isobaric compounds, use of isotopic clusters, spectral deconvolution and data processing speed that hamper the correct identification in some pesticide/matrix combinations. These drawbacks should be improved in the future for its effective implementation in routine residue analysis.

Identification in residue analysis based on liquid chromatography with tandem mass spectrometry: Experimental evidence to update performance criteria

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) is one of the most widely used techniques for identification (and quantification) of residues and contaminants across a number of different chemical domains. Although the same analytical technique is used, the parameters and criteria for identification vary depending on where in the world the analysis is performed and for what purpose (e.g. determination of pesticides, veterinary drugs, forensic toxicology, sports doping). The rationale for these differences is not clear and in most cases the criteria are essentially based on expert opinions rather than underpinned by experimental data. In the current study, the variability of the two key identification parameters, retention time and ion ratio, was assessed and compared against requirements set out in different legal and guidance documents. The study involved the analysis of 120 pesticides, representing various chemical classes, polarities, molecular weights, and detector response factors, in 21 different fruit and vegetable matrices of varying degrees of complexity. The samples were analysed non-fortified, and fortified at 10, 50 and 200 μg kg(-1), in five laboratories using different LC-MS/MS instruments and conditions. In total, over 135,000 extracted-ion chromatograms were manually verified to provide an extensive data set for the assessment. The experimental data do not support relative tolerances for retention time, or different tolerances for ion ratios depending on relative abundance of the two product ions measured. Retention times in todays chromatographic systems are sufficiently stable to justify an absolute tolerance of ±0.1 min. Ion ratios are stable as long as sufficient response is obtained for both product ions. Ion ratio deviations are typically within ±20% (relative), and within ±45% (relative) in case the response of product ions are close to the limit of detection. Ion ratio tolerances up to 50% did not result in false positives and reduced the false negative rate for pesticides with product ions in the low S/N range to <5%. Without ion ratio criterion, two false positives were obtained in 105 non-fortified samples. Although the study has been conducted for pesticides residues in fruits and vegetables, the impact of these findings is believed to extend towards other application areas and possibly support adjustment or consolidation of criteria across other analytical domains.

A sensitive and efficient method for routine pesticide multiresidue analysis in bee pollen samples using gas and liquid chromatography coupled to tandem mass spectrometry

Several clean-up methods were evaluated for 253 pesticides in pollen samples concentrating on efficient clean-up and the highest number of pesticides satisfying the recovery and precision criteria. These were: (a) modified QuEChERS using dSPE with PSA+C18; (b) freeze-out prior to QuEChERS using dSPE with PSA+C18; (c) freeze-out prior to QuEChERS using dSPE with PSA+C18+Z-Sep; and (d) freeze-out followed by QuEChERS using dSPE with PSA+C18 and SPE with Z-Sep. Determinations were made using LC-MS/MS and GC-MS/MS. The modified QuEChERS protocol applying a freeze-out followed by dSPE with PSA+C18 and SPE clean-up with Z-Sep was selected because it provided the highest number of pesticides with mean recoveries in the 70-120% range, as well as relative standard deviations (RSDs) typically below 20% (12.2% on average) and ensured much better removal of co-extracted matrix compounds of paramount importance in routine analysis. Limits of quantification at levels as low as 5μgkg(-1) were obtained for the majority of the pesticides. The proposed methodology was applied to the analysis of 41 pollen bee samples from different areas in Spain. Pesticides considered potentially toxic to bees (DL50<2μg/bee) were detected in some samples with concentrations up to 72.7μgkg(-1), which could negatively affect honeybee health.

Determination of pesticides in edible oils by liquid chromatography-tandem mass spectrometry employing new generation materials for dispersive solid phase extraction clean-up.

The goal of this work was to evaluate the efficiency of several sorbents on removal fats from edible oils (olive, soya and sunflower) during the clean-up step for posterior determination of 165 pesticides by UHPLC-QqQ-MS/MS system. The extraction procedure employed in this work was the citrate version of QuEChERS method followed by a step of freezing out with dry ice and clean-up evaluation using i) PSA with magnesium sulfate (d-SPE); ii) magnesium sulfate and Z-sep sorbent (d-SPE); iii) Z-sep (column SPE) and iv) Agilent Bond Elut QuEChERS Enhanced Matrix Removal-Lipid (EMR-Lipid). After evaluation of the recovery results at 10, 20 and 50μgkg(-1), the EMR-Lipid showed important advantages comparing to the other sorbents evaluated, such as better recovery rates and RSD%. The method was validated at the three concentrations described above. Analytical curves linearity was evaluated by spiking blank oil samples at 10, 20, 50, 100 and 500μgkg(-1). The method demonstrated good recoveries values between the acceptable range of 70-120% and RSD%<20 for most of evaluated pesticides. In order to evaluate the performance of the method, this same procedure was employed to other oils such as soya and sunflower with very good results.