Food Safety: Determining Fungicides In Fruits

The identification and quantification of fungicides in fruits and vegetables is important. Increasingly, liquid chromatography/mass spectrometry (LC/MS) is being used for the analysis of many polar fungicides of the European Union (EU) Directives and is rapidly becoming an accepted technique in pesticide residue analysis for regulatory monitoring. In particular LC/MS works well on the various families of fungicides that are polar and labile. Thus, LC/MS methods may be preferred over the older gas chromatography/mass spectrometry (GC/MS) methods.

For example, the use of MS and MS/MS for pesticides (including fungicides) in food has been reviewed and there are currently about 100 published papers dealing with the analysis of pesticides in food. Of these, there are approximately 25 papers that deal with fungicides by LC/MS, the majority published in the early 2000s. Thus, LC/MS is an emerging technology for the analysis of food products. Interestingly though, none of the reported papers in this most recent review deal with the use of liquid chromatography/time-of-flight (LC/TOFMS) and accurate mass analysis for fungicides in food.
Thus, there is an important need for research studies and methods development on the analysis of pesticides in food by accurate mass using LC/TOFMS. Our study is one of the first of its kind to examine the new Agilent LC/MSD TOF time-of-flight mass spectrometer for the analysis of fungicides (carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole) in food. (The full report can be found at www.agilent.com/chem, Agilent technologies publication 5989-1924eN.) This topic was chosen because of the relevance of these fungicides and their significant use on apples, oranges, lemons, melons, tomatoes, broccoli, and peppers.

The following is an abbreviation of the application and illustrates the data for carbendazim. The sample was prepared using the
QuEChERS method.

Results
Figure 1 shows the LC separation of five fungicides: carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole. The sample contained the fungicides at 0.125 milligrams/kilogram (ppm) in a fortified lemon extract. The window of extraction was clean, which is commonly a feature of accurate mass extraction of ions, where the width of the window of extraction may be narrowed to (0.02 amu) or ~100 ppm. In figure 1, the window of extraction was 0.1 amu.

At concentrations as low as 0.05 milligrams/kilogram, the extracted ions (m/z 192, 202, 346, 388, and 404) still yielded clean chromatograms testifying to the importance of accurate mass and its ability to give clean extracted ion chromatograms (EICs) with a narrow mass window. This accurate mass window is reflected in the determinations of the accurate mass of the protonated molecule of each of the fungicides. Table 1 shows the elemental composition, the exact mass, and errors, in mDa and ppm, for each of the fungicides at the 0.50 mg/kg concentration. Mass accuracy was always better than 2 ppm in all the fruit and vegetable matrices, except for dimethomorph.

These results show that the use of continuous calibration is effective for accurate mass across an order of magnitude concentration range in complex fruit and vegetable matrices.

Figure 2 shows the fragmentation pathway of carbendazim, based on accurate mass spectra with pure standards in a methanol solution.

The fragmentation ions of the LC/MSD TOF were verified with the LC/MSD Trap spectra at MS2 with identical chromatographic conditions. With carbendazim, the MS2 shows that the protonated molecule loses methanol (32 Da) to give the m/z 160 fragment ion (Figure 2). This fragment ion of m/z 160 is also seen in the LC/MSD TOF.

Linearity and detection limits
Calibration curves were established for the five fungicides using both the LC/MSD TOF and LC/MSD Trap over the analyte concentration range of interest, which was from 0.01 milligrams/kilogram to 0.5 milligrams/kilogram (ppm) in solvent and in each of the fruit and
vegetable matrices (orange, lemon, melon, broccoli, and pepper). Results showed the similarity among matrices with the LC/MSD TOF and LC/MSD Trap, given the variability in fruit and vegetable matrices. For example, Figure 3 shows the standard curve for the fungicide carbendazim. This figure demonstrates that there was little or no matrix suppression in the LC/MSD TOF system for the fruit extracts up to 0.5 ppm of parent compound for carbendazim.

The best accuracy for reporting concentrations was with the standard curve made up in matrix for each fruit or vegetable. This was the procedure that was used for unknown analysis in real fruit and vegetable samples. Furthermore, Figure 3 also shows that the standard curve was linear across the range of concentration tested with correlation coefficients of 0.990 to 0.999 for all matrices tested (Table 2). Similar results for standard curves were seen with the LC/MSD Trap, with correlation coefficients typically of 0.990±0.001.

Table 3 shows the limits of detection (LODs) of the fungicides for various matrices. Typically, the values of Table 3 vary from 0.001 ppm to 0.010 ppm. The European standard for pesticides with no regulatory standard is 0.010 ppm. Thus, the LC/MSD TOF is sufficiently sensitive to detect these compounds in all matrices.

Conclusions
• LC/MSD TOF analysis is a useful tool for identification of fungicides in fruits and vegetables and is a new tool for environmental food chemistry.

• Quantitation is easily possible over 2 orders of magnitude with accuracy better than 3 ppm, typically less than 2 ppm, and in this work was 1 ppm in ESI+ ion for most compounds.

• Elemental composition of fungicides and fragment ions are possible with LC/MSD TOF, and further confirmed with LC/MSD Trap.

• LODs of the five fungicides in fruit and vegetables were from 0.001 to 0.010 micrograms/gram. These concentrations are equal to or better than the EU directives for controlled fungicides in fruits
and vegetables.

Click here for Figure 1
Click here for Figure 2
Click here for Figure 3

Click here for Table 1, Table 2, Table 3