rapid screening of residual pesticides on fruits and ... · rapid screening of residual pesticides...

Research Article

Received: 19 August 2014 Revised: 19 October 2014 Accepted: 26 October 2014 Published online in Wiley Online Library

Rapid Commun. Mass Spectrom. 2015, 29, 163–170

Rapid screening of residual pesticides on fruits and vegetablesusing thermal desorption electrospray ionizationmass spectrometry

Christopher Shiea1, Yeou-Lih Huang1, De-Lin Liu2, Chih-Chang Chou2, Jo-Han Chou2,Peng-Yu Chen2, Jentaie Shiea2,3 and Min-Zong Huang2*1Department of Medical Laboratory Science and Biotechnology, Kaohsiung Medical University, Kaohsiung, Taiwan2Department of Chemistry, National Sun Yat-Sen University, Kaohsiung, Taiwan3Department of Medicinal and Applied Chemistry, Kaohsiung Medical University, Kaohsiung, Taiwan

RATIONALE: Conventional mass spectrometry is encumbered by laborious and inconvenient sample pretreatment.Ambient thermal desorption electrospray ionization mass spectrometry (TD-ESI-MS) is most noted for its rapid, simple,and sensitive detection capabilities. In this study, TD-ESI-MS was used to rapidly characterize residual pesticides on thesurfaces of fruits and vegetables.METHODS: A direct sampling probe was used to obtain analytes from sample surfaces. MS and MS/MS analyses wereperformed on fruits and vegetables via TD-ESI-MS. External calibration curves and reproducibility tests were performedusing liquid pesticide standards. Pesticide decay and distribution on samples was studied, as well as the removal ofresidual pesticides via soaking in water or detergent baths.RESULTS: Since sample pretreatment was unnecessary, an analysis was completed in approximately 15 s or less, with novisible sample damage. Mass spectra were obtained for 22 pesticides. Linear calibrations (R2 from 0.9414–0.999) hadlimits of detection as low as 0.5 μg·L–1, with satisfactory reproducibilities for liquids and solids. Pesticides on samplesurfaces decayed over 2 weeks under ambient conditions. Residual pesticides localized at the fruit peel. Detergent bathsremoved more pesticide than water baths.CONCLUSIONS: TD-ESI-MS was used to rapidly screen residual pesticides in liquids and solids. Pesticides were foundon fruits and vegetables, where the decay, distribution, and removal of pesticides on samples were also explored. Due toshort analysis times, the technique allows for high-throughput analyses for applications in food and environmentalsafety. Copyright © 2014 John Wiley & Sons, Ltd.

(wileyonlinelibrary.com) DOI: 10.1002/rcm.7086

Pesticides are used throughout the world to protect fruits andvegetables from pests such as microbes, insects, and molds, butthese are also toxic to humans.[1–5] Symptoms of exposure topesticides range from fatal or severe neurotoxic effects, cancer,organ damage, interference with motor and physiologicalfunctions, and neurological, developmental, and autoimmunedisorders.[3] The use of pesticides in agriculture is thereforeregulated by food and environmental safety agencies, wheregeneral maximum residue limits or tolerances for pesticides infoods range from 0.01 to 100 mg·L–1.[4,5] Pesticides such asimazalil and thiabendazole,[2] carbaryl, carbofuran, chlorpyrifos,cypermethrin, diazinon, dimethoate, flutolanil, imidacloprid,methamidophos,methomyl, and permethrin are used in differentcountries.[3] On the other hand, pesticides like carbofuran,methomyl, methamidophos, carbofuran, carbendazim, paraquat,permethrin are severely restricted in the U.S. and E.U. due totheir high toxicities.[5] However, their use on crops in othercountries and regions means that it is necessary to analyzeimported fruits and vegetables for pesticides. Conventionally,liquid chromatography/mass spectrometry (LC/MS) and gas

* Correspondence to: M.-Z. Huang, Department of Chemistry,National Sun Yat-Sen University, Kaohsiung, Taiwan.E-mail: [email protected]

Rapid Commun. Mass Spectrom. 2015, 29, 163–170

1

chromatography/mass spectrometry (GC/MS) are used tocharacterize residual pesticides. The limits of detection (LODs)of residual pesticidesusing either technique can reachμg·L–1 levelswith the use of tandem mass spectrometry (MS/MS) and high-resolution mass spectrometry.[4,6,8] Although these techniquespossess a high degree of sensitivity and reproducibility, laboriousand time-costly sample pretreatment procedures such as solventextraction, filtration, concentration, fractionation, and derivati-zation are often necessary.[1,5–7] The coupling of GC/MS andLC/MS with increasingly efficient sample pretreatment methodssuch as QuEChERS (an acronym for an extractionmethod knownas Quick, Easy, Cheap, Effective, Rugged, and Safe) has signifi-cantly improved its use in food and environmental safetyinspections,[4,8,9] but is still constrained by necessary and incon-venient sample pretreatment steps.

Given the limitations of conventional GC/MS and LC/MSanalyses, ambient mass spectrometry (AMS) has establisheditself as a faster and simpler alternative. Samples are analyzedat room temperature and atmospheric pressure with high saltand matrix tolerances, obviating many pretreatment steps andtherefore allowing for high-throughput analysis.[6] Examplesof AMS techniques include desorption electrospray ionization(DESI),[10] direct analysis in real time (DART),[11] electrospraylaser desorption ionization (ELDI),[12] and low-temperatureplasma probe (LTP).[13] Since their development, AMS

Copyright © 2014 John Wiley & Sons, Ltd.

63

C. Shiea et al.

164

techniques have been used to characterize various analytes infood such as melamine in milk and pesticides on fruits andvegetables.[1,2,4,5,14–18] Minimal or no sample pretreatment isrequired, therefore enhancing the applicability of AMS forin situ food safety checks.[19] However, due to their sourcesetups, many AMS techniques can only analyze samples thatfit in the confines between the sampling and ionizationsource and the inlet to the mass spectrometer.[6] Large andirregularly shaped samples have to be cut down to anappropriate shape and size prior to analysis, resulting insample damage. An AMS technique that can analyzesamples regardless of sample shape and size can thusincrease analytical efficiency.We have recently reported an AMS technique known as

thermal desorption-electrospray ionization mass spectrometry(TD-ESI-MS) that fulfilled the aforementioned criteria, where asampling probe was used to collect analytes prior to thermaldesorption and electrospray ionization in an enclosed sourcefollowed byMS detection.[20] With the use of a sampling probe,TD-ESI-MS can rapidly and conveniently analyze samplesregardless of shape and size, which makes it suitable foranalyzing fruits and vegetables of different shapes and sizescompared to other AMS and conventional MS techniques. Inaddition, no pretreatment methods such as QuEChERSwere needed prior to sample analysis. Given these advantages,TD-ESI-MS was therefore used for the rapid screening ofresidual pesticides on fruits and vegetables. The quantitativecapabilities of the technique were further explored for analyzingpesticides in solution. In addition, the distribution of residualpesticides on food samples, the decay of residual pesticides overtime, and the removal of residual pesticides via soaking werealso investigated.

EXPERIMENTAL

Reagents and standards

HPLC-grade methanol was purchased from Merck (Darmstadt,Germany), while reagent-grade glacial acetic acid was pur-chased from J.T. Baker (Phillipsburg, NJ, USA). Distilleddeionized water (purified using a Milli-Q Plus apparatus fromMillipore, Molsheim, France) was used to prepare standardsample solutions. Pesticide standards were purchased fromSinon Corporation (Taichung, Taiwan) with different purities(in parentheses) including chlorpyrifos (40.8%), imidacloprid

Figure 1. Schematic diagram of the TD-ESI-Mwere (a) obtained using a sampling probe swdesorbed from the probe after placement in aplume prior to MS detection. The samplingusing a high-temperature flame.

wileyonlinelibrary.com/journal/rcm Copyright © 2014 John Wi

(9.6%), methamidophos (50%), methomyl (40%), carbaryl(85%), dimethoate (44%), diazinon (60%), cypermethrin(12.5%), permethrin (10%), paraquat (41%), halosulfuron-methyl (75%), thiabendazole (40%) and carbendazim (60%).The pesticides used in this study are commonly used andcommercially available in Taiwan, with different types(carbamates, organochlorines, organophosphates) and classes(insecticides, fungicides, herbicides). Supplementary Fig. S1(see Supporting Information) lists the structures, chemicalformulas and molecular weights of these pesticides. The abovereagents were used without further purification. Fruits andvegetables were purchased from local stores and markets andwere analyzed without pretreatment.

Instrumentation

The TD-ESI-MS technique was set up in the same configurationas described in detail in our previous publication.[20] A TD-ESIsource comprised of a direct sampling probe, thermal desorptionunit, and electrospray ionization interfacewas coupled to a triplequadrupole mass spectrometer (Agilent 6410B, Santa Clara, CA,USA) for MS and MS/MS analyses. As shown in Fig. 1, asampling probe with a commercially available stainless steelinoculating loop (radius 2 mm) was used to collect analytesfrom solid surfaces and in liquids. The tip of the sampling probe(2.5 ×10–2 cm thick as approximated using a ruler) was sweptacross solid sample surfaces over a distance of 1 cm (measuredusing a ruler) so that the sampling area was calculated to beapproximately 2.5 ×10–2 cm2 per sweep. For liquid samples,the probe was dipped in and removed from the sample solutionso that the loop held approximately 2 μL of sample solution.After sampling, the probe was inserted in the TD-ESI source.The desorption temperature of the TD-ESI source was setbetween 250 and 280 °C using a temperature controller (ANLYAT-502, Taipei, Taiwan). A metal tubing for the nitrogen gaswas attached to the heated oven in order to preheat the gas priorto entering the desorption area. The preheated nitrogen streamwas flowed from the top of the TD unit towards the ESIplume at a rate of 0.5 L·min–1. An electrospray solutioncomprised of 50%methanol (v/v) andwater with 0.1% aceticacid (v/v) was flowed through a fused-silica capillary. Ahigh voltage (–4.5 kV) was applied to the capillary to induceelectrospray ionization via solution conduction. After eachsample analysis, the sampling probe loop was cleaned byburning it with a high-temperature flame from a handheldbutane torch for 3–4 s. The absence of sample analytes was

S analytical procedure. Sample analytesept across solids or dipped in liquids, (b)heated TD unit, and (c) ionized in an ESIprobe was cleaned in between analyses

ley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2015, 29, 163–170

Screening of pesticides on fruits and vegetables using TD-ESI-MS

16

confirmed by checking the mass spectrum of the cleaned probe.Sampling, thermal desorption, electrospray ionization, detection,and probe cleaningwere completed in approximately 15 s or lessfor each analysis.

Calibration curves and reproducibility tests

Pesticide standard solutions were prepared at 1000 mg·L–1 indistilled deionized water and diluted accordingly. Standardswere analyzed immediately after preparation to avoidhydrolysis or degradation. Ten replicates for each calibrationpoint were examined. Two pesticides were randomly chosenfrom the available analytes. An aqueous standard of methomyl(1 mg·L–1) was used for reproducibility tests for liquid samples,while an aqueous standard of methamidophos (5 mg·L–1) wasapplied and dried on tomato skins for reproducibility tests forsolid samples. Since the methamidophos spike was dried andsampled as a solid, it was converted to 10 ng by multiplyingthe methamidophos concentration (5 mg L–1) by the samplingvolume (2 μL). Similar to previous experiments, the probewas swept across the tomato over a measured distanceof 1 cm (sampling area ca. 2.5 × 10–2 cm2) per analysis.

MSandMS/MSdetection of pesticides on fruits and vegetablesand in standard solutions

MS andMS/MS analyses of pesticide standardswere performedin the positivemode to obtain spectra for pesticide precursor andproduct ions. MS/MS analyses were conducted in multiplereaction monitoring (MRM) mode with collision energiesranging from 5–30 eV. For MS/MS analyses, azoxystrobin,boscalid, dimethomorph, imazalil, spirodiclofen, tebufenozide,and trifloxystrobin were sampled from the respective fruits andvegetables on which they were found. Pesticides and theirproduct ions were confirmed via comparison of their massspectra with those from an analytical protocol by the TaiwanFood and Drug Administration.Different layers of an orange were cut and sampled to

investigate the distribution of residual pesticides in fruits. Thesampling probe was swept over orange surfaces over a distanceof 1 cm (sampling area ca. 2.5 × 10–2 cm2). The peak area ofthiabendazole in each different layer of the orange wascompared with that of an aqueous thiabendazole standard forsemi-quantitative comparison. The thiabendazole concentrationin the liquid standard (5 mg·L–1) was multiplied by itssampling volume (2 μL) to yield 10 ng of thiabendazole; theconverted mass was used in semi-quantitative comparisonswith other thiabendazole masses found in different layersof the orange.

Decrease in pesticide concentrations on tomatoes over time

Standard solutions (1000mg·L–1) of thiabendazole, halosulfuron-methyl, methamidophos, methomyl, carbaryl, and chlorpyrifoswere prepared separately, after which they were sprayed onone tomato each over a square area of 16 cm2 using a spraybottle. One tomato was used for each pesticide, for a total of sixtomatoes; the absence of pesticides was confirmed for eachtomato prior to pesticide application. The probe was sweptacross the tomato over a measured distance of 1 cm (samplingarea ca. 2.5×10–2 cm2) per analysis. Sampling within the16 cm2 square area was performed from top to bottom and fromleft to right to avoid resampling. Pesticide solutions were dried

Rapid Commun. Mass Spectrom. 2015, 29, 163–170 Copyright © 2014 J

and stored at room temperature and lighting for 16 days, whereTD-ESI-MS analyses were performed on a daily basis to studythe decomposition of residual pesticides over time, with threereplicates per sample.

Decrease in pesticide concentrations on oranges after soaking

Unwashed oranges were placed in 1 L water or detergent(hand soap) baths for 5 min, after which the sampling probewas swept across the surface of the orange to characterizeresidual thiabendazole. To check if detergent baths wouldremove more pesticide, an unwashed orange was placed ina 1 L detergent bath for 5 min, after which the sampling probewas also used to check for reduction of residual pesticide. Tenreplicates were analyzed for each orange, where the samplingprobe was swept across the surface of the oranges over adistance of 1 cm (sampling area ca. 2.5 ×10–2 cm2). The peakarea of thiabendazole detected on each orange was comparedwith that of a 5 mg·L–1 standard solution to determine theamount of thiabendazole on the orange. The 5 mg·L–1

standard was multiplied by the sampled volume of 2 μLand converted to 10 ng for semi-quantitative comparisonwith thiabendazole on the orange (expressed in g). Theresidual thiabendazole on samples over areas of 2.5×10–2 cm2

was calculated by comparing the peak areas of both thethiabendazole standard (5 mg·L–1) and samples using thefollowing proportion:

Peak area for standard10 ng

¼ Peak area for orangeX ng

The amount of pesticide on each orange was set asunknown. The average surface area of the oranges (with anaverage radius of 4 cm) was assumed to be spherical[23] andcalculated to be 201.1 cm2, which was divided by thesampling area (2.5 × 10–2 cm2) to yield a value of 8042.5. Theamount of residual pesticide found within the sampling areafor each orange was multiplied by this value to obtain thetotal amounts of thiabendazole (in μg) on the surface of eachorange. The percentages of residual pesticide were calculatedfor water- and detergent-soaked oranges relative to theunwashed orange (which was assigned a residual pesticidepercentage of 100%).

RESULTS AND DISCUSSION

Calibration curves and reproducibility tests

Since a direct sampling probe was used for sampling, samplescould be analyzed regardless of shape and size and were notdamaged during analysis. Memory effects from previoussamples were easily removed from the probe simply byburning it with a high-temperature flame. A typical sampleanalysis, including sampling, desorption, ionization, detection,and cleaning, was approximately 15 s or less. Calibrationcurves for the aqueous pesticide standards are shown in Fig. 2.The enclosed TD-ESI source greatly reduced variations inanalyte introduction so that pesticide calibration curvesshowed high linearity, albeit with varying linear ranges. Forexample, the linear ranges for carbendazim, methamidophos,and thiabendazole were 1–25, 1–50, and 100–5000 μg·L–1,respectively. Estimated limits of detection (LODs) for the

ohn Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/rcm

5

Figure 2. Calibration curves for 12 aqueous pesticide standards.

C. Shiea et al.

166

pesticides are shown in Table 1, where LODs of most pesticidessuch as methomyl and carbendazim were at sub-μg·L–1 levels,while others such as paraquat and permethrin had LODs of

Table 1. Results from MS and MS/MS analyses of pesticides

Type AnalyteMass(Da)

Fungicide Azoxystrobin 403.39Boscalid 342.03Carbendazim 191.19Dimethomorph 387.12Flutolanil 323.31Imazalil 297.18Thiabendazole 201.25Trifloxystrobin 408.37

Insecticide Carbofuran 221.25Carbaryl 201.22Chlorpyrifos 350.59Cypermethrin 416.3Diazinon 304.35Dimethoate 229.26Imidacloprid 255.66Methamidophos 141.1Methomyl 162.21Permethrin 391.29Spirodiclofen 410.11Tebufenozide 352.47

Herbicide Halosulfuron-methyl 434.81Paraquat 257.16

The numbers written in bold indicate the major product ion(Supporting Information) for MS/MS spectra). Limits of detectiof the pesticides were approximately 3:1. Since TD-ESI-MS anacalculated by adding the mass of a proton to the molecular wei


100 μg·L–1. Differences in volatility, thermal stability, andproton affinity among the studied pesticides may account forthe varying LODs.[5] Reproducibility tests using 10 replicates

MS(m/z)

MS/MS(m/z)

LOD(μg·L–1)

404 372, 344 -343 307, 140 -192 160 0.5388 301 -324 262, 242 -298 255, 201, 176, 159 -202 65, 92, 131, 175 1409 206, 186 -222 165 -202 145 2.5352 97, 125, 153, 200, 296, 324 5416 191 50305 97, 153, 169 5230 88, 125, 171, 199 25256 175, 209, 212 10142 94, 125 1163 65, 88, 106, 122 0.5392 314, 332, 342 100411 313, 71 -353 297 -435 182, 403 50187 97, 157, 172 100

s of the respective pesticides (see Supplementary Fig. S3on (LODs) were determined when the signal-to-noise ratioslyses were conducted in the positive mode, m/z ratios wereght (MW) of each analyte.



or methamidophos on solids (10 ng) and methomyl in liquids(1 mg·L–1) yielded relative standard deviations (RSDs) of13.2% and 6.2%, respectively (Supplementary Fig. S2, seeSupporting Information). Methamidophos and methomyl willbe of interest in quantification studies as they are severelyrestricted in the US and the EU. It should be noted that sinceeach analysis took less than 15 s, experiments for the calibrationfcurves and reproducibility tests were completed in a fraction ofthe time required for conventional GC/MS and LC/MS analyses.These results, in addition to the rapidity and simplicity of theanalytical procedure, demonstrated the suitability of TD-ESI-MSfor rapid quantification of liquids under ambient conditions.

MSandMS/MSdetection of pesticides on fruits and vegetablesand in standard solutions

Figure 3 shows the mass spectra of residual pesticides detectedon the surfaces of local and imported fruits and vegetables (seeSupplementary Fig. S4 (Supporting Information) for moremass spectra from real samples). Supplementary Fig. S3 (seeSupporting Information) shows the results of MS/MS analysesof pesticide standard solutions and residual pesticides onsample surfaces including azoxystrobin, boscalid, imazalil,spirodiclofen, tebufenozide, and trifloxystrobin. Table 1 showsthe results of MS and MS/MS analyses of pesticides fromreal samples and aqueous standards. The inhomogeneousdistribution of pesticides on sample surfaces was taken intoaccount, so that the sampling probe was swept across differentareas of each sample.[5] It is worth noting that the analyteswere detected with no interference from the chemicalcompounds in the juice of fruits or vegetables, since sampleswere obtained from sample surfaces without solvent-basedextraction that would have otherwise extracted chemicalcompounds in the samples.[2,21]

Several pesticides including carbaryl, carbofuran, cypermethrin,diazinon, halosulfuron-methyl, imidacloprid, methamidophos,methomyl, paraquat, and permethrin were found on some

Figure 3. MS spectra from analysis of local and imcabbage, red pepper, bellfruit, tomato, and dawere imported.


samples in this study. Although these pesticides are restricted ordiscontinued in the United States or Europe due to their hightoxicities to ecosystems, these pesticides are still used on fruitsand vegetables in many countries that do not have restrictionson the aforementioned pesticides. High-throughput analysis ofpesticides from a large number of imported consumables is stillnecessary to prevent the distribution of foods with banned orrestricted pesticides.

Sincemost pesticides are applied on the surfaces of fruits, it isa legitimate concern whether these pesticides can migrate intothe fruit flesh.[17] Therefore, the extent of the permeation of aresidual pesticide into fruit was investigated using TD-ESI-MS. Figure 4 shows the distribution of thiabendazolethroughout the layers of an orange sliced into layers, whereresidual thiabendazole was detected on the surface of the peeland in the rind layer between the peel and the flesh.[2] The peakarea of thiabendazole on the peel (ca. 95% of the 10 ng standard)was estimated to be 9.5 ng, while the peak area of thiabendazolein the rind (ca. 25% of the 10 ng standard) was estimated to be2.5 ng. No pesticide was detected in the fruit layers of theorange, where the entry of pesticide into the fruit may havebeen mostly prevented by the harder peel and rind layers.

Decrease in pesticide concentrations on tomatoes over time

The processing of certain fruits and vegetables involves systemicor post-harvest pesticide application.[17] This experiment wasconducted in order to study the decay of pesticides on treatedfruits using TD-ESI-MS. Based on results shown in Fig. 5, theconcentrations of all pesticides spiked on the surfaces oftomatoes gradually decreased over the span of 10–16 days.Due to differences in volatilities, the degree of photodegradationand oxidation in air, the pesticides exhibited various decayrates.[22] Chlorpyrifos had the lowest initial peak area, whilethiabendazole had the highest. Methamidophos was found todecay in the shortest time, followed by methomyl andchlorpyrifos, carbaryl, halosulfuron-methyl, and thiabendazole.

ported fruits and vegetables. The crown daisy,te were local, while the orange and apple


167

Figure 4. Distribution of thiabendazole throughout an imported orange. Differentlayers of an orange were sampled (areas indicated by black dots). Thiabendazoleintensities from orange analyses (blue) were compared with those of a 5 mg·L–1

thiabendazole standard solution (red). The concentration of the standard (5 mg·L–1)was multiplied by the amount of sampled standard (2 μL) to yield 10 ng ofthiabendazole. Thiabendazole in the orange layers was approximated via peak areacomparisons between the standard and the different layers.

Figure 5. Decay of residual pesticides under ambient conditions over time.Colored bars for each pesticide represent the peak areas obtained from EICs forthe respective pesticides (See Table 1 for m/z ratios of the analyzed pesticides).

C. Shiea et al.

168

Decrease in pesticide concentrations on fruits and vegetablesafter soaking

Fruits and vegetables are commonly cleaned prior to con-sumption, which should dramatically remove most residualpesticides on their surfaces. In order to confirm this pheno-menon, the effects of soaking fruits on the intensities of residualpesticides were investigated using TD-ESI-MS. In addition tothe qualitative detection of residual pesticides on samples,semi-quantitative analysis was also performed to estimatepesticide concentrations on solids. The results of thesecalculations are shown in Table 2, which includes the calculatedpeak areas, RSDs, percentages of remaining thiabendazole, andtotal amounts of residual pesticide on the oranges. Sample RSDsranged from 20.1–38.0%, increasing as analyte concentrationsdecreased. Figures 6(a), 6(b), and 6(c) show the extracted ionchromatograms (EICs) for thiabendazole at m/z 202 on an


unwashed orange, an orange soaked in a water bath, and anorange soaked in a detergent bath, respectively. The amount ofthiabendazole on the water-soaked orange (Fig. 6(b)) was66.6% of the unwashed orange. The amount of thiabendazoleon the detergent-soaked orange (Fig. 6(c)) was 13.6% of theunwashed orange. Thiabendazole on oranges decreased withsoaking, where detergent baths were more effective than waterbaths at removing residual pesticide. Based on the decrease inazoxystrobin signal intensities on a grape, the amount of residualpesticide on a fruit was found to decrease with longer waterrinses (Supplementary Fig. S5 see Supporting Information).Water-based cleaning of fruits did not seem to be as effective asdetergent-based cleaning for removing residual pesticide onfruits and possibly vegetables.[2] The results of these experimentsindicate that detergent-based cleaning or longer cleaning timesare effective in reducing pesticides on fruits and vegetables.


Figure 6. Effects of soaking oranges on signal intensities of residual thiabendazole (shown as EICsfor thiabendazole atm/z 202). Prior to TD-ESI-MS analysis, samples were (a) unwashed, (b) placedin a 1 L water bath for 5 min, or (c) placed in a 1 L detergent bath for 5 min. (d) A 5 mg L–1

thiabendazole aqueous standard was also analyzed for semi-quantitative comparisons ofthiabendazole in (a)–(c). The concentration of the standard (5mg·L–1)wasmultiplied by the amountof sampled standard (2 μL) to yield 10 ng of thiabendazole. Thiabendazole in the orange layerswasapproximated via peak area comparisons between the standard and the different layers.

Table 2. Summary of results from the pesticide soaking experiment

SamplePeak area(counts)

SD(counts)

RSD (%)(n =10)

Residualpesticide (%)

Totalconc. (μg)

No wash 2027833.2 407315.1 20.1 100.0 60.6Water (5 min) 1350736.5 365475.9 27.1 66.6 40.4Water + detergent (5 min) 275165.5 104616.1 38.0 13.6 8.2ppm (aq) 2690438.4 447861.8 16.6 - 10.0

Intensities, peak areas, and standard deviations (SDs) are expressed in counts. Residual pesticide percentages are relative tothe amount of thiabendazole on the unwashed orange based on peak area and total concentration. The concentration of thestandard (5 mg·L–1) was multiplied by the amount of sampled standard (2 μL) to yield 10 ng of thiabendazole. Thiabendazolein the orange layers was approximated via peak area comparisons between the standard and the different layers.


16

CONCLUSIONS

We have demonstrated the usability of TD-ESI-MS for the rapidcharacterization of pesticides on the surfaces of fruits andvegetables and from liquids without sample pretreatment. Thenotable use of a sampling probe obviates the need for the massspectrometer to accommodate the sample. The sample is insteadconformed to the mass spectrometer without inconvenient and


unnecessary sample resizing or reshaping. As a result, thepresence, decay, and removal of residual pesticides on a varietyof samples were rapidly studied. TD-ESI-MS therefore allows forhigh-throughput analyses and can be used for application in foodand environmental safety checks with large number of samples.

Calibration curves for certain aqueous pesticide standardsyielded good linearity and limits of detection as low as0.5 μg·L–1. RSDs from pesticide reproducibility tests of


9

C. Shiea et al.

170

liquids and solids were less than 7% and 14%, respectively(Supplementary Fig. S2 see Supporting Information). TD-ESI-MShas great potential for high-throughput quantification of a largenumber of liquid samples due to the precision and rapidity ofthe technique. The ability of the probe to be separated from thesource allowed for a greater freedom of mobility, which meantthat it could be taken outside the laboratory for in situ sampling.In spite of these advantages, TD-ESI-MS is hindered by thedifficulty of pesticide quantification on solids due to thenonhomogeneous distribution of pesticides on sample surfaces.TD-ESI-MS is therefore currently only able to performqualitativeanalyses of pesticides on solids. In conclusion, TD-ESI-MS iscapable of rapid pesticide screening for pesticides due to itssimple samplingwithout sample pretreatment, and has potentialapplications in quantification and qualification of pesticides inliquids and solids, respectively.

REFERENCES

[1] J. F. Garcia-Reyes, A. U. Jackson, A. Molina-Diaz,G. R. Cooks. Desorption electrospray ionization massspectrometry for trace analysis of agrochemicals infood. Anal. Chem. 2009, 81, 820.

[2] N. Malaj, Z. Ouyang, G. Sindona, G. R. Cooks. Analysis ofpesticide residues by leaf spray mass spectrometry. Anal.Methods 2012, 4, 1913.

[3] M. Sanborn, K. Bassil, C. Vakil, K. Kerr, K. Ragan. Systematicreview of pesticide health effects.OCFP Pesticides Rev. 2012, 1.

[4] H. V. Botitsi, S. D. Garbis, A. Economou, D. F. Tsipi. Currentmass spectrometry strategies for the analysis of pesticidesand their metabolites in food and water matrices. MassSpectrom. Rev. 2010. DOI: 10.1002/mas.20307.

[5] M. C. Jecklin, G. Gamez, D. Touboul, R. Zenobi. Atmosphericpressure glow discharge desorption mass spectrometry forrapid screening of pesticides in food. Rapid Commun. MassSpectrom. 2008, 22, 2791.

[6] L. P. Li, B. S. Feng, J. W. Yang, C. L. Chang, Y. Bai,H. W. Liu. Applications of ambient mass spectrometryin high-throughput screening. Analyst 2013, 138, 3097.

[7] J. Fenoll, P. Hellin, C. M. Martinez, P. Flores, S. Navarro.Determination of 48 pesticides and their main metabolitesin water samples by employing sonication and liquidchromatography-tandem mass spectrometry. Talanta 2011,85, 975.

[8] Y. Sapozhnikova. Evaluation of low-pressure gas chromatography-tandem mass spectrometry method for the analysisof >140 pesticides in fish. J. Agr. Food Chem. 2014. DOI:10.1021/jf404389e.

[9] A.Wilkowska,M. Biziuk. Determination of pesticide residuesin food matrices using the QuEChERS methodology. FoodChem. 2011, 125, 803.

[10] Z. Takats, J. D. Wiseman, G. R. Cooks. Mass spectrometrysampling under ambient conditions with desorptionelectrospray ionization. Science. 2004, 125, 471.


[11] R. B. Cody, J. A. Laramee, H. D. Durst. Versatile new ionsource for the analysis of materials in open air underambient conditions. Anal. Chem. 2005, 77, 2297.

[12] J. Shiea, M. Z. Huang, H. J. Hsu, C. Y. Lee, C. H. Yuan, I. Beech,J. Sunner. Electrospray-assisted laser desorption/ionizationmass spectrometry for direct ambient analysis of solids. RapidCommun. Mass Spectrom. 2005, 19, 3701.

[13] J. D. Harper, N. A. Charipar, C. C. Mulligan, X. Zhang,G. R. Cooks, Z. Ouyang. Low-temperature plasma probe forambient desorption ionization. Anal. Chem. 2008, 80, 9097.

[14] S. E. Edison, L. A. Lin, L. Parrales. Practical considerations forthe rapid screening for pesticides using ambient pressuredesorption ionisation with high-resolution mass spectrometry.Food Addit. Contam. A 2011, 28, 1393.

[15] L. Luosujarvi, S. Kanerva, V. Saarela, S. Franssila,R. Kostiainen, T. Kotiaho, T. J. Kauppila. Environmentaland food analysis by desorption atmospheric pressurephotoionization-mass spectrometry. Rapid Commun. MassSpectrom. 2010, 24, 1343.

[16] M. Li, B. Hu, J. Li, R. Chen, X. Zhang, H. Chen. Extractiveelectrospray ionization mass spectrometry toward in situanalysis without sample pretreatment. Anal. Chem. 2009,81, 7724.

[17] S. Soparawalla, F. K. Tadjimukhamedov, J. S. Wiley,Z. Ouyang, G. R. Cooks. In situ analysis of agrochemicalresidues on fruit using ambient ionization on a handheldmass spectrometer. Analyst 2011, 136, 4392.

[18] S. E. Edison, L. A. Lin, B. M. Gamble, J. Wong, K. Zhang.Surface swabbing technique for the rapid screening forpesticides using ambient pressure desorption ionizationwith high-resolution mass spectrometry. Rapid Commun.Mass Spectrom. 2011, 25, 127.

[19] M. Farré, D. Barceló. Analysis of emerging contaminants infood. Trends Anal. Chem. 2013, 43, 240.

[20] M. Z. Huang, C. C. Zhou, D. L. Liu, S. S. Jhang, S. C. Cheng,J. Shiea. Rapid characterization of chemical compounds inliquid and solid states using thermal desorptionelectrospray ionization mass spectrometry. Anal. Chem.2013, 85, 8956.

[21] M. M. Gomez-Ramos, C. Ferrer, O. Malato, A. Aguera,A. R. Fernandez-Alba. Liquid chromatography-high-resolutionmass spectrometry for pesticide residue analysis in fruit andvegetables: screening and quantitative studies. J. Chromatogr. A2013, 1287, 24.

[22] L. Kantiani, M. Llorca, J. Sanchis, M. Farre, D. Barceló.Emerging food contaminants: a review. Anal. Bioanal. Chem.2010, 398, 2413.

[23] M. Clayton, N. D. Amos, N. H. Banks, R. H. Morton.Estimating peach palm fruit surface area using allometricrelationships. New Zeal. J. Crop. Hort. 2010, 23, 345.

SUPPORTING INFORMATION

Additional supporting information may be found in theonline version of this article at the publisher’s website.


rapid screening of residual pesticides on fruits and ... · rapid screening of residual pesticides...

Documents