Simultaneous LC-MS/MS quantitation of 180 pesticides in tea 

Hitha P P¹, Sashank Pillai¹, Holly Lee², Craig M. Butt³

¹SCIEX, India; ²SCIEX, Canada; ³SCIEX, USA

Abstract

This technical note describes the quantitation of 180 pesticides in tea. The SCIEX Triple Quad 4500 system achieved sub-ng/mL in-vial limits of quantitation (LOQs) in both aqueous and matrix-matched calibration with suitable mean accuracy and precision (n = 3) (Figure 1). For the matrix-matched calibration standards, the in-sample equivalent LOQ range was 1.25 to 100 ng/g. Apparent and absolute recoveries were evaluated at 10 ng/g and 100 ng/g spikes in a composite tea matrix. Apparent and absolute recoveries ranged from 70 to 128% and 70 to 111%, respectively, with overall precision ranging from 1.9 to 20 %CV (n = 5). Most target pesticides (74%) exhibited acceptable matrix effects (<±30%). Analysis of locally purchased tea revealed the detection of acetamiprid, carbendazim, dimethoate, piperonyl butoxide, thiacloprid and thiamethoxam.

Key benefits of the SCIEX Triple Quad 4500 system for the quantitation of pesticides in tea
 

• Good quantitative performance at sub-to-low ppb LOQs. SCIEX QTRAP 4500 system achieved sub-to-low ng/mL in-vial LOQs (0.025–2 ng/mL) in aqueous and matrix-matched calibration standards. In-sample equivalent LOQ ranged from 1.25 to 100 ng/g for the matrix-matched standards (n = 3). 
• Robust method performance in matrix spikes. Majority of pesticides showed good apparent and absolute recoveries, with further improvement observed for some pesticides when using matrix-matched calibration standards. 
• Method applicability in commercial tea: Thiamethoxam, acetamiprid, carbendazim, thiacloprid and piperonyl butoxide were positively detected in 5 locally purchased tea samples. 

Introduction

Tea is one of the most widely consumed beverages in the world and possesses many health benefits, including cardiovascular disease prevention and antioxidant, as well as anti-carcinogenic and anti-inflammatory effects.¹ Further, tea is an important economic commodity for many countries and is traded globally. However, the pesticide residues applied to tea can be transferred to the beverage during brewing and represent a potential human health risk while consuming. Tea matrices are known to be complex and analytically challenging due to the presence of co-extractables such as pigments, caffeine, organic acids and sugars.² Therefore, it is important to develop sensitive and robust analytical methods for measuring pesticides in tea. 

The European Union (EU) and several countries, including the United States, China, India and Japan, have set maximum residue limits (MRLs) for pesticides in tea. There is no global consensus on pesticide MRL values in tea, and the MRL list and level vary by country. Therefore, it is crucial that analytical methods are broad in scope and flexible to meet the changing needs of individual nations. In this technical note, a sample preparation and LC-MS/MS method is developed for 180 pesticides in tea, covering a broad range of pesticide classes. 

Methods

Standard preparation: The standard mix of 206 pesticides was purchased from TRC Canada. The intermediate stock solutions were prepared in 50:50 (v/v) of acetonitrile/water. Calibration standards were prepared in 90:10 (v/v) acetonitrile/water with 0.01% (v/v) formic acid at levels ranging from 0.025 to 50 ng/mL. In addition, matrix-matched calibration standards were prepared through post-extraction spikes, following the sample preparation procedure described below. 

Matrix spike recovery experiments: A homogenate of 5 different teas was prepared by mixing 20 g of each sample and thoroughly mixing. This homogenate was used for the method development experiments. Method recovery was determined by performing matrix pre- and post-extraction spike experiments (n = 5) at 10 ng/g and 100 ng/g. In addition, procedural blanks comprised of water were processed (n = 3). The apparent and absolute recoveries were calculated to evaluate the method performance.

Sample preparation: Tea samples were extracted using QuEChERS, and the extracts were cleaned using dispersive solid phase extraction (dSPE). Briefly, 2 g of tea was weighed into a polypropylene tube and soaked with 5 mL of water for 15 min. Then, 10 mL of 0.1% (v/v) formic acid in acetonitrile was added, followed by 4 g of magnesium sulfate (MgSO₄), 1 g of sodium chloride (NaCl) and 1.5 g of disodium hydrogen citrate sesquihydrate. The solution was vortexed and centrifuged at 4500 rpm for 5 minutes. 1 mL of the supernatant was transferred to a clean vial containing 150 mg of MgSO₄, 50 mg of primary secondary amine, 50 mg of C18 and 50 mg of graphitized carbon black. The vial was vortexed and centrifuged at 4500 rpm for 5 min. The supernatant was diluted 10x with water before analysis. 

Chromatography: Chromatographic separation was performed on an ExionLC AD system using a Kinetex Biphenyl column (2.6 μm, 100 mm x 3 mm, P/N 00D-4622-Y0). The flow rate was 0.4 mL/min, and the mobile phase and gradient conditions used are shown in Table 1. The injection volume was 10 μL and the column oven temperature was set to 30°C. The autosampler temperature was set to 10°C.  

Mass spectrometry: Samples were analyzed using the SCIEX Triple Quad 4500 system with electrospray ionization, and data were acquired using scheduled multiple reaction monitoring (sMRM) mode with polarity switching. The target cycle time was 1650 milliseconds, with a minimum and maximum dwell time of 10 and 20 milliseconds, respectively. These conditions ensured that at least 10-12 data points across the chromatographic peak were acquired (Figure 2). Optimized source parameters are presented in Table 2. Optimized compound-specific parameters were used with two transitions per analyte.

Data processing: All data were acquired and processed using the SCIEX OS software (version 3.3.1).

Sensitivity, accuracy, precision and linear dynamic range in solvent-based and matrix-matched calibration standards

Initial method development showed that 180 out of the 206 targeted pesticides displayed good chromatographic peak shape and sensitivity. This trend was expected due to the diversity of physical-chemical properties in the large pesticide mix. For example, some macrocyclic lactones, such as avermectin, are susceptible to thermal degradation in the ion source. Therefore, final method development and application only considered these 180 pesticides. 

The quantitative performance of the solvent-based and matrix-matched calibration standards was investigated for the 180 pesticides using the SCIEX Triple Quad 4500 system (n = 3 injections). Overall, the in-vial LOQs ranged from 0.025 to 2 ng/mL in both calibration standard sets. For the matrixmatched calibration standards, the in-sample equivalent LOQ range was 1.25 to 100 ng/g. The LOQ was selected based on the 2 MRM transitions, retention time, ion ratio, mean accuracy between 70-130% and precision %CV ≤20%. The exceptions were for thiofanox and trichlorfon which only showed one transition. The pesticide panel covered diverse structures and physical-chemical properties, resulting in the observed sensitivity levels. Further, the chosen source parameters represented the optimized conditions for most of the analyte panel and may not have been ideal for some analytes. The LOQ density plots are presented in Figure 1. Overall, the distribution of the in-vial LOQs are similar between the solvent and matrix-match standards, demonstrating that matrix effects did not significantly impact sensitivity. Further, the density plots show that the majority of the in-vial LOQs were between 0.1 and 1 ng/mL (in-sample equivalent of 5-100 ng/g). 

Both calibration sets showed good accuracy and precision at the LOQ level. Mean LOQ accuracy ranged between 70-129% and 70-128% for the solvent-based and matrix-matched calibration standards, respectively. Further, mean LOQ precision ranged between 0.60-16% and 0.10-16% for the solvent-based and matrix-matched calibration standards, respectively. In addition, good linearity was observed with r² values ≥0.99 for all pesticides in the matrix-matched standards and for most of the solvent-based standards, except for etaconazole, etoxazole, and furathiocarb (r² ≥0.985). 

Quantitative performance in tea matrix spikes  

The apparent and absolute recoveries, and matrix effects were determined at the 10 ng/g and 100 ng/g spiking levels. The apparent recovery was calculated from the pre-extraction spike samples and quantified against either the solvent-based or matrix-matched calibration standards. When quantifying against solvent-based standards, the apparent recovery considers the analyte extraction recovery and matrix effects. 

The recovery trends for the 10 and 100 ng/g spikes are presented in Figure 3. Considering the apparent recoveries when quantified against the solvent-based calibration standards (left-side panel, Figure 3), recoveries for most pesticides ranged between 50-100% in 10 ng/g spikes. Further, the range of apparent recoveries increased to 60-100% in the 100 ng/g spikes. However, when quantified against the matrix-matched standards, the apparent recoveries were predominantly between 60-110% for the 10 ng/g and between 75- 105% in the 100 ng/g spikes, respectively (middle panel, Figure 3). Typically, the acceptable recovery range is 70-130%, although some regulations allow for 50-150% recovery at lower spike levels. Overall, these results demonstrate that good recoveries were achieved for the large pesticide panel in tea. Further, the spike experiments showed that matrix-matched calibration standards were necessary to some compounds which exhibit high matrix effects.

The absolute recovery was calculated as the ratio of the pre- to post-extraction spike area counts and this parameter evaluates the sample preparation extraction efficiency. Both spiking levels showed good absolute recoveries with most pesticides ranging from 70-95% for the 10 ng/g spike, and 65-100% for the 100 ng/g spike (right-side panel, Figure 3). These results demonstrate that the QuECHERS/dSPE method resulted in good extraction recovery for most pesticides studied. 

Figure 4 shows example XICs for azoxystrobin in the recovery samples as well as matrix blank and calibration standard LOQ

Matrix effects 

The matrix effects were assessed in the 100 ng/g spike samples since the apparent recovery experiments demonstrated improvement when quantifying against the matrix-matched calibration standards. The matrix effects were calculated based on the peak areas from the 100 ng/g post-extraction spike and the equivalent solvent calibration standards (2 ng/mL) using the following equation:  

Figure 5 shows the distribution of the matrix effects for the pesticides. Most compounds were within the ±30% range indicating slight matrix effects, while 46 pesticides exhibited matrix suppression or enhancement outside of this range

Method applicability in real-world tea sample

The sample preparation and LC-MS/MS method were applied to 5 tea samples purchased from a local store. Few pesticides were detected in the tea samples (Table 3 and Figure 6). The detected pesticides included acetamiprid, carbendazim, dimethoate, piperonyl butoxide, thiacloprid and thiamethoxam. Interestingly, most of the detected pesticides (acetamiprid, thiacloprid, thiamethoxam) were from the neonicotinoid class and may indicate the prevalent use of this pesticide class on tea. These pesticides were not detected in the procedural blanks, confirming their presence in the tea. The calculated concentrations in the tea range from the low- to 100s of mg/kg levels. However, the thiamethoxam concentrations are most likely not accurate due to the high matrix effects observed (- 81%). 

Conclusion

This technical note demonstrated:

• A sample preparation and LC-MS method for the analysis of 180 pesticides in tea matrix using the SCIEX Triple Quad 4500 system.
  
  
• Sub- to low-ng/mL in-vial LOQs (0.025-2 ng/mL) in both aqueous and matrix match calibration standards. In-sample equivalent LOQs ranged from 1.25 to 100 ng/g for the matrix-matched standards.
  
  
• Most of the pesticides showed good apparent and absolute recoveries, but recovery of some pesticides improved when using matrix-matched calibration standards
  
  
• Detection of thiamethoxam, acetamiprid, carbendazim, thiacloprid and piperonyl butoxide in 5 locally purchased tea samples demonstrating method applicability to real-world tea samples

References
 

1. Fernandes, I.D.A.A.; Maciel, G.M.; Bortolini, D.G.; Pedro, A.C.; Rubio, F.T.V.; de Carvalho, K.Q. The bitter side of teas: Pesticide residues and their impact on human health. Food Chem. Toxicol. 2023, 179, 113955. DOI: 10.1016/j.fct.2023.113955
   
   
2. Chen, G.; Cao, P.; Liu, R. A multi-residue method for fast determination of pesticides in tea by ultra performance liquid chromatography-electrospray tandem mass spectrometry combined with modified QuEChERS sample preparation procedure. Food Chemistry 2011, 125(4), 1406-1411. DOI: 10.1016/j.foodchem.2010.10.017