Standard kresoxim-methyl (98.0% purity) was purchased from Dr. Ehrenstorfer GmbH (Augsburg, Germany). A suspension emulsion (SE) containing 40% kresoxim-methyl was obtained from Shanxi Biaozheng Crop Science Co. Ltd (Shanxi, China). HPLC-grade acetonitrile was purchased from J. T. Baker (Phillipsburg, NJ, United States). MS-grade formic acid was obtained from Aladdin Industrial Corporation (Shanghai, China). Analytical grade activated anhydrous magnesium sulfate and sodium chloride were obtained from Merck (Darmstadt, Germany). Primary secondary amine (PSA; 40–60 μm in size) and octadecylsilane (C₁₈; 40–60 μm in size) were purchased from Agela Technologies Company (Tianjin, China). Wahaha pure water was used (Wahaha Group Co., Ltd., Hangzhou, China).

Field trials to investigate the dissipation of kresoxim-methyl and residues in banana and soil were carried out in Guangzhou (GZ: 113°E, 22°N, oceanic subtropical monsoon climate) in Guangdong Province, and Fangchenggang (FCG: 108°E, 21°N, oceanic monsoon climate) in Guangxi Province in 2016 according to the “Guidelines for Pesticide Residue Field Trials” (NY/T 788-2004) issued by the Ministry of Agriculture of China. Plots where kresoxim-methyl had never been previously applied were selected. The average rainfall in GZ and FCG was 291 mm and 465 mm, respectively; the average temperature was 23°C and 33°C, respectively, and the average relative humidity was 69% and 78%, respectively. Trials were conducted from September 9 to December 14 in GZ and from May 12 to July 19 in FCG. The characteristic properties of the soil in the fields were as follows: sandy loam in GZ with an organic matter content of 23.0 g kg⁻¹ and a pH of 6.71, and clay in FCG with an organic matter content of 19.9 g kg⁻¹ and a pH of 5.83.

Forty percent kresoxim-methyl suspension emulsion (SE) dissolved in water was sprayed to banana plots and soil plots once, at a dosage of 750 mg/kg (1.5 times the recommended dosage) when the banana reached 50% of the size of a ripe banana. Each banana plot contained three banana trees, and the soil plot measured 30 m². In addition, there was a 1-m buffer zone between plots, each plot was set in triplicate, and no-pesticide application was treated with clear water. Representative banana and soil samples were randomly collected from each plot to evaluate dissipation at 2 h (calculated as the original concentration) and 1, 3, 7, 10, 14, 21, 28, 35, and 42 days after spraying.

The kresoxim-methyl formulation (40% SE) was applied to the final residual experimental plots at a low-dose level of 500 mg/kg (the recommended dosage) and a high-dose level of 750 mg/kg (1.5 times the recommended dosage) three and four times for two treatments. The re-treatment interval was 7 days, which was the recommended minimum re-treatment interval. Terminal residue samples were randomly collected from each plot at different pre-harvest intervals (PHIs; 35 and 42 days) to investigate its distribution in the whole banana, pulp, and peel. Twelve bananas were randomly collected with a knife from each treatment plot.

Representative banana samples were cut into smaller pieces according to the four-point method. Samples of 0.5 kg were homogenized in a food processor (Hobart FP-400, United States). Soil samples were thoroughly mixed, and 200 g of subsample was used for kresoxim-methyl analyses. All collected samples were stored in a freezer at −20°C until analysis. There was no need to consider the storage stability problem of analytical samples, because the kresoxim-methyl was stable at −20°C at least 18 months according to the JMPR report 1999.

The sample (5 g) was weighed in a 50-ml centrifuge tube, 20 ml of acetonitrile was added, and the screw cap was closed and vigorously shaken for 1 min using a vortex mixer at maximum speed. Next, 1 g of NaCl and 4 g of anhydrous MgSO₄ were added, vortexed for 1 min, and centrifuged for 2 min at 5,000 rpm. An aliquot of 2 ml was drawn from the supernatant and added in a pre-prepared 5-ml tube with 50 mg of PSA and 50 mg of C₁₈; then, this tube was immediately shaken by hand, vortexed for 10 s, and centrifuged for 2 min at 5,000 rpm. Finally, the extract was filtered through a filter membrane (0.22 µm, Millex-GV, Millipore, Bedford, MA, United States) and then analyzed by HPLC-MS/MS. A blank sample was collected and subjected to pretreatment using the same procedure described above.

A high-performance liquid chromatography system (Agilent 1,200, United States) tandem triple quadrupole mass spectrometer (AB SCIEX 4000Q TRAP, United States) was employed to separate and quantify kresoxim-methyl. The analyses were separated by a poroshell-120 EC-C₁₈ column (150 mm × 3.0 mm, 2.7 μm, Agilent, United States). The oven temperature was set at 35°C; the flow rate and injection volume were 0.35 ml/min and 5 μl. The mobile phase consisted of (A) acetonitrile and (B) 0.1% (v/v) formic acid aqueous solution, and the ratio of solvent A to solvent B was 75:25. This current speed was kept constant throughout the 7-min analysis time.

Electrospray ionization was carried out using the positive ion mode (ESI⁺). The capillary voltage was set at 5.0 kV. The temperature of the source was 550°C. Nitrogen was used as the nebulizer and drying gas at 50 psi and 350°C. Multiple reaction monitoring (MRM) was selected. The selected precursor ion was m/z 314.0 with product quantitative and qualitative ions of m/z 116.1, m/z 222.0, and m/z 235.4, respectively, when the corresponding collision energy levels were set at 17.3 V, 18.9 V, and 24.3 V. Under the above conditions, the kresoxim-methyl retention time was approximately 3.9 min (total run time = 7 min).

The matrix effect is usually considered to be a major drawback in trace-level analysis. Matrix-matched calibration is an efficient method to avoid matrix effects (ME) (Utture et al., 2011; Wiest et al., 2011; Rahman et al., 2015). ME were calculated using the following equation: ME ( % ) = ( ( m matrix / m solvent ) ‐ 1 ) × 100 %

Where m_matrix and m_solvent are the slope of the calibration curves in the matrix and in the pure solvent, respectively. An ME with a positive (negative) value indicates that the analyte response is enhanced (suppressed) by the matrix; a zero value indicates that the analyte response is removed by the matrix. The result showed that the ME of the whole banana and peel were all negative values (−3.9% and −36%), which indicated that they caused signal suppression. In contrast, the ME of the pulp caused signal enhancement. The calibration curves were constructed by plotting standard concentrations against the response of the chromatographic peak, and the correlation coefficients for all the calibration curves were greater than 0.999 (Supplementary Table S1).

Recoveries were determined to evaluate the accuracy and precision of this method. Banana and soil samples spiked at three levels of fortification (0.01, 0.1, and 1.0 mg kg⁻¹), with five replicates for each level (Table 1), and the selectivity of the method is illustrated in Figure 2. The average recoveries ranged from 90.3% to 97.7% with RSDs of 2.29%–4.14% in the whole banana, from 90.6% to 96.4% with RSDs of 2.74%–3.03% in pulp, from 85.9% to 96.3% with RSDs of 3.03%–4.56% in peel, and from 94.2% to 99.7% with RSDs of 1.89%–5.77% in soil, respectively. LOQs were defined as the minimum fortified concentrations of analyte in the matrix with a signal-to-noise ratio of 10, which was 0.12 μg kg⁻¹ for the whole banana, pulp, peel, and soil. LODs were 0.04 μg kg⁻¹, which could be calculated at a signal-to-noise ratio of 3.

The initial concentrations of kresoxim-methyl residues in the whole banana were 3.33 mg kg⁻¹ in GZ and 2.84 mg kg⁻¹ in FCG. The half-lives were 5.7 and 4.8 days in GZ and FCG. The degradation rate was gradually decreased with the time elapsed, and the concentrations of kresoxim-methyl reached more than 55% of the initial concentration after 14 days. These values are lower than those observed in other studies, where the half-lives of kresoxim-methyl were 10.6–14.5 days in greenhouse strawberry (Chen et al., 2018) and 5–10 days in wheat (Yu and Cai, 2010). The half-lives of kresoxim-methyl in banana were higher than cucumber (1.5–2.1 days) (Liu et al., 2012) and were similar to the half-lives of 4.58–4.77 days in apple (Malhat et al., 2013). The dissipation rate of kresoxim-methyl in GZ was slower than in FCG. The experiment in GZ was conducted from September to December, and from May to July in FCG; thus, the variety of banana and growth dilution may be major factors during the dissipation. The precipitation could not be an important factor due to the low water solubility of kresoxim-methyl (2 mg/L, 20°C). The temperature and humidity in GZ (15.6°C, 64%) were lower than those in FCG (26.5°C, 78%); hence, it may cause the lower initial concentration and short half-lives in FCG. Some chemical and physical factors like experiment plot, growth speed, chemical properties of pesticide, and light could play a significant role in the degradation of fungicide. The half-life (t _1/2) and other statistical parameters of dissipation were calculated from the experimental data, and the results are summarized in Table 2 and Figure 3.

The initial kresoxim-methyl residues in soil were 1.43 mg kg⁻¹ in GZ and 1.23 mg kg⁻¹ in FCG. The half-lives of kresoxim-methyl in soil were 5.5 and 6.5 days in GZ and FCG, respectively. The dissipation rate of kresoxim-methyl in soil was faster than in apple soil (12–13 days) (Zhang et al., 2008), strawberry soil (8–12 days) (Li et al., 2009), and wheat soil (7–13 days) (Yu and Cai, 2010). The results showed that different chemical and physical factors of soil like pH, cation exchange capacity, and microorganism might influence the adsorption and leaching of kresoxim-methyl in soil, which led to different dissipation rates in soil.

The terminal residue concentrations of kresoxim-methyl in bananas at 35 and 42 days after the last application were investigated. When kresoxim-methyl formulation was applied 3 times and 1.5 times the recommended dosage, the residues were 0.12–0.54 mg kg⁻¹ and 0.13–0.79 mg kg⁻¹, respectively, based on PHI (35 days). It was easily observed that the concentrations under different dosages followed the trend that residues increased with the increase in the applied dosage. When kresoxim-methyl was applied three and four times the recommended dosage based on PHI (42 days), the concentrations of kresoxim-methyl were 0.03–0.08 mg kg⁻¹ and 0.12–0.25 mg kg⁻¹, respectively. There were positive relations between residues and application times. When kresoxim-methyl was applied three times the recommended dosage based on different PHI (35 days and 42 days), the kresoxim-methyl residues were 0.12–0.54 mg kg⁻¹ and 0.03–0.08 mg kg⁻¹, respectively. It revealed that the residues were lower with a longer harvest interval.

The terminal residues were 0.03–0.08 mg kg⁻¹, 0.06–0.17 mg kg⁻¹, and <0.01 mg kg⁻¹ in the whole banana, peel, and pulp, respectively, when kresoxim-methyl was applied three times the recommended dosage based on PHI (42 days). The residues in peel were much more than the whole banana, in which the concentration in peel was about 2 times that of the whole banana. It revealed that most of the residues were concentrated in the peel (Table 3). The peel accounts for about 35% of the whole fruit weight (Vu et al., 2017). The peel has been traditionally used for the treatment of burns, inflammation, and so on (Pereira and Maraschin, 2015). Thus, high levels of residue in the peel should be paid attention due to its broad use in the medical field and animal forage.

The residues of kresoxim-methyl in the whole banana ranged from 0.03 to 0.08 mg kg⁻¹ (high residue, HR) when kresoxim-methyl was applied three times the recommended dosage based on PHI (42 days), which was higher than the MRL value of EU (0.01 mg kg⁻¹) and lower than Japan (5 mg kg⁻¹). In China, the principle of pesticide maximum residue limit standard is based on at least double the HR value with rounding up; thus, the MRL value for kresoxim-methyl in banana is 0.5 mg kg⁻¹.

The surface of the pulp was covered with peel, and kresoxim-methyl SE was applied to the peel, so the pulp is not directly exposed to the fungicide. First, the fungicide needs to pass through a wax layer of banana before slowly entering the pulp. The kresoxim-methyl is a fat-soluble fungicide due to its Kow logP (3.4, pH 7, 25°C) (the e-pesticide manual, Version 5.0), which indicated that it is adsorbed in the peel. Other factors like wax thickness, rainfall, and temperature will also influence the residue levels in the peel. Hence, the residues that permeated into the pulp were very low. The above results showed that the kresoxim-methyl residues were all below 0.01 mg kg⁻¹ in pulp, which was far below the MRL value in China.

The measure of the potential risk to humans due to the presence of pesticide in banana is assessed through the estimation of the risk quotient (RQ) (Tables 4 and 5). The Chinese dietary structure, registered crops in China, and the corresponding MRLs recommended by various countries and organizations were taken into account to accurately assess the NEDI. If there was no STMRi, the corresponding MRLs would be adopted for risk assessment. The selection standards of residue limits followed the principle of China first, CAC second, United States third, Australia fourth, Korea fifth, EU sixth, and Japan last, and at the same time, the selection standards of residue limits needed to satisfy the value with the greatest risk. The STMR of kresoxim-methyl in banana at the PHI of 35 and 42 days were 0.22 mg kg⁻¹ and 0.10 mg kg⁻¹, respectively (Supplementary Table S2). The total NEDI was 0.6232 mg (PHI 35) and 0.6178 mg (PHI 42) in various food classifications, and the RQ was 2.47% and 2.45%. The results showed that the RQ values of kresoxim-methyl in banana were all below 100% of the recommended dosage after being applied three and four times, which indicated that the residual amounts of kresoxim-methyl in banana were not hazardous to people.