Optimization of an Analytical Method for Indoxacarb Residues in Fourteen Medicinal Herbs Using GC–μECD, GC–MS/MS and LC–MS/MS

Abstract

Pesticide residue analysis in medicinal herbs is a challenging task because of the matrix effect and its influence on quantitative analysis despite the continuous development of several new analytical methods and instrumentations. In this study, a modified QuEChERS method was developed for the analysis of indoxacarb residue in medicinal herbs by using the conventional instrument, gas chromatography micro-electron-capture-detector (GC–μECD), and comparing it with gas chromatography–tandem mass spectrometry (GC–MS/MS) and liquid chromatography–tandem mass spectrometry (LC–MS/MS). Samples were extracted with acetonitrile and purified using an NH₂ cartridge. The optimized method efficiently removes the co-extractives and offered a limit of quantification of 0.01 mg kg⁻¹. The GC–μECD analysis results of indoxacarb in seven medicinal herbs out of fourteen species at a fortification level of 0.01 mg kg⁻¹ showed a recovery range of 79.7–117.6%, while the rest showed recovery > 120%. Similarly, the recovery of indoxacarb by GC and LC–MS/SM were 74.1–105.9 and 73.0–99.0%, respectively, with a relative standard deviation of <20%. Matrix effects for the majority of medicinal herbs analyzed by GC–MS/MS were >±20%. Whereas the results for LC–MS/MS were <20%, which was within the acceptable range according to the SANTE/11312/2021 guidelines. Considering the performance of the method and alignment with the regulatory guidelines, LC–MS/MS is recommended for the analysis of indoxacarb in selected medicinal herbs.

1. Introduction

Korean medicinal herbs are widely used for the treatment and prevention of various diseases [1]. Previously these herbs were naturally collected; however, currently, they are artificially cultivated [2]. During cultivation, pests such as lepidopterans and sucking insects attack most of the medicinal herbs and cause damage [3]. Therefore, indoxacarb [CAS No. 144171-61-9] is a new oxadiazine insecticide registered against lepidopteran larvae and sucking insects on some medicinal herbs as well as various crops, fruits, and vegetables [4,5,6]. It mainly acts on the nervous system by blocking the voltage-dependent sodium channel in insect nerve cells resulting in paralysis and death of larvae [4,7]. The residues of indoxacarb are of substantial concern in terms of safety; thus, authorities must regulate the use of this potentially harmful insecticide in medicinal herbs. To analyze residues of indoxacarb in medicinal herbs, a multi-residue analytical method recommended by the Korean Ministry of Food and Drug Safety (MFDS) for food is used. However, one of the challenging tasks in that method is the matrix effect and its influence on quantitative analysis. The presence of numerous compounds with polarity close to those of analytes may hinder adequate clean-up without which there may be causing chromatographic interferences [8]. The task is more challenging for medicinal herbs due to the presence of various pigments as well as a large number of constituents [9]. Thus, sample preparation is a crucial step required before instrumental analysis in the development of a method for quantification purposes [10].

At present, conventional sample preparation methods for pesticide analysis in medicinal herbs mainly include pressurized liquid extraction, sorbent-based methods such as solid-phase extraction (SPE), supercritical fluid extraction, solid-phase microextraction techniques, matrix solid-phase dispersion, and QuEChERS (quick, easy, cheap, effective, rugged, and safe) [11,12,13,14,15]. Nowadays, the QuEChERS method developed by Anastassiades et al. [13] has become one of the most common methods for pesticide residue analysis in a variety of complex matrices including medicinal herbs due to its simplicity, low cost, the use of a smaller volume of organic solvents, high throughput and high efficiency [10]. It consists of two steps, first extraction based on partitioning of analytes between an aqueous and an organic layer via salting-out effect, and second SPE to further clean the extract using different sorbents, such as ammonia, silica, florisil, etc. to remove interfering substances [13]. Pesticide residue in various food and environmental samples are rapidly analyzed by several methods such as paper chromatography, colorimetric, surface-enhanced Raman spectroscopy (SERS), a highly sensitive immunoassay, molecular-imprinted polymer-grafted paper-based multi-disc micro-disc plate (MIP method), smartphone-based detection, paper-based visual detection and chronopotentiometry [16,17,18,19,20]. However, traditional identification and quantification of pesticides were carried out by gas chromatography (GC) with a variety of sensitive detectors such as micro electron capture detector (μECD), mass spectrometry (MS), and tandem mass spectrometry (MS/MS) [15,21]. Besides GC–MS/MS, which has some limitations, a perfect complement is high or ultra-high-pressure liquid chromatography (HPLC, UHPLC) coupled to MS/MS [22]. Despite the continuous development of several new analytical methods and instrumental techniques, one of the major problems in pesticide residue analysis is the matrix effect and its influence on qualitative and quantitative determination, particularly in the analysis of complex matrices [13]. Nevertheless, different methods were proposed to correct matrix effects, including extensive sample clean-up, the use of analyte protectants, internal standards, dilution, and the use of sophisticated instruments [13]. However, the shortcomings of these methods are that they are of high cost, labor-intensive, time-consuming, and still require performing matrix-matched calibrations. Therefore, an existing knowledge gap needs to be filled by developing specific methods with one extraction and one clean-up suitable for various medicinal herb matrices and detectors.

Thus, this work aimed to modify the QuEChERS method for the analysis of indoxacarb in the complex matrices of fourteen medicinal herbs by using the conventional instrument GC–μECD and comparing it with sophisticated instruments GC–MS/MS and LC–MS/MS. Further, matrix effects arising in GC–MS/MS and LC–MS/MS were evaluated by comparing solvent and matrix-matched calibration curves.

2. Materials and Methods

2.1. Chemical and Reagents

The analytical standard of indoxacarb (>98% purity) was purchased from Sigma Aldrich, Saint Louis, MO, USA, HPLC grade water acetonitrile, n-hexane, acetone, and dichloromethane were supplied by Thermo Fisher Scientific, Waltham, MA, USA. Sodium chloride (NaCl) was provided by Duksan Pure Chemicals Co., Ltd., Ansan, the Republic of Korea, and magnesium sulfate (MgSO₄) was purchased from Yakuri Pure Chemical Co., Ltd., Kyoto, Japan. SPE cartridges filled with aminopropyl and florisil were purchased from Phenomenex, Torrance, CA, USA. Stock solutions of indoxacarb were prepared in acetonitrile and stored at −20 °C before use.

2.2. Medicinal Herbs

A total of 14 raw medicinal herbs (Ostericum koreanum (Max.) Kitagawa, Ligusticum jeholense Nakai, Platycodon grandiflorum, Angelica gigas Nakai, Liriope muscari Decne, Cynanchum wilfordii Maxim, Angelica dahurica Fisch, Atractylodes ovata Thunb, Achyranthes japonica Nakai, Lithospermum erythrorhizon Siebold & Zucc, Rehmannia glutinosa Libosch, Cnidium officinale Makino, Phaseolus coccineus, and Radix astragali) were purchased from online distribution store in the Republic of Korea. Each sample was pulverized, sieved through a 2.0 mm sieve to obtain homogenized samples, and stored at −20 °C before analysis. These samples were then used as control, spiked samples for recovery, and matrix-matched calibration.

We confirm all medicinal herbs used in the current work comply with the IUCN Policy Statement on Research Involving Species at Risk of Extinction and the Convention on the Trade in Endangered Species of Wild Fauna and Flora.

2.3. Sample Preparation

Extraction was carried out following a modified QuEChERS multi-residue method described by Anastassiades et al. [13]. Briefly, 4 mL water was added to 2.0 g of homogenized samples and equilibrated for 30 min at room temperature. Samples were then extracted with 10 mL acetonitrile by shaking for 1 min. Subsequently, 1.0 g MgSO₄ and 1.5 g of NaCl were added and shaken for 1 min followed by centrifugation at 4000 rpm for 10 min. The amount of 2.5 mL supernatant was transferred to another tube and concentrated to dryness under a vacuum rotary evaporator at 40 °C and re-dissolved in 5 mL of acetone/n-hexane (10/90, v/v) mixture (Figure S1).

2.4. SPE Clean-Up

Clean-up was carried out by using the NH₂ SPE cartridge. The NH₂ SPE cartridge was conditioned by pre-washing with 5 mL of n-hexane. The extracted sample was passed through the cartridge and the cartridge was washed with a 10 mL dichloromethane/n-hexane (10/90, v/v) mixture. Finally, the analyte was eluted with a 10 mL dichloromethane/n-hexane (40/60, v/v) mixture. The collected elute was concentrated to dryness under a vacuum rotary evaporator at 40 °C and redissolved in 1 mL acetonitrile and transferred to a vial for instrumental analysis.

2.5. Instrumentation

2.5.1. GC–μECD Analysis

Indoxacarb was analyzed by using gas chromatograph Agilent model 7890A (GC) equipped with a micro-electron capture detector (μECD). A DB-5 capillary column J&W 122-5032 (30 m × 250 µm, 0.25 µm) was used for separation. The GC-μECD parameters are provided in Supplementary Materials (Table S1). A sample volume of 2.0 microliter was injected into the instrument in split mode (5:1) using an autosampler. The inlet port and detector temperature were set at 260 and 310 °C, respectively. The oven temperature was initially maintained at 90 °C for 1 min, increased to 220 °C at a rate of 25 °C min⁻¹ and held for 2 min, and then finally increased to 300 °C at 25 °C min⁻¹, and was held for 10 min. Helium was used as a carrier gas with a flow rate of 1.2 mL min⁻¹, while nitrogen (99.99% purity) was used as a make-up gas at a flow rate of 59.8 mL min⁻¹.

2.5.2. GC–MS/MS Analysis

The separation of indoxacarb was performed by DB-5MS 122-5532UI (30 m × 250 µm, 0.25 µm) with a gas chromatographic system coupled to tandem mass spectrometry (GC–MS/MS-TQ8050; Shimadzu Corporation, Kyoto, Japan). The GC–MS/MS parameters and selected product ions are shown in Table S2. A sample volume of 2.0 microliter was injected into the instrument in splitless mode using an autosampler. The injection port temperature was 280 °C. The oven temperature programmed for peak separation was as follows: initially maintained at 80 °C for 1 min, increased to 220 °C at a rate of 25 °C min⁻¹ and held for 2 min, and then finally increased to 300 °C at 25 °C min⁻¹, and was held for 10 min. Helium was used as a carrier gas with a flow rate of 1.2 mL min⁻¹. For detection of indoxacarb, an MS/MS system equipped with electron ionization (EI) was used. The MS/MS parameters were as follows: MS transfer line temperatures: 280 °C; ionization source temperature: 250 °C; EI energy: 70 eV and multiple reaction monitoring mode (MRM) was used.

2.5.3. LC–MS/MS Analysis

Quantitative determination of indoxacarb was carried out by using a UHPLC system (Dionex Ultimate 3000, Thermo Fisher Scientific, Waltham, MA, USA) coupled with tandem mass spectrometry (MS/MS) (TSQ quantum access max, Thermo Science, Waltham, MA, USA). The UHPLC–MS/MS parameters and selected product ions are shown in Table S3. Chromatographic separation was performed by reversed-phase C18 column (Imtak, 100 mm × 3.0 μm, 2.0 mm). The sample injection volume was 1.0 µL. Water (solvent A) and methanol (solvent B) containing 0.1% formic acid and 5 mM ammonium format were used as mobile phases. The gradient elution started at 5% of (B), increased to 95% at 1.5 min, and was held for 4.5 min, following this mobile phase, the composition was set-back to initial conditions and maintained for 1.5 min for column equilibration. An MS/MS system (TSQ quantum ultra, Thermo Science, Waltham, MA, USA) equipped with an electrospray ionization source operating in positive mode (ESI+) was used. The MS parameters were as follows: spray voltage: 3500 V; sheath gas flow: 20 AU; capillary temperature: 200 °C and auxiliary gas flow rate: 10 AU and vaporization temperature: 220 °C.

2.6. Method Validation

The analytical method was validated in terms of different performance criteria such as linearity, accuracy and precision, limit of detection (LOD), limit of quantitation (MLOQ), and matrix effect (ME). The linearity of the calibration curve was evaluated by the values of the correlation coefficient (R²). The accuracy and precision were obtained in terms of recovery (70–120%) and repeatability (n = 3). The recoveries were determined by adding the analyte at the concentration level of 0.01 mg L⁻¹ in the control sample and processed by the method described earlier in Section 2.3 and Section 2.4. Using GC–μECD the recoveries were calculated by calibration curve (ranged: 0.02 mg kg⁻¹ to 1.0 mg kg⁻¹) while using GC–MS/MS and LC–MS/MS the recoveries were calculated by comparing the response of the analyte in samples with response in an equivalent amount of standard solutions prepared in a matrix. The repeatability expressed as the relative standard deviation (RSD) of the analyzed samples was calculated from three repetitions. LOD was evaluated by injecting the indoxacarb standard at different concentrations and the concentration at which the peaks could be identified at a signal-to-noise ratio of >3 was considered LOD. The LOQ was calculated by Equation (1) taking into consideration the following factors: the instrument limit of detection, volume of extraction solvent, injection volume, dilution factor, and sample amount [23,24,25].

$$\mathrm{A}\left(\mathrm{ng}\right)\times \frac{\mathrm{B}\left(\mathrm{mL}\right)}{\mathrm{C}\left(\mathsf{\mu }\mathrm{L}\right)}\times \frac{\mathrm{D}\left(\mathrm{mL}\right)}{\mathrm{E}\left(\mathrm{mL}\right)}=\mathrm{LOQ}\left(\mathrm{mg}·{\mathrm{kg}}^{-1}\right)$$

(1)

where A: instrument detection limit; B: extraction solvent volume; C: injection volume; D: dilution factor; and E: sample amount.

2.7. Matrix Effect

Matrix effect (ME) occurs due to the presence of potentially interfering substances in the sample matrix, which affects the response of the analytes. The ME in this study was evaluated by making 3-point calibration curves at concentrations levels of 0.01, 0.05, and 0.1 mg kg⁻¹ in medicinal herbs matrix and in pure solvent. The ME was then calculated by comparing the slope of the calibration curves prepared in medicinal herb matrix and in pure solvent using Equation (2). Negative values of the matrix effect indicate suppression of the signal, and positive values indicate enhancement. To better understand the results, the values were categorized into three groups: matrix effect will be considered soft when the values are <±20%; whereas it will be considered medium when the values are >±20 and <±50%, similarly, MEs will be considered strong when the values are >±50% [26].

$$\mathrm{ME}(\%)=\left(\frac{\mathrm{Slope} \mathrm{in} \mathrm{matrix}}{\mathrm{Slope} \mathrm{in} \mathrm{solvent}}-1\right)\times 100$$

(2)

3. Results and Discussion

3.1. Selection of Extraction Method

In this study, the modification of the QuEChERS method was essential since the original method is intended for samples with a moisture content of more than 75% and a sample amount of 10 g [22]. Due to the presence of a large number of constituents in medicinal herbs, extraction steps are necessary to minimize or eliminate the matrix effect. One way to minimize the matrix effect is to use a small amount of the sample; therefore, a weight of 2 g of the dry herb was used. Medicinal herb samples belong to matrices with low moisture content, adding water to hydrate the samples before extraction is a key to achieving better recoveries of pesticides [22]. Another important parameter to optimize is the choice of extracting solvent. In the extraction of pesticides from a solid sample, the extracting solvent must have a good dissolving ability for pesticides and great permeability into the matrix [16]. Among the various solvents used in the QuEChERS extraction method, acetonitrile was selected because of its potential for better recoveries, less co-extracted components of the matrix, and thus less chromatographic interference during GC and LC analysis [27,28].

3.2. Optimization of SPE Condition

Routine analysis of pesticide residues in complex medicinal herb matrices requires suitable sorbents for adequate clean-up, which is critical for the removal of co-extracted substances. Solid-phase extraction (SPE) in a multi-residue analysis of pesticides has the advantages of simplification of the clean-up process, the ability for automation, and reduction of organic solvents [29]. Various types of SPE adsorbent materials (such as silica gel, alumina, and florisil) are commercially available and were applied to the multi-residue analysis. In an initial attempt to clean the extracts, the clean-up process previously described by the Korean Ministry of food and drug safety was applied using florisil SPE cartridge for removing a large number of co-extractive substances present in medicinal herbs. However, a huge amount of co-extractive interfering peaks were observed in the chromatograms, hence, it was not possible to effectively remove the interference using a florisil SPE cartridge. Therefore, it was found necessary to investigate the cleaning efficiency of the NH₂ SPE cartridge, which is a weak anion exchange cartridge used for the removal of a wide variety of coextracted compounds [30]. Fortunately, in the preliminary experiment, it was observed that no significant peaks could interfere with target analytes and good recoveries were obtained when the NH₂ cartridge was used. To eliminate the co-extractives effectively and achieve complete elution of pesticide, the elution efficiency of dichloromethane and n-hexane mixture with different ratios were tried and the final clean-up procedure involved pre-washing of the NH₂ cartridge with 5 mL n-hexane, loading of the extracted sample, successive washing of the NH₂ cartridge with 10 mL dichloromethane/n-hexane 10:90 (v/v) to remove co-extractives, and elution of the target pesticide with 10 mL dichloromethane/n-hexane 40:60 (v/v). This procedure attained efficient sample clean-up and resulted in satisfactory recovery.

3.3. Method Validation

The optimized analytical method for indoxacarb in 14 types of medicinal herbs using GC–μECD, GC–MS/MS, and LC–MS/MS was evaluated. Different parameters such as linearity (R²), limit of quantification (LOQ) recovery, and precision (expressed as RSD) were the characteristics of the developed method.

3.3.1. GC–μECD Analysis

Multi-residue analysis of pesticides in complex matrices is commonly carried out by chromatography coupled with tandem mass spectrometry [31]. The cost of such instruments for quantitative analysis is considered unaffordable for some laboratory’s budgets [22]. Therefore, the optimized method in this study was first evaluated by GC–μECD, which is a conventional instrument and easily available in common analytical laboratories. Furthermore, GC–μECD is the most common detector used for the analysis of halogenated compounds, due to its selectivity toward these compounds. In this study, the GC–μECD method allowed a good detection of indoxacarb in medicinal herbs. The results in Figure S2 showed that the developed method provided a clean baseline and there were no interfering peaks found around the expected retention time of the target compound in spiked and unspiked samples, showing that the method was specific. The analyzed parameters indicated excellent linearity of the 7-point calibration curve (0.02 mg kg⁻¹ to 1.0 mg kg⁻¹) with a regression coefficient (R²) of 0.99. The sensitivity of the method was obtained from the slope of the calibration curve using the regression equation (Table S4). A better slope in the calibration curve showed better sensitivity of the method. The LOD and LOQ were 0.005 and 0.01 mg kg⁻¹, respectively. The average recovery of indoxacarb in 7 medicinal herbs out of 14 species at a fortification level of 0.01 mg/kg⁻¹ was in the range of 79.7 to 117.6% with a relative standard deviation (RSD) values less than 20% (

Table 1. Recovery (%) of indoxacarb in 7 medicinal herbs analyzed by GC–μECD.

Medicinal Herbs	Average Recovery (±SD *)
P. coccineus	89.4 ± 3.51
Radix astragali	97.8 ± 12.32
A. japonica	85.6 ± 3.88
P. grandiflorum	107.4 ± 9.70
L. muscari	114.6 ± 3.54
C. wilfordii	79.7 ± 8.60
R. glutinosa	117.6 ± 3.08

* SD = standard deviation.

and Table S5). These results are in agreement with SANTE/11312/2021 [32]. The SANTE guidelines stated that recovery between 70.0 and 120.0% is considered acceptable for pesticide residue analysis with an RSD below 20.0%. In contrast, the rest of the medicinal herbs (seven species) showed unsatisfactory recovery (>120%) that could not satisfy the SANTE guideline’s acceptable criterion. Therefore, another sophisticated instrumental method for the analysis of herbal medicines species was sought; however, GC–MS/MS was selected for this purpose.

3.3.2. GC–MS/MS Analysis

Since GC–μECD is only suitable for seven species of medicinal herbs under the study, it was necessary to search for an appropriate instrument for indoxacarb analysis in medicinal herbs with good recoveries and low matrix effect. Recently, advanced analytical tools have been used for the analysis of chemicals in complex matrices [33]. Among these tools, GC–MS/MS is increasingly getting attention due to its high resolution, good sensitivity, and low detection limits [34]. Therefore, the currently developed method was further evaluated by GC–MS/MS. Using GC–MS/MS, the developed method showed a clean baseline with no interfering peaks around the expected retention time of indoxacarb in spiked and control samples (Figure S3). The LOD and LOQ were 0.005 mg kg⁻¹ and 0.01 mg kg⁻¹, respectively. The LOQ of the developed method satisfies the MFDS guidelines for pesticide residue analysis which indicated that LOQ should be less than 0.05 mg kg⁻¹ or below the maximum residual limit [24]. The recoveries were calculated by comparing the peak area of indoxacarb in the sample extract with the equivalent amount of the standard solution prepared in the matrix. The recovery of indoxacarb at a fortification level of 0.01 mg kg⁻¹ was in an acceptable range of 74.1 to 105.9% with RSD below 10% (

Table 2. Recovery (%) of indoxacarb in 14 medicinal herbs analyzed by GC–MS/MS.

Medicinal Herbs	Average Recovery (±SD *)
R. glutinosa	85.2 ± 5.6
A. ovata	83.0 ± 4.0
L. jeholense Nakai	105.9 ± 6.3
O. koreanum Kitagawa	81.5 ± 2.7
A. dahurica	86.5 ± 0.6
C. wilfordii	86.5 ± 1.0
L. erythrorhizon	85.7 ± 2.4
P. coccineus	82.6 ± 2.9
Radix astragali	100.4 ± 4.0
A. japonica	74.1 ± 2.1
C. officinale Makino	86.4 ± 5.9
P. grandiflorum	79.2 ± 1.8
L. muscari	81.2 ± 3.3
A. gigas Nakai	85.8 ± 1.8

* SD = Standard deviation.

and Table S6). Overall, the accuracy and precision of the method fulfill all the requirements of SANTE/11312/2021; however, due to the high matrix effect (>±20%), it is recommended that GC–MS/MS is not a suitable instrument for the analysis of indoxacarb in selected medicinal herbs. Therefore, an alternative instrument was suggested by selecting LC–MS/MS.

3.3.3. LC–MS/MS Analysis

After GC–μECD and GC–MS/MS analysis, the optimized method was evaluated by LC–MS/MS. As a result, using LC–MS/MS the chromatograms (Figure 1 and Figure S4) showed no interfering peaks present at the same retention time of indoxacarb for all transition ions in the LC–MS/MS system, suggesting that the extraction, clean-up, and instrument conditions were adequate, and the method was selective. The LOD and LOQ were 0.005 mg kg⁻¹/kg and 0.01 mg kg⁻¹, respectively. The recoveries were calculated by the method mentioned in Section 3.3.2. The average recoveries of indoxacarb from medicinal herbs at a fortification level of 0.01 mg kg⁻¹ were in an acceptable range of 73.0 to 99.0% with an RSD below 10% (

Table 3. Recovery (%) of indoxacarb in 14 medicinal herbs analyzed by LC–MS/MS.

Medicinal Herbs	Average Recovery (±SD *)
R. glutinosa	99.0 ± 3.7
A. ovata	73.0 ± 2.5
L. jeholense Nakai	91.3 ± 2.5
O. koreanum Kitagawa	75.8 ± 2.2
A. dahurica	75.2 ± 1.6
C. wilfordii	90.0 ± 13.9
L. erythrorhizon	94.9 ± 2.9
P. coccineus	84.2 ± 4.3
Radix astragali	93.7 ± 3.2
A. japonica	84.5 ± 6.4
C. officinale Makino	96.0 ± 4.4
P. grandiflorum	75.5 ± 4.8
L. muscari	88.1 ± 2.0
A. gigas Nakai	80.5 ± 5.7

* SD = Standard deviation.

and Table S7). The validation procedure showed that the developed method is selective, accurate, and precise, giving recoveries of 70–120% with RSD ≤ 10%. Therefore, it is recommended that LC–MS/MS is a suitable instrument for the analysis of indoxacarb in medicinal herbs.

3.3.4. Matrix Effect

The presence of co-extractive components in complex matrices such as medicinal herbs is one of the major problems in pesticide residue analysis because it can suppress or enhance the chromatographic signals resulting in uncertainty in quantitative results [22]. As described earlier in Section 2.2, a total of 14 medicinal herbs were selected in this study. To evaluate the matrix effect, three concentration levels (0.01, 0.05, and 0.1 mg kg⁻¹) of indoxacarb in triplicate were analyzed in each of the chosen matrices by GC–MS/MS and LC–MS/MS. Excellent linearity was obtained for all the matrices (R² > 0.99). The method was sensitive as indicated by the slope of the calibration curve (Table S8). The matrix effect was obtained by comparing the slope of the three-point calibration curves prepared in each medicinal herb matrix and in pure solvent. The matrix effect can be positive when the slope of the calibration curve in the matrix is greater than in pure solvent. It can be negative when the slope of the calibration curve in the matrix is smaller than the calibration curve in the pure solvent [35]. The results of the matrix effect determined by optimized QuEChERS extraction using GC–MS/MS and LC–MS/MS detection are presented in Figure 2 and Table S9.

In the case of the GC–MS/MS analysis study, the matrix effect of three medicinal herbs for indoxacarb was lower than ±20% indicating a soft matrix effect while the matrix effect of eleven medicinal herbs was higher than ±20% showing a medium effect. None of the medicinal herbs showed a strong matrix effect. In GC–MS/MS, signal enhancement was commonly observed as it is typical for the GC to induce enhancement resulting from blocking the active sites of the inlet and the column by matrix components, leading to an increase in the transfer of the analyte to the detector [16]. Applying LC–MS/MS analysis, the matrix effect of twelve medicinal herbs for indoxacarb was lower than ±20%, indicating a soft matrix effect while the matrix effect of the remaining two medicinal herbs was higher than ±20% indicating a medium effect. Similar to GC–MS/MS, no strong effect was observed in any medicinal herb. Generally, in LC–MS/MS detection, matrix effect values were found smaller than those of GC–MS/MS. In LC–MS/MS, signal suppression was prevalent and appeared in eight matrices. Such suppression phenomenon in LC–MS/MS mainly originates from the competition between matrix components and analytes to obtain available protons as the amount of proton is limited [36]. However, the slight enhancement effect in six matrices is contrary to our expectations, and the possible reason for this is still unknown and there is no clear hypothesis explaining the cases of few signal enhancement [37]. The matrix effect evaluated by GC–MS/MS and LC–MS/MS was varied; in GC, the effect was observed higher than LC–MS/MS, which suggests that the matrix effect not only depends on the analyte and matrix but also on instrumental techniques. The obtained result indicated that for the majority of medicinal herbs analyzed by LC–MS/MS, the matrix effect was not significant (<20%), which was within the acceptable range according to the SANTE guideline. Therefore, this study recommends LC–MS/MS for the analysis of indoxacarb in selected medicinal herbs.

4. Conclusions

A modified QuEChERS method was developed for the analysis of indoxacarb in fourteen medicinal herbs using GC–μECD, GC–MS/MS, and LC–MS/MS. The GC–μECD analysis results indicated good validation parameters of the method for seven medicinal herbs out of fourteen while the rest (seven species) could not satisfy the acceptance criteria of the SANTE/11312/2021 guidelines. However, in the case of GC–MS/MS and LC–MS/SM, the validation parameters showed that the developed method is selective, accurate, and precise, giving recoveries of 70–120%, RSD ≤ 10%, and LOQ values of 0.01 mg kg⁻¹. The matrix effect results for the majority of medicinal herbs analyzed by GC–MS/MS exhibited matrix effects higher than ±20%, which could not satisfy the SANTE guideline’s acceptance criteria. Whereas the matrix effect analyzed by LC–MS/MS for the majority of medicinal herbs was <20%, which was within the acceptable range according to the SANTE guidelines. Based on the findings seven species of medicinal herbs (P. coccineus, Radix astragali, A. japonica, P. grandiflorum, L. muscari, C. wilfordii and R. glutinosa) can be analyzed by GC–μECD. All tested medicinal herbs (R. glutinosa, A. ovata, L. jeholense Nakai, O. koreanum Kitagawa, A. dahurica, C. wilfordii, L. erythrorhizon, P. coccineus, Radix astragali, A. japonica, C. officinale Makino, P. grandiflorum, L. muscari and A. gigas Nakai) can be analyzed with either GC–MS/MS or LC–MS/MS. However, GC–MS/MS exhibits a matrix effect, while in the LC–MS/MS no matrix effect was observed. Therefore, this study recommends LC–MS/MS for the analysis of indoxacarb in selected medicinal herbs.