Determination of Glyphosate, Glufosinate, and Their Major Metabolites in Tea Infusions by Dual-Channel Capillary Electrophoresis following Solid-Phase Extraction

In this study, two analytical procedures were developed and validated using dual-channel capillary electrophoresis-coupled contactless conductivity detection (CE-C4D) followed by solid-phase extraction (SPE) for simultaneous determination of glyphosate (GLYP), glufosinate (GLUF), and their two major metabolites, aminomethylphosphonic acid (AMPA) and 3-(methylphosphinico) propionic acid (MPPA), respectively, in a popular beverage such as tea infusions. GLYP, GLUF, and AMPA were analyzed in the first channel using background electrolyte (BGE) of 1 mM histidine (His) adjusted to pH 2.75 by acetic acid (Ace). In contrast, MPPA was quantified in the second channel with a BGE of 30 mM His adjusted to pH 6.7 by 3-(N-morpholino) propanesulfonic acid (MOPS) and 10 µM of cetyltrimethylammonium bromide (CTAB). In addition, the samples of tea infusions were treated using SPE with 10 mL of 0.5 mM HCl in methanol as eluent. At the optimized conditions, the method detection limit (MDL) of GLYP, GLUF, AMPA, and MPPA is 0.80, 1.56, 0.56, and 0.54 μg/l, respectively. The methods were then applied to analyze four target compounds in 16 samples of tea infusions. GLYP was found in two infusion samples of oolong tea with concentrations ranging from 5.34 to 10.74 µg/L, and GLUF was recognized in three samples of green tea infusion in the range of 45.1–53.9 µg/L.


Introduction
Since its commercialization by Monsanto Company (USA) under the name Roundup in 1974, glyphosate (GLYP) has become the world's most commonly used herbicide for broadleaf weed control and as a plant growth regulator [1]. Besides GLYP, glufosinate (GLUF), a similar phosphorus herbicide, is also widely used because of its broad spectrum of action, including against GLYP-resistant weeds [2]. On the other hand, according to the World Health Organization (WHO), glyphosate has potential carcinogenic effects [3].
Furthermore, acute glufosinate exposure can cause toxicities for the central nervous and respiratory systems [4].
is increases the risk of human exposure to these herbicides and raises concerns about human health and environmental hazards. As a result, the European Union has stringent regulations regarding the existence of pesticides in the water meant for human consumption, such as a maximum concentration of 0.1 µg/L for each pesticide and a total concentration of less than 0.5 µg/L [17].
Tea is one of the most popular beverages in the world after water, and it is brewed by steeping tea leaves/buds in hot/boiling water to make tea infusions. e tea is cultivated in old orchards, which provide an ideal setting for weed development [18], which is progressively controlled by herbicides like GLYP or GLUF. Consequently, there is increasing concern that tea infusions may be a source of herbicide residues' exposure to humans. erefore, it is essential to determine the concentration of GLYP and GLUF and their primary metabolites aminomethylphosphonic acid (AMPA) and 3-(methylphosphinico)propionic acid (MPPA) in tea infusions [19].
It is challenging to simultaneously analyze GLYP, GLUF, and their metabolites in aqueous matrices due to their characteristics of highly polar, lack of chromogenic and fluorescent groups. e methods of liquid chromatography (LC) [20][21][22][23][24][25][26], high-performance liquid chromatography (HPLC) [19], and ultra-high performance liquid chromatography (UHPLC) [27][28][29][30] coupled to tandem mass spectrometry (MS/MS) are commonly used methods, although these methods usually need the derivatization. On the other hand, the derivatization step can be omitted when ion chromatography coupled to tandem mass spectrometry (IC-MS/MS) is used [31]. However, using LC, GC, or IC-MS/MS increases analytical costs and requires skilled technicians. In contrast, capillary electrophoresis using contactless conductivity detection (CE-C 4 D) is likely the more appropriate method with its advantages of simple and low cost [32].
Although some research [33][34][35][36][37][38] has used the CE technique to determine GLYP residues with GLUF, AMPA, and MPPA in environmental and drinking water samples, the use of C 4 D detection is few and lacks updates. Hong Heng See et al. used the CE-C 4 D method and electrokinetic injection technique to analyze those substances in drinking water with a linear range of 0.01-0.1 µM [39,40]. However, the analytical procedures in these works have only been illustrated for the relatively simple matrix of tap water and are difficult to apply to more complex samples. Another study used the software to determine thermodynamic acidity constants and limit ionic mobilities of glyphosate and its metabolites and application in the separation of glyphosate [41]. Regardless, GLUF and MPPA were not the targets of that work. erefore, to fill the gap, this study aimed to develop the application of the dual-channel CE-C 4 D combined with solid-phase extraction (SPE) to simultaneously determine GLYP, GLUF, AMPA, and MPPA in tea infusions. A new approach using a two-channel CE system to simultaneously analyze these compounds is in only one run. In addition, by analyzing on two different channels, the process of optimization was simpler.
A purpose-made CE instrument with two channels independently working was used for all experiments. e setup and operation of the CE system were demonstrated in previous publications [42][43][44]. e capillaries used in this system were fused silica obtained from BGB Analytik AG (Böckten, Switzerland) with 50 µm ID and 365 µm OD. Before use, the capillaries were preconditioned with 0.1 M NaOH for 10 min and deionized water for 10 min prior to flushing with the BGE solutions.
e SPE system from Supelco (USA) was used for SPE producers using six mL Oasis WAX cartridges (containing 500 mg sorbents). In addition, a vacuum evaporator system of BUCHI (Switzerland) and MGS-2200 nitrogen sample concentrator (Eyela, Japan) were used to concentrate the SPE extract.

Sample Collection and Sample Treatment.
Sixteen samples of tea products were collected from the typical supermarkets in Hanoi in the last quarter of 2021. e sample preparation was adapted from Phan i, LA. et al. (2020) [45], and in particular, 1.0 gram of dry tea leaves of each sample was weighed and transferred to a glass flask, and then, 100 mL boiled water was added. e flask was kept for 10 min, and then, the aqueous solution was filtered through a PTFE filter with 0.45 µm size.
An Oasis WAX cartridge was preconditioned with 5 mL of MeOH followed by 5 mL of deionized water. en, 100 mL of tea infusion sample was loaded through at a 1 mL/ min flow rate. Next, the SPE cartridge was washed with 5 mL of deionized water and 5 mL of methanol and then dried for 10 min. Next, the target compounds in the cartridge were eluted with 10 mL of 0.5 M HCl in MeOH at a rate of 0.5 mL/ min. Finally, the extract was dried under a gentle nitrogen stream at 60 C, and the obtained solid residue was dissolved in 1 mL of deionized water. e final solution was filtered through a 0.2 μm nylon membrane filter before CE-C 4 D analysis.

Analytical Procedures.
In this study, GLYP, GLUF, and AMPA were simultaneously determined on the first channel of the dual CE-C 4 D system using the BGE composed of 1 mM His adjusted to pH 2.75 by Ace and a voltage of +20 kV from the detection end of the capillary. e second channel was used to analyze MPPA with the optimized BGE of 30 mM His/MOPS (pH � 6.7) and then added 10 µM of CTAB and an applied voltage of +20 kV. Both capillaries of the two channels have the exact total and effective lengths of 70 and 62 cm, respectively.

Development of the CE-C4D
Methodology for Determination of GLYP, GLUF, and MPPA. In the selection of BGE, the primary consideration is the ionization characteristic of the analytes. Based on the acid-base property of the four target substances, GLYP, GLUF, and AMPA can exist as the cation (pH < pK a1 , the -NH 2 group is protonated to becomes -NH 3 + ) or anion (pH > pK a2 , two acidic O-H groups are deprotonated), while MPPA can only be analyzed as anion (pH > pK a ) due to the lack of an amino group. Besides, the BGEs composed of low His concentrations at acidic pHs have been shown to be good separation efficient of phosphate ions [46]. Accordingly, BGEs composed of 4 mM His adjusted to pH 3.0 by different commonly used acids (Ace, Cit, For, Lac, and Suc) were investigated ( Figure 1).
As shown in Figure 1(a), AMPA signals in all conditions were minimal because this compound mainly existed in the form of a zwitterion at pH 3.0, resulting in a total charge of approximately zero. erefore, AMPA should be surveyed with the different BGE conditions at a higher pH on another channel while three remaining analytes were separated in one capillary. e BGE with the composition of His/Ace produced a stable baseline, higher peaks, and better separation for GLYP, GLUF, and MPPA, and as a significant result, this buffer was chosen for further modification of BGE concentrations and pH, as it was previously proven that the signals in CE-C 4 D may also be dependent on these parameters. Figure 1(b) illustrates the effect of various His concentrations ranging from 0 to 6 mM on the separation performance of GLYP, GLUF, and MPPA. As it can be seen, an increase in BGE concentration was associated with a more extended separation time due to the increase in the velocity of solutions at higher concentrations. In His absence, the acquired signals were markedly reduced, possibly due to a decrease in background conductivity. With 1 mM of His, the highest resolutions and peak areas of GLYP, GLUF, and MPPA were obtained. Moreover, Figure 1(c) represents the result of varying the pH from 2.50 to 3.50 on the separation performance between GLYP, GLUF, and MPPA. When the pH of BGE decreased, the migration time and separation performance increased due to the increasing velocity, as mentioned. e best resolutions were achieved at the condition of pH 2.50, but at the same time, the peak heights were the worst for all analytes. erefore, the optimum condition was selected with a pH of 2.75.

Development of the CE-C4D
Methodology for Determination of AMPA. As mentioned, for analysis of AMPA (with pKa values of 1.8; 5.4; and 10) on the second channel, the selected BGE solution must have a pH higher than pK a2 to ensure that most of these analytes exist in anion form. A study by Hong et al. showed that the pH range of 5.8-6.8 was suitable for AMPA detection [40]. However, in that work, the used His/MES buffer needs to be added methanol at a concentration of 3% (v/v) to separate AMPA and GLUF. erefore, BGEs composed of His/MOPS were used to provide the higher separation in this study. In addition, the separation of an anion at high pH is strongly affected by the electroosmotic flow (EOF). Consequently, CTAB was added to BGE solutions to reduce the hindrance of EOF. At the initial condition, the pH and CTAB concentration in buffer solution were fixed at 6.5 and 20 μM, respectively. e impact on AMPA detection performance of His amounts ranging from 6 to 54 mM was examined. As is shown in Figure 2(a), increasing the His portion resulted in a longer analysis time, and the best shape and highest signal of AMPA were obtained at the concentration of 30 mM. e pH of 30 mM His/MOPS buffers was then explored between 6 and 7, agreeing with previous studies. e results in Figure 2(b) indicated that a higher pH caused a longer migration time.
is observation is attributable to the strength of the EOF, which has a more suppressive effect on AMPA migration at greater pH. In comparison, the decrease in MOPS concentration due to the drop in pH yielded a reduction in background conductivity but an enhancement in height peaks. Additionally, the pH 6.7 condition produced a better signal at an appropriate analysis period and was chosen for AMPA determination. Finally, the effect of CTAB concentration, the EOF modifier, was investigated in the range of 0-75 µM. As illustrated in Figure 2(c), when there was no CTAB, the analysis time was extremely long (1500 seconds), owing to the EOF's influence. e migration times lowered, and the AMPA signals decreased as CTAB concentrations increased. As a result, a CTAB concentration of 10 µM was chosen as the optimum condition.

Method Optimization for Solid-Phase Extraction.
Due to their acid-base properties, GLYP, GLUF, AMPA, and MPPA could be dissociated into anions at the pH of tea infusions (5.30-6.03) [47], so it was suitable for using ion-exchange SPE columns, such as WAX columns. erefore, in this study, the analytes were retained on the WAX cartridges as anions by ion-pair interaction with the sorbent of ammonium ion and eluted by an acidic eluent, HCl/MeOH. Two factors were explored for optimizing the SPE technique for sample enrichment: the HCl concentration and the eluent volume. e efficiencies of the SPE process were calculated as the ratio of the measured concentrations to the initial spiked concentration (10 µg/L).
First, the effect of HCl concentration was researched between 0.1 and 1 mM (see Figure 3(a)). It was discovered that raising the concentration of HCl to 0.5 M raised the extraction effectiveness of all target chemicals and then declined at higher concentrations. ese findings substantiated the choice of the 0.5 M HCl/MeOH as optimum eluent. Subsequently, the HCl/MeOH volume was studied between 5 and 20 mL to determine the extraction efficiency. As shown in Figure 3(b), the extraction efficiency of the target compounds rose as the solvent volume increased from 5 to 10 mL and remained almost constant above 10 mL. As a result, the SPE operations were performed using a volume of 10 mL of 0.5 M HCl/MeOH. At the discovered optimal conditions, the SPE recoveries of GLYP, GLUF, AMPA, and MPPA were 91.5, 87.2, 92.5, and 92.7%, respectively, and the enrichment factors were 100 for all analytes. Table 1 summarizes the salient performance data for the determinations of GLYP, GLUF, AMPA, and MPPA with the developed CE-C 4 D methodologies.

Method Validation.
e method detection limit (MDL) and method quantitative limit (MQL) for all target chemicals in the tea infusion sample were at the ppb level within the acceptable range of 0.54-5.19 µg/L. e comparison of results with other methods is shown in Table 2, and it can be observed that, though higher than in chromatographic methods, the MDL values found in this study were similar to those of previous studies using the CE technique. e linear ranges were from MQLs to 400, 200, 200, and 1,600 µg/L for GLYP, GLUF, MPPA, and AMPA, respectively. e calibration curve of each analyte was established in the range of 5.0-100 µg/L with good linearities of R 2 > 0.998. According to regulation standards [48], the maximum allowable limit   for GLYP in drinking water is 0.1 µg/L and in dry tea leaves is 1 mg/kg, which corresponds to 10 µg/L in tea infusion using the developed analytical procedure of brewing 1.0 g of tea leaves in 100 mL. us, the MDL and MQL for GLYP in this study indicated that although the developed analytical method could not meet the EU standard for drinking water, it met the objectives for detecting GLYP residues in tea leaves. e intra-assay precision was assessed by comparing the relative standard derivation (RSD) values for peak area and migration time measurements (n � 11) taken within one day under optimum conditions and with a standard solution of 2 mg/L for each analyte. e interassay precision was measured using the same solution but over a seven-day period (n � 7). e recovery was determined using the standard addition method with the concentrations of 5, 10, and 20 µg/L spiked in a real matrix of the TD2 sample. e RSD values of migration time and peak area of each analyte were less than 7% and 9% in the intra-and interday precisions, respectively. e majority of obtained values for repeatability and reproducibility met the AOAC accuracy criterion [49]. e recoveries of four target compounds ranged between 80.6 and 99.6%, indicating that the procedure produced dependable results with high precisions. Table 3 and Figure 4, three tea infusions prepared from green teas (TX1, TX3, and TX4) were detected with GLUF in the range of 45.1-53.9 µg/L, while there were two samples of oolong tea infusions (TOL3 and TOL4) detected with GLYP at concentrations of 5.34-10.74 µg/L. AMPA and MPPA were not detected in all tea infusion samples. Although the CE-C 4 D methodology can be used to screen for GLYP or  GLUF residues and control tea product quality, specific analytical techniques are usually required for further validation before such mandatory reporting is performed.

Conclusions
Two analytical procedures using a dual-channel CE-C 4 D system effectively developed a convenient and cost-effective approach for quantifying GLYP, GLUF, AMPA, and MPPA in tea infusions. e targeted analytes were extracted and concentrated by SPE and then were analyzed on two separate CE channels. e MDLs for GLYP, GLUF, AMPA, and MPPA were 0.80, 1.56, 0.56, and 0.54 μg/l, respectively. GLYP was found in two infusion samples of oolong tea with concentrations ranging from 5.34 to 10.74 µg/L, and GLUF was recognized in three samples of green tea infusion in a total of 16 analyzed samples at concentrations in the range of 45.1-53.9 µg/L. e research has contributed to the development of CE-C 4 D applications in food and beverage quality control, and the developed method can be applied to local laboratories with modest budgets and limited expertise.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this article.