Development of a MOF-based SPE method combined with GC–MS for simultaneous determination of alachlor, acetochlor and pretilachlor in field soil

In this work, a rapid, highly selective, reusable and effective method was developed for simultaneous determination of alachlor, acetochlor and pretilachlor in field soil by GC–MS coupled with MIL-101 based SPE. Main factors affecting the SPE by using MIL-101 were optimized. Moreover, by comparing with the other commercial materials such as C18, PSA and Florisil, the MIL-101(Cr) exhibited excellent adsorption performance, which aimed at amide herbicides. On the other hand, method validation displayed excellent method performance, achieving good linearities with r2 ≥ 0.9921, limits of detection between 0.25–0.45 μg kg−1, enrichment factors ≥ 89, matrix effect in the range of ± 20%, recoveries between 86.3% and 102.4%, and RSD lower than 4.38%. The developed method was successfully applied to the determination of amide herbicides in soil taken from the wheat, corn and soybean field at different depths, where the concentration of alachlor, acetochlor and pretilachlor were in the range of 0.62–8.04 μg kg−1. It was demonstrated that the more depth of soil, the lower of three amide herbicides. This finding could be proposed a novel method to detect the amide herbicides in the agriculture and food industry.


Introduction
Amide herbicides are a sort of herbicides widely used in production of corn, peanut, soybean, cotton and other crops, which has the advantages of wide weeding spectrum, outstanding effect, and low cost (Qu et al., 2017;Zhao et al., 2009). However, most amide herbicides, especially alachlor and acetochlor, are carcinogenic, teratogenic and sterile (Gao et al., 2022a). Furthermore, amide herbicides residual in the soil could damage crops (Yang et al., 2021) and affect adjacent systems, such as fish, water bird and other aquatic animals (Qu et al., 2017) by water cycle system. Therefore, it is necessary to establish determination method of amide herbicide residues.
Until now, different assay method including gas chromatography-electron capture detector (GC-ECD) (Chen et al., 2022), GC-MS (Yu et al., 2019; et al., 2008), high-performance liquid chromatography (HPLC) (Xie et al., 2016), HPLC-MS (Liu et al., 2020), fluorescence (FL) analysis (Zha et al., 2021), and photoelectronchemical (PEC) analysis (Jin et al., 2017) have been developed for qualitative and quantitative analysis of amide herbicides residues. Among these detection methods, GC-MS has become one of the most effective means to analyze pesticide residues in complex matrix as a result of its strong separation ability, high sensitivity and simple operation.
Nevertheless, as a result of their trace amount and complex matrix, some preconcentration and clean-up processes are required. Current sample pretreatment techniques such as accelerated solvent extraction (ASE) (Cara et al., 2021), dispersive liquid-liquid microextraction (DLLME) (Zhao et al., 2009), LLME (Zhao et al., 2008), SPE (Gao et al., 2022b;Nevado et al., 2007) and solid phase microextraction (SPME) (Zhang et al., 2014) have been reported for determination of amide herbicides residues in different samples. Nowadays, with the development of biotechnology, immunosorbent chromatography (ICA) (Chen et al., 2009)and molecular imprinting technology (MIT) (Yu et al., 2019) have also been used for the purification of amide herbicide.
Developed from the liquid-solid extraction technology, SPE is a sample pretreatment technology mainly used for the separation, enrichment and purification of analytes (Nevado et al., 2007), which has the advantages of high recoveries and less organic solvent consumption. For SPE process, the adsorption performance of adsorbent material is the most critical factor. Although common SPE cartridges on the market such as Bond Elut-ENV SPE cartridges (Nevado et al., 2007), C18 (Qu et al., 2017), HLB (Gao et al., 2022b), MCX (Liu et al., 2020), GCB (Yu et al., 2019) and PSA (Yu et al., 2019) are widely applied to the pretreatment of amide herbicide, adsorbent material with better adsorption capacity need to be developed. Hu et al. prepared a new adsorbent to adsorb nitrogenous pollutants in water including amide herbicides, which is boronate affinity (Hu et al., 2014). At the same time, other adsorbent materials such as graphene have also been used for the enrichment of amide herbicides (Zhang et al., 2014).
As Metal-Organic Frameworks (MOFs) gradually attract people's attention, their functions are better developed (Wang et al., 2019). MOFs are a kind of porous material developed rapidly in the past two decades which have large surface and controlled surface-functionalization by selecting different metal ions or organic bridging ligands area (Zhou et al., 2019). In general, metal ions are used as connection points, while organic ligands are used to form a spatial 3D extension (Wu et al., 2020). Based on their special structure and properties, MOFs are widely used in catalysis (Jing et al., 2020;Miyake et al., 2020), energy storage (Tate et al., 2016;Yang et al., 2013;Zheng et al., 2018), separation (Wu et al., 2020) and enrichment (Wang et al., 2019). In previous literatures, MOFs showed excellent adsorption properties to glycopeptides (Pu et al., 2019), immunoprotein G (Ig) (Hu et al., 2020), organic dyes (Wu et al., 2021;Xu et al., 2019) and organic UV filters, etc. (Nurerk et al., 2020). However, among the numerous MOFs, MIL-101 is gaining increasing attention (Zhang et al., 2013;Zheng et al., 2018) as a result of its superior adsorption and low cost, which has widely used in the pretreatment of compounds (Nurerk et al., 2020).
To our best knowledge, there is no relevant literature that applied MIL-101(Cr) to the pretreatment of amide herbicides in soil matrices. Therefore, for accurate determination of trace amount of amide herbicides in soil, MIL-101(Cr) was prepared and selected as SPE absorbent coupled with GC-MS analysis. Surprisingly, the SPE-GC-MS method based on MIL-101(Cr) was demonstrated to be accurate and feasible, which presented advantages of simplicity, reusability and environmental friendliness compared with other conventional methods.
A Bruker Alpha spectrometer (Ettlingen, Germany) was employed to collect Fourier transform infrared (FT-IR) spectra. The crystal structure of the absorbent was recorded using a SMARTLAB9 X-ray Diffractometer (Rigaku, Tokyo, Japan). The SEM images were obtained from a FEI Inspect F50 SEM (FEI (China), Shanghai, China). The SPE process was carried on a LC-CQ-12F SPE device (LICHEN, Shanghai, China). A GC-MS equipment 7890B-5977A (Agilent Technologies (China), Shanghai, China) and a DB-5MS column (Agilent Technologies (China), Shanghai, China) was used to carry out the separation and analysis of herbicides. Data analysis and management of the GC-MS system were done with Agilent Chem Station software.
After cooling down, the green solid was separated from water by a centrifuge (10,000 rpm, 5 min), and the precipitation was washed ten times alternately with hot ethanol and N,N-dimethylformamide. Finally, the green solid was dried at 80 ℃ and kept in an airtight container.

Samples
Soil samples were taken from the wheat field (A), corn field (B) and soybean field (C) (Kunming, China) at different depths (0-10 cm, 10-20 cm, 20-30 cm and 30-40 cm). After natural drying, the soil was ground and passed through a 40-60 mesh sample sieve. Then, 10 g of soil was weighed and blended with 20 mL of methanol followed by an ultrasonic extraction for 30 min. Finally, the supernatant was separated through a centrifuge (10,000 rpm, 8 min) and collected at 4 ℃.
For the blank soil sample, the soil was heated at 300 ℃ in an oven to remove possible analytes, and then it was put into a glass bottle after cooling.
Solid phase extraction procedure Extraction process was depicted in Fig. 1. Briefly, 200 mg of MIL-101 green powder was put into a 12 mL empty SPE column connected to the SPE device. Before extraction process, 10 mL of acetone and 10 mL of dichloromethane were loaded into the column successively. Afterward, 50 mL of sample (in five times) was loaded through the absorbent with vacuum. Subsequently, 30 mL of EAC: ACE (1:1, v/v) was loaded through the absorbent with gravity. Finally, the eluent was blown to dryness by nitrogen and dissolved in 0.5 mL of dichloromethane.

GC-MS analysis
A DB-5MS column (Agilent Technologies (China), Shanghai, China) was used to separate analytes. With its flow rate at 1.0 mL min −1 , high purity helium was used as carrier gas, and injector temperature was set at 260 ℃. Started at 40 ℃, the oven temperature program raised to 180 ℃ (20 ℃ min −1 ) and kept isothermally for 4 min. Afterwards, it raised to 200 ℃ (3 ℃ min −1 ) and maintained for 3 min. Then it increased to a final temperature of 280 ℃ (20 ℃ min −1 ) and maintained for 5 min. For mass spectrometry, a single quadrupole mass analyzer was used to perform a full scan of 45-500 amu. The temperature of electron impact ion source (EI) and quadrupole was set at 230 ℃ and 150 ℃.
For each amide herbicide, the ion with the highest abundance was selected and used for quantitative analysis, and the second as well as the third was used for the identification of compounds.

Method validation
Based on determination of linearity, LODs, precision, and recoveries, the developed method was validated. The linearity and LODs were calculated from the calibration curve, while the precision was calculated in terms of RSD values (intra-and inter-day). For each herbicide, the concentration used for the calibration curve was 0.05, 0.5, 1.0, 2.0, 3.0, 4.0 and 5.0 μg mL −1 . The recovery at three spike levels (10, 20 and 50 μg L −1 ) was determined to indicate the accuracy of the method, and each spike level was repeated three times. To assess the extraction efficiency of the developed method, the enrichment factor (EF) was calculated according to Eq. (1): where [Analyte]org is the of analyte in organic solvent after extraction while [Analyte]aq is the concentration of analyte in soil sample before extraction.
Due to the numerous compounds in soil which may interfere with the analysis of target compounds, the matrix effects (ME) need to be evaluated. The blank soil sample was treated according to the same extraction process and then spiked with the standards at the same concentration as the standard curve in organic solvent. The matrix effects were calculated according to Eq. (2): where k matrix means slope of standard curve in matrix while k solvent means the slope of standard curve in organic solvent. The ME (%) needs to be within ± 20% to meet the performance acceptance criteria.

Real sample detection
Different soil samples (taken in Kunming, China) were determined using the MIL-101-based SPE-GC-MS method. Firstly, 50 mL of the sample was loaded through the absorbent with vacuum. Once all the sample was passed through the column, the analytes were eluted with 30 mL of eluent, EAC: ACE (1:1, v/v), which was slowly blown by nitrogen and dissolved in 0.5 mL of dichloromethane before analysis of GC-MS.

Characterization
In order to confirm that MIL-101 was successfully synthesized, the absorbent was characterized by FT-IR, SEM, and XRD. The results of FT-IR spectrum, as depicted in Fig. 2, confirmed that the ligand was successfully connected. Owing to the stretching vibration C-O of terephthalic acid, there was a band at 1400 cm −1 in the spectrum, and the bands at 1600 cm −1 (C = C), 1100 cm −1 and Vol.: (0123456789) 750 cm −1 (C-H) meant the appearance of benzene. The XRD pattern (Fig. 3) showed the crystalline property of the absorbent, which was consistent with previous literature (Bromberg et al., 2012). As for the surface morphology, the SEM image (Fig. 4) indicated that the adsorbent has excellent crystallization properties and high porosity. The crystal had a regular octahedral structure, and the particle size was between 100-500 nm, which was consistent with the previous results.

Optimization of SPE
Once the absorbent was synthesized, it was applied in a SPE procedure for the extraction of target analytes from soil samples. In order to achieve the ideal effect, variables affecting the extraction procedure were optimized. The mixed solution (0.1 μg mL −1 ) was used to optimize SPE process three times. In addition, ANOVA was carried out to investigate the significant difference of each factor.

Composition of the eluent
As the most critical factor during the elution procedure, the polarity of the eluent determines the effect of elution. Only when the polarity of the eluent is close to that of the analytes, the analytes will be completely eluted. Due to the weak polarity of amide herbicides, acetonitrile, dichloromethane, ethyl acetate, acetone and EAC: ACE (1:1, v/v) were selected. Table S5 showed the results of one-way ANOVA, indicating that the difference of eluents was significant. (p ≤ 0.001) As illustrated in Fig. S10, recoveries reached the maximum value (more than 90%) when using EAC: ACE (1:1, v/v) as eluent. Thus, EAC: ACE (1:1, v/v) was chosen as the best eluent.

pH, NaCl and eluent volume
After the composition of eluent was selected, effects of three main factors, sample pH, NaCl addition and eluent volume, were also investigated. The eluent volume (A) was assessed at three levels (20, 30 and 50 mL), the sample pH (B) at four levels (3,5,7,9), and the NaCl (C) at three levels (0.0, 5 and 10% (w/v)). Table S6 showed the results for the multifactor ANOVA study. As can be seen, factor A (eluent volume) was the most influential. (p ≤ 0.001) Some of interactions of factor A with factor B or C had statistically significant while factor B and C as well as their interactions were no significant. Figure 5 showed the Pareto charts which clearly pointed out the most important factors. Obviously, factor A accounted for more than 50% of all factors, indicating that the eluent volume was the most important factor affecting the extraction process among these factors. However, factor B, C and BC accounted for less than 30%, which indicated that the pH and addition of NaCl affected the experimental results to a small extent.
Considering the effect of eluent volume, insufficient amount of eluent would lead to incomplete elution, while excess eluent would cause impurities to be eluted, as well as the use of too much organic reagent. Herein, the experiments were conducted with elution volumes of 20 mL, 30 mL and 50 mL. As shown in Fig. 6c, when the elution volume exceeded 30 mL, further addition of eluent did not have much effect on the recoveries. Consequently, 30 mL was selected as the most suitable elution volume. Additionally, according to Fig. 6a and b, pH was adjusted to 5.0 and no NaCl was added to all the samples for further analyses.

Sample breakthrough volume
Since inappropriate amount of sample could result in the reduction of recovery, the sample volume needs to be explored. In this section, 200 mg of MIL-101 were loaded into a cartridge. Afterwards, 40 mL, 50 mL, 60 mL and 70 mL of the samples were loaded to the cartridges followed by the SPE based on MIL-101. As shown in Fig. S11, when more than 50 mL of sample was passed through, recoveries decreased, which was probably attributed to the saturation of the adsorption sites. As a result, a sample breakthrough volume of 50 mL was chosen for SPE process.

Method validation
Under the optimized conditions, several relevant parameters were investigated to appraise the analytical performance of the current method. The example of the GC-MS SIM chromatogram of black soil sample spiked at 400 µg mL −1 was presented in Fig. 7, where it can be seen that there was no interfering signal in the retention time area of target analytes.
Three herbicide standard solutions in dichloromethane at 0.05, 0.5, 1.0, 2.0,3.0, 4.0,5.0 µg mL −1 were prepared and analyzed by the established GC-MS method. As can be seen in Table 1, for three herbicides, regression coefficients were all greater than 0.99, indicating excellent linear relationships.
For determination of the limit of detection (LOD), herbicides at different concentrations were added to blank soil sample. The LODs were calculated as the concentration at which the signal intensity was three times that of the noise. Results obtained showed that method LODs ranged from 0.25 to 0.45 µg kg −1 for all herbicides, resulting from the strong adsorption to these herbicides of MIL-101.
Although good linear relationships and low LODs were validated, influence of interfering substances in matrix was required to be investigated, which was calculated as matrix effect (ME). In this study, ME (%) of three herbicides were all in the range of ± 20%, which indicated that the influence of matrix on the analytes was negligible, attributing to the good cleaning ability of MIL-101(Cr).
In addition to excellent performance in purification, the absorbent has strong adsorption capacity for analytes, which was reflected in its high enrichment factor. For three herbicides, EF were calculated between 89 and 98 when only 200 mg of absorbent was used.
We also examined the method precision by adding 100 μg kg −1 herbicides into blank soil sample within a day (n = 3) and three days (n = 6). The RSD values were in the range of 2.23% -4.38% (intra-day) and 2.07%-3.05% (inter-day), which demonstrated the high reproducibility of the experiment.
To verify the accuracy of the method, blank soil sample was spiked at 10, 20 and 50 μg kg −1 prior to analysis. As shown in Table 1, recoveries for alachlor and acetochlor ranged from 89.0 to 102.3% with RSD values less than 3.4%, indicating appropriate accuracy of the method. Although recoveries for pretilachlor were lower than alachlor and acetochlor, the values between 86.3 and 90.7% were acceptable.
While obtaining good accuracy and precision, the proposed method provided low LODs and matrix effects, suggesting that the developed method for analysis of amide herbicides residues in soil was efficient, precise and reliable.

Reusability of MIL-101(Cr)
Despite validation of some basic parameters, reusability of MIL-101(Cr) was investigated in response to the purpose of green chemistry. After each extraction, the absorbent was washed alternately with ethyl acetate, acetone and acetonitrile and put into a 100 ℃ oven for three hours. Subsequently, recoveries of three herbicides was determined to assess the reusability of MIL-101(Cr). As shown in Fig. S12, recoveries hardly changed after five cycles, indicating the high stability and reusability of MIL-101(Cr). Thus, MIL-101(Cr) was able to be reused at least five times, which made this method more convenient and economical.

Comparison with C18, PSA and Florisil
For the verification of the good adsorption of MIL-101(Cr), the extraction efficiency was compared with that of three commercial cartridges (C18, PSA and Florisil). As the most commonly used adsorbent, C18  has strong hydrophobicity as a result of the saturated aliphatic chain with 18 C atoms. PSA has two amino groups and has strong ion exchange capacity, which can effectively remove fatty acids and sugars in matrix. As for Florisil, it is mainly used for adsorption of polar compounds such as ethanol, dyes, pesticide, and nitrogen-containing compounds. As shown in Fig. 8, at the spiked concentration of 0.1 μg mL −1 , the recoveries of the MIL-101 group (≥ 90%) were higher than those of other groups. Its high adsorption to amide herbicides might be attributed to the coordination of unsaturated Cr 3+ in MIL-101 with Nitrogen atoms in amide herbicides. (Nurerk et al., 2020).Among the three commercial cartridges, the Florisil achieved great recoveries (≥ 80%), while the C18 and PSA, especially PSA, didn't work very well.

Comparison with other reports
Comparison between the present work and those already published was enumerated in Table 2. It turned out that the developed method showed competitive parameters such as recoveries and RSD values. Besides, as a result of its excellent adsorption property and reusability, a small amount of MIL-101 absorbent was used, achieving good results compared with other materials. On the other hand, it is a relatively environmentally friendly method which only consumes a small amount of harmful solvents and reusable absorbent.

Application in real soil samples
Alachlor, acetochlor and pretilachlor, especially acetochlor, are frequently used in agriculture. Once left in the soil, they could damage crops and affect adjacent systems. In this study, wheat, corn and soybean fields were selected as a result of their large consumption of amide herbicides. Alachlor, acetochlor and pretilachlor in 12 soil samples were determined using the MIL-101-based SPE-GC-MS method. Figure 9 exhibited the detection results. Among these herbicides,  acetochlor was used most frequently and detected in all three fields. Furthermore, it was found that the deeper the soil, the lower the herbicide concentration, which might be attributed to their migration activity and degradation by underground microorganisms.

Conclusions
In this study, a MIL-101-based SPE-GC-MS method was developed for the determination of amide herbicides in field soil. After optimization of parameters affecting extraction, the method exhibited excellent sensitivity, satisfactory repeatability and good accuracy. In addition, it is worth mentioning that the MIL-101(Cr) absorbent could be reused at least five times, which reduced waste and save analysis time. Compared with other commercial cartridges, the MIL-101(Cr) has stronger adsorption to amide herbicides, which is attributed to the coordination of unsaturated Cr 3+ in MIL-101 with Nitrogen atoms in amide herbicides. (Nurerk et al., 2020). The developed method has been successfully applied to the pretreatment and quantification of amide herbicides in wheat, corn and soybean field soil, where alachlor and acetochlor were detected. Owing to the strong adsorption of MIL-101(Cr), our method offers a strategy for the analysis of other compounds containing heteroatoms.
Funding This work was financially supported by the Programs for Open Project from Yunnan Key Laboratory of Tobacco Chemistry (Development and application of plant-derived thermo-aromatic substances in HNB, No. 2019539200340164), Pu'er Quality and Technical Supervision Comprehensive Inspection Center (Fan Chunlin Expert Workstation, No. 20210903) and Non-subject project of R&D Center of China Tobacco Yunnan Industry Co., Ltd (Study on the application of aroma components of five yun-produced characteristic tobacco for heating cigarettes, No.KY-800322066.04).
Data availability Data will be made available on reasonable request.

Declarations
Competing interests The authors declare no competing interests.

Ethics approval Not necessary.
Consent to participate Informed consent not applicable.