A method for detecting pesticide in rice based on QuEChERS and UV-visible spectrophotometry

Abstract

The application discloses a pesticide detection method for rice based on QuEChERS and ultraviolet-visible spectrophotometry, and takes three pesticides, i.e., bensulfuron-methyl, propanil and cypermethrin, as examples to carry out research on a rapid screening method for multiple pesticide residues in rice. The improved QuEChERS method is used to process rice to obtain a matrix liquid, matrix standard solution and solvent standard solution are prepared, the quantitative linear range of each pesticide and the matrix effect are obtained. In the quantitative concentration range, mixed solutions of the three pesticides in four combinations are prepared, the types of the pesticides in the rice are determined through support vector machine qualitative analysis, and the concentrations of the pesticides are predicted through partial least squares regression analysis, so that the pesticide residues in the rice can be effectively and rapidly detected.

Description

Technical Field

[0001] This invention relates to a detection method, specifically a method for detecting pesticides in rice based on a combination of QuEChERS and ultraviolet-visible spectrophotometry. Background Technology

[0002] Globally, over 60% of crops are related to grains, and rice is a staple food for nearly 60% of the population. While pesticides effectively control weeds and pests and increase crop yields, problems such as the illegal use of prohibited or restricted pesticides, irrational pesticide use, the addition of prohibited or restricted ingredients to pesticides, and inadequate monitoring systems lead to excessive pesticide residues. These residues can accumulate in animals and humans through bioaccumulation in the food chain, causing acute or chronic poisoning, seriously threatening human health and posing a significant safety hazard to the food industry.

[0003] The rapid pace of pesticide product updates, the increasing variety of pesticides, the diverse structures and properties of pesticides, the complexity of sample matrices, the increasingly stringent pesticide residue limits, and the high timeliness requirements for screening all pose significant challenges to trace and ultra-trace pesticide residue analysis. Developing simple, accurate, efficient, high-throughput, and highly sensitive rapid screening technologies for multiple pesticide residues has become a current research hotspot. Pesticide residue analysis mainly includes sample pretreatment and detection techniques. The core issues in pesticide residue analysis are how to remove background interference, ensure the response of target compounds, and perform non-targeted qualitative analysis.

[0004] Classic sample pretreatment methods, such as liquid-liquid extraction, Soxhlet extraction, and activated carbon adsorption, suffer from drawbacks including long processing times, numerous steps, potential analyte loss, high consumption of organic solvents, and high costs. These methods have been gradually replaced by faster, more efficient, and environmentally friendly pretreatment methods. Currently, solid-phase extraction (SPE) is widely used in domestic laboratories. However, the experimental process is cumbersome and time-consuming, requiring extensive column activation and solvent elution, resulting in high costs and significant environmental pollution. While newly reported sample pretreatment techniques such as solid-phase microextraction, supercritical fluid extraction, microwave-assisted extraction, and molecularly imprinted acceptor synthesis offer advantages such as high efficiency, speed, safety, simplicity, and less environmental pollution, they still cannot meet the technical requirements for simultaneous detection of the vast majority of pesticides with low limits of quantitation.

[0005] The QuEChERS method integrates extraction and purification, and is characterized by its speed, simplicity, low cost, high recovery rate, and few extraction steps, making it suitable for real-time pesticide residue detection. For nearly two decades since its inception, pesticide residue analysts both domestically and internationally have been dedicated to its application in rice pesticide residue detection. Nguyen TD; Lee BS; Lee BR; Lee DM; Lee G.-H.; A multiresidue method for the determination of 109 pesticides in rice using the Quick Easy Cheap Effective Rugged and Safe (QuEChERS) sample preparation method and gas chromatography / mass spectrometry with temperature control and vacuum concentration[J], Rapid Communications in Mass Spectrometry, 2007, 21:3115-3122. In the literature, Nguyen et al. prepared samples at low temperature, which improved the recovery rate of heat-sensitive pesticides; during the extraction process, Koesukwiwat U.; Sanguankaew K.; Leepipatpiboon N.; Rapid determination of phenoxy acid residues in rice by modified QuEChERS extraction and liquid chromatography-tandem mass spectrometry[J], Analytica Chimica In Acta, 2008, 626:10-20, Koesukwiwat et al. added citrate-sodium citrate buffer solution and compared the mixing effects of different solid adsorbents, finding that the adsorbent had a significant impact on the purification effect of the target compound; Pareja L.; Cesio V.; Heinzen H.; Fernández-Alba AREvaluation of various QuEChERS based methods for the analysis of herbicides and other commonly used pesticides in polished rice by LC-MS / MS[J], Talanta, 2011, 83: 1613-1622. In the literature, Pareja et al. added acetate-sodium acetate buffer solution to improve the recovery rate of alkaline and acidic pesticides in rice; Mastovska K.; Dorweiler KJ; Lehotay SJ; Wegscheid J.S.; Szpylka KA; Pesticide multiresidue analysis in cereal grains using modified QuEChERS method combined with automated direct sample introduction GC-TOFMS and UPLC-MS / MS techniques[J], Journal of Agricultural and Food Chemistry, 2010, 58: 5959-5972.

[0006] In the literature, Mastovska et al. used an automatic shaker to replace the manual shaking method, ensuring thorough mixing of the samples; Kaczyński P.; B.;One-step QuEChERS-based approach to extraction and cleanup in multiresidue analysis of sulfonylurea herbicides in cereals by liquid chromatography-tandem mass spectrometry[J], Food Analytical Methods, 2017, 10:147-160. In the literature, Kaczyński et al. combined the extraction and purification steps into one step, and used a one-step method to process the sample, obtaining ideal results in the analysis of 23 sulfonylurea herbicides in cereals. I.; S.; M.; D.; ; A.;Evaluation of LC-MS / MS methodology fordetermination of 179multi-class pesticides in cabbage and rice by modifiedQuEChERS extraction[J],Food Control,2021,123:107693In the literature While dilution injection has replaced purification, simplifying sample processing and reducing analyte loss during sample preparation, the reduced amount of analyte injected into the instrument lowers the method's sensitivity. Due to its superior performance, the QuEChERS method has become the preferred method for cereal sample pretreatment in the current Chinese standard GB23200.113-2018, "Determination of Residues of 208 Pesticides and Their Metabolites in Plant-Derived Foods by Gas Chromatography-Mass Spectrometry".

[0007] Although rice is the most commonly consumed food worldwide, there are relatively few reports on methods for detecting pesticide residues in this complex sample matrix. High-performance liquid chromatography (HPLC), liquid chromatography-mass spectrometry (LC-MS), gas chromatography (GC), and gas chromatography-mass spectrometry (GC-MS) are common analytical techniques. GC and HPLC mainly rely on retention time for qualitative analysis, which has limited screening capabilities for multiple pesticide residues and is easily interfered with by complex sample matrices. GC and GC-MS are not suitable for analyzing highly polar, high molecular weight, and thermally unstable compounds. The main advantage of MS lies in its very high separation efficiency and equally high detection sensitivity. This makes LC-MS more popular for detecting multiple pesticide residues. Its analytical level is generally below g / L, meeting the EU's requirements for high sensitivity, selectivity, and specificity in food testing, making it an important analytical method. Patent CN113376298A describes a method for rapidly determining the residues of four pesticides in rice. This method utilizes ultra-high performance liquid chromatography-tandem mass spectrometry to determine the residues of the four pesticides. It is rapid, simple, and accurate in determining the residues of the four pesticides in rice. However, this detection method still has the disadvantages of being cumbersome and time-consuming in terms of rapid on-site identification, and is not suitable for rapid on-site screening and analysis of large batches of samples.

[0008] my country's GB 2763-2021 standard, "Maximum Residue Limits for Pesticides in Food," specifies the maximum residue limits for 31 pesticides in rice. It is impossible for a single type of rice to contain so many pesticides simultaneously. However, various types of rice may contain a wide variety of pesticides. Current detection methods consume a large amount of pesticide standards, resulting in low detection efficiency. Therefore, developing an analytical detection technology that can overcome these shortcomings and conducting research on rapid screening methods for multiple pesticide residues in rice has become an urgent technical problem for those skilled in the art. Summary of the Invention

[0009] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a method for detecting pesticide residues in rice based on the combination of QuEChERS and ultraviolet-visible spectrophotometry. This method can effectively and rapidly detect pesticide residues in rice.

[0010] To address the above technical problems, this invention provides a method for detecting pesticides in rice based on QuEChERS coupled with ultraviolet-visible spectrophotometry, specifically including the following steps:

[0011] S1: Rice samples treated using the QuEChERS method

[0012] Take Wuchang organic rice, grind it using a grinder, pass it through a 20-mesh sieve, and set it aside. Weigh the ground rice powder, put it into a plastic centrifuge tube, add deionized water, vortex to mix, let it stand, add acetic acid acetonitrile solution, anhydrous magnesium sulfate, and sodium acetate, cover the centrifuge tube, shake vigorously to dislodge the lower layer of solid, then vortex to mix, put it into a centrifuge, centrifuge, pipette the supernatant into a centrifuge tube, add anhydrous magnesium sulfate and C18, vortex to mix, centrifuge, dilute the supernatant with acetonitrile by 1:1 to obtain the rice matrix solution; S2: Preparation of solvent standard solution

[0013] Standard working solutions of bensulfuron-methyl, propargyl and cypermethrin were prepared using acetic acid acetonitrile solution as solvent.

[0014] S3: Preparation of matrix standard solution

[0015] The rice matrix solution prepared in S1 was used to prepare matrix standard working solutions for three pesticides: bensulfuron-methyl, propargyl, and cypermethrin; S4: Preparation of mixed pesticide matrix solution.

[0016] Using the orthogonal experimental design, four concentration levels were set for each pesticide within a quantitative range. Rice matrix solutions containing a mixture of bensulfuron-methyl (A), propargite (B), and cypermethrin (C) were prepared, resulting in four combinations: AB, AC, BC, and ABC.

[0017] S5: The spectral acquisition method uses a 0.5% acetonitrile acetate solution as a reference. The test solution, i.e. the prepared matrix standard solution, is placed in the sample cell of a 1cm cuvette. The spectrum is scanned from 350nm to 190nm, and absorbance data is collected at 1nm intervals.

[0018] S6: Qualitative Identification Methods for Pesticide Types

[0019] SVM is used to build an intelligent recognition model. All data in the spectrum are used as independent variables. The combination of four pesticides corresponds to four types of samples. The Gaussian kernel function is used as the kernel function. The gamma and c parameters in the SVM model are optimized by grid search technology. Finally, the recognition rate of the validation set is used to compare the performance of the models.

[0020] S7: Methods for Quantifying Pesticide Concentration

[0021] Partial least squares regression was used to fit the training set samples for each combination, and quantitative analysis models for AB, AC, BC, and ABC combinations were established respectively. The peak values ​​corresponding to the entire band were used as independent variables and the concentration was used as dependent variables to establish quantitative analysis models and quickly detect the concentration of the prediction set samples for each combination.

[0022] The technical solution further defined in this invention is:

[0023] Furthermore, in the aforementioned method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry, the acetonitrile acetate solution added in step S1 is an acetonitrile acetate solution with a volume concentration of 1%.

[0024] In the aforementioned method for detecting pesticides in rice based on the combination of QuEChERS and UV-Vis spectrophotometry, the centrifuge speed in step S1 is 4000-4200 r / min, and the centrifugation time is 5-6 min.

[0025] In the aforementioned method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry, the solvent in step S2 is an acetonitrile acetate solution with a volume concentration of 0.5%.

[0026] In the aforementioned method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry, the concentrations of the standard working solutions of bensulfuron-methyl, propargyl, and cypermethrin in step S2 are 0.078, 0.156, 0.3125, 0.625, 1.25, 2.5, 5, 10, 20, 40, and 80 mg / L, respectively.

[0027] In the aforementioned method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry, the concentrations of the matrix standard working solutions for bensulfuron-methyl, propargyl, and cypermethrin in step S2 are 0.3125, 0.625, 1.25, 2.5, 5, 10, 20, and 40 mg / L, respectively.

[0028] In the aforementioned method for detecting pesticides in rice based on the combination of QuEChERS and UV-Vis spectrophotometry, the test solution is tested three times in step S5, and the data obtained are processed using the Matlab 2010b platform.

[0029] The technical effect is that measuring three times can effectively reduce the error of experimental measurements.

[0030] In the aforementioned method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry, the specific parameter settings in step S6 are as follows: the range of C is limited to [2]. －8 ,2 8 The range of gamma is [2]. －8 ,2 8 Five-fold cross-validation was performed, with step values ​​of 0.5 for both C and gamma. The final step value for the accuracy discretization of the parameter selection results was 4.5, and the model recognition rate was 90.625%.

[0031] The beneficial effects of this invention are:

[0032] The Support Vector Machine (SVM) used in this invention was first proposed by Cortes and Vapnik in 1995. It is based on the VC dimension theory and the principle of structural risk minimization in statistical learning theory. It seeks the optimal trade-off between model complexity and learning ability based on limited sample information, aiming to achieve the best generalization capability. It exhibits unique advantages in solving small-sample, nonlinear, and high-dimensional pattern recognition problems.

[0033] Partial least squares regression (PLSR) is a multivariate statistical analysis method that combines principal component analysis and canonical correlation analysis. Because PLSR effectively addresses the problem of multicollinearity among high-dimensional variables, it is widely used in the selection of spectral variables and the establishment of quantitative spectral analysis models.

[0034] First, the type of pesticide in the sample is initially determined by SVM, and then the pesticide content is accurately predicted by PLSR. The QuEChERS sample pretreatment technology and ultraviolet-visible spectrophotometric analysis and detection technology combined with chemometrics method established in this invention can rapidly detect pesticide residues.

[0035] my country's GB 2763-2021 standard, "Maximum Residue Limits for Pesticides in Food," specifies the maximum residue limits for 31 pesticides in rice. It's impossible for a single type of rice to contain so many pesticides simultaneously; however, different types of rice may contain a wider variety of pesticides. To conserve pesticide standards and improve detection efficiency, it's necessary to first conduct rapid screening of the rice samples to be tested, followed by accurate quantification of positive samples. The QuEChERS method allows for rapid sample pretreatment, while UV-Vis spectrophotometry offers fast scanning speeds. Combining these two methods can improve analytical efficiency and is expected to become a primary technology for rapid pesticide residue detection. This study investigated a rapid screening method for multiple pesticide residues in rice using a combination of QuEChERS sample pretreatment and UV-Vis spectrophotometry. Taking three pesticides—benzylsulfuron (A), propargite (B), and cypermethrin (C)—that are limited to rice in GB 2763 and have similar UV-Vis spectrophotometric peaks and are difficult to distinguish, as examples, the QuEChERS method was used to process the samples to prepare rice matrix solutions. Matrix and solvent standard solutions were then prepared to obtain the linear concentration ranges and matrix effects for each pesticide. The linear ranges for A, B, and C were 0.625-40, 0.3125-20, and 0.625-20 mg / L, respectively, with matrix effects of -0.51%, -3.33%, and -16.08%, respectively. Within a quantitative concentration range, a mixed solution of three pesticides was prepared, resulting in four combinations. Qualitative analysis using SVM can identify the pesticide combinations in rice with a recognition rate exceeding 90%. Further PLSR analysis can predict the concentration of each pesticide in the mixed pesticide solution with an identification error of less than 5.36%. The QuEChERS method enables rapid sample pretreatment, while UV-Vis spectrophotometry offers fast scanning speed. Combining these two methods improves analytical efficiency. By establishing a supervised identification model for pesticide composition using SVM and then predicting pesticide content using PLSR, rapid detection of pesticide residues is possible. Attached Figure Description

[0036] Figure 1 This is a spectrum of absorbance values ​​of bensulfuron-methyl at different concentrations in the bensulfuron-methyl standard working solution in this embodiment of the invention;

[0037] Figure 2 The images show the absorbance values ​​of different concentrations of *Pteris vittata* in the standard working solution of *Pteris vittata* in this embodiment of the invention.

[0038] Figure 3 This is a spectrum of absorbance values ​​of cypermethrin at different concentrations in the cypermethrin standard working solution in this embodiment of the invention;

[0039] Figure 4 This is a spectrum of absorbance values ​​of bensulfuron-methyl at different concentrations in the bensulfuron-methyl matrix standard working solution in this embodiment of the invention;

[0040] Figure 5 This is a spectrum of absorbance values ​​at different concentrations of propantheline in the standard working solution of propantheline matrix in an embodiment of the present invention;

[0041] Figure 6 This is a spectrum of absorbance values ​​of cypermethrin at different concentrations in the cypermethrin matrix standard working solution in this embodiment of the invention;

[0042] Figure 7 The spectrum of the matrix standard solutions of bensulfuron-methyl (A), propargite (B), and cypermethrin (C) in the embodiments of the present invention is at 5 mg / L.

[0043] Figure 8 The images show the spectra of four combined pesticides in this embodiment of the invention. Detailed Implementation

[0044] The following were used in the examples: Wuchang organic rice (Wuchang Jinggong Rice Food Co., Ltd.), bensulfuron-methyl, propargyl, cypermethrin (analytical standard, Shanghai Aladdin Biochemical Technology Co., Ltd.), acetonitrile, acetic acid, anhydrous magnesium sulfate, sodium acetate (analytical grade, Chengdu Kelong Chemical Co., Ltd.), deionized water (grade I, 18.02 MΩ·cm, Ron's reagent), and octadecylsilane-bonded silica gel (C18, 40-60 μm, Shanghai Xinhu Experimental Equipment Co., Ltd.).

[0045] T6 UV-Vis spectrophotometer (Beijing Purkinje General Instrument Co., Ltd.); JYS-M01 grinder (Joyoung Co., Ltd.); FA124C electronic analytical balance, LC-LX-L50C centrifuge, LC-Vortex-P2 jog mixer (Lichen Technology).

[0046] The present invention can be better understood from the following embodiments. However, those skilled in the art will readily understand that the descriptions in the embodiments are for illustrative purposes only and should not, and will not, limit the invention as detailed in the claims.

[0047] Example 1

[0048] This embodiment provides a method for detecting pesticides in rice based on QuEChERS combined with ultraviolet-visible spectrophotometry, characterized by the following steps:

[0049] S1: Rice samples treated using the QuEChERS method

[0050] Weigh 100g of Wuchang organic rice, grind it in a grinder, pass it through a 20-mesh sieve, and set it aside. Weigh 5g of rice flour, put it into a 50mL plastic centrifuge tube, add 10mL of deionized water, vortex mix for 2min, let it stand for 30min, add 15mL of 1% acetonitrile acetate solution, 6g of anhydrous magnesium sulfate, and 1.5g of sodium acetate, cap the centrifuge tube, shake it vigorously to loosen the lower solid layer, vortex mix for 2min, put it into a centrifuge, centrifuge at 4200r / min for 5min, pipette 8mL of the supernatant into a 15mL centrifuge tube, add 1.2g of anhydrous magnesium sulfate and 0.4g of C18, vortex mix for 1min, centrifuge at 4000r / min for 5min, dilute the supernatant with acetonitrile by 1:1 to obtain the rice matrix solution;

[0051] S2: Preparation of standard solvent solution

[0052] Using a 0.5% (v / v) acetonitrile acetate solution as the solvent, standard working solutions of bensulfuron-methyl, propargyl, and cypermethrin were prepared with concentrations of 0.078, 0.156, 0.3125, 0.625, 1.25, 2.5, 5, 10, 20, 40, and 80 mg / L, respectively; S3: Preparation of matrix standard solutions.

[0053] The rice matrix solution prepared in S1 was used to prepare matrix standard working solutions of three pesticides: bensulfuron-methyl, propargyl, and cypermethrin, with concentrations of 0.3125, 0.625, 1.25, 2.5, 5, 10, 20, and 40 mg / L, respectively.

[0054] S4: Preparation of Mixed Pesticide Matrix Solution

[0055] Using the orthogonal experimental design, four concentration levels were set for each pesticide within a quantitative range. Rice matrix solutions containing a mixture of bensulfuron-methyl (A), propargite (B), and cypermethrin (C) were prepared, resulting in four combinations: AB, AC, BC, and ABC.

[0056] To verify the accuracy of the qualitative and quantitative analysis models, solutions were prepared for training and validation models, respectively. The specific preparation methods are shown in Table 1.

[0057] Table 1 Concentrations of each pesticide in the mixed pesticide matrix solution

[0058]

[0059]

[0060] S5: The spectral acquisition method uses a 0.5% acetonitrile acetate solution as a reference. The test solution, i.e., the prepared matrix standard solution, is placed in the sample cell of a 1cm cuvette. The spectrum is scanned from 350nm to 190nm, and absorbance data is collected at 1nm intervals. To reduce the experimental measurement error, each solution is measured three times. The obtained data is processed based on the existing Matlab 2010b platform.

[0061] S6: Qualitative Identification Methods for Pesticide Types

[0062] SVM (Support Vector Machine) is used to build an intelligent recognition model. All data from the spectrum are used as independent variables. The combination of four pesticides corresponds to four types of samples. The Gaussian kernel function [RBF (Radial basis function)] is used as the kernel function. The gamma and c parameters in the SVM model are optimized through grid search technology. Finally, the recognition rate of the validation set is used to compare the performance of the models.

[0063] S7: Methods for Quantifying Pesticide Concentration

[0064] Partial least squares regression (PLSR) was used to fit the training set samples 1-16 for each combination, and quantitative analysis models were established for the AB, AC, BC, and ABC combinations respectively. The peak values ​​corresponding to the entire band were used as independent variables and the concentration was used as dependent variables to establish quantitative analysis models and quickly detect the concentration of the prediction set samples for each combination.

[0065] In this implementation, the solvent standard solution was analyzed as follows:

[0066] The wavelength corresponding to the maximum absorbance of bensulfuron-methyl exhibits a red shift with increasing solution concentration, with strong absorbance at 237 nm. Specific data can be found in [reference needed]. Figure 1 As shown in Table 2;

[0067] Table 2 Absorbance values ​​of bensulfuron-methyl at different concentrations

[0068] Concentration mg / L absorbance value at 237nm 80 2.488 40 1.564 20 0.798 10 0.401 5 0.206 2.5 0.112 1.25 0.060 0.625 0.036 0.3125 0.017

[0069] The wavelength corresponding to the maximum absorbance of dimethoate does not change with concentration and remains at 252 nm. (See details...) Figure 2 As shown in Table 3;

[0070] Table 3 Absorbance values ​​of different concentrations of propanil

[0071] Concentration mg / L absorbance value at 252nm 40 2.780 20 1.624 10 0.825 5 0.414 2.5 0.210 1.25 0.106 0.625 0.051 0.3125 0.028 0.156 0.017

[0072] The wavelength corresponding to the maximum absorbance of cypermethrin tends to red-shift with increasing solution concentration, with the strongest absorbance at 226 nm. See details... Figure 3 And as shown in Table 4;

[0073] Absorbance values of cypermethrin at different concentrations

[0074]

[0075]

[0076] At the above wavelengths, a working curve of the solvent standard solution of the pesticide was established to obtain the quantitative concentration range, as shown in Table 5;

[0077] Table 5 Linear range, correlation coefficient, linear equation and matrix effect of 3 pesticides

[0078]

[0079] Analysis of matrix standard solution:

[0080] Some substances in rice can interfere with the analysis process of the target substance and affect the accuracy of the analysis results. These effects and interferences are called matrix effects (matrix effect, ME). The evaluation of matrix effects can be calculated according to formula (1). When -20% < ME < 20%, it is a weak matrix effect. When -50% < ME ≤ -20% or 20% ≤ ME < 50%, it is a medium matrix effect. When ME ≤ -50% or ME ≥ 50%, it is a strong matrix effect;

[0081] ME = (km / ks - 1) × 100% (1)

[0082] In the formula, km and ks are the slopes of the matrix standard curve and the solvent standard curve, respectively.

[0083] Most of the existing methods for treating rice samples by QuEChERS method are to provide test samples for mass spectrometry analysis. In this invention, ultraviolet-visible spectrophotometry is adopted. Acetic acid has strong absorption and will interfere with sample detection. Therefore, on the basis of the QuEChERS method in GB 23200.113, the obtained rice matrix solution is diluted 1 time with acetonitrile to reduce the influence of acetic acid, and a 0.5% acetic acid acetonitrile solvent is used as the test reference solution; the ethylenediamine-N-propylsilylated silica gel (PSA) adopted in the QuEChERS method in GB 23200.113 will adsorb part of acetic acid, resulting in the volume concentration of acetic acid in the rice matrix solution being lower than 0.5% and negative absorbance values appearing in the spectrogram. Therefore, the QuEChERS method in GB 23200.113 is improved, and PSA is not used in the purification stage to obtain the rice matrix solution. The specific operation is as described in S1.

[0084] To evaluate the merits of the improved QuEChERS method for UV-Vis spectrophotometry, a matrix effect assessment was conducted. Using rice matrix solution prepared by the improved QuEChERS method as a solvent, matrix standard solutions of three pesticides were prepared. Working curves for the matrix standard solutions of bensulfuron-methyl, propargyl, and cypermethrin were established at 237 nm, 252 nm, and 226 nm, respectively. Details are shown in Tables 6-8 and 6-8. Figure 4-6 ;

[0085] Table 6 Absorbance values ​​of bensulfuron-methyl matrix solution at different concentrations

[0086] Concentration mg / L absorbance value at 237nm 0.625 0.185 1.25 0.213 2.5 0.268 5 0.375 10 0.576 20 0.962 40 1.713

[0087] Table 7 Absorbance values ​​of different concentrations of the propagule matrix solution

[0088] Concentration mg / L absorbance value at 252nm 0.3125 0.276 0.625 0.303 1.25 0.361 2.5 0.465 5 0.690 10 1.063 20 1.824

[0089] Table 8 Absorbance values ​​of cypermethrin matrix solution at different concentrations

[0090] Concentration mg / L absorbance value at 226 nm 0.625 0.356 1.25 0.385 2.5 0.434 5 0.514 10 0.694 20 0.952

[0091] The matrix effect was calculated based on the above data, and the results are shown in Table 5. The matrix effect of all three pesticides was relatively weak. Figure 7 The spectra of these three pesticides in matrix standard solutions at a concentration of 5 mg / L are shown. It can be seen that their spectra almost completely overlap, making it difficult to perform qualitative and quantitative analysis of these three pesticides simultaneously.

[0092] Qualitative identification analysis often involves the use of several or even dozens of pesticides during rice cultivation. Given the numerous possible mathematical combinations, the likelihood of pesticide residues in rice is extremely high. Sample identification is divided into unsupervised and supervised classification, with supervised classification generally performing better than unsupervised classification. Supervised classification typically requires a large number of samples with known categories; therefore, obtaining prior knowledge about the samples is crucial. This embodiment uses three pesticides with highly overlapping spectra—benzylsulfuron (A), propargite (B), and cypermethrin (C)—as examples. Based on the linear ranges of the three pesticides obtained in Table 5, four concentration levels were set for each pesticide within the linear range: 0.625, 2.5, 10, and 30 mg / L for A; 0.3125, 1.25, 5, and 15 mg / L for B; and 0.625, 2, 5, and 15 mg / L for C. Training set solutions were prepared for pairwise combinations of pesticides (AB, AC, BC) according to a two-factor, four-level orthogonal experiment. The three-pesticide combination (ABC) was prepared according to a three-factor... For the four-level orthogonal experiment, training set solutions were prepared, requiring 16 mixed solutions of different concentrations, as shown in Table 1. Four concentration levels were then set as validation set solutions: 1.25, 5, 15, and 20 mg / L for A; 0.625, 2.5, 10, and 12.5 mg / L for B; and 1.25, 2.5, 10, and 12.5 mg / L for C. Eight validation set solutions were prepared for each pesticide combination according to Table 1. The 16 training set samples from each combination solution were used to train the model, and the 8 validation set samples were used to verify the accuracy of the established model.

[0093] like Figure 8 The spectra of the training set solutions of the four pesticide combinations (a total of 64 spectra) are quite similar to each other. To improve the recognition accuracy, SVM is introduced to establish an intelligent recognition model, using all the data from the spectra as independent variables. The four pesticide combinations correspond to four types of samples. Using RBF (Radial Basis Function) as the kernel function, the gamma and c parameters are optimized through grid search technology. Finally, the recognition rate of the validation set is used to compare the performance of the models. The specific parameter settings are as follows, with the range of C limited to [2]. －8 ,2 8 ], the range of g is [2 －8 ,2 8 Five-fold cross-validation was performed with a step value of 0.5 for both C and g. The final parameter selection results graph shows a step value of 4.5 for the accuracy discretization. The model recognition rate was 90.625%, better than 90%, indicating that when given an unknown solution sample, the established SVM classification model can first determine the combination of pesticides contained in the sample, and then further quantitative analysis can be performed.

[0094] Quantitative analysis:

[0095] The PLSR was used to fit the training set samples 1-16 for each combination, and quantitative analysis models were established for the AB, AC, BC, and ABC combinations respectively. The peak values ​​corresponding to the whole band were used as independent variables and the concentration was used as dependent variables to establish quantitative analysis models. The concentration of the prediction set samples for each combination was predicted. The results are shown in Table 9. The model predicts the concentration of two-component mixed pesticides relatively accurately, while the error for three-component mixtures is slightly larger, but both are within 5.36%, which shows a relatively ideal quantitative prediction capability.

[0096] Table 9 Quantitative Analysis Results of PLSR

[0097]

[0098] 1 RDP S For single-component relative deviation, RDP S (%)=100 / n∑(│x pred -x act │ / x act ), x pred x represents the predicted concentration. act This represents the actual concentration, and n is the number of predicted samples, which is 8 here.

[0099] 2 RDP T For the total relative deviation, RDP T (%) = 100 / m∑RDP S m represents the number of material components in the sample, with AB, AC, and BC having 2 components and ABC having 3 components.

[0100] The UV-Vis spectra of three pesticides—benzylsulfuron, propargite, and cypermethrin—studied in this invention are highly similar, yet excellent results were achieved with a recognition rate exceeding 90% and a concentration prediction error below 5.36%. Other pesticides with limits set in national standards have significantly different structures, resulting in different spectra. Therefore, the method established in this invention is applicable to the analysis of other pesticides. Furthermore, most national standards use liquid chromatography-mass spectrometry (LC-MS), while the UV-Vis spectrophotometry method used in this paper offers shorter detection time and higher detection efficiency. It should be noted that in the QuEChERS method, 5g of rice flour is weighed, and a 15mL solution is obtained. The linear range of the UV-Vis spectrophotometer for detecting substances is higher than 0.3125mg / L. Therefore, pesticides in rice can only be detected when the pesticide content is higher than 0.9375mg / kg. That is, the limits of propargite (2mg / L), isoprothiolane (1mg / L), carbendazim (2mg / L), flufenoxuron (1mg / L), methyl chlorpyrifos (5), methyl pyrimidine phosphate (1mg / L), carbaryl (1mg / L), and fenitrothion (1mg / L) in GB 2763 can be detected by the method established in this invention.

[0101] To address the challenges of diverse pesticide residues, low concentrations, significant interference from sample matrix components, and slow detection speed, this invention employs a combined QuEChERS sample pretreatment technique and UV-Vis spectrophotometry to develop a high-throughput rapid screening method for multiple pesticide residues in rice. Taking three pesticides—benzylsulfuron (A), propargite (B), and cypermethrin (C)—which are limited to rice in GB 2763 and have similar UV-Vis spectroscopic peaks, making them difficult to distinguish, as examples, the QuEChERS method was used to process the sample to prepare a rice matrix solution. Matrix and solvent standard solutions were then prepared to obtain the linear concentration ranges and matrix effects for each pesticide. The linear ranges for A, B, and C were 0.625-40, 0.3125-20, and 0.625-20 mg / L, respectively, with matrix effects of -0.51%, -3.33%, and -16.08%, respectively. Within a quantitative concentration range, mixed solutions of three pesticides were prepared, resulting in four combinations. Qualitative analysis using SVM (Supervised Visualization) could identify the pesticide combinations in rice with a recognition rate exceeding 90%. Further PLSR (Plastic Pest Detection and Reduction) analysis could predict the concentration of each pesticide in the mixed pesticide solution, with an identification error of less than 5.36%. The QuEChERS method allows for rapid sample pretreatment, while UV-Vis spectrophotometry offers fast scanning speed. Combining these two methods improves analytical efficiency. By establishing a supervised model for pesticide composition using SVM and then predicting pesticide content using PLSR, rapid detection of pesticide residues is possible.

[0102] In addition to the embodiments described above, the present invention may have other implementations. All technical solutions formed by equivalent substitution or equivalent transformation fall within the protection scope claimed by the present invention.

Claims

1. A method for detecting pesticides in rice based on QuEChERS coupled with ultraviolet-visible spectrophotometry, characterized in that, Specifically, the following steps are included: S1: Rice samples treated using the QuEChERS method Take Wuchang organic rice, grind it into powder using a grinder, pass it through a 20-mesh sieve, and set it aside. Weigh the ground rice powder, put it into a plastic centrifuge tube, add deionized water, vortex mix, let it stand, add acetonitrile acetate solution, anhydrous magnesium sulfate, and sodium acetate, cover the centrifuge tube, shake it vigorously to loosen the lower layer of solid, then vortex mix again, put it into a centrifuge, centrifuge, pipette the supernatant into a centrifuge tube, add anhydrous magnesium sulfate and C18, vortex mix, centrifuge, dilute the supernatant with acetonitrile by 1 time to obtain rice matrix solution; S2: Preparation of standard solvent solution Standard working solutions of bensulfuron-methyl, propargyl and cypermethrin were prepared using acetic acid acetonitrile solution as solvent. S3: Preparation of matrix standard solution The rice matrix solution prepared in S1 was used to prepare matrix standard working solutions for three pesticides: bensulfuron-methyl, propargyl and cypermethrin. S4: Preparation of Mixed Pesticide Matrix Solution Using the orthogonal experimental design, four concentration levels were set for each pesticide within a quantitative range. Rice matrix solutions containing a mixture of bensulfuron-methyl, propargite, and cypermethrin were prepared, with bensulfuron-methyl solution as solution A, propargite solution as solution B, and cypermethrin solution as solution C. There were a total of four combinations: AB, AC, BC, and ABC. S5: Spectral Acquisition Methods Using a 0.5% (v / v) acetonitrile acetate solution as a reference, the test solution, i.e. the prepared matrix standard solution, was placed in the sample cell of a 1 cm cuvette, and the spectrum was scanned from 350 nm to 190 nm, with absorbance data collected at 1 nm intervals. S6: Qualitative Identification Methods for Pesticide Types SVM is used to build an intelligent recognition model. All data from the spectrum are used as independent variables. The combination of four pesticides corresponds to four types of samples. The Gaussian kernel function is used as the kernel function. The gamma and c parameters in the SVM model are optimized by grid search technology. Finally, the recognition rate of the validation set is used to compare the performance of the models. S7: Methods for Quantifying Pesticide Concentration Partial least squares regression was used to fit the training set samples for each combination, and quantitative analysis models for AB, AC, BC, and ABC combinations were established respectively. The peak values ​​corresponding to the entire band were used as independent variables and the concentration was used as dependent variables to establish quantitative analysis models and quickly detect the concentration of the prediction set samples for each combination.

2. The method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry according to claim 1, characterized in that: The acetic acid acetonitrile solution added in step S1 is an acetic acid acetonitrile solution with a volume concentration of 1%.

3. The method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry according to claim 1, characterized in that: In step S1, the centrifuge speed is 4000-4200 r / min, and the centrifugation time is 5-6 min.

4. The method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry according to claim 1, characterized in that: In step S2, the solvent is an acetonitrile acetate solution with a volume concentration of 0.5%.

5. The method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry according to claim 1, characterized in that: In step S2, the concentrations of the standard working solutions of bensulfuron-methyl, propargyl, and cypermethrin are 0.078, 0.156, 0.3125, 0.625, 1.25, 2.5, 5, 10, 20, 40, and 80 mg / L, respectively.

6. The method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry according to claim 1, characterized in that: In step S2, the matrix standard working solution concentrations of bensulfuron-methyl, propargyl, and cypermethrin are 0.3125, 0.625, 1.25, 2.5, 5, 10, 20, and 40 mg / L, respectively.

7. The method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry according to claim 1, characterized in that: In step S5, the test solution is tested three times, and the data obtained is processed based on the Matlab2010b platform.

8. The method for detecting pesticides in rice based on QuEChERS coupled with UV-Vis spectrophotometry according to claim 1, characterized in that: The specific parameter settings in step S6 are as follows: the range of C is limited to [2]. －8 ,2 8 ], the range of gamma is [2 －8 ,2 8 Five-fold cross-validation was performed, with step values ​​of 0.5 for both C and gamma. The final step value for the accuracy discretization of the parameter selection results was 4.5, and the model recognition rate was 90.625%.

Images