A Molecularly Imprinted Polymer Electrochemiluminescence Sensor Based on AuNPs@Ru-ZIF-8 for the Rapid Detection of Cyhalothrin Residues in Lycium barbarum L.

Abstract

Lycium barbarum L. is a widely used medicinal and edible Chinese medicinal material. However, with consumers’ heightened concern for health and food safety, pesticide residues have become one of the major challenges affecting its quality and safety. Cyhalothrin is a pyrethroid insecticide and a typical type of pesticide with excessive pesticide residues in Lycium barbarum L. Rapid detection of pesticide residues is an effective way to ensure the quality and safety of traditional Chinese medicinal materials. In this work, a molecularly imprinted polymer electrochemiluminescence (ECL) sensor based on gold nanoparticles (AuNPs)@Ru-ZIF-8 was constructed for rapid detection of cyhalothrin residues. The prepared cyhalothrin molecularly imprinted polymers (MIPs) were used as a recognition element and modified on the surface of a glassy carbon electrode (GCE) by an electrochemical polymerization method. AuNPs were utilized to promote the excitation of Ru(bpy)₃²⁺ and TPrA in the ECL system, which improved the observability of the light signal. The GCE modified with the metal–organic frameworks (MOFs) ZIF-8 was employed to increase the specific surface area and enhance the electron transfer capacity on the electrode, thereby improving the sensing sensitivity of the sensor. In addition, the luminescent reagent of Ru(bpy)₃²⁺ was introduced into the synthesis process of ZIF-8, which caused Ru(bpy)₃²⁺ to be tightly bound around it and enhanced the stability of the sensor. Under optimal conditions, the linear detection range of the sensor is 1 × 10⁻¹~1 × 10⁴ nM, with a limit of detection (LOD) of 10 pM. The accuracy of the ECL-MIP sensor has been verified through spiked recovery experiments and actual sample detection. This study has opened up a new approach to rapid detection of pesticide residues in traditional Chinese medicinal materials used for both food and medicine.

1. Introduction

Lycium barbarum L. is rich in natural active ingredients such as polysaccharides, carotenoids and betaine, and has significant therapeutic effects, high clinical safety, and multi-target health regulation characteristics such as oxidation, immunomodulation and anti-tumor auxiliary effects [1]. In recent years, Lycium barbarum L. has attracted much attention in the domestic and foreign pharmaceutical and health product markets. Studies have confirmed that the compound preparation of Lycium barbarum L. can enhance the immune function of cancer patients through mechanisms such as activating T cells, regulating cytokine secretion and improving chemotherapy tolerance. Its market demand has driven the rapid growth of large-scale planting and export trade in major production areas such as Ningxia and Xinjiang. However, intensive cultivation models have led to a significant increase in the incidence of Lycium barbarum L. pests and diseases, and in order to ensure yield, growers generally rely on the highly effective cyhalothrin insecticide.

However, the abuse of pesticides and lack of supervision have led to excessive cyhalothrin residues in Lycium barbarum L. [2,3], and it is difficult to completely degrade during processing. The acceptable daily intake (ADI) of cyhalothrin for humans ranges from 0 to 0.02 mg kg⁻¹ of body weight. In accordance with the National Food Safety Standards, the maximum residue limit (MRL) for traditional Chinese medicinal materials such as Lycium barbarum L. is set at 0.01 mg kg⁻¹. These residues can enter the human body through biological enrichment, interfering with neurotransmitter function and increasing the risk of chronic diseases. Examples include skin or mucosal damage, dizziness, and muscle tremors; in severe cases, symptoms such as convulsions and pulmonary edema may even develop. Moreover, cyhalothrin is particularly toxic to honeybees and highly toxic to aquatic arthropods like fish, shrimp, and crabs, thereby disrupting the aquatic food chain. All these issues have become key constraints on the sustainable development of the industry [4,5].

Traditional detection methods include chromatography, immunoassay, and spectroscopy. Chromatography features high sensitivity and good accuracy but suffers from high detection costs and reliance on professional personnel. Immunoassay is low-cost and easy to perform, yet it has low sensitivity and is susceptible to interference. Spectroscopy is low-cost, convenient, and rapid in detection, but its sensitivity is insufficient. Compared with traditional detection methods, the electrochemiluminescence (ECL) method is highly favored due to its high sensitivity and low background signal. For example, ECL biosensors have shown high sensitivity and specificity for acetamiprid detection [6]. ECL is an analytical technology that triggers chemiluminescence through electrochemical reactions on the surface of the electrode. Its essence is a synergistic process between electron transfer at the electrochemical interface and the energy release of the excited state of the luminescent material [7]. Its core mechanism is as follows. Driven by an external potential, a luminescent substance (such as a metal complex) fixed on the electrode surface undergoes a redox reaction, and a high-energy excited-state intermediate is generated through electron transfer. When the excited-state molecules relax to the ground state, photons are released to generate detectable optical signals [8,9]. With its wide dynamic response range, extremely low background interference (no exogenous light source) and ultra-high sensitivity (can reach the single molecule level), ECL has become a key technology in the fields of environmental pollutant screening, clinical diagnosis and food trace analysis [10].

Among many ECL systems, tri(2,2′-bipyridine)ruthenium(II)(Ru(bpy)₃²⁺) has become the most widely used luminescent material due to its excellent light stability, reversible electrochemical behavior and efficient luminescence quantum yield. To optimize its ECL intensity, it is often necessary to introduce co-reactants (such as triproamine, TPA) for constructing an “oxidation–reduction” dual circulation path. During anodization, TPA generates strong reducing free radicals (TPA*) during anodization, which can reduce the intermediate of Ru(bpy)₃³⁺ to the excited state Ru(bpy)₃²⁺, significantly amplifying the luminescent signal [11,12]. Based on this, researchers have developed a variety of ECL sensors driven by the Ru(bpy)₃²⁺/TPA collaborative system.

It is worth noting that in recent years, MIPs have been introduced into ECL sensing design as highly specific biometric elements [13]. ECL-MIP sensors have been developed in which molecularly imprinted membranes are modified via the sol-gel method to enhance their specificity and selectivity. The fabricated ECL-MIP sensors were successfully applied to the determination of 17β-estradiol in water samples, with a recovery rate ranging from 88.7% to 105.0% [14]. Such molecules can accurately capture target objects through structural changes, thereby regulating the electron transfer efficiency or steric hindrance effect of the ECL reaction interface and realizing the controllable output of the signal (such as “signal on” or “signal off” mode), greatly expanding the application potential of ECL technology in complex biological sample detection.

The metal–organic framework material of ZIF-8 (zeolite-type porous crystals constructed with Zn²⁺ coordinated with 2-methylimidazole) has unique advantages in building high-sensitivity sensors due to its high specific surface area (~1600 m² g⁻¹), adjustable pore size (0.34–1.16 nm) and surface functionalization capabilities [15,16,17]. Its three-dimensional pores can serve as an ideal carrier for molecular recognition sites. Through pre-designed ligands or post-modification strategies (such as the introduction of functional groups such as thiol groups and amino groups) [18], ZIF-8 can selectively adsorb targets (such as heavy metal ions, biomarkers or volatile organics) and change the photo/electrochemical response characteristics of the material through host–guest interaction to achieve signal transduction. For example, in the field of electrochemical sensing, electrode materials constructed in a composite with carbon nanotubes (CNTs), ZIF-8/CNTs, can improve the electrical conductivity and biocompatibility of the GCE and enhance the detection performance through the following synergistic mechanisms. At the signal amplification level, CNTs accelerate electron transfer, while ZIF-8’s high loading enhances redox-active site exposure [19,20]. In fluorescence sensing, quantum dots coated with ZIF-8 (QDs@ZIF-8) can form a “fluorescence switch” system. When the target object (such as antibiotic residue) enters the ZIF-8 channel, it undergoes electron transfer or energy resonance effects with the surface of the quantum dot, resulting in fluorescence quenching or enhancement, and the detection limit can reach the nano-molar level [21,22].

In this work, first, according to the synthesis path based on ZIF-8 [23], ECL luminescent substances were added during the synthesis of ZIF-8. This approach effectively prevented the loss of luminescent substances during the detection process. Secondly, the excellent adsorption characteristics of ZIF-8 were utilized, and gold nanoparticles (AuNPs) were encapsulated on the surface of ZIF-8, aiming to compensate for the disadvantages of poor conductivity, and the second was to promote the excitation of the ECL luminescent system [24,25]. In pesticide residue detection, AuNPs are widely employed to construct ECL sensors and aptasensor platforms, benefiting from their merits of signal amplification and enhanced specific recognition. They enable trace-level and accurate detection of multiple categories of pesticides, including organophosphates, carbamates, and pyrethroids. In addition, the GCE was modified by using multi-walled carbon nanotubes (MWCNTs), improving the conductivity of the electrode [26]. Molecular imprinted polymers were synthesized using cyhalothrin as the template molecule and methacrylic acid (MAA) as a functional monomer. When the molecularly imprinted polymer specifically recognizes and captures pesticide molecules, the intensity of ECL decreases, accurately measuring the concentration of cyhalothrin based on changes in the light signal [27,28,29].

2. Experimental Section

2.1. Reagents and Instruments

The reagents and instrument types used during the experiments are given in the Supplementary Materials. Origin (v2024) was used in this study.

2.2. Preparation of AuNPs

The synthesis of AuNPs was conducted following the sodium citrate reduction method [30] with experimental modifications. Initially, 0.01 g of chloroauric acid trihydrate (HAuCl₄·3H₂O) was dissolved in 100 mL of ultrapure water to prepare a 0.01% (w/v) HAuCl₄ solution. This solution was transferred into a 250 mL sterilized beaker and heated to boiling point under continuous magnetic stirring. Subsequently, 10 mL of 1% (w/v) trisodium citrate solution was rapidly injected into the boiling mixture under vigorous agitation. An immediate chromatic transition from pale yellow to burgundy was observed, indicating the reduction of auric ions to metallic gold nanoparticles. The reaction system maintained thermal agitation for 15 min until achieving chromatic stability. Post-heating continuous stirring facilitated cooling to ambient temperature. The resultant AuNP colloid was quantitatively transferred to a 100 mL volumetric flask and diluted to 50 mL with ultrapure water, designated as Solution A.

2.3. Preparation of AuNPs@Ru-ZIF-8

For the synthesis of Ru-incorporated ZIF-8 (Ru-ZIF-8) [23], an electrochemical luminescent complex (Ru(bpy)₃Cl₂·6H₂O) was integrated during the solvent-based crystallization process. Solution B was prepared by dissolving 2.97 g zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O) and 5 mg Ru(bpy)₃Cl₂·6H₂O (hereafter referred to as the Ru-complex) in 100 mL methanol. Solution C contained 3.28 g of 2-methylimidazole (C₄H₆N₂) dissolved in 100 mL of methanol. Under ambient conditions, Solutions A and B were homogeneously mixed under constant magnetic stirring at 600 rpm, followed by the rapid injection of 20 mL Solution C. The ternary mixture underwent continuous agitation for 24 h, during which progressive turbidity development signaled ZIF-8 crystal nucleation. After 30 min of gravitational sedimentation, phase separation was enhanced through centrifugal processing (10 min at 5000 rpm, 4 °C). The pale red precipitate underwent three successive washings with methanol, each followed by centrifugation. The purified product was re-dispersed in 100 mL ultrapure water and stored in light-proof containers under ambient conditions.

2.4. Construction of Electrochemiluminescence Aptasensor

GCE underwent sequential polishing using 0.3 µm and 0.05 µm alumina slurries under mechanical abrasion until achieving a mirror-like surface finish. Post-polishing, the electrode was rigorously rinsed with ultrapure water to eliminate residual alumina particles. Subsequent purification involved sequential immersion in ethanol and deionized water, each subjected to 5 min ultrasonication for organic/ionic contaminant removal, followed by nitrogen stream drying. A homogeneous AuNPs@Ru-ZIF-8 composite suspension was prepared through 30 min ultrasonic dispersion, from which a 5 mL aliquot was separated. Precise deposition of 5 µL suspension onto the GCE surface yielded an AuNPs@ZIF-8 modified electrode after ambient drying. For molecular imprinting preparation, 4.5 mg of cyhalothrin standard was dissolved in 10 mL of acetonitrile to formulate a 1 mM template solution under continuous magnetic agitation. Sequential addition of 5 mg MAA and 20 mg ethylene glycol dimethacrylate (EGDMA) into the template solution achieved complete monomer dissolution. The polymerization precursor solution underwent 10 min of degassing via ultrasonication to remove dissolved oxygen. A three-electrode system was configured with the modified GCE as the working electrode, platinum wire as the counter electrode, and Ag/AgCl as the reference electrode. Electropolymerization was performed through 10 cyclic voltammetric scans (−1.0 V to 1.5 V vs. Ag/AgCl, 50 mV s⁻¹ scan rate). Post-polymerization, the electrode received thorough ultrapure water rinsing to eliminate unreacted monomers. Template molecule extraction involved triple methanol ultrasonication (10 min per cycle), followed by nitrogen drying. The resultant sensor, designated as the ECL-MIP sensor [7], was stored in light-proof containers under an inert atmosphere. All the above steps are illustrated in Scheme 1.

2.5. Electrochemiluminescence Measurement

Electrochemical measurements included cyclic voltammetry (CV), differential pulse voltammetry (DPV) and ECL analysis. CV measurements were performed in the potential range of −0.2 to 0.6 V and at a scan rate of 50 mV s⁻¹. DPV measurements were performed in the potential range of −2 to 0 V with a pulse period of 0.5 s. A total of 0.1 M tri-n-propylamine (TPrA) was added to the PBS solution as a co-reactor of Ru(bpy)₃²⁺. The constructed ECL-MIP sensor was used as a working electrode, the platinum wire electrode was used as a counter electrode, and the Ag/AgCl electrode was used as a reference electrode to construct a three-electrode ECL detection system. The voltage of the electrochemiluminescence workstation was set in the voltage range of −0.2 to 0.6 V. The CV scan was performed at a scanning rate of 50 mV s⁻¹, and the photovoltaic high voltage was set at 600 V. The ECL-MIP sensor was used as the working electrode.

2.6. Sample Pretreatment

The sample was first subjected to initial treatment. A total of 5 g of crushed Lycium barbarum L. was taken, and 50 mL of ultrapure water was added to the sample and extracted by ultrasound at a constant temperature for 30 min and then centrifuged at 5000 rpm for 5 min to obtain the aqueous phase as well as the solid residue. The solid residue was then added to 10 mL of ethyl acetate for the second extraction, and the organic phase was obtained by centrifugation at 5000 rpm for 5 min after 30 min of ultrasonic extraction. The organic phase was transferred to a centrifuge tube containing 50 mg of C18 adsorbent and 150 mg of anhydrous MgSO₄ for dispersive solid-phase extraction. C18 adsorbs volatiles through hydrophobicity, and anhydrous MgSO₄ removes residual water (vortex shaking for 2 min followed by centrifugation at 5000 rpm for 5 min), and the supernatant was collected. The purified supernatant was concentrated to 5 mL by nitrogen blowing to obtain the cyhalothrin assay sample, which was preserved.

3. Results and Discussions

3.1. Phase and Structure Analysis of Experimental Materials

The morphology of Ru-ZIF-8 was observed by scanning electron microscopy (SEM), and AuNPs were further observed by transmission electron microscopy (TEM). The elemental composition of the materials was analyzed using energy dispersion X-ray spectroscopy (EDS) [31]. EDS was used to analyze the elemental composition of the materials. The bonding between the physical phases and structures of the synthesized materials was analyzed using X-ray photoelectron spectroscopy (XPS).

Figure 1 shows the TEM image of AuNPs. Based on the TEM observations, it can be seen that the AuNPs are roughly spherical and dispersed in the field of view. The size distribution of the particles is uneven, and some of the particles have an aggregation phenomenon, forming small clusters. The scale bar in the lower right corner of the image is 100 nm, and the size of the particles is estimated to be about 10–20 nm. It can be observed from the image that most of the particles have a more complete morphology, but some of the particles have irregularities at the edges, which may be related to the adsorption of surfactants during the preparation process or defects during the growth of the particles. However, they can be completely used to construct ECL-MIP sensors.

Figure 2 illustrates the SEM images of (A) ZIF-8, (B) Ru-ZIF-8, and (C) MIPs/AuNPs@Ru-ZIF-8, presenting the surface morphology characteristics of the different materials. Figure 2D–F illustrate the EDS elemental distribution of Zn, Ru, and Au. The EDS images show the distribution of a specific element on the sample surface by means of color coding, where each bright spot represents the location of the detected specific element. Figure 2A clearly shows the typical morphology of the ZIF-8 material. The ZIF-8 crystals show a regular rhombic dodecahedral structure with smooth crystal surfaces and clear, sharp edges. The crystal size distribution is relatively uniform, indicating that the synthesized ZIF-8 has good crystallinity and homogeneity. Figure 2B shows the morphology of Ru-ZIF-8. Compared with pure ZIF-8, the crystal morphology of Ru-ZIF-8 basically maintains the characteristics of a rhombic dodecahedron, but the surface of the crystals is slightly roughened, which may be due to the introduction of Ru elements having a certain effect on the growth of ZIF-8 crystals. Nevertheless, the crystal structure of Ru-ZIF-8 still maintains a high degree of integrity. Figure 2C shows the morphology of MIPs/AuNPs@Ru-ZIF-8. Compared with Ru-ZIF-8, the surface roughening of MIPs/AuNPs@Ru-ZIF-8 was further increased, and obvious agglomeration appeared on some surfaces. This suggests that the polymerization of MIPs significantly affected the crystal morphology of Ru-ZIF-8, which may have led to the alteration of the crystal surface structure and the appearance of agglomeration behavior. Overall, the SEM images clearly demonstrate the surface morphology features of ZIF-8, Ru-ZIF-8 and MIPs/AuNPs@Ru-ZIF-8, revealing the effects of the introduction of Ru elements and MIPs on the ZIF-8 crystal structure. Figure 2D shows the distribution of Zn elements, with the highlights appearing yellow in color and uniformly dispersed on a black background. This indicates that the Zn element in the sample is uniformly distributed throughout the field of view, corroborating the existence of the ZIF-8 main frame nodes. The uniform distribution of the bright spots implies that the element is highly dispersed in the sample. Figure 2E shows the distribution of the Ru element in the sample, with the bright spots appearing orange–red in color and again evenly dispersed on the black background. The different colors of the bright spots in (E) compared to (D) suggest that another specific element, Ru, is distributed in the sample, and the uniform distribution of the bright spots suggests that the element Ru is also well dispersed in the sample. Figure 2F shows the elemental distribution of Au, with the bright dots appearing blue or cyan in color and evenly dispersed on a black background. The color of the highlights in (F) is again different compared to (D) and (E), suggesting that another specific element, Au, is distributed in the sample, and the uniform distribution of the highlights suggests that this element is also well dispersed in the sample. Overall, the EDS elemental distribution map clearly shows the distribution of three different elements. The uniform distribution of the bright dots indicates that these elements are well dispersed in the respective samples. By comparing the colors of the bright spots in different samples, the different elemental compositions of the samples can be distinguished, and the EDS results provide an important basis for further analysis of the elemental composition and distribution of the samples.

The morphology of the composites was examined using SEM, while the specific dimensions were further observed using TEM. The elemental compositions of the materials were analyzed using EDS. Additionally, the elemental composition of the composites was determined through ultraviolet–visible spectroscopy analysis (UV-vis) and XPS to investigate the bonding between the physical phase and structure of the synthesized materials in this study.

The surface elemental composition and chemical states of AuNPs@Ru-ZIF-8 composites were analyzed in detail by XPS. Split-peak fitting is a common technique used in XPS analysis to decompose a complex spectrum into multiple components, thus obtaining more detailed information about the chemical state of a particular element. By analyzing the fine spectra of individual elements, the chemical states and interactions of the elements in the Ru@ZIF-8 material can be determined. The full spectrum in Figure 3A confirms the presence of the elements gold (Au), carbon (C), ruthenium (Ru), nitrogen (N) and zinc (Zn) in the material. Figure 3B shows the fine spectrum of elemental N in the 1s orbital, and the split-peak fitting plot gives three states of elemental N. These are Zn-N at the binding energy of 339.5 eV, which proves the main structure of ZIF-8, and uncoordinated nitrogen at 398.5 and 400.7 eV [15], i.e., nitrogen contained in imidazole ligands. Analysis of the spectral peak areas indicates that most imidazole ligands are involved in constructing the overall skeleton of ZIF-8. Figure 3C shows the fine spectrum of elemental Au in the 4f orbital; the peak at 84.1 eV represents the AuNPs in the sample, while the peak at 87.2 eV represents the oxidation of some of the AuNPs, which is an unavoidable phenomenon in the analysis of samples. From the fine spectrum of element C in the 1s orbital in Figure 3D, the peaks at 284.8, 286.4 and 287.9 eV can be clearly observed, which represent the C-C, C-N and C-O bonds, respectively [15], and their presence also proves that the synthesis of ZIF-8 material is successful. The fine spectral image of element Ru in the 3p orbital clearly shows two peaks at 458.9 and 462.1 eV, as shown in Figure 3E, in which the smaller area of the peaks represents the luminescent material Ru(bpy)₃²⁺, while the larger area may be the divalent Ru after reduction by AuNPs. Figure 3F shows the XPS image of element Zn, and the peak at 1022.1 eV is the XPS image of the ZIF-8 material. The peak at eV is Zn²⁺, which is clearly observed as a node of the ZIF-8 framework.

3.2. Electrochemical and Electrochemiluminescent Validation

Each assembly process of the ECL-MIP sensor was analyzed in detail using CV, DPV and ECL, respectively. As shown in Figure 4A, the CV plots demonstrate the curve changes in GCE at different modification stages, which clearly reveals the effect of each step of modification on the peak current.

As shown in Figure 4A, firstly, the peak current significantly increased after modifying MWCNTs (curve a), and the peak current of GCE reached 150 μA after the modification of MWCNTs. This indicates that MWCNTs greatly enhanced the electron transfer ability of the electrodes by virtue of their high surface area and high conductivity, which laid an excellent basic performance for the sensor. Secondly, after the addition of AuNPs@RuZIF-8 on top of MWCNTs (curve b), the peak currents decreased to ~140 μA (0.25 V) for the anode and −140 μA (0.05 V) for the cathode, which can be attributed to the poor electrical conductivity of ZIF-8 partially offsetting the electrical conductivity of both MWCNTs and AuNPs. However, the peak current was lower than that of the simple Ru-ZIF-8 (curve c). This is attributed to the poorer conductivity of ZIF-8 partially offsetting the conductivity of MWCNTs and AuNPs. However, compared with Ru-ZIF-8 alone (curve c), the addition of AuNPs still increases the current, reflecting the improved conductivity of AuNPs. Then, after immobilization of MIPs by electropolymerization (curve d), the peak currents further decreased to about 117 μA (0.25 V) at the anode and −117 μA (0.05 V) at the cathode, which may be attributed to the insulating nature of MIPs hindering the electron transfer and also proved that MIPs were successfully immobilized to provide the basis for specific detection. Subsequently, after the addition of cyhalothrin (curve f), its binding to the specific sites of MIPs continued to reduce the peak currents to about 95 μA (0.25 V) at the anode and −95 μA (0.05 V) at the cathode, reflecting that the molecularly blocked sites of cyhalothrin limited the electron transfer and verified the specific detection capability of the sensor. Finally, the selectivity of the sensor was further confirmed by non-imprinted polymer (NIP) control experiments, in which NIPs had no significant effect on the electrochemical response (curve e), ruling out non-specific interference.

DPV analysis was done simultaneously for the assembly process of the ECL-MIP sensor, as shown in Figure 4B. The DPV signal trends throughout the assembly process had the same results as in the CV analysis (Figure 4A). As shown in the ECL curves in Figure 4C, the peak current is significantly enhanced after modification with MWCNTs (curve a). Furthermore, in comparison with Ru-ZIF8/MWCNTs/GCE without AuNPs modification (curve c), the ECL intensity of AuNPs@Ru-ZIF8/MWCNTs/GCE modified with AuNPs (curve b) is distinctly increased. Subsequent electropolymerization of MIPs resulted in a decrease in ECL signal (curve d). The ECL intensity of the eluted MIPs/AuNPs@Ru-ZIF-8/MWCNTs/GCE (curve d) was not significantly different from that of the NIPs/AuNPs@Ru-ZIF-8/MWCNTs/GCE obtained by electric polymerization without the addition of template molecules (curve e).

3.3. Optimizing Experimental Parameters

The experiments mainly examined the effects of the electropolymerization cycle, the pH of the polymerization solution, the molar ratio of template molecules to functional monomers, and the elution time on the sensors of the ECL-MIPs, and the error lines in the images indicate the error intervals under the same experimental conditions three times.

The experimental results showed that the ECL intensity showed an increasing and then decreasing trend with the increase in the number of polymerization cycles, as shown in Figure 5A. Specifically, the ECL intensity gradually increases when the polymerization period increases from 5 to 15, which indicates that the responsiveness of the sensor to the target molecules is enhanced with the increase in the polymer film thickness. However, when the polymerization period exceeds 15 and continues to increase to 20, the ECL intensity begins to decrease, which indicates that an excessively thick polymer membrane is instead detrimental to the sensitivity of the sensor. Too thin polymer membranes are easily damaged during the elution process, leading to insufficient blotting cavities and thus reduced sensitivity, while too thick polymer membranes may hinder the complete removal of template molecules, which also affects sensitivity. Therefore, through the experimental data, 15 cycles were determined to be the optimal cycle for electropolymerization, and the molecularly imprinted polymer membranes prepared under this condition had the best performance.

As shown in Figure 5B, the variation in ECL intensity with pH shows a typical bell-shaped curve, indicating the existence of an optimal pH value for maximum sensor response. Specifically, a significant upward trend in the ECL intensity was observed during the gradual increase in pH from 6.5 to 8.0 and peaked at pH 8.0. This phenomenon suggests that the polymerization reaction effectively promotes the formation of high-quality molecularly imprinted polymer films in this pH range, which improves the recognition of target analytes by the sensor. However, the ECL intensity decreased rapidly when the pH exceeded 8.0 and continued to increase to 9.0. This suggests that the structure of the polymer film is distorted at too high pH conditions, leading to a decrease in sensor sensitivity. The molar ratio of template molecules to functional monomers has a significant effect on the formation of molecularly imprinted cavities.

As shown in Figure 5C, the ECL intensity showed a clear peak with the molar ratio. At lower molar ratios, there is an excess of functional monomers, resulting in an insufficient number of imprinted cavities, which reduces the binding ability of the sensor to the target molecule. As the molar ratio increases, the number of imprinted cavities gradually increases and the ECL intensity rises. However, when the molar ratio is too high, there is an excess of template molecules, which may lead to uneven distribution of the imprinted cavities or even residual template molecules, thus reducing the sensitivity of the sensor. The experimental data showed that the MIPs formed by polymerization with a molar ratio of 1:4 provided the best conditions for the formation of imprinted cavities and achieved the optimal ECL signal response. The elution process is a critical step to remove the template molecules from the polymer membrane, which directly affects the sensor’s ability to recognize the target analyte.

As shown in Figure 5D, a significant increase in ECL intensity with elution time was observed for shorter elution times (10 to 20 min). This suggests that the residual template molecules in the polymer membrane gradually decrease with increasing elution time, thus releasing more imprinted cavities and improving the binding ability of the sensor to the target analyte. However, when the elution time exceeded 25 min, the ECL intensity reached a plateau and no longer increased significantly. This suggests that the template molecules have been basically completely removed under this condition, and further prolongation of the elution time has a limited effect on the enhancement of the sensor performance. However, excessive elution time may adversely affect the structure of the polymer membrane. Excessive elution may lead to the destruction of the pore structure of the polymer membrane or even the collapse of the imprinted cavity, thus reducing the sensitivity and selectivity of the sensor. Therefore, an elution time of 25 min was determined to be the optimal condition. The condition effectively removes the template molecules while maintaining the integrity of the polymer membrane, thus achieving the highest ECL signal response.

In summary, 15 cycles of electric polymerization, a polymerization solution pH of about 8.0, a molar ratio of template molecules to functional monomers of 1:4, and an elution of 25 min were selected as the optimal experimental conditions.

3.4. Aptasensor Performance Analysis

The detection performance of the ECL-MIP sensors was further tested under optimal detection conditions. The test was performed by adding nine sets of different concentrations of cyhalothrin to the ECL-MIP sensor, and the test results are shown in Figure 6A. In the figure, the concentration progressively decreases from the first curve to the last one, with the blue curve corresponding to the highest concentration and the red one the lowest. The concentration of cyhalothrin (C_Cyh) showed a negative correlation with the intensity of ECL, I, in the range of 1 × 10⁻¹ to 1 × 10⁴ nM (Figure 6B), with a correlation coefficient R² of 0.992 and a regression equation of I = 9931 − 1528 LgC_Cyh. The limit of detection (LOD) was 10 pM.

3.5. Stability, Reproducibility and Specificity

Eighteen identical ECL-MIP sensors were pre-prepared to evaluate their stability under optimal experimental conditions. Tests were performed using cyhalothrin at a concentration of 1 nM. The 18 ECL-MIP sensors were randomly divided into six groups. One group was randomly selected by dropping 3 μL of 1 nM cyhalothrin standard solution onto three ECL-MIP sensors and waiting for natural drying to perform ECL analysis. The remaining five groups of ECL-MIP sensors were randomly sampled every 3 days, and the test results are shown in Figure 7A. After comparing the six sets of data, the intensity of ECL varied within 600 intervals within 18 days, with an RSD of 2.82%. The results indicate that the ECL-MIP sensors have good stability. Four ECL-MIP sensors were prepared under optimal experimental conditions, and 3 μL of 1 nM cyhalothrin standard solution was added and waited for drying to be tested immediately, and the change in ECL was utilized to assess the repeatability of the ECL-MIP sensors. The experimental results are shown in Figure 7B; the relative standard deviation RSD = 1.06% for the four sensors, indicating that the ECL-MIP sensors have good repeatability. The specificity of the ECL-MIP sensors was separately tested using different pesticides: (α) mixed pesticides, (β) cyhalothrin, (γ) acetamiprid, (δ) carbofuran, (ε) trichlorfon, (ζ) propoxypyrin, (η) malathion, (θ) chlorpyrifos, (ι) thiamethoxam, (κ) buthionate and (λ) imidacloprid. As the results in Figure 7C show, the ECL intensities of γ, δ, ε, ζ, η, θ, ι, κ, and λ are much higher than those of α and β, while the ECL intensities of α and β are not similar to each other, which fully demonstrates the good specificity of the ECL-MIP sensors.

3.6. Analysis of Lycium barbarum L.

The analysis results of the LC-MS/MS method [32] were compared. The standard sample addition method was used to add 0, 10 and 100 ng mL⁻¹ diluted cyhalothrin solutions to the pretreated Lycium barbarum L. samples, and the detection was carried out using the ECL-MIP sensor. The experimental data are shown in

Table 1. Test results of cyhalothrin content in samples (n = 3).

Added Amount (ng mL−1)	Method	Detection Value (ng mL−1)	RSD (%)	Recovery Rate (%)
0	LC-MS/MS	0	-	-
	ECL-MIPs	0	-	-
10	LC-MS/MS	10.16	3.13	101.60
	ECL-MIPs	10.23 ± 0.28	4.00	102.30
100	LC-MS/MS	107.02	5.10	107.02
	ECL-MIPs	106.50 ± 4.60	4.55	116.50

, and the recoveries of the ECL-MIP sensor ranged from 100.08% to 116.50%. Referring to the results of the LC-MS/MS method, it can be proved that the prepared sensor has the same detection ability for real samples. According to the sample processing method in Chapter 2, Lycium barbarum L. from different origins was grouped and independently observed three times to obtain the average ECL values. The ECL was calculated according to the standard curve equation, and the results are shown in

Table 2. Realistic ECL results of samples from different sources.

Source	ECL Detection Value	Detection Value (ng mL−1)
Ningxia	9170	1.42
Gansu	9020	1.79
Qinghai	9200	1.34
Xinjiang	8950	1.98

. The four different origins of Lycium barbarum L. met the requirements of the current Chinese national standard GB2763-2021 Minimum Detectable Amounts (MDAMs) in foods. In addition, research articles published within the last five years were compared. The ECL sensor showed a larger linear range and lower LOD compared to colorimetric and fluorescence methods, and the present work had a lower LOD compared to EC and previous ECL sensors (see

Table 3. Comparison of ECL-MIP sensors with previously published cyhalothrin detection.

Methods	Linear Range	LOD	References
Colorimetric	2.22~800.25 nM	6.25 nM	[33]
Fluorescence	0.10~3000 μM	24 nM	[10]
EC	0~1.50 μM	10 nM	[28]
ECL	1 × 10−1~1 × 104 nM	10 pM	this work

).

4. Conclusions

This experiment combined MIT and ECL analysis methods to successfully construct an ECL-MIP sensor for detection of cyhalothrin residues in Lycium barbarum L. The ECL-MIP sensor used ZIF-8 with an extremely high specific surface area and ordered microporous structure as the basic framework, and the luminescent material Ru(bpy)₃²⁺ was simultaneously compounded in ZIF-8. The improved Ru-ZIF-8 composite avoided the Ru(bpy)₃²⁺ loss in the electrochemiluminescence system of Ru(bpy)₃²⁺ and TPrA systems. The introduction of AuNPs had enhanced the ability of Ru-ZIF-8 to catalyze the luminescence reaction of the Ru(bpy)₃²⁺ and TPrA system. MIPs produced by electric polymerization had specificity in identifying cyhalothrin. The developed sensor can linearly operate in the detection range of 1 × 10⁻¹~1 × 10⁴ nM and has an LOD of 10 pM (S/N = 3). In actual sample analysis, the cyhalothrin recovery rate of ECL-MIP sensors ranges from 97% to 115.5%. The prepared ECL-MIP sensors have good selectivity, repeatability and stability, providing a reliable method for detecting cyhalothrin residues in Lycium barbarum L.