Novel Dual-Signal SiO2-COOH@MIPs Electrochemical Sensor for Highly Sensitive Detection of Chloramphenicol in Milk

In view of the great threat of chloramphenicol (CAP) to human health and the fact that a few producers have illegally used CAP in the food production process to seek economic benefits in disregard of laws and regulations and consumer health, we urgently need a detection method with convenient operation, rapid response, and high sensitivity capabilities to detect CAP in food to ensure people’s health. Herein, a molecularly imprinted polymer (MIP) electrochemical sensor based on a dual-signal strategy was designed for the highly sensitive analysis of CAP in milk. The NiFe Prussian blue analog (NiFe-PBA) and SnS2 nanoflowers were modified successively on the electrode surface to obtain dual signals from [Fe(CN)6]3−/4− at 0.2 V and NiFe-PBA at 0.5 V. SiO2-COOH@MIPs that could specifically recognize CAP were synthesized via thermal polymerization using carboxylated silica microspheres (SiO2-COOH) as carriers. When the CAP was adsorbed by SiO2-COOH@MIPs, the above two oxidation peak currents decreased at the same time, allowing the double-signal analysis. The SiO2-COOH@MIPs/SnS2/NiFe-PBA/GCE sensor used for determining CAP was successfully prepared. The sensor utilized the interactions of various nanomaterials to achieve high-sensitivity dual-signal detection, which had certain innovative significance. At the same time, the MIPs were synthesized using a surface molecular imprinting technology, which could omit the time of polymerization and elution and met the requirements for rapid detection. After optimizing the experimental conditions, the detection range of the sensor was 10−8 g/L–10−2 g/L and the limit of detection reached 3.3 × 10−9 g/L (S/N = 3). The sensor had satisfactory specificity, reproducibility, and stability, and was successfully applied to the detection of real milk samples.


Introduction
Chloramphenicol (CAP) is a broad-spectrum antibiotic commonly used to treat bacterial infections caused by Gram-positive and Gram-negative bacteria [1]. Unfortunately, CAP ultimately enters and causes serious harm to the body [2,3]. In view of the huge threat of CAP to human health, many countries such as the United States, China, and Canada have issued decrees to explicitly prohibit the use of CAP in the food production process [4]. It is worrying that there are still many producers who pursue economic benefits by illegally using CAP in the food production process, regardless of the regulations and the health of consumers. Therefore, a detection method with convenient operation, rapid response, and high sensitivity capabilities is needed to detect CAP in food to preserve people's health.
Among the various detection methods, electrochemical methods have received extensive attention due to their fast detection speeds, ease of operation, and low detection

Experimental Section
The following components are introduced in the Supplementary Materials, namely the instruments and medicines

Preparation of SiO 2 -COOH@MIPs/SnS 2 /NiFe-PBA/GCE Sensor
We synthesized NiFe-PBA, SnS 2 , and SiO 2 -COOH@MIPs according to the steps in Scheme 1a-c (the specific synthesis process is described in detail in the Supplementary Materials). For the sake of obtaining a 1.0 mg/mL NiFe-PBA solution, the NiFe-PBA was dissolved in ultrapure water and sonicated for 15 min. Following the same operation steps, a 1.0 mg/mL SnS 2 nanoflower solution was attained. Then, 10 µL of NiFe-PBA solution was dropped on the surface of the polished glassy carbon electrode to dry naturally at room temperature. After the NiFe-PBA solution was completely dry, 10 µL of SnS 2 nanoflower solution was modified onto the NiFe-PBA/GCE surface. Ultimately, the sensor was obtained by modifying 10 µL of 10 mg/mL SiO 2 -COOH@MIPs on the SnS 2 /NiFe-PBA/GCE surface (Scheme 1d).

Preparation of SiO2-COOH@MIPs/SnS2/NiFe-PBA/GCE Sensor
We synthesized NiFe-PBA, SnS2, and SiO2-COOH@MIPs according to the steps in Scheme 1a-c (the specific synthesis process is described in detail in the Supplementary Materials). For the sake of obtaining a 1.0 mg/mL NiFe-PBA solution, the NiFe-PBA was dissolved in ultrapure water and sonicated for 15 min. Following the same operation steps, a 1.0 mg/mL SnS2 nanoflower solution was attained. Then, 10 μL of NiFe-PBA solution was dropped on the surface of the polished glassy carbon electrode to dry naturally at room temperature. After the NiFe-PBA solution was completely dry, 10 μL of SnS2 nanoflower solution was modified onto the NiFe-PBA/GCE surface. Ultimately, the sensor was obtained by modifying 10 μL of 10 mg/mL SiO2-COOH@MIPs on the SnS2/NiFe-PBA/GCE surface (Scheme 1d).

Characterization of the Prepared Materials
The morphological features of NiFe-PBA were characterized using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As shown in Figure 1A,B, the as-synthesized NiFe-PBA material had a typical nanocube structure with a smooth surface and uniform size, with an average size of 100 nm. By comparing the main diffraction peaks in the x-ray diffraction (XRD) detection results for NiFe-PBA with the

Characterization of the Prepared Materials
The morphological features of NiFe-PBA were characterized using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As shown in Figure 1A,B, the as-synthesized NiFe-PBA material had a typical nanocube structure with a smooth surface and uniform size, with an average size of 100 nm. By comparing the main diffraction peaks in the x-ray diffraction (XRD) detection results for NiFe-PBA with the data for PB (JCPDS No. 73-0687), we could more intuitively know that the NiFe-PBA had been successfully prepared and had high crystallinity ( Figure 1C).  The ingredients in NiFe-PBA were further analyzed via X-ray photoelectron spectroscopy (XPS). The full measured spectrum of NiFe-PBA ( Figure 1D) showed that there were five elements (Ni, Fe, C, N, and O) in the NiFe-PBA. Six peaks can be observed in the Figure 1E. The peaks of Ni 2+ 2p 3/2 and Ni 2+ 2p 1/2 appeared at 856.2 eV and 873.9 eV, respectively, and the two peaks at 858.3 eV and 875.1 eV belonged to Ni 3+ 2p 3/2 and Ni 3+ 2p 1/2 , respectively. The remaining two peaks were noticeable satellite peaks (863.2 eV and 880.7 eV). In the spectrum of Fe 2p, five peaks could be observed ( Figure 1F). The peaks of Fe 2+ 2p 3/2 and Fe 2+ 2p 1/2 appeared at 708.4 eV and 720.9 eV, respectively, and the two peaks at 709.7 eV and 723.4 eV belonged to Fe 3+ 2p 3/2 and Fe 3+ 2p 1/2 , respectively. The remaining peak was a more obvious satellite peak (711.1 eV), which was consistent with previous reports in the literature [29]. The XPS spectra demonstrated that the two elements existed in NiFe-PBA in Ni 3+/2+ and Fe 3+/2+ valence states. The three peaks shown in the C 1s spectrum ( Figure 1G) at 284.8, 286.4, and 294.1 eV correspond to C=C, C-N, and C=O, respectively. There was only one peak in the N 1s spectrum at 398 eV ( Figure 1H).

SnS 2 Nanoflowers
In this study, SEM, TEM, mapping, XPS, XRD, and the Randles-Sevcik equation were used to fully characterize and analyze the properties of SnS 2 nanoflowers. The relevant experimental data are shown in Supplementary Material 3.1.2.

Electrochemical Investigation of SiO 2 -COOH@MIPs/SnS 2 /NiFe-PBA/GCE
The electrochemical performances of electrodes modified with different materials were characterized using the differential pulse voltammetry (DPV) technique. As shown in Figure 2A, only one oxidation peak appeared in the DPV curve of the bare electrode. Dropping the poorly conductive NiFe-PBA onto the bare electrode surface resulted in a decrease in the oxidation peak current at 0.20 V and the appearance of a new oxidation peak at 0.50 V. However, the peak value at 0.50 V was too minute to achieve the purpose of detection. Therefore, we introduced SnS 2 nanoflowers with good electrical conductivity, and the three-dimensional flower-like structure of the SnS 2 nanoflowers increased the specific surface area and provided more attachment sites. Notably, when SnS 2 nanoflowers were decorated on the NiFe-PBA/GCE surface, the peak current of the oxidation peaks at 0.20 V and 0.50 V was significantly enhanced and reached a maximum value. Since the as-prepared SiO 2 -COOH @MIPs did not have electrical conductivity, the oxidization peak-to-peak current at [Fe(CN) 6 ] 3−/4− and NiFe-PBA were both weakened when the SiO 2 -COOH@MIPs were decorated on the SnS 2 /NiFe-PBA/GCE surface. However, the surfaces of SiO 2 -COOH@MIPs had abundant imprinted cavities similar in structure and size to the template molecule after elution, which provided channels for electron transfer. The two peak currents were less weakened. When the as-prepared electrode was placed in a certain concentration of CAP solution, the CAP was adsorbed specifically to fill the imprinted cavities on the SiO 2 -COOH@MIPs, which hindered the electron transfer. The peak currents of the two oxidation peaks were significantly reduced. The SiO 2 -COOH@MIPs/SnS 2 /NiFe-PBA/GCE sensor had been successfully fabricated as demonstrated by the above tests. Combined with Figure 2B, this allowed us to better prove the above conclusion.

Electrochemical Investigation of SiO2-COOH@MIPs/SnS2/NiFe-PBA/GCE
The electrochemical performances of electrodes modified with different materials were characterized using the differential pulse voltammetry (DPV) technique. As shown in Figure 2A, only one oxidation peak appeared in the DPV curve of the bare electrode. Dropping the poorly conductive NiFe-PBA onto the bare electrode surface resulted in a decrease in the oxidation peak current at 0.20 V and the appearance of a new oxidation peak at 0.50 V. However, the peak value at 0.50 V was too minute to achieve the purpose of detection. Therefore, we introduced SnS2 nanoflowers with good electrical conductivity, and the three-dimensional flower-like structure of the SnS2 nanoflowers increased the specific surface area and provided more attachment sites. Notably, when SnS2 nanoflowers were decorated on the NiFe-PBA/GCE surface, the peak current of the oxidation peaks at 0.20 V and 0.50 V was significantly enhanced and reached a maximum value. Since the as-prepared SiO2-COOH @MIPs did not have electrical conductivity, the oxidization peak-to-peak current at [Fe(CN)6] 3-/4-and NiFe-PBA were both weakened when the SiO2-COOH@MIPs were decorated on the SnS2/NiFe-PBA/GCE surface. However, the surfaces of SiO2-COOH@MIPs had abundant imprinted cavities similar in structure and size to the template molecule after elution, which provided channels for electron transfer. The two peak currents were less weakened. When the as-prepared electrode was placed in a certain concentration of CAP solution, the CAP was adsorbed specifically to fill the imprinted cavities on the SiO2-COOH@MIPs, which hindered the electron transfer. The peak currents of the two oxidation peaks were significantly reduced. The SiO2-COOH@MIPs/SnS2/NiFe-PBA/GCE sensor had been successfully fabricated as demonstrated by the above tests. Combined with Figure 2B, this allowed us to better prove the above conclusion.
The different assembly steps of the sensor were further characterized using EIS, and the results are shown in Figure 2C. Compared with the bare electrode, when the NiFe-PBA with poor conductivity was modified on the electrode surface, the impedance increased and the diameter of the semicircle became larger. When SnS2 with good conductivity was further modified on the electrode surface, the impedance decreased and the diameter of the semicircle became smaller. When the electrode surface was further modified with SiO2-COOH@MIPs, the impedance increased due to the poor conductivity of SiO2-COOH@MIPs. As the adsorption process went on, the imprinted cavity on the electrode surface was occupied, which hindered the transfer of electrons and further increased the impedance. The results of the EIS analysis were similar to those of the CV and DPV analyses, which further proved that the sensor preparation process was successful.

Optimization of Conditions
Here, we define the ΔID value (ΔID = ΔIprobe + ΔIsubstrate) that represents the sum of the current differences between elution and incubation. The different assembly steps of the sensor were further characterized using EIS, and the results are shown in Figure 2C. Compared with the bare electrode, when the NiFe-PBA with poor conductivity was modified on the electrode surface, the impedance increased and the diameter of the semicircle became larger. When SnS 2 with good conductivity was further modified on the electrode surface, the impedance decreased and the diameter of the semicircle became smaller. When the electrode surface was further modified with SiO 2 -COOH@MIPs, the impedance increased due to the poor conductivity of SiO 2 -COOH@MIPs. As the adsorption process went on, the imprinted cavity on the electrode surface was occupied, which hindered the transfer of electrons and further increased the impedance. The results of the EIS analysis were similar to those of the CV and DPV analyses, which further proved that the sensor preparation process was successful.

Optimization of Conditions
Here, we define the ∆I D value (∆I D = ∆I probe + ∆I substrate ) that represents the sum of the current differences between elution and incubation.

Molar Ratio of CAP to MAA
The ability of MIPs to re-adsorb CAP varied with the molar ratio of CAP to MAA ( Figure 3A). When the M CAP /M MAA ratio increased from 2:1 to 1:4, the ∆I D value steadily increased. However, when the molar ratio of CAP to MAA increased from 1:4 to 1:6, the ∆I D value gradually decreased. This phenomenon could be explained by the copolymerization of the crosslinker and functional monomer during the actual reaction. The presence of excessive functional monomers could lead to polymerization between the functional monomers and reduce the number of imprinted cavities with specific recognition ability, thereby reducing the adsorption capacity of the MIP. We could conclude that the optimal M CAP .M MAA ratio was 1:4. The ability of MIPs to re-adsorb CAP varied with the molar ratio of CAP to MAA ( Figure 3A). When the MCAP/MMAA ratio increased from 2:1 to 1:4, the ΔID value steadily increased. However, when the molar ratio of CAP to MAA increased from 1:4 to 1:6, the ΔID value gradually decreased. This phenomenon could be explained by the copolymerization of the crosslinker and functional monomer during the actual reaction. The presence of excessive functional monomers could lead to polymerization between the functional monomers and reduce the number of imprinted cavities with specific recognition ability, thereby reducing the adsorption capacity of the MIP. We could conclude that the optimal MCAP.MMAA ratio was 1:4.  Figure 3B shows the effect of the molar ratio of the template molecule CAP to the crosslinker EGDMA on the ability of MIPs to re-adsorb CAP. The ΔID value also changed significantly with the change in the amount of crosslinking agent added. The ΔID value   Figure 3B shows the effect of the molar ratio of the template molecule CAP to the crosslinker EGDMA on the ability of MIPs to re-adsorb CAP. The ∆I D value also changed significantly with the change in the amount of crosslinking agent added. The ∆I D value continually increased when the M CAP :M EGDMA ratio rose from 1:5 to 1:20. This was because the more EGDMA we added, the more CAP was bound when forming the MIPs.
It was common sense that the more EGDMA we added, the more the CAP formed MIPs and the more specific recognition cavities formed after elution. However, the ∆I D values continued to decrease when the M CAP :M EGDMA ratio rose from 1:20 to 1:30 because the excess crosslinker interfered with the ability of the MIPs to specifically recognize CAP. By contrasting the ∆I D value, the optimal molar ratio of CAP, MAA, and EGDMA was determined to be 1:4:20.

Adsorption Time
The performance of the sensor was also affected by the adsorption time. As the adsorption time steadily increased, the ∆I D value also increased gradually, which could be attributed to the fact that a large number of target molecules were specifically adsorbed into the cavity of the MIPs with the increase in time, hindering the transfer of electrons. The maximum value of ∆I D was reached at 20 min ( Figure 3C). When the adsorption time continued to extend, we found that the ∆I D value decreased slightly. This phenomenon could be attributed to the expansion and deformation of MIPs caused by prolonged immersion, resulting in the detachment of some target molecules from the MIPs.

Elution Times
The elution times were also one of parameters that needed to be optimized for the sensor because the number of imprinted cavities with specific recognition target molecules was directly affected by the elution times. According to the UV absorption spectrum of the eluate in Figure 3D (inset), the UV absorption peak (278 nm) of CAP in the fifth eluate was close to zero, indicating that the template molecule CAP could be completely removed after five elution phases. However, combined with the statistical graph in Figure 3D, the ∆I D reached the maximum value in the third elution phase. When the elution continued, the ∆I D showed a significant downward trend. This might have been because the excessive number of elution phases reduced the imprinted cavity, which could specifically recognize the target molecule, by disrupting the 3D imprinted structure of the MIPs. Therefore, three elution phases was selected as the optimal number.

Adsorption Studies
The absorbance values of different concentrations (0.01 mg/mL-0.1 mg/mL) of CAP standard solutions were measured using a UV spectrophotometer, and the results are shown in Figure 4A. Then, taking the concentration of the CAP solution as the abscissa and the absorbance value as the ordinate, the standard curve was drawn (R 2 = 0.999), as shown in the illustration in Figure 4A.

Adsorption Studies
The absorbance values of different concentrations (0.01 mg/mL-0.1 mg/mL) of CAP standard solutions were measured using a UV spectrophotometer, and the results are shown in Figure 4A. Then, taking the concentration of the CAP solution as the abscissa and the absorbance value as the ordinate, the standard curve was drawn (R 2 = 0.999), as shown in the illustration in Figure 4A.

Adsorption Isotherm Analysis
The adsorption isotherm curve was one of the important factors for studying the adsorption mechanism, which could reflect the interactions between the template molecules and SiO 2 -COOH@MIPs during the adsorption process [30]. The equilibrium adsorption capacity (Q e ) of SiO 2 -COOH@MIPs was 44.3 mg/g, which was 3.6 times higher than that of SiO 2 -COOH@NIPs (12.1 mg/g) ( Figure 4B). Compared with SiO 2 -COOH@NIPs, the high specific binding ability of SiO 2 -COOH@MIPs to CAP could be attributed to the large number of imprinted cavities in the SiO 2 -COOH@MIPs, which could specifically recognize CAP.
The Scatchard equation is a commonly used analytical method to analyze the binding properties of MIPs. The Scatchard equation formula is as follows [31]: where Q m (mg/g) is the maximum adsorption capacity of the binding site and K d (mg/L) is the dissociation constant of the binding site.
By performing a Scatchard model analysis on the adsorption isotherm, the fitting curve as shown in Figure 4C,D could be obtained, and the data obtained via fitting and calculation are recorded in Table S1. From Figure 4C, we found that there were two fitting curves with different slopes in the Scatchard model analysis of the CAP adsorption by SiO 2 -COOH@MIPs, indicating that there might be two different binding sites for the adsorption of CAP by SiO 2 -COOH@MIPs. A further analysis showed that the K d values of the two fitted curves were 1.286 mg/L and 3.595 mg/L, respectively. It could be concluded that the two binding sites exhibited high affinity and low affinity, respectively. Comparing the K d value of the SiO 2 -COOH@NIPs with those of the SiO 2 -COOH@MIPs, we were surprised to find that the K d value of the SiO 2 -COOH@NIPs was much larger than the two K d values of the SiO 2 -COOH@MIPs. This phenomenon occurred because the SiO 2 -COOH@NIPs did not specifically recognize and bind to the imprinted cavity of CAP.
The adsorption behavior of the SiO 2 -COOH@MIPs was further analyzed using Langmuir and Freundlich isothermal models ( Figure 4E,F), where the Langmuir model was suitable for monolayer adsorption and the Freundlich model was suitable for multilayer adsorption. The data obtained from the fitting calculation are recorded in Table S2. The Langmuir formula and the Freundlich formula are as follows [32]: ln Q e = ln K F + ln C e n where K L (L/mg) is the adsorption constant of the Langmuir formula, K F (L/mg) is the adsorption constant of the Freundlich formula, and n is the constant of the adsorption strength of the polymer. According to Table S2, the value of 1/n was 0.366 (0.1 < 1/n < 1) [33]. This indicated that the SiO 2 -COOH@MIPs had good adsorption capacity for CAP under the optimal experimental conditions. At the same time, the table shows that the correlation value of the Langmuir model (R 2 = 0.988) was smaller than that of the Freundlich model (R 2 = 0.994), which indicated that the process of CAP adsorption was more suitably described by the Freundlich model, while the adsorption process was dominated by multi-molecular layer adsorption.

Adsorption Kinetics Analysis and Selective Adsorption Analysis
In this study, the adsorption kinetics and selective adsorption of SiO 2 -COOH@MIPs and SiO 2 -COOH@NIPs were investigated. The pseudo-first-order kinetics equation and pseudo-second-order kinetics equation were used to fit the adsorption data. The specific data are recorded in Supplementary Material 3.4.2.

Capability Assessment of the SiO 2 -COOH@MIPs/SnS 2 /NiFe-PBA/GCE Sensor
Under the optimum tentative conditions, the sensor was used to detect CAP solutions with different concentrations (10 −8 g/L-10 −2 g/L) to test the capacity of the sensor to detect and analyze CAP. When the CAP concentration gradually increased, the oxidization peak-to-peak current of the [Fe(CN) 6 ] 3−/4− and NiFe-PBA gradually decreased; that is, the I D value gradually decreased ( Figure 5A). There was a good linear relationship between the obtained I D value and the logarithm of the CAP concentration. The linear equation was I D = 174.48 − 14.02 Lgc (R 2 = 0.991), and the LOD was 3.3 × 10 −9 g/L (S/N = 3).
To demonstrate the merits of dual-signal MIPs sensors, we also prepared single-signal SiO 2 -COOH@MIPs/SnS 2 /GCE sensors ( Figure 5B) as well as SiO 2 -COOH@NIPs/SnS 2 / NiFe-PBA/GCE sensors ( Figure 5C) as controls. It can be seen from Figure 5B that the detection range of the single-signal sensor was 10 −8 g/L-10 −2 g/L, and the linear equation was I probe = 132.01 − 9.71 Lgc (R 2 = 0.982). By comparing the sensitivity of the single-signal sensor (9.71 µA/(g/L)) and the dual-signal sensor (14.02 µA/(g/L)), the sensitivity of the dual-signal sensor was better, and higher sensitivity detection could be achieved. It can be seen from Figure 5C that after the sensor modified by SiO 2 -COOH@NIPs was adsorbed in distinct concentrations of CAP solution, no obvious linear relationship between the I D value and CAP concentration was found. Table 1 summarizes some of the published methods used for CAP detection. From the comparison data, it was obvious that the sensor designed in this paper had better detection ability.    To explore the specificity of the SiO 2 -COOH@MIPs/SnS 2 /NiFe-PBA/GCE sensor and the SiO 2 -COOH@NIPs/SnS 2 /NiFe-PBA/GCE sensor, this study measured the ∆I D value in a mixed solution containing 10 −4 g/L CAP and 10 −3 g/L of interfering substances. As shown in Figure 5D, the ability of the SiO 2 -COOH@MIPs to specifically recognize the CAP was not affected by the interfering substances, and the ability of the SiO 2 -COOH@MIPs to recognize the CAP was much higher than the SiO 2 -COOH@NIPs.
For the purpose of testing the stability of the sensor ( Figure 5E), we prepared 24 SiO 2 -COOH@MIPs/SnS 2 /NiFe-PBA/GCE electrodes under optimal and identical conditions, divided them into 6 groups, and stored them at 4 • C in a refrigerator. We removed a group of electrodes every three days, put them into the CAP solution at a concentration of 1.0 × 10 −3 g/L for adsorption, and calculated the ∆I D value. After 15 days, the ∆I D value of the last set of electrodes remained at 93.91% of the initial value. The above series of experiments fully demonstrated the excellent stability of the sensor. The reproducibility test was carried out using 6 electrodes ( Figure 5F). The sensor prepared under optimal conditions was immediately placed in a CAP solution at a concentration of 10 −3 g/L for adsorption and the ∆I D value was calculated. The relative standard deviation (RSD) of the ∆I D value of the 6 GCEs was merely 1.85%, indicating that the sensor had excellent reproducibility.

Real Sample Analysis
We tested the performance of the constructed sensor in the detection of real samples using the standard additive method. Table 2 shows the detection results of the sensor in the actual sample. The recovery range of the standard addition was 97.7~104.1%, and the RSD was less than 4.68%. The detection results showed that the SiO 2 -COOH@MIPs/SnS 2 /NiFe-PBA/GCE sensor had gratifying detection accuracy and convincing reliability in real sample detection, and was a promising CAP detection tool.

Conclusions
In this paper, a molecularly imprinted electrochemical sensor based on a dual-signal analysis was constructed for the detection of chloramphenicol in milk. The prepared SiO 2 -COOH@MIPs/SnS 2 /NiFe-PBA/GCE using the combination of molecular imprint-ing technology and electrochemical sensing technology had a wider detection range (10 −8 g/L-10 −2 g/L), a lower LOD (3.3 × 10 −9 g/L), and higher sensitivity (14.02 µA/(g/L)), with good specificity and reproducibility, satisfactory stability, and excellent recovery (97.7~104.1%) in actual sample detection at the same time. This paper was based on the interactions of multiple nanomaterials to achieve dual-signal high-sensitivity detection, which had certain innovative significance. At the same time, the detection method proposed in this paper also had certain reference significance for the detection of target substances in other fields. However, molecularly imprinted electrochemical sensors also face some challenges, such as the incomplete elution of template molecules and difficulties with single-synthesis methods. Therefore, exploring new elution methods and synthesis methods in future work is still an important research direction.  Table S1: Scatchard parameters of SiO 2 -COOH@MIPs and SiO 2 -COOH@NIPs for the adsorption of CAP. Table S2: Langmuir and Freundlich parameters of SiO 2 -COOH@MIPs and SiO 2 -COOH@NIPs for the adsorption of CAP. Table S3: Pseudo-first-order and Pseudo-second-order parameters of SiO 2 -COOH @MIPs for the adsorption of CAP [27,28,30,[41][42][43][44][45][46].