Electrochemical sensor based on a poly(para-aminobenzoic acid) film modified glassy carbon electrode for the determination of melamine in milk
Ya Ting Liu,Jian Deng,Xi Lin Xiao,Li Ding,Ya Li Yuan,Hui Li,Xiu Ting Li,Xiao Na Yan,Li Li Wang
DOI: https://doi.org/10.1016/j.electacta.2011.02.088
IF: 6.6
2011-01-01
Electrochimica Acta
Abstract:A novel and simple sensor is developed in this paper for melamine detection, which is based on an electropolymerized molecularly imprinted polymer (MIP) of para -aminobenzoic acid ( p ABA). The poly( para -aminobenzoic acid) (P- p ABA) film was deposited in a p ABA solution by potentiodynamic cycling of potential with and without the template (melamine) on a glassy carbon electrode. The surface feature of the modified electrode was characterized by electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV). The molecular imprinted sensor was tested by differential pulse voltammetry (DPV) to verify the changes in redox peak currents of hexacyanoferrate. Several important parameters controlling the performance of the P- p ABA were investigated and optimized. In the optimal conditions, the relative redox peak currents of hexacyanoferrate were linear. The concentration of melamine ranged from 4.0 μM to 0.45 mM, with a linear correlation coefficient of 0.9992. The detection limit was 0.36 μM (S/N = 3). The MIP sensor was successfully applied to the determination of melamine in milk products and showed high selectivity, sensitivity, and reproducibility. The results of this research demonstrate that it is feasible to use the molecular imprinting methodology when preparing sensing devices for analytes that are electrochemically inactive. Keywords Molecularly imprinted polymer Poly( para -aminobenzoic acid) Electropolymerization Melamine Milk 1 Introduction Melamine (1,3,5-triazine-2,4,6-triamine), a kind of triazine analogue with three amino groups, is an industrial chemical used in the production of melamine-formaldehyde resins [1–4] . Because it contains a substantial amount of nitrogen (66% by mass), it has been used as a filler for protein-rich diets by unethical manufacturers. This was mirrored in the pet food incident in early 2007 and the milk products scandal recently when melamine was added to raw materials to obtain high protein contents [5] . Melamine has a low oral acute toxicity, but the chronic administration of high concentrations can induce renal pathology and even death, especially in babies and children [6] . Melamine in combination with cyanuric acid is able to form insoluble melamine cyanurate crystals in the kidney, which causes renal failure. This situation prompted the US Food and Drug Administration (FDA), the European community, and other countries and regions to establish the criteria of maximum residue limits for melamine in various food products. Standard limits of 1 ppm for melamine in infant formula and 2.5 ppm in other milk products have been introduced by many countries [7] . Therefore, a sensitive and reliable method is urgently needed for the determination of melamine in food and, particularly, in dairy products for children. Conventionally, the estimation of melamine in feed and food has been performed using chromatography [8–13] . Several other analytical techniques have been applied to detect melamine in milk products, including enzyme immunoassays with high performance liquid chromatography/mass spectroscopy (HPLC/MS) [14] , solid phase extraction [3] , ultraviolet spectroscopy (UV) [15,16] , Raman spectroscopy [17,18] , electrophoresis [19–21] , mass spectrometry coupled with other techniques [22,23] , and chemiluminescence (CL) [2] . However, most of these techniques are not widely available in ordinary analytical laboratories. Compared with the methods mentioned above, chemosensors have the characteristics of simplicity, low cost, accuracy, sensitivity and high stability. However, so far only a few papers [1,4,24–26] have reported using chemosensors to determine melamine at solid electrodes; among these chemosensors, molecularly imprinted polymer (MIP)-based sensors [24–26] have received considerable attention in the fields of electroanalytical chemistry because of their predetermined selectivity for target molecules, robustness of recognition, high stability, low cost, and easy preparation. Liang et al. [4] developed a potentiometric sensor based on an MIP for the determination of melamine in milk. They used methacrylic acid (MAA) as a functional monomer and ethylene glycol dimethacrylate (EGDMA) as a cross-linking agent to fabricate a polymeric membrane-based ion-selective electrode for the detection of melamine in milk, with a limit of detection (LOD) of 1.6 μM. A melamine piezomicrogravimetric chemosensor using a MIP film was constructed by Pietrzyk et al. [24] . The film was prepared by the electropolymerization of melamine complexed by the functional monomer of a bis(bithiophene) derivative bearing an 18-crown-6 substituent and cross-linking monomer of bithianaphthene, with a LOD of approximately 5 nM. Electrosynthesis of MIP-based sensors allows the generation of a rigid, uniform and compact molecularly imprinted layer with good adherence to a transducer surface of any shape and size. Moreover, the thickness and the density of the polymer layer can also be well controlled with the electropolymerization conditions (e.g., applied potential, the number of CV cycles). This feature gives the possibility of creating a direct communication between the coating and the surface of the transducer in a very simple way for the development of chemical sensors [27–29] . Many MIP-based sensors have been developed by electropolymerization employing either a single monomer, such as phenol [30] , aminophenol [31–33] , mercaptobenzimidazole [34,35] , pyrrole [27,36–40] , metal porphyrin [28,41] , o -phenylenediamine [42–44] , or multiple monomers to acquire a copolymer [29,45–47] . In this paper, the assembly of an electrochemical sensor, which is based on the electropolymerization of para -aminobenzoic acid ( p ABA) using molecular imprinting technology, is described for the detection of melamine for the first time. p -ABA was chosen as the monomer for electropolymerization because it has been shown that it can be easily electropolymerized on various substrate materials and form films with good chemical and mechanical stability [48–54] . Meanwhile, p ABA can establish hydrogen bonds with melamine, which facilitates the recognition and selectivity for the analyte. The morphologic feature of the imprinting and non-imprinting poly( para -aminobenzoic acid) (P- p ABA) membrane-modified glassy carbon electrode (GCE) was observed by electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV). Considering that melamine is electro-inactive [1] over the studied potential range, an electroactive substance, hexacyanoferrate, was chosen as the redox probe of the P-pABA film coated electrodes in solutions containing the analyte. In this study, differential pulse voltammetry (DPV) was selected as the transduction method due to its higher sensitivity compared with CV. 2 Experimental 2.1 Reagents and materials All chemicals were of analytical grade and used as received without further purification. Melamine, p -aminobenzoic acid, sodium sulfate, potassium ferricyanide, acetonitrile, and other reagents were obtained from Tianjin Kermel Chemical Reagent Co., Ltd. The milk analyzed in this work was obtained from a local supermarket. Twice-distilled water (1.0 μs/cm) was used throughout all the experiments. Stock solutions of 0.01 M melamine and 0.1 M p ABA were made by dissolving 0.1261 g (1 mmol) melamine and 1.3714 g (10 mmol) p ABA in 100 ml of a 1:1 (v/v) acetonitrile–water solution, respectively. The background solution used in the measurement experiments was a 1:1 (v/v) acetonitrile–water solution with 0.2 M Na 2 SO 4 and 2 mM hexacyanoferrate. The extraction solvent was prepared by mixing acetonitrile and water in a volume ratio of 1:1. 2.2 Apparatus A CHI 660C Electrochemical Workstation (CHI Instrument Company, Shanghai, China) was used for electrochemical measurements. A conventional three-electrode system was employed with a bare or P- p ABA-modified glassy carbon electrode (3.0 mm in diameter) as the working electrode, a platinum electrode as the counter electrode, and a saturated calomel electrode (SCE, 0.2415 V vs. SHE) as the reference electrode. All potentials reported in this article were referenced to the SCE. All measurements were carried out at room temperature. The electrolyte solutions were thoroughly deaerated by purging with high-purity nitrogen and kept under a nitrogen atmosphere. 2.3 Sample pretreatment The liquid milk was pretreated according to the following procedure. Five grams of accurately weighed (precision 0.1 mg) milk was placed into a 50 ml polypropylene centrifuge tube, to which 40 ml of extraction solvent was added. After 10 min of shaking, the mixture was centrifuged at 4000 rpm for 10 min, and the supernatant was filtered through a 0.22 μm filter membrane. The filtrate was diluted with extraction solvent to 50 ml to obtain the samples for detection. Because melamine-contaminated milk cannot be purchased from the market anymore, the milk was spiked with appropriate amounts of melamine standard solution directly. 2.4 Preparation of P- p ABA modified electrodes Prior to modification, the GCE was polished with emery paper and chamois leather containing 0.3 and 0.05 μm Al 2 O 3 slurry, respectively, and then thoroughly rinsed ultrasonically with HNO 3 , ethanol, and doubly distilled water for 5 min in turn. Then the electrode was cycled between −0.5 and 2.0 V in 0.5 M H 2 SO 4 at 100 mV/s until a reproducible voltammogram was obtained. P- p ABA with imprinting of melamine-modified GCE (MIP/GCE) was prepared by potentiodynamic cycling of potential for 15 scans in the range from −0.8 to 1.0 V at a scan rate of 50 mV/s from a 1:1 acetonitrile–water solution, which contained 5 mM p ABA and 1 mM melamine and was purged with nitrogen for 10 min before using. For comparison, a non-imprinted polymer/GCE (NIP/GCE) was fabricated following the same procedure, but in the absence of the template. The MIP and NIP sensors were rinsed three times, each in 5 ml of extraction solvent for 5 min after electropolymerization, to remove the imprinting molecules and adsorbates on the surface of the membrane with gentle agitation provided by a magnetic stirrer. Meanwhile, the sensors were washed three times using the same procedure to remove melamine during the experiment. 2.5 Experimental measurements A standard three-electrode cell connected to the CHI660C was used for electrochemical measurements. After being immersed in the melamine solution for 1 min, the electrode was washed carefully with double-distilled water and dried under a nitrogen stream. The CV experiment was performed in the 10 ml background solution, and the potential range was taken from −0.2 to 0.6 V at 50 mV/s. Current measurements were performed using DPV in a potential range between −0.2 and 0.6 V, a modulation amplitude of 50 mV, a pulse width of 50 ms and a step potential of 5 mV. The sensor response was obtained from the change of oxidation current (Δ i ) of hexacyanoferrate before and after the sensor was immersed in each concentration of the analyte solution. Impedance measurements were carried out at a potential of 0.175 V over the frequency range from 100 mHz to 100 kHz, using an alternating potential of 5 mV. 3 Results and discussion 3.1 Electrosynthesis of molecularly imprinted and non-imprinted poly( p ara-aminobenzoic acid) Fig. 1 shows a typical cyclic voltammogram recorded during the electropolymerization of p ABA on the GCE in the presence of melamine. In the first cycle, when the potential ranged from −0.80 to 1.00 V, an irreversible oxidation peak appeared with an anodic peak potential at 0.94 V, which could be attributed to the formation of an amino cation radical [48–50,52] . This anodic peak current decreased rapidly and disappeared in the potential range of 0.90–1.00 V after the completion of the first several cycles. As can be seen, a pair of redox peaks were observed at 0.38 and 0.31 V, which may be ascribed to the oxidation and reduction of the film, and the current intensity of the redox pair became lower as the number of scans increased until reaching a certain limit, which is in accordance with the results reported by other authors [48,52] and reflects the continuous growth of the low conducting layer around the electrode. These observations indicate that a P- p ABA film was electrodeposited on the surface of the GCE. As a result, an adherent blue polymer film was formed on the GCE surface. No notable differences were observed in the cyclic voltammogram with and without the present of melamine (not shown), which can be explained by the fact that melamine does not have any electroactivity on the GCE in the potential window chosen for the polymerization. During the electrodeposition of the molecularly imprinted polymer, the template melamine molecules are trapped in the polymer matrix because of the molecular imprinting effect. The ring-nitrogen atoms and the amino groups of the melamine molecule can form hydrogen bonds with both the carboxyl group and the amino group of pABA units. Additionally, one can reasonably speculate that the benzene ring of P- p ABA can provide recognition sites through a “π–π stacking” interaction with the aromatic structure of melamine ( Scheme 1 ). As a result, chain branching and cross linking in P- p ABA generate a three-dimensional matrix with niches containing the template melamine, and the imprinting process creates a microenvironment for the recognition of the melamine molecule based on shape selection and complementarity of the functional groups. 3.2 Optimization of the parameters influencing performance of the imprinted sensor The sensor behavior is influenced by the conditions in the electrodeposition and measurement process. The polymerization parameters, such as the concentration of monomer and template, applied potential, electropolymerization cycles, and scan rate, extraction method and exposure time, determine the thickness and physical stability of the polymer film produced on the GCE and are expected to have an effect on the sensor response. All the optimization experiments were performed using DPV and the experimental conditions were the same as those described in the experimental measurements section. The measured current difference, namely, the change of the oxidized peak current (Δ i ) of hexacyanoferrate before and after the sensor was immersed in a 0.1 mM melamine solution, was used as the evaluation indicator. 3.2.1 Effect of the monomer concentration The sensor behavior was found to be dependent on the monomer concentration employed during the electropolymerization. This is because of the ability to bind the template molecules onto the coating film depends on the number of imprints, and in turn the monomer concentration should be proportional to the thickness of the deposit and the amount of imprinted molecule (template) in the polymer matrix [27,40] . To examine the effect of the monomer concentration on the response of both MIP/GCE and NIP/GCE to melamine, the films were grown in solutions of constant melamine concentration and varying p ABA concentrations in the range of 3–10 mM by cycling the potential between −0.8 V and 1.0 V for the same cycles. The results are shown in Fig. 2 A . It was observed that the response of the MIP/GCE to melamine increased as the p ABA concentration increased up to 5 mM, then decreased when the p ABA concentration increased further. Increasing the monomer concentration may cause a rapid polymerization and increase the sensor sensitivity. However, a high concentration of the monomer might lead to the generation of a bulky membrane, which causes some of the recognition sites to be not easily accessible [36,38] and consequently causes a nonselective electrochemical response to the template [29] . To improve the sensitivity of melamine detection, the difference between the responses of the MIP/GCE and the NIP/GCE to melamine should be as large as possible (Δ i MIP/GCE ≫ Δ i NIP/GCE ) [27,40] . Therefore, 5 mM was selected as the optimum monomer concentration in this work. 3.2.2 Effect of the template concentration The template concentration in the polymerization mixture is also a significant factor in the response of the sensor. Fig. 2 B shows the effect of the template concentration on the response of the MIP/GCE during the electropolymerization of p ABA in the range of 0.5–1.5 mM of melamine. The response of the MIP/GCE to melamine increased with the increase of the template concentration between 0.5 and 1.0 mM. There was a small decrease in the response of the MIP/GCE above 1.0 mM. The sensor produced at a lower template concentration was less sensitive; this is probably due to the small number of recognition sites formed in the polymer matrix. The sensor prepared in a higher concentration of template molecule also showed poor sensitivity. This may be because the melamine molecules could have formed a complex with most of the monomer molecules, lowering the likelihood of cross-linking and the consequent entrapment of the template molecule in a tri-dimensional matrix [38] . Based on the results, the optimum template concentration was chosen to be 1.0 mM and the molecular ratio of monomer to template was 5–1. 3.2.3 Effect of the applied potential To avoid overoxidation of the synthesized polymer, different potential intervals were studied for carrying out the electropolymerization. The best results were obtained when potential range intervals from −0.80 to 1.00 V were employed. At the lower positive potential, such as 0.90 V, it seems that the anodic current was not very evident. On the other hand, when positive potential values of 1.10 V or higher were used, the oxidation wave became unstable, probably due to overoxidation of the newly electrodeposited polymer. In fact, the highest response of MIP/GCE to melamine was obtained by applying a potential interval from −0.80 to 1.00 V during the electrosynthesis of P- p ABA. Therefore, the optimum upper applied potential was selected to be 1.00 V. 3.2.4 Effect of the electropolymerization cycles and rate It was found that the number of CV cycles applied to the cell during the electropolymerization affected the sensitivity of the sensor ( Fig. 2 C). There was a considerable decrease in the performance of the MIP/GCE below and above 15 cycles. Many more cycles led to more extensive electropolymerization, resulting in a thicker sensing film with less accessible imprinted sites [27,40] . It was observed that the response of the NIP/GCE to melamine increased with an increasing number of cycles, with less sensitivity than that of the MIP. The highest current difference between the MIP/GCE and NIP/GCE for melamine was obtained by applying 15 cycles in the electropolymerization. Therefore, the optimum number of the polymerization cycles was determined to be 15. It has been reported that the morphology of polymer films depends on the electropolymerization rate [35] . The film formed on the surface of the electrode may be compact and smooth in a slow polymerization process, while becoming rough and porous when its growth rate is increased. However, when the scan rate was too low, the template would be firmly entrapped in the polymer and could hardly be removed. It was observed that the optimal scan rate is 50 mV/s. 3.2.5 Effect of extraction Many methods can be used to remove the template in the imprinting polymer, such as supercritical fluid desorption [55] , microwave assisted extraction [55] , overoxidation of the polymer [40] , redox of the template in the polymer [27,44,47] , and extraction with a solvent [2–4,28,29,31–38,41–45,47] . Extraction with a solvent is the most common method, and a solvent can strongly interact with the polymer, making it swoln for template cleaning. In this study, alcohol, ultra-pure water, methanol–acetic acid (8:2, v/v), acetonitrile–acetic acid (8:2, v/v), and an acetonitrile–water solution (1:1, v/v) were each applied to remove the template. The results show that the 1:1 acetonitrile–water solution can remove the template most quickly and completely because melamine can easily dissolve in the 1:1 acetonitrile–water solution, while the other extraction solvents can only partially remove the template. It was found that stirring the imprinting electrode in 5 ml 1:1 acetonitrile–water solution three times, each for 5 min, could remove the template completely, which was verified electrochemically. So this extraction procedure was used in our work. 3.2.6 Effect of exposure time As the exposure of the template-extracted MIP to the analyte solution involved in the measurements is another critical factor for the performance of imprinted sensor, the effect of the exposure time was investigated. After the template was removed from the MIP film, the sensor was immersed in a 1:1 acetonitrile–water solution containing 0.1 mM melamine for different time periods and then put into the background solution to study the peak current quantitatively. It was observed that the peak current decreased sharply with increasing exposure time due to partial binding sites in MIP films occupied by melamine and that a stable response was obtained after immersion for 1 min, which indicates the rapid recognition ability and high affinity of the imprinted sensor to the target molecule. Therefore, an exposure time of 1 min was selected whenever the measurement was made with the MIP sensor. 3.3 Electrochemical characterization of the imprinted sensor Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were used to characterize the imprinted sensor. Cyclic voltammograms of 2 mM hexacyanoferrate solutions containing 0.2 M Na 2 SO 4 on the bare GCE, NIP/GCE and MIP/GCE are shown in Fig. 3 A . A couple of typical redox peaks of hexacyanoferrate appear on the bare GCE (curve 1). Only a very small background response, however, is observed on the NIP/GCE (curve 2) over the potential range, due to the fact that the film coated on the electrode was compact and less conductive than the GCE and there are almost no channels for the active probe to approach the GCE surface. It is also noted that compared with the bare GCE, an obvious current decrease appeared when using the MIP/GCE after the template removal (curve 3). This may result from the low conductivity of the film coated on GCE, though the cavity caused by the removal of melamine molecule forms some channels for hexacyanoferrate to arrive at the electrode surface. Curve 4 shows the same MIP/GCE after being saturated with 0.1 mM melamine, where only a small peak appears due to partial binding sites in MIP films occupied by melamine. EIS is also an efficient tool for studying the interface properties of surface-modified electrodes. The diameter of the semicircle is usually equal to the charge-transfer resistance ( R ct ), which controls the charge-transfer kinetics of the redox probe at the electrode interface [1,28,34] . Fig. 3 B illustrates the typical results of electrochemical impedance spectra (presented in the form of the Nyquist plot) of the bare GCE, NIP/GCE and MIP/GCE, respectively, with hexacyanoferrate as an electroactive probe. It is obvious that the bare GCE exhibits a straight line, which reveals that there is almost no heterogeneous charge-transfer resistance on the bare GCE surface, while the R ct of the NIP/GCE is much larger, indicating that the compact film of low conductivity acts as a definite kinetic barrier for the charge transfer. Similar to the NIP/GCE, the R ct of the MIP/GCE after being immersed in 0.1 mM melamine for 1 min was also higher because imprinted molecular cavities exist in it, and the cavities are occupied by melamine molecules. Therefore, these cavities are unavailable for the hexacyanoferrate redox probe to diffuse through. A significant decrease in R ct can be observed on the MIP/GCE after removing the template because the cavities in the film are not occupied by the template and are, therefore, available for the charge transfer. These results are accordant with CV assays as described in detail above. 3.4 Performance of the imprinted sensor 3.4.1 DPV response and calibration curve In this study, DPV was employed for the quantitative determination of melamine, which is relatively sensitive compared to the conventional CV method. Differential pulse voltammograms in the background solution were performed after the MIP/GCE was immersed in solutions containing melamine of different concentrations for 1 min. When the MIP/GCE was immersed in the solution containing melamine, the cavities in the film were partially occupied by melamine, which led to the decrease of current signal produced by hexacyanoferrate. The higher the concentration of melamine was, the lower the current would be. Fig. 4 A shows the DPV curves of the MIP/GCE for melamine in a concentration range from 4.0 μM to 0.75 mM. As shown in Fig. 4 A, the peak current decreases as the concentration of melamine increases, which suggests that more and more binding sites in the film are occupied by melamine molecules. A dependence of the relative oxidization peak current change of hexacyanoferrate and the melamine concentration is exhibited in Fig. 4 B. Considering the absolute currents may vary with the surface areas of the employed electrodes and the distant between the work electrode and the auxiliary electrode in each experiment, relative current change (Δ i / i 0 ) was therefore used to indicate the current response [44] . Here, Δ i = i 0 − i c , i 0 and i c is the current when the concentration of melamine is 0 and c (with the unit of 0.1 mM), respectively. The relative current change tended to be stable when the concentration of melamine was up to 0.45 mM, which indicates that the imprinting sites are gradually occupied by melamine molecules and have approached the point of “saturation”. The linear region ranged from 4.0 μM to 0.45 mM with LOD of 0.36 μM (S/N = 3). The linear calibration graphs of (Δ i / i 0 ) versus concentration of melamine (c) can be described by the following equation, y = Δ i / i 0 = 0.1344 c (0.1 mM) + 0.0445, and the correlation coefficient is calculated to be 0.9992. 3.4.2 Affinity and selectivity of the imprinted sensor MIPs have selectivity towards the template molecule. The selective recognition is based on the interaction between the template and the imprinting sites. To verify the selectivity of the P- p ABA MIP sensor to melamine determination, some possible ingredients in milk including glycine, tyrosine, phenylalanine, histidine, tryptophan (which may result from the hydrolysis of protein), and ascorbic acid were used to conduct interference experiments on the determination of melamine with an equivalent concentration (0.1 mM). The results are shown in Table 1 . The selectivity coefficient was defined as K = Δ i r /Δ i m , where Δ i r is the corresponding relative current change of 0.1 mM interfering material solution and Δ i m is the relative current change of sample solution containing 0.1 mM melamine. Almost no interference (<2.5%) was observed with tyrosine, phenylalanine, tryptophan, and ascorbic acid, which are bigger than melamine in molecular size and cannot enter into the imprinting sites of the MIP film. Only a slight interference was observed with histidine, which is somewhat similar to melamine in structure but still mismatches the imprinting sites and shows good selectivity. Comparatively, interference produced by glycine was slightly greater than other substances, which may be explained by the fact that glycine is smaller than melamine in molecular size and has a chance of approaching the imprinting sites. Given the above observations, we can conclude that the recognition sites formed in the polymerized film have the capability to distinguish target molecules through their size, shape and functional group distribution. 3.4.3 Reproducibility, repeatability and storage stability of the imprinted sensor To investigate the reproducibility of the proposed technique, three sensors fabricated independently under the same conditions were examined in a 2 mM hexacyanoferrate solution after being immersed in 0.1 mM melamine for 1 min. The response shows a relative standard deviation (RSD) of 3.4%, indicating excellent reproducibility. The repeatability was investigated for 0.1 mM melamine as well, using the same sensor. The calculated RSD was about 4.2% ( n = 5). This good repeatability reveals that binding of melamine to the “cavity” is reversible and the MIPs sensor could be regenerated and used repeatedly. After the electrode was exposed to air for 30 days at room temperature and used at least 60 times, it retained 90.45% of its original response, suggesting acceptable storage stability. 3.4.4 Determination of melamine in milk samples To investigate the potential practical application of the MIP sensor, it was applied to determine melamine in liquid milk bought from supermarkets. Because the existing milk in the market is free of melamine, the samples were spiked with certain amounts of melamine standard solution directly. The RSD of the results acquired by the proposed sensor ranged from 2.42% to 4.38% and the recovery ranged from 95.6% to 105.2% ( Table 2 ), suggesting an acceptable detection result. Therefore, the developed imprinted sensor could be reliable and effective for the determination of melamine in real samples. Table 3 compares the results of melamine determination obtained by several methods. Compared with several current techniques for the assay of melamine, the MIP/GCE sensor has some advantages. The LOD of this method is lower than that using the MIP/potentiometric sensor [4] , liquid chromatography (LC) [8] , UV [15] , and capillary zone electrophoresis (CZE) [19] . It has the same selectivity as the oligonucleotides/Au sensor [1] and the MIP/CL method [2] , and the LOD is not as low as for the oligonucleotides/Au sensor [1] and the acoustic sensor [24] . However, the method has the advantage of cheap reagents and being easy to fabricate. The instrumentation required by the MIP/GCE sensor is simpler and cheaper than that required for LC [8] , HPLC–MS/MS [10] , and gas chromatography–mass spectroscopy/mass spectroscopy (GC–MS/MS) [12] . It should be noted that the recovery obtained in this method meets the requirements for determining melamine in real samples. 4 Conclusions A new approach for the fabrication of a MIP/GCE was presented to determine melamine in milk using a differential pulse voltammetry method. The modified electrode exhibited a low detection limit, wide linear concentration range, good repeatability, and fast response for the detection of melamine. Excellent selectivity, good reproducibility, and high stability were achieved by combining a molecular imprinting technique and the electrodeposition of a poly( p ara-aminobenzoic acid) film on the GCE. In addition, the fabrication procedure is very simple, rapid, and inexpensive. The proposed method was applied successfully to determine melamine in milk products with good precision and accuracy, and it can be used as a routine method for the analysis of melamine in food samples. The strategy proposed in this work is potentially applicable to the electrochemical determination of other important molecules without electrochemical activities. Acknowledgements The authors acknowledge the financial support of Hunan Technology Department , China (Project No. 2010FJ3020 ), and Hunan Education Department , China (Project No. CX2010B378 ). The study was also supported by the Natural Science Foundation of Hunan Province (Project No 10JJ3024). References [1] Q. Cao H. Zhao L.X. Zeng J. Wang R. Wang X.H. Qiu Y.J. He Talanta 80 2009 484 [2] J.H. Yu C.C. Zhang P. Dai S.G. Ge Anal. Chim. Acta 651 2009 209 [3] H.H. Yang W.H. Zhou X.C. Guo F.R. Chen H.Q. Zhao L.M. Lin X.R. Wang Talanta 80 2009 821 [4] R.N. Liang R.M. Zhang W. Qin Sens. Actuators B: Chem. 141 2009 544 [5] J.S. Chen Chin. Med. J. 122 3 2009 243 [6] N. Guan Q.F. Fan J. Ding Y.M. Zhao J.Q. Lu Y. Ai G.B. Xu S.N. Zhu C. Yao L.N. Jiang J. Miao H. Zhang D. Zhao X.Y. Liu Y.N. Yao Engl. J. Med. 360 11 2009 1067 [7] F.X. Sun W. Ma L.G. Xu Y.Y. Zhu L.Q. Liu C.F. Peng L.B. Wang H. Kuang C.L. Xu Trends Anal. Chem. 29 11 2010 1239 [8] R. Muñiz-Valencia S.G. Ceballos-Magana D. Rosales-Martinez R. Gonzalo-Lumbreras A. Santos-Montes A. Cubedo-Fernandez-Trapiella R.C. Izquierdo-Hornillos Anal. Bioanal. Chem. 392 2008 523 [9] H.W. Sun L.X. Wang L.F. Ai S.X. Liang H. Wu Food Control 21 2010 686 [10] Y.T. Wu C.M. Huang C.C. Lin W.A. Ho L.C. Lin T.F. Chiu D.C. Tarng C.H. Lin T.H. Tsai J. Chromatogr. A 1216 2009 7595 [11] X.M. Xu Y.P. Ren Y. Zhu Z.X. Cai J.L. Han B.F. Huang Y. Zhu Anal. Chim. Acta 650 2009 39 [12] H. Miao S. Fan Y.N. Wu L. Zhang P.P. Zhou J.G. Li H.J. Chen Y.F. Zhao Biomed. Environ. Sci. 22 2009 87 [13] GB/T 22388-2008, Determination of melamine in raw milk and dairy products, October 7, 2008. [14] B. Kim L.B. Perkins R.J. Bushway S. Nesbit T.T. Fan R. Sheridan V. Greene J. AOAC Int. 91 2 2008 408 [15] J. Rima M. Abourida T. Xu I.K. Cho S. Kyriacos J. Food Compos. Anal. 22 2009 689 [16] Y.T. Liu J. Deng L. An J. Liang F. Chen H. Wang Food Chem. 126 2 2011 745 [17] Y. Cheng Y.Y. Dong J.H. Wu X.R. Yang H. Bai H.Y. Zheng D.M. Ren Y.D. Zou M. Li J. Food Compos. Anal. 23 2010 199 [18] M. Lin L. He J. Awika L. Yang D.R. Ledoux H. Li A. Mustapha J. Food Sci. 73 8 2008 129 [19] C.W. Klampfl L. Andersen M. Haunschmidt M. Himmelsbach W. Buchberger Electrophoresis 30 2009 1 [20] N. Yan L. Zhou Z.F. Zhu X.G. Chen J. Agric. Food Chem. 57 3 2009 807 [21] C. Zhai W. Qiang J. Sheng J.P. Lei H.X. Ju J. Chromatogr. A 1217 2010 785 [22] S.P. Yang J.H. Ding J. Zheng B. Hu J.Q. Li H.W. Chen Z.Q. Zhou X.L. Qiao Anal. Chem. 81 2009 2426 [23] L. Zhu G. Gamez H.W. Chen K. Chingin R. Zenobi Chem. Commun. 2009 559 [24] A. Pietrzyk W. Kutner R. Chitta M.E. Zandler F. D'Souza F. Sannicolò P.R. Mussini Anal. Chem. 81 2009 10061 [25] M.C. Blanco-López M.J. Lobo-Castañón A.J. Miranda-Ordieres P. Tu- ñón-Blanco Trends Anal. Chem. 23 1 2004 36 [26] A. Ramanavičius A. Ramanavičienė A. Malinauskas Electrochim. Acta 51 2006 6025 [27] L. Özcan Y. Şahin Sens. Actuators B: Chem. 127 2007 362 [28] A. Gómez-Caballero A. Ugarte A. Sánchez-Ortega N. Unceta M. Aranzazu Goicolea R.J. Barrio J. Electroanal. Chem. 638 2010 246 [29] A. Gómez-Caballero M.A. Goicolea R.J. Barrio Analyst 130 2005 1012 [30] T.L. Panasyuk V.M. Mirsky S.A. Piletsky O.S. Wolfbeis Anal. Chem. 71 1999 4609 [31] J.P. Li J. Zhao X.P. Wei Sens. Actuators B: Chem. 140 2009 663 [32] H.P. Liao Z.H. Zhang H. Li L.H. Nie S.Z. Yao Electrochim. Acta 49 2004 4101 [33] J. Zhang Y.Q. Wang R.H. Lv L. Xu Electrochim. Acta 55 2010 4039 [34] A. Aghaeia M.R.M. Hosseinia M. Najafi Electrochim. Acta 55 2010 1503 [35] J.L. Gong F.C. Gong G.M. Zeng G.L. Shen R.Q. Yu Talanta 61 2003 447 [36] B.S. Ebarvia S. Cabanilla F. Sevilla III Talanta 66 2005 145 [37] H.J. Liang T.R. Ling J.F. Rick T.C. Chou Anal. Chim. Acta 542 2005 83 [38] D.R. Albano F. Sevilla III Sens. Actuators B: Chem. 121 2007 129 [39] V. Syritski J. Reut A. Menaker R.E. Gyurcsányi A. Öpik Electrochim. Acta 53 2008 2729 [40] S.P. Özkorucuklu Y. Şahin G. Alsancak Sensors 8 2008 8463 [41] E. Mazzotta C. Malitesta Sens. Actuators B: Chem. 148 2010 186 [42] L. Feng Y.J. Liu Y.Y. Tan J.M. Hu Biosens. Bioelectron. 19 2004 1513 [43] Z.L. Cheng E.K. Wang X.R. Yang Biosens. Bioelectron. 16 2001 179 [44] Y. Liu Q.J. Song L. Wang Microchem. J. 91 2009 222 [45] A. Gómez-Caballero N. Unceta M.A. Goicolea R.J. Barrio Sens. Actuators B: Chem. 130 2008 713 [46] M.S. Wu T.C. Wen A. Gopalan Mater. Chem. Phys. 74 2002 58 [47] H. Peng C.D. Liang A.H. Zhou Y.Y. Zhang Q.J. Xie S.Z. Yao Anal. Chim. Acta 423 2000 221 [48] F. Xu M.N. Gao G.Y. Shi L. Wang W. Zhang J. Xue L.T. Jin J.Y. Jin Anal. Chim. Acta 439 2001 239 [49] Y.Z. Zhang J. Wang M.L. Xu Colloid. Surf. B 75 2010 179 [50] K.J. Huang C.X. Xu W.Z. Xie W. Wang Colloid. Surf. B 74 2009 167 [51] J.Y. Ye J.Y. Liu Z.Q. Zhang J.M. Hu S.J. Dong Y.H. Shao J. Electroanal. Chem. 508 2001 123 [52] F. Xu M.G. Gao L. Wang G.Y. Shi W. Zhang L.T. Jin J.Y. Jin Talanta 55 2001 329 [53] A. Benyoucef F. Huerta J.L. Vázquez E. Morallon Eur. Polym. J. 41 2005 843 [54] L. Li D.F. Sun Z.P. Wang X.Y. Song S.X. Sun Solid State Sci. 11 2009 1040 [55] A. Ellwanger L. Karlsson P.K. Owens C. Berggren C. Crecenzi K. Ensing S. Bayoudh P. Cormack D. Sherrington B. Sellergren Analyst 126 2001 784