Determination of malachite green in fish based on magnetic molecularly imprinted polymer extraction followed by electrochemiluminescence

Talanta 142 (2015) 228–234

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Determination of malachite green in fish based on magnetic molecularly imprinted polymer extraction followed by electrochemiluminescence Baomei Huang a,b, Xibin Zhou b, Jing Chen b, Guofan Wu b, Xiaoquan Lu b,n a

College of Chemistry & Chemical Engineering, MianYang Normal University, MianYang 621000, China Key Laboratory of Bioelectrochemistry & Environmental Analysis of Gansu Province, College of Chemistry & Chemical Engineering, Northwest Normal University, Lanzhou 730070, China

b

art ic l e i nf o Article history: Received 17 February 2015 Received in revised form 15 April 2015 Accepted 17 April 2015 Available online 27 April 2015 Keywords: Magnetic molecularly imprinted polymers Malachite green Electrochemiluminescence Quenching Fish

a b s t r a c t A novel procedure for selective extraction of malachite green (MG) from fish samples was set up by using magnetic molecularly imprinted polymers (MMIP) as the solid phase extraction material followed by electrochemiluminescence (ECL) determination. MMIP was prepared by using Fe3O4 magnetite as magnetic component, MG as template molecule, methacrylic acid (MAA) as functional monomer and ethylene glycol dimethacrylate (EGDMA) as crosslinking agent. MMIP was characterized by SEM, TEM, FT-IR, VSM and XRD. Leucomalachite green (LMG) was oxidized in situ to MG by 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ). And then MMIP was successfully used to selectively enrich MG from fish samples. Adsorbed MG was desorbed and determined by ECL. Under the optimal conditions, calibration curve was good linear in the range of 0.29–290 μg/kg and the limit of detection (LOD) was 7.3 ng/kg (S/N¼ 3). The recoveries of MMIP extraction were 77.1–101.2%. In addition, MMIP could be regenerated. To the best of our knowledge, MMIP coupling with ECL quenching of Ru(bpy)32 þ /TPA for the determination of MG has not yet been developed. & 2015 Elsevier B.V. All rights reserved.

1. Introduction Malachite green (MG) is a cationic triphenylmethane dye, which has been used by the aquaculture industry to combat ectoparasites and control fungus on fish eggs, fingerlings and adult fish from 1933 [1]. At present, the use of MG in aquaculture has aroused serious concerns because it has been reported that MG may cause human carcinogenesis and mutagenesis [2,3]. Leucomalachite green (LMG, Fig. 1) can be formed by the metabolic reduction of MG and it is considered even more hazardous than MG because their degradation paths are similar but the half-life of LMG is much longer than that of MG [4,5]. However, due to their low cost and high efficacy, they are still used in some parts of the world. According to the European Commission (2004), the analytical methods developed should meet a minimum required performance limit of 2 ppb (μg/kg) for the sum of MG and LMG in fish [6]. To date, many methods have been proposed for this purpose, such as liquid chromatography [2,7–13], capillary electrophoresis [14], spectrophotometric method [15,16], n

Corresponding author. Tel.: þ 86 931 7971276; fax: þ 86 931 7971323. E-mail address: [email protected] (X. Lu).

http://dx.doi.org/10.1016/j.talanta.2015.04.053 0039-9140/& 2015 Elsevier B.V. All rights reserved.

etc. However, some of these methods have the shortcomings of being time-consuming, expensive, complex and less sensitivity and narrow linear response range. Therefore, it is highly desirable to develop a sensitive, low-cost and highly effective method for the determination of MG in fish. Electrochemiluminescence (ECL) [17,18] has been attracting considerable attention for the past several decades due to its inherent high sensitivity, simplified setup and so on. Ru(bpy)32 þ /TPA is the most widely used system in the research of ECL owing to its strong luminescence, wide linear range, etc. Our previous study has found that MG has high quenching efficiency over Ru(bpy)32 þ /TPA [19], which is promising to provide a highly sensitive method for the detection of MG. However, the application of this method is limited by its poor selectivity. Due to the highly complex matrices and low levels of MG in fish, a selective and sensitive sample preparation method is required for the isolation or enrichment of the analyte. This can be done by solidphase extraction (SPE) with molecularly imprinted polymer (MIP). MIP is artificial polymer with a predetermined selectivity for target molecule which is formed in the presence of target molecules and removed by proper solvents [20,21]. Owing to the chemical,

B. Huang et al. / Talanta 142 (2015) 228–234

Fig. 1. Chemical structures of MG (A) and LMG (B).

mechanical, thermal stability and high selectivity for the template molecules, MIP as selective solid-phase adsorbent for MG extraction has been used in recent years [7–9,22–24]. In 2007, Yan et al. [24] first reported the preparation of MG–MIP for separating MG from seafood, water and other matrices. Later, some researchers [7–9,22,25] proposed MG–MIP solid phase extraction combining with HPLC for the detection of MG from fish samples. Guo et al. [26] also proposed a successful MG–MIP combining with luminol–ECL for the determination of MG and LMG in fish. But all the proposed MIPs had been used in cartridge mode, which often resulted in a tedious column packing procedure, high backpressure, a large amount of organic solvents and time-consuming. Recently, magnetic adsorbent based on Fe3O4 as the core has attracted intensive interest in sample treatment because it can be easily separated by external magnet to avoid complicated process of filling column and pretreatment in traditional SPE. Furthermore, MIP is coated on Fe3O4, so the resulting polymer's recognition sites are situated at the surface of the magnetic material and it is easy to prepare and chemically stable. Coupling magnetic material with MIP (MMIP) is possible to combine the advantages of them such as high selectivity, specificity and sensitivity to the target molecules. To the best of our knowledge, this is the first paper combining MMIP with ECL for the determination of trace MG. The aim of the present study is to propose a novel, selective and effective method for the detection of trace level of MG in complicated matrix based on ECL quenching of Ru(bpy)32 þ /TPA and high selectivity of MMIP. A core–shell MMIP was synthesized using Fe3O4@SiO2 as magnetic supporter and MG as template. The adsorption capacity, recognition site and selectivity of MMIP were characterized in detail. The results showed that MMIP displayed specific recognition towards MG and high magnetism which could be collected and separated fast by external magnetic field. The MMIP was used to separate and enrich MG from fish samples and followed by ECL monitoring.

2. Experimental 2.1. Chemicals and apparatus Tris(2,2-bipyridyl) dichlororuthenium(II) hexahydrate (Ru(bpy)3 Cl2  6H2O) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Malachite green (MG), aluminum oxide (Al2O3), crystal violet (CV), leucomalachite green (LMG), 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) and tripropylamine (TPA) were purchased from Aladdin (Shanghai, China). FeCl2  4H2O, FeCl3  6H2O and 3-methacryloxypropyltrimethoxy-silane (MPS) were purchased from Shuangshuang chemical company limited (Yantai, China). Ammonia (NH3  H2O, 25%, w/w) and acetic acid were purchased from Guangfu (Tianjin, China). Tetraethyl orthosilicate (TEOS) was obtained from Sinopharm (Shanghai, China). Ethanol, methanol, chloroform,

229

methacrylic acid (MAA), ethyleneglycol dimethacrylate (EGDMA) azoisobutyronitrile (AIBN) and acetonitrile were obtained from Kermel (Tianjin, China). Phosphate buffer solution (PBS) was prepared using disodium hydrogen phosphate (Na2HPO4) and sodium dihydrogen phosphate (NaH2PO4), and the buffer was adjusted to appropriate pH with sodium hydroxide (NaOH) and phosphoric acid (H3PO4). The standard stock solution of MG was prepared at the concentration of 1 mM in double distilled water, and the standard stock solution of LMG at a concentration of 1.0 mM was prepared in acetonitrile, working standard solutions of MG and LMG were prepared fresh daily by stepwise dilution of standard stock solutions with 0.01 mol/L NaOH and acetonitrile respectively. They were stored at 4 °C in dark. Sample solution was prepared by diluting stock solution with phosphate buffer just before measurements. All other reagents were of analytical reagent grade and used without further purification. Cyclic voltammograms and ECL experiments were performed using a model MPI-A electrochemiluminescence analyzer system (Xi'an Remax Electronic Science & Technology Co. Ltd., Xi'an, China). A three-electrode system was used with a Φ3 mm glass carbon disk (GC) as the working electrode, an Ag/AgCl as the reference electrode (in saturated KCl solution), and a platinum wire as an auxiliary electrode. And all potentials were measured and reported according to this reference electrode. A home-made cylindroid quartz cell was used as the ECL cell [28], and was directly placed in front of the photomultiplier tube. A high voltage of 800 V was supplied to the PMT for intensity determination. UV–vis absorption spectrum was taken with a UV-1102 UV–vis spectrophotometer (Tianmei Scientific Instrument Co. Ltd., Shanghai, China). The infrared spectrum (IR) was measured by FT-IR spectrometer (Bruker Vertex 70v). TEM was taken using a Hitachi H-8100 electron microscope at an accelerating voltage of 200 kV (Hitachi, Tokyo, Japan). The surface morphology was observed through scanning electron microscopy (SEM) on JSM6701F (Japan Electron Optics Company). X-ray diffraction (XRD) study was conducted by D/max2400 (Pannlytical X'Pert PRO). 2.2. Preparation of nano-Fe3O4@SiO2 Fe3O4 nanoparticles were prepared by the co-precipitation method [29]. 5.41 g of FeCl3  6H2O and 1.27 g of FeCl2  6H2O were dissolved in 100 mL water. Then, the mixture was added into 10 mL of ammonia solution (25%, v/v) dropwise under N2 atmosphere at 90 °C with mechanical stirring (700 rpm). After 60 min, the black nano-Fe3O4 was collected with the help of an external magnet and washed several times with water and ethanol. Then nano-Fe3O4 was modified with SiO2 according to the paper of Chen et al. [30]. Briefly, nano-Fe3O4 (1 g), ammonia (3 mL) and TEOS (1 mL) were added into 150 mL of 50% aqueous alcohol and the mixture was stirred for 12 h. The prepared Fe3O4@SiO2 was dried under vacuum at 60 °C and then modified with MPS which introduced polymerizable double bonds. 1 g of Fe3O4@SiO2 was dispersed into 50 mL toluene containing 1 mL MPS, and the mixed solution was refluxed at 90 °C for 24 h under the protection of nitrogen. The surface-modified product (Fe3O4@SiO2–MPS) was collected by an external magnetic field and then washed with ethanol and dried under vacuum at 60 °C. 2.3. Preparation of MMIP and MNIP The preparation of MMIP was shown in Scheme 1. MG as template molecule, MAA as functional monomer and EGDMA as crosslinker, MMIP was prepared by bulk polymerization. 2 mmol MAA and 0.5 mmol MG were mixed in 30 mL of acetonitrile and prepolymerized under stirring at 150 rpm for 12 h at room temperature. Then, 10 mmol EGDMA and 20 mg AIBN were added into the mixture. The mixture was degassed and purged with nitrogen for

230

B. Huang et al. / Talanta 142 (2015) 228–234

2.7. Preparation of fish samples

Scheme 1. Schematic representation of the possible process of MG MMIP.

10 min. Subsequently, 200 mg Fe3O4@SiO2–MPS was added into this solution followed by sonication for 30 min. The polymerization was performed at 60 °C under N2 atmosphere for 24 h. In the control experiment, the non-imprinted polymer (MNIP) was prepared in the same procedure without the addition of template molecules. The product was separated by external magnetic field and washed thoroughly with a mixture of methanol/acetic acid (volume ratio 9:1) and redistilled water for three times to elute the template MG. Finally, these polymers were dried under vacuum at 60 °C for 12 h. 2.4. Adsorption studies To investigate the adsorption capacity, adsorption isotherm and kinetic experiments were performed. Adsorption isotherms were carried out as follows: 10 mg MMIP or MNIP was equilibrated with a series of MG concentrations (10, 20, 50, 100, 200 and 400 μM) in 5 mL aqueous solution at room temperature. The mixtures were shaken for 12 h at room temperature. Then MMIP (or MNIP) was separated by an external magnetic field and the supernatant was analyzed by UV–vis at 617 nm. The dynamics method was the same as adsorption isotherm method, except for different times (20, 40, 60, 90, 120, 150, 180 and 240 min) at a constant concentration (200 μM). The adsorption capacity (Q) of MG was calculated from

Q=

(C0 − C e )VM m

(1) 1

where Q (mg g ) is the amount of MG adsorbed onto the unit of MMIP or MNIP, C0 and Ce (μM) are the initial and equilibrium concentrations of MG, respectively. M is the molecule weight of MG. V (mL) is the volume of initial solution and m (g) is the weight of MMIP (MNIP).

The fish samples were purchased from local supermarket. Before extraction, individual fish was beheaded, boned, skinned, minced, crushed and homogenized. Homogenized blank tissue samples were spiked with MG and LMG at different mass ratios indicated in the text, and incubated 30 min at ambient temperature prior to the extraction and further process as described below [7,31]. The fish sample extraction method used in this study was according to a previous publication [32] with minor modifications. 5.0 g of homogenized tissue or 5.0 g of spiked sample (homogenized blank tissue sample was spiked with MG and LMG at different mass ratios indicated in the text) was placed into a 50 mL Teflon centrifuge tube. 20 mL of acetonitrile, 0.5 mL of 100 g/L hydroxylamine hydrochloride, 0.5 mL of 0.1 M ammonium acetate and 0.20 mL of 1 M p-toluenesulfonic acid were added. The sample was then vortexmixed for 1 min. Subsequently, 2 g of NaCl and 10 g of neutral Alumina-N were added, followed by vortexing for 5 min. Then, the mixture was centrifuged for 10 min at 4000 rpm, and the supernatant was collected in a 50 mL tube. Residual tissues were repeatedly extracted twice with 10 mL of acetonitrile each time. The three supernatants were collected and concentrated approximately to 1 mL, followed by the addition of 4 mL of 0.001 M DDQ solution which oxidized the colorless LMG to MG [13]. 50 mg MMIP was added to the above mixture, and then shaken at room temperature for 20 min. A magnet was used to separate MMIP from the solution. Then, the MMIP was eluted with 1 mL of methanol/acetic acid (9:1, v/v) by sonication for 15 min. Supernatants were completely evaporated to dryness and redissolved in 250 μL of water with 3 min of mixing vortex. Finally, 100 μL of the resulting solution was used for ECL analysis. 2.8. ECL analysis Prior to ECL measurements, the GC working electrode was polished with 0.3 and 0.05 μm Al2O3 slurry on silk cloth. Firstly, 200 μL of 5 mM Ru(bpy)32 þ , 100 μL of 1 mM TPA and certain concentration sample solution were added to a 5.0 mL volumetric flask, Then, they were diluted to 1000 μL with 0.1 M PBS (pH¼ 8.5), and 600 μL mixed solution was transferred to the ECL cell. A cyclic voltammetry was scanned in the range of 0.2–1.4 V with the scan rate of 100 mV/s, and the ECL signal was then recorded. Determination was based on the ECL intensity ΔI, (ΔI¼ I0 I, where I0 is the background ECL intensity of Ru(bpy)32 þ /TPA in the absence of MG, and I is the intensity in the presence of MG). All experiments were carried out at room temperature.

2.5. Adsorption selectivity

3. Results and discussion

The adsorption selectivity of MMIP was investigated with crystal violet (CV) as a structure analog of MG template. 5 mL of 200 μM individual standard mixture prepared with water was agitated with 10 mg MMIP or MNIP for 120 min, the concentration of CV in supernatant was measured at the wavelength of 590 nm and the following experiment was the same as Section 2.4.

3.1. Characterization of MMIP

2.6. Reusability of MMIP Regeneration experiment was conducted to investigate whether MMIP could be desorbed/released and reused. After adsorption of MG onto MMIP, MMIP was regenerated with methanol/ acetic acid (9:1, v:v) and methanol successively, dried in vacuum. Then the regenerated MMIP was reused to adsorb MG in the next cycle adsorption experiment.

The morphology and size of MMIP were investigated with SEM and TEM images (in the Supplemental information Fig. S1). TEM indicated that the imprinting layer was apparent. And SEM showed that MMIP was of uniform spherical morphology. FT-IR spectra of Fe3O4, Fe3O4@SiO2 and MMIP were compared in Fig. S2. In Fe3O4 curve, the peak at 567 cm  1 was attributed to Fe–O bond. In Fe3O4@SiO2 curve, Fe–O bond and Si–O–Si bond were at 571 and 1092 cm  1, respectively. In MMIP curve, Fe–O bond and Si–O–Si bond were at 580 and 1099 cm  1, respectively, these results proved that Fe3O4 was embedded in these materials. The weak peaks of C ¼O and C ¼C at 1635 and 1602 cm  1, respectively, demonstrated that MAA was cross-linked [33]. Fig. S3 compared XRD patterns of Fe3O4, Fe3O4@SiO2 and MMIP. In the 2θ region of 5–80°, six characteristic peaks for Fe3O4

B. Huang et al. / Talanta 142 (2015) 228–234

(2θ ¼30.2°, 35.5°, 43.1°, 53.4°, 57.0°, and 62.6°) were all observed in the three curves. The results revealed that the crystal structure of Fe3O4 remained stable in polymerization process and Fe3O4 was incorporated into MMIP [34]. VSM was employed to study magnetic properties of MMIP. The magnetic hysteresis loop of MMIP was shown in Fig. 2. It was apparent that there was no hysteresis for three curves, suggesting that Fe3O4, Fe3O4@SiO2 and MMIP were superparamagnetic. The saturation magnetization of Fe3O4, Fe3O4@SiO2 and MMIP was 59.13,

231

21.64 and 17.42 emu g  1, respectively. In the inserted photograph of Fig. 2, with the external magnetic field, the brown dispersed MMIP was easily attracted to the wall of vial. The result showed that MMIP kept enough magnetic response to meet the need of magnetic separation. 3.2. Adsorption kinetics In order to examine the adsorption rate, the dynamic method was carried out. Also, the pseudo-first-order and the pseudo-second-order equations were applied to investigate the adsorption mechanism. The pseudo-first-order equation can be expressed as [35]

ln (Q e − Qt ) = ln Q e − k1t

(2)

The pseudo-second-order equation can be expressed as [36]

1 t t = + Qe Qt k2 Q e 2

Fig.2. Hysteresis loops of Fe3O4, Fe3O4@SiO2, and MMIP. The inset shows the separation and re-dispersion processes of a solution of MMIP in the absence (left) and presence (right) of an external magnetic field.

(3)

where Qe is the equilibrium capacity and Qt is the adsorption capacity at different time. t is the sorption time. k1 and k2 are the rate constants of pseudo-first-order and pseudo-second-order adsorption, respectively. As seen in Fig. 3A, the adsorption capacity of MG by MMIP increased rapidly in the first 120 min and then gradually until equilibrium. In addition, the adsorption amount of MMIP for MG was higher than that of MNIP, which was attributed to the good imprinting effect of MMIP for MG. From Fig. 3B and C, it could be seen that the kinetic data was better fitted to pseudo-second-order equation according to the correlation coefficient.

Fig. 3. (A) Adsorption dynamics curve of MMIP and MNIP. (B) Pseudo-first-order curves of MMIP and MNIP. (C) Pseudo-second-order curves of MMIP and MNIP. (D) Intraparticle diffusion curves of MMIP and MNIP.

232

B. Huang et al. / Talanta 142 (2015) 228–234

Considering that the pseudo-second-order model could not identify the diffusion mechanism, intraparticle diffusion model was tested to identify the diffusion mechanism [37,38]:

Q t =ki t 1/2

(4)

where ki is the intraparticle diffusion rate constant. If the intraparticle diffusion is the rate-controlling step, it should be a straight line passing through the origin. As shown in Fig. 3D, the intraparticle diffusion plot for MMIP exhibited multi-linearity, indicating that the intraparticle diffusion was not the only rate controlling step, but also other processes might control the rate of adsorption. This kind of multilinearity was also observed in the adsorption of MG onto MNIP.

expressed as follows [39]:

Q max − Q Q = Ce KD

(5)

where Q is the amount of MG bound to polymers at equilibrium; Ce is the free MG concentration at equilibrium; Qmax is the maximum binding amount and KD is the dissociation constant. The values of KD and Qmax can be calculated from the slope and intercept of the linear line plotted in Q/Ce versus Q. As seen from Fig. 4B, the Scatchard plot for MMIP consisted of two linear parts with different slopes, indicating that there were

3.3. Adsorption isotherms and Scatchard analysis The adsorption isotherms of MG on MMIP and MNIP were plotted in Fig. 4A. It was clearly demonstrated that the adsorption amount of MG on MMIP and MNIP first increased sharply, and then slightly and finally reached equilibrium. The adsorption capacity of MMIP for MG was much higher than that of MNIP, which suggested that the imprinted cavities of MMIP formed might cause the high affinity binding for the template. In order to further evaluate the specificity, equilibrium data were analyzed by Scatchard equation. Because it could indicate how many kinds of binding sites exist in MMIP. Scatchard plot equation is Fig. 5. Adsorption selectivity of MMIP. Adsorbents dose: 10 mg, initial MG and CV. Concentration: 200 μM, solution volume: 5 mL, contact time: 2 h.

Fig. 6. Effect of recycle times on binding amount of MMIP.

Fig. 4. (A) Isothermal adsorption curves of MG on MMIP and MNIP. (B) Starchard analysis curves of MMIP and MNIP.

Fig. 7. Effect of pH on ECL intensity.

B. Huang et al. / Talanta 142 (2015) 228–234

two classes of binding sites in MMIP and the binding sites in MMIP were heterogeneous. While the Scatchard plot for MNIP was a single straight line which indicated that the binding sites of MNIP were homogeneous [40].

233

3.5. Reusability of MMIP Fig. 6 showed the adsorption capacity of MMIP for MG in adsorption–desorption cycles. The fifth rebinding amount of MG on MMIP was still as high as 70% of the first one, indicating that MMIP had certain ability to reuse and regenerate.

3.4. Adsorption selectivity 3.6. Optimization of ECL conditions The binding of MMIP for MG was compared to CV because both of them had similar chemical structures and often coexisted in fish samples. As shown in Fig. 5, comparison of MMIP binding with MNIP showed that MMIP was significantly specific for MG. Comparatively, MNIP was affected obviously by the interferent.

The ECL intensity was affected by a number of factors, including the mode of applied potential, scan rate and the pH of buffer solution. In order to obtain a higher sensitivity of ECL, all these factors were optimized. Different modes of applied potential were tried. Results showed that good and steady ECL signal in ECL quenching was obtained using cyclic voltammetry mode. Different scan rates were used to examine the ECL behavior of the Ru(bpy)32 þ /TPA–MG system. It was found that the maximum ΔI could be obtained at a scan rate of 100 mV s  1. Solution pH was known to have a considerable effect on aqueous ECL intensity [41]. In order to find out the optimal pH condition for the system, a series of PBS with different pH values were used. As shown in Fig. 7, ΔI increased with the pH value till the maximum ΔI appeared at the pH value of 8.5 and then decreased slightly. At low pH values (pHo5), both Ru(bpy)32 þ and Ru(bpy)32 þ /TPA had very low ECL. Thus, the optimized pH value was set at 8.5 for further experiments. 3.7. ECL analytical method validation

Ru(bpy)32 þ /TPA

Fig. 8. ECL–E curves of the system with different concentrations of MG. Concentration of MG (nM) (a) 0, (b) 0.8, (c) 16, (d) 200, (e) 400, and (f) 800. Inset: the linear relationship between the ΔI and the concentration of MG/CV. Table 1 Determination of MG residue in real samples (n¼ 6) by the ECL method. Sample

Spiked (μg/kg)

Found (μg/kg)

RSD (%)

Recovery (%)

1 2 3

3.65 (MG/LMG¼ 5/5)a 3.65 (MG/LMG¼ 2/8) 3.65 (MG/LMG¼ 8/2)

3.08 3.36 3.11

7.4 9.1 7.6

77.1–91.9 83.0–101.2 77.6–92.8

a MG/LMG ¼ 5/5 means the solution was obtained by mixing MG and LMG in a mass ration of 5:5. Similarly MG/LMG¼ 2/8 and MG/LMG¼ 2/8 mean mass ratios of 2/8 and 2/8.

Based on previous studies [19,42], we knew that trace MG showed efficient quenching over Ru(bpy)32 þ /TPA, the high Ksv (1.2  105 M  1) implied that the method was sensitive, and the quenching mechanism was also proposed. Resonance energy transn fer from excited-state luminophore Ru(bpy)32 þ to MG and dynamic quenching are suggested as the quenching mechanism. Under the optimized conditions, ΔI was well linear with the concentration of MG in the range of 0.29–290 μg/kg (Fig. 8). The regression equation of MG was ΔI¼ 1955þ11.6C (nM) with a correlation coefficient of 0.996. The limit of detection (LOD) was 0.036 μg/kg for MG (S/N¼3). Considering the concentration of MG in fish samples would increase 5 times in MMIP–SPE procedure, the method detection limit of MG in fish samples could be 7.3 ng/kg. These values were lower than

Table 2 Comparing performance of literature reported with our proposed MMIP–ECL. Method

Linear range

LOD

Recovery (%)

Reference

HPLC–UV HPLC–MS/MS

MGþLMG: 2.5–2000 μg/kg MG: 5–500 ng/mL LMG: 1–100 ng/mL MG: 5–1000 ng/mL MG: 10.0–250 ng/mL LMG: 10.0–250 ng/mL MGþLMG: 0.5–250 ng/mL MG: 9.9–800 ng/mL MGþLMG: 1–100 ng/g MG: 1  10  8–5  10  7 M MG:0.1–10 μg/L LMG: 0.1–10 μg/L MG: 0.05–200 μg/L MGþLMG: 0.13–15 ng/mL MGþLMG: 0–100 ng/mL MGþLMG: 1  10  11–5  10  9 M MGþLMG: 0.29–290 μg/kg

1 μg/kg 0.13 ng/mL 0.06 ng/mL 5 ng/mL 0.13 ng/mL 0.12 ng/mL 0.28 ng/mL 2.9 ng/mL 3 ng/kg 4.1  10  9 M 0.007 μg/kg 0.006 μg/kg 0.05 μg/L 0.15 ng/mL 0.16 ng/mL 3.0  10  12 M 7.3 ng/kg

71.0–101.0 71.7–113.3 89.5–105.0 76.8–93.7 89.8–95.8 90.6–97.5 92.5–104.5 93.14–104.6 86–110 94–98 88.2–94.6 82.1–97.8 70.73–77.61 71.8–95.8 70.3–85.7 84.5–96.6 77.1–101.2

[1] [32]

HPLC–UV HPLC–UV UV–vis UV–vis HPLC–MS/MS UV–vis LC–MS/MS HPLC–UV HPLC–UV HPLC–UV Luminol–ECL Ru(bpy)32 þ –ECL

MGþ LMG: after LMG was oxidized to MG, the sum of MG and LMG was determined.

[23] [7] [27] [15] [8] [16] [2] [9] [12] [13] [26] This paper

234

B. Huang et al. / Talanta 142 (2015) 228–234

2 μg/kg for the sum of MG and LMG proposed by the European Commission, indicating that this method was very sensitive for the determination of MG. The proposed method was validated using a series of fish samples. No MG at detectable levels was found in fish samples. The accuracy of the method was expressed as recovery, which was carried out by analyzing blank samples spiking with standard MG and LMG at different mass ratios indicated in Table 1. The recovery rate of MG in fish samples was higher than 77.1%, indicating the proposed method was accurate. To further investigate the performance of the proposed method, we compared the results with other methods in Table 2. It could be seen that this method exhibited remarkable advantages, such as higher sensitivity, wider linear range and lower detection limit.

4. Conclusions In this study, a novel, selective and sensitive method for the determination of trace levels of MG from fish samples was proposed based on combining highly sensitive ECL with highly selective MMIP. MMIP was prepared as selective solid-phase extraction of MG in fish samples. MMIP was characterized by TEM, SEM, XRD, FI-IR and VSM. The recognition capacity of MMIP for MG was 2.5 times than that of MNIP. These results showed that MMIP had strong magnetic responsiveness and excellent selection properties. MMIP was used to extraction of MG from fish samples using an external magnetic field. The proposed method based on MMIP–ECL was successfully applied to selective enrichment and determination of MG in spiked fish samples. The high recovery proved that the method was valid for the analysis of MG in fish samples. In a word, the analytical method was satisfactory due to the efficient pretreatment of sample.

Acknowledgments This work was supported by the National Natural Science Foundation of China (Nos. 21175108, 21165016, and 21165015), the Science and Technology Support Projects of Gansu Province (Nos. 090GKCA036 and 1011GKCA025), the Key Laboratory of Polymer Materials of Gansu Province, China and Key Fund Project of Sichuan Provincial Department of Education (13ZA0112).

Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.talanta.2015.04.053.

References [1] A.A. Bergwerff, P. Scherpenisse, J. Chromatogr. B 788 (2003) 351–359. [2] Y. Tao, D. Chen, X. Chao, H. Yu, P. Yuanhu, Z. Liu, L. Huang, Y. Wang, Z. Yuan, Food Control 22 (2011) 1246–1252. [3] S. Srivastava, R. Sinha, D. Roy, Aquat. Toxicol. 66 (2004) 319–329. [4] L. Chen, Y. Lu, S. Li, X. Lin, Z. Xu, Z. Dai, Food Chem. 141 (2013) 1383–1389. [5] A. Stammati, C. Nebbia, I.D. Angelis, A.G. Albo, M. Carletti, C. Rebecchi, F. Zampaglioni, M. Dacasto, Toxicol. In Vitr. 19 (2005) 853–858. [6] C.D. 2002/657/EC, Off. J. Eur. Commun. 221 (2002) 28. [7] C. Long, Z. Mai, Y. Yang, B. Zhu, X. Xu, L. Lu, X. Zou, J. Chromatogr. A 1216 (2009) 2275–2281. [8] M.J. Martínez Bueno, S. Herrera, A. Uclés, A. Agüera, M.D. Hernando, O. Shimelis, M. Rudolfsson, A.R. Fernández-Alba, Anal. Chim. Acta 665 (2010) 47–54. [9] Z. Lian, J. Wang, Mar. Pollut. Bull. 64 (2012) 2656–2662. [10] C. Nebot, A. Iglesias, R. Barreiro, J.M. Miranda, B. Vázquez, C.M. Franco, A. Cepeda, Food Control 31 (2013) 102–107. [11] H.R. Rajabi, O. Khani, M. Shamsipur, V. Vatanpour, J. Hazard. Mater. 250–251 (2013) 370–378. [12] J. Xie, T. Peng, D.D. Chen, Q.J. Zhang, G.M. Wang, X. Wang, Q. Guo, F. Jiang, D. Chen, J. Deng, J. Chromatogr. B 913–914 (2013) 123–128. [13] A.A. Fallah, A. Barani, Food Control 40 (2014) 100–105. [14] C.H. Tsai, J.D. Lin, C.H. Lin, Talanta 72 (2007) 368–372. [15] L. An, J. Deng, L. Zhou, H. Li, F. Chen, H. Wang, Y. Liu, J. Hazard. Mater. 175 (2010) 883–888. [16] M. Bahram, F. Keshvari, P. Najafi-Moghaddam, Talanta 85 (2011) 891–896. [17] A.J. Bard, Electrogenerated Chemiluminescence, Marcel Dekker, New York, 2004. [18] M.M. Richter, Chem. Rev. 104 (2004) 3003–3036. [19] B. Huang, X. Zhou, Z. Xue, G. Wu, J. Du, D. Luo, T. Liu, J. Ru, X. Lu, Talanta 106 (2013) 174–180. [20] W. Yang, F. Jiao, L. Zhou, X. Chen, X. Jiang, Appl. Surf. Sci. 284 (2013) 692–699. [21] L. Chen, X. Zhang, Y. Xu, X. Du, X. Sun, L. Sun, H. Wang, Q. Zhao, A. Yu, H. Zhang, L. Ding, Anal. Chim. Acta 662 (2010) 31–38. [22] Y.H. Li, T. Yang, X.L. Qi, Y.W. Qiao, A.P. Deng, Anal. Chim. Acta 624 (2008) 317–325. [23] L.Q. Su, S. Qiao, W.B. Zhang, Chin. Chem. Lett. 18 (2007) 229–232. [24] S. Yan, Z. Gao, Y. Fang, Y. Cheng, H. Zhou, H. Wang, Dyes Pigment. 74 (2007) 572–577. [25] C. Long, Z. Mai, B. Zhu, X. Zou, Y. Gao, X. Huang, J. Chromatogr. A 1203 (2008) 21–26. [26] Z. Guo, P. Gai, T. Hao, J. Duan, S. Wang, J. Agric. Food Chem. 59 (2011) 5257–5262. [27] A. Afkhami, R. Moosavi, T. Madrakian, Talanta 82 (2010) 785–789. [28] X. Lu, H. Wang, J. Du, B. Huang, D. Liu, X. Liu, H. Guo, Z. Xue, Analyst 137 (2012) 1416–1420. [29] L. Fan, C. Luo, Z. Lv, F. Lu, H. Qiu, J. Hazard. Mater. 194 (2011) 193–201. [30] F.F. Chen, X.Y. Xie, Y.P. Shi, Talanta 115 (2013) 482–489. ́ M. Šafařıková, ́ [31] I. Šafařık, Water Res. 36 (2002) 196–200. [32] K.C. Lee, J.L. Wu, Z. Cai, J. Chromatogr. B 843 (2006) 247–251. [33] F.F. Chen, X.Y. Xie, Y.P. Shi, J. Chromatogr. A 1252 (2012) 8–14. [34] J. Pan, H. Yao, L. Xu, H. Ou, P. Huo, X. Li, Y. Yan, J. Phys. Chem. C 115 (2011) 5440–5449. [35] Y.S. Ho, G. McKay, Water Res. 34 (2000) 735–742. [36] Y.S. Ho, G. McKay, Process Biochem. 34 (1999) 451–465. [37] Y.S. Ho, G. McKay, Process Saf. Environ. Prot. 76 (1998) 183–191. [38] B.H. Hameed, M.I. El-Khaiary, J. Hazard. Mater. 153 (2008) 701–708. [39] E.M. Perdue, C.R. Lytle, Environ. Sci. Technol. 17 (1983) 654–660. [40] C. Piao, L. Chen, J. Chromatogr. A 1268 (2012) 185–190. [41] H. Qiu, J. Yan, X. Sun, J. Liu, W. Cao, X. Yang, E. Wang, Anal. Chem. 75 (2003) 5435–5440. [42] B. Huang, X. Zhou, Z. Xue, X. Lu, Anal. Chem. 51 (2013) 107–116.