DOI:
10.1039/C6RA03787A
(Paper)
RSC Adv., 2016,
6, 41675-41686
A combination of β-cyclodextrin functionalized magnetic graphene oxide nanoparticles with β-cyclodextrin-based sensor for highly sensitive and selective voltammetric determination of tetracycline and doxycycline in milk samples†
Received
10th February 2016
, Accepted 11th April 2016
First published on 13th April 2016
Abstract
Food safety problems caused by tetracycline antibiotics residues are a significant concern due to their great risk to human health even at trace levels. In the current work, a highly sensitive and selective determination of similar chemical composition tetracycline (TC) and doxycycline (DC) antibiotics in milk samples was developed by offline solid phase extraction using β-cyclodextrin functionalized magnetic graphene oxide nanoparticles (β-CD-MGONPs) followed by differential pulse voltammetric determination with a β-cyclodextrin modified carbon paste (β-CD-MCP) sensor. β-CD-MGONPs was prepared, characterized by XRD, FTIR, TEM, and TGA and applied for the removal of TC/DC using batch adsorption experiments. The effects of medium acidity, equilibration time, adsorbent dose, initial antibiotic concentration, and ionic strength were extensively investigated using UV-Vis spectrophotometry at optimal adsorption conditions. The experimental data are well described by the Langmuir isotherm and pseudo-second-order kinetic model. The maximum adsorption capacity was found to be 666.7 mg g−1 for TC and 769.2 mg g−1 for DC. Furthermore, loss in the removal efficiency was 3.5–5.6% after three adsorption–desorption cycles. Then, the electrochemical response of the β-CD-MCP sensor following β-CD-MGONPs extraction was investigated by differential pulse voltammetry (DPV) and cyclic voltammetry (CV). Under optimal electrochemical parameters, anodic differential pulse voltammetry (ADPV) at +320.0 mV and cathodic differential pulse voltammetry (CDPV) at −800.0 mV were selectively used to detect TC and DC, respectively. The response currents of the β-CD-MCP sensor exhibited a linear relationship towards TC/DC concentrations ranging from 0.5–90.0 ng L−1. The limit of detection (LOD) of TC or DC was calculated as 0.18 ng L−1. Other validation parameters confirmed the adequate applicability of the proposed system to detect TC/DC in milk samples. The proposed platform showed advantages of simplicity, rapidity, reliability, and low cost compared to other previously published reports.
1. Introduction
Recently, food safety problems, especially those caused by antimicrobial residues in animal origin foods, have received great interest from researchers. Tetracycline (TCs) antibiotics are one of the greatest concerns of these food contaminants because of their broad-spectrum activity and low cost compared with other available antibiotics. TCs are a group of bacteriostatic compounds that act by inhibiting protein biosynthesis and are broadly used in the treatment of respiratory and urinary tract infections. Currently, over 20 tetracyclines (TCs) are widely approved for use in food producing animals at sub-therapeutic doses as growth promoters with tolerances ranging from 0.1 to 2 mg L−1 in commonly consumed animal tissue and products.1–3 Tolerances for less commonly consumed tissue reach 12 mg L−1.4 Tetracycline (TC) and doxycycline (DC) (Fig. 1, similar in chemical composition) are members of the TCs that acted by binding to the 30 S ribosomal sub-particle in a bacterial cell. Improper use of doxycycline in food producing animals can result in residues in animal tissues, which can be toxic and dangerous for human health and may generate resistance of many microorganisms to antibiotics. In dairy cows, significant percentages of administered TCs are excreted in milk. To protect human health from exposure to these drug residues in milk, the EU has established a maximum residue limit (MRL) of 100 μg kg−1 for TC, while in the USA, the food and drug administration (FDA) has established a MRL of 300 μg kg−1 for the combined residues of TC. DC is not licensed in the aforementioned countries for use in animals that produce milk for human consumption, so no MRL has been set for this drug. Sample preparation, including isolation, clean-up, and preconcentration of TCs residues from their matrices are the crucial parameters for the analytical determination of TCs in samples of different matrices. Capillary electrophoresis (CE),4,5 spectrophotometry,6,7 chemiluminescence,8–10 high performance liquid chromatography (HPLC),11–13 and microbiological assays14 are used for the analytical determination of TCs residues in food individually or consecutively according to the nature of the samples as well as the matrix effect.15 However, some of these techniques suffer from various drawbacks regarding tedious preparation steps. Therefore, it is necessary to establish a new method with simplicity, high selectivity, and sensitivity for determination of TC/DC in complex samples.
|
| Fig. 1 Chemical structures of TC and DC. | |
Solid phase extraction (SPE) is a significant separation and preconcentration technique to reduce the influence of interferences with additional advantages of simplicity, low cost, and ease of use. Lately, magnetic solid phase extraction (MSPE) based on functionalized graphene oxide nanosorbents have attracted considerable attention16–20 in biological applications, and in analytical applications,21 e.g. preconcentration of organic compounds from large volume samples.22–25 The magnetic GO nanosorbents offered fast separation modes as well as the ease of removal of highly dispersed GO together with the analyte species from the sample solution.26 Moreover, GO has high hydrophilicity, as it contains a huge number of oxygen containing functional groups (hydroxyl, epoxide, and carboxylic acid groups) which could readily link with a highly selective agent such as cyclodextrins (CDs). CDs are cyclic oligosaccharides composed of glucopyranose units and can be represented as a truncated cone structure with a hydrophobic cavity. The inner cavity diameters of α-, β-, and γ-CDs are ca. 5.7, 7.8, and 9.5 A, respectively. The depth of the cavity is 7.8 A. These dimensions coupled with an apolar cavity give CDs the ability to include hydrophobic guest molecules within the cavity based on size. β-Cyclodextrin has the capability to bind in a selective manner with various organic compounds by the formation of stable inclusions in its hydrophobic cavity.27 Magnetic graphene oxide–β-cyclodextrin nanocomposite (β-CD-MGONPs) has a special structure with unique properties, including (i) ease of separation and recovery by its magnetic properties, (ii) hydrophilicity of the GO, and (iii) hydrophobicity of the β-cyclodextrin moiety. Furthermore, it can be combined with an electrochemical sensor to detect a trace amount component in complex matrices. An electrochemical sensor showed attractive advantages of simplicity, sensitivity, and label-free to analyze molecules in complex matrixes.28–30 However, there are also several problems involving non-specific binding, poor regeneration, and limited response for electrochemical sensor applications. Fortunately, these problems can be overcome by using highly selective sensing materials.
Molecular recognition at the surface of solid materials has attracted interest of researchers who are trying to realize functional materials for bio and chemical sensors.31–39 β-cyclodextrin can be viewed as a molecular receptor that can accommodate a wide variety of organic guest molecules to form stable host–guest inclusion complexes in their hydrophobic cavity, showing high molecular selectivity.28,29 Furthermore, β-cyclodextrin was impregnated in the carbon paste electrode forming a β-cyclodextrin modified carbon past sensor (β-CDMCP sensor) in order to increase the selectivity of measurements in commercial samples.
In the present work, by using β-cyclodextrin in both the β-CD-MGONPs and β-CDMCP sensor, double selectivity and high sensitivity were achieved to remove and detect TC/DC in milk samples. First, β-CD-MGONPs was prepared, characterized by XRD, FTIR, TEM and TGA. Second, the optimization of medium acidity, equilibration time, adsorbent dose, initial antibiotic concentration, and ionic strength through offline solid phase extraction were extensively investigated using UV-Vis spectrophotometry. Then, electrochemical behavior of TC/DC at the β-CD-MGONPs was investigated by cyclic voltammetry (CV) and differential pulse voltammetry (DPV). Method validation parameters including selectivity, precision, accuracy, linearity, and detection limits were studied.
2. Materials and methods
2.1. Materials
All chemicals were of analytical grade and used without further purification. Doxycycline hydrochloride (98%), and tetracycline hydrochloride (95%) were obtained from Sigma (USA). Graphite powder was provided by Fluka (Buchs, Switzerland). Ferric chloride hexahydrate (FeCl3·6H2O), ferrous chloride tetrahydrate (FeCl2·4H2O), and β-cyclodextrin (β-CD) were purchased from Sigma-Aldrich (USA). Milk samples were supplied from local markets. Different varieties of milk were analyzed: (a) natural milk, (b) skimmed milk, 0% fat (fresh and high-temperature, short-time pasteurized milk), (c) semi-skimmed milk, 1.2% fat (fresh and high-temperature, short-time pasteurized milk) and (d) full-fat milk, 3.0% fat. All solutions throughout the experiments were prepared by dissolving the respective analytical grade reagent in deionized water with a resistivity not less than 18.0 M Ωcm, which was provided by a Milli-Q system (USA).
2.2. Instrumentation
All voltammetric investigations were performed in a 10.0 mL glass voltammetric cell using a commercial available electrode stand (Metrohm, Switzerland). The electrode was connected via an IME-663 module (Netherlands). Potentials were controlled using a 3-electrode configuration comprising a β-CDMCP sensor working electrode (3 mm diameter, Metrohm), an Ag/AgCl (3.00 M KCl) reference electrode, and a platinum wire counter electrode. The β-CDMCP sensor was prepared according to the method described by our group27 in which the graphite powder and paraffin oil were prepared in a ratio of 70:30% (w/w) followed by the addition of β-CD solution. The modified carbon paste was then filled carefully into a cavity of a carbon paste electrode. The treatment of electrode surfaces after each measurement with doubly distilled water (10 s) increased the stability of the membrane. In 10 deposition/measurement/regeneration cycles, the response could be reproduced with about 2% relative standard deviation. UV-Vis measurements were performed by a UV-Vis double beam model UVD-3500 (Labomed Inc., USA) spectrophotometer. The pH values were measured using a Fischer Scientific pH meter model 810 (USA) equipped with a combined glass electrode, which was calibrated regularly with buffer solutions (pH at 4.0 and 7.0) at 22 ± 2 °C.
2.3. Synthesis of β-CD-MGONPs
GO was prepared by oxidizing graphite powder according to a modified Hummers' method.40 MGONPs were prepared by in situ chemical co-precipitation of Fe2+ and Fe3+ in alkaline solution in the presence of GO. In brief, 50 mg of GO was added to 50 mL water and the dispersed solution was sonicated for 30 min. Then 0.5 g of FeCl3·6H2O and 0.2 g of FeCl2·4H2O were dissolved in 25 mL water and the resultant mixture was added to the GO solution. Finally, 75.0 mL of 25% NH4OH was added within 25–30 min to precipitate the Fe3O4 at 40 °C under N2 gas with magnetic stirring until the pH of solution mixture increased to 10–11. The black MGONPs were collected using an external magnet and washed six times with 300 mL of water. To prepare the β-cyclodextrin functionalized magnetic graphene oxide, 10 mL of 5 mg mL−1 MGO suspension was mixed with 10 mL of 5 mg mL−1 β-CD aqueous solution. After being vigorously shaken, the vial was heated in a water bath at 60 °C for 3.5 h.41,42 β-CD-MGO was precipitated, separated by a magnetic field for six cycles to remove the excess β-CD, washed with 100 mL water three times, 50 mL of anhydrous ethanol three times, and finally dried in a vacuum at 50 °C overnight.
2.4. Characterization of β-CD-MGONPs
β-CD-MGONPs were examined by Siemens, USA, ( Bruker Co, USA) D5000 X-ray diffractometer (XRD) at room temperature. In order to confirm the coating of GO and β-cyclodextrin onto Fe3O4 nanoparticles, the compositions of the nanoparticles were analyzed in dried KBr powder by recording the infrared spectra over the range of 4000–400 cm−1 using a Fourier transform infrared (FTIR) spectrophotometer (EQUINOX55, Bruker Co). The morphology and particle size of the synthesized nanoparticles were investigated by high-resolution transmission electron microscopy (HRTEM) using a Philips EM208 microscope operated at an acceleration voltage of 200 kV. The adsorbed mass of CDs on the surface of MGONPs was measured by thermogravimetric analysis (TGA) with a Polymer Lab TGA-1500, from 25 °C to 600 °C using a heating rate of 10 °C min−1 in a N2 atmosphere.
2.5. TC and DC adsorption studies
All batch adsorption experiments were carried out with a shaking rate of 250 rpm. The effect of the nanosorbent dosage on TC/DC removal was conducted using 25 mL of TC/DC solution (10 mg L−1, pH 7.0) with 1.0, 5.0, 10.0, 15.0, 20.0, 25.0, and 50.0 mg of nanosorbent at 25 °C for 20 min. The influence of pH on TC/DC removal was studied in the pH range of 2–10 with a contact time of 20 min using 20.0 mg of β-CD-MGONPs and 10 mg L−1 of TC/DC solution at 25 °C. The solution pH was adjusted by addition of proper amounts of 0.1 M HCl and 0.1 M NaOH solutions. In order to investigate the effect of contact time on the adsorption of TC/DC for kinetic studies, 25 mL of TC/DC solution (10 mg L−1 at pH 7.0) was shaken with 20 mg of nanosorbent at 25 °C for predetermined time intervals. TC/DC adsorption isotherms were determined by agitating 25 mL of TC/DC solutions with various initial concentrations (1–25 mg L−1) at pH 7.0 with 20.0 mg of β-CD-MGONPs at 25 °C until equilibrium was attained. In all experiments, the magnetic nanosorbent was separated from TC/DC solution by an external magnet. The absorbance of samples was determined using a double beam UV-Vis spectrophotometer at a wavelength of 356 nm for TC and 345 nm for DC. After equilibrium, the percentage of TC/DC removal (% R) and the equilibrium adsorption capacity (qe) was calculated according to eqn (1) and (2). |
| (1) |
|
| (2) |
where Ci, Ce (mg L−1) are the initial and equilibrium concentrations of TC and DC, respectively, qe (mg g−1) is the equilibrium adsorption capacity, V is the TC/DC solution volume, and m (g) is the mass of the magnetic nanosorbent.
2.6. β-CD-MGONPs coupled with β-CDMCP for TC and DC determination in milk samples
For extraction and preconcentration of the analytes in milk samples,40,43 20 mg of the magnetic nanosorbent was added into a beaker. The particles were conditioned with 5.0 mL of methanol in an ultrasonic bath for 5 min. Then, the nanosorbent was magnetically isolated and washed twice with 10.0 mL of deionized water for 3 min, and the supernatant was discarded. Milk samples included natural, skimmed, or semi-skimmed stored in plastic containers at 4 °C in the dark until used within four days in this study. A 10.0 mL sample of milk was mixed in a 50 mL plastic centrifuge tube with 30 mL of the McIlvane/EDTA solution. The solution was agitated for 1 min and centrifuged for 15 min (4000 rpm) or until the protein precipitated. The precipitate was disposed of and the solution was used for subsequent magnetic solid-phase extraction. Then, 25.0 mL of the treated milk sample solution was mixed with the pre-activated magnetic nanosorbent and the pH was adjusted to 7.0 by 0.1 M HCl or 0.1 M NaOH. TC/DC removal was conducted by shaking for 20 minutes at rate of 250 rpm. An external magnet was applied to isolate the nanosorbent with the adsorbed analytes. The liquid phase was decanted, while the solid phase was washed three times with 5.0 mL of water. The solution was collected directly into an electrochemical cell containing 5 mL of phosphate buffer (0.1 mol L−1, pH 7.0) to be detected with β-CDMCP by DPV.
3. Results and discussion
3.1. Morphology and structural characterization of β-CD-MGONPs
The particle size and morphology information of β-CD-MGONPs was studied by TEM images, as shown in Fig. 2A. The TEM image of the magnetic nanosorbent revealed irregular monodispersed nano-particles with a rough spherical surface and large surface area. The particle size distribution (PSD) of β-CD-MGONPs depicts a very narrow size distribution of nanoparticles with a mean particle size of 10 ± 3 nm.
|
| Fig. 2 (A) TEM image of β-CD-MGOMNPs, (B) FTIR spectra of GO and magnetic nanosorbents, (C) XRD of β-CD-MGOMNPs, and (D) TGA of β-CD-MGOMNPs. | |
Fig. 2B exhibits the FTIR spectra of GO, MGONPs, and β-CD-MGONPs. The wave numbers (cm−1) of the different absorption bands and their assignments for all nanosorbents are compiled in Table S1.† GO and MGONPs both showed the O–H stretching vibration adsorption peak at 3227.3 and 3220.2 cm−1, respectively. As to graphene oxide, the peaks at 1711.2 and 1400.2 cm−1 were the CO stretching vibration peaks of carboxyl and carbonyl; the peak at 1583 cm−1 was attributed to the stretching vibration of aromatic CC; the peaks at 1170.1 and 1121.4 cm−1 were ascribed to the C–O stretching vibration of epoxy and alkoxy groups.44 These peaks demonstrated the existence of carboxyl and epoxy groups in graphene oxide. Moreover, the Fe–O characteristic stretching vibration peak was observed at 550.3 cm−1 for MGONPs and 554.3 cm−1 for β-CD-MGONPs which proved that Fe3O4 nanoparticles were successfully anchored onto GO and β-CD-MGONPs. The major characteristic peaks that support the immobilization of β-cyclodextrin on the MGONPs can be assigned as 1027.2 cm−1 (R-1, 4-bond skeleton vibration of β-CD) and 3425 cm−1 (O–H stretching vibration).
Fig. 2C shows the wide-angle XRD patterns of MGONPs. The broad band centered at 2θ = 22° can be assigned to the presence of abundant oxygen-containing functional groups on GO and presence of intercalated water molecules between the layers which is similar to the reported value of graphene.45 Six characteristic peaks indicate the existence of the magnetite Fe3O4. The intense diffraction peaks indexed to (220), (311), (400), (422), (511), and (440) planes appear at 2θ = 30.04, 35.42, 43.03, 53.45, 57.00, and 62.64°, respectively.15,46,47 The small peak at 2θ = 18.30° might be related to β-CD.48 The broad nature of the XRD patterns can be ascribed to the small size of MNPs.49
TGA experiments were performed to investigate the amount of GO and β-CD bounded onto the surface of Fe3O4 nanoparticles and obtain information on their thermal stability. A typical thermogram of the magnetic nanosorbent specifies four stages of weight loss as presented in Fig. 2D. The weight loss, 15.5% observed in the first stage (27.6–108.7 °C), is due to the desorption of surface water molecules. In the second stage (109.2–286.2 °C), the weight loss is about 7.3 wt%, which was assigned to the thermal degradation of β-CD coated onto the surface of Fe3O4 nanoparticles. The weight losses in the third and fourth stages (467.8–600 °C) were about 12.1 and 17.8 wt%, respectively, and may be attributed to decomposition of GO residues.50,51 Therefore, these results suggest that β-CD-MGONPs are stable below 286.2 °C.
3.2. Adsorption studies
3.2.1. Effect of pH. The distribution of TC/DC species in aqueous solution and the surface characters of β-CD-MGONPs are highly pH-dependent. TC/DC are amphoteric molecules that have three acid dissociation constants (3.0–3.3, 7.8–8.0, and 9.2–9.6).52,53 The cationic species (H3TC+ or H3DC+) are predominant below pH 3, a zwitterion (H2TC or H2DC) exists between pH 3.0 and 8.0, and anionic forms (HTC−/TC2−or HDC−/DC2−) are predominant above pH 8.0. It is expected that uptake of TC/DC by Fe3O4–GO–β-CD MNPs is pH dependent. The adsorption behavior of TC/DC by the MNPs could result from a combination of speciation of TC/DC, surface charge properties of β-CD-MGONPs, and the possible complexation between TC/DC and β-CD-MGONPs. In this work, the influence of solution pH on TC/DC removal by β-CD-MGONPs was investigated via batch sorption experiments using 10 mg L−1 TC/DC solutions and presented in Fig. 3A. As observed in Fig. 3A, low percentage removal for TC/DC is observed at low pH values (2–3) and both DC/TC are well adsorbed to the surface of β-CD-MGONPs in the pH range 6–8 but both are best extracted at pH 7. At low pH values (2–3), the low uptakes of TC/DC are explained in terms of the strong repulsive forces experienced between TC/DC and β-CD-MGONPs. This is due to protonation of the dimethyl-ammonium group in the cationic species of TC/DC and the functional groups of GO and β-CD in the nanosorbent. As well, it is well known that cyclodextrin can be easily hydrolyzed under acidic conditions to a series of oligosaccharides ranging from an opened ring down to glucose.27 Likewise, at higher pH (9–10), electrostatic repulsion between deprotonated TC/DC molecules and negatively charged β-CD-MGONPs appeared. Thus, at extreme pH conditions, TC/DC uptakes could be determined by the electrostatic interactions. At relatively neutral pH (6–8), electrostatic attraction of functional groups and/or complexation of TC/DC and β-CD-MGONPs appeared and contributed to the adsorption of TC/DC.
|
| Fig. 3 (A) Effect of pH for 30 min contact time, (B) effect of contact time at pH = 7, (C) effect of β-CD-MGONPs dosage (mg), and (D) effect of ionic strength at pH = 7.0 for 30 min contact time, on the percentage removal of 10 mg L−1 TC and DC at 25.0 °C. | |
3.2.2. Effect of contact time. The effect of contact time on the TC/DC sorption processes by β-CD-MGONPs was studied at various time intervals (1, 5, 10, 15, 20, 25, 30, 35, 40, and 50 min) at pH 7.0 and represented in Fig. 3B. DC showed a steeper slope than that of TC, due to difference in the removal efficiencies of MNPs towards TC/DC. DC sorption is found to proceed via two steps in 30 minutes to attain the equilibrium time. The first step is characterized to produce 60–65% removal in the first 10 min. The second step is established in 10–30 min to complete saturation of MNPs surface. TC sorption was carried out in a single step to attain the equilibrium time within 30 min. Thus, one can conclude from this study that the interaction between TC/DC and β-CD-MGONPs is very fast.
3.2.3. Effect of β-CD-MGONPs dosage. The amount of β-CD-MGONPs controls the available number of active sites and, in turn, affects the MNPs–sorbate interactions. The effect of β-CD-MGONPs dosage on the percentage removal of TC/DC was evaluated at the optimized conditions (pH 7.0 and 30 min of contact time) and the results of this study are shown in Fig. 3C. It is evident from the graph that the maximum percentage removal was established for TC/DC using 20 mg of β-CD-MGONPs. The gradual increase in the percentage removal values at lower dosages (1–15 mg) is mainly attributed to the high availability of TC or DC species in solution compared to the active interaction sites. At higher dosages (25–100 mg), the decrease in the percentage removal values is due the formation of aggregates during sorption, causing a direct decrease in the effective uptake with respect to the exposed surface area at higher MNPs dosages.
3.2.4. Effect of ionic strength. Ionic strength is an essential parameter that controls the electrostatic attractions between TC/DC species and the adsorption sites in the MNPs. Fig. 3D shows the effect of ionic strength on percentage removal of TC/DC by β-CD-MGONPs. The removal efficiency is reduced from 96.0% to 81.4% for DC and from 93.0% to 77.3% for TC, as the ionic strength is increased with 0 to 0.5 M NaCl. This is attributed to the interaction of positively charged sodium ions with the deprotonated hydroxyl and carboxylic groups in MNPs. The results confirm that electrostatic interaction is a dominant route in the mechanism of the sorptive removal of TC/DC by β-CD-MGONPs.
3.2.5. Adsorption kinetics. Four different models were used to analyze the kinetic data of TC and DC on the β-CD-MGONPs. Lagergren pseudo-first order, pseudo-second-order,54,55 Elovich,56 and intra-particle diffusion57 kinetics models were applied to fit the experimental data. All the kinetic equations and their parameters are shown in Table 1. Based on the correlation coefficients (r2) and comparing the calculated qe values from the different kinetic models with the experimental value qe, it is concluded that the pseudo-second order is the most suitable kinetic model for sorption of TC and DC onto β-CD-MGONPs. In addition, DC adsorption by MNPs rapidly increased during the first 10 min, followed by a slower adsorption process, and the equilibrium was reached within 30 min. It is proposed that an external mass transfer dominates the initial adsorption stage, while in the second stage the rate-limiting step is intra-particle diffusion, causing a slower adsorption rate.
Table 1 Kinetic parameters used to analyze the adsorption of TC (10 mg L−1) and DC (10 mg L−1) onto β-CD-MGONPs at 25.0 °C
Kinetic model and parameters |
TC |
DC |
Kinetic equation |
qe (exp.) (mg g−1) |
11.609 |
11.984 |
|
Lagergren pseudo-first order equation |
qe (calc.) (mg g−1) |
8.870 |
48.960 |
ln(qe − qt) = lnqe − k1t, qe and qt are the sorption capacity at equilibrium and at time t (min), k1 is the first order rate constant |
k1 (min−1) |
0.0686 |
0.1005 |
R2 |
0.9691 |
0.9716 |
Pseudo-second-order equation |
qe (calc.) (mg g−1) |
11.765 |
12.048 |
, k2 is the second order rate constant |
k2 (min−1) |
377.032 |
474.686 |
R2 |
0.9703 |
0.9828 |
Elovich equation |
qe (calc.) (mg g−1) |
9.795 |
10.923 |
, α is the initial adsorption rate β is the desorption constant |
α (mg g−1 min−2) |
3.992 |
2.290 |
β (g mg−1 min−1) |
0.5149 |
0.4278 |
R2 |
0.8810 |
0.9390 |
Intra-particle diffusion equation |
qe (calc.) (mg g−1) |
9.953 |
11.039 |
, C is the intra-particle diffusion rate constant. kint is the intra-particle order rate constant |
kint |
1.503 |
1.739 |
C |
3.232 |
3.262 |
R2 |
0.9832 |
0.9680 |
3.2.7. Desorption and reusability of β-CD-MGONPs. In practical applications, the recycling studies are essential to determine regeneration performance of the nanosorbent. Based on the previous study of pH effect, the adsorption of TC/DC onto β-CD-MGONPs is insignificant at pH 1.0. Therefore, the loaded MNPs were re-suspended in 0.1 mol L−1 of HCl solution after sonication for 5 min to desorb TC or DC from the surface of MNPs with the aid of an external magnet. It was found that the percentage of recovery of the nanosorbent is 98.7% ± 2.6 in the case of TC and 99.2% ± 1.8 for DC. To estimate the reusability of MNPs, the adsorption–desorption was repeated three times, where the loss in the removal efficiency was found to be 4.3–5.6%, and 3.5–5.1% for TC and DC, respectively. These results indicate that β-CD-MGONPs has good reusability and stability.
3.2.8. Mechanism of adsorption of TC and DC. Based on the special structure and high surface area of the magnetic nanosorbent, as well as the excellent hydrophilicity of GO and β-CD, the adsorptive mechanism of TC/DC onto β-CD-MGONPs was proposed to be via a π–π donor–accepter and cation–π bonding interactions.58,59 The π–π stacking mechanism was considered as a dominant driving force to explain the interaction of four aromatic rings in the molecular structure of TC/DC to GO surface. On the other hand, the cation–π bonding is controlled by electrostatic attraction between cationic forms of tetracyclines (TC+/DC+) and the permanent quadrupole of the π-electron-rich aromatic structure in GO.60,61
3.3. Voltammetric behavior of TC and DC at β-CDMCP sensor
TC/DC was detected at the concentration of 500.0 ng L−1 in the presence of 0.05 mol L−1 phosphate buffer as supporting electrolyte, pH 7.0 at the plain carbon paste electrode (CPE, not shown). The electrochemical response of TC/DC under such conditions was rather poor with broad peak shape. The sensitivity and selectivity of electrochemical analysis can be enhanced using chemically modified electrodes. Cyclodextrins (CDs) were used as sensing material.27,28 In this case, the preconcentration of analyte on the electrode surface was based on the formation of a supramolecular complex (intercalation) between the CD host and the analyte. β-CD modified carbon paste sensor (β-CDMCP sensor) was used for the determination of several organic compounds with good selectivity and sensitivity.27,28 It was observed that β-CD caused a significant increase in the current response (about ten-fold) with a little shift to less positive potential. The signal enhancement and potential shifting can be interpreted as due to the formation of an inclusion complex between β-CD and TC/DC when its molecule entered into the CD cavity. The nature of the electrochemical process was carried out by applying cyclic voltammetry (CV) in the potential range from −1500 to 1500 mV (vs. Ag/AgCl). TC exhibited one cathodic peak at −730.0 mV and one anodic peak at +300.0 mV as indicated in Fig. 5A. The oxidation peak of TC may be due to the phenolic group being very close to the delocalized first aromatic ring from the left side (Fig. 1). On the other hand, DC shows one cathodic peak at −870.0 mV as indicated in Fig. 5B. For the effect of scan rate on 900.0 ng L−1 of TC/DC, it was observed that the maximum CV peak currents exhibited a linear dependence on the scan rate (correlation coefficient of 0.996) in the region of 100–1000 mV s−1, which is typical for the signal of a surface-attached reactant. This phenomenon indicates that the electrochemical process is controlled by an adsorption process, which increased the surface concentration of this species to different extents, thus increasing the current sensitivity. Repeating the scan immediately, the anodic and cathodic peaks decreased both in solution with the TC or DC (no stirring was applied between the scans) and without the TC or DC, which indicates the dissociation of the guest analyte molecule from its complex with host β-CD. Thus, the voltammetric technique with a rapid scan rate was required for such measurements. As well, in order to achieve a highly sensitive sensor, the selection of a proper electrochemical technique is of great importance. Therefore, differential pulse voltammetry (DPV) as a sensitive method was selected and used for further investigation. In addition, the electroactive surface area (A) of the β-CDMCP sensor compared with non-modified CPE was determined by CV in 5.0 mM of [Fe(CN)6]4−/3− solution containing 0.1 M KCl at different scan rates (v) according to the Randles–Sevcik equation.62 The obtained value of electroactive surface area was found to be 17.23 mm2 for the β-CDMCP sensor compared to 9.85 mm2 for CPE. Therefore, the electroactive surface area of the modified working electrode was increased by 74.9% with respect to the non-modified electrode, which provided effective evidence for a larger number of available electrochemical active sites on the surface of β-CDMCP sensor as expected.
|
| Fig. 5 Cyclic voltammograms of TC (900.0 ng L−1) (A) and DC (900.0 ng L−1) (B) in the presence of 0.05 mol L−1 phosphate buffer, pH 7.0 on β-CDMCP sensor at scan rate 500 mV s−1. | |
3.4. DPV of TC and DC at β-CDMCP sensor
The DP anodic and/or cathodic voltammetric behavior of TC/DC on the β-CDMCP sensor was investigated in 0.05 mol L−1 phosphate buffer as supporting electrolyte, pH 7.0. The influence of electrochemical parameters known to affect the differential pulse voltammograms, viz. pulse amplitude, pulse width, and scan rate were studied. During the study, each variable was changed, while the others were kept constant. The variables of interest were studied over the ranges 25–100 mV, 30–100 ms, and 10–100 mV s−1 for pulse amplitude, pulse width, and scan rate, respectively. The best sensitivity and well-shaped waves with relatively narrow peak widths were obtained when the values of 50 mV, 35 ms, and 50 mV s−1 were chosen for pulse amplitude, pulse width, and scan rate, respectively. Under the same electrochemical parameters, anodic differential pulse voltammetry (ADPV) at +320.0 mV was selectively used to detect TC in real samples. However, cathodic differential pulse voltammetry (CDPV) at −800.0 mV was used to detect DC in real samples.
3.5. The validation of TC and DC by β-CD-MGONPs coupled with β-CDMCP sensor
Under the optimal conditions, validation of the method was performed according to the recommendations of ICH guidelines.63 Dependence of the anodic peak current of TC and the cathodic peak current of DC on their concentration was investigated by DPV in milk samples. Peak currents were linearly related to the TC/DC concentrations in the range of 0.5 to 90.0 ng L−1 as shown in Table 3. For a proper practical application in milk samples, the concentrations of TC and DC were determined by using the standard addition method in this proper linear range. If any food sample has high concentrations of TC or DC, as is expected in rare cases, the appropriate dilution of sample in electrolyte is the proper solution to achieve precise measurements. To evaluate the repeatability of the modified electrode, all measurements were performed on the same electrode surface. However, in the case of electrode reproducibility, after each measurement the modified electrode was refreshed as described in the experimental section and reused again. Reproducibility was checked by 9 successive measurements within three consecutive days using 10.0 ng L−1 of TC/DC under optimized conditions; the relative standard deviation (RSD) of less than 2% was obtained. The limit of detection (LOD) (signal to noise ratio = 3) of 0.18 ng L−1 for TC/DC was easily achieved using the foregoing optimal conditions. The detection limit was also calculated theoretically by using the equation dl = 3SD/a27,28 where SD is the standard deviation of the blank solution (n = 10) and “a” is the slope of the calibration curve; it was found to be 0.15 ng L−1. The limit of quantification (LOQ) considered as the lowest concentration of each single compound providing a signal-to-noise ratio of approximately 10 and found to be 0.56 ng L−1 (RSD = 1.54%). The achieved linearity limits, LODs, and LOQs in the current work were lower than the reported MRL by FDA as mentioned in Section 1. Furthermore, the stability of the proposed modified electrode was checked by analyzing the limit of quantification of TC/DC during 48 h and comparing the collected results with data collected from freshly prepared standard solutions. The relative standard deviations were calculated to be ±1.2%. These results indicate that the modification can lead to a stable and reusable modified surface. The accuracy of the proposed method was also confirmed by using standard reference material of TC/DC and applying the optimized analytical approaches with three spiking replicates at three concentration levels covering the linearity range (1.00, 40.0, and 80.0 ng L−1). The spiked milk solutions were extracted on β-CD-MGONPs, analyzed, and the obtained mean recoveries of data collected by replicating the procedure within three consecutive days; results ranged from 87.45 to 108.31%.
Table 3 Linearity, limit of detection (LOD) and limit of quantification (LOQ) of the proposed method
Parameters |
TC |
DC |
Linearity (working range) (ng L−1) |
0.50–90.0 |
Correlation coefficient (r2) |
0.997 |
0.995 |
Intercept |
0.021 ± 0.008 |
0.043 ± 0.006 |
Slope |
0.34 ± 0.03 |
0.36 ± 0.02 |
Limit of detection LOD (ng L−1) |
0.18 |
Limit of quantification LOQ (ng L−1) |
0.55 |
The electrooxidation of TC in the presence of a member of quinolone group antibiotics (ciprofloxacin) under similar conditions was studied at/on the surface of modified electrode. The dependence of the oxidation current of ciprofloxacin in different concentrations up to 200 fold at a fixed concentration of TC (40.0 ng L−1) was investigated using DPV. As well, the electroreduction of DC (40.0 ng L−1) in the presence of TC (500.0 ng L−1) was also studied at/on the surface of the modified electrode. The results indicated that there is no influence on the peaks of the investigated compound. Furthermore, the effect of gelatin and some surfactants (e.g. CTAB, SDS, and Triton X-100), which can affect the adsorptive response through competitive coverage, was investigated. These surface-active compounds had no influence until their concentrations exceeded 50.0 μg L−1. Moreover, the proposed method was compared with other recently developed techniques for the extraction and detection of TC/DC in milk samples, as presented in Table 4. The results of this comparison indicated that the sensitivity and selectivity of the proposed method is satisfactory for evaluating TC/DC in commercial milk samples.
Table 4 Different techniques used for the extraction and detection of TC and DC residues in milk samples in literature compared to the proposed method. TC on selected adsorbent
Mode of extraction/technique |
Linear range |
LOD |
Recoveries (%) |
Reference |
TC |
DC |
TC |
DC |
TC |
DC |
Dispersive-SPME/HPLC with diode array detection |
20–400 ng g−1 |
10.2 ng g−1 |
35.3 ng g−1 |
87.8–91.4 |
87.8–91.4 |
64 |
Fluorescence detection (AgNPs–Eu3+ probe) |
0.04–8 μM |
— |
5.5 μM |
— |
94.8–119.0 |
— |
65 |
Fluorescence detection (CDs@MIPs probe) |
0.02–14 μM |
— |
20 μM |
— |
97.3–105.3 |
— |
66 |
MSPE based on phenyl silica adsorbent/capillary electrophoresis |
27–250 μg L−1 |
25–250 μg L−1 |
9 μg L−1 |
5 μg L−1 |
94.2–99.8 |
15 |
Ultrasound-assisted dispersive extraction/HPLC with diode array detection |
50–150 μg kg−1 |
9.3 μg kg−1 |
18.7 μg kg−1 |
90.20–106 |
91.7–107 |
67 |
MSPE based on β-CD-MGONPs with β-CD based sensor/differential pulses voltammetry |
0.50–90.0 ng L−1 |
0.18 ng L−1 |
55 ng L−1 |
92.1–105.0 |
Present work |
The determination of TC/DC (spiked concentration) in four milk samples (natural, skimmed, semi-skimmed, and full fat) was performed by three approaches: first, the solid phase extraction by β-CD-MGONPs followed by UV-Vis spectrophotometry measurements; second, by direct DPV measurements of TC/DC without prior extraction; and third, the combination of β-CD-MGONPs with the β-CD-MCP sensor for the consecutive solid phase extraction and DPV measurement of TC/DC. Accordingly, the UV-Vis spectrophotometry measurements and β-CD-MCP sensor, when used alone, were not competent to measure ultra-trace TC/DC concentrations (1.5 ng L−1) in milk samples. However, when β-CD-MGONPs was coupled with the β-CD-MCP sensor, the sensing system was sensitive to respond TC or DC concentration (1.0 ng L−1) in all types of milk samples studied (see Fig. 6). Interestingly, when the dual preconcentration of TC or DC was done first on the β-CD-MGONPs and second on the β-CD-MCP sensor, the current response was enhanced more than with those obtained with the sensor only.
|
| Fig. 6 Representative differential pulse voltammograms of TC (anodic polarization, 1.0 ng L−1) (A) and DC (cathodic polarization, 1.0 ng L−1) (B) in natural milk sample. Other experimental conditions as sited in the text. | |
In order to evaluate the applicability of the proposed method, various amounts of TC/DC were added to the samples, and the results showed that TC/DC in milk samples was assayed with satisfactory recoveries of 92.1–105.5% and RSD% of 0.74–3.86 as listed in Table 5. The peak of an analyte was identified by using the standard addition method. In all cases, the application of the t-test for the slopes of the calibration curves showed no significant statistical differences. Consequently, there is no evidence of systematic error affecting the determination of TC/DC in milk by the proposed method. From these results, it can be proved that the combined β-CD-MGONPs with β-CD-MCP sensor is useful for the determination of TC or DC in milk samples.
Table 5 Recoveries of TC and DC from different milk samples by β-CD-MGONPs with β-CD-MCP sensor
Samples |
Added (ng L−1) |
Found (ng L−1) |
Recoveries (%) |
RSD% (n = 3) |
TC |
DC |
TC |
DC |
TC |
DC |
TC |
DC |
Natural milk |
1.00 |
0.93 |
0.95 |
93.00 |
94.57 |
3.54 |
2.68 |
40.00 |
37.12 |
37.14 |
92.80 |
92.93 |
2.57 |
2.43 |
80.00 |
84.40 |
84.45 |
105.50 |
104.88 |
1.87 |
1.91 |
Skimmed milk |
1.00 |
0.99 |
0.99 |
99.00 |
99.36 |
1.87 |
1.47 |
40.00 |
39.32 |
39.35 |
98.30 |
98.41 |
0.99 |
1.03 |
80.0 |
79.83 |
79.84 |
99.79 |
99.85 |
0.74 |
0.71 |
Semi-skimmed milk |
1.00 |
0.97 |
0.98 |
97.00 |
97.60 |
2.17 |
2.20 |
40.00 |
40.15 |
40.20 |
100.38 |
100.46 |
1.17 |
1.33 |
80.00 |
81.15 |
80.04 |
101.44 |
101.62 |
0.93 |
0.95 |
Full-fat milk |
1.00 |
0.93 |
0.96 |
93.00 |
93.37 |
3.86 |
3.57 |
40.00 |
36.84 |
36.83 |
92.10 |
92.61 |
2.86 |
2.71 |
80.0 |
83.99 |
82.68 |
104.99 |
105.20 |
2.12 |
2.00 |
4. Conclusions
The present study reports a novel method for determining TC/DC in milk samples with combination of β-CD-MGONPs and the β-CD-MCP sensor. Accordingly, β-CD has the main role for the dual preconcentration during the extraction on β-CD-MGONPs and voltammetric detection with the β-CD-MCP sensor. Moreover, β-cyclodextrin appeared to have a great ability to easily differentiate between TC and DC in complex matrices due to its selective inclusion interaction inside its cavity followed by electron transfer onto the surface of a working electrode. The strategy could be based on the selective extraction on β-CD-MGONPs via batch adsorption followed by the sensitive DPV quantification via adsorption on a β-CD-MCP sensor. Therefore, the proposed platform in the current study showed the advantages of simplicity, rapidity, reliability, and low cost compared to other hitherto published reports for the determination of TC/DC in milk samples.
References
- C. R. Anderson, H. S. Rupp and W.-H. Wu, J. Chromatogr. A, 2005, 1075, 23 CrossRef CAS PubMed.
- European Commission Regulation no. 2377/90.
- M. Kuhne, D. Ihnen, G. Moller and O. Agthe, J. Vet. Med., A, 2000, 47, 379 CrossRef CAS PubMed.
- P. Kowalski, J. Pharm. Biomed. Anal., 2008, 47, 487 CrossRef CAS PubMed.
- L. Nozal, L. Arce, B. M. Simonet, A. Ríos and M. Valcárcel, Anal. Chim. Acta, 2004, 517, 89 CrossRef CAS.
- J. Kurittu, S. Lönnberg, M. Virta and M. Karp, J. Agric. Food Chem., 2000, 48, 3372 CrossRef CAS PubMed.
- U. Saha, A. K. Sen and T. K. Das, Talanta, 1990, 37, 1193 CrossRef CAS PubMed.
- A. Townshend, W. Ruengsitagoon, C. Thongpoon and S. Liawruangrath, Anal. Chim. Acta, 2005, 541, 105 CrossRef CAS.
- C. Lau, J. Lu and M. Kai, Anal. Chim. Acta, 2004, 503, 235 CrossRef CAS.
- A. Pena, L. P. Palilis, C. M. Lino, M. I. Silveira and A. C. Calokerinos, Anal. Chim. Acta, 2000, 405, 51 CrossRef CAS.
- H. Oka, Y. Ito and H. Matsumoto, J. Chromatogr. A, 2000, 882, 109 CrossRef CAS PubMed.
- D. J. Fletouris, J. E. Psomas and N. A. Botsoglou, J. Agric. Food Chem., 1990, 38, 1913 CrossRef CAS.
- W. H. Farrington, J. Tarbin, J. Bygrave and G. Shearer, Food Addit. Contam., 1991, 8, 55 CrossRef CAS PubMed.
- S. Sczesny, H. Nau and G. Hamscher, J. Agric. Food Chem., 2003, 51, 697 CrossRef CAS PubMed.
- I. S. Ibarra, J. A. Rodriguez, J. M. Miranda, M. Vega and E. Barrado, J. Chromatogr. A, 2011, 1218, 2196 CrossRef CAS PubMed.
- V. K. Gupta, M. L. Yola, N. Atar, Z. R. Üstündağ and A. O. Solak, J. Mol. Liq., 2014, 191, 172 CrossRef.
- N. Atar, T. Eren and M. L. Yola, Thin Solid Films, 2015, 590, 156 CrossRef CAS.
- N. Atar, T. Eren, B. Demirdögen, M. L. Yola and M. O. Çağlayan, Ionics, 2015, 21, 2285 CrossRef CAS.
- N. Atar, T. Eren, M. L. Yola, H. Karimi-Maleh and B. Demirdögen, RSC Adv., 2015, 5, 26402 RSC.
- N. Atar, T. Eren, M. L. Yola, H. Gerengi and S. Wang, Ionics, 2015, 21, 3185 CrossRef CAS.
- A. A. Yakout and H. M. Albishri, J. Taiwan Inst. Chem. Eng., 2015, 55, 180 CrossRef CAS.
- M. Safarikova and I. Safarik, J. Magn. Magn. Mater., 1999, 194, 108 CrossRef CAS.
- V. K. Gupta, N. Atar, M. L. Yola, Z. Üstündağ and L. Uzun, Water Res., 2014, 48, 210 CrossRef CAS PubMed.
- M. L. Yola, N. Atar, T. Eren, H. Karimi-Maleh and S. Wang, RSC Adv., 2015, 5, 65953 RSC.
- G. Kotan, F. Kardaş, Ö. A. Yokuş, O. Akyıldırım, H. Saral, T. Eren, M. L. Yola and N. Atar, Anal. Methods, 2016, 8, 401 RSC.
- L. Fan, C. Luo, M. Sun, H. Qiu and X. Li, Colloids Surf., B, 2013, 103, 601 CrossRef CAS PubMed.
- Y. J. Guo, S. J. Guo, Y. M. Zhai, S. J. Dong and E. K. Wang, ACS Nano, 2010, 4, 4001 CrossRef CAS PubMed.
- D. Abd El-Hady, Anal. Chim. Acta, 2007, 593, 178 CrossRef CAS PubMed.
- D. Abd El-Hady and N. Abo El-Maali, Microchim. Acta, 2008, 161, 225 CrossRef.
- D. Abd El-Hady and A. K. Youssef, Anal. Chim. Acta, 2013, 772, 68 CrossRef CAS PubMed.
- M. L. Yola, T. Eren and N. Atar, Electrochim. Acta, 2014, 125, 38 CrossRef CAS.
- M. L. Yola, V. K. Gupta and N. Atar, Mater. Sci. Eng., C, 2016, 61, 368 CrossRef CAS PubMed.
- N. Atar, M. L. Yola and T. Eren, Appl. Surf. Sci., 2016, 62, 315 CrossRef.
- M. L. Yola, V. K. Gupta, T. Eren, A. E. Şen and N. Atar, Electrochim. Acta, 2014, 120, 204 CrossRef CAS.
- M. L. Yola, N. Atar, Z. Üstündağ and A. O. Solak, J. Electroanal. Chem., 2013, 698, 9 CrossRef CAS.
- V. K. Gupta, N. Atar, M. L. Yola, M. Eryılmaz, H. Torul, U. Tamer, İ.-H. Boyacı and Z. Üstündağ, J. Colloid Interface Sci., 2013, 406, 231 CrossRef CAS PubMed.
- V. K. Gupta, M. L. Yola, M. S. Qureshi, A. O. Solak, N. Atar and Z. Üstündağ, Sens. Actuators, B, 2013, 188, 1201 CrossRef CAS.
- M. L. Yola, T. Eren and N. Atar, Sens. Actuators, B, 2015, 210, 149 CrossRef CAS.
- O. Akyildirim, H. Medetalibeyoglu, S. Manap, M. Beytur, F. S. Tokali, M. L. Yola and N. Atar, Int. J. Electrochem. Sci., 2015, 10, 7743 CAS.
- W. S. Hummers Jr and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339 CrossRef.
- R. Liang, C. Liu, X. Meng, J. Wang and J. Qiu, J. Chromatogr. A, 2012, 1266, 95 CrossRef CAS PubMed.
- H. A. Zadeh and Z. Talleb, Talanta, 2015, 134, 387 CrossRef PubMed.
- K. Aguilar-Arteaga, J. A. Rodriguez and E. Barrado, Anal. Chim. Acta, 2010, 674, 157 CrossRef CAS PubMed.
- V. Chandra, J. Park, Y. Chun, J. W. Lee, I. C. Hwang and K. S. Kim, ACS Nano, 2010, 4, 3979 CrossRef CAS PubMed.
- P. Bradder, S. K. Ling, S. Wang and S. Liu, J. Chem. Eng. Data, 2011, 56, 138 CrossRef CAS.
- J. Hu, D. Shao, C. L. Chen, G. D. Sheng, J. X. Li, X. K. Wang and M. Nagatsu, J. Phys. Chem. B, 2010, 114, 6779 CrossRef CAS PubMed.
- Q. Xiao, X. Tan, L. Ji and J. Xue, Synth. Met., 2007, 157, 784 CrossRef CAS.
- L. X. Song, H. M. Wang, X. Q. Guo and L. Bai, J. Org. Chem., 2008, 73, 8305 CrossRef CAS PubMed.
- E. Delahaye, V. Escax, N. E. Hassan, A. Davidson, R. Aquino, V. Dupuis, R. Pelzynski and Y. L. Raikher, J. Phys. Chem. B, 2006, 110, 26001 CrossRef CAS PubMed.
- Z.-L. Wang, D. Xu, Y. Huang, Z. Wu, L.-M. Wanga and X.-B. Zhang, Chem. Commun., 2012, 48, 976 RSC.
- D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff, Chem. Soc. Rev., 2010, 39, 228 RSC.
- R. A. Figueroa, A. Leonard and A. A. Mackay, Environ. Sci. Technol., 2004, 38, 476 CrossRef CAS PubMed.
- Q. Liu, Y. Zheng, L. Zhong and X. Cheng, J. Environ. Sci., 2015, 28, 29 CrossRef PubMed.
- Y. S. Ho, Scientometrics, 2004, 59, 171 CrossRef CAS.
- Y. S. Ho and G. McKay, Process Biochem., 1999, 34, 451 CrossRef CAS.
- S. H. Chien and W. R. Clayton, Soil Sci. Soc. Am. J., 1980, 44, 265 CrossRef CAS.
- S. K. Srivastava, R. Tyagi and N. Pant, Water Res., 1989, 23, 1161 CrossRef CAS.
- Y. Gao, Y. Li, L. Zhang, H. Huang, J. Hua, S. M. Shah and X. Su, J. Colloid Interface Sci., 2012, 368, 540 CrossRef CAS PubMed.
- L. Ji, W. Chen, L. Duan and D. Zhu, Environ. Sci. Technol., 2009, 43, 2322 CrossRef CAS PubMed.
- M. E. Parolo, M. C. Savini, J. M. Valles, M. T. Baschini and M. J. Avena, Appl. Clay Sci., 2008, 40, 179 CrossRef CAS.
- C. Gu, K. G. Karthikeyan, S. D. Sibley and J. A. Pedersen, Chemosphere, 2007, 66, 1494 CrossRef CAS PubMed.
- A. J. Bard and L. R. Faulkner, Electrochemical methods: fundamentals and applications, Wiley, New York, 2nd edn, 1980 Search PubMed.
- Commission Decision 2002/657/EC of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results, OJ 2002, L 221, p. 8.
- W.-H. Tsai, T.-C. Huang, J.-J. Huang, Y.-H. Hsue and H.-Y. Chuang, J. Chromatogr. A, 2009, 1216, 2263 CrossRef CAS PubMed.
- H. L. Tan and Y. Chen, Sens. Actuators, B, 2012, 173, 262 CrossRef CAS.
- J. Hou, H. Li, L. Wang, P. Zhang, T. Zhou, H. Ding and L. Ding, Talanta, 2016, 146, 34 CrossRef CAS PubMed.
- E. Karageorgou, M. Armeni, I. Moschou and V. Samanidou, Food Chem., 2014, 150, 328 CrossRef CAS PubMed.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra03787a |
|
This journal is © The Royal Society of Chemistry 2016 |