Sensitive and selective determination of aqueous triclosan based on gold nanoparticles on polyoxometalate/reduced graphene oxide nanohybrid

The widespread use of triclosan (TCS) in household cleaning products, medical devices and personal care poses a potential risk to the ecological system and human health due to its release into sediments, surface water and ground water resources and chronicle toxicity to aquatic organisms. A novel molecularimprinted electrochemical sensor based on gold nanoparticles decorating polyoxometalate (H3PW12O40)/reduced graphene oxide was developed for determination of trace TCS in wastewater. Reduced graphene oxide (rGO) was functionalized by polyoxometalate (POM) through electrostatic interaction between the POM and rGO nanosheets to produce a photocatalyst (POM/rGO) in aqueous solution. Gold nanoparticles (AuNPs) were further deposited on the POM/rGO without using any reducing agent and the prepared nanomaterial (AuNPs/POM/rGO) was employed to modify a glass carbon (GC) electrode (AuNPs/POM/rGO/GC) under infrared light. Several techniques, X-ray photoelectron spectroscopy (XPS), reflection-absorption infrared spectroscopy (RAIRS), scanning electron microscopy (SEM), and transmission electron microscopy (TEM), were used for electrode characterization. TCS imprinted film was generated on AuNPs/POM/rGO/GC via polymerization of phenol and TCS and characterized by cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS). The sensor was found to have a linear detection range and a limit of TCS at 0.5–50.0 nM and 0.15 nM, respectively. The molecular imprinted sensor was applied to wastewater and lakewater samples and demonstrated effective performance as compared to other complicated methods.


Introduction
TCS is a typical chemical that has been used in pharmaceuticals and personal products 1 such as surgical suture materials or hand soaps, deodorants, toothpastes, antiseptic-creams, plastics, foodstuffs and functional clothing for over 40 years. 2 Its widespread use has led to the release of TCS into wastewater, sediments and many water sources. 3 TCS is chronically toxic to aquatic organisms and its presence in wastewaters may affect the ecosystem and human health. 2 Several analytical methods have been reported to detect TCS using gas chromatography-tandem mass spectrometry, liquid chromatography-mass spectrometry (LC-MS), liquid chromatography/electrospray ionization tandem mass spectrometry and voltammetrics. 4-10 But these techniques have some disadvantages such as expensive apparatus and complicated operation. 11,12 A rapid and sensitive method to detect triclosan is thus important to ensure human and environment safety. In recent years, various nanosensors have been reported for selective, sensitive and rapid determination of toxic compounds, biomolecules and drugs. [12][13][14][15] In the past few years, graphene has become an intensive interest of scientists all over the world due to its stability and high surface area. 16,17 Graphene has honeycomb-like structure via sp 2 hybridization in one-atom-thickness. 18 Currently, graphene oxide (GO) is widely produced by chemical oxidation of graphite and used as a precursor to graphene. GO can be reduced by thermal treatment or chemical reduction to form rGO, 19 and rGO has been used for fuel cells, drug detection and sensors. [20][21][22] Many papers have also reported the chemical, 23 optical, 24 adsorption 25 and electronic properties 26 of various nanostructured metals. AuNPs are utilized as electrode surface for sensors 11,27 to increase the surface area and rate of electron transfer. In addition, polyoxometalates (POMs) are polyatomic anionic ion clusters composing of d-block transitional metal-oxides, and they have multiple redox behavior and photo-electrochemical properties. 28,29 POMs are a class of photoactive materials used in homogeneous reactions or heterogeneous processes. In reduced forms, their electron and proton transfer and/or storage abilities make them act as efficient donors or acceptors of several electrons without structural change. POMs also have been shown to serve as reducing and capping agents for metal nanostructures. 17 The molecular imprinting technique is widely used for molecular recognition 30 via the polymerization of target molecules, forming specic cavities. 24,27 From those above materials and molecular imprinting technique, various sensors can be fabricated.
There is no report about determination of TCS by using a molecular imprinting method based on the nanomaterials including rGO and AuNPs. Jiang et al. prepared AuNPs on GO surface by using polyethylenimine as a reducing reagent, followed by L-cysteine immobilization through an Au-S bond. Aer the preparation, the nanocomposite was applied as a novel ZIC-HILIC material to achieve highly selective enrichment of glycopeptides from biological samples. 31 In this study, AuNPs were synthesized under the UV light on rGO surface with POM as reducing and stabilizing reagent. We then prepared a TCS imprinted electrochemical sensor based on AuNPs deposition on POM functionalized rGO. The developed imprinted electrochemical sensor shows high sensitivity and selectivity in wastewater measurement.

Instrumentation
Differential pulse voltammetry (DPV) and CV were carried out on an electrochemical station (IviumStat, US) equipped with a C3 cell stand. Electrochemical impedance spectroscopy (EIS) data were acquired at 10 mV wave amplitude from 0.1 to 100 kHz and at an electrode potential of 0.195 V. The infrared spectra were obtained from a Bruker Tensor 27 FT-IR. XPS analysis was performed on a PHI 5000 Versa Probe X-ray photoelectron spectrometer (F ULVAC-PHI, Inc., Japan/USA). TEM images were obtained on a JEOL 2100 HRTEM instrument (JEOL Ltd, Tokyo, Japan) and SEM images were obtained on a ZEISS EVO 50 analytic microscope (Germany).

Cleaning of glass carbon (GC) electrodes
All GC electrodes were rst polished by 0.1 and 0.05 mm alumina successively and then the electrodes were sonicated in pure water and IPA + MeCN solution (50 : 50 by v/v) to remove unreacted materials from the surface. The reference electrode was a Ag/AgCl/KCl (sat) and the counter electrode was a Pt wire.

Preparation of rGO
GO was prepared according to the protocol in our previous papers. 14 The as-prepared GO was dispersed into water (200 mL) with addition of hydrazine hydrate (4 mL, 80 wt%) and was heated at 100 C for 24 h in an oil bath. The rGO was collected by vacuum ltration. Finally, 20 mL of AuNPs/POM/rGO (0.5 mg mL À1 ) was dropped on the GC electrode and then the modied electrode (AuNPs/ POM/rGO/GC) was dried under an infrared heat lamp.

Preparation of imprinted TCS sensors
The preparation of TCS imprinted sensors is illustrated in Scheme 1. Firstly, TCS molecular imprinted polymer (MIP) lm on AuNPs/POM/rGO/GC electrode (MIP/AuNPs/POM/rGO/GC) was prepared by CV for 20 cycles using 80 mM phenol as a monomer in a phosphate buffer solution (pH 7.0) containing 20 mM TCS at a scan rate of 100 mV s À1 between 0.0 V and +1.0 V. Aer electropolymerization, the electrode was dried at room temperature. For comparison, MIP/GC, MIP/rGO/GC and MIP/ POM/rGO/GC electrodes were also prepared with same way. A non-polymer imprinted electrode (NIP) was prepared without using TCS for a control experiment like the preparation of MIP. To break up the electrostatic interactions between phenol monomer and polar groups of the TCS, we used 1.0 M NaCl as desorption agent in a batch system. A TCS imprinted electrode was dipped into 25 mL of the 1.0 M NaCl aqueous solution and was swung in a bath (200 rpm) at room temperature for 20 min. Aer that, the electrode was washed with ultra pure water and dried in nitrogen gas under vacuum (200 mmHg, 25 C). The MIP electrodes were stored in a closed box without uctuations of temperature and pressure. In addition, the voltammograms were obtained in an insulation cabinet for avoiding temperature and pressure uctuation to affect the sensor response.

Preparation of wastewater samples
Wastewater samples were collected from an industrial wastewater pool in Izmir, Turkey, using pre-cleaned amber glass bottles. Lakewater samples were collected from Van Lake in Turkey. The sample bottles were lled without headspace and immediately placed in coolers lled with icepacks and transferred to the laboratory for storage at 4 C and analysis within one week. Before analysis, the collected wastewater and lakewater samples were centrifuged again at 4500 rpm for 5 min and ltrated by a 0.45 mm syringe lter. The ltrates were then diluted with 0.1 M phosphate buffer solution (pH 7.0) for analysis.

Characterization of electrode surface
TEM image of AuNPs/POM/rGO shows that the particle sizes of AuNPs are very similar at the mean diameter of 8-9 nm (Fig. 1A). The AuNPs are presented in dark dots on a lighter-shaded substrate of planar POM/rGO sheets. The creased nature of rGO is highly benecial in providing a high surface area on GC electrodes. In addition, C, Au, O, W and P peaks have been observed in EDX analysis (Fig. 1B), conrming the formation of AuNPs/POM/rGO nanohybrid. The IR spectra of the AuNPs/ POM/rGO also show the formation of the nanohybrid (Fig. 1C). The bands around 3200 cm À1 and 1600 cm À1 suggested the oxygen-containing functional groups of rGO. The peaks around 1580 cm À1 can be attributed to the stretching vibrations of C]O groups of the rGO sheets. Fig. 1C conrms the POM attached on rGO planes. The bands around 1050 cm À1 and 1350 cm À1 are referred to metal-oxygen groups of POM/ rGO. The formation of POM/rGO may be explained with the electrostatic interaction between POM and rGO via strong adsorption. 28,32 The formation of AuNPs/POM/rGO was further examined by XPS. The peaks of C 1s , P 2p , Au 4f and W 4f conrmed the formation of AuNPs/POM/rGO nanohybrid (Fig. 1D). The Au 4f 7/2 peak at 82.5 eV conrms the presence of AuNPs and the signal at 87.2 eV can be attributed to free gold nanoparticles. 14 SEM characterization was performed to evaluate the morphologies of the electrode surfaces in step by step modication. Fig. 2A displays that GC electrode has smooth surface. Fig. 2B shows the layers of rGO indicating high surface area of modied GC surface while Fig. 2C presents the POM/rGO/GC electrode surface. For AuNPs/POM/rGO on GC electrode, an intensive layer was observed covering the surface (Fig. 2D). An electrodeposition layer by electro polymerization of phenol covered the MIP/AuNPs/POM/rGO/GC electrode. These images indicate that the imprinted electrochemical sensor is accomplished (Fig. 2E). Moreover, AuNPs/POM/rGO/GC was regular spheres while the surface was rough. Compared with POM/rGO/ GC, the existence of AuNPs could not only enhance the adsorption capacity but also conducive to the formation of MIP/ AuNPs/POM/rGO/GC, playing the role of a framework for the formation of MIP/AuNPs/POM/rGO/GC. Some granular substances were attached on the surface of MIP/AuNPs/POM/ rGO/GC, indicating the formation of MIPs through electrochemical polymerization.
Electro polymerization was performed by CV in a phosphate buffer solution (0.1 M, pH 7.0) and the voltammograms are presented in Fig. 3. It was clearly demonstrated that the currents decreased with number of the cycles. The oxidation of phenol was recorded as the irreversible peak at the potential of 0.65 V on the rst scan. During continuous scanning, the current of the reduction peak decreased and then disappeared. This showed MIP lm formation on the AuNPs/POM/rGO/GC electrode.

Characterization of electrode impedance
EIS of bare GC electrode displays a small semicircle at high frequencies. The value of charge transfer resistance (R ct ) of the bare GC electrode was calculated to be 100 ohm (curve a of Fig. 4A). When the rGO was coating on the bare GC electrode, the value of R ct was calculated as 70 ohm (curve b of Fig. 4A). This is clearly indicative that the rGO layer increases the electron transfer rate. When the POM was coating on rGO/GC electrode, the value of R ct was found to be 58 ohm (curve c of Fig. 4A). These performances were attributed to the large surface area and the synergistic effect of POM and rGO. The EIS of AuNPs/POM/rGO/GC electrode presents in a straight line, the characteristic of a diffusional limiting step (curve d of Fig. 4A). Hence, it is clear that AuNPs/POM/rGO nanocomposite In addition, aer the electrochemical polymerization of phenol monomer on AuNPs/POM/rGO/GC electrode, the MIP/

Characterization of voltammetrics of electrodes
DPV showed the responses of TCS at different electrodes ( Fig. 4C and D). The MIP/AuNPs/POM/rGO/GC electrode shows no background current signal in 0.1 M phosphate buffer (pH 7.0) (curve a of Fig. 4C). Aer rebinding of TCS (10.0 nM TCS), it shows a much higher peak at about 0.65 V (curve c of Fig. 4C). However, the NIP/AuNPs/POM/rGO/GC electrode shows a small current signal (curve b of Fig. 4C). This indicates that the non-specic interaction of TCS is weak and the response aer MIP is very strong. The performances of different MIP sensors were also compared by DPV (Fig. 4D). It is shown that the performance of MIP/AuNPs/POM/rGO/GC electrode (curve d of Fig. 4D) is better than that of MIP/POM/rGO/GC, MIP/rGO/GC and MIP/GC electrodes (curves c, b and a of Fig. 4D) due to more effective surface area.

Optimization of fabrication and analytical conditions
The effects of the concentration of AuNPs/POM/rGO on MIP/ AuNPs/POM/rGO/GC electrode were rst tested. Initially, with the increasing concentration of AuNPs/POM/rGO up to 0.5 mg mL À1 , the peak current of TCS increased and reached a maximum at 7 mA. However, aer the concentration exceeded  0.5 mg mL À1 , the peak current of TCS (10.0 nM) is decreased (Fig. 5A). Hence, 0.5 mg mL À1 of AuNPs/POM/rGO was selected as the optimum amount. The pH of the medium also produces a signicant inuence on the polymeric lm. 11,14 Fig. 5B demonstrates the DPV peak current in the pH range of 5.0-9.0. The maximum signal was appearing at pH 7.0. TCS molecules show different electrochemical oxidation behaviors to the polymeric lm at different pHs. The DPV response of TCS increased with solution pH up to 7.0 and decreased subsequently. Aer the solution pH exceeded 7.0, the decrease of the peak current may be owing to the dissociation of the phenolic moiety.
The inuence of TCS to phenol monomer molar ratio was also studied (Fig. 5C). The peak current of TCS achieved a maximum at the ratio of 1 : 4. This was linked with the available binding sites. At low amount of phenol monomer, the available binding sites were less. According to the results, the signal of TCS increased when the amount of monomer increased to 80.0 mM. The increase was resulted from increase of the number of binding site. However, at a high concentration of phenol monomer, the non-specic interactions of TCS-monomer could occur, reducing the specic response. Fig. 5D shows the variation of DPV responses at different elution time. The TCS peak current shows increasing with the elution time, reaching a maximum at 20 min, and then it remained stable aer 20 min, indicating that the elution of TCS was completed during 20 min. Thus, the optimal elution time at 20 min was taken. Fig. 5E shows the effect of temperature on DPV responses in the range of 5-30 C. As shown in Fig. 5E, the highest peak current occurred at 20 C. Aer that, it remained stable. Therefore, the experiment temperature was chosen as 20 C.

The linear detection range of TCS
The differential pulse voltammograms at varying TCS concentrations (Fig. 6A) show that the peak currents increased with increasing TCS concentration. For each point of the calibration graph, six independent measurements were obtained and the mean value was used. The linear regression equation of TCS (Fig. 6B) was obtained as y ¼ 0.471x + 1.593. From the equation, limits of TCS quantication (LOQ) and its detection (LOD) were found to be 5.0 Â 10 À10 M and 1.5 Â 10 À10 M, respectively. 27 Moreover, the recovery experiments in wastewater and lakewater samples were conducted using different TCS concentrations ( Table 1). The recovery rate of 98.9-100% shows excellent recovery of the developed TCS imprinted electrochemical sensor. For a comparison, LC-MS as a sensitive method was further performed 14 and no signicant difference between the LC-MS and DPV was found based on the Wilcoxon test (T calculated > T tabulated , p > 0.05) ( Table 2).      (Fig. 7C). Thus, we produced a highly selective sensor via creating binding sites that are specic to the target molecule. For reproducibility study, six different MIP/AuNPs/POM/ rGO/GC electrodes were prepared under the same condition and tested in TCS detection and analysis. Aer that, each MIP electrode was applied to wastewater samples for TCS analysis. According to the obtained results, the relative standard deviation (RSD) is 0.3% in 10.0 nM TCS.
The stability of MIP/AuNPs/POM/rGO/GC electrode was also checked. Aer 30 days, the signal was found to be approximate 98.8% of the original value which suggests its excellent longterm stability. Table 3 presents a comparison of the sensor performance in terms of linear range and LOD with other analytical methods. It is seen that the developed sensor showed a much lower limit of detection.

Conclusion
A new TCS imprinted electrochemical sensor based on AuNPs/ POM/rGO modied GC electrode was prepared and tested for determination of trace TCS in aqueous solution. The prepared sensor exhibits high selectivity and sensitivity in TCS detection with a detection limit of 0.15 nM. It demonstrates analytic capability comparable to other complicated methods but it offers simple and efficient application in target detection from wastewater and lakewater samples.