Determination of allura red using composites of water-dispersible reduced graphene oxide-loaded Au nanoparticles based on ionic liquid

ABSTRACT A facile route for producing reduced graphene oxide (RGO)-loaded Au nanoparticles based on ionic liquids (IL) has been proposed, in which the as-prepared RGO can be dispersed stably in water. With the assistance of IL, Au nanoparticles were uniformly and densely absorbed on the surfaces of the IL functionalised reduced graphene oxide (IRGO), forming a new composite of IRGO/Au with high dispersibility. This IRGO/Au composite enhanced its electrochemical signal obviously in the measurement of allura red in foods and exhibited a wider linear response ranging from 0.297 (0.0006 μmol L−1) to 99.3 μg L−1 (0.2 μmol L−1) with lower detection limit of 0.213 μg L−1 (0.00043 μmol L−1) at a signal-to-noise ratio of 3. To further study the practical applicability of the proposed sensor, the modified electrode was successfully applied to detect allura red in five kinds of common foods and the assay results were in a good agreement with the reference values detected by HPLC.


Introduction
Synthetic dyes including amaranth [1,2], ponceau 4R [3], sunset yellow [4], tartrazine [5], brilliant blue [6] and allura red [7,8] have been widely added into foodstuffs to make them appealing. Allura red, as a water-soluble synthetic dye, is employed to give fascinating red colour in many manufactured foods including drinks, snacks, jelly, haw flakes, marble chocolate, candy and dairy products. However, when consumed in excessive amounts of allura red, it may lead to asthma and allergies [9,10]. Given that its health risk due to its azo and aromatic ring groups, the detection of its content in foods is extremely important. In China, its maximum limit in soft drinks is 0.1 gL −1 (GB2760-2014) [8]. In fact, as far as we know, the allura red has been forbidden in many countries, including France, Belgium, Switzerland and Denmark [9]. Allura red can undergo degradation in beverage when exposed to irradiation [11,12]. So, we think the contents of allura red added in foods had better be controlled. And construct simple detection method is more significative. Compared with high performance liquid chromatography [13], capillary electrophoresis [14] and spectrophotometry method [15], building electrochemical sensor is a more good choice.
Hybridizing graphene with metal nanoparticles has been widely explored as electrode materials [16] because the integration of them not only can enhance their electronic and electrochemical properties but also can offer special features in the new hybrid [17][18][19]. Generally, there are two approaches to load metal nanoparticles on graphene including in situ synthesis approach [20,21] and electrostatic adsorption method [22]. However, the serious problem associated with the two approaches is the aggregation of graphene [23][24][25] and given that its economical accessibility and mass production, the reduced graphene oxide (RGO), obtained by reduction of graphene oxide (GO) [6,26,27], is also versatile material. Nevertheless, natural hydrophobicity and restacking of RGO [28,29] also extremely restricts its application in loading nanoparticles evenly. Especially, the toxic agents employed would lead to potential environmental hazards during the chemical reduction process [30][31][32].
To prohibit the irreversible aggregation tendency of RGO, in 2007, Li and co-workers first reported that water-dispersed RGO can be obtained from reduction of GO by hydrazine in the presence of ammonia via an electrostatic repulsion mechanism [33], and ionic liquid (IL)-functionalised RGO was also reported by Yang et al. in 2009 [34] and Shen et al. in 2011 [35]. Still, when it comes to practical production, the merits of these approaches may be discounted due to the lengthy process and potential environmental threat.
Herein, we present a green and facile route for the synthesis of water-dispersible RGO based on IL of 1-allyl-3-methylimidazolium chloride (abbreviated as IRGO) [36] which was used to load Au nanoparticles (Au) via the self-assembly of IRGO and Au through electrostatic adsorption method. Compared with IL (IL-NH 2 )-functionalised RGO [34], it is worth noting that the IRGO prepared here shows better dispersibility in water (up to 0.6 mg mL −1 ) and completely dispersed IRGO/Au composites were prepared by us. The integration of the Au nanoparticles and IRGO, and the good dispersibility of the composite make it good electrode material for electrochemical sensor which offers high sensitivity for the determination of allura red in foods.

Reagents
All reagents used in this work were of analytical grade and used as received without further purification. Allura red, gold acid chloride trihydrate and sodium borohydride were purchased from Aladdin (Shanghai, China). 1-Allyl-3-methylimidazolium chloride was purchased from Sinopharm Chemical Reagent Co. (Shanghai, China). Drinks of Bacardi Breezer Peach, RIO cocktail and Rainbow cocktail were purchased from a local market (Taiyuan, China). A series of 0.1 mol L −1 BR buffer solutions of different pH values were prepared according to a previous report [6]. Double distiled water was used throughout the investigation.

Instruments
All electrochemical experiments were performed on a CHI660d electrochemical workstation from CH Instruments Inc. (Shanghai, China), with a conventional three-electrode system including a IRGO/Au composite modified glass carbon electrode (IRGO/Au/GCE) as a working electrode, a Pt wire counter electrode and a saturated calomel electrode reference electrode. The transmission electron micrograph (TEM) measurements and UV-visible spectrum (UV-vis) were carried out on a JEM-2100 high-resolution transmission electron microscopy (JEOL, Japan) and Shimadzu UV-3600 UV-vis spectrophotometer (Shimadzu, Japan). HPLC experiment was carried out on a EX 1600 HPLC system containing an UV-vis detector (Shanghai, China) and a C18 column was used. The mobile phase CH 3 COONH 4 solution 1% (m/v), adjusted to pH 7.5 by addition of NaOH solution 10% (m/v), was delivered at flow rate of 1.5 mL min −1 . In addition, the detection was monitored at 507 nm and the injection volume was set at 20 μL according to previous reports [13].

Synthesis of IRGO/Au composites
GO was synthesised by a modified Hummers' method as reported elsewhere from graphite powder [37]. Among various methods proposed in the past, the chemical reduction of GO in solution aimed at the large-scale production of RGO. Unfortunately, this strategy has encountered the issues of aggregation and restacking of RGO, which is harmful to load metal nanoparticles on it uniformly ( Figure 1, Route 1).
Herein, a facile route for the synthesis of water-dispersible RGO, based on 1-allyl-3methylimidazolium chloride, was presented according to a previous report [38]. The key steps of the presented synthesis process are illustrated in Route 2 of Figure 1. First, the GO aqueous solution was added to IL of the 1-allyl-3-methylimidazolium chloride at 180°C under continuous stirring. The colour of the liquid mixture changed from yellowbrown to black, indicating the occurrence of GO reduction. The mixture was allowed to react for 2 h at 180°C (with constantly stirring). After the homogeneous mixture cooled down to room temperature, the solution was vacuum-filtered and washed by copious deionised water until there was no precipitate in the filtrate if titrated by silver nitrate to get IRGO. The stable IRGO dispersions were obtained through redispersing the IRGO in water.
1-mL IRGO at 0.1 mg mL −1 was added to 5 mL Au nanoparticles with the size of 15 nm which was prepared according to procedures reported before [39], followed by stirring at room temperature for 15 min. The resultant was centrifuged at 8000 rpm for 10 min and then thoroughly washed with water. IRGO/Au nanocomposite was obtained.

Preparation of the IRGO/Au-modified GCE (IRGO/Au/GCE)
The bare GCE was carefully polished by alumina slurry, washed successively by nitric acid (1:1, v/v), deionised water and acetone for 3 min. Then, 7.5 μL of IRGO/Au dispersion was dropped on the freshly prepared bare GCE surface and dried under the infrared lamp. For comparison, IRGO/GCE and Au/GCE were fabricated under the similar procedures.

Measurement procedures
Firstly, the allura red solution is prepared by diluting dye standard solution (10-3 mol L −1 ) with BR buffer solution (0.1 mol L −1 , pH 5.0). Then the IRGO/Au/GCE was immersed in the above solution. After 5 min accumulation, the square wave voltammetry (SWV) responses at frequency of 10 Hz were recorded. Before or after every measurement, the IRGO/Au/GCE was cleaned by double distiled water and the potential scan was repeated successively in a blank solution to regenerate the electrode surface.

Characterisation of IRGO/Au composite
The aqueous dispersion of IRGO/Au composite (curve b) was first characterised by UVvis spectroscopy and compared with GO aqueous dispersions (curve a) as shown in Figure 2(A). It can be seen from the curve b when the GO was reduced that the strong absorption peak of GO dispersion at 230 nm (curve a) corresponding to the π-π* transition of aromatic C-C bonds obviously red shifted to a higher wave-length of 270 nm and the shoulder peak at 300 nm of GO (curve a) also disappeared, suggesting that GO is reduced and the π-conjugation network is restored. For the Au nanoparticles prepared in this work with size of 15 nm, the typical surface plasmon band is around 517 nm [39]. So, we really saw another peak appeared at 517 nm from curve b. These results suggested that Au nanoparticles may load onto IRGO surface.
The as-obtained composites were further confirmed by TEM. As shown in Figure 2(B), the IRGO was almost transparent and displayed wrinkled shapes without obvious aggregation and there were many well-dispersed dark dots with diameter of 15 nm on the surface of the wrinkled IRGO, which also can be clearly seen from the TEM images at high magnification which was showed in Figure S1.Therefore, the TEM results verify excellent dispersion of Au nanoparticles on the IRGO surface evenly. This was attributed to the high dispersibility of the IRGO prepared and its photograph has been taken in the inset of Figure S2(A). There was no sediment observed in the vial even settling for 1 month at room temperature.
The UV-vis spectra of IRGO aqueous solutions with different concentrations were shown in Figure S2(A) to prove the high dispersibility of the IRGO. The plot of the absorbance at 270 nm versus the concentration is shown in Figure S2(B). A good linear concentration range was from 0.1 to 0.6 mg mL −1 with a correlation coefficient of 0.997. Obviously, the absorbance of the RGO obeys Beer's law. It was also found that the absorbance of the IRGO was not changed noticeably after 1 month. However, the absorbance at the nominal concentration above 0.7 mg mL −1 severely deviates from Beer's law, because of heterogeneous dispersion [38]. So the IRGO can be redispersed stably in water up to 0.6 mg mL −1 and the high dispersibility of the IRGO makes Au nanoparticles successfully loaded on it.

Electrochemical behaviour of allura red on modified GCEs
The electrochemical response of allura red at the IRGO/Au/GCE was first examined using cyclic voltammetry (CV), which is shown in Figure S3. A pair of redox peaks (O1/R2) is observed at first CV cycle. During the second potential sweeps, the oxidation peaks (O1) gradually decrease, and the reduction peaks (R2) gradually increase. From the comparison, it is clear that the oxidation signal is more sensitive than the reduction signal. So, considering high sensitivity and good simplicity, SWV single sweep was used to study the electrochemical behaviour of allura red.  enhancement effect to allura red. Based on the peak current and the active areas of electrodes, the peak current densities of bare GCE, Au/GCE, IRGO/GCE and IRGO/Au/GCE were calculated to be 0.05, 0.18, 1.07 and 2.88 μA mm −2 for allura red, respectively. Moreover, the peak potential of allura red at IRGO/Au/GCE shifted more negatively. These phenomena indicated that IRGO/Au/GCE could greatly facilitate the oxidation of allura red. It is also expected that the cooperation of IRGO and Au nanoparticles will increase the electro-catalytic active area and promote the electron transfer. The possible mechanism of the synergistic effect is ascribed as follows: First, the incorporation of IL with RGO can not only prevent the coagulation of RGO but also form a conductive interconnection network. Second, the Au nanoparticles enhanced the catalytic activity of IRGO/GCE by increasing the peak current due to the increasing electronic conductivity and effective surface area.

Optimization of the condition for IRGO/Au/GCE fabrication
The influence of the amount of IRGO/Au composite on the oxidation currents of allura red is displayed in Figure 4(A). The peak currents of 0.04 μmol L −1 allura red increased with the increasing amount of IRGO/Au composite from 0.0 to 7.5 μL, indicating that electron transfer process strongly depends on film thickness, afterwards the peak currents decreased as the amount further increased. Consequently, 7.5 μL of the IRGO/ Au composite was chosen for the subsequent experiments. Figure 4(B) shows the dependence of peak current on the accumulation time. The oxidation peak currents of allura red increased sharply within the first 3 min, and then continued with moderate slope up to 5 min. Finally, after 5 min, the oxidation peak current tended to a constant. Considering sensitivity and speed, an accumulation time of 5 min was chosen. The effect of pH value of BR buffer solution on the electrochemical behaviour of 0.8 μmol L −1 allura red was investigated over the pH range of 4.0-10.0 with CV method (Figure 4(C)). It was found that the peak current reached its maximum value at pH 5.0, and then decreased slightly when the pH exceeded 5.0. Therefore, BR buffer solution of pH 5.0 was used as the supporting electrolyte in all determination. Meanwhile, the oxidation peak potentials shifted negatively with the increase of pH value. In the inset of Figure 4(C), the relationship between oxidation potentials (E pa ) and pH could be expressed by the equation: E pa (V) = −0.028 pH + 0.92 (R 2 = 0.997), which indicated that proton involved in the oxidation reaction of allura red. Figure 4(D) shows effect of scan rates on the CV response of 0.6 μmol L −1 allura red. It can be seen from the inset of Figure 4(D) that the oxidation peak currents of allura red showed a linear relationship versus the scan rate in the range of 20-300 mV s −1 with the regression equation of I pa (μA) = 12.72ν +0.53 (R 2 = 0.999), indicating a typical adsorption-controlled process at the IRGO/Au/GCE. Figure 5(A) shows the relationship between the oxidation peak currents and the concentration of allura red in 0.1 mol L −1 pH 5.0 BR buffer solutions. It can be seen from Figure 5(B) that the SWV peak current of allura red increased with the increasing concentration of allura red. Under the optimum experimental conditions, the oxidation current of allura red was proportional to its concentration range from 0.297 (0.0006 μmol L −1 ) to 99.3 μg L −1 (0.2 μmol L −1 ). The linear regression equation was expressed as I pa (μA) = 0.266 + 193 C (μmol L −1 ) with a correlation coefficient of 0.998. The limit of detection is 0.213 μg L −1 (0.00043 μmol L −1 ) at the signal to noise ratio of 3:1. Comparisons of the proposed IRGO/Au/GCE performance with other electrochemical methods reported previously [7,8,[40][41][42] have been displayed in Table 1. It can be found that the proposed method here shows a relatively wide linear range and a lower limit of detection, which is sensitive enough to detect allura red in foods.

Reusability, reproducibility and stability of IRGO/Au/GCE
The regeneration of the IRGO/Au/GCE was carried out by CV from 0.3 to 1.0 V in pH 5 blank BR buffer solution for three times. It can be seen from Figure S4(A) that the asrenewed IRGO/Au/GCE for the measurement of 0.08 μmol L −1 allura red could restore 94.6% of the initial value after six assay runs, indicating high repeatability. Figure S4(B) gives the CV curves of 0.06 μmol L −1 allura red at six independently IRGO/Au/GCEs. The   reproducibility of the relative standard deviation was calculated to be 5.08%, indicating acceptable fabrication reproducibility of the electrode. The IRGO/Au/GCE for the measurement of 0.05 μmol L −1 allura red retained 80.0% of its original peak current value after 15 days of storage at 25 ± 2°C, which proved good stability.

Interference
The tolerance limit was defined as the maximum concentration of the interferences that caused an error less than ±5%. Under optimised experimental conditions, the effects of common interferent on the peak current of allura red at IRGO/Au/GCE were evaluated. Figure S(5) showed that the peak current of allura red has little change after the addition of different interferent. So, there is nearly no influence on the detection of 0.08 μmol L −1 allura red was found after the addition of 1000-fold concentrations of K + , Na + and Mg 2+ ( Figure S5(A)); 500-fold concentrations of phenol, glucose and glutathione ( Figure S5(B)); 100-fold concentrations Vitamin C and potassium sorbate ( Figure S5(C)); 20-fold concentrations of citric acid and sodium citrate ( Figure S5(D)) and 5-fold concentrations of sunset yellow and Amaranth ( Figure S5(E)).

Practical application
In order to verify the applicability of the proposed sensor in real samples, IRGO/Au/GCE was utilised for determination of allura red in five kinds of common foods including Milk chocolate, Lang strawberry jelly, Cranberry grain tea, Strawberry apple juice and Bacardi Breezer. For the first two, they should be pretreated following the steps below before detection: five grains of Milk chocolate and a Lang strawberry jelly were weighed, respectively, and then 25.0 mL ethanol was added. After 15 min ultrasoniccation, the mixture was filtrated with a 0.22-μm organic filter membrane and the liquid phase was collected in a 25.0-mL volumetric flask for detection. For the remaining three, they were used directly without any pretreatment.
To verify the practical application of this new developed method, 3 mL of each of food samples were diluted to 20.0 mL with BR (pH 5.0) and detected using the IRGO/Au/ GCE. The obtained results are shown in Table 2 which was also compared with the test results by HPLC. The assay results obtained by both methods showed a good agreement, indicating good accuracy of the proposed method.

Conclusion
A green route to prepare water-dispersible RGO-loaded Au nanoparticles based on IL has been developed. RGO shows good dispersibility in water due to the assistance of IL which resulted in uniformly absorbed of Au nanoparticles on the surfaces of the IRGO. The composite of IRGO/Au remarkably improved electrochemical activity towards allura red because of the synergistic effect of IRGO and Au nanoparticles. This method also applied in the detection of allura red in real sample and obtained satisfactory results.

Disclosure statement
No potential conflict of interest was reported by the authors.