A reduced graphene oxide-β-cyclodextrin nanocomposite-based electrode for electrochemical detection of curcumin

Curcumin is a polyphenolic compound with anti-oxidative and anti-cancer properties that is obtained from turmeric plants. Several studies have demonstrated that cancer cells are not killed unless they are exposed to 5–50 mM of curcumin. Consequently, it is vital to control the concentration of curcumin in cancer therapy. In this study, a sensitive electrochemical sensor was fabricated based on a beta-cyclodextrin–reduced graphene oxide (β-CD–rGO) nanocomposite for measuring curcumin concentration. The effects of experimental factors were investigated and the optimum parametric conditions were determined using the Taguchi optimization method. The β-CD–rGO modified electrode exhibited good electrochemical properties for curcumin detection. The results of differential pulse voltammetry experiments unveiled that the sensor shows a linear response to curcumin concentration over the range of 0.05–10 mM with a detection limit of 33 nM and sensitivity of 4.813 μA μM−1. The fabricated sensor exhibited selectivity in the presence of other electroactive species, e.g., propranolol, clomipramine and clonazepam.

CM suppresses cancer cell proliferation through the inhibition of inducible nuclear factor kappa B (NF-kB). The anti-inammatory effect of CM is a consequence of the reduction of NF-kB, cyclooxygenase 2 (COX2) and tumour necrosis factora (TNF-a). 6,7 In vitro studies have shown that the cancer cells are not killed unless they are exposed to 5-50 mM of CM. 8,9 There are several methods for CM detection, such as HPLC, [10][11][12] UV uorescence 13 and electrochemical methods. [14][15][16] Electrochemical sensors possess high sensitivity, portability, rapid measurement, simplicity and need a small quantity of sample. 17,18 Commercial applications conrm the attractive advantages of these biosensors. 19,20 In recent years, carbon nanomaterials and nanostructures have received signicant attention as electrochemical sensors and biosensors due to their biocompatibility, high surface area, good electrical properties and chemical stability. 18,21,22 Graphene-based nanomaterials with large surface area, excellent electron transportation and high thermal conductivity have shown great potentials for electrochemical biosensors. 23,24 In this regard, graphene-based electrochemical sensors have been used for detecting glucose, 25 dopamine, 26 hemoglobin 27 and heavy metal ions. 28 Based on the unique properties of graphene (high surface area and superconductivity) and b-CD (supramolecular recognition due to their capability of forming inclusion complexes with many hydrophobic guest molecules [29][30][31][32] ), the integration of graphene and b-CD can introduce a new nanocomposite which extends individual properties of both materials. CM, as a hydrophobic drug, can form inclusion complexes with b-CD molecules. 33,34 In the present study, beta-cyclodextrin-reduced graphene oxide (b-CD-rGO) nanocomposite was introduced for the rst time to build up a highly sensitive electrochemical sensor for quantifying CM. The nanocomposite was characterized using Fourier transform infrared spectroscopy, high-resolution transmission electron microscopy, electrochemical impedance spectroscopy, cyclic voltammetry and differential pulse voltammetry. Finally, the effects of various experimental parameters on the performance of the fabricated sensor were investigated through two different methods (e.g., Taguchi and individual common optimization).

Material and reagents
CM, b-CD and graphite powder were purchased from Sigma Aldrich, Germany. Hydrazine solution, NaOH, HCl, dimethyl sulfoxide (DMSO), H 2 O 2 , H 2 SO 4 , H 3 PO 4 , K 2 HPO 4 , KH 2 PO 4 and KMnO 4 were obtained from Merck, Germany. Glassy carbon (GC) electrode was provided by Azar Electrode, Tabriz, Iran. All chemicals were of analytical grade and used without further purication. Hydrogen chloride (HCl) and sodium hydroxide (NaOH) were used for pH adjustment. Double distilled water was used throughout the work. All electrochemical experiments were carried out at room temperature 25 AE 0.1 C.

Instrument and measurement methods
Fourier transform infrared (FTIR) spectroscopic measurements were carried out using a 6300 JASCO FTIR Spectrometer (Japan). All electrochemical measurements were carried out at room temperature using potentiostat/galvanostat Autolab PGSTAT (Eco Chemie, Utrecht, Netherlands; driven with NOVA soware). These measurements were carried out with a conventional three-electrode system consisting of modied/ unmodied GC electrode as a working electrode, a platinum wire as an auxiliary electrode, and an Ag/AgCl (3 M KCl) electrode as a reference electrode. Electrochemical Impedance Spectroscopy (EIS) was run in a 0.1 M KCl solution containing 5 mM K 3 [Fe(CN) 6 ]/K 4 [Fe(CN) 6 ] (1 : 1). EIS measurements were recorded under an oscillation potential of 5 mV over a frequency range of 10-0.1 Hz. Differential Pulse Voltammetry (DPV) was performed in 0.1 M phosphate buffer solution (PBS) (pH 7.4) with an amplitude of 25 mV and a pulse width of 0.05 s. To record DPV plots of CM, 1 mM CM was dissolved in DMSO, and the modied electrode was immersed in it for 30 minutes, then carefully washed with distilled water. Finally, DPV was performed at the potential range of 0.1-0.9 V, and current was plotted as a function of potential. High-resolution transmission electron microscopy (HRTEM) images were obtained using a JEM-2100F machine, Japan, operating at an accelerating voltage of 200 kV.

Synthesis of graphene oxide
Graphene oxide (GO) nanosheets were synthesized from graphite akes by a modied Hummer's method. 35 Typically, a mixture of H 2 SO 4 /H 3 PO 4 (360/40 mL) was added to a mixture of graphite powder (3 g) and KMnO 4 (1.8 g). The reaction mixture was warmed up to 50 C and stirred for 12 h. The solution was cooled down to room temperature, and then, 200 mL of ice and 2.6 mL hydrogen peroxide (30%) were added to the mixture. The mixture was centrifuged, and supernatant was decanted away. The remaining solid material aer multiple washing process (with ether, 30% HCl, ethanol and water) was vacuum-dried for 24 h to obtain GO powder.

Synthesis of b-CD-rGO nanocomposite
To fabricate b-CD-rGO nanocomposite, 10 mg GO and 20 mg b-CD were dispersed in deionized water (20 mL). Then, the solution was mixed with 300 mL ammonia solution and 20 mL hydrazine solution. Aer continuous stirring for a few minutes, the homogenous solution was placed in water bath (60 C) for 4 h; the stable black dispersion was obtained. The dispersion was ltered with a nylon membrane (0.22 mm) to obtain b-CD-rGO nanocomposite. 36,37

Preparation of modied electrodes
For pre-treatment of b-CD-rGO, 100 mg of b-CD-rGO powder was dispersed in 10 mL water and sonicated for 20 minutes to obtain a homogenous solution of b-CD-rGO. A bare GC electrode was polished with 0.3 mm and 0.05 mm alumina powder, and carefully washed with ethanol and distilled water. The modied glassy carbon electrode b-CD-rGO (b-CD-rGO/GC) was constructed by drop coating of 2 mL b-CD-rGO solution (2 mg mL À1 ) onto the surface of GC electrode and dried in air.

Experimental design
Taguchi method 38 is a powerful tool for designing experiments and analyzing the effect of control factors. We dened four important factors for the optimization of CM percentage, as follow: incubation time (minutes), b-CD/GO mass ratio, b-CD-rGO concentration (g mL À1 ), and electrolyte pH (Table SI-1, ESI †). The level of factors was selected by varying them in a range according to the experimental optimization conguration. The present design includes four control factors, three factors with three levels, and one with two levels, which total the experiments to 18.  [39][40][41] Fourier-transform infrared spectroscopy (FTIR) spectra of b-CD, GO and b-CD-rGO nanocomposite are shown in Fig. 1B. FTIR spectra of GO conrms the presence of oxygen containing groups: C-O vibration at 1051 cm À1 , C-O-C vibration at 1178 cm À1 , C-OH vibration at 1404 cm À1 , C]O in carboxylic groups at 1738 cm À1 , and OH stretching vibration at 3429 cm À1 . 36,42 The major bands for b-CD molecules are located at 707, 756, 857 and 943 cm À1 . These peaks reveal the presence of ring vibration (characteristic peaks) for b-CD. The major absorption peak is located at 3396 cm À1 that is assigned to OH stretching vibration. 36,43 When b-CD molecules were introduced to rGO surface, b-CD-rGO nanocomposite was formed. As shown in the FTIR spectra of b-CD-rGO nanocomposite, several characteristic peaks of b-CD molecules are observed which indicates that b-CD molecules attached to the surface of GO. It is also seen that the OH stretching vibration exhibited typical red-shi when hydrogen bonding was formed. 44 When b-CD molecules were assembled on the surface of GO, the OH stretching vibration (3427 cm À1 ) in b-CD-rGO exhibited red-shi relative to OH stretching vibration in b-CD (3396 cm À1 ). From the FTIR data, both GO and b-CD had multiple OH-groups, so, when the b-CD molecules were introduced to the GO solution, hydrogen bonding was formed; this nding is in good accordance with previous results. 36,45 HRTEM images determined the GO sheets with ake-like shape ( Fig. 2A), which is the evidence of successful GO sheets synthesis. 46,47 The HRTEM image of b-CD-rGO (Fig. 2B) indicates the successful linkage of b-CD molecules with the surface of rGO. It also displays a homogenous surface with uniform distribution of b-CD on the surface of rGO.

Surface analysis of modied electrode
The semicircle portion of Nyquist diagram addresses the electron charge resistance; the charge transfer resistance can be directly measured via the semicircle diameter. This method is a useful way to study the surface properties of modied electrode. Fig. 3 shows the Nyquist plots of GC electrode, GO modied GC electrode (GC/GO), b-CD modied GC electrode (GC/b-CD) and b-CD-rGO modied GC electrode (GC/b-CD-rGO) in the presence of 10 mM K 3 Fe(CN) 6 /K 4 Fe(CN) 6 and 0.3 M KCl. For the GO modied GC electrode and b-CD modied GC electrode, the semicircle portion of corresponding Nyquist diagrams dramatically increased in comparison with bare GC electrode as the semiconducting properties of GO and b-CD modied GC electrode increased. Also, the electrostatic repulsion between negative oxygen groups on the surface of GO and b-CD molecules with negatively charged electrochemical probe (K 3 Fe(CN) 6 /K 4 Fe(CN) 6 ) increased the charge transfer resistance. When the GC electrode was modied with the b-CD-rGO nanocomposite layer, the semicircle portion dramatically decreased compared to bare GC electrode; this phenomenon is attributed to good electronic properties of b-CD-rGO nanocomposite. Thus, b-CD-rGO nanocomposite can greatly increase the electron transfer kinetics and provide a suitable environment for electron transfer.

Electrochemical behavior of curcumin
CM has functional phenolic hydroxyl groups and methoxy groups which can be oxidized at the surface of electrode. The hydroxyl groups, present at the benzene rings, can easily  undergo an oxidation process. 48,49 In fact, the electrochemical oxidation of CM has two-steps. 50 Fig. 4A shows one cyclic voltammogram of GC/b-CD-rGO electrode in the absence of CM and two successive cycles in the presence of 1 mM CM, both in the phosphate buffer solution (pH 7.4). An irreversible oxidation peak I and a pair of reduction/oxidation peak (II/III) are observed in the presence of CM. In the rst cycle, CM exhibits an oxidation peak I and a reduction peak II with E paI ¼ 0.7 V and E pcII ¼ 0.21 V. In the second cycle, a new anodic peak III appears with E paIII ¼ 0.31 V while the oxidation peak I is disappeared. According to others' ndings, the oxidation peak I is an irreversible step and its active group comes from the product of irreversible reaction. 51,52 The electrochemical behaviors of GC (voltammogram a), GC/ b-CD (b), GC/GO (c) and GC/b-CD-rGO (d) electrodes are compared in Fig. 5A, aer incubation with 1 mM CM in 0.1 mM phosphate buffer solution, free of CM, at the scan rate of 50 mV s À1 . Inset Fig. 5A shows cyclic voltammograms of these electrodes in 0.1 mM phosphate buffer solution, free of CM, at the similar scan rate. No oxidation current was observed at the voltammograms of electrodes before incubation with CM. Aer incubation with CM, an obvious CM oxidation peak was observed at the surface of GC/b-CD-rGO electrode. The oxidation peak of CM appeared at 0.7 V vs. Ag/AgCl electrode which was attributed to the phenolic groups of CM. This was an irreversible oxidation process, in good accordance with previous reports. 53,54 A very small oxidation peak at the surface of GC/GO and GC/b-CD was also observed. These results demonstrated that CM intensely accumulated at the surface of GC/b-CD-rGO during the incubation, and, b-CD-rGO modied electrode signicantly increased the sensitivity of electrode for CM sensing/measuring.
The electrochemical oxidation properties of GC/b-CD-rGO electrodes were investigated using DPV. This technique is an effective method for measuring the electroactive species. Fig. 5B shows differential pulse voltammograms of GC/b-CD-rGO electrode before (voltammograms a) and aer accumulation with 1 mM CM (b) in phosphate buffer electrolyte solution (pH 7), free of CM. Aer the incubation with CM, an obvious electrochemical oxidation peak was observed which was originated from the accumulation of CM at the surface of electrode.

Scan rate study
The scan rate study is an important stage in characterizing the electrochemical behaviour of CM for the GC/b-CD-rGO electrode. The effect of potential scan rate (n) on both cathodic and anodic peak currents was investigated by cyclic voltammetry at different scan rates. Fig. 6A displays the cyclic voltammograms of 50 mM CM for GC/b-CD-rGO electrode at the scan rate ranged from 20 to 300 mV s À1 . Both cathodic and anodic peaks are linearly proportional with the scan rate. The linear regression equations are i pa (mA) ¼ 0.40491n (mV s À1 ) + 0.595 (mA) (R 2 ¼ 0.9981), and i pc (mA) ¼ À0.07365n (mV s À1 ) + 0.2846 (mA) (R 2 ¼ 0.999), conrming the reaction as a surface-conned process. 55 3.5 Optimization of experimental conditions for detecting curcumin 3.5.1 Effect of incubation time. The effect of CM incubation time for the modied electrode was investigated via the oxidation current peak of 1 mM CM. For this purpose, the b-CD-rGO modied GC electrode was immersed in the CM solution for 15, 30, 45, 60 and 75 minutes, and the corresponding differential pulse voltammograms were recorded in the phosphate buffer electrolyte solutions (pH 7), free of CM. As shown in Fig. 7A, the maximum oxidation peak belongs to the 45 minutes incubation time, therefore, 45 minute time was selected as the optimum incubation time for the CM characterization.
3.5.2 Effect of pH. The effect of pH on the electrochemical behavior of CM in the GC electrode modied with b-CD-rGO nanocomposite, was investigated aer incubation with 1 mM CM in phosphate buffer electrolyte solutions, free of CM, with the pH range 4.0-10 using DPV (Fig. 7B). For all pH levels, an obvious electrochemical signal was observed. The oxidation peak potential of CM shied positively when the pH increased; this indicated that the electrochemical oxidation of CM was proton transfer dependant. The maximum oxidation current for CM was obtained at pH 7.
3.5.3 Characterizing of optimized mass ratio b-CD/GO. The effect of mass ratio b-CD/GO on the electrochemical oxidation of CM was investigated in the synthesis process of b-CD-rGO nanocomposite. b-CD-rGO nanocomposite was prepared with different mass ratios of b-CD/GO (0.05/1, 0.1/1, 0.25/1, 0.5/1, 1/1 and 1/2), and their corresponding b-CD-rGO modied electrodes. Aer incubation with 1 mM CM, the electrochemical behavior of these electrodes was examined in phosphate buffer electrolyte solution (pH 7), free of CM. As shown in Fig. 7C, the oxidation peak current of CM decreased with an increase of b-CD/GO mass ratio; this was attributed to non-conductive property of b-CD molecules. When b-CD/GO ratio was in the range of 0.1/1, the oxidation peak current maximized. At the lower b-CD/GO mass ratio, the oxidation current of CM decreased, due to the low concentration of b-CD which acted as a receptor for trapping CM. Thus, 0.1/1 was selected as the optimum state for sensor fabrication.
3.5.4 Effect of b-CD-rGO quantity on surface modication and electrode performance. Modifying the amount of b-CD-rGO in the electrode surface can change the b-CD-rGO thickness and function of the electrode surface. To cover this, DPV of electrodes aer incubation with CM were recorded for GC electrode modied for different amounts of 1, 2, 3, 4, 5 and 6 mg mL À1 b-CD-rGO suspension. The oxidation peak current increased by increasing the concentration of composite from 1 to 4 mg mL À1 (Fig. 7D). For the concentrations up to 4 mg mL À1 , the oxidation peak current decreased, because by increasing the concentration of composite at the surface of electrode, the material was wasted. Therefore, 4 mg mL À1 was selected as the optimum concentration for fabricating the modied electrode.

Taguchi design results
Taguchi design method was employed to study the effects of electrolyte pH and incubation time of CM on the modied electrode, as well as the effect of b-CD/GO mass ratio and b-CD-rGO quantity on the GC/b-CD-rGO electrode response. This method maintains the interactions among experimental parameters. Table SI-2 † shows the orthogonal array of experimental runs and their response factors according to experimental runs. Based on these results, corresponding differential pulse voltammograms of 18 experimental runs were recorded (data not shown). The main aim of experimental design is to optimize important control factors, including the CM characterization. The graphs in Fig. 8 were used to determine the optimum parameters for CM characterization. The optimum values of pH, incubation time, GO/b-CD mass ratio and b-CD-rGO concentration were found 7, 60 minutes, 0.1 and 4 mg mL À1 , respectively. At these optimum conditions, the best response (160 mA) for CM determination was achieved. The optimum conditions from the Taguchi design are in good agreement with the data obtained by individual investigation of each parameter.

Detection of curcumin
The relationship between oxidation peak current and CM concentration in the specied optimum condition was examined using DPV method.  Table 1 represents the comparison between b-CD-rGO sensors typical electrochemical ones for detecting CM concentration.

Selectivity of sensor
Selectivity is one of the most important features of sensors. The selectivity of b-CD/rGO to CM was investigated with the presence of some antidepressant drugs, such as propranolol, clomipramine and clonazepam, which have polycyclic structure similar to CM. For DPV experiments, 0.1 mM of each compounds was dissolved in DMSO, and GC/b-CD-rGO electrode was immersed in it for 45 minutes, then, washed with distilled water. Fig. 10A shows the differential pulse voltammograms of GC/b-CD-rGO electrode aer incubation with these drugs, under the optimum condition, in 0.1 M phosphate buffer solution free of these compounds at the scan rate of 10 mV s À1 . Fig. 10B shows the plot of net response of GC/b-CD-rGO electrode against each drug, at a constant potential 0.5 V vs. Ag/AgCl electrode. As seen, these species caused no remarkable interference for characterizing CM. Thus, b-CD-rGO sensor can selectively and sensitively detect CM without any remarkable interference.

Conclusion
Herein, a simple, affordable, sensitive and selective CM electrochemical sensor was fabricated based on b-CD-rGO modied GC electrode. CM can form inclusion complex with b-CD molecules in b-CD-rGO nanocomposite while graphene nanosheets can accelerate electron transfer at the surface of modied electrode. Experimental results showed that b-CD-rGO nanocomposite greatly increased the electron transfer  kinetics and provided a suitable environment for constructing modied electrodes for sensory and biosensory applications. The mass ratio b-CD/rGO and the thickness of b-CD-rGO layer had great effects on the electrochemical characterization of CM. Experimental data and Taguchi method were used to nd the optimized pH, incubation time, b-CD/ rGO mass ratio and the concentration of electrode modi-er. Good agreement was obtained between experimental data and the results from Taguchi experimental design. These ndings revealed that the optimized pH, incubation time, b-CD/rGO mass ratio and the concentration of electrode modier were independent in the sensor fabrication. Under the obtained optimum conditions (7, 60 minutes, 0.1 and 4 mg mL À1 for, respectively, pH, preconcentration time, b-CD/rGO mass ratio and b-CD-rGO concentration), the fabricated electrochemical sensor had a very good analytical performance in comparison with similar electrochemical CM sensors. The fabricated sensor has a potential for various electrochemical sensors.