Fullerene/MWCNT/Nafion Modified Glassy Carbon Electrode for the Electrochemical Determination of Caffeine

Herein, a convenient method based on a fullerene/multiwalled carbon nanotube/Nafion modified glassy carbon electrode (fullerene/MWCNT/Naf/GCE) for the electrochemical determination of caffeine (CAF) is reported. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were used to study ionic exchange properties and conductivity the proposed electrode using [Fe(CN)6] redox couple. Caffeine gave an irreversible oxidation peak around +1.33 V (vs. Ag/AgCl reference electrode) in HClO4 (pH 1). The linear dependence of the peak current with the square root of the scan rate showed that the electron transfer process is controlled by diffusion. After optimization of key analytical parameters involved in differential pulse voltammetry (DPV), the oxidation peak current varied linearly with CAF concentration in the range of 10 to 1000 μM. A detection limit of 7.289 × 10 M (S/N = 3) was found. Kinetic and chronocoulometric studies were also performed to characterize the diffusion of CAF. The developed electrode exhibited good stability and was easily regenerated. The influence of some potential interfering compounds such as dopamine, uric acid, glucose and sulfite ions on the anodic peak current of CAF was also examined. The proposed method was successfully employed in the determination of CAF in some commercial drugs.


Introduction
The analysis of drugs is of key importance in drug quality control. In this line, How to cite this paper: Tajeu the development of sensitive, simple and rapid methods for the determination of active ingredients in drugs is a daily task for researchers working in analytical sciences. Food and pharmaceutical industries use caffeine (Scheme 1) in a wide range of products such as tea, coffee, soft drinks and medications [1] [2]. Most neuroprotective and metabolic benefits commonly associated with coffee and tea consumption have often been linked to caffeine content of these beverages [3]. Caffeine promotes gastric acid secretion and diuresis; and stimulates the central nervous [4] [5] [6] and the cardiovascular systems [7] [8]. It is however known that a high concentration of caffeine is toxic for the human body: it can cause nausea, trembling, nervousness, depression, hyperactivity and mutation effects [7] [9]. It constitutes a risk factor for cardiovascular diseases and asthma disease [4] [10]. The consumption of caffeine above 200 mg/day may lead to death [11]. Therefore, several effective methods have been developed for the detection and quantification of caffeine in various samples. Amongst these methods, the most used are simple chromatographic methods [12] [13] [14], coupled in some cases with mass spectrometry [15] [16]. These techniques are usually expensive and time-consuming when compared to electrochemical methods [17]. However, the high positive oxidation potential of caffeine limits the use of common electrodes such as glassy carbon electrode [18] and carbon paste electrode [19]. Nafion (Naf) is widely used as chemical modifier due to its capacity to decrease the oxidation potential of some organic compounds. It is a perfluorinated sulfonated cation exchanger polymer widely used in electrochemistry to modify electrodes because it possesses many excellent properties such as chemical inertness, thermal stability, capacity, mechanical strength, huge cation exchange and antifouling properties [20]. Several nafion-based composite electrodes were used for the quantification of caffeine, that include nafion/modified boron-doped diamond [21], nafion/carbon nanotube [22], nafion/graphene modified electrode [23], nafion modified carbon electrode [24], nafion/graphite nanoplatelets [25], nafion/attapulgite [26] and nafion-ruthenium oxide pyrochlore modified glassy carbon electrode [27].
transfer and to their strong electrocatalytic ability in electrochemical reactions [27]. Thus, their application as new effective electrocatalysts for various chemical and biochemical reactions and in the design of electrochemical sensors is undergoing increasing investigation [28]. Therefore, it is expected that, by combining fullerenes carbon nanotubes and nafion as unique electrode modifier, a synergistic effect of these components could lead to a sensitive sensor.
The present work describes the determination of caffeine via differential pulse voltammetry (DPV) using a glassy carbon electrode (GCE) modified by fullerene/MWCNT and nafion. Prior to its use for electroanalytical studies, the thin film electrode (fullerene/MWCNT/Naf) was characterized by electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV). Some parameters were investigated to obtain the best conditions for caffeine determination.

GCE Surface Modification
Prior to its modification by drop-coating, each GCE was polished with alumina pastes on a billiard cloth. They were then placed in a 1:1 ethanol-water solution and properly cleaned in a sonicator for 10 min to eliminate any remaining alumina particles. 3 mg of fullerene/MWCNT were dispersed in 1 mL Nafion (1%) and then in 1 mL ethanol by ultrasonication to give homogeneous fullerene/ MWCNT/Naf and fullerene/MWCNT dispersions. The GCE surface was coated with 2 µL of each resulting dispersion and dried at room temperature for 15 min to obtain the working electrode (hereafter referred as fullerene/MWCNT/Naf/ GCE). For comparison purposes, a bare GCE covered by a thin film of nafion (Naf/GCE) was also prepared.

Electrochemical Procedure
Cyclic voltammograms were recorded in HClO 4 solution containing a given concentration of CAF, in the potential range from +0.6 V to +1.8 V. For trace analysis of caffeine, DPV was performed using the following optimized parameters: pulse amplitude 50 mV, step potential 5 mV, initial potential +1.0 V, final potential +1.7 V, equilibrium time 5 s. The GCE before and after its modification was characterized by EIS set at the frequency range from 1 kHz to 0.01 Hz, and with an amplitude of 15 mV.
For real sample analysis, two pharmaceutical preparations were purchased locally. The analyzed solutions were obtained by dissolving a capsule of Brupanax and Pipadol (Extra) drugs in HClO 4 . The corresponding molar concentration of the solutions was calculated and a chosen volume of each of them directly added to the measurement cell. For chronocoulometric measurements, the potential was stepped from 0.0 V to +1.4 V for a period of 1 min with a sample time of 0.2 s. tice that, on fullerene/MWCNT/Naf/GCE the peak current is significantly higher than those registered on the bare GCE and on Naf/GCE. The oxidation process of CAF on all electrodes is irreversible, which is consistent with other reports [29]. The oxidation mechanism of CAF has been revealed by Hansen et al. [30].

Electrochemical Behavior of CAF at the Fullerene/MWCNT/Naf/GCE
Compared with the GCE andie, Naf/GCE, the oxidation potential at the fullerene/MWCNT/Naf/GCE shifted negatively from 1.522 to 1.393 V. This phenomenon may be an evidence of catalytic effect of fullerene/MWCNT, associated to its large specific surface area [31] [32].

Electrochemical Impedance Spectroscopy (EIS)
EIS is a valuable method to characterize the electron transfer properties of the electrode surface during the modification process [33]. Figure 2 presents the EIS curves recorded in 1.0 mM [Fe(CN) 6 ] 3−/4− using the bare GCE, Naf/GCE and fullerene/MWCNT/Naf/GCE. The resistance of charge transfer (R ct ) deriving from the recorded Nyquist plots with the lowest value (14 kΩ•cm 2 ) was obtained on the unmodified GCE (inset in Figure 2). Upon the modification of this later using only a nafion film, the R ct increased to 18.77 kΩ•cm 2 and further to 64.31 kΩ•cm 2 on fullerene/MWCNT/Naf/GCE. This is indicative of a better electrode surface coverage with a non-conductive film from the bare GCE to fullerene/ MWCNT/Naf/GCE.

Chronocoulometric Study
The electrochemical effective surface area of the GCE and fullerene/MWCNT/ Naf/GCE was calculated from the slope of the plot of Q versus t 1/2 obtained by chronocoulometry ( Figure 3 Based on the slope of the linear relationship between Q and t 1/2 (Figure 3(A)),   (Figure 3(B)).

Effect of Scan Rate
The effect of potential scan rate (v) on the DPV current density (j pa ) and peak potential (E pa ) on fullerene/MWCNT/Naf/GCE in 0.1 M HClO 4 (pH 1) + 10 −3 M CAF was investigated by CV ( Figure 4). As observed, E pa shifted to a positive value with an increase of the scan rate, along with a concurrent increase in anodic peak current. A plot of logj pa vs logv (result not shown) was used to determine whether the oxidative behavior of CAF was due to diffusion or adsorption. The slope of 0.87 > 0.5 for a theoretical diffusion-controlled process allowed to conclude that a mixed behavior showing both adsorption and diffusion-controlled processes were involved [36]. When the scan rate increased, the anodic peak potential at the fullerene/MWCNT/Naf/GCE shifted positively. For further verification of the value of α, the Laviron equation for an irreversible electrode process was used [37]: where E 0 is formal potential, k 0 is heterogeneous electron transfer rate constant, American Journal of Analytical Chemistry

Effect of Supporting Electrolyte and Its pH
Usually, the nature of the electrolyte solution affects the electrochemical behavior of the analyte, as it was observed in previous studies for CAF [38] [39] [40] [41]. The electrooxidation of CAF involving protons, different decimolar acidic solutions (H 2 SO 4 , HClO 4 , CH 3 COOH, and HCl) were tested for the differential pulse voltammetric determination of CAF on fullerene/MWCNT/Naf/GCE. Although a peak was recorded for all investigated solutions, HClO 4 generated the best signal in terms of shape and intensity. This allowed to use 0.1 M HClO 4 as supporting electrolyte for further experiments. The redox behavior of most biomolecules is affected by the pH of supporting electrolyte [42]. In this study, the effect of pH on peak current (j pa ) and peak potential (E pa ) of CAF was evaluated within the acidic pH range (0.5 -2.5). From pH 0.5, the peak current increased slightly at pH 1, then decreased from pH 1.0 to pH 2.5 ( Figure 5(A)). one can also observe that the anodic peak potential of CAF shifted towards less positive values as the pH changes from 0.5 to 2.5 ( Figure 5(B)). Therefore, pH 1.0 was chosen for the subsequent experiments ( Figure 5).

Interference Study, Repeatability and Stability of the Electrode
The next step in the work focused on the evaluation of the specificity of the sensor when operating in the presence of interfering agents frequently found in commercialized drinks and drugs. A CAF solution (10 −4 mol•L −1 ) spiked with various species was evaluated under the same experimental conditions in 0.1 M HClO 4 by DPV. The selected interfering compounds include citric acid, sulfite ions, glucose and dopamine that were added in the electrochemical cell, with concentrations 100 to 200-fold higher than that of CAF (10 −4 mol•L −1 ). The tolerance limit, defined as the concentration ratio of the additive over CAF causing ±5.0% relative error was used to measure the interfering effect. The results obtained (Table 1) showed that the oxidation peak of CAF is slightly affected by citric acid as more than 100-fold excess of this compound did not significantly influence the peak current of caffeine. However, dopamine and sulfite ions were the most interfering agents, somewhat limiting the selectivity of the proposed sensor for media containing high amount of such interfering compounds.

Variation of CAF Concentration, Calibration Graph and Analytical Application
The DPV curves of CAF in the concentration range from 10 to 1000 µM at fullerene/MWCNT/Naf/GCE in 0.1 M HClO 4 (pH 1.0) are shown in Figure 6. The recorded peak current increased linearly with an increase in the concentration of CAF. A calibration plot of the oxidation peak current against CAF concentration is linear, with a correlation coefficient of 0.999 and a detection limit of 72.89 nM (S/N = 3).
As compared with other modified electrodes, it is obvious that the present method is comparable or better in the determination of CAF since it provides a significantly low detection limit ( Table 2). On the other hand, the preparation of the modified electrode is simple and reproducible. The applicability of the developed method to the determination of caffeine in real samples was investigated using the standard addition method. The experimentally detected values and the labeled values are compared in Table 3. It was found that the results obtained using fullerene/MWCNT/Naf/GCE are in a good agreement with the content label.

Conclusion
This work showed the ability of fullerene/MWCNT/Naf/GCE electrode towards the electrochemical quantification of caffeine. The obtained results showed that the electrocatalytic oxidation of caffeine at the surface of the modified electrode occurs at a potential of about 1.33 V. The anodic peak current of caffeine was proportional to the concentration in the range from 10 to 1000 µM, with a detection limit 72.89 nM and correlation coefficient of 0.999. The simple preparation, good stability, excellent reproducibility, sensitivity and rapid analysis make the developed method suitable for routine determination of caffeine in real samples.