MWCNTs were supplied from DropSens, HCON(CH₃)₂ (68%, Sigma-Aldrich), and NaNO₂ (98%, Fluka) were of analytical quality. The phosphate-buffered solution (PBS) was elaborated by mixing 0.1 M NaH₂PO₄ and Na₂HPO₄ solutions (98%, Sigma-Aldrich). Deionized water was employed for all preparations.

The electrochemical study was carried out at room temperature in a three-electrode arrangement in the standard cell; the data were collected with a potentiostat/galvanostat (model PGZ402, Radiometer Analytical, France) coupled with a PC-computer. A saturated calomel electrode (SCE) (Hg/Hg₂Cl₂) as a reference, Pt wire as an auxiliary electrode, and MWCNTs/Pt as a working electrode (Ø = 2 mm) were used on the electrochemical cell. The pH was measured with a digital pH meter (model inoLab^® pH 7110).

The working electrode was mechanically polished to a mirror-like surface with Al₂O₃ (0.3 µm) slurry on a polishing cloth. After being washed with distilled water, the bare electrode was sonicated (5 min) in ethanol and dried at ambient temperature. A suspension was prepared by dispersing 1 mg MWCNTs in 1 mL DMF under sonication for 30 min. The MWCNT-modified Pt was prepared by casting a known volume (3 µL) of the mentioned suspension on the Pt electrode surface using a micropipette and left to dry at room temperature. The surface was electrochemically cleaned in supporting electrolyte by CV between –0.25 V and 1.5 V for several cycles after each run to remove the adsorbed substances. Then the electrode surface was rinsed with distilled water.

MWCNTs/GCE (glassy carbon electrode) was electrochemically activated in PBS solution (0.1 M, pH 7.0) under scanning potential between –0.25 and 1.2 V at 50 mVs^–1 until a reproducible cyclic voltammogram was obtained. The sensor was transferred into a 10 mL glass standard cell filled with PBS solution, and aliquots of the stock solution were added. The calibration graph was constructed from the SW voltammograms plotted by successive additions from aliquots of the stock solution to 10 mL of PBS solution (0.1 M, pH 7.0); each concentration was measured in triplicate. All potential were referred to the SCE at an ambient temperature.

Figure 1a illustrates the voltammograms of the bare electrode and MWCNTs/Pt in 0.1 M PBS (pH 7.0) in the absence and the presence of 1.0 mM NO₂^– at a scan rate of 50 mVs^–1. As expected, no signal could appear on the bare Pt and modified electrode in pure 0.1 M PBS at pH 7.0. However, in the addition of 1.0 mM NO₂^–, an oxidation peak appears at 0.70 V and no reduction peak appears in the reversal scan. The cyclic voltammogram of 1.0 mM NO₂^– at the bare Pt under identical conditions is given for comparison. As shown, the oxidation peak is obtained at about 0.77 V, and no reduction peak appears in the reversal scan.

In addition, when the obtained SWV results were compared (Figure 1b), the MWCNTs/Pt shows a higher response current, which presents an increase of more than four times in the voltammetric response compared to the bare Pt electrode. This indicates that the MWCNTs/Pt improves the electrochemical reaction of NO₂^– compared with bare Pt electrode. The enhanced signal is probably due to the facile electron transfer with an improved rate and larger active surface area of the modified electrode.

The effect of MWCNT amounts on the peak current response of 1 mM NO₂^– was examined (Figure 1c). In this regard, different volumes of the suspension were cast on the bare Pt. The best response of NO₂^– was obtained when a 3 μL drop was used. Therefore, the amount of MWCNTs chosen for the preparation of the sensor was fixed at 3 μL.

The cyclic voltammograms of bare Pt and MWCNTs/Pt electrodes were carried out in 1.0 mM [Fe(CN)₆]^3–/4– containing 0.1 M KCl at a scan rate of 50 mVs^–1 (Figure 2a). A pair of redox peaks were observed at both unmodified and modified electrodes. The peak current at the modified electrode increases dramatically compared with that observed at the bare Pt electrode, suggesting that MWCNTs/Pt exhibits faster electron transfer kinetics compared with the bare Pt electrode. Using the Randles-Sevcik relation [20]:

Where n, C_o, A, v, and D_o are the number of electrons involved, the substrate concentration (1.0 mM), the active surface area, the scan rate, and the diffusion coefficient of ferrocenium/ferrocene (7.6 × 10^–6 cm/s [21]) respectively. The surface area of MWCNTs/Pt and bare Pt electrode are respectively 0.042 and 0.035 cm². The results clearly show that the electrode modified with MWCNTs film has a larger active surface area, which results in higher sensitivity.

EIS analysis of the bare Pt and MWCNTs/Pt was performed to explain the electron transfer changes between the modified electrode and the solution interface in the presence of 10 mM [Fe(CN)₆]^3–/4– containing 0.1 M KCl. In the EIS plots (Figure 2b), the high-frequency region of the impedance real part presents the solution resistance (R_s). The semicircular part at high frequency illustrates the electron transfer resistance (R_ct) while the straight line at the low-frequency ranges is ascribed to the diffusion of K₃Fe(CN)₆/K₄Fe(CN)₆ as the redox probe (Warburg impedance, W) [22–25]. Moreover, a constant phase element (CPE) is introduced to take into account the topological imperfections of the materials caused by the surface roughness of the electrode. The impedance (Z_CPE) is characterized by a slight deviation from pure capacitive behavior and is then given by [22, 23]:

Where j = (–1)^1/2, ω = 2πf being f the frequency of the applied ac potential, Q=Cωmax1-α and α is fractional and its experimental value is between 0.5 (for an ideally porous electrode) and 1 (for a perfect smooth electrode). The EIS parameters were fitted by Zview using the Randles electrical circuit (Figure 2b, insert). The results of the simulations are reported in Figure 2b (curves in red) which show a very good correspondence with the experimental data. The values of each component are shown in Table 1. As seen, the bare Pt electrode has exhibited a large semicircle with a high R_ct value, while the diameter of the semicircle of the MWCNTs/Pt electrode is much smaller with a low R_ct value compared to the bare electrode. By contrast, the values of Q are opposed to the R_ct values, which indicates that the electronic transport and mass transport resistance of MWCNTs/Pt electrodes is much smaller than those of bare Pt electrodes. Moreover, the W value of the redox probe at the bare Pt electrode is 597.8 kΩ and is decreased to 41.37 kΩ at the modified electrode, which proves that the MWCNTs film may increase the diffusion rate of the electroactive species towards the electrode surface.

EIS was also used to investigate the charge transfer property of different electrodes in 0.1 M PBS (pH = 7.0) aqueous solution having 1 mM NO₂^– (Figure 2c). By fitting the Randles equivalent electrical circuits, the R_ct caused by the faradaic reactions at the MWCNTs/electrolyte interface has a smaller semicircle diameter of 8.3 kΩ compared to 141.8 kΩ recorded for the bare Pt electrode. This feature suggests that MWCNTs improve the surface conductivity and hence the electron transfer kinetics of the analyte. Therefore, the MWCNTs/Pt-modified electrode was considered to be used in the application of nitrite electrochemical determination.

The influence of scan rate on nitrite oxidation at MWCNTs/Pt was examined to investigate the kinetic mechanism of nitrite oxidation behavior. Figure 3a shows a cyclic voltammogram of the NO₂^– at various sweeping rates (5–300 mVs^–1) in 0.1 M PBS solution at pH 7.0 using MWCNTs/Pt. The current responses are increased linearly with the square root of the scan rates (v^1/2) (Figure 3b). This supports that the oxidation of NO₂^– is governed by the diffusion of the analyte at the MWCNTs/Pt electrode.

To support these results, the dependence of log I_p on log v was investigated (Figure 3c). Normally, the slope is close to 0.5 for a diffusion-controlled and 1.0 for an adsorption-controlled process [26]. The dependence of log I_p with log v for the nitrite oxidation peak gives a slope of 0.47 with a correlation coefficient close to unity (0.99), corroborating a diffusion-controlled electron transfer process of NO₂^– on the proposed sensor.

Therefore, the number of electrons (n) involved in the reaction can be extracted from the relation [27]:

Where E_p is the peak potential, and E_p/2 is the half-peak potential. Thus, αn was found to be 1.0. The coefficient α is equal to 0.5 for an irreversible system. The number of exchanged electrons in the NO₂^– oxidation is equal to 2.0, involving two electrons in the NO₂^– oxidation to NO₃^– in agreement with previous reports on nitrite oxidation [28–32]. Hence, the following reaction is suggested:

Figure 4 depicts the current responses obtained at different pH values for the electro-oxidation of 1 mM NO₂^– on MWCNTs/Pt. As can be seen, the current response augments from pH 5.0 to 7.0, reaching a maximal pH of 7.0 and then decreases up to pH 9.0. Such results indicate that the pH domain is small in both acid and alkaline solutions, while the neutral pH is a suitable medium for the electro-catalytic oxidation of NO₂^– [32]. Hence, subsequent detection of NO₂^– was performed in PBS solution buffered at pH 7.0.

The electrocatalytic oxidation of NO₂^– on MWCNTs/Pt was also studied by ChAm in 0.1 M PBS (pH 7.0) having 10 µM NO₂^– at an applied potential of 0.7 V (Figure 5). Plotting the net current against t^-1/2 exhibits a linear relation (Figure 5, insert a), with a dominant diffusion-controlled process. By using the slope of this plot, the diffusion coefficient (D) of NO₂^– can be extracted from the Cottrell relation [33]:

Where C_o is the bulk concentration. The diffusion coefficient of NO₂^– is found to be 1.44 × 10^–5 cm²/s which is close to formerly reported values [34, 35]. This technique was also used to estimate the catalytic rate constant (k) for the interface reaction between the nitrite and the modified electrode according to equation [36]:

Where I_cat and I_L are the amperometric currents of the MWCNTs/Pt with and without NO₂^– respectively, γ (= kC_ot) is the argument of the error function, and t is the elapsed time (s). When γ > 1.5, erf (γ^1/2) is very close to 1 and the simplified equation becomes:

From the slope of the (I_cat/I_L) versus t^1/2 plot (Figure 5, insert b), the value of k was obtained as 1.43 × 10⁴ M^–1s^–1.

SWV is a pulse-sensitive method that could attenuate the background noise, allowing a higher sensitivity than the CV for low analyte concentrations [37]. SWV was performed in 0.1 M PBS (pH = 7.0) using a pulse duration time of 0.05 s, pulse amplitude of 25 mV, frequency of 50 Hz, a scan rate of 10 mVs^–1, and a step potential of 10 mV. A well-defined nitrite oxidation peak is observed at 0.67 V as shown in Figure 6a. The peak current increases linearly with increasing the NO₂^– amount in the concentration ranges from 1.0 to 1,000 µM (Figure 6b). The LOD of the fabricated sensor is found as 0.1 µM taking LOD = 3 s/m, where s is the standard deviation of the current of low NO₂^– concentration and m is the slope of the calibration graph (n = 5) [38]. The LOD is well below the maximum allowable limit for NO₂^– in potable water (43.48 μM) according to the WHO directives [3]. The relative standard deviations (RSD) were evaluated at 1.5% for five separate voltammograms of 50 μM NO₂^–. Therefore, our sensor is better suited to assess NO₂^– concentrations in real water. The analytical parameters of the elaborated electrode are listed in Table 2. The electrocatalytic performance of MWNCNTs/Pt for the oxidation of NO₂^– was superior or comparable to previously reported results as displayed in Table 3. The improved sensing performance of MWCNTs/Pt can be attributed to their good conductivity and large surface area. These results for the MWCNTs/Pt system reveal a good analytical performance in terms of simplicity and efficiency for the detection of this analyte.

To test the stability of the MWCNTs/Pt electrode, the current responses of 50 μM NO₂^– were followed for 5 consecutive days. The oxidation current of NO₂^– remains practically unchanged with an RSD not exceeding 1.5% for five successive tests, showing clearly that the prepared electrode possesses good stability. Additionally, the reproducibility of the proposed sensor was tested with five different fabricated electrodes in 50 μM NO₂^–, and the RSD of the current signal was obtained as 2.04%, suggesting the good reproducibility of this electrochemical sensor.

In order to investigate the anti-interference property of the modified electrode, the effects of common interfering minerals and organic substances on the response sensor were examined. The electrode response to a 50 µM tested solution of NO₂^– in 0.1 M PBS (pH 7.0) is 1.26 µA. After adding 100-fold excess amounts of Na⁺, Cl^–, K⁺, Zn²⁺, Ba²⁺, Mn²⁺, glucose, and ascorbic acid to the tested nitrite solution, no interference effect for the nitrite determination was observed (Table 4). The relative error values for the response were smaller than ± 5%, suggesting that the MWCNTs/Pt sensing electrode has good selectivity toward NO₂^– detection.

To verify the practical applicability of the proposed sensor, NO₂^– was analyzed on MWCNTs/Pt. Known concentrations of NO₂^– were added to both mineral and source waters. The recovery study was realized thanks to the standard addition technique and the method can be applied for the NO₂^– detection in real effluents (Table 5).