Introduction

Nitrite is an inorganic anion composed of a nitrogen atom bonded to two oxygen atoms. Its use as a preservative in food industry and as a corrosion inhibitor during industrial water preparation has been highlighted [1,2,3,4,5]. The effect of high concentrations of nitrite in environmental water bodies and human health has been discussed [3, 6, 7]. The maximum limit of nitrite in water samples should not exceed 3 mg L−1, according to World Health Organization [8]. Hence, it is necessary to develop a simple and selective sensor for continuous monitoring of nitrite so that if recommended limits by regulatory bodies are exceeded then remediation techniques can be sought quickly without posing environmental harm.

Literature survey reveals that different instrumental techniques have been used for nitrite determination with great success [9,10,11,12,13]. However, time-consuming, high costs of the equipments, and the need for qualified technicians have been cited as some drawbacks [14]. In an attempt to rectify this, electrochemists have shifted to voltammetric techniques as an option as evidenced by working principles which are easy to follow when monitoring environmental samples [15, 16]. Furthermore, voltammetric techniques involving exploitation of chemically modified electrodes (CMEs) incorporating different modifiers [2, 3, 5, 15,16,17,18,19,20,21,22] have been reported for electro-oxidation of nitrite.

The use of carbon nanotubes in electrochemical sensing of numerous environmental pollutants has received attention [14, 15, 23] mainly due to their small particle size, enhanced voltammetric peaks, and high electronic conductivity [24, 25]. On the other hand, metal and metal oxide nanoparticles are the mostly widely employed nanomaterials due to excellent and catalytic properties. TiO2 has been used in sensor fabrication [20, 26,27,28,29] due to low cost, non-toxicity, large surface area, strong adsorptive ability, high uniformity and excellent catalytic activity [29]. In continuation of our studies involving CMEs [14], carbon nanotubes were used as substrates for anchoring TiO2 nanoparticles in order to provide influence on morphology and electrochemical responses of nitrite through oxidation process since there is no interference from nitrate and molecular oxygen [15, 16, 30]. The ease of fabrication, simplicity, and low-detection limit are the main advantages of this sensor over the previous reported ones.

Experimental

Materials

Phosphate buffer solutions (as supporting electrolyte) with different pH values were prepared by mixing standard stock solutions of 0.10 M Na2HPO4 and 0.10 M NaH2PO4. NaNO3, K3Fe(CN)6, K4Fe(CN)6, N,N-dimethylformamide (DMF), TiO2 nanopowder and MWCNT were obtained from Sigma-Aldrich (South Africa). All solutions were prepared using ultra Millipore water from Milli-Q Water Systems (Millipore Corp. Bedford, MA, USA). The purification of MWCNTs to remove metal oxide catalysts was followed as reported [2].

Equipment

Transmission electron microscopy (TEM) image was obtained from a Zeiss Libra TEM 120 model operated at 90 kV using carbon-coated 200 mesh grids. The energy dispersive X-ray spectrometer (EDS) images were obtained using a TESCAN Vega TS 5136LM electron microscope. Cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS), chronoamperometry, linear sweep voltammogram (LSV), and differential pulse voltammetry (DPV) were performed using an Autolab potentiostat PGSTAT 302F (Eco Chemie, Utrecht, Netherlands) equipped with NOVA 1.10 software.

Preparation of TiO2/MWCNT nanocomposite

An appropriate ratio of MWCNT:TiO2 (3:1 w/w) was mixed and then suspended in 80 mL of Millipore water through sonication for 30 min. The suspension obtained after sonication was magnetically stirred for 12 h and the precipitate obtained thoroughly washed with Millipore water using centrifugation process for 30 min and then dried prior to electrode modification.

Modification of electrodes

A three-electrode system was used with glassy carbon electrode (GCE) as working electrode, an Ag/AgCl (saturated KCl) reference electrode and platinum wire auxiliary electrode. The GCE was polished to a mirror finish with alumina slurry (0.3 μm). The electrode was sonicated in Millipore water for 5 min during the three successive cleaning stages and finally dried in a stream of nitrogen. The drop dry technique was employed for electrode modification. TiO2, MWCNT, and TiO2/MWCNT composite (1 mg mL−1) in DMF were used as electrode modifiers to give TiO2/GCE, MWCNT/GCE, and TiO2/MWCNT/GCE which were stored at room temperature when not in use.

Results and discussion

Microscopic characterization

A monodispersion of TiO2 nanoparticles with almost homogeneous size is confirmed by TEM image (Fig. 1). Black dots represent nanoparticles of average size 19 nm that are present on the image as depicted on the corresponding histogram (inset) showing size distribution. These characteristic features are good in electrochemical sensing.

Fig. 1
figure 1

TEM image for TiO2 modifier. Corresponding histogram (inset)

Full size image

EDS was further used to show the elemental composition of MWCNT (Fig. 2a) and TiO2/MWCNT composite (Fig. 2b). Carbon was mainly observed upon modification with MWCNT alone. However, titanium and oxygen emerged after introduction of the nanosized TiO2. The aluminum signal evident in the EDS spectrum could be coming from the aluminum sample holder used during analysis.

Fig. 2
figure 2

EDS for (a) MWCNT (b) TiO2/MWCNT composite

Full size image

CV and EIS studies on modified GCE

Cyclic voltammetry (Fig. 3a) and electrochemical impedance spectroscopy (Fig. 3b) were employed to study the electron transfer ability of the modification process in 1 mM [Fe(CN)6]3−/4−containing 0.1 M KCl. Table 1 shows the difference between anodic and cathodic peak potential (ΔE p) of the modified electrodes. It has been highlighted that a low value of ΔE p reflects good electron transfer for the redox probe [23]. The trend in electron transfer in the redox probe is: TiO2/GCE (102 mV) < bare GCE (92 mV) < MWCNT/GCE (85 mV) < TiO2/MWCNT/GCE (80 mV). The results confirmed that anchoring TiO2 on MWCNT offered fast electron transfer ability.

Fig. 3
figure 3

a CVs of the modified GCE in the presence of 1 mM [Fe (CN)6]3−/4− in 0.1 M KCl at a scan rate of 50 mV s−1. b EIS behavior of modified GCE measured by impedance in the presence of 1 mM [Fe (CN)6]3−/4− in 0.1 M KCl. c Plot of i pa and R ct vs. modified GCEs. d Bode plots for a: GCE, b: TiO2, c: MWCNT, and d: TiO2/MWCNT/GCE

Full size image
Table 1 Summary of electrochemical parameters of modified electrodes
Full size table

In the EIS technique, the semicircle part at higher frequencies represents the electron transfer limited process and its diameter is equated to the electron transfer resistance (R ct). From the Nyquist plots (Fig. 3b) and Table 1, the GCE (295 Ω) displays a small semi-circle indicating a low transfer resistance. The TiO2/GCE (390 Ω) has a larger diameter suggesting that TiO2 acted as insulating layer and barrier. The MWCNT/GCE (190 Ω) showed less resistance. An almost straight line and a small circle portion which was quite small was observed on the TiO2/MWCNT/GCE and might be due to increase in reactive area, reduced interfacial resistance, and the composite making the electron transfer faster. Such small R ct values characterized by relatively small differences between them for different probes have been reported before for probes designed for different analytes [31, 32]. The obtained results are in good agreement with peak current (i pa) values obtained from CV measurements (Fig. 3c). The information from the Bode plots (Fig. 3d) further supports that modified surfaces have different behaviors since their phase angles shifted to different frequencies.

Nitrite detection

Influence of pH

The pH effect on i pa of nitrite (87 μM) at TiO2/MWCNT/GCE was investigated using CVs in 0.1 M PBS (pH 2–10) at a scan rate of 100 mV s−1. The i pa generated from voltammograms increased from pH 2 to 7 and then decreased to pH 10 (Fig. 4). The i pa below pH 6 is due to the instability of the nitrite in acidic solution and hence its consequent conversion to nitrate [26, 30]. The lack of protons [18, 33] in supporting electrolyte above pH 7 has been suggested as a cause for decrease of i pa and similar trends with different probes have been reported [5, 15, 26, 34, 35]. Consequently, all the subsequent experiments were carried out in PBS at pH 7for the oxidation of nitrite.

Fig. 4
figure 4

Effect of solution pH on the i pa of 87 μM NO2 − at the TiO2/MWCNT/GCE in 0.1 M PBS

Full size image

Cyclic voltammetry detection of nitrite

The voltammogram in 0.1 M PBS (pH 7) showing only background current before addition of nitrite using GCE probe is shown (Fig. 5a). The overpotentials ranging from 750 to 1100 mV for the oxidation of nitrite is shown by the probes prepared. It can be observed that the use of MWCNT in combination with TiO2 significantly lowered the overpotential, compare the GCE (1100 mV) and TiO2/GCE (1080 mV) with TiO2/MWCNT/GCE (750 mV). A negative shift of 250 mV of TiO2/MWCNT/GCE compared to GCE was deduced. However, the background corrected currents also improved, compare TiO2/MWCNT/GCE (i pa = 18.8 μA) with GCE (i pa = 7.8 μA), TiO2/GCE (i pa = 4.9 μA), and MWCNT/GCE (i pa = 12.0 μA) (Table 1 and Fig. 5a, b inset). The enhancement in i pa of TiO2/MWCNT/GCE is due to good conductivity of nanotubes. The behavior of nitrite on TiO2/MWCNT/GCE was checked through successive scans (Fig. 6). The electrode was characterized by a very small reduction in i pa showing more stability and suggesting more usability.

Fig. 5
figure 5

CVs of GCE in the absence of NO2 − (a) TiO2/GCE, (b) GCE, (c) MWNCT/GCE, (d) TiO2/MWCNT/GCE, and (e) in the presence of 87 μM NO2 − in 0.1 M PBS (pH 7.0) at 100 mV s−1; error bar = ± S.D. and n = 3. Inset B: bar graph of i pa vs. modified electrodes

Full size image
Fig. 6
figure 6

Continuous cyclic voltammograms for TiO2/MWCNT/GCE in 87 μM NO2 − in 0.1 M PBS (pH 7.0) at 100 mV s−1

Full size image

Kinetic studies for nitrite detection

The detection of nitrite at different scan rates on the modified electrode was investigated (Fig. 7). E pa shifted to more positive values with increasing scan rate (v) showing that the reaction is irreversible [14]. Further, a linear dependence of i pa on v 1/2 from 10 to 175 mV s−1 (Fig. 7 inset) [36] suggests a diffusion-controlled process on the electrode.

Fig. 7
figure 7

CVs of TiO2/MWCNT/GCE in 0.1 M PBS (pH 7.0) containing 87 μM nitrite, (a–h) correspond to 10, 25, 50, 75, 100, 125, 150, and 175 mV s−1. Inset: i pa vs. v 1/2, error bar = ± S.D., and n = 3

Full size image

The electrons involved in oxidation process were calculated using information form CV. Using E pa = 750 mV, E p/2 = 693 mV and substituting into Eq. 1, αn a = 0.84 which finally gives “n” 1.67 (ca. 2) by assuming electron transfer coefficient (α) of 0.5 [37].

$$ {E}_{\mathrm{p}}-{E}_{\mathrm{p}/2}=47.7/\upalpha {n}_{\mathrm{a}}\ \mathrm{mV}\ \mathrm{at}\ {25}^{{}^{\circ}}\mathrm{C} $$
(1)

Linear sweep voltammetry detection of nitrite was investigated from 1.0 to 18 μM at the modified electrode (Fig. 8). The i pa increased with a slight negative shift in E pa upon each successive addition of nitrite. A plot of i pa vs. [NO2 −] gave a linear relationship (Fig. 8 inset A). Furthermore, the linear relationship from a plot of log i pa vs. log [NO2 −] (Fig. 8 inset B) implies first-order kinetics with respect to the analyte.

Fig. 8
figure 8

LSVs obtained for nitrite in the concentrations ranging from 1.0 to 18 μM (a–g), when nitrite was added in steps of 3 μM each in 0.1 M PBS (pH 7). Inset A, shows calibration plot and B shows plot of log i pa vs. log [NO2 −]; error bar = ± S.D. and n = 3

Full size image

Chronoamperometry

The diffusion coefficient (D) was calculated using information from chronoamperometry (Fig. 9). A plot of i vs. t −1/2 (Fig. 9a, inset) with the best fits for different concentrations of nitrite is shown. Additional, plot of slopes vs. [NO2 −], is also given (Fig. 9b, inset). From the resulting slope and Cottrell equation [36], an average value of D (= 2.15 × 10−6 cm2 s−1) was deduced and is close to those reported [1, 2].

Fig. 9
figure 9

Chronoamperograms obtained at TiO2/MWCNT/GCE in 0.1 M PBS (pH 7.0) containing different concentration of NO2 −. Curves of 1–5 correspond to 0.02, 0.05, 0.15, 0.25, and 0.35 mM NO2 − for a potential step of 0.95 V. Inset: a Plots of i vs. t −1/2. b Slope vs. [NO2 −]/mM

Full size image

Differential pulse voltammetry

To demonstrate the ability of TiO2/MWCNT composite as good sensor candidate for voltammetric traces is one of the important characteristics. In this work, differential pulse voltammograms (Fig. 10) gave a detection limit (LOD) of 0.011 μM calculated using 3 s/b, where s is the standard deviation of the i pa of blank (n = 10) and b is the slope of the calibration plot. It can also be seen that current increased linearly with nitrite concentration in the range of 0.02 to 600 μM (Fig. 10, inset) and a sensitivity of 0.289 μA μM−1 is reported. The results are much improvement to literature values (Table 2) as shown by good stability of composite on the electrode.

Fig. 10
figure 10

DPV as a function of different nitrite concentrations (a) blank (PBS pH 7.0), (b) 0.02 μM, (c) 5 μM, (d) 10 μM, (e) 15 μM, (f) 100 μM, (g) 200 μM, (h) 300 μM, (i) 400 μM, (j) 500 μM, and (k) 600 μM. Inset: corresponding calibration plot of i pa vs. [NO2 −, error bar = ± S.D., and n = 3

Full size image
Table 2 Comparison of different chemically modified electrodes for the determination of nitrite with the present fabricated sensor
Full size table

Interference study

Interference measurements were performed using the mixed-solution method (mixing the interferents at different concentrations with 200 μM nitrite in 0.1 M PBS pH 7.0) using differential pulse voltammetry. The peak current change was less than 10% after 100-fold concentration of Mg2+, Ca2+, Zn2+, Cu2+, and Na+ and 20-fold concentration of ascorbic acid and urea. It was observed that metal ions have minimal effect on the detection of nitrite at the sensor surface. This was attributed to excellent solubility of the metals nitrites in aqueous media (see Table 3). The obtained results and proposed sensor have a good selectivity during electrochemical analysis.

Table 3 Effect of interferents
Full size table

Analytical applications

The practical applicability of the proposed method using the TiO2/MWCNT/GCE sensor was investigated by analysis of nitrite in samples from a dam receiving waste from treatment plant. The samples were collected from three different locations in the dam and analyzed without any treatment. The standard addition method was used for determination. A 1.0-mL water sample was added into each of the series of the 10-mL volumetric flasks. Nitrite standard solutions of different concentrations were added to the flask, which were made up to volume with 0.1 M PBS (pH 7.0). An aliquot of 3.0 mL of the solution was placed in a cell for determination and the results are shown in Table 3. The recoveries were in the range of 96.9–98.9%, which confirmed better analysis of nitrite in aqueous solution. For validation, a UV-Vis was used, and relative error values of less than 10% were obtained (Table 4).

Table 4 Results for determination of nitrite in water samples (n = 3)
Full size table

Conclusion

A stable sensor for nitrite was fabricated by drop dry method using TiO2/ MWCNT composite and showed appreciably low detection limits. Good recoveries were obtained and ranged from 96.9–98.9%. The constructed electrode offers a decrease on overvoltage of nitrite oxidation. The interference study showed no significant changes in the detection of nitrite. The results obtained demonstrate that the sensor could be used as a management tool for assessing the quality of wastewater.