Fabrication of the ePAD
An ePAD was fabricated by the sticking of a patterned self-adhesive glossy paper onto the surface of a commercial type of wax paper. The different steps of the fabrication process are shown in Fig. S1 (Supporting Information). The patterned self-adhesive glossy paper was used as a stencil to control the size and the shape of the electrode system. Furthermore, this layer could provide a hydrophobic area on the ePAD. Basically, the reproducibility of the surface area of an electrode is a critical parameter for the development of an electrochemical sensor. The laser cutting technique provides a precise stencil from glossy adhesive paper. The reproducibility of the surface area of electrodes also depends on laser cutter parameters, including power and speed. In this study, the effects of these parameters on the reproducibility of the patterns were evaluated by the ImageJ software. The mean calculated for the surface area in ten patterns and the relative standard deviations (RSD) are listed in Tables S1 and S2. The RSD values calculated for the optimum parameters (i.e. the power of 12.5% and the speed of 10 mm/s) were around 4%. Therefore, this method can be applied in the reproducible construction of ePADs.
In the next step (Fig. S1e,f), the electrodes and the connections were fabricated by a manual screen-printing process. The as-fabricated ePAD was mechanically stable and flexible (Fig. 1A). The thickness of the layer of graphite was investigated using SEM (Fig. 1B–D). It could be controlled easily by adjusting the pattern thickness. The SEM image showed the uniform printing of the graphite layer on the paper surface with a thickness of 40.5 ± 1.9 μm. The reproducible paper-based electrochemical system could, thus, be fabricated using the proposed method.
The surface morphology of the working electrode before and after the deposition of AuNPs was observed by SEM (Fig. S2). The AuNPs were successfully distributed onto the graphite surface with a sphere-like structure. Figure S3a shows the X-ray diffraction spectra of the Au/ePAD. The diffraction peaks located at 2θ = 43.2° correspond to (100) crystal planes of graphite, while the diffraction peaks at around 2θ = 36.3°, 38.1°, 50.4°, and 74.1° can be assigned to (111), (200), (220) and (311) planes of the Au nanoparticles respectively27,28. These results suggest the formation of Au nanoparticles on the graphite working electrode surface. Furthermore, as the energy-dispersive X-ray spectrum of the Au/ePAD suggests in Fig. S3b, the working electrode surface was composed of Au and carbon.
Electrochemical investigation of the ePAD
Cyclic voltammetry was conducted to investigate the electrochemical performance of the ePAD in the presence of 1.0 mM Fe(CN)3−/4− containing 0.1 M KCl at a scan rate of 100 mV s−1. The cyclic voltammogram in the absence of electroactive species (curve a, Fig. 2A) showed a low background current without an oxidation or reduction peak. This indicates that graphite ink components are not electroactive in the CVs potential range. However, in the presence of 1.0 mM [Fe(CN)6]4−/3−, a pair of redox peaks appeared with the ΔEp of 176 mV (curve b in Fig. 2A). This suggests that [Fe(CN)6]4−/3− has a quasi-reversible behavior. In the literature, ΔEp values higher than 150 mV have been reported for graphite-based electrodes19,29,30, which is attributed to the high electrical resistance of the binder31. Furthermore, the ratios of Ipa to Ipc for the 1st and the 20th cycles were 1.05 and 1.02, indicating the stability of ferricyanide ions produced at the surface of the electrode. It is noteworthy that no absorption or adsorption of [Fe(CN)6]4−/3− ions occurred on the graphite working electrode of the ePAD.
Electrochemical stability is an important parameter for the development of various electrochemical sensors, especially paper-based sensors. In this case, after 20 cycles on the ePAD, the peak potentials did not have a noticeable change, and the peak currents only varied for less than 2% (Fig. 2B). These results suggest that ePADs have acceptable electrochemical stability to be applied as sensors.
The day-to-day reproducibility of inks was investigated by the measurement of the peak currents for the ePADs constructed on different days (Fig. S4). The RSD values calculated for different inks were around 5%, indicating that the process of ink preparation had good reproducibility. Moreover, the RSDs of five replicates from a single-batch ink were less than 3%.
The electrochemical characterization of ePAD, Au/ePAD, and NTX–Au/ePAD was performed by cyclic voltammetry in the presence of 1.0 mM [Fe(CN)6]4−/3− containing 0.1 M KCl at a scan rate of 100 mV s−1. A pair of well-defined redox peaks was observed at the ePAD (Fig. 3A). When the Au nanoparticles were deposited on the graphite working electrode, the electroactive surface was increased, and the peak current was enhanced consequently. After the modification of AuNPs with NTX (NTX–Au/ePAD), due to the acceleration of the electron transfer between [Fe(CN)6]4−/3− in solution and the electrode surface, the peak current improved significantly; the value of ΔEp for the NTX–Au/ePAD (ΔEp = 149 mV) was smaller than that for the Au/ePAD (ΔEp = 160 mV) and the ePAD (ΔEp = 176 mV).
The interface properties of NTX–AuNPs on the ePAD surface were evaluated using electrochemical impedance spectroscopy (EIS). Figure 3B shows the Nyquist plots for the ePAD, the Au/ePAD and the NTX–Au/ePAD. The Rct was decreased as AuNPs and NTX were added to the ePAD and the Au/ePAD respectively, suggesting that the presence of NTX on AuNPs effectively enhanced the electron transfer of the electrode. These results confirm the assembly of NTX and AuNPs on the surface of the working electrode.
In order to understand the nature of the redox currents, the cyclic voltammograms of the NTX–Au/ePAD were recorded at various scan rates. Figure 3C shows the cyclic voltammograms of the NTX–Au/ePAD in the presence of 1.0 mM Fe(CN)3−/4− containing 0.1 M KCl at different scan rates (ν). The figure also denotes the dependence of cathodic and anodic peak currents on the square root of scan rates. The linearity of the cathodic and anodic peak currents vs. the ν1/2 suggests that electrochemical reactions on NTX–Au/ePADs are diffusion-controlled rather than adsorption-controlled.
Electrochemical behaviors of nitrite at the NTX–Au/ePAD
The electrochemical activity of ePADs for the oxidation of nitrite was investigated with CVs in the presence of various concentrations of a phosphate buffer solution (pH 6.0) in the potential range between 0.0 to 1.0 V at a scan rate of 50 mV s−1 (Fig. 4A,B). The Au/ePAD presented an oxidation peak for 1.0 mM nitrite in a 0.1 M phosphate buffer solution (pH 6.0) at 805 mV. After the modification of AuNPs with NTX, the oxidation peak current increased obviously (Ip had a 61% improvement) with a 120-mV negative shift at the peak potential (Fig. 4A). The significant improvement in the nitrite oxidation signals indicated that NTX–AuNPs could effectively enhance the electrochemical performance of the ePAD.
The oxidation of nitrite in a low concentration of the supporting electrolyte (0.001 M phosphate buffer, pH 6.0) was investigated (Fig. 4B). The oxidation current of 0.1 mM nitrite at the Au/ePAD sharply decreased as the supporting electrolyte concentration decreased at 0.001 M. Nevertheless, in the presence of NTX, the oxidation current did not change with the concentration of the supporting electrolyte (Fig. 4B). To investigate the transfer of nitrite ions to the electrode surface, the oxidation of 1.0 mM nitrite was recorded at various scan rates. Figure 4C shows the cyclic voltammograms of the NTX–Au/ePAD in the presence of 1.0 mM nitrite ions in a 0.1 M phosphate buffer solution (pH 6.0) at different scan rates. The linearity of the oxidation currents vs. the scan-rate1/2 suggested that the electrochemical oxidation of nitrite was controlled by diffusion from the solution to the surface of the working electrode (Inset a of Fig. 4C). Also, a linear correlation between currents vs. the scan-rate1/2 was observed in the 0.001 M PBS, which indicated that the oxidation of nitrite was controlled by diffusion from the solution to the surface of electrode (Inset b of Fig. 4C). Furthermore, the diffusion coefficient (D) for the nitrite ions at the NTX–Au/ePAD was calculated by the Randles–Sevcik equation32. The value of the diffusion coefficient for nitrite ions on the NTX–Au/ePAD was 1.44 × 10−5 cm2 s−1.
Based on the results, the presence of NTX on AuNPs leads to the acceleration of electron transfer kinetics with a low ionic strength. Under weakly supported conditions, the migration of ions leads to electroneutrality with changes in the concentration profiles of the ions on the electrode surface. In the typical concentration of the supporting electrolyte, the ions migrate in the electrolyte and perturb the electric field. Therefore, analytes with a low concentration can diffuse on the electric double layer23. While the electron transfer of nitrite ions in the absence of NTX decreases by this effect, the positively charged amino group on the NTX provides a potential gradient on the electrode surface, similar to the one obtained in a typical concentration of the supporting electrolyte33. Furthermore, through an electrostatic interaction, this positively charged NTX leads to the concentration of nitrite ions near the surface, as shown in Fig. 5.
In order to confirm this effect, the working electrode of the ePAD was modified with asparagusic acid (AGA) instead of NTX in the same experimental conditions (Fig. 4D). Asparagusic acid is a derivative of disulfide compounds with a carboxylic acid functional group. As it can be seen in Fig. 4D, there is no peak current for the oxidation of 1.0 mM nitrite with AGA–AuNPs. The peak current disappeared with the adsorption of AGA on AuNPs, indicating that the access of nitrite to the electrode surface was suppressed by AGA. This may be attributed to the electrostatic repulsion of the carboxylate group of AGA on the AuNPs surface (Fig. 5).
Optimization of the experimental conditions
The effects of the experimental conditions including pH, incubation time and concentration of NTX on the performance of the NTX–Au/ePAD were investigated. The corresponding details are presented in Fig. S5. The effect of pH on the electrocatalytic oxidation of 1.0 mM nitrite was investigated using the NTX–Au/ePAD at different pH values between 2 and 9. Based on the results, with an increase in the pH from 2.0 to 6.0, the peak currents increased and then decreased (Fig. S5). The maximum peak current occurred at pH 6. The decrease in the oxidation peaks at pH values lower than 6 was due to the decomposition of nitrite ions to NO3–, as reported in previous papers34,35. Besides, the sharp decrease in the currents at a pH value higher than 6 was due to the lack of protons36.
The effect of the incubation time of NTX on the oxidation of 1.0 mM nitrite was also investigated, and the optimum time of 7 min was obtained. The impact of pH on the NTX absorbed on the AuNPs was also evaluated. The oxidation peak currents of nitrite increased with an increase in the pH of NTX solutions. This may be due to the dependence of NTX ionization on pH, which leads to increased hydrophobicity of NTX at higher pH values, as well as the accelerating effect of the NTX adsorbed on the AuNPs. Therefore, pH 9 was used for the NTX solutions.
Calibration curve of nitrite
Figure 6A shows the DPV curves of the NTX–Au/ePAD in different concentrations of nitrite in a 0.1 M phosphate buffer solution (pH 6.0). The peak currents of the nitrite oxidation increased with an increase in the nitrite concentration. Besides, the calibration plot of the nitrite concentration versus the oxidation peak currents presented in Fig. 6B indicates a linear relationship between the peak currents and the nitrite concentration in a range from 0.05 to 1,400 μM with the linear regression equation of Ip (μA) = 0.1194 Cnitrite (μM) + 5.84. The limit of detection calculated from the calibration plot (3sb/slope) was found to be 20 ± 2 nM. The calibration plot in the weakly supported media (0.001 M PBS) was obtained too (Fig. 6C). The regression equation was Ip(μA) = 0.1173 Cnitrite(μM) + 5.44. The slopes of the calibration curves for nitrite in the normally supported media (0.1119 μA μM−1) and the weakly supported media (0.1173 μA μM−1) were virtually the same, which indicates that the determination of nitrite is possible in weakly supported media too. Table S3 summarizes the analytical performance of the proposed method based on an NTX–Au/ePAD with some electrochemical sensors for nitrite. The proposed NTX–Au/ePAD provides a sensitive technique for the determination of nitrite in real samples.
The selectivity, repeatability, stability and analytical applications of NTX–Au/ePADs
The selectivity of the NTX–Au/ePAD was performed in the presence of various interfering chemicals that may co-exist with nitrite in samples. A 20-fold concentration of NaCl, Na2CO3, MgCl2, CuSO4, FeCl2, FeCl3, Na2HPO4, CH3COONa and Na2CO3 and a tenfold concentration of cysteine, ethanol, methanol, uric acid, ascorbic acid, glucose and urea were added to a mixture of 0.1 mM nitrite. Then, the NTX–Au/ePAD was used for measurements. Based on the results, no relative signal change over ± 5% could be observed, which suggests that the NTX–Au/ePAD was not affected by the interfering chemicals that co-existed with nitrite in the samples.
The reproducibility of the method for the oxidation of 50 μM nitrite was investigated using twelve NTX–Au/ePADs. The relative standard deviation (RSD) was 4.6%, indicating the acceptable reproducibility of the NTX–Au/ePAD. Moreover, the RSD for three determinations of 50 µM nitrite using a single NTX–Au/ePAD was 2.6%, indicating the good repeatability of the device. In fact, a laser cutter provides a precise and rapid prototyped pattern which makes it easy to print the ink on the wax paper and to control the thickness and the surface area of the printed ink.
Furthermore, the NTX-Au/ePAD was used to measure nitrite in real samples. As presented in Table S4, the response obtained from the NTX-Au/ePAD was comparable with those obtained by the Griess protocol. T-tests indicated no significant differences between the NTX-Au/ePAD and the Griess method at a confidence level of 95%.