Electrochemical Sensing of Hydrogen Peroxide Using Prussian Blue@poly(p-phenylenediamine) Coated Multi-walled Carbon Nanotubes

Multi-walled carbon nanotubes (MWCNTs, 95% purity, diameter 10–15 nm) were purchased from ILJIN Nanotech Co. (Korea). L-cysteine was purchased from Yakuri Pure Chemical Co. (Japan). Ammonium persulfate ((NH₄)₂S₂O₈, APS, ≥98.0%), p-phenylenediamine (C₆H₈N₂, pPD, ≥99.0%), potassium ferricyanide (K₃Fe(CN)₆, 99+%), potassium ferrocyanide (K₄Fe(CN)₆, 99%), iron(III) chloride (FeCl₃, 97%), potassium chloride (KCl), acetic acid (99.7 wt% in H₂O), hydrogen peroxide (H₂O₂, 30 wt% in H₂O), Nafion (~5% in a mixture of lower aliphatic alcohols and water), ascorbic acid, uric acid, and glucose were purchased from Sigma-Aldrich Co. (USA). All the chemicals were used without further purification. All the solutions were prepared with deionized water (DIW) obtained from an ultra-pure water purification system (Human Co., Korea) with a resistivity of not less than 18.2 MΩ cm. All the measurements were carried out at room temperature.

The surface morphology of composites was characterized by field emission transmission electron microscopy (FE-TEM, JEOL JEM-F200, Japan). TEM samples were prepared by placing a drop of the dispersion on a carbon-coated copper grid. Fourier transform infrared spectra (FT-IR) analyses were performed using a Perkin-Elmer Spectrum 2000 FT-IR spectrometer in the range from 400 to 4000 cm^-1 with a KBr pellet. Powder X-ray diffraction (XRD) analyses were performed using a New D8 advance diffractometer. Electrochemical impedance spectroscopy (EIS) measurements were performed using a CHI 760B electrochemical analyzer (CH Instrument, Inc., USA), and cyclic voltammetry (CV) and amperometry were performed using a CHI 842B electrochemical analyzer (CH Instrument, Inc., USA) with the conventional three-electrode system. A bare or modified glassy carbon electrode (GCE, d = 3.0 mm) was used as the working electrode. A platinum (Pt) electrode served as the auxiliary electrode, and a silver/silver chloride (Ag/AgCl) electrode filled with 3 M KCl served as the reference electrode. All ultrasonic cleaning was performed using a US-2510 Ultrasonic Cleaner (Branson, USA). Before each electrochemical measurement, solutions were thoroughly deoxygenated by bubbling N₂ gas through the solution for at least 10 min to remove the dissolved oxygen.

2.0 mg of Pristine MWCNTs were added to 17.6 mL of DIW and well dispersed by a 30 min sonication. To this solution, 2.0 mL of 0.1 M pPD and 0.4 mL of 0.1 M APS were added at room temperature under vigorous magnetic stirring for 24 hours. The dark purple solution was carefully filtered through a membrane filter and washed with DIW. The collected precipitate was dried overnight in an oven at 40℃. Finally, the product was dispersed in an acidic solution at pH 3.5.

1 mM of K₃Fe(CN)₆ was added to 25 mL of 0.1 M KCl solution and adjusted to pH 2.7. To this solution, 0.8 mL of dispersed MWCNTs@PpPD solution was added under magnetic stirring for 5 min. 25 mL of 1 mM FeCl₃ was added slowly to this mixture with continuous stirring at room temperature for 1 hour. Subsequently, 2 μL of 30% H₂O₂ was added to the mixed solution. The color of this solution gradually changed from yellowish green to dark blue. After 3 hours, the obtained dispersion was filtered, rinsed with DIW several times, and finally redispersed in 10 mL of DIW.

Before modification, the glassy carbon electrode was polished to acquire a mirror-like surface with fine alumina powders (particle size: 0.3 μm and 0.05 μm) on a polishing pad. The electrode was cleaned and sonicated with DIW for 2 min, in sequence. After drying under a gentle stream of N₂ gas, the modified electrode was prepared by dropping 8 μL of the MWCNTs@PpPD-PB dispersion onto the purified GCE surface and dried in an oven at 40℃ for 2 hours. Then, 2 μL of Nafion (0.5%) was placed on the surface of the modified electrode, and it was dried again in an oven at 40℃ for 1 hour.

Fig. 1 shows the synthetic route for MWCMT@PpPD-PB. At first, pristine MWCNTs were dispersed in a neutral solution through sonication. pPD monomers were added to the solution, and APS was further added resulting in a dark-violet solution. And then, the solution containing MWCNTs@PpPD nanocomposite was added to an acidic solution containing K₃Fe(CN)₆, and FeCl₃ was dripped very slowly. The snapshots of the products for each step are shown in Fig. S1.

Figure1.

Synthesis of MWCNTs@PpPD-PB nanocomposites.

Fig. 2 shows the XRD patterns and FE-TEM images of MWCNTs, MWCNTs@PpPD, and MWCNT@PpPD which suggests the morphologies and sizes of the as-synthesized nanocomposites. The peaks centered at 2θ values of 26.0° and 43.0° correspond to the (002) and (100) reflection planes of graphite from the MWCNTs, respectively.³⁸ The diffraction pattern of the MWCNTs@PpPD nanofibers shows five big sharp peaks (marked as asterisk marks) and eight minor peaks between 2θ = 8-30°.³⁹

Figure2.

XRD pattern (a) of MWCNTs and TEM images (b-c) of MWCNTs at different magnifications, XRD pattern (d) of MWCNTs@PpPD and TEM images (e-f) of MWCNTs@PpPD at different magnifications, XRD pattern (g) of MWCNTs@PpPD-PB and TEM images (h-i) MWCNTs@PpPD-PB at different magnifications.

Compared to the XRD pattern of the MWCNTs@PpPD, the peaks of MWCNTs@PpPD-PB are observed at 2θ values of 17.5°, 24.8°, 35.3°, 39.6° and 43.7° correspond to the (200), (220), (400), (420), and (422) reflection planes, respectively.⁴⁰ All the reflections can be indexed as a pure face-centered-cubic phase of Fe₄[Fe(CN)₆]₃, which demonstrates the successful synthesis of PB on the modified MWCNTs. Furthermore, the average diameter of the PBNPs was ~15.4 nm, calculated by the Scherrer formula.⁴¹ It is analogous to the diameter of the PBNPs of MWCNTs@PpPD nanocomposite determined from the TEM images, which were measured at 10.2-17.5 nm.

Fig. S2 shows the FT-IR spectra of MWCNTs, MWCNTs @PpPD, and MWCNTs@PpPD-PB which investigate the chemical components of the as-synthesized samples. After the polymerization of pPD on the surface of MWCNTs, the characteristic absorption peaks of PpPD were observed. The broad absorption peaks located at 3444 cm^-1 and the shoulder peak located at 3216 cm^-1 could be because of the N–H stretching vibration of the –NH group in MWCNTs @PpPD. Furthermore, in another broad adsorption peak assigned approximately from 1650 to 1300 cm^-1, the two strong peaks located at 1606 and 1545 cm^-1 can be attributed to the C=N and C=C stretching vibration of the phenazine ring structure, respectively.⁴² The other peak located at 1494 cm^-1 corresponds to the stretching of the benzene ring. The two small peaks at 1262 and 1208 cm^-1 are assigned to the imine C=N stretching vibration of the benzoid and quinoid imine units (–C=N–)⁴², proving that PpPD was successfully polymerized on MWCNTs. Compared to the MWCNTs@PpPD, the MWCNTs@PpPD-PB nanocomposite showed a strong absorption peak at 2083 cm^-1 which is attributed to the C≡N stretching absorption band in the [Fe²⁺–CN–Fe³⁺] structure of PB. An absorption peak at 502 cm^-1 was also observed, because of the formation of [Fe²⁺–CN–Fe³⁺] structure.

The inset in Fig. 3 is the Randles equivalent circuit model used for fitting the measurement data, where R_s for electrolyte resistance, C_dl for double-layer capacitance, R_ct for electron transfer resistance, and Z_w for Warburg impedance. R_ct represents the electron-transfer-limited process at the electrode/electrolyte interface and is equal to the semicircle diameter in the Nyquist diagrams. It is measured in a background solution, containing 5 mM [Fe(CN)₆]^3- and 5 mM [Fe(CN)₆]^4- in 0.1 M KCl at the frequency range from 10 mHz to 100 kHz at 0.19 V (vs. Ag/AgCl).

Fig. 3 shows the impedance spectra represented as the Nyquist plots (Z′ vs Z") for (a) bare GCE, (b) GCE/PB, (c) GCE/MWCNTs, (d) GCE/MWCNTs@PpPD, and (e) GCE/MWCNTs@PpPD-PB. R_ct of the bare GCE was 127.14 Ω with an almost straight tail line. When MWCNTs were cast on the GCE, R_ct decreased to 42.66 Ω, indicating that the MWCNTs could prompt the charge transfer. Furthermore, when the GCE electrode surface was coated with PBNPs, R_ct value increased considerably to 421.79 Ω, indicating that the PBNPs on the GCE obstructed the charge transfer of the electrochemical process at the electrode surface. After the MWCNTs@PpPD was dropped over the GCE, the diameter of the high-frequency semicircle was significantly reduced to 9.06 Ω as a R_ct value. R_ct value of MWCNTs@PpPD-PB was 42.43 Ω, which is greater than that of the MWCNTs@PpPD, indicating that PB interrupted the charge transfer. Note that GCE/MWCNTs@PpPD-PB showed a steeper Warburg slope at low-frequency limit compared to others. This is due to the increased significance of chargetransfer resistance and double-layer capacitance as the frequency rises.⁴³

Figure3.

Nyquist plots of (a) bare GCE, (b) GCE/PB, (c) GCE/MWCNTs, (d) GCE/MWCNTs@PpPD and (e) GCE/MWCNTs@PpPD-PB in 0.1 M KCl solution containing 5 mM [Fe(CN)6]^3-and 5 mM [Fe(CN)₆]^4-.

Fig. 4(a) shows the CVs of the bare GCE, GCE/MWCNTs, GCE/MWCNTs@PpPD GCE/MWCNTs-PB, and GCE/MWCNTs@PpPD-PB in the absence of H₂O₂. The redox peaks were observed in the CVs of MWCNTs-PB and MWCNTs@PpPD-PB modified GCE, whereas no redox pair peaks were observed for the bare GCE, GCE/MWCNTs and GCE/MWCNTs@PpPD. Between the observed redox peaks, the peak current of MWCNTs@PpPD-PB modified GCE was much larger than that of MWCNTs-PB modified GCE, probably because of the high electrical conductivity of the coated pPD polymer on the MWCNTs. After adding 5 mM H₂O₂, the CVs of the bare GCE, GCE/MWCNTs, and GCE/MWCNTs@PpPD did not change (Fig. 4(b)). Meanwhile, the CVs of the MWCNTs-PB and MWCNTs @PpPD-PB modified GCE show that the reduction peak significantly increased, but the oxidation peak decreased slightly. These results indicate the presence of PBNPs on the modified GCE.

Figure4.

CVs of (A) bare GCE, (B) GCE/MWCNTs, (C) GCE/MWCNTs@PpPD, (D) GCE/MWCNTs-PB, and (E) GCE/MWCNTs @PpPD-PB in the (a) absence of H₂O₂ and (b) present of H₂O₂ from -0.3 to 0.6 V in 0.1 M KCl solution (pH 2.7) at scan rate of 50 mV/s.

Fig. 5 shows the CVs of GCE/MWCNTs-PB and GCE/MWCNTs@PpPD-PB in N₂-saturated 0.1 M KCl (pH 2.7) at different concentrations of H₂O₂. The cathodic peak current value increases with the increasing concentration of H₂O₂ because of the PB. Moreover, Fig. S3 shows that the reduction current of the MWCNTs@PpPD-PB-modified GCE is about twice as high as that of the MWCNTs-PB-modified GCE.

Figure5.

CVs of (a) GCE/MWCNTs-PB and (b) GCE/MWCNTs@PpPD-PB in N₂-saturated 0.1M KCl aqueous solution (pH 2.7) in the presence of H₂O₂ with different concentrations (from (A) to (E) : 1, 2, 3, 4, and 5 mM) at a scan rate of 50 mV/s.

Fig. 6 shows the correlation between the current and the scan rate. The electrochemical effect of the scan rate on the electrocatalytic H₂O₂ cathodic current was investigated to evaluate the kinetics of the electrochemical performance of the as-synthesized MWCNTs nanocomposites. The cathodic peak currents increased linearly in proportion to the square root of the scan rate with a correlation coefficient of 0.9978 as the scan rate was increased from 10 to 200 mV/s (Fig. S4). This result signifies that the process of electrocatalytic reaction is diffusion-controlled.

Figure6.

CVs of MWCNTs@PpPD-PB nanocomposite modified glassy carbon electrode in N₂-saturated 0.1M KCl aqueous solution (pH 2.7) in the presence of 1 mM H₂O₂ at different scan rates (from (a) to (e) : 10, 20, 50, 100, 200 mV/s).

Fig. 7 shows the comparative amperometric response of GCE/MWCNTs-PB and GCE/MWCNTs@PpPD-PB for successive sensing of H₂O₂ into stirred N₂-saturated 0.1 M KCl (pH 2.7) aqueous solution at an applied potential of +0.15 V (vs. Ag/AgCl). The reduction current value showed a rapid response to the subsequent addition of H₂O₂ and quickly reached the 95% steady-state current, which was less than 2 sec. In the MWCNTs-PB modified GCE and the MWCNTs@PpPD-PB modified GCE, the current for both increased with the addition of H₂O₂, and cathodic current of the latter was greater than the former.

Figure7.

Amperometric response of the (a) MWCNTs-PB and (b) MWCNTs@PpPD-PB nanocomposite modified glassy carbon electrode to successive injection of H₂O₂ in 0.1 M KCl aqueous solution (pH 2.7) under stirring. Applied potential: +0.15 V.

The amperometric response was unstable with the addition of H₂O₂ probably because of the weak interaction between the MWCNTs and the PBNPs. To solve this problem, the MWCNTs were effectively coated by PpPD and strongly attached to the PBNPs. Therefore, the modified electrode composed of the MWCNTs, PpPD, and PB showed a better response to the addition of H₂O₂.

Fig. S5 shows the calibration curve of the reduction peak currents versus the concentration of H₂O₂, which presents the linear response region of MWCNTs-PB and MWCNTs @PpPD-PB. MWCNTs@PpPD-PB shows a linear response range from 0.005 mM to 2.225 mM with a correlation coefficient of 0.9992, and the sensitivity of the sensor was 583.6 μAmM^-1cm^-2, which was much higher than that of a previous study estimated from GCE/MWCNTs-PB and GCE/MWCNTs/Ppy/PB.^44-45 Moreover, the limit of detection of GCE/MWCNTs@PpPD-PB was calculated as 0.95 μM with a signal-to-noise ratio of 3.

Table 1 summarizes the electrochemical characteristics of various modified electrodes containing PB for sensing H₂O₂, indicating that GCE/MWCNTs@PpPD-PB has a high sensitivity and a wide linear range in comparison to those of other reported modified electrodes, proving improved electrical conductivity and good stability.

An amperometric measurement was performed to assess the selectivity of the as-synthesized sensor. Fig. 8 shows that the current generated in the presence of 0.1 mM of interfering substances, such as citric acid, ascorbic acid, and L-cysteine. These interfering substances could be ignored relative to the current attributed solely to 0.1 mM H₂O₂ in N₂-saturated 0.1 M KCl solution (pH 2.7) at an applied potential of 0.15 V, indicating that this sensor possessed an excellent anti-interference capacity.

Figure8.

Interference-free behavior of the MWCNTs@PpPD-PB nanocomposite modified GCE by the amperometric trace recorded in N₂-saturated 0.1 M KCl solution (pH 2.7) under stirring. Injection of 0.1 mM H₂O₂, citric acid (CA), ascorbic acid (AA), and L-Cysteine (L-Cys). Applied potential: +0.15 V.

The reproducibility was investigated by the CV experiments on five modified electrodes under the same condition, showing excellent performance as evidenced by a relative standard deviation (R.S.D) of 2.5%. Fig. S6 shows the stability examined by monitoring the current response in 1 mM H₂O₂ solution after effectively cycling the modified electrode. After 50 cycles, the reduction of the electrode signal was only 2.6%, and after 50 additional cycles, the loss of signal reached only 2.9%. To determine the life span of the prepared sensor, the as-synthesized electrodes were stored in a refrigerator at 4℃, and the electrode signal decreased to only 97.1% of its original peak current in 1 mM H₂O₂ after 20 days. In brief, MWCNTs@PpPD-PB on the GCE showed excellent selectivity, reproducibility, and stability. The current response becomes stable in less than 2 sec, which indicates the rapid response of the MWCNTs@PpPD-PB on the GCE towards H₂O₂.

The performance of MWCNTs@PpPD-PB nanocomposites on the GCE was analyzed by applying it to the determination of H₂O₂ present in the household disinfectant. Using amperometry measurement to obtain the response of the household disinfectant, the experiment was conducted under the same conditions as the conventional amperometry experimental conditions. The exact H₂O₂ concentration in the household disinfectant was measured using the standard addition method.

The concentrations obtained by stepwise increasing the concentration and the R.S.D are listed in Table 2. The recoveries of H₂O₂ were found to be in the range of 95-100%, indicating that the modified electrode is suitable for real sample analyses.