Electrochemical Detection of H2O2 using Bi2O3/Bi2O2Se Nanocomposites

1. Introduction

Electrochemical detection of hydrogen peroxide (H₂O₂) is a widely used technique in various fields, including analytical chemistry, biochemistry, and environmental monitoring. It relies on the principal of producing a measurable electrical signal due to oxidation and reduction of H₂O₂ at the surface of the electrode. This method is appealing and gaining attention in recent studies over the other traditional techniques like chromatography [1], colorimetry [2], fluorimetry [3] etc. due to its remarkable advantages such as high sensitivity, compatibility, low cost and real time analysis.

Electrochemical detection is usually carried out using enzymatic and non-enzymatic methods. However, enzymatic electrochemical sensors are expensive, less stable due to enzyme denaturation, shows poor reproducibility and require tedious purification process. Because of these drawbacks, the non-enzymatic H₂O₂ electrochemical sensors [4] are in demand as they offer high reproducibility, selectivity and sensitivity. Moreover, they are more stable, easy to use and inexpensive.

Nowadays, metal oxide nanoparticles are used for H₂O₂ sensors [5] which offers distinctive catalytic and electrical properties and excellent stability for the redox reaction of hydrogen peroxide. In this respect, bismuth-based materials like bismuth oxide [6,7], bismuth chalcogenide [8,9,10,11] bismuth oxychloride [12] etc. have gained significant interest in the field of energy storage, photocatalysis and electrochemical sensing applications. Another emerging bismuth containing nanomaterial which shows distinguished thermal, chemical and optoelectronic properties is Bi₂O₂Se. Bismuth oxyselenide is a two-dimensional layered material where [Bi₂O₂]²⁺ and Se²⁻ ions are held together by weak electrostatic forces [13,14].

Till date, various synthesis methods have been developed to produce Bi₂O₂Se nanosheets, and among these; chemical vapor deposition [15], pulsed laser deposition [16], one-pot hydrothermal method [17], and solution-phase method [18] are most popular. In most cases selenourea is utilized as the Se source. In this work, we synthesized Bi₂O₂Se using an approach developed by Chitara et al. [19,20] with slight modifications. Here, as a selenium source we used selenium powder in place of selenourea. This modified protocol provided an easy, scalable and cost-effective synthetic recipe to obtain Bi₂O₂Se nanosheets. The use of selenium powder ($13/g) instead of selenourea ($93/g) significantly reduces the cost of synthesis. Moreover, the toxicity of selenourea is also significantly higher. Our work also identifies the optimal conditions to grow pure phase Bi₂O₂Se nanosheets from Se powder under mild solution-processing conditions. Furthermore, we explore the efficacy of the intermediate Bi₂O₃/Bi₂O₂Se (or Bi₂O_xSe_y) nanocomposites (prepared by varying the synthesis time from 10 minutes to 7 days) towards H₂O₂ sensing.

2. Material and Methods

2.1. Materials and Reagents

Bismuth nitrate pentahydrate (Bi(NO₃)₃·5H₂O) (Thermo Scientific, USA), hydrazine hydrate (N₂H_4ꞏH₂O, Hydrazine, 64%) (Thermo Scientific, USA), Selenium powder from Alfa Aesar, Potassium hydroxide (Fischer Scientific, USA), Sodium hydroxide (Fischer Scientific, USA), ethylenediaminetetraacetic acid disodium salt (EDTA, Across Organics), Nafion solution (5% in lower alcohols, Sigma Aldrich), Isopropyl alcohol (Merck), De-ionized water (DI water, <18 Ω) were used for the experiments. H₂O₂ (30%, Fischer Scientific, USA) was used as analyte. Phosphate-buffered saline (PBS, pH 7.4, 1X molarity) was used as an electrolyte medium.

2.2. Synthesis

The Bi₂O_xSe_y nanomaterials were synthesized using a solution-phase method at room temperature. 1 g Bi(NO₃)₃·5H₂O was dissolved in 200 ml of deionized water and sonicated for 30 minutes. Then another solution containing 80 mg of selenium powder with 10 ml hydrazine hydrate was added into the bismuth nitrate solution. The solution turns black. In this mixture, 3 g of EDTA was added and the stirring was continued for another 10 minutes. Finally, 1.2 g of KOH and 3.2 g of NaOH was added. The solution was vigorously stirred for certain reaction time. The obtained precipitate was then centrifuged at 5000 rpm, washed with water and ethanol, and dried in an oven at 80 ⁰C. To obtain variable compositions of Bi₂O₃/Bi₂O₂Se (Bi₂O_xSe_y) nanocomposites, the final reaction time (post addition of KOH and NaOH) was varied viz. 10 minutes, 3 hours, 6 hours, 18 hours, and 7 days. The samples are denoted as BOSe-10 min, BOSe-3 h, BOSe-6 h, BOSe-18 h, and BOSe-7 days, respectively. The instrument details for characterization are provided in the supporting information file (ESI).

2.3. Modification of GCE and Electrochemical Sensing

For H₂O₂ sensing, we used a 3-electrode configuration with glassy carbon electrode (GCE, 3 mm diameter), platinum wire (Pt) as the counter electrode and Ag/AgCl (3.5 M KCl) as the reference electrode. The voltammetric cell was placed in a cell stand (C3, BAS Inc.) equipped with controlled stirring and Argon gas purging. Cyclic voltammetry (CV), and differential pulse voltammetry (DPV) signals were recorded to examine the sensing performance towards H₂O₂ and other analytes (uric acid, NaCl, ascorbic acid (AA) and dopamine).

For electrochemical analyses dispersed solutions of the nanocomposites were prepared by mixing 5 mg of the samples in isopropanol/DI water (9:1 ratio), along with 100 μL of Nafion solution as binder. This dispersion (2.5 μL) was then drop cast on polished GCE. For the electrochemical measurements, 15 mL of phosphate buffer saline (PBS 1X, pH 7.4) was taken in the voltammetric cell and purged with argon gas for 30 min, with stirring at 400 rpm. During the electrochemical measurements, the stirring and purging was momentarily paused to record the CV and DPV scans. The potential window was 0.6 to -0.9 V vs Ag/AgCl. The scan rate used was 10 mV/s. The pulse parameters for DPV, i.e., height, width, period and increment were 50 mV, 0.01 s, 0.1 s and 10 mV, respectively. The DPV scans were recorded without applying any inversion during the cathodic and anodic sweeps. The initial CV and DPV scans of the electrode, in blank electrolyte, i.e., without dopamine addition were performed, prior to sensing tests. 150 µL of H₂O₂ from 2 µM stock solution was added. The effective concentration of H₂O₂ in the electrolyte was 100 times lower. Purging and stirring were continued for 5 min for each addition of H₂O₂. The stock solutions of 2, 20, 200, 1000 and 10000 µM were used for H₂O₂ sensing. The effective concentration at each addition of H₂O₂ was calculated and used in the sensitivity (S) and limit of detection (LOD) evaluation.

For the selectivity tests, analytes of 10 mM concentrations (stock) were used. 150 µL of these stock solutions were sequentially added. The concentration of the interferant analytes (uric acid, NaCl, AA and dopamine) was purposefully kept high to study the efficacy of the sample towards H₂O₂ detection.

3. Results and Discussion

3.1. Structural and Morphological Characterization

The powder XRD patterns are shown in Figure 1. The standard data from Inorganic Crystal Structure Database (ICSD) are also provided for reference. The sample prepared with reaction time of 10 min, showed the formation of α-Bi₂O₃ (ICSD code: 2374), i.e., monoclinic structure with Space group P 121/c1. With increasing reaction time, the formation of Bi₂O₂Se-type phase is observed. The intensity of the Bi₂O₂Se component increases with longer reaction time and finally at 7 days, it completely transforms into single phase. The crystal structure of the Bi₂O₂Se is tetragonal (I 4/mmm space group, ICSD code: 2903) similar to that reported by Ghosh et al. [18]. For simplicity we refer to intermediate nanocomposite with varying Se content as Bi₂O_xSe_y.

The FE-SEM images at two different scales are shown in Figure 2. The BOSe-10 min sample has predominant micron-size rod-like structures with very fine agglomerates on the surface. The lack of sheet structure indicates absence of any Bi₂O₂Se. With increased reaction time (BOSe-3 and 6 h), the sheets appear to evolve from the surface of these rod-like structures. At around 18 h, the rods completely transform into agglomerated sheet-like structures. The sheets become well defined for BOSe-7 days, indicating complete transformation to Bi₂O₂Se. The elemental mapping and EDX analyses for the representative BOSe-6 h and BOSe-7 days are provided in Figures S2 and S3 and Tables S1 and S2 in the ESI. Clearly, the elemental analyses for BOSe-7 days sample shows good match with the expected stoichiometry of Bi₂O₂Se. The BOSe-6 h sample has much lower Se content, in line with inference from XRD. The XPS analysis for this sample is also provided as Figure S4 and Table S3 in ESI, which could possibly be more accurate than EDX.

The normalized UV-Vis-NIR diffuse reflectance spectra (UV-Vis-NIR DRS) of the samples are shown in Figure 3. The BOSe-7 days sample shows slightly different reflectance pattern (indicating a different phase) than others. The composite samples show higher absorption in the NIR region, but lower absorption in the UV-visible region. Nevertheless, the band edge is similar, leading to similar direct band gaps which are estimated using the Kubelka-Munk function [21,22]. Interestingly, from the Tauc plots we can observe a slightly higher area below the linear region, for BOSe- 6h which indicates slightly higher proportion of defects in this sample compared to BOSe-7 days. This is understandable considering more interfaces for the nanocomposite.

3.2. Growth Mechanism

Based on the evidence from XRD, SEM, EDX and XPS we can hypothesize the growth mechanism of the Bi₂O_xSe_y samples with this synthesis protocol. Equations (1)–(4), shown below, indicate the possible chemical reactions involved in the formation of Bi₂O₂Se nanosheets, similar to that suggested by Chitara et al. [19]. According to their mechanism, Bi(NO₃)₃ can undergo hydrolysis to produce BiONO₃. Furthermore, Selenium powder is reduced by hydrazine to produce Se²⁻ ions in highly alkaline medium. Finally, the self-assembly of the oppositely charged (Bi₂O₂)²⁺ and Se²⁻ layers results in the formation of Bi₂O₂Se nanosheets.

$$\mathrm{B}\mathrm{i}{\left(\mathrm{N}{\mathrm{O}}_3\right)}_3+{\mathrm{H}}_2\mathrm{O}\to \mathrm{B}\mathrm{i}\mathrm{O}\mathrm{N}{\mathrm{O}}_3+2{\mathrm{H}}^{+}+2\mathrm{N}{\mathrm{O}}_3^{-}$$

(1)

$$2\mathrm{B}\mathrm{i}\mathrm{O}\mathrm{N}{\mathrm{O}}_3+\mathrm{O}{\mathrm{H}}^{-}\to {\mathrm{B}\mathrm{i}}_2{\mathrm{O}}_3+{\mathrm{H}}^{+}+2\mathrm{N}{\mathrm{O}}_3^{-}$$

(2)

$$2\mathrm{S}\mathrm{e}+{\mathrm{N}}_2{\mathrm{H}}_4+4\mathrm{O}{\mathrm{H}}^{-}\to 2{\mathrm{S}\mathrm{e}}^{2-}+{\mathrm{N}}_2+4{\mathrm{H}}_2\mathrm{O}$$

(3)

Based on our evidence in this work, it appears that this mechanism may not be complete. We suggest that the availability of Se^2- ions could be a rate limiting factor and the replacement of the Se^2- ions may be much slower, unlike previously anticipated. Therefore, we can expect the following two outcomes:

• Large availability of Se^2-

  $$2\mathrm{B}\mathrm{i}\mathrm{O}\mathrm{N}{\mathrm{O}}_3+{\mathrm{S}\mathrm{e}}^{2-}\to {\mathrm{B}\mathrm{i}}_2{\mathrm{O}}_2\mathrm{S}\mathrm{e}+2\mathrm{N}{\mathrm{O}}_3^{-}$$

  (4)
• Low availability of Se^2-

  $$4\mathrm{B}\mathrm{i}\mathrm{O}\mathrm{N}{\mathrm{O}}_3+\mathrm{O}{\mathrm{H}}^{-}+{\mathrm{S}\mathrm{e}}^{2-}\to {\mathrm{B}\mathrm{i}}_2{\mathrm{O}}_2\mathrm{S}\mathrm{e}+{\mathrm{B}\mathrm{i}}_2{\mathrm{O}}_3+{\mathrm{H}}^{+}+4\mathrm{N}{\mathrm{O}}_3^{-}$$

  (5)

Another possible route could be through the formation of hydroxide or oxyhydroxide, as below:

$$2\mathrm{B}\mathrm{i}{\left(\mathrm{N}{\mathrm{O}}_3\right)}_3+3{\mathrm{H}}_2\mathrm{O}\to {\mathrm{B}\mathrm{i}}_2{\mathrm{O}}_3+6{\mathrm{H}}^{+}+6\mathrm{N}{\mathrm{O}}_3^{-}$$

(6)

$${\mathrm{B}\mathrm{i}}_2{\mathrm{O}}_3+{\mathrm{H}}_2\mathrm{O}\to 2\mathrm{B}\mathrm{i}{\left(\mathrm{O}\mathrm{H}\right)}_3$$

(7)

$${\mathrm{B}\mathrm{i}}_2{\mathrm{O}}_3+12{\mathrm{O}\mathrm{H}}^{-}\to 2\mathrm{B}\mathrm{i}{\left(\mathrm{O}\mathrm{O}\mathrm{H}\right)}_3+3{\mathrm{H}}_2\mathrm{O}$$

(8)

$$2\mathrm{B}\mathrm{i}{\left(\mathrm{O}\mathrm{H}\right)}_3+{\mathrm{S}\mathrm{e}}^{2-}\to {\mathrm{B}\mathrm{i}}_2{\mathrm{O}}_2\mathrm{S}\mathrm{e}+6{\mathrm{H}}^{+}+2{\mathrm{O}}_2$$

(9)

$$2\mathrm{B}\mathrm{i}{\left(\mathrm{O}\mathrm{O}\mathrm{H}\right)}_3+{\mathrm{S}\mathrm{e}}^{2-}\to {\mathrm{B}\mathrm{i}}_2{\mathrm{O}}_2\mathrm{S}\mathrm{e}+6{\mathrm{H}}^{+}+5{\mathrm{O}}_2$$

(10)

However, more work needs to be done to prove the mechanism, which will be undertaken in future.

3.3. Electrochemical Sensing Performance

The CV and DPV scans were performed to analyze the sensitivity by applying a potential from 0.6 to -0.9 V against Ag/AgCl electrode. 150 µL of hydrogen peroxide solution with different concentrations (ranging from 2-10000 µM stock) were then added into 15 mL of argon purged PBS (pH 7.4, 1X). The effective concentrations were calculated to range from 0.02 to 88.2 µM. In between each addition, stirring and purging was employed to help mixing and maintaining the inert atmosphere inside the cell. For each composite, before addition of H₂O₂ blank reading, i.e., without H₂O₂ was performed. Figure S1 (ESI) depicts the cyclic voltammograms of the different samples. The change in the redox peak patterns is clearly visible among the samples and also with increasing H₂O₂ additions. However, we refrain from in-depth evaluation of sensitivity (S) and LOD using CV as the peak patterns are complex with multiple peaks.

The DPV technique is more sensitive than CV [23] as it excludes the capacitive contribution. The changes in the current values become significant even at low concentrations. Moreover, in this work (Figure 4), the peak profile is highly simplified (a single peak in the potential window of 0 to -0.9 V) unlike CV. Therefore, for the evaluation of sensitivity (S) and limit of detection (LOD), we consider the DPV scans with this peak centered around -0.75 V. The LOD and S can be obtained from the plot of the peak current with respect to the concentration of H₂O₂ [23,24], as per the equations below.

$$\mathrm{L}\mathrm{O}\mathrm{D}=3.3\ast \left({\displaystyle \frac{\mathsf{\sigma }}{\mathrm{k}}}\right)$$

(11)

$$\mathrm{S}={\displaystyle \frac{\mathrm{k}}{\mathrm{A}}}$$

(12)

where, σ is the standard deviation of the response, k is the slope of the linear fits, and A is the area of the working electrode in cm².

The evaluated sensitivity and LOD values are given in Table 1. The BOSe-6 h sample shows the highest sensitivity of 152.7 µA µM^-1 cm^-2 and a low LOD of 1.62 µM. The linear range was found to be 0.02–15 µM. It is interesting to note that the sensitivity values are significantly less for the sample prepared in 10 min (final stirring post addition of KOH and NaOH) and pure phase Bi₂O₂Se. These results clearly indicate that the efficacy of the composite Bi₂O_xSe_y (in particular with reaction time of 6 h) is much higher as compared to the pure phase Bi₂O₂Se, towards H₂O₂ sensing.

For the selectivity test, we utilize the BOSe-6 h sample. CV and DPV scans were recorded with sequential addition of the high concentration of interfering analytes in the same electrolyte. As seen in Figure 5, the addition of uric acid, NaCl and Ascorbic acid shows negligible change in the current signal. However, we observe some change when dopamine is added. The DPV peak response for dopamine is usually around 0 V (a small signal is observed, Figure 5 (b)). The change is more significant with H₂O₂ addition (around -0.8 V), indicating a higher selectivity for H₂O₂ over dopamine.

4. Conclusion

Bi₂O_xSe_y nanocomposites with varying Se content were successfully synthesized using a solution-phase method by varying the reaction time. Morphological and structural characterization confirm the progressive changes occur in the nanocomposites with increasing reaction time. The evidence suggests transformation of α-Bi₂O₃ rods to Bi₂O₂Se nanosheets. Electrochemical studies performed using CV and DPV showed that Bi₂O₃/Bi₂O₂Se nanocomposites with an optimum reaction time of 6 hours, demonstrates remarkable sensitivity (152.7 µA µM^-1 cm^-2) and selectivity towards H₂O₂ and can be a potential non-enzymatic sensor.