Enhancing Photocatalytic Activity with the Substantial Optical Absorption of Bi₂S₃-SiO₂-TiO₂/TiO₂ Nanotube Arrays for Azo Dye Wastewater Treatment

Abstract

One-dimensional TiO₂ nanotube arrays (TNAs) were vertically aligned and obtained via the electrochemical anodization method. In this study, Bi₂S₃-TiO₂-SiO₂/TNA heterojunction photocatalysts were successfully prepared with different amounts of Bismuth(III) sulfide (Bi₂S₃) loading on the TNAs by the successive ionic layer adsorption and reaction (SILAR) method and characterized by X-ray diffraction (XRD) patterns, field-emission scanning electron microscope–energy-dispersive spectroscopy (FESEM-EDS), Fourier transform infrared (FTIR) spectra, ultraviolet-visible diffuse reflectance spectra (UV–Vis/DRS), and electrochemical impedance spectroscopy (EIS) techniques. The photocatalytic performances of the samples were investigated by degrading Basic Yellow 28 (BY 28) under visible-light irradiation. Optimization of the condition using the response surface methodology (RSM) and central composite rotatable design (CCRD) technique resulted in the degradation of BY 28 dye, showing that the catalyst with 9.6 mg/cm² (designated as Bi₂S_3(9.6)-TiO₂-SiO₂/TNA) showed the maximum yield in the degradation process. The crystallite size of about 17.03 nm was estimated using the Williamson–Hall method. The band gap energies of TiO₂-SiO₂/TNA and Bi₂S_3(9.6)-TiO₂-SiO₂/TNA were determined at 3.27 and 1.87 eV for the direct electronic transitions, respectively. The EIS of the ternary system exhibited the smallest arc diameter, indicating an accelerated charge transfer rate that favors photocatalytic activity.

1. Introduction

Nowadays, due to the rapid development of various industries and population growth, environmental pollution has become one of the most substantial challenges of human society [1]. Utilizing solar energy, a clean, powerful, inexhaustible, affordable, and renewable energy source, to eliminate pollutants is regarded as the best solution to critical problems [2]. Wastewater colored with synthetic organic dye that contains nitrogen as the azo group (N=N) from various industries, especially the paper and textile industries, is essentially soluble in water and is considered the main source of pollution, causing significant environmental issues. Consequently, in today’s world, it appears to be essential to eliminate residual dyes in industrial effluents. Within the different groups of industrial wastewater dyes, basic dyes are visible at a low concentration and exhibit pronounced, vigorous color intensity [3,4]. BY28, a cationic dye that is hazardous to the environment, is frequently used for polyacrylonitrile and to color cotton [3].

Photocatalysts have been developed and used in environmental areas for industrial wastewater elimination [5]. The semiconductor titanium dioxide (TiO₂) is widely used, being generally applied in heterogeneous photocatalysis, to remove pollution wastewater. However, due to its photocatalytic properties and inherent defects, the practical applications of TiO₂ are significantly limited [6]. The rapid recombination of photogenerated electron and hole pairs leads to a lower rate of desired chemical transformations due to the absorbed light energy and low quantum efficiency [7,8,9]. It is worth noting that photocatalysts play an essential role in photocatalytic processes that regulate the efficiency of sunlight utilization and photocatalytic ability. Creating hybrid heterojunction designs and manufacturing photocatalysts based on semiconductors to increase photocatalytic activity have attracted much attention [10,11,12]. In a conventional type II heterojunction system, the electrons generated by light in the conduction band (CB) of a semiconductor with high reduction potential (i.e., more negative) can be transferred to the CB of a semiconductor with low reduction potential (i.e., less negative) due to more negative CB edge potentials. However, the resulting photoexcited electrons and holes are less active, in accordance with the lower oxidation capacity [13]. In a direct Z-scheme photocatalytic system, the photogenerated electrons on the CB of a semiconductor which has low reduction potential (i.e., less negative) will recombine with the photogenerated holes on the valance band (VB) of a semiconductor which has low oxidation potential (i.e., less positive), giving photogenerated electrons and holes high reduction and oxidation abilities for participating in photocatalytic redox reactions [9,11]. Notably, Z-scheme photocatalysts, compared to conventional heterojunction-type photocatalysts, have been designed to improve the separation efficiency of photogenerated charge carriers. They could play a critical role in improving charge transfer, oxidation potential, and photocatalytic efficiency [11]. This structure enables electrons and holes to collect on different semiconductors, extending the lifetime of the charge carriers and utilizing a broader range of light wavelengths, which ultimately increases light harvesting efficiency. For this purpose, Bi₂S₃ (band gap energy (E_g) of 1.27 eV)-based Z-scheme heterojunction semiconductors with more negative conduction edge potentials have gained attention for the production of TiO₂-based Z-scheme photocatalysts [13], due to their high conduction potential, excellent photoactivity in visible light, and photosensitization, which thoroughly meets the requirements for the construction of photocatalysts based on Z-schematic heterojunctions [14,15,16,17].

For the photodegradation of organic dyes, researchers have resorted to constructing heterojunction systems as an emerging technology. According to the energy band matching principle, these systems are suitable for the fabrication of heterojunction photocatalysis in Bi₂S₃, which is excited in the visible region with a narrow bandgap energy, and semiconductors, like TiO₂, which are activated in the ultraviolet region, which promotes the spatial separation of the excited electron/hole pairs [1,17]. Bai [18] focused on the solvothermal synthesis of a nanocomposite of Bi₂S₃ and TiO₂ that had improved visible-light photocatalytic activity. Arjmand et al. [19] investigated the photocatalytic activity of TiO₂/CNT/BiOBr/Bi₂S₃ in the degradation of Acid black 1721 under visible-light irradiation. As a result, complete decolorization was achieved at an initial pH of 3, a dye concentration of 30 mg/L, a reaction time of 40 min, and a catalyst dosage of 1 g/L, while COD removal was 99% after 120 min. Lin et al. [20] reported an improved photocatalytic performance of a Bi₂S₃-TiO₂ nanotube array heterostructure under visible-light irradiation by sensitizing a wide-band-gap semiconductor with a narrow-band-gap semiconductor. Sun et al. [21] prepared Bi₂S₃/BiVO₄/TiO₂ heterostructure photocatalysts using continuous hydrothermal methods, and the photocatalytic activity of the sample was investigated by degrading methyl orange under visible-light irradiation. Compared to TiO₂ nanorod arrays, Bi₂S₃/BiVO₄/TiO₂ photocatalysts extend the photocatalytic degradation efficiency of Bi₂S₃/BiVO₄/TiO₂ during the catalytic decomposition of methyl orange to 76.3%, which is approximately 4 times higher than that of bare TiO₂ nanorod arrays. Drmosh et al. [22] synthesized a Bi₂S₃/MoS₂/TiO₂ heterostructure using a facile microwave-assisted hydrothermal method. The photocatalytic performance of the as-prepared nanocomposite was evaluated by monitoring the photocatalytic degradation of methylene blue under sunlight. The photocatalytic performance of Bi₂S₃/MoS₂/TiO₂ is much better than that of TiO₂, MoS₂, or Bi₂S₃. In this work, we investigated the performance of Bi₂S₃-SiO₂-TiO₂/TiO₂ nanotube array heterojunction photocatalysts in degrading BY 28 under visible-light irradiation. The enhanced performance was mainly attributed to the synergistic effect of Bi₂S₃ and SiO₂, which improved the dispersion of TNAs [2]. The introduction of SiO₂ into the TiO₂ system significantly enhances the stability of photocatalysts and the dispersivity. The supporting catalysts, using SiO₂-TiO₂ compared to single-component oxide supports, possess the benefits of significantly larger surface areas, enhanced acidic sites that improve redox abilities, and better degradation of pollutants [8]. Meanwhile, TNAs, which were vertically formed on the titanium (Ti) substrate by electrochemical anodization, were used as scaffolds to construct Bi₂S₃ and TiO₂ nanoparticles to increase the photocatalytic performance of the photocatalyst. TNAs with smooth surfaces possess a one-dimensional (1D) geometry that facilitates charge carrier transfer. Also, paths along the TNA walls reduced charge carrier diffusion, thereby decreasing the occurrence of charge losses [6,7]. In addition, TNAs have shown better photocatalytic efficiency than other morphologies, possessing excellent properties, such as a high orientation; a large internal surface area, preventing the recombination of photogenerated electrons and holes under irradiation; and a short diffusion length to the conductive substrate for charge carrier transfer [8,9]. On the other hand, Bi₂S₃, which is advantageous for increasing efficient visible-light absorption, photogenerated carriers, and photocatalytic performance, also enhances electron–hole separation efficiency and a reduction in the recombination speed of electron–hole pairs, thereby improving the efficient capture of photogenerated charge carriers [23].

In this research, we report a novel photocatalyst, Bi₂S₃-TiO₂-SiO₂/TNA, developed using a successive ion layer adsorption and reaction (SILAR) technique to enhance the photocatalytic performance of the photocatalyst. Scheme 1 shows the experimental setup and strategy adopted for the preparation of Bi₂S₃-TiO₂-SiO₂/TNA. The samples were characterized, and their photocatalytic efficiency was tested for the degradation of BY 28 dye solution as a model of textile waste under visible light. Subsequently, the RSM-CCRD approach was employed to evaluate the interactions between the influencing experimental variables and to reduce the number of experiments needed to achieve the optimal photodegradation response. The effects of Bi₂S₃ loading on the TNAs, initial concentration of dye, and pH were investigated.

2. Experimental Section

2.1. Materials

The materials used in this study were of reagent grade, such as tetrabutylorthotitanate (Ti (OC₄H₉)₄, 98%), acetyl acetone (C₅H₈O₂, 99%), hydrochloric acid (HCl, 37%), ethanol (C₂H₅OH, 99%), tetraethylorthosilicate (TEOS, 99%), ammonium fluoride (NH₄F, 99%), ethylene glycol (C₂H₆O₂, 99%), and bismuth nitrate (Bi(NO₃)₃.₅H₂O, 99%). Titanium (Ti) plates (99.5% pure) were applied as a base metal with the dimensions of 1 × 4 cm and a thickness of 1 mm. All chemical reagents used in this work were purchased from Sigma–Aldrich (St. Louis, MO, USA) and Merck (Darmstadt, Germany).

2.2. Preparation of TNAs

A traditional electrochemical anodization process at a constant voltage was used to prepare the TNAs [20,24]. Firstly, Ti foils with 100- and 600-grit sandpaper were thoroughly polished, washed successively using acetone, isopropyl alcohol, and ethanol in ultrasound for 10 min, and dried in ambient air. Anodization experiments were carried out in an electrolyte bath consisting of 0.5 wt.% NH₄F and 2.5 vol.% DI water in ethylene glycol [24]. The anodization was performed at a constant voltage of 60 V for 1 h, and the gap between the anode and cathode was 10 mm, while a Ti plate was chosen as the anode and platinum was used as the cathode. After the electrochemical process, the as-prepared samples were rinsed several times in DI water and then crystallized the amorphous nanotubes in the anatase phase and annealed in a furnace at a temperature of 500 °C, maintained for 2 h.

2.3. Fabrication of Bi₂S₃-TiO₂-SiO₂/TNA

TiO₂-SiO₂ was synthesized with the sol–gel technique, which is similar to previous reports. The details are as follows [25,26]. Firstly, a mixture of 10 mL of ethanol, 2.5 mL of tetrabutylorthotitanate, 2.5 mL of tetraethylorthosilicate, and 2.5 mL of acetylacetone was mixed at room temperature (23 °C) and stirred for 45 min; then, 2 mL of DI water was added to the reaction mixture, followed by stirring for 10 min. Subsequently, under continuous stirring for 2 h, the pH of the suspension was adjusted to 1.8. Finally, the mixture was aged at room temperature for 24 h.

TiO₂-SiO₂/TNA prepared by the immersion dip-coating method [27]. For this purpose, the TNAs were transferred into the TiO₂-SiO₂ sol solution, removed after 15 min, and dried in ambient air. The procedure of dipping, drying, and heating was repeated 3 times, for 30 s each time, and the covered TNAs were dried at 60 °C, for 2 h. Subsequently, the film was annealed by a furnace at a temperature of 500 °C for 2 h.

Bi₂S₃-TiO₂-SiO₂/TNA photocatalysts with different amounts of Bi₂S₃ were assembled uniformly onto the TiO₂-SiO₂/TNA using the SILAR technique in Bi₂S₃ solution with various concentrations (0.13, 0.20, 0.30, 0.40, 0.46 M). Briefly, each deposition cycle consisted of four distinct steps: (i) immersing the TiO₂-SiO₂/TNA in a solution containing the Bi³⁺ precursor for 5 min; (ii) rinsing the film in ethylene glycol for 1 min to eliminate the loosely bounded ions and drying it in ambient air; (iii) immersing the film in the S²⁻ precursor for 5 min for the adsorption of sulfur ions that will react with pre-adsorbed bismuth ions; and (iv) rinsing the substrate again in ethylene glycol for 1 min to eliminate the loosely bound ions and then drying it at ambient temperature. To investigate the growth of Bi2S3 thin films, each procedure was repeated three times. Eventually, the films were dried at 80 °C for 3 h [28]. Scheme 1 illustrates the experimental setup and depiction outlining the SILAR technique for the preparation of Bi₂S₃-TiO₂-SiO₂/TNA.

2.4. Instruments

The X-ray diffraction (XRD) patterns were obtained using an XRD analytical diffractometer (X’PertPro, using a CuKα radiation source with λ = 1.54 Å; Ni-filtered, the Netherlands). The surface morphologies of the prepared samples were imaged using a field-emission scanning electron microscope (FESEM, INSPECT F50 FEI). To reveal the compositions of Bi2S3, TiO₂, and SiO₂ nanoparticles doped into the TNA surface, energy-dispersive spectroscopy (EDS) was used. The ultraviolet-visible diffuse reflectance spectra (UV–Vis/DRS) were recorded using a UV–vis spectrophotometer (Shimadzu UV-2450). A Fourier transform infrared (FTIR) spectrum was performed using a Nicolet Model Nexus 470 FTIR spectrometer in the range of 400–4000 cm⁻¹. Electrochemical impedance spectroscopy (EIS) experiments (AUTOLAB PGSTAT302N/FRA2) to investigate the photoelectrochemical behaviors of photocatalysts were conducted in three-electrode compartments in a frequency range from 100 kHz to 10 mHz at OCP with an amplitude signal of 10 mV. The electrolyte was 0.2 M Na₂SO₄ solution, with a saturated calomel electrode, platinum foil, and the as-prepared samples serving as the reference, counter, and working electrodes, respectively.

2.5. Photocatalytic Performance Tests

The decomposition of the azo dye BY 28 by the prepared photocatalysts was evaluated in the presence of an Osram lamp (HQI-BT 400 W/D/DE40FLH1, 400 W). The wavelength range for the lamp was 380 to 750 nm. The distance between the solution and the light source was set to 20 cm. For the photodegradation step, the Bi₂S₃-TiO₂-SiO₂/TNA photocatalyst (with a surface area of 1 × 4 cm²) was vertically immersed in 25 mL of BY 28 (22.5 mg/L) aqueous solution and separated by a distance of 20 cm from the light source. Photocatalytic tests were performed with a magnetic stirrer for 30 min in the dark before the test to evaluate the adsorption/desorption equilibrium. The concentration of the degraded solution was determined by taking samples every 30 min and then examined using a UV–Vis spectrophotometer over 180 min. The maximum absorbance (λ_max) of BY 28 was at the wavelength of 453 nm [29]. The change in the intensity of the absorbance peak of BY 28 during the photocatalytic process was recorded. The photocatalytic efficiency of the azo dye solution as a percentage was determined according to the following equation [16,17,30]:

$$\mathrm{B}\mathrm{Y} 28 \mathrm{p}\mathrm{h}\mathrm{o}\mathrm{t}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y} \left(\%\right)= {\displaystyle \frac{C_0-C}{C_0}} \times 100$$

(1)

where C₀ is the concentration of BY 28 at 0 min after 30 min of adsorption–desorption equilibrium and C is the BY 28 concentration at time t. The first-order rate constant (kr) was determined from the slope of the fitted line using the Langmuir–Hinshelwood kinetic equation, as below [14,19]:

$$-\mathrm{l}\mathrm{n}{\displaystyle \frac{C}{{\mathrm{C}}_0}}= {\mathrm{K}}_{\mathrm{r}}\mathrm{t}$$

(2)

2.6. Experimental Design and Statistical Analysis

The RSM model is a reliable statistical tool with important application in the optimization and analysis of the significance of various operational parameters, as it reduces the number of required experiments and highlights the interactions between process variables [31]. Minitab 16 software and a CCRD approach were used for the analysis of variance (ANOVA) of the regression parameters of the predicted response surface models based on CCRD-RSM to determine the interaction effects of three independent parameters, such as the amount of Bi₂S₃ loaded on the TNA, dye concentration, and pH, which have evident effects on the photocatalytic activity. These parameters and their levels are presented in

Table 1. Experimental variables and their levels in CCRD.

Factors	Levels
	Low	Central (0)	High
	−1	0	+1
X1: [Bi3+] (M)	0.30	0.40	0.50
X2: [BY 28] (mg/L)	10	20	30
X3: pH	4	6	8

.

The following quadratic polynomial model employed a correlation between the operating parameters and the photodegradation efficiency [32].

$$\mathrm{Y}(\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{l}\mathrm{o}\mathrm{r}\mathrm{i}\mathrm{z}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}\left(\%\right))={\mathsf{\beta }}_0+\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathsf{\beta }}_{\mathrm{i}}{\mathrm{X}}_{\mathrm{i}}+\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathsf{\beta }}_{\mathrm{i}\mathrm{i}}{\mathrm{X}}_{\mathrm{i}}^2+\sum _{\mathrm{i}=1}^{\mathrm{n}}\sum _{\mathrm{j}=1}^{\mathrm{n}}{\mathsf{\beta }}_{\mathrm{i}\mathrm{j}}{\mathrm{X}}_{\mathrm{i}}{\mathrm{X}}_{\mathrm{j}}$$

(3)

where Y is the response model, β₀ is the constant of the model, X_i and X_j are the coded values of the independent variable, β_i, β_ii, and β_ij correspond to the linear, quadratic, and cubic coefficients, respectively, and n is the number of independent parameters [33].

3. Results and Discussion

3.1. Characterization of Prepared Films

3.1.1. Structural Properties

The TiO2-SiO2/TNA and Bi₂S_3(9.6)-TiO₂-SiO₂/TNA films were characterized by X-ray diffraction patterns to determine the phase structure and composition of the samples. As demonstrated in Figure 1A, two of the photocatalyst films had significant 2θ peaks at 25.70°, 37.72°, 38.61°, 48.54°, 53.17°, and 62.80°, which correspond to (101), (103), (112), (105), (200), and (105), which matches well with the anatase phase of TiO₂ (JCPDS 21–1272) [9,12]. It can be seen that the XRD pattern of TiO₂-SiO₂/TNA and Bi₂S_3(9.6)-TiO₂-SiO₂/TNA films had the characteristic (101) crystal plane of the anatase type of TiO₂ [1]. After the deposition of Bi₂S₃ by the SILAR method, a new diffraction peak located at 35.54° appeared in Bi₂S_3(9.6)-TiO₂-SiO₂/TNA, which could correspond to the (240) lattice plane of Bi₂S₃ (JCPDS 84-0279) [34]. No signal was observed corresponding to SiO₂ on Bi₂S_3(9.6)-TiO₂-SiO₂/TNA, consistent with previously reported literature [26,27,28]. This may be attributable to the amorphous nature of the synthesized SiO₂ and the incorporation of Si within the crystalline lattice of TiO₂ as an interstitial atom [35]. According to the Scherrer formula and Williamson–Hall (W-H) method, the crystallite size (D) of the sample could be analyzed from the X-ray line broadening of the anatase (1 0 1) reflection of the samples using Equation (4) [34,35]:

$$D={\displaystyle \frac{(k \lambda )}{\beta cos\theta }}$$

(4)

where k, β, λ, and θ denote a constant (k = 0.9), the full width at the half maximum (FWHM) in radians, the wavelength of the X-ray (λ = 0.15406 nm), and the Bragg angle of reflection in radians, respectively. The Williamson–Hall method is a simplified integral width method that separates the contributions of diffraction line broadening, according to the following formula, Equation (5) [15,36];

$${\mathsf{\beta }}_{\mathrm{h}\mathrm{k}\mathrm{l}}\mathrm{c}\mathrm{o}\mathrm{s}{\mathsf{\theta }}_{\mathrm{h}\mathrm{k}\mathrm{l}}= {\displaystyle \frac{\mathrm{k}\mathsf{\lambda }}{\mathrm{D}}}+\mathsf{\eta } \mathrm{s}\mathrm{i}\mathrm{n}{\mathsf{\theta }}_{\mathrm{h}\mathrm{k}\mathrm{l}}$$

(5)

β_hkl is the instrument-corrected FWHM, k is a constant (k = 0.9), D is crystallite size, λ is the wavelength of the X-ray source, and η is the strain of the lattice [5,9,23]. By plotting 4sinθ values along the X-axis against βcosθ on the Y-axis and X-axis, respectively, the plot demonstrates the crystallite size of the sample according to the equation (Figure 1B). By applying the Scherer formula and the Williamson–Hall model to the diffraction patterns, the average particle sizes of 34.41 nm and 17.03 nm, respectively, were calculated for the Bi₂S_3(9.62)-TiO₂-SiO₂/TNA sample.

3.1.2. FT-IR Spectra

The chemical structures of Bi₂S_3(9.6)-TiO₂-SiO₂/TNA were characterized by the FT-IR spectrum (Figure 2). The FT-IR spectrum of the coating in Figure 2 displays the Si-O bonds at 500–600 cm⁻¹ and 700–800 cm⁻¹ [34]. In each of the two spectra, the formation of Si-O-Si and Si-O-Ti bands was confirmed, with peaks observed at 1080 and 925 cm⁻¹. Saravanan et al. [37] studied the FTIR spectra of a silicon dioxide/titanium dioxide thin film prepared using the sol–gel synthesis process and found that the asymmetric Si–O–Si vibration peak appears at 1073 cm⁻¹. It can be observed that the spectrum of the sorbent was indexed by distinguished peaks at around 3400 cm⁻¹ and 1631 cm⁻¹, which are ascribed to the bending vibration of O–H [38] and the O-H bending mode of the surface hydroxyl groups or free water adsorbed on the surface. The FT-IR spectrum of Bi₂S_3(9.6)-TiO₂-SiO₂/TNA (Figure 2B) demonstrated a decrease in some bands, which suggests that Bi₂S₃ is strongly linked with TiO₂-SiO₂/TNA. It also indicates that the stronger interactions between Bi₂S₃ and TiO₂-SiO2/TNA could be related to the shifts in band positions and changes in the intensity of Bi₂S_3(9.6)-TiO₂-SiO₂/TNA.

3.1.3. SEM-EDX Studies

A cross-sectional FE-SEM of the surface morphology of the Bi₂S_3(9.6)-TiO₂-SiO₂/TNA sample is displayed in Figure 3A–C. As evident from the figure, well-defined TiO₂ with a tubular structure appeared to be anodized at 60 V. As can be seen, the tubes presented a top-end-open, vertically oriented, and uniform structure (Figure 3D). The SEM images show that the nanotube arrays were covered by Bi₂S₃, TiO₂, and SiO₂ layers, so the surface morphology of the supported TNA became rough. EDX analysis data are listed in Figure 3D and confirm that the main elements composing the Bi₂S_3(9.6)-TiO₂-SiO₂/TNA sample were Ti, O, Si, Bi, and S.

3.1.4. Band Gap Energy and UV–Vis/DRS Analysis

To evaluate the optical properties of the TiO₂-SiO₂/TNA and Bi₂S_3(9.6)-TiO₂-SiO₂/TNA samples, UV–vis/DRS was applied, as exhibited in Figure 4A. The spectra of TiO₂-SiO₂/TNA indicated that the sample only absorbed UV light at wavelengths less than 400 nm due to the larger photon energy, which showed a steep absorption edge and had no absorbance in the visible region [37,38]. Similar phenomena have been reported by other researchers. Wang et al. [5] investigated the absorbance spectra of TiO₂-ZrO₂-SiO₂ and TiO₂-SiO₂ films. All the films only had strong absorption at wavelengths less than 300 nm. The absorbance between 300 nm and 350 nm sharply decreased. Wang et al. [39] carried out UV–Vis/DRS measurements to investigate the optical properties of SiO₂- TiO₂/TNTs. SiO₂- TiO₂/TNT film predominantly demonstrates a photoresponse at wavelengths less than 380 nm, which is associated with the absorption threshold of anatase TiO₂. Wanag et al. [40] investigated the UV-vis absorption spectra of a SiO₂/TiO₂ composite system for which strong absorbance only in the ultraviolet range of 250–375 nm was observed, and the process of photoexcitation was observed from the lower valence band to the upper CBis. Additionally, TiO₂-SiO₂/TNA modified with Bi₂S₃ exhibits a noticeable increase in the visible-light region, resulting from the combination of Bi₂S₃ and TiO₂-SiO₂/TNA, which indicates a significant red-shift effect [41]. The reflectance spectra of samples from the UV–Vis spectrophotometer and the plot of the modified Kubelka–Munk function were used. The Kubelka–Munk function is calculated by ${\left(\alpha h\nu \right)}^n={A(h\nu -E_g)}^{n/2}$, where α denotes the absorption coefficient, h is the Planck constant, ν corresponds to light frequency, A is a constant, E_g is the band gap, and n depends on the type of transition, where for direct allowed transition, n is equal to 2, and for an indirectly allowed transition, n is equal to ½ [42]. The E_g of semiconductor materials can be calculated by the extrapolation of the linear section of the ($\alpha h\nu$)² plot versus photon energy over the wavelength range of 200–700 nm. The results are shown in Figure 4B, and the energy band gap of TiO₂-SiO₂/TNA and Bi₂S_3(9.6)-TiO₂-SiO₂/TNA was determined, which can be reasonably assumed to be 3.27 eV and 1.87 eV, respectively [43].

These results indicate that the energy gap of Bi₂S_3(9.6)-TiO₂-SiO₂/TNA was narrower than that of TiO₂-SiO₂/TNA and, with improving efficiency of solar spectrum absorption, decreased the energy needed to excite the light-induced electrons. This catalyst exhibits noteworthy advantages because it utilizes more than two components [3] through the presence of Bi₂S_3(9.6)-TiO₂-SiO₂/TNA. The support of SiO₂ reduces the particle size of TiO₂ and improves its dispersive capacity, making more surface area of TiO₂ available to absorb more light. At the same time, SiO₂ has a strong reflective effect on light, and it is believed that the light reflected by SiO₂ could be absorbed by the photocatalyst on the surface of SiO₂ [4]. The relatively stronger light absorption promotes the creation of charge carriers in the photocatalytic reaction that can be efficiently enhanced to generate hydroxyl radicals (OH^•) and superoxide radicals (O₂^•−) on the surface [6]. It is noteworthy that the incorporation of Bi₂S₃ into TiO₂-SiO₂/TNA shows a favorable response to visible light because it has a small band gap and a superior absorption coefficient and promotes photocatalytic performance [7].

3.1.5. Electrochemical Analysis

To gain deep insight into the properties and photocatalytic mechanism of the Bi₂S₃-TiO₂-SiO₂/TNA heterojunction and investigate the charge transfer resistance on the interface during the photocatalytic process, EIS was recorded. Figure 5A,B illustrate the Nyquist and Bode log |Z| curves of EIS analysis. As is well known, the shorter radius of the arc of EIS spectra is indicative of a lower interfacial resistance and a more rapid charge transfer impedance. As the corresponding EIS results depicted in Figure 5A show, the Bi₂S_3(9.6)-TiO₂-SiO₂/TNA photocatalyst presented a shorter radius of the arc, reflecting the smallest charge transfer resistance on the interface and the higher charge transfer yield. The results agree with previous literature [8,9]:

3.2. Photocatalytic Performance of Photocatalysts

To evaluate the photocatalytic activity of the different samples by the photodegradation of azo dye, the reaction was carried out in a dye suspension with a concentration of 22.5 mg/L and a natural pH. Prior to the photocatalytic experiment, to determine the surface adsorption extent, experiments were carried out in dark conditions in the presence of the catalyst. The adsorption extent of BY 28 on the photocatalyst was determined to be about 15% after 180 min of contact time. Figure 6A shows a performance comparison for the development of the photocatalytic degradation of TiO₂/TNA, TiO₂-SiO₂/TNA, Bi₂S₃-TiO₂/TNA, and Bi₂S_3(9.6)-TiO₂-SiO₂/TNA. As can be observed from the degradation results, the Bi₂S_3(9.6)-TiO₂-SiO₂/TNA heterojunction exhibited higher photocatalytic activity over time than TiO₂/TNA, TiO₂-SiO₂/TNA, and Bi₂S₃-TiO₂/TNA within 180 min of irradiation. In contrast, the photocatalytic activity of TiO₂/TNA was not noticeable. Moreover, these results show that TiO₂-SiO₂/TNA and Bi₂S₃-TiO₂/TNA had almost lower degradability [2,4,43,44].

Using the linear correlation between ln(C/C₀) and the irradiation time, the first-order kinetic rate constant could be determined, whereby k could be calculated for all samples displayed in Figure 6B. The rate constant k was 0.025 min⁻¹ for the photocatalytic degradation of BY 28 by the obtained Bi₂S_3(9.6)-TiO₂-SiO₂/TNA, which is more than that of TiO₂/TNA, TiO₂-SiO₂/TNA and Bi₂S₃-TiO₂/TNA, especially as k was 0.0027, 0.0045, 0.0083, and min−1, which is 8 and 6 times that of TiO₂/TNA and TiO₂-SiO₂/TNA and 3 times that of Bi₂S₃-TiO₂/TNA, respectively. The results show that the deposition of Bi₂S₃ was beneficial for improving the photocatalytic activity of TiO₂-SiO₂/TNA photocatalysts. It can be seen that the reaction rate constant of TiO₂-SiO₂/TNA was 2 times higher than that of TiO₂/TNA, indicating the effect of the SiO₂, TiO₂, and TNA ternary components for enhancing the photocatalytic ability of TiO₂-SiO₂/TNA. Additionally, the introduction of SiO₂ can improve the specific surface area and light absorption, further promoting photocatalytic performance [8,23].

Table 2. Comparison of BY 28 dye degradation efficiency among catalysts.

Photocatalysts	Sources of Irradiation	Time (min)	Degradation (%)	Ref.
1.5 Mg–TiO2/SrTiO3	Visible-light irradiation	150	97%	[31]
Y2O3, CeO2, and ZnO/PdO	Visible-light irradiation	120	94%	[45]
TiO2/CoCr2O4/SrTiO3	Visible-light irradiation	120	98%	[46]
N-S doped TiO2	Visible-light irradiation	180	98%	[47]
Zn/Tu-TiO2	Visible-light irradiation	240	95%	[48]
N/F/S-doped TiO2	Visible-light irradiation	360	94%	[49]
Sr/S/N doped TiO2 nanolayers on glass orbs	Visible-light irradiation	480	91%	[50]
Bi2S3-TiO2-SiO2/TNA	Visible-light irradiation	180	95.45%	This work

shows a comparison of the results for dye removal efficiency with various photocatalysts. Notably, when using powder to photodegrade pollutants in wastewater, the performance of the photocatalytic systems increased, due to the strong oxidizing effect and large surface contact with the pollutants [19,36]. However, separation problems were encountered following repeated utilization. Comparing TiO₂ nanotube array structures with photocatalyst nanopowders demonstrates that TNAs have higher surface-to-volume ratios and larger specific surface areas, which not only increases electron transfer as well as the interpenetration of holes, but also enhances the separation and transferring efficient of the photogenerated electron–hole (e⁻–h⁺) pairs [3,10], offering great potential for enhancing the performance of photocatalytic systems [41].

3.3. Degradation Parameter Optimization Using RSM

In the present study, parameter optimization through RSM for the degradation of azo dye was determined. A five-level, three-factor design was implemented here. A total of 17 runs were necessary to improve the three individual parameters through the RSM. Consequently, the relationship between the degradation of BY 28 and the independent variable parameters was established and expressed by appropriate quadratic equations in the form of coded factors (Equation (6)):

$${\mathrm{Y}}_{\mathrm{B}\mathrm{Y}28}=88.62+6.90 {\mathrm{X}}_1+4.55 {\mathrm{X}}_2+17.24 {\mathrm{X}}_3-4.95{ \mathrm{X}}_1{\mathrm{X}}_1-20.51{ \mathrm{X}}_2{\mathrm{X}}_2-8.46{ \mathrm{X}}_3{\mathrm{X}}_3+2.12{ \mathrm{X}}_1{\mathrm{X}}_2-6.78{ \mathrm{X}}_1{\mathrm{X}}_3+4.87 { \mathrm{X}}_2{\mathrm{X}}_3$$

(6)

The analysis of variance (ANOVA) showed that the mathematical model can be applied to samples with a probability level of 95% and quantifies the dominance of the control factor. The test results of Bi₂S₃-TiO₂-SiO₂/TNA for BY 28 photodegradation are listed in

Table 3. ANOVA table from CCRD for BY 28 photodegradation.

Item	Degrees of Freedom	Sum of Squares	F-Value	p-Value
Model	9	10,396.8	120.05	0.000
Linear	3	4996.3	173.08	0.000
X1	1	651.4	67.69	0.000
X2	1	283.5	29.47	0.001
X3	1	4061.4	422.07	0.000
Square	3	4907.7	170.01	0.000
X1*X1	1	176.6	18.36	0.004
X2*X2	1	4744.7	493.08	0.000
X3*X3	1	808.0	83.97	0.000
2-Way Interaction	3	492.7	17.07	0.001
X1*X2	1	36.1	3.75	0.094
X1*X3	1	267.2	27.76	0.001
X2*X3	1	189.4	19.69	0.003
Error	7	67.4
Lack of Fit	5	31.8	0.36	0.847
Pure Error	2	35.6
Total	16	10,464.1

, which shows the status of considerable parameters and details of the ANOVA. To ensure the applicability of the model, the lack of fit was checked, which, with a p-value of 0.082 and an F-value of 0.76, indicated that the model was not significant regarding pure error, suggesting that the model effectively describes the measured response data within the experimental region. F-values and p-values (p < 0.01) are criteria for determining the significance of the less/more input variable effect on the measured response of the RSM model. F-values higher than the critical value (p < 0.05) indicate the significance of the quadratic model or role of the parameter [44]. p-values < 0.05 indicate a confidence level of 95%. In the present study, a p-value < 0.0001 demonstrated that the proposed model terms were significant, which shows that the second-order polynomial model fit the experimental results well [43]. From Table 3, it can be seen that the selected model was highly significant, with an F-value of 120.0 and a p-value of less than 0.0001, which is obvious from Fisher’s “F” test [41,42]. In addition, the regression coefficient (R²) for the photodegradation of BY 28 was 0.9936% and thus close to one, which indicates the suitability of the applied model and indicates that it reflected a good correlation between the observed and the predicted values. In addition, the values of R² and the adjusted R² were very close to each other (R²: 0.9936, Adj. R²: 0.9853), indicating a negligible amount of variation in the predicted and actual experimental data. The results presented in

Table 4. Experimental design for optimization of significant variables.

Run	Independent Variables			% Degradation
	X1: [Bi3+] (M)	X2: Dye Concentration (mg/L)	X3: pH	Experimental Value	Predicted Value
1	1	1	1	84.21	83.58
2	1	1	−1	53.66	52.73
3	−1.68	0	0	67.33	66.41
4	1	−1	−1	47.13	48.16
5	0	0	0	92.57	88.68
6	−1	1	1	79.35	78.56
7	−1	1	−1	22.14	23.56
8	1.68	0	0	90.25	88.06
9	0	0	0	88.69	88.68
10	1	−1	1	61.76	62.09
11	0	0	1.68	95.45	93.09
12	−1	−1	−1	27.65	27.43
13	−1	−1	1	61.85	65.52
14	0	1.68	0	38.28	37.42
15	0	−1.68	0	25.64	22.61
16	0	0	0	84.14	88.68
17	0	0	−1.68	36.63	35.11

indicate that a maximum degradation of 95.45% was attained, which closely aligns with the predicted response from the software (93.09%). The correspondence between the predicted response and experimental results demonstrates the applied model’s validity and its proficiency in forecasting the maximum degradation of BY 28. Figure 7A illustrates the relationship between the predicted responses versus experimental responses to the degradation of the dye, confirming the results were in close agreement. The normal plots of the studentized residuals are shown in Figure 7B. As shown in the plot, except for a few data points, most of the generated data points in the plot were scattered over a straight region, and no individual residuals exceeded the residual variance, confirming that the obtained data have a linear relationship and the errors along a zero mean have a normal distribution. This also confirmed the excellent suitability of the model for degradation purposes [51].

3.4. Key Parameter Optimization Using RSM

The three-dimensional (3D) surface plot generated from the RSM analysis displays a comparison among the contribution of key parameters in the design space. Figure 8A–C represent the effect of three independent parameters: the amount of Bi₂S₃ loaded on the TNA, dye concentration, and pH. In Figure 8A, interactions between the amount of Bi₂S₃ loaded on the TNA and the dye concentration on the response are shown. Five Bi₂S₃-TiO₂-SiO₂/TNA catalysts were prepared, and their Bi³⁺ content was determined by atomic absorption spectroscopy by digesting the catalysts in an HF + HNO₃ + H₂O₂ mixture. The summarized results are shown in

Table 5. Amount of Bi₂S₃ loaded on the TNA.

Catalysts				C (M)		Mass Loading (mg/cm2)
				Bi3+		Bi2S3
Bi2S3(6.1)-TiO2-SiO2/TNA				0.13		6.1
Bi2S3(9.6)-TiO2-SiO2/TNA				0.20		9.6
Bi2S3(11.4)-TiO2-SiO2/TNA				0.30		11.4
Bi2S3(13.7)-TiO2-SiO2/TNA				0.40		13.7
Bi2S3(15.8)-TiO2-SiO2/TNA				0.46		15.8
The Bi2S3-TiO2-SiO2/TNA photocatalytic process under optimal conditions in degradation of BY 28.
Dyes	X1		X2		X3		Deg. (%)
	Coded	Actual (M)	Coded	Actual (mg/L)	Coded	Actual
BY 28	0.18	0.30	0.25	22.5	1.03	8.34	95.45

. It was observed that, initially, as the Bi₂S₃ content increased, with a Bi³⁺ concentration of 0.30 M, the degradation of the dye increased. These results show that increasing the Bi₂S₃ content enhanced the degradation efficacy of the dye by increasing the number of active sites that can be activated to generate reactive radicals, including superoxide radicals and hydroxyl, thereby facilitating an enhancement in the degradation rate of BY 28 [52]. However, when the Bi₂S₃ content increased to 15.82, the degradation efficacy of the dye decreased. Produced hydroxyl radicals can be recombined into less reactive H₂O₂ at higher doses [10,44,53]. These results reveal that an optimum amount of Bi₂S₃ was required to obtain high BY 28 degradation. Based on the results obtained from the RSM-CCRD technique in the photodegradation of BY 28 dye, the catalyst consisting of 9.6 mg/cm² (designated as Bi₂S_3(9.6)-TiO₂-SiO₂/TNA) showed an increase in degradation efficiency compared to other synthesized catalysts. Hence, the Bi₂S_3(9.6)-TiO₂-SiO₂/TNA catalyst was used in the following experiments and designated as the “optimized catalyst”.

According to Figure 8B, the degradation process for BY 28 elimination was enhanced at a concentration of 22.5 mg/L. According to the collision theory, the probability of dye molecules colliding with the surface of the catalyst increases when the dye concentration rises to a certain level, leading to a high degradation efficiency [17,21]. Notably, at low dye concentrations, this probability is low, and as the concentration increases, it increases to the optimum concentration. The degradation efficiency was further enhanced by increasing the dye concentration to levels higher than the optimal range; high levels of BY 28 molecules can have a screening effect by absorbing some of the illuminated photons [54]. Thus, the degradation yield decreases due to the low production of free radicals and the saturation of the catalyst surface. Therefore, low and high BY 28 concentrations can result in a lower efficiency of photochemical degradation [52].

Figure 8C displays the 3D surface plot of the response to the significant quadratic effects of Bi₂S₃ and pH, indicating that the response increased the degradation efficiency with an increase in pH from 6 to 8 for the degradation of BY 28 dye. The surface of a catalyst exhibits a negative charge when the pH is basic, and a positive charge when the pH is acidic [41,42,45]. BY 28 dye is classified as a cationic azo dye, which has surfaces with a positive charge [52]. Because the surface charge of the Bi₂S_3(9.6)-TiO₂-SiO₂/TNA photocatalyst became negatively charged at a solution pH greater than 6, the dye’s degradation was observed to be higher in basic solutions compared to acidic solutions. An increase in the pH value in the solution, producing large numbers of H+ and thereby occupying the active sites on the surface of the photocatalyst and creating electrostatic interactions between the dye and H⁺ ions, led to a reduction in the interaction rate between the photocatalyst and the dye, which resulted in a lower degradation yield. At high pH values, the catalyst surface gradually became negatively charged [55]. The high pH value was favorable for the adsorption of cationic dyes due to electrostatic interaction, increasing the contact rate between the catalyst and the organic dye molecule [5], and assisted higher photogenerated charge carrier excitation, which resulted in a greater photocatalytic yield. The alkaline solution not only contained a higher concentration of OH⁻ ions but influenced the amount of hydroxyl radicals (OH^⦁) formed. The OH⁻ ions were adsorbed onto the catalyst’s surface and subsequently oxidized to OH^⦁ radicals by the holes during the photoexcitation process [41]. The hydroxyl radicals, which were an active species in the photocatalytic yield of BY 28, are recognized as a highly reactive species that is a non-selective and extremely strong oxidant, resulting in the breakdown of various organic compounds [4]. At pH values higher than 8 (especially at high basic pH values), the repulsive force between the anionic BY 28 molecules, which were most abundant at these pH values, and the negatively charged catalyst surface repelled BY 28 molecules, resulting in a sharp decrease in degradation efficiency [8,9].

The identification of the operational parameters for the optimization of the process response has significant importance. RSM was used for process optimization using the different parameters listed in Table 4, with the objective of maximizing the degradation efficiencies of the dye.

The assessment of Chemical Oxygen Demand (COD) revealed a decline from 111 to 21 mg/L, thereby suggesting that the Bi₂S_3(9.6)-TiO₂-SiO₂/TNA photocatalyst has significant potential for the implementation of a photocatalytic system in the enhancement of water resource purification.

3.5. Energy Level Structure and the Photocatalytic Degradation Mechanism

Theoretical calculations based on Mulliken electronegativity theory [11] were used to determine the relative band edge positions of Bi₂S_3(9.6)-TiO₂-SiO₂/TNA, as shown:

$$E_{VB}=\chi -E_c+{\displaystyle \frac{1}{2}}{E}_g$$

(7)

$$E_{CB}=E_{VB}-E_g$$

(8)

Specifically, the valance band potential (E_VB), conduction band potential (E_CB), band gap of the semiconductor (E_g), free electron energy of the standard hydrogen electrode potential (E_c = 4.5 eV), and absolute electronegativity of the corresponding semiconductor (X) were used to estimate the energies of the bands [52]. A possible photocatalytic mechanism during a photocatalytic reaction for Bi₂S_3(9.6)-TiO₂-SiO₂/TNA heterojunction was estimated. There were two charge transfer mechanisms: a conventional type II heterojunction and a direct Z-scheme heterojunction (Figure 9A,B). Irradiation by visible light can excite both TiO₂ and Bi₂S₃ to generate charge carriers (e⁻, h⁺). In a typical type II heterojunction system (Figure 9A), the charge carriers (e⁻, h⁺) migrate through the typical charge transfer process [56,57].

The CB of TiO₂ hosted the photoexcited electrons (e_CB⁻), and the CB potential of TiO₂ (−0.29 eV) was more positive and not adequate to reduce O₂ to O₂^•− (O₂/O₂^•− = −0.33 eV vs. NHE) as an active radical species to degrade dye [56,57]. In addition, the VB of Bi₂S₃ hosted the photogenerated holes (h_VB⁺), and the position of the VB was well above the standard potential of H₂O/HO^• (OH^•/OH⁻ = 2.40 eV vs. NHE), and it is an inactive band which leads to no HO^• production [55,57,58]. In the second mechanism in Figure 9B, based on the potential edges in the Z-scheme mechanism, in the charge transfer pathway, the photogenerated charges migrated the photoexcited electrons at the CB of TiO₂ transfer and combined with the photogenerated holes to the VB of Bi₂S₃, resulting in the photogenerated electrons in the CB of Bi₂S₃ and holes in the VB of TiO₂. Therefore, the abundant accumulated electrons in Bi₂S₃ with strong reducing power could rapidly release CB to the adsorbed O₂ to generate O₂^•−, while the holes were formed in the TiO₂ valance band, which can not only react with pollutants but also generate OH^• by reacting with H₂O, which oxidizes BY 28 to H₂O and CO₂. As a result, the formation of Z-schemes in the photocatalytic system could explain the improvement of photocatalytic performance for organic pollutants [55,56,57].

3.6. Reusability of Bi₂S_3(9.6)-TiO₂-SiO₂/TNA

The reuse experiment was conducted to evaluate the practicability and cost-effectiveness of using photocatalysts for environmental remediation [43,51]. The photocatalyst was reused for multiple runs at the optimal conditions [59]. The results are shown in Figure 10. The reused photocatalyst was washed and dried at a temperature of 100 °C and used in further experiments in fresh conditions. The results indicate that the efficiency of photocatalyst degradation decreased to about 9% of the initial photocatalytic activity after five reuse runs.

4. Conclusions

In summary, TiO₂, SiO₂, and Bi₂S₃ nanoparticles were loaded on TNAs (Bi₂S₃-TiO₂-SiO₂/TNA) and used for the degradation of BY 28 under visible light. The XRD results indicate that TiO₂-SiO₂/TNA and Bi₂S_3(9.6)-TiO₂-SiO₂/TNA films have the characteristics of the (101) crystal plane of the anatase type of TiO₂. UV–Vis/DRS measurements of TiO₂-SiO₂/TNA modified with Bi₂S₃ exhibited a noticeable increase in the visible-light region, which provided effective visible-light absorption and electron excitation. The Nyquist curves of the EIS analysis showed the Bi₂S_3(9.6)-TiO₂-SiO₂/TNA photocatalyst had a shorter radius of arc, reflecting the smallest charge transfer resistance on the interface and the higher charge transfer yield. In this research paper, CCRD-RSM was employed for the optimization of the key parameters in BY 28 photocatalytic degradation. The RSM model was found to have an R² and adjusted R² value of 0.9936 and 0.9853, respectively, and could significantly (p < 0.0001) predict the response variables. Using Bi₂S_3(9.6)-TiO₂-SiO₂/TNA, the optimum conditions were set as 0.30 M, 22.5 mg/L, and 8.34 for the amount of Bi₂S₃ loaded on the TNA, dye concentration, and pH, respectively, with a maximum degradation efficiency (%) of 95.45%. The amount of Bi₂S₃ loaded on the TNA is a critical factor in the photocatalytic performance of Bi₂S₃-TiO₂-SiO₂/TNA due to the enhancement of light absorption and the ability of the Bi₂S₃ to absorb more visible light and promote the photocatalytic performance to generate a higher number of electrons and holes. On the other hand, SiO₂ has an evidently enhanced effect on the photocatalytic performance of Bi₂S_3(9.6)-TiO₂-SiO₂/TNA because it considerably enhances the thermal stability and dispersivity and increases the surface area. The Bi₂S_3(9.6)-TiO₂-SiO₂/TNA photocatalyst has two mechanisms: the conventional type II heterojunction and direct Z-scheme systems. The direct Z-scheme mechanism is suitable for the production of superoxide due to the accumulating electrons in the CB of Bi₂S₃. In addition, the direct Z-scheme can facilitate charge transportation and separation during photocatalytic reactions, which suppresses the electron–hole recombination rate, improving the photodegradation ability of BY 28.