Construction of Honeycomb-like ZnO/g-C₃N₅ Heterojunction for MB Photocatalytic Degradation

Abstract

In this study, a combination of calcination and hydrothermal methods was employed to synthesize a honeycomb-like ZnO/g-C₃N₅ (ZCN) heterojunction in situ. The ZCN heterojunction photocatalyst exhibits remarkable photocatalytic degradation performance, achieving a 97% methyl blue (MB) degradation rate with the rate constant of 0.0433 min⁻¹ (almost twice that of ZnO). Optical performance tests reveal that the ZCN heterojunction broadens the absorption edge to 710 nm and enhances the charge carrier separation. The presence of abundant oxygen vacancies, as revealed by X-ray photoelectron spectroscopy analysis, effectively suppresses the recombination of photogenerated electron–hole pairs. Furthermore, density functional theory simulations indicate that the combination of ZnO and g-C₃N₅ creates an internal electric field due to their differing work functions. This leads to the formation of a Z-scheme heterojunction that effectively suppresses charge carrier recombination and preserves the strong redox capabilities of ZnO and g-C₃N₅. Finally, electron spin resonance results indicate that ${\mathrm{O}}_2^{-}$ and OH are the primary active radicals involved in the degradation process. This study introduces a potential approach for the development of highly efficient Z-scheme photocatalysts for water treatment applications.

1. Introduction

Rapid industrialization has substantially resulted in the increased presence of various dyes in water systems [1]. These dye pollutants pose significant risks to human health and environmental sustainability due to their carcinogenic properties and limited biodegradability in the environment [2]. To resolve these problems, various techniques have been employed for water treatment to eliminate organic pollutants, such as membrane separation [3], chemical oxidation [4], coagulation [5], and adsorption [6]. However, these methods are expensive and complex and may even result in the generation of toxic by-products. Hence, it is essential to develop effective techniques to overcome these drawbacks.

In recent years, semiconductor-based photocatalysis has gained attention as a promising and sustainable strategy to tackle this issue. During the photocatalysis process, electron–hole pairs are generated in the conduction band (CB) and valence band (VB) of a semiconductor upon irradiation with simulated sunlight [7]. Subsequently, the electron in the CB reacts with oxygen (O₂) to form superoxide radical anions (O²⁻), while the hole in the VB interacts with water to produce hydroxyl radicals (OH). These two highly reactive radicals are considered vital species in degrading organic pollutants in water. Various semiconductor photocatalysts such as zinc oxide (ZnO) [8], titanium dioxide (TiO₂) [9], and cadmium sulfide (CdS) [10] have been successfully applied in water purification. Among them, ZnO has emerged as a highly promising material for both solar energy conversion and the photodegradation of organic pollutants, thanks to its low toxicity, abundant availability, excellent chemical and optical stability, superior photocatalytic efficiency, and affordability [11,12]. Yashni et al. synthesized ZnO nanoparticles from Citrus sinensis peels to degrade Congo red (CR) under UV irradiation. The results showed that the ZnO/UV system achieved 96.73% CR decolorization and 96.70% COD removal under optimized operation parameters (0.171 g of ZnO nanoparticles, pH 6.43, and 5 mg·L⁻¹ of CR) [13]. Sahu et al. prepared ZnO, SnO₂, and MoS₂ to degrade dye pollutants like Rhodamine B (RhB) and CR under simulated sunlight irradiation, among which ZnO (88.02% RhB, 76.56% CR) exhibited higher degradation performance than the other two photocatalysts (80.11% RhB and 74.25% CR for SnO₂; 78.63% RhB and 73.14% CR for MoS₂) [14]. However, the following two issues still hinder the practical applicability of ZnO, which urgently needs to be resolved: (i) ZnO primarily responds to ultraviolet light (bandgap of 3.20 eV), resulting in minimal utilization of visible light. (ii) The photoinduced electron–hole pairs in pure ZnO have a short lifetime [15].

Graphite nitride carbon (g-C₃N₅) is considered a novel semiconductor with significant potential in energy and environmental applications, thanks to its stability, eco-friendliness, and responsiveness to visible light irradiation [16]. Moreover, the nitrogen-rich conjugated structure diminishes the barriers for electron transitions and enhances electron mobility, thereby augmenting its potential utility in the photocatalytic degradation of aqueous pollutants [17]. However, g-C₃N₅ is inherently limited by considerable carrier recombination and a lack of sufficient reactive sites, which restricts its photocatalytic efficiency. Many researchers have overcome the above drawbacks via constructing heterojunctions. Guan et al. designed a g-C₃N₅/Ti₃C₂ binary heterojunction to improve the photocatalytic performance of g-C₃N₅ [18]. After coupling with Ti₃C₂, the composite displayed 81.6% antibiotic tetracycline degradation within 60 min, which is much higher than that of g-C₃N₄/Ti₃C₂ (65.3%), Ti₃C₂ (50%), g-C₃N₄ (60.7%), and g-C₃N₅ (33.1%). They believe that the enhanced photoactivity can be attributed to the improvements in the light harvest, the charge carrier separation, and the surface reaction rate caused by constructing the heterojunction with Ti₃C₂. Chen et al. constructed a g-C₃N₅/Bi₂MoO₆ heterojunction for aflatoxin B1 photocatalytic degradation [19]. After coupling Bi₂MoO₆ with g-C₃N₅, the photodegradation efficiency of AFBi was improved to 92% within 75 min. The integration of g-C₃N₅ enhances the separation and migration efficiency of photogenerated carriers in Bi₂MoO₆. It has been reported that a ZnO nanosphere with a large specific surface area and reactive sites will facilitate the transfer and recombination of photogenerated electrons [20]. Therefore, integrating g-C₃N₅ nanosheets with ZnO nanospheres to establish a heterojunction may be an effective strategy to enhance the photocatalytic efficiency of sole ZnO and g-C₃N₅, which is the primary aim of this work.

2. Materials and Methods

2.1. Chemicals and Materials

Zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O), trisodium citrate dehydrate (C₆H₅Na₃O₇·2H₂O), urea (CH₄N₂O), 3-Amino-1,2,4-triazole (3-AT), methylene blue (MB), and ethanol (EtOH) were bought from Aladdin Biochemical Technology Co., Ltd., Shanghai, China. 5,5-dimethyl-1-pyrroline nitrogen oxide (DMPO) was bought from Alfa Aesar Chemical Co., Ltd., Tianjin, China. All experimental reagents were used directly without further purification. Deionized (DI) water (18.25 MΩ/cm) was used in the experiments.

2.2. Synthesis of Catalysts

2.2.1. Synthesis of g-C₃N₅

Figure 1a illustrates the synthesis process of bulk carbon nitride (g-C₃N₅). This material was prepared using a modified one-step thermal polymerization method [21]. Specifically, 6 g of 3-AT was placed in a sealed crucible and heated in a muffle furnace. The material was calcined at 550 °C for 4 h, with a heating rate of 5 °C per minute. After cooling, the resulting orange powder was ground and collected as the reference sample.

2.2.2. Synthesis of ZnO and ZnO/g-C₃N₅

ZnO was synthesized using an improved hydrothermal method, building on the approach outlined in the literature [22]. The synthesis procedure is as follows (Figure 1b): 0.05 mol of Zn(NO₃)₂·6H₂O, 0.1 mol of CH₄N₂O, and 0.005 mol of C₆H₅Na₃O₇·2H₂O were first dissolved in 75 mL of deionized water. The solution was stirred for 5 min, then transferred to a 100 mL autoclave. It was heated at 120 °C for 6 h to facilitate the reaction. After cooling to room temperature, the resulting product was thoroughly washed three times with deionized water and ethanol. Finally, the material was subjected to annealing at 350 °C for 45 min in a tubular furnace. Then, the white powder was collected as another reference sample. The synthesis process of ZnO/g-C₃N₅ (ZCN) is similar to that of ZnO except for adding g-C₃N₅ powder at an optimized mass ratio of ZnO/g-C₃N₅ = 1:0.8 in the precursor solution (Figure S1). Finally, the yellow powder was obtained as an experimental sample (Figure 1c).

2.3. Characterization of Samples

X-ray powder diffraction (XRD) patterns were obtained with a D8 advanced X-ray diffractometer (Bruker, Karlsruhe, Germany) equipped with Cu Kα radiation at 40 kV and 40 mA. Fourier transform infrared spectroscopy (FTIR) was performed using a Nicolet iS20 spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) to analyze surface functional groups. Morphological features were examined via field emission scanning electron microscopy (FE-SEM) using a Merlin compact SEM (Zeiss, Oberkochen, Germany). Specific surface areas and pore size distributions were determined through Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda (BJH) methods with a JW-BK200 instrument (JWGB, Beijing, China). The valence states of elements were analyzed by X-ray photoelectron spectroscopy (XPS) using an ESCALAB Xi+ spectrometer (Thermo Fisher Scientific, Waltham, USA). UV-vis diffuse reflectance spectra (UV-vis DRS) in the range of 200–800 nm were recorded with a UV2600 spectrometer (Shimadzu, Kyoto, Japan). Photoluminescence (PL) spectra were measured using an F-7100 fluorescence spectrometer (Hitachi, Tokyo, Japan), and electron spin resonance (ESR) signals were recorded with an A300 spectrometer (Bruker, Karlsruhe, Germany).

2.4. Degradation Experiments

The photocatalytic MB decomposition tests with different catalysts were carried out under the exposure of a 300 W Xe lamp. Typically, the photocatalysts (0.2 g·L⁻¹) were dispersed in the 50 mL MB solution (30 mg·L⁻¹) using ultrasonic treatment, followed by magnetic stirring to reach adsorption-desorption equilibrium. The simulated light source was then turned on to begin the reaction. At 10 min intervals, 6 mL samples were collected, filtered through a 0.22 μm polyether sulfone membrane, and analyzed using a UV-visible spectrophotometer at 664 nm. To ensure reliability and minimize measurement errors, each experiment was conducted in triplicate.

2.5. Density Functional Theory (DFT) Calculations

The DFT simulations were performed in the Materials Studio software using the CASTEP module. For the construction of surface models, the thickness of the vacuum slab was set to 20 Å. ZnO (101) and g-C₃N₅ (002) were constructed as input models. The exchange-correlation interaction was defined by generalized gradient approximation (GGA) parameterized with the Perdew-Burke-Ernzerhof (PBE) functional. The energy cutoff was set to 630 eV, and the Monkhorst-Pack k-point mesh was set to Gamma point. The atom convergence tolerance was set to 2 × 10⁻⁶ eV [23].

3. Results and Discussion

3.1. Characterization of Catalysts

The phase structure of the catalysts was determined by XRD analysis. As illustrated in Figure 2a, the peaks of ZnO at 31.7°, 34.5°, 36.2°, 47.3°, 56.5°, 62.7°, and 67.9° correspond to the crystal planes (100), (002), (101), (102), (110), (103), and (112), which are in perfect agreement with the hexagonal wurtzite crystalline phase (JCPDS NO. 36-1451) [24]. In the case of g-C₃N₅, a distinct peak at 27.4° is assigned to the (002) plane, corresponding to the π–π interlayer stacking of carbon nitride [25]. The XRD pattern of the synthesized ZCN sample reveals distinct diffraction peaks characteristic of ZnO (denoted by rhombus symbols) and g-C₃N₅ (denoted by a star symbol). This confirms that the in situ synthesis of ZnO on g-C₃N₅ successfully forms a heterojunction without altering the fundamental structures of either ZnO or g-C₃N₅. Furthermore, FTIR spectroscopy was used to investigate the surface functional groups of the catalysts. As illustrated in Figure 2b, ZnO exhibits a broad peak in the range of 430–530 cm⁻¹, which is attributed to the Zn-O stretching vibrations [26]. The peaks of ZnO observed in the range of 1430–1580 cm⁻¹ are assigned to the stretching vibrations of C=O and the bending vibrations of O-H, respectively [27]. For g-C₃N₅, a sharp characteristic peak at 810 cm⁻¹ corresponds to the condensed C-N heterocycles of the triazine moiety [28]. The peaks within the range of 1240–1630 cm⁻¹ are associated with C=N and C-N stretching vibrations in the triazole and triazine rings [29], while the characteristic peaks at 3000 and 3600 cm⁻¹ are ascribed to residual -NH₂ groups [30]. Notably, the ZCN exhibits the characteristic peaks of both g-C₃N₅ and ZnO, without the appearance of any additional peaks, indicating the successful synthesis of the composite material with high purity.

The SEM images of ZnO, g-C₃N₅, and ZCN samples are shown in Figure 3. As illustrated in Figure 3a, ZnO exhibits a spherical morphology with an average diameter of approximately 5–7 µm. In Figure 3b, the g-C₃N₅ nanostructure displays a block-like structure stacked by nanosheets. Figure 3c,d reveals that the introduction of g-C₃N₅ alters the morphology of ZnO to a certain extent, leading to the formation of a porous honeycomb-like structure.

The surface area characteristics of ZnO, g-C₃N₅, and ZCN samples were analyzed by using BET analysis based on N₂ adsorption-desorption isotherms. As shown in Figure 4, all the samples exhibit a typical type IV isotherm with an H3 hysteresis loop [31], indicating the existence of lamellar or sheet-like structures. From the BJH pore size distribution plots (the insets of Figure 4), it can be observed that they are mesoporous materials. As listed in Table S1, ZnO demonstrates a significantly higher adsorption capacity compared to the other two samples, owing to its porous spherical structure. After the construction of a heterojunction between ZnO and g-C₃N₅, a slight reduction in the S_BET value is observed, which can be attributed to the fact that a small amount pores of ZnO are occupied by g-C₃N₅.

3.2. Photocatalytic Performance

As illustrated in Figure 5, the photocatalytic performance of the ZnO, g-C₃N₅, and ZCN samples was investigated by degrading the MB under simulated sunlight irradiation. As shown in Figure 5a, 16% MB can be degraded within 70 min in the absence of photocatalysts. Notably, the MB removal efficiencies are significantly enhanced after the introduction of photocatalysts. After 70 min of photocatalysis, the MB degradation rates in the pure g-C₃N₅ system and ZnO system are 70% and 78%, respectively, while the ZCN system achieves a significantly higher degradation rate exceeding 97%. To further illustrate the changes in organic matter concentration during the degradation process, the time-dependent absorbance variations of the MB solution were monitored at a 10 min interval using the ZCN catalyst as a representative, as displayed in Figure 5b. The corresponding degradation kinetics were also calculated, and the results are shown in Figure 5c,d. The kinetic constant of the ZCN system is 0.0433 min⁻¹, which is 1.9 times and 2.6 times that of the ZnO system and g-C₃N₅ system, respectively. In scientific research, the evaluation of catalytic performance predominantly focuses on the catalyst’s stability and its recyclability. As shown in Figure 5e,f, the MB removal efficiency by ZCN remained over 88% after four cycles under simulated sunlight, demonstrating the excellent stability of ZCN. However, for the fifth cycle, the MB degradation efficiency declined to approximately 78%, which could be attributed to catalyst loss, photo-corrosion during recycling, and competitive adsorption on the catalyst surface. Subsequently, we compared our research findings with those of previous studies, as summarized in

Table 1. Comparison of MB degradation performance of ZCN with some reported ZnO and g-C₃N₅ catalytic systems.

System	Condition	Time	Degradation Percentage	Ref.
ZCN	300 W Xe lamp, [MB]0 = 30 mg·L−1, [Catalyst]0 = 0.2 g·L−1	70 min	97%	This work
TiO2/ZnO/rGO	300 W Xe lamp, [MB]0 = 20 mg·L−1, [Catalyst]0 = 0.5 g·L−1	180 min	~80%	[8]
Cu-ZnO	300 W Xe lamp, [MB]0 = 5 mg·L−1, [Catalyst]0 = 0.5 g·L−1	90 min	~81%	[32]
ZnO: Eu (10%)-M	300 W Xe lamp, [MB]0 = 10 mg·L−1, [Catalyst]0 = 1 g·L−1	150 min	~90%	[15]
3wt % Cu-doped ZnO (CuZn3)	300 W Xe lamp, [MB]0 = 10 mg·L−1, [Catalyst]0 = 0.1 g·L−1	180 min	92%	[33]
ZnO/SnO2/carbon	300 W Xe lamp, [MB]0 = 25 mg·L−1, [Catalyst]0 = 3.6 g·L−1	300 min	~50%	[34]
ZnO-Cu-CuO/C3N4	300 W Xe lamp, [MB]0 = 30 mg·L−1, [Catalyst]0 = 0.2 g·L−1	120 min	~75%	[35]
ZnO/g-C3N4	300 W Xe lamp, [MB]0 = 5 mg·L−1, [Catalyst]0 = 0.4 g·L−1	100 min	~80%	[36]
TiO2/g-C3N5	500 W Xe lamp, [MB]0 = 20 mg·L−1, [Catalyst]0 = 2.5 g·L−1	180 min	97%	[37]
g-C3N5/MK30	300 W Xe lamp, [MB]0 = 10 mg·L−1, [Catalyst]0 = 0.2 g·L−1	40 min	90%	[38]

. The results indicated that it took more than 100 min to achieve 80% efficiency or higher in approximately two-thirds of the reports, regardless of whether ZnO or g-C₃N₅ was used in composite materials. Our work demonstrated a degradation efficiency exceeding 97% within 70 min with the ZCN photocatalyst.

These results demonstrate the superior photocatalytic performance of the ZCN heterojunction compared to ZnO and g-C₃N₅, which can be employed as a promising material applied for the photocatalytic degradation of organic compounds.

3.3. Photocatalytic Mechanism Exploration

The photocatalytic performance of photocatalysts is primarily influenced by their light absorption capacity and the efficiency of photogenerated electron–hole pair separation under illumination [39]. Therefore, the enhanced degradation mechanism of the honeycomb-like ZCN heterojunction was investigated via these two fundamental perspectives.

To determine the light absorption capacity of the materials, the UV-vis DRS spectra of ZnO, g-C₃N₅, and ZCN were recorded (Figure 6a). ZnO only absorbs UV light with an absorption edge located at 395 nm, which aligns well with previously reported findings for ZnO [40]. The absorption edges of g-C₃N₅ and ZCN are approximately 720 and 710 nm, respectively. In comparison to ZnO, the absorption spectrum of ZCN exhibits a significant redshift, signifying an enhanced visible light absorption range for ZnO upon the formation of a heterojunction with g-C₃N₅. The bandgap values of the synthesized samples were further determined using the Tauc model, described by the following equation [41]:

$$\alpha \mathrm{h}\nu ={\mathrm{A}(\mathrm{h}\nu -{\mathrm{E}}_{\mathrm{g}})}^{\mathrm{n}/2}$$

(1)

where α is the absorption coefficient, hν is the photon energy, E_g represents the bandgap energy, A is a proportionality constant, and n depends on the type of electronic transition (n = 1 for direct leaps, while n = 4 for indirect leaps). According to previous studies, ZnO is a direct bandgap semiconductor, characterized by an n value of 1, while g-C₃N₅ acts as an indirect bandgap semiconductor with an n value of 4 [42,43]. As illustrated in Figure 6b,c, the bandgap energy (E_g) values for ZnO, g-C₃N₅, and ZCN were determined from the Tauc plots by extrapolating the linear tangent to the abscissa. The estimated E_g values are 3.20, 2.16, and 2.79 eV for ZnO, g-C₃N₅, and ZCN, respectively. Based on the valence band XPS (VB-XPS) spectra (Figure 6d,e), the conduction band (CB) positions of the catalysts were calculated and are presented in Figure 6f. The narrower energy gap of g-C₃N₅ suggests its enhanced capability for efficient potential-driven electron transfer to ZnO.

When semiconductors are exposed to light, electrons in the VB are excited to the CB, resulting in the generation of electron-hole pairs. However, if these charge carriers are not effectively utilized, they tend to recombine rapidly, leading to their annihilation. Thus, effective separation of photogenerated electrons and holes is essential for semiconductors to achieve enhanced photocatalytic performance under light irradiation. PL is a well-established method for characterizing the recombination efficiency of photogenerated electron-hole pairs, and thus the room-temperature measurements were performed using an excitation wavelength of 320 nm, as depicted in Figure 7. ZCN presents two emission peaks at 380 and 435 nm, corresponding to the characteristics of ZnO and g-C₃N₅, respectively [44]. Compared to ZnO and g-C₃N₅, ZCN exhibits the significantly lower PL intensity, indicating the lowest recombination rate of photogenerated electrons. This suggests that the honeycomb-like heterojunction formed through the in situ synthesis of ZnO on g-C₃N₅ significantly enhances the photogenerated electron separation efficiency of ZnO.

To gain deeper insight into the electron transfer process between ZnO and g-C₃N₅, XPS was employed to compare the spectra of ZCN with those of pure ZnO and g-C₃N₅. Figure 8a shows the high-resolution Zn 2p spectrum of ZnO, and it displays the characteristic peaks at 1021.98 and 1045.07 eV, ascribed to the Zn 2p_1/2 and Zn 2p_3/2 orbits, respectively [45]. The corresponding peaks shift to 1021.88 and 1044.94 eV for ZCN, suggesting electron transfer from g-C₃N₅ to ZnO in the composite, resulting in the formation of a built-in electric field from g-C₃N₅ to ZnO. The O 1s spectrum of ZnO displayed in Figure 8b can be deconvoluted into three peaks at 530.28, 531.75, and 534.04 eV, attributed to lattice oxygen bound to Zn, oxygen vacancies, and adsorbed H₂O, respectively [46,47]. The change in binding state of O is generally not employed to determine the built-in electric field due to the low intensity of the peak compared to Zn according to previous research [27]. Nevertheless, the peak of oxygen vacancies (531.75–532.10 eV) is significantly enlarged for ZCN. It indicates an enhanced probability of electron capture by defective energy levels in the ZCN. These findings further confirm that the honeycomb-like ZCN heterojunction exhibits excellent separation efficiency for photogenerated electron–hole pairs. The C 1s spectra are displayed in Figure 8c, and the characteristic peak of g-C₃N₅ at 284.86 eV is attributed to graphitic carbon (C-C), while the peaks at 286.60, 288.18, and 293.45 eV belong to C-NH₂ groups, sp²-hybridized carbon (N=C-N), and π–π^* bonds, respectively [48,49]. The N 1s spectrum of g-C₃N₅ in Figure 8d can be divided into four peaks at 398.55, 399.54, 400.68, and 404.57 eV ascribed to the terminal C-N=C, tertiary N-(C)₃, amino groups, and π–π^* bond, respectively [49]. The binding energies of C 1s and N 1s in the ZCN both shift toward higher energies to different degrees, indicating electron loss from C and N in the ZCN. Again, this result confirms that electrons transfer from g-C₃N₅ to ZnO.

The work function (Φ) of a catalyst is a key parameter related to its charge transfer ability, which is related to the contact potential difference (ΔV) expressed by the following equation [50]:

$$\Delta \mathrm{V}=\mathsf{\Phi }-\mathsf{\phi }$$

(2)

where φ is the work function of the instrument. ΔV is associated with the binding energy of electrons, which influences the kinetic energy of free electrons. It can be calculated through the distance of inflection points (IPs) in the VB-XPS spectrum [51]. As shown in Figure 6d,e, the ΔV for ZnO and g-C₃N₅ is 1.2 and 0.8 eV, respectively. This indicates that g-C₃N₅ has a higher Fermi energy level (EF) and smaller work function than ZnO, suggesting that electrons can transfer readily from g-C₃N₅ to ZnO.

Moreover, the EF of ZnO and g-C₃N₅ was estimated by DFT calculations, as illustrated in Figure 9a,b. The larger EF of ZnO signifies that the electrons are more likely to escape from g-C₃N₅ to ZnO at the interface. Consequently, as shown in Figure 9c,d, upon contact between ZnO and g-C₃N₅, electrons from g-C₃N₅ are transferred to ZnO until their Fermi levels reach equilibrium. This process results in the formation of an internal electric field (IEF) at their interface and, thereby, the band bending of ZnO and g-C₃N₅. This promotes the recombination of electrons in the CB of ZnO (CB_ZnO) with holes in the VB of g-C₃N₅ (VB_CN) under simulated sunlight irradiation, effectively suppressing the recombination of photogenerated charge carriers within their respective semiconductors. Finally, the photogenerated carriers at the interface establish a Z-scheme pathway, with accumulated electrons and holes localized at the CB of g-C₃N₅ (CB_CN) and the VB of ZnO (VB_ZnO), respectively.

To further elucidate the Z-scheme pathway mechanism, ESR spectroscopy was employed by using DMPO as the capture agent for ${\mathrm{O}}_2^{-}$ and OH. The DMPO-${\mathrm{O}}_2^{-}$ patterns displayed in Figure 10a clearly exhibit a distinct signal corresponding to ${\mathrm{O}}_2^{-}$ $({{\mathrm{E}}^0}_{{\mathrm{O}}_2/{\mathrm{O}}_2^{-}}$ = −0.16 V) [52], generated by the g-C₃N₅ (E_CB-CN = −0.67 V) of the ZCN heterojunction during the photocatalytic process. As shown in Figure 10b, a DMPO-OH signal can be seen, indicating that water is oxidized to OH under light irradiation. Given that the VB position of g-C₃N₅ (E_VB-CN = 1.49 V) is lower than the oxidation potential of water $({{\mathrm{E}}^0}_{{\mathrm{H}}_2\mathrm{O}/\mathrm{O}\mathrm{H}}$ = 2.3 V) and the VB position of ZnO (E_VB-ZnO = 2.78 V) is higher, the ZnO-generated holes can oxidize water, but the g-C₃N₅-generated holes cannot. Thus, only the VB of ZnO contributes to the generation of OH during the photocatalytic degradation of MB using the ZCN heterojunction. Obviously, these results not only confirm the Z-scheme pathway mechanism but also prove that ${\mathrm{O}}_2^{-}$ and OH were involved in the degradation process.

Based on the aforementioned results and findings, a photocatalytic mechanism is suggested and depicted in Figure 11. Upon simulated sunlight irradiation (hv), electrons in the ZCN heterojunction are excited from the VB to the CB. The IEF at the interface facilitates the recombination of electrons from the CB of ZnO (${\mathrm{e}}_{\mathrm{ZnO}}^{-}$) with holes in the VB of g-C₃N₅ (${\mathrm{h}}_{\mathrm{CN}}^{+}$). The remaining holes in the VB of ZnO (${\mathrm{h}}_{\mathrm{ZnO}}^{+}$) and the electrons in the CB of g-C₃N₅ (${\mathrm{e}}_{\mathrm{C}\mathrm{N}}^{-}$) then participate in redox reactions, resulting in the degradation of MB pollutants into water and carbon dioxide. This photodegradation process is summarized by Equations (3)–(8) [20].

$${\mathrm{g-C}}_3{\mathrm{N}}_5+\mathrm{hv}\to {\mathrm{e}}_{\mathrm{CN}}^{-}+{\mathrm{h}}_{\mathrm{CN}}^{+}$$

(3)

$$\mathrm{ZnO}+\mathrm{hv}\to {\mathrm{e}}_{\mathrm{ZnO}}^{-}+{\mathrm{h}}_{\mathrm{ZnO}}^{+}$$

(4)

$${\mathrm{e}}_{\mathrm{ZnO}}^{-}+{\mathrm{h}}_{\mathrm{CN}}^{+}\to \mathrm{Z}-\mathrm{scheme} \mathrm{pathway} \mathrm{recombination}$$

(5)

$${\mathrm{e}}_{\mathrm{CN}}^{-}+{\mathrm{O}}_2\to {\mathrm{O}}_2^{-}$$

(6)

$${\mathrm{h}}_{\mathrm{ZnO}}^{+}+{\mathrm{H}}_2\mathrm{O} \to \mathrm{OH}+{\mathrm{H}}^{+}$$

(7)

$${\mathrm{O}}_2^{-}/\mathrm{OH}/{\mathrm{h}}_{\mathrm{ZnO}}^{+}+\mathrm{MB}\to \mathrm{intermediates}{+\mathrm{CO}}_2+{\mathrm{H}}_2\mathrm{O}$$

(8)

4. Conclusions

In this study, a honeycomb-like ZCN heterojunction with a Z-scheme energy band configuration was synthesized over g-C₃N₅ in situ using a combination of calcination and hydrothermal methods. The ZCN photocatalyst exhibited significantly enhanced kinetics (k = 0.0433 min⁻¹) for MB degradation, which were 1.9 times that of pure ZnO (k = 0.0223 min⁻¹). The remarkable photocatalytic performance of ZCN can be attributed to two key factors. First, the ZCN heterojunction facilitates efficient photon absorption up to 710 nm, which broadens the adsorption edge of ZnO to the visible light region. Second, the Z-scheme heterojunction effectively suppresses charge carrier recombination, while it preserves the strong redox capacity of both g-C₃N₅ and ZnO, as confirmed by ESR experimental data and DFT simulations. This study presents a potential approach for the design and development of highly efficient Z-scheme photocatalysts for environmental remediation. Nevertheless, this study was conducted under laboratory conditions without testing on actual wastewater, which highlights the need for future research to address this gap.