In the past several decades, the use of molybdates as functional materials has been attracting great attention because of their favorable properties and complex chemistry. A variety of molybdates have been synthesized for different purposes. For example, divalent transition metal molybdates NiMoO₄ and CoMoO₄ demonstrated excellent good rate capability and high specific capacitances when used as electrodes for electrochemical capacitors [1, 2]. Meanwhile, divalent metal molybdates with the general formula MMoO₄ (M = Ca, Mg, Zn) are semiconductors that have drawn considerable interest because of their potential applications in luminescence, catalysis, capacitors, and scintillation detectors [3-5]. Among divalent metal molybdates, the synthesis and use of zinc molybdates are attractive because zinc cations are recognized as 'inorganic nodes' for the design of porous inorganic compounds and metal semiconductors [6]. Zinc molybdates possess complex structures and their general chemical formula can be expressed as nZnO·mMoO₃ (n = 3, 2, or 1 and m = 2, 3, or 1), in which the molybdenum cation has a valence of +6 or +4. Compounds with different stoichiometric ratios, including ZnMoO₄, Zn₂Mo₃O₈, and Zn₃Mo₂O₉, have also been reported [7]. In particular, ZnMoO₄ crystals show similarities with Aurivillius-type compounds, and have different crystalline phases, such as α-ZnMoO₄ (JCPDS No. 35-0765), β-ZnMoO₄ (JCPDS No. 25-1024), and ZnMoO₄∙0.8H₂O (JCPDS No.25-1025). Among the three phases, α-ZnMoO₄ has a triclinic structure and is characterized by the structural group [ZnO₆/MoO₄] with tetrahedral MoO₄^2- anions. The other two phases possess monoclinic structures described by distorted [ZnO₆]/[MoO₆] groups with octahedral MoO₆^6- anions [8]. These materials are semiconducting; triclinic α-ZnMoO₄ was reported to have a bandgap of 3.3 eV, while monoclinic β-ZnMoO₄ was characterized with bandgaps of around 3.13-3.51 eV [9] and 2.74-2.85 eV [8]. To date, ZnMoO₄ has been used to effectively treat methyl orange [10], Victoria blue R, and phenol [8] under visible or ultraviolet (UV) light.

China has the largest source of molybdenum in the world (> 19.6 Mt) [11]. The vast majority of the molybdenum is used in the production of high-strength steels employed in bridges, power plants, and pipelines. An alternative value-added use of molybdenum is in decontamination materials. In the present study, β-ZnMoO₄ and graphitic carbon nitride (g-C₃N₄)-modified β-ZnMoO₄ are synthesized to explore the feasibility of using these materials for the photocatalytic degradation of organic pollutants. We combine β-ZnMoO₄ with g-C₃N₄ because it is a low-cost light-active material that has been reported to enhance the photocatalytic performance of composites via a type Ⅱ or Z-scheme mechanism [12]. Meanwhile, g-C₃N₄ itself is a photocatalyst that is responsive to visible light. The introduction of g-C₃N₄ to β-ZnMoO₄ may improve the photocatalytic performance of the resulting ZnMoO₄/g-C₃N₄ composite in the visible light region via the formation of heterogeneous junctions. Multiple approaches to synthesize g-C₃N₄ composites have been reported, including in situ hydrothermal treatment [13-16], ultrasonication [17-20], one-step heating [21-28], and precipitation [29, 30]. Han et al. [31] prepared g-C₃N₄-hybridized TiO₂ nanofibers by an electrospinning process combined with a heat-etching method. Liao and colleagues fabricated a graphene oxide/g-C₃N₄ composite photocatalyst using a sonochemical approach [19]. Although various g-C₃N₄-based composites have been prepared by the above-mentioned methods, few studies have investigated the effect of the synthesis method on the performance of the resulting materials.

Therefore, in the present study, we use different methods to synthesize composite catalysts, and explore the effect of synthesis method on the performance of catalysts in which β-ZnMoO₄ and g-C₃N₄ serve as the host material and dopant material, respectively. Sulfamethazine (SMZ), as one of the most commonly used antibiotics, is selected as a model pollutant because of its presence in natural water streams and adverse effects on living beings [32, 33]. The performance of the composite photocatalysts is tested by SMZ degradation experiments. The kinetics and degradation pathways of SMZ over the course of degradation are accordingly studied. We hope that this study will not only deepen our understanding of ZnMoO₄-based photocatalysts, but also give clues to find simple and effective strategies to obtain high-performance composite photocatalysts.

Analytical-grade chemicals including zinc nitrate hexahydrate, sodium molybdate dihydrate, p-benzoquinone (BQ), triethanolamine (TEOA), sodium azide (NaN₃), t-butanol (TBA), formic acid, and ethanol, along with methanol (high-performance liquid chromatography (HPLC) grade) were all purchased from Shanghai Sinopharm Chemical Reagent Co., Ltd. Melamine (99%) and SMZ (99%) were purchased from Shanghai Aladdin Industrial Corporation. Ultrapure water was used throughout this work.

β-ZnMoO₄ powder was synthesized by hydrothermal reaction under various conditions [8]. For the β-ZnMoO₄ powder obtained at 180 ℃ (denoted β-ZnMoO₄-180), Zn(NO)₃∙6H₂O and Na₂MoO₄∙2H₂O with a molar ratio of 1: 1 were used as the starting materials, which were separately dissolved in ultrapure water and then mixed under constant stirring. The mixture was transferred to a Teflon-lined autoclave and heated at 180 ℃ for 12 h. The crude powder was collected, washed with water and ethanol several times, and then dried at 60 ℃ for 12 h to give β-ZnMoO₄-180. For the β-ZnMoO₄ powder obtained at 280 ℃ (β-ZnMoO₄-280), a 1: 2 molar ratio of Mo to Zn was used and the hydrothermal conditions were 280 ℃ for 24 h. We obtained g-C₃N₄ by a thermal polycondensation process in a tube furnace [34]. The precursor melamine powder was transferred into a ceramic crucible and heated at a rate of 5 ℃/min to 300 ℃ for 2 h; further deammoniation treatment was subsequently performed at 520 ℃ for 2 h.

The g-C₃N₄-modified ZnMoO₄ (β-ZnMoO₄/g-C₃N₄-X%) samples were prepared by in situ hydrothermal synthesis. In this case, g-C₃N₄ powder was added to the Na₂MoO₄ solution with a g-C₃N₄/Mo molar ratio of 1: 100, 2: 100, 3: 100, 4: 100, 5: 100, or 8: 100, and then the subsequent steps were the same as those used to synthesize β-ZnMoO₄. In addition, β-ZnMoO₄/g-C₃N₄-X%-U samples were also prepared by an ultrasonication method. An appropriate amount of g-C₃N₄ was completely dispersed in methanol assisted by ultrasonication. The as-prepared β-ZnMoO₄ powder was subsequently added and the mixture was stirred in a fume hood for 24 h. The resulting powder was collected by centrifugation and then dried at 60 ℃ for 12 h.

The photocatalytic experiments were carried out under the irradiation of an 18-W light-emitting diode (LED). The central wavelength of the LED was 380 nm, and it was placed at the bottom of a double-jacketed glass reactor. A given amount of catalyst (100 mg) was added to SMZ aqueous solution (200 mL, 2 mg/L), and the mixture was mechanically stirred at 25 ℃. Prior to irradiation, the mixture was constantly stirred under dark conditions for 30 min to ensure the adsorption equilibrium was reached. At specific time intervals, 2 mL aliquots were withdrawn. To remove the particles, each sample was filtered through a 0.45 μm filter. To investigate the photodegradation mechanism, inhibition tests were carried out using different radical scavengers. BQ (0.2 mmol/L), TBA (0.2 mol/L), NaN₃ (0.012 mol/L), and TEOA (0.1 mol/L) were used as scavengers for superoxide radicals (·O₂^-), hydroxyl radicals (·OH), singlet oxygen (¹O₂), and holes (h⁺), respectively [22].

Phase identification of samples was conducted by X-ray diffraction (XRD; Shimadzu X-ray 6000). The morphologies of samples were obtained by field-emission scanning electron microscope (FE-SEM; Zeiss SIGMA) and transmission electron microscopy (TEM; JEOL JEM-2010). Fourier transform infrared (FTIR) spectra were acquired by a Nicolet Avatar 360 spectrometer with KBr pellets in the 4000-400 cm^-1 region. Specific surface areas were determined with the Brunauer-Emmett-Teller (BET) equation using an adsorption apparatus (Micromeritics ASAP2460, USA). X-ray photoelectron spectroscopy (XPS) was performed on a Thermo Scientific Escalab 250Xi spectrometer. The granulometric distribution of each sample was determined by a particle size analyzer (Microtrac S3500). UV-vis diffuse reflection spectroscopy (DRS) was performed on a Shimadzu UV-2550 spectrophotometer using BaSO₄ as the reference. The concentration of ammonia nitrogen (NH₄⁺-N) in each sample was measured by Nessler's reagent spectrophotometry. The Mott-Schottky analysis and electrochemical impedance spectroscopy (EIS) were performed with a three-electrode configuration in 0.1 mol/L Na₂SO₄ solutions, with a saturated calomel electrode (SCE) as the reference electrode. For these measurements, working electrodes were prepared by mixing and kneading the as-prepared catalysts with acetylene black (8:1) and binder (polytetrafluoroethylene, 35 wt%) along with several drops of ethanol. The resulting slurry was rolled into film using a rolling machine. The film was cut into a square (1×1 mm), which was then pressed onto a titanium mesh. Measurements were performed on an electrochemical workstation (CHI-660b, Shanghai Chenhua Co., Ltd, China).

The degradation products of SMZ were determined by HPLC (LC-15C, Shimadzu) using a system equipped with a UV absorbance detector. HPLC was conducted on a C18 reversed-phase column using a mobile phase of 30:70 methanol:water (containing 0.1% formic acid), with a flow rate of 1 mL/min, and detection wavelength of 270 nm. The identification of intermediates was performed by liquid chromatography-mass spectrometry (LC-MS; Agilent 6460 triple quadrupole mass spectrometer) with an electrospray ionization source. A reversed-phase column (Agilent Eclipse Plus C18) was used with a flow rate of 0.3 mL/min, and the elution was accomplished with 20:80 methanol:water (containing 0.1% formic acid). The total organic carbon (TOC) decay of solutions was determined on an Elementar Vario TOC analyzer. The inorganic ions in the solutions during photocatalytic degradation were measured by ion chromatography (Metrohm 930).

Fig. 1(a) presents the XRD patterns of ZnMoO₄ crystals prepared under different conditions. All peaks were indexed to the crystalline structure of β-ZnMoO₄ (JCPDS No. 25-1024) [8]. No diffraction peaks from other crystalline phases were detected, indicating that β-ZnMoO₄ of good quality was successfully synthesized by the two hydrothermal procedures used. Fig. 1(b) shows the XRD patterns of the β-ZnMoO₄/g-C₃N₄ composites prepared with different g-C₃N₄/Mo molar ratios. The pure g-C₃N₄ sample presented two obvious peaks consistent with g-C₃N₄. The (002) peak was typical of interplanar stacking in conjugated aromatic systems, and the other one at 13° (100) was associated with the in-plane structural packing motif of the tri-s-triazine units [35]. For the β-ZnMoO₄-180/g-C₃N₄ composites, the main peak at 27.5° from g-C₃N₄ was observed, and the intensities of the peaks increased with the g-C₃N₄/Mo ratio. Compared with that of the pristine g-C₃N₄, the main (002) peak of g-C₃N₄ in the composite shifted slightly (by 0.3°), indicating the interlayer spacing decreased. This observation was also reported by other scholars [31]; a plausible reason for this phenomenon is that the interaction of g-C₃N₄ with ZnMoO₄ influences its interplanar spacing.

SEM images of the samples revealed that the β-ZnMoO₄-180 sample produced using a 1:1 Mo/Zn molar ratio had a sheet-like structure, and the sheets further assembled into flower-like shapes (Fig. 2(a)). Meanwhile, β-ZnMoO₄-280 prepared with a 1: 2 Mo/Zn molar ratio has an irregular shape with larger corners (Fig. 2(d)), which is similar to the morphology of β-ZnMoO₄ crystals reported by Jiang et al. [8]. The different initial contents of the starting materials and the varied hydrothermal conditions affected the rate and direction of crystal growth in the samples, leading to the formation of β-ZnMoO₄ crystals with dissimilar morphologies. In addition, the surfaces of these microcrystals are very smooth, with low BET surface areas of 0.883 m²/g for β-ZnMoO₄-180 and 0.491 m²/g for β-ZnMoO₄-280. Fig. 2(b) and (c) show a selected-area electron diffraction (SAED) pattern and high-resolution transmission electron microscopy (HRTEM) image of the β-ZnMoO₄-180 sample, respectively. The lattice spacings in Fig. 2(c) are 0.56, 0.47, and 0.34 nm, which correspond to the lattice spacings of the (010), (100), and (110) planes of the β-ZnMoO₄ crystals, respectively. The corresponding SAED pattern demonstrates that the β-ZnMoO₄-180 sample is monocrystalline. The SAED pattern of β-ZnMoO₄-280 (Fig. 2(e)) indicates this sample has a polycrystalline nature, and its HRTEM image (Fig. 2(f)) has lattice spacings of 0.36, 0.24, and 0.23 nm, corresponding to the lattice spacings of the (110), (002), and (200) planes, respectively. These results are in good accordance with those of the XRD pattern in Fig. 1(a). Fig. 2(g) and (h) show the morphology of the g-C₃N₄ prepared in this study. The pristine g-C₃N₄ existed in the form of small lumps, and its lamellar structure can be observed. This kind of structure makes g-C₃N₄ easy to split into smaller blocks during ultrasonication.

Fig. 3 illustrates the morphologies of g-C₃N₄-modified β-ZnMoO₄ prepared with different methods. For the samples prepared by the in situ hydrothermal method (Fig. 3(a), (c), and (e)), g-C₃N₄ particles were successfully loaded onto the surface of β-ZnMoO₄-180, whereas no g-C₃N₄ particles were observed for the β-ZnMoO₄-280/g-C₃N₄ sample. The FTIR spectra (see Fig. S1 in Supporting Information (SI)) of the β-ZnMoO₄-280/g-C₃N₄ samples also suggested that no g-C₃N₄ could be detected on the surface of the β-ZnMoO₄-280 crystals. This result is contradictory to the results of XRD in Fig. 1(b), in which the β-ZnMoO₄-280/g-C₃N₄ sample showed the characteristic peak of g-C₃N₄ at 27.5°. This observation suggests that the detected g-C₃N₄ might be incorporated into the structure of the composite, rather than modifying its surface. For the composites prepared by the ultrasonic method (Fig. 3(b), (d), and (f)), g-C₃N₄ particles were observed on the surface of the β-ZnMoO₄ crystals, and the loading of g-C₃N₄ particles increased with the g-C₃N₄/Mo ratio. This experimental result suggests that the g-C₃N₄ particles can attach to the surface of β-ZnMoO₄ crystals through the ultrasonic method. Therefore, extra experiments were carried out to examine why g-C₃N₄ was not observed on the β-ZnMoO₄-280/g-C₃N₄ sample. Fig. S2 in the SI shows the concentrations of NH₄⁺-N in the solutions after the hydrothermal treatment of g-C₃N₄ suspension for different periods. The inset in Fig. S2 reveals that suspended g-C₃N₄ particles still existed in the solution after treatment at 280 ℃ for 0.5 h, and the solution become clear after treatment for 1 h. Meanwhile, the NH₄⁺-N concentration in the solutions increased, indicating some of the nitrogen in g-C₃N₄ was converted into dissolved NH₄⁺-N. The inset proves that during hydrothermal treatment at 280 ℃, the decomposition of g-C₃N₄ particles progressed. Some of the particles might be incorporated into β-ZnMoO₄ crystals in the early stage of the hydrothermal process, while the rest of the g-C₃N₄ particles eventually decomposed, preventing the existence of surface-attached g-C₃N₄ particles. This may explain why the characteristic diffraction peak of g-C₃N₄ was observed in the XRD pattern of the β-ZnMoO₄-280/g-C₃N₄ sample (Fig. 1(b)), but g-C₃N₄ particles were not observed on the surface of the samples in the SEM images (Fig. 3(e)).

XPS analysis was used to investigate the chemical composition of the as-prepared β-ZnMoO₄ and β-ZnMoO₄/g-C₃N₄ composites. As shown in Fig. 4, the spin-orbit splitting of the Mo 3d level gave rise to 3d_5/2 and 3d_3/2 levels with an energy separation of 3.2 eV, and the binding energies of 232.6 and 235.8 eV corresponded to Mo 3d_5/2 and Mo 3d_3/2, respectively. The Mo 3d_5/2 peak revealed that Mo possessed a hexavalent oxidation state and the Zn 2p_3/2 peak at 1021.3 eV revealed Zn had a divalent oxidation state [36]. The characteristic binding energy of 530.1 eV for O 1s revealed O was present as a metal oxide. Regarding the g-C₃N₄-modified β-ZnMoO₄ prepared by the in situ hydrothermal synthesis method, there was no evident difference between the binding energies of Zn, Mo, and O with or without g-C₃N₄. In Fig. 4(c) and (d), the small deviation of Mo 3d and O 1s may be caused by the different experimental conditions. Jiang et al. [8] reported a similar observation in their XPS analysis of different β-ZnMoO₄ samples.

The particle size distribution curves of β-ZnMoO₄ with different morphologies are shown in Fig. 5. Based on the distribution curves, the β-ZnMoO₄ samples prepared under different hydrothermal reaction conditions with flower-like and irregular shapes possessed very similar average particle sizes (D₅₀) of 15.02 and 17.80 μm, respectively. Finer particles were found for the β-ZnMoO₄-280 sample, as can also be observed in the SEM images (Fig. 2). The β-ZnMoO₄-180 sample had a narrower size distribution than that of β-ZnMoO₄-280 because of its more regular sheet-like structure.

The adsorption and photodegradation of SMZ by pristine β-ZnMoO₄ samples were studied. Fig. 6(a) reveals that the two kinds of β-ZnMoO₄ show very weak capability to uptake SMZ in the dark. The small BET surface area of the β-ZnMoO₄ particles explains their poor SMZ adsorption ability. Fig. 6(b) shows the photocatalytic activities of the β-ZnMoO₄ and g-C₃N₄ particles for SMZ degradation. After irradiation for 3 h, the degradation efficiencies of the β-ZnMoO₄-180 sample with flower-like morphology, β-ZnMoO₄-280 crystals with irregular corners, and g-C₃N₄ were 71%, 93%, and 100%, respectively. With increasing reaction temperature and time, the crystallinity of β-ZnMoO₄ increased and its defect content decreased, leading to the enhanced photocatalytic activity of the β-ZnMoO₄-280 sample over that of β-ZnMoO₄-180 [8, 37-39].

The photocatalytic activities of the β-ZnMoO₄/g-C₃N₄ composites are presented in Fig. 6(c) and (d). The photocatalytic degradation of SMZ by the composites follows pseudo-first-order kinetics. In the case of g-C₃N₄-modified β-ZnMoO₄-180 (Fig. 6(c)), the photodegradation experiments revealed that this composite showed higher photocatalytic activity than that of β-ZnMoO₄-180. When the g-C₃N₄/Mo ratio was 3:100, the maximum photodegradation rate was achieved. Fig. 6(d) illustrates that the β-ZnMoO₄-280/g-C₃N₄ composites showed lower photocatalytic activities than that of β-ZnMoO₄-280; namely, the introduction of g-C₃N₄ lowered the performance of this catalyst. Fig. 6(e) and (f) show the photocatalytic activities of β-ZnMoO₄ and g-C₃N₄-modified β-ZnMoO₄ photocatalysts prepared by the ultrasonic method. With increasing g-C₃N₄ content, the photodegradation abilities of the two kinds of β-ZnMoO₄ improved slightly. For the β-ZnMoO₄-180/g-C₃N₄-U samples (Fig. 6(e)), considerably enhanced degradation of SMZ was only observed for the composite sample with 8% g-C₃N₄. For the β-ZnMoO₄-280 samples (Fig. 6(f)), modification with g-C₃N₄ by the ultrasonic method only weakly influenced the photocatalytic degradation of SMZ. The above results reveal that all the surface-modified β-ZnMoO₄/g-C₃N₄ samples were better than the pristine β-ZnMoO₄ samples at SMZ photodegradation. In particular, the β-ZnMoO₄-180/g-C₃N₄ sample with 3% g-C₃N₄ prepared by the in situ hydrothermal method exhibited high photodegradation efficiency, indicating this synthesis method allowed good bonding between the β-ZnMoO₄ and g-C₃N₄ particles. Retardation of photoactivity following g-C₃N₄ addition was only found for the β-ZnMoO₄-280/g-C₃N₄ sample prepared by the in-situ hydrothermal method. Different from the surface-attached g-C₃N₄, the structural incorporation of g-C₃N₄ into β-ZnMoO₄-280 may decrease its crystallinity, resulting in its poor photocatalytic activity.

The growth mechanism of the β-ZnMoO₄ and β-ZnMoO₄/g-C₃N₄ composites synthesized by the hydrothermal method are shown in Fig. 7. The growth of zinc molybdate crystals mainly involves the following stages [8, 40]: (a) nucleation of the first clusters or nuclei, (b) self-assembly of small crystallites, and (c) crystal growth. The electrostatic attraction between Zn²⁺ and MoO₄^2- promotes the nucleation process that produces the initial β-ZnMoO₄ precipitates or nuclei. The hydrothermal conditions increase the effective collision frequency of the nanoparticles in the suspension. At a reaction temperature of 180 ℃, the nanoparticles aggregate to form plate-like β-ZnMoO₄ microcrystals. At 280 ℃, the intensified self-aggregation of particles compared with that at 180 ℃ resulted in the formation of β-ZnMoO₄ microparticles with irregular shape. When using the in situ hydrothermal method to synthesize the g-C₃N₄-modified β-ZnMoO₄ composites, temperature-dependent results were obtained because of the different growth pathways of β-ZnMoO₄ crystals at different temperatures. While g-C₃N₄ bound to the surface of flower-like β-ZnMoO₄-180, it was included inside the β-ZnMoO₄-280 crystals of the β-ZnMoO₄-280/g-C₃N₄ sample. The structural incorporation of g-C₃N₄ into the composite may lower the crystallinity of β-ZnMoO₄-280, resulting in its decreased photodegradation ability.

To elucidate the photodegradation mechanism of SMZ by β-ZnMoO₄ and the β-ZnMoO₄/g-C₃N₄ composites, photocatalytic experiments were carried out in the presence of different radical scavengers (Fig. 8). First, TBA, a widely used ·OH scavenger, was added to investigate whether it affected the degradation of SMZ. For both β-ZnMoO₄-180 and β-ZnMoO₄-180/g-C₃N₄ samples, TBA caused only a slight decrease of photocatalytic ability. When a scavenger for ¹O₂ (NaN₃) was added to the photocatalytic experiments, similar results were obtained. These results indicated that ·OH and ¹O₂ were not the primary oxidative species in the photodegradation of SMZ. By contrast, the addition of TEOA or BQ markedly retarded SMZ photodegradation. TEOA and BQ are scavengers for h⁺ and ·O₂^-, respectively, indicating that the photodegradation mechanism of the samples was mainly associated with h⁺ and ·O₂^-.

UV-vis DRS of β-ZnMoO₄ are presented in Fig. 9(a). The absorption edge of ZnMoO₄ was in the visible region, corresponding to band gaps (E_g) of 3.078 eV for β-ZnMoO₄-180 and 3.038 eV for β-ZnMoO₄-280. The conduction band (CB) of the β-ZnMoO₄ samples was determined by Mott-Schottky analysis. As shown in Fig. 9(b), the flat-band (FB) potential of β-ZnMoO₄-180 was estimated to be -1.08 V vs. SCE, and the positive slope of this linear plot suggests the n-type character of this sample. It has been reported that the CB potentials of n-type semiconductors are 0.1-0.2 V more negative than their FB potential [41]. Assuming that the CB edge is 0.1 V more negative than the FB potential, the estimated CB edge for β-ZnMoO₄-180 is -1.18 V vs. SCE (-0.94 V vs. normal hydrogen electrode (NHE)), and the estimated valence band (VB) is 1.90 V vs. SCE (2.14 V vs. NHE). Likewise, the estimated CB edge for β-ZnMoO₄-280 is -1.27 V vs. SCE (-1.03 V vs. NHE), and its VB is 1.77 V vs. SCE (2.01 V vs. NHE). UV-vis DRS and Mott-Schottky plots for g-C₃N₄ are shown in Fig. S3 and S4 in the SI; its determined CB and VB positions are -1.04 and 1.77 V vs. NHE, respectively. The band structures of β-ZnMoO₄ and g-C₃N₄ are summarized in Fig. 9(c).

The EIS analysis of β-ZnMoO₄-180 and β-ZnMoO₄-180/g-C₃N₄-3% was performed to investigate the effect of g-C₃N₄ modification on electrode behavior. Generally, a smaller arc size in a Nyquist plot indicates lower charge transfer resistance on an electrode surface [15, 42]. The arc radius of the Nyquist plot of β-ZnMoO₄-180/g-C₃N₄-3% was smaller than that of the β-ZnMoO₄-180 sample under illumination (Fig. S5 in the SI), suggesting that the separation and transfer efficiency of photoinduced electron-hole pairs was increased through an interfacial interaction between β-ZnMoO₄-180 and g-C₃N₄ particles in the composite. The proposed photocatalytic degradation process of the β-ZnMoO₄/g-C₃N₄ composite is illustrated in Fig. 9(c). The difference of chemical potential between the two semiconductor units results in band bending at the heterogeneous interface in the composite. This band bending induces a built-in electric field, which can drive the migration of photogenerated electrons and holes in opposite directions. Under illumination, photogenerated holes would form when electrons were excited from the VB to the CB of both β-ZnMoO₄ and g-C₃N₄. The photogenerated electrons in g-C₃N₄ could easily migrate to the β-ZnMoO₄ surface. The CB potential of β-ZnMoO₄ is more negative than the redox potential of the O₂/·O₂^- couple (-0.33 eV vs. NHE), which can produce ·O₂^- radicals from dissolved O₂. At the same time, the holes generated in the VB of β-ZnMoO₄ could migrate to g-C₃N₄. However, the VB potential of g-C₃N₄ is more negative than the redox potential of the ·OH/OH^- couple (1.99 eV vs. NHE), thus it cannot lead to the production of ·OH from the holes [13], which was also demonstrated in the photocatalytic experiments with radical scavengers (Fig. 8(b)). Therefore, the heterogeneous interface in the β-ZnMoO₄/g-C₃N₄ composite markedly promoted the separation of photogenerated electron-hole pairs, thereby improving its SMZ photodegradation efficiency. For the β-ZnMoO₄-180/g-C₃N₄ samples, the in situ hydrothermal synthesis ensured favorable bonding between the g-C₃N₄ particles and β-ZnMoO₄-180 sheets. As a result, the highest degradation efficiency of SMZ was observed for the composite with 3% g-C₃N₄ (Fig. 6(c)). By contrast, for the samples prepared by the ultrasonic method, the bonding of the g-C₃N₄ particles on the surface of the β-ZnMoO₄-180 sheets may be not as strong as that of the samples prepared by the in situ hydrothermal method. Therefore, a larger g-C₃N₄ content of 8% was needed to achieve an evident enhancement effect of g-C₃N₄ on SMZ degradation for composites prepared by the ultrasonic method.

The photodegradation of SMZ and the formation of several intermediates were studied by LC-MS analysis. Fig. 10(a) shows the total ion chromatogram obtained after 60 min of photodegradation by β-ZnMoO₄-180/g-C₃N₄. In this study, a molecular ion peak at a mass-to-charge (m/z) of 279 was observed at 5.265 min, which was attributed to SMZ. Seven products with different m/z ratios, as indicated by peak A to H in Fig. 10a, were identified and are labeled in Table S1. Fig. 10(b) reveals that the TOC content of the solution decreased by 20% after 9 h of photocatalytic degradation, which indicated that SMZ was mineralized; that is, inorganic ions such as SO₄^2- and NO₃^- may be generated as the final degradation products. The LC-MS analysis results confirmed that SO₄^2- and NO₃^- were present in the solution with concentrations of 1.575 and 0.175 mg/L, respectively, after 6 h of photocatalytic degradation.

A number of studies have determined that degradation of sulfonamides occurs through cleavage at various positions [43], as shown in Fig. 11. In this study, compound d (m/z = 215) was observed, which is most likely a product of -SO₂-extrusion (cleavage at the γ and δ positions in Fig. 11). Several possible pathways for the formation of extrusion products with m/z = 215 have been proposed [33, 44, 45]. Subsequently, the h⁺ and ·O₂^- attacked the carbon-nitrogen bond between the benzene and pyrimidine rings in compound d, resulting in two products, a and c (Fig. 11), as also identified by Guo et al [44]. The product with m/z = 265 identified at 1.245 min (compound b) and that with m/z = 239 identified at 2.326 min (compound e) correspond to products of amino group extrusion (cleavage at position β) or methyl group extrusion. Another pathway involves the benzene and pyrimidine rings of SMZ being attacked by reactive oxygen species to cause hydroxylation, which has been reported in other studies [33]. Moreover, the continuous attack of the hydroxylated SMZ by reactive species results in the loss of an amino or methyl group, corresponding to the product with m/z = 283 (compound f). Intermediate f might also be formed by the hydroxylation of compound b with ·OH. Subsequently, these intermediates may decompose into smaller molecules, ending up as environmentally benign compounds, such as SO₄^2- and NO₃^-, which were detected by ion chromatography.

In summary, β-ZnMoO₄ catalysts with different morphologies were synthesized via the hydrothermal method. Both the in situ hydrothermal and ultrasonic methods were used to fabricate g-C₃N₄-modified β-ZnMoO₄ composites. Our results indicated that β-ZnMoO₄ microparticles with a unique microstructure were good platforms for the introduction of visible-light-responsive g-C₃N₄. The pristine β-ZnMoO₄-280 sample obtained showed higher photocatalytic activity than β-ZnMoO₄-180. In the case of β-ZnMoO₄-180/g-C₃N₄ composites, the surface-modified samples all showed enhanced photocatalytic activity relative to that of the pristine β-ZnMoO₄-180 sample. By contrast, the structural incorporation of g-C₃N₄ into β-ZnMoO₄ crystals prepared at 280 ℃ lowered their crystallinity, thereby deteriorating the performance of these composite catalysts. Study of the photocatalytic mechanism of the composites revealed that h⁺ and ·O₂^- were the dominant oxidative species in the SMZ photodegradation process. The enhanced photocatalytic performance of the composites was attributed to the higher separation efficiency of the photogenerated electron-hole pairs at their heterogeneous interfaces. Compared with the ultrasonic method, the hydrothermal synthesis at 180 ℃ resulted in a more favorable interaction between g-C₃N₄ and β-ZnMoO₄, and the optimal g-C₃N₄ content was 3%. SMZ was effectively photodegraded by the β-ZnMoO₄/g-C₃N₄ composites, and the degradation mechanism involved the extrusion of -SO₂-and amino groups. This work highlights the importance of selecting an appropriate method to prepare composite photocatalytic materials based on g-C₃N₄. The abundant molybdenum resources in China may be used to prepare catalytic materials for various applications in environmental and energy fields.