Abstract

In this work, graphitic carbon nitride (g-C3N4)/titanium dioxide (TiO2) nanoparticles with heterostructures were synthesized in situ from a mixture of melamine and peroxo-titanium complexes in a calcination process. The TiO2 nanoparticles are well-dispersed on the g-C3N4 nanosheets. The prepared TiO2/g-C3N4 composites have a heterostructure and excellent photocatalytic activity for decomposing methylene blue (MB) under visible light irradiation. The as-obtained g-C3N4 embroiled with TiO2 has a much larger surface area than its components (66.7 and 6.6 m2·g−1 for TiO2 and g-C3N4 against 95.5–143.8 m2·g−1 for the composite, respectively). It enhances the separation of photo-generated charge carriers. The TiO2/g-C3N4 photocatalytic degradation of MB was investigated in aqueous heterogeneous suspensions. The experimental kinetic data for the photocatalytic process follow the pseudo-first-order kinetic model. Furthermore, TiO2/g-C3N4 retains high photocatalytic activity after four reaction cycles. In addition to prompt removal of the color, the TiO2/g-C3N4 photocatalyst can oxidize MB almost completely to final oxidation products. The pathway of MB decomposition was also addressed. Additionally, the TiO2/g-C3N4 photocatalytic system was employed to eliminate other typical organic pigments, such as malachite green, methyl blue, and methyl red. The TiO2/g-C3N4 material, with remarkable dye degradability, is a promising catalyst in industrial textile treatment and can find applications in light-harvesting systems.

1. Introduction

Recently, graphitic carbon nitride (g-C3N4) has become prominent as a stable photocatalyst and gained significant attention [1] owing to its relevant bandgap (2.7 eV) [2] and, therefore, can be active under visible light. In addition, the negative potential (−1.12 eV vs. normal hydrogen electrode (NHE)) of the conduction band (CB) enables g-C3N4 to form heterojunctions with other semiconductors. Despite its disadvantages, such as small surface area and high recombination rate of photoinduced charge carriers, g-C3N4 is still a potential candidate for hybrid heterostructures with wide bandgap semiconductors [3].

Besides various semiconducting materials, such as SnO2, ZnO, ZnS, WO3, and V2O5 [4], TiO2 has attracted tremendous attention in photocatalytic applications because of its low cost, high stability, nontoxic nature, and environmental friendliness. Although TiO2 has been used in several photocatalytic processes of wastewater treatment and air purification [5], its wide bandgap (3.2 eV) [6] limits its application only to the ultraviolet region of visible light, which is around only 4% of the solar spectrum. To overcome this disadvantage, doping other nonmetallic or trace metallic elements, such as N, P, Fe, Ni, and V, to develop composite materials with narrowband energy has been studied intensively [4, 7, 8]. Numerous studies have presented the feasibility of synthesizing TiO2/g-C3N4 photocatalysts [9]. Using g-C3N4 with TiO2 to form a heterojunction is one of the efficient approaches for either narrowing bandgap or enhancing photoinduced electron–hole pair separation. For example, Liu et al. [10] synthesized TiO2/g-C3N4 from TiO2 and urea and utilized this material to promote the photocatalytic reduction of U(VI) in water. Zhang et al. [11] and Alcudia-Ramos et al. [12] fabricated TiO2/g-C3N4 from tetrabutyl titanate or TiOSO4 and urea to enhance photocatalytic hydrogen evolution under visible light. Titanium alkoxides, TiCl3, and TiCl4 in a hydrochloric acid aqueous solution or Ti(SO4)2 in a sulfuric acid aqueous solution are mainly used as a titanium source [11, 12]. However, titanium alkoxides, such as tetra-isopropoxide, are highly unstable, flammable, and readily hydrolyze in a moist atmosphere. Moreover, titanium chlorides in these acidic solutions are difficult to handle because of their aggressive and toxic nature. For solving these issues, insensitive approaches have been investigated in developing water-soluble titanium complexes, such as peroxo-titanium oxalate and peroxo-titanium complexes [13], for synthesizing titanium-dioxide-based materials [14]. To our best knowledge, reports on the TiO2-based materials synthesized via water-soluble titanium species are limited. Therefore, we developed a strategy for preparing TiO2/g-C3N4 from a water-soluble titanium complex and melamine solution to form a homogeneous mixture in this study. The peroxo–hydro titanium complexes were synthesized from titanium oxide in a hydrothermal process, followed by the formation of water-soluble titanium complexes with hydrogen peroxide. TiO2/g-C3N4 was then prepared by evaporating the homogeneous complex solution, followed by calcination. The photocatalytic activity of TiO2/g-C3N4 was studied via the kinetics, mineralization, photocatalytic mechanism, and decomposition pathway of methylene blue.

2. Experimental

2.1. Materials

Anatase (TiO2, 99%) and melamine (C3H6N6, ≥99.5%) were purchased from Merck. Potassium iodide (KI, ≥99%), sodium chloride (NaCl, ≥99.5%), sodium hydroxide (NaOH, ≥96%), hydrochloric acid (HCl, 38%), hydrogen peroxide (H2O2, ≥30%), and isopropanol (CH3CH(OH)CH3, ≥99.8%) were obtained from Xilong, China. Methylene blue (C16H18N3SCl·3H2O, MB), methyl red (C15H15N3O2, MR), methyl blue (C37H27N3Na2O9S3, MyB), malachite green (C28H30N2O3, ≥96.5%, MG), and p-benzoquinone (C6H4O2, ≥99%) were provided by Sigma–Aldrich. All the chemicals were of analytical grade and used as received without any further purification.

2.2. Synthesis of g-C3N4

According to Yan et al. [1], the g-C3N4 material was synthesized from melamine. Briefly, 10 g of melamine powder was weighed, placed in a mortar, and finely ground. The powder was put into a porcelain cup and calcined at 500°C for 4 hr with a heating rate of 5°/min. The powder was cooled to ambient temperature and finely grounded to obtain a yellow product, denoted as g-C3N4.

2.3. Synthesis of Peroxo-Titanium Complexes

According to Chau et al. [15] and Le et al. [16], the peroxo–hydroxo titanium(IV) complex was synthesized. Briefly, 0.25 g of TiO2 powder was dispersed into 12.5 mL of a 20 M NaOH solution under sonication for about 15 min. Then, the entire solution was transferred to a Teflon flask and heated at 130°C for 10 hr. After cooling to ambient temperature, the white solid was separated and rinsed with distilled water and a 0.1 M HCl solution to completely remove the alkali then dried at 80°C for 2 hr (around 0.189 g of dried TiO2 obtained). The 4.62 mg of hydrothermal-treated TiO2 was added 35 mL of 30% H2O2 at 90°C, stirring for 1 hr to obtain a clear yellow solution (the solution is stable at 10°C for several days). The obtained peroxo-titanium complexes solution has a composition of 4.62 mg TiO2/35 mL determined with the gravity method.

2.4. Synthesis of TiO2/g-C3N4

TiO2/g-C3N4 composites were synthesized via an aqueous solution process: a mixture of titanium peroxo-titanium complex solution (35 mL; 4.62 mg TiO2) and (0.00; 1.16; 4.62; 18.48 mg) with known TiO2/melamine mass ratios was sonicated for 1 hr, followed by calcination at 500°C for 4 hr to obtain a TiO2/g-C3N4 composite. The mixtures are denoted as (10/0)TiO2/g-C3N4, (8/2)TiO2/g-C3N4, (5/5)TiO2/g-C3N4, (2/8)TiO2/g-C3N4, and (0/10)TiO2/g-C3N4, where the fraction in the parentheses is the TiO2/melamine mass ratio. The synthetic diagram is shown in Scheme 1.

2.5. Material Characterization

X-ray diffraction (XRD) was performed on a D8 Advance device (Bruker, USA) with a Cu/Kα radiation source and λ = 0.154056 nm. Scanning electron microscopy (SEM) images were collected by using an SEM Hitachi-S-4800, Japan, scanning electron microscope. FT infrared (FTIR) spectra for the samples were obtained on an IRAffinity-1S-Shimadzu spectrometer (Japan). Specific surface areas were determined from N2 adsorption–desorption isotherms in liquid nitrogen at 77 K by using a Micromeritics ASAP 2020 analyzer. UV–vis DR spectra were obtained on a UV2600-Shimadzu spectrometer (Japan).

2.6. Catalytic Activity

The catalytic activity of the obtained materials was studied via the decomposition of methylene blue (10 ppm) on the catalyst (0.2 g/L). The mixture was stirred in the dark for 40 min to reach the adsorption/desorption equilibrium, then illuminated with an Osram, 160 W filament lamp (filter cutoff λ < 420 nm). The MB concentration in the supernatant was determined by using UV–vis spectroscopy at a maximum wavelength of 664 nm. The pH value of the dye solutions was adjusted with 0.1 M HCl or 0.1 M NaOH solutions. In addition, the concentration of malachite green was determined at 617 nm and methyl red at 521 nm. The UV–vis spectra were recorded on a Spectro UV-2650 spectrometer (Labomed, Inc., USA). The chemical oxygen demand (COD) was determined by using a Thermoreactor Lovibond ET108 analyzer according to the dichromate method. The intermediates that appear during MB degradation were extracted and accumulated by chloroform, then determined by the ion trap method in combination with EIS ionized mass spectroscopy on a LC/MSD Trap-SL ion trap mass spectrometer. ICP analysis for titanium was performed using a 7800 ICP-MS instrument (Agilent Technologies, Santa Clara, CA, USA).

3. Results and Discussion

Figure 1 presents the XRD patterns of the products synthesized under different conditions. Anatase (TiO2) reacts with NaOH to create several forms of sodium titanate at 130°C and 10 hr. The characteristic peaks of Na2Ti3O7 indexed as (011), (–311), (–313), and (–105); Na2Ti6O13 indexed as (201), (110), (401), (402), (313), (223), and (713) and Na2Ti9O19 indexed as (003), (006), (007), (60–1), and (020) according to JCPDS 31-1329, JCPDS 73-1398, JCPDS 33-1293, respectively, are present in the XRD patterns (Figure 1(a)). The sodium titanates after being rinsed with HCl until neutral filtrates change to anatase TiO2 (JCPDS 21-1272) (Figure 1(b)). The preparation of water-soluble peroxo-titanium complexes is associated with the formation of peroxo-titanium complexes, such as [Ti(O2)(OH)]+, [Ti(O2)(OH)2], and [Ti(O2)(OH)3], which dimerize and result in various oxo–peroxo–hydroxo di-titanium complexes, such as Ti2(O)(O2)2(OH)2, [Ti2(O)(O2)2(OH)3], and [Ti2(O)(O2)2(OH)4]2−. These dimeric complexes can undergo further self-condensation to form oligomers. As a result, polynuclear oxo–peroxo–hydroxo titanium complexes precipitate as [Ti2(O)(O2)2(OH)2]n [17] (Equations (1)–(4)).

In the synthesis of the TiO2/g-C3N4 composite, different peroxo-titanium complexes/melamine mass ratios were used. Figure 2(a) presents the XRD patterns of pristine g-C3N4, TiO2, and TiO2/g-C3N4 composites. For the pristine g-C3N4 and TiO2, the characteristic peaks of the g-C3N4 phase are observed on the XRD pattern of (0/10)TiO2/g-C3N4 at 13° and 27°, indexed as (100) and (002) (JCPDS 87-1526). These two peaks are likely assigned to the structure of the tri-s-triazine unit with interplanar spacing and the conjugated aromatic system [18]. This observation confirms the presence of g-C3N4 in the sample. The XRD pattern of (10/0)TiO2/g-C3N4 exhibits characteristic peaks at 25.2°, 37.8°, 48.0°, 53,8°, 62.6°, 70.3°, and 75.0° corresponding to (101), (004), (200), (105), (204), (220), and (215) Miller index, respectively (JCPDS 21-1272). These peaks suggest the TiO2 anatase form. The characteristic peaks of both TiO2 and g-C3N4 are observed in the XRD patterns of TiO2/g-C3N4 composites. The increasing amount of g-C3N4 augments the intensity of its characteristic peaks.

The textural properties of the obtained composites were studied via their nitrogen adsorption/desorption isotherms (Figure 2(b)). All the isotherm curves of g-C3N4, TiO2, and its composite belong to the IV type according to the IUPAC classification with the H3 hysteresis loop at relative pressures 0.6–1, suggesting the mesoporous structures formed from fine particles. The surface area of g-C3N4 and TiO2 is similar to what was reported previously [19, 20]. It is worth noting that the composites have a significantly larger surface area than g-C3N4 and TiO2 (95.5–143.8 m2·g−1 compared with 66.7 and 6.6 m2·g−1) (Table 1).

Table 1 shows that the surface area of the composite results mainly from the mesoporous area. Both g-C3N4 and TiO2 enhance the microporous and mesoporous spaces. The reason could be that the initial precursor of both titania and melamine in the liquid form is favorable for creating a homogeneous mixture. During heating, the gas products are generated to form the porous structure of TiO2/g-C3N4. The (2/8)TiO2/g-C3N4 sample exhibits the largest surface area (143.8 m2·g−1), higher than the TiO2/g-C3N4 samples synthesized from titanium n-butoxide and urea (62.1 m2·g−1) [21] and TiO2 powder and melamine (97.26 m2·g−1) [22].

Figure 3(a) presents the FTIR spectra of g-C3N4, TiO2, and TiO2/g-C3N4 composites. For bare g-C3N4 and TiO2, the vibration at 810 cm−1 is related to the C–N ring of the tri-s-triazine unit. The peaks at 1,238, 1,316, and 1,402 cm−1 are attributed to the stretching vibrations of the C–NH bond in g-C3N4. The peaks at 1,637 and 3,000–3,500 cm−1 are assigned to the stretching vibration of C = N [21, 23] and –N–H or = N–H bond [23, 24]. The stretching vibrations of the Ti–O–Ti bonds at 470 cm−1 are typical for TiO2 [21]. The vibrations of both g-C3N4 and TiO2 in the FTIR spectra of TiO2/g-C3N4 composites again confirm the composite formation between g-C3N4 and TiO2.

The (2/8)TiO2/g-C3N4 composite was also characterized from Raman spectra (Figure 3(b)3(d)). The characteristic peaks of g-C3N4 at 470, 543, 702, 746, 978, and 1,147 cm−1 are observed, which are similar to those as reported by Bai et al. [25]. For TiO2, the vibration modes at 144 cm−1 (Eg), 197 cm−1 (Eg), 399 cm−1 (B1g), 513 cm−1 (A1g), and 639 cm−1 (Eg) belong to anatase [26, 27]. For the TiO2/g-C3N4 composite, the vibrations of TiO2 and g-C3N4 observed in the Raman spectra confirm the coexistence of TiO2 and g-C3N4.

UV–vis diffuse reflectance spectra were employed to evaluate the bandgap of the material. Compared with pristine TiO2, TiO2/g-C3N4 displays a significant red shift because of the band alignment of coupled TiO2 and g-C3N4 (Figure 4(a)). This shift suggests that introducing g-C3N4 into TiO2 could enhance the visible light absorption available to TiO2/g-C3N4. The bandgap of g-C3N4 and TiO2/g-C3N4 was estimated from the Tauc equation [28]: α() = A(Eg)n/2, where α is the absorption coefficient; is the energy of photon; Eg is the bandgap of the semiconductor; and A is the constant. The index n depends on the properties of the transition of semiconductors: n = 1 for direct transition and n = 4 for indirect transition semiconductors. TiO2 and g-C3N4 are indirect bandgap semiconductors. Thus, the value of n for g-C3N4, TiO2, and TiO2/g-C3N4 is 4. The calculated bandgaps are 3.20, 2.80, 2.75, 2.70, and 2.65 eV for (10/0)TiO2/g-C3N4, (8/2)TiO2/g-C3N4, (5/5)TiO2/g-C3N4, (2/8)TiO2/g-C3N4, and (0/10)TiO2/g-C3N4, respectively (Figure 4(b)). It is possible that g-C3N4, a nitrogen-rich molecule, also acts as a nitrogen source to form TiO2−xNx. The substitution of N in the TiO2 lattice results in a visible light response owing to an additional N level above the TiO2 VB that triggers the bandgap to narrow to 0.55 eV. These results confirm that the formation of heterojunction between g-C3N4 and TiO2 could narrow the bandgap of pristine TiO2 and enhance visible light availability.

The elemental composition in the composite was studied by EDX-mapping. Figure 5 illustrates the EDX-mapping of the (2/8)TiO2/g-C3N4 sample. The electron image shows that this sample has a porous structure that agglomerates in large particles of several micrometers (Figure 5(a)). EDX spectrum exhibits the existense of C, N, and Ti elements (Figure 5(b)). Figure 5(c)5(f) shows that C, N, O, and Ti are evenly dispersed in the material.

The morphology of TiO2, g-C3N4, and (2/8)TiO2/g-C3N4 is shown in Figure 6. Figure 6(a) shows the layered structure of g-C3N4 with a nanometer stacking that is typical for the C3N4 graphite layers. The morphology of TiO2 consists of titania fibers with a 2 × 100 nm diameter and irregular sheets around 50–100 nm in diameters. Composite TiO2/g-C3N4 (Figure 6(c)) is formed through embroidering the titania fiber and the g-C3N4 layers in a nanoscale, suggesting the formation of the heterojunction interface in the TiO2/g-C3N4 composite.

3.1. Visible-Light-Driven Photocatalyst for Methylene Blue (MB) Decomposition
3.1.1. Kinetics of MB Decolorization

Figure 7(a) displays the photocatalytic activity of the g-C3N4, TiO2, and TiO2/g-C3N4 composites. The experiment was conducted in two stages: adsorption in the dark and photocatalytic decomposition of MB. The adsorption lasted for 40 min to ensure the adsorption/desorption equilibrium, followed by visible-light irradiation for 40 min. It is obvious that the pristine g-C3N4 and TiO2 possess either low adsorption capacity or poor decolorization, indicating that these materials do not catalyze MB decomposition. The color of the MB solution remains almost unchanged after 40 min of irradiation, suggesting MB is stable under these conditions. The TiO2/g-C3N4 composites substantially enhance MB decolorization compared with the components. The MB decolorization rate increases with the increasing amount of g-C3N4 and peaks for the (2/8)TiO2/g-C3N4 sample. Therefore, this sample was selected for further experiments.

The leaching of active metal species to the liquid phase was studied by removing the catalyst after 50 min of illumination, and the MB decolorization was studied. It was found that, although illumination continued for up to 120 min (Figures 7(b) and 7(c)), the MB decolorization did not occur in the absence of the catalyst. Only a small amount of leaching titanium was also detected in the supernatant (0.336 mg·L−1) by ICP analysis. This fact confirms that (2/8)TiO2/g-C3N4 is a heterogeneous catalyst in MB photocatalytic degradation.

The kinetics of MB decomposition is shown in Figure 7(d). The process was conducted in two stages, as described above. The (2/8)TiO2/g-C3N4 sample exhibits adsorption equilibrium after around 40 min with an equilibrium adsorption efficiency of ∼8%–10%, depending on the initial concentration. The higher the MB concentration, the larger the adsorption capacity is because of the increasing driving force. The kinetics of photocatalytic decomposition is widely described by using the Langmuir and Hinshelwood (L–H) mechanism [2931]. In this case, MB adsorption on the catalyst surface follows the Langmuir isotherm, and the adsorption equilibrium maintains during the photocatalytic reaction, that is, the adsorption rate is larger than the reaction rate of the electrons or holes, which is the rate-determining step. The reactions and kinetics equation are expressed as follows:where kf and kb are the forward and backward adsorption rate coefficients; kr is the rate constant of the photocatalytic reaction; C0e (ppm) is the MB concentration at the equilibrium of dark adsorption; and kapp (min−1) is the apparent rate constant.

The value of kapp can be obtained from the linear plot of vs. t. It was found that the fitted lines of against irradiation time exhibit high linearity (R2 = 0.9679), and these values indicate that the MB photodegradation apparently fits well with pseudo-first-order kinetics [32]. The pseudo rate constant at different MB concentrations is listed in Table 2. A comparison of kapp with those of previous reports is also presented. It was found that the rate constant of MB decolorization in this paper is a bit higher than those of previous reports, confirming the advantage of synthesizing TiO2-based materials from water-soluble titanium complexes.

3.1.2. Effect of pH and Some Scavengers on MB Decomposition

The pH effect of dark adsorption and photocatalytic decomposition of MB is shown in Figure 8(a). When pH > 3.8 (pKa of MB 3.8 [40]), MB is positively charged. The pHPZC value of TiO2/g-C3N4 determined with the pH shift method is 7.4. At pH less than pHPZC, the material’s surface carries a positive charge because of protonation. The MB degradation efficiency increases gradually with pH from 4 to 10. When pH > 7.4, the electrostatic interaction between the material’s negatively charged surface and positively charged MB causes decolorization efficiency to increase.

The results of free radical scavenging experiments are shown in Figure 8(b). The experiments were conducted to determine the free radicals formed during irradiation, causing MB decomposition. The three free radical scavengers used in the research are based on Tu et al. [41]: potassium iodide to capture h+, isopropanol to capture OH, and benzoquinone to capture O2•−. The results indicate that benzoquinone significantly reduces MB decolorization, followed by isopropanol and potassium iodide, indicating that free radicals play an essential role in the MB degradation in the following order: O2•− > OH > h+.

3.1.3. Photocatalytic Degradation Pathway of MB

The UV–vis absorption spectrum of MB solutions is shown in Figure 8(c). The absorption peaks at 245 and 292 nm correspond to the excitation of π-electrons from the π interaction in the benzene ring, while those at 615 and 664 nm are assigned to the benzene ring and the heteropolyaromatic linkage [42]. The maximum absorption occurs at 664 nm. The intensity of the adsorption peaks decreases with increasing illumination time, and the color of the MB solution almost disappears after 100 min of illumination. This decolorization reveals that the MB structure was destroyed during illumination. The mineralization of carbons is also evidenced by the COD values, as shown in Figure 8(d). COD gradually decreases from an initial 271 to 11.1 mg·L−1 after 100 min of irradiation and 8.3 mg·L−1 after 120 min. This decolorization is associated with the aromatic ring opening with the transient formation of carboxylic acids, followed by the evolution of CO2 according to the “photo-Kolbe” reaction: R–COO + h+ ⟶ R + CO2. Methylene blue is broken down into harmless or less harmful substances, such as carbon dioxide (CO2) and water (H2O), via certain intermediates.

The degradation of MB was also studied quantitatively by using HPLC before and after degradation. The peak area reveals that an insignificant amount of MB remains after the 100 min illumination (only 3.7% remains (Figure 9)).

The intermediate compounds were identified by using liquid chromatography–mass spectrometry (LC–MS), as shown in Figure 10. Based on the mass/charge ratio (m/z) in the mass spectra, we proposed the compounds and their structural formulas (Scheme 2). The signal at m/z of 284 presents before degradation is assigned to MB. Then, the cleavage of each methyl group on amine groups, in turn, leads to the formation of Azure B, Azure A, and thionin, with m/z of 268, 252, and 225, respectively (Figure 10(a)10(e)). These products result from the reduction of the methyl group in the MB molecule via photocatalytic degradation [35, 43]. The peak at 1.8 min (Figure 10(f)) corresponding to m/z 192 is attributed to HO–C6H3–SO3H(NH3+). Accordingly, the decomposition is due to the cleavage of the bonds in the C–S+=C and C–N=C functional groups. The C–S+=C group is attacked by HO radicals and oxidized to the C–S(=O)–C sulfoxide group, causing the formation of a central imino group (–NH) at the para-position of the middle aromatic ring.

The H atoms for the formation of C–H and N–H bonds may result from the deprotonation by photogenerated electrons to form H radicals [44]. The saturation of the two amino groups by H radicals produces substituted aniline. Then, the amino group can be replaced by an OH radical, forming the corresponding phenol and releasing an NH2 radical producing NH3 and ammonium ions. The sulfoxide group can be attacked by HO radicals for the second time to produce sulfone. At the same time, the HO radicals cause ring cleavage; then, the sulfone group can be further attacked by HO radicals to produce sulfonic acid. The peak at 2.1 min (Figure 10(g)) corresponding to m/z of 164 is assigned to (HO)2C6H3–N(CH3)–CHO. Accordingly, the first step is due to the cleavage of the bonds in C–S+=C and C–N=C in Azure B, which are then attacked by HO radicals to form NH2 and NH2 groups. They are further attacked by HO radicals and form –OH groups in phenol, and the –CH3 group in N(CH3)2 forms the –CHO group. Thus, the obtained structure is consistent with the structure of the fragment with m/z of 167, as proposed by Houas et al. [44]. The peak at 2.7 min (Figure 10(h)) corresponding to m/z of 137 has the proposed structure as (HO)2C6H3–NH–CH3. Accordingly, the –CHO group can be attacked by HO radicals to form the –COOH group. Then, the N–COOH group is attacked by h+ to form N, and it combines with H+ to form the –NH– group.

3.1.4. Mechanism of MB decomposition

The valence band (VB) and CB potential edges of g-C3N4 and TiO2 were calculated according to the equations proposed by Xu and Schoonen [45] and Lin et al. [46] as follows:where ECB is the CB edge energy, EVB is the VB edge energy, χ is the electronegativity (χ = 4.73 eV for g-C3N4 and χ = 5.81 eV for TiO2 [45]), Ee is the free energy of electrons with respect to the NHE (Ee = 4.5 eV [47]), and Eg is the bandgap energy of the semiconductor.

The values of ECB and EVB for g-C3N4, calculated from Equations (7) and (8), are –1.11 and 1.55 eV, and those for TiO2 are –0.29 and 2.91 eV, respectively. The position of the edge energy of g-C3N4 and TiO2 is shown in Scheme 3. First, MB is adsorbed by TiO2/g-C3N4, and its aromatic rings interact with g-C3N4 or TiO2 on the surface via ππ bonding [48]. Under visible-light illumination, the MB molecule can be excited, and the photoelectrons in HOMO of MB (–0.82 eV) migrate to LUMO of MB (–1.04 eV) (Equation (9)) [49]. These photoelectrons could transfer to the CB of TiO2. The VB electrons of g-C3N4 would be excited, and the photoelectron–hole pairs are, thus, generated. These photoelectrons have a chance to transfer to the CB of TiO2, which enhances the photoelectron–hole separation (Equation (10)). The electrons and holes could react with oxygen and water to form free radicals (Equations (11) and (12)). They, in turn, can decompose MB to different products with decreasing molecular mass under prolonged light illumination. MB can be completely mineralized to inorganic substances (CO2, H2O, Cl, NO3, SO42−, and NH4+), similar to what was reported by Houas et al. [44].

The following photocatalytic degradation mechanism for MB decomposition on the TiO2/g-C3N4 catalyst is illustrated as follows:

3.2. Recyclability of TiO2/g-C3N4

The catalyst’s reuse after a catalytic reaction is a critical requirement besides its high catalytic activity. The used catalyst was regenerated by calcining in the nitrogen atmosphere at 500°C for 3 hr. The SEM observation of the catalyst after the dark adsorption following the dye degradation is shown in Figures 11(a) and 11(b). As seen from the figure, the particles of the catalyst present the large agglomerates resulted in the combination of fine particles and dyes molecules. After the third recycle, fine particles are separated clearly indicating that the MB dyes are removed from the surface of (2/8)TiO2/g-C3N4 due to the photocatalytical degradation of MB. The MB decomposition efficiency after three recycling is found as 98.5%, 93.0%, 87.5%, and 82.5%. It is obvious that the efficiency decreases gradually after each recycling. It drops only around 6% after three recyclings, compared with the first reaction. The XRD patterns of the catalyst remain unchanged, indicating that the catalyst is stable and promising for treating dyes wastewaters (Figure 11(c)).

3.3. Photocatalytic Activity TiO2/g-C3N4 for Other Dyes

To confirm the photochemical degradation ability of TiO2/g-C3N4 materials for other dyes under visible illumination, we also conducted other experiments with malachite green (MG), methyl blue (MyB), and methyl red (MR). The assessment of color degradation was based on the change in the main absorption peak’s intensity at λmax of 617 nm for MG, 607 nm for MyB, and 521 nm for MR. Figure 11(d) shows that the TiO2/g-C3N4 material exhibits high catalytic ability in the decomposition of these organic dyes, especially the cationic ones. The photochemical degradation efficiency of MG with the cationic configuration is 100% after 80 min of illumination. Methyl blue with the anionic configuration decomposes by 97.8% after 100 min, and MR with the neutral configuration decomposes by 65.9% after 120 min. This result shows that TiO2/g-C3N4 is possibly a photocatalyst working in the visible light region to decompose organic dyes in aqueous media.

4. Conclusion

A novel and facile approach to the in situ synthesis of TiO2/g-C3N4 nanoparticles with heterostructures and enhanced photocatalytic activity was demonstrated. The composites were synthesized from a homogeneous mixture of melamine and a water-soluble titanium (IV) complex, which initiates an intimate interfacial contact between TiO2 nanoparticles and g-C3N4 nanosheets with a relatively large surface area. The obtained catalyst exhibits excellent catalytic activity toward methylene blue decomposition. The dye was degraded via a free radical mechanism in which the radical activity followed the order O2•− > OH > h+. First, methylene blue was degraded by opening its central aromatic rings, then the oxidation of their subsequent metabolites, and finally, the evolution of CO2. In addition, TiO2/g-C3N4 also exhibited efficient photocatalytic degradation toward other dyes, such as malachite green, methyl red, and methyl blue. The TiO2/g-C3N4 material is a promising catalyst for treating diluted wastewater in textile industries.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Funding

This work was financially supported by Hue University, Vietnam, under project code DHH 2021-04-151 and Van Lang University, Vietnam.

Acknowledgments

The authors (Nguyen Thi Thanh Tu and Dang Thi Ngoc Hoa) thank the financial support from Hue University and Van Lang University.