N-TiO 2- δ /g-C 3 N 4 Dual Photocatalysts for Efficient Oxytetracycline Hydrochloride Photodegradation and CO 2 Photoreduction

Abstract

A series of $x\%$ (wt) N-TiO_2-δ /g-C₃N₄ composites was synthesized by calcination and hydrothermal methods (labeled $x$ TiCN, $x$ : 5, 10, and 15). All composites were characterized by X-ray diffraction, Fourier transform infrared spectroscopy, UV-vis diffuse reflectance spectroscopy, transmission electron microscopy, and X-ray photoelectron spectroscopy. The photocatalytic activity of these composites was evaluated through oxytetracycline hydrochloride (denoted as OTC) photodegradation and CO₂ photoreduction. The $x$ TiCN composites exhibited higher OTC photodegradation than bulk g-C₃N₄. 10TiCN was slightly more active than 5TiCN and 15TiCN, with a photodegradation yield of 97% after 5 h of light irradiation and constant rate of 0.647 h^-1. For CO₂ photoreduction, it was observed that 5TiCN exhibited the highest activity among the synthesized composites, with 7.0 ppm CH₄ formed. This CH₄ concentration was 7.8 times higher than the concentration formed by bulk g-C₃N₄ (0.9 ppm). A $Z$ -scheme mechanism was proposed to explain the enhanced photocatalysis by $x\%$ (wt) N-TiO_2-δ /g-C₃N₄ composites. The $Z$ -scheme structure increased redox ability, caused better separation of photogenerated electron-hole pairs, and broadened the light absorption zone of the photocatalysts.

1. Introduction

Antibiotics are widely used to control bacterial infections in medical, agricultural, and veterinary sectors [1]. Oxytetracycline hydrochloride (OTC) is a commonly used tetracycline antibiotic. Large amounts of OTC have been released into the environment due to its extensive use [2, 3]. It has a long half-life due to its naphthacene core, persisting in the environment for long periods of time. Its environmental residue, especially in water sources, being carcinogenic, and causing antibiotic resistance in bacteria, is dangerous for human health and the ecosystem [3, 4]. Along with antibiotic water pollution, air pollution has also increased in recent times. With rapid worldwide industrial development, large amounts of CO₂ are released into the atmosphere, causing the greenhouse effect and global warming [5, 6]. Climate change, due to pollution, has caused severe meteorological phenomena, such as typhoons and floods. Thus, CO₂ emission reduction and its environmental remediation are an urgent necessity.

Among different ways to solve these above pollution problems, photocatalysis is an attracted one in recent times [7–12]. The photocatalysis can oxidize OTC to nontoxic compounds [13] and reduce CO₂ to useful chemical compounds, such as CH₄ and CH₃OH [14], reducing water and air pollution. It is a simple and green process, requiring only light irradiation for catalyst activation. The TiO₂ photocatalyst is commonly used because of its high photocatalytic efficiency, good stability, and nontoxicity [15, 16]. However, it is only activated under 380 nm light irradiation. UV light intensity in the sunlight spectrum is approximately 5%; therefore, the TiO₂ photocatalyst requires a UV light source, increasing the process cost. To overcome this drawback, numerous studies report reducing the TiO₂ bandgap, activating TiO₂ at longer wavelengths (such as visible light), or combining TiO₂ with narrow bandgap semiconductors. In the first strategy, the TiO₂ oxide is doped with nonmetallic elements (N, F) or transition metals (Fe, Co) [17–22]. In the second method, TiO₂ is combined with other semiconductors, such as CuO, BiOBr, metal Au, Pt, or both [23–29]. When combining with an oxide-owned narrower bandgap energy, the new composite will be activated by a longer light wavelength, such visible light, and prevent electron-hole pair recombination. With metal deposition, the metallic phase conduction band attracts free electrons, reducing the recombination of electron-hole pairs.

Recently, graphitic carbon nitride (g-C₃N₄), a nonmetallic photocatalyst, has attracted immense interest [30–36]. Synthesized by facile and cost-effective methods, it exhibits high chemical and thermal stability. In particular, it exhibits potential for antibiotic-photocatalytic degradation in aqueous solutions and CO₂ photoreduction [37, 38]. A low bandgap of 2.7 eV enables their visible-light activation. However, rapid recombination of photogenerated electron-hole pairs is a limitation of the g-C₃N₄ photocatalyst. The combining of g-C₃N₄ with other semiconductors to prevent this recombination is an interesting strategy [39].

Among the aforementioned photocatalyst improvement strategies, the construction of a heterojunction between two semiconductors is a promising method. In this structure, the photogenerated electron-hole pairs are separated into different zones, thus preventing recombination and enhancing photocatalytic activity. Further, the carrier transfer behavior has been altered by several photocatalytic mechanisms such as type-I, type-II, and, quite recently, the $Z$ -scheme [40]. Among these mechanisms, the $Z$ -scheme has garnered considerable attention as it not only exhibits electron-hole pair separation in two different zones but also enhances the redox properties. For instance, Guan et al. demonstrated that the activity of 20% LaFeO₃/BiOBr was ~21.0 and~1.3 times that of bare LaFeO₃ and BiOBr, respectively [41]. Additionally, in 2022, Cheng’s group also revealed that the optimal photocatalyst 20% Bi₄Ti₃O₁₂/CdS exhibited an activity 1.6 and 3.3 times that of bare CdS and Bi₄Ti₃O₁₂, respectively [42]. An improved photocatalytic activity was also observed on the Bi₂O₂CO₃ photocatalyst [43]. The enhanced composite photoactivity was explained by improved photoexcited carrier separation in the composite. Therefore, with a similar targeted photocatalyst structure, TiO₂- and g-C₃N₄-based composites were also developed, which exhibited positive results. Wang et al. demonstrated that the tetracycline (TCL) photodegradation on TiO₂@g-C₃N₄ was 75% and 12% greater than that on TiO₂ and g-C₃N₄, respectively [44]. Similarly, for CO₂ photoreduction, Reli et al. showed a twofold increase in CH₄ formation on TiO₂/g-C₃N₄ (0.3/1) as compared to g-C₃N₄ [45]. These studies indicate that along with the nature of coupled semiconductors, the morphology and interface interaction between two semiconductors play an important role in photocatalytic improvement.

In view of the above, this study demonstrates the coupling of TiO₂ and g-C₃N₄ to obtain heterojunction photocatalysts, $x\%$ TiO₂/g-C₃N₄. On the one hand, TiO₂ has the disadvantage of a large bandgap energy, $E_g=3.2\text{eV}$ ; on the other hand, g-C₃N₄, despite having a lower bandgap energy, $E_g=2.7\text{eV}$ , demonstrates a fast recombination of electron-hole pairs. Therefore, the formation of a heterojunction composite through semiconductor coupling could allow overcoming the disadvantages of each constituent. Additionally, the CB/VB potential position of TiO₂ (-0.17 V/+3.0 V) and g-C₃N₄ (-1.15 V/+1.5 V) [44] would allow the establishment of a staggered band structure and the formation of a $Z$ -scheme photocatalyst. The hydrothermal method, prevalently employed for tuning the morphology of synthesized compounds, was used to prepare the composite. Subsequently, calcination was conducted to enhance the interaction between TiO₂ and g-C₃N₄ such as an in situ nitrogen doping into TiO₂. The photocatalytic activity was estimated through OTC oxidation in the liquid phase. OTC is the most stable compound in the TCL group, and as per our knowledge, no studies on OTC oxidation were reported using this composite type. To further evaluate the synthesized photocatalyst composites, CO₂ photoreduction was also investigated in the gas phase.

2. Materials and Methods

2.1. Materials

Chemical compounds, melamine (C₆H₆N₆) (Sigma-Aldrich), C₁₆H₃₆O₄Ti (Sigma-Aldrich), CH₃COOH (China), C₂H₅OH (China), TiO₂ (Evonik P25), and oxytetracycline hydrochloride (OTC) (Sigma-Aldrich), of analytical purity were used as obtained.

2.2. Synthesis of g-C₃N₄ and N-TiO_2-δ /g-C₃N₄ Composites

2.2.1. Synthesis of g-C₃N₄ and N-TiO₂

Melamine calcination at 550°C, for 3 h, in nitrogen gas medium, was used to synthesize g-C₃N₄.

The nitrogen-doped TiO₂ (N-TiO₂) synthesis was inspired by the work of Viswanath et al. [46]. In this typical procedure, titanium (IV) butoxide in ethanol solution and melamine in hot water-ethanol (1 : 3 volume ratio) solution were mixed, then stirred for 24 hours and aged for 5 days. The obtained gel was dried and calcined at 400°C for 3 hours.

2.2.2. Synthesis of N-TiO_2-δ /g-C₃N₄

Calculated amounts of g-C₃N₄, titanium (IV) butoxide, and acetic acid (1 : 30 in volume) were mixed in a 250 ml beaker, with 15 min magnetic stirring, and autoclaved at 140°C for 12 h. The resulting solution was centrifuged and rinsed several times with ethanol. The solid obtained was oven dried at 60°C for 24 h and calcined at 400°C for 3 h, under nitrogen gas. The $x\%$ (wt) N-TiO_2-δ /g-C₃N₄ composites were denoted as $x$ TiCN ( $x$ : 5, 10, and 15).

2.3. Photocatalytic Procedure

2.3.1. OTC Photodegradation

Into a beaker containing 100 ml 10 ppm OTC solution (C₀), g-C₃N₄ or $x$ TiCN composite (0.1 g) was added with stirring (Figure 1(a)) and left in the dark for 1 h to attain adsorption equilibrium. The mixture was then illuminated with a 200 W LED lamp. Every 1 hour, 5 ml sample was taken, filtered, and analyzed by a UV-vis spectrophotometer to detect OTC content ( $C_t$ ) of the reaction mixture. The photostability test was performed 5 times. The separation of the catalyst and the reaction solution was carried out by centrifugation (6000 rpm).

Figure 1

Scheme of the photocatalytic reactor used for OTC oxidation (a) and CO₂ reduction (b).

(b)

2.3.2. CO₂ Photoreduction

Into a glass beaker (5 cm diameter) containing 15 ml deionized water, g-C₃N₄ or $x$ TiCN composite (0.1 g) was added. After 15 min stirring, the mixture was evaporated in an oven, at 70°C, to obtain a well-dried, homogeneously dispersed powder. The catalyst-containing beaker was placed in a handmade closed stainless steel reactor (169 cm³ volume), equipped with a 6 cm diameter quartz window, and purged with high-purity (99.999%) 500 ml/min CO₂ flow for 30 min. The reactor was illuminated with a 150 W Xenon lamp (Newport model 67005) for 18 h. Gaseous products were analyzed using a gas valve system connected to a gas chromatograph, equipped with a thermal conductivity detector and flame ionization detector (TCD-FID) (Scion 456) (Figure 1(b)).

2.4. Characterizations

X-ray diffraction (XRD) (Bruker D8), Fourier-transform infrared spectroscopy (FTIR) (8101M Shimadzu), N₂ adsorption-desorption (TriStar 3000-Micromeritics), UV differential reflectance spectroscopy (UV-DRS) (Jaco V-530), transmission electronic microscopy (TEM, JEM 1400 – Plus Jeon), and X-ray photoelectron spectroscopy (XPS) (Thermo Scientific MultiLab 2000) were used to characterize g-C₃N₄ and $x$ TiCN composites.

3. Results and Discussion

3.1. Structural Characterization

XRD patterns of $x$ TiCN composites are shown in Figure 2(a). The formation of the g-C₃N₄ crystalline phase, with a characteristic peak at 27.8°, was observed in all XRD patterns, confirming the g-C₃N₄ structure after composite synthesis. Characteristic peaks at 25.5°, 38.0°, 48.1°, 54.2°, and 62.8°, corresponding to the TiO₂ anatase phase, were observed in the XRD patterns of all composites. In the N-TiO₂ XRD patterns, a small quantity of the rutile phase was recognized at 2 theta of 27.5 and 36.1°. To identify different components, FTIR characterization was carried out, and the spectra are shown in Figure 2(b). In the g-C₃N₄ spectrum, the peak at 802 cm^-1 was ascribed to the s-triazine bending mode [30]. The peaks at 1228 and 1311 cm^-1 were attributed to the C–N stretching vibrations [47, 48]. The peaks at 1392, 1535, and 1625 cm^-1 originated from the –C=N stretching vibrations in aromatic rings. The broad band at 3000–3500 cm^-1 corresponded to the stretching vibrations of the absorbed water hydroxyl group (–O–H) and terminal amino groups (–NH₂) [47, 48]. Finally, the broad peak at 534 cm^-1 was attributed to the Ti–O bond vibrations [49].

Figure 2

XRD patterns (a) and FTIR spectra (b) of g-C₃N₄ and $x$ TiCN composites.

(b)

The light absorption abilities of the composites were analyzed by UV-vis DR spectra. Figure 3(a) shows the UV-vis DR spectra of g-C₃N₄ and $x$ TiCN. The bandgap energies, calculated using the Kubelka–Munk function (results are shown in Figure 3(b)), were 2.58, 2.60, 2.63, 2.66, and 2.90 eV for g-C₃N₄, 5TiCN, 10TiCN,15TiCN, and N-TiO₂, respectively. Increasing the TiO₂ content broadened the composite bandgap energy owing to a greater contribution of the large band gap energy by TiO₂ (3.2 eV) as compared to g-C₃N₄ (2.58 eV). However, generally, all photocatalyst composites could be activated by visible light. The corresponding differential curves of UV-vis DR spectra are displayed in Figure 3(c). The absorption edges (λ _abs) of pristine TiO₂ (P25) and g-C₃N₄ are at wavelengths of 396.3 and 437.5 nm, respectively. In contrast, the $x$ TiCN composites exhibit a slight shift of the g-C₃N₄ adsorption edge peak to a shorter wavelength, while that of TiO₂ slightly shifts to a longer wavelength. The longer-wavelength-shifted adsorption edge of the TiO₂ constituent is possibly due to nitrogen doping in TiO₂ during the composite synthesis. These observations suggest an interaction between g-C₃N₄ and TiO₂, forming in situ-doped N-TiO₂. In the case of 5TiCN, the observed adsorption edge peaks of TiO₂ andg-C₃N₄ correspond to wavelengths of 402.0 and 435.3 nm, respectively. Based on the relationship $E_g=1240/{\lambda }_{\text{abs}}$ , the calculated bandgap energies ( $E_g$ ) of TiO₂ and g-C₃N₄ are 3.08 and 2.85 eV, respectively. For the N-TiO₂ sample prepared by the hydrothermal method, the differential curve of the UV-vis DR spectrum exhibited a single weak and broad peak at 392.5 nm (or $E_g=3.16\text{eV}$ ). This peak is attributed to the adsorption edge of TiO₂.

Figure 3

UV-vis DR spectra (a), energy bandgap determination by Kubelka–Munk function (b), and its corresponding differential curves (c) of g-C₃N₄, N-TiO₂, and $x$ TiCN composites; N₂ adsorption-desorption isotherms and corresponding pore distribution curves of g-C₃N₄ (solid line) and 5TiCN composite (dash line) (d).

(b) (c) (d)

The g-C₃N₄ and 5TiCN composites were selected to characterize its specific surface area, one of the important properties of heterogeneous catalysis. The results are presented in Figure 3(d). It is observed that two samples show the type-4 isotherms with H3 hysteresis loop, which indicate the presence of a mesopore. This is suitable with the obtained pore distribution curves. The BET surface areas are 18 and 20 m²/g for g-C₃N₄ and 5TiCN, respectively. Hence, the composite preparation did not seem to change the g-C₃N₄ structure.

The interaction between the TiO₂ and g-C₃N₄ phases is better understood from the TEM images presented in Figure 4. It is observed that the TiO₂ particles (dark areas in Figures 4(a) and 4(b)) were formed in various sizes and shapes. Some TiO₂ particles were deposited on the g-C₃N₄ layer (e.g., position of cycle 1), while the others were covered by g-C₃N₄ multisheets, forming a core-shell structure (e.g., position of cycle 2), thereby enhancing the interaction surface between the TiO₂ and g-C₃N₄ phases.

Figure 4

TEM images of 5TiCN.

The elemental composition and oxidation state of the catalyst influence catalytic performance. Therefore, the 5TiCN composite was characterized using XPS, as shown in Figure 5.

Figure 5

X-ray photoelectron spectra of 5TiCN: (a) survey spectra; (b) high-resolution spectra of Ti2p, (c) high-resolution spectra of O1s; (d) high-resolution spectra of N1s; (e) high-resolution spectra of C1s.

(b) (c) (d) (e)

Three peaks at binding energies of 458.5, 459.1, and 464.3 eV, corresponding to orbitals Ti⁴⁺2p_3/2, Ti³⁺2p_1/2, and Ti⁴⁺2p_1/2, respectively, were observed in the Ti2p high-resolution spectrum [50, 51]. As reported by Jia et al., binding energies of two prominent peaks (at 458.5 and 464.3 eV) exhibited 0.2 eV shifts compared to spectral peaks of pure TiO₂. This could be due to the substitution of O^2- by N^3-, leading to the formation of N-Ti-O bonds [52]. This indicated N doping of TiO₂ during composite synthesis, decreasing the TiO₂ bandgap energy, leading to visible-light activation [17, 18]. The small peak of Ti³⁺2p_1/2 confirmed oxygen vacancies in TiO₂ [51]. Hence, TiO_2-δ is a more accurate molecular formula than TiO₂.

Deconvolution peaks for the O1s spectrum exhibited two peaks at 531 and 532 eV, assigned to O-H of absorbed water and the Ti–O bond [53], respectively. N1s spectrum exhibited three peaks at 398.6, 399.7, and 400.8 eV, ascribed to the sp²C of C-N=C, tertiary N of N-(C)₃ group, and N-C=N bonds, respectively. The C1s spectrum exhibited three peaks at 284.9, 286.3, and 288.2 eV, corresponding to C-C, C-NH₂, and N-C=N bonds, respectively [53, 54].

3.2. Evaluation of Photocatalytic Activity

3.2.1. Photooxidation of OTC

Before performing the photocatalytic test, the adsorption equilibria were carried out (Figure 6(a)). The results demonstrated that all composites reached rapidly the adsorption equilibrium after only about 15 minutes, while 60 minutes was required for N-TiO₂. The calculation indicated that the equilibrium adsorption quantities of OTC at 60 min were 11%, 14%, 17%, 12%, and 88% for g-C₃N₄, 5TiCN, 10TiCN, 15TiCN, and N-TiO₂, respectively. It remarked that there was a strong adsorption phenomenon of OTC on N-TiO₂.

Figure 6

OTC adsorption equilibrium (a) on $x$ TiCN and UV-vis spectra of OTC during test time on 5TiCN composite (b).

(b)

The photocatalytic activity of the catalysts was investigated through OTC photodegradation and CO₂ photoreduction. Figure 6(b) shows the photodegradation of OTC and UV-vis spectra of OTC solutions during test time, using 5TiCN. OTC concentrations were calculated from the absorbance intensity of UV-vis spectra at the 357 nm wavelength.

Figure 7 shows OTC photodegradation efficiency and kinetics. All composite photocatalysts exhibited excellent OTC degradation activity, with yields of 93%, 97%, and 92% for 5TiCN, 10TiCN, and 15TiCN, respectively. For N-TiO₂, the adsorption phenomenon dominated, reaching 90% adsorbed OTC quantity after 1 hour of equilibrium, and the efficiency of OTC removal increased only ~4% when turning on the light for 5 hours. Hence, the photocatalytic reaction on N-TiO₂ was negligible. This behavior was possibly due to the formation of the melon structure formed during the synthesis, besides the process of nitrogen doping on TiO₂ [55]. The melon structure, which is not a semiconductor, could cover TiO₂/N-TiO₂ particles, preventing the photocatalytic process. A blank test (without catalyst) was also carried out for comparison. In the blank test, the OTC concentration decreased about 9% by photolysis. Hence, after 5 h light irradiation, the 10TiCN composite was slightly more active than those of 5TiCN and 15TiCN. It is noted that the adsorption phenomenon contributes a small part in the conversion calculation, only 14% in the case of 5TiCN as mentioned in the above adsorption equilibrium study. For these OTC conversion reactions, reaction kinetics were described by the following equation [56]: $\begin{array}{}\text{(1)}& \ln \left(\frac{C_0}{C_t}\right)=kt+\ln \left(\frac{C_0}{C_0^{\prime }}\right),\end{array}$ where $C_0$ , $C_0^{\prime }$ , and $C_t$ are the initial, equilibrium, and time $t$ concentrations of OTC during the test, respectively. Rate constants ( $k$ ) were 0.389, 0.457, 0.647, and 0.451 h^-1 for g-C₃N₄, 5TiCN, 10TiCN, and 15TiCN, respectively (from fitted lines in Figure 7(b)). The reaction kinetic on N-TiO₂ was not investigated as the catalytic activity was negligible with respect to the adsorption phenomenon. Based on rate constant values, composite activities followed the following order: 10TiCN > 5TiCN ≈ 15TiCN > g-C₃N₄. The rate constant of 10TiCN was approximately 1.7 times that of g-C₃N₄.

Figure 7

(a) Photodegradation efficiency of OTC on $x$ TiCN composites versus irradiation time and (b) kinetic of OTC degradation reaction on $x$ TiCN composites.

(b)

There are no reports of OTC photodegradation on TiO₂-g-C₃N₄ composites, while there are very few reports on tetracycline (TCL) photodegradation using such composites. Wang’s research group reported 20 ppm TCL solution photodegradation using composites, such as TiO₂@g-C₃N₄ heterojunction, TiO₂@g-C₃N₄ core-shell quantum heterojunction, and TiO_2-x/ultrathin g-C₃N₄/TiO_2-x direct $Z$ -scheme heterojunction [44, 57, 58]. TCL photodegradation rate on the TiO₂@g-C₃N₄ core-shell quantum heterojunction composite was 2 and 2.3 times higher than those on TiO₂ and bulk g-C₃N₄, respectively. Photoactivity of TiO_2-x/ultrathin g-C₃N₄/TiO_2-x direct $Z$ -scheme heterojunction system was 20.1 and 1.3 times higher than those of TiO₂ and g-C₃N₄, respectively. Wang et al. reported a 3 times higher TCL degradation rate on N-TiO₂/O-doped N-vacancy g-C₃N₄ than on N-vacancy g-C₃N₄ [59]. Rao et al. reported a 4.4 times higher rate constant of 10 ppm TCL solution photodegradation using a hierarchical structure of anatase-rutile TiO₂/g-C₃N₄ (ARC), compared to g-C₃N₄ [60].

In general, the TiO₂-g-C₃N₄ heterojunction structure improved photoactivity remarkably compared to those of bulk TiO₂ and g-C₃N₄. OTC photodegradation yield reached up to 97%, with a 1.7 times higher rate constant with 10TiCN than with g-C₃N₄.

The influence of pH on the photoactivity was evaluated. The researches focused on 5TiCN (Figure 8(a)). After 5 hours of irradiation, the photoactivity at pH 11 and 5 was slightly improved, in comparison with pH 7. In the study of Yu et al., the isoelectric point of g-C₃N₄ was about pH 5 [61]. It means that the g-C₃N₄ surface charge is positive at $\text{pH}<5$ and negative at $\text{pH}>5$ . The OTC has the $\text{pK}{\text{a}}_1=3.22$ , $\text{pK}{\text{a}}_2=7.68$ , and $\text{pK}{\text{a}}_3=8.94$ [62]. Therefore, the OTC exhibited a positive charge at $\text{pH}<3.22$ in the form of H₃TC⁺ and negative one at $\text{pH}>9$ in the main form of OTC^-2. Thus, the charges on the g-C₃N₄ surface and OTC molecules in experimental pH conditions are the same, which prevent adsorption of OTC on the g-C₃N₄ surface by electrostatic repulsion force. That could explain why the pH 3 and pH 7 did not influence strongly on the OTC conversion as observed.

Figure 8

Photodegradation efficiency of OTC in function of pH (a) and photostability test on 5TiCN (b).

(b)

The stability tests were carried out for the 5TiCN composite (Figure 8(b)). The OTC conversion decreased from 93.4% to 89.4% after 5 cycles. This is the promising result in an application view.

3.2.2. Photoreduction of CO₂

To investigate the dual photocatalytic behavior of N-TiO_2-δ /g-C₃N₄ composites, they were used in the photoreduction of CO₂ by H₂O in the gaseous phase (Figure 9). Concentration of CH₄ (the only product detected) was monitored to analyze the photocatalytic behavior of the synthesized composites. 5TiCN exhibited the highest activity, with 7.0 ppm CH₄ concentration, followed by 10TiCN (5.2 ppm CH₄), 15TiCN (4.3 ppm CH₄), and g-C₃N₄ (0.9 ppm CH₄). No product was detected in the N-TiO₂ test. As mentioned in the OTC photooxidation results above, this low photoactivity of N-TiO₂ could be explained by the formed melon structure that covered TiO₂/N-TiO₂ particles and possibly the lower CB potential position of N-TiO₂ than the standard reduction potential of CO₂/CH₄ [55]. The detected CH₄ concentration was quite low over g-C₃N₄. Thus, the test was performed three times, and average value was taken. Hence, the CH₄ concentration was 7.8 times higher for 5TiCN (composite exhibiting maximum photocatalysis) than for pristine g-C₃N₄.

Figure 9

CH₄ concentration on g-C₃N₄ and $x$ TiCN composite.

There are very few reports on photoreduction using a photocatalyst formed by TiO₂ and g-C₃N₄ [45, 63–69]. CO₂ photoreduction in the gaseous phase [45, 63–65] and liquid phase [66–69] has been reported. Gas phase studies, with CH₄ and CO products, indicate better photocatalytic activity for the TiO₂-g-C₃N₄-combined photocatalyst, compared to bulk TiO₂ and g-C₃N₄. Zhou et al. reported a 4 times higher CO formation on g-C₃N₄-N-TiO₂ (14.73 μmol) than on P25 (TiO₂) [63]. In liquid phase photoreductions, besides CH₄ and CO, other oxygenated hydrocarbons (CH₃OH, HCOOH, and CH₃COOH) are formed. TiO₂-g-C₃N₄-combined photocatalysts exhibit also higher photoactivity compared to single-phase TiO₂ or g-C₃N₄. Badiei et al. reported an 11.3 μmol·g^-1·h^-1 CH₃OH formation for g-C₃N₄@TiO₂, which was 5 and 10 times higher than those for g-C₃N₄ and P-25 (TiO₂), respectively [66]. Lu et al. reported a 283.9 μmol·h^-1·g^-1 CO formation for 2D g-C₃N₄/TiO₂, which was 292.2, 6.8, and 5.7 times higher than those for TiO₂, bulk g-C₃N₄, and mechanically mixed TiO₂/g-C₃N₄, respectively [67]. These results cannot be compared due to different experimental conditions. However, all studies indicate higher photoactivity using TiO₂-g-C₃N₄-combined photocatalysts. In this study, CH₄ production on 5TiCN was 7.8 times higher than that on bulk g-C₃N₄.

3.2.3. Photocatalytic Mechanism

An outstanding photocatalytic activity was exhibited by the $x$ TiCN composites during OTC photodegradation and CO₂ photoreduction as compared to the pristine g-C₃N₄ and N-TiO₂. Hence, the coupling mechanism of g-C₃N₄ and TiO₂ is worth investigating. From the UV-vis DR spectrum, the semiconductor conduction/valance edge energies ( $E_{\text{CB}}/E_{\text{VB}}$ ) can be calculated using the electronegativity theory [70]. According to this theory, we have the following empirical formulas: $\begin{array}{}\text{(2)}& E_{\text{CB}}=X-E^e-0.5E_g,E_{\text{VB}}=E_{\text{CB}}+E_g,\end{array}$ where $X$ , $E^e$ , and $E_g$ are the semiconductor electronegativity (5.81 eV for TiO₂ [71] and 4.72 eV for g-C₃N₄ [30]), free electron energy corresponding to the hydrogen scale (4.5 eV), and semiconductor bandgap energy, respectively. As determined above from the differential curve of the UV-vis DR spectrum for 5TiCN, the bandgap energies are 3.08 and 2.85 eV, respectively, for the N-TiO_2-δ and g-C₃N₄ constituents. Using equation (2), $E_{\text{CB}}/E_{\text{VB}}$ is -0.23 V/+2.85 V for N-TiO_2-δ and -1.21 V/+1.64 V for g-C₃N₄ (-0.14/+2.76 for separately synthesized N-TiO₂). With these calculated $E_g$ and $E_{\text{CB}}/E_{\text{VB}}$ values, the energy band diagram before the coupling of the N-TiO_2-δ and g-C₃N₄ phases is presented in Figure 10(a). As thoroughly discussed by Yang [70], when coupling two semiconductors, the built-in electric field formation and photoexcited carrier transfer occur by different routes depending on the semiconductor types, their Fermi level, and CB/VB potential positions. In $x$ TiCN composites, N-TiO₂ and g-C₃N₄ behave as n-type semiconductors [72, 73], and thus, their Fermi level is near the top conduction edge energy. As the N-TiO_2-δ Fermi level is higher than that of g-C₃N₄, spontaneous electron diffusion from N-TiO_2-δ to g-C₃N₄ occurs during their coupling and generates a built-in electric field with positive charge accumulation in the g-C₃N₄ interface zone and a negative one in the N-TiO_2-δ interface zone. This diffusion is gradually suppressed by the built-in electric field itself, and finally, a thermal equilibrium state is established in the N-TiO_2-δ /g-C₃N₄ heterojunction [70, 74]. On irradiating the composite, photogenerated electron (e^-) and hole (h⁺) pairs formed on the CB and VB of each constituent. Subsequently, the built-in electric field promoted the photogenerated electrons on the CB of N-TiO_2-δ to the VB of g-C₃N₄. This transfer is displayed in Figures 10(b) and 10(c) and is well-known as the $Z$ -scheme mechanism. In this system, the charge transfer between two phases caused the separation of photogenerated electrons and holes into different zones—electrons in the CB of g-C₃N₄ and holes in the VB of TiO₂—preventing their recombination or, simply put, prolonging their lifetime. Notably, this type of charge transfer enriched the photogenerated electrons on the negative CB of g-C₃N₄ and photogenerated holes on the positive VB of TiO₂, thereby strengthening the redox property of the composite. Additionally, a low bandgap of g-C₃N₄ broadened the light absorption region and improved the light utilization efficiency. All these factors contributed to the photocatalytic enhancement of N-TiO_2-δ /g-C₃N₄ composites.

Figure 10

Schematic illustrations of the energy band structures of N-TiO_2-δ and g-C₃N₄ before contact (a), N-TiO_2-δ /g-C₃N₄ heterojunction composite for OTC photooxidation mechanism (b), and CO₂ photoreduction mechanism (c).

(b) (c)

Note that the type-II and $Z$ -scheme mechanisms entail a similar photocatalyst composite structure. However, in the type-II mechanism, reduction takes place on the CB of the semiconductor with a less negative CB potential. Hence, for the $x$ TiCN composite, if the type-II mechanism had occurred, the CO₂ would have been reduced on the CB of the N-TiO_2-δ constituent. This is unlikely as the reduction potential of the N-TiO_2-δ constituent (-0.23 V) is less negative than that of the CO₂/CH₄ (-0.24 V) [14]. Nevertheless, the obtained experimental results show a remarkably higher CH₄ content formation than that of bare g-C₃N₄. This evidence suggests that the $x$ TiCN photocatalyst composites entail the $Z$ -scheme mechanism. It is also found that CO₂ photoreduction could not take place on separately synthesized N-TiO₂ as its CD potential position (-0.14 V) is less negative than that of CO₂/CH₄ (-0.24 V).

The photogenerated electron-hole pair transfers are summarized as follows:

$\text{N}‐{\text{TiO}}_{2‐\delta }+h\nu ⟶{\text{e}}^{-}\left(\text{N}‐{\text{TiO}}_{2‐\delta }\left(\text{CB}\right)\right)+{\text{h}}^{+}\left(\text{N}‐{\text{TiO}}_{2‐\delta }\left(\text{VB}\right)\right)$

$\text{g}‐{\text{C}}_3{\text{N}}_4+h\nu ⟶{\text{e}}^{-}\left(\text{g}‐{\text{C}}_3{\text{N}}_4\left(\text{CB}\right)\right)+{\text{h}}^{+}\left(\text{g}‐{\text{C}}_3{\text{N}}_4\left(\text{VB}\right)\right)$

${\text{e}}^{-}\left(\text{N}‐{\text{TiO}}_{2‐\delta }\left(\text{CB}\right)\right)⟶{\text{h}}^{+}\left(\text{g}‐{\text{C}}_3{\text{N}}_4\left(\text{VB}\right)\right)\left(\text{charge}\text{transfer}\right)$

Then, for oxidation of OTC,

${\text{e}}^{-}\left(\text{g}‐{\text{C}}_3{\text{N}}_4\left(\text{CB}\right)\right)+{\text{O}}_2⟶{{}^{\cdot }\text{O}}_2$

${\text{h}}^{+}\left(\text{N}‐{\text{TiO}}_{2‐\delta }\left(\text{VB}\right)\right)+\text{OTC}⟶\text{byproduct}+{\text{CO}}_2+{\text{H}}_2\text{O}$

or for reduction of CO₂,

$8{\text{e}}^{-}\left(\text{g}‐{\text{C}}_3{\text{N}}_4\left(\text{CB}\right)\right)+{\text{CO}}_2+8{\text{H}}^{+}⟶{\text{CH}}_4+2{\text{H}}_2\text{O}$

$8{\text{h}}^{+}\left({\text{TiO}}_{2‐\delta }\left(\text{VB}\right)\right)+8{\text{H}}_2\text{O}⟶8{\text{H}}^{+}+8^{\cdot }\text{OH}$

$\begin{array}{c}{8}^{\cdot }\text{OH}⟶4{\text{H}}_2\text{O}+2{\text{O}}_2\end{array}$

4. Conclusions

Here, $\text{x}\%$ (wt) N-TiO₂- _δ /g-C₃N₄ composites were synthesized, and their textural and structural properties were analyzed by XRD, FTIR, UV-DRS, TEM, and XPS. All composites showed better activity than bulk g-C₃N₄ towards OTC photodegradation and CO₂ photoreduction. OTC photodegradation yields were higher for the composites (93%, 97%, and 92%, on 5TiCN, 10TiCN, and 15TiCN, respectively) than for bulk g-C₃N₄ (86%). Among these catalysts, the 10TiCN showed the highest rate constant of 0.647 h^-1. In CO₂ photoreduction, CH₄ was the only product detected. CH₄ concentrations of 0.9, 7.0, 5.2, and 4.3 ppm were detected using bulk g-C₃N₄, 5TiCN, 10TiCN, and 15TiCN, respectively. CH₄ formation of 5TiCN was 7.8 times higher than that of bulk g-C₃N₄. Enhanced composite photoactivities were attributed to their $Z$ -scheme mechanism. With this structure, charge transfer between N-TiO_2-δ (CB) and g-C₃N₄ (VB) occurred, leading to recombination prevention of photogenerated electron-hole pairs and stronger redox abilities. The interesting obtained result above indicates the promising novel dual photocatalysts. Tuning composite morphology (e.g., specific surface area and porosity) and contact surface area between components (e.g., good dispersion phase and core-shell system) could enhance photoactivity of $x\%$ (wt) N-TiO_2-δ /g-C₃N₄ further.